Arno de Klerk Fischer–Tropsch Refining
Further Reading Arpe, Hans-Jurgen ¨
Kolb, G.
Industrial Organic Chemistry
Fue...
199 downloads
2143 Views
5MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Arno de Klerk Fischer–Tropsch Refining
Further Reading Arpe, Hans-Jurgen ¨
Kolb, G.
Industrial Organic Chemistry
Fuel Processing
5th, completely revised edition
for Fuel Cells
2010
2008
Hardcover
Hardcover
ISBN: 978-3-527-32002-8
ISBN: 978-3-527-31581-9
Wijngaarden, Ruud I./Westerterp, K. Roel/Kronberg, Alexander/Bos, A.N.R. (Eds.)
Deublein, D., Steinhauser, A.
Industrial Catalysis Optimizing Catalysts and Processes
Biogas from Waste and Renewable Resources An Introduction 2011
2011 Hardcover
Hardcover ISBN: 978-3-527-32798-0
ISBN: 978-3-527-31837-7
Olah, G. A., Goeppert, A., Prakash, G. K. S.
Beyond Oil and Gas: The Methanol Economy
H¨aring, H.-W. (Ed.)
Industrial Gases Processing 2008 Hardcover
2010
ISBN: 978-3-527-31685-4
Softcover ISBN: 978-3-527-32422-4
Barbaro, P., Bianchini, C. (Eds.)
Catalysis for Sustainable Energy Production 2009 Hardcover ISBN: 978-3-527-32095-0
Stolten, D. (Ed.)
Hydrogen and Fuel Cells Fundamentals, Technologies and Applications 2010 Hardcover ISBN: 978-3-527-32711-9
Al-Qahtani, Khalid Y./Elkamel, Ali
Planning and Integration of Refinery and Petrochemical Operations 2010 Hardcover ISBN: 978-3-527-32694-5
Elvers, B. (Ed.)
Handbook of Fuels 2008 Hardcover ISBN: 978-3-527-30740-1
Arno de Klerk
Fischer–Tropsch Refining
The Author Arno de Klerk University of Alberta Chemical and Materials Engineering Edmonton, Alberta, T6G 2V4 Canada
All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . 2011 Wiley-VCH Verlag & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany
All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Cover Design Formgeber, Eppelheim Typesetting Laserwords Private Limited, Chennai, India Printing and Binding Printed in Singapore Printed on acid-free paper Print ISBN: 978-3-527-32605-1 ePDF ISBN: 978-3-527-63562-7 oBook ISBN: 978-3-527-63560-3 ePub ISBN: 978-3-527-63561-0
V
To my wife, Ch`erie, who loves and supports me so much.
VII
Contents Preface Part I
XIX Introduction
1
1 1.1 1.2 1.2.1 1.2.2 1.2.3 1.3 1.4 1.4.1 1.4.2 1.4.3 1.5 1.5.1 1.5.2 1.5.3
Fischer–Tropsch Facilities at a Glance 3 Introduction 3 Feed-to-Syngas Conversion 4 Feed Logistics and Feed Preparation 5 Syngas Production 5 Syngas Cleaning and Conditioning 7 Syngas-to-Syncrude Conversion 8 Syncrude-to-Product Conversion 10 Upgrading versus Refining 10 Fuels versus Chemicals 11 Crude Oil Compared to Syncrude 12 Indirect Liquefaction Economics 14 Feed Cost 14 Product Pricing 15 Capital Cost 17 References 19
2 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.3 2.3.1 2.4
Refining and Refineries at a Glance 21 Introduction 21 Conventional Crude Oil 22 Hydrocarbons in Crude Oil 23 Sulfur Compounds in Crude Oil 23 Nitrogen Compounds in Crude Oil 25 Oxygenates in Crude Oil 25 Metals in Crude Oil 26 Physical Properties 27 Products from Crude Oil 28 Boiling Range and Product Quality 29 Evolution of Crude Oil Refineries 31
VIII
Contents
2.4.1 2.4.2 2.4.3 2.4.4 2.4.5 2.4.6
First-Generation Crude Oil Refineries 32 Second-Generation Crude Oil Refineries 33 Third-Generation Crude Oil Refineries 36 Fourth-Generation Crude Oil Refineries 39 Petrochemical Refineries 43 Lubricant Base Oil Refineries 44 References 46 Part II
Production of Fischer–Tropsch Syncrude
49
3 3.1 3.2 3.2.1 3.2.2 3.3 3.3.1 3.3.2 3.3.3 3.3.4 3.3.5 3.4 3.4.1 3.4.2 3.4.3 3.4.4 3.4.5 3.5 3.5.1 3.6 3.6.1 3.7
Synthesis Gas Production, Cleaning, and Conditioning 51 Introduction 51 Raw Materials 51 Natural Gas 51 Solid Carbon Sources 52 Syngas from Natural Gas 53 Natural Gas Cleaning 55 Adiabatic Prereforming 55 Steam Reforming 56 Adiabatic Oxidative Reforming 56 Gas Reforming Comparison 57 Syngas from Solid Carbon Sources 58 Gasification of Heteroatoms 59 Low-Temperature Moving Bed Gasification 60 Medium-Temperature Fluidized Bed Gasification 62 High-Temperature Entrained Flow Gasification 64 Gasification Comparison 66 Syngas Cleaning 66 Acid Gas Removal 67 Syngas Conditioning 69 Water Gas Shift Conversion 69 Air Separation Unit 70 References 71
4 4.1 4.2 4.3 4.3.1 4.3.2 4.3.3 4.4 4.4.1 4.4.2 4.4.3
Fischer–Tropsch Synthesis 73 Introduction 73 Fischer–Tropsch Mechanism 74 Fischer–Tropsch Product Selectivity 77 Probability of Chain Growth 78 Hydrogenation versus Desorption 80 Readsorption Chemistry 81 Selectivity Manipulation in Fischer–Tropsch Synthesis Fischer–Tropsch Catalyst Formulation 81 Fischer–Tropsch Operating Conditions 83 Fischer–Tropsch Reaction Engineering 84
81
Contents
4.5 4.5.1 4.5.2 4.5.3 4.5.4 4.5.5 4.5.6 4.5.7
Fischer–Tropsch Catalyst Deactivation 88 Poisoning by Syngas Contaminants 89 Volatile Metal Carbonyl Formation 90 Metal Carboxylate Formation 91 Mechanical Catalyst Degradation 92 Deactivation of Fe-HTFT Catalysts 93 Deactivation of Fe-LTFT Catalysts 93 Deactivation of Co-LTFT Catalysts 95 References 99
5 5.1 5.2 5.2.1 5.2.2 5.3 5.3.1 5.3.2 5.3.3 5.3.4 5.3.5
Fischer–Tropsch Gas Loop 105 Introduction 105 Gas Loop Configurations 107 Open Gas Loop Design 107 Closed Gas Loop Design 108 Syncrude Cooling and Separation Pressure Separation 110 Cryogenic Separation 110 Oxygenate Partitioning 111 HTFT Syncrude Recovery 113 LTFT Syncrude Recovery 114 References 116 Part III
109
Industrial Fischer–Tropsch Facilities
6 6.1 6.2 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.4.5 6.5
German Fischer–Tropsch Facilities 119 Introduction 119 Synthesis Gas Production 119 Fischer–Tropsch Synthesis 121 Normal-Pressure Synthesis 122 Medium-Pressure Synthesis 125 Gas Loop Design 127 Carbon Efficiency 128 Fischer–Tropsch Refining 128 Refining C3 –C4 Crude LPG 129 Refining Carbon Gasoline 130 Refining of Condensate Oil 132 Refining of Waxes 135 Aqueous Product Refining 136 Discussion of the Refinery Design 137 References 138
7 7.1 7.2
American Hydrocol Facility 141 Introduction 141 Synthesis Gas Production 142
117
IX
X
Contents
7.3 7.4 7.4.1 7.4.2 7.5
Fischer–Tropsch Synthesis 143 Fischer–Tropsch Refining 145 Oil Product Refining 146 Refining Aqueous Product 149 Discussion of the Refinery Design 150 References 151
8 8.1 8.2 8.2.1 8.2.2 8.3 8.3.1 8.3.2 8.3.3 8.4 8.4.1 8.4.2 8.4.3 8.4.4 8.4.5 8.5 8.5.1 8.5.2 8.5.3 8.5.4 8.6
Sasol 1 Facility 153 Introduction 153 Synthesis Gas Production 154 Lurgi Dry Ash Coal Gasification 154 Rectisol Synthesis Gas Cleaning 155 Fischer–Tropsch synthesis 157 Kellogg HTFT synthesis 157 Arge LTFT Synthesis 159 Gas Loop Design 162 Fischer–Tropsch Refining 163 Kellogg HTFT Oil Refining 163 Arge LTFT Oil Refining 165 Aqueous Product Refining 166 Coal Pyrolysis Product Refining 169 Synthetic Fuel Properties 170 Evolution of the Sasol 1 Facility 172 Changes in Synthesis Gas Production 172 Changes in Fischer–Tropsch Synthesis 173 Changes in Fischer–Tropsch Refining 174 Changes in Coal Pyrolysis Product Refining 177 Discussion of the Refinery Design 177 References 179
9 9.1 9.2 9.2.1 9.2.2 9.3 9.3.1 9.4 9.4.1 9.4.2 9.4.3 9.4.4 9.4.5 9.5 9.5.1
Sasol 2 and 3 Facilities 181 Introduction 181 Synthesis Gas Production 182 Lurgi Dry Ash Coal Gasification 182 Synthesis Gas Cleaning 182 Fischer–Tropsch Synthesis 183 Gas Loop Design 184 Fischer–Tropsch Refining 186 Synthol HTFT Condensate Refining 188 Synthol HTFT Oil Refining 192 Aqueous Product Refining 194 Coal Pyrolysis Product Refining 196 Synthetic Fuel Properties 198 Evolution of Sasol Synfuels 199 Changes in Synthesis Gas Production 201
Contents
9.5.2 9.5.3 9.5.4 9.5.5 9.5.6 9.5.7 9.5.8 9.6
Changes in Fischer–Tropsch Synthesis 201 Changes in Fischer–Tropsch Condensate Refining 202 Extraction of Linear 1-Alkenes 204 Changes in Fischer–Tropsch Oil Refining 205 Changes in Fischer–Tropsch Aqueous Product Refining 210 Changes in Coal Pyrolysis Product Refining 211 Synthetic Jet Fuel 212 Discussion of the Refinery Design 212 References 214
10 10.1 10.2 10.2.1 10.2.2 10.3 10.3.1 10.4 10.4.1 10.4.2 10.4.3 10.5 10.5.1 10.5.2 10.6
Mossgas Facility 217 Introduction 217 Synthesis Gas Production 218 Natural Gas Liquid Recovery 218 Gas Reforming 218 Fischer–Tropsch Synthesis 220 Gas Loop Design 221 Fischer–Tropsch Refining 222 Oil Refining 222 Aqueous Product Refining 225 Synthetic Fuel Properties 227 Evolution of the PetroSA Facility 227 Addition of Low-Temperature Fischer–Tropsch Synthesis 227 Changes in the Fischer–Tropsch Refinery 227 Discussion of the Refinery Design 228 References 229
11 11.1 11.2 11.3 11.4 11.4.1 11.4.2 11.5 11.6
Shell Middle Distillate Synthesis (SMDS) Facilities 231 Introduction 231 Synthesis Gas Production in Bintulu GTL 232 Fischer–Tropsch Synthesis in Bintulu GTL 233 Fischer–Tropsch Refining in Bintulu GTL 235 Oil Refining 235 Aqueous Product Treatment 238 Pearl GTL Facility 238 Discussion of the Refinery Design 239 References 239
12 12.1 12.2 12.3 12.4 12.4.1 12.4.2
Oryx and Escravos Gas-to-Liquids Facilities 241 Introduction 241 Synthesis Gas Production in Oryx GTL 242 Fischer–Tropsch Synthesis in Oryx GTL 243 Fischer–Tropsch Refining in Oryx GTL 244 Oil Refining 244 Aqueous Product Treatment 247
XI
XII
Contents
12.5
Discussion of the Refinery Design 247 References 248 Part IV
Synthetic Transportation Fuels
249
13 13.1 13.2 13.3 13.3.1 13.3.2 13.3.3 13.3.4 13.3.5 13.3.6 13.3.7 13.3.8 13.3.9 13.4 13.5
Motor-Gasoline 251 Introduction 251 Motor-Gasoline Specifications 252 Motor-Gasoline Properties 253 Octane Number 253 Density 259 Volatility 259 Fuel Stability 261 Alkene Content 261 Aromatic Content 262 Sulfur Content 262 Oxygenate Content 262 Metal Content 263 Aviation-Gasoline 264 Future Motor-Gasoline Specification Changes References 266
14 14.1 14.2 14.2.1 14.2.2 14.3 14.3.1 14.3.2 14.3.3 14.3.4 14.3.5 14.3.6 14.3.7 14.3.8 14.4
Jet Fuel 269 Introduction 269 Jet Fuel Specifications 270 Synthetic Jet Fuel 271 Fuel for Military Use 272 Jet Fuel Properties 273 Net Heat of Combustion 274 Density and Viscosity 275 Freezing Point Temperature 276 Aromatic Content and Smoke Point 276 Sulfur and Acid Content 278 Volatility 278 Stability 278 Elastomer Compatibility and Lubricity 279 Future Jet Fuel Specification Changes 280 References 280
15 15.1 15.2 15.3 15.3.1 15.3.2
Diesel Fuel 283 Introduction 283 Diesel Fuel Specifications 284 Diesel Fuel Properties 286 Cetane Number 286 Density and Viscosity 290
265
Contents
15.3.3 15.3.4 15.3.5 15.3.6 15.3.7 15.3.8 15.3.9 15.4 15.5
Flash Point 290 Lubricity 290 Aromatic Content 292 Sulfur Content 292 Cold-Flow Properties 293 Stability 294 Elastomer Compatibility 294 Diesel Fuel Additives That Affect Refinery Design 295 Future Diesel Fuel Specification Changes 296 References 297 Part V
16 16.1 16.2 16.2.1 16.2.2 16.3 16.3.1 16.3.2 16.3.3 16.3.4 16.3.5 16.3.6 16.3.7 16.4 16.4.1 16.4.2 16.4.3 16.4.4 16.4.5 16.4.6 16.5 16.5.1 16.5.2 16.5.3 16.5.4 16.6 16.6.1 16.6.2 16.6.3 16.6.4 16.7
Refining Technology 301
Refining Technology Selection 303 Introduction 303 Hydrotreating 305 Hydrogenation of Alkenes 306 Hydrodeoxygenation 307 Addition and Removal of Oxygen 308 Dehydration 308 Etherification 309 Hydration 309 Esterification 310 Carbonyl Aromatization 310 Hydroformylation 311 Autoxidation 311 Alkene Conversion 312 Double Bond Isomerization 312 Metathesis 314 Skeletal Isomerization 314 Oligomerization 315 Aliphatic Alkylation 316 Aromatic Alkylation 317 Alkane Conversion 319 Hydroisomerization 319 Hydrocracking 320 Naphtha Reforming and Aromatization 321 Dehydrogenation 322 Residue Conversion 323 Catalytic Cracking 323 Visbreaking 324 Thermal Cracking 324 Coking 326 Fischer–Tropsch Refining Technology Selection References 328
326
XIII
XIV
Contents
17 17.1 17.2 17.2.1 17.2.2 17.2.3 17.3 17.3.1 17.3.2 17.3.3 17.4 17.4.1 17.4.2 17.4.3
Dehydration, Etherification, and Hydration 335 Introduction 335 Dehydration 336 Reaction Chemistry 339 Catalysis 340 Syncrude Process Technology 341 Etherification 343 Reaction Chemistry 345 Catalysis 346 Syncrude Process Technology 347 Hydration 347 Reaction Chemistry 349 Catalysis 349 Syncrude Process Technology 350 References 350
18 18.1 18.2 18.2.1 18.2.2 18.3 18.3.1 18.3.2 18.3.3 18.4 18.4.1 18.4.2 18.4.3 18.4.4
Isomerization 353 Introduction 353 Reaction Chemistry 354 Alkene Skeletal Isomerization 354 Alkane Hydroisomerization 356 Skeletal Isomerization 357 Butene Isomerization Catalysis 358 Pentene Isomerization Catalysis 359 Syncrude Process Technology 360 Hydroisomerization 360 Butane Hydroisomerization Catalysis 362 C5 –C6 Naphtha Hydroisomerization Catalysis 362 Heavy Alkane and Wax Hydroisomerization Catalysis Syncrude Process Technology 364 References 366
19 19.1 19.2 19.3 19.3.1 19.3.2 19.3.3 19.3.4 19.3.5 19.3.6 19.4
Oligomerization 369 Introduction 369 Reaction Chemistry 372 Catalysis 374 Solid Phosphoric Acid 375 H-ZSM-5 Zeolite 378 Amorphous Silica–Alumina 380 Acidic Resin 381 Homogeneous Nickel 383 Thermal Oligomerization 384 Syncrude Process Technology 385 References 388
364
Contents
20 20.1 20.2 20.3 20.3.1 20.3.2 20.3.3 20.4
Aromatic Alkylation 393 Introduction 393 Reaction Chemistry 395 Catalysis 396 Aromatic Alkylation with Ethene 397 Aromatic Alkylation with Propene 399 Aromatic Alkylation with C4 and Heavier Alkenes 401 Syncrude Process Technology 403 References 405
21 21.1 21.2 21.2.1 21.2.2 21.2.3 21.3 21.3.1 21.4 21.4.1 21.4.2 21.5 21.5.1 21.5.2
Cracking 407 Introduction 407 Reaction Chemistry 410 Thermal Cracking 410 Catalytic Cracking 414 Hydrocracking 416 Thermal Cracking 419 Syncrude Processing Technology 421 Catalytic Cracking 421 Catalysis 423 Syncrude Processing Technology 425 Hydrocracking 427 Catalysis 430 Syncrude Processing Technology 434 References 436
22 22.1 22.2 22.3 22.3.1 22.3.2 22.3.3 22.4 22.4.1 22.4.2 22.4.3 22.5 22.5.1 22.5.2 22.5.3
Reforming and Aromatization 441 Introduction 441 Thermal Naphtha Reforming 443 Conventional Catalytic Naphtha Reforming 444 Reaction Chemistry 444 Catalysis 447 Syncrude Processing Technology 449 Monofunctional Nonacidic Pt/L-Zeolite Naphtha Reforming 450 Reaction Chemistry 451 Catalysis 452 Syncrude Processing Technology 453 Aromatization 454 Reaction Chemistry 456 Catalysis 457 Syncrude Processing Technology 460 References 461
23 23.1
Chemical Technologies 465 Introduction 465
XV
XVI
Contents
23.2 23.2.1 23.2.2 23.2.3 23.2.4 23.3 23.3.1 23.3.2 23.3.3 23.3.4
Production of n-1-Alkenes (Linear α-Olefins) 466 Extraction of 1-Pentene and 1-Hexene 467 Extraction of 1-Octene 470 Production of 1-Octene from 1-Heptene 473 Distillate-Range n-1-Alkene Extraction 474 Autoxidation 474 Autoxidation Regimes 477 Reaction Chemistry 478 Fischer–Tropsch Wax Oxidation 480 Syncrude Process Technology 484 References 485 Part VI
Refinery Design 489
24 24.1 24.2 24.2.1 24.2.2 24.2.3 24.2.4 24.3 24.3.1 24.3.2 24.3.3 24.3.4 24.4 24.4.1 24.4.2 24.4.3 24.4.4 24.4.5
Principles of Refinery Design 491 Introduction 491 Refinery Design Concepts 491 Characteristic of the Refining Business 491 Complex Systems and Design Rules 493 Refining Complexity 495 Refining Efficiency 496 Conceptual Refinery Design 497 Linear Programming 497 Hierarchical Design 498 Technology Preselection 498 Carbon-Number-Based Design 499 Real-World Refinery Design 500 Refinery Type 501 Refinery Products and Markets 501 Refinery Feed Selection 502 Refinery Location 503 Secondary Design Objectives 506 References 508
25 25.1 25.2 25.2.1 25.2.2 25.2.3 25.3 25.3.1 25.3.2 25.3.3 25.3.4
Motor-Gasoline Refining 509 Introduction 509 Gap Analysis for Syncrude to Motor-Gasoline 510 Motor-Gasoline Specifications 510 Carbon Number Distribution 511 Composition and Quality 512 Decisions Affecting Motor-Gasoline Refining 514 Chemicals Coproduction 514 Fate of C2 –C4 Hydrocarbons 515 Fate of the Residue and Wax 516 Fate of the Aqueous Product 517
Contents
25.3.5 25.3.6 25.3.7 25.4 25.4.1 25.4.2 25.5 25.5.1 25.5.2 25.5.3
Alkane-Based Naphtha Refining 518 Technology Selection 519 Co-refining 521 Motor-Gasoline Refining from HTFT Syncrude 522 HTFT Motor-Gasoline Design Case I 522 HTFT Motor-Gasoline Design Case II 526 Motor-Gasoline Refining from LTFT Syncrude 529 LTFT Motor-Gasoline Design Case I 529 LTFT Motor-Gasoline Design Case II 534 LTFT Motor-Gasoline Design Case III 537 References 539
26 26.1 26.2 26.2.1 26.2.2 26.2.3 26.3 26.3.1 26.3.2 26.3.3 26.3.4 26.4 26.4.1 26.4.2 26.5 26.5.1
Jet Fuel Refining 541 Introduction 541 Gap Analysis for Syncrude to Jet Fuel 541 Jet Fuel Specifications 541 Carbon Number Distribution 542 Composition and Quality 542 Decisions Affecting Jet Fuel Refining 544 Fate of C2 –C4 Hydrocarbons 544 Fate of the Residue and Wax 545 Technology Selection 546 Co-refining 547 Jet Fuel Refining from HTFT Syncrude 548 HTFT Jet Fuel Design Case I 549 HTFT Jet Fuel Design Case II 552 Jet Fuel Refining from LTFT Syncrude 553 LTFT Jet Fuel Design Case I 555 References 558
27 27.1 27.2 27.2.1 27.2.2 27.2.3 27.2.4 27.3 27.3.1 27.3.2 27.3.3 27.3.4 27.3.5 27.3.6 27.4
Diesel Fuel Refining 559 Introduction 559 Gap Analysis for Syncrude to Diesel Fuel 560 Diesel Fuel Specifications 560 Carbon Number Distribution 561 Composition and Quality 562 Density–Cetane–Yield Triangle 563 Decisions Affecting Diesel Fuel Refining 564 Fate of C2 –C4 Hydrocarbons 565 Fate of the Residue and Wax 565 Fate of the Aqueous Product 565 Technology Selection 566 Co-refining 567 Dealing with the Density–Cetane–Yield Triangle Diesel Fuel Refining from HTFT Syncrude 570
569
XVII
XVIII
Contents
27.4.1 27.5 27.5.1 27.5.2
HTFT Diesel Fuel Design Case I 570 Diesel Fuel Refining from LTFT Syncrude LTFT Diesel Fuel Design Case I 573 LTFT Diesel Fuel Design Case II 576 References 578
28 28.1 28.2 28.2.1 28.2.2 28.3 28.3.1 28.3.2 28.3.3 28.3.4 28.4 28.4.1 28.4.2 28.4.3 28.4.4 28.4.5 28.5 28.5.1 28.5.2 28.5.3
Chemicals and Lubricant Refining 581 Introduction 581 Petrochemical and Lubricant Markets 582 Petrochemicals 582 Lubricants 584 Overview of Chemicals Refining Concepts for Syncrude 585 Alkane-Based Refining 585 Aromatics Production 586 Alkene and Oxygenate Recovery 587 Fuels and Chemicals Coproduction 588 Fischer–Tropsch-Based Petrochemical Refining 591 Alkane Refining 591 Light Alkene Refining 592 Linear 1-Alkene Refining 594 Aromatics Refining 595 Oxygenate Refining 597 Fischer–Tropsch-Based Lubricant Base Oil Refining 597 Group III Lubricant Refining 598 Group IV Lubricant Refining 599 Lubricant Base Oil Refining 600 References 601 Index 603
573
XIX
Preface Life sometimes takes one on a journey that is quite unanticipated. After graduation, I had dreams of making the world a better place by becoming a forensic scientist. Three years later, life took me down a different path. As a young process engineer the objective changed and I had to console myself with trying to make the world a better place in a different way. Thus started an industrial career in the refining of synthetic liquids. The energy business is tremendously dependent on the crude oil price, which by all accounts seems to be inherently unpredictable. The crude oil price holds the synthetic liquids industry to ransom, as it fluctuates in response to many global forces. Today, coal-to-liquids and gas-to-liquids are economical processes – tomorrow it may not be. So, it goes on and on, and has been going on for many decades. When an idealistic individual is confronted with the realities of the energy business and the fickleness of decisions related to the continuous quest for power and money, it can become frustrating. It is not possible to develop technology for the refining of synthetic liquids in phase with the waxing and waning of the oil price. This in turn leads to inefficient refining practices and creates false impressions about the refining of synthetic liquids. Research is not amenable to the stop–start–stop–start cycles dictated by the economic fortunes of the synthetic liquids industry. The wheel has been reinvented many times over, as know-how is lost in times when indirect liquefaction is not economical. This book is an attempt to present and preserve some of the thinking around the refining of Fischer–Tropsch syncrude in the hope that it will help bridge the stop–start–stop–start interest in indirect liquefaction by Fischer–Tropsch synthesis. There are no other works on this topic, except for the occasional chapter in works on Fischer–Tropsch synthesis. The catalysis related to Fischer–Tropsch refining has been discussed in a recent book by Ed Furimsky and myself that is titled ‘‘Catalysis in the refining of Fischer–Tropsch syncrude.’’ There was a deliberate attempt to avoid duplication of effort and overlap with the aforementioned work. This book focuses on the application of catalysis, the processes, the refining technologies, and the refinery design associated with Fischer–Tropsch syncrude. During the writing of this book, some decisions had to be made. The book also had to deal with shortcomings in the reported literature that could not be overcome by the author’s experience in this field. The intent is not to apologize for these decisions and shortcomings but rather to make the reader aware of them. 1) Throughout the book, the International Union of Pure and Applied Chemistry (IUPAC) chemical nomenclature was employed. This may create a slightly unfamiliar feel for many
XX
Preface
2)
3)
4)
5)
6)
7)
readers from the industry and maybe even some readers from the academia. It is a common occurrence to refer to paraffins (not alkanes) and olefins (not alkenes). Yet, having waded through a fair bit of the older literature on Fischer–Tropsch in writing this book, one appreciates the value of having a consistent nomenclature. It was too often necessary to scrounge around to establish what compound or mixture has been described by a colloquial term that had been in common use 80 years ago, but is quite unfamiliar at present. As concession and in order to improve readability, commonly used trivial names and terms were provided in brackets with the IUPAC nomenclature. In cases where the trivial name is unambiguous and recognized in IUPAC nomenclature, the more familiar name was adopted, for example, o-xylene instead of 1,2-dimethylbenzene. In chemical structures, hydrogen atoms are not indicated unless it improves readability. The symbol ‘‘R’’ denotes an alkyl group or hydrogen and the symbol ‘‘M’’ denotes a metal atom. ` The Systeme International d’Unit´es (SI units) were used, albeit with some exceptions. Temperature is reported in degrees Celsius (◦ C) and not Kelvin (K). The conversion from degrees Celsius to Kelvin is easy, just add 273.15. Kinematic viscosity is reported in centistokes (cSt) and not square meter per second (1 cSt = 1 × 10−6 m2 ·s−1 = 1 mm2 ·s−1 ). Not all rates were converted to a per second basis and more familiar time periods were employed for production capacities and flow rates. Since the topic of the book is on refining, it also became clear that the unit of barrels per day (bbl/day) cannot be avoided (1 bbl = 0.158 987 3 m3 ). In the chapters that discuss transportation fuel specifications, the measurement of fuel properties mainly refers to the American Society for Testing and Materials (ASTM) standard test methods. There are of course equivalent methods from the Institute of Petroleum (IP), International Standards Organization (ISO), and various national institutes. Reference to the one rather than the other implies no value judgment. Transportation fuel specifications are country dependent and are ever changing. No attempt was made to provide an anthology of global specifications, which would in any case become outdated rather quickly. The European motor-gasoline (EN228:2004) and diesel fuel (EN590:2004) specifications were selected as the basis for discussion, with reference to some other specifications, including the World Wide Fuel Charter (WWFC). The same applies to jet fuel, where the DEF-STAN 91-91 Issue 6 has been selected as the basis for discussion. There is no implicit value judgment. The discussion focuses on the fundamentals and the specifications are only illustrative in nature. Refining consists of conversion and separation processes. In the book, there is a definite bias toward conversion processes. This does not imply that separation is less important than conversion, but in many instances the challenge in fuel refining is not efficient separation, but efficient conversion. In petrochemical refining, the roles are sometimes reversed. The bias toward conversion goes hand in hand with the focus. The effort that has been expended in literature to correctly identify and quantify compounds varies considerably. If some compounds or compound classes have not been mentioned in conjunction with a specific topic, it does not necessarily imply that these compounds were not present. In Fischer–Tropsch literature, the oxygenates and especially the aqueous products tend to be ignored or are considered with less care than is bestowed on the organic product. Where possible this bias was rectified, but this was not possible in all instances.
Preface
8) The book ‘‘Catalysis in the refining of Fischer–Tropsch syncrude’’ contains an in-depth discussion on the catalysis needed for the refining of Fischer–Tropsch syncrude. It also contains a review of the patent literature on syncrude refining. References to patent literature and catalysis literature have therefore been kept to a minimum. Nevertheless, some discussion of catalysis in the context of refining could not be avoided, since it is critical to the success of syncrude refining. 9) Although every effort has been made to provide a comprehensive discussion of refining, this book is not a general text on oil refining. Process flow diagrams and schematics have consequently not been provided for every technology, and there was a deliberate attempt not to duplicate material readily available in reference texts on crude oil refining. Details related to general issues, such as the pressure and energy balance over fluid catalytic cracking units, were therefore not discussed unless it had a direct bearing on syncrude refining. 10) A number of sections were devoted to the relationship between crude oil refining, transportation fuel specifications, and syncrude refining. Yet, the focus throughout was on Fischer–Tropsch syncrude refining. It was assumed that the reader has at least a superficial knowledge of the conversion processes employed in crude oil refining. If this is not the case, the narrative will be somewhat more taxing to follow, but should still be understandable. Edmonton, AB, Canada, December 2010
Arno de Klerk
XXI
1
Part I Introduction
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
3
1 Fischer–Tropsch Facilities at a Glance 1.1 Introduction
Industrial Fischer–Tropsch facilities are currently only used for coal-to-liquid (CTL) and gas-to-liquid (GTL) conversion. The purpose of such facilities is to convert solid or gaseous carbon-based energy sources into products that may be used as fuels or chemicals. Although Fischer–Tropsch synthesis lies at the heart of the conversion, it is actually only a small part of the overall process. The process can be divided into three steps (Figure 1.1): feed-to-syngas conversion, syngas-to-syncrude conversion, and syncrude-to-product conversion. Generically, this is called indirect liquefaction, because the feed is first transformed into synthesis gas (syngas) and the syngas is then transformed into products. From Figure 1.1 it can be seen that the type of feed materials that can be converted in the first step is not restricted to coal and natural gas. The conversion of biomass in a biomass-to-liquids (BTLs) process and waste in a waste-to-liquids (WTLs) process can likewise be considered. Collectively, all of these processes are referred to as feed-to-liquids (XTLs) conversion processes. The raw feed material limits the technology selection for the feed-to-syngas conversion step, but not for the subsequent steps. Once the feed has been converted into syngas, which is a mixture of carbon monoxide (CO) and hydrogen (H2 ), the syngas can be conditioned to serve as feed for any syngas-to-syncrude conversion technology. Fischer–Tropsch synthesis is not the only possible technology for the conversion of syngas into a synthetic crude oil (syncrude), but together with syngas-to-methanol conversion [1], Fischer–Tropsch synthesis is industrially the most relevant. This book deals with the third step in Figure 1.1, namely, the refining of the syncrude into final marketable products, and it specifically deals with the refining of Fischer–Tropsch syncrude as the title suggests. Since methanol is also a product of Fischer–Tropsch synthesis, the refining of methanol as syncrude component is covered too. The representation in Figure 1.1 does not do justice to the complexity of indirect liquefaction. Whole texts have been devoted to aspects of the indirect liquefaction process, such as coal gasification [2, 3], Fischer–Tropsch technology [4–7], and the catalysis of Fischer–Tropsch syncrude refining [8]. This chapter provides only an overview of Fischer–Tropsch facilities. It shows how the component parts are linked together and why they are interdependent. In subsequent chapters, each one of the topics is revisited in more depth, in order to present the detail that is necessary to comprehensively deal with the topic of this book, namely, Fischer–Tropsch refining. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
4
1 Fischer–Tropsch Facilities at a Glance
Feed
Feed-to-syngas
Coal Gasification Natural gas Reforming Biomass Partial oxidation Waste Oil shale Oil sands Figure 1.1
Syngas-to-syncrude
Syncrude-to-product
Products
Fischer-Tropsch Syngas-to-methanol Kölbel Engelhardt Syngas-to-oxygenates
Refinery processes
Fuels Chemicals
Overall indirect liquefaction process for feed-to-liquids (XTL) conversion.
1.2 Feed-to-Syngas Conversion
Feed-to-syngas conversion is an energy-intensive operation and also the most expensive step in indirect liquefaction. Many of the advantages that are related to the feed-to-syngas conversion step do not depend on subsequent processing. It is these advantages that make indirect liquefaction attractive, despite its poorer energy efficiency than direct liquefaction [9–11]. 1) Feed diversity. One of the major advantages of indirect liquefaction over direct liquefaction is the wide selection of feed materials that can be used. In addition to coal and natural gas, it is possible to employ almost any other carbon source as feed material. The conversion of biomass and waste are attractive concepts, since biomass represents a renewable source of energy and waste conversion represents the beneficial reuse of discarded material. Waste products that can be considered include domestic and industrial waste, for example, discarded plastic containers, old tires, and asphalthenes from carbon rejection processes. However, feed diversity is not the same as feed flexibility. The design of the feed-to-syngas conversion step has to be based on a specific feed slate and it generally has little feed flexibility beyond its designed range of feed compositions. 2) Mineral rejection. Indirect liquefaction has the inherent ability to process and separate carbon matter from mineral matter in mineral-containing carbon sources. Oil shales, peat, coal, and oil sands are all mineral-containing carbon sources. Such solid feed materials are typically converted in gasifiers to produce syngas. Once the carbon in these carbon sources has been oxidized to carbon monoxide, separation of the gaseous products from the mineral matter is easily achieved. The physical state of the rejected mineral matter depends on the gasification technology that was employed and it may be a dry ash or a slag. 3) Heteroatom removal. Carbon-containing feed material usually contains other elements in addition to carbon and hydrogen. When the feed is converted into a raw synthesis gas, heteroatoms in the feed are also converted into gaseous compounds, such as hydrogen sulfide (H2 S), carbonyl sulfide (COS), and ammonia (NH3 ). When the raw synthesis gas is purified, these heteroatom-containing compounds are removed to produce a pure synthesis gas, consisting of only carbon monoxide and hydrogen. With the exception of oxygen, all other heteroatoms are therefore removed during syngas purification. The removal of heteroatoms benefits the syncrude refinery, since the syncrude now only contains Cx Hy Oz -compounds.
1.2 Feed-to-Syngas Conversion
1.2.1 Feed Logistics and Feed Preparation
It is convenient to look at the carbon-containing feed merely as a feed process stream. In the case of natural gas feed that is already available from a pipeline supply, this may be a good approximation, but it is an oversimplification in most other cases. The steps involved in obtaining and preparing feed for indirect liquefaction are more complex (Figure 1.2). The carbon source is not always concentrated, as it is in the case of a natural nonrenewable resource such as coal. Biomass-derived feed is not concentrated at a single point of origin. Biomass has a low energy density and the feed logistics involved in collecting and transporting the biomass from its origin to the indirect liquefaction facility significantly adds to the cost and complexity of the process. Feed pretreatment and logistics are generally costlier than the direct operating cost of indirect liquefaction to produce Fischer–Tropsch syncrude. It can account for up to a third of the total production cost of the whole facility [12]. For natural gas, the feed logistics may be a significant factor in deciding whether to invest in indirect liquefaction or not. Natural gas can be directly distributed by pipeline as fuel gas, or it can be compressed and distributed as liquefied natural gas (LNG). All raw materials, including natural gas, require some form of feed pretreatment before they are suitable for conversion into syngas. The nature of the pretreatment is directly linked to the method of syngas production. It is prudent to select the syngas production technology with this in mind, since feed pretreatment can be a significant cost component. 1.2.2 Syngas Production
All syngas production technologies involve some form of partial oxidation (Chapter 3). It is convenient to consider the production of syngas from gaseous and solid carbon sources separately. Irrespective of the feed, the syngas production technology must be compatible with the feed and it should ideally be selected to meet the syngas requirements of the syngas-to-syncrude conversion technology. As rule of thumb, one aims for a H2 :CO ratio of around 2 in the syngas. The exact H2 :CO ratio that is required depends on the Fischer–Tropsch technology and the design of Fischer–Tropsch gas loop. The H2 :CO ratio can also be adjusted during syngas conditioning (Section 1.2.3).
Resource
Recovery
Transport
Feed preparation
Coal Natural gas Biomass Waste Oil shale Oil sands
Mining Drilling Harvesting Collection
Railroad Pipeline Trucking Shipping
Crushing Conditioning Slurrying Compacting Dewatering Milling Sieving
Figure 1.2
Feed logistics and preparation for indirect liquefaction.
Feed
5
6
1 Fischer–Tropsch Facilities at a Glance
Natural gas is already gaseous and it has no associated mineral matter to contend with. The two main conversion technologies for feed-to-syngas conversion are steam reforming and adiabatic oxidative reforming. Steam reforming is the dominant process for hydrogen production in refineries, and it is able to convert hydrocarbon feed materials ranging from natural gas to heavy naphtha. A steam reformer is essentially a reactor that consists of a fired heater with catalyst-filled tubes placed in the radiant zone of the fired heater. The heat needed for reforming, which is an endothermic conversion, is externally supplied by burning a fuel in the fired heater. The feed consists of a mixture of hydrocarbons and steam (H2 O). The syngas thus produced has a high H2 :CO ratio; a H2 :CO > 2 is typical. When syngas is prepared for Fischer–Tropsch synthesis, steam can be partially substituted by carbon dioxide (CO2 ) to lower the H2 :CO ratio in the syngas [13]. Adiabatic oxidative reforming produces a syngas with a lower H2 :CO ratio; a H2 :CO ratio in the range 1.6–1.9 is typical. The feed consists of a methane-rich hydrocarbon source, an oxidant (air or oxygen), and, in some instances, steam. The heat needed for reforming is directly supplied by combustion of part of the feed. This allows for a more compact design than a steam reformer. However, in the case of oxygen-fired reformers, it has the disadvantage of requiring an associated air separation unit (ASU), which is not required by a steam reformer. Solid feed materials have to be gasified in order to produce syngas. Gasification processes can be classified in terms of gas outlet temperature or reactor properties. These two classifications go hand in hand (Table 1.1) [3]. Low-temperature gasification typically employs a moving bed and has a gas outlet temperature of 425–650 ◦ C. The carbon-containing feed is fed from the top and the oxidizing gas is fed at the bottom. In this countercurrent flow arrangement, the hot ash at the bottom of the bed preheats the oxidizing gas before it enters the gasification zone. Gasification takes place in the middle of the bed. As the hot syngas produced in the gasification zone moves upward through the bed, it preheats and devolatilizes the carbon-containing feed at the top of the bed. Much of the heat recovery therefore takes place in the gasifier. Owing to the lower temperature in the top layer of the gasifier, pyrolysis liquids are coproduced during low-temperature gasification. This is an important distinguishing feature of low-temperature gasification that has implications for downstream refining. The refinery receives not only syncrude from the syngas-to-syncrude Table 1.1
Classification of gasification technologies for feed-to-syngas conversion and their main attributes.
Attribute
Temperature of syngas (◦ C) Reactor technology Particle size of feed (mm) Oxidant demand Steam demand Pyrolysis products in gas H2 :CO ratio in syngas
Gasification technology Low temperature
Medium temperature
High temperature
425–650 Moving bed 6–50 Low High Yes >2 : 1 to <1 : 1
900–1050 Fluidized bed 6–10 Moderate Moderate Possibly <1 : 1
1250–1600 Entrained flow <0.1 High Low No ∼1 : 2
1.2 Feed-to-Syngas Conversion
conversion step but also pyrolysis liquids from gasification. The nature of the pyrolysis liquids depends on the feed material, and biomass pyrolysis produces a different product than that produced by coal pyrolysis. The composition of the pyrolysis products also vary with gasifier type, gasifier operation, and the nature of the specific feed. For example, different coals yield different coal pyrolysis products [14]. The syngas composition is also dependent on the gasifier type and operation, with H2 :CO ratios varying from more than 2 : 1 to less than 1 : 1. Medium-temperature gasification is usually conducted in a fluidized bed and has a gas outlet temperature of 900–1050 ◦ C (lower for biomass as feed). Although the carbon-containing feed is fed countercurrent to the oxidizing agent, as is the case with moving-bed gasifiers, the fluidized bed is well mixed. The temperature profile in the fluidized bed is fairly uniform and it approximates continuous stirred tank reactor behavior. Carbon conversion is therefore lower than that in other gasification types. Some pyrolysis products may be produced during gasification, but considerably less than that produced by using low-temperature gasification. The syngas composition is typically rich in CO and has a H2 :CO ratio of less than 1 : 1. High-temperature gasification is conducted in entrained flow processes with a gas outlet temperature of 1250–1600 ◦ C. The flow of the carbon-containing feed and oxidizing gas is cocurrent. The principal advantage of this type of gasifier is that it produces a clean syngas that is free of pyrolysis products, thereby simplifying downstream gas cleanup and refining. The syngas is CO-rich and typically has a H2 :CO ratio of around 1 : 2. From the preceding discussion, it is important to reiterate that the feed-to-syngas conversion step has a direct impact on the associated refinery design. The coproduction of pyrolysis liquids in some gasification technologies increases the volume and complexity of the feed material that has to be refined. This is not necessarily detrimental, and the corefining of pyrolysis liquids may be synergetic. 1.2.3 Syngas Cleaning and Conditioning
Syngas cleaning is required to remove compounds that are Fischer–Tropsch catalyst poisons from the syngas. The most important and universal poison for Fischer–Tropsch catalysts is sulfur, but other species such as nitrogen-containing compounds, oxygen, chlorides, and bromides may also lead to catalyst deactivation. In addition to Fischer–Tropsch catalyst poisons, CO2 may also be removed during syngas cleaning. The extent of CO2 removal depends on the syngas cleaning technology employed. The degree of CO2 removal that is required depends on the optimum gas loop design. Cleaning technologies that remove most sulfur compounds and CO2 can be selected. Depending on the feed material, the raw syngas also contains NH3 . The NH3 readily dissolves in the aqueous product obtained from cooling the raw syngas and can be removed almost quantitatively by water washing. When low-temperature gasification technology is employed for syngas generation, the pyrolysis products that are coproduced during gasification need to be recovered during syngas cleaning. The pyrolysis products are mostly heavier organic compounds that can be separated as liquids on syngas cooling. These compounds naturally distribute between the less polar organic liquid phase and the more polar aqueous phase according to their polarity. (An analogous partitioning
7
8
1 Fischer–Tropsch Facilities at a Glance
Fuel gas External recycle Water gas shift Syngas
Fischer−Tropsch
Stepwise cooling and separation
Gas reforming Internal recycle
Syncrude
Figure 1.3 Syngas conditioning that may involve one or more of the steps shown: water gas shift conversion, gas reforming, and gas recycle after Fischer–Tropsch synthesis.
of compounds in Fischer–Tropsch syncrude takes place on cooling after synthesis.) The liquid products obtained by condensation likely contain particulate carbon matter from gasification. This has some implications for the further upgrading of these pyrolysis products in the refinery. Although gas cleaning can be viewed as a feed pretreatment step before Fischer–Tropsch synthesis, gas conditioning also benefits from syngas cleaning. Syngas conditioning is necessary to adjust the H2 :CO ratio of the synthesis gas to meet the requirements of Fischer–Tropsch conversion. This is performed by a combination of one or more of the following: water gas shift (WGS) conversion, methane reforming, and gas recycle after Fischer–Tropsch synthesis (Figure 1.3).
1.3 Syngas-to-Syncrude Conversion
Once the feed-to-syngas conversion has been completed and the syngas has been cleaned and conditioned, by whatever combinations of technology, the syngas can be converted into syncrude. It was pointed out that Fischer–Tropsch synthesis is not the only syngas-to-syncrude conversion technology, but, in line with the title of the book, it is the only technology that is discussed. There are basically three syncrude types that are produced commercially, which one can choose from, and it is convenient to classify Fischer–Tropsch synthesis accordingly. The three types of syntheses are iron-based high-temperature Fischer–Tropsch (Fe-HTFT), iron-based low-temperature Fischer–Tropsch (Fe-LTFT), and cobalt-based low-temperature Fischer–Tropsch (Co-LTFT). Syncrude compositions that are representative of each type are given in Table 1.2. Within each syncrude type, there is variation caused by reactor technology, operation, catalyst, and catalyst deactivation. In this respect, syncrude is analogous to crude oil, where the composition of a specific crude oil may vary not only between sources but also within a source between locations and with time [15]. Despite the limited industrial application of Fischer–Tropsch synthesis thus far, there have been a plethora of technologies that were industrially applied (Table 1.3). Many of these are still used commercially. The selection of a catalyst and reactor technology combination is consequently
1.3 Syngas-to-Syncrude Conversion Syncrude compositions that are typical of iron-based high-temperature Fischer–Tropsch (Fe-HTFT), iron-based low-temperature Fischer–Tropsch (Fe-LTFT), and cobalt-based low-temperature Fischer–Tropsch (Co-LTFT) syntheses.
Table 1.2
Product fraction
Carbon range
Compound class
Syncrude compositiona (mass%) Fe-HTFT
Fe-LTFT
Co-LTFT
Tail gas
C1 C2
Alkane Alkene Alkane
12.7 5.6 4.5
4.3 1.0 1.0
5.6 0.1 1.0
LPG
C3 –C4
Alkene Alkane
21.2 3.0
6.0 1.8
3.4 1.8
Naphtha
C5 –C10
Alkene Alkane Aromatic Oxygenate
25.8 4.3 1.7 1.6
7.7 3.3 0 1.3
7.8 12.0 0 0.2
Distillate
C11 –C22
Alkene Alkane Aromatic Oxygenate
4.8 0.9 0.8 0.5
5.7 13.5 0 0.3
1.1 20.8 0 0
Residue/wax
C22 +
Alkene Alkane Aromatic Oxygenate
1.6 0.4 0.7 0.2
0.7 49.2 0 0
0 44.6 0 0
Aqueous product
C1 –C5
Alcohol Carbonyl Carboxylic acid
4.5 3.9 1.3
3.9 0 0.3
1.4 0 0.2
a The syncrude composition is expressed as the total mass of product from Fischer–Tropsch synthesis, excluding inert gases (N2 and Ar) and water gas shift products (H2 O, CO, CO2 , and H2 ). Zero indicates low concentration and not necessarily a total absence of such compounds.
not straightforward. Ultimately, the aim is to produce a specific product slate. One can select the feed-to-syngas and syngas-to-syncrude technologies to yield a syncrude that provides the most efficient feed material to refine to the desired product slate. The analogy with crude oil still holds, since the crude oil basket that is selected for a crude oil refinery, both before design and during operation, also aims to achieve the most efficient conversion of crude oil to the desired product slate [16]. Even so, Fischer–Tropsch synthesis is more versatile, since it is in principle possible to design the Fischer–Tropsch technology (reactor, catalyst, and operation) with a specific refining objective in mind. Unfortunately, in the author’s experience, Fischer–Tropsch technology is developed independent of the ultimate refining objective. The different technologies for Fischer–Tropsch synthesis are discussed in detail in Chapter 4.
9
10
1 Fischer–Tropsch Facilities at a Glance Table 1.3 Fischer–Tropsch technologies that have been applied industrially. The reactor technology, year of first commercial use, and name of the process are given.
Fe-HTFT
Fe-LTFT
Co-LTFT
Fixed fluidized bed (1951, Hydrocol)
Fixed bed (1955, Arbeitsgemeinschaft Ruhrchemie-Lurgi)
Fixed bed (1936, German normal-pressure)
Circulating fluidized bed (1955, Kellogg Synthol)
Slurry bed (1993, Sasol slurry bed process)
Fixed bed (1937, German medium-pressure)
Circulating fluidized bed (1980, Sasol Synthol)
–
Fixed bed (1993, Shell middle distillate synthesis)
Fixed fluidized bed (1995, Sasol Advanced Synthol)
–
Slurry bed (2007, Sasol slurry bed process)
1.4 Syncrude-to-Product Conversion
The product from feed-to-syngas and syngas-to-syncrude conversions is a synthetic crude oil (Table 1.2) that is comparable but different to conventional crude oil. One may therefore argue that it is not necessary to associate syncrude-to-product conversion with a Fischer–Tropsch based XTL process. However, syncrude has some drawbacks compared to crude oil when it is considered as the final product. Crude oil production has to contend with associated gas production, but Fischer–Tropsch syncrude production has to contend with even more product phases. In the case of HTFT syncrude, the syncrude is composed of three different product phases, namely, gaseous, organic liquid, and aqueous. In the case of LTFT syncrudes, there is a fourth phase, namely, organic solids. In fact, less than half of the syncrude from any of the Fischer–Tropsch technologies practised today are available as a liquid organic product. It is only by an upgrading step or proper syncrude refining that more of the syncrude can be transformed into useful products. A Fischer–Tropsch based process without syncrude-to-product conversion is clearly a very inefficient XTL process. The syncrude-to-product conversion step is where the value addition takes place. 1.4.1 Upgrading versus Refining
The difference between upgrading and refining is that upgrading produces intermediate products that must still be refined to produce final products and refining produces final products. The Fischer–Tropsch industry seemingly has come full circle, starting with Co-LTFT and upgrading, moving to Fe-HTFT and Fe-LTFT with refining, and returning to Co-LTFT with upgrading [17, 18]. Yet, this oversimplifies the situation. The German Co-LTFT syncrude was sufficiently upgraded so that it could be blended with coal liquids and crude-oil-derived fuels for use as final products without further refining. The other product fractions were refined to
1.4 Syncrude-to-Product Conversion
chemicals. The same is true for the products from the Co-LTFT-based Shell Middle Distillate Synthesis (SMDS) process, where the distillate is blended with crude-oil-derived diesel fuel as a final product, while the rest of the products are refined to n-alkane (paraffin) solvents, waxes, and lubricating oils. All industrial Fischer–Tropsch facilities have at least partial refining for the syncrude-to-product conversion step. It is consequently more appropriate to identify three levels of syncrude-to-product conversion: 1) Upgrading. In upgrading, all of the products are destined for further refining (not just blending) before becoming final products. The product from an upgrader is a higher quality crude oil and it is marketed as such. 2) Partial refining. When partial refining takes place, at least some of the products are refined to final products, or are destined for blending (not refining) to produce final products. 3) Stand-alone refining. All the products from stand-alone refineries are final products, irrespective of whether they are on-specification transportation fuels, commodity chemicals, or speciality products. In patent literature, there are suggestions on how to upgrade Fischer–Tropsch syncrude into a transportable liquid product [8]. However, philosophically one has to question the use of Fischer–Tropsch synthesis in conjunction with only an upgrader for such a purpose. If the main purpose of an XTL conversion facility is to employ indirect liquefaction to produce a transportable liquid that can be refined elsewhere, syngas-to-methanol synthesis is far more appropriate than Fischer–Tropsch synthesis [1]. 1.4.2 Fuels versus Chemicals
The nature of Fischer–Tropsch syncrudes is such that it lends itself to the recovery and refining of some chemicals. The abundance of alkenes (olefins), oxygenates, and n-alkanes presents the refinery designer with a smorgasbord of options for both extractive and synthetic approaches to the refining of chemicals [19–25]. The production of chemicals from syncrude will be covered in Chapter 28. The chemicals that can be produced in an XTL facility are not only limited to those from Fischer–Tropsch synthesis. The feed-to-syngas conversion step, among others, creates opportunities for chemical coproduction too: 1) Chemicals from gasification liquids. Low-temperature gasification yields pyrolysis products that can contain valuable chemicals typical of the feed source. For example, phenol, benzene, and naphthalene from coal, or methanol from wood. 2) Permanent gases. ASUs associated with oxygen-blown gasification and gas reforming processes produce nitrogen (N2 ) as by-product. In addition, some noble gases, such as argon (Ar), neon (Ne), krypton (Kr), and xenon (Xe), may be separated. 3) Sulfur- and nitrogen-based chemicals. Hydrogen sulfide (H2 S) and ammonia (NH3 ) are recovered during syngas cleaning. Sulfur- and nitrogen-based compounds may also be recovered from gasification liquids.
11
12
1 Fischer–Tropsch Facilities at a Glance
4) Carbon dioxide (CO2 ). Depending on the future legislative framework around carbon emissions, the separation and recovery of CO2 during syngas cleaning and conditioning may be a benefit [26]. Pure CO2 also has value as a chemical. Chemicals generally command a higher price than fuels, and from an economic point of view it makes sense to produce as much chemicals as possible. Yet, at this point it is important to add some words of caution. Some of the high-value chemicals that are commercially produced from Fischer–Tropsch syncrude have rather small markets. In the case of linear α-olefins (1-hexene and 1-octene) and some oxygenates (1-propanol), 20–30% of global demand is being satisfied from a single Fe-HTFT facility [22]. The scope for recovery of such chemicals in future facilities is therefore rather limited. Furthermore, the sale of commodity chemicals such as ethene and propene depends on the market associated with downstream petrochemical facilities. Although both gases have large global markets, the market may not be accessible when the Fischer–Tropsch facility has been sited to employ a cheap but remote carbon source as feed. Historically, investment in Fischer–Tropsch facilities was mainly motivated by energy security and not by economic considerations. As a result, most Fischer–Tropsch facilities were primarily designed to produce transportation fuels. With time, the original motivation and need may dissipate and based purely on economic considerations there may be significant impetus to convert from fuels into chemicals production. A case in point is the Sasol 1 facility in South Africa. It was strategically motivated to produce transportation fuels for the local market, but it was later converted into a petrochemical production facility for economic reasons. A flexible refinery design that includes the production of both fuels and chemicals has some advantages over a refinery that produces only fuels or only chemicals [27, 28]. This will also become apparent in subsequent chapters, where the conversion of syncrude into fuels and chemicals will be considered. Fundamentally, on a molecular level, fuel and chemical coproduction makes sense. Some molecules can easily be converted into either chemicals or fuels, but some other molecules have efficient refining pathways to only one of the two products. Forcing the conversion of a molecule into a product that requires a less efficient refining pathway, is inherently wasteful and violates a number of green chemistry principles (prevent waste, maximize atom economy, and increase energy efficiency). This principle has also been applied to recommend refining technologies for the efficient conversion of Fischer–Tropsch syncrude [29]. The nature of the conversion processes that lead to efficient refining of syncrude is such that the coproduction of fuels and chemicals occurs naturally, often yielding molecules that can be employed either as fuels or as chemicals. 1.4.3 Crude Oil Compared to Syncrude
Some analogies (as well as some differences) between Fischer–Tropsch syncrudes and conventional crude oils have been pointed out. Viewing syncrude as just another crude oil type would simplify the discussion on syncrude-to-product conversion, which then becomes a discussion of crude oil refining. In fact, technologies such as hydrocracking, hydrotreating, and oligomerization that are found in Fischer–Tropsch refineries, are also found in crude oil refineries. Since syncrude and crude oil are both refined to the same products and are subject to the same
1.4 Syncrude-to-Product Conversion Comparison between the product compositions obtained from Fischer–Tropsch synthesis and those found in a typical conventional crude oil.
Table 1.4
Compound class
HTFT
LTFT
Crude oila
Alkanes (paraffins) Cyclo-alkanes (naphthenes) Alkenes (olefins) Aromatics Oxygenates Sulfur compounds Nitrogen compounds Organometallics Water
>10% <1% Major product 5–10% 5–15% None None Carboxylates Major by-product
Major product <1% >10% <1% 5–15% None None Carboxylates Major by-product
Major product Major product None Major product <1% O (heavy) 0.1–5% S <1% N Phorphyrines 0–2%
a
There is considerable variation between different crude oil types and some crude oils may fall outside the boundaries indicated.
performance criteria, the analogy seems to hold. However, at the detail level the analogy quickly breaks down and some fundamental differences between syncrudes and crude oils can be pointed out (Table 1.4) [30]. From a refining perspective, the following attributes of Fischer–Tropsch syncrude, which are different to that of crude oil, have a significant impact on refinery technology selection and refining approach: 1) 2) 3) 4) 5) 6)
Multiple product phases High oxygenate content High alkene content and especially 1-alkene content High concentration of linear products (little cyclic compounds) Absence of sulfur and nitrogen compounds Metal carboxylates.
Conventional crude oil refining technologies can be used with Fischer–Tropsch syncrude, but this often requires feed pretreatment to eliminate compound classes that are not compatible with the technology, or modifications to the technology to deal with Fischer–Tropsch syncrude peculiarities. Direct application of crude oil refining technologies to Fischer–Tropsch derived streams, that is, without pretreatment or modification, generally leads to poor or inefficient operation, and in some cases it leads to complete failure of the technology. There is a subset of conversion processes that perform well with syncrude [29]. Likewise, there is a subset of catalysts that perform well with syncrude [8, 31]. Not all of these catalysts and conversion processes are commonly found in crude oil refineries. It therefore stands to reason that the design of efficient Fischer–Tropsch refineries will differ from that of efficient crude oil refineries [30, 32]. Taken together, these differences form the justification for this book on Fischer–Tropsch refining.
13
14
1 Fischer–Tropsch Facilities at a Glance
1.5 Indirect Liquefaction Economics
The economics of indirect liquefaction is strongly affected by three aspects, namely, the cost of the carbon-containing feed material, product pricing, and the capital cost of the indirect liquefaction facility. There is a complex interrelationship between these factors, and what is pertinent to the present discussion is the role of the refinery in the overall economics. 1.5.1 Feed Cost
Superficially, the feed cost is just a significant cost component of the operating cost of a Fischer–Tropsch facility. As such, it influences the economics. To reduce the feed cost, one can make use of a cheap feed source, but feed price is not something that can intrinsically be controlled. What can be controlled is the design of the facility, and through the design the efficiency of feed conversion can be controlled. Efficiency is a catch-all term that in itself is too broad to provide clear engineering guidance. Control can be exerted over the feed cost through feed selection based on conversion efficiency (not just its price), the carbon efficiency of the design, and the energy efficiency of the design. 1) Feed selection based on the nature of the feed. Not all carbon sources have been created equally and the feed cost expressed in terms of money per unit energy, or money per unit mass, is insufficient to describe the impact of the feed cost on the economics. It is easier to convert less complex and hydrogen-rich feed materials, such as natural gas, than complex materials, such as coal or biomass, by indirect liquefaction. The feed-to-syngas conversion is easier, there are less by-products, and there is less waste (heteroatoms and mineral matter) associated with the feed material. 2) Carbon efficiency. The carbon efficiency specifically relates to the percentage of carbon in the feed that is incorporated into the product. By definition, the remaining carbon is lost during the conversion in various ways, ranging from energy production to refining losses. Thus far not much attention has been paid to carbon efficiency, yet carbon efficiency is a measure of the success in performing the transformation from one type of carbon-based energy carrier to another. Ultimately, the aim of indirect liquefaction is not the conversion of a carbon-based feed into energy, but into useful products. The carbon efficiency for a Fischer–Tropsch based CTL facility is around 28–34% (Table 1.5) [9–11, 33, 34]. The values are in agreement with the author’s experience in dealing with industrial HTFT-based CTL facilities [35]. 3) Thermal efficiency (energy efficiency). The thermal efficiency is the percentage of the energy in the feed that is converted into energy output as power (electricity) and as products. It generally has a higher value than the carbon efficiency, since some of the energy needed in feed-to-syngas conversion, as well as energy produced during syngas-to-syncrude conversion, can be recovered as steam and converted into power. It has been reported that Fischer–Tropsch based GTL facilities have thermal efficiencies of around 60%, whereas those of CTL facilities are around 50% [33]. This illustrates the impact that the nature of the feed has on the efficiency of the process. However, it is also very much dependent on the
1.5 Indirect Liquefaction Economics Carbon efficiencies reported for Fischer–Tropsch based coal-to-liquids facilities, with a coal feed rate around 20 000 tons per day on a dry, ash-free basis.
Table 1.5
Carbon efficiency (%)
32a 33c 28c 32 34 34
Technology description
References
Gasification
Fischer–Tropsch
Refinery
–b Lurgi dry-bottom –b Texaco entrained flow Texaco entrained flow General Electric entrained flow
HTFT HTFT HTFT HTFT LTFT LTFT
Excluded Included –b Excluded Excluded Included
[9] [10] [11] [33] [33] [34]
a Liquid
products’ efficiency is based on energy value. not listed in source reference. c Liquid products and pipeline gas counted toward carbon efficiency. b Data
design. For example, thermal efficiencies of 70 and 58%, respectively, have been reported for HTFT based CTL facilities that coproduce substitute natural gas as the major product [9, 10]. The nature of the feed, the carbon efficiency, and the thermal efficiency all have some bearing on each other, but the relationship with each other and the way in which each affects the impact of feed cost on the economics are intricate. Nevertheless, carbon efficiency and thermal efficiency emerge as two critical measures that determine how sensitive the economics will be to feed cost. These two measures affect the economics differently, since it relates feed cost to product pricing in different ways. The product pricing for energy (electricity), fuels, and chemicals is different and not necessarily based purely on either carbon or energy content. Simply put, the value of feed-to-electricity conversion is different to the value of feed-to-products conversion. The feed cost has to be substantially lower than the price of the products to offset the loss associated with the feed-to-product conversion. 1.5.2 Product Pricing
The product pricing determines the income generated by the Fischer–Tropsch facility. Even if all other variables are held constant, any change in the relative amounts of the products being produced will change the income. This economic reality is well established in crude oil refining, where the refinery design is optimized to produce the highest value product slate that is possible from the crude oil feed selected. The design reflects the anticipated product pricing, but flexibility is built into the design to allow for an imperfect prediction of the future. The flexibility in the refinery design is exploited during refinery operation to respond to changes in product pricing and to keep on optimizing the product slate for maximum income. The choice of feed-to-syngas conversion technology and Fischer–Tropsch technology for syngas-to-syncrude conversion influences the products that can be produced, but it does not determine the product slate. The output from Fischer–Tropsch synthesis is syncrude, and as with
15
16
1 Fischer–Tropsch Facilities at a Glance
crude oil refining, the product slate is determined by the refinery design and refinery operation. The Fischer–Tropsch refinery ultimately determines the extent of value addition. Product pricing is never constant. More importantly, relative product pricing also changes over time. These changes cannot be controlled and are determined by global and local factors. What can be controlled is the design of the facility and through the design the flexibility to respond to changes in product pricing. 1) Feed-to-electricity conversion. The emphasis that is placed on thermal efficiency, as opposed to carbon efficiency, during XTL design determines the amount of power (electricity) produced relative to the amounts of fuels and chemicals. There is little flexibility that can be built into the design to manipulate this ratio, short of modifying the actual design. The product pricing of electricity is determined by local conditions. Exporting power is usually regulated, regional, and dependent on the local infrastructure for electricity supply and distribution. The decision to export electricity as one of the main products from an XTL facility should not be taken lightly. Electricity is not traded on the global market and it locks the XTL facility into a regional supply agreement and regional product pricing. It is also more likely that the price of electricity is related to the feed cost for the XTL facility. The economics of the feed-to-electricity conversion is therefore linked to the cost effectiveness of XTL versus conventional power generation. The danger of viewing XTL as a power producer should therefore be apparent. Why employ syngas as an intermediate if you want to generate steam to drive a turbine? Philosophically speaking, the main aim of XTL is not power generation, although power generation may be a valuable (inevitable) by-product. 2) Feed-to-products conversion. A Fischer–Tropsch refinery can be designed to offer product flexibility. The choice between fuels and chemicals has already been discussed (Section 1.4.2). From an economic perspective, this decision is also dependent on the location of the XTL facility, although less so than with electricity generation. The product price must be discounted by the logistic cost of transporting the product to the market. The logistic cost is small if the products are supplied in the local market. Fuels often have the advantage that there is local demand, whereas chemicals may not have this advantage. The product price difference between chemicals and fuels can quickly be eroded if the XTL facility is located far from chemical markets. Furthermore, some chemicals are difficult to transport, and it may be impractical to sell such products as chemicals to a remote market. 3) Crude oil price. The notion of upgrading syncrude to be sold as a synthetic crude oil has been discussed (Section 1.4.1). In such a case, the product price is directly linked to the crude oil price. The product prices of fuels and chemicals are also influenced by the crude oil price, since it is the main feed cost component in most current petrochemical facilities. The pricing of energy as a product reveals that price is primarily determined by the nature of the energy carrier or product and not by its energy content. In fact, if this had not been the case, XTL would make no sense at all. It is also important to note that some products have a negative price associated with a disposal cost or tax. In this respect, the product price (value or cost) of CO2 and water, which are two of the main products from Fischer–Tropsch based XTL, may become an important factor in determining XTL economics. The difference between produce price and feed cost defines the value addition of XTL and correlates well with the production income and economic viability of XTL facilities.
1.5 Indirect Liquefaction Economics
1.5.3 Capital Cost
There is a large capital cost associated with indirect coal liquefaction based on Fischer–Tropsch synthesis (Table 1.6) [9, 11, 36, 37]. Such processes can be economical if the difference between the product price and feed cost is large enough, although the size of investment required is a serious obstacle to widespread commercial interest. Considering the complexity of a Fischer–Tropsch based CTL facility, the differences in the capital estimates are reasonable. The breakeven crude oil price for a 50 000 bbl/day crude-oil-equivalent Fischer–Tropsch based CTL facility in 2007 was reportedly around US$50–70 [34]. These values are in agreement with author’s experience. The capital cost associated with GTL facilities is less, since the conversion of natural gas into synthesis gas is less complex. From Table 1.6, it is clear that coal preparation and gasification contribute 30% or more of the capital cost, even before taking the cost of utilities, gas cleaning, and air separation into account, which brings the total for syngas preparation to more than 70% of the capital cost for CTL. Capital cost data for Fischer–Tropsch based GTL are listed in Table 1.7 [38, 39]. There is a significant difference in the estimates, with one being double the other. On the basis of the CTL estimates, one is inclined to accept the higher cost estimate as the more realistic of the two. Cost data on recent and current GTL construction projects do not clarify the picture. In fact, they highlight an extreme sensitivity to the economic climate and location [40]. The Oryx GTL facility has been completed at a cost close to the lower estimate in Table 1.7. However, the Oryx GTL design employed a new and unproven Fischer–Tropsch technology, which had technical Table 1.6 Capital cost of Fischer–Tropsch based coal-to-liquids facilities producing 50 000 bbl/day crude oil equivalent liquid products.
Description
Fischer–Tropsch CTL capital cost [9]
[11]
[36]
[37]
Capital cost (US$/bbl daily capacity, 2010)
81 000
78 000a
97 000
85 000
Capital cost distribution (%) Coal preparation and gasification Syngas cleaning and conditioning Air separation Heat recovery and utilities Fischer–Tropsch, gas recovery, refining
30 11 21 15 23
42 15 17 –c 26
30 17 17 18 18d
37 20b 10 16 17
a Production of large volumes of liquid petroleum gas and substitute natural gas, which were not counted toward liquid products for calculation. b Cost of water gas shift reactor included with cost of Fischer–Tropsch synthesis. c Not separately listed. d Refinery contributed to only 5% of the overall capital cost; the refinery included a hydrocracker, naphtha hydrotreater, distillate hydrotreater, naphtha hydroisomerization unit, and catalytic naphtha reformer.
17
18
1 Fischer–Tropsch Facilities at a Glance Table 1.7
Capital cost of Fischer–Tropsch based gas-to-liquids facilities.
Description
Fischer–Tropsch GTL capital cost [36]
[37]
Capital cost (US$/bbl daily capacity, 2010) Liquid product production capacity (bbl/day)
62 000 43 000
33 000 34 000
Capital cost distribution (%) Syngas generation, including air separation Fischer–Tropsch synthesis Upgrading and refining Off-sites, utilities, and other units
47 15 9a 29
30 15 10b 45
a Refinery consists of wax hydrocracker, distillate hydrotreater, naphtha hydrotreater, naphtha hydroisomerization, catalytic naphtha reformer, butane hydroisomerization, and aliphatic alkylation unit. b Refinery consists of only a wax hydrocracker.
problems during start-up. This prevented the facility from reaching full production capacity and additional capital was required to address the technical problems. The actual capital cost was consequently higher due to the increase in capital spending and reduction in production capacity. The committed capital in Shell’s Pearl GTL project in the same location over time increased to US$10 billion in 2007 (about US$70 000 per daily barrel) and further increases were expected [41]. The capital cost of the Escravos GTL facility, using the same technology as Oryx GTL, was reported to have increased to US$6 billion (about US$180 000 per daily barrel) [42]. At the close of 2009, it was announced that the wax production capacity at the Sasol 1 site would be doubled at a cost of US$1.1 billion (about US$200 000 per daily barrel) [43]. The global economy displayed some extreme behavior during the 2005–2010 period, which is partly responsible for the variance in cost estimates. The only consistent indicator is that the capital cost of Fischer–Tropsch based indirect liquefaction facilities is high. It should be added that this is for facilities based on current commercial technology. There are interesting developments in the field of smaller scale Fischer–Tropsch based designs for BTL. Out of necessity such designs have to make trade-offs to reduce complexity and size. By doing so, some opportunities in process integration and intensification are exploited, which are not seen in larger scale designs. The adage ‘‘economy-of-scale’’ may therefore not hold true. The cost associated with indirect liquefaction should be evaluated in context. Other competing technologies for the production of synthetic fuels are also capital intensive (Table 1.8) [44]. These values are without the cost of refining. The capital cost associated with the refining of the Fischer–Tropsch syncrude is small (around 10%) compared to the total capital cost of the indirect liquefaction facility. This has some important implications for refinery design: 1) Maximize carbon efficiency in the refinery. Any carbon contained in the Fischer–Tropsch syncrude has already required 90% of the total capital cost to produce. Designing a refinery that leaves syncrude unrefined in order to save on the refinery capital cost makes little sense.
References Table 1.8
Capital cost estimates of competing technologies for synthetic fuel production.
Synthetic fuel technology
Capacity (bbl/day equivalent)
Capital cost (US$/bbl daily capacity, 2010)
50 000 50 000 50 000
84 000 78 000 66 000
10 250b
230 000
10 250b 10 250b
58 000 44 000
Direct coal liquefaction Oil shale recovery Oil sands recovery Coal gasification for SNGa Advanced gasifier design Underground gasification Horisontal coal seams Steeply dipping coal seems a
SNG, substitute natural gas. BTU gas, about 60 GJ equivalent.
b Medium
2) Indirect liquefaction facilities are inherently complex. Although complex refinery design is not advocated, the complexity of the refinery design has little impact on the overall complexity of the project. 3) Produce final products. There is a trap in performing incremental economics on a Fischer–Tropsch refinery, since it ignores the significant capital and operating costs associated with the production of the syncrude. For example, if you cannot justify spending 10% more to refine naphtha to a motor-gasoline, how did you justify spending nine times as much on making the naphtha in the first place? References 1. Olah, G.A., Goeppert, A., and Prakash, G.K.S.
2.
3. 4.
5.
6.
(2006) Beyond Oil and Gas: The Methanol Economy, Wiley-VCH Verlag GmbH, Weinheim. Rezaiyan, J. and Cheremisinoff, N.P. (2005) Gasification Technologies. A Primer for Engineers and Scientists, Taylor & Francis, Boca Raton, FL. Higman, C. and Van der Burgt, M. (2008) Gasification, 2nd edn, Elsevier, Amsterdam. Anderson, R.B., K¨olbel, H., and Ralek, M. (1984) The Fischer–Tropsch Synthesis, Academic Press, Orlando, FL. Steynberg, A.P., Dry, M.E. (eds) (2004) Fischer–Tropsch Technology, Studies in Surface Science and Catalysis, Vol. 152, Elsevier, Amsterdam. Davis, B.H. and Occelli, M.L. (eds) (2007) Fischer–Tropsch Synthesis, Catalysts and Catalysis, Studies in Surface Science and Catalysis, Vol. 163, Elsevier, Amsterdam.
7. Davis, B.H. and Occelli, M.L. (eds) (2009) Ad-
8.
9. 10.
11.
12.
vances in Fischer–Tropsch Synthesis, Catalysts and Catalysis, Taylor & Francis, Boca Raton. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Nowacki, P. (1979) Coal Liquefaction Processes, Noyes Data Corporation, Park Ridge, NJ. Mangold, E.C., Muradaz, M.A., Ouellette, R.P., Rarah, O.G., and Cheremisinoff, P.N. (1982) Coal Liquefaction and Gasification Technologies, Ann Arbor Science Publishers, Ann Arbor. De Malherbe, R., Doswell, S.J., Mamalis, A.G., and De Malherbe, M.C. (1983) Synthetic Fuels From Coal, Fortschritt-Berichte der Verein Deutscher Ingenieure Zeitschriften, Reihe 3, Vol. 79, VDI-Verlag, D¨usseldorf. Zwart, R.W.R., Boerrigter, H., and Van der Drift, A. (2006) The impact of biomass pretreatment on the feasibility of overseas biomass conversion to Fischer – Tropsch products. Energy Fuels, 20, 2192–2197.
19
20
1 Fischer–Tropsch Facilities at a Glance 13. Aasberg-Petersen, K., Christensen, T.S.,
14. 15.
16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
26. 27.
28.
Dybkjær, I., Sehested, J., Østberg, M., Coertzen, R.M., Keyser, M.J., and Steynberg, A.P. (2004) Synthesis gas production for FT synthesis. Stud. Surf. Sci. Catal., 152, 258–405. Gavalas, G.R. (1982) Coal Pyrolysis, Coal Science & Technology, Vol. 4, Elsevier, Amsterdam. Wauquier, J.-P. (ed.) (1995) Petroleum Refining, Crude Oil. Petroleum Products. Process Flowsheets, Vol. 1, Editions Technip, Paris. Favennec, J.-P. (ed.) (2001) Petroleum Refining, Refinery Operation and Management, Vol. 5, Editions Technip, Paris. De Klerk, A. (2008) Refining of Fischer-Tropsch syncrude: lessons from the past. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 53 (2), 105–109. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 331–364. Meintjes, J. (1975) Sasol 1950–1975, Tafelberg, Cape Town. Dry, M.E. (1987) Chemicals produced in a commercial Fischer-Tropsch process. ACS Symp. Ser., 328, 18–33. Gregor, J.H. (1990) Fischer-Tropsch products as liquid fuels or chemicals. An economic evaluation. Catal. Lett., 7, 317–332. Collings, J. (2002) Mind Over Matter. The Sasol Story: A Half-century of Technological Innovation, Sasol, Johannesburg. Steynberg, A.P., Nel, W.U., and Desmet, M.A. (2004) Large scale production of high value hydrocarbons using Fischer-Tropsch technology. Stud. Surf. Sci. Catal., 147, 37–42. Redman, A. (2005) Production of olefins and oxygenated compounds from Fischer-Tropsch. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd179. De Klerk, A., Dancuart, L.P., and Leckel, D.O. (2005) Chemicals refining from Fischer-Tropsch synthesis. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd185. Schrag, D. (2009) Coal as a low-carbon fuel? Nat. Geosci., 2, 818–820. Chadwick, J.L. (1977) Economics of Chemical Refineries. SRI Process Economics Program Report 107. Stanford Research Institute, Menlo Park, CA. (1995) Petroleum Refining Profitability. SRI Process Economics Program Report 215, Stanford Research Institute, Menlo Park, CA.
29. De Klerk, A. (2008) Fischer-Tropsch refining:
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40. 41. 42. 43. 44.
technology selection to match molecules. Green Chem., 10, 1249–1279. De Klerk, A. (2007) Environmentally friendly refining: Fischer-Tropsch versus crude oil. Green Chem., 9, 560–565. De Klerk, A. (2009) Catalysts important in the refining of Fischer-Tropsch syncrude to fuels. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 116–117. De Klerk, A. (2011) Fischer-Tropsch fuels refinery design. Energy Environ. Sci., DOI: 10.1039/c0ee00692k. Steynberg, A.P. and Nel, H.G. (2004) Clean coal conversion options using Fischer-Tropsch technology. Fuel, 83, 765–770. Williams, R.H., Larson, E.D., Liu, G., and Kreutz, T.G. (2009) Fischer-Tropsch fuels from coal and biomass: strategic advantages of once-through (‘‘polygeneration’’) configurations. Energy Procedia, 1, 4379–4386. De Klerk, A. (2010) Indirect liquefaction carbon efficiency. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 55 (2), 338–339. Kreutz, T.G., Larson, E.D., Liu, G., and Williams, R.H. (2008) Fischer-Tropsch fuels from coal and biomass. Proceedings of the 25th Annual Pittsburgh Coal Conference, 29 September – 3 October, 2008, Pittsburgh, PA. Rahmim, I.I. (2008) GTL, CTL finding roles in global energy supply. Oil. Gas J., 106 (12), 22–31. Choi, G.N., Kramer, S.J., Tam, S.S., and Fox, J.M. III (1997) Design and economics of a Fischer-Tropsch plant for converting natural gas to liquid transportation fuels. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 42 (2), 667–671. Dry, M.E. and Steynberg, A.P. (2004) Commercial FT process applications. Stud. Surf. Sci. Catal., 152, 406–481. Forbes, A. (2007) Surge in interest a long time coming. Pet. Econ., 74 (1), 19–20. Forbes, A. (2007) Reality check. Pet. Econ., 74 (7), 30. (2008) Sasol cuts stake in Escravos GTL as costs rise to $6bn. Pet. Econ., 75 (10), 30. (2009) Sasol will double synthetic wax. Chem. Eng. News, 87 (49), 23. Barbiroli, G. and Mazzaracchio, P. (1995) Synthetic fuel technologies as strategic pathways. Energy Sources, 17, 595–604.
21
2 Refining and Refineries at a Glance 2.1 Introduction
The purpose of a refinery is a simple one, namely, to take a raw material and transform it into a more useful and valuable product that is desired in the marketplace. This description is valid for all refineries, immaterial of whether it converts crude oil to fuels or metal bearing ores into refined metals. In this chapter, we focus on oil refining. There are three components of refining and refinery design (Figure 2.1): 1) Raw material. At present, conventional crude oil dominates the refining scene. It is by far the preferred raw material for refining. Although we refer to crude oil by using a collective term, there are many different types of crude oils with widely differing properties (Section 2.2). Other raw materials are gaining market share, but these alternative oil sources still constitute a small fraction (<5%) of the raw material used for oil refining. Some of the important alternative raw materials are unconventional crude oils, such as oil sand-derived bitumen, Fischer–Tropsch syncrudes, coal liquids, and bio-oils. 2) Transformation process (refining). The transformation of the raw material into products takes place in the refinery (or in an upgrader if the product is only a higher quality crude oil intermediate). The products determine the refinery classification, which can be as follows: energy refineries, nonenergy refineries, and mixed-type refineries. The actual transformation is performed by a combination of refining technologies (Section 2.3), each of which consists of various conversion and separation steps. To accomplish this transformation process efficiently, the refinery must be designed with specific raw materials and products in mind. It is very important to know the properties of the raw materials in relation to those of the products. When products and raw materials are well matched on a molecular level, refining becomes more efficient, which is a central theme in this book. On the detail level, the refinery design is also dependent on location. 3) Products. The ultimate aim of refining is to produce products that are more useful and valuable than the raw materials. The main product types are transportation fuels (energy refinery), petrochemicals and lubricants (nonenergy refineries), and intermediates (upgrader). The product slate is usually selected to supply a specific market. The products and markets must be considered together, since the products determine what will be made, while the Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
22
2 Refining and Refineries at a Glance
Raw material Crude oil Oil sands bitumen FT syncrude Coal liquids Bio-oils
Figure 2.1
Transformation process Conversion processes Separation processes
Products Chemicals Fuels Oils Intermediates
Refinery
Key elements of refining.
markets determine the product specifications. Less frequently, the product slate is dictated by strategic considerations, or it is selected to exploit a specific raw material advantage. The objective of this chapter is threefold. Firstly, it is to give an overview of crude oil refining, which is more familiar territory than Fischer–Tropsch refining. Secondly, it is to highlight the relationship between raw materials, products, and refining. Lastly, it is to show that refineries are dynamic and how refinery design has had to evolve over time to embrace new technologies in order to meet the changing demands imposed by products, markets, legislation, and politics.
2.2 Conventional Crude Oil
The roots of refining lie in conventional crude oil. It is difficult to follow the developments in refining without a feel of the raw material that has to be transformed. There are many texts dealing with crude oil composition, and the selection referred to here is not intended to be exhaustive [1–3]. The approach taken in this chapter is to highlight the main compound classes and properties of crude oil that influence the design of the transformation process to produce refined products. No two crude oils are the same and there are in the order of 150 different crude oils being traded globally at present [4]. There is not a single comprehensive and generally valid classification system for crude oils. Many classification systems have been proposed, but the varied nature of crude oil makes generalization difficult. Nevertheless, crude oils can be grouped by making reference to three attributes that significantly affect refining: 1) Hydrocarbon class. Crude oils are grouped as paraffinic, naphthenic, aromatic, or asphaltic based on the main hydrocarbon class present in the crude oil. Strictly speaking, this is a mixed classification, since ‘‘asphaltic’’ is not a hydrocarbon class but a solubility class. When a crude oil is asphaltic, one can expect a significant heavy fraction, but by being asphaltic it usually implies that the lighter fractions are mainly aromatic in character [5]. 2) Distillation profile. The distillation profile is a quantitative measure of the volume fraction of the oil that distills in a specific boiling range. On the basis of the amount of material that boils in the distillate and naphtha range, a crude oil can be classified as heavy or light (Crude oil distillation terminology is discussed in Section 2.3.1.) The distillation profile is a quantitative measure, but the terms ‘‘heavy’’ and ‘‘light’’ are relative measures, and heaviness is often used to distinguish between crude oils from the same region that have different distillation profiles. The density can also give an indication of heaviness and is more often used for classification purposes. On the basis of density, crude oils are light (<825 kg·m−3 ), medium (825−875 kg·m−3 ), heavy (875−1000 kg·m−3 ), or extra-heavy (>1000 kg·m−3 ) [2].
2.2 Conventional Crude Oil
3) Heteroatom content. The most important heteroatoms in crude oil in the order of importance are S > N > O. High-sulfur crude oils require more intensive refining than low-sulfur crude oils. When a crude oil contains a high amount of sulfur, it is called ‘‘sour’’ crude, whereas low-sulfur crude oils (<0.5 mass%) are called ‘‘sweet.’’ Nitrogen-containing compounds inhibit acid catalysis and are more refractory to remove from the crude oil during refining. Oxygenates tend to occur only in the heavier crude oil fractions and do not pose a serious refining challenge, with the exception of the acid content. The total acid number (TAN) is an indication of the corrosiveness of the crude oil. Crude oil with a TAN > 1 mg KOH·g−1 is considered a high-TAN crude oil. The metal content of crude oil is also important, but it is not used for classification. Despite the usefulness of grouping crude oils based on key properties, it is the detail that matters in refinery design. It is therefore worthwhile to look more closely at each property and point out some implications for refining. 2.2.1 Hydrocarbons in Crude Oil
The main hydrocarbon compound classes that are present in conventional crude oil are alkanes (paraffins), cycloalkanes (naphthenes), and aromatics, with very little or no alkenes (olefins). It is also unlikely that dienes and alkynes are present in crude oil [3]. As is often the case, there are exceptions. For example, Pennsylvanian crude oils may contain up to 3% alkenes. Although unrefined crude oil is usually devoid of unsaturated aliphatic compounds, such compounds are produced during refining. The alkanes found in crude oil tend to have a high linear hydrocarbon content. This affects the fuel quality of straight-run products. In this respect, paraffinic crude oils share some refining needs with Fischer–Tropsch syncrude, since both are rich in linear hydrocarbons. The cycloalkanes most frequently encountered in crude oil have five- or six-membered rings, which are present as monocyclic and multicyclic compounds. In some cases, multicyclic compounds with four- and five-membered rings are found, which serve as biochemical markers. In general, cycloalkanes are desirable compounds from a refining perspective. Mono-, di-, and polynuclear aromatics are found in percentage level concentrations in most crude oils. The aromatic content of straight-run naphtha considerably varies between crude oils, but it is on average around 10%. Aromatic species may also incorporate heteroatoms, which will be discussed separately. 2.2.2 Sulfur Compounds in Crude Oil
Sulfur is the heteroatom that is most commonly found in crude oil and the total sulfur is usually in the range of 0.1–5 mass% depending on the crude and seldom falls outside the range of 0.05–6.0 mass% [3]. Examples of very-high-sulfur crude oils can be found: the 14% sulfur crude from Rozel Point in Utah, USA. Sulfur is the least desirable heteroatom from a transportation fuel perspective. The distribution of sulfur compounds is not equal over the distillation range and the sulfur content generally
23
24
2 Refining and Refineries at a Glance Table 2.1 Distribution of sulfur compounds over the distillation range of a conventional crude oil containing 1.2 mass% sulfur.
Distillation range (◦ C)
S content (mass%)
70–180 160–240 230–350 350–550 >550
Sulfur compound distribution (mass%)
0.02 0.2 0.9 1.8 2.9
a Benzothiophenes
Thiols
Sulfides
Thiophenes
Othera
50 25 15 5 Trace
50 25 15 5 Trace
Trace 35 35 30 10
– 15 35 60 90
and heavy sulfides.
increases with boiling point (Table 2.1) [6]. The compound classes of the sulfur species also differ, and the heavier boiling fractions contain the more refractory sulfur compound classes. The sulfur can be present in an inorganic form, as elemental sulfur, carbonyl sulfide (COS), or hydrogen sulfide (H2 S), or it can be present in an organic form. Free elemental sulfur is not often found in crude oil, but when it is, it can be present as a suspension or it can be dissolved in the oil. Hydrogen sulfide is found in the reservoir gas, but the amount dissolved in the oil is typically less than 50 µg·g−1 . The organic classes most often encountered are sulfides, thiols, and thiophenes (Figure 2.2). Sulfides can either be mono–sulfides (R–S–R), or disulfides (R–S–S–R). Both acyclic and cyclic sulfides, including sulfides that are attached to aromatic rings, such as thiaindane, are present. Thiols (mercaptans) are the sulfur equivalents of alcohols. Owing to the presence of the –SH functional group, the thiols are weak acids (much more than alcohols) [7]. These compounds are also extremely malodorous. The thiols are mainly present in the lighter boiling fractions, but their concentration in crude oil is usually quite low. Thiophenes are aromatic compounds containing sulfur as part of the ring structure and may take the form of thiophene, benzothiophene, dibenzothiophene, and even heavier thiophene derivatives. Most of the sulfur in crude oil occurs in this form. The thiophenes in general and the sterically hindered dibenzothiophenes in particular present a significant challenge to deep hydrodesulfurization to reduce the sulfur content to levels <10 µg·g−1 for transportation fuel [8].
S R R' Sulfide
S
S S R R' Disulfide
R S Thiaindane
S
Thiophene Benzothiophene
SH
Thiol
S Dibenzothiophene
Figure 2.2 Sulfur-containing compound classes found in conventional crude oil.
2.2 Conventional Crude Oil
NH2 R
R
Primary amine
R''
N
N R
R'
Secondary amine
N
O R'
H Indole
N
C
R
Tertiary amine
N
H Pyrrole
N
H
NH2
Amide
N H Carbazole
N Isoquinoline
N
Pyridine
Quinoline
Acridine
Figure 2.3
Nitrogen-containing compound classes found in conventional crude oil.
2.2.3 Nitrogen Compounds in Crude Oil
Nitrogen-containing compounds are mostly found in crude oil fractions boiling above 250 ◦ C. The nitrogen content depends on the crude oil origin and is usually less than 0.5 mass%, but it can be as high as 2 mass% [3]. Basic, neutral, and acidic nitrogen compounds are present, with the main nitrogen-containing compound classes being amines, amides, pyrroles, and pyridines (Figure 2.3) [2]. Crude oil also contains nitrogen compounds of the porphyrin family, which are able to chelate metal ions (Section 2.2.5). Amines are basic nitrogen compounds and are found as primary, secondary, and tertiary amines depending on the number of alkyl groups attached to the nitrogen. The basic nitrogen-containing compounds inhibit catalysis over acidic refining catalysts and are especially undesirable. Amides are acidic nitrogen compounds and can be found as both aliphatic and aromatic compounds. Pyrroles are aromatic compounds containing nitrogen as part of a five-membered ring structure and may be present as pyrrole, indole, carbazole, and even heavier derivatives of pyrrole. These are all neutral compounds, despite the presence of the N–H functionality. This is the result of the lone-pair electrons on nitrogen being delocalized in the aromatic π-electron cloud and they are consequently not available for sharing with acids. In most of the resonance structures of pyrrole, the nitrogen carries a positive charge. Pyridine and its derivatives such as quinoline, isoquinoline, and acridine are basic nitrogen compounds, with the nitrogen forming part of a six-membered aromatic ring. These compounds and their heavier derivatives occur throughout the distillation range. As in the case of the amines, these compounds are undesirable and inhibit conversion over acidic refining catalysts. 2.2.4 Oxygenates in Crude Oil
The oxygen-containing compounds in crude oil are probably the least discussed of the main heteroatom compounds. This is understandable, because oxygenates are generally present in low
25
26
2 Refining and Refineries at a Glance O
O OH O Furan
Figure 2.4
R Phenol
C
O R'
Ester
C OH R Carboxylic acid
Oxygenate classes found in conventional crude oil.
concentrations, 0.05–1.5 mass% [3], and apart from the organic acids, these have little impact on the refining severity required. The main oxygenate classes present in crude oil are furans, phenols, esters, and carboxylic acids (Figure 2.4). Furans are the oxygen equivalent of thiophenes and contain oxygen as part of an aromatic ring structure. Phenols are aromatic alcohols, but are more acidic than aliphatic alcohols due to the additional resonance stabilization of the phenoxide ion. Phenols (pKa ≈ 10) are much weaker acids than the carboxylic acids (pKa ≈ 5) in crude oil and should, strictly speaking, not be counted toward the TAN of the crude oil. Esters are present in low concentration. Both aliphatic and aromatic esters can be found in crude. Carboxylic acids are not evenly distributed over the whole boiling range [2]. In lighter cuts, some linear aliphatic carboxylic acids can be found, but usually most of the carboxylic acids in crude oil are naphthenic acids (five- and six-membered ring cycloalkane carboxylic acid derivatives). The naphthenic acids include a variety of polycyclic structures. The acid composition is very much dependent on the crude oil origin and age and younger crude oils are richer in linear carboxylic acids, rather than naphthenic acids [9]. The organic acids can cause corrosion problems when processing crude at elevated temperature, and this is exacerbated by high flow conditions. Few refineries are capable of refining high-TAN crude oils. The inability to process high-TAN crude oils is not related to the conversion properties of the organic acids, but it is due to the metallurgical limitations of the refinery equipment (carbon steel is readily corroded by high-TAN crude oils). The corrosion products are iron carboxylates and naphthenates of the general formula Fe(RCO2 )2 . The nature of the naphthenic acids is important and cases of severe corrosion by low-TAN crude oils, with a TAN of 0.2–0.3 mg KOH·g−1 , have been reported [10]. However, the relationship between corrosion, TAN, and the nature of the naphthenic acids is far from clear [11]. 2.2.5 Metals in Crude Oil
The main metal impurities in crude oil are nickel and vanadium and, to a lesser extent, iron. Other metals, such as mercury and arsenic, may also be present in the oil phase in low concentration. The metal content can vary from 1 to 1000 µg·g−1 depending on the crude oil. These metals are found in the heavier fractions and end up in the residue processing units of a refinery. A high metal content is undesirable, since it forms deposits on refining catalysts, thereby reducing catalyst lifetime. Some of the metals are trapped in molecules of the porphyrin family (Figure 2.5), which act as efficient chelating agents for ions like Ni2+ and VO+ . However, not all Ni and V are contained
2.2 Conventional Crude Oil Figure 2.5 N
Metal-containing porphyrin structure.
N M
N
N
M = Ni2+, VO+
in porphyrin structures [12]. The carboxylic acids (linear and naphthenic acids) may likewise act as effective metal carriers, and the formation of iron carboxylates and naphthenates as corrosion products during crude oil refining has been pointed out (Section 2.2.4). Metal soaps can also be formed with alkali metals and alkali earth metals, which have been implicated in severe oil production problems related to deposit and emulsion formation [13]. In addition to the metals contained in organic compounds, metals are present in the form of inorganic salts. Sodium, magnesium, and calcium chlorides are the most common, but sulfates and carbonates can also be found. Salt levels vary, with most crude oils having an inorganic salt level of 25−350 µg·g−1 of NaCl equivalent. As in the case of organic salts, the major portion of the metals finds its way to the residue fraction, causing problems in the refinery by fouling fuel oil burners and undermining product quality. Magnesium and calcium chlorides start to hydrolyze at 120 ◦ C in the presence of water (e.g., MgCl2 + 2H2 O → Mg(OH)2 + 2HCl), which can lead to severe corrosion at higher temperatures. 2.2.6 Physical Properties
The vapor pressure of the crude oil as it is produced can be as high as 2 MPa. Storage and transportation of crude oil with such a high vapor pressure is uneconomical. To reduce the vapor pressure, the crude oil goes through a number of pressure reduction stages, which are basically flash drums in series. The pressure is reduced from the wellhead pressure to less than 0.1 MPa. The associated gas that is obtained during the pressure reduction is separated from the crude, leaving the crude oil with a vapor pressure of less than 0.1 MPa. The low content of C4 and lighter compounds in crude oil is consequently an artifact of the way in which crude oil is produced. The flashing of the associated gas influences the boiling point distribution and physical properties of a crude oil. The importance of the distillation profile has already been pointed out. Other important physical properties of crude oil are as follows: 1) Density. The density of crude oil is mostly in the range of 800–1000 kg·m−3 . It gives an indirect indication of the composition of the crude and distillation range, with less dense crude oils being more hydrogen rich and lighter boiling. In the oil industry, the density is often expressed in degrees API and the conversion to SI units is given by Equation 2.1, where density (ρ) is in units of kg·m−3 . ◦
API = 141.5/(0.001 · ρ) − 131.5
(2.1) ◦
2) Pour point. Crude oil pour points usually range between −60 and +30 C. The pour point of crude oil is important from a processing point of view. It gives an indication of the minimum temperature at which a crude oil can still be pumped, although it is based on an oversimplification, since pumpability is also affected by pumping itself (pumping inhibits crystallization processes, thereby increasing fluidity). Crude oil does not congeal at a
27
28
2 Refining and Refineries at a Glance
Atmospheric residue Extract with n-heptane (C7:oil = 30 : 1)
Soluble
Insoluble Soluble
Maltene Figure 2.6
Asphaltene
Extract with benzene
Insoluble Sediment
Solubility classification of crude oil residue.
specific temperature. Crude oil is a complex mixture of compounds and different compounds crystallize out of solution at different temperatures to gradually increase resistance to flow. The thermal history of crude plays a role in determining the pour point and must be known. Heating a crude to 45–65 ◦ C redissolves some of the paraffinic seed crystals, thereby lowering the pour point, while heating it to 100 ◦ C raises the pour point because of loss of light material by evaporation. 3) Viscosity. Like pour point, viscosity is important from a processing point of view, because it influences the pumping cost and the pressure drop in pipelines. The kinematic viscosity of crude oils varies over a wide range, having values of less than 10 to more than 5000 mm2 ·s−1 (cSt) when measured at 20 ◦ C. The viscosity–temperature relationship depends on the composition of the crude. Paraffinic crudes show a rapid increase in viscosity with a decrease in temperature, while naphthenic crudes have a more gradual response. 4) Carbon residue. The carbon residue is determined by destructive distillation in the absence of air. It gives an indication of the coke-forming propensity of the crude oil and is useful for determining the best way to upgrade residue fractions. 5) Asphaltene content. The asphaltene content is an indication of the solubility of the crude oil compounds in a paraffinic as opposed to an aromatic matrix (Figure 2.6) [2]. The refining of crude oil leads to changes in its composition and is often aimed at increasing its hydrogen content. This may adversely affect the solubility of some compounds in the crude oil and lead to precipitation. The asphaltene content is determined by precipitating the asphaltenes from an aromatic (benzene) solution with a light alkane. Industrial practice varies and propane, butane, and light naphtha range alkanes can all be used for asphaltene precipitation. The reported asphaltene content depends on the alkane used for precipitation.
2.3 Products from Crude Oil
Now that the properties and composition of conventional crude oil have been described, we can look at the properties and composition of the products produced from crude oil. This is not as straightforward as describing the crude oil. The products and properties of the products
2.3 Products from Crude Oil
produced from crude oil changed over time. The evolution of crude oil refining and how it was influenced by changes in the market for different products are described in the next section (Section 2.4). The products produced from crude oil can be broadly categorized into three groups, with the products within each group having some general characteristics: 1) Fuels. Heating and transportation fuels are by far the largest volume products produced from crude oil. Fuels are energy carriers and all fuels are ultimately combusted. Fuels are separated into different types on the basis of their boiling range and not composition. It is therefore possible to produce two fuels of the same type but with very different compositions. As such, fuels are by definition mixtures and not pure compounds. The quality of a fuel is related to its combustion properties in the application for which it has been intended. This is important, because the perceived quality of a fuel depends on its performance in a specific application; it is not an intrinsic characteristic of the fuel. Furthermore, a fuel type may have different applications, which imposes different quality requirements even though it is still the same fuel. Fuels for the same application may also have different quality requirements depending on the geographic location and are regulated by government agencies through fuel specifications. 2) Petrochemicals. The chemical industry makes use of a small number of commodity chemicals (mainly short-chain alkenes and monocyclic aromatics) from which a bewildering array of more specialized products are produced. Most commodity chemicals are pure compounds, or mixtures with well-defined composition. The quality of a chemical is related to its purity and the nature of the impurities that may be present. The quality of the chemical can therefore be ascertained independent of the application, which may qualify or disqualify the chemical for a specific use. This is fundamentally different from fuels, where quality is very application specific. 3) Lubricants. Lubricant base oils, such as fuels, are mixtures of compounds. Base oils are differentiated on the basis of viscosity and viscosity index (a measure of how much the viscosity changes with temperature). There is a correlation between the base oil classification and its boiling range, but boiling range is not used for classification as in the case of transportation fuels. Lubricant base oil classification is discussed in Section 28.2.2. In terms of quality, one has to distinguish between lubricant base oils and lubricants. Lubricant base oils are used to produce lubricants. The quality of a lubricant base oil can largely be evaluated independent of its final application, but the performance of a lubricant is judged by its performance in a specific application. 2.3.1 Boiling Range and Product Quality
The terminology associated with different boiling fractions from crude oil (Figure 2.7) can be confusing and to make matters worse, some of the products have overlapping boiling ranges. It may be even more disconcerting to realize that the boiling range associated with a particular product is not precisely defined. Some of this apparent confusion clears up when considering the history of crude oil refining, and hopefully the narrative in Section 2.4 throws light on the subject.
29
2 Refining and Refineries at a Glance Figure 2.7 Typical crude oil distillation fractions.
Vacuum residue 550
430 Boiling point temperature (°C)
30
Atmospheric residue
Vacuum gas oil
Heavy gas oil
360
Heavy cycle oil
320 Light gas oil 260
Light cycle oil
Distillate
220
Kerosene
180 160
Heavy gasoline Heavy naphtha
80
Light gasoline Light naphtha
20 0 −50
Liquid petroleum gas Gas
On a technical level, the origin of any product-related boiling range can be traced back to the quality requirements for that product. The boiling range indicates that the inclusion of any lighter or heavier material adversely affects one or more performance characteristics of the product. Consequently, as the quality requirements of a product change, so does the most appropriate boiling range. Unless the boiling range of a product is fixed by specification, the refiner is free to include materials outside the traditional boiling range of that product in the product. The caveat is that the refiner must still be able to meet the quality requirements. For most products, there is a minimum quality requirement and in the case of transportation fuels these quality requirements are set by the legislated fuel specifications. If the quality requirements are not met, the product cannot be sold into the market under that name. This also applies individually to any material that falls within the boiling range of that product. For example, a naphtha fraction may have the correct boiling range to be sold as a motor-gasoline, but it can only be sold as such if the naphtha meets the minimum quality standards set for motor-gasoline.
2.4 Evolution of Crude Oil Refineries
It is important to note that not all countries have the same specifications for the same products. A product of acceptable quality in one country may not be marketable in another country.
2.4 Evolution of Crude Oil Refineries
Much of the value that is associated with crude oil is not intrinsic to the crude oil itself but to the products that can be produced from it. The phenomenal success of Rockefeller can partly be attributed to his insight that crude oil has a price and that it has little value unless it is refined [14]. A purely historical account of the exploitation of crude oil and its ascendancy to become a key global resource is fascinating and well worth the read [14, 15]. The growth in demand for crude oil as an energy resource is closely linked to the history of innovation in the fields of transportation and refining. Here, the objective is not to focus on history per se, but rather to employ history as narrative to explain how crude oil refining and refineries developed and evolved over time. Soon after the production of crude oil started, it entered a period of rapid growth. World consumption of crude oil doubled every decade since 1880 until the ‘‘Oil Crisis’’ in 1973 (Figure 2.8). This unchecked growth in crude oil consumption was partly due to the comparative price stability of crude oil during this period; the crude oil price remained at US$1–2 per barrel. Over time, the market for crude-oil-derived products also changed dramatically. As a natural consequence, crude oil refineries had to change in design and complexity to keep up with the changes in the market. As crude oil became available from more and more sources, refineries had to cope not only with changes in products and product specifications but also with changes in raw materials and global refining economics.
Global crude oil consumption/kt/a
10 000 000
1 000 000 1973 "Oil Crisis" 100 000
10 000
1000
100 1870
Figure 2.8
1890
1910
1930 1950 Year
Global crude oil consumption from 1870 to 2010.
1970
1990
2010
31
32
2 Refining and Refineries at a Glance
2.4.1 First-Generation Crude Oil Refineries
In the middle of the nineteenth century, the market for lamp oil (kerosene) was already well established as an important power source for illumination. These were the days before electric lamps, which only started to be introduced in the 1880s [16]. The main source of lamp oil was whales. In 1850, a process for making ‘‘coal oil’’ was patented in Scotland. This process was developed to augment the dwindling supplies of whale oil by producing kerosene from coal pyrolysis. The ability to also produce kerosene from crude oil was demonstrated in 1854, by showing that kerosene can be obtained by crude oil distillation. The availability of crude oil was a limiting factor, but this started to change when Drake in 1859 demonstrated that one could drill for oil just as one drills for water. By 1862, kerosene from oil had completely displaced kerosene from coal [14]. There is some irony in this, since the replacement of a renewable resource (whales) by a nonrenewable resource (crude oil) saved the whales from being hunted to extinction. The first crude oil refineries consisted of a pot-still to boil the crude oil so that the middle-boiling kerosene fraction (180–300 ◦ C) could be collected and sold as lamp oil. Until 1910, all crude fractionation had been done by batch distillation. Only then did continuous-operating boilers appear and the first true distillation columns were built only in the 1920s. Initially, in the 1870s, many refiners were interested only in producing the maximum amount of kerosene from the crude oil, throwing away the various by-products that were produced. Andrews of Standard Oil was one of the more skillful refiners of his time and he worked hard to achieve more benefits from the by-products. The naphtha, which is lighter boiling than kerosene, was used as fuel to heat the crude oil stills. The heavy material that remained after the kerosene was recovered was sold to gas plants as alternative fuel to coal. Andrews even persuaded people to consider the heavy oils as lubricants instead of vegetable oils [14]. Around the same time, in 1872, Otto invented the internal combustion engine by employing gas as fuel. Subsequently, Daimler and Benz devised a more practical implementation of the invention. They realized that the naphtha fraction from crude oil distillation would be a better fuel than gas and adapted the Otto engine for such use. The spark-ignition system and water cooling for automobiles can both be credited to Benz, and it laid the necessary foundation for modern automobile engine technology [16]. This created a market for the light-boiling fraction of the first-generation distillation-only refineries. These two distillation products, naphtha (gasoline or petrol) and kerosene (lamp oil), were in high demand from 1890 onward, but really took off with the boom in the automobile industry after the introduction of the Ford model T in 1908. Most European automobiles were built for sportsmen and enthusiasts, and not for the ordinary man who just needed transport. Ford had the insight to provide transportation for the common man and 15 million Ford model T’s were sold between 1908 and 1927. It took longer for the heavier boiling fractions (>300 ◦ C) from crude oil to gain the same market pull as with the naphtha (<180 ◦ C) for gasoline and kerosene (180–300 ◦ C) for lamp oil. Diesel invented the compression-ignition engine in 1893, but it took some time before its use became widespread. Despite its better thermodynamic efficiency than a spark-ignition Otto engine, a compression-ignition engine is heavier and more expensive. This made its use for
2.4 Evolution of Crude Oil Refineries
automobiles less attractive, and it was better suited for heavy vehicles and ships. Fuel oil, both ‘‘diesel oil’’ and a residual fuel called ‘‘boiler fuel,’’ started replacing coal as energy source on ships in the 1900s [17]. It was also possible to use the heavier material from the pot-still distillation for lubrication. 2.4.2 Second-Generation Crude Oil Refineries
The problem faced by refiners at the beginning of the twentieth century was that the market did not require the products in the same ratio as they were obtained from crude oil distillation. After the Wright brothers demonstrated powered flight in 1903, there was rapid growth in the aviation industry in the 1910s, which gave significant impetus for the development of more powerful engines. The automobile manufacturing industry also became more advanced. Improvements in engine design in turn required a better quality fuel to deal with the performance requirements that were imposed by engines operating at higher compression ratios. The ability of a fuel to withstand autoignition when compressed, also called its antiknock behavior, became an important fuel quality measure. Thus, the concept of an octane number was born (Section 13.3.1). Highly branched alkanes, branched alkenes, and aromatics were found to possess good antiknock behavior and were desirable components in gasoline. Straight-run naphtha was no longer good enough to be used as gasoline. These two aspects, namely, the difference between market demand and distillation yield and the need for better fuel quality, were destined to become recurring themes in crude oil refining. Upgrading crude oil by distillation alone (first-generation refinery) was no longer adequate and it heralded the start of the second phase in refinery development. The market for petroleum-derived products was changing rapidly. Electricity gradually replaced town gas and lamp oil as the main energy source for lighting. This eroded the market for kerosene. In practice, this caused crude oil refiners to blend the bottom part of the kerosene fraction into motor-gasoline. This is why older literature on refining often refers to motor-gasoline as the C5 -204 ◦ C (C5 -400 ◦ F) fraction. The heavier fraction of the kerosene was not separated from the heavy product recovered by atmospheric distillation and the 204–427 ◦ C (400–800 ◦ F) cut became known as distillate, or diesel fuel used for compression-ignition engines. Powered flight became commonplace. Aircraft made use of spark-ignition engines and a market for aviation-gasoline evolved in parallel to that for motor-gasoline. The quality demands placed on aviation-gasoline were more severe than those for motor-gasoline. Aircraft engines operated at high compression ratios and required very high-octane gasoline. Aviation-gasoline also needed to be clean and stable on account of the disastrous consequences associated with in-flight engine failure. In response to these demands, second-generation refineries are characterized by the introduction of conversion processes. Conversion processes were able to change the quality of the crude oil derived products, as well as their boiling range. During this period, many important refining technologies, such as hydrotreating, isomerization, oligomerization, cracking, and alkylation (Table 2.2), were developed [6, 18–20]. Although the development of refineries, catalysts, and the chemical engineering discipline started off independently, they quickly came together to create second-generation refineries.
33
34
2 Refining and Refineries at a Glance Table 2.2
Development and commercialization of some important refining technologies.
Year
Process
Original catalyst
1920s 1933 1934 1936 1930s 1940 1942 1942 1949 1959 1964 1967
Hydrogenation Paraffin isomerization Oligomerization (catalytic polymerization) Catalytic cracking (Houdry) Aliphatic alkylation Catalytic cracking Aliphatic alkylation Fluid catalytic cracking Catalytic reforming Hydrocracking Zeolite fluid catalytic cracking Multimetallic reforming
Ni–kieselguhr AlCl3 Solid phosphoric acid (SPA) Montmorillonite H2 SO4 Amorphous silica-alumina (ASA) HF Amorphous silica-alumina (ASA) Pt–Al2 O3 Ni/Mo– and Co/Mo–silica–alumina Faujasite (Y-zeolite) Pt/Re– and Pt/Ir–Al2 O3
Second-generation refineries generally used a simple topping–reforming refinery flowscheme (Figure 2.9). This is basically a first-generation refinery with a thermal reforming unit added. During thermal reforming, the low-octane-number, heavier alkanes are thermally cracked at temperatures above 540 ◦ C to increase the octane number of the product by decreasing the alkane content and increasing the aromatic and alkene content (Section 22.2) [21]. Depending on the refinery, the design sometimes included additional upgrading units: visbreaking (mild thermal cracking to improve the fluidity of heavy fuel fractions), thermal or catalytic cracking to convert some of the residues into lighter products, and oligomerization to convert the light alkenes from thermal reforming, visbreaking, and cracking into olefinic motor-gasoline. The thermal upgrading of the heavier fractions allowed them to be used as fuel oil, and, depending on the raw material used, also as lubrication oils and greases. Atmospheric residue further served Fuel gas LPG recovery Atmospheric distillation
Gas Naphtha
Liquid petroleum gas
Acid gas treatment Thermal reforming
Gasoline
Oligomerization Crude oil
Diesel fuel
Distillate Cracker
Residue Figure 2.9
Visbreaker
Fuel oil
Simple second-generation topping–reforming refinery typical of the pre-1950s.
2.4 Evolution of Crude Oil Refineries
as feed for asphalt production. Asphalt and blown asphalt served as important materials for road construction [22]. Details on thermal refining technologies can be found in contemporary textbooks [23, 24]. One of the by-products of thermal upgrading in second-generation refineries was the gaseous hydrocarbons. Methane production was not seen as detrimental, because light hydrocarbon gases could be sold as town gas. The problem was the heavier gases and especially the C3 –C4 gases. These gaseous hydrocarbons tend to condense under pipeline conditions that were used for town gas distribution, but were too volatile to be included in motor-gasoline. Initially, refiners tried to blend as much of the C3 –C4 hydrocarbons as possible into the motor-gasoline, which became dangerously volatile on account of this practice. In 1904, the liquid petroleum gas (LPG) industry developed in Germany, where ‘‘Blaugas’’ was sold in pressurized containers. This ultimately led to the development of the LPG industry to market the ‘‘troublesome’’ C3 –C4 hydrocarbons from refineries. The development of the LPG industry was slow before 1932. This was partly due to a scarcity of raw materials for the construction of pressurized containers and partly due to a lack of understanding of the thermodynamic properties of propane–butane mixtures [25]. With the increased demand for high-octane aviation-gasoline during the Second World War, refinery configurations that specifically incorporated technology for producing high-octane gasoline range molecules evolved. Alkene oligomerization using solid phosphoric acid, as well as aliphatic alkylation using H2 SO4 and later HF, became important technologies in this regard. However, after the Second World War the market for high-octane aviation-gasoline decreased significantly, leaving refineries that invested in such upgrading units with excess high-octane gasoline capacity. These refineries were well ahead of their time and in some cases such refineries already evolved to become the equivalent of fourth-generation refineries (Section 2.4.4). Many refineries did not make use of all the technologies that were on offer, but rather opted to add tetraethyl lead (TEL) to the gasoline to improve the octane number sufficiently for use as motor-gasoline. In fact, the development of alkyl lead as octane booster in the 1920s allowed many small refiners to keep on operating simple second-generation topping–reforming refineries. In 1949, catalytic reforming technology was introduced. This brought a significant change in refining and it was the enabling technology that was needed to make hydrogen more readily available for refining. The thermal reforming units of the older topping–reforming designs were gradually replaced by catalytic units. Thermal cracking units were likewise replaced by catalytic cracking units. These were not the only changes. The composition of the motor-gasoline reflected changes in refining technology as refineries relied more on catalytic conversion and less on thermal conversion (Table 2.3) [21]. The 1950s also saw a significant shift in aircraft design from spark-ignition engines to turbine engines. Kerosene turned out to be an ideal fuel for turbine engines, which allowed refiners to reallocate the straight-run material. The heavy gasoline fraction and light distillate fractions could now be used for jet fuel. This was a fortuitous development. The production of jet fuel led to an improvement in the quality of the motor-gasoline and diesel fuel, purely by virtue of distillation. With hydrogen from the catalytic reformer and the kerosene fraction being allocated to jet fuel, second-generation topping–reforming refineries got a new lease on life. It was possible to survive most of the fuel quality changes until the beginning of the 1970s by employing TEL addition in combination with sweetening (reactive thiol removal) and hydrogenation (Figure 2.10) [2]. As in the earlier topping–reforming designs, additional units may have been included as shown in Figure 2.9 and reflected in Table 2.3.
35
36
2 Refining and Refineries at a Glance Table 2.3 Changes in the composition and quality of motor-gasoline over time in the US refineries as catalytic technologies were introduced.
Description Blending components (vol%) Straight-run naphtha Thermal conversion processes Alkylate and polymer gasolines Catalytic cracking Catalytic naphtha reforming Motor-gasoline quality Research octane number (RON), unleaded Lead addition (g Pb l−1 ) RON, leaded
1940
1950
1960
1972
50 46 2 2 0
40 32 7 20 1
19 10 10 31 30
12 4 13 38 33
64.3 0.40 74.6
75.3 0.58 85.1
86.2 0.53 94.0
89.3 0.61 96.8
LPG recovery Atmospheric distillation
Gas Naphtha Kerosene
Crude oil
Gas oil
Acid gas treatment Naphtha hydrotreater
Catalytic reforming
Fuel gas Propane Butane
Gasoline
Sweetening
Jet fuel
Distillate hydrotreater
Diesel fuel
Residue Figure 2.10 Second-generation topping–reforming refinery (without residue upgrading) of the period 1950–1970.
2.4.3 Third-Generation Crude Oil Refineries
The transition from second-generation to third-generation refineries was mainly related to the inclusion of vacuum distillation and residue upgrading capacity. This change was brought about by one of the most important historical events in the crude oil industry, the ‘‘Oil Crisis.’’ In the early 1970s, crude oil prices were at their lowest, with a barrel of Arabian Light crude costing less than US$2. When the Oil Crisis hit at the end of 1973, crude oil prices increased sixfold in a short period of time. In January 1974, a barrel of Arabian Light crude was around US$12, which was followed by another significant increase at the end of the decade, pushing the crude oil price beyond US$30 per barrel. This is an order of magnitude change in less than a decade. Crude oil as a raw material was clearly no longer a cheap commodity. Before we
2.4 Evolution of Crude Oil Refineries
consider the impact of the Oil Crisis on crude oil refinery design, let us have a brief look at what precipitated this momentous event. A small number of very large international oil companies, informally working together, controlled most of the world’s markets for crude oil in the 1950s. By 1959, these companies were under economic pressure from competition with each other and outsiders to maintain their stranglehold on the crude oil market. To increase their profits, these companies decided to reduce their tax commitments to the producing countries. This was achieved by posting lower market prices for the crude oil on which tax was calculated. The reaction was unexpectedly strong. One of the unintended consequences was the formation of the Organization of Petroleum Exporting Countries (OPEC) at the instigation of Venezuela. During the 1960s, internal disagreements among member countries failed to result in any coherent collective action, but OPEC survived mostly as a result of the political solidarity felt by the Arab countries. It was this solidarity that caused the production cuts in 1973, which led to the Oil Crisis. It was seen as a way to retaliate at the pro-Israel Western world after the Six Day War in 1967 and, finally, the Yom Kippur War that started on 6 October 1973 [26]. Irrespective of how history played out, it is clear that price of crude oil remained artificially low and it was only a matter of time before market forces would cause the price to increase. Crude oil is a nonrenewable energy source and there is, by definition, only a finite supply. Under pressure of continuous demand, the exploitation of any natural resource will increase over time, but because it is a finite resource, it will ultimately pass through a maximum, which will be followed by a natural decline (Figure 2.11). This is the fundamental reality that underpins the concept of ‘‘peak oil,’’ which seeped into public awareness through the work of Hubbert. Hubbert successfully predicted that crude oil production in the United States would reach its peak between 1967 and 1973. This point was indeed reached in 1970 and psychologically the Oil Crisis came at a sensitive time. 15
Q (t ) = 80
Q total 1 + a.e−b.t
12
60
9
40
6
20
3
0
0 0
2
4
6
8
10
12
Time (arbitrary units) Figure 2.11 Annual production () and cumulative production (•) of a finite resource, such as crude oil, according to the Hubbert-model description.
14
16
Annual production, Q (t ) (% of total)
Cumulative production (% of total)
100
37
38
2 Refining and Refineries at a Glance
LPG recovery Acid gas treatment Atmospheric distillation
Gas Naphtha Kerosene
Crude oil
Residue
Vacuum distillation
Fuel gas Propane Butane
Distillate
Naphtha hydrotreater
Reforming
Gasoline
Sweetening
Kerosene
Distillate hydrotreater
Diesel
Fluid catalytic cracking
Visbreaking
Sweetening
Fuel oil
Figure 2.12 Third-generation topping–reforming–cracking– visbreaking refinery that included vacuum distillation and more extensive residue upgrading. Although a coking unit has not been indicated, many refineries included coking instead of, or in addition to, visbreaking.
The increase in oil price affected not only the price of white products (motor-gasoline, jet fuel, and diesel fuel) but also the price of fuel oil. This resulted in a drop in the use of fuel oil and a slowdown in the growth of white product consumption (Figure 2.8). Refiners now had to produce more white products and less fuel oil, but light crude oils were becoming scarcer and more costly. Since the residue fraction of most crude oils constitutes close to 50% of the total volume of crude oil, better use of the residue fraction became an economical imperative. This gave rise to the third generation of crude oil refineries that included significant residue upgrading capacity (Figure 2.12). Such refineries included a vacuum distillation column for residue fractionation and typically used topping–reforming–cracking or topping–reforming–cracking–visbreaking configurations. Fundamentally, the responsible use of all crude oil fractions was long overdue. The economic pressure that caused the increased emphasis on residue upgrading was just the catalyst. Many smaller refineries did not have the infrastructure or capital to make the transition from secondto third-generation refineries and, in the long run, were closed down. Conversely, some refiners opted to install vacuum distillation capacity and residue upgrading decades before the Oil Crisis. Surprisingly, in the United States many of these second-generation refineries survived the 1970s without adding residue upgrading capability. This was due to the local political situation and specifically the implementation of the Emergency Petroleum Allocation Act. Refiners with
2.4 Evolution of Crude Oil Refineries
access to local cheaper crude oil had to pay a fee to refiners processing more expensive imported crude oil. A system of ‘‘entitlements’’ was implemented, which favored small refiners that did not have the capacity to refine their ‘‘entitlement’’ and could sell their ‘‘entitlement’’. Despite the panic, in real terms the motor-gasoline price in the United States increased to 1955 prices and there was no incentive to curb demand. This artificial politically engineered protection of inefficient refiners was corrected in 1981. As a result, many refineries in the United States were closed down, but with little real loss in refining capacity [27]. Nevertheless, not all refineries have opted to change with the times, and intermediate product exchange agreements were used by some refiners to stay in business without spending the capital to move to the next generation of refining. 2.4.4 Fourth-Generation Crude Oil Refineries
The evolution of fourth-generation refineries was a direct consequence of vehicle emission control standards. It became quite clear that air quality was rapidly deteriorated because of automotive emissions, which was especially apparent in densely populated cities. In the United States, the first Clean Air Act was passed in 1963. The Clean Air Act established the authority of the federal government to regulate air pollution. Emission control standards for hydrocarbon and carbon monoxide emissions were formalized in the Clean Air Act Amendments of 1970, with legislative enforcement commencing in 1975. Automobile manufacturers were faced with a challenge and a new emission control device was invented, the tail pipe catalytic converter. The catalytic converter did a splendid job of meeting the new emission regulations, but there was a snag. The catalytic converter performance was degraded by 50% in less than 35 000 km of use. The performance loss could be correlated to the quality of the motor-gasoline used as fuel [28]. The ubiquitous use of TEL in motor-gasoline to meet the octane number specifications was about to come to an end. TEL was not only a human health hazard but also a poison to tail pipe catalytic converters. To meet the new emission standards, motor-gasoline had to be unleaded. This requirement translated into law and thus unleaded motor-gasoline became part of the requirements to meet the new vehicle emission standards. The addition of TEL to motor-gasoline provided an octane number increase of five to seven units. In refineries, the need to forgo the use of TEL forced refiners to have a serious look at the quality of the products being produced. The octane number shortfall had to be overcome by refining in order to increase the quality of motor-gasoline so that the octane number specification could be met without TEL. The threat to refinery operation was not immediate, since the phaseout of lead was a gradual process. Initially, the market for unleaded motor-gasoline was small, allowing many refineries to tweak their blending to produce both unleaded and leaded motor-gasoline, or remain suppliers of leaded motor-gasoline only. Yet, the change was inexorable and refineries had to make the transition to unleaded motor-gasoline. This implied a change in the refining process. Refineries that invested in technologies to meet the demand for high-octane aviation-gasoline in the 1940s were well positioned to make the transition to unleaded motor-gasoline. It was quickly realized that alkane quality was the key to meeting the octane demands and that isomerizing C4 –C6 alkanes provided a cheap way to increase the octane number.
39
40
2 Refining and Refineries at a Glance
Fuel gas LPG recovery Atmospheric distillation
Acid gas treatment
Gas
C5/C6 isomerization
Naphtha
Naphtha hydrotreater
Reforming
Distillate
Residue
Gasoline Kerosene
Kerosene
Crude oil
Propane Butane
Diesel fuel
Distillate hydrotreater Fluid catalytic cracking
Sweetening
Residue hydroconversion
Etherification Oligomerization
Vacuum distillation
Coking
Visbreaking
Deasphalting
Alkylation Fuel oil Coke Asphalt
Figure 2.13 Generic fourth-generation refinery that includes significant upgrading of all crude oil fractions. There are many different possible configurations within a refinery design.
The alkenes produced by the high-temperature residue upgrading units also opened synthetic routes for the production of high-octane motor-gasoline. Three technologies, namely, aliphatic alkylation, naphtha hydroisomerization, and etherification, emerged as important for the production of high-octane unleaded motor-gasoline. Fourth-generation refineries (Figure 2.13), typical of the 1990s, used topping–reforming–cracking–visbreaking–alkylation–isomerization or topping–reforming–cracking–coking–alkylation–isomerization schemes, which frequently included other refining technologies, such as etherification, too. Some refineries employed oligomerization units, which also helped to meet the octane requirements, but alkenes became a valued commodity in refineries. The petrochemical industry evolved in parallel with the transportation fuel industry. Fuel refineries became an important source of chemicals, such as propene [29]. Propene and other petrochemicals were no longer available for fuels refining if the refinery could sell them to a petrochemical producer at a higher price. However, as we see later, history led many refiners down a path of refinery development that would ultimately end with new oligomerization capacity. Air quality improvement by lowering automotive emissions extended well beyond the phasing out of lead. Research initiatives, such as the ‘‘Air Quality Improvement Research Program’’ (AQIRP), which was a combined effort between the automotive and oil industries, evaluated
2.4 Evolution of Crude Oil Refineries Table 2.4
Motor-gasoline properties that affect vehicle emissions and air quality.
Fuel property Hydrocarbon Lower vapor pressure Lower sulfur content Lower alkene content Lower aromatic content Higher oxygenate content
Less Less – – –
Exhaust emissions CO NOx Less Less – – Less
– Less Less – –
Urban ozone Toxics – Less – Less –
Less Less Less – –
ways to make motor-gasoline a less polluting fuel (Table 2.4) [28]. The study provided legislators with a scientific framework for driving down automotive emissions, and from the 1990s onward these recommendations made their way into motor-gasoline specifications. The implications for refineries were better octane numbers at reduced levels of sulfur, alkenes, aromatics, and volatile hydrocarbons. The addition of oxygenates to motor-gasoline became a saga on its own. The addition of oxygenates was aimed at reducing CO emissions (Table 2.4). In Europe, the industrial production of methyl tertiary butyl ether (MTBE) as fuel ether started in 1973 [30]. This development was mainly in response to the phaseout of TEL, since MTBE has a high octane number and could compensate for the octane deficiency. As time went by, the inclusion of oxygenates in motor-gasoline was regulated by the EN228 European motor-gasoline specifications, but the inclusion of oxygenates never became mandatory. In the United States, the history of oxygenate inclusion in motor-gasoline played out differently. When the Clean Air Act Amendments was signed into law in 1990, oxygenated gasoline and reformulated gasoline (RFG) were introduced with a mandatory minimum oxygenate content. On a scientific level, oxygenates were required to reduce CO emissions, but on the political level mandatory oxygenate inclusion was a legislation that was in favor of the corn-ethanol lobby. However, refiners and fuel distributors found that the high vapor pressure and demixing tendency of ethanol made it a less desirable oxygenate than MTBE. Furthermore, the use of MTBE rather than ethanol resulted in lower engine emissions, making it a sensible choice. This resulted in a tremendous increase in installed etherification capacity [31]. Yet, failure to enforce the laws regulating underground storage facilities led to MTBE ending up in ground water as a result of leaking storage tanks. In 1999, politicians geared up to ban the use of MTBE in motor-gasoline, which happened shortly thereafter [32, 33]. This left many refiners with defunct MTBE capacity and heralded a return to alkene oligomerization as high-octane technology by retrofitting the MTBE units for isobutene dimerization. The different motor-gasoline blending components that may be available in a fourth-generation crude oil refinery are listed in Table 2.5 [4]. It is pointed out that some of the properties are dependent on unit operation, cut point, or feed composition. motor-gasoline was by no means the only source of automotive emissions. The use of diesel fuel also contributed to the overall deterioration of air quality. Some of the problems associated with motor-gasoline (Table 2.4) were not an issue with diesel fuels. Diesel fuel is a heavier
41
42
2 Refining and Refineries at a Glance Table 2.5 Properties of motor-gasoline components that may be found in a fourth-generation crude oil refinery.
Refinery stream Straight-run blending components Butanes Light straight-run (LSR) naphtha (C5 -80 ◦ C) Heavy straight-run naphtha Products from naphtha refining Isomerate, once-through LSR naphtha Reformate, 94 RON Reformate, 98 RON Alkylate, C3 feed Alkylate, C3 –C4 feed Alkylate, C4 feed Polymer gasoline Products from heavy-end upgrading Hydrocracker light naphtha Hydrocracker heavy naphtha Coker naphtha Visbreaker, light thermal naphtha FCC (fluid catalytic cracker) naphtha, 95–150 ◦ C FCC hydrogenated light naphtha FCC hydrogenated heavy naphtha Oxygenated blending components Ethanol 2-Methoxy-2-methylpropane (MTBE) 2-Ethoxy-2-methylpropane (ETBE) 2-Methoxy-2-methylbutane (TAME)
RON
MON
RVP (kPa)
93 66 62
92 62 59
420 77 7
83 94 98 91 95 97 97
81 84 87 87 93 96 84
93 19 15 39 34 32 60
83 68 67 80 92 83 90
82 67 60 73 77 81 81
89 8 25 68 10 96 3
132 118 118 111
106 101 102 98
76 62 28 10
The research octane numbers (RONs), motor octane numbers (MONs), and Reid vapor pressures (RVPs) are all typical blending values.
fuel and volatile hydrocarbon emissions related to fuel vapor pressure were of no concern. The alkenes in motor-gasoline are produced by residue upgrading and diesel fuel is usually devoid of alkenes, which implies that alkene content is not an issue. However, because diesel fuel is heavier than motor-gasoline, it inherently has a higher sulfur content (Table 2.1). The maximum allowable sulfur content in diesel fuel was decreased over time. At present, in many countries, the maximum sulfur content of diesel fuel is less than 10 µg·g−1 . This made the sulfur removal strategy of refineries and severity of hydrotreating, in particular, critical to the production of on-specification diesel fuel. As a result, fourth-generation refineries (Figure 2.13) include more extensive hydroprocessing of the heavier fractions. An unintended beneficial consequence of the more stringent sulfur specification was that it facilitated many other improvements in diesel fuel quality. Deeper hydrodesulfurization caused more aromatic saturation, and, therefore, improved the cetane number. Cetane number is a quality characteristic of diesel fuel and it is a measure of how susceptible the fuel is to
2.4 Evolution of Crude Oil Refineries
compression ignition (Section 15.3.1). The upper boiling point temperature of diesel fuel was reduced to limit hydrogen use. Many of these changes were legislated, but in refinery design it was in any case incorporated to facilitate sulfur removal. Hydrogen availability became a bottleneck in many fourth-generation crude oil refineries. This is even more so in refineries serving the European market, where the demand ratio for motor-gasoline and diesel fuel is biased to diesel fuel. The catalytic reforming unit is often the only source of hydrogen in the refinery. With the maximum aromatic content of motor-gasoline being regulated, the amount of hydrogen that can be produced is limited. As the ratio of motor-gasoline to diesel fuel shifts toward diesel fuel, the need for hydrogen increases, but the amount that can be provided by catalytic reforming decreases. 2.4.5 Petrochemical Refineries
The evolution of crude oil refineries (Sections 2.4.1–2.4.4) was mainly influenced by the development of the fuel industry, and the discussion therefore focused mainly on fuels. The evolution of the petrochemical industry took place in parallel. One can chart the changes in the petrochemical industry to show how the demand for chemicals changed over time, but the petrochemical market is much smaller than the market for fuels [34]. Consequently, petrochemicals were less influential than fuels in terms of crude oil refinery design. In terms of market size, the most important chemical commodities are the light alkenes, namely, ethene and propene, and the light aromatics, namely, benzene, toluene, and the xylenes (BTXs) (Table 28.1). Petrochemical facilities are organized along these two compoundclass-based value chains [1]: 1) C2 –C4 alkenes. Alkenes are derived from cracking operations, which may be thermal or catalytic. Thermal cracking, fluid catalytic cracking, steam cracking, visbreaking, and coking are all potential sources of short-chain alkenes. These short-chain alkenes form the building blocks for many other commodity chemicals. 2) C6 −C8 aromatics. BTX aromatics are produced by catalytic reforming, which is followed by an aromatics extraction. These aromatics form the commodity base from which other derivatives can be produced. It is clear that the main classes of petrochemicals are coproduced during crude oil refining intended to produce fuels. Yet, the concept of a crude-oil-based stand-alone chemical refinery for the production of only high-value chemicals and not fuels is appealing. An example of such a stand-alone petrochemical refinery is shown in Figure 2.14 [35]. Most of the main petrochemical commodities have carbon numbers that are less than C10 . However, even in very light crude oils, more than half the material is heavier than C10 and much of the associated gas that would have been valuable in petrochemical production is no longer associated with the crude oil. Substantial refining effort is required to convert all the heavy crude oil components into lighter petrochemicals. Depending on the crude oil, it may be possible to refine the atmospheric residue into lubricant base oil (Section 2.4.6). Stand-alone petrochemical facilities for the production of alkenes and aromatics usually make use of light hydrocarbons and not unrefined crude oil as the feed material. The light hydrocarbons may be obtained from natural gas and associated natural gas liquids, or it may be obtained from
43
44
2 Refining and Refineries at a Glance
Atmospheric distillation
Fuel gas
LPG recovery
Gas
Alkene refinery Alkene recovery Crude oil
Thermal cracking
Gas oil
Ethene Propene Butenes Butadiene
Hydrotreater
Fuel oil
Residue Aromatics refinery Aromatics extraction
Naphtha reforming
Benzene Dealkylation or disproportionation
Toluene Xylenes
Naphtha Figure 2.14
Naphtha hydrotreater
Heavy aromatics
Stand-alone petrochemical refinery.
upgraded crude oil, such as partially refined crude, chemical intermediates, and commodity chemicals. 2.4.6 Lubricant Base Oil Refineries
The refining of crude oil to produce lubricating base oil is on the opposite side of the refining spectrum to petrochemical refineries, although both are nonenergy refineries. Whereas petrochemical refineries produce products mainly in the lower than C10 carbon number range, base oil refineries produce products mainly in the heavier than C20 (atmospheric residue) range. A stand-alone base oil refinery has a problem that is analogous but inverse to that of a stand-alone petrochemical refinery. Unless a crude oil is extremely heavy, with very little straight-run naphtha and distillate, it is difficult to economically justify a stand-alone base oil refinery with no coproduction of fuels. The most important properties of base oils are viscosity, viscosity index, cold-flow properties, and stability. Two distillation steps are typically required for base oil production from crude oil. Atmospheric distillation is used to produce an atmospheric residue that serves as feed for base oil production, and vacuum distillation is used to produce fractions having a specific viscosity range. The vacuum gas oil must still be refined further in order to produce lubricant base oil. The three main approaches to lubricant base oil refining have not changed much over time:
2.4 Evolution of Crude Oil Refineries
Atmospheric distillation
Crude oil
Fuels refinery
Transportation fuels
Vacuum distillation
Solvent dewaxing
Lubricant base oil Waxes
Solvent deasphalting Figure 2.15
Asphalt
Solvent-based lubricant base oil refining.
1) Eliminate undesirable compounds from the crude oil by separation to leave only compounds with good base oil character. There are three major classes of undesirable substances that must be dealt with, namely, asphaltenes that may precipitate, waxes that have poor cold-flow properties, and compounds that have low viscosity index or poor stability [36]. Historically, base oil manufacturing involved treatment with sulfuric acid, sodium hydroxide, and clay, which were effective in the removal of asphaltenes, chemically unstable compounds, and low-viscosity-index material. This treatment is followed by the elimination of wax in a separate processing step to produce the lubricant base oil. Owing to the complex action of sulfuric acid and the by-products produced, this approach has been replaced by physical methods employing solvent-based separation, such as deasphalting, extraction with a polar solvent, and solvent dewaxing. A general solvent-based processing scheme is shown in Figure 2.15 [23]. The same basic solvent refining operations are still in use [4]. 2) Convert undesirable compounds into desirable ones by appropriate conversion processes. To further improve base oil properties beyond the properties related to the crude oil, hydroprocessing and catalytic dewaxing are required. Catalytic dewaxing involves the hydrocracking and hydroisomerization of the n-alkane waxes in the vacuum gas oil, which reduces the sulfur content, improves the cold-flow properties, and increases the viscosity index. 3) Render undesirable compounds harmless by neutralizing their negative impact through the addition of an additive. This is an inevitable part of lubricant production. This is not really a separate refining solution to produce lubricant base oils per se. It requires a very special raw material to directly convert straight-run vacuum gas oil into lubricant base oil by additive addition only. Over time, the quality and performance requirements for base oils increased, like those of transportation fuels. With successive generations of fuel refineries (Sections 2.4.1–2.4.4), the refining that was employed to produce higher quality lubricant base oils became more sophisticated. Conversion-based lubricant base oil production processes advanced with progress in catalysis. Yet, many of the older physical refining processes are still in use.
45
46
2 Refining and Refineries at a Glance
References 1. Jones, D.S.J. (2006) in Handbook of Petroleum
2.
3.
4.
5.
6.
7.
8. 9.
10. 11.
12.
13.
14. 15. 16. 17.
Processing (eds D.S.J. Jones and P.R. Pujad´o), Springer, Dordrecht, pp. 1–45. Wauquier, J.-P. (ed.) (1995) Petroleum Refining Crude Oil. Petroleum Products. Process Flowsheets, Vol. 1, Editions Technip, Paris. Speight, J.G. (2007) The Chemistry and Technology of Petroleum, 4th edn, CRC Press, Boca Raton, FL. Gary, J.H., Handwerk, G.E., and Kaiser, M.J. (2007) Petroleum Refining. Technology and Economics, 5th edn, CRC Press, Boca Raton, FL. Kalichevsky, V.A. and Kobe, K.A. (1956) Petroleum Refining with Chemicals, Elsevier, Amsterdam. Leprince, P. (ed.) (2001) Petroleum Refining, Conversion Processes, Vol. 3, Editions Technip, Paris. March, J. (1985) Advanced Organic Chemistry, 3rd edn, John Wiley & Sons, Inc., New York, pp. 220–222. Ho, T.C. (2004) Deep HDS of diesel fuel: chemistry and catalysis. Catal. Today, 98, 3–18. Fan, T.-P. (1991) Characterization of naphthenic acids in petroleum by fast atom bombardment mass spectrometry. Energy Fuels, 5, 371–375. Bailey, G. and Palmer, J.W. (2002) Crude oil corrosion. Hydrocarbon Eng., 7 (3), 89–90. ´ Laredo, G.C., L´opez, C.R., Alvarez, R.E., Castillo, J.J., and Cano, J.L. (2004) Identification of naphthenic acids and other corrosivity-related characteristics in crude oil and vacuum gas oils from a Mexican refinery. Energy Fuels, 18, 1687–1694. Dunning, H.N., Moore, J.W., Bieber, H., and Williams, R.B. (1960) Porphyrin, nickel, vanadium, and nitrogen in petroleum. J. Chem. Eng. Data, 5, 546–549. Gallup, D.L., Curiale, J.A., and Smith, P.C. (2007) Characterization of sodium emulsion soaps formed from production fluids of Kutei Basin, Indonesia. Energy Fuels, 21, 1741–1759. Tugendhat, C. and Hamilton, A. (1975) Oil. The Biggest Business, Eyre Methuen, London. Yergin, D. (2003) The Prize. The Epic Quest for Oil, Money and Power, Free press, New York. Larsen, E. (1969) A History of Invention, J. M. Dent & Sons, London. Winkler, M.F. (2003) in Fuels and Lubricants Handbook: Technology, Properties, Performance, and Testing (ed. G.E. Totten), ASTM, West Conshohocken, PA, pp. 145–168.
18. Ipatieff, V.N., Corson, B.B., and Egloff, G.
19.
20.
21.
22. 23. 24. 25.
26.
27. 28.
29.
30.
31.
32. 33.
34.
(1935) Polymerization, a new source of gasoline. Ind. Eng. Chem., 27, 1077–1081. Iverson, J.O. and Schmerling, L. (1958) in Advances in Petroleum Chemistry and Refining, Vol. 1 (eds K.A. Kobe and J.J. Mc Ketta Jr.), Interscience, New York, pp. 336–385. Bartholomew, C.H. and Farrauto, R.J. (2005) Fundamentals of Industrial Catalytic Processes, 2nd edn, John Wiley & Sons, Inc., Hoboken. Sterba, M.J. and Haensel, V. (1976) Catalytic reforming. Ind. Eng. Chem. Prod. Res. Dev., 15, 2–17. Kastens, M.L. (1948) Paving asphalt from California crude oil. Ind. Eng. Chem., 40, 548–557. Nelson, W.L. (1949) Petroleum Refinery Engineering, 3rd edn, McGraw-Hill, New York. Bell, H.S. (ed.) (1959) American Petroleum Refining, 4th edn, D. Van Nostrand, Princeton, NJ. Falkiner, R.J. (2003) in Fuels and Lubricants Handbook: Technology, Properties, Performance, and Testing (ed. G.E. Totten), ASTM, West Conshohocken, PA, pp. 31–59. Odell, P.R. (1975) Oil and World Power. Background to the Oil Crisis, 3rd edn, Taplinger, New York. Aalund, L.R. (2005) Phantom refining capacity. Oil Gas J., 103 (43), 15. Colucci, J.M. (2004) Fuel quality – an essential element in vehicle emission control. Proceedings of the ASME Internal Combustion Engine Division Technical Conference, Long Beach, CA, 2004. Walther, M. (2003) Refinery sources will fill the future ‘propylene gap’. Oil Gas J., 101 (4), 52–54. Ancillotti, F. and Fattore, V. (1998) Oxygenate fuels: market expansion and catalytic aspect of synthesis. Fuel Process. Technol., 57, 163–194. Swain, E.J. (1999) U.S. MTBE production at a record high in 1998. Oil Gas J., 97 (24), 99–101. Parkinson, G. (1999) All sides pumped up for MTBE ban. Chem. Eng., 106 (6 ), 49. Lamberth, R. (2004) 2003 was a year of transition for the MTBE, fuels industry. Oil Gas J., 102 (2), 52–58. Davis, S. (2008) Petrochemical Industry Overview, SRI Chemical Economics Handbook. Marketing Research Report Section 350, Stanford Research Institute, Menlo Park, CA.
References 35. Chadwick, J.L. (1977) Economics of Chemical
Refineries, SRI Process Economics Program Report 107, Stanford Research Institute, Menlo Park, CA.
36. Kalichevsky, V.A. (1938) Modern Methods of Re-
fining Lubricating Oils, ACS Monograph Series, Vol. 76, Reinhold, New York.
47
49
Part II Production of Fischer–Tropsch Syncrude
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
51
3 Synthesis Gas Production, Cleaning, and Conditioning 3.1 Introduction
Any carbon source can potentially be used as raw material for the production of syngas (CO and H2 ). The preparation of syngas as feed material for Fischer–Tropsch synthesis is by far the most expensive part of any feed-to-liquids (XTL) facility. When the raw material is a solid, such as coal, the delivery of clean and conditioned syngas to Fischer–Tropsch synthesis can easily account for more than 70% of the total capital cost. The relative capital cost for natural gas conversion to syngas is less, and natural gas to syngas conversion is technologically less challenging. The preparation of syngas as feed material for Fischer–Tropsch synthesis is a specialized topic of tremendous scope. Chapter 1 provided an overview of the steps involved in converting the raw material into a feed material for syngas production, as well as the basic steps involved in syngas production, cleaning, and conditioning. The purpose of this chapter is to consider syngas production, cleaning, and conditioning in more detail and specifically highlight the aspects that will affect refining.
3.2 Raw Materials 3.2.1 Natural Gas
Natural gas is obtained from natural gas reservoirs, which is called nonassociated gas, and as a by-product from crude oil production, which is called associated gas. The associated gas is obtained when the light hydrocarbons that are dissolved in the crude oil are volatilized when the crude oil pressure is reduced from well head pressure to less than 0.1 MPa. Depending on the heavier hydrocarbon content in the natural gas, it is termed lean or rich (Table 3.1) [1]. In natural gas reservoirs, the nonassociated natural gas always contains some heavier hydrocarbons, which are called natural gas liquid (NGL). The amount and composition of the associated NGL depend on the origin of the natural gas. When NGL is present, this n-alkane-rich raw material can be co-refined with the syncrude. The presence of associated NGL is beneficial for Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
52
3 Synthesis Gas Production, Cleaning, and Conditioning Table 3.1
Natural gas composition.
Compounds
Methane Ethane Heavier hydrocarbons Carbon dioxide Nitrogen Thiols, H2 S, and COS
Nonassociated gas (vol%)
Associated gas (vol%)
Lean
Heavy
Lean
Heavy
95.7 0.3 – – 4 0.002
87.9 5.3 3.2 – 3.6 0.002
89.6 7.3 0.7 1.6 0.8 <0.001
84.9 6.6 6.2 1.5 0.8 0.002
the carbon efficiency of the overall gas-to-liquid (GTL) process, and it makes little sense to use it for syngas production. When a source of natural gas can be tried into a natural gas distribution network, the natural gas directly becomes a useful energy carrier in its own right. This is the case in many regions of the world. In such instances, it is environmentally irresponsible to consider the natural gas as raw material for a GTL process. The loss of carbon associated with the transformation of the natural gas to another energy carrier cannot be justified. However, if the natural gas is available in a location where it is not useful as energy carrier, GTL conversion can be justified. This point is explicitly made, because it has implications for the refining pathway of methane produced during syngas generation, Fischer–Tropsch synthesis, and syncrude refining. Natural gas components are potentially useful and valuable energy carriers when the infrastructure permits. The point is illustrated by the coal-to-gas conversion facility in North Dakota, which is operated by the Dakota Gasification Company [2]. At this facility, lignite coal is converted into synthetic natural gas (SNG), which is the primary product from this operation. Since the capital cost associated with coal-to-gas conversion is 70–80% of that of a Fischer–Tropsch-based coal-to-liquid (CTL) facility, the value of SNG is less than, but comparable to that of syncrude in that location. Depending on the location, it may be better to co-produce natural gas components as final products rather than using it as raw material for syngas production. 3.2.2 Solid Carbon Sources
There are many different solid carbon sources that can be considered as raw materials for syngas production. The carbon sources that are typically considered are coal, petroleum coke, biomass, and waste. These different solid raw materials differ widely in composition, and even within each type there is much variation (Table 3.2). The heteroatom content indirectly reduces the carbon efficiency that can be obtained with a specific raw material. Depending on the nature of the heteroatom content, it may also add significantly to the load during syngas cleaning. The values in Table 3.2 have been expressed on a mineral- and moisture-free basis, but the raw materials are definitely not mineral and moisture free. The presence of moisture affects the heating value of the raw material and, depending on the gasification technology, it may or may not be detrimental. At gasification conditions, the water will participate in the water gas shift (WGS)
3.3 Syngas from Natural Gas Composition of different carbon-containing raw materials on a moisture- and mineral-free basis.
Table 3.2
Carbon source
H:C ratio
Heteroatom content (mass%)
Coal, anthracite Coal, bituminous Coal, subbituminous Coal, lignite Peat Biomass, lignin-rich Biomass, cellulosic Biomass, oils, and fats Waste, plastic
0.25–0.5 0.6–0.8 ∼0.8 ∼0.8 1.1–1.2 ∼1.2 1.6–1.7 ∼1.9 ∼2.0
1–5 5–15 15–25 25–35 30–40 25–30 ∼50 10–15 <1
reaction (Equation 3.1) to increase hydrogen production. CO + H2 O H2 + CO2 , H298K = −41 kJ·mol−1
(3.1)
The mineral matter is a different kettle of fish. All of the mineral matter will ultimately be rejected as ash or slag. The creation of solid waste is inevitable. The mineral matter also reduces the thermal efficiency of the process, since it has to be heated up to gasification temperatures, but heat recovery from the mineral matter is difficult. Handling of solids is normally more expensive than liquid or gas handling. The transportation of the solid raw material from its point of origin to the XTL facility, as well as its subsequent pretreatment, needs to be considered very carefully. The added complexity and associated capital and operating cost contribution of raw material transportation and pretreatment should not be underestimated. 3.3 Syngas from Natural Gas
Natural gas is converted into syngas by gas reforming. The two main types of gas reforming are steam reforming (Section 3.3.3) and adiabatic oxidative reforming (Section 3.3.4). The main difference between these two types of reforming is the way in which energy is supplied to drive the reforming process. Steam reforming is a catalytic process. Natural gas and steam, with or without CO2 addition, is converted by endothermic reforming reactions into CO and H2 (Equations 3.2 and 3.3) and heat is supplied externally to the process. CH4 + H2 O CO + 3H2 , H298K = +206 kJ·mol−1 −1
CH4 + CO2 2CO + 2H2 , H298K = +247 kJ·mol
(3.2) (3.3)
Adiabatic oxidative reforming, irrespective of whether it employs a catalyst or not, takes place by exothermic partial combustion of the natural gas to produce CO (Equation 3.4) to supply heat for
53
54
3 Synthesis Gas Production, Cleaning, and Conditioning
the endothermic reforming reactions (Equations 3.2 and 3.3). 1 CH4 + 1 O2 → CO + 2H2 O, H298K = −519 kJ·mol−1 2
(3.4)
In both processes, the WGS reaction (Equation 3.1) takes place, and the final product approaches equilibrium at the reaction conditions. One of the most important side reactions is the coking reaction, which may take place by CO disproportionation by the reverse Boudouard reaction (Equation 3.5), or by catalyzed dehydrogenation of the natural gas feed (Equation 3.6). 2CO C + CO2 , H298K = −172 kJ·mol−1
(3.5)
CH4 C + 2H2 , H298K = +75 kJ·mol−1
(3.6)
The steam that is co-fed with the natural gas for reforming has the added benefit of suppressing the coking reaction. The risk of coking is increased when C2 and heavier hydrocarbons are present in the natural gas. A typical process (Figure 3.1) for the production of syngas from natural gas employs gas cleaning (Section 3.3.1) and prereforming (Section 3.3.2) as pretreatment steps before gas reforming. The objective of gas cleaning is to remove potential reforming catalyst poisons from the natural gas feed. Not all of the adiabatic oxidative reforming technologies require a catalyst, and gas cleaning is not a prerequisite in such cases. However, it turns out that precleaning the natural gas is not only effective for protecting the reforming catalyst but also is an efficient way to ensure that the syngas produced does not require subsequent cleaning to remove Fischer–Tropsch catalyst poisons. Gas cleaning before reforming has therefore become standard industrial practice. Prereforming is not an essential processing step, but it holds advantages for the overall process. The main objective of prereforming is to convert the C2 and heavier hydrocarbons, thereby reducing the risk of coking during the main gas reforming process. Steam
Natural gas
ZnO Gas reforming
Hydrogen
Ni/Al2O3 Hydrogenation
Sulfur absorption
Prereformer
Figure 3.1 Generic process flow diagram of a gas reforming section that converts natural gas into clean but unconditioned syngas. Heat integration is not indicated. The process flow associated with gas reforming depends on the reforming technology that is selected.
Reformer
Clean syngas (unconditioned)
3.3 Syngas from Natural Gas
3.3.1 Natural Gas Cleaning
Industrial gas reforming catalysts are mainly supported nickel-based catalysts. The main catalyst poisons for such nickel-based reforming catalysts are [3] given below: 1) Sulfur-containing compounds, which are present in the natural gas (Table 3.1). 2) Halogen-containing compounds, which may be present in the natural gas. 3) Arsenic, which is seldom present in natural gas. In order to avoid poisoning of the reforming catalyst, these catalyst poisons must be removed from the natural gas. A two-step strategy is employed (Figure 3.1). The natural gas is hydrogenated to remove any traces of unsaturated compounds and to convert all sulfur species into hydrogen sulfide (H2 S). At the same time, the halogen-containing compounds are also hydrogenated to produce the corresponding hydrogen acid (HX, X = F, Cl, Br, and I). In the second step, H2 S and HX are quantitatively removed. The concentration of compounds in the natural gas that has to be hydrogenated is very low and only a small amount of hydrogen or hydrogen-containing gas needs to be premixed with the natural gas feed. The mixed feed is preheated to 350–400 ◦ C and hydrogenation is conducted over a sulfided base metal catalyst, such as NiMo/Al2 O3 or CoMo/Al2 O3 . Various metals and metal oxides can be considered for sulfur capture [4, 5]. Industrially, zinc oxide (ZnO) is preferred. It has the further advantage of being easily regenerable by roasting, because zinc does not form high-temperature, stable sulfur-containing species. The thermodynamic equilibrium of the sulfur removal reaction (Equation 3.7) is favorable and far to the right. ZnO + H2 S ZnS + H2 O, Keq,623K = 106
(3.7)
The H2 S in natural gas can consequently be removed to extremely low levels (ng·g−1 levels). The equilibrium is improved by operating at lower temperatures, but there is a trade-off with the kinetics of sulfur removal, which becomes slower. The ZnO absorbers are usually operated in the same temperature range as the hydrogenation reactor. When H2 and CO2 are present in the feed, the equilibrium removal of H2 S is lower. The natural gas composition (Table 3.1) determines the CO2 content, but the H2 content can be optimized. Strategies to achieve maximum H2 S removal are discussed in the literature [1]. 3.3.2 Adiabatic Prereforming
Adiabatic prereforming is a form of low-temperature steam reforming. The function of an adiabatic prereformer is to convert the C2 and heavier hydrocarbons that are present in the natural gas into syngas (Equation 3.8) to lower the risk of irreversible coking (Equation 3.9) at higher reforming temperatures. Cn H2n+2 + nH2 O → nCO + (2n + 1)H2
(3.8)
Cn H2n+2 → nC + (n + 1)H2
(3.9)
55
56
3 Synthesis Gas Production, Cleaning, and Conditioning
The feed is preheated and then steam reformed over a supported nickel catalyst (Figure 3.1). The reforming reaction is endothermic, but the adiabatic temperature profile is not necessarily monotonically decreasing. Methanation, the reverse reaction of methane reforming (Equation 3.2), and the WGS reaction (Equation 3.1) can also take place. These are exothermic reactions and may lead to the adiabatic temperature profile passing through a minimum due to steam reforming, before increasing toward the reactor outlet due to methanation and WGS [1, 6]. 3.3.3 Steam Reforming
Steam reforming employs steam to convert methane over a nickel-based catalyst (Ni/Al2 O3 ) into hydrogen, carbon monoxide, and carbon dioxide (Equations 3.1–3.3). The reactions are endothermic and equilibrium limited. Thermodynamically, high temperature, low pressure, and a high steam to methane ratio favor methane conversion into syngas. Heat is supplied externally by combustion of a fuel to drive the reforming reaction. A steam reformer does not require molecular oxygen (O2 ) from air, or pure oxygen from an air separation unit (ASU), as a process stream. Oxygen is only required as a utility stream for fuel combustion. Steam reforming produces a syngas that is rich in hydrogen, and the H2 :CO ratio is typically well above 5 : 1. This ratio is much higher than is required by Fischer–Tropsch synthesis. The H2 :CO ratio can be reduced by co-feeding CO2 (Equation 3.3), as well as by using a lower steam to methane ratio in the feed. The extreme case is dry reforming of methane, where methane is reformed with only CO2 . At low steam to methane ratios, the catalyst selection is critical, because unpromoted Ni/Al2 O3 catalysts are prone to carbon formation under such conditions [7, 8]. The operating conditions are quite severe, with a furnace inlet temperature of 540–580 ◦ C and outlet temperature of 820–880 ◦ C. For economic reasons, the operating pressure is usually around 2.0–2.5 MPa and it is a trade-off between volumetric productivity and methane conversion. The steam to methane ratio is usually 2.5–5.0 [3]. A high steam to methane ratio lowers the risk of carbon deposition and increases methane conversion, but lowers the volumetric productivity. The main advantages of steam reforming for syngas production are the following: The reforming reagent (H2 O) can easily be recovered by condensation and it does not increase the volumetric flow to downstream units. 2) No air separation unit is required. 3) Steam reforming provides an easy pathway for CO2 conversion, and co-feeding CO2 is a useful strategy to produce a syngas composition closer to the usage ratio required by Fischer–Tropsch synthesis. 4) The energy that is required for reforming can be supplied by combusting lower quality feed materials. 1)
3.3.4 Adiabatic Oxidative Reforming
Adiabatic oxidative reforming of natural gas into syngas can take place by one of the three methods [1]:
3.3 Syngas from Natural Gas
• Homogeneous syngas production. An oxidant (air or oxygen) is mixed with the gaseous feed (natural gas and optionally steam) and subjected to high temperature, typically in the order of 1300–1400 ◦ C. These processes are also referred to as homogeneous gasification or partial oxidation (POX) processes. No catalyst is employed. The POX process is followed by a section to recover heat from the raw synthesis gas, as well as a section to remove carbon (soot) produced during POX. The H2 :CO ratio is typically in the range 1.6–1.9. • Heterogeneous syngas production. The natural gas, steam, and an oxidant are mixed before being passed over a catalyst. The catalyst is responsible for the POX of the gaseous feed, as well as for the reforming reaction. These processes are also referred to as catalytic partial oxidation (CPO). Conversion approaches that of steam reforming. • Combined homo- and heterogeneous syngas production. As in the other two cases, the oxidant is mixed with the natural gas and steam. It is first partially combusted in a burner to raise the temperature of the mixture. The product from partial oxidation is then passed over a catalyst to complete the conversion into syngas. This sequence is employed in autothermal reforming (ATR) technology. Steam is required to avoid carbon formation and to increase the H2 :CO ratio of the syngas. The H2 :CO ratio is typically above 2 and depends on the steam to methane ratio, temperature, and CO2 content of the feed. The main advantages of adiabatic oxidative reforming processes are the following: 1) Much larger single stream units are possible than is possible with steam reforming. 2) Adiabatic reformers are more compact in size. 3) Higher outlet gas temperatures can be achieved, which allows higher methane conversion at lower steam to methane ratios.
3.3.5 Gas Reforming Comparison
Gas reforming technology cannot be selected in isolation and it has to be matched to the gas loop design of the Fischer–Tropsch process. Current industrial Fischer–Tropsch facilities employ steam reforming, partial oxidation, and ATR for the production of syngas. The main characteristics of these processes are compared in Table 3.3. It has been reported that adiabatic oxidative reforming is more cost effective than steam reforming for large-scale natural gas conversion [1]. Specifically, it has been reported that ATR holds the most promise for GTL conversion applications [9]. Compared to steam reformers, the additional cost of air separation units is offset by the economy of scale that can be achieved by the larger single-train capacity and the more compact size of adiabatic oxidative reformers. Yet, it is not clear whether this analysis holds true for all applications. From a refining perspective, the two most important differences are hydrogen and nitrogen availability. Adiabatic oxidative reforming technology, which requires air separation, enables diversification of the refinery into nitrogen-based chemicals. Hydrogen availability allows export of hydrogen to the refinery.
57
58
3 Synthesis Gas Production, Cleaning, and Conditioning Table 3.3 Comparison of gas reforming technologies that are industrially applied in combination with Fischer–Tropsch synthesis.
Description Feed O2 :C ratio (mol mol−1 ) Feed H2 O:C ratio (mol mol−1 ) Outlet temperature (◦ C) Outlet pressure (MPa) Outlet CH4 content (mol%) Product H2 :CO ratio (mol mol –1 )a Product CO2 :CO ratio (mol mol –1 )a a
Steam reforming
Partial oxidation
0 2.5–5.0 820–880 2.0–2.5 3–5 4–7 0.5–1
0.55–0.65 0–0.15 1300–1400 2.5–4.0 0.1 1.6–1.9 0.05–0.1
Autothermal reforming 0.55–0.6 1.5–2.5 950–1050 ∼2.5 0.5–1 2.5–3.5 0.2–0.3
0.6 0.6 1020–1065 2.5–2.9 0.5–1.2 2.2–2.3 0.2
Product compositions with no additional CO2 co-feed to the reformer.
3.4 Syngas from Solid Carbon Sources
The conversion of solid carbon sources into syngas takes place by gasification. Gasification takes place in the range 800–1800 ◦ C in the presence of a substoichiometric amount of oxidant (typically O2 ) and a moderator (typically H2 O). At gasification temperatures, the description of carbon gasification parallels that of methane reforming and becomes independent of the carbon source. The nature of the raw material influences the gasification process mainly by its solid handling characteristics and its behavior in the gasifier before gasification temperatures are reached. This becomes very important in biomass gasification and low-temperature gasification processes in general. There are a large number of gasification technologies that have been developed over the years, mainly for the gasification of coal [10–14]. The principle can be extended from coal to other solid carbon-containing feed materials. The gasifier design affects the temperature history and environment of a solid particle as it moves through the gasifier, which in turn affects the gasification chemistry. These effects will be considered when each gasifier type is discussed (Sections 3.4.2–3.4.4). In order to illustrate the basic chemistry of gasification, we will consider the changes taking place as a solid particle is heated up to the gasification temperature independently of the gasifier type. 1) Drying. Surface moisture and free water are evaporated to dry the solid feed. This is an endothermic process and the energy requirement is related to the heat of vaporization of water (Hvap,3 MPa = 1.8 MJ· kg−1 ) at gasification pressure. 2) Autoxidation. In the presence of oxygen, autoxidation reactions start almost immediately. At low temperature (<200 ◦ C), oxygen is incorporated into the carbon-containing molecules as hydroperoxides and other oxygenate functionalities. This renders the material more thermally labile. As the temperature increases, it becomes only the first step leading to partial or complete combustion. 3) Pyrolysis. Pyrolysis is the thermal cracking of molecules in the absence of oxygen, but in the context of gasification the term is more loosely used to indicate thermal cracking in
3.4 Syngas from Solid Carbon Sources
4)
5)
6)
7)
the presence of oxygen also. The temperature at which pyrolysis commences is dependent on the nature of the molecules in the solid particle. The bond dissociation energy of each bond determines the temperature at which homolytic bond scission can take place [15–17]. Molecules containing heteroatoms are more labile, and pyrolysis starts becoming significant at temperatures above 300 ◦ C, although it may actually start at lower temperatures (100–150 ◦ C). In the temperature range 300–450 ◦ C, pyrolysis has to compete with free radical recombination reactions, also called retrograde condensation reactions (which must not be confused with the term used to describe phase behavior). During retrograde condensation, heavier products are formed that are more stable than the feed. However, as the temperature is raised above 450 ◦ C, retrograde condensation becomes insignificant. Devolatilization. Like drying, some lighter organic molecules are vaporized from the solid particle as the temperature is increased. Lighter molecules are also produced by pyrolysis. Devolatilization is therefore the liberation and vaporization of organic compounds from the solid particle, most of which is formed by pyrolysis. Secondary pyrolysis. The lighter products produced by pyrolysis and devolatilization will continue to be subjected to thermal decomposition as the temperature increases above 450 ◦ C. These high-temperature thermal cracking reactions are similar to the conversion that is purposefully conducted in thermal crackers. The high-temperature chemistry of light hydrocarbons has been well described in the classic text by Egloff [18]. As the temperature approaches and exceeds the gasification temperature (>800 ◦ C), complex molecules break down completely to yield C1 products. Oxidation. The extent of oxidation increases with temperature and oxygen availability, and at gasification temperatures carbon is converted to either CO or CO2 . In regions of high O2 availability and temperature, complete combustion takes place, whereas when O2 availability is limited, partial combustion takes place. Reforming. When the solid carbon has been reduced to C1 products through secondary pyrolysis and oxidation, the gasification chemistry is similar to that described for gas reforming (Section 3.3). The product composition can then be determined from the equilibrium conversion of the WGS (Equation 3.1) as well as Boudouard (Equation 3.5) and methanation (Equation 3.6) reactions [14]. This description is somewhat of an oversimplification, because it describes only hydrogen, carbon, and oxygen. Raw materials used for gasification generally include other heteroatoms, such as sulfur and nitrogen too. The chemistry of these compounds will be discussed separately (Section 3.4.1).
3.4.1 Gasification of Heteroatoms
The main heteroatoms in solid carbon sources are oxygen, nitrogen, and sulfur. Other trace compounds are present both in organic matter and associated with mineral matter. Halogens, arsenic, and metals such as mercury can also be gasified. Sulfur is converted to hydrogen sulfide (H2 S) and, under typical gasification conditions, 93–98% of the gasified sulfur will be in this form. The remainder is in the form of carbonyl sulfide (COS) [14]. The ratio between H2 S and COS is determined by the equilibration of the
59
60
3 Synthesis Gas Production, Cleaning, and Conditioning
reaction with carbon oxides (Equations 3.10 and 3.11). H2 S + CO COS + H2 , H298K = −7 kJ·mol−1 H2 S + CO2 COS + H2 O, H298K = +34 kJ·mol
(3.10) −1
(3.11)
When the gasifier technology is such that pyrolysis products do not pass through the gasification region in the gasifier, it is possible to obtain other sulfur species. For example, when low-temperature moving bed gasification technology (Section 3.4.2) is employed, the pyrolysis products may contain thiophenes, thiols, and carbon disulfide. Nitrogen is introduced into the gasifier by the raw material and in combination with the oxidation. In air-blow gasification, the amount of molecular nitrogen (N2 ) is high, but even in oxygen-fired gasification ‘‘pure’’ O2 contains some Ar and N2 . Molecular nitrogen may be oxidized in the gasification zone and later be reduced to ammonia, but its contribution to ammonia production is not significant. Most of the non-N2 species in the syngas are hydrogen cyanide (HCN) and ammonia (NH3 ) produced by decomposition of nitrogen compounds in the feed. When the nitrogen is bound in aromatic rings, decomposition yields mainly HCN, whereas the principal product from aliphatic nitrogen is NH3 [14]. In practice, much of the nitrogen in coal feed is refractory, and even in a moving bed gasifier much of the bound nitrogen is converted to N2 in the gasification section [19]. The NH3 concentration in raw synthesis gas produced from feed materials with more aliphatic nitrogen is expected to be higher. Halogen-containing compounds that are gasified are likely to be present as the acid in the gas. As the raw syngas is cooled down below 250 ◦ C, these acids will react with ammonia to produce ammonium salts (Equation 3.12). HX + NH3 → NH4 X(s),X = F, Cl, Br, I
(3.12)
Metals can be liberated and gasified in the form of the elemental compound, such as Hg, or as volatile metal carbonyls such as Fe(CO)5 and Ni(CO)4 . Elements that can be reduced to form volatile compounds, such as AsH3 , may also be found in the gas phase. These compounds are present in trace levels, but when the syngas production is large, the cumulative amount quickly becomes significant. Gas cooling and cleaning must make provision for these compounds if they are present in the raw syngas. 3.4.2 Low-Temperature Moving Bed Gasification
Low-temperature gasification does not take place at a low temperature, but it has a low gas outlet temperature. The outlet gas temperature is typically in the range 425–650 ◦ C. This is a result of the gasifier design, which employs the hot syngas from the gasification zone to pyrolize and preheat the incoming solid feed (Figure 3.2). The flow of solid raw material feed and gas is countercurrent, and this has three important consequences. The first consequence is that pyrolysis liquids are always coproduced by low-temperature gasification, because the hot gas strips pyrolized products from the solid feed material before it reaches the gasification and combustion zones. The second consequence is that the exiting gas can entrain fine material and the solid feed must have a particle size of at least 6 mm. The third consequence is that much of the heat recovery takes place in the gasifier. The solid feed is preheated and pyrolized with the hot raw gas before it exits the gasifier, while
3.4 Syngas from Solid Carbon Sources
Feed lock Solid feed
Raw gas
Pyrolysis Gasifier
Gasification Steam oxygen Ash/ slag Figure 3.2
Combustion
O2 + H2O 0
400
800
1200
Temperature (°C)
Low-temperature moving bed gasification process.
the hot ash/slag preheats the steam and oxygen before the ash/slag exits the gasifier. This makes low-temperature gasification thermally very efficient. When the heating value of the liquid products is included, the cold gas efficiency (Equation 3.13) of low-temperature gasifiers is better than that of medium- or high-temperature gasifiers. Cold gas efficiency =
Higher heating value of product gas Heating value of feed
(3.13)
There are two important low-temperature moving bed gasification technologies that produce very different raw gas compositions (Table 3.4) [14]: 1) Dry ash moving bed gasification. Industrially, the Lurgi dry ash moving bed gasifier is extensively used for coal to syngas conversion. The solid feed enters the gasifier through a feed lock-hopper that is isolated from the gasifier during loading. The loading and pressurization of the feed lock-hopper is a periodic process, as is the feed to the gasifier. When the solid feed enters the gasifier, which is operated at 2.5–3.0 MPa, it falls on top of the hot bed of the feed material already in the gasifier. As it slowly moves down the bed, the feed passes through zones where it is dried, devolatilized, pyrolized, gasified, and combusted. Material that has a tendency to agglomerate and cake cannot be processed, because it will impede movement of the bed. At the bottom of the bed, the mineral matter (ash) that cannot be combusted is cooled by the incoming oxygen and steam before it is discharged through the ash lock. The ash lock, like the feed lock-hopper, is operated on a periodic basis to discharge the ash. The dry ash mode of operation places a restriction on the mineral composition of the feed material. The mineral matter should have a melting point above the maximum operating temperature in the gasifier to avoid sintering or melting of the ash. 2) Slagging moving bed gasification. The British Gas/Lurgi (BGL) slagging gasifier is a variation on the Lurgi dry gas gasifier design. The feed lock-hopper and the operation of the gasifier are similar. However, at the bottom of the bed the mineral matter is allowed to melt and form a molten slag. This allows the use of solid feed materials with a lower ash melting temperature and thereby a higher temperature gasifier operation. The steam to feed ratio is also lower than in dry bottom gasifiers, and as a consequence the raw gas has a lower H2 :CO
61
62
3 Synthesis Gas Production, Cleaning, and Conditioning Table 3.4 Comparison of the feed requirements and product composition of different low-temperature moving bed gasifiers.
Description Feed (kg/normal m3 H2 + CO) Bituminous coal, maf a Steam Oxygenb Raw gas composition, maf (mol%) H2 CO CO2 CH4 C2 and heavier hydrocarbon gasc H2 S + COS NH3 N2 Pyrolysis liquids (kg/normal m3 H2 + CO)
Lurgi dry bottom gasifier
BGL slagging gasifier
0.75 1.93 0.40
0.52 0.20 0.33
42.15 15.18 30.89 8.64 0.79 1.31 0.36 0.68 0.06
31.54 54.96 3.46 4.54 0.48 1.31 0.36 3.35 0.01
a Bituminous coal composition on moisture- and ash-free (maf) mass basis: 77.3% C, 5.9% H, 4.3% S, 1.4% N, and 11.1% O. b 1.43 kg O /normal m3 O . 2 2 c Excludes pyrolysis liquids.
ratio (Table 3.4). The method of steam and oxygen injection further allows the injection of solid fines at the bottom of the gasifier. When low-temperature moving bed gasification technology is employed for syngas generation, it has a significant impact on the refinery design. The refinery design is no longer independent of the raw material, since the raw material influences the composition of the pyrolysis liquids that are coproduced during gasification. The pyrolysis liquids increase the carbon efficiency of the overall process and provides refining synergies, albeit at the expense of added complexity. Depending on the raw material, various chemicals can be recovered from the aqueous and organic products. 3.4.3 Medium-Temperature Fluidized Bed Gasification
Most medium-temperature gasification processes are fluidized bed processes. With intensive mixing between the oxygen, steam, and solid feed in the fluidized bed, the fluidized bed approximates a continuous stirred tank reactor (CSTR). Operating temperatures are usually in the range 950–1050 ◦ C, but can be lower for more reactive solid feed materials. The raw gas leaving the gasifier is consequently at the same temperature as the gasifier (Figure 3.3).
3.4 Syngas from Solid Carbon Sources
Raw gas Feed lock Solid feed
Gasifier O2 + H2O
Steam oxygen 0 Ash
400
800
1200
Temperature (°C)
Figure 3.3 Medium-temperature fluidized bed gasification process, as illustrated by a fixed fluidized bed gasifier.
The nature of the fluidization depends on the mean gas velocity. Gasifiers with a low mean gas velocity are operated as fixed fluidized beds, but at higher gas velocities circulating fluidized bed and entrained bed operation are possible. Medium-temperature gasifiers discharge the noncombustible mineral matter as ash. The ash melting temperature of the solid feed is important, and the gasifier must be operated below the ash melting temperature to avoid slag formation, which can lead to defluidization of the bed. Some technologies allow some agglomeration of the ash to facilitate ash removal. The ash is discharged in a highly leachable form, which poses some challenges for solid waste management [14]. Because of the lower operating temperature of the gasifier (the gasification actually takes place at a lower temperature than in low-temperature moving bed gasifiers), carbon conversion is incomplete (<97%). Medium-temperature fluidized bed gasifiers are therefore mainly used to convert very reactive feed materials, such as peat, low-rank coals, and biomass. In order to improve conversion at the comparatively low operating temperature, smaller particles are favored, but particle size affects fluidization. Smaller particles can be processed than in moving bed gasifiers, but maintaining fluidization requires careful balancing of the particle size distribution and gas velocity. The main advantage of fluidized bed gasifiers is their high turndown flexibility and the ability to process variable loads. Fluidized bed gasifiers can also be employed in conjunction with sorbent materials to capture sulfur from high-sulfur solid feed materials. By judicious operation, up to 90% of the sulfur can be retained by the bed [20]. Typical performance data of fluidized bed gasifiers with different feed materials are presented in Table 3.5 [12, 14]. The raw gas produced by medium-temperature gasification is not necessarily free of pyrolysis products. When it is the case, it is inconvenient, since the amount of pyrolysis products thus produced is insufficient to create refining synergy, but is just significant enough to have an impact of the refinery if it has to be co-refined.
63
64
3 Synthesis Gas Production, Cleaning, and Conditioning Table 3.5 Performance data for medium-temperature fluidized bed gasifiers operated at different conditions and with different feed materials.
Description
Biomass
Temperature (◦ C) Pressure (MPa) Feed (kg/normal m3 H2 + CO) Solid feed, maf Steam Oxygen Raw gas composition, maf (mol%) H2 CO CO2 CH4 H2 S + COS N2 + Ar
1100 3
a
0.89 0 1.75a 18.9 31 6.7 2.1 <0.1 41.3
Lignite coal 815–980 0.1 0.47 1.38 1.62 38.4 35.3 21.9 1.8 1.5 1.1
Bituminous coal 1100 3
1100 3
0.78 0 0.48
0.52 0.20 0.46
32.8 56.7 6.2 2.6 0.2 1.5
37.3 52 5.3 3.5 0.3 1.6
Air (not O2 ) as oxidant; the O2 feed content is 0.41 kg/normal m3 H2 + CO.
3.4.4 High-Temperature Entrained Flow Gasification
High-temperature entrained flow gasifiers operate at high gasification temperatures, typically above 1400 ◦ C. The solid raw material and oxygen are co-currently fed to the gasifier, and the flow direction can be either up or down depending on the technology. The solid raw material may be fed as a water slurry or as a dry, dense fluidized medium. In the case of a dry feed, steam can be added separately. The gas outlet temperature is the same as the gasification temperature, because the feed is co-currently entrained in the gas (Figure 3.4). In high-temperature entrained flow gasifiers, the gasification temperature exceeds the ash melting point, and all entrained flow gasifiers are slagging. Very high carbon conversion can be achieved, but at the expense of a high oxygen to feed ratio. Despite the high gasification temperature, the residence time in the gasifier is short because of the co-current flow configuration. To ensure complete carbon conversion, the solid feed must be ground to a small size and feed preparation is costlier than for the other gasifier types. However, there is little restriction on the feed material, apart from its ash content. High ash feed materials quickly deteriorate the cold gas efficiency of entrained flow gasifiers [11]. The economic limit for slurry-feed gasifiers is around 20% ash, and for dry-feed gasifiers it is around 40% [14]. The raw syngas produced by entrained flow gasifiers is free of any pyrolysis products, which simplifies downstream processing. The raw syngas composition obtained from high-temperature entrained flow gasification is more or less insensitive to the feed and the gasification technology employed (Table 3.6) [14]. The methane production in entrained flow gasifiers is generally low, and the H2 :CO ratio is usually around 1 : 2. However, solid feed materials that are rich in oxygenates will produce more CO2 and even less H2 .
3.4 Syngas from Solid Carbon Sources Feed slurry Oxygen
O2
Solid feed
Gasifier
0 Raw gas and slag
400
800
1200
Temperature (°C)
Figure 3.4 High-temperature entrained flow gasification process, as illustrated by a slurry-feed down-flow gasifier. Table 3.6 Performance of dry-feed, high-temperature entrained flow gasification with different feed materials.
Description
Feed (kg/normal m3 H2 + CO) Coal, maf Steam Oxygen Raw gas composition, maf (mol%) H2 CO CO2 H2 S + COS N2 + Ar
North Dakota lignite
Illinois subbituminous
South African bituminous
German anthracite
1.07 0 0.51
0.51 0.06 0.37
0.48 0.09 0.37
0.44 0.13 0.37
26 62 10 0.1 2
34 63 1 0.4 2
33 64 1 0.3 2
31 65 1 0.2 2
The different entrained flow gasifier technologies can be classified in many ways, but the most important differentiating feature is the way in which the feed is introduced, as given below: 1) Dry-feed entrained flow gasification. The flow direction has an impact on the temperature profile within the gasifier. Up-flow gasification approximates CSTR behavior, and the temperature profile is almost uniform throughout the gasifier. Down-flow gasification has a profile closer to that shown in Figure 3.4. The gas purity from either design is high, usually with little CO2 and CH4 in the raw syngas. This has some advantages for downstream syngas cleaning (Section 3.5.1). 2) Slurry-feed entrained flow gasification. The main advantages of slurry feed are that it is easier to introduce the solid feed material under pressure into the gasifier and that it is not necessary to dry the coal. However, the water thus introduced must be evaporated in
65
66
3 Synthesis Gas Production, Cleaning, and Conditioning
the gasifier, which lowers the cold gas efficiency. About 20–25% more oxygen is required to supply combustion heat. The raw syngas from a slurry-feed gasifier usually has a higher H2 :CO ratio and more CO2 than that from a dry-feed gasifier. The gasifier also requires a longer residence time for the same conversion as a dry-feed gasifier, because part of the reactor volume is required for evaporation. Preheating the slurry feed to around 300 ◦ C before introducing it into the gasifier can reduce the performance degradation associated with water in the slurry feed. 3.4.5 Gasification Comparison
Gasification technology selection is subject to gasifier compatibility with the feed [20]. The most important feed properties that will determine which gasifier types can be used are as follows: 1) Ash content. If the ash content is high, high-temperature gasification becomes inefficient. 2) Ash fusion temperature. Depending on the melting temperature of the ash, the ash can be removed either as a dry ash or as a slag. The gasifier type must be selected accordingly. 3) Reactivity. Medium-temperature fluidized bed gasification requires more reactive feed to compensate for its lower gasification temperature. 4) Agglomeration. Feed materials that tend to agglomerate (or cake) when heated up cannot be used in gasifiers where it will impede bed movement or cause defluidization. 5) Particle size. The particle size distribution determines the preferred gasifier type or, conversely, feed pretreatment must ensure that an appropriate particle size distribution is achieved. Low-temperature moving bed gasifiers require large (>6 mm) particles, whereas high-temperature entrained flow gasifiers require small (<0.1 mm) particles. Once a list of gasification technologies that are compatible with the solid feed material has been compiled, the final selection can be made on the basis of other performance criteria. A summary of the different gasification technologies is presented in Table 1.1. There are many trade-offs. There is no single gasification technology that will be the most economical or practical in every instance [1]. In large-scale coal-based facilities, low-temperature gasification has advantages that are related to its higher carbon efficiency and the value of the pyrolysis liquids. In the author’s experience with existing and conceptual CTL facilities, this translates into an economic advantage, despite the added complexity. For smaller scale applications and specifically biomass-to-liquid applications, it may not be possible to exploit these advantages to the same extent.
3.5 Syngas Cleaning
The syngas produced by natural gas reforming does not require additional cleaning to remove sulfur, because the sulfur is removed before gas reforming (Section 3.3.1). The syngas produced by natural gas reforming contains CO2 (Table 3.3) and, depending on the Fischer–Tropsch gas loop design, one may want to consider CO2 removal.
3.5 Syngas Cleaning
CO2 + H2S Gas cleaning Solid feed
Gasifier
Steam, O2
Pure syngas Gas liquor Dusty tar
Figure 3.5 Synthesis gas cleaning from gasification when the raw synthesis gas contains condensable liquids, such as the product from low-temperature coal gasification.
The raw syngas produced by gasification of solid feed materials requires more cleaning. In the worst case, the raw syngas may contain significant quantities of steam, pyrolysis products, solids, light hydrocarbons, methane, sulfur, nitrogen-containing gases (mainly H2 S, COS, and NH3 ), and CO2 . These complex ‘‘dirty’’ raw syngas streams are typical of low-temperature coal gasification and much of the literature on the cleaning of such worst case syngas can be found in the literature on coal gas. When the syngas is cleaned in a Fischer–Tropsch-based facility, the refinery may gain carbon efficiency by exploiting the opportunities presented by such ‘‘dirty’’ gas. The overall cleaning process is shown in Figure 3.5. In the first step, the solids are removed by scrubbing the gas with condensable products from the gas. The operating conditions at which this scrubbing operation takes place will determine whether water will be condensed or not. Even if water is not condensed in the first scrubbing stage, it usually condenses shortly thereafter. The product is a two-phase liquid that contains fine particulate matter. The aqueous product from gasification is called gas liquor and the organic product is called tar. The gas liquor contains dissolved NH3 , H2 S, and CO2 , in addition to water-soluble compounds such as light oxygenates and phenolics (the composition depends on the raw material used for gasification). Technologies for the recovery of ammonia and chemicals from gas liquor were already well established nearly a century ago [21]. The tar composition depends strongly on the raw material that is gasified. The raw gas that remains after condensation contains H2 S, COS, and light hydrocarbons, in addition to gases such as H2 , CO, CO2 , Ar, and N2 . This raw gas is then further cleaned to remove the sulfur compounds, CO2 , and light hydrocarbons to very low levels. The pure synthesis gas that is obtained from this intensive cleaning step still contains some inert gases, but it consists mainly of H2 and CO. The H2 :CO ratio may not necessarily be suitable for Fischer–Tropsch synthesis and subsequent synthesis gas conditioning may be required. 3.5.1 Acid Gas Removal
Processes for the removal of acid gases (CO2 and H2 S) operate on the same basic principle, namely, the selective absorption of the target compounds from the gas stream. Solid absorbents, such as activated carbon, molecular sieves, and metal oxides (Section 3.3.1), can be used when the concentration of acid gases is low (parts per million levels). Solid absorbents are not cost effective for treating raw synthesis gas from gasification which contains percentage levels of
67
68
3 Synthesis Gas Production, Cleaning, and Conditioning Table 3.7
Solvents and processes for acid gas removal.
Solvent for absorption
Monoethanolamine (MEA) Diethanolamine (DEA) Potassium carbonate Methanol N-Methylpyrrolidone (NMP) Polyethylene glycol dimethyl ether
Type
Chemical Chemical Chemical Physical Physical Physical
Effective for absorption of CO2
H2 S
COS
Thiols
Yes Yes Yes Yes Yes Yes
Yes Yes Yes Yes Yes Yes
No No Yes Yes Yes Yes
No No No Yes Yes Yes
Process name
MEA DEA Benfield Rectisol Purisol Selexol
acid gases. In such cases, liquid absorbents that can be easily regenerated are more effective. Industrially, only liquid-based absorption processes are used for the purification of raw gas from gasification. Liquid absorption processes can be classified by the action of the absorption medium (Table 3.7) [22]: 1) Chemical absorption. In petroleum refinery processes, chemical absorption is almost exclusively used for acid gas removal [23]. Chemical absorption entails a chemical reaction between the acid gases and the solvent. These are normally acid–base reactions. The reaction must be reversible to allow solvent regeneration and chemical absorption to be equilibrium driven. Low temperatures and high pressures promote absorption reactions, whereas the opposite conditions promote regeneration. Since the driving force for absorption is reaction chemistry, a rate law (typically second order) describes the rate of absorption. The overall rate is dependent on mass transfer from the gas to the liquid phase and the reaction within the liquid phase. 2) Physical absorption. Physical absorption is accompanied by a smaller energy change than chemical absorption and is consequently more easily reversible. Since absorption is based on solubility, physical absorption processes tend to be less selective. This is advantageous when both acidic (CO2 and H2 S) and comparatively nonacidic (thiols and light hydrocarbons) components must be removed simultaneously. To achieve high syngas purity, as is required by Fischer–Tropsch synthesis, the removal of sulfur compounds must be almost quantitative. In order to achieve such high removal efficiency with a physical absorbent, multiple absorption stages are required. The driving force for absorption is described by Henry’s Law, and it is consequently related to the partial pressure of the acid gas compounds. The overall rate is dependent on the mass transfer from the gas to the liquid phase, with saturation solubility being the limiting case. The simultaneous removal of CO2 and H2 S by virtually all solvent systems is advantageous from a syngas cleaning perspective. However, it also implies that, when the solvent is regenerated, CO2 and H2 S are desorbed together. If the CO2 concentration is high compared to that of H2 S, the subsequent removal of H2 S from CO2 becomes more difficult and less efficient. It is environmentally unacceptable to release H2 S into the atmosphere. In this respect, gasifiers that
3.6 Syngas Conditioning
produce less CO2 have an advantage for gas cleaning, because the H2 S concentration in the mixed desorption gas is higher. When the H2 S concentration is sufficiently high, sulfur removal can be performed by standard sulfur removal technology, such as the Claus process.
3.6 Syngas Conditioning
Syngas conditioning refers to any adjustments made in the composition of the pure syngas to match the H2 :CO ratio requirements of Fischer–Tropsch synthesis. The workhorse of syngas conditioning is WGS conversion (Section 3.6.1). In the next chapter (Chapter 4) it will be seen that some Fischer–Tropsch catalysts are also WGS active and that such catalysts can in principle perform the WGS conversion in the synthesis reactor. The conditioning of the pure syngas is not a sequential process (Figure 1.3): in fact, far from it. The gas loop design is integral to the application of the Fischer–Tropsch technology and has ramifications for the refinery too. Gas loop design is a specialized topic in its own right and cannot be discussed in isolation [24]. Gas loop design will be discussed in Chapter 5 and as part of the description of different Fischer–Tropsch facilities (Chapters 6–12). 3.6.1 Water Gas Shift Conversion
The WGS reaction (Equation 3.1) allows the H2 :CO ratio to be adjusted by adding either H2 O or CO2 . The reaction is equilibrium limited, and the equilibrium constant KWSG varies with the temperature T (K) according to Equation 3.14, which is valid over the temperature range 315–480 ◦ C (588–753 K) [25]. KWGS = exp [(4578/T) − 4.33]
(3.14)
WGS conversion is carried out in a fixed bed reactor. The aim is normally to increase the H2 :CO ratio. The operating temperature is determined by the choice of catalyst. 1) High-temperature WGS. Most high-temperature WGS catalysts are based on Fe2 O3 −Cr2 O3 . Conversion typically takes place at a reactor inlet temperature of 300–360 ◦ C and outlet temperature of 400–500 ◦ C if the shift reactor is used to increase H2 by the addition of steam. Operating at high temperature increases the reaction rate but reduces the equilibrium conversion. At temperatures below 300 ◦ C, Fe2 O3 −Cr2 O3 catalysts become very inefficient for WGS conversion. 2) Low-temperature WGS. The need for low-temperature WGS is mainly one of increasing the equilibrium conversion to produce more H2 . Low-temperature WGS catalysts are based on Cu–ZnO/Al2 O3 . Operation takes place at lower temperatures, typically in the range 190–250 ◦ C. As the temperature is lowered, the equilibrium improves, but the reaction rate decreases. For compact applications, noble metal catalysts can be employed for WGS, and some alternative catalyst types have been reviewed by Ratnasamy and Wagner [25]. All WGS catalysts are sensitive
69
70
3 Synthesis Gas Production, Cleaning, and Conditioning
to sulfur poisoning, and upstream syngas cleaning not only protects the Fischer–Tropsch catalyst, but also the WGS catalyst.
3.7 Air Separation Unit
In adiabatic oxidative reforming, as well as in gasification, a choice of oxidant has to be made. The trade-off between air and pure O2 is complex. When air is used, the syngas volume is significantly increased by the N2 in the air, which remains as a diluent in the syngas. This decreases the efficiency and increases the size of all downstream equipment in the gas loop. When pure O2 is used, the diluting effect of N2 is avoided, but it requires an ASU to produce the pure O2 . All current large-scale industrial applications of Fischer–Tropsch technology employ pure O2 for syngas production. The main compounds in dry air on a molar basis are N2 (78.08%), O2 (20.95%), and Ar (0.93%). Other compounds include the noble gases He, Ne, Kr, and Xe (25 ppm), CO2 (350 ppm), and CH4 (3 ppm). Water is present at variable levels, as are trace quantities of other compounds. The main elements of an ASU are shown in Figure 3.6 [26]. The air is compressed to a pressure that is sufficient to partially liquefy the air on adiabatic expansion, typically around 0.6–0.7 MPa. Before the air can enter the cryogenic section, water and CO2 must be removed, because these compounds will freeze and produce solids during cryogenic cooling. Hydrocarbons and trace impurities (pollutant gases) must be removed to avoid the formation of explosive mixtures with O2 and to reduce corrosion. Adsorption systems are used in swing operation to achieve air purification. The clean air is liquefied in a refrigeration cycle. In most modern air separation plants, refrigeration takes place by a Claude cycle, or a variation of the Claude cycle. There is much similarity with the well-known Linde or Joule–Thompson cycle, but it is more efficient. Once the air has been liquefied, it can be distilled to purify the O2 . An integrated double-column system
O2 N2 Refrigeration Air
Compression
Purification
0.1 MPa
0.6 MPa
Cryogenic distillation Figure 3.6
Simplified process flow diagram of an air separation unit.
References
is employed, which consists of a low-pressure and a high-pressure column, the reboiler of the former serving as condenser for the latter. The boiling point difference between N2 (−195.8 ◦ C) and O2 (−182.7 ◦ C) is large, but the boiling point of Ar (−185.7 ◦ C) is very close to that of O2 . It is possible to obtain N2 in very high purity, but obtaining very pure O2 is more difficult. This is not a major issue for syngas production, although it implies that some inert material will be present as diluent in the gas loop, which has implications for the gas loop design (Chapter 5). The noble gases and N2 are by-products from air separation and can be recovered and marketed as chemical coproducts.
References 1. Aasberg-Petersen, K., Christensen, T.S.,
2.
3.
4.
5.
6.
7.
8.
9.
10.
Dybkjær, I., Sehested, J., Østberg, M., Coertzen, R.M., Keyser, M.J., and Steynberg, A.P. (2004) Synthesis gas production for FT synthesis. Stud. Surf. Sci. Catal., 152, 258–405. Stelter, S. (2001) The New Synfuels Energy Pioneers. A History of Dakota Gasification Company and the Great Plains Synfuels Plant, Dakota Gasification Company, Bismarck, ND. Bourbonneux, G. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 451–501. Protasio, S.F. and St. Pierre, G.R. (1985) The use of metals and metal oxides for sulfur capture in combustion systems. Coal Sci. Technol., 9, 467–484. Kwon, K.C., Crowe, E.R., and Gangwal, S.K. (1997) Reactivity of metal oxide sorbents for removal of sulfur compounds from coal gases at high temperature and pressure. Sep. Sci. Technol., 32, 775–792. Christensen, T.S. (1996) Adiabatic prereforming of hydrocarbons – an important step in syngas production. Appl. Catal. A, 138, 285–309. Xu, J., Yeung, C.M.Y., Ni, J., Meunier, F., Acerbi, N., Fowels, M., and Tsang, S.C. (2008) Methane steam reforming for hydrogen production using low water-ratios without carbon formation over ceria coated Ni catalysts. Appl. Catal. A, 345, 119–127. Fan, M.-S., Abdullah, A.Z., and Bhatia, S. (2009) Catalytic technology for carbon dioxide reforming of methane to synthesis gas. ChemCatChem, 1, 192–208. Bakkerud, P.K. (2005) Update on synthesis gas production for GTL. Catal. Today, 106, 30–33. Howard-Smith, I. and Werner, G.J. (1976) Coal Conversion Technology, Noyes, Park Ridge, NJ.
11. Mangold, E.C., Muradaz, M.A., Ouellette, R.P.,
12.
13.
14. 15.
16.
17.
18. 19.
20.
21.
22.
Rarah, O.G., and Cheremisinoff, P.N. (1982) Coal Liquefaction and Gasification Technologies, Ann Arbor Science Publishers, Ann Arbor. Kuo, J.C.W. (1984) in The Science and Technology of Coal and Coal Utilization (eds B.R. Cooper and W.A. Ellingson), Plenum Press, New York, pp. 163–230. Rezaiyan, J. and Cheremisinoff, N.P. (2005) Gasification Technologies. A Primer for Engineers and Scientists, Taylor & Francis, Boca Raton. Higman, C. and Van der Burgt, M. (2008) Gasification, 2nd edn, Elsevier, Amsterdam. McMillen, D.F. and Golden, D.M. (1982) Hydrocarbon bond dissociation energies. Ann. Rev. Phys. Chem., 33, 493–532. Berkowitz, J., Ellison, G.B., and Gutman, D. (1994) Three methods to measure RH bond energies. J. Phys. Chem., 98, 2744–2765. Blanksby, S.J. and Ellison, G.B. (2003) Bond dissociation energies of organic molecules. Acc. Chem. Res., 36, 255–263. Egloff, G. (1937) The Reactions of Pure Hydrocarbons, Reinhold, New York. Bunt, J.R. and Waanders, F.B. (2008) An understanding of the behaviour of a number of element phases impacting on a commercial-scale Sasol-Lurgi FBDB gasifier. Fuel, 87, 1751–1762. Collot, A.-G. (2006) Matching gasification technologies to coal properties. Int. J. Coal. Geology, 65, 191–212. Wilson, P.J. Jr. and Wells, J.H. (1945) in Chemistry of Coal Utilization (ed. H.H. Lowry), John Wiley & Sons, Inc., New York, pp. 1371–1481. Richardson, I.M.J. and O’Connell, J.P. (1975) Some generalizations about processes to absorb acid gases and mercaptans. Ind. Eng. Chem. Process. Des. Dev., 14, 467–470.
71
72
3 Synthesis Gas Production, Cleaning, and Conditioning 23. Legrand, C. and Castel, J. (2001) in Petroleum
25. Ratnasamy, C. and Wagner, J.P. (2009) Water
Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 575–621. 24. Dry, M.E. and Steynberg, A.P. (2004) Commercial FT process applications. Stud. Surf. Sci. Catal., 152, 406–481.
gas shift catalysis. Catal. Rev.-Sci. Eng., 51 (3), 325–440. 26. Gunardson, H. (1998) Industrial Gases in Petrochemical Processing, Marcel Dekker, New York.
73
4 Fischer–Tropsch Synthesis 4.1 Introduction
Fischer–Tropsch synthesis collectively refers to processes for the conversion of synthesis gas to synthetic crude oil (syncrude). The name pays tribute to the inventors, the German scientists Franz Fischer and Hans Tropsch, who were awarded a patent on 22 July 1925 for the catalytic conversion of carbon monoxide and hydrogen into heavier hydrocarbons and oxygenates [1]. In its simplest form, the main reactions during Fischer–Tropsch synthesis can be represented by Equations 4.1–4.6: Alkenes: nCO + 2nH2 → (CH2 )n + nH2 O
(4.1)
Alkanes: nCO + (2n + 1)H2 → H(CH2 )n H + nH2 O
(4.2)
Alcohols: nCO + 2nH2 → H(CH2 )n OH + (n − 1)H2 O
(4.3)
Carbonyls: nCO + (2n − 1)H2 → (CH2 )nO + (n − 1)H2 O
(4.4)
Carboxylic acids: nCO + (2n − 2)H2 → (CH2 )nO2 + (n − 2)H2 O, n > 1
(4.5)
Water gas shift: CO + H2 O → CO2 + H2
(4.6)
The syncrude produced by Fischer–Tropsch synthesis, like conventional crude oil, does not refer to a single product. The composition of the syncrude depends on the Fischer–Tropsch catalyst and how the synthesis is conducted. As a consequence, the synthesis step directly influences the syncrude quality. The syncrude composition can be manipulated with Fischer–Tropsch refinery design in mind, and it is not an immutable input parameter. In fact, efficient Fischer–Tropsch refinery design depends on the judicious selection and manipulation of the syncrude properties to best match the refining objectives. However, Fischer–Tropsch synthesis imposes some restrictions on the extent to which the syncrude composition can be manipulated. For this reason, it is important to have a fundamental understanding of Fischer–Tropsch synthesis. In this chapter, the parameters that govern Fischer–Tropsch product selectivity are discussed. The focus is on opportunities that exist for the manipulation of Fischer–Tropsch synthesis to achieve a specific refining objective. The Fischer–Tropsch catalyst and process is viewed as a tunable entity that can be selected, much like crude oil selection prior to conventional crude oil refinery design. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
74
4 Fischer–Tropsch Synthesis
4.2 Fischer–Tropsch Mechanism
Despite the seeming simplicity of carbon monoxide hydrogenation during Fischer–Tropsch synthesis, the reaction network is complex and the mechanism is not resolved. Nevertheless, the mechanism is a good place to start to describe Fischer–Tropsch synthesis, since it highlights many of the parameters that influence the composition of the Fischer–Tropsch syncrude. Detailed discussions on the mechanism of Fischer–Tropsch synthesis can be found by many of the famous Fischer–Tropsch researchers, such as Anderson [2], Dry [3], and Davis [4]. The Fischer–Tropsch reaction is initiated by the adsorption of carbon monoxide (CO) on the catalyst surface. Once that has happened, chain propagation and chain termination can take place by various routes. Chemisorption of CO is therefore an important step in all mechanisms. Differences in mechanistic descriptions are mainly related to the nature of the monomer that is formed on the catalyst surface (Figure 4.1) and the way in which chain growth takes place (Figure 4.2). The original carbide mechanism proposed by Fischer and Tropsch suggested that CO is dissociatively adsorbed on the catalyst surface to form metal carbides. The carbon atoms of the metal carbides are then individually hydrogenated to the adsorbed CH2 species, which can combine to form products. One very attractive feature of this description is the M−CH2 −CH2 −M intermediate (Figure 4.2a). If chain growth occurs by insertion into one of the M−CH2 bonds, desorption would yield a cycloalkene and hydrogenation would yield a cycloalkane, thereby explaining the cyclic aliphatic hydrocarbons in the Fischer–Tropsch syncrude.
O C (a)
+H2
+H(ad)
C
CH
+H(ad)
CH2
−H2O
O O C
C (b)
+H2
C
+H2
CH2
−H2O
O C
+H2
H
OH C
(c) O
O H C
H
C
(d) Figure 4.1 Differences in the mechanistic representation of the monomer responsible for chain growth during Fischer–Tropsch synthesis. Representations (a) and (b) are variations of the original carbide mechanism that involves CO dissociation on the catalyst surface to produce a metal–carbon
species. The metal–carbon species is subsequently hydrogenated. The oxygenate mechanism (c) and the formaldehyde variation of the oxygenate mechanism (d), both involve CO adsorption without dissociation. These representations differ in the way in which adsorbed CO is hydrogenated.
4.2 Fischer–Tropsch Mechanism
C
+2H2
C
CH2 CH2
CH2
CH2
(a) CH2 HC CH2
HC
CH2
CH2 +CH2(ad)
HC CH2
HC
CH3
CH2
HC CH
(b) H OH H OH C
C
H −H2O
OH
C
C
OH +H2
CH3
C
(c) H OH C (d)
+H2
+CO
O
C CH2
C CH2
OH +H2
C CH3
−H2O
OH
O H (e)
CH2
O
C
+H2
CH2
+H2
CH3
+CO
O
O
C CH3
C
CH3
−H2O
Figure 4.2 Difference in the mechanistic representation of chain growth during Fischer–Tropsch synthesis. Chain growth in the carbide mechanism (a) takes place through bonding of adsorbed metal=CH2 intermediates (the metal=CH2 is actually M−CH2 −M with the carbon sp3 hybridized). In a variation thereof (b), the initial bonding takes
place between adsorbed metal≡CH and metal=CH2 intermediates. In the original oxygenate mechanism (c), chain growth takes place through bonding of the metal–enol species, but in a variation thereof (d), chain growth by direct CO insertion is possible. In the formaldehyde mechanism (e), chain growth is also by direct CO insertion.
Many aspects of the original carbide mechanism have since been disproved, but it has been difficult to unambiguously rule out surface carbon as an intermediate [2]. With advances in surface science instrumentation, it was found that CO adsorption on single-crystal metal surfaces indeed produced a surface covered in carbon, with only a little oxygen being present. This led to a revival of mechanistic descriptions invoking chemisorbed carbon. However, it is unlikely that dissociative CO adsorption is a dominant pathway in Fischer–Tropsch synthesis, and it is more likely that Fischer–Tropsch synthesis takes place through an oxygenate intermediate [4]. A detailed oxygenate (enol-based) mechanism was proposed by Storch, Golumbic, and Anderson [5]. In the original proposal, chain growth took place by bonding of metal–enol species on the catalyst surface. This required CO to be adsorbed and hydrogenated before it can participate in chain growth. One possible variation of the mechanism involves direct CO insertion followed by hydrogenation. Formaldehyde intermediates as surface oxygenates have been suggested by the work of Fahey [6], as well as the observed similarities between Fischer–Tropsch synthesis and hydroformylation. This mechanistic description involves direct CO insertion. The main difference with respect to the enol-based mechanism is in the nature of the monomer, which is the product from hydrogenation by dissociated hydrogen (H, not H2 ). It was pointed out by Maitlis that the coupling of sp2 - and sp3 -hybridized carbon species are energetically more favorable than the coupling of two sp3 -hybridized carbon species [7].
75
76
4 Fischer–Tropsch Synthesis
This led to the suggestion that the insertion takes place through an alkenyl mechanism, where the first step involves the coupling of metal≡CH and metal=CH2 intermediates to obtain a metal−CH=R working species. The similarity between the key intermediates in the proposals of Maitlis (metal−CH=R) and Fahey (metal−CH=O) is clear and it highlights the importance of having an sp2 -hybridized carbon for chain growth. Despite these differences in the detailed description of Fischer–Tropsch synthesis, a simplified reaction network may be developed on the basis of the oxygenate mechanism (Figure 4.3). This is useful to illustrate the product diversity obtained during Fischer–Tropsch synthesis, without claiming that it is mechanistically accurate. It also enables some general deductions to be made about the parameters that can influence product selectivity. The main chain termination reactions leading to final products are hydrogenation and desorption. Any catalyst properties or operating conditions that favor these termination steps will result in a lighter product. Conversely, any catalyst properties or operating conditions that favor CO adsorption will increase the concentration of adsorbed species and thereby increase the chance of chain propagation to result in a heavier product. It can also be seen that alkanes, 1-alkenes, primary alcohols, carboxylic acids, and aldehydes are all primary Fischer–Tropsch products. Thus far, the description of Fischer–Tropsch synthesis focused only on CO adsorption as a method of chain initiation. Although this may be true at low conversion and high CO partial pressure, it is possible for the Fischer–Tropsch products to re-adsorb on the catalyst. In fact, readsorption of oxygenates compete successfully with CO adsorption and, at high enough temperature, the oxygenate interconversion reactions are at equilibrium (for example,
CO CO (g)
H2 (g)
H2C
O
CH3CHO
CH3OH
C2H4 CH2
H2 (g)
H2 (g)
HC
C O
CH2
H2 (g)
CO H2O (g)
CH
O H2 (g)
H2 (g)
H2 (g)
CH4
CH3COOH
HC CH2
CH3
CH3
CH2
H2O (g)
C2H5OH
Figure 4.3 Simplified reaction network based on the oxygenate mechanism of Fischer–Tropsch synthesis, showing the formation of the primary products, namely alkanes, alkenes, alcohols, aldehydes, and carboxylic acids.
CH3 etc. CH
H2O (g) H2 (g)
C2H6
CH3
CO
C O
4.3 Fischer–Tropsch Product Selectivity
Equations 4.7 and 4.8) [8, 9]. C2 H5 OH + H2 O CH3 COOH + H2
(4.7)
C2 H5 OH CH3 CHO + H2
(4.8)
From Fischer–Tropsch refining catalysis studies, we know that oxygenates often dominate surface chemistry by competitive adsorption against hydrocarbons [10]. It is hardly surprising to find that the surface chemistry during Fischer–Tropsch synthesis is also subject to such competitive adsorption of oxygenates. Davis [4] pointed out that the iron-based Fischer–Tropsch conversion of alcohols to alkanes with one less carbon than the alcohol may take place through a carboxylate intermediate. This would also explain the direct formation of CO2 and account for chain initiation by CO2 . It has further been reported [3] that the introduction of acetic acid during Fischer–Tropsch synthesis over iron-based catalysts leads to increased production of acetone. This is consistent with metal carboxylate decomposition. Of the main oxygenate classes formed during Fischer–Tropsch synthesis, only ketones and sec-alcohols are secondary products. These products can be explained by the competitive adsorption and reaction of primary oxygenates. Ketone formation by carboxylate decomposition has already been mentioned, and since most of the ketones are methyl ketones, acetic acid plays an important role in ketone formation. The sec-alcohols can be produced by partial hydrogenation of the ketones. An interesting inverse relationship was found between the selectivity to internal alkenes and carboxylic acid production [11]. This inverse relationship between internal alkenes and carboxylic acids seems to generally hold true, irrespective of the catalyst type. It has been postulated that the same sites responsible for double bond isomerization of 1-alkenes to internal alkenes can be converted to sites responsible for carboxylic acid formation [11]. However, the probability of double bond isomerization by readsorption of the alkene on the Fischer–Tropsch catalyst seems to be chain length dependent. In studies with 14 C-labeled compounds, Davis and coworkers found that pentene isomers did not undergo double bond isomerization during Fischer–Tropsch synthesis when co-fed with syngas, but that decene did [12]. An alternative interpretation of these results focusing on oxygenates is that the increased conversion of carboxylic acids follows the following reaction sequence [8]: carboxylic acids → ketones → sec − alcohols → internal alkenes. If in fact oxygenate conversion chemistry is as important in Fischer–Tropsch synthesis as it is in the refining catalysis of syncrude [10], carboxylic acid and ketone interconversion reactions can be used to explain the formation of internal alkenes and possibly some of the branching in Fischer–Tropsch syncrude. Mechanistically, the picture is more complex, and multiple reaction pathways have been suggested to describe selectivity effects like branching [13]. 4.3 Fischer–Tropsch Product Selectivity
The mechanism highlighted some important aspects that govern the product selectivity during Fischer–Tropsch synthesis, namely, the probability of chain growth, hydrogenation activity, and readsorption chemistry of products. The product selectivity determines the composition of the syncrude and directly influences syncrude refining and refinery design.
77
78
4 Fischer–Tropsch Synthesis
4.3.1 Probability of Chain Growth
The carbon number distribution obtained during Fischer–Tropsch synthesis is determined by the probability of chain growth on the catalyst, which is also called the α-value of the catalyst. If the reaction environment is kept constant, a relative comparison of α-values would be a direct measure of the likelihood that a Fischer–Tropsch catalyst would catalyze chain propagation, rather than chain termination. There is a well-defined interrelationship between the carbon number distribution of the products formed during Fischer–Tropsch synthesis and the probability of chain growth over the catalyst. The carbon number n and the probability of chain growth α are related to the molar fraction of each carbon number in the product xn by Equation 4.9 [14]. xn = (1 − α)·α (n−1)
(4.9)
This relationship describes the carbon number distribution in the syncrude, which is called the Schulz–Flory equation. It has later been extended to account for branching by Anderson while retaining the same basic form [15]. The carbon number distribution described by the Schulz–Flory equation is commonly referred to as the Anderson–Schulz–Flory (ASF) distribution. In order to allow the calculation of the α-value from the molar fraction of each carbon number, the relationship is often expressed in logarithmic form (Equation 4.10). log xn = n·log α + log[(1 − α)/α]
(4.10)
The probability of chain growth can also be calculated from experimental data by taking the molar ratio of any two carbon numbers in the product (Equation 4.11) [3]. xn /xj = α (n−j)
(4.11)
Despite the mathematical elegance of the ASF description of the Fischer–Tropsch carbon number distribution, in practice there are deviations from the ASF distribution (Figure 4.4). Many explanations have been forwarded [16]. The three most prominent deviations are [14] the following: 1) Methane selectivity. The methane selectivity during Fischer–Tropsch synthesis is usually higher than that predicted by the ASF distribution. In the literature, various reasons for the high methane selectivity have been proposed, which include mass transfer limitations favoring thermodynamics (i.e., methane production), sites only active for hydrogenation, and hydrogenolysis. High methane selectivity is not entirely surprising if one considers the mechanism. The first chain growth step is bimolecular and more difficult than just hydrogenation of a metal=CH2 intermediate. Hydrogenation of metal=CH2 intermediates (Figures 4.1 and 4.2) would lead to desorption, unlike hydrogenation of metal−CH=R intermediates, which may not necessarily lead to desorption. Furthermore, hydrogenation of metal=CH2 intermediates would not only remove the adsorbed intermediate as methane but also prevent chain growth with an adjacently adsorbed specie. The hydrogenation of metal=CH2 not only increases methane yield but also decreases chain growth of other adsorbed intermediates.
ASF plot, log (xn)
−0.5
2.5
−1.0
α1
−1.5
2.0
1.5
α2
−2.0
1.0
−2.5
0.5 0
10
20 30 Carbon number, n
40
Composition of C3 and heavier (mass%)
4.3 Fischer–Tropsch Product Selectivity
50
Figure 4.4 Anderson–Schulz–Flory (ASF) semilogarithmic plot of the molar fraction of each carbon number (xn ) versus carbon number (n) and the corresponding mass percentage in the C3 and heavier fraction. The data is typical of a low-temperature Fischer–Tropsch syncrude.
2) C2 selectivity. The selectivity to C2 hydrocarbons is usually much lower than is predicted by the ASF distribution. In addition to the lower C2 selectivity, the ethene to ethane ratio is also much lower than anticipated from the alkene to alkane ratio of higher carbon numbers. Explanations offered in the literature focus mainly on secondary reactions of ethene. One explanation that is particularly appealing in its simplicity, is the doubling of chain growth probability of a C2 intermediate of the form M−CH2 −CH2 −M (Figure 4.2a), which allows growth from either end. Then the lower ethene to ethane ratio can be ascribed to the difference in desorption rate versus hydrogenation rate to cause desorption. However, in the case of high-temperature Fischer–Tropsch (HTFT) synthesis, it was pointed out by Weitkamp and Frye that the anomalously low C2 hydrocarbon yield did not represent a deviation from the Schulz–Flory equation (Equation 4.9) if one took the contribution of C2 oxygenates into account [8]. 3) Two α-value distribution. Low-temperature Fischer–Tropsch (LTFT) catalysts that produce a considerable amount of heavier hydrocarbons do not have a single α-value, but two. This is suggestive of a vapor-liquid equilibrium effect. The first α-value (α1 ) describes the carbon number distribution lower than C8 , while the second α-value (α2 ) describes the distribution of the heavier than C12 fraction. In between, there is a crossover region where there are different contributions κi of each α-value (Equation 4.12). xn = κ1 ·(α1 )(n−1) + κ2 ·(α2 )(n−1)
(4.12)
The notion of a single α-value, or even two different but constant α-values, is the product of mathematical convenience. In practice, the α-value seems to slowly increase with chain length and there is a slight curvature to ASF plots. A description of chain-length-dependent chain growth probability αn was developed by Botes (Equation 4.13) [17]: αn = [1 + τP + τO ·exp(−k·n)]−1
(4.13)
79
80
4 Fischer–Tropsch Synthesis
In this equation, the production of alkanes (paraffins) is given by the ratio of the rate of hydrogenation to the rate of chain growth τP . The production of alkenes (olefins) are given by the ratio of the rate of desorption to the rate of chain growth τO . However, the desorption rate is not constant, but decreases with increasing chain length, which is modeled as an Arrhenius exponent k multiplied by the chain length n. The model not only describes the chain-length-dependent change in α-value but also predicts the alkene to alkane ratio for each carbon number (Equation 4.14). (alkenes/alkanes)n = (τO /τP )·exp(−k·n)
(4.14)
This model description was validated for Fe-LTFT data and could be extrapolated to give accurate predictions of methane, ethene, and ethane selectivity. Although this chain-length-dependent desorption model is not a complete selectivity model, it accurately captures some of the relationships between chain length and selectivity to alkenes and alkanes. Oxygenates follow the same regular relationship as described for hydrocarbons. The alcohols apparently follow the same distribution as the (n + 1) hydrocarbons [3]. This relationship between alcohols and hydrocarbons seems to be valid for iron-based Fischer–Tropsch only and may be related to alcohol decomposition [4]. 4.3.2 Hydrogenation versus Desorption
The manner in which chain growth is terminated influences the Fischer–Tropsch product selectivity to a large extent. Chain growth can be terminated by desorption without complete hydrogenation, complete hydrogenation followed by desorption, and hydrogenolysis. Fischer–Tropsch products are not only hydrocarbons, but also include different classes of oxygenates. The retention of an oxygenate functionality in the growing chain presupposes that the chain is not fully hydrogenated. Like the alkenes, the relative concentration of oxygenates in each carbon number fraction decreases rapidly with increasing chain length. If we extend the principles introduced by the model description in Equation 4.13 to oxygenates, the different oxygenate classes can also be viewed as compounds produced mainly by desorption and not by hydrogenation. Although chain-length-dependent desorption can explain the decrease in oxygenate concentration with increasing chain length, it does not account for the different oxygenate classes. The partial hydrogenation of carbonyls to alcohols, or dehydrogenation of alcohols to carbonyls, does not necessarily lead to desorption, but it influences the oxygenate composition. There are also hydration and dehydration reactions to consider, as well as oxygenate interconversion reactions that require oxygenate adsorption. It can therefore be said that the hydrogenation and desorption activity affects chain growth and that the ratio of hydrogenation to desorption activity affects the product selectivity within each carbon number. In addition to hydrogenation, one should also consider the hydrogenolysis activity of the Fischer–Tropsch metal. The mechanism of hydrogenolysis over Fe and Co is similar to that of Ni, namely α-scission of the end-adsorbed species, but with the difference that Fe and Co are capable of deep hydrogenolysis [18]. Unlike Ni, where desorption of the two hydrogenolysis fragments occur after hydrogenolysis, Fe and Co have the ability to unzip the chain and produce
4.4 Selectivity Manipulation in Fischer–Tropsch Synthesis
methane as the main product. The hydrogenolysis activity of Co is greater than that of Fe, but both metals perform hydrogenolysis in the same way. 4.3.3 Readsorption Chemistry
The description of Fischer–Tropsch synthesis and the kinetics of Fischer–Tropsch synthesis usually focus only on water gas shift (WGS) compounds (H2 , CO, H2 O, and CO2 ) [14, 19–21]. This creates the false impression that the products produced during Fischer–Tropsch synthesis are effectively inert with respect to readsorption and reaction. Although this may be a reasonable assumption at high H2 and CO partial pressures, competitive adsorption takes place and becomes increasingly important as syngas conversion increases. One has to consider both the Fischer–Tropsch active metal and, where relevant, metal oxides present as unreduced metal sites, as promoters, or as support. Generally, CO is strongly adsorbed by metals and can be used to regulate hydrogenation selectivity at low concentration: for example, the selective hydrogenation of ethyne to ethene [22]. Alkenes and hydrocarbons in general are not very effective at competing with CO for metal site adsorption. Oxygenates are a different matter altogether. Oxygenate catalysis over Fischer–Tropsch catalysts may be responsible for many of the nontraditional Fischer–Tropsch products observed. For example, aromatics can be produced by carbonyl condensation, and esters can be produced by carboxylic acid and alcohol condensation. As pointed out before, such oxygenate reactions may also introduce branching, form internal alkenes, ketones, and sec-alcohols.
4.4 Selectivity Manipulation in Fischer–Tropsch Synthesis
Selectivity manipulation has to take catalyst activity into account, although Fischer–Tropsch catalyst activity per se is unimportant to Fischer–Tropsch syncrude refining. Fischer–Tropsch selectivity can be manipulated by the catalyst formulation, operating conditions, and reaction engineering principles. The product selectivity is also influenced by catalyst deactivation (Section 4.5). The objective of this section is to show how each Fischer–Tropsch technology has manipulated these variables to drive syncrude selectivity. Conceptually, it will also be shown how new Fischer–Tropsch technology can be devised to tailor the syncrude to a specific refining objective. 4.4.1 Fischer–Tropsch Catalyst Formulation
Literature on the synthesis and performance of catalysts for Fischer–Tropsch synthesis is abundant [3, 5, 19, 23, 24]. Of the many Fischer–Tropsch active materials, the most often studied are Fe, Co, Ni, Ru, and ThO2 . Good Fischer–Tropsch activity requires metals that allow dissociative CO adsorption and have good H2 adsorption behavior. A nice summary of the activity of different transition metals for Fischer–Tropsch synthesis can be found in the work
81
82
4 Fischer–Tropsch Synthesis
by Perego and coworkers from Eni [25]. Only Fe and Co are industrially applied. By selecting the Fischer–Tropsch active metal that forms the basis for the catalyst formulation, many of the important catalyst properties are determined. Some important repercussions are the following: 1) Hydrogenation activity. The different metals active for the Fischer–Tropsch reaction have different hydrogenation activities. Fe is the least active for hydrogenation, followed by Co, Ni, and Ru. Ni and Ru are both very active methanation catalysts and produce heavy hydrocarbons only at low temperatures. At similar conditions, the products from Fe is more olefinic and contains more oxygenates than that from Co, because Fe is a less hydrogenating metal. This has important implications for the refinery. Alkenes enable refining technologies that are not accessible to alkanes. Oxygenates are also reactive molecules, but require more careful selection of refining technologies and catalysts [10, 26]. 2) WGS activity. An important difference between Fe- and Co-based Fischer–Tropsch catalysts is the ability of Fe-based catalysts to catalyze the WGS reaction (Equation 4.6). Co-based Fischer–Tropsch catalysts have virtually no WGS activity. Although this difference has little impact on the refinery, the implications for gas loop design are tremendous. The WGS reaction allows the interconversion of CO, H2 , CO2 , and H2 O. Over Fe-based Fischer–Tropsch catalysts, CO2 can be consumed and it is possible to conduct Fischer–Tropsch synthesis with H2 -rich syngas and CO2 . Over Co-based Fischer–Tropsch catalysts, CO2 is a final product. The H2 :CO operating range of Fe-based Fischer–Tropsch catalysts is therefore much wider than that of Co-based catalysts [25]. 3) Sensitivity to promoters. One way in which the properties of the Fischer–Tropsch active metal can be changed is by adding promoters. The promoters act as modifying agents of the catalyst. Co-based Fischer–Tropsch catalysts are not sensitive to promoters during high-pressure operation [9] or to the support being used [27]. Fe-based Fischer–Tropsch catalysts are affected by promoters, and the addition of strong alkaline salts of Na and K are required for the preparation of useful Fischer–Tropsch catalysts. Alkali promotion increases amount of CO adsorbed and thereby the chain growth probability. With increasing alkaline addition on supported and unsupported Fe catalysts, the activity passes through a maximum, while the α-value and the carboxylic acid selectivity increase [3, 9]. The changes in carbon number distribution and oxygenate content of the syncrude affect refinery design. 4) Catalyst stability. Fischer–Tropsch catalyst deactivation affects both activity and selectivity during synthesis. These changes in syncrude composition affect the refinery, albeit not always in a negative way [28]. The general perception is that Fe-based Fischer–Tropsch catalysts deactivate at a higher rate than their Co-based counterparts. This perception is challenged by some old and new time-on-stream performance data, which shows that some Fe-LTFT catalysts stabilize after some time on stream with little further deactivation [5, 29, 30]. Catalyst deactivation is discussed separately (Section 4.5). 5) Support material. The Fischer–Tropsch active metal may be reduced as a bulk-activity catalyst, as in the case of many iron-based Fischer–Tropsch catalysts. The Fischer–Tropsch active metal may also be supported on a carrier material, typically when the metal is more expensive, like Co, Ni, and Ru. Generally, a metal oxide support (SiO2 , Al2 O3 , TiO2 , or ZrO2 ) is employed. The support material affects catalyst performance through its effect on metal dispersion, accessibility, mechanical strength, attrition resistance, and the strength of the metal–support interaction.
4.4 Selectivity Manipulation in Fischer–Tropsch Synthesis
6) Bifunctionality. The support material may also be used to modify the Fischer–Tropsch reaction by providing another conversion pathway. Zeolites and acidic supports have been investigated for their ability to modify the synthesis reaction in various ways. Shape selectivity was imposed on Fischer–Tropsch synthesis by placing cobalt clusters active for Fischer–Tropsch synthesis within zeolite pores [31]. The acidity of zeolites and other acidic materials has been employed to modify the carbon number distribution and degree of branching by acid catalysis in parallel with Fischer–Tropsch synthesis by using an acidic support for the Fischer–Tropsch catalyst [32, 33]. Rather than combining both functions in one catalyst, a mixture of two catalysts was also used for this purpose [34–37]. Separate Fischer–Tropsch synthesis and acid catalysis have been proposed to overcome the challenge of matching Fischer–Tropsch synthesis conditions with that of acid-catalyzed conversion, as well as to overcome deactivation by promoter migration and differences in deactivation rate [35, 38]. However, when the synthesis and acid catalysis are conducted in separate reactors, the acid catalysis can be considered a first refining conversion step. 4.4.2 Fischer–Tropsch Operating Conditions
Under typical operating conditions, synthesis reactions are far from being in thermodynamic equilibrium [3]. In this respect, short-chain oxygenate interconversion by readsorption during HTFT synthesis is an exception. The operating conditions can be used to manipulate the carbon number distribution in various ways but, irrespective of how the operating conditions are manipulated, broadly speaking the ASF distribution is not violated. The operating conditions can also be used to manipulate the product selectivity within each carbon number. The impact of the main operating variables is as follows (Table 4.1): 1) Temperature. Temperature affects the desorption rate of products on the catalyst surface. Desorption is an endothermic process and an increase in temperature increases the desorption rate and thereby chain termination to less hydrogenated products (alkenes and oxygenates). The rate of hydrogenation, which can also cause chain termination to more Table 4.1
Influence of Fischer–Tropsch operating conditions on product selectivity.
Selectivity parameter
Carbon number distribution Methane selectivity Alkene selectivity Oxygenate selectivity Aromatic selectivity Syngas conversion a Some b The
Operating parameter being increased Temperature
Pressure
Space velocity
H2 : CO ratio
Lower α-value Increases –b –b Increases Increases
Higher α-value Decreases –b Increases –b Increases
No changea Decreases Increases Increases Decreases Decreases
Lower α-value Increases Decreases Decreases Decreases –b
change is possible if secondary reactions like re-incorporation, hydrogenation, or cracking are significant. direction of change depends on a more complex relationship.
83
84
4 Fischer–Tropsch Synthesis
2)
3)
4)
5)
hydrogenated products (alkanes), is increased by an increase in temperature too. The net effect is to lower the α-value and increase the desorption rate, resulting in a carbon number distribution that favors shorter chain products. The relative increase in desorption rate without hydrogenation and hydrogenation leading to desorption determines whether the product will become more olefinic or not. The reaction rate of oxygenates by readsorption and interconversion also increases with temperature. Pressure. The total pressure, in combination with the syngas composition, determines the partial pressure of the reagents, H2 and CO. Since CO is more strongly adsorbed onto the catalyst than H2 , the concentration of CO on the catalyst surface increases when the pressure is increased. High CO concentration on the catalyst surface promotes chain growth and increases the observed α-value of the catalyst. A high CO concentration also favors CO incorporation and consequently the production of oxygenates. Pressure has another important consequence that is related to the difference in the kinetic description of Feand Co-based Fisher–Tropsch synthesis. At high pressure and low to moderate per pass conversion (low to moderate H2 O partial pressure), the productivity of Fe-based catalysts is higher than that of Co-based catalysts [39]. Space velocity. By increasing the space velocity, secondary reactions are reduced. The products from Fischer–Tropsch synthesis spend less time in contact with the catalyst and the probability of readsorption and reaction becomes less. This reduces hydrogenation of alkenes and oxygenates, as well as the interconversion of oxygenates. Increasing the space velocity also affects conversion and, by decreasing conversion, the CO partial pressure at the outlet of the reactor is higher, which further helps to reduce secondary reactions. One would not expect the carbon number distribution to be meaningfully affected unless secondary reactions involving either re-incorporation or cracking are significant. Usually such reactions are not, but instances where they are, have been reported [40]. However, with increase in space velocity, direct CO hydrogenation to produce methane seems to be reduced somewhat relative to syncrude production. Synthesis gas composition. During Fischer–Tropsch synthesis, H2 and CO are consumed in a ratio that is dependent on the products being formed (Equations 4.1–4.6). As the products become heavier, the usage ratio approaches 2, irrespective of the compound class being formed. Ideally, the synthesis gas composition of the feed should match the usage ratio because, if it does not, the H2 :CO ratio will change during synthesis. The effect of the H2 :CO ratio is therefore not only an effect of the ratio in the syngas feed but also that of the usage ratio during synthesis (see also Section 4.4.3). In general, if the H2 :CO ratio increases, the driving force for hydrogenation increases. As hydrogenation is increased, the probability of chain growth decreases, as well as the selectivity to alkenes and oxygenates. Catalyst activation. Although Fischer–Tropsch catalyst activation is strictly speaking not an operating condition, it has an impact on the overall performance, selectivity, and deactivation behavior of the catalyst.
4.4.3 Fischer–Tropsch Reaction Engineering
The aim of reaction engineering is to find the best combination of catalyst, reactor, and operating philosophy to attain a specific goal. Although reaction engineering principles can be employed to
4.4 Selectivity Manipulation in Fischer–Tropsch Synthesis
manipulate Fischer–Tropsch selectivity, reactor design and operation cannot be manipulated as independent variables. It is emphatically stated that the Fischer–Tropsch catalyst formulation, operating regime, and reactor technology must be developed together and with a specific syncrude composition in mind to achieve an optimal Fischer–Tropsch technology. The different parameters that are typically considered and optimized during reaction engineering are discussed individually, but they are by no means independent. 1) Reactor type. There are three main reactor types that are employed for Fischer–Tropsch synthesis, namely, fixed bed, slurry bed, and fluidized bed. The variables to consider in reactor selection have been discussed in detail by Sie and Krishna [41]. Fixed bed reactors approximate plug flow reactors (PFRs). The composition and conditions of the reaction medium change continuously along the length of the catalyst bed. The driving force for synthesis is maximized and, in the absence of heat and mass transfer limitations, fixed bed reactor technology is the most efficient reactor type for synthesis. Slurry bed and fluidized bed reactors both approximate continuous stirred tank reactors (CSTRs) better than PFRs. The nature of the design influences the degree of mixing, and one can expect that a fixed fluidized bed would display more CSTR behavior than a circulating fluidized bed. The choice of reactor has many consequences (Table 4.2) and the impact on refining is significant. Among others, at the same level of conversion, one can expect the product from a fixed bed reactor to be more hydrogenated (less alkene and oxygenates) than that from a slurry bed or fluidized bed reactor. For example, during Fe-LTFT synthesis at 248 ◦ C, 0.8 MPa, and 50–60% CO conversion, the C2 −C4 alkene to alkane ratio for fixed bed synthesis was 0.09, 0.9, and 1.2 respectively, whereas for slurry bed synthesis it was 3.7, 5.6, and 4.5 [42]. 2) Reaction phase. The reaction phase is not limiting in fixed bed synthesis, but it limits the α-value and operating range in slurry bed and fluidized bed synthesis. Fluidized bed operation requires the reaction to take place in the gas phase. The α-value of fluidized bed Fischer–Tropsch synthesis must out of necessity be low (α < 0.70) to avoid condensation Table 4.2
Characteristics of the main reactor types used for Fischer–Tropsch synthesis.
Description
Nature of the reactor Reaction phase Catalyst particle size (mm) Mass transfer limitation Heat transfer limitation On-line catalyst replacement Catalyst mechanical strength Catalyst–product separation Scale-up risk (lab to plant) Scale-up economy of scale Feed poisoning Feed turn down limitation
Fixed bed Multitubular
Microchannel
PFR g or g+l >2 High High No Low Easy Low Medium–Low Local None
PFR g or g+l <0.1 Low Low No Low Easy Low Low Local None
Slurry bed
Fluidized bed Fixed fluidized Circulating
CSTR g+l <0.1 Medium Low Possible Medium Difficult Medium High Global Catalyst settling
CSTR g <0.1 Medium-low Medium-low Possible High Fairly easy Medium Very high Global Defluidization
CSTR g <0.1 Medium-low Medium-low Possible High Fairly easy Medium High Global Defluidization
85
86
4 Fischer–Tropsch Synthesis
3)
4)
5)
6)
and bed defluidization at the reaction conditions. The converse is true of slurry bed operation, which requires the presence of a significant liquid-phase product. The α-value of slurry bed Fischer–Tropsch synthesis must, out of necessity, be high to allow the formation of a liquid-phase product at the reaction conditions. The manipulation of the carbon number distribution to improve refining is therefore restricted by the reaction phase requirement by some reactor types. Mass transfer. The intrinsic rate of Fischer–Tropsch synthesis under typical industrial operating conditions is high and mass transfer can be limiting. Mass transfer limitations may also affect selectivity if there is a significant change in H2 :CO ratio in the catalyst particle or if slow product diffusion increases the probability of secondary reactions. In LTFT synthesis, where the liquid phase increases mass transfer resistance compared to gas-phase HTFT synthesis, catalyst productivity can be increased by employing a reactor technology that reduces mass transfer resistance. For example, in LTFT synthesis mass transfer is improved on moving from a multitubular fixed bed to slurry bed to washcoated structured supports and microchannel fixed bed operation [43, 44]. Mass transfer is also affected by the hydrodynamic regime, which plays an important role especially in slurry bed reactors [41]. Heat transfer. Fischer–Tropsch synthesis is very exothermic (H298K = −140 to −160 kJ · mol−1 CO converted, depending on the product). Heat transfer can quickly become constraining, and Fischer–Tropsch reactors are effectively massive heat exchangers. Providing sufficient heat transfer area is a major consideration in the design of Fischer–Tropsch reactors. The effectiveness of heat transfer determines the temperature profile in the reactor as well as the temperature profile within each catalyst particle. The temperature profile in turn affects selectivity, which has an impact on the refinery. In most industrial designs, the reaction heat is removed by steam production. In this respect, HTFT synthesis has an important advantage over LTFT synthesis, since the steam produced by reaction at around 320–340 ◦ C is high-pressure steam, whereas the steam produced by reaction at 220–230 ◦ C is only medium-pressure steam. Catalyst replacement. The loading and the unloading of catalyst in fixed bed reactors have to take place off-line. This is a time-consuming task and fixed bed reactor technology is best employed with Fischer–Tropsch catalysts that have a low deactivation rate or are amenable to in situ regeneration. An advantage of slurry bed and fluidized bed reactors is that the catalyst can be replaced on-line during normal operation. By definition, this allows control over the average catalyst age in the reactor, which in turn can be manipulated to increase the overall efficiency of synthesis and refining. It also allows the refinery to be designed for a fairly constant syncrude composition. In order to achieve the same level of stability over time with fixed bed reactor technology, multiple reactors must be employed in parallel and the catalyst age distribution must be controlled to approximate a constant average catalyst age. Catalyst mechanical strength. Depending on the reactor type, different demands are placed on the mechanical strength of the catalyst. This in turn affects mass transfer and activity. It has an impact on the refinery only if there is significant mechanical attrition (Section 4.5.4). In fixed bed reactors, the crushing strength of the particles is important, but once loaded, the mechanical demands placed on the catalyst during operation are very small. In slurry bed and fluidized bed reactors, the catalyst is in constant movement and frequent impact of catalyst particles with each other and with the equipment is to be expected. This places limitations on catalyst design.
4.4 Selectivity Manipulation in Fischer–Tropsch Synthesis
7) Catalyst–product separation. In fixed bed reactors, the catalyst remains stationary in the reactor and catalyst–product separation is not an issue. In slurry bed and fluidized bed reactors, a separation step is required to remove the catalyst from the product. The high-temperature gas–solid separation by cyclones, which is required during fluidized bed operation, is similar to the cyclone separation used in fluid catalytic cracking technology. Although such separation is not without its challenges, gas–solid separation is fairly easy compared to the liquid–solid separation required during slurry bed operation. Catalyst–product separation is one of the main detractors of slurry bed technology. Inefficient catalyst separation has an impact on the refinery, since it requires the downstream units to cope with solids in the syncrude. Fine particles can form stable colloidal suspensions that can be destabilized during refining, causing deep-bed plugging of fixed bed reactors [45]. 8) Per pass conversion. The extent of CO conversion in the reactor will have an impact on many aspects of the reactor design, as well as on the overall gas loop design. Central to this is the usage ratio and the change in partial pressure of the WGS compounds. When the Fischer–Tropsch catalyst has little WGS activity, the syngas feed must have a H2 :CO ratio close to the usage ratio in the synthesis reactor (Table 4.3). If this is not the case, the H2 :CO ratio will change with increasing conversion, ultimately affecting selectivity and catalyst stability. The per pass conversion has to be controlled at a level that will ensure sufficient productivity in the reactor (the kinetics depends on H2 and CO partial pressure) and avoid catalyst deactivation (Section 4.5). When the Fischer–Tropsch catalyst is WGS active, the H2 :CO ratio changes in response to the composition of the syngas, and CO2 can also be employed as carbon source. The same principles apply, but for WGS-active catalysts the stoichiometric ratio must be expressed in terms of CO2 as well (Equation 4.15), which can be rearranged to express it in terms if the Ribblett ratio (Equation 4.16) [9]. Stoichiometric ratio: (H2 − CO2 )/(CO + CO2 ) ≈ 2
(4.15)
Ribblett ratio: (H2 )/(2CO + 3CO2 ) ≈ 1
(4.16)
9) Reactor configuration. Fischer–Tropsch reactors, like other reactors, can be configured in different ways by placing reactors in series and/or parallel. When reactors are employed
Table 4.3
Usage ratio of H2 :CO for the production of different Fischer–Tropsch primary products.
Product
H2 :CO usage
Alkanes Alkenes Alcohols Carbonyls Carboxylic acids a Formic
(2n + 1)/n 2 2 (2n − 1)/n (2n − 2)/n
H2 :CO usage ratio based on product carbon number C1
C2
C4
C10
C25
C50
3 – 2 1 –a
2.5 2 2 1.5 1
2.25 2 2 1.75 1.5
2.1 2 2 1.9 1.8
2.04 2 2 1.96 1.92
2.02 2 2 1.98 1.96
acid production from WGS compounds requires CO + H2 O or CO2 + H2 .
87
88
4 Fischer–Tropsch Synthesis
in series, intermediate product separation and conditioning can be used to manipulate the selectivity. In the case of reactors that approximate CSTR behavior, using reactors in series can increase productivity. This follows from first principles: PFR behavior can be approximated by multiple CSTRs in series. 10) Scale-up. There are two scale-up issues related to the reactor technology. The first is the risk involved in scaling up from small-scale to large-scale equipment. The second is the economy of scale that can be achieved by reducing the reactor cost per unit product produced. The least scale-up risk is found in fixed bed reactors, where a single-tube or single-channel experiment at the appropriate scale is representative of the whole. Conversely, the least economy of scale can be achieved with fixed bed reactors. In absolute terms, the potential cost saving of using a slurry bed reactor compared to a fixed bed reactor for LTFT synthesis is large, but the impact on the overall economics of the facility is small [41]. When evaluating the economy of scale that can be achieved during scale-up, it is also prudent to evaluate the complexity, robustness, and maintenance aspects associated with the scale-up of each reactor type. Project and production risks can be introduced by selecting an overly large reactor that is difficult to construct, transport, and maintain. Scale-up issues related to the Fischer–Tropsch reactor do not affect the refinery design. 11) Syngas production upsets. Syngas production upsets can affect the syngas feed rate, quality and/or composition, and/or quality. Changes in the syngas feed rate do not seriously affect the stability of a fixed bed reactor. Slurry bed reactors need sufficient syngas flow to keep the catalyst in suspension. Fluidized bed reactors are the least robust and there is a lower threshold beyond which bed defluidization takes place. A deterioration in syngas quality will have a localized effect on a fixed bed reactor (top of the catalyst bed will be affected), but will have a global effect on slurry bed and fluidized bed reactors (total catalyst bed will be affected). The impact of syngas composition depends mainly on the Fischer–Tropsch catalyst and not so much on the reactor technology.
4.5 Fischer–Tropsch Catalyst Deactivation
Fischer–Tropsch catalysts, like all other catalysts, deactivate over time. The mechanism of deactivation depends on the catalyst type and its use. Over the years, a significant body of literature has developed on catalyst deactivation. Bartholomew [46] grouped catalyst deactivation into six categories, most of which are applicable to Fischer–Tropsch catalysts [3]: 1) Poisoning, which is caused by the strong chemisorption of species on the catalyst. 2) Fouling, which is caused by the deposition of material onto the catalyst. 3) Thermal degradation, which can either be by sintering (loss of active area) or by the chemical transformation of catalytically active phases to inactive phases. 4) Loss of active material from the catalyst by volatilization or leaching, which is caused by reaction with the feed or by products from reaction. 5) Chemical transformation of the catalytically active phase or support material by interaction with the feed to render the catalyst less active. 6) Mechanical degradation of the catalyst particle by crushing, attrition, or erosion.
4.5 Fischer–Tropsch Catalyst Deactivation
4.5.1 Poisoning by Syngas Contaminants
All Fischer–Tropsch catalysts are deactivated by sulfur compounds. The nature of the sulfur compound plays a role in determining the rate of adsorption and thereby the rate of deactivation of the catalyst. Hydrogen sulfide (H2 S) is rapidly adsorbed, whereas carbonyl sulfide (COS) and thiols (Cx H2x+1 SH) are more slowly adsorbed, causing different sulfur distribution profiles in fixed bed applications [3]. Hydrogen sulfide and organic sulfur are permanent Fischer–Tropsch catalyst poisons, which means that Fischer–Tropsch catalyst poisoning by sulfur is not reversible by reduction with a sulfur-free gas. It has been reported that sulfur dioxide (SO2 ) is only a temporary poison for iron-based catalysts [23], but it is unlikely that sulfur will be present in syngas as SO2 (Chapter 3). Equilibrium data for the formation of metal sulfides over the most prominent Fischer–Tropsch metals are presented in Table 4.4 [24]. In practice, Fischer–Tropsch catalysts have a lower tolerance for sulfur than is indicated by the equilibrium data [3, 23]. Gas cleaning before Fischer–Tropsch synthesis (Sections 3.3.1 and 3.5) therefore aims to remove sulfur to the lowest possible level in the synthesis gas. Sulfur deactivation of an Fe-HTFT catalyst at 320 ◦ C is noticeable at a sulfur content of only 0.4 mg · m−3 and it becomes very low only at levels of 0.1 mg · m−3 or less [3]. The situation is worse at lower temperatures. It was reported that the rate of activity decline of an industrial fixed bed Fe-LTFT catalyst decreased discernibly when the sulfur content of the synthesis gas was decreased from 0.14 to 0.01 µg S · g−1 [47]. There is consequently not a minimum tolerance level for sulfur, but rather a continuous benefit in decreasing the syngas sulfur content. Deactivation is mainly due to a loss of activity, but increased methane selectivity has also been reported [48]. Nevertheless, examples have been cited where beneficial effects of limited sulfur poisoning were observed [23, 24, 49, 50]. Of specific importance from a refining perspective is the reported increase in alkene selectivity [49, 50]. This is a consequence of the reduced hydrogenation activity of the catalyst. Since sulfur chemisorption is cumulative, it is not clear whether these beneficial effects can be realized during prolonged industrial operation. Current industrial experience indicates that such sulfur poisoning is unlikely to have a long-term benefit. Chloride and bromide ions (but not fluoride ions) are poisons for iron-based catalysts, and it is likely that other highly electronegative elements will act as poisons for Fischer–Tropsch Table 4.4 Sulfidation–reduction equilibria of different Fischer–Tropsch active metals in temperature range 400–700 K.
Equilibrium, pK = log10 [H2 S/H2 ]
Reaction
FeS + H2 Fe + H2 S 0.11 Co9 S8 + H2 0.11 Co + H2 S NiS + H2 Ni + H2 S 0.5 RuS2 + H2 0.5Ru + H2 S
400 K
500 K
600 K
700 K
–8.3 –8.6 –6.9 –7.6
–6.4 –6.4 –5.1 –5.5
–5.2 –5.0 –3.9 –4.1
–4.3 –4.0 – –3.1
89
90
4 Fischer–Tropsch Synthesis
catalysts [3]. It has been reported that chloride ions present during the preparation of iron-based Fischer–Tropsch catalysts also have a detrimental effect on catalyst activity [5]. The impact of nitrogen-containing compounds that could be present in synthesis gas seems to depend on the Fischer–Tropsch active metal. Ammonia (NH3 ) leads to a decrease in conversion over Co-LTFT catalysts but not over iron-based Fischer–Tropsch catalysts [51]. Hydrogen cyanide (HCN) is also expected to decrease conversion over Co-LTFT catalysts, because its strength of adsorption is comparable to that of CO [51], but the extent of inhibition depends on the rate of HCN hydrogenation. It has been found that HCN had no inhibitory effect on syngas-to-methanol conversion over a Cu/ZnO/Al2 O3 catalyst, because it was rapidly hydrogenated to methyl amine (CH3 NH2 ) [52]. If the same happens over a Co-LTFT catalyst, HCN may, in fact, not be inhibitory.
4.5.2 Volatile Metal Carbonyl Formation
The operating envelope of Fischer–Tropsch synthesis is limited by the potential loss of active metal from the catalyst due to the formation of volatile metal carbonyls. Metal carbonyl formation is promoted by increasing CO pressure and decreasing temperature (Table 4.5) [5]. The possibility that metal will be lost due to carbonyl formation is higher during LTFT synthesis than during HTFT synthesis. Under typical operating conditions for Fe-LTFT and Co-LTFT synthesis (200–240 ◦ C and 2–3 MPa), the equilibrium carbonyl concentration is very low and unlikely to be a significant contributor to Fischer–Tropsch catalyst deactivation. In this respect, the risk of metal loss is higher with Ni- and Ru-based Fischer–Tropsch catalysts. Nickel-based catalysts have a higher equilibrium concentration of metal carbonyl species at low temperature, whereas the high pressures associated with ruthenium-based synthesis promote carbonyl formation [2]. Metal carbonyls may also be formed upstream of Fischer–Tropsch synthesis when the syngas is in contact with process vessels or piping at lower temperatures. It has been reported that iron carbonyls thus formed caused the deactivation of a Co-LTFT catalyst, but it is not a problem when employing an iron-based Fischer–Tropsch catalyst [51].
Table 4.5
Property data of some metal carbonyl compounds formed from Fischer–Tropsch active metals.
Metal carbonyl
Fe(CO)5 [Co(CO)4 ]2 Ni(CO)4 Ru(CO)5
Melting point (◦ C)
–21 50 –25 –22
Decomposition temperature (◦ C)
150 53 60 –15
Equilibrium, pK = log10 [M(CO)n /CO] 100 ◦ C
200 ◦ C
–3.5 – –1.4 –
–9.3 – – –
4.5 Fischer–Tropsch Catalyst Deactivation
4.5.3 Metal Carboxylate Formation
Metal carboxylates are produced during Fischer–Tropsch synthesis through the action of syngas (Figure 4.5) [53], chain initiation by CO2 [4], and readsorption of carboxylic acids on the catalyst. Once formed, the surface carboxylate species are quite stable even in the presence of hydrogen. The stability of the C1 −C3 carboxylates of the Fischer–Tropsch active metals Fe, Co, and Ni were investigated by infrared spectroscopy by heating the samples under a hydrogen atmosphere to 200 ◦ C [53]. It was found that at 200 ◦ C the iron ethanoate intensity was about one-half of the original intensity and that the iron propanoate intensity was only moderately reduced. Cobalt acetate was stable and little diminished at 200 ◦ C, but cobalt propanoate mostly decomposed. Nickel propanoate showed no change at 200 ◦ C. An indication of the stability of carboxylates during Fischer–Tropsch synthesis can be gained by looking at their thermal decomposition behavior (Table 4.6) [54–60], although this disregards the effect of hydrogenation. Despite the scatter in the reported decomposition data, the short-chain carboxylates of Fe, Co, and Ni all decompose in the temperature range 230–300 ◦ C under an inert atmosphere. The temperature of maximum ‘‘corrosiveness’’ of the carboxylic acids depends on the metal. For Fe, maximum corrosiveness is reached in the range 260–300 ◦ C, with iron carboxylate decomposition becoming dominant at temperatures above 300 ◦ C [61]. For Ni, maximum corrosiveness is reached at around 200 ◦ C [59]. Less has been reported about the carboxylate leaching of Co-based catalysts.
R C
O
+H2
Figure 4.5
R
R
R
R
CH2
CH2 O
CH2
O
+CO
d−O
C d+
O
C
CH2 O
O
O
Formation of a surface carboxylate from CO and H2 during Fischer–Tropsch synthesis.
Onset of thermal decomposition (Td ) of Fe, Co, and Ni carboxylates under inert atmosphere.
Table 4.6
Compound
Ethanoate ◦
Fe(III) carboxylate Co(II) carboxylate Ni(II) carboxylate
a Onset
C
Propanoate ◦
Td ( C)
Reference
Td ( C)
Reference
278 280 255a 260a 280 284
[41] [43] [44] [44] [46] [47]
230 230 300 285 280 295
[42] [43] [45] [45] [46] [47]
of thermal decomposition in air.
91
92
4 Fischer–Tropsch Synthesis
Hydrogenation studies have indicated that the environment of the metal also plays an important role in catalyst stability against acid leaching. Hydrogenation of ethanoic acid to ethanal over different supported Fe catalysts resulted in different deactivation behavior, with Fe/C being deactivated almost completely within 6 h [62]. Fe powder and Fe/SiO2 catalysts were more stable. Hydrogenation selectivity to ethanal was high at <280 ◦ C, but the selectivity profile changed dramatically as the decomposition temperature for iron ethanoate was approached. Carboxylic acid leaching is not only limited to the Fischer–Tropsch active metal. Alkali and other promoters may also be leached. During Fe-HTFT synthesis, the two most common metals present in the syncrude are Fe and Na [63]. In the case of supported Fischer–Tropsch catalysts, the support material may also be subjected to carboxylic acid leaching. High levels of aluminium from the leaching and/or attrition of an industrial Co/Pt/Al2 O3 Fischer–Tropsch catalyst has been found, which led to a spate of patent applications dealing with aluminium removal from LTFT wax [64–67]. Metal carboxylate formation is not a major contributor to Fischer–Tropsch catalyst deactivation, but it is a significant source of metal-related problems in downstream operations and in the refinery [28, 68].
4.5.4 Mechanical Catalyst Degradation
Fischer–Tropsch synthesis in slurry bed and fluidized bed reactors expose the Fischer–Tropsch catalyst to mechanical abrasion and attrition by collision of the catalyst particles with each other and with the equipment. Fine particulate matter that is produced by such mechanical catalyst degradation can be difficult to remove. When such fine particulate matter derived from a Fischer–Tropsch catalyst makes its way to the refinery, it can cause serious operating problems, especially if it is present as a colloidal suspension [45]. The need for mechanical strength of Fischer–Tropsch catalysts to be used in fixed bed reactors (crushing strength) and fluidized bed reactors (attrition resistance) has historically been appreciated, but surprisingly not the need for attrition resistance for slurry bed operation. When Anderson tabulated the Fischer–Tropsch catalyst requirements for different reactor types, mechanical strength was listed as ‘‘unimportant’’ for slurry bed operation [23]. In later studies, the importance of attrition resistance during Fischer–Tropsch slurry bed operation was realized and a number of studies evaluated catalyst attrition performance for slurry bed operation [25, 69–72]. It has been reported that Al2 O3 is the most attrition-resistant support for Co-LTFT catalysts, but that under typical Fischer–Tropsch operating conditions alumina exposed to the reaction environment is hydrated to a more mechanically fragile state, compromising catalyst integrity [25]. The production of <1 µm fines may have several side effects, including the failure of solid–liquid separation associated with Fischer–Tropsch synthesis, foam formation inside the Fischer–Tropsch slurry bed reactor, and problems in the downstream refinery [25]. This type of attrition tripped up the first commercial implementation of slurry bed Co-LTFT synthesis, limiting production to only 20% of design capacity in the first year after commissioning [73]. Attrition has not been reported to be a problem for industrial slurry bed operation with Fe-LTFT.
4.5 Fischer–Tropsch Catalyst Deactivation
4.5.5 Deactivation of Fe-HTFT Catalysts
On account of the high operating temperature of Fe-HTFT processes, carbonaceous deposits are readily be formed on the Fischer–Tropsch catalyst during operation. Such carbonaceous deposits can cause deactivation by fouling. Indeed, after weeks on stream, there is on an atomic basis more carbon than iron in an operational fused Fe-HTFT catalyst [19]. These carbonaceous deposits are due to heavy aromatics produced during Fischer–Tropsch synthesis, as well as due to elemental free carbon that is produced by the Boudouard reaction (Equation 4.17) or the equivalent reaction by reduction with H2 (Equation 4.18). 2CO → C (solid) + CO2
(4.17)
CO + H2 → C (solid) + H2 O
(4.18)
Yet, the activity decline is relatively small compared to the amount of carbon deposited on the catalyst, and fouling is neither the dominant deactivation mechanism in Fe-HTFT synthesis nor is it permanent. Activity can be restored by hydrogenation of fouled Fe-HTFT catalysts at temperatures above 350 ◦ C. However, carbon deposition is not innocuous. The most active Fischer–Tropsch sites, which are associated with K2 O promoters, are fouled first. This diminishes the effective alkali promotion of the catalyst and causes a change in activity and selectivity [3]. Carbon deposition is also a significant indirect contributor to permanent deactivation. An analysis of aged Fe-HTFT catalysts showed that small crystals of iron carbide are found embedded in carbon from the Boudouard reaction [74]. Alkali promoter, which is important for selectivity control, is likewise found. Since the carbon-rich material is mechanically weaker and less dense than the rest of the fused iron catalyst, it is more readily lost by attrition in the fluidized bed reactor. Inefficient cyclone separation of the less dense carbon- and alkali-rich material from the product gas causes permanent loss of this material from the reactor [19]. The loss of alkali metal promoter is one of the main reasons for the selectivity change toward lighter and less olefinic products with time on stream (Table 4.7) [75]. This change in syncrude composition is detrimental to the refinery. The gradual increase in the degree of oxidation is not a significant deactivation mechanism of Fe-HTFT catalysts. Oxidation is restricted mainly to the core of larger catalyst particles during normal operation. However, under abnormal conditions where the combined H2 O and CO2 partial pressure is high, more extensive oxidation may take place that can lead to an increased rate of catalyst deactivation. Dry [3] stated that there is no reason to associate catalyst aging with the carbide type or content. A working Fe-HTFT catalyst, even though it has been fully reduced initially, consists mainly of magnetite (Fe3 O4 ) and different iron carbide phases. Although the iron carbides are more resistant to oxidation than reduced iron, further oxidation inevitably takes place during synthesis. Yet, as indicated before, this is not a significant deactivation mechanism. 4.5.6 Deactivation of Fe-LTFT Catalysts
The deactivation of Fe-LTFT catalysts has recently been reviewed by De Smit and Weckhuysen [50]. Good overviews of Fe-LTFT deactivation can also be found in the works by Dry [3, 19].
93
4 Fischer–Tropsch Synthesis Table 4.7 Selectivity change during industrial Fe-HTFT synthesis with increasing time on stream.
Selectivity (mass%)a
Compound or fraction
Methane Ethene Ethane Propene Propane Butenes Butanes C5 and heavier condensate Light oil Decanted oil Aqueous product a Product
Start of run
Average
End of run
7 4 3 10 1 7 1 6 40 14 7
10 4 6 12 2 8 1 8 35 7 7
13 3 9 13 3 9 2 9 30 2 7
selectivity excluding WGS compounds.
100
100
80
80
60
60
40
40
20
20
0
0 0
100
200 300 400 Time on stream (h)
500
Figure 4.6 Deactivation of a Fe-LTFT catalyst by oxidation with time on stream. Syngas conversion () decreases as the percentage of Fe that is present as Fe2 O3 increases. The Fe2 O3 content has been calculated from Mo¨ ssbauer measurements at 20 K () and at 300 K (♦).
600
Fe as Fe2O3 (%)
Unlike Fe-HTFT catalyst deactivation, Fe-LTFT catalyst deactivation is primarily caused by oxidation. The decrease in syngas conversion with increasing oxidation is well illustrated by the study of Davis and coworkers (Figure 4.6) [76]. They presented M¨ossbauer analyses of Fe-LTFT catalysts at different times on stream, as well as the corresponding syngas conversion data. This clearly showed that syngas conversion decreased with increasing oxidation.
Syngas conversion (%)
94
4.5 Fischer–Tropsch Catalyst Deactivation
The effect of oxidation is especially apparent in fixed bed reactors where the gas becomes progressively more oxidizing (higher H2 O partial pressure) as conversion increases over the catalyst bed. The increase in catalyst oxidation is also accompanied by a loss in active BET surface area and an increase in the average crystallite size due to hydrothermal sintering [19]. The BET area decreases over time from 200 m2 · g−1 for freshly reduced and carbided precipitated (SiO2 supported) Fe-LTFT catalysts to around 50 m2 · g−1 for spent Fe-LTFT catalysts [3]. The deactivation of Fe-LTFT catalysts in slurry bed reactors occurs by the same processes. The main difference is that in fixed bed reactors it is mainly the bottom part of the bed that is affected, whereas the complete bed is affected in slurry reactors. Over time, as the Fe-LTFT catalyst loses its activity, the decrease in activity is accompanied by a change in selectivity toward lighter products. Contrary to expectation, a decrease in wax production cannot be attributed to a loss in the K2 O alkali promoter. Although wax selectivity decreases with time on stream, it increases over the bed length during fixed bed synthesis. The highest wax selectivity was reportedly found at the reactor outlet, which implies that the wax selectivity is not correlated to activity. Maximum activity was found in the middle of the bed [19]. As an Fe-LTFT catalyst deactivates, the shift toward to lighter products (lower α-value) is not necessarily accompanied by a decrease in olefinicity. Reports can be found that range from a slight decrease in olefinicity to an increase in alkene and oxygenate selectivity with progressive deactivation of Fe-LTFT catalysts [5, 77, 78]. The transient kinetic studies performed by Schulz and coworkers indicated that initially Fe-LTFT becomes less hydrogenating with time on stream, leading to an increase in product olefin content [79]. Whether this trend is perpetuated during deactivation is not clear. A deactivation process that decreases chain growth may do so by either improving hydrogenation rate (decrease in alkene selectivity) and/or by increasing the desorption rate (increase in alkene selectivity). Two studies that are of particular interest are those conducted in the Ruhrchemie pilot plants (Table 4.8) [5] and more recently the studies by Janse van Vuuren and coworkers [29, 30]. In both cases, an Fe-LTFT catalyst was operated for an extended period. As expected, the products became lighter and the methane selectivity increased with time on stream. However, after some time the methane selectivity stabilized and remained constant over time, whereas the light gas (C2 −C4 ) selectivity kept on increasing. The alkene and oxygenate selectivity also kept on increasing with time on stream. LTFT catalyst deactivation that results in an increased desorption rate, but without an increase in hydrogenation rate, is actually advantageous for product refining [28]. In such a case, the syncrude from the ‘‘equilibrated’’ Fe-LTFT catalyst is preferable to the syncrude from a fresh Fe-LTFT catalyst. Furthermore, the ‘‘equilibrated’’ Fe-LTFT has a very low rate of deactivation. This challenges the notion that Fe-LTFT catalysts have a shorter lifetime than Co-LTFT catalysts. 4.5.7 Deactivation of Co-LTFT Catalysts
The deactivation of Co-LTFT catalysts has been ascribed to various mechanisms. The review by Holmen and coworkers provides an excellent summary of Co-LTFT deactivation [80]. The most prominent deactivation mechanisms are carbon deposition and fouling, oxidation and mixed oxide formation, and changes in the Co crystallite size due to processes such as sintering
95
96
4 Fischer–Tropsch Synthesis Table 4.8 Ruhrchemie pilot-plant data for the German medium-pressure Fischer–Tropsch process operated at 215–220 ◦ C and 1 MPa with a K2 O/Fe–Cu–CaO–kieselguhr catalyst. Feed conversion was maintained at 64–66% of H2 + CO. The recycle ratio was 3, resulting in a H2 :CO ratio in combined feed of 1.6 : 1. The usage ratio was close to 1 : 1.
Description
Production with increasing time on stream (h) 0–724
Fresh feed space velocity (h−1 ) Selectivity (mol% CO converted) CH4 CO2 Yield (g/m3 of H2 + CO) C3 and C4 C5 and heavier Selectivity (mass%) C5 -200 ◦ C 200–320 ◦ C 320–460 ◦ C >460 ◦ C Alkenes and oxygenates (vol%) C5 -200 ◦ C 200–320 ◦ C
95
724–1471 100
1471–2249
2249–2923
2923–3550
87
84
80
4 33.9
4 31.4
4.6 33.6
4.3 31
4.5 32.3
10.9 106
12.6 92
12 90
13 88
15 89
31.7 15.2 16.5 36.6
29.6 16.2 17.8 36.4
30.5 15.1 18.3 37.2
29.7 17.8 16.5 35.6
31.8 15.3 19.9 33
68 43
69 49
71 53
72 54
72 55
and agglomeration (coalescence). Other deactivation mechanisms that may be active include carbidization, surface reconstruction, leaching of metals, and catalyst attrition. There are two deactivation periods. The first period is associated with rapid short-term deactivation early in the reaction cycle until the catalyst has stabilized, and this period is in the order of a 100 h (a few days to weeks). The second period is much longer and is associated with the usual time-on-stream deactivation found during Co-LTFT synthesis, which determines the overall cycle length before catalyst rejuvenation is necessary or the catalyst has to be replaced. Carbon deposition and fouling can cause Co-LTFT catalyst deactivation in various ways by forming a physical barrier or by chemically changing the nature of the catalyst. Of the different types of carbon that can be formed, deactivation has been ascribed to the formation of a polymeric inactive form of carbon that prevents access to the catalytic sites [81, 82]. It is not a permanent deactivation mechanism, and rejuvenation or regeneration of the catalyst to remove the carbonaceous deposits can restore access to the catalytic sites. Carbon deposition affects not only Co-LTFT catalyst activity but also its selectivity. The impact of carbon on selectivity and specifically on increased methane selectivity is illustrated by the accelerated deactivation data presented in Figure 4.7 [81]. A similar increase in methane selectivity was observed after upset conditions with a Co/Al2 O3 Fischer–Tropsch catalyst [83].
4.5 Fischer–Tropsch Catalyst Deactivation
Methane selectivity (mass%)
16 14 12 10 8 6 4
Hydrogen flow stopped for 2 h to accelerate carbon deposition
2 0 0
20
40
60
80
100
120
140
Time on stream (h) Figure 4.7 Increase in methane selectivity caused by carbon deposition during accelerated aging of a Co/Pt/Al2 O3 catalyst.
Oxidation as deactivation mechanism of Co-LTFT catalysts is a more contentious topic. The view has been expressed by some that oxidation can been ruled out as a deactivation mechanism under realistic Fischer–Tropsch synthesis conditions with a Co/Pt/Al2 O3 catalyst [84]. Yet, for the same catalyst it has been reported that long-term exposure to higher water partial pressure may have increased the rate of longer term irreversible deactivation, although it did not cause a step change in activity [85]. Soled and coworkers expressed a different view [86]. It was found that Co oxidation was responsible for a gradual loss of Co-LTFT catalyst activity. The rate of activity loss due to oxidation was partly determined by the severity of catalyst reduction before synthesis. An AGC-21 Co-LTFT catalyst reduced at 225 ◦ C deactivated by 10% over a period of 200 days, but if the same catalyst was reduced at 275 ◦ C, it deactivated by only 6% over the same period. Although deactivation by Co oxidation was reversible, its impact on catalyst stability was more insidious. Cobalt oxidation eventually leads to some irreversible deactivation by causing both mixed metal oxide formation and agglomeration. Smaller Co crystallites are the first to be completely oxidized and the rate of reversible deactivation by Co oxidation is therefore dependent on the crystallite size distribution. Unlike the reduced Co, the cobalt oxides and oxyhydroxides can wet the support. In this way, oxidized Co can form mixed oxides Cox MOy with the support, such as cobalt silicates and cobalt aluminates [86, 87]. The formation of these mixed metal oxides is not possible when Co is in its reduced state. Although the oxidation can be reversed by reductive rejuvenation (rehydrogenation by H2 of the catalyst, typically at 200–225 ◦ C), once formed, the mixed metal oxides are not reduced during rejuvenation and constitute a permanent loss of Co [86]. The process of oxidation and rejuvenation is illustrated in Figure 4.8 [86]. Oxidation of the Co is diffusion controlled. As the outer layer is oxidized, the metal migrates to the outside of the oxide layer. As this process is repeated, the crystallite grows in size, forming a hollow dome of cobalt oxide on the inside. These structures were observed by TEM. Ultimately, when most of the Co is oxidized, there is a significant decrease in catalyst activity. When the catalyst is rejuvenated by rehydrogenation, the CoO dome is reduced to form a cluster of smaller Co crystallites. Although
97
98
4 Fischer–Tropsch Synthesis
Co
CoO Co
CoO
Co
Side view (cross section) Top view
Co CoO (a)
(b)
(c)
(d)
Figure 4.8 The process of Co-LTFT deactivation by oxidation. The fresh catalyst (a) is oxidized to form a layer of CoO on the surface (b). The Co migrates through the oxide layer to the outer surface (c), but eventually the complete crystallite is oxidized to a
(e) hollow dome-shaped CoO particle (d). When this particle is rehydrogenated during reductive catalyst rejuvenation, a number of smaller Co crystallites are formed (e).
this temporarily restores catalyst activity, these smaller Co crystallites can coalesce, not only with themselves but also with adjacent Co crystallites. The rejuvenation increases dispersion, but it results in smaller particles that are too closely spaced and eventually leads to increased agglomeration, which is irreversible. The smaller Co crystallites are also more readily oxidized and may have sizes below 6 nm, which is the threshold for good Fischer–Tropsch activity [88]. In fixed bed Co-LTFT experiments, it was found that the agglomeration of Co increased toward the end of the bed where the water partial pressure and oxidation rate are the highest [86]. The Co crystallite size can also change during synthesis independently of oxidation– rejuvenation cycles. In general, when Co crystallites are closely spaced, growth of the Co crystallites takes place during synthesis. This leads to irreversible loss of activity. In catalysts that have initially smaller crystallites, the crystallites grow faster to larger clusters than in catalysts with larger but better spaced crystallites [86]. The agglomeration process takes place by coalescence and not by Ostwald ripening. No evidence was found of crystallites reducing in size, while other crystallites increased in size, as one would expect from Ostwald ripening. This type of behavior has also been observed during Fischer–Tropsch synthesis over a Co/Pt/Al2 O3 catalyst at 230 ◦ C and 2 MPa in a 100 bbl/day slurry bed demonstration unit [89]. It was found that the smaller crystallites agglomerated over time to cause a shift in crystallite size distribution to large crystallites, with a concomitant decrease in metal dispersion and surface area. This data also indicated that the sintering or agglomeration did not take place by Ostwald ripening. With increasing Co-LTFT catalyst deactivation, there is a gradual change in product selectivity to lighter products and specifically methane (Table 4.9) [5]. This change in syncrude composition with time on stream is detrimental to the refinery [28]. The increase in methane selectivity with time on stream can be the result of many factors, including an increase in temperature to maintain constant conversion [5], sulfur poisoning [48], and fouling by carbonaceous deposits [81, 83]. This increase is reportedly not related to permanent catalyst deactivation [90]. After repeated reaction–regeneration cycles with a Co/Al2 O3 catalyst, the relative methane selectivity was not influenced despite some irrecoverable activity loss for Fischer–Tropsch synthesis. It was reported that there is a Co-LTFT-catalyst-independent linear relationship between increased methane selectivity and decreased C5 and heavier product selectivity in the methane
References Ruhrchemie pilot-plant data for the German medium-pressure Fischer–Tropsch process operated at 204–223 ◦ C and 0.7 MPa with a Co–ThO2 –MgO–kieselguhr catalyst. Feed conversion was maintained at 62–64% of H2 + CO. The recycle ratio was 3, resulting in a H2 :CO ratio in combined feed of 0.8 : 1.
Table 4.9
Description
Fresh feed space velocity (h−1 ) Operating temperature (◦ C) Selectivity (mol% CO converted) CH4 CO2 Yield (g/m3 of H2 + CO) C3 and C4 C5 and heavier Selectivity (mass%) C5 -200 ◦ C 200–320 ◦ C 320–460 ◦ C >460 ◦ C Alkenes and oxygenates (vol%) C5 -200 ◦ C 200–320 ◦ C
Production with increasing time on stream (months) 1
2
3
4
5
6
134 204
124 207
121 211
120 214
113 218
101 223
5 0.2 14 101
7.1 1.8
7.9 1.4
9.4 2.7
12.6 2.9
15.1 3.9
14 97
14 99
15 92
17 84
23 68
48.6 20.4 21.4 9.6
52.2 25.6 17 5.2
52.6 26.4 16.1 4.9
56.1 24.4 16 3.5
57.6 26 13 3.4
67.6 24.3 6.4 1.7
61.5 44.5
62 42.7
59.5 42
58.5 40.3
55.6 40
53.8 34.2
selectivity range 6–9% (Equation 4.19) [91]. (C5 + selectivity) = −2.68(CH4 selectivity) + 105.4
(4.19)
The implication of this relationship is that there is a significant increase in C2 −C4 selectivity with an increase in methane selectivity.
References 1. Steynberg, A.P. (2004) Introduction to
4. Davis, B.H. (2001) Fischer–Tropsch
Fischer–Tropsch technology. Stud. Surf. Sci. Catal., 152, 1–63. 2. Anderson, R.B. (1956) in Catalysis, Hydrocarbon Synthesis, Hydrogenation and Cyclization, Vol. IV (ed. P.H. Emmett), Reinhold, New York, pp. 257–371. 3. Dry, M.E. (1981) in Catalysis Science and Technology, Vol. 1 (eds J.R. Anderson and M. Boudart), Springer-Verlag, Berlin, pp. 159–255.
synthesis: current mechanism and futuristic needs. Fuel Process. Technol., 71, 157–166. 5. Storch, H.H., Golumbic, N., and Anderson, R.B. (1951) The Fischer–Tropsch and Related Syntheses, John Wiley & Sons, Inc., New York. 6. Fahey, D.R. (1981) Rational mechanism for homogeneous hydrogenation of carbon monoxide to alcohols, polyols, and esters. J. Am. Chem. Soc., 103, 136–141.
99
100
4 Fischer–Tropsch Synthesis 7. Maitlis, P.M. and Zanotti, V. (2009) The role
8.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19. 20.
of electrophilic species in the Fischer–Tropsch reaction. Chem. Commun., 1619–1634. Weitkamp, A.W. and Frye, C.G. (1953) (Products of the hydrogenation of carbon monoxide) relation of product composition to reaction mechanism. Ind. Eng. Chem., 45, 363–367. Dry, M.E. (2004) Chemical concepts used for engineering purposes. Stud. Surf. Sci. Catal., 152, 196–257. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Janse van Vuuren, M.J., Govender, G.N.S., Kotze, R., Masters, G.J., and Pete, T.P. (2005) The correlation between double bond isomerization, water gas shift and acid production during Fischer–Tropsch synthesis. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 50 (2), 200–202. Tau, L.-M., Dabbagh, H.A., and Davis, B.H. (1990) Fischer–Tropsch synthesis: 14 C tracer study of alkene incorporation. Energy Fuels, 4, 94–99. Gaube, J. and Klein, H.-F. (2009) in Advances in Fischer–Tropsch Synthesis, Catalysts and Catalysis (ed B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 199–214. Van der Laan, G.P. and Beenackers, A.A.C.M. (1999) Kinetics and selectivity of the Fischer–Tropsch synthesis: a literature review. Catal. Rev.-Sci. Eng., 41, 255–318. Anderson, R.B. (1978) Schulz-Flory equation. J. Catal., 55, 114–115. Puskas, I. and Hurlbut, R.S. (2003) Comments about the causes of deviations from the Anderson-Schulz-Flory distribution of the Fischer–Tropsch reaction products. Catal. Today, 84, 99–109. Botes, F.G. (2007) Proposal of a new product characterization model for the iron-based low-temperature Fischer–Tropsch synthesis. Energy Fuels, 21, 1379–1389. Matsumoto, H., Saito, Y., and Yoneda, Y. (1971) The classification of metal catalysts in hydrogenolysis of hexane isomers. J. Catal., 22, 182–192. Dry, M.E. (2004) FT catalysts. Stud. Surf. Sci. Catal., 152, 533–600. Botes, F.G. and Breman, B.B. (2006) Development and testing of a new macro kinetic expression for the iron-based low-temperature Fischer–Tropsch reaction. Ind. Eng. Chem. Res., 45, 7415–7426.
21. Botes, F.G., Van Dyk, B., and McGregor,
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
C. (2009) The development of a macro kinetic model for a commercial Co/Pt/All2 O3 Fischer–Tropsch catalyst. Ind. Eng. Chem. Res., 48, 10439–10447. Trimm, D.L., Liu, I.O.Y., and Cant, N.W. (2010) The selective hydrogenation of acetylene over a Ni/SiO2 catalyst in the presence and absence of carbon monoxide. Appl. Catal. A, 374, 58–64. Anderson, R.B. (1956) in Catalysis, Hydrocarbon Synthesis, Hydrogenation and Cyclization, Vol. IV (ed. P.H. Emmett), Reinhold, New York, pp. 29–255. Anderson, R.B., K¨olbel, H., and Ralek, M. (1984) The Fischer–Tropsch Synthesis, Academic Press, Orlando, FL. Perego, C., Bortolo, R., and Zennaro, R. (2009) Gas to liquids technologies for natural gas reserves valorization: the Eni experience. Catal. Today, 142, 9–16. De Klerk, A. (2008) Fischer–Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. Iglesia, E. (1997) Design, synthesis, and use of cobalt-based Fischer–Tropsch synthesis catalysts. Appl. Catal. A, 161, 59–78. De Klerk, A. (2010) Deactivation in Fischer–Tropsch synthesis and its impact on refinery design. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 55 (1), 86–89. Janse van Vuuren, M.J., Huyser, J., Kupi, G., and Grobler, T. (2008) Understanding Fe-LTFT selectivity changes with catalyst age. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 53 (2), 129–130. Janse van Vuuren, M.J., Huyser, J., Grobler, T., and Kupi, G. (2009) in Advances in Fischer–Tropsch Synthesis, Catalysts and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 229–241. Fraenkel, D. and Gates, B.C. (1980) Shape-selective Fischer–Tropsch synthesis catalyzed by zeolite-entrapped cobalt clusters. J. Am. Chem. Soc., 102, 2478–2480. Egiebor, N.O., Cooper, W.C., and Wojciechowski, B.W. (1989) Synthesis of motor fuels from HY-zeolite supported Fischer–Tropsch iron catalysts. Appl. Catal., 55, 47–64. Idem, O.R., Katikaneni, S.P.R., Sethuraman, R., and Bakhshi, N.N. (2000) Production of C4 hydrocarbons from modified Fischer–Tropsch synthesis over Co-Ni-ZrO2 /sulfated-ZrO2 hybrid catalysts. Energy Fuels, 14, 1072–1082.
References 34. Song, X. and Sayari, A. (1996) Sulfated zirconia
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
as a cocatalyst in Fischer–Tropsch synthesis. Energy Fuels, 10, 561–565. Botes, F.G., and B¨ohringer, W. (2004) The addition of HZSM-5 to the Fischer–Tropsch process for improved gasoline production. Appl. Catal. A, 267, 217–225. Liu, Z.-W., Li, X., Asami, K., and Fujimoto, K. (2006) High performance Pd/beta catalyst for the production of gasoline-range iso-paraffins via a modified Fischer–Tropsch reaction. Appl. Catal. A, 300, 162–169. Subiranas, A.M. and Schaub, G. (2009) Combining Fischer–Tropsch (FT) and hydrocarbon reactions under FT reaction conditions: model compound and combined-catalyst studies. Int. J. Chem. React. Eng., 7, A31. Varma, R.L., Bakhshi, N.N., Mathews, J.F., and Ng, S.N. (1987) Performance of dual-reactor system for conversion of syngas to aromatic-containing hydrocarbons. Ind. Eng. Chem. Res., 26, 183–188. Steynberg, A.P., Dry, M.E., Davis, B.H., and Breman, B.B. (2004) Fischer–Tropsch reactors. Stud. Surf. Sci. Catal., 152, 64–195. Iglesia, E., Reyes, S.C., and Madon, R.J. (1991) Transport-enhanced α-olefin readsorption pathways in Ru-catalyzed hydrocarbon synthesis. J. Catal., 129, 238–256. Sie, S.T. and Krishna, R. (1999) Fundamentals and selection of advanced Fischer–Tropsch reactors. Appl. Catal. A, 186, 55–70. Satterfield, C.N., Huff, G.A. Jr., Stenger, H.G., Carter, J.L., and Madon, R.J. (1985) A comparison of Fischer–Tropsch synthesis in a fixed bed reactor and a slurry reactor. Ind. Eng. Chem. Fundam., 24, 450–454. Visconti, C.G., Tronconi, E., Lietti, L., Groppi, G., Forzatti, P., Crisriani, C., Zennaro, R., and Rossini, S. (2009) An experimental investigation of Fischer–Tropsch synthesis over washcoated metallic structured supports. Appl. Catal. A, 370, 93–101. Lerou, J.J., Tonkovich, A.L., Silva, L., Perry, S., and McDaniel, J. (2010) Microchannel reactor architecture enables greener processes. Chem. Eng. Sci., 65, 380–385. Wang, S., Chung, K.H., Masliyah, J.H., and Gray, M.R. (1999) Deposition of fine particles in packed beds at hydrotreating conditions: role of surface chemistry. Ind. Eng. Chem. Res., 38, 4878–4888. Bartholomew, C.H. (2001) Mechanism of catalyst deactivation. Appl. Catal. A, 212, 17–60.
47. Dry, M.E. (1983) in Applied Industrial Catalysis,
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
Vol. 2 (B.E. Leach), Academic Press, New York, pp. 167–213. Visconti, C.G., Lietti, L., Forzatti, P., and Zennaro, R. (2007) Fischer–Tropsch synthesis on sulphur poisoned Co/Al2 O3 catalyst. Appl. Catal. A, 330, 49–56. Kritzinger, J.A. (2002) The role of sulfur in commercial iron-based Fischer–Tropsch catalysis with focus on C2 -product selectivity and yield. Catal. Today, 71, 307–318. De Smit, E. and Weckhuysen, B.M. (2008) The renaissance of iron-based Fischer–Tropsch synthesis: on the multifaceted catalyst deactivation behaviour. Chem. Soc. Rev., 37, 2758–2781. Van Steen, E. and Claeys, M. (2008) Fischer– Tropsch catalysts for the biomass-to-liquid process. Chem. Eng. Technol., 31, 655–666. Quinn, R., Dahl, T.A., and Toseland, B.A. (2004) An evaluation of synthesis gas contaminants as methanol synthesis catalyst poisons. Appl. Catal. A, 272, 61–68. Blyholder, G., Shihabi, D., Wyatt, W.V., and Bartlett, R. (1976) Adsorption and interaction of C2 H4 , H2 , CO, and organic acids on Fe, Co, and Ni. J. Catal., 43, 122–130. Jewur, S.S. and Kuriacose, J.C. (1977) Studies on the thermal decomposition of ferric acetate. Thermochim. Acta, 19, 195–200. Bassi, P.S., Randhawa, B.S., and Jamwal, H.S. (1983) M¨ossbauer study of the thermal decomposition of some iron(III) monocarboxylates. Thermochim. Acta, 62, 209–216. Perinet, G. and Le Van, M. (1968) Sur les enthalpies et produits de d´ecomposition thermique des alcanoates de cobalt anhydres en atmosph´ere inerte. C. R. Acad. Sci. Paris C, 266, 732–734. Leicester, J. and Redman, M.J. (1962) Thermal decomposition of the nickel and cobalt salts of aliphatic acids. J. Appl. Chem., 12, 357–366. Bassi, P.S., Jamwal, H.S., and Randhawa, B.S. (1983) Comparative study of the thermal analyses of some transition metal(II) propionates. Part I. Thermochim. Acta, 71, 15–24. Pienaar, A.D. and De Klerk, A. (2008) Nickel catalyst stability toward carboxylic acids. Ind. Eng. Chem. Res., 47, 4962–4965. Le Van, M. (1965) Etude microcalorim´etrique des sels hydrat´es de cuivre et de nickel, III. Formiates, ac´etate, propionate de nickel. Bull. Soc. Chim. Fr., 366–371. Y´epez, O. (2007) On the chemical reaction between carboxylic acids and iron, including
101
102
4 Fischer–Tropsch Synthesis
62.
63.
64.
65. 66.
67.
68.
69.
70.
71.
72.
73.
74. 75.
76.
the special case of naphthenic acid. Fuel, 86, 1162–1168. Rachmady, W. and Vannice, M.A. (2002) Acetic acid reduction to acetaldehyde over iron catalysts: I. Kinetic behaviour. J. Catal., 208, 158–169. Swart, J.S., Czajkowski, G.J., and Conser, R.E. (1981) Sasol upgrades Synfuels with refining technology. Oil Gas J., 79 (35), 62–66. Kuperman, A., Bul, L.M., O’Rear, D.J., and Kuehne, D. (2008) Process for removing aluminum contaminants from Fischer–Tropsch feed streams using dicarboxylic acid. US Patent 7,416,656. Botha, J.M. (2006) Treatment of hydrocarbons. US Patent WO 2006/005084. Botha, J.M., Leckel, D.O., Visagie, J.L., Preston, H., and Smook, D. (2006) Treatment of hydrocarbons. US Patent WO 2006/005085. Vogel, A.P. and Nel, H.G. (2006) Treatment of high molar mass hydrocarbon streams. US Patent WO 2006/053350. De Klerk, A. (2008) Hydroprocessing peculiarities of Fischer–Tropsch syncrude. Catal. Today, 130, 439–445. Zhao, R., Goodwin, J.G. Jr., and Oukaci, R. (1999) Attrition assessment for slurry bubble column reactor catalysts. Appl. Catal. A, 189, 99–116. O’Brien, R.J., Xu, L., Bao, S., Raje, A., and Davis, B.H. (2000) Activity, selectivity and attrition characteristics of supported iron Fischer–Tropsch catalysts. Appl. Catal. A, 196, 173–178. Wei, D., Goodwin, J.G., Oukaci, R., and Singleton, A.H. Jr. (2001) Attrition resistance of cobalt F-T catalysts for slurry bubble column reactor use. Appl. Catal. A, 210, 137–150. Bukur, D.B. (2005) Attrition studies with catalysts and supports for slurry phase Fischer–Tropsch synthesis. Catal. Today, 106, 275–281. Anonymous (2008) GTL: oryx breakthrough and oil-price surge lift industry spirits. Pet. Econ., 75 (6), 36–38. Dry, M.E. (1990) Fischer–Tropsch synthesis over iron catalysts. Catal. Lett., 7, 241–252. Hoogendoorn, J.C. (1973) Experience with Fischer–Tropsch synthesis at Sasol. Clean Fuels from Coal Symposium, Chicago, pp. 353–365. Sarkar, A., Seth, D., Dozier, A.K., Neathery, J.K., Hamdeh, H.H., and Davis, B.H. (2007) Fischer–Tropsch synthesis: morphology, phase
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
transformation and particle size growth of nano-scale particles. Catal. Lett., 117, 1–17. Bukur, D.B., Koranne, M., Lang, X., Rao, K.R.P.M., and Huffman, G.P. (1995) Pretreatment effect studies with a precipitated iron Fischer–Tropsch catalyst. Appl. Catal. A, 126, 85–113. Hao, Q.-L., Liu, F.-X., Wang, H., Chang, J., Zhang, C.-H., Bai, L., Xiang, H.-W., Li, Y.-W., Yi, F., and Xu, B.-F. (2007) Effect of reduction temperature on a spray-dried iron-based catalyst for slurry Fischer–Tropsch synthesis. J. Mol. Catal. A, 261, 104–111. Schulz, H., Schaub, G., Clayes, M., and Riedel, T. (1999) Transient initial kinetic regimes of Fischer–Tropsch synthesis. Appl. Catal. A, 186, 215–227. Tsakoumis, N.E., Rønning, M., Borg, Ø., Rytter, E., and Holmen, A. (2010) Deactivation of cobalt based Fischer–Tropsch catalysts: a review. Catal. Today, 154, 162–182. Moodley, D.J. (2008) On the deactivation of cobalt-based Fischer–Tropsch synthesis catalysts. PhD thesis. Technische Universiteit Eindhoven, Eindhoven. Moodley, D.J., Van de Loosdrecht, J., Saib, A.M., and Niemantsverdriet, J.W. (2009) in Advances in Fischer–Tropsch Synthesis, Catalysts and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 49–81. Gruver, V., Zhan, X., Engman, J., and Robota, H.J. (2004) Deactivation of a Fischer–Tropsch catalyst through the formation of cobalt carbide under laboratory slurry reactor conditions. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (2), 192–194. Van de Loosdrecht, J., Balzhinimaev, B., Dalmon, J.-A., Niemantsverdriet, J.W., Tsybulya, S.V., Saib, A.M., Van Berge, P.J., and Visagie, J.L. (2007) Cobalt Fischer–Tropsch synthesis: deactivation by oxidation? Catal. Today, 123, 293–302. Botes, F.G. (2009) Influences of water and syngas partial pressure on the kinetics of a commercial alumina-supported cobalt Fischer–Tropsch catalyst. Ind. Eng. Chem. Res., 48, 1859–1865. Soled, S.L., Kiss, G., Baumgartner, J., and Kliewer, C. (2010) Reversible and irreversible changes in Co Fischer–Tropsch catalysts during synthesis. 21st Canadian Symposium on Catalysis, May 9–12, 2010, Banff, AB, Canada, p. 35.
References 87. Hilmen, A.M., Schanke, D., Hanssen, K.F., and
Holmen, A. (1999) Study of the effect of water on alumina supported cobalt Fischer–Tropsch catalysts. Appl. Catal. A, 186, 169–188. 88. Bezemer, G.L., Bitter, J.H., Kuipers, H.P.C.E., Oosterbeek, H., Holewijn, J.E., Xu, X., Kapteijn, F., Van Dillen, A.J., and De Jong, K.P. (2006) Cobalt particle size effects in the Fischer–Tropsch reaction studied with carbon nanofiber supported catalysts. J. Am. Chem. Soc., 128, 3956–3964. 89. Overett, M.J., Breedt, B., Du Plessis, E., Erasmus, W., and Van de Loosdrecht, J. (2008) Sintering as a deactivation mechanism for an
alumina supported cobalt Fischer–Tropsch synthesis catalyst. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 53 (2), 126–128. 90. Zhan, X., Arcuri, K., Huang, R., Agee, K., Engman, J., and Robota, H.J. (2004) Regeneration of cobalt-based slurry catalysts for Fischer–Tropsch synthesis. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (2), 179–181. ˚ 91. Rytter, E., Skagseth, T.H., Eri, S., and Sjastad, A.O. (2010) Cobalt Fischer–Tropsch catalysts using nickel promoter as a rhenium substitute to suppress deactivation. Ind. Eng. Chem. Res., 49, 4140–4148.
103
105
5 Fischer–Tropsch Gas Loop 5.1 Introduction
The Fischer–Tropsch gas loop refers to the way in which syngas flows around Fischer–Tropsch synthesis. An overview of the principles and practical considerations in gas loop design can be found in the work by Dry and Steynberg [1]. The design of the gas loop has a marked influence on the carbon efficiency of indirect liquefaction by Fischer–Tropsch synthesis, since it controls the flow and destination of syngas carbon. The gas loop also has a tremendous impact on the refinery, and Fischer–Tropsch refinery design actually starts in the gas loop. Gas loop design is an optimization problem. Central to the gas loop is the Fischer–Tropsch synthesis. The input to and output from Fischer–Tropsch synthesis influence the gas loop design and are in turn influenced by the gas loop design. Synthesis and gas loop designs are not independent; optimization of Fischer–Tropsch synthesis and the gas loop is an iterative process. Some of the key aspects that must be considered during gas loop design are as follows: 1) Fischer–Tropsch per pass conversion. The per pass syngas conversion determines how much unconverted syngas remains in the product from Fischer–Tropsch synthesis. It also affects the syngas conditioning requirements. At a low per pass syngas conversion, it is not that critical to prepare a syngas feed that is close to the usage ratio (Section 4.4.3). The importance of providing conditioned syngas that is close to the usage ratio becomes increasingly important with increasing conversion during Fischer–Tropsch synthesis (Figure 5.1). At a high per pass conversion, the syngas composition in the feed in relation to the usage ratio affects the effective H2 :CO ratio, or Ribblett ratio (Equation 4.16) for water-gas-shift-active Fischer–Tropsch catalysts. In turn, it affects catalyst deactivation rate, kinetics, product selectivity, and the syngas compound that is the limiting reagent. 2) Syngas conditioning. The purpose of syngas conditioning is to match the syngas H2 :CO feed ratio to the H2 :CO usage ratio (Section 3.6). The level to which the syngas must be conditioned dictates the configuration and type of syngas conditioning units required. 3) Fate of unconverted syngas. It is only in an ideal world where Fischer–Tropsch synthesis can be operated at complete conversion with a feed at its usage ratio. This implies that the product from Fischer–Tropsch synthesis contains some unconverted H2 and CO. Recycling the unconverted gas can increase the overall syngas conversion, but even by doing so the product still contains some unconverted syngas compounds. The fate of the unconverted Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
5 Fischer–Tropsch Gas Loop
2.4 Fischer−Tropsch H2:CO feed ratio = 1.6
2 H2:CO ratio at outlet
106
1.6 H2:CO usage ratio: 1.4 1.5 1.6 1.7 1.8
1.2
0.8 0
10
20
30
40
50
60
70
80
90
CO conversion (%) Figure 5.1 Impact of per pass CO conversion during Fischer–Tropsch synthesis on the syngas conditioning requirements of the feed, as illustrated for a H2 :CO feed ratio of 1.6.
syngas is determined by the gas loop design. On the one extreme, the gas is purified to recover the unconverted syngas as pure H2 and CO that can be recycled. On the other extreme, the unconverted syngas is used as fuel gas. This design decision clearly affects the complexity of the gas loop, as well as the carbon efficiency of synthesis. It also has a knock-on effect on product recovery. 4) Product recovery. After Fischer–Tropsch synthesis, the product gas can be cooled down under pressure to knock out condensable material. The condensable material can be recovered by gas–liquid separation. Even volatile light hydrocarbons in the C3 –C4 range can be recovered by proper design. For example, at 45 ◦ C both propane and propene can be substantially recovered by condensation at a pressure of less than 2 MPa. However, to recover the C1 –C2 hydrocarbons, cryogenic separation is required. The decision to employ cryogenic separation in the gas loop improves the efficiency of light hydrocarbon separation and it enables the separation of H2 and CO for recycling. This comes at added complexity and cost, but improves the overall carbon efficiency of the process. The recovery of a methane-rich product also enables the export of synthetic natural gas, or may warrant the inclusion of a gas reformer (Section 3.3) in the gas loop to convert the methane into synthesis gas. 5) Syncrude cooling and separation design. The design of the stepwise cooling and product separation after Fischer–Tropsch synthesis determines the composition of the feed material to the Fischer–Tropsch refinery. This is one of the advantages of Fischer–Tropsch refining over conventional crude oil refining, because the gas loop essentially prefractionates the syncrude [2]. The efficiency of product cooling and separation also has an impact on downstream units in the gas loop itself.
5.2 Gas Loop Configurations
The gas loop designs of industrial Fischer–Tropsch facilities will be discussed later (Chapters 6–11). The objective of this chapter is to give an overview of gas loop design in general, and the focus is on how gas loop design decisions affect the Fischer–Tropsch refinery.
5.2 Gas Loop Configurations
In its most rudimentary form, the gas loop configuration refers to the process flow of the syngas, without making reference to any specific units modifying the composition beyond Fischer–Tropsch synthesis and product recovery. The gas that remains after product recovery is called the tail gas. Gas loop configurations are accordingly classified as open or closed loop designs depending on the fate of the tail gas (Figure 5.2). 1) Open loop design. In a pure open loop design, the tail gas is not recycled. This is the simplest possible gas loop design and involves a once-through flow of the synthesis gas. In such a gas loop, all unconverted syngas is ‘‘lost’’ with respect to further Fischer–Tropsch conversion, although the unconverted syngas may still be employed as fuel gas. 2) Closed loop design. In a closed loop design, the tail gas is recycled. In its most extreme form, the product recovery removes all nonsyngas components and the tail gas can be recycled to extinction. In such an ideal gas loop, all syngas is ultimately converted and recovered as the product. 5.2.1 Open Gas Loop Design
The main advantage of an open gas loop design is its simplicity of construction and operation. No buildup of inert compounds is possible and there are no recycle loops to complicate the design. Since the tail gas is ‘‘lost’’ for further synthesis, there must be a good reason not to recycle the tail gas. An open loop configuration may be considered in the following instances: 1) 2) 3)
The conversion of syngas during Fischer–Tropsch synthesis is very high. Syngas is produced by air blown gasification or adiabatic oxidative reforming. Fischer–Tropsch synthesis is combined with an electricity cogeneration facility.
When the syngas conversion is very high, little syngas remains to be recycled and the cost associated with recycling may not be justified. One may, therefore, consider a configuration with
(a) Open: Syngas
Fischer−Tropsch
Tail gas Product recovery
Syncrude
Tail gas (b) Closed: Syngas Figure 5.2
Fischer−Tropsch
Product recovery
Syncrude
Fischer–Tropsch gas loop configurations: (a) open gas loop and (b) closed gas loop.
107
108
5 Fischer–Tropsch Gas Loop
more than one Fischer–Tropsch reactor in series, optionally with intermediate product recovery, in order to simplify the gas loop design. This may become especially attractive if synthesis generation has been simplified by selecting air as an oxidant, instead of pure O2 from an air separation unit. In such cases, the amount of inert gas (N2 ) present in the syngas makes it impractical to recycle the tail gas. When the tail gas can be beneficially employed as a fuel gas or used for electricity cogeneration, there may be instances where the cost and added complexity of tail gas recycling are not warranted. Philosophically it is difficult to justify an open gas loop with cogeneration, since the ultimate objective of feed-to-liquid conversion is not cogeneration. The unconverted syngas in the tail gas has a high associated CO2 footprint and using it for cogeneration may be economical but it is difficult to justify from a carbon efficiency perspective. Some gas loop designs are called open when a limited recycle is employed and a significant portion of the tail gas is used as fuel gas. According to the definition, such designs are strictly speaking ‘‘closed’’ and are discussed in the following section. 5.2.2 Closed Gas Loop Design
Most Fischer–Tropsch gas loop designs are likely to include some form of recycle. Tail gas recycle streams come in two varieties (Figure 5.3): 1) Internal recycle. Tail gas that is directly recycled to Fischer–Tropsch synthesis, without passing through any conversion units, is called an internal recycle. An internal recycle is stripped of Fischer–Tropsch products in the product recovery section, but it is otherwise not chemically modified. The relative composition of the light compounds, including unconverted syngas, is the same as that at the outlet of the Fischer–Tropsch reactor. 2) External recycle. Tail gas that passes through one or more separation steps and/or conversion units before being recycled to Fischer–Tropsch synthesis, is called an external recycle. In an external recycle, the H2 :CO ratio of the syngas is typically adjusted by syngas conditioning, and in the case of a methane-rich tail gas, it may even be used as feed for syngas generation. Adding an internal recycle to a Fischer–Tropsch gas loop design does not add much complexity. Gas loop configurations with only an internal recycle are referred to as open gas loop designs
External recycle
Tail gas processing
Internal recycle
Syngas Figure 5.3
Fischer−Tropsch
Purge gas
Tail gas
Product recovery
Syncrude
Closed gas loop design showing internal recycle and external recycle of tail gas.
5.3 Syncrude Cooling and Separation
by some authors. Adding an external recycle usually adds complexity to the gas loop design. The external recycle may be processed in units that are already present, or it may entail new units that are used only for the processing of tail gas in the external recycle. The units that are typically considered are gas reforming, acid gas removal, and water gas shift conversion (Chapter 3). Gas reforming is useful for the conversion of methane in the tail gas to syngas, acid gas removal is useful for removing CO2 produced during synthesis, and water gas shift is useful for conditioning the syngas by adjusting the composition. Closed gas loops must contain a purge gas stream (Figure 5.3). This is necessary to avoid the buildup of inert compounds such as Ar and N2 . These inert gases enter the gas loop with the syngas, and to achieve steady state operation, these inert gases must leave the gas loop at the same rate. The concentration of inert gases in the purge gas is usually low, and depending on its heating value, the purge gas can be used as fuel gas.
5.3 Syncrude Cooling and Separation
The combined product that is obtained after Fischer–Tropsch synthesis contains unconverted syngas and the syncrude produced by syngas conversion. The gas loop itself is mainly concerned with the flow of the syngas. The product recovery section is where the gas loop interfaces with the Fischer–Tropsch refinery. It is here, in the product recovery section, that the hot products from Fischer–Tropsch synthesis are cooled down in order to separate the syncrude from the unconverted syngas. From the perspective of gas loop design, syncrude cooling and separation is necessary to limit the amount of syncrude in the internal recycle (Figure 5.3) and separation efficiency is all that matters. From a refining perspective, syncrude cooling and separation is a potentially valuable prefractionation of the syncrude that can improve the efficiency of the Fischer–Tropsch refinery design. It is strongly advocated that the engineering of the product recovery section should be performed as part of the refinery design, with the gas loop design merely specifying the separation efficiency that is required. Fischer–Tropsch synthesis determines the composition of the syncrude. The impact of different variables on the actual syncrude composition has been discussed (Chapter 4). The product from high-temperature Fischer–Tropsch (HTFT) synthesis is completely in the gas phase when it leaves the synthesis reactor. No reactor-specific prefractionation occurs and the design of the product recovery section is solely responsible for determining the prefractionation of the syncrude. The product from low-temperature Fischer–Tropsch (LTFT) synthesis is not only much heavier but also prefractionated in the reactor. A waxy product, which is a liquid at reactor conditions, exists as a condensed phase in the synthesis reactor and its cut point is not determined by the design of the product recovery section. The gas-phase product that leaves the LTFT reactor is the only fraction of the syncrude where the design of the product recovery section can influence the way in which it is prefractionated. The product recovery section after HTFT synthesis (Section 5.3.4) is consequently different in a number of respects from the product recovery section after LTFT synthesis (Section 5.3.5).
109
110
5 Fischer–Tropsch Gas Loop
5.3.1 Pressure Separation
In most gas loop designs, most of the syncrude is recovered at synthesis pressure. This is mainly done for two reasons. Firstly, it allows the internal recycle of closed loop designs to remain near synthesis pressure, thereby reducing the recompression cost associated with recycling. Secondly, it allows most C3 and heavier hydrocarbons to be recovered by condensation at temperatures of 45 ◦ C and above. Normal cooling water can be used in the final cooling stage, and it is not necessary to resort to cryogenic cooling. Another potential advantage of syncrude recovery under pressure is related to energy efficiency. Depending on the pressure of the gas loop, the condensation temperature of water that is produced during Fischer–Tropsch synthesis may be sufficiently high to enable steam generation. On a mass basis, more than half of the syngas that is converted ends up as water (CO + 2H2 → CH2 + H2 O). The energy efficiency of the process can be improved if the condensation energy from the water can be recovered as steam. Product recovery under pressure also has an inherent disadvantage. The relative volatility difference between compounds decreases with increasing pressure. This causes the quality of separation to deteriorate. The syncrude fractions produced by stepwise cooling have broad carbon number distributions and much carbon number overlap. This decreases the value of prefractionation for refining. Nevertheless, by careful design of the stepwise cooling, it is still possible to generate fractions that improve refining efficiency. 5.3.2 Cryogenic Separation
Cryogenic separation in general improves the carbon efficiency of the process. Methane and ethane are the most hydrogen-rich compounds in the Fischer–Tropsch syncrude. This implies that they are especially valuable compounds from a carbon efficiency point of view, even more so when the raw material has a low H:C ratio [3]. Ethene is a valuable chemical in its own right, with many industrial uses [4]. Ethane can be converted to ethene with high selectivity, and the recovery of C2 hydrocarbons provides a good basis for petrochemical production. The recovery and separation of C1 –C2 hydrocarbons is a design decision. The design of a generic closed gas loop with cryogenic separation is shown in Figure 5.4. An internal recycle can be taken from the tail gas before cryogenic separation. This may seem unusual, since it is possible to recover pure syngas from cryogenic separation. It is a trade-off between the costly separation of H2 and COx from the C1 –C2 hydrocarbons and the additional Fischer–Tropsch reactor volume ‘‘wasted’’ by recycling C1 –C2 hydrocarbons. Ethene, of course, may be incorporated into heavier products during Fischer–Tropsch synthesis, and strictly speaking, it is not inert [5]. The CO2 that is present in the tail gas must be removed before cryogenic separation to avoid problems with solidification during cooling below −50 ◦ C. The C1 –C2 hydrocarbons and syngas compounds can be separated by cryogenic distillation. Cryogenic separation may include several distillation columns employing cooling cycles with heat integration to minimize the refrigeration cost [6]. The overhead product from the last cryogenic separation step contains the syngas compounds and forms the external recycle. This stream also contains the inert gases that have
5.3 Syncrude Cooling and Separation Purge gas PSA H2
Syngas-rich
External recycle
Cryogenic separation
CH4
Gas reformer
C2 hydrocarbons
Acid gas removal
CO2
Internal recycle Fischer−Tropsch
Syngas
Syncrude
Product recovery
Figure 5.4 Generic closed gas loop design with cryogenic separation, pressure swing absorption (PSA) for hydrogen recovery, and gas reforming of the methane to supplement syngas production.
to be purged. To reduce the loss of syngas to the purge, hydrogen can be recovered by pressure swing absorption (PSA) before the final gas purge. The methane that is recovered during cryogenic separation can be reformed to produce additional syngas. Alternatively, the methane can be exported as synthetic natural gas. 5.3.3 Oxygenate Partitioning
Oxygen is more electronegative than carbon, which results in a dipole moment in oxygenate compounds. The dipole moment arises due to a separation of charge (Figure 5.5), with the oxygen becoming more negative than neutral and the adjacent carbon becoming more positive than neutral. This separation of charge may have an inductive effect on bonds further away from the oxygen-carbon bond. The inductive effect rapidly decreases with distance from the oxygen functionality.
H
H δ+ C
δ− O H
H Dipole moment (separation of charge)
H H δδ+ δ+ H C C H
δ− O H
H
H
H
H
H
H
H
C
C
C
C
C
H H H δδ+ δ+ C C C
H
H
H
H
H
H
0 Inductive effect
Figure 5.5 Polar nature of oxygenates resulting from the dipole moment caused by separation of charge and its inductive effect to polarize bonds further from the oxygen-carbon bond.
Apolar
H
δ− O H
H + Polar
−
111
112
5 Fischer–Tropsch Gas Loop
In oxygenates that are not symmetrical with respect to oxygen, the dipole moment renders one side of the molecule more positive and the other side more negative. When this polarization extends over the whole molecule, we think of it as a polar compound. As the chain length of the molecule increases, the fraction of the molecule that is polarized decreases and the compound becomes more apolar (Figure 5.5). Aliphatic hydrocarbons have no dipole moment and are apolar. When a compound is condensed in the presence of two liquid phases, one being polar and the other being apolar, it preferentially dissolves in the phase best fitting its nature. Polar molecules dissolve in the polar phase and apolar molecules dissolve in the apolar phase. The situation becomes more complicated when the molecules have both polar and apolar character. Such compounds are called surface active, because they can satisfy both the polar and apolar natures by arranging themselves at the interface between the two phases. Fischer–Tropsch syncrude mainly consists of hydrocarbons, but it also contains percentage levels of oxygenates (Table 1.2). The main oxygenate classes in the syncrude are alcohols, carboxylic acids, and carbonyls. The short-chain oxygenates are very polar compounds with little apolar hydrocarbon character. The polarity of oxygenates decreases with increasing chain length, but even the long-chain oxygenates display some surface-active behavior. During the stepwise cooling of the syncrude, the condensing hydrocarbons form an apolar organic phase and the condensing short-chain oxygenates and water form a polar aqueous phase. The interface region can accommodate very little oxygenates and most of the oxygenates must dissolve either in the organic phase or in the aqueous phase. This causes a partitioning of oxygenates between the two liquid phases, with some oxygenates being found in both phases. Under equilibrium conditions, compounds partition between the two liquid phases in such a way that the Gibbs free energy is minimized. The C2 –C4 oxygenates preferentially dissolve in the aqueous phase, whereas the heavier oxygenates preferentially dissolve in the organic phase (Table 5.1) [7]. This does not imply that C2 –C4 oxygenates are found in the aqueous phase only, nor does it imply that the heavier oxygenates are exclusively found in the organic phase. Table 5.1 Normal boiling point temperature and phase preference of C1 –C7 linear oxygenate isomers that may be encountered in Fischer–Tropsch syncrude. The lighter oxygenates preferentially partition to the aqueous phase and the division line indicates the transition between preferred water solubility and preferred oil solubility.
Normal boiling point of linear isomers (◦ C)
Carbon number 1 2 3 4 5 6 7
n-Alkane −164 −89 −42 −0.5 36 69 98
Alcohol 65 79 82–97 100–117 116–138 135–158 160–176
Aldehyde −21 21 49 76 103 128 153
Ketone – – 56 80 101–102 125–128 144–151
Carboxylic acid 101 118 141 166 186 205 223
5.3 Syncrude Cooling and Separation
It is further noted from Table 5.1 that there is a significant boiling point difference between oxygenates and hydrocarbons of the same carbon number. There is carbon number difference of 2 for carbonyls, 4 for alcohols, and 6 for carboxylic acids. This has important ramifications for the refinery, because a loss of oxygenate functionality decreases the boiling point. The boiling point decreases by 50–150 ◦ C depending on the oxygenate class that was converted, thereby causing boiling point broadening. Depending on the engineering of the product recovery section, equilibration may not take place. Transport limitations and insufficient contact time between the different phases result in oxygenate partitioning that is far from equilibrium. It may also result in variable partitioning of oxygenates, which can wreak havoc in downstream units in the gas loop and in the Fischer–Tropsch refinery. Oxygenates can upset acid gas removal [8]. Oxygenates also affect refining catalysis [9]. Particular care should be taken in the detailed engineering design to ensure that all the short-chain carboxylic acids report to the aqueous phase. Short-chain carboxylic acids are corrosive and the process equipment that is exposed to syncrude containing such acids should preferably be constructed from stainless steel. The corrosion products from short-chain carboxylic acids are detrimental to refinery operation [10]. 5.3.4 HTFT Syncrude Recovery
The syncrude from HTFT synthesis is completely in the vapor phase when leaving the Fischer–Tropsch reactor. Syncrude recovery takes place by stepwise cooling of the hot product gas after synthesis. The temperature at which each fraction is collected and the way in which the cooling is performed are matters related to design. The product fractions collected in a typical industrial installation are (Figure 5.6) as follows: 1) Decanted oil. The heaviest oil condenses first and it is called the decanted oil (DO). With proper design, the DO can be used to trap the fine Fischer–Tropsch catalyst particles that have not been removed from the product gas at the reactor outlet. It can also be used to trap
Tail gas Tail gas separation
Condensates Light oil
HTFT
Syngas
Aqueous product (“ reaction water ”) Decanted oil Waste oil (“gunk ”)
Figure 5.6 Generic HTFT syncrude recovery section showing the product fractions obtained by stepwise cooling in a typical industrial operation.
113
114
5 Fischer–Tropsch Gas Loop
2)
3)
4)
5)
the metal carboxylates produced by corrosion and catalyst leaching. This stream may contain some distillate-range compounds, but it is primarily a broad cut that is representative of atmospheric residue in crude oil terminology. Provision is normally made for the catalyst particles to settle out, so that most of the catalyst can be removed in a waste oil stream called ‘gunk’. The decanted oil is the clear oil. Light oil. The light oil is cocondensed with the water to form two liquid phases. It is in this recovery step that oxygenate partitioning (Section 5.3.3) takes place. The light oil that is recovered by condensation is saturated with dissolved light gases. To indicate that the light oil is saturated with dissolved gases, it is called unstabilized light oil (ULO), and once the dissolved light gases have been stripped from the oil, it is called stabilized light oil (SLO). Typically, the light oil mainly contains naphtha and distillate-range material. It is a reactive syncrude fraction that is rich in alkenes and oxygenates. Aqueous product. The aqueous product, or reaction water, is cocondensed with the light oil. The oxygenate content in the aqueous product can be as high as 6% [11]. Ideally, all corrosive short-chain carboxylic acids should be recovered in this product phase. In addition to the short-chain carboxylic acids, the aqueous product contains C1 –C4 alcohols and carbonyl compounds. Condensates. If the light oil was recovered in a single step, the tail gas will still contain a significant fraction of the C5 –C7 hydrocarbons in addition to normally gaseous hydrocarbons. It is possible to recover the C3 –C4 liquid petroleum gas (LPG) fraction, as well as the C5 –C7 fraction, from the tail gas without resorting to cryogenic cooling. The syncrude thus obtained is called condensate, irrespective of whether it has been obtained by cryogenic separation or not. The main difference between the C5 –C7 light naphtha fraction in the condensate and the same fraction in the light oil is their respective oxygenate concentration. The condensate fraction has very little oxygenates compared to the equivalent light oil fraction which is a consequence of oxygenate partitioning (Section 5.3.3). This makes refining of the condensates easier than that of the light oil. Tail gas. The syncrude content of the tail gas depends on the manner in which the condensates are recovered. Recovery of condensates by cryogenic separation (Section 5.3.2) yields a tail gas that is substantially free of syncrude, whereas a recovery by ambient pressure distillation yields a tail gas that contains most of the C1 –C2 hydrocarbons produced during Fischer–Tropsch synthesis.
5.3.5 LTFT Syncrude Recovery
The syncrude from LTFT synthesis leaves the reactor in two separate streams. One stream is in the liquid phase and mainly contains molten wax. The other stream is in the vapor phase and it contains the unconverted syngas and lighter syncrude compounds. Syncrude recovery from the vapor phase takes place in a fashion analogous to the syncrude recovery from HTFT synthesis (Section 5.3.4), namely, by stepwise cooling of the hot product gas after synthesis. It is likewise a matter of design that determines at what temperature each fraction is collected and how the cooling is performed. The product fractions collected in a typical industrial installation are (Figure 5.7) as follows:
5.3 Syncrude Cooling and Separation
Tail gas Cold condensate
LTFT
Syngas
Aqueous product (“reaction water”) Hot condensate Wax
Figure 5.7 Generic LTFT syncrude recovery section showing the product fractions obtained by stepwise cooling in a typical industrial operation.
1) Wax. The heavy hydrocarbons in the syncrude are present as a liquid product at the operating conditions of LTFT synthesis. This product is directly obtained from the Fischer–Tropsch reactor. Depending on the reactor technology, the wax can be separated outside of the reactor, which enables further cooling and condensation of vapor-phase products (not shown in Figure 5.7). However, there is a lower temperature limit. The wax fraction should not be cooled down below its congealing point. In the case of high α-value LTFT catalysts, the heavy wax fraction can have a congealing point above 100 ◦ C. 2) Hot condensate. The heaviest vapor fraction that is condensed after the wax is still very waxy, and it has to be kept hot to avoid congealing at ambient conditions. The unfortunate nomenclature that has evolved for this lighter wax fraction is ‘‘hot condensate.’’ This should not be confused with the condensates in HTFT nomenclature. 3) Cold condensate. Further condensation of the vapor results in two liquid phases and is analogous to light oil condensation after HTFT synthesis. It is in this recovery step that oxygenate partitioning (Section 5.3.3) takes place. The cold condensate is effectively a light oil and again the choice of nomenclature is unfortunate. 4) Aqueous product. The aqueous product is also called reaction water, and it is cocondensed with the cold condensate. Although the oxygenate content in an LTFT aqueous product is lower than that of HTFT, there may still be around 2–3% oxygenates in this fraction. The same dictum as for HTFT applies; it is preferable to recover all short-chain carboxylic acids in the aqueous product. Alcohols and carboxylic acids are the main oxygenate classes dissolved in the water. 5) Tail gas. The design and separation intensity of cold condensate recovery determine which hydrocarbons remain in the tail gas. It is possible to recover the remaining light naphtha and C3 –C4 LPG from the tail gas without cryogenic separation. The same principles of recovery of light hydrocarbons from HTFT tail gas (Section 5.3.4) apply to LTFT tail gas. The main difference is that the light hydrocarbon fraction in LTFT syncrude is considerably less than that in HTFT syncrude. However, the contribution of light gases to the total syncrude increases with Fischer–Tropsch catalyst deactivation (Section 4.5) and a decision regarding the level of tail gas processing should bear this in mind.
115
116
5 Fischer–Tropsch Gas Loop
References 1. Dry, M.E. and Steynberg, A.P. (2004) Commer-
2.
3.
4. 5.
6.
7. Weast, R.C. (1988) CRC Handbook of Chemistry cial FT process applications. Stud. Surf. Sci. and Physics, CRC Press, Boca Raton, FL. 8. Hanekom, P. and Gibson, P. (1997) ContamCatal., 152, 406–481. De Klerk, A. (2007) Environmentally friendly ination of Benfield CO2 removal system by refining: Fischer-Tropsch versus crude oil. Green carboxylic acid salts. Ammonia Plant Saf., 37, Chem., 9, 560–565. 281–289. 9. De Klerk, A. and Furimsky, E. (2010) Catalysis De Klerk, A. (2010) Indirect liquefaction carbon in the Refining of Fischer–Tropsch Syncrude, Royal efficiency. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Society of Chemistry, Cambridge. Chem., 55 (2), 338–339. 10. De Klerk, A. (2008) Hydroprocessing peculiariMiller, S.A. (1969) Ethylene and Its Industrial ties of Fischer-Tropsch syncrude. Catal. Today, Derivatives, Ernest Benn, London. 130, 439–445. Tau, L.-M., Dabbagh, H.A., and Davis, B.H. 11. Hoogendoorn, J.C. and Salomon, J.M. (1957) (1990) Fischer-Tropsch synthesis: 14 C tracer Sasol: World’s largest oil-from-coal plant. III. Br. study of alkene incorporation. Energy Fuels, 4, Chem. Eng., 368–373. 94–99. Chauvel, A. and Lefebvre, G. (1989) Petrochemical Processes. 1. Synthesis-gas Derivatives and Major Hydrocarbons, Editions Technip, Paris.
117
Part III Industrial Fischer–Tropsch Facilities
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
119
6 German Fischer–Tropsch Facilities 6.1 Introduction
Construction of the first industrial-scale Fischer–Tropsch facility by Ruhrchemie AG, started in 1935 in Germany [1]. It was soon thereafter followed by the construction of more Fischer–Tropsch-based facilities, which can broadly be grouped into two categories based on the operating pressure of the Fischer–Tropsch synthesis section. The first industrial installations operated at below 0.1 MPa using the ‘‘normal-pressure synthesis’’ (Ger. ‘‘Normaldruck-Synthese’’) or atmospheric process. Later on, a ‘‘medium-pressure synthesis’’ (Ger. ‘‘Mitteldruck-Synthese’’) process was employed, which operated at 0.5–1.5 MPa (75–220 psi). The operating temperature of synthesis was usually in the range 180–200 ◦ C [2], making these both low-temperature Fischer–Tropsch (LTFT) processes. The German Fischer–Tropsch facilities all employed coal as the raw material and a total of nine plants were constructed in Germany (Table 6.1) [1, 3, 4]. A further five plants were constructed under license outside Germany, one in France, three in Japan, and one in China. At the beginning of 1944, the coal-to-liquids industry provided close to two-thirds of the German fuel and oil consumption, with Fischer–Tropsch synthesis producing 7%, coal hydrogenation 40%, benzene 5%, and coal tar distillation 12% [1]. Many of the plants were damaged by the Allied bombing offensive during the latter half of 1944 and early 1945. After the Second World War, only Brabag, Essener, and Krupp started up again and the last of these were shut down for economic reasons in 1962 [5]. There is diversity in the design of the different Fischer–Tropsch facilities. All produced transportation fuels and some additionally produced lubricant base oils and chemicals. The subsequent discussion gives an overview of the technology associated with the German Fischer–Tropsch technology. Rather than discuss each plant separately, the refining of the syncrude is discussed generically. 6.2 Synthesis Gas Production
The majority of facilities produced their synthesis gas from coke, not coal [2]. Specifically, the plants in the Ruhr area employed coke and coker gas for syngas production, whereas the plants in central Germany directly employed brown coal for the production of syngas [6]. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
120
6 German Fischer–Tropsch Facilities Table 6.1
German Fischer–Tropsch facilities constructed in the period 1935–1939.
Plant
Brabag Essener Hoesch Kruppc Rheinpreussen Ruhrchemie Schaffgotsch Viktor Wintershall
Location
Ruhland-Schwarzheide Bergkamen Dortmund Wanne-Eickel Homberg Sterkrade-Holten Deschowitz-Beuthen Castrop-Rauxel Lutzkendorf-Mucheln
Coal type as feed
Lignite Bituminous Bituminous – Bituminous Bituminous Lignite Bituminous Lignite
Production (kt a−1 )a Normalpressure
Mediumpressure
164.6 80.0 0 48.3 67.2 18.1 0 37.7 11.5
0 0 46.0 11.9 0 44.5 26.4 0 0
Products
Fuel Fuel Fuelb Fuel Fuel, acidsb Fuel, lubricant Fuel Fuel Fuel
a Based
on actual production figures for 1942. amount of lubricants coproduced. c First stage of synthesis is normal-pressure, second stage is medium-pressure. b Small
The first step in synthesis gas production was to convert the coal into coke using a coking oven (Figure 6.1) [7]. The coke obtained from the coking oven was then gasified with O2 and steam to produce a syngas with H2 :CO ratio around 1 : 1. For example, at the Rheinpreussen facility, coke gasification took place in Koppers high-temperature-entrained flow gasifiers (Section 3.4.4) [2]. In almost all of the German facilities, the syngas was cleaned in a two-step process. The first ‘‘coarse purification’’ step employed the standard iron oxide process, which was a standard process for H2 S removal at that time. The H2 S was removed by passing the raw syngas over a large bed of iron oxide (Fe2 O3 ) at 40–50 ◦ C and near atmospheric pressure [8]. The iron oxide was usually dispersed on a porous material such as wood shavings. Regeneration took place by reoxidation with air and subsequent extraction of the sulfur [9]. This was an unwieldy, but widely adopted technology, which required a large plot space owing to the size of the multiple iron oxide boxes. In the second ‘‘fine purification’’ step, the organic sulfur compounds were removed by passing the syngas over an Fe2 O3 –Na2 CO3 catalyst (100 : 16 mixture by mass [9]) at a temperature of 175 ◦ C and increasing the temperature with catalyst age up to 280 ◦ C. The addition of around 0.2 vol% O2 to the synthesis gas was beneficial for both sulfur removal steps [2]. The target specification for sulfur was less than 2 mg·m−3 syngas [9]. The water gas shift (WGS) catalyst employed for syngas conditioning was a typical Fe2 O3 –Cr2 O3 -based high-temperature WGS catalyst. The WGS catalyst contained 38.5% Fe2 O3 , 18.2% CaO, 5.4% Cr2 O3 , 5.2% MgO, 18.0% H2 O, and various other minor constituents [2]. It was reported that coking the coal beforehand had some additional advantages. The carbonization of coal in a coke-oven produced two important by-products apart from coke. When the coal is pyrolyzed, the coal pyrolysis liquids (coal tar) could be recovered as a light oil rich in aromatic compounds. A typical light oil obtained by condensation from coke-oven
6.3 Fischer–Tropsch Synthesis
Coker gas Coal tar separation
Coal
Thermal reformer (noncatalytic)
H2:CO > 2:1 Light oil (coal tar)
Aromatic oil
Fischer−Tropsch synthesis
Coke-oven
Syncrude and tail gas
Coke O2 steam
Gasifier
Water gas shift
H2:CO = 1:1
Iron oxide process
Final purifier (Fe2O3−Na2CO3)
Sulfur removal Figure 6.1 facilities.
Generic process flow diagram of synthesis gas production in German Fischer–Tropsch
gas contained 57% benzene and 13% toluene as the major components, the remainder being sulfur compounds, aliphatic hydrocarbons, and heavier aromatics [10]. These coal liquids were valuable blending components in fuels production (Sections 6.4.2 and 6.4.3). Coker gas consisted mainly of H2 (50–55%), CO (5–10%), and CH4 (25–30%) on a dry basis, the rest being combustion gases and inerts [11]. The coke-oven gas could be thermally reformed in a noncatalytic process with steam at high temperature (1200 ◦ C) to produce a hydrogen-rich syngas. By doing so, it reduced the amount of gas from coke gasification that had to be subjected to WGS in order to obtain an H2 :CO ratio of 2 : 1 for Fischer–Tropsch synthesis [2].
6.3 Fischer–Tropsch Synthesis
Much of the development leading to the industrial application of Fischer–Tropsch synthesis has to be credited to Otto Roelen. He was responsible for the development of the cobalt-based catalyst that became the standard Fischer–Tropsch catalyst in all Fischer–Tropsch facilities based on the German Ruhrchemie technology [1]. The initial Co-LTFT catalyst consisted of a mixture of Co, ThO2 , and kieselguhr in a ratio of 100 : 18 : 100. The composition was modified in 1938 to include magnesium oxide, to become the standard catalyst throughout the Second World War. This catalyst employed Co, ThO2 , MgO, and kieselguhr in a ratio of 100 : 5 : 8 : 200 [12]. The particle size of the catalyst was 2–3 mm [13]. The bulk density of the catalyst was such that the Co content was about 0.1 kg·m−3 of reactor volume [9].
121
122
6 German Fischer–Tropsch Facilities
Three modes of deactivation were identified, namely, pore blockage by the product wax, poisoning by sulfur and carbon deposits, and sintering. No oxidation or cobalt silicate formation was observed [12]. The absence of oxidation is likely a consequence of the low-temperature and low water partial pressure operation of the German Fischer–Tropsch technology. The medium-pressure process also experienced metal leaching [13], which is further discussed in Section 6.3.2. 6.3.1 Normal-Pressure Synthesis
Fischer–Tropsch synthesis by the normal-pressure process was carried out in the operating range 180–200 ◦ C and 30 kPa. The gas hourly space velocity (GSHV) was on the order of 100 h−1 at the inlet. The normal-pressure Fischer–Tropsch reactors were about 5 m long, 2.5 m wide, and 1.5 m high. Each reactor consisted of tubes and heat transfer plates, with the catalyst being loaded on the shell-side between the heat transfer plates (Figure 6.2) [2, 9, 13]. Water was circulated in the tubes to regulate the temperature. The catalyst bed was typically operated at a temperature 5–8 ◦ C higher than that inside the tubes. The reactor temperature could be controlled to within 1 ◦ C in the range 170–200 ◦ C by regulating the water pressure, using a boiler principle. The tubes could withstand a pressure of 3 MPa. The heat release during normal-pressure Fischer–Tropsch synthesis was about 150 kJ·m−3 synthesis gas, which is equivalent to about 1.5 MJ·kg−1 of product [9]. Each reactor had a total internal volume of 12 m3 for catalyst and tubes, and each reactor contained 900 kg of Co [1]. In order to achieve a reasonable synthesis gas conversion, normal-pressure plants employed two or three Fischer–Tropsch synthesis stages in series. After each reaction stage, the condensable products were removed by direct condensation in a spray cooler (Figure 6.3), through which water formed during synthesis was also removed that in turn reduced oxidation of the catalyst in the later stages of synthesis [9]. Another advantage was that the volume of the gas feed to the following stage was reduced, which improved the utilization of reactor volume (Table 6.2) [9].
Catalyst (between plates) Water tubes
Plate Outside 34 mm Inside 29 mm
Figure 6.2
Internals of a normal-pressure Fischer–Tropsch reactor.
6.3 Fischer–Tropsch Synthesis
First-stage synthesis
Tail gas
Second-stage synthesis
Carbon absorber
LTFT
LTFT Crude LPG
Syngas
Carbon gasoline
LTFT
Condensate oil Aqueous product Catalyst wax
Figure 6.3 Synthesis section of a typical two-stage normal-pressure German Fischer–Tropsch facility. Dashed lines indicate discontinuous flow. Table 6.2 Performance of two-stage normal-pressure Fischer–Tropsch synthesis over a Co–ThO2 –MgO–kieselguhr catalyst at the Ruhrchemie facility.
Description Gas volume (m3 )a Syngas volume (m3 ) Overall syngas conversion (%) Overall CO conversion (%) Gas composition (vol%) H2 CO CO2 CH4 N2 Hydrocarbons a
Inlet
After stage 1
After stage 2
100 79.9 – –
50.7 26.1 67 78
34.7 8.9 89 96
53.2 26.7 14.4 0.4 5.3 0
33.7 17.8 29.6 7 10.4 1.5
16 9.7 44.2 13.2 14.9 2
All data arbitrarily expressed on the basis of 100 m3 total gas feed to the first stage.
More reactors were required for the first stage of conversion than were needed for the second stage of conversion. For example, at Ruhrchemie plant, the normal-pressure synthesis section employed 18 reactors for the first stage and 9 reactors for the second. Under normal-pressure operation, the Co-LTFT catalyst typically had a lifetime of about four to six months. Deactivation because of blockage by wax required rejuvenation every
123
124
6 German Fischer–Tropsch Facilities
700 h. Rejuvenation entailed catalyst washing at 100 ◦ C by spraying it with kerosene to extract the wax. The product thus obtained was called catalyst wax. The catalyst could also be treated with H2 at 200 ◦ C to extend its useful operating lifetime [13]. At the normal rate of catalyst deactivation, a temperature increase of 1 ◦ C was needed to maintain constant conversion. The product recovery section at the end of the last synthesis stage entailed condensation and absorption. The product was first cooled down by a spray cooler to yield an oil and a water fraction, before the uncondensed gas was passed over a bed of activated carbon. The gaseous hydrocarbons were adsorbed onto the activated carbon bed from which the product was recovered batch-wise by steaming (Figure 6.3). The desorbed product from activated carbon adsorption was then stabilized by pressure distillation to yield an ‘‘active carbon gasoline’’ and a ‘‘crude liquid petroleum gas (LPG).’’ The composition of the syncrude fractions from a typical normal-pressure process is given in Table 6.3 [9]. In addition to the hydrocarbon products, the syncrude also contained oxygenates, mostly alcohols and carboxylic acids, which gave it a characteristic smell. The amount of syncrude that dissolved into the aqueous product was very little (<1%). The aliphatic hydrocarbons in the naphtha fraction from normal-pressure Co-LTFT synthesis have been analyzed in detail by Friedel and Anderson (Table 6.4) [14]. The high degree of linearity of the acyclic aliphatic hydrocarbons is typical of Fischer–Tropsch synthesis. The 1-alkene content is quite low for Fischer–Tropsch synthesis, and the composition of the alkene fraction indicates that a significant amount of secondary reactions took place during the normal-pressure process.
Table 6.3 Composition of the various syncrude streams obtained from German normal-pressure Fischer–Tropsch synthesis. The aqueous product contains <1% of the syncrude.
Description Mass of total syncrude (%) Composition (%) H2 , CO, and CO2 Methane Ethane and ethene Propane and propene C4 hydrocarbons C5 hydrocarbons C6 -180 ◦ C fraction 180–230 ◦ C fraction (Kogasin I) 230–320 ◦ C fraction (Kogasin II) 320–460 ◦ C fraction (Slack wax) >460 ◦ C fraction (Hard wax)
Catalyst wax
Condensate oil
Carbon gasoline
Crude LPG
2
40
50
8
– – – – – – – – – 20–30 70–80
– – – – – – 1–3 35–40 30–35 20 1
– – – 1 5–15 15–20 70–75a – – – –
15–40 2–3 1–2 15–20 20–40 10–20 3–5 – – – –
a In the source reference the carbon gasoline did not add to 100%; only the C fraction (20–25%) was tabulated and the 6 heavier gasoline was omitted from the table.
6.3 Fischer–Tropsch Synthesis Composition of the acyclic aliphatic hydrocarbons in the C5 –C8 naphtha from the German normal-pressure Fischer–Tropsch process.
Table 6.4
Composition Alkane selectivity (vol%) Linear Methyl branched Alkene selectivity (vol%) 1-Alkene cis-Alkene (internal) trans-Alkene (internal) a
C5
C6
C7
C8
95 5
90 10
88 12
85 15
36 25 39
28 30 42
18 30 52
– – –
a a a
Not reported.
6.3.2 Medium-Pressure Synthesis
The yield of wax and liquid products from Co-LTFT synthesis improved when the pressure was increased (Table 6.5) [7]. The medium-pressure process was developed to exploit this yield increase by operating at 180–200 ◦ C and 0.5–1.5 MPa. The GSHV was on the order of 100 h−1 at the inlet. It was also found that the catalyst lifetime improved to six to nine months by operating at a higher pressure, because of the solvent action of the condensed products in the reactor [13]. The medium-pressure Fischer–Tropsch reactor (Figure 6.4) was of a different design compared to the normal-pressure reactor (Figure 6.2). The reactor consisted of a vertical pressure vessel, 4.5 m in height and 2.7 m in diameter. On the inside it was fitted with 2044 double jacketed Table 6.5 Effect of pressure on the product yield obtained with a Co–ThO2 –kieselguhr catalyst averaged over a four-week period.
Product yield (g·m−3 of CO + H2 )a
Pressure (MPa)
0 0.15 0.5 1.5 5 15 a Syngas b All
C5 -200 ◦ C
>200 ◦ C oil
Oil-free wax
Wax and liquidsb
35 46 30 30 19 19
35 40 48 34 34 32
9 14 56 66 50 24
109 122 139 135 128 98
volume at normal conditions. solid and liquid products, including LPG.
125
126
6 German Fischer–Tropsch Facilities
Detail of double tube
Catalyst (in annulus) Water
24 mm 48 mm Figure 6.4
Internals of a medium-pressure Fischer–Tropsch reactor.
tubes. The catalyst was loaded into the annular space between the shell and the inner tube [2, 13]. Each medium-pressure reactor had space for about 10 m3 of catalyst, which was equivalent to about 1 ton of Co [9]. The configuration of the synthesis section was otherwise very similar to that of the normal-pressure process (Figure 6.3). Multiple stages were employed for the same reasons as in the case of the normal-pressure process, thus achieving a high syngas conversion (Table 6.6) [13]. The GHSV was increased in each stage, with the last stage having double the space velocity of the first stage. The lower H2 :CO ratio was also found to be beneficial in compensating for the lower alkene selectivity and to improve the overall conversion efficiency of the process (Section 6.3.4). Table 6.6 Performance of three-stage medium-pressure (0.9–1.0 MPa) Fischer–Tropsch synthesis over a Co–ThO2 –MgO–kieselguhr catalyst at the Ruhrchemie facility.
Description Gas volume (m3 )a Syngas volume (m3 ) Overall syngas conversion (%) Overall CO conversion (%) Gas composition (vol%) H2 CO CO2 CH4 N2 Hydrocarbons a All
Inlet
After stage 1
After stage 2
After stage 3
100 87.4 – –
49.6 34.2 61 61
26.9 11.3 87 91
18.9 4.8 94 98
52.7 34.7 6.9 0.4 5.2 0
33.9 35 14.7 5 10.5 0.9
18.2 23.9 28 11.3 17.4 1.15
11.5 14 36.3 15.7 21 1.25
data arbitrarily expressed on the basis of 100 m3 total gas feed to the first stage.
6.3 Fischer–Tropsch Synthesis
The higher operating pressure required some changes to the detailed design of the synthesis section. Many of these changes were necessitated by the corrosiveness of the short-chain carboxylic acids that were formed during medium-pressure Fischer–Tropsch synthesis. Parts of the plant were constructed with mild steel, which is not resistant to corrosion by aqueous solutions of carboxylic acids. In order to prevent such corrosion, condensation of the aqueous product phase had to be prevented. The gas lines leaving the synthesis reactor were insulated to prevent condensation and the gas from interstage spray coolers was preheated to 160 ◦ C [13]. After 90 h time on stream at 180 ◦ C and 1 MPa, the wax produced during the medium-pressure process turned black. With increasing time on stream, the ash content of the wax went through a maximum, decreasing to zero after 360 h from the start of the run. The ash contained carbon and mineral matter derived from the catalyst. In parallel, some cobalt leached into the aqueous product and remained at a level of 25–50 µg·g−1 even though the ash content in the wax decreased [13]. The increase in operating pressure resulted in a change in the syncrude composition between the normal- and medium-pressure processes (Table 6.7) [7, 9]. The increased pressure resulted in the formation of heavier and more hydrogenated products. The medium-pressure process was better suited for diesel fuel, chemicals, lubricants, and wax, whereas the quantity and quality of the gasoline deteriorated significantly. 6.3.3 Gas Loop Design
High conversions of syngas (90–95%) were obtained in both the normal- and the medium-pressure Fischer–Tropsch processes (Tables 6.2 and 6.6). The use of an open gas loop design is therefore not surprising. There was a desire to increase alkene production from the medium-pressure process, which would improve the quality of the gasoline, as well as its synthetic uses. It was found that by recycling the syngas after the first-stage conversion, the alkene content could be significantly increased without a loss in the overall product yield. In fact, the space–time yield was increased during recycling (Ger. ‘‘Kreislauf’’) operation. Despite the advantages of this closed loop design employing an internal recycle, it happened too late to be generally introduced [9]. Table 6.7 Product distribution obtained from German normal- and medium-pressure Fischer–Tropsch processes.
Product fraction
Tail gas C3 –C4 (LPG) C5 -180 ◦ C (Gasoline) 180–230 ◦ C (Kogasin I) 230–320 ◦ C (Kogasin II) 320–460 ◦ C (Slack wax) >460 ◦ C (Hard wax)
Normal-pressure process
Medium-pressure process
Yield (%)
Alkenes (%)
Yield (%)
Alkenes (%)
16 12 39 14 9 7 3
– 43 37 18 8 – –
14 9 22 20 11 15 9
– 40 24 10 – – –
127
128
6 German Fischer–Tropsch Facilities Table 6.8 Carbon efficiency of the German three-stage medium-pressure Fischer–Tropsch process under different operating conditions.
Description
Coke requirement (kg/kg product recovered) H2 + CO volume in feed (%) Overall syngas conversion (%) Average catalyst lifetime (h) Product yield (g/m3 )a of CO + H2 Total liquids and solids Liquids and solids, excluding LPG (C3 –C4 ) a Syngas
Feed gas H2 :CO ratio 2:1
1.6 : 1
4.99 81 90.8 4015
4.35 86.5 95 2745
126.2 138.4
156.6 144.4
volume at normal conditions.
The following specific advantages of introducing an internal recycle in the first conversion stage have been noted [4]: 1) 2) 3) 4) 5)
Higher concentration of alkenes in the product. Synthesis less sensitive to operating disturbances and variation in the syngas feed. Increased yield of products per volume of syngas. Lower catalyst cost and requirement of a small number of units for the same production. Shorter start-up period required to bring a synthesis reactor on line.
6.3.4 Carbon Efficiency
The US Navy Technical Mission in Europe has reported the carbon efficiencies of the three-stage German medium-pressure process using different operating conditions (Table 6.8) [4]. Operating at a lower H2 :CO ratio was beneficial for carbon efficiency, but the Co-LTFT catalyst lifetime was shortened. It is important to reiterate that the feed material is coke (Section 6.2). On a coal feed basis, the overall carbon efficiency of the process is higher, since additional liquid products from coal pyrolysis are coproduced in the coke-oven during coke production.
6.4 Fischer–Tropsch Refining
The main product fractions obtained from stepwise cooling after Fischer–Tropsch synthesis is indicated in Figure 6.3. It is convenient to discuss the refining of the Co-LTFT syncrude by tracing the upgrading pathway of each of these fractions. Some of the main design features found in the Fischer–Tropsch refineries associated with the normal- and medium-pressure processes
6.4 Fischer–Tropsch Refining
Tail gas Carbon absorber
LPG Oligomerization or hydration
Crude LPG
Oligomers
Kogasin I (180−230 °C) Kogasin II (230−320 °C) Slack wax (>320 °C)
Gasoline Gasoline
Carbon gasoline Spray cooler
Alcohols
Paraffin oxidation
Diesel fuel Diesel fuel
Fatty acids
NaOH wash Sweating / deoiling
Medium wax
Condensate oil Waste water
Aqueous product
Syngas
LTFT
Catalyst wax
Solvent recovery
Tm = 50−70 °C Steam stripping (vacuum)
Figure 6.5
Tm > 90 °C
Medium wax Hard wax
Generic German Fischer–Tropsch refinery. Dashed lines indicate discontinuous flow.
are illustrated in Figure 6.5; it is a generic refinery design and does not represent any specific refinery design. 6.4.1 Refining C3 –C4 Crude LPG
The German Fischer–Tropsch facilities upgraded the C3 –C4 syncrude fraction (crude LPG or ‘‘Gasol’’) in different ways. The alkene content in this fraction was very important, because it enabled two important upgrading pathways, namely, oligomerization and hydration. Some of the Fischer–Tropsch facilities, such as the Viktor-plant, made use of liquid phosphoric acid (H3 PO4 )-catalyzed oligomerization to convert the C3 –C4 alkenes into heavier alkenes. The method practised in Germany used a reactor consisting of three silvered tubes connected in series in a single water-cooled reactor. The tubes were 5 m high and had a diameter of 0.18 m. Each tube was loaded with 40 kg of acid and 70 kg of oligomer to about one-third of
129
130
6 German Fischer–Tropsch Facilities
its height. The lighter alkenes were then pumped through the tubes. The process had an inlet pressure of 6 MPa and outlet pressure of 4 MPa. Reaction took place at 180–200 ◦ C and an alkene conversion of 90% was achieved. Entrained phosphoric acid was separated from the product and recycled. Phosphoric acid catalyzed oligomerization is an important Fischer–Tropsch refining technology and it is discussed in detail in Section 19.3.1. Despite the high quality of the gasoline derived from this process, it was not practised on a large scale for the conversion of Fischer–Tropsch syncrude. The production of gasoline by Fischer–Tropsch synthesis was not favored in Germany at that time, and much of the gasoline was produced by direct coal liquefaction [2]. The C3 –C4 alkenes could, in principle, also be oligomerized using AlCl3 , as was done in the case of lubricant manufacturing. However, the quality of the lubricating oil deteriorates when lighter materials are employed [15]. Although these alkenes could, in principle, be coprocessed with the alkenes derived from thermal cracking of heavier Fischer–Tropsch fractions (Section 6.4.3), it is not clear whether this was done in practice. The C3 –C4 fraction was also employed as a feedstock for the production of alcohols by indirect hydration using the sulfuric acid process. In this process, the sulfuric acid solution dissolves the alkenes resulting in the formation of sulfuric acid esters (Equation 6.1), which is then hydrolyzed in a subsequent step to yield the alcohols (Equation 6.2) [16]. 2Cn H2n + H2 SO4 → (Cn H2n+1 )2 SO4
(6.1)
(Cn H2n+1 )2 SO4 + 2H2 O → 2Cn H2n+1 OH + H2 SO4
(6.2)
Oligomerization of the alkenes is a side reaction during this process. These olefinic oligomers make a good-quality gasoline blending stock, and the oligomers thus produced were blended into the gasoline. This is a nice example to demonstrate the synergy between hydration and Fischer–Tropsch refining. In isolation, alkene oligomerization as a side reaction may be viewed as a drawback, but in the context of refining it is not a drawback at all. The alkane-rich C3 –C4 fraction that remained after alkene conversion by either oligomerization or hydration was compressed and bottled as LPG. 6.4.2 Refining Carbon Gasoline
The straight-run properties of the carbon gasoline, also called ‘‘A-K’’ gasoline, were very dependent on the operating conditions of the Fischer–Tropsch reactors. The degree of branching in Fischer–Tropsch naphtha is generally low and the naphtha from the German Fischer–Tropsch process was no exception (Table 6.4). The n-alkanes have very low octane numbers and make poor gasoline. The only components in the straight-run material that had high octane numbers were the alkenes and specifically the internal alkenes, which were a major product in the normal-pressure process. The alkene content of the carbon gasoline was higher in the normal-pressure process than in the medium-pressure process (Table 6.7), and this was reflected in the octane numbers of gasoline from the two processes (Table 6.9) [2, 4, 13, 17, 18]. The German motor-gasoline specifications at that time required a motor octane number (MON), after tetraethyl lead (TEL) addition, of at least 72 and a minimum density of 720 kg·m−3 . The minimum density specification increased to 740 kg·m−3 if both TEL and aromatics were
6.4 Fischer–Tropsch Refining Research octane number (RON), motor octane number (MON), and density of different fractions from the straight-run naphtha of German normal- and medium-pressure Fischer–Tropsch synthesis.
Table 6.9
Naphtha fraction
C5 -90 ◦ Ca C5 -110 ◦ C C5 -110 ◦ Ca C5 -130 ◦ Ca C5 -140 ◦ C C5 -150 ◦ C C5 -160 ◦ Ca C5 -180 ◦ C (bauxite treated) C5 -200 ◦ C C4 –C10 (175 ◦ C) C4 –C11 (200 ◦ C) a Range
Normal-pressure process
Medium-pressure process
RON
MON
RON
MON
– – – – – 57 – – 43 52 49
73 67 66–68 58 62 55 49 50 40 – –
– – – – – 38 – – 25 28 25
– – – – – – – 25 – – –
Density (kg·m−3 )
656 – 670–671 683 – 660 694 – 704 689 693
based on T95 distillation temperature; the end point is typically 10–20 ◦ C higher.
added to the gasoline [3]. The straight-run Fischer–Tropsch gasoline fell short of both octane number and density specifications. It was therefore a common practice to blend the poor-quality synthetic gasoline from Fischer–Tropsch synthesis with the aromatic-rich coal tar (‘‘benzole’’) and alcohols in line with the then-existing German motor-gasoline specifications [2]. A mild caustic washing step was also required as a polishing step to remove the dissolved carboxylic acids from the gasoline [2]. The Fischer–Tropsch gasoline was of too poor quality to be considered for aviation-gasoline. Different refining strategies were evaluated at that time to improve the quality of the gasoline from the German Fischer–Tropsch processes. The studies focused on the heavy naphtha fraction, because the lighter boiling fraction inherently was of a better quality (for an explanation see Chapter 13). In fact, the heavy naphtha was sometimes employed as a blending component for diesel fuel (Section 6.4.3) in order to avoid degrading the quality of the gasoline. Universal Oil Products (UOP) investigated the upgrading of the syncrude from a German normal-pressure process [17]. The origin of the Fischer–Tropsch syncrude was not disclosed, but at that stage only normal-pressure facilities were in operation. The light naphtha fraction (C5 , 110 ◦ C) was removed by distillation to leave the heavier C8 –C11 naphtha for further upgrading. The C8 –C11 naphtha had an MON of 40 and density of 732 kg·m−3 and could clearly benefit from further refining. It was thermally reformed (catalytic reforming had not yet been invented) at 550 ◦ C and 5 MPa to increase its octane number to 57–62 depending on the residence time. Reforming was accompanied by cracking, and the reformate yield was 69–74 vol% of the feed. The production of ‘‘polymer gasoline’’ from the lighter alkenes by oligomerization over solid phosphoric acids was also reported (Section 19.3.1). Thermal cracking at temperatures above 650–700 ◦ C by the true-vapor-phase (T-V-P) process was evaluated for the conversion of the heavy naphtha into a more olefinic, higher octane
131
132
6 German Fischer–Tropsch Facilities
number product [18]. The design of the T-V-P process is such that it uses thermal cracking to produce light alkenes, but then allows the alkenes to react by thermal oligomerization (Section 19.3.6) to produce a heavier, higher octane number gasoline. This allowed the C3 –C4 fraction from Fischer–Tropsch synthesis to be co-fed to the T-V-P process, in which the C3 –C4 alkenes were converted to heavier olefinic products during thermal oligomerization of the process. The final gasoline blend between carbon gasoline and gasoline from T-V-P process cracking had an MON (before TEL addition) in the range 63–70 depending on the ratio of the blend. Despite the demonstrated ability of this process to upgrade naphtha, the cracking plant that was installed at the Ruhrchemie facility was not used for this purpose [2]. Hot clay treating of the naphtha was found to increase the octane number, and an increase of up to 20 units could be achieved by Ruhrchemie [3]. This improvement was obtained by double bond isomerization of the 1-alkenes to internal alkenes (Section 16.4.1), a process that is especially well suited for the upgrading of Fischer–Tropsch-derived olefinic naphtha. 6.4.3 Refining of Condensate Oil
The condensate oil contained little naphtha (Table 6.3), and in this respect the stepwise cooling of the Fischer–Tropsch products resulted in useful pre-fractionation of the syncrude. The distillate was recovered from the heavier wax fraction by atmospheric distillation (Figure 6.5). Three products were obtained: Kogasin I, a kerosene fraction (180–230 ◦ C); Kogasin II, a distillate fraction (230–320 ◦ C); and slack wax, a waxy paraffinic bottom product from atmospheric distillation. Oxygenate partitioning between the condensate oil and aqueous product was not at equilibrium. A caustic wash step was included to neutralize (Equation 6.3) and remove carboxylic acids from the condensate oil before atmospheric distillation. R – COOH + NaOH(aq) → R – COO− Na+ (aq) + H2 O
(6.3)
The cutpoint of the Kogasin II fraction was typical of a light diesel fuel (equivalent to the US D-1 diesel fuel). By operating the boiler of the atmospheric distillation column at a temperature close to 320 ◦ C, problems with thermal cracking of the reactive Fischer–Tropsch syncrude in the boiler was avoided. Two types of diesel fuel were prepared from the Fischer–Tropsch synthesis, namely, a light and a heavy diesel fuel (Table 6.10) [2, 3, 19]. The light diesel fuel was a mixture of heavy naphtha (Section 6.4.2) and Kogasin I and was used as a winter diesel fuel. The heavy diesel was a mixture of Kogasin I and Kogasin II and was used as a summer diesel fuel. The heavy diesel fuel contained 2% oxygenates, mainly alcohols, carbonyls, and carboxylic acids, with some esters and phenolic compounds [19]. It was found that after removal of the polar compounds the cetane number of the heavy Fischer–Tropsch diesel fuel increased from 80 to 88 [19]. This is likely due to the removal of the phenolic compounds. Phenolic compounds are known to improve storage stability by inhibiting oxidation. The phenolic compounds form stable free radical species to prevent the propagation of free radicals [20], and in the same way it inhibits compression ignition (for an explanation see Chapter 15). Owing to the high n-alkane content of the diesel fuel, it generally had a high cetane number. Under laboratory test conditions, the cetane numbers obtained with the
6.4 Fischer–Tropsch Refining Selected properties of light and heavy diesel fuel from German Fischer–Tropsch synthesis and the German ‘‘Sonder Diesel Kraftstoff’’ (SDK) specifications in the 1940s.
Table 6.10
Fuel property
Light diesel fuel
Distillation range (◦ C) Cetane number Density at 15 ◦ C (kg·m−3 ) Flash point (◦ C) Cloud point (◦ C) Pour point (◦ C) Viscosity at 40 ◦ C (cSt) Alkene content (g Br/100 g)
155–250 75–78 743–749 27–49 – <–37 – –
Heavy diesel fuel
165–255a >70 745 >57 – −38 – –
195–310b 80 772c 78 0 −1 2.1d 6.9
SDK specification
200–300a 85 762 75 – −9 – –
– 45 min. 810–865 55 min. –10 max. –30 max. – –
a Light
diesel (1943) and heavy diesel (1938) from the Ruhrchemie facility. ` Heavy diesel fuel from German medium-pressure Fischer–Tropsch process operated under license by Carrieres Kuhlmann in Harnes, France. c Density reported at 20 ◦ C is 768 kg·m−3 . d Source value in Saybolt Second Units (SSU) at 100 ◦ F: 1 cSt = 1 mm2 ·s−1 = [0.0022·(SSU) − 1.8/(SSU)]·100. b
Table 6.11 Comparison of distillate fractions obtained during laboratory evaluations of the Co–ThO2 –MgO–kieselguhr catalyst at 175–200 ◦ C and H2 :CO ratio of 2 : 1 for the normal- (near atmospheric) and medium-pressure (0.7 MPa) processes.
Fuel property
Cetane number Density at 15 ◦ C (kg·m−3 ) Pour point (◦ C) Alkene content (vol%)
Normal-pressure process
Medium-pressure process
C11 –C18
C12 –C19
C11 –C18
C12 –C19
100 760 −18 15
100 766 −9 13
100 760 −7 10
105 766 −2 8
Co–ThO2 –MgO–kieselguhr catalyst was even higher and the main difference between the products from normal- and medium-pressure syntheses was the cold-flow properties (Table 6.11) [13]. As in the case of gasoline, the diesel fuel produced by Fischer–Tropsch synthesis fell short of the German diesel fuel specifications of that time. Compared to a 47 cetane number crude-oil-derived diesel fuel, the Fischer–Tropsch-derived diesel fuel led to a 5% higher volumetric fuel consumption and had a 25% higher exhaust gas temperature [3]. The Co-LTFT-derived diesel fuel on its own was not considered a good diesel fuel [2]. The Fischer–Tropsch material was blended with crude oil or coal-derived distillates. Such blends typically contained around 40–45% Co-LTFT material.
133
134
6 German Fischer–Tropsch Facilities
In some of the German Fischer–Tropsch facilities, Kogasin II was also employed as a feed material for the production of lubricating oil [21]. Converting this material into lubricating oil was not just a war-time expedient, but it was considered to be its proper use [6]. Two processes were primarily applied to convert the Kogasin II into feed material for lubricating oil production, namely, thermal cracking and chlorination. Thermal cracking was performed at the Ruhrchemie facility. The Kogasin II was cracked in a Dubbs thermal cracking unit at 500–520 ◦ C and 0.4 MPa. Steam was passed into the cracking unit after heating and before it entered the expansion chamber, to limit thermal oligomerization. The alkene-rich product was then dried and oligomerized with AlCl3 in a batch reactor. The quality of the product was controlled by gradually increasing the oligomerization temperature. Spindle oil was produced by oligomerization at 40–100 ◦ C and aviation bright stock was produced by oligomerization at 15–60 ◦ C (Table 6.12) [21]. These oils were recovered after posttreatment by vacuum distillation. The Kogasin II from the German normal-pressure process was preferred for the production of lubricating oil, because the Kogasin II from the medium-pressure process contained more branched material [21]. It was also pointed out that the Kogasin II had to be treated before cracking to remove traces of Co derived from the Fischer–Tropsch catalyst, because the Co catalyzed undesirable side reactions [2]. At the Rheinpreussen facility, the Kogasin II was batch-wise chlorinated at 80–120 ◦ C at near atmospheric pressure. The chlorinated oil was then used as an alkylating agent for the Friedel–Crafts alkylation of naphthalene. Aromatic alkylation was carried out in a batch reactor at 70–100 ◦ C using AlCl3 as the catalyst. After posttreatment, different oil fractions were recovered by vacuum distillation (Table 6.12) [21]. Higher viscosity index (VI) oils were reportedly prepared by alkylation with chlorinated paraffins having an average carbon number of C16 [3]. These oils had a VI of 105 and pour point of −7 ◦ C. Lubricant base oils were also prepared from thermally cracked Fischer–Tropsch products in Japan. The properties were very similar, with a VI of 101, density of 865 kg·m−3 , and pour point of −22 ◦ C [22]. Table 6.12 Properties of lubricating oils prepared from Kogasin II by thermal cracking followed by AlCl3 -catalyzed oligomerization and by chlorination followed by AlCl3 -catalyzed alkylation of naphthalene.
Property
Density at 20 ◦ C (kg·m−3 ) Pour point (◦ C) Flash point (◦ C) Viscosity index Viscosity (cSt)a At 38 ◦ C (100 ◦ F) At 50 ◦ C (122 ◦ F) At 99 ◦ C (210 ◦ F) a Source
Cracking–oligomerization
Chlorination–alkylation
Spindle oil
Aviation bright stock
Spindle oil
Turbine oil
Cylinder oil
845 −51 193 –
865 −26 321 ∼110
901 −4 171 53
928 −26 202 49
965 −4 274 61
– 15 –
– 290 –
16 – 3
45 – 5
52 – 10
values in Saybolt Second Units (SSU) – see Table 6.9 for conversion.
6.4 Fischer–Tropsch Refining
In the Ruhrchemie facility, some of the alkenes produced by cracking were converted by the ‘‘OXO’’ process to alcohols [2]. The process entails hydroformylation of alkenes obtained by cracking with H2 and CO to produce aldehydes (Equation 6.4), which is then hydrogenated to alcohols (Equation 6.5). R–CH=CH2 + H2 + CO → R–CH2 –CH2 –CH=O
(6.4)
R–CH2 –CH2 –CH=O + H2 → R–CH2 –CH2 –CH2 –OH
(6.5) ◦
In the hydroformylation step, the alkenes were converted to aldehydes at 135–150 C and 15–20 MPa using the standard Fischer–Tropsch catalyst, but without MgO. The alkenes were pre-fractionated into different carbon number cuts (C11 –C12 , C13 –C14 , C15 –C16 , and C17 ) that were converted separately to enable separation of the alkenes and aldehydes after reaction by distillation. The hydrogenation step could also be performed with the Fischer–Tropsch catalyst, but a cheaper Ni-based hydrogenation catalyst was preferred. Some of the distillate range material was sold and was employed as feed for the production of ‘‘mersol’’ detergents. The detergents were obtained by light-induced free radical sulfochlorination (Equation 6.6) followed by saponification [6, 9]. R–H + SO2 + Cl2 → R–SO2 Cl + HCl
(6.6)
With Kogasin II feed, it was also possible to carry out sulfochlorination in the dark [9]. The Kogasin II apparently contained sufficient hydroperoxides to initiate the free radical reaction. Hydroperoxide formation does not take place during Fischer–Tropsch synthesis, but it is a natural consequence of exposing the reactive syncrude to air during transport and storage. Other uses and applications of distillate range Fischer–Tropsch products that have been considered in Germany at that time, have been discussed by Freerks [3]. 6.4.4 Refining of Waxes
The major portion of the waxes produced during normal-pressure Fischer–Tropsch synthesis was obtained by the periodic removal of waxes from the catalyst by solvent washing. In the medium-pressure process, this wax fraction was removed because of the solvent action of the product that was present at high pressure. Slack wax was also recovered by atmospheric distillation of the condensate oil. The heavy wax product was separated by steam stripping, and the wax from the lighter fractions was recovered by sweating. Typical properties of Fischer–Tropsch waxes thus obtained are given in Table 6.13 [23]. Detailed analysis of the Fischer–Tropsch waxes revealed that the hard wax contained n-alkanes with carbon numbers over C150 and a melting point of about 117 ◦ C [24]. The slack wax fraction from atmospheric distillation and the waxy paraffins (Gatsch) that remain after sweating to recover the heavier waxes, served as feed materials for the production of chemicals. These conversions were not necessarily performed in the Fischer–Tropsch refinery. The material was mainly used for two processes, namely, cracking and oxidation. Cracking of the slack wax fraction was performed in a way similar to Kogasin II cracking (Section 6.4.3). The aim was to produce lighter linear 1-olefins for the manufacture of lubricants and detergents and oxidation. For example, slack wax was sold and converted in the I.G.-P¨olitz plant to a lubricating oil by thermal cracking and AlCl3 oligomerization [21].
135
136
6 German Fischer–Tropsch Facilities Table 6.13 Properties of different waxes obtained from Fischer–Tropsch synthesis at the Ruhrchemie facility. Catalyst wax is from normal-pressure synthesis. The other waxes were obtained by distillation and/or sweating of product from medium-pressure synthesis.
Properties Congealing point (◦ C) Clear melting point (◦ C) Density at 15 ◦ C (kg·m−3 ) Mean molecular mass (g·mol−1 ) Acid content (mg KOH/g) Alkene content (g I/100 g)
Catalyst wax
Soft wax
Medium wax (Block wax)
Hard wax
87–91 – 904 ∼550 – 3.5
42.5 44 – – 0.14 –
50–52 53 899 380 0.03 2.5
90–93 110 921 600 0.1 2.0
One of the important uses of the 320–460 ◦ C slack wax fraction was as a crude paraffin feed for the production of fatty acids by atmospheric oxidation [25]. The fatty acids were used as raw materials for the production of other products such as soaps and edible fats [2, 6]. Oxidation was performed in a special wax oxidation kettle that was constructed using aluminium with an alloy steel head to limit oxidation by the corrosive short-chain carboxylic acids. Manganese (such as KMnO4 ) and other metal carboxylates were employed as oxidation catalysts and oxidation was carried out at 140–160 ◦ C with air as the oxidant. The heavier waxes were refined by bleaching and activated carbon absorption to remove color components. The application of these waxes, which were substitutes for ceresin waxes (Tm = 68–72 ◦ C, ρ = 920–940 kg·m−3 [26]), were dependent on their congealing point. Typical applications included the manufacturing of candles, insulation materials, and polishes. Part of the heavier wax fractions was also oxidized to produce emulsifiable waxes [6]. The aim of this type of oxidation is to incorporate oxygenate functionality into the wax without degrading the chain. Autoxidixation without chain degradation requires less severe conditions than are required for oxidation of slack waxes to produce fatty acids. Autoxidation of waxes and lighter alkanes is discussed later in more detail (Section 23.3). 6.4.5 Aqueous Product Refining
The German Fischer–Tropsch processes did not produce much water-soluble oxygenates. At the Ruhrchemie facility, the amount of water-soluble light alcohols produced was 1.16 g/kg of C3 and heavier products. The most abundant oxygenate class in the aqueous product was the alcohols (Table 6.14) [13]. In addition to the alcohols, carboxylic acids, aldehydes, and esters were also found in the aqueous product, but in smaller amounts. The aqueous product was refined only in the Hoesch facility. The combined aqueous product from steam desorption of the activated carbon gasoline and wash coolers was distilled to condense the oil. This concentrated the organic material from 0.5 mass% in the aqueous product to 70% in the overhead products from the column, of which 8–10% was nonalcohol oxygenates.
6.5 Discussion of the Refinery Design Alcohol composition of the Fischer–Tropsch aqueous products from the Ruhrchemie and Hoesch facilities.
Table 6.14
Compound
Methanol Ethanol Propanols Butanols C5 –C6 alcohols a C –C 4 6
Alcohol composition (%) Ruhrchemie
Hoesch
10 15 20 40a –
10 25 30 25 10
alcohols.
The alcohol thus recovered was about 0.6% of the total syncrude and about 45% of the total organic matter in the aqueous product [13]. The carboxylic acids were also recovered at the Hoesch facility. The neutralized carboxylic acids obtained from the oil condensate and from the bottom of the alcohol recovery columns were acidified with H2 SO4 to regenerate the carboxylic acids, and were extracted with a light aromatic solvent. The carboxylic acid mixture consisted of 30% C2 –C4 , 30% C5 –C9 , 35% C10 –C20 , and 5% heavier carboxylic acids. This constituted a recovery of 0.4% of the total syncrude and about 30% of the total organic matter in the aqueous product [13].
6.5 Discussion of the Refinery Design
The diversity in the German Fischer–Tropsch refinery designs make it quite clear that there are many ways to refine syncrude. Some design decisions make sense only in historic context. The refinery designs must be evaluated in terms of the product specifications at that time. One should also bear in mind that in the late 1930s very few catalytic processes were available (Table 2.2). As a consequence, a fair number of the refining technologies that were selected for upgrading the Fischer–Tropsch syncrude employed free radical mechanisms. Despite the limitations imposed by the technology at that time, there are valuable lessons to be learnt from these designs, which are also applicable to modern Fischer–Tropsch refinery design [27]. 1) The refineries were designed to produce transportation fuels, chemicals, and lubricating oils. The ratio of liquid fuels to chemicals and speciality products were 72 : 28 [6]. This resulted in considerable product diversity. The value of syncrude as a clean raw material for chemical production was realized, and many of the senior German officials involved with Fischer–Tropsch technology were at pains to point out that syncrude has far more potential for chemical production than for fuels [2]. 2) Alkenes are valuable. The alkenes were important for gasoline quality, chemical manufacturing and for the production of lubricating oils. It was not that alkanes were not
137
138
6 German Fischer–Tropsch Facilities
3)
4)
5)
6)
7)
8)
9)
10)
useful, but apart from the heavy waxes, much effort was expended on activating the alkanes by processes such as cracking and chlorination to enable transformations that could be performed directly with alkenes. The quality of the fuels produced from Co-LTFT synthesis was generally poor and failed to meet the German fuel specifications. The naphtha and distillate fractions may have been referred to as gasoline and diesel fuel, but were actually only blending components for final fuels. The coal liquids coproduced in the coke-ovens during syngas production was quite valuable in providing an aromatic blending stock to upgrade the Fischer–Tropsch liquids. There was synergy between the coal liquids and Co-LTFT syncrude for fuels and lubricating oil production. The German normal-pressure process with its lighter product (lower α-value) and higher alkene content was better suited for fuels production than the medium-pressure process. About 75% of the normal-pressure Co-LTFT syncrude was naphtha and distillate, and an even higher liquid fuel yield (>80%) was possible by conversion of the C3 –C4 olefins. The stepwise cooling of the syncrude as the first refining step pre-fractionated the syncrude in a sensible way that simplified refinery design. Intermediate syncrude recovery also enabled the use of Fischer–Tropsch reactors in series, which improved the syngas conversion and allowed a less complex gas loop design. When comparing the Fischer–Tropsch refinery with a generic second-generation crude oil refinery (Section 2.4.2) of that time, different refining technologies were important. In the refining of Fischer–Tropsch syncrude, the key technologies were oligomerization, aromatic alkylation, hydration, and cracking. Of these, only oligomerization and cracking were employed in more advanced conventional crude oil refineries, such as those that produced aviation-gasoline. The refinery design had to make provision for dealing with oxygenates. Oxygenates were also considered valuable and oxygenate production was increased by processes such as oxidation, hydration, and hydroformylation. However, despite their value, much care had to be taken in processing carboxylic acid containing streams. The carboxylic acids also dictated material selection for process equipment. Metals and metal carboxylates from Fischer–Tropsch catalyst degradation ended up in refinery streams. In processes where such materials were detrimental, specific provision had to be made for pretreating the feed. Although the Fischer–Tropsch aqueous product contained <1% of the syncrude, part of the oxygenates, including carboxylic acids, was recovered in one of the refineries.
References 1. Stranges, A.N. (2007) A history of the
3. Freerks, R. (2003) Early efforts to upgrade
Fischer-Tropsch synthesis in Germany 1926−45. Stud. Surf. Sci. Catal., 163, 1–27. 2. Weil, B.H. and Lane, J.C. (1949) The Technology of the Fischer-Tropsch Process, Constable, London.
Fischer-Tropsch reaction products into fuels, lubricants and useful materials. AIChE Spring National Meeting, New Orleans, 2 April, Paper 86d. 4. U. S. Navy Technical Mission in Europe (1945) The synthesis of hydrocarbons and chemicals
References
5.
6.
7.
8.
9. 10.
11.
12.
13.
14.
15. 16.
from CO and H2 . Technical Report No. 248-45, September 1945. B¨ussemeier, B., Frohning, C.D., and Cornils, B. (1976) Lower olefins via Fischer-Tropsch. Hydrocarbon Process., 55 (11), 105–112. Hall, C.C. (1947) The production of oil from coal by the Fischer-Tropsch process. J. Junior Inst. Eng., 57, 309–334. Komarewsky, V.I., Riesz, C.H., and Estes, F.L. (1945) The Fischer-Tropsch Process. An Annotated Bibliography, Institute of Gas Technology, Chicago. Gollmar, H.A. (1945) in Chemistry of Coal Utilization (ed. H.H. Lowry), John Wiley & Sons, Inc., New York, pp. 947–1007. Asinger, F. (1968) Paraffins Chemistry and Technology, Pergamon, Oxford. Glowacki, W.L. (1945) in Chemistry of Coal Utilization (ed. H.H. Lowry), John Wiley & Sons, Inc., New York, pp. 1136–1231. Powell, A.R. (1945) in Chemistry of Coal Utilization (ed. H.H. Lowry), John Wiley & Sons, Inc., New York, pp. 921–946. Casci, J.L., Lok, C.M., and Shannon, M.D. (2009) Fischer-Tropsch catalysis: the basis for an emerging industry with origins in the early 20th century. Catal. Today, 145, 38–44. Storch, H.H., Golumbic, N., and Anderson, R.B. (1951) The Fischer-Tropsch and Related Syntheses, John Wiley & Sons, Inc., New York. Friedel, R.A. and Anderson, R.B. (1950) Composition of synthetic liquid fuels. I. Product distribution and analysis of C5 –C8 paraffin isomers from cobalt catalyst. J. Am. Chem. Soc., 72, 1212–1215. Asinger, F. (1968) Mono-Olefins Chemistry and Technology, Pergamon, Oxford. Fielding, J.C. (1973) in Propylene and Its Industrial Derivatives (ed. E.G. Hancock), Ernest Benn, London, pp. 214–272.
17. Egloff, G., Nelson, E.S., and Morrell, J.C. (1937)
18.
19.
20. 21. 22.
23.
24.
25.
26.
27.
Motor fuel from oil cracking production by the catalytic water gas reaction. Ind. Eng. Chem., 29, 555–559. Snodgrass, C.S. and Perrin, M. (1938) The production of Fischer-Tropsch coal spirit and its improvement by cracking. J. Inst. Pet. Technol., 24, 289–301. Ward, C.C., Schwartz, F.G., and Adams, N.G. (1951) Composition of Fischer-Tropsch diesel fuel. Cobalt catalyst. Ind. Eng. Chem., 43, 1117–1119. Scott, G. (1965) Atmospheric Oxidation and Antioxidants, Elsevier, Amsterdam. Horne, W.A. (1950) Review of German synthetic lubricants. Ind. Eng. Chem., 42, 2428–2436. Givens, E.N., LeViness, S.C., and Davis, B.H. (2007) Synthetic lubricants: Advances in Japan up to 1945 based on Fischer-Tropsch derived liquids. Stud. Surf. Sci. Catal., 163, 29–36. Gall, D. (1947) The characteristics of Fischer-Tropsch waxes. Inst. Petrol. Rev., 1, 336–338. Koch, H. and Ibing, G. (1943) Composition of paraffin waxes from the Fischer-Tropsch water-gas synthesis. Pet. Refiner, 22 (9), 89–96. Withers, J.G., West, H.L., and Ruhr-Chemie, A.G. (1946) Sterkrade Holten Interrogation of Dr. O. Roelen at Wimbleton November 15th and December 20th 1945. British Intelligence Objective Sub-Committee Target No. 30/5.01, 1946. Hawley, G.G. (1971) The Condensed Chemical Dictionary, 8th edn, Van Nostrand Reinhold, New York. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 331–364.
139
141
7 American Hydrocol Facility 7.1 Introduction
Subsequent to the Second World War, much technical information about the German Fischer–Tropsch technology (Chapter 6) became available in the United States. During this period Hydrocarbon Research Inc. developed an American version of the Fischer–Tropsch process, called the Hydrocol process [1]. The Hydrocol process used an iron-based Fischer–Tropsch catalyst in a fixed fluidized bed reactor. The process was piloted at Olean, New York. This work led to the construction of the first industrial-scale Fischer–Tropsch facility in the United States by the Carthage Hydrocol Company in Brownsville, Texas. It was the first industrial application of iron-based high-temperature Fischer–Tropsch (Fe-HTFT) synthesis. It was designed to convert 75 500 m3 ·h−1 natural gas into 38 m3 ·h−1 (5800 bbl/day) motor-gasoline, 8 m3 ·h−1 (1200 bbl/day) diesel fuel, and 2.8 t·h−1 (150 000 lb/day) of oxygenates [1]. The Hydrocol plant was in commercial operation during the period 1951–1957, and was shut down mainly for economical reasons. A second Hydrocol-type facility was planned by the Standolind Oil and Gas Company and it would have been constructed in Kansas, but it was never built [2–4]. Around the same time Philips Petroleum Company started the development of an iron-based fluidized bed Fischer–Tropsch process. This process was piloted at their research facilities in Bartlesville, Oklahoma, but the process was never commercialized [5, 6]. Various other companies also conducted research programs at that time and some are still active in the field of Fischer–Tropsch research [7]. Apart from the commercial interest in Fischer–Tropsch technology in the United States, the strategic value of this technology was recognized at the government level. Prior to the Second World War, Fischer–Tropsch research was funded by the US government. This process was accelerated on 5 April 1944 when the ‘‘Synthetic Liquids Fuels Act’’ was passed by the United States Congress. The US Bureau of Mines was tasked by the United States Congress to find an alternative to crude oil as a source for transportation fuel [3]. Fischer–Tropsch and direct coal liquefaction were investigated in parallel, since it was unclear which of these technologies were the most economical. Demonstration-scale operation was terminated in 1953 after the discovery of substantial new oil reserves in the United States, Canada, and Middle East. Pilot plant studies were continued at the Pittsburgh Energy Technology Centre and a number of different Fischer–Tropsch technologies were investigated [8]: Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
142
7 American Hydrocol Facility
1) Oil circulation process. This was an expanded fixed-bed low-temperature Fischer–Tropsch (LTFT) technology that used a fused iron catalyst. By the circulation of a part of the liquid product from Fischer–Tropsch synthesis in an up-flow mode, catalyst agglomeration was avoided and heat transfer was improved. 2) Hot gas recycle process. This work followed on the work performed with fixed-bed German Fischer–Tropsch synthesis, where the value of increasing the recycle was demonstrated (Section 6.3.3). This was developed further as a fixed-bed Fe-HTFT process. 3) Slurry bed process. Improved heat management by the liquid reactor product, as was employed in the oil circulation process, was taken one step further. Smaller iron catalyst particles were suspended in liquid product through which syngas was bubbled. It was an Fe-LTFT process. 4) Fluidized bed process. A nitrided iron catalyst was used in a fluidized bed under typical LTFT conditions (238–253 ◦ C) to produce a lighter oxygenate-rich syncrude. Although the developments in the United States that took place in parallel with the construction and operation of the Hydrocol facility did not directly impact the Hydrocol design, it highlights the issues that were being grappled with at that time. Most important was the complete shift from Co-based Fischer–Tropsch synthesis to Fe-based Fischer–Tropsch synthesis. The rest of the chapter deals exclusively with the industrial Hydrocol facility.
7.2 Synthesis Gas Production
The Hydrocol plant made use of natural gas as feedstock, and was the first industrial gas-to-liquids (GTL) facility based on Fischer–Tropsch synthesis (Figure 7.1) [1, 9].
Air
Air separation
N2
O2 Steam
Partial oxidation
Syngas
Gas scrubber
Oil scrubber Tail gas
Internal recycle Gas
Oil
Oil HTFT
Reaction water and dissolved oxygenates Natural gas Figure 7.1
Natural gas liquids
Aqueous product Natural gasoline
Hydrocol synthesis gas generation and Fischer–Tropsch gas loop.
7.3 Fischer–Tropsch Synthesis
The first step in the process was to separate the condensable liquids from the natural gas. The natural gas liquids served as a blending material for motor-gasoline production (Section 7.4.1). The natural gas was converted into synthesis gas by noncatalytic partial oxidation of the methane with pure O2 . The oxidative reforming of the natural gas was performed at 2 MPa in two 56 m3 vessels lined with refractory bricks. The space velocity was equivalent to 110 kg·m−3 reformer volume per hour [10]. Steam addition was regulated to prevent soot formation. The synthesis gas thus produced had an H2 :CO ratio of approximately 2 : 1 [11]. The air separation unit was the largest in the world at that time. The natural gas contains low levels of sulfur (Table 3.1). Yet, despite the presence of low levels of sulfur in the syngas, the syngas employed for Hydrocol synthesis was not purified [1, 4]. The syngas was not conditioned either, although this was not necessary considering its H2 :CO ratio of close to 2:1, and the syngas was supplied to the Fischer–Tropsch section as is. The Hydrocol Fischer–Tropsch technology made these shortcuts in syngas purification and conditioning possible.
7.3 Fischer–Tropsch Synthesis
The Fischer–Tropsch catalyst used in the Hydrocol process was a fused iron catalyst that was also used for ammonia synthesis. It consisted of 97% Fe3 O4 (magnetite), 2.5% Al2 O3 , and 0.5% K2 O. This catalyst was later replaced with a natural magnetite containing 0.5% K2 O. The fused catalyst was ground to a fine powder (0.045–0.45 mm) and was completely reduced at 350–460 ◦ C in hydrogen before use [10]. The Hydrocol technology is classified as a high-temperature Fischer–Tropsch (HTFT) process. The Hydrocol plant used two 5 m diameter fixed fluidized bed synthesis reactors with a height of 18 m [12, 13]. Each reactor was loaded with about 200 tons of iron catalyst. A heat exchanger was suspended in the bed to control the temperature by producing 2.1–2.8 MPa steam. The surface of the fluidized catalyst bed remained fairly smooth during operation, although local eruption of bubbles could reportedly be observed. Gas–solid separation was efficient and little catalyst was carried over with the product gas. The reactor was operated in the range 305–345 ◦ C and 2.7 MPa, with a linear gas velocity of 0.2 m·s−1 . The fresh syngas feed gas hourly space velocity was 2000–3000 h−1 and a syngas conversion of 90–96% was achieved [12]. About 0.15 kg C3 and heavier hydrocarbons were produced per 1 m3 of synthesis gas converted. The overall reactor productivity was around 180 kg·h−1 /m3 of catalyst [10]. The syncrude from Hydrocol synthesis consisted mainly of C3 –C4 and naphtha range products (Table 7.1) [9, 11]. The production of heavy products had to be avoided, since heavy products would cause condensation, catalyst agglomeration, and fluidized bed slumping in the synthesis reactor. This is a general restriction that is imposed on Fischer–Tropsch synthesis conducted in fluidized bed reactors. The α-value required for HTFT synthesis in fluidized bed reactors is <0.7. One of the advantages of fixed fluidized bed operation was the ability to add and remove catalyst while in operation. The ability to replace catalyst while the reactor was on-stream enabled
143
144
7 American Hydrocol Facility Table 7.1 Syncrude composition of the Hydrocol process, excluding unrecovered C1 –C2 hydrocarbons in the tail gas.
Syncrude fraction
LPG (C3 –C4 ) Naphtha (C5 -204 ◦ C) Distillate Residue Aqueous producta
Syncrude composition (mass%) Recovered
Oil product
26 45 6 3 20
32 56 8 4 –
a Source reported aqueous products as 21 749 000 lb a−1 /1000 bbl/day hydrocarbons.
the omission of a sulfur removal step after syngas generation (Section 7.2). In actual industrial operation, the ability to exchange catalyst while in operation proved to be difficult. The Fischer–Tropsch gas loop (Figure 7.1) included an internal recycle. The hot vapor-phase product from Fischer–Tropsch synthesis was cooled down and separated in a product separator drum into an aqueous product phase, organic (oil) product phase, and tail gas. The oil product and tail gas were subsequently water washed in direct contact scrubbers. In the gas scrubber, the tail gas was water washed to remove residual light oxygenates. The washed tail gas contained mainly CO2 , some unconverted syngas, and light hydrocarbons. The Hydrocol facility employed a closed gas loop design and part of the washed tail gas was recycled to the Fischer–Tropsch reactor as an internal recycle. The internal recycle to fresh syngas feed ratio was about 1.5 : 1 [12]. The gas loop did not include cryogenic separation, and recovery of light hydrocarbons was restricted to pressure separation after syncrude cooling. In the oil scrubber, the organic product was water washed to remove light oxygenates and specifically carboxylic acids. It was pointed out that corrosion-resistant materials were required in many parts of the process and that hot processing of carboxylic acid containing streams was best performed in stainless steel [9]. The wash water from the gas scrubber and the oil scrubber was combined with the Fischer–Tropsch aqueous product from the primary product separator drum. The combined aqueous product flow was equivalent to 2.4 m3 /m3 oil and the aqueous product contained 5–10% dissolved oxygenates [9]. The water-soluble oxygenates in the aqueous product were mainly ethanol and acetic acid, with lesser amounts of acetone, ethanal (acetaldehyde), and higher alcohols being produced. The oil product produced by the Hydrocol process was rich in linear 1-alkenes. The syncrude had a low degree of branching (Table 7.2) [10, 11], with the methyl branched hydrocarbons being mainly 2-methyl isomers [9]. In addition to the hydrocarbons in the oil product it also contained oxygenates and aromatics (Table 7.3) [10, 14]. The heavier fraction contained a substantial amount of aromatics and was not wax-like as in the case of LTFT synthesis. Up to 30% aromatic was found in the C17 and heavier Hydrocol syncrude [10].
7.4 Fischer–Tropsch Refining Table 7.2
Carbon number C4 C5 C6 C7 C8
Table 7.3
Carbon number
C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 a Oxygenates
Degree of branching in the naphtha fraction of Hydrocol syncrude. Linear
Methyl branched
Dimethyl branched
Cyclic
89.4 80.9 75.9 60.2 55.4
10.6 18.8 20.0 29.3 36.6
– 0.0 0.4 1.7 2.4
0.0 0.3 3.7 8.8 5.6
Distribution of compound classes in the C3 –C12 oil fraction of Hydrocol syncrude. Boiling fraction (◦ C)
– – 15–40 40–75 75–104 104–138 138–162 162–187 187–200 200–220
Composition by carbon number (mass%) Alkanes
Alkenes
Aromatics
Oxygenatesa
20.2 15.5 15.4 13.2 10.5 9.6 1.8 6.0 9.3 5.4
79.8 84.5 81.8 80 80.2 76.2 79 79.3 74 76.5
– – 0 0.2 2.1 6.0 6.2 7.2 5.8 3.6
– – 2.8 6.6 7.2 8.2 13 7.5 10.9 14.5
were calculated by difference.
7.4 Fischer–Tropsch Refining
The aim of the Hydrocol process was to produce motor-gasoline with a better than 80% yield from the oil product. This was an important statement of intent. The product distribution of the oil product (Table 7.1) is such that the Fischer–Tropsch refinery required conversion processes that would increase the amount of material in the naphtha boiling range. The refinery also required conversion processes that would upgrade the naphtha to meet the motor-gasoline specifications of the 1950s. A typical regular motor-gasoline in that period had a research octane number (RON) of 87 and a motor octane number (MON) of 83 after the addition of tetraethyl lead (TEL). The amount of TEL that could be added was limited to 1.3 kg·m−3 (3 ml·gal−1 ). Many of the properties were not regulated and the specifications were only suggestions [15].
145
146
7 American Hydrocol Facility
Bauxite treatment
Debutanizer
C3 −C4
Oligomerization (SPA)
LPG
Polymer gasoline Oil
Gasoline Atmospheric distillation
Diesel fuel
Natural gasoline Oxygenates (chemicals)
Aqueous product
Acid water Figure 7.2
Hydrocol refinery design.
In line with the uncomplicated design of the syngas production section and Fischer–Tropsch gas loop, the refinery design has little complexity (Figure 7.2). Yet, the design reflected much thought. Apart from normal oil product separation, the refinery included conversion and separation steps that addressed issues specific to the Hydrocol process: 1) Removal of the unwanted oxygenates from the organic product, especially the carboxylic acids that were known to cause problems [16]. 2) Increase of the octane number of the straight-run naphtha (RON = 62, MON = 68 [11]) to a higher value, so that the required octane number can be achieved by TEL addition. 3) Conversion of the gaseous alkenes (C3 –C4 ) into liquid products, since it constituted about one-third of the oil product from Hydrocol synthesis. The conversion was such that most of the product from C3 –C4 conversion boiled in the naphtha range. 4) Recovery of the alcohols and other valuable oxygenates dissolved in the aqueous product to be sold as chemicals [9]. 7.4.1 Oil Product Refining
The Hydrocol refinery design (Figure 7.2) departed significantly from conventional crude oil refinery design. Refining did not start by atmospheric distillation. It was recognized that the oxygenates in the Fischer–Tropsch syncrude co-boils with hydrocarbons that are two to four carbon numbers heavier than the oxygenates. Any refinery conversion that converts oxygenates into hydrocarbons will cause a change in the boiling point of the converted oxygenates by upward of 50 ◦ C. In order to avoid such boiling range broadening, the refinery design included a unit to convert oxygenates into hydrocarbons before atmospheric distillation. Another objective of this design approach was to reduce the corrosivity of the syncrude by converting the carboxylic acids in the oil. The complete oil fraction was treated over bauxite at a temperature around 400 ◦ C. Bauxite is an alumina-rich natural material with mild acidity. The bauxite treatment step was a commercial
7.4 Fischer–Tropsch Refining Research octane number (RON) and motor octane number (MON) of straight-run Hydrocol naphtha (C5 -204 ◦ C) before and after bauxite treatment.
Table 7.4
Motor-gasoline blend
Naphtha Naphtha + 1 ml TEL/gala Naphtha + 3 ml TEL/gala a
RON
MON
Straight run
Bauxite treated
Straight run
Bauxite treated
68 79 84
87 93 94
62 70 74
76 80 82
Tetraethyl lead (TEL) addition: 1 ml TEL/gal = 0.44 g l−1 .
process, called the Perco-process, which was used as a sulfur removal step in oil refineries [17]. Bauxite treatment was also employed for oxygenate removal from the syncrude produced in the Philips Petroleum HTFT pilot plant [18]. Although the bauxite treatment was not able to remove all of the oxygenates, very good results were obtained with this process. All the alcohols and esters were converted to hydrocarbons (mainly alkenes), and the combined conversion of carboxylic acids and carbonyls was around 90% [16]. In addition to the conversion of most of the oxygenates into alkenes, the process also increased the octane number of the naphtha (Table 7.4) [11]. The Hydrocol syncrude is rich in alkenes, but the alkenes are mainly linear 1-alkenes that have lower octane numbers than linear internal and branched alkenes. The bauxite treatment resulted in double bond isomerization of the alkenes (Section 16.4.1), which caused an increase in the octane number of the product. Owing to the nature of the alumina-based catalyst, little cracking took place and the carbon number distribution was not changed. The product from the bauxite treatment step was stabilized in a debutanizer column. The C3 –C4 overhead product served as feed for alkene oligomerization over a solid phosphoric acid catalyst in a UOP CatPoly process. Solid phosphoric acid oligomerization (Section 19.3.1) is fairly insensitive to the feed material and operating conditions used, and produces mainly products in the naphtha range. It was therefore a good refining choice considering that the aim of the Hydrocol refinery was to produce motor-gasoline as the main product. The oligomerization was conducted at 205 ◦ C, 3.5 MPa, and liquid hourly space velocity of 1.15 h−1 . More than 90% of the alkenes in the feed was converted and good-quality olefinic ‘‘polymer gasoline’’ was obtained (RON = 95.4, MON = 82.4). The naphtha to distillate ratio of the product oligomers was 88 : 12 [11]. The stabilized oil, which was the bottom product from the debutanizer column, was separated into naphtha and distillate fractions. These fractions formed the base stock for the motor-gasoline and diesel fuel from the Hydrocol refinery. The final motor-gasoline consisted of a blend containing 64% bauxite-treated naphtha, 25% ‘‘polymer gasoline’’ from oligomerization, and 11% ‘‘natural gasoline’’ (n-butane rich) from natural gas separation (Table 7.5) [11]. In the 1950s, a typical crude-oil-derived motor-gasoline contained 20–40% alkenes and 5–30% aromatics [15]. The Hydrocol motor-gasoline was therefore very olefinic compared to the quite olefinic motor-gasolines at that time. Table 7.5 also indicates that the Hydrocol refinery is MON constrained; that is, the ability of the refinery to improve the
147
148
7 American Hydrocol Facility Table 7.5 Selected properties of the motor-gasoline obtained from the Hydrocol refinery and that of a typical 1950s US regular motor-gasoline.
Fuel property RON (clear) RON (+3 ml TEL/gal) MON (clear) MON (+3 ml TEL/gal) Density (kg·m−3 ) Reid vapor pressure (kPa) Alkene content (g Br/100 g) T90 distillation (◦ C)
Hydrocol gasolinea
US regular gasoline
91.4 97.2 80.2 84.1 717 63 118 172
– >87b – >83b 730–740 65–100 30–50 <200
a Source
indicated 11% n-butane instead of ‘‘natural gasoline.’’ Usual addition in the range of 1–2.5 ml TEL/gal to reach RON = 87 and MON = 83. b
Table 7.6 Selected properties of the light diesel fuel obtained from the Hydrocol process and that of the typical 1950s US D-1 and D-2 diesel fuels.
Fuel property
Cetane number Density at 15 ◦ C (kg·m−3 ) Pour point (◦ C) Olefin content (g Br/100 g) Distillation (◦ C) T10 T50 T90 a Light
Hydrocol diesel fuela
US diesel fuel
Straight run
Hydrogenated
D-1
D-2
45–50 806 −9 to −15 47
71 806 −1 2
40 825 <–7 –
35–40 845–865 <–7 –
204 232 304
227 260 327
215–230 – ∼270
215–230 – 305–320
diesel fuel without heavy diesel (residue) fraction.
octane number of the gasoline is limited by its ability to increase the MON. This is typical of refineries that produce motor-gasolines with a high alkene content. The distillate fraction was not upgraded any further and was directly used as diesel fuel. It has been shown that the cetane number of the diesel fuel could be improved by mild hydrogenation (Table 7.6) [19], but this was not done. The cetane number was above that of typical crude-oil-derived diesel fuels of the 1950s, and hydrogenation would have caused problems with the cold-flow properties. It should be pointed out that the diesel fuel properties in Table 7.6 are that of the light diesel fuel fraction. When the heavier material was included, the cetane number and density were increased, but the cold-flow properties become worse.
7.4 Fischer–Tropsch Refining Composition of the oxygenates in the Hydrocol aqueous product and an indication of their chemical market in the United States in 1948.
Table 7.7
Compound
Alcohols Methanol Ethanol 1-Propanol 2-Propanol (isopropanol) Butanols C5 and heavier Aldehydes Ethanal (acetaldehyde) Propanal (propionaldehyde) Butanal (butyraldehyde) Ketones Acetone 2-Butanone (MEK) 2-Pentanone 2-Hexanone Carboxylic acids Ethanoic acid (acetic) Propanoic acid (propionic) Butanoic acid (butyric) C5 and heavier a
Concentration (mass%)
Capacity (t a−1 )a
Relative to US market (%)
0.3 36.9 8.7 0.8 4.0 1.2
70 9170 2160 190 990 310
<1 5 130 <1 2 11
6.0 2.2 2.1
1500 560 520
2 – –
b
7.5 2.2 0.9 0.2
1860 540 220 60
2 5 – –
b
18.1 4.7 3.4 0.8
4490 1180 840 190
6 – – –
b
b
b b b
Capacity in ton per year (t a−1 ) based on 150 000 lb/day for 7000 bbl/day Hydrocol capacity. data for US market not available.
b Compound-specific
7.4.2 Refining Aqueous Product
The Hydrocol aqueous product consisted of the primary aqueous-phase product separated after Fischer–Tropsch synthesis, as well as the wash water from the gas and oil scrubbers (Figure 7.2). The water-soluble oxygenates found in the aqueous product contained alcohols, carbonyls, and carboxylic acids (Table 7.7) [9]. The upgrading of the aqueous product was investigated. It was soon realized that many of the carbonyls and alcohols formed azeotropes on distillation. The chemical potential of HTFT could be exploited in principle, but it required extensive investigations to be achieved in practice, mainly owing to the absence of binary and ternary phase diagrams at that time. Distillation of the aqueous product enabled recovery of about 9% of the feed material as an oxygenate-rich overhead product. At this point, recovery of the nonacidic oxygenates was almost complete, and the carboxylic acids and most of the water were recovered as bottom products. In terms of the distillation curve, the cut was made just before azeotropes of carboxylic acids and
149
150
7 American Hydrocol Facility
water appear in the overhead product. Although details of the subsequent separation steps have not been provided, some inherent difficulties with separation and the reactivity of the chemical products have been pointed out [9]. These include the oxidation of aldehydes to carboxylic acids (Equation 7.1), formation of cyclic ethers (Equation 7.2), and the reaction between aldehydes and alcohols to form acetals (Equation 7.3). R–CH=O + 1/2 O2 → R–COOH
(7.1)
3R–CH=O (–C(RH)O–)3
(7.2)
2 R–CH2 –OH + R–CH=O R–C(–O–CH2 –R)2 + H2 O
(7.3)
Problems related to oxygenate cross-contamination on the ability to achieve chemical product specifications have also been discussed [9]. These are all very real problems a refiner must overcome in order to produce on-specification chemical products from a Fischer–Tropsch aqueous product. 7.5 Discussion of the Refinery Design
The Hydrocol facility had a very specific design objective that was well met by its Fischer–Tropsch refinery design. The properties of the HTFT syncrude were exploited to its best advantage, resulting in a very efficient refinery design. Bauxite treatment and solid phosphoric acid oligomerization were the only two conversion units in the Hydrocol refinery (Figure 7.2). Yet, the Hydrocol refinery was capable of producing a similar quality motor-gasoline as the Sasol 2 and 3 designs (Chapter 9) 30 years hence. The crude oil refining approach followed for Sasol 2 and 3 did not exploit the syncrude properties and it resulted in a more complex and costly refinery design to achieve essentially the same product quality as the Hydrocol refinery [20]. Additional learning points from the Hydrocol refinery design are the following: 1) Refinery conversion was dominated by alkenes and oxygenates. The refinery design beneficially exploited these reactive compound classes to meet the refining objectives. Potential processing problems associated with oxygenates were recognized, but where possible oxygenates were used to their best advantage. 2) The success of the bauxite treatment step is notable. Bauxite is a natural alumina-rich catalyst possessing mild acidity. It is not an advanced catalyst. Yet, owing to the reactive nature of the syncrude, this was ideal for converting, without overconverting, the syncrude. In fact, side-product formation beyond the intended isomerization and oxygenate removal reactions was minimal. Another subtle benefit of incomplete oxygenate conversion over bauxite was the retention of some polar compounds to provide boundary layer lubricity in the fuels. 3) The positioning of the bauxite treatment step before atmospheric distillation was insightful. It not only avoided subsequent boiling range broadening, but also reduced the risk of thermal decomposition in the feed boiler of the atmospheric distillation column. 4) The stated aim of the Hydrocol refinery was to produce mainly motor-gasoline. The selection of solid phosphoric acid oligomerization for upgrading the olefinic C3 –C4 material was therefore an excellent technology choice. It produced oligomers with a narrow carbon number distribution, high naphtha selectivity, and good motor-gasoline properties. This technology enabled the refinery to achieve a motor-gasoline yield close to 80%, the stated design objective of the refinery.
References
5) The Fischer–Tropsch refinery easily met the quality requirements for transportation fuels of that time. 6) Chemical production from the HTFT aqueous product is attractive, but in practice it may be constrained by market size (Table 7.7). Although many of the chemicals in HTFT syncrude have large markets, it is not true of all chemicals. The chemical market can be quickly saturated if multiple Fischer–Tropsch facilities are constructed with chemical extraction [21]. One should take cognizance of this market limitation during the refinery design. There are also other practical considerations. For example, the chemical potential of the carboxylic acids was noted, but no suggestions were made on how to economically recover these compounds from the dilute aqueous product. References 1. Keith, P.C. (1946) Gasoline from natural gas. 2. 3.
4.
5. 6.
7.
8.
9.
10. 11.
Oil Gas J., 45 (6), 102–112. Anonymous (1947) Hydrocol process surveyed before WPRA. Chem. Eng. News, 25, 1044. Kastens, M.L., Hirst, L.L., and Dressier, R.G. (1952) An American Fischer-Tropsch plant. Ind. Eng. Chem., 44, 450–466. Batchelder, H.R. (1962) in Advances in Petroleum Chemistry and Refining, Vol. 5 (eds K.A. Kobe and J.J. McKetta Jr.), Interscience, New York, pp. 3–80. Alden, R.C. (1947) Conversion of natural gas to liquid fuels. Pet. Eng., 18 (1), 148–158. Helmers, C.J., Clark, A., and Alden, R.C. (1948) Catalytic treatment of synthetic gasoline. Oil Gas J., 47 (26), 86–92. Davis, B.H. (2002) Overview of reactors for liquid phase Fischer-Tropsch synthesis. Catal. Today, 71, 249–300. Baird, M.J., Schehl, R.R., Haynes, W.P., and Cobb., J.T. Jr. (1980) Fischer-Tropsch processes investigated at the Pittsburgh energy technology center since 1944. Ind. Eng. Chem. Prod. Res. Dev., 19, 175–191. Eliot, T.Q., Goddin, C.S., and Pace, B.S. (1949) Chemicals from hydrocarbon synthesis. Chem. Eng. Progress, 45, 532–536. Asinger, F. (1968) Paraffins Chemistry and Technology, Pergamon, Oxford. Bruner, F.H. (1949) Synthetic gasoline from natural gas. Composition and quality. Ind. Eng. Chem., 41, 2511–2515.
12. Dry, M.E. (1981) in Catalysis Science and
13.
14.
15.
16.
17.
18.
19.
20.
21.
Technology, Vol. 1 (eds J.R. Anderson and M. Boudart), Springer-Verlag, Berlin, pp. 159–255. Steynberg, A.P., Dry, M.E., Davis, B.H., and Breman, B.B. (2004) Fischer-Tropsch reactors. Stud. Surf. Sci. Catal., 152, 64–195. Clark, A., Andrews, A., and Fleming, H.W. (1949) Composition of a synthetic gasoline. Ind. Eng. Chem., 41, 1527–1532. Noel, H.M. (1959) in American Petroleum Refining (ed. H.S. Bell), 4th edn, D. Van Nostrand, Princeton, NJ, pp. 299–307. Schlesinger, M.D. and Benson, H.E. (1955) Upgrading Fischer-Tropsch products. Ind. Eng. Chem., 47, 2104–2108. Helmers, C.J. and Brooner, G.M. (1948) Catalytic desulfurization and reforming of naphthas over bauxite. Pet. Process., 3, 133–138. Helmers, C.J., Clark, A., and Alden, R.C. (1948) Catalytic treatment of synthetic gasoline. Oil Gas J., 47 (26), 86–92. Tilton, J.A., Smith, W.M., and Hockberger, W.G. (1948) Production of high cetane number diesel fuels by hydrogenation. Ind. Eng. Chem., 40, 1269–1273. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 331–364. Ryan, P. (1945) The Synthol process. Oil Gas J., 43 (47), 264–268.
151
153
8 Sasol 1 Facility 8.1 Introduction
The Fischer–Tropsch work of P. C. Keith of Hydrocarbon Research Inc., which formed the basis for the Hydrocol process (Chapter 7), was also of interest to Anglovaal, the Anglo-Transvaal Consolidated Investment Company. In 1945, Anglovaal officially informed the South African government of its intentions to build a coal-to-liquids (CTL) plant in South Africa. Although a license to produce synthetic fuels was officially given to Anglovaal in 1949, it became apparent that a venture of this magnitude and cost would be difficult to finance without government support. It was also realized that synthetic fuels facility was in national interest, since South Africa had no proven oil reserves. This led to the formation of a new company, the South African Coal, Oil and Gas Corporation (Afr. ‘‘Suid-Afrikaanse Steenkool, Olie-en Gaskorporasie’’), better known as Sasol. On 26 September 1950, Sasol became a public company, with Dr. P. E. Rousseau as its managing director [1]. During this period, the technology selection for the synthetic fuels process was already under way. Five proposals were considered, of which only the proposals of the Arbeitsgemeinschaft Ruhrchemie-Lurgi (Arge) in Germany and the M. W. Kellogg Company in the United States were found to be technically sound enough. It is interesting to note that P. C. Keith originally worked for M. W. Kellogg before forming the Hydrocarbon Research Inc. company. The high-temperature Fischer–Tropsch (HTFT) technology used in the Hydrocol process and the HTFT process offered by Kellogg were different, but related. It was realized that the HTFT Kellogg proposal would be much cheaper to implement, but that the low-temperature Fischer–Tropsch (LTFT) Arge proposal brought all the German industrial experience to the table and that the latter was therefore a much safer option. Eventually both proposals were accepted. The Sasol 1 facility was constructed using two-thirds American technology and one-third German technology. Initially it was thought to run the processes in parallel, but separately. However, it was soon realized that there was considerable synergy between the HTFT Kellogg and LTFT Arge technologies. In order to exploit the synergism between HTFT and LTFT synthesis, the two processes were integrated. Construction of the Sasol 1 facility started in the middle of 1952 at a place that is now called Sasolburg. It is located about 100 km south of Johannesburg, close to the Vaal River. The Vaal River is one of the largest rivers in South Africa and this was a desirable location for the Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
154
8 Sasol 1 Facility
new facility, because it would have access to water for the CTL process. The location was also selected because of its proximity to the Sigma coal mine, which would provide raw material for gasification. Commissioning of the air separation plant, power generation plant, and coal gasifiers started in 1954, but the Fischer–Tropsch synthesis section and refinery were commissioned only the following year. The Kellogg HTFT section was commissioned on 23 August 1955, and the Arge LTFT section on 26 September of the same year. By 1 November 1955, the refinery already produced the first synthetic motor-gasoline for the local market. In the 1950s coal was a cheap raw material at US$0.60 per ton compared to crude oil at US$1.50 per barrel. Yet, despite the price difference in raw material cost with that of crude oil, the facility failed to turn a profit until 1960. This was due to some of the initial technical difficulties that were experienced after commissioning [2]. The Sasol 1 facility is still in operation. Over the years it has seen many changes, although some of the original equipment is still in use. The discussion will therefore take place in two parts. Initially, the discussion will be limited to the original Sasol 1 design (Sections 8.2–8.4), which will be followed by a discussion of the changes over time (Section 8.5).
8.2 Synthesis Gas Production 8.2.1 Lurgi Dry Ash Coal Gasification
The original Sasol 1 plant was designed to use coal as feed material. Because of the high ash content (20–40%) of the subbituminous coal employed as feed, low-temperature moving bed Lurgi dry ash gasifiers (Section 3.4.2) were selected. A total of nine gasifiers were installed to convert the coal in the presence of oxygen and steam into synthesis gas. When the Sasol 1 gasification section was built, it was the largest in the world. The Lurgi dry ash gasifiers were operated at 2.7 MPa pressure, with a product gas outlet temperature in the range 300–650 ◦ C. The original gasifiers were called Lurgi Mark I and had an internal diameter of 3.6 m. These gasifiers are different from subsequent larger designs. Each gasifier was designed to produce around 25 000 m3 ·h−1 raw gas (gas volume at normal conditions). Over time, output was improved to average around 35 000 m3 ·h−1 . The steam and O2 consumption per 1000 m3 raw gas was 850 and 160 m3 , respectively [2]. Because of the high ash content of the coal, ash disposal was an important part of the solid waste management of the facility. The ash was transported as an ash–water slurry in sluice ways to a screening and separation plant. The coarse ash was disposed of on heaps, whereas the fine ash (particles <0.4 mm) had to be processed through fine ash dams, which are analogous to tailing dams. The clarified water from the tailings dams were reused for ash transport. The air separation unit that provided O2 to the coal gasifiers delivered O2 at a purity of 99.1% [3]. This implies that some of the N2 that was present in the raw syngas originated from the oxidant and not from the coal. From a refining perspective, the most important aspect of the gasifier selection is the coproduction of coal pyrolysis products with the raw synthesis gas. These products must be
8.2 Synthesis Gas Production Waste heat boiler
Coal
Precooler
Lurgi dry ash coal gasifier
Raw syngas
LP steam Gas liquor Boiler feed water
Raw gas coolers
Cooling water
Raw gas
Raw gas
Gas liquor
Inert gases
130 °C
O2
30 °C
Expansion vessels
Gas liquor coolers
Steam 70 °C
Expansion vessels
Oil separator
Neutral oil
Ash
Gas liquor Tar separator
Tar Sludge
Figure 8.1
Synthesis gas production by low-temperature coal gasification in Lurgi dry ash gasifiers.
separated from the raw synthesis gas (Figure 8.1). Separation of the neutral oil, gas liquor, and tar was based on liquid density differences. The liquid products also contained dissolved light hydrocarbons, NH3 , and CO2 . The degassed coal pyrolysis product was rich in aromatics, phenols, tar acids, and creosotes and were refined to produce fuels and chemicals (Section 8.4.4).
8.2.2 Rectisol Synthesis Gas Cleaning
The raw gas from low-temperature coal gasification contained 60% H2 and CO, 29% CO2 , 9% CH4 , the remainder being N2 , Ar, H2 S, hydrocarbons, and other contaminants. The composition of the sulfur-containing compounds in the raw gas was 97% H2 S, 2% CH3 SH, and 1% COS and CS2 [2]. Since the sulfur-containing compounds are Fischer–Tropsch catalyst poisons, these compounds were removed from the raw syngas during syngas cleaning. Syngas cleaning was performed in a Rectisol unit. The Rectisol process was a new process in 1955 and the unit at Sasol 1 was the first commercial installation [4]. The installation consisted of three identical scrubbing trains followed by a common regeneration section. The Rectisol process uses physical absorption of the acid gases under pressure in methanol at cryogenic conditions. This process was found to be very effective for the removal of all sulfur-containing compounds, in addition to removing about 98% of the CO2 [3].
155
156
8 Sasol 1 Facility
Rectisol wash train
Main wash column
Flare
Flare
I
II
−35 °C
Raw syngas
Pre-wash column
Stack
Fine wash column
Pure syngas
Flare
III − 45 °C IV −54 °C
Main wash expansion column
V −59 °C VI −63 °C
Flare
Stack
VII −70 °C
Expansion vessel
Methanol
Steam condensate
Fine wash stripping column
Rectisol naphtha
Naphtha separator
Regeneration Water/ methanol mixture
Naphtha
Naphtha stripper
Methanol
Methanol distillation Water
Figure 8.2
Raw synthesis gas cleaning by the Rectisol acid gas removal process.
Detailed descriptions of the Rectisol process can be found in the literature [5, 6]. Since one of the refinery feed streams originated from this process, a brief description is included. The Rectisol process as applied at Sasol 1 consisted of three steps (Figure 8.2): 1) Prewash. In the prewash step, the condensable hydrocarbons, oxygenates, and organic sulfur compounds were removed from the gas as liquid products. The liquid product thus produced is called Rectisol naphtha and it was co-refined with the other coal pyrolysis products (Section 8.4.4). 2) Main wash. In the main wash step, the CO2 was liquefied under pressure. Most of the H2 S, COS, and CS2 , as well as 95% of the CO2 were removed. 3) Fine wash. The fine wash removed the remaining sulfur compounds (COS and H2 S) in addition to some CO2 . The pure syngas that was obtained from the Rectisol unit contained at most 40 µg·m−3 sulfur. This was of sufficient purity to be used for Fischer–Tropsch synthesis without further purification. A typical composition of the purified syngas feed after the Rectisol unit was 85% H2 and CO, 1% CO2 , 13% CH4 , and 1% N2 and Ar [2]. When the methanol was expanded to unload the absorbed gases, an acid gas stream containing mainly 1.5% H2 S and 98.5% CO2 was produced. The sulfur in this gas was not recovered, but discharged directly to the atmosphere [3].
8.3 Fischer–Tropsch Synthesis
8.3 Fischer–Tropsch Synthesis 8.3.1 Kellogg HTFT Synthesis
The Kellogg Synthol process for HTFT synthesis made use of a fused-iron catalyst that was operated in a circulating fluidized bed (CFB) reactor (Figure 8.3). This technology is referred to as Kellogg HTFT synthesis in order to differentiate the American CFB Fischer–Tropsch reactor technology from the modified ‘‘Synthol’’ CFB reactor technology developed by Sasol in the 1970s. Although the latter technology became known as Synthol the term Synthol had been in colloquial use since the 1940s [7]. The CFB Fischer–Tropsch reactors had been newly developed by Kellogg and it had only been tested on a 0.1 m diameter pilot plant scale before being scaled up for the Sasol 1 facility. The industrial scale reactors were 2.3 m in diameter and 46 m in height [8]. Two Kellogg CFB reactors were installed, each with a production capacity equivalent to 2000 bbl/day [9]. Typical operating conditions proposed by Kellogg for Fischer–Tropsch synthesis were 290–340 ◦ C and 1.9 MPa [10]. The syngas was hydrogen rich and contained 61.8% H2 , 22.4% CO, 7.2% CO2 , the remainder being methane and inerts [11]. The syngas feed was fed at the bottom of the reactor at a pressure of 2.2 MPa, where it came into contact with a stream of catalyst, somewhat analogous to the operation of a fluid catalytic cracker unit. The catalyst was entrained by the syngas into the reaction zone. The process could reliably achieve 85% conversion of CO Product gas
Cyclone
Heat exchangers
Catalyst hopper Standpipe with catalyst
Slide valve
Entrained flow of gas and catalyst
Syngas
Figure 8.3
Kellogg circulating fluidized bed reactor for HTFT synthesis.
157
158
8 Sasol 1 Facility
and CO2 at fresh feed rates of 100 000 m3 ·h−1 . Two banks of heat exchangers in the reaction zone removed 30–40% of the reaction heat. The hot product gas left the reactor via cyclones that removed the entrained catalyst particles and returns them to the catalyst hopper. The exit gas temperature was typically in the range 320–360 ◦ C [8]. The temperature increase over the reactor was 15–20 ◦ C [9]. The catalyst recirculation rate was estimated at more than 4000 t·h−1 , with a loss due to fines of about 0.0015% [10]. One of the design advantages cited for the use of a CFB reactor was the ability to change catalyst while the unit is in production. Although this was the design intent, the Kellogg reactor technology did not live up to this promise and the operating procedure was eventually changed to campaign mode operation to make the technology work [1]. The run length was limited by carbon buildup on the catalyst and each run lasted about 40–50 days before the reactor had to be unloaded, cleaned, and loaded with fresh catalyst [8, 12]. This placed a serious constraint on equipment availability, and in 1960 a third CFB reactor was installed. Although the Kellogg CFB reactors proved to be flexible in terms of dealing with fluctuations in flow rate, composition, temperature, and pressure of the synthesis gas, the design had serious operational difficulties associated with the catalyst circulation rate. There was a small operating window that allowed stable performance. At high catalyst flow rates the intercooler tubes of the reactor were plugged, resulting in a loss of heat transfer and increased erosion of the tubes that were not plugged. At low catalyst flow rates, bridging of the catalyst in the standpipe occurred, causing an interruption in production [13]. The process originally used a fused-iron catalyst made from a magnetite ore (Allenwood ore) [14], similar to that for Hydrocol (Chapter 7). However, the Hydrocol operation was a fixed fluidized bed, not a CFB, and the fused catalyst supplied by Kellogg was not without its own problems [1]. During CFB operation, the catalyst is turbulently transported at linear velocities of several meters per second. The mechanical strength of the catalyst is therefore of paramount importance. If the catalyst is not robust enough, catalyst attrition will result in excessive fines formation (Section 4.5.4). Fines formation increases catalyst loss, as well as a change in the fluidization behavior [2, 12]. The Kellogg HTFT technology used in the Sasol 1 facility was licensed technology, and Sasol at that stage was a new company. In order to overcome some of the operational problems associated with the Kellogg technology, the Sasol Research department was established in 1957. Much effort was devoted to understanding the behavior and requirements of fused-iron Fischer–Tropsch catalysts. The suitability of local sources of magnetite was investigated, and mill scale from the nearby steelworks in Vanderbijlpark was used to replace the imported Allenwood ore. The preparation of fused-iron catalyst and the influence of chemical promoters (e.g., K2 O) and structural promoters (e.g., MgO, Al2 O3 ) have been described in the literature [8, 12, 14]. After production and milling of the catalyst to the correct particle size distribution, it is reduced with hydrogen in a fixed bed reactor at 350–450 ◦ C over a period of two days. The unreduced catalyst has a surface area of only 1 m2 ·g−1 , but on reduction this could be increased from 2 to 30 m2 ·g−1 , depending on the type of promoters that were used. The product distribution from the Kellogg HTFT synthesis is given in Table 8.1 [15]. When this is expressed in terms of only C3 and heavier hydrocarbons, it amounts to 32% C3 –C4 liquid petroleum gas (LPG), 53% naphtha, 8% distillate, and 7% residue. As expected, the Kellogg HTFT syncrude is very similar to that of Hydrocol HTFT syncrude (Table 7.1). The product was quite olefinic and contained oxygenates and aromatics too (Table 8.2) [15]. The oxygenates partitioned
8.3 Fischer–Tropsch Synthesis Table 8.1
Composition of Kellogg CFB Fe-HTFT syncrude.
Compound
Syncrude composition (C atom%)a
Methane Ethene Ethane Propene Propane Butenes Butanes C5 –C11 C12 –C18 C19 –C21 C22 –C30 C31 and heavier Aqueous product Nonacid chemicals Carboxylic acids
10 4 6 12 2 8 1 39 5 1 3 2 6 1
a Average compositions and actual compositions change over time with operating conditions and catalyst age.
Table 8.2 Compound classes in naphtha and kerosene fractions of Kellogg CFB Fe-HTFT syncrude.
Compound class
n-Alkanes iso-Alkanes Alkenes Aromatics Alcohols Carbonyls
Syncrude composition (%) C5 –C10
C11 –C14
7 6 70 5 6 6
9 6 60 15 5 5
between the aqueous and organic product phases. Most of the lighter oxygenates ended up in the aqueous product. 8.3.2 Arge LTFT Synthesis
The Arge process for LTFT synthesis is, at the time of writing, still in operation. It made use of a precipitated iron-based Fischer–Tropsch catalyst. The catalyst was loaded in multitubular fixed bed reactors that were operated in down-flow mode (Figure 8.4).
159
160
8 Sasol 1 Facility Syngas
Water
Steam
Tubes filled with catalyst
Product gas
Wax product
Figure 8.4
Arge fixed bed reactor for LTFT synthesis.
Five Arge reactors were installed in the Sasol 1 plant with a design capacity of 550 bbl/day. This is equivalent to a reactor production capacity of approximately 18 000 tons per year per reactor [9]. Every Arge reactor contained 2052 tubes, and each tube was 12 m long with 50 mm internal diameter [8, 10, 16]. The total catalyst charge was about 40 m3 , and the design fresh feed gas hourly space velocity was 500 h−1 [8]. The process was designed to use a H2 :CO ratio of 1.7–1.8 : 1 [10]. The reaction temperature could be controlled and maintained almost isothermally by regulating the pressure at which the boiler feed water was allowed to evaporate on the outside of the reactor tubes. Typical operating conditions of Fischer–Tropsch synthesis during start-up were 200–230 ◦ C and a gas inlet pressure of 2.5–2.7 MPa. During operation, the temperature was increased by 25–30 ◦ C before end-of-run conditions were reached [10]. Run lengths exceeding 350 days could be achieved but the run length was typically dictated by the product selectivity that was desired [8] (see also Section 4.5.6.). The precipitated iron catalyst was produced on site by dissolving iron in nitric acid and precipitating it by adding sodium carbonate. The potassium promoter was added before extrusion. The catalyst paste was extruded to produce catalyst particles of about 3 mm diameter and 10 mm length and these extrudates were then dried. The catalyst was reduced at around 220 ◦ C and then coated in wax to prevent reoxidation before being loaded into the reactor. Commissioning of the reactor was conducted under a hydrogen atmosphere to prevent the formation of volatile iron carbonyls [14]. The Arge LTFT syncrude composition was typical of a high α-value Fischer–Tropsch product (Table 8.3) [8, 15]. The oil product contained alkenes and alcohols as major compound classes in addition to alkanes that was the dominant compound class (Table 8.4) [8, 15]. Little aromatics
8.3 Fischer–Tropsch Synthesis Table 8.3
Composition of Arge fixed bed Fe-LTFT syncrude.
Compound
Methane Ethene Ethane Propene Propane Butenes Butanes C5 –C11 C12 –C18 C19 –C23 C24 –C35 C35 and heavier Aqueous product Nonacid chemicals Carboxylic acids
Syncrude composition (C atom%)a [8]
[15]
2 0.1 1.8 2.7 1.7 3.1 1.9 18 14 7 20 25
5 0.2 2.4 2 2.8 3 2.2 22.5b 15b 6b 17b 18b
3 0.2
3.5 0.4
a Average composition and actual compositions change over time with operating conditions and catalyst age. b Reported for the ranges C –C , C –C , C –C , 5 12 13 18 19 21 C22 –C30 , and C31 and heavier.
Table 8.4 Compound classes in the naphtha and distillate fractions of Arge fixed bed Fe-LTFT syncrude.
Compound class
Syncrude composition (%) C5 –C11 [8]
n-Alkanes iso-Alkanes Alkenes Aromatics Alcohols Carbonyls Carboxylic acids
57 3 32 0 7 0.6 0.4
C5 –C12 [15] 50 3 40 0 6 1 –
C12 –C18 [8] 60 5 25 0 6 <1 0.05
C13 –C18 [15] 60 5 28 0 6 1 –
were found. Ketones and carboxylic acids were minor components. As the product became heavier, it became more paraffinic. The medium wax fraction (C24 –C35 ) contained 10% alkenes [8]. It should be mentioned that the catalyst formulation for the precipitated Arge Fe-LTFT catalyst was improved over time to increase the α-value. A more recent analysis of a broad-cut straight-run Arge LTFT wax with average carbon number of C43 revealed the following compound classes [17]: 88% alkanes, 5.7% alkenes, 0.3% aromatics, and 6% oxygenates. The density was 920 kg·m−3 and it contained only 2.3% branched material.
161
162
8 Sasol 1 Facility
8.3.3 Gas Loop Design
The original Sasol 1 gas loop design is interesting, because it combined HTFT and LTFT synthesis reactors that operate at different temperatures and pressures and with different H2 :CO ratios into a single gas loop (Figure 8.5) [3, 10]. The Arge LTFT reactors required a higher pressure and had a lower synthesis gas conversion than the Kellogg HTFT reactors. The tail gas from Arge synthesis was partly recycled to the Arge reactors to ensure the correct H2 :CO ratio (LTFT internal recycle) and partly sent to the gas reformer in the HTFT gas loop. The tail gas from Kellogg synthesis was partly recycled to the Kellogg reactors (HTFT internal recycle) and partly sent to the gas reformer (HTFT external recycle). In this way, the high CO conversion during HTFT synthesis was exploited to convert the unreacted CO from LTFT synthesis. The gas reformer in the HTFT gas loop was employed to convert the methane-rich gas into synthesis gas. It was a catalytic partial oxidation type gas reformer with a nickel catalyst operating at around 1000 ◦ C [2]. The feed gas typically contained around 20–25% CH4 , 45–50% H2 , 20% CO, and the remainder CO2 and inerts (e.g., Ar and N2 ). The product from the gas reformer contained <5% CH4 [3]. A flow diagram with flow rates and compositions of the different streams in the HTFT–LTFT gas loop has been reported by Hoogendoorn [9].
Raw syngas
Rectisol
Arge tail gas
Cold separation
C 3 – C4 Naphtha
Pure gas
Fuel gas (purge)
2.2 MPa
Gas absorption
Arge tail gas 35 °C
2.5 MPa
Cooling water
External recycle
Gas wash tower Lean oil
Cold condensate Arge LTFT
Water
Reaction water Reactor wax
Carbonate
Hot condensate
Water and spent carbonate
Kellogg tail gas
C3–C4 Internal recycle
Debutanizer
145 °C 1.6 MPa
40 °C
Light oil Reaction water
Naphtha Kellogg HTFT O2 Steam
Figure 8.5
Gas reforming
Combined HTFT–LTFT gas loop of the original Sasol 1 design.
Decanted oil Gunk
8.4 Fischer–Tropsch Refining
8.4 Fischer–Tropsch Refining
The flow scheme of the original Sasol 1 design is very complex. The feed material being sent to the refinery originates from coal gasification (coal pyrolysis products), syngas cleaning (Rectisol naphtha), and the various stepwise condensation fractions from HTFT and LTFT syntheses. The overall Fischer–Tropsch refinery actually consisted of four different refining sections, namely, tar workup, Kellogg oil workup, Arge oil workup, and chemical workup (Figure 8.6). The Kellogg oil workup, Arge oil workup, and chemical workup sections together constituted the Fischer–Tropsch refinery. Unlike the gas loop design, the refinery design had very little integration, apart from a common aqueous product workup section and transfer of the C3 –C4 fraction from the Arge oil workup to the Kellogg oil workup. The tar workup section refined the coal pyrolysis products from low-temperature gasification. These by-products from syngas production (Figure 8.1) and syngas cleaning (Figure 8.2) were similar to those derived from the coke ovens associated with many of the German Fischer–Tropsch facilities (Section 6.2). The tar workup section was consequently not really part of the Fischer–Tropsch refinery, but required due to the gasification technology that was selected.
8.4.1 Kellogg HTFT Oil Refining
The HTFT Kellogg product refinery consisted of two conversion units, namely, a clay-treater and an oligomerization unit (Figure 8.7) [18]. The similarity with the design of the Hydrocol refinery (Figure 7.2), which employed bauxite treating and oligomerization as the only two conversion units, is self-evident.
Gas liquor Gasification
Neutral oil Tar workup
Tar Rectisol
Naphtha
Kellogg HTFT
Aqueous product
Oil
Kellogg oil workup
Figure 8.6
Chemical workup
Arge LTFT
Aqueous product
C3 – C4
Oil
Arge oil workup
High-level flow diagram of the original Sasol 1 refinery.
163
164
8 Sasol 1 Facility
Oil wash tower Cracked gas (To gas absorber)
Water
425 °C
Naphtha
Claytreaters
Atmospheric distillation
Light oil Aqueous product (To chemical workup)
(From debutanizer bottoms)
Motor-gasoline Vacuum flash
Diesel Water
Decanted oil
Heavy oil Steam
>345 °C
Water
Gas (To external recycle)
C3 – C 4 LPG Catalytic polymerization (oligomerization) Depropanizer
Polymer gasoline
Debutanizer
Figure 8.7
Kellogg HTFT workup section of the original Sasol 1 refinery.
The clay-treater had a function similar to the Bauxite treatment unit used in the Hydrocol process (Section 7.4.1). The objective was to remove oxygenates by converting them into alkenes and to improve the octane number of the linear 1-alkene-rich Kellogg HTFT syncrude by double bond isomerization. The unit had an operating temperature of 425 ◦ C and employed a silica–alumina acid catalyst. It is likely that the catalyst was either an acidified montmorillonite or/and acidified beidellite, which are naturally occurring zeolites. Unlike bauxite, which is alumina-rich and less acidic, the more active silica–alumina catalyst resulted in cracking of the syncrude. Cracked gas produced in the clay-treater was sent back to the gas loop. The cracked gas stream was combined with the product from the gas wash tower (Figure 8.5), which was then routed to the gas absorber and debutanizer, where the C3 –C4 fraction was recovered and returned to the Kellogg workup section. All C3 –C4 molecules were routed to the oligomerization (‘‘catalytic polymerization’’) unit. The feed gas to the process contained 60–70% alkenes, which were converted into a high octane olefinic motor-gasoline and some distillate. The oligomerization technology originally selected for the Sasol 1 refinery was that of the Polymer Corporation, which employed a copper pyrophosphate and charcoal catalyst [19, 20]. The process was very similar to the better known Universal Oil Product (UOP) Catalytic Polymerization process, which employed a solid phosphoric acid catalyst (Section 19.3.1). The copper pyrophosphate-based technology was probably selected because Kellogg developed a similar process based on this type of catalyst [21].
8.4 Fischer–Tropsch Refining
There were four reactors in the oligomerization unit, each 1.2 m in diameter and 7.6 m in height. Every reactor contained five catalyst beds with interstage quench capability. Part of the feed was compressed to 6–7 MPa and preheated to 220 ◦ C before being introduced to the reactors, while the remainder of the feed was used as reactive interstage quenches. The reactors were configured to operate as a pair of two reactors in series. When fresh catalyst was loaded in one of the reactors, the reactor with fresh catalyst was always configured to be the second reactor in series to boost conversion. By doing so, the risk of temperature excursions due to rapid oligomerization over a fresh catalyst was reduced [18]. The bottom product from the two debutanizer columns (Figures 8.5 and 8.7), the steam-stripped decanted oil, and the degassed product from clay treatment were combined. This combined oil stream served as feed to the oil fractionator, together with the oligomers produced from the conversion of the C3 –C4 fraction. The oil fractionator was an atmospheric distillation unit that produced motor-gasoline, diesel, and fuel oil cuts as final products.
8.4.2 Arge LTFT Oil Refining
There are three conversion units in the Arge workup section, namely, bauxite treatment, catalytic cracking, and wax hydrogenation (Figure 8.8) [18].
C3 – C4 (To Kellogg workup)
C3 – C4
Bauxite treatment (Hot refining)
C3 – C4 C5 – C7 naphtha
Arge tail gas cold separation
Arge tail gas
Naphtha
Motor-gasoline
Atmospheric distillation
C8 + naphtha
Power paraffin NaOH wash
Cold condensate
Water wash
Diesel Fuel oil
Hot condensate Waxy oil
C3 – C4
<320 °C
Thermal cracking (Paraforming)
Reactor wax
Gatsch
Medium wax
H2
320 – 370 °C
Steam stripper
Soft wax
De-oiling
370 – 460 °C
H2
Steam
High vacuum distillation
Wax hydrogenation >460 °C
Figure 8.8
Arge LTFT workup section of the original Sasol 1 refinery.
Hard wax
165
166
8 Sasol 1 Facility
Bauxite treatment, which is similar to that used in the Hydrocol refinery (Section 7.4.1), converts the oxygenates and double bond isomerizes the alkenes to produce a higher octane motor-gasoline as final product. Only the C5 –C7 fraction was bauxite treated, because it was the only fraction where the final product had a high enough octane number to be used as motor-gasoline. The process was operated at 400 ◦ C, slightly lower than the clay-treaters in the Kellogg workup, but the aim of both processes was the same. It is interesting that, on the Arge-side of the refinery, bauxite treatment was selected instead of clay treatment as in Kellogg workup (Section 8.4.1). Although deoxygenation and isomerization by alumina-rich catalysis (bauxite) is better, the selection of two different technologies to perform the same task underscores the lack of integration in the refinery design. The Paraformer, or paraffin reformer plant, was a thermal cracker that converted the hard wax to medium and soft wax, as well as the waxy oil (slack wax) fractions to diesel and motor-gasoline [18]. The conversion and product selectivity during thermal cracking is controlled by varying the temperature of the process. The C3 –C4 cracked gas is sent to the Kellogg workup, while the naphtha and distillate range product is caustic washed with the cold condensate before being distilled with the hot condensate into naphtha, power paraffin (kerosene), diesel fuel, fuel oil, and waxy oil fractions. The kerosene, diesel fuel, and fuel oil are final products, while the naphtha is first bauxite treated before becoming motor-gasoline as the final product. Waxy oil and reactor wax were steam stripped to remove the lighter than 320 ◦ C material, before it was deep vacuum distilled to produce soft, medium, and hard wax fractions. The waxes contained small amounts of alkenes and oxygenates that had to be removed by hydrogenation. Wax color was also an issue, which could be improved by hydrogenation to meet the required Saybolt color specification [22]. Three reactors were used, two for hard wax hydrogenation and one to hydrogenate the lighter fractions [18]. Hydrogenation was performed at 260 ◦ C, 5 MPa, and a liquid hourly space velocity (LHSV) of 0.3–0.5 h−1 . Unsulfided nickel hydrogenation catalysts were used for this task [2]. Operation of these hydrogenation reactors was early on changed to the up-flow mode to improve temperature control. Wax hydrogenation is not very exothermic, but the refinery H2 was occasionally contaminated with CO, which resulted in severe temperature excursions. By changing to up-flow mode of hydrogenation, this operational difficulty was overcome. The de-oiling of the wax was performed by dissolving the wax in a suitable solvent at 60 ◦ C and then slowly cooling it down in a scraper cooler [18]. The wax fractions that crystallized out could be recovered by filtration. The solvent typically used in such applications is 2-butanone (methyl ethyl ketone). Details of the solvent dewaxing technology at that time can be found in the literature [23]. Much effort has gone into the characterization of the different wax fractions that could be obtained from these refining steps. Foremost was the measurement of the wax crystallinity and degree of branching [24–28]. 8.4.3 Aqueous Product Refining
The oxygenates in the Fischer–Tropsch aqueous product are classified as either nonacid chemicals (Table 8.5) [2, 9] or as carboxylic acids (Table 8.6) [9]. The nonacid chemicals are
8.4 Fischer–Tropsch Refining Nonacid chemicals in the aqueous products from Kellogg HTFT and Arge LTFT synthesis.
Table 8.5
Compound
Nonacid chemicals (mass%) Kellogg HTFTa
Alcohols Methanol Ethanol 1-Propanol 2-Propanol (isopropanol) 1-Butanol 2-Butanol 2-Methyl-1-propanol (isobutanol) 1-Pentanol 2-Pentanol Other alcohols Aldehydes Ethanal (acetaldehyde) Propanal (propionaldehyde) Butanal (butyraldehyde) Pentanal Ketones Acetone 2-Butanone (MEK) Pentanones
Arge LTFT
1.4 55.6 12.8 3.0 4.2 0.8 4.2 1.2 0.1 0.6
24 50 11 0 6 0 0 4 0 0
3.0 1.0 0.6 0.2
2b
10.6 3.0 0.8
2 1 0
a Source’s b Value
nonacid chemical composition adds up to 103.1%. for all aldehyde species in the product.
Table 8.6 Composition of the carboxylic acids in the aqueous product of Kellogg HTFT synthesis.
Compound
Mass%
Ethanoic acid (acetic) Propanoic acid (propionic) Butanoic acid (butyric) C5 and heavier acids
70 16 9 5
typically aldehydes, ketones, and alcohols, with a distribution much like that of the Hydrocol aqueous product (Table 7.7). The combined aqueous product from both Kellogg HTFT and Arge LTFT synthesis contained about 6% oxygenates dissolved in water. The chemical workup section consisted of two parts, namely chemical recovery and solvent recovery (Figure 8.9) [18]. The aqueous product was first distilled in the primary distillation column, which was stainless steel lined on account of the carboxylic acids in the feed. To prevent any carboxylic acids from
167
168
8 Sasol 1 Facility
Chemicals recovery
Solvent recovery
Aldehyde tower Carbonyl stripper
Acetaldehyde Water
Primary distillation
Ketone tower
Crude carbonyl mixture
Alkali 25% H2O
Water
Aqueous product Wastewater
Crude alcohol mixture
Alkali Steam Water
Ethanol or ethanol– propanol mixture
Water removal (Keyes process)
Alcohol distillation
Acetone, butanone, butanal and 12% water
Alcohol – water splitter Methanol tower
Methanol, ethanol and 95% water
Methanol Aldehyde hydrogenation
H2
H2 Water Wet ethanol
Hydrogenation
Heavy alcohols
Figure 8.9 Chemical workup section in the original Sasol 1 refinery that processed the combined aqueous product from Kellogg HTFT and Arge LTFT synthesis.
distilling into the overhead product, an alkali solution was injected on two trays near the top of the column. The bottom product from the column was also treated with alkali to neutralize the carboxylic acids. The overhead product from primary distillation of the aqueous product was rich in oxygenates and contained 25% water. The overhead product served as feed for the carbonyl stripper column. The carbonyl stripper obtained a crude separation between the alcohols and carbonyls. The overhead product from the carbonyl stripper contained mostly carbonyls (aldehydes and ketones), esters, and methanol, which were further refined in the solvent recovery section. The bottom product from the carbonyl stripper contained mostly water, ethanol, and the heavier alcohols. This crude alcohol fraction was further purified in the chemical recovery section. The first step in alcohol purification was water removal by azeotropic distillation with benzene. This is a textbook purification called the Keyes process [29]. The water-free alcohol stream was then mildly hydrotreated to convert the residual aldehydes and ketones into alcohols. The final alcohol purification step was used to produce ethanol or an ethanol–propanol mixture as fuel alcohols depending on market demand. It was not enough to purify the heavy alcohols by continuous distillation, and the mixture was sent to the solvent recovery section where the alcohols were separated by batch distillation (not shown in Figure 8.9).
8.4 Fischer–Tropsch Refining
In the solvent recovery section, the overhead product from the carbonyl stripper was distilled in various steps to successively remove ethanal, mixed carbonyls (mainly acetone, 2-butanone, and butanal), water, methanol, and ethanol. Of these, only methanol was a final product. The ethanal was hydrogenated to ethanol at 150 ◦ C and 3.8 MPa over an unsulfided nickel catalyst. The reactor had three catalyst beds with interstage quenching. It is interesting to note that preheating of the ethanal was done in the presence of hydrogen, presumably to limit reactions such as aldol condensation. The product was then combined with the wet ethanol from the methanol tower and processed with the other alcohols in the chemical recovery section. The mixed carbonyls, like the heavy alcohols, were further separated by batch distillation. The aqueous effluent that contained about 1% carboxylates was concentrated in an evaporator to yield a 50% salt solution. The alkali was regenerated from this solution, but the acids were not recovered, although this intent was expressed [30]. It is not clear when the practice to neutralize all the acids were discontinued. Later on it was reported that the wastewater obtained as bottom product from the primary distillation column was biologically treated to destroy the dissolved carboxylic acids [2]. After biological treatment, the wastewater was returned to the Vaal River. The total water intake of the Sasol 1 facility was 63 000 m3 /day, of which 31 500 m3 /day was returned to the Vaal River [31]. 8.4.4 Coal Pyrolysis Product Refining
Coal pyrolysis products are rich in aromatic and phenolic material. The phenolic material and NH3 were concentrated in the aqueous phase (gas liquor). The aromatic hydrocarbons and heavier heteroatom-containing aromatics were concentrated in the organic phase (Rectisol naphtha, neutral oil, and tar). These streams were refined separately from the Fischer–Tropsch syncrude. The only integration between the tar workup section and the Fischer–Tropsch refinery was during fuel blending. The aromatic naphtha fraction was employed as a blending component in the final motor-gasoline. The refining of the aqueous and organic phase took place in two separate sections of the tar workup section (Figure 8.10). The gas liquor had an average concentration of 0.85% dissolved NH3 and 0.18% phenolic compounds (tar acids). The technologies to refine these products were well established in the coal industry [32]. In the Phenosolvan process, the phenolic material was selectively extracted with butyl acetate at a pH of 8.5 to produce a crude phenol mixture that contained 40% phenol, 30% cresols, 7% xylenols, and 23% higher boiling tar acids [30]. The NH3 was steam stripped from the phenol–lean gas liquor and separated from the dissolved CO2 . The CO2 was vented to atmosphere and the phenol–lean gas liquor was disposed of through biological treatment. A flow diagram of this process can be found in the paper by Hoogendoorn and Salomon [30]. The ammonia stream thus purified contained around 6% NH3 . The ammonia was converted into ammonium sulfate fertilizer. The material in the organic phase was fractionated by atmospheric distillation. The naphtha fraction contained some water, which was phase separated before the naphtha was fed to the coal tar naphtha hydrotreater. The naphtha was preheated and vaporized in the presence of hydrogen. Hydrotreating was conducted in the vapor phase in a fixed bed reactor at 315–370 ◦ C and 5 MPa with interbed hydrogen quenching. Hydrotreating removed most of the heteroatoms,
169
170
8 Sasol 1 Facility Tar – oil separation Gas liquor
Gas liquor
Ammonium sulfate production
NH3
Phenol extraction (Phenosolvan)
(NH4)2SO4 Crude phenols
Tar/oil Gas naphtha hydrogenation
Hydrogenated naphtha distillation
H2
Water wash
H2SO4 wash
NaOH wash
Water wash
Aromatic gasoline
Atmospheric distillation
Rectisol naphtha
Neutral creosote
Creosote
Water
Neutral oil Tar
Road tar prime Vacuum distillation
Pitch
Figure 8.10 Tar workup section in the original Sasol 1 refinery that processed the coal pyrolysis products from synthesis gas production and purification.
and it was specifically used to remove gum-forming substances, sulfur, and phenols. The hydrogenated product was cleaned by alkali and acid washing before being distilled. Although only the fractionation scheme for aromatic motor-gasoline is shown (Figure 8.10), some of the hydrogenated naphtha was further separated by batch distillation to prepare benzene, toluene, xylene, neutral oil, and heavy naphtha fractions for the solvent market [30]. The bottom product from the atmospheric distillation of the coal liquids was vacuum distilled to produce creosotes, road tar prime, and pitches. Heavy fractions that could not be sold were discharged into a tar pit, which is the equivalent of a tailings pond for coal tar.
8.4.5 Synthetic Fuel Properties
The straight-run octane number of Kellogg HTFT naphtha is considerably better than that of Arge LTFT naphtha (Table 8.7). Both naphtha streams were significantly upgraded by deoxygenation and isomerization on account of their olefinicity [2]. The refining of LTFT syncrude contributed little to the final motor-gasoline product. The ‘‘polymer gasoline’’ from C3 to C4 alkene oligomerization was a good quality motor-gasoline and a desirable blending component. The
8.4 Fischer–Tropsch Refining Density, research octane number (RON), and motor octane number (MON) of the motor-gasoline blending components in the original Sasol 1 refinery.
Table 8.7
Motor-gasoline blending components
Density (kg·m−3 )
RON
MON
Straight-run HTFT C5 –C11 naphtha Straight-run LTFT C5 –C11 naphtha Clay-treated C5 –C11 HTFT naphtha Bauxite-treated C5 –C7 LTFT naphtha Polymer gasoline (oligomerization) Hydrogenated coal tar naphtha Synthetic ethanola
710–720 710–720 710–720 680–690 710–720 ∼790 ∼790
∼65 ∼35 ∼86 ∼65 95–97 80–85 ∼120
– – – – 82 70–75 ∼100
a Not
a pure ethanol stream.
Table 8.8 Selected properties of the final diesel fuel produced by the original Sasol 1 refinery.
Fuel property Cetane number Density (kg·m−3 ) Viscosity (cSt) Compound classes (vol%) Alkanes Alkenes Aromatics
Diesel fuela 55–65 800 2–3 65–75 20–30 5–10
a Source indicated that it is from Sasol Arge LTFT-only, but the composition indicates that it is HTFT and LTFT combined.
hydrogenated coal tar naphtha did not have a very high octane number despite its high aromatics content. This was due to the hydrogenation of the n-alkenes to n-alkanes, which was detrimental to the octane number of the coal tar naphtha. For over a decade, it was also blending practice to include 15% ethanol from the chemical workup section into the motor-gasoline. With tetraethyl lead (TEL) addition, the final blend easily met the quality requirements for motor-gasoline in South Africa at that time. A regular motor-gasoline typically required a research octane number (RON) of 87 and a motor octane number (MON) of 83. In fact, it was reported that the product from Kellogg HTFT workup had an RON of 86–90 even before lead addition [11]. The straight-run cetane numbers of C12 –C18 Kellogg and Arge distillates were 55 and 75, respectively [2]. Meeting diesel fuel quality requirements was therefore not a problem (Table 8.8) [33]. The lowest quality diesel fuel component was the small amount of kerosene-range material that was produced in the oligomerization process. This material typically had a cetane number of less than 30, but excellent cold flow properties.
171
172
8 Sasol 1 Facility
8.5 Evolution of the Sasol 1 Facility
The original Sasol 1 facility was designed to produce synthetic fuels. One would expect the facility to have changed over time to comply with changes in transportation fuel specifications (Chapter 2). Yet, the main change driver that shaped the changes to the facility had been the extraction of chemicals. Today, Sasol 1 produces mostly chemicals, and the transformation form a combined fuels and chemicals refinery to a chemical refinery is complete. Although it was the main change driver, changes were by no means restricted to the refinery. Despite its age, the facility will reportedly continue to grow. Plans to double the production capacity by 2014 have been announced [34]. 8.5.1 Changes in Synthesis Gas Production
The design requirements for industrial facilities in the 1950s did not place the same emphasis on emission control and on limiting the environmental footprint as is current design practice. Air quality was undermined by the practice of venting all sulfurous emissions to atmosphere, including the H2 S-containing off-gas from Rectisol syngas purification (Section 8.2.2). Early in the 1970s, changes were made to reduce the environmental impact of the original Sasol 1 design. Electrostatic precipitators were added to the power stations and these were commissioned in October 1972. In 1973, a Stretford process was constructed to convert the H2 S in the Rectisol off-gas to elemental sulfur. However, the gas composition and material of construction proved incompatible and this unit was never successfully commissioned [35]. More specifically, it failed to work because of the microbial growth in the water and the wrong type of wood used as packing material. In later years, another attempt was made to reduce the H2 S emissions by developing the Sasol Clean Air Technology (SCAT) [36]. The SCAT process employed the same catalyst as was used for Arge LTFT synthesis. It was piloted in the 1990s, but this technology was never commercialized. Inhabitants of Sasolburg had to wait until 2004 for the H2 S emissions to be reduced when Sasol 1 was converted from a CTL facility into a gas-to-liquids (GTL) facility [37]. Over the years, the gasification section became a test bed for gasification technology. Lurgi Mark IV (3.85 m diameter) and later Lurgi Mark V (5 m diameter) coal gasifiers were successfully tested at Sasol 1 [2]. The Lurgi Mk IV gasifier had a coal throughput of 40 t·h−1 and could produce around 73 000 m3 ·h−1 (normal) gas, whereas the Lurgi Mk V gasifier had a coal throughput of 60 t·h−1 and could produce around 115 000 m3 ·h−1 (normal) gas [38]. In 2004, a pipeline was completed to import natural gas from Mozambique. In anticipation of the natural gas pipeline, a new oxygen plant and a new syngas generation section were constructed at Sasol 1. The new synthesis gas generation section employed autothermal reformers to convert the natural gas into synthesis gas. The Lurgi gasifiers were kept in operation alongside the autothermal reformers for some time but has since been shut down. Although the coal gasification section was decommissioned, the tar workup section was kept in operation by importing product from the Sasol Synfuels facility (Chapter 9). By converting the feed source of the facility from a local coal source to an imported gas source, Sasol 1 lost its strategic importance for South Africa as an independent means of fuel and chemical production.
8.5 Evolution of the Sasol 1 Facility
8.5.2 Changes in Fischer–Tropsch Synthesis
The original design capacity of the Sasol 1 facility was equivalent to 6750 bbl/day. In current terms, even for the Fischer–Tropsch industry, this is a small-scale operation. As a consequence it became the test bed for new Fischer–Tropsch reactor technologies developed by Sasol. In the 1980s, a 1 m diameter demonstration unit was commissioned to test the Sasol Advanced Synthol (SAS) fixed fluidized bed technology and in 1990 the Sasol Slurry Bed Process (SSBP) [39]. The success of these two projects led to significant changes at all Sasol’s operations. A commercial-scale slurry bed reactor (Figure 8.11) was commissioned at Sasol 1 in 1993 [39]. The 5 m diameter slurry bed reactor was designed for a synthesis gas capacity of 182 000 m3 ·h−1 (normal) with on-line removal and addition of catalyst. It used a similar precipitated Fe-based Fischer–Tropsch catalyst as the Arge LTFT reactors. Typical operating conditions were 245 ◦ C and 2.1 MPa. During normal operation, the catalyst is suspended in the liquid-phase product (molten wax) through which the syngas is bubbled. As more wax is produced, liquid is drawn off and the Fischer–Tropsch catalyst is removed from the wax by filtration inside the reactor and secondary filtration outside the reactor. The product slate from the slurry bed process is different to that obtained with the fixed bed Arge reactors, although the catalyst essentially has the same formulation. The production of heavier material (>320 ◦ C) was increased from 50–55% for Arge to 55–60% for the SSBP (Table 8.9) and the product became more olefinic and more linear (Table 8.10) [40, 41]. The α-value of the Fe-LTFT SSBP product is 0.95. A typical H2 :CO usage ratio for the Fe-LTFT catalyst under slurry bed bubble column operation is around 1.55 [42]. Since a slurry bed reactor approximates a continuous stirred tank reactor (CSTR), the Fischer–Tropsch catalyst effectively sees only the tail gas composition. The sensitivity of reactor performance to the feed gas composition is dependent on the per pass conversion (Figure 5.1). The water partial pressure in the reactor is also determined by the per pass conversion. The water partial pressure influences the deactivation rate of the Fe-LTFT catalyst,
Tail gas Tail gas Internal recycle
Cold condensate Reaction water Quench column Hot condensate Filter Secondary filtration
Reactor wax
Synthesis gas
Figure 8.11
Slurry bed reactor
Slurry bed reactor at Sasol 1.
173
174
8 Sasol 1 Facility Table 8.9 Composition of Fe-LTFT syncrude from slurry bed synthesis at 240 ◦ C and 2 MPa.
Compound
Syncrude composition (mass%)a
Methane Ethene Ethane Propene Propane Butenes Butanes C5 –C6 C7 -160 ◦ C 160–350 ◦ C >350 ◦ C Aqueous product a
4 0.5 1 2.5 0.5 3 1 7 9 17.5 50 4
Composition depends on the catalyst age.
Table 8.10 Compound classes in the naphtha and distillate fractions of slurry bed Fe-LTFT syncrude.
Compound class
n-Alkanes iso-Alkanes Alkenes Aromatics Oxygenates
LTFT syncrude composition (%) C5 –C12
C13 –C18
28 1 64 0 7
42 2 50 0 6
and the H2 O to H2 + CO ratio should preferably be kept below 0.15. This implies that per pass conversion in an Fe-LTFT slurry bed reactor is determined by catalyst deactivation rate. The commissioning of the slurry bed reactor paved the way for converting Sasol 1 to a chemical production facility. The product slate could be simplified by having only LTFT synthesis. The original Kellogg HTFT synthesis reactors were decommissioned and the Fischer–Tropsch synthesis section became LTFT-only. 8.5.3 Changes in Fischer–Tropsch Refining
The production of chemicals in the Sasol 1 facility started in 1958 with the production of ammonium sulfate from the ammonia recovered from the gas liquor (Section 8.4.4). In the 1990s, when the Fe-LTFT slurry bed reactor was commissioned and the Kellogg CFB Fe-HTFT
8.5 Evolution of the Sasol 1 Facility Gascor town gas
Pure gas
Flare
Hydrogenation Dew point correction
−20 °C
Atmospheric distillation
Dew point condensate
Tail gas
Arge reactors SSBP reactor
C5 – C6 paraffins
H2
C7 – C8 paraffins
Paraffins
C9 – C10 paraffins
Cold condensate
C11– C13 paraffins
Fifth column
Waksol B
C14– C17 paraffins Waksol B
Hot condensate Waksol A
Reactor wax
H2 Waksol A
Vacuum distillation (four columns)
Waksol A Medium wax hydrogenation
Medium wax
Medium wax
M5
Hard wax Air hydrogenation
Oxidized wax Hydrogenated wax
H8 /H16
SPD 1&2
Hard wax SPD 4
C80 wax C105 wax
Reactor wax (SSBP only)
SPD 3
C80M wax C105M wax
Figure 8.12 Fischer–Tropsch oil refinery configuration of Sasol 1 in the mid-1990s after changing over from HTFT–LTFT synthesis to LTFT-only synthesis.
reactors were decommissioned, Sasol 1 became a chemical production facility (Figure 8.12). Many changes took place in the intervening years and the conversion into a chemicals refinery was not such an abrupt change as is suggested by a comparison of the original Sasol 1 design (Figures 8.7 and 8.8) with that after the conversion (Figure 8.12). Major changes took place in the 1960s after a decision in 1962 to expand production in the direction of chemicals. Some of these additions to the chemicals production capacity in the 1960s were [1, 35] as follows: 1) 2) 3) 4) 5)
The nitrogen from the air separation unit was converted into ammonia in an ammonia synthesis plant that was commissioned in 1963. Butadiene and styrene were produced and sold for synthetic rubber manufacturing. In 1964, Gaskor was founded to supply local industries with methane-rich gas and a pipeline to the steel works in Vanderbijlpark was completed in 1966. In 1966, the first naphtha cracker was built to produce ethene for Safripol (Afr. ‘‘Suid-Afrikaanse Poli-olefiene,’’ Engl. Transl. ‘‘South African Polyolefins’’). The first cracker was followed by a second cracker in 1969 to keep up with ethene demand for the production of high-density polyethylene (HDPE).
With all the chemicals extraction and production units, the product diversity from the Sasol 1 facility became impressive. It included motor-gasoline, diesel fuel, kerosene, fuel oils, LPG,
175
176
8 Sasol 1 Facility
bitumen, fuel gas, ethene, propene, butadiene, styrene, liquid N2 , liquid O2 , Ar, CO2 , various grades of waxes, oxidized waxes, C2 –C5 alcohols, acetone, butanone (MEK), creosote, tar acids (phenols, cresols, xylenols), aromatic solvents, aliphatic solvents (paraffins), and ammonia derivatives [1]. The construction of the Natref inland conventional crude oil refinery next to Sasol 1 created new fuels blending opportunities. On 5 July 1971, a new formulation of motor-gasoline was marketed [1]. The combination of crude-oil- and syncrude-derived fuels proved to be remarkably beneficial. It allowed the smaller volume of gasoline from Sasol 1 to be blended away in the crude oil gasoline [43], thereby avoiding further fuels-related refinery upgrading. During the commissioning of Sasol 2 and 3 (Chapter 9), a pipeline was constructed between Secunda and Sasolburg. With the large quantity of ethene that became available, it was no longer economically viable to keep the two naphtha crackers at the Sasolburg site operational and these were decommissioned in 1983. Demand for ethene kept on increasing. In 1988, one of the naphtha crackers was recommissioned as an ethane cracker, to convert ethane from Sasol 2 and 3 into ethene [35]. An explosives business was established in 1984 to add more value to the ammonia. A porous ammonium nitrate plant was commissioned in Sasolburg in 1985. The low crude oil price in the 1990s and the phasing out of TEL as motor-gasoline additive to improve octane number, contributed to the decision to phase out fuels production at Sasol 1 [45]. The conversion to an LTFT chemical facility focused heavily on the value of linear paraffinic material, and refining occurred mainly by hydrogenation and distillation. This is evident from a closer examination of the process flow diagram (Figure 8.12). The atmospheric distillation unit separated the lighter fractions into Waksol B and paraffins. The paraffins are not, as the name implies, exclusively n-alkanes; they are straight-run LTFT naphtha and distillate (Tables 8.4 and 8.10). The feed material to the atmospheric distillation unit was a combination of caustic-washed cold condensate (not shown in Figure 8.12, see Figure 8.8 for details), dew point condensate (product recovery from the tail gas by cooling down to −20 ◦ C), and the overhead product from the ‘‘fifth column.’’ The paraffins stream was hydrogenated to turn it into an n-alkane-rich product, which was separated into different carbon number cuts. These paraffinic cuts were sold to the solvent market. The Waksol B is a waxy oil and was sold to the fuels market. The ‘‘fifth column,’’ so called because it was the fifth column that formed part of the high-vacuum distillation section. It had a bottom temperature of 215 ◦ C and was operated at 14 kPa absolute pressure to avoid wax cracking. The Waksol A (or gatsch) overhead product was a waxy oil that was sold to the fuel market. The bottom product was a heavier wax product that was further purified in the high-vacuum distillation section. The bottom product from the fifth column was combined with the reactor wax to serve as feed to the four columns in the high-vacuum distillation section. In the high-vacuum distillation section, the wax was fractionated into waxes with different congealing points. The medium wax that was recovered as an overhead product from the second and third columns had a congealing point of 51 ◦ C and was combined with the overhead product from the fourth column that had a congealing point of 62 ◦ C. This combined medium wax product was hydrogenated over a Ni-based catalyst at 160 ◦ C and 2–3 MPa. The hydrogenated medium wax was sold to the market, typically for use in candle manufacturing.
8.6 Discussion of the Refinery Design
The bottom product from high-vacuum distillation was a hard wax, which had a slightly brown color. The wax was hydrogenated in two reactors in series over a Ni-based catalyst at 230 ◦ C and 4.5–5.5 MPa. These reactors were operated in a flooded up-flow mode, with hydrogen being added only to the first reactor. Part of the hydrogenated hard wax production was converted into various grades of oxidized waxes by partial air oxidation in batch reactors (Section 23.3). When the slurry bed reactor was installed, the wax distillation capacity became insufficient. A new section making use of short path distillation (SPD) was constructed. The SPD method of distillation is different from conventional vacuum distillation and it allowed the production of hard wax cuts with congealing points of 80 ◦ C (C80 and C80M waxes) and 105 ◦ C (C105 and C105M waxes). These columns were subsequently replaced by an even more efficient distillation configuration. Other chemical production facilities were added to the Sasol 1 portfolio to upgrade oxygenates recovered from the Fischer–Tropsch aqueous product. This included plants such as the synthesis of 4-methyl-2-pentanone (MIBK, methyl isobutyl ketone), and 4-methyl-2-pentanol (methyl isobutyl alcohol) from acetone over Pd on an acidic resin catalyst (Amberlyst CH28).
8.5.4 Changes in Coal Pyrolysis Product Refining
With the conversion of Sasol 1 from a CTL facility to a GTL facility, the tar products obtained by low-temperature gasification in Lurgi gasifiers were no longer produced. This would have implied closure of the tar processing facilities at Sasol 1, but it was decided to import material and keep these facilities operational. In fact, production was expanded with the commissioning of a new Tar Naphtha Phenol Extraction (TPNE) plant to increase tar acid production [46].
8.6 Discussion of the Refinery Design
The combination of HTFT and LTFT refineries in a single facility was not the result of engineering optimization, but it turned out to be serendipitous despite the limited level of integration. The Kellogg HTFT refinery mimicked the Hydrocol refinery (Chapter 7) and had similar success with the refining of HTFT syncrude to on-specification transportation fuels for the local market. The Arge LTFT refinery had some of the design elements found in the German Fischer–Tropsch refineries (Chapter 6). The ‘‘troublesome’’ waxy oil (slack wax) fraction was thermally cracked to produce more useful product fractions, and the light olefinic hydrocarbons were co-refined with the abundant C3 –C4 hydrocarbons from HTFT synthesis. The aqueous products were also co-refined to produce chemicals and an alcoholic blending component for the motor-gasoline. Some observations follow from each refinery section, the combined HTFT–LTFT refinery, and the subsequent changes to the refinery: 1) The clay-treater in the Kellogg oil workup section employed a more active and acidic catalyst than bauxite, which was the catalyst used by the Hydrocol facility for the same purpose. This turned out to be an unfortunate decision. HTFT syncrude is very reactive, and alumina-rich catalysts with their much milder acidity are more appropriate catalysts for syncrude refining
177
178
8 Sasol 1 Facility
2)
3) 4)
5)
6)
7)
8)
[44]. Unfortunately, this mistake was repeatedly made when the importance of less active catalysts for syncrude conversion was not appreciated (see Chapter 9). The decanted oil was steam stripped and not processed in the clay-treaters. This was a thoughtful approach for a number of reasons. The decanted oil is rich in aromatics and low in oxygenates; clay treating would have little benefit. The aromatics would have promoted coking of the catalyst, especially with the acidic catalyst used. The decanted oil also contained most of the metal impurities in the syncrude and processing it in the clay-treater would have resulted in bed plugging. The atmospheric distillation unit was the last step in HTFT syncrude refining. This avoided problems with boiling range broadening and the risk of thermal cracking in the reboiler. Although the refinery design only partially integrated the refining of HTFT and LTFT syncrude, there were definite benefits in combining these two Fischer–Tropsch processes. The syncrudes were complementary in flow rate and to a lesser extent in composition. The volume of HTFT residue was small compared to the total syncrude production. As a result, the refinery units processing HTFT residue were disproportionately small compared to the other refinery units. The same problem was found with the light hydrocarbons and oxygenates in the aqueous product from LTFT synthesis. By combining HTFT and LTFT syncrude, proper economy of scale could have been achieved in both light and heavy product upgrading. Fuels and chemicals were produced from the HTFT–LTFT refinery. With the design employed for the Sasol 1 refinery, the HTFT syncrude was better suited for fuels refining, whereas the LTFT syncrude was better for chemicals refining. The original Sasol 1 refinery design established conversion pathways for alkanes, alkenes, and oxygenates in the Fischer–Tropsch refinery, as well as refining pathways for the coal pyrolysis products. In later years, when the facility was converted from HTFT–LTFT to LTFT-only and from CTL to GTL, the refinery design no longer exploited the synthetic value of the alkenes. Although the production volume of alkenes decreased, it is not clear why alkene refining was abolished. This design decision is even more surprising when considering that coal pyrolysis product refining was retained and expanded, even though coal pyrolysis products were no longer being produced at the Sasol 1 site at all. From a Fischer–Tropsch refining perspective, the decision to abolish alkene refining was a step in the wrong direction; it reduced the efficiency of the facility. Although the product slate of the current Sasol 1 refinery suggests that it is a chemicals-only facility, it is not entirely true. Over time, the facility changed from a stand-alone synthetic fuels and chemical refinery to a chemical refinery that is reliant on other facilities to upgrade part of the LTFT syncrude. Indirectly it still produced fuels, but as lower value synthetic crude oil intermediates that are co-refined with conventional crude oil in order to produce transportation fuels. It seems that LTFT refining is reliant on blending or co-refining with other material in order to process all of the LTFT syncrude into final products. Three examples to support the observation can be cited from the facilities discussed thus far. All of the German Co-LTFT syncrude could be converted into final products by blending it with coal liquids (Sections 6.4.2 and 6.4.3). All of the Arge Fe-LTFT syncrude could be converted into final products by co-refining it with HTFT syncrude. All of the Arge and SSBP Fe-LTFT syncrude could be converted into final products by co-feeding part of the LTFT syncrude with crude oil in a conventional oil refinery.
References
References 1. Meintjes, J. (1975) Sasol 1950–1975, Tafelberg, 2.
3.
4.
5. 6.
7.
8.
9.
10.
11.
12.
13.
14. 15.
16.
Cape Town. Dry, M.E. (1983) in Applied Industrial Catalysis, vol. 2 (ed. B.E. Leach), Academic Press, New York, pp. 167–213. Hoogendoorn, J.C. and Salomon, J.M. (1957) Sasol: World’s largest oil-from-coal plant. I. Br. Chem. Eng., 238–244. Wainwright, H.W., Kane, L.J., Wilson, M.W., Shale, C.C., and Ratway, J. (1956) Purification of synthesis gas. Removal of dust, carbon dioxide, and sulfur compounds. Ind. Eng. Chem., 48, 1123–1133. Hochgesand, G. (1970) Rectisol and Purisol. Ind. Eng. Chem., 62 (7), 37–43. Weiss, H. (1988) Rectisol wash for purification of partial oxidation gases. Gas Sep. Pur., 2 (4), 171–176. Weil, B.H. and Lane, J.C. (1949) The Technology of the Fischer-Tropsch Process, Constable, London. Dry, M.E. (1981) in Catalysis Science and Technology, vol. 1 (eds J.R. Anderson and M. Boudart), Springer-Verlag, Berlin, pp. 159–255. Hoogendoorn, J.C. (1973) Experience with Fischer-Tropsch synthesis at Sasol. Clean Fuels Coal Symp., 353–365. Hoogendoorn, J.C. and Salomon, J.M. (1957) Sasol: World’s largest oil-from-coal plant. II. Br. Chem. Eng., 308–312. Garrett, L.W. Jr. (1960) Gasoline from coal via the Synthol process. Chem. Eng. Prog., 56 (4), 39–43. Steynberg, A.P., Espinoza, R.L., Jager, B., and Vosloo, A.C. (1999) High temperature Fischer-Tropsch synthesis in commercial practice. Appl. Catal. A, 186, 41–54. Holtkamp, W.C.A., Kelly, F.T., and Shingles, T. (1977) Circulating fluid bed catalytic reactor for the Fischer-Tropsch synthesis at Sasol II. ChemSA, 44–45. Dry, M.E. (2004) FT catalysts. Stud. Surf. Sci. Catal., 152, 533–600. Hoogendoorn, J.C. (1975) New applications of the Fischer-Tropsch process. Clean Fuels and Coal Symposium II, pp. 343–358. Espinoza, R.L., Steynberg, A.P., Jager, B., and Vosloo, A.C. (1999) Low temperature Fischer-Tropsch synthesis from a Sasol perspective. Appl. Catal. A, 186, 13–26.
17. Gregor, J.H. (1990) Fischer-Tropsch products
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
as liquid fuels or chemicals. An economic evaluation. Catal. Lett., 7, 317–332. Hoogendoorn, J.C. and Salomon, J.M. (1957) Sasol: World’s largest oil-from-coal plant. III. Br. Chem. Eng., 368–373. Armistead, G. Jr. (1946) Modern refining processes. 5. Nonselective catalytic polymerization has important postwar utility. Oil Gas J., 44 (48), 131–136. Steffens, J.H., Zimmerman, M.U., and Laituri, M.J. (1949) Correlation of operating variables in catalytic polymerization. Chem. Eng. Prog., 45 (4), 269–278. Asinger, F. (1968) Mono-Olefins Chemistry and Technology, Pergamon, Oxford. Bolder, F.H.A. (2007) Fischer-Tropsch wax hydrogenation over a sulfided nickel-molybdenum catalyst. Energy Fuels, 21, 1396–1399. Bell, H.S. (ed.) (1959) American Petroleum Refining, 4th edn, D. Van Nostrand, Princeton, NJ. pp. 308–344. Le Roux, J.H. (1969) Fischer-Tropsch waxes. I. An infra-red method for the determination of crystallinity. J. Appl. Chem., 19, 39–42. Le Roux, J.H. (1969) Fischer-Tropsch waxes. II. Crystallinity and physical properties. J. Appl. Chem., 19, 86–88. Le Roux, J.H. (1969) Fischer-Tropsch waxes. IV. Qualitative and quantitative investigation of chain branching. J. Appl. Chem., 19, 230–234. Le Roux, J.H. (1970) Fischer-Tropsch waxes. V. Study of branching and its effect on crystallinity using an improved infra-red method. J. Appl. Chem., 20, 203–207. Le Roux, J.H. and Dry, L.J. (1972) Fischer-Tropsch waxes. VI. Distribution of branches and mechanism of branch formation. J. Appl. Chem. Biotechnol., 22, 719–726. Henley, E.J. and Seader, J.D. (1981) Equilibrium-Stage Separation Operations in Chemical Engineering, John Wiley & Sons, Inc., New York, pp. 101–102. Hoogendoorn, J.C. and Salomon, J.M. (1957) Sasol: World’s largest oil-from-coal plant. IV. Br. Chem. Eng., 418–419. Department of Water Affairs, (1986) Bestuur van die Waterhulpbronne van die Republiek van Suid-Afrika (Engl. Transl. ‘‘Management of the water resources of the Republic of South Africa’’), Government printer, Pretoria, p. 3.28.
179
180
8 Sasol 1 Facility 32. Wilson, P.J. Jr. and Wells, J.H. (1945) in Chem-
33.
34. 35.
36.
37.
38. 39.
istry of Coal Utilization, (ed. H.H. Lowry), John Wiley & Sons, Inc., New York, pp. 1371–1481. Hoogendoorn, J.C. (1982) Producing automotive fuels from coal in South Africa. Hydrocarbon Process., 61 (5), 34E–34Q. (2009) Sasol will double synthetic wax. Chem. Eng. News, 87 (49), 23. Collings, J. (2002) Mind Over Matter. The Sasol Story: A Half-Century of Technological Innovation, Sasol, Johannesburg. Mashapa, T.N., Rademan, J.D., and Janse van Vuuren, M.J. (2007) Catalytic performance and deactivation of precipitated iron catalyst for selective oxidation of hydrogen sulfide to elemental sulfur in the waste gas streams from coal gasification. Ind. Eng. Chem. Res., 46, 6338–6344. Dancuart, L.P. and Steynberg, A.P. (2007) Fischer-Tropsch based GTL technology: a new process? Stud. Surf. Sci. Catal., 163, 379–399. Higman, C. and Van der Burgt, M. (2008) Gasification, 2nd edn, Elsevier, Amsterdam. Steynberg, A.P. (2004) Introduction to Fischer-Tropsch technology. Stud. Surf. Sci. Catal., 152, 1–63.
40. Jager, B. and Espinoza, R. (1995) Advances in
41.
42.
43. 44.
45.
46.
low temperature Fischer-Tropsch synthesis. Catal. Today, 23, 17–28. Dry, M.E. (2004) Chemical concepts used for engineering purposes. Stud. Surf. Sci. Catal., 152, 196–257. Janse van Vuuren, M., Huyser, J., Grobler, T., and Kupi, G. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 229–241. Dry, M.E. (1982) Hydrocarbon Process., 61 (8), 121–124. De Klerk, A. (2010) in Catalysis, vol. 23 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 331–364. Lund, G., Davis, W., Coogan, E., De Wit, G., and Naidoo, R. (2006) Extraction of tar acids from naphtha streams: Tar Naphtha Phenolic Extraction (TNPE) plant. Chem. Technol. (S. Afr.), 10–12.
181
9 Sasol 2 and 3 Facilities 9.1 Introduction
The Sasol 2 and 3 facilities are products of the 1973 ‘‘Oil Crisis.’’ On 30 November 1973, the South African government requested Sasol to investigate the technical and economic viability of building one or more coal-to-liquid facilities in the Republic of South Africa. The study was completed by early January 1974, and on 3 December 1974 the intent to construct a new coal-to-liquids facility, Sasol 2, was publicly announced. In December 1978, the South African government repeated the study request. A report detailing the cost of doubling Sasol 2, then still known as the Sasol 2 expansion project, was provided to the government in January 1979. The decision to proceed with Sasol 3 was taken in February 1979 and was partly motivated by the transitions that happened in Iran, when the Shah of Iran fled the country [1]. The intent with both facilities was clear; it was a means to provide energy security for South Africa. Sasol 2 and 3 were grassroots facilities just like Sasol 1. These two facilities were erected adjacent to each other at the place that is presently called Secunda. Secunda is about 100 km east of Johannesburg. Construction of Sasol 2 started in 1976, and commissioning of the steam plant took place early in 1979. The synthesis and refining sections of Sasol 2 were commissioned in 1980, and the facility produced its first hydrocarbon product on 1 March 1980. The first marketable products were produced on 25 April 1980. The last Fischer–Tropsch synthesis train was only commissioned in January 1981, and Sasol 2 reached design production capacity in 1982. Construction of Sasol 3 started in 1979 already, and took place alongside that of Sasol 2 and was aided tremendously by the activities and infrastructure already in place. Commissioning of Sasol 3 took place in parallel with the startup of Sasol 2 and already commenced in 1981, with the synthesis and refinery sections being commissioned in 1982. Sasol 3 produced its first hydrocarbons on 10 May 1982 and first marketable products by 1 July 1982. Design production capacity was reached at Sasol 3 on 16 February 1983, only a year after Sasol 2. As in the case of Sasol 1 (Chapter 8), the Sasol 2 and 3 facilities in Secunda are still in operation. The original designs of Sasol 2 and 3 are very similar. Minor improvements were made to units in Sasol 3 on the basis of lessons learnt from Sasol 2, and these will be pointed out only when they led to significant differences between the two facilities. Over the years, both facilities saw many changes and the facilities will likely continue to change. The discussion is therefore given Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
182
9 Sasol 2 and 3 Facilities
in two parts. The original Sasol 2 and 3 design is discussed first (Sections 9.2–9.4), which is then followed by a discussion of the changes that took place (Section 9.5).
9.2 Synthesis Gas Production 9.2.1 Lurgi Dry Ash Coal Gasification
Much has been learnt from the Sasol 1 operation about the conversion of South African high ash coals in Lurgi dry ash gasifiers. The same type of gasifier was consequently selected for the Sasol 2 and 3 facilities. Since the commissioning of the Sasol 1 plant, many improvements were made to the Lurgi gasifier design and, by the time Sasol 2 was constructed, an improved Lurgi Mark IV test gasifier was already in operation at the Sasol 1 site. The operation of a Lurgi Mk IV gasifier since 1978 at Sasol 1 allowed sufficient time for optimization as well as the commissioning of these gasifiers at Sasol 2 to proceed without significant delay [2]. At total of 80 Lurgi Mark IV gasifiers were installed at Secunda, 40 per facility. Each gasifier was 3.85 m in diameter and was able to processes about 1000 tons of coal per day. This yielded 55 000 m3 ·h−1 (normal) raw synthesis gas [2–4]. Since the Lurgi gasifiers are low-temperature gasifiers, coal pyrolysis products are coproduced with the synthesis gas. The separation methodology followed to separate the raw synthesis gas from these pyrolysis products was very similar to that employed in the Sasol 1 design (Figure 8.1). The design was more involved (Figure 9.1), with additional coolers and tar separators forming part of the design. Tar filtration literally became the black sheep of this operation. 9.2.2 Synthesis Gas Cleaning
Purification of the raw gas from the coal gasification section was performed by physical absorption of CO2 and H2 S in cold methanol in a Rectisol process. This process has already been described in some detail (Section 8.2.2). Because of the size of the Secunda facility, more than one Rectisol train was built. The size of the Rectisol trains was limited by the maximum size of the vessels that could be transported by road [2]. The practice of releasing the CO2 and H2 S directly into atmosphere, as in the Sasol 1 facility, was no longer acceptable. An environmentally more responsible way of dealing with the H2 S emissions had to be incorporated in the design. The concentration of H2 S in the CO2 was low and a Claus-type sulfur recovery unit could not be considered for sulfur removal. The H2 S from the Rectisol units were converted into elemental sulfur and recovered in a Sulfolin process. The sulfur recovery units in the original design for Sasol 2 and 3 were for technical reasons not commissioned and were modified after construction to become Sulfolin units [2]. These units were commissioned in 1986. Sulfolin is a modified Stretford process that was developed by Linde to resolve the incompatibility issues previously experienced with the standard Stretford process (see Section 8.5.1) when
9.3 Fischer–Tropsch Synthesis
Gasification
Gas cooling
Waste Preheat boiler cooler Wash cooler
Trim cooler
180 °C
45 °C
Air coolers 160 °C
Lurgi Mk IV gasifier
Tarry gas liquor coolers
Dusty gas liquor coolers
90 °C
Recycle
Tar
35 °C
Oil separator
Mud liquor
Secondary tar separator
Mud liquor Tertiary tar separator
(from Phenosolvan)
Gas liquor
Gas liquor
Gas liquor separation
Coal tar filtration Figure 9.1
Raw syngas
Oily gas liquor coolers Oil
Primary tar separator
Final cooler
Crude tar
Synthesis gas production by low-temperature coal gasification in Lurgi dry ash gasifiers.
it was used with Rectisol off-gas after low-temperature coal gasification. The modified process uses the same basic chemistry. In a Stretford-type process, the H2 S is absorbed in a sodium carbonate (Na2 CO3 ) solution (Equation 9.1). In this solution, the H2 S is oxidized by vanadium(V) to produce elemental sulfur (Equation 9.2). The elemental sulfur can be recovered from the solution. The vanadium(IV) formed during this step is converted back to vanadium(V) by reoxidation with air (Equation 9.3) to make it a catalytic process [5, 6]. H2 S + Na2 CO3 (aq) NaHS(aq) + NaHCO3 (aq)
(9.1)
2NaSH(aq) + 4NaVO3 (aq) + H2 O → Na2 V4 O9 (aq) + 4NaOH(aq) + 14 S8 ↓
(9.2)
Na2 V4 O9 (aq) + 2NaOH(aq) + O2 → 4NaVO3 (aq) + H2 O
(9.3)
9.3 Fischer–Tropsch Synthesis
The Fischer–Tropsch synthesis section employed high-temperature Fischer–Tropsch (HTFT) conversion exclusively. The circulating fluidized bed (CFB) design of the Kellogg reactors (Figure 8.3) formed the basis of the new reactors, called Sasol Synthol reactors. The CFB design was optimized, and many of the design problems of the original Kellogg reactors were sorted out [7]. The new CFB design was also larger and produced the equivalent of 7500 bbl/day of product [8]. Typical operating conditions of the Synthol reactors were 330–360 ◦ C and 2.5 MPa, which was a
183
184
9 Sasol 2 and 3 Facilities Table 9.1 Typical composition of the syncrude from HTFT synthesis in Sasol Synthol circulating fluidized bed reactors.
Compound
Methane Ethene Ethane Propene Propane Butenes Butanes C5 –C6 alkenes C5 –C6 alkanes C7 -160 ◦ C 160–350 ◦ C >350 ◦ C >525 ◦ C Aqueous product Nonacid chemicals Carboxylic acids
Syncrude composition (mass%) [11]
[12]
[13]
[14]
11 4 3.4 11.4 1.4 9.3 2 12.8 2.6 17.2 13 5.4 –
10 4 4 12 2 9 2 51b – – – – –
10 4 3.5 12 2 9 2 – – 37c 11c 3c 0.5
11 7.5a – 13a – 11a – 51.5b – – – – –
6.5d –
5 1
6d –
6d –
a
Alkanes and alkenes combined. C5 and heavier product. c Source ranges are C -190, 190–400, and 400–525 ◦ C. 5 d Total nonacid and acid chemicals in the aqueous product. b Total
slightly higher pressure than the Kellogg CFB reactors [9]. A total of eight Synthol reactors were installed at Sasol 2 and the same number at Sasol 3, giving the two facilities a combined design production capacity of 120 000 bbl/day crude oil equivalent. The fused-iron based Fischer–Tropsch catalyst itself also harked back to the original catalyst for the Kellogg reactors. The catalyst was not the same though. Over the years, the catalyst formulation has been improved as the effect of various promoters and manufacturing technique were better understood, although the basic formulation remained very similar. The state of the art at that time was documented by Dry [10]. Typical product distributions from commercial Synthol HTFT synthesis is given in Table 9.1 [11–14]. The average composition from Synthol HTFT and Kellogg HTFT was very similar, since the catalyst and operation of these units were alike, albeit not the same. The syncrude composition was also dependent on the average Fischer–Tropsch catalyst age, which was be controlled by on-line addition and removal of catalyst from the reactors. 9.3.1 Gas Loop Design
A closed gas loop with internal recycle, external recycle, and cryogenic separation was employed to maximize carbon efficiency in the gas loop (Figure 9.2).
9.3 Fischer–Tropsch Synthesis H2 PSA Purge gas
CH4
Cold separation Methane reformer
CO2
Third stage Second stage cooling
H2
Cryogenic separation
cooling First stage Dryer cooling
C2 rich gas Propene and Condensate 3 Condensate 2
Benfield
Condensate 1 Washed tail gas
External recycle
HTFT synthesis
Water
Tail gas wash
Internal recycle
Synthol HTFT Air cooler
Water cooler
Wash water
Tail gas
Light oil Aqueous product
Pure gas
Decanted oil
Figure 9.2
Decanted oil
Gas loop design of Sasol 2 and 3.
The pure synthesis gas from Rectisol syngas cleaning was mixed with the internal and external recycle streams to produce the feed to HTFT synthesis. The Fischer–Tropsch catalyst is water gas shift active and the feed composition was controlled on the basis of the Ribblett ratio (Equation 4.16), which was kept close to 1. The actual H2 :CO ratio at the inlet of the synthesis reactors was around 4 : 1 and the CO conversion was very high. The internal recycle consisted of the tail gas produced after knocking out the products that could be condensed by water cooling. These condensed products were sent to the refinery. Hydrogen from the pressure swing absorption (PSA) unit and the products from methane reforming with oxygen over a nickel catalyst were combined to constitute the external recycle. To enable the recovery of hydrogen and methane, the gas loop included a ‘‘cold box’’ where these compounds were cryogenically separated from the heavier hydrocarbons. The inclusion of a CO2 removal step in the tail gas processing was essential, since CO2 solidifies at the cryogenic conditions needed to separate H2 and C1 –C2 hydrocarbons. The CO2 removal was performed in a Benfield unit. The Benfield CO2 removal process is a chemical absorption process. It employs a hot potassium carbonate wash. The CO2 is absorbed into the water to form carbonic acid (H2 CO3 ) and is limited by the solubility of CO2 in the water (Equation 9.4). The absorption equilibrium is continuously driven to the right by the acid–base reaction between the carbonic acid and the basic potassium carbonate (K2 CO3 ), which consumes the carbonic acid as it is formed (Equation 9.5) [15, 16]. CO2 + H2 O H2 CO3 (aq)
(9.4)
H2 CO3 (aq) + K2 CO3 (aq) 2KHCO3 (aq)
(9.5)
185
186
9 Sasol 2 and 3 Facilities
The tails gas from Benfield was scrubbed in an amine unit. The tail gas was dried by passing the gas over a molecular sieve before it entered cryogenic separation [2]. Since both the Benfield hot carbonate wash and the amine unit contained bases, it was necessary to achieve complete removal of oxygenates from the tail gas during the washing step. The detrimental effect of oxygenates on such units has been documented [17]. The cold box flow diagram shown in Figure 9.2 is simplified considerably. The heat integration found in such units is quite complex [18]. A large portion of the light products from HTFT synthesis was condensed by stepwise cooling in the ‘‘cold box.’’ The light products were not separated in carbon number fractions during cryogenic separation. Some propene as well as the C2 and lighter products were separated though. The heavier hydrocarbons were condensed as mixtures and sent to the refinery as the Condensate 1, 2, and 3 streams. Since the Condensate 1, 2, and 3 streams were produced by progressive cooling, the composition of Condensate 1 (C3 –C7 ) was such that it contained heavier hydrocarbons than Condensate 2 (C4 -rich), which in turn contained heavier hydrocarbons than Condensate 3 (C3 -rich). The ethene and some propene that were separated as pure compounds were used as feed to produce polymers. The methane was recycled to the oxidative methane reformer where it was converted into synthesis gas. The synthesis gas from the methane reformer together with some H2 -rich gas from cryogenic separation formed the external recycle in the gas loop. Not all of the H2 -rich gas was recycled. Some H2 was recovered from this stream by pressure swing absorption (PSA) for use in the refinery. The remainder formed the purge gas through which the inert gases (mainly N2 and Ar) in the gas loop was purged together with some hydrogen.
9.4 Fischer–Tropsch Refining
The Sasol 2 and 3 refineries each consisted of four different refinery sections (or value chains as they are called in Sasol), namely condensate workup, oil workup, chemical workup, and tar workup (Figure 9.3). The organic phase was refined in the condensate and oil workup sections and the aqueous phase was refined in the chemical workup section. Together, these three sections constituted the Fischer–Tropsch refinery. The organizational separation of the condensate workup and oil workup sections is indicated (Figure 9.4). In the original design, the separation was quite clear and these sections are discussed separately. In terms of feed streams, it relied heavily on the stepwise cooling in the Fischer–Tropsch gas loop to separate the material going to each section. This separation was not perfect and some of the light material that was present in the unstabilized light oil (ULO) was removed during light oil fractionation and transferred to the condensate workup section. Although a single column in Figure 9.4 represents the fractionation of light oil, the design of oil fractionation was more complex. The ULO from the Fischer–Tropsch gas loop was stabilized (debutanized) to produce stabilized light oil (SLO). The SLO is also sometimes referred to as Synthol light oil and the stabilized oil was fractionated to produce different SLO cuts. The condensate and oil workup sections were designed to meet the following objectives [19]: 1) Convert normally gaseous C3 and C4 alkenes into liquid range products.
9.4 Fischer–Tropsch Refining
NH3 synthesis
Air separation
N2 H2
Gas liquor O2
Gasification
Tar workup
Neutral oil
O2
Tar
PSA Rectisol
Methane reforming
Purge gas
CH4
Syngas
Naphtha Sulfolin
H2 S + CO2
Syngas S8
H2 rich Cryogenic separation
Benfield
Synthol HTFT
Oil Condensate 1, 2, and 3 Condensate workup
Aqueous product
CO2
Chemical workup
CO2 Wastewater
C5 – C6
Oil workup
Figure 9.3 Block flow diagram of the original Sasol 2 and 3 facilities showing the feed origin and flow to the various refinery sections.
2) Remove contaminants from Synthol oil to protect downstream catalysts and yield suitable products. 3) Upgrade the quality of the motor-gasoline to meet octane specifications. The Fischer–Tropsch oil refinery design incorporated many design elements that were typical of a third-generation crude oil refinery (Section 2.4.3). This was a significant departure from previous Fischer–Tropsch refinery designs (Chapters 6–8) that took more cognizance of the syncrude properties. Despite the increase in refining complexity for the Sasol 2 and 3 designs, the fuel properties that could be achieved by a crude oil refining approach was mediocre (Table 9.2) [20, 21]. It can be retrospectively argued that, with more than two decades of HTFT refining experience, Sasol might have known better than to adopt a crude oil refinery design. The real impact of the decision was not apparent then, since the crude oil refinery design adopted for Sasol 2 and 3 produced refined transportation fuels that met South African 1980 fuel specifications (Section 9.4.5). One aspect of the refinery design that may not be apparent from Figure 9.4 was the tremendous flexibility to change the motor-gasoline to diesel fuel production ratio. It has been reported that by changing the cut-point temperatures and altering the operation of alkene oligomerization and hydrocracking (hydrodewaxing), the motor-gasoline to diesel fuel ratio could be changed from around 10 : 1 to 1 : 1 [3].
187
188
9 Sasol 2 and 3 Facilities
Condensate workup
LPG / butane
Catalytic polymerization (oligomerization)
Condensate 1, 2, and 3
C5 –C6
Motor-gasoline Olefin hydrotreater
Isomerization
Heavy alcohols
Motor-gasoline Illuminating paraffin/ light diesel fuel Motor-gasoline
(from chemical workup)
Oil workup C5 –C6
Light oil
C7 –C10
Naphtha hydrotreater
Catalytic reformer
Motor-gasoline Heavy naphtha
>C10
Diesel fuel Decanted oil
LVGO
Distillate hydrotreater
Distillate hydrocracker
Motor-gasoline
HVGO Waxy oil
Fuel oil
Figure 9.4 Original Sasol 2 and 3 oil refinery design, which was organized into condensate workup and oil workup sections.
As noted during the discussion of the original Sasol 1 refinery design (Section 8.4), the tar workup section refined the products from low-temperature coal pyrolysis that were separated from the raw synthesis gas. The tar refinery section is strictly speaking not part of the Fischer–Tropsch refinery. The products from the tar refinery were blended with the Fischer–Tropsch-derived synthetic fuels. Integration with the tar refining section occurred only during fuel blending. A tar refinery will not be required if a high-temperature coal gasification technology is used, or when natural gas is used as feed. 9.4.1 Synthol HTFT Condensate Refining
In the original design the condensate workup section was responsible for refining the C3 –C6 hydrocarbons. Most of this material was condensed from the tail gas during cold separation (Figure 9.2). The condensate streams thus obtained contained little oxygenates and was rich in alkenes. The condensate workup section had design features that were similar to those of the Hydrocol refinery (Section 7.4) and the Kellogg oil refinery at Sasol 1 (Section 8.4.1). The C3 –C4 material was oligomerized to produce heavier liquid hydrocarbons, and the C5 –C6 material was isomerized
9.4 Fischer–Tropsch Refining Density, research octane number (RON), motor octane number (MON), and cetane number from key streams in the original Sasol 2 and 3 Fischer–Tropsch oil refinery.
Table 9.2
Sasol 2 and 3 refinery product
Motor-gasoline Olefinic SPA oligomers Hydrogenated SPA oligomers Hydrogenated SPA <150 ◦ C naphtha C5 –C6 straight-run naphtha C5 –C6 isomerate Reformate (RON 81 severity) Reformate (RON 87 severity) Hydrocracker naphthaa Diesel fuel Hydrogenated SPA oligomers Hydrogenated straight-run distillateb Residue hydrodewaxing/crackinga
Selected fuel properties Density (kg·m−3 )
RON
MON
Cetane number
745 730 700 680 680 765 770 720
95–96 64 75 83 88 81 87 77
81–82 70 72 72 77 75 80 71
– – – – – – – –
765 810 860
– – –
– – –
34 55 67
a Distillate b
selective cracker (DSC). Distillate hydrotreater (DHT).
and deoxygenated to improve its motor-gasoline quality. The conversion strategy was similar, but it incorporated some changes that affected the outcome significantly and for the worse: 1) Oligomerization of the C3 –C4 alkenes was followed by a hydrogenation unit to hydrogenate the kerosene range oligomers as well as part of the naphtha range oligomers (Figure 9.5). The latter was required in order to reduce the olefinicity of the motor-gasoline produced at the facility. A target value of maximum 30% alkenes in the motor-gasoline was set on the basis of road tests [19] although this was not a legislated requirement. 2) Only the C5 –C6 material was slated for catalytic isomerization and deoxygenation in a process analogous to bauxite treatment and clay treatment. The unit was not designed to process the complete oil product, and feed was restricted to C5 –C6 material. This feed restriction is typical of catalytic isomerization units to limit cracking. Yet, the amount of C7 and heavier material entering the unit was controlled by the separation efficiency in the gas loop and the operation of the light oil fractionation in the oil workup section. Neither of these separation steps yielded sharp carbon number cuts. 3) A more modern and acidic catalyst was selected for this unit. This catalyst selection was made despite the reactive nature of the feed, which consisted mainly of alkenes and oxygenates, with <20% alkanes. The catalyst was too active for the feed. The alkene oligomerization unit took its feed from the Condensate 1, 2, and 3 streams. It was fractionated in a feed debutanizer column, with the C3 –C4 overhead fraction being used as feed for oligomerization and the C5 –C6 bottom fraction being sent to the isomerization unit. Universal Oil Product (UOP) Catalytic Polymerization technology was selected for alkene oligomerization.
189
190
9 Sasol 2 and 3 Facilities
>C2 material
Water injection
To LPG recovery
Oligomerization reactors
(ethylene column)
Condensate 3
Product debutanizer
Condensate 2 Light naphtha
Water
(Overheads from SLO splitter)
C5 – C6 naphtha (to isomerization)
Water LPG quench
Feed debutanizer H2
H2
To DHT
H2
Fuel gas Recycle H2
Motorgasoline
Hydrogenation reactors Diesel fuel Olefinic feed from oligomerization
Figure 9.5 Alkene oligomerization and oligomer hydrogenation units in the condensate workup section. No naphtha recycle or separation of olefinic motor-gasoline is shown.
Oligomerization was catalyzed by a solid phosphoric acid (SPA) catalyst (Section 19.3.1). The feed water content was adjusted by water injection to control the SPA catalyst hydration level. This practice was later discontinued. The oligomerization unit (Figure 9.5) consisted of two trains, each with four reactors having five catalyst beds with interbed quenches. The oligomerization reaction is highly exothermic, and part of the liquid petroleum gas (LPG) fraction of the feed (propane and butane) was recycled to keep the temperature rise over the reactor within design parameters. At Sasol 2, the oligomerization units were configured in such a way that it was possible to recycle part of the motor-gasoline fraction to boost overall distillate make. This was not done at Sasol 3, and the oligomerization units therefore vary in design and operation. The unhydrogenated motor-gasoline has good octane numbers (Table 9.3) [19], which is fairly insensitive to variations in feed composition. The same is not true for the hydrogenated motor-gasoline from SPA oligomerization, which is very sensitive to the feed composition [22]. The octane number of the hydrogenated motor-gasoline was considerably lower (Table 9.3). The distillate range oligomers, as well as part of the naphtha range oligomers, were hydrogenated. The alkene hydrogenation unit, called the Polyhydrotreater, consisted of three reactors (Figure 9.5). Each reactor was loaded with a single bed of sulfided Co/Mo–Al2 O3 catalyst and was operated in the range 260–350 ◦ C and 5 MPa. This resulted in almost complete saturation of the alkenes in the feed material. The quality of the distillate was improved, but neither the motor-gasoline nor the diesel fuel was of high quality (Table 9.3).
9.4 Fischer–Tropsch Refining Selected fuel properties of the unhydrogenated (olefinic) and hydrogenated naphtha and distillate fractions from oligomerization of C3 –C4 HTFT syncrude over solid phosphoric acid.
Table 9.3
Fuel property
Naphtha Olefinic
Density (kg·m−3 ) Alkene content (g Br/100 g) RON RON + 3 ml TEL/gala MON MON + 3 ml TEL/gala Cetane number Distillation (◦ C) T10 T50 T90 a Tetraethyl
Distillate (kerosene) Hydrogenated
Olefinic
Hydrogenated
740.5 134 96.3 100.3 82.3 87.0 –
746.7 131 94.5 99.9 80.9 85.3 –
729.4 1.3 63.7 86.8 70.6 88.8 –
777.4 108 – – – – 20.9
764.4 0.99 – – – – 34.4
103 134 176
114 139 180
109 137 178
186 189 201
182 192 203
lead (TEL) addition: 1 ml TEL/gal = 0.44 g l−1 .
The C5 –C6 condensates, C5 –C6 light oil from atmospheric distillation in the oil workup section, and heavy alcohols from the chemical workup section were combined to form the feed to the C5 –C6 isomerization unit. The objective was to convert the oxygenates into alkenes and to isomerize the alkenes in order to increase the octane number of the product. The isomerized product was used directly as a blending component in motor-gasoline (Table 9.2). The process employed two single bed reactors in parallel that were operated in a swing mode. While the one reactor was converting C5 –C6 feed material, the other reactor was regenerated by controlled coke burn-off. The unit was operated at 370–410 ◦ C and close to atmospheric pressure. The isomerization catalyst, HZ-1, was a rare-earth-exchanged Y-zeolite in a silica–alumina matrix. The reason for the selection of HZ-1 over an alumina-rich catalyst with its proven performance for this type of conversion is not clear. The decision turned out to be disastrous. The HZ-1 was too acidic and too active to be employed with such a reactive Fischer–Tropsch feed. The catalyst resulted in excessive cracking and was rapidly deactivated. Ultimately, the unit was shut down and converted into a dehydration unit. The original design intent was clear, but it failed on two accounts. Firstly, the catalyst selected for this application was ill suited for the conversion of Fischer–Tropsch syncrude and a much milder catalyst, such as alumina, would have been a better choice. It indicated a lack of appreciation of the differences between the C5 –C6 fraction derived from Fischer–Tropsch syncrude and an equivalent cut derived from conventional crude oil. Secondly, the process failed because the feed was not limited to C5 –C6 naphtha. Material that is C7 or heavier would not only have cracked readily but also increased the rate of deactivation by coking. Both effects were observed, because the limitation on C7 and heavier material was not appreciated. Separation in the gas loop and the oil workup section was not tightly controlled to limit the content of C7 and
191
192
9 Sasol 2 and 3 Facilities
heavier material in the condensate as well as the overhead product from the SLO splitter in the oil workup section. 9.4.2 Synthol HTFT Oil Refining
The oil workup section (Figure 9.4) was designed purely along the lines of a third-generation crude oil refinery. An atmospheric distillation unit fractionated the SLO into naphtha and distillate fractions. The naphtha was processed in a naphtha hydrotreater (NHT) followed by a catalytic reformer (Platformer), and the distillate was processed in a distillate hydrotreater (DHT). The atmospheric residue was vacuum distilled and processed in a distillate-selective cracker (DSC). The reboiler temperature in the atmospheric distillation column was limited to 300–320 ◦ C to avoid thermal cracking of the syncrude being distilled. This made proper separation between distillate and residue fractions difficult. The situation was exacerbated by the small residue fraction, which would have resulted in a bottom product with very low flow rate. In order to balance column operation, more material was included in the bottom product. This resulted in a very inefficient overall separation, which was made worse by the boiling point broadening that occurred in the downstream units. The separation strategy made little sense in terms of the downstream refinery requirements. It was almost as if the refinery design had to include up-front atmospheric and vacuum distillation units just because it was standard practice in crude oil refinery design. It will become apparent from the subsequent discussion that this approach failed to realize the same benefits as in crude oil refining. The C7 –C10 fraction from atmospheric distillation of the SLO was used as feed material for catalytic reforming to increase its octane number. The feed was pretreated in an NHT. The main function of the NHT was hydrodeoxygenation (HDO). The oxygenates and water were removed from the naphtha before entering the catalytic naphtha reformer because water and water-forming substances caused catalyst deactivation as well as corrosion problems that were associated with the use of a chlorided reformer catalyst. The reactor configuration of the NHT (Figure 9.6) was similar to that used for the oligomer hydrotreaters shown in Figure 9.5. The same catalyst (sulfided Co/Mo–Al2 O3 ) was initially used, but the operating conditions were more severe. Typical operating conditions were 320–420 ◦ C and 6.0–6.5 MPa. Hydrotreating occurred in the vapor phase. Heat management was very important, and the heat of reaction in the first and second reactors was used to preheat and vaporize the feed to the second and third reactors. The adiabatic temperature rise in the first reactor was managed by recycling some of the hydrogenated product. The hydrogenated product was fractionated to remove the light naphtha and gas that were produced as a result of distillation broadening. The production of ‘‘heavy naphtha’’ was a consequence of poor separation in the atmospheric distillation unit (SLO splitter). The need for additional product fractionation after the NHT illustrates the futility of atmospheric distillation before deoxygenation of Fischer–Tropsch syncrude. The catalytic naphtha reformer was built using UOP Platforming technology. The unit is of the continuous catalyst regeneration (CCR) type. These types of units are well described in the literature [23–25]. The Sasol 2 and 3 units were designed for operation at 540 ◦ C and 1 MPa with a chlorinated bimetallic Pt/Re–Al2 O3 catalyst. Although a catalytic naphtha reformer is generally considered a source of high octane motor-gasoline, the low aromatic and cycloalkane (naphthenic)
9.4 Fischer–Tropsch Refining
Fresh H2 Naphtha H2
H2 recycle
H2
H2
Purge
Naphtha hydrotreaters
Hydrogenated product recycle
Water
Fresh feed
Product fractionator Hydrogenated product
Gas Light naphtha Hydrogenated naphtha (to catalytic reformer)
Heavy naphtha Figure 9.6
Naphtha hydrotreater showing the heat management.
content of the Fischer–Tropsch feed resulted in a much lower conversion to aromatics than normally found. The N + 2A value of the HTFT feed is <30, making it a very lean naphtha. As a result, the liquid yield decreased significantly with an increase in operating severity. For example, at the operating severity required to produce research octane number (RON) 87, only 84% liquid yield was obtained [19]. In practice, the reformer was run at the lowest severity necessary to meet fuel specifications. The poor performance of standard catalytic naphtha reforming catalysts with Fischer–Tropsch feed is discussed later on (Chapter 22). There was a definite benefit in using a heavier Fischer–Tropsch naphtha cut as feed for the catalytic naphtha reformer. For the same RON of the product, the difference in the C5 and heavier liquid yield between a 180 and a 205 ◦ C end point naphtha feed was 15 vol% [19]. However, irrespective of the feed fraction employed, a significant yield loss was incurred to produce the reformate. The light products co-produced with the reformate were separated from the reformate in a stabilizer column that formed part of the unit. The DHT processed a mixture consisting of light vacuum gas oil (LVGO) and heavy vacuum gas oil (HVGO) from vacuum distillation, a side cut from atmospheric distillation and the heaviest product from the oligomerization reactors. It is not clear why feed fractionation was followed by feed recombination before the DHT. The DHT reactor contained four catalyst beds with interstage quenches and was operated in the range 290–380 ◦ C and 5.5–6.3 MPa. This unit was operated with a standard sulfided hydrotreating catalyst. The product from the DHT was fractionated. The naphtha fraction was routed to the NHT to ultimately be used as feed for the catalytic naphtha reformer. The distillate fraction became diesel fuel (Table 9.2) and the bottom fraction served as feed to the DSC. Again,
193
194
9 Sasol 2 and 3 Facilities
the need for product fractionation after the DHT unit highlighted the futility of fractionation before deoxygenation. The DSC was much smaller than the DHT, since HTFT technology produced little heavy material. Surprisingly, the original Sasol 2 and 3 designs overestimated the heavy-end fraction from HTFT synthesis. Although two DSC reactors were built, one at each facility, only one was really needed. The DSC used Mobil technology, and the reactor contained a single catalyst bed operated at 300–415 ◦ C and 4.0–3.6 MPa. The catalyst was a zeolite-based selective hydrodewaxing catalyst [3], likely a Ni/H-ZSM-5 catalyst. The product was fractionated to produce a poor quality motor-gasoline and good diesel fuel (Table 9.2), as well as small fuel oil fraction. Again a product fractionation column was needed after the conversion unit to overcome boiling range broadening. The viscosity of the product that had to be distilled was quite high, which complicated the operation of this column. The design intent of the DSC unit was to recycle the residue product to extinction and by doing so, to convert the entire residue fraction into mainly distillate. This unit failed to substantially convert the HTFT residue on account of its high aromatic content. Although the residue fraction is referred to as a waxy oil, it bore little resemblance to the n-alkane-rich product obtained during low-temperature Fischer–Tropsch (LTFT) synthesis. Neither the operating conditions nor the catalyst selection was appropriate for the feed material. The NHT, DHT, and DSC reactors are all fixed bed hydrotreaters. This caused some operating problems with the syncrude, which invariably contained some metals, notably Fe and Na [19]. It was found that, during normal operation, these units were pressure drop constrained rather than activity constrained [26]. 9.4.3 Aqueous Product Refining
The distribution of the water-soluble oxygenates in the aqueous product is given in Table 9.4 [27, 28]. In many respects, the composition of the Synthol HTFT aqueous product was similar to that of the Kellogg HTFT aqueous product (Tables 8.4 and 8.5). This was to be expected, since the designs were very similar. The sequence of primary separation at the Sasol 2 and 3 chemical workup sections (Figure 9.7) was the same as that of Sasol 1 (Figure 8.9). The oxygenates in the reaction water were concentrated by the primary distillation column in the overhead fraction. The bottom fraction consisted of a dilute acidic solution and contained predominantly light carboxylic acids (1–2%). Most of the bottom fraction was sent for biotreatment, and some was employed in the tail gas wash step after the Synthol reactors. The carboxylic acids were not recovered, but were biologically destroyed in activated sludge units [28]. The overhead fraction from the primary distillation column contained about 20% water and was further separated in the carbonyl stripper column to produce an overhead product that was rich in carbonyls (ketones and aldehydes) and a bottom product that was rich in alcohols. The overhead product from the carbonyl stripper was sent to the carbonyl recovery section. The separation and processing sequence was similar to that of the Sasol 1 chemical workup (Section 8.4.3). The ethanal (acetaldehyde) was recovered in the aldehyde column and hydrogenated to produce more ethanol. In the ketone column, the ketones were recovered overhead, while
9.4 Fischer–Tropsch Refining Composition of oxygenates in the aqueous product of Synthol HTFT synthesis.
Table 9.4
Compound
Alcohols Methanol Ethanol 1-Propanol 2-Propanol (isopropanol) 1-Butanol 2-Butanol 2-Methyl-1-propanol (isobutanol) 1-Pentanol 2-Pentanol Other alcohols Aldehydes Ethanal (acetaldehyde) Propanal (propionaldehyde) Other aldehydes Ketones Acetone 2-Butanone (MEK) Pentanones Other ketones Carboxylic acids Ethanoic acid (acetic) Propanoic acid (propionic) Butanoic acid (butyric) Other carboxylic acids a
Composition (mass%) [25]
[26]
1.2 46.4 10.7 2.5 3.5 0.7 3.5 1.0 0.1 0.5
0.9 49.7 14.5a – 6.3a – – 1.8a – –
2.5 0.8 0.5
4.5b – –
8.9 2.5 0.7 0.2
9.0 2.7 0.9 –
9.7 2.2 1.2 0.7
6.7 1.5 0.9 0.5
All alcohol isomers of the same carbon number. aldehydes.
b All
the bottom product, which contained mainly water, methanol, and some ethanol, was further separated in subsequent columns. Unlike in the Sasol 1 design, the ketones were further purified to produce acetone and butanone (methylethyl ketone, MEK) as final products. The bottom product of the carbonyl stripper was sent to the alcohol recovery section, which was similar to that used at Sasol 1. The water was removed by azeotropic distillation with benzene. This was followed by selective hydrogenation of the ketones in the alcohol mixture to produce alcohols. In the Sasol 2 chemical workup section, the alcohol mixture was then separated to produce ethanol, mixed propanols, mixed butanols, and a heavier alcohol mixture. Not all of these steps were included in the Sasol 3 alcohol recovery section. The acid water and other wastewater streams were biologically treated and kept in storage dams for reuse on the facility. The water system at Sasol 2 and 3 was designed to be a zero-effluent system. It was found that the water quality requirements for reuse on the plant was less stringent
195
196
9 Sasol 2 and 3 Facilities
Carbonyl recovery Acetone
Aldehyde H2 hydrogenation
Butanone Ethanol
Primary separation Carbonyls + CH3OH
Aqueous product Acid water
Aldehyde tower
Alcohols + water
Carbonyl stripper Primary separation column
Methanol
Water Ethanol
Alcohol recovery Ethanol (95%) Water removal (Benzene azeotropic distillation)
Water
Figure 9.7
C5+ ketones
Ketone tower
C3 alcohols
H2
C4 alcohols Ketone hydrogenation
C5+ alcohols
Aqueous product refining in the original chemical workup section of Sasol 2.
than the wastewater treatment required in order to return the water to the Vaal River system [2]. Most of the water was disposed of through the cooling towers. 9.4.4 Coal Pyrolysis Product Refining
Both Sasol 2 and Sasol 3 included ammonia synthesis (Figure 9.3) as part of their original design [4]. This made sense, since a large quantity of pure nitrogen was available from the air separation units as a by-product and at essentially no additional cost. Phenolic material and ammonia were recovered in a Phenosolvan process. The Phenosolvan process is similar to that described for the Sasol 1 plant (Section 4.4.1). The manufacture and recovery of ammonia was extended to the production of nitrogen-based chemicals. The nitrogen-based chemicals can be grouped into two categories fertilizers for the agricultural industry and explosives for the mining industry. The production of nitrogen-based chemicals took place in downstream facilities and did not form part of the refinery design. The designs of the Sasol 2 and Sasol 3 tar workup sections (Figure 9.8) were similar, except for the addition of a high-pressure creosote hydrogenation section to the tar refinery of Sasol 3 [1]. This unit processed the heavy tar from both Sasol 2 and 3 and was one of the few points of integration between the original Sasol 2 and 3 facilities. The feed to the tar distillation unit consisted of the phenolic pitch, after phenol extraction in the Phenosolvan unit, and crude tar from the coal tar filtration unit (Figure 9.1). These feeds were combined with a heavy recycle stream from the coal tar naphtha hydrogenation unit before being
9.4 Fischer–Tropsch Refining
Coal tar naphtha hydrotreater
H2
Rectisol naphtha H2
H2
Tar distillation Light naphtha
Creosote hydrotreater H2
Crude tar Phenolic pitch
Heavy naphtha
Water 180 °C stripper Tar distillation
Naphtha
Residue oil
Water
Medium creosote
H2
H2
H2 H2
H2
H2
H2
H2
Naphtha
Heavy creosote
Residue oil
Distillate Pitch
Pitch
Creosote (from Sasol 2)
Figure 9.8 Coal pyrolysis product refining in the original tar workup section at Sasol 3. The tar workup section of Sasol 2 did not include a creosote hydrotreater.
distilled. A typical product distribution from tar distillation consisted of six fractions, namely 10% light naphtha, 8% heavy naphtha, 25% medium creosote, 13% heavy creosote, 5% residue oil, and 39% pitch. The light naphtha, heavy naphtha, and Rectisol naphtha from the Rectisol unit were combined to serve as feed for the coal tar naphtha hydrogenation unit. The feed was preheated to around 210 ◦ C and flashed to separate the naphtha range material from the heavier boiling material. Separation of the heavier material took place in the first reactor, which was operated in up-flow mode. The heavier material, called residue oil, was recycled to the tar distillation unit, while the naphtha vapor was hydrogenated. The first reactor was loaded with a mild hydrotreating catalyst and was mainly used for diene saturation. Apart from saturation of the most reactive species, little conversion took place in the first catalyst bed. Most of the hydrotreatment was performed in the second reactor, which had two catalyst beds with an interstage hydrogen quench. Typical operating conditions for the second reactor were a reactor inlet at around 230 ◦ C and 5 MPa, and a total adiabatic temperature rise over the catalyst beds in excess of 100 ◦ C. Originally, a sulfided Co/Mo–Al2 O3 hydrotreating catalyst was employed in the main catalyst beds of this unit. The product from the coal tar NHT is a hydrogenated naphtha consisting of mainly alkanes and aromatics. Despite the high aromatic content of this naphtha fraction, it had a low octane number (Table 9.5). Most of the benzene in the final motor-gasoline also came from this product. The creosote hydrotreater at Sasol 3 was designed for a production capacity of 360 000 ton per year, which was sufficient to process creosote from both facilities [1]. The objective of this unit was to convert the medium creosote, heavy creosote, and residue oil from the tar distillation units
197
198
9 Sasol 2 and 3 Facilities Table 9.5 Density, research octane number (RON), motor octane number (MON), and cetane number of products in the tar workup section.
Refinery product
Motor-gasoline Hydrogenated coal tar naphtha Hydrogenated creosote naphtha Diesel fuel Hydrogenated creosote distillate
Selected fuel properties Density (kg·m−3 )
RON
MON
Cetane number
795 770–775
80–83 71–72
73–76 70–72
– –
875
–
–
35
into blending components for motor-gasoline and diesel fuel. This required deep hydrotreating, and the creosote hydrotreater was operated at 280–380 ◦ C and 18.5 MPa hydrogen pressure [29]. Although this was sufficient to remove most heteroatoms, it was accompanied by considerable hydrodearomatization (HDA). The fuel properties from this unit was rather poor (Table 9.5), but the high density of the distillate made it a welcome diesel blending component. The creosote diesel was the main source of diesel density in the refinery. 9.4.5 Synthetic Fuel Properties
The refining and blending philosophy that governed the design of the Sasol 2 and 3 facilities is best summarized by quoting Hoogendoorn, a previous General Manager of Sasol, directly [30]: ‘‘When one produces synthetic motor fuels on a large scale, then it is essential that the refined products can be marketed as specification products without the necessity of blending with products from crude-oil refineries.’’ The fuel quality that could be achieved by refining HTFT syncrude with a crude oil refining approach was adequate for the time (Table 9.6) [12, 30]. The South African fuel specifications were not very stringent. The addition of tetraethyl lead made it comparatively easy to meet motor-gasoline octane specifications. The tetraethyl lead requirements were reduced by adding the alcohols recovered from the aqueous product to the motor-gasoline [31]. The cetane number requirement for diesel fuel was also quite low and easily achieved by the refinery. Many of the other specifications (not listed in Table 9.6) were mainly concerned with fuel stability, distillation profile, and corrosivity. A salient point worth mentioning is that there was (and still is) a large market for fuels close to the Sasol 2 and 3 facilities in the Pretoria–Johannesburg region. In this market, Sasol 2 and 3 had a logistic advantage over the coastal crude oil refineries. Furthermore, the cities of Pretoria and Johannesburg are at an altitude of 1500–1600 m above sea level. Because of the high altitude of this region, the octane number requirement for motor-gasoline is lower than that required at sea level. In practice it meant that Sasol 2 and 3 had a captive market for mainly RON 87 and 93 grades of motor-gasoline. There was no need to produce RON 98 motor-gasoline, which was a coastal grade.
9.5 Evolution of Sasol Synfuels Selected properties of the motor-gasoline and diesel fuel refined from Fe-HTFT syncrude and coal liquids at Sasol 2 that was marketed as final products in South Africa, as well as the South African fuel specifications at that time.
Table 9.6
Fuel property
Liquid product yield (%) Density (kg·m−3 ) RON (+0.3 g Pb/l)a MON (+0.3 g Pb/l)a Cetane number Viscosity (cSt) Cold filter plugging point (◦ C) Flash point (◦ C) Reid vapor pressure (kPa) Distillation (◦ C) T50 Final boiling point (FBP) Composition (%) Alkanes Alkenes Aromatics Sulfur
Motor-gasoline
Diesel fuel
Sasol 2
Specification
Sasol 2
Specification
55–65 720 93 85 – – – – –
– – 93b – – – – – <75
35–45 800 – – 47 1.85 –12 59 –
– – – – 35 1.8–5.8 <–7c >65 –
90 200
77–115 <215
219 378
– 385
49 24 27 0.0001
– – – <0.15
96–98 0 2–4 0.0002
– – – 0.5
a Tetraethyl
lead addition, maximum allowed by specification is 0.836 g Pb/l. different grades were marketed: RON 87, 93, and 98 (coastal areas only). c Specification is for the pour point. b Three
9.5 Evolution of Sasol Synfuels
Sasol 2 and 3 are no longer known by these names. The two refineries became so integrated that they are now called Sasol West and Sasol East and the refinery complex is collectively known as Sasol Synfuels. Unlike Sasol 1, it remained a coal-to-liquids facility and it also remained an HTFT facility. Fuel production at Sasol Synfuels year on year contributed more than half of the overall profit posted by the Sasol group of companies, making synthetic fuel production the single most important asset and source of cash for growth for Sasol. The significant changes in fuel specifications that took place in the rest of the world during the period 1980–2000 drove transition to fourth-generation crude oil refineries (Section 2.4.4), but initially had little impact on the synthetic fuel refining business. This situation is changing though and its full effect is still to be felt as the South African fuel specifications move closer to European standards that have been used as a benchmark. In turn, this will inevitably lead to future changes in the refinery. Thus far, most of the changes at Sasol Synfuels were brought about by chemical extraction from the Fischer–Tropsch syncrude. Although the facility has a logistic disadvantage for chemical
199
200
9 Sasol 2 and 3 Facilities
exports due to its inland location, transportation cost is more than offset by the cheap raw material cost. The syncrude that serves as raw material for chemicals is the same raw material as that for fuels, and generally the cost of extraction is less than the cost of on-purpose synthesis. Inevitably, the increasing level of chemical extraction also affected fuel refining at Sasol Synfuels. Many of the opportunities for chemical extraction that were exploited, improved the quality of the fuels being produced by removing molecules from the syncrude that had poor fuel properties. Chemical production from syncrude seemed like a clear refining benefit for fuels production, but there were unintended consequences. Some general impacts on refining, which are not specific to the transformation of Sasol 2 and 3, and that have to be considered, are the following: 1) Increasing chemical extraction comes at the cost of decreased refining flexibility, decreased operability and increased refinery complexity. If the chemical extraction is tightly integrated in the fuels refinery, any upsets in chemical production will also affect fuels production. 2) The benefit of increased fuel quality by chemical extraction can be realized only if the chemical production facility becomes an independent consumer. The chemical production facility must guarantee the extraction of that chemical. If this is not the case, the fuel refinery derives no benefit. In fact, it comes at an additional cost to the fuel refinery, because it should have operating modes to meet fuel specifications with and without chemical extraction and it should have strategies to make the transition between the modes of operation. The engineering of reincorporation pathways then becomes very important [32]. As the number of chemical extraction facilities increases, the combinations and permutations that have to be allowed for also increases. 3) Tightly integrating chemical extraction and fuel refining reduces the overall refinery stability. The probability of smooth operation is the product of the individual probabilities of smooth operation. One way to mitigate the impact of instability caused by chemical extraction units is to have sufficient storage tank capacity available to store feed destined for chemical extraction and to have sufficient downstream capacity in the chemical extraction unit itself. This mitigation strategy becomes less successful if the integration is so tight that fuel refinery operation depends on return streams from chemical extraction. 4) Fuel blending sensitivity to meet fuel specifications increases as chemical extraction increases. This happens in two ways. Firstly, the volume of the fuel that is available for blending decreases. This implies that any fixed volume being added or removed to the fuel increases in percentage relative to the total volume of fuel, which has decreased due to chemical extraction. Normal refinery upsets therefore have a bigger impact on fuel blending, because the total volume of fuel that is being produced has become smaller. Another effect of shrinking the fuel volume is that the concentration of regulated compounds, such as benzene, increases. Secondly, chemical extraction usually targets a single compound of a single class. This not only affects the distillation profile of the fuels, but also the relative volume of fuel types. In the case of motor-gasoline specifically, it directly affects the ability of the refinery to meet fuel specifications, because there are limits on the maximum alkene, aromatic, and oxygenate contents in motor-gasoline. By removing a chemical from one compound class, it automatically increases the relative concentration of all the other compound classes. 5) Some chemicals have poor fuel properties, but have efficient refining pathways to be upgraded to good quality fuels. By removing such chemicals from the fuel, these refining pathways may be closed, or may become unattractive because the amount of material that can benefit from such upgrading is insufficient to warrant such a unit.
9.5 Evolution of Sasol Synfuels
6) When a large volume of material is extracted as chemicals, it may alter the composition of the base fuel to such an extent that it invalidates the refinery fuel blending model. If this is not rectified, it may result in suboptimal refining and blending decisions. 9.5.1 Changes in Synthesis Gas Production
Synthesis gas is still produced by coal gasification with Lurgi dry ash gasifiers. However, with the construction of a natural gas pipeline to supply Sasol 1 with natural gas (Section 8.5.1), natural gas also became available to Sasol Synfuels. Future capacity expansion at Sasol Synfuels may well be with natural gas rather than with coal. As in the case of Sasol 1, by doing so it will undermine the original intent of providing energy security. 9.5.2 Changes in Fischer–Tropsch Synthesis
In June 1995, an 8 m diameter Sasol Advanced Synthol (SAS) fixed fluidized bed reactor (Figure 9.9) was commissioned at Secunda. The synthesis gas is introduced at the bottom of the reactor. The gas flow rate is controlled to create a fluidized bed, the lower limit being set by fluidization and the upper limit set by entrainment. Catalyst particles that are entrained by the product gas is separated and returned to the fluidized bed by cyclones in the reactor. On-line catalyst addition and removal is possible and it is used to maintain reactor operation at a constant average catalyst age. The principle of fixed fluidized bed operation is straightforward, but it placed more demands on the control of the catalyst particle size distribution and gas flow rate than the Synthol CFB reactors. The SAS fixed fluidized bed technology proved to be more efficient than the Synthol CFB technology, although both use the same fused-iron catalyst and produce a similar product slate [8]. Product gas Cyclone
Water
Steam
Fixed fluidized bed
Gas distributor Syngas
Figure 9.9 reactor.
Sasol advanced synthol (SAS) fixed fluidized bed
201
202
9 Sasol 2 and 3 Facilities
From a historical point of view, commercial HTFT synthesis came full circle, since the American Hydrocol facility also employed fixed fluidized bed reactor technology (Section 7.3). The success of the first SAS reactor led to the decision in March 1996 to change the Fischer–Tropsch synthesis section and replace all the old gooseneck Synthol CFB reactors with SAS reactors. The synthesis sections at Sasol West and East each had two trains of four Synthol CFB reactors, and it was decided to install one 8 m diameter and one 10.7 m diameter SAS reactor per train. The first new SAS reactor came on stream in September 1998 and the last one in February 1999. A ninth 8 m diameter SAS reactor was added in 2001 to serve as backup [4]. Another modification that was later made was to place some SAS reactors in series. This increased the overall syngas conversion, as one would expect from reaction engineering fundamentals. With the Sasol Synfuels synthesis gas composition and operating conditions, an 8 m diameter SAS has a capacity of around 11 000 bbl/day and a 10.7 m one around 20 000 bbl/day. The name plate capacity of Sasol Synfuels therefore remained close to 120 000 bbl/day crude oil equivalent. 9.5.3 Changes in Fischer–Tropsch Condensate Refining
The commissioning of a 120 000 tons per year (tpa) polypropylene plant at Secunda during late 1989–early 1990 heralded the start of a new chemicals production era. The polypropylene plant was built using BASF technology. The capacity of this plant was later increased to 220 000 tpa [4]. The Synfuels refineries were affected by the addition of a propene–propane splitter and extraction of some propene from the HTFT condensate streams. Further extraction of propene would follow in later years to supply feed to chemical projects like the production of acrylic acid and 1-butanol. Although the acrylic acid and 1-butanol plants were constructed at the Midlands site in Sasolburg (across the road from Sasol 1), they use propene supplied from the Synfuels facility. The second major expansion in the polymer field came in 2006 with the addition of new polyethylene and polypropylene facilities as part of project Turbo [33]. Most of the additional ethene and propene were to be supplied from a new catalytic cracker unit (Section 9.5.5), but it also resulted in increase propene extraction from the HTFT condensate. A header system was constructed in the condensate workup section, which in principle made separate C3 and C4 refining possible. The increased extraction of propene allowed more butane-rich operation of the SPA-catalyzed oligomerization units. Separate C3 and C4 oligomerization does not affect the quality of olefinic motor-gasoline, but C3 -only oligomerization improves distillate yield [34], whereas C4 -only oligomerization improves the octane number of hydrogenated motor-gasoline [22]. The alkene oligomerization units in the original Sasol 2 and 3 design operated with a mixed C3 –C4 feed, which often included some C5 on account of poor separation. The advantage offered by C4 -only oligomerization for hydrogenated motor-gasoline production was not exploited, and the quality of the hydrogenated motor-gasoline was correspondingly poor (Table 9.3). As more and more propene was extracted for chemical use, the butene content of the C3 –C4 mixture increased and as a consequence the 2-methylpropene (isobutene) content was increased as well. This caused a gradual improvement in the quality of the hydrogenated motor-gasoline from SPA-catalyzed oligomerization over time (Figure 9.10) [35].
9.5 Evolution of Sasol Synfuels
90 87.8 86.1
Octane number
85
C4-only operation
80 Sasol Synfuels in 2008 75
79.6 77.8
77.0 74.4
Sasol Synfuels in 2001 70.6
70 Original Sasol 2 and 3 65
RON MON
63.7
60 0
10 20 30 40 50 Propene content of the C3 –C4 alkenes in the feed (mass%)
60
Figure 9.10 Improvement in the quality of hydrogenated motor-gasoline obtained by alkene oligomerization over solid phosphoric acid due to the increased extraction of propene from the mixed C3 –C4 Fischer–Tropsch feed material.
Although ethene is not normally processed in the condensate workup section, a Dimersol E unit was constructed to convert excess ethene into C4 and C6 alkenes. The 1-butene was oligomerized with the C3 –C4 condensate fraction, and the 1-hexene was processed with the C5 –C6 condensate material. The technology was licensed from the Institut Franc¸ais du P´etrole (IFP). It is a homogeneous oligomerization process that is catalyzed by a Ni-based Ziegler-type complex [36]. The Dimersol E unit was a risk-mitigation project associated with downstream polymer production. The on-stream availability of the polyethylene facilities was a concern, and the Dimersol E unit was used to convert the ethene during upset conditions. However, it was expensive to keep the plant idling, and alternative uses for the plant had been actively sought in the late 1990s. The unit was eventually permanently decommissioned. The C5 –C6 isomerization unit did not perform well for the reasons outlined earlier (Section 9.4.1). It was shut down and reused as an alcohol dehydration unit. Mixed alcohols from the aqueous product refinery were dehydrated over an η-alumina catalyst at around 400 ◦ C and near-atmospheric pressure [37]. The alkenes that were produced by alcohol dehydration were processed with the Fischer–Tropsch condensate streams. This was only a temporary arrangement until the alcohols could be sold in the solvent market. Applying an η-alumina catalyst for the conversion of the C5 –C7 SLO naphtha had been considered [38] but it was never implemented. One of the concerns was carboxylic acid formation. In addition to the changes in refinery configuration, the units saw changes in catalyst selection. Some changes were made because of performance: for example, the C84/3 SPA catalyst from S¨ud-Chemie for the alkene oligomerization units [39]. Changes were also made in the hydrotreating catalyst for alkene hydrogenation in the Polyhydrotreater.
203
204
9 Sasol 2 and 3 Facilities
9.5.4 Extraction of Linear 1-Alkenes
The extraction of n-1-alkenes (linear α-olefins) from HTFT syncrude is probably the best known expansion drive in chemical production at the Sasol Synfuels facility. The condensate and oil streams of Sasol West and East served as feed material for extraction. The short-chain n-1-alkenes, namely, 1-butene, 1-hexene, and 1-octene, are used as comonomers in the production of polyethylene. These molecules are usually produced by ethene oligomerization and there is a significant production cost advantage to extract the n-1-alkenes directly from HTFT syncrude compared to production from ethene [40]. There was consequently a clear incentive to develop extraction technology. Over the years, Sasol developed much expertise in the extraction of n-1-alkenes from Fischer–Tropsch syncrude, and it is one of the few fields where Sasol did not rely on licensed technology from others. Processes were developed for the extraction of 1-pentene, 1-hexene, 1-heptene, and 1-octene (Chapter 23). In 1992, a pilot plant was constructed for the purification of 1-pentene and 1-hexene from HTFT condensate. After demonstrating the process on pilot scale it was commercialized. The commercial extraction process has been described in the literature [40]. The extraction units are not all the same. Extraction of 1-hexene from the light oil is more involved than from HTFT condensate, because it contains percentage level oxygenates that must be removed by N-methyl pyrrolidone (NMP) extraction before the etherification step in the process. The first commercial implementation of 1-hexene extraction used HTFT condensate as feed, which was simpler to purify. A twin-train plant was constructed at Sasol Synfuels to produce 100 000 tpa 1-pentene and/or 1-hexene. After debottlenecking, the capacity was increased to 140 000 tpa. In 2000, a third 1-hexene train was added with a capacity of 80 000 tpa. This gave Sasol about 25% of the global market share for 1-hexene [4]. The marketing of 1-pentene was not successful despite the benefits of using it as a comonomer [41]. The risk of having a single supplier of a particular chemical apparently outweighed the benefits of using it. This highlighted two important problems associated with n-1-alkene extraction from Fischer–Tropsch syncrude. Firstly, the global market for these molecules is limited, and there is limited scope for future HTFT facilities to capitalize on the same technology. This is a problem that has been pointed out before, but in conjunction with HTFT oxygenates (Section 7.5) [42]. Secondly, it may be possible to extract useful chemicals that enable new applications (i.e., odd-numbered n-1-alkenes), but nobody likes to be dependent on a single global supplier. The impact of 1-hexene (RON = 76.5, motor octane number, MON = 63.4 [43]) extraction on the motor-gasoline quality was viewed as positive. However, the octane number of 1-hexene can easily be increased by isomerization, and 1-hexene is an example of a chemical that has poor fuel properties but has efficient refining pathways to be upgraded to good quality fuels. The most abundant of the ethers coproduced during the 1-hexene extraction, namely 2-methoxy-2-methylpentane, has a blending RON and MON of 85 and 86, respectively [44]. Etherification of the C6 material did not markedly improve the quality of the motor-gasoline. The next n-1-alkene that was targeted for extraction was 1-octene. Unlike 1-hexene, all the 1-octene is found in the light oil. The purification of 1-octene required a different process configuration and it is quite different to that employed for 1-hexene extraction. The first 1-octene extraction train was commissioned in 1999 and had a capacity of 50 000 tpa. The process has been described [45], and from a refining perspective the first step is of
9.5 Evolution of Sasol Synfuels
importance. The oxygenates that co-boil with 1-octene contain carboxylic acids. The oxygenates could therefore not be extracted with NMP, which is a base, before the carboxylic acids were removed. The carboxylic acids were neutralized with potassium carbonate (K2 CO3 ) and removed by a water-wash step. However, the amphophilic nature of the potassium carboxylates was underestimated. Parts per million levels of the potassium propanoate and potassium butanoate ended up in the organic phase and this caused severe operational problems in the NHT downstream from the 1-octene extraction unit [46, 47]. The NHT developed serious pressure drop problems due to the potassium carboxylates. The problem could eventually be overcome. It also prompted the development of a new approach for the second 1-octene extraction train. In the second 1-octene train, the K2 CO3 neutralization step could be eliminated by employing azeotropic distillation with an ethanol–water mixture as entraining solvent [45, 48]. The oxygenates are not entrained with this mixture, with the result that the carboxylic acids and other oxygenates can be removed in a single step. All the off-cuts produced during the extraction processes were returned to the oil refinery. Surprisingly, the oxygenate-rich material was not processed separately but recombined with the hydrocarbons, nullifying the potential advantage of separate processing. A third 1-octene train was commissioned in 2008. This unit differed considerably from the other two 1-octene extraction trains. It did not extract 1-octene but extracted 1-heptene. The 1-heptene that is extracted from the light oil is then hydroformylated to produce octanal, which is hydrogenated to 1-octanol and then dehydrated over γ -alumina to produce the 1-octene [49, 50]. The extraction of 1-heptene (RON = 54.5, MON = 50.7 [43]) and 1-octene (RON = 28.7, MON = 34.7 [43]) improved motor-gasoline quality. Both molecules had an existing refining pathway for the production of good quality motor-gasoline by catalytic naphtha reforming. Contrary to their straight-run octane numbers, the extraction of 1-heptene was more beneficial to fuel refining than the extraction of 1-octene. The C7 hydrocarbons are one of the most troublesome Fischer–Tropsch carbon numbers to efficiently refine to motor-gasoline [32]. The know-how to purify linear 1-alkenes for the comonomer market was also applied to the C12 –C13 range of materials. These detergent range alkenes were purified and then hydroformylated with CO and H2 over a classic Rh-based catalyst. The aldehydes thus produced were hydrogenated to alcohols for the detergent alcohol market [51]. The Safol detergent alcohol plant was commissioned in 2002. 9.5.5 Changes in Fischer–Tropsch Oil Refining
Sasol Synfuels benefited from the lenient South African fuel specifications. The refinery was not designed to cope with the changes in international fuel specifications that were being introduced elsewhere in the world. It still had to make the transition from a third-generation to a fourth-generation refinery (Section 2.4.4). Most of the early changes in the Fischer–Tropsch refinery were related to catalyst replacements in the existing units. The Fischer–Tropsch syncrude favored catalysts with mild activity because of the reactive nature of the alkene- and oxygenate-rich syncrude. The trend in crude oil refining was to continuously increase the activity of catalysts and, particularly in the field of hydrotreating, many advances were made. Throughout its history, the naphtha and distillate hydrotreaters employed typical sulfided crude oil hydrotreating catalysts. This necessitated not
205
206
9 Sasol 2 and 3 Facilities
C2-rich
Ethene splitter
Propene splitters
Condensate 2 + 3 Condensate 1
C3 /C4 alkene oligomerization
Alkene hydrotreater
C5 skeletal isomerization
TAME - C5 etherification
C5 –C6
C5 –C6
1-Hexene extraction C5/C6 CD hydrotreater
C5-raffinate C6-offcuts
C11 –C14 >C15
Decanted oil
Hydroformylation
LVGO
Motor-gasoline Motor-gasoline 1-Octene
Naphtha hydrotreater
C7 –C10
TAME 1-Hexene
1-Octene extraction Light oil (SLO)
Ethene Propene LPG/butane Motor-gasoline Distillate* Motor-gasoline
>C10
Heavy naphtha
Distillate hydrotreater
HVGO
Catalytic reformer
Distillate hydrocracker
Motor-gasoline Diesel fuel Detergent alcohols Diesel fuel Motor-gasoline Diesel fuel Fuel oil
Figure 9.11 Condensate and oil workup sections of the Sasol Synfuels refinery in 2004. The stream indicated with an asterisk (*) was also employed for the production of semisynthetic jet fuel.
only sulfur addition to the otherwise sulfur-free syncrude, but also necessitated careful evaluation of catalysts as newer and more active catalysts replaced the older, less active catalysts. Apart from the chemical extraction units, very few refinery units were added to improve fuel quality before 2004 (Figure 9.11). Compared to the original design of the condensate and oil workup sections (Figure 9.4), only the C5 –C6 refining pathway was changed. The C5 –C6 isomerization unit was decommissioned and new units for the refining of the C5 material were installed. A catalytic distillation column was used to saturate dienes in the feed before it was etherified in a subsequent catalytic distillation column. In this ‘‘CD Hydro’’ unit, the hydrogenation was conducted over a Pd-based catalyst which is active for double bond isomerization (Section 16.2.1). This resulted in the double bond isomerization of some 1-pentene to 2-pentenes with higher octane numbers. The etherification process employed methanol and produced 2-methoxy-2-methylbutane (tertiary amyl methyl ether, TAME). These units were commissioned before the year 2000. Subsequently, a pentene skeletal isomerization unit was added to increase the TAME production. This unit employed an alumina-catalyzed skeletal isomerization process that was licensed from IFP [52]. The process was designed to operate at 400–500 ◦ C, 0.2–0.4 MPa (absolute), and liquid hourly space velocity (LHSV) of 1–2 h−1 and it employed CCR to maintain catalyst activity. It performed well with the Fischer–Tropsch C5 feed, and early operational issues related to incomplete catalyst regeneration [53] were quickly resolved.
9.5 Evolution of Sasol Synfuels
The first major change in South African fuel specifications came with the phasing out of leaded motor-gasoline. Initially, the impact of lead phaseout was limited. It coincided with the introduction of ‘‘Dual fuel’’ by Sasol. The Dual fuel was a lead replacement gasoline (Table 9.7) [54] which substituted tetraethyl lead with methylcyclopentadienyl manganese tricarbonyl (MMT). The octane number of the motor-gasoline could therefore still be improved by metal-based additives and without refinery changes. This was only an interim measure, since the South African Government announced its intentions to remove lead from motor-gasoline and align South African fuel quality requirements with international fuel specifications and emission standards. The fuel specification changes that were legislated in June 2006 [55] required no intentional metal addition to motor-gasoline, although provision was made for lead replacement fuel. In other respects, the only other major change was a reduction in the maximum sulfur content to 0.05%, which did not affect the Fischer–Tropsch refinery. The fuel quality approached European 1993 Euro-2 specifications (Table 9.7). Presumably, it is only a matter of time before the South African fuel specifications move closer to 1999 Euro-3 and the 2004 Euro-4 motor-gasoline (EN228) and diesel fuel (EN590) specifications [54]. In 2003, Sasol announced Project Turbo [56]: ‘‘It is anticipated that the South African Government, in line with new international environmental trends, will introduce in January 2006 stricter new legislation in order to terminate the production and sale of both leaded petrol and higher-sulfur diesel. The Government is also expected to promulgate further legislation aimed specifically at reducing the levels of benzene and aromatics in petrol after 2006. Sasol will have to invest an estimated R7 billion to modify Sasol Synfuels’ liquid fuel refining and blending operations and to establish additional new plant[s] aimed at increasing the octane rating of our synthetic petrol. The major Project Turbo expenditure, revolving primarily around an estimated investment of R5.5 billion to install a selective catalytic cracker . . . this one does not meet the required [financial] returns . . . however, the new selective catalytic cracker will generate additional large quantities of monomer feedstock (ethylene and propylene). This will enable lucrative expansion opportunities for the production and international marketing of polyethylene and polypropylene by Sasol Polymers. These envisaged monomer and polymer expansions, costing about R6 billion, will yield high returns after 2006 and will counter the impact of the investments undertaken for Sasol Synfuels.’’ The intended changes to the Synfuels refinery for Project Turbo (Figure 9.12) fell behind schedule and were not completed in time to meet the 2006 fuel specification changes. A number of low cost modifications were suggested to meet the specifications [57]: Fractionating the hydrogenated coal tar naphtha to obtain a high octane <130 ◦ C benzene and toluene-rich light naphtha cut. It would remove the heavier n-alkanes present in the hydrogenated coal tar naphtha from the motor-gasoline, thereby increasing the octane number. The 130–180 ◦ C cut was earmarked as feed for the catalytic naphtha reformer, whereas the >180 ◦ C material would end up in the diesel fuel. This would integrate the tar workup section and the Fischer–Tropsch refinery. Instead of installing a new column, the hydrogenated coal tar naphtha was routed to the product fractionation after the alkene hydrogenation section, that is, the Polyhydrotreater (Figure 9.5). The hydrogenated motor-gasoline from this unit improved by 9 octane number units. 2) Sending additional C6 alkenes to the catalytic distillation column used in the light oil feed preparation section before 1-hexene extraction (Section 9.5.4). This catalytic distillation unit
1)
207
208
9 Sasol 2 and 3 Facilities Table 9.7 Properties of the motor-gasoline and diesel fuel produced by Sasol Synfuels in relation to the South African motor-gasoline (SANS 1598) and diesel fuel (SANS 342) specifications in 2006.
Fuel property
Density at 20 ◦ C (kg·m−3 ) RON, minimum MON, minimum Cetane number, minimum Viscosity at 40 ◦ C (cSt) Cold filter plugging point (◦ C) Flash point (◦ C) Reid vapor pressure (kPa) Maximum distillation (◦ C) Initial boiling point (IBP) T10 T50 T90 FBP, final boiling point Content, maximum (%) Alkenes Aromatics, total Benzene Polynuclear aromatics Sulfur Oxygenates, as O
Motor-gasoline
Diesel fuel
Lead replacement
Unleaded
Specification
Diesel
Specification
723 93 83 – – – – 66
729 93 83 – – – – 67
710–785 91/93/95a 81/83/85a – – – – 45–75
829 – – 55 2.2 −6 77 –
800–860 – – 45 2.2–5.3 <−4 >62 –
35 51 89 155 201
35 53 96 155 200
Report <65 77–115 <185 <210
192 228 277 348 394
– – – 362 Report
30 25 –b – <0.001 0.05
30 29 –b – <0.001 0.14
30 50 5 – 0.05 3.7c
– 25 – <1 <0.0005 –
– – – – 0.05 –
a Three
grades of motor-gasoline. Lead replacement fuel available only in RON 93 and 95. reported, normally around 3–5%, to the higher end with SCC operation. c The maximum addition of ethanol is 8% and of ether is 20% (inland)/15% (coastal). b Not
had similar operation as the ‘‘CD Hydro’’ unit for the C5 material and it employed an isomerizing hydrogenation catalyst. By increasing the amount of C6 material going to this unit, the octane number could be improved. Specifically, the C6 cut rich in 3-methyl-1-pentene and 4-methyl-1-pentene was targeted. 3) Increasing the severity and feed rate to the catalytic naphtha reformers. As mentioned before, the 130–180 ◦ C fraction of the hydrogenated coal tar naphtha was to be routed to the reformer, thereby increasing the feed rate. This capacity increase was possible because the 1-octene extraction units (Section 9.5.4) removed material from the feed to the reformers. The proposed increase in severity was accompanied by an increase in octane number of the reformate to RON 95, albeit at a lower liquid yield. It required the catalytic naphtha reformers to operate close to their maximum design temperature of 543 ◦ C, but under the circumstances this was warranted. 4) Modifying the C5 refining pathway to increase the throughput. This allowed additional production of high octane C5 s.
9.5 Evolution of Sasol Synfuels
C2-rich
Ethene splitter
Propene splitters
Condensate 2 + 3 Condensate 1
C3 /C4 alkene oligomerization
Alkene hydrotreater
C5 skeletal isomerization
TAME - C5 etherification
C5 –C6
C5 CD hydrotreater C5
C5 –C6
1-Hexene extraction
C5 -raffinate C6
Catalytic cracker
C7
1-Octene extraction Light oil (SLO)
Naphtha hydrotreater
C7 –C10
HydroC11–C14 formylation >C15
Decanted oil
LVGO
Distillate hydrotreater
HVGO
Catalytic reformer
Heavy naphtha
Distillate hydrocracker
Ethene Propene LPG/butane Motor-gasoline Distillate* Motor-gasoline TAME Motor-gasoline 1-Hexene Ethene Propene Motor-gasoline 1-Octene Motor-gasoline Distillate* Detergent alcohols Distillate* Motor-gasoline Diesel fuel Fuel oil
Figure 9.12 Condensate and oil workup sections of the Sasol Synfuels refinery in 2006, after the construction of Project Turbo but before the extraction of 1-heptene. The streams indicated with an asterisk (*) are employed for diesel fuel and/or fully synthetic jet fuel.
5) Other backup contingency plans included flash point optimization of various refinery cuts and the addition of ethanol from the aqueous product refinery to the motor-gasoline. It was not necessary to implement all these modifications, and the refinery was able to meet the new fuel specifications in time and without the changes associated with Project Turbo (Table 9.7). This was a remarkable achievement and the team responsible for these changes received an award from the South African Institution of Chemical Engineers for this work. By looking at the molecules and refining Fischer–Tropsch syncrude as syncrude, not crude oil, a significant gain a fuel quality could be achieved. It also demonstrated that the large capital expenditure associated with Project Turbo and the selective catalytic cracker under construction at that time [33] was not necessary. The construction of a high-temperature fluid catalytic cracker (FCC) as main refinery intervention to meet future fuel specification was aimed at production of chemicals rather than fuels [33]. There was a desire to increase polyethylene and polypropylene production, and for that more ethene and propene were needed. The refining philosophy underlying this decision was to convert the lowest quality oxygenate-rich straight-run C6 –C7 Fischer–Tropsch naphtha into chemicals, namely ethene and propene, rather than upgrading the quality of the fuel per se. The dangers inherent in such an approach have been discussed (Section 9.5). The risk was somewhat offset by the anticipated coproduction of an aromatic motor-gasoline. In fact, the coproduction
209
210
9 Sasol 2 and 3 Facilities
of an aromatic motor-gasoline by converting the less refractory lower octane species mimics thermal reforming technology (Section 22.2). The main difference between thermal reforming and catalytic cracking in this application was that some aromatic synthesis would also take place during catalytic cracking, albeit at the expense of alkene production (Section 21.4). The KBR Superflex Selective Catalytic Cracking (SCC) technology [58, 59] was selected for this purpose. It was claimed that the SCC technology could obtain an overall ethene and propene yield of 50–70%, with an ethene to propene ratio of 1 : 2 [59]. With once-through operation based on a light FCC naphtha feed, 15% ethene and 30% propene were obtained [58]. With Fischer–Tropsch naphtha, the product slate was 16% ethene, 25% propene, 10% butenes, 2% C5 and heavier alkenes, and 24% aromatics, the rest being alkanes, CO, CO2 , H2 O, and coke. The SCC unit at Sasol Synfuels was designed for a production of 250 000 tpa propene [59]. This was the first commercial implementation of the Superflex SCC technology. Commissioning of the SCC took much longer than expected, and at the time of writing successful achievement of the design intent has not yet been reported. The basic operation of the SCC technology is similar to that of short-contact-time FCC, but the catalyst is selected to be propene selective. The temperature of operation is 50–80 ◦ C higher than ordinary FCC and the end-of-riser temperature is >600 ◦ C [59]. At such high temperatures both catalytic and thermal cracking takes place, which explains the high ethene yield claimed for the technology. By analogy, one would also expect a higher benzene yield than is found during normal FCC operation. Fundamentally, the selection of a carbon rejection technology for use with a Fischer–Tropsch naphtha is surprising and not considered efficient [60]. On the basis of the technology fundamentals (Section 21.4), problems can be anticipated. The products that were desired, namely ethene, propene, and an aromatic-rich motor-gasoline, require hydrogen rejection relative to the feed, not carbon rejection. The implication is that hydrogen rejection must occur to final products and that the aromatic-rich motor-gasoline is likely to contain a significant amount of alkanes. The product then has the same product deficiency as encountered with hydrogenated coal tar naphtha, namely, an octane number lower than anticipated on the basis of the aromatic content only. Another implicit consequence of using a carbon rejection technology to perform hydrogen rejection is that the product slate becomes very sensitive to the alkane content of the feed. The selection of the SCC technology is also surprising considering the stated aim of Project Turbo. Legislation aimed at reducing the levels of benzene and aromatics in motor-gasoline was anticipated [56]. Yet, based on the fundamentals, the SCC technology is likely to increase the production of both benzene and aromatics. In fact, the motor-gasoline is quite aromatic (Table 9.7). Contrary to the stated aim, the post-Project Turbo refinery (Figure 9.12) does not include a refining pathway to reduce the aromatics or benzene content. Technology to reduce the refinery benzene level was evaluated on a commercial scale [61] but not implemented. Whatever the ultimate technology decision, in future further refinery changes will have to be made when legislation calls for a reduction in benzene. 9.5.6 Changes in Fischer–Tropsch Aqueous Product Refining
The original design of the chemical workup section to refine the aqueous product was quite extensive, and many oxygenate chemicals were produced as final products. Even so, the chemical
9.5 Evolution of Sasol Synfuels
workup section was also expanded to produce high purity solvents and chemicals for the solvent market. The following expansions are in operation: 1) In June 1996, a 1-propanol purification plant was commissioned with a capacity of 45 000 tpa. At the time of construction, it was sufficient to supply 30% of the global demand for this compound [4]. 2) In September 1999, a high purity (99.99%) ethanol plant was commissioned [4]. 3) In May 2001, a process was commissioned for the production of ethyl ethanoate (ethyl acetate) from ethanol. The conversion of ethanol into ethyl ethanoate takes place over a CuO-based catalyst at 220–245 ◦ C and 1.4 MPa [62]. It is followed by a selective hydrogenation reactor to convert butanone (a by-product) into 2-butanol over a Ru/C catalyst at 60–120 ◦ C and 4 MPa. This is necessary for purification of the ethyl ethanoate, which is co-boiling with butanone. Other chemical production facilities were also constructed, but these have since been decommissioned. Ethanal (acetaldehyde) was used to produce crotonaldehyde by successive aldol condensation and dehydration steps. The crotonaldehyde could then be hydrogenated to produce 1-butanol. The aldol condensation step was difficult to control, causing the production of a lot of off-specification products. This plant was eventually shut down, and a 1-butanol plant based on propene hydroformylation was built in Sasolburg. Presently, the ethanal is hydrogenated to produce ethanol, using the same approach as in the original Sasol 1 design (Figure 8.9). A carboxylic acid recovery pilot plant was built in the chemical workup section to recover ethanoic acid (acetic acid) and propanoic acid (propionic acid). These acids are dissolved in the water-rich bottom product of the primary separation column (Figure 9.7). The bottom product contains about 1–2% carboxylic acids and is quite corrosive. It was found that the carboxylic acids could be selectively extracted with 2-methyl-2-methoxypropane (methyl tertiary butyl ether, MTBE). However, the pilot plant was plagued by corrosion problems and equipment failures. This resulted in poor on-stream time. Furthermore, to scale-up this process to full commercial scale would have required a large MTBE inventory as well as a very large diameter extractor. The process was very energy intensive, and this carboxylic acid recovery technology was never taken beyond the pilot plant stage [4]. 9.5.7 Changes in Coal Pyrolysis Product Refining
With the inclusion of NH3 production as part of the original design, expansion into downstream chemicals based on ammonia was logical. However, the timing for the expansion of fertilizer production was unfortunate, since South Africa experienced a severe drought that put a dampener on agricultural activity [4]. Nevertheless, in the long run all turned out well. Expansion of nitrogen-based chemicals also took place in the explosives business. In 1985, a cartridge emulsion explosives plant was commissioned in Secunda. Facilities were also constructed at Sasolburg (Section 8.5.3). A delayed coker plant was added in the tar workup section in the late 1990s, mostly with the aim of producing anode coke. The design was very flexible and allowed the production of various types and grades of coke. However, the stated objective could not be met. Because of the thermal expansion coefficient of the coke, the coke was not suitable for use as anode coke.
211
212
9 Sasol 2 and 3 Facilities
The delayed coker unit affected the tar workup section. It produced typical by-product streams that had to be refined. The coker naphtha was processed in the coal tar naphtha hydrogenation units, and coker gas oil was processed in the creosote hydrotreater. As in the case of the Fischer–Tropsch refinery, the development of more active catalysts for crude oil refining led to catalyst changes in the coal tar naphtha hydrotreater and in the creosote hydrotreater. More active catalysts were not always desirable and led to a decrease in product quality. Proper adiabatic laboratory testing also became imperative to avoid the selection of overactive hydrotreating catalysts, which would result in temperature runaways during commercial operation. 9.5.8 Synthetic Jet Fuel
Fischer–Tropsch products from Sasol Synfuels have been qualified as blending components for the production of jet fuel (Section 14.2.1). The hydrogenated kerosene range product from SPA-catalyzed alkene oligomerization, called isoparaffinic kerosene (IPK) was initially the only product allowed in a blend with crude-oil-derived material for the production of semisynthetic jet fuel. The qualification was later expanded to include the heavy naphtha from the NHT and light distillate from the DHT to produce fully synthetic jet fuel [63].
9.6 Discussion of the Refinery Design
Unlike previous Fischer–Tropsch refineries (Chapters 6–8), many catalytic refining technologies reached maturity well before Sasol 2 and 3 was designed. Unfortunately, many of the lessons that could have been learnt from previous designs were disregarded. The refinery design and many of the subsequent changes approached Fischer–Tropsch syncrude refining as if it was a conventional crude oil. This crude oil refining approach led to many problems and poor overall efficiency of the design [21]. When the syncrude was recognized for what it was, considerable gains could be achieved with minimal effort [57]. Although the facility initially focused on transportation fuel production, within the first decade of its operation there was a shift in interest toward chemical production. If Project Turbo can achieve the design intent, the ratio of fuels to chemicals production will be around 40 : 60. It was shown that, even with a simple refinery, HTFT syncrude yielded 37% chemicals and with some increase in refining complexity 40–50% of the HTFT syncrude could be converted into chemicals [64]. The increased focus on chemical extraction and production is therefore understandable. As with previous Fischer–Tropsch refineries, much can be learnt from the design, operation, and subsequent changes to the Sasol 2 and 3 facilities: 1) The distillation distribution of HTFT syncrude is uneven. The split between C3 –C4 , naphtha, distillate, and residue is approximately 32 : 44 : 17 : 7. This is very different to the distillation profile of crude oil and requires a different separation and processing approach. 2) Throughout its history, the facility suffered from poor separation. A number of design decisions contributed to the poor separation. The stepwise cooling in the gas loop did not take refinery needs into consideration. The fractionation steps in the refinery itself did
9.6 Discussion of the Refinery Design
3)
4)
5)
6)
7)
8)
9) 10) 11)
12)
not achieve clean separations. This may partly be due to oxygenates that complicate the vapor–liquid equilibrium during separation. The combination of an atmospheric distillation unit and a vacuum distillation unit served little purpose, because subsequent oxygenate conversion caused boiling point broadening. It was also an ill-conceived separation strategy, because HTFT syncrude contains little atmospheric residue. Poor separation undermined the product quality and yield from many refinery units. It highlighted a general Fischer–Tropsch syncrude refining principle: syncrude must be refined by carbon number, not by boiling range as is often the case in crude oil refining. Deoxygenation was not performed on the syncrude as a primary refinery conversion step. Although this enabled subsequent n-1-alkene extraction, it necessitated product distillation after every conversion unit. This caused a tremendous increase in complexity compared to the Hydrocol and Sasol 1 HTFT refineries. Even in instances where oxygenates were separated during chemical extraction, the benefit of separation was not exploited. Inattention to oxygenates in syncrude can cause many problems and care must be taken during refinery catalyst selection [65]. Oxygenates are reactive, produce water, have a rich chemistry, and strongly and competitively adsorb on catalytic surfaces. Selecting too active catalysts can undermine performance. The failure of the C5 –C6 isomerization process is a case in point. Overactive catalysts were also detrimental to the performance of the tar workup section, where over-hydrogenation degraded fuel quality. The refinery is inherently alkene-rich (contrary to crude oil refineries). Although chemical extraction reduced the amount of ethene, propene, 1-hexene, 1-heptene, and 1-octene that had to be refined to fuels, the motor-gasoline was still very olefinic. The key to good quality motor-gasoline production is to produce good quality alkanes. In many refinery naphtha streams, the n-alkane content reduced the octane number of otherwise high octane number components significantly. It was also necessary to hydrogenate some naphtha range alkenes, because the motor-gasoline was too olefinic. Yet, despite the crude oil refining approach followed, the refinery was surprisingly lacking in conversion pathways to produce high octane alkanes. The impact of alkene hydrogenation was especially apparent during hydrogenation of the oligomers from SPA oligomerization. The quality of unhydrogenated olefinic motor-gasoline from alkene oligomerization over SPA is insensitive to the feed and operating conditions. This changes completely when the product is hydrogenated and the quality of the motor-gasoline becomes extremely sensitive to the feed composition and operating conditions. This was not appreciated in the original design and the production of alkylate equivalent product by separate propene and butene oligomerization was not realized. Alkene oligomerization is one of the most important conversion processes in the Fischer–Tropsch refinery. There are synergies between coal pyrolysis product and Fischer–Tropsch refining. These synergies are not limited to product blending. The aqueous product is a rich source of chemicals, but recovery of the carboxylic acids is challenging. Some of these oxygenates have small global markets, and throughout the history of the facility some of these oxygenates were converted or blended into fuels. Chemical extraction can be beneficial for fuel quality. However, this benefit can be realized only if chemical extraction is guaranteed. When a fuel refinery is tightly integrated with chemicals units, no benefit of chemical extraction can accrue, because there must be a
213
214
9 Sasol 2 and 3 Facilities
refining pathway to process the chemicals when they are not extracted. In such a case, chemical production complicates fuel refining and reduces refinery operability, stability, and flexibility.
References 1. Wessels, P. (1990) Crescendo tot Sukses. Sasol
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13. 14.
1975–1987 (Engl. Transl. ‘‘Crescendo to success. Sasol 1975–1987’’), Human & Rousseau, Cape Town. Marriott, J.N. (1986) Sasol process technology – the challenge of synfuels from coal. ChemSA, 12 (8), 174–178. Dry, M.E. (1983) in Applied Industrial Catalysis, vol. 2 (ed. B.E. Leach), Academic Press, New York, pp. 167–213. Collings, J. (2002) Mind Over Matter. The Sasol Story: A Half-Century of Technological Innovation, Sasol, Johannesburg. Fenton, D.M. and Gowdy, H.W. (1979) The chemistry of the Beavon sulfur removal process. Environ. Int., 2 (3), 183–186. Vermaire, S. and De Haan, R. (1988) Influence of sodium carbonate/vanadium concentration ratios on vanadate (V) equilibria and on the reoxidation of V(IV) in a hydrogen sulfide removal process. Ind. Eng. Chem. Res., 27, 1242–1245. Holtkamp, W.C.A., Kelly, F.T., and Shingles, T. (1977) Circulating fluid bed catalytic reactor for the Fischer-Tropsch synthesis at Sasol II. ChemSA, 3 (3), 44–45. Steynberg, A.P., Espinoza, R.L., Jager, B., and Vosloo, A.C. (1999) High temperature Fischer-Tropsch synthesis in commercial practice. Appl. Catal. A, 186, 41–54. Dry, M.E. (1996) Practical and theoretical aspects of the catalytic Fischer-Tropsch process. Appl. Catal. A, 138, 319–344. Dry, M.E. (1981) in Catalysis Science and Technology, vol. 1 (eds J.R. Anderson and M. Boudart), Springer-Verlag, Berlin, pp. 159–255. Dry, M.E. (1984) High yield high quality diesel from Fischer-Tropsch process. ChemSA, 10 (2), 286–290. Hoogendoorn, J.C. (1982) Producing automotive fuels from coal in South Africa. Hydrocarbon Process., 61 (5), 34E–34Q. Joiner, J.R. and Kovach, J.J. (1982) Sasol two and Sasol three. Energy Prog., 2 (2), 66–68. Linnig, D.A., Mako, P.F., and Samuel, W.A. (1982) Coal to oil and gas. Sasol one, two, and three. Energy Process. Can., 74 (1), 49–54.
15. Bartoo, R.K. (1984) Removing acid gas by the
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Benfield process. Chem. Eng. Prog., 80 (10), 35–39. Sanyal, D., Vasishtha, N., and Saraf, D.N. (1988) Modeling of carbon dioxide absorber using hot carbonate process. Ind. Eng. Chem. Res., 27, 2149–2156. Hanekom, P. and Gibson, P. (1997) Contamination of Benfield CO2 removal system by carboxylic acid salts. Ammonia Plant Saf., 37, 281–289. Chauvel, A. and Lefebvre, G. (1989) Petrochemical Processes. 1. Synthesis-Gas Derivatives and Major Hydrocarbons, Technip, Paris, pp. 149–154. Swart, J.S., Czajkowski, G.J., and Conser, R.E. (1981) Sasol upgrades Synfuels with refining technology. Oil Gas J., 79 (35), 62–66. Marais, Z.-M. (1991) Sintese en evaluering van ’n reeks petrolmengkomponente (Engl. Transl. ‘‘Synthesis and evaluation of a series gasoline blending components’’). M.Sc. Dissertation, University of the Orange Free State, Bloemfontein, South Africa. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, FL, pp. 331–364. De Klerk, A., Engelbrecht, D.J., and Boikanyo, H. (2004) Oligomerization of Fischer-Tropsch olefins: effect of feed and operating conditions on hydrogenated motor-gasoline quality. Ind. Eng. Chem. Res., 43, 7449–7455. Doolin, P.K., Zalewski, D.J., and Oyekan, S.O. (2004) in Catalytic Naphtha Reforming, 2nd edn (eds G.J. Antos and A.M. Aitani), Marcel Dekker, New York, pp. 433–457. Lapinski, M., Baird, L., and James, R. (2004) in Handbook of Petroleum Refining Processes, 3rd edn (ed. R.A. Meyers), McGraw-Hill, New York, pp. 4.3–4.31. Gary, J.H., Handwerk, G.E., and Kaiser, M.J. (2007) Petroleum Refining. Technology and Economics, 5th edn, CRC Press, Boca Raton, FL, pp. 207–220.
References 26. De Klerk, A. (2008) Hydroprocessing peculiari-
27.
28.
29.
30.
31. 32.
33. 34.
35.
36.
37.
38.
39.
40.
ties of Fischer-Tropsch syncrude. Catal. Today, 130, 439–445. Kronseder, J.G. and Bogart, M.J.P. (1979) Coal, liquefaction, South Africa’s Sasol II. Encycl. Chem. Process. Des., 9, 299–328. Dry, M.E. (1987) Chemicals produced in a commercial Fischer-Tropsch process. ACS Symp. Ser., 328, 18–33. Leckel, D.O. (2006) Catalytic hydroprocessing of coal-derived gasification residues to fuel blending stocks: effect of reaction variable and catalyst on hydrodeoxygenation (HDO), hydrodenitrogenation (HDN), and hydrodesulfurization (HDS). Energy Fuels, 20, 1761–1766. Hoogendoorn, J.C. (1982) Gasoline from coal – the pioneering South African experience. Energy Prog., 2 (1), 32–36. Dry, M.E. (1988) The Sasol route to chemicals and fuels. Stud. Surf. Sci. Catal., 36, 447–456. Dancuart, L.P., De Haan, R., and De Klerk, A. (2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532. (2005) Gaint project turbo advances. Chem. Technol. (S. Afr.), 13–16. De Klerk, A. (2006) Distillate production by oligomerization of Fischer-Tropsch olefins over solid phosphoric acid. Energy Fuels, 20, 439–445. De Klerk, A. and De Vaal, P.L. (2008) Alkylate technology selection for Fischer-Tropsch syncrude refining. Ind. Eng. Chem. Res., 47, 6870–6877. Chauvin, Y., Hennico, A., Leger, G., and Nocca, J.L. (1990) Upgrading of C2 , C3 , C4 olefins by IFP dimersol technology. Erdol, Erdgas, Kohle, 106 (7/8), 309–315. Nel, R.J.J. and De Klerk, A. (2007) Fischer-Tropsch aqueous phase refining by catalytic alcohol dehydration. Ind. Eng. Chem. Res., 46, 3558–3565. Bolder, F.H.A. and Mulder, H. (2006) Dehydration of alcohols in the presence of carbonyl compounds and carboxylic acids in a Fischer-Tropsch hydrocarbons matrix. Appl. Catal. A, 300, 36–40. Coetzee, J.H., Mashapa, T.N., Prinsloo, N.M., and Rademan, J.D. (2006) An improved solid phosphoric acid catalyst for alkene oligomerization in a Fischer-Tropsch refinery. Appl. Catal. A, 308, 204–209. (2001) Linear Alpha Olefins, SRI Process Economics Program Report 12D, Stanford Research Institute, Menlo Park, CA.
41. Joubert, D. and Tincul, I. (2002) Ethylene
42. 43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
copolymers with Fischer-Tropsch olefins. Macromol. Symp., 178, 69–79. Ryan, P. (1945) The synthol process. Oil Gas J., 43 (47), 264–268. (1988) Physical Constants of Hydrocarbons and Non-Hydrocarbon Compounds, ASTM Data Series DS 4B, 2nd edn, ASTM, Philadelphia, PA. De Klerk, A. (2004) Etherification of C6 Fischer-Tropsch material for linear α-olefin recovery. Ind. Eng. Chem. Res., 43, 6349–6354. Hahn, T. (2003) The octene train II purification process. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd013. Van Eeden, J.A. (2003) Short chain potassium carboxylate phase behaviour. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd058. De Klerk, A. (2009) Reaction engineering in Fischer-Tropsch context. E V Murphree Award symposium in honour of Prof. M. P. Dudukovi. 237th American Chemical Society National Meeting, March 22–26, 2009, Salt Lake City, USA. Diamond, D., Hahn, T., Becker, H., and Patterson, G. (2004) Improving the understanding of a novel complex azeotropic distillation process using a simplified graphical model and simulation. Chem. Eng. Process., 43, 483–493. McGurk, K. (2003) From 1-heptene to 1-octene: a new production route. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd082. Makgoba, N.P., Sakuneka, T.M., Koortzen, J.G., Van Schalkwyk, C., Botha, J.M., and Nicolaides, C.P. (2006) Silication of γ -alumina catalyst during the dehydration of linear primary alcohols. Appl. Catal. A, 297, 145–150. Redman, A. (2005) Production of olefins and oxygenated compounds from Fischer-Tropsch. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd179. Duplan, J.-L., Amigues, P., Verstraete, J., and Travers, C. (1996) Kinetic studies of the skeletal isomerization of n-pentenes over the ISO-5 process catalyst. Proc. Ethylene Prod. Conf., 5, 429–449. Smook, D. and De Klerk, A. (2003) Catalyst mechanical properties – case study.
215
216
9 Sasol 2 and 3 Facilities
54.
55.
56. 57.
58.
59.
Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cdP09. Kamara, B.I. and Coetzee, J. (2009) Overview of high-temperature Fischer-Tropsch gasoline and diesel quality. Energy Fuels, 23, 2242–2247. South African Government Gazette, Department of Minerals and Energy; Petroleum Products Act (120/1977): Regulations: Petroleum products specifications and standards (2006) vol. 492 no. 28958 on 23 June. (2003) Sasol Annual Report, Sasol, Johannesburg, p. 52. Van Wyk, A.M., Moola, M.A., De Bruyn, C.J., Venter, E., and Collier, L. (2007) Molecule management for clean fuels at Sasol. Chem. Technol. (S. Afr.), 4–10. Len, A.S. and Pavone, T. (2004) An alternative option for producing light olefins. Pet. Technol. Q., 9 (1), 89–93. Eng, C., Heidenreich, S., Swart, S., and M¨oller, F. (2005) Clean fuels and petrochemicals at Sasol via Superflex. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd122.
60. De Klerk, A. (2008) Fischer-Tropsch refining:
61.
62.
63.
64.
65.
technology selection to match molecules. Green Chem., 10, 1249–1279. Sakuneka, T.M., Nel, R.J.J., and De Klerk, A. (2008) Benzene reduction by alkylation in a solid phosphoric acid catalyzed olefin oligomerization process. Ind. Eng. Chem. Res., 47, 7178–7183. Colley, S.W. and Tuck, M.W.M. (2003) in Catalysis in Application, (eds S.D. Jackson, J.S.J. Hargreaves, and D. Lennon), Royal Society of Chemistry, Cambridge, pp. 101–107. UK Ministry of Defence (2008) Turbine Fuel, Aviation Kerosine Type, Jet A-1. NATO Code: F-35. Joint Service Designation: AVTUR. Defence Standard 91-91, Issue 6; UK Ministry of Defence, 8 April 2008. De Klerk, A., Dancuart, L.P., and Leckel, D.O. (2005) Chemicals refining from Fischer-Tropsch synthesis. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd185. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge.
217
10 Mossgas Facility 10.1 Introduction
Despite the success of Sasol 1 (Chapter 8) and later Sasol 2 and 3 (Chapter 9) to reduce South Africa’s dependence on imported oil, energy security was still a point of concern for the South African government. In 1984, the Mossgas project was initiated to investigate the conversion of gas and associated natural gas liquids (NGL) to transportation fuel. The government gave the final go-ahead for the project in 1986 based on the gas find in the Bredasdorp Basin, off the coast of Mossel Bay. The offshore platform landed its first gas in March 1992, and construction of the Mossgas facility was completed by the middle of 1992, achieving full production in January 1993 [1, 2]. Since start-up, the facility has been running at an average on-stream factor of >90% [3]. The Mossgas facility was designed to produce 33 000 bbl/day of products. Actual production is higher, about 36 000 bbl/day [4, 5]. Of the total, 10 500 bbl/day came from associated NGL, which were produced with the natural gas at a rate of 70 m3 ·h−1 [1]. The remainder, 22 500 bbl/day, were produced by Fischer–Tropsch synthesis. The project had capital cost of close to US$2.4 billion at that time, but with the large unexpected drop in the oil price in the 1990s to less than US$20 bbl−1 , the South African government was severely criticized for ‘‘wasting’’ tax payer’s money. Despite the criticism, the project had a production cash cost of US$9–10 bbl−1 , which resulted in a significant operating profit [1]. The Petroleum Oil and Gas Corporation of South Africa (PetroSA) was formed in January 2002 by merging the state-owned petroleum assets. The Mossgas gas-to-liquids facility is now operated by PetroSA and it is no longer referred to as Mossgas [5]. The facility is still in operation. Unlike the Sasol facilities, the PetroSA facility was not significantly modified (Section 10.5), and its modern refinery design makes it well suited for fuels production. The product slate is 90% fuels and 10% chemicals. The main fuel products are on-specification motor-gasoline and diesel fuel for the South African market (Table 10.1) [2]. The main chemicals are 135 000 m3 per year of alcohols and 60 000 m3 per year of specialty distillates that are exported [4]. The latter is a product of a low aromatics distillate hydrotreater (DHT) that was later added to the refinery. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
218
10 Mossgas Facility Table 10.1
Final fuel products from the Mossgas refinery.
Final product
Propane Liquid petroleum gas Motor-gasoline (unleaded RON 95) Kerosene (illuminating paraffin) Diesel fuelb Fuel oil (low sulfur) Total fuels production
Production volumea (m3 /day)
(bbl/day)
85 265 2760 600 1400 100 5210
530 1665 17 360 3775 8800 630 32 760
a Does b The
not include the contribution of LTFT synthesis (since 2005). production of low aromatic diesel (since 2003) is not shown.
10.2 Synthesis Gas Production 10.2.1 Natural Gas Liquid Recovery
The natural gas is produced by an offshore production platform about 85 km south of Mossel Bay. The offshore production platform supplies the onshore Mossgas facility with a mixture of natural gas and NGL. In 2005, the reported production rate of natural gas was 230 000 m3 ·h−1 (normal) and that of NGL was 65 m3 ·h−1 [5]. The natural gas and NGL are separated in the NGL recovery section. Part of the natural gas can also be diverted for liquefaction in a liquefied natural gas (LNG) plant. The purpose of the LNG plant is to provide a backup supply of natural gas during upsets at the offshore production platform. The natural gas backup is sufficient for 24 h of production [1, 3]. Typical compositions of the natural gas, NGL, and products from the NGL recovery section are given in Table 10.2. The products from the NGL recovery section are gas, which is used as gas reformer feed material; a C4 fraction; and the gas condensates (Figure 10.1) [2]. The gas condensates derived from the NGL should not be confused with the Fischer–Tropsch condensates. The gas condensates from the NGL recovery section are further fractionated to produce naphtha feed to the refinery, a straight-run diesel blending component, and a fuel oil fraction. 10.2.2 Gas Reforming
The methane-rich gas from the NGL recovery section is desulfurized. The desulfurization takes place in two desulfurization reactors loaded with ZnO. The natural gas feed was not hydrotreated before desulfurization. This led to some problems during early operation, because the ZnO was ineffective in removing organic sulfur compounds. The situation was exacerbated by the
10.2 Synthesis Gas Production Composition of the natural gas and natural gas liquids (NGL) from the offshore production platform, as well as the products from the NGL recovery section.
Table 10.2
Compound
H2 S N2 CO2 H2 O Methane Ethane Propane n-Butane Isobutane Pentanes Hexanes C7 -120 ◦ C 120–180 ◦ C 180–400 ◦ C >400 ◦ C
Offshore feed (mass%)
Products from NGL recovery (mass%)
Natural gas
NGL
Gas
C4 s
8 µg·g−1 2.2 5.4 70 µg·g−1 70.5 11.0 6.4 3.1a
– 0.0 1.2 60 µg·g−1 6.0 5.2 8.7 10.4a
8 µg·g−1 2.3 5.9 <0.1 74.8 12.9 3.9 0.2 – – – – – – –
– – – – – – 0.5 61.3 37.1 1.1 – – – – –
1.4b – – – – –
68.5b – – – – –
Condensate – – – – – – – 0.3 <0.1 16.8 7.1 21.2 21.1 29.2 4.3
a Total b
butanes. All C5 and heavier hydrocarbons.
formation of organic sulfur compounds by the methanol that was injected at the offshore platform to prevent hydrate formation in the pipeline. The problem was eliminated by loading some CoMo–Al2 O3 on top of the ZnO bed and including a recycle with 3 vol% H2 to the feed. This hydrogenates the organic sulfur compounds to produce H2 S, which is then efficiently removed by the ZnO [6]. The desulfurized gas forms the feed to the gas reformers. The Mossgas facility employs the Lurgi Combined Reforming Process, which is a two-step reforming approach. The primary reformer is a tubular steam reformer. Only part of the natural gas feed is steam reformed. Steam reforming is mainly used for hydrogen production, since it produces a H2 :CO ratio well above that required for Fischer–Tropsch synthesis. The product gas from steam reforming is combined with the remainder of the natural gas feed and the tail gas from Fischer–Tropsch synthesis that is employed as external recycle (Figure 10.1). The combined feed contained 34.5% H2 , 4.8% CO, 7.0% CO2 , 26.3% H2 O, 22.9% CH4 , and 1.8% N2 , the remainder being heavier hydrocarbons [7]. The combined feed is converted with steam and O2 in a secondary adiabatic oxidative (autothermal) reformer. Steam to carbon ratios down to 1.4 were employed [7]. The synthesis gas outlet temperature from the autothermal reformer is 960 ◦ C, which is used to generate 11 MPa steam at a production rate of 480 t·h−1 . This steam is used to drive a 90 MW turbine generator [5]. The secondary reformers experienced corrosion by metal dusting, which led to some equipment failure. The reason for this type of corrosion and the preventative measures taken have been documented [6, 7].
219
220
10 Mossgas Facility
N2
Air
Air separation
NGL recovery
CO2 removal
PSA H2
O2
LNG plant Natural gas and NGL
CO2
CH4 rich
Autothermal reformer
Synthol HTFT
Steam reformer
Tail gas processing H2 , C1 – C2
Fuel gas H2 Aqueous product SLO DO C3 condensate C4 condensate C5 + condensate
Purge gas External recycle Mixed C4 s
NGL C4 s NGL naphtha
Gas condensate
NGL distillate NGL fuel oil
Figure 10.1 Mossgas natural gas liquid (NGL) recovery, synthesis gas generation, and Fischer–Tropsch gas loop.
There are three identical reforming trains serving the three Synthol Fischer–Tropsch reactors. The synthesis gas to each Synthol reactor has a H2 :CO ratio of around 3 : 1 [1]. A production of 850 000 m3 ·h−1 of synthesis gas can be sustained under normal operating conditions [5].
10.3 Fischer–Tropsch Synthesis
The Mossgas facility was the first gas-to-liquids facility to be built since Hydrocol (Chapter 7) and, like the Hydrocol facility, it made use of high-temperature Fischer–Tropsch (HTFT) synthesis. The Fischer–Tropsch technology that was selected for the Mossgas project was the tried-and-tested Sasol Synthol circulating fluidized bed (CFB) reactor design, which was then in operation at Sasol 2 and 3 (Section 9.3). The design decision was made in 1988 and was just not in time to benefit from the new Sasol Advanced Synthol (SAS) fixed fluidized bed design (Section 9.5.2). Sasol was only willing to license the HTFT Synthol CFB reactor technology. Nevertheless, PetroSA claimed better on-stream availability with the Synthol CFB reactors than was obtained with the SAS reactors. The same fused-iron Fischer–Tropsch catalyst and similar operating conditions (330–360 ◦ C and 2.5 MPa) as the Synthol reactors at the Sasol 2 and 3 facilities were employed. A constant catalyst age is maintained in the reactors by online catalyst addition and removal. The product distribution from HTFT synthesis is therefore similar to that reported for Sasol 2 and 3 (Tables 9.1 and 9.4). The rated design capacity of the three Synthol reactors was 7500 bbl/day each. However, the Mossgas Synthol reactors produced close to 8000 bbl/day [8]. This increase in capacity is not due to the mechanical design but due to the synthesis gas composition. The synthesis gas in the
10.3 Fischer–Tropsch Synthesis
Mossgas facility contained less inert material, which resulted in a higher partial pressure of the syngas components. This in turn caused the reactor productivity to increase. Typical production figures from the Mossgas HTFT synthesis section are 50 m3 ·h−1 C3 –C4 gas, 100 m3 ·h−1 light oil, and 200 m3 ·h−1 aqueous product containing 10% extractable oxygenates. The reaction heat during Fischer–Tropsch synthesis is used to generate 450 t·h−1 steam, most of which is used as feed to the gas reformers. Catalyst consumption is equivalent to 15 000 tons per year of iron ore [5]. 10.3.1 Gas Loop Design
About 30% of the synthesis gas from the secondary reformer is sent to a Benfield unit for CO2 removal (Figure 10.1). CO2 is removed by chemical absorption in a hot potassium carbonate solution [9]. The chemistry of the Benfield process has been discussed previously (Section 9.3.1). Part of the CO2 -free synthesis gas from the Benfield unit serves as feed to the pressure swing absorption (PSA) unit for refinery hydrogen production. The rest of the CO2 -free synthesis gas is combined with the CO2 -containing synthesis gas and is used for Fischer–Tropsch synthesis. The Fe-HTFT catalyst is active for the water gas shift reaction, and Fischer–Tropsch synthesis is controlled by the Ribblett ratio (Equation 4.16) rather than the H2 :CO ratio. The syngas composition can be adjusted by the extent of CO2 removal in relation to the H2 :CO ratio of the synthesis gas. The syncrude is recovered from the hot gaseous product after HTFT synthesis by stepwise cooling (Figure 10.2). The heaviest fraction is the decanted oil, which is obtained from the first Tail gas C3 condensate Tail gas processing
C4 condensate C5 + condensate
Hot quench tower
Tail gas
Light oil
Synthol product Slurry oil
Mixersettler units
Aqueous product Decanted oil
Gunk
Figure 10.2
Reaction water
Solid waste
Syncrude recovery after Fischer–Tropsch synthesis by stepwise cooling.
221
222
10 Mossgas Facility
condensation product after removing the suspended catalyst particles in mixer-settlers. Further cooling yields the stabilized light oil (SLO), aqueous product, and tail gas. Tail gas processing does not include cryogenic separation. There are no downstream petrochemical facilities close to the Mossgas facility to exploit the C2 hydrocarbons, and the refinery has not been designed to process the C2 hydrocarbons. The H2 , CH4 , ethane, and ethene in the tail gas are therefore not recovered as separate products. The mixture is instead recycled to the secondary reformer. The heavier condensate products in the tail gas are separated by pressure distillation to yield C3 , C4 , and C5 + condensate streams. Unlike the HTFT condensate 1, 2, and 3 streams at Sasol 2 and 3, the Mossgas condensates are proper distillation cuts, making Fischer–Tropsch refining more efficient. The C3 condensate contains 87% propene, 12% propane, 0.2% C2 , and 0.8% C4 hydrocarbons. The C4 condensate contains 85% butenes and 13% butanes, and the remainder is C3 and C5 hydrocarbons. The C5 + condensate contains less than 1% C4 material.
10.4 Fischer–Tropsch Refining
The processing of NGL is integrated with the refining of the Fischer–Tropsch syncrude in the Mossgas facility. There is no separate workup section dealing with NGL. The Fischer–Tropsch-derived condensates are likewise integrated with the refining of the Fischer–Tropsch oil fraction. The refinery therefore consists of only two sections, namely an oil workup section to refine all the hydrocarbons and a chemical workup section to refine the water-soluble oxygenates in the reaction water. 10.4.1 Oil Refining
There is tight integration between the Fischer–Tropsch syncrude and NGL in the Mossgas oil workup section (Figure 10.3) [2]. Although the Mossgas refinery was constructed hardly a decade after the Sasol 2 and 3 refineries, the design is modeled on a more modern fourth-generation crude oil refinery (Section 2.4.4). This can clearly be seen from the quality of many of the refinery products (Table 10.3). The coprocessing of NGL makes the design better suited to an oil refining approach; yet, the design took far more cognizance of the nature of the Fischer–Tropsch products. This made the design more efficient. The refinery does not have separate atmospheric and vacuum distillation columns, on account of the small residue fraction of the HTFT syncrude. Light oil distillation is conducted in a near-atmospheric distillation column under slight vacuum. This avoids thermal decomposition problems in the reboiler and achieves proper separation between distillate and residue. The refinery also boasts of an oligomerization process that was specifically designed for the upgrading of Fischer–Tropsch alkenes to distillate, namely the Conversion of Olefins to Distillates (COD) technology [10]. The abundance of butanes from the NGL recovery section and abundance of butenes from the HTFT condensate made alkylate production a natural choice. The n-butane is hydroisomerized
10.4 Fischer–Tropsch Refining
NGL C4s
C4 isomerization
Motor-gasoline (butane)
Alkylate
Motor-gasoline (alkylate)
Aliphatic alkylation
Isobutane
C4 condensate C3 condensate
n-Butane
Alkene oligomerization Naphtha
C5 + condensate
C5 – C6
Distillate
NGL naphtha
C5 – C6 hydroisomerization C5 – C6
C5 –120 °C
SLO DO NGL fuel oil
Motor-gasoline (olefinic)
120 –180 °C
Naphtha hydrotreater
Catalytic reformer
Motor-gasoline (isomerate) Motor-gasoline (reformate)
Kerosene
Distillate hydrotreater
Diesel fuel
NGL distillate Fuel oil Figure 10.3
Mossgas oil workup section.
Density, research octane number (RON), motor octane number (MON), and cetane number from key streams in the Mossgas oil refinery.
Table 10.3
Mossgas refinery product
Selected fuel properties Density
Motor-gasoline Olefinic COD naphtha C5 –C6 isomerate HF alkylate Reformate (RON 95 severity) Butane Diesel fuel Hydrogenated COD distillate Hydrogenated straight-run distillatea a Product
(kg·m−3 )
RON
MON
Cetane number
740 650 720 770–780 580–585
81–85 84 94–95 95 94
74–75 82 92–93 84–85 90
– – – – –
790–800 820
– –
– –
51–54 56
is a mixture: 80% from the distillate hydrotreater (DHT) and 20% NGL distillate.
to produce isobutane over a chlorinated Pt–Al2 O3 catalyst [11]. The UOP Butamer technology was selected for this purpose. The isobutane is reacted with the HTFT-derived butenes in a hydrofluoric acid (HF) catalyzed aliphatic alkylation process [12]. The alkylation technology was also licensed from UOP. Aliphatic alkylation produces a product that is rich in trimethylpentanes (iso-octanes). Alkylate is a high octane number paraffinic motor-gasoline blending component.
223
224
10 Mossgas Facility
The selection of an HF-catalyzed aliphatic alkylation process is surprising. The HTFT syncrude is rich in 1-butene, which is the worst butene feed for HF alkylation [13]. The HF alkylation of 1-butene (RON = 94 and MON = 92) has a lower octane number than the equivalent H2 SO4 alkylation of 1-butene (RON = 98 and MON = 94). The higher octane number of H2 SO4 -catalyzed alkylation is due to double bond isomerization of the 1-butene. Nevertheless, alkylate produced by HF alkylation is still good quality motor-gasoline. The HTFT C3 condensate, HTFT C5 + condensate, and C5 –C8 (C5 -120 ◦ C) overhead fraction from oil distillation are combined to provide an alkene-rich feed for oligomerization. The COD technology is an oligomerization process that has been developed by the South African Central Energy Fund and S¨ud-Chemie [10]. The process is in many respects similar to the Mobil Olefins to Gasoline and Distillate (MOGD) process [14]. The COD technology also uses an H-ZSM-5 zeolite catalyst (S¨ud-Chemie COD-9) to catalyze oligomerization [15]. The Fischer–Tropsch-derived feed is rich in alkenes and contains oxygenates at percentage level. The technology therefore serves as an oligomerization and deoxygenation process. The boiling point broadening that takes place as a result of deoxygenation is of no consequence, since the alkenes produced by deoxygenation are in any case oligomerized with the Fischer–Tropsch alkenes. The oxygenates in the feed posed challenges, the foremost being severe organic acid corrosion in the COD unit [5]. The acid-catalyzed pathway for the conversion of ketones into carboxylic acids is quite prevalent in the catalysis of Fischer–Tropsch syncrude [16]. Two distinct operating regimes were found during COD operation. At a weighted-average bed temperature <280 ◦ C, the aqueous product from oligomerization contained 1.1 mg KOH/g acids, and at >280 ◦ C the aqueous product contained only 0.1 mg KOH/g acids [17]. The oligomerization unit has three reactor trains in parallel, each consisting of three reactors in series [5, 17]. This configuration allows one of the reactor trains to be taken off-line at a time for in situ catalyst regeneration. During oligomerization, coking deactivates the catalyst, and catalyst regeneration is accomplished by controlled carbon burnoff. For distillate mode operation, the operating conditions are typically 200–320 ◦ C and 5.5 MPa. Typical COD production figures are 1100 m3 ·h−1 (normal) fuel gas, 4 m3 ·h−1 propane, 2 m3 ·h−1 butane, 15 m3 ·h−1 olefinic naphtha, and 46 m3 ·h−1 distillate [2]. The olefinic naphtha from the COD oligomerization process (Table 10.3) is of a lower quality than that from phosphoric acid-catalyzed oligomerization employed in other Fischer–Tropsch refineries (Chapters 6–9). The fuel properties are inherent to the type of catalysis used, and ZSM-5 produces a more linear product due to the pore constraining geometry of the catalyst [18]. The C5 –C6 naphtha fraction is routed to the naphtha hydrotreater (NHT). The heavier olefinic COD naphtha is used as a blending component in the motor-gasoline, but part of it can also be processed through the NHT (Figure 10.3). The distillate from the COD process is hydrogenated in the DHT. The hydrogenated COD distillate is of high quality (Table 10.3). The NHT is employed as feed pretreatment for both the C5 –C6 hydroisomerization and catalytic naphtha reforming units. The feed to the NHT is a combination of light oil (120–180 ◦ C boiling range), COD naphtha, NGL naphtha, and naphtha from the DHT. The NHT is licensed from UOP and has the dual function of hydrodeoxygenation (HDO) and alkene saturation. The removal of oxygenates is especially important. Both processes downstream from the NHT are sensitive to water, because they employ chlorinated Pt–Al2 O3 -based catalysts.
10.4 Fischer–Tropsch Refining
Many of the operating problems that were found on the NHT were related to the oxygenates in the syncrude feed to this unit [5]: 1) Carboxylic acid corrosion of the feed pumps to the unit. 2) Fouling of the NHT feed–product heat exchangers. 3) High pressure drop developing over the NHT reactors due to metals originating from the Fischer–Tropsch catalyst. The hydroisomerization unit uses licensed UOP Penex technology [19]. The n-alkane-rich C5 –C6 material is hydroisomerized to increase the content of branched alkanes and thereby the octane number of the product. The isomerate is used as a blending component in motor-gasoline. Once-through operation of this type of C5 –C6 hydroisomerization unit typically yields an isomerate with RON = 84 and MON = 82 [19]. Higher octane numbers can be achieved by separating and recycling the n-alkanes in the isomerate. The catalytic naphtha reformer uses licensed UOP Platforming technology [20]. The reformate is a blending component in the final motor-gasoline. The distillate fraction from oil distillation, together with the olefinic COD distillate, forms the feed for the DHT. The DHT technology has been licensed from IFP. The function of the DHT is HDO and alkene saturation. The hydrotreated product is fractionated downstream from the reactor. The naphtha fraction (<180 ◦ C), which is formed mainly due to boiling point broadening on hydrogenating the oxygenates in the distillate, is routed to the NHT unit. A kerosene cut (180–230 ◦ C) is made to produce illuminating paraffin, and a gas oil cut is made for blending into diesel fuel. The heaviest fraction is not further upgraded and the sulfur-free fuel oil is the final product. 10.4.2 Aqueous Product Refining
The chemical workup section processes a Fischer–Tropsch aqueous product of a similar composition to that of Sasol 2 and 3 (Table 9.4). The refining approach is different though, and the chemical workup section has been simplified by converting all carbonyl compounds into alcohols (Figure 10.4). The aqueous product is separated in the primary distillation column into an oxygenate-rich overhead product and a water-rich bottom product. Rather than separating the water-soluble Primary distillation
Aqueous product
Carbonyl hydrogenation
Oxygenate rich
H2 Alcohols and water
Methanol Alcohol recovery Water
Light alcohols Heavy alcohols
Acid water
Wastewater Figure 10.4
Aqueous product refining in the Mossgas chemical workup section.
225
226
10 Mossgas Facility
organic compounds in the overhead product, the carbonyls are partially hydrogenated to produce alcohols [1]. The carbonyl to alcohol hydrogenation is performed with a Ni/SiO2 –Al2 O3 catalyst (S¨ud-Chemie G-134). This catalyst was found to be very stable in this application [21]. The product from partial hydrogenation contains a mixture of alcohols and water. The water is removed, and the alcohols are separated into mainly two fractions, a light alcohol mixture and a heavy alcohol mixture. The methanol content in HTFT syncrude is very low. The light alcohol mixture contains ethanol and 2-propanol as main constituents. In the original design, this stream was used as a motor-gasoline blending component, but it has since been marketed as solvent. The heavy alcohol mixture contains 1-propanol and butanol isomers as main constituents. This product is sold as solvent, but it can also be added to the diesel fuel to produce a 5% oxygenated diesel [22]. The bottom product from the primary distillation column (Figure 10.4) is acid water, and it contains mainly water with 1–2% carboxylic acids and a small amount of heavy oxygenates. The carboxylic acids in the acid water are substantially converted into CH4 and CO2 in an anaerobic biological water treatment system. The bio-derived gas is used in the facility, thereby effectively increasing the carbon efficiency of the process. Although the commissioning time for the anaerobic system was long due to the slow growth rate of the microorganisms, it is clearly a successful and carbon efficient Fischer–Tropsch wastewater treatment strategy.
Table 10.4 Selected properties of the motor-gasoline and diesel fuel refined from Fe-HTFT syncrude and NGL at Mosgass.
Fuel property
Liquid product yield (%) Density at 15 ◦ C (kg·m−3 ) RON MON Cetane number Cold filter plugging point (◦ C) Distillation (◦ C) IBP (initial boiling point) T90 FBP (final boiling point) Composition (%) Aromatics Sulfur a Contains
Motor-gasoline
Diesel fuel
Leaded
Unleaded
Typical
Oxygenateda
66 710–720 97/93b 89b – –
66 710–720 95 85 – –
34 810 – – 51–53 –14
– 806.5 – – 49.3 –
– – –
– – –
220 320 360
81 318 363
– <0.01c
– –
14–16 <0.001c
15.9 <0.001c
5% heavy alcohols from the chemical workup section. addition as tetraethyl lead (TEL) of 0.4 g Pb/l, equivalent to 1 ml TEL/gal. c Sulfur level usually well below 10 µg·g−1 . b Lead
10.5 Evolution of the PetroSA Facility
10.4.3 Synthetic Fuel Properties
The Mossgas was commissioned when South African fuel specifications still allowed the addition of tetraethyl lead as octane booster for motor-gasoline. However, Mossgas was the first refinery in South Africa to produce certified unleaded 95 octane motor-gasoline [1]. As in the case with the original Sasol 2 and 3 designs, the blending design allowed for the inclusion of alcohols in the fuel. Transportation fuel properties are given in Table 10.4 [1, 2, 22]. Some of the subsequent changes to the Mossgas refinery (Section 10.5) influenced the properties of the diesel fuel. The properties of the diesel fuel have been reported as >54 cetane number, <5% aromatics content, and <5µg·g−1 sulfur content [5].
10.5 Evolution of the PetroSA Facility 10.5.1 Addition of Low-Temperature Fischer–Tropsch Synthesis
In January 2005, a new 1000 bbl/day Fischer–Tropsch synthesis unit was commissioned at the PetroSA facility. This unit employed low-temperature Fischer–Tropsch (LTFT) synthesis. The technology was developed and licensed by Statoil of Norway and used a Co-LTFT catalyst in a slurry bubble column reactor [5]. The composition of synthesis gas from the reforming section, which has a high H2 :CO ratio, was adjusted in a membrane unit to match the usage ratio over the Co-LTFT catalyst. The syncrude from LTFT synthesis included waxes and extended the product slate of the PetroSA facility to heavier products. Although the LTFT synthesis section is small compared to the HTFT synthesis section, it effectively converted the facility to an HTFT–LTFT operation. 10.5.2 Changes in the Fischer–Tropsch Refinery
There is a value-added market for low aromatic distillate (LAD) range of products as indoor heating fuels and specialty fluids, such as drilling fluids. Initially, part of the distillate range material was toll-hydrogenated off-site to produce LAD, and in 2000 approval was given to construct a 70 000 ton per year LAD plant in the refinery. The LAD unit was commissioned in August 2003 [4]. In the LAD unit, a noble metal catalyst is employed for deep hydrogenation of the distillate to convert all the aromatics into cycloalkanes [4]. The impact of extracting 1-hexene from the HTFT syncrude before the feed is sent to COD oligomerization has been evaluated [15, 17]. As one would anticipate from the ZSM-5 acid catalysis of the COD process, the extraction of 1-hexene had little effect on the overall process and product distribution. The main impact was lower production, because the flow rate of alkenes was reduced by removing 1-hexene from the feed.
227
228
10 Mossgas Facility
The possible extraction of specialty chemicals in future has been indicated by PetroSA, specifically the recovery of propene, 1-hexene, 1-octene, and carboxylic acids [5].
10.6 Discussion of the Refinery Design
The Mossgas refinery was modeled on a fourth-generation oil refinery, but care was taken to adapt the design for syncrude. It produces Euro-4 type (EN228:2004) motor-gasoline and, apart from the low density of the diesel fuel, the diesel fuel is also Euro-4 (EN590:2004) compliant. Overall, the refinery is less complex than a fourth-generation oil refinery and demonstrates that it is easier to refine HTFT syncrude to on-specification fuels than crude oil. It also demonstrates the value of taking syncrude properties into account during refinery design, rather than imposing a crude oil design on syncrude [23]. Important aspects of the refinery design are the following: 1) Good separation in the gas loop had many benefits in the refinery. By routing the right carbon numbers to the right units, refining efficiency was increased. The principle of carbon number refining has been applied to the lighter products. Proper separation of the HTFT condensate and NGL condensate allowed clean C4 refining. The separation strategy also enabled future chemical extraction of propene and 1-hexene with little disruption to the refinery. The use of a near-atmospheric distillation enabled clean separation between distillate and residue while reducing the risk of thermal decomposition in the reboiler. 2) Carbon number broadening was not avoided during hydrotreating of the heavier fractions. This was the consequence of a trade-off. It was necessary to separate the light naphtha (C5 -120 ◦ C) from the light oil for oligomerization. Rather than just topping the light oil, the light oil was fractionated into different boiling fractions, the naphtha and distillate being hydrotreated separately. The resulting boiling point broadening could have been avoided by topping and then hydrotreating all of the light oil before distillation. 3) Alkenes and oxygenates were central to the refinery design. 4) Alkene oligomerization of a broad cut (C3 –C8 ) of the HTFT naphtha was an important aspect of the refinery design. This unit increases refinery flexibility, since its design in principle allows the processing of heavier material and different carbon number ratios in the feed. This is a consequence of the catalysis selected for this process, which was specifically developed to convert Fischer–Tropsch-derived naphtha. 5) Aqueous product refining was considerably simplified by partial hydrogenation of the carbonyls to alcohols before further separation. By converting the carbonyls to alcohols, it also reduced the number of oxygenate products and improved the economy of scale of alcohol recovery. This strategy also allowed flexibility in terms of fuels and chemicals production. PetroSA demonstrated blending pathways for inclusion of the alcohols into fuels. 6) Carboxylic acids and metal carboxylates were important species to consider in the refinery design. The metallurgy of equipment employed for oxygenate refining had to be selected to cope with carboxylic acids. This was also true of units in which carboxylic acids were produced from ketones. Metal carboxylates caused pressure drop problems. Wastewater treatment exploited the carboxylic acids through anaerobic digestion to improve the overall carbon efficiency of the process.
References
7) The refinery design included conversion units to produce high octane number paraffinic motor-gasoline blending components. Such units are invaluable for the production of high octane unleaded motor-gasoline. 8) The refinery design highlights the benefits of combined HTFT–LTFT syncrude refining. The NGL liquids have much in common with LTFT syncrude, and the later addition of LTFT synthesis integrated well with the overall design. Both NGL and LTFT syncrude consist mainly of n-alkanes. The Co-LTFT syncrude contains some alkenes and oxygenates, especially in the lighter fractions, but these can be synergistically co-refined with the HTFT syncrude, which contains even more alkenes and oxygenates. 9) The refinery design focused on fuels production, but it is clear that with little additional effort considerable chemical potential can be realized. This has not yet come to pass, but the refinery design is robust and able to produce on-specification fuels with and without chemical extraction. This implies that chemical extraction can be implemented as a value addition strategy, and many of the risks outlined before (Section 9.5) can be avoided. References 1. Terblanche, K. (1997) The Mossgas challenge. 2.
3.
4.
5.
6.
7.
8.
9.
Hydrocarbon Eng., 2–4. Steyn, C. (2001) The role of Mossgas in Southern Africa. 2nd Sub-Saharan Africa Catalysis Symposium, November 5–7, 2001, Swakopmund, Namibia, Paper 2. Terblanche, K. (1999) Value added synthetic fluids key to Mossgas’ success. Oil Gas J., 97 (49), 48–52. Mabena, N. (2005) Operating the worlds largest GTL facility (Natural gas-to-liquids). Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg. p. cd187. Minnie, R., Knottenbelt, C., Clur, J., Grond, W., Karodia, M., and De Wet, H. (2005) in Fundamentals of Gas to Liquids, 2nd edn, Petroleum Economist, London, pp. 22–24. De Wet, H., Minnie, R.O., and Davids, A.J. (1998) Postcommissioning operating experience at the Mossgas plant. Ammonia Plant Saf., 38, 64–72. Holland, M.L. and De Bruyn, H.J. (1996) Metal dusting failures in methane reforming plant. Int. J. Pres. Ves. Pip., 66, 125–133. Steynberg, A.P., Espinoza, R.L., Jager, B., and Vosloo, A.C. (1999) High temperature Fischer–Tropsch synthesis in commercial practice. Appl. Catal. A, 186, 41–54. Bartoo, R.K. (1984) Removing acid gas by the Benfield process. Chem. Eng. Progr., 80 (10), 35–39.
10. K¨ ohler, E., Schmidt, F., Wernicke, H.J.,
11.
12.
13.
14.
15.
16.
17.
18.
De Pontes, M., and Roberts, H.L. (1995) Converting olefins to diesel–the COD process. Hydrocarbon Technol. Int., 37–40. Travers, C. and Leprince, P. (2001) Petroleum Refining, Conversion Processes, Vol. 3, Editions Technip, Paris, pp. 229–256. Joly, J.-F. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 257–289. De Klerk, A. (2008) Fischer–Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. Quann, R.J., Green, L.A., Tabak, S.A., and Krambeck, F.J. (1988) Chemistry of olefin oligomerization over ZSM-5 catalyst. Ind. Eng. Chem. Res., 27, 565–570. Minnie, O.R., Petersen, F.W., and Samadi, F.R. (2003) Effect of 1-hexene extraction on the COD process conversion of olefins to distillate. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd083. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Minnie, O.R. (2006) The effect of 1-hexene extraction on the COD process. MTech dissertation, University of South Africa, Pretoria. De Klerk, A. (2007) Properties of synthetic fuels from H-ZSM-5 oligomerisation of Fischer–Tropsch type feed materials. Energy Fuels, 21, 3084–3089.
229
230
10 Mossgas Facility 19. Cusher, N.A. (2004) in Handbook of Petroleum
Refining Processes, (ed. R.A. Meyers), 3rd edn, McGraw-Hill, New York, pp. 9.15–9.27. 20. Lapinski, M., Baird, L., and James, R. (2004) in Handbook of Petroleum Refining Processes (ed. R.A. Meyers), 3rd edn, McGraw-Hill, New York, pp. 4.3–4.31. 21. Nel, R.J.J. and De Klerk, A. (2007) Fischer–Tropsch aqueous phase refining by
catalytic alcohol dehydration. Ind. Eng. Chem. Res., 46, 3558–3565. 22. Knottenbelt, C. (2002) Mossgas ‘‘gas-to-liquids’’ diesel fuels – an environmentally friendly option. Catal. Today, 71, 437–445. 23. De Klerk, A. (2009) in Advances in Fischer–Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 331–364.
231
11 Shell Middle Distillate Synthesis (SMDS) Facilities 11.1 Introduction
Shell started with the development of their gas-to-liquids (GTLs) technology in the 1970s. The work was motivated by an increase in the discovery of gas reserves in remote locations. Natural gas is a useful energy carrier in its own right, but the markets for natural gas are not always accessible from remote gas fields. The low energy density of natural gas necessitates its liquefaction in some way to enable transport if markets cannot be reached by pipeline. The reasoning behind the decision to develop an indirect liquefaction process has been stated as follows [1]: 1) 2) 3)
4)
5)
In cases where liquefied natural gas (LNG) or pipeline transport to natural gas markets is not feasible, GTLs technology can enable transportation. The production of synthetic fuels holds a strategic benefit for the country where it is produced by providing energy security. Synthetic fuel production may attract local government subsidies, or the products may sell at a premium owing to the high cost of importing fuels, especially to remote and inland locations. The inherent quality of synthetic fuels produced by low-temperature Fischer–Tropsch (LTFT) synthesis, being sulfur-free and aromatics-free, allows the fuels to be used as quality improvers for conventional crude-oil-derived fuels. Globally, the long-term demand for distillates may be more than that for motor-gasoline.
The development phase of the technology started in 1983 with the construction of a pilot plant at the Shell Research and Technology Centre in Amsterdam. This work culminated in the design of the first industrial-scale plant based on the Shell Middle Distillate Synthesis (SMDS) technology. The site for the first application of the SMDS technology was Bintulu, Malaysia. The Shell MDS (Malaysia) Snd Bhd company was formed in April 1986 with 60% share holding by Shell Gas BV, 10% by the Sarawak State Government, 10% by Petronas, and 20% by Diamond Gas Holdings, a subsidiary of Mitsubishi Corporation. The natural gas source was the offshore Central Luconia gas fields [2]. Shell completed the project specification for the SMDS facility at Bintulu in June 1988, and in 1989 JGC Corporation of Japan was selected as main contractor for the project. The official groundbreaking ceremony took place on 3 November 1989. Construction was completed in May Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
232
11 Shell Middle Distillate Synthesis (SMDS) Facilities
1993 and the facility was commissioned. The first shipment of products from the Bintulu facility took place in September 1993 [2, 3]. The Bintulu facility was designed for a production capacity of 12 500 bbl/day. Initial on-stream availability was low compared to the 98% availability currently achieved. The plant trips were mainly due to problems with synthesis gas generation. The Fischer–Tropsch synthesis and refinery sections operated smoothly from the start [3]. Production was interrupted when an explosion took place in the air separation unit at the Bintulu facility on 25 December 1997. Subsequently, the facility was reconstructed and recommissioned in 2000. During the statutory shutdown in 2003, some minor modifications were made, which increased the production capacity to 14 700 bbl/day [3]. The second facility to make use of the SMDS technology is, at the time of writing, under construction at the Ras Laffan Industrial City in Qatar. The Pearl GTL project was announced in July 2006 and has a design capacity of 140 000 bbl/day of synthetic liquids. The Pearl GTL facility will consist of two back-to-back GTL trains of 70 000 bbl/day each. In addition to the synthetic liquids, Pearl GTL will also process 120 000 bbl/day of associated natural gas liquids (NGL), making it a 260 000 bbl/day facility. The foundation stone was laid on 26 February 2007 [3–5]. The two SMDS facilities, Bintulu GTL (Sections 11.2–11.4) and Pearl GTL (Section 11.5), are discussed separately. The SMDS technology is described for the Bintulu facility. The Pearl GTL facility makes use of the same basic design, but on a much larger scale and with some scale-up of the individual processes. Once the Pearl GTL facility is fully commissioned, which is expected in 2011, it will be the largest GTL process in the world.
11.2 Synthesis Gas Production in Bintulu GTL
Synthesis gas production at the Bintulu GTL facility is accomplished by the reforming of natural gas (Figure 11.1). This is the most expensive part of the SMDS process, constituting 50–60% of the total capital cost [6]. The natural gas feed rate is around 150 000 m3 ·h−1 (normal) [7]. The natural gas is desulfurized by a combination of hydrogenation and sulfur absorption (Figure 3.1). During hydrogenation, the low levels of organic sulfur compounds are converted into their equivalent alkanes and H2 S. The H2 S is then removed from the natural gas by absorption on ZnO [2]. Most of the synthesis gas is produced using the Shell gasification process (SGP). The SGP was developed in the 1950s using heavy oil as feed material. There are many units in industrial operation, using a variety of raw materials as feed. The SGP technology is based on noncatalytic thermal reforming of natural gas at >1300 ◦ C, and operation at pressures up to 6.5 MPa is possible. The raw synthesis gas produced from natural gas typically contains about 1% CH4 and 2% CO2 , with 1% inert gases (N2 and Ar) [3, 8]. The H2 :CO ratio is 1.7 : 1 [6, 8]. Waste heat recovery forms part of the process. The primary syngas cooler generates steam (up to 12 MPa), which cools the syngas down to about 340 ◦ C [8]. After heat recovery, the cooled synthesis gas is cleaned by water scrubbing and passed over guard beds to remove trace components that may affect Fischer–Tropsch synthesis [6]. The SGP is claimed to be a robust technology with a unit availability better than 99% [3]. At Bintulu high availability was achieved through spare gasifier capacity.
11.3 Fischer–Tropsch Synthesis in Bintulu GTL
Air separation
Natural gas
O2
Partial oxidation reformer (SGP)
Water gas shift
H2 :CO = 1.7
Sulfur removal
H2 :CO > 3
Steam reformer
PSA
H2
Syngas
C1 – C4
Fischer– Tropsch synthesis
Syncrude C5 +
Figure 11.1 Synthesis gas generation from natural gas by partial oxidative reforming in the Shell gasification process (SGP) and steam reforming to increased the H2 :CO ratio.
The H2 :CO ratio obtained by the SGP is lower than the Fischer–Tropsch usage ratio of 2.15 : 1 [9]. The hydrogen deficiency is made up by converting some of the natural gas in a steam reformer to a more hydrogen-rich syngas. The steam reformer operates at around 850 ◦ C and converts the methane over a nickel catalyst into a syngas with a H2 :CO ratio >3 : 1 [10]. Part of the synthesis gas production from the steam reformer is used for natural gas desulfurization, as well as to produce H2 for the refinery by pressure swing absorption.
11.3 Fischer–Tropsch Synthesis in Bintulu GTL
The approach taken by Shell in the development of the Fischer–Tropsch technology for SMDS was quite different to the industrial practice at that time. It was a move away from iron-based high-temperature Fischer–Tropsch (Fe-HTFT) synthesis to cobalt-based low-temperature Fischer–Tropsch (Co-LTFT) synthesis. Implicit in this decision was, as the name ‘‘Shell Middle Distillate Synthesis’’ suggests, the production of paraffinic distillate range products. Fischer–Tropsch catalyst development focused not only on activity but also on producing a catalyst with an α-value of 0.90 or higher [11]. Data from the literature suggested that on average it was easier to obtain a high α-value with a Co-based catalyst than with an Fe-based catalyst [1, 10]. It was also believed that the catalyst lifetime that could be achieved with Co-based catalysts would be longer than that achieved with Fe-based catalysts. For this purpose, a Co/Zr/SiO2 catalyst was developed [12]. The development of the new Fischer–Tropsch catalyst was also intrinsically linked to the preselection of the reactor type. Details of the reactor selection and the fundamental work that were performed in this regard have been published by Sie and coworkers [1, 11, 13]. The main
233
11 Shell Middle Distillate Synthesis (SMDS) Facilities
factor that resulted in the selection of a multitubular fixed bed reactor was one of timing. The scale-up of multitubular fixed bed reactors is comparatively simple, since each tube is essentially a complete reactor. Piloting performed on a single reactor tube can be representative of commercial operation. The proven stability of the Arge multitubular fixed bed reactors (Section 8.3.2) confirmed that these reactors were stable and easy to operate. The selection of a multitubular fixed bed reactor design therefore had a low risk factor [11]. Other beneficial properties of multitubular fixed bed reactors that were cited were the following: 1) High volumetric utilization of reactor space by the Fischer–Tropsch catalyst. A fixed bed catalyst can occupy around 60% of the reactor volume. 2) The ability to have a catalyst gradient along the reactor. 3) No need for solid–liquid separation. 4) Minimal catalyst attrition. One of the main drawbacks of fixed bed reactors is the inability to add and remove a catalyst during operation. By developing a stable Co-LTFT catalyst, catalyst lifetime was not a concern. The Shell Co-based Fischer–Tropsch catalyst required rejuvination every 9–12 months [3]. Rejuvenation takes places in situ and the catalyst has an overall useful lifetime of around five years [6]. The Shell Fischer–Tropsch synthesis process is called heavy paraffin synthesis (HPS) in the literature. An overall synthesis gas conversion of up to 96% has been reported, with a C5 and heavier product selectivity in the range 90–95% [6, 9]. The original Co-LTFT catalyst had an α-value of close to 0.9 based on its published product distribution (Figure 11.2) [10]. The catalyst age at which this distribution was obtained has not been reported. A newer generation Co-LTFT catalyst has since been developed and has been applied in the Bintulu facility on a retrofit basis since 2001 [3]. The Bintulu facility has four fixed bed multitubular reactors. Each reactor shell contains around 26 000 tubes [3] with a rated capacity of around 3000 bbl/day each [6]. Typical operating conditions 0
6 5
−1
4 a = 0.89
3
−2
2 1
−3
0 0
5
10
15 20 25 Carbon number
30
Figure 11.2 Carbon number distribution of the C5 and heavier products () obtained with the original Co-LTFT catalyst for SMDS, as well as the corresponding Anderson–Schulz–Flory (ASF) plot (•).
35
40
ASF plot, log (xn)
Syncrude composition (mass%)
234
11.4 Fischer–Tropsch Refining in Bintulu GTL
are 220 ◦ C and 2.5 MPa [12]. The product gas from Fischer–Tropsch synthesis is cooled down to recover the C5 and heavier oil fraction. The lighter products (C1 –C4 ) are used as fuel gas. The C5 and heavier oil fraction and wax are refined to final products (Section 11.4.1). The aqueous product is treated as wastewater (Section 11.4.2). The combination of the very hydrogenating, high α-value Co-LTFT catalyst in a fixed bed reactor technology that promotes hydrogenation resulted in the production of a syncrude that is very paraffinic. The alkenes and oxygenate (mainly alcohol) content of the syncrude is the lowest for any large-scale, industrially applied Fischer–Tropsch technology. This matched the design intent of the SMDS process.
11.4 Fischer–Tropsch Refining in Bintulu GTL 11.4.1 Oil Refining
The refinery section employed to upgrade the products from LTFT synthesis is remarkably simple (Figure 11.3). It contains only two processing units, namely, a hydrocracker and a hydrotreater. This purely hydroprocessing approach fits well with the syncrude that is very paraffinic in nature. Hydrocracking has four refining objectives: alkene hydrogenation, hydrodeoxygenation (HDO), hydrocracking, and hydroisomerization [1]. The hydrocracker is operated in the range 300–350 ◦ C and 3–5 MPa with a noble metal hydrocracking catalyst developed by Shell. Noble metal catalysts benefit from sulfur-free operation. The process has a low gas make, which reportedly is less than Fuel gas
Co-LTFT synthesis
(to steam reformer)
Hydrocracker Syngas
C1 – C4
Naphtha Kerosene Distillate
>C5
Wax Waxy raffinate
Waxy raffinate
Hydrotreater C5 –C10 C10 –C13 C14 –C17 >C17
Aqueous product
Wax production
Waxes Wastewater
Figure 11.3
Fischer–Tropsch refinery at the SMDS Bintulu facility.
235
236
11 Shell Middle Distillate Synthesis (SMDS) Facilities Table 11.1
Properties of the hydrocracked products produced at the SMDS Bintulu facility.
Property
Naphtha
Kerosene
Gas oil
Density at 15 ◦ C (kg·m−3 ) Boiling range (◦ C) Cetane index Viscosity at 40 ◦ C (cSt) Smoke point (mm) Flash point (◦ C) Freezing point (◦ C) Composition Aromatics (vol%) Sulfur (µg·g−1 )
690 43–166 – – – – –
738 155–191 58 – >50 42 −47
750 150–250 – – >100 – −47
776 184–357 76a 2.7 – – –
780 201–358 75 – – 88 – b,c
0 <3
<0.1 <10
– –
<0.05 <2
<0.1 <15
a
Cetane number is 81. pour point of 250–360 ◦ C material is −10 ◦ C. c The cloud point is 1 ◦ C and the cold filter plugging is −2 ◦ C. b The
2%. The hydrocracking catalyst has demonstrated good stability with Fischer–Tropsch syncrude, and the catalyst in the Bintulu hydrocracker has operated for more than eight years without the need for regeneration [3]. Depending on the hydrocracking severity, the product composition can be varied from 15% lights and naphtha, 25% kerosene, and 60% gas oil in the gas oil mode to 25% lights and naphtha, 50% kerosene, and 25% gas oil in kerosene mode. Some product properties are given in Table 11.1 [1, 6, 7, 14, 15]. The fuel-related products are only blending materials and not final fuels. This is in line with the design intent of the SMDS process (Section 11.1). Depending on the severity of the hydrocracker, the heavy naphtha/kerosene is also employed as paraffinic chemical (Table 11.2) when it is not suitable as kerosene for blending into jet fuel. The light gases were combined with those from Fischer–Tropsch synthesis and recycled to the steam reformer. The >360 ◦ C boiling material from the hydrocracker could be recycled to extinction, or recovered as a final product called waxy raffinate. The waxy raffinate contains 18–26% wax, which can be removed by solvent dewaxing in order to produce lubricant base oils [16]. The material has a high viscosity index and low Noack volatility. Solvent dewaxing is not performed in the Bintulu facility, but the material is exported to Shell refineries in France and Japan for lubricant base-oil production (Table 11.3) [17]. The hydrotreater employs a nickel-based catalyst [3]. Its main functions are alkene hydrogenation and HDO. The hydrotreater is employed to produce chemicals and causes no isomerization in the product. Waxes and paraffinic products (n-alkane-rich fractions) are the main chemicals that are produced. The lighter hydrocarbons are fractionated into C5 –C10 for solvents, C10 –C13 for detergents, and C14 –C17 for flame-retardant materials (Table 11.2) [16]. These are all sharp cuts, with little material outside of the carbon number range. Further refining steps to produce these end-product applications do not form part of the Bintulu oil refinery. The heavier than C17 fraction is separated into four wax fractions by vacuum distillation (Table 11.4) [16].
11.4 Fischer–Tropsch Refining in Bintulu GTL Properties of the paraffinic chemicals produced at the SMDS Bintulu facility by hydroprocessing Fischer–Tropsch syncrude.
Table 11.2
Property
Paraffinic product
Density at 15 ◦ C (kg·m−3 ) Boiling range (◦ C) Viscosity at 25 ◦ C (cSt) Flash point (◦ C) Pour point (◦ C) Composition n-Alkanes (mass%) Alkenes (g Br/100 g) Sulfur (µg·g−1 ) a Heavy
C5 –C10
C10 –C13
C14 –C17
C8 –C12 a
690 35–160 0.6 – –
750 190–230 1.7 75 −20
775 250–280 3.3 110 5
735 155–195 1.5 43 −35
91 0.01 0
96 0.005 0
95 0.0055 0
50 0.02 0
naphtha/kerosene from hydrocracker operated at low severity; low iso:n ratio.
Table 11.3
Lubricant base oils produced by chemical modification of waxes produced in the SMDS process.
Property
Wax-derived chemically modified mineral oil
Density at 15 ◦ C (kg·m−3 ) Kinematic viscosity at 40 ◦ C (cSt) Kinematic viscosity at 100 ◦ C (cSt) Viscosity index Pour point (◦ C) Flash point (◦ C) Noack volatility (mass%)
Table 11.4
GTL-2
GTL-3
GTL-5
GTL-7
795 5.0 1.7 – −59 159 –
805 10.4 2.8 122 −45 207 28.8
818 20.1 4.5 144 −21 238 7.8
820 38.4 7.0 147 −21 260 2.0
Properties of the waxes produced as chemicals at the SMDS Bintulu facility by hydrogenation.
Property
Wax product
Congealing point (◦ C) Oil content (mass%)a a Determined
SX30
SX50
SX70
SX100
31 5
50 2.5
70 0.4
98 0.1
by ASTM D721, butanone (MEK) solubility at −32 ◦ C.
237
238
11 Shell Middle Distillate Synthesis (SMDS) Facilities
The heaviest wax grades are fractionated in wiped film evaporators, also known as short path distillation columns, which limit the residence time of the wax in the column. Distillation takes place at very low pressures (6–12 Pa) and allows the temperature to be kept below 280 ◦ C [2]. By doing so, the risk of thermal degradation is minimized. The heavier wax grades thus produced are treated in a hydrofinishing step to ensure that the color and odor specifications are met. 11.4.2 Aqueous Product Treatment
The oxygenates that are dissolved in the aqueous product from Fischer–Tropsch synthesis constitute around 1% of the total syncrude. These oxygenates are not recovered, and the total Fischer–Tropsch aqueous product is considered wastewater. It is treated in a flotation/flocculation unit to remove suspended solids before it is sent to a biotreater. In the biotreater, the dissolved organics are consumed to a level that allows discharge of the wastewater as surface water [14]. Whereas the production of Fischer–Tropsch syncrude from coal is a net consumer of water, production of syncrude from natural gas is a net producer of water. The SMDS facility in Bintulu produces 1500–2000 m3 /day of water. It was pointed out that this is a potential advantage in arid climates [2, 14].
11.5 Pearl GTL Facility
The feed material to the Pearl GTL facility is 2.4 million m3 ·h−1 (normal) natural gas and 120 000 bbl/day crude oil equivalent of associated NGL, liquid petroleum gas, and ethane [18]. The feed makeup is analogous to that used for the Mossgas facility (Chapter 10), and the SMDS process will be preceded by an NGL recovery section. The associated liquids will provide analogous synergies and integration opportunities with the Fischer–Tropsch syncrude as noted before, except that in this case the integration will be with LTFT syncrude. The Co-LTFT syncrude and NGL are quite similar in that both materials consist mainly of n-alkanes. The synthesis gas production and Fischer–Tropsch synthesis sections are scaled-up versions of the SMDS process at Bintulu. In the Pearl GTL facility, a very modular approach has been followed. Each of the two 70 000 bbl/day GTL trains is further divided into four modular sections per train. The following scale-up factors relative to Bintulu GTL have been applied for Pearl GTL [19]: 1) Air separation units, 1.13 scale-up and only one unit per train. 2) Partial oxidative reformers (SGP), 3.5 scale-up. 3) Multitubular fixed bed Fischer–Tropsch synthesis, 1.12 scale-up and six reactors per train. The refinery will include hydrocracking and fractionation facilities as well as the hydrogenation and purification sections to produce chemical products [18]. Despite the dramatic increase in the overall cost estimate of the project (Section 1.5.3), Pearl GTL is expected to be a very profitable venture [19].
References
11.6 Discussion of the Refinery Design
The SMDS process signified an important departure in Fischer–Tropsch refinery design from that of a stand-alone refinery to produce final products to that of a partial refinery to produce final products and blending materials. This was the original SMDS design intent. The Shell group has extensive conventional crude oil refining infrastructure and a global distribution network for fuels and chemicals. Converting the gas into transportable products made sense in this context. The refinery design of the SMDS process is compact, which is in line with the potentially remote location of the facility, and with the objective to produce mainly distillate as the name of the process suggests. Other points that can be highlighted are the following: 1) The refinery relies on prefractionation of the syncrude in the gas loop by stepwise condensation. Distillation takes place only after hydroprocessing (hydrocracking or hydrotreating), thereby avoiding boiling point broadening. 2) Clean product cuts between carbon numbers are made when producing chemicals. Care is also taken with the separation to avoid any thermal degradation of the products. 3) Blending materials intended for fuels and chemicals are produced. There is a clear refining pathway for all products, even though the intermediate products are not refined on site. 4) The refinery design is flexible and the ratio between fuel intermediates and chemicals can be varied considerably. The main constraint is unit capacity, and the refinery is otherwise free of constraints related to fuels blending. 5) Catalyst selection and the design of the refining units take the properties of the syncrude into consideration. Hydroprocessing is performed with unsulfided noble- and base metal catalysts, which does not require the addition of a sulfiding agent to the sulfur-free syncrude. The units are operated at conditions in line with the nature of the feed, and operation is generally milder than in crude oil refining. 6) No specific provision is made to deal with alkenes and oxygenates, which are converted into alkanes by hydroprocessing. It should be added that Co-LTFT synthesis in a fixed bed reactor as applied in the SMDS process is very hydrogenating and the syncrude is consequently very paraffinic in nature. 7) The refinery does not upgrade the C1 –C4 hydrocarbons, which are recycled to the steam reformer. The alkene content of this fraction is low, and there is little benefit in employing alkene oligomerization or any other alkene conversion technology to upgrade this material into liquid products. 8) Because of the hydrogenating nature of Co-LTFT synthesis, a small amount water-soluble oxygenates is produced. The aqueous product is treated as wastewater rather then being refined. References 1. Sie, S.T., Senden, M.M.G., and Van Wechem,
H.M.W. (1991) Conversion of natural gas to transportation fuels via the Shell Middle Distillate Synthesis process (SMDS). Catal. Today, 8, 371–394.
2. (1992) The Shell Middle Distillate Synthesis
Plant at Bintulu, Shell MDS (Malaysia) Sdn Bhd brochure. 3. Overtoom, R., Fabricius, N., and Leenhouts, W. (2009) in Proceedings of the 1st Annual Gas
239
240
11 Shell Middle Distillate Synthesis (SMDS) Facilities
4. 5. 6.
7.
8.
9.
10.
11.
12.
Processing Symposium (eds H.E. Alfadala, G.V.R. Reklaitis, and M.M. El-Halwagi.), Elsevier, Amsterdam, pp. 378–386. Brown, A. (2006) Towards sustainability. Hydrocarbon Eng., 11 (10), 22–24. (2007) Pearl GTL plant foundation stone laid. Oil Gas J., 105 (10), 9. Schrauwen, F.J.M. (2004) in Handbook of Petroleum Refining Processes (ed. R.A. Meyers), 3rd edn, McGraw-Hill, New York, pp. 15.25–15.40. Senden, M.M.G., Punt, A.D., and Hoek, A. (1998) Gas-to-liquids processes: current status and future prospects. Stud. Surf. Sci. Catal., 119, 961–966. Higman, C. and Van der Burgt, M. (2008) Gasification, 2nd edn, Elsevier, Amsterdam, pp. 149–155. Smith, R. and Asaro, M. (2005) Fuels of the Future. Technology Intelligence for Gas to Liquids Strategies, Stanford Research Institute, Menlo Park, CA. Eilers, J., Posthuma, S.A., and Sie, S.T. (1990) The Shell middle distillate synthesis process (SMDS). Catal. Lett., 7, 253–269. Sie, S.T. (1998) Process development and scale-up. IV. Case history of the development of a Fischer-Tropsch synthesis process. Rev. Chem. Eng., 14 (2), 109–157. Moodley, D.J., Van de Loosdrecht, J., Saib, A.M., and Niemantsverdriet, J.W. (2009) in Advances
13.
14.
15.
16.
17.
18.
19.
in Fischer– Tropsch Synthesis, Catalysts and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 49–81. Sie, S.T. and Krishna, R. (1999) Fundamentals and selection of advanced Fischer-Tropsch reactors. Appl. Catal. A, 186, 55–70. Tijm, P.J.A., Van Wechem, H.M.H., and Senden, M.M.G. (1993) The Shell Middle Distillate Synthesis project new opportunities for marketing natural gas. Alternate Energy ’93 Conference, April 27–30, 1993, Colorado Springs. Tijm, P.J.A. (1994) Shell middle distillate synthesis: the process, the plant, the products. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 39 (4), 1146–1150. Ansorge, J. (1997) Shell Middle Distillate Synthesis: Fischer-Tropsch catalysis in natural gas conversion to high quality products. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 42 (2), 654–663. Henderson, H.E. (2006) Gas to liquids, in Synthetics, Mineral Oils, and Bio-Based Lubricants. Chemistry and Technology, Chapter 19 (ed. L.R. Rudnick), Taylor & Francis, Boca Raton. Fabricius, N. (2005) Fundamentals of Gas to Liquids, 2nd edn, Petroleum Economist, London, pp. 12–14. Forbes, A. (2008) Pearl GTL – a gem of a project, so long as it works. Pet. Econ., 75 (4), 8–11.
241
12 Oryx and Escravos Gas-to-Liquids Facilities 12.1 Introduction
The Oryx and Escravos gas-to-liquids (GTL) facilities were partly modeled on the Shell Middle Distillate Synthesis (SMDS) process (Chapter 11) and design philosophy. These facilities do not produce final products, and the associated refinery is even simpler. The original intent was to produce mainly distillates but without the coproduction of chemicals, although subsequent projects were envisioned to expand production in the direction of lubricating base oils [1, 2]. The technologies that make up the GTL design come from three different companies and are the same for both Oryx GTL and Escravos GTL processes [3, 4]: 1) Natural gas reforming, licensed by Haldor Topsøe. 2) Fischer–Tropsch synthesis, licensed by Sasol Technology. 3) Syncrude refining, licensed by ChevronTexaco. The Oryx GTL venture started in July 1997, when a memorandum of understanding was signed between Qatar Petroleum, Sasol, and Philips Petroleum. The agreement called for the construction of a 20 000 bbl/day Fischer–Tropsch-based GTL plant in Ras Laffan, Qatar. Philips Petroleum withdrew from the project after the collapse of oil prices in 1998. By mid-2001, a new joint venture agreement was reached, with Qatar Petroleum having a 51% stake and Sasol a 49% stake [5]. Natural gas from the Al Khaleej field was earmarked as feed. The natural gas supply infrastructure was already developed by a joint venture between Qatar Petroleum and ExxonMobil [6]. This allowed front-end engineering and design of a 34 000 bbl/day facility to proceed, which was called Oryx GTL. The product slate that was anticipated is given in Table 12.1 [7]. Oryx GTL achieved mechanical completion in 2006 and was officially opened in June 2006. The commissioning phase started in mid-2006 and the production of the first products was announced in February 2007 [8]. In mid-2007, it was reported that commissioning problems in the Fischer–Tropsch synthesis section constrained output to 7000−10 000 bbl/day and that additional downstream equipment was required to overcome the problem [9]. By early 2008, production reached 20 000 bbl/day, and by August 2008 the average production was around 24 000 bbl/day [10, 11]. The Escravos GTL facility initially was a joint venture between the Nigerian National Petroleum Company, with a 25% stake, and Chevron Nigeria, with a 75% stake [12]. In a subsequent Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
242
12 Oryx and Escravos Gas-to-Liquids Facilities Table 12.1
Design product slate for the 34 000 bbl/day Oryx
GTL facility. Product
Production (bbl/day)
Liquid petroleum gas Naphtha Distillate
1 000 7 000–9 000 24 000–26 000
agreement between Chevron and Sasol, the latter acquired a 37.5% stake but subsequently reduced it to 10% after the problems with Oryx GTL and the escalating cost of the project [11]. The engineering, procurement, and construction contract was awarded in April 2005, and from the outset the project faced problems with timing and the location. The facility has a design capacity of 34 000 bbl/day and it is a clone of Oryx GTL. Civil unrest and necessary front-end engineering design changes were cited as the main causes for the delay in targeted completion from 2009 to 2011 [11]. Since Oryx GTL and Escravos GTL are similar in basic design, the subsequent discussion focuses only on the former.
12.2 Synthesis Gas Production in Oryx GTL
The facility has been designed to process around 390 000 m3 ·h−1 (normal) of lean natural gas [8]. Synthesis gas production employs the same conversion steps as the SMDS process (Section 11.2) but uses different technology (Figure 12.1). The natural gas feed is conditioned by removing sulfur over a ZnO guard bed and removing the coke precursors with a copper catalyst. The cleaned gas is preheated in a prereformer by heat Methane reforming Autothermal reforming
PSA Steam
Fuel gas
Steam O2
LTFT synthesis
Tail gas processing Tail gas
70 °C
Natural gas
Fuel gas
External recycle Internal recycle
Pre reformer
H2
Heavy end recovery
Figure 12.1
CO2 stripper Cold condensate
Wax Water
Stack gas
Hot condensate
70 °C
Gas conditioning
CO, CO 2
Reaction water
Aqueous product Wax
Synthesis gas generation and Fischer–Tropsch gas loop in the Oryx GTL facility.
12.3 Fischer–Tropsch Synthesis in Oryx GTL
exchange with the hot product gas from the autothermal reformer (ATR). In the prereformer, the natural gas is reformed with steam to convert the C2 and heavier hydrocarbons in the natural gas to syngas (Section 3.4.2). This reduces the risk of carbon formation in the ATR. Prereforming is followed by adiabatic oxidative reforming in the ATR with pure O2 and steam. The O2 is supplied from an air separation unit. The reformers and air separation units constituted 30% of the capital cost of the project [13]. A separate steam reformer is used to convert some of the natural gas into a hydrogen-rich synthesis gas. The hydrogen is recovered in a pressure swing absorption (PSA) unit for use in the refinery, while the hydrogen lean product is sent to the fuel gas system. The steam reformer produces a more H2 -rich syngas than the ATR and is used to adjust the H2 :CO ratio of the synthesis gas. The syngas is cooled to about 70 ◦ C to knock out the water and water-soluble gases in a water-wash column before it is used as feed for Fischer–Tropsch synthesis.
12.3 Fischer–Tropsch Synthesis in Oryx GTL
The Fischer–Tropsch synthesis section makes use of the Sasol slurry-phase distillate (SPD) process, which is a new technology adapted from the Fe-LTFT slurry bed process that was commercialized in the Sasol 1 facility (Section 8.5.2). The same basic reactor technology is employed, but the SPD process uses a newly developed a cobalt-based low-temperature Fischer–Tropsch (Co-LTFT) catalyst [5]. Typical operating conditions are around 230 ◦ C and 2.5 MPa. The Co/Pt/Al2 O3 LTFT catalyst is manufactured in a new catalyst preparation plant at De Meern, The Netherlands. The catalyst manufacturing facility is operated by a Sasol–Engelhard (now BASF) joint venture. The Co/Pt/Al2 O3 catalyst is claimed to be more resistant to reoxidation by water, allowing it to be more productive under high per pass conversion operation [14]. Catalyst stability was evaluated under industrially relevant conditions during reaction in a 100 bbl/day slurry bed unit. The unit was operated with a H2 :CO ratio of 2 : 1 at 230 ◦ C, 2 MPa, and a H2 + CO conversion of 50–70%. Activity of the Co/Pt/Al2 O3 catalyst was halved after 50 days on stream [15]. Oryx GTL uses two slurry bed reactors, each weighing 2100 tons. The reactors are about 60 m in height and almost 10 m in diameter [13, 16]. The reactors were prefabricated and shipped because of the difficulty of on-site assembly [5]. Installation also required specialized lifting equipment. Construction and installation of this type of Fischer–Tropsch technology can consequently easily become critical path items during project execution. The Fischer–Tropsch synthesis section represents about 15% of the total capital cost of the facility [13]. The selection of a slurry bed reactor in combination with a Co-LTFT catalyst is somewhat surprising. One of the main advantages of a slurry bed reactor over a fixed bed reactor is the ability to add and remove catalyst while the reaction is in operation. Yet, one of the advantages of Co-LTFT synthesis is its long catalyst lifetime. On-line catalyst addition and removal is therefore not necessary. Although a slurry bed reactor has good heat transfer properties, it requires liquid–solid separation and a catalyst with good attrition resistance. The problems that were encountered during start-up of the Oryx GTL Fischer–Tropsch section was related to attrition of the Co-LTFT catalyst. A fine sediment was formed due to catalyst attrition. This caused clogging of downstream equipment and reduced throughput to a fraction
243
244
12 Oryx and Escravos Gas-to-Liquids Facilities
of the design capacity [9, 10]. These problems also resulted in an increased metal content in the syncrude, which gave rise to a spate of patenting activity in the field of metal removal from Fischer–Tropsch wax to enable refining [17]. The deactivation mechanisms have been described earlier (Sections 4.5.3 and 4.5.4). The syncrude composition from Co-LTFT synthesis varies with the age of the catalyst, and a typical composition can be found in Table 1.2. The product from LTFT synthesis is filtered in the reactor to separate the catalyst from the hydrocarbon product. The catalyst remains in the reactor and the hot wax goes through a secondary filtration step before being sent to the refinery. It is at this point in the process where new technology had to be developed to reduce the metals that remained in the wax. The wax-free gaseous products are cooled down to about 70 ◦ C to condense hydrocarbons and water (Figure 12.1). The hydrocarbon fraction (hot condensate) is liquid–liquid phase separated from the aqueous product (reaction water). The aqueous product contains some dissolved oxygenates, such as methanol. The gaseous product (tail gas) that is not used for an internal recycle in the Fischer–Tropsch gas loop is chilled to condense the C3 and heavier hydrocarbons as well as some water that remained in the tail gas. The liquid product from this separation is called the heavy end recovery stream. The heavy end recovery stream and the hot condensate are passed through a CO2 stripper column, where the dissolved CO and CO2 are removed before the oil is sent to the refinery as a cold condensate. Uncondensed tail gas contains mainly C1 –C2 hydrocarbons, H2 , CO, and CO2 . Part of this product is recycled directly to the ATR as an external recycle, with the rest being purged for use as fuel gas. Some possibilities for process intensification based on reactor hydrodynamics have been noted [18]. The estimated capacity increase is 25–30%. In order to realize these benefits in practice, the cooling capacity of the slurry bed reactor will also have to be increased, and methods to accomplish this were suggested. Considering the complexity of the reactor construction, it is not clear whether it will be possible to realize these benefits with the installed equipment at the Oryx GTL facility.
12.4 Fischer–Tropsch Refining in Oryx GTL 12.4.1 Oil Refining
The refinery design of Oryx GTL is very simple and consists of a single conversion unit, a hydrocracker (Figure 12.2). It contributes only 10% to the overall capital cost of the facility [13]. It is only a partial refinery and it does not produce transportation fuels or chemicals, only intermediate products and LPG. The refinery section receives two feed streams from the Fischer–Tropsch gas loop, namely wax and cold condensate. The wax and the cold condensate are combined to serve as feed to the hydrocracker. The wax hydrocracker uses the ChevronTexaco Isocracking technology and catalyst. The hydrocracking catalyst is commercially available and it consists of sulfided base metals on an acidic support. Since the Fischer–Tropsch syncrude is sulfur-free, operation of this unit
12.4 Fischer–Tropsch Refining in Oryx GTL
Sulfiding agent
Hydrocracker
H2
Main fractionator
Purge
Wax
Stabilizer column
LPG
Cold HP separator
Cold condensate
Naphtha Distillate
Hot HP separator >360 °C
Bottom recycle
Figure 12.2
Refinery section in the Oryx GTL facility.
requires that a sulfiding agent must be co-fed in order to keep the catalyst in a sulfided state. Typical operating conditions are a liquid hourly space velocity (LHSV) of 1.2 h−1 , 350 ◦ C, and 7 MPa, with the temperature being adjusted to keep the per pass conversion at around 65%. The product from hydrocracking is distilled to produce LPG (3–7%), naphtha (20–30%), and distillate (65–75%), with the unconverted >360 ◦ C waxy product being recycled to the hydrocracker feed [7]. The straight-run naphtha has a research octane number of 50–55 and has a high content of n-alkanes (Table 12.2) [7]. Since the naphtha fraction forms part of the cold condensate, the
Table 12.2
Characteristics of different naphtha fractions produced from Co-LTFT synthesis and
refining. Property Density at 20 ◦ C (kg·m –3 ) Cetane number Cloud point (◦ C)b Flash point (◦ C) Distillation (◦ C) IBP (initial boiling point) T10 T50 T90 FBP (final boiling point) Composition (mass%)c n-Alkanes Branched alkanes Alkenes Aromatics Oxygenates (alcohols) Sulfur a The
Straight run
Hydrotreateda
710 – –51 –9
683 42.7 –54 –18
688 30 –35 –21
58 94 118 141 159
60 83 101 120 133
49 79 101 120 131
53.2 1.2 35 0 10.7 <0.0001
90.1 8.3 1.5 0.1 0 <0.0001
Oryx GTL refinery does not include a hydrotreater. point predicted from differential scanning calorimeter analysis. c Source reference values do not all add to 100%. b Cloud
Hydrocracked
28.6 66.7 0 0.5 0 <0.0001
245
246
12 Oryx and Escravos Gas-to-Liquids Facilities
hydrocracker acts as hydrotreater, hydroisomerization unit, and hydrocracker for this fraction. Hydrocracking of the distillate and wax fractions further increases the volume of naphtha. The properties of the final product are very dependent on the operation of the hydrocracker and the properties of the hydrocracking catalyst. Since all of the naphtha passes through the hydrocracker, the product is paraffinic and the n-alkane content is reduced by hydroisomerization. It was reported that the naphtha makes a good steam-cracking feedstock [19], and the intention was to market it as such. The distillate has almost no sulfur, and the degree of branching is determined by the hydrocracker. The hydrocracked distillate has a high cetane number, but lacks the density to meet EN590:2004 diesel fuel specifications (Table 12.3) [20–22]. The distillate can be fractionated into kerosene and gas oil, but this does not form part of the original Oryx GTL refinery design. The distillate that is produced is a diesel fuel blending stock.
Table 12.3 Properties of different hydrocracked distillate range products from pilot plant operation and the industrial Oryx GTL facility.
Property
Sasol SPD pilot plant Kerosene
Density at 15 ◦ C (kg·m−3 ) Cetane number Viscosity at 40 ◦ C (cSt) Cold filter plugging point (◦ C) Freezing point (◦ C) Smoke point (mm) Flash point (◦ C) Lubricity, HFRRc (µm) Net heating value (MJ kg –1 ) Distillation (◦ C) IBP (initial boiling point) T10 T50 T90 FBP (final boiling point) Composition Alkenes (g Br/100 g) Aromatics (mass%) Sulfur (mass%) Acidity (mg KOH/g)
Distillate
769 72–73 2.0 –19 –15 – 58 617 43.8
771a 87 2.4 –7 – – 65 – –
154 168 191 254 267
151 182 249 325e 334
168 206 268 345 364
0 0 <0.01 0.009
0.6 0.1 <0.0001 0.001
– <1 <0.0005 –
747 – 4.2b – –48 >50 45 500d 44.1
at 20 ◦ C. at −20 ◦ C. c HFRR, high-frequency reciprocating rig. d Lubricity by ball-on-cylinder lubricity evaluator (BOCLE), ASTM D5001. e T95 distillation point. a Density
b Viscosity
Oryx GTL
Distillate
12.5 Discussion of the Refinery Design Table 12.4
Fischer–Tropsch aqueous product from Co-LTFT synthesis.
Compound Water Nonacid oxygenates Carboxylic acids Hydrocarbons Inorganic matter
Aqueous product (mass%) 98.89 1 0.09 0.02 <0.005
12.4.2 Aqueous Product Treatment
Approximately 1.3 tons of water is produced for every 1 ton of hydrocarbon. The Fischer–Tropsch aqueous product is consequently a major product stream. It contains some dissolved oxygenates, which are mainly light alcohols (Table 12.4) [23]. The reaction water is separated by distillation into an alcohol-rich overhead product that is incinerated and a carboxylic acid–containing water-rich product that is biologically degraded to purify the water. No oxygenates are recovered from the reaction water. However, considering the separation step that is already present in the facility, such recovery is readily achievable.
12.5 Discussion of the Refinery Design
Superficially, the Oryx GTL refinery employed an analogous hydrocracker-based refinery design as the SMDS process but there are notable differences between the two refinery designs. The syncrude to the Oryx GTL refinery contains more alkenes and oxygenates than that to the SMDS refinery. Although this provides the syncrude with more synthetic capability, it was not exploited by the refinery design. In fact, the Oryx GTL refinery design did not exploit the properties of the syncrude in any way, whereas the refinery technology selection and design of the SMDS refinery capitalized on the properties of the syncrude. The capital cost contribution of the Oryx GTL refinery is therefore less than that in the SMDS process, but the refining advantages provided by the Co-LTFT syncrude were forfeited. The concept of partial refining was taken to the extreme by the Oryx GTL design and it approaches that of being just an upgrader. 1) Most of the alcohols in the aqueous product is recovered from the water effluent, but not refined to final products. 2) With the refinery shown in Figure 12.2, it is not possible to produce higher value chemical products. The refinery lacks a separate hydrotreater and the supporting separation infrastructure to produce chemicals such as waxes and paraffinic solvents. 3) The decision to produce only intermediate fuel-based products is surprising considering the trend in all other industrial Fischer–Tropsch facilities to coproduce chemicals. Refining is
247
248
12 Oryx and Escravos Gas-to-Liquids Facilities
the step in any Fischer–Tropsch-based facility where most of the value addition takes place with the least amount of capital. The Sasol 1 facility and the SMDS Bintulu facility are much smaller than Oryx GTL, but include more complex refineries, which supports the contention that there is economic merit in more extensive syncrude refining. 4) The hydrocracker design employs a sulfided base metal hydrocracking catalyst developed for crude oil hydrocracking. This requires the addition of sulfur to an otherwise sulfur-free feed. It also requires hydrocracker operation under more severe conditions (>350 ◦ C and 7 MPa) than is possible with a noble metal hydrocracking catalyst (300–350 ◦ C and 3–5 MPa). A comparison between the two catalyst types clearly showed the benefit of noble metal hydrocracking in terms of product selectivity, especially at high per pass conversion [17]. References 1. Turner, C. (2005) Fundamentals of Gas to Liq-
2.
3. 4. 5.
6. 7.
8. 9. 10.
11. 12.
13. 14.
uids, 2nd edn, Petroleum Economist, London, pp. 5–7. Dancuart, L.P. and Steynberg, A.P. (2007) Fischer-Tropsch based GTL technology: a new process? Stud. Surf. Sci. Catal., 163, 379–399. Cook, P. (2003) Fundamentals of Gas to Liquids, 1st edn, Petroleum Economist, London, p. 46. Jay, M. (2003) Fundamentals of Gas to Liquids, 1st edn, Petroleum Economist, London, p. 45. Collings, J. (2002) Mind Over Matter. The Sasol Story: A Half-century of Technological Innovation, Sasol, Johannesburg. Daya, A. (2006) In vogue. Pet. Econ., 73 (4), 27–28. Dancuart, L.P., De Haan, R., and De Klerk, A. (2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532. (2007) Oryx plant produces GTL products for first time. Oil Gas J. 105 (5), 10. Forbes, A. (2007) Reality check. Pet. Econ., 74 (7), 30. (2008) GTL: Oryx breakthrough and oil-price surge lift industry spirits. Pet. Econ., 75 (6), 36–38. (2008) Sasol cuts stake in Escravos GTL as costs rise to $6bn. Pet. Econ., 75 (10), 30. Fraser, K. (2005) Fundamentals of Gas to Liquids, 2nd edn, Petroleum Economist, London, pp. 15–16. Halstead, K. (2008) Oryx GTL from conception to reality. Nitrogen+Syngas, 292, 43–50. Espinoza, R.L., Steynberg, A.P., Jager, B., and Vosloo, A.C. (1999) Low temperature Fischer-Tropsch synthesis from a Sasol perspective. Appl. Catal. A, 186, 13–26.
15. Saib, A.M., Borgna, A., Van de Loosdrecht, J.,
16. 17.
18.
19.
20.
21.
22.
23.
Van Berge, P.J., and Niemantsverdriet, J.W. (2006) XANES study of the susceptibility of nano-sized cobalt crystallites to oxidation during realistic Fischer-Tropsch synthesis. Appl. Catal. A, 312, 12–19. Forbes, A. (2007) Surge in interest a long time coming. Pet. Econ., 74 (1), 19–20. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer– Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Vogel, A., Steynberg, A.P., and Breman, B. (2007) Intensification of commercial slurry phase reactors. Stud. Surf. Sci. Catal., 167, 61–66. Dancuart, L.P., Mayer, J.F., Tallman, M.J., and Adams, J. (2003) Performance of the Sasol SPD naphtha as steam cracking feedstock. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 48 (2), 132–138. Lamprecht, D. and Roets, P.N.J. (2004) Sasol Slurry Phase Distillate semi-synthetic aviation turbine fuel. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 426–430. Lamprecht, D., Dancuart, L.P., and Harrilall, K. (2007) Performance synergies between low-temperature and high-temperature Fischer-Tropsch diesel blends. Energy Fuels, 21, 2846–2852. Kamara, B.I. and Coetzee, J. (2009) Overview of high-temperature Fischer-Tropsch gasoline and diesel quality. Energy Fuels, 23, 2242–2247. Dry, M.E. and Steynberg, A.P. (2004) Commercial FT process applications. Stud. Surf. Sci. Catal., 152, 406–481.
249
Part IV Synthetic Transportation Fuels
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
251
13 Motor-Gasoline 13.1 Introduction
Motor-gasoline (petrol) is the oldest transportation fuel and has been in use almost since the development of the spark-ignition engine, or Otto engine, in 1876. Many of the changes in the design of refineries (Chapter 2) were associated with changes in the requirements for motor-gasoline. In fact, motor-gasoline is the most demanding of the transportation fuels in terms of its refining requirements. As a final product, the properties of motor-gasoline are governed on a regional basis by fuel specifications despite the global nature of the automotive industry. One would have anticipated a global standard, but fuel specifications have technical, environmental, and political origins. Transportation fuels are major consumer products and they exert a tremendous influence on the economy and the environment. It is therefore not surprising to find that fuel specifications are on the political agenda. The refiner will always be at the mercy of politically driven changes in fuel specifications. Politically motivated changes are part of the cost of doing business in the transportation fuels sector. As such, these changes are unpredictable and cannot be planned for much beyond making a refinery flexible. One of the important driving forces behind changes in fuel specifications is an increasing awareness of the impact that transportation and, by implication, transportation fuels have on the environment. The fuel itself during storage and dispensing contribute some emissions to the environment, which should be limited. Ultimately, the transportation fuel will be combusted and its elemental composition is a good indicator of what products will be emitted as combustion products into the atmosphere. The engine technology is responsible for the efficient combustion of the fuel. Engine design therefore imposes some technical requirements on the fuel quality to enable clean combustion. The emission performance of the vehicle also depends on the postcombustion cleanup of the tailpipe emissions. The performance of these systems can likewise be influenced by the fuel and impose restrictions on the fuel composition. These changes in fuel specifications can usually be anticipated well before they are imbedded in legislation. The paper by Colucci [1], which details the history of changes in fuels specifications in the United States, is well worth the read. Fuel specifications with a technical origin are governed by the requirements of the engine technology. Many vehicle owners know little or nothing about motor-gasoline. It is assumed that the motor-gasoline they put in their car or truck will allow the engine to start and keep on Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
252
13 Motor-Gasoline
running despite the variable demands imposed on the engine by different driving conditions. Vehicle manufacturers want their engines to live up to these expectations. Engine performance is dependent on fuel quality, which in turn can be related to the properties and composition of the fuel. Fuel specifications that govern technical criteria change with engine technology but, due to their technical origin, these changes can be anticipated. The purpose of this chapter is to provide an overview of motor-gasoline and is followed by analogous chapters dealing with jet fuel (Chapter 14) and diesel fuel (Chapter 15). More detailed discussions on transportation fuels can be found in the literature [2, 3]. The purpose of each Chapters 13–15 is to give an account of the following: 1) Current fuel specifications and how these specifications relate to the molecular requirements of the fuel. 2) Direction of anticipated changes that will in future affect refinery design. 3) Demands placed on Fischer–Tropsch refineries by fuel property requirements.
13.2 Motor-Gasoline Specifications
Motor-gasoline specifications are location- and time dependent. It is therefore difficult to provide a comprehensive refining guideline that is universally valid. Rather than focus on an exhaustive set of region-specific fuel specifications, a more general discussion will be given. Two important guidelines are referred to in order to illustrate the general trends: 1) European Union (EU) fuel specifications (Table 13.1). These fuel specifications are legislated and determine the minimum fuel quality standards in member nations of the EU. It was developed to meet the technical engine requirements and to improve the air quality by setting emission standards. National specifications based on the EN228 specifications may allow other octane grades or require more stringent specifications. 2) World-Wide Fuel Charter (WWFC) guidelines (Tables 13.2 and 13.3) [4]. The WWFC is a product of the joint efforts of the European Automobile Manufacturers Association, Alliance of Automobile Manufacturers, Engine Manufacturers Association, and the Japan Automobile Manufacturers Association. It captures the fuel quality needs of vehicle manufacturers, but these guidelines are not legislated fuel specifications. Guidelines for motor-gasoline specifications in the United States are listed in the ASTM D 4814 standard test method [5]. An integral component of the final on-specification motor-gasoline is the fuel additive package employed [3]. Additives provide important fuel properties and address issues not always covered by the fuel specifications, such as detergency, corrosion prevention, and static charge dissipation. Additives can also be used to improve the fuel performance with respect to specifications that are regulated. For example, storage stability can be improved by antioxidant addition, and octane number can be improved by the addition of octane improvers (octane enhancers). The use of metal-based octane improvers such as tetraethyl lead (TEL) and methylcyclopentadienyl manganese tricarbonyl (MMT) reduces the refining effort required to meet fuel specifications. Although there are still some countries where these additives are allowed and
13.3 Motor-Gasoline Properties Selected motor-gasoline specifications (EN228) legislated in the European Union and the year in which each standard was adopted.
Table 13.1
Property
Euro-2 EN228:1993 (January 1994)
Euro-3 EN228:1999 (January 2000)
Euro-4 EN228:2004 (January 2005)
725–780 95 85 35–70 215 360 5
720–775 95 85 45–60 210 360 5
720–775 95 85 45–60 210 360 5
–b –b 5 500 0.013c –b
18 42 1 150 0.005c 2.7
18 35 1 50/10 0.005c 2.7
3 5 10 7 10 15 10
3 5 10 7 10 15 10
3 5 10 7 10 15 8
Density at 15 Ž C (kgÐm3 ) RON, min. MON, min. Vapor pressure at 37.8 Ž C (kPa)a Final boiling point (Ž C), max. Oxidation stability (min), min. Gum content, washed (mg/100 ml), max. Hydrocarbon content (vol%), max. Alkenes Aromatics Benzene Sulfur content (µgÐg1 ), max. Lead content (gÐl1 ), max. Oxygen (mass%), max. Oxygenate content (vol%), max. Methanol Ethanol 2-Propanol tert-Butanol Isobutanol C5 and heavier ethers Other ethers a Region
specific based on the climate and season. Not regulated by the fuel specification. c No intentional addition. b
employed, the intentional addition of metal-based compounds is not a sustainable practice in future. For the purpose of refinery design, additives will not be considered as a means of simplifying the design.
13.3 Motor-Gasoline Properties 13.3.1 Octane Number
In a spark-ignition engine, a spark initiates combustion. The spark is timed in such a way that the highest efficiency can be achieved. The combustion is thus controlled and it requires that the fuel does not autoignite during the process. When a fuel autoignites, the combustion rate
253
254
13 Motor-Gasoline Table 13.2 Motor-gasoline fuel categories defined by the WWFC to differentiate markets based on the stringency of their emission legislation.
Category 1
2 3 4
Table 13.3
Description Markets with no or minimal emission control requirements. Fuel specifications are based primarily on fundamental vehicle/engine performance criteria Markets with stringent emission control or other demands, for example, Euro-2 or US Tier 1 fuels Markets with advanced emission control or other demands, for example, Euro-3 and Euro-4 Markets with further advanced requirements for emission control to enable sophisticated after-treatment technologies to meet future needs, for example, Euro-4 and US EPA Tier 2 fuels
Selected motor-gasoline quality guidelines proposed by the WWFC.
Specification
Category 1
Category 2
Category 3
Category 4
Density at 15 Ž C (kgÐm3 ) RON, min. MON, min. Vapor pressure at 37.8 Ž C (kPa)a Final boiling point (Ž C), max. Oxidation stability (min), min. Gum content, washed (mg/100 ml), max. Hydrocarbon content (vol%), max. Alkenes Aromatics Benzene Sulfur content (µgÐg1 ), max. Lead content (gÐl1 ), max. Oxygen (mass%), max. Oxygenate content (vol%), max. Methanol Ethanol C3 and heavier alcohols C5 and heavier ethers Other ethers
715–780 91/95/98 82/85/88 45–60 215 360 5
715–770 91/95/98 82 1/2/85/88 45–60 195 480 5
715–770 91/95/98 82 1/2/85/88 45–60 195 480 5
715–770 91/95/98 82 1/2/85/88 45–60 195 480 5
–b 50 5 1000 0.4c 2.7
20 40 2.5 200 –c 2.7
10 35 1 30 –c 2.7
10 35 1 5–10 –c 2.7
0 10 0.1 –b –b
0 10 0.1 –b –b
0 10 0.1 –b –b
0 10 0.1 –b –b
a Region
specific based on the climate and season. regulated by the fuel specification. c Where lead addition is still permitted (category 1), otherwise no intentional addition. b Not
13.3 Motor-Gasoline Properties
and associated pressure increase is much higher, giving rise to a knocking or pinking noise. Engine knocking over an extended period will destroy the engine. The ability of a fuel to resist autoignition during engine operation is a key motor-gasoline quality parameter. The autoignition resistance of a motor-gasoline fuel is quantified by its octane number. It is a relative measure and compares the autoignition behavior of the fuel to that of a binary mixture of n-heptane and 2,2,4-trimethylpentane. By definition, the octane number of n-heptane is 0 and that of 2,2,4-trimethylpentane is 100. There are two octane measures commonly used: 1) Research octane number (RON). The autoignition behavior is measured at mild conditions, which are representative of low and medium engine speeds. Test conditions are 600 rpm, no prewarming of the mixture, and constant ignition timing. The RON is evaluated according to the ASTM D 2699 standard test method (previously ASTM D 908) [6]. In older literature, reference is also made to this as the F1 test method. A too low RON leads to ‘‘acceleration knocking.’’ 2) Motor octane number (MON). The autoignition behavior is measured at severe conditions that represent high speed and high load. Test conditions are 900 rpm, prewarming of the mixture to 150 Ž C, and variable ignition timing. The MON is evaluated according to the ASTM D 2700 standard test method (previously ASTM D 357) [7]. In older literature, this is also referred to as the F2 test method. A too low MON leads to ‘‘high-speed knocking.’’ The difference between the RON and MON of a motor-gasoline is called the sensitivity of the octane number. Because both measures are important in characterizing the resistance to autoignition of a fuel, the octane number can also be expressed as a road octane number or antiknock index (Equation 13.1). Anti-knock index D
1 Ð(RON C MON) 2
(13.1)
The octane value indicated on service station pumps is either the anti-knock index, like in the United States, or the RON of the motor-gasoline. The octane number requirement of an engine is affected by many parameters, which include the compression ratio of the engine, condition of the engine, atmospheric conditions, and driving style. An increase in altitude or humidity, and a decrease in ambient temperature can reduce the octane number requirements. Numerous more complex correlations have been developed to relate the road octane performance to laboratory-determined RON and MON values [8]. In order to relate the octane number to the molecular properties of the motor-gasoline, it is important to understand the reactions that lead to autoignition. Autoignition and the oxidation of fuel are free radical chain reactions involving air, which ultimately lead to the conversion of the fuel into stable oxidation products (mainly CO2 and H2 O). There are two properties to consider. The first is the susceptibility of a compound to autoxidation. Autoxidation is the noncatalyzed reaction of molecular oxygen (O2 ) in air with a molecule and usually proceeds by forming a hydroperoxide, which is thermally labile and subsequently decomposes [9]. The second is the stability of the radical product formed by decomposition and its tendency to propagate the free radical chain reaction. Autoignition requires thermal decomposition that is accelerated by autoxidation and productive propagation of the free radical chain. Both processes are needed. The octane number of a compound will consequently be high if it can resist either or both: that is, resist autoxidation or produce free radicals that are stable to retard propagation or that terminate propagation.
255
256
13 Motor-Gasoline
The octane numbers of many pure compounds have been measured [10]. These can be used to calculate the octane number of a blend based on their volumetric contribution. The simplest blending model calculates the octane number of the blend as a linear combination of the volume of each pure compound (Vi ) and the octane number of the pure compound (Equations 13.2 and 13.3): (13.2) RON D (Vi ÐRONi ) Vi MON D
(Vi ÐMONi ) Vi
(13.3)
This type of blending model tends to be conservative in its estimation of the overall octane number of the blend, because it does not take any synergistic effects into account. If a compound with a low octane number is mixed into a blend that will retard free radical propagation, the mixture may have a higher octane number than expected based on linear volumetric blending. So, although the compound with low octane number readily leads to autoignition on its own, the blend more than compensates for this by effectively stabilizing the free radicals produced. One may therefore calculate a blending octane number for a pure compound in a specific fuel mixture. The blending octane number is the octane number of the pseudo-pure compound that would have given the octane number of the blend when a linear blending calculation (Equations 13.2 and 13.3) is performed. However, linear blending is not always conservative, and negative deviations may also occur. In the absence of experimental data, it is best to make provision for possible deviations from linear blending. Deviations of RON values are usually small and within 0.5 of a unit, whereas deviations of MON values are larger, typically 1–2 units [8]. It is customary for refineries to develop blending models based on their specific fuel components. When doing modifications to a refinery design, it is important to remember that a major change will affect the blending model. The blending octane number of a compound or refinery stream depends on the properties of the fuel with which it will be blended. For example, the blending octane numbers of low-octane n-alkanes are considerably higher when blended with a high-octane base stock [11]. The powerful effect of compounds that retard free radical propagation can be seen from the success of TEL addition to improve octane number. Although TEL readily decomposes on heating to produce radicals, the lead (Pb) thus produced is a radical scavenger that retards propagation and results in smoother combustion (Equation 13.4). ž
Pb(CH2 CH3 )4 Pb C 4 CH2 CH3
(13.4)
Let us now examine the octane numbers of the different compound classes that are allowed in motor-gasoline: 1) Alkanes. The octane numbers of the alkanes decrease with increasing carbon number and decreasing degree of branching (Table 13.4) [10]. The octane number sensitivity of the alkanes is low, and the RON and MON values are usually close to each other. The position of the branching is very important and, when a molecule has significant n-alkane character, it has a correspondingly low octane number. This is a direct consequence of the high autoxidation susceptibility of n-alkanes. The aliphatic hydrocarbons in Fischer–Tropsch syncrude is predominantly linear. The n-alkanes have very low octane numbers and, in order
13.3 Motor-Gasoline Properties Research octane numbers of selected C3 –C8 alkanes to illustrate the effect of carbon number and degree of branching.
Table 13.4
Carbon number
C3 C4 C5 C6 C7 C8
Research octane number (RON) Linear
2-Methyl
2,3-Dimethyl
2,2,3-Trimethyl
112 93.8 61.7 24.8 0 <0
– 101.3 92.3 73.4 42.4 20.6
– – – 103.5 91.1 71.3
– – – – 112 109.6
to produce motor-gasoline, the n-alkane content must be substantially reduced. Conversion by hydroisomerization (Chapter 18) and reforming (Chapter 22) are two useful refining pathways in this regard. The relationship between the octane number of the alkane and the skeletal structure of the isomers also explains the stringent demands placed on alkene oligomerization (Chapter 19) technology when the alkenes are to be hydrogenated. 2) Alkenes. As a compound class, the alkenes generally have higher octane numbers than the corresponding alkanes, except for highly branched molecules where the alkanes outperform the alkenes. The octane numbers are less sensitive to the skeletal structure of the alkenes but are quite sensitive to the position of the double bond in linear isomers (Table 13.5) [10]. The n-1-alkenes have much lower octane numbers than the other double bond isomers. The alkenes in Fischer–Tropsch syncrude are predominantly n-1-alkenes and the reason for the success of double bond isomerization as a refining technology is quite evident. The octane number sensitivity of alkenes is higher, especially in the more branched alkenes. The degree of branching that is required to have a high RON is less for alkenes than for Octane numbers of selected C6 –C8 alkenes to illustrate the effect of carbon number, position of the double bond, and degree of branching.
Table 13.5
Compound
1-Alkene trans-2-Alkene trans-3-Alkene 2-Methyl-1-alkene 2-Methyl-2-alkene 2,3-Dimethyl-1-alkene 2,3-Dimethyl-2-alkene
C6 -isomers
C7 -isomers
C8 -isomers
RON
MON
RON
MON
RON
MON
76.4 92.7 94.0 94.2 97.8 101.3 97.4
63.4 80.8 80.1 81.5 83.0 82.8 80.5
54.5 73.4 89.8 90.7 91.6 99.3 97.5
50.7 68.8 79.3 78.8 79.2 84.2 80.0
28.7 56.3 72.5 70.2 – 96.3 93.1
34.7 56.5 68.1 66.3 – 83.6 79.3
257
258
13 Motor-Gasoline Table 13.6 Octane numbers of selected aromatic hydrocarbons in the motor-gasoline boiling range.
Compound Benzene Toluene Ethylbenzene o-Xylene m-Xylene p-Xylene Cumene o-Ethyltoluene 1,3,5-Trimethylbenzene sec-Butylbenzene o-Cymene p-Diethylbenzene a No
RON
MON
–a 120.1 107.4 –a 117.5 116.4 113.1 102.5 >120.3 106.7 106.0 106.0
115.0 103.5 97.9 100.0 115.0 109.6 99.3 92.1 106.0 95.7 96.0 95.2
value reported in ASTM DS 4B.
alkanes. This explains the octane insensitivity of olefinic motor-gasoline and why many olefinic motor-gasoline blends have an RON around 95 and a MON around 80. 3) Cycloalkanes. The cycloalkanes have octane numbers that roughly approximate that of the corresponding acyclic alkenes that would be obtained on ring opening. 4) Aromatics. Most aromatic compounds in the motor-gasoline boiling range have high octane numbers (Table 13.6) [10]. This is related to their autoxidation resistance and, in the case of aromatics with longer alkyl chains, the retarding effect of the decomposition products. For example, cumene is quite susceptible to autoxidation, but the aromatic decomposition product is a phenoxy radical, which effectively delocalizes the unpaired electron in the π-electron system of the aromatic. Methyl aryl ethers can therefore also be used as octane number improvers [12]. The syncrude from high-temperature Fischer–Tropsch (HTFT) synthesis contains some aromatics, but in syncrude from low-temperature Fischer–Tropsch (LTFT) synthesis aromatics are practically absent (Table 1.2). In order to produce motor-gasoline, it is necessary to produce additional aromatic compounds by reforming. 5) Oxygenates. The main oxygenate classes considered as fuel additives are alcohols and ethers. Ketones and carbonates have also been considered as octane improvers [13, 14]. Oxygenates in the motor-gasoline boiling range are usually high octane number compounds, but not always. The autoignition resistance of oxygenates are related to their propagation propensity, and the octane number is consequently dependent on the structure of the hydrocarbon backbone of the oxygenate. For example, it has been reported that n-butanol and n-pentanol did not result in an improvement of octane number of an unleaded test gasoline [15]. It has also been found that some ethers in the motor-gasoline range have quite low blending octane numbers (Table 17.2). Fischer–Tropsch syncrude contains oxygenates, and light alcohols can be recovered from the aqueous product. Octane improvement by
13.3 Motor-Gasoline Properties
oxygenate addition makes sense in a Fischer–Tropsch refining context. Having said that, a word of caution: oxygenates are politically sensitive motor-gasoline components and there is a risk involved in relying on oxygenates to meet octane requirements. The history of 2-methoxy-2-methylpropane (MTBE) addition to motor-gasoline in the United States is a case in point (Section 2.4.4). Various attempts have been made to correlate the octane number to molecular properties [16] and compound classes [17]. These correlations may, on average, yield good predictions. These models were not evaluated with Fischer–Tropsch-derived motor-gasoline per se, although the models are such that the origin of the motor-gasoline should not matter. However, significant variation of octane numbers within a specific group may result in considerable error if the refining process is predisposed to a specific skeletal structure. Although fuel specifications require a minimum octane number for the total fuel, it is important that the fuel should have a good octane number over the whole distillation range. This is especially important for older vehicles with carburetor engines, since differential distillation may occur in the carburetor, causing the engine to see more of the lighter boiling fraction during warm-up. If the fuel has a poor octane number in the light boiling range, the engine may experience knocking, despite the fact that the fuel meets the overall octane number requirement. Another octane-related aspect that influences refinery design is the way in which the minimum octane number requirement is legislated. The European fuel specifications (Table 13.1) have minimum requirements for both RON and MON. In the United States, fuel specifications regulate the antiknock index, which provides an additional degree of freedom during refinery design. 13.3.2 Density
Although density is included in motor-gasoline specifications, it is not an important property for engine operation or protection. It has an influence on the energy value of the fuel and therefore on vehicle performance and fuel economy. In a refinery, the real importance of fuel density is one of economics. Fuel is sold on a volumetric basis, and marketing a fuel with the lowest possible density is advantageous to the refiner (least mass sold for the highest income). In practice, density is a consequence of the refinery design rather than the converse being true. Fischer–Tropsch-derived naphtha is usually less dense than crude-oil-derived naphtha. However, on-specification motor-gasolines produced from syncrude and crude oil have little difference in composition or density [18]. 13.3.3 Volatility
Volatility is characterized by the distillation profile and vapor pressure of the motor-gasoline. The volatility of the motor-gasoline should be matched to the average ambient conditions. Summer and winter grades have different volatilities. Since lighter and more volatile compounds generally have higher octane numbers (Section 13.3.1), a refinery is typically designed to meet summer-grade motor-gasoline specifications. Winter-grade specifications can be met by
259
260
13 Motor-Gasoline
including more butanes in the fuel and cutting back on reformer severity (if it is allowed by the refinery H2 balance). The same principle applies to Fischer–Tropsch refineries, except that the reformer operation is not constrained by the H2 balance, because H2 can be obtained from the Fischer–Tropsch gas loop. The distillation profile of motor-gasoline is measured by the ASTM D 86 standard test method [19]. The profile can be expressed either as T-points or E-points. The T-point terminology refers to the temperature at which a certain volume percentage of material has been distilled. For example, the T10 refers to the distillation temperature at which 10% (by volume) of the material has been distilled. The E-point terminology refers to the percentage of fuel that has been distilled at a specific temperature. For example, E100 refers to the volume percentage of fuel that is distilled overhead at 100 Ž C. The need for T10, T50, and T90 characteristics is related to various aspects of engine performance. If the T10 is too high, the fuel exhibits poor cold-starting behavior and when it is too low, problems with vapor lock, hot starting, and evaporative losses can be encountered. If the T50 is too high, the fuel will have poor short-trip economy, will exhibit rough acceleration, and have poor warm-up performance, while, if the T50 is too low, injector icing can occur (fuel too volatile and flashes over the injector port). If the T90 is too high, combustion deposits may form in the engine, and if it is too low, it will result in poor long-trip fuel economy. The distillation profile has a direct influence on the driveability index (DI), which is a measure of engine performance during the warm-up period. A lower DI means better performance and lower exhaust emissions during the warm-up period. The oxygenate-corrected expression for DI is given by (Equation 13.5) DI D 1.5Ð(T10) C 3Ð(T50) C (T90) C 11Ð(vol% oxygenates)
(13.5)
The maximum vapor pressure of a fuel is separately regulated. The vapor pressure is an indication of evaporative emissions during filling, which is an environmental, safety, and health concern. Fuel specifications typically refer to the Reid vapor pressure (RVP) of a fuel, which is the vapor pressure at 37.8 Ž C (100 Ž F) as measured by the ASTM D 323 standard test method [20]. The RVP decreases with increasing carbon number (Table 13.7) [10]. Table 13.7
Vapor pressure at 37.8 Ž C (100 Ž F) of C4 –C8 hydrocarbons.
Compound class
n-Alkane Branched alkane n-1-Alkene n-Alkene (internal) Branched alkene Cycloalkane Aromatic
Reid vapor pressure (kPa) C4
C5
C6
C7
C8
357 500 436 315–344 445 235 –
107 141–253 132 104–106 99–182 68 –
34 42–68 41 34–37 29–90 23–31 22
11 14–23 14 12–13 12–31 10–18 7
4 5–8 5 4–5 5–11 3–10 2–3
13.3 Motor-Gasoline Properties
13.3.4 Fuel Stability
The storage stability of a fuel is related to its oxidation stability. Although care is taken to store motor-gasoline with the least possible exposure to air, air cannot be excluded during storage and transfer. Over time, the motor-gasoline can be autoxidized, even at ambient conditions. This period can be shortened by the presence of some metal ions. Oxidation reactions take place by a free radical mechanism. There is typically an induction period during which radical initiation takes place but little oxidation occurs. After the induction period, oxidation results in fuel degradation by the formation of oxygenates, such as hydroperoxides, alcohols, and carbonyls. Some of the oxygenates (notably carbonyl compounds) readily form heavier compounds, or gums, that are detrimental to engine cleanliness. The oxidation stability of motor-gasoline is measured in terms of its induction period. The induction period is determined under conditions that accelerate oxidation, as defined by the ASTM D 525 standard test method [21]. Although the oxidation stability of a fuel is very important, it is not a fuel property that is determined by refining. The compound classes generated during refining may improve or undermine the inherent oxidation stability of the fuel, but oxidation stability is typically addressed by fuel additives, such as metal deactivators, oxidation inhibitors, and radical scavengers [9, 22]. Gums are mainly formed by reactive oxygenates and dienes during motor-gasoline storage. These heavy organic products are usually soluble in the motor-gasoline. On evaporation during normal engine operation, the heavier products may remain as a sticky residue. In fuel testing, a distinction is made between potential gums and existing gums. The potential gum relates to oxidation stability and the presence of reactive compounds that will readily form heavier products. The potential gum content is determined by the ASTM D 873 standard test method [23]. The existing gum content of motor-gasoline leaving a refinery is generally very low and it is determined by the ASTM D 381 standard test method [24]. Fischer–Tropsch syncrude contains various oxygenates. Many of these oxygenates are not detrimental to storage stability or gum formation. Nevertheless, care must be taken to remove the carbonyl compounds, which are prone to aldol condensation that result in gums. 13.3.5 Alkene Content
Evaporative emissions of light alkenes contribute to ozone formation. This is the primary reason for legislation to reduce the alkene content in fuels. Alkenes in fuel may also lead to gum formation and engine deposits, but this is mostly caused by dienes, not the monoolefins. The alkene content is a critical fuel specification for Fischer–Tropsch refineries, since the gasoline-range material produced by Fischer–Tropsch synthesis is olefinic in nature. Producing an olefinic motor-gasoline from Fischer–Tropsch syncrude is the refining pathway that requires the least effort. This can also be seen from the commercial synthetic motor-gasoline produced by Fischer–Tropsch refining at Sasol Synfuels (Table 9.7), which has an alkene content approaching 30% [18].
261
262
13 Motor-Gasoline
13.3.6 Aromatic Content
Aromatic molecules are good octane components with a high energy density. The negative attributes of aromatics in motor-gasoline are mainly related to its benzene content. Benzene is a known human carcinogen, and limiting the benzene content of fuel reduces not only the toxic tailpipe emissions but also benzene emissions due to evaporative losses at filling stations. Benzene has a high vapor pressure (Table 13.7). Motor-gasoline specifications limit the total aromatic content (which includes benzene), as well as the benzene content specifically. Heavier aromatics have been shown to result in engine deposits, particularly combustion chamber deposits, which increase exhaust emissions. The heavy aromatic content is indirectly controlled by specifying the final boiling point (FBP) of motor-gasoline. HTFT syncrude contains some aromatics, including benzene, whereas LTFT syncrude contains almost no aromatics. The limitation on aromatic content is therefore not an inherent issue for syncrude. However, during refining to produce motor-gasoline, additional aromatics will typically be produced. The refinery design determines the nature and amount of aromatics in the final motor-gasoline. A high benzene content can be avoided by judicious selection of refining technologies, or alternatively, when benzene is produced in significant quantity, the benzene can be alkylated (Chapter 20). 13.3.7 Sulfur Content
Sulfur is present in motor-gasoline because it is present in all conventional crude oils (Section 2.2.2). Sulfur increases exhaust emissions, increases ozone formation potential, adversely affects exhaust gas oxygen sensors, and inhibits catalytic activity of exhaust gas catalytic converters. No positive effects of having sulfur have been noted and there is general consensus that sulfur should be eliminated from motor-gasoline. Fischer–Tropsch-derived motor-gasoline has an inherent advantage, since all sulfur is removed in the pretreatment step before Fischer–Tropsch synthesis (Chapter 3). Unless sulfur is added, Fischer–Tropsch-derived motor-gasoline has zero sulfur. The use of sulfided hydroprocessing catalysts during refining and blending with sulfur-containing material, such as coal liquids, may introduce a small amount of sulfur into the motor-gasoline. The level of sulfur thus introduced is easily kept below the specification limit. 13.3.8 Oxygenate Content
Motor-gasoline produced from conventional crude oil does not usually include oxygenates. In a crude oil refinery, oxygenates can be viewed as a high-volume chemical additive to motor-gasoline, since it relies on external acquisition, or on-purpose synthesis of the required oxygenates. The high octane numbers of some alcohols and ethers make them attractive fuel blending components. Fuel ethers are, from a technical point of view, more attractive than alcohols
13.3 Motor-Gasoline Properties
because fuel ethers have a lower vapor pressure and do not have the demixing tendency of alcohols [25]. When alcohols are blended with motor-gasoline, it is imperative that the fuel system is water-free. In the presence of water, the short-chain alcohols naturally partition to the aqueous phase, as is found in Fischer–Tropsch refineries (Section 5.3.3). Comparative data on the impact of oxygenates on air quality also indicated that fuel ethers resulted in lower volatile organic emissions, NOx , ozone precursors, and carcinogens, compared to ethanol [26]. Yet, despite the advantages of fuel ethers over alcohols, one of the most frequently used fuel ethers, MTBE, was banned from use in parts of the United States [27]. The banning of MTBE was not due to technical reasons, but as a result of poor storage tank care. Leaking storage tanks caused MTBE to contaminate the groundwater. Since the root cause has not been addressed, it will be interesting to see how public perception in the United States is affected when ethanol leaks into the groundwater. Oxygenate inclusion is not as politicized everywhere in the world. For example, fuel ethers (including MTBE) are allowed by the European specifications (Table 13.1). However, it illustrates the potential volatility of legislation governing the oxygenate content of fuel. Fischer–Tropsch syncrude contains oxygenates. The light alcohols can be recovered from the aqueous product. Fuel ethers can be synthesized from recovered alcohols, or the alcohols can be used to synthesize fuel ethers (Chapter 17). The amount and species of oxygenated compounds that can be included in motor-gasoline are regulated. Since the alcohols have a market as chemicals also, it is not necessary for a Fischer–Tropsch refiner to include the oxygenates in motor-gasoline. Despite the oxygenate content of Fischer–Tropsch syncrude, it is easy to adhere to the oxygenate specifications. 13.3.9 Metal Content
The history of the introduction and phasing out of TEL as octane improver has been discussed (Chapter 2). In some countries, lead replacement fuels have been marketed. This was partly motivated by refiners who wanted to avoid the cost associated with refinery upgrades and partly by concern from motor manufacturers about valve seat recession (VSR). VSR takes place as a result of the action of the hot exhaust gases leaving the cylinder head under conditions of continuous high load or high-speed operation. The hot gases ‘‘erode’’ the valve seat on the cylinder head. Historically, the risk VSR was reduced by lead that bound to the valve seat and thereby protected it from VSR. With proper cooling, as is typical in modern engines, and with mixed-mode driving, the possibility of VSR is greatly reduced. Some additives based on phosphorous, alkali metals, and manganese have been employed to provide lead replacement motor-gasoline for older vehicles. In practice, VSR has not been seen in countries where lead was phased out. Arguments for adding metals to motor-gasoline seem tenuous and, in countries where lead was phased out, the policy of ‘‘no intended metal addition’’ was adopted. There is no need for metal addition to Fischer–Tropsch-derived fuels.
263
264
13 Motor-Gasoline
13.4 Aviation-Gasoline
Aviation-gasoline is used as fuel for spark-ignition engines on aircraft. It is therefore a gasoline, just like motor-gasoline, but it has more stringent requirements due to the performance demands placed on aircraft engines. It is not just a spark-ignition engine fuel, but it is an aviation fuel that has many specifications in common with jet fuel, which is an aviation turbine fuel (Chapter 14). The requirements for aviation-gasoline are given by ASTM D 910 standard (Table 13.8) [28]. There are different grades, and for the higher quality grades the RON rating is replaced by a performance number. The performance number evaluates the autoignition behavior under rich-mixture supercharged conditions and is related to the RON (Equation 13.6) [29]: Performance number D 2800Ð(128 RON)1
(13.6)
In order to achieve high performance numbers, the addition of TEL is allowed in aviation-gasoline. However, unlike lead addition to motor-gasoline, TEL must be mixed with at least a stoichiometric proportion of 1,2-dibromoethane. 1,2-Dibromoethane acts as a lead scavenger to prevent lead oxide deposits in the engine. It reacts with the lead oxide to form lead bromide and lead oxybromides, which are volatile compounds that can be removed with the rest of the combustion products during engine operation [29]. In order to produce aviation-gasoline in a Fischer–Tropsch refinery, the naphtha refining units must be selected in such a way that a very high-quality base stock is produced. This requires good Table 13.8
Selected aviation-gasoline specifications.
Property
Aviation-gasoline grade
Performance number, min. RON, min. MON, min. Vapor pressure at 37.8 Ž C (kPa) Freezing point (Ž C) Potential gum content (mg/100 ml), max. Net heat of combustion (MJÐkg1 ), min. Sulfur content (µgÐg1 ), max. Lead content (gÐl1 ), max Distillation profile (Ž C) E10, max. E40, min. E50, max. E90, max. FBP, max. (E10 C E50), min. a Calculated
80
91
100
100LL
– 87.0 80.0 38–49 58 6 43.5 500 0.14
– 98.0 91.0 38–49 58 6 43.5 500 0.56
130.0 106.5a 99.5 38–49 58 6 43.5 500 1.12
130.0 106.5a 99.5 38–49 58 6 43.5 500 0.56
75 75 105 135 170 135
75 75 105 135 170 135
75 75 105 135 170 135
75 75 105 135 170 135
from the performance number requirement.
13.5 Future Motor-Gasoline Specification Changes
quality, high-octane paraffinic blend components over the whole gasoline distillation range, as well as sufficient aromatics.
13.5 Future Motor-Gasoline Specification Changes
Refinery designers must anticipate change and, where possible, allow flexibility in their design to deal with such anticipated changes. Making predictions about future motor-gasoline specification changes is helpful in making design decisions, but these are educated guesses and should be used as guidelines only. The statements will be motivated, and it is left to the reader to decide how much credence each prediction deserves. 1) Octane number. With the phasing out of lead, some refineries had to invest in high-octane motor-gasoline generating capacity. As was the case with aviation-gasoline after the Second World War, these changes left refineries with ‘‘surplus’’ capacity to produce high-octane motor-gasoline. The public perception was that octane was a good all-round measure of motor-gasoline performance, and it was exploited by marketers. The high-octane generating capacity was used to produce fuels with higher than legislated octane numbers. This trend of high-octane unleaded motor-gasoline, typically RON 98 and MON 88, may grow in countries where the fuel market is deregulated and the pump price is determined by supply and demand. This in turn may provide an opportunity for manufacturers to exploit the benefits of higher octane number fuels, such as producing engines with higher compression ratios. Producing high-octane motor-gasoline by Fischer–Tropsch refining is not a problem, although it is more challenging to do so from LTFT syncrude than from HTFT syncrude. 2) Alkene content. In crude oil refineries, alkenes are mainly produced in high-temperature conversion units, such as fluid catalytic crackers. Alkene availability in crude oil refineries is therefore limited. However, olefinic motor-gasoline blending components that have a high octane number, but also a significant fraction of pentenes, are regularly included in the final motor-gasoline. The short-chain alkenes increase exhaust emissions, notably the ozone-forming emissions. Specification changes to limit the alkene content in motor-gasoline are really aimed at minimizing light alkenes. The impacts on crude oil refineries are not significant, and it is unlikely that there will be much resistance to specification changes limiting the alkene content to 10% or even less. The impact on Fischer–Tropsch refining is significant because syncrude inherently is alkene-rich. This not a serious problem for new Fischer–Tropsch refineries, which can be designed accordingly, but it is a problem for existing refineries. 3) Aromatic content. Although the maximum allowable aromatic content decreased in successive motor-gasoline specifications, there is a high cost associated with the hydrogen needed to decrease the aromatic content further. In addition to this, a decrease in high-octane aromatics implies that higher octane alkanes are required. This is difficult to achieve and can hardly be justified considering the limited impact the aromatic content has on emissions. However, it is expected that the trend in benzene reduction will continue. In future, the maximum benzene content in motor-gasoline will likely be less than 1%. Benzene is a known carcinogen and is responsible for much of the emissions associated with aromatics. There are consequently good reasons to limit benzene. The technology exists to keep refinery
265
266
13 Motor-Gasoline
benzene levels low and in many instances much lower than the legislated requirements. The Environmental Protection Agency (EPA) claims that 70% of the benzene emissions in the United States are due to transportation fuels and has suggested that the benzene content in motor-gasoline must be limited to 0.62% by volume [30]. 4) Sulfur content. The marked reduction in sulfur content is likely to continue in future. Sulfur provides no benefit to motor-gasoline, and it only serves to increase emissions. Its reduction is environmentally sound. Many refineries already opted for installing sulfur removal capacity that allows them to produce fuels with 10 µgÐg1 or less sulfur. Reducing the legislated sulfur content much beyond this level becomes a challenge for reactor hydrodynamics and refinery compliance. 5) Metal content. The complete phasing out of lead and other metal-containing additives has already taken place in many countries. This trend is likely to continue and also for good environmental reasons. Aviation-gasoline, which initially escaped the lead ban, is likely to become low-lead or lead-free in future. Some lead-free grades are already available, but considering that grades 100 and 100LL have a minimum MON requirement of 99.5, phasing out lead is more difficult, because it requires a very good base stock. 6) Oxygenate content. The drive to include fuel from renewable sources provides a strong political incentive to increase the maximum oxygenate content of motor-gasoline. Changes in this direction can be seen in directives such as 2009/30/EC [31]. Future changes in oxygenate content may also be accompanied by compound specific restrictions; for example, oxygenates must be included as ethanol or ethanol derivatives. The best way to deal with many of these anticipated changes is to design refineries with very good paraffinic motor-gasoline-producing conversion units. Alkanes are the only compound class that is not restricted by motor-gasoline specifications. Alkanes are also inherently the lowest octane number compounds in a motor-gasoline blend. By ensuring that the alkanes have a high octane number, the refinery is supplied with a good base stock for motor-gasoline and aviation-gasoline blending.
References 1. Colucci, J.M. (2004) Fuel quality – an essential
element in vehicle emission control. Proceedings of the ASME Internal Combustion Engine Division Technical Conference, Long Beach, CA. 2. Totten, G.E. (ed.) (2003) Fuels and Lubricants Handbook: Technology, Properties, Performance, and Testing, ASTM, West Conshohocken, PA. 3. Dabelstein, W., Reglitzky, A., Sch¨ utze, A., and Reders, K. (2008) in Handbook of Fuels (ed. B. Elvers), Wiley-VCH Verlag GmbH, Weinheim, pp. 97–195. 4. (2006) World-Wide Fuel Charter, 4th edn, Alliance of Automobile Manufacturers, Washington DC.
5. ASTM (2004) D 4814 – 04b. Standard specifi-
6.
7.
8.
9.
cation for automotive spark-ignition engine fuel, ASTM, West Conshohocken, PA. ASTM (2004) D 2699 – 04a. Standard test method for research octane number of spark-ignition engine fuel, ASTM, West Conshohocken, PA. ASTM (2004) D 2700 – 04a. Standard test method for motor octane number of spark-ignition engine fuel, ASTM, West Conshohocken, PA. Annable, W.G. and Pless, L.G. (1962) in Advances in Petroleum Chemistry and Refining, vol. VI (eds K.A. Kobe and J.J. McKetta Jr.), John Wiley & Sons, Inc., New York, pp. 3–70. Scott, G. (1965) Atmospheric Oxidation and Antioxidants, Elsevier, Amsterdam.
References 10. (1988) Physical Constants of Hydrocarbons and
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Non-Hydrocarbon Compounds, ASTM Data Series DS 4B, 2nd edn, ASTM, Philadelphia, PA. Heck, R.H. (1989) Contribution of normal paraffins to the octane pool. Energy Fuels, 3, 109–111. Singerman, G.M. (1982) Gasoline extenders/ octane improvers from coal liquids. Energy Prog., 2 (2), 100–106. Pacheco, M.A. and Marshall, C.L. (1997) Review of dimethyl carbonate (DMC) manufacture and its characteristics as a fuel additive. Energy Fuels, 11, 2–29. Golombok, M. (1999) Tautomerism and octane quality in carbonyl-containing oxygenates. Ind. Eng. Chem. Res., 38, 3776–3778. Yacoub, Y., Bata, R., Gautam, M., and Martin, D. (1997) The performance characteristics of C1 -C5 alcohol-gasoline blends with matched oxygen content in a single cylinder SI engine. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 42 (2), 723–727. Albahri, T.A. (2003) Structural group contribution methods for predicting the octane number of pure hydrocarbon liquids. Ind. Eng. Chem. Res., 42, 657–662. Ghosh, P., Hickey, K.J., and Jaffe, S.B. (2006) Development of a detailed gasoline composition based-model octane model. Ind. Eng. Chem. Res., 45, 337–345. Kamara, B.I. and Coetzee, J. (2009) Overview of high-temperature Fischer-Tropsch gasoline and diesel quality. Energy Fuels, 23, 2242–2247. ASTM (2005) D 86 – 05. Standard test method for distillation of petroleum products at atmospheric pressure, ASTM, West Conshohocken, PA. ASTM (1999) D 323 – 99a. Standard test method for vapor pressure of petroleum products (Reid method), ASTM, West Conshohocken, PA. ASTM (2005) D525 – 05. Standard test method for oxidation stability of gasoline (induction period method), ASTM, West Conshohocken, PA.
22. Ingold, K.U. (1961) Inhibition of the autoxida-
23.
24.
25.
26.
27.
28.
29.
30. 31.
tion of organic substances in the liquid phase. Chem. Rev., 61, 563–589. ASTM (2002) D 873 – 02. Standard test method for oxidation stability of aviation fuels (potential residue method), ASTM, West Conshohocken, PA. ASTM (2009) D 381 – 09. Standard test method for gum content in fuels by jet evaporation, ASTM, West Conshohocken, PA. Travers, P. (2001) in Petroleum Refining, Conversion Processes, vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 291–319. Hodge, C. (2003) Comment: more evidence mounts for banning, not expanding, use of ethanol in US gasoline. Oil Gas J, 101 (38), 18–20. Lamberth, R. (2004) 2003 was a year of transition for the MTBE, fuels industry. Oil Gas J, 102 (2), 52–58. ASTM (2004) D 910 – 04a. Standard specification for aviation gasolines, ASTM, West Conshohocken, PA. Bacha, J., Barnes, F., Franklin, M., Gibbs, L., Hemighaus, G., Hogue, N., Lesnini, D., Lind, J., Maybury, J., and Morris, J. (2000) Aviation Fuels, Chevron, San Ramon, CA. Hogue, C. (2007) Less benzene in gasoline. Chem. Eng. News, 85 (8), 8. (2009) Directive 2009/30/EC of the European Parliament and of the Council of 23 April 2009 amending Directive 98/70/EC as regards the specification of petrol, diesel and gas-oil and introducing a mechanism to monitor and reduce greenhouse gas emissions and amending Council Directive 1999/32/EC as regards the specification of fuel used by inland waterway vessels and repealing Directive 93/12/EEC. Off. J. Eur. Union, L140, 88.
267
269
14 Jet Fuel 14.1 Introduction
Aviation turbine fuel, or jet fuel, was introduced during the Second World War for military aircraft and its use has since become widespread for both military and civilian aircraft. It displaced aviation-gasoline (Section 13.4) as the main fuel type for aircraft, as the power source of most aircraft changed from spark-ignition engines to turbine engines. Jet fuel is a kerosene range fuel. Since jet fuel is used in a turbine engine, it requires very different fuel properties than aviation-gasoline. Combustion is not initiated in a closed combustion chamber, but must be sustained in an open chamber. The energy content and combustion quality of the fuel are key performance properties. The jet fuel is directly and continuously combusted with compressed air in a combustion chamber and the hot gases are used to drive a turbine. Poor combustion will not only lead to energy loss and high hydrocarbon emissions, but also to the generation of particulate matter that can damage the turbine. Poor engine performance in an aircraft has more disastrous consequences than poor engine performance during road transportation. Furthermore, aircraft are not restricted to a single country. The specifications for jet fuel are therefore mostly of an international nature, without much of the local variations in legislation seen for motor-gasoline and diesel fuel. For the same reasons, it is difficult to make changes to jet fuel specifications, and these specifications are not subject to political pressures in the same way as motor-gasoline and diesel fuel are. From a refining point of view, jet fuel is an easy fuel to produce. It has less risk of significant specification changes, but it has a smaller market base. Depending on the refinery design, the kerosene range material used for jet fuel can be blended with road transportation fuels and fluctuations in jet fuel demand can be easily absorbed. The structure and aim of this chapter is similar to that of the chapters dealing with motor-gasoline (Chapter 13) and diesel fuel (Chapter 15). The jet fuel specifications will be related to the molecular requirements of the fuel and the demands placed on Fischer–Tropsch refineries. More detailed descriptions of jet fuels can be found in the literature [1, 2].
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
270
14 Jet Fuel
14.2 Jet Fuel Specifications
The jet fuel originally used for turbine engines was illuminating kerosene, which was produced for wick lamps. It was initially thought that turbine engines are relatively insensitive to fuel properties and kerosene was chosen because of its availability. Later, a wider cut that included some lighter material was used, but it was abandoned for general use because of greater evaporative losses at high altitude and the safety risks involved in dealing with such a volatile fuel. The British DEF STAN 91-91 (previously DERD 2494) [3] is widely recognized as the international standard to specify the civil aviation turbine fuel Jet A-1 (Table 14.1). Country-specific jet fuel specifications are closely aligned with the Jet A-1 specification published in the DEF STAN 91-91. The United States still uses Jet A for national air travel, which is defined by the ASTM D 1655 standard [4]. The main difference between Jet A and Jet A-1 is in the freezing point requirement.
Table 14.1
Selected specifications for civilian aviation turbine fuels.
Property Net heat of combustion (MJ kg−1 ), min. Density at 15 ◦ C (kg·m−3 ) Freezing point (◦ C), max. Vapor pressure (kPa), max. Flash point (◦ C), min.b Viscosity at −20 ◦ C (cSt), max. Smoke point (mm), min. Existing gums (mg/100 ml), max. Lubricity, BOCLE (mm), max.b Composition, max. Aromatic content (vol%)b Naphthalene content (vol%)e Sulfur content (mass%) Thiol content (mass%) Acid content (mg KOH/g) Distillation (◦ C), max.b IBP T10 T20 T50 T90 FBP a Not
Jet A-1
Jet A
Jet B
42.8 775–840 −47 –a 38 8 25c 7 0.85d
42.8 775–840 −40 –a 38 8 25c 7 –a
42.8 751–802 −50 21 –a –a 25 7 –a
25 3 0.3 0.003 0.1
25 3 0.3 0.003 0.1
25 3 0.3 0.003 –a
Report 205 –a Report Report 300
–a 205 –a Report Report 300
–a –a 145 190 245 –a
limited by the specification. specification requirements for Fischer–Tropsch-derived synthetic jet fuel. c Requirement can be lowered to 19 mm (Jet A-1) or 18 mm (Jet A) if used in conjunction with naphthalene content. d Only a requirement if the jet fuel contains >95% hydroprocessed material. e Not a separate requirement if the smoke point is higher than 25 mm. b Different
14.2 Jet Fuel Specifications
Jet A has a maximum freezing point specification of −40 ◦ C, whereas for Jet A-1 it is −47 ◦ C, which makes Jet A-1 more suitable for long international flights. The wider cut, more volatile Jet B is used only in Arctic regions, mainly due to its better cold-flow properties. The specifications listed in Table 14.1 do not include all the stability and additive requirements. Additives are strictly regulated. In Jet A-1, antioxidants are required in any fuel composition that has been hydroprocessed. The approved antioxidants for aviation fuel are hindered phenols with a maximum allowable concentration of 24 mg·l−1 . The addition of a metal deactivator is allowed, and the only approved metal deactivator is N,N -disalicylidene-1,2-propane diamine. Jet fuel additives are a very important part of producing final on-specification jet fuel. These molecules are added during blending and do not affect jet fuel refining. The limited discussion is therefore not a reflection on their importance; additives are crucial, but it is a reflection on their impact on refinery design. 14.2.1 Synthetic Jet Fuel
The marketing effort to differentiate Fischer–Tropsch-derived fuels from crude-oil-derived fuels had some unintended consequences for jet fuel derived from syncrude. The perception was created that the molecules in syncrude are somehow different from their counterparts in crude oil. Although the jet fuel refined from HTFT syncrude falls well within the composition and specification limits defined for crude-oil-derived jet fuel, it required a lengthy qualification process before it was allowed for use in jet fuel. In addition to the specifications for Jet A-1 refined from conventional crude oil, Jet A-1 refined from syncrude had to meet a subset of more stringent requirements (Table 14.2) [3]. The ‘‘marketing’’ differentiation of synthetic jet fuel had a knock-on effect on the ASTM D 1655 specification also [4]: ‘‘Jet fuels containing synthetic hydrocarbons have been previously allowed under Specification D 1655. However, the fraction of these hydrocarbons was not limited, and Supplementary civil aviation turbine fuel specifications specific to synthetic jet fuels containing material that were obtained from Fischer–Tropsch synthesis.
Table 14.2
Property
Fischer–Tropsch material (vol%), max. Flash point (◦ C) Lubricity, BOCLE (mm), max. Aromatic content (vol%) Distillation (◦ C), min. (T50 – T10) (T90 – T10) a No
Jet A-1 specifications Semisynthetic
Fully synthetic
50 –a 0.85 8–25
100 38–50 0.85 8–25
– –
a a
20 40
additional specification restrictions beyond that required for conventional Jet A-1.
271
272
14 Jet Fuel
there were no requirements or restrictions placed on either these hydrocarbons or the final blend. It has been recognized that synthetic blends represent a potential departure from experience and from key assumptions on which the fuel property requirements . . . have been based.’’ The ASTM D 1655 specification was aligned with the DEF STAN 91-91, and the potential use of non-Sasol Fischer–Tropsch-derived jet fuel as civil aviation turbine fuel was effectively scuttled. The first synthetic jet fuel blend to be qualified as fuel for civilian aircraft was a semisynthetic jet fuel. This blend was allowed to contain at most 50% material derived from Fischer–Tropsch synthesis. In fact, the DEF STAN 91-91 specification is explicit in limiting the Fischer–Tropsch-derived material in semisynthetic jet fuel to hydrogenated oligomers from the Sasol Synfuels facility (Figure 9.11). These hydrogenated oligomers, which are also called iso-paraffinic kerosene (IPK), are obtained from solid phosphoric acid-catalyzed oligomerization of C3 –C4 HTFT condensate followed by hydrogenation over a sulfided base metal catalyst. Since this qualification process was completed before 2004, many of the subsequent change to the refinery (Section 9.5) altered the composition of the IPK. The DEF STAN 91-91 specification was also explicit in prohibiting synthetic aromatics in this material. It is not clear why these restrictions were imposed, since there are no obvious technical grounds for this. In fact, considering the stricture of the specification, it is not clear whether future semisynthetic jet fuel production by Sasol Synfuels is technically in violation of the DEF STAN 91-91 specification due to changes in refinery configuration and operation. Fully synthetic jet fuel has been qualified for use since April 2008 [3]. As in the case of the semisynthetic jet fuel, the composition of fully synthetic jet fuel is limited specifically to a blend of light distillate, heavy naphtha, and IPK (Figure 9.12). The light distillate and heavy naphtha streams are hydrotreated, straight-run HTFT products. Again, the technical grounds for the stricture are not clear. The qualification process nevertheless highlighted some important requirements that should hold true for synthetic and crude-oil-derived jet fuel. These are as follows: 1) The jet fuel must have a reasonably wide and even boiling point distribution. The temptation to selectively remove certain carbon numbers as chemicals and the selective addition of large volumes of narrow boiling range materials must be avoided. 2) Fuel systems that were exposed to aromatics cannot be reliably used with aromatics-free fuels. There must be a minimum amount of aromatics. 3) Synthetic jet fuel is naturally more refined and attention must be paid to lubricity, which is typically destroyed during hydroprocessing. 14.2.2 Fuel for Military Use
Military aircraft rely on the supply logistics of their country or alliance of origin. In this respect, such aircraft are not subject to the same need for international specifications as civilian aircraft. Different subspecifications may therefore exist to accommodate specific military applications (Table 14.3) [2]. Fischer–Tropsch-derived fuels, with their associated supply security independent of crude oil, are attractive for military applications. The concept of a single Battlefield Use Fuel of the Future (BUFF) is also a way to simplify supply logistics. Since kerosene is mainly a light distillate
14.3 Jet Fuel Properties United States military jet propulsion (JP) aviation turbine fuels and their corresponding maximum freezing point and minimum flash point specifications.
Table 14.3
Fuel
Freezing point (◦ C)
Flash point (◦ C)
RVP (kPa)
Fraction
Comments
JP-1 JP-2 JP-3 JP-4 JP-5 JP-6 JP-7 JP-8
−60 −60 −60 −72 −46 −54 −43 −47
43 – – – 60 – 60 38
– <14 34–48 14–21 – – – –
Kerosene Wide cut Wide cut Wide cut Kerosene Kerosene Kerosene Kerosene
Obsolete, introduced in 1944 Obsolete, introduced in 1945 Obsolete, introduced in 1947 Air Force fuel (Jet B analog) Navy fuel (aircraft carrier use) Obsolete, XB-70 program in 1956 Lower volatility, higher thermal stability Air Force fuel (Jet A-1 analog)
RVP, Reid vapor pressure.
fraction, a paraffinic kerosene can in principle be refined to meet specifications for both turbine engines and compression-ignition engines [5]. No standard has been agreed upon for such a BUFF, but it is clear that Fischer–Tropsch-derived kerosene fractions may be well suited for such an application. A number of studies have reported efforts to refine Fischer–Tropsch syncrude and coal liquids to JP-5 type jet fuels [6–10].
14.3 Jet Fuel Properties
The composition of a large number of kerosene blends, mostly in the 190–230 ◦ C boiling range, were compared to jet fuel properties routinely used to specify jet fuel quality [11]. The kerosene fractions were obtained from different crude oils and coal liquefaction. This work is remarkable in that it was able to clearly show which compound classes are required to produce on-specification jet fuel. The compound classes were lumped into three main groups: the n-alkanes [n], the branched alkanes and cycloalkanes [BC], and the aromatics [Ar]. Each of the jet fuel properties was linearly correlated to the jet fuel composition expressed in terms of the mass fraction of the aforementioned three compound classes (Equation 14.1). Jet fuel property = a1 [n] + a2 [BC] + a3 [Ar] + a4
(14.1)
These correlations are helpful in understanding the relationship between jet fuel properties and the composition of the blend. Using these relationships, it is possible to construct a diagram that identifies the compositional space allowed by jet fuel specifications for synthetic Jet A-1 (Figure 14.1). The analysis has one inherent shortcoming when applied to Fischer–Tropsch-derived synthetic jet fuel, which relates to the lumping of branched alkanes and cycloalkanes as the [BC] group. The cycloalkanes have comparable freezing points to the branched alkanes, but
273
274
14 Jet Fuel
[n]
Minimum aromatics Maximum aromatics Minimum density On-specification synthetic Jet A-1
Smoke point Maximum density Maximum freezing point
[BC]
[Ar]
Figure 14.1 Relationship between jet fuel composition and the synthetic jet fuel specification domain. The composition is defined by grouping the n-alkanes as [n], the branched alkanes and cycloalkanes as [BC], and the aromatics as [Ar].
cycloalkanes on average have a 70–80 kg·m−3 higher density than the branched alkanes. For crude-oil- and coal-liquid-derived jet fuels, which contain a significant amount of cycloalkanes, the minimum density specification is not constraining (Figure 14.1), but for Fischer–Tropsch liquids that may not be the case. Nevertheless, it does not detract from the analysis by Cookson et al. [11]. 14.3.1 Net Heat of Combustion
The net heat of combustion is the amount of energy released when the fuel is completely oxidized to produce carbon dioxide and water vapor. The standard method to determine the net heat of combustion is the ASTM D 4809 [12], which employs oxidation in a bomb calorimeter. It should be noted that the aviation turbine fuel specification places a limit on the minimum gravimetric energy content, not the volumetric energy content (Table 14.1). A jet fuel with a low density and that is rich in alkanes will have a high gravimetric energy content. Fischer–Tropsch-derived jet fuels fit this description well and will easily meet the energy specification. Conversely, a fuel with a high density that is rich in aromatics will have a high volumetric energy content but low gravimetric energy content. In practice, there is a trade-off between using a fuel with a high gravimetric energy content as opposed to a fuel with a high volumetric energy content.
14.3 Jet Fuel Properties
1) Military applications. The fuel tank space in an aircraft is limited, which imposes a volumetric constraint on flight range. Carrying fuel with a high volumetric energy content (high density) will improve the flight range for a given tank capacity. In military applications, a fuel with high volumetric energy content is consequently desirable. The extreme case is jet fuel for missiles, where a very dense fuel is beneficial. For example, JP-10 has an energy density of 39.4 MJ·l−1 (For Jet A-1, it is typically around 35.1 MJ·l−1 ) [1]. 2) Civilian applications. Selecting a fuel for commercial aircraft involves more trade-offs. Although the fuel tank capacity on civilian aircraft is also limited, aircraft usually take on only enough fuel to reach their destination, with some additional fuel to give it an adequate safety margin. Selecting a fuel with a high gravimetric energy content (low density) is more efficient, because the energy weighs less and fuel tank capacity is not constraining. However, fuel is sold volumetrically. This implies that a jet fuel with high gravimetric energy content costs more for the same energy content as a jet fuel with high volumetric energy density. As a consequence, in most cases a fuel with high volumetric energy density (high density) is preferred, but airliners seldom have the luxury to pick the energy content of their fuel.
14.3.2 Density and Viscosity
The fuel viscosity and, to a lesser extent, the fuel density influence the spray pattern and droplet size of the fuel when it is injected under high pressure into the combustion chamber of a turbine engine. The fuel system is designed to produce a fine spray that can easily evaporate when it is mixed with the hot air from the compression section. Density is not really an independent variable. The jet fuel specifications allow a wide density range, and in practice density is not controlled, but a result of the limitations placed on the fuel composition. Fischer–Tropsch-derived materials produce jet fuels on the lower end of the density scale, mainly due to its high alkane and low cycloalkane content. The spray pattern and the droplet size distribution are important to the performance of a turbine engine. When the viscosity is increased and the density is decreased, it will change the droplet size distribution to produce larger droplets [13]. If the droplets are too large it affects the performance of the engine. Larger droplets also pose a safety risk, because it makes it difficult to relight an engine in flight after flameout has occurred. (Flameout occurs when the air velocity exceeds the flame propagation velocity.) Another effect of higher viscosity is to increase the pressure drop in the fuel lines. This is detrimental to aircraft operation for a number of reasons. The pump duty required to maintain a constant fuel flow is increased if the viscosity is too high, and the fuel pump may not be able to supply the engine with the required amount of fuel. High viscosity will also reduce the cooling efficiency of the jet fuel as heat exchange fluid on aircraft (Section 14.3.7). The maximum viscosity in jet fuel is consequently regulated by specification. On the lower end of the viscosity scale, too low a viscosity is not good either. The hydrodynamic lubricity of a fuel is influenced by its viscosity. A jet fuel with too low a viscosity may cause excessive wear on pump parts and flow control units. Although there is no minimum limit for viscosity; the distillation profile effectively sets a lower viscosity limit.
275
276
14 Jet Fuel
Materials with high density and molecular mass usually have a higher liquid viscosity. The heavy end of the distillation profile strongly influences the viscosity, and the inclusion of material that has a boiling point higher than 260 ◦ C is limited by the viscosity requirements. 14.3.3 Freezing Point Temperature
The freezing point is defined as the highest temperature at which all compounds in the fuel are still in the liquid phase. The primary reference method for freezing point determination of aviation fuel is ASTM D 2386 [14], but other more modern methods may also be used. Since the jet fuel is a mixture of many compounds, the fuel mixture does not become a solid at the freezing point. As the temperature is lowered below the freezing point of the fuel, compounds will start crystallizing out of solution to create a slush of fuel and solid hydrocarbons. This affects the ability of the fuel to be moved from the fuel tanks to the engine. Jet fuel remains pumpable only to between 4 and 15 ◦ C below its freezing point [1]. Fischer–Tropsch syncrude is rich in linear materials, which have high freezing points (Table 14.4). In order to meet the freezing point specification of jet fuel, the straight-run kerosene fraction of syncrude must be hydroisomerized (Chapter 18). Paraffinic kerosene suitable for jet fuel can also be produced from lighter material by oligomerization (Chapter 19) followed by hydrogenation, or from heavier material by hydrocracking (Chapter 21). 14.3.4 Aromatic Content and Smoke Point
Fuel components that have a tendency to form carbonaceous particles during the early stages of combustion should be avoided. In the event that such particles are formed, these particulates must be completely consumed before leaving the combustion chamber. The formation of such carbonaceous particles in the combustion chamber is detrimental in two ways. Firstly, the particles become incandescent at the high temperature and pressure conditions of the combustion chamber, which can cause hot spots on the chamber wall due to the high additional heat transfer rate. This may cause cracks and lead to premature engine failure. The particles can also block the holes in the combustion chamber wall that supply air to the combustion section, thereby disrupting the flow pattern. Secondly, if these particles are not consumed in the combustion chamber, they impinge on the turbine blades and stators, causing erosion of the turbine section [1]. Such particles are also responsible for visible smoke. Aromatic compounds and especially naphthalenic compounds are more prone to the formation of such particles than aliphatic hydrocarbons. Both the total aromatic content and the total naphthalenic content of jet fuel are therefore regulated by the aviation turbine fuel specifications. The smoke point is a measure of the tendency of a fuel to form particles (black smoke) during combustion, and the method for its determination is described by the ASTM D 1322 standard test method [15]. It is a simple test wherein the fuel is burned in a wick-fed lamp, and the smoke point is the maximum flame height that can be achieved without smoke being formed. The aromatic content of Fischer–Tropsch syncrude is low, and even in the heaviest aromatic fraction of HTFT syncrude the dinuclear aromatic content is low. Fischer–Tropsch-derived jet fuels are inherently more hydrogen-rich, have a high smoke point, and burn clean. In comparison
14.3 Jet Fuel Properties Table 14.4
Selected physical properties of C9 –C15 hydrocarbons.
Compound
n-Alkanes n-Nonane n-Decane n-Undecane n-Dodecane n-Tridecane n-Tetradecane n-Pentadecane Branched alkanes 2-Methyloctane 2-Methylnonane 2,5-Dimethyloctane 2-Methyldecane 2,3-Dimethylnonane 2-Methylundecane 2,5-Dimethyldecane 2-Methyldodecane 2,4-Dimethylundecane 2-Methyltridecane 2,3-Dimethyldodecane 2-Methyltetradecane 2,4-Dimethyltridecane Cycloalkanes n-Propylcyclohexane cis-1,2-Diethylcyclopentane cis-Decalin n-Butylcyclopentane n-Butylcyclohexane n-Pentylcyclohexane n-Hexylcyclohexane n-Heptylcyclohexane n-Octylcyclohexane Aromatics n-Propylbenzene Cumene n-Butylbenzene sec-Butylbenzene p-Cymene p-Diethylbenzene n-Pentylbenzene n-Hexylbenzene n-Heptylbenzene n-Octylbenzene a Density
reported at 20 ◦ C.
Formula
Boiling point (◦ C)
Density at 15 ◦ C (kg·m−3 )
Freezing point (◦ C)
C9 H20 C10 H22 C11 H24 C12 H26 C13 H28 C14 H30 C15 H32
150.8 174.2 195.9 216.3 235.2 253.8 270.7
721.9 734.2 744.5 752.7 761.7 763.3 772.2
−53.5 −29.6 −25.6 −9.6 −5.4 5.9 9.9
C9 H20 C10 H22 C10 H22 C11 H24 C11 H24 C12 H26 C12 H26 C13 H28 C13 H28 C14 H30 C14 H30 C15 H32 C15 H32
143.3 167 158 189.9 186.8 210.2 198.1 229.4 216.8 247.4 245.9 264 250
717.7 730.6 737a 736.9a 747.1a 745.6a 747a 753.3a 754.6a 760.4a 768.4a 766a 767.1a
−80.4 −74.7 −84.5 −49.5 −117.7 −46 −60.4 −26 −75.5 −26.5 −57.9 −8.9 −44
C9 H18 C9 H18 C10 H18 C10 H20 C10 H20 C11 H22 C12 H24 C13 H26 C14 H28
156.7 153.6 195.8 156.6 177 203 224 244.9 263.6
793.6a 800.4 901.8 810.3 799a 804a 808a 811.2 817.2
−94.9 −118 −43 −108 −79 −57.5 −43 −30.5 −19.7
C9 H12 C9 H12 C10 H14 C10 H14 C10 H14 C10 H14 C11 H16 C12 H18 C13 H20 C14 H22
159.2 152.4 183.3 173.3 177.1 183.8 205.4 226.1 246.1 264.4
868.3 868.5 866 866.2 860.7 866.3 862.9 862.1 860.8 860.2
−99.5 −96 −88 −75.5 −67.9 −42.8 −75 −61 −48 −36
277
278
14 Jet Fuel
with crude-oil-derived JP-8 (Jet A-1 type) jet fuel, Fischer–Tropsch-derived synthetic jet fuel and blends thereof have considerably lower particulate matter emissions [16]. In general, the combustion quality of jet fuel is related to its hydrogen content, and in Fischer–Tropsch-derived jet fuels the hydrogen content is usually higher than most crude-oil-derived jet fuels. A high smoke point and low particulate emissions are therefore to be expected on the basis of the high hydrogen content. 14.3.5 Sulfur and Acid Content
The limitation on the sulfur content in jet fuel is mostly to reduce SOx emissions, but it has the added advantage of reducing microbial growth during fuel storage. A key aspect to limiting biological activity is limiting free water, but reducing elemental nutrients like sulfur also helps. In terms of fuel quality, the limitation placed on the thiol (mercaptan) content is more crucial. Thiols and organic acids can cause corrosion of some engine and fuel system components and are therefore limited by jet fuel specifications. Fischer–Tropsch syncrude is sulfur-free but contains carboxylic acids. The acid content can be reduced by hydrotreating (Chapter 16). 14.3.6 Volatility
The volatility of jet fuel is regulated by specification of the distillation profile. In the case of synthetic jet fuel, this specification is more stringent (Table 14.2). The distillation profile indirectly controls other properties also, such as the minimum viscosity and combustion quality. By specifying a minimum slope for the distillation profile of synthetic jet fuel, the gradual evaporation of the fuel over the whole boiling range is ensured. By doing so, it reduces the risk of flash evaporation associated with narrow boiling synthetic compounds, which would cause the formation of fuel-rich and air-poor pockets in the turbine engine. 14.3.7 Stability
There are two stability aspects to consider for jet fuel. The first is storage stability, which refers to the ability of the fuel to resist autoxidation during storage. The second is thermal stability, which refers to the ability of the fuel to resist the formation of particulates, gums, and deposits when the fuel comes into contact with hot surfaces. The thermal stability of jet fuel is very important, since the fuel serves as a heat exchange medium in the engine and airframe to cool engine oil, hydraulic fluid, and air conditioning equipment. It is a key fuel property, especially in high-performance aircraft, such as used by the military. Thermal stability is measured by the ASTM D 3241 standard test method [17]. Fischer–Tropsch hydrocarbons generally improve the thermal stability of jet fuel blends under mild oxidative stress conditions [18]. However, this beneficial effect is dependent on the amount of prestressing. Prestressing reduces the amount of antioxidants in the fuel, which leads to faster subsequent oxidation [19]. Phenol (not hindered alkyl phenols) has been implicated
14.3 Jet Fuel Properties
in the formation of deposits during thermal stability testing, and a mechanism involving phenol oxidation was proposed [20] that described the observations with various crude oil and Fischer–Tropsch semisynthetic jet fuel mixtures reasonably well [18]. One of the advantages of Fischer–Tropsch-derived synthetic jet fuel for naval aviation is the very low copper migration [8]. This is a direct consequence of the low sulfur level in Fischer– Tropsch-derived jet fuel. Copper ions are not the most efficient ions to catalyze hydroperoxide decomposition [21–23], but by reducing the dissolution of copper, fuel stability is improved. The contamination of syncrude with dissolved iron, either from the Fischer–Tropsch catalyst or from carboxylic acid corrosion, must be prevented. Even at very low levels, iron can degrade the stability of jet fuel [24]. In this respect, cobalt is an even more potent autoxidation catalyst [23]. Metal contaminants from Fischer–Tropsch syncrude must consequently be avoided in synthetic jet fuel production. 14.3.8 Elastomer Compatibility and Lubricity
During normal operation with crude-oil-derived jet fuel, the elastomeric materials in contact with the fuel absorb some of the jet fuel components, which results in seal swelling. This is a natural and expected consequence of exposing the elastomers to the fuel. It is therefore not an issue that has to be addressed by the jet fuel specifications. Elastomer swelling also takes place while in contact with synthetic jet fuels. However, the amount of swelling in the presence of crude-oil-derived jet fuel and a paraffinic fuel can be quite different [25, 26]. When an aircraft switches from conventional to synthetic jet fuel, or vice versa, it is important that the elastomeric materials retain their degree of swelling. When the degree of elastomeric swelling changes, it can cause leaks and this must be avoided. Semisynthetic jet fuel, irrespective of whether it is produced from IPK, avoids elastomer compatibility problems by blending with conventional jet fuel. In a more general sense, this holds true for semisynthetic jet fuel blends with any paraffinic Fischer–Tropsch-derived material. Although the DEF STAN 91-91 specifications recognize only HTFT-derived IPK, in this regard there is little difference between HTFT-derived IPK and an analogous IPK derived from hydrocracked LTFT material. The qualification of fully synthetic jet fuel from HTFT syncrude (Section 14.2.1) demonstrated that a synthetic jet fuel that naturally contains aromatic compounds does not result in elastomer compatibility problems. The minimum aromatic specification in synthetic jet fuel is to ensure elastomer compatibility. However, it was pointed out that it is not only the aromatic content but also the presence of some oxygenates in low concentration that is needed for the aromatics to achieve the required volume swell [25]. Crude-oil-derived jet fuel contains low concentrations of sulfur compounds that perform the same function. Severely hydrotreated jet fuels may have different elastomer swelling characteristics, even though they contain percentage levels of aromatics. The heteroatom-containing compounds in jet fuel, irrespective of their origin, have another important function. Polar compounds provide boundary layer lubrication. In the case of Fischer–Tropsch-derived jet fuel, the low levels of oxygenates remaining in the jet fuel after refining are likely to be responsible for improved lubricity and elastomer swelling characteristics [27]. Severe hydrotreatment can remove these oxygenates and lead to poor lubricity as well as problems with differential elastomer swelling, despite the presence of aromatics.
279
280
14 Jet Fuel
The inclusion of a lubricity specification for severely hydrotreated conventional jet fuels and synthetic jet fuels is therefore understandable. This is a lesson that was learnt the hard way with diesel fuels (Section 15.3.4). Lubricity is determined by the ball-on-cylinder lubricity evaluator (BOCLE), as outlined in the ASTM D 5001 standard test method [28].
14.4 Future Jet Fuel Specification Changes
The nature of aircraft and jet engines, as well as the air transport industry in general, is such that changes in aviation turbine fuel specifications are difficult to make. There is a significant safety aspect involved, and any change in jet fuel specification must be qualified on different engine types before it can be accepted as safe for use. It is doubtful that significant changes in the specification of Jet A-1 will be seen. The most likely specification to come under scrutiny is the sulfur content. The specification limit of 0.3% is very high compared to the drive toward 0.001% in other transportation fuel types. For the crude oil refining industry, such a change will require a significant intervention, since a large portion of the jet fuel is produced by sweetening, and not by hydrotreating. Sweetening processes oxidatively convert thiols into disulfides and are not always followed by a disulfide extraction step. Even so, the sweetened kerosene will still contain other sulfur species, even when sweetening is followed by extraction. A different conversion step will be needed to remove the remaining sulfur. Fischer–Tropsch-derived jet fuels are not affected by such a change and may even benefit indirectly. As the sulfur level in conventional jet fuels is decreased, the trace compound differences between synthetic and conventional jet fuels will become smaller. This may pave the way for a single jet fuel standard, irrespective of the refining origin. Military applications of jet fuel may also help to unify jet fuel specifications. The development of BUFF is pertinent. Should the development of a BUFF be successful, it is possible that the qualifying tests done on military aircraft might give enough credence to also qualify such fuels for civilian aircraft. This in turn could provide the impetus to align the Jet A-1 specifications with BUFF specifications. It is not yet clear whether bio-derived fuels will try to make inroads in the jet fuel market. With sufficient political support this may well happen and, if it does, it may also catalyze a unification of Jet A-1 specifications for conventional and synthetic fuel types.
References 1. Bacha, J., Barnes, F., Franklin, M., Gibbs, L.,
Hemighaus, G., Hogue, N., Lesnini, D., Lind, J., Maybury, J., and Morris, J. (2000) Aviation Fuels, Chevron, San Ramon, CA. 2. Bishop, G.J. (2008) in Handbook of Fuels (ed. B. Elvers), Wiley-VCH Verlag GmbH, Weinheim, pp. 321–341. 3. UK Ministry of Defence (2008) Turbine Fuel, Aviation Kerosine Type, Jet A-1. NATO Code:
F-35. Joint Service Designation: AVTUR. Defence Standard 91-91, Issue 6; UK Ministry of Defence, 8 April 2008. 4. ASTM (2005) D 1655 – 05. Standard specification for aviation turbine fuels, ASTM, West Conshohocken, PA. 5. Forest, C.A. and Muzzell, P.A. (2005) Fischer-Tropsch fuels: why are they of interest
References
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
to the United States military? SAE Tech. Pap. Ser., 2005-01-1807. Freerks, R.L. and Muzzell, P.A. (2004) Production and characterization of synthetic jet fuel produced from Fischer-Tropsch hydrocarbons. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 407–410. Muzzell, P.A., Freerks, R.L., Baltrus, J.P., and Link, D.D. (2004) Composition of syntroleum S-5 and conformance to JP-5 specification. Prepr. Pap.-Am. Chem. Soc., Div. Pet. Chem., 49 (4), 411–413. Chang, P.H., Colbert, J.E., Hardy, D.R., and Leonard, J.T. (2004) Evaluation of Fischer-Tropsch synthetic fuels for United States Naval applications. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 414–417. Lamprecht, D. (2007) Fischer-Tropsch fuel for use by the U.S. military as battlefield-use fuel of the future. Energy Fuels, 21, 1448–1453. Heyne, J.S., Boehman, A.L., and Kirby, S. (2009) Development of a drop-in unifuel/single battlefield fuel of high thermal stability. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 35–37. Cookson, D.J., Lloyd, C.P., and Smith, B.E. (1987) Investigation of the chemical basis of kerosene (jet fuel) specification properties. Energy Fuels, 1, 438–447. ASTM (2005) D 4809 – 00. Standard test method for heat of combustion of liquid hydrocarbon fuels by bomb calorimeter (precision method), reapproved in 2005, ASTM, West Conshohocken, PA. Calvert, S. (1984) in Handbook of Air Pollution Technology (eds S. Calvert and H.M. Englund), Wiley-VCH Verlag GmbH, New York, p. 235. ASTM (2005) D 2386 – 05. Standard test method for freezing point of aviation fuels, ASTM, West Conshohocken, PA. ASTM (2002) D 1322 – 97. Standard test method for smoke point of kerosene and aviation turbine fuel, reapproved in 2002, ASTM, West Conshohocken, PA. Corporan, E., DeWitt, M.J., Belovich, V., Pawlik, R., Lynch, A.C., Gord, J.R., and Meyer, T.R. (2007) Emissions characteristics of a turbine engine and research combustor burning a Fischer-Tropsch jet fuel. Energy Fuels, 21, 2615–2626. ASTM (2005) D 3241 – 05b. Standard test method for thermal oxidation stability of aviation turbine fuels (JFTOT procedure), ASTM, West Conshohocken, PA.
18. Sobkowiak, M., Griffith, J.M., Wang, B., and
19.
20.
21.
22. 23.
24.
25.
26.
27.
28.
Beaver, B. (2009) Insight into the mechanisms of middle distillate fuel oxidative degradation. Part 1: on the role of phenol, indole, and carbazole derivatives in the thermal oxidative stability of Fischer-Tropsch/Petroleum jet fuel blends. Energy Fuels, 23, 2041–2046. Jones, G.E., Balster, L.M., and Balster, W.J. (1998) Effects of pre-stressing on the autoxidation of aviation fuel. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 43 (3), 346–348. Beaver, B., Gao, L., Burgess-Clifford, C., and Sobkowiak, M. (2005) On the mechanisms of formation of thermal oxidative deposits in jet fuels. Are unified mechanisms possible for both storage and thermal oxidative deposit formation for middle distillate fuels? Energy Fuels, 19, 1574–1579. Mesrobian, R.B. and Tobolsky, A.V. (1961) in Autoxidation and Antioxidants, vol. 1 (ed. W.O. Lundberg), Wiley-VCH Verlag GmbH, New York, pp. 107–131. Scott, G. (1965) Atmospheric Oxidation and Antioxidants, Elsevier, Amsterdam, pp. 88–92. Sheldon, R.A. and Kochi, J.K. (1981) Metal-Catalyzed Oxidations of Organic Compounds, Academic Press, New York, pp. 38–43. Pickard, J.M. and Jones, E.G. (1997) Catalysis of jet-a fuel autoxidation by Fe2 O3 . Energy Fuels, 11, 1232–1236. Graham, J.L., Striebich, R.C., Minus, D.K., and Harrison, W.E. (2004) The swelling of selected O-ring materials in jet propulsion and Fischer-Tropsch fuels. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem, 49 (4), 435–439. Graham, J.L., Minus, D.K., and Harrison, W.E. (2007) An investigation of aviation components material compatibility with blends of a Fischer-Tropsch-derived jet fuels with petroleum derived jet fuels. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 52 (2), 454–456. Link, D.D., Baltrus, J.P., Zandhuis, P.H., and Hreha, D. (2004) Separation and identification of oxygenates as suspected performance-enhancers for synthetic jet fuels. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 418–421. ASTM (2003) D 5001 – 03. Standard Test Method for Measurement of Lubricity of Aviation Turbine Fuels by the Ball-On-Cylinder Lubricity Evaluator (BOCLE), ASTM, West Conshohocken, PA.
281
283
15 Diesel Fuel 15.1 Introduction
Diesel fuel refers to the fuel that is used in compression-ignition engines (diesel engines). It is generally material boiling in the C11 –C22 hydrocarbon range. The requirements for diesel fuel are often diametrically opposed to that of motor-gasoline (Chapter 13). In motor-gasoline, it is important to suppress autoignition of the fuel to allow spark ignition to be correctly timed for good engine performance. In diesel fuel, it is important that the fuel autoignites and that the delay between fuel injection and the start of combustion is short. Compression-ignition engines operate at high compression ratios, typically around 15–17 : 1. This improves the thermodynamic efficiency of the engine, and on average compression-ignition engines are more efficient than spark-ignition engines. The engine itself is heavier, and traditionally it has been applied mainly for heavier vehicles. This situation has changed and, despite the higher cost of compression-ignition engines as compared to spark-ignition engines, there has been a marked shift in preference for diesel-powered passenger vehicles in Europe. This change is partly due to an increase in environmental awareness. Better engine efficiency leads to lower fuel consumption and can be translated into less CO2 emitted per distance traveled. The move to vehicles with lower rated emissions (in grams of CO2 per kilometer) is supported by the policies of the European Union. This shift in transportation fuel preference for the passenger vehicle market had a marked impact on refineries. The hydrogen in a typical conventional crude oil refinery is obtained from catalytic naphtha reforming. As the demand for motor-gasoline decreases, the need for reformate also decreases, which in turn reduces hydrogen production. Yet, refining crude oil to diesel fuel requires more hydrogen than refining crude oil to motor-gasoline. In each refinery there is a minimum motor-gasoline to diesel fuel ratio beyond which refining becomes inefficient and costly. The shift toward diesel fuel has created an imbalance in refinery production and in practice this imbalance in Europe is corrected by exporting motor-gasoline and importing diesel fuel. In the long run, it will be interesting to see how this might affect diesel fuel specifications, which over time have made diesel fuel production more difficult. Diesel fuel, like motor-gasoline, is a major consumer product and exerts a tremendous influence on the economy and the environment. Diesel fuel specifications are consequently subject to technical, environmental, and political pressures (Section 13.1). Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
284
15 Diesel Fuel
This chapter will employ a similar format as the preceding two chapters that dealt with motor-gasoline and jet fuel (Chapters 13 and 14). More detailed descriptions of fuels and fuel specifications can be found in the literature [1, 2]. The following aspects will be discussed in this chapter: 1) Current diesel fuel specifications, regional differences, and how the specifications relate to the molecular requirements of the fuel. 2) Direction of anticipated changes that will affect refinery designs in future. 3) The considerable difference in distillate properties from high-temperature Fischer–Tropsch (HTFT) synthesis and low-temperature Fischer–Tropsch (LTFT) synthesis. 4) The different demands placed on HTFT and LTFT refineries by the diesel fuel property requirements.
15.2 Diesel Fuel Specifications
Compression-ignition engines are robust and can operate reliably with heavy fuels of quite a varied composition. There are divergent views about diesel fuel quality and specifications. This can be seen when comparing the European EN590 diesel fuel specifications, the American ASTM D 975 specification, and the World Wide Fuel Charter (WWFC) guidelines (Tables 15.1 and 15.2). Table 15.1
American and European diesel fuel specifications.
Specification
D-1
D-2
ASTM D 975 Density at 15 ◦ C (kg·m−3 ) Cetane number, min. Viscosity at 40 ◦ C (cSt) Flash point (◦ C), min. Lubricity, HFRR at 60 ◦ C (µm), max. Hydrocarbon content (mass%), max. Total aromatics Polycyclic aromatics Sulfur content (µg·g−1 ), max. Water content (µg·g−1 ), max. Distillation (◦ C) T90, min. T90, max. T95, max. FAME addition (vol%), max.d a Not
Euro-2
Euro-3
Euro-4
EN590:1993
EN590:1999
EN590:2004
–a 40b 1.3–2.4 38 520
–a 40b 1.9–4.1 52 520
820–860 49 2.0–4.5 55 460
820–845 51 2.0–4.5 55 460
820–845 51c 2.0–4.5 55 460
–a –a 500/15 500
–a –a 500/15 500
–a 11 2000/500 200
–a 11 350 200
–a 11 50/10 200
–a 288 –a –a
282 338 –a –a
–a –a 370 –a
–a –a 360 –a
–a –a 360 5
regulated by the fuel specification. operation at higher altitude and low ambient temperature may require a higher cetane number; minimum cetane index requirement is also 40. c Minimum cetane index of 46 is required. d Fatty acid methyl ester (FAME) inclusion allow the blending of renewable material. b Engine
15.2 Diesel Fuel Specifications Table 15.2
Selected diesel fuel quality guidelines proposed by the WWFC.
Specification
Category 1
Category 2
Category 3
Category 4
Density at 15 ◦ C (kg·m−3 ) Cetane number, min. Cetane index, min.a Viscosity at 40 ◦ C (cSt) Flash point (◦ C), min. Lubricity, HFRR at 60 ◦ C (µm), max. Hydrocarbon content (mass%), max. Total aromatics Polycyclic aromatics Sulfur content (µg·g−1 ), max. Water content (µg·g−1 ), max. Distillation (◦ C), max. T90 T95 FBP FAME addition (vol%), max
820–860 48 48/45 2.0–4.5 55 400
820–850 51 51/48 2.0–4.0 55 400
820–840 53 53/50 2.0–4.0 55 400
820–840 55 55/52 2.0–4.0 55 400
–b –b 2000 500
25 5 300 200
20 3 50 200
15 2 10 200
–b 370 –b 5
340 355 365 5
320 340 350 5
320 340 350 0
a The higher value must be used if no cetane number improver is used; the lower value is valid only for the base fuel before addition of a cetane number improver. b Not specified by the guidelines.
Some trends can be seen from the changes in the European EN590 over time as well as the different WWFC categories (Table 13.2) [3]. There is a steady increase in the cetane number (CN) requirement for diesel fuel. This is mainly related to reported beneficial effects, such as lower crank time and lower hydrocarbon and CO emissions. Other changes are also motivated by emission reduction. In order to limit deactivation of tailpipe catalytic converters and reduce exhaust emissions in general, the maximum sulfur content in diesel fuel was dramatically reduced. The maximum polycyclic aromatic content was decreased in tandem with a lowering of the T95 distillation temperature to reduce particulate emissions. The ASTM D 975 specifications [4] are much less restrictive, although provision was made for low-sulfur diesel fuels. The diesel fuel specifications listed in Tables 15.1 and 15.2 are not a complete set of specifications. Cold-flow properties, which have not been listed, are important specifications and are dependent on the climate where the diesel fuel will be used. There are also other specifications, such as fuel stability requirements, that have not been listed. As in the case of motor-gasoline and jet fuel, final on-specification diesel fuel includes additives [2]. Additives address some of the properties that are not regulated by the fuel specifications, such as detergency, antifoaming properties, and water de-emulsifying behavior. There are also major diesel fuel specification requirements that can be addressed through the judicious use of additives. Since these additives make refining less onerous and consequently have a direct impact on refinery design, such additives will be discussed separately (Section 15.4).
285
286
15 Diesel Fuel
15.3 Diesel Fuel Properties
The molecular properties of a diesel fuel have a significant impact on engine performance and emissions. Even for Fischer–Tropsch-derived diesel fuels that have almost identical ASTM D 975 properties, observed NOx and particulate matter (PM) emissions can be quite different [5]. The point made by Bisio and Atkinson is pertinent and it is well stated [5]: ‘‘Often it is not appreciated that the properties of (and the nature and distribution of molecular species) in Fischer–Tropsch diesel are established in the downstream refinery processes . . . and not in the Fischer–Tropsch reactor.’’ Cookson, Lloyd, and Smith applied the same property correlation methodology as that employed for jet fuel (Section 14.3) to diesel fuel [6]. Diesel fuel properties were linearly correlated to the mass percentage of [n] the n-alkanes, [BC] the branched alkanes and cycloalkanes, and [Ar] the aromatics (Equation 14.1). It was noted that an expansion of the correlation to differentiate between mononuclear and dinuclear aromatics did not improve its predictive ability much. A more complex, but analogous approach has successfully been followed to relate PM emissions to the molecular properties of diesel fuels [7]. Using the approach of Cookson, Lloyd, and Smith [6], a specification domain can be constructed for diesel fuel that shows the compositional space allowed by diesel fuel specifications (Figure 15.1). As in the case of jet fuels, the inherent shortcoming of lumping branched alkanes and cycloalkanes into one parameter, [BC], is that it is not representative of Fischer–Tropsch-derived diesel fuel. The key difference between branched alkanes and cycloalkanes is density. It turns out that density is a critical property for meeting EN 590 type and WWFC type diesel fuels [8]. The diesel fuel density specification is constraining and, although Figure 15.1 gives a reasonable qualitative indication of the specification domain, it does not reflect the density deficiency associated with a low cycloalkane content.
15.3.1 Cetane Number
The cetane number is a measure of the compression–ignition delay of a fuel. The ignition delay is the time difference between injecting the fuel into the combustion chamber and the autoignition of the fuel. If the ignition delay is too long, it leads to noisy combustion, very high pressures in the combustion chamber, and increased NOx emissions. The ignition delay is one of the key quality parameters for a diesel fuel. The CN is a derived property value, which is a measure of the ignition delay of a fuel. A low value signifies a long compression–ignition delay, while a high value signifies a short delay. The CN is measured on a test engine in accordance with the ASTM D 613 standard test method [9]. It is defined on an arbitrary scale where n-hexadecane (n-cetane) has a value of 100 and 1-methylnaphthalene has a value of 0. Since 1962, 1-methylnaphthalene has been replaced by 2,2,4,4,6,8,8-heptamethylnonane as a primary reference fuel. The 2,2,4,4,6,8,8-heptamethylnonane has a cetane value of 15 as measured relative to that of the original definition of CN.
15.3 Diesel Fuel Properties
[n]
Minimum cetane number
On-specification EN590-type diesel fuel
Minimum density
Maximum cold-flow properties
[Ar]
[BC] Figure 15.1 Relationship between diesel fuel composition and the EN590 diesel fuel specification domain. The composition is defined by grouping the n-alkanes as [n], the branched alkanes and cycloalkanes as [BC], and the aromatics as [Ar].
The ASTM D 613 is an expensive and material-intensive test method. A derived CN can also be obtained from the more recently developed Ignition Quality Tester (IQT) [10], which measures the time from the start of fuel injection into a constant-volume combustion chamber to the start of combustion. The derived CN from the IQT is one-to-one correlated with the CN. The inverse relationship between CN and the ignition delay is the opposite of octane number (Section 13.3.1), which is a measure of autoignition resistance. The autoignition resistance is directly related to ignition delay. CN is correlated to the blending value of the motor octane number by (Equation 15.1) [11]: Cetane number = 58.9 − 0.47 · MON
(15.1)
As anticipated by this correlation, the compound classes that display poor octane numbers (Section 13.3.1) are those that have high CNs (Table 15.3) [12]. The CNs for isostructural aliphatic hydrocarbons increase with increasing carbon number [13], as one would expect from the change in octane number. Various derived methods have been developed to calculate the ignition delay properties of diesel fuels without having to resort to the difficult, time-consuming, and costly ASTM D 613 method for CN determination. Crude oil refiners often make use of the cetane index as an alternative to the CN. The cetane index is a calculated value based on the density and distillation properties of a diesel fuel. It correlates with the CN before the addition of cetane improvers
287
288
15 Diesel Fuel Table 15.3
Cetane numbers of various hydrocarbon compounds.
Compound class
Compound
n-Alkane
n-Decane n-Dodecane n-Tetradecane n-Hexadecane n-Octadecane 2,2,4,4,6,8,8-Heptamethylnonane 7,8-Dimethyltetradecane 5-Butyldodecane 1-Decene 1-Dodecene 1-Tetradecene 1-Hexadecene 1-Octadecene 5-Butyl-4-dodecene n-Butyldecalin n-Octyldecalin 3-Cyclohexylhexane trans-Decalin 2,6-Dimethylnaphthalene 1-n-Butylnaphthalene Tetralin n-Butyltetralin n-Hexylbenzene n-Octylbenzene
Branched alkane
n-1-Alkene
Branched alkene Cycloalkane
Aromatic
Cetane number 77 87 95 100 106 15 40 45 56 71 80 86 90 45 31 31 36 46 −13 6 13 16 26 32
(Section 15.4) and has a limited range of applicability (32.5 < CN < 56.5). It is generally not valid for Fischer–Tropsch-derived diesel fuels, since the correlation is implicit in assuming an average composition based on the refining of crude oil. Many similar correlations can be found in the literature, employing the physical and chemical characteristics of the diesel fuel [13–17]. Of these, the methods based on nuclear magnetic resonance (NMR) spectrometry has been found to be especially useful for very paraffinic distillates, typical of LTFT wax hydrocracking, and hydrogenated oligomers from alkene oligomerization. For good compression-ignition engine performance, a diesel fuel should have a CN that exceeds a minimum CN, which places an upper limit on the ignition delay. From the different specification values (Tables 15.1 and 15.2), it is clear that there is no consensus about the minimum value. Since the distillates refined from Fischer–Tropsch syncrude tend to have high CNs, the minimum CN requirement is not critical to Fischer–Tropsch refining. An aspect that is more interesting is the maximum CN, where LTFT distillates have an advantage over petroleum-derived diesel fuel. The view has been expressed that an ideal diesel fuel should have the highest possible CN [18]. This view was motivated by the better combustion performance and lower emissions that can, in principle, be obtained by high CN diesel fuels. An almost diametrically opposed view is expressed in the ASTM D 975 specifications in order to motivate the recommended
15.3 Diesel Fuel Properties
minimum CN requirement [4]: ‘‘Increase in cetane number over values actually required does not materially improve engine performance. Accordingly, the cetane number specified should be as low as possible to assure maximum fuel availability.’’ One may argue that synthetic diesel fuel availability is not necessarily constrained by a high CN. It would therefore be best to maximize CN whenever possible to take maximum advantage of the reduced emissions from high CN diesel fuels. However, in practice there is an upper limit to CN beyond which emission performance may be degraded. Tests with a high-speed direct-injection diesel engine indicated poorer PM emission performance by an 81 CN diesel fuel than a well-formulated 53 CN diesel fuel [7]. This was ascribed to the very short ignition delay that resulted in combustion commencing before sufficient fuel–air mixing occurred. The definition of CN is such that it is mainly a measure of the time required for free radical initiation leading to ignition. The CN of a diesel fuel can therefore be improved by blending the fuel with a thermally labile compound. Typical cetane improvers are compounds such as peroxides and alkyl nitrates. However, if CN was only a measure of initiation, the CN gain obtained by thermally labile compounds should have been independent of the fuel matrix, which is not the case. Ignition requires radical propagation, and a low CN base stock cannot be improved much by the addition of a cetane improver. If we translate these properties to the molecular requirements for CN, molecules will have a high CN if they are thermally labile or readily autoxidized. At this point, we can also anticipate that there are potential trade-offs involved in having a high CN and good fuel stability (Section 15.3.8). For example, the CN can be improved by autoxidation [19], and one may expect a slight increase in CN after prolonged storage. Resasco and coworkers made an exhaustive study of the relationship between CN and molecular properties in order to suggest refining pathways to improve diesel fuel quality [12]. It was found that the most effective description resulted from dividing the hydrocarbons into two groups: the first group consisting of the alkanes and cycloalkanes and the second group consisting of the alkenes and aromatics. Some common features that correlated well with CN emerged from the analysis, despite differences in the detail level description: 1) The highest positive charge on a hydrogen atom in the molecule. 2) Ovality of the molecule, expressed as the actual surface area relative to that of an equivalent sphere. 3) The number of –CH2 – groups in the molecule. 4) Some description of the connectivity. The study yielded valuable insights relevant to crude oil refining. The study also indirectly explained why straight-run HTFT and LTFT syncrudes have high CNs. In the case of HTFT syncrude, the alkyl aromatics and mixtures of linear and branched hydrocarbons (alkanes and alkenes) all have reasonable CNs, which can be further improved by mild hydrotreating. In the case of LTFT syncrude, which contains mainly n-alkanes, the CN is of course extremely high. The alcohols in Fischer–Tropsch syncrude can also be employed as diesel fuel components. The alcohols can be directly used as diesel fuel extenders [20]. However, of the oxygenate classes, the linear ethers have been found to provide the best compromise between CN and cold-flow properties [21]. Methoxymethane (dimethyl ether, DME) is the simplest of the linear ethers, with a CN of 55–60 and it has been extensively investigated as conventional diesel fuel substitute [22]. Longer chain linear ethers have been successfully tested as blend components with HTFT diesel
289
290
15 Diesel Fuel
fuel [23]. Etherification (Chapter 17) is a potentially useful refining technology for diesel fuel production in a Fischer–Tropsch refinery. 15.3.2 Density and Viscosity
The injection of diesel fuel into a compression-ignition engine is controlled either by a solenoid valve or by positive volumetric displacement. The density and viscosity of the fuel determine the performance of the injection system. It affects not only the energy value of the material that is injected, but also the droplet size distribution (Section 14.2.2). If the diesel density and viscosity are not within the narrow range for which the engine was calibrated, it will degrade the engine performance and increase emissions [3]. The progressive narrowing of both the density and viscosity range of diesel fuel specifications is therefore intentional. Fischer–Tropsch syncrude inherently has low density and viscosity. The diesel fuel obtained from HTFT syncrude is at the lower end of the acceptable range, but that from LTFT syncrude is usually well below the density of crude-oil-derived diesel fuels. When LTFT syncrude is hydroprocessed, the acyclic aliphatic nature of the syncrude is retained and the distillate has a high CN but low density. The low density is advantageous for the Fischer–Tropsch refiner, but not for the consumer. It was found that LTFT distillate had a 2.7% better gravimetric energy efficiency than crude-oil-derived diesel fuel, but a 7.2% lower density [24]. On a volumetric basis, it implies that LTFT distillate results in 5% higher volumetric fuel consumption for the same energy delivery. Refining strategies have been suggested to improve the density of Fischer–Tropsch-derived distillates without degrading the CN to an unacceptable level [8]. Such interventions are necessary only for stand-alone refineries that have to produce on-specification diesel fuel as the final product. In most situations, the density deficiency can be overcome by blending with petroleum products or pyrolysis products such as coal liquids. 15.3.3 Flash Point
The flash point of diesel fuels is mainly controlled from a safety point of view, since it gives an indication of the tendency of the fuel to form a flammable mixture with air. It is extensively used to assess the fire risk associated with the storage of potentially flammable products. The flash point is controlled by distillation, and, by increasing the temperature of the initial boiling point of the diesel, the flash point can also be increased. 15.3.4 Lubricity
Diesel fuel pumps that do not have an external lubricating system rely on the lubricity of the fuel to reduce wear. Inadequate lubricity can result in excessive pump wear and even pump failure. Historically, lubricity of diesel fuel was not an issue. The importance of specifying lubricity became apparent only after ultralow sulfur diesel (sulfur content of <50 µg·g−1 ) was introduced. Shortly after the introduction of the ultralow sulfur diesel into the Swedish market in 1992,
15.3 Diesel Fuel Properties
diesel-powered passenger cars started experiencing problems with their fuel pumps. Cars using the Bosch rotary pump reported failures within 3000 to 10 000 km, while other manufacturers reported reduced pump performance at short service life [25]. It turned out that the natural lubricating properties of the diesel were destroyed during the severe hydroprocessing to reduce the sulfur content of the diesel fuel. Diesel lubricity is determined by the ASTM D 6079 standard test method [26], which makes use of a high-frequency reciprocating rig (HFRR). Although other test methods can be used, this test has been shown to have a good correlation with actual automotive diesel fuel pump wear. The loss of lubricity during the production of low-sulfur diesel from crude oil seemed to indicate that lubricity is linked to the sulfur-containing species. Superficially, this contention was supported by the observation that hydroprocessed sulfur-free Fischer–Tropsch distillates also had poor lubricity [27]. However, it was shown not to be the case. The heteroatom-derived lubricity-improving properties of compounds in diesel fuel followed the trend O > N >> S [28]. Straight-run Fischer–Tropsch syncrude should consequently have good boundary layer lubricity, which it indeed has. Boundary layer lubricity improvement is not the only lubricity requirement; hydrodynamic lubricity that is related to viscosity is also important. Boundary layer lubricity is imparted by surface-active species. A critical surface concentration is necessary to provide adequate lubricity. The nonlinear response of diesel fuel lubricity in response to the addition of a lubricity improver is to be expected (Figure 15.2) [29]. The lubricity-enhancing properties of various oxygenates were investigated [25, 28–33]. Not all classes of oxygenates are equally effective as lubricity providers, and it has been shown that the lubricity of biodiesel, which is rich in methyl esters, can itself be improved by mild oxidation [34]. When different oxygenate classes of the same chain length were added to diesel fuel, the lubricity improved in relation to the polarity of the oxygenate used: ethers < ketones < methyl esters < 1-alcohols < aldehydes < carboxylic acids [28].
Lubricity, HFRR wear scar (µm)
600 Sunflower oil Olive oil Corn oil Used frying oil
500
400
300
200 0
0.2
0.4 0.6 Concentration (%)
Figure 15.2 Lubricity response of diesel fuel to the addition of different amounts of methyl esters obtained from different oils.
0.8
1
291
292
15 Diesel Fuel
During the refining of Fischer–Tropsch syncrude, it is important not to over-hydrogenate the distillate. With proper catalyst selection and operation, it is possible to hydrotreat syncrude in such a way that enough natural lubricity is retained to meet the diesel fuel specifications [35]. Bypassing hydroprocessing steps with some of the straight-run Fischer–Tropsch distillate can also markedly improve lubricity. Because of the nonlinear response of lubricity to surface-active material (Figure 15.2), this does not have to be a large fraction of the syncrude. 15.3.5 Aromatic Content
According to the diesel fuel specifications, a molecule is considered an aromatic if it contains at least one benzene ring in its structure. Consequently, there are a wide variety of molecules, with very different fuel properties, that can be classified as aromatics. This is why the fuel specifications differentiate between total aromatics and polynuclear aromatics, which are aromatics with more than one benzene ring. The total aromatic content determines the flame temperature during combustion. A higher aromatic content leads to higher PM and NOx emissions, but this does not imply that aromatics are the sole source of these emissions. In a study by ExxonMobil and Toyota using a modern high-speed direct-injection engine, it was found that PM showed a statistically significant correlation with aromatics, cycloalkanes, CN, and density [7]. The aromatics in straight-run Fischer–Tropsch syncrude and in the hydroprocessed material from HTFT and LTFT synthesis are mainly mononuclear (Table 15.4) [24, 36, 37]. 15.3.6 Sulfur Content
The sulfur in diesel fuel has a negative influence on engine performance and exhaust emissions. It contributes significantly to the formation of fine PM through the formation of sulfates in the exhaust stream and atmosphere. It can also lead to corrosion and wear of engine systems and reduce the efficiency of the exhaust emission after-treatment systems. The maximum sulfur content in diesel fuel has therefore, over time, been reduced to very low levels. The change in the sulfur specification of diesel fuel has no impact on Fischer–Tropsch refining, because the syncrude is sulfur-free. However, it influenced the marketing of Fischer–Tropsch
Table 15.4
Aromatic content of different Fischer–Tropsch-derived distillates.
Origin of distillate
Aromatic content (mass%) Total
SMDS distillate Sasol SPD distillate Straight-run HTFT distillate Hydrotreated HTFT distillate a No
1.4 0.47 27 22.45
Mononuclear
Dinuclear
Trinuclear
1 0.44 26 22.2
0.4a 0.03 0.9 0.24
– <0.01 0.1 <0.01
distinction made between di- and trinuclear aromatics.
15.3 Diesel Fuel Properties
products. The sulfur specification has become so stringent that it is no longer possible to use Fischer–Tropsch distillate as a low-sulfur blend stock with crude oil to help crude oil refiners meet the sulfur specification. 15.3.7 Cold-Flow Properties
The specifications for cold-flow properties are region-specific. It is a very important fuel specification. As the temperature of a diesel fuel is lowered, the highest cetane number components, that is the n-alkanes, tend to crystallize out of solution as a wax. The wax may block the fuel filter and fuel lines, rendering engine operation difficult or impossible. The cold-flow properties of a diesel are therefore defined by wax-related tests, such as the cloud point (CP) and cold filter plugging point (CFPP). CP is the temperature at which the diesel fuel becomes cloudy due to the formation of small wax crystals in the fuel. The observed turbidity is due to light scattering by the small crystals. This is not an accurate measure of the operability of the fuel, which may remain pumpable and filterable at temperatures below the CP. The CP is determined by the ASTM D 2500 standard test method [38]. The advantage of the CP is that it gives a conservative estimate of the temperature limit where solid formation becomes significant. The CFPP is a filterability test that gives a better reflection of the temperature at which operational problems will occur during actual engine operation. The CFPP will in general be lower than the CP, since it will be possible for small wax crystals to pass through the fuel filter without blinding or blocking it. The cold-flow behavior can be modified by additives (Section 15.4), but ultimately the diesel fuel must be refined to produce molecules with acceptable cold-flow properties. The cold-flow properties are related to the freezing points of the compounds in the mixture and to the mutual solubility of the compounds in each other. Freezing point generally increases with molecular mass, and the heavier fraction of the diesel fuel is the more likely fraction to determine cold flow. The compound classes that have the worst cold-flow properties are the n-alkanes and unsubstituted aromatics. Much of the discussion on jet fuel freezing point (Section 14.3.3) applies. Although the cold-flow criterion is less strict, the material is heavier. HTFT syncrude has a small heavy-end fraction and, despite being called a waxy oil, it is rich in mononuclear alkyl aromatic compounds (Table 15.4). Ideally, one would like to hydroisomerize the alkyl groups of such molecules without causing dealkylation. The aliphatic compounds include both n-alkanes and branched alkanes. The cold-flow properties of the aliphatic compounds can also be improved by hydroisomerization (Chapter 18). LTFT syncrude contains a large fraction of n-alkane waxes that are solids at ambient conditions. The straight-run distillate fraction is also rich in n-alkanes and, except for the light distillate fraction, it requires hydroisomerization to be fluid. It has been reported that poor cold flow is the largest development obstacle to direct blending of straight-run distillate [39]. Direct application of straight-run material avoids yield losses associated with further hydroprocessing, and it retains the beneficial syncrude properties, such as lubricity. When most of the distillate is produced by LTFT wax hydrocracking (Chapter 21), the cold-flow properties are determined by the isomerizing nature of the hydrocracking catalyst.
293
294
15 Diesel Fuel
15.3.8 Stability
The stability of diesel fuel is measured in terms of its oxidation, storage, and thermal stabilities. Oxidation stability is expressed in terms of the insoluble material formed during oxidation as described by the ASTM D 2274 standard test method [40]. It is stated that this method is not applicable to fuels containing a significant amount of material derived from nonpetroleum sources. Storage stability requires a longer test and is performed in accordance with the ASTM D 4625 standard test method [41]. Thermal stability can be determined using the ASTM D 6468 standard test method [42]. Diesel fuel with a high CN inherently has a high autoxidation propensity, which undermines its storage stability. The situation is more precarious when thermal stability is considered. CN improvers are thermally labile compounds and will, by definition, decompose at elevated temperatures. Hydroperoxides which are formed during autoxidation are such compounds and are indeed employed as cetane improvers. Stability is therefore an issue for very high CN Fischer–Tropsch-derived fuels that are readily autoxidized. In fact, stability has been pointed out as one of the greatest concerns associated with diesel fuel blends containing Fischer–Tropsch material [43]. The way in which the stability is measured is important though. Deposit formation subsequent to oxidation requires three sequential steps: initiation by autoxidation, propagation of the free radicals, and termination in such a way that the products will form deposits on filtration. There is also a time factor involved. The autoxidation sequence may be interrupted at any point to improve stability, but by doing so the CN is undermined. However, the duration of the test sets a limit on the amount of inhibition that is required. The volume of oxidation inhibitor that is added (or naturally present) in the fuel is therefore also of consequence. During the oxidative process, the oxidation inhibitor is consumed and, once consumed, autoxidation can readily take place. Adequate oxidation stability of diesel fuel is therefore a trade-off between the duration of stability required and the extent of CN reduction that can be endured. Stability is not exclusively related to the oxygenate content of a fuel, but oxygenates and oxygenate chemistry play a dominant role. Straight-run Fischer–Trosch syncrude contains oxygenates and, even in conventional crude-oil-derived diesel fuels, oxygenates are employed as lubricity improvers and CN improvers. Balancing oxygenate content and stability is important and necessary to produce a good diesel fuel. Very paraffinic Fischer–Tropsch distillates, as is obtained from LTFT syncrude (not HTFT syncrude), has the added disadvantage of decreased dissolving power. The alkanes are poor solvents for heavier oxidation products, which increases precipitation of insoluble matter. 15.3.9 Elastomer Compatibility
Elastomer compatibility issues have already been described for jet fuels (Section 14.3.8). The difference in seal swelling that is caused by changing from an aromatic diesel fuel to a more paraffinic diesel fuel has been evaluated for Fischer–Tropsch distillates [44]. LTFT distillate has a much lower aromatic content than HTFT distillate (Table 15.4) and blending of HTFT and LTFT
15.4 Diesel Fuel Additives That Affect Refinery Design Elastomer compatibility of standard nitrile butadiene rubber when changing from a US D-2 diesel fuel that was used as reference standard to other diesel fuels.
Table 15.5
Property
Diesel fuel US D-2
Fuel properties Density (kg·m−3 ) Cetane number Viscosity at 40 ◦ C (cSt) Flash point (◦ C) Cold filter plugging point (◦ C) Total aromatics (mass%) Polynuclear aromatics (mass%) Sulfur content (µg·g−1 ) Elastomer property (% change) Mass Thickness Hardness
858 41 2.6 39 −14 34.5 5.9 400 0c 0c 0c
UK EN590
HTFTa
LTFTb
833 56 3.4 91 −24 18 4.2 11
824 52 2.3 71 −3 30.8 2.6 2
765 72 2.0 59 −19 0.14 0 1
−9.4 −4.9 6.1
−7.7 −3.3 5.2
−12.3 −4.9 8.5
a HTFT
diesel fuel from Sasol Synfuels, which contains some hydrogenated coal liquids. LTFT distillate prepared by Chevron at their Richmond pilot plant facilities. c Reference fuel used to evaluate elastomer compatibility, by definition zero. b Hydrocracked
distillate is advantageous [27]. However, the aromatic content is not the only aspect influencing the changes in elastomer mass, dimension, hardness, and tensile strength. The polar material in the diesel fuel also influences elastomer properties. This was especially apparent from the elastomer compatibility comparison between South African HTFT-derived diesel fuel, United Kingdom EN590-compliant diesel fuel, and United States D-2 diesel fuel (Table 15.5) [44]. Although the HTFT and US D-2 diesel fuels were closer to each other in terms of aromatics content, the elastomer compatibility behavior of HTFT and the lower aromatic EN590 diesel fuel was better matched. Elastomer compatibility is of concern only in countries where mixed fuel availability enables fuel users to change from one fuel type to another. In this respect, LTFT-derived distillates benefit from blending with crude oil or HTFT distillates to reduce the risk of incompatibility.
15.4 Diesel Fuel Additives That Affect Refinery Design
Additives are compounds that are added to the diesel fuel base stock during final blending in order to produce on-specification diesel fuels [2]. When additives can be employed to meet fuel specifications that would otherwise require more refining effort, it is usually cheaper to do so rather than to increase the refining intensity. In such cases, the refinery design is directly affected by the ability to improve product quality by the additives. Such diesel fuel additives effectively
295
296
15 Diesel Fuel
reduce the specifications that must be met by the refinery design, and for design purposes it is equivalent to relaxing some of the fuel specifications. The design decision to rely on additives to address specific limitations must be consciously made. Once it is decided to employ additives to avoid the associated refining cost to upgrade the fuel to the required specification limit, blending becomes dependent on the additive to meet specification. It will then no longer be possible for the refinery design to meet the fuel specifications without the additive. There are mainly three diesel fuel additive classes that can make refining less onerous: 1) CN improvers, also known as ignition improvers, are able to increase the CN of the diesel fuel. These additives include compounds such as alkyl nitrates (e.g., 2-ethylhexyl nitrate) and peroxides. The CN that can be achieved with a fixed volume addition of the CN improver ironically increases as the quality of the base diesel fuel improves. A diesel fuel with a CN in the range 40–50 will typically register a CN increase of 3–4 when dosing 0.05 vol% of the CN improver as additive. Because the CN improver reduces the actual ignition delay, the improvement will be reflected by the ASTM D 613 standard test method [9], as well as by the ASTM D 6890 (IQT-based) standard test method [10]. However, it will not affect the calculated cetane index measurement, because the additive volume is too small to substantially affect the density and distillation profile of the diesel fuel. 2) Lubricity additives are surface-active compounds and are required only in small volumes to provide sufficient boundary layer lubrication (Section 15.3.4). It compensates for the effects of severe hydroprocessing, which destroys the surface-active compounds. In a crude oil refinery where the sulfur specification necessitates severe hydrotreating, reliance on a lubricity additive is inevitable. The same is not true in a Fischer–Tropsch refinery where it is not necessary to ensure hydrodeoxygenation to the same level. 3) Cold-flow improvers affect the way in which the n-alkanes in the diesel fuel crystallize at low temperatures. The CP can be lowered only by increasing the solubility of the n-alkanes in the fuel. CP depressants can be added and a lowering of around 3 ◦ C can be achieved at a dosage level of 500 µg·g−1 of an appropriate depressant [2]. Since the CP is less important than the CFPP as a measure of engine operability at low temperatures, flow improvers that lower the CFPP are more important additives for diesel fuel. Flow improvers do not affect the CP, but modify the crystallization process and the nature of the wax crystals that are formed. By forming more, but smaller wax crystals that do not agglomerate, the crystals can still pass through the fuel filter and do not affect fuel flow to the engine. The use of cold-flow improvers as additives allow the n-alkane content of the diesel fuel to be higher and thus reduce the level of refining that is required in order for the cold-flow requirements to be met. This also results in a CN benefit, because less n-alkanes have to be isomerized to branched alkanes. Cold-flow improvers are therefore indirectly also CN improvers. 15.5 Future Diesel Fuel Specification Changes
The most significant change in diesel specifications by far was the drastic reduction in sulfur content. At a maximum sulfur content of 10 µg·g−1 , refineries are approaching the limit of sulfur removal. Beyond that it becomes difficult to reduce sulfur when considering practical limitations, such as hydrodynamic effects in fixed bed reactors.
References
Many of the other diesel specification changes have been incremental. The CN is nudging up, the density and viscosity ranges are narrowing, and the heavy end of the distillation profile is becoming lighter. These changes are significant, but not disruptive. Indirectly, these changes reduced the polynuclear aromatic content of diesel fuel. Whether more stringent limitations of the polynuclear aromatic content of diesel fuel will serve a purpose is doubtful, unless a very low limit is set. Deeper hydrogenation may not be economical even at high crude oil prices and in practice refiners may opt to reduce the cut point of diesel fuel rather than to increase hydrogenation severity. The only specification that is of concern to Fischer–Tropsch refiners is the lower specification limit on diesel fuel density. There is a density–cetane–yield triangle that limits on-specification diesel fuel production in high yield from Fischer–Tropsch syncrude [8]. This is discussed in detail in Chapter 27, which deals with diesel fuel refining. The density deficiency can be overcome by blending with heavier distillates, but it is an issue for stand-alone refineries where such blending is limited or logistically challenging. There is a definite trend, spurred by environmental and political reasoning, to include renewable material in diesel fuel. Much work has been done on esterification of plant and animal fats and oils as blending components in diesel. Diesel fuel specifications already make provision for limited inclusion of fatty acid methyl esters (FAMEs). However, it has been reported that blends of LTFT distillate, FAME, and petroleum have storage stability issues [45]. Future legislation concerning the inclusion of bio-derived oxygenates in diesel fuel is likely to be a subject of political expediency. Fischer–Tropsch refiners may have the option of including bio-derived material in the raw material used as feed for synthesis gas production. Legislation to ensure the inclusion of a minimum amount of renewable material in diesel fuel may therefore be met either by using the biomass in syncrude production or by co-refining the bio-derived products in the refinery. As such, co-refining of biomass can more easily be accommodated in a Fischer–Tropsch refinery than a crude oil refinery, since a Fischer–Tropsch refinery has to process oxygenates in any case. This may resolve some of the blending issues related to oxygenated products in diesel fuel.
References 1. Totten, G.E. (ed.) (2003) Fuels and Lubricants
2.
3.
4.
5.
Handbook: Technology, Properties, Performance, and Testing, ASTM, West Conshohocken, PA. Dabelstein, W., Reglitzky, A., Sch¨utze, A., and Reders, K. (2008) in Handbook of Fuels (ed. B. Elvers), Wiley-VCH Verlag GmbH, Weinheim, pp. 97–195. (2006) World-Wide Fuel Charter, 4th edn, Alliance of Automobile Manufacturers, Washington. ASTM (2005) D 975 – 05. Standard specification for diesel fuel oils, ASTM, West Conshohocken, PA. Bisio, A. and Atkinson, C. (2002) Shouldn’t we know the molecular composition of
6.
7.
8.
9.
Fischer-Tropsch diesel fuels. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 47 (2), 496–497. Cookson, D.J., Lloyd, C.P., and Smith, B.E. (1988) Investigation of the chemical basis of diesel fuel properties. Energy Fuels, 2, 854–860. Androulakis, I.P., Weisel, M.D., Hsu, C.S., Qian, K., Green, L.A., Farrell, J.T., and Nakakita, K. (2005) An integrated approach for creating model diesel fuels. Energy Fuels, 19, 111–119. De Klerk, A. (2009) Can Fischer-Tropsch syncrude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. ASTM (2005) D 613 – 05. Standard test method for cetane number of diesel fuel oil, ASTM, West Conshohocken, PA.
297
298
15 Diesel Fuel 10. ASTM (2004) D 6890 – 04. Standard test method
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
determination of ignition delay and Derived Cetane Number (DCN) of diesel fuel oils by combustion in a constant volume chamber, ASTM, West Conshohocken, PA. Morris, W. (2007) Method relates diesel cetane, octane ratings. Oil Gas J., 105 (45), 58–60. Santana, R.C., Do, P.T., Santikunaporn, M., Alvarez, W.E., Taylor, J.D., Sughrue, E.L., and Resasco, D.E. (2006) Evaluation of different reaction strategies for the improvement of cetane number in diesel fuels. Fuel, 85, 643–656. Ghosh, P. and Jaffe, S.B. (2006) Detailed composition based model for predicting the Cetane number of diesel fuels. Ind. Eng. Chem. Res., 45, 346–351. Bailey, B.K., Russell, J.A., Wimer, W.W., and Buckingham, J.P. (1986) Cetane number prediction from proton-type distribution and relative hydrogen population. SAE Tech. Pap. Ser., 861521. O’Connor, C.T., Forrester, R.D., and Scurrell, M.S. (1992) Cetane number determination of synthetic diesel fuels. Fuel, 71, 1323–1327. Ladommatos, N. and Goacher, J. (1995) Equations for predicting the cetane number of diesel fuels from their physical properties. Fuel, 74, 1083–1093. Yang, H., Ring, Z., Briker, Y., McLean, N., Friesen, W., and Fairbridge, C. (2002) Neural network prediction of cetane number and density of diesel fuel from its chemical composition determined by LC and GC-MS. Fuel, 81, 65–74. Hardenberg, H.O. (1980) Thoughts on an ideal diesel fuel from coal. S. Afr. Mech. Eng., 30, 34–47. Hashimoto, K., Ikeda, M., Arai, M., and Tamura, M. (1996) Cetane number improvement of diesel fuel by autoxidation. Energy Fuels, 10, 1147–1149. Knottenbelt, C. (2002) Mossgas ‘‘gas-to-liquids’’ diesel fuels – an environmentally friendly option. Catal. Today, 71, 437–445. Pecci, G.C., Clerici, M.G., Giavazzi, F., Ancilotti, F., Marchionna, M., and Patrini, R. (1991) Oxygenated diesel fuels. Part 1 – structure and properties correlation. 9th International Symposium on Alcohols Fuels, pp. 321–326. Olah, G.A., Goeppert, A., and Prakash, G.K.S. (2006) Beyond Oil and Gas: The Methanol Economy, Wiley-VCH Verlag GmbH, Weinheim. Van Heerden, J., Botha, J.J., and Roets, P.N.J. (1998) Improvement of diesel performance with the addition of linear ethers to diesel fuels. 12th
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
International Symposium on Alcohols Fuels, pp. 188–199. Wu, T., Huang, Z., Zhang, W.-G., Fang, J.-H., and Yin, Q. (2007) Physical and chemical properties of GTL-diesel fuel blends and their effects on performance and emissions of a multicylinder DI compression ignition engine. Energy Fuels, 21, 1908–1914. Anastopoulos, G., Lois, E., Karonis, D., Zanikos, F., and Kalligeros, S. (2001) A preliminary evaluation of esters of monocarboxylic fatty acid on the lubrication properties of diesel fuel. Ind. Eng. Chem. Res., 40, 452–456. ASTM (2005) D 6079 – 04. Standard Test Method for Evaluating Lubricity of Diesel Fuels by the High-frequency Reciprocating Rig (HFRR), ASTM, West Conshohocken, PA. Lamprecht, D., Dancuart, L.P., and Harrilall, K. (2007) Performance synergies between low-temperature and high-temperature Fischer-Tropsch diesel blends. Energy Fuels, 21, 2846–2852. Knothe, G. and Steidley, K.R. (2005) Lubricity of components of biodiesel and petrodiesel. The origin of biodiesel lubricity. Energy Fuels, 19, 1192–1200. Anastopoulos, G., Lois, E., Serdari, A., Zanikos, F., Stournas, S., and Kalligeros, S. (2001) Lubrication properties of low-sulfur diesel fuels in the presence of specific types of fatty acid derivatives. Energy Fuels, 15, 106–112. Hughes, J.M., Mushrush, G.W., and Hardy, D.R. (2002) Lubricity-enhancing properties of soy oil when used as a blending stock for middle distillate fuels. Ind. Eng. Chem. Res., 41, 1386–1388. Anastopoulos, G., Lois, E., Zannikos, F., Kalligeros, S., and Teas, C. (2002) The tribological behavior of alkyl ethers and alcohols in low sulfur automotive diesel. Fuel, 81, 1017–1024. Geller, D.P. and Goodrum, J.W. (2004) Effects of specific fatty acid methyl esters on diesel fuel lubricity. Fuel, 83, 2351–2356. Hu, J., Du, Z., Li, C., and Min, E. (2005) Study on the lubrication properties of biodiesel as fuel lubricity enhancers. Fuel, 84, 1601–1606. Wain, K.S. and Perez, J.M. (2002) Oxidation of biodiesel fuels for improved lubricity. Proc. ICE Spring Tech. Conf., 38, 27–34. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Schaberg, P.W., Myburgh, I.S., Botha, J.J., Roets, P.N.J., and Dancuart, L.P. Diesel engine
References
37.
38.
39.
40.
emissions with Sasol Slurry Phase Distillate fuel. 11th World Clean Air and Environment Congress, 14–18 September 1998, Durban, South Africa, pp. 6F–63. Leckel, D.O. (2009) Hydroprocessing CTL diesel from high-temperature Fischer-Tropsch syncrude and pyrolysis tar oil. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 125–126. ASTM (2005) D 2500 – 05. Standard test method for cloud point of petroleum products, ASTM, West Conshohocken, PA. Suppes, G.J., Terry, J.G., Burkhart, M.L., and Cupps, M.P. (1998) Compression-ignition fuel properties of Fischer-Tropsch syncrude. Ind. Eng. Chem. Res., 37, 2029–2038. ASTM (2003) D 2274 – 03a. Standard test method for oxidation stability of distillate fuel oil (accelerated method), ASTM, West Conshohocken, PA.
41. ASTM (2004) D 4625 – 04. Standard test method
42.
43.
44.
45.
for middle distillate fuel storage stability at 43 ◦ C (110 ◦ F), ASTM, West Conshohocken, PA. ASTM (2008) D 6468 – 08. Standard test method for high temperature stability of middle distillate fuels, ASTM, West Conshohocken, PA. O’Rear, D.J., Bacha, J.D., and Tiedemann, A.N. (2004) Thermally stable blends of Fischer-Tropsch and LCO diesel fuel components. Energy Fuels, 18, 682–684. Lamprecht, D. (2007) Elastomer compatibility of blends of biodiesel and Fischer-Tropsch diesel. SAE Tech. Pap. Ser., 2007-01-0029. Mushrush, G.W., Willauer, H.D., Bauserman, J.W., and Williams, F.W. (2009) Incompatibility of Fischer-Tropsch diesel with petroleum and soybean biodiesel blends. Ind. Eng. Chem. Res., 48, 7364–7367.
299
301
Part V Refining Technology
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
303
16 Refining Technology Selection 16.1 Introduction
The refinery designs associated with industrial Fischer–Tropsch facilities (Chapters 6–12) highlighted the need to look at the molecules in syncrude and determine the refining strategy accordingly. In instances where a crude oil refining approach was employed in the design of a Fischer–Tropsch refinery, the designs were inefficient. Inattention to detail in the selection of catalysts and/or refining technologies also degraded performance. Two important change drivers affected successive Fischer–Tropsch refinery designs, as well as modifications to operational refineries. Foremost were the changes in transportation fuel specifications (Chapters 13–15). As the requirements for producing on-specification motor-gasoline, jet fuel, and diesel fuel changed, refining strategies had to be adapted to deliver products of acceptable composition and quality. The second change driver was refining of chemicals. The extraction of chemicals as higher valued products than fuels became a common thread in many of the industrial Fischer–Tropsch facilities. Central to refinery design and the ability to produce the desired products (fuels and/or chemicals) is the selection of appropriate refining technology. By doing so, Fischer–Tropsch refining can be more efficient than crude oil refining [1]. Some guidelines can be derived from the preceding chapters: 1) Carbon-number-based refining works better for Fischer–Tropsch syncrude than refining based on boiling range. This has its origin in the molecular properties of the syncrude, where specific carbon number ranges are compatible with specific refining technologies. Carbon-number-based refining [2], managing the molecules [3], or technology selection to match molecules [4] are all different ways of saying that one has to refine syncrude based on molecular properties and not on the boiling range. 2) Boiling range broadening results from deoxygenation of syncrude. This is one of the main detractors from boiling-range-based refining. 3) Catalyst selection for the refining of Fischer–Tropsch syncrude [5] is as important as selecting the conversion technology. The reactive nature of Fischer–Tropsch syncrude, due to its significant alkene and oxygenate content, requires milder catalysts. Four catalyst classes have been identified as being especially important in Fischer–Tropsch refining [6]: alumina, solid phosphoric acid (SPA), nonacidic Pt/L-zeolite, and mildly acidic Pt/SiO2 –Al2 O3 . Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
304
16 Refining Technology Selection
These are not the only catalysts that can be used, but represent the catalyst types that, in conjunction with appropriate hydrotreating catalysts, allow the design of an efficient stand-alone Fischer–Tropsch refinery capable of producing on-specification transportation fuels (Figure 16.1). 4) The cost of a Fischer–Tropsch refinery is typically less than 15% of the total capital cost of a Fischer–Tropsch-based facility. The capital needed to produce the syncrude exceeds the cost of refining by far. Carbon-efficient refining consequently makes economic and environmental sense and it requires the inclusion of conversion technologies for the aqueous product and C1 –C4 material in the refinery design. 5) Learn the lessons from the past [7, 8]. Fischer–Tropsch refiners have a much smaller experience base to draw on than crude oil refiners. It is important not to repeat the mistakes previously made. 6) By its nature, Fischer–Tropsch syncrude contains molecules that have value as chemicals [9–13]. Through proper refining, there are also opportunities to synthetically produce C2 Dehydration (alumina)
Hydrotreater
C3 +
Aqueous product
Syngas
Fischer –Tropsch gas loop
C2 C3
C4
Light oil product
Alkene hydration (SPA) Aromatic alkylation (SPA)
Motor-gasoline (ethanol) Motor-gasoline Jet fuel
Oligomerization (SPA)
Hydrotreater
Motor-gasoline Jet fuel
Hydroisomerization (Pt /SiO2 –Al2O3)
Motor-gasoline
Benzene
C5 Hydrotreater
C6 –C8
C9 +
Naphtha reforming (nonacidic Pt /L)
Motor-gasoline
Hydroisomerization (Pt/SiO2 –Al2O3)
Jet fuel Diesel fuel
Figure 16.1 Stand-alone high-temperature Fischer–Tropsch (HTFT) refinery design, illustrating the use of the four key catalyst types needed to produce on-specification transportation fuels.
16.2 Hydrotreating
chemicals. By selecting appropriate refining technologies, products can be produced in an efficient manner that can be applied as fuels or chemicals. Refinery technology selection is discussed in terms of the conversion processes employed. This serves as an introduction to the chapters that describe selected conversion processes important to high-temperature Fischer–Tropsch (HTFT) and low-temperature Fischer–Tropsch (LTFT) refining in more detail (Chapters 17–23). The conversion processes that have been selected for detailed discussion are those that provide significant opportunity to Fischer–Tropsch refining. Detailed accounts of catalyst and technology selection for the refining of Fischer–Tropsch syncrude can be found in the literature [4, 5]. The conversion processes typically associated with crude oil refining have been briefly discussed in Chapter 2. If one were to include commodity petrochemical refining, the number of conversion processes grows somewhat. The list of conversion processes that are potentially relevant to Fischer–Tropsch refining is more extensive. There are two reasons for this. The first is the abundance of alkenes and oxygenates, which allows refining pathways not equally accessible in a crude oil refinery. The second is the chemical nature of the syncrude, which invites the use of technologies not normally associated with mainstream crude oil refining. There is unfortunately not an obvious classification system for conversion processes, and treatises on refining tend to discuss them in a sequential manner [14–18]. In order to give some structure to the discussion on the application of different conversion processes to Fischer–Tropsch syncrude, the processes were organized into five categories, despite the inherent overlap between some: hydrotreating, addition and removal of oxygen, alkene conversion, alkane conversion, and residue conversion processes.
16.2 Hydrotreating
Hydrotreating is a very basic upgrading step and involves the use of hydrogen to remove heteroatoms, increase the H:C ratio of the feed material, and partially convert functional groups by selective hydrogen addition. The reactor technology and catalyst/s are selected in line with the intended application [5, 19, 20]. In a refinery, hydrotreating is employed as a feed pretreatment step, or as a product polishing step. When it is used for feed pretreatment, the objective is to remove contaminants in the feed that will adversely affect downstream processes, or to transform material in such a way that it will be amenable to downstream processing. When it is used as a product polishing step, the objective is to adjust the quality of the final product in line with product specifications. It is customary to refer to the nature of the hydrotreating action as if it is a separate process, although these processes occur in parallel: 1) Hydrogenation (HYD) of alkenes. This is a very important hydrotreating function in Fischer–Tropsch refining (Section 16.2.1). 2) Hydrodearomatization (HDA) [21]. The polynuclear aromatic content of LTFT and HTFT material is well below diesel fuel specifications (Table 15.4), and deep HDA to remove such compounds is not important in Fischer–Tropsch refining. The HDA of benzene is one way of dealing with refinery benzene, but is not preferred when alternatives exist [22, 23].
305
306
16 Refining Technology Selection
3) Hydrodemetallization (HDM) [24, 25]. The metals and their levels present in syncrudes depend mainly on the Fischer–Tropsch synthesis (Section 4.5). Metals may also be present as corrosion products or produced in conversion processes. The metals are generally present as metal carboxylates, and HDM catalysts are ineffective in their removal [26]. 4) Hydrodesulfurization (HDS) [27, 28]. This is the most important hydrotreating function in crude oil refineries, but Fischer–Tropsch syncrude is sulfur-free. 5) Hydrodenitrogenation (HDN) [29–31]. The importance of HDN in a crude oil refinery depends on the nitrogen content of the crude oils being processed. Fischer–Tropsch syncrude contains no nitrogen compounds. 6) Hydrodeoxygenation (HDO). Complete oxygen removal and the selective partial HYD of oxygenates are important hydrotreating functions in Fischer–Tropsch refining (Section 16.2.2). 16.2.1 Hydrogenation of Alkenes
The oxygenate content and nature of the oxygenates in the feed material that must be hydrogenated will influence the HYD catalyst selection. Carboxylic acids are aggressive to reduced base metal catalysts [32]. The carboxylic acids leach the active metal when the operating temperature is below the metal carboxylate decomposition temperature. Noble metals have a higher leaching resistance, but are far more active and not practical for bulk alkene HYD. Noble metal HYD catalysts are useful for selective partial HYD applications. When carboxylic acids are present in a stream that requires bulk alkene HYD, upstream acid removal should be considered. This is not always practical, and in cases where the hydrotreating involves alkenes and oxygenates, sulfided base metal catalysts can be considered (Section 16.2.2). Applications that involve alkene HYD in a Fischer–Tropsch refinery are the following: 1) Partial HYD of alkenes in order to avoid downstream processing problems. When the feed material contains dienes, the removal of the dienes by selective HYD is often required to avoid the formation of gums (heavy products from diene oligomerization). Diene removal can be combined with double bond isomerization to increase the octane number if the product becomes a blending component for motor-gasoline. Some metals are efficient for partial HYD and double bond isomerization [33]. The reported order of isomerization activity of reduced metals is Pd > Ni > Rh, Ru > Os, Ir, Pt [34]. When double bond isomerization takes place with partial HYD, it is not only beneficial for downstream processing but also serves as an upgrading step (Section 16.4.1). 2) Partial HYD of alkenes in order to comply with final product requirements. Motor-gasoline specifications may limit the amount of alkenes that can be included in the final product (Section 13.2). Because of the alkene-rich nature of syncrude, this may require the HYD of some alkenes. Ideally, one would like to selectively hydrogenate the alkene isomers that will result in the least octane number loss. Unfortunately, the most branched alkenes which have the highest hydrogenated octane number are also the most difficult to hydrogenate. In single carbon number cuts, it is best to fully hydrogenate with a bypass, rather than partially hydrogenate the total stream (Figure 16.2) [35]. The same strategy does not hold true for mixed carbon number streams, because lighter alkenes have higher hydrogenated octane numbers than isostructural heavier alkenes. One may also exploit unsulfided and sulfided HYD to increase catalyst lifetime and improve selectivity [35, 36].
16.2 Hydrotreating
100
1-Octene
Conversion (%)
80
60 2,4,4-Trimethyl pentenes
Partial (90%) hydrogenation
RON = 42.6
Complete hydrogenation
RON = 44.5
50 : 50 feed mixture
40
10% bypass 20 20
40
60
80
100
Temperature (°C) Figure 16.2 Single carbon number alkene hydrogenation configurations illustrated by the hydrogenation of an equimolar mixture of 1-octene and mixed 2,4,4-trimethylpentenes over a Pd/C catalyst.
3)
Complete HYD of alkenes in order to avoid downstream catalyst deactivation. This is typically a feed pretreatment step that is considered before a process that employs a bifunctional catalyst, with both metal and acid functionality. This is an easy way to ensure downstream stability, but it does not exploit the syncrude properties to its full potential. In some cases, the heat of alkene HYD can be exploited while controlling the alkene partial pressure and avoiding catalyst deactivation by thoughtful design [37]. 4) Complete HYD of alkenes to comply with final product requirements. HYD of Fischer– Tropsch waxes and paraffinic solvents are typical applications [38]. It is also required in the production of iso-paraffinic kerosene (IPK) for synthetic jet fuel production (Section 14.2.1). 16.2.2 Hydrodeoxygenation
HDO catalysis has been reviewed by Furimsky [39, 40]. As in the case of alkene HYD, catalyst selection is constrained by the presence of carboxylic acids. The heavier carboxylic acids are not as aggressive as the short-chain carboxylic acids, and accounts of unsulfided base metal–containing catalysts that were used for carboxylic acid hydrotreatment can be found in the literature [41–43]. Nevertheless, the use of unsulfided base metal catalysts with carboxylic acid–containing feed runs the risk of catalyst deactivation by leaching. Noble metal catalysts are typically not considered for bulk HDO and HYD of syncrude because noble metals are too active. Catalysts that are too active make heat management difficult and can easily result in excessive temperature excursions. Sulfided base metal catalysts are quite resistant to acid leaching and have been used for bulk HDO and HYD of Fischer–Tropsch syncrude [26, 44, 45]. The main disadvantage of sulfided
307
308
16 Refining Technology Selection
catalysts for syncrude HYD is that it requires the addition of sulfur in the form of a sulfiding agent to an otherwise sulfur-free feed. Bulk hydrotreating of syncrude find application in the following: 1) Feed pretreatment by partial oxygenate conversion. The most common Fischer–Tropsch application is carbonyl to alcohol HYD. It can be used to simplify aqueous product refining by converting the aldehydes and ketones into alcohols (for example, Section 10.4.2). It is also extensively used in the production of chemicals from syncrude. 2) Feed pretreatment by HDO for downstream refinery processes that are sensitive to oxygenates. A typical application is hydrotreating before catalytic naphtha reforming. One may also consider hydrotreating as a pretreatment step before hydrocracking. It has been shown that oxygenates influence hydrocracking selectivity [46], and the yield advantage of hydrotreating before mild hydrocracking to produce diesel fuel from HTFT syncrude has been demonstrated [47]. 3) HYD to comply with product requirements. Such HYD is typically performed on straight-run syncrude that will be included in final products. For example, distillate hydrotreating to produce diesel fuel from HTFT syncrude. Controlling the level of HYD is quite important, since properties like diesel fuel lubricity can be retained by proper catalyst selection and HDO control [5].
16.3 Addition and Removal of Oxygen 16.3.1 Dehydration
Alcohols are the dominant oxygenate class in LTFT syncrude and also one of the main oxygenate classes in HTFT syncrude. In order to simplify refining, carbonyl compounds can be selectively hydrogenated to alcohols, thereby increasing the alcohol concentration in Fischer–Tropsch syncrude even further. Alcohol dehydration is an important syncrude refining technology and is discussed in detail later (Chapter 17). Dehydration takes place over an acidic catalyst. Catalyst selection and operating conditions determine the extent of dehydration. Mild dehydration leads to etherification (Section 16.3.2), whereas more severe dehydration produces alkenes. It is an equilibrium-limited reaction, and the reverse reaction, namely, alkene hydration (Section 16.3.3) to produce alcohols is also possible. A number of dehydration applications have been noted: 1)
Complete dehydration of alcohols from the Fischer–Tropsch aqueous product has been suggested as a way to simplify aqueous product refining [48]. By converting the oxygenates via alcohols into alkenes, the alkenes can be co-refined with the rest of the alkenes in the syncrude. 2) Dehydration of specific n-1-alcohols to produce n-1-alkenes as chemicals [49]. 3) Low-temperature alkane conversion into alkenes [50]. The alkanes are activated by autoxidation (Section 16.3.5), which is followed by partial HYD to alcohols and subsequent dehydration to alkenes.
16.3 Addition and Removal of Oxygen
4) Partial dehydration of alcohols to produce fuel ethers for both motor-gasoline and diesel fuel. The etherification of specifically the heavier n-1-alcohols produces linear ethers that are good diesel fuel blending components [51].
16.3.2 Etherification
Fuel ethers can be prepared from the reaction of alcohols with alkenes, or by the reaction of alcohols with alcohols by partial dehydration. Since alcohols and alkenes are abundant in syncrude, etherification has a natural feed advantage. In the case of the etherification of alkenes with alcohols, the C=C of the alkene must be on a tertiary carbon. The branched alkenes in syncrude is less than the n-alkenes, and skeletal isomerization (Section 16.4.3) can be employed to increase the degree of branching of the alkenes. Etherification is an important syncrude refining technology and it is discussed in more detail in Chapter 17. Etherification is equilibrium limited and is catalyzed by an acidic catalyst. In crude oil refining, the etherification of alcohols and alkenes found widespread use in motor-gasoline, but encountered complications of a political nature (Section 2.4.4). Nevertheless, fuel ethers are still in use as octane improvers for motor-gasoline. Etherification of alkenes with alcohols has two industrial Fischer–Tropsch applications: 1) The purification of n-1-alkenes as chemicals from HTFT syncrude employs etherification [52]. Methanol reacts with the alkene isomers that are close boiling to 1-pentene and 1-hexene to produce ethers that can be removed by conventional distillation. 2) Conventional etherification of branched alkenes with alcohols is used to produce fuel ethers as octane improvers for motor-gasoline [53]. This is usually performed in conjunction with alkene skeletal isomerization (Section 16.4.3) to improve etherification yield.
16.3.3 Hydration
The hydration of alkenes to alcohols is usually performed to produce alcohols for chemical use. In a Fischer–Tropsch refinery, hydration has one very specific application, namely, the conversion of ethene into a liquid product, ethanol, when an HTFT facility is remote from a petrochemical market. The problem is analogous to that in gas-to-liquid conversion, where the otherwise useful natural gas cannot be exploited as an energy carrier unless it is converted into a liquid product (Section 3.2.1). Alkene hydration is equilibrium limited and the equilibrium does not favor hydration, but dehydration. It is an acid-catalyzed reaction. It fits well within a Fischer–Tropsch refinery, because the hydration technology can benefit from the infrastructure in the aqueous product refinery to process and purify the raw hydration product. Although the application of hydration is very refinery-specific, it is important and is discussed in more detail later (Chapter 17).
309
310
16 Refining Technology Selection
OH+
O +
R ′R
OH OH
Figure 16.3
+H
− H+
R
OH
R
O H ′R
′R
OH+ OH
− H+
O H
+ H+
O R
O
′R
H2O
Acid-catalyzed esterification of a carboxylic acid with an alcohol.
16.3.4 Esterification
Carboxylic acids and alcohols can react with each other to produce esters (Figure 16.3). This reaction is not used in crude oil refining, but it is extensively used in the refining of biomass. The direct transesterification of plant oils and animal fats with methanol produces fatty acid methyl esters (FAME) and 1,2,3-propantriol (glycerol) [54]. These esters can be used as renewable diesel fuel additive and are allowed as blending components in conventional diesel fuels (Table 15.1). Although analogous esters can be produced from syncrude, the concentration of heavy carboxylic acids in syncrude is low. The esterification of carboxylic acids with alcohols is an equilibrium-limited reaction that is catalyzed by acids. (Direct transesterification of bio-derived oils and fats with methanol is acidor base catalyzed.) The potential application of esterification in a Fischer–Tropsch refinery is not to produce fuel blending components per se, but rather as a means to refine carboxylic acids: 1)
The removal of carboxylic acids in syncrude by converting the acids into esters has been investigated [55]. As a feed pretreatment step, the conversion of acids into a neutral oxygenates enables the use of acid-sensitive downstream technology and unsulfided base metal hydrotreating catalysts. 2) Gas-phase esterification of carboxylic acids in the gas loop can potentially reduce the amount of carboxylic acids that are dissolved in the reaction water. Once the carboxylic acids are dissolved and end up in the Fischer–Tropsch aqueous product, further refining and recovery become difficult. The possibility of realizing this in practice is unfortunately low, mainly due to the low equilibrium conversion to esters [56], competitive adsorption, and high risk of side reactions. 3) HDO of carboxylic acids can be facilitated by converting the acids into methyl esters. 16.3.5 Carbonyl Aromatization
Carbonyl aromatization is not a commercial refining technology. Aldehydes and ketones can repeatedly undergo aldol condensation and dehydration. When this process is acid catalyzed, the trimer unit can self-condense to produce an aromatic compound (Figure 16.4). Ethanal yields benzene [57], whereas the C3 and heavier carbonyl compounds yield alkyl aromatics. For example, propanal produces 1,3,5-trimethylbenzene (mesitylene) [58], as does propanone (acetone) [59]. In mixtures of carbonyl compounds, the products will not necessarily be radially symmetric. The aromatization of carbonyl compounds is a potentially useful reaction to increase aromatics production from the Fischer–Tropsch aqueous product [60].
16.3 Addition and Removal of Oxygen
O
O R
+
R
+ H+
OH O
− H+
R
O
O R
+
− H+
R O R
H
− H+
+ H2 O
O
+ H+
R R
R
R
OH O
R R
R
R
+ H+
R
O
+ H+
R
− H+
R R
+ H2O
R
R + H2O
+
R Figure 16.4
Carbonyl aromatization by repeated aldol condensation and dehydration.
16.3.6 Hydroformylation
In a Fischer–Tropsch refinery, all the reagents necessary for hydroformylation are readily available, namely, alkenes, CO, and H2 . The hydroformylation reaction (Equation 16.1) is related to Fischer–Tropsch synthesis, and it involves chain growth by the addition of CO. R–CH=CH2 + CO + H2 → R–CH2 –CH2 –CHO
(16.1)
The reaction is conducted in the liquid phase with a homogeneous catalyst that is based on either Co or on Rh [61]. There is a trade-off involved in selecting the hydroformylating metal. When employing a Co-based catalyst, a high operating pressure (>10 MPa) is required. When an Rh-based catalyst is employed, which is about two orders of magnitude more reactive than Co-based catalysts, a lower pressure (∼1.5 MPa) is needed, but the catalyst is considerably more expensive [62]. Hydroformylation technology is generally too expensive to be considered for fuel refining and its extensive use by Sasol (Section 9.5) is for the production of chemicals. 16.3.7 Autoxidation
The reaction of oxygen in air with hydrocarbons takes place at ambient conditions, but at such conditions it is a slow reaction. This is called autoxidation and this is the reaction that causes fuel degradation during storage (Sections 14.3.7 and 15.3.8). However, autoxidation can also be beneficial and it has found application in the production of chemicals from Fischer–Tropsch waxes (Sections 6.3.3 and 8.5.3). The application of autoxidation for chemical production is discussed in more depth later (Section 23.3). Autoxidation is a free radical process and requires no catalyst, although a catalyst can be used to increase the reaction rate. It is typically conducted at moderate conditions, <180 ◦ C and low pressure [50, 63]. It should not be confused with autoxidation at more severe conditions that is used for asphalt hardening [64]. Although autoxidation is not extensively used in crude oil refining at present, in future it has considerable potential for oxidative desulfurization
311
312
16 Refining Technology Selection
and oxidative denitrogenation [65–70]. In addition to oxidative heteroatom removal crude oil refineries may share some of the syncrude refining applications: 1) Lubricity and cetane number of fuels can be improved by autoxidation. The hydroperoxides are thermally labile and are cetane improvers. The oxygenates that are formed by oxygen incorporation are surface-active and improve boundary layer lubricity. 2) The low-temperature activation of alkanes to produce alkenes has already been mentioned [50]. 3) Chemical applications where the products from syncrude autoxidation are useful oxygenates, such as oxidized waxes. 4) Autoxidation under more severe conditions can be used to oxidatively degrade heavy organic waste streams to produce lighter oxygenates. Such oxygenates are biologically more easy to degrade, thereby reducing the environmental footprint of the process. If anaerobic digestion is employed, some of the carbon may even be recovered as methane, which can be converted into syngas again. Anaerobic digestion to produce methane is industrially employed in conjunction with HTFT refining by PetroSA (Chapter 10). 16.4 Alkene Conversion 16.4.1 Double Bond Isomerization
When highly olefinic motor-gasoline was allowed as commercial fuel, the easiest way to achieve a good quality motor-gasoline from syncrude was by double bond isomerization. The syncrude is rich in n-1-alkenes, which have the lowest octane numbers of all alkenes (Table 13.5). By isomerizing the 1-alkenes to internal alkenes, a large gain in octane number is achieved. This strategy works very well across all carbon numbers, with large gains in octane number that is possible from a very simple conversion process. Double bond isomerization can be catalyzed by acids, bases, free radicals, and metals in the presence of hydrogen (Figure 16.5). Potential catalysts have been listed in the extensive review of Dunning [71]. Newer literature includes studies on catalyst classes such as acidic resins [72], acidic zeolites [73], basics zeolites [74], and mixed oxides [75]. Although double bond isomerization is often regarded as a facile reaction, this is not always the case. There must be sufficient thermodynamic driving force. Isomerization of 1-alkenes to internal alkenes may be a thermodynamically favorable reaction, but the surface intermediate that causes isomerization to take place must also favor isomerization. As the carbon chain length of the alkene becomes longer, steric effects may decrease the driving force for isomerization of the adsorbed species and may even promote isomerization to the terminal position. In the past, double bond isomerization was extensively used in Fischer–Tropsch refining (Chapters 6–9). Bauxite treatment, or clay treatment over alumina-based catalysts, was an effective way to improve motor-gasoline quality. Unfortunately, double bond isomerization no longer has the same appeal as it had in the past. There are three inherent drawbacks to double bond isomerization. The first is that the octane numbers that can be achieved by just double bond isomerization is limited despite the significant
16.4 Alkene Conversion + H+
(a) R
(b) R
−H
+
− H+ + H+
+
R
R
− H+ + H+
−
(d) R
+ H+ − H+
R
− H•
R
− H•
R
+ H–Ni
R
− H–Ni
−
R
+ H•
(c) R
R
R
H
H
− H –Ni
R
Ni Ni Ni
+ H•
Ni Ni Ni
+ H –Ni
R
R
Figure 16.5 Double bond isomerization catalyzed by (a) acids, (b) bases, (c) free radicals, and (d) metals in the presence of hydrogen.
gain in octane number. The cost per octane number gain is very low, making it an attractive conversion process, but the ultimate octane number that can be achieved makes it impractical for C7 and heavier alkenes. The second drawback is related to the nature of the conversion. Double bond isomerization upgrades only the alkenes, and the gain is proportionally less as the alkane content of the syncrude increases. It also produces an olefinic product, and the octane number gain will be completely lost if the alkene is hydrogenated; no structural change takes place and n-1-alkenes are still linear after double bond isomerization. The third drawback is related to the blending value of the material for motor-gasoline. Motor-gasoline specifications no longer allow a high alkene content or the addition of tetraethyl lead, and double bond isomerization produces a highly olefinic material of only moderate quality compared to the octane number requirements of final on-specification motor-gasoline. For the reasons mentioned, the prospect for future application of double bond isomerization in a Fischer–Tropsch refinery is slim, despite the fact that it is a proven technology for syncrude [4]. Instances where it may still find applications are as follows: 1)
Upgrading of a straight-run syncrude stream that is earmarked for direct blending into motor-gasoline. Since such a feed will in any case be blended directly, any improvement in octane number is valuable and the drawbacks mentioned do not apply. 2) Pretreatment of feed before a process that is sensitive to the double bond position. For example, 3-methyl-1-butene is not reactive for etherification, but the double bond isomers 2-methyl-1-butene and 2-methyl-2-butene are. Some gain can also be found in isomerizing 1-butene to 2-butenes before aliphatic alkylation using HF as catalyst (Section 16.4.5). 3) Conversion in combination with metathesis (Section 16.4.2) to shift the carbon number distribution by disproportionation of two carbon chain lengths that are presenting refining difficulties. 4) Enabling separation of compounds for chemical use. For example, purification of 1-hexene from Fischer–Tropsch syncrude by distillation requires the reactive removal of close boiling
313
314
16 Refining Technology Selection
isomers. Double bond isomerization is often sufficient to change the boiling point to enable separation by distillation. 16.4.2 Metathesis
Alkene disproportionation or metathesis (Figure 16.6) is employed in chemical production to produce propene from ethene and 2-butene, different 1-alkenes, and specialized polymer products [76]. It is of potential interest in a Fischer–Tropsch refinery, because many of the commercial metathesis technologies employ ethene for disproportionation. It presents an alternative refining pathway for stand-alone Fischer–Tropsch facilities far from petrochemical consumers and, like hydration (Section 16.3.3), is able to convert ethene into heavier transportable products. Metathesis is also capable of changing the carbon number distribution around its mean while retaining the same average molecular mass in the product as in the feed. The most commonly used heterogeneous catalysts for metathesis are Re2 O7 (Meta-4 process of IFP/Axens), MoO3 (SHOP, Shell Higher Olefins Process), and WO3 (OCT, Olefins Conversion Technology of ABB Lummus). These catalyst systems require regular regeneration. Some applications of metathesis in a Fischer–Tropsch refinery that can be envisioned are given below: 1)
Light and heavy fractions that do not have convenient refining pathways can be disproportionated to a product of intermediate carbon number. In this way, the carbon number distribution can be shifted to a distribution that is more useful for refining. Because of the high 1-alkene content, this type of application may require a double bond isomerization step. 2) The production of high-quality polyalphaolefin (PAO) lubricating oil from n-1-olefins can be increased by metathesis of internal alkenes with ethene. This is especially useful for the production of PAO lubricating oil from alkanes via autoxidation [50], which yields mainly internal alkenes on dehydration. 3) Chemical applications can use metathesis technologies. 16.4.3 Skeletal Isomerization
Skeletal isomerization is the process whereby the skeletal structure of an alkene is rearranged so that the product is more branched than the feed. Increasing the degree of branching of alkenes has a number of benefits. Foremost is the increase in synthetic value of the branched alkenes for fuel production. Etherification of alkenes with alcohols (Section 16.3.2) requires branched alkenes, which can be provided from a skeletal isomerization unit. During oligomerization (Section 16.4.4), branched alkenes also produce more branched dimers and oligomers, which are especially useful if the material has to be hydrogenated. Alkylate-equivalent R
R
R
R
′R
R
R
+
+ ′R
Figure 16.6
′R
′R
′R
Alkene disproportionation (metathesis).
′R
16.4 Alkene Conversion
high-octane paraffinic motor-gasoline can be produced by selective dimerization and HYD of methylpropene (isobutene) [77, 78], and is an environmentally friendly alternative to aliphatic alkylation (Section 16.4.5). Oligomerization of skeletally isomerized alkenes therefore yields higher octane number hydrogenated motor-gasoline and the branched kerosene range material has a low freezing point making it suitable for jet fuel. Industrial processes for skeletal isomerization have been developed mainly for the conversion of n-butenes to isobutene [79, 80] and n-pentenes to methylbutenes [80, 81]. Hexenes can in principle also be skeletally isomerized [82] because hexenes are as resistant to cracking as pentenes, but they are less often used for etherification and oligomerization. Heptene and heavier alkenes are difficult to skeletally isomerize without significant cracking taking place. Skeletal isomerization is acid catalyzed and it is difficult to prevent side reactions. Alkenes are reactive molecules and, once the alkenes have been skeletally isomerized, other acid-catalyzed side reactions are accelerated. Skeletal isomerization is an integral part of many conversion processes, such as hydroisomerization, hydrocracking, and catalytic cracking. In fact, skeletal isomerization is one of the key reactions in acid-catalyzed hydrocarbon conversion and it is discussed in detail later (Chapter 18). In a Fischer–Tropsch refinery, skeletal isomerization of alkenes finds application in the following instances: 1) Skeletal isomerization in combination with etherification to produce high-octane fuel ethers for motor-gasoline. 2) Skeletal isomerization in combination with selective dimerization and HYD to produce high-octane motor-gasoline. However, alkylate-equivalent material can also be produced without a separate skeletal isomerization unit [83]. 16.4.4 Oligomerization
Oligomerization is used to refer to one or more consecutive alkene addition reactions and includes dimerization, which involves only a single addition. Short-chain alkenes can be oligomerized to produce longer chain alkenes. In this way, the carbon number distribution can be shifted to heavier products. Oligomerization is employed in some crude oil refineries, but alkene availability restricts its widespread use. Alkene availability is not constraining in a Fischer–Tropsch refinery, and oligomerization has been identified as a key refining technology for Fischer–Tropsch syncrude [7, 8]. By careful matching of the feed, catalyst, and operating conditions, a variety of high-quality products can be produced, which range from motor-gasoline to lubricating oils. Oligomerization is discussed in depth later on (Chapter 19). Oligomerization as is mostly applied in refining context is an acid-catalyzed and very exothermic reaction. Metal-catalyzed oligomerization can also be performed, but it is better suited for chemical applications. There are many potential applications of oligomerization in Fischer–Tropsch refining: 1) High-octane olefinic motor-gasoline can be produced by oligomerization of a wide range of alkenes. This type of conversion is the application of oligomerization most often found in crude oil refineries.
315
316
16 Refining Technology Selection
2) It can produce alkylate-equivalent high-octane paraffinic motor-gasoline by selective alkene oligomerization followed by HYD. Butenes are usually the feed material for this type of conversion, also called indirect alkylation, and the alkylate can be produced in two different ways. One way is to skeletally isomerize the butene to produce isobutene, which can then be selectively dimerized to trimethyl pentenes over an acidic resin or SPA catalyst [77, 78]. The other way is to dimerize the 1-butene-rich syncrude directly over SPA at low temperature to produce mainly trimethylpentenes [83, 84]. 3) It can produce IPK for synthetic jet fuel. Oligomerization naturally introduces branching in the product, which gives the hydrogenated kerosene fraction very good cold-flow properties. 4) By selecting an appropriate oligomerization technology that produces more linear distillate range material [85–87], the oligomers can be hydrogenated to yield a high cetane number distillate. 5) Lubricating base oils can be prepared by oligomerization of n-1-alkene-rich fractions. 6) Various chemicals can be produced by oligomerization: among others, the alkene feed for hydroformylation (Section 16.3.6) to produce plasticizer and detergent alcohols, and alkenes for benzene alkylation (Section 16.4.6) to produce detergents.
16.4.5 Aliphatic Alkylation
Aliphatic alkylation involves the condensation reaction of isobutane with alkenes to produce a heavier, highly branched alkane mixture called alkylate. Alkylate is one of the main high-octane paraffinic motor-gasoline blending components in crude oil refineries (Table 2.5). As technology, this is its only purpose. It is a homogeneous acid-catalyzed process employing either hydrofluoric acid (HF) or sulfuric acid (H2 SO4 ) as catalyst [88]. The products that are obtained are sensitive to the alkene feed and operating conditions, but either catalyst is capable of producing a high-octane paraffinic product (Table 16.1) [89, 90]. Table 16.1 Octane number of alkylate produced by HF and H2 SO4 alkylation of isobutane with different alkenes.
Alkene feed
Propene 1-Butene 2-Butenes Isobutene n-Pentenes Methylbutenes Pentenes (mixed) a Average
value for pentenes.
H2 SO4 alkylation
HF alkylation
RON
MON
RON
MON
89 97.8 97.8 93.2 91 91.2 –
87.1 93.9 93.9 90.3 88 88.8 –
91–92 94.4 97.8 95.9 82.5a – 90–91.5
89.5–90 91.6 94.6 93.4 – – 89–90
16.4 Alkene Conversion
F− + HF
(a) Initiation
+
F− (b) Hydride transfer
(c) Alkylation
(d) Propagation by hydride transfer
+
+
+
F−
+
+
F−
+
+ F−
+
+
F−
+
F−
+
Figure 16.7 Aliphatic alkylation mechanism, which involves hydride transfer from isobutane to another carbocation in order to make isobutane the primary carbocation source for alkylation.
Although aliphatic alkylation has been grouped under the alkene conversion technologies, it is also an alkane conversion technology. High-quality alkylate can be produced only if the isobutane becomes the carbocation source for alkylation. In order to ensure that the isobutane is activated by hydride transfer to another carbocation (Figure 16.7), aliphatic alkylation is conducted with a large excess of isobutane at low temperature. Typical operating conditions for H2 SO4 -catalyzed alkylation is 0–10 ◦ C at an isobutane:alkene ratio of 5–8, and for HF-catalyzed alkylation it is 10–40 ◦ C at an isobutane:alkene ratio of 10–15 [89]. The large excess of isobutane also limits the reaction of alkenes with each other, as well as alkene-related side reactions. The development of a solid-acid-catalyzed process for aliphatic alkylation has been ongoing for decades. Strong acidity is required to activate the isobutane and to promote hydride transfer from the isobutane to the carbocation-covered catalyst surface, but strong acidity easily leads to rapid catalyst deactivation [91, 92]. Progress is being made with solid-acid-catalyzed aliphatic alkylation, but it has yet to be adopted by industry. In most refineries, the motor-gasoline becomes off-specification without alkylate. Adopting a new technology in such a sensitive area is a risk, and liquid-acid-catalyzed processes remain the preferred aliphatic alkylation technologies despite their large environmental footprint. Unless there is an additional source of isobutane, aliphatic alkylation is not well matched with syncrude, which contains too much alkenes and too little alkanes in the C4 fraction. Technologies for alkylate productions were evaluated, and it was concluded that ‘‘indirect alkylation’’ by oligomerization (Section 16.4.4) and alkene HYD was preferred over aliphatic alkylation [83]. This is a significant difference between syncrude and crude oil refining, where aliphatic alkylation is key to the latter. 16.4.6 Aromatic Alkylation
The addition of an alkene to an aromatic is called aromatic alkylation. In the petrochemical industry, aromatic alkylation is widely used for the production of commodity chemicals such
317
318
16 Refining Technology Selection
as ethyl benzene and isopropyl benzene (cumene) [93–95]. It is not a process often found in fuel refineries, although it was suggested as an effective refining technology to lower refinery benzene levels [22, 23]. The maximum benzene content in motor-gasoline is regulated. The main advantage of employing aromatic alkylation to reduce the refinery benzene level is that it does not destroy the octane value of benzene, but retains it in the motor-gasoline. Aromatic alkylation is an acid-catalyzed process. In petrochemical production, acid-catalyzed side reactions of the alkene, such as oligomerization, is limited by operating at a high aromatic:alkene ratio. When this process is used within a fuel refinery, it is not necessary to follow the same operating philosophy, since alkene oligomers (polymer gasoline, Table 2.5) is a well-known motor-gasoline blending component. Depending on alkene availability in the refinery, one may even combine alkene oligomerization and benzene alkylation in a single process [96]. Aromatic alkylation is an important refining technology for syncrude and is discussed in detail in Chapter 20. There are a number of potential applications for aromatic alkylation in Fischer–Tropsch refining, and it is a complementary technology to nonacidic Pt/L-zeolite naphtha reforming (Section 16.5.3): 1)
2) 3)
4)
5)
6)
7)
Refinery benzene can be alkylated in a stand-alone unit, or by co-feeding the benzene with an alkene feed to an oligomerization unit. The former is preferred when aromatic alkylation is used in conjunction with nonacidic Pt/L-zeolite naphtha reforming, which produces benzene specifically for alkylation. A stand-alone unit also provides the refinery with the flexibility to produce fuels or chemicals from the benzene. Benzene alkylation in an oligomerization unit is useful when the benzene level in the naphtha fraction is <5%. Fully synthetic jet fuel production can be achieved in a single unit by combining aromatic alkylation and alkene oligomerization [97]. Linear alkyl benzenes (LABs) can be produced by aromatic alkylation with longer chain alkenes. The LAB product is a chemical, but it is also a high cetane number, high-density compound that in a pinch can be used to meet diesel fuel specifications [98]. It is possible to produce alkyl phenols. Although phenol is present only in a very low concentration in syncrude, there are pathways to phenol in Fischer–Tropsch refining. Phenol is industrially produced from cumene, which is also produced by aromatic alkylation. In coal-to-liquids facilities, phenol may be obtained from the gas liquor and coal tar naphtha produced during coal gasification (Chapter 3). The hindered phenols are commonly used antioxidants, and many of the hindered alkyl phenols display good antioxidant properties [99–101]. The alkylated phenols can also be employed as high-octane motor-gasoline blending components [102]. When a Fischer–Tropsch refinery has an associated coal tar refinery, the alkylation of aromatic coal tar naphtha holds many quality benefits [103]. Central to the derived benefit is the ability to exploit the alkenes, which would otherwise be hydrotreated to produce low octane number alkanes. Aromatic alkylation can be employed to shift the boiling range distribution in the refinery by converting naphtha range aromatics and alkenes into distillate. Depending on the nature of the alkene, such heavier alkyl aromatics have good jet fuel properties and acceptable diesel fuel properties. The refining of ethene in a stand-alone refinery far from petrochemical markets has been pointed out before. Aromatic alkylation of ethene is a useful refining pathway to convert the
16.5 Alkane Conversion
8)
ethene into a liquid product. Mono-alkylation to yield ethyl benzene can be used for chemical or motor-gasoline production, whereas multiple alkylation is useful for jet fuel production. Some synthetic lubricating oils can be prepared with an aromatic core and long-chain alkyl groups. Such oils are prepared by aromatic alkylation.
16.5 Alkane Conversion 16.5.1 Hydroisomerization
Hydroisomerization is an analogous process to skeletal isomerization, but for alkanes. The alkane product is more branched than the feed. The conversion of butane into isobutane and the conversion of n-alkane-rich light straight-run (LSR) naphtha into an isomerized naphtha (isomerate) are two processes often encountered in crude oil refining. Isobutane is used as feed material for aliphatic alkylation (Section 16.4.5) and isomerate is a final motor-gasoline blending component (Table 2.5). Heavier material can also be isomerized, and crude oil processes like catalytic dewaxing is essentially a hydroisomerization process. The main difference between skeletal isomerization and hydroisomerization is that the latter requires a bifunctional catalyst with metal and acid sites. The metal sites are required to dehydrogenate the alkanes to alkenes, which can then be skeletally isomerized on the acid sites. The isomerized alkenes can then be hydrogenated to alkanes. By doing so, the risk of side reactions occurring is reduced, because the amount of alkenes present at any given time is low. As in the case of skeletal isomerization, this is a fundamental step in the catalytic conversion of alkanes and is discussed in more depth later (Chapter 18). Hydroisomerization is an equilibrium-limited conversion. The equilibrium limitation can be overcome by separating and recycling the less isomerized material. There are various technologies for doing so [104]. The quality of the final product is determined by the overall degree of isomerization. For once-through operation, the conversion process determines the overall conversion, but with recycle the conversion step is of secondary importance. The use of hydroisomerization in a Fischer–Tropsch refinery is quite similar to that in a crude oil refinery and the applications depend on the carbon number of the feed: 1) Isobutane production for aliphatic alkylation is usually considered only if there is a significant source of n-butanes in the refinery to match the butenes. 2) Hydroisomerization of C5 –C6 naphtha produces a good quality paraffinic motor-gasoline blending component. When a single carbon number feed is used, recycling of less isomerized material becomes easier. It has also been shown that hydroisomerization technology can be employed without pretreatment with straight-run Fischer–Tropsch naphtha that is rich in n-alkenes [37]. 3) The cold-flow properties of distillate range material can be improved by mild hydroisomerization. With longer chain feed materials, there is always the risk of cracking, which is why mildly acidic catalysts perform so much better for the hydroisomerization of heavier n-alkane feed materials than their more acidic counterparts. Jet fuel requires sufficient isomerization of the kerosene to meet the freezing point specification, and there are not really any adverse
319
320
16 Refining Technology Selection
effects of a high degree of isomerization. In diesel fuel there is a trade-off between cold-flow properties and cetane number, and the degree of hydroisomerization must be controlled. 4) Lubricating base oils can be prepared from Fischer–Tropsch waxes by hydroisomerization to improve the cold-flow properties of the otherwise solid material [105, 106]. 16.5.2 Hydrocracking
Hydrocracking is strictly speaking a residue conversion technology, but due to its central role in LTFT wax upgrading it is better classified as an alkane conversion technology in the present context. In a crude oil refinery, hydrocracking has three functions. It is a hydrogen addition technology that removes heteroatoms, increases the H:C ratio of the product relative to the feed, and performs some HDA. Additionally it improves the cold-flow properties, and decreases average the boiling point of the material. Like hydroisomerization (Section 16.5.1), it requires a bifunctional catalyst, and hydrocracking is always accompanied by hydroisomerization. Because hydrocracking catalysts contain acid sites, hydrocracking is often preceded by a hydrotreater to remove most of the nitrogen bases that can undermine hydrocracker performance [107]. The hydrocracking of HTFT residue is analogous to crude oil hydrocracking in many respects, except for the absence of sulfur. The amount of HTFT residue is small and it contains some metal carboxylates that reduce the run length of fixed bed hydrocracking by pressure drop, rather than activity decline [26]. Hydrocracking of Fischer–Tropsch waxes is very different to typical crude oil hydrocracking (Table 16.2) [108, 109]. It is near isothermal, is not prone to the formation of carbonaceous deposits, and can be conducted at less severe conditions. Hydrocracking with mildly acidic unsulfided Pt/SiO2 –Al2 O3 catalysts enabled distillate selectivities of 75–80% at 70% wax conversion [110–112], which exceed that obtainable by sulfided base metal hydrocracking catalysts typically used for crude oil. More recently, Calemma and coworkers reported distillate selectivities around Table 16.2 Typical processing conditions for conventional crude oil hydrocrackers, mild crude oil hydrocrackers, and Fischer–Tropsch wax hydrocrackers.
Description
Operating conditions Temperature (◦ C) Pressure (MPa) Space velocity (h−1 ) H2 :feed ratio (m3 ·m−3 ) Conversion properties Cracking conversion (%) H2 consumption (mass% of feed) Heat release
Conventional hydrocracking
Mild hydrocracking
LTFT wax hydrocracking
350–430 10–20 0.2–2 800–2000
380–440 5–8 0.2–2 400–800
325–375 3.5–7 0.5–3 500–1800
70–100 1.4–4 Exothermic
20–40 0.5–1 Exothermic
20–100 <1 Near isothermal
16.5 Alkane Conversion
85% at close to complete conversion using a Pt/SiO2 –Al2 O3 catalyst [113]. Hydrocracking of Fischer–Tropsch syncrude is discussed in detail in Chapter 21. The main applications of hydrocracking in a Fischer–Tropsch refinery are as follows: 1) Jet fuel production by hydrocracking and hydroisomerization of distillate range and heavier material to isomerized paraffinic kerosene range material. 2) Production of distillate from C23 and heavier material. The distillate can then be used as a high cetane number, low-density blending component in diesel fuel; 3) Manufacturing of lubricant base oil by mild hydrocracking and isomerization of heavier than distillate syncrude fractions. Lubricant base oils can be produced from LTFT syncrude [105, 106], as well as from HTFT syncrude [114]. 16.5.3 Naphtha Reforming and Aromatization
Naphtha reforming has been a key refining unit in crude oil refineries for many decades (Sections 2.4.2–2.4.4). Initially, thermal naphtha reforming processes were employed [115], which were later superseded by catalytic naphtha reforming processes. In both cases the main products were a high-octane aromatic motor-gasoline and hydrogen. Catalytic naphtha reforming is a type of aromatization process. Catalytic naphtha reforming differs from the broader class of aromatization processes by its feed requirements. Aromatization processes are capable of producing aromatics from a wide range of feed materials, which include C5 and lighter hydrocarbons, whereas naphtha reforming is restricted to the aromatization of C6 and heavier hydrocarbons. The catalyst types that are used for naphtha reforming and aromatization are different, the latter requiring the ability to promote chain growth. Hydrogen production by catalytic naphtha reforming is of paramount importance in crude oil refineries, but less so in Fischer–Tropsch refineries where H2 is available from the Fischer–Tropsch gas loop. Standard catalytic naphtha reforming use chlorinated, platinum-promoted alumina (Pt/Cl− /Al2 O3 ) catalysts [116]. Fischer–Tropsch naphtha performs poorly with Pt/Cl− /Al2 O3 -catalyzed reforming on account of its high linear hydrocarbon content [4]. Conversely, Fischer–Tropsch syncrude performs very well with nonacidic Pt/L-zeolite reforming for the same reason. Catalytic naphtha reforming and the differences between the technologies are discussed in detail later (Chapter 22). Aromatization processes are normally found in petrochemical complexes and not in fuel refineries. During aromatization, the alkene partial pressure is much higher than during catalytic reforming and catalyst regeneration is more frequent. Although most aromatization technologies were developed to convert propane and butanes into aromatics, heavier feed materials and alkene-rich feed materials can also be converted. There are a number of applications for catalytic naphtha reforming and aromatization in a Fischer–Tropsch refinery: 1) The refinery C5 and heavier liquid yield can be improved by the aromatization of C3 –C4 alkanes, which would otherwise end up as liquid petroleum gas (LPG) [2]. 2) Aromatics can be produced from low octane number syncrude fractions that are difficult to upgrade by catalytic naphtha reforming. If the syncrude fractions contain oxygenates, it is
321
322
16 Refining Technology Selection
best to consider deoxygenation before aromatization; otherwise, there is a risk of premature catalyst deactivation [117]. 3) Typical reformate production by catalytic naphtha reforming provides a high octane number aromatic motor-gasoline blending component. The selection of the feed material and reforming technology is important. When a large amount of benzene is coproduced, the benzene must have an upgrading pathway, for example, by aromatic alkylation (Section 16.4.6). 4) Benzene, toluene, and xylenes (BTXs) can be produced as commodity chemicals. Such production is especially efficient using nonacidic Pt/L-zeolite reforming of hydrotreated C6 –C8 Fischer–Tropsch naphtha. Aromatization technology can also be employed for this purpose, but then the feed selection may be different. 16.5.4 Dehydrogenation
Dehydrogenation converts alkanes into their corresponding alkenes. The HYD of alkenes (Section 16.2.1) is a reversible metal-catalyzed reaction and, under high hydrogen partial pressure, typical for hydrotreating, the equilibrium favors HYD. When alkanes are brought into contact with a metal-promoted catalyst (e.g., Pt/Al2 O3 ) at high temperature, low pressure, and low partial pressure of H2 , dehydrogenation occurs (Equation 16.2) R–CH2 –CH2 –R R–CH=CH–R + H2
(16.2)
The operating temperature for metal-catalyzed dehydrogenation depends on the feed material. Heavier alkanes are dehydrogenated at around 450 ◦ C [118], but lighter alkanes (C3 –C5 ) require a temperature above 500 ◦ C [119]. Dehydrogenation of light alkanes can also be promoted by the use of O2 , which increases the thermodynamic driving force by removing H2 as water (Equation 16.3). R–CH2 –CH2 –R + 12 O2 → R–CH=CH–R + H2 O
(16.3)
Oxidative dehydrogenation takes place at high temperature and short contact time, and the process benefits from using a metal catalyst, such as Pt/Al2 O3 [120]. Unlike the exclusively metal-catalyzed reaction, oxidative dehydrogenation takes place by a free radical mechanism and the catalyst serves as a combustion promoter rather than a dehydrogenation catalyst. A lower temperature pathway for the dehydrogenation of alkanes to alkenes that is based on indirect oxidative dehydrogenation has been suggested [50], but it is not clear whether this pathway is more economical. A number of applications of dehydrogenation in a Fischer–Tropsch refinery can be envisioned: 1) Isobutene can be produced from butane by hydroisomerization in conjunction with dehydrogenation. Hydroisomerization of butane is more selective than skeletal isomerization of butene and this production route has a selectivity advantage. Syncrude does not contain much butane, but butane availability may be increased from other sources, including refining processes such as hydrocracking. 2) In locations where there is no market for LPG, the propane and butane can be dehydrogenated and refined with the propene and butenes present in syncrude.
16.6 Residue Conversion
3) Dehydrogenation can in principle be used to convert all alkanes that do not have a useful refining pathway into alkenes, which generally have more potential refining pathways. 4) Petrochemical refineries can make use of dehydrogenation to increase the amount of ethene and propene in the refinery. In the case of HTFT synthesis, ethane is as abundant as ethane, and better economy of scale for downstream ethene processing can be achieved by dehydrogenation. In this application, dehydrogenation is an alternative technology to steam cracking (Section 16.6.3). 5) Longer chain n-alkanes can be dehydrogenated for chemical applications that can employ linear internal alkenes, such as LAB production or the production of detergent alcohols [118]. The linear internal alkenes may also be double bond isomerized to produce n-1-alkenes [121].
16.6 Residue Conversion 16.6.1 Catalytic Cracking
Fluid catalytic cracking (FCC) is widely used in crude oil refining to upgrade residue fractions. Residue fractions are cracked to produce lighter products, but that is not the only purpose of FCC. The product has a higher H:C ratio than the feed, and the lighter cracking products contain a large amount of alkenes. Furthermore, the naphtha fraction has a reasonable octane number on account of its aromatic and alkene content (Table 2.5). Most of the alkenes needed for motor-gasoline production come from the FCC unit, making it a very important unit in a modern crude oil refinery. The feed properties influence the yield and quality of the product and, although it is a residue upgrading technology, it benefits from better quality feed. Hydrotreating the feed to an FCC unit can significantly reduce the heteroatom content in the final products, without adversely affecting the alkene yield from the unit [17]. The best type of feed material for FCC is a feed rich in cycloalkanes. An aromatic feed with high heteroatom and metal content will increase catalyst consumption and decrease yield [122]. Very heavy and aromatic feed materials are better upgraded by coking (Section 16.6.4). Catalytic cracking is an acid-catalyzed process. In a typical FCC, the cracking process is started by bringing the feed into contact with the very hot catalyst from the regenerator. The initial cracking reactions are thermal, but subsequent reactions are dominated by acid-catalyzed cracking. It is a carbon rejection process, and heavier aromatic compounds are deposited as coke on the catalyst. The catalyst is rapidly deactivated, and catalyst activity is restored by combusting the coke in the catalyst regenerator. Catalyst regeneration also provides the heat that drives the process, thereby beneficially employing the energy value of the carbon-rich deposits on the catalyst. Fischer–Tropsch syncrude is a comparatively clean feed for FCC, which can in principle be applied to convert heavy syncrude fractions into lighter olefinic products. It does not make sense to employ FCC as a carbon rejection technology with syncrude, because syncrude is already hydrogen rich.
323
324
16 Refining Technology Selection
A number of studies have been published on FCC of Fischer–Tropsch wax [123–128]. It has been reported that an LTFT refinery making use of FCC as opposed to hydrocracking (Section 16.5.2) for wax conversion is more economical [127]. Catalytic cracking has also been applied with HTFT naphtha (Section 9.5.5), but with less success. Short-chain alkanes are quite resistant to acid-catalyzed cracking. Catalytic cracking is discussed in detail later (Chapter 21). Some applications of FCC in Fischer–Tropsch refining are as follows: 1) Increasing the alkene content in an LTFT refinery by FCC of wax. This can facilitate motor-gasoline production, as well as enable other opportunities requiring more alkenes in the syncrude. 2) Converting heavier syncrude fractions from either HTFT or LTFT syncrude into lighter products to increase chemical and/or fuel production based on alkenes and aromatics. 3) The partial deoxygenation of syncrude as a feed pretreatment step before further refining. This would also modify the syncrude properties. It is analogous in concept to bauxite treatment, but more in line with suggestions to combine Fischer–Tropsch synthesis with acid-catalyzed conversion to improve the straight-run motor-gasoline properties of the syncrude [129–132].
16.6.2 Visbreaking
Historically, crude oil residues were mainly upgraded by thermal processes and, even after the development of catalytic cracking (Section 16.6.1), thermal conversion processes were still extensively used. The original objective of the technology was to avoid waste (cutter stock) by converting the heavy material into fuel oil [133]. Over time, this objective has changed somewhat, with increasing emphasis being placed on incremental naphtha and distillate production at the expense of fuel oil yield. The mildest form of thermal upgrading is visbreaking, which is a mild thermal cracking of residue. The residue is heated to 430–490 ◦ C, and the product is separated into light gas, naphtha, distillate, and residue fractions. The furnace exit temperature is set by the nature of the feed and the desired conversion level. The alkene content of the lighter products is fairly high. When atmospheric residue is employed as feed, visbreaking is a simple technology to obtain the maximum amount of naphtha and distillate while complying with residual fuel oil stability and viscosity specifications. When vacuum residue is employed as feed, the aim is to reduce the viscosity of the residue sufficiently so that it can be used as a fuel oil [115, 134]. Lower viscosity residue for fuel oil applications is the main product from visbreaking, and the conversion specifically targets the thermal cracking of long-chain alkanes in the residue to achieve this objective [17]. As such, it is a refining technology that has no real application in a Fischer–Tropsch refinery. 16.6.3 Thermal Cracking
Thermal cracking of crude-oil-derived residues has been effectively replaced by catalytic cracking (Section 16.6.1), which is a more efficient cracking technology. Thermal cracking (steam cracking)
16.6 Residue Conversion Cracking of Fischer–Tropsch wax. A comparison between thermal cracking and hydrocracking over a sulfided NiMo/SiO2 –Al2 O3 catalyst.
Table 16.3
Description
Yield of C1 –C4 (mass%) Yield of C5 -370 ◦ C (mass%) Naphtha (C5 –C9 ):Distillate (C10 –C22 ) Operating conditions Temperature (◦ C) Pressure (MPa) LHSV (h−1 ) H2 :wax (m3 ·m−3 , normal)
Thermal cracking
Hydrocracking
With H2
No H2
2 20 –a
1 21 1:9
3 28 1:7
442 2 1 250
476 0 8 0
370 3.5 1 1500
a Naphtha to distillate ratio not reported, but alkene content listed by source as 39% n-1-alkenes and 3% other alkene isomers.
at severe conditions is still applied for the conversion of light alkane gases and naphtha into ethene and propene for petrochemical applications. Alkane-rich feed materials are in general good thermal cracking feedstocks, and the cracking behavior of the alkanes are well documented in the classic work by Egloff [135]. The thermal cracking of heavier alkanes yields a product that is rich in n-1-alkenes and resembles an oxygenate-free HTFT syncrude. Fischer–Tropsch waxes were in the past upgraded by thermal cracking at mild conditions (Sections 6.4.3 and 8.4.2). Fuel production by thermal cracking of LTFT waxes was evaluated, and it was found to be less efficient than hydrocracking although it required no hydrogen co-feed and had a lower gas make (Table 16.3) [136]. Thermal cracking is discussed in more detail in Chapter 21. Thermal cracking is better suited than the other cracking technologies for the production of chemicals and chemical precursors from syncrude. The following petrochemical applications are noted: 1)
2) 3) 4) 5)
6)
It was demonstrated that LTFT naphtha is a good feed material for steam cracking [137]. This has been suggested as an upgrading pathway for the naphtha from partial Fischer–Tropsch refining (Section 12.4.1). Light alkanes (C2 –C4 ) can be used as steam cracking feedstock for petrochemical production. Wax and slack wax cracking can be a source of n-1-alkenes for applications such as PAO lubricant and detergent production [138, 139]. Mild cracking may be used to change wax properties for chemical applications [136]. The partial thermal cracking of Fischer–Tropsch wax on its own and in combination with polyethylene waste plastic can be employed to produce a base stock for hydroisomerization to lubricating base oil [140]. Thermal cracking can be used to lower the viscosity and improve the cold-flow properties of LTFT syncrude to enable transportation by pipeline. This is a typical upgrader-type application, which enables the syncrude to be sold as a synthetic crude oil for refining elsewhere.
325
326
16 Refining Technology Selection
16.6.4 Coking
When thermal cracking of a residue is initiated and the product is then kept at a temperature around 450–500 ◦ C for a prolonged period, it produces hydrogen-enriched light products and an aromatic coke. Coking is a carbon rejection technology and the objective is twofold. Products in the fuel boiling range is produced with a higher H:C ratio than the feed. At the same time, carbon and heteroatoms are rejected as coke, thereby reducing the H2 requirement of the refinery. The coke itself may also be a valuable product. The quality and potential applications for which the coke may be suited depend on the quality of the residue [141]. It is therefore not possible to convert a poor quality residue by coking into hydrogen-enriched light products and good quality coke. The coking propensity of a feed is measured in terms of its Conradson carbon residue, as determined by the ASTM D 189 standard test method [142]. With the exception of HTFT residue, most Fischer–Tropsch syncrude fractions have a low coking propensity. This is understandable, because syncrude has a H:C ratio around 2, which makes carbon rejection very difficult. Coking has little application in a Fischer–Tropsch refinery, although it may be applied in coal-to-liquids facilities that coproduce coal liquids during low-temperature gasification. 16.7 Fischer–Tropsch Refining Technology Selection
Refinery technology selection goes hand in hand with refinery design. Any of the refinery technologies that were discussed can be made to work with Fischer–Tropsch syncrude. The operative construct is ‘‘made to work.’’ If we apply a Green chemistry and engineering approach, refining technologies can be preselected on the basis of their goodness of fit with syncrude and their overall environmental footprint. Selection criteria have been developed for the evaluation of refining technologies for use with Fischer–Tropsch syncrude (Table 16.4) [4]: 1) Fischer–Tropsch syncrude compatibility: There are two aspects that govern compatibility between syncrude and a refining technology. The first is the amount of feed pretreatment that is required to make the technology work. The second is that the chemistry is efficient for the conversion of the molecules in the feed. This is often catalyst dependent, and a specific refining technology may have different syncrude compatibility issues when different catalysts are used. A three-point scale is employed: ‘‘Good’’ indicates that syncrude has an advantage compared to crude oil when used with this technology and catalyst combination. ‘‘Average’’ means that syncrude is equally well suited to this technology as crude oil and that the same effort in feed pretreatment is required. ‘‘Poor’’ indicates that syncrude requires significantly more feed pretreatment than crude oil to make it compatible, or that the purpose of the technology is not well aligned with syncrude composition. For example, it has been indicated that FCC performs well with LTFT wax as feed, but it is a carbon rejection technology and carbon rejection does not make sense in the context of wax refining. If you want a light alkene-rich syncrude, select HTFT synthesis, it is likely to be more efficient than converting an LTFT syncrude into an HTFT syncrude in the refinery. 2) Waste generation: Waste generation is an indication of the extent of waste that is generated by the technology in relation to a typical refining process. A three-point scale is employed: ‘‘Low’’ refers to processes that generate the same or less waste than the norm, where one
16.7 Fischer–Tropsch Refining Technology Selection Refinery technologies evaluated for processing of Fischer–Tropsch syncrude in terms of compatibility, environmental impact, and recommended use.
Table 16.4
Refining technology
Hydrotreating Alkene hydrogenation Hydrodeoxygenation Addition and removal of oxygen Alcohol dehydration Etherification (alcohol + alcohol) Etherification (alcohol + alkene) Alkene hydration Esterification Carbonyl aromatization Hydroformylation Autoxidation Alkene conversion Double bond isomerization Metathesis Butene skeletal isomerization Pentene skeletal isomerization Alkene di-/oligomerization
Aliphatic alkylation Aromatic alkylation Alkane conversion Butane hydroisomerization C5 –C6 hydroisomerization
Hydrocracking Catalytic reforming Aromatization
Catalyst
Syncrude compatibility
Environmental impact Waste
Chemicals
Energy use
Sulfided Unsulfided Sulfided Unsulfided
Average Good Average Average
Low Low Low Low
DMDSa None DMDS None
Low Low Moderate Low
Alumina Acidic resin Acidic resin H3 PO4 /support Solid acid Solid acid Rh/Co-complex Thermal
Good Average Average Average Average Average Good Good
Low Low Low Moderate Low Low Low Low
None None None H3 PO4 None None Catalyst None
Moderateb Low Low Moderate Low Low Low Low
Alumina Metal oxide Ferrierite Alumina Acidic mol sieve Acidic resin H-ZSM-5 MFS/TON ASA SPA Homogeneous Thermal HF H2 SO4 SPA Zeolites
Good Average Average Good Averagee Average Good Average Good Good Average Good Poorf,g Poorf,g Good Average
Low Low Lowd Lowd Low Low Low Low Low Low Moderate Low Moderate High Low Low
None None None None None None None None None None Catalyst/NaOH None HF H2 SO4 None None
High Moderatec High High Moderate Low Moderate Low Low Low Low High Low Low Low Moderate
Pt/Cl – /Al2 O3 Pt/Cl – /Al2 O3 Pt/metal oxide Pt/mordenite Sulfided Unsulfided Pt/Cl – /Al2 O3 Nonacidic Pt/L M/H-ZSM-5
Average Poore Poorg Average Average Good Poorg,h Good Average
Low Low Low Low Low Low Lowd Low Low
Chloro-alkane Chloro-alkane None None DMDS None Chloro-alkane None None
Moderate Low Low Low High High High High High
(continued overleaf )
327
328
16 Refining Technology Selection Table 16.5
(continued )
Refining technology
Residue conversion Catalytic cracking Visbreaking Thermal cracking Coking
Catalyst
Zeolite Thermal Thermal Thermal
Syncrude compatibility
Poori Poor Good Poor
Environmental impact Waste
Chemicals
Energy use
Low Low Low Low
None None None None
High High High High
a DMDS = dimethyl
disulfide (CH3 –S–S–CH3 ) on the feed: ethanol dehydration (high), other alcohols (moderate). c Depends on metal oxide catalyst: Re O (low), MoO (moderate), and WO (moderate/high). 2 7 3 3 d Side product formation is 10% or more. e Depends on feed origin: oxygenates adsorb on catalyst to decrease its operating window. f Depends on butane availability, usually butane:butene 1. g Oxygenates and water must be removed from feed to limit catalyst deactivation. h High linear hydrocarbon content and low cycloalkane and aromatic content (N + 2A < 30). i Works well with wax, but attempts carbon rejection with a hydrogen-rich feed. b Depends
anticipates solid waste from catalyst change and some by-products during conversion. Most processes fall within this category. ‘‘Moderate’’ indicates that the technology produces waste in excess of the norm. ‘‘High’’ is reserved for the waste management which in relation to the technology is significant. One would select processes with ‘‘moderate’’ or ‘‘high’’ waste generation only if they provide a conversion that is essential to the refinery design. 3) Chemical addition: Some refining processes rely on the addition of chemicals to work. Often only low levels of addition are required. It is a judgment call whether the additional hazard and the associated increase in environmental footprint are acceptable or not. 4) Energy requirements: The carbon efficiency of Fischer–Tropsch-based processes is related not only to the efficiency of converting raw material carbon into final products, but also to the overall carbon balance over the system. Any net energy flow has an equivalent carbon cost, irrespective of how the energy was generated or consumed. A three-point scale is employed to rank energy use related to operating temperature to ‘‘low,’’ ‘‘moderate,’’ and ‘‘high,’’ where ‘‘high’’ is for processes requiring temperatures >350 ◦ C. Care should be taken in interpreting these designators, since a high-temperature process with good heat management can easily be more efficient than a low-temperature process due to the quality of the heat (higher temperature is more useful) and the net energy flow. The qualitative assessment provided in Table 16.4 is a guideline but is not a substitute for sound engineering judgment. References 1. De Klerk, A. (2007) Environmentally
friendly refining: Fischer-Tropsch versus crude oil. Green Chem., 9, 560–565.
2. Dancuart, L.P., De Haan, R., and De Klerk, A.
(2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532.
References 3. Van Wyk, A.M., Moola, M.A., De Bruyn, C.J.,
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Venter, E., and Collier, L. (2007) Molecule management for clean fuels at Sasol. Chem. Technol. (S. Afr.), April, 4–10. De Klerk, A. (2008) Fischer-Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. De Klerk, A. (2009) Catalysts important in the refining of Fischer-Tropsch syncrude to fuels. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 116–117. De Klerk, A. (2008) Refining of Fischer-Tropsch syncrude: lessons from the past. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 53 (2), 105–109. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, FL, pp. 331–364. Dry, M.E. (1987) Chemicals produced in a commercial Fischer-Tropsch process. ACS Symp. Ser., 328, 18–33. Gregor, J.H. (1990) Fischer-Tropsch products as liquid fuels or chemicals. An economic evaluation. Catal. Lett., 7, 317–332. Steynberg, A.P., Nel, W.U., and Desmet, M.A. (2004) Large scale production of high value hydrocarbons using Fischer-Tropsch technology. Stud. Surf. Sci. Catal., 147, 37–42. Redman, A. (2005) Production of olefins and oxygenated compounds from Fischer-Tropsch. Preprints of the 18th World Petroleum Congress, September 25–29, Johannesburg, p. cd179. De Klerk, A., Dancuart, L.P., and Leckel, D.O. (2005) Chemicals refining from Fischer-Tropsch synthesis. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd185. Leprince, P. (ed.) (2001) Petroleum Refining, Conversion Processes, Vol. 3, Editions Technip, Paris. Meyers, R. A. (ed.) (2004) Handbook of Petroleum Refining Processes, 3rd edn, McGraw-Hill, New York. Jones, D.S.J. and Pujad´o, P.R. (eds) (2006) Handbook of Petroleum Processing, Springer, Dordrecht. Gary, J.H., Handwerk, G.E., and Kaiser, M.J. (2007) Petroleum Refining. Technology and Economics, 5th edn, CRC Press, Boca Raton, FL.
18. Speight, J.G. (2007) The Chemistry and Tech-
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29. 30. 31.
32.
33. 34.
nology of Petroleum, 4th edn, CRC Press, Boca Raton, FL. Delmon, B. (1993) New technical challenges and recent advances in hydrotreatment catalysis. A critical updating review. Catal. Lett., 22, 1–32. Furimsky, E. (1998) Selection of catalysts and reactors for hydroprocessing. Appl. Catal. A, 171, 177–206. Stanislaus, A. and Cooper, B.H. (1994) Aromatic hydrogenation catalysis: a review. Catal. Rev.-Sci. Eng., 36, 75–123. Goezler, A.R., Hernandez-Robinson, A., Ram, S., Chin, A.A., Harandi, M.N., and Smith, C.M. (1993) Refiners have several options for reducing gasoline benzene. Oil Gas J., 91 (13), 63–69. De Klerk, A. and Nel, R.J.J. (2008) Benzene reduction in a fuels refinery: an unconventional approach. Energy Fuels, 22, 1449–1455. Dautzenberg, F.M. and De Deken, J.C. (1987) Modes of operation in hydrodemetallization. ACS Symp Ser., 344, 233–256. Toulhoat, H., Szymanski, R., and Plumail, J.C. (1990) Interrelations between initial pore structure, morphology and distribution of accumulated deposits, and lifetimes of hydrodemetallisation catalysts. Catal. Today, 7, 531–568. De Klerk, A. (2008) Hydroprocessing peculiarities of Fischer-Tropsch syncrude. Catal. Today, 130, 439–445. Chianelli, R.R. (1984) Fundamental studies of transition metal sulfide hydrodesulfurization catalysts. Catal. Rev.-Sci. Eng., 26, 361–393. Ho, T.C. (2004) Deep HDS of diesel fuel: chemistry and catalysis. Catal. Today, 98, 3–18. Ho, T.C. (1988) Hydrodenitrogenation catalysis. Catal. Rev.-Sci. Eng., 30, 117–160. Perot, G. (1991) The reactions involved in hydrodenitrogenation. Catal. Today, 10, 447–472. Prins, R., Zhao, Y., Sivasamkar, N., and Kukula, P. (2005) Mechanism of C-N bond breaking in hydrodenitrogenation. J. Catal., 234, 509–512. Pienaar, A.D. and De Klerk, A. (2008) Nickel catalyst stability toward carboxylic acids. Ind. Eng. Chem. Res., 47, 4962–4965. Asinger, F. (1968) Mono-Olefins Chemistry and Technology, Pergamon, Oxford, pp. 1039–1040. Pecque, M. and Maurel, R. (1969) Hydrog´enation catalytique. III. Cin´etique de
329
330
16 Refining Technology Selection
35.
36.
37.
38.
39.
40. 41.
42.
43.
44.
45.
46.
47.
48.
l’hydrog´enation avec isom´erisation. Bull. Soc. Chim. Fr., 6, 1878–1882. De Klerk, A. and Strauss, M.J. (2008) Opportunities for efficiency improvement in high temperature Fischer-Tropsch hydroprocessing units. Prepr. Pap.-Am. Chem. Soc., Div. Fuel. Chem., 53 (1), 313–314. Br´emaud, M., Vivier, L., P´erot, G., Harl´e, V., and Bouchy, C. (2005) Hydrogenation of olefins over hydrotreating catalysts. Promotion effect on the activity and on the involvement of H2 S in the reaction. Appl. Catal. A, 289, 44–50. Lamprecht, D. and De Klerk, A. (2009) Hydroisomerization of 1-pentene to iso-pentane in a single reactor. Chem. Eng. Commun., 196, 1206–1216. Overtoom, R., Fabricius, N., and Leenhouts, W. (2009) in Proceedings of the 1st Annual Gas Processing Symposium (eds H.E. Alfadala, G.V.R. Reklaitis, and M.M. El-Halwagi), Elsevier, Amsterdam, pp. 378–386. Furimsky, E. (1983) Chemistry of catalytic hydrodeoxygenation. Catal. Rev.-Sci. Eng., 25, 421–458. Furimsky, E. (2000) Catalytic hydrodeoxygenation. Appl. Catal. A, 199, 147–190. Tahara, K., Nagahara, E., Itoi, Y., Nishiyama, S., Tsuruya, S., and Masai, M. (1997) Liquid-phase hydrogenation of carboxylic acid on supported bimetallic Ru-Sn-alumina catalysts. Appl. Catal. A, 154, 75–86. Aly, M. and Baumgarten, E. (2001) Hydrogenation of hexanoic acid with different catalysts. Appl. Catal. A, 210, 1–12. Rachmady, W. and Vannice, M.A. (2002) Acetic acid reduction by H2 on bimetallic Pt-Fe catalysts. J. Catal., 209, 87–98. De Klerk, A. (2001) Hydrotreating in a Fischer-Tropsch refinery. 2nd Sub-Saharan Africa Catalysis Symposium, November 5–7, 2001, Swakopmund, Namibia, paper 10. Lamprecht, D. (2007) Hydrogenation of Fischer-Tropsch synthetic crude. Energy Fuels, 21, 2509–2513. Leckel, D.O. (2007) Selectivity effect of oxygenates in hydrocracking of Fischer-Tropsch waxes. Energy Fuels, 21, 662–667. Leckel, D.O. (2009) Hydroprocessing Euro 4-type diesel from high-temperature Fischer-Tropsch vacuum gas oils. Energy Fuels, 23, 38–45. Nel, R.J.J. and De Klerk, A. (2007) Fischer-Tropsch aqueous phase refining by
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
catalytic alcohol dehydration. Ind. Eng. Chem. Res., 46, 3558–3565. McGurk, K. (2003) From 1-heptene to 1-octene: a new production route. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd082. Bolder, F.H.A., De Klerk, A., and Visagie, J.L. (2009) Hydrogenation of oxidized wax and a process to produce olefins from paraffins by autoxidation, selective hydrogenation, and dehydration. Ind. Eng. Chem. Res., 48, 3755–3760. Nel, R.J.J. and De Klerk, A. (2009) Dehydration of C5 -C12 linear 1-alcohols over η-alumina to fuel ethers. Ind. Eng. Chem. Res., 48, 5230–5238. De Klerk, A. (2004) Etherification of C6 Fischer-Tropsch material for linear α-olefin recovery. Ind. Eng. Chem. Res, 43, 6349–6354. Travers, P. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 291–319. L´opez, D.E., Goodwin, J.G. Jr., Bruce, D.A., and Lotero, E. (2005) Transesterification of triacetin with methanol on solid acid and base catalysts. Appl. Catal. A, 295, 97–105. Bolder, F.H.A. (2009) Removal of carboxylic acids in Fischer-Tropsch hydrocarbon product. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 54 (1), 1–3. Lee, D.G., Yan, Y., and Chandler, W.D. (1994) Measurement of equilibrium constants for the formation of esters from aliphatic carboxylic acids and alcohols. Anal. Chem., 66, 32–34. Chang, Y.-C. and Ko, A.-N. (2000) Vapor phase reactions of acetaldehyde over type X zeolites. Appl. Catal. A, 190, 149–155. Hoang, T.Q., Zhu, X., Sooknoi, T., Resasco, D.E., and Mallinson, R.G. (2010) A comparison of the reactivities of propanal and propylene on HZSM-5. J. Catal., 271, 201–208. Salvapati, G.S., Ramanamurthy, K.V., and Janardanarao, M. (1989) Selective catalytic self-condensation of acetone. J. Mol. Catal., 54, 9–30. Nel, R.J.J. and De Klerk, A. (2009) Overview of Fischer-Tropsch aqueous product refining strategies. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 118–119. Beller, M., Cornils, B., Frohning, C.D., and Kohlpaintner, C.W. (1995) Progress in hydroformylation and carbonylation. J. Mol. Catal. A, 104, 17–85.
References 62. Gates, B.C., Katzer, J.R. and Schuit, G.C.A.
63.
64.
65.
66.
67.
68.
69.
70.
71. 72.
73.
74.
(1979) Chemistry of Catalytic Processes, McGraw-Hill, New York, pp. 140–147. De Klerk, A. (2003) Continuous-mode thermal oxidation of Fischer-Tropsch waxes. Ind. Eng. Chem. Res., 42, 6545–6548. Chelton, H.M., Traxler, R.N., and Romberg, J.W. (1959) Oxidized asphalts in a vertical pilot plant. Ind. Eng. Chem., 51, 1353–1354. Attar, A. and Corcoran, W.H. (1978) Desulfurization of organic sulphur compounds by selective oxidation. 1. Regenerable and nonregenerable oxygen carriers. Ind. Eng. Chem. Prod. Res. Dev., 17 (2), 102–109. Shiraishi, Y., Tachibana, K., Hirai, T., and Komasawa, I. (2002) Desulfurization and denitrogenation process for light oils based on chemical oxidation followed by liquid-liquid extraction. Ind. Eng. Chem. Res., 41, 4362–4375. Babich, I.V. and Moulijn, J.A. (2003) Science and technology of novel processes for deep desulfurization of oil refinery streams: a review. Fuel, 82, 607–631. Ishihara, A., Wang, D., Dumeignil, F., Amano, H., Qian, E.W., and Kabe, T. (2005) Oxidative desulfurization and denitrogenation of a light gas oil using an oxidation/adsorption continuous flow process. Appl. Catal. A, 279, 279–287. Da Coneicao, L., De Almeida, C.L., Egues, S., Dallago, R.M., Paroul, N., Do Nascimento Filho, I., De Souza, W.F., and Pergher, S.B.C. (2005) Preliminary study of the oxidation of nitrogen compounds of gas oil from Brazilian petroleum. Energy Fuels, 19, 960–963. Ukkirapandian, V., Sadasivam, V., and Sivasankar, B. (2008) Oxidation of dibenzothiophene and desulphurization of diesel. Pet. Sci. Technol., 26, 423–435. Dunning, H.N. (1953) Review of olefin isomerization. Ind. Eng. Chem., 45, 551–564. Karinen, R.S., Lylykangas, M.S., and Krause, A.O.I. (2001) Reaction equilibrium in the isomerization of 2,4,4-trimethyl pentenes. Ind. Eng. Chem. Res., 40, 1011–1015. Kondo, J.N. and Domen, K. (2003) IR observations of adsorption and reactions of olefins on H-form zeolites. J. Mol. Catal. A, 199, 27–38. Li, J. and Davis, R.J. (2003) On the use of 1-butene double-bond isomerisation as probe reaction on cesium-loaded zeolite X. Appl. Catal. A, 239, 59–70.
75. Rosenberg, D.J., Bachiller-Baeza, B., Dines,
76. 77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
T.J., and Anderson, J.A. (2003) Nature of surface sulfate species and the generation of active sites on silica-zirconia mixed-oxide catalysts. J. Phys. Chem. B, 107, 6526–6534. Mol, J.C. (2004) Industrial applications of olefin metathesis. J. Mol. Catal. A, 213, 39–45. Meister, J.M., Black, S.M., Muldoon, B.S., Wei, D.H., and Roeseler, C.M. (2000) Optimize alkylate production for clean fuels. Hydrocarbon Process., 79 (5), 63–75. Birkhoff, R. and Nurminen, M. (2004) in Handbook of Petroleum Refining Processes, 3rd edn (ed. R.A. Meyers), McGraw-Hill, New York, pp. 1.3–1.9. Wise, J.B. and Powers, D. (1994) Highly selective olefin skeletal isomerization process. ACS Symp. Ser., 552, 273–285. Ozmen, S.M., Abrevaya, H., Barger, P., Bentham, M., and Kojima, M. (1993) Skeletal isomerization of C4 and C5 olefins for increased ether production. Fuel Reformulat., 3 (5), 54–59. Duplan, J.-L., Amigues, P., Verstraete, J., and Travers, C. (1996) Kinetic studies of the skeletal isomerization of n-pentenes over the ISO-5 process catalyst. Proc. – Ethylene Prod. Conf., 5, 429–449. Li, D., Li, M., Chu, Y., Nie, H., and Shi, Y. (2003) Skeletal isomerization of light FCC naphtha. Catal. Today, 81, 65–73. De Klerk, A. and De Vaal, P.L. (2008) Alkylate technology selection for Fischer-Tropsch syncrude refining. Ind. Eng. Chem. Res., 47, 6870–6877. De Klerk, A., Engelbrecht, D.J., and Boikanyo, H. (2004) Oligomerization of Fischer-Tropsch olefins: effect of feed and operating conditions on hydrogenated motor-gasoline quality. Ind. Eng. Chem. Res., 43, 7449–7455. Garwood, W.E. (1983) Conversion of C2 -C10 to higher olefins over synthetic zeolite ZSM-5. ACS Symp. Ser., 218, 383–396. K¨ohler, E., Schmidt, F., Wernicke, H.J., De Pontes, M., and Roberts, H.L. (1995) Converting olefins to diesel – the COD process. Hydrocarbon Technol. Int., (Summer), 37–40. De Klerk, A. (2007) Properties of synthetic fuels from H-ZSM-5 oligomerization of Fischer-Tropsch type feed material. Energy Fuels, 21, 3084–3089. Corma, A. and Mart´ınez, A. (1993) Chemistry, catalysts, and processes for isoparaffin-olefin
331
332
16 Refining Technology Selection
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
alkylation. Actual situation and future trends. Catal. Rev.-Sci. Eng., 35, 483–570. Joly, J.-F. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 257–289. Nielsen, R.H. (2001) Alkylation for Motor Fuels. SRI Process Economics Program Report 88B, Stanford Research Institute, Menlo Park, CA. Weitkamp, J. and Traa, Y. (1999) Isobutane/butene alkylation on solid catalysts. Where do we stand? Catal. Today, 49, 193–199. Nayak, S.V., Ramachandran, P.A., and Dudukovic, M.P. (2010) Modeling of key reaction pathways: Zeolite catalyzed alkylation processes. Chem. Eng. Sci., 65, 335–342. Degnan, T.F. Jr., Smith, C.M., and Venkat, C.R. (2001) Alkylation of aromatics with ethylene and propylene: recent developments in commercial processes. Appl. Catal. A, 221, 283–294. Perego, C. and Ingallina, P. (2002) Recent advances in the industrial alkylation of aromatics: new catalysts and new processes. Catal. Today, 73, 3–22. Schmidt, R.J. (2005) Industrial catalytic processes – phenol production. Appl. Catal. A, 280, 89–103. Sakuneka, T.M., Nel, R.J.J., and De Klerk, A. (2008) Benzene reduction by alkylation in a solid phosphoric acid catalyzed olefin oligomerization process. Ind. Eng. Chem. Res., 47, 7178–7183. Sakuneka, T.M., De Klerk, A., Nel, R.J.J., and Pienaar, A.D. (2008) Synthetic jet fuel production by combined propene oligomerization and aromatic alkylation over solid phosphoric acid. Ind. Eng. Chem. Res., 47, 1828–1834. De Klerk, A. (2009) Can Fischer-Tropsch syncrude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. Rosenwald, R.H., Hoatson, J.R., and Chenicek, J.A. (1950) Alkyl phenols as antioxidants. Ind. Eng. Chem., 42, 162–165. Wasson, J.I. and Smith, W.M. (1953) Effect of alkyl substitution on antioxidant properties of phenols. Ind. Eng. Chem., 45, 197–200. Nixon, A.C., Minor, H.B., and Calhoun, G.M. (1956) Effect of alkyl phenols on storage and manifold stability of gasolines. Ind. Eng. Chem., 48, 1874–1880.
102. Singerman, G.M. (1982) Gasoline extended/
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
octane improvers from coal liquids. Energy Progress, 2 (2), 100–106. De Klerk, A. and Nel, R.J.J. (2007) Phenol alkylation with 1-octene on solid acid catalysts. Ind. Eng. Chem. Res., 46, 7066–7072. Travers, C. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 229–256. Calemma, V., Peratello, S., Stroppa, F., Giardino, R., and Perego, C. (2004) Hydrocracking and hydroisomerization of long-chain n-paraffins. Reactivity and reaction pathway for base oil formation. Ind. Eng. Chem. Res., 43, 934–940. Kobayashi, M., Saitoh, M., Togawa, S., and Ishida, K. (2009) Branching structure of diesel and lubricant base oils prepared by isomerization/hydrocracking of Fischer-Tropsch waxes and alpha-olefins. Energy Fuels, 23, 513–518. Scherzer, J. and Gruia, A.J. (1996) Hydrocracking Science and Technology, Taylor & Francis, Boca Raton, FL. Billon, A. and Bigeard, P.-H. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Technip, Paris, pp. 333–364. Bouchy, C., Hastoy, G., Guillon, E., and Martens, J.A. (2009) Fischer-Tropsch waxes upgrading via hydrocracking and selective hydroisomerization. Oil Gas Sci. Technol. - Rev. IFP, 64 (1), 91–112. Calemma, V., Peratello, S., Pavoni, S., Clerici, G., and Perego, C. (2001) Hydroconversion of a mixture of long chain n-paraffins to middle distillate: effect of the operating parameters and products properties. Stud. Surf. Sci. Catal., 136, 307–312. Calemma, V., Correra, S., Perego, C., Pollesel, P., and Pellegrini, L. (2005) Hydroconversion of Fischer-Tropsch waxes: assessment of the operating conditions effect by factorial design experiments. Catal. Today, 106, 282–287. Leckel, D.O. (2007) Noble metal wax hydrocracking catalysts supported on high-siliceous alumina. Ind. Eng. Chem. Res., 46, 3505–3512. Calemma, V., Gambaro, C., Parker, W.O. Jr., Carbone, R., Giardino, R., and Scorletti, P. (2010) Middle distillates from hydrocracking of FT waxes: Composition, characteristics and emission properties. Catal. Today, 149, 40–46. Morgan, P.M., Van der Merwe, D.G., Goosen, R., Leckel, D.O., Saayman, H.M., Loubser, H., and Botha, J.J. (1999) The
References
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
production of high quality base oils from Sasol hydrocrackates. S. Afr. Mech. Eng., 49 (4), 11–13. Cooper, T.A. and Ballard, W.P. (1962) in Advances in Petroleum Chemistry and Refining, Vol. VI (eds K.A. Kobe and J.J. Mc Ketta Jr.), John Wiley & Sons, Inc., New York, pp. 171–238. Antos, G. J. and Aitani, A. M. (eds) (2004) Catalytic Naphtha Reforming, 2nd edn, Marcel Dekker, New York. De Klerk, A. (2003) in Catalysis in Application (eds S.D. Jackson, J.S.J. Hargreaves, and D. Lennon), Royal Society of Chemistry, Cambridge, pp. 24–31. Muller, G.R. and Arakawa, M. (1971) Linear Alpha Olefins. SRI Process Economics Program Report 12A, Stanford Research Institute, Menlo Park, CA, pp. 103–132. Sanfilippo, D. (2000) Dehydrogenation of paraffins. Key technology for petrochemicals and fuels. CatTech, 4 (1), 56–72. Beretta, A., Forzatti, P., and Ranzi, E. (1999) Production of olefins via oxidative dehydrogenation of propane in autothermal conditions. J. Catal., 184, 469–478. De Klerk, A., Hadebe, S.W., Govender, J.R., Jaganyi, D., Mzinyati, A.B., Robinson, R.S., and Xaba, N. (2007) Linear α-olefins from linear internal olefins by a boron-based continuous double-bond isomerization process. Ind. Eng. Chem. Res., 46, 400–410. Bonifay, R. and Marcilly, C. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 169–227. Kobolakis, I. and Wojciechowski, B.W. (1985) The catalytic cracking of a Fischer-Tropsch synthesis product. Can. J. Chem. Eng., 63, 269–277. Abbot, J. and Wojciechowski, B.W. (1985) Catalytic cracking on HY and HZSM-5 of a Fischer-Tropsch product. Ind. Eng. Chem. Prod. Res. Dev., 24, 501–507. Leib, T.M., Kuo, J.C.W., Garwood, W.E., Nace, D.M., Derr, W.R., and Tabak, S.A. (1988) Upgrading of Fischer-Tropsch waxes to high quality transportation fuels. Proceedings AIChE Annual Meeting, Washington, DC, paper 61d. Schwartz, M.M. (1995) The Selective Catalytic Cracking of Fischer-Tropsch Liquids to High Value Transportation Fuels. DOE Contract no. DE-AC22– 91PC90057 - Final Report.
127. Choi, G.N., Kramer, S.J., Tam, S.S.,
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
Fox, J.M. III, and Reagan, W.J. (1996) Fischer-Tropsch indirect coal liquefaction design/economics – mild hydrocracking vs. fluid catalytic cracking. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 41 (3), 1079–1083. Dupain, X., Krul, R.A., Makkee, M., and Moulijn, J.A. (2005) Are Fischer-Tropsch waxes good feedstocks for fluid catalytic cracking units? Catal. Today, 106, 288–292. Varma, R.L., Bakhshi, N.N., Mathews, J.F., and Ng, S.N. (1987) Performance of dual-reactor system for conversion of syngas to aromatic-containing hydrocarbons. Ind. Eng. Chem. Res., 26, 183–188. Egiebor, N.O., Cooper, W.C., and Wojciechowski, B.W. (1989) Synthesis of motor fuels from HY-zeolite supported Fischer-Tropsch iron catalysts. Appl. Catal, 55, 47–64. Song, X. and Sayari, A. (1996) Sulfated zirconia as a cocatalyst in Fischer-Tropsch synthesis. Energy Fuels, 10, 561–565. Botes, F.G. and B¨ohringer, W. (2004) The addition of HZSM-5 to the Fischer-Tropsch process for improved gasoline production. Appl. Catal. A, 267, 217–225. Joshi, J.B., Pandit, A.B., Kataria, K.L., Kulkarni, R.P., Sawarkar, A.N., Tandon, D., Ram, Y., and Kumar, M.M. (2008) Petroleum residue upgrading via visbreaking: a review. Ind. Eng. Chem. Res., 47, 8960–8988. Leprince, P. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 365–379. Egloff, G. (1937) Reactions of Pure Hydrocarbons, ACS Monograph Series, Vol. 73, Reinhold, New York. De Klerk, A. (2007) Thermal cracking of Fischer-Tropsch waxes. Ind. Eng. Chem. Res., 46, 5516–5521. Dancuart, L.P., Mayer, J.F., Tallman, M.J., Adams, J., and Sasol, S.P.D. (2003) Performance of the naphtha as steam cracking feedstock. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 48 (2), 132–138. Dazeley, G.H. and Hall, C.C. (1948) Production of olefins by the cracking of Fischer-Tropsch waxes and their conversion into lubricating oils. Fuel, 27 (2), 50–55. Horne, W.A. (1950) Review of German synthetic lubricants. Ind. Eng. Chem., 42, 2428–2436.
333
334
16 Refining Technology Selection 140. Miller, S.J., Shah, N., and Huffman, G.P.
(2005) Conversion of waste plastic to lubricating base oil. Energy Fuels, 19, 1580–1586. 141. Swindell, R. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 381–407.
142. ASTM (2005) D 189 – 05. Standard test method
for conradson carbon residue of petroleum products, ASTM, West Conshohocken, PA.
335
17 Dehydration, Etherification, and Hydration 17.1 Introduction
The oxygenate refining technologies that are most relevant to Fischer–Tropsch syncrude all involve alcohols as either feed, intermediate, or product. In many cases, the reactions are reversible and equilibrium limited and the compounds that are interconverted are alcohols, ethers, alkenes, and water (Figure 17.1). Of these reactions, dehydration (Section 17.2), etherification (Section 17.3), and hydration (Section 17.4) are important Fischer–Tropsch refining technologies. Hydrolysis, which is the scission of a C–O bond by the reaction with water, is of minor importance in syncrude refining and it is not discussed separately. The oxygenate reaction network and conversion processes are also relevant to biomass refining, but in the case of biomass, hydrolysis is indeed very important. Unlike syncrude, biomass is rich in ether bonds (C–O–C), and hydrolysis enables depolymerization to produce shorter chain oxygenates that are more easily refined.
Alcohol dehydration − H2O
2 R
OH
+ H2O
2
R
Alkene hydration Ether + H2 O hydrolysis
− H2O
Alcohol etherification
Alcohol − H2O dehydration
+ H2 O
Alkene hydration
Ether dehydration
R
O
R
− H2O
R
OH +
+ H2O
R
Alkene etherification
Figure 17.1 Reaction network showing the interconversion of alcohols, ethers, alkenes, and water. This simplified reaction network does not show isomers or side reactions. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
336
17 Dehydration, Etherification, and Hydration
OH R
OH
+ H+ − H+
R
O
O + H+
+ H2O
R′ R′
(a)
− H+
(b)
R′
+ H 2O
R′
Figure 17.2 Dehydration reactions important to Fischer–Tropsch refining: (a) alcohol dehydration to produce an alkene and (b) dehydration of an aldol condensation product to produce an enone.
17.2 Dehydration
Dehydration is a deoxygenation reaction that removes water from an alcohol to produce an alkene. It is the simplest of the deoxygenation reactions and provides a clean chemistry to convert alcohols into hydrocarbons. Three dehydration reactions are important in Fischer–Tropsch refining: direct alcohol dehydration to alkenes, bimolecular partial alcohol dehydration to ethers, and aldol condensation of carbonyl compounds followed by dehydration to produce enones (Figure 17.2). Crude oil usually does not contain alcohols, and alcohol dehydration is not a refining technology that is found in crude oil refineries. This is in contrast to Fischer–Tropsch synthesis, where alcohols are primary products that are present in both the oil phase and the aqueous phase. The applications of alcohol dehydration in Fischer–Tropsch refining have already been touched on (Section 16.3.1): 1) Aqueous product refining can be simplified by dehydration to convert most of the oxygenates into hydrocarbons, which can then be refined with the oil product. The simplest approach is to subject the complete aqueous product to dehydration, but it is unlikely to be efficient due to the low oxygenate concentration, typically <10%. The efficiency can be improved by distillation of the aqueous product to obtain an oxygenate-rich overhead product and water-rich bottom product. The oxygenate-rich overhead product can be dehydrated without further separation (Figure 17.3). The alcohols will be converted into alkenes and some of the carbonyl compounds will be converted into enones and aromatics Dehydration Aqueous product
Hydrocarbons (to oil refinery)
Oxygenate rich
Water recycle Acid water
Wastewater
Figure 17.3 Dehydration of mixed alcohols and carbonyls in the Fischer–Tropsch aqueous product, with partial recycle of the water-rich product to increase deoxygenation and improve heat management.
17.2 Dehydration
Carbonyl aromatization Aromatics Aqueous product
Oxygenate rich
Carbonyl rich
Wastewater Alcohol dehydration
Acid water
Alkenes Alcohol rich
Wastewater Figure 17.4 Separate dehydration of the alcohol- and carbonyl-rich fractions in the Fischer–Tropsch aqueous product, with partial recycle of the water-rich product in each case to increase deoxygenation and improve heat management.
(Figure 16.4). The carbonyl conversion is bimolecular and is thus less efficient in a diluted mixture. The product water can be partly recycled to increase conversion and to improve heat management (Section 17.2.3). Conversion in this manner leads to incomplete deoxygenation, but deoxygenation is sufficient to produce an oil phase for separation from the water phase to simplify the refinery design. 2) A variation on complete dehydration involves a second separation step to achieve a rough separation of the alcohols and carbonyls (Figure 17.4). The alcohols can then be dehydrated to produce alkenes, while the carbonyls can be aromatized (Section 16.3.5). The latter conversion involves dehydration in combination with aldol condensation and benefits from the higher concentration of carbonyls in the feed to promote the required bimolecular reactions. Aromatics production from carbonyl compounds has been proposed as a fuel-refining pathway for Fischer–Tropsch syncrude [1], as well as biomass-derived products [2]. 3) Another variation on complete dehydration that has been suggested [3] involves partial hydrogenation to first convert all the carbonyl compounds into alcohols (Figure 17.5). The alcohol–water mixture can then be dehydrated to alkenes, achieving almost complete deoxygenation. One may also selectively separate some alcohols that are directly of value as fuels or chemicals before dehydrating the remainder of the alcohols. For example, one may want to separate methanol or ethanol as fuel alcohol or etherification feed before dehydrating of the heavier alcohols to their corresponding alkenes. 4) In an analogous manner to point 1, one may treat the syncrude oil phase to deoxygenate it by dehydration. Depending on the catalyst selection, deoxygenation takes place to different degrees, but it is generally incomplete. This approach has been extensively used in the past (Chapters 6–9) to deoxygenate and double bond isomerize the alkenes in the syncrude [4, 5]. It has also been evaluated for deoxygenation of syncrude without isomerization [6].
337
338
17 Dehydration, Etherification, and Hydration
Partial carbonyl hydrogenation Oxygenate rich
Aqueous product
Acid water
H2 Alcohol dehydration Alkenes Alcohol rich
Water recycle
Wastewater Figure 17.5 Partial hydrogenation of carbonyls in the Fischer–Tropsch aqueous product to alcohols followed by dehydration of mixed alcohols to produce alkenes. Partial recycling of the water-rich product is possible to increase deoxygenation and improve heat management.
5) Dehydration can be employed as a selective conversion step in chemical production. For such applications, the catalyst selection and operating conditions are critical in order to achieve high selectivity to the desired products. The dehydration of n-1-alcohols to produce n-1-alkenes [7] is practised commercially. 6) The relative ease of dehydration of alcohols can also be exploited for the selective dehydration of secondary and tertiary alcohols in the presence of primary alcohols [3]. This facilitates separation of alcohols for chemical production and reduces the complexity of the product for subsequent further separation by distillation. 7) Alcohol etherification is a partial dehydration reaction (Figure 17.1). It is useful for the production of fuel ethers [8, 9]. Of the oxygenate classes that can be added to diesel fuel to produce an oxygenated fuel, linear ethers were found to provide the best compromise between cetane number and cold-flow properties [10]. The properties of some of these fuel ethers that can be employed as diesel fuel extenders are given in Table 17.1 [9–11]. The C5 and heavier n-1-alcohols present in Fischer–Tropsch syncrude can be exploited to produce such diesel Table 17.1
Selected properties of diesel fuel range ethers prepared by partial dehydration of alcohols.
Compound
Boiling point (◦ C) Density at 20 ◦ C (kg·m−3 )
Blending properties in diesel fuel Cetane number Cloud point (◦ C) CFPP (◦ C)
Butoxybutane Pentoxypentane Hexoxyhexane Heptoxyheptane Octoxyoctane
142 187–190 228–229 259 286–287
CFPP, cold filter plugging point.
769 783 793 801 806
85 109 117/118 117 118
−20 −22/−20 −7/−5 −7 −17
−13 −20 −5 −5 −15
17.2 Dehydration
Dehydration/ dimerization Straight-run LTFT naphtha
Naphtha recycle
Alkane rich
Ethers + alkenes
Naphtha
Distillate Wastewater
Figure 17.6 Partial dehydration of alcohols to ethers in combination with alkene dimerization to increase distillate production from straight-run Fischer–Tropsch naphtha.
fuel extenders and was successfully tested as blending components in high-temperature Fischer–Tropsch (HTFT) diesel fuel [12]. Although methanol etherification to produce methoxymethane (DME, dimethyl ether) is less important in the Fischer–Tropsch context, it is nevertheless an important conversion of this type. Light alcohols may also be etherified to produce fuel ethers for motor-gasoline and 2-(1-methylethoxy)-propane (DIPE, diisopropyl ether) which is produced by 2-propanol etherification and is a high octane number fuel ether [13]. 8) Dehydration in combination with alkene dimerization as side reaction can be employed to convert naphtha range alcohols and alkenes into distillate range products (Figure 17.6) [9]. The conditions that favor alcohol etherification also favor alkene dimerization; both are bimolecular acid-catalyzed reactions. Partial dehydration of the alcohols to ethers is preferable, but in a syncrude matrix any dehydration product will just add to the alkenes available for dimerization. This application is especially well suited for improving the distillate yield during low-temperature Fischer–Tropsch (LTFT) syncrude refining. In this way, the dehydration process doubles as an oligomerization process (Chapter 19). There are many synergies and advantages: dehydration is endothermic and dimerization is exothermic; the alkane matrix improves heat management and reduces side reactions by dilution; both alkenes and oxygenate reactivity are beneficially exploited. 9) The etherification of alcohols in combination with hydration of alkenes in Fischer–Tropsch syncrude has been proposed as a method to separate and recover alcohols from the straight-run product [14]. By etherifying the alcohols, it is possible to recover the ethers by distillation, and the alcohols can then be regenerated by hydrolysis of the ethers. 10) Alcohol dehydration is an integral step in the low-temperature conversion of alkanes to alkenes [15]. 17.2.1 Reaction Chemistry
Dehydration is an acid-catalyzed reaction. The classic description of alcohol dehydration by a Brønsted acid involves protonation of the alcohol group (Figure 17.7). The protonated alcohol group (−OH2 + ) is a good leaving group and the intermediate undergoes heterolysis to yield water and a carbocation. The carbocation thus formed is the same as that obtained by alkene protonation and is capable of the same side reactions.
339
340
17 Dehydration, Etherification, and Hydration
OH R
R′
+
+ H+ − H+
R
OH2 R′
R
H2O + R′
− H+ + H+
R
R′ + H2O
H Figure 17.7
Mechanism for classic Brønsted acid-catalyzed alcohol dehydration.
The ease of alcohol dehydration follows the same order as the stability of the carbocation intermediate: tertiary > secondary > primary [16]. The temperature required to produce alkenes by alcohol dehydration increases with increasing carbon number, at least from C2 to C4 [3]. A monotonic increase in dehydration rate with carbon chain length was also reported by Stauffer and Kranich [17], but they found that the activation energy was constant at 129 kJ·mol−1 for the dehydration of the C2 –C6 n-1-alcohols. Although this seems contradictory, there is more than one dehydration pathway (Figure 17.1). The two-step dehydration through an ether intermediate is slower than direct dehydration to an alkene, although both pathways ultimately yield the same product. Alcohol dehydration is an endothermic reaction. The standard heat of reaction for the n-1-alcohols is around 35 kJ·mol−1 and for ethanol it is 45 kJ·mol−1 . There is consequently a significant adiabatic temperature decrease associated with dehydration, and heat management is very important during industrial operation (Section 17.2.3). Dehydration is a reversible reaction, but the equilibrium favors dehydration over hydration (Section 17.4) even in the presence of an excess of water. As can be seen from the basic reaction network (Figure 17.1), alcohols can be dehydrated directly to alkenes, or can be partially dehydrated to ethers. Ethers are an intermediate product and the etherification yield passes through a maximum as alcohol conversion is increased. Only the primary alcohols form ethers in high yield under suitable conditions. This is related to the stability of the carbocation intermediate. According to the classic reaction mechanism (Figure 17.7), dehydration of a primary alcohol yields a primary carbocation, which rapidly rearranges to a more stable intermediate. The formation of a primary carbocation can be avoided by alcohol etherification, and in the case of primary alcohols there is a driving force to promote etherification. At this point, it should be pointed out that the classic mechanism does not provide an accurate mechanistic picture of alcohol dehydration over some important catalysts (Section 17.2.2). Nevertheless, the stability of the carbocation intermediate is a good indicator of the etherification propensity of an alcohol. Tertiary alcohols readily form stable carbocation intermediates and are easily dehydrated directly to alkenes. Even under conditions of high pressure and low temperature that would favor alcohol etherification, it is difficult to obtain ethers by dehydration of tertiary alcohols. It is somewhat easier to obtain ethers from secondary alcohols, but not as easily as from primary alcohols. 17.2.2 Catalysis
Over the years, a large number of homogeneous and heterogeneous catalysts have been investigated for the dehydration of alcohols [18]. The nature of the reaction requires the catalyst to be stable in the presence of water. A review of the literature on alcohol dehydration indicates that the catalysts most often employed for dehydration are the following:
17.2 Dehydration
1) Amorphous metal oxide catalysts. Alumina is the dominant catalyst of this type and it has been extensively studied and industrially used for the conversion of Fischer–Tropsch alcohols [19]. 2) Acidic resin catalysts. Although these catalysts are inhibited by water, they are capable of alcohol dehydration at low temperatures and have been extensively studied for the partial dehydration of alcohols to produce ethers [20]. 3) Zeolite catalysts. Some studies on zeolite conversion of heavier alcohols can be found, but the main application of zeolites for alcohol dehydration is for methanol conversion over H-ZSM-5 [21]. The catalyst influences the mechanism of alcohol dehydration, and of specific importance is how the intermediate is stabilized. Fischer–Tropsch syncrude contains mainly n-1-alcohols. The implication that a primary carbocation intermediate is involved in dehydration (Figure 17.7), even fleetingly, is doubtful. A more in-depth look at the dehydration mechanism over alumina shows that no primary carbocation intermediate is formed and that the surface intermediate is indeed capable of desorbing 1-alkenes from 1-alcohols, as is found in practice [19, 22]. Oxygenates and water adsorb more strongly onto the polar surface of acidic catalysts than hydrocarbons. The more polar compounds in the reaction mixture inhibit the acid-catalyzed side reactions of alkenes. As alcohol conversion increases, the partial pressure of alkenes also increases and thereby the possibility of side reactions. Since water is also an inhibitor, side reactions involving alkenes are usually limited. Double bond isomerization may still take place because of adsorbed water, rather than successful competitive alkene adsorption, through the hydration-dehydration sequence (Equation 17.1): 1-Alkene + H2 O → 2-alcohol → 2-alkene + H2 O
(17.1)
It is important to suppress this reaction in some chemical applications of alcohol dehydration. Although this is an acid-catalyzed reaction, it occurs on the same site as alcohol dehydration, which is a reversible reaction. Strong acid sites may be neutralized, but the introduction of more basic sites may promote alcohol dehydrogenation as side reaction (Equation 17.2): Alcohol (RCH2 OH) → carbonyl (RCHO) + H2
(17.2)
Some acid catalysts have acid–base site pairs that catalyze dehydrogenation to produce carbonyl compounds. The carbonyl compounds can successfully compete with the other oxygenates to adsorb on the catalyst and may give rise to a myriad of side reactions [23]. With proper catalyst selection, such side reactions can be minimized, but side reactions involving oxygenates are difficult to completely avoid. 17.2.3 Syncrude Process Technology
The design or selection of an appropriate process technology for the dehydration of Fischer– Tropsch alcohols is strongly influenced by the desired product. The catalyst choice is crucial, since it determines the operating range, selectivity profile, side reactions, and deactivation behavior. The following aspects are important to bear in mind for alcohol dehydration:
341
342
17 Dehydration, Etherification, and Hydration
1) Alcohol etherification benefits from high-pressure or liquid-phase operation at low temperature. When the aim is to produce ethers, select an acid catalyst capable of operating at low temperature (<150 ◦ C), such as an acidic resin. Acidic resin catalysts are unfortunately inhibited by water, requiring continuous water removal or low per pass conversion during the engineering design. As such, resin catalysts are better at alcohol dehydration in the oil phase, where the water and oxygenate content of the syncrude is lower. Resin catalysts are also employed as selective dimerization catalysts, and dimerization is moderated by polar compounds [24, 25]. However, resin catalysts are known to selectively convert branched alkenes only even when in the presence of n-alkenes. Although this suggests that resin catalysts may not be suitable for n-alkene conversion in syncrude, acidic resins are quite capable of converting n-alkenes, albeit at a lower rate [26]. Acidic-resin-catalyzed dehydration is recommended for use with LTFT naphtha. Using such catalysts with HTFT naphtha requires more care due to the high alkene and nonalcohol oxygenate content, which may lead to deactivation. 2) Complete alcohol dehydration benefits from high temperature and low pressure. If the alcohol content in the feed is high, the catalyst must have a high tolerance for water. Alumina is a proven catalyst for Fischer–Tropsch aqueous product dehydration. The required operating temperature is determined by the feed composition, but is usually >300 ◦ C for complete dehydration [3]. Ethanol is the most difficult of the alcohols to dehydrate and requires a higher operating temperature to achieve a similar space–time dehydration yield as the heavier alcohols. 3) Alcohol dehydration is endothermic. Heat management is critical, and the engineering design must make provision for interstage heating and lower per pass conversion with recycling, or the feed must contain nonalcohol material to provide sufficient heat capacity to limit the adiabatic temperature decrease (Figure 17.8). There is a trade-off between the reactor size increase due to water co-feeding to limit the temperature decrease associated with complete alcohol conversion and the added complexity of the design to accommodate interstage heating or product separation with recycling. Although pure ethanol recycling requires the smallest reactor volume, one of the main advantages of dehydration is that it can avoid alcohol–water separation. In this respect, complete alcohol conversion with only water recycling is advantageous. 4) Small quantities of side products may be formed during dehydration. At high temperature over amorphous metal oxide catalysts and zeolites, some dehydrogenation may take place. The carbonyl compounds may in turn be converted into carboxylic acids. Carboxylic acids not only affect the material selection of the processing equipment but also build up in the water from the dehydration reaction when it is recycled [3]. Analogous problems are encountered if the Fischer–Tropsch feed contains carbonyls or carboxylic acids. 5) Water inhibits the formation of carbonaceous deposits, and alumina is quite stable during alcohol dehydration. At high per pass conversion in the absence of a diluent, carbonaceous deposits may form and will ultimately cause catalyst deactivation. When using more acidic catalysts than alumina, such as zeolites, the same principles apply but the risk of deactivation is higher [27]. Fischer–Tropsch feeds usually contain other oxygenates and alkenes that can accelerate deactivation. The engineering design should make provision for regeneration or catalyst replacement in line with the feed, catalyst choice, and operating philosophy.
17.3 Etherification
430
Temperature (°C)
410 Pure ethanol, fresh:recycle = 1 : 5 390 Water:ethanol = 1 : 1 Fresh:recycle = 1 : 2
370
350
Pure ethanol, no recycle
Water:ethanol = 10 : 1, No ethanol recycle
330 0
20
40 60 Overall ethanol conversion (%)
80
100
Figure 17.8 Adiabatic temperature decrease during ethanol dehydration as function of overall conversion (not per pass conversion) for different recycle configurations. All ratios indicated are molar ratios.
17.3 Etherification
Etherification refers to the formation and an ether bond. Alcohol etherification, which entails the formation of an ether bond by the partial dehydration of two alcohol molecules, was discussed in the previous section (Section 17.2). In this section, the discussion is limited to etherification by the reaction of an alcohol with an alkene (Figure 17.1). In crude oil refining, the etherification of alkenes with alcohols is employed to produce fuel ethers for oxygenated motor-gasoline [13]. The use of ethers as high octane blending components was discussed (Section 2.4.4 and Chapter 13), and the most common fuel ethers are shown in Figure 17.9. The production of fuel ethers from heavier than C5 alkenes that are present in the light naphtha from fluid catalytic cracking of heavy crude oil fractions [28, 29], and fluid catalytic cracking of Fischer–Tropsch waxes [30], has been reported. Not all ethers perform equally well as blending components in Fischer–Tropsch syncrude, and it was shown that the branching structure of the alkene plays an important role in the octane number that can be achieved (Table 17.2) [31]. The application of etherification in Fischer–Tropsch refining is limited by the low concentration of branched alkenes present in straight-run syncrude. The two main industrial O
O
2-Methoxy-2methylpropane (MTBE) Figure 17.9
2-Ethoxy-2methylpropane (ETBE)
O 2-Methoxy-2 methylbutane (TAME)
O 2-Ethoxy-2methylbutane (TAEE)
Fuel ethers commonly used as high octane number motor-gasoline blending components.
343
344
17 Dehydration, Etherification, and Hydration Table 17.2 Selected physical properties and motor-gasoline blending properties of potential fuel ethers with HTFT gasoline and a primary standard gasoline (40% n-heptane; 60% 2,2,4-trimethylpentane). Blending properties were determined using 20% ether blends.
Compound
Ether propertiesa
HTFT gasoline blendb
Boiling Density RON MON RVP point (◦ C) (kg·m−3 ) (kPa) No ether addition Ethoxyethane (DEE) 2-Ethoxypropane 2-(1-Methylethoxy)-propane (DIPE) 2-(1-Methylpropoxy)-2-butane 2-Methoxy-2-methylpropane (MTBE) 2-Ethoxy-2-methylpropane (ETBE) 2-Methoxy-2-methylbutane (TAME) 2-Ethoxy-2-methylbutane (TAEE) 2-Propoxy-2-methylbutane 2-(1-Methylethoxy)-2-methylbutane 2-Methoxy-2-methylpentane 3-Methoxy-3-methylpentane 2-Methoxy-2,3-dimethylbutane 2-Methoxy-2-methylheptane 3-Methoxy-3,5-dimethylhexane
– 34 54 76 113 55 67 85 93 116 107 105 105 103 145 137
– 717 726 725 763 742 743 744 768 774 766 781 794 795 796 805
82 −11 34 102 86 114 117 108 104 102 113 85 102 110 52 98
76 0 64 106 93 106 110 108 103 96 110 86 99 99 57 94
48 111 55 32 0 52 32 15 0 0 −1 −2 −3 −5 −4 −5
Primary standard blend RON
MON
RVP (kPa)
60 −40 −40 98 58 145 134 126 102 96 121 69 97 111 14 89
60 −40 15 125 86 135 145 145 125 116 130 87 118 123 36 103
14 111 57 35 1 52 34 17 1 1 0 0 0 −1 0 −1
a
Physical properties as determined experimentally and reported in the source reference. HTFT gasoline consisted of a blend containing 40% reformate (RON 81), 13% C5 –C6 straight-run naphtha, 20% C5 –C6 isomerate, 25% hydrogenated SPA < 150 ◦ C naphtha, and 2% creosote naphtha (see Tables 9.2 and 9.5 for details). b The
applications of etherification in conjunction with Fischer–Tropsch syncrude has been mentioned before (Section 16.3.2): 1) The production of fuel ethers for addition to motor-gasoline is similar to that found in crude oil refineries. Alkenes and alcohols are available as straight-run products from Fischer–Tropsch syncrude, but enable fuel ether production only on a small scale. A skeletal isomerization unit (Chapter 18) is typically required to increase the branched alkene content of the syncrude in order to increase ether production. However, the amount of alcohols that can be recovered from the Fischer–Tropsch aqueous product may be insufficient to meet the demands of a large etherification unit. 2) Etherification is industrially employed in the purification of 1-pentene and 1-hexene from HTFT syncrude (Sections 9.5.4 and 23.2.1). In the case of 1-pentene, it is used to convert 2-methyl-1-butene, which is close boiling to 1-pentene, and in the case of 1-hexene it is used to convert 2-methyl-1-pentene and 2-ethyl-1-butene, which are close boiling to 1-hexene [32]. By converting the close boiling isomers to their corresponding ethers, purification is facilitated. In an analogous way, etherification can also be employed in the purification of 1-butene.
17.3 Etherification
17.3.1 Reaction Chemistry
The reaction of an alcohol with an alkene is acid catalyzed and equilibrium limited. The reaction is exothermic, with standard heat of reaction around −35 to −40 kJ·mol−1 for the etherification of isobutene. The etherification equilibrium is favored by a lower temperature (Table 17.3) [33–36]. Alcohol addition to the alkene follows the Markovnikov rule, which states that the acid hydrogen will be attached to the carbon with the highest amount of hydrogen atoms. This is just a different way of saying that the carbocation formed by protonation of the alkene by the acid will be the most stable; the charge will be carried on the most substituted carbon. The mechanism is a textbook electrophilic addition reaction mechanism (Figure 17.10). Since the reaction chemistry involves a carbocation intermediate, it is anticipated that the reactivity of alkenes will follow the same order as predicted from carbocation stability. This is indeed the case. Etherification of alkenes takes place only when the carbon double bond of the alkene is on a tertiary carbon. It is therefore possible to selectively convert branched alkenes meeting this requirement in a mixture of alkenes. It is also possible to perform etherification with minimal acid-catalyzed side reactions, even when the feed mixture has a high concentration of alkenes. Equilibrium constants for the liquid-phase etherification of C4 –C5 alkenes with methanol and ethanol.
Table 17.3
Etherification product
Equilibrium constant at different temperatures
2-Methoxy-2-methylpropane (MTBE) 2-Ethoxy-2-methylpropane (ETBE) 2-Methoxy-2-methylbutane (TAME) Methanol + 2-methyl-1-butene Methanol + 2-methyl-2-butene 2-Ethoxy-2-methylbutane (TAEE) Ethanol + 2-methyl-1-butene Ethanol + 2-methyl-2-butene
50 ◦ C
60 ◦ C
70 ◦ C
80 ◦ C
118 ± 9 58 ± 10
70 ± 5 41 ± 7
48 ± 4 26 ± 3
32 ± 2 18 ± 1
21 ± 1 13 ± 2
[33] [34]
65 ± 5 5.8 ± 0.4
40 ± 3 4.1 ± 0.3
31 ± 2 3.2 ± 0.2
21 ± 1 2.4 ± 0.2
16 ± 1 1.9 ± 0.1
[35] [35]
45 3.4
31 2.6
21 1.9
15 1.5
– –
[36] [36]
+ H+
OH
− H+
OH2+
OH
− H+
+
HO
+ H+
Reference
40 ◦ C
OH2+
Figure 17.10 Mechanism of etherification of alkenes by alcohols, as illustrated by the acid-catalyzed reaction of methanol with 2-methylbutenes to produce 2-methoxy-2-methylbutane (TAME).
− H+
OH+
+ H+
O
345
346
17 Dehydration, Etherification, and Hydration
Etherification of alkenes with alcohols exploits one of the characteristics of acid catalysis in the presence of oxygenates. The alcohols are more polar and stronger nucleophiles than the alkenes. The alcohols outcompete the alkenes for interaction with the acid catalyst and are preferentially protonated. Alcohols effectively inhibit protonation of the alkenes by the acid, in addition to being reagents for the reaction. This has two important consequences. The first is that the alcohol becomes the proton carrier. The acidity of the protonated alcohol (alcoxonium ion) is lower than that of the acid catalyst, thereby moderating the reaction severity in an otherwise very reactive feed. The second consequence is that the protonation of the alkene by the protonated alcohol by definition takes place in close proximity to the alcohol. Once the alcohol has protonated the alkene, the alcohol becomes a nucleophile and the probability of nucleophilic addition of the alcohol to the carbocation is high. Acid-catalyzed side reactions become significant only when the alcohol concentration becomes too low to prevent direct protonation of the alkenes by the acid catalyst. It has been shown that, with decreasing alcohol concentration, the mechanism gradually changes from protonation of the alkenes by a protonated alcohol to protonation directly by the catalyst [37]. In the latter case, the carbocation is not formed in association with an alcohol, and the reaction probability is determined by the concentration of different species in the mixture. In Fischer–Tropsch syncrude there is a further possible complication, namely, the presence of other oxygenate classes than just alcohols. The nonalcohol oxygenate classes can successfully compete with the alcohols for interaction with the acid catalyst, result in side reactions and lower the activity coefficient of the alcohol used for etherification [38]. The nett effect is that the oxygenates present in syncrude inhibits etherification and increases the formation of side products. Water can also successfully compete with alcohols to be protonated, which then leads to hydration of the alkene [39]. Alkene hydration will be discussed separately (Section 17.4). Even when the feed is water-free, some alcohol etherification by partial alcohol dehydration (Section 17.2) can take place to produce water. It is clear that there is a trade-off between a high alcohol concentration to limit alkene-related side reactions and an increase in alcohol-related side reactions. 17.3.2 Catalysis
The industrial application of etherification is dominated by acidic resin catalysts and mainly sulfonic acid copolymers of polystyrene and divinylbenzene [13]. Acidic resin catalysts perform well at low temperature, which favors the reaction equilibrium for etherification (Table 17.3). The importance of a sufficiently high alcohol concentration to ensure reaction by a protonated alcohol has already been mentioned. With Fischer–Tropsch syncrude, it has been found that a methanol to reactive alkene ratio of more than 2 : 1 is required during etherification over acidic resin catalysts to limit side reactions and specifically gum formation, which leads to catalyst deactivation [32]. The alcohols also swell the resin catalyst to improve accessibility and moderate acid strength by solvating the hydrogen-bonded sulfonic acid groups [40]. Etherification can be performed using other acidic catalysts, but most other acid catalyst types cannot compete with the activity of acidic resins at low temperature. Of the zeolites, only zeolite H-Beta has comparable activity for etherification to acidic resins at low temperature [41].
17.4 Hydration
17.3.3 Syncrude Process Technology
The process technology employed for the etherification of Fischer–Tropsch-derived alkenes is similar to that employed in crude oil refining. Fixed bed or catalytic distillation technology can be employed. Some aspects specific to syncrude that need to be considered in the detailed design are the following: 1) The syncrude feed is likely to contain other oxygenates and water. Even when such compounds are present in low concentration, it will result in side product formation and cause some inhibition of the etherification reaction [38]. 2) Aldehydes and ketones are capable of aldol condensation to form heavier products (gums). Aldol condensation can be catalyzed by both acids and bases. Sulfonic acid groups involved in acid catalysis can, in their deprotonated form, catalyze aldol condensation by abstracting the acidic α-hydrogen from carbonyl compounds. The involvement of an acid site in catalysis may rule out other acid-catalyzed side reactions, but it does not preclude the deprotonated site from catalyzing aldol condensation. 3) Gum formation can be suppressed by having a sufficiently high alcohol content in the feed. It has been reported that prolonged etherification of syncrude at an alcohol to reactive alkene ratio of 0.5 resulted in gradual catalyst deactivation by gum formation [32]. Such heavy products can be formed by aldol condensation and alkene oligomerization. 4) The swelling of acidic resin catalysts by alcohols can likewise be caused by other polar compounds. There is a natural partitioning of oxygenates between the bulk liquid phase and the resin, analogous to the organic–aqueous phase partitioning found during syncrude recovery in the gas loop (Section 5.3.3). When nonalcohol oxygenates are present in the syncrude, the resin catalyst will gradually swell to a volume larger than anticipated from just alcohol-based swelling. Unless sufficient reactor volume is allowed for such additional expansion, catalyst swelling may result in equipment damage. 5) Double bond isomerization takes place during etherification and is a natural consequence of the mechanism (Figure 17.10). Although etherification through a secondary carbocation intermediate is very limited, some double bond isomerization may still take place. In applications sensitive to n-alkene isomerization, it is detrimental to operate at low space velocity. The reaction is slow, but double bond isomerization of the n-alkenes will gradually take place.
17.4 Hydration
Water addition to alkenes, or the hydration of alkenes, is analogous to etherification. Instead of an alcohol, water is used as ‘‘etherification’’ agent. Hydration is not a conventional crude oil refining technology, but it is employed for petrochemical production. The hydration of ethene to ethanol [42], and the hydration of propene to 2-propanol [43], are both practised industrially [44]. One would not normally consider hydration as a fuel-refining technology. There is one exception. When ethene is produced in significant quantities in a facility that is remote and does not have access to downstream petrochemical consumers of ethene, a refining pathway
347
348
17 Dehydration, Etherification, and Hydration
must be provided. In such circumstances, hydration can be employed as conversion technology to refine ethene to ethanol. Ethanol is a liquid product that is easily transportable and can be used as transportation fuel, commodity chemical, or as an intermediate in processes such as etherification (Section 17.3). Hydration is of course not the only technology that can be employed to refine ethene to a liquid product. Aromatic alkylation to produce alkyl aromatics (Chapter 20) and alkene oligomerization to produce heavier alkene oligomers (Chapter 19) can also be considered, depending on the refining objectives and design. In a Fischer–Tropsch facility, hydration has some benefits though, as given below: 1) Ethanol is a product from Fischer–Tropsch synthesis. By producing more ethanol, the economy of scale for ethanol refining is improved, without adding additional complexity to the ethanol refining pathway in the design. 2) Side products from ethene hydration are mainly ethoxyethane, ethanal, ethene oligomers and traces of carboxylic acids, and ketones [42]. These side products all have analogs in syncrude. 3) Although the side products produced during ethene hydration are minor in quantity, they have an impact on process efficiency and ethanol purity. The additional downstream processing that is required to ensure adequate ethanol quality adds much complexity to a stand-alone hydration plant [44]. In a Fischer–Tropsch refinery, much of this added
Ethene recycle
Ethene Water
Alkenes
Ethene hydration
Ethoxyethane
Ethanol, water, oxygenates
Partial carbonyl hydrogenation
Aqueous product
H2
Oxygenate rich
Alcohol-rich mixture
Alcohol rich Water recycle Acid water
Figure 17.11 Integration of ethene hydration with a Fischer–Tropsch aqueous product refinery to produce an alcohol-rich mixture.
Wastewater
17.4 Hydration
complexity can be avoided by integrating the ethene hydration unit with the aqueous product refinery (Figure 17.11). 17.4.1 Reaction Chemistry
The chemistry of alkene hydration is analogous to that of alkene etherification. Protonated water (H3 O+ ) is an effective proton carrier, and alcohols are formed instead of ethers. Hydration is equilibrium limited, and the equilibrium favors dehydration (Section 17.2.1). The reaction is exothermic and the hydration equilibrium is favored by low temperature. From a kinetic point of view, hydration has one advantage over dehydration at low temperature, namely that hydration yields the alcohol directly, but that dehydration may proceed by partial dehydration to yield an ether. The ether can be further dehydrated, but it may also be hydrolyzed (Figure 17.1). This does not change the equilibrium conversion, but it holds some practical advantage for hydration. The equilibrium constant for ethene hydration to ethanol, Khyd , rapidly increases with a decrease in temperature T (K) (Equations 17.3 and 17.4) [45, 46]. Liquid phase: ln Khyd = 5460/T − 15.3
(17.3)
Gas phase: ln Khyd = 2100/T − 6.195
(17.4)
The equilibrium constant for ethanol etherification by partial dehydration to ethoxyethane, Kether , likewise increases with a decrease in temperature (Equation 17.5) [47]. Liquid phase: ln Kether = 2580/T − 3.14
(17.5)
The classic Brønsted acid-catalyzed reaction mechanism of dehydration (Figure 17.2) and by definition the reverse reaction, that is, hydration, requires the formation of a primary carbocation intermediate for ethene hydration. This is unlikely and it is therefore anticipated that an acid catalyst that will be able to successfully hydrate ethene, preferably at low temperature, should also be able to avoid the formation of a primary carbocation intermediate. 17.4.2 Catalysis
Of the many heterogeneous catalysts that have been evaluated for ethene hydration, phosphoric acid is favored for industrial hydration [18, 46]. As anticipated, this is due to the ability of phosphoric acid to avoid a primary carbocation intermediate. Vapor-phase hydration over phosphoric acid does not proceed through a phosphoric acid ester intermediate, unlike most other alkene reactions. The ethene reacts with the hydrated phosphoric acid to produce a loosely associated intermediate to directly yield the alcohol (Figure 17.12) [19]. Ethene can of course also form an ethyl phosphoric acid ester. Ethyl phosphoric acid esters are intermediates for oligomerization, which is a side reaction during dehydration. Under very dilute acidic conditions, the ethyl phosphoric acid ester may be hydrolyzed to liberate ethanol, but this reaction pathway is not significant under practical hydration conditions.
349
350
17 Dehydration, Etherification, and Hydration
H 2C
CH2 H 2C
HO
H
O
O
H
P
O
H
HO
OH Figure 17.12 Table 17.4
Catalyst
H
O
O
H
P
O
CH2
CH2
H
HO
OH
H
O
HO
O
H
O
P
O
HO
P
OH
OH
OH
Mechanism of ethene hydration over hydrated phosphoric acid to produce ethanol. Typical operating conditions for hydration of ethene and propene over different acid catalysts. Alkene
Operating conditions Temperature (◦ C)
H3 PO4 /SiO2 H3 PO4 /SiO2 W2 O5 Acidic resin
H 3C
H 3C
CH2
Ethene Propene Propene Propene
290 180 270 130–150
Pressure (MPa)
Water:alkene
7 1 25 6.5–10
0.6 : 1 1:1 15 : 1 12.5–15 : 1
Hydration performance Alkene conversion (%)
Alcohol selectivity (%)
4.7 5.7 47 75
97 96 98.8 96.5
The main disadvantage of phosphoric acid on silica as catalyst is the loss of phosphoric acid due to vaporization. Although the actual amount is low, it still requires constant phosphoric acid addition and a downstream design that can deal with the phosphoric acid. Propene is easier to hydrate that ethene, because it contains a secondary carbon. In addition to phosphoric acid supported on silica (H3 PO4 /SiO2 ), other catalysts are also employed for hydration, such as tungsten oxides (WO3 and W2 O5 ) and acidic resins. Typical operating conditions for the different catalysts are given in Table 17.4 [46]. 17.4.3 Syncrude Process Technology
There is no difference between the ethene hydration technology employed for crude-oil-derived ethene and Fischer–Tropsch-derived ethene. The main difference is in the integration opportunities that exist in a Fischer–Tropsch refinery (Figure 17.11), which are not available in a crude oil refinery. References 1. Nel, R.J.J. and De Klerk, A. (2009) Overview
of Fischer-Tropsch aqueous product refining strategies. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 118–119. 2. Hoang, T., Zhu, X., Lobban, L.L., Mallinson, R.G., and Resasco, D.E. (2010) Catalytic
conversion of small oxygenates from bio-oil to alkylaromatics on zeolites. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 55 (2), 331–332. 3. Nel, R.J.J. and De Klerk, A. (2007) Fischer-Tropsch aqueous phase refining by
References
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
catalytic alcohol dehydration. Ind. Eng. Chem. Res., 46, 3558–3565. Helmers, C.J., Clark, A., and Alden, R.C. (1948) Catalytic treatment of synthetic gasoline. Oil Gas J., 47 (26), 86–92. Schlesinger, M.D. and Benson, H.E. (1955) Upgrading Fischer-Tropsch products. Ind. Eng. Chem., 47, 2104–2108. Bolder, F.H.A. and Mulder, H. (2006) Dehydration of alcohols in the presence of carbonyl compounds and carboxylic acids in a Fischer-Tropsch hydrocarbons matrix. Appl. Catal. A, 300, 36–40. McGurk, K. (2003) From 1-heptene to 1-octene: a new production route. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa. p. cd082. Marchionna, M., Patrini, R., Giavazzi, F., and Pecci, G.C. (1996) Linear ethers as high quality components for reformulated diesel fuels. Prepr. Pap.-Am. Chem. Soc., Div. Pet. Chem., 41 (3), 585–589. Nel, R.J.J. and De Klerk, A. (2009) Dehydration of C5 -C12 linear 1-alcohols over η-alumina to fuel ethers. Ind. Eng. Chem. Res., 48, 5230–5238. Pecci, G.C., Clerici, M.G., Giavazzi, F., Ancilotti, F., Marchionna, M., and Patrini, R. (1991) Oxygenated diesel fuels. Part 1 – structure and properties correlation. 9th International Symposium on Alcohols Fuels, pp. 321–326. Bringu´e, R., Tejero, J., Iborra, M., Fit´e, C., Izquierdo, J.F., and Cunill, F. (2008) Study of the chemical equilibrium of the liquid-phase dehydration of 1-hexanol to dihexyl ether. J. Chem. Eng. Data, 53, 2854–2860. Van Heerden, J., Botha, J.J., and Roets, P.N.J. (1998) Improvement of diesel performance with the addition of linear ethers to diesel fuels. 12th International Symposium on Alcohols Fuels, pp. 188–199. Travers, P. (2001) in Petroleum Refining, Conversion Processes, vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 291–319. O’Rear, D.J., Harris, T.V., and Chen, C.-Y. (2008) Recovery of alcohols from Fischer-Tropsch naphtha and distillate fuels containing the same. US Patent 7, 402,187, July 22, 2008. Bolder, F.H.A., De Klerk, A., and Visagie, J.L. (2009) Hydrogenation of oxidized wax and a process to produce olefins from paraffins by
16.
17.
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
autoxidation, selective hydrogenation, and dehydration. Ind. Eng. Chem. Res., 48, 3755–3760. Morrison, R.T. and Boyd, R.N. (1987) Organic Chemistry, 5th edn, Allyn and Bacon, Boston, pp. 287–293. Stauffer, J.E. and Kranich, W.L. (1962) Kinetics of the catalytic dehydration of primary alcohols. Ind. Eng. Chem. Fundam., 1 (2), 107–111. Winfield, M.E. (1960) in Catalysis, Oxidation, Hydration, Dehydration and Cracking Catalysts, vol. VII (ed. P.H. Emmett), Reinhold, New York, pp. 93–182. De Klerk, A. (2011) Key catalyst types for the efficient refining of Fischer– Tropsch syncrude: alumina and phosphoric acid, in Catalysis, vol. 23 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge, UK, pp. 1–49. Tejero, J., Fit´e, C., Iborra, M., Izquierdo, J.F., Bringu´e, R., and Cunill, F. (2006) Dehydration of 1-pentanol to di-n-pentyl ether catalyzed by a microporous ion-exchange resin with simultaneous water removal. Appl. Catal. A, 308, 223–230. Chang, C.D. (1983) Hydrocarbons from methanol. Catal. Rev.-Sci. Eng., 25, 1–118. Shi, B. and Davis, B.H. (1995) Alcohol dehydration: mechanism of ether formation using an alumina catalyst. J. Catal., 157, 359–367. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer-Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Honkela, M.L. and Krause, A.O.I. (2003) Influence of polar components in the dimerization of isobutene. Catal. Lett., 87, 113–119. Talwalkar, S., Chauhan, M., Aghalayam, P., Qi, Z., Sundmacher, K., and Mahajani, S. (2006) Kinetic studies on the dimerization of isobutene with ion-exchange resin in the presence of water as a selectivity enhancer. Ind. Eng. Chem. Res., 45, 1312–1323. De Klerk, A. (2006) Reactivity differences of octenes over solid phosphoric acid. Ind. Eng. Chem. Res., 45, 578–584. Aguayo, A.T., Gayubo, A.G., Atutxa, A., Olazar, M., and Bilbao, J. (2002) Catalyst deactivation by coke in the transformation of aqueous ethanol into hydrocarbons. Kinetic modeling and acidity deterioration of the catalyst. Ind. Eng. Chem. Res., 41, 4216–4224. Wyczesany, A. (1995) Chemical equilibria in the process of etherification of light FCC gasoline with methanol. Ind. Eng. Chem. Res., 34, 1320–1326.
351
352
17 Dehydration, Etherification, and Hydration 29. Rihko, L.K. and Krause, A.O.I. (1996) Etherifica-
30.
31.
32.
33.
34.
35.
36.
tion of FCC light gasoline with methanol. Ind. Eng. Chem. Res., 35, 2500–2507. Reagan, W.J. (1994) Gasoline range ether synthesis from light naphtha products of fluid catalytic cracking of Fischer-Tropsch wax. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 39 (2), 337–342. Marais, Z.-M. (1991) Sintese en evaluering van’n reeks petrolmengkomponente. (Engl. Transl. ‘‘Synthesis and evaluation of a series gasoline blending components’’) M.Sc. Dissertation, University of the Orange Free State, Bloemfontein, South Africa. De Klerk, A. (2004) Etherification of C6 Fischer-Tropsch material for linear α-olefin recovery. Ind. Eng. Chem. Res., 43, 6349–6354. Izquierdo, J.F., Cunill, F., Vila, M., Tejero, J., and Iborra, M. (1992) Equilibrium constants for methyl tert-butyl ether liquid-phase synthesis. J. Chem. Eng. Data, 37, 339–343. Izquierdo, J.F., Cunill, F., Vila, M., Iborra, M., and Tejero, J. (1994) Equilibrium constants for methyl tert-butyl ether and ethyl tert-butyl ether liquid-phase synthesis using C4 olefinic cut. Ind. Eng. Chem. Res., 33, 2830–2835. Rihko, L.K., Linnekoski, J.A., and Krause, A.O.I. (1994) Reaction equilibria in the synthesis of 2-methoxy-2-methylbutane and 2-ethoxy-2-methylbutane in the liquid phase. J. Chem. Eng. Data, 39, 700–704. Kitchaiya, P. and Datta, R. (1995) Ethers from Ethanol. 2. Reaction equilibria of simultaneous tert-amyl ethyl ether synthesis and isoamylene isomerisation. Ind. Eng. Chem. Res., 34, 1092–1101.
37. Ancillotli, F., Mauri, M.M., Pescarollo, E., and
38.
39.
40.
41.
42.
43.
44.
45.
46. 47.
Romagnoni, L. (1978) Mechanisms in the reaction between olefins and alcohols catalyzed by ion exchange resins. J. Mol. Catal., 4, 37–48. Smook, D. and De Klerk, A. (2006) Inhibition of etherification and isomerization by oxygenates. Ind. Eng. Chem. Res., 45, 467–471. Linnekoski, J.A., Krause, A.O.I., and Struckmann, L.K. (1998) Etherification and hydration of isoamylenes with ion exchange resin. Appl. Catal. A, 170, 117–126. Thornton, R. and Gates, B.C. (1974) Catalysis by matrix-bound sulfonic acidic groups: olefin and paraffin formation from butyl alcohols. J. Catal., 34, 275–287. Goodwin, J.G. Jr., Natesakhawat, S., Nikolopoulos, A.A., and Kim, S.Y. (2002) Etherification on zeolites: MTBE synthesis. Catal. Rev.-Sci. Eng., 44, 287–320. Miller, S.A. (ed.) (1969) Ethylene and its Industrial Derivatives, Ernest Benn, London, pp. 690–801. Fielding, J.C. (1973) in Propylene and its Industrial Derivatives (ed. E.G. Hancock), Ernest Benn, London, pp. 214–234. Elkin, L.M., Fong, W.S., and Morse, P.L. (1979) Synthetic Ethanol and Isopropanol. SRI Process Economics Program Report 53A, Stanford Research Institute, Menlo Park, CA. Cope, C.S. and Dodge, B.F. (1959) Equilibria in the hydration of ethylene at elevated pressures and temperatures. AIChE J., 5, 10–16. Izumi, Y. (1997) Hydration/hydrolysis by solid acids. Catal. Today, 33, 371–409. Cope, C.S. (1964) Equilibria in the hydration of ethylene and of propylene. AIChE J., 10, 277–281.
353
18 Isomerization 18.1 Introduction
Isomerization manipulates the structure of a molecule, without changing its carbon number or molecular mass. In a refinery, isomerization is employed to improve one or more properties of a fraction, without significantly changing its boiling point. Historically, both thermal and catalytic methods were employed to isomerize hydrocarbons [1]. There are six main types of isomerization that can be performed, but only some are important in Fischer–Tropsch syncrude refining: 1) Double bond isomerization. This type of isomerization is restricted to alkenes, and it does not affect the skeletal structure of the molecule. The isomerization causes the position of the C=C double bond to change, or the configuration of the double bond to change from cis to trans, or vice versa. It is useful only if the molecule remains an alkene. The position of the double bond affects the properties, reactivity, and usefulness of the alkene. In chemical and lubricating oil applications, 1-alkenes are preferred, but in many fuel applications it is preferable to have the double bond at an internal position. Double bond isomerization has been extensively used in Fischer–Tropsch refining technology in the past, but has less use in modern Fischer–Tropsch refineries (Section 16.4.1). It is still an important isomerization reaction though. 2) Conformational isomerization. The simplest form of conformational isomerization is the textbook interconversion of cyclohexane between ‘‘boat’’ and ‘‘chair’’ conformations. Such conformational isomerization is thermally induced and it is not always reversible: for example, denaturing of protein. It does not have much application in refining, although the ability to induce or prevent crystal domains from forming is relevant in some applications. The conformation of heavy molecules is also relevant in feed pretreatment for XTL conversion: for example, cellulose hydrolysis. 3) Stereochemical isomerization. Stereochemistry is of paramount importance in the pharmaceutical industry. Inverting the stereochemistry of a molecule is difficult and usually it has to be synthesized in such a way that the desired isomer can be obtained. It is not a type of isomerization considered in refining, although it may be relevant to niche chemical production. 4) Skeletal isomerization. Broadly speaking, any isomerization of the structure of the carbon backbone of a molecule is skeletal isomerization. In the present context, skeletal isomerization Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
354
18 Isomerization
refers specifically to skeletal isomerization of aliphatic molecules that takes place in the absence of hydrogenation, to differentiate it from hydroisomerization. The aim of skeletal isomerization is to increase the degree of branching of a molecule, and refinery applications of alkene skeletal isomerization are usually limited to butenes and pentenes (Section 16.4.3). In crude oil refineries, it is a technology mainly employed for motor-gasoline production and, considering the alkene-rich nature of syncrude, it is a potentially useful refining technology to consider (Section 18.3). 5) Hydroisomerization. When skeletal isomerization is conducted in conjunction with dehydrogenation and hydrogenation to enable skeletal isomerization of alkanes, it is called hydroisomerization. Like skeletal isomerization, the aim is to increase the degree of branching of the molecules. Hydroisomerization is extensively used in crude oil refining to produce fuels, lubricating oils, and chemicals (Section 16.5.1). Naphtha range molecules are hydroisomerized mainly to produce good quality paraffinic motor-gasoline, whereas hydroisomerization of heavier fractions is mainly aimed at improving cold-flow properties. The same applications are relevant for Fischer–Tropsch syncrude refining, and hydroisomerization is an important syncrude refining technology. Subsequent discussion of hydroisomerization (Section 18.4) will be organized on the basis of feed material, since different catalysts and technologies are used for each. 6) Aromatic isomerization. Alkyl aromatics can be isomerized to change the relative position of the alkyl groups on the aromatic ring. Although fuel properties are affected by the isomer conformation of the alkyl aromatics, aromatic isomerization is considered only for chemical applications. For example, the isomerization of xylenes and cresols finds practical use in chemical refining. Most types of isomerization are acid catalyzed and also take place as side reactions during acid-catalyzed conversion processes. Isomerization can take place during oligomerization (Chapter 19), aromatic alkylation (Chapter 20), cracking (Chapter 21), and reforming (Chapter 22). The underlying reaction chemistry involved in isomerization as a side reaction is the same as for the stand-alone technologies that are described here.
18.2 Reaction Chemistry 18.2.1 Alkene Skeletal Isomerization
The Brønsted acid-catalyzed skeletal isomerization of alkenes takes place through a protonated cyclopropane (PCP) intermediate [2–4]. The isomerization mechanism is shown in Figure 18.1. Evidence for skeletal isomerization through a larger protonated cycloalkane intermediate has also been presented as a minor mechanistic pathway [3]. Protonation of a carbon double bond on a linear carbon chain results in the formation of a secondary carbocation. Since a secondary carbocation is not as stable as a tertiary carbocation, the intermediate can gain additional stabilization by rearrangement to convert the secondary carbocation into a tertiary carbocation. When this rearrangement takes place through a PCP structure, additional stabilization can be obtained from the different contributing resonance
18.2 Reaction Chemistry
H H R
+
H H R
R′
+ H+ − H+
H R′
H
H H
H R′
+
H
H R′
R
R
R
H
H R′
R
R′ +
H
R
H H
R
+
+
R′
+
+
H
R′
R′
R +
− H+ + H+
− H+ + H+
R′
R
R′ R
Figure 18.1 Mechanism of Brønsted acid-catalyzed isomerization. Skeletal isomerization involves a protonated cyclopropane (PCP) intermediate, which may also result in double bond isomerization.
structures. Each resonance structure shown in Figure 18.1 contains the pentacoordinated carbocation on a different carbon. Either double bond isomerization or skeletal isomerization may result from the PCP intermediate. Skeletal isomerization is an equilibrium-limited reaction and the isomerization is exothermic, typically in the range −5 to −20 kJ·mol−1 . The heat release is toward the upper range when isomerizing n-1-alkenes. Lower temperatures favor branched isomers and there is an equilibrium advantage to perform skeletal isomerization at the lowest practical temperature. The need for R and R in Figure 18.1 to be alkyl groups, rather than just hydrogen, is evident from the mechanism. If R and R were not alkyl groups, the intermediate carbocation species would have included primary carbocations. This leads to a very important conclusion, namely that skeletal isomerization requires a molecule with a chain length of five or more carbon atoms in order to avoid the formation of a primary carbocation intermediate. It is hardly surprising to find that it is much easier to skeletally isomerize n-pentenes to methylbutenes than it is to skeletally isomerize n-butenes to methylpropene (isobutene). Skeletal isomerization of butene represents a special case. Unless intimate contact of butene with the catalyst surface stabilizes the transition-state intermediate, it is predicted that skeletal isomerization would form a primary carbocation, which is very unlikely. This lies at the heart of the debate on a monomolecular versus bimolecular mechanism for the skeletal isomerization of butene [5]. According to the monomolecular mechanism, the butene strongly bonds to the catalyst surface, or carbon on the catalyst surface, thereby converting the α-carbon into a secondary carbon (Figure 18.2a). Rearrangement resulting in an intermediate with a positive charge on the α-carbon is then not a problem, because the α-carbon is secondary. According to the bimolecular mechanism, butene first dimerizes to produce an octene, which is then skeletally isomerized through a PCP intermediate, before it cracks to produce isobutene as the product (Figure 18.2b). This avoids the issue of a primary carbocation. In practice, both mechanisms are operative. The bimolecular mechanism of butene skeletal isomerization highlights two important side reactions of alkenes that can take place during isomerization:
355
356
18 Isomerization
(a)
(b)
H
+ H+ − H+
+ H+ − H+
− H+
+
+
+
+
+
H
+ H+
− H+ + H+
Figure 18.2 Butene skeletal isomerization mechanisms, showing the (a) monomolecular isomerization pathway and (b) bimolecular isomerization by dimerization and cracking.
1) Dimerization. Alkene dimerization is a bimolecular reaction favored by a high partial pressure of alkenes. In order to avoid dimerization, skeletal isomerization must be conducted at the lowest practical pressure. There is consequently a trade-off between volumetric reactor productivity and the alkene partial pressure. In this respect, hydroisomerization has an advantage over skeletal isomerization, since the hydroisomerization mechanism provides a way other than the system pressure to regulate the partial pressure of alkenes (Section 18.2.2). 2) Cracking. The isomerized C8 intermediate shown in Figure 18.2b can readily crack to produce two lighter alkenes. Cracking takes place by β-scission of the carbocation intermediate. The cracking reaction becomes easier as the degree of branching of the alkene increases and especially if β-scission leads to the formation of another tertiary carbocation. It is later shown (Section 21.2.2) that cracking of C6 and lighter alkenes is difficult, because β-scission of most C6 and lighter carbocations requires the formation of a primary carbocation intermediate. Cracking as side reaction during skeletal isomerization is therefore dependent on the carbon chain length and it can become a significant side reaction during the isomerization of C7 and heavier molecules. Fischer–Tropsch syncrude contains oxygenates, and acid-catalyzed side reactions involving oxygenates can occur during skeletal isomerization. In fact, the oxygenates are generally stronger nucleophiles than alkenes and will preferentially adsorb onto the catalyst [5]. Dehydration, etherification, and hydration (Chapter 17) are all acid-catalyzed reactions. Depending on the nature of the skeletal isomerization catalyst, water and oxygenates present in syncrude may also lead to catalyst inhibition and catalyst deactivation [6]. 18.2.2 Alkane Hydroisomerization
Alkanes are not reactive for isomerization and before skeletal isomerization can take place, the alkanes must first be activated in some way. During hydroisomerization, the alkanes are first dehydrogenated to alkenes, which can then be skeletally isomerized (Section 18.2.1) before the isomerized alkene is hydrogenated to yield an isomerized alkane as the final product (Figure 18.3). Hydroisomerization employs metal-catalyzed dehydrogenation and hydrogenation in addition to acid-catalyzed isomerization. Alkane activation takes place on different catalytic sites to the isomerization reaction. This makes catalyst design more complicated, but it also holds some advantages for selectivity. Depending on the acid support, the support material itself
18.3 Skeletal Isomerization
R
R′
− H2
R
R′
+ H+ +
+ H2
−H
Metal
Acid
R
H +
R′
− H+ +
+H
Acid
R
R′
+ H2 − H2
R′ R
Metal
Figure 18.3 Hydroisomerization mechanism, which requires metal sites for dehydrogenation and hydrogenation and acid sites for skeletal isomerization.
may be sufficiently active to catalyze alkane isomerization [7]. The general and catalyst-specific isomerization mechanisms have been reviewed by Ono [8]. Hydroisomerization is equilibrium limited and slightly exothermic. The heat of reaction at standard conditions is typically less than −10 kJ·mol−1 . The isomerization equilibrium favors branched products at lower temperatures. The probability of side reactions involving alkenes can be controlled by the combination of catalyst design and operating conditions. Under reaction conditions, most of the material is present as alkanes and not as alkenes. The alkene partial pressure is determined by the metal-catalyzed dehydrogenation–hydrogenation equilibrium, which is decreased by an increase in hydrogen partial pressure and a decrease in operating temperature. By reducing the alkene partial pressure, bimolecular side reactions, such as dimerization, are suppressed. There is a trade-off, of course, because the skeletal isomerization rate depends on the alkene partial pressure and, by lowering the alkene partial pressure, reactor productivity is decreased. Furthermore, the monomolecular side reactions of the alkenes, such as cracking, are not reduced by lowering the alkene partial pressure. In order to reduce cracking, the acid strength of the catalyst can be reduced and the ratio of metal sites to acid sites can be increased. Lowering the acid strength also affects skeletal isomerization, since cracking and isomerization have a common intermediate. By increasing the ratio of metal to acid sites, the probability of hydrogenating a branched alkene before it can re-adsorb onto an acid site is increased. This reduces the risk of cracking branched products, although it cannot prevent successive isomerization and cracking before desorption. This topic is covered in more detail during the discussion of hydrocracking (Section 21.2.3). A strong hydrogenation function may lead to some metal-catalyzed side reactions, such as hydrogenolysis.
18.3 Skeletal Isomerization
The skeletal isomerization of alkenes is usually limited to C4 –C6 material, which somewhat restricts its usefulness as a general-purpose refining technology. The cracking propensity of C7 and heavier alkenes effectively prevents its application to heavier feed materials. There is little difference in the purpose of skeletal isomerization between refineries (Section 16.4.3): 1) It is used in combination with etherification technology (Section 17.3) to produce high-octane fuel ethers as blending component for motor-gasoline. There is a natural synergy between skeletal isomerization and etherification, because etherification selectively converts only the branched alkenes with the double bond on a tertiary carbon. This makes it easy to recycle unconverted alkenes to the skeletal isomerization unit, because etherification improves
357
358
18 Isomerization
Skeletal isomerization
Etherification
Alkene recycle
Alkene
Light alcohol ether (side product) Fuel ether
Alcohol
Oligomers (side product)
Figure 18.4 Skeletal isomerization combined with etherification, which exploits the inherent separation performed by etherification to facilitate recycling unconverted alkenes.
separation (Figure 18.4). In this way, the equilibrium limitation of skeletal isomerization is overcome without the need for a separation unit after skeletal isomerization. 2) Skeletal isomerization can also be employed in combination with oligomerization. Branched alkenes are not only more reactive for oligomerization but also result in more branched products. However, the octane number of branched alkenes is not very sensitive to the degree of branching (Table 13.5). Since branched alkenes have octane numbers in a fairly narrow range, it makes sense to use skeletal isomerization before oligomerization only if the product will be hydrogenated to produce a highly branched paraffinic product. Butene is the only alkene considered for this purpose and, even in this application, there are alternative refining pathways that are preferable to skeletal isomerization followed by dimerization [9]. Dimers of C5 alkenes are not branched enough to produce a high-quality paraffinic motor-gasoline, and no additional benefit is derived from increasing the degree of branching of kerosene for jet fuel beyond the branching that will in any case result from direct dimerization of n-pentenes. The same argument is valid for C6 alkenes. 18.3.1 Butene Isomerization Catalysis
Of the catalysts that were investigated for butene skeletal isomerization, ferrierite is considered the most selective one [10]. Close to 80% of equilibrium conversion can be achieved, but the overall efficiency is somewhat eroded by side product formation (Table 18.1) [11]. At these conditions, the equilibrium isobutene concentration is around 50%. Butene skeletal isomerization requires an operating temperature around 350 ◦ C and higher. Over time, the catalyst activity is gradually lost due to coking of the catalyst [12]. Cycle lengths close to 500 h have been reported for commercial processes [11]. Catalyst activity can be restored by controlled oxidative regeneration. The oxygenate content of a C4 cut in a Fischer–Tropsch refinery is quite low, typically parts per million. In a recycle configuration with an etherification unit (Figure 18.4), some oxygenates may be introduced in the feed by the alkene recycle, but with proper design the oxygenate content of the recycle should be very low. The effect of oxygenates on butene skeletal isomerization over ferrierite was reported for water and alcohols [13]. In both cases, conversion was suppressed through competitive adsorption.
18.3 Skeletal Isomerization Butene skeletal isomerization over ferrierite in the Llyondell process for different cycle lengths.
Table 18.1
Compounds
Light gases Propene n-Butenes (unconverted) Isobutene C5 and heavier alkenes
Average yield (mass%) 290 h
480 h
<0.1 0.5 55.9 39.6 4.0
<0.1 0.3 60.0 36.9 2.8
Butanol had a more severe effect than water, and over the reported test period butanol caused the butene conversion to decrease by 16% (absolute) as compared to 4% (absolute) for water. 18.3.2 Pentene Isomerization Catalysis
Pentene skeletal isomerization is a less demanding reaction, and a wider selection of catalysts is available. Commercial processes are available with ferrierite [11], alumina [14], and acidic molecular sieves as catalysts [15]. The performance of these catalysts differs markedly with respect to side product formation, resistance to oxygenates, and operating conditions (Table 18.2) [6, 11, 14].
Pentene skeletal isomerization over ferrierite (Llyondell), alumina (IFP/Axens), and acidic molecular sieve (UOP) catalyzed processes.
Table 18.2
Description
Operating conditions Temperature (◦ C) Pressure (MPa) Space velocity (h –1 ) Cycle length Co-feed Performance Equilibrium approach (%) Methylbutene yield (%) Side product yield (%)
Ferrierite
Alumina
HTFT feed
No oxygenates
380–430 <0.1 ∼2a 2 wkb None
400–500 0.1–0.3 1–2 CCR None
320–380 1.7 2 1–2 mo H2 c
255–380 1.7 2–4 10–12 mo H2 c
∼80 61 1–2
>85 55–65 10–15
75–85 55–65 ∼5
80–90 65–75 2–4
CCR, continuous catalyst regeneration. Not reported, typical value obtained from other studies. b Different cycle lengths reported, all on the order of weeks, not months. c Co-feeding H :pentene at a molar ratio of 2 : 1–3 : 1. 2 a
Acidic molecular sieve
359
360
18 Isomerization
The benefits that can be derived from a lower operating temperature are clear from the performance of the acidic molecular sieve catalyst (Table 18.2), which outperforms the other catalysts in most respects. This catalyst is unfortunately sensitive to oxygenates and water. These strongly adsorb onto the catalyst, causing catalyst deactivation and these oxygen-containing compounds require a temperature of 320 ◦ C to be desorbed [6]. Unless the oxygenates are removed from a Fischer–Tropsch feed, performance is measurably degraded, as can be seen from the comparison between the high-temperature Fischer–Tropsch (HTFT) feed and the same feed material with the oxygenates removed. The skeletal isomerization activity of alumina is improved by the addition of water and oxygenates [13]. This is related to the hydration of the alumina surface [16]. An alumina-catalyzed process is industrially employed for the skeletal isomerization of HTFT-derived pentenes (Section 9.5.5). 18.3.3 Syncrude Process Technology
There is little difference between the application of skeletal isomerization in a crude oil refinery and in a Fischer–Tropsch refinery. However, differences in the feed composition, which are mainly related to the oxygenate content and the nature of the other impurities in the feed, have a significant impact on technology selection. 1) Oxygenates and water are beneficial for alumina but detrimental to the performance of other skeletal isomerization catalysts. When applying skeletal isomerization technology with syncrude as the feed material, oxygenates and water must be removed from the feed unless an alumina-catalyzed process is selected. 2) Syncrude has a high n-1-alkene content. The n-1-alkenes are more reactive and have a higher associated heat release during skeletal isomerization than other alkenes. This situation is exacerbated by the oxygenates in syncrude. The operating conditions must be adjusted to compensate for the higher reactivity of syncrude feed. For example, during the skeletal isomerization of HTFT pentenes over alumina, problems were experienced during commissioning due to incomplete catalyst regeneration and coke buildup [17]. This could be overcome by adjusting the operation of the regenerator.
18.4 Hydroisomerization
Mechanistically, there is a significant difference in the isomerization of C4 , C5 –C6 , and the C7 and heavier alkanes (Section 18.2). This is reflected in the classification of hydroisomerization technologies. Hydroisomerization technology is divided into four categories based on the carbon number of the feed material, namely, n-butane, C5 –C6 naphtha, C7 naphtha, and heavier alkanes including waxes. This classification reflects differences in catalyst selection, application objectives, and technology. In addition to stand-alone units, hydroisomerization is often performed as a beneficial side reaction during hydrocracking (Chapter 21). The isomerization of heavier alkanes can be performed by hydrocracking or hydroisomerization, and the decision of selecting one over the other is related to the desirability of shifting the carbon number distribution.
18.4 Hydroisomerization
The application of hydroisomerization technology in crude oil refineries parallels that in Fischer–Tropsch refineries. The applications are linked to the carbon number of the feed (Section 16.5.1): 1) Hydroisomerization of n-butane to produce methylpropane (isobutane) is mainly performed to provide feed for aliphatic alkylation (Section 16.4.5). It provides a useful refining pathway for butanes and enables the production of high-octane motor-gasoline. In a Fischer–Tropsch refinery, butane is not naturally abundant. Butane hydroisomerization is industrially applied in Fischer–Tropsch refining (Section 10.4.1), but it relies on an external source of n-butane as feed. 2) Isobutane can also be used to indirectly produce isobutene. Hydroisomerization of butane is more selective and it is conducted at less severe conditions than skeletal isomerization of butene (Section 18.3.1). The isobutane can then be dehydrogenated (Section 16.5.4) to produce isobutene [18]. The same principle can also be applied to C5 material. Single-step hydroisomerization and dehydrogenation of n-pentane to methylbutenes has been reported [19]. 3) The primary aim of C5 –C6 naphtha hydroisomerization is to increase the octane number of this cut for use as paraffinic motor-gasoline (Table 18.3) [20]. Typical once-through hydroisomerization yields a 79–84 RON, which can be increased to 88–92 RON with full recycle. The octane sensitivity is low, and the MON is typically only 2–4 units lower. 4) Refinery benzene can be converted in a C5 –C6 naphtha hydroisomerization unit as long as the benzene content in the feed is not too high. It has been shown that Pt/MOR hydroisomerization catalysts perform well with a feed containing 4.9% benzene [21]. This refining strategy loses the octane value of the benzene and consumes much H2 . Nevertheless, the co-refining of benzene in a naphtha hydroisomerization unit is convenient to adjust the Octane numbers and Reid vapor pressure (RVP) at 37.8 ◦ C of the C5 –C6 alkanes.
Table 18.3
Compound
C5 alkanes n-Pentane 2-Methylbutane 2,2-Dimethylpropane Cyclopentane C6 alkanes n-Hexane 2-Methylpentane 3-Methylpentane 2,2-Dimethylbutane 2,3-Dimethylbutane Cyclohexane Methylcyclopentane
RON
MON
Average octane gain relative to n-alkane
RVP (kPa)
61.7 92.3 85.5 101.3
62.6 90.3 80.2 84.9
– 29.2 20.7 31.0
107 141 253 68
24.8 73.4 74.5 91.8 103.5 83.0 91.3
26.0 73.5 74.3 93.4 94.3 77.2 80.0
– 48.1 49.0 67.2 73.5 54.7 60.3
34 47 42 68 51 23 31
361
362
18 Isomerization
benzene content in motor-gasoline if a separate unit is not justified or an alternative refining pathway is not available. 5) The C7 alkanes are one of the most troublesome Fischer–Tropsch fractions to upgrade [22]. Hydroisomerization of C7 naphtha holds considerable promise. In order to achieve a reasonable octane number, a dibranched C7 product is required. This is challenging, because dibranched C7 hydrocarbons are readily cracked (Section 18.2). 6) Hydroisomerization is ideally suited for improving the cold-flow properties of straight-run distillate range material. Branched alkanes are necessary to meet the freezing point specification of jet fuel. Some branching is also required to meet the cold-flow specifications of diesel fuel. In both cases, cracking must be avoided. 7) Lubricating base oils can be prepared by mild hydrocracking of Fischer–Tropsch waxes, and the vacuum gas oil fraction (slack wax) can also be converted into lubricant base oil by hydroisomerization [23, 24]. 18.4.1 Butane Hydroisomerization Catalysis
The principal catalyst for n-butane hydroisomerization is Pt/Cl− /Al2 O3 . The acidity of the catalyst is maintained by constant chlorination by a chloroalkane co-feed. This requires the feed to be water- and oxygenate free. Chlorine can be leached from the chlorided alumina by water, which leads to catalyst deactivation due to loss of strong acidity. Water reacts with the chlorine to produce hydrochloric acid (HCl), which is corrosive and can cause downstream refining problems. The need for proper feed pretreatment when Fischer–Tropsch-derived butane is used is clear. Nevertheless, with proper feed pretreatment, this technology can be applied in a Fischer–Tropsch refinery (Section 10.4.1). Butane hydroisomerization is conducted at 180–220 ◦ C, 1.5–2.0 MPa, 2 h−1 space velocity, and 0.5–2.0 H2 :butane ratio. Hydroisomerization is equilibrium limited and near-equilibrium conversion is achieved. Side product formation is <2% [25]. 18.4.2 C5 –C6 Naphtha Hydroisomerization Catalysis
A larger variety of catalysts are in industrial use for the hydroisomerization of C5 –C6 naphtha than for butane (Table 18.4) [26]. The most active catalyst type for hydroisomerization is Pt/Cl− /Al2 O3 . This implies that the alkanes are converted at the lowest temperature, where the isomerization equilibrium is most favorable. The once-through yield of branched alkanes is the highest with this catalyst type. The Pt/Cl− /Al2 O3 catalyst is unfortunately also the most sensitive to feed impurities, and it is imperative that water and oxygenates are excluded from the feed for the reasons already mentioned (Section 18.4.1). The advantage that is gained during once-through operation is somewhat offset by the increased cost associated with feed pretreatment and the need for a continuous chloroalkane co-feed to keep the catalyst in its chlorinated state. Industrially, this type of catalyst has been successfully used for the hydroisomerization of HTFT pentanes (Section 10.4.1), demonstrating that with proper feed pretreatment the feed sensitivity of Pt/Cl− /Al2 O3 can be managed.
18.4 Hydroisomerization Catalyst types industrially used for the hydroisomerization of C5 –C6 naphtha.
Table 18.4
Description Operating conditions Temperature (◦ C) Pressure (MPa) Space velocity (h –1 ) H2 :alkane ratio Feed requirements Water tolerance (µg·g−1 ) Sulfur tolerance (µg·g−1 ) Benzene tolerance (%) C7 alkane tolerance (%) Once-through yield (%) 2-Methylbutane/C5 2,2-Dimethylbutane/C6
Pt/Cl – /Al2 O3 Pt/SO4 2− /ZrO2
Pt/MOR
130–150 1.5–3.5 1–3 1–2
180–210 1.5–3.5 1–3 1–2
250–280 1.5–3.5 1–3 1–1.5
0 0 2 2
20 20 2 2
200 200 10 5
68–72 21
65–71 20.5
63–67 19
Sulfated zirconia-based catalysts (Pt/SO4 2− /ZrO2 ) were developed for naphtha hydroisomerization to achieve low-temperature operation while having some tolerance for feed impurities. Although the operating temperature is higher and the once-through performance is not as good as that of Pt/Cl− /Al2 O3 catalysts, feed pretreatment is less demanding and no chloroalkane co-feed is required. Zeolite-based hydroisomerization that employs Pt/mordenite catalysts are more robust than any of the other catalyst types, but requires a higher operating temperature. The once-through performance of Pt/MOR is consequently worse, but very long catalyst lifetimes (>10 years) have been achieved with this catalyst type. Pt/MOR is the recommended catalyst type for the hydroisomerization of Fischer–Tropsch naphtha on account of its stability and tolerance for water and oxygenates, as well as its ability to handle a higher concentration of benzene. The equilibrium limitations associated with once-through operation can be overcome by recycle operation. The strength of the metal and acid functionality, as well as the operating conditions, also has some other implications: 1) A strong metal functionality may result in increased hydrogenolysis. The nature of hydrogenolysis over Pt is such that it proceeds in a manner analogous to β-scission over acid catalysts. Methane is not a preferred product, and hydrogenolysis of n-hexane over Pt yields products in the order propane > ethane ≈ n-butane > methane ≈ n-pentane [27]. 2) When the operating temperature is increased, the catalyst becomes more active for dehydrogenation, which is an endothermic reaction. This promotes skeletal isomerization but also acid-catalyzed side reactions, such as cracking and oligomerization. 3) A strong acid function affects cracking propensity. This is important when the feed contains cycloalkanes that can be ring-opened by cracking. The amount of cyclohexane and benzene in the feed is preferably limited to 1–2 mass% [25], in order to avoid the potential octane number decrease associated by their conversion to acyclic hydrocarbons (Table 18.3).
363
364
18 Isomerization
18.4.3 Heavy Alkane and Wax Hydroisomerization Catalysis
Many catalyst types have been tested for the hydroisomerization of C7 and heavier alkanes and waxes [5, 28–30]. These heavier alkane feed materials are susceptible to cracking, and hydroisomerization catalysts are also potential hydrocracking catalysts. Hydroisomerization is conducted at a lower temperature than hydrocracking, and hydroisomerization catalysts are optimized to increase isomerization while suppressing cracking. Some general catalyst properties that are important for good hydroisomerization catalysts are the following: 1) Mild acidity is preferred for hydroisomerization. Mild acidity reduces the probability of cracking as the degree of branching increases. Strong acid sites promote cracking, which is not desirable in a hydroisomerization catalyst. 2) There is an optimum ratio between metal sites and acid sites. This ratio does not refer to the number of sites, but rather the activity ratio of the sites. If the metal:acid balance is too low, the cracking probability increases. If the metal:acid balance is too high, conversion is lower and the risk of hydrogenolysis increases. 3) The pore structure of the acid support affects the selectivity to multibranched isomers. For example, it was found that 12-membered ring zeolites in general have a higher selectivity to dibranched alkanes than 10-membered ring zeolites. In this respect, the performance of mesoporous silica–alumina was between that of 12- and 10-membered ring zeolites [31]. 4) Higher hydroisomerization yields can be achieved with catalysts that are promoted with noble metals, such as Pt and Pd [29]. The higher dehydrogenation–hydrogenation activity of noble metals requires fewer metal sites for the same metal site activity. This in turn allows better metal site dispersion and increases hydroisomerization selectivity. It also allows space for more mild acid sites and thereby increases overall catalyst activity, without resorting to stronger acidity to achieve the optimum metal:acid balance. 5) High catalyst accessibility provided by a suitable large pore structure is needed to accommodate long-chain alkanes. 18.4.4 Syncrude Process Technology
The application of hydroisomerization in a Fischer–Tropsch refinery has one advantage over crude oil refining in that the syncrude is sulfur free. However, on the downside, syncrude process technology has to cope with oxygenates and water present in the syncrude, as well as a high alkene content in the naphtha fraction. Some aspects to consider are the following: 1) Oxygenates are reactive compounds, and standard feed drying technology may not be suitable for oxygenate removal to the level required by the more sensitive hydroisomerization catalysts. In general, it is recommended to make use of oxygenate-tolerant Pt/MOR catalysts (e.g., S¨ud-Chemie Hysopar) for naphtha hydroisomerization. 2) Recycle operation during naphtha hydroisomerization allows the production of a better quality isomerate. Recycle operation also makes the hydroisomerization unit insensitive to equilibrium conversion and thereby compensates for the higher operating temperature required by oxygenate-tolerant Pt/MOR catalysts. All material can be isomerized, and the
18.4 Hydroisomerization
actual equilibrium conversion affects only the size of the recycle stream. Although recycle operation is more costly than once-through operation, the increased isomerate quality can potentially offset the added cost of recycling elsewhere in the refinery. For example, it has been shown that a considerable saving is possible by lowering the reformer severity when a higher quality isomerate is produced [32]. 3) Recycle operation requires the separation of the near-equilibrium mixture of linear and branched alkanes. Separating the n-alkanes from the branched alkanes is comparatively easy for a single-carbon-number feed, but much more involved for a C5 –C6 mixture. This is not a problem unique to syncrude, but considering the large variety of refining pathways available to syncrude, serious consideration should be given to single-carbon-number hydroisomerization. Furthermore, the straight-run C5 –C6 naphtha fraction constitutes more than 50 mass% of the total C5 –C10 naphtha fraction in HTFT syncrude and about 40% in LTFT syncrude. It may be difficult to include all of the C5 –C6 material in motor-gasoline as isomerates, which lends further support to single-carbon-number hydroisomerization. 4) Single-carbon-number hydroisomerization is beneficial from a fundamental point of view too. When mixed C5 –C6 naphtha is hydroisomerized, isomerization of the C5 naphtha is equilibrium limited, whereas the isomerization of C6 naphtha is kinetically limited. The yield of 2-methylbutane is close to equilibrium at each of the operating temperatures indicated in Table 18.4, but 2,2-dimethylbutane is quite far from equilibrium. For example, the equilibrium concentration of 2,2-dimethylbutane in an acyclic C6 alkane mixture is around 35% at 270 ◦ C and more than 50% at 150 ◦ C [33]. During single-carbon-number hydroisomerization, the space velocity can be adjusted to optimize conversion for a specific carbon number. 5) Straight-run syncrude naphtha is alkene rich, and hydroisomerization is an alkane conversion technology. Conventional wisdom dictates that the syncrude should be hydrotreated before being hydroisomerized, which will make it a costly syncrude refining technology. However, one can take advantage of the alkene-rich nature of syncrude. When straight-run syncrude is fed directly to a Pt/MOR hydroisomerization process, the heat of alkene hydrogenation can be utilized to lower the fresh feed preheater temperature from 260 to 100–150 ◦ C (Figure 18.5) [34]. The partial pressure of alkenes at the reactor inlet is lowered by the H2
H2 recycle
Hydroisomerization
Syncrude
Isomerate
n-Alkane recycle
Figure 18.5
Direct hydroisomerization of an alkene-rich Fischer–Tropsch syncrude.
365
366
18 Isomerization
paraffinic recycle, and the low inlet temperature further reduces the risk of catalyst coking. The hydroisomerization design is governed by heat management, which implies that the amount of paraffinic recycle is determined not by hydroisomerization conversion but by the alkene content of the fresh feed. For operation with C5 feed, it may be possible to increase the space velocity in the hydroisomerization reactor to balance the per pass conversion with the size of the recycle. For operation with C6 feed, a higher per pass 2,2-dimethylbutane yield may be obtained.
References 1. Condon, F.E. (1958) in Catalysis, Alkylation,
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Isomerization, Polymerization, Cracking and Hydroreforming, vol. VI (ed. P.H. Emmett), Reinhold, New York, pp. 43–189. Weitkamp, J. (1982) Isomerization of long-chain n-alkanes on a Pt/CaY zeolite catalyst. Ind. Eng. Chem. Prod. Res. Dev., 21, 550–558. Martens, J.A. and Jacobs, P.A. (1990) Evidence for branching of long-chain n-alkanes via protonated cycloalkanes larger than cyclopropane. J. Catal., 124, 357–366. Sie, S.T. (1993) Acid-catalyzed cracking of paraffinic hydrocarbons. 3. Evidence for the protonated cyclopropane mechanism from hydrocracking/hydroisomerization experiments. Ind. Eng. Chem. Res., 32, 403–408. De Klerk, A. and Furimsky, E. (2010) Catalysis in the refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Cowley, M. (2006) Skeletal isomerization of Fischer-Tropsch-derived pentenes: the effect of oxygenates. Energy Fuels, 20, 1771–1776. Chu, H.Y., Rosynek, M.P., and Lunsford, J.H. (1998) Skeletal isomerization of hexane over Pt/H-Beta zeolites: Is the classical mechanism correct? J. Catal., 178, 352–362. Ono, Y. (2003) A survey of the mechanism in catalytic isomerization of alkanes. Catal. Today, 81, 3–16. De Klerk, A. and De Vaal, P.L. (2008) Alkylate technology selection for Fischer-Tropsch syncrude refining. Ind. Eng. Chem. Res., 47, 6870–6877. Choudhary, V.R. (1974) Catalytic isomerization of n-butene to isobutene. Chem. Ind. Dev., 8 (7), 32–41. Wise, J.B. and Powers, D. (1994) Highly selective olefin skeletal isomerization process. ACS Symp. Ser., 552, 273–285. De Jong, K.P., Mooiweer, H.H., Buglass, J.G., and Maarsen, P.K. (1997) Activation
13.
14.
15.
16.
17.
18.
19.
20.
21.
and deactivation of the zeolite ferrierite for olefin conversion. Stud. Surf. Sci. Catal., 111, 127–138. Rossini, S., Catani, R., Cornaro, U., Guercio, A., Miglio, R., and Piccoli, V. (1997) Oxygenates and nitriles effects on some catalysts for the skeletal isomerization of C4 and C5 olefins. Prepr. Pap.-Am. Chem. Soc., Div. Pet. Chem., 42 (3), 588–591. Duplan, J.-L., Amigues, P., Verstraete, J., and Travers, C. (1996) Kinetic studies of the skeletal isomerization of n-pentenes over the ISO-5 process catalyst. Proc. Ethylene Prod. Conf., 5, 429–449. Ozmen, S.M., Abrevaya, H., Barger, P., Bentham, M., and Kojima, M. (1993) Skeletal isomerization of C4 and C5 olefins for increased ether production. Fuel Reformulat., 3 (5), 54–59. De Klerk, A. (2011) Key catalyst types for the efficient refining of Fischer–Tropsch syncrude: alumina and phosphoric acid, in Catalysis, vol. 23 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge, pp. 1–49. Smook, D. and De Klerk, A. Catalyst mechanical properties – case study. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cdP09. Sanfilippo, D. (2000) Dehydrogenation of paraffins. Key technology for petrochemicals and fuels. CatTech, 4 (1), 56–72. Herrera, G., Lardiz´abal, D., Mart´ınez, V.H.C., and Elgu´ezabal, A.A. (2001) Dehydroisomerization of n-pentane to isopentene on molecular sieves impregnated with platinum. Catal. Lett, 76, 161–166. (1988) Physical Constants of Hydrocarbons and Non-Hydrocarbon Compounds, ASTM Data Series DS 4B, 2nd edn, ASTM, Philadelphia, PA. ´ Perger, Hancs´ok, J., Holl´o, A., Debreczeni, E., J., and Kall´o, D. (1999) Benzene saturating
References
22.
23.
24.
25.
26.
27.
isomerisation. Stud. Surf. Sci. Catal., 125, 417–424. Dancuart, L.P., De Haan, R., and De Klerk, A. (2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532. Calemma, V., Peratello, S., Stroppa, F., Giardino, R., and Perego, C. (2004) Hydrocracking and hydroisomerization of long-chain n-paraffins. Reactivity and reaction pathway for base oil formation. Ind. Eng. Chem. Res, 43, 934–940. Kobayashi, M., Saitoh, M., Togawa, S., and Ishida, K. (2009) Branching structure of diesel and lubricant base oils prepared by isomerization/hydrocracking of Fischer-Tropsch waxes and alpha-olefins. Energy Fuels, 23, 513–518. Travers, C. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 229–256. Weyda, H. and K¨ohler, E. (2003) Modern refining concepts – an update on naphtha-isomerization to modern gasoline manufacture. Catal. Today, 81, 51–55. Matsumoto, H., Saito, Y., and Yoneda, Y. (1970) Contrast between nickel and platinum catalysts in hydrogenolysis of saturated hydrocarbons. J. Catal., 19, 101–112.
28. Okuhara, T. (2004) Skeletal isomerization of
29.
30.
31.
32.
33.
34.
n-heptane to clean gasoline. J. Jpn. Pet. Inst., 47 (1), 1–10. Deldari, H. (2005) Suitable catalysts for hydroisomerization of long-chain normal paraffins. Appl. Catal. A, 293, 1–10. Akhmedov, V.M. and Al-Khowaiter, S.H. (2007) Recent advances and future aspects in the selective isomerization of high n-alkanes. Catal. Rev.-Sci. Eng., 49, 33–139. Corma, A., Martinez, A., Pergher, S., Peratello, S., Perego, C., and Bellusi, G. (1997) Hydrocracking-hydroisomerization of n-decane on amorphous silica-alumina with uniform pore diameter. Appl. Catal. A, 152, 107–125. Kuchar, P.J., Bricker, J.C., Reno, M.E., and Haizmann, R.S. (1993) Paraffin isomerization innovations. Fuel Process. Technol., 35, 183–200. Rossini, F.D. (1972) Chemical thermodynamics in the petroleum industry. J. Inst. Pet., 58, 279–286. Lamprecht, D. and De Klerk, A. (2009) Hydroisomerization of 1-pentene to iso-pentane in a single reactor. Chem. Eng. Commun., 196, 1206–1216.
367
369
19 Oligomerization 19.1 Introduction
Alkene oligomerization was identified as one of the key Fischer–Tropsch refining technologies [1, 2]. This is understandable considering the large amount of alkenes in syncrude (Table 1.2) and especially the amount of alkenes in the lighter syncrude fractions. It is difficult to conceive of a high-temperature Fischer–Tropsch (HTFT) refinery without an alkene oligomerization unit to convert the gaseous alkenes into liquid products. Although low-temperature Fischer–Tropsch (LTFT) synthesis produces less alkenes, there are still percentage levels of alkenes present in the gaseous products from LTFT synthesis. Furthermore, the total amount of light alkenes that are produced may increase with time on stream as the LTFT catalyst deactivates (Sections 4.5.5–4.5.7). Depending on the Fischer–Tropsch catalyst, such deactivation may actually be to the advantage of the refinery if it employs oligomerization or alkylation technology [3]. Oligomerization refers to one or more consecutive addition reactions between alkenes. In this book, alkene addition is collectively referred to as oligomerization, unless it is important to indicate a specific multiple of the addition reaction. It is a catch-all term that includes dimerization (addition reaction of two alkenes), trimerization, tetramerization, and higher multiples of addition reactions. Refinery processes employing oligomerization is often colloquially called polymerization processes, which should not be confused with true polymerization to produce plastics such as polyethylene and polypropylene. The term polymerization will not be employed here, despite its ubiquitous use in oligomerization and refining literature. As refining technology, oligomerization was originally developed to convert the gaseous products from crude oil cracking operations into liquid products. Oligomerization units started appearing in second-generation crude oil refineries (Section 2.4.2), and some refineries to this day still have oligomerization units. The main product from these units is high-octane olefinic motor-gasoline. In Fischer–Tropsch refineries, alkene availability is not constraining, but in crude oil refineries the limited availability of alkenes restricts the continued widespread use of oligomerization technology. The capital and operating cost of aliphatic alkylation (Section 16.4.5) as alternative is more than double that of oligomerization, and, as a result, oligomerization is still a refining technology selected by some crude oil refiners despite its higher alkene demand [4]. There are some general objectives when oligomerization technology is employed, but in most cases the application (Section 16.4.4) is closely linked to the selection of the oligomerization technology and catalyst. In this respect, oligomerization is different from most other refining Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
370
19 Oligomerization Table 19.1 Alkene oligomerization technologies relevant for fuel refining. Examples of some commercially available technologies are given.
Catalyst
Technology
Supplier
Main fuels application
Solid phosphoric acid
CatPoly InAlk Polynaphthaa Selectopola Octol-Ab MOGD COD Emogas NExOCTANE Dimersol Gc
UOP UOP IFP/Axens IFP/Axens H¨uls/UOP ExxonMobil PetroSA ExxonMobil Fortum Oy IFP/Axens
Motor-gasoline, jet fuel Motor-gasoline Diesel fuel Motor-gasoline Motor-gasoline Diesel fuel Diesel fuel Motor-gasoline, jet fuel Motor-gasoline Motor-gasoline
Amorphous silica-alumina Montmorillonite H-ZSM-5 zeolite H-ZSM-22 or -57 zeolite Acidic resin Homogeneous nickel a
Also available with a zeolite-based catalyst. available with an Ni-promoted catalyst for butene dimerization to a more linear product. c Dimersol E for ethene oligomerization and Dimersol X for butene dimerization to a more linear product. b Also
technologies. For most refinery conversion processes, the products are broadly speaking similar, irrespective of the technology or catalyst selection. For oligomerization, different applications require different catalyst and technology combinations (Table 19.1) [5]. Potential applications of different oligomerization technologies and catalysts with Fischer– Tropsch syncrude include the following: 1) The most important general application of oligomerization is to convert normally gaseous alkenes into liquids. Preferably, this conversion should target a specific product, but in the coarsest sense it can be used to increase the liquid syncrude yield from a Fischer–Tropsch facility. It is consequently a useful technology for a syncrude upgrader, as well as for a refinery to improve its carbon efficiency. 2) Some oligomerization technologies can double as deoxygenation technologies. In such applications, straight-run oil fractions can be oligomerized and deoxygenated to produce olefinic products with little remaining oxygenates that can more easily be fractionated and refined. By doing so, the boiling point broadening is significantly reduced. One specific implementation of this type is the conversion of the total syncrude product over an oligomerization and deoxygenation active acid catalyst directly after Fischer–Tropsch synthesis [6]. 3) The naphtha to distillate ratio in the refinery can be manipulated through oligomerization. The selection of the oligomerization technology will favor one over the other, but some oligomerization catalysts also allow the manipulation of the naphtha to distillate ratio through the operating conditions of oligomerization. 4) The primary product of alkene oligomerization in most refineries is good quality olefinic motor-gasoline, also called polymer gasoline (Table 2.5). For this purpose solid phosphoric acid (SPA) catalyzed oligomerization is preferred, although alternative catalysts can be selected.
19.1 Introduction
5) Selective isobutene dimerization followed by alkene hydrogenation is called indirect alkylation, because it produces an alkylate-equivalent product, namely a high-octane paraffinic motor-gasoline [7, 8]. Indirect alkylation units became increasingly prevalent after the inclusion of 2-methoxy-2-methylpropane (MTBE) in motor-gasoline was prohibited in parts of the United States [9]. This made some installed etherification capacity for MTBE production redundant. These units and their associated isobutene feed became prime candidates for conversion into resin-catalyzed isobutene dimerization processes. Selective isobutene dimerization can also be performed over SPA. Straight-run Fischer–Tropsch syncrude is lean in isobutene, and isobutene dimerization should typically be considered in conjunction with another refining technology that produces isobutene, such as skeletal isomerization (Chapter 18) or catalytic cracking (Chapter 21). 6) Another form of indirect alkylation is selective n-butene dimerization over SPA followed by alkene hydrogenation [10]. This technology is well matched to Fischer–Tropsch syncrude, because it does not require isobutene. The oligomerization and hydrogenation steps can also be combined in a single unit by employing a metal-promoted SPA catalyst, for example, Fe/NiO/SPA [11, 12]. 7) Most oligomerization technologies produce a kerosene range product that is branched. This is a natural consequence of the oligomerization mechanism. When this product is hydrogenated, it becomes an iso-paraffinic kerosene (IPK), which is a good jet fuel blending component. The branching is required for the demanding cold-flow characteristics of jet fuel (Section 14.3.3). The IPK derived from SPA-catalyzed Fischer–Tropsch alkene oligomerization is the only synthetic product that has been qualified for inclusion in semisynthetic jet fuel, and it is also a component qualified for fully synthetic jet fuel (Section 14.2.1). 8) Oligomerization and aromatic alkylation (Chapter 20) can be combined in a single unit to produce a synthetic jet fuel meeting Jet A-1 specifications and the minimum aromatic requirement for synthetic jet fuel [13]. This process makes use of an SPA catalyst. 9) Oligomerization at higher temperatures with H-ZSM-5 is able to generate aromatics as a side reaction during oligomerization. It is therefore in principle possible to operate such an oligomerization process in a way so that the process, even without an aromatic co-feed, can meet the minimum aromatic requirement for synthetic jet fuel. It was reported that the hydrogenated kerosene from H-ZSM-5 oligomerization met all standard Jet A-1 specifications [14]. With mild metal promotion of the zeolite, this process may even be further improved. 10) Refinery benzene reduction can be accomplished by co-feeding the benzene with alkenes to an appropriate acid-catalyzed refinery process, of which oligomerization has been shown to be the most suitable [15]. This process has been successfully demonstrated on industrial scale in a Fischer–Tropsch refinery using an SPA-catalyzed oligomerization unit [16]. 11) Distillate range material can be produced by oligomerization. Depending on the oligomerization technology, it is possible to convert gaseous and naphtha range alkenes into distillate [17]. The distillate properties depend on the technology. When ZSM-5-catalyzed oligomerization is employed, the branching in the distillate is limited and the hydrogenated distillate has a high cetane number and good cold-flow properties [18–20].
371
372
19 Oligomerization
12) Selective dimerization of 1-butene over a Ni-containing catalyst gives a product that has a low degree of branching [21]. The octene mixture thus produced is employed as feed material for hydroformylation and hydrogenation to produce isononanol, which is a plasticizer alcohol. The synergy between this application and a Fischer–Tropsch-based facility is obvious, with 1-butene, CO, and H2 being readily available. 13) Various homogeneous catalyst systems are industrially employed for ethene oligomerization to produce a range of n-1-alkenes for various chemical uses [22]. There are also specific homogeneous catalysts for the selective trimerization of ethene to 1-hexene, which is a comonomer used in plastics [23], and selective tetramerization of ethene to 1-octene is being commercialized [24]. 14) In the past, detergent range alkenes have been produced by tetramerization of propene over SPA [25]. The degree of branching of this product is very high, and more linear alkenes from other routes are generally preferred. In a Fischer–Tropsch refinery, selective dimerization of hexenes and heptenes may be considered for this purpose. 15) Synthetic polyalphaolefin (PAO) lubricants are produced by the oligomerization of n-1-alkenes. Lubricant base oils have been prepared from Fischer–Tropsch-derived alkenes [26]. Lubrication oil production by oligomerization of Fischer–Tropsch 1-alkenes is still of interest [27]. 19.2 Reaction Chemistry
The reaction chemistry of alkene oligomerization cannot be described in generic terms without some reference to the catalyst or process. There is not a single mechanism and, even when alkene addition is catalyzed by the same general mechanism, different catalysts may still produce products with very different characteristics [28]. Broadly speaking, there are four different mechanisms of industrial relevance for alkene oligomerization (Figure 19.1): 1) Classic Whitmore-type carbocation mechanism. Brønsted acid-catalyzed alkene oligomerization takes place through this mechanism. The first step in this mechanism involves the protonation of an alkene to yield a free carbocation intermediate. This carbocation intermediate is capable of all the side reactions associated with such chemistry. Depending on the length of the carbon chain, double bond isomerization, skeletal isomerization, and cracking by β-scission may take place (Figure 18.2). These are all monomolecular reactions and are always in competition with the addition reaction that is bimolecular. After the alkene addition reaction, the addition intermediate is still a carbocation, which is still capable of further alkene addition as well as the aforementioned side reactions. The final product is an alkene and it is also capable of further protonation and further reaction. Zeolites and acidic resin catalysts oligomerize alkenes by a carbocation mechanism. 2) Ester-based mechanism. Some acid catalysts form strong formal σ -bonds with the protonated intermediate, and the mechanism involves a polarized acid ester, rather than the equivalent carbocation. Transitions that would otherwise involve a primary carbocation intermediate become possible when the α-carbon of the alkene is bonded to the acid, since the α-carbon is no longer a primary carbon. The polarized intermediates are weaker electrophiles than
19.2 Reaction Chemistry
(b)
+
+ H+
(a)
−H
− H+
+
+
+
+ H3PO4
− H3PO4 d+
− H3PO4
O d− O
P OH OH
(c)
+ H+
+ H3PO4
O O
P
OH
OH +H
−H
L (d) L
+
L
M
L M H
L
−L
L
L
L
M
L M
M H
L L
M b -H
1,2-Insertion
b Hydride elimination
Figure 19.1 Oligomerization mechanisms illustrated insertion mechanism where L denotes an arbitrary by propene dimerization: (a) classic Whitmore-type ligand. Stereochemistry, side reactions, and rearcarbocation mechanism, (b) ester-based mechanism, rangements of the intermediates are not shown. (c) free radical mechanism, and (d) organometallic
carbocations and reactions can be more selective, but typical acid-catalyzed side reactions can still occur. Phosphoric acid–based catalysts oligomerize alkenes by an ester mechanism. 3) Organometallic insertion mechanism. Some metals are capable of forming coordination complexes with alkenes. This chemistry forms the basis for Ziegler–Natta and metallocene polymerization catalysis. Chain growth proceeds by 1,2-insertion, and β-hydride elimination terminates it. The metal most often encountered in general alkene oligomerization is Ni. For ethene oligomerization specifically, a variety of metal complexes are used, which include metals such as Ni, Al, Zr, Ti, and Cr. The ligands on the organometallic catalyst can be designed to catalyze oligomerization in a very specific way, making it useful for very selective catalysis. Homogeneous and heterogeneous nickel-based catalysts can oligomerize alkenes by the insertion mechanism to yield more linear products than is found during acid catalysis. 4) Free radical mechanism. The free radical oligomerization of alkenes is initiated by the formation of a free radical species that propagates by addition to another alkene. The chain growth is terminated when the radical recombines with another radical, or abstracts a hydrogen atom to initiate a new radical chain. This chemistry is also encountered in free radical polymerization. The process is thermally initiated, and co-feeding compounds with low bond dissociation energy, such as peroxides, can lower the initiation temperature.
373
374
19 Oligomerization
The side reactions occurring during free radical oligomerization are usually fewer than for acid-catalyzed oligomerization. Double bond isomerization occurs by intramolecular hydrogen transfer (Figure 16.5). Skeletal isomerization and cracking is not prevalent, except at high temperatures (>400 ◦ C) where C–C bond scission becomes significant. There is no catalyst involved in free radical oligomerization. Alkene oligomerization is highly exothermic, and heat management is crucial in/during oligomerization. Each alkene addition reaction is accompanied by a heat release of 85–105 kJ·mol−1 at standard conditions. To put this into perspective, during ethene oligomerization, every 1% conversion is equivalent to an adiabatic temperature increase of 12–13 ◦ C. Oligomerization is favored by low temperature and high pressure, as one would expect from Le Chatelier’s principle. It is not customary to think of oligomerization as an equilibrium-limited process, but at high temperature, oligomerization is subject to equilibrium limitations. At low temperature (<200 ◦ C), equilibrium favors oligomerization, and the reaction rate is usually not high enough to reach the equilibrium limitation. Over acid catalysts, there is a transition between kinetically limited oligomerization and thermodynamically controlled oligomerization in the temperature region 200–250 ◦ C. The point of transition is determined by the activity of the catalyst; more active catalysts will reach the transition point at a lower temperature.
19.3 Catalysis
There are many oligomerization technologies to choose from for fuels refining (Table 19.1) and even more for chemicals refining. Over the years, there have been different waves of technology development, starting with thermal (noncatalytic) oligomerization, which was followed by catalytic oligomerization employing phosphoric acid and other liquid acid catalysts. Later on, various amorphous and zeolitic silica–alumina catalysts were developed, as well as acidic resin catalysts. Each catalyst type enabled the production of new or different products and contributed to the smorgasbord of options, rather than displacing the older technologies or catalysts. In fact, for many Fischer–Tropsch-based applications, SPA is still the catalyst of choice. Overviews of homogeneous and heterogeneous oligomerization catalysis can be found in books and reviews spanning many decades [5, 29–37]. Only some of the catalyst types are discussed here. For some commercial technologies, there is insufficient information available to perform a proper evaluation of the catalyst type for syncrude conversion. In order the help with technology selection for refining by oligomerization, some general pointers are given below: 1) Carbon number distribution. Both catalyst and operating conditions affect the carbon number distribution and thereby the yield of the different product types. The most important controlling factors are kinetic limitations at lower temperatures, thermodynamic limitations at higher temperatures, and catalyst-related effects that promote desorption over adsorption (or radical transfer in the case of thermal oligomerization). 2) Alkene composition. Some catalyst types are sensitive to the composition of the alkenes in the feed. Such feed sensitivity can be in the form of preferential conversion of particular
19.3 Catalysis
3)
4)
5)
6)
7)
alkenes, or it can affect the nature of the oligomerization product. If the product is formed in such a way that the carbon double bond is in a sterically hindered position, further reaction may be difficult. Concentration of inerts. The amount of inert material can have a dramatic effect on oligomerization selectivity. Inert material reduces the probability of multiple addition reactions before desorption. Increasing the inert concentration may therefore cause a shift in product distribution to lighter products. Inert material also helps with heat management. More inert material reduces the adiabatic temperature rise, which may be beneficial for oligomerization selectivity. Oxygenates. Not all oligomerization catalysts tolerate oxygenates. Oxygenates are reactive compounds and are more strongly adsorbed by acid catalysts. The oxygenates may also be catalyst poisons for some catalyst types, and the water generated during deoxygenation reactions may affect catalyst performance. Catalyst geometry. Catalysts with a pore-constrained geometry can affect the nature of the oligomerization products through transition-state selectivity or by imposing diffusion limitations. This can be useful if the degree of branching in the product must be limited. Acid strength. The operating temperature window, propensity for side reactions, and catalyst deactivation rate can all be affected by the acid strength of the catalyst. In some cases, strong acid sites may activate alkanes for aliphatic alkylation reactions, or catalyze hydrogen transfer reactions. Mechanism. The dominant reaction mechanism is dependent on the catalyst, but temperature can affect it. Selecting the most appropriate oligomerization mechanism for the products required is an effective way to narrow the technology selection.
19.3.1 Solid Phosphoric Acid
SPA catalysts are actually supported liquid-phase catalysts. The SPA catalyst is prepared by impregnating kieselguhr (diatomaceous earth), which is a natural silica source, with phosphoric acid and then extruding and calcining the mixture. Alkene oligomerization takes place in the viscous phosphoric acid layer on the support. Reaction takes place by an ester mechanism (Figure 19.1b) and the catalytic behavior is complex [38]. Catalysis is influenced by the amount of water or oxygenates in the feed, which affects the hydration (acid strength) of the catalyst. Catalysis is also affected by the chain length of the alkene and operating conditions. Shorter alkenes form stable phosphoric acid esters, and the alkyl phosphoric acid esters of the shorter alkenes become the primary reaction intermediates. This affects the product selectivity. The relative stabilities of the different alkyl phosphoric acid esters also determine intramolecular rearrangement and the likelihood that it will participate in an addition reaction. In most refineries that employ SPA-catalyzed oligomerization, the primary product is high-octane olefinic motor-gasoline. The conversion is insensitive to the feed composition and operating conditions (Table 19.2) [39–43]. Similar observations were reported for the oligomerization of Fischer–Tropsch-derived alkenes (Table 9.3). Different C2 –C4 alkenes converted at different conditions result in olefinic motor-gasolines within a narrow octane number range, usually with a RON of 95–97 and MON of 81–82. This reflects the insensitivity of octane number
375
376
19 Oligomerization Table 19.2 Olefinic motor-gasoline produced from different alkene feedstocks by phosphoric acid–catalyzed oligomerization.
Description
Phosphoric acid oligomerization with different alkenes Ethene [39] Propene [40] n-C4 [41] Mixed C4 [42] C2 –C4 [43] C3 –C4 [43]
Operating conditions Temperature (◦ C) Pressure (MPa) Performance Alkene conversion (mass%) Naphtha selectivity (mass%) Olefinic motor-gasoline Density at 15.6 ◦ C (kg·m –3 ) RON MON Reid vapor pressure (kPa) T90 distillation (◦ C) T95 distillation (◦ C) FBP distillation (◦ C)
296–324 3.6
204–232 1.7
177 0.8
205–230 5.5
204–232 1.4
204 0.7
65–73 59–71
90–95 100a
72 77
75–95 100
67–94 100a
72–89 100a
711 96 82 45 – 183 203
732 94.6 81.0 69 193 – 226
708 – 81 – – ∼175 –
732 97.1 84.0 14 128 – 198
734 – 82 59 198 – 249
722 – 81 76 207 – 237
a Complete product employed as olefinic motor-gasoline, at a 175 ◦ C cut point; the naphtha to distillate ratio is around 83 : 17.
to the branching structure in branched alkenes. The alkene oligomers produced by different feeds and at different conditions are not isostructural. In a Fischer–Tropsch refinery, there is a need to produce a good quality alkylate-equivalent blending component for motor-gasoline. The considerable impact of feed composition and operating conditions on SPA oligomerization becomes apparent when the olefinic motor-gasoline is hydrogenated (Table 19.3) [44]. There are some important aspects of SPA-catalyzed oligomerization to be kept in mind when oligomerizing syncrude to produce a hydrogenated motor-gasoline: 1) Hydrogenated RON and MON are very sensitive to the structure of the oligomer. 2) The best quality hydrogenated motor-gasoline is produced by dimerization of butenes with as little lighter and heavier alkenes in the feed as possible [44]. For mixed alkene feed materials, the hydrogenated octane numbers are much poorer. 3) It is not necessary to have a high isobutene content in the C4 feed to obtain a good quality motor-gasoline, because SPA catalyzes butene skeletal isomerization through the ester mechanism at low temperature [45]. Nevertheless, increasing the isobutene content will improve the quality of the motor-gasoline beyond which can be obtained by n-butenes only. 4) Low temperature (160 ◦ C or less) combined with a high level of phosphoric acid hydration for weaker acidity produces the best quality of motor-gasoline from butene oligomerization [46]. The same conditions also increase the naphtha to distillate ratio, thereby maximizing the yield of high-quality product.
19.3 Catalysis Hydrogenated motor-gasoline quality produced by alkene oligomerization over SPA at 3.8 MPa and with a cut point of 185 ◦ C between naphtha and distillate. Catalyst hydration was in the range 104–108% H3 PO4 .
Table 19.3
Description
Feed composition Propene n-Butenes Isobutene C5 alkenes Alkanes Operating conditions Temperature (◦ C) Space velocity (h –1 ) Performance Alkene conversion (mass%) Naphtha selectivity (mass%) Hydrogenated motor-gasoline Density at 20 ◦ C (kg·m –3 ) RON MON
Phosphoric acid oligomerization with different alkenes C3 -only
C3 –C4
C3 –C5
C4 -rich
56 0 0 0 44
57 17 2 0 24
26 32 3 5 34
26 32 3 5 34
0 51 5 4 40
0 51 5 4 40
0 61 6 4 29
200 1.3
191 –
180 0.7
160 0.7
180 0.7
160 0.7
160
85 –
– –
81 82
75 82
93 78
82 83
63 91
– 47 –
– 56 –
720 70 76
720 74 78
719 81 83
720 86 88
719 90 89
2
5) Although motor-gasoline quality and yield are improved by higher catalyst hydration, a hydration level lower than 103% H2 PO4 may lead to increased phosphoric acid loss due to leaching [30]. 6) The oligomerization temperature should be decreased as the isobutene content in the feed increases. Very high octane number alkylate-type products can be obtained when operating at low temperature (<140 ◦ C) with an isobutene-rich feed: for example, the UOP InAlk-process (‘‘indirect alkylation’’) [7]. SPA-catalyzed oligomerization can also produce a highly branched kerosene range product with very low freezing point (<–47 ◦ C), which makes it a good blending material for jet fuel. Although some jet fuels can be produced by butene oligomerization, mixed C3 –C4 and C3 feed materials are better for jet fuel production. Oligomerization over SPA mechanistically limits the carbon number distribution. The solubility of heavier alkenes in the phosphoric acid phase is limited, and longer chain alkenes do not form sufficiently stable phosphoric acid esters [38]. Both factors contribute to limit the distillate yield from SPA oligomerization. Although naphtha range alkenes can be converted over SPA, it is a poor catalyst for naphtha range alkene conversion for the reasons mentioned. It has been found that the >175 ◦ C distillate yield from the oligomerization of a mixed alkene syncrude feed with naphtha recycling is a function of the propene conversion only (Equation 19.1) [47]. Distillate yield = 0.7 · [propene converted]
(19.1)
377
378
19 Oligomerization
Complete conversion of C3 -only feed into distillate is possible with recycle operation [25]. The distillate thus produced is essentially a kerosene. Distillates produced by SPA oligomerization have low cetane numbers (typically 25–30) but excellent cold-flow properties. Such distillates make good jet fuel but poor diesel fuel. SPA-catalyzed oligomerization should therefore not be considered for diesel fuel production, although the production of diesel fuel from Fischer–Tropsch syncrude is practised industrially (Section 9.4.1). Although phosphoric acid is used as catalyst for ethene hydration (Section 17.4), SPA is not well suited for oligomerization of oxygenate-rich syncrude. Oxygenates typically present in syncrude affect SPA oligomerization in a number of ways: 1) Oxygenates inhibit oligomerization [48]. The oxygenates are not only stronger nucleophiles than the alkenes, but are also more soluble in the phosphoric acid phase. Oxygenate conversion often leads to dehydration. The water thus formed increases the hydration state of the SPA catalyst and reduces its acid strength. 2) Oxygenates participate in various acid-catalyzed side reactions [49]. 3) Oxygenates undermine the structural integrity of the SPA catalyst [48]. This is not a problem at low oxygenate content but, as the oxygenate content increases and water release increases, phosphoric acid leaching increases (minor effect), and the mechanical strength of the catalyst decreases. In contact with Fischer–Tropsch naphtha, this leads to collapse of the catalyst bed. The average catalyst lifetime obtained during industrial operation of SPA with Fischer–Tropsch syncrude is 620 kg product per kilogram catalyst [50] and, with proper operation, lifetimes on the order of 600–900 kg product per kilogram catalyst can be achieved [51]. In comparison to other heterogeneous catalysts used for oligomerization, the lifetime of SPA is low. This is cited as one of the primary reasons for selecting other catalyst types. Because SPA is a cheap catalyst, it is not regenerated, which raises concern over the environmental impact of SPA. Yet, the process is, environmentally speaking, surprisingly benign. The catalyst is produced from natural silica source without the need of templating molecules, and the spent catalyst can be disposed of in a beneficial way. By neutralizing the spent SPA catalyst with ammonia, ammonium phosphate fertilizer can be produced for agricultural use [52]. This is practised on a commercial scale in South Africa with the spent SPA catalyst from the Fischer–Tropsch syncrude refinery. Spent SPA therefore does not generate solid waste. 19.3.2 H-ZSM-5 Zeolite
Oligomerization over H-ZSM-5 (MFI-type zeolite) is usually conducted around 200–320 ◦ C and 5 MPa. At lower temperatures, H-ZSM-5 catalyzes oligomerization with limited cracking, resulting in the formation of oligomers that are multiples of the monomer [53]. Under such conditions, the catalyst is rapidly deactivated by the accumulation of heavy oligomers that restrict access to the catalyst. By operating at a sufficiently high temperature, oligomerization and cracking reactions equilibrate and are thermodynamically controlled. This has some important consequences for oligomerization:
19.3 Catalysis
1) The carbon number distribution becomes equilibrated and is not sensitive to that of the feed. It has been shown that feed materials ranging from C2 to C10 alkenes all result in oligomerization products with similar carbon number distribution [17]. 2) Extraction of chemicals from a feed to an oligomerization unit will not affect the yield of the different product fractions, because the carbon number distribution is equilibrated. This principle was demonstrated by a study to ascertain the impact of 1-hexene extraction from the HTFT feed to an H-ZSM-5 oligomerization unit [54]. 3) Although the carbon number distribution is equilibrated, it is a limited ‘‘equilibrium,’’ which does not extend to the alkene isomers. The feed insensitivity with respect to carbon number distribution does not imply insensitivity with respect to isomer distribution. The impact of feed composition, catalyst age, and operating conditions on motor-gasoline quality can be significant [20]. 4) The naphtha to distillate ratio of the product can be controlled by the operating conditions. There are bounds to the control though. At high pressure, the reaction may again become kinetically constrained and the heavy oligomers predicted from thermodynamic control may not be obtained [55]. At high temperature, hydrogen transfer reactions may take place, leading to the formation of aromatics and alkanes. These reactions become signification at temperatures above 300 ◦ C. This can be advantageous for jet fuel production. 5) A wide range of feed materials can be employed for oligomerization, and oligomerization over H-ZSM-5 is not restricted to gaseous alkenes. In fact, it was a specific design intent of the ‘‘Conversion of Olefins to Distillate’’ (COD) process to convert gaseous and naphtha range alkenes in Fischer–Tropsch syncrude [18]. Oligomerization takes place by classic Brønsted acid catalysis (Figure 19.1), but the zeolite pore structure of H-ZSM-5 imposes a geometric constraint on oligomerization. The H-ZSM-5 catalyst has an MFI zeolite structure with 0.51 × 0.55 nm and 0.53 × 0.56 nm channels [56]. The pore-constrained geometry reduces the degree of branching of the oligomers by imposing transition-state selectivity on the mechanism [53]. Reactant and transition-state selectivity effects of H-ZSM-5 are also discussed in Section 21.4.1. The lower degree of branching has a direct impact on the product quality (Table 19.4) [14, 18–20, 57]. The more linear structure of the distillate range material is advantageous for diesel fuel production. The lower degree of branching improves the cetane number, while the presence of at least some branching improves the cold-flow properties of the diesel fuel. The olefinic motor-gasoline properties is not as good as that obtained by SPA oligomerization (Table 19.2), but as is explained, the lower degree of branching is only partly to blame for the poorer octane number. The octane number of the olefinic motor-gasoline seems extraordinarily low for a branched olefinic product, even though the degree of branching is not very high. This is a consequence of the feed selection. When a naphtha range syncrude is employed as feed material for oligomerization, the naphtha range alkanes are not converted and are retained in the naphtha fraction. Since syncrude contains mainly linear alkanes, the resulting olefinic naphtha range product has a lower octane number on account of the alkanes. Despite the low coking propensity of H-ZSM-5 catalysts, the catalyst has to be regenerated every three to six months by controlled coke burnoff. The catalyst lifetime extends over multiple cycles. With time on stream and as the catalyst ages over time, the strongest acidic sites are deactivated first. This is beneficial, since it allows skeletal isomerization of the shorter chain
379
380
19 Oligomerization Table 19.4 Olefinic motor-gasoline and hydrogenated diesel fuel properties obtained from the H-ZSM-5-catalyzed oligomerization of various alkene-rich feed materials.
Fuel properties
MOGD process FCC C3 –C4
Olefinic motor-gasoline Density at 15.6 ◦ C (kg·m –3 ) RON MON Hydrogenated diesel fuel Density at 15.6 ◦ C (kg·m –3 ) Cetane number Viscosity at 40 ◦ C (cSt) Cold filter plugging point (◦ C) Pour point (◦ C) T90 distillation (◦ C) FBP distillation (◦ C)
COD process HTFT C3 –C6
H-ZSM-5 pilot plant Propene
HTFT C5 –C6
HTFT C7 –C9
730 92 79
– 81–85 74–75
738 85 75
– 80 –
– 76–80 –
– 66–67 –
779 52–56 2.5 – <–50 342 –
787a 52–54 2.55 <–35 <–51 320 358
801a 51 – – – 323 361
– 40 – <–25 – 365 388
– 45–48 – <–25 – 345–352b –
– 41–43 – – – –
at 20 ◦ C. distillation.
a Density b T95
alkenes to improve the octane number without cracking, and it reduces distillate cracking to improve the distillate yield. Single-pass distillate yield is around 65%, which can be improved to around 84% with naphtha recycle operation [14]. Oligomerization activity over H-ZSM-5 is reduced by water and oxygenates in the alkene feed [58]. Water and alcohols result in catalyst inhibition, and catalyst activity is restored when the water and alcohols are removed from the feed. Carboxylic acids and ketones result in stronger inhibition as well as deactivation. This can be explained by the acid-catalyzed interconversion of carbonyls and carboxylic acids, combined with carbonyl aromatization (Figure 16.4). When alcohols are present in the mixture, the deactivating effect of carboxylic acids is suppressed [59]. In industrial practice, H-ZSM-5 oligomerization is applied with oxygenate-containing HTFT feed despite some inhibition. 19.3.3 Amorphous Silica–Alumina
Amorphous silica–alumina (ASA), like H-ZSM-5, is composed of SiO2 and Al2 O3 , but there are a number of important differences that affect oligomerization. In order to differentiate these two catalysts, let us look at the differences that are expected to affect oligomerization behavior: 1) ASA by definition does not have the crystalline nature of H-ZSM-5. Oligomerization is not subject to the transition-state limitations imposed by a pore-constraining geometry. The products should therefore be more branched than those obtained by H-ZSM-5.
19.3 Catalysis
2) The acid strength of ASA is lower than that of H-ZSM-5 with the same SiO2 to Al2 O3 ratio. It has also been found that lower acid strength sites are in fact more effective for oligomerization than higher acid strength sites [60]. As a consequence, the activity of ASA may not be lower, but the cracking propensity is likely to be lower and one would anticipate that a heavier product can be produced at comparable operating conditions. 3) The ratio of hydrogen transfer to oligomerization activity of ASA is an order of magnitude higher than that of H-ZSM-5 [61]. This implies that ASA is more prone to produce alkanes, dienes, cycloalkanes, cycloalkenes, and aromatics. The density of the product from ASA-catalyzed oligomerization is likely to be higher than that from H-ZSM-5. 4) Oligomerization over ASA should take place by the standard Brønsted acid-catalyzed carbocation mechanism (Figure 19.1a). However, it has been reported that the ASA oligomerization is cis-selective [62], and that the dimerization products from 1-butene have a higher blending octane number than those from 2-butenes [63]. There is consequently some macroscopic evidence that the oligomerization mechanism over ASA may involve some ester-like intermediates (Figure 19.1b). The behavior predicted from catalysis fundamentals was indeed observed during the oligomerization of Fischer–Tropsch syncrude over ASA catalysts (Table 19.5) [64, 65]. The higher degree of branching is reflected in the fuel properties, and the distillate in general was much heavier. The hydrogen transfer activity was quite high, and at 175 ◦ C the alkane yield reached 5% and monotonically increased with increasing temperature [64]. This reduced the distillate yield and motor-gasoline quality by converting naphtha range alkenes into alkanes. With Fischer–Tropsch feed materials that were low in oxygenates, catalyst cycle lengths of 100–120 days could be obtained with a start-of-run temperature of 180 ◦ C [64]. At similar start-of-run conditions, feed materials containing 1–4% oxygenates resulted in cycle lengths of 80–90 days [65]. Oligomerization activity could be completely restored at the end of each cycle by controlled oxidative regeneration. It was also recommended that the start-of-run temperature of ASA catalysts should be reduced to 110 ◦ C with Fischer–Tropsch-derived feeds. Under such start-up conditions, cycle lengths of eight months have been obtained with oxygenate-containing feed [65]. ASA catalysts are tolerant to oxygenates, are able to perform partial deoxygenation of the feed, and are active for carbonyl and carboxylic acid interconversion at typical operating conditions. Oligomerization over ASA competes with H-ZSM-5 as oxygenate-tolerant distillate producing technology. The distillate yield over ASA is lower, and the ultimate yield that can be obtained by recycle operation varies from 63 to 76% depending on the feed [64]. The main difference is in distillate density, a quality that is lacking in Fischer–Tropsch distillate for diesel fuel production [66]. Although ASA-derived distillate has a much higher density, the cetane number is also much lower, detracting from its use for the production of synthetic diesel fuel. 19.3.4 Acidic Resin
The use of acidic resin catalysts for oligomerization is mainly aimed at the production of high octane number alkylate-equivalent motor-gasoline by selective isobutene dimerization. There are many different types of acidic resins that can be employed, and sulfonated
381
382
19 Oligomerization Table 19.5 Fuel properties of products obtained by the oligomerization of different Fischer–Tropsch-derived feed materials over amorphous silica–alumina.
Description
ASA oligomerization with different alkene feeds
Operating conditions Temperature (◦ C) Pressure (MPa) Space velocity (h –1 ) Feed oxygenate content (%) Distillate yield, >177 ◦ C (%) Olefinic motor-gasoline Density at 15.6 ◦ C (kg·m –3 ) RON MON Reid vapor pressure (kPa) Alkene content (g Br/100 g) Hydrogenated motor-gasoline RON Diesel fuel, hydrogenated Density at 15.6 ◦ C (kg·m –3 ) Cetane number Viscosity at 40 ◦ C (cSt) Cold filter plugging point (◦ C) Pour point (◦ C) T90 distillation (◦ C) FBP distillation (◦ C) a Unhydrogenated
C3 –C6 HTFT condensate
C7 –C10 SPA oligomers
C5 -105 ◦ C HTFT alumina treated
C5 -105 ◦ C HTFT water washed
140–235 6 0.5 <0.01 65–67
180–210 3.5 0.5 0.05 52–60
180–230 6 0.5 0.5–0.8 52–57
225–280 6 0.5 0.5–3.6 52–55
707 92–94 71–72 72 82–114
708–714 78–82 – – 44–62
– 76 – – 66
– 74 – – 44
76–85
–
–
–
810–816a 28–29 2.8–3.4a –15a
809–810 29–30 3.5–3.6 <–20
810 37 2.5 <–20
810 37 2.8 –
329–348 434–451
363–364 462–463
347 452
346 448
properties.
styrene–divinylbenzene-based resins are commonly used in industrial oligomerization processes. The process is moderated by the addition of polar compounds, typically 2-methyl-2-propanol (tert-butanol), to maximize dimerization selectivity and limit heavy oligomer formation [67, 68]. Dimerization is conducted in the liquid phase at <100 ◦ C. The product from selective isobutene dimerization can be hydrogenated to give a high-octane alkylate-equivalent product that is rich in trimethylpentanes. The motor-gasoline from this type of indirect alkylation has an RON of 99–101 and MON of 96–99 [8]. Oligomerization of a mixture of n-alkenes and branched alkenes results in selective conversion of the branched alkenes, with little n-alkene conversion taking place. Acidic resin catalysts are quite capable of isomerizing and oligomerizing n-alkenes, but, under typical selective dimerization conditions, these reactions are effectively suppressed. Acidic resin–catalyzed oligomerization of syncrude can therefore be performed, and the products are anticipated to be typical of those formed by a Brønsted acid–catalyzed carbocation mechanism (Figure 19.1a). There
19.3 Catalysis
is consequently no specific benefit to be derived from using acidic resin catalysts for the oligomerization of syncrude. 19.3.5 Homogeneous Nickel
Homogeneous catalysts can be designed to enable very selective conversion. Catalyst addition rate can also be varied to compensate for feed rate or feed poisons, and such processes can deal well with variable load operation. The main drawback of homogeneously catalyzed processes is the separation of the catalyst from the product. The industrial homogeneous-catalyzed oligomerization processes make use of a nickel-based Ziegler-type catalyst system, and alkene oligomerization takes place by the organometallic insertion mechanism (Figure 19.1d) [21]. Typical process conditions are 40–50 ◦ C, 1–3 MPa, and 1–5 h residence time [35]. A number of process variants are commercially available to produce naphtha range products (Table 19.6) [35, 69, 70]: 1) Dimersol E is used for the oligomerization of ethene and the light off-gas (C2 –C3 alkenes) from fluid catalytic cracking or steam cracking units to produce olefinic motor-gasoline. 2) Dimersol G is used for the oligomerization of propene and C3 –C4 alkene mixtures to produce olefinic motor-gasoline. 3) Dimersol X is used for butene dimerization to produce octenes with a low degree of branching for plasticiser alcohol manufacturing. The olefinic motor-gasoline quality obtained by this process is low. Reduced catalyst consumption and improved selectivity can be achieved by employing the Difasol modification of this process, which conducts oligomerization in a biphasic medium with ionic liquids [71, 72]. The organometallic catalyst is sensitive to any impurities that will complex with the nickel; among others, it is sensitive to dienes, alkynes, water, and sulfur, which should not exceed 5–10 µg g−1 . Attention must be paid to feed pretreatment when using any of these technologies with Fischer–Tropsch-derived feed. The Dimersol E process was successfully employed for variable load ethene conversion from HTFT synthesis (Section 9.5.3). Unless there are significant variable Olefinic motor-gasoline properties obtained by homogeneous nickel-catalyzed oligomerization of C2 –C4 alkenes. Table 19.6
Fuel properties Alkene feed range Olefinic motor-gasoline Density at 15.6 ◦ C (kg·m –3 ) RON MON Reid vapor pressure (kPa) a Calculated
from composition data.
Dimersol E
Dimersol G
Dimersol Xa
C2
C3 –C4
C4
720 93–94 79–80 –
692–700 96–96.5 80–82 48
730 79 71 5
383
384
19 Oligomerization
load constraints, there is no advantage in using homogeneous nickel oligomerization for olefinic motor-gasoline production over the use of SPA oligomerization. There is a competitive advantage to employ the Dimersol X and Difasol technologies for chemical production from Fischer-Tropsch-derived butenes. Fischer–Tropsch butenes have a low isobutene content and high 1-butene content, which increases selectivity to linear dimers. 19.3.6 Thermal Oligomerization
Thermal oligomerization is a noncatalytic process following a free radical mechanism (Figure 19.1c). The mechanism is not skeletally isomerizing, and skeletally isomerized products are formed only at higher temperatures (>400 ◦ C) due to the combined effects of thermal oligomerization and thermal cracking. On the basis of its mechanism, some attributes of thermal oligomerization can be inferred: 1) 2) 3)
4)
5)
6)
Thermal oligomerization is less efficient than catalytic oligomerization. This is one of the reasons for the displacement of thermal by catalytic processes in general. Oxygenates are beneficial for thermal oligomerization on account of their lower homolytic bond dissociation energies [73]. The oxygenates will act as radical initiators. The oligomerization product will have a low degree of branching. The degree of branching will reflect the carbon number and degree of branching of the feed material. Lighter and more branched alkenes will lead to a more branched product, whereas a more linear and heavier alkene feed will lead to a less branched product. The olefinic motor-gasoline quality from thermal oligomerization will be lower than that from acid-catalyzed processes. Under thermal oligomerization conditions that do not promote substantial aromatization, olefinic motor-gasoline from thermal oligomerization has an MON of 76–78 [74, 75]. The distillate range product will benefit from the low degree of branching. Even with a comparatively light C5 –C6 Fischer–Tropsch feed, the distillate obtained by thermal oligomerization had a cetane number of ≥54 after hydrogenation [76]. Thermal oligomerization should be able to produce good quality lubricant base oils from longer chain alkenes. The mechanism suggests that oligomerization of n-1-alkenes should yield PAO-like lubricant base oils. Since the alkenes in Fischer–Tropsch syncrude are mainly n-1-olefins, it should make a good feedstock for lubricant base oil production. This was indeed found in practice, with purified n-1-alkenes and cracked Fischer–Tropsch wax producing good quality lubricating oils (Table 19.7) [77]. Although C5 –C6 HTFT naphtha is a poor feed material for lubricating oil production, it was nevertheless possible to produce an oil with a viscosity index of 100 [76]. Equal or better results were reported with HTFT material from the Hydrocol process (Chapter 7) [77].
Thermal oligomerization of Fischer–Tropsch gaseous and naphtha range alkenes gives products that have characteristics similar to that of H-ZSM-5 oligomerization (Table 19.4). However, thermal oligomerization is not as efficient as catalytic oligomerization and it requires higher operating temperatures than catalytic oligomerization. Attempts to reduce the operating temperature by making use of radical initiators such at di-tertiary-butyl peroxide failed because
19.4 Syncrude Process Technology Thermal oligomerization at around 320 ◦ C and 10 h residence time to produce lubricant base oils from different alkene-based feed materials.
Table 19.7
Description
Product yield (mass%) Atmospheric distillate + gases Vacuum distillate Lubricating oil Lubricating oil propertiesb Density at 15.6 ◦ C (kg·m –3 ) Viscosity index Viscosity at 37.8 ◦ C (cSt) Viscosity at 98.9 ◦ C (cSt) Pour point (◦ C) a
Thermal oligomerization with different alkene feeds Ethene
1Hexene
1Octene
1Decene
1Dodecene
1Hexadecene
Cracked wax
– – 40
74 16 10
25 30 45
2 56 42
5 45 50
15 30 55a
– – 34
837.8 115 58.0 8.1 <–34
829.4 85 16.6 3.4 <–54
833.8 128 29.8 5.4 <–57
835.3 143 31.9 6.1 –29
829.4 ∼154 27.7 5.8 –4
836.8 146 39.0 7.4 +24
– 137 – 4.8 –34
Thermal oligomerization at 350 ◦ C. oils were unhydrogenated.
b Lubricating
of low initiator productivity [78]. Thermal oligomerization is therefore not recommended for the production of transportation fuels. However, thermal oligomerization has potential as technology for the production of lubricant base oils from straight-run Fischer–Tropsch syncrude. There is no need to remove oxygenates, which simplifies the process considerably. The oxygenates are actually beneficial. In this respect, thermal oligomerization has an advantage over catalytic routes for lubricant base oil production. Catalyst systems like AlCl3 [26] and BF3 [79], which were previously used for lubricant base oil production from Fischer–Tropsch n-1-alkenes, are sensitive to water and oxygenates. Using such catalytic technologies to produce lubricant base oils from syncrude requires extensive feed pretreatment. 19.4 Syncrude Process Technology
There are three important differences in syncrude composition relative to crude oil that are pertinent to oligomerization: alkene availability is comparatively high; the dominant alkene species for each carbon number is the n-1-alkene; and there is usually oxygenates or water present with the alkenes. Some guidelines are given below: 1) Oligomerization technologies that rely on a branched alkene feed are at a disadvantage when used in a Fischer–Tropsch refinery. The disadvantage may be overcome by employing skeletal isomerization (Chapter 18) or catalytic cracking (Chapter 21) technologies that produce branched alkenes. However, if such conversion processes must be installed in order to make the oligomerization technology work, one should reconsider the choice
385
386
19 Oligomerization
2)
3)
4)
5)
6)
of oligomerization technology. Often, there will be a different but overall more efficient refining pathway. Catalysts and technologies that are very sensitive to oxygenates or water are at a disadvantage when applied in a Fischer–Tropsch refinery. In some instances, the alkenes are available from a processing step that removes oxygenates and water (e.g., ethene purification) and then oxygenate tolerance is not an issue. However, if extensive feed pretreatment is required to make the oligomerization technology work, one should reconsider the choice of oligomerization technology. Conversely, some oligomerization catalysts are also effective deoxygenation catalysts. Some oligomerization technologies benefit from n-1-alkenes in the feed. Depending on the refining objective, it is worthwhile considering such oligomerization technologies first. These technologies are oligomerization of butenes over SPA to produce motor-gasoline, oligomerization of butenes over homogeneous nickel to produce chemicals, and thermal oligomerization of heavier naphtha and distillate range straight-run syncrude to produce lubricant base oils. The carbon number distribution of the feed to an oligomerization process can be critical to the quality of the product. In such cases, care should be taken to properly fractionate the feed. Alkene oligomerization is very exothermic, and heat management is crucial. Some catalysts have a threshold temperature where mechanistic changes take place. In such cases, it is important to keep the temperature within the desired mechanistic regime. When the oligomerization selectivity is very temperature sensitive, or when it is important to maximize dimerization selectivity, better control can be achieved by employing a reactive quench with dilute alkene feed (Figure 19.2). In this way, the reactive short-chain alkene feed is always diluted in an alkane matrix and the alkene concentration is lower. It also allows similar-sized catalyst beds to be employed, which holds an advantage for mechanical integrity of the catalyst. Different bed heights place a disparate amount of mechanical stress on the catalyst at the bottom of the largest catalyst bed.
Oligomerization Alkene-rich fresh feed
Alkane-rich recycle
LPG or light alkanes Olefinic product Figure 19.2 Reactive quenching with a dilute alkene feed to the top of the oligomerization reactor as a strategy to improve heat management and dimerization selectivity.
19.4 Syncrude Process Technology
7) In cases where cycle length is restricted by pressure drop, it is often found that the top of the first catalyst bed is fouled. It may also happen that the cycle length is determined by deactivation of the first catalyst bed. In these instances, it is advantageous, from an operational point of view, to place the first oligomerization bed in a separate vessel. One should also consider having two such beds in parallel for swing-mode operation. Metal carboxylates in the syncrude feed and reactive species, such as oxygenates, may justify such a configuration. 8) Oligomerization by SPA is a good technology for producing high octane number olefinic motor-gasoline, high octane number alkylate-equivalent paraffinic motor-gasoline, and high-quality jet fuel. All these products can be produced in high yield. SPA-catalyzed oligomerization can also be employed to coprocess benzene to reduce refinery benzene levels. For these applications, SPA-based oligomerization is a proven and recommended technology with Fischer–Tropsch syncrude. 9) In refinery designs where SPA-catalyzed oligomerization is employed for alkenes spanning more than one carbon number, one should consider the use of two separate oligomerization units. Butenes should be oligomerized separately from other alkenes if hydrogenated motor-gasoline is a final product. 10) Oligomerization by H-ZSM-5 is a good technology for producing high-quality distillate range material in high yield. It has the further advantage that it is capable of processing oxygenate-containing Fischer–Tropsch naphtha, as well as gaseous alkenes. It can therefore double as an oligomerization and deoxygenation process. For these applications, it is a recommended Fischer–Tropsch oligomerization technology and it has been industrially proven with Fischer–Tropsch syncrude. 11) When an oligomerization technology is employed with a naphtha range syncrude feed, one must carefully consider the effect of the naphtha range alkanes on the final naphtha range product. The alkanes are not converted. When the product is a blending component for motor-gasoline, it may be worthwhile to fractionate the naphtha range product in such a way that the fraction containing the feed alkanes is separated from the rest of the product (Figure 19.3). This alkane-containing fraction may then be separately refined to upgrade its Oligomerization Recycle
Syncrude
Light key: C7 -alkenes
Heavy key: C7 -alkanes
Light key: C6-alkenes Heavy key: C7 -alkanes
Figure 19.3 Oligomerization with naphtha range feed, illustrating the separation of feed alkanes in the product for separate downstream processing and partial recycle to aid heat management. This type of configuration is beneficial if olefinic motor-gasoline is one of the products from oligomerization.
Alkane-rich (to refinery) Product (to product fractionation) Syncrude (to refinery)
387
388
19 Oligomerization
octane number. Ideally, one would like to restrict the alkanes in the feed to C6 and lighter material, but this is not always possible. Part of the alkane-containing product fraction may also be used as recycle to improve heat management. 12) Product fractionation after oligomerization can be employed to improve the properties of the products. For example, if both olefinic and paraffinic motor-gasoline are produced, a cut point at 105–110 ◦ C will maximize the octane number that can be obtained by hydrogenating the overhead fraction. This cut point allows the trimethylpentenes to be separated from the dimethylhexenes and less branched products. In an analogous way, it is possible to find optimum cut points for lighter and heavier carbon number cuts. There are two learning points from this example. Firstly, the cut point affects quality, and, secondly, it may be worthwhile investing in sharper product fractionation after oligomerization. 13) Acid-catalyzed oligomerization processes are capable of carbonyl and carboxylic acid interconversion. If the syncrude feed contains either compound class, the material of construction of the oligomerization reactor and downstream equipment should be selected to cope with corrosive short chain carboxylic acids. References 1. De Klerk, A. (2008) Refining of
2.
3.
4.
5.
6.
7.
8.
Fischer–Tropsch syncrude: lessons from the past. Prepr. Pap.-Am. Chem. Soc., Div. Pet. Chem., 53 (2), 105–109. De Klerk, A. (2009) in Advances in Fischer–Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, FL, pp. 331–364. De Klerk, A. (2010) Deactivation in Fischer–Tropsch synthesis and its impact on refinery design. Prepr. Pap.-Am. Chem. Soc., Div. Pet. Chem., 55 (1), 86–89. Gary, J.H., Handwerk, G.E., and Kaiser, M.J. (2007) Petroleum Refining. Technology and Economics, 5th edn, CRC Press, Boca Raton, FL. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Botes, F.G. and B¨ohringer, W. (2004) The addition of HZSM-5 to the Fischer–Tropsch process for improved gasoline production. Appl. Catal. A, 267, 217–225. Meister, J.M., Black, S.M., Muldoon, B.S., Wei, D.H., and Roeseler, C.M. (2000) Optimize alkylate production for clean fuels. Hydrocarbon Process., 79 (5), 63–75. Birkhoff, R. and Nurminen, M. (2004) in Handbook of Petroleum Refining Processes, 3rd edn (ed. R.A. Meyers), McGraw-Hill, New York, pp. 1.3–1.9.
9. Lamberth, R. (2004) 2003 was a year of transi-
10.
11.
12.
13.
14.
15.
16.
tion for the MTBE, fuels industry. Oil Gas J., 102 (2), 52–58. De Klerk, A. and De Vaal, P.L. (2008) Alkylate technology selection for Fischer–Tropsch syncrude refining. Ind. Eng. Chem. Res., 47, 6870–6877. Ipatieff, V.N. and Komarewsky, V.I. (1937) Hydro-polymerization. J. Am. Chem. Soc., 59, 720–722. Ipatieff, V.N. and Komarewsky, V.I. (1937) Iso¨octane production by simultaneous polymerization and hydrogenation. Ind. Eng. Chem., 29, 958. Sakuneka, T.M., De Klerk, A., Nel, R.J.J., and Pienaar, A.D. (2008) Synthetic jet fuel production by combined propene oligomerization and aromatic alkylation over solid phosphoric acid. Ind. Eng. Chem. Res., 47, 1828–1834. Tabak, S.A. and Krambeck, F.J. (1985) Shaping process makes fuels. Hydrocarbon Process., 64 (9), 72–74. De Klerk, A. and Nel, R.J.J. (2008) Benzene reduction in a fuels refinery: an unconventional approach. Energy Fuels, 22, 1449–1455. Sakuneka, T.M., Nel, R.J.J., and De Klerk, A. (2008) Benzene reduction by alkylation in a solid phosphoric acid catalyzed olefin oligomerization process. Ind. Eng. Chem. Res., 47, 7178–7183.
References 17. Garwood, W.E. (1983) Conversion of C2 -C10
18.
19.
20.
21.
22.
23.
24. 25. 26. 27.
28.
29.
30.
31. 32.
to higher olefins over synthetic zeolite ZSM-5. ACS Symp. Ser., 218, 383–396. K¨ohler, E., Schmidt, F., Wernicke, H.J., De Pontes, M., and Roberts, H.L. (1995) Converting olefins to diesel – the COD process. Hydrocarbon Technol. Int., (Summer), 37–40. Knottenbelt, C. (2002) Mossgas ‘‘gas-to-liquids’’ diesel fuels – an environmentally friendly option. Catal. Today, 71, 437–445. De Klerk, A. (2007) Properties of synthetic fuels from H-ZSM-5 oligomerization of Fischer–Tropsch type feed material. Energy Fuels, 21, 3084–3089. Chodorge, J.A., Cosyns, J., Hughes, F., and Olivier-Bourbigou, H. (1997) Dimerisation of C4 olefins for the manufacture of isononanols. DGMK-Conference on ‘‘C4 Chemistry – Manufacture and use of C4 Hydrocarbons’’, Aachen, Germany, pp. 227–233. (2001) Linear Alpha Olefins. SRI Process Economics Program Report 12D, Stanford Research Institute, Menlo Park, CA. Dixon, J.T., Green, M.J., Hess, F.M., and Morgan, D.H. (2004) Advances in selective ethylene trimerisation – a critical overview. J. Organomet. Chem., 689, 3641–3668. Tullo, A.H. (2009) Sasol mulls Gulf Coast octene plant. Chem. Eng. News, 87 (47), 16. Jones, E.K. (1954) Cumene and tetramer production. Pet. Refiner, 33 (12), 186–187. Horne, W.A. (1950) Review of German synthetic lubricants. Ind. Eng. Chem., 42, 2428–2436. De Klerk, A. (2005) Oligomerization of 1-hexene and 1-octene over solid acid catalysts. Ind. Eng. Chem. Res., 44, 3887–3893. Schmerling, L. and Ipatieff, V.N. (1950) The mechanism of the polymerization of alkenes. Adv. Catal., 2, 21–80. Oblad, A.G., Mills, G.A., and Heinemann, H. (1958) in Alkylation, Isomerization, Polymerization, Cracking and Hydroreforming, Catalysis, Vol. VI (ed. P. H. Emmett), Reinhold, New York, pp. 341–406. McMahon, J.F., Bednars, C., and Solomon, E. (1963) in Advances in Petroleum Chemistry and Refining, Vol. 7 (eds K.A. Kobe and J.J. Mc Ketta Jr.), John Wiley & Sons, Inc., New York, pp. 285–321. Asinger, F. (1968) in Mono-olefins chemistry and technology, Pergamon, Oxford. O’Connor, C.T. and Kojima, M. (1990) Alkene oligomerization. Catal. Today, 6 (3), 329–349.
33. Skupinska, J. (1991) Oligomerization of
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
α-olefins to higher oligomers. Chem. Rev., 91, 613–648. ˚ S.G. (1999) in Sanati, M., H¨ornell, C., and J¨aras, Catalysis, Vol. 14 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge, pp. 236–287. Leprince, P. (2001) in Petroleum Refining, Conversion Processes, Vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 321–331. Pujad´o, P.R. and Ward, D.J. (2006) in Handbook of Petroleum Processing (eds D.S.J. Jones and P.R. Pujad´o), Springer, Dordrecht, pp. 372–399. De Klerk, A. (2008) Fischer–Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. De Klerk, A. (2011) in Catalysis, Vol. 23 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge, pp. 1–49. Ipatieff, V.N. and Corson, B.B. (1936) Gasoline from ethylene by catalytic polymerization. Ind. Eng. Chem., 28, 860–863. Deeter, W.F. (1950) Propylene polymerization for motor-gasoline production. Oil Gas J., 48 (46), 252–258. Ipatieff, V.N. and Schaad, R.E. (1938) Mixed polymerization of butenes by solid phosphoric acid catalyst. Ind. Eng. Chem., 30, 596–599. Ipatieff, V.N., Schaad, R.E., and Shanley, W.B. (1953) in The Science of Petroleum, Synthetic Products of Petroleum, Vol. 5, Part II (eds B.T. Brooks and A.E. Dunstan), Oxford University Press, London, pp. 14–23. Ipatieff, V.N., Corson, B.B., and Egloff, G. (1935) Polymerization, a new source of gasoline. Ind. Eng. Chem., 27, 1077–1081. De Klerk, A., Engelbrecht, D.J., and Boikanyo, H. (2004) Oligomerization of Fischer–Tropsch olefins: effect of feed and operating conditions on hydrogenated motor-gasoline quality. Ind. Eng. Chem. Res., 43, 7449–7455. De Klerk, A. (2004) Isomerization of 1-butene to isobutene at low temperature. Ind. Eng. Chem. Res., 43, 6325–6330. De Klerk, A., Leckel, D.O., and Prinsloo, N.M. (2006) Butene oligomerization by phosphoric acid catalysis: Separating the effects of temperature and catalyst hydration on product selectivity. Ind. Eng. Chem. Res., 45, 6127–6136. De Klerk, A. (2006) Distillate production by oligomerization of Fischer–Tropsch olefins over solid phosphoric acid. Energy Fuels, 20, 439–445.
389
390
19 Oligomerization 48. Mashapa, T.N. and De Klerk, A. (2007) Solid
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
phosphoric acid catalysed conversion of oxygenate containing Fischer–Tropsch naphtha. Appl. Catal. A, 332, 200–208. De Klerk, A., Nel, R.J.J., and Schwarzer, R. (2007) Oxygenate conversion over solid phosphoric acid. Ind. Eng. Chem. Res., 46, 2377–2382. Coetzee, J.H., Mashapa, T.N., Prinsloo, N.M., and Rademan, J.D. (2006) An improved solid phosphoric acid catalyst for alkene oligomerization in a Fischer–Tropsch refinery. Appl. Catal. A, 308, 204–209. De Klerk, A. and Prinsloo, N.M. (2009) Solid phosphoric acid: Have the sands of time run out for this steadfast catalyst? Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 211–212. Van der Merwe, W. (2010) Conversion of spent solid phosphoric acid catalyst to environmentally friendly fertilizer. Environ. Sci. Technol., 44, 1806–1812. De Klerk, A. (2007) Properties of synthetic fuels from H-ZSM-5 oligomerisation of Fischer–Tropsch type feed materials. Energy Fuels, 21, 3084–3089. Minnie, O.R., Petersen, F.W., and Samadi, F.R. (2003) Effect of 1-hexene extraction on the COD process conversion of olefins to distillate. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd083. Quann, R.J., Green, L.A., Tabak, S.A., and Krambeck, F.J. (1988) Chemistry of olefin oligomerization over ZSM-5 catalyst. Ind. Eng. Chem. Res., 27, 565–570. Baerlocher, C., Meier, W.M., and Olson, D.H. (2001) Atlas of Zeolite Framework Types, 5th edn, Elsevier, Amsterdam, pp. 184–185. Tabak, S.A., Krambeck, F.J., and Garwood, W.E. (1986) Conversion of propylene and butylene over ZSM-5 catalyst. AIChE J., 32, 1526–1531. O’Connor, C.T., Langford, S.T., and Fletcher, J.C.Q. (1992) The effect of oxygenates on the propene oligomerization activity of ZSM-5. Proceedings of the 9th International Zeolite Conference, Montreal, Canada, 1992, pp. 467–474. Chang, C.D., Chen, N.-Y., Koenig, L.R., and Walsh, D.E. (1983) Synergism in acetic acid/methanol reactions over ZSM-5 zeolites. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 28 (2), 146–152.
60. Holm, V.C.F., Bailey, G.C., and Clark, A. (1959)
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
Acidity studies of silica-alumina catalysts. J. Phys. Chem., 63, 129–133. Miller, S.J. (1987) Olefin oligomerization over high silica zeolites. Stud. Surf. Sci. Catal., 38, 187–197. Lucchesi, P.J., Baeder, D.L., and Longwell, J.P. (1959) Stereospecific isomerization of butene-1 to butene-2 over SiO2 -Al2 O3 catalyst. J. Am. Chem. Soc., 81, 3235–3237. Golombok, M. and De Bruijn, J. (2000) Dimerization of n-butenes for high octane gasoline components. Ind. Eng. Chem. Res., 39, 267–271. De Klerk, A. (2006) Oligomerization of Fischer–Tropsch olefins to distillates over amorphous silica-alumina. Energy Fuels, 20, 1799–1805. De Klerk, A. (2007) Effect of oxygenates on the oligomerization of Fischer–Tropsch olefins over amorphous silica-alumina. Energy Fuels, 21, 625–632. De Klerk, A. (2009) Can Fischer–Tropsch syncrude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. Marchionna, M., Di Girolamo, M., and Patrini, R. (2001) Light olefins dimerization to high quality gasoline components. Catal. Today, 65, 397–403. Honkela, M.L. and Krause, A.O.I. (2003) Influence of polar components in the dimerization of isobutene. Catal. Lett., 87, 113–119. Benedek, W.J. and Mauleon, J.-L. (1980) Dimersol scores high in commercial operation. Oil Gas J., 78 (17), 77–83. Chauvin, Y., Hennico, A., Leger, G., and Nocca, J.L. (1990) Upgrading of C2 , C3 , C4 olefins by IFP Dimersol technology. Erdoel, Erdgas, Kohle, 106 (7/8), 309–315. Favre, F., Foresti`ere, A., Hughes, F., Olivier-Bourbigou, H., and Chodorge, J.A. (2005) Butenes dimerization: from monophasic Dimersol to biphasic Difasol. Oil Gas Eur. Mag., 31 (2), 83–87. Foresti`ere, A. and Favre, F. (2005) in Multiphase Homogeneous Catalysis, Vol. 2 (eds B. Cornils, W.A. Herrmann, I.T. Horv´ath, W. Leitner, S. Mecking, H. Olivier-Bourbigou, and D. Vogt), Wiley-VCH Verlag GmbH, Weinheim, pp. 547–560. Blanksby, S.J. and Ellison, G.B. (2003) Bond dissociation energies of organic molecules. Acc. Chem. Res., 36, 255–263. Cooke, M.B., Swanson, H.R., and Wagner, C.R. (1935) Thermal process for polymerizing
References olefin-bearing gases. Refiner Nat. Gas. Manuf., 14 (11), 497–505. 75. Carey, J.S. (1936) Commercial aspects of the unitary thermal polymerization process. Refiner Nat. Gas. Manuf., 15 (12) 549–554, 576. 76. De Klerk, A. (2005) Thermal upgrading of Fischer–Tropsch olefins. Energy Fuels, 19, 1462–1467. 77. Seger, F.M., Doherty, H.G., and Sachanen, A.N. (1950) Noncatalytic polymerization of olefins
to lubricating oils. Ind. Eng. Chem., 42, 2446–2452. 78. Cowley, M. (2007) Oligomerisation of alkenes by radical initiation. Org. Process Res. Dev., 11, 286–288. 79. Ranwell, A. (1994) Die oligomerisasie van Sasol alfa-olefienfraksies. (Engl. Transl. ‘‘The oligomerisation of Sasol alpha-olefin fractions’’). M.Sc. dissertation, Rand Afrikaans University, Johannesburg, South Africa.
391
393
20 Aromatic Alkylation 20.1 Introduction
Aromatic alkylation entails the addition of an alkene to an aromatic and can also be performed with other alkylating agents, such as alcohols. The subsequent discussion is focused on aromatic alkylation by alkenes. Aromatic alkylation should not be confused with aliphatic alkylation (Section 16.4.5), which is often employed in crude oil refining. Although both reactions involve the addition of an alkene to another hydrocarbon, the catalysts, chemistry, and process technologies involved are very different. Fuel refineries seldom have an aromatic alkylation unit, which is normally associated only with petrochemical production. In future, this may change as the maximum allowable benzene content in motor-gasoline keeps decreasing. Aromatic alkylation is one of the more promising conversion technologies to reduce refinery benzene, since it is not accompanied by a deterioration of the fuel properties and can be applied to mixtures [1, 2]. As with all alkene-based refinery conversion processes, it is reliant on alkenes, which is not as readily available in crude oil refineries as it is in Fischer–Tropsch refineries. In this respect, Fischer–Tropsch syncrude has an advantage in being alkene-rich. Aromatic alkylation can be considered as conversion technology in a Fischer–Tropsch refinery design for a number of reasons: 1) Ethene is a major primary product from HTFT synthesis and it is a valuable commodity chemical. However, when an HTFT facility is far from petrochemical markets, the ethene must be converted on-site. We have encountered this challenge before (Section 17.4). Depending on the size and logistics of the facility, the best refining pathway may be to convert the ethene into a transportation fuel blending component. This is also true of LTFT facilities, where there is less ethene that can be extracted for petrochemical production. Alkylation of aromatics with ethene is one of the preferred refining pathways to incorporate ethene into fuels. 2) Refinery benzene levels can be reduced by coprocessing the benzene with alkenes in other acid-catalyzed refinery units [2]. It is especially well matched with oligomerization (Chapter 19). This application has been successfully demonstrated on an industrial scale with HTFT syncrude in a solid phosphoric acid (SPA)-catalyzed oligomerization process [3]. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
394
20 Aromatic Alkylation
3) 4)
5)
6)
7)
8)
9)
It is a convenient retrofit for an existing Fischer–Tropsch or crude oil refinery. It may also be included as benzene reduction strategy in new refineries. Synthetic jet fuel can be produced in a single step by combining aromatic alkylation and alkene oligomerization processes [4]. Additional fuel blending flexibility can be obtained by including aromatic alkylation with C4 and lighter material in a refinery design. For example, cumene has excellent motor-gasoline properties (RON = 113; MON = 99.3), it is a useful component in synthetic jet fuel, and it can even be included as a density improver in diesel fuel (ρ = 686.5 kg m−1 , cetane number = 15). In an HTFT refinery, a significant fraction of the naphtha is light boiling as a result of the ASF distribution of the Fischer–Tropsch syncrude. Although the pentenes have other convenient refining pathways, the vapor pressure of the motor-gasoline may still be high due to the abundance of pentenes. By employing some of the pentenes for aromatic alkylation, the vapor pressure of the motor-gasoline can be reduced, while producing compounds that have use as low vapor pressure solvents or distillate fuel additives. The inherently low density of Fischer–Tropsch distillates can be increased to meet diesel fuel specifications by blending alkyl aromatics [5]. There is a trade-off involved between cetane number and density, but cetane number increases with increasing chain length of the alkene used for alkylation. Linear alkyl benzenes (LABs) are normally produced to manufacture alkyl benzene sulfonate detergents. Although conventional LAB production is not seen as an economically viable pathway to address fuel quality, such technology may nevertheless be considered in the absence of viable alternatives. Co-refining of aromatics from crude oil or coal liquids by alkylation can benefit from the alkene-rich nature of Fischer–Tropsch syncrude. Coal liquids produced by low-temperature coal gasification (Section 3.4.2) to generate syngas can benefit significantly from aromatic alkylation. This is a difficult application but, with judicious catalyst selection, the benzene, phenol, and alkene content in coal liquids can be reduced before hydrotreating. By doing so, refinery benzene is reduced, less n-alkenes are hydrogenated to low octane number n-alkanes, and the phenol is converted into compounds useful as oxidation inhibitors. Aromatic alkylation also reduces hydrogen use and improves the quality of the final product [6]. Some alkyl aromatics can double as both fuels and chemicals. An aromatic alkylation unit can introduce fuel–chemical production flexibility into a refinery design to increase refinery profitability. Since the alkyl aromatic can be employed as a fuel component also, there is a natural reintegration pathway in the refinery. The application of aromatic alkylation as chemical technology in a fuel refinery can increase transportation fuel blending robustness. In the case of motor-gasoline blending, the alkyl aromatics can provide an additional high octane number blending material. When such high octane number alkyl aromatics are blended into the motor-gasoline, it may be possible to reduce the severity of the catalytic naphtha reformer (Section 22.3). The catalytic naphtha reformer is one of the most expensive units to operate in a refinery. By reducing reformer severity, the aromatic content of the motor-gasoline is lowered and a considerable operating cost saving is realized. It may also be possible to adjust motor-gasoline that is below specification on octane number and/or density. An additional benefit from having aromatic alkylation as chemical technology in a fuel refinery is that it can provide fallback blending
20.2 Reaction Chemistry
material to adjust the density and/or aromatic content of off-specification jet fuel and diesel fuel. 10) Alkanes can be employed as alkylating agents with aromatics by combining a metal functionality with the acid functionality of the aromatic alkylation catalyst [7]. When applied to light alkanes, it is possible to increase the overall liquid yield from the refinery. In this application, the alkylation reaction continuously removes alkenes by dehydrogenation over the metal and thereby drives the dehydrogenation equilibrium. 11) Alcohols can be used as alkylating agents with catalysts tolerant of alcohols and water. This is an alternative to alcohol dehydration (Section 17.2) for converting light alcohols from the Fischer–Tropsch aqueous product. This application has the added benefit of combining an endothermic and an exothermic processing step to make overall heat management much easier. Water is also an effective reaction moderator and diluent to potentially improve selectivity. 12) Aromatic alkylation can be combined with various other liquid-phase acid-catalyzed processes that employ alkene-rich feed materials in order to reduce the buildup of heavy material on the catalyst surface. The aromatics are good solvents and aromatic co-feeding can be of benefit to catalysts in fouling service.
20.2 Reaction Chemistry
In organic chemistry terms, aromatic alkylation is the electrophilic substitution of an aromatic molecule and it is often referred to as Friedel–Crafts alkylation (Figure 20.1). A detailed description of the rich chemistry associated with aromatic alkylation by hydrocarbons can be found in reviews [8]. The reaction is catalyzed by both Brønsted and Lewis acids. Alkylation is promoted by activating groups (electron-donating groups) attached to the aromatic. Typical activating groups are alkyl (−R), ether (−OR), and alcohol (−OH) groups. One of the most important consequences of this is that aromatic alkylation by alkenes activates the aromatic and promotes further alkylation. Multiple alkylation of the aromatic can be undesirable, but it is favored by the reaction chemistry. The side reactions that take place during aromatic alkylation are mostly acid-catalyzed conversions of the feed and products (Figure 20.2). Alkenes are reactive feed materials, and side reactions such as double bond isomerization, skeletal isomerization, oligomerization, and cracking can be expected. Depending on the nature of the alkene, some of these reaction types may be irrelevant or unlikely to occur. For example, propene is unaffected by double bond and skeletal +
+ H+ − H+ +
− H+
H
+ H+
Figure 20.1 Aromatic alkylation reaction over an acidic catalyst as illustrated by the alkylation of benzene with propene to produce isopropylbenzene (cumene).
395
396
20 Aromatic Alkylation Dimerization +C5 −C5
Double bond isomerization
C10 alkene
Cracking Skeletal isomerization Aromatic alkylation
+
−
Dealkylation (cracking)
+ H2
Hydrogen transfer
+C5
Isomerization
Aromatic alkylation
C5 alkane
C5 alkyl C5 alkyl
Figure 20.2 Reaction network of acid-catalyzed reactions that can take place during aromatic alkylation as illustrated by the alkylation of benzene with pentene.
isomerization and it is resistant to cracking, but propene is prone to oligomerization. The alkyl aromatic product can also be subject to acid-catalyzed side reactions such as isomerization, transalkylation, and dealkylation. The alkylation of aromatics with alkenes is very exothermic. The calculated heat of reaction at standard conditions is −95 to −105 kJ·mol−1 . This is comparable to the heat of reaction of alkene oligomerization. At typical aromatic alkylation conditions, equilibrium favors aromatic alkylation, and dealkylation becomes important only around 300 ◦ C. The composition of the different alkyl aromatic isomers in relation to each other is also governed by equilibrium. This refers to the degree of alkylation (e.g., mono-, di-, and tri-alkylated) and positional isomers (e.g., ortho, meta, and para substitution). As with other acid-catalyzed reactions, there is some control over the equilibrium through the feed composition, operating conditions, and catalyst.
20.3 Catalysis
The carbon chain length of the feed determines the type of catalyst that is best suited for aromatic alkylation. Subsequent discussion of the catalysis and process technology will be classified in terms of the alkene used for alkylation. The distinction based on chain length is related to a number of factors: 1) Reactivity of the alkene. The acid-catalyzed mechanism of aromatic alkylation (Figure 20.1) requires protonation of the alkene as the first step. With ethene as feed, this will give rise to a primary carbocation, which is not readily formed and requires strong acidity. It was found that only strong acid sites are active for benzene alkylation with ethene and that weaker acid sites are completely inactive for this reaction [9]. Ethene alkylation also benefits from a mechanistic route that forms a carbocation intermediate that is not a primary carbocation. Propene and longer chain alkenes can form secondary or tertiary carbocations that are more
20.3 Catalysis
2)
3)
4)
5)
stable and that are more readily formed. Less acidic catalysts can be employed for aromatic alkylation by propene and longer chain alkenes. Cracking propensity of the alkene. Generally speaking, C2 –C6 alkenes are resistant to acid-catalyzed cracking [10]. This is a distinction we previously encountered in isomerization (Chapter 18), where the cracking propensity of the carbocation intermediate determined what catalyst types could be employed. When C7 and longer chain alkenes are employed for alkylation, as is found in LAB manufacture, catalyst selection is more limited [11]. Cracking propensity of the alkyl aromatic. Although the C2 –C6 alkenes can be converted over very acidic catalysts without much cracking, once alkylated, the carbon chain length effectively grows by one carbon number. The aromatic nucleus is quite resistant to cracking, but the alkyl group is now polarized by the aromatic nucleus. Although the alkyl group is no longer an alkene, one has to take the cracking susceptibility of the alkyl aromatic into account during catalyst selection. Pore size restrictions based on molecular size. The molecular diameter of the alkylated aromatic product is affected by the size of the alkylating alkene. In catalysts with a small pore diameter, it may be possible for the reagents to enter, but not for the more bulky products to either form or leave, thereby effectively disqualifying such catalysts from use. Pore size can also play a role in limiting coke formation in the same way. The pore-constraining geometry of zeolites can be exploited to benefit aromatic alkylation. When the pore size of a zeolite is selected appropriately in relation to the alkene, it is possible to favor specific isomers or limit multiple alkylation [12, 13]. Exploiting pore size for selectivity becomes more difficult as the alkene chain length increases. Phase behavior. The chain length of the alkene in conjunction with the operating temperature and pressure determines whether the reaction takes place in the gas or liquid phase. Some catalysts may benefit from high-temperature gas-phase operation. For example, H-ZSM-5 performs well for gas-phase ethylbenzene synthesis, because the catalyst has a low coking tendency and at high temperature the heavier products are readily cracked to improve selectivity and catalyst lifetime.
20.3.1 Aromatic Alkylation with Ethene
When aromatic alkylation is catalyzed by the traditional AlCl3 –HCl Friedel–Crafts catalyst, the reaction mechanism involves intermediates that do not require the protonated ethene to form an unstable primary carbocation (Figure 20.3). This explains the industrial success of the traditional liquid-phase AlCl3 –HCl Friedel–Crafts catalyst, which is still widely used in industry [14, 15].
d+
H
H
H
H
H Cl AlCl3 d+ d−
Figure 20.3
H + −
H H Cl
H
H H
H
H H
H
H
H
H
AlCl3
H Cl
AlCl3
H
H H Cl d− AlCl3
H Cl AlCl3 d+ d−
Benzene alkylation with ethene over a traditional AlCl3 –HCl Friedel–Crafts alkylation catalyst.
397
398
20 Aromatic Alkylation Table 20.1
Catalysts industrially employed for benzene alkylation with ethene to produce ethylbenzene.
Description
AlCl3 –HCl
H-ZSM-5
H-Y
Operating conditions Temperature (◦ C) Phase Benzene:ethene ratio Transalkylation required Catalyst life (yr) Technology suppliers
160 Liquid 2.5 No –b MonsantoLummus
390–440 240–270 Gas Liquid 7.6 7.2 Yes Yes 1 1 ExxonMobil- LummusBadger Unocal-UOPc
MCM-22
Betaa
<300 Liquid 4 Yes 3 ExxonMobilRaytheon
<300 Liquid 4–6 Yes 2 LummusUOP
a Catalyst
EBZ-500, presumably an H-Beta-based catalyst. Continuous addition of catalyst possible to compensate for deactivation. c An H-Y catalytic distillation process is also available with better catalyst life. b
The main drawback of this process is the corrosiveness of the catalyst, which requires the conversion to be carried out in enameled or glass lined reactors [16]. The catalyst is very active for the reaction, which can be conducted at a much lower temperature and with much lower aromatic:alkene ratio than other aromatic alkylation processes for ethene (Table 20.1) [16]. Close to equilibrium conversion is achieved. Continuous addition and removal of the catalyst maintains constant activity, and catalyst lifetime does not constrain operation. The catalyst is also active for transalkylation. It is therefore possible to recycle multiply alkylated products back into the alkylation reactor for disproportionation (Equation 20.1.). Diethylbenzene + benzene 2 ethylbenzene
(20.1)
In the petrochemical industry, mono-alkylation is very important, since the main product desired from the aromatic alkylation with ethene is ethylbenzene. In a Fischer–Tropsch refining application, the selectivity to mono-alkylated products will also be important for chemical production, but it is of lesser importance for fuel production and multiple alkylation is even desirable in some instances. Vapor-phase alkylation of aromatics with ethene over SPA and amorphous silica-alumina (ASA) were investigated, but were not able to compete with AlCl3 –HCl [17]; this is even before the development of zeolite-catalyzed processes. There is a slight once-through selectivity advantage for mono-alkylation obtained over SPA, but SPA has much higher butylbenzene selectivity. SPA does not catalyze transalkylation, and multiply alkylated products cannot be recycled for disproportionation. A number of zeolite catalysts have proven to be effective for ethylation of benzene (Table 20.1). The most important of these are H-Y (FAU), H-ZSM-5 (MFI), MCM-22 (MWW), and H-Beta (BEA). In all cases, a transalkylation reactor is required to increase the yield of mono-alkylated products by disproportionation. Of these catalysts, MCM-22 is the most selective for mono-alkylation [16]. All aromatic alkylation catalysts used with ethene are sensitive to impurities in the feed material. Water reacts vigorously with AlCl3 –HCl to destroy the catalyst and foul the reactor. Oxygenates likewise cause catalyst deactivation [18]. Water and oxygenates must therefore rigorously be
20.3 Catalysis
excluded from AlCl3 –HCl-catalyzed alkylation. The zeolite catalysts are less sensitive to water, and zeolite catalysts can be used for alkylation of aromatics with ethanol, which produces water as a by-product (Equation 20.2). Benzene + ethanol (C2 H5 OH) → ethylbenzene + H2 O
(20.2)
Strong oxygenate adsorption by carbonyl compounds, as well as acid-catalyzed side reactions of the carbonyl compounds, will cause zeolite catalyst deactivation. 20.3.2 Aromatic Alkylation with Propene
Up until 1992, almost all cumene was produced using AlCl3 –HCl- or SPA-catalyzed alkylation processes (Table 20.2) [15, 16, 19]. Most of the advantages and drawbacks of AlCl3 –HCl are the same as for ethylbenzene production (Section 20.3.1). Compared to SPA, the biggest advantage of AlCl3 –HCl for chemical production is the low oligomerization selectivity. Near-equilibrium conversion is obtained for benzene and toluene alkylation with propene to produce cumene and cymene, respectively. Alkylation of toluene can be conducted at even lower temperatures (60–80 ◦ C) than benzene alkylation. SPA is an effective catalyst for benzene and toluene alkylation with propene. Of all the industrially applied catalysts for aromatic alkylation with propene, SPA has the lowest selectivity for multiple alkylation. A large fraction of the industrial production of cumene is performed using SPA [19]. Catalysts industrially employed for liquid-phase benzene alkylation with propene to produce cumene.
Table 20.2
Description
AlCl3 –HCl
SPA
Mordenitea
MCM-22
Beta
Operating conditions 110 195–225 170 <180 <180 Temperature (◦ C) 2–4 4 Benzene:propene ratio 2.4 5 –b Transalkylation required No Noc Yes Yes Yes 1 – >2e 2–4e Catalyst life (yr) –d Cumene yield (%)f Benzene to cumene 98.8 97.8g >99 >99 >99 98.6 98.7 99.9 Propene to cumene 98.5 96.2g Technology suppliers Monsanto-Lummus UOP Dow-Kellogg Badger (ExxonMobil) Enichem, UOP a Highly
dealuminated H-MOR, Si:Al around 70 : 1. Not reported. c SPA has a low selectivity for multiple alkylation; transalkylation is optional. d Continuous addition of catalyst possible to compensate for deactivation. e Cycle length around two years; the catalyst is ex situ regenerable. f With recycle operation and/or transalkylation. g Once-through: benzene to cumene yield is 94.6% and propene to cumene yield is 91.2%. b
399
400
20 Aromatic Alkylation
Catalysis by SPA requires control of the catalyst hydration level [20], and the catalysis principles are the same as discussed before (Section 19.3.1) [21]. Processes based on zeolite-catalyzed aromatic alkylation with propene have been commercialized for dealuminated mordenite (MOR), MCM-22, and H-Beta (Table 20.2). The choice of operating conditions affects the relative performance of MCM-22 and H-Beta. Reports can be found favoring H-Beta over MCM-22 [16, 22], as well as reports favoring MCM-22 over H-Beta [15, 23]. Zeolite catalysts require transalkylation to improve mono-alkylation selectivity, and a separate transalkylation reactor is required in order to disproportionate the 5–15% multiply alkylated products. The selection of a catalyst for aromatic alkylation with propene is not clear-cut. The cost of zeolite-based alkylation processes is within 5% similar to the cost of SPA-catalyzed alkylation [19]. Cost is not a differentiating feature, except with respect to AlCl3 –HCl, which is more capital intensive. Despite the industrial success of SPA for aromatic alkylation with propene, the selection of a zeolite catalyst instead of SPA is justified on the basis of the following perceived limitations [15, 16, 22]: 1) The biggest drawback of SPA for chemical production is its high oligomerization selectivity – SPA is a very good catalyst for alkene oligomerization (Chapter 19). It is not a drawback for fuel production and, for applications in a Fischer–Tropsch refinery, the higher oligomerization selectivity is not an issue. 2) Aromatic alkylation over SPA requires a higher aromatic:alkene ratio for selective alkylation. This is relevant for chemical production constrained by alkene availability, but not for fuel production, or in Fischer–Tropsch refinery context. The need to operate at higher aromatic:alkene ratios is necessary only to limit multiple alkylation and oligomerization; it is not necessary to protect the SPA catalyst. An industrial test run in a Fischer–Tropsch refinery demonstrated that SPA performed well with an aromatic:alkene ratio as low as 0.07 [3]. 3) The SPA catalyst is not regenerable and must be disposed of, increasing the environmental impact compared to zeolites. This is in fact a misconception. A lifecycle analysis of SPA compared to that of zeolites indicates that SPA has a smaller environmental impact than zeolites. SPA is manufactured from kieselguhr, which is a naturally occurring form of silica. The silica and alumina do not have to be dissolved for synthesis as in the case of zeolites, nor do they require templating molecules. Furthermore, the spent SPA is neutralized with ammonia to produce fertilizer for agricultural use [24], and it is not a waste product at all. 4) Some phosphoric acid is lost that can cause downstream corrosion, which is not a problem with zeolite catalysts. This is an unavoidable drawback, which becomes worse when applied with oxygenate-containing syncrude. However, when zeolites are applied with syncrude, it requires rigorous exclusion of carbonyl compounds to avoid similar problems due to carboxylic acid formation. In a Fischer–Tropsch refinery, SPA has more advantages than drawbacks compared to zeolites. In fact, the benefit that can be derived from combined oligomerization and aromatic alkylation (Section 20.1) lends strong support to SPA as the preferred catalyst for aromatic alkylation by propene. There is also the issue of catalyst deactivation. With proper feed preparation, SPA
20.3 Catalysis Influence of typical impurities from coal liquids on the conversion of cresols over H-ZSM-5 at 350 ◦ C and 6 MPa.
Table 20.3
Impurity evaluated Pyridine Benzonitrilea Aniline Indole Thiophenol Benzothiophene
Feed concentration (µg·g –1 ) 200 3700 200 200 850 850
Extent of poisoning
Recovered activity with clean feed
Strong Strong Weak effect Very weak effect No effect Very weak effect
Partial Complete Higherb Higherb No effect Slightly higherb
a Decomposed b After
to yield NH3 , and poisoning was likely by NH3 rather than benzonitrile itself. introduction of clean feed, there was a period that activity was better than the baseline with clean feed only.
has a shorter lifetime than zeolites, but this presupposes proper feed preparation. The carbonyl compounds in syncrude can cause zeolite deactivation and lead to downstream corrosion. The coprocessing of material from synthesis gas generation may also introduce low levels of heteroatom impurities, which are far more detrimental to zeolites than SPA. An evaluation of the suitability of zeolite-catalyzed alkylation must take cognizance of these nontraditional feed impurities that are encountered in a Fischer–Tropsch refinery. The impact of some typical impurities from coal liquids obtained by low-temperature coal gasification on the activity of H-ZSM-5 illustrates the point (Table 20.3) [25]. Poisoning effects typically increase with decreasing operating temperature, because there is less driving force for desorption. The impact of impurities at typical aromatic alkylation conditions (<200 ◦ C) will be considerably more severe than that observed at 350 ◦ C. 20.3.3 Aromatic Alkylation with C4 and Heavier Alkenes
A wide variety of catalysts have been reported for the alkylation of aromatics with different C4 –C6 alkenes, but not many industrial applications are apparent. One of the few important applications is the alkylation of phenol and p-cresol with isobutene to produce oxidation inhibitors. This reaction is mostly performed using tert-butanol as alkylating agent, but alkylation with isobutene has been reported with catalysts such as acidic resins [26], sulfated zirconia [27], and tris-(2-tert-butylphenolato) aluminium [28]. The latter is also industrially applied. In a Fischer–Tropsch refining context, the use of aromatic alkylation was investigated using various C4 –C6 alkenes, employing various solid acid catalysts, but mostly SPA [6, 29–32]. None of these applications was industrially applied. In applications with a low aromatic to alkene ratio, as is relevant to refinery applications of aromatic alkylation, the alkylation to oligomerization selectivity is strongly affected by branching. Over SPA, branching was a prerequisite for oligomerization and, even in an alkene-rich matrix, linear alkenes were preferentially alkylated. However, once alkenes were skeletally isomerized, oligomerization selectivity became much higher than alkylation selectivity, even in equimolar
401
402
20 Aromatic Alkylation
mixtures [32]. This change in selectivity is partly due to the mechanism whereby aromatic alkylation catalysis takes place. Aromatic alkylation catalysis takes place by an Eley–Rideal mechanism (reaction occurs between an adsorbed alkene and aromatic in the bulk phase), except in pore-constrained catalysts [12, 23]. Once skeletal isomerization enables oligomerization, the alkenes and aromatics in the bulk phase have to compete for addition to the polarized intermediate. Although aromatics are stronger nucleophiles than alkenes, over SPA the alkenes are more effective. When aromatic alkylation is performed over SPA with a broad range of alkenes, the nature of the naphtha becomes very important. The short-chain alkenes will be the main alkylating alkenes. Aromatic conversion and alkyl aromatic selectivity will depend on the degree of branching of the naphtha range alkenes present in the feed (Table 20.4) [33]. When the naphtha range alkenes are mainly linear, which is typical of Fischer–Tropsch syncrude, the naphtha range alkenes will act as a diluent, with some alkylation of the heavier alkenes taking place. When the naphtha is highly branched, it is active for oligomerization and little alkylation by heavier alkenes takes place. The branched material is also more susceptible to cracking (depolymerization) [34], which is a source of tert-butyl aromatics in the product.
Table 20.4 Alkylation of benzene with a mixed Fischer–Tropsch C3 –C4 alkene feed to which naphtha range material was added.
Description
Naphtha range feed C7 –C11 n-alkanes 1-Hexene 2,3-Dimethylbutene 1-Octene 2,4,4-Trimethylpentene
Conversion (%) Propene Butene Benzene Naphtha:distillate ratio Mono:di-alkylation ratio
95.82 86.8 76.73 82 : 18 4.6
99 92 66 84 : 16 5.0
98 95 42 74 : 26 14.3
96 87 73 87 : 13 5.1
100 95 46 72 : 28 3.8
Product selectivity (%) Olefin oligomers Cumene sec-Butyl benzene tert-Butyl benzene Diisopropylbenzenes Hexyl benzenes C7 -alkylbenzenesa C8 -alkylbenzenesa C9 -alkylbenzenesa
88.2 8.3 1.4 0 1.5 0 0.6 0 0
88.8 7.2 1.5 0.3 0 0.4 0 1.4 0.4
93.8 3.4 1.1 1.3 0 0 0 0.2 0.2
89.0 7.8 1.4 0.0 1.3 0 0.5 <0.1 0
91.1 2.0 1.5 3.5 0.6 0 0.8 0.4 0.1
Alkylation is performed over solid phosphoric acid in a fixed bed reactor at 160–190 ◦ C, 3.8 MPa, and liquid hourly space velocity (LHSV) of 1 h−1 . The mixed alkene feed consisted of 15% propene, 7% butenes, 20% propane/butane, 4% benzene, and 54% naphtha (as indicated). a Mainly di-alkylated benzenes.
20.4 Syncrude Process Technology Catalysts employed for liquid-phase benzene alkylation with heavy alkenes (C10 –C14 ) to produce linear alkyl benzenes.
Table 20.5
Description Operating conditions Temperature (◦ C) Benzene:alkene ratio Product quality Overall alkyl linearity (%) 2-Phenylalkanes (%) Tetralins (%) a
AlCl3 –HCl
HF
F – /SiO2 –Al2 O3
60–75 8
40–60 8
– –
98 –a <1.0
92–94 15–18 <0.3
94–95 >25 <0.5
a a
Not reported.
Aromatic alkylation with C7 and heavier alkenes is mainly performed to produce LABs. LABs are the primary raw materials for the production of LAB sulfonate detergents, which is a large-volume commodity chemical. There are two significant challenges associated with the acid-catalyzed alkylation of aromatics with long-chain alkenes. The first is to suppress isomerization. For detergent applications, the linearity and position of the phenyl in the alkyl benzene are both important to the final product quality. This is also true for applications involving LAB as diesel fuel additives. The second is to avoid cracking. It stands to reason that it will be easy to suppress cracking if isomerization can be suppressed, since β-scission is slow when only secondary carbocation intermediates are involved (Section 21.2.2). However, once alkylation of the aromatic takes place, the carbon attached to the aromatic ring becomes a tertiary carbon. Suppression of isomerization and cracking go hand in hand, but cracking is not entirely dependent on isomerization. The industrial application of aromatic alkylation by long-chain alkenes is performed almost exclusively with liquid hydrofluoric acid (HF) [11, 16, 35]. One solid acid–catalyzed technology (Detal) employing a fluorided silica–alumina catalyst has been demonstrated on an industrial scale [11, 16]. The solid catalyst requires periodic mild rejuvenation, and over a test period of 14 months the catalyst showed little deactivation. Typical performance data is provided in Table 20.5 [11, 16]. The application of HF or fluorided silica–alumina with Fischer–Tropsch-derived feed materials will require rigorous oxygenate and water removal.
20.4 Syncrude Process Technology
It is important to realize that most aromatic alkylation technologies have been developed for petrochemical production. One also has to keep in mind that these technologies were designed for facilities employing crude oil as feedstock and where the alkene supply is limited. Commercially available aromatic alkylation technologies are therefore optimized to increase the
403
404
20 Aromatic Alkylation
yield of specific mono-alkylated aromatic products (the desired chemical product) and to increase alkene selectivity to alkylation products (the alkene is the reagent in short supply). When considering aromatic alkylation technology for Fischer–Tropsch refining, high alkene consumption is less critical and it is even desirable in some instances. Depending on the refining objective, a process design optimized for chemical production may actually be suboptimal in a Fischer–Tropsch refinery geared toward fuel production. It is important to realize that within a Fischer–Tropsch refining context aromatic alkylation is not only about the production of alkyl aromatics, but it is also about the conversion of gaseous alkenes into liquid products, the avoidance of alkene conversion into alkanes, and the conversion of refinery benzene (and phenol from associated coal liquids). In all cases, it is best to adapt the technology to suit the refining requirements. Some specific aspects that need to be considered when applying aromatic alkylation technology with syncrude are the following: 1) The contaminants that are present in Fischer–Tropsch syncrude are different from those found in a crude oil refinery. It is likely that the light alkenes (C2 –C3 ) that are used for alkylation are quite pure, but water and volatile oxygenates may still be present. The amount of impurities will increase with carbon number, and heavier alkenes will contain oxygenate impurities. Feed preparation should be matched to the requirements of the catalyst system. Whenever possible, it is advisable to select a catalyst that is not sensitive to trace amounts of water or oxygenates. 2) Zeolite and other solid acid catalysts that have strong acid sites can be poisoned or inhibited by oxygenates and other trace impurities. This is particularly relevant when coprocessing material such as coal liquids that is not Fischer–Tropsch-derived. The temperature of desorption of heteroatom-containing compounds that are strongly interacting with the catalyst surface is related to the acid site strength. When the operating temperature of the catalyst is low, typically in order to limit acid-catalyzed side reactions, the operating temperature may be insufficient to desorb strongly adsorbed compounds. Such strongly adsorbed compounds may not normally be perceived as catalyst poisons, but acid sites will be titrated in relation to their strength at a given temperature. (The same principle is applied in temperature-programmed desorption analysis of catalysts to ascertain the acid strength distribution.) Over time, even trace amounts of strongly adsorbing material will affect catalytic activity. 3) The suppression of skeletal isomerization is necessary in alkylation technology where the linearity of the alkylated aromatic is important, for example, LABs. It is also beneficial for suppressing oligomerization and cracking. Although this requirement is not specific to syncrude refining, it remains an important consideration when dealing with C4 and heavier alkenes that can be skeletally isomerized. When dealing with syncrude in which alkene availability is higher, this requirement becomes more crucial, because one may want to operate at a lower aromatic to alkene ratio. 4) When aromatic alkylation is selected as a technology to convert ethene into liquid products, the aromatic compounds may be the limiting reagents. It is likely to be the case in HTFT refining, where the syncrude contains around 5% ethene (Table 1.2) on a mass basis. On a molar basis, ethene is much more abundant. Multiple alkylation of the aromatic compound is desirable in this application. No transalkylation reactor will be needed in such a design.
References
5) Aromatic alkylation in a Fischer–Tropsch refinery can operate at much lower aromatic to alkene ratios and without a transalkylation reactor. Operating at high aromatic to alkene ratio and with a transalkylation reactor should be considered only if the oligomerization product has poor properties, or if the mono-alkylated aromatic is the desired product. 6) Designing an aromatic alkylation unit with the flexibility to accommodate lower aromatic to alkene ratios can come in handy to improve refinery robustness. When it is necessary to work away refinery feeds with a higher than usual benzene content, or to provide material to increase the aromatic content of diesel fuel, or to increase the density of fuels in general, the refinery will benefit from operating the aromatic alkylation unit at lower aromatic to alkene ratio. Conversely, the refinery can benefit from operating the aromatic alkylation unit at higher aromatic to alkene ratio if the opposite effect is desired. This type of flexibility may be possible in a Fischer–Tropsch refinery, but is more difficult to realize in a crude oil refinery. 7) Aromatic alkylation to reduce refinery benzene levels benefit from a low aromatic to alkene ratio, because the probability of benzene alkylation is increased. This is a fuels application, and it is not necessary to purify the benzene beforehand, nor is oligomerization as side reaction an issue. What is of importance is the nature of the alkenes and the possible contaminants for the reasons mentioned before. 8) It is generally beneficial to perform aromatic alkylation with a reactive alkene-containing quench (Figure 19.2); this improves alkylation selectivity and heat management. 9) For the production of synthetic jet fuel by combined alkylation and oligomerization, it is necessary to control the aromatic to alkene ratio in such a way that the kerosene fraction contains between 8 and 25% aromatics. This is the specification requirement for synthetic jet fuel (Section 14.2.1).
References 1. Goezler, A.R., Hernandez-Robinson, A., Ram,
S., Chin, A.A., Harandi, M.N., and Smith, C.M. (1993) Refiners have several options for reducing gasoline benzene. Oil Gas J., 91 (13), 63–69. 2. De Klerk, A. and Nel, R.J.J. (2008) Benzene reduction in a fuels refinery: an unconventional approach. Energy Fuels, 22, 1449–1455. 3. Sakuneka, T.M., Nel, R.J.J., and De Klerk, A. (2008) Benzene reduction by alkylation in a solid phosphoric acid catalyzed olefin oligomerization process. Ind. Eng. Chem. Res., 47, 7178–7183. 4. Sakuneka, T.M., De Klerk, A., Nel, R.J.J., and Pienaar, A.D. (2008) Synthetic jet fuel production by combined propene oligomerization and aromatic alkylation over solid phosphoric acid. Ind. Eng. Chem. Res., 47, 1828–1834.
5. De Klerk, A. (2009) Can Fischer-Tropsch syn-
6.
7.
8.
9.
crude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. De Klerk, A. and Nel, R.J.J. (2007) Phenol alkylation with 1-octene on solid acid catalysts. Ind. Eng. Chem. Res., 46, 7066–7072. Sealy, S. and Traa, Y. (2005) Direct alkylation of toluene with ethane on bifunctional zeolite catalysts. Appl. Catal. A, 294, 273–278. Patinkin, S.H. and Friedman, B.S. (1964) in Friedel-Crafts and Related Reactions, Alkylation and Related Reactions, Part 1, Vol. II (ed. G.A. Olah), John Wiley & Sons, Inc., New York, pp. 1–288. Sun, X., Wang, Q., Xu, L., and Liu, S. (2004) The role of acid strength of zeolites in liquid-phase alkylation of benzene with ethylene. Catal. Lett., 94, 75–79.
405
406
20 Aromatic Alkylation 10. Buchanan, J.S., Santiesteban, J.G., and Haag,
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
W.O. (1996) Mechanistic considerations in acid-catalyzed cracking of olefins. J. Catal., 158, 279–287. Kocal, J.A., Vora, B.V., and Imai, T. (2001) Production of linear alkylbenzenes. Appl. Catal. A, 221, 295–301. Smirniotis, P.G. and Ruckenstein, E. (1995) Alkylation of benzene or toluene with MeOH or C2 H4 over ZSM-5 or β zeolite: effect of the zeolite pore openings and of the hydrocarbons involved on the mechanism of alkylation. Ind. Eng. Chem. Res., 34, 1517–1528. Raj, K.J.A., Malar, E.J.P., and Vijayaraghavan, V.R. (2006) Shape-selective reactions with AEL and AFI type molecular sieves alkylation of benzene, toluene and ethylbenzene with ethanol, 2-propanol, methanol and t-butanol. J. Mol. Catal. A, 243, 99–105. (1993) Styrene. SRI Process Economics Program Report 33C, Stanford Research Institute, Menlo Park, CA. Degnan, T.F. Jr., Smith, C.M., and Venkat, C.R. (2001) Alkylation of aromatics with ethylene and propylene: recent developments in commercial processes. Appl. Catal. A, 221, 283–294. Perego, C. and Ingallina, P. (2002) Recent advances in the industrial alkylation of aromatics: new catalysts and new processes. Catal. Today, 73, 3–22. Landau, R., Lidov, R.E., Habeshaw, J., and Roper, A.N. (1975) in Benzene and its Industrial Derivatives (ed. E.G. Hancock), Ernest Benn, London, pp. 285–343. Donaldson, J.W. (1969) in Ethylene and its Industrial Derivatives (ed. S.A. Miller), Ernest Benn, London, pp. 876–941. Cesar, M.A. and Chou, G. (1999) Cumene, SRI Process Economics Program Report 219, Stanford Research Institute, Menlo Park, CA. Cavani, F., Girotti, G., and Terzoni, G. (1993) Effect of water in the performance of the ‘‘solid phosphoric acid’’ catalyst for alkylation of benzene to cumene and for oligomerisation of propene. Appl. Catal. A, 97, 177–196. De Klerk, A. (2011) in Catalysis, Vol. 23 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge, pp. 1–49. Schmidt, R.J. (2005) Industrial catalytic processes – phenol production. Appl. Catal. A, 280, 89–103. Corma, A., Mart´ınez-Soria, V., and Schnoeveld, E. (2000) Alkylation of benzene with short-chain
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
olefins over MCM-22 zeolite: catalytic behaviour and kinetic mechanism. J. Catal., 192, 163–173. Van der Merwe, W. (2010) Conversion of spent solid phosphoric acid catalyst to environmentally friendly fertilizer. Environ. Sci. Technol., 44, 1806–1812. B¨ohringer, W., Dwyer, C.L., and Fletcher, J.C.Q. (2006) Acid catalysed conversion of phenolics – influence of impurities from coal gasification. Proceedings of the South African Chemical Engineering Congress, September 20–22, 2006, Durban, South Africa, OralPDO03. Santacesaria, E., Silvani, R., Wilkinson, P., and Carr`a, S. (1988) Alkylation of p-cresol with isobutene catalyzed by cation-exchange resins: a kinetic study. Ind. Eng. Chem. Res., 27, 541–548. Yadav, G.D. and Thorat, T.S. (1996) Kinetics of alkylation of p-cresol with isobutylene catalysed by sulfated zirconia. Ind. Eng. Chem. Res., 35, 721–731. K¨upper, F.-W. (2004) A new mechanism – key for an improved synthesis of 2,6-di-tertbutylphenol. Appl. Catal. A, 264, 253–262. Cowley, M., De Klerk, A., and Nel, R.J.J. (2005) Amylation of toluene by solid acid catalysis. Ind. Eng. Chem. Res., 44, 5535–5541. Pienaar, A.D., Nel, R.J.J., and De Klerk, A. (2006) Upgrading Fischer–Tropsch products by alkylation of benzene over solid phosphoric acid. Proceedings of the 16th Saudi Arabia-Japan Joint Symposium, 5–6 November 2006, Dhahran, Saudi Arabia, pp. 57–66. Cowley, M., De Klerk, A., Nel, R.J.J., and Rademan, J.D. (2006) Alkylation of benzene with 1-pentene over solid phosphoric acid. Ind. Eng. Chem. Res., 45, 7399–7408. Nel, R.J.J. and De Klerk, A. (2007) Selectivity differences of hexene isomers in the alkylation of benzene over solid phosphoric acid. Ind. Eng. Chem. Res., 46, 2902–2906. Sakuneka, T.M., De Klerk, A., and Nel, R.J.J. (2009) Unpublished results (approved for release by Sasol on 28 Jan 2009). Ipatieff, V.N. and Pines, H. (1936) Alkylation accompanying depolymerization. J. Am. Chem. Soc., 58, 1056. Jones, K. (1975) in Benzene and its Industrial Derivatives (ed. E.G. Hancock), Ernest Benn, London, pp. 446–472.
407
21 Cracking 21.1 Introduction
Cracking is the dominant type of conversion for heavy-end (residue) upgrading. It involves the decomposition of molecules to yield products that are of a lower molecular mass than the feed. Cracking technology is widely used in crude oil refineries for the production of transportation fuels and in petrochemical refineries for the production of ethene and propene. Thermal cracking is also the earliest refining technology. It was accidentally discovered in 1861 in a small New Jersey refinery, after a batch distillation was left unattended and the temperature increased to cause the oil to break down into lighter products [1]. Many of the subsequent thermal refining technologies, such as visbreaking (Section 16.6.2), thermal cracking (Section 16.6.3), and coking (Section 16.6.4) that were developed, employed the same basic cracking principle. The temperature is increased sufficiently to cause homolytic bond dissociation. Such thermal technologies formed the basis of a typical second-generation thermal refinery (Figure 2.9). Catalytic cracking is arguably the most important petroleum refining process in the history of the industry [2]. The efficiency of thermal cracking was greatly improved by the use of acid catalysts. This enabled much higher residue conversion and achieved better quality naphtha range products. Although catalytic cracking and thermal cracking both cause molecular decomposition, the mechanism of these two cracking types are different and, as a consequence, the products. Acid catalysts employed for residue cracking were rapidly deactivated. In catalytic cracking technology, the ability to restore catalytic activity after cracking became as important as the cracking process itself. The original Houdry fixed bed catalytic cracking process employed multiple catalytic converters in parallel to accommodate the 10 min operating cycles of reaction–purging–regeneration [3]. The fixed bed process was followed shortly thereafter by the development of moving catalyst bed technology and a transported catalyst technology known as fluid catalytic cracking (FCC). The latter became the dominant catalytic cracking technology in refining and is widely used in crude oil refineries. Catalytic cracking has three key objectives:
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
408
21 Cracking
1) The conversion of residue into lighter and more valuable products. This ‘‘bottom of the barrel’’ upgrading function is shared with all cracking technologies employed for fuels refining and is an inherent attribute of cracking. 2) The production of light alkenes specifically. Catalytic cracking is the primary source of alkenes in most crude oil refineries. Unlike Fischer–Tropsch syncrude, there are little or no alkenes in most crude oils (Section 2.2.1). Since many important crude oil refining technologies require alkenes (Chapter 16), cracking became indispensable in crude oil refining. 3) Enrichment of the hydrogen content of the lighter products. Catalytic cracking is a carbon rejection technology. In this way, the H:C ratio of the products is improved. The coke that is formed on the catalyst also captures some of the metals, which are thereby removed from the product. The strength of catalytic cracking is also a cause for weakness. The light products are hydrogen-enriched and the alkenes are valuable to further refining, but the products still contain heteroatoms, such as sulfur. The sulfur content in the final fuel products must be very low (typically 10–15 µgÐg1 or less) and FCC is not an effective technology for heteroatom removal. This gave rise to the development of another class of cracking technology, namely, hydrocracking. The three types of cracking are compared in Table 21.1 [3–5]. Hydrocracking is a specific type of acid-catalyzed cracking process that is conducted in parallel with hydrogenation. Instead of carbon rejection, as in the case of catalytic cracking, hydrocracking uses hydrogen addition. The technology was originally developed in the 1950s to convert residues into motor-gasoline [6]. In effect, hydrocracking is a form of process intensification that accomplishes more than one upgrading step by employing a bifunctional catalyst. Bifunctional hydrocracking catalysts have Table 21.1 Comparison of ‘‘mixed phase’’ thermal cracking (Dubbs-type unit), fluid catalytic cracking, and conventional hydrocracking of crude oil to produce products for further refining to transportation fuels.
Description Operation Temperature (Ž C) Pressure (MPa) Space time H2 use (% of feed mass)a Per pass yield (mass%)a Gas Naphtha Distillate Residue Coke
Thermal cracking
Catalytic cracking
Hydrocracking
450–540 1.5–4.0 >15 min 0
480–550 0.1–0.3 <10 s 0
350–430 10–20 >30 min ¾2
14 27 2 57 0
22 49 16 8 5
4 13 44 41 0
a These are typical values taken from the literature for residual oil conversion. The actual numbers are dependent on the operating conditions and nature of the feed.
21.1 Introduction
both acid and metal functionalities. Yet, describing it in such simple terms fails to capture all of the refining objectives that can be met by hydrocracking. The clich´e that the whole is more than the sum of its parts is true of hydrocracking: 1) Hydrocracking is a cracking process. Cracking implies the breaking of C–C bonds to decrease the average molecular mass and boiling point of products relative to that of the feed. It is primarily a heavy-end upgrading technology, because it capably transforms atmospheric residue into lighter products with lower boiling points. Therefore, one of the objectives of hydrocracking is to increase the transportation fuel yield in a refinery. 2) Hydrocracking is a hydrotreating process (Section 16.2). Hydrogen addition increases the hydrogen to carbon ratio of the products relative to that of the feed. Alkenes are hydrogenated (HYD) to alkanes and, although aromatics are more difficult to hydrogenate, hydrocracking catalysts are usually capable of hydrodearomatization (HDA) also. By doing so, the stability and cetane number of the distillate range products are increased. In the hydrocracking process, heteroatoms are removed, and depending on the feed material, hydrodeoxygenation (HDO), hydrodesulfurization (HDS), and hydrodenitrogenation (HDN) all take place. Another objective of hydrocracking is therefore to improve product quality. 3) Hydrocracking is a hydroisomerization process (Section 18.4). The degree of branching of the products is increased relative to that of the feed. This improves the cold-flow properties of the products. 4) Hydrocracking is an alkane activation process. Alkanes are not very reactive and, by incorporating a metal functionality in the catalyst, it is possible to dehydrogenate alkanes to alkenes, thereby effectively activating the feed. Alkenes are reactive species that can readily be converted to lighter products at less severe operating conditions than alkanes. Hydrocracking therefore lowers the temperature required for catalytic cracking. 5) Hydrocracking is a carbon retaining process. Hydrocracking catalysts have a much lower rate of deactivation than catalytic cracking catalysts. Coke precursors are hydrogenated to reduce coke formation, thereby extending the catalyst lifetime from seconds (as in FCC) to years. In crude oil refining, there are a number of factors that will influence the decision to make use of hydrocracking. Foremost is the need for a high yield of good quality distillate [4]. Hydrocracking is the only heavy-end upgrading technology that delivers good quality distillate as a main product. Fuel refineries that target distillates as the main product are likely to prefer hydrocracking over other heavy-end upgrading technologies. Hydrocracking is also the heavy-end cracking technology that is best for the conversion of aromatic-rich heavy petroleum fractions into lighter products [5]. It is difficult to achieve good conversion of aromatic-rich feed materials into lighter products by catalytic cracking only, which is better suited for the conversion of more paraffinic fractions. Hydrocracking is also more effective for complete heteroatom removal, since the combination of hydrogenation and cracking allows ring opening of refractory heteroatom-containing compounds. There are some important drawbacks to hydrocracking. It requires the availability of hydrogen, which may be a constraining factor in some crude oil refineries. It does not produce alkenes and is therefore of limited value for motor-gasoline production.
409
410
21 Cracking
21.2 Reaction Chemistry 21.2.1 Thermal Cracking
Thermal cracking relies on high temperature to provide sufficient energy for the rupture of molecular bonds. These cracking reactions occur by homolytic bond dissociation, and the temperature required to break the bonds depends on the bond strength as expressed by the homolytic bond dissociation energies (BDEs) (Table 21.2) [7].
Homolytic bond dissociation energies (BDEs) of different hydrocarbon bond types at 25 Ž C.
Table 21.2
Homolytic bond dissociation reactiona
Bond type
BDE (kJ·mol−1 )
Alkanes CH4 ! CH3 ž C Hž C2 H6 ! CH3 CH2 ž C Hž C3 H8 ! (CH3 )2 CHž C Hž C4 H10 ! (CH3 )3 Cž C Hž C2 H6 ! CH3 ž C ž CH3 C3 H8 ! CH3 CH2 ž C ž CH3 C4 H10 ! (CH3 )2 CHž C ž CH3 C4 H10 ! CH3 CH2ž C ž CH2 CH3 C5 H12 ! (CH3 )3 Cž C ž CH3 C5 H12 ! (CH3 )2 CHž C ž CH2 CH3 C6 H14 ! (CH3 )2 CHž C ž CH(CH3 )2 C6 H14 ! (CH3 )3 Cž C ž CH2 CH3 C7 H16 ! (CH3 )3 Cž C ž CH(CH3 )2 C8 H18 ! (CH3 )3 Cž C ž C(CH3 )3
C–H C–H C–H C–H C–C C–C C–C C–C C–C C–C C–C C–C C–C C–C
439 423 413 404 377 372 371 368 366 364 358 358 346 329
C–H C–H C–C C–C C–C C–C C–C
463 372 424 418 415 320 315
C–H C–H C–C C–C
472 375 433 325
Alkenes CH2 DCH2 ! CH2 DCHž C Hž CH2 DCHCH3 ! CH2 D CHCHž2 C Hž CH2 DCHCH3 ! CH2 DCHž C ž CH3 CH2 DCHC2 H5 ! CH2 DCHž C ž CH2 CH3 CH2 DCHC3 H7 ! CH2 DCHž C ž CH(CH3 )2 CH2 DCHC2 H5 ! CH2 DCHCH2 ž C ž CH3 CH2 DCHC3 H7 ! CH2 DCHCH2 ž C ž CH2 CH3 Aromatics C6 H6 ! C6 H5 ž C Hž C6 H5 CH3 ! C6 H5 CH2 ž C Hž C6 H5 CH3 ! C6 H5 ž C ž CH3 C6 H5 CH2 CH3 ! C6 H5 CH2 ž C ž CH3 a Note:
When a radical specie has been indicated as ‘‘CHž ,’’ the radical is on the C not the H.
21.2 Reaction Chemistry
(a) R CH2 CH2 R′ CH2 CH CH2 (b) R CH2 +
R′
R CH2 + CH2 CH (c) R CH2 +
∆
CH3 R′
R CH2 CH2 CH
CH2 R′
R CH2 +
R CH2 CH2 R′
∆
CH2 CH CH2
R R′
CH3 +
CH2 CH CH2 +
R′
R′
CH2 R′
R CH2 CH2 CH R′
CH2 R′
R CH2 CH2 CH
R CH2 + CH2 CH
R′
R′
R′
R CH2 CH2 R′
Figure 21.1 Thermal cracking mechanism showing some representative (a) initiation, (b) propagation, and (c) termination reactions.
Homolytic bond dissociation initiates free radical decomposition. Each dissociation step produces two free radicals (Figure 21.1a). Not all of the cracking has to take place through the primary initiation step though. Cracking reactions can also propagate (Figure 21.1b). Propagation can continue and, while the free radical reactions continue to propagate, primary initiation continues. Importantly, propagation can proceed by self-decomposition. During self-decomposition the C–C bond on the β-position relative to the radical is most likely to break. Such radical-induced β-scission was used to successfully explain the products from thermal cracking [8]. This mechanistic description has been formalized and forms the basis of the Rice–Kossiakoff description of thermal cracking [9]. Ultimately, thermal cracking reactions terminate by the recombination of free radicals (Figure 21.1c). Thermal cracking is quite different to catalytic cracking in its reaction rate over time behavior. Under isothermal conditions, the rate of thermal cracking rapidly increases during the propagation period and the maximum thermal cracking rate is reached only some time after thermal cracking was initiated. Conversely, the maximum cracking rate during catalytic cracking occurs when the reagent concentration is the highest, usually at or near the time the feed comes into contact with the catalyst. Some features of the reaction chemistry involved in thermal cracking that are of importance to note are the following: 1) There is no catalyst involved in thermal cracking. Compared to catalytic cracking at the same temperature, the rate of thermal cracking is usually slower at temperatures below 550–600 Ž C. At higher temperatures, thermal cracking becomes more effective than catalytic cracking, because catalyst accessibility and mass transport do not limit the thermal cracking rate. This has important consequences for catalytic cracking at high temperature [10]. 2) Because there is no catalyst involved, metal and other contaminants cannot deactivate thermal cracking. This is quite beneficial, and it has been shown that metal contaminants present in low-temperature Fischer–Tropsch (LTFT) wax could effectively be removed during thermal cracking. The metals were deposited as a separate phase [11].
411
412
21 Cracking
3) Initiation of thermal cracking in a hydrocarbon feed material that contains 1-alkenes occurs at a lower temperature. This is mainly due to the lower BDE of the C–C bond that would yield an allyl radical (ž CH2 –CHDCH2 ), as can be seen from Table 21.2. The allyl radical is ž resonance stabilized (ž CH2 –CHDCH2 $ CH2 DCH–CH2 ) and is readily produced during thermal cracking. 4) The thermal cracking of alkanes produces 1-alkenes during self-dissociation after initiation (Figure 21.1b). The mechanism of thermal cracking inherently produces n-1-alkenes from n-alkanes, and historically thermal cracking was the primary source of n-1-alkenes in the industry. 5) In Fischer–Tropsch syncrude, there are also oxygenates present. These oxygenates have quite similar BDEs as the hydrocarbons (Table 21.3) [12], albeit somewhat lower on average for linear species. Some oxygenate species, such as the peroxides, have very low BDEs. 6) The bond strength is affected by radical formation [7, 13]. This is well illustrated by the following two examples. The BDE to form an ethyl radical (423 kJÐmol1 from Table 21.2) is much more than the subsequent dissociation to yield ethene (Equation 21.1), and the BDE of propanone to form acetyl and methyl radicals (340 kJÐmol1 from Table 21.3) is considerably more than the subsequent dissociation of the acetyl radical to liberate CO (Equation 21.2) [13]. ž
CH2 CH3 ! CH2 DCH2 C H , BDE25Ž C D 149 kJÐmol1
(21.1)
(CO)CH3 ! CO C CH3 , BDE25Ž C D 40 kJÐmol1
(21.2)
ž
ž
ž
7) Radical rearrangement may also take place during thermal cracking (Figure 16.5c). Such radical rearrangement reactions are by 1,4- and 1,5-hydrogen transfer and do not result in skeletal rearrangement [14]. 8) At lower temperatures and higher pressures, radical addition may become significant (Section 19.3.6). This is also called thermal oligomerization or retrograde condensation. It effectively sets the lower temperature limit for thermal cracking and thermal cracking must therefore be conducted at temperatures above 450 Ž C, unless the thermal cracking process employs a strategy to limit free radical addition reactions. 9) Thermal cracking is capable of producing molecular hydrogen (H2 ) as the product of propagation (Equation 21.3) and termination (Equation 21.4). ž
ž
RH C H ! R C H2 ž
2H ! H2
(21.3) (21.4)
10) It is possible to produce diradical species [15]. Such diradicals have two active centers for propagation, but can also terminate by intramolecular radical recombination. Sophisticated thermal cracking models that can be found in the literature take both primary thermal cracking reactions and secondary cracking and recombination reactions into account [14, 16]. All these models are based on the Rice–Kossiakoff description of thermal cracking [9]. According to this description, a pool of small alkyl radicals will be produced by homolytic bond dissociation. The alkyl radicals will eventually reach a steady-state concentration when radical formation and radical termination are in equilibrium. These alkyl radicals will remove hydrogen atoms from the other molecules, producing a large radical that will rapidly decompose by C–C
21.2 Reaction Chemistry Homolytic bond dissociation energies (BDEs) of different oxygenate species at 25 Ž C.
Table 21.3
Homolytic bond dissociation reaction
Bond type
BDE (kJ·mol−1 )
Alcohols CH3 OH ! ž CH2 OH C Hž CH3 CH2 OH ! CH3 (ž CH)OH C Hž CH3 OH ! CH3 Ož C Hž CH3 CH2 OH ! CH3 CH2 Ož C Hž CH3 OH ! CH3 ž C HOž CH3 CH2 OH ! CH3 CHž2 C HOž
C–H C–H O–H O–H O–C O–C
393 389 437 436 386 383
C–H C–C O–C O–C
389 361 349 344
C–H C–H C–H C–C C–C C–C
364 360 411 345 340 361
O–H O–O O–O O–O
365 213 157 194
O–H O–C
443 346
O–H O–C
362 267
Ethers CH3 OCH3 ! ž CH2 OCH3 C Hž CH3 CH2 OCH3 ! ž CH2 OCH3 C ž CH3 CH3 OCH3 ! CH3 Ož C ž CH3 CH3 CH2 OCH2 CH3 ! CH3 CH2 Ož C ž CH2 CH3 Aldehydes and ketones CH2 O ! ž CHO C Hž CH3 CHO ! ž (CO)CH3 C Hž CH3 (CO)CH3 ! ž CH2 (CO)CH3 C Hž CH3 CHO ! C ž CHO C ž CH3 CH3 (CO)CH3 ! ž (CO)CH3 C ž CH3 CH3 CH2 (CO)CH3 ! ž CH2 (CO)CH3 C ž CH3 Peroxides H2 O2 ! HOOž C Hž H2 O2 ! HOž C ž OH CH3 OOCH3 ! CH3 Ož C ž OCH3 (CH3 )3 CCH2 OOH ! (CH3 )3 CCH2 Ož C ž OH Carboxylic acids and esters CH3 COOH ! CH3 COOž C Hž CH3 COOCH2 CH3 ! CH3 COOž C ž CH2 CH3 Phenol and phenyl ethers C6 H5 OH ! C6 H5 Ož C Hž C6 H5 OCH3 ! C6 H5 Ož C ž CH3
bond rupture at the β-position from the hydrogen-deficient carbon (Figure 21.1b). The product distribution from thermal cracking is mainly determined by this self-dissociative behavior of the radicals, which, with each successive self-cracking event, produces another smaller radical and a 1-alkene. Deviations from the Rice–Kossiakoff description occur at high temperature when secondary reactions occur and equilibration takes place. At temperatures above 650 Ž C, thermal cracking reactions reach an equilibrium after about 5 s [17]. Under such conditions, the product distribution is determined by Gibbs free energy minimization. As the temperature and residence time are
413
414
21 Cracking
increased, secondary reactions then become more prominent, which include reactions typical of thermal reforming, such as the condensation and aromatization of ethene (Equations 21.5 and 21.6) [18]. 2CH2 DCH2 ! CH2 DCH–CHDCH2 C H2
(21.5)
CH2 DCH2 C CH2 DCH–CHDCH2 ! C6 H6 (benzene) C 2H2
(21.6)
Under mild conditions that one would typically consider for the thermal cracking of Fischer–Tropsch waxes, secondary reactions are minor and the Rice–Kossiakoff description is valid. Thermal cracking of wax then follows first-order kinetics [19]. At pressures below 1 MPa, a reaction order of 1.3 reportedly describes the kinetics better [20]. The thermal cracking rate constant k(n) increases with the carbon chain length n of the molecule. The chain-length-dependent kinetics is described by the empirical correlation of Voge and Good (Equation 21.7) [21]: k(n) D (n 1)Ð(1.57 ð n 3.9) ð 105
(21.7)
The heavier wax molecules will consequently have a higher rate of cracking than the shorter chain hydrocarbons being produced. It implies that thermal cracking can be employed to remove the heaviest wax fraction from a mixture and that, at intermediate wax conversion, a bimodal product distribution will develop. This has been experimentally observed. Although the original chain length dependence of thermal cracking kinetics (Equation 21.7) was developed for C4 –C16 alkanes, it could successfully be extended to model thermal cracking of Fischer–Tropsch waxes in the range C20 –C120 [11]. The carbon number distribution of the products can be adequately described by combining the Voge and Good kinetics with the Rice–Kossiakoff description of cracking. 21.2.2 Catalytic Cracking
Catalytic cracking is an acid-catalyzed reaction. Cracking proceeds differently for alkanes and alkenes. Alkanes are difficult to activate, and it requires very strong acidity to directly protonate the alkane to produce a pentacoordinated carbocation intermediate (Figure 21.2). This pentacoordinated carbocation (carbonium ion) intermediate cracks by α-scission, which is also called the Haag–Dessau mechanism of alkane protolysis [22]. The products from α-scission are an alkane H H
+ H+
R
R′
+
−H
R
+
R′
H Pentacoordinated carbocation
HH R
+
R′
H a-Scission
Figure 21.2 Catalytic cracking by protolysis of an alkane. This involves direct protonation by an acid catalyst to form a pentacoordinated carbocation that subsequently cracks by α-scission to yield an alkane and an alkene.
− H+
R
+ H3C
R′
21.2 Reaction Chemistry
and a classic tricoordinated carbocation (carbenium ion) intermediate, which on deprotonation yields an alkene. There are two important consequences of cracking by protolysis. The first is that it is possible to produce C1 –C2 hydrocarbons that are difficult to form by the usual acid-catalyzed cracking of alkenes. The second is that one of the products from protolysis is an alkene. Further reaction will favor acid-catalyzed cracking of the alkene, which is much easier to convert than acid-catalyzed protolysis of the alkane. For example, over H-ZSM-5 at 510 Ž C, acid-catalyzed cracking of alkanes is three orders of magnitude slower than the cracking of alkenes [23]. As a consequence, the alkanes will not be substantially converted, whereas the alkenes will be readily cracked to light products that are mainly C6 and lighter alkenes. The contribution of protolysis to overall cracking is determined by the availability of other carbocation-creating pathways, and it is especially important during initial conversion before a carbocation-covered catalyst surface is created [24]. The description of catalytic cracking thus far considered only the monomolecular processes, but cracking can also proceed via a bimolecular process once some carbocation intermediates have been formed. The main difference between the monomolecular and bimolecular processes is that the monomolecular process requires proton transfer from the catalyst to the alkene, whereas the bimolecular process requires hydride transfer from the alkane to another carbocation intermediate. This mechanism of alkane activation is central to aliphatic alkylation catalysis (Figure 16.7). The activation energy for hydride transfer is less than that required to protonate an alkane in order to create a pentacoordinated carbocation [25]. For hydride transfer to occur, there must be sufficient space to allow the bimolecular reaction to proceed. The ratio of direct alkane protonation by the Haag–Dessau mechanism and hydride transfer, strongly correlates to the pore dimensions of the acid catalyst. The relative contributions are about equal at a pore diameter just over 0.7 nm, whereas the contribution of direct protonation is an order of magnitude more as the pore diameter decreases to just under 0.6 nm [26]. Alkenes are readily converted over acid catalysts. The alkene CDC double bond is easily protonated to produce a carbocation intermediate. The carbocation intermediate is capable of double bond isomerization (Figure 16.5a), skeletal isomerization (Figure 18.1), and addition reactions such as oligomerization (Figure 19.1a) and aromatic alkylation (Figure 20.1). The reaction network associated with acid-catalyzed cracking is quite complex. High temperature and low pressure thermodynamically favor cracking over addition reactions. It will also be shown that skeletal isomerization is a key supporting reaction that is necessary to accelerate catalytic cracking. The rate of cracking is determined by the relative stabilities of the carbocation intermediates involved in the conversion (Figure 21.3). The most stable carbocation pair is two tertiary carbocations, and type A is the most facile mode of cracking. The cracking rate decreases as the stability of the intermediates become less. On the basis of cracking over H-ZSM-5, the relative rates of types B to E cracking were established: type B (120) > type C (40) × type D (2) > type E (1) [23]. These numbers are only indicative of the relative rates, but clearly illustrate that branching is effectively a prerequisite for cracking to occur at an appreciable rate. On very acidic catalysts, the less favorable cracking modes are more accessible, whereas less acidic catalysts will be restricted to the more favorable cracking modes.
415
416
21 Cracking
Type A (tert to tert )
Type B1 (sec to tert )
Type B2 (tert to sec)
Type C (sec to sec)
Type D (sec to primary)
Type E (tert to primary) Figure 21.3
R
+
R′
R
R
+
R′
R
R
+
R′
R
R
+
R′
R
R
+
+
+
+
R′
+
+
R′
+
+
R′
R
+
+
R′
R′
+
+
R
R′
R′
+
+
R
R′
Different modes of catalytic cracking by β-scission of a carbocation intermediate.
These observations explain why it is difficult to catalytically crack C6 and lighter alkenes. It also explains why C1 –C2 products are normally not appreciably formed during catalytic cracking, except under severe conditions. Hydrogen transfer is another important process that takes place during catalytic cracking. It is in this way that acid catalysts are capable of producing aromatics at high temperature. During hydrogen transfer, one molecule of hydrogen (H2 ) is effectively transferred between two molecules (Equation 21.8) and can ultimately lead to aromatization (Equation 21.9). 2 alkenes ! alkane C diene
(21.8)
5 alkenes ! 4 alkanes C aromatic
(21.9)
One molecule is dehydrogenated while the other molecule is hydrogenated, and the process of hydrogen transfer must not be confused with hydride transfer, which is a propagation step. Hydrogen transfer does not involve C–C bond breaking, but it affects the product selectivity by changing the relative ratio of alkanes, alkenes, and aromatics. Acid catalysts in general have poor hydrogen desorbing capability and, when aromatics are formed, hydrogen transfer normally plays a key role. Corma and coworkers pointed out that, in instances where molecular hydrogen (H2 ) was reported during acid-catalyzed cracking, thermal cracking could not be ruled out [24]. 21.2.3 Hydrocracking
The main difference between hydrocracking and catalytic cracking is that a hydrocracking catalyst contains metal-catalyzed dehydrogenation and hydrogenation activity in addition to acid-catalyzed cracking activity (Figure 21.4). The situation is analogous to that of skeletal isomerization and hydroisomerization (Chapter 18). The ability to dehydrogenate alkanes to
21.2 Reaction Chemistry − H2
R′
R
+ H2
+ H+
R′
R
Metal site
− H+
+
R
Acid site +
R
R′
R
− H2 + H2
R′
R
R′
R′
+ H+ − H+
R
R'
+
Unlikely
R
R′
+
R
R′
+
b-Scission − H2
R
R′
+ H2 R
+ H+
R′ − H+
R
+
R′ Likely R
+
R′
Figure 21.4 Hydrocracking mechanism that involves dehydrogenation over a metal site to produce an alkene, before standard acid-catalyzed cracking occurs. The alkene products are again hydrogenated over a metal site to produce alkanes.
alkenes makes the alkenes available for acid-catalyzed cracking, and milder acidity is necessary than is required for direct alkane cracking by protolysis. The alkene availability is also limited by the dehydrogenation–hydrogenation equilibrium, which reduces the probability of bimolecular side reactions of alkenes. Since the actual cracking steps are the same as for acid-catalyzed cracking of alkenes, one anticipates similar behavior. The dependence on branching structure and the stability of the carbocation intermediates (Figure 21.3) are also observed during hydrocracking. The molecules must first be isomerized to a dibranched structure before hydrocracking occurs. This was beautifully illustrated by the work of Steijns and Froment [27]. They proposed different reaction models for the hydroisomerization and hydrocracking of n-alkanes. When these different models were subjected to discrimination based on experimental data, it was found that all the other models could either be rejected or simplified to a single model description (Figure 21.5). The model description disallowed cracking of n-alkanes and monobranched alkanes, requiring all hydrocracking to proceed through dibranched isomers. The metal functionality introduces an added dimension to the reaction chemistry. It enables metal-catalyzed side reactions and adds a few complications to the mechanistic description: 1) Some metals, such as Ni [28–31] and Pt [28, 29, 32, 33], are prone to hydrogenolysis, which is an analogous reaction to protolysis, but it is the result of C–C bond cleavage by hydrogen. However, there is a significant difference between the hydrogenolysis mechanism and selectivities obtained over Ni and Pt [28]. Hydrogenolysis over Ni takes place by α-scission of the terminally adsorbed carbon to produce methane (CH4 ) as side product. Hydrogenolysis of Pt takes place by a mechanism analogous to β-scission and the methane selectivity is low.
417
418
21 Cracking
R
R′
n -Alkane
R
R′ Monobranched alkane
R
R′
Dibranched alkane
Cracking products Figure 21.5 Dominant model description for hydrocracking of n-alkanes. Discrimination using experimental data caused all alternative model descriptions to be rejected or simplified to this model.
2)
Metals, such as Pt, can also catalyze skeletal isomerization, which is believed to occur mainly through a cyclic intermediate [32–34]. This also implies that the metal is able to effect ring closure and hydrogenolysis of the ring to cause ring opening. 3) Dehydrogenation of cycloalkanes to aromatics and the direct dehydrocyclization of alkanes to aromatics are metal-catalyzed reactions. Pt is a good catalyst for the direct dehydrocyclization of alkanes to cycloalkanes [35, 36] and aromatics [37]. Aromatization is discussed in more depth in the next chapter (Chapter 22). 4) The dynamic dehydrogenation–hydrogenation of the hydrocarbons has an important consequence for the identity of the alkenes. During acid catalysis (no metal sites), a carbocation on the surface may desorb by deprotonation and the position of the CDC in the desorbed product will indicate where the carbocation had been. The position of the carbocation, although it no longer exists, is thus retained by the position of the CDC in the desorbed product. When the alkene is readsorbed, the carbocation can be re-created on the same carbon or at most one carbon distant from the original carbocation. There is limited mobility of the carbocation between successive adsorption and desorption cycles on acid sites of the catalyst. This is not the case when a metal function is present on the catalyst. The position of the CDC in the desorbed product will be lost after hydrogenation. Subsequent dehydrogenation of the alkane to form an alkene again is unlikely to occur in the same position. In fact, dehydrogenation will more likely take place on a linear part of the molecule than at a position of branching. The position of the CDC may not be in close proximity to existing branches. On short-chain alkanes, the impact of this process may be limited, but on long-chain Fischer–Tropsch waxes it enables the virtual CDC position to ‘‘move’’ over distances, which is unrealistic by double bond isomerization. It also implies that the carbocation formed after dehydrogenation is probably secondary rather than tertiary. This follows from the classical description of hydrocracking. 5) During hydrocracking, the dominant compound class is the alkanes. Alkene availability is not dependent on the extent of cracking, but on the dehydrogenation–hydrogenation equilibrium. As in the case of pure acid-catalyzed cracking, hydride transfer is responsible for the direct activation of branched alkanes. Hydride transfer from a branched alkane to a carbocation on the catalyst surface will produce a tertiary carbocation from the branched alkane. The intermediate thus formed is indistinguishable from the carbocation formed after skeletal isomerization, but it is likely to be different from the carbocation formed after dehydrogenation.
21.3 Thermal Cracking
6) Hydrogenation of branched alkenes is more difficult than n-alkenes, because the double bond is sterically more hindered. Alkenes produced by cracking and that have a CDC on a tertiary carbon will therefore be longer lived than the n-alkenes produced by either cracking or dehydrogenation. 7) The different pathways by which a carbocation is formed determine how likely it is that a second branch will form in the proximity of the first branch. If successive skeletal isomerization steps occur without desorption, or if the carbocation is formed on the tertiary carbon by hydride transfer from the branched alkane and is followed by isomerization, the cracking probability is increased. Branching in close proximity enables type B cracking (Figure 21.3). However, if the classical description of hydrocracking holds true and a single skeletal isomerization step is followed by either cracking or desorption of the alkene, hydroisomerization may introduce branching that is quite distant from each other on long-chain alkanes, thereby having a low probability of cracking. This principle was demonstrated by the conversion of different methyl branched alkanes [38]. Hydroisomerization of 2-methyloctane to dibranched products produced an isomer distribution that was in good agreement with the statistical distribution expected from isomerization of a mixture of secondary carbocations. A comparatively low concentration of multibranched material is typically observed in a hydrocracking product from Fischer–Tropsch waxes over selective wax hydrocracking catalysts [39, 40]. It has been found that, during the hydrocracking of Fischer–Tropsch wax over a Pt/SiO2 –Al2 O3 catalyst, the amount of multibranched alkanes exceeded the amount of monobranched alkanes only in the heavier product fraction, that is, C18 –C19 and heavier, irrespective of the conversion [40]. This is somewhat surprising if the classical description of hydrocracking holds true and every isomerization step is followed by either cracking or hydrogenation. The first isomerization step is unlikely to lead to cracking (Figure 21.4). If isomerization is followed by hydrogenation, the steric hindrance caused by branching will decrease the probability that subsequent dehydrogenation will regenerate the same alkene. There is no guarantee that dehydrogenation will form a CDC bond in the vicinity of the branching. It should therefore be possible to produce highly isomerized products from Fischer–Tropsch waxes before type B or C cracking becomes possible. This is not what is generally observed. In order to better explain the experimental observations, it is proposed that the transition state of the intermediate is different (Figure 21.6). Instead of the protonated cyclopropane (PCP) intermediate leading to isomerization by forming a noncyclic carbocation, the PCP may lead to isomerization by forming another PCP intermediate. When this occurs, the second PCP intermediate is in a position to effect isomerization that is perfectly positioned for cracking by β-scission. This would explain the comparative lack of dibranched isomers often observed during hydrocracking, as well as the low branched to linear product ratio in the lighter fractions.
21.3 Thermal Cracking
The thermal cracking of alkanes is a straightforward method of converting the alkanes into alkenes without the risk of skeletal isomerization. In an LTFT refinery, where there is an abundance of n-alkanes, thermal cracking provides a convenient pathway for n-1-alkene
419
21 Cracking
Step1 − H2
R′
R
+ H2
R′
R
Fast
+
R
+ H+ − H+
+
R
Fast
R′
R′
Step 2a R
+
− H+
+
R′
R′
R
+H
+
+ H2
R′
R
− H2
Slow
Sterically dehydrogenation is favoured further from branching
Step 2b +
R′
R
+
R′
R
R′
+
R′
+
R
Likely
H+
+
H+
R
R′
R
−
420
R
Figure 21.6 Hydrocracking mechanism to explain the low amount of dibranched products, as well as low branched to linear light product formation during Fischer–Tropsch wax hydrocracking. Step 1 is similar to classic hydrocracking, but recognizing that step 2a leads to a product that is less likely
R′
R
+
R′
to regenerate a CDC at branch, or close to branch, step 2b is proposed as alternative pathway. Step 2b enables isomerization from PCP to PCP intermediate, leading to a dibranched intermediate that is likely to crack by β-scission.
production. Depending on the severity of the thermal cracking, a mild decrease in carbon number distribution can be achieved, right through to deep cracking for ethene and propene production. Ethene oligomerization has replaced n-alkane cracking as the main source of n-1-alkenes in industry because ethene is more readily available, it is less energy intensive (discounting the energy requirements of ethene production), and the product contains fewer impurities than the product from wax cracking [41]. However, considering the purity of alkanes derived from syncrude and its availability in an LTFT refinery, there is no reason why thermal cracking should not be an efficient refining pathway to chemicals. Mild conditions are required for the thermal cracking of Fischer–Tropsch wax, and mild thermal cracking were historically used to refine some Fischer–Tropsch fractions (Sections 6.3.3 and 8.4.2) [42–47]. Thermal cracking is a less efficient pathway than catalyzed conversion to fuels [11]. Potential applications of thermal cracking in Fischer–Tropsch refining include the upgrading of alkanes in LTFT syncrude. These applications have been outlined before (Section 16.6.3). High-temperature Fischer–Tropsch (HTFT) syncrude contains little heavy material, and the heavy fraction is aromatic rich. In principle, one can employ HTFT material with severe thermal cracking to produce ethene and propene, but there is little feedstock advantage in doing so.
21.4 Catalytic Cracking
21.3.1 Syncrude Processing Technology
The absence of a catalyst makes thermal technology quite robust, and the properties of the feed determine how thermal cracking should be implemented. When straight-run Fischer–Tropsch waxes are used as feed material, the waxes will also contain low concentrations of alkenes, oxygenates (mainly alcohols), and metal contaminants in the form of metal carboxylates. The feed-related operating issues that need to be considered are the following: 1) Feed reactivity. Alkenes and oxygenates lower the temperature needed for the initiation of thermal cracking on account of their lower homolytic BDEs (Tables 21.2 and 21.3). This is an advantage of using Fischer–Tropsch-derived feed material. 2) Coking propensity. The hydrogen-rich nature of alkanes reduces the probability of thermal aromatization. The increase in H2 on dehydrogenation increases the probability of radical termination by H2 . The same is not true of oxygenate-rich HTFT syncrude, where the carbonyl compounds are capable of rapid thermal decomposition and recombination at comparatively low temperatures. High-temperature thermal cracking is governed by thermodynamic equilibrium and the probability of coking reactions can be calculated. At high temperatures, the metal surfaces of process equipment may contribute to catalytic dehydrogenation of compounds. This is not unique to thermal cracking of Fischer–Tropsch materials. 3) Non-carbon deposits. The metal contaminants present in the wax will form metal-rich deposits during thermal conversion of Fischer–Tropsch-derived materials [11, 48]. Unless the design makes allowance for dealing with such deposits, it may cause operational problems related to obstructions inside the process equipment. 4) Product slate. On account of the oxygenates present in syncrude, the products will, among others, include water gas shift compounds (H2 , CO, CO2 , and H2 O). Downstream units must be designed to process typical cracking products in addition to the water gas shift compounds. One may consider using the noncondensable gas as co-feed to Fischer–Tropsch synthesis.
21.4 Catalytic Cracking
A number of applications of FCC technology with syncrude have been noted (Section 16.6.1). Most of the interest was directed at the conversion of LTFT waxes [49–56]. In principle, FCC is even better suited for the conversion of HTFT residue, which better resembles crude oil residue, but it is such a small syncrude fraction that it is difficult to justify an FCC unit. Industrially, FCC has been applied with HTFT material, not for the conversion of residue, but for the conversion of low octane number HTFT naphtha (Section 9.5.5) [57]. The stated aim was to ‘‘. . . convert low-octane components into high-octane components, while also producing substantial quantities of ethylene and propylene . . .’’ [58]. This implied the conversion of naphtha range alkanes. There were no reports that this implementation was successful, and based on the fundamentals of catalytic cracking, it is likely that this technology failed to meet the stated design
421
422
21 Cracking
objectives. The reason for this deduction will become apparent as we consider catalytic cracking of syncrude in more depth. FCC is a carbon rejection technology, which is an awkward proposition for syncrude applications. Conversion of hydrogen-rich feed materials creates production issues not seen in the FCC of crude oil residue fractions. These issues must be addressed by the syncrude processing technology (Section 21.4.2): 1) A hydrogen balance indicates that the products from FCC of a hydrogen-rich syncrude will contain a higher concentration of light alkanes in the gaseous and naphtha fractions than is normally found during FCC of crude oil residue. This is a consequence of hydrogen transfer in an attempt to reject carbon, for which FCC technology is designed. The following example serves to illustrate the point. If a C36 n-alkane is cracked to a 1 : 1 mixture of C3 and C6 hydrocarbons, the alkane content of the product is 12 1/2% (Equation 21.10), but if it is cracked and one of the C6 alkenes is converted into an aromatic, the alkane content of the product increases to 50% (Equation 21.11). C36 H74 ! 4C6 H12 C 3C3 H6 C C3 H8
(21.10)
C36 H74 ! C6 H6 C 3C6 H12 C 4C3 H8
(21.11)
2) In general, the more hydrogen-rich the syncrude, the more difficult it will become to form coke on the cracking catalyst. This implies that the cracking catalyst will deactivate at a slower rate during normal FCC operation. At similar operating conditions as crude-oil-derived residue, a higher conversion will therefore be achieved with syncrude. For example, at similar conditions, catalytic cracking of gas oil and LTFT wax yielded conversions of 61.6 and 88.1%, respectively [53]. Superficially it may be seen as a benefit, but it creates problems if the objective of FCC operation is to generate intermediate products, such as propene. When the feed conversion increases, the production of intermediates passes through a point of maximum yield, and then decreases as a result of increasing consumption by secondary reactions. 3) The heat necessary for FCC operation is supplied by the combustion of the coke on the acid catalyst in the regenerator. If insufficient coke is deposited on the catalyst during operation, there is insufficient heat generated during regeneration. 4) There is a significant difference in the cracking rate naphtha range alkanes and alkenes. Qualitatively, the relative conversion rates and reactions were given by Buchanan: alkene isomerization × alkene cracking × alkane cracking > aromatization [59]. The relative reaction rates of alkenes and alkanes were also quantitatively reported for H-ZSM-5 (Table 21.4) [23]. Although this data is for a low-activity catalyst, it illustrates the effect of carbon number and compound class on relative conversion rates well. Catalytic cracking can therefore be applied to naphtha range alkenes, but converting naphtha range alkanes requires more severe conditions. Naphtha range alkanes can be converted by thermal cracking, but make poor catalytic cracking feed material. 5) The ease of cracking increases with the heaviness of the feed material. This can be seen from data in Table 21.4. Applying FCC with wax is therefore very different to FCC of naphtha. Naphtha range alkanes are more refractory and require very severe conditions to be converted. Hydrogen rejection to produce naphtha range alkanes effectively transforms those molecules into inert species.
21.4 Catalytic Cracking Cracking rate constants for naphtha range alkenes and alkanes over a low-activity H-ZSM-5 catalyst at 510 Ž C and 1.3 kPa hydrocarbon partial pressure. The catalyst and conditions were selected to avoid bimolecular reactions and diffusion effects.
Table 21.4
Carbon number
C5 C6 C7 C8
Rate constant (s−1 ) Alkene
Alkane
9.5 231 1823 5732
0.30 0.84 1.49 2.25
Rate ratio of alkene to alkane conversion
32 275 1220 2550
21.4.1 Catalysis
The FCC of Fischer–Tropsch waxes have been investigated over amorphous silica–alumina (ASA) and the zeolites Beta (BEA), H-Y (FAU), and H-ZSM-5 (MFI). Differences in the activity of various cracking catalysts are masked by the high reactivity of Fischer–Tropsch wax as feed material. High conversion was obtained even at very low catalyst to wax ratio (Table 21.5) [53]. Irrespective of the catalyst employed for FCC of wax, it was found that for each catalyst the conversion and product distribution was insensitive to the operating severity and the condition of the catalyst (i.e., catalyst age) [51, 53]. The coke make was also consistently low (<1.5%) for all catalysts tested. Despite the other operating insensitivities, the different cracking catalyst types influenced the nature of the products (Table 21.6) [53]. The catalyst property that affects the product distribution most is pore geometry. The catalysts in Table 21.6 have been arranged from the least pore size constrained (FAU 0.74 ð 0.74 nm) to the most pore size constrained (MFI 0.51 ð 0.55 to 0.53 ð 0.56 nm) [60]. The following specific Effect of catalyst to wax ratio on the conversion of Fe-LTFT medium wax in a micro-yield unit at 470 Ž C over different zeolite catalysts.
Table 21.5
Catalyst to wax ratio
3 0.75 0.375 0.187 0.09
Conversion of Fe-LTFT medium wax (mass%) Equilibrium H-Y
Steamed H-Beta
Steamed H-ZSM-5
– 77 š 6 80 š 4 66 š 6 –
92 90 š 1 87 š 3 72 š 12 –
92 š 2 86 š 3 87 š 2 84 66 š 1
423
424
21 Cracking Table 21.6 Effect of different catalyst types on the product yield and unhydrogenated gasoline quality at 83–84% wax conversion in a micro-yield unit at 470 Ž C.
Description Product yield (mass%) C2 and lighter Propene Propane n-Butenes Isobutene Butanes n-Pentenes Methylbutenes Pentanes C6 -221 Ž C >221 Ž C Coke Gasoline properties RON MON
H-Y
H-Beta
H-ZSM-5
0.6 7.4 0.8 7.4 5.8 3.7 4.0 7.7 3.6 41.7 17.0 0.3
0.6 8.9 0.9 8.3 9.4 3.6 4.3 9.2 2.2 35.8 16.8 0.2
1.5 17.5 2.7 15.1 12.3 3.6 4.1 9.8 2.0 15.3 16.2 0.1
85.2 76.2
84.4 74.6
84.4 76.0
observations can be made dependent on and independent of the catalyst type on the catalytic cracking of Fischer–Tropsch wax: 1) All catalysts seem to ‘‘nibble’’ bits off the long-chain wax molecules [49]. This is understandable, because it would be unlikely for a wax molecule to enter into the catalyst without encountering an acid site on its way in. This has two important implications. The first is that for long-chain alkanes the cracking rate will be chain length independent, as was indeed experimentally observed [49]. The second implication is that the cracking probability is not equal along the carbon chain. 2) The carbon number distribution of the product from H-Y with its wider pores is mainly C3 –C10 , whereas the product distribution from H-ZSM-5 with its narrower pores is predominantly C3 –C6 . It is less likely for a wax molecule on its way into the wider pore zeolite catalyst to immediately encounter an acid site, and the ‘‘nibbles’’ from the wider pore zeolite are larger. The differences in product carbon number distribution reflect the different probabilities of being protoned as the wax molecule enters the zeolite. A dramatic demonstration of this effect can be seen from polyethylene cracking over H-ZSM-5 at 400 Ž C, which yielded C3 –C4 hydrocarbons as main product [61]. 3) The catalytic cracking products at around 80% wax conversion (Table 21.6) are mostly in the C3 –C12 carbon number range. Little or no H2 and C1 –C2 hydrocarbons are formed, indicating that the contribution of thermal cracking and protolysis were limited. 4) The ratio of branched to linear hydrocarbons in the >220 Ž C boiling material increased with increasing conversion over H-ZSM-5, while it decreased over H-Y [50]. This is not due to skeletal isomerization of the wax, but to reactant shape selectivity. It is more difficult for
21.4 Catalytic Cracking
5)
6)
7)
8)
9)
10)
branched alkanes in the feed to enter the smaller pores of H-ZSM-5 than the n-alkanes. The n-alkanes were thus more readily converted. The wider pores of H-Y zeolite did not impose the same diffusion restriction on branched alkanes and there were no reactant shape selectivity effects. Since branched alkanes intrinsically have a higher cracking rate than n-alkanes, the branched material was converted at a higher rate over H-Y than the n-alkanes. The pore-constrained geometry of H-ZSM-5 is also responsible for transition-state selectivity (Section 19.3.2). This caused all but the lightest products produced by catalytic cracking to be less branched than that obtained by wider pore acid catalysts. The effect of transition-state selectivity is not apparent in the C4 –C5 fraction. The position of branching affected the cracking conversion rate of branched alkanes over H-ZSM-5, which was not the case with H-Y. In H-ZSM-5, the 2-methyl and 3-methyl branched alkanes were more readily converted than alkanes with methyl branching closer to the center [50]. This is also a reactant shape selectivity effect due to the more pore-constrained geometry of H-ZSM-5 compared to H-Y. The octane number of the naphtha fraction is generally low. The cracking of Fischer–Tropsch wax over H-ZSM-5, H-Beta, and H-Y catalysts mostly resulted in a RON in the range 84–88 and a MON in the range 74–76 [53]. The rate of formation of multinuclear aromatics to produce coke is lower on H-ZSM-5 than on H-Y. This is especially apparent when looking at initial selectivity. The limited formation of coke is a result of the transition-state selectivity imposed by the pore constraints in H-ZSM-5, which hinders repeated cyclization necessary for the formation of multinuclear aromatics [62]. Although the FCC of wax did not show a strong dependence on the acidity of the catalysts [53], acid strength as well as acid site density has an effect on cracking rate and coking rate [63]. These effects will be more apparent during the cracking of naphtha range alkanes that are more refractory. Hydrothermal dealumination of catalytic cracking catalysts by steaming takes place during reaction and regeneration [64]. Oxygenates will be converted by FCC to, among other products, water. The water in itself is not expected to have a significant impact on the rate of hydrothermal dealumination. However, it has been postulated that dealumination caused by oxygenate dehydration directly on the catalytic surface may be considerably higher than the rate of dealumination by steaming [65]. Since the water is produced as an adsorbed reaction intermediate on an acidic alumina site, desorption may occur, but it may also lead to bond formation and dealumination.
21.4.2 Syncrude Processing Technology
Processing hydrogen-rich Fischer–Tropsch syncrude in an FCC unit requires some modifications that deviate from standard practice with residue feed materials. Applying FCC technology with Fischer–Tropsch syncrude requires significant modifications of the technology and in different ways for different feed and operating conditions:
425
426
21 Cracking
1) Fischer–Tropsch waxes are readily converted, and high conversion can be achieved at low catalyst to wax ratio (Table 21.5). Since cracking is an endothermic reaction, there is a practical lower limit to the catalyst to wax ratio from a heat balance perspective and one may want to consider increasing the addition of inert material as heat carrier only. Decreasing the catalyst to wax ratio without substituting the catalyst with inert material to maintain a similar particle size distribution and solids to wax ratio will also lead to scale-up issues. Dudkovi´c and coworkers experimentally demonstrated that the solids to fluid ratio in the riser affects the axial velocity profile [66]. As the mass flux is decreased, the degree of back-mixing is increased, which implies that a fraction of the catalyst particles spend a much longer time in the riser. At a low catalyst to wax ratio, this may lead to an even lower effective catalyst to wax ratio and a broad distribution of coke content on the catalyst. This will affect both cracking performance and the performance of the regenerator. When hydrodynamics is taken into consideration, it recommended that wax FCC applications employ higher catalyst to wax ratios than suggested by the conversion requirements. 2) If the operation of the FCC unit is not adjusted to obtain sufficient coke laydown on the cracking catalyst, it will be necessary to provide an additional source of heat in the regenerator section. With a hydrogen-rich feed, it is unlikely that sufficient coke can in fact be formed on the catalyst. During the extensive test work conducted at Amoco [53], a 1% coke yield was seldom exceeded. 3) Introducing additional fuel in the regenerator may seem a simple modification of the technology, but it is not. Coke has very different combustion characteristics to that of liquid fuels, even a very heavy fuel. Achieving complete catalyst regeneration and avoiding methane formation due to thermal cracking of the liquid fuel in the regenerator are easier said than done. Introducing a solid carbonaceous fuel is less convenient, but in the long run it may lead to more stable regenerator operation. 4) High-temperature FCC runs the risk of an increased contribution from thermal cracking. Corma and coworkers have shown that an increase in thermal cracking during FCC drastically lowers the light alkene yield [67]. This is very detrimental to petrochemical applications of FCC with Fischer–Tropsch syncrude. It is also detrimental to fuels applications, since the naphtha range product will have an even lower octane number despite the increase in aromatics content. 5) Employing FCC with paraffinic naphtha feed results in a naphtha range product with lower octane number than that obtained by FCC of wax. With wax, the unhydrogenated naphtha product typically had an RON of 84–88 and MON of 74–76, but with naphtha feed the product had an RON of 74–79 and MON of 65–70 [53]. These values were obtained by cracking in the temperature range 470–520 Ž C, and increasing the temperature had little effect on the quality of the naphtha. 6) Naphtha range alkanes are difficult to crack, and even at very high temperature the conversion of Fischer–Tropsch naphtha will be strongly dependent on the alkene and oxygenate content of the naphtha feed. Aromatization at very high temperature may increase the octane number of the naphtha fraction somewhat, but typically at the expense of alkene conversion into alkanes. There is consequently a trade-off between petrochemical and fuel production by FCC of syncrude. 7) In petrochemical applications, FCC at severe conditions may be contemplated. If such operation falls outside the temperature range of typical FCC operation (Table 21.1), care
21.5 Hydrocracking
8)
9)
10)
11)
12)
must be taken to confirm the higher temperature characteristics of the equipment. Both physical and chemical aspects need to be considered. The thermal expansion coefficients and strength of the metals and refractory bricks developed for typical FCC operation must be evaluated for higher temperature operation. A mismatch in thermal expansion or mechanical strength can lead to problems. The activity of metal surfaces for dehydrogenation and coking should also be evaluated. Oxygenates may also undergo dehydrogenation and hydrogenation reactions [68]. Conducting FCC at high temperature, where thermal cracking becomes significant beyond the initial thermal shock at the bottom of the riser, affects reaction engineering considerably. When the contribution of thermal cracking cannot be ignored, cracking may continue after the riser and disengagement of the catalyst, because thermal cracking does not require a catalyst. The total residence time before product cooling quenches free radical reactions now becomes critical. Until the product is quenched, free radical reactions may continue. The time–temperature profile is likewise critical. As the product is cooled down, it passes through the temperature region where thermal oligomerization (Section 19.3.6) takes place. Radical recombination reactions may lead to the formation of heavy oils and carbonaceous deposits when the product spends time in the 300–450 Ž C temperature region. High-temperature operation will also lead to changes in product selectivity. Some products, such as benzene, dienes, CH4 , and H2 , are more abundant in products from thermal cracking than in products from conventional FCC. This not only affects downstream refining but also increases the dry gas flow from the FCC. The impact of increased benzene production on the refining requirements to meet motor-gasoline specifications is self-evident. Additionally, oxygenates produce CO and CO2 . Feed preheating for high-temperature FCC is also challenging. Some thermal cracking of the syncrude can be initiated at temperatures above 300–320 Ž C. If linear velocities are too low, heavy products will be deposited on the exchanger surfaces, causing fouling problems. This is especially likely with alkene- and oxygenate-rich syncrude feeds. Oxygenates present in the syncrude may be more aggressive to the cracking catalyst than steam. Hydrothermal dealumination by oxygenates may lead to a higher catalyst consumption rate. Although high oxygenate conversion can be achieved in an FCC unit, the product may contain some carboxylic acids. Downstream processing equipment must be designed to handle acidic products.
21.5 Hydrocracking
In Fischer–Tropsch refining, hydrocracking has emerged as an important conversion technology for the upgrading of LTFT waxes in gas-to-liquids facilities (Chapters 11 and 12). It is industrially employed in a number of commercial facilities. Hydrocracking of LTFT waxes is different to typical crude oil hydrocracking technology in a number of respects (Table 16.2) [4, 69]. Fischer–Tropsch wax hydrocracking requires milder conditions than crude oil hydrocracking, and on account of its low heteroatom
427
428
21 Cracking
and aromatics content, LTFT wax hydrocracking consumes little hydrogen and is almost isothermal. HTFT residues, on the other hand, more closely resemble aromatic crude oil fractions, albeit without sulfur- and nitrogen-containing compounds. One should therefore not confuse the mild processing requirements of LTFT wax hydrocracking with the more severe requirements for HTFT residue hydrocracking. The inclusion of a hydrocracker in a Fischer–Tropsch refinery is typically considered for the following applications (Section 16.5.2): 1)
2)
3)
4)
5)
The use of wax hydrocracking to produce distillate blend stock is the best known and industrially the most practised application of hydrocracking in Fischer–Tropsch refining. The distillate that can be produced from hydrocracking LTFT wax has a low aromatic content (<2%), high cetane number (>70), and low density (<780 kgÐm3 ) [39, 70, 71]. However, when hydrocracking of LTFT wax is considered for diesel fuel production, it is important to define a blending (or refining) strategy to produce final on-specification diesel fuel. The distillate that is produced by wax hydrocracking does not meet EN590:2004 diesel fuel specifications, and refining such material to meet these specifications is not trivial [72]. The kerosene fraction produced by hydrocracking of LTFT wax can be optimized for a high branched to linear alkane ratio (iso:n-ratio). A high hydrocracking conversion is required to produce a kerosene with high iso:n-ratio, since the degree of isomerization of the kerosene fraction from wax hydrocracking is low at low conversion [40]. Material with a sufficiently high iso:n-ratio has a low freezing point and can be employed as synthetic jet fuel component, or as battlefield-use fuel of the future (BUFF) [73–75]. Hydrocracking can also be used for the production of synthetic jet fuel from HTFT syncrude, and, in general, hydrocracking is important for maximizing synthetic jet fuel production from Fischer–Tropsch syncrude. The hydrocracking of coal liquids for synthetic jet fuel production in combination with Fischer–Tropsch-derived kerosene has also been considered [76]. Lubricant base oil manufacturing requires hydroisomerization (Section 18.4) and only mild hydrocracking. When dealing with a heavy LTFT wax, it can be beneficial to perform mild hydrocracking, which preferentially converts the heavier wax fraction and inevitably results in hydroisomerization [77, 78]. The production of base oils by hydrocracking HTFT material has also been reported [79]. Although the HTFT residue constitutes only a small fraction of HTFT syncrude, mild hydrocracking facilitates the production of on-specification EN590:2004 diesel fuel, not just distillate blend stock (Table 21.7) [80]. HTFT residues can also be combined with coal liquids from low-temperature coal gasification (Section 3.4.2) to be co-refined in a hydrocracking process. It was suggested that it is better to perform the hydrocracking of such material in an ebulated bed or slurry-phase hydrocracker [81]. Both feed materials are aromatic and contain compounds that lead to deposit formation and pressure drop problems in a fixed bed hydrocracker. Ebulated bed or slurry phase hydrocracking is a good technology for increasing the liquid yield from a coal-to-liquids Fischer–Tropsch facility without adding complexity to the refinery (Figure 21.7). An analogous configuration may be employed in biomass-to-liquids and waste-to-liquid facilities.
21.5 Hydrocracking Mild hydrocracking of the light and heavy vacuum gas oil fractions of HTFT syncrude over a sulfided Ni/H-ZSM-5 catalyst. The wide boiling range is due to poor industrial separation.
Table 21.7
Description
HTFT feed
Product distribution (mass%) <190 Ž C naphtha 190–360 Ž C distillate 360–490 Ž C vacuum gas oil >490 Ž C vacuum residue Hydrotreated distillate properties Density at 20 Ž C (kg m3 ) Cetane number Cloud point (Ž C) Acid number (mg KOH/g) Viscosity at 40 Ž C (cSt) Flash point (Ž C) Lubricity, HFRR wear (µm) Total aromatics (mass%) Multinuclear aromatics (mass%) T95 distillation (Ž C)
Tail gas processing
Fischer– Tropsch
16 68 16 <1 802.3 60 –3 0.02 2.2 – 488 22.3 0.3 363
Mild HTFT hydrocracking 350 ◦ C
360 ◦ C
370 ◦ C
26 59 14 1
30 54 15 1
36 51 12 1
817.4 56 –14 0.018 2.3 92 486 20.8 <0.5 363
832.1 49 –64 0.01 2.4 93 363 25.4 <0.5 353
Refinery Condensates
Light oil
Fischer– Tropsch refinery units
Fuel blending
Aqueous product Decanted oil
Tar refinery units
Syngas conditioning Coal naphtha
Coal
821.6 55 –27 0.01 2.3 92 488 23.2 <0.5 358
Gasification
Coal liquids
Slurryphase hydrocracking
Figure 21.7 Slurry bed hydrocracking of a combined Fischer–Tropsch and coal liquid feed to overcome problems associated with dissolved metals and solid particulate matter in the heavy oil fractions of a coal-to-liquids facility.
Fuels
429
21 Cracking
21.5.1 Catalysis
Hydrocracking catalysts must have metal and acid sites. This follows from the mechanism of hydrocracking (Section 21.2.3). Many different catalysts have been investigated for n-alkane hydrocracking in general and for Fischer–Tropsch wax hydrocracking in particular [65]. Based on the metal functionality, there are, broadly speaking, three classes of catalysts that are relevant for the hydrocracking of syncrude: 1) Sulfided base metals on acidic support materials. These are typical crude oil hydrocracking catalysts and require the addition of a sulfiding agent to keep the metals in their sulfided state. Such catalysts are industrially applied in some Fischer–Tropsch facilities with HTFT and LTFT syncrudes (Chapters 9 and 12). The catalysts of this type that have most often been studied are NiMo and NiW on ASA [82]. It is possible to employ zeolite support materials, as is the case with crude oil [82, 83], but zeolites tend to be too strongly acidic for Fischer–Tropsch syncrude. The reasons for this are explained later. 2) Unsulfided base metals on acidic support materials. Such catalysts were tested with model n-alkane feed materials, but hydrogenolysis and deactivation by oxygenates present obstacles to industrial application. 3) Unsulfided noble metal catalysts. This class of catalysts is applied industrially in some Fischer–Tropsch facilities with LTFT syncrude (Chapter 11). Hydrocracking catalysts based on Pt/SiO2 –Al2 O3 were found to be very good for wax cracking, and higher distillate selectivities can be obtained at high conversion than over sulfided base metal catalysts (Figure 21.8) [39, 84]. Good performance with HTFT syncrude has also been reported [80]. 80 Distillate
Pt/Siral75 (Leckel) PtW/Siral75 (Leckel) Pt/ASA (Calemma, et al.) Sulfided base metal
70 60 Yield (mass%)
430
50 40 30 20 10 Naphtha
0 0
10
20
30
40
50
60
70
80
Conversion of C22 and heavier material (%) Figure 21.8 Hydrocracking of LTFT wax over Pt- and PtW-promoted silicated amorphous silica–alumina (Siral75) catalysts, Pt-promoted amorphous silica–alumina (ASA) catalyst, and a commercially available sulfided base metal hydrocracking catalyst.
90
100
21.5 Hydrocracking
It is possible to employ zeolite support materials, but as in the case of sulfided base metal catalysts, zeolites tend to have too much strong acidity for distillate-selective hydrocracking. One of the main parameters affecting the performance of a bifunctional catalyst for hydrocracking is the ratio of metal sites to acid sites. This follows directly from the reaction mechanism (Figure 21.4), which requires alkane dehydrogenation over a metal site before isomerization or cracking can take place on an acid site. Although it is convenient to refer to the metal-to-acid site ratio as the ratio of the number of sites of each, it does not take into account how effective the sites are. The ratio of the turnover frequency (TOF) of the metal-to-acid sites is more important, because the TOF also accounts for differences in the activity of the different sites of each type. At low metal-to-acid site ratio, there is not enough metal site activity to quickly establish quasi-equilibrium between alkanes and alkenes on the catalyst surface. The number of acid sites that can take part in catalysis is dependent on alkene availability, because protolysis is usually only a minor cracking pathway at hydrocracking conditions. When the dehydrogenation of alkanes to alkenes is the rate-limiting step, many of the acid sites are interacting with alkanes (not alkenes) and are effectively not taking part in the catalysis. For example, it has been found that the n-heptane conversion over Pt/FAU catalysts decreased as the number of acid sites was increased for a constant metal site loading when the metal-to-acid site ratio was 1 : 35 or less [85]. The acid sites were effectively alkene-starved. When the metal-to-acid site ratio was increased beyond 1 : 35, conversion remained constant. As pointed out before, it is actually the metal-to-acid TOF ratio that is important. If the metal sites are not very active, a higher ratio of metal-to-acid sites will be required than would be necessary for a more active metal. The same holds true for the acid support material. Hydrogenation activity increases in the order sulfided NiMo < sulfided NiW < PtPd [82]. Acid cracking activity increases in the order Al2 O3 < halogenated Al2 O3 < ASA < zeolites [82]. One would therefore expect that sulfided base-metal-promoted hydrocracking catalysts would require a higher metal loading than unsulfided noble metal catalysts to achieve the same metal-to-acid TOF ratio. For example, it was found that, when the same acid support was impregnated with different transition metals to achieve the same metal-to-acid site ratio, there were notable differences in performance between the catalysts [86]. The nature of the metal is also important and can affect product selectivity by catalyzing reactions specific to that metal. Reduced nickel’s propensity for hydrogenolysis [31] and platinum’s ability to catalyze ring closure [36] are two pertinent examples. When a catalyst has sufficient metal site activity, catalyst activity can be increased by increasing the activity (TOF) of the acid sites while keeping the metal-to-acid site ratio constant [85]. The activity and strength of acid sites are often correlated, but not necessarily so. Activity refers to the TOF of the site for the reaction of interest. Strength refers to the ease with which reactions can be catalyzed and the type of reactions that can be catalyzed. Increasing the strength of an acid site may enable reaction pathways that are not available to weaker acid sites, and by doing so the selectivity can be changed. Very strong acid sites are capable of catalyzing protolysis (Figure 21.2) and can convert alkanes directly. Weaker acid sites can perform cracking of alkenes only by β-scission (Figure 21.3) and cannot convert alkanes directly. In Fischer–Tropsch applications, the best selectivity is obtained when cracking takes place only by β-scission on weaker acid sites. For example, sulfided base metal and unsulfided noble-metal-promoted ASA catalysts resulted in less overcracking and better product properties than zeolite-based hydrocracking catalysts (Table 21.8) [80].
431
432
21 Cracking Table 21.8 Hydrocracking of the heavy vacuum gas oil fraction of HTFT syncrude over different catalysts at 5 MPa and LHSV of 0.5 h1 . The base metal catalysts were sulfided and the noble metal catalyst was unsulfided.
Description
HTFT feed Pt/MoO3 /ASA NiMo/ASA NiW/zeolite-#1 NiW/zeolite-#2 390 ◦ C
Product distribution (mass%) <190 Ž C naphtha 190–360 Ž C distillate >360 Ž C vacuum gas oil Hydrotreated distillate properties Density at 20 Ž C (kgÐm3 ) Cetane number Cloud point (Ž C) Acid number (mg KOH/g) Viscosity at 40 Ž C (cSt) Lubricity, HFRR wear (µm) Total aromatics (mass%) Multinuclear aromatics (mass%)a T95 distillation (Ž C) a The
0 64 36
16 74 15
843.9 53 –11 12 3.6 251 34 – 343
829 56 –15 1.5 2.8 340 22 – 325
370 ◦ C
370 ◦ C
370 ◦ C
25 70 5
30 65 5
47 53 0
814.5 60 –10 0.02 2.5 422 15.8 2.9 325
806.5 61 –10 0.02 2.3 357 13.6 3.1 319
820.1 61 –9 0.03 2.7 458 22.4 5.4 333
multinuclear aromatics consist of 90% or more dinuclear aromatics.
An ‘‘ideal’’ hydrocracking catalyst is one that limits secondary cracking reactions. To put it in another way, an ‘‘ideal’’ hydrocracking catalyst allows an alkene intermediate only one acid-catalyzed transformation before the alkene is hydrogenated to an alkane. Such hydrocracking behavior is typified by a symmetric carbon number distribution around the average carbon number of the feed, with equal selectivity to all C4 to Cn3 products derived from a Cn feed and somewhat lower selectivity to C3 and Cn2 products [87]. The higher the metal-to-acid ratio of a hydrocracking catalyst, the more ‘‘ideal’’ the catalyst becomes (within limits, too high a metal content may result in an increase in metal-catalyzed side reactions). For such ‘‘ideal’’ hydrocracking catalysts, the relationship between product selectivity and conversion appears to be independent of the operating conditions [88]. Using a 0.5% Pt/USY catalyst in the operating range 180–240 Ž C and 0.7–10 MPa, it was found that with n-decane as feed, hydroisomerization and hydrocracking selectivities were functions of conversion only. ‘‘Ideal’’ hydrocracking is related to the ability of the catalyst to establish a quasi-equilibrium between the alkanes and alkenes on the catalyst surface. The ability to do so is not just dependent on the catalyst. A catalyst with high metal-to-acid ratio and that exhibits ‘‘ideal’’ hydrocracking under one set of operating conditions may not be able to do so under another set of operating conditions. Four operating effects were identified from kinetic analysis that can lead to ‘‘nonideal’’ hydrocracking [89]: 1) Low pressure. As the pressure is decreased, the partial pressure of H2 decreases. This favors the dehydrogenation of alkanes to produce alkenes and causes the carbocation concentration
21.5 Hydrocracking
on the catalyst surface to increase. This increases the rate of the acid-catalyzed reactions but decreases the ability of the metal sites to hydrogenate the products before secondary reactions can take place. Lowering the pressure effectively lowers the TOF of the metal sites and, by doing so, effectively changes the metal-to-acid TOF ratio. This is of importance to Fischer–Tropsch hydrocracking, which can be performed at much lower pressures than crude oil hydrocracking. Stable LTFT wax hydrocracking operation has been reported at 3.5 MPa and stable operation at 1 MPa was observed, but no extended lifetime testing was performed [90]. The ability to perform wax hydrocracking at low pressure is partly due to the low aromatization propensity of wax under hydrocracking conditions [91]. 2) High temperature. The overall activation energy needed for metal-catalyzed reactions (dehydrogenation and hydrogenation) is lower than that required for acid-catalyzed reactions (isomerization and cracking) at low temperature. As temperature is increased, this relationship changes. As in the case of low pressure, hydrogenation becomes increasingly difficult at high temperature, which favors dehydrogenation. The ability of the metal sites to hydrogenate the products before secondary reactions can take place is undermined by high temperature, which leads to ‘‘nonideal’’ hydrocracking behavior. 3) Low H2 :hydrocarbon ratio. The H2 :hydrocarbon ratio affects H2 availability. At a low H2 :hydrocarbon ratio, the partial pressure of H2 effectively decreases as the amount of hydrocarbons in the gas phase increases. 4) Heavier feed. The reactivity of n-alkanes for hydrocracking increases with increasing chain length [92]. A definite conversion bias based on the heaviness of the feed has been observed with Fischer–Tropsch wax [93]. As the carbon number of the feed increases, the enthalpy of chemisorption becomes more negative. In mixtures, the longer chain hydrocarbons are more strongly adsorbed on the catalyst than the shorter chain hydrocarbons. This results in a higher concentration and residence time of longer chain hydrocarbons on the metal sites, which may cause an apparent imbalance in the metal-to-acid TOF ratio. This is a competitive adsorption effect, with heavier compounds being favored for adsorption [94]. The metal-to-acid TOF ratio can also be affected by compounds that preferentially adsorb onto either metal sites or acid sites. Aromatics preferentially adsorb onto acid sites to increase the effective metal-to-acid ratio [95]. Alcohols likewise adsorb preferentially onto acid sites to increase the effective metal-to-acid ratio [96]. The same is also true of nitrogen bases, which preferentially adsorb onto acid sites to increase the metal-to-acid ratio. Although sulfur is necessary for base metal catalysts, it lowers the hydrogenation activity of the metal function, and some metal sulfides have acidic properties. Sulfur compounds therefore result in an effective decrease in the metal-to-acid ratio. When Fischer–Tropsch syncrude is hydrocracked over a sulfided hydrocracking catalyst, the sulfur content becomes an adjustable parameter, which can be exploited to manipulate selectivity during operation [97]. Carboxylic acids preferentially adsorb on metal sites to cause an effective decrease in the metal-to-acid ratio [96]. Only some of these species are present in Fischer–Tropsch syncrude. The influence of oxygenates is relevant to the hydrocracking of syncrudes, whereas the influence of aromatics is only relevant to the hydrocracking of HTFT material. Another influence of the operating conditions that is less obvious from a kinetic analysis, is the impact on reactor hydrodynamics. Pressure, temperature, H2 :hydrocarbon ratio, and the heaviness of the feed all affect the vapor–liquid equilibrium. Vapor–liquid equilibrium determines the volumetric ratio between the vapor and liquid phases and thereby it influences
433
434
21 Cracking
hydrodynamic parameters important for reaction, such as liquid holdup. The vapor–liquid equilibrium also determines how each compound will partition between phases. It stands to reason that the conversion of compounds that are mainly in the vapor phase will be limited relative to compounds that are mainly in the liquid phase, since the hydrocracking catalyst is liquid filled. 21.5.2 Syncrude Processing Technology
The hydrocracking of HTFT residue and LTFT wax is quite different, because the >360 Ž C material from these syncrudes are very different. HTFT residues contain around 25% aromatics (Tables 21.6 and 21.7), whereas LTFT wax consists of mainly n-alkanes. Some aspects to consider when applying hydrocracking technology to Fischer–Tropsch syncrude are the following: 1) The metal carboxylates in the heavier fractions of HTFT syncrude make residue conversion units pressure drop constrained rather than activity constrained [98]. By hydrocracking the HTFT residue on its own or in combination with pyrolysis products in an ebulated or slurry bed reactor (Figure 21.7) can reduce operating problems with particulate matter and dissolved metals. The same applies to the hydrocracking of LTFT waxes containing dissolved metals or colloidal suspensions of metals. 2) Hydrocracking is not a preferred technology to produce good quality motor-gasoline. It is best suited for distillate and lubricant base oil production. 3) For distillate-selective hydrocracking, less acidic supports are better. In general, ASA supports give a better distillate yield than zeolite supports [82]. This was confirmed during HTFT hydrocracking (Tables 21.7 and 21.8), where the sulfided base metal and unsulfided noble metal catalysts using an ASA support resulted in better distillate yields and quality. 4) Industrially there are two opposing views about the preferred metal for wax hydrocracking catalysts. The Sasol–Chevron facilities (Chapter 12) make use of sulfided base metal hydrocracking catalysts. Such catalysts are cheaper, but less selective than noble metal catalysts and their use requires the addition of sulfur to the otherwise sulfur-free syncrude. The Shell facilities (Chapter 11) make use of unsulfided noble metal hydrocracking catalysts. Most developments in the field of Fischer–Tropsch hydrocracking is focused on unsulfided noble-metal-promoted catalysts, such as the work by Calemma and coworkers at ENI on Pt/SiO2 –Al2 O3 [39]. In fact, Pt/SiO2 –Al2 O3 was identified as one of the four key catalyst types in Fischer–Tropsch refining [99]. 5) Hydrotreating before hydrocracking can improve selectivity during hydrocracking of HTFT syncrude [80] as well as LTFT syncrude [73]. The extent of the benefit depends on the oxygenate and aromatic content of the syncrude, as well as on the composition of the catalyst. In conventional hydrocrackers, the cost of Pt is a major contributor to the overall cost of the unit and hydrocracking catalysts are typically optimized to minimize the amount of Pt. This makes the hydrocracking catalyst sensitive to operating effects, which may cause ‘‘nonideal’’ behavior (Section 21.5.1) [89]. Using a hydrocracking catalyst with slightly more Pt can reduce the impact of such feed effects without prior hydrotreating. 6) The severity of hydrocracking affects the product selectivity. Under very mild conditions where hydroisomerization is dominant, lubricant base oils can be produced. As the severity is increased, the probability of producing naphtha is increased. This is related to the carbon number distribution of the material, which determines the selectivity [93]. This
21.5 Hydrocracking
7)
8)
9)
10)
relationship can be modeled [100] and, superficially speaking, it codifies the product selectivity in terms of the cracking probability at C3 to Cn2 positions along the carbon chain [87]. As the carbon chain becomes longer, the probability of producing naphtha by cracking theoretically decreases. Since the carbon number distribution becomes lighter with increasing hydrocracking conversion, the increase in naphtha selectivity with conversion is a natural consequence of the mechanism. The per pass conversion, and decision to implement once-through or recycle operation, both have an impact on the product yields and the degree of isomerization in each product fraction. For the highest degree of isomerization, a low per pass conversion with recycle is recommended. The fundamental basis for this follows from Figure 21.6. Introducing an isolated branch with subsequent hydrogenation is likely to increase product branching, because subsequent dehydrogenation, isomerization, and cracking may not take place in proximity to the branch, thereby ensuring additional branching in the product. The implementation of product recycle in hydrocracking should take the nature of the product into account. The ‘‘unconverted’’ material that is recycled is likely to be isomerized and have a lighter carbon number distribution than the feed. Similar product yield can be achieved only if the recycle is introduced into the reactor at a position where it is comparable to the carbon number distribution of the partially converted heavy fraction from the fresh feed (Figure 21.9). In fact, there is a distinct selectivity advantage to fractionate the product in more than just a >360 Ž C fraction, so that the light vacuum gas oil can be recycled to a lower position in the hydrocracker than the heavy vacuum gas oil and bottom product (Figure 21.10). The higher reactivity of long-chain n-alkanes in the hydrocracker can be exploited during once-through operation by introducing heavy feed at different positions along the bed. This configuration is similar to a reactive quench in oligomerization (Figure 19.2) or aromatic alkylation. By doing so, a heavy hydrocarbon fraction is present throughout the reactor to suppress overconversion of the lighter material. Fischer–Tropsch wax hydrocracking is near isothermal. The engineering design must make sufficient provision for preheating to ensure that the hydrocracker can be operated over the operating range of the catalyst. No quench streams or interbed cooling is needed for wax hydrocracking.
Fresh H2 Wax
H 2 recycle
Purge LPG Naphtha Distillate
>360 °C recycle
Figure 21.9 Simple wax hydrocracker recycle configuration that introduces the recycle lower in the reactor to match partial conversion of the fresh feed. Feed heating is not shown.
435
436
21 Cracking
Fresh H2 H 2 recycle
Wax
Purge LPG
Figure 21.10 Complex wax hydrocracker recycle configuration for better product quality and distillate yield, which employs finer product fractionation to better match recycle position to partial conversion of the fresh feed. The order of feed introduction from the top of the reactor is fresh wax feed, vacuum residue (VR), heavy vacuum gas oil (HVGO), and light vacuum gas oil (LVGO). Feed heating is not shown.
Naphtha Distillate
LVGO
>360 °C recycle
HVGO VR
11) The low coking propensity of Fischer–Tropsch waxes allows hydrocracker operation at lower hydrogen pressures. By operating at a lower hydrogen pressure, it is also possible to lower the operating temperature of hydrocracking [90]. 12) The hydrogen consumption during LTFT wax hydrocracking is much less than during HTFT residue or crude oil hydrocracking. This is due to the low heteroatom and aromatic content in the wax, with little additional H2 being consumed apart from hydrogenation of the cracking products.
References 1. Leprince, P. (2001) in Petroleum Refining,
2.
3.
4.
5.
6.
7. Blanksby, S.J. and Ellison, G.B. (2003) Bond Conversion Processes, vol. 3 (ed. dissociation energies of organic molecules. Acc. P. Leprince), Editions Technip, Paris, Chem. Res., 36, 255–263. 8. Rice, F.O. (1931) The thermal decomposition of pp. 365–379. Oblad, A.G. (1983) The contributions of Euorganic compounds from the standpoint of free gene J. Houdry to the development of catalytic radicals. I. Saturated hydrocarbons. J. Am. Chem. Soc., 53, 1959–1972. cracking. ACS Symp. Ser., 222, 61–75. Bell, H.S. (ed.) (1959) American Petroleum Refin9. Kossiakoff, A. and Rice, F.O. (1943) Thering, 4th edn, D. Van Nostrand, Princeton, NJ, mal decomposition of hydrocarbons, resonance stabilization and isomerization pp. 240–263. Billon, A. and Bigeard, P.-H. (2001) in Petroleum of free radicals. J. Am. Chem. Soc., 65, 590–595. Refining, Conversion Processes, vol. 3 (ed. P. 10. Melancon, S., Le Van Mao, R., Kletnieks, P., Leprince), Technip, Paris, pp. 333–364. Gary, J.H., Handwerk, G.E., and Kaiser, M.J. Ohayon, D., Intem, S., Saberi, M.A., and McCann, D. (2002) Selective deep catalytic (2007) Petroleum Refining Technology and Ecocracking process of hydrocarbon feedstocks for nomics, 5th edn, CRC Press, Boca Raton, FL, the production of light olefins: II. Cooperative pp. 121–179. Sullivan, R.F. and Scott, J.W. (1983) The develeffect of thermal cracking and catalytic reactions opment of hydrocracking. ACS Symp. Ser., 222, observed in a 1-zone reactor. Catal. Lett., 80, 293–313. 103–109.
References 11. De Klerk, A. (2007) Thermal cracking of
12.
13.
14.
15. 16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
26.
Fischer-Tropsch waxes. Ind. Eng. Chem. Res., 46, 5516–5521. McMillen, D.F. and Golden, D.M. (1982) Hydrocarbon bond dissociation energies. Ann. Rev. Phys. Chem., 33, 493–532. Berkowitz, J., Ellison, G.B., and Gutman, D. (1994) Three methods to measure RH bond energies. J. Phys. Chem., 98, 2744–2765. Ranzi, E., Dente, M., Pierucci, S., and Biardi, G. (1983) Initial product distributions from pyrolysis of normal and branched paraffins. Ind. Eng. Chem. Fundam., 22, 132–139. Borden, W.T. (ed.) (1982) Diradicals, John Wiley & Sons, Inc., New York. Sundaram, K.M. and Froment, G.F. (1978) Modeling of thermal cracking kinetics. 3. Radical mechanisms for the pyrolysis of simple paraffins, olefins and their mixtures. Ind. Eng. Chem. Fundam., 17 (3), 174–182. Linden, H.R. and Peck, R.E. (1955) Gaseous product distribution in hydrocarbon pyrolysis. Ind. Eng. Chem., 47, 2470–2474. Schneider, V. and Frolich, P.K. (1931) Mechanism of formation of aromatics from lower paraffins. Ind. Eng. Chem., 23, 1405–1410. Bragg, L.B. (1941) Rate of cracking of paraffin wax. Ind. Eng. Chem., 33, 376–380. Worrell, G.R., Chapman, R.F., and Woinsky, S.G. (1967) Kinetics of wax cracking for alpha-olefin production. Ind. Eng. Chem. Proc. Des. Dev., 6 (1), 89–95. Voge, H.H. and Good, G.M. (1949) Thermal cracking of higher paraffins. J. Am. Chem. Soc., 71, 593–597. Kotrel, S., Kn¨ozinger, H., and Gates, B.C. (2000) The Haag-Dessau mechanism of protolytic cracking of alkanes. Microporous Mesoporous Mater., 35–36, 11–20. Buchanan, J.S., Santiesteban, J.G., and Haag, W.O. (1996) Mechanistic considerations in acid-catalyzed cracking of olefins. J. Catal., 158, 276–287. Corma, A., Planelles, J., S´anchez-Mar´ın, J., and Tom´as, F. (1985) The role of different types of acid site in the cracking of alkanes on zeolite catalysts. J. Catal., 93, 30–37. Van Santen, R.A. and Kramer, G.J. (1995) Reactivity theory of zeolitic Brønsted acidic sites. Chem. Rev., 95, 637–660. Corma, A. and Orchill´es, A.V. (2000) Current views on the mechanism of catalytic cracking. Microporous Mesoporous Mater., 35–36, 21–30.
27. Steijns, M. and Froment, G. (1981) Hydroiso-
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
merization and hydrocracking. 3. Kinetic analysis of rate data for n-decane and n-dodecane. Ind. Eng. Chem. Prod. Res. Dev., 20, 660–668. Matsumoto, H., Saito, Y., and Yoneda, Y. (1970) Contrast between nickel and platinum catalysts in hydrogenolysis of saturated hydrocarbons. J. Catal., 19, 101–112. Matsumoto, H., Saito, Y., and Yoneda, Y. (1971) The classification of metal catalysts in hydrogenolysis of hexane isomers. J. Catal., 22, 182–192. Pines, H. and Postl, W.S. (1957) Modified nickel catalyst for the hydrogenolysis and alkylation of aromatic hydrocarbons. J. Am. Chem. Soc., 79, 1769–1770. Dos Santos, L.T., Herbst, M.H., and Pereira, M.M. (2004) Nickel activation for hydrogenolysis reaction on USY zeolite. Catal. Lett., 92, 81–86. Corolleur, C., Tomanova, D., and Gault, F.G. (1972) The mechanisms of hydrogenolysis and isomerization of hydrocarbons on metals. VII. Isomerization of labeled hexanes and hydrogenolysis of methyl (13 C) cyclopentane on a 10% platinum-alumina catalyst. J. Catal., 24, 401–416. Santacesaria, E., Gelosa, D., Carra, S., and Adami, I. (1978) Influence of metal dispersion on n-hexane reactions over platinum-alumina catalysts. Ind. Eng. Chem. Prod. Res. Dev., 17, 68–71. Barron, Y., Maire, G., Muller, J.M., and Gault, F.G. (1966) The mechanisms of hydrogenolysis and isomerization of hydrocarbons on metals. II. Mechanisms of isomerisation of hexanes on platinum catalysts. J. Catal., 5, 428–446. Coonradt, H.L. and Garwood, W.E. (1964) Mechanism of hydrocracking. Reactions of paraffins and olefins. Ind. Eng. Chem. Process Des. Dev., 3, 38–45. Chaar, M.A. and Butt, J.B. (1994) Conversion of n-heptane over Pt- and Pd/SAPO-11 catalysts. Appl. Catal. A, 114, 287–293. Davis, B.H. and Venuto, P.B. (1969) Paraffin dehydrocyclization. Distribution of aromatic products obtained with ‘‘nonacidic’’ supported Pt catalysts. J. Catal., 15, 363–372. Weitkamp, J. and Farag, H. (1978) Isomerization of the methylnonanes and 2-methyloctane on a bifunctional zeolite catalyst. Acta Phys. Chem., 24, 327–333. Calemma, V., Gambaro, C., Parker, W.O. Jr., Carbone, R., Giardino, R., and Scorletti, P. (2010) Middle distillates from hydrocracking
437
438
21 Cracking
40.
41.
42.
43.
44.
45. 46.
47.
48.
49.
50.
51.
52.
53.
of FT waxes: composition, characteristics and emission properties. Catal. Today, 149, 40–46. Gamba, S., Pellegrini, L.A., Calemma, V., and Gambaro, C. (2010) Liquid fuels from Fischer-Tropsch wax hydrocracking: Isomer distribution. Catal. Today, 156, 58–64. (2001) Linear Alpha Olefins. SRI Process Economics Program Report 12D, Stanford Research Institute, Menlo Park, CA. Snodgrass, C.S. and Perrin, M. (1938) The production of Fischer-Tropsch coal spirit and its improvement by cracking. J. Inst. Petrol. Technol., 24, 289–301. Dazeley, G.H. and Hall, C.C. (1948) Production of olefins by the cracking of Fischer-Tropsch waxes and their conversion into lubricating oils. Fuel, 27 (2), 50–55. Weil, B.H. and Lane, J.C. (1949) The Technology of the Fischer-Tropsch Process, Constable, London. Horne, W.A. (1950) Review of German synthetic lubricants. Ind. Eng. Chem., 42, 2428–2436. K¨onnecke, H.G. and Gawalek, G. (1956) Uber die spaltung von, F.-T.-Weichparaffin (Engl. Transl. ‘‘On the cracking of F.-T. soft wax’’). Chem. Tech. Berlin, 8 (10), 603–604. Huibers, D.T.A. and Waterman, H.I. (1960) Die produktverteilung bei der thermischen crackung aliphatischer kohlenwasserstoff-fraktionen (Engl. Transl. ‘‘The product distribution from thermal cracking of aliphatic hydrocarbon fractions’’). Brennst.-Chem., 41 (10), 297–299. De Klerk, A. (2005) Thermal upgrading of Fischer-Tropsch olefins. Energy Fuels, 19, 1462–1467. Kobolakis, I. and Wojciechowski, B.W. (1985) The catalytic cracking of a Fischer-Tropsch synthesis product. Can. J. Chem. Eng., 63, 269–277. Abbot, J. and Wojciechowski, B.W. (1985) Catalytic cracking on HY and HZSM-5 of a Fischer-Tropsch product. Ind. Eng. Chem. Prod. Res. Dev., 24, 501–507. Leib, T.M., Kuo, J.C.W., Garwood, W.E., Nace, D.M., Derr, W.R., and Tabak, S.A. (1988) Upgrading of Fischer-Tropsch waxes to high quality transportation fuels. Proceedings AIChE Annual Meeting, Washington, DC, Paper 61d. Reagan, W.J. (1994) Gasoline range ether synthesis from light naphtha products of fluid catalytic cracking of Fischer-Tropsch wax. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 39 (2), 337–342. Schwartz, M.M. (1995) The Selective Catalytic Cracking of Fischer-Tropsch Liquids to High
54.
55.
56.
57.
58. 59.
60.
61.
62.
63.
64.
Value Transportation Fuels, DOE Contract no. DE-AC22– 91PC90057 - Final report. Choi, G.N., Kramer, S.J., Tam, S.S., Fox, J.M. III, and Reagan, W.J. (1996) Fischer-Tropsch indirect coal liquefaction design/economics – mild hydrocracking vs. fluid catalytic cracking. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 41 (3), 1079–1083. Dupain, X., Krul, R.A., Makkee, M., and Moulijn, J.A. (2005) Are Fischer-Tropsch waxes good feedstocks for fluid catalytic cracking units? Catal. Today, 106, 288–292. Dupain, X., Krul, R.A., Schaverien, C.J., Makkee, M., and Moulijn, J.A. (2006) Production of clean transportation fuels and lower olefins from Fischer-Tropsch synthesis waxes under fluid catalytic cracking conditions. The potential of highly paraffinic feedstocks for FCC. Appl. Catal. B, 63, 277–295. Eng, C., Heidenreich, S., Swart, S., and M¨oller, F. (2005) Clean fuels and petrochemicals at Sasol via Superflex. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd122. (2005) Gaint project turbo advances. Chem. Technol. (S. Afr.), 13–16. Buchanan, J.S. (1991) Reactions of model compounds over steamed ZSM-5 at simulated FCC reaction conditions. Appl. Catal., 74, 83–94. Baerlocher, C., Meier, W.M., and Olson, D.H. (2001) Atlas of Zeolite Framework Types, 5th edn, Elsevier, Amsterdam. Ohkita, H., Nishiyama, R., Tochihara, Y., Mizushima, T., Kakuta, N., Morioka, Y., Ueno, A., Namiki, Y., Tanifuji, S., Katoh, H., Sunazuka, H., Nakayama, R., and Kuroyanagi, T. (1993) Acid properties of silica-alumina catalysts and catalytic degradation of polyethylene. Ind. Eng. Chem. Res., 32, 3112–3116. Kucherov, A.V., Slinkin, A.A., Gitis, K.M., and Isagulants, G.V. (1988) Direct observation of transition state selectivity: low temperature cyclization of dienes inside H-Mordenite and H-ZSM-5 channels. Catal. Lett., 1, 311–316. You, Y.S., Shim, J.-S., Kim, J.-H., and Seo, G. (1999) Liquid-phase degradation of polyethylene wax over mordenite catalysts with different Si/Al molar ratios. Catal. Lett., 59, 221–227. Blasco, T., Corma, A., and Mart´ınez-Triguero, J. (2006) Hydrothermal stabilization of ZSM-5 catalytic-cracking additives by phosphorus addition. J. Catal., 237, 267–277.
References 65. De Klerk, A. and Furimsky, E. (2010) Catalysis
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Bhusarapu, S., Al-Dahhan, M.H., Dudukovi´c, M.P., Trujillo, S., and O’Hern, T.J. (2005) Experimental study of the solids velocity field in gas-solid risers. Ind. Eng. Chem. Res., 44, 9739–9749. Corma, A., Melo, F.V., Sauvanaud, L., and Ortega, F.J. (2004) Different process schemes for converting light straight run and fluid catalytic cracking naphthas in a FCC unit for maximum propylene production. Appl. Catal. A, 265, 195–206. Riley, M. and Anthony, R.G. (1987) Decomposition of decanol and decanal over Ni- and H-ZSM-5. J. Catal., 103, 87–94. Bouchy, C., Hastoy, G., Guillon, E., and Martens, J.A. (2009) Fischer-Tropsch waxes upgrading via hydrocracking and selective hydroisomerization. Oil Gas Sci. Technol. - Rev. IFP, 64 (1), 91–112. Tijm, P.J.A. (1994) Shell middle distillate synthesis: the process, the plant, the products. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 39 (4), 1146–1150. Leckel, D.O. and Liwanga-Ehumbu, M. (2006) Diesel-selective hydrocracking of an iron-based Fischer-Tropsch wax fraction (C15 -C45 ) using a MoO3 -modified noble metal catalyst. Energy Fuels, 20, 2330–2336. De Klerk, A. (2009) Can Fischer-Tropsch syncrude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. Freerks, R.L. and Muzzell, P.A. (2004) Production and characterization of synthetic jet fuel produced from Fischer-Tropsch hydrocarbons. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 407–410. Forest, C.A. and Muzzell, P.A. (2005) Fischer-Tropsch fuels: why are they of interest to the United States military? SAE Tech. Pap. Ser., 200501–1807. Lamprecht, D. (2007) Fischer-Tropsch fuel for use by the U.S. military as battlefield-use fuel of the future. Energy Fuels, 21, 1448–1453. Roets, P.N.J., Moses, C.A., and Van Heerden, J. (2004) Physical and chemical properties of fully synthetic aviation turbine fuel from coal. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 403–406. Calemma, V., Peratello, S., Stroppa, F., Giardino, R., and Perego, C. (2004) Hydrocracking and hydroisomerization of long-chain
78.
79.
80.
81.
82. 83.
84.
85.
86.
87.
88.
89.
n-paraffins. Reactivity and reaction pathway for base oil formation. Ind. Eng. Chem. Res., 43, 934–940. Kobayashi, M., Saitoh, M., Togawa, S., and Ishida, K. (2009) Branching structure of diesel and lubricant base oils prepared by isomerization/hydrocracking of Fischer-Tropsch waxes and alpha-olefins. Energy Fuels, 23, 513–518. Morgan, P.M., Van der Merwe, D.G., Goosen, R., Leckel, D.O., Saayman, H.M., Loubser, H., and Botha, J.J. (1999) The production of high quality base oils from Sasol hydrocrackates. S. Afr. Mech. Eng., 49, 11–13. Leckel, D.O. (2009) Hydroprocessing Euro 4-type diesel from high-temperature Fischer-Tropsch vacuum gas oils. Energy Fuels, 23, 38–45. Leckel, D.O. (2009) Diesel production from Fischer-Tropsch: the past, the present, and new concepts. Energy Fuels, 23, 2342–2358. Ward, J.W. (1993) Hydrocracking processes and catalysts. Fuel Process. Technol., 35, 55–85. Scherzer, J. and Gruia, A.J. (1996) Hydrocracking Science and Technology, Taylor & Francis, Boca Raton. Leckel, D.O. (2007) Noble metal wax hydrocracking catalysts supported on high-siliceous alumina. Ind. Eng. Chem. Res., 46, 3505–3512. Giannetto, G.E., Perot, G.R., and Guisnet, M.R. (1986) Hydroisomerization and hydrocracking on n-alkanes. 1. Ideal hydroisomerization PtHY catalysts. Ind. Eng. Chem. Prod. Res. Dev., 25, 481–490. Welters, W.J.J., Van der Waerden, O.H., Zandbergen, H.W., De Beer, V.H.J., and Van Santen, R.A. (1995) Hydrocracking of n-decane over zeolite-supported metal sulfide catalysts. 1. CaY-supported transition metal sulfides. Ind. Eng. Chem. Res., 34, 1156–1165. Martens, J.A. (1986) Attempts to rationalize the distribution of hydrocracked products. I Qualitative description of the primary hydrocracking modes of long chain paraffins in open zeolites. Appl. Catal., 20, 239–281. Steijns, M., Froment, G., Jacobs, P., Uytterhoeven, J., and Weitkamp, J. (1981) Hydroisomerization and hydrocracking. 2. Product distributions from n-decane and n-dodecane. Ind. Eng. Chem. Prod. Res. Dev., 20, 654–660. Thybaut, J.W., Narasimhan, C.S.L., Denayer, J.F., Baron, G.V., Jacobs, P.A., Martens, J.A., and Marin, G.B. (2005) Acid-metal balance of a hydrocracking catalyst: Ideal versus nonideal behavior. Ind. Eng. Chem. Res., 44, 5159–5169.
439
440
21 Cracking 90. Leckel, D.O. (2007) Low-pressure hydrocracking
91.
92.
93.
94.
95.
reactions (WNiPd/CeY-Al2 O3 ). Fuel Process. of coal-derived Fischer-Tropsch waxes to diesel. Technol., 89, 808–818. 96. Leckel, D.O. (2007) Selectivity effect of oxyEnergy Fuels, 21, 1425–1431. Leckel, D.O. (2009) A closer look at the intrinsic genates in hydrocracking of Fischer-Tropsch low aromaticity in diesel hydrocracked from waxes. Energy Fuels, 21, 662–667. 97. Leckel, D.O. (2009) H2 S effects in low-temperature Fischer-Tropsch wax. Energy base-metal-catalyzed hydrocracking of Fuels, 23, 32–37. Weitkamp, J. (1975) The influence of chain Fischer-Tropsch wax. Energy Fuels, 23, length in hydrocracking and hydroisomerization 2370–2375. 98. De Klerk, A. (2008) Hydroprocessing peculiariof n-alkanes. ACS Symp. Ser., 20, 1–27. Calemma, V. and Gambaro, C. (2010) Hydrocties of Fischer-Tropsch syncrude. Catal. Today, racking of Fischer-Tropsch wax: effect of feed 130, 439–445. 99. De Klerk, A. (2009) Catalysts important in the composition. Prepr. Pap.-Am. Chem. Soc., Div. refining of Fischer-Tropsch syncrude to fuels. Fuel Chem., 55 (2), 439–440. Denayer, J.F., Baron, G.V., Souverijns, W., Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 Martens, J.A., and Jacobs, P.A. (1997) Hydroc(1), 116–117. racking of n-alkane mixtures on Pt/H-Y zeolite: 100. Gambaro, C. and Calemma, V. (2010) Hydrocracking different Fischer-Tropsch waxes: model Chain length dependence of the adsorption and simulations. Prepr. Pap.-Am. Chem. Soc., Div. the kinetic constants. Ind. Eng. Chem. Res., 36, Fuel Chem., 55 (2), 378–379. 3242–3247. Tailleur, R.G. and Nascar, J.R. (2008) The effect of aromatics on paraffin mild hydrocracking
441
22 Reforming and Aromatization 22.1 Introduction
Aromatics are an important compound class in petrochemical refining and aromatics are an essential part of the main transportation fuels, motor-gasoline, jet fuel, and diesel fuel. Although many crude oils do not contain a high concentration of straight-run aromatics in the naphtha range, naphtha range aromatic compounds are generated during crude oil refining. Heavier crude oil fractions typically contain percentage levels of aromatic compounds. The Fischer–Tropsch refiner must also produce fuels with sufficient aromatics to ensure elastomer compatibility, even when a fuel blend can be synthesized that meets the general fuel specifications without aromatics. Elastomer compatibility problems arise when fuels with little or no aromatics are interchangeably used with typical crude-oil-derived fuels (Sections 14.3.8 and 15.3.9). High-temperature Fischer–Tropsch (HTFT) syncrude contains percentage level straight naphtha range aromatics, but low-temperature Fischer–Tropsch (LTFT) syncrudes contain very little straight-run aromatics, if at all (Table 22.1) [1]. The aromatic content of HTFT naphtha is comparable to that of low aromatic crude oils, but that of LTFT syncrude is well below the concentration range encountered in most crude oils. From a refining perspective, it is not the straight-run aromatic content that differentiates syncrude from crude oil, but the low concentration of cycloalkanes. The cycloalkanes are excellent aromatic precursors and can be easily converted into aromatics by conventional catalytic naphtha reforming. The naphtha reforming unit has been a key refinery unit from second-generation crude oil refineries onwards (Chapter 2). In crude oil refineries, catalytic naphtha reforming became not only a useful source of aromatics for motor-gasoline production but also the primary source of refinery hydrogen. In a Fischer–Tropsch refinery, the emphasis on H2 is less, but the applications requiring aromatics production (Section 16.5.3) are very important. The lack of cyclic material in Fischer–Tropsch syncrudes has a significant impact on the efficiency of conventional catalytic reforming (Section 22.3.1). The performance of conventional catalytic naphtha reforming technology with HTFT naphtha is very poor (Section 9.4.2), and the performance with LTFT naphtha is even worse. A different approach is clearly called for in the refining of syncrude. Rather than focusing on the shortcomings of syncrude, more efficient refining pathways can be suggested that exploit the properties of syncrude within the context of a Fischer–Tropsch refinery [3]: Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
442
22 Reforming and Aromatization Table 22.1 Analyses of hydrogenated naphtha fractions that were obtained from industrial fixed bed Fe-LTFT and circulating fluidized bed Fe-HTFT operation at the Sasol 1 facility and laboratory Co-LTFT synthesis at 0.1 MPa using a Co–ThO2 –kieselguhr Fischer–Tropsch catalyst.
Carbon number
C5
Product
Fe-LTFT 220 ◦ C
Fe-LTFT Industrial average
Fe-LTFT 241 ◦ C
n-Pentane
94.9
93.1
91.2
80.5
87.8
2-Methylbutane
5.1
6.7
8.8
18.8
12.2
Cyclopentane C6
C7
C8
C9
C10
a Original b These
Hydrogenated composition per carbon number (%)a Fe-HTFT 320 ◦ C
Co-LTFT 190 ◦ C
–
0.1
–
0.7
–
n-Hexane
94.0
90.5
88.3
70.7
80.6
Methylpentanes
5.4
8.6
10.3
24.7
19.3
2,3-Dimethylbutane
0.3
0.3
0.4
0.8
0.1
Methylcyclopentane + benzene
0.3
0.6
1.0
3.6
–
n-Heptane
94.1
90.6
86.3
58.7
73.6
Methylhexanes
4.6
7.2
10.1
27.8
25.6
3-Ethylpentane
0.2
0.4
0.5
0.8
0.4
Dimethylpentanes
0.4
0.7
0.9
2.2
0.4
C7 -Cycloalkanes
0.6
1.0
1.9
7.0
–
Toluene
<0.1
0.1
0.3
3.5
–
n-Octane
93.8
90.1
84.7
53.6
67.9
Methylheptanes
5.0
7.2
10.6
27.9
29.2
3-Ethylhexane
0.3
0.6
0.6
1.5
1.7 1.2
Di/tri-branched C8 -aliphatics
0.4
0.9
1.3
3.9
C8 -Cycloalkanes
0.4
0.9
1.8
7.9
–
C8 -Aromatics
0.2
0.3
1.0
5.2
–
n-Nonane
94.2
90.9
84.3
50.3
63.3
Methyloctanes
4.6
6.2
10.7
29.3
33.0 3.7
Other branched C9 -aliphatics
0.8
1.4
3.2
9.1
C9 -Cycloalkanes
0.3
1.3
1.4
6.8
–
C9 -Aromatics
0.1
0.2
0.4
4.5
–
n-Decane
93.8
92.9
83.8
50.0
60.4
Methylnonanes
4.9
5.3
10.8
30.3
33.6
Other branched C10 -aliphatics
1.1
1.3
3.5
10.1
6.0
C10 -Cycloalkanes + aromatics
<0.1
0.1
–
3.2
–
Unidentified C10 compoundsb
0.2
0.4
1.9
6.4
–
data from Ref. [2]. compounds are likely to be mainly cycloalkanes and aromatics.
22.2 Thermal Naphtha Reforming
1) Catalytic reforming over a nonacidic Pt/L-zeolite catalyst is very selective for the conversion of n-alkanes, specifically to aromatics. The n-alkane content of syncrude is high (Table 22.1) and this type of reforming is well suited for converting Fischer–Tropsch naphtha into aromatics. One of the main drawbacks of the nonacidic Pt/L-zeolite catalyst is its extreme sensitivity to sulfur poisoning [4]. Sulfur must typically be removed to levels of less than 10 ng·g−1 (<10 ppb S). Again, syncrude has an advantage since it is sulfur free and it is not necessary to include elaborate sulfur removal equipment as feed pretreatment. 2) Light hydrocarbons are inevitably produced during Fischer–Tropsch synthesis (Table 1.2). Additional light hydrocarbons may be produced during refining by conversion processes such as cracking (Chapter 21). The alkenes have many refining pathways, but the same is not true of the alkanes, and often a substantial portion of the C3 –C4 alkanes end up as liquid petroleum gas (LPG). One of the attractive refining options for C3 –C4 alkanes is aromatization [5]. Although it is a less efficient technology than naphtha reforming, it is capable of converting normally gaseous alkanes into a liquid aromatic product.
22.2 Thermal Naphtha Reforming
Naphtha reforming is the mainstay of motor-gasoline refining. In early refineries, naphtha reforming started out as a thermal process. Thermal reforming was conducted at close to 540 ◦ C and 6 MPa (usually in the range 3.5–7.0 MPa) [6, 7]. The purpose of the process was to increase the octane number of the heavier naphtha and the conversion philosophy was simple. In a typical straight-run crude oil naphtha, there are some aromatics, but the octane number is low on account of the high concentration of very low octane number alkanes. The octane number of the alkanes drops precipitously with increase in carbon number (Table 13.4) unless the degree of branching is very high. Since the alkanes and cycloalkanes in straight-run crude oil are not very branched, their octane numbers are low. The octane number can be increased by removing the low octane number compounds from the heavy naphtha. The thermal cracking rate of n-alkanes increases with carbon number (Section 21.2.1). By exposing the heavy naphtha to thermal cracking conditions, the heavier paraffinic naphtha is thermally cracked to produce lighter alkanes and alkenes, without affecting the aromatics. The lowest octane number compounds are thus converted into lighter, higher octane number compounds. By doing so, the aromatics concentration in the heavy naphtha boiling range (75–200 ◦ C) is increased and so is the octane number. Thermal reforming is essentially a thermal cracking process, and the temperature of operation is not severe enough to lead to substantial aromatics synthesis. Thermal reforming processes therefore had two important drawbacks. Firstly, by thermally cracking the heavier aliphatic molecules, the liquid yield was substantially reduced. Secondly, little additional aromatics were produced during the thermal reforming process and as a consequence the octane number could not be increased much beyond an RON 85 [7]. The net effect of thermal reforming was to increase the motor-gasoline octane number, but it also increased its volatility and alkene content, while decreasing the overall motor-gasoline yield. Typically, a gasoline yield of around 70% at RON 80 could be obtained with thermal reforming [8].
443
444
22 Reforming and Aromatization
Thermal reforming technology was rapidly replaced by catalytic naphtha reforming after its introduction (Table 2.4) and it is presently of historical interest only. It is interesting that this thermal reforming type of thinking was revived in 2003 by Sasol for the upgrading of HTFT naphtha (Section 9.5.5). The philosophy behind the refinery modification was the same as that of thermal reforming, namely to convert the lower octane number compounds in order to increase the octane number of the remainder. A better quality motor-gasoline could be produced by thermal cracking in the high-temperature fluid catalytic cracking (FCC) unit because some aromatic synthesis took place, but the refractory nature of the C6 –C7 alkanes in the feed limited the overall octane number gain.
22.3 Conventional Catalytic Naphtha Reforming
The incentive for a catalytic reforming process was derived from the shortcomings of the thermal reforming process. Specifically, there was a need to increase the amount of aromatics. The first commercial catalytic naphtha reforming process was commissioned in 1949, the UOP Platforming process [7]. The term platforming is often colloquially used to refer to halogenated Pt/Al2 O3 -based conventional catalytic naphtha reforming. In parallel with the halogenated Pt/Al2 O3 -based catalytic reforming technology, a technology was developed that made use of MoO3 /Al2 O3 in a FCC-type unit. This technology, called fluid hydroforming, overcame the coking problems encountered during low-pressure catalytic reforming by employing continuous catalyst regeneration [8]. Eventually, the ‘‘platforming’’ type of catalytic reforming was adopted as the industry standard. The objectives of conventional catalytic naphtha reforming are to convert low octane number heavy naphtha with the least amount of yield loss into a high octane number motor-gasoline, while coproducing H2 for refinery use. Each one of these objectives is important. Yield is important from an economic perspective, the high octane number is important to provide a high-quality blending material for motor-gasoline, and the H2 yield is important as primary source of H2 for hydroprocessing units. Feedwise, the main drawback of conventional catalytic naphtha reforming is that it is inefficient at converting light naphtha into aromatics. In general, the rate of conversion increases with carbon number, with C6 < C7 < C8 ≈ C9 ≈ C10 . Fischer–Tropsch syncrude performs poorly with conventional catalytic naphtha reforming technology. There is a fundamental mismatch between the feed properties and the nature of the conversion process. The reasons for the inefficient conversion of syncrude will become apparent during the discussion of the reaction chemistry. 22.3.1 Reaction Chemistry
In order to produce aromatic compounds as primary products, the conversion must at least involve dehydrogenation. Dehydrogenation is equilibrium limited, it is endothermic, and it is accompanied by an increase in the number of moles in the product. In order to favor the dehydrogenation, equilibrium reforming has to be conducted at high temperature and the lowest
22.3 Conventional Catalytic Naphtha Reforming R
− H2
R
+ H2
R
− H2
R
+ H2
− H2 + H2 R + H2
R
R
− H2
R
+ H2
Figure 22.1 Reaction network showing the main reactions taking place during conventional catalytic reforming over a halogenated Pt/Al2 O3 -based catalyst. Hydrogen production and consumption are indicated, but not stoichiometrically balanced.
practical pressure. The most basic catalytic reforming reaction to produce aromatics is the direct dehydrogenation of a cyclohexane ring (Equations 22.1 and 22.2). C6 H14 (cyclohexane) C6 H6 (benzene) + 4H2 , H298K = +206 kJ·mol−1
(22.1)
C7 H16 (methylcyclohexane) C7 H8 (toluene) + 4H2 , H298K = +205 kJ·mol−1 (22.2) Direct dehydrogenation of cyclohexane isomers is a metal-catalyzed reaction and it is the fastest of the reforming reactions. However, catalytic naphtha reforming would be only of limited value if this had been the only reaction that could be performed. Conventional catalytic naphtha reforming is a bifunctional process that involves acid and metal catalysis (Figure 22.1). A number of reactions are possible [7–13]: 1) Direct dehydrogenation of cycloalkanes. It is metal catalyzed and an H2 producing reaction. 2) Hydrodecyclization of alkanes. It is metal catalyzed and an H2 producing reaction. The efficiency of this reaction increases with increasing chain length from C6 to C9 . This is related to the position of alkane adsorption to enable ring closure to a cyclopentane or cyclohexane isomer. For hydrodecyclization to take place, the alkane must be adsorbed onto a carbon in position j that allows j, ( j + 4) or j, ( j + 5) ring closure. For example, a ring close is not possible for hexane if it is adsorbed on C3 or C4 , because only one to five, one to six, or two to six ring closure is possible. 3) Hydroisomerization of alkanes and cycloalkanes. It is a bifunctional process and the reaction chemistry has been discussed in Section 18.4. Of specific importance in catalytic reforming is the interconversion of cyclopentane and cyclohexane isomers. The equilibrium favors methylcyclopentane over cyclohexane in a ratio around 9 : 1 at reforming conditions [8]. 4) Hydrocracking of alkanes and cycloalkanes. It is a bifunctional process and the reaction chemistry has been discussed in Section 21.2.3. It is an H2 consuming reaction. 5) Hydrogenolysis. This is mainly a metal-catalyzed side reaction which is responsible for the production of CH4 and C2 H6 , although Sterba and Haensel noted that it is not unlikely that free radical–type reactions may have a contribution [7]. It is an H2 consuming reaction. 6) Dimerization and cracking. These are acid-catalyzed reactions. The short-chain alkenes present in the reaction mixture may dimerize (Section 19.2) and crack (Section 21.2.2) to change the carbon number distribution of the product.
445
446
22 Reforming and Aromatization
7) Alkylation and disproportionation. These are acid-catalyzed reactions and affect the length and degree of branching on aromatics (Section 20.2). The reaction network (Figure 22.1) shows many of the competing pathways. The most significant trade-off is between dehydrocyclization that leads to aromatics formation and hydrocracking that leads to the formation of lighter products. At low H2 partial pressures aromatics formation is promoted, whereas at higher H2 partial pressures hydrocracking becomes more significant. Hydrocracking is not promoted by an increase in H2 pressure. On the contrary, it is just that in competition with hydrodecyclization it becomes kinetically more favorable. Hydrocracking is a net consumer of H2 and hydrodecyclization is a net producer of H2 . The H2 yield is consequently very dependent on the ratio between hydrodecyclization and hydrocracking. Hydrocracking leads to an increase in octane number by removing low octane number alkanes from the gasoline, but it is not a desirable reaction. Hydrocracking is not desirable because it decreases the liquid yield and the H2 production. It is a reaction that must be tolerated, because hydroisomerization is desirable. As hydroisomerization increases, so does the probability of hydrocracking. The various reactions that contribute to octane number improvement, liquid yield loss, and H2 yield can be illustrated by Figure 22.2. From this representation, it is clear that the feed has a dramatic impact on the performance of a conventional catalytic naphtha reformer. The easiest octane number gain for the least liquid yield loss and maximum H2 gain occurs by dehydrogenation of cycloalkanes. After that, hydroisomerization of the alkanes can increase the octane number with almost no liquid yield loss. Thereafter, it becomes more difficult to achieve a further gain in octane number without significant liquid yield and H2 yield loss. The yield loss that is incurred for a given reformate octane rating is affected by the octane number of the hydrogenated straight-run feed and by the cycloalkane content of the feed. Specifically, the gain that can be made by just dehydrogenation is an indication of how severe the reformer operation must be to achieve the gains made possible by hydroisomerization, dehydrocyclization, and hydrocracking. This gave rise to the N + 2A metric (Equation 22.3) to establish the quality of naphtha feed materials for conventional catalytic reforming. N + 2A = (vol% cycloalkanes) + 2(vol% aromatics)
(22.3)
The aromatic (A) content is an indication of the straight-run octane number, whereas the cycloalkane (N, naphthene) content is an indication of the octane number increase that is possible by dehydrogenation only. The N + 2A of Fischer–Tropsch-derived naphtha is typically less than 30, making it a lean naphtha, which is difficult to convert into a high octane number reformate without significant yield loss. This has been found in practice also, with a RON 87 reformate obtained from HTFT naphtha having an associated liquid yield of only 84% [14]. The situation is nicely summarized by Gregor from UOP [15]: ‘‘The paraffinic nature of FT products is . . . a disadvantage for the reformer. Reformers are designed to maximize aromatics yield, and this objective becomes increasingly difficult as the paraffinicity of the naphtha increases. The lean naphtha by petroleum standards is a rich naphtha by FT standards.’’
22.3 Conventional Catalytic Naphtha Reforming
Reason for yield loss
Main reactions during naphtha reforming Hydroisomerization
100
Yield (%)
Cycloalkane to aromatic
Alkane to aromatic
Dehydrogenation Dehydrocyclization
n -Alkane to iso-alkane
H2 gain due to dehydrocyclization
Heavy naphtha to light naphtha and gas
Hydrocracking
H2 yield
H2 loss due to hydrocracking
RON Straight-run naphtha RON
Liquid yield Operating range Hydrogen yield
Figure 22.2 Relationship between octane number gain, liquid yield, and H2 yield for the main reactions taking place during conventional catalytic naphtha reforming.
22.3.2 Catalysis
Developments in conventional naphtha reforming catalysis were directed to overcome four fundamental obstacles. The first and foremost was carbon formation on the catalyst. Historically, this had been the main obstacle to the development of the technology and was one of the reasons way the FCC-type fluid hydroforming technology was developed. Aromatics production is favored by lower H2 pressure and high temperature and so is the formation of coke. There was a trade-off between catalyst lifetime and aromatics production, which was governed by the H2 partial pressure. Platinum-based catalysts were active enough to achieve reasonable catalyst life and aromatics production, but coke formation could not be eliminated. Better catalyst stability could be achieved at higher H2 partial pressures, but it also presented a problem. The second problem was hydrogenolysis. Pt catalyzes hydrogenolysis and the hydrogenolysis activity increases with increasing H2 partial pressure. The high activity of Pt had to be tempered in some way to reduce hydrogenolysis without affecting dehydrogenation and dehydrocyclization activity. The third obstacle was cost. It was important to ensure good Pt dispersion on the reforming catalyst to obtain maximum activity for the least amount of platinum. With time, during reaction and during regeneration the Pt metal sites tended to increase in size to reduce the dispersion, which undermined activity and selectivity. Ways had to be found to retain good Pt
447
448
22 Reforming and Aromatization Table 22.2
Operating characteristics of the main conventional catalytic naphtha reforming technologies.
Description
Semiregenerative
Cyclic
Continuous regeneration
Temperature (◦ C)a Pressure (MPa) Reactor configuration Cycle length (mo) Typical reformate RON
480–550 1.4–3.5b Fixed bed 6–12 85–100
480–550 1.4–2.1 Fixed bed <1 100–104
480–550 0.3–1.7 Moving bed Continuous 95–108
a Highest b
operating temperature is determined by reactor metallurgy. Operation at 2.8–3.5 MPa with Pt/Cl− /Al2 O3 ; 1.4–2.1 MPa possible with PtRe/Cl− /Al2 O3 .
dispersion and even to redisperse Pt after regeneration. Although cyclohexane dehydrogenation is not structure sensitive, coking is, and by redispersing the metal the larger flat metal surfaces required for coking are reduced while the overall surface area for dehydrogenation is increased. The fourth obstacle related to acidity. Similar to the case of the bifunctional catalysts employed for hydroisomerization and hydrocracking, a balance had to be struck between the metal site and acid site activity, as well as the acid strength of the support. Too strong an acidity increased cracking reactions and reduced the liquid yield, whereas too weak an acidity limited alkyl cyclopentane to cyclohexane isomerization. Early on, it was realized that an SiO2 –Al2 O3 support is too acidic and that an Al2 O3 -only support is too weakly acidic. An appropriate level of acidity could be obtained by halogenating Al2 O3 . Some of the first successful reforming catalysts were fluorinated platinum-promoted alumina (Pt/F− /Al2 O3 ). Currently, reforming catalysts employ chlorinated alumina as support material. The relative importance of each of these challenges is influenced by the reactor technology. Catalytic naphtha reforming technologies are classified into three types based on how catalyst regeneration is conducted: semiregenerative, cyclic (fully regenerative), and continuous regenerative (Table 22.2) [16]. Semiregenerative units must be shut down for regeneration. It is the most often used technology and the least expensive. Operation is planned for a predetermined cycle length, which is typically 6–12 months. The severity of the operation in the unit is limited to increase the cycle length. Although more severe operation is possible, in practice semiregenerative units are operated to produce reformate with RON 85–100, depending on the N + 2A of the naphtha. The reactors are fixed bed and the catalyst shape and mechanical strength can be optimized for such operations. The rate of catalyst deactivation by coking is of paramount importance in these units, because it determines the cycle length. Catalysts for semiregenerative reforming are therefore optimized to minimize coke production. A more stable catalyst that was less prone to coke formation than monometallic Pt/Cl− /Al2 O3 was developed by introducing a second metal to alloy with platinum. Bimetallic PtRe/Cl− /Al2 O3 has a lower coking rate, and the Pt:Re ratio can be optimized for the intended operating conditions. Reactor loading can also be optimized, and different catalysts can be used in different beds. Being a fixed bed technology, the first catalyst bed must bear the brunt of catalyst deactivation due to feed impurities. The better sulfur tolerance of monometallic Pt/Cl− /Al2 O3 may be put to good use in the first catalyst bed.
22.3 Conventional Catalytic Naphtha Reforming
Cyclic units differ from semiregenerative units in one respect: that is, the availability of a spare fixed bed reactor to allow one reactor to be taken off-line for regeneration without shutting the unit down. This enables cyclic units to be operated at lower H2 partial pressure and under more severe conditions, because cycle lengths of as little as one week can be managed. Reformate of up to RON 100–104 is achievable, and a higher liquid yield is possible due to lower pressure operation. A variety of catalysts have been used with cyclic units, including Pt, PtRe, and PtSn/Cl− /Al2 O3 , as well as trimetallic catalysts on halogenated alumina. Continuous regenerative units make use of moving bed reactor technology. The catalyst flows under gravity from one bed to another until it is withdrawn from the last bed to be regenerated. The catalyst is continuously withdrawn, regenerated, and returned to the top of the first catalyst bed in the naphtha reformer. Higher severity and lower pressures are possible, and it is possible to produce reformate of RON 95–108. The rapid coke deposition on the catalyst can be tolerated because the catalyst is continuously being regenerated and returned to the reactor. Application of a catalyst in a moving bed reactor imposes restrictions on catalyst morphology (it has to be spherical) and requires the catalyst to be more attrition resistant. The more severe operation requires the catalyst to be more resistant to metal sintering and agglomeration. Redispersion of the metal in the regenerator is possible though. Most catalysts for continuous regenerative units are based on PtSn/Cl− /Al2 O3 . These catalysts are further modified by reducing the Pt content and acidity of the support in response to the much lower operating pressures that are possible. 22.3.3 Syncrude Processing Technology
Some important catalyst and technology characteristics that have an impact on the application of conventional catalytic naphtha reforming with Fischer–Tropsch feed are the following: 1) In semiregenerative and cyclic reformers, the hydrogenolysis activity of freshly reduced reforming catalysts can be suppressed by partial sulfiding the metal (0.05–0.06% S) [17]. During normal operation, sulfur is a strong inhibitor. Pt/Cl− /Al2 O3 catalysts can tolerate parts per million S in the feed, whereas PtRe/Cl− /Al2 O3 catalysts are more sensitive and feed sulfur has to be around 0.25–1 µg·g−1 or lower. When applying semiregenerative or cyclic reforming with Fischer–Tropsch syncrude, naphtha hydrotreatment with a sulfided catalyst is actually beneficial, because it introduces low levels of S (<1 µg·g−1 ) into the reformer feed, making desulfurization of the catalyst more gradual. This reduces the risk of increased hydrogenolysis activity early in a reactive cycle. This is not applicable to continuous regenerative reformers. 2) The alumina support material derives its acidity from chlorination. Constant chlorine addition is required, typically by co-feeding CCl4 or Cl2 C=CCl2 [12]. The compounds are reduced to produce HCl, which chlorinates the alumina. This makes the catalyst sensitive to water, oxygenates, and nitrogen bases. The latter is not present in syncrude, but both dissolved water and oxygenates are ubiquitous in Fischer–Tropsch naphtha. Care must be taken in the feed pretreatment to limit water and oxygenates in the reformer feed. An equivalent water content of 10–20 µg·g−1 is apparently beneficial to develop acidity on the alumina, but a higher water content leaches Cl from the catalyst [13]. 3) Industrial applications of conventional catalytic reforming technology with Fischer–Tropsch syncrude employs a naphtha hydrotreater as feed pretreatment step (Chapters 9 and 10).
449
450
22 Reforming and Aromatization
n -1-Alkenes
Syncrude
Extraction of n -1-alkenes
Naphtha reformer
Deoxygenation
Hydrotreater
H2 + gas Reformate
Oxygenates Figure 22.3 Integration of n-1-alkene recovery from Fischer–Tropsch syncrude with a catalytic naphtha reformer to exploit the benefit of additional heat release by the oxygenate-free unhydrogenated syncrude.
In the naphtha hydrotreater, the high heat release is a significant design consideration [18]; yet, in the reformer the conversion is highly endothermic and each catalyst bed requires an interstage heater. In continuous regenerative reformers, where a slight increase in catalyst coking can be tolerated, co-feeding deoxygenated but unhydrogenated syncrude could have a significant advantage. It would not only reduce the heating cost but also increase conversion due to the higher average bed temperature that can be maintained. This type of application is especially well suited for implementation in conjunction with n-1-alkene recovery (Figure 22.3), where oxygenate separation is performed as part of the chemical extraction technology (Chapter 23). 4) The mechanism of reforming over conventional Pt-based chlorinated alumina catalysts requires a rich naphtha with high N + 2A ratio to obtain a product with high RON without too much liquid yield loss. Fischer–Tropsch syncrude makes a very poor feed and, although it can be used with syncrude, it is not the reforming technology of choice [3].
22.4 Monofunctional Nonacidic Pt/L-Zeolite Naphtha Reforming
Conventional catalytic naphtha reforming is hugely successful and has been almost universally adopted in crude oil refineries. Despite the industrial success of the technology, there were some shortcomings that led to the investigation of alternative reforming catalysts not based on Pt/Cl− /Al2 O3 . Primarily, there was a need to find a catalyst that could dehydrocyclize acyclic alkanes to aromatics, which are the slowest to react over conventional naphtha reforming catalysts. The landmark work by Bernard on alkali-exchanged Pt/L-zeolites [19] led to the development of a new reforming technology that was based on a monofunctional nonacidic Pt/L-zeolite catalyst. The Aromax technology from Chevron Phillips Chemical company [20] and RZ-Platforming technology from UOP [21], both employ monofunctional reforming catalysts. Compared to conventional bifunctional reforming (Section 22.3), monofunctional reforming demonstrated very high selectivities for aromatization and performed exceptionally well with n-alkanes, including the C6 –C7 n-alkanes (Figure 22.4) [22]. This was a considerable bonus, since conventional catalytic naphtha reforming is poor at converting light naphtha.
22.4 Monofunctional Nonacidic Pt/L-Zeolite Naphtha Reforming
Relative aromatization rate (%)
100
80
60
40
20
PtBaK/L-zeolite PtRe(S)/Cl/alumina
0 5
6
7
8
9
10
Carbon number of n -alkane feed Figure 22.4 Comparison of the relative aromatics production rates of n-alkanes over nonacidic PtBaK/L-zeolite and PtRe(S)/Cl− /Al2 O3 reforming catalysts at 493 ◦ C, 0.8 MPa, and 18 h−1 LHSV.
The advantage of the monofunctional catalyst for the reforming of Fischer–Tropsch syncrude with its significant linear hydrocarbon content (Table 22.1) is apparent. Hydrotreated HTFT feed material was successfully piloted with the Aromax catalyst and produced higher than predicted H2 and aromatics yields [3]. In fact, the use of monofunctional reforming technology to refine Fischer–Tropsch products has been advocated for quite some time [3, 5, 23, 24]. 22.4.1 Reaction Chemistry
The reaction chemistry over monofunctional Pt only, simplifies the reaction network of naphtha reforming considerably. By eliminating the acid-catalyzed pathways, the bifunctional pathways are also eliminated. Hydroisomerization, hydrocracking, dimerization, and cracking, as well as alkylation and disproportionation, are all dependent on catalyst acidity. Although these reactions no longer take place over monofunctional nonacidic platinum catalysts, some metal-catalyzed reactions that were of minor importance over bifunctional reforming catalysts now become much more prominent [25]: 1) Dehydrogenation. Pt is a very good dehydrogenation–hydrogenation catalyst. An equilibrium between dehydrogenated and hydrogenated species is established, and of specific importance is the dehydrogenation of cyclohexane-based species to their equivalent aromatics. This is the main aromatization reaction over monofunctional reforming catalysts and is the same as what has been encountered over bifunctional reforming catalysts. 2) Dehydrocyclization. The ability of Pt to cyclize alkanes by dehydrogenation is central to its ability to convert acyclic compounds into aromatics. It is the same process as was discussed for bifunctional catalysts (Section 22.3.1), which involves j, (j + 4) or j, (j + 5)
451
452
22 Reforming and Aromatization
− H2 R
Figure 22.5
Pt
R + H2 Pt
H2
R R
Metal-catalyzed isomerization by dehydrocyclization and hydrogenolysis.
ring closure to yield cyclopentanes and cyclohexanes. Once a cyclohexane is formed, further dehydrogenation to the aromatic is possible. 3) Hydrogenolysis. The cleavage of a C–C bond through the action of H2 can produce lighter products, including C1 –C2 alkanes. In the case of cyclic aliphatics, the cleavage of a C–C bond on the ring results in ring opening, rather than the hydrolytic cleavage of the molecule into smaller fragments. This is an important process, because it is the pathway by which cyclopentanes can be ring-opened. Cyclopentanes cannot be aromatized, and, by ring opening, these molecules are again available for potential j, (j + 5) ring closure to yield cyclohexanes that can be aromatized. 4) Isomerization. Ring closure by dehydrocyclization followed by ring opening by hydrogenolysis can lead to skeletal isomerization (Figure 22.5). This may increase or decrease the degree of branching depending on the intermediate and the position of hydrogenolysis. 22.4.2 Catalysis
Metal-catalyzed ring closure of acyclic alkanes over Pt, without the contribution of support acidity, was reported by a number of researchers [26, 27]. Thus germinated the idea to develop a reforming catalyst without acidity and consequently without the yield loss associated with hydrocracking. Coke formation was a problem on nonacidic Pt/Al2 O3 and Pt/SiO2 catalysts, and catalyst activity was quickly lost at low pressure. This could be corrected by alloying the Pt with a second metal, and, in laboratory tests, nonacidic PtSnK/Al2 O3 showed some promise [28]. Another approach to reduce coking rate was to impose geometric constraints on the metal sites, so that it would be difficult to form heavy carbonaceous deposits and coke. It is this approach that led to the work of Bernard on alkali-exchanged Pt/L-zeolites [19]. The remarkable selectivity for n-hexane conversion to benzene (Figure 22.4), which was higher than the aromatization rate of the heavier n-alkanes, is not what one would expect on statistical grounds. The 1,6-ring closure occurred at a high rate. A number of explanations have been forwarded [25]: 1) The ‘‘confinement’’ model explained the selectivity in terms of the narrow pore structure of L-zeolite (LTL), which orientates the n-hexane in such a way that 1,6-adsorption and ring closure is favored [29]. It is specifically the high selectivity of Pt/L-zeolite for 1,6-ring closure over 1,5-ring closure that makes it such a successful catalyst compared to alternative support materials [30]. This requires the Pt to be in-channel and not out-of-channel in order to achieve high selectivity [31]. The confinement effect is not limited to L-zeolite, and the selectivity is related to the orientation of the molecule as it approaches the Pt rather than the zeolite morphology per se. It has been found that both selectivity and deactivation rate could be related to the ratio of in-channel and out-of-channel Pt in L-zeolite catalysts [32].
22.4 Monofunctional Nonacidic Pt/L-Zeolite Naphtha Reforming
2)
The selectivity was also related to the Pt cluster size [33]. It was argued that the L-zeolite achieves its high aromatization selectivity by stabilizing very small Pt clusters, rather than just by the effect of confinement and molecular orientation. In an unconstrained environment, the relative rate of dehydrocyclization to hydrogenolysis increased with increasing metal dispersion (decrease in metal loading) [34, 35]. 3) Metal–support interaction to increase the electron density of Pt has been noted in combination with the stabilization of small Pt clusters to explain the high aromatization selectivity of n-hexane [36]. Fischer–Tropsch-derived material is sulfur free, but unless it is deeply hydrotreated, it contains oxygenates and alkenes. The effect of alcohols, ketones, aldehydes, and CO on PtK/L-zeolite conversion of n-hexane was reported to be the same [37]. The oxygenates were introduced at percentage levels in the feed and suppressed n-hexane conversion, decreased selectivity to benzene and C1 –C5 products, and increased selectivity to n-hexenes. Co-feeding water had no effect. A near-linear relationship was found between n-hexane conversion and benzene selectivity, irrespective of whether the changes were induced by oxygenates, coking, or operating conditions. No increase in deactivation was observed with oxygenate-containing feed. Feeding an alkene, instead of alkane, made no difference to the behavior of PtK/L-zeolite [37]. 22.4.3 Syncrude Processing Technology
Fischer–Tropsch syncrude is closer to an ideal feed material for nonacidic Pt/L-zeolite reforming than crude-oil-derived feed on account of being sulfur free. The high n-alkane content is also a benefit, but it does not necessarily contain more n-alkanes than a crude-oil-derived light straight-run naphtha. Some specific observations for the application of monofunctional reforming with syncrude are given below: 1) When a naphtha hydrotreater is employed as feed pretreatment for monofunctional nonacidic Pt/L-zeolite-based catalytic reforming, it is preferable to employ an unsulfided hydrotreating catalyst. The nonacidic Pt/L-zeolite catalyst is extremely sensitive to sulfur [4], and the advantage of syncrude as feed is eroded by purposefully introducing sulfur during the feed pretreatment step. 2) The level of hydrotreating that is required for feed pretreatment is less stringent than for conventional catalytic naphtha reforming of syncrude, because the nonacidic Pt/L-zeolite is only inhibited by oxygenates, and water reportedly has no effect [37]. Nevertheless, it is still recommended to reduce the level of water and oxygenates in the feed. 3) The same process intensification as suggested for conventional catalytic naphtha reforming (Figure 22.3) can be applied to monofunctional catalytic reforming to exploit the heat of hydrogenation of alkenes. In fact, the extent of deoxygenation that is required is less, and alumina-catalyzed deoxygenation may be sufficient feed pretreatment. 4) The aromatic compounds produced in monofunctional reforming are by direct dehydrocyclation, and Cn alkanes produce Cn aromatics. By controlling the carbon number distribution of the naphtha feed, the carbon number distribution of the aromatics is controlled. Some hydrogenolysis takes place, and the selectivity of lighter aromatics is somewhat higher than
453
454
22 Reforming and Aromatization
5)
6)
7)
8) 9)
the feed carbon number distribution. This implies that benzene will be coproduced even if C6 naphtha is excluded from the feed. Whenever monofunctional reforming is employed, it must be in conjunction with a benzene refining strategy. Excluding C6 material from the feed will not avoid benzene coproduction, and it is actually advantageous to employ C6 -containing naphtha as reformer feed. The highest aromatization selectivity is achieved with n-hexane, which forms benzene in high yield (>80%). There is an important difference though, because the benzene is available as a benzene-rich product and not as diluted benzene. Chemical extraction of the benzene may be profitable, and aromatic alkylation (Chapter 20) has an excellent technology fit on account of the alkene-rich nature of syncrude. The high benzene availability from C6 reforming in a Fischer–Tropsch refinery is an advantage and not a refining problem. The higher cost of monofunctional catalytic naphtha reforming restricted industrial application to chemical production. Industrially, the feed preference is for C6 –C8 naphtha, because it has the most advantage over conventional naphtha reforming for aromatics production (Figure 22.4). This does not imply that the technology cannot employ heavier feed, but in a crude oil refining context it would not make sense. In a Fischer–Tropsch refinery, it is the preferred naphtha reforming technology and it is more efficient for heavier syncrude naphtha than conventional naphtha reforming. The feed range is determined by the refining requirements and may include C9 –C10 naphtha. A Fischer–Tropsch refinery with a monofunctional nonacidic Pt/L-zeolite-based naphtha reformer is likely to export H2 to the Fischer–Tropsch gas loop. In this way, the refinery can increase the syncrude production by Fischer–Tropsch synthesis. Monofunctional naphtha reforming is the most efficient refining pathway for the difficult-to-refine C7 naphtha fraction [5]. When a wide-cut naphtha is employed as feed, it enables the refinery to develop a balanced petrochemical portfolio that includes aromatic chemicals. This is normally a shortcoming in some suggested petrochemical applications of Fischer–Tropsch synthesis [38]. It is also a useful technology alternative to thermal cracking (Chapter 21) for value addition to LTFT naphtha.
22.5 Aromatization
The synthesis of ZSM-5 in the 1970s enabled a number of applications. Among others, it led to the development of the M-forming processes [22, 39, 40]. M-forming was based on the same premise as thermal reforming, but with the difference that H-ZSM-5 was able to catalyze aromatization. It was found that the relative cracking rates of light alkanes (C5 –C7 ) over H-ZSM-5 were inversely proportional to their octane numbers. Low octane number alkanes such as n-heptane and n-hexane were more easily cracked than the higher octane branched alkanes. This was very beneficial, because refiners did not have a convenient refining pathway for C7 naphtha. The C7 naphtha was not easily converted by conventional catalytic naphtha reforming (Section 22.3) and could not be effectively hydroisomerized either (Section 18.4). The main drawback of H-ZSM-5 aromatization of light naphtha was the trade-off between aromatics production and liquid yield. As a light naphtha reformer application, M-forming
22.5 Aromatization Table 22.3
Aromatization of light naphtha over H-ZSM-5 in the M-forming processes.
Description
Operating conditions Temperature (◦ C) Pressure (MPa) H2 :feed molar ratio WHSVa (h−1 ) Reformate Liquid yield (mass%) 1 2 (RON + MON) Aromatics in liquid (%) Aromatics yield (%)
M-forming
M2-forming
C6 -80 ◦ C light naphtha
n-Pentane
n-Hexane
Propene
FCC C5 -110 ◦ C
Straight-run C6 -110 ◦ C
315 2.8 7 –
315 2.8 7 –
575 <0.1 0 1
538 <0.1 0 1
538 <0.1 0 68
550 <0.1
550 <0.1
1.3
1.3
78b 90 42.7 33
74.5b 93 46.8 35
– – – 45
– – – 45
– – – 58
54.3 – 99.4 54
47.3 – 92.6 43.8
a Weight b
hourly space velocity (WHSV). Liquid yield reported in volume %.
operating at 315 ◦ C, 2.8 MPa, and H2 :naphtha molar ratio of 7 : 1 achieved an overall aromatics yield around 33–35% depending on the space velocity (Table 22.3) [39, 40]. In applications where aromatics, rather than reformate, was the main objective, the naphtha could be recycled to extinction. In the M2-forming process, it was possible to obtain a 45% yield of aromatics from light naphtha n-alkanes (Table 22.3), but it required different conditions: 538–575 ◦ C and near-atmospheric pressure. The per pass yields were around 31–33%, in the same range as that from M-forming. Higher yields could be obtained with olefinic feed materials. At these more severe conditions, the H-ZSM-5 deactivated by coking within days, reaching 10 mass% coke on the catalyst after 42 h. This required cyclic operation, with oxidative regeneration following each aromatization cycle. It was reported that the catalyst remained stable over the 28 cycles that it was evaluated. The M-forming approach was evaluated for the conversion of HTFT naphtha at 375–425 ◦ C and 0.1–0.5 MPa over H-ZSM-5 [41]. Depending on the operating conditions, the liquid yield increased over time from 55–80 to 80–90%, while the aromatics content decreased from 42% (highest value) to around 6%. This is equivalent to an aromatics yield of 32% at best, which monotonically decreased with increasing time on stream. Irrespective of the feed, the maximum possible aromatics yield that could be obtained by aromatization of naphtha over H-ZSM-5 was limited, and it came at a significant liquid yield penalty. It was marginally better than thermal reforming, but the technology still employed the same basic philosophy (Section 22.2). One way of overcoming this intrinsic liquid yield penalty was not to start with naphtha as feed. LPG is not always a convenient refinery product. The C3 –C4 alkanes are very hydrogen rich and are not as valuable as transportation fuels [42]. Recovering some H2 and converting the LPG into liquid products is enticing, even more so in a petrochemical complex where dehydrogenated compounds (alkenes and aromatics) are the main products.
455
456
22 Reforming and Aromatization Light alkanes CH4 C2H6 R
Acid (protolysis)
R
R
R
+ H2
R
R Alkanes
Figure 22.6
+ H2 Metal
R
R Alkenes
− H2 Metal
R
R
− H2 R
Acid (hydrogen transfer)
Acid
R Light alkenes
− H2
R
Metal
Acid
Cycloalkanes
Aromatics
Reaction network of aromatization over metal-promoted H-ZSM-5.
The problem with LPG as feed is that it is very refractory. The conversion rate of n-alkanes over H-ZSM-5 decreases with decreasing carbon number [43]. This is understandable because H-ZSM-5 conversion of LPG effectively requires protolytic catalytic cracking (Section 21.2.2) of propane and butane to generate alkenes before any aromatization can take place. Aromatization technology therefore makes use of metal-promoted H-ZSM-5, which is a bifunctional catalyst [22]. In this respect, aromatization catalysts are analogous to conventional naphtha reforming catalysts, but in aromatization catalysts the acid functionality plays a more important and dominant role. 22.5.1 Reaction Chemistry
Aromatization employs metal- and acid-catalyzed conversions (Figure 22.6), but the reaction chemistry is predominantly that of H-ZSM-5 acid catalysis. The metals employed for aromatization are not strongly dehydrogenating or hydrogenating. Aromatization and cracking are both endothermic reactions that result in an increase in the molar volume of the product. These reactions are reversible, and the aromatization equilibrium is favored by high temperature and low pressure. Aromatization processes are usually conducted in the gas phase at temperatures in the range 450–520 ◦ C and pressure below 1 MPa. The conversion of hydrocarbons over H-ZSM-5 results in a limited ‘‘equilibrium’’ that is restricted to the equilibration of the carbon number distribution [44]. Alkene-based feed materials in the C2 –C10 range were all equilibrated to carbon number distributions that were quite similar. This dynamic equilibrium is achieved by acid-catalyzed cracking (Section 21.2.2) and oligomerization of the alkenes (Section 19.3.2), which take place in parallel. The same cracking–oligomerization principle applies to alkanes, although the rate of alkane conversion is much slower than that of the alkenes. For example, over H-ZSM-5 at 538 ◦ C, complete conversion of n-hexane required a space velocity of 1 h−1 , whereas complete propene conversion could be achieved at a space velocity of 68 h−1 (Table 22.3) [39]. The rate-limiting step in alkane conversion
22.5 Aromatization
over H-ZSM-5 is protolytic cracking. In the case of metal-promoted H-ZSM-5, the metal loading determines whether the first dehydrogenation of the alkane to alkene is rate limiting, or whether cracking is rate limiting [45]. The dominance of the acid catalysis over aromatization catalysts effectively eliminates the feed-specific restrictions and advantages observed during catalytic naphtha reforming. Cyclic hydrocarbons and heavier naphtha range aliphatics are rapidly ‘‘equilibrated,’’ and aromatization takes place from this equilibrated pool of shorter chain alkenes (Figure 22.6). The many acidand metal-catalyzed reactions that can take place have been discussed previously, and what is of specific interest is the aromatization reaction chemistry: 1) Aromatization of n-hexane and 1-hexene over metal-promoted H-ZSM-5 does not take place by direct dehydrocyclization. Both compounds are cracked to lighter alkenes, which are oligomerized to produce, among others, cyclic products. The cycloalkanes produced by oligomerization of alkenes are the precursors for aromatics [45–47]. The mechanism of the actual aromatization reaction by dehydrogenation is similar to that of conventional catalytic reforming (Figure 22.1). 2) When cyclohexane and cyclohexene are used as feed, only some of the cyclic C6 material is directly converted into aromatics to produce benzene [48]. The amount of benzene increased as the metal loading of the H-ZSM-5 catalyst was increased, although the total yield of aromatics decreased somewhat. The remainder of the aromatics was produced through cracking, oligomerization, and cyclization of heavier than C6 alkenes. The rate of cracking and oligomerization is much higher than that of dehydrogenation [49]. Much of the cyclic material succumbs to cracking before dehydrogenation can take place, which explains why heavier aromatics are formed preferentially, even when using cyclic C6 feed material. 3) Aromatization over H-ZSM-5 also requires the formation of a cyclic intermediate, but instead of metal-cyclized dehydrogenation, aromatization of the cyclic intermediate takes place by hydrogen transfer to light alkenes [47]. Again the role of a cyclic intermediate is important, but even when cyclic intermediates are in the feed, much of the cyclic material is destroyed during the rapid equilibration that takes place over H-ZSM-5. 4) Cyclization of the alkenes preferentially takes place from C7 and heavier material, which does not involve an unfavorable carbocation intermediate [47]. The C7 and heavier material is more susceptible to cracking, and this effectively limits the aromatization rate, which requires cyclization by dehydrogenation. 5) Benzene is not readily produced by oligomerization, cyclization, and dehydrogenation, because C6 requires a primary carbocation intermediate for acid-catalyzed cyclization. Benzene is nevertheless produced over ZSM-5 catalysts, of which some benzene is formed by transalkylation, rather than direct aromatization. 22.5.2 Catalysis
Aromatization catalysts have been exclusively based on H-ZSM-5. One of the reasons for this is the pore-constrained geometry of H-ZSM-5, which favors the production of mononuclear aromatics. Because it is difficult to form multinuclear aromatic structures, coking is inherently
457
458
22 Reforming and Aromatization
limited, although some coking inevitably occurs. Of equal importance is the coking capacity of H-ZSM-5. In comparison with other often used zeolites, H-ZSM-5 has a much larger coke capacity, and a high concentration of coke is required before significant deactivation occurs [50]. In combination, these two effects allow longer cycle lengths for H-ZSM-5 during aromatization than other zeolites. Although H-ZSM-5 is quite capable of aromatization on its own (Table 22.3), there are disadvantages to aromatization by acid catalysis only. Acid catalysts perform dehydrogenation in the absence of a metal function by hydrogen transfer to alkenes, not by desorption of molecular H2 [51]. The maximum aromatics yield is consequently limited by the hydrogen balance, which requires hydrogen to be rejected as light hydrocarbons. By including a metal function, this limitation can be overcome. The metal function in aromatization catalysts fulfills three important functions: 1)
Alkanes are activated by dehydrogenation, which immediately creates reactive intermediates for aromatization. This increases the rate of conversion, because it reduces the reliance on protolytic cracking to initiate the reaction. 2) The metal function is responsible for desorption of hydrogen as H2 . By doing so, the aromatics yield can be increased, because hydrogen rejection no longer requires alkenes for hydrogen transfer. Furthermore, it enables the aromatization unit to become a source of H2 for the refinery. 3) Coke formation is reduced because the metal function is able to hydrogenate heavy condensed aromatic structures that form on the catalyst surface. This extends the cycle length during severe operation. The coke that ultimately forms is more hydrogen rich and less refractory, making oxidative catalyst regeneration easier. The activity of the metal determines not only the rate of dehydrogenation, but also the rate of hydrogenation [49]. Poorer aromatization performance has been reported when using Pt instead of less hydrogenating metals [22]. A very active metal increases the rate of hydrogenation, and the rates of both intermediate hydrogenation and aromatics hydrogenation are much higher. This is desirable during hydroisomerization and hydrocracking, but it is undesirable during aromatization, where it is important to keep the reaction pool unsaturated. Less hydrogenating metals are consequently employed for aromatization catalysts. The two main catalysts employed for industrial scale aromatization are Ga/H-ZSM-5 (e.g., Cyclar process) and Zn/H-ZSM-5 (e.g., Alpha process). Despite the metal promotion, the catalysts deactivate within days by coke deposition. These processes are therefore characterized by periodic operation. Each production cycle, which is on the order of two days, is followed by a regeneration cycle during which the coke on the catalyst is oxidatively removed and catalyst activity is restored. During oxidative regeneration, some water is generated which causes hydrothermal dealumination of the zeolite and results in gradual but permanent catalyst deactivation [52]. Numerous reaction–regeneration cycles are nevertheless possible. The rate of permanent activity loss from cycle to cycle is dependent on the degree of ‘‘steaming,’’ which is determined by the exposure time to water vapor and the water partial pressure during regeneration [53]. Typical production figures for the Cyclar process is provided in Table 22.4 [15, 22]. Liquid yields around 65% are typical for aromatization over metal-promoted ZSM-5, and it is related to
22.5 Aromatization Table 22.4
Aromatization of propane and butane over Ga/H-ZSM-5 in the Cyclar process.
Description
Reformate Liquid yield (mass%) RON MON Product yield (mass%) Hydrogen Methane Ethane + ethene C3 –C4 hydrocarbons Benzene Toluene C8 aromatics C9 and heavier aromatics a
Product yield (mass%) Propane
Butane
FT C3 –C4 a
63.6 111.6 100.2
67.5 111.9 100.6
65.4 – –
6.0 18.8 11.1 0.5 19.3 26.6 11.1 6.6
5.5 16.4 10.3 0.3 18.0 28.9 13.4 7.2
3.8 30.8
16.7 27.7 13.4 7.7
Fischer–Tropsch C3 –C4 fraction containing alkenes.
the limited ‘‘equilibrium’’ of the carbon number distribution in the system. The liquid yield can be increased somewhat by increasing the operating pressure, but at the expense of decreasing the aromatics yield. With limited variation, the liquid yield is independent of feed, and similar liquid yields are obtained with naphtha range feed. This implies that LPG has an advantage over naphtha as feed material, because it is accompanied by an increase in liquid yield rather than a decrease, as in the case of naphtha range feed. Alkenes present no problem for aromatization catalysts and are beneficial from a conversion rate perspective (Table 22.3). Pilot plant studies of the Cyclar process with typical alkene-containing Fischer–Tropsch HTFT and LTFT LPG fractions resulted in easier conversion [15]. Oxygenates can also be converted over H-ZSM-5 catalysts (Section 19.3.2) despite some oxygenate inhibition. On the basis of these observations, it was anticipated that oxygenate-containing straight-run Fischer–Tropsch syncrude can be aromatized over metal-promoted H-ZSM-5 catalysts. However, in practice the oxygenates undermined catalyst stability. Aromatization of HTFT naphtha over Zn/H-ZSM-5 has been investigated [54]. The feed contained around 3 mass% O as oxygenates and about 50% alkenes. An aromatics yield in the range 50–55% was obtained over a number of reaction cycles and, despite some permanent deactivation, the naphtha conversion remained above 90%. The aromatics yield then dropped precipitously and stabilized at a yield of around 35%, typical of an H-ZSM-5 catalyst. The syncrude effectively converted the catalyst from a Zn/H-ZSM-5 to an H-ZSM-5 catalyst. The alkenes in the Fischer–Tropsch feed presented no problem, but the oxygenates destroyed the metal function of the catalyst and the oxygenates increased the dealumination rate significantly. At the high operating temperatures associated with aromatization, oxygenates are clearly very destructive to the metal-promoted aromatization catalyst.
459
460
22 Reforming and Aromatization
22.5.3 Syncrude Processing Technology
The application of aromatization technology with naphtha range feed material is associated with a significant liquid yield loss. This liquid yield loss is not related to the origin of the feed but is inherent to the operating conditions of the process. This alone suggests that aromatization should not normally be considered for the upgrading of Fischer–Tropsch naphtha. Fischer–Tropsch naphtha also contains oxygenates that were found to be destructive to metal-promoted H-ZSM-5 aromatization catalysts [54]. It is possible to conceptualize a process involving deoxygenation of the syncrude naphtha before aromatization, but the added feed pretreatment cost combined with the liquid yield loss makes aromatization very unattractive for the refining of Fischer–Tropsch naphtha. Aromatization technology can be employed with LPG feed material in a Fischer–Tropsch refinery. This application is very similar to the use of aromatization with LPG in a crude oil refinery. The gaseous C3 –C4 syncrude fraction contains very little oxygenates, and refinery LPG streams are generally oxygenate free. This is the preferred application of aromatization in Fischer–Tropsch refining. The syncrude-specific comments pertain to the potential application of aromatization with C4 and lighter feed material: 1) Alkenes that are present in the feed to an aromatization unit drastically reduce the residence time required for conversion. Alkenes react much faster than alkanes (Table 22.3). A smaller aromatization reactor is required, which can be operated at a higher space velocity than an equivalent unit employing a paraffinic feed. 2) When employing a straight-run Fischer–Tropsch material, it is a worthwhile precaution to pass it over an adsorbent to remove water and oxygenates that may be present in the feed. Both processes lead to permanent catalyst deactivation, even though it may not noticeably affect the cycle length, activity, or selectivity toward aromatics production in the short term. Water in the feed increases hydrothermal dealumination during the reaction cycle. Oxygenates in the feed not only produces water, but are also responsible for destroying the metal function of the aromatization catalyst. 3) When operating with an oxygenate-free feed, the overall catalyst lifetime, or number of reaction–regeneration cycles that can be performed before catalyst replacement, is determined by dealumination during oxidative catalyst regeneration. The production of water vapor during coke combustion is unavoidable, but the water partial pressure is actually controlled by the flow rate and water content of the N2 -rich diluent gas. The N2 is used to dilute the air that is used for combustion in order to control the O2 concentration. Commercial technology makes use of an N2 recycle, because N2 is not readily available in crude oil refineries. It is impractical to remove all the water vapor from this high-temperature N2 -rich gas recycle stream. The application of aromatization within a Fischer–Tropsch refinery has an advantage, because the air separation units associated with synthesis gas production (Chapter 3) also produces N2 as by-product. It is consequently not necessary to employ an N2 recycle in a Fischer–Tropsch refinery. The N2 from air separation can be used on a once-through basis to reduce the water partial pressure during regeneration, thereby increasing the catalyst lifetime. 4) The aromatic liquid product that is produced from an LPG feed is virtually free of aliphatic compounds (Table 22.4). This makes it an excellent feed for petrochemical applications and
References
H2
Pressure swing absorption
Fuel gas
Gas Gas
Aromatization
Liquid Gas
Fischer−Tropsch synthesis and gas loop
Aromatic alkylation
Tail gas
Liquid
Aromatic liquids Syncrude liquids
Figure 22.7 Application of aromatization technology in combination with aromatic alkylation for the recovery of liquid products from the Fischer–Tropsch tail gas. Depending on syncrude recovery in the gas loop design, the tail gas feed may contain up to C4 hydrocarbons.
5)
similar benefits can be derived from the aromatic product as noted for monofunctional reforming over nonacidic Pt/L-zeolite (Section 22.4.3). An application within a Fischer–Tropsch refinery that has significant benefit is the use of aromatization as tail gas conversion technology. Depending on the nature of the gas loop, aromatization in combination with aromatic alkylation (Chapter 20) can be an efficient way to recover value from the tail gas (Figure 22.7). This proposed application is especially advantageous in facilities where there are no downstream petrochemical consumers to make use of products such as ethene, or when the gas loop design does not recover the light hydrocarbons from the tail gas. When processing C2 material, the operation must be adapted to compensate for the lower feed reactivity. Under identical operating conditions, the reported conversion rate and aromatics yield from ethane is less than that from propane and butane over both Zn/H-ZSM-5 and Ga/H-ZSM-5 catalysts [55]. This is to be expected, since ethane is more difficult to convert (it has no secondary carbons) and it is more hydrogen rich. Following on the trend for conversion by the Cyclar process (Table 22.4), a higher H2 yield and lower aromatics yield should be expected from C2 -rich feed materials. The H2 can be used in the refinery or recycled to the Fischer–Tropsch gas loop.
References 1. Anderson, R.B., K¨ olbel, H., and Ralek, M.
4. Fukunaga, T. and Ponec, V. (1995) The
(1984) The Fischer– Tropsch Synthesis, Academic Press, Orlando, FL, pp. 178–181. 2. Pichler, H., Schulz, H., and K¨ uhne, D. (1968) Brennst. Chem., 49, 344. 3. De Klerk, A. (2008) Fischer-Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279.
nature of the high sensitivity of Pt/KL catalysts to sulfur poisoning. J. Catal., 157, 550–558. 5. Dancuart, L.P., De Haan, R., and De Klerk, A. (2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532.
461
462
22 Reforming and Aromatization 6. Bell, H. S. (ed.) (1959) American Petroleum
7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Refining, 4th edn, D. Van Nostrand, Princeton, NJ, pp. 292–298. Sterba, M.J. and Haensel, V. (1976) Catalytic reforming. Ind. Eng. Chem. Prod. Res. Dev., 15 (1), 2–17. Ciapetta, F.G., Dobres, R.M., and Baker, R.W. (1958) in Catalysis, Alkylation, Isomerization, Polymerization, Cracking and Hydroreforming, vol. VI (eds P.H. Emmett), Reinhold, New York, pp. 495–692. Haensel, V. and Berger, C.V. (1958) in Advances in Petroleum Chemistry and Refining, vol. I (eds K.A. Kobe and J.J. Mc Ketta Jr.), Interscience, New York, pp. 387–427. Gates, B.C., Katzer, J.R., and Schuit, G.C.A. (1979) Chemistry of Catalytic Processes, McGraw-Hill, New York, pp. 184–324. Davis, B.H. (1999) Alkane dehydrocyclization mechanism. Catal. Today, 53, 443–516. Martino, G. (2001) in Petroleum Refining, Conversion Processes, vol. 3 (ed. P. Leprince), Editions Technip, Paris, pp. 101–168. Antos, G. J. and Aitani, A. M. (eds) (2004) Catalytic Naphtha Reforming, 2nd edn, Marcel Dekker, New York. Swart, J.S., Czajkowski, G.J., and Conser, R.E. (1981) Sasol upgrades Synfuels with refining technology. Oil Gas J., 79 (35), 62–66. Gregor, J.H. (1990) Fischer-Tropsch products as liquid fuels or chemicals. An economic evaluation. Catal. Lett., 7, 317–332. Aitani, A.M. (2004) in Catalytic Naphtha Reforming (eds G.J. Antos and A.M. Aitani), 2nd edn, Marcel Dekker, New York, pp. 477–495. Doolin, P.K., Zalewski, D.J., and Oyekan, S.O. (2004) in Catalytic Naphtha Reforming, 2nd edn (eds G.J. Antos and A.M. Aitani), Marcel Dekker, New York, pp. 433–457. De Klerk, A. (2001) Hydrotreating in a Fischer-Tropsch refinery. 2nd Sub-Saharan Africa Catalysis Symposium, November 5–7, 2001, Swakopmund, Namibia, paper 10. Bernard, J.R. (1980) Proceedings of the 5th International Zeolite Conference, Heyden, London, pp. 686–695. Tamm, P.W., Mohr, D.H., and Wilson, C.R. (1988) Octane enhancement by selective reforming of light paraffins. Stud. Surf. Sci. Catal., 38, 335–353. Swift, J.D., Moser, M.D., Russ, M.B., and Haizmann, R.S. (1995) The RZ Platforming
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
process: something new in aromatics technology. Hydrocarbon Technol. Int., (Autumn) 86–89. Hughes, T.R., Jacobson, R.L., and Tamm, P.W. (1988) Catalytic processes for octane enhancement by increasing the aromatics content of gasoline. Stud. Surf. Sci. Catal., 38, 317–333. Dry, M.E. (1996) Practical and theoretical aspects of the catalytic Fischer-Tropsch process. Appl. Catal. A, 138, 319–344. Dry, M.E. (2004) Present and future applications of the Fischer-Tropsch process. Appl. Catal. A, 276, 1–3. Pa´al, Z. (2004) in Catalytic Naphtha Reforming, 2nd edn (eds G.J. Antos and A.M. Aitani), Marcel Dekker, New York, pp. 35–74. Barron, Y., Maire, G., Muller, J.M., and Gault, F.G. (1966) The mechanisms of hydrogenolysis and isomerization of hydrocarbons on metals. II. Mechanisms of isomerisation of hexanes on platinum catalysts. J. Catal., 5, 428–446. Davis, B.H. and Venuto, P.B. (1969) Paraffin dehydrocyclization. Distribution of aromatic products obtained with ‘‘nonacidic’’ supported Pt catalysts. J. Catal., 15, 363–372. Davis, B.H. (1976) Paraffin dehydrocyclization. VII. Comparison of aromatic distribution from n-octane conversion over Pt at atmospheric and 200 psig. J. Catal., 42, 376–380. Derouane, E.G. and Vanderveken, D.J. (1988) Structural recognition and preorganization in zeolite catalysis: Direct aromatization of n-hexane on zeolite L-based catalysts. Appl. Catal., 45, L15–L22. Miller, J.T., Agrawal, N.G.B., Lane, G.S., and Modica, F.S. (1996) Effect of pore geometry on ring closure selectivities in platinum L-zeolite dehydrocyclization catalysts. J. Catal., 163, 106–116. Tauster, S.J. and Steger, J.J. (1990) Molecular die catalysis: Hexane aromatization over Pt/KL. J. Catal., 125, 387–389. Jentoft, R.E., Tsapatsis, M., Davis, M.E., and Gates, B.C. (1998) Platinum clusters supported in zeolite LTL: Influence of catalyst morphology on performance in n-hexane reforming. J. Catal., 179, 565–580. Mielczarski, W., Hong, S.B., Davis, R.J., and Davis, M.E. (1992) Aromatization of n-hexane by platinum-containing molecular sieves. II. n-Hexane reactivity. J. Catal., 134, 359–369.
References 34. Davis, B.H., Westfall, G.A., and Naylor, R.W.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
(1976) Paraffin dehydrocyclization. V. The influence of Pt loading on the aromatic selectivity. J. Catal., 42, 238–246. Santacesaria, E., Gelosa, D., Carra, S., and Adami, I. (1978) Influence of metal dispersion on n-hexane reactions over platinum-alumina catalysts. Ind. Eng. Chem. Prod. Res. Dev., 17 (1), 68–71. Menacherry, P.V. and Haller, G.L. (1998) Electronic effects and effects of particle morphology in n-hexane conversion over zeolite-supported platinum catalysts. J. Catal., 177, 175–188. Dry, M.E., Nash, R.J., and O’Connor, C.T. (1998) The effect of oxygenates on the n-hexane aromatization activity of Pt/KL. Proceedings of the 12th International Zeolite Conference, Baltimore, pp. 2557–2564. Steynberg, A.P., Nel, W.U., and Desmet, M.A. (2004) Large scale production of high value hydrocarbons using Fischer-Tropsch technology. Stud. Surf. Sci. Catal., 147, 37–42. Chen, N.Y. and Yan, T.Y. (1986) M2 Forming – A process for aromatization of light hydrocarbons. Ind. Eng. Chem. Process Des. Dev., 25, 151–155. Chen, N.Y., Garwood, W.E., and Heck, R.H. (1987) M-Forming process. Ind. Eng. Chem. Res., 26, 706–711. Madikizela-Mnquanqeni, N.N. (2009) Improving the gasoline quality of liquid products from the Fischer-Tropsch process. Catal. Today, 143, 126–131. Gary, J.H., Handwerk, G.E., and Kaiser, M.J. (2007) Petroleum Refining. Technology and Economics, 5th edn, CRC Press, Boca Raton, FL. Narbeshuber, T.F., Vinek, H., and Lercher, J.A. (1995) Monomolecular conversion of light alkanes over H-ZSM-5. J. Catal., 157, 388–395. Garwood, W.E. (1983) Conversion of C2 -C10 to higher olefins over synthetic zeolite ZSM-5. ACS Symp. Ser., 218, 383–396. Kanai, J. and Kawata, N. (1988) Aromatization of n-hexane over ZnO/H-ZSM-5 catalysts. J. Catal., 114, 284–290.
46. Csicsery, S.M. (1979) Dehydrocyclodimeriza-
47.
48.
49.
50.
51.
52.
53.
54.
55.
tion. Ind. Eng. Chem. Proc. Des. Dev., 18 (2), 191–197. Mole, T., Anderson, J.R., and Creer, G. (1985) The reaction of propane over ZSM-5-H and ZSM-5-Zn zeolite catalysts. Appl. Catal., 17, 141–154. Urda, A., Tel’biz, G., and Sandulescu, I. (2001) Reaction pathways for the aromatisation of cyclohexane and cyclohexene on Zn/H-ZSM-5 zeolites. Stud. Surf. Sci. Catal., 135, 4017–4024. Biscardi, J.A. and Iglesia, E. (1999) Reaction pathways and rate-determining steps in reactions of alkanes on H-ZSM5 and Zn/H-ZSM5 catalysts. J. Catal., 182, 117–128. Mihindou-Koumba, P.C., Cerqueira, H.S., Magnoux, P., and Guisnet, M. (2001) Methylcyclohexane transformation over HFAU, HBEA, and HMFI zeolites: II. Deactivation and coke formation. Ind. Eng. Chem. Res., 40, 1042–1051. Tsang, C.M., Dai, P.-S.E., Mertens, F.P., and Petty, R.H. (1994) Saturation of olefinic products by hydrogen transfer during propylene conversion. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 39 (3), 367–371. De Lucas, A., Canizares, P., Dur´an, A., and Carrero, A. (1997) Dealumination of HZSM-5 zeolites: Effect of steaming on acidity and aromatization activity. Appl. Catal. A, 154, 221–240. Sano, T., Suzuki, K., Shoji, H., Ikai, S., Okabe, K., Murakami, T., Shin, S., Hagiwara, H., and Takaya, H. (1987) Dealumination of ZSM-5 zeolites with water. Chem. Lett. Jpn., 1421–1424. De Klerk, A. (2003) in Catalysis in Application (eds S.D. Jackson., J.S. J. Hargreaves, and D. Lennon), Royal Society of Chemistry, Cambridge, pp. 24–31. Yakerson, V.I., Vasina, T.V., Lafer, L.I., Sytnyk, V.P., Dykh, G.L., Mokhov, A.V., Bragin, O.V., and Minachev, K.M. (1989) The properties of zinc and gallium containing pentasils – the catalysts for the aromatisation of lower alkanes. Catal. Lett., 3, 339–346.
463
465
23 Chemical Technologies 23.1 Introduction
It is not possible to comprehensively cover chemical refining technologies in a single chapter or even in a single book. One only has to look at some of the standard references on chemicals such as ethene [1], propene [2], benzene [3], and toluene and xylenes [4] to appreciate the breadth of the field. The books by Asinger on alkane and alkene conversion technologies [5, 6] are likewise founts of information on conversion processes relevant to typical Fischer–Tropsch compounds. The scope of the present discussion is very limited by comparison and focuses on n-1-alkene production and autoxidation. The extraction and synthesis of n-1-alkenes from syncrude is one of the few instances where considerable effort has been expended to develop Fischer–Tropsch specific technologies. Autoxidation is discussed because of its wide general applicability to syncrude and its potential use in alkane activation at moderate processing conditions. Industrially, the refining of Fischer–Tropsch syncrude to chemicals can be divided into two main categories: 1) Chemical extraction from Fischer–Tropsch syncrude. These are processes that were devised to extract and purify compounds already present in the syncrude. Some of these technologies are standard separations that may also be encountered in crude oil refineries and petrochemical complexes. For example, the purification of ethene and propene, the purification of ethanol from an ethanol–water mixture, and vacuum fractionation of wax are industrially practised in crude-oil-based facilities too. In all of these examples, the separations are not trivial, and the product matrix is not necessarily unique to syncrude. The extraction technologies that are discussed are those that are unique to the separation of chemicals from a typical syncrude matrix. These technologies are all aimed at n-1-alkene extraction (Sections 23.2.1 and 23.2.2). 2) Chemical synthesis from Fischer–Tropsch syncrude. Syncrude, like conventional crude oil, can form the basis for a petrochemical complex. In many instances, the technologies that are employed require purified feed to meet specific product or technology requirements. After purification, syncrude-derived feed may have different trace level compounds compared to crude oil, but it does not require a different type of petrochemical technology. Depending on the trace impurities, it may render syncrude incompatible with some of the chemical technologies, but normally some benefit can be derived from the sulfur- and nitrogen-free Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
466
23 Chemical Technologies
nature of syncrude. Short of reviewing chemical production in general, there are few syncrude-specific conversion technologies. The discussion on conversion technologies for chemical production is limited to the conversion of 1-heptene into 1-octene (Section 23.2.3), the production of alcohols from distillate-range n-1-alkenes (Section 23.2.4), and autoxidation (Section 23.3). Alkene to alcohol conversion by hydroformylation is widely practised, and in a Fischer–Tropsch context, it is differentiated by the feed purification that is required. Autoxidation has been applied with syncrude since the first industrial Fischer–Tropsch facilities were constructed. It has a much wider applicability than just for chemical production. Autoxidation is considered as a refining technology with a broad range of potential applications in Fischer–Tropsch syncrude refining. Some of the technologies that were discussed in previous chapters also have use in the production of chemicals. To name a few, the oxygenate conversions (Chapter 17), wax hydroisomerization (Section 18.4), and aromatic alkylation (Chapter 19), all double as both fuel refining and petrochemical technologies. A review of these technologies is not repeated here.
23.2 Production of n-1-Alkenes (Linear α-Olefins)
The technologies for the production of n-1-alkenes (linear α-olefins) from Fischer–Tropsch syncrude are niche applications. The global market size of all even-numbered n-1-alkenes, excluding 1-butene, is in the order of 4 million tons per year, and the main uses are comonomers for polyethylene production, hydroformylation to produce plasticizer- and detergent-range alcohols and polyalphaolefin (PAO) synthetic lubricants (Table 23.1) [7]. Most of the industrial production of n-1-alkenes is by ethene oligomerization, and it has two important consequences for the n-1-alkene market. Firstly, the market developed around even-numbered n-1-alkenes only and although odd-numbered n-1-alkenes may be useful, such compounds are not produced by ethene oligomerization. Secondly, most (not all) ethene oligomerization processes yield a Poisson-like distribution of n-1-alkenes, rather than a specific carbon number. The product values of the different fractions are therefore interdependent. Table 23.1
Industrial applications of n-1-alkenes.
Carbon number range
Applications of n-1-alkenes
C4 –C8 C6 –C8 C6 –C10 C10 –C12 C10 –C12 C10 –C16 C10 –C18 C10 –C30 C C20 –C30 C
Polymers and polyethylene comonomers Carboxylic acids and thiols Plasticizer alcohols Polyalphaolefins and lubricant additives Amine oxides and amines Detergent alcohols and nonionic surfactants Surfactants Oil field chemicals Wax replacements
23.2 Production of n-1-Alkenes (Linear α-Olefins)
The extraction and purification of n-1-alkenes from Fischer–Tropsch syncrude are different, because both odd- and even-numbered n-1-alkenes can in principle be extracted. However, attempts to market the odd-numbered 1-pentene met with limited success, because it was a single supplier chemical [8]. Some specifics relevant to the application of n-1-alkene recovery from Fischer–Tropsch syncrude are as follows: 1) All n-1-alkene extraction technology has been developed for recovery from high-temperature Fischer–Tropsch (HTFT) syncrude only. These technologies can be adapted for use with low-temperature Fischer–Tropsch (LTFT) syncrudes, but neither the carbon number distribution nor the alkene content favors n-1-alkene recovery. 2) Unless there is a dramatic increase in the global market for n-1-alkenes, it is unlikely that future Fischer–Tropsch facilities will be able to exploit the niche opportunity that existed in the Sasol Synfuels facility (Section 9.5). This does not imply that one cannot exploit the beneficial properties of the n-1-alkenes within the refinery, for example, for lubricating base oil production. It is just that there is a high risk of oversupply when the n-1-alkenes are considered final products. 3) Some of the main advantages of n-1-alkene extraction from syncrude are that the production cost is not linked to the ethene price and that single carbon number production is possible. The latter advantage is being eroded by the development of single carbon number n-1-alkene technologies from ethene [9]. 4) The complexity of extraction increases with increasing carbon number. The complexity is also dependent on the origin of the syncrude fraction, with HTFT condensate fractions being cleaner than HTFT light oil fractions (Figure 5.6). 5) There are different impurities in the n-1-alkenes purified from Fischer–Tropsch syncrude than those in the n-1-alkenes produced from ethene. This may affect some applications. In this respect, the production of n-1-alkenes from ethene has an advantage, because the ethene is more easily purified than syncrude and the ethene oligomerization products are consequently free of nonhydrocarbon contaminants. This situation is analogous to the advantage that indirect liquefaction technology has over direct liquefaction technology – it is easier to purify synthesis gas (or ethene in this case) than it is to purify a solvent-refined liquid, catalytic liquefaction liquid, or pyrolysis oil (or syncrude in this case). 6) In LTFT refineries, the production of n-1-alkenes by thermal cracking (Chapter 21) may be a contender for HTFT n-1-alkene extraction. Although thermal cracking of LTFT syncrude produces a range of products as is produced by ethene oligomerization, it has the same feed cost advantage as HTFT syncrude over ethene. 23.2.1 Extraction of 1-Pentene and 1-Hexene
The same process is employed for 1-pentene and 1-hexene extraction from Fischer–Tropsch syncrude. The commercial extraction process has been described in literature (Figure 23.1) [10]. The subsequent description will use 1-hexene extraction as an example. The first step in the process is to prepare a C6 -cut that is rich in 1-hexene (not shown in Figure 23.1). The design of this step depends on the origin of the feed in the Fischer–Tropsch gas loop (Table 23.2) [11]. When the material is derived from the HTFT condensate, the C6 -cut can be prepared by topping and tailing the feed. The condensate material is not oxygenate free,
467
468
23 Chemical Technologies Methanol recovery
Methanol
Light products
HTFT C6-cut
Alkanes 1-Hexene Cycloalkenes Oxygenates
Recycle of lean NMP
Etherification Figure 23.1
C6-narrow cut
Heavy products
Extractive distillation
Extraction of 1-hexene from Fischer–Tropsch syncrude.
Composition of typical HTFT C6 -cut material obtained from different stages of cooling in the Fischer–Tropsch gas loop.
Table 23.2
Compound
C5 and lighter 4-Methyl-1-pentene (4M1P) 3-Methyl-1-pentene (3M1P) Methylpentanes 2,3-Dimethyl-butene 4-Methyl-cis-2-pentene 2-Methyl-1-pentene (2M1P) 1-Hexene n-Hexane 2-Ethyl-1-butene (2E1B) 2- and 3-Hexene isomers Methyl-cyclopentenes 2-Methyl-2-pentene (2M2P) C7 and heavier Oxygenates
Composition of C6 -cut (mass%) HTFT condensate
HTFT light oil
1.2 9.2 10.9 4.5 0.8 0.2 4.2 55.7 6.5 0.5 2.0 0.6 0.8 2.2 0.7
0.0 3.4 3.4 5.4 0.7 0.0 5.3 68.7 2.0 0.7 0.8 0.6 0.1 0.1 8.8
but it contains much less oxygenates than the material in the HTFT light oil. When 1-hexene is extracted from light oil, which contains percentage level oxygenates, it is necessary to first remove the oxygenates in the C6 -cut. The next step is an etherification process, which benefits from an oxygenate-free feed. The oxygenates inhibit etherification and double bond isomerization, thereby reducing the effectiveness of this unit [12]. Even though the HTFT condensate is not subject to oxygenate removal in industrial practice, there is a benefit to be derived from doing so.
23.2 Production of n-1-Alkenes (Linear α-Olefins)
N
−3 O Density = 1027 kg m Melting point = −24 °C Boiling point = 202 °C
Figure 23.2
Structure of N-Methyl-2-pyrrolidone (NMP).
On account of their polar nature, oxygenates have a strong affinity for the polar extractant N-methyl-2-pyrrolidone (NMP) (Figure 23.2) [13]. NMP also has a high density, which makes it ideal for liquid–liquid extraction applications. NMP can therefore be considered as a solvent for extractive removal of oxygenates from HTFT naphtha, and it is indeed industrially applied in n-1-alkene purification [10, 14, 15]. The main drawback of NMP as solvent for Fischer–Tropsch applications is its basic nitrogen. In syncrude fractions that contain carboxylic acids, subsequent recovery of the NMP is very difficult [14]. Although this is not a problem in 1-hexene purification, it is an issue in 1-octene purification (Section 23.2.2). The C6 -cut is already rich in 1-hexene, but for comonomer applications, the purity must be very high. The alkene isomers that are close boiling to the n-1-alkene present a separation challenge, because these isomers cannot be separated by temperature (distillation) or polarity (extraction). For n-1-alkene extraction, the close boiling alkene isomers (Table 23.3) [16, 17] must be reactively converted to isomers that are less close boiling. This is accomplished by etherification with methanol (Section 17.3). During the etherification reaction, alkenes with a CDC on a tertiary carbon are selectively converted into the corresponding methyl ether or are double bond isomerized. The combined conversion to ethers and double bond isomers that are not close boiling to the n-1-alkene is quite high. In the case of 1-hexene, the conversion of 2-ethyl-1-butene can be >99% and the conversion of 2-methyl-1-pentene can be 92–96% [10, 11]. The conversion is equilibrium limited. The product from the etherification reactor is separated to produce a light product, a heavy product and a C6 -narrow cut for further purification. This separation can be performed in two distillation columns or in a single divided wall column. The distillation takes place at near-atmospheric conditions [10]. The methanol is recovered and recycled to the etherification reactor. The C6 -narrow cut contains around 94% 1-hexene and further purification is performed by NMP extraction, using an NMP:C6 mass ratio of 15 : 1 [10]. The purpose of the extractive distillation is to remove the close boiling alkane isomers, and other compounds that are still present in small concentrations. Three extractive distillation columns are employed in series. Hydrocarbon isomers present in HTFT syncrude that are close boiling to 1-pentene and 1-hexene, respectively.
Table 23.3
C5 hydrocarbons 3-Methyl-1-butene 2-Methylbutane 1-Pentenea 2-Methyl-1-butene n-Pentane cis-2-Pentene a
Boiling point (◦ C) 20.0 27.8 30.0 31.2 36.1 36.9
C6 hydrocarbons 2-Methylpentane 2-Methyl-1-pentene 3-Methylpentane 1-Hexene 2-Ethyl-1-butene cis-3-Hexene
Underlined entries indicate the target compound for purification.
Boiling point (◦ C) 60.3 60.7 60.3 60.3 64.7 66.4
469
470
23 Chemical Technologies
Most of the alkanes are recovered overhead from the first column. The purified 1-hexene, about 98.5% pure, is recovered overhead from the second column. The third column serves as a solvent stripper, where the cycloalkenes and oxygenates that contaminate and build up in the NMP are removed from the NMP solvent. The NMP is then recycled for extraction. 23.2.2 Extraction of 1-Octene
The complexity of syncrude increases with carbon number. This is partly due to the proliferation of isomers that is possible for each carbon number. A more important aspect from an extraction point of view is oxygenate partitioning (Section 5.3.3). The carbon number difference between hydrocarbons and oxygenates in the same boiling range is around 2–6 depending on the oxygenate class. The C7 –C10 heavy naphtha fraction of the HTFT light oil is just heavy enough so that there are C3 –C6 oxygenates partitioning into the oil phase that are coboiling with the heavy naphtha. The concentration of oxygenates is in the same order of magnitude as that of the C5 –C6 light naphtha from the light oil (Table 23.2), but the heavy naphtha also contains carboxylic acids. This presents a problem for NMP extraction, which requires the carboxylic acids to be removed first. One approach to acid removal is to neutralize the carboxylic acids with an appropriate water-soluble base and then remove the carboxylate salts in the aqueous phase. This methodology was applied in the design of the first 1-octene extraction unit at Sasol Synfuels (Figure 23.3) [14]. The neutralization process employed a potassium carbonate (K2 CO3 ) in water solution (Equation 23.1). 2 RCOOH C K2 CO3 (aq) C 2 RCOO KC (aq) C H2 O C CO2
(23.1)
The design intent was for neutralization to be limited to the potassium carboxylate and potassium bicarbonate (KHCO3 ), which could be regenerated in a thermal regenerator [14]. The bicarbonate is still a base and as indicated by Equation 23.1, the reaction can proceed to completion, converting the carbonate into water and carbon dioxide. It has been reported that the industrial thermal regenerator used for this process was prone to sludge buildup, which had to be disposed of [14]. This resulted in a loss of potassium from the system. Another problem with the K2 CO3 neutralization approach is that the potassium carboxylates are surface active. The carboxylates in the carbon number range of the butanoates are amphophilic (i.e., they have hydrophilic and hydrophobic character). The phase behavior of C2 –C6 potassium carboxylates in water–alcohol mixtures indicates that the carboxylates may be found in both the aqueous and organic phases [18]. Separation by water extraction is not necessarily quantitative and the potassium carboxylates in the organic phase may cause downstream problems, as was indeed industrially found (Section 9.5.4) [18, 19]. Once the carboxylic acids are removed from the light oil, a C8 -cut can be prepared (Figure 23.3). Topping and tailing of the light oil takes place in a divided wall distillation column, instead of two columns, and produces a C8 -cut. The C8 -cut contains hydrocarbons and nonacid oxygenates. The oxygenate compounds are removed by extractive distillation using NMP as a heavy-boiling solvent. The oxygenates can then be separated from the NMP and the NMP is recycled.
23.2 Production of n-1-Alkenes (Linear α-Olefins)
HTFT naphtha
C7 and lighter Oxygenates
C7 –C10
K2CO3 (aq)
Recycle of NMP Thermal regenerator
Extractive distillation
C9 and heavier
Oxygenate free C8 -cut
Methylheptene 1-Octene Cycloalkenes
Recycle of NMP
n -Octane Superfractionation Figure 23.3
Extractive distillation
Extraction of 1-octene from Fischer–Tropsch syncrude, original Train I design.
The oxygenate-free C8 -cut is then processed in two superfractionation columns to remove close boiling lighter and heavier compounds from the 1-octene (Table 23.4) [16, 17]. The C8 -narrow cut thus produced is rich in 1-octene. The C8 -narrow cut from superfractionation is further purified by extractive distillation with NMP to remove the cycloalkenes that are still present in the 1-octene. The cycloalkenes are undesirable when 1-octene is employed as comonomer [14]. The final 1-octene purity is better than 97%. A different approach has been suggested to remove the oxygenates from the Fischer–Tropsch light oil, without first removing the carboxylic acids. It was found that all of the oxygenates could be removed from the C8 -cut by azeotropic distillation with ethanol and water [15]. By doing so, the two-step oxygenate removal can be consolidated in a single step, leading to a more efficient design. It also avoids potential problems with metal carboxylates. The azeotropic distillation methodology was followed in the design of the second 1-octene extraction unit at Sasol Synfuels (Figure 23.4) [14, 20]. The separation of the oxygenate-free C8 -cut by superfractionation and NMP extraction is analogous to that described before. The main difference between this part of the design of the first (Figure 23.3) and second (Figure 23.4) 1-octene extraction units, is heat integration and optimization. An overall energy saving of 30% was achieved. This helped to offset the additional processing cost associated with the production of a higher quality 1-octene product, and the net overall energy consumption of the second 1-octene extraction unit is 15% higher than that of the first [14].
471
472
23 Chemical Technologies Table 23.4 Some of the close boiling hydrocarbon isomers to 1-octene that may affect 1-octene extraction.
Boiling point (◦ C)
C8 hydrocarbons 2-Methylheptane 2-Methyl-1-heptene 3-Methylheptane 2-Ethyl-1-hexene 1-Octenea 1,2,3-Trimethylcyclopenteneb trans-4-Octene cis-4-Octene 2-Methyl-2-heptene cis-3-Octene trans-3-Octene cis-2-Octene trans-2-Octene n-Octane
117.6 118.2 119.0 120.0 121.3 121.6 122.3 122.5 122.6 122.9 123.3 125.0 125.6 125.7
a Underlined b There
entry indicates the target compound for purification. are various C3 -alkylcyclopentanes close boiling to 1-octene.
C7 and lighter
HTFT naphtha
C7 –C10
C8-cut
Azeotropic distillation Oxygenates C9 and heavier
Oxygenate free C8-cut
Methylheptene 1-Octene Cycloalkenes
Recycle of NMP
n -Octane Superfractionation Figure 23.4
Extractive distillation
Extraction of 1-octene from Fischer–Tropsch syncrude, Train II design.
23.2 Production of n-1-Alkenes (Linear α-Olefins)
23.2.3 Production of 1-Octene from 1-Heptene
The process to produce 1-octene from Fischer–Tropsch-derived 1-heptene consists of five steps (Figure 23.5) [21]. The first step is to extract the 1-heptene from the Fischer–Tropsch syncrude and purify it to a sufficient degree so that it is compatible with the hydroformylation technology. The extraction process is similar to that described for 1-octene extraction (Section 23.2.2), but it is not necessary to produce such a narrow 1-heptene cut, since purification of the final product is performed in the last step of the process. The hydroformylation reaction is conducted in the liquid phase with a homogeneous catalyst, usually a Co- or Rh-based organometallic catalyst (Section 16.3.6). Being a homogeneous catalytic process, catalyst recovery and recycling are important to the design and more so when an expensive catalyst is employed. Hydroformylation of 1-heptene to octanal (Equation 23.2) is the first of three conversion steps, and it is followed by hydrogenation of the octanal to 1-octanol (Equation 23.3) and then the dehydration of the 1-octanol to 1-octene (Equation 23.4). CH3 (CH2 )4 –CHDCH2 C CO C H2 ! CH3 (CH2 )4 –CH2 –CH2 –CHDO
(23.2)
CH3 (CH2 )5 –CH2 –CHDO + H2 ! CH3 (CH2 )5 –CH2 –CH2 –OH
(23.3)
CH3 (CH2 )5 –CH2 –CH2 –OH ! CH3 (CH2 )5 –CHDCH2 C H2 O
(23.4)
Since the process involves a stepwise synthesis, the efficiency of the overall process is multiplicative. One advantage of the many reaction steps is that each conversion step enables the efficient removal of some impurities [21]. For example, better separation of close boiling isomers to 1-octene can be achieved before the final dehydration of 1-octanol to 1-octene. The final product polishing step requires only the separation of some internal and branched alkenes produced during the dehydration reaction. This configuration places an additional selectivity burden on the dehydration reaction and specifically on the avoidance of double bond isomerization. The dehydration reaction and its associated catalysis make it difficult to avoid double bond isomerization at high per pass conversion (Section 17.2.2). Although alumina is the preferred dehydration catalyst [22], it is capable of complex reaction chemistry [23]. HTFT naphtha
Extraction of 1-heptene
H2
H 2O
Hydroformylation
CO + H2
Selective hydrogenation
Alcohol dehydration
1-Octene purification
1-Octene Naphtha (to refinery)
Figure 23.5
Conversion of 1-heptene from Fischer–Tropsch syncrude into 1-octene.
473
474
23 Chemical Technologies
23.2.4 Distillate-Range n-1-Alkene Extraction
The same basic principles that have been applied to the extraction of n-1-alkenes from heavy naphtha (Section 23.2.2) can also be applied to distillate-range materials. The Fischer–Tropsch syncrude matrix is more complex and obtaining a pure n-1-alkene is more difficult. In some applications, a partial purification is all that is required. For example, the production of detergent-range alcohols from Fischer–Tropsch distillate requires alkene purification but not n-1-alkene separation. Hydroformylation of the mixed alkene product, followed by aldehyde to alcohol hydrogenation (similar to what is shown in Figure 23.5), yields an alcohol product that reflects the degree of branching of the Fischer–Tropsch syncrude. Specifically, the Rh-catalyzed hydroformylation of HTFT distillate-range alkenes, as industrially applied at Sasol Synfuels (Section 9.5), produces a product with little branching in the 2- and 3-position relative to the alcohol functionality [20].
23.3 Autoxidation
Autoxidation refers to conversion processes involving air oxidation without a catalyst. It is different from selective catalytic oxidation reactions that require the presence of a catalyst to proceed. Selective catalytic oxidation is performed in many petrochemical processes [24, 25]. Although the mechanism of some catalytic oxidation processes involves free radical chemistry, autoxidation processes proceed exclusively by free radical chemistry and autoxidation are inherently less selective than catalytic oxidation. The main advantages of autoxidation over catalytic oxidation are process simplicity, low sensitivity to contaminants and a lower risk of complete oxidation to CO2 . Selective oxidation in general, which includes autoxidation, has the further advantage that it employs air as O2 source, which is cheap and readily available and does not require very severe operating conditions. There are different autoxidation regimes (Section 23.3.1), which have very different industrial applications. The regimes that are of interest in Fischer–Tropsch refining are autoxidation reactions taking place at moderate to low temperature, where some control over the product selectivity can be achieved. Autoxidation naturally takes place at ambient conditions too and is responsible for product degradation during storage (Sections 14.3.7 and 15.3.8) [26, 27]. This type of autoxidation is usually undesirable, although the chemistry is the same. This again highlights the importance of selectivity and control in autoxidation. The challenge of selectivity in autoxidation of hydrocarbons has been outlined by Emanuel [28], who pointed out that selectivity control is possible through, among others, limited initiation, differences in gas-phase and liquid-phase kinetics, and exploitation of the sequential nature of the reaction network. Selectivity control and limitations to selectivity control are explored during the discussion on reaction chemistry (Section 23.3.2). Fundamentally, there is synergy between autoxidation and Fischer–Tropsch refining. The Fischer–Tropsch syncrude already contains oxygenates. Introduction of additional oxygenate functionality into syncrude through autoxidation does not affect the refining complexity. Potential refining applications of autoxidation in general, have already been discussed (Section 16.3.7).
23.3 Autoxidation
Fischer–Tropsch syncrude-specific refining applications of autoxidation by air that can be considered are as follows: 1) Improving the quality of diesel fuel. Surface-active polar compounds provide boundary layer lubricity, and oxygen is the most effective of the heteroatoms in providing such lubricity (Section 15.3.4). Mild autoxidation can introduce alcohol and carbonyl functionalities to provide lubricity. Although carboxylic acids are more effective [29], the production of carboxylic acids through autoxidation involves C–C scission and runs the risk of producing corrosive short-chain carboxylic acids. In addition to boundary layer lubricity, the hydroperoxides formed during autoxidation also provide cetane number improvement [30]. The latter is typically not required for Fischer–Tropsch-derived diesel fuel. 2) Improving the density of diesel fuel. In countries where the diesel fuel specification includes a lower density limit for better engine emission control, it becomes difficult to produce on-specification diesel fuel from Fischer-Tropsch syncrude in high yield [31]. One of the strategies for improving diesel fuel density is autoxidation, which not only provides lubricity and cetane number as mentioned before, but also results in increased density (Table 23.5) [32]. 3) Lubricant base oil production. Thermal oligomerization of distillate-range Fischer–Tropsch syncrude fractions can produce good quality lubricant base oil (Section 19.3.6). This type of conversion requires high temperature. Attempts to perform thermal oligomerization at lower temperatures were abandoned because of low productivity of the peroxide-initiator, which made the process uneconomical [33]. Autoxidation can introduce hydroperoxides into syncrude at a lower cost than traditional initiators such as di-tert-butyl peroxide (DTBP). Combined autoxidation and thermal oligomerization is an alternative pathway for lubricant base oil production from straight-run Fischer–Tropsch distillates that are rich in n-1-alkenes. The reaction proceeds at low temperature and a 10% yield of peroxide oil has been obtained within 2 h during autoxidation of 1-hexadecene at 130 Ž C [34]. In fact, this type of oxidative oligomerization is well known for the conversion of unsaturated bio-oils by air blowing into protective coatings, oilcloth, linoleum, and ink [35]. Change in Fischer–Tropsch distillate properties after autoxidation at 160 Ž C for 90 min with oxygen.
Table 23.5
Fuel property
Distillate yield from feed (mass%) Density (kgÐm3 ) Oxygen content (mass%) Lubricity, HFRR wear scar (µm) Distillation profile (Ž C) T10 T50 T90
Fischer–Tropsch-derived distillate Unoxidized
Autoxidized
– 773.2 0.2 570
95 811.5 5.4 365
238 304 352
237 304 364
475
476
23 Chemical Technologies
Alcohols Alkanes
Autoxidation
Selective hydrogenation
Dehydration
Alkane to alcohol Alkane to carbonyl
Carbonyl to alcohol
Alcohol to alkene
Alkenes
Figure 23.6 Conversion of alkanes by selective autoxidation and hydrogenation into alcohols or by subsequent dehydration into alkenes.
4) Alkanes can be activated by autoxidation at low temperature. Once the alkanes have been converted into oxygenates, further conversion is easier. A conceptually simple process has been proposed to employ selective autoxidation for the production of alcohols and alkenes (Figure 23.6) [36]. There are similarities with the 1-heptene to 1-octene conversion process (Figure 23.5), with autoxidation replacing n-1-alkene extraction and hydroformylation and the product being an internal alcohol or alkene. The dehydration step can also be replaced with another acid-catalyzed conversion, for example, aromatic alkylation. If the last step is omitted, alcohols are obtained as final products. Depending on the chain length of the starting material, the alcohols can be employed as solvents, plasticizers, or detergents. Not all solvent applications require 1-alcohols. 5) Carboxylic acid (fatty acid) production. Prolonged autoxidation of syncrude, or autoxidation under more severe conditions, yields carboxylic acids as products. This type of conversion has been extensively used in the past for the production of carboxylic acids and carboxylate soaps from Fischer–Tropsch slack wax (Chapter 6) [5]. Changes in the soap market after the Second World War made autoxidation of slack wax for soap manufacturing uneconomical and it was discontinued [37]. However, the same process may also be applied to produce lighter carboxylic acids. Ethanoic acid (acetic acid) and propanoic acid (propionic acid) are valuable commodity chemicals and these light carboxylic acids are more easily recovered from syncrude autoxidation products than by extraction from the Fischer–Tropsch aqueous product. 6) Wax oxidation to produce oxidized waxes. There are many applications of oxidized wax products, which include emulsifiers, polishes, ink carriers, and surface coating for various applications. This process is industrially practised with Fischer–Tropsch waxes [38, 39], and is discussed in more detail (Section 23.3.3). 7) Autoxidation for the conversion of refinery waste products. Heavy or solid-rich waste products can be autoxidized to produce lighter oxygenates that can be recovered as products. Because autoxidation is noncatalytic, contaminants that would otherwise foul catalytic systems do not affect autoxidation. Even when oxidation inhibitors are present, it is possible to break down organic compounds by prolonged autoxidation. Autoxidation can also be employed during wastewater treatment to reduce the chemical oxygen demand by preoxidizing the organic species. This has the added benefit of making the material more susceptible to conversion by microorganisms.
23.3 Autoxidation
23.3.1 Autoxidation Regimes
The three main variables that affect product selectivity within any autoxidation application are temperature, oxygen availability, and duration of exposure. Different autoxidation regimes can be identified for different applications. On the two extremes are the very-high-temperature oxidation processes that are employed for total or partial combustion, as is found during synthesis gas production (Chapter 3) and the ambient-temperature autoxidation that affects product degradation during storage. Of practical interest in refining are the autoxidation regimes in the temperature range 75–475 Ž C: 1) High temperature and low oxygen availability. High-temperature (>300 Ž C) thermal conversion does not require O2 . These conditions are typical of thermal residue upgrading technologies such as visbreaking (Section 16.6.2) and coking (Section 16.6.4). Toward the lower end of this temperature range, radical recombination reaction, or retrograde condensation, dominates. In the absence of a hydrogen-donor solvent, chain degradation takes place with the production of refractory aromatic products and coke production is promoted. When a suitable hydrogen-donor solvent is available, these conditions can be employed for direct liquefaction. For example, the direct thermal liquefaction of coal is usually conducted in the temperature range 425–460 Ž C [40]. The rate of conversion is increased by the introduction of some O2 to produce oxygenates. The oxygenates have lower homolytic bond dissociation energies (Tables 21.2 and 21.3) and are more readily thermally cracked. This autoxidation regime has little value in the Fischer–Tropsch refinery, although it may have use in the coproduction of products from the raw material used for synthesis gas generation. 2) High temperature and high oxygen availability. At temperatures >300 Ž C and in the presence of sufficient O2 , a carbon-based material is rapidly oxidatively degraded. Under these conditions, carboxylic acids and light hydrocarbons are readily produced by a combination of oxidation and thermal cracking. In the absence of a suitable diluent, retrograde condensation results in the formation of more refractory aromatic products, but with sufficient exposure time these compounds may ultimately also be oxidatively degraded. This is the type of autoxidation employed in gasification and reforming. 3) Medium temperature (180–300 ◦ C). This autoxidation regime is used in processes such as asphalt oxidation to decrease its softening point (harden the asphalt) [41]. The asphalt conversion is regulated by a combination of oxygen availability and exposure time. The autoxidation strips out some lighter material and promotes retrograde condensation of the more reactive, oxidized asphalt to produce a more refractory product. In this way, much of the oxygen that has been incorporated during autoxidation is again lost through chain degradation and the production of lighter oxygenates, such as carboxylic acids. To employ this autoxidation regime for oxygenate production, retrograde condensation must be suppressed or prevented when thermal cracking becomes significant or when the feed is alkene rich. Alkene-rich feed autoxidation in this temperature regime leads to a rapid increase in viscosity due to oxidative oligomerization [35]. 4) Moderate temperature (150–180 ◦ C). The moderate-temperature autoxidation regime is employed for the production of oxidized waxes and other oxygen-functionalized hydrocarbons (Section 23.3.3). Selectivity can be regulated somewhat by oxygen availability and exposure time, but the temperature is generally too high to eliminate the production of secondary
477
478
23 Chemical Technologies
oxidation products. Some chain degradation takes place and carboxylic acids and esters are usually produced in abundance, but the temperature is too low for thermal cracking and retrograde reactions are not significant in alkane-based feed materials. Alkene-based feed materials undergo oxidative oligomerization. 5) Low temperature (<150 ◦ C). Low-temperature autoxidation with low oxygen availability is selective toward primary oxidation products (alcohols, carbonyls, and hydroperoxides). Although oxygen is incorporated into the product, the conditions are such that little chain degradation takes place. Operating in this regime can be very selective, but with continued exposure secondary reactions eventually take place. By limiting the exposure time and the overall extent of conversion, the probability of secondary reactions can be reduced. When oxygen availability is increased in combination with exposure time, secondary oxidation products (acids and esters) are produced along with some chain degradation. However, the temperature is too low for thermal cracking, and with alkane-rich feed materials, retrograde reactions are not significant. For alkene-rich materials, the upper temperature threshold of low-temperature autoxidation is 130 Ž C, not 150 Ž C, and below this temperature there is a measurable induction period before oxidative oligomerization takes place [42]. 23.3.2 Reaction Chemistry
Autoxidation takes place by a free radical mechanism. As is the case with all free radical reactions, there are three distinct reaction steps: initiation, propagation, and termination. In the case of autoxidation, initiation is through the action of molecular oxygen (O2 ) and it can take place even at ambient conditions, albeit very slowly. Initiation takes place when molecular oxygen abstracts a hydrogen atom from the hydrocarbon (Equation 23.5). ž
ž
RH C O2 ! R C OOH
(23.5)
This is a slow reaction even at moderate temperatures (150–180 Ž C). Two radicals are generated. The concentration of the hydroperoxy radical (ž OOH) remains low and it rapidly abstracts a hydrogen to form hydrogen peroxide (Equation 23.6). ž
ž
RH C OOH ! R C H2 O2
(23.6)
The alkyl radical (Rž ) that is formed by initiation (Equation 23.5) and the subsequent hydroperoxy radical reaction (Equation 23.6) determine the rate of O2 uptake. If this radical is stable, O2 uptake is slow, since further O2 uptake is autocatalytic (Equations 23.7–23.9). ž
R C O2 ! ROO ž
ž
(23.7) ž
RH C ROO ! R C ROOH
(23.8)
ž ž RCHDCH2 C ROO ! 0R(C )H–CH2 –OOR
(23.9)
0
The activation energy for the O2 uptake reaction (Equation 23.7) is very small and the O2 uptake reaction is generally fast [27]. The main products from initiation are, therefore, the alkylperoxy radical (ROOž ) and the hydroperoxide. The concentration of the alkylperoxy radical is determined by the extent of the initiation reaction (Equation 23.5). This is a slow reaction, and despite the importance of the alkylperoxy radical, the alkylperoxy radical is not the species that governs
23.3 Autoxidation
Oxygenate products
O2 consumption
Concentration
O2 conversion
Organic peroxides
Figure 23.7 Progress of autoxidation over time, indicating the induction period and buildup of organic peroxides, whereafter the rates of O2 consumption and product formation rapidly increase.
Time Induction period
autoxidation. Autoxidation is governed by the hydroperoxides (Equation 23.8) and alkylperoxides (Equation 23.9). During the induction period, as the autoxidation progresses, the concentration of hydroperoxides and alkylperoxides gradually increases (Figure 23.7). During the induction period, before the rate of autoxidation becomes significant, some oygenates (mainly ketones and alcohols) are formed as final products. The rate of O2 consumption then rapidly increases as the peroxides decompose and increase the concentration of free radicals in the reaction mixture. The chemistry of the organic peroxides is extensive and it is described in the monumental work edited by Swern [43–45]. The organic peroxides all contain the labile peroxide bond (O–O). Decomposition of the peroxide bond, by either homolytic bond dissociation (Equation 23.10) or other processes, such as radical-induced dissociation (Equation 23.11), determines the rate and nature of the products formed by autoxidation. ž
ž
ROOH ! RO C OH ž
(23.10) ž
ROOH C ROO ! ROH C RO C O2
(23.11)
At temperatures <180 Ž C, the contribution of thermally induced dissociation is not as significant as induced dissociation, either homolytically or otherwise. To quote Haitt [46]: ‘‘. . .the evidence appears to offer incontrovertible proof that the homolysis of RO2 H can be assisted by olefins, aldehydes and ketones, carboxylic acids, alcohols, amines and amine salts, sulfur compounds, bromide ions, aralkanes, or another molecule of hydroperoxide; in other words, by practically anything.’’ There are two important observations relating to autoxidation reaction chemistry of initiation and autocatalytic O2 consumption. Firstly, the contribution of the induction period to the overall product yield is minimal. The conditions during this time may be harsher to speed up the autoxidation. The induction period can also be practically eliminated by the use of small amounts of metal ions [47, 48] or, as Hiatt put it, ‘‘practically anything’’ that assists in peroxide homolysis [46]. Secondly, the organic peroxides are the common intermediates of autoxidation products. The primary products from hydroperoxide autoxidation are alcohols and ketones (Figure 23.8). The reactions involve a mixture of propagation and termination steps and the selectivity is influenced by the reaction parameters. Low temperature favors alcohol formation over ketone formation. This is due to the lower free radical concentration and the higher probability of hydrogen abstraction by the alkoxy radical (ROž ). In n-alkanes, the position of the oxygenate
479
480
23 Chemical Technologies
OH ″RH
OOH R
C
R′
H
R
R
C
R′ + ″R
H O
O − HO
C
R′
″ROO
R
C
R′ + ″ROOH
H O ″RO
Figure 23.8
R
C
R′ + ″ROH
Alcohols and ketones as primary autoxidation products.
functionality is of equal probability along the chain, with the exception of terminal carbons, which are not oxidized as primary products [49], (i.e., little or no 1-alcohols and aldehydes are formed). In alcohols, the hydrogen on the carbon containing the alcohol functionality can be oxidatively removed to produce ketones. This is a secondary reaction, but the product is indistinguishable from ketones that are formed as primary autoxidation products. The ketones are also key intermediates during the formation of acids and esters. Autoxidation results in the introduction of a peroxide functionality on the carbon adjacent to the carbonyl group. Carboxylic acids and esters are secondary autoxidation products. The carboxylic acids and esters are usually produced in a close to 1 : 1 ratio. During the autoxidation of Fischer–Tropsch waxes, it was found that carboxylic acid formation was completely suppressed at low O2 availability and low temperature, but esters were still produced in low concentration [39]. The overall reaction network for hydrocarbon autoxidation is presented in Figure 23.9. Further oxidation of the secondary products yields CO2 . The CO2 is partly produced by the decarboxylation of the carboxylic acids, but not exclusively so [50]. Oxidative decarboxylation of carboxylic acids proceeds only to a limited extent. It was also shown that esters could be decarboxylated. 23.3.3 Fischer–Tropsch Wax Oxidation
A number of studies described the autoxidation of Fischer–Tropsch waxes specifically [38, 39, 51–53]. Industrially, Fischer–Tropsch wax oxidation is conducted by batch-mode autoxidation. The autoxidation takes place in a bubble column reactor with internal heating and cooling coils for heat management. The process, strictly speaking, is a semibatch process, with the Alcohols Hydrocarbon
O2
Peroxides
Carboxylic acids Ketones Esters
Figure 23.9
Reaction network of hydrocarbon autoxidation.
23.3 Autoxidation
LTFT wax
Oxidized wax
Step 1 Wax preheating
Step 2 Autoxidation (initiation)
Step 3 Wax cooling
Step 4 Autoxidation (propagation)
N2
Air
Air
Air
Figure 23.10 Workflow of commercial batch-mode oxidized wax production at Sasol 1. (1) The loaded wax is preheated to the autoxidation temperature under N2 . (2) The first phase of autoxidation is mainly needed to initiate autoxidation. (3) The temperature
LiOH or Ca(OH)2
Step 5 Wax cooling
N2
is then adjusted to the autoxidation temperature for the propagation period. (4) The second phase of autoxidation during which most of the wax oxidation takes place. (5) The oxidized wax is cooled down under nitrogen for unloading.
Fischer–Tropsch wax being a stationary phase (batch), while air is continuously bubbled through the molten wax. The steps involved in batch-mode wax oxidation are shown in Figure 23.10 [36]. The wax conversion is performed in two steps. The induction of the autoxidation process is slow in the absence of a catalyst or initiator. For this reason, the first oxidation period is usually conducted at a higher temperature to reduce the duration of the induction period. Most of the autoxidation takes place during the second oxidation period, which controls the oxidation selectivity and final product properties. Batch times may be preset, or can be determined based on the analysis of the product. The duration of the second autoxidation period is typically adjusted in accordance with the wax conversion, as indicated by the acid number of the oxidized wax. Some wax grades are saponified after autoxidation to convert some of the acids into soaps (metal carboxylates). This takes place at the end of the second period of autoxidation. The determination of the acid number [54] is a comparatively easy laboratory method that can be performed in a short period of time. Most other oxidized wax analysis methods are quite cumbersome [55–57]. Assessing the extent of oxidation by the acid number is of course not a true reflection of the actual extent of oxidation (Section 23.3.2). Spectrometric methods based on infrared (IR) and near-infrared (NIR) analyses have been successfully employed for the characterization of oxidized Fischer–Tropsch waxes [39, 58, 59]. Such spectrometric methods provide a more accurate reflection of the overall extent of oxidation and are less time consuming than wet chemical analyses. The properties of the oxidized wax are mainly determined by the temperature programme, operating conditions, and wax feed. Some variation is introduced in product quality by residual oxidized wax from the previous batch that affects initiation time; corrosion products can catalyze autoxidation [47, 48], and handling conditions on unloading and storage of the oxidized wax may allow further autoxidation. The production procedure (Table 23.6) and properties (Table 23.7) of some of the commercial grades of oxidized Fischer–Tropsch waxes produced by batch-mode autoxidation are given [36]. In general, the nature of batch-mode autoxidation is such that oxygen availability is not controlled during the second autoxidation period. Carboxylic acids are consequently produced. Apart from the carboxylic acid functionality in the oxidized wax, oxidation also results in the formation of short-chain carboxylic acids and other oxygenates. The short-chain oxygenates are stripped from the oxidized wax by the air that is bubbled through the reactor and the off-gas is
481
482
23 Chemical Technologies Table 23.6 Production procedure for different commercial oxidized waxes according to the workflow shown in Figure 23.10.
Step
1 2
3 4
5
a Highly
Description
Fe-LTFT wax feed Saponification chemical Wax preheating time (min) Autoxidation, phase 1 time (min) Autoxidation temperature (Ž C) Wax cooling time (min) Autoxidation, phase 2 time (min) Autoxidation temperature (Ž C) Acid number (mg KOH/g) Unloading temperature (Ž C) Average batch time (min)
Oxidized wax grades
Saponified oxidized wax grades
A1
A6a
A28.1
A2
A3
A14
A28.2
H2-wax None
H2-wax None
C105-wax None
H2-wax LiOH
H2-wax LiOH
H2-wax Ca(OH)2
C105-wax LiOH
10–15
10–15
10–15
10–15
10–15
10–15
10–15
80
80
80
80
80
80
80
175
180
175
175
175
175
175
40
0
40
40
40
40
40
>80
>80
>80
>80
>80
>80
>80
140
180
140
140
140
140
140
27–29
35–45
27–29
21–23
25–28
25–27
21–23
100–120
130
110–130
110–120
120–130
100–120
110–120
400
–
450
380
470
–
390
oxidized wax, no nitrogen used during its production.
rich in oxygenates. These oxygenates and light alkanes resulting from oxidative cleavage of the wax must be recovered before the N2 -rich air can be vented to atmosphere. The composition of the off-gas depends on the severity of autoxidation (Table 23.8) [36]. This stream is corrosive on account of the short-chain carboxylic acids. In a Fischer–Tropsch refinery, the condensed material can in principle be coprocessed with the Fischer–Tropsch aqueous product. Continuous-mode wax oxidation has been suggested as an alternative to batch-mode wax oxidation [39]. The continuous-mode process involves the cocurrent feed of wax and air in a pipe-reactor (a long process line rather than a reactor vessel). With this type of design, there is the possibility to adjust the temperature over the length of the reactor and there is also the possibility of bleeding in air at different positions along the length of the reactor (analogous to interbed injection in fixed-bed reactors). This type of autoxidation and reactor design can overcome some of the problems associated with batch-mode wax oxidation:
23.3 Autoxidation Typical properties of different commercial oxidized waxes that are produced from Sasol Fe-LTFT-derived H2- and C105-waxes.
Table 23.7
Property
LTFT wax feed H2
Acid number <0.1 (mg KOH/g) Ester number – (mg KOH/g) Penetration at <0.1 25 Ž C, ASTM D1321 (mm) Congealing point, 98 ASTM D 938 (Ž C) Melting point, – ASTM D 127 (Ž C) 217 Latent heat of fusion (kJÐkg1 )
Oxidized wax grades
C105
A1
A6
Saponified oxidized wax grades
A28.1
A2
A3
A14
–
27
37
28
11
10
7
–
28
65
27
14
24
26
<0.1
0.6
104
87
–
96.5
245
192
2.5
–
79
94 –
126
– 174
<0.4
0.25
89
91
101
103.5
182
185
A28.2 12 –
0.2
–
86
97 –
179
– 175
Specific heat capacity (kJÐkg1 ÐK1 ) At 25 Ž C At 130 Ž C At 170 Ž C Average molecular mass (gÐmol1 )
1.8 2.3 2.5
1.9 2.7 2.8
2.3 3.0 2.9
3.1 2.6 2.6
2.2 2.6 2.5
785
1100
700
720
890
940 – – –
950 – – –
950 825 819 813
950 842 830 820
940
2.5 3.8 3.8
2.1 2.8 2.6
2.6 – 1.6
2.1 3.0 2.8
750
780
750
1090
960 814 805 797
980 829 824 821
980 846 836 832
960 – – 800
32 22 15
44 38 32
36 28 21
Density (kgÐm3 ) At 25 Ž C At 100 Ž C At 110 Ž C At 120 Ž C
– 810 –
Viscosity (cP) At 125 Ž C At 135 Ž C At 145 Ž C Surface tension at 130 Ž C (NÐm1 )
– 6–10 –
– 18 –
–
–
12 10 8.5 0.0284
25 20 17 0.0291
19 16 14 0.0222
20 14 11.5 0.0288
0.0336
0.0294
0.0281
1) The initiation period and product properties in continuous-mode oxidation are not dependent on variables such as the batch-to-batch variations in cleanliness of the reactor and product handling. 2) Control over oxidation selectivity in larger scale batch reactors is limited by equipment constraints such as heating and cooling rates between periods, operating range for airflow, and bubble hydrodynamics. These are not constraints in continuous-mode oxidation.
483
484
23 Chemical Technologies Table 23.8 Typical composition of the off-gas from batch-mode autoxidation of Fischer–Tropsch waxes. Normal operation refers to the production of oxidized waxes such as A1 and severe operation refers to the production of oxidized waxes such as A6.
Description Flow rate (kgÐh1 per ton wax) Organics Water Remaining air Organics composition (mass%) Alcohols Carbonyls Carboxylic acids Alkanes Other organics
Normal operation
Severe operation
9 27 1124
19 52 1089
4 5 74 12 5
5 5 72 13 5
3) In continuous-mode autoxidation, the oxygen availability is inherently limited by consumption and can be regulated by intermediate addition. This avoids overoxidation and provides selectivity control that is more difficult to achieve with batch-mode oxidation. To achieve the same conversion-dependent control in batch-mode autoxidation, the air must be diluted with nitrogen and the ratio must be controlled over time. 4) Hydrocarbon chain degradation and carboxylic acid formation can be almost completely suppressed during continuous-mode oxidation, which is an advantage for autoxidation applications that rely on alcohol intermediates. 5) It is possible to produce oxidized wax grades with continuous-mode autoxidation that cannot be produced by batch-mode autoxidation. 6) In batch-mode oxidation, where air is introduced at the bottom of the bubble column reactor, foaming presents an operating problem. The oxidized wax is surface active, because the oxygen-containing functionality is polar. This causes the already oxidized wax to concentrate at the gas–liquid interface and stabilize the air bubbles. This not only leads to stable foam formation but also keeps the oxidized wax in contact with air, which in turn can result in overoxidation. 7) Continuous operation and steady state control are possible in continuous-mode autoxidation, whereas batch-mode control is time dependent.
23.3.4 Syncrude Process Technology
The application of autoxidation with a paraffinic Fischer–Tropsch-derived feed is no different from the general application of autoxidation to alkanes. The situation is different when applying
References
autoxidation to straight-run syncrude, which contains alkenes and oxygenates. A few general observations are as follows: 1) The presence of low levels of metals in syncrude, specifically iron or cobalt, has a potentially beneficial effect on autoxidation, by increasing the rate of peroxide decomposition. Many of the other nonalkane species may fulfill a similar role. This may result in autoxidation with a very short induction period. 2) The oxygenates present in syncrude are susceptible to further autoxidation (Figure 23.9). 3) The alkenes present in syncrude are not only susceptible to autoxidation but are likely to result in oxidative oligomerization. This is advantageous only when oil production is the objective.
References 1. Miller, S.A. (ed.) (1969) Ethylene and Its Indus2. 3. 4.
5. 6. 7.
8.
9.
10.
11.
12.
13.
trial Derivatives, Ernest Benn, London. Hancock, E.G. (ed.) (1973) Propylene and Its Industrial Derivatives, Ernest Benn, London. Hancock, E.G. (ed.) (1975) Benzene and Its Industrial Derivatives, Ernest Benn, London. Hancock, E.G. (ed.) (1982) Toluene, the Xylenes and their Industrial Derivatives, Elsevier, Amsterdam. Asinger, F. (1968) Paraffins Chemistry and Technology, Pergamon, Oxford. Asinger, F. (1968) Mono-Olefins Chemistry and Technology, Pergamon, Oxford. Camara Greiner, E.O., Gubler, R., and Inoguchi, Y. (2004) Linear Alpha Olefins, Chemical Economics Handbook Marketing Research Report, Stanford Research Institute, Menlo Park, CA. Collings, J. (2002) Mind Over Matter. The Sasol Story: A Half-century of Technological Innovation, Sasol, Johannesburg. Dixon, J.T., Green, M.J., Hess, F.M., and Morgan, D.H. (2004) Advances in selective ethylene trimerisation – a critical overview. J. Organomet. Chem., 689, 3641–3668. (2001) Linear Alpha Olefins. SRI Process Economics Program Report 12D, Stanford Research Institute, Menlo Park, CA. De Klerk, A. (2004) Etherification of C6 Fischer-Tropsch material for linear α-olefin recovery. Ind. Eng. Chem. Res., 43, 6349–6354. Smook, D. and De Klerk, A. (2006) Inhibition of etherification and isomerization by oxygenates. Ind. Eng. Chem. Res., 45, 467–471. Hawley, G.G. (1971) The Condensed Chemical Dictionary, 8th edn, Van Nostrand Reinhold, New York.
14. Hahn, T. (2003) The octene train II purifica-
15.
16.
17.
18.
19.
20.
21.
tion process. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd013. Diamond, D., Hahn, T., Becker, H., and Patterson, G. (2004) Improving the understanding of a novel complex azeotropic distillation process using a simplified graphical model and simulation. Chem. Eng. Process., 43, 483–493. Weast, R.C. (ed.) (1988) CRC Handbook of Chemistry and Physics, CRC Press, Boca Raton, FL. (1988) Physical Constants of Hydrocarbons and Non-Hydrocarbon Compounds, ASTM Data Series DS 4B, 2nd edn, ASTM, Philadelphia, PA. Van Eeden, J.A. (2003) Short chain potassium carboxylate phase behaviour. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd058. De Klerk, A. (2009) Reaction engineering in Fischer-Tropsch context. E V Murphree Award symposium in honour of Prof. M. P. Dudukovi´c. 237th American Chemical Society National Meeting, March 22–26, 2009, Salt Lake City, USA. Redman, A. (2005) Production of olefins and oxygenated compounds from Fischer-Tropsch. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd179. McGurk, K. (2003) From 1-heptene to 1-octene: a new production route. Proceedings of the South African Chemical Engineering Congress, September 3–5, 2003, Sun City, South Africa, p. cd082.
485
486
23 Chemical Technologies 22. Makgoba, N.P., Sakuneka, T.M., Koortzen, J.G.,
23.
24. 25.
26.
27. 28.
29.
30.
31.
32.
33.
34.
35.
36.
Van Schalkwyk, C., Botha, J.M., and Nicolaides, C.P. (2006) Silication of γ -alumina catalyst during the dehydration of linear primary alcohols. Appl. Catal. A, 297, 145–150. De Klerk, A. (2011) Key catalyst types for the efficient refining of Fischer–Tropsch syncrude: alumina and phosphoric acid, in Catalysis, vol. 23 (ed. J.J. Spivey), Royal Society of Chemistry, Cambridge, pp. 1–49. Hucknall, D.J. (1974) Selective Oxidation of Hydrocarbons, Academic Press, London. Gates, B.C., Katzer, J.R., and Schuit, G.C.A. (1979) Chemistry of Catalytic Processes, McGraw-Hill, New York. Nixon, A.C. (1962) in Autoxidation and Antioxidants, vol. II (ed. W.O. Lundberg), Interscience, New York, pp. 695–856. Scott, G. (1965) Atmospheric Oxidation and Antioxidants, Elsevier, Amsterdam. Emanuel, N.M. (1965) in The Oxidation of Hydrocarbons in the Liquid Phase (ed. N.M. Emanuel), Macmillan, New York, pp. 1–31. Knothe, G. and Steidley, K.R. (2005) Lubricity of components of biodiesel and petrodiesel. The origin of biodiesel lubricity. Energy Fuels, 19, 1192–1200. Hashimoto, K., Ikeda, M., Arai, M., and Tamura, M. (1996) Cetane number improvement of diesel fuel by autoxidation. Energy Fuels, 10, 1147–1149. De Klerk, A. (2009) Can Fischer-Tropsch syncrude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. Naegeli, D.W., Childress, K.H., Moulton, D.S., and Lacey, P.I. (2002) Method for producing oxygenates fuels. US Patent 6, 488,727. Cowley, M. (2007) Oligomerisation of alkenes by radical initiation. Org. Process Res. Dev., 11, 286–288. Norton, C.J. and Drayer, D.E. (1968) in Oxidation of Organic Compounds, Liquid-Phase, Base-Catalyzed and Heteroatom Oxidations, Radical Initiation and Interactions, Inhibition (Advances in Chemistry Series, 75), Vol. I (ed. F.R. Mayo), American Chemical Society, Washington, DC, pp. 78–92. Sims, R.P.A. and Hoffman, W.H. (1962) in Autoxidation and Antioxidants, vol. II (ed. W.O. Lundberg), Interscience, New York, pp. 629–682. Bolder, F.H.A., De Klerk, A., and Visagie, J.L. (2009) Hydrogenation of oxidized wax and a process to produce olefins from paraffins by
37. 38. 39.
40.
41.
42.
43. 44. 45. 46.
47.
48.
49.
50.
51.
52.
autoxidation, selective hydrogenation, and dehydration. Ind. Eng. Chem. Res., 48, 3755–3760. Rosen, B.H. (1960) Wax oxidation. Ind. Eng. Chem., 52, 14–16. LeRoux, J. H. and Oranje, S. (eds) (1984) Fischer-Tropsch Waxes, Sasol, Sasolburg. De Klerk, A. (2003) Continuous-mode thermal oxidation of Fischer-Tropsch waxes. Ind. Eng. Chem. Res., 42, 6545–6548. Whitehurst, D.D., Mitchell, T.O., and Farcasiu, M. (1980) Coal liquefaction. The Chemistry and Technology of Thermal Processes, Academic Press, New York. Chelton, H.M., Traxler, R.N., and Romberg, J.W. (1959) Oxidized asphalts in a vertical pilot plant. Ind. Eng. Chem., 51, 1353–1354. Hess, P.S. and O’Hare, G.A. (1950) Oxidation of linseed oil. Temperature effects. Ind. Eng. Chem., 42, 1424–1431. Swern, D. (ed.) (1970) Organic Peroxides, vol. I, Wiley-Interscience, New York. Swern, D. (ed.) (1971) Organic Peroxides, vol. II, Wiley-Interscience, New York. Swern, D. (ed.) (1972) Organic Peroxides, vol. III, Wiley-Interscience, New York. Hiatt, R. (1971) in Organic Peroxides, vol II (ed. D. Swern), Wiley-Interscience, New York, pp. 1–151. Emanuel, N.M., Maizus, Z.K., and Skibida, I.P. (1969) The catalytic activity of transition metal compounds in the liquid-phase oxidation of hydrocarbons. Angew. Chem. Int. Ed., 8 (2), 97–107. Sheldon, R.A. and Kochi, J.K. (1981) Metal-Catalysed Oxidations of Organic Compounds, Academic Press, New York, pp. 33–70. Bashkirov, A.N., Kamzolkin, V.V., Sokova, K.M., and Andreyeva, T.P. (1965) in The Oxidation of Hydrocarbons in the Liquid Phase (ed. N.M. Emanuel), Macmillan, New York, pp. 183–193. Berezin, I.V., Berezkina, L.G., and Nosova, T.A. (1965) in The Oxidation of Hydrocarbons in the Liquid Phase (ed. N.M. Emanuel), Macmillan, New York, pp. 110–124. Luyt, A.S. (1985) Die oksidasiekinetika van ’n Fischer-Tropsch-was (Engl. Transl. ‘‘The oxidation kinetics of a Fischer-Tropsch wax’’). PhD thesis, Potchefstroom University for Christian Higher Education, Potchefstroom. Breet, E.L.J., Luyt, A.S., and Oranje, S. (1990) Oxidation kinetics of a Fischer-Tropsch wax. Part 1. A simple method for measuring oxygen consumption. S. Afr. J. Chem., 43 (3/4), 83–86.
References 53. Breet, E.L.J. and Luyt, A.S. (1991) Oxidation
57. Dressler, F. and Uhde, E.D. (1976) Bestimmung kinetics of a Fischer-Tropsch wax. Part 2. Inducfunktioneller gruppen in oxidierten paraffintion period as first macroscopic process stage. S. waschen (Engl. Transl. ‘‘Determination of the Afr. J. Chem., 44 (4), 101–104. functional groups in oxidised paraffin waxes’’). 54. ASTM (1998) D 1386 – 98. Standard Test Method Fette, Seifen, Anstrichm., 78 (6), 235–238. for Acid Number (Empirical) of Synthetic and Nat- 58. Prinsloo, N.M. (1992) Fast characterisation of Fischer-Tropsch waxes by modern quantitative ural Waxes, ASTM, West Conshohocken, PA. 55. Link, W.E. and Formo, M.W. (1962) in AuIR techniques for laboratory and on line applitoxidation and Antioxidants, vol. I (ed. W.O. cation. PhD thesis, University of South Africa, Lundberg), Interscience, New York, pp. Pretoria. 59. De Klerk, A. (2004) Influence of stainless steel 367–416. 56. Brink, A. and Haasbroek, P.P. (1971) Analon the autoxidation of Fischer-Tropsch waxes. ysis of oxidised waxes I: determination of Ind. Eng. Chem. Res., 43, 6898–6900. hydroxyl number. Fette, Seifen, Anstrichm., 73 (10), 608–610.
487
489
Part VI Refinery Design
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
491
24 Principles of Refinery Design 24.1 Introduction
One only has to look at the diversity of crude oil refineries to realize that there is nothing like a standard refinery configuration [1]. The same is true for Fischer–Tropsch refineries (Chapters 6–12). So, how does one go about designing a refinery? It is quite possible to deal with refinery design in similar terms as any other discipline dealing with inputs and outputs. The principle underlying refinery design is straightforward (Figure 24.1). In refinery design, it is necessary to devise a transformation (separation and conversion processes) that will convert the input (raw feed materials) into a desired output (product slate). The same principle is encountered during process design in general, as well as in disciplines such as computer software engineering, business planning, and manufacturing. All of these disciplines deal with the transformation of an input to a desired output. Different systems approaches have been advocated and these are discussed to evaluate their applicability to refinery design (Section 24.3). However, despite all the systems approaches that have been advocated, there is no simple strategy that can be followed and by which refineries are designed. In essence, refinery design is a creative process, which can be aided by a systems approach but cannot be efficiently replaced by it. On the basis of the author’s experience, there are a few concepts that are important to a successful refinery design; some are philosophical in nature and others are more down-to-earth: 1) There are characteristics and peculiarities of the refining business that must be understood for a refinery design to succeed (Section 24.2.1). 2) Refineries are complex systems (Sections 24.2.2 and 24.2.3). 3) Refining efficiency matters both to the environment and to the bottom line (Section 24.2.4).
24.2 Refinery Design Concepts 24.2.1 Characteristic of the Refining Business
Gary et al. [1] gave a very nice description of the characteristics of the refining business: Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
492
24 Principles of Refinery Design
Input Syngas production
Fischer−Tropsch synthesis
Imported material
Figure 24.1
Transformation
Pyrolysis liquids
Syncrude
Fischer−Tropsch refinery and product blending
Output
Pipeline gas LPG Motor-gasoline Jet fuel Diesel fuel Fuel oil Chemicals Lubricants Intermediates Waste
Gas condensates, crude oil products, bio-derived oils, alcohols, etc.
Principle of refinery design.
1) Every refinery is unique. There is no such thing as a cookie-cutter refinery design. 2) Crude oils are all different. Every refinery is designed with a specific mixture of crude oils in mind. The crude oil selection affects the sizes of the different units, the types of products that can be produced, and even the metallurgy of the equipment. The same is true for Fischer–Tropsch syncrude. In Fischer–Tropsch refineries, there is better control over the syncrude, but the syncrude composition is affected by Fischer–Tropsch catalyst deactivation and operating conditions (Chapter 4). 3) Refineries are capital intensive, long lived, and very specific assets. The decision to construct a refinery is a long-term proposition and needs to make provision for potential future changes in the economy and product requirements. The 1955 Sasol 1 facility (Chapter 8) is a good example of a long lived asset. 4) Refineries change over time. The evolution of crude oil refineries has been described in Chapter 2. It illustrates the dynamic nature of the refining business, which is influenced by raw material availability and changes in product requirements. Fischer–Tropsch refineries likewise changed over time (Sections 8.5 and 9.5). 5) Refining complexity differ. The number of units and diversity of products determine the complexity of a refinery. It is very wrong to assume that a specific raw material is associated with a certain refining complexity. This is illustrated by a comparison of different Fischer–Tropsch refineries employing similar syncrude, for example, Hydrocol (Chapter 7), Sasol 2 and 3 (Chapter 9), and Mossgas (Chapter 10). 6) Most refinery products are commodities. Commodities are undifferentiated from those of a competitor and are sold on the basis of price. The marketing desire to differentiate fuels may have unintended consequences. The extremely restrictive synthetic jet fuel specifications in DEF STAN 91-91 (Section 14.2.1) is a case in point. Differentiation caused more hassles than good. 7) Transportation fuels are mainly sold in regional markets. The local market situation determines the type and volume of products that can be sold. 8) Refinery optimization involves a multitude of trade-offs. This is related to the complexity of the system. In practice, refinery units are optimized and not always with respect to the total
24.2 Refinery Design Concepts
refinery. Total refinery optimization is more difficult, because it may occasionally require some units to operate less optimally for the total refinery to perform more optimally. This is a hard sell to the engineer in charge of the unit that must sacrifice for the greater good. 9) Refining is energy intensive. In this respect, a Fischer–Tropsch refinery has a benefit over a crude oil refinery, which must process more refractory materials [2]. In a Fischer–Tropsch facility, the energy intensity of the refinery pales in comparison to syngas generation, but that does not detract from the fact that the refinery is also energy intensive. 10) Commodity prices are volatile. This is an extremely important aspect for syncrude refiners. The raw material cost for Fischer–Tropsch synthesis is not necessarily linked to the crude oil price. In times of low crude oil price, it may be difficult to show a profit, but in times of high crude oil price, syncrude may have a significant cost advantage. 11) Refinery economics is complicated. This hardly comes as a surprise considering the complexity of refining. It also reflects a multitude of trade-offs. The main implication for refinery design is that a flexible refinery has a hidden economic benefit. In a Fischer–Tropsch facility, the ability to shift the production emphasis between fuels and chemicals, as well as between different fuel types, allows the refinery to respond to market and pricing changes. 24.2.2 Complex Systems and Design Rules
Refineries are complex systems and the refinery designer must respect this fact. This also implies that one must pay attention to the detail. This is quite clear from a superficial comparison between crude oil and Fischer–Tropsch refining. Refining technologies and strategies that work well with crude oil do not necessarily work well with Fischer–Tropsch syncrude. There are even differences between the syncrudes that may lead to completely different designs. This is also true for catalysis in the refining of syncrude [3]. This can be formally stated as the first rule of refinery design: Rule 1: The devil is in the details. In complex systems, one should be careful not to employ a purely mathematical approach to design. Not all refinery design decisions are subject to the two process engineering favorites: optimization and economic evaluation. Refinery optimization and economics are important, but these carry a risk that is typical of complex systems. Optimization of a refinery design tends to reduce refinery flexibility. If the refinery design is robust, it will be amenable to optimization later on, but an inherently poor design may appear viable through optimization. When this happens, the optimized conceptual design may look like a workable refinery design on paper, but it may be an untenable design in practice. This is something that should be avoided during the conceptual design phase. Rule 2: Avoid optimization during conceptual refinery design. The lessons that can be learnt from the wonderful book on Systemantics by Gall [4] should not be lost on refinery designers. Because refineries are complex systems, they are subject to all the peculiarities associated with complex systems. To quote Gall: ‘‘A complex system that works is
493
494
24 Principles of Refinery Design
invariably found to have evolved from a simple system that worked.’’ Applied to refinery design it means that a simple unoptimized refinery design must work for an optimized refinery to work. Rule 3: Any design must be as simple as practicable. A refinery design must also obey the laws of nature. The literature on refining is not perfect and experimental data are sometimes misleading due to analytical error or other happenstance. It is prudent to verify mass, atom, and energy balances. Conversion chemistry is not simple and no amount of wishful thinking will make it so. The chemistry defines the fundamental purpose of the conversion process. Whenever a technology is applied to achieve a goal that is contrary to its fundamental purpose, the design is likely to fail. The belief that it will work is called religious engineering and it may compromise the design. Rule 4: The design must be fundamentally sound. The behavior of complex systems is also closely linked to chaos theory, as illustrated by the well-known novels on chaos theory by Crichton [5, 6]: Jurassic Park and The Lost World. You cannot engineer a complex system and hope to anticipate or predict all the possible outcomes. In such a complex system as a refinery, there are just too many variables and interrelationships. One may be able to describe these interrelationships, but it is hard to predict the outcome beforehand. This can be formulated as the fifth rule of refinery design. Rule 5: Complex behavior may be accurately described, but it is difficult to predict. This rule holds some implications for refinery design and operation. If a complex system has been created directly and it is not based on a fundamentally simple system, its behavior may be unpredictable. This in turn creates operability problems. Even with a proper model, the response of the refinery can only be calculated after the fact. It is not possible to calculate all possibilities beforehand. The unintuitive behavior may have serious safety ramifications: ‘‘The mode of failure of a complex system cannot ordinarily be predicted from its structure’’ [4]. A simple mathematical example illustrates the potential impact of underestimating the behavior of complex systems. A Mandelbrot set is generated by plotting the number of iterations that is required for a simple equation (Equation 24.1) to reach a threshold criterion (Equation 24.2). zj+1 = (zj )2 + c zj > 2
(24.1) (24.2)
The value of c, which is a complex number, can be represented as x,y-coordinates with the real-value plotted on the x-axis and the imaginary value plotted on the y-axis. The Mandelbrot set is a three-dimensional picture (Figure 24.2), which is represented in two dimensions as a contour plot of j for various values of c. Despite the simplicity of the mathematical description, it is impossible to a priori predict the answer. Furthermore, even a small change in the value of c can significantly change the answer, thereby making even interpolation dangerous. There is beauty in Figure 24.2, but also a warning.
24.2 Refinery Design Concepts Figure 24.2 Mandelbrot set showing the significant amount of variation when the zoom level is increased. Clockwise from the top left, the values of c being plotted are (−2 + 1.5i to 1–1.5i), (−0.65–0.51i to −0.45–0.71i), (−0.625–0.668i to −0.605–0.688i), and (−0.615–0.683i to −0.6137–0.6843i).
Unless a refinery design is clearly in a stable region, small changes in feed composition or refinery operation may render the refinery inoperable or incapable of meeting product specifications. It is not advocated that refinery design should steer clear of a mathematical description, quite the contrary. However, it is advocated that conceptual refinery design must be as simple as practicable and make fundamental sense. As already cautioned, a mathematically optimum refinery design runs the risk of being brilliant only on paper. A distinction must be made between conceptual refinery design (Section 24.3) and real-world refinery design (Section 24.4). Last and perhaps the most difficult realization about complex systems is that a complex system hardly ever operates at a single design point. Although the design of a complex system is performed with absolute numbers, in reality there are no absolutes. The concept of steady-state operation is a fallacy. It is a convenient fallacy, because it allows us to calculate our design performance in terms of absolute numbers. This leads us to the last rule of refinery design. Rule 6: Calculate in absolutes, but believe in variability. Consequently, the robustness of a refinery design is not a ‘‘nice-to-have,’’ but an imperative. It is critical to understand how much change a design can tolerate and still be able to perform its designated function. 24.2.3 Refining Complexity
In order to strive for the lowest practicable refinery complexity, it is important to define what is meant by refinery complexity. Complexity is a measure of the interconnectedness of the refining units. A refinery with more units but less interunit transfers is less complex than a refinery with fewer units but more interunit transfers. In mathematical terms, this can be expressed using digraph theory [7]. The refinery can be represented as a digraph, with each refinery unit being counted as a node (nodej = 1). For this calculation, nodes are not counted if they are just sources (i.e., feed streams entering the refinery) or just sinks (i.e., products leaving the refinery). The least complex refinery unit (nodej ) has only one process stream entering the unit (indegreej = 1) and only one
495
496
24 Principles of Refinery Design
process stream leaving the unit (outdegreej = 1). The refinery complexity can then be expressed in terms of a complexity factor (CF), which is the average number of additional streams entering or leaving each refinery unit (Equation 24.3). (indegreej + outdegreej − 2) CF = (24.3) (nodej ) A refinery design that has a CF of zero has the least overall complexity possible. It is useful to include the blending operation as a separate unit in order to capture the complexity of product blending as part of the overall refinery complexity. The same calculation procedure may also be performed on each refining unit to determine the complexity of the technology, by taking each unit operation in the technology as a node. The CF has implications for refinery design. For the least complexity, it is required that the design must employ proper molecule management by ensuring that the feed to each separation and conversion unit is well matched to that unit. Molecule management heavily relies on proper and clean separation. Sloppy separation causes increased refinery complexity, because molecules that are routed to the wrong units must be rerouted to the correct unit, thereby increasing interconnectivity. It may also necessitate pre- or post-treatment steps that could otherwise have been avoided; this affects refining efficiency (Section 24.2.4). Because complexity is a measure of interconnectedness, it also has implications for refinery operability and stability. Refinery stability and operability increase with decreasing complexity, because an upset in one unit is less likely to affect the operation of other units in a complex way. The number of steps involved in the refinery is not reflected in the complexity calculation (Equation 24.3). The number of refining steps affects efficiency, but not complexity. 24.2.4 Refining Efficiency
Refinery efficiency can be expressed in terms of the average number of times that an atom visits a refining unit before it ends up as a final product. Essentially, it measures the number of times the work has to be performed on an atom before it can leave the refinery. Efficiency can be calculated in terms of an efficiency factor (EF) by adding the mass flow of every unit in the refinery (unit capacity) and dividing it by the total refinery capacity, which is the mass flow of all the products leaving the refinery (Equation 24.4): (unit capacity)j EF = (24.4) (refinery capacity) The same calculation procedure that is used in the case of complexity, may also be performed on each refining technology to determine its efficiency. There is another aspect of efficiency that is not captured by the EF, which is the amount of wastage. The green chemistry principles of atom economy and prevention of waste are important efficiency metrics. These values can be calculated in terms of carbon efficiency (Equation 24.5) and E-factor (Equation 24.6) [8], as measures of the wastage. (mass of C in products) (mass of C in feed) (mass of waste generated) E–factor = (mass of products generated) Carbon efficiency =
(24.5) (24.6)
24.3 Conceptual Refinery Design
24.3 Conceptual Refinery Design
Irrespective of the complexity inherent in refinery design, some basic steps can be identified that are common to all design approaches. In the development of any process configuration, three aspects are implicitly considered (Figure 24.1), which are as follows [9, 10]: 1) Feed description (input). This defines the nature of the feed material that must be refined. This refers not only to the Fischer–Tropsch syncrude, but also to the material coproduced during syngas generation, or the feed that is imported with the raw material employed for syngas generation. 2) Product description (output). This includes the product slate required and the specifications of each product. Intermediate products also have specifications, even if those specifications are self-defined. 3) Refining technology (transformation). This includes the different separation and conversion processes that can be considered for the refinery design. This draws on the know-how of the refinery design engineer. It also includes the supporting utility infrastructure that is needed by each technology. By knowing the feed and product description and having knowledge of the transformations, there is sufficient information for the development of a refinery configuration. With this level of detail, only conceptual studies are possible. It is nevertheless useful, since it allows the refinery configuration to be studied divorced from factors that influence real-world refineries. In this way, the limitations and sensitivities of different refinery configurations can be probed, without getting bogged down in the additional complexity introduced by issues related to the specific refinery location. It has been pointed out before that refinery design is essentially a creative process and that there are different tools and methodologies to guide the designing process. It is instructive to look at some of the approaches that can be followed for refinery designs. 24.3.1 Linear Programming
Linear programming can be employed to make a selection between different proposed refinery configurations based on specific design constraints [11]. The design constraints, such as the product specifications, are criteria that must be satisfied. Refinery configurations that cannot satisfy the design constraints are eliminated. Further discrimination between different designs can be made with respect to objective functions, such as the highest profitability, the lowest capital cost, and the maximum yield of a particular product. For linear programming to have value in discriminating between different refinery designs, it presupposes the accurate modeling of the various refinery units used in the proposed configurations. Poor representation of the detail in any unit risks violation of the first rule of designing complex systems. For example, when modeling a Fischer–Tropsch refinery with coal pyrolysis product co-feed, the content of cyclic material increases. Unless the models for units such as naphtha hydroisomerization and naphtha reforming incorporate cycloalkane sensitivity, the calculation will not reflect the impact of the change on octane number. Linear programming
497
498
24 Principles of Refinery Design
Address first (a) Feed characterization (b) Product specifications
Figure 24.3 Design hierarchy that indicates the order in which different aspects of a refinery design must be addressed.
Conversion processes Separation process Heat integration network
Address last Utilities
will then fail to recognize refinery configurations that exploit this change in composition. The value of linear programming is thus directly related to the quality of the model description. Linear programming is well suited for optimization and is capable of solving complex multivariable optimization problems. Applying it as tool in refinery selection is acceptable as long as the proposed refinery configurations are fundamentally sound and conceptually simple. Using it as a tool to fine tune the design process, violates the second rule of designing complex systems. 24.3.2 Hierarchical Design
A design hierarchy has been proposed for refinery debottlenecking, which can also be used to guide design (Figure 24.3) [12]. In the hierarchical design approach, the feed flow is used to determine design bottlenecks, and it can also be used for analyzing subsystems of the refinery. For example, hydrogen availability can be used to guide refinery design [13]. The hierarchical approach is a very logical approach and it formalizes the principles of refinery design. First, you have to know what you want to refine and what you want to refine it to – the input and output in Figure 24.1. Then, you can select the transformation steps. Each transformation requires feed within a specific range and produces a product that needs to be separated for blending and sale. Heat integration and the utilities that are ultimately required to make the design work are the last to be addressed. Depending on the level of detail needed in the conceptual refinery design, heat integration and utilities may not even be addressed. In essence, hierarchical design is akin to the Michael Jackson programming method [14], where the data flow ‘‘in’’ is transformed to ‘‘out.’’ This transformation determines the structure of the software design and by analogy the refinery design. 24.3.3 Technology Preselection
The method of technology preselection simplifies the design process by ruling out refining technologies and restricting the design to a limited number of preferred technologies. This
24.3 Conceptual Refinery Design
assists the design engineer by focusing further conceptual thinking on only the transformations that are allowed. The limited set of technologies may suggest specific configurations based on their respective feed requirements and product properties. For example, if only four naphtha conversion processes have been preselected, blending calculations will suggest the range of product ratios that yields on-specification products. This in itself may suggest a specific configuration in which each of these technologies may be employed. This approach was applied to perform a conceptual refinery design for future refineries that would have less of an environmental impact than typical current crude oil refineries [15]. The refining technologies were preselected on the basis of their environmental footprints, and the refinery design was subsequently performed on the basis of a logical ordering of the more limited set of refining technologies. Various selection criteria may be used. For example, political sanctions may prevent the licensing of technologies from the countries that imposed the sanctions. The refinery owner may also have specific preferences. Utility limitations can likewise rule out some technologies. Whatever the selection criteria, the creative design process is streamlined by avoiding configurations that will have to be rejected. In Fischer–Tropsch refining, the approach of technology preselection based on green chemistry principles is advocated [16]. The aim is to avoid the use of technologies that either inherently have a large environmental footprint or that perform poorly with Fischer–Tropsch syncrude. Inefficient refinery configurations due to feed-technology incompatibility are thus avoided. This method is especially helpful for refinery design engineers that are familiar with crude oil refinery design, because it allows the creative design process to focus only on technologies that will work well, or even have an advantage in converting syncrude. From the discussion in Chapter 16, it is clear that the list of preferred technologies for syncrude refining (Table 16.4) is quite different to that typically considered during conventional crude oil refining. 24.3.4 Carbon-Number-Based Design
All refining technologies are designed to process a specific carbon number (or boiling point) range as feed material. Conversion processes for motor-gasoline production, or chemicals, are generally speaking more feed sensitive than residue-conversion processes. This is reflected by the distillation cuts produced in a conventional crude oil refinery and an analogous situation exists in a Fischer–Tropsch refinery. The principle remains the same; namely, refining technologies have preferred feed ranges. Carbon-number-based refinery design allocates refining technologies to each carbon number based on the technology’s feed range. A refinery configuration can then be developed by conceptually routing all of the material through a conceptual carbon number separator, which then redirects the material to an appropriate conversion unit or allocates it as a final product (Figure 24.4). The products from each conversion either go to blending to become a final product, or are recycled to the carbon number separator. If one looks at a conventional crude oil refinery (Chapter 2), the atmospheric distillation unit and the vacuum distillation unit effectively produce some degree of carbon number separation based on boiling point. Conceptual carbon number design reflects reality. This type of approach has also been suggested for Fischer–Tropsch syncrude refining [17]. In a Fischer–Tropsch facility,
499
500
24 Principles of Refinery Design
Conceptual carbon number separation C2 C3 C4
Example: Technologies for C6 refining (principle is applied to each carbon number) Blending
Final product
C5 C6
Technology #1
C7 Syncrude or Crude oil
C8 C9
Technology #2
etc.
C10 C11−C15 C16−C22 C23−460 °C 460−550 °C >550 °C Aqueous
Figure 24.4 Carbon-number-based refinery design conceptually separates the refinery feed into carbon numbers or boiling ranges, for which each carbon number can be associated with the most efficient refining technologies. The products either go to blending to become final products or are recycled to the carbon number separator.
syncrude recovery in the Fischer–Tropsch gas loop enables a preseparation. This preseparation can easily be turned into a real carbon number separation for the lighter fractions, where carbon number separation matters most.
24.4 Real-World Refinery Design
Real-world refinery designs, as opposed to conceptual studies, have the aim of producing a practical refinery design for a specific purpose. A real-world refinery design is location specific. There are many factors influencing real-world refinery designs and when this added layer of detail is considered in the design process, the design becomes unique. It should be emphasized that beyond the conceptual stage, there is no such thing as a generic or even general refinery design, despite many real refinery designs being close to each other.
24.4 Real-World Refinery Design
In this section, some of the important factors affecting real-world refinery designs are discussed. It is shown that this is not just an added layer of detail that is superimposed on the conceptual refinery design, but that it dictates the design by influencing the feed and product description, as well as the process selection. The arbitrary feed, product, and process selection that can be made during conceptual design now becomes a requirement or a consequence of factors such as location, market, and politics. 24.4.1 Refinery Type
The selection of refinery type is a business decision and constitutes the primary design objective. There are three main classes of products from refining (Section 2.3): fuels, chemicals, and lubricants. It is also possible to combine these different products in a single mixed-type refinery that potentially have some economic benefits [18]. The decision remains the one that is made by the business as an investment decision and that fits into its overall business strategy. The conversion of a refinery from fuels to chemicals is quite feasible, as was demonstrated by the evolution and transformation of some Fischer–Tropsch refineries (Chapters 8 and 9). There are some risks involved when integrating chemical production with fuel production (Section 9.5), but potentially there are also many rewards. 24.4.2 Refinery Products and Markets
The types of products and the market into which each of the products will be sold, forms part of the business decision to build a specific type of refinery (Section 24.4.1). The identification of a market gap or strategic positioning within a specific market may drive this investment decision. The products and markets should be considered together, since the products determine what will be made, while the markets determine the product specifications. For example, a refinery producing fuels for the North American market looks different than a refinery targeting the European market and both considerably differ from a refinery supplying the African market. In principle, all refinery designs should aim to meet a future demand in a market requiring future fuel specifications. Anticipating changes in fuel specifications is therefore not merely a mental game, but could have a huge impact on the cost and complexity of a refinery design. This places the discussion on future trends in fuel specifications (Sections 13.5, 14.4, and 15.4) into perspective. It is inevitable that there will be a number of years that will pass between the design, construction, and commissioning of a refinery. During this period, there may be changes in the market, because predictions about the future are inherently imprecise. The importance of refinery design flexibility cannot be overemphasized, since it is almost inevitable that even a new refinery may have to be modified. It is almost guaranteed that a refinery will have to be modified in some way during its existence to keep up with changes in fuel specifications. Understanding the market is sometimes much more important than adhering to, or anticipating specifications. This is especially true of speciality markets for petrochemicals, but may also be true of fuels. A classic example can be found in the history of the oil industry [19]. When Marcus Samuel (founder of Shell) in the 1890s took on John D. Rockefeller (of Standard Oil) in the kerosene market in the Far East, his plan hinged on the efficiency of shipping kerosene
501
502
24 Principles of Refinery Design
in bulk. It was cheaper to ship the kerosene in bulk, rather than in tins that were packed in wooden crates as was the case with Standard Oil’s product. Both the tins and wooden crates added to the cost of getting the kerosene to the market. However, when the kerosene was placed in the market at half the cost of Standard Oil’s, nobody wanted it. It turned out that the tins played a vital role in the Far Eastern way of life at that time, since the tins could be reused for the manufacturing of household items like buckets, cups, and much more. The tins were therefore more important to the market than the kerosene itself. A container factory had to be built in order to sell the kerosene. Product preference and markets can also be influenced by contractual obligations. These obligations can be in the form of a supply agreement that necessitates the production of one or more products, as well as restrictions that may lock the refinery into a specific market, or prevent the refinery from selling its products in a market. Such agreements and obligations are a double-edged sword. The following example, which is related to syncrude refining, illustrates the point. In South Africa, there was a supply agreement between the synthetic fuels producer Sasol and the crude oil refiners in the country. According to this supply agreement, the crude oil refiners were required to purchase a certain minimum amount of synthetic fuels, irrespective of the market conditions, thereby guaranteeing a market for the product. In exchange, Sasol could not enter the retail market on its own and a limitation was placed on the volume of synthetic fuels that could be produced. Among others, this affected the decision to increase chemical production from syncrude (Section 9.5). The supply agreement was terminated at the end of December 2003 at the request of Sasol. When this happened, the production volume of synthetic fuels was no longer restricted, but the synthetic fuels were no longer guaranteed of a market either. With no guaranteed market, it was found that the products produced by the Sasol Synfuels refinery were not properly aligned with the local market. It left the synthetic fuels refinery exposed to fluctuations in the retail market for the first time since its construction. Furthermore, since the fuel price in South Africa is regulated, it was not possible to market any surplus fuel at a lower retail price. Refinery operation had to respond to these changes in the supply and demand and had to respond in direct competition with the crude oil refiners. 24.4.3 Refinery Feed Selection
In a crude oil refinery, the feedstock, or basket of crudes that is selected, has a significant impact on the profitability of the refinery. The range of crudes that can be processed is determined during the refinery design phase and flexibility to deviate from the design basis is determined by processing constraints. For example, if the atmospheric and vacuum distillation units were designed for Arabian Light (54% lighter than 350 ◦ C material), it is quite possible to exchange it with a crude such as Iranian Light (55% lighter than 350 ◦ C material), but not with Algerian Hassi Messaoud (75% lighter than 350 ◦ C material) or South American Bachequero (30% lighter than 350 ◦ C material) [10]. The price differential between low-quality heavy crudes and good-quality light crudes can easily exceed US$10 per barrel. However, the refining infrastructure and operating cost required to process poorer crudes may not always justify the selection of cheaper feed [20]. The nature of the refinery feed determines the refining effort to produce the required product slate. The feed selection must match the product requirements. For example, it would make better
24.4 Real-World Refinery Design
sense to select Venezuelan Tia Juana or Bachequero crudes that naturally yield good lubricating oil properties as feed for a lubricant refinery, rather than selecting a crude oil that requires significant processing to achieve the same result. This is true for Fischer–Tropsch syncrudes too, where the syncrude selection affects the ease of refining to produce specific products. For example, it makes more sense to employ low-temperature Fischer–Tropsch (LTFT) syncrude for lubricating oil production by hydroisomerization than high-temperature Fischer–Tropsch (HTFT) syncrude. It may also happen that exploiting a specific feed material drives the decision to build a refinery. This is often true of facilities that do not make use of crude oil. When the decision is motivated by energy security, the feed material will typically be selected to match the national resource base. Other factors, including political expediency, may also suggest the use of specific feed materials. In such cases, the feedstock is preselected and is the driver for the refinery design, rather than a marketing opportunity or specific product demand. A word of caution is prudent. When a refinery is designed on political preference or based on a trend, it is critical to ensure that the design is robust enough to weather a reversal in political opinion or change in trend. For every trend, there is a potential trend killer. It may be easy to spot the trend, but not always so easy to spot the trend killer. For example, with instability in oil supply, coal-to-liquid (CTL) technology becomes a viable alternative means of energy security, but a potential trend killer is the perceived link between CO2 emissions and climate change. It does not matter whether the correlation between CO2 and climate change is causal, or whether it is just two variables that are both correlated with time. If it is a political reality, then a reversal in political opinion will cause the once favored feed material to become disfavored. In a Fischer–Tropsch refining environment, the feed material has less of an impact on the refinery design. Irrespective of whether it is a biomass-to-liquid (BTL), CTL, gas-to-liquid (GTL), or waste-to-liquid (WTL) facility, the feed is first converted into synthesis gas. The indirect conversion makes all feed materials equivalent with respect to the refinery, unless they are associated with coproducts such as associated natural gas liquids and pyrolysis oils. It is therefore the choice of Fischer–Tropsch technology that determines the properties of the refinery feed that has to be refined, and not the raw material used as feed for syngas generation. 24.4.4 Refinery Location
The importance of selecting the location of the refinery is like the three P’s of property investment – position, position, and position. In general, refineries are located close to the source of the feed material, or on trade routes with easy access to the feed material. Most crude oil refineries are consequently situated on the coast with easy access to shipping for supply of the feed and export of the products. Inland refineries are less common and require either a local feed source or access to a pipeline for supply. In the case of Fischer–Tropsch based refineries, it has thus far been the practice to situate the facility close to the feed source. The locations of commercial Fischer–Tropsch facilities are indicated in Figure 24.5. Solid-material handling is generally more difficult and expensive than fluid handling, making it almost imperative to situate CTL and BTL refineries close to the source of the feedstock. To a lesser degree, the same applies to GTL refineries in the absence of pipeline
503
504
24 Principles of Refinery Design
Escravos Escravos (GTL)
Ras Laffan Oryx (GTL) Pearl (GTL) Bintulu Shell (GTL)
Secunda Sasol Synfuels (CTL) Mossel Bay Petro SA (GTL)
Figure 24.5
Sasolburg Sasol 1 (GTL)
Location of Fischer–Tropsch facilities.
infrastructure, since GTL facilities need to compete with liquefied natural gas (LNG), which is effectively the transportation alternative. Yet, in selecting a location there are other aspects to consider that may outweigh the proximity to the feed material. The location directly impacts on the refinery design and operation in terms of design details and cost. Some location-specific aspects to consider are as follows: 1) Climate. This determines the insulation, heating, and cooling requirements. In extreme climates, such as those found in the Arctic or desert regions, special measures are required to deal with either heating or cooling. Climate-dependent measures transcend the equipment design. For example, the lubricating oils for rotational equipment need to be suitable for the climate. The material of construction is also influenced by the climate, with atmospheric corrosion being generally higher in coastal regions than in inland locations. The spate of hurricanes in 2005 that affected the Gulf of Mexico also illustrated the potential impact of extreme weather phenomena on refineries [21]. 2) Geology. The geology of the refinery site may require special measures to be taken in site preparation and construction. Examples of such measures include the strengthening of foundations for protection against an earthquake and the use of deep foundations in marshy grounds to ensure that construction is supported by bedrock. 3) Natural resources. The lack of sufficient water or water of acceptable quality to make use of a standard cooling water design may add to the refinery cost. An example of this is the refineries in desert regions that employ salt water cooling systems. 4) Environmental sensitivity. Working in a sensitive ecology can markedly affect construction and operating practices. This is also true for facilities that are situated close to communities. It may be required to use quiet rotating equipment, ensure that emissions are lower than legal limits, and invest in plant beautification. Communities that are not properly informed may even rebel at perceived emissions, such as steam. Construction in such areas needs to be planned very carefully, since future expansion might be limited. End-of-life site remediation and the environmental impact of operations may dictate some design
24.4 Real-World Refinery Design
5)
6)
7)
8)
9)
decisions, such as on-site effluent treatment and investment in inherently low-emission technologies only. Utility access. Utilities may not be readily available at the location of the facility. In areas where potable water supply and sewerage works are not available, construction of these utility systems may have to be undertaken as part of the facility. In locations where the power grid is already taking a strain, it may be prudent to invest in on-site electricity generation. Dependence on third-party utility supply is always associated with some risk. Although contracts may cover production loss due to supply interruption, it will not cover equipment damage. In regions where there is a risk of utility interruption (including climate-induced utility interruption), the refinery design needs to take the effects of unplanned shutdowns into account and consider the recommissioning time during optimization. Tight heat integration may make a refinery more energy efficient, but may also prolong the time needed for recommissioning after a shutdown. Location factor. Cost estimators use the location factor as a measure to indicate the impact of the location on the construction cost. The location factor tends to be of greater significance for inland locations and for locations far from a significant skill base. If the refinery is not situated close to a commercial harbor with the infrastructure to off-load large vessels, the supply and transport of equipment can become an issue. Unit sizes may have to be restricted to facilitate road transportation and it will also impact the construction schedule. Although this seems like a once-off impact, it is not. Siting a refinery in a remote location where living conditions are not considered desirable can result in much higher operating cost too. It may be necessary to pay more for labor (need for ‘‘location allowances’’), work force productivity may be lower, and it may also be more difficult to attract and retain skilled personnel. This can be a very serious consideration. For example, a marked performance deterioration was seen at the Sasol 1 facility when a key personnel was transferred to the Sasol 2 and 3 facilities [22]. Legislation. Refinery design is subject to various laws. Environmental legislation may require investment in technologies to meet specific emission limits, require CO2 sequestration, or even prevent the use of some technologies, such as HF-catalyzed aliphatic alkylation. The intellectual property protection provided is likewise an important consideration, since inadequate licensor protection might limit the basket of technologies that licensors are willing to license for the refinery design at that location. Legislation also governs operation and profitability through labor laws, tax laws, competition laws, as well as religious and public holidays in the country where the refinery is sited. Politics and governance. The local, national, and international politics of the region could have an impact on the refinery design. Preferred suppliers, sanctions, and boycotting of suppliers affect the technology selection. Global politics may also influence the refinery design. For example, the US government placed restrictions on the supply of technology to some countries, which implies that a refinery design will be limited in its technology selection. Other issues that may affect refinery construction and operation include security, crime, labor unrest, government integrity, and local bribery practices. Marketing logistics. Although the product and market selection can be performed independent of the refinery site selection (product can be exported), the latter will impact the refinery design. The location imparts a location advantage or disadvantage for the different products. For example, a fuels refinery close to an international airport has a location
505
506
24 Principles of Refinery Design
advantage for jet fuel production, which is likely to be more profitable than producing diesel fuel with its more diffuse market distribution. In such a case, a refinery design favoring jet fuel over diesel fuel production would make more sense. The refinery configuration is not only affected by the market for final products. The configuration may also be affected by the market for intermediates and blending components. When a refinery is close to other refineries and petrochemical producers, business agreements can be put in place to simplify the refinery design. By interrefinery exchange of intermediate products, which cannot be sold to consumers, final products can be prepared by blending. In this way, refineries can decrease the capital investment required to refine products to specification. It may also be beneficial to deliberately sell a specific product as an intermediate, rather than refining it. However, transportation of intermediates over long distances can present problems. Depending on the product, the transportation routing may be dictated by local legislation governing hazardous substances. For example, transportation of carcinogenic substances such as benzene may be prohibitively expensive and routing benzene through some countries may even be illegal. Reactive intermediates may also degrade during transportation and require repurification closer to the market. For example, it is best to repurify n-1-alkenes that are extracted from Fischer–Tropsch syncrude after long distance transportation. 10) Intellectual property. Unless patent protection for a technology was filed very widely, some countries have likely been omitted. The presence or absence of relevant patents in a specific country may either create an opportunity or provide an obstacle to the design. A competing company with strong patent portfolio in a country may hinder the design of an efficient refinery by negating the use of key technologies. 24.4.5 Secondary Design Objectives
The importance of secondary design objectives should not be underestimated, since it is these objectives that influence the details of the design. The secondary design objectives are set by the business to guide the refinery design in order to meet the current financial situation, set strategic direction, and reflect the corporate culture. The corporate culture is an important influence and it may invite inventiveness, stifle creativity, foster environmental responsibility, or just be greedy. The secondary design objectives are sometimes erroneously called a ‘‘wish list’’, but these are more often than not a ‘‘must list’’ for the project. There are many possible secondary design objectives, of which only some are discussed. Examples of typical secondary design objectives are as follows: 1) Minimize capital expenditure. This directive is necessary when the gearing of the company is high and additional capital cannot be raised through share issues. The impact of this on a refinery design is to invest in the least expensive configuration that will meet the primary design objective. Typical side effects of spending the least amount of capital are reduced flexibility (more units operating at or close to nameplate capacity), increased operating cost (more labor-intensive units with less automation), reduced on-line availability (less redundant equipment, such as spare pumps), and increased maintenance cost (less expensive materials of construction and lower quality of equipment).
24.4 Real-World Refinery Design
2) Maximize net present value (NPV). In a refinery, there is a limited refining margin available to generate profit in a regulated environment. Any capital spent on the refinery to keep up with changing fuel specifications by definition has a negative net present value (NPV). Such expenditures are seen as part of the cost of staying in business. The NPV can only be positively affected if the yield of final products per unit volume of feed is increased, a cheaper feed stock can be used to produce the same products, or the product slate is changed to contain more high-value products. In this respect, the options open to the refinery designer are limited by the restrictions placed on the refinery type, feed selection, and product slate. In general, the highest NPV can be obtained by producing more nonenergy products, such as chemicals. 3) Smallest environmental footprint. It is wrong to state that more environmentally friendly technologies cost more, although superficially speaking they generally do. The environmental impact of refining technologies can (and should) be considered in refinery technology selection (Table 16.4). Often there is more than one way to achieve a specific outcome during refinery design. In the author’s opinion, environmentally friendly design is part of responsible engineering practice and it should always be a design objective. 4) Maximum liquid product volume. In fuels refining, this strategy forms the basis for maximizing refinery profitability. Fuels are sold by volume, and an increase in the volume of final products per unit volume of feed usually translates into increased profits. This will drive refinery technology selection away from conversion processes that reduce the liquid volume of products and will tend to drive refinery design to lighter products, such as motor-gasoline, unless there is a significant negative price differential with heavier products. 5) Refinery flexibility. Energy markets are cyclic, as typified by the ‘‘summer driving season’’ and ‘‘winter heating market.’’ The same holds true for chemical markets. Refineries are exposed to these variations. It is a business decision to either invest in the capital necessary for refinery flexibility or to sell the production surplus at discounted prices. Investing in refinery flexibility has other potential benefits too, but these are not immediately reflected on the balance sheet. It will be easier to meet future specification changes (smaller chance of specific units constraining refinery upgrades), plant upsets are easier to deal with (rerouting of streams is possible), and plant shutdowns are less constraining (more capacity to work away product). Added refinery flexibility may come at the cost of increased complexity, but this is not always the case. 6) Least refinery ‘‘complexity’’. The most elegant refinery designs are those where the refinery has the most flexibility with the least complexity and highest efficiency. Cost estimators define ‘‘complexity’’ in terms of the type and/or number of units in a plant or refinery, which can then be used to estimate the capital cost [23, 24]. This may be a sensible definition for purposes of capital cost estimation, but it does not capture the interdependency of units or the actual amount of work that has to be performed (operating cost) to make a final product. A distinction must be drawn between complexity (Section 24.2.3) and efficiency (Section 24.2.4). It is therefore important to clarify as to what is meant by ‘‘complexity’’, since different meanings will affect the design in different ways. 7) Shortest time to completion. The time pressure on a project schedule is driven by economics, which can be linked to the time-value of money, and sometimes a transient window of opportunity. Whatever the business reason, the construction schedule of a refinery can only be reduced by selecting commercial refining technologies with a low construction
507
508
24 Principles of Refinery Design
complexity, which is not related to refinery complexity. Technologies requiring specialized manufacturing, exotic construction materials, and highly skilled artisans to assemble are automatically disqualified from the design. 8) Contractual obligations. Agreements, such as joint ventures, may lock the refinery design into the use of specific technologies.
References 1. Gary, J.H., Handwerk, G.E., and Kaiser, M.J.
2.
3.
4. 5. 6. 7.
8. 9. 10.
11.
12.
13.
(2007) Petroleum Refining. Technology and Economics, 5th edn, CRC Press, Boca Raton, FL. De Klerk, A. (2007) Environmentally friendly refining: Fischer-Tropsch versus crude oil. Green Chem., 9, 560–565. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Gall, J. (1979) Systemantics, Fontana/Collins, Glasgow. Crichton, M. (1991) Jurassic Park, Arrow, London. Crichton, M. (1995) The Lost World, BCA, London. Skvarcius, R. and Robinson, W.B. (1986) Discrete Mathematics with Computer Science Applications, Benjamin-Cummings, Menlo Park, CA. Sheldon, R.A. (2007) The E factor: fifteen years on. Green Chem., 9, 1273–1283. Wauquier, J.-P. (ed.) (1995) Petroleum Refining, Technip, Paris. Jones, D.S.J. and Pujad´o, P.R. (eds) (2006) Handbook of Petroleum Processing, Springer, Dordrecht. Maiti, S.N., Eberhardt, J., Kundu, S., Cadenhouse-Beaty, P.J., and Adams, D.J. (2001) How to efficiently plan a grassroots refinery. Hydrocarbon Process., 80 (6), 43–49. Zhang, J., Zhu, X.X., and Towler, G.P. (2001) A level-by-level debottlenecking approach in refinery operation. Ind. Eng. Chem. Res., 40, 1528–1540. Alves, J.J. and Towler, G.P. (2003) Analysis of refinery hydrogen distribution systems. Ind. Eng. Chem. Res., 41, 5759–5769.
14. Bell, D., Morrey, I., and Pugh, J. (1987) Soft-
15. 16.
17.
18.
19.
20.
21. 22.
23.
24.
ware Engineering. A Programming Approach, Prentice-Hall, Englewood Cliffs, NJ. Chen, N.Y. (2001) An environmentally friendly oil industry? Chem. Innov., 31 (4), 10–21. De Klerk, A. (2008) Fischer-Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. Dancuart, L.P., De Haan, R., and De Klerk, A. (2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532. Chadwick, J.L. (1977) Economics of Chemical Refineries, SRI Process Economics Program Report 107, Stanford Research Institute, Menlo Park, CA. Tugendhat, C. and Hamilton, A. (1975) Oil. The Biggest Business, Eyre Methuen, London, pp. 48–49. Snow, N. (2004) Are upgrades to process lower-quality crudes worth the investment? World Refin., 14 (8), 52. Tippee, B. (2006) Previews of 2006 reflect new variable: hurricanes. Oil Gas J., 104 (1), 20–28. Wessels, P. (1990) Crescendo tot Sukses. Sasol 1975–1987 (Engl. Transl. ‘‘Crescendo to success. Sasol 1975–1987’’), Human & Rousseau, Cape Town. Zevnik, F.C. and Buchanan, R.L. (1963) Generalized correlation of process investment. Chem. Eng. Prog., 59 (2), 70–77. Viola, J.L. (1981) Estimate capital costs via a new, shortcut method. Chem. Eng., April 6, 80–86.
509
25 Motor-Gasoline Refining 25.1 Introduction
Motor-gasoline is the most demanding of all the transportation fuel types to refine. One only has to look at the complexity of a generic fourth-generation crude oil refinery (Figure 2.13) to realize that any fuel refinery that includes motor-gasoline production will require a number of conversion units just for motor-gasoline production. Considering this complexity, it is understandable that there has been a trend not to produce final on-specification transportation fuels in the gas-to-liquids facilities constructed since the year 2000 [1]. There is a natural distribution of transportation fuel types in a refinery that is determined by the composition and carbon number distribution of the syncrude. In the refinery, these attributes may be manipulated to produce a product distribution closer to what is desired from a marketing perspective. This is where the secondary refinery design objectives (Section 24.4.5) play a major role in guiding the conceptual design. A number of different aspects are explored in the design of motor-gasoline refineries to illustrate how the secondary objectives affect the design process: 1) Low-complexity refinery design, which is somewhat akin to low capital cost design, but not necessarily so. This highlights the inherent strengths and weaknesses of syncrudes for the production of motor-gasoline. 2) Robust and flexible refinery design, which allows changes in the product distribution during blending. It is valuable when considering the operability and future operation of the refinery. Although robustness and flexibility go hand in hand, there is a difference between the two. A robust design allows operation away from the design point (nonideal operation), whereas flexibility allows the design point to be changed, thereby redefining how the refinery should operate. 3) Maximum motor-gasoline refinery design, which pushes the motor-gasoline production to the extreme. This type of design has been explored before [2]. It highlights the limitations involved in producing motor-gasoline from syncrude. This is the first of three chapters that deal with the main transportation fuel types, namely motor-gasoline, jet fuel (Chapter 26), and diesel fuel (Chapter 27). In order to avoid repetition, each chapter focuses on only one fuel type. This is an artificial segregation and in any fuels refinery one would have to consider all fuels together. The motor-gasoline refinery designs will therefore include the other fuel types, but the design decisions leading to a specific jet fuel Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
510
25 Motor-Gasoline Refining
or diesel fuel configurations are discussed only with respect to their impact on motor-gasoline production.
25.2 Gap Analysis for Syncrude to Motor-Gasoline
The overall refining methodology for refining syncrude to motor-gasoline is the same as for the other transportation fuels and in practical terms two tasks must be accomplished in parallel: 1) The carbon number distribution must be adjusted to produce more material in the motor-gasoline range, preferably from material that is either normally gaseous or a residue. In maximum motor-gasoline refining, one would extend this to include material from the other transportation fuel types. 2) The molecular composition must be manipulated in an efficient way to ensure that the motor-gasoline meets fuel specifications. This is a quality adjustment and is specific to the needs of a particular design. Since motor-gasoline is a mixture, there is considerable leeway in the choice of molecules that will be used to meet the fuel quality requirements. In order to determine what must be accomplished during the refining of syncrude into motor-gasoline, a gap analysis is performed. The essential elements in a gap analysis are the feed and product properties. The product property requirements are defined by the motor-gasoline specifications, whereas the syncrude feed properties can be described in terms of the carbon number distribution and composition. 25.2.1 Motor-Gasoline Specifications
The properties, composition, and specifications of motor-gasoline have already been discussed in detail (Chapter 13). It was pointed out that motor-gasoline specifications are location specific. Some general features can, nevertheless, be identified from different specifications (Tables 13.1 and 13.3). In order to keep the discussion general, some generic motor-gasoline specifications have been selected as quality target for the motor-gasoline refinery designs (Table 25.1). The European EN228:2004 standards (Table 13.1) were used as benchmark. The most important deviation from this specification is the lower limit on benzene content (0.6 vol%), which is in line with the proposed change by the Environmental Protection Agency (EPA) [3]. In a society that is worried about unproven claims about the effect of chemicals at part per million levels in some products, accepting even 0.6 vol% of a known carcinogen in a commodity consumer product is surprising. The variability in oxygenate content, mainly spurred by political pressure, will be a sensitivity that is considered in the refinery designs. The use of an antiknock index ( 12 ·RON + 12 ·MON), rather than a strict individual RON and MON specification, will likewise be considered as a sensitivity.
25.2 Gap Analysis for Syncrude to Motor-Gasoline Motor-gasoline specifications employed for syncrude refinery designs.
Table 25.1
Fuel property
Fuel specification
Restrictions on boiling range Density at 15 ◦ C (kg·m−3 ) Reid vapor pressure (kPa) Restrictions on composition RON MON Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (mass% O) Sulfur content (µg·g−1 ) Metal content
Minimum
Maximum
720 –
775 60
95 – 85 – – 18 – 35 – 0.6 – 2.7 – 10 No intentional addition
25.2.2 Carbon Number Distribution
The amount of straight-run syncrude that is in the correct boiling range for motor-gasoline production is determined by the chain growth probability (α-value) of Fischer–Tropsch synthesis. An estimate of the fraction of straight-run naphtha (C5 –C10 ) can be obtained from the Anderson–Schulz–Flory distribution for different chain growth probabilities (Figure 25.1). Yield of straight-run naphtha (mass%)
60 50 40 Industrial HTFT operation
30 20 10
Industrial LTFT operation
0 0.4
0.5
0.6 0.7 0.8 Fischer – Tropsch chain growth probability
Figure 25.1 Yield of straight-run naphtha (C5 –C10 ) in the C3 and heavier hydrocarbon fraction from Fischer–Tropsch synthesis as calculated from the Anderson–Schulz–Flory distribution.
0.9
1
511
512
25 Motor-Gasoline Refining
It indicates that there will be less refining work required to maximize motor-gasoline production from high-temperature Fischer–Tropsch (HTFT) syncrude compared to low-temperature Fischer–Tropsch (LTFT) syncrude. However, it gives no indication of the amount of refining work that is required to convert lighter and heavier material into the naphtha boiling range to increase motor-gasoline production. The easiest way to convert gaseous material into liquid products is by acid-catalyzed alkene conversion. In this respect, HTFT syncrude has an advantage, because the gaseous C2 –C4 material is alkene rich. Less hydrogenated LTFT syncrudes, such as slurry-bed-derived Fe-LTFT syncrude, are in turn easier to convert than Co-LTFT and fixed bed Fe-LTFT syncrude. Depending on the nature of the Fe-LTFT catalyst, Fe-LTFT catalyst deactivation may be beneficial for refining [4], because the α-value decreases and the alkene content of the C2 –C4 material increases. Heavier material can be cracked to produce naphtha range material. It is easier to crack aliphatic than aromatic residues and LTFT syncrude has an advantage over HTFT syncrude. 25.2.3 Composition and Quality
The naphtha that has to be converted into motor-gasoline comes from two sources, namely, straight-run Fischer–Tropsch naphtha and naphtha produced during other conversion processes in the refinery. The latter is a design outcome and here we consider only the composition and quality of the straight-run naphtha. Refining challenges that can be identified on the basis of the composition of straight-run Fischer–Tropsch syncrude are as follows: 1) Motor-gasoline specifications limit the maximum content of all compound classes, except the alkanes, which are the key to good quality motor-gasoline (Figure 25.2). The alkenes, aromatics, and oxygenates are usually high octane number compounds, whereas the octane numbers of the alkanes depend strongly on the degree of branching (Table 13.4). However, straight-run Fischer–Tropsch naphtha contains a large fraction of linear hydrocarbons, mainly n-alkanes and n-1-alkenes, and none of these have high octane numbers. Historically, alkene-rich Fischer–Tropsch naphtha fractions could be refined by double bond isomerization (Section 16.4.1), but with the restriction on alkene content, this approach is of limited value. It is clear that the alkane quality is key to motor-gasoline refining, and one or more refining technologies to produce high octane number alkanes will be required in any successful design. No limit
Alkanes
Figure 25.2
Maximum content in motor-gasoline 18%
35%
15%
Alkenes
Aromatics
Oxygenates
RON ~95 MON ~80
RON ~110 MON ~100
RON ~115 MON ~100
Alkane quality as key to motor-gasoline refining.
25.2 Gap Analysis for Syncrude to Motor-Gasoline
Volume of straight-run naphtha (%)
40 Fischer–Tropsch chain growth probability
HTFT
35
a = 0.6 a = 0.7 a = 0.8 a = 0.9
30 25 20 LTFT
15 10 5 0 4
5
6
7
8
9
10
11
Carbon number Figure 25.3
Bias towards lighter products in straight-run Fischer–Tropsch naphtha.
2) The Anderson–Schulz–Flory carbon number distribution of the straight-run naphtha fraction can be expressed in terms of volume (Figure 25.3). This shows that syncrude is biased toward lighter products and that this bias decreases with increasing α-value. This has two implications for the production of motor-gasoline. Firstly, the vapor pressure of the straight-run HTFT motor-gasoline will be high and it is likely to become a constraining fuel specification. Secondly, it will be necessary to add heavier naphtha range material to the straight-run HTFT motor-gasoline in order to balance the motor-gasoline distillation profile. 3) The straight-run syncrude is rich in acyclic aliphatic material. The combined effect of the syncrude composition and bias toward lighter naphtha range material is that the straight-run product has a density that is below the lower limit of the motor-gasoline density specification. The calculated density of the straight-run naphtha fraction based on acyclic aliphatic compounds only increases from 675 to 700 kg·m−3 as the α-value increases from 0.6 to 0.9. In the case of HTFT syncrude, the straight-run naphtha fraction contains aromatics, which increases the density by 5–10 kg·m−3 . The oxygenates present in HTFT and LTFT syncrude also contribute to increase the density somewhat. It is clear that some aromatics will have to be produced in order to increase the motor-gasoline density. 4) A large proportion of the straight-run alkenes present in the Fischer–Tropsch syncrude is n-1-alkenes. The n-1-alkenes have poor octane numbers compared to the rest of the alkenes and require some refining. In the past, this was accomplished by double bond isomerization over an alumina-rich catalyst (Section 16.4.1) and it allowed HTFT syncrude to be easily upgraded to motor-gasoline quality. With the limitations on alkene content in motor-gasoline, this is no longer possible. When straight-run syncrude is to be included in the final motor-gasoline blend, double bond isomerization and deoxygenation over an alumina catalyst is still recommended, but it is no longer a major refining pathway. 5) There are too little aromatics in the motor-gasoline. Naphtha reforming played a key role in the refining of crude oil ever since motor-gasoline became an important product (Section 2.4.2). From second-generation crude oil refineries onwards, naphtha reforming became a fixture in
513
514
25 Motor-Gasoline Refining
the refinery design. In early Fischer–Tropsch refineries, the alkenes fulfilled the same role as aromatics. However, with the restriction on alkene content, more demanding octane number requirements, and phaseout of tetraethyl lead, this design approach is no longer available to the Fischer–Tropsch refiner. Motor-gasoline refining consequently requires additional aromatics production. 6) Although the inclusion of some oxygenates are allowed, not all the oxygenates in syncrude are desirable components for motor-gasoline. The carbonyl compounds are susceptible to gum-forming reactions, such as aldol condensation. The carboxylic acids in the motor-gasoline boiling range are too corrosive for inclusion in the motor-gasoline. It will be necessary to devise a refining strategy that deals with not only the oxygenates but also the boiling point broadening caused by converting the oxygenates. Historically, alumina-catalyzed double bond isomerization and deoxygenation served this purpose. Hydrotreating can be considered as an alternative. The decision depends on the need to retain alkenes after the deoxygenation step or not.
25.3 Decisions Affecting Motor-Gasoline Refining
There are a number of issues specific to Fischer–Tropsch syncrude refining that must be decided during the planning stages of the design. These design decisions will all have an impact on the design and affect the carbon efficiency, environmental footprint, robustness, flexibility, and complexity of the refinery. Motor-gasoline refining is usually affected the most. Many of the petrochemicals are in the naphtha boiling range, and the C2 –C4 hydrocarbons have direct bearing on refining technologies for motor-gasoline production. The oxygenates in the Fischer–Tropsch aqueous product are likewise in the naphtha boiling range and some can be considered for fuel blending, or can be converted into products that can be refined to motor-gasoline. 25.3.1 Chemicals Coproduction
If one looks at the composition of Fischer–Tropsch syncrudes, it is easy to spot many molecules and boiling fractions that are present in percentage levels that are also petrochemicals. For example, ethene, propene, ethanol, acetic acid, n-1-alkenes, and waxes are all petrochemical products (Chapter 28). Most of these petrochemicals are in the motor-gasoline boiling range or affect motor-gasoline refining. A conscious decision must be made whether to extract petrochemicals or not. A conscious decision must also be made about the extent of petrochemical extraction and whether the petrochemicals are considered permanent products. If chemicals are to be extracted, the general observations about the impact of chemical extraction on fuel refining have to be considered (Section 9.5). For convenience, the essence of the main points that were made are repeated below: 1)
Increasing chemical extraction increases refining complexity and potentially comes at the cost of decreased refining flexibility and operability.
25.3 Decisions Affecting Motor-Gasoline Refining
2) When the chemicals that are extracted are low-quality fuel components, the benefit of increased fuel quality derived by chemical extraction can be realized only if extraction is guaranteed. 3) Tightly integrating chemical extraction and fuel refining reduces overall refinery stability. 4) Fuel blending becomes more sensitive to changes, because a increase in chemical production leads to a decrease in the volume of fuel that is available for blending. 5) Some refining pathways are closed when extracting material as chemicals. 6) Large-scale chemical extraction affects the fuel matrix and thereby the blending properties of the fuel. When the facility is far from petrochemical markets, it may be more difficult to justify extraction of chemicals such as ethene and propene, but liquid products are more easily transported. The temptation to extract chemicals due to their higher value than fuels will always be there. Even when the decision is made to produce fuels only, it is only prudent to keep the potential future extraction of chemicals in mind during Fischer–Tropsch refinery design. Fischer–Tropsch facilities have a history of chemical coproduction. 25.3.2 Fate of C2 –C4 Hydrocarbons
The C2 –C4 hydrocarbons are naturally gaseous products. The design of the gas loop and specifically the design of the syncrude recovery section of the gas loop (Section 5.3) determine the degree of separation and recovery. The C3 –C4 hydrocarbons can be recovered by condensation under pressure at ambient conditions, but recovery of the C2 hydrocarbons requires cryogenic separation. In the absence of a clear incentive to recover the C2 hydrocarbons, the added complexity and cost of cryogenic tail gas separation may at first glance seem unattractive. Yet, considering that the capital cost of a Fischer–Tropsch refinery is only 10–15% of the total capital cost of a Fischer–Tropsch facility, it does not make sense to use 85–90% of the capital cost to produce a product and then not refine it. When one views the cost in this way, cost saving in the refinery at the expense of carbon efficiency is ridiculous. The extent of C2 –C4 hydrocarbon recovery in the gas loop has far-reaching consequences for the Fischer–Tropsch refinery design. The fate of the C2 –C4 hydrocarbons and their implications are as follows: 1) The carbon efficiency of the facility is directly influenced by the extent of tail gas recovery. For example, considering the typical carbon efficiency of syngas generation and Fischer–Tropsch synthesis [5], every 1 mol of ethane or ethene in the tail gas, already came at the cost of around 4–5 mol CO2 that were produced. 2) Chemical coproduction may drive the recovery and purification of ethene and propene. This determines whether these compounds have to be refined or not. There are also practical considerations. The separation between propane and propene is such that propene recovery is not 100% and the refinery may still have to process a propene–propane mixture. Petrochemical production may include technologies such as aromatization (Section 22.5). The extent and nature of chemical coproduction therefore markedly affects the composition of the C2 –C4 material that has to be refined to fuels.
515
516
25 Motor-Gasoline Refining
3) Is liquid petroleum gas (LPG) a viable refinery product? The answer to this will have a significant impact on the value and fate of propane and butanes. 4) Light alkenes are enabling compounds for a number of refining technologies. The light alkenes are employed in key technologies for the production of motor-gasoline, such as aliphatic alkylation, oligomerization, and etherification. In fact, producing on-specification motor-gasoline becomes difficult, if not impossible, without access to C3 –C4 alkenes. The butenes may well be the single most important alkene of all the alkenes for motor-gasoline refining. 5) When the C2 –C4 hydrocarbons are available for refining, their degree of prefractionation influences the separation strategy and refining technology selection. Processing pure alkenes may result in heat management issues, whereas alkane–alkene mixtures may affect recycling and the size of the conversion units. 6) The refining of ethene is a special case. Reference was made to ethene refining in the discussions on technologies for hydration (Section 17.4), oligomerization (Chapter 19), and aromatic alkylation (Chapter 20). The selection of a technology for ethene refining will depend on the overall refinery design, but it is also influenced by the composition of the stream containing ethene. One can expect different designs when refining ethene as pure ethene, an ethene–ethane mixture, or C2 and lighter tail gas, which avoids cryogenic separation. Devising a suitable conversion technology for converting a C2 and lighter tail gas that contains unconverted syngas is sorely needed. Such a technology has the opportunity to improve the carbon efficiency of Fischer–Tropsch-based facilities and other facilities that deal with impure ethene-containing streams. 25.3.3 Fate of the Residue and Wax
The amount of residue in HTFT syncrude is very small and it is very aromatic. Irrespective of the refining pathway selected for it, it will not have a significant impact on motor-gasoline production. Hydrocracking may produce low octane number naphtha, but this can be co-refined with the straight-run naphtha if its quality is too low for direct blending. The situation is very different when LTFT syncrude is the refinery feed. About half the syncrude mass is contained in the atmospheric residue as a paraffinic wax. The type of cracking conversion (Chapter 21) that is needed will depend on the following: 1) The extent to which chemical and lubricant base oil coproduction is allowed. Waxes as petrochemicals and lubricant base oils are higher value products than transportation fuels. If the production of these products is more important than motor-gasoline production, the technology decision will be based on the former. In the present context it is not the case. 2) The importance of fuels production, and motor-gasoline specifically. When there is emphasis on motor-gasoline, cracking will be performed in such a way that it will favor naphtha production. This can be achieved by increasing the cracking severity, or by the selection of a cracking technology that predominantly produces naphtha. 3) The amount of light alkenes that is required to produce motor-gasoline may dictate the type of cracking conversion selected. Here the nature of the Fischer–Tropsch syncrude plays a role, because it determines the amount of light alkenes available in the straight-run product.
25.3 Decisions Affecting Motor-Gasoline Refining
It is the need for light alkenes that led to the conclusion that fluid catalytic cracking (FCC) of wax is better than hydrocracking for fuel production [6]. It is also the need for light alkenes and their value in the refinery that led to the conclusion that of the LTFT technologies, Fe-LTFT has a refining advantage, both in terms of its straight-run product and in the way the deactivation of Fe-LTFT affects the syncrude composition [4].
25.3.4 Fate of the Aqueous Product
The largest volume liquid product from Fischer–Tropsch synthesis is water. When the water is condensed during syncrude recovery, the polar oxygenates partition between the aqueous and organic phases (Section 5.3.3). This very important separation, which is always part of Fischer–Tropsch syncrude recovery, determines what compounds will dissolve in the water and what compounds will dissolve in the oil. Since water is the largest volume product, the syncrude in the Fischer–Tropsch aqueous product is always present as a dilute aqueous solution. As in the case of the C2 –C4 hydrocarbons (Section 25.3.2), there may not be an immediately apparent incentive to recover the dissolved oxygenates from the aqueous product. The decision to recover oxygenates or not, affects the carbon efficiency of the facility in much the same way. The level of carbon efficiency that can be attained hinges on the level of oxygenate recovery that is included in the design and how the wastewater is processed (Figure 25.4). When no oxygenates are recovered, the total aqueous product is treated as a wastewater. Aerobic conversion leads to no carbon recovery, but some carbon may be recovered as methane by anaerobic digestion of the dissolved oxygenates by suitable microorganisms. By recovering (a)
(b)
Fischer–Tropsch gas loop
Fischer–Tropsch gas loop
Aqueous product
CH4
Aerobic wastewater treatment
Water
Anaerobic wastewater treatment
Water
Aqueous product
(c)
(d)
Fischer–Tropsch gas loop
Fischer–Tropsch gas loop
Aqueous product
Aqueous product
Oxygenate recovery
Oxygenates
Wastewater treatment
Water
Oxygenate recovery
Oxygenates
Acid recovery Wastewater treatment
Figure 25.4
Carboxylic acids
Water
High-level options for Fischer–Tropsch aqueous product refining.
517
518
25 Motor-Gasoline Refining
some or most oxygenates from the aqueous product, the relative contribution of wastewater treatment can be reduced. The oxygenates that are lower boiling than water can easily be recovered as an overhead product during conventional atmospheric distillation. The subsequent recovery of oxygenates from the concentrated oxygenate stream (overhead product) is more efficient, because further separation and conversion is not affected by the bulk of the water. The bottom product still contains the oxygenates that are higher boiling than water, most of which are carboxylic acids. Recovering carboxylic acids from a dilute aqueous product is not easy, and there is no obvious technology to do so in an efficient manner. The design of the aqueous product recovery and refining is divorced from the gas loop design, although anaerobic methane production during wastewater treatment may yield methane to be processed in the gas loop. The selective condensation of carboxylic acids before condensing the bulk of the water during syncrude cooling has been evaluated, but it is not possible. The carboxylic acid content in the vapor phase is too little and, once the water starts condensing, there is little preferential partitioning between carboxylic acids and water in the vapor and liquid. 25.3.5 Alkane-Based Naphtha Refining
It was pointed out that the alkane quality is the key to motor-gasoline refining (Figure 25.2). The minimum quality of the alkane fraction of motor-gasoline that is necessary to meet fuel specifications can be calculated. Since the nonalkane compound classes on average have better octane numbers than the alkanes, a theoretical blend with the least amount of alkanes will determine the minimum alkane quality allowable. If the alkenes, aromatics, and oxygenates are at their maximum specification limit (Table 25.1), 32 vol% alkanes are required to complete the blend. The alkane quality necessary to meet the octane requirements is an RON around 70 and an MON around 65. This minimum is valid for all refineries producing motor-gasoline. One would of course not design a refinery to be at the blending limit, and the average quality of the alkanes that is required by a realistic refinery design must be better than the lower limit indicated. There are some specific challenges when refining alkanes for inclusion in motor-gasoline, only some of which are specific to Fischer–Tropsch refining: 1) In a refinery, the alkanes are not always segregated from the other compound classes. This becomes a challenge when low octane number alkanes are present in mixtures with other compounds and when the alkanes cannot be easily separated or upgraded. It is typically the case when dealing with a reformate in which there are alkanes present in a mixture with aromatics. The alkanes lower the octane number of the reformate, and the concentration of alkanes depends on the technology selection and operation (Chapter 22). There are ways to separate the alkanes and aromatics [7], but this separation is seldom practised in fuels refineries. The challenge is not unique to Fischer–Tropsch syncrude refining. 2) The refining technologies that are available for producing good quality naphtha range alkanes are limited to butanes, isomerates, and alkylates. This limitation is not unique to Fischer–Tropsch syncrude refining. Some butanes can be directly blended into the motor-gasoline until the vapor pressure limit is reached. The butane content in a typical motor-gasoline is 5–10 vol%. Hydroisomerization of C5 –C6 naphtha is the main technology for light naphtha refining. Aliphatic alkylation is an important crude oil refining technology
25.3 Decisions Affecting Motor-Gasoline Refining
for producing high-octane heavy naphtha, but in a Fischer–Tropsch refinery aliphatic alkylation is not recommended in the absence of a butane/isobutane source [8]. In such cases, indirect alkylation by selective dimerization followed by hydrogenation is preferred. 3) There is no refining technology for the upgrading of C7 –C10 alkanes into high octane number alkanes.
25.3.6 Technology Selection
The main refining technologies that can be used to produce motor-gasoline are shown on the carbon-number-based technology map (Figure 25.5). Not all of the technologies are equally well suited for Fischer–Tropsch syncrude. Technology preselection based on the discussion in Chapter 16 and in the literature [9] can whittle down the list for efficient syncrude refining. Within each conversion type, there are many technology choices, and in some instances the application may be very sensitive to the actual technology choice. The first rule of refinery design is very applicable – the devil is in the details. Heavier than naphtha range carbon numbers have not been shown in the technology map (Figure 25.5), because the products derived from cracking usually benefit from further refining before they are included in motor-gasoline. In this way, the cracking products only contribute to the feed in the carbon number range shown. This does not mean that the cracked products cannot be directly included in motor-gasoline, but the blending decision depends on its properties in the context of overall fuel quality. Alcohols
C2
Alkene hydration
C3
Aromatic alkylationa
Aromatics
C4
Aromatization
Aromatics
C5
C4 hydroisomerization
C6
Oligomerization
Aliphatic alkylation
Alkanes Alkenes
Alkene hydrogenation
Alkanes
C7 Skeletal isomerization
Etherificationb
Ethers
C8 C5 – C6 hydroisomerization
C9 C10 a b
Naphtha hydrotreating
Naphtha reforming
Requires a source of aromatics, preferably benzene. Requires a source of alcohols, preferably methanol or ethanol.
Figure 25.5
Technology map for motor-gasoline refining.
Alkanes Aromatics
519
520
25 Motor-Gasoline Refining
With the possible exception of naphtha from FCC, cracking technology is not selected to produce motor-gasoline per se. In crude oil refineries, the light alkenes produced during cracking conversion is very important for motor-gasoline production, and in this way cracking is very important. In a Fischer–Tropsch refinery, alkenes are primary products and cracking technology is not required as an alkene source. Nevertheless, motor-gasoline production from LTFT syncrude benefits greatly from the FCC of wax [10]. There is insufficient alkenes and naphtha range material in straight-run LTFT syncrude to produce motor-gasoline as the main product. When FCC of wax is employed, the cracking product resembles HTFT syncrude, but with the advantage that the products are deoxygenated and more branched. Some general observations about the technology selection for motor-gasoline refining from Fischer–Tropsch syncrudes are the following: 1) The refining of ethene and propene must be considered together. When chemical extraction is not considered, or possible, these molecules must be refined to fuel. Ethene is more difficult to refine, and the most efficient refining pathway for ethene is aromatic alkylation. However, aromatic alkylation depends on aromatics availability and can be considered only when sufficient aromatics and preferably benzene are available from aromatization or reforming. Alternatively, hydration can be considered to produce ethanol, which is a commodity chemical and a fuel. When ethene is not employed for aromatic alkylation, at least some propene must be reserved for benzene alkylation in order to reduce refinery benzene levels, irrespective of whether benzene is produced on purpose or as a side product. The remainder of the propene can be oligomerized to produce a heavy olefinic motor-gasoline. Since there is a limitation on the alkene content of motor-gasoline, the excess product can be hydrogenated and blended into jet fuel or diesel fuel depending on the oligomerization technology selection. 2) The C4 alkenes are central to motor-gasoline refining. The recommended refining pathway for butene conversion is the production of alkylate-equivalent motor-gasoline [8]. There are many other potential pathways (Figure 25.5), but none as attractive. The future viability of motor-gasoline refining relies heavily on the ability to produce good quality paraffinic motor-gasoline. 3) The fate of butane depends on the vapor pressure of the base motor-gasoline and the possibility of producing LPG. Aromatization can be considered if a large quantity of butane and propane is available. 4) Hydroisomerization is the preferred refining pathway for C5 naphtha and, depending on the refining needs, may also be considered for C6 naphtha. The advantage of C5 -only hydroisomerization is that it makes recycling easier and cheaper. The alkene fraction of C5 material can be etherified, but the application of any technology for the production of fuel oxygenates runs the risk of future politically motivated specification changes. The history of methyl tertiary butyl ether (MTBE) (Section 2.4.4) should not be forgotten. 5) The refining technology choice for the refining of heavy naphtha (C7 –C10 ) to motor-gasoline is very limited. In fact, there is only catalytic naphtha reforming and, in the case of syncrude, preferably monofunctional nonacidic Pt/L-zeolite-based reforming (Section 22.4). 6) When monofunctional nonacidic Pt/L-zeolite-based reforming technology is employed with C6 hydrocarbons, benzene is a major product. Although this provides a platform for the production of chemicals (Chapter 28), the maximum benzene content in motor-gasoline
25.3 Decisions Affecting Motor-Gasoline Refining
7)
8)
9)
10)
is limited. Aromatic alkylation is a preferred refining technology in combination with reforming, since it retains the octane value of the benzene, it can process all refinery sources of benzene, it upgrades aromatics and alkenes, and it may be employed as a technology for chemical coproduction. The use of catalytic cracking to convert low-quality heavy naphtha is not recommended for the technical reasons that were outlined before (Section 21.4). It violates the fourth rule of refinery design – the design must be fundamentally sound. In a syncrude refinery, there is a shortage of material (quality and quantity) in the heavy naphtha range. Although it is possible to convert the heavy naphtha into aromatics to improve quality, the volume of heavy naphtha must also be supplemented. Oligomerization to produce olefinic motor-gasoline, or indirect alkylation by selective butene dimerization and hydrogenation to produce paraffinic motor-gasoline, is important to balance the motor-gasoline. When the quality of the heavy naphtha is constraining motor-gasoline blending to meet specifications, it is possible to blend some of the C9 –C10 fraction into jet fuel. However, Jet A-1 has a maximum T10 distillation specification of 205 ◦ C, which limits the volume of heavy naphtha that can be included. Straight-run Fischer–Tropsch naphtha contains percentage level oxygenates. It is therefore inadvisable to have an atmospheric distillation unit (ADU) as the first unit in the refinery. In order to avoid boiling point broadening, it is worthwhile to consider deoxygenating or hydrotreating the light oil first, before separating it into different boiling fractions for refining. This is viable only when the downstream refining technologies do not derive a benefit from the alkenes or oxygenates in the feed. In such cases where alkenes or oxygenates must be preserved, some prefractionation can be considered.
25.3.7 Co-refining
The technology that is employed for synthesis gas generation (Chapter 3) may lead to the coproduction of products that can be co-refined with the Fischer–Tropsch syncrude. This is the case when low-temperature gasification technology is employed, which leads to the coproduction of pyrolysis products. The nature of the pyrolysis products reflects the composition of the raw material used as feed to the gasifier. The gasifier technology will also affect the composition, since it determines the extent of secondary pyrolysis reactions taking place. Examples of such co-refining in industrial Fischer–Tropsch facilities have been pointed out (Sections 6.4.2, 6.4.3, 8.4.4, and 9.4.4). The co-refining can be limited to blending or separate refining followed by blending, or involve coprocessing with the syncrude. The pyrolysis products from biomass, coal, and waste are different [11–14], and each may offer specific advantages (or disadvantages) during blending, separate refining, or coprocessing. For example, the high content of cyclic material in many pyrolysis products can be used to improve the N + 2A value of syncrude for conventional catalytic naphtha reforming (Section 22.3). It has also been reported that phenolic coal liquids can be refined by aromatic alkylation to produce high octane number motor-gasoline extenders [15]. The raw material itself may be accompanied by products that are more efficiently co-refined than being gasified or reformed to produce synthesis gas. These accompanying products are
521
522
25 Motor-Gasoline Refining
related to the raw material, like the natural gas liquids that are associated with natural gas, which is employed for synthesis gas production. Examples of such co-refining in industrial Fischer–Tropsch facilities have been pointed out (Sections 7.2 and 10.2.1). As in the case of pyrolysis liquids, the co-refining can be limited to blending or separate refining followed by blending, or involve coprocessing with the syncrude. The advantages and disadvantages for co-refining are related to the nature of the associated products. Whatever the coproduced or associated product, a decision must be taken on how this material will be processed and integrated with the syncrude. There is limited literature on the subject [16], which is partly due to its specificity.
25.4 Motor-Gasoline Refining from HTFT Syncrude
The carbon cost associated with not refining the light hydrocarbons and aqueous product from HTFT synthesis is very high. Designs that do not take these feed materials into account are not considered. All HTFT designs are therefore based on a closed gas loop with methane recycle. Cryogenic separation of the C2 hydrocarbons is implicit in such a gas loop design. The HTFT design cases (Sections 25.4.1 and 25.4.2) have been selected to illustrate specific aspects of motor-gasoline refining from HTFT syncrude. No optimization was performed, and many alternative configurations are possible. 25.4.1 HTFT Motor-Gasoline Design Case I
The first HTFT refinery design that will be considered (Figure 25.6) is a simple and straightforward design that was performed on the basis of the recommended technologies for Fischer–Tropsch syncrude (Section 25.3.6). To keep it simple, a standard HTFT gas loop design was employed for the syncrude recovery (Figure 5.6), but with an implicit prefractionation of the C3 –C4 and C5 –C6 HTFT condensate streams. In this design, aromatic alkylation was ruled out as technology for ethene refining, because the molar flow of ethene is three times more than the molar flow of benzene from the reforming unit. Ethene hydration was therefore selected as ethene refining technology. Hydrating the ethene produces ethanol in excess of the blending requirements for motor-gasoline. This is not a problem, since ethanol is a large-volume commodity chemical that is easily transportable and can be sold as ethanol. The aqueous refining strategy matched the oil refining strategy. Ethene hydration technology can be integrated with carbonyl hydrogenation and alcohol purification. By doing so, there is no need to duplicate the units necessary for product workup in the ethene hydration unit. In the design, the mixed C3 and heavier alcohols from carbonyl hydrogenation are dehydrated to alkenes, which can then be refined with the rest of the straight-run syncrude. It is also possible to retain the alcohols as chemicals. Without changing the basic design of Figure 25.6, the cut point and blending of streams were changed and units were turned down to assess the impact on the product distribution and
25.4 Motor-Gasoline Refining from HTFT Syncrude Carbonyl hydrogenation
Aqueous product Ethene hydration
C2
Wastewater C3+ alcohol dehydration
Condensate C5 –C6
Light oil
LPG
C3
C3 SPA oligomerization
C4
C4 SPA oligomerization
Gasoline
Alkene hydrogenation
C5 hydroisomerization
Benzene
Gasoline Jet fuel Butanes Gasoline
Nonacidic Pt/L reforming
H2 LPG Gasoline
Mild hydrocracking
Jet fuel Diesel
Hydrotreater
Decanted oil
Alcohols Gasoline Jet fuel
Aromatic alkylation
HTFT gas C –C 3 4 loop
Ethanol
Figure 25.6 HTFT motor-gasoline refinery design case I. The refinery design is based on recommended technologies and the gas loop includes cryogenic C2 separation with methane recycle.
product quality (Table 25.2). From this analysis, it can be seen that the simple design is quite robust and flexible. The following cases were considered: 1) Base case design. Oxygenates were not included in the motor-gasoline blend. The ethanol that was produced by ethene hydration and aqueous product refining is listed as a petrochemical product. Because of the hydration unit, the refinery is a net consumer of water (disregarding the much larger volume of reaction water produced by HTFT synthesis). All transportation fuels meet the specification. 2) No dehydration. When the dehydration unit is shut down, the volume of propene and butene to the refinery decreases. The C3 and heavier alcohols are now available as liquid petrochemical products. The cut points between motor-gasoline and distillate products were not adjusted. Despite the change, all transportation fuels still meet specifications. Since the motor-gasoline blend is ethanol free, it implies that the ethene hydration, carbonyl hydrogenation, and alcohol dehydration units can all be shut down simultaneously without significantly affecting transportation fuel production. 3) Ethanol addition. The volume and quality of motor-gasoline production can be increased relative to the base case by including 5 vol% ethanol into the fuel. Importantly, the vapor pressure specification was not violated when ethanol was added. 4) Maximum motor-gasoline. In order to produce the maximum amount of motor-gasoline, a number of changes relative to the base case can be implemented. By routing the hydrocracker naphtha directly to the motor-gasoline, the volume loss associated with aromatization in
523
524
25 Motor-Gasoline Refining Table 25.2
Products from HTFT motor-gasoline refinery design case I shown in Figure 25.6.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t−1 ) H2 and fuel gas (kg·t−1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m−3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m−3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m−3 )
Base case
0.5 7.2 4.4 58.6 8.1 6.3 11.6 0.0 4.5 –1.1
No dehydration
0.5 7.2 4.4 56.3 7.1 6.3 14.1 0.0 6.6 –2.4
Ethanol added
0.5 7.2 4.4 61.7 8.1 6.3 8.4 0.0 4.5 –1.1
Max gasoline
0.3 7.1 4.3 62.7 9.0 4.9 8.2 0.0 4.5 –1.1
0.083 0.782 0.104 0.076
0.082 0.751 0.091 0.076
0.083 0.822 0.104 0.076
0.081 0.840 0.116 0.059
1.045 116 77
1.000 141 77
1.085 84 77
1.097 82 74
97.2 89.8 749 53 17.6 33.1 0.3 0.0
97.3 90.2 750 55 14.3 34.5 0.3 0.0
98.4 90.3 751 58 16.7 31.5 0.3 4.9
97.1 89.1 747 57 16.5 28.3 0.2 5.0
776 21.9
780 25.0
776 21.9
777 19.8
55 822
55 822
55 822
51 825
a On a syncrude basis, excluding H , CO, CO , H O, and CH from HTFT synthesis (closed gas loop with cryogenic 2 2 2 4 separation). b Syncrude waste products (e.g., metal-containing fuel oil) and unrecoverable material (e.g., carboxylic acids in wastewater).
the reformer is avoided. A small additional gain can be made by blending more distillate into jet fuel to free up alkyl aromatics for motor-gasoline blending. Since this reduces the vapor pressure and increases the base blending volume, 5 vol% ethanol can be added to increase the volume further. Further gains can, in principle, be made by blending some
25.4 Motor-Gasoline Refining from HTFT Syncrude
light kerosene into the motor-gasoline, but this was not done. Extreme interventions, such as increasing the operating severity of the hydrocracker to produce more naphtha, were likewise not considered. The designs had some common features that are worthwhile pointing out. Firstly, it is a net producer of H2 . The H2 can be used to improve the carbon efficiency of the Fischer–Tropsch gas loop, or it can be exported as a product. Secondly, although the emphasis was on motor-gasoline production, all the transportation fuels met specifications. Doing so is not especially difficult for jet fuel, but meeting the minimum density specification for diesel fuel can be a challenge (Chapter 27). The motor-gasoline was well above the minimum octane number specifications. Thirdly, the natural ratio of motor-gasoline to distillate products from HTFT syncrude refining is slightly over 80 : 20. This matches the design objective of the first commercial HTFT facility, the Hydrocol plant in the United States (Chapter 7). Fourthly, even though it was not the design intent to produce chemicals, the design naturally yielded around 10 mass% chemicals. The chemical portfolio was restricted to ethanol. The refinery produces a large amount of fuel gas (7%) and much organic material is labeled as unrecovered or waste products (5–7%). When we analyze the refinery design for shortcomings and potential areas of improvement, the following emerge: 1)
2)
3)
4)
5)
Ethane-rich product is not refined, although it is available as a very pure C2 hydrocarbon stream. The ethane-rich stream represents 70% of the refinery product that ends up as fuel gas. It is a good feed material for thermal cracking to produce ethene, and refining of this product should be considered. Using it as a fuel gas is wasteful. This is a simple addition to the design, and in the context of Figure 25.6 the size of the feed streams to the ethene hydration unit and C3 –C4 separation will increase. The H2 recovery from the reformer product gas by pressure-swing absorption was assumed to be 85%. The fate of the excess H2 production and the need to perform separation on the whole stream depend on the use of the excess H2 . If it will be reintegrated in the Fischer–Tropsch gas loop, there may be some gain in efficiency. This can be easily accomplished within the design and it does not affect the refinery configuration. Fuel oil is produced during the mild hydrocracking of the distillate and residue fractions. In the product listing (Table 25.2), the fuel oil is grouped with the waste products, but it can be used as lubricant base oil [17]. It is a small stream, but nevertheless contributes 40% of the waste in the base case design, or about 2% of the syncrude to the refinery. This does not require a change in refinery configuration, but it requires an additional hydrotreating step to reduce the aromatics content in order to produce American Petroleum Institute (API) Group III lubricant base oil. Lubricant refining is discussed in Chapter 28. No carboxylic acid recovery takes place in the aqueous product refinery. This is a technology deficiency rather than a design deficiency (Section 25.3.4). The inability to recover oxygenates of higher boiling point than that of water from the aqueous product contributes 60% to the waste, or about 3% of the syncrude to the refinery. The separation strategy is potentially suboptimal. Any distillation of material that will be deoxygenated in the refinery before deoxygenation will result in boiling point broadening. The design intent of placing the C5 separation before the hydrotreater was to beneficially use the heat of alkene hydrogenation in the C5 hydroisomerization unit [18]. By implication, a
525
526
25 Motor-Gasoline Refining
Pt/MOR catalyst must be selected (Section 18.3.2). The best placement of C5 separation is an issue to be considered at the next level of detail. 6) The design of the oil hydrotreater and subsequent separation is also potentially suboptimal. Hydrotreating such a broad range of products in a single unit may not be the best way to proceed. Heavier products are hydrogenated more deeply than light products when hydrotreated as a mixture [19]. This needs to be addressed at the next level of detail. The routing of distillate to the mild hydrocracker increases the naphtha yield. In a motor-gasoline refinery this is advantageous, but it may not be the best overall design decision. The hydrogenated straight-run distillate may be used directly as blending components in the fuel. This topic is considered in subsequent chapters. 7) A conscious design decision was made to include the C6 naphtha in the feed to the nonacidic Pt/L-zeolite-based reformer. Hexane is the most efficient feed for such a unit and the decision is technically sound. It has some repercussions though. Excess H2 is produced that is, strictly speaking, not needed by the refinery. The refinery produces benzene, which has an advantage as a chemical platform, but requires alkylation to be included into motor-gasoline. The overall refinery product becomes more aromatic and has a higher density, thereby reducing the volume of fuel being produced. The octane number of the motor-gasoline is high, and the additional octane value of the C6 -derived aromatics is not necessary to meet fuel specifications, although it allows more volatile and lower octane number compounds to be blended into the motor-gasoline.
25.4.2 HTFT Motor-Gasoline Design Case II
In the second HTFT design case, we consider the impact of the C6 refining strategy. Instead of sending the C6 hydrocarbons to the nonacidic Pt/L-zeolite-based reformer to produce benzene, it is hydroisomerized in combination with the C5 hydrocarbons (Figure 25.7). In addition to this, aspects that are related to the refining of the C5 and heavier material and that were highlighted during the discussion in Section 25.4.1 are incorporated in the design. Hydroisomerization of the combined C5 –C6 naphtha necessitates a more complicated recycle design within the hydroisomerization unit. Recycling with a single-carbon-number feed can be performed by distillation, but molecular sieve separation or separation by multiple distillation columns is required for C5 –C6 mixed material. The design employs a single unit, but there may be benefit in two hydroisomerization units, one for C5 and one for C6 (analogous to SPA oligomerization), because recycling is easier in a single-carbon-number hydroisomerization unit. It also allows the decoupling of equilibrium-limited C5 conversion from kinetic-limited C6 conversion. This is at the next level of detail and does not change the basic refinery configuration. Dehydrating the C3 and heavier alcohols has not been included in the base case design, thereby simplifying the aqueous product refinery. The aromatic alkylation unit has also been omitted from the design, because refinery benzene levels are much lower. An oil hydrotreater converts all the C5 and heavier material into hydrocarbons. This avoids boiling point broadening, but recognizing that in practice, it may be better to employ two hydrotreaters and that there may still be benefit in routing the unhydrogenated C5 –C6 naphtha to the hydroisomerization unit (see comments in Section 25.4.1).
25.4 Motor-Gasoline Refining from HTFT Syncrude Carbonyl hydrogenation
Aqueous product
C3 HTFT C3 –C4 gas loop
Wastewater
Ethene hydration
C2
C4
LPG Gasoline
C3 SPA oligomerization C4 SPA oligomerization
Alkene hydrogenation
Light oil Decanted oil
Gasoline Jet fuel Butanes
C5 – C6 hydroisomerization
Gasoline
Nonacidic Pt/L reforming
H2 LPG Gasoline
Mild hydrocracking
Jet fuel Diesel
Condensate, C5 –C6 Hydrotreater
Ethanol Alcohols
Figure 25.7 HTFT motor-gasoline refinery design case II. The refinery design employs C6 hydroisomerization instead of C6 reforming and benzene alkylation, leading to a more paraffinic product slate. The gas loop includes cryogenic C2 separation and methane recycling.
An analysis of the design in Figure 25.7 shows that it is less robust than that in Figure 25.6 on account of the lower aromatic content of the products. This can be seen from the variations that were investigated to evaluate robustness (Table 25.3): 1) Base case design. All transportation fuels meet specification. However, in order to meet the motor-gasoline specifications, the base case design had to include ethanol in the fuel blend. In the base case, the excess ethanol and the C3 and heavier alcohols are petrochemical products. Although no C6 material is sent to the reformer, some benzene is produced by hydrogenolysis. There is consequently a risk involved, should the benzene selectivity in the reformer increase, because the benzene content is close to the specification limit. In the event that this should happen, two mitigating strategies can be implemented. The excess benzene can be co-fed to the hydroisomerization unit or, alternatively, the excess benzene can be co-fed to the C3 SPA oligomerization unit [20]. 2) C3 and heavier alcohol dehydration added. In North American refineries, where it is not necessary to meet the RON and MON specifications separately, the C3 and heavier alcohols can be dehydrated to alkenes as shown in Figure 25.6. The refinery design becomes RON-constrained with the additional propene from dehydration, because the volume of C3 -derived SPA oligomers becomes too much to be easily blended. The maximum volume of olefinic motor-gasoline is limited by the alkene specification. The maximum volume of isoparaffinic kerosene is limited by the minimum density requirement of the jet fuel. In order to strike a balance, some of the oligomers can be included into the motor-gasoline as a hydrogenated product, but hydrogenated propene oligomers have poor octane numbers
527
528
25 Motor-Gasoline Refining Table 25.3
Products from HTFT motor-gasoline refinery design case II shown in Figure 25.7.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t−1 ) H2 and fuel gas (kg·t−1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m−3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m−3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m−3 )
Base case
0.1 6.9 4.9 59.6 8.5 5.0 10.9 0.0 6.6 −2.4
Dehydration added
0.0 6.9 4.7 62.4 9.5 5.0 8.2 0.0 4.5 −1.1
No ethanol
0.1 6.9 4.2 57.2 8.5 5.0 14.1 0.0 6.6 −2.4
0.091 0.818 0.110 0.061
0.087 0.857 0.123 0.060
0.078 0.791 0.110 0.061
1.079 109 69
1.127 82 69
1.039 141 69
95.0 88.4 729 60 16.2 21.4 0.5 4.9
94.5 87.9 728 60 17.9 20.4 0.5 5.0
93.7 87.9 723 60 16.7 22.2 0.5 0.0
778 18.6
775 16.6
778 18.6
51 824
51 824
51 824
a On a syncrude basis, excluding H , CO, CO , H O, and CH from HTFT synthesis (closed gas loop with cryogenic 2 2 2 4 separation). b Syncrude waste products (e.g., metal-containing fuel oil) and unrecoverable material (e.g., carboxylic acids in wastewater).
(Table 19.3). Despite the seemingly small change, there are just too little aromatics to make the blends. This leads us to a valuable general observation about syncrude refining: it becomes difficult (if not impossible) to produce on-specification transportation fuels with less than 20 vol% aromatics in the transportation fuel boiling range.
25.5 Motor-Gasoline Refining from LTFT Syncrude
3) No ethanol addition. Reducing or excluding ethanol from the motor-gasoline is no longer an option, because the ethanol is needed to meet octane number requirements. The production of H2 in the refinery has diminished as a result of the decrease in reformer feed. In general, blending is more constrained, with many fuel properties at or near the specification limit. Although the present design (Figure 25.7) contains fewer units than the case I design (Figure 25.6), the decrease in robustness is telling. Etherification was not considered because the amount of ethanol that could be included in the motor-gasoline was not vapor pressure constrained.
25.5 Motor-Gasoline Refining from LTFT Syncrude
The concentration of the C2 –C4 hydrocarbons in LTFT syncrude is less than that in HTFT syncrude. The concentration of these compounds increases with deactivation of the Fischer–Tropsch catalyst. Recovery of at least the C3 –C4 material is necessitated by the value of this fraction for the production of motor-gasoline. The decision to separate and recover the C2 hydrocarbons is less obvious. The carbon efficiency of the facility is decreased when the C2 hydrocarbons are not recovered, but overall it may be more practical to recycle most of the C2 and lighter tail gas to a reformer. All the LTFT designs will be based on a closed gas loop with C2 and lighter recycle. The design basis for LTFT refining is therefore different than for HTFT refining. The lower C1 –C2 hydrocarbon and carboxylic acid selectivity of LTFT compared to HTFT synthesis translates into a higher volumetric product yield, even though these compounds are not recovered. The LTFT design cases (Sections 25.5.1–25.5.3) have been selected to illustrate specific aspects of motor-gasoline refining from LTFT syncrude. 25.5.1 LTFT Motor-Gasoline Design Case I
The selection of the wax conversion technology is central to LTFT refinery design. Half of the syncrude is in the C22 and heavier fraction. In fact, the products obtained by wax conversion, rather than the straight-run syncrude, characterize the properties of the syncrude that has to be refined. In the first LTFT design case, a fluid catalytic cracker is employed to convert the heavier alkanes into lighter material (Figure 25.8). It is a simple and straightforward design based on the recommended technologies for Fischer–Tropsch syncrude (Section 25.3.6). To keep it simple, the configuration used a standard LTFT gas loop design (Figure 5.7), which additionally included recovery of the C3 –C4 hydrocarbons from the tail gas. Apart from the wax FCC, the LTFT motor-gasoline refinery design (Figure 25.8) has the same units as in the HTFT motor-gasoline cases (Figures 25.6 and 25.7). The wax FCC in the design operated with a recycle to limit the products to C14 and lighter. The aqueous product refinery is smaller, and its significance in the overall refinery depends on the LTFT technology. The amount of oxygenates in the Fe-LTFT aqueous product is less than that in the Fe-HTFT case (Table 1.2). In the case of Co-LTFT, the decision to recover oxygenates from
529
530
25 Motor-Gasoline Refining Carbonyl hydrogenation
Aqueous product
Wastewater C3+ alcohol dehydration
C3 LTFT C –C 3 4 gas loop
C4
Methanol Ethanol
LPG Gasoline
C3 SPA oligomerization C4 SPA oligomerization
Alcohols
Alkene hydrogenation
Gasoline Jet fuel Butanes Fuel gas
Wax FCC
Wax
Hydrotreater
Hot and cold condensate
C5–C6 hydroisomerization
Gasoline
Nonacidic Pt/L reforming
H2 LPG Gasoline Jet fuel Diesel
Figure 25.8 LTFT motor-gasoline refinery design case I. The refinery design is based on recommended technologies and wax fluid catalytic cracking. The gas loop includes C3 –C4 recovery with recycling of the C1 –C2 hydrocarbons in the tail gas.
the aqueous product becomes less obvious, since the alcohols contained in the aqueous product is <2% of the syncrude. As a matter of principle (maximum carbon efficiency), the oxygenates in the aqueous product are recovered. In LTFT syncrude, the alcohol content is much higher than the carbonyl content. There is also substantially more methanol in the aqueous product from LTFT synthesis than in the HTFT aqueous product. A carbonyl hydrotreater is nevertheless included in the design to ensure that even low levels of carbonyls are converted into alcohols. The carbonyls can produce carboxylic acids and aldol condensation products in downstream processes if not dealt with appropriately. The volume of C3 and heavier alcohols is small, and a small dehydration unit simplifies the product slate by making the heavier alcohols available as alkenes. These alcohols can alternatively be recovered as final products. An important design decision was to opt for C6 hydroisomerization, as in the second HTFT design case (Section 25.4.2). This is very much a Fischer–Tropsch syncrude refining approach and does not give due recognition to the significant contribution from the fluid catalytic cracker. Yet, it avoids the on-purpose production and further processing of refinery benzene. Only a small amount of benzene is produced in the FCC and by hydrogenolysis in the reformer. The refining decision also recognizes that the carbon number distribution is such that the FCC product and LTFT syncrude contain more material for reforming in the C7 and heavier naphtha than was the case for HTFT syncrude.
25.5 Motor-Gasoline Refining from LTFT Syncrude
Without changing the basic configuration of the design in Figure 25.8, the cut point and blending of streams were changed to assess the impact of changes on the product distribution and product quality (Table 25.4). The following cases were considered: 1) Base case design. A syncrude from slurry bed Fe-LTFT technology was employed for the base case design. The ethanol that was recovered from the aqueous product was blended into the motor-gasoline. The methanol was considered a final product, and the C3 and heavier alcohols were dehydrated to alkenes. The straight-run distillate range products were hydrotreated and blended into jet fuel and diesel fuel. Although the motor-gasoline met specifications, the branching of the kerosene range product might be insufficient to meet the freezing point specification of jet fuel. The gas oil range product had a high cetane number and low density, which is a refining issue that is discussed in Chapter 27. 2) Co-LTFT design. The base case design was retained, but the Fischer–Tropsch technology was changed to fixed bed Co-LTFT with the same α-value. The main impact of this on the syncrude is that the straight-run material became more hydrogenated and that the oxygenate content in the aqueous product was less. The fuel qualities were comparable with those of Fe-LTFT, the motor-gasoline from Co-LTFT refining having a slightly lower octane number. 3) Maximum motor-gasoline. The motor-gasoline yield can be increased by co-feeding the distillate in the LTFT hot condensate with the wax to the FCC. This has the added advantage of overcoming the quality issues related to the distillate. Some jet fuel is still produced from the FCC, but now the kerosene is isomerized and meeting the freezing point specification of jet fuel is no longer a problem. 4) No aqueous refinery. The complexity of the refinery can be reduced by omitting the aqueous product refinery and treating the total product as wastewater. Losing the ethanol as a blending component reduces the octane number of the motor-gasoline and results in a yield loss from the refinery. The lack of aqueous product recovery is clearly detrimental. The wax FCC unit acted as a buffer to many of the changes that were evaluated. At this point, a word of caution. The application of FCC technology to Fischer–Tropsch wax is not without its challenges (Section 21.4.2). The robustness of the design is potentially a reflection of the simplifying assumptions that were made in the description of wax FCC over an equilibrium USY catalyst. It may well be true that the product distribution is insensitive to the carbon number distribution of Fischer–Tropsch wax feed, but as cautioned before, ‘‘The devil is in the details.’’ There are other common features that can be pointed out. Firstly, all variations on the design were net producers of H2 . Secondly, the distillate range products were no longer as easily refined to on-specification products as with HTFT syncrude. Thirdly, by converting the wax into lighter products, much of the natural chemical potential of LTFT syncrude was destroyed. Fourthly, it was possible to produce on-specification motor-gasoline in high yield from LTFT syncrude. This dispels the myth that LTFT is not good for motor-gasoline production. In fact, the best case achieved a motor-gasoline to distillate ratio of 90 : 10 (Table 25.4). When we analyze the refinery design for shortcomings and potential areas of improvement, the following emerge: 1) The refinery effectiveness is based on only the aqueous product and C3 and heavier hydrocarbons from LTFT synthesis. The assumption that the C1 –C2 hydrocarbons can be recycled to the Fischer–Tropsch gas loop also implies a purge to avoid the buildup of inert
531
532
25 Motor-Gasoline Refining Table 25.4
Products from LTFT motor-gasoline refinery design case I shown in Figure 25.8.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t−1 ) H2 and fuel gas (kg·t−1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m−3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m−3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m−3 )
Base case (Fe-LTFT)
Co-LTFT
Max gasoline
No aqueous refining
0.6 1.9 6.2 61.5 6.9 20.3 1.0 0.0 1.1 0.5
0.8 2.0 8.5 60.1 6.5 20.9 0.4 0.0 0.6 0.4
0.9 2.4 6.6 78.6 8.9 0.0 1.0 0.0 1.2 0.5
0.6 1.9 5.8 59.5 6.7 20.3 0.0 0.0 4.8 0.3
0.116 0.836 0.088 0.263
0.158 0.816 0.084 0.270
0.124 1.068 0.115 –
0.109 0.812 0.087 0.263
1.303 10 25
1.329 4 27
1.307 10 33
1.270 0 26
95.2 89.0 736 60 6.6 28.8 0.1 3.2
94.9 88.9 736 60 5.6 30.3 0.1 1.3
95.0 88.9 736 59 6.3 29.0 0.1 2.5
94.3 88.7 733 60 6.3 29.6 0.1 0.0
775 16.2
777 17.6
775 16.0
775 16.5
76 773
76 773
– –
76 773
a On a syncrude basis, excluding H , CO, CO , H O, and C –C hydrocarbons from LTFT synthesis (closed gas loop 2 2 2 1 2 with C3 –C4 recovery from the tail gas). b Syncrude waste products (e.g., coke on FCC catalyst) and unrecoverable material (e.g., carboxylic acids in wastewater).
material. This potential loss of material as a fuel gas is not reflected in the analysis. The exclusion of cryogenic separation and C2 hydrocarbon refining therefore creates a positively biased picture of LTFT syncrude as refinery feed. It does not detract from the refinery design, but with a different set of gas loop assumptions the design may have been more complicated without reflecting much gain.
25.5 Motor-Gasoline Refining from LTFT Syncrude
2) The production of distillates that do not meet transportation fuel specifications, and how to rectify the situation is addressed in subsequent chapters (Chapters 26 and 27). 3) The motor-gasoline is very close to specification. This is the consequence of some design decisions, foremost being the decision to route the C6 naphtha to a hydroisomerization unit instead of to the nonacidic Pt/L-zeolite-based reformer. The improved quality from the C4 SPA oligomerization unit due to the higher isobutene content in the feed is beneficial for motor-gasoline quality [21], but its volume is rather limited. The quality of the alkanes in the motor-gasoline is mainly determined by C5 –C6 hydroisomerization product. 4) The oxygenate content of the motor-gasoline was not close to specification and, in addition to ethanol, LTFT syncrude contains methanol. The methanol was not utilized in the refinery design. 5) It is important to recover oxygenates from the aqueous product, but the aqueous product refinery was disproportionately elaborate in relation to its size. The alcohols that are used in motor-gasoline refining can be recovered by separation. This separation is more complex when the alcohols are present in a mixture with a carbonyl compounds. It is consequently difficult to dispense with the carbonyl hydrogenation unit when the alcohols are recovered for refinery use. If the alcohols are not required as alcohols, it may be worthwhile to investigate the impact of co-feeding the mixture of alcohols and carbonyl compounds to one of the acid-catalyzed processes in the refinery, such as the FCC. 6) Two of the three largest units in the refinery are high-temperature endothermic conversion processes, namely, FCC and naphtha reforming. The refinery yield profile does not reflect the associated utility cost of these units. In the case of FCC, the additional amount of fuel required will be substantial, because wax cracking does not lay down enough coke on the catalyst to close the heat balance (Section 21.4). 7) A substantial fraction of the syncrude requires a sequence of three conversion units in series to be refined: FCC, hydrotreating, and hydroisomerization or naphtha reforming. 8) The skeletal isomerization of the C5 –C6 alkenes that takes place in the FCC is not exploited by the downstream refining processes. There are alcohols available from the aqueous product recovery, which can be employed with the isomerized C5 -alkenes to produce fuel ethers, for example, tertiary amyl methyl ether (TAME) or tertiary amyl ethyl ether (TAEE). The motor-gasoline is not alkene-constrained and the C6 material from the FCC can be directly included in the motor-gasoline. 9) The configuration of C5 –C6 hydroisomerization is not necessarily optimal, although it may arguably be at the next level of detail. Some benefit can be implicitly derived by introducing the hydrotreated C5 –C6 naphtha feed at the product separation step, rather than at the feed to the hydroisomerization reactor. However, by hydrotreating the FCC product before hydroisomerization, the potential benefit of olefinicity is lost [18]. One can also consider conducting the hydroisomerization of C5 and C6 material in separate units to make recycle operation easier. 10) A large volume of LPG is produced. With an appropriate market, it is not a problem, but one may have to consider alternative refining pathways if LPG is not a desirable product.
533
534
25 Motor-Gasoline Refining
25.5.2 LTFT Motor-Gasoline Design Case II
The motor-gasoline quality is strongly influenced by the refining strategy for the C5 –C6 naphtha. In the second LTFT motor-gasoline design case (Figure 25.9), due attention is paid to the composition of the C5 –C6 fraction, which contains around 80% FCC naphtha and 20% straight-run LTFT naphtha. The strategy can be summarized as follows: The C5 FCC naphtha is rich in methylbutenes, and methanol is available from the aqueous product. Etherification is not a complicated process and the methanol can be completely consumed by TAME production. No skeletal isomerization unit is required and it is only necessary to recycle methanol in the unit. Ether production can be once-through, because the methylbutenes are available in excess of the methanol availability. It is of course possible to import additional methanol to fully exploit the methylbutenes, but this has not been done. 2) The C5 -raffinate from etherification, which contains unconverted C5 alkenes and alkanes, can be combined with the hydrotreated straight-run C5 LTFT naphtha. This combined feed
1)
Carbonyl hydrogenation
Ethanol Alcohols Wastewater
Aqueous product
C3 C3 –C4
C4
LPG Gasoline
C3 SPA oligomerization C4 SPA oligomerization
Methanol
LTFT gas loop
Alkene hydrogenation
Gasoline Jet fuel Butanes
Etherification
Fuel ether
C5 -raffinate
C5
Wax FCC
Wax
C6
C5
Gasoline Fuel gas C5 hydroisomerization
Gasoline
Benzene Hydrotreater
Hot and cold condensate
Nonacidic Pt / L reforming
H2 LPG Gasoline Jet fuel Diesel
Figure 25.9 LTFT motor-gasoline refinery design case II. The refinery design exploits the properties of the fluid catalytic cracker C5 –C6 naphtha rather than imposing a syncrude design strategy on this fraction, thereby improving the refining efficiency. The gas loop includes C3 –C4 recovery with recycling of the C1 –C2 hydrocarbons in the tail gas.
25.5 Motor-Gasoline Refining from LTFT Syncrude
can then be hydroisomerized, beneficially exploiting the heat of hydrogenation of the alkenes in the mixture [18]. The only reason why the LTFT is hydrotreated beforehand is to simplify the processing of the LTFT cold condensate. On the next level of detail, it is worthwhile to investigate whether the C5 -stripper column can be justified before the hydrotreater to recover the olefinic C5 LTFT naphtha. 3) The C6 naphtha is not combined with the C5 naphtha. This makes recycling of n-pentanes in the C5 hydroisomerization unit easy. It allows the efficient and near-complete conversion of C5 naphtha to methylbutane. The reactor can be optimized for pentanes without being burdened by kinetic constraints imposed by hexane hydroisomerization. 4) The hydrotreated straight-run C6 LTFT naphtha is routed to the nonacidic Pt/L-zeolite-based naphtha reformer, with the heavier naphtha obtained during hydrotreating. The hydrotreated C6 LTFT naphtha is rich in n-hexane and will be converted with high efficiency and selectivity to benzene. Yet, the volume of benzene that is produced in this way is only around 1% of the motor-gasoline. It can easily be co-fed with propene in an SPA-catalyzed oligomerization unit to produce propene oligomers and cumene [20]. By doing so, it reduces the amount of C9 naphtha, which would otherwise have to be included in the motor-gasoline as an olefinic blending component. The conversion of straight-run C6 naphtha into benzene is doubly useful, because it increases the aromatics content slightly and reduces the alkene content of the motor-gasoline slightly. 5) The C6 naphtha from the wax FCC is very olefinic and branched. The C6 FCC naphtha has good motor-gasoline properties, but its direct inclusion in the motor-gasoline is limited by the alkene specification. Yet, with the reduced amount of C9 naphtha from propene oligomerization there is sufficient blending space to directly include the C6 FCC naphtha into the motor-gasoline. Overall, the design in Figure 25.9 is more efficient than that in Figure 25.8, with less material passing through the three conversion units. It also demonstrates that motor-gasoline quality does not have to be improved by more refining (Table 25.5), but can be improved more efficiently by looking at the molecules and refining accordingly. Similar variations within the design were evaluated to assess the impact of changes on the product distribution and product quality: 1) Base case design. The design is based on slurry bed Fe-LTFT syncrude. The aqueous product refinery includes carbonyl hydrogenation followed by alcohol separation. The methanol and ethanol are both employed in motor-gasoline blending, the methanol for TAME and the ethanol for direct blending. The heavier alcohol stream is so small that it can be blended with the diesel fuel [22], although it is shown as a separate chemical product. The motor-gasoline quality in this design (Figure 25.9) is improved compared to the equivalent case I design shown in Figure 25.8. However, the straight-run distillate range products are unlikely to meet fuel specifications, and the same comments apply (Section 25.5.1). 2) Co-LTFT design. There is little impact when the syncrude is changed from Fe-LTFT to Co-LTFT, because the product from wax FCC dominates the motor-gasoline refinery design. 3) Maximum motor-gasoline. The distillate quality issues can be addressed by increasing the motor-gasoline production, which is achieved by co-feeding the straight-run LTFT distillate in the LTFT hot condensate with the wax to the FCC. This avoids diesel fuel production, at the same time producing on-specification motor-gasoline and jet fuel.
535
536
25 Motor-Gasoline Refining Table 25.5
Products from LTFT motor-gasoline refinery design case II shown in Figure 25.9.
Description
Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t−1 ) H2 and fuel gas (kg·t−1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m−3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m−3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m−3 )
Base case (Fe-LTFT)
1.0 2.0 5.7 58.5 10.4 20.3 0.1 0.0 1.6 0.3
CoLTFT
1.1 2.1 8.3 56.8 9.7 20.9 0.0 0.0 0.8 0.3
Max gasoline
1.3 2.5 5.9 75.1 13.1 0.0 0.1 0.0 1.7 0.3
No aqueous refining
0.9 2.0 5.9 55.4 10.4 20.3 0.0 0.0 4.8 0.3
0.103 0.781 0.133 0.263
0.149 0.761 0.124 0.270
0.107 1.006 0.168 –
0.106 0.746 0.133 0.263
1.281 1 30
1.304 0 32
1.282 1 38
1.247 0 29
98.7 90.3 749 60 14.9 31.6 0.0 8.5
97.9 89.9 747 60 15.4 33.3 0.0 3.6
98.3 90.0 746 60 15.8 31.5 0.0 6.6
96.8 89.4 743 60 15.6 33.1 0.0 0.0
779 22.5
782 24.7
775 16.0
779 22.5
76 773
76 773
– –
76 773
a On a syncrude basis, excluding H , CO, CO , H O, and C –C hydrocarbons from LTFT synthesis (closed gas loop 2 2 2 1 2 with C3 –C4 recovery from the tail gas). b Syncrude waste products (e.g., coke on FCC catalyst) and unrecoverable material (e.g., carboxylic acids in wastewater).
25.5 Motor-Gasoline Refining from LTFT Syncrude
4) No aqueous refinery. The motor-gasoline specifications were still met when the oxygenates were not recovered from the LTFT aqueous product, and etherification was omitted from the design. This reflects the robustness of the design.
25.5.3 LTFT Motor-Gasoline Design Case III
The extensive use of alkane-based conversion technologies for motor-gasoline refining from LTFT syncrude suggests that the wax can be hydrocracked, instead of being cracked in a fluid catalytic cracker. Selecting a hydrocracker has less risk associated with the conversion, because a number of industrial Fischer–Tropsch facilities employ wax hydrocracking. Since the objective is motor-gasoline production, the hydrocracker would have to be operated at higher severity, with recycling of the gas oil and wax in order to increase the naphtha production. This configuration forms the basis for the third LTFT design case (Figure 25.10), which explores the usefulness of hydrocracking as opposed to FCC for wax conversion to produce motor-gasoline. The significant amount of C5 –C6 material and lack of aliphatic heavy naphtha to balance the motor-gasoline cause the motor-gasoline to be vapor pressure constrained. The vapor pressure constraint is such that no butanes can be blended into the motor-gasoline and no alcohols can be added. The inclusion or exclusion of aqueous product refining has no impact on motor-gasoline production. The lack of aliphatic heavy naphtha is a direct consequence of the limited alkene availability. Most of the light alkenes have to be employed as alkylating alkenes to reduce refinery benzene. Apart from the straight-run syncrude, there is no additional source of alkenes. The amount of light alkenes in LTFT syncrude is limited, and the combined oligomerization
C3 –C4
C5 –C6 LTFT gas loop
SPA alkylation oligomerization
LPG Gasoline Jet fuel
C5 – C6 hydroisomerization
Gasoline
Benzene Hydrotreater
Hot and cold condensate
C6 –C10
Nonacidic Pt/L reforming
H2 LPG Gasoline Jet fuel Diesel
Hydrocracker
Wax C15 and heavier
Figure 25.10 LTFT motor-gasoline refinery design case III. The refinery design is based on wax hydrocracking (instead of FCC). The gas loop includes C3 –C4 recovery with recycling of the C1 –C2 hydrocarbons in the tail gas.
Jet fuel
537
538
25 Motor-Gasoline Refining
and aromatic alkylation unit must be operated like an aromatic alkylation unit with some oligomerization. A hydrocracker-based refinery design for motor-gasoline is on the borderline of viability (Table 25.6) and the yield of motor-gasoline is much less than for an FCC-based design. The following cases were considered to determine the robustness of the design: Table 25.6
Products from LTFT motor-gasoline refinery design case III shown in Figure 25.10.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t−1 ) H2 and fuel gas (kg·t−1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m−3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m−3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m−3 )
Base case (Fe-LTFT)
0.0 1.2 13.4 23.0 37.1 20.3 0.0 0.0 4.6 0.3
Co-LTFT
0.1 1.2 15.9 23.5 36.2 20.9 0.0 0.0 1.9 0.3
Max gasoline
0.2 1.3 14.9 29.4 49.3 0.0 0.0 0.0 4.6 0.3
0.242 0.311 0.479 0.263
0.286 0.318 0.468 0.270
0.268 0.397 0.636 –
1.295 0 12
1.342 0 13
1.301 0 15
95.9 89.8 741 60 3.6 34.7 0.1 0.0
95.5 89.8 739 61 1.5 34.5 0.1 0.0
96.7 89.6 741 60 12.6 32.9 0.2 0.0
775 15.2
775 14.8
775 13.9
76 773
76 773
– –
a On a syncrude basis, excluding H2 , CO, CO2 , H2 O, and C1 –C2 hydrocarbons from LTFT synthesis (closed gas loop with C3 –C4 recovery from the tail gas). b Syncrude waste products and unrecoverable material (e.g., carboxylic acids in wastewater).
References
1) Base case design. The design employed slurry bed Fe-LTFT syncrude as feed and did not include any aqueous product recovery. This can be added, but due to the vapor pressure constraints on the motor-gasoline, the oxygenates recovered from the aqueous product cannot be integrated into the fuel pool. There are too little alkenes for etherification to be a viable low-vapor-pressure option. The refinery design itself is straightforward, with only five conversion units. The motor-gasoline just meets fuel specifications, but the distillate range products have the same shortcomings noted in the other LTFT design cases (as discussed in Section 25.5.1). 2) Co-LTFT design. The light alkene content in the straight-run syncrude is even lower than in the base case design. This affects the amount of aliphatic heavy naphtha that is produced, thereby increasing the vapor pressure of the motor-gasoline slightly. Since most of the material is processed through the hydrotreater and hydrocracker, the impact of the alkene content on the straight-run syncrude is limited. The inability to include alcohols from the aqueous product, likewise renders the design insensitive to the lower oxygenate content in the Co-LTFT aqueous product. 3) Maximum motor-gasoline. All the distillate range material can be fed to the hydrocracker to convert the gas oil and ensure that the kerosene range material is adequately hydroisomerized. This ensures that the freezing point specification is met. It also makes more material available to increase the aromatics content and density by feeding additional material, including C6 naphtha, to the nonacidic Pt/L-based reformer. This allows the production of on-specification motor-gasoline and jet fuel. The additional aromatics relieve the vapor pressure constraint slightly, and the inclusion of a small amount of ethanol or butanes is possible. The yield of motor-gasoline is still low compared to the yield of jet fuel. Hydrocracking is consequently not the preferred wax conversion technology when motor-gasoline production is important. Although on-specification motor-gasoline can be produced in a hydrocracker-based LTFT refinery, hydrocracking is better suited for a refinery design with jet fuel as main product. Another consequence of hydrocracking for wax upgrading is the inevitable increase in propane and butane production. The yield of LPG increases significantly when hydrocracking is compared to FCC, where the C3 –C4 fraction from FCC is olefinic and can be converted into motor-gasoline. Unless LPG is a desirable final product, it is worthwhile to include a conversion technology to convert light alkanes in the refinery design. References 1. De Klerk, A. (2009) in Advances in
Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, FL, pp. 331–364. 2. De Klerk, A. (2011) Fischer-Tropsch fuels refinery design. Energy Environ. Sci., 4, 1177–1205. 3. Hogue, C. (2007) Less benzene in gasoline. Chem. Eng. News, 85 (8), 8. 4. De Klerk, A. (2010) Deactivation in Fischer-Tropsch synthesis and its impact on
refinery design. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 55 (1), 86–89. 5. De Klerk, A. (2010) Indirect liquefaction carbon efficiency. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 55 (2), 338–339. 6. Choi, G.N., Kramer, S.J., Tam, S.S., Fox, J.M. III, and Reagan, W.J. (1996) Fischer-Tropsch indirect coal liquefaction design/economics mild hydrocracking vs. fluid catalytic cracking. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 41 (3), 1079–1083.
539
540
25 Motor-Gasoline Refining 7. Feldman, J. (1970) in Refining Petroleum for
8.
9.
10.
11.
12.
13. 14.
15.
Chemicals, Advances in Chemistry Series 97 (eds L.J. Spillane and H.P. Leftin), ACS, Washington, DC, pp. 242–263. De Klerk, A. and De Vaal, P.L. (2008) Alkylate technology selection for Fischer-Tropsch syncrude refining. Ind. Eng. Chem. Res., 47, 6870–6877. De Klerk, A. (2008) Fischer-Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. Leib, T.M., Kuo, J.C.W., Garwood, W.E., Nace, D.M., Derr, W.R., and Tabak, S.A. (1988) Upgrading of Fischer-Tropsch waxes to high quality transportation fuels. Proceedings AIChE Annual Meeting, Washington, DC, p. 61d. Albright, L.F. and Crynes, B.L. (eds) (1976) Industrial and Laboratory Pyrolyses, ACS Symposium Series 32, ACS, Washington, DC. Jones, J.L. and Radding, S.B. (eds) (1978) Solid Wastes And Residues. Conversion by Advanced Thermal Processes, ACS Symposium Series 76, ACS, Washington, DC. Nowacki, P. (1979) Coal Liquefaction Processes, Noyes Data Corporation, Park Ridge, NJ. Mohan, D., Pittman, C.U. Jr., and Steele, P.H. (2006) Pyrolysis of wood/biomass for bio-oil: a critical review. Energy Fuels, 20, 848–889. Singerman, G.M. (1982) Gasoline extenders/octane improvers from coal liquids. Energy Progr., 2 (2), 100–106.
16. Dancuart, L.P., De Haan, R., and De Klerk, A.
17.
18.
19.
20.
21.
22.
(2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532. Morgan, P.M., Van der Merwe, D.G., Goosen, R., Leckel, D.O., Saayman, H.M., Loubser, H., and Botha, J.J. (1999) The production of high quality base oils from Sasol hydrocrackates. S. Afr. Mech. Eng., 49, 11–13. Lamprecht, D. and De Klerk, A. (2009) Hydroisomerization of 1-pentene to iso-pentane in a single reactor. Chem. Eng. Commun., 196, 1206–1216. De Klerk, A. and Strauss, M.J. (2008) Opportunities for efficiency improvement in high temperature Fischer-Tropsch hydroprocessing units. Prepr. Pap.-Am. Chem. Soc., Div. Fuel. Chem., 53 (1), 313–314. Sakuneka, T.M., Nel, R.J.J., and De Klerk, A. (2008) Benzene reduction by alkylation in a solid phosphoric acid catalyzed olefin oligomerization process. Ind. Eng. Chem. Res., 47, 7178–7183. De Klerk, A., Engelbrecht, D.J., and Boikanyo, H. (2004) Oligomerization of Fischer-Tropsch olefins: effect of feed and operating conditions on hydrogenated motor-gasoline quality. Ind. Eng. Chem. Res., 43, 7449–7455. Knottenbelt, C. (2002) Mossgas ‘‘gas-to-liquids’’ diesel fuels – an environmentally friendly option. Catal. Today, 71, 437–445.
541
26 Jet Fuel Refining 26.1 Introduction
The refining of jet fuel from Fischer–Tropsch syncrude is not a very complicated task. In a conventional oil refinery, jet fuel is often produced by desulfurization alone [1]. Jet A-1 specifications are not nearly as demanding on a molecular level as that for motor-gasoline. Most of the complexity in a jet fuel refinery is consequently not due to jet fuel refining, but due to the conversion units needed to meet the specifications of other transportation fuel types. The Fischer–Tropsch refiner is burdened with additional specifications that are not imposed on crude oil refiners (Section 14.2.1). The additional specification requirements can be divided into two types. The first are quantitative constraints on the jet fuel properties (Table 14.2). These were necessary to ensure compatibility, such as the minimum aromatic content required, between crude-oil-derived and Fischer–Tropsch-syncrude-derived jet fuels. The second set of constraints is qualitative and places a restriction on the refinery configuration. The latter is ignored in the subsequent discussion based on the justification provided in Chapter 14. Jet fuel refining is discussed by focusing on the fuel properties of Jet A-1. Once refined to specification, a Fischer–Tropsch-derived jet fuel should have properties that make it indistinguishable from those of crude-oil-derived jet fuel, which is the ultimate intent of the jet fuel specifications. Discussion on jet fuel refining follows the same structure as that employed for the discussion of motor-gasoline refining in Chapter 25. Points of overlap are not repeated, and in this chapter the focus is specifically on how to refine syncrude to jet fuel. 26.2 Gap Analysis for Syncrude to Jet Fuel 26.2.1 Jet Fuel Specifications
Jet fuel specifications and how the specifications relate to the molecular properties of the kerosene were discussed in detail (Chapter 14). The specifications for Jet A-1 (Table 14.1), which is the most important jet fuel for commercial aviation, are used as the basis for the discussion. In addition, the requirements for synthetic Jet A-1 (Table 14.2) are adhered to. Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
542
26 Jet Fuel Refining
Freezing point is notoriously difficult to estimate with any amount of accuracy [2]. Out of necessity, statements about the freezing point of mixtures are based on the physical properties of the pure compounds available from the literature. This discounts solubility and mixture behavior. 26.2.2 Carbon Number Distribution
The boiling range of kerosene is usually around 160–260 ◦ C. Depending on the requirements of other fuels, it may be extended to 160–290 ◦ C [1]. For synthetic Jet A-1, the range is fixed by a hard specification on the upper end, with maximum final boiling point temperature of 300 ◦ C, and an indirect specification on the lower end, with a maximum flash point of 50 ◦ C. Nominally, in terms of n-alkane carbon number, it defines a cut between n-C9 and n-C17 . Although the specification allows the heavy material (260–290 ◦ C) to be included, the inclusion of heavier material affects the diesel fuel properties; the diesel fuel becomes very heavy. On the light end, 10% or more material must be in the 210 ◦ C and lower boiling range. The base stocks that in the past were used to prepare synthetic jet fuels from Fischer–Tropsch syncrude varied in carbon number distribution depending on the conversion technology employed. Syntroleum S-5 synthetic jet fuel had a carbon number distribution ranging from C9 to C18 and aimed to meet the heavier JP-5 carrier-based jet fuel specifications, with a minimum flash point of 60 ◦ C [3, 4]. A Fischer–Tropsch-derived JP-8 military jet fuel, which is similar to Jet A-1, was tested with C8 –C16 (mainly C9 –C15 ) material [5]. Synthetic jet fuels produced from the Sasol slurry phase distillate process were restricted to C9 –C15 (mainly C10 –C14 ) material [6]. Generally speaking, the carbon number distribution is limited on the light end by flash point requirements and on the heavy end by the freezing point requirements. However, jet fuels that are based only on isoparaffinic kerosene (IPK) require aromatic addition and have insufficient density to meet jet fuel specifications. It is therefore likely that the carbon number distribution on the light end will also be constrained by density requirements and not only by the flash point. The amount of straight-run syncrude that is in the correct boiling range for jet fuel production can be estimated from Anderson–Schulz–Flory distribution for different chain growth probabilities (Figure 26.1). It indicates that a Fischer–Tropsch catalyst with a chain growth probability (α-value) around 0.84 is optimal to maximize straight-run kerosene production. 26.2.3 Composition and Quality
The composition of Fischer–Tropsch-derived synthetic jet fuel is dominated by the alkanes, followed by the aromatics that must be included to meet the minimum aromatic specification of 8 vol%. The aromatics are also helpful to increase density. The lack of cycloalkanes in syncrude causes synthetic jet fuel to have a density at the lower end of the specification range of 775–840 kg·m−3 . Freezing point is central to the production of jet fuel. The freezing point and density of branched alkanes in the C11 –C15 range are shown in Figure 26.2. It illustrates the refining challenge involved in producing on-specification synthetic jet fuel. The boiling point of alkanes of the same carbon number decreases as the degree of branching increases, whereas the freezing
Yield of straight run kerosene (mass%)
26.2 Gap Analysis for Syncrude to Jet Fuel
45 40
Kerosene range
35
C9 – C17 C9 – C15
30
C10 – C14
25 20 15 Industrial LTFT operation
Industrial HTFT operation
10 5 0 0.5
0.6
0.7 0.8 0.9 Fischer–Tropsch chain growth probability
1
Figure 26.1 Yield of different straight-run kerosene cuts in the C3 and heavier hydrocarbon fraction from Fischer–Tropsch synthesis as calculated from the Anderson–Schulz–Flory distribution.
−20 Freezing point (°C)
840
C11 C12 C13 C14 C15 Density
−40
820 800
−60
780
−80
760
−100
740
−120
Density (kg •m−3)
0
720 170
190
210
230
250
270
Boiling point (°C) Figure 26.2 Freezing point and density of branched alkane isomers in the C11 –C15 kerosene range. Solid symbols are methyl branched isomers and open symbols are dimethyl branched isomers.
point on average decreases slightly with increase in the degree of branching. The densities of the C11 –C15 alkane isomers are close to or below the minimum jet fuel density. On the basis of the composition of straight-run Fischer–Tropsch syncrude, the following refining requirements can be identified:
543
544
26 Jet Fuel Refining
1) Jet fuel requires aromatics. High-temperature Fischer–Tropsch (HTFT) syncrude has the advantage over low-temperature Fischer–Tropsch (LTFT) syncrude in that it already contains some aromatics. Even so, the aromatics content in HTFT kerosene is sufficient for the straight-run kerosene to meet Jet A-1 specifications, but it is too little to accommodate much IPK. It will be necessary to produce additional kerosene range aromatics to meet the minimum aromatics requirement and to meet the minimum density requirement. The latter is likely to be the governing specification that determines the amount of aromatics that will have to be included in the synthetic jet fuel blend. 2) A large fraction of Fischer–Tropsch syncrude consists of molecules with a linear carbon structure. These molecules will have to be isomerized in order to meet the freezing point requirements (Table 14.4). When producing kerosene range material from lighter alkenes by oligomerization, the required level of branching will be introduced as a natural consequence of the conversion. Solid phosphoric acid (Section 19.3.1) is the prototypic ideal oligomerization catalyst for jet fuel production. Even with an oligomerization catalyst that limits the degree of branching, such as H-ZSM-5 (Section 19.3.2), the kerosene range oligomers will likely contain sufficient branching to meet the jet fuel freezing point specification of −47 ◦ C. When producing kerosene range material from heavier hydrocarbons, it is important to select a cracking catalyst that is very isomerizing. Cracking catalysts that are not very isomerizing limit the amount of heavier hydrocarbons that can be included in the jet fuel, as illustrated by the work of Lamprecht [6]. 3) Oxygenates are not allowed in jet fuel, except in the form of additives. The syncrude will therefore have to be deoxygenated. This will destroy the natural lubricity of syncrude, and synthetic jet fuel, like severely hydroprocessed crude-oil-derived jet fuel, requires a lubricity additive in the final blend. 4) Alkenes are not specifically excluded from jet fuel, but stability requirements preclude alkene inclusion. 26.3 Decisions Affecting Jet Fuel Refining
Many of the important decisions affecting motor-gasoline production (Section 25.3) affect jet fuel production in an analogous way. Additional discussion on jet-fuel-specific issues is included only for some of the following topics that are applicable to jet fuel too: 1) 2) 3) 4) 5)
Chemicals coproduction (Section 25.3.1) Fate of the C2 –C4 hydrocarbons (Section 25.3.2) Fate of the residue and wax (Section 25.3.3) Fate of the aqueous product (Section 25.3.4) Co-refining (Section 25.3.7).
26.3.1 Fate of C2 –C4 Hydrocarbons
Propene and butene can both be oligomerized to produce a kerosene range product with adequate branching. Solid phosphoric acid performs well in this application, because it limits the carbon
26.3 Decisions Affecting Jet Fuel Refining
number distribution of the product. When hydrogenated, this product is called isoparaffinic kerosene, which is one of the main blending components approved for semisynthetic and fully synthetic jet fuels (Section 14.2.1). The recovery of the C3 –C4 hydrocarbons is consequently very important to enable synthetic jet fuel production. Syncrudes that contain more C3 –C4 alkenes have an advantage over syncrude that are more paraffinic, or has less material in this boiling range. The combined oligomerization and alkylation of aromatics in a single unit makes it possible to produce an on-specification jet fuel directly [7]. The impact of ethene recovery is more pronounced in an HTFT refinery than in an LTFT refinery. The preferred refining pathway for ethene in a jet fuel refinery is not hydration, as in the case of motor-gasoline. Recovery of ethene as alkylating alkene for aromatics has much more value in maximizing jet fuel production [8]. In this application, it is not necessary to separate the ethene and ethane. Aromatic alkylation with ethene not only produces jet fuel range aromatics through multiple alkylation of the aromatics but also frees the propene for IPK production. By converting ethene into alkyl aromatics, it also provides some blending flexibility, since the alkyl aromatics have good motor-gasoline properties. Likewise, in the case of benzene alkylation, the mono-alkylated aromatic, which is ethylbenzene, has value as a petrochemical product. The C2 –C4 alkanes are not converted during either oligomerization or aromatic alkylation. If LPG is an acceptable product, there is a clear product outlet for propane and butane but not for all of the ethane. Ethane is really a petrochemical precursor and, unless the refinery includes a thermal cracker for chemical coproduction, the potential applications for ethane are limited. Ethane can be recycled to the Fischer–Tropsch gas loop for conversion into syngas, it can be employed as fuel gas, and some ethane can be mixed with methane to produce substitute natural gas (SNG). 26.3.2 Fate of the Residue and Wax
The HTFT residue can be converted into an aromatic jet fuel component by hydrocracking. It is a small fraction of the syncrude and has a limited impact on jet fuel refining. On account of the large fraction of LTFT syncrude that is in the wax product, the selection of the wax conversion technology has a significant impact on jet fuel refining. The preferred refining pathway for jet fuel production depends more on the other products that need to be produced than on jet fuel itself. The following needs to be considered in selecting the cracking technology: 1) Hydrocracking is the most direct and efficient way to produce jet fuel from wax and it preferably requires a very isomerizing hydrocracking catalyst. The hydroisomerization that takes place during cracking is necessary to meet the freezing point specification. It also serves as a hydrotreatment step, thereby improving stability and reducing the heteroatom content. Depending on the severity of the hydrocracking, it is a useful technology for coproducing some distillate and lubricant base oil. 2) Catalytic cracking does not yield much jet fuel directly, but the light alkenes can be oligomerized and hydrogenated to produce jet fuel. The kerosene fraction from catalytic cracking is isomerized and only needs to be hydrogenated. Overall, this is less efficient than hydrocracking because it requires two or more conversion steps instead of just one. Despite
545
546
26 Jet Fuel Refining
the lower efficiency of catalytic cracking for jet fuel production, it may still be the preferred wax conversion pathway if motor-gasoline coproduction is important. 3) Thermal cracking is useful for petrochemical production. The light alkenes can be oligomerized to kerosene range material, but the straight-run kerosene from thermal cracking is not skeletally isomerized. It is consequently the least preferred of the cracking technologies for jet fuel production.
26.3.3 Technology Selection
In order to produce jet fuel from syncrude, the refining technologies have to address the low aromatics content, high alkane linearity, removal of oxygenates, and improvement in the yield of kerosene range material. Using carbon-number-based refining, four carbon number ranges can be identified that must be considered for jet fuel refining (Figure 26.3): 1) Light hydrocarbons. The technologies considered for light hydrocarbons are mainly aromatization, aromatic alkylation, and oligomerization. Aromatization is a way to produce much needed aromatics from light hydrocarbons. The advantage of aromatization technology is that it can convert alkanes and alkenes. When employed in conjunction with hydrocracking of heavier syncrude, it provides a pathway for converting the light alkane fraction into aromatics. The light alkenes can be used as alkylating agents for aromatics to produce kerosene range aromatics from naphtha range aromatics. The synergy between aromatization and aromatic alkylation technology was pointed out earlier (Figure 22.7). Oligomerization in combination with alkene hydrogenation is a good source of IPK. C2 Aromatic alkylation
Aromatics
C3 C4
Aromatics
Aromatization
C5 C6 – C8
Oligomerization
Alkene hydrogenation
Naphtha hydrotreating
Naphtha reforming
Alkanes Aromatics
C9 – C10 Hydroisomerization
Alkanes
Hydrocracking
Alkanes
C11 – C15 >C15 Figure 26.3
Technology selection for the refining of syncrude to jet fuel.
26.3 Decisions Affecting Jet Fuel Refining
2) Naphtha. Depending on the technology selection for the light hydrocarbons, catalytic naphtha reforming (preferably by nonacidic Pt/L-zeolite-based technology) or aromatization can be employed to produce aromatics for jet fuel. Aromatization is not the preferred technology for naphtha because of the associated liquid yield loss, but co-feeding the naphtha with the light hydrocarbons can be considered as strategy to avoid duplicating aromatics-producing technologies in the refinery. Naphtha range alkenes can also be oligomerized to kerosene, but it requires appropriate catalyst selection. Zeolitic or amorphous silica–alumina-based catalysts are preferred for this application. The heavy (C9 –C10 ) naphtha can also be considered as light kerosene for jet fuel refining. 3) Kerosene. The kerosene range material is in the right boiling range for jet fuel production, but on account of the high linear hydrocarbon content, its freezing point is too low. Hydroisomerization is the most appropriate technology to adjust the properties of the kerosene fraction for jet fuel. It increases the degree of branching to lower the freezing point and removes heteroatoms and saturates the alkenes. 4) Gas oil and heavier material. Hydrocracking is the most appropriate technology for jet fuel production. The selection of the cracking technology may, however, be influenced by other factors (Section 26.3.2). Hydrocracking performs the same functions as hydroisomerization, in addition to cracking to shift the carbon number distribution to lower carbon numbers. 26.3.4 Co-refining
The two main compound classes lacking in straight-run Fischer–Tropsch syncrude are cycloalkanes and aromatics. Both classes on average have good cold-flow properties and both classes are complementary to linear and branched alkanes with respect to density (Table 26.1). A minimum amount of aromatics is required by synthetic jet fuel specifications (Table 14.2), but the specifications do not require any cycloalkanes. The value of cycloalkanes in jet fuel is related to their generally good cold-flow properties combined with their density. Because of the high acyclic aliphatic content of straight-run syncrude and the IPK derived from alkene oligomerization and wax hydrocracking, Fischer–Tropsch-derived kerosene range material mostly has a low density. Cycloalkanes are good density boosters and they move the kerosene density away Average density and freezing point properties of compound classes acceptable in jet fuel.
Table 26.1
Description
Jet A-1 specifications Compound class n-Alkanes Branched alkanes Cycloalkanes Aromatics (mononuclear)
Properties of kerosene range material Density (kg·m –3 )
Freezing point (◦ C)
775–840
<–47
720–770 720–770 790–820 860–870
–50 to +10 –100 to −10 –110 to −40 –100 to −30
547
548
26 Jet Fuel Refining
from the lower density specification limit, thereby relieving a constraint on jet fuel blending and increasing the robustness of the refinery. In the absence of cycloalkanes, aromatics are responsible for all of the density increase necessary to meet specification, and the volume of aromatics in the final blend may be dictated by density rather than the minimum aromatics specification. The benefit of blending Fischer–Tropsch kerosene with crude-oil-derived kerosene has been demonstrated by the success of semisynthetic jet fuel (Section 14.2.1). There is also a significant advantage that can be obtained by blending Fischer–Tropsch-derived kerosene with coal-liquid-derived kerosene fractions. The synthetic jet fuel produced from direct coal liquefaction is rich in cycloalkanes and aromatics and has a correspondingly high density [9–11]. The coal liquids that are produced as by-product during low-temperature coal gasification can fulfill the same role. The naphtha range aromatics may also be alkylated to produce kerosene range alkyl aromatics for blending. This type of co-refining has been successfully demonstrated on an industrial scale [12]. The coproduction of coal liquids with Fischer–Tropsch liquids is consequently desirable both from a blending and co-refining perspective.
26.4 Jet Fuel Refining from HTFT Syncrude
The amount of kerosene and heavier material in HTFT syncrude is very small compared to the amount of syncrude in the naphtha and lighter range. Synthetic jet fuel production from HTFT syncrude is mainly a building-up process. Most of the kerosene range material must be synthesized from lighter fractions, and the alkene content of the syncrude determines the yield that can be obtained. The main technologies that will be required to produce jet fuel from lighter material are alkene oligomerization followed by hydrogenation to produce IPK, reforming or aromatization to produce aromatics, and aromatic alkylation to convert naphtha range aromatics into kerosene range aromatics. Luckily, these conversion processes are also valuable for motor-gasoline production. The straight-run kerosene and heavier material can be converted into jet fuel by mild hydrocracking and hydroisomerization. The severity of this operation will determine the yield of kerosene range material. Hydrocracking can in principle convert all of the heavier material into kerosene range and lighter products. It is consequently possible to conceive a Fischer–Tropsch jet fuel refinery that employs only the five technologies mentioned: namely, oligomerization, aromatization (or reforming), aromatic alkylation, hydrotreating, and hydrocracking. This type of minimalist design will be explored in Section 26.4.1 to confirm whether this is indeed the case. The second HTFT jet fuel design case (Section 26.4.2) will explore the design considerations and limitations when maximum jet fuel yield is the main objective. As in the refining of motor-gasoline from HTFT syncrude (Chapter 25), the high carbon cost associated with light hydrocarbons and the aqueous product precludes designs that do not refine these streams. All HTFT designs are consequently based on a closed gas loop with methane recycle, where cryogenic recovery of the C2 hydrocarbons is implicit in the gas loop design.
26.4 Jet Fuel Refining from HTFT Syncrude
26.4.1 HTFT Jet Fuel Design Case I
A minimalist design with the least amount of conversion units to produce jet fuel from Fischer–Tropsch syncrude was reported before [8]. It highlighted the importance of the design decisions surrounding ethene refining and the selection of the alkene oligomerization technology. It also confirmed that the technologies highlighted in Section 26.4 are very important for jet fuel refining. The robustness of a design based on these technologies is explored, without resorting to such an extremely minimalist design. A jet fuel design is presented that employs technologies with good HTFT fit (Figure 26.4). It is minimalist with respect to hydrocarbon conversion, keeping to the jet fuel technologies that were highlighted (Section 26.4), but also includes an aqueous product refinery. The design illustrates the role of the prominent technologies for jet fuel refining and highlights the overlap and constraints that are imposed on the refining of other products. Ethene is used as alkylating agent for benzene to produce mainly diethylbenzenes for addition to the jet fuel. The amount of ethene exceeds the amount of benzene that is available, and in this respect it is not an optimal choice. By operating the benzene alkylation unit at low aromatic to alkene ratio, some ethene oligomerization will likely take place, which is actually desirable in the current context. Carbonyl hydrogenation
Aqueous product
Wastewater C3+ alcohol dehydration
HTFT gas loop
C3 C3 – C4
C4
Alcohols Fuel gas Gasoline Jet fuel
Aromatic alkylation
C2
Ethanol
LPG Gasoline
C3 SPA oligomerization Alkene hydrogenation
C4 SPA oligomerization
Condensate C5 – C6
Gasoline Jet fuel Butanes Gasoline
C5 Benzene Nonacidic Pt / L reforming
Light oil Hydrotreater
Decanted oil
Hydrocracking >C11
Figure 26.4 HTFT jet fuel refinery design case I. The refinery design is minimalist, but based on recommended technologies. The gas loop includes cryogenic C2 separation with methane recycle.
H2 LPG Gasoline Jet fuel Jet fuel Diesel
549
550
26 Jet Fuel Refining
The combination of selective carbonyl hydrogenation and dehydration of the C3 and heavier alcohols boosts propene production and thereby jet fuel production. The ethanol is retained as a motor-gasoline blending component. In this way, it is possible to recover and refine the oxygenates that are lighter boiling than water. Solid phosphoric acid–based oligomerization of propene followed by hydrogenation yields high-quality IPK. The oligomerization of butene is conducted separately to produce some jet fuel, but mainly to produce an alkylate equivalent product for motor-gasoline. The volume of kerosene produced by light alkene oligomerization can be increased by judicious selection of the catalyst, process configuration, and operating conditions for oligomerization. The advantage of the present design is flexibility and product quality. In order to increase kerosene production, the hydrocracker can be operated at more severe conditions, with all heavier than kerosene material being recycled to extinction. Without changing the basic design shown in Figure 26.4, changes were evaluated to probe the robustness of the design. The cut point and blending of streams were manipulated, and units were turned down to assess the impact on the product distribution and product quality (Table 26.2). The following cases were considered: 1) Base case design. The refining strategy that was employed, converted ethene, aqueous product oxygenates, and heavier than kerosene range material into jet fuel and motor-gasoline. The oligomerization technology was not selected for extreme jet fuel production. A ratio of jet fuel to motor-gasoline of <1 can be obtained from the oligomerization of the lighter than kerosene range alkenes over an amorphous silica–alumina or even zeolite catalyst (Chapter 19). In the base case, only the C3 –C4 alkenes were converted into kerosene range material over SPA. Most of the C5 material was included as a straight-run product, and the C6 and heavier material was reformed in a nonacidic Pt/L-zeolite-catalyzed process. All C6 and heavier material was hydrotreated before distillation to avoid boiling point broadening. Jet fuel and motor-gasoline met the fuel specifications. However, the motor-gasoline properties were close to the specification limits on many properties, including RON, aromatic content, and alkene content. The blending is sensitive to the volume of straight-run C5 material that is included in the final motor-gasoline, and in this respect the design is not very robust. 2) No alcohol dehydration. The C3 and heavier alcohols have value as petrochemical products. When these alcohols are not dehydrated to alkenes, the jet fuel yield is lower, but it does not adversely affect the product quality. 3) Mild hydrocracking. The cracking severity of the hydrocracker can be reduced to produce jet fuel and diesel fuel from the straight-run HTFT distillate and residue fractions. It causes the jet fuel to become richer in aromatics, and overall the refinery runs close to the maximum aromatics specification. Meeting the jet fuel specification presented no problem for the design shown in Figure 26.4. It is therefore a robust jet fuel refinery design. It also confirms that only five conversion units are important for jet fuel refining from HTFT syncrude: namely oligomerization, reforming, aromatic alkylation, hydrotreating, and hydrocracking. Yet, one has to look at the overall design. Motor-gasoline was the main product, not jet fuel. It was somewhat surprising to find that on-specification motor-gasoline could be produced by a jet fuel refinery design. This was mainly due to the selection of solid phosphoric acid–based oligomerization technology. The blending constraints for motor-gasoline can easily be overcome
26.4 Jet Fuel Refining from HTFT Syncrude Table 26.2
Products from HTFT jet fuel refinery design case I shown in Figure 26.4.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Nett productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t –1 ) Fuel gas (kg·t –1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m –3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m –3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m –3 )
Base case
No dehydration
0.2 6.8 4.9 46.0 35.6 0.0 1.6 0.0 3.0 1.9
0.2 6.8 4.9 46.0 32.3 0.0 4.1 0.0 5.0 0.6
Mild hydrocracking
0.3 6.6 3.9 48.2 29.0 5.4 1.6 0.0 3.1 1.9
0.092 0.610 0.457 –
0.092 0.610 0.413 –
0.073 0.635 0.371 0.065
1.159 16 69
1.114 41 70
1.144 16 70
95.1 87.1 754 60 17.1 32.6 0.3 4.9
95.1 87.1 754 60 17.1 32.6 0.3 4.9
95.8 87.5 759 60 17.5 34.6 0.3 4.9
779 21.2
783 23.4
782 24.9
– –
– –
61 835
a On a syncrude basis, excluding H2 , CO, CO2 , H2 O, and CH4 from HTFT synthesis (closed gas loop with cryogenic separation). b Syncrude waste products (e.g., metal-containing fuel oil) and unrecoverable material (e.g., carboxylic acids in wastewater).
by adding an appropriate refining unit to upgrade the C5 naphtha which was included as a straight-run product. Analysis of the design in Figure 26.4 shows that the C5 naphtha is actually a pivot point in the refinery. Including the C5 naphtha in the motor-gasoline has many consequences for both motor-gasoline and jet fuel production. The most important impact on jet fuel production is the
551
552
26 Jet Fuel Refining
decrease in yield. The C5 naphtha not only represents 10% of the overall C2 and heavier HTFT syncrude, but also requires additional aromatics to balance the motor-gasoline. A large volume of material is therefore locked into motor-gasoline. The C5 naphtha significantly increases the vapor pressure and alkene content of the motor-gasoline. The increased vapor pressure has the knock-on effect of reducing the volume of butane that can be blended into the fuel. All of these factors contribute to the observed constraints in motor-gasoline blending. Now that we have established the most important conversion units and the pivotal role of the C5 refining pathway, it is worthwhile to consider the maximum extent of jet fuel production that is possible. 26.4.2 HTFT Jet Fuel Design Case II
It has been reported that, by selecting a distillate producing oligomerization technology, the ratio of kerosene to naphtha range material can be increased considerably [8]. The disadvantages of this approach were also noted. It led to a complex design, with inefficient refining pathways for some carbon numbers that had to pass through multiple units before becoming a final product. The motor-gasoline quality that is obtained from H-ZSM-5 (Table 19.4) and amorphous silica–alumina (Table 19.5) based oligomerization is comparatively poor even as an olefinic motor-gasoline. Although the objective is to produce maximum jet fuel, it is not prudent to create a refining problem for motor-gasoline production while doing so. A solid phosphoric acid oligomerization process can be operated to have a high kerosene selectivity, especially if the cut point is around C9 . The two main jet fuel properties affected by a low cut point are density and flash point. Since the oligomers will not be the only jet fuel component and aromatics need to be added to meet the minimum aromatics specification of synthetic jet fuel, one would have to keep an eye on these properties, but they are not immediately problematic. The high alkene content in the C5 naphtha can therefore be exploited in conjunction with propene to produce kerosene in high yield. Furthermore, some aromatics can be co-fed for alkylation in combination with oligomerization [7, 12], in order to address some of the concerns noted. It is possible to oligomerize the complete C3 –C5 fraction together, but the desirability of separate solid phosphoric acid–catalyzed dimerization of butenes for motor-gasoline production has been demonstrated (Chapter 25). Although it seems like an increase in refinery complexity, it is only minor. The volume of alkenes to be oligomerized and the associated engineering issues related to heat management dictate that multiple reactors will have to be employed in parallel. Prefractionating the feed increases only the separation requirements, but the payback in terms of product quality (Table 19.3) is significant [13]. A design employing this oligomerization strategy for the C5 naphtha to increase jet fuel yield is shown in Figure 26.5. The design in Figure 26.5 is much less constrained than the design in Figure 26.4, with little added complexity. The jet fuel yield is also increased significantly (Table 26.3). Yet, the fundamentally sound jet fuel design philosophy has been retained and, when the same scenarios are evaluated as before (Section 26.4.1), it confirms that the refinery design is much more robust in addition to having a higher jet fuel yield. In fact, by slightly adjusting the volume of material routed to the reformer, it is possible to meet the motor-gasoline specifications even without ethanol addition (Table 26.3). This makes on-specification fuel blending independent of the aqueous product refinery.
26.5 Jet Fuel Refining from LTFT Syncrude Carbonyl hydrogenation
Aqueous product
Wastewater C3+ alcohol dehydration
HTFT gas loop
C3 –C4
C3
Mixed SPA oligomerization
C4
C4 SPA oligomerization
Condensate C5 –C6
Alcohols Fuel gas Gasoline Jet fuel
Aromatic alkylation
C2
Ethanol
LPG Gasoline Alkene hydrogenation
C5 Benzene Nonacidic Pt / L reforming
Light oil Hydrotreater
Decanted oil
Hydrocracking >C9
Gasoline Jet fuel Butanes H2 LPG Gasoline Jet fuel Jet fuel Diesel
Figure 26.5 HTFT jet fuel refinery design case II. The refinery design is very similar to design case I, but C5 is the pivot for additional jet fuel production. The gas loop includes cryogenic C2 separation with methane recycle.
Diesel fuel production undermines the jet fuel yield and constrains the blending, requiring the straight-run C9 –C10 fraction to be reformed in order to meet motor-gasoline specifications. 26.5 Jet Fuel Refining from LTFT Syncrude
Producing jet fuel in high yield from LTFT syncrude is easier than from HTFT syncrude. It has been shown that a hydrocracker-based design for motor-gasoline production from LTFT syncrude produces jet fuel as major product (Section 25.5.3). Fundamentally, the conversion sequence is attractive. A heavy and predominantly aliphatic hydrocarbon-based feed is hydrocracked to yield a product in the kerosene range. The kerosene is isomerized sufficiently to meet the jet fuel freezing point specification and can be directly employed as an IPK base stock for jet fuel blending. The material that has been overcracked to naphtha can be catalytically reformed to produce aromatics for alkylation and direct blending into the IPK to satisfy the density and minimum aromatic content requirements for synthetic jet fuel. Wax hydrotreating and wax hydrocracking, which cause wax hydroisomerization, are also the technologies that form the basis for petrochemical and lubricant base oil production from LTFT syncrude (Chapter 28). Coproducing chemicals and lubricants decrease the jet fuel yield. The severity of the hydrocracker and what fraction of material is recycled to the hydrocracker determine what will be produced.
553
554
26 Jet Fuel Refining Table 26.3
Products from HTFT jet fuel refinery design case II shown in Figure 26.5.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t –1 ) Fuel gas (kg·t –1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m –3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m –3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m –3 )
Base case
No dehydration
Mild hydrocracking
–0.1 6.8 4.2 30.0 51.8 0.0 2.4 0.0 3.0 1.9
0.0 6.8 4.2 30.0 48.5 0.0 4.9 0.0 5.1 0.5
0.1 6.6 3.2 31.4 45.9 5.4 2.4 0.0 3.1 1.9
No ethanol addition
0.0 6.8 4.0 31.8 48.5 0.0 4.0 0.0 3.0 1.9
0.080 0.397 0.668 –
0.079 0.397 0.624 –
0.061 0.412 0.592 0.065
0.076 0.417 0.625 –
1.144 24 67
1.100 49 67
1.130 24 67
1.117 40 68
95.9 89.1 756 60 6.4 29.4 0.4 5.0
95.9 89.1 756 60 6.4 29.4 0.4 5.0
96.3 89.4 762 59 6.3 31.8 0.4 4.8
95.0 88.7 761 53 6.1 34.2 0.4 0.0
775 20.2
777 21.6
775 22.7
777 21.3
– –
– –
61 835
– –
a On a syncrude basis, excluding H2 , CO, CO2 , H2 O, and CH4 from HTFT synthesis (closed gas loop with cryogenic separation). b Syncrude waste products (e.g., metal-containing fuel oil) and unrecoverable material (e.g., carboxylic acids in wastewater).
If jet fuel is the only fuel that has to meet the specifications, it is in principle possible to produce a high yield of jet fuel with an aromatics-rich naphtha as a coproduct by employing only four conversion units in the jet fuel refinery (Figure 26.6). In this design, aromatic alkylation and oligomerization are combined and an analogous type of design has been presented before [8], indicating that a 75 : 25 ratio between jet fuel and aromatic naphtha is possible. Aqueous product
26.5 Jet Fuel Refining from LTFT Syncrude
Alcohols Aqueous
Wastewater SPA alkylation oligomerization
C3 –C4
Hydrotreater
Jet fuel
C5 LPG and naphtha
LTFT gas loop
Hot and cold condensate
Benzene and some toluene
M/H-ZSM-5 aromatization
H2 Fuel gas Naphtha Jet fuel
Hydrocracker
Jet fuel
>C9
Wax C15 and heavier
Figure 26.6
Uncomplicated LTFT jet fuel refinery design with aromatic naphtha and alcohol coproduction.
refining is limited to the recovery of oxygenates lighter boiling than water. Since the refinery does not produce motor-gasoline, the alcohol-rich oxygenate product, such as the aromatic naphtha, is a potential petrochemical feed material. The main shortcoming of the design shown in Figure 26.6 is its inability to produce final on-specification transportation fuels. In order to be useful as a strategic asset, for example, to produce jet fuel for the military [14], the naphtha range material should also be a final fuel for ground transportation use. Because of the increasing interest in this specific application (see also Section 14.2.2), it will be an important objective in the development of an LTFT jet fuel refinery design. 26.5.1 LTFT Jet Fuel Design Case I
Producing Jet A-1 from LTFT syncrude is not challenging. LTFT jet fuel refinery design is consequently not primarily concerned with the strategy for jet fuel production but in what way jet fuel can be produced so that the motor-gasoline also meets specification. For reasons that are apparent in Chapter 27, diesel fuel production is avoided. One of the endearing aspects of the uncomplicated LTFT jet fuel refinery design shown in Figure 26.6 is that it does not coproduce LPG. The LPG is converted into aromatics using an aromatization technology (Section 22.5), rather than producing aromatics by catalytic naphtha reforming. In applications focused on military transportation, this is a bonus because LPG is not a desired product. Furthermore, the fuel gas produced by aromatization can potentially be recycled with the C2 and lighter tail gas from Fischer–Tropsch synthesis to increase syngas production. It is possible to design an LTFT jet fuel refinery based on nonacidic Pt/L-zeolite reforming that produces on-specification jet fuel and motor-gasoline. Such designs have been presented in the literature [8], and is shown in Figure 25.10. Since severe hydrocracking inevitably produces
555
556
26 Jet Fuel Refining Carbonyl hydrogenation
Aqueous
Ethanol Alcohols Wastewater Butane Gasoline
SPA alkylation oligomerization
C3 –C4 C5
LTFT gas loop
Hydrotreater
Jet fuel
C5 –C6 hydroisomerization
Gasoline
M/H-ZSM-5 aromatization
H2 Fuel gas Gasoline Jet fuel
Light aromatics LPG Hot and cold condensate
C6 –C8 C5 –C6
>C9 Hydrocracker
Jet fuel
Wax C15 and heavier
Figure 26.7 LTFT jet fuel refinery design case I. The gas loop includes C3 –C4 recovery with recycle of the C1 –C2 hydrocarbons in the tail gas.
C3 –C4 alkanes and the straight-run LTFT syncrude contains C3 –C4 alkanes, LPG is a major by-product. There is consequently considerable incentive to select an aromatization technology over a reforming technology for the production of aromatic hydrocarbons when LPG is not a desired product. As in the case of HTFT jet fuel refining, the C5 –C6 material forms the pivot point for balancing the refinery design (Figure 26.7). In order to produce on-specification motor-gasoline, it is only necessary to add a hydroisomerization unit to the design presented in Figure 26.6. The hydroisomerization unit isomerizes the C5 –C6 products from hydrocracking (and optionally the straight-run alkanes from the LTFT syncrude). This unit requires full recycle to achieve the highest possible octane number, since it becomes the main source of high octane alkanes in the motor-gasoline. Additional octane number improvement is provided by blending ethanol from the aqueous product into the motor-gasoline. This is adequate for producing on-specification motor-gasoline and jet fuel from Fe-LTFT syncrude but not from Co-LTFT syncrude (Table 26.4). There are just too little alkenes in Co-LTFT to alkylate the benzene and produce sufficient olefinic motor-gasoline to meet the octane number specification. The lack of an acceptable quality aliphatic heavy naphtha blending component causes the motor-gasoline to fail the octane number specification. Furthermore, as the Co-LTFT catalyst deactivates the refining, the situation will become worse. The opposite
26.5 Jet Fuel Refining from LTFT Syncrude Table 26.4
Products from LTFT jet fuel refinery design case I shown in Figure 26.7.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg t –1 ) Fuel gas (kg t –1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg m –3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg m –3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg m –3 )
Fe-LTFT
–0.6 3.0 0.0 35.2 59.4 0.0 1.4 0.0 1.5 0.2
Co-LTFT
–0.6 3.3 0.0 31.9 64.1 0.0 0.4 0.0 0.6 0.2
– 0.477 0.764 –
– 0.433 0.821 –
1.240 14 24
1.254 4 27
95.3 87.7 739 59 13.3 33.2 0.1 4.7
93.5 87.0 735 59 6.3 35.0 0.1 2.5
777 17.8
781 21.2
– –
– –
a On a syncrude basis, excluding H2 , CO, CO2 , H2 O, and C1 –C2 hydrocarbons from LTFT synthesis (closed gas loop with C3 –C4 recovery from the tail gas). b Syncrude waste products and unrecoverable material (e.g., carboxylic acids in wastewater).
is true for Fe-LTFT, where deactivation of the Fe-LTFT catalyst will benefit refining [15], and specifically benefit the design presented in Figure 26.7. Jet fuel production is clearly more efficient from a more olefinic Fe-LTFT syncrude than from a more paraffinic Co-LTFT syncrude. The oxygenates that are recovered from the aqueous product and that are not used in transportation fuel blending are listed as chemical by-products. Depending on the purpose of the refinery, these oxygenates may be used as refinery fuel.
557
558
26 Jet Fuel Refining
References 1. Bishop, G.J. (2008) in Handbook of Fuels (ed.
2.
3.
4.
5.
6.
7.
8.
B. Elvers), Wiley-VCH Verlag GmbH, Weinheim, pp. 321–341. Reid, R.C., Prausnitz, J.M., and Poling, B.E. (1987) The Properties of Gases and Liquids, 4th edn, McGraw-Hill, New York. Freerks, R.L. and Muzzell, P.A. (2004) Production and characterization of synthetic jet fuel produced from Fischer-Tropsch hydrocarbons. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 407–410. Muzzell, P.A., Freerks, R.L., Baltrus, J.P., and Link, D.D. (2004) Composition of Syntroleum S-5 and conformance to JP-5 specification. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 411–413. Corporan, E., DeWitt, M.J., Belovich, V., Pawlik, R., Lynch, A.C., Gord, J.R., and Meyer, T.R. (2007) Emissions characteristics of a turbine engine and research combustor burning a Fischer-Tropsch jet fuel. Energy Fuels, 21, 2615–2626. Lamprecht, D. (2007) Fischer-Tropsch fuel for use by the U.S. military as battlefield-use fuel of the future. Energy Fuels, 21, 1448–1453. Sakuneka, T.M., De Klerk, A., Nel, R.J.J., and Pienaar, A.D. (2008) Synthetic jet fuel production by combined propene oligomerization and aromatic alkylation over solid phosphoric acid. Ind. Eng. Chem. Res., 47, 1828–1834. De Klerk, A. (2011) Fischer-Tropsch fuels refinery design. Energy Environ. Sci., 4, 1177–1205.
9. Schobert, H.H., Beaver, B., Rudnick, L.,
10.
11.
12.
13.
14.
15.
Santoro, R., Song, C., and Wilson, G. (2004) Progress toward development of coal-based JP-900 jet fuel. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 493–497. Berkhous, S.K. and Schobert, H.H. (2004) Freeze point determination of prototype JP-900 fuels. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 49 (4), 510–512. Burgess Clifford, C. and Schobert, H.H. (2007) Development of coal-based jet fuel. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 52 (2), 403–404. Sakuneka, T.M., Nel, R.J.J., and De Klerk, A. (2008) Benzene reduction by alkylation in a solid phosphoric acid catalyzed olefin oligomerization process. Ind. Eng. Chem. Res., 47, 7178–7183. De Klerk, A., Engelbrecht, D.J., and Boikanyo, H. (2004) Oligomerization of Fischer-Tropsch olefins: Effect of feed and operating conditions on hydrogenated motor-gasoline quality. Ind. Eng. Chem. Res., 43, 7449–7455. Forest, C.A. and Muzzell, P.A. (2005) Fischer-Tropsch fuels: why are they of interest to the United States military? SAE Tech. Pap. Ser., 2005–01–1807. De Klerk, A. (2010) Deactivation in Fischer-Tropsch synthesis and its impact on refinery design. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 55 (1), 86–89.
559
27 Diesel Fuel Refining 27.1 Introduction
Divergent opinions have been voiced about the quality and the desirability of the diesel fuel produced from Fischer–Tropsch distillates. The diesel fuel produced by the original German cobalt-based low-temperature Fischer–Tropsch (LTFT) processes was considered an inferior fuel. The Fischer–Tropsch-derived diesel fuel did not meet the German diesel fuel specifications of that time, resulting in a higher fuel consumption, and had to be blended with crude-oil- or coal-liquid-derived distillates to make it acceptable as diesel fuel (Section 6.3.3) [1, 2]. Conversely, from a mechanical engineering perspective, an alkane-rich distillate makes an excellent fuel for compression-ignition engines [3]. Numerous studies appeared in the literature dealing with the lower exhaust emission profile from LTFT distillates, but not all studies were equally favorable. The combustion chemistry in a compression-ignition engine is complex, and emission properties vary with load and timing [4]. The adage is true: ‘‘The same elephant viewed from different angles looks differently.’’ The diesel fuel prepared from high-temperature Fischer–Tropsch (HTFT) distillate resembles that prepared from crude-oil-derived distillate. The refinery designs for producing motor-gasoline (Chapter 25) and jet fuel (Chapter 26) from Fischer–Tropsch syncrude produced little or no diesel fuel. This may seem at variance with the focus on distillate production seen in recent Fischer–Tropsch-based gas-to-liquids facilities (Chapters 11 and 12). The distillate obtained from LTFT synthesis has a very high cetane number. Yet, there is a difference between distillate, which meets the boiling point criteria for compression-ignition fuel, and on-specification diesel fuel that has to meet all the fuel specifications associated with fuel for compression-ignition engines. In this chapter, we explore diesel fuel refining. As in the two previous chapters, the objective is the production of on-specification transportation fuel from syncrude by appropriate refinery design. Like the other transportation fuels, the carbon number distribution of the syncrude must be shifted to the distillate range and the quality of the distillate must be improved to meet fuel specifications.
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
560
27 Diesel Fuel Refining
27.2 Gap Analysis for Syncrude to Diesel Fuel 27.2.1 Diesel Fuel Specifications
The properties of diesel fuel, diesel fuel specifications, and how the molecular properties relate to the specifications are discussed in Chapter 15. Diesel fuel specifications are country-dependent and change over time. In order to keep the discussion general but quantitative, a subset of specifications will be employed for refinery design (Table 27.1). The European EN590:2004 specification (Table 15.1) is used as benchmark. However, a lower polycyclic aromatic content is used as suggested by directive 2009/30/EC of the European parliament and of the council of the European Union [5]. The most challenging diesel fuel specification for syncrude refining is diesel fuel density. Not all diesel fuel specifications include a lower density limit, since it is not an issue when refining diesel fuel from crude oil. The lower volumetric fuel economy obtained with LTFT distillate is directly related to the fuel density. The density also influences the stoichiometry of the air–fuel mixture during combustion. Unless an engine has been specifically calibrated for LTFT distillate, the LTFT distillate will result in leaner combustion than crude-oil-derived diesel fuel. In the absence of a lower density limit in fuel specifications, it is easy to refine Fischer–Tropsch syncrude in high volume to diesel fuel. When there is a lower density limit, it becomes a constraining specification (Section 27.2.4). All of the diesel fuel specifications are important, but not all are easy to deal with in a quantitative manner. For example, it is important to ensure that the diesel fuel has good cold-flow properties and viscosity, but estimating the cold-flow properties or the liquid viscosity of pure compounds and mixtures is associated with large errors [6]. The discussion is therefore limited to qualitative statements about such properties based, where possible, on data available from the literature. Because of the storage stability issues noted in the literature [7], refinery designs excluded blending of bio-derived material as fatty acid methyl esters (FAME). It is nevertheless discussed as an avenue for diesel fuel improvement. It should also be noted that FAME blending can Table 27.1
Diesel fuel specifications employed for syncrude refinery designs.
Fuel property
Restrictions on boiling range Density at 15 ◦ C (kg·m−3 ) T95 distillation (◦ C) Restrictions on composition Cetane number Polycyclic aromatic content (vol%) Sulfur content (µg·g−1 )
Fuel specification Minimum
Maximum
820 –
845 360
51 – –
– 8 10
27.2 Gap Analysis for Syncrude to Diesel Fuel
potentially reduce some specification constraints of syncrude-derived diesel fuel, for example, diesel density. 27.2.2 Carbon Number Distribution
The amount of straight-run syncrude that is in the distillate boiling range can be estimated from the Anderson–Schulz–Flory distribution for different chain growth probabilities (Figure 27.1). Jet fuel production affects the availability of material for diesel fuel production. In the absence of jet fuel production, the maximum amount of straight-run distillate is obtained from synthesis over a Fischer–Tropsch catalyst with a chain growth probability (α-value) around 0.88. Additional distillate can be produced during refining by oligomerization of olefinic material lighter than the distillate and cracking of material heavier than the distillate. The actual amount of distillate that can be produced by such a simple distillate refinery (Figure 27.2) depends somewhat on the technologies selected, but increases in the order Fe-HTFT < Co-LTFT < Fe-LTFT [8]. Although LTFT syncrude with a higher α-value contains less straight-run distillate, with an appropriate hydrocracking technology it is possible to produce distillate range syncrude with higher selectivity (∼75%) than by optimizing the α-value for straight-run production. The alkene content of LTFT syncrude also contributes to the yield of the distillate that can be produced. The distillate yield from alkene oligomerization (∼65%) also exceeds the yield increase that can be obtained by optimizing the α-value. Refining technology is consequently more selective for increasing distillate production than manipulating Fischer–Tropsch synthesis.
Yield of straight run distillate (mass%)
40 Distillate range C11−C22 C15−C22
35 30 25
Industrial HTFT operation
20 15 10
Industrial LTFT operation
5 0 0.6
0.7
0.8
0.9
Fischer−Tropsch chain growth probability Figure 27.1 Yield of different straight-run distillate cuts in the C3 and heavier hydrocarbon fraction from Fischer–Tropsch synthesis as calculated from the Anderson–Schulz–Flory distribution.
1
561
562
27 Diesel Fuel Refining
<100 °C Aqueous
FT gas loop
C 3 −C 4
Wastewater
Oligomerization
C5 −C10
C5 −C10
Oil >C11
Wax/residue Figure 27.2
C3 −C4
Hydrocracker
C11 −C22
LPG
Naphtha Distillate
>C23
Generic Fischer–Tropsch refinery configuration to increase distillate yield.
Other technologies that result in an increase in molecular mass can likewise be employed to produce distillate range material from lighter feeds, for example, esterification, etherification, and aromatic alkylation. The upper boiling range of diesel fuel is directly constrained by the T95 boiling point requirement. The heaviness of the distillate is also indirectly regulated by properties such as the maximum density and the maximum viscosity, both of which increase as the amount of heavier material increases. The minimum flash point specification limits the amount of kerosene that can be used for jet fuel production and indirectly requires the inclusion of some lighter boiling material in the distillate. The lower boiling range is indirectly constrained by the minimum density and minimum viscosity requirements. The cetane numbers of lighter material are lower than that of heavier material and can impose an indirect constraint on the amount of lighter material that can be in included in the distillate. 27.2.3 Composition and Quality
Fischer–Tropsch syncrudes are sulfur free and do not have a cetane deficiency. The conversion strategies that are necessary to upgrade crude oil [9] are therefore not directly relevant to syncrude. In fact, the lack of sufficient cyclic structures in LTFT syncrude is an issue, because it causes a density deficiency. This is not a problem in HTFT syncrude, although careful refining is needed to ensure that the density-providing compounds are not destroyed during hydroprocessing [10]. There are many similarities between the composition and quality requirements for diesel fuel and jet fuel (Section 26.2.3). The requirements can be discussed in terms of compound classes: 1) The high content of linear hydrocarbons in syncrude requires that the feed must be hydroisomerized in order to improve the cold-flow properties. A high degree of isomerization is not required and is actually detrimental, because it lowers the cetane number. There is an inevitable trade-off between the amount of high-cetane-number n-alkanes with poor cold-flow properties that can be retained in diesel fuel and the amount of moderate-cetane-number
27.2 Gap Analysis for Syncrude to Diesel Fuel
2)
3)
4)
5)
branched alkanes with good cold-flow properties that are required. The isomerization requirement is also influenced by the amount of kerosene that is used for jet fuel production. By removing the lighter fraction of the distillate, the cold-flow properties of the heavier distillate become worse. Conversely, by removing the lighter fraction, the density and viscosity of the remaining heavier distillate increase, which is highly desirable in the case of straight-run LTFT-derived distillate. Aromatic compounds are not required in diesel fuel and are actually not desirable due to their generally low cetane numbers (Table 15.3). Yet, there are two reasons why at least some aromatics are preferable to an aromatic-free diesel fuel. Firstly, a lack of aromatic compounds contributes to problems with elastomer compatibility (Section 15.3.9). Secondly, aromatics have high densities and improve the overall density of syncrude-derived diesel fuel. Mildly hydroprocessed HTFT-derived diesel fuel contains sufficient aromatics to be interchangeable with crude-oil-derived EN590:2004-compliant diesel fuel. The same is not true of hydroprocessed LTFT-derived distillate, which contains almost no aromatics. As a compound class, the cycloalkanes have reasonable cetane numbers and adequate density to make them desirable diesel fuel components. The lack of cycloalkanes is one of the main reasons why it is difficult to refine Fischer–Tropsch syncrude to on-specification EN590:2004-type diesel fuel in high yield [8]. Alkenes are not prohibited in diesel fuel, but there is every reason to hydrogenate alkenes to alkanes before blending it into diesel fuel. When the properties of isostructural alkanes and alkenes are compared, the alkanes have better cetane numbers and better fuel stability. Oxygenates are included in diesel fuel as additives. Among others, the oxygenates provide boundary layer lubricity. It is therefore advantageous to retain some of the oxygenates in the diesel fuel by milder hydrotreating [11], in order to preserve the inherently good lubricity of syncrude. Depending on the oxygenate class, oxygenates may also improve the cetane number and increase the density. Alcohols have been used as diesel fuel extenders [12], and linear fuel ethers can be employed as cetane improvers [13]. Long-chain linear carboxylic acids are good lubricity providers [14], and esters are explicitly allowed in diesel fuel as FAME.
27.2.4 Density–Cetane–Yield Triangle
The lack of cycloalkanes in Fischer–Tropsch-derived distillate gives rise to a phenomenon called the density–cetane–yield triangle [8]. In essence it states that it is difficult to refine Fischer–Tropsch syncrude to diesel fuel in high yield to an EN590:2004-compliant diesel fuel that has both a minimum cetane number (>51) specification and a minimum density (820–845 kg·m−3 ) specification. Any two of these criteria can be met without too much refining effort. For example, it is possible to produce a distillate in high yield with adequate cetane number, or with adequate cetane number and density, but not meeting all three criteria simultaneously. The refining of HTFT syncrude to on-specification diesel fuel is a case in point. With proper hydroprocessing, HTFT syncrude can be converted into a diesel fuel of adequate cetane number and adequate density, but the overall diesel fuel yield is low in comparison to the other fuel types. The point is also illustrated by the hydrocracking of LTFT wax. LTFT syncrude can be converted in high yield to a high-cetane-number distillate, but the distillate has a low density (∼780 kg·m−3 ).
563
27 Diesel Fuel Refining
120
n -Alkanes 1-Alkenes n -Alkylbenzenes Branched alkanes Cycloalkanes
100 Cetane number
564
80 60 40 20 0 720
Figure 27.3
740
760
780 800 820 Density at 20 °C (kg m−3)
840
860
880
Cetane–density relationship of various compounds in the distillate boiling range.
There is a trade-off between cetane number and density for different compound classes (Figure 27.3). The value of cycloalkanes as constituent of diesel fuel is clear, since these are the only compounds that readily meet cetane number and density specifications. The cycloalkanes also have good cold-flow properties. An important synergistic effect is that both cetane number and density generally increase with carbon number (boiling point) within each compound class. Coproducing jet fuel in a Fischer–Tropsch refinery is therefore helpful in meeting both cetane number and density specifications of diesel fuel, albeit at the expense of yield. The loss of yield is anticipated by the density–cetane–yield triangle. Overcoming the density–cetane–yield triangle is central to the production of on-specification diesel fuel in a Fischer–Tropsch refinery. In countries where the diesel fuel does not have a minimum density requirement, or has a lower minimum density requirement, it is much easier to produce on-specification diesel fuel because only two of the criteria have to be met. Unfortunately, it has to be conceded that on a molecular level Fischer–Tropsch syncrude is unsuitable for the production of EN590:2004-type diesel fuel in high yield. Ways in which this deficiency can be overcome are explored in Section 27.3.4.
27.3 Decisions Affecting Diesel Fuel Refining
There are design decisions that affect all transportation fuels. Many of the important points discussed with reference to motor-gasoline production (Section 25.3) are also applicable to diesel fuel production. The same is true of the discussion on jet fuel production (Section 26.3). In order to avoid duplication, reference is made to all the topics that are relevant, but only the aspects that are specific to diesel fuel production are expanded upon: 1) Chemicals coproduction (Section 25.3.1).
27.3 Decisions Affecting Diesel Fuel Refining
2) 3) 4) 5)
Fate of the C2 –C4 hydrocarbons (Sections 25.3.2 and 26.3.1). Fate of the residue and wax (Sections 25.3.3 and 26.3.2). Fate of the aqueous product (Section 25.3.4). Co-refining (Section 25.3.7).
27.3.1 Fate of C2 –C4 Hydrocarbons
Oligomerization is an important technology to increase the yield of distillate from the C10 and lighter syncrude. This includes the C2 –C4 hydrocarbons. However, unlike jet fuel, a high degree of branching is not desirable in diesel fuel. Oligomerization technologies that do not employ pore-constraining catalysts are likely to produce highly branched oligomers from the light alkenes. Highly branched distillates will have excellent cold-flow properties, but correspondingly low cetane numbers. Oligomerization with H-ZSM-5-based technology is recommended to convert C3 –C4 alkenes into distillate. The MFI-type zeolite is pore constrained and yields a distillate with adequate cetane number despite the light alkene feed material employed (Table 19.4). Ethene is more difficult to oligomerize over solid acid catalysts, but can in principle produce a linear product with high cetane number, not unlike the heavier alkenes present in the straight-run syncrude. Unfortunately, there is no convenient technology for ethene oligomerization to distillate specifically; such technologies are all aimed at n-1-alkene petrochemical production [15]. It makes little sense then to hydrogenate the material to distillate if it is one of the few products that attracted Fischer–Tropsch-specific technology development (Section 23.2). 27.3.2 Fate of the Residue and Wax
Choosing between the different cracking technologies for converting LTFT wax to distillate is somewhat academic. Fluid catalytic cracking does not yield a significant distillate (light cycle oil) fraction from wax (Table 21.6), whereas hydrocracking does (Figure 21.8). Thermal cracking has been compared to hydrocracking for distillate production, and it was found to be less efficient although the product is more linear [16]. Hydrocracking is therefore the obvious choice for producing distillate from LTFT wax. The same arguments apply to HTFT syncrude. The HTFT residue can be directly converted into on-specification diesel fuel by hydrocracking [10]. The residue is a small fraction of the total syncrude and, although it can easily be refined to diesel fuel, it does not contribute much to the overall refining of HTFT syncrude. 27.3.3 Fate of the Aqueous Product
The C3 and heavier oxygenates recovered from the Fischer–Tropsch aqueous product are employed as diesel fuel extenders in conjunction with Fischer–Tropsch-derived distillate (Table 10.4) [12]. The carbonyls are first partially hydrogenated to produce alcohols, and the alcohol mixture is then blended with the diesel fuel. These alcohols are mainly in the C3 –C5 range. The blending
565
566
27 Diesel Fuel Refining
of ethanol and ethanol–methanol mixtures with diesel fuel has also been reported [17], although not specifically with Fischer–Tropsch distillates. An alternative pathway involves etherification. The C5 and heavier alcohols can be etherified to produce linear fuel ethers [13, 18], but C3 –C4 alcohols are considered too light for this purpose. Etherification is better suited for the refining of oil-phase oxygenates. 27.3.4 Technology Selection
The technology selection for diesel fuel refining is very dependent on whether the diesel fuel must conform to a minimum density specification or not. These two cases are considered separately. When there is no restriction placed on the minimum density of the diesel fuel, hydrocracking, oligomerization, and hydrotreating are the only three conversion units needed to refine syncrude to diesel fuel. The conversion to distillate can be performed as shown in Figure 27.2. In more practical terms, the flow diagram should include a hydrotreater to improve the storage stability and increase the cetane number of the straight-run distillate range products, as well as the distillate obtained from alkene oligomerization. The technology selection is consequently straightforward (Figure 27.4). In countries where the cold-flow requirements for diesel fuel are arduous, it may be necessary to perform mild hydroisomerization on the straight-run distillate. When there is a restriction on the minimum allowable density of the diesel fuel, the technology selection must overcome the limitations described as the Fischer–Tropsch density–cetane–yield triangle (Section 27.2.4). A too low density of diesel fuel is not a problem encountered in crude oil refining; the opposite is often the case. Consequently, there are no refining technologies that have been developed specifically to overcome this problem. Blending and co-refining (Section 27.3.5) with materials such as crude oil and coal-derived liquids are obvious solutions to this problem. These materials are rich in cyclic hydrocarbons that can be refined to cycloalkane-rich distillates with adequate density and cetane number. The cetane number and density will also increase with increasing jet fuel production, albeit with an associated loss in diesel fuel yield. C3 C4
Oligomerization
C5 − C10 C11− C22 >C22
Alkene hydrogenation
Alkanes
Hydrotreating
Alkanes
Mild hydroisomerization
Alkanes
Hydrocracking
Figure 27.4 Technology selection for the refining of syncrude to diesel fuel when there is no minimum diesel density requirement.
Alkanes
27.3 Decisions Affecting Diesel Fuel Refining
When it is important for the refinery design to be stand-alone and independent of external blending materials, technologies and refining strategies must be selected to deal with the density–cetane–yield triangle. The following technology-based strategies were proposed to increase distillate density without undermining the cetane number of the distillate [8]: 1) Cycloalkane synthesis. The synthesis of cycloalkanes by dehydrocyclization of alkanes is one of the steps in catalytic reforming (Chapter 22). The driving force for further catalytic dehydrogenation is thermodynamic, and cycloalkanes can in principle be recovered as an intermediate product. Cycloalkanes may be recovered as the main product from conversion in the operating range 250–400 ◦ C; aromatics become the dominant product at higher temperatures [19]. Higher temperature operation is possible, but only when the conversion per pass is limited to allow the recovery of cycloalkanes as intermediate products. 2) Linear alkyl benzene(LAB) synthesis. The technology for LAB synthesis by aromatic alkylation (Chapter 20) is well established [20]. Since LABs are large-volume commodity petrochemicals, it is doubtful whether it would make sense to synthesize LAB and then use it as a diesel fuel additive. 3) Alcohol refining. It has been noted that alcohols can be used as diesel fuel extenders (Section 27.3.3). Heavy alcohols can help increase the density, but the alcohols obtained from the aqueous product are too light to have a beneficial effect. As the chain length of the alcohol increases, cold flow becomes an issue. This problem can be overcome by turning the alcohols into linear fuel ethers, but the distillate range fuel ethers typically have densities of less than 810 kg·m−3 (Table 17.1). 4) Autoxidation of distillate. When distillate range material is autoxidized (Section 23.3), the incorporation of an oxygenate functionality into the hydrocarbons results in an increase in the density [21]. There is an accompanying increase in cetane number and lubricity too, but a decrease in storage stability. 5) Oligomerization technology selection. Oligomerization over an amorphous silica–alumina catalyst results in a distillate with much higher density (Table 19.5) on account of some cycloalkane formation. This comes at the expense of a lower cetane number, invoking the spectre of the density–cetane–yield triangle. 27.3.5 Co-refining
Blending and co-refining of non-Fischer–Tropsch-derived material is by far the easiest way to overcome the limitation described by the density–cetane–yield triangle. It also addresses some of the other concerns related to hydroprocessed LTFT distillate specifically, namely, elastomer incompatibility due to the lack of aromatics and polar compounds [22], emissions produced on account of the too high cetane number that results in combustion before complete air–fuel mixing [23], and the poorer volumetric fuel economy due to the low fuel density. Crude-oil-derived diesel fuel is the most readily available blending stock. In the past, such a blend would have been beneficial for the crude-oil-derived diesel fuel, because the sulfur-free Fischer–Tropsch distillate would reduce the sulfur content significantly. At present many crude-oil-derived diesel fuels contain less than 10 µg·g−1 sulfur and there is little benefit in this regard when mixed with Fischer–Tropsch distillate. Fischer–Tropsch distillate may still be
567
568
27 Diesel Fuel Refining
Natural gas
Gas reformer
H2 Gasifier
Oxygenates
Fischer− Tropsch synthesis
Syngas purification and conditioning
Coal biomass waste
Purge
Tail gas
Wastewater
Syncrude Pyrolysis oil Fuel gas
Naphtha
LPG Crude oil
Distillate
REFINERY
ADU Vacuum gas oil
Gasoline Diesel fuel Lubricant base oil
VDU Vacuum residue
Figure 27.5
Chemicals
Integrated crude oil and Fischer–Tropsch refinery.
beneficial in mixtures with heavy crude-oil-derived diesel fuels in order to reduce the density. By doing so, the lack of adequate density in the Fischer–Tropsch distillate is synergistically exploited. Other benefits of blends between crude-oil-derived diesel fuel and LTFT distillate have also been reported [24]. The advantage of co-refining LTFT syncrude with crude oil has been outlined by Gregor [25]. By coprocessing the vacuum gas oil and Fischer–Tropsch wax in a hydrocracker, a better quality distillate can be obtained from the combined feed. Other advantages that are not directly related to diesel fuel have also been noted. A Fischer–Tropsch facility that is integrated with a crude oil refinery has many potential synergies (Figure 27.5). Coal liquids offer similar advantages to crude-oil-derived diesel fuel for blending with Fischer–Tropsch distillates. Blends of coal-liquid-derived distillates and LTFT distillates were shown to have improved fuel economy and material compatibility [26]. Historically, the blending of coal liquids and LTFT distillates was common practice for the production of diesel fuel [1, 2]. Two industrial Fischer–Tropsch facilities were designed to co-refine coal liquids and syncrude (Chapters 8 and 9), and blends of HTFT distillate and coal distillate are marketed commercially as diesel fuel in South Africa [27–29]. The addition of biofuel in the form of FAME has some benefit, although it would be necessary to address the storage stability issues reported in the literature [7]. The FAME that are derived from different oils all have high densities and adequate cetane numbers (Table 27.2) [30]. By blending FAME with Fischer–Tropsch distillate, the density can be increased without undermining the cetane number of the blend.
27.3 Decisions Affecting Diesel Fuel Refining Selected properties of fatty acid methyl esters (FAME) prepared from different oils that are important for diesel fuel.
Table 27.2
FAME source
Sunflower oil Corn oil Used frying oil Olive oil
Density at 20 ◦ C (kg·m –3 ) 885.3 885.8 882.9 880.1
Cetane number
57.5 65 59 61
Viscosity at 40 ◦ C (cSt) 4.4 4.5 4.5 4.7
Cold filter plugging point (◦ C) –2 –7 –4 –6
The co-refining of HTFT and LTFT syncrudes is not a new concept, and it was practised on an industrial scale for many decades (Chapter 8). The combined refining of HTFT and LTFT syncrude has specific advantages for diesel fuel production and was considered for the proposed Mafutha coal-to-liquids project [31]. The HTFT syncrude fulfills the same purpose as crude oil as a source of an aromatic heavy fraction. However, due to the limited volume of distillate and heavier material in HTFT syncrude, the yield benefit is less than the yield benefit that can be obtained by co-refining a heavier feed material, such as crude oil. 27.3.6 Dealing with the Density–Cetane–Yield Triangle
There are no conventional refining pathways to produce diesel fuel in high yield from Fischer–Tropsch syncrude while meeting both the density and cetane number specifications. In order to overcome the limitations of the density–cetane–yield triangle, some less conventional refining approaches, product blending, and co-refining were explored (Sections 27.3.4 and 27.3.5). Of these options, blending is by far the easiest solution. The question still remains, what yield of on-specification diesel fuel can be obtained by conventional refining of syncrude? The refining challenge can be outlined as follows [32]: Most conventional refining technologies to produce distillate from syncrude produce either distillate high in density and low in cetane number, like aromatics, or distillate low in density and high in cetane number, like alkanes. It is this trade-off that gives rise to the density–cetane–yield triangle. Refinery design to enable any further increase in the diesel fuel yield relies heavily on the lever rule (Figure 27.6). Two streams must be blended in such a way that the targeted minimum density (Equation 27.1) and minimum cetane number (Equation 27.2) can be met. ρ = y·ρ1 + (1 − y)·ρ2 = 820 kg·m−3
(27.1)
CN = z·CN1 + (1 − z)·CN2 = 51
(27.2)
The volume fraction y indicates the maximum amount of low-density, high-cetane-number material that can be blended with a high-density, low-cetane-number material. Conversely, the volume fraction z indicates the minimum amount of the low-density, high-cetane-number material that must be blended with the high-density, low-cetane-number material. If y ≥ z, then the blend will meet specifications, but if y < z, then the blend will not be able to meet specifications.
569
570
27 Diesel Fuel Refining
∆1
(1 − y )
y
Low
r1
∆2 High
∆1
∆2
r
y=
r2
∆2
(1 − z)
z
Low
CN2
r2 − r r2 − r1
∆1 High
∆2
z=
CN
∆1
CN1
CN − CN2 CN1 − CN2
Figure 27.6 Application of the lever rule to the refining of diesel fuel from syncrude by balanced blending of a fuel that has a low density (ρ1 ) and high cetane number (CN1 ) with a fuel that has a high density (ρ2 ) and low cetane number (CN2 ).
When a conversion process produces a distillate that is either density- or cetane number constraining, a complementary conversion process must be selected that produces a distillate with the opposite constraint. The lever rule can then be applied to see whether the materials can be blended in a ratio that allows the diesel fuel specifications to be met. If two complementary processes can be found, the yield of diesel fuel can be increased. As mentioned before, this refining challenge becomes a mute point when the diesel fuel specification does not require a minimum density.
27.4 Diesel Fuel Refining from HTFT Syncrude
The distillate and residue fractions from HTFT syncrude can be hydroprocessed to produce an on-specification diesel fuel that meets both the minimum density and cetane number requirements (Table 27.1). The volume of diesel fuel that can be produced is equivalent to about 10% of the syncrude. This hydroprocessed diesel fuel forms the base blending material for further diesel fuel production. Any further increase in diesel fuel yield with conventional refining technology must employ the strategy outlined in Section 27.3.6. As mentioned before (Sections 25.4 and 26.4), there is a high carbon cost associated with the HTFT light hydrocarbons and the HTFT aqueous product. HTFT refinery design is therefore based on a closed gas loop with cryogenic C2 hydrocarbon recovery and methane recycle. 27.4.1 HTFT Diesel Fuel Design Case I
By employing only conventional refining technologies and FAME addition, it was shown that a refinery can be designed where about 25% of the liquid transportation fuel was on-specification EN590:2004 diesel fuel [32]. The hydroprocessed HTFT distillate and residue was employed as base blending material for the diesel fuel, and the yield was increased by blending hydrotreated
27.4 Diesel Fuel Refining from HTFT Syncrude Carbonyl hydrogenation
Aqueous product
Wastewater Fuel gas Gasoline Jet fuel Diesel
Aromatic alkylation
C2
LPG Gasoline
H-ZSM-5 oligomerization
C3 −C4
Alcohols
HTFT gas loop
Condensate C5 −C6 C5 Benzene
C6 −C8
Light oil
Alkene hydrogenation
Jet fuel Diesel
C5 Hydroisomerization
Gasoline
Nonacidic Pt/L reforming
H2 LPG Gasoline
>C6
Jet fuel
Hydrotreater
Decanted oil Hydrocracking >C11
Jet fuel Diesel
Figure 27.7 HTFT diesel fuel refinery design case I. The gas loop includes cryogenic C2 separation with methane recycle.
H-ZSM-5 oligomers and alkyl aromatics. The coproduction of jet fuel assisted the increase in diesel fuel yield, because it removed the lower density and lower cetane number fraction from the alkane-based distillate. An analogous design is presented (Figure 27.7). Diesel fuel production takes place by mild hydrocracking of the straight-run distillate and residue. Additional diesel fuel is produced by blending the distillate fraction from H-ZSM-5-based alkene oligomerization with alkyl aromatics. The H-ZSM-5-derived oligomer stream is the high-cetane, low-density material, and the alkyl aromatics stream is the low-cetane, high-density material. The alkyl aromatics employed were di- and triethylbenzenes obtained from benzene alkylation with ethene. These compounds have cetane numbers around 10 and densities around 870 kg·m−3 . Aromatic alkylation with ethene allows the ethene to be converted into a liquid product, and using ethene as alkylating agent is not necessarily the best choice from a diesel fuel perspective. Other technologies are necessary to meet the fuel specifications for motor-gasoline and jet fuel. The Fischer–Tropsch aqueous product is refined to alcohols by partial hydrogenation of the carbonyl compounds. The C3 and heavier alcohols can in principle be employed as diesel fuel extenders, and the ethanol can be used to improve the octane number of the motor-gasoline.
571
572
27 Diesel Fuel Refining
The amount of diesel fuel that can be produced depends mainly on the design and operation of the hydrocracker the fraction of light distillate that is employed for jet fuel production, and the target cetane number before additive addition. The impacts of these variables are shown (Table 27.3) as different operating scenarios of the Figure 27.7 refinery design: Table 27.3
Products from HTFT diesel fuel refinery design case I shown in Figure 27.7.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsc Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t –1 ) Fuel gas (kg·t –1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m –3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m –3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m –3 )
Base case
0.0 7.3 7.3 36.9 21.1 15.2 6.5 0.0 5.1 0.5
Cetane number 47
0.2 7.0 7.2 38.2 12.5 22.7 6.5 0.0 5.1 0.5
FAME addition
0.0 7.3 7.3 36.9 19.0 18.2 6.5 −0.9b 5.1 0.5
0.133 0.493 0.270 0.186
0.132 0.510 0.160 0.277
0.133 0.493 0.245 0.222
1.082 65 73
1.079 65 72
1.093 56 73
95.1 87.7 749 60 17.1 33.3 0.3 0.0
95.2 87.8 750 59 16.5 32.5 0.3 0.0
95.1 87.7 749 60 17.1 33.3 0.3 0.0
778 22.3
779 24.2
776 22.1
51 820
47 820
51 820
a On a syncrude basis, excluding H , CO, CO , H O, and CH from HTFT synthesis (closed gas loop with cryogenic 2 2 2 4 separation). b Fatty acid methyl ester imported for blending. c Syncrude waste products (e.g., metal-containing fuel oil) and unrecoverable material (e.g., carboxylic acids in wastewater).
27.5 Diesel Fuel Refining from LTFT Syncrude
1) Base case design. The C3 and heavier alcohols from aqueous product refining were considered as petrochemicals and were not added to the diesel fuel. The hydrocracker design is based on a NiMo/SiO2 –Al2 O3 catalyst (Table 21.8) [10]. 2) Lower cetane number. Cetane number improvers (Section 15.4) can be employed to increase the cetane number of the final diesel fuel. Addition of 0.05 vol% of an appropriate cetane improver will increase the cetane number of the diesel fuel by around 3–4 points. This does not affect the cetane index. There is a minimum cetane index requirement too. The refinery design was evaluated at a cetane number of 47, which corresponds to a cetane index above the minimum requirement. The assumption is that the cetane number deficiency of 4 points can be rectified by the addition of a cetane improver. Lowering the cetane number requirement allows the diesel fuel yield to be improved at the expense of jet fuel production. 3) FAME addition. The addition of 5 vol% of FAME allows a slight increase in diesel fuel yield by cutting more jet fuel into the diesel fuel, because the FAME (Table 27.2) increases both the density and cetane number of the diesel fuel. All of the scenarios highlighted the constrained nature of diesel fuel production, as well as the negative impact that it has on the blending flexibility of other fuel types. HTFT syncrude is not well suited for diesel fuel production beyond that obtained from mild hydrocracking of the distillate and residue fractions.
27.5 Diesel Fuel Refining from LTFT Syncrude
Unlike HTFT syncrude, there is no refining pathway that directly yields on-specification diesel fuel when there is a minimum density specification (Table 27.1). Yet, a higher yield of on-specification diesel fuel and distillate can be refined from LTFT syncrude [32]. Since not all countries have a minimum density limitation on diesel fuel, two design cases will be considered. In the first design case (Section 27.5.1), the diesel fuel yield will be maximized without a minimum density constraint. This design case differs from the distillate refining strategy (Figure 27.2) in that all transportation fuels must meet specification. The diesel fuel in this case is not a distillate – it is a diesel fuel – with the caveat that there is no minimum density specification. The design is therefore not subject to the density–cetane–yield triangle. In the second design case (Section 27.5.2), the refinery design will produce as much diesel fuel as possible with a minimum density constraint. 27.5.1 LTFT Diesel Fuel Design Case I
Producing a diesel fuel with acceptable cetane number and cold-flow properties from LTFT syncrude is not difficult when diesel fuel density does not have to be considered. The design approach shown in Figure 27.2 will ensure that diesel fuel is produced in high yield. The need to produce transportation fuels that are all on-specification necessitates the inclusion of at least one conversion unit that is capable of producing aromatics. This has the added advantage of making the refinery design less reliant on the Fischer–Tropsch gas loop for its
573
574
27 Diesel Fuel Refining Carbonyl hydrogenation
Aqueous
C3 – C4
Alcohols Wastewater LPG Gasoline
SPA alkylation oligomerization
C5
LTFT gas loop
Hydrotreater
Jet fuel
C5 hydroisomerization
Gasoline
Benzene
Hydrotreater
C6 – C8
Nonacidic Pt / L reforming
Hot and cold condensate
Hydrocracker
Wax
H2 Fuel gas Gasoline Jet fuel Diesel
Diesel
C23 and heavier
Figure 27.8 LTFT diesel fuel refinery design case I, which maximizes diesel fuel yield when there is no minimum density specification requirement. The gas loop includes C3 –C4 recovery with recycling of the C1 –C2 hydrocarbons in the tail gas.
hydrogen requirements. Excess aromatics can be added to the diesel fuel, since the paraffinic diesel fuel will have a high cetane number. The design of a diesel fuel refinery is quite straightforward, and it incorporates only hydrocracking, oligomerization, hydrotreating, aromatization (reforming), aromatic alkylation, and hydroisomerization units (Figure 27.8). The diesel fuel yield can be further increased by incorporating an oligomerization unit to convert the naphtha range alkenes into distillate (Figure 27.9). Oligomerization can be performed with an amorphous silica–alumina- or H-ZSM-5-based technology. In both variations of this design (Figures 27.8 and 27.9) the aqueous product is recovered and partially hydrogenated to produce alcohols. The alcohols may be added to the fuel as a diesel fuel extender. It is clear that, when there is no minimum density specification, a high diesel fuel yield can be obtained with both Fe-LTFT and Co-LTFT syncrudes (Table 27.4). The jet fuel has a much lower cetane number than the diesel fuel, but considering the high cetane number of the distillate, it is possible to include additional jet fuel into the diesel fuel. The constraining specification is the diesel fuel viscosity, which is lowered when additional kerosene range material is included in the diesel fuel.
27.5 Diesel Fuel Refining from LTFT Syncrude Products from LTFT diesel fuel refinery design case I, which produces diesel fuel without a minimum density specification requirement. The base case design is shown in Figure 27.8 and the design with an added naphtha oligomerization unit to increase diesel fuel yield is shown in Figure 27.9.
Table 27.4
Description
Figure 27.8 design Fe-LTFT
Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsb Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t –1 ) Fuel gas (kg·t –1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m –3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m –3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m –3 )
−0.1 0.4 7.7 14.4 14.4 58.3 3.2 0.0 1.4 0.3
Co-LTFT
0.0 0.4 10.0 13.8 13.7 60.0 1.3 0.0 0.6 0.3
Figure 27.9 design Fe-LTFT
−0.2 0.1 7.3 10.7 6.3 71.0 3.2 0.0 1.4 0.2
Co-LTFT
−0.1 0.1 9.6 12.1 4.1 72.1 1.3 0.0 0.6 0.2
0.139 0.192 0.185 0.750
0.181 0.185 0.177 0.771
0.132 0.145 0.081 0.913
0.174 0.162 0.052 0.926
1.266 32 3
1.313 13 4
1.271 32 –1
1.315 13 0
95.4 87.3 752 42 15.0 30.4 0.4 0
95.3 87.8 746 60 7.1 31.1 0.4 0
98.4 90.5 742 59 16.1 32.9 0.3 0
98.8 90.5 746 59 17.7 33.8 0.3 0
776 17.7
777 17.7
775 19.0
779 17.1
72 778
72 778
69 778
69 778
a On a syncrude basis, excluding H , CO, CO , H O, and C –C hydrocarbons from LTFT synthesis (closed gas loop 2 2 2 1 2 with C3 –C4 recovery from the tail gas). b Syncrude waste products and unrecoverable material (e.g., carboxylic acids in wastewater).
575
576
27 Diesel Fuel Refining Carbonyl hydrogenation
Aqueous
C3 –C4
Wastewater LPG Gasoline
SPA alkylation oligomerization
C5 LTFT gas loop
Hot and cold condensate
Alcohols
Hydrotreater
Jet fuel
C5 hydroisomerization
Gasoline
C6 –C10
Oligomerization
Hydrotreater
Benzene
C6 –C8
Nonacidic Pt / L reforming
Hydrocracker
Wax
H2 Fuel gas Gasoline Jet fuel Diesel
Diesel
C23 and heavier
Figure 27.9 LTFT diesel fuel refinery design case I with an additional naphtha oligomerization unit to boost diesel fuel yield, when there is no minimum density specification requirement. The gas loop includes C3 –C4 recovery with recycling of the C1 –C2 hydrocarbons in the tail gas.
27.5.2 LTFT Diesel Fuel Design Case II
When the diesel fuel must meet minimum density and cetane number requirements, the refinery design is subject to the limitations described by the density–cetane–yield triangle. A refinery design strategy can be followed that employs the lever rule to balance high-cetane-number, low-density material with low-cetane-number, high-density material (Section 27.3.6). The high cetane number of the gas oil fraction from LTFT wax hydrocracking provides enough leverage in combination with alkyl aromatics to just meet the diesel fuel specifications (Figure 27.10). It illustrates the dilemma of refining paraffinic material to diesel fuel. Central to the design is the selection of the cracking unit. Refinery designs for the production of on-specification EN590:240 diesel fuel from LTFT syncrude that have been proposed in the literature [8, 32] employed fluid catalytic cracking for wax conversion. In each instance, diesel fuel was not the major product, and the selection of fluid catalytic cracking instead of hydrocracking was dictated by the refining requirements of the other transportation fuels. The high distillate
27.5 Diesel Fuel Refining from LTFT Syncrude
yield that can be obtained by wax hydrocracking is forfeited because it is not possible to produce on-specification diesel fuel in high yield. The operating window for on-specification diesel fuel production is small. The diesel fuel just meets specification (Table 27.5), and the base case design includes 5% FAME addition. The diesel
Table 27.5
Products from LTFT diesel fuel refinery design case II shown in Figure 27.10.
Description Product distribution (mass%)a H2 Fuel gas LPG Motor-gasoline Jet fuel Diesel fuel Petrochemicals produced Petrochemicals consumed Unrecovered organicsc Water to/from refining Liquid fuels (m3 ·t –1 )a LPG Motor-gasoline Jet fuel Diesel fuel Net productiona Liquid fuels (m3 ·t –1 ) Petrochemicals (kg·t –1 ) Fuel gas (kg·t –1 ) Motor-gasoline properties RON MON Density at 15 ◦ C (kg·m –3 ) Reid vapor pressure (kPa) Alkene content (vol%) Aromatic content (vol%) Benzene content (vol%) Oxygenate content (vol%) Jet fuel properties Density at 15 ◦ C (kg·m –3 ) Aromatic content (vol%) Diesel fuel properties Cetane number Density at 15 ◦ C (kg·m –3 ) a On
Fe-LTFT
Co-LTFT
0.8 1.2 6.6 33.2 31.2 22.9 3.2 −1.1b 1.7 0.3
1.0 1.2 8.9 35.6 28.4 23.6 1.3 −1.2b 0.9 0.3
0.120 0.444 0.401 0.279
0.161 0.478 0.363 0.288
1.244 20 20
1.290 1 22
98.1 90.6 747 60 10.4 34.0 0.5 0.0
98.0 90.3 744 60 12.7 32.5 0.5 0.0
779 22.9
782 24.7
51 820
51 820
a syncrude basis, excluding H2 , CO, CO2 , H2 O, and C1 –C2 hydrocarbons from LTFT synthesis (closed gas loop with C3 –C4 recovery from the tail gas). b Fatty acid methyl ester imported for blending. c Syncrude waste products (e.g., coke on FCC catalyst) and unrecoverable material (e.g., carboxylic acids in wastewater).
577
578
27 Diesel Fuel Refining Carbonyl hydrogenation
Aqueous
Wastewater LPG Gasoline
SPA alkylation oligomerization
C3 – C4
LTFT gas loop
Alcohols
C5
C3 – C4
C5
Hydrotreater
Jet fuel Diesel
C5 hydroisomerization
Gasoline
Benzene Hot and cold condensate
Nonacidic Pt / L reforming Hydrotreater
H2 Fuel gas Gasoline Jet fuel Diesel
Wax FCC
Fuel gas
Wax
Figure 27.10 LTFT jet fuel refinery design case II, which maximizes diesel fuel yield when there is a minimum density specification requirement. The gas loop includes C3 –C4 recovery with recycling of the C1 –C2 hydrocarbons in the tail gas.
fuel is very aromatic and, although the aromatics are almost exclusively mononuclear aromatics, the design in Figure 27.10 is only of theoretical interest. Although the design is capable of producing some on-specification diesel fuel, it is clearly not an elegant solution, and in practice blending or co-refining is preferable.
References 1. Weil, B.H. and Lane, J.C. (1949) The Technol-
5. (2009) Directive 2009/30/EC of the European
ogy of the Fischer-Tropsch Process, Constable, London. 2. Freerks, R. (2003) Early efforts to upgrade Fischer-Tropsch reaction products into fuels, lubricants and useful materials. AIChE Spring National Meeting, 2 April, 2003, New Orleans, paper 86d. 3. Hardenberg, H.O. (1980) Thoughts on an ideal diesel fuel from coal. S. Afr. Mech. Eng., 30, 34–47. 4. Boehman, A.L., Szybist, J.P., Song, J., Zello, V., Alam, M., and Miller, K. (2004) Combustion characterization of GTL diesel fuel. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 49 (2), 714–716.
Parliament and of the Council of 23∼April 2009 amending Directive 98/70/EC as regards the specification of petrol, diesel and gas-oil and introducing a mechanism to monitor and reduce greenhouse gas emissions and amending Council Directive 1999/32/EC as regards the specification of fuel used by inland waterway vessels and repealing Directive 93/12/EEC. Off. J. Eur. Union, L140, 88. 6. Reid, R.C., Prausnitz, J.M., and Poling, B.E. (1987) The Properties of Gases and Liquids, 4th edn, McGraw-Hill, New York. 7. Mushrush, G.W., Willauer, H.D., Bauserman, J.W., and Williams, F.W. (2009) Incompatibility of Fischer-Tropsch diesel with petroleum and
References
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
soybean biodiesel blends. Ind. Eng. Chem. Res., 48, 7364–7367. De Klerk, A. (2009) Can Fischer-Tropsch syncrude be refined to on-specification diesel fuel? Energy Fuels, 23, 4593–4604. Santana, R.C., Do, P.T., Santikunaporn, M., Alvarez, W.E., Taylor, J.D., Sughrue, E.L., and Resasco, D.E. (2006) Evaluation of different reaction strategies for the improvement of cetane number in diesel fuels. Fuel, 85, 643–656. Leckel, D.O. (2009) Hydroprocessing Euro 4-type diesel from high-temperature Fischer-Tropsch vacuum gas oils. Energy Fuels, 23, 38–45. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Knottenbelt, C. (2002) Mossgas ‘‘gas-to-liquids’’ diesel fuels – an environmentally friendly option. Catal. Today, 71, 437–445. Pecci, G.C., Clerici, M.G., Giavazzi, F., Ancilotti, F., Marchionna, M., and Patrini, R. (1991) Oxygenated diesel fuels. Part 1 – Structure and properties correlation. 9th International Symposium on Alcohols Fuels, 1991, pp. 321–326. Knothe, G. and Steidley, K.R. (2005) Lubricity of components of biodiesel and petrodiesel. The origin of biodiesel lubricity. Energy Fuels, 19, 1192–1200. (2001) Linear Alpha Olefins, SRI Process Economics Program Report 12D, Stanford Research Institute, Menlo Park, CA. De Klerk, A. (2007) Thermal cracking of Fischer-Tropsch waxes. Ind. Eng. Chem. Res., 46, 5516–5521. Ribeiro, N.M., Pinto, A.C., Quintella, C.M., Da Rocha, G.O., Teixeira, L.S.G., Guariero, L.L.N., Do Carmo Rangel, M., Veloso, M.C.C., Rezende, M.J.C., Da Cruz, R.S., De Oliveira, A.M., Torres, E.A., and De Andrade, J.B. (2007) The role of additives for diesel and diesel blended (ethanol or biodiesel) fuels: a review. Energy Fuels, 21, 2433–2445. Marchionna, M., Patrini, R., Giavazzi, F., and Pecci, G.C. (1996) Linear ethers as high quality components for reformulated diesel fuels. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 41 (3), 585–589. Steiner, H. (1956) in Catalysis Vol. IV, Hydrocarbon synthesis, hydrogenation and cyclization,
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31. 32.
(ed. P.H. Emmett), Reinhold, New York, pp. 529–560. Kocal, J.A., Vora, B.V., and Imai, T. (2001) Production of linear alkylbenzenes. Appl. Catal. A, 221, 295–301. Naegeli, D.W., Childress, K.H., Moulton, D.S., and Lacey, P.I. (2002) Method for producing oxygenates fuels. Patent US 6,488,727. Lamprecht, D. (2007) Elastomer compatibility of blends of biodiesel and Fischer-Tropsch diesel. SAE Tech. Pap. Ser., 2007−01–0029. Androulakis, I.P., Weisel, M.D., Hsu, C.S., Qian, K., Green, L.A., Farrell, J.T., and Nakakita, K. (2005) An integrated approach for creating model diesel fuels. Energy Fuels, 19, 111–119. Wu, T., Huang, Z., Zhang, W.-G., Fang, J.-H., and Yin, Q. (2007) Physical and chemical properties of GTL-diesel fuel blends and their effects on performance and emissions of a multicylinder DI compression ignition engine. Energy Fuels, 21, 1908–1914. Gregor, J.H. (1990) Fischer-Tropsch products as liquid fuels or chemicals. An economic evaluation. Catal. Lett., 7, 317–332. Lamprecht, D. (2009) Blending performance of hydroprocessed coal pyrolysis products in low-temperature Fischer-Tropsch diesel. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 120–121. Leckel, D.O. (2009) Hydroprocessing CTL diesel from high-temperature Fischer-Tropsch syncrude and pyrolysis tar oil. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem., 54 (1), 125–126. Kamara, B.I. and Coetzee, J. (2009) Overview of high-temperature Fischer-Tropsch gasoline and diesel quality. Energy Fuels, 23, 2242–2247. Leckel, D.O. (2009) Diesel production from Fischer-Tropsch: the past, the present, and new concepts. Energy Fuels, 23, 2342–2358. Anastopoulos, G., Lois, E., Serdari, A., Zanikos, F., Stournas, S., and Kalligeros, S. (2001) Lubrication properties of low-sulfur diesel fuels in the presence of specific types of fatty acid derivatives. Energy Fuels, 15, 106–112. (2008) Sasol proposes Project Mafutha CTL plant. Oil Gas J., 106 (4), 10–11. De Klerk, A. (2011) Fischer-Tropsch fuels refinery design. Energy Environ. Sci., 4, 1177–1205.
579
581
28 Chemicals and Lubricant Refining 28.1 Introduction
Chemicals and lubricants are the higher value, lower volume cousins of the transportation fuels. There is consequently an economic incentive to produce these higher value products when the refinery (and market) permits. Central to the refining of syncrude to chemicals and lubricants is whether these products can be produced on their own, or whether these products must be produced in conjunction with fuel refining. This is not clear cut, and a similar question has been raised in relation to crude oil [1]: ‘‘Whether or not the production of chemicals from crude oil can be independent of the production of fuels has still not been determined.’’ The conundrum can be traced back to the molecular makeup of the syncrude or crude oil. There is a subset of refining technologies that lend themselves to the production of petrochemicals. Likewise, some compounds are easily refined to chemicals and lubricants. However, the selectivity to chemicals and lubricants, even from well-designed petrochemical technologies and appropriate feed materials, is never 100%. Chemicals are usually single-compound products with stringent quality specifications. Nonchemical products cannot be blended away into chemicals, as is the case with products that are mixtures, such as fuels. Lubricants, being a mixture of compounds, are more forgiving products, but there is still a limit to which nonlubricant material can be blended away. So, what do you do with the material that is neither a chemical nor a lubricant? Unless one can recycle the material to be converted into chemicals or lubricants, fuel production becomes inevitable. The high value of chemicals and lubricants allows a higher rejection rate of material as refinery fuels, but there is a limit to the amount of refinery fuel required. Ultimately, some material may have to be sold as heating or transportation fuels. Furthermore, materials that were produced during chemical and lubricant refining may be difficult to convert into either of these products, thereby reducing the efficiency of the refinery, if one wanted to do so. This seems to be the case for crude oil [2], but is it true of syncrude as well? The gradual conversion of some of the industrial Fischer–Tropsch facilities to increase the production of chemicals (Chapters 8 and 9), as well as the intentional production of chemicals and lubricants in new facilities (Chapter 11), indicates that syncrude has considerable potential for chemical and lubricant production. The extent to which syncrude can be converted into chemicals and lubricants will be discussed. Concepts that were proposed in the literature for chemical and lubricant refining from syncrude are evaluated (Section 28.3). Then the extent to which syncrude can efficiently be refined to chemicals and lubricants is explored (Sections 28.4 and 28.5). Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
582
28 Chemicals and Lubricant Refining
28.2 Petrochemical and Lubricant Markets 28.2.1 Petrochemicals
The petrochemical market is much smaller than the energy market. It was estimated that the global energy consumption in 2007 was 465 EJ (441 quadrillion BTU), of which 298 EJ was derived from fossil fuels and 27 EJ was employed for chemical production [3]. The main commodities and how each is related to the raw materials employed for production are shown in Figure 28.1 [3]. Despite its comparatively small market share in the overall energy sector, there are many opportunities for the conversion of Fischer–Tropsch syncrude into petrochemicals. Petrochemicals are higher value products and are economically more enticing products than transportation fuels. Herein lies a risk too. A common fallacy that is sometimes perpetrated in evaluating chemical opportunities is that the market will not respond to the additional production capacity. This is not true, since the global installed capacity often exceeds the global consumption. The logistics cost associated with chemical production will favor producers based on their proximity to the consumers. Gaining market share by selling product at a lower price can lead to a price war, which quickly erodes profitability. The prospective petrochemical refinery designer should be aware of ‘‘shut-down economics.’’ This requires careful competitor analysis to determine for how little a chemical can be sold from an existing but competing production facility, without running the facility at a loss. The risk of a hostile response to entry into a specific petrochemical market increases with decreasing market size. Raw material
Feed fraction
Commodities
Methane
1.1 EJ
Products
Methanol, etc.
Natural gas Natural gas liquids
4.5 EJ
Ethane, propane, butane condensates
Ethene, propene, butadiene, etc.
Refinery off-gas
Crude oil Refinery liquids
21.2 EJ
Naphtha gas oil
Coal
Coal pyrolysis liquids
0.08 EJ
Benzene, toluene, xylenes, etc.
Figure 28.1 Petrochemical production showing the origin and energy flow in 2007 and indicating the seven highest volume primary commodity chemicals.
Plastics, solvents, surfactants, elastomers, coatings, fibres, etc.
28.2 Petrochemical and Lubricant Markets
A Fischer–Tropsch-based petrochemical facility carries a large capital cost burden, making it more vulnerable to a hostile response. For a new Fischer–Tropsch facility, the actual petrochemical production cost must include the operating cost and the cost of capital. An established Fischer–Tropsch facility runs a lower risk when it enters a niche market for petrochemicals, but when designing a new Fischer–Tropsch-based petrochemical facility it is better to focus on large-volume commodity chemicals where there is less risk of a hostile market. The focus in the remainder of the chapter is therefore on large-volume commodity chemicals, while acknowledging that there are many opportunities for smaller volume petrochemicals. The largest volume primary commodity chemicals are listed in Table 28.1 [3]. The market is growing and, by producing any of these petrochemicals from Fischer–Tropsch syncrude, it is unlikely to upset the market or price. Of these, ethene and propene are primary Fischer–Tropsch products and constitute 15–20% of a typical HTFT (high-temperature Fischer–Tropsch) syncrude [4], but much less of LTFT (low-temperature Fischer–Tropsch) syncrude. The primary commodity chemicals are used in the manufacture of various derived petrochemicals, although the derived petrochemicals are not exclusively obtained from the primary commodity chemicals. These primary commodity chemicals are also called first-generation intermediates [5]. The derived petrochemicals in general have a higher value than the primary commodities. Since some of these derived products also have large markets, further value addition within a Fischer–Tropsch petrochemical refinery can be considered without increasing the risk. This may be especially desirable when the derived product has a large petrochemical market and good transportation fuel properties. Such chemicals provide an easy reincorporation pathway into the refinery by direct fuel blending, making the refinery design more robust. The market sizes of some derived petrochemicals that may be of interest in the design of a Fischer–Tropsch petrochemical refinery are given in Table 28.2 [6]. These are of course not the Global consumption of the largest volume primary commodity chemicals for the period 2005–2007.
Table 28.1
Primary commodity
Global consumption (million tons per year) 2005
Alkenes Ethane Propene Butadiene Aromatics Benzene Toluene Xylenes Oxygenatesa Methanol
2006
2007
105.6 66.7 9.5
110.1 70.7 9.7
114.6 73.5 10.1
37.5 18.8 35.2
38.5 19.8 37.6
40.6 20.9 40.7
36.7
38.9
40.6
a Ethanol is used as chemical and fuel and has therefore not been included.
583
584
28 Chemicals and Lubricant Refining Table 28.2 Products and markets relevant to petrochemical production from Fischer–Tropsch syncrude.
Petrochemical product
Cumene Detergent alcohols Ethanoic acid (acetic acid) Ethanola Linear alkanes, C9 –C17 (paraffins) Linear 1-alkenes (linear α-olefins) Linear alkylbenzene sulfonates Phenol Plasticizer alcohols 1-Butanol 2-Ethylhexanol Isononanol Propanoic acid Waxes a Volume-based
Global consumption Million tons per year
Year
11.9 2.0 10.6 57.3 2.9 2.7 4.1 8.6 8.0 2.8 2.8 1.0 0.3 3.0
2007 2008 2009 2008 2007 2006 2008 2007 2008 2008 2008 2008 2009 2005
market for chemicals and fuels: 72 681 million liters per year in 2008.
only chemicals that can be considered, but it gives an indication of the opportunities that exist for the refining Fischer–Tropsch syncrude to petrochemicals. 28.2.2 Lubricants
The lubricant base oil market size is on the order of 40 million m3 /year [7]. Compared to the petrochemicals, it is a much smaller market. It is nevertheless of interest to the Fischer–Tropsch refiner, because high-quality lubricant base oils can be prepared from LTFT waxes [8] as well as from n-1-alkenes in the heavy naphtha and kerosene fractions of HTFT and LTFT syncrude. The American Petroleum Institute (API) classifies lubricant base oils into five groups (Table 28.3). The viscosity index (VI), which is calculated in accordance with the ASTM D 2270 method [9], is a key quality parameter. The VI is a measure of the temperature sensitivity of the kinematic viscosity of the oil. A high-VI oil retains more of its viscosity as the temperature is increased than a low-VI oil. VI is not the only performance parameter. Among others, good cold-flow properties, kinematic viscosity, volatility, lubricity, and oxidation stability are also important quality parameters. Good biodegradability and low toxicity are also desirable properties. Lubricant base oils can be produced from Fischer–Tropsch syncrude in API groups III–V by using different feed fractions and conversions: 1) API group III lubricant base oils. These oils can be prepared from LTFT waxes [8] and HTFT residues [10]. In the case of LTFT-derived base oils, the branched alkane content can be
28.3 Overview of Chemicals Refining Concepts for Syncrude Table 28.3
American Petroleum Institute (API) classification of lubricant base oils.
API classification
Sulfur content (mass%)
Group I Group II Group III Group IV Group V
>0.03a ≤0.03 ≤0.03
Saturates (mass%)
<90a ≥90 ≥90 All polyalphaolefin (PAO) oilsb All oils not included in Groups I–IV
Viscosity index 80–119 80–119 ≥120
a Group
I, if either sulfur or saturates content criteria are met. Includes polyinternalolefin (PIO) oils. Although PIO oils are not recognized as a separate group by API, it has been suggested that PAO and PIO oils should not be grouped together. b
manipulated through the selection and operation of hydrocracking and hydroisomerization, or by processes such as solvent dewaxing to reduce the n-alkane content. 2) API group IV lubricant base oils. The n-alkenes in Fischer–Tropsch syncrude can be oligomerized with an appropriate technology to produce either polyalphaolefin (PAO) or polyinternalolefin (PIO) lubricant base oils. Depending on the oligomerization technology, it may be possible to directly convert the alkenes within the syncrude matrix, making subsequent separation easier. The quality of such oils is likely to be lower than that of the oils prepared by purification (Section 23.2) followed by oligomerization. 3) API group V lubricant base oils. Fischer–Tropsch syncrude contains oxygenates and these may be converted into different types of lubricant base oils [11]. The most pertinent types are esters and neutral trialkyl phosphate esters.
28.3 Overview of Chemicals Refining Concepts for Syncrude 28.3.1 Alkane-Based Refining
LTFT syncrude consists mainly of heavier alkanes (Table 1.2), with less alkenes and oxygenates. A chemical refinery concept was developed on the basis of the alkane-rich nature of LTFT syncrude that exploited this property of the syncrude to produce exclusively alkane-based chemicals [12]. Product upgrading is by hydroprocessing only, with no aqueous product recovery. Typical products from the alkane-based LTFT chemical refinery are the following: 1) Waxes, which are used in application such as candles, coatings, inks, and packaging. 2) Synthetic lubricants, which include automotive lubricant applications, gear oils, hydraulic fluids, and greases. 3) Mixed paraffins, which are apolar, odor-free solvents for paints, cleaner products, insecticides, and drilling fluids.
585
586
28 Chemicals and Lubricant Refining
4) n-Alkanes, which are feed materials for further chemicals production, such as detergent alcohols, linear alkyl benzenes, and plasticizers. 5) Liquid petroleum gas (LPG), which is a fuel. On the basis of the straight-run content of syncrude in the C5 and heavier range, such an alkane-based, hydroprocessing-only LTFT refinery can yield 80–85% chemicals and 5–10% LPG fuels as main products. The exact numbers depend on the syncrude properties and the efficiency of the refinery units. Although LPG is produced as a fuel product, the concept goes a long way to achieve the goal of a stand-alone Fischer–Tropch-based petrochemical complex. The changes to industrial LTFT facilities (Chapter 8) and the design of new industrial LTFT facilities (Chapters 11 and 12) strongly indicate that an alkane-based petrochemical refinery is the preferred strategy for LTFT refining. 28.3.2 Aromatics Production
Aromatics constitute about one-third of the market for commodity petrochemicals, making aromatics a key compound class for chemicals production. The main aromatic commodities are the benzene, toluene, and xylene (BTX) group of compounds [13, 14]. The aromatic commodities that do not contain exclusively methyl substituents, are usually derived from the BTX compounds by alkylation with alkenes (Chapter 20) [15, 16]. Depending on the origin of the aromatics, ethylbenzene may also be present in percentage levels, and in such cases the C6 –C8 aromatics are sometimes referred to as benzene– toluene–ethylbenzene–xylenes (BTEX). Benzene itself is not a major product from conventional catalytic naphtha reforming. Much of the benzene produced is derived from toluene disproportionation. HTFT syncrude contains some aromatics (Table 1.2), but even so the BTX content is low. In the case of LTFT syncrude, there are very little aromatics. In most petrochemical complexes, the aromatics are produced by conventional catalytic naphtha reforming at severe conditions. However, syncrude makes very poor feed material for conventional catalytic naphtha reforming (Section 22.3.1). Conversely, the alkenes in the C3 –C4 syncrude make aromatization more efficient. An aromatics-based refining concept for a gas-to-liquids facility has been proposed for the aromatization of C3 –C4 natural gas liquids, straight-run C3 –C4 Fischer–Tropsch syncrude, and the LPG produced during the refining of the syncrude (Figure 28.2) [17]. Aromatics production can be extended to the naphtha range material. One of the methods for chemicals production from Fischer–Tropsch syncrude that has been suggested is to produce BTX aromatics by monofunctional naphtha reforming over a nonacidic Pt/L-zeolite catalyst [18]. This type of reforming has a high aromatics selectivity and matches well with syncrude as feed (Section 22.4) [19]. Although the combination of LPG and naphtha aromatization was not specifically mentioned, such a combination could form the basis for aromatics production in a Fischer–Tropsch-based petrochemicals complex. Whether a further aromatic advantage may be obtained by selecting K¨olbel–Engelhardt-type synthesis [20, 21] rather than traditional Fischer–Tropsch synthesis is not clear. During K¨olbel–Engelhardt synthesis, the aromatic content of the syncrude is higher but the degree of branching increases, which is less favorable for monofunctional naphtha reforming.
28.3 Overview of Chemicals Refining Concepts for Syncrude Fischer – Tropsch synthesis and gas loop
C 1 – C2
C 3 – C4 Recovery of natural gas liquids
Natural gas
C 3 – C4
Liquids
C5 and heavier
Fischer – Tropsch liquids refinery
Refined products
C3 – C4 C 1 – C2 Aromatization
H2
Aromatics Figure 28.2 Aromatic petrochemicals by the integration of an aromatization technology (Zn- or Ga/H-ZSM-5 catalyzed) into a Fischer–Tropsch-based gas-to-liquids facility.
28.3.3 Alkene and Oxygenate Recovery
Alkenes and oxygenates are the most obvious petrochemicals that can be recovered from Fischer–Tropsch syncrude. The opportunity to recover alkenes and oxygenates as chemicals depends on the nature of the syncrude and increases in the order Co-LTFT < Fe-LTFT < Fe-HTFT [22]. This sequence is related to the concentration of alkenes and oxygenates and specifically the concentration in the chemically relevant C1 –C15 range. The Fischer–Tropsch catalyst and technology combination can be specifically selected to maximize the chemical content in the syncrude. For example, it was suggested that an Ru-LTFT-based process can be used to achieve close to 50% n-1-alkenes and n-1-alcohols in the straight-run C4 –C18 product fraction [17]. The associated chemicals refining concept that was proposed was to separate the oxygenates from the syncrude first and then to separate the alkanes and alkenes (Figure 28.3) [17]. Although an oxygenate separation strategy was not suggested, it was indicated that the alkane–alkene separation can be performed with the UOP Olex process. Conceivably, one could employ oxygenate separation in a way related to the technologies employed for n-1-alkene recovery from HTFT syncrude (Section 23.2).
Hydrocarbons Fischer–Tropsch synthesis and Oil gas loop
UOP Olex process
Oxygenate separation
Aqueous Oxygenate recovery
Oxygenates
Figure 28.3 Alkene and oxygenate petrochemicals by separation of the total Fischer–Tropsch syncrude based on polarity.
Alkenes Alkanes Oxygenates
587
588
28 Chemicals and Lubricant Refining
28.3.4 Fuels and Chemicals Coproduction
The philosophy behind the design shown in Figure 28.3 is that Fischer–Tropsch syncrude is a treasure chest filled with valuable chemicals that can be harvested. This is of course true, but recovering these molecules and purifying them to the required level are not always easy. Potential chemical applications of all compounds in syncrude, even when properly separated, are not self-evident either. A more pragmatic approach has been followed in the design of industrial facilities that refine fuels with chemicals coproduction. Compounds that do not have a petrochemical market can then be converted into fuels, which are mixtures. This leads to the design of a complex refinery with many different products. Such large-scale recovery of chemicals from Fischer–Tropsch syncrude was part of the original Sasol 1 design and the subsequent chemicals-directed modifications of the facility in the 1960s [23]. An analogous approach was also followed at Sasol Synfuels [24], as can be seen from the increased complexity of the associated refinery (Figures 9.7 and 9.12). A large number of products are commercially produced from Fischer–Tropsch-based facilities. These include lubricating oils, waxes, alkanes, various oxygenates (e.g., alcohols, ketones, and oxidized waxes), and alkenes (ethene, propene, and n-1-alkenes), as well as associated chemical products from coal liquids, syngas generation (sulfur and nitrogen based compounds), and air separation [25]. In the literature on the chemical potential of Fischer–Tropsch syncrude [26, 27], a catalog of chemicals and chemical intermediates can be found: 1) 2)
3)
4)
5)
6)
7)
Ethene, which can be obtained by cryogenic separation from the Fischer–Tropsch tail gas, is a major product in HTFT syncrude. Ethene has numerous applications [28]. Propene, which is the most abundant chemical in HTFT syncrude, is also present in percentage level quantity in Fe-LTFT and syncrude from deactivated Fe-LTFT synthesis. Propene can be recovered by distillation under pressure and it has numerous industrial applications [29]. 1,3-Butadiene can be produced from the n-butene isomers by catalytic dehydrogenation. Butene is a major product in HTFT syncrude and, like propene, may also be recovered from Fe-LTFT syncrude. Butadiene is used in the production of synthetic butadiene–styrene rubbers. Linear 1-alkenes (linear α-olefins) can be separated from HTFT syncrude, and industrial applications have been listed in Table 23.1. The concentration in LTFT syncrude is much lower than in HTFT syncrude. Waxes can be separated from LTFT syncrudes. The wax fraction constitutes about one-half of the syncrude from LTFT synthesis. Various wax-based products can be prepared by hydrogenation and autoxidation. Alcohols, aldehydes, ketones, and carboxylic acids can be recovered mainly from the aqueous product, but in principle also from the oil product (Figure 28.3). Of these compounds, the carboxylic acids are the most difficult to recover efficiently, because the acids are present in low concentration in the water that is produced during Fischer–Tropsch synthesis. Associated chemical products can also be recovered. These products are by-products in a Fischer–Tropsch-based facility, but are not directly derived from Fischer–Tropsch synthesis.
28.3 Overview of Chemicals Refining Concepts for Syncrude
In a coal-to-liquids facility, it may be possible to recover tar acids (e.g., phenol and cresols), BTX aromatics, sulfur, and ammonia. When an air separation unit is employed, argon and nitrogen are by-products. The large volume of CO2 produced in all Fischer–Tropsch facilities may also be employed for chemical use, for example, in carbonated drinks and fire extinguishers. High-level block flow diagrams that do not show individual units have been proposed to indicate the chemicals that can be produced from HTFT, LTFT, and combined HTFT–LTFT refineries (Table 28.4) [30]. This work indicates that HTFT is better than LTFT for chemical production, both in volume and in value. In some of these proposed chemicals refinery concepts, fluid catalytic cracking (FCC) plays an important role. The assumption based on tests with Fischer–Tropsch material was made that FCC could produce 10–20% ethene and 15–30% propene from catalytic cracking of syncrude [30]. The use of FCC technology with syncrude has been discussed in detail (Section 21.4) and some yield trade-offs have been pointed out. The proposed concepts (Table 28.4) also coproduce transportation fuels as well as some intermediates. The tacit assumption was made that the motor-gasoline and diesel fuel obtained after chemical extraction will meet specification to be sold as final fuels. Table 28.4
Mass balances for combined fuels and chemicals refinery concepts using different syncrudes.
Products
Relative product value Chemical products (mass%) Ethene Ethane Propene Butenes n-Alkanes and LAB Oxygenates Other, for example, n-1-alkenes Fuel products (mass%) LPG Naphtha Motor-gasoline Diesel fuelc,d a
Sasol SPD
Combined fuels and chemicals refining LTFTa
HTFT–LTFTa
HTFTb
100
130
150
150
– – – – – – – 0
– – – – 10 – 2 12
6 – 11 4 3 6 3 33
10 3 19 13 – 12 4 61
2 30 – 68 100
3 18 – 67 88
3 7 8 49 67
4 – 13 23 40
Concept requires external source of benzene. not add up to 100% in source reference. c Stated to be diesel fuel, but may not meet minimum density requirements of specifications such as EN590:2004. d A portion of this material can also be refined to lubricating oils. b Does
589
590
28 Chemicals and Lubricant Refining
Enumerating the chemical possibilities is one thing, but realizing it in practice is another. Transportation fuel specifications have become more restrictive over time, and producing on-specification fuels require increasingly complex refineries (Chapter 2). This dictum applies to Fischer–Tropsch refineries also, and producing on-specification fuels is not trivial (Chapters 25–27). In fact, the extraction of some chemicals listed in Table 28.4, such as the butenes, makes the production of on-specification transportation fuels difficult. The generic problems associated with chemical extraction in a fuels refinery were outlined before (Section 9.5). Although the impact of extraction is often beneficial to fuels refining, it is important to provide a reincorporation pathway if such extraction is not guaranteed [31]. Chemical reincorporation may well be necessary in a stand-alone refinery. When fuel refining is assumed to take place in parallel with chemical production, it is best to evaluate the integrated concept and confirm that the fuels will indeed meet the specification requirements. Refinery designs were presented for fuels and chemicals co-refining to indicate the level of chemical coproduction that can be achieved from HTFT syncrude without jeopardizing fuel specifications (Figures 28.3 and 28.4) [4]. These refinery designs were able to meet EN228:2004 motor-gasoline specifications (Table 13.1) and the distillate was of similar quality to that produced by the PetroSA HTFT refinery (Table 10.4). The distillate does not meet European EN590:2004 diesel fuel specifications but meets American D-2 specifications (Table 15.1). The refinery designs focused only on commodity chemicals with significant global markets, and niche products, such as n-1-alkenes, were deliberately excluded from consideration. It was
Ethene
Ethene
C2-rich tail gas
Thermal cracking
Ethane
LPG Propene
C3
HTFT condensate
C4
C5
Oligomerization (SPA)
Hydrotreater
Gasoline Distillate Gasoline
C5 Hydroisomerization
Gasoline Aromatic alkylation
Cumene
Benzene C6
Light oil
Figure 28.4
H2, LPG, gasoline
>C8 >C9
Decanted oil Aqueous product
Reforming (nonacid Pt/L)
Gasoline Distillate
Hydrotreater/ hydrocracker
Oxygenate recovery
Alcohols Ketones
Simple HTFT fuels and chemicals refinery with 37% yield of chemicals.
28.4 Fischer–Tropsch-Based Petrochemical Refining Ethene
Ethene
C2-rich tail gas
Thermal cracking
Ethane
LPG Propene
C3
C4
Oligomerization (SPA)
HTFT condensate
Hydrotreater
Gasoline TAME
Etherification
C5 skeletal isomerization
C5
Benzene
C6
C5 Hydroisomerization
Gasoline
Aromatic alkylation
Cumene H2, LPG, gasoline
Reforming (nonacid Pt/L)
Light oil C9 –10 >C9
Alcohols
Aromatic alkylation
Distillate LAB
Gasoline Distillate
Hydrotreater/ hydrocracker
Decanted oil Aqueous product
Figure 28.5
Hydroformylation
C11 –14 >C15
Gasoline Distillate
Oxygenate recovery
Alcohols Ketones
Complex HTFT fuels and chemicals refinery with 49% yield of chemicals.
found that with a simple HTFT combined fuels and chemicals refinery (Figure 28.4), 37% of the syncrude could be recovered as chemicals, and with a more complex refinery (Figure 28.5) 49% of the syncrude could be recovered as chemicals (Table 28.5) [4]. In both instances, the objective was to employ refining technologies that had good synergy with Fischer–Tropsch syncrude, rather than increasing the chemicals yield with technologies that did not have an inherently good fit.
28.4 Fischer–Tropsch-Based Petrochemical Refining 28.4.1 Alkane Refining
Most alkane-based petrochemicals are mixtures of alkanes with specific volatility (boiling range) requirements, or melting point properties in the case of waxes. The refining of alkane-based
591
592
28 Chemicals and Lubricant Refining Table 28.5 Mass balances for HTFT combined fuels and chemicals refinery concepts shown in Figures 28.4 and 28.5.
Products
Combined fuels and chemicals refining (mass%) Figure 28.4
Commodity chemicals LPG Motor-gasoline Distillate Nonrecoverable productsa a Organic
37 6 39 13 5
Figure 28.5 49 6 31 10 4
products not recovered and wastewater produced during oxygenate refining.
petrochemicals is dominated by separation processes to obtain the different product fractions. Hydroprocessing is the only type of conversion that is required. Hydrotreating ensures that all reactive functional groups are removed and is also useful in addressing properties such as color and odor. Hydroisomerization may be required to adjust properties such as hardness (penetration) and cold flow. A typical alkane-based petrochemical refinery is shown in Figure 8.12. The petrochemical production of alkanes is best performed using LTFT syncrude as feed material. The Fischer–Tropsch technology affects the syncrude properties, and not all LTFT or HTFT syncrudes have similar properties or are equally well suited for alkane production. Nevertheless, generally speaking, LTFT syncrudes have a number of advantages over HTFT syncrude for alkane-based petrochemicals: 1) 2) 3) 4) 5)
It contains a large fraction of n-alkane waxes. It contains little aromatics. The oxygenates are mainly n-alcohols, which are easily converted into n-alkanes. It is alkane rich, rather than alkene rich. It has a higher ratio of linear to branched isomers.
Because of the large wax fraction in LTFT syncrude, an important petrochemical alternative is lubricant base oil production (Section 28.5). 28.4.2 Light Alkene Refining
The carbon number distribution of HTFT syncrude is such that a large fraction of the straight-run syncrude is in the C2 –C4 range. This naturally favors light alkene recovery. Even though LTFT syncrude has less material in the C2 –C4 range, the light alkenes are still present in percentage levels. Of the LTFT technologies, slurry bed Fe-LTFT is preferred for light alkene production, and deactivation of the Fe-LTFT catalyst over time results in an added selectivity advantage. Fischer–Tropsch-based petrochemical facilities can be designed to exploit advantages that are inherent to the technology. In some respects it makes syncrude a better raw material for
28.4 Fischer–Tropsch-Based Petrochemical Refining
light alkene production than either crude oil or natural gas. The specific advantages include the following: 1) The C2 –C4 hydrocarbons are free from heteroatom contaminants. 2) Cryogenic separation, which is needed for the production of ethene, also benefits the Fischer–Tropsch gas loop design and the overall carbon efficiency of the facility. 3) The straight-run C2 –C4 alkenes are primary products. 4) By including a methane reformer in the Fischer–Tropsch gas loop, methane can be converted back into synthesis gas. In this way, methane produced during petrochemical refining and Fischer–Tropsch synthesis is not a product with no further potential for upgrading. 5) The light alkanes can be refined to light alkenes in an analogous way to crude-oil-derived light alkanes, but it is sulfur free and reportedly has a selectivity advantage [32]. 6) Light alkenes can also be produced with high selectivity by dehydration of alcohols in the Fischer–Tropsch aqueous product [33]. 7) Fischer–Tropsch synthesis can be tailored to increase light alkene selectivity. The straight-run alkene production can be increased by decreasing the degree of hydrogenation that takes place during Fischer–Tropsch synthesis: for example, by selecting a reactor with less plug-flow behavior or using a Fischer–Tropsch catalyst that is less hydrogenating. The production of straight-run C2 –C4 hydrocarbons can be further increased by adjusting the chain growth probability (α-value) of the Fischer–Tropsch catalyst (Figure 28.6). A decrease in the α-value comes at the expense of increased methane production, as well as increased water-soluble oxygenate selectivity. The maximum yield of straight-run C2 –C4 hydrocarbons is therefore not necessarily defining the optimum operating point.
70
C2 –C4
60
Yield (mass%)
50 40 C1
30 20
C5 –C10 (naphtha)
10 C11 and heavier
0 0.3
0.4
0.5 0.6 0.7 Fischer–Tropsch chain growth probability
Figure 28.6 Yield of light hydrocarbons from Fischer–Tropsch synthesis as calculated from the Anderson–Schulz–Flory distribution. The yield of C1 –C2 was estimated on the basis of the C3 yield.
0.8
0.9
593
594
28 Chemicals and Lubricant Refining
Purge Methane reformer
H2
PSA
CH4 + H2 Ethane
Tail gas
Cryogenic separation
Ethene
C3 –C4 separation
Propene Butenes Butadiene
C3 –C4 alkanes
Syngas
Fischer – Tropsch synthesis
Thermal cracking
Pyrolysis oil
Oil Alcohol dehydration Carbonyl hydrogenation
Aqueous
Figure 28.7
Alcohols Wastewater
Stand-alone Fischer–Tropsch-based petrochemical facility producing light alkenes.
The ability to manipulate the product distribution from Fischer–Tropsch synthesis (Section 4.4), in combination with an appropriate petrochemical refinery design, provides a unique opportunity for light alkene production. It is possible to envisage a stand-alone petrochemical facility that produces mainly C2 –C4 alkenes (Figure 28.7). The recovery of straight-run Fischer–Tropsch alkenes is combined with additional alkene production in the syncrude refinery. The main unit is a thermal cracker, which converts various feed materials into light alkenes. There is synergy in such a combination of Fischer–Tropsch synthesis with thermal cracking, because the light gas separation section can be shared. By purifying the Fischer–Tropsch tail gas together with the cracked gas, the overall capital cost of the facility is reduced. The feed to the thermal cracker is a combination of the straight-run C5 and heavier oil product and the C2 –C4 alkanes from light alkene recovery. In the design shown in Figure 28.7, the aqueous product is refined separately and only the tail gas separation section is shared. The carbonyl compounds in the aqueous product are selectively hydrogenated to produce alcohols, which are then dehydrated with the alcohols that are already present in the aqueous product. A high selectivity of ethene from ethanol and propene from propanol can be obtained in this way. A simplification of the design involves the conversion of the aqueous product oxygenates with the other material in the thermal cracker, albeit at lower selectivity to the targeted alkenes. 28.4.3 Linear 1-Alkene Refining
Straight-run Fischer–Tropsch syncrude contains alkenes that are petrochemical products. The extraction of these n-1-alkenes is practised industrially, and the technologies for such extraction
28.4 Fischer–Tropsch-Based Petrochemical Refining
were discussed (Chapter 23). It is possible to recover n-1-alkenes from both HTFT and LTFT syncrude, but for each carbon number the concentration of the n-1-alkenes is lower in LTFT than HTFT syncrude. HTFT syncrude is the preferred feed material for the production of n-1-alkenes. Although the n-1-alkene market is very lucrative, designing a Fischer–Tropsch-based petrochemical facility for the production of n-1-alkenes should take into account the following: 1) The total market for all n-1-alkenes (Table 28.2) is much smaller than that of the primary commodity chemicals (Table 28.1). 2) Recovering n-1-alkenes from syncrude must compete with n-1-alkene synthesis from ethene. The complexity of the purification of n-1-alkenes from syncrude increases with the carbon number. The trade-off is analogous to that between direct and indirect liquefaction. Direct recovery of n-1-alkenes is more efficient, but further refining to remove impurities is more demanding. Producing n-1-alkenes indirectly from ethene has an inherent purity advantage, although the initial production cost is higher. 3) Fischer–Tropsch synthesis determines the n-1-alkene content in the syncrude. Since production relies on refinery separation, not conversion, the yield is fixed. 4) Petrochemical production of n-1-alkenes targets only a fraction of the syncrude, and for the most part the design of such a petrochemical facility will be determined by the refining of the remainder of the syncrude. One would not design a facility primarily for the production of n-1-alkenes. 28.4.4 Aromatics Refining
There are four technologies that can be considered for the production of aromatics from Fischer–Tropsch syncrude: Conventional catalytic naphtha reforming over Pt/Cl− /Al2 O3 -based catalysts (Section 22.3). The majority of the C6 –C8 aromatic petrochemicals are produced from crude oil in this way. Because of the low N + 2A of syncrude, it is not an efficient technology for syncrude conversion into aromatics. 2) Monofunctional naphtha reforming over nonacidic Pt/L-zeolite-based catalysts (Section 22.4). Industrially, some of the C6 –C8 aromatic petrochemicals are produced from crude-oil-derived naphtha in this way, despite the extreme sulfur sensitivity of the catalyst. This technology has an excellent technology fit with syncrude, not only due to the sulfur-free nature of the syncrude but also due to the high linear hydrocarbon content. It has the further advantage that benzene can directly be produced in high yield from hexane, rather than through toluene disproportionation. 3) Aromatization of hydrocarbons over Ga- or Zn/H-ZSM-5-based catalysts (Section 22.5). Industrially, some of the C6 –C8 aromatic petrochemicals are produced in this way. This type of aromatic conversion can be efficiently integrated with a Fischer–Tropsch gas loop for tail gas aromatization (Figure 22.7). It is also a useful alternative to thermal cracking for the conversion of light alkanes into petrochemicals. 4) Carbonyl aromatization over acidic catalysts (Section 16.3.5). This conversion is not industrially applied for aromatics production. Although it is capable of producing some aromatics 1)
595
596
28 Chemicals and Lubricant Refining
Purge Methane reformer
PSA
CH4 rich
Syngas
Ethane Fischer – Tropsch synthesis
Tail gas
H2
Cryogenic separation
Ethene
C3 – C4 separation
Propene Butenes Butadiene
C3 –C4 alkanes
Oil
Aromatization
C5
Aromatics
Hydrotreater
C6 –C10 Reforming
Gas Oil
Gas
Aromatics separation
Aromatics
Aliphatics
Thermal cracking
Aqueous
Wastewater
Figure 28.8 Stand-alone Fischer–Tropsch-based petrochemical facility producing light alkenes and aromatics.
from HTFT syncrude, only ethanal yields BTX aromatics, but ethanal is present in too small an amount to contribute significantly to petrochemical production. These aromatics-producing technologies all make use of C10 and lighter material. A stand-alone Fischer–Tropsch-based petrochemical complex for aromatics production only is therefore impractical. A cracking unit to convert the heavier material is suggested, which implies that alkene and aromatics should preferably be coproduced, as is often found in petrochemical facilities. The α-value can be manipulated to yield mainly C10 and lighter products (Figure 28.6) to avoid a cracking unit, but when doing so the light olefinic syncrude still favors alkene and aromatics coproduction. Conceptually, alkene and aromatics production goes hand in hand, and it opens downstream refining pathways for the production of commodities such as ethylbenzene, styrene, and cumene. A stand-alone Fischer–Tropsch-based petrochemical facility (Figure 28.8) has many design features in common with a crude-oil-based facility (Figure 2.14). However, integration of light hydrocarbon separation with the Fischer–Tropsch gas loop, the lighter carbon number distribution, and the more reactive sulfur-free syncrude make such a facility more efficient. The ratio of alkene to aromatics production can be changed by routing more material to the cracker and including only a single aromatics production technology.
28.5 Fischer–Tropsch-Based Lubricant Base Oil Refining
28.4.5 Oxygenate Refining
Industrial designs for the recovery and separation of oxygenates from the Fischer–Tropsch aqueous product were discussed (Sections 7.4.2, 8.4.3, 9.4.3, and 10.4.2). Aqueous product refining concepts were also discussed in the context of fuels refinery design (Section 25.3.4). Oxygenates can be recovered from the Fischer–Tropsch oil product by the separation strategies employed for n-1-alkene separation (Section 23.2). However, many of the heavier oxygenates are not large-volume commodities, and the concentration of plasticizer and detergent range alcohols present in straight-run syncrude is comparatively low. Unless the Fischer–Tropsch technology is selected to produce a larger concentration of alcohols, the alcohol concentration is insufficient to justify extraction and purification as petrochemicals. If the objective is to produce alcohols specifically as the primary synthesis product, one should consider syngas-to-methanol technology [34] instead of Fischer–Tropsch synthesis. There are also various accounts of oxygenate-selective Fischer–Tropsch-type catalysts in the literature if oxygenates other than methanol is of interest. However, oxygenate recovery and purification from a complex matrix is not trivial, and this may not be the most efficient route for the large-scale production of heavier oxygenate-based chemicals. Oxygenates can be synthesized from hydrocarbons with good selectivity, and industrially most of the heavier oxygenates are prepared in this way. Specific synthetic routes that can be considered in conjunction with syncrude are the following: 1) Alcohol synthesis by hydroformylation of alkenes (Section 16.3.6). Syncrude contains not only n-alkenes in significant concentration, but H2 and CO are available from the syngas in the Fischer–Tropsch gas loop. Short-chain alcohols can also be synthesized by direct hydration of alkenes (Section 17.4). 2) Oxidized hydrocarbons through autoxidation of alkanes (Section 23.3). There is no clear syncrude advantage, except with respect to feed availability and maybe purity for some specific cuts. There are many other possibilities, and syncrude can in principle be employed as feed material for any of the oxygen-containing petrochemicals that are synthesized from crude oil.
28.5 Fischer–Tropsch-Based Lubricant Base Oil Refining
Lubricant base oil is a heavy product. Processing a full-range Fischer–Tropsch syncrude to lubricant base oil in a stand-alone facility will have to be combined with either fuels or petrochemicals production in order to be efficient. The main classes of lubricants that can be easily produced in large volume from syncrude are the following: 1) API Group III lubricant base oils prepared from the Fischer–Tropsch residue fraction by hydroisomerization and mild hydrocracking. LTFT syncrude is the preferred feed material and it does not require extensive hydrodearomatization (HDA) as in the case of HTFT syncrude. The volume of lubricant that can be produced is determined by the syncrude employed, but additional lubricant can be produced by oligomerization.
597
598
28 Chemicals and Lubricant Refining
2) API Group IV lubricant base oils prepared from the alkenes in syncrude. The n-1-alkenes can be converted into PAO lubricant base oils, the quality being dependent on the carbon number range of the alkenes and the internal alkene content. PIO lubricant base oils can be readily prepared. In both cases, the syncrude must be deoxygenated and an appropriate oligomerization catalyst must be selected. The volume of the latter can be increased through autoxidation and dehydration of the alkanes to increase the internal alkene content. The volume of PAO lubricant is limited by the type of syncrude employed. The volume of PIO lubricant can be readily increased through refining. It is possible to produce some API group V lubricant base oils, but not in the same yields or with the same ease as the group III and IV lubricant base oils. Since the group V lubricant base oils are all synthetic oils, in principle one can use syncrude like crude oil as feedstock, but in such syntheses the syncrude has no specific advantage over crude oil. The two classes of lubricant base oils that have a potential feedstock advantage are the esters and neutral trialkyl phosphate esters. In both cases, the feedstock advantage comes from the 1-alcohols in the syncrude, the volume of which may be increased through synthesis (Section 28.4.5). Ester-based lubricants employ naphtha and distillate range 1-alcohols. Lubricant base oil can be produced by in situ esterification, which has processing synergy with hydroformylation as a feed preparation step, albeit using longer chain alcohols [35]. Neutral trialkyl phosphate esters are also produced from naphtha and distillate range 1-alchols. Industrially, the dominant route for synthesis is the reaction of the alcohols with phosphorus oxychloride [36]. 28.5.1 Group III Lubricant Refining
The refining of LTFT waxes into lubricating base oil combines the removal of one of the undesirable base oil constituents, namely waxes, with the production of a high-VI oil. The n-alkane waxes have a good VI in the liquid state, but are undesirable due to their high melting points. By introducing some branching into the molecular structure, the melting point is drastically lowered, but this is accompanied by a decrease in the VI of the isomerized product. There is a trade-off between the need for better cold-flow properties and retaining the inherently high VI of the n-alkanes. The VI of lubricant base oils prepared by the hydroisomerization and mild hydrocracking of LTFT wax is determined by the distillation range and the average number of branches per molecule in the oil. A correlation was developed to relate the average carbon number (ACN) and the average branching number (ABN) of the oil to the VI (Equations 28.1 and 28.2) [37]. VI = −0.0008χ 2 + 0.7599χ − 17.449 χ = (ACN)2 (ABN)−1
(28.1) (28.2)
The correlation was found to be independent of the aliphatic hydrocarbon feedstock and also of the operating conditions of hydroprocessing [37]. This makes sense, because it correlates molecular properties with the physical properties. When the correlation is used in conjunction with the cold-flow requirements, which is also related to the ACN and the ABN, it is possible to calculate the optimum degree of hydroisomerization.
28.5 Fischer–Tropsch-Based Lubricant Base Oil Refining
It can also be shown that a lubricating base oil produced from the atmospheric residue of LTFT syncrude will meet API group III specifications unless the degree of isomerization is extremely high. A dibranched C22 alkane has a calculated VI of 120. VIs of 150 and higher that were obtained by the hydroisomerization of thermally cracked Fischer–Tropsch wax and polyethylene waste plastic [38] are therefore understandable. The refining of HTFT residue to lubricant base oil has been demonstrated [10], and the refining is akin to the hydroprocessing of crude-oil-derived residues. 28.5.2 Group IV Lubricant Refining
It is only the alkene feed material employed for the production of PAO and PIO lubricant base oils that differs, namely n-1-alkenes and internal alkenes. The same catalysts and technologies can be otherwise employed. The alkenes are oligomerized (Chapter 19) and then hydrogenated. The challenge is to prevent skeletal isomerization, cracking, and double bond isomerization as side reactions during oligomerization. Brønsted acid–catalyzed processes are understandably unsuitable for PAO production, since double bond isomerization will result in the conversion of the n-1-alkenes into internal alkenes to produce PIO lubricant base oil. The selection of the catalyst [39] or the use of free radical oligomerization [40] is critical to the quality of the final product. The VI of the PAO lubricant base oil improves with increasing carbon number of the n-1-alkene used as oligomerization feed (Table 19.7). This can be understood in terms of the methylene chain index description of Kobayashi and coworkers [37], who related the VI to the chain length of branching structures. The methylene chain index is calculated from the sum of the methylene length (ML) contributions of all CH2 structures in the molecule (Equation 28.3) Methylene chain index = (MLi )1.5
(28.3)
Primary, tertiary, and quaternary carbons are not counted. It is only the lengths of the CH2 -containing branches that are counted (Figure 28.9). The longer the chain length of the n-alkyl branches, the higher the methylene chain length index, which is also an indirect measure of the ACN. The methylene chain length index is proportional to the VI (Equation 28.4). VI ∝ 2.4 · (Methylene chain index) ML = 5 2 1
ML = 2
4 3
ML = 6
2 5
(28.4)
1
1
3 2
5 4
6
1 2
ACN = 22 ABN = 2
ML = 3 3
Methylene chain index = 51.5 + 21.5 + 61.5 + 31.5 = 33.9 Figure 28.9 Calculation of the methylene chain index for aliphatic lubricant base oils as defined by Kobayashi and coworkers.
599
600
28 Chemicals and Lubricant Refining
Although this is just an empirical relationship of some quantitative value, it has considerable qualitative value. For example, it explains the increase in VI of PAO oil with increasing length of the alkene monomer, as well as their superior VI compared to those produced from internal alkene monomers of the same chain length. As such, it is useful as a guidance for developing a refining strategy to convert distillate range alkanes into PIO lubricant base oils. These distillate range alkanes are not suitable for conversion into API group III base oils, but can be converted into internal alkenes by catalytic dehydrogenation or autoxidation followed by an appropriate deoxygenation step [41]. The internal alkanes can then be oligomerized to PIO oils. On the basis of the methylene chain index (Equation 28.3), a PIO oil requires an internal alkene feed that is two carbon numbers heavier than the n-1-alkene feed for a PAO oil to achieve a comparable VI. In an LTFT lubricant base oil refinery, the contribution of group IV oils will be less than that of group III oils (Section 28.5.1). LTFT syncrude contains some n-1-alkenes that can be converted into PAO lubricant base oils. The amount of n-1-alkenes can be further increased by dehydration of the 1-alcohols if required. The distillate range alkanes can be converted into internal alkenes for conversion into PIO oils. In an HTFT lubricant base oil refinery, the relative abundance of alkenes favors the production of PAO and PIO oils over that of group III oils. 28.5.3 Lubricant Base Oil Refining
The heavier carbon number distribution of LTFT syncrude makes it a better feed material for lubricant base oil refining than HTFT syncrude. Most of the distillate and heavier material in the syncrude can be converted into lubricant base oils (Figure 28.10). The atmospheric residue and wax can be mildly hydrocracked and hydroisomerized to produce high-VI API group III lubricant base oils. Part of the distillate range material can also be Fischer – Tropsch synthesis
Oil
Tail gas Aqueous
Chemicals fuels
Fischer–Tropsch refinery
C5 –C10
C11 –C22
Dehydration/ deoxygenation
Oligomerization
Dehydrogenation
Oligomerization
PAO oil
>C22
PIO oil Hydrocracking/ hydroisomerization
Group III oil
Wax
Figure 28.10
Stand-alone Fischer–Tropsch-based lubricant base oil facility.
References
converted into lubricant base oils. The n-1-alkenes in the straight-run distillate, as well as those produced from 1-alcohol dehydration, can be oligomerized to API group IV PAO lubricant base oils. The alkanes can then be converted into internal alkenes, which can be oligomerized to PIO lubricant base oils. The naphtha and lighter material are not suitable for conversion into high-quality lubricant base oil. However, the naphtha and lighter fractions of syncrude make an excellent feed for petrochemical conversion (Section 28.4). It is not practical to design a stand-alone Fischer–Tropsch lubricant base oil refinery without an associated petrochemical or fuels refinery.
References 1. Spillane, L.J. (1970) in Refining of Petroleum
2.
3.
4.
5.
6. 7.
8.
9.
10.
for Chemicals, Advanced Chemical Series, Vol. 97 (eds L.J. Spillane and H.P. Leftin), ACS, Washington, DC, pp. vii–vix. Chadwick, J.L. (1977) Economics of Chemical Refineries. SRI Process Economics Program Report 107, Stanford Research Institute, Menlo Park, CA. Davis, S. (2008) Petrochemical Industry Overview. SRI Chemical Economics Handbook, Marketing Research Report Section 350, Stanford Research Institute, Menlo Park, CA. De Klerk, A., Dancuart, L.P., and Leckel, D.O. (2005) Chemicals refining from Fischer-Tropsch synthesis. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd185. Chauvel, A. and Lefebvre, G. (1989) Petrochemical Processes Technical and Economic Characteristics, Synthesis Gas Derivatives and Major Hydrocarbons, Vol. 1, 2nd edn, Technip, Paris. SRI Chemical Economics Handbook, marketing research reports (various). Stipanovic, A.J. (2003) in Fuels and Lubricants Handbook: Technology, Properties, Performance, and Testing (ed. G.E. Totten), ASTM, West Conshohocken, PA, pp. 169–184. Henderson, H.E. (2006) Gas to liquids, in Synthetics, Mineral oils, and Bio-Based Lubricants. Chemistry and Technology, Chapter 19 (ed. L.R. Rudnick), Taylor & Francis, Boca Raton, FL. ASTM (2004) D 2270 – 04. Standard practice for calculating viscosity index from kinematic viscosity at 40 and 100 ◦ C, ASTM, West Conshohocken, PA. Morgan, P.M., Van der Merwe, D.G., Goosen, R., Leckel, D.O., Saayman, H.M., Loubser, H., and Botha, J.J. (1999) The production of high
11.
12.
13. 14.
15.
16.
17.
18.
19.
20.
21.
quality base oils from Sasol hydrocrackates. S. Afr. Mech. Eng., 49, 11–13. Buenemann, T.F., Boyde, S., Randles, S., and Thompson, I. (2003) in Fuels and Lubricants Handbook: Technology, Properties, Performance, and Testing (ed. G.E. Totten), ASTM, West Conshohocken, PA, pp. 249–266. Schubert, P.F., Bayens, C.A., and Weick, L. (2001) Development progresses for GTL fuels, speciality products. Oil Gas J., 99 (11), 69–73. Hancock, E.G. (ed.) (1975) Benzene and its Industrial Derivatives, Ernest Benn, London. Hancock, E.G. (ed.) (1982) Toluene, the Xylenes and their Industrial Derivatives, Elsevier, Amsterdam. Degnan, T.F. Jr., Smith, C.M., and Venkat, C.R. (2001) Alkylation of aromatics with ethylene and propylene: recent developments in commercial processes. Appl. Catal. A, 221, 283–294. Perego, C. and Ingallina, P. (2002) Recent advances in the industrial alkylation of aromatics: new catalysts and new processes. Catal. Today, 73, 3–22. Gregor, J.H. (1990) Fischer-Tropsch products as liquid fuels or chemicals. An economic evaluation. Catal. Lett., 7, 317–332. Dry, M.E. (1996) Practical and theoretical aspects of the catalytic Fischer-Tropsch process. Appl. Catal. A, 138, 319–344. De Klerk, A. (2008) Fischer-Tropsch refining: technology selection to match molecules. Green Chem., 10, 1249–1279. Chaffee, A.L. and Loeh, H.J. (1985) Aromatic hydrocarbons from the K¨olbel-Engelhardt reaction. Appl. Catal., 19, 419–422. Chaffee, A.L. and Loeh, H.J. (1986) The K¨olbel-Engelhardt reaction over a silica supported nickel catalyst. Variation of product distributions with reaction conditions. Appl. Catal., 26, 123–139.
601
602
28 Chemicals and Lubricant Refining 22. Redman, A. (2005) Production of olefins and
23. 24.
25.
26.
27. 28. 29. 30.
31.
32.
oxygenated compounds from Fischer-Tropsch. Preprints of the 18th World Petroleum Congress, September 25–29, 2005, Johannesburg, p. cd179. Meintjes, J. (1975) Sasol 1950–1975, Tafelberg, Cape Town. De Klerk, A. (2009) in Advances in Fischer-Tropsch Synthesis, Catalysts, and Catalysis (eds B.H. Davis and M.L. Occelli), Taylor & Francis, Boca Raton, pp. 331–364. De Klerk, A. and Furimsky, E. (2010) Catalysis in the Refining of Fischer–Tropsch Syncrude, Royal Society of Chemistry, Cambridge. Dry, M.E. (1987) Chemicals produced in a commercial Fischer-Tropsch process. ACS Symp. Ser., 328, 18–33. Dry, M.E. (1988) The Sasol route to chemicals and fuels. Stud. Surf. Sci. Catal., 36, 447–456. Miller, S. A. (ed.) (1969) Ethylene and its Industrial Derivatives, Ernest Benn, London. Hancock, E.G. (ed.) (1973) Propylene and its Industrial Derivatives, Ernest Benn, London. Steynberg, A.P., Nel, W.U., and Desmet, M.A. (2004) Large scale production of high value hydrocarbons using Fischer-Tropsch technology. Stud. Surf. Sci. Catal., 147, 37–42. Dancuart, L.P., De Haan, R., and De Klerk, A. (2004) Processing of primary Fischer-Tropsch products. Stud. Surf. Sci. Catal., 152, 482–532. Dancuart, L.P., Mayer, J.F., Tallman, M.J., and Adams, J. (2003) Performance of the Sasol SPD naphtha as steam cracking feedstock. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 48 (2), 132–138.
33. Nel, R.J.J. and De Klerk, A. (2007)
34.
35.
36.
37.
38.
39.
40.
41.
Fischer-Tropsch aqueous phase refining by catalytic alcohol dehydration. Ind. Eng. Chem. Res., 46, 3558–3565. Tijm, P.J.A., Waller, F.J., and Brown, D.M. (2001) Methanol technology developments for the new millennium. Appl. Catal. A, 221, 275–282. Bolder, F.H.A. (2009) Removal of carboxylic acids in Fischer-Tropsch hydrocarbon product. Prepr. Pap.-Am. Chem. Soc., Div. Petrol. Chem., 54 (1), 1–3. Harris, J.S. (1952) in Phosphoric Acid, Phosphates and Phosphatic Fertilizers, 2nd edn (ed. W.H. Waggaman), Reinhold, New York, pp. 486–506. Kobayashi, M., Saitoh, M., Ishida, K., and Yachi, H. (2005) Viscosity properties and molecular structure of lube base oil prepared from Fischer-Tropsch waxes. J. Jpn. Petrol. Inst., 48 (6), 365–372. Miller, S.J., Shah, N., and Huffman, G.P. (2005) Conversion of waste plastic to lubricating base oil. Energy Fuels, 19, 1580–1586. Brennan, J.A. (1980) Wide-temperature range synthetic hydrocarbon fluids. Ind. Eng. Chem. Prod. Res. Dev., 19, 2–6. Seger, F.M., Doherty, H.G., and Sachanen, A.N. (1950) Noncatalytic polymerization of olefins to lubricating oils. Ind. Eng. Chem., 42, 2446–2452. Bolder, F.H.A., De Klerk, A., and Visagie, J.L. (2009) Hydrogenation of oxidized wax and a process to produce olefins from paraffins by autoxidation, selective hydrogenation, and dehydration. Ind. Eng. Chem. Res., 48, 3755–3760.
603
Index
a acid-catalyzed oligomerization, in acid interconversion 388 acid gas removal 67–69 – solvents and processes for 68 acid number, determination of 481 acid strength, in alkene oligomerization 375 acidic molecular sieve, pentene skeletal isomerization 359 acidic resin catalysts 341 – in alkene oligomerization 381–383 active carbon gasoline 124 adiabatic oxidative reforming 6, 53–54, 56–57 – advantages of 57 adiabatic prereforming 55–56 agglomeration 66 – in Co-LTFT catalysts 98 air pollution 39, 40–41 Air Quality Improvement Research Program (AQIRP) 40–41 air separation unit (ASU) 70–71, 70 ‘‘A-K’’ gasoline 130 Al Khaleej field 241 alcohol 136–137, 565, 588 – in biomass refining 310 – composition 137 – etherification 338 – and motor gasoline 263 – as primary autoxidation products 480 – refining 567 aldehydes 588 aldol condensation 336 aliphatic alkylation 223, 316–317 aliphatic hydrocarbons 124, 125 alkane(s) – conversion – – dehydrogenation 322–323 – – hydrocracking 320–321
– – hydroisomerization 319–320 – – naphtha reforming 321–322 – in crude oil 23 – hydrocracking of 445 – hydrodecyclization of 445 – hydroisomerization of 356–357, 445 – octane numbers of 256–257 – refining 585–586, 591–592 – – and motor-gasoline refining 518–519 alkene(s) 137–138, 147, 563 – to alcohol conversion by hydroformylation 466 – in aromatic alkylation – – cracking propensity of 397 – – phase behavior 397 – – pore size restrictions 397 – – reactivity of 396–397 – complete HYD 307 – composition, in alkene oligomerization 374–375 – conversion – – aliphatic alkylation 316–317 – – aromatic alkylation 317–319 – – double bond isomerization 312–314 – – metathesis 314 – – oligomerization 315–316 – – skeletal isomerization 314–315 – cyclization of 457 – disproportionation. see metathesis – hydration 309 – hydrogenation 213 – octane numbers of 257 – oligomerization 189–190, 190, 213 – – acidic resin catalysts 381–383 – – amorphous silica–alumina (ASA) 380–381 – – applications of 370–372 – – of broad cut of HTFT naphtha 228
Fischer–Tropsch Refining, First Edition. Arno de Klerk. 2011 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2011 by Wiley-VCH Verlag GmbH & Co. KGaA.
604
Index alkene(s) (contd.) – – classic Whitmore-type carbocation mechanism 372 – – ester-based mechanism 372–373 – – free radical mechanism 373–374 – – heat management in 386 – – homogeneous nickel catalysts 383–384 – – H-ZSM-5 zeolite 378–380 – – organometallic insertion mechanism 373 – – SPA catalyst 375–378 – – syncrude process technology 385–388 – – thermal oligomerization 384–385 – and oxygenate recovery 587, 587 – partial HYD 306 alkene skeletal isomerization 354–356 alkyl aromatics 394 – cracking propensity of 397 alkylation 446 alkylperoxy radical (ROO) 478 α-value – of catalyst 78 – two, distribution 78 alumina (Al2 O3 ) – and attrition resistance 92 – pentene skeletal isomerization 359 η-alumina catalyst 203 American Hydrocol facility 141–151 – Fischer–Tropsch synthesis 143–145 – – gas loop 144 – – gas scrubber 144 – – oil scrubber 144 – Fischer–Tropsch refining 145–150 – – aqueous product 148–149 – – oil product 146–148 – refinery design 146, 146, 150–151 – syncrude – – branching 145 – – compound classes, distribution 145 – syngas production 142–143, 142 American Petroleum Institute (API) classification of lubricant base oils 584, 585 amides, in crude oil 25 amines, in crude oil 25 ammonia (NH3 ) 7 – in catalyst deactivation 90 amorphous metal oxide catalysts 341 amorphous silica–alumina (ASA), in alkene oligomerization 380–381 Anderson–Schulz–Flory (ASF) distribution 78, 79, 513, 542, 561 Anglo-Transvaal Consolidated Investment Company 153 aniline 401 antiknock behavior 33
aqueous product 114, 115, 151, 222, 565, 571 – diesel fuel and 565–566 – and motor-gasoline refining 517–518 – refining 136–137, 148–149, 166–169, 194–196, 196, 210–211, 225–226, 228 – – alcohol purification 168 – – chemical workup 167, 168 – – high-level options for 517 – – water-soluble oxygenates, composition 194 Arge LTFT oil refining 165–166, 165 – bauxite 166 – thermal cracking 166 – wax hydrogenation 166 Arge LTFT synthesis 159–161, 160 – compound classes 161 – gas loop 162 – syncrude composition 161 aromatic alkylation 317–319 – with C4 alkenes 401–403 – catalysis 396–397 – as conversion technology 393–395 – with ethene 397–399 – with heavy alkenes 401–403 – and isomerization 354 – with propene 399–401 – reaction chemistry of 395–396 – syncrude process technology 403–405 aromatic isomerization 354 aromatics 547, 563 – in crude oil 23 – octane numbers of 258 – production 586, 587, 595–596 – refining 595–596 aromatization 454, 460, 546 – catalysis 457–459 – reaction chemistry 456 – syncrude processing technology 460–461 aromax technology 450 ash content 66 ash fusion temperature 66 asphalt 35 asphaltene content in crude oil 28 associated gas 51 ASTM D 86 standard test method 260 ASTM D 323 standard test method 260 ASTM D 525 standard test method 261 ASTM D 613 standard test method 287, 296 ASTM D 873 standard test method 261 ASTM D 910 standard test method 264 ASTM D 975 specifications 285, 288 ASTM D 1322 standard test method 276 ASTM D 1655 standard 270, 271–272 ASTM D 2274 standard test method 294 ASTM D 2386 standard testing method 276
Index ASTM D 2500 standard test method 293 ASTM D 3241 standard test method 278 ASTM D 4625 standard test method 294 ASTM D 4809 standard test method 274 ASTM D 5001 standard test method 280 ASTM D 6079 standard test method 291 ASTM D 6468 standard test method 294 ASTM D 6890 standard test method 290 atmospheric distillation 44, 169, 170, 176, 178 atmospheric distillation unit (ADU) 521 atmospheric oxidation 136 attrition of catalysts 92 autoignition 253, 255, 256 – resistance 255, 258, 287 automobiles 32 autothermal reforming (ATR) technology 57, 58, 243, 244 autoxidation 58, 258, 294, 311–312, 466, 474, 479 – of distillate 567 – Fischer–Tropsch wax oxidation 480–484 – reaction chemistry 478–480 – regimes 477–478 – syncrude process technology 484–485 – of waxes 136 average branching number (ABN) 598 average carbon number (ACN) 598 aviation – gasoline 33, 35, 264–265 – – specifications 264 – turbine engines in 35 – turbine fuel 269 aviation bright stock 134
b ball-on-cylinder lubricity evaluator (BOCLE) 246 base oils. see lubricant base oils battlefield use fuel of the future (BUFF) 272, 280 bauxite 146–147, 147, 150, 166 Benfield CO2 removal process 185 benzene 262, 457 benzene, toluene, and the xylenes (BTXs) 43, 586 benzene–toluene–ethylbenzene–xylenes (BTEX) 586 benzonitrile 401 benzothiophene 401 bimolecular partial alcohol dehydration 336, 337 biomass derived feed, logistics and pretreatment 5 biomass-to-liquids (BTLs) process 3 blending octane number 256 boiler fuel 33 boiling range broadening 303 boiling ranges of crude oil 29–31
bond dissociation energies (BDEs) 410, 410, 412–413, 413 branched alkanes 273 – oligomerization process in 385–386 BritishGas/Lurgi (BGL) slagging gasifier 61 bromide ions, in catalyst deactivation 89–90 Brønsted acid-catalyzed alcohol dehydration, mechanism for 340 Brønsted acid–catalyzed processes 599 1,3-butadiene 588 butene 544 – isomerization catalysis 358–359, 362 butoxybutane 338
c C2 –C4 alkenes 43 C2 –C4 hydrocarbons – diesel fuel and 565 – jet fuel and 544–545 – and motor-gasoline refining 515–516 C2 –C4 selectivity 79 – and methane selectivity 98–99 C3 –C4 Crude LPG – refining 129–130 C5 –C6 naphtha 551–552 – hydroisomerization 360 – isomerization catalysis 362–363 C6 −C8 aromatics 43 C6 -cut 469 – preparation of 467–468 C7 naphtha – hydroisomerization 360 C8 -cut 470–474 C84/3 SPA catalyst 203 capital cost 17–19, 19 capital expenditure, minimizing 506 carbide mechanism, for chain growth 74 carbon deposition – and Co-LTFT catalyst deactivation 96 – and Fe-HTFT catalyst deactivation 93 – and methane selectivity 97 carbon dioxide (CO2 ) 7, 12 – removal of 67–69, 185 carbon efficiency 14, 15, 496 – and capital cost 18 – in German facilities 128, 128 carbon-efficient refining 304 carbon gasoline, refining 130–132 carbon number – and hydroisomerization technology 361–362 – -based refining 303 – – refinery design 500 carbon number broadening 228
605
606
Index carbon number distribution 561–562 – in alkene oligomerization 374 – and chain growth probability 78 – in oligomerization process 386 carbon rejection technology 210 carbon residue of crude oil 28 carbonates 258 carboxylic acids 137, 228, 480, 481, 588 – in aqueous products 167–168, 167 – in biomass refining 310 – in crude oil 26, 27 – leaching, and catalyst deactivation 92 Carthage Hydrocol Company 141 catalyst deactivation, Fischer–Tropsch 88–99 – Co-LTFT catalysts, deactivation of 95–99 – Fe-HTFT catalysts, deactivation of 93 – Fe-LTFT catalysts, deactivation of 93–95 – mechanical catalyst degradation 91 – metal carboxylate formation 90–91 – poisoning by syngas contaminants 89–90 – volatile metal carbonyl formation 90 catalyst formulation, Fischer–Tropsch – and product selectivity manipulation 81–83 – – bifunctionality 83 – – catalyst stability 82 – – hydrogenation activity 82 – – promoters, sensitivity to 82 – – support material 82 – – WGS activity 82 catalyst geometry, in alkene oligomerization 375 catalyst poisons, removal of 7 catalyst wax 124 catalyst(s) – nickel-based reforming 55 – selection 303 catalytic cracking 210, 407, 414–416, 421, 545–546 – catalysis 423–425 – syncrude processing technology 425–427 catalytic dewaxing 45 catalytic distillation 206 catalytic naphtha reformer 192–193 catalytic partial oxidation (CPO) 57 catalytic reforming technology 35 central luconia 231 cetane–density relationship of various compounds 564 cetane index 287 cetane number 42–43 – improvers 573 chain growth 74–76 – mechanistic representation of 74–75 – probability 78–80 – – C2 selectivity 79
– – methane selectivity 78 – – two α-value distribution 79 chain termination reactions 76 chemical absorption 68 chemical extraction 200–201, 213–214 chemicals 11–12 – coproduction, and motor-gasoline refining 514–515 – extraction from Fischer–Tropsch syncrude 465 – from gasification liquids 11 – synthesis from Fischer–Tropsch syncrude 465–466 chemicals and lubricant refining 581 – Fischer–Tropsch-based lubricant base oil refining 597 – – group III lubricant refining 598–599 – – group IV lubricant refining 599–600 – – lubricant base oil refining 600–601 – Fischer–Tropsch-based petrochemical refining 591 – – alkane refining 591–592 – – aromatics refining 595–596 – – light alkene refining 592–594 – – linear 1-alkene refining 594–595 – – oxygenate refining 596 – lubricant base oils 584–585 – petrochemicals 582–584, 582, 584 – syncrude, chemicals refining concepts for – – alkene and oxygenate recovery 587, 587 – – alkane-based refining 585–586 – – aromatics production 586, 587 – – fuels and chemicals coproduction 588–591 chemical technologies 465 – autoxidation 474 – – Fischer–Tropsch wax oxidation 480–484 – – reaction chemistry 478–480 – – regimes 477–478 – – syncrude process technology 484–485 – n-1-alkenes, production of 466 – – distillate-range n-1-alkene extraction 474 – – 1-octene, extraction of 470–472 – – 1-octene production from 1-heptene 473 – – 1-pentene and 1-hexene, extraction of 467–470 Chevron Nigeria 241–242 Chevron Phillips Chemical company 450 Chevron Texaco Isocracking 244 chloride ions, in catalyst deactivation 89–90 chlorination 134, 134 circulating fluidized bed (CFB) reactor 157–158, 183 – design 220 civilian turbine fuels 270 classic Whitmore-type carbocation mechanism, in alkene oligomerization 372 clay-treater 163, 164, 177
Index Clean Air Act 39, 41 closed gas loop design 107, 108–109 cloud point (CP) 293 coal 154 – carbonization of 120 coal liquids 138, 404, 568 coal oil 32 coal pyrolysis product refining 169–170, 170, 177, 196–198, 197, 211–212 – tar workup section products 198 coal-to-liquids (CTLs) 503 cobalt (Co) – hydrogenolysis 80–81 – see also catalyst formulation cobalt-based low-temperature Fischer–Tropsch (Co-LTFT) 8, 9, 10, 10, 11, 233, 234–235, 238, 239, 243–244 – catalyst 121 – – deactivation of 95–99, 99 coker gas 120 coking reaction 54, 326 cold condensate 115 cold filter plugging point (CFPP) 293 combined homo- and heterogeneous syngas production 57 completion, shortest time to 507–508 compression-ignition engines 283, 284, 288, 290 condensate 114 – cold 115 – hot 115 – refining 202–203 condensate oil, refining 132–135 ‘‘confinement’’ model 452 conformational isomerization 353 continuous catalyst regeneration (CCR) 192 continuous regenerative units 449 continuous stirred tank reactor (CSTR) 62, 85, 173 contractual obligations 508 conventional catalytic naphtha reforming 444, 445 – catalysis 447–449 – reaction chemistry 444–447 – syncrude processing technology 449–450 conventional crude oil 22–28 – hydrocarbons in 23 – metals in 26–27, 27 – nitrogen compounds in 24, 24 – oxygenates in 25–26, 26 – physical properties of 27–28 – sulfur compounds in 23–24, 24 – see also crude oil conversion of olefins to distillates (COD) technology 222 – in alkene oligomerization 379
conversion process 33 co-refining – diesel fuels 567–569 – jet fuel 547–548 – and motor-gasoline refining 521–522 Co–ThO2 –kieselguhr catalyst – product yield, pressure effect on 125 Co–ThO2 –MgO–kieselguhr catalyst – distillate fractions, comparison 133 cracking 356, 407, 445 – catalytic cracking 421 – – catalysis 423–425 – – syncrude processing technology 425–427 – hydrocracking 427 – – catalysis 430–434 – – syncrude processing technology 434–436 – and isomerization 354 – rate of 415 – reaction chemistry 410 – – catalytic cracking 414–416 – – hydrocracking 416–419 – – thermal cracking 410–414 – of slack wax fraction 135 – thermal cracking 419 – – syncrude processing technology 421 creosote hydrotreater 197–198 crude oil 10, 492 – density of 27 – -derived diesel fuel 567 – -derived kerosene and Fischer–Tropsch kerosene, blending 548 – global consumption of 31 – price 16 – products from 28–31 – residue, classification of 28 – vs. syncrude, comparison 12–13, 13 – see also conventional crude oil crude oil refining/refineries 409 – conventional crude oil 22–28 – – products from 28–31 – evolution of 31–45 – – first generation 32–33 – – fourth generation 39–43, 40 – – lubricant base oil refineries 44–45 – – petrochemical refineries 43–44 – – second generation 33–35, 34, 36 – – third generation 36–39, 38 – fluid catalytic cracking 323 – naphtha reforming 321–322 – and oligomerization 315–316 – thermal cracking 324–325 cryogenic separation 106 cumene 258 cyclar process 458
607
608
Index cyclic units 449 cycloalkane(s) 273, 457, 547, 563 – in crude oil 23 – direct dehydrogenation of 445 – hydrocracking of 445 – hydroisomerization of 445 – octane numbers of 258 – synthesis 567 cyclohexane 457 cyclohexene 457 cyrogenic separation 110–111, 111
d Dakota Gasification Company 52 De Meern 243 decanted oil (DO) 113, 178 DEF STAN 91–91 specifications 270, 272, 279 dehydration 336–339 – and alkene dimerization 339 – catalysis 340–341 – reaction chemistry 339–340 – syncrude process technology 341–342 density–cetane–yield triangle 563–564, 569–570 desorption 76, 83 – vs. hydrogenation 80–81 devolatilization 59 diamond gas holdings 231 1,2-dibromoethane 264 diesel fuel 33, 41–42, 132, 133, 138, 148, 283 – additives affecting refinery design 295–296 – boiling range of 562 – future specification changes 296–297 – production 571 – properties 171, 208, 286 – – aromatic content 292 – – cetane number 286–290 – – cold-flow properties 293 – – density and viscosity 290 – – elastometer compatibility 294–295 – – flash point 290 – – lubricity 290–292 – – stability 294 – – sulfur content 292–293 – residue and wax, 565 – specifications 284–285, 560–561 – – American and European 284 – see also jet fuels diesel fuel refining 559 – application of lever rule to 570 – decisions affecting 564 – – aqueous product 565–566 – – C2 –C4 hydrocarbons 565 – – co-refining 567–569
– – density–cetane–yield triangle, dealing with 569–570 – – residue and wax 565 – – technology selection 566–567, 566 – gap analysis for syncrude to diesel fuel 560 – – carbon number distribution 561–562 – – composition and quality 562–563 – – density–cetane–yield triangle 563–564 – – diesel fuel specifications 560–561, 560 – heavy. see heavy diesel fuel – from HTFT syncrude 570–573, 571, 572 – from LTFT syncrude 573–578, 574, 575, 577 – light. see light diesel fuel – refinery design 574 dimerization 356, 445 Dimersol E unit 203 direct alcohol dehydration 336 disproportionation 446 distillate 33 distillate hydrotreater (DHT) 192, 193, 217, 225 distillate-range n-1-alkene extraction 474 distillate-selective cracker (DSC) 192, 194 distillation – atmospheric 44 – broadening 192 – fractions of crude oil 30 – profile – – and crude oil classification 22 – – of motor-gasoline 260 – vacuum 36, 38, 44 di-tert-butyl peroxide (DTBP) 475 double bond isomerization 353 driveability index (DI) 260 dry ash moving bed gasification 61 dry-feed entrained flow gasification 65 drying 58 Dual fuel 207
e E-factor 496 efficiency factor (EF) 496 electricity, export of 16 Emergency Petroleum Allocation Act 38 energy efficiency. see thermal efficiency entrained flow gasifiers 64–66 environmental footprint, smallest 507 E-point terminology 260 Escravos GTL facility 17 ester(s) 480 – -based lubricants 598 – -based mechanism, in alkene oligomerization 372–373 – in crude oil 26 ethanal 211
Index ethane 110, 545 ethanoic acid (acetic acid) 476 ethene 43, 203, 549, 588 – in aromatic alkylation 397–399 – in HTFT synthesis 393 etherification 289, 343–344 – catalysis 346 – reaction chemistry 345–346 – syncrude process technology 347 2-ethoxy-2-methylbutane (TAEE) 345 2-ethoxy-2-methylpropane (ETBE) 345 ethyl ethanoate 211 ethylbenzene 586 – production 399 – – catalysts used in 398 European Union (EU) fuel specifications existing gum 261 external recycle 108, 108 ExxonMobil 241
252
f fatty acid methyl esters (FAMEs) 297, 560–561, 568, 569 feed – cost 14–15 – description 497 – diversity 4 – flexibility 4 – selection 14, 15 feed-to-electricity conversion 16 feed-to-liquids (XTLs) conversion 3 – and power generation 16 feed-to-syngas conversion 4–8 – feed, logistics and preparation of 5, 5 – syngas cleaning and conditioning 7–8, 8 – syngas production 5–7 ferrierite, pentene skeletal isomerization 359 fifth column 176 final boiling point (FBP), of motor-gasoline 262 first generation crude oil refineries 32–33 first-generation intermediates 583 Fischer, Franz 73 Fischer–Tropsch distillate 559, 567–568 Fischer–Tropsch facilities 3–19 – feed-to-syngas conversion 4–8 – – feed, logistics and preparation of 5, 5 – – syngas cleaning and conditioning 7–8, 8 – – syngas production 5–7 – indirect liquefaction economics 14–19 – – capital cost 17–19, 19 – – feed cost 14–15 – – product pricing 15–16 – syncrude-to-product conversion 10–13 – – fuels versus chemicals 11–12
– – upgrading and refining 10–11 – syngas-to-syncrude conversion 8–10 Fischer–Tropsch refining/refinery 222, 303, 541 – alcohol dehydration in, applications of 336–339 – alkene oligomerization 369 – American Hydrocol facility 145–150 – – aqueous product 148–149 – – oil product 146–148 – aqueous product refining 225–226, 247 – Bintulu GTL – – aqueous product treatment 238 – – oil refining 235–238 – chemical addition 328 – cost of 304 – energy requirements 328 – evaluated technologies for 327–328 – German 128–137 – – aqueous product 136–137 – – C3 –C4 Crude LPG 129–130 – – carbon gasoline 130–132 – – condensate oil 132–135 – – waxes 135–136 – oil refining 222–225, 244–246 – oligomerization, applications of 315–316 – Sasol 1 facility 163–171, 163, 174–177, 175 – – aqueous product refining 166–169 – – Arge LTFT oil refining 165–166 – – coal pyrolysis product refining 169–170 – – Kellogg HTFT oil refining 163–165 – – synthetic fuel properties 170–171 – – tar workup 161 – Sasol 2 and 3 facilities 186–199 – – aqueous product refining 194–196 – – coal pyrolysis product refining 196–198 – – synthetic fuel properties 198 – – synthol HTFT condensate refining 188–192, 189 – – synthol HTFT oil refining 192–194 – syncrude compatibility 326 – synthetic fuel properties 227 – waste generation 326, 328 – see also industrial Fischer–Tropsch facilities Fischer–Tropsch syncrude 244, 256–258, 261, 263, 276, 278, 279, 286, 289, 290–292, 562 – oxygenate refining technologies 335 Fischer–Tropsch synthesis 73–99, 201–202, 220, 243–244 – American Hydrocol facility 143–145 – – syncrude composition 144 – Bintulu GTL 233–235 – catalyst deactivation 88–99 – – Co-LTFT catalysts, deactivation of 95–99 – – Fe-HTFT catalysts, deactivation of 93 – – Fe-LTFT catalysts, deactivation of 93–95 – – mechanical catalyst degradation 91
609
610
Index Fischer–Tropsch synthesis (contd.) – – metal carboxylate formation 90–91 – – poisoning by syngas contaminants 89–90 – – volatile metal carbonyl formation 90 – gas loop design 221–222 – in German facilities 121–128 – – carbon efficiency 128 – – gas loop design 127–128 – – medium-pressure synthesis 125–127 – – normal-pressure synthesis 122–124, 123 – mechanism 74–77 – – carbide 74 – – oxygenate 75, 76, 76 – product selectivity 77–81 – – chain growth probability 78–80 – – hydrogenation vs. desorption 80–81 – – manipulation 81–88 – – readsorption chemistry 81 – Sasol 1 facility 157–162, 173–174 – – Arge LTFT Synthesis 159–161 – – gas loop design 162 – – Kellogg HTFT synthesis 157–159 – Sasol 2 and 3 facilities 183–186 – – gas loop design 184–186, 185 – – syncrude composition 184 Fischer–Tropsch wax oxidation 480–484 fixed bed reactors 85 fluid catalytic cracking (FCC) 407, 422, 444, 517, 589 – in crude oil refining 323 – and thermal cracking 324–325 fluid hydroforming 444 fluidized bed gasifiers 62–63, 85, 142 fouling – of Co-LTFT catalysts 96 – of Fe-HTFT catalysts 93 fourth generation crude oil refineries 39–43, 40 free radical mechanism, in alkene oligomerization 373–374 freezing point 276 Friedel–Crafts alkylation 395 fuel ethers 262–263 – properties of 338 fuel(s) 11–12, 29, 32 – blending flexibility, in aromatic alkylation 394 – and chemicals coproduction 588–591 – quality 200 furans, in crude oil 26 fused-iron catalyst 153, 184
g gap analysis – for syncrude to diesel fuel 560 – – carbon number distribution 561–562 – – composition and quality 562–563
– – density–cetane–yield triangle 563–564 – – diesel fuel specifications 560–561 – for syncrude to jet fuel 541 – – carbon number distribution 542 – – composition and quality 542–544 – – jet fuel specifications 541–542 – for syncrude to motor-gasoline 510 – – carbon number distribution 510–511 – – composition and quality 511–514 – – motor-gasoline specifications 510 gas cleaning, natural gas 54, 55 gas hourly space velocity (GSHV) 122 gasification 6, 58 – classification of 6 – of heteroatoms 59–60 – high-temperature entrained flow gasification 64–66 – low-temperature moving bed gasification 60–62 – medium-temperature fluidized bed gasification 62–63 – technologies, selection/comparison 66 gas liquor 67 gas loop, Fischer–Tropsch 105–115 – configurations 107–109 – – closed gas loop design 107, 108–109 – – open gas loop design 107–108, 107 – design 69, 221–222 – – in German facilities 127–128 – – Sasol 1 facility 162, 162 – – Sasol 2 and 3 facilities 184–186, 185 – syncrude, cooling and separation 109–115 gas oil and heavier material 547 gaseous hydrocarbons 35 gas reforming 53, 57, 109 – technologies, comparison of 58 gas-to-liquids (GTLs) process 52, 142, 231 – Fischer–Tropsch refining, in Bintulu GTL – – aqueous product treatment 238 – – oil refining 235–238 – Fischer–Tropsch synthesis, in Bintulu GTL 233–235 – Oryx 241–248 – Pearl GTL facility 238 – syngas production in Bintulu GTL 232–233 generic Fischer–Tropsch refinery configuration 562 gasoline 32, 138 – aviation 33, 35 German Fischer–Tropsch facilities 119–138, 120 – Fischer–Tropsch refining 128–137 – – aqueous product 136–137 – – C3 –C4 Crude LPG 129–130 – – carbon gasoline 130–132 – – condensate oil 132–135 – – waxes 135–136
Index – refinery design 137–138 – syngas production 119–121, 121 – Fischer–Tropsch synthesis 121–128 – – carbon efficiency 128 – – gas loop design 127–128 – – medium-pressure synthesis 125–127 – – normal-pressure synthesis 122–124, 123 gums 261
h H2:CO ratio 5, 84, 87 Haag–Dessau mechanism 414 halogen-containing compounds, gasification of 60 heavier alkanes, hydroisomerization 360, 364 heavy diesel fuel 132 – properties 133 heavy end recovery stream 244 heavy paraffin synthesis (HPS) 234 heavy vacuum gas oil (HVGO) 193 1-heptene, extraction 204, 205, 467–470, 468 heptoxyheptane 338 heteroatoms – and crude oil classification 23 – gasification of 59–60 – removal of 4 heterogeneous syngas production 57 hexoxyhexane 338 high-temperature entrained flow gasification 64–66, 65 – dry-feed entrained flow gasification 65 – performance of 65 – slurry-feed entrained flow gasification 65–66 high-temperature Fischer–Tropsch (HTFT) process 143m 258, 262, 271, 272, 279, 289, 293–295, 304, 305, 544, 583, 586, 590 –592 – distillate 559 – product 109 – syncrude refining 339, 563 – – co-refining 569 – – diesel fuel refining from 570–573 – – jet fuel refining from 548–553 – – recovery 113–114, 113 high-temperature fluid catalytic cracker (FCC) 209 high-temperature gasification 6, 7 high-temperature water gas shift 69 high-vacuum distillation 176–177 Hoesch facility 136 homogeneous gasification 57 homogeneous nickel catalysts, in alkene oligomerization 383–384 homogeneous syngas production 57 homolytic bond dissociation 411 hot condensate 115 hot gas recycle process 142
hydration 129, 130, 347–349 – catalysis 349–350 – reaction chemistry 349–348 – syncrude process technology 350 hydride transfer 415 hydrocarbon autoxidation, reaction network of 480 hydrocarbon class, and crude oil classification 22 hydrocarbon compounds, cetane numbers of 288 Hydrocarbon Research Inc. 141 hydrocarbons 278 – in crude oil 23 – gaseous 35 – vs. oxygenates, boiling point difference 112 hydrocarbons, light 546 – yield of 593 Hydrocol process 141 hydrocracking 235–236, 244–246, 320–321, 408, 409, 416–419, 420, 427, 428, 446, 545, 547, 548, 565 – catalysis 430–434 – and hydroisomerization 360 – ideal 432 – nonideal 432 – syncrude processing technology 434–436 hydrodearomatization (HDA) 198, 305, 597 hydrodemetallization (HDM) 306 hydrodenitrogenation (HDN) 306 hydrodeoxygenation (HDO) 192, 224, 306, 307–308 hydrodesulfurization (HDS) 306 hydroformylation, alkene to alcohol conversion by 135, 466 hydrogen cyanide (HCN), in catalyst deactivation 90 hydrogenated motor-gasoline, alkene oligomerization 377 hydrogenation (HYD) 74–76, 189, 190, 190 – of alkenes 305 – vs. desorption 80–81 – wax 166 hydrogenolysis 80–81, 445 hydrogen sulfide (H2 S) 183 – in catalyst deactivation 89 – emissions, reducing 172 – removal of 67–69 hydroisomerization 319–320, 354, 360–362 – butane hydroisomerization catalysis 362 – C5 –C6 naphtha isomerization catalysis 362–363 – heavy alkane hydroisomerization catalysis 364 – syncrude process technology 364–366 – wax hydroisomerization catalysis 364 hydroperoxides 255, 294, 478 hydrotreater 236, 246 hydrotreating 169–170, 305–306 – alkenes, hydrogenation of 306–307 HZ-1 catalyst 191
611
612
Index H-ZSM-5, 457–458, 565 – aromatization over 454, 457 – -derived oligomer stream 571 H-ZSM-5 zeolite catalyst 224 – in alkene oligomerization 378–380
i ignition improvers 296 ignition quality tester (IQT) 287 income. see product pricing indirect alkylation 371 indirect liquefaction 3, 3 – economics 14–19 – – capital cost 17–19, 19 – – feed cost 14–15 – – product pricing 15–16 indole 401 industrial Fischer–Tropsch facilities 303 – see also Fischer–Tropsch refinery inerts, concentration of – in alkene oligomerization 375 integrated crude oil and Fischer–Tropsch refinery 568 internal recycle 108, 108, 128 iron (Fe) – hydrogenolysis 80–81 – see also catalyst formulation iron-based high-temperature Fischer–Tropsch (Fe-HTFT) synthesis 8, 9, 10, 141, 143–144 – catalysts 221 – – deactivation of 93 – industrial, selectivity change during 94 iron-based low-temperature Fischer–Tropsch (Fe-LTFT) synthesis 8, 9, 10 – catalysts, deactivation of 93–95, 96 isomerization – aromatic isomerization 354 – conformational isomerization 353 – double bond isomerization 353 – hydroisomerization 354 – skeletal isomerization 353–354 – stereochemical isomerization 353 isoparaffinic kerosene (IPK) 212, 272, 279, 542, 545
j Jet A-1 specifications 541, 542 jet fuel refining 541 – decisions affecting 544 – – C2 –C4 hydrocarbons, fate of 544–545 – – co-refining 547–548 – – residue and wax, fate of 545–546 – – technology selection 546–547, 546 – gap analysis for syncrude to jet fuel 541 – – carbon number distribution 542
– – composition and quality 542–544 – – jet fuel specifications 541–542 – from HTFT syncrude 548–553, 549, 551, 553 – from LTFT syncrude 553–557 jet fuels 269, 286, 574 – production 561 – properties 273–274 – – aromatic content and smoke point 276, 278 – – density and viscosity 275–276 – – elastometer compatibility and lubricity 279–280 – – freezing point temperature 276 – – net combustion heat 274–275 – – stability 278–279 – – sulfur and acid content 278 – – volatility 278 – specifications 270–271 – – fuel for military use 272–273 – – future changes 280 – – synthetic jet fuel 271–272, 274 – see also diesel fuel jet propulsion (JP) aviation turbine fuels, United States military 273 JGC Corporation 231 JP-8 military jet fuel, Fischer–Tropsch-derived 542
k KBR Superflex Selective Catalytic Cracking (SCC) technology 210 Kellogg HTFT oil refining 163–165, 164 – clay-treater 163, 164 – oligomerization 163, 164–165 Kellogg HTFT synthesis 157–159, 157 – compound classes 159 – gas loop 162 – product distribution 159 kerosene 32, 547 – crude-oil-derived 548 kerosene cut 225 kerosene range product – from SPA-catalyzed oligomerization 377 ketone(s) 77, 258, 588 – as primary autoxidation products 480 Keyes process 168 kogasin I 132 kogasin II 132 – conversion into lubricating oil 134 K¨olbel–Engelhardt-type synthesis 586
l lamp oil. see kerosene lead bromide 264 lead oxybromides 264 least refinery ‘‘complexity’’ light alcohols 339
507
Index light alkenes 43 – refining 592–594 light aromatics 43 light diesel fuel 132 – properties 133, 148 light oil 114 light vacuum gas oil (LVGO) 193 linear 1-alkene(s) 588 – extraction of 204–205 – refining 594–595 linear alkyl benzene (LAB) synthesis 567 linear hydrocarbons 562 liquefied natural gas (LNG) 218, 231 liquid absorption 68 – chemical absorption 68 – physical absorption 68 liquid petroleum gas (LPG) 35, 124, 443, 455, 555, 585 low aromatic distillate (LAD) range of products 227 low-complexity refinery design 509 low-temperature Fischer–Tropsch (LTFT) process 231, 235, 243, 244, 258, 262, 288, 290, 293, 294 – distillate 559 – jet fuel refinery design 574, 576, 578 – refinery 305 – syncrude 544, 583 – – diesel fuel refining from 573–578 – – jet fuel refining from 553 – – product 109 – – recovery 114–115, 115 – – refining 339 low-temperature gasification 6–7, 6 low-temperature moving bed gasification 60–62, 61 – dry ash moving bed gasification 61 – feed requirements and product composition, comparison 62 – slagging moving bed gasification 61–62 low-temperature water gas shift 69 lubricant base oil refining/refineries 44–45, 597 – group III lubricant refining 598–599 – group IV lubricant refining 599–600 – lubricant base oil refining 600–601 – solvent-based 45 lubricant base oils 29, 134, 584–585 lubricating oil 134 – properties 134 Lurgi dry ash coal gasification – Sasol 1 facility 154–155, 155 – – ash disposal 154 – Sasol 2 and 3 facility 182 Lurgi dry ash moving bed gasifier 61 Lurgi Mark I coal gasifiers 154 Lurgi Mark IV gasifier 182
Lurgi Mark V coal gasifiers 172 Lurgi Mark VI coal gasifiers 172
m magnesium oxide 121 maximum liquid product volume 507 maximum motor-gasoline refinery design 509 mechanical catalyst degradation 92 medium-pressure synthesis, German 125–127, 126 – product distribution 127 – RON, MON and density of naphtha 131 medium-temperature gasification 6, 7 – fluidized bed gasification 62–63, 63 – – performance data of 64 mercaptans. see thiols ‘‘mersol’’ detergents 135 metal carbonyl compounds formation, and catalyst deactivation 90, 90 metal carboxylates 228 – formation, and catalyst deactivation 91–92, 94 – – hydrogenation 92 – – thermal decomposition, onset of 91 metals – in crude oil 26–27, 27 – gasification of 60 – hydrogenolysis 80–81 metathesis, in alkene conversion 314 methanation 56 methane 110 – selectivity 78 – – and C2-C4 selectivity 98–99 2-methoxy-2-methylbutane (TAME) 345 2-methoxy-2-methylpropane (MTBE) 259, 263, 345 methoxymethane (dimethyl ether, DME) 289 methyl aryl ethers 258 methylcyclopentadienyl manganese tricarbonyl (MMT) 207, 252 methylene chain index 599, 599 methyl tertiary butyl ether (MTBE) 41, 520 MFI-type zeolite 565 M-forming processes 454, 455, 455 mineral rejection 4 Mitsubishi Corporation 231 mixed paraffins 585 mobil olefins to gasoline and distillate (MOGD) process 224 Mossgas facility 217 – discussion of refinery design 228–229 – Fischer–Tropsch refining 222 – – aqueous product refining 225–226 – – oil refining 222–225 – – synthetic fuel properties 227 – Fischer–Tropsch synthesis 220
613
614
Index Mossgas facility (contd.) – – gas loop design 221–222 – PetroSA facility, evolution of 227 – – changes in Fischer–Tropsch refinery 227–228 – – low-temperature Fischer–Tropsch synthesis, addition of 227 – syngas production 218 – – gas reforming 218–220 – – natural gas liquid recovery 218 motor gasoline (petrol) 33, 35, 145, 147, 150, 171, 213, 251, 283, 285 – components, properties of 42 – composition changes in 35, 36 – high-octane 39, 40 – leaded 207 – properties of 41, 148, 208 – – alkene content 261 – – aromatic content 262 – – aviation gasoline 264–265 – – density 259 – – fuel stability 261 – – metal content 263 – – octane number 253–259 – – oxygenate content 262–263 – – sulfur content 262 – – volatility 259–260 – RON, MON and density in Sasol 1 refinery 171 – specifications 252–253, 253, 512 – – future changes 265–266 motor-gasoline refining 509 – decisions affecting 514 – – alkane-based naphtha refining 518–519 – – aqueous product 517–518 – – C2 –C4 hydrocarbons 515–516 – – chemicals coproduction 514–515 – – co-refining 521–522 – – residue and wax 516–517 – – technology selection 519–521 – gap analysis for syncrude to motor-gasoline 510 – – carbon number distribution 510–511 – – composition and quality 511–514 – – motor-gasoline specifications 510 – from HTFT syncrude 522–529, 523, 524, 527, 528 – from LTFT syncrude 529–539, 530, 532, 534, 536, 537, 538 – and syncrude refinery designs 511 – technology map for 519 motor octane number (MON) 131, 171, 255, 287
n n-1-alkenes 513, 595 – applications of 466 – extraction 204 – production of 466
– – 1-octene, extraction of 470–472 – – 1-octene production from 1-heptene 473 – – 1-pentene and 1-hexene, extraction of 467–470 – – distillate-range n-1-alkene extraction 474 n-alkanes 585 – hydrocracking of 418 naphtha 547 – reforming 513 – – thermal 443–444 – see also gasoline naphtha hydrotreater (NHT) 192, 193, 224 naphthenic acids, in crude oil 26 natural gas 51–52, 231, 242–243 – composition 52 – logistics and pretreatment 5 – syngas production from 53–58, 54, 142–143, 144 – – adiabatic oxidative reforming 56–57 – – adiabatic prereforming 55–56 – – gas cleaning 54, 55 – – gas reforming technologies, comparison 57, 58 – – steam reforming 56 natural gas liquid (NGL) 51, 217, 232 – recovery 218 natural gasoline 147 n-butane 222 – aromatization of 459 – hydroisomerization 360 net present value (NPV), maximizing 507 neutral trialkyl phosphate esters 598 Ni/SiO2 –Al2 O3 catalyst 226 nickel (Ni) – -based reforming catalysts 55 – in crude oil 26 Nigerian National Petroleum Company 241 nitrogen – -based chemicals 11 – compounds – – in catalyst deactivation 90 – – in crude oil 24, 24 – gasification of 60 N-methyl-2-pyrrolidone (NMP) 469, 469 N-methyl pyrrolidone (NMP) extraction 204 nonacid chemicals, in aqueous products 166–167, 167 nonassociated gas 51 normal-pressure synthesis, German 122–124, 123, 138 – aliphatic hydrocarbons, composition 125 – product distribution 127 – RON, MON and density 131 – syncrude fractions, composition 124 nuclear magnetic resonance (NMR) spectrometry 288
Index
o octane number 33 1-octene, extraction 204–205, 470–472, 471, 472, 473 octoxyoctane 338 oil circulation process 142 oil crisis 36–37, 37, 181 oil refining 146–148, 205–210, 222–225 olefinic motor-gasoline, from phosphoric acid-catalyzed oligomerization 376 olefinic naphtha 224 oligomerization 129–130, 134, 134, 147, 150, 163, 164–165, 189–190, 190, 202, 203, 565 – acid-catalyzed 388 – in alkene conversion 315–316 – and isomerization 354 – technology, selection 567 open gas loop design 107–108, 107 operating conditions, Fischer–Tropsch – and product selectivity manipulation 83–84, 83 – – catalyst activation 84 – – pressure 84 – – space velocity 84 – – syngas composition 84 – – temperature 83–84 organic acids, in crude oil 26 organic sulfur, in catalyst deactivation 89 Organization of Petroleum Exporting Countries (OPEC) 37 organometallic insertion mechanism, in alkene oligomerization 373 Oryx and Escravos gas-to-liquids 241, 242 – Fischer–Tropsch refining in – – aqueous product treatment 247 – – oil refining 244–246 – refinery design discussion 247–248 – syngas production in 242–243 – Fischer–Tropsch synthesis 243–244 Oryx GTL facility 17 oxidation 59 – atmospheric 136 – and Co-LTFT catalyst deactivation 97–98, 98 – and Fe-HTFT catalyst deactivation 93 – and Fe-LTFT catalyst deactivation 94–95, 94 – stability 261 ‘‘OXO’’ process 135 oxygen, addition and removal of – autoxidation 311–312 – carbonyl aromatization 310, 311 – dehydration 308–309 – esterification of 310 – etherification 309 – hydration of 309 – hydroformylation 311
oxygenate mechanism 76, 76 – for chain growth 75 oxygenate partitioning 111–113, 111 oxygenate refining 335, 596 oxygenated gasoline 41 oxygenates 138, 149, 459, 544, 587, 563 – in alkene oligomerization 375 – in crude oil 25–26, 26 – vs. hydrocarbons, boiling point difference – octane numbers of 258–259 – readsorption 76–77 – in SPA oligomerization 378
112
p P. C. Keith of Hydrocarbon Research Inc. 153 paraffins 176 – reformer plant 166 – waxy 135 Paraformer 166 partial oxidation (POX) processes 57, 58 partial refining 11 particle size distribution 66 particulate matter (PM) emissions 286, 289, 292 ‘‘peak oil’’ 37 Pearl GTL project 232, 238 1-pentene, extraction of 467–470 pentene isomerization catalysis 359–360 pentene skeletal isomerization 206 pentoxypentane 338 per pass conversion 105, 106 Perco-process 147 permanent gases 11 petrochemical refining/refineries 43–44, 591 – alkane refining 591–592 – aromatics refining 595–596 – light alkene refining 592–594 – linear 1-alkene refining 594–595 – oxygenate refining 596 – stand-alone 44 petrochemicals 29, 582–584, 582, 584 Petroleum Oil and Gas Corporation of South Africa (PetroSA) 217 Petronas 231 PetroSA facility, evolution of 227 – changes in Fischer–Tropsch refinery 227–228 – low-temperature Fischer–Tropsch synthesis, addition of 227 phenolic compounds 132 phenols 278–279 – in crude oil 26 Phenosolvan process 169, 196 Philips Petroleum Company 141 phosphoric acid oligomerization 130, 147, 150, 552 physical absorption 68
615
616
Index platforming 444 platinum-based catalysts 447 plug flow reactors (PFRs) 85 poisoning by syngas contaminants, and catalyst deactivation 89–90 polyalphaolefin (PAO) 585 Polyhydrotreater 190, 207 polyinternalolefin (PIO) 585 Polymer Corporation 164 polymer gasoline 147, 170, 370 potential gum 261 pour points of crude oil 27–28 powered flight 33 precipitated-iron catalysts 159, 160 prereforming 54, 243 – adiabatic 55–56 pressure separation 110 pressure swing absorption (PSA) 243 primary design objective 501 product description 497 product fractionation, after alkene oligomerization 388 product pricing 15–16 product selectivity, Fischer–Tropsch synthesis 77–81 – chain growth probability 78–80 – hydrogenation vs. desorption 80–81 – manipulation 81–88 – – catalyst formulation 81–83 – – operating conditions 83–84 – – reaction engineering 84–88 – readsorption chemistry 81 products – from crude oil 28–31 – – quality and boiling range 29–31 – recovery 106 – – see also syncrude, cooling and separation – of refining 21–22 Project Turbo 207 propane, aromatization of 459 propanoic acid (propionic acid) 476 propene 43, 202, 544, 588 – in aromatic alkylation 399–401 proton transfer 415 Pt/L-zeolite naphtha reforming, monofunctional nonacidic 450 – catalysis 452–453 – reaction chemistry 451–452 – syncrude processing technology 453–454 pyridine 401 – in crude oil 25 pyrolysis 7, 58–59 pyrroles, in crude oil 25
q Qatar Petroleum 241
r raw materials – crude oil refining 21 – for syngas production 51–53 reaction engineering, Fischer–Tropsch – and product selectivity manipulation 84–88 – – catalyst mechanical strength 86 – – catalyst replacement 86 – – catalyst–product separation 87 – – heat transfer 86 – – mass transfer 86 – – per pass conversion 87 – – reaction phase 85–86 – – reactor configuration 87–88 – – reactor type 85, 85 – – scale-up issues 88 – – syngas production upsets 88 reaction mechanism, in alkene oligomerization 375 reaction water 114, 115 reactivity, gasification 66 readsorption – chemistry 81 – of oxygenates 76–77 real-world refinery design 500 – refinery feed selection 502–503 – refinery location 503–506, 504 – – climate 504 – – environmental sensitivity 504–505 – – geology 504 – – intellectual property 506 – – legislation 505 – – location factor 505 – – marketing logistics 505–506 – – natural resources 504 – – politics and governance 505 – – utility access 505 – refinery products and markets 501–502 – refinery type 501 – secondary design objectives 506–508 rectisol naphtha 156 rectisol syngas cleaning 155–156 – fine wash 156 – main wash 156 – prewash 156 refinery benzene levels – in aromatic alkylation 393–394 refinery design 491, 569 – American Hydrocol facility 150–151 – in Bintulu GTL 239 – concepts – – characteristic of refining business 491–493
Index – – complex systems and design rules 493–495 – – refining complexity 495–496 – – refining efficiency 496 – conceptual 497 – – carbon-number-based design 499–500 – – hierarchical design 498 – – linear programming 497–498 – – technology preselection 498–499 – German facility 137–138 – Mossgas facility 228–229 – Oryx 247–248 – principles of 492 – for producing motor-gasoline and jet fuel 559 – real-world refinery design 500 – – refinery feed selection 502–503 – – refinery location 503–506 – – refinery products and markets 501–502 – – refinery type 501 – – secondary design objectives 506–508 – robust and flexible 509 – Sasol 1 facility 177–178 – Sasol 2 and 3 facilities 212–214 – strategy 576 refinery flexibility 507 refining/refineries 21–45, 497 – components of 21, 22 – crude oil vs. syncrude 12–13, 13 – partial 11 – process 21 – stand-alone 11 – see also crude oil refining/refineries reforming 59, 441 – aromatization 454 – – catalysis 457–459 – – reaction chemistry 456 – – syncrude processing technology 460–461 – conventional catalytic naphtha reforming 444 – – catalysis 447–449 – – reaction chemistry 444–447 – – syncrude processing technology 449–450 – and isomerization 354 – monofunctional nonacidic Pt/L-zeolite naphtha reforming 450 – – catalysis 452–453 – – reaction chemistry 451–452 – – syncrude processing technology 453–454 – thermal naphtha reforming 443–444 reformulated gasoline (RFG) 41 rehydrogenation, of Co-LTFT catalysts 97–98, 98 Reid vapor pressure (RVP) 260, 273 rejuvenation 123–124 religious engineering 494 research octane number (RON) 131, 255–257, 264
residue and wax – diesel fuel and 565 – jet fuel and 545–546 – and motor-gasoline refining 516–517 – see also wax(es) residue conversion – catalytic cracking 323–324 – coking 326 – Fischer–Tropsch refining technology selection 326–327 – thermal cracking 324–325 – visbreaking 324 residue oil 197 residue upgrading capacity 36, 38 retrograde condensation reactions 59 Ribblett ratio 87 Roelen, Otto 121 RZ-Platforming technology from UOP 450
s Sarawak State Government 231 Sasol 1 facility 153–178 – evolution of 172–177 – – coal pyrolysis product refining, changes in 177 – – Fischer–Tropsch refining, changes in 174–177 175 – – Fischer–Tropsch synthesis, changes in 173–174 – – syngas production, changes in 172 – Fischer–Tropsch refining 163–171, 163 – – aqueous product refining 166–169 – – Arge LTFT oil refining 165–166 – – coal pyrolysis product refining 169–170 – – Kellogg HTFT oil refining 163–165 – – synthetic fuel properties 170–171 – Fischer–Tropsch synthesis 157–162 – – Arge LTFT Synthesis 159–161 – – gas loop design 162 – – Kellogg HTFT synthesis 157–159 – refinery design 177–178 – syngas production 154–156 – – Lurgi dry ash coal gasification 154–155 – – rectisol syngas cleaning 155–156 Sasol 2 and 3 facilities 181–214 – Fischer–Tropsch refining 186–199 – – aqueous product refining 194–196 – – coal pyrolysis product refining 196–198 – – synthetic fuel properties 198 – – synthol HTFT condensate refining 188–192, 189 – – synthol HTFT oil refining 192–194 – Fischer–Tropsch synthesis 183–186 – – gas loop design 184–186, 185 – – syncrude composition 184 – refinery design 212–214 – Sasol synfuels, evolution of 199–212
617
618
Index Sasol 2 and 3 facilities (contd.) – – coal pyrolysis product refining, changes in 211–212 – – Fischer–Tropsch aqueous product refining, changes in 210–211 – – Fischer–Tropsch condensate refining, changes in 202–203 – – Fischer–Tropsch oil refining, changes in 205–210 – – Fischer–Tropsch synthesis, changes in 201–202 – – linear 1-alkenes, extraction of 204–205 – – syngas production, changes in 201 – – synthetic jet fuel 212 – syngas production 182–183, 183 – – Lurgi dry ash coal gasification 182 – – syngas cleaning 182–183 Sasol 2 expansion project 181 Sasol Advanced Synthol (SAS) 173, 220 – fixed fluidized bed reactor 201, 201 Sasol Clean Air Technology (SCAT) 172 Sasol East 199 Sasol Slurry Bed Process (SSBP) 173 Sasol Synfuels 261, 262 – evolution of 199–212 – – coal pyrolysis product refining, changes in 211–212 – – condensate and oil workup sections 206, 209 – – Fischer–Tropsch aqueous product refining, changes in 210–211 – – Fischer–Tropsch condensate refining, changes in 202–203 – – Fischer–Tropsch oil refining, changes in 205–210 – – Fischer–Tropsch synthesis, changes in 201–202 – – linear 1-alkenes, extraction of 204–205 – – syngas production, changes in 201 – – synthetic jet fuel 212 – reactors 183 Sasol West 199 Schulz–Flory equation 78 sec-alcohols 77 second generation crude oil refineries 33–35, 34, 34, 36 Second World War – high octane aviation gasoline, demand for 35 secondary pyrolysis 59 Secunda 181 semiregenerative units 448 semisynthetic jet fuel 279 sensitivity, of octane number 255 Shell Gas BV 231 Shell gasification process (SGP) 232–233 Shell MDS (Malaysia) Snd Bhd company 231
Shell Middle Distillate Synthesis (SMDS) process 11, 231, 241 – Fischer–Tropsch refining, in Bintulu GTL – – aqueous product treatment 238 – – oil refining 235–238 – Fischer–Tropsch synthesis, in Bintulu GTL 233–235 – Pearl GTL facility 238 – refinery design 239 – syngas production in Bintulu GTL 232–233 Shell Research and Technology Centre 231 Shell’s Pearl GTL facility 17 short path distillation (SPD) 177, 238 single-carbon-number hydroisomerization 365 skeletal isomerization 314–315, 353–354, 357–358 – butene isomerization catalysis 358–359 – hydroisomerization 319–320 – pentene isomerization catalysis 359–360 – syncrude process technology 360 slack wax 132, 135–136, 177 slagging moving bed gasification 61–62 slurry bed reactors 85, 142, 173, 173 – composition 174 – compound classes 174 slurry-feed entrained flow gasification 65–66 slurry-phase distillate (SPD) process, Sasol 243 smoke point 276 solid carbon sources 52–53 – composition 53 – syngas production from 58–66 – – gasification technologies, comparison 66 – – heteroatoms, gasification of 59–60 – – high-temperature entrained flow gasification 64–66 – – low-temperature moving bed gasification 60–62 – – medium-temperature fluidized bed gasification 62–63 solid phosphoric acid (SPA) catalyst 190 – in alkene oligomerization 375–378 – key aspects of 376–377 ‘‘sour’’ crude 23 South African Coal, Oil and Gas Corporation 153 sp2 - and sp3 -hybridized carbon coupling, for chain growth 75–76 spark-ignition engine 251, 253, 264 spindle oil 134 stabilized light oil (SLO) 114, 222 stabilized oil 147 stand-alone Fischer–Tropsch-based petrochemical facility 594, 596, 600 stand-alone refining 11 Standolind Oil and Gas Company 141 steam reforming 6, 53, 56, 58 – advantages of 56
Index stereochemical isomerization 353 stoichiometric ratio 87 storage stability 261, 278 Stretford process 182 substitute natural gas (SNG) 545 sulfides, in crude oil 24 Sulfolin process 182 sulfur 7 – -based chemicals 11 – gasification of 59–60 – compounds – – in catalyst deactivation 89, 89 – – in crude oil 23–24, 24, 24 sulfur dioxide (SO2), in catalyst deactivation 89 surface active compounds 112 ‘‘sweet’’ crude 23 syncrude – chemicals refining concepts for – – alkane-based refining 585–586 – – alkene and oxygenate recovery 587, 587 – – aromatics production 586, 587 – – fuels and chemicals coproduction 588–591 – cooling and separation 109–115 – – cyrogenic separation 110–111, 111 – – design 106 – – HTFT syncrude recovery 113–114 – – LTFT syncrude recovery 114–115 – – oxygenate partitioning 111–113, 111 – – pressure separation 110 – vs. crude oil, comparison 12–13, 13 – process technology 484–485 – recovery 138 syncrude-to-product conversion 10–13 – fuels versus chemicals 11–12 – upgrading and refining 10–11 syngas – cleaning 7–8, 8, 51, 66–69, 67, 120 – – acid gas removal 67–69 – – Sasol 2 and 3 facility 182–183 – conditioning 7–8, 8, 51, 69–70, 105 – – water gas shift (WGS) conversion 52–53, 54, 69–70 – unconverted 105–106 syngas production 5–7, 51, 201 – American Hydrocol facility 142–143 – in Bintulu GTL 232–233 – gas reforming 218–220 – German Fischer–Tropsch facilities 119–121, 121 – from natural gas 53–58, 54 – – adiabatic oxidative reforming 56–57 – – adiabatic prereforming 55–56 – – gas cleaning 54, 55 – – gas reforming technologies, comparison 57, 58 – – steam reforming 56
– – – – – – – – – – – – –
natural gas liquid recovery 218 Oryx 242–243 raw materials 51–53 Sasol 1 facility 154–156, 172 – Lurgi dry ash coal gasification 154–155, 155 – rectisol syngas cleaning 155–156 Sasol 2 and 3 facilities 182–183, 183 – Lurgi dry ash coal gasification 182 – syngas cleaning 182–183 from solid carbon sources 58–66 – gasification technologies, comparison 66 – heteroatoms, gasification of 59–60 – high-temperature entrained flow gasification 64–66 – – low-temperature moving bed gasification 60–62 – – medium-temperature fluidized bed gasification 62–63 syngas-to-syncrude conversion 8–10 synthesis gas. see syngas synthetic fuel, properties 170–171, 198, 199, 227 synthetic jet fuel 212, 271–272, 274 Synthetic Liquids Fuels Act 141 synthetic lubricants 585 synthetic natural gas (SNG) 52 synthol HTFT condensate refining 188–192, 189 – naphtha and distillate fractions, fuel properties 191 synthol HTFT oil refining 192–194 synthol reactors 157 syntroleum S-5 synthetic jet fuel 542
t tail gas 107, 114, 115, 144, 186 – processing 222 – recycle systems 108, 108 tail pipe catalytic converter 39 tar 67 Tar Naphtha Phenol Extraction (TPNE) 177 technology selection – for diesel fuel refining 566–567, 566 – for jet fuel refining 546–547, 546 – and motor-gasoline refining 519–521 tetraethyl lead (TEL) 35, 39, 198, 252, 256, 263, 264 thermal cracking 134, 134, 166, 324–325, 407, 410–414, 411, 419, 546, 565 – syncrude processing technology 421 thermal efficiency 14–15 thermal naphtha reforming 443–444 thermal oligomerization, in alkene oligomerization 384–385 thermal reforming 210 thermal stability 278 thiols, in crude oil 24 thiophenes, in crude oil 24 thiophenol 401
619
620
Index third generation crude oil refineries 36–39, 38 total acid number (TAN) 23 T-point terminology 260 Tropsch, Hans 73 true-vapor-phase (T-V-P) process 131–132 turbine engines, in aviation 35
u United States, Fischer–Tropsch facility in. see American Hydrocol facility Universal Oil Products (UOP) 131 – Catalytic Polymerization process 164, 189 – Penex technology 225 – Platforming technology 192 unstabilized light oil (ULO) 114 upgrading 11 US Bureau of Mines 141
v vacuum distillation 36, 38, 44 valve seat recession (VSR) 263 vanadium 183 – in crude oil 26 vapor pressure – of crude oil 27 – of fuel 260 Viktor-plant 129 visbreaking 324
viscosity index (VI) 134, 584, 598, 599 viscosity of crude oil 28
w Waksol A 176 Waksol B 176 waste-to-liquids (WTLs) process 3 water, alkene hydration 346 water gas shift (WGS) catalyst 120 water gas shift (WGS) reaction 52–53, 54, 69–70 water-soluble oxygenates 136, 149 – composition of 149 wax(es) 115, 585, 588 – hydrogenation 166 – hydroisomerization 360 – refining 135–136 – properties 136 wax hydroisomerization catalysis 364 wax hydrotreating and wax hydrocracking 553 waxy oil 194 waxy raffinate 236 World Wide Fuel Charter (WWFC) guidelines 252, 254, 285
z zeolite catalysts 341 zeolites, in coprocessing zinc oxide (ZnO) 55
404