High Temperature Corrosion and Materials Applications

Name ///sr-nova/Dclabs_wip/High Temp/5208_1.pdf/Chap_01/ 26/10/2007 1:51PM Plate # 0 High-Temperature Corrosion And Ma...

Author: George Y. Lai

88 downloads 1310 Views 18MB 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

DOWNLOAD PDF

Name ///sr-nova/Dclabs_wip/High Temp/5208_1.pdf/Chap_01/

26/10/2007 1:51PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p1 DOI: 10.1361/hcma2007p001

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 1

Introduction METALS AND ALLOYS will react during high-temperature service with the surrounding environment, resulting in high-temperature corrosion. In gaseous environments, high-temperature corrosion is deﬁned as the corrosion that takes place above the maximum temperature at which acids condense and dew-point corrosion takes place. Although a majority of hightemperature corrosion reactions take place at temperatures above 500 °C (930 °F), severe high-temperature corrosion has been encountered in many cases at temperatures below 500 °C (930 °F). In waste-to-energy boilers, for example, carbon and low-alloy steels have experienced severe ﬁreside corrosion problems in the waterwalls of the boilers at the tube metal temperatures of approximately 260 to 315 °C (500 to 600 °F). This book is intended primarily for engineers and metallurgists who are concerned with hightemperature materials problems in the following industries: aerospace/gas turbine, chemical processing, reﬁning and petrochemical, fossil-ﬁred power generation, coal gasiﬁcation, waste-toenergy industry, pulp and paper, heat treating, mineral and metallurgical processing, and others. The technical data presented in this book are pertinent to “real” materials problems related to the aforementioned industries. The book will also be useful for both undergraduate and graduate students who are interested in studying or pursuing research on the subject of hightemperature corrosion. The book covers eight basic modes of hightemperature corrosion. A brief description of thermodynamics is included for most chapters to help readers to understand the corrosion reactions. The external stresses (or strains) can cause alloys to suffer preferential corrosion penetration attack in a certain corrosive environment, such as sulﬁdizing environments. In addition, external stresses or residual stresses can cause the alloy to suffer brittle, intergranular cracking when exposed to the lower end of the intermediate temperatures for certain alloys. This type of cracking is frequently referred to as “reheat

cracking,” “stress-relaxation cracking,” or “strain-age cracking” (for nickel-base alloys). Both of these subjects are covered in Chapter 14, “Stress-Assisted Corrosion and Cracking.” The subject of erosion and erosion/corrosion is also reviewed with an attempt to offer readers general guidance on materials selection and application. Discussion also includes hydrogen attack of carbon steels in boilers and reﬁnery equipment. Finally, extensive discussion on the materials problems in coal-ﬁred boilers, oil-ﬁred boilers, waste-to-energy boilers, and black liquor recovery boilers is included. In summary, the subjects covered extensively in this book include:

Oxidation Nitridation Carburization and metal dusting Corrosion by halogen and hydrogen halides Sulﬁdation Hot corrosion Molten salt corrosion Liquid metal corrosion and embrittlement Erosion and erosion/corrosion Stress-assisted corrosion and cracking Hydrogen attack Coal-ﬁred boilers Oil-ﬁred boilers and furnaces Waste-to-energy boilers and waste incinerators Black-liquor recovery boilers

The focus of this book is on commercial alloys, including both generic and proprietary alloys. Most data are presented to reveal alloy ranking and thus serve as a general guide to materials selection and application. Engineers can thus use the data and information to compare alloys that are commercially available. The effects of alloying elements, temperature, and environmental conditions on the corrosion behavior of alloys are also discussed, providing information about the capability of an alloy in terms of useful temperature limitation. Trademarks for alloys and alloy manufacturers are listed in Appendix 1. The compositions of alloys are tabulated in Appendix 2.

pg 1

Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/

26/10/2007 12:10PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p3-4 DOI: 10.1361/hcma2007p003

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 2

Challenges in Materials Applications for High-Temperature Service IN MANY INDUSTRIAL SYSTEMS, plant operating conditions can be quite complex; it is rather difﬁcult to use laboratory tests to simulate plant conditions. However, laboratory tests can provide good general guidance for making preliminary alloy selection. In situ ﬁeld testing or ﬁeld trials of candidate alloys in the operating plant provides the best way of obtaining the corrosion information that can be reliably used for ﬁnal materials selection. During the preliminary alloy selection process, it is important to evaluate not only the high-temperature corrosion resistance of the alloy, but also its mechanical properties such as tensile and creep-rupture strengths. The microstructural changes at the application temperatures such as thermal stability of the alloy, should also be considered. For example, duplex stainless steels are known to suffer 475 °C (885 °F) embrittlement caused by the formation of alpha prime (α0 ) coherent precipitates. Accordingly, these stainless steels should be avoided for use as a structural component at temperatures approximately, above 340 °C (650 °F). ASME Codes may have lower maximum service temperature limits for these alloys. Consideration should also be given to fabrication issues, such as weldability and welding procedures, annealing heat treatments, postweld heat treatment (PWHT) and stress relieving, and codes and standards requirements. The availability of the alloy can also be an issue. It is not uncommon to ﬁnd that some alloys are no longer commercially available in stock due to a number of reasons, which may include poor market demands in the past, difﬁculty in manufacturing, and so forth. In some cases, the alloy may only be available on order for a whole production heat, which can be tens of thousands of pounds of material. Another important factor is the alloy price. A cost

analysis needs to be conducted to balance the material cost with the expected life for the component to ensure the alloy is cost effective. Often the life-cycle cost is a better criterion than the initial material cost in making an alloy selection. Selection of an appropriate ﬁller metal for welding is important for component fabrication involving welding. Normally, it is a simple process when the candidate alloy has a ﬁller metal with matching chemical composition. However, many high-temperature alloys do not have ﬁller metals with matching chemistries. The widely used Fe-Ni-Cr alloy 800H is a good example. Many heat-resistant cast alloys also do not have matching ﬁller metals. Thus, when no matching chemistry ﬁller metal is available for welding, it is critical to select a ﬁller metal that not only possesses excellent weldability but, also exhibits comparable or better high-temperature corrosion resistance along with comparable strengths, thermal stability, and other relevant properties. Some fabricators sometimes use weldability to select a ﬁller metal without considering the resistance of the weld metal to the speciﬁc hightemperature corrosive environment in the end application. This can lead to premature failures. For example, because of their good weldability, high nickel ﬁller metals, such as ﬁller metal alloy 82 (ERNiCr-3), are sometimes used for welding the alloys that are to be in service in sulﬁdizing environments. This can cause preferential sulﬁdation attack at the weld joint because of the relatively poor sulﬁdation resistance of high nickel alloys. Welding can still be an issue for some hightemperature alloys even with matching ﬁller metals. This is because some high-temperature alloys contain many alloying elements for various metallurgical reasons, such as improving the resistance to a certain mode of high-temperature

pg 3

Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/

26/10/2007 12:10PM Plate # 0

4 / High-Temperature Corrosion and Materials Applications

corrosion, increasing tensile and creep-rupture strengths, or increasing wear resistance. Increasing the levels of some of these alloying elements can increase the difﬁculty in the weldability of the alloy. For example, an alloy containing high silicon, high aluminum, high carbon, or very high chromium can be difﬁcult to weld even though a matching ﬁller metal is available. For construction of a component, engineers have the option to consider whether a wrought alloy or a cast alloy will be more suitable metallurgically and/or economically for the intended high-temperature application. Engineers may also consider a totally different approach to address the high-temperature corrosion issue for some existing plant equipment that has suffered corrosion. In reﬁneries, many reactor vessels, such as crude towers, hydrocrackers, and hydrodesulfurizers, are made of clad plates with a corrosion-resistant cladding in original installations. Cladding can be corroded after years of operation. One common approach is to refurbish the corroded vessels by applying a corrosion-resistant weld overlay instead of replacing it with a new construction. This approach has been adopted in the boiler industry in recent years to address the severe corrosion problems with the waterwalls of boilers in waste-to-energy boilers, coal-ﬁred boilers, basic oxygen furnace hoods in steel mills, and so forth. With automatic controls for gas metal arc welding machines, a large scale of weld overlay can be applied in vessels or boilers with engineering quality. Laser cladding can also be applied in the shop on large equipment such as waterwall panels. Coextruded composite tubes with a corrosion-resistant alloy cladding on the outer

diameter have long been available for construction of waterwalls as well as superheaters in boilers. Composite tubes manufactured by a spiral weld overlaying process have been made available in recent years. These composite tubes use the outer diameter cladding for providing corrosion protection and the substrate base tube for the load-bearing structural part. Most of these composite tubes are used in superheaters and reheaters in boilers with metal temperatures being likely less than about 650 °C (1200 °F). Many furnace tubes used in petrochemical processing, such as ethylene cracking furnace tubes, are exposed to temperatures higher than 980 °C (1800 °F) and carburizing gas streams on the internal diameter (ID) of the tube, application of composite tubes with a carburization- and coking-resistant alloy cladding on the tube ID can potentially increase the operating temperature and/or prolong the tube life. Aluminide coatings reportedly have been used in ethylene cracking furnace tubes. At the writing of this book, it appears no commercial companies in the United States provide aluminizing coating services for ethylene furnace tubes or pipes. Another diffusion coating, chromized coating, has also reportedly been used in boilers. Both of these diffusion coatings are very thin. Coatings have been highly successful in providing protection against oxidation and hot corrosion for the high-temperature components, such as airfoils, in gas turbines. The coatings used involve aluminide coatings, overlay MCrAlY coatings by vapor deposition processes (e.g., electron beam physical vapor deposition), and ceramic thermal barrier coatings (e.g., stabilized ZrO2). Coatings are considered sacriﬁcial and are to be replaced periodically.

pg 4

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p5-66 DOI: 10.1361/hcma2007p005

31/10/2007 12:43PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 3

Oxidation 3.1 Introduction Oxidation is the most important hightemperature corrosion reaction. Metals or alloys are oxidized when heated to elevated temperatures in air or in highly oxidizing environments, such as combustion atmospheres with excess air or oxygen. Many metallic components, such as retorts in heat treat furnaces, furnace heater tubes and coils in chemical and petrochemical plants, waterwalls and superheater/reheater tubes in boilers, and combustors and transition ducts in gas turbines, are subject to oxidation. For many industrial processes, combustion involves relatively “clean” fuels such as natural gas or No. 1 or No. 2 fuel oil. These fuels generally have low concentrations of contaminants, such as sulfur, chlorine, alkali metals, and vanadium. In many cases, excess air is used to ensure complete combustion of the fuel. The combustion products thus consist primarily of O2, N2, CO2, and H2O. Although alloys in these environments are oxidized by oxygen, other combustion products, such as H2O, may play an important role in affecting the oxidation behavior of the alloy. The presence of N2 in the combustion gas stream can cause signiﬁcant internal nitridation attack under certain conditions, which is discussed in Chapter 4 “Nitridation.” Oxidation can also take place in a “reducing” environment (i.e., the environment with a low oxygen potential created by the combustion under a substoichiometric condition). When combustion takes place under stoichiometric or substoichiometric conditions, the resultant environment becomes “reducing.” This type of environment is generally characterized by low oxygen potentials. Under this condition, the oxygen potential of the environment is typically controlled by pH2 =pH2 O or pCO =pCO2 ratio, and the oxidation kinetic is generally slow. The development of a protective oxide scale can be

sluggish for most alloys. As a result, the effects of corrosive contaminants can become more pronounced, resulting in other modes of hightemperature corrosion. For example, if the sulfur level in the environment is high, sulﬁdation then becomes the predominant mode of corrosion, even though oxidation also takes part in the corrosion reaction. Thus, a majority of hightemperature corrosion problems in reducing environments are caused by modes of corrosion attack other than oxidation. Most industrial environments have sufﬁcient oxygen activities (or potentials) to allow oxidation to participate in the high-temperature corrosion reaction regardless of the predominant mode of corrosion. In fact, the alloy often relies on the oxidation reaction to develop a protective oxide scale to resist corrosion attack, such as sulﬁdation, carburization, hot corrosion, and so forth. The oxidation behavior under these conditions is discussed in other chapters dealing with different modes of corrosion attack. There is a large spectrum of engineering alloys available for applications in different temperature ranges. This chapter presents a large oxidation database for a wide spectrum of engineering alloys, ranging from carbon and Cr-Mo steels, which serve the low end of the temperature spectrum, to superalloys serving the highest temperature regime. The data are organized by alloy groups to help readers to compare alloys within the same alloy group and also to compare alloys between different alloy groups. The focus is to present comparison data, thus allowing readers to consider candidate alloys for applications in the temperature regime of interest. Also included are some important metallurgical and environmental factors that can affect the oxidation behavior of the alloy. It should be noted that a majority of the database has been generated in laboratory tests that were conducted at temperatures higher than those at which the tested alloys would normally be

pg 5

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

pg 6

6 / High-Temperature Corrosion and Materials Applications

used. The intent for this approach was to determine the oxidation behavior of alloys within relatively short test durations. For example, many tests have been conducted at 980 to 1200 °C (1800 to 2200 °F) for stainless steels, Fe-Ni-Cr alloys, and Ni-Cr alloys. This may result in the alloy performance ranking based more on scaling (metal loss) than on internal oxidation attack which the alloys would most likely have encountered at lower application temperatures. An example is given in Fig. 3.1, which illustrates an actual ﬁeld experience with a furnace heater coil made of a Ni-Cr alloy that had been in service for about 4 to 5 years at temperatures less than 900 °C (1650 °F), suffering extensive internal oxidation attack with very little scaling (metal loss) (Ref 1). Few long-term tests have been conducted at temperatures of 650 to 980 °C (1200 to 1800 °F) where most stainless steels, Fe-Ni-Cr alloys, and Ni-Cr alloys are used in high-temperature applications. Furthermore, many test results were presented as weight changes instead of actual measurements of the damage to the metal, such as the total depth of oxidation attack including both metal loss (thickness reduction) and internal penetration.

3.2 Thermodynamic Considerations 3.2.1 Formation of Oxides Thermodynamically, an oxide is likely to form on a metal surface when the oxygen potential (pO2 ) in the environment is greater than the oxygen partial pressure in equilibrium with the oxide. The oxygen partial pressure in equilibrium with the oxide can be determined from the standard free energy of formation of the oxide. Consider the reaction: M+O2 ÐMO2

ð3:1Þ

aMO2 DG =7RT ln aM pO2

ð3:2Þ

Assuming the activities of the metal and the oxide are unity, Eq 3.2 becomes: DG =RT ln pO2

ð3:3Þ

Then

pO2 =eDG

=RT

ð3:4Þ

Equation 3.4 permits the determination of the oxygen partial pressure in equilibrium with the

0.5 mm

Fig. 3.1

A Ni-Cr alloy furnace heater coil suffering extensive internal oxidation attack with little surface scaling after service for 4 to 5 years at temperatures below 900 °C (1650 °F). Source: Ref 1

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

pg 7

Chapter 3:

partial pressure of oxygen in equilibrium with Cr2O3 at 1000 °C (1830 °F) is about 10−21 atm from Fig. 3.2. This implies that the formation of Cr2O3 is favored thermodynamically at 1000 °C in environments with oxygen potentials higher than 10−21 atm. When the environment is “reducing” (e.g., the environment generated by stoichiometric or substoichiometric combustion), the oxygen potential is controlled by either pH2 =pH2 O or pCO =pCO2 ratio. The oxygen potential can be determined by the reaction:

oxide from the standard free energy of formation. The standard free energies of formation of selected oxides as a function of temperature are shown in Fig. 3.2. The ﬁgure also allows quick determination of the oxygen partial pressure (pO2 ) in equilibrium with the oxide. This oxygen partial pressure can be read by drawing a straight line from the point marked “O” on the left vertical axis of Fig. 3.2 through the free-energy line of the oxide at the intersecting point with the temperature of interest. This line continues to extend until it intersects with the pO2 scale located at the right-hand side and bottom of the Fig. 3.2. The intersecting point shows the oxygen partial pressure in equilibrium with the oxide of interest. If the oxygen partial pressure in the environment is greater than the oxygen partial pressure in equilibrium with the oxide, the oxide is likely to form on the metal surface. Conversely, the oxide is not likely to form. For example, the

2H2 +O2 Ð2H2 O

ð3:5Þ

The standard free energy of formation is related to the partial pressures of hydrogen, oxygen, and water by: p2H2 O 2 pH2 pO2

DG =7RT ln

H2/H2O ratio

10–8

CO/CO2 ratio

10–8

10–6

ð3:6Þ

pO

10–4

10–4

10–6

!

2

10–2 10–2

0

O

1

–100

O4 Fe 3

–200

+O

4

2

=

O3 e2 6F

oO

NiO

+ Ni

O2

M

O2 o+

1 10–2

= 2C

1

2C

=2

10–4

M

2

–300

102 102

∆G °=RT In pO2 (kJ/mole O2)

–400 –500

H C

–600

4- Cr + 3

–700

Si +

M

2 Cr2O 3 =O2 3

O2

M

I 2O 3

O I+

2

B

104

4- A 3

B O O Ca Mg =2 2 B 108 = 2 O + 2 + O 2Ca g 2M Change of state Element Oxide M M Melting point M M B Boiling point B 1010

–900

10–14 106 10–16

M

–1000 –1100 –1200

200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 10 Temperature, °C CO/CO2 ratio 10–14 10–12 10

10–10

10–12

106

2- A = 3

10–6 10–8

104

iO 2 =S

–800

10–18 108 10–20

0

OK

10–22

H2/H2O ratio

10–20010–100 10–70 10–60 10–50 10–42 10–38 10–34 10–30 10–28 10–26 10–24

pO

2

Fig. 3.2

Oxidation / 7

Standard free energies of formation of selected oxides as a function of temperature. Source: Ref 2

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

pg 8

8 / High-Temperature Corrosion and Materials Applications

Rearranging the equation results in:

pO2 =eDG

=RT

1 (pH2 =pH2 O )2

ð3:7Þ

Thus, the oxygen partial pressures at various temperatures can be determined as a function of pH2 =pH2 O values. The pH2 =pH2 O value in equilibrium with the oxide can be read from Fig. 3.2, using the method discussed previously, except that the starting point for the straight line is “H” and the pH2 =pH2 O value is determined from the H2/H2O scale. For example, the oxygen potential, in terms of pH2 =pH2 O , in equilibrium with Cr2O3 at 1000 °C, is about 5 × 103 from Fig. 3.2. Thus, Cr2O3 is likely to form at 1000 °C when the pH2 =pH2 O ratio in the environment is less than 5×103. The equilibrium reaction for an environment whose oxygen potential is controlled by pCO =pCO2 is: 2CO+O2 Ð 2CO2

ð3:8Þ

The corresponding oxygen potential is:

pO2 =eDG

=RT

1 (pCO =pCO2 )2

ð3:9Þ

The pCO =pCO2 value can be read from Fig. 3.2 using the method discussed previously, with the exception that a straight line is drawn from point “C” to the CO/CO2 scale. Thus, it is possible to obtain the oxygen potential of the environment in terms of pO2 , pH2 =pH2 O , pCO =pCO2 , and the oxygen partial pressure in equilibrium with the oxide of interest from Fig. 3.2, to determine whether or not the oxide is likely to form thermodynamically. Figure 3.2 also illustrates the relative stability of various oxides. The most stable oxides have the largest negative values of ΔG°, or the lowest value of pO2 , or the highest values of pH2 =pH2 O and pCO =pCO2 . It is clear from Fig. 3.2 that oxides of iron, nickel, and cobalt, which are the alloy bases for the majority of engineering alloys, are signiﬁcantly less stable than the oxides of some solutes (e.g., chromium, aluminum, silicon, etc.) in engineering alloys. When one of these solute elements is added to iron, nickel, or cobalt, internal oxidation of the solute is expected to occur if the concentration of the solute is relatively low. As the solute concentration increases to a sufﬁciently high level, oxidation of the solute will be changed from internal oxidation to external oxidation, resulting in an oxide scale

that protects the alloy from rapid oxidation. This process is known as “selective oxidation.” The majority of iron-, nickel-, and cobalt-base alloys rely on selective oxidation of chromium to form a Cr2O3 scale for oxidation resistance. Some hightemperature alloys use aluminum to form an Al2O3 scale for oxidation resistance. Most oxides exhibit high melting points and remain in a solid state for the temperature range in which the alloys are used. If the oxide is present as a liquid state, catastrophic oxidation can occur. Since many engineering alloys contain many alloying elements for various metallurgical reasons, formation of oxides that become liquid at the service temperature should be prevented. Table 3.1 shows the melting points of selected oxides of alloying elements commonly found in high-temperature alloys. Most oxides remain solid until they reach extremely high temperatures. Oxides of molybdenum (MoO3) and vanadium (V2O5), however, exhibit very low melting points. Vanadium (V), which is a strong carbide former, is often used in alloy steels for increasing the strength of the material. However, the amount used typically is quite small and is not likely to form V2O5. Molybdenum (Mo) is also a strong carbide former and is used in a small amount to strengthen low-alloy steels (e.g., CrMo steels). It is unlikely these steels will be affected by MoO3-related oxidation problems. However, molybdenum is an effective alloying element for improving the resistance of the alloy to aqueous corrosion. Some stainless steel grades contain molybdenum, with superaustenitic stainless steels containing much higher levels of molybdenum. Some nickel-base alloys contain very high levels of molybdenum for either aqueous corrosion resistance or solid-solution

Table 3.1 Melting points of selected oxides for alloying elements commonly found in high-temperature alloys Qxide

αAl2O3 CoO Cr2O3 FeO Mn3O4 MoO3 Nb2O5 NiO SiO2 TiO2 V2O5 WO3 Source: Ref 3

Melting point, °C (°F)

2015 (3659) 1935 (3515) 2435 (4415) 1420 (2588) 1705 (3101) 795 (1463) 1460 (2660) 1990 (3614) 1713 (3115) 1830 (3326) 690 (1274) 1473 (2683)

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

pg 9

Chapter 3:

Oxidation / 9

volatile CrO3 (Ref 5–8). Table 3.2 shows the weight-loss data for Cr2O3 pellets after heating to 1000 to 1200 °C (1830 to 2190 °F) in dry O2 due to formation of gaseous CrO3 by oxidation of Cr2O3 (Ref 5). Caplan and Cohen (Ref 5) also observed that moisture promoted volatilization of Cr2O3. Asteman et al. (Ref 9) indicated that high vapor pressure of CrO2(OH)2 can form by reacting Cr2O3 with H2O in O2-containing environments. The theoretical calculated partial pressure of CrO2(OH)2 as a function of temperature for the O2-containing environment with pO2 =0:9 atm and pH2 O =0:1 atm is shown in Fig. 3.4 (Ref 9).

strengthening. The formation of MoO3 and its effect on oxidation are discussed in section 3.2.2 “Volatility of Oxides” and section 3.4.17 “Catastrophic Oxidation.” 3.2.2 Volatility of Oxides Some oxides exhibit high vapor pressures at very high temperatures (e.g., above 1000 °C, or 1830 °F). Oxide scales become less protective when their vapor pressures are high. Figure 3.3 shows vapor pressures of several refractory metal oxides exhibiting high vapor pressures at temperatures above 1000 °C (1830 °F) (Ref 4). Vanadium is typically used in small quantities as a carbide former in alloy steels. Thus, the volatility of VO2 is generally of no concern in oxidation of alloys. Molybdenum (Mo) and tungsten (W) are often used as alloying elements in signiﬁcant amounts in Ni- or Co-base alloys as solution-strengthening elements. Formation of WO3 or MoO3 may occur under certain conditions in some alloy systems, particularly in alloy systems containing insufﬁcient chromium for forming a protective Cr2O3 scale. A majority of engineering alloys rely on the Cr2O3 scale to provide resistance to oxidation. When heated to very high temperatures (i.e., above 1000 °C), Cr2O3 can react with O2 to form

Table 3.2 Weight loss of Cr2O3 on heating in dry O2 and Ar environments Run

Temperature, °C (°F)

Time, h

Gas

Gas ﬂow, mL/min

Weight loss, mg

1 2 3 4 5 6 7 8 9

1100 (2010) 1200 (2190) 1200 (2190) 1200 (2190) 1200 (2190) 1200 (2190) 1200 (2190) 1200 (2190) 1200 (2190)

20 20 20 20 20 20 42 66 115

Dry O2 Dry O2 Dry O2 Dry O2 Dry O2 Dry O2 Dry O2 Dry Ar Dry Ar

200 10 10 20 200 200 200 200 192

0.6 2.1 1.3 1.8 2.3 2.6 8.0 0 0

Source: Ref 5

Temperature, °C 1

1500

1000

600

10–2

WO3 CrO3

10–4

Vapor pressure, atm

MoO3 10–6 10

VO2

–8

10–10 10–12 10–14 10–16 10–18 0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

Inverse temperature, 103/T (K–1)

Fig. 3.3

Vapor pressures of several refractory metal oxides exhibiting high vapor pressures at temperatures above 1000 °C (1830 °F). Source: Ref 4

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

pg 10

10 / High-Temperature Corrosion and Materials Applications

–2

–4 CrO2(OH)2, g

Log p, X, atm

–6

–8

–10

–12 CrO3, g

–14

–16 600

700

800

900

1000

1100

1200

1300

Temperature, K Theoretical partial pressures of CrO2(OH)2 and CrO3 as a function of temperature for the environment with pO2 =0:9 atm and pH2 O =0:1 atm: Source: Asteman et al. (Ref 9)

For service temperatures above 1200 °C (2190 °F), the increasing volatility of oxides can progressively cause the oxide to lose its protective capability. SiO2 and Al2O3 are the only two oxides that are capable of forming a very protective barrier against oxidation at temperatures above 1200 °C (Ref 9). However, the SiO2 scale may lose some protective capability by forming gaseous SiO at low oxygen partial pressures (Ref 10).

Inverse log Parabolic Oxide mass, m

Fig. 3.4

Log

Linear

Time, t

3.3 Kinetic Considerations The kinetics of oxidation of metals and alloys generally follow several reaction rates. Most reactions follow a parabolic rate. Some reactions follow a linear rate. Some other reaction kinetics may include logarithmic and inverse logarithmic rates. These reaction kinetics are illustrated schematically in Fig. 3.5 (Ref 11), and a brief summary, based on the article by Danielewski (Ref 11), is presented in sections 3.3.1 to 3.3.3.

Fig. 3.5

Different oxidation kinetics. Source: Ref 11

decreases with increasing time due to the increasing diffusion distance for ions. The oxidation rate is, thus, inversely proportional to the thickness of the oxide scale: X 2 =k′t

ð3:10Þ

where X is the oxide scale thickness, t is the exposure time, and k′ is the parabolic constant; when t = 0, X = 0.

3.3.1 Parabolic Kinetics When the oxide scale forms on the metal surface, the oxidation reaction is controlled by the diffusion of ions through the oxide scale, which is in turn controlled by the chemical potential gradient as a driving force. As the thickness of the oxide scale increases, the rate of oxidation

3.3.2 Linear Kinetics When the oxide scale forming on the metal surface provides no protection barrier due to oxide cracking and spalling, volatile oxides, and molten oxidation products, the oxidation rate generally remains constant with increasing

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

pg 11

Chapter 3: Oxidation / 11

time. The linear oxidation kinetic rate can be expressed by:

The inverse logarithmic rate can be expressed by the following equation:

X =kl t

1=X =b7ki log t

ð3:11Þ

where X is the mass (or thickness) of the oxide, t is the exposure time, and kl is the linear rate constant; when t = 0, X = 0. 3.3.3 Logarithmic and Inverse Logarithmic Kinetics

where b and ki are constants.

3.4 Oxidation in Air, O2, and “Clean” Combustion Atmospheres 3.4.1 Carbon and Cr-Mo Steels

At very low temperatures when the oxide ﬁlm forms on the metal surface, the oxidation rate usually follows either a logarithmic or inverse logarithmic rate. The driving force for the oxidation is the electric ﬁeld across the oxide ﬁlm. The logarithmic rate can be expressed by: X =ke log (at+1)

ð3:13Þ

ð3:12Þ

where ke and a are constants.

Carbon and Cr-Mo steels are the most widely used engineering materials and are used extensively for high-temperature applications in power generation, chemical and petrochemical processing, petroleum reﬁning, pulp and paper industry, industrial heating, and metallurgical processing. At temperatures below 570 °C (1060 °F), iron (Fe) oxidizes to form Fe3O4 and Fe2O3. Above

120

110

100 1400 °F

90

Weight-loss, mg/cm2

80

70 1200 °F 60

50

40

30

20

1000 °F

10

800 °F

0

0

100

200

300

400

500

600

700

Time, h

Fig. 3.6

Oxidation behavior of plain low-carbon steel in air at 430, 540, 650, and 760 °C (800, 1000, 1200, and 1400 °F). Source: Ref 12

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

12 / High-Temperature Corrosion and Materials Applications

570 °C (1060 °F), it oxidizes to form FeO, Fe3O4, and Fe2O3. The oxidation behavior of carbon steel in air at 430, 540, 650, and 760 °C (800, 1000, 1200, and 1400 °F) is summarized in Fig. 3.6 (Ref 12). At 430 and 540 °C (800 and

Average penetration/side, mils

10.5

Calculated continuing penetration rate

53 mpy*

9.0 Carbon steel - 1200 °F

7.5 6.0 4.5

5.3 mpy* A242 type 1 HSLA steel - 1200 °F

3.0

Carbon steel - 1000 °F

1.5

2.3 mpy*

A242 type 1 HSLA steel - 1000 °F 1.0 mpy*

0

200

400

600

800 1000

Exposure time, h

Fig. 3.7

Oxidation of carbon steel and high-strength low-alloy (HSLA) steel in air. Source: Ref 13, reproduced

from Ref 14

1000 °F), carbon steel showed very little weight gain after exposure for 500 h. As the temperature was increased to 650 °C (1200 °F), the oxidation rate was signiﬁcantly increased. Carbon steel suffered rapid oxidation at 760 °C (1400 °F), exhibiting essentially a linear rate of oxidation attack. Vrable et al. (Ref 13) reported oxidation data for carbon steel and high-strength low-alloy (HSLA) steel, as shown in Fig. 3.7 (Ref 14). At 650 °C (1200 °F), carbon steel suffered an oxidation rate of about 1.3 mm/yr, or 53 mpy (mils per year). The oxidation rate is expected to be much higher when exposed to temperatures higher than 650 °C (1200 °F). Recent test results by John (Ref 15) showed that carbon steel exhibited about 0.25 mm/yr (10 mpy) of oxidation at 604 °C (1120 °F). Figure 3.7 also illustrates that HSLA steel is more oxidation resistant than carbon steel, presumably due to minor alloying elements such as manganese, silicon, chromium, and nickel. Cr-Mo steels are used at higher temperatures than carbon steel because of higher tensile and creep-rupture strengths as well as better microstructural stability. Molybdenum and chromium provide not only solid-solution strengthening but also carbide strengthening. Low-alloy steels with chromium and silicon additions exhibit better oxidation resistance than carbon steel. The beneﬁcial effects of chromium and silicon additions to carbon steel are summarized in Fig. 3.8 (Ref 16). Silicon is very effective in improving the oxidation resistance of Cr-Mo steels. Addition of 1.5% Si to 5Cr-0.5Mo steel signiﬁcantly improved its oxidation resistance. The most important alloying element for improving oxidation resistance is chromium. As shown in Fig. 3.8, for 0.5% Mo steels, increasing chromium from 1 to 9% signiﬁcantly increases oxidation resistance. The 7Cr-0.5Mo and 9Cr-1Mo steels showed negligible oxidation rates at temperatures up to 680 °C (1250 °F) and 700 °C (1300 °F), respectively. Further increases in chromium improve oxidation resistance even more. Alloys become martensitic or ferritic grades of stainless steels (400 series) when chromium content is increased to 12% or higher. 3.4.2 Martensitic, Ferritic, and Austenitic Stainless Steels

Fig. 3.8

Effects of chromium and/or silicon on the oxidation resistance of steels in air. Source: Ref 16

The superior oxidation resistance of martensitic and ferritic stainless steels to that of carbon and Cr-Mo steels is illustrated in Fig. 3.9 (Ref 17). As chromium content in the straight

pg 12

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

Chapter 3: Oxidation / 13

Fig. 3.9

Oxidation resistance of carbon, low-alloy, and stainless steels in air after 1000 h at temperatures from 590 to 930 °C (1100 to 1700 °F). Source: Ref 17

chromium steels increases from 9 to 25%, resistance to oxidation improves signiﬁcantly. The 25Cr steel (Type 446) is the most oxidation resistant among the 400 series stainless steels, due to the development of a continuous Cr2O3 scale on the metal surface. In Fe-Cr alloys, it appears that a minimum of approximately 18wt % Cr is needed to develop a continuous Cr2O3 scale against further oxidation attack (Fig. 3.10) (Ref 18). Cyclic oxidation studies conducted by Grodner (Ref 19) also revealed that Type 446 was the best performer in the 400 series stainless steels, followed by Types 430 (14–18Cr), 416 (12–14Cr), and 410 (11.5–13.5Cr) (Fig. 3.11). Figure 3.11 shows that Fe-12Cr steels, such as Types 410 and 416, showed increased rates of cyclic oxidation above 760 °C (1400 °F). At 650 °C (1200 °F) in air, cycling from 650 to 300 °C, 12Cr-1Mo steel (X20 CrMoV 12 1) steel

exhibited a thin, adherent (Fe,Cr)2O3 scale, as observed by Walter et al. (Ref 20). The growth of the (Fe,Cr)2O3 scale as a function of the accumulated isothermal hold time up to 1000 h is shown in Fig. 3.12 (Ref 20). John (Ref 15) reported that Fe-12Cr steel (Type 410) exhibited an air oxidation rate of 0.25 mm/yr (10 mpy) at 832 °C (1530 °F). Another ferritic stainless steel, 18SR (about 18% Cr), was found to be as good as, and sometimes better than, Type 446 (25% Cr), as illustrated in Tables 3.3 and 3.4 (Ref 21). This was attributed to the addition of 2% Al and 1% Si to the alloy. Furthermore, both of these ferritic stainless steels, Type 446 and 18SR, showed better cyclic oxidation resistance than some austenitic stainless steels, such as Type 309 and 310, when cycled to 980 to 1040 °C (1800 to 1900 °F), as shown in Table 3.4.

pg 13

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 14

14 / High-Temperature Corrosion and Materials Applications

Fe2O3 Fe3O4 FeO Fe 10–6 Fe2O3 Fe3O4 Parabolic rate constant, g2 · cm4 · s1

FeO 10–7

Fe – 2Cr

Fe/Cr oxide

Fe2O3 (Fe,Cr)2O3

10–8

Fe – 9Cr

Fe2O3

Fe3O4

10–9

Cr2O3 Fe – 16Cr

FeFe(2 ... x)CrxO4

10–10 Cr2O3 Fe – 28Cr 10–11 0

10

20

30

40

50

60

70

80

90

100

Alloy chromium content, wt%

Fig. 3.10

Effect of chromium content on oxidation of Fe-Cr alloys at 1000 °C (1830 °F) in 0.13 atm O2. Source: Ref 18

When the service temperature is above 650 °C (1200 °F), ferritic stainless steels, which have a body-centered cubic (bcc) crystal structure, drastically lose their elevated-temperature strength (both tensile and creep-rupture strength). As a result, the application of ferritic stainless steels becomes limited at higher temperatures. At these temperatures, alloys with a face-centered cubic (fcc) crystal structure are preferred because of their higher creep-rupture strength. Nickel is added to Fe-Cr steels to stabilize the austenitic structure. The austenitic structure is inherently stronger and more creep resistant than the ferritic structure (Ref 22). The 300 series austenitic stainless steels have been widely used for high-temperature components in various industries because of their strength and high-temperature corrosion resistance, including oxidation resistance. These alloys exhibit higher elevated-temperature

strength than do ferritic stainless steels. Furthermore, they do not suffer 475 °C (885 °F) embrittlement or ductility-loss problems in thick sections and in heat-affected zones as do ferritic stainless steels. Nevertheless, some austenitic stainless steels can suffer some ductility loss upon long-term exposure to intermediate temperatures (e.g., 540 to 800 °C, or 1000 to 1470 °F) due to sigma-phase formation. The oxidation resistance of two austenitic stainless steels, Types 309 and 310, is compared with that of a number of ferritic stainless steels in Fig. 3.11. Several austenitic stainless steels are compared in Fig. 3.13 (Ref 23). Nickel improves the resistance of alloys to cyclic oxidation. Moccari and Ali (Ref 24) also observed the similar beneﬁcial effects of nickel in improving the oxidation resistance of alloys. Brasunas et al. (Ref 25) studied the oxidation behavior of about 80 experimental Fe-Cr-Ni alloys exposed

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 15

Chapter 3: Oxidation / 15

Oxidation resistance of several stainless steels as a function of temperature. Source: Ref 19

Thickness of the oxide layer (d ), µm

Fig. 3.11

40 35 30 25 20 15 10 5 0 0

100

200

300

400

500

600

700

800

900

1000

Testing time (t ), h

Fig. 3.12

Oxidation behavior of 12Cr-1Mo steel (X20 CrMoV 12 1) at 650 °C (1200 °F) in air with every 8 h of exposure specimens being cycled from 650 to 300 °C. Source: Ref 20

to air-H2O mixture at 870 to 1200 °C (1600 to 2190 °F) for 100 and 1000 h. They observed that increases in nickel in excess of 10% in alloys containing 11 to 36% Cr improved the oxidation resistance of the alloys. John (Ref 15) reported that Type 304 and 310 exhibited 0.25 mm/yr (10 mpy) of oxidation attack (both metal loss and

internal oxidation penetration) at 893 and 982 °C (1640 and 1800 °F), respectively, in air. In Fig. 3.13, several high-nickel alloys were found to be more resistant to oxidation than austenitic stainless steels. The oxidation behavior of highnickel Fe-Ni-Cr alloys and Ni-base alloys is discussed in later sections of this chapter.

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

16 / High-Temperature Corrosion and Materials Applications

Table 3.3 Cyclic oxidation resistance of several stainless steels in air cycling to 870 to 930 °C (1600 to 1700 °F) temperature range Specimen weight changes after indicated cycles, mg/cm2 Alloy

288 cycles

409 + Al Destroyed 430 9.9 22-13-5 0.5 442 0.7 446 0.3 309 0.3 18SR 0.3

480 cycles

750 cycles

958 cycles

… Destroyed −3.0 1.2 0.4 −4.6 0.4

… … −18.8 1.5 0.2 −23.7 0.5

… … −35.7 1.5 0.1 −32.6 0.6

Note: 15 min in furnace and 15 min out of furnace. Source: Ref 21

Table 3.4 Cyclic oxidation resistance of several stainless steels in air cycling to 980 to 1040 °C (1800 to 1900 °F) temperature range Specimen weight changes after indicated cycles, mg/cm2 Alloy

130 cycles

368 cycles

561 cycles

753 cycles

1029 cycles

446 18SR 309 310

0.4 0.7 −24.2 1.5

0.5 1.1 −77.5 −11.3

−0.2 1.5 −178.3 −29.3

7.0 2.2 −242 −62.8

−19.4 3.0 −358 −107

Note: 15 min in furnace and 15 min out of furnace. Source: Ref 21

In evaluating materials for automobile emission-control devices, such as thermal reactors and catalytic converters, Kado et al. (Ref 26) and Michels (Ref 27) have carried out cyclic oxidation tests on various stainless steels. In cyclic oxidation tests performed by Kado et al. (Ref 26) in still air at 1000 °C (1830 °F) for 400 cycles (30 min in the furnace and 30 min out of the furnace), Types 409 (12Cr), 420 (13Cr), and 304 (18Cr-8Ni) suffered severe attack. Type 420 (13Cr) was completely oxidized after only 100 cycles, although the sample did not show any weight changes. Alloys that performed well under these conditions were Types 405 (14Cr), 430 (17Cr), 446 (25Cr), 310 (25Cr-20Ni), and DIN 4828 (19Cr-12Ni-2Si), as illustrated in Fig. 3.14. When cycled to 1200 °C (2190 °F) for 400 cycles (30 min in the furnace and 30 min out of the furnace), all the alloys tested except F-1 alloy (Fe-15Cr-4Al) suffered severe oxidation attack (Fig. 3.15). This illustrates the superior oxidation resistance of alumina formers (i.e., alloys that form Al2O3 scales when oxidized at elevated temperatures). The data also illustrate that for temperatures as high as 1200 °C (2190 °F), Cr2O3 oxide scales can no longer provide adequate oxidation resistance. Oxidation of Fe-Cr-Al alloys is discussed in Section 3.4.7.

Kado et al. (Ref 26) also investigated oxidation behavior in a combustion environment that simulated the gasoline engine. Their test involved air-to-fuel ratios of 9 to 1 and 14.5 to 1 and regular gasoline that contained 0.01wt% S. Exhaust gas taken from the exhaust manifold was mixed with air before being piped into a furnace retort where tests were performed. Test specimens were exposed to the mixture of exhaust gas and air. The gas mixture consisted of 72.4% N2, 9.7% H2O, 9.93% O2, 8% CO2, and 507 ppm NOx, when an air-to-fuel ratio of 14.5 to 1 was used for combustion, while that coming from the combustion using an air-to-fuel ratio of 9 to 1 consisted of 70.6% N2, 13.7% H2O, 3.21% O2, 12.5% CO2, and 34 ppm NOx. The total accumulated test duration was 200 h (400 cycles with 30 min in the hot zone and 30 min out of the hot zone). Test results along with air oxidation data are summarized in Fig. 3.16. There were no signiﬁcant differences between air and exhaust gas test environments when tested at 800 °C (1470 °F). All the alloys tested showed negligible attack except Type 304, which exhibited much more severe attack in the exhaust environment. When the test temperature was increased to 1000 °C (1830 °F), all the 400 series stainless steels with less than 17% Cr (i.e., Types 409, 405, 410, and 430) and Type 304 exhibited signiﬁcantly more oxidation attack in the exhaust environment. Type 310, Type 446, DIN 4828 (Fe-19Cr-12Ni-2Si), F-1 alloy (Fe15Cr-4Al), A-1 alloy (Fe-16Cr-13Ni-3.5Si), and A-2 alloy (Fe-20Cr-13Ni-3.5Si) performed well. At 1200 °C (2190 °F), all the alloys tested suffered severe oxidation attack with the exhaust environment being more aggressive than air. These authors attributed the enhanced attack to the presence of sulfur in the exhaust gas environment, although low-sulfur (0.01%) gasoline was used for testing. Sulfur segregation to the scale/metal interface was detected. In a study by Michels (Ref 27), the engine combustion atmosphere was also found to be signiﬁcantly more corrosive than the air-10% H2O environment. The engine combustion exhaust gas contained about 10% H2O along with 2% CO, 0.33 to 0.55% O2, 0.05 to 0.24% hydrocarbon, and 0.085% NOx. The balance was presumably N2 (not reported in the paper). The engine exhaust gas was piped into a furnace retort where the tests were performed. The results, which were generated in the air-10%H2O and the engine exhaust environment, are shown in Fig. 3.17. After exposure to the air-10%H2O

pg 16

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 17

Chapter 3: Oxidation / 17

10

625 600 80Ni-20Cr 60Ni-15Cr

0

800 22Cr-32Ni

–10 Type 310 25Cr-20Ni

– 20

Change in weight, %

19Cr-14Ni 20Cr-25Ni Type 309 23Cr-13Ni – 30

– 40 Type 347 18Cr-8Ni(Cb)

– 50 Type 304 18Cr-8Ni

– 60

– 70 0

200

400

600

800

1000

Hours of cyclic exposure (15 min heating – 5 min cooling)

Fig. 3.13

Cyclic oxidation resistance of several stainless steels and nickel-base alloys in air at 980 °C (1800 °F). Source: Ref 23

environment at 980 °C (1800 °F) for 102 h, Type 309, Type 310, 18SR, alloy OR-1 (Fe-13Cr-3Al), alloy 800, and alloy 601 were all relatively unaffected. On the other hand, only 18SR and alloy 601 were relatively unaffected by the engine exhaust gas environment, with alloy OR-1, Type 309, Type 310, and alloy 800 suffering severe oxidation attack. The sulfur content

in the gasoline used in this test was not reported. The relatively high gas velocity, about 6.1 to 9.2 m/s (20 to 30 ft/s) was considered by the author to be one of the possible factors responsible for much higher oxidation attack in the engine exhaust gas test. Oxidation data generated in combustion atmospheres is relatively limited. No systematic

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 18

18 / High-Temperature Corrosion and Materials Applications

Fig. 3.14

Cyclic oxidation resistance of several ferritic and austenitic stainless steels in still air at 1000 °C (1830 °F) for up to 400 cycles (30 min in furnace and 30 min out of furnace). Source: Ref 26

409 0 F–1 (Fe-15Cr-4AI)

Weight change, mg/cm2

430 –100

DIN 4828 (19Cr-12Ni-2Si)

420 –200

310

304

446

–300

0

100

200

300

400

Number of cycles

Fig. 3.15

Cyclic oxidation resistance of several ferritic and austenitic stainless steels in still air at 1200 °C (2190 °F) for up to 400 cycles (30 min in furnace and 30 min out of furnace). Source: Ref 26

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 19

Chapter 3: Oxidation / 19

1200 °C (2190 °F)/400 cycles 100

80

60

40

Thickness loss, %

20

0

1000 °C (1830 °F)/400 cycles

100

80

60

40

20

0 800 °C (1470 °F)/400 cycles 20

) .5

Si

) 3N

(F e-

20

C

r-1

3N (F e-

i-3

.5 i-3

31 A– 2

16

C

r-1

D

A– 1

In air In exhaust gas (R = 9)

Si

0S

0

IN

48

31

28

4 30

6

0

44

r-4

F–

1

(F e-

15

C

43

) AI

0 41

5 40

40

9

0

Fig. 3.16

Comparison of cyclic oxidation resistance between air and gasoline engine exhaust gas environments at 800, 1000, and 1200 °C (1470, 1830, 2190 °F) for 400 cycles (30 min in hot zone and 30 min out of hot zone). Alloy F-1 suffered localized attack at 1200 °C in engine exhaust gas. Source: Ref 26

studies have been reported that varied combustion conditions, such as air-to-fuel ratios. In combustion atmospheres, the oxidation of metals or alloys is not controlled by oxygen only. Other combustion products, such as H2O, CO, CO2, N2, hydrocarbon, and others, are expected to inﬂuence oxidation behavior. When air is used for combustion, nitridation in conjunction with oxidation can occur in combustion atmospheres under certain conditions. This nitridation/ oxidation behavior of alloys is discussed in

Chapter 4 “Nitridation.” The presence of water vapor can be an important factor in affecting oxidation behavior of alloys. The effect of water vapor on the oxidation resistance of alloys is covered in Section 3.4.15. 3.4.3 Surface Chemistry versus Bulk Chemistry It is important to point out that the surface chemistry may not be the same as the reported

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

20 / High-Temperature Corrosion and Materials Applications

Fig. 3.17

Cyclic oxidation resistance of several ferritic and austenitic stainless steels in (a) air-10H2O at 980 °C (1800 °F) cycled every 2 h, and (b) gasoline engine exhaust gas at 980 °C (1800 °F) cycled every 6 h. Source: Ref 27

“bulk” chemistry. Some of the manufacturing processes involved in the production of an alloy product, such as plate, sheet, or tubular products, may result in lower chromium contents at or near the surface of the product. This is particularly important for austenitic stainless steels, since the chromium speciﬁcation range for austenitic stainless steels can be at or near the borderline for forming a continuous Cr2O3 scale. For example, ASTM A 213/A 213M (or ASME SA213/SA-213M) speciﬁcation for the chromium range is 18.0 to 20.0% for Type 304, 16.0 to 18.0% for Type 316, 17.0 to 20.0% for Type 321, and 17.0 to 20.0% for Type 347. The lower end of the chromium content in these alloys is essentially at the minimum level that is considered to be required for forming a continuous Cr2O3 scale when exposed to elevated temperatures. A slight surface depletion in chromium due to some manufacturing processes may result in the condition that a continuous, protective Cr2O3 scale cannot be formed, thus resulting in accelerated oxidation due to formation of nonprotective iron oxides. Some manufacturing processes are prone to producing the ﬁnal ﬁnished product with surface depletion of chromium. This surface depletion

becomes more critical for sheet products or thinwall tubular products because of much higher percentage of the surface-depletion zone in the overall thickness of the component. In alloy manufacturing, annealing is required after each cold-rolling step for sheet product manufacturing (or cold pilgering for reduction in thickness and sizes for tubular product manufacturing) to soften the metal for further cold-reduction steps until a ﬁnal product is produced. Stainless steels are typically annealed in the temperature range of 1010 to 1120 °C (1850 to 2050 °F) (Ref 28). When annealing is performed in air, heavy chromium oxide scales form on the metal surface. As a result, the matrix immediately underneath the oxide scales can be depleted in chromium. Figure 3.18 shows the concentration proﬁle of chromium near the surface of the plate of alloy AL-6XN (Fe-21Cr-24Ni-6.5Mo-0.2N) after annealing in air at 1120 and 1175 °C (2050 and 2150 °F), respectively (Ref 29). Oxidation during annealing at either temperature resulted in a signiﬁcant chromium depletion near the surface of the plate. It is quite common to perform annealing in air during manufacturing of stainless steels. This is commonly referred to as “black annealing,” as opposed to “bright

pg 20

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 21

Chapter 3: Oxidation / 21

annealing,” which is performed in a protective atmosphere, such as hydrogen atmosphere. The oxide scales on the alloy plate, sheet, or tubing are generally removed by acid pickling. This manufacturing process can often produce ﬂat products (plate or sheet) as well as tubular products with surface depletion of chromium. When bright annealing is performed, the alloy surface is protected from oxide scale formation. As a result, chromium depletion at and near the product surface is minimized. Surface depletion in chromium becomes increasingly critical as the thickness of the component, such as sheet or tubular product,

decreases. Figure 3.19 shows an example of surface depletion of chromium in a thin-gage commercial heat-exchanger tube made from Type 321 in the as-fabricated condition (Ref 30). The manufacturing process involved was not known. The Type 321 tube is shown to exhibit depletion in chromium near the tube surface when analyzed by energy-dispersive x-ray spectroscopy (EDX). The analysis was terminated at approximately about 0.5 μm from the tube surface. It is expected that the chromium content would decrease further if the analysis was performed at locations closer to the surface. After service for 6 months as a recuperator tube at

22 As hot rolled

Chromium content, wt%

20

18

Annealed 1175 °C Annealed 1120 °C

16 AL-6XN 14 2

0

8

6

4

10

12

14

16

18

20

Distance, µm

Fig. 3.18

Surface deletion of chromium near the surface of the plate of alloy AL-6XN (Fe-21Cr-24Ni-6.5Mo-0.2N) after annealing in air at 1120 and 1175 °C (2050 and 2150 °F), respectively. Source: Ref 29

18.4

Cr concentration, wt%

18.2 18 17.8 17.6 17.4 17.2 17 16.8 0

1

2

3

4

5

6

7

8

9

10

Distance from tube OD surface, µm

Fig. 3.19

Surface depletion of chromium observed in a thin-gage commercial heat-exchanger tube in the as-fabricated condition made from Type 321. Source: Ref 30

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

22 / High-Temperature Corrosion and Materials Applications

metal temperatures of approximately 620 to 670 °C (1150 to 1240 °F) for preheating air, signiﬁcant oxide spalling and scaling was observed on the air side of the heat-exchanger tube. Figure 3.20 shows heavy oxide scales formed on the side of the tube exposed to the incoming air after 6 months of service (Ref 30). The oxide scales were analyzed by scanning electron microscopy/energy-dispersive x-ray spectroscopy (SEM/EDX) analysis; the results are shown in Fig. 3.21. The analysis showed that the outermost oxide layer was essentially iron oxides with very little chromium. Thus, the stainless steel recuperator no longer exhibited adequate resistance to oxidation because of its failure to form and maintain a protective Cr2O3 scale. Once a protective Cr2O3 scale is no longer present on the stainless steel surface, iron oxides then take over. This results in scaling and accelerating oxidation. This is often referred to as “breakaway” oxidation (or corrosion). “Breakaway” oxidation is discussed in Section 3.4.13. Stainless steels manufactured by different producers can exhibit different chemical compositions and different surface characteristics. This is illustrated in Fig. 3.22 (Ref 30). Two Type 321 heat-exchanger tubes, manufactured by two different suppliers (supplier A and B), were tested in the ﬁeld in the same recuperator as described previously for preheating air at approximate metal temperature of 620 to 670 °C (1150 to 1240 °F) for about 1008 h. Figure 3.22 (a) shows the formation of mushroom-type oxide nodules on the surface of the Type 321 tube, produced by Supplier A. This is considered the initiation of breakaway oxidation. As shown in Fig. 3.22(a), the mushroom-type oxide nodule consisted of two layers of oxides, light grayish outer oxide scale and darkish inner oxide layer. The rest of the metal surface was still protected

by a thin, adherent oxide scale, which was not clearly revealed in the ﬁgure at about 400× magniﬁcation. Once the formation of the light, grayish outer oxide scale formed on the metal surface, the inner oxidation took place with accelerated growth. The outer oxide scale was found to be iron-rich oxides with little chromium, as shown in Fig. 3.23 (Ref 30). With the formation of the iron-rich oxide scale at the outer layer, which failed to provide protection, the inner Fe-Cr oxides were found to penetrate into the metal, causing a relatively massive internal oxidation penetration (Fig. 3.23). In the same test, the other Type 321 heat-exchanger tube (produced by supplier B), which was subjected to the same test condition and duration, was found to exhibit a thin, adherent oxide scale with no evidence of mushroom-type oxide nodules, as shown in Fig. 3.22(b). The oxide scale was analyzed by SEM/EDX, showing an Fe-Cr oxide scale formed on the metal surface (Fig. 3.24). Bulk chemical compositions of two tubes were analyzed; results are shown in Table 3.5. The chemical analysis results showed that the chromium in the Type 321 tube from supplier A was essentially at the lower end of the speciﬁcation range. Any depletion in chromium near the surface could result in a chromium level that is below the speciﬁcation limit. On the other hand, the tube from supplier B contained much higher chromium and was found to be much more resistant to oxidation under the same test conditions. It appears that some stainless steel manufacturers produce their products with leaner chemistry in terms of major alloying elements, such as chromium and nickel. For hightemperature oxidation and other modes of corrosion, stainless steel with the chromium in the lower end of the speciﬁcation range can be potentially more prone to breakaway oxidation

25 µm

Fig. 3.20

Heavy oxide scales formed on the side of Type 321 recuperator tube that was exposed to the incoming air after 6 months of service with the metal temperatures approximately 620 to 670 °C (1150 to 1240 °F). This tube was from the same batch of tubes that shows surface chromium depletion (Fig. 3.19). Source: Ref 30

pg 22

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 23

Chapter 3: Oxidation / 23

A

B C D

E 25 µm (a)

30 µm

Fig. 3.21

Scanning electron micrograph (backscattered electron image) showing the oxide scales formed on the outside diameter of the heat-exchanger tube (from the same batch of tubes that showed surface chromium depletion) exposed to air for 6 months. Energy-dispersive x-ray spectroscopy (EDX) analysis was performed to determine the chemical compositions at different locations, marked “A” to “E.” The results (wt%) of the EDX analysis are summarized below (minor elements not included). Source: Ref 30 A: 1.9% Cr, 97.6% Fe. B: 44.2% Cr, 44.4% Fe, 6.8% Ni. C: 48.8% Cr, 40.4% Fe, 4.4% Ni. D: 28.3% Cr, 46.4% Fe, 21.0% Ni. E: 38.7% Cr, 45.3% Fe, 10.2% Ni.

or corrosion. A brief discussion of this important issue is presented in Section 3.4.5. 3.4.4 Surface Conditions As discussed in Section 3.4.3 “Surface Chemistry versus Bulk Chemistry,” the concentration of chromium at and near the surface of the alloy product plays a signiﬁcant role in the oxidation of stainless steels such as Types 304, 316, 321, 347, and so forth. This is because the chromium concentration of these stainless steels is at the lower end of the chromium range that is generally required to form a continuous Cr2O3 scale when heated to elevated temperatures.

25 µm (b)

Fig. 3.22

Type 321 heat-exchanger tubes, which were manufactured by two different alloy suppliers, were tested in the same facility as described previously for preheating air at approximate metal temperature of 620 to 670 °C (1150 to 1240 °F) for about 1008 h. (a) Supplier A. (b) Supplier B. Note the tube from supplier A showed the initiation of accelerated oxidation attack (a) as opposed to the tube from supplier B showing no sign of accelerated oxidation attack (b).

Manufacturing processes can greatly inﬂuence the surface chemistry of an alloy product. Stainless steels can be ﬁnished into the ﬁnal product by bright annealing (i.e., annealing is performed in a protective atmosphere, such as hydrogen environment or dissociated ammonia environment). This process generally produces a product with minimal depletion of chromium at or near the surface. On the other hand, when the alloy product is ﬁnished by black annealing (i.e., annealing is performed in air or combustion atmosphere in the furnace) and followed by acid pickling, there is a good chance that the alloy product may exhibit a surface depletion of chromium. This is particularly important for thin-gage sheet products or thin tubular products.

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 24

24 / High-Temperature Corrosion and Materials Applications

In most laboratory oxidation tests, the test specimens are typically prepared by grinding and polishing with different grits of emery papers prior to testing. The objective of grinding and polishing the test specimens to a certain surface ﬁnish condition is to keep the surface condition of all test specimens constant in order to compare the oxidation behavior of different alloys. However, the mechanical forces of grinding and polishing can produce a thin cold-worked layer on the specimen surface. This cold-worked structure at the surface layer can signiﬁcantly enhance the diffusion of chromium from the interior to the surface of the metal to form

1

chromium oxide scales when heated to elevated temperatures, thus increasing the oxidation resistance of the alloy. For some critical applications involving a thin-gage sheet (or foil) product or a thin tubular product, testing should be carried out on the specimen that retains the surface condition of the product without prior surface grinding or mechanical polishing. Electropolishing, which is not commonly used to improve the surface ﬁnish of the alloy product for high-temperature services, may cause surface depletion of chromium for the product. Nevertheless, some investigators may use electropolishing to prepare the surface condition of the test specimens. Table 3.6 shows some comparison oxidation data generated in wet O2 between the wet ground and electropolished surface conditions for several stainless steels (Ref 31). 3.4.5 Today’s Stainless Steels Some stainless steel producers may manufacture stainless steels at the bottom of the

2

1 3

2 4

13 µm 13 µm

Fig. 3.23

Scanning electron micrograph (backscattered electron image) showing the oxide scales formed on the outside diameter of Type 321 tube (from supplier A) exposed to air at approximately 620 to 670 °C (1150 to 1240 °F) for 1008 h. Energy-dispersive x-ray spectroscopy (EDX) analysis was performed to determine the chemical compositions at different locations as marked 1 to 4 in the oxides. The results (wt%) of the EDX analysis are (minor elements not included): 1: 12% Cr, 86% Fe. 2: 52% Cr, 24% Fe, 18% Ni. 3: 32% Cr, 25% Fe, 38% Ni. 4: 37% Cr, 52% Fe, 7% Ni.

Table 3.5

Fig. 3.24

Scanning electron micrograph (backscattered image) showing the oxide scales formed on the outside diameter of Type 321 tube (from supplier B) exposed to air at approximately 620 to 670 °C (1150 to 1240 °F) for 1008 hours. EDX analysis was performed to determine the chemical compositions at different locations, marked as No. 1 and 2, in the oxides. The results (wt%) of the EDX analysis are (minor elements not included): 1: 35% Cr, 50% Fe, 6% Ni. 2: 21% Cr, 59% Fe, 4% Ni.

Chemical compositions of Type 321 tubes from Suppliers A and B Composition, wt%

Supplier

A B

C

Cr

Ni

Ti

Mn

Si

Mo

Cu

P

S

Fe

0.043 0.072

17.01 17.77

8.85 8.81

0.32 0.38

1.04 0.94

0.54 0.76

0.35 0.21

0.26 0.09

0.032 0.035

<0.005 0.018

bal bal

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 25

Chapter 3: Oxidation / 25

speciﬁcation range for key alloying elements, such as chromium, to reduce materials cost. Kelly (Ref 32) discusses this point in his 1992 paper, as shown in Tables 3.7 and 3.8. Accordingly, the chromium content can be insufﬁcient to maintain a continuous chromium oxide scale during prolonged service, or when subjected to thermal cycling, or overheating conditions, thus promoting breakaway oxidation. These “lean” stainless steels can be further “aggravated” by the surface depletion of chromium resulting from manufacturing processes that may involve excessive pickling after “black” annealing (annealing in air or combustion atmosphere), during successive reductions in cold rolling in ﬂat product manufacturing, or pilgering in tubular manufacturing. This can result in a product, particularly a thin-gage sheet or tube, in which the chromium concentration at the surface is too low to form or maintain a continuous chromium Table 3.6 Metal losses of several stainless steels tested at 648 °C (1200 °F) for 168 h in wet O2 Alloy

Surface ﬁnish

Metal loss after descaling, mg/cm2

304

Wet ground Electropolished Wet ground Electropolished Wet ground Electropolished Wet ground Electropolished

3.6 5.9 1.2 6.5 1.8 6.6 2.3 6.4

321 316 347

oxide scale during service. As a result, iron oxides and isolated nonprotective Fe-Cr oxide nodules can develop on the metal surface, thus resulting in breakaway oxidation. 3.4.6 Special, Proprietary Austenitic Stainless Steels There are several commercial proprietary heatresistant alloys that belong to the austenitic stainless steel group in terms of chromium and nickel contents, but with the addition of silicon for increasing the resistance of the alloy to oxidation and other high-temperature corrosion attack. These alloys include 253MA® (Fe21Cr-11Ni-1.7Si-0.17N-0.04Ce) and 85H® (Fe18.5Cr-14.5Ni-3.5Si-0.2C-1.0Al). Figure 3.25 shows the cyclic oxidation resistance of 253MA in air at 1090 °C (2000 °F) compared with that of Type 309 and some higher-alloyed Fe-NiCr alloys (e.g., 800H and RA330) (Ref 33). Figure 3.26 shows more cyclic oxidation tests comparing 253MA with 800H and 353MA at 1093 and 1150 °C (2000 and 2100 °F) (Ref 34). Figure 3.27 shows the oxidation resistance of 253MA compared with austenitic stainless steels, such as 18-10Ti (Type 321), 20-12Si (DIN 1.4828, 20Cr-12Ni-2Si), and 25-20 (Type 310), and 353MA and Ni-Cr alloy 601 as a function of temperature (Ref 35). Figure 3.28 shows that 12

Source: Ref 31

Composition, wt%

ASTM A 240 Pre-1965 heats Current production

Cr

Ni

18.00–20.00 18.7 18.3

8.00–10.50 9.9 9.0

Source: Ref 32

10

RA309

800H RA310

Weight gain, mg/cm2

Table 3.7 Chromium and nickel contents for Type 304 between pre-1965 heats and current production heats

1090 °C (2000 °F) 20 h cycles in air

8 RA330 RA333 6 RA601

4 RA 253 MA

Table 3.8 Chromium, nickel, and molybdenum contents for Type 316 between pre-1965 heats and current production heats Composition, wt%

ASTM A 240 Pre-1965 heats Current production Source: Ref 32

2

0

Cr

Ni

Mo

16.00–18.00 17.9 16.3

10.00–14.00 12.4 10.2

2.00–3.00 2.4 2.1

Fig. 3.25

100

300 200 Exposure time, h

400

500

Cyclic oxidation resistance of 253MA in air at 1090 °C (2000 °F) with 20 h cycles compared with that of Type 309 and some higher-alloyed Fe-Ni-Cr alloys (e.g., 800H and RA330) up to 500 h testing. Source: Ref 33

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

26 / High-Temperature Corrosion and Materials Applications

specimen was cathodically descaled, instead of weight gain. The oxidation resistance of RA330 along with other alloys in the 20–25Cr/30–40Ni group of Fe-Ni-Cr alloys is discussed in Section 3.4.8.

160 870 °C (1600 °F)

1090 °C (2000 °F)

120

1150 °C (2100 °F)

100 80 60

Weight gain, mg/cm2

140

980 °C (1800 °F)

40 20 0 RA 253 MA

RA 353 MA RA800H/AT

Fig. 3.26

Cyclic oxidation tests performed on 253MA, 353MA, and 800H at 871, 982, 1093, and 1150 °C (1600, 1800, 2000, and 2100 °F) for 1640 h with specimens cycling to room temperature every 160 h. Source: Ref 34

3

20 - 12Si

Weight increase, g/m2 h

18 - 10Ti

253 MA

2 153 MA

25 - 20

1

Alloy 601

353 MA 0 1000

1100

1200

Temperature, °C Isothermal oxidation rates (g/m2 h) obtained by testing for 45 h at various temperatures for 253MA and 353MA compared with some austenitic stainless steels, such as 18-10Ti (Type 321), 20-12Si (DIN 1,4828, 20Cr-12Ni-2Si), 2520 (Type 310), and nickel alloy 601. 1.0 g/m2 = 0.1 mg/cm2. Source: Ref 35

Fig. 3.27

RA85H was better than alloy 800H, but not as good as some higher-alloyed Fe-Ni-Cr alloys (e.g., RA330, 353MA) at 1150 °C (2100 °F) in air with 164 h cycles (Ref 33). In a long-term cyclic oxidation test in air at 1090 °C (2000 °F) for times up to 1 year with specimens cycling to room temperature after every 20 h of exposure, both 253MA and 85H were not as good as RA330, as shown in Table 3.9 (Ref 36). The oxidation resistance in this test was measured by metal loss, which was determined after the

3.4.7 Fe-Cr-Al Alloys Aluminum is a very effective alloying element in improving the resistance of the alloy to oxidation and other high-temperature corrosion attack. The alloy normally requires about 4% Al to form a continuous Al2O3 scale. The Al2O3 scale provides exceptional protection against oxidation attack. Figure 3.29 shows parabolic rate constants of some important oxides, such as Al2O3, Cr2O3, SiO2, FeO, NiO, and CoO (Ref 2). Al2O3 exhibits the lowest parabolic rate constants. When the metal is heated to 1200 °C (2200 °F) and higher, the Cr2O3 scale, which exhibits high growth rates as well as forming volatile CrO3, becomes essentially nonprotective. Under these very high temperature conditions, Al2O3 scale provides excellent protection against oxidation. Because of extremely slow growth rates at low and intermediate temperatures, Al2O3 scale provides less protection at these temperatures. As a result, high-temperature alloys that are designed to form Al2O3 scale for very high temperatures also contain adequate chromium to form Cr2O3 scale for intermediate temperatures. Some commercial electrical resistance heating elements are made of Fe-Cr-Al alloys, such as Kanthal® alloys, which rely on the formation of the Al2O3 scale for applications up to 1400 °C (2550 °F) (Ref 37). For example, some of the Kanthal alloys that are available in wire, strip, and ribbon product forms are Kanthal A-1 (Fe-22Cr-5.8Al), AF (Fe-22Cr-5.3Al), and D (Fe-22Cr-4.8Al). Since these wrought alloy products are essentially ferritic alloys, they exhibit low creep-rupture strengths when the temperature exceeds 650 °C (1200 °F) and cannot be used for high-temperature structural components. Thus, the electrical resistance heating elements made of these alloys have to be properly supported to avoid creep deformation, such as sagging. These Kanthal wires can be used in arc or ﬂame spraying to produce an oxidationresistant coating or in weld overlay cladding by using gas metal arc welding (GMAW) process. A powder metallurgy (P/M) process was used to produce a new alloy product, Kanthal APM. This P/M-produced Kanthal APM alloy was

pg 26

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:44PM Plate # 0

pg 27

Chapter 3: Oxidation / 27

120 1150 °C (2100 °F) 164 h cycles in air 164 100

80

80

74

Weight gain, mg/cm2

72

60

40

36

23 20

20

RA85H

RA310

RA330

RA333

RA 353MA

RA600

800H

Total exposure time, 1640 h

Fig. 3.28

Cyclic oxidation tests at 1150 °C (2100 °F) for 1640 h in air, with specimens cycling to room temperature every 164 h for Type 310, RA85H, 800H, 353MA, RA330, and nickel-base alloys 600 and RA333. Source: Ref 33

Table 3.9 Isothermal and cyclic oxidation tests in air at 1090 °C (2000 °F) Metal loss, μm (mils) Alloy

Exposure time

Cyclic(a)

Isothermal

85H

3000 h 1 year 3000 h 1 year 3000 h 1 year 3000 h

883 (34.8) 2665 (105) 1033 (40.7) 3730 (147) 683 (26.9) 2005 (79) …

41 (1.63) … 22 (0.88) … 44 (1.74) … 34 (1.34)

253MA RA330 310

(a) Specimens were cycled to room temperature after every 20 h of exposure. Source: Ref 36

reported to exhibit higher creep-rupture strengths (Ref 38). Other commercial Fe-Cr-Al alloys include ALFA-I™ (Fe-13Cr-3Al), ALFA-II™ (Fe-13Cr4Al), and ALFA-IV™ (Fe-20Cr-5Al-Ce) developed by Allegheny Ludlum (Ref 39), and

Fecralloy® (Fe-16Cr-4Al-0.3Y) developed by U.K. Atomic Energy Authority (Ref 40). Allegheny Ludlum (Ref 39) conducted cyclic oxidation tests in air by repeatedly resistance heating a thin ALFA IV foil (0.050 mm, or 0.02 in., thick) to various temperatures for 2 min until it failed. The cycles to failure of the foil are summarized in Table 3.10. No deﬁnition of the cycle to failure was given in the report. It is believed that the failure of the foil was deﬁned as when oxidation penetrated through the thin foil. ALFA IV uses a rare earth element, cerium, for enhancement of the Al2O3 scale. Very few studies have been conducted on the effectiveness of cerium on the adhesion of the Al2O3 scale. More studies have been conducted on the effects of yttrium, zirconium, and other reactive elements. In the Fecralloy alloy, the rare earth element yttrium is added to increase the adhesion of the

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 28

28 / High-Temperature Corrosion and Materials Applications

Al2O3 scale. Yttrium and reactive elements such as zirconium have been found to increase the adhesion of the Al2O3 scale formed on Fe-Cr-Al alloys, thus improving the oxidation resistance of the alloy (Ref 41–44). An Fe-Cr-Al-Y alloy can be strengthened by the oxide-dispersion strengthening (ODS) mechanism to signiﬁcantly increase its elevated-temperature strengths by the mechanical alloying process (Ref 45). One such commercial ODS alloy is MA956 (Fe-20Cr4.5Al-0.5Y2O3). Oxidation resistance of some ODS alloys is discussed in Section 3.4.10 on ODS alloys as well as in Section 3.4.12 on oxidation in high-velocity combustion gas streams.

Temperature, K 1350

1400

1300

1250 FeO

Log k, g2 cm–4s–1

–6

Table 3.10 Cycles to failure of a thin foil (0.050 mm, or 0.002 in., thick) of ALFA IV repeatedly resistance heated to the indicated temperatures for 2 min in still air from room temperature

NiO

–10

Cr2O3 –12

Fig. 3.29

As the nickel content in the Fe-Ni-Cr system increases from austenitic stainless steels to a group of iron-base alloys with 20–25Cr and 30–40Ni, the alloys become more stable in terms of metallurgical structure and more resistant to creep deformation (i.e., higher creep-rupture strengths). In general, this group of alloys also exhibits better oxidation resistance. Some of the wrought alloys in this group are 800H/800HT (Fe-21Cr-32Ni-Al-Ti), RA330 (Fe-19Cr-35Ni1.2Si), HR120 (Fe-25Cr-37Ni-0.7Nb-N), AC66 (Fe-27Cr-32Ni-0.8Nb-Ce), 353MA (Fe-25Cr35Ni-1.5Si-Ce), and 803 (Fe-26Cr-35Ni-Al-Ti). The oxidation resistance of 353MA compared with RA330, 800H, several stainless steels, and nickel-base alloys is shown in Fig. 3.26 to 3.28. Figure 3.30 shows the comparison of 353MA

CoO

–8

–14

3.4.8 Fe-Ni-Cr Alloys (20–25Cr/30–40Ni Alloys)

SiO2 AI2 O 3

Temperature, °C (°F)

7.0

7.8 7.6 7.4 1 × 104, K–1 T

7.2

8.0

8.2

Cycles to failure

1090 (2000) 1150 (2100) 1200 (2200) 1260 (2300)

Parabolic rate constants of several oxides. Source: Ref 2

2600 1000 460 200

Source: Ref 39

4 253MA

Weight change, mg/cm2

2 0 353MA HK 4M

–2 –4 –6 –8

Inco 803 –10

HP45-Nb 500

1500

2500

3500

Time, h

Fig. 3.30

Isothermal oxidation tests in air at 1000 °C (1830 °F) for 353MA, alloy 803, HK4M (Fe-25Cr-25Ni-0.3Al-0.4Ti), and HP45Nb (Fe-24Cr-37Ni-1.4Si-1.2Nb). Source: Ref 46

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 29

Chapter 3: Oxidation / 29

and 803 in comparison with some casting alloys with similar chromium and nickel contents when tested in air at 1000 °C (1830 °F) (Ref 46). Additional oxidation data on 353MA, 803, HR120, 800H, and RA330 are summarized in Fig. 3.31 to 3.35 (Ref 47–49). Particular emphasis should be placed on long-term oxidation tests (Fig. 3.33 and 3.35), where some alloys are shown to suffer breakaway oxidation after some incubation time of relatively little or mild weight losses. More discussion on breakaway oxidation is presented in Section 3.4.13. 3.4.9 Ni-Cr/Co-Cr Superalloys In many Ni-Cr alloys, many alloying elements, such as those for solid-solution strengthening (e.g., Mo, W) and precipitation strengthening (e.g., Al, Ti, Nb), are added into the alloys to provide strengthening of the alloy at elevated temperatures. Many of these alloys are commonly referred to as “superalloys.” The superalloys also include oxide dispersion strengthened (ODS) alloys, which are brieﬂy discussed in Section 3.4.10. Similar to Fe-Cr-Al alloys, aluminum is also used as an alloying element in Ni-Cr alloys to improve the oxidation resistance. Although it generally requires a minimum of 4% Al in order

for a Ni-Cr alloy to form a protective Al2O3 scale, the addition of less than 4% Al can also signiﬁcantly improve the oxidation resistance of the alloy. Alloy 601 with only about 1.3% Al shows excellent oxidation resistance, as shown in Fig. 3.32 to 3.34 and Fig. 3.36 to 3.38. In both Fig. 3.37 and 3.38, alloy 601GC was included in the testing to compare with alloy 601. Alloy 601GC is essentially alloy 601 with small additions of titanium, zirconium, and nitrogen for grain-size control involving nitrides and carbonitrides. The development of alloy 601GC was intended to prevent excessive grain coarsening when exposed to 1100 °C (2010 °F) and higher with the formation of nitrides and carbonitrides of zirconium, titanium, and Zr + Ti (Ref 51). In long-term oxidation testing (for up to close to 500 days) in air at 1000 °C (1830 °F), alloy 601GC was found to be better than alloy 601, as shown in Fig. 3.38. The improvement of oxidation resistance in alloy 601GC over alloy 601 may be the result of ﬁner grain sizes as well as the presence of zirconium, a reactive element that has been found to be beneﬁcial in improving oxidation resistance for Fe-Cr-Al alloys by several investigators cited previously. Even though the presence of about 1.4% Al in alloy 601 is beneﬁcial in improving oxidation resistance, the oxide scales formed on alloy 601

50 N08811

R30566

40

Weight gain, mg/cm2

N08330 30 S35315 20

N06333 HR-120

10

0

0

500

1000

1500

2000

2500

3000

Time, h

Fig. 3.31

Cyclic oxidation tests for 353MA (S35315), HR120, RA330 (N08330), 800HT (N08811), RA333 (N06333), and 556 (R30556) in air at 1090 °C (2000 °F) with specimens cycling to room temperature once a week. Source: Ref 47

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 30

30 / High-Temperature Corrosion and Materials Applications

350

300

N08810

Weight gain, mg/cm2

250 HR-120 200

150

100

N06601

50

0

0

500

1000

1500

2000

N06045

2500

S35315

3000

Time, h

Fig. 3.32

Cyclic oxidation tests for 353MA (S35315), HR120, 800H (N08810), 601 (N06601), and 45TM (N06045) in air at 1150 °C (2100 °F) with specimens cycling to room temperature once a week. Source: Ref 47

100

100 Incoloy alloy DS

UNS N06601

0

0

UNS N06601

–100 Incoloy alloy 803

–200 –300

UNS N06600

–400 –500

UNS N08330

–600 UNS N08810

–700

Mass change, mg/cm2

Mass change, mg/cm2

Incoloy alloy DS

–100

UNS N06600

–200

UNS N08330

Incoloy alloy 803

–300 – 400 – 500 UNS N08810

–800 0

5000

10,000

15,000

20,000

– 600 0

Exposure time, h

Fig. 3.33

Oxidation tests for 803, RA330 (N08330), 800H (N08810), alloy DS, and 600 (N06600) and 601 (N06601) in air + 5% H2O at 1000 °C (1830 °F). Source: Ref 48

are primarily chromium-rich oxides. Chromiumrich oxide scales are prone to scaling, cracking, and spalling, particularly under cyclic conditions at very high temperatures, such as 1100 °C (2010 °F) and higher. This is illustrated in Fig. 3.39 for alloy 601 tested in air for 1056 h at 850 °C (1560 °F), 1000 °C (1830 °F), 1100 °C (2010 °F), and 1200 °C (2190 °F) (Ref. 52). For testing at 850, 1000, and 1100 °C, specimens were cycled to room temperature from the test

1000

2000

3000

4000

5000

6000

Exposure time, h

Fig. 3.34

Oxidation tests for 803, RA330 (N08330), 800H (N08810), and alloy DS in compared with Ni-Cr alloys 600 (N06600) and 601 (N06601) in air + 5% H2O at 1100 °C (2010 °F). Source: Ref 48

temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed from the hot zone every 16 h. In addition to scaling and spalling of external oxide scales when exposed to 1100 and 1200 °C (2010 and 2190 °F), signiﬁcant internal oxide penetration was also observed. However, with slightly higher aluminum content (i.e., about 2%), the

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 31

Chapter 3: Oxidation / 31

0

HR120

Weight change, mg/cm

2

–50.0

–100.0

RA85H –150.0

800HT

–200.0 0

90

180

270

360

450

540

630

720

Time, days

Fig. 3.35

Oxidation tests for HR120, 800H, and RA85H in still air at 980 °C (1800 °F) for times up to 720 days with specimens being cooled to room temperature every 30 days for weight measurement. Source: Ref 49

alloy can signiﬁcantly reduce its scaling and spalling of external oxide scales as well as internal oxide penetration. This is illustrated in Fig. 3.40, which shows the oxide scales formed on alloy 602CA (Ni-25Cr-9Fe-2.2Al-0.2C0.06Zr-0.08Y) after the same test conditions as alloy 601 as shown in Fig. 3.39. The improved oxidation resistance of alloy 602CA may also be attributed to the presence of yttrium and zirconium. The beneﬁcial effects of yttrium and zirconium along with other reactive elements in aluminum-containing alloys are discussed later in the chapter. The cyclic oxidation resistance of alloy 602CA compared with alloy 800H and alloy 601 as well as several nickel- and cobalt-base superalloys at temperatures from 750 to 1200 °C (1380 to 2190 °F) are summarized in Table 3.11 (Ref 53). With further increase in aluminum to 2.8% in the same alloy composition series, alloy 603GT (Ni-25Cr-9Fe-2.8Al-0.22C0.1Zr-0.1Y) exhibits more compact, adherent oxide scales (Fig. 3.41) under the same test conditions as Fig. 3.40 (Ref 52). Ni-Cr alloys containing about 4% Al or higher form a very protective Al2O3 scale when heated to very high temperatures. This is illustrated in

Fig. 3.42, where alloy 214 (Ni-16Cr-4.5Al-Y) was compared with alloy 601 and alloy 800H in cyclic oxidation tests performed in still air at 1150 °C (2100 °F) with specimens cycling to room temperature once a day except weekends (Ref 54). Alloy 214 showed essentially no weight loss after 42 days of testing, while alloy 601 suffered a linear weight loss. Figure 3.43 shows the cross section of the specimen between alloy 214 and alloy 601 oxidation tested in air at 1090 °C (2000 °F) for 1008 h. The Al2O3 oxide scale formed on alloy 214 remains very protective even when the temperature is increased to close to its incipient melting point, as shown in Fig. 3.44 and 3.45 (Ref 55). Figure 3.44 shows the alloy 214 coupon after exposure to ﬂowing air at 1320 °C (2400 °F) for 200 h with the specimen being cycled to room temperature every 24 h. The incipient melting temperature for alloy 214 is believed to be slightly higher than 1345 °C (2450 °F) (Ref 56). The specimen showed no evidence of oxidation attack (Fig. 3.44). SEM/EDX examination of the cross section of the tested specimen showed a thin aluminum-rich oxide scale formed on the metal surface (Fig. 3.45).

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 32

32 / High-Temperature Corrosion and Materials Applications

As shown in Fig. 3.29, in addition to Cr2O3 and Al2O3, SiO2 also exhibits very low parabolic rate constants and can provide an effective barrier to oxidation attack at very high temperatures.

Fig. 3.36

Cyclic oxidation resistance of alloy 601 (Ni-23Cr14Fe-1.4Al) compared with alloy 600 (Ni-16Cr8Fe) and alloy 800 (Fe-20Cr-32Ni-0.4Al-0.4Ti) in air at 1090 °C (2000 °F) with specimens being in hot zone for 15 min and out of hot zone for 5 min. Source: Ref 50

Birks and Pettit (Ref 57) suggested that both Al2O3 and SiO2 are capable of providing adequate oxidation resistance above 1200 °C (2200 °F). There are a number of commercial high-temperature alloys that contain relatively high levels of silicon for improving hightemperature corrosion resistance. However, the amount of silicon addition to an alloy that can be manufactured into a wrought product or casting product is quite limited. Silicon can make an alloy difﬁcult to cast and also very difﬁcult to weld. For some commercial, wrought alloys, silicon levels are up to 3.5%. The formation of a continuous, external SiO2 scale was not observed for these commercial Fe-Ni-Cr alloys or Ni-Cr alloys in a way that a continuous, external Al2O3 scale forms on Fe-Cr-Al alloys or Ni-Cr-Al alloys. A Ni-Cr-Co-Si alloy, HR160, was developed in late 1980s for applications in severe sulﬁdizing environments (Ref 58). The oxidation resistance of HR160 alloy (Ni-28Cr-30Co2.75Si-0.5Ti-0.5Nb) was found to be quite comparable to alloy 601 and signiﬁcantly better than alloy 800HT when tested for time up to about 1 year in air at 1090 °C (2000 °F), as shown in Fig. 3.46 (Ref 59). However, in contrast to Ni-Cr-Al alloy 214 that forms an external Al2O3 with essentially no internal oxide or void formation when exposed to very high temperatures, such as 1200 °C (2200 °F), HR160 forms external Cr2O3/SiO2 oxides scales with internal oxide/void formation (Ref 49). Figure 3.47 shows oxidation data for alloys 214 and HR160 along with several other nickel- and iron-base alloys. The oxidation tests were conducted at

100

Mass change, mg/cm2

0 601 –100

601GC

–200

330

–300 RA85H –400

0

50

100

150

200

Exposure time, days

Fig. 3.37

Oxidation tests in air + 5% H2O at 1100 °C (2010 °F) for up to 200 days for alloys 601 (Ni-23Cr-14Fe-1.4Al), 601GC (Ni-23Cr-14Fe-1.3Al-0.2Ti-0.2Zr-0.05N), RA85H (Fe-18Cr-14Ni-3.5Si-1Al), and RA330 (Fe-19Cr-35Ni-1.2Si). Specimens were cooled to room temperature and weighed at the indicated data points. Source: Ref 51

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 33

Chapter 3: Oxidation / 33

10

Mass change, mg/cm2

0

–10

601GC

–20

–30 330 601

–40

–50 0

100

300

200

400

500

600

Exposure time, days

Fig. 3.38

Oxidation tests in air + 5% H2O at 1000 °C (1830 °F) for times up to close to 500 days for alloys 601 (Ni-23Cr-14Fe-1.4Al), 601GC (Ni-23Cr-14Fe-1.3Al-0.2Ti-0.2Zr-0.05N), and RA330 (Fe-19Cr-35Ni-1.2Si). Specimens were cooled to room temperature and weighed at the indicated data points. Source: Ref 51

(a)

(b)

Fig. 3.39

20 µm

(c)

(d)

Alloy 601 tested in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F), (c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref 52. Courtesy of ThyssenKrupp VDM

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

34 / High-Temperature Corrosion and Materials Applications

(a)

20 µm

(c)

(d)

(b)

Fig. 3.40

Alloy 602CA (Ni-25Cr-9Fe-2.2Al-0.2C-0.06Zr-0.08Y) tested in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F), (c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM

Table 3.11 Weight change data (mg/m2h) from cyclic oxidation tests in air for 1200 h at indicated temperatures Test temperature Alloy

602CA X 800H 625 601 617 188

750 °C (1380 °F)

850 °C (1560 °F)

1000 °C (1830 °F)

1100 °C (2010 °F)

1200 °C (2190 °F)

+0.4 +1 +7 +1 +1 +4 +1

+3 +8 +8 +6 +10 +12 +4

+12 +5 −24 −100 +7 +19 +7

+7 −5 −162 −1410 −24 −19 −302

−310 … … … −820 … …

Note: Specimens were held at the test temperature for 16 h followed by cooling to room temperature. For testing at 750 to 1100 °C, specimens were cycled by furnace cooling and furnace heat up (about 1.5 hours heat up). Cycling for 1200 °C testing involved air cooling and inserting the specimen directly to the furnace hot zone. Source: Ref 53

1200 °C (2200 °F) in air for almost 1 year. HR160 showed extensive internal attack with relatively little metal loss due to external oxidation, compared with the alumina-former alloy

214. The internal attack observed in alloy HR160 was caused by formation of internal oxides and voids. Alloy 214, on the other hand, showed essentially no metal loss due to the formation of a compact external Al2O3 scale. Also shown in Fig. 3.47 is a high-silicon ironbase alloy, RA85H (Fe-19Cr-15Ni-3.5Si-1Al), suffering much more extensive internal attack under the same test conditions. The oxidation resistance data for alloy RA85H is also shown in Table 3.9 and Fig. 3.28 and 3.37. Figure 3.48 shows the oxide morphology of a high-silicon Ni-Cr-Fe alloy, 45TM (Ni-27Cr-23Fe-2.7Si), after oxidation testing in air for 1154 h at 850 to 1200 °C (1560 to 2190 °F) (Ref 52). Extensive internal void formation was observed in the specimen tested at 1100 °C (2010 °F). The oxidation data in comparing 45TM with other high-temperature alloys is shown in Fig. 3.32.

pg 34

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

Chapter 3: Oxidation / 35

(a)

20 µm

(b)

(c)

(d)

Fig. 3.41

Alloy 603GT (Ni-25Cr-9Fe-2.8Al-0.22C-0.13Zr-0.01Y) tested in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F), (c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM

In addition to maximizing the oxidation resistance by adjusting the level of chromium and/or aluminum, or silicon, a majority of high-temperature alloys have been developed to attain elevated-temperature strengths by alloying with many other elements. A large number of superalloys have been developed in response to the demands of gas turbine engines for critical operating conditions involving high stresses and high temperatures. In response to the demands for high stresses at intermediate temperatures, one group of wrought superalloys is strengthened by precipitation strengthening with Ni3X precipitates (X represents aluminum, titanium, niobium, etc.) along with solid-solution strengthening using molybdenum or tungsten. These alloys include 718, R-41, X-750, Waspalloy, and Nimonic 80A. Some of the applications for these alloys in gas turbines include compressors, diffusers, turbine disks and cases, heat shields, exhaust system, thrust reversers, and turbine shroud rings. Most alloys in this group are used in the heat treated conditions to

take advantage of the precipitation strengthening. Most heat treating procedures are performed in the intermediate temperature range. As a result, the applications for these alloys are in the intermediate temperature range to prevent overaging of the strengthened precipitates. Oxidation of these alloys at intermediate temperatures generally does not present a signiﬁcant issue in terms of the performance. Another important group of wrought superalloys is generally classiﬁed as solidsolution-strengthened alloys. Solid-solutionstrengthening alloying elements typically are molybdenum and tungsten. These alloys are also strengthened with carbides. This group of alloys is typically used in stationary components with lower mechanical stresses than rotating parts such as disks and blades. However, operating temperatures for these alloys are generally higher. Typical applications in gas turbines are combustors, transition ducts, and afterburners. The alloys require good combined properties of creep and fatigue strengths, fabricability,

pg 35

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 36

36 / High-Temperature Corrosion and Materials Applications

weldability, thermal stability, and oxidation resistance. Some of these alloys are being used increasingly in nongas turbine industries. +20 214 0

–20

Weight change, mg/cm2

–40

–60 601 –80

–100

–120

–140 800H

–160

–180

–200

10

20

30

40

50

Exposure time, days

Fig. 3.42

Cyclic oxidation resistance of alloy 214 compared to alloys 601 and 800H in still air at 1150 °C (2100 °F) cycled once a day every day except weekends. Source: Ref 54

sample surface

Fig. 3.43

For cast nickel-base superalloys, substantial amounts of aluminum and titanium are used to produce a large volume fraction of gamma prime (γ′) precipitates to further increase strengthening of the alloy. They also contain large amounts of refractory elements, such as molybdenum and tungsten for solid-solution strengthening, along with boron, zirconium, carbon, and hafnium for grain-boundary strengthening. Since these alloys do not require hot and cold working during manufacturing, they can be designed to contain maximum amounts of these alloying elements to attain the maximum strength requirements. These cast nickel-base superalloys include IN713C, IN713LC, IN738X, IN100, B-1900, Rene 80, IN792, MAR-M246, and MAR-M247. Lacking precipitation-strengthening phases in cobalt-base alloys, cast cobalt-base superalloys derive their strengths from solid-solution strengthening as well as carbide strengthening. These alloys typically contain high levels of carbon. Some of these alloys are X-40 (or alloy 31), MAR-M509, WI-52, and MAR-M302. They are widely used for high-pressure vanes. Another group of superalloys are referred to as the ODS alloys. These alloys are produced by the mechanical alloying process and are strengthened by oxide dispersoids (Ref 45). MA956 and MA754 were originally developed for gas turbine combustors and stator vanes, respectively (Ref 60), and MA6000 was developed for rotor blades (Ref 61). ODS alloys are much more difﬁcult to fabricate than conventional wrought alloys. Joining can present a particularly signiﬁcant challenge to this group of alloys. ODS

sample surface

Cross sections of the specimens for alloy 601 (a) and alloy 214 (b) after oxidation testing in flowing air at 1090 °C (2000 °F) for 1008 h (samples were cycled to room temperature once every week). Samples were cathodically descaled to remove oxide scale prior to metallographic mounting. Top edge of the micrograph represents the original specimen surface. Source: Ref 55

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 37

Chapter 3: Oxidation / 37

HR160

0 Weight change, mg/cm2

–100 601

–200 –300 –400 –500 –600 –700 –800

800HT

–900 100

200 Time, days

300

400

Fig. 3.46

Fig. 3.44

Alloy 214 specimen tested in air at 1320 °C (2400 °F), which was about 25 °C below the incipient melting point of the alloy, for 200 h with the specimen being cycled to room temperature every 24 h. Source: Ref 55

Fig. 3.45

Scanning electron micrograph showing the adherent aluminum-rich oxide scale formed on alloy 214 after exposure in flowing air at 1320 °C (2400 °F) for 200 h with the specimen being cycled to room temperature every 24 h. EDX analysis was performed at three different locations, marked 1, 2, and 3, in the aluminum oxide scale. The results of the analysis (wt%) are shown in: Area. 1: 69.9% Al, 17.0% Cr, 10.2% Ni, 1.1% Fe, and 1.8% Zr. Area 2: 97.8% Al, 1.3% Cr, 0.7% Ni, and 0.2% Zr. Area 3: 98.3% Al, 0.2% Cr, and 1.5% Ni. Source: Ref 55

alloys are brieﬂy described in Section 3.4.10. More information about superalloys for gas turbines is available in Ref 62 to 66. Long-term oxidation data in air at an intermediate temperature were generated by Barrett (Ref 67) for 33 alloys, from ferritic stainless steels to superalloys. Tests were performed at 815 °C (1500 °F) for 10,000 h with 1000 h cycles (a total of 10 cycles). The results are shown in Fig. 3.49. Alloys, which contain no or

Oxidation resistance of HR160 alloy (Ni-28Cr30Co-2.75Si-0.5Ti-0.5Nb) compared with alloys 601 and 800HT when tested for up to about 1 year in air at 1090 °C (2000 °F). Source: Ref 59

low chromium content, such as nickel-base alloy B (Ni-28Mo) and alloy N (Ni-7Cr-16.5Mo) and Type 409 (Fe-11Cr), can suffer severe oxidation. It was surprising to ﬁnd that Type 321 exhibited poor oxidation resistance compared with Types 304, 316, and 347. The author offered no explanation in the paper. As is discussed in other sections, austenitic stainless steels, such as Types 304, 347, and 321, containing a borderline level of chromium, are very sensitive to the chromium content for the bulk chemistry as well as the surface chemistry in the alloy. When the surface depletion of chromium occurs for a stainless steel product whose bulk chromium content is at the low end of the speciﬁcation, breakaway oxidation is likely to take place, thus resulting in severe oxidation attack. The majority of the oxidation data reported so far has been presented in terms of weight changes (mg/cm2) as a function of times or temperatures. However, it is impossible to use the weight change data (either weight gain or weight loss) to estimate the life of the component due to oxidation attack. The oxidation data that is of engineering importance is the depth of oxidation attack, which includes the depth of metal loss and the depth of internal oxidation attack. The total depth of the oxidation attack (or the depth of the metal affected) is responsible for the reduction of the load-bearing capability for the component. A large oxidation database in terms of the depth of oxidation attack for commercial alloys from stainless steels to high-alloy Fe-Ni-Cr alloys (20–25Cr/30–40Ni alloys), and nickel- and

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

38 / High-Temperature Corrosion and Materials Applications

Average metal affected, mm

8 7

>6.35

>6.35

>6.35

>6.35

RA85H

617

HR120

800HT

Average internal attack Metal loss

6 5 4 3 2 1 0

214

HR160

230

601

Fig. 3.47 Oxidation data in terms of metal loss, resulting from external oxide scales, and internal attack, resulting from internal oxide and/or void formation, for alumina-former alloy 214 and chromia/silica-former alloy HR160 along with several other nickeland iron-base alloys, generated at 1200 °C (2200 °F) in air for 360 days. Source: Ref 49

(a)

20 µm

(b)

(c)

(d)

Fig. 3.48

A high-silicon Ni-Cr-Fe alloy, 45TM (Ni-27Cr-23Fe-2.7Si), after oxidation testing in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F), (c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM

cobalt-base superalloys, is summarized in Table 3.12 (Ref 68). Tests were conducted in ﬂowing air (30 cm/min) at 980, 1095, 1150, and 1200 °C (1800, 2000, 2100, and 2200 °F) for

1008 h. The specimens were cooled to room temperature for visual examination once a week (168 h). Specimens from sheet products were surface ground to maintain the uniform surface

pg 38

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

Chapter 3: Oxidation / 39

Fig. 3.49

Long-term oxidation tests (10,000 h) in air at 815 °C (1500 °F) with 1000 h cycles (a total of 10 cycles to room temperature for the entire test) for iron-, nickel-, and cobalt-base alloys. Also included is the upper limit of the metal loss for isothermal tests (i.e., 10,000 h without cycling to room temperature) for same alloys. Source: Ref 67

condition for all test specimens. Some of the observations of the data are summarized:

Types 304 and 316 specimens were consumed completely at 1095, 1150, and 1200 °C (2000, 2100, and 2200 °F). Type 304 was found to be much more resistant than Type 316 at 980 °C (1800 °F). Type 446 was not as good as Type 310 at all test temperatures. Type 446 specimens were consumed at 1150 and 1205 °C (2100 and 2200 °F). Type 310 appeared to be slightly better than RA330 and 800H. For applications at high temperatures, many superalloys contain numerous alloying elements for increasing the elevated-temperature strength of the alloy. Molybdenum and tungsten are common alloying elements for providing solidsolution strengthening for increasing the creeprupture strength of the alloy. Two iron-base superalloys, Multimet alloy (Fe-20Ni-20Co21Cr-3Mo-2.5W-1.0Nb+Ta) and alloy 556

(Fe-20Ni-18Co-22Cr-3Mo-2.5W-0.6Ta-0.02La0.02Zr), are good examples. However, the oxides of both molybdenum and tungsten (MO3 and WO3) exhibit high vapor pressures at very high temperatures, as shown in Fig. 3.3. Multimet alloy suffered rapid oxidation attack at 1150 and 1200 °C (2100 and 2200 °F), with specimens completely consumed at both temperatures. However, formation of the volatile oxides of MO3 and WO3 can be minimized by modiﬁcation of some key alloying elements in Multimet alloy. The development of alloy 556 was aimed at improving the oxidation resistance of Multimet alloy without losing the elevated-temperature strength by making some modiﬁcation of alloying elements in Multimet alloy. The modiﬁcation involved a slight increase in chromium, a decrease in cobalt, replacement of niobium with tantalum, and addition of a rare-earth element, lanthanum, and a reactive element, zirconium, but the same amounts of molybdenum and tungsten were kept. The result was a much more oxidation-resistant alloy, alloy 556, at 1095

pg 39

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 40

40 / High-Temperature Corrosion and Materials Applications

and 1150 °C (2000 and 2100 °F), although the alloy still suffered rapid oxidation at 1200 °C (2200 °F). As shown in the Table 3.12 nickel- and cobaltbase alloys containing molybdenum and/or tungsten were also found to suffer rapid oxidation at very high temperatures (i.e., specimens were consumed during the tests). Specimens of nickel-base alloys that were consumed at 1200 °C (2200 °F) were alloy S (14% Mo), alloy X (9% Mo, 0.6% W), and alloy 625 (9% Mo, 3.5% Nb). Some of those nickel-base alloys containing molybdenum and/or tungsten that were not consumed at 1200 °C (2200 °F) were alloy 230 (14% W), alloy 617 (9% Mo), and RA333 (3% Mo, 3% W). From these two different sets of oxidation behavior at very high temperatures, one can design nickel-base alloys (relying on chromium oxide scales) containing molybdenum and tungsten for elevatedtemperature strengthening to resist oxidation resistance at very high temperatures by adjusting other alloying elements.

For nickel-base alloys containing high levels of molybdenum and/or tungsten, it is believed that increasing chromium is probably the most important factor in suppressing rapid oxidation involving molybdenum and/or tungsten. Some nickel-base precipitation-strengthened alloys containing high titanium as well as molybdenum that were consumed at 1200 °C (2200 °F) were Waspaloy (4.3% Mo, 3.0 Ti), René 41 (10% Mo, 3.0 Ti), and alloy 263 (6% Mo, 2.2Ti). Titanium was found to be very active in oxide-scale formation. Figure 3.50 illustrates the oxide scale formed on alloy 263 after exposure to air for 1 h at 1200 °C (2200 °F), showing mainly titanium-rich oxides and Cr-Ti oxides. The formation of titanium-rich oxides apparently disrupts the Cr2O3 scale. Nagai et al. (Ref 69) found that titanium was detrimental to the oxidation resistance of Ni-20Cr alloy. In the Fe-CrAl system, however, the addition of 1% Ti to Fe18Cr-6Al was found to improve resistance in cyclic oxidation in air at 950 °C (1740 °F) (Ref 70). It is not clear whether the beneﬁcial

Table 3.12 Results of oxidation tests for various alloys at indicated temperatures in flowing air (30 cm/min) for 1008 h 980 °C (1800 °F)

Alloy

Metal loss, mm (mils)

1095 °C (2000 °F)

Average metal affected, mm (mils)

Metal loss, mm (mils)

1150 °C (2100 °F)

Average metal affected, mm (mils)

Metal loss, mm (mils)

1205 °C (2200 °F)

Average metal affected, mm (mils)

Metal loss, mm (mils)

Average metal affected, mm (mils)

214

0.0025

(0.1) 0.005

(0.2)

0.0025

(0.1)

0.0025

(0.1)

0.005

(0.2)

0.0075

(0.3)

0.005

(0.2)

0.018

(0.7)

601 600

0.013 0.0075

(0.5) 0.033 (0.3) 0.023

(1.3) (0.9)

0.03 0.028

(1.2) (1.1)

0.067 0.041

(2.6) (1.6)

0.061 0.043

(2.4) (1.7)

0.135 0.074

(5.3) (2.9)

0.11 0.13

(4.4) (5.1)

0.19 0.21

(7.5) (8.9)

230 S 617 333 X 671

0.0075 0.005 0.0075 0.0075 0.0075 0.0229

(0.3) (0.2) (0.3) (0.3) (0.3) (0.9)

0.018 0.013 0.033 0.025 0.023 0.043

(0.7) (0.5) (1.3) (1.0) (0.9) (1.7)

0.013 0.01 0.015 0.025 0.038 0.038

(0.5) (0.4) (0.6) (1.0) (1.5) (1.5)

0.033 0.033 0.046 0.058 0.069 0.061

(1.3) (1.3) (1.8) (2.3) (2.7) (2.4)

0.058 0.025 0.028 0.05 0.11 0.066

(2.3) (1.0) (1.1) (2.0) (4.5) (2.6)

0.086 0.043 0.086 0.1 0.147 0.099

(3.4) 0.11 (1.7) >0.8I (3.4) 0.27 (4.0) 0.18 (5.8) >0.9 (3.9) 0.086

(4.5) 0.20 (31.7) >0.8I (10.6) 0.32 (7.1) 0.45 (35.4) >0.9 (3.4) 0.42

(7.9) (31.7) (12.5) (17.7) (35.4) (16.4)

625 Waspaloy R-4I 263

0.0075 0.0152

(0.3) 0.018 (0.6) 0.079

(0.7) (3.1)

0.084 0.036

(3.3) (1.4)

0.12 0.14

(4.8) (5.4)

0.41 0.079

(16.0) (3.1)

0.46 0.33

(18.2) >1.2 (13.0) >0.40

(47.6) >1.2 (15.9) >0.40

(47.6) (15.9)

0.0178 0.0178

(0.7) 0.122 (0.7) 0.145

(4.8) (5.7)

0.086 0.089

(3.4) (3.5)

0.30 0.36

(11.6) (14.2)

0.21 0.18

(8.2) (6.9)

0.44 0.41

(17.4) >0.73 (16.1) >0.91

(28.6) >0.73 (35.7) >0.91

(28.6) (35.7)

188 25 150 6B

0.005 0.01 0.01 0.01

(0.2) (0.4) (0.4) (0.4)

0.015 0.018 0.025 0.025

(0.6) (0.7) (1.0) (1.0)

0.01 0.23 0.058 0.35

(0.4) (9.2) (2.3) (13.7)

0.033 0.26 0.097 0.39

(1.3) 0.18 (10.2) 0.43 (3.8) >0.68 (15.2) >0.94

(7.2) 0.2 (16.8) 0.49 (26.8) >0.68 (36.9) >0.94

(8.0) (19.2) (26.8) (36.9)

(21.7) (37.9) (46.1) (36.8)

(21.7) (37.9) (46.1) (36.8)

556 Multimet

0.01 0.01

(0.4) 0.028 (0.4) 0.033

(1.1) (1.3)

0.025 0.226

(1.0) (8.9)

0.067 0.29

(2.6) 0.24 (11.6) >1.2

(9.3) 0.29 (47.2) >1.2

(11.6) >3.8 (47.2) >3.7

800H RA330 310 316 304 446

0.023 0.01 0.01 0.315 0.14 0.033

(0.9) (0.4) (0.4) (12.4) (5.5) (1.3)

0.046 (1.8) 0.14 0.11 (4.3) 0.02 0.028 (1.1) 0.025 0.36 (14.3) >1.7 0.21 (8.1) >0.69 0.058 (2.3) 0.33

(5.4) 0.19 (0.8) 0.17 (1.0) 0.058 (68.4) >1.7 (27.1) >0.69 (13.1) 0.37

(7.4) 0.19 (7.5) 0.23 (6.7) 0.041 (1.6) 0.22 (2.3) 0.075 (3.0) 0.11 (68.4) >2.7 (105.0) >2.7 (27.1) >0.6 (23.6) >0.6 (14.5) >0.55 (21.7) >0.55

>0.55 >0.96 >1.I7 >0.94

>0.55 >0.96 >1.I7 >0.94

(150.0) >3.8 (146.4) >3.7

(8.9) 0.29 (11.3) 0.35 (8.7) 0.096 (3.8) 0.21 (4.4) 0.2 (8.0) 0.26 (105.0) >3.57 (140.4) >3.57 (23.6) >1.7 (68.0) >1.73 (21.7) >0.59 (23.3) >0.59

(150.0) (146.4) (13.6) (8.3) (10.3) (140.4) (68.0) (23.3)

Note: 3304 cm3/min of ﬂow rate in a 1.75 in. diam furnace tube. The moisture was removed from the air by a ﬁlter prior to entering into the furnace tube. Specimens were cathodically descaled for measurement of the metal loss. The average metal affected is the sum of the metal loss and the depth of internal attack. The depth of internal attack was measured by metallography. Source: Ref 68

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 41

Chapter 3: Oxidation / 41

effect of titanium for oxidation resistance is only for alumina formers such as in this case (Fe18Cr-6Al), but not for chromia formers in Ni20Cr alloy. Niobium is another alloying element that may be detrimental to alloy oxidation resistance at very high temperatures. The relatively poor oxidation resistance of alloy 625 at 1095 and 1150 °C (2000 and 2100 °F) can be attributed to niobium. Cobalt-base alloys with tungsten, such as alloy 188 (Co-22Cr-22Ni-14W-0.04La), alloy 25 (Co-20Cr-10Ni-15W), and alloy 6B (Co-30Cr4.5W-1.2C), suffered rapid oxidation at 1205 °C (2200 °F). A cobalt-base alloy, alloy 150 (Co27Cr-18Fe), containing no tungsten also suffered rapid oxidation attack at 1205 °C (2200 °F). Again, the oxidation of a cobalt-base alloy can be signiﬁcantly improved with some modiﬁcation of alloying elements. Alloy 25 with 15% W exhibits excellent creep-rupture strengths at high temperatures. However, because of the high level of tungsten, the alloy suffers high oxidation rates at very high temperatures, such as 1095 and 1150 °C (2000 and 2100 °F). With slight increase in chromium and nickel along with the addition of lanthanum, the result of the modiﬁcation was alloy 188. As shown in Table 3.12, alloy 188 exhibits signiﬁcantly better oxidation resistance than alloy 25 at 1095 and 1150 °C (2000 and 2100 °F). The best alloy among those investigated was an alumina former, alloy 214 (Ni-16Cr-4.5Al-Y). The depth of oxidation attack (average metal affected) was found to be less than 0.025 mm (1.0 mils) after 1008 h at temperatures up to

Fig. 3.50 Scanning electron micrograph showing the early stage of oxidation in air at 1200 °C (2200 °F) for 1 hour for alloy 263, revealing titanium-rich and Cr-Ti oxides on the outermost oxide scale. Area 1: 28.1% Cr, 70.9% Ti, 0.8% Co, 0.2% Ni. Area 2: 57.0% Cr, 36.8% Ti, 2.8% Co, 25% Ni, 1.0% Fe.

1205 °C (2200 °F). The oxidation resistance of alloy 214 is also presented in Fig. 3.42 to 3.45 and 3.47. The oxidation data presented in Table 3.12 were generated with a weekly cycle (168 h). When the cyclic frequency was increased to 25 h cycles, oxidation rates were increased for all the alloys tested. However, some alloys are more sensitive to cyclic oxidation than others. The effect of thermal cycling on the oxidation resistance of various alloys in air at 1095 °C (2000 °F) is illustrated in Table 3.13, which compares once-a-week (168 h) cycle data (Ref 68) with 25 h cycle data (Ref 71). All the data are presented in terms of the average depth of metal affected, which represents the metal loss plus the depth of internal oxidation attack. Both sets of the data were generated under the same test conditions using the same test furnaces and test procedures except the differences in cyclic frequencies. Some chromia formers, such as alloys 230, S, 188, 556, and 310, showed good resistance to thermal cycling. A long-term oxidation test program was undertaken to test alloys up to 2 years at 980, 1095, and 1150 °C (1800, 2000, and 2100 °F)

Table 3.13 Comparative oxidation resistance of various alloys in flowing air between 168 h and 25 h cycles at 1095 °C (2000 °F) Total depth of attack, mm (mils)

Extrapolated oxidation rate, mm/yr (mpy)

Alloy

1008 h/168 h

1050 h/25 h

168 h cycles

25 h cycles

214 601 600 671 230 S G-30 617 RA333 625 Waspaloy 263 188 25 150 6B 556 Multimet 800H RA330 310 446

0.003 (0.1) 0.066 (2.6) 0.041 (1.6) 0.061 (2.4) 0.003 (1.3) 0.003 (1.3) 0.122 (4.8) 0.046 (1.8) 0.058 (2.3) 0.122 (4.8) 0.137 (5.4) 0.361 (14.2) 0.003 (1.3) 0.259 (10.2) 0.097 (3.8) 0.394(15.5) 0.066 (2.6) 0.295 (11.6) 0.188 (7.4) 0.170 (6.7) 0.058 (2.3) 0.368 (14.5)

0.025 (1.0) 0.297 (11.7) 0.185 (7.3) 0.584 (23.0) 0.086 (3.4) 0.061 (2.4) 0.203 (8.0) 0.267 (10.5) 0.130 (5.1) 0.414 (16.3) 0.414 (16.3) 0.478 (18.8) 0.058 (2.3) 0.490 (19.3) 0.353 (13.9) >0.800 (31.5) 0.117 (4.6) 0.381 (15.0) 0.406 (16.0) 0.442 (17.4) 0.112 (4.4) 0.655 (25.8)

0.025 (1) 0.58 (23) 0.36 (14) 0.53 (21) 0.28 (11) 0.28 (11) 1.07 (42) 0.41 (16) 0.51 (20) 1.07 (42) 1.19 (47) 3.12 (123) 0.28 (11) 2.26 (89) 0.84 (33) 3.43 (135) 0.58 (23) 2.57 (101) 1.63 (64) 1.47 (58) 0.51 (20) 3.20 (126)

0.20 (8) 2.49 (98) 1.55 (61) 4.88 (192) 0.71 (28) 0.51 (20) 1.70 (67) 2.24 (88) 1.09 (43) 3.45 (136) 3.45 (136) 3.99 (157) 0.48 (19) 4.09 (161) 2.95 (116) >6.68 (263) 0.97 (38) 3.18 (125) 3.40 (134) 3.68 (145) 0.94 (37) 5.46 (215)

Note: 3304 cm3/min of ﬂow rate in a 1.75 in. diam furnace tube. The moisture was removed from the air by a ﬁlter prior to entering into the furnace tube. Specimens were cathodically descaled for measurement of the metal loss. The total depth of attack is the sum of the metal loss and the depth of internal attack. Metal loss was measured by cathodically descaling the oxide scale prior to measurement of the specimen thickness. Source: Ref 71

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 42

42 / High-Temperature Corrosion and Materials Applications

(Ref 49). Test specimens, which were cut from plate products, had dimensions of 1.27 cm (thick) by 2.54 cm (width) by 2.54 cm (length). Tests were conducted in a box furnace with still air; specimens were removed from the furnace every 30 days to allow air cooling. Weights were then measured to determine the time to initiation of breakaway oxidation. For block specimens (e.g., 1.25 cm, or 0.5 in., thick specimens), oxide scales were cracking, breaking, and spalling with “bing” noises as soon as the specimens were removed from the hot zone in the furnace. Cracking noises would not stop until almost all the oxide scales were broken and spalled off. Thus, no cathodic descaling was necessary to remove the oxide scales for measurement of the specimen thickness at the end of the test. When similar oxidation testing was performed with thin test coupons (about 3.2 mm, (0.125 in.) or

200

Weight gain, mg/cm2

0 –200 –400 –600 –800 –1000

0 30 60 90 120 150 180 210 240 270 300 330 360 Days

Fig. 3.51

The oxidation behavior of alloy 800H tested in still air at 1095 °C (2000 °F) involving a thick, blocky specimen (1.25 cm, or 0.5 in., thick) cycling to room temperature every 30 days for weight measurement, showing the alloy was under protective scales initially for about 30 days and then suffered breakaway oxidation. Courtesy of Haynes International, Inc.

thinner), no “cracking” noises were heard during specimen cooling from the hot zone, and oxide scales remained on the specimen surface for these thin test coupons. Figure 3.51 shows weight change data for alloy 800H in this longterm testing at 1095 °C (2000 °F) using thick, blocky specimens, showing breakaway oxidation with linear weight loss after 60 days of exposure. Some alloys showed no breakaway oxidation even after 2 years of testing. Figure 3.35 shows no breakaway oxidation for HR120 after 2 years of testing at 980 °C (1800 °F), while 800H and RA85H suffered breakaway oxidation. HR160 and 601 showed no breakaway oxidation after one year of testing at 1095 °C (2000 °F), while alloy 800H suffered breakaway oxidation (Fig. 3.46). At the end of testing for 2 years (720 days) at 980 °C (1800 °F) and 1 year (360 days) at 1093, 1150, and 1200 °C (2000, 2100, and 2200 °F), specimens were cut, mounted, and polished for metallographic determination of the depth of internal attack. Metal loss was determined by subtracting the original specimen thickness from the thickness after testing. The data generated from blocky specimens are summarized in Tables 3.14 to 3.16. The annual oxidation rates in terms of the total depth of oxidation attack are included in Tables 3.14 to 3.16 at 980, 1090, and 1150 °F (1800, 2000, and 2100 °F), respectively. These oxidation rate values would be considered to be quite reasonable, since the test duration was almost 2 years for 980 °C (1800 °F) and about 1 year for 1090 and 1150 °C testing. At 980 °C (1800 °F), alloys 230, 617, HR120, 556, and HR160 exhibited oxidation rates of less than 10 mpy. At 1090 °C (2000 °F), only alloy 230 exhibited about 10 mpy of oxidation rate, while other alloys tested exhibited more than 20 mpy. At 1150 °C (2100 °F), all alloys tested exhibited more than 30 mpy of

Table 3.14 Oxidation of several high temperature alloys in still air at 980 °C (1800 °F) for 720 days with specimens cycling to room temperature every 30 days Alloy

230 617 HR120 556 HR160 601 RA85H 800HT

Weight change, mg/cm2

Metal loss, mm (mils)

Total depth of attack, mm (mils)

Oxidation rate, mm (mpy)

−1.4 1.0 −33.7 −19.8 −51.2 −9.9 −122.2 −417.8

0.00254 (0.1) 0 0.04060 (1.6) 0.02286 (0.9) 0.0635 (2.5) 0.0127 (0.5) 0.16002 (6.3) 0.52578 (20.7)

0.14732 (5.8) 0.23876 (9.4) 0.30988 (12.2) 0.38608 (15.2) 0.42418 (16.7) 0.56896 (22.4) 1.36398 (53.7) 2.02692 (79.8)

0.0508 (2) 0.127 (5) 0.1524 (6) 0.2032 (8) 0.2286 (9) 0.2794 (11) 0.6858 (27) 1.0414 (41)

Note: Total depth of attack = metal loss + internal attack. Source: Ref 49

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 43

Chapter 3: Oxidation / 43

Table 3.15 Oxidation of several high-temperature alloys in still air at 1095 °C (2000 °F) for 360 days with specimens cycling to room temperature every 30 days Alloy

230 556 RA330 HR160 HR120 601 800HT RA85H

Weight change, mg/cm2

Metal loss, mm (mils)

Total depth of attack, mm (mils)

Oxidation rate, mm/yr (mpy)

−42.1 −298.1 −405.0 −73.1 −665.7 −110.5 −893.9 −348.1

0.04826 (1.9) 0.36322 (14.3) 0.50800 (20.0) 0.09144 (3.6) 0.82804 (32.6) 0.13716 (5.4) 1.12522 (44.3) 0.45466 (17.9)

0.27178 (10.7) 0.53848 (21.2) 0.60706 (23.9) 0.73660 (29.0) 0.96520 (38.0) 1.14554 (45.1) 1.29540 (51.0) 2.03962 (80.3)

0.27940 (11) 0.55880 (22) 0.60960 (24) 0.73660 (29) 0.99060 (39) 1.16840 (46) 1.32080 (52) 2.05740 (81)

Note: Total depth of attack = metal loss + internal attack. Source: Ref 49

Table 3.16 Oxidation of several high-temperature alloys in still air at 1150 °C (2100 °F) for 360 days with specimens cycling to room temperature every 30 days Alloy

230 617 HR120 HR160 800H 601 RA85H

Weight change, mg/cm2

Metal loss, mm (mils)

Total depth of attack, mm (mils)

Oxidation rate, mm/yr (mpy)

−249.7 −452.7 −894.2 −155.5 −1315.6 −258.7 −389.2

0.28194 (11.1) 0.54102 (21.3) 1.10998 (43.7) 0.19304 (7.6) 1.65608 (65.2) 0.32004 (12.6) 0.50800 (20.0)

0.83680 (34.0) 0.94488 (37.2) 1.34620 (53.0) 1.49098 (58.7) 1.78562 (70.3) 1.84912 (72.8) 2.40792 (94.8)

0.88900 (35) 0.96520 (38) 1.37160 (54) 1.52400 (60) 1.80340 (71) 1.87960 (74) 2.4384 (96)

Note: Total depth of attack = metal loss + internal attack. Source: Ref 49

oxidation rate. At 1200 °C (2200 °F), aluminaformer alloy 214 showed little or no oxidation attack (Fig. 3.47). These data are valuable in providing readers with the oxidation data in terms of the total depth of oxidation attack in air based on the actual measurements of the specimens after 1 to 2 years of testing. Since the data were generated from thick, blocky specimens, caution should be used when the data are being considered for application in thin-gage sheets or foils. This is related to the reservoir effect of a solute alloying element for the formation of a protective oxide scale. The discussion of the reservoir issue and the oxidation in thin foils is presented later. In this test program (Ref 49), in addition to the metal loss caused by formation of external oxide scales and internal attack caused by formation of internal oxides and/or voids, the weight-loss values were also determined. The weight loss of the specimens is related mostly to the metal loss resulting from the removal of the external oxide scales and is not signiﬁcantly affected by the formation of internal oxides and/or voids. The weight-loss values of the alloys tested are plotted against their corresponding metal-loss values at 980, 1090, and 1150 °C (1800, 2000, and 2100 °F), revealing a nice straight line correlation (Fig. 3.52). Alloys tested were 230, 617, 601, 556, HR160, HR120 RA330, 800HT, and

RA85H. These alloys are primarily chromia formers. This correlation may be useful in making rough estimates of the metal loss for an alloy that showed only weight-loss data. The depth of oxidation attack was also investigated by John (Ref 15) for a wide variety of commercial alloys in isothermal air oxidation testing. Table 3.17 summarizes his data in terms of the temperature at which the oxidation rate reaches 10 mpy. Figure 3.53 illustrates some oxidation data in terms of oxide penetration as a function of test temperature in air after 1 year for some alloys (Ref 15). Table 3.18 shows the depth of oxidation attack of various heat-resistant alloys after cyclic oxidation tests at 1100 °C (2010 °F) in air + 5% H2O for 504 h with specimens cycling out of the furnace every 15 min (Ref 72). Lai et al. (Ref 73) reported the oxidation data generated from a ﬁeld test inside a radiant tube ﬁred with natural gas with an average temperature of 1010 °C (1850 °F) (Table 3.19). The test rack containing coupons of various alloys was exposed for about 3000 h. Many chromia formers, such as alloys 601, 230, 556, 310, 600, and RA330, were found to perform well, with extrapolated oxidation rates of less than 0.5 mm/ yr (20 mpy). Type 304, however, suffered severe oxidation attack with an extrapolated oxidation rate of more than 4.4 mm/yr (>175 mpy). The

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 44

44 / High-Temperature Corrosion and Materials Applications

70 60

1.50

1.00

40 30 20

Metal loss, mm

Metal loss, mils

50

0.50

10 0 0

200

400

600

800

1000

1200

1400

Weight loss, mg/cm2 Correlation between weight loss (mg/cm2) and the depth of metal loss (mils) for commercial alloys that are primarily chromia formers tested in air at 980 °C (1800 °F)/720 days, 1095 °C (2000 °F)/360 days, and 1150 °C (2100 °F)/360 days. Alloys tested were 230, 617, 601, 556, HR160, HR120, RA330, 800HT, and RA85H. 1.0 mil = 0.0254 mm

Fig. 3.52

Table 3.17 The oxidation rate of total depth of attack was reached after 1 year in air Maximum temperature for 0.25 mm/yr (10 mpy), °C (°F)

Carbon steel Copper Nickel 9Cr-1Mo 410 304 617 803 625 800H 601GC DS 230 310 RA330 446 556 HR120 253MA 602CA MA956 214

… C11000 N02270 S50400 S41000 S30400 N06617 … N06625 N08810 … … N06230 S31000 S33000 S44600 R30556 … S30815 … S67956 N07214

604 (1120) 677 (1250) 782 (1440) 799 (1470) 832 (l530) 893 (1640) 938 (1720) 954 (1750) 960 (1760) 966 (1770) 977 (1790) 977 (1790) 982 (1800) 982 (l800) 999 (1830) 1010 (1850) 1010 (1850) 1010 (1850) 1082 (1980) 1121 (2050) >1150 (>2100) >1150 (>2100)

AISI 410

AISI 304 10

0.3 Alloy 617 Nickel AISI 310 Alloy 800 H

Penetration, mils

UNS No.

2.5

9Cr -1Mo Carbon steel

Penetration, mils

Alloy

102

1 0.03

0.1 1000

1200

1400

1600

1800

2000

Temperature, °F

Source: Ref 15

Fig. 3.53

alumina former, alloy 214, showed little or no oxidation attack with an extrapolated oxidation rate of about 0.076 mm/yr (3 mpy). In another ﬁeld test (Ref 74), a test rack containing coupons of various alloys was placed in a natural-gas-ﬁred furnace used for reheating ingots and slabs of nickel- and cobalt-base alloys. The test was conducted for about 113 days at temperatures varying from 1090 to 1230 °C (2000 to 2250 °F), with frequent cycles to 540 °C

Oxidation penetration (metal loss + internal attack) as a function of test temperature for 1 year in air for a variety of commercial alloys. Source: Ref 15

(1000 °F) during furnace idling. The results are summarized in Table 3.20. All the chromia formers tested suffered severe oxidation attack. The alumina former (alloy 214), however, exhibited little attack. Examination of the oxide scale formed on alloy 214 was found to consist of essentially aluminum-rich oxides.

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 45

Chapter 3: Oxidation / 45

Table 3.18 Cyclic oxidation resistance of various heat-resistant alloys at 1100 °C (2010 °F) for 504 h in air-5H2O Speciﬁc weight change (descaled), mg/cm2 Alloy

ACI grade HK 310SS 800 601 617 X RA333 IN-814 188 MA-956

Mean

Range

Metal loss, mm (mils)

Maximum attack, mm (mils)

−105.8 −149.0 −168.6 −11.0 −13.5 −20.0 −30.5 −3.5 −25.0 −1.0

−98 to −124 −92 to −235 −83 to −223 −6.3 to −17.2 −6.5 to −17.5 −10.0 to −29.5 … −2.5 to −4.9 −13.0 to −40.5 −0.3 to −1.5

0.25 (9.9) 0.31 (12.2) 0.39 (15.4) <0.02 (0.8) … 0.05 (2.0) … 0 0.03 (1.2) 0.01 (0.4)

0.35 (13.8) 0.38 (15.0) 0.59 (23.2) 0.12 (4.7) … 0.25 (9.9) … <0.01 (0.4) 0.15 (5.9) 0.02 (0.8)

Note: 15 min in furnace and 5 min out of furnace. Source: Ref 72

Table 3.19 Results of field test in a natural-gas-fired radiant tube at 1010 °C (1850 °F) for 3000 h Alloy

214 601 230 556 310 600 RA330 800H 309 304

Metal loss, mm (mils)

Maximum metal affected(a), mm (mils)

Oxidation rate, mm/yr (mpy)

0.003 (0.1) 0.023 (0.9) 0.028 (1.1) 0.018 (0.7) 0.041 (1.6) 0.018 (0.7) 0.048 (1.9) 0.12 (4.7) 0.50 (19.7) >1.5 (60)(b)

0.025 (1) 0.076 (3) 0.10 (4) 0.10 (4) 0.10 (4) 0.15 (6) 0.15 (6) 0.30 (12) 0.50 (20) >1.5 (60)(b)

0.076 (3) 0.23 (9) 0.30 (12) 0.30 (12) 0.30 (12) 0.46 (18) 0.46 (18) 0.89 (35) 1.5 (58) >4.4 (175)

(a) Metal loss + maximum internal penetration. (b) Sample was consumed. Source: Ref 73

Table 3.20 Results of field test in a natural-gas-fired furnace for reheating nickeland cobalt-base alloy ingots and slabs for 113 days at 1090 to 1230 °C (2000 to 2250 °F) with frequent cycles to 540 °C (1000 °F) Alloy

214 RA330 601 600 800H 310SS 304SS 316SS 446SS

Metal loss, mm (mils)

Maximum metal affected(a), mm (mils)

0.013 (0.5) 0.39 (15.5) 0.18 (7.2) 0.64 (25.0) >0.79 (31.0)(b) >1.0 (41.0)(b) >1.5 (60.0)(b) >1.6 (63.0)(b) >0.61 (24.0)(b)

0.11 (4.5) 0.65 (25.5) 0.95 (37.2) 1.1 (45.0) >0.79 (31.0)(b) >1.0 (41.0)(b) >1.5 (60.0)(b) >1.6 (63.0)(b) >0.61 (24.0)(b)

(a) Metal loss + internal penetration. (b) Samples were consumed. Source: Ref 74

3.4.10 Oxide-Dispersion-Strengthened (ODS) Alloys Oxide-Dispersion Strengthened alloys use very ﬁne oxide particles that are uniformly distributed throughout the matrix to provide excessive strengthening at very high temperatures. These oxide particles, typically yttrium oxide, do not react with the alloy matrix so no coarsening or dissolution occurs during the exposure to very high temperatures, thus maintaining the strengthening of the alloy. This group of superalloys is produced using specialty powders that are manufactured by the mechanical alloying process. These powders are essentially composite powders with each particle containing a uniform distribution of submicron oxide particles in an alloy matrix. The process of producing these ODS powders involves repeated fracturing and rewelding of a mixture of powder particles in vertical attritors or horizontal ball mills (Ref 75). Alloy powders are then canned, degassed, and hot extruded, followed by hot working and

annealing to produce a textured microstructure (Ref 75). Alloys are available in mill products such as bar, plate, sheet, and so forth, or custom forgings. Some ODS alloys are shown in Table 3.21 (Ref 75). The oxidation behavior of some of these ODS alloys tested in air containing 5% H2O at 1200 °C (2190 °F) is shown in Fig. 3.54 (Ref 75). The oxidation behavior of MA956 compared with those of several ironand nickel-base alloys at 1100 °C (2010 °F) is shown in Fig. 3.55 (Ref 76). Additional oxidation data for some ODS alloys is presented in Section 3.4.12. 3.4.11 Effect of Oxygen Concentration on Oxidation Air atmosphere consists primarily of oxygen and nitrogen with some water vapor and small amounts of inert gases, such as argon, neon, and helium. Dry air consists of essentially 21% O2

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

pg 46

46 / High-Temperature Corrosion and Materials Applications

Table 3.21

Nominal chemical compositions of several ODS alloys

Alloy

Ni

Fe

Cr

Al

Ti

W

Mo

Ta

Y2O3

B

Zr

MA754 MA758 MA760 MA6000 MA956

bal bal bal bal …

… … … … bal

20 30 20 15 20

0.3 0.3 6.0 4.5 4.5

0.5 0.5 … 2.5 0.5

… … 3.5 4.0 …

… … 2.0 2.0 …

… … … 2.0 …

0.6 0.6 0.95 1.1 0.5

… … 0.01 0.01 …

… … 0.15 0.15 …

Composition, wt%

Note: All alloys contain 0.05% C. Source: Ref 75

50 Mass change, mg/cm2

MA 760 MA 754

–50 –100

0 –0.5 –1.0 –1.5 –2.0

–150 MA 6000 –200 –250 0

10

20

30

40

50

–2.5 –3.0 –3.5 60

Mass change, Ib/in.2 × 10–3

0.5

MA 956 0

Exposure time, days

Fig. 3.54

Oxidation behavior of several ODS alloys in air containing 5% H2O at 1200 °C (2190 °F).

Source: Ref 75

and 78% N2 with about 1% inert gases. In combustion, the concentration of oxygen may vary. Also, in some processes, oxygen may be the only gaseous component to which the equipment is exposed. Thus, the effect of oxygen concentration on the oxidation behavior of alloys may need to be evaluated for some applications. John (Ref 15, 77) investigated the oxidation behavior of a wide variety of commercial alloys in N2-O2 mixtures with the concentration of oxygen varying from 1 to 100%. The data generated at 871 °C (1600 °F) are shown in Fig. 3.56 (Ref 77), and data generated at 927 °C (1700 °F) are shown in Fig. 3.57 (Ref 15). The data presented in Fig. 3.56 were based on the 1152 h testing, while the data in Fig. 3.57 were based on tests after 1 year. At 871 °C (1600 °F), Type 304 was the only alloy that showed signiﬁcant increase in oxidation attack from about 0.25 mm/yr (10 mpy) in N2-21%O2 to close to 2.5 mm/yr (100 mpy) in 100% O2. All other alloys in the ﬁgure showed about 10 mpy or less of oxidation attack at three different levels of O2 concentrations (1%, 21%, and 100%). Figure 3.57 shows the oxidation behavior of a number of alloys at 927 °C (1700 °F). The alloys that were found to increase oxidation attack with increasing oxygen

Fig. 3.55

Cyclic oxidation resistance of ODS alloy MA956 compared with alloy 601, HK alloy, alloy 800, and Type 310. Source: Ref 76

concentration included 9Cr-1Mo steel, 410, 304, and 617, while carbon steel, nickel, 800H, and 310 were relatively unaffected by oxygen concentrations. In Fig. 3.57, Type 304 was found to exhibit approximately 0.25 mm/yr (10 mils) of attack at 927 °C (1700 °F) after 1 year in 100% O2, while close to 2.5 mm/yr (100 mpy) of attack was extrapolated based on 1152 h exposure at 871 °C (1600 °F) in 100% O2, as shown in Fig. 3.56. Extrapolation from short-term tests here showing a higher oxidation rate at lower temperature could be an issue here. More longterm tests are needed. It is also of practical interest to perform long-term tests to evaluate the oxidation behavior in 100% O2 environments for some alloys that are to be used in chemical processes involving 100% O2. 3.4.12 High-Velocity Combustion Gas Streams Oxidation of alloys can signiﬁcantly increase under high-velocity gas streams. Combustors and transition ducts in the gas turbine are subject to such conditions. These components are also

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:45PM Plate # 0

Chapter 3: Oxidation / 47

103

AISI 304 AISI 310

102

956MA

Penetration rate, mpy

10

253MA HR120

1

800HT AISI 446

10–1

617 10–2

230 214

10–3

601GC 10–4 10–2

1

10–1 Partial pressure of O2, atm

Fig. 3.56

602CA 556

Effect of oxygen concentration in the N2-O2 mixture on the oxidation penetration (metal loss + internal attack) at 871 °C (1600 °F) for 1152 h. 1.0 mil = 0.025 mm. Source: Ref 77

subject to severe thermal cycling, particularly for gas turbines in airplane engines. Laboratory burner rigs have been developed to evaluate this type of oxidation, often referred to as “dynamic oxidation,” under the condition of very high gas velocities. Some of these dynamic oxidation burner rigs are described elsewhere (Ref 78–83). Lai (Ref 82) investigated a wide range of alloys—from stainless steels to superalloys—in a burner rig that generated a combustion gas stream with 0.3 Mach (100 m/s) velocity. The specimens were held in a carousel-type holder rotating at 30 rpm with respect to the combustion gas stream. Every 30 min the carousel was withdrawn from the hot zone, and quenched to less than 260 °C (500 °F) by a blast of cold air, and then automatically reinserted back into the hot zone. The specimens were subject to severe thermal cycling. Combustion was generated using No. 2 fuel oil with an air-to-fuel ratio of 50 to 1, producing a high-velocity (0.3 Mach, or 100 m/s) test gas. The tests were conducted at 1090 °C (2000 °F) with 30 min cycles, and the results are tabulated in Table 3.22. At 1090 °C (2000 °F) with a high-velocity gas stream plus severe thermal cycling, most alloys suffered signiﬁcant metal loss, which constituted a large portion of the total depth of oxidation attack. With protection by an aluminum oxide scale, alloy 214 suffered very little

attack. The alloy showed no sign of breakaway oxidation after 500 h. Figure 3.58 shows the oxide scale formed after testing for 500 h (1000 cycles) (Ref 82). The scale consisted of aluminum-rich oxides. After 1000 h (2000 cycles) of testing, the scale remained aluminumrich. The maximum metal affected (metal loss +maximum internal penetration) remained about the same after 1000 h compared to after 500 h (Ref 82). The test results generated at 980 °C (1800 °F) for 1000 h (2000 cycles) are shown in Table 3.23 (Ref 82). Unlike 1090 °C (2000 °F) testing (Table 3.22), testing at 980 °C (1800 °F) resulted in internal oxidation and nitridation in addition to metal loss. Internal nitridation penetrated deeper into metal interior than internal oxidation penetration. Table 3.23 included only internal oxidation penetration data (i.e., maximum metal affected = metal loss + internal oxidation penetration). Limited tests were conducted to determine the effect of thermal cycling by testing at 980 °C (1800 °F) for 1000 h with 30 min cycling and without thermal cycling in dynamic oxidation testing (Ref 82). As expected, thermal cycling primarily contributed metal loss portion of the oxidation attack. The results are summarized in Table 3.24 (Ref 82). Total attack presented in Table 3.24 was based on metal loss and internal oxidation penetration. Since internal

pg 47

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

48 / High-Temperature Corrosion and Materials Applications

10

3

Carbon steel

10

9Cr-1Mo

2

Penetration, mils

10

Penetration, mm

AISI 410 1 Nickel Alloy 800 H

10

AISI 310

10

AISI 304

–1

Alloy 617

–2

1 10

–3

10

–2

10

–1

1

10

pO , atm 2

Fig. 3.57

Effect of oxygen concentration in the N2-O2 mixture on the oxidation penetration (metal loss + internal attack) after 1 year at 927 °C (1700 °F) for various commercial alloys. 1.0 mils = 0.025 mm. Source: Ref 15

nitridation attack was found to penetrate deeper into the alloy than internal oxidation attack does for many alloys, the total depth of attack for many alloys under dynamic oxidation test conditions was more than that reported in Tables 3.23 and 3.24. The oxidation/nitridation behavior of various alloys under dynamic oxidation test conditions is discussed in Chapter 4 “Nitridation.” Hicks (Ref 83) performed dynamic oxidation tests with 170 m/s gas velocity at 1100 °C (2010 °F) with 30 min cycles for several wrought chromia-former superalloys and an ODS alumina-former (MA956). Alumina former MA956 was found to be considerably better than chromium formers, such as alloys 230, 86, 617, 188, and 263. His results are shown in Fig. 3.59.

MA956 along with some ODS alloys was tested by Lowell et al. (Ref 78) with 0.3 Mach gas velocity at 1100 °C (2010 °F) with 60 min cycles. ODS alloys tested included MA956 (Fe19Cr-4.4Al-0.6Y2O3), HDA8077 (Ni-16Cr4.2Al-1.6Y2O3), TD-NiCr (Ni-20Cr-2.2ThO2) and STCA264 (Ni-16Cr-4.5Al-1Co-1.5Y2O3). Also included in the test was physical vapor deposition (PVD) coating of Ni-15Cr-17Al-0.2Y on MAR-M-200 alloy (Ni-9Cr-10Co-12W-1Nb5Al-2Ti). Their results are shown in Fig. 3.60. MA956 and HDA8077 as well as PVD Ni-Cr-AlY coating were found to perform well. No explanation was offered in the paper for STCA264, which did not perform as well as HDA8077 although both alloys had similar chemical compositions.

pg 48

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 49

Chapter 3: Oxidation / 49

Table 3.22 Dynamic oxidation resistance of iron-, nickel-, and cobalt-base alloys in high-velocity combustion gas stream (0.3 mach velocity) with 30 min cycles at 1090 °C (2000 °F) for 500 h Alloy

214 230 RA333 188 556 X RA330 S 600 310 601 617 800H 625 Multimet

Metal loss, mm (mils)

Maximum metal affected(a), mm (mils)

0.013 (0.5) 0.056 (2.2) 0.10 (4.0) 0.19 (7.5) 0.22 (8.7) 0.23 (9.0) 0.28 (10.9) 0.30 (11.8) 0.44 (17.2) 0.54 (21.2) 0.27 (10.7) 0.32 (12.4) 0.77 (30.5)(b) >0.79 (31.0)(c) 1.25 (49.1)(d)

0.046 (1.8) 0.15 (5.7) 0.22 (8.7) 0.27 (10.7) 0.30 (11.7) 0.34 (13.5) 0.35 (13.6) 0.39 (15.2) 0.53 (20.7) 0.61 (24.1) 0.61 (24.0) 0.61 (24.0) 0.86 (34.0)(b) >0.79 (31.0)(c) 1.42 (55.8)(d)

Note: Gas velocity was 0.3 mach (100 m/s, or 225 mph); samples were cycled to less than 260 °C (500 °F) once every 30 min; 50 to 1 air-to-fuel ratio; two parts No. 1 fuel oil and one part No. 2 fuel oil. Internal nitridation occurred in some alloys, but is not included in the current data. See section 4.3.3 in Chapter 4. (a) Metal loss + maximum internal, penetration. (b) Extrapolated from 400 h; sample was about to be consumed after 400 h. (c) Sample was consumed in 500 h. (d) Extrapolated from 225 h; sample was about to be consumed after 225 h. Source: Ref 82

Table 3.23 Dynamic oxidation resistance of iron-, nickel-, and cobalt-base alloys in high-velocity combustion gas stream at 980 °C (1800 °F) for 1000 h Alloy

214 230 188 556 X S RA333 625 617 RA330 Multimet 800H 310 600 601 304 316

Metal loss, mm (mils)

Maximum metal affected(a), mm (mils)

0.010 (0.4) 0.020 (0.8) 0.028 (1.1) 0.043 (1.7) 0.069 (2.7) 0.079 (3.1) 0.064 (2.5) 0.12 (4.9) 0.069 (2.7) 0.20 (7.8) 0.30 (11.8) 0.31 (12.3) 0.35 (13.7) 0.31 (12.3)(b) 0.076 (3.0) >9.0 (354)(c) >9.0 (354)(c)

0.031 (1.2) 0.089 (3.5) 0.107 (4.2) 0.158 (6.2) 0.163 (6.4) 0.17 (6.6) 0.18 (7.0) 0.19 (7.6) 0.27 (10.7) 0.30 (11.8) 0.38 (14.8) 0.39 (15.3) 0.42 (16.5) 0.45 (17.8)(b) 0.51 (20.0) >9.0 (354)(c) >9.0 (354)(c)

Note: Gas velocity was 0.3 mach (100 m/s, or 225 mph); samples were cycled to less than 260 °C (500 °F) once every 30 min; 50 to 1 air-to-fuel ratio; two parts No. 1 fuel oil and one part No. 2 fuel oil. Internal nitridation occurred in some alloys, but is not included in the current data. See section 4.3.3 in Chapter 4. (a) Metal loss + maximum internal penetration. (b) Extrapolated from 917 h; sample was about to be consumed after 917 h. (c) Extrapolated from 65 h; sample was consumed in 65 h. Source: Ref 82

3.4.13 Breakaway Oxidation In Fe-Cr, Fe-Ni-Cr, Ni-Cr, and Co-Cr alloy systems, the formation of an external Cr2O3 oxide scale provides the oxidation resistance for

Fig. 3.58

Oxide scales formed on alloy 214 in a high-velocity gas stream (0.3 Mach velocity) with 30 min cycles at 1090 °C (2000 °F) for 500 h. Area 1: 96.5% Al, 1.5% Cr, 0.1% Fe, 1.9% Ni. Area 2: 75.2% Al, 6.2% Cr, 2.6% Fe, 16.0% Ni. Area 3: 95.8% Al, 1.0% Cr, 0.1% Fe, 3.1% Ni. Area 4: 53.0% Al, 2.8% Cr, 9.2% Fe, 35.0% Ni. Source: Ref 82

Table 3.24 Effect of thermal cycling in dynamic oxidation behavior of several nickel-base alloys at 980 °C (1800 °F) for 1000 h No thermal cycling

Thermal cycling

Alloy

Metal loss, mm

Total attack(a), mm

Metal loss, mm

Total attack(a) mm

230 617 X 263

0.04 0.03 0.03 0.07

0.11 0.16 0.12 0.21

0.07 0.17 0.16 0.32

0.16 0.24 0.23 0.42

(a) Metal loss + internal oxidation. Source: Ref 82

the alloy. The growth of the Cr2O3 oxide scale follows a parabolic rate law as the exposure time increases. As the temperature increases, the oxide scale grows faster. The growth of the Cr2O3 oxide scale requires a continuous supply

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

50 / High-Temperature Corrosion and Materials Applications

0

Total depth of oxidation, mm/side

MA 956

230

0.1

617 0.2 86 188 (1971 Cast)

0.3

X

263

188 (1980 Cast) 0.4 (C) 0

200

400

600

800

Number of cycles

Fig. 3.59

Dynamic oxidation resistance of several wrought superalloys including MA956 alloy in highvelocity combustion gas stream (170 m/s) at 1100 °C (2010 °F) with 30 min cycles. Source: Ref 83

Specific weight change, mg/cm2

10

Coated MAR-M-200

0 MA 956 –10

HDA8077

–20 –30 –40

STCA-264

–50 0

400 800 1200 1600 2000 2400 2800 3200 Cycles

Fig. 3.60

Dynamic oxidation tests at 1100 °C (2010 °F) in a Mach 0.3 gas stream with each cycle consisting of 1 h at temperature followed by quenching to ambient temperature for 3 min. Source: Ref 78

of chromium from the alloy interior diffusing to the oxide/metal interface. Continued oxidation can eventually deplete chromium in the alloy matrix immediately underneath the oxide scale. When the chromium concentration in the alloy matrix immediately beneath the oxide scale is reduced to below a critical concentration, the alloy matrix no longer has adequate chromium to reform a protective Cr2O3 oxide scale when the scale cracks or spalls due to oxide growth stresses

or thermal cycling. Once this occurs, fastgrowing, nonprotective iron oxides, or nickel oxides or cobalt oxides (i.e., oxides of base metal) form and grow on the alloy surface. Breakaway oxidation, thus, initiates, and the alloy begins to undergo oxidation at a rapid rate. This is illustrated in Fig. 3.51. The alloy thus requires the level of chromium immediately underneath the chromium oxide scale to have a critical level to allow the chromium oxide scale to reheal. Gleeson (Ref 84) presented air cyclic oxidation data for three chromia formers tested at 982 °C (1800 °F) for up to 360 days. Also presented were the corresponding chromium concentration analyzed by EDX on the surface of the metal when the oxide scale was spalled off from the test specimen. The data are presented in Fig. 3.61. In this test program, thick, blocky specimens (13 mm thick, 25 mm wide, and 25 mm long) instead of typically thin coupons were used. The specimens were cycled to room temperature once every 30 days for weight measurement. Oxide scales were found to completely spall off while the specimens were removed from the furnace for cooling to room temperature with “popping” noises being heard during cooling (Ref 85). Figure 3.61 shows that alloy 230 exhibited very little weight loss with no evidence of breakaway oxidation after 360 days at 982 °C (1800 °F). The chromium concentration of the alloy immediately underneath the spalled oxide scale was found to remain at about 16% with no sign of decreasing with increasing exposure time. For HR120 alloy, the weight loss data also revealed no evidence of breakaway oxidation up to 360 days of exposure. The corresponding chromium concentration of the alloy on the surface underneath the spalled oxide scale remained approximately about 18 to 20% up to 240 days, and then dropped to about 13% after 360 days. Alloy 800HT, on the other hand, showed breakaway oxidation after 180 days. The corresponding chromium concentration of the alloy on the surface immediately underneath of the spalled oxide scale after 180 days was found to be about 10%. With continuing oxidation, alloy 800HT suffered linear weight loss and continued the decrease in chromium concentration to 8% when the exposure reached to 360 days. However, when alloy 800HT was oxidized after 90 days showing no sign of weight loss, the chromium concentration of the alloy underneath the oxide scale was about 11%. The data appear to suggest

pg 50

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 51

Chapter 3: Oxidation / 51

that the critical chromium concentration of the alloy underneath the oxide scale is approximately 11% for alloy 800HT. When the chromium concentration underneath the oxide scale is below 11%, the reformation of a protective chromium-rich oxide scale is not possible, thus resulting in breakaway oxidation. In the same oxidation test program at 1150 °C (2100 °F), similar analysis on the chromium concentration proﬁle underneath the oxide scale was performed for alloys 230, HR160 and HR120 with the data presented in Fig. 3.62 and 3.63 (Ref 86). Figure 3.62 shows the weight-loss data for three alloys up to 360 days. The chromium concentration proﬁle from the surface immediately underneath the

230 Weight change, mg/cm2

0.0 HR120

–100.0

–200.0 800HT

–300.0 0

100

(a)

200 Time, days

300

400

spalled oxide scale to the alloy interior after 360 days of testing was determined using EDX analysis with the results shown in Fig. 3.63. The chromium concentration immediately underneath the spalled oxide scale for both alloys 230 and HR120 was well below 10%, while that of alloy HR160 was about 10%. Alloy HR120 showed sign of breakaway oxidation after 90 days of exposure. Continuing oxidation testing resulted in a linear weight-loss rate. Alloy 230 showed signs of breakaway oxidation after 240 days. HR160, however, showed a linear weight loss up to 360 days, although suffering the least weight-loss rate, with no clear sign of breakaway oxidation. Oxidation of alloy HR160 is involved the formation of Cr2O3 and SiO2. This may explain that HR160, although continuing to lose weight, still showed no sign of breakaway oxidation after 360 days at 1150 °C (2100 °F). In a long-term oxidation study of Fe-20Cr-25Ni alloy in CO2 containing 1% CO, 300 ppm H2O, and 300 ppm CH4 at 1023 to 1173 K (750 to 900 °C), Evans et al. (Ref 87) found the critical chromium level for rehealing of chromium oxide scales to be about 16%. The minimum level of chromium needed to maintain a protective chromium oxide scale to prevent breakaway oxidation may vary from environment to environment and from alloy to alloy. To prolong the time for the initiation of breakaway oxidation, it is necessary to have an adequate reservoir for chromium immediately below the oxide scale to provide adequate chromium to maintain a protective chromium oxide

28.0

Weight change, mg/cm2

Surface Cr concentration, wt%

0

HR160

24.0 HR120 20.0 230 16.0

12.0

–250.0 230 –500.0 HR120 –750.0

800HT 8.0 (b)

0

100

200

300

400

–1000.0 0

Time, days

60

120

180

240

300

360

Time, days

Fig. 3.61

Weight changes as a function of exposure time in long-term cyclic oxidation tests in air at 982 °C (1800 °F) for alloys 800HT, HR120, and 230 (a), and the corresponding changes in the surface chromium concentration (measured after the scale was spalled off) as a function of exposure time (b). Source: Ref 84

Fig. 3.62

Weight changes as a function of exposure time for alloys 230, HR160, and HR120 in air oxidation tests at 1150 °C (2100 °F) with thermal cycling to room temperature for weight measurement once every 30 days. Source: Ref 86

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

52 / High-Temperature Corrosion and Materials Applications

80.0 70.0 Ni Concentration, wt%

60.0 50.0 40.0 30.0 Cr 20.0 W

10.0 0 0

200

400

600

800

1000

1200

Distance from surface, µm

(a) 50.0

Ni

Concentration, wt%

40.0

Co

30.0

20.0

Cr

10.0

0 0

400

800

1200

1600

2000

Distance from surface, µm

(b) 60.0

Concentration, wt%

50.0 Ni

40.0

30.0

Fe

20.0

Cr

10.0

0 0 (c)

Fig. 3.63

200

400

600

800

1000

Distance from surface, µm

Concentration profiles for (a) alloy 230, (b) HR160, and (c) HR120 after air oxidation tests for 360 days at 1150 °C (2100 °F), as shown in Fig. 3.62. Source: Ref 86

scale or to reheal the oxide scale that suffered local cracking or failure. Brady et al. (Ref 88) proposed that chromium carbides might provide a reservoir of chromium for maintaining the growth and rehealing of chromium oxide scales. These authors observed that an as-cast Fe-15Cr0.5C formed a nonprotective iron-rich oxide scale when exposed in O2 at 850 °C (1562 °F). When the alloy was forged at 1150 °C (2100 °F) to produce a uniformly distributed ﬁne carbide phase, the forged alloy showed a thin protective Cr2O3 scale under the same test condition (Ref 88). During oxidation testing of this forged sample, the ﬁne chromium-rich carbides are dissolved into the underlying alloy substrate to continue supplying chromium to maintain and reheal the chromium oxide scales (Ref 88). For alumina formers, such as Fe-Cr-Al alloys, and Fe-Cr-Al-base and Ni-Cr-Al-base ODS alloys, breakaway oxidation occurs when aluminum concentration underneath the Al2O3 scale has reduced to a critical level such that healing of the Al2O3 is no longer possible, thus resulting in the formation of nonprotective, fast-growing oxides of base metals (e.g., iron oxides or nickel oxides). The breakaway oxidation due to rapid growth of iron oxides or nickel oxides becomes essentially a life-limiting factor. This critical aluminum concentration was found to be about 1.0 to 1.3% for Fe-Cr-Al-base ODS alloys (e.g., MA956, ODM751) at 1100 to 1200 °C (2012 to 2192 °F) (Ref 89, 90). These values were obtained from foil specimens (0.2 to 2 mm thick) tested in still air at 1100 to 1200 °C. For the nonODS Fe-20Cr-5Al alloy, this critical aluminum concentration was found to be higher (about 2.5%) at 1200 °C (Ref 89). Since the breakaway oxidation is related to aluminum reservoir in the alloy, and the aluminum reservoir becomes a critical issue when the component is made of thin sheet or foil. Because of excellent oxidation resistance at very high temperatures, there is increasing interest in looking at alumina formers for products that require thin foils, such as honeycomb seals in gas turbines, metallic substrates for automobile catalyst converters, and recuperators in microturbines. The oxidation behavior of several commercial alumina formers in thin foils is summarized in Section 3.4.14. For alumina formers to improve their resistance to breakaway oxidation, yttrium is frequently used to increase the adhesion of the aluminum oxide scale. Other alloying elements that are known to increase the adhesion of the aluminum oxide scale include zirconium and

pg 52

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 53

Chapter 3: Oxidation / 53

hafnium. Quadakkers (Ref 91) shows that both MA956 (Fe-20Cr-4.5Al-0.5Y2O3) and Aluchrom (Fe-20Cr-5Al-0.01Y) exhibited much more cyclic oxidation resistance than Fe-20Cr5Al when tested at 1100 °C in synthetic air with a hourly cycle to room temperature (Fig. 3.64). Addition of Y2O3 to an alumina former has a similar beneﬁcial effect as yttrium alloying element. Klower and Li (Ref 92) studied the oxidation resistance of Fe-20Cr-5Al alloys in 10 different compositions containing various amounts of yttrium ranging from 0.045 to 0.28%. All 10 compositions contained 0.002% S, and eight compositions contained 0.04 to 0.06% Zr with two compositions containing no zirconium. Cyclic oxidation tests were performed at 1100 and 1200 °C (2012 and 2192 °F), respectively, with each cycle consisting of 96 h at temperature and rapid air cooling to room temperature. These authors concluded that the yttrium addition of about 0.045% was sufﬁcient to prevent the oxide scales from spalling and when the yttrium concentration was increased to more than 0.08%, substantial internal oxidation could occur, resulting in rapid metal wastage, as shown in Fig. 3.65 (Ref 92). Sulfur in the alloy is known to play a very signiﬁcant role in the adhesion of the aluminum oxide scale to the alloy substrate for alumina formers. The role of yttrium is believed to prevent the preferential segregation of sulfur in the alloy to the scale/metal interface to weaken the adhesion of the oxide scale (Ref 93–95). Reducing the concentration of sulfur in a Ni-Cr-Al alloy can signiﬁcantly improve the oxidation resistance of the alloy. Smeggil (Ref 96) compared cyclic oxidation resistance between the

normal purity Ni-Cr-Al alloys (approximately 30 to 40 ppm S) with the high-purity Ni-Cr-Al alloys (approximately 1 to 2 ppm S), showing a signiﬁcant improvement in cyclic oxidation resistance when sulfur in the alloy was signiﬁcantly reduced. This is illustrated in Fig. 3.66 (Ref 96). Also demonstrated in the ﬁgure is the beneﬁcial effect of yttrium addition to the normal purity Ni-20Cr-12Al alloy, showing signiﬁcant improvement in the cyclic oxidation resistance of the alloy without reducing the sulfur content in the alloy. Sulfur has been found to segregate to the oxide/alloy interface during oxidation in FeCr-Al alloys (Ref 97, 98). The role of yttrium is believed to tie up sulfur at the oxide/metal interface, thus improving the oxide-scale adhesion (Ref 96). 3.4.14 Thin Foils There are some industrial applications that require thin-gage sheet materials or thin foils for construction of some critical components. As the component thickness decreases, oxidation becomes a major limiting factor for its service life. When the component is made of thin foil, prolonging the incubation time before the initiation of breakaway oxidation is the controlling factor for extending the service life of the component. Thus, as applications are being pushed toward higher and higher temperatures, alloys that form aluminum oxide scales can offer tremendous advantages in performance over those alloys that form chromium oxide scales. In gas turbine applications, one important component made of a thin foil is a turbine seal ring assembly that controls the turbine tip

Depth of internal oxidation, µm

Weight change, mg/cm2

4 Fe-20Cr-AI MA956 Aluchrom

3 2 1 0 –1 0

200

400

600

800

1000

1200

800

1100 °C 1200 °C

600 400 200 0 0

Fig. 3.64

Cyclic oxidation resistance of MA956, Aluchrom, and Fe-20Cr-5Al tested in synthetic air at 1100 °C (2012 °F) with an hourly cycle (each cycle consisted of 56 min heating and 4 min cooling). Source: Ref 91

0.1

0.2

0.3

Yttrium, wt%

Time, h

Fig. 3.65

Maximum internal oxidation depth as a function of yttrium content in the alloys after 3000 h of cyclic oxidation tests at 1100 and 1200 °C (2012 and 2192 °F) in air. Source: Ref 92

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 54

54 / High-Temperature Corrosion and Materials Applications

Mass change/unit area, mg/cm2

0

High-purity NiCrAI –8 Normal-purity NiCrAIY –16

–24

Normal-purity NiCrAI

–32 0

20

40

60

80

100

Number of 1 h cycles

Fig. 3.66

Cyclic oxidation resistance of the normal purity Ni-20Cr-12Al (30 to 40 ppm S), the high-purity Ni-20Cr-12Al (1 to 2 ppm S) and the normal purity Ni-20Cr-12Al-Y at 1180 °C. Source: Ref 96

clearance for improving thermal efﬁciency. The seal ring assembly is typically constructed out of a honeycomb seal brazed onto a superalloy casing. The traditional alloys used for honeycomb seals are chromia formers, such as nickel-base alloy X. Lai (Ref 99) evaluated honeycomb samples made of wrought alloy 214 (Ni-16Cr-3Fe-4.5Al-Y) and alloy X (Ni-22Cr18.5Fe-9Mo) in dynamic burner rig testing that simulated gas turbine hot gas conditions. Tests were conducted at 954 °C (1750 °F) in a highvelocity combustion gas stream (0.3 Mach or 100 m/s) with cycling every 30 min. Both alloy X and alloy 214 honeycomb samples were made of 0.076 mm (3 mils) foils. The alloy X honeycomb sample was completely oxidized (destroyed) after 154 h of testing, while the alloy 214 honeycomb sample was unaffected after 317 h when the test was terminated. Figure 3.67 shows the condition of both samples after tests. Simms et al. (Ref 100) conducted extensive oxidation studies on several commercial foil materials with different thicknesses (from 0.05 to 0.127 mm) in a simulated combustion environment (nominally N2-14%O2-3%CO2-7.8%H2O) at 950 to 1250 °C (1742 to 2282 °F). Test specimens were cycled to room temperature every 100 h when tested at 950 and 1050 °C (1742 and 1922 °F), every 40 h at 1150 °C (2102 °F), every 20 h at 1250 °C (2280 °F). Each specimen was

contained in an individual alumina crucible so the spalled oxides could be included in the weight measurement. All the specimens were preoxidized in air at 1050 °C (1922 °F) for 1 h prior to oxidation testing. The alloys tested included Kanthal AF (Fe-20Cr-5Al-0.05Y0.08Zr wrought alloy), Aluchrom YHf (Fe-20Cr5.8Al-0.04Y-0.05Zr-<0.1Hf wrought alloy), PM2000 (Fe-20Cr-5.5Al-0.5Y2O3 ODS alloy), and Haynes 214 alloy (Ni-3Fe-16Cr-4.5Al0.01Y wrought alloy). The 0.127 mm thick Haynes 214 foil specimens were the only foils exhibiting no breakaway oxidation over the test duration of 300 h (15 cycles) at 1250 °C (2282 °F). Kanthal AF, Aluchrom YHf and PM2000 foil specimens with 0.127 mm thick suffered breakaway oxidation well before completing 300 h (15 cycles) test duration. Thinner foils suffered breakaway oxidation in shorter times than thicker foils. As the test temperature was lowered, the life of the foil was progressively longer. At 950 °C, tests were conducted only on alloys 214 and PM2000. Both alloys exhibited no breakaway oxidation for up to 1500 h (15 cycles) for all the thicknesses (0.05 to 0.127 mm) tested. Klower (Ref 101) studied the effect of the foil thickness varying from 0.049 to 0.25 mm (2 to 10 mils) on breakaway oxidation of Fe20Cr-5Al alloys, which contained rare-earth metals (mischmetal, such as Ce), in air at

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

Chapter 3: Oxidation / 55

(a)

(b)

Fig. 3.67 Honeycomb samples after testing at 950 °C (1750 °F) for 154 h for the alloy X honeycomb sample (a) and 317 h for the 214 honeycomb sample (b) in a high-velocity combustion gas stream (0.3 Mach or 100 m/s) generated by a dynamic burner rig. The samples were also subjected to rapid quenching from the test temperature to less than 260 °C (500 °F) for 2 min every 30 min. Both honeycomb samples were made of 0.076 mm (3 mil) foils. Courtesy of Haynes International, Inc. 1100 °C (2012 °F). The specimens were cooled to room temperature for weighing every 96 h. The time to breakaway oxidation increased with increasing foil thickness, as illustrated in Fig. 3.68 (Ref 101). The incubation time for breakaway oxidation is dependent on the amount of aluminum in the reservoir for alumina formers. The total amount of aluminum in the reservoir of the alloy increases with increasing foil thickness. Thus, it takes much longer for a thicker foil to reduce the concentration of aluminum below the critical level such that rehealing of the protective oxide scale is not possible, resulting in breakaway oxidation.

Thin foils of oxidation-resistant alloys including stainless steels, Fe-Ni-Cr alloys and nickel-base alloys have been extensively evaluated for high-temperature recuperators in microturbines (Ref 102–104). The use of a recuperator for preheating the incoming air for combustion can signiﬁcantly increase the efﬁciency of the microturbine. The oxidation tests were carried out on foils with about 100 µm (0.1 mm, or 4 mils) thick in air containing 10% H2O, which was to simulate the level of H2O that would be present in the exhaust gas stream in microturbines. It was found that air containing 10% H2O was signiﬁcantly more aggressive than dry air (i.e., laboratory air) (Ref 102–104). The

pg 55

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

56 / High-Temperature Corrosion and Materials Applications

60 0.072 50 Mass change, g/m2

Foil thickness, mm 0.25 0.165 0.125 0.091 0.072 0.058 0.05 0.049

0.058

40

0.05 0.091

30 0.049 20 10 0 0

500

1000

1500

2000

2500

3000

3500

Exposure time, h

Fig. 3.68

Weight gain as a function of exposure time in air at 1100 °C (2012 °F) for Fe-20Cr-5Al foils with various thicknesses. The alloy also contained 0.015% Mischmetal. Source: Ref 101

effect of water vapor on oxidation behavior of alloys is discussed in next section. Pint (Ref 104) reported that Type 347 foil (100 µm, or 4 mils, thick) suffered accelerated oxidation (or breakaway oxidation) after less than 2000 h of exposure at 650 °C (1200 °F) in air containing 10% H2O. The foil was also subjected to thermal cycling every 100 h. Under the same test condition, alloy HR120, and alloy 625 showed no accelerated oxidation (or breakaway oxidation) after more than 8000 h. Both alloys HR120 and 625 exhibited a slight mass loss of about less than 0.2 mg/cm2 after 8000 h of exposure. The author attributed this small mass loss to volatilization of CrO2(OH)2 that formed during exposure of the water vapor in the environment (Ref 104). Both alloys HR120 and 625 showed that the oxidation behavior at 700 °C (1292 °F) was similar to that at 650 °C (1200 °F), exhibiting very little mass loss after more than 8000 h (Ref 104). At 800 °C (1472 °F), alloy HR120 showed no sign of accelerated oxidation after 7500 h. Alloy 625 again showed excellent oxidation resistance with little mass loss at 800 °C (1472 °F) for times up to 6000 h when the test was terminated (Ref 104). Both HR120 and 625 exhibited excellent oxidation resistance in air containing 10% H2O at 650 to 800 °C (1200 to 1472 °F). However, Type 347 performed poorly at 650 °C (1200 °F) in air containing 10% H2O. Nevertheless, when Type 347 foil was tested at 650 °C (1200 °F) in laboratory air (i.e. dry air) for up to 40,000 h, the alloy showed a thin oxide scale with no sign of

accelerated oxidation (Ref 105). The effect of water vapor on oxidation of alloys is discussed in the next section. 3.4.15 Effect of H2O on Oxidation In high-temperature combustion atmospheres, water vapor is invariably present in the environment. In some cases, the level of water vapor in the environment can be signiﬁcant. The effect of water vapor on the oxidation of alloys is thus an important factor in the alloy selection process. Most oxidation data are generated in laboratory air, which generally contain low levels of water vapor. Tuck et al. (Ref 106) investigated the effect of water vapor on oxidation of iron in air and air with 5%, 10%, 15%, and 20% H2O at 800 and 1000 °C (1472 and 1832 °F) for times up to 200 min. Figure 3.69 shows the weight changes as a function of time at 800 °C (1472 °F), indicating essentially no effects on oxidation by the presence of water vapor up to 20% (Ref 106). Also, no water vapor effects were observed when tested at 1000 °C (1832 °F). The tests, however, were conducted at such high temperatures that carbon steels would never be used. In boilers burning fuels containing large amount of moisture (e.g., 15 to 40% in coal) (Ref 107), a large amount of H2O is expected to be present in the combustion atmosphere in a coal-ﬁred boiler. The ﬂue gas analysis from a utility pulverized coal-ﬁred boiler showed approximately 10% H2O (Ref 108). Carbon and low-alloy steels,

pg 56

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 57

Chapter 3: Oxidation / 57

40

Weight gain, mg/cm2

30

20

10 Air 20% steam Air 15% steam Air 10% steam Air 5% steam Dry air 0

50

100

150

200

250

Time, min

Fig. 3.69

Effect of water vapor on oxidation of iron in air and air with 5%, 10%, 15%, and 20% H2O at 800 °C (1472 °F) for times up to 200 min. Source: Ref 106

which are used for the construction of the boiler waterwalls in the combustion zone, have performed well under normal oxidizing conditions. More detailed discussion on oxidation and high-temperature corrosion of carbon and low-alloy steels in coal-ﬁred boilers is presented in Chapter 10. Segerdahl et al. (Ref 109, 110) studied the oxidation behavior of 11Cr steel (X20 11Cr1MoV) in O2 and O2 with 10% and 40% H2O at 600 °C (1112 °F). In tests at 600 °C (1112 °F) for 700 h at a gas ﬂow rate of 0.5 cm/s, the steel showed no weight changes for 100% O2 and O2-10H2O, while it suffered breakaway oxidation after 336 h when tested in O2-40% H2O (Ref 109). The steel exposed to dry O2 exhibited a protective (Cr,Fe)2O3 scale after testing, while the steel showed hematite in the outer layer and FeCr spinel in the inner layer when exposed to O2-40H2O mixtures (Ref 109). When tests were conducted at 600 °C (1112 °F) for 168 h with test gas velocity at 0.5, 1.0, 2.5, 5, and 10 cm/s, the steel showed no weight changes in 100% O2 and O2-10% H2O at all gas velocities, while it suffered accelerated weight gain at 5 and 10 cm/s when tested in O2-40% H2O (Ref 110). Asteman et al. (Ref 9) studied oxidation behavior of Type 304L at 873 K (600 °C) in O2

and O2 with 10% H2O. In the O2 environment, the weight of the alloy increased initially with increasing time, which was then followed by a decrease in the rate of increase with further exposure time, indicating the formation of a protective oxide scale. As for the environment consisting of O2 with 10% H2O, the weight of the alloy initially increased and then decreased with increasing exposure time. The authors observed a brownish layer of deposits in the exhaust end of the furnace tube downstream from the test specimen. The analysis of the deposits revealed a high concentration of chromium. They proposed the weight loss in the O2-10% H2O environment was due to the evaporation of CrO2(OH)2 that formed on the alloy as a result of water vapor. The author showed that the theoretical partial pressure of CrO2(OH)2 in the O2-10%H2O environment was quite high at low temperatures, as shown in Fig. 3.4 (Ref 9). The effect of water vapor on the oxidation behavior of Type 321 at 800 °C (1472 °F) was illustrated in Fig. 3.70, showing Type 321 foil suffered severe oxidation attack in air containing 10% H2O compared with the sample tested in dry air (laboratory air) (Ref 111). The ﬁgure also shows a severe oxidation attack for Type 347 foil tested at 800 °C (1472 °F) in air with 10%

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

58 / High-Temperature Corrosion and Materials Applications

10 347, 800 °C 10% H2O

9 8

321, 800 °C 10% H2O

Mass change, mg/cm2

7 6

321, 700 °C 10% H2O

5 4 3 2

347, 700 °C 10% H2O

1

321, 800 °C dry air

0 0

100

200

300

400

500

600

700

800

900

1000

Time, h

Fig. 3.70

Oxidation behavior of Type 321 and 347 foils (100 µm thick) at 700 and 800 °C (1292 and 1472 °F) in dry air and air with 10% H2O. Specimens were cycled to room temperature every 100 h. Source: Ref 111

H2O. The detrimental effect of water vapor in air on the oxidation resistance of Type 347 is clearly illustrated in Fig. 3.71 (Ref 112). Water vapor has been found to be detrimental to the oxidation resistance of Type 310 (Fig. 3.72) (Ref 114), alloy X (Fig. 3.73) (Ref 114), and CMSX-4 (Fig. 3.74) (Ref 115). Both Type 310 and alloy X are chromia formers, while CMSX-4 (Ni-9.5Co-6.4Cr-5.7Al-6.3W-6.5Ta2.9Re) is an alumina former. Onal et al. (Ref 116) investigated the cyclic oxidation for a number of alumina-formers cyclic oxidation in both dry air and air containing 30% H2O at temperatures from 700 to 1000 °C (1292 F to 1832 °F). The authors proposed that water vapor reduced the adhesion of aluminum oxide scale, thus causing oxide spallation (Ref 116). 3.4.16 Intermetallic Compounds Tremendous interest has been generated among academic researchers in intermetallic materials and their structures, properties, and high-temperature corrosion behaviors. However, industrial applications of these materials are still limited. Limited data are presented here to brieﬂy

compare some promising nickel aluminides (Ni3Al) or nickel aluminide-base materials with some commercial nickel-base alloys. Readers are referred to Ref 117 for more information about the oxidation and corrosion of various types of intermetallic alloys. Comparison oxidation tests were performed in air at 1150 and 1200 °C (2100 and 2200 °F) between two promising Ni3Al-base nickel aluminides, IC-50 and IC-218 (developed by Oak Ridge National Laboratory), and nickelbase alloys, alloy 214 (alumina former) and alloys X and 230 (chromia formers) (Ref 55). Test results for IC-50 (Ni-11.3Al-0.6Zr0.02B), IC-218 (Ni-8.5Al-7.8Cr-0.8Zr-0.02B), alloy 214 (Ni-4.5Al-16Cr-3Fe-0.01Y), alloy X (Ni-22Cr-18Fe-9Mo-0.6W), and alloy 230 (Ni-22Cr-14W-2Mo-0.02La) are summarized in Table 3.25. Test results on other alloys under the same test conditions are shown in Table 3.12. Samples of IC-218 were completely turned into oxides after 1008 h at 1150 and 1200 °C (2100 and 2200 °F), although they still maintained the sample shape. Their specimen weight gain data did not reveal the complete oxidation of the specimen. IC-50 suffered internal oxidation

pg 58

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 59

Chapter 3: Oxidation / 59

347 foil, 40,000 h, air

50 µm Cu-plating

(a)

Cu-plating

347 foil, 40,000 h, air

10 µm

(b)

347 foil, 10,000 h, air+10% H2O

50 µm (c)

Fig. 3.71

Optical micrographs showing a thin oxide scale formed on Type 347 foil when tested in laboratory air after 40,000 h at 650 °C (1200 °F) (a) and (b) and thick unprotective oxide scales formed on Type 347 foil when tested in air containing 10% H2O after only 10,000 h at the same temperature (c). Source: Ref 112. Courtesy of Oak Ridge National Laboratory

10

Weight change, mg/cm2

0

–50

–100 0.01% H2O 5% H2O 10% H2O –150

0

200

400

600 Time, h

Fig. 3.72

800

1000 Descaled

Effect of water vapor in air on the oxidation resistance of Type 310 at 1100 °C (2010 °F) cycled every 100 h. Source: Ref 113

that penetrated through the specimen thickness, as illustrated in Fig. 3.75. Again, the weight change data failed to reveal the severity of the oxidation attack for IC-50. This example emphasizes the fact that specimen weight gain (or weight change) data can be misleading. Metallographic examination of the test specimens is also required to determine the total depth of attack, which included both metal loss and internal penetration. Comparing IC-50 and IC-218 with alloy 214, it is interesting to note that both nickel aluminides, with signiﬁcantly higher aluminum contents (twice of that in alloy 214), failed to produce a protective aluminum oxide scale. SEM/EDX analyses of the scales formed on both nickel aluminide samples after testing at 1150 °C (2100 °F) revealed both nickel- and aluminumrich oxides for IC-50 and mostly nickel-rich oxides for IC-218 (Fig. 3.76) (Ref 55). It is believed that chromium is needed to increase aluminum activities in the alloy for forming a continuous aluminum oxide scale. In his studies on oxidation of IC-50 in air at 1000 °C (1832 °F), Natesan (Ref 118) found that NiO was

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 60

60 / High-Temperature Corrosion and Materials Applications

10

+1

0

0

CMSX4 in dry air –1 ∆m/A, mg/cm2

∆W/A, mg/cm2

–2

0.01% H20 5% H20

–50

10% H20

–3

CMSX4 in wet air

–4 –5 –6 –7 –8

–100 0

200

400

600

800

Time, h

–9

1000 Descaled

Fig. 3.73

Oxidation of alloy X at 1100 °C (2010 °F) in air containing various amounts of H2O. Cycling to room temperature every 100 h for weighing. Source: Ref 114

0

100

200

300

400

Time, h

Fig. 3.74

Cyclic oxidation behavior of CMSX4 (Ni-9.5Co6.4Cr-5.7Al-6.3W-6.5Ta-2.9Re) in dry air and air containing 0.1 atm H2O (10% H2O in 1.0 atm test gas), cycling once an hour with 45 min at 1100 °C (2012 °F) and 15 min cooling. Source: Ref 115

Table 3.25 Results of oxidation tests on nickel aluminides and nickel-base alloys in air for 1008 h at indicated temperatures 1150 °C (2100 °F)

1200 °C (2200 °F)

Material

Weight gain, mg/cm2

Metal loss, mm (mils)

Average metal affected(a), mm (mils)

IC-50 IC-218 214 230 X

2.7 16.7 1.2 33.5 68.5

0.008 (0.3) >0.36 (14)(c) 0.005 (0.2) 0.058 (2.3) 0.11 (4.5)

>0.38 (15)(b) >0.36 (14)(c) 0.0075 (0.3) 0.086 (3.4) 0.147 (5.8)

Weight gain, mg/cm2

Metal loss, mm (mils)

Average metal affected, mm (mils)

8.8 61.5 1.3 81.2 169(d)

0.02 (0.8) >0.36 (14)(c) 0.005 (0.2) 0.11 (4.5) >0.9 (35.4)(e)

>0.38 (15)(b) >0.36 (14)(c) 0.018 (0.7) 0.20 (7.9) >0.9 (35.4)(e)

Note: Flowing air 30 cm/min (472 cm3 min in a 1.75 in. diam tube); cycled to room temperature once a week (168 h cycles). (a) Metal loss + average internal penetration. (b) Internal penetration through thickness. (c) Sample was completely oxidized. (d) Specimen weight gain after 504 h. (e) Extrapolated from 504 h; specimen was completely oxidized (consumed) in 504 h. Source: Ref 55

the only oxide formed on the sample; no Al2O3 was detected. Burner rig dynamic oxidation tests under a high-velocity combustion gas stream (0.3 Mach or 100 m/s) were also performed on IC-50 compared with commercial alloys (Ref 55). The nickel aluminide IC-50 suffered signiﬁcantly more severe oxidation attack than nickel-base alloys 214 (alumina former) and 230 (chromia former) after testing at 1090 °C (2000 °F) for 500 h with 30 min cycles. IC-50 suffered more than 0.38 mm (15 mils) of oxidation attack, compared with 0.05 mm (1.8 mils) for alloy 214 and 0.14 (5.7 mils) for alloy 230. The data for other commercial alloys tested under the same conditions are shown in Table 3.22. The nickel aluminide was very susceptible to internal oxidation. Signiﬁcant internal oxidation was observed after only 50 h of testing. SEM/EDX analysis of the scale formed on the

50 h tested sample revealed mainly nickel-rich oxides. 3.4.17 Catastrophic Oxidation As temperature increases, metals and alloys generally suffer increasingly higher rates of oxidation. When the temperature is excessively high, metals and alloys can suffer rapid oxidation. There is, however, another mode of rapid oxidation that takes place at relatively lower temperatures. This mode of rapid oxidation, which is often referred to as “catastrophic oxidation,” is associated with the formation of a liquid oxide. The liquid oxide disrupts and dissolves the protective oxide scale, causing the alloy to suffer rapid oxidation at relatively low temperatures. Leslie and Fontana (Ref 119) observed an unusually rapid oxidation for Fe25Ni-16Cr alloy containing 6% Mo when heated

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

pg 61

Chapter 3: Oxidation / 61

Original Sample Thickness

33.3 µm

Fig. 3.76

Nickel-rich oxides formed on nickel aluminide IC218 after 1008 h at 1150 °C (2100 °F) in air with 168 h cycles. Area 1: 11.7% Al, 80.0% Ni, 8.3% Cr. Area 2: 18.8% Al, 49.0% Ni, 32.2% Cr. Area 3: 22.7% Al, 57.1% Ni, 20.2% Cr. Source: Ref 55

200 µm

Fig. 3.75

Specimen cross section of IC-50 after oxidation testing in air at 1150 °C (2100 °F) for 1008 h with 168 h cycles. Samples were descaled prior to metallographic mounting. Source: Ref 55

in static air to 900 °C (1650 °F). The same alloy exhibited good oxidation resistance when heated to the same temperature in ﬂowing air. They postulated that the rapid oxidation was due to the accumulation of gaseous MoO3 on the metal surface. The oxidation is accelerated by the thermal dissociation of MoO3 into MoO2 and O. Meijering and Rathenau (Ref 120), Brasunas and Grant (Ref 121), and Brennor (Ref 122), however, attributed this to the presence of a liquid oxide phase. The MoO3 oxide melts at about 795 °C (1463 °F). The 19Cr-9Ni steel suffered catastrophic oxidation attack in the presence of MoO3 at 770 °C (1420 °F). This temperature was very close to the eutectic temperature of MoO2-MoO3-Cr2O3, which was reported to be 772 °C (1420 °F) (Ref 123).

Other mixed oxides involving MoO3 that exhibit low melting points are MoO2-MoO3 (778 °C, or 1435 °F), MoO2-MoO3-NiO (764 °C, or 1400 °F), Fe2O3-MoO3 (730 °C, or 1345 °F), Fe2O3-Fe-MoO3 (725 °C, or 1335 °F), V2O5MoO3 (610 to 718 °C, or 1130 to 1330 °F), Na2O-MoO3 (499 °C, or 930 °F), and Cu2OMoO2-MoO3 (470 °C, or 880 °F) (Ref 123). Other oxides, such as PbO and V2O5, can also cause metals or alloys to suffer catastrophic oxidation in air at intermediate temperatures of 640 to 930 °C (1200 to 1700 °F) (Ref 124). PbO and V2O5 melt at 888 and 690 °C (1630 and 1270 °F), respectively. The deleterious effect of lead oxide was believed to be related to exhaust-valve failures in gasoline engines. Gasoline additives were a primary source for lead compounds. Vanadium is an important contaminant in residual or heavy fuel oils. Therefore, V2O5 plays a signiﬁcant role in oil ash corrosion in oil-ﬁred boilers, which is discussed in Chapter 11. Sawyer (Ref 124) indicated that accelerated oxidation of Type 446 stainless steel in the presence of lead oxide can proceed at temperatures where the liquid phase does not exist. Experiments carried out by Brasunas and Grant (Ref 125) showed that 16Cr-25Ni-6Mo alloy specimens placed adjacent to, but not in contact with, 0.5 g samples of WO3 oxides suffered accelerated oxidation attack when tested in air at 868 °C (1585 °F), which is well below the melting point of WO3 (i.e., 1473 °C, or 2683 °F).

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

62 / High-Temperature Corrosion and Materials Applications

Because molybdenum and tungsten are very important solid-solution-strengthening alloying elements, many superalloys containing either or both have been developed since Leslie and Fontana ﬁrst observed catastrophic oxidation in 1948. Some of these alloys, including alloys X, R-41, 625, 617, S, 230, 25, and 188, have been used successfully in service for the hot section of gas turbine engines. Many of them have also been used successfully in heat treating, chemical processing, and related industries. The most effective way to alleviate the potential catastrophic oxidation problem is to avoid a stagnant condition for the gaseous atmosphere.

3.5 Oxidation/Nitridation in Air and Combustion Atmospheres Nitridation can take place in conjunction with oxidation in oxidizing atmospheres including air and combustion environments. This can result in signiﬁcant internal nitridation penetration and affect the mechanical properties of the alloy. This topic is covered in detail in Chapter 4 “Nitridation.”

3.6 Summary Oxidation data for carbon and low-alloy steels, ferritic stainless steels, austenitic stainless steels, Fe-Cr-Ni alloys with 20-25Cr/30-40Ni, Fe-Cr-Al alloys, and superalloys including ODS alloys are presented. The data are presented in such a way for readers to make comparisons between alloys within the same alloy group or between alloys from different alloy groups. The chapter focuses on long-term oxidation behavior. Oxidation resistance of some nickel aluminides is also compared with that of several commercial nickel-base superalloys. A large portion of commercial hightemperature alloys relies on chromium for forming a protective chromium oxide scale to resist oxidation attack. The temperature range in which the chromium oxide scale is effective in providing adequate oxidation resistance varies from 540 to 1090 °C (1000 to 2000 °F). This group of alloys is frequently referred to as chromia formers. An adequate supply of chromium from the alloy interior to the metal surface to form and maintain a continuous, protective chromium oxide scale is necessary for an alloy to maintain its oxidation-resistant capability under

the service condition. Once the chromium supply drops below the critical level to maintain and reheal the chromium oxide scale, breakaway oxidation is initiated followed by rapid growth of oxides of base metals, such as iron, nickel, or cobalt oxides. Detailed discussion on the effect of chromium on the oxidation behavior and breakaway oxidation of various alloys is presented. Effects of other minor elements on the oxidation behavior of chromia formers are also discussed. A relatively small number of commercial high-temperature alloys rely on aluminum for forming a protective aluminum oxide scale to resist oxidation attack at very high temperatures. This group of alloys that form aluminum oxide scales are typically referred to as alumina formers. The oxidation behavior of alumina formers that include Fe-Cr-Al, Ni-Cr-Al, and ODS alloys is presented. The effect of aluminum as well as yttrium and sulfur on the oxidation resistance of alumina formers is discussed. Discussion also includes oxidation under high-velocity gas streams, oxidation of thin foils, effect of surface depletion of chromium in austenitic stainless steels, effect of water vapor, and catastrophic oxidation (oxidation under molten oxides). Most oxidation data were generated at 980 to 1200 °C (1800 to 2200 °F). However, many industrial applications are in the temperature range of 650 to 980 °C (1200 to 1800 °F), which are below the test temperatures at which most data were generated. More long-term oxidation data need to be generated at 650 to 980 °C (1200 to 1800 °F) for stainless steels, Fe-Ni-Cr and some simple Ni-Cr alloys to provide a more reliable database at intended application temperatures.

REFERENCES

1. G.Y. Lai, unpublished data, 2004 2. N. Birks and G.H. Meier, Introduction to High Temperature Oxidation of Metals, Edward Arnold Ltd., London, 1983 3. R.C. Weast, Ed., Physical Constants of Inorganic Compounds, Handbook of Chemistry and Physics, 65th ed., The Chemical Rubber Company, 1984, p B68– B161 4. C.E. Ramberg, P. Beatrice, K. Kurokawa, and W.L. Worrell, Mater. Res. Soc. Proc., Vol 322, C.L. Briant, J.J. Petrovic, B.P. Bewlay, A.K. Vasudevan, and H.H. Lipsitt,

pg 62

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

Chapter 3: Oxidation / 63

5. 6.

7. 8.

9.

10.

11.

12.

13.

14. 15.

16. 17. 18.

Ed., Materials Research Society, 1994, p 243 D. Caplan and M. Cohen, The Volatilization of Chromium Oxide, J. Electrochem. Soc., Vol 108 (No. 5), 1961, p 438 E.A. Gulbranson and S.A. Jansson, in Heterogenous Kinetics at Elevated Temperatures, G.R. Belton and W.L. Warrell, Ed., Plenum Press, 1970, p 181 H.C. Graham and H.H. Davis, Oxidation/ Vaporization Kinetics of Cr2O3, J. Am. Ceram. Soc., Vol 54 (No. 2), 1971, p 89 C.A. Stearns, F.J. Kohl, and G.C. Fryburg, Oxidative Vaporization Kinetics of Cr2O3 in Oxygen from 1000 to 1300 °C, J. Electrochem. Soc.: Solid-State Sci. Technol., Vol 121 (No. 7), 1974, p 945 H. Asteman, J.E. Svensson, L.G. Johansson, and M. Norell, Indication of Chromium Oxide Hydroxide Evaporation During Oxidation of 304L at 873 K in the Presence of 10% Water Vapor, Oxid. Met., Vol 52 (No. 1/2), 1999, p 95 N. Birks and F.S. Pettit, Environmental Effects During Application of Materials at Temperatures Above 1200 °C, Mater. Sci. Eng., A143, 1991, p 187 M. Danielewski, Kinetics of Gaseous Corrosion Processes, in Corrosion: Fundamentals, Testing and Protection, Vol 13A, ASM Handbook, ASM International, 2003, p 97 W.R. Patterson, in Designing for Automotive Corrosion Prevention, Proceedings P-78 (Troy, MI), Nov 8–10, 1978, Society of Automotive Engineers, p 71 J.B. Vrable, R.T. Jones, and E.H. Phelps, “The Application of Corrosion-Resistant High-Strength Low-Alloy Steels in the Chemical Industry,” presented at Fall Meeting, American Society for Metals, Chicago, 1977 R.T. Jones, in Process Industry Corrosion, NACE, 1986, p 373 R.C. John, Compilation and Use of Corrosion Data for Alloys in Various HighTemperature Gases, Paper No. 73, Corrosion/99, NACE International, 1999 A.W. Zeuthen, Heating, Piping and Air Conditioning, Vol 42 (No. 1), 1970, p 152 H.E. McGarrow, Ed., The Making, Shaping and Treating of Steel, United States Steel Corp., 1971, p 1136 I.G. Wright, in Corrosion, Vol 13, Metals Handbook, ASM International, 1987, p 97

19. A. Grodner, Weld. Res. Counc. Bull., No. 31, 1956 20. M. Walter, M. Schutze, and A. Rahmel, Behavior of Oxide Scales on 12Cr-1Mo Steel During Thermal Cycling, Oxid. Met., Vol 39 (No. 5/6), 1993, p 389 21. S.B. Lasday, Ind. Heat., March 1979, p 12 22. O.D. Sherby, Acta Metall., Vol 10, 1962, p 135 23. H.E. Eiselstein and E.N. Skinner, STP 165, ASTM, 1954, p 162 24. A. Moccari and S.I. Ali, Br. Corros. J., Vol 14 (No. 2), 1979, p 91 25. A.S. Brasunas, J.T. Gow, and O.E. Harder, Proc. ASTM, Vol 46, 1946, p 870 26. S. Kado, T. Yamazaki, M. Yamazaki, K. Yoshida, K. Yabe, and H. Kobayashi, Trans. Iron Steel Inst. Jpn., Vol 18 (No. 7), 1978, p 387 27. H.T. Michels, Met. Eng. Q., Aug 1974, p 23 28. J. Douthett, Heat Treating of Stainless Steels, in Heat Treating, Vol 4, ASM Handbook, ASM International, 1991, p 769 29. J.F. Grubb, in Proc. International Conf. on Stainless Steels, Iron and Steel Institute of Japan, Tokyo, 1991, p 944 30. G.Y. Lai, Unpublished data, 2005 31. W.E. Ruther and S. Greenberg, J. Electrochem. Soc., Vol 111, 1964, p 1116 32. J.C. Kelly, Today’s Versus Yesterday’s Stainless Steels, Mater. Perform., March 1992 33. “Heat Resistant Alloy Speciﬁcations and Operating Data,” Bulletin 150, Rolled Alloys, Temperance, MI, 1997 34. “RA 353MA alloy,” Bulletin No. 1069, Rolled Alloys, Temperance, MI 35. “Avesta 353MA,” Information 9270, Avesta AB, S-77480 Avesta, Sweden 36. G.R. Rundell, Evaluation of A High Silicon-Aluminum Alloy for Waste Incineration and Fluidized Bed Combustion Systems, Performance of High Temperature Materials in Fluidized Bed Combustion Systems and Process Industries, P. Ganesan and R.A. Bradley, Ed., ASM International, 1987, p 193 37. “Kanthal® Electrical Resistance and High Temperature Alloys,” Kanthal Alloys Data Sheet, Sandvik, Sweden 38. R. Berglund and B. Jonsson, New Powder Metallurgical Fe-Cr-Al Alloy for High Temperature and Corrosion Resistance in Furnace Uses, Ind. Heat., Oct 1989, p 21

pg 63

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

64 / High-Temperature Corrosion and Materials Applications

39. “Allegheny Ludlum Data Sheet,” Allegheny Ludlum, Pittsburgh, PA, 1990 40. P.T. Moseley, K.R. Hyde, B.A. Bellamy, and G. Tappin, The Microstructure of the Scale Formed During the High Temperature Oxidation of a Fecralloy® Steel, Corros. Sci., Vol 24 (No. 6), 1984, p 547 41. E. Tsuzi, The Role of Yttrium on the Oxide Adherence of Fe-24Cr Base Alloys, Metall. Trans. A, Vol 11A, Dec 1980, p 1965 42. T.A. Ramanarayanan, M. Raghavan, and R. Petkovic-Luton, Metallic Yttrium Additions to High Temperature Alloys: Inﬂuence on Al2O3 Scale Properties, Oxid. Met., Vol 22 (No. 3/4), 1984, p 83 43. J.L. Pandey, S. Prakash, and M.L. Mehta, Effect of Zirconium Concentration on High Temperature Cyclic Oxidation Behavior of Fe-15Cr-4Al at 1150 °C, J. Electrochem. Soc.: Solid-State Sci. Technol., Vol 135 (No. 1), Jan 1988, p 209 44. J.L. Pandey, S. Prakash, and M.L. Mehta, Effects of Varying the Zirconium Concentration and 1 wt.% Y on High Temperature Oxidation of Fe-15wt.%Cr-4wt.% Al Alloy under Isothermal and Cyclic Conditions, J. Less-Common Met., Vol 159, 1990, p 23 45. J.S. Benjamin, Metall. Trans., Vol 1, 1970, p 2943 46. M. Lundberg, L.-P. Bergmark, and M. Ramberg, Mechanical and Chemical Properties of 353MA—A Seamless Tube for High-Temperature Petrochemical Applications, 1999 Stainless Steel World Conf. Proc., Book 2, KCI Publishing BV, Zutphen, The Netherlands, 1999, p 563 47. J.C. Kelly and J.D. Wilson, Oxidation Rates of Some Heat Resistant Alloys, Heat-Resistant Materials II: Conf. Proc. Second International Conference on HeatResistant Materials, K. Natesan, P. Ganesan, and G. Lai, Ed., ASM International, 1995, p 53 48. P. Ganesan, G.D. Smith and C.S. Tassen, Mechanical Properties and Corrosion Resistance of Incoloy Alloy 803, Applications and Materials Performance: Proc. Nickel-Cobalt 97 International Symposium, F.N. Smith, J.F. McGurn, G.Y. Lai, and V.S. Sastri, Ed., The Metallurgical Society of CIM, Montreal, Canada, 1997, p 97 49. M.A. Harper, J.E. Barnes, and G.Y. Lai, Paper No. 132, Corrosion/97, NACE International, 1997

50. “Inconel Alloy 601 Brochure,” Huntington Alloys, Inc., 1969 51. P. Ganesan, G.D. Smith, and C.S. Tassen, Performance of A New Alloy in High Temperature Service, Paper No. 234, Corrosion/93, NACE, 1993 52. D.C. Agarwal, ThyssenKrupp VDM unpublished data 53. D.C. Agarwal and U. Brill, Performance of Alloy 602CA (UNS N06025) in High Temperature Environments up to 1200 °C, Paper No. 521, Corrosion/2000, NACE International, 2000 54. G.Y. Lai, J. Met., Vol 37 (No. 7), July 1985, p 14 55. G.Y. Lai, unpublished results, Haynes International, Inc., 1988 56. “Haynes Alloy No. 214,” H-3008B, Haynes International, Inc. Kokomo, IN 57. N. Birks and F.S. Pettit, Environmental Effects During Application of Materials at Temperatures above 1200 °C, Mater. Sci. Eng., Vol A143, 1991, p 187 58. G.Y. Lai, Sulﬁdation-Resistant Co-Cr-Ni Alloys with Critical Contents of Silicon and Cobalt, U.S. Patent No. 4711763, Dec 1987 59. G.Y. Lai, Meeting the Challenge of Materials Development for Coal Combustion Plants, Mater. High Temp., Vol 11 (No. 1–4), 1993, p 143 60. W. Crawford, in Proc. Conf. Frontiers of High Temperature Materials II, London, Inco Alloys International, May 1983, p 272 61. R.F. Singer, in Proc. Conf. Frontiers of High Temperature Materials II, London, Inco Alloys International, May 1983, p 336 62. Superalloys Source Book, M.J. Donachie, Jr., Ed., American Society for Metals, 1984 63. C.T. Sims, in Proc. Fifth International Symposium on Superalloys (Seven Springs, Champion, PA), Metallurgical Society of AIME, 1984, p 399 64. W. Betteridge and W.W.K. Shaw, Mater. Sci. Technol., Vol 3, 1987, p 682 65. B.H. Kear and E.R. Thompson, Science, Vol 208, May 23, 1980, p 847 66. M.J. Donachie and S.J. Donachie, Superalloys: A Technical Guide, 2nd ed., ASM International, 2002 67. C.A. Barrett, in Proc. Conf. Environmental Degradation of Engineering Materials, M.R. Louthan, Jr. and R.P. McNitt, Ed., Virginia Polytechnic Institute, 1977, p 319 68. M.F. Rothman, Cabot Corporation internal report, 1985

pg 64

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

Chapter 3: Oxidation / 65

69. H. Nagai, M. Okaboyashi, and H. Mitani, Trans. Jpn. Inst. Met., Vol 21, 1980, p 341 70. G.M. Kim, E.A. Gulbranson, and G.H. Meier, in Proc. Conf. Fossil Energy Materials Program, ORNL/FMP 87/4, May 19–21, 1987, R.R. Judkins, Ed., Oak Ridge National Laboratory, 1987, p 343 71. G.Y. Lai, unpublished results, Haynes International, Inc., 1988 72. R.H. Kane, J.W. Schultz, H.T. Michels, R.L. McCarron, and F.R. Mazzotta, Ref 30 of the paper by R.H. Kane in Process Industries Corrosion, B.J. Moniz and W.I. Pollock, Ed., NACE, 1986, p 45 73. G.Y. Lai, M.F. Rothman, and D.E. Fluck, Paper No. 14, Corrosion/85, NACE, 1985 74. J.J. Barnes and S.K. Srivastava, Paper No. 527, Corrosion/89, NACE, 1989 75. J.J. deBarbadillo and J.J. Fischer, Dispersion-Strengthened Nickel-Base and IronBase Alloys, in Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, Metals Handbook, ASM International, 1990, p 943 76. R.H. Kane, G.M. McColvin, T.J. Kelly, and J.M. Davidson, Paper No. 12, Corrosion/ 84, NACE, 1984 77. R.C. John, Oxidation Studies of Commercial Alloys at 871–1093 °C (1600– 2000 °F), in Heat-Resistant Materials II—Conf. Proc. Second International Conference on Heat-Resistant Materials, K. Natesan, P. Ganesan, and G. Lai, Ed., ASM International, 1995, p 41. 78. C.E. Lowell, D.L. Deadmore, and J.D. Whittenberger, Long-Term High-Velocity Oxidation and Hot Corrosion Testing of Several NiCrAl and FeCrAl Base Oxide Dispersion Strengthened Alloys, Oxid. Met., Vol 17 (No. 3/4), 1982, p 205 79. M.F. Rothman, “Oxidation Resistance of Gas Turbine Combustion Materials,” Paper No. 85-GT-10, presented at the Gas Turbine Conference (Houston, TX), March 18–21, 1985, ASME, 1985 80. J.V. Wright, “The Effects of Gas Velocity and of Temperature on the Oxidative Response of Selected Sheet Superalloys,” Paper No. 88-GT-281, presented at the Gas Turbine and Aeroengine Congress (Amsterdam, The Netherlands), June 6–9, 1988, ASME, 1988 81. U. Brill and T.I. Haubold, Corrosion Behaviour of Some Gas Turbine Alloys under High Velocity Burnt Fuels, Paper

82. 83. 84.

85. 86.

87. 88.

89.

90.

91.

92.

93. 94. 95. 96.

No. 522, Corrosion/2000, NACE International, 2000 G.Y. Lai, unpublished results, Haynes International, Inc., 1988 B. Hicks, Mater. Sci. Technol., Vol 3, Sept 1987, p 772 B. Gleeson, High-Temperature Corrosion of Metallic Alloys and Coatings, in Corrosion and Environmental Degradation, Vol II, Materials Science and Technology, M. Schutze, Ed., Wiley-VCH, Weinheim, Germany, 2000, p 173 G.Y. Lai, unpublished results, Haynes International, Inc., 1996 B. Gleeson and M.A. Harper, The LongTerm, Cyclic-Oxidation Behavior of Selected Chromia-Forming Alloys, Oxid. Met., Vol 49 (No. 3/4), 1998, p 373 H.E. Evans, D.A. Hilton, R.A. Holm, and S.J. Webster, Oxid. Met., Vol 14, 1980, p 235 M.P. Brady, B. Gleeson, and I.G. Wright, Alloy Design Strategies for Promoting Protective Oxide-Scale Formation, JOM, Jan 2000, p 16 W.J. Quadakkers and K. Bongartz, The Prediction of Breakaway Oxidation for Alumina forming ODS Alloys Using Oxidation Diagrams, Werkst. Korros., Vol 45, 1994, p 232 I. Gurrappa, S. Weinruch, D. Naumenko, and W.J. Quadakkers, Factors Governing Breakaway Oxidation of FeCrAl-Based Alloys, Mater. Corros., Vol 51, 2000, p 224 W.J. Quadakkers, Growth Mechanisms of Oxide Scales on ODS Alloys in the Temperature Range 1000–1100 °C, Werkst. Korros., Vol 41, 1990, p 659 J. Klower and J.G. Li, Effects of Yttrium on the Oxidation Behavior of Iron-ChromiumAluminum Alloys, Mater. Corros., Vol 47, 1996, p 545 J.G. Smeggil, A.W. Funkenbusch, and N.S. Bornstein, High Temp. Sci., Vol 20, 1985, p 163 A.W. Funkenbusch, J.G. Smeggil, and N.S. Bornstein, Met. Trans. A, Vol 16, 1985, p 1164 J.G. Smeggil, A.W. Funkenbusch, and N.S. Bornstein, Met. Trans. A, Vol 17, 1986, p 923 J.G. Smeggil, Some Comments on the Role of Yttrium in Protective Oxide Scale Adherence, Mater. Sci. Eng., Vol 87, 1987, p 261

pg 65

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:46PM Plate # 0

66 / High-Temperature Corrosion and Materials Applications

97. P.Y. Hou and J. Stringer, Oxide Scale Adhesion and Impurity Segregation at the Scale/Metal Interface, Oxid. Met., Vol 38 (No. 5/6), 1992, p 323 98. P.Y. Hou, Compositions at Al2O3/ FeCrAl Interfaces after High Temperature Oxidation, Mater. Corros., Vol 51, 2000, p 329 99. G.Y. Lai, “Several Modern Wrought Superalloys for Gas Turbine Applications,” Paper 96-TA-030, presented at ASME Turbo Asia ’96 (Jakarta, Indonesia), Nov 5–7, 1996 100. N.J. Simms, R. Newton, J.F. Norton, A. Encinas-Oropesa, J.E. Oakey, J.R. Nicholls, and J. Wilber, Mater. High Temp., Vol 20 (No. 3), 2003, p 439 101. J. Klower, Factors Affecting the Oxidation Behaviour of Thin Fe-Cr-Al Foils, Mater. Corros., Vol 49, 1998, p 758 102. P.J. Maziasz, B.A. Pint, R.W. Swindeman, K.L. More, and E. Lara-Curzio, “Advanced Stainless Steels and Alloys for High Temperature Recuperators,” DOE/CETC/ CANDRA Workshop on Microturbine Applications (Calgary, Alberta, Canada), Jan 21–23, 2003 103. B.A. Pint, “The Effect of Water Vapor on Cr Depletion in Advanced Recuperator Alloys,” GT2005-68495, ASME Turbo Expo 2005 (Reno-Tahoe, Nevada), June 6–9, 2005 104. B.A. Pint, Stainless Steels with Improved Oxidation Resistance for Recuperators, J. Eng. Gas Turbines Power, Vol 128, 2006, p 1 105. B.A. Pint, “The Effect of Water Vapor on Cr Depletion in Advanced Recuperator Alloys, GT2005-68495,” ASME Turbo Expo 2005, (Reno-Tahoe, Nevada), June 6–9, 2005 106. C.W. Tuck, M. Odgers, and K. Sachs, Scaling Rates of Pure Iron and Mild Steel in Oxygen, Steam, Carbon Dioxide in the Range 850°–1000 °C, Anti-Corrosion, June 1966, p 14 107. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, 1992, p 13-2 108. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, 1992, p T-17 109. K. Segerdahl, J.E. Svensson, and L.G. Johansson, The High Temperature Oxidation of 11% Chromium Steel: Part

110.

111.

112. 113.

114.

115.

116.

117. 118. 119.

120. 121. 122. 123.

124. 125.

I—Inﬂuence of pH2 O , Mater. Corros., Vol 53, 2002 p 247 K. Segerdahl, J.E. Svensson, and L.G. Johansson, The High Temperature Oxidation of 11% Chromium Steel: Part II— Inﬂuence of Flow Rate, Mater. Corros., Vol 53, 2002, p 479 B.A. Pint, R. Peraldi, and P.F. Tortorelli, The Effect of Alloy Composition on the Performance of Stainless Steels in Exhaust Gas Environments, Paper No. 03499, Corrosion/2003, NACE International, 2003 B.A. Pint, private communication, 2007 R.L. McCarron and J.W. Schultz, in Proc. Symp. High Temperature Gas-Metal Reactions in Mixed Environments, AIME, 1973, p 360 C.C. Clark and W.R. Hulsizer, Superalloys Development for Gas Turbines Operating in the Marine Environment, Conf. Proc., Gas Turbine Materials Conference, Naval Ship Engineering Center, 1972, p 35 C. Sarioglu et al., The Adhesion of Alumina Films to Metallic Alloys and Coatings, Mater. Corros., Vol 51, 2000, p 358 K. Onal, M.C. Maris-Sida, G.H. Meier, and F.S. Pettit, Water Vapor Effects on the Cyclic Oxidation Resistance of Alumina Forming Alloys, Mater. High Temp., Vol 20 (No. 3), 2003, p 327 G. Welsch and P.D. Desai, Ed., Oxidation and Corrosion of Intermetallic Alloys, Purdue University, 1996 K. Natesan, Oxid. Met., Vol 30 (No. 1/2), 1988, p 53 W.C. Leslie and M.C. Fontana, Paper No. 26, 30th Annual Convention of ASM (Philadelphia, PA), Oct 25–29, 1948 J.K. Meijering and G.W. Rathenau, Nature, Vol 165, Feb 11, 1950, p 240 A.D. Brasunas and N.J. Grant, Iron Age, Aug 17, 1950, p 85 S.S. Brennor, J. Electrochem. Soc., Vol 102 (No. 1), Jan 1955, p 16 J.H. DeVan, “Catastrophic Oxidation of High Temperature Alloys,” ORNL-TM-51, Oak Ridge National Laboratory, Oak Ridge, TNn, Nov 10, 1961 J.W. Sawyer, Trans. TMS-AIME, Vol 221, 1961, p 63 A. de S. Brasunas and N.J. Grant, Trans. ASM, Vol 44, 1950, p 1133

pg 66

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p67-96 DOI: 10.1361/hcma2007p067

30/10/2007 3:55PM Plate # 0

pg 67

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 4

Nitridation 4.1 Introduction In air or combustion atmospheres containing nitrogen, nitridation can take place under certain exposure conditions. In most cases, oxidation dominates the high-temperature corrosion reaction. However, nitridation can take place for some alloys when oxide scales no longer provide protection. The alloys that are particularly susceptible to oxidation/nitridation attack are those containing strong nitride formers, such as titanium and aluminum. Many high-temperature nickel-base alloys containing both aluminum and titanium are strengthened by γ′ phase, Ni3(Al,Ti). For these alloys, nitridation by forming internal nitrides of aluminum and titanium can deplete the surface layer with aluminum and titanium, thus weakening the alloy. Under a high-velocity combustion gas stream with severe thermal cycling, similar to the conditions in “ﬂying” gas turbines (aircraft engines), nitridation can be particularly severe in oxidation/nitridation attack. In nitrogen-base atmospheres, such as N2 or N2-H2, metals and alloys can also suffer nitridation attack. This type of atmosphere is often used as a protective atmosphere in heat treating and sintering operations. Molecular nitrogen can be severely nitriding for many metals and alloys, particularly when temperatures are sufﬁciently high. Ammonia (NH3) is a commonly used nitriding gas for case hardening at temperatures from 500 to 590 °C (925 to 1100 °F) (Ref 1). Furnace equipment and components repeatedly subjected to these service conditions frequently suffer brittle failures as a result of nitridation attack. Carbonitriding is another important method of case hardening that produces a surface layer of both nitrides and carbides. The process is typically carried out at 700 to 900 °C (1300 to 1650 ° F) in ammonia, with additions of carbonaceous gases, such as CH4 (Ref 2). Thus, the heat treat retort, ﬁxtures, and other furnace equipment are subject to both nitridation and carburization.

Cracked ammonia (i.e., ammonia that is completely dissociated into H2 and N2) provides an economical protective atmosphere for processing metals and alloys. Many bright annealing operations for stainless steels use a protective atmosphere consisting of N2 and H2, generated by dissociation of ammonia. With three parts H2 and one part N2 produced in cracked ammonia, nitridation is less critical for the heat treating equipment. In the chemical processing industry, nitriding environments are generated by processes employed for production of ammonia, nitric acid, melamine, and nylon 6-6 (Ref 3, 4). Ammonia is produced by reacting nitrogen with hydrogen over a catalyst at temperatures of typically 500 to 550 °C (930 to 1020 °F) and pressures of 200 and 400 atm. Commercial processes for ammonia synthesis are discussed in detail in Ref 5. The converter, where the ammonia synthesis reaction takes place, may suffer nitridation attack. Brittle failure of the welds for the waste heat boiler of an ammonia plant has been reported by Van der Horst (Ref 6). These tube-to-tube sheet welds were high-nickel alloy 182, which suffered nitridation attack. Production of nitric acid involves the oxidation of ammonia over a platinum gauze catalyst at temperatures of about 900 °C (1650 °F) (Ref 5). The catalyst grid support structure and other processing components in contact with ammonia may also be susceptible to nitridation attack. Figure 4.1 shows the nitrided structure of a nickel-base alloy catalyst grid support after two years of service in a nitric acid plant.

4.2 Thermodynamic Considerations When metal is exposed to nitrogen gas at elevated temperatures, nitridation proceeds according to: 1=2N2 (gas)=[N] (dissolved in metal)

ð4:1Þ

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 68

68 / High-Temperature Corrosion and Materials Applications

[%N] = k(pN2 )1=2

ð4:2Þ

where k is the equilibrium constant and pN2 is the partial pressure of N2 in the atmosphere. In nitrogen-base atmospheres, the nitriding potential is proportional to (pN2 )1=2 . Increasing the nitrogen partial pressure (or nitrogen concentration) increases the thermodynamic potential for nitridation. Molecular nitrogen is less aggressive than ammonia in terms of nitridation of metals. However, when metal is heated to excessively high temperatures (e.g., 1000 °C (1830 °F) or higher), nitridation by molecular nitrogen can become a serious material issue. Under oxidation/nitridation conditions, nitrogen molecules permeate through cracks and pores and reach the metal underneath the oxide scales when the oxide scale is no longer protective. Nitridation then proceeds by dissociation of

100 µm

nitrogen molecules and absorption of nitrogen atoms by the metal following Reaction 4.1. Nitridation often takes place in the metal at the vicinity of cracks developed under creep conditions in air or N2-containing combustion atmospheres. In this case, oxides are often associated with the crack under creep deformation in air. Oxidation consumes the oxygen molecules from air, which penetrates into the crack, thus depleting oxygen and increasing nitrogen potential (or concentration) around the crack. As a result, nitridation takes place in the vicinity of the creep crack. When the environment is ammonia (NH3) or contains NH3, metals or alloys may undergo rapid nitridation reactions. It is precisely for this reason that NH3 is frequently used for case hardening. Ammonia is metastable and dissociates into molecular N2 and molecular H2 when heated to elevated temperatures. Once NH3 is completely dissociated into N2 and H2, the nitriding potential is deﬁned by Reaction 4.3. To increase nitrogen absorption by steel, molecular NH3 should be allowed to dissociate on the steel surface, thus allowing dissociated atomic nitrogen to be dissolved at the metal surface (Ref 7, 8). Thus, to increase nitridation reactions, it is necessary to bring as much fresh, uncracked NH3 as possible in contact with the surface of the metal to minimize the production of molecular nitrogen. At temperatures below 600 °C (1110 °F) and at high gas ﬂow rates, the production of nitrogen is minimized and the nitrogen solubility at the surface of iron is determined by (Ref 8):

(a)

100 µm (b)

Fig. 4.1

Alloy X (Ni-22Cr-18.5Fe-9Mo-0.6W) catalyst grid support structure bar after 2 years of service in a nitric acid plant. (a) Internal nitride precipitates (about 20 mils in depth) containing mainly chromium-rich nitrides along with some carbides formed during thermal aging. (b) Microstructure in the unaffected interior containing mainly carbides due to thermal aging at the service temperature

NH3 $ 3=2 H2 +[N] (dissolved in Fe)

ð4:3Þ

[%N]=k(pNH3 =(pH2 ))3=2

ð4:4Þ

where k is the equilibrium constant, and pNH3 , and pH2 are partial pressures of NH3 and H2, respectively. The nitriding potential is proportional to pNH3 =( pH2 )3=2 . Increasing ammonia partial pressure (or concentration) in the atmosphere increases the thermodynamic potential for nitridation. When nitrogen in the metal exceeds its solubility limit, nitrides will then precipitate out. The nitrides of important alloying elements for engineering alloys are tabulated in Table 4.1 (Ref 9). For iron, nickel, and cobalt, three important alloy bases for high-temperature alloys, only iron forms stable nitrides. No nitrides of nickel and cobalt have been reported. The relative stabilities among various nitrides can

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 69

Chapter 4: Nitridation / 69

alloys with nickel ranging from 0 to 35% (Ref 19). In Fe-18Cr-Ni-N system, increasing nickel reduces the solubility of nitrogen, as shown in Fig. 4.3 (Ref 11). It is also shown in Table 4.2 that the solubility of nitrogen in stainless steels increased with increasing temperature. Alloys with higher nitrogen solubilities generally exhibit less resistance to nitridation attack. e 4N

+10

2F

N H

3

0

2

N o2

–10

2M

2 2C

2 1/

–40

Iron Chromium Molybdenum Tungsten Aluminum Titanium Niobium Tantalum Zirconium Hafnium Silicon Vanadium Boron Manganese Magnesium

Fe4N CrN MoN WN AlN TiN NbN TaN ZrN HfN Si3N4 VN BN Mn4N Mg3N

bN

–60 2

a

3N 2

–70

VN N2

C

AI

N 2

2

Ta

–100 N

2

–120

Ti

N

2

–130

Zr

2M + N2 = 2MN

–140 Nb4N3 Ta3N5

0

500

1000

Hf4N3

Fig. 4.2 Mn3N

102

N2 partial pressure, bar

Nickel, %

20

γ + Cr2N

γ 10 α+γ 0.1

0.2

0.3

1000 °C (1830 °F)

γ + CrN

1 γ + Cr2N

0.1 10 –2

γ

γ+π

10 –3 10 –4 0

α + γ + Cr2N

2000

Standard free energy of formation for selected nitrides. Source: Ref 10

10

0

1500

Temperature, °C

Source: Ref 9

0 α

BN

g3

N

–90

V 2N Mn2N

2

M

–110

Hf3N2

N

2

Fe2N Cr2N Mo2N W 2N Ti2N Nb2N Ta2N

N4 Si 3

–50

Table 4.1 Nitrides of important alloying elements for engineering alloys Nitrides

Cr

r 2N

–30

–80

Element

N

–20

∆G°, Kcal

be compared in terms of their free energies of formation, as illustrated in Fig. 4.2 (Ref 10). Physical-chemical properties of some of these nitrides can be found in Ref 9. The types of nitrides that are likely to form in the alloy can be predicted by examining the phase-stability diagram. Figure 4.3 shows a phase-stability diagram of Fe-18Cr-Ni-N system at 900 °C (1650 °F), indicating phase regions of Cr2N in γ (or α+γ) phase as a function of nickel and nitrogen contents (Ref 11). Nitride phases formed in alloys are also dependent on nitrogen partial pressure ( pN2 ), as shown in Fig. 4.4 for Ni-Cr-N system at 1000 °C (1830 °F) (Ref 12). Nitrogen solubility in the alloy is important in affecting the nitridation resistance of the alloy. Table 4.2 summarizes some nitrogen solubility data for iron, stainless steels, and nickel alloy (Ref 13–18). Iron and stainless steels exhibit signiﬁcantly higher nitrogen solubility than a nickel-base alloy (Ni-20Fe). Nickel has been found to decrease the nitrogen solubility in Fe-Ni

0.4

γ+α 10

20

30

40

50

60

Cr concentration, wt %

Nitrogen, %

Fig. 4.3

Phase stability diagram for Fe-18Cr-Ni-N system at 900 °C (1650 °F). Source: Ref 11

Fig. 4.4

Phase stability diagram for the Ni-Cr-N system as a function of N2 partial pressure at 1000 °C (1830 °F). Source: Ref 12

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

70 / High-Temperature Corrosion and Materials Applications

Table 4.2

Nitrogen solubility in metals and alloys

Nitrogen, wt%

0.06 0.26 max 0.02 (at pN2 : 1 atm) 0.125 0.177 0.190 0.258 0.281 0.18 0.18 0.21 0.26 0.26 0.0001 (at pN2 : 1 atm)

Metal or alloy

α-Fe γFe Fe-10%Ni Type 304 Type 304 Type 304 Type 304 Type 304 Fe-18Cr-12Ni-2Ti Fe-18Cr-12Ni-2Ti Fe-18Cr-12Ni-2Ti Fe-18Cr-12Ni-2Ti Fe-18Cr-12Ni-2Ti Ni-20%Fe

Temperature, °C (°F)

502 (936) γ region of Fe-C 1000 (1832) 538 (1000) 593 (1100) 927 (1700) 954 (1749) 981 (1800) 985 (1805) 1040 (1905) 1093 (2000) 1150 (2100) 1210 (2210) 1000 (1832)

Ref

13 14 15 16 17 17 17 17 18 18 18 18 18 15

4.3 Internal Nitridation in Oxidizing Environments In air or oxidizing combustion environments, oxidation usually dominates high-temperature corrosion reactions. However, under certain conditions, alloys can suffer internal nitridation attack along with oxidation. Internal nitridation attack, when it occurs, can penetrate farther into the metal interior than oxidation, thus signiﬁcantly affecting the creep-rupture behavior of the alloy by accelerating the creep crack growth. Discussion of internal nitridation under no external stresses and under creep conditions is presented in sections 4.3.1 and 4.3.2. 4.3.1 Internal Nitridation in Air under No External Stress Some high-temperature alloys that contain strong nitride formers such as aluminum and titanium can suffer internal nitridation even in air environments. In an air oxidation study for two nickel-base alloys, IN939 (Ni-22Cr-20Co-3.8Ti1.4Al-2W-1Nb-1.3Ta) and IN738LC (Ni-16Cr9Co-3.5Ti-3.3Al-1.8Mo-1Nb-1.8Ta) at 700, 900, and 1100 °C (1290, 1650, and 2010 °F). Litz et al. (Ref 20) observed internal titanium nitrides (needle shape) formed in front of internal aluminum oxides that formed underneath the external oxide scales. In an oxidation study of alloy 800HT (Fe-21Cr-32Ni-0.5Al-0.5Ti) in air at 980 °C (1800 °F) for about 2 years (720 days), Harper et al. (Ref 21) observed Widmanstätten acicular chromium-rich nitrides along with aluminum nitrides that formed below chromiumrich oxides. Lai (Ref 22) observed internal aluminum nitrides (needle shape) that formed

underneath the external oxide scale and internal oxides in alloy 601 exposed to a furnace oxidizing atmosphere for about 4 to 5 years at temperatures probably between 760 and 870 °C (1400 and 1600 °F), as shown in Fig. 4.5. Severe thermal cycling that causes cracking and spalling of oxide scales can also result in severe internal nitridation. Han and Young (Ref 23) conducted cyclic oxidation tests by heating the specimens to 1100 °C (2010 °F) in still air for 1 h followed by cooling to room temperature for 15 min then repeating the cycle again for 260 cycles. The alloys investigated were Ni-24 to 38%Cr-14 to 25%Al. The specimens suffered severe oxide scale spallation. The internal nitridation attack was found to be extensive, and the nitridation zone consisted of AlN beneath Cr2O3 and Al2O3, then AlN+Cr2N, and then AlN in the deepest region (Ref 23). Douglas (Ref 24) indicated that the diffusivity of nitrogen appears to be two orders of magnitude greater than that of oxygen in nickel or nickel alloys. Table 4.3 summarizes the diffusion coefﬁcients of nitrogen in nickel and iron alloys compared with those of oxygen and carbon in nickel, based on the diffusivity data from Rubly and Douglas (Ref 25, 26), Grabke and Peterson (Ref 27), Park and Alstetter (Ref 28), and Gruzin et al. (Ref 29). The diffusivity of nitrogen is also on the same order of magnitude as that of carbon as shown in Table 4.3. It is thus not surprising to ﬁnd internal nitrides were advancing in front of internal oxides. 4.3.2 Internal Nitridation at Creep Cracks in Air Environment During creep testing in air, extensive internal nitridation can develop in the vicinity of cracks. Brickner et al. (Ref 30) found that types 302, 304, and 310 stainless steels showed signiﬁcant nitridation after creep-rupture testing in air at 870 °C (1600 °F) in less than 1000 h. Acicular nitrides (believed to be chromium nitrides) in a Widmanstätten pattern were found to form extensively in the vicinity of microcracks, as shown in Fig. 4.6 (Ref 30). Extensive nitridation was conﬁrmed by the chemical analysis of the tested specimens for nitrogen, which showed the nitrogen content was increased from about 0.058% before testing to 0.30 to 0.53% after creep-rupture testing (Ref 30). Extensive internal nitrides were also observed in the vicinity of creep cracks in alloy 253MA (Fe-21Cr-11Ni)

pg 70

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 71

Chapter 4: Nitridation / 71

(a) 0.1 mm

1 2

(c) 10 µm

(b) 0.0010 in Formation of internal aluminum nitrides beneath external oxide scales and internal oxides in alloy 601 after exposing to a furnace oxidizing atmosphere for approximately 4 to 5 years in a temperature range of 760 to 870 °C (1400 to 1600 °F). (a) Optical micrograph showing the external oxide scales and the internal oxides, and then the chromium denuded zone immediately below, followed by internal nitrides underneath the denuded zone. (b) Optical micrograph at higher magnification showing internal nitrides. (c) SEM (backscattered electron image) showing internal aluminum nitrides and the EDX analysis of nitrides. Results of the semiquantitative EDX analysis (at.%) on internal aluminum nitrides are summarized as:

Fig. 4.5

Phase 1 Phase 2

Table 4.3

41.5% Al, 24.7% Ni, 10.6% Cr, 6.8% Fe, 5.5% Ti, and 10.0% N 58.0% Al, 13.1% Ni, 6.1% Cr, 3.6% Fe, 0.8% Ti, and 17.3% N

Diffusion coefficients of nitrogen, oxygen, and carbon. Diffusion coefﬁcient, cm2/s

Temperature, °C (°F)

700 (1290) 800 (1470) 900 (1650)

N in Ni-Alloys(a)

N in Fe-20Ni(b)

O in Ni(c)

C in Ni(d)

9.5 × 10−9 to 2.3 × 10−8 3.2 × 10−8 to 8.5 × 10−8 1.4 × 10−7 to 4.0 × 10−7

1.17 × 10−8 3.86 × 10−8 1.47 × 10−7

7.4 × 10−11 5.05 × 10−10 2.3 × 10−9

3.19 × 10−9 1.47 × 10−8 0.55 × 10−7

(a) Ref 25, 26. (b) Ref 27. (c) Ref 28. (d) Ref 29

after creep-rupture testing at 900 °C (1652 °F) for 11,800 h in air (Ref 31), in 800H after creep-rupture testing at 900 to 1000 °C (1650 to 1830 °F) in air (Ref 32), and in alloy 800H during the creep crack growth testing at 1000 °C (1830 °F) (Ref 33). The nitrides identiﬁed were Cr2N in 253MA (Ref 31), Cr2N (major) and CrN (minor) in 800H creep-ruptured specimens

(Ref 32), and Cr2N and AlN in 800H creep-crack growth specimens (Ref 33). Hoffman and Lai (Ref 34) investigated an alloy 800HT pigtail that suffered cracking after about 7.5 years of service in a hydrogen reformer. The pigtail section, which was exposed to air at approximately 850 °C (1565 °F) at the outside of the reformer furnace, was found to show

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 72

72 / High-Temperature Corrosion and Materials Applications

extensive blocky precipitates along grain boundaries and acicular precipitates in the matrix in the vicinity of cracks at the tube outer-diameter side (exposed to air). Samples were cut from this pigtail section and solution-annealed in a furnace for 1 h at 1093, 1149, and 1204 °C (2000, 2100, and 2200 °F), respectively. Microstructural examination of these samples indicated that both blocky, grain-boundary phases and acicular phases in the matrix remained in the microstructure and were not put back into solution, suggesting those phases were nitrides instead of carbides. Also, chemical analysis of the samples from the pigtail indicated that carbon content remained about the same as that of the material before service (about 0.07%), while nitrogen content was about 0.27 wt% with a nominal nitrogen content of about 0.02% prior to service. The process gas in the tube contained essentially no nitrogen (typically about 0.01%). Thus, the nitrogen ingress into the tube was primarily from air from the outside diameter side of the pigtail. Using scanning electron microscopy with energy-dispersive x-ray spectroscopy (SEM/EDX) analysis, acicular phases were found to be enriched in aluminum, while blocky phases were enriched in chromium; the former was believed to be aluminum nitride and the latter chromium nitride. Figure 4.7 shows the acicular aluminum nitrides and blocky chromium nitrides that remained in the microstructure after solution annealing at 1150 °C (2100 °F) for 1 h for the sample from the straight section of the pigtail (Ref 34). Figure 4.8 shows extensive nitride formation in the vicinity of creep cracks in

the bend section of the pigtail from another hydrogen reformer (Ref 34). When creep cracks initially develop at the metal surface during creep testing in air, oxidation occurs at the crack surface including the crack tip. The oxide scales formed on the crack surface become nonprotective due to creep deformation, thus causing the oxygen potential to decrease signiﬁcantly with concurrent increase in nitrogen potentials at the oxide/metal interface. As a result, nitrogen is absorbed by the metal and is diffused into the metal in the vicinity of cracks to form internal nitrides.

100 µm

Fig. 4.7

Acicular aluminum nitrides and blocky chromium nitrides, which formed in the vicinity of the creep cracks in alloy 800HT pigtail in a hydrogen reformer, were not dissolved into solution after the sample was resolution annealed at 1150 °C (2100 °F) for 1 h. Source: Ref 34

100 µm

Fig. 4.6 Acicular nitrides (believed to be chromium nitrides) in a Widmanstätten pattern formed in the vicinity of creep cracks in Type 302SS after creep-rupture testing at 870 °C (1600 °F) in less than 1000 h. Original magnification, 500×. Source: Ref 30

Fig. 4.8

Extensive aluminum and chromium nitrides formed in the vicinity of creep cracks in the bend section of an alloy 800H pigtail in another hydrogen reformer. Source: Ref 34

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 73

Chapter 4: Nitridation / 73

4.3.3 Oxidation and Nitridation in Combustion Atmospheres High-temperature alloys that are exposed to a high-velocity, oxidizing combustion gas stream at high temperatures are susceptible to internal nitridation attack. In investigating transition duct component failures in a land-based gas turbine, Swaminathan and Lukezich (Ref 35) observed that alloy 617 (Ni-22Cr-12.5Co-9Mo-1.2Al) had suffered severe oxidation and nitridation attack from both the air side (outside diameter side of the transition duct) and the combustion side (inside diameter side of the duct) after service for slightly less than 2 years (14,000 h). Extensive internal nitridation from both air and combustion gas sides of alloy 617 transition duct is shown in Fig. 4.9 (Ref 35). Alloy 230 (Ni-22Cr-14W2Mo-0.3Al-La) was also tested for 16,000 h as a transition duct, suffering similar oxidation/nitridation attack (Ref 35). However, no aluminum nitrides were observed in alloy 230. Signiﬁcant nitrogen pickup was observed from both transition ducts. Results of the chemical analyses of nitrogen from samples at the exit end of the transition duct and at the location far away from the exit for both alloy 617 and 230 transition ducts are shown in Table 4.4 (Ref 35). Lai (Ref 34) used a high-velocity dynamic burner rig test to simulate a gas turbine combustion environment. The simulated combustion gas stream was generated by burning fuel oil (a mixture of two parts No. 1 fuel and one part No. 2 fuel) with an air-to-fuel ratio of approximately 50 to 1 in a laboratory burner rig. Most of the air for combustion was from a compressor. When combusted with fuel oil, a

OD

ID

high-velocity combustion gas stream with about 0.3 Mach (100 m/s) was generated. Specimens were loaded in a carousel specimen holder that rotated at 30 rpm during testing to ensure all the specimens were subjected to the same test conditions. Furthermore, the specimens were subjected to severe thermal cycling once every 30 min by lowering the carousel from the test chamber followed by rapid fan-air cooling to below 260 °C (500 °F) for 2 min before returning the carousel back to the test chamber. A schematic of this dynamic burner rig is shown in Fig. 4.10. The combustion gas was determined to consist of 76% N2, 13% O2, 6% CO2, and 5% H2O. The test on alloy 617 produced severe internal nitridation, with the microstructure very similar to that observed by Swaminathan and Lukezich (Ref 35) from the transition duct in a land-based gas turbine power plant. Figure 4.11 shows the microstructure of an alloy 617 specimen after testing at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. Extensive needle-shape aluminum nitrides were observed. Some blocky chromium nitrides were observed to form right below the external oxide scales. Aluminum nitrides were found to penetrate farther into the metal interior than chromium nitrides. The oxide scales were found to be porous and nonprotective. Alloy 230 was included in the test and found to show less nitridation attack under the same test condition. Nitridation in alloy 230 involved only the formation of internal chromium nitrides below internal chromium oxides and chromium denuded zone with no aluminum nitrides. Two other common combustor alloys, alloys X and 263, were also included in the test. Figures 4.12 and 4.13 show the microstructures of alloys X and 263, respectively, after testing at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. Alloy X showed mainly internal chromium nitrides, while alloy 263 showed mainly tiny needle-shaped nitrides, presumably titanium

Table 4.4 Nitrogen contents at different locations of alloy 617 and alloy 230 transition ducts

Fig. 4.9

Cross section (2.5 mm, or 0.1 in.) of an alloy 617 (Ni22Cr-12.5Co-9Mo-1.2Al) transition duct after service for less than 2 years (about 14,000 h) in a land-based gas turbine, showing extensive formation of both aluminum and chromium nitrides from both air side (outside diameter of the transition duct) and the combustion gas side (inside diameter of the duct). Source: Ref 35

Transition duct/service

Location

Alloy 617/14,000 h

Exit Far away from exit Exit Far away from exit

Alloy 230/16,000 h Source: Ref 35

Nitrogen, wt%

0.24 0.004 0.22 0.05

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 74

74 / High-Temperature Corrosion and Materials Applications

Thermocouple for steady-state control

Thermocouple for recording specimen temperature history

50 mm square insulated flame tunnel Specimen temperature measured by pyrometer

Compressed inlet air 425 °C (800 °F)

Fuel

Combustor Rotating shaft

Blower (thermal shock)

Thermal cycle

Fig. 4.10

The dynamic burner rig used by Lai (Ref 36) for simulating a gas turbine combustion environment in evaluating the oxidation/nitridation behavior of gas turbine combustor alloys. Courtesy of Haynes International, Inc.

50 µm

50 µm

Fig. 4.11

Some blocky chromium nitrides and extensive acicular aluminum nitrides formed in alloy 617 after testing in the dynamic burner rig at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. The combustion gas stream with about mach 0.3 (100 meter/s) consisted of 76% N2, 13% O2, 6% CO2, and 5% H2O. Source: Ref 36

nitrides. Similar to both alloy 617 and 230, alloys X and 263 also exhibited nonprotective, porous oxide scales.

Fig. 4.12

Extensive internal chromium nitrides formed in alloy X after testing in the dynamic burner rig at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. The combustion gas stream with about mach 0.3 (100 m/s) consisted of 76% N2, 13% O2, 6% CO2, and 5% H2O. Source: Ref 36

X-ray diffraction analysis of the oxide scales formed on alloys 230, 617, and X was performed with the results summarized in Table 4.5. The results showed that NiO and NiCr2O4 along with

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 75

Chapter 4: Nitridation / 75

Cr2O3 oxides made up the external oxide scales for alloy 230, and NiO and Cr2O3 made up the oxide scales for alloys 617 and X. With the formation of NiO oxides, the alloys were no longer protected by Cr2O3 oxide scales. An electrolytic extraction technique was used to extract precipitate phases in the tested specimens for analysis, and the results are summarized in Table 4.5, revealing essentially nitride phases in all three alloys. Alloy 230 also showed M6C carbides, which were the carbides in the alloy in the as-solution-annealed condition. Bulk nitrogen

contents for the original samples (before testing) and those after testing are summarized in Table 4.6, showing signiﬁcant nitrogen adsorption for alloys 263, 617, and X. Alloy 230 showed little nitrogen adsorption. Bulk carbon contents for the samples before and after testing were also determined, and the results clearly showed that carburization was not involved (Table 4.7). The overall test results in terms of weight loss (due to oxidation), metal loss (due to oxidation), internal oxidation, internal nitridation, and total depth of attack are summarized in Table 4.8. In continuing his testing program for the same simulated gas turbine environment involving the same four combustor alloys (i.e., 230, 617, 263, and X) at the same test temperature and duration Table 4.6 Results of bulk nitrogen analysis before and after testing at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling Alloy

Original nitrogen, wt%

Nitrogen after testing, wt%

230 263 617 X

0.05 0.004 0.03 0.04

0.06 0.42 0.52 0.57

Source: Ref 36

50 µm

Fig. 4.13

Extensive internal nitrides (believed to be titanium nitrides) formed in alloy 263 after testing in the dynamic burner rig at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. The combustion gas stream with about mach 0.3 (100 m/s) consisted of 76% N2, 13% O2, 6% CO2, and 5% H2O. Source: Ref 36

Table 4.5 Results of x-ray diffraction analysis on oxide scales and extracted precipitate phases for alloys 230, 617, and X after testing at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling Alloy

Surface oxide scales

230

NiCr2O4 (strong) Cr2O3 (medium) NiO (medium) NiO (strong) Cr2O3 (medium) NiO (strong) Cr2O3 (weak)

617 X Source: Ref 36

Table 4.7 Results of bulk carbon analysis before and after testing in the dynamic burner rig at 980 °C (1800 °F) for 1000 h with 30 minute thermal cycling Alloy

Original carbon, wt%

Carbon after testing, wt%

230 263 617 X

0.09 0.06 0.05 0.08

0.09 0.03 0.04 0.01

Source: Ref 36

Table 4.8 Test results in terms of weight loss, depth of oxidation penetration, depth of nitridation, and total depth of attack after testing at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling Alloy

Weight change, mg/cm2

Metal loss, mm

Internal oxidation, mm

230 617 263 X

−6.8 −80.1 −219.4 −107.1

0.07 0.17 0.32 0.16

0.09 0.07 0.10 0.07

Extraction residues

M6C (strong) Cr2N (medium) AlN (strong) TiN (medium weak) CrN (medium strong) Cr2N (medium strong)

Internal nitridation, mm

0.17 >0.41(b) 0.29 >0.40(b)

Total attack(a), mm

0.24 >0.58(b) 0.61 >0.56(b)

Note: 1.0 mm=39.4 mils. (a) Metal loss + internal oxidation or internal nitridation (whichever is greater). (b) Internal nitridation through thickness. Source: Ref 36

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 76

76 / High-Temperature Corrosion and Materials Applications

(i.e., 980 °C for 1000 h), Lai (Ref 37) examined the effect of the thermal cycling on the internal nitridation. In this test, no thermal cycling was involved. The results of this test (Ref 37) were then compared with those in the earlier test (Ref 36). The results of the high-velocity dynamic burner rig test at 980 °C (1800 °F) for 1000 h without thermal cycling are summarized in Table 4.9. The results of chemical analysis showing nitrogen content before and after testing are summarized in Table 4.10. In comparing the test results with thermal cycling and those without thermal cycling, thermal cycling signiﬁcantly accelerated oxidation attack by causing oxide spallation, as shown in Fig. 4.14. Thermal cycling was also found to accelerate nitridation attack, as shown in Fig. 4.15. Among the four combustor alloys, alloy 230, however, was least affected by thermal cycling. The test program was extended to include some iron-base alloys under the same test conditions using the same dynamic burner rig (Ref 38). The results showed that iron-base alloys suffered signiﬁcantly more nitridation attack than nickel-base alloys. Figure 4.16(a) shows the microstructure of alloy 556 (Fe-22Cr20Ni-18Co-3Mo-2.5W-0.6Ta-0.2N-La), revealing signiﬁcant internal nitridation attack with formation blocky chromium nitrides after testing

at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. The nitrogen content was found to increase to 1.27% after testing from the original 0.13% before testing. Figure 4.16(b) shows the microstructure of Type 310 stainless steel (SS) after testing at 980 °C (1800 °F) for 1000 h with 30 min thermal cycling, revealing signiﬁcant internal nitridation attack with formation of blocky chromium nitrides. The

Table 4.9 Test results in terms of weight loss, depth of oxidation penetration, depth of nitridation, and total depth of attack after testing at 980 °C (1800 °F) for 1000 h without thermal cycling

Fig. 4.14

Metal loss, mm

Internal oxidation, mm

Internal nitridation, mm

Total attack(a), mm

230 617 263 X

−3.7 −12.1 −19.3 −5.0

0.039 0.031 0.072 0.030

0.065 0.126 0.135 0.086

0.18 >0.55(b) 0.16 0.10

0.22 >0.58(b) 0.23 0.13

Note: 1.0 mm = 39.4 mils. (a) Metal loss +internal oxidation or internal nitridation (whichever is greater). (b) Internal nitridation through thickness. Source: Ref 37

Table 4.10 Results of bulk nitrogen analysis before and after testing at 980 °C (1800 °F) for 1000 h without thermal cycling Alloy

Original nitrogen, wt%

Nitrogen after testing, wt%

230 263 617 X

0.05 0.004 0.03 0.04

0.065 0.097 0.179 0.075

Source: Ref 37

263

Weight change, mg/cm2

–40

–60

–80

–100

–219 Non cyclic test Cyclic test Comparison weight change data between the thermal cycling test (30 min cycles) and no thermal cycle test during the dynamic burner rig testing at 980 °C (1800 °F) for 1000 h for alloys 230, X, 617, and 263. Source: Ref 36, 37

Nitrogen, wt%

Weight change, mg/cm2

617

X

–20

Cyclic test

0.6 0.4 0.2 0

Nitrogen, wt%

Alloy

230

0

230

X

617

263

Non cyclic test

0.4 0.2 0

230

X

617

263

Original After testing

Fig. 4.15

Comparison nitrogen gain data between the thermal cycling test (30 min cycles) and no thermal cycle test during the dynamic burner rig testing at 980 °C (1800 °F) for 1000 h for alloys 230, X, 617, and 263. Source: Ref 36, 37

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 77

Chapter 4: Nitridation / 77

nitrogen content increased from the original 0.03% before testing to 1.69% after testing. Alloy 800H was also shown to suffer severe nitridation attack with chromium nitrides and needle-shaped aluminum nitrides (Fig. 4.16c). The oxide scales formed on these iron-base alloys after testing were porous and nonprotective. At 870 °C (1600 °F), internal nitridation attack was found to be signiﬁcantly reduced.

Figure 4.17 shows the microstructures of alloys 230, 617, and X after testing at 870 °C (1600 °F) for 2000 h with 30 min cycles. Some aluminum nitrides and chromium nitrides were observed in alloy 617, and only some chromium nitrides were observed in alloy X, while no nitrides were observed in alloy 230, as shown in Fig. 4.17

(a)

(a)

20 µm

50 µm

(b)

(b)

(c) (c)

Fig. 4.16

Extensive internal blocky chromium nitrides formed in alloy 556 (a), Type 310 (b), and alloy 800H (c) after the dynamic burner rig testing at 980 °C (1800 °F) for 1000 h with 30 min cycles. Courtesy of Haynes International, Inc.

Fig. 4.17

Alloys 230 (a), 617 (b), and X (c) after the dynamic burner rig testing at 870 °C (1600 °F) for 2000 h with 30 min cycles. Alloy 230 revealed no nitrides, alloy 617 showed both chromium nitrides (blocky phases) and aluminum nitrides (needle phases), and alloy X showed only blocky chromium nitrides. Internal oxides were observed for all three alloys, and all three alloys showed porous and nonprotective external oxide scales. Courtesy of Haynes International, Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 78

78 / High-Temperature Corrosion and Materials Applications

(Ref 38). Oxide scales for three alloys were found to be porous and nonprotective. Iron-base alloys, such as Type 310SS, on the other hand, were found to suffer severe internal nitridation attack, as shown in Fig. 4.18 (Ref 38). The difference in the internal nitridation attack between nickel-base and iron-base alloys is likely to be caused by the differences in nitrogen solubilities in two different alloy systems with much lower nitrogen solubilities in nickel-base alloys. The above test data were generated from Ni-Cr and Fe-Ni-Cr alloys. These alloys are chromia formers (i.e., alloys forming Cr2O3 oxide scale). These chromia formers are susceptible to oxidation/nitridation attack in varying degrees under gas turbine combustion conditions. For very high temperatures and harsh oxidizing combustion conditions, alumina formers (i.e., alloys forming Al2O3 oxide scale) are better performers. Lai (Ref 38) investigated two alumina formers; one was wrought alloy 214 (Ni-16Cr-3Fe-4.5Al-Y) and the other oxidedispersion-strengthened alloy (produced by powder metallurgy) MA956 (Fe-20Cr-4.5Al0.5Y2O3). Due to much more tenacious aluminum oxide scales, these two alloys were tested at 1150 °C (2100 °F) for 200 h with 30 min cycles. The test results showed no nitridation in either alloy. Figure 4.19 shows the cross section of

alloy 214 tested specimen, and Fig. 4.20 shows the cross section of alloy MA956 tested specimen. Both alloys showed ﬁngerlike preferential oxidation penetration. The preferential oxidation penetration was the result of thermal stresses developed from severe thermal cycling from 1150 to less than 260 °C (2100 to <500 °F) every 30 min. No preferential oxidation penetration was observed under the same test condition without thermal cycling. Scanning electron microscopy with energy-dispersive x-ray spectroscopy (SEM/EDX) analysis of the oxide 2 3

1

10 µm

Fig. 4.19

Scanning electron micrograph showing the oxide scale of alloy 214 after testing in the dynamic burner rig at 1150 °C (2100 °F) with 30 min cycle. The results of the energy-dispersive x-ray spectroscopy (EDX) analysis of the oxide scale are summarized: 1, aluminum oxide; 2, aluminum oxide; and 3, Al-rich (Ni,Cr) oxide

2 1 3 4

10 µm 20 µm

Fig. 4.18

Extensive blocky chromium nitrides formed in Type 310SS after testing in the dynamic burner rig testing at 870 °C (1600 °F) for 2000 h with 30 min cycles. Courtesy of Haynes International, Inc.

Fig. 4.20

Scanning electron micrograph showing the oxide scale of alloy MA956 after testing in the dynamic burner rig at 1150 °C (2100 °F) with 30 min cycle. The results of the energy-dispersive x-ray spectroscopy (EDX) analysis of the oxide scale are summarized: 1, Fe-Al-rich oxide; 2–4, aluminum oxide

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 79

Chapter 4: Nitridation / 79

scales showed that both alloys exhibited an aluminum oxide scale. For MA956 some Fe-Al-rich oxide phases were observed to form on the top of the aluminum oxide scale. A model for oxidation/nitridation reactions for chromia formers (Ni-Cr and Fe-Ni-Cr alloys) in oxidizing combustion atmospheres was proposed by Lai (Ref 36). This model, as schematically illustrated in Fig. 4.21, involves the following reaction steps: 1. Chromium oxides form a protective scale initially on the alloy surface in oxidizing combustion atmospheres. 2. Cracks, pores, and other defects develop in the chromium oxide scales after thermal cycling and/or long-term exposure. 3. Chromium oxide scales become porous and nonprotective with nickel oxides forming in Ni-Cr alloys and iron oxides forming in ironbase alloys. 4. Both O2 and N2 molecules from the combustion gas stream permeate through the oxide scales and reach the metal underneath. 5. Oxidation of the metal surface results in lower oxygen potential with a concurrent increase in nitrogen potential. 6. Nitridation then occurs following the reaction: 1=2N2 (gas) $ N (solution); the concentration of nitrogen absorbed in the metal is then proportional to the nitrogen potential ( pN2 ) by: [%N]=k( pN2 )1=2

where k is an equilibrium constant. 7. Nitrogen dissolves into the alloy and diffuses into the metal interior. Increasing nitrogen concentration in the alloy with

increasing exposure times eventually leads to the formation of nitrides in the alloy once the solubility limits for CrN, Cr2N, AlN, and/or TiN are exceeded.

4.4 Nitridation in NH3-H2O Environments A NH3-H2O mixture has been considered in the Kalina cycle for power generation (Ref 39). Very little published data are available for the corrosion behavior of engineering alloys in this type of environment. Grabke et al. (Ref 40) investigated the corrosion behavior of a number of commercial alloys in the NH3-30%H2O gas mixture at 500 °C (930 °F). Several interesting results were obtained from this investigation. One of the most interesting observations was that signiﬁcant nitridation attack and severe intergranular cracking were observed in alloy 600 (Ni-16Cr-8Fe) after only 200 h of exposure, as shown in Fig. 4.22. Alloy 600 has been known to be one of the most nitridation-resistant alloys in ammonia environments, and the alloy has been widely used in ammonia plants (data are presented in Section 4.5). Alloy 800 (Fe-22Cr-32Ni-Al-Ti) was also found to suffer severe nitridation attack and intergranular cracking (Fig. 4.23). For the ferritic stainless steel, Fe-18Cr (Sicromal), a combination of severe intergranular nitridation, cracking, and oxidation attack caused rapid disintegration of the alloy in 200 h (Fig. 4.24). An austenitic stainless steel (Fe-18Cr-9Ni) was also found to suffer severe nitridation attack. However, no

2 CrN Cr2N

2

Log pN , atm

–2

Test environment

–6 AIN

Cr2O3

–10 Cr

–14 –18

AI

–22 –55

–45

AI2O3 50 µm –35

–25

–15

–5

5

Log pO , atm 2

Fig. 4.21

Schematic showing a model for internal nitridation attack in high-temperature alloys in a simulated combustion environment. Source: Ref 36

Fig. 4.22

Optical micrograph showing the cross section of alloy 600 (Ni-16Cr-8Fe) after exposure to the NH3-30%H2O gas mixture at 500 °C (930 °F) for only 200 h, revealing severe nitridation attack and intergranular cracking. Source: Ref 40

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 80

80 / High-Temperature Corrosion and Materials Applications

cracking was observed in the austenitic stainless steel. They found formation of very ﬁne CrN precipitates in these alloys. The authors (Ref 40) pointed out that in these tests (low pressure) NH3 should decompose largely to N2 and H2, while in the Kalina process (under high pressures) decomposition of NH3 is considered to be negligible. It is believed that at the test temperature of 500 °C, a signiﬁcant amount of NH3 is believed to be retained without decomposition. Robo (Ref 4) found that about 60% NH3 was decomposed, leaving about 40% NH3 at the exhaust end at 525 °C (980 °F) in his laboratory testing using 100% NH3 in the inlet gas. Barnes and Lai (Ref 41) found a larger decomposition of NH3 (100% NH3 in the inlet test gas and 30% NH3 in the exhaust) when tested at a slightly higher temperature, 650 °C. The corrosion behavior of alloys in the NH3-H2O mixture as observed by Grabke et al. (Ref 40) appears to be different from that of the NH3-H2 mixture, as is discussed in the next section.

(a)

4.5 Nitridation in NH3 and H2-N2-NH3 Environments Ammonia (NH3) is thermally unstable. It can readily dissociate into N2 and H2 at elevated temperatures. There appears to be no published data available on the dissociation rate as functions of temperature and pressure. However, several laboratory measurements have indicated that dissociation rates were extremely fast at high temperatures. Table 4.11 shows some dissociation data generated in laboratory test furnaces by measuring the amount of NH3 at the exhaust end of the tube furnace when 100% NH3 was entered into the furnace tube. Both sets of data were measured in laboratory setups with pressures being close to 1 atm. In order to avoid catalyst reactions by metals, the measurement of ammonia dissociation at 650, 980, and 1090 °C was made with no metallic samples in the furnace tube, which was composed of high-purity alumina (Ref 38). No published high-pressure dissociation values are available. Thus, for laboratory test data even with 100% NH3 as the

200 µm

30 µm

Fig. 4.24

Optical micrograph showing the cross section of an Fe-18Cr (Sicromal alloy) tested specimen that suffered combination of severe nitridation, oxidation, and intergranular cracking after the exposure to the NH3-30%H2O gas mixture at 500 °C (930 °F) for 200 h. Source: Ref 40

50 µm (b)

Fig. 4.23 Severe nitridation attack and intergranular cracking in alloy 800 (Fe-22Cr-32Ni-Al-Ti) after exposure to the NH3-30%H2O gas mixture at 500 °C (930 °F) for 200 h. (a) Surface appearance of the tested specimen showing intergranular cracks. (b) Cross section of the tested specimen. Source: Ref 40

Table 4.11 NH3 content measured at the exhaust end of the tube furnace at different temperatures when 100% NH3 was injected NH3 at inlet, vol%

Temperature, °C (°F)

100

525 650 980 1090

(a) Ref 4. (b) Ref 41

°C (980 °F) °C (1200 °F) °C (1800 °F) °C (2000 °F)

NH3 at exhaust, vol%

40(a) 30(b) <5(b) <5(b)

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 81

Chapter 4: Nitridation / 81

inlet gas, the corrosion reactions generally involve both N2 and NH3. At very high temperatures, such as 980° and 1090 °C (1800 and 2000 °F), nitridation is most likely involved in the reaction with N2 because of rapid dissociation of NH3. Verma et al. (Ref 42) reported that an ammonia cracker unit, used to produce nitrogen and hydrogen, failed after 1000 h of operation. The preheater tubes (operating at 350 to 400 °C, or 660 to 750 °F) were made of Type 304SS, while the furnace tubes (operating at about 600 °C, or 1110 °F) were made of Type 310SS. Both suffered severe nitridation attack. To select an alternate alloy, nitriding tests were performed on various alloy samples at 600 °C (1110 °F) in an environment consisting of 6 to 8% NH3, 75.77 to 77.5 wt% N2, and 16.25 to 16.5 wt% H2. Test results are summarized in Table 4.12. The alloys that performed well include Types 347, 316, 321, SLX-254, and HV-9A. Type 347 was the best performer, having a linearly extrapolated penetration rate of about 0.13 mm/yr (5 mpy). Alloy 800, which contains more nickel than any of the above stainless steels, did not perform as well. Furthermore, Type 304 was found to suffer attack two orders of magnitudes higher than that of Type 316L. The results also showed that titanium suffered severe nitridation attack, which resulted in severe sample cracking. Both carbon steel and 1Cr-0.5Mo steel suffered decarburization after only 50 h. Ammonia (NH3) is produced by synthesis from hydrogen and nitrogen at high pressures and elevated temperatures. The “heart” of the process is the ammonia “converter,” where hydrogen and nitrogen combine. Signiﬁcant corrosion issues are associated with the converter and the internal components inside the

converter. The converters operate at high pressures (130–350 atm or 800–1000 atm) and temperatures up to 650 °C (1200 °F) (Ref 43). Cihal (Ref 43) discussed the major corrosion problems—hydrogen attack and nitridation—for the ammonia converter. The converter usually consists of a vessel with a catalyst basket and an interchanger inside the vessel. Because of high-pressure, high-temperature hydrogen in the converter, early converters were constructed out of a thick-wall steel vessel with an inner carbon steel lining and vent holes through the vessel wall. Thus, the inner carbon steel lining was the only part suffering hydrogen attack, while the main thick-wall vessel was unaffected by highpressure, high-temperature hydrogen (Ref 43). Hydrogen attack is the damage of steel by the reaction of hydrogen with cementite (Fe3C) in steel to form methane gas (CH4), resulting in formation of microcracks and ﬁssures as well as decarburization in steel. (Hydrogen attack is reviewed and discussed in Chapter 17.) Later designs of the converter allowed the cold inlet gas ﬂowing along the vessel wall to keep the vessel cold, thus eliminating the potential hydrogen attack problem for the vessel (Ref 43). Cihal (Ref 43) indicated that the internal components made of carbon steels exhibited a short life due to hydrogen attack. Alloy steels containing chromium were more resistant to hydrogen attack, but had suffered severe embrittlement problems due to nitridation attack. Figure 4.25 shows intergranular cracking in the nitrided layer of an alloy steel (0.12C5.6Cr-0.42Mo) after exposure to the synthesis gas inside the converter at 325 atm and 450 to 500 °C (840 to 930 °F) for 4380 h (Ref 43). An alloy steel containing a strong nitride former such as titanium, such as alloy steel with

Table 4.12 Nitridation attack of various alloys in an ammonia-bearing environment at 600 °C (1110 °F) for indicated exposure times Penetration depth of nitrtdiltion attack, mm (mils) Alloy

Carbon steel 1Cr-0.5Mo steel Titanium 304 316L 329 310 321 347 SLX-254(b) HV-9A(c) 800

50 h

100 h

300 h

600 h

1000 h

1500 h

Decarb. Decarb. 0.0066 (0.3) … 0.02 (0.8) … … 0.013 (0.5) … 0.013 (0.5) 0.01 (0.4) 0.02 (0.8)

Decarb. Decarb. 0.0133 (0.5) 0.013 (0.5) 0.02 (0.8) 0.066 (2.6) 0.03 (1.2) 0.013 (0.5) 0.013 (0.5) 0.013 (0.5) 0.10 (3.9) 0.10 (3.9)

Decarb. 0.033 (1.3) 0.233 (9.2) 0.013 (0.5) 0.02 (0.8) 0.10 (3.9) 0.13 (5.1) 0.013 (0.5) 0.013 (0.5) 0.026 (1.0) 0.10 (3.9) 0.20 (7.9)

Decarb. 0.033 (1.3) 0.266 (10.5) 0.03 (1.2) 0.03 (1.2) 0.10 (3.9) 0.16 (6.3) 0.016 (0.6) 0.013 (0.5) 0.026 (1.0) 0.10 (3.9) …

Decarb. 0.033 (1.3) Cracked(a) 0.06 (2.4) 0.04 (1.6) 0.20 (7.9) 0.33 (13.0) 0.06 (2.4) 0.02 (0.8) 0.06 (2,4) 0.10 (3.9) 0.20 (7.9)

Decarb. 0.3 (11.8) Cracked(a) 4.2 (165) 0.04 (1.6) 0.40 (15.7) 0.40 (15.7) 0.06 (2.4) 0.02 (0.8) 0.06 (2.4) 0.10 (3.9) 0.20 (7.9)

Decarb.: decarburized. (a) Nitridation through thickness. (b) SLX-254: Fe-19.7Cr-24.5Ni-4.35Mo-1.43Cu. (c) HV-9A: Fe-21.2Cr-24.6Ni-3.8Mo-1.5Cu. Source: Ref 42

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 82

82 / High-Temperature Corrosion and Materials Applications

0.05% C, 2.9% W, and 0.54% Ti, exposed to the same converter environment under the same test conditions as the 0.12C-5.6Cr-0.42Mo steel (as shown in Fig. 4.25) was found to show no cracking. Nitridation resistance of various alloys was studied by Moran et al. (Ref 44) in an ammonia converter and preheater line. The results are summarized in Table 4.13. Corrosion rates were found to depend strongly on the concentration of ammonia. Type 304, for example, suffered corrosion rates that increased from 0.02 to 2.5 mm/yr (0.6 to 99 mpy) as the concentration of NH3 was increased from 5 to 6% (in the ammonia converter) to 99% (in the ammonia preheater line) at about 500 °C (930 °F). In an ammonia converter with about 5 to 6% NH3 and 490 to 550 °C (910 to 1020 °F), all stainless steels tested (i.e., 430, 446, 302B, 304, 316, 321, 309, 314, 310, and 330) showed negligible nitridation attack, with corrosion rates of about 0.03 mm/yr (1 mpy) or less. For the plant ammonia line (preheater exit), which was exposed to 99% NH3, stainless steels, such as 446, 304, 316, and 309, suffered severe nitridation attack, with corrosion rates of about 2.54 mm/yr (100 mpy) or more. Moran et al. (Ref 44) found that Type 316 suffered signiﬁcantly more attack than Type 304, contrary to the observations of Verma et al. (Ref 42). Robo (Ref 4) reported the performance of several alloys in a Topsoe-type ammonia converter. Most of the components made of Type 304, exposed to temperatures up to 500 °C (930 °F) with ammonia concentration up to 20%, exhibited negligible nitridation rates (0.4 to

4 mpy). One Type 304 sample showed a slightly higher corrosion rate (10 mpy), presumably due to a higher temperature. Alloy 600 (Ni-Cr-Fe alloy) had signiﬁcantly better nitridation resistance than stainless steels, with corrosion rates 1 or 2 orders of magnitude lower. The results are summarized in Table 4.14 (Ref 4). McDowell (Ref 45) reported ﬁeld test results performed in a Casale converter (540 °C, or 1000 °F, and 11 ksi) for 1 and 3 years. These results are summarized in Table 4.15. AISI 502 (5Cr steel) was extremely susceptible to nitridation attack, with more than 2.54 mm (0.1 in., or 100 mils) of nitridation depth in a year. Results showed a general trend of increased resistance to nitridation as nickel content in the alloy increased. One striking observation was that after 3 years of exposure, the alloys showed essentially similar depths of nitridation attack as they did after 1 year. Robo (Ref 4) reported kinetic data for Type 304 and alloys 600 and 625 in laboratory tests performed with pure ammonia as the inlet test gas. The nitridation for Type 304 was found to follow a linear rate law at 525 °C (980 °F) for up to 1000 h. The maximum thickness of the nitride layer (in the form of scale) measured metallographically is shown in Fig. 4.26 as a function of time. A growth rate of about 0.37 µm/h was observed. This corresponds to about 3240 µm/yr (128 mpy). This rate is signiﬁcantly higher than those observed in ammonia converters. The ammonia concentration in this test (reportedly, 40% ammonia was dissociated Table 4.13 Corrosion behavior of various alloys in an ammonia converter and plant ammonia line Corrosion rate. mm/yr (mpy) Alloy

430 446 302B 304 316 321 309 314 310 330 (0.47Si) 330 (1.00Si) 600 80Ni-20Cr Ni

Fig. 4.25

Intergranular cracking in the nitrided layer of an alloy steel (0.12C-5.6Cr-0.42Mo) after exposure to the synthesis gas inside the converter at 325 atm and 450 to 500 °C (840 to 930 °F) for 4380 h. Source: Ref 43

Ammonia (converter)(a)

Plant ammonia line(b)

0.022 (0.90) 0.028 (1.12) 0.019 (0.73) 0.015 (0.59) 0.012 (0.47) 0.012 (0.47) 0.006 (0.23) 0.003 (0.10) 0.004 (0.14) 0.002 (0.06) 0.001 (0.02)

… 4.18 (164.5) … 2.53 (99.5) >13.21 (520) … 2.41 (95) … … … 0.43 (17.1) 0.16 (6.3) 0.19 (7.4) 2.01 (79.0)

(a) 5 to 6% NH3, 29164 h at 490 to 550 °C (910 to 1020 °F), and 354 atm (5200 psi) ( Haber-Bosch converter). (b) 99. 1% NH3, 1540 h at 500 °C (930 °F). Source: Ref 44

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 83

Chapter 4: Nitridation / 83

Table 4.14 Converter No.

1 2 3

4

Corrosion behavior of various alloys in an ammonia converter Component

Material

Gas temperature, °C

Lining Plate, 2nd bed Bolt Wire mesh, 2nd bed Perforated plate. 1st bed Inner shell, 2nd bed Perforated plate. 2nd bed Center tube, 2nd bed Nut, bottom Bolt, bottom Wire mesh ThermoweII

304 304 302 Alloy 600 304 304 304 304 304 403 Alloy 600 304

525 475 … 520 500 440 440 485 480 480 c.500 c.500

NH3

15–20 15–20 … … 13 8–10 8–10 16 16 16 3.5

Time of operation, yr

Thickness of nitride, μm (mils)

4 7 7 7 5 5 5 5 5 5 4 8

1000 (39.4) 100 (3.9) 375 (14.8) 8 (0.3) 270 (10.6) 45 (1.8) 60 (2.4) 440 (17.3) 260 (10.2) 540 (21.3) 6 (0.2) 200 (7.9)

Average nitriding, µm/yr (mpy)

250 (9.8) 14 (0.6) 54 (2.1) 1 (0.04) 54 (2.1) 9 (0.4) 12 (0.5) 88 (3.5) 52 (2.0) 108 (4.3) 1.5 (0.06) 25 (1.0)

Note: Topsoe-type ammonia converter operated at 22 MPa (3.2 ksi). Source: Ref 4

Table 4.15 Depth of nitridation for various alloys after 1 and 3 years in a Casale ammonia converter

500

Nitridation depth, mm (mils)

502 (5Cr steel) 446 304 316 321 347 309 310 800 804 (30Cr-42Ni) 600 Nickel 200

1 yr

3 yr

2. 88 (113.2) 1.06 (41.7) 1.08 (42.7) 0.46 (18. 2) 0.46 (18.3) 0.49 (19.2) 0.24 (9.5) 0.22 (8.8) 0.14 (5.4) 0.03 (1.2) 0.16 (6.4) None

Completely nitrided 1.15 (45.3) 1.12 (44.0) 0.48 (18.7) 0.60 (23.6) 0.45 (17.6) 0.24 (9.6) 0.23 (9.2) 0.13 (5.3) 0.03 (1.2) 0.16 (6.4) None

Note: Operated at 538 °C (1000 °F) and 76 MPa (11 ksi). Source: Ref 45

in the test furnace) was signiﬁcantly higher than in the ammonia converters. The rate of about 3.25 mm/yr (128 mpy) was, however, of the same order of magnitude as that observed by Moran et al. (Ref 44) in the plant ammonia line (about 100 mpy for Type 304). At 700 °C (1290 °F), Robo (Ref 4) found that the reaction rates for Type 304, alloy 600, and alloy 625 can be described by: X =kt n

2[N] Dt c[m]

300

200

100

0

0

500

1000

Exposure time, h

Fig. 4.26

Nitriding depth of Type 304SS in ammonia (100% in the inlet gas and 60% in the exhaust) at 525 °C (980 °F) as a function of exposure time. Source: Ref 4

ð4:5Þ

where X is thickness of the nitrided layer in µm, k is reaction constant, t is time in hours, n is 0.66 for Type 304 and 0.26 for alloys 600 and 625. Jack (Ref 46) developed a kinetic model based on the models for internal oxidation, to describe the growth of internal penetration (X) in the absence of iron nitride formation: X 2=

400 Thickness of nitride layer, µm

Alloy

ð4:6Þ

where [N] is the surface nitrogen concentration (at.%), [m] is the alloy element concentration

(at.%), γ is the ratio of nitrogen to alloy element in the nitride phase, D is diffusivity of nitrogen, and t is time. This model predicts that fast nitriding rates can be achieved by increasing the ammonia content in the gas mixture and thus the surface nitrogen concentration. Another important factor in nitriding kinetics is the concentration of the alloy element that forms internal nitrides. The model predicts that nitriding depth is inversely proportional to the concentration of the nitrideforming alloy element.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 84

84 / High-Temperature Corrosion and Materials Applications

The nitridation behavior of a wide variety of commercial alloys in ammonia was extensively investigated by Barnes and Lai (Ref 41). The alloys tested included stainless steels, Fe-Ni-Cr alloys, and nickel- and cobalt-base superalloys. Many alloys tested contained alloying elements (e.g., Al, Ti, Zr, Nb, Cr, Mo, W, Fe, etc.) that form nitrides. The test results generated at 650, 980, and 1090 °C (1200, 1800, and 2000 °F) are summarized in Tables 4.16 to 4.18. It was found that nickel-base alloys are generally more nitridation resistant than iron-base alloys. Increasing nickel content generally improves the resistance of the alloy to nitridation attack. Increasing cobalt content appears to have the same effect. When nitrogen absorption was plotted against Ni + Co content in the alloy for the 650 °C (1200 °F) test data, resistance to nitridation improves with increasing Ni + Co content up to about 50 wt%, as shown in Fig. 4.27. Further increases up to about 75% did not seem to affect the nitridation resistance of the alloy. For maximum resistance to nitridation attack at 650 °C (1200 °F), it appears that alloys with at least 50% Ni or Ni + Co are most suitable. This is in general agreement with the results reported by Moran et al. (Ref 44). Their results suggested that improvement in nitridation resistance began to level off at about 40% Ni, with no improvement resulting from further increases in nickel up to about 80%. Pure nickel, however, showed signiﬁcantly lower nitridation resistance (Ref 43). At 980 °C (1800 °F), a slightly different relationship was observed, as shown in

Table 4.17 Nitridation resistance of various alloys in ammonia at 980 °C (1800 °F) for 168 h Alloy

214 600 S 601 230 617 HR-160 188 625 6B 253MA 25 X RA333 RA330 800H 825 150 MULTIMET 316 556 304 310 446

Alloy

C-276 230 HR-160 600 625 RA333 601 188 S 617 214 X 825 800H 556 316 310 304

Alloy base

Nitrogen absorption, mg/cm2

Depth of nitride penetration, mm (mils)

Nickel Nickel Nickel Nickel Nickel Nickel Nickel Cobalt Nickel Nickel Nickel Nickel Nickel Iron Iron Iron Iron Iron

0.7 0.7 0.8 0.8 0.9 1.0 1.1 1.2 1.3 1.3 1.5 1.7 2.5 4.3 4.9 6.9 7.4 9.8

0.02 (0.6) 0.03 (1.2) 0.01 (0.5) 0.03(1.3) 0.01 (0.5) 0.03 (1.0) 0.03 (1.0) 0.02 (0.6) 0.03 (1.1) 0.03 (1.0) 0.04 (1.5) 0.04 (1.5) 0.06 (2.2) 0.10 (4.1) 0.09 (3.5) 0.19 (7.3) 0.15 (6.0) 0.21 (8.4)

Note: 100% NH3 in the inlet gas and 30% NH3 in the exhaust gas. Source: Ref 41

Depth of nitride penetration, mm (mils)

Nickel Nickel Nickel Nickel Nickel Nickel Nickel Cobalt Nickel Cobalt Iron Cobalt Nickel Nickel Iron Iron Nickel Cobalt Iron Iron Iron Iron Iron Iron

0.3 0.9 0.9 1.2 1.4 1.5 1.7 2.3 2.5 3.1 3.3 3.6 3.2 3.7 3.9 4.0 4.3 5.3 5.6 6.0 6.7 7.3 7.7 12.9

0.04 (1.4) 0.12 (4.8) 0.18 (7.2) 0.17 (6.6) 0.12(4.9) 0.38 (15.0) 0.18 (7.2) 0.19(7.4) 0.17 (6.9) 0.15 (5.8) 0.48 (19.0) 0.26(10.4) 0.19 (7.4) 0.42(16.4) 0.52 (20.6) 0.28 (11.1) 0.58 (23.0) 0.38 (15.1) 0.35 (13.6) 0.52 (20.3) 0.37 (14.7) >0.58 (23.0) 0.38 (15.1) >0.58 (23.0)

Note: 100% NH3 in the inlet gas and less than 5% NH3 (detection limit) in the exhaust gas. Source: Ref 41

Table 4.18 Nitridation resistance of various alloys in ammonia at 1090 °C (2000 °F) for 168 h Alloy

Table 4.16 Nitridation resistance of various alloys in ammonia at 650 °C (1200 °F) for 168 h

Alloy base

Nitrogen absorption mg/cm2

600 214 S 230 25 617 188 HR-160 601 RA330 625 316 304 X 150 556 446 6B MULTIMET 825 RA333 800H 253MA 310

Alloy base

Nitrogen absorption, mg/cm2

Depth of nitride penetration, mm (mils)

Nickel Nickel Nickel Nickel Cobalt Nickel Cobalt Nickel Nickel Iron Nickel Iron Iron Nickel Cobalt Iron Iron Cobalt Iron Nickel Nickel Iron Iron Iron

0.2 0.2 1.0 1.5 1.7 1.9 2.0 2.5 2.6 3.1 3.3 3.3 3.5 3.8 4.1 4.2 4.5 4.7 5.0 5.2 5.2 5.5 6.3 9.5

0 0.02 (0.7) 0.34 (13.4) 0.39 (15.3) >0.65 (25.5) >0.56 (22) >0.53 (21) 0.46 (18) >0.58 (23) >0.56 (22) >0.56 (22) >0.91 (36) >0.58 (23) >0.58 (23) 0.51 (20) >0.51 (20) >0.58 (23) >0.64 (25) >0.64 (25) 0.58 (23) >0.71 (28) >0.76 (30) >1.5 (60) >0.79 (31)

Note: 100% NH3 in the inlet gas and less than 5% (detection limit) in the exhaust gas. Source: Ref 41

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 85

Chapter 4: Nitridation / 85

15

8

Nitrogen absorption, mg/cm2

Nitrogen absorption, mg/cm2

10

6

4

2

0

0

20

40

60

10

5

0

80

0

20

40

60

80

Ni + Co, wt%

Ni + Co, wt%

Fig. 4.28

Fig. 4.27

Effect of the Ni + Co content in iron-, nickel-, and cobalt-base alloys on nitridation resistance at 650 °C (1200 °F) for 168 h in ammonia (100% NH3 in the inlet gas and 30% NH3 in the exhaust). Source: Ref 41

Effect of the Ni + Co content in iron-, nickel-, and cobalt-base alloys on nitridation resistance at 980 °C (1800 °F) for 168 h in ammonia (100% NH3 in the inlet gas and <5% NH3 in the exhaust). Source: Ref 41

Fig. 4.28. Nitrogen absorption was reduced drastically with an initial 15% Ni (or Ni + Co). As Ni + Co content increased from 15 to 50%, no drastic improvement in nitridation resistance was noted. Further increases in Ni + Co content in excess of about 50% caused a sharp improvement. Alloys with Ni (or Ni + Co) in excess of about 60% showed the most resistance to nitridation. In this test program (Ref 41), no alloys with 80% Ni (or Ni + Co) and higher were tested. It is generally believed that the beneﬁcial effect of nickel or cobalt in increasing nitridation resistance is caused by the reduced solubility of nitrogen in the alloy. Nickel and cobalt were found to reduce the solubility of nitrogen in iron (Ref 15, 47). Morphology of nitrides formed in alloys as a result of exposure to NH3 is widely different between low- and high-temperature exposures. Nitridation at low temperatures (e.g., 650 °C, or 1200 °F) generally results in a surface nitride layer. For iron-base alloys, the surface nitride layer consists of mostly iron nitrides (Fe2N or Fe4N), while for nickel- and cobalt-base alloys, the nitride layer consists of mainly CrN. Hightemperature exposures, on the other hand, result in formation of internal nitrides, which are mostly CrN, Cr2N, (Fe,Cr)2N, AlN, and TiN. Table 4.19 summarizes the results of the x-ray diffraction analysis performed on the surfaces of the selected test specimens tested at different temperatures. Figure 4.29 illustrates the morphology of the surface nitride layer formed at

Table 4.19 Phases detected from the x-ray diffraction analysis performed on the surfaces of test specimens after exposure to NH3 at temperatures indicated for 168 h Phases Alloy

304SS 800H 556 230 188

650 °C (1200 °F)

980 °C (1800 °F)

1090 °C (2000 °F)

Fe2N (Fe3Ni)N (Fe3Ni)N CrN CrN

CrN CrN CrN CrN CrN

(Cr,Fe)2N1-x (Cr,Fe)2N1-x (Cr,Fe)2N1-x (Cr,Mo)12(Fe, Ni)8-xN4-z Cr2N, CrN

Note: At 650 °C (1200 °F), 100% NH3 in the inlet gas and 30% NH3 in the exhaust. At 980 and 1090 °C (1800 and 2000 °F), 100% NH3 in the inlet gas and <5% NH3 in the exhaust. Source: Ref 41

650 °C (1200 °F) for iron-base alloys, Type 446SS (Fe-25Cr) and Type 304SS (Fe-18Cr8Ni), and nickel-base alloys, alloys 600 (Ni16Cr-8Fe), 625 (Ni-22Cr-9Mo-3.5Nb-3Fe), X (Ni-22Cr-18.5Fe-9Mo-0.6W), and C-276 (Ni16Cr-5Fe-16Mo-4W). Type 446SS containing about 25% Cr with no nickel exhibited a thick nitride layer (about 0.72 mm, or 28 mils, thick) formed on the alloy surface, as shown in Fig. 4.29(a). With about 8% Ni in the alloy, Type 304SS showed a signiﬁcantly thinner nitride layer (about 0.2 mm, or 0.008 in., 8 mils thick) formed on the alloy surface. For four nickelbase alloys, alloys 600, 625, X, and C-276, an extremely thin nitride layer (about 1.2 to 2.2 µm thick) was found to form on the alloy surface. The iron content in these four nickel-base alloys varies from about 3 to 19%.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 86

86 / High-Temperature Corrosion and Materials Applications

When exposed to high temperatures, the alloy forms internal nitrides due to increased diffusivity of nitrogen. Figure 4.30 shows internal nitrides formed in Type 446SS, Type 304SS, alloy 800H (Fe-21Cr-32Ni-0.4Al-0.4Ti), and alloy 188 (Co-22Cr-22Ni-14W-La). For Type 304SS, alloy 800H, and alloy 188, nitrides appeared to be etched differently with the nitrides formed near the surface etched darker than those in the interior. The nitrides (etched darker) that formed in the surface zone are believed to be CrN. The x-ray diffraction analysis of the test specimen surface showed CrN for these three

alloys (Table 4.19). As nitrogen diffuses farther into metal interior, nitrogen activities become lower, thus forming Cr2N. Thus, the nitrides (etched lighter) formed in the metal interior are believed to be Cr2N. Ni-Cr alloys containing aluminum or titanium (or both) form internal nitrides of not only chromium but also aluminum or titanium (or both) at high temperatures. Both aluminum and titanium are stronger nitride formers, and AlN and TiN can form and penetrate farther into the metal interior than CrN and Cr2N. When alloys contain relatively low aluminum (e.g., about

(a)

200 µm

(d)

(b)

200 µm

(e)

(c)

Fig. 4.29

200 µm

(f)

200 µm

200 µm

200 µm

Optical micrographs showing typical nitride morphology of a surface nitride layer that formed on the alloy surface when exposed to NH3 (100% NH3 in the inlet gas and 30% NH3 in the exhaust) for 168 h at 650 °C (1200 °F) for (a) Type 446, (b) Type 304, (c) alloy 600, (d) alloy 625, (e) alloy X, and (f) alloy C-276. Courtesy of Haynes International, Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 87

Chapter 4: Nitridation / 87

1%), such as alloys 601 and 617, a signiﬁcant amount of internal aluminum nitrides formed in the alloy, as shown in Fig. 4.31(a). Also observed in alloy 601 (Fig. 4.31a) are chromium nitrides (blocky-type phases) that formed near the alloy surface. For alloys containing high concentrations of aluminum (e.g., 4.5% Al in alloy 214), aluminum nitrides formed on the alloy surface, as shown in Fig. 4.31(b). Al2O3 oxide is also believed to form on the alloy 214 surface.

4.6 Nitridation in N2 Atmosphere Metals and alloys are also susceptible to nitridation attack in N2 or N2-H2 environments, particularly at high temperatures. The N2 or N2-H2 atmosphere is commonly used as a protective atmosphere in heat treating and sintering operations. Figure 4.32 shows extensive nitridation attack of Type 314 wire mesh belt in a sintering furnace after 2 to 3 months of

(a)

50 µm

(c)

(b)

50 µm

(d)

Fig. 4.30

50 µm

50 µm

Optical micrographs showing typical nitride morphology in form of internal nitrides penetrating into the metal interior when exposed to NH3 for 168 h at 980 °C (1800 °F) for (a) Type 446, (b) Type 304, (c) alloy 800H, and (d) alloy 188. Courtesy of Haynes International, Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

88 / High-Temperature Corrosion and Materials Applications

service at 1120 °C (2050 °F) in the N2-10%H2 atmosphere. Type 314SS (Fe-26Cr-20Ni-2Si) is commonly used for wire mesh furnace belts. Smith and Bucklin (Ref 48) investigated nitridation reactions in 100% N2 for several iron- and nickel-base alloys. Their results, generated at 980, 1090, and 1200 °C (1800, 2000, and 2200 °F), are tabulated in Table 4.20. As shown in the table, the nitridation kinetics in 100% N2 is extremely rapid. Even nickel-base alloys were found to suffer severe nitridation attack even when the temperature was reduced to 980 °C (1800 °F). Both AlN and Cr2N were found in alloys 600 and 800 after exposure to 100% N2 at 1200 °C (2200 °F) for 100 h. For RA330, only Cr2N was detected after exposure to the same test conditions. Ganesan and Smith (Ref 49) identiﬁed the nitride phases formed near the surface of the test specimens after exposure at 980 °C (1800 °F) for 1008 h in pure nitrogen atmosphere using x-ray diffraction. The major phases identiﬁed are summarized in Table 4.21 (Ref 49).

200 µm (a)

200 µm (b)

Fig. 4.31

(a) Extensive internal aluminum nitride (long needle phase) formation in alloy 601 and (b) insignificant AlN formation in alloy 214 after exposure to NH3 at 1090 °C (2000 °F) for 168 h

For alloys containing 20% or more chromium, Cr2N-type nitrides were found in the region near the surface, except alloy 600, which contains only about 16% Cr. This is in agreement with the phase stability diagram at 1000 °C in terms of pN2 versus Cr content in Ni-Cr alloys as shown in Fig. 4.4, which shows CrN is the most likely nitride in Ni-16Cr alloy (Ref 12). Barnes and Lai (Ref 50) conducted an extensive nitridation study in pure nitrogen atmosphere for iron-, nickel-, and cobalt-base alloys at 1090 °C (2000 °F) for 168 h. Test results in terms of nitrogen absorption (mg/cm2) and the depth of nitridation are summarized in Table 4.22. As a result of rapid nitridation kinetics under the test condition, nitridation attack penetrated through the thickness of the test specimen for many alloys. Due to different thicknesses for different alloys, the ranking of alloy performance in terms of nitridation depths became difﬁcult for most alloys tested. (The thickness of the test specimen varied from alloy to alloy, because of the use of whatever sheet products were available for preparation of test specimens.) Iron-base alloys, the last group from RA330 to Type 310SS, suffered the worst nitridation attack. Two cobalt-base alloys, alloys 188 (Co-22Cr-22Ni-14W-La) and 150 (Co-27Cr18Fe), exhibited poor resistance, with alloy 150 (high Cr and no Ni) showing extremely poor nitridation resistance similar to iron-base alloys. The nitride phases formed in alloys were analyzed using x-ray diffraction performed on the chemical extraction residues obtained from the test specimens. Selected alloys (six nickelbase alloys, two cobalt-base alloys, and one iron-base alloy) were analyzed, and the x-ray diffraction analysis results are summarized in Table 4.23. All the alloys except alloy 214 exhibited Cr2N nitrides. No internal nitrides were observed in alloy 214, which contains 4.5% Al. The alloy 214 specimen showed only surface Al2O3 and AlN phases, as analyzed by x-ray diffraction analysis performed on the surface scales of the specimen. For nickel-base alloys containing low levels of aluminum, such as alloys 601 and 617 (both contain about 1.3% Al), extensive AlN nitrides formed in metal interior. Figure 4.33 shows a through-thickness nitrided alloy 617 specimen, exhibiting extensive needle-shaped internal AlN nitrides along with Cr2N nitrides. Figure 4.34 shows needle-shaped internal AlN nitrides as well as Cr2N nitrides at high magniﬁcation in alloy 601. The addition of 4.5% Al to a nickel-base alloy can provide

pg 88

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 89

Chapter 4: Nitridation / 89

a very effective protection against nitridation attack at high temperatures. The nitride formed on alloy 214 surface after 168 h was too thin to be identiﬁed. However, after 500 h of exposure, the surface scale was found to be composed of AlN and Al2O3, with AlN predominating (Ref 50). The nitrogen absorbed after 168 h

was 0.3 mg/cm2, and no further increase in nitrogen absorption was observed after 504 h. Comparing alloy 214 with another nickel-base alloy containing little aluminum, such as alloy 230, in 100% N2 at 1090 °C (2000 °F) for up to 500 h of exposure clearly showed the superior resistance of alloy 214 against nitridation attack

1.3 mm (50 mils) 100 µm

Fig. 4.32

Optical micrograph showing extensive internal chromium nitrides that formed in the entire cross section of a wire sample obtained from a Type 314 wire mesh belt in a sintering furnace after service for 2 to 3 months at 1120 °C (2050 °F) in N2-10% H2. Courtesy of Haynes International, Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 90

90 / High-Temperature Corrosion and Materials Applications

Table 4.20 Alloy

600 601 800 520 330 DS 314SS

Nitridation resistance of iron- and nickel-base alloys in pure nitrogen 982 °C (1800 °F)/1008 h(a) nitrided depth, mm (mils)

1093 °C (2000 °F)/900 h(b) nitrided depth, mm (mils)

1204 °C (2200 °F)/100 h(c) nitrided depth, mm (mils)

1.30 (51) 1.55 (61) 1.85 (73) … 2.57 (101) … >3.81 (150)

1.85 (73) 2.79 (110) >3.81 (150) >3.81 (150) >3.81 (150) >3.81 (150) >3.81 (150)

2.16 (85) >3.81 (150) >3.81 (150) … >3.81 (150) … …

(a) Specimens were cycled to room temperature once every 24 h for the ﬁrst 3 days and then weekly for the remainder of the test. (b) Specimens were cycled to room temperature once every 96 h (4 days). (c) Isothermal exposure. Source: Ref 48

Table 4.21 Major phases formed in the nearsurface region of the test specimens after exposure to 100% N2 at 980 °C (1800 °F) for 1008 h, as determined by x-ray diffraction

Table 4.23 Results of x-ray diffraction analysis of extraction residues obtained from specimens after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h

Alloy

Alloy

Major phases

Type 314SS Type 330SS Alloy 800 Alloy 601 Alloy 600

(Cr,Fe)2N (Cr,Fe)2N (Cr,Fe)2N, AlN (Cr,Fe)2N, AlN CrN

Source: Ref 49

Table 4.22 Nitrogen absorbed (mg/cm2) and the average depth of internal nitridation for iron-, nickel-, and cobalt-base alloys after exposure in 100% N2 at 1090 °C (2000 °F) for 168 h Alloy

214 600 230 HR160 X 617 601 188 150 RA330 RA85H 556 HR120 253MA 800H 800HT Type 310 SS

Nitrogen absorbed, mg/cm2

Depth of internal nitridation, mm

0.2 1.1 2.7 3.9 6.0 5.1 7.2 3.7 9.0 6.6 8.5 9.0 9.6 10.0 10.3 11.4 12.3

0.0 0.41 0.46 1.19 0.63 >0.58 >0.59 >0.51 >0.80 >1.52 >1.44 >1.52 >0.86 >1.50 >1.50 >1.46 >0.79

214 (a) 230 600 601 617 HR160 188 150 RA85H

Phases detected

AlN, Al2O3 Cr2N, (Cr,Mo)12(Fe, Ni)8-xN4-z, M6C Cr2N, TiN Cr2N, AlN Cr2N, AlN CrN, Cr2N Cr2N Cr2N Cr2N, AlN

(a) Surface analysis. Source: Ref 50

Source: Ref 50

Fig. 4.33

with AlN/Al2O3 surface scales, as illustrated in Fig. 4.35. Alloy 150 (Co-27Cr-18Fe) suffered nitridation attack as severe as that experienced by some iron-base alloys. Figure 4.36 shows a through-thickness nitrided alloy 150 compared with Type 310SS. Extensive Cr2N nitrides were

Optical micrographs showing a through-thickness nitridation attack for alloy 617, a nickel-base alloy containing about 1.3%Al, after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h. Note extensive blocky chromium nitrides and needle-shaped aluminum nitrides. Magnification bar represents 200 μm. Courtesy of Haynes International, Inc.

observed throughout the alloy 150 specimen cross section.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:56PM Plate # 0

pg 91

Chapter 4: Nitridation / 91

Nickel-base alloys are in general more resistant to nitridation attack than iron-base alloys. This is illustrated in Fig. 4.37 comparing alloy X with 253MA after 168 h in 100% N2 at 1090 °C (2000 °F). Similar ﬁndings were observed in nitridation studies in nitrogen atmospheres by Smith and Bucklin (Ref 48) and Tjokro and Young (Ref 51). Tjokro and Young (Ref 51) investigated a number of commercial alloys in N2-5%H2 at 1100 and 1200 °C. Their results

10 µm

showed the nitridation rate constants decreased with increasing nickel concentration, as illustrated in Fig. 4.38.

(a)

200 µm

(b)

200 µm

Fig. 4.34

Scanning electron micrograph (backscattered image) showing internal chromium nitrides (blocky phases) and aluminum nitrides (long needle-shaped phases) formed in alloy 601, a nickel-base alloy containing about 1.3% Al, after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h. Source: Ref 50

4.50 230 alloy

N absorbed, mg/cm2

4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 0.00

214 alloy 200.00

400.00

600.00

Time, h

Fig. 4.35

Nitridation kinetic data for alloy 214 (nickel-base alloy containing 4.5% Al) and alloy 230 (nickelbase alloy containing little aluminum) after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h. Source: Ref 50

Fig. 4.36

Optical micrographs showing through-thickness nitridation attack for (a) Type 310SS (Fe-25Cr20Ni) and (b) alloy 150 (Co-27Cr-18Fe) after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h. Courtesy of Haynes International, Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:57PM Plate # 0

92 / High-Temperature Corrosion and Materials Applications

4.7 Nitridation Kinetics between NH3 and N2 Atmospheres Nitridation data generated in 100% NH3 (Ref 41) and those generated in 100% N2 (Ref 50) were compared. Since both test programs were generated using the same test apparatus and procedures, and both tests were carried out by same technicians, the laboratoryto-laboratory variation was signiﬁcantly minimized. Thus, the comparison between these two sets of test results could yield a more meaningful comparison in terms of the difference in environments. The test results generated at 1090 °C

(a)

200 µm

(2000 °F) for 168 h in 100% NH3 (Ref 41) and 100% N2 (Ref 50) are tabulated in Table 4.24. The data are also presented in terms of nitrogen absorption as a function of Ni + Co content in the alloy (Fig. 4.39). The results clearly indicated that the nitrogen atmosphere was a more severe nitriding environment than the ammonia environment at 1090 °C (2000 °F). The amount of nitrogen absorbed in N2 environment was more than double that in NH3 environment for many alloys. Figure 4.40 shows two nitrided alloy 601 specimens, one exposed to NH3 environment and the other to N2 environment. The N2 environment caused signiﬁcantly more internal nitride formation than for the NH3 environment. More chromium nitrides (blocky shaped) and aluminum nitrides (needle shaped) formed in the N2 environment than in the NH3 environment. Ammonia readily dissociates to one part N2 and three parts H2 at 1090 °C (2000 °F). With the test system used in the study by Barnes and Lai (Ref 41), 100% NH3 was fed into the alumina test tube with no test specimens inside, and the exhaust gas was measured to contain less than 5% NH3, which was the detection limit of the the apparatus used for measuring NH3 (Table 4.11). It is believed most, if not all, of the ammonia had been dissociated into H2 and N2 before the test gas was in contact with the test specimens. The NH3 test environment was essentially a cracked ammonia, which was dissociated into H2 and N2. Thus, the nitridation potential ( pN2 ) in the NH3 test environment (0.25 atm) was much lower than that in the N2 test environment (1.0 atm). As a result, the N2 test environment was found to produce more severe nitridation attack for most of the alloys tested (Table 4.24 and Fig. 4.39).

4.8 Summary

(b)

Fig. 4.37

200 µm

Optical micrographs showing a through-thickness nitrided Fe-20Cr-10Ni-1.7Si-Ce alloy 253MA (a) and a better nitridation resistant Ni-22Cr-9Mo-18Fe-0.6W alloy X after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h. Courtesy of Haynes International, Inc.

Nitridation behavior of metals and alloys in (a) air, (b) gas-turbine combustion gas, (c) NH3-H2O, (d) NH3, and (e) N2 environments is reviewed. Nitridation attack can occur in air and oxidizing, combustion environments. Under certain conditions, alloys can suffer oxidation/ nitridation attack. Internal nitridation attack is much more prevalent in a high-velocity combustion gas stream with thermal cycling. In NH3H2O environments, alloys appear to behave differently under nitridation attack. Extensive review is carried out on the behavior of metals

pg 92

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:57PM Plate # 0

pg 93

Chapter 4: Nitridation / 93

6000 Intragranular 1000 °C (1830 °F) 1100 °C (2010 °F)

309S

4000

kp, µm2/h

310S

153MA

2000

RA330 AC66 800

253MA

353MA IN601 0

0

0.200

0.400

0.600

0.800

1.000

XNi' (a) 12 Intergranular 1000 °C (1830 °F) 1100 °C (2010 °F)

153MA

kp, 103 µm2/h

253MA

800

309S 310S

6

RA330 AC66

353MA IN601

0

0

0.200

0.400

0.600

0.800

1.000

XNi' (b)

Fig. 4.38

Nitridation rate constants as a function of the alloy’s nickel concentration when tested in N2-5%H2 at 1000 and 1100 °C (1830 and 2010 °F). Source: Ref 51

and alloys in NH3 and N2 environments. Comparative resistance to nitridation attack for a wide variety of alloys is presented.

REFERENCES

1. Metals Handbook, Vol 2, 8th ed., American Society For Metals, 1964, p 149 2. Metals Handbook, Vol 2, 8th ed., American Society For Metals, 1964, p 119 3. G.L. Swales, Behavior of High Temperature Alloys in Aggressive Environments, Proc.

1979 Petten International Conference, I. Kirman et al., Ed., The Metals Society, London, 1980, p 45 4. K. Rorbo, Environmental Degradation of High Temperature Materials, Series 3, No. 13, Vol 2, The Institution of Metallurgists, London, 1980, p 147 5. R.N. Shreve, The Chemical Process Industries, McGraw-Hill, 1956 6. J.M.A. Van der Horst, Corrosion Problems in Energy Conversion and Generation, C.S. Tedmon, Jr., Ed., The Electrochemical Society, 1974

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:57PM Plate # 0

pg 94

94 / High-Temperature Corrosion and Materials Applications

Table 4.24 Nitrogen absorption in NH3 and N2 environments at 1090 °C (2000 °F) for 168 h Nitrogen absorption, mg/cm2 Alloy

214 600 230 617 601 X 188 150 556 RA330 800H 253MA

NH3

N2

0.2 0.3 1.5 1.9(a) 2.6(a) 3.8(a) 2.0(a) 4.1 4.2(a) 3.1(a) 5.5 6.3(a)

0.0 1.1 2.7 5.1(a) 7.2(a) 6.0(a) 3.7(a) 9.0(a) 9.0(a) 6.6(a) 10.3(a) 10.0(a)

(a) Nitrided all the way through specimens. Source: Ref 41, 50

13 1090 °C (2000 °F) / 168 h N2 12

(a)

200 µm

(b)

200 µm

NH3

Nitrogen absorption, mg/cm2

10

8

6

4

2

0

0

20

40

60

80

Ni or Ni + Co, wt%

Fig. 4.39

Nitrogen absorption as a function of Ni + Co content in the alloy for 100% NH3 and 100% N2 environments at 1090 °C (2000 °F) for 168 h. Source: Ref 41, 50

7. M.B. Bever and C.F. Floe, Source Book on Nitriding, American Society For Metals, 1977, p 125 8. B.J. Lightfoot and D.H. Jack, Source Book on Nitriding, American Society For Metals, 1977, p 248 9. K.N. Strafford, Corros. Sci., Vol 19, 1979, p 49 10. T. Rosenquist, Principles of Extractive Metallurgy, McGraw-Hill, 1974

Fig. 4.40

Optical micrographs showing both chromium nitride and aluminum nitride (needle-shaped phase) formed in alloy 601 after exposure to (a) 100% NH3 and (b) 100% N2 at 1090 °C (2000 °F) for 168 h. Courtesy of Haynes International, Inc.

11. T. Masumoto and Y. Imai, J. Jpn. Inst. Met., Vol 33, 1969, p 1364 12. H.J. Christ, S.Y. Chang, and U. Krupp, Thermodynamic Characteristics and Numerical Modeling of Internal Nitridation of Nickel Base Alloys, Mater. Corros., Vol 54 (No. 11), 2003, p 887

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:57PM Plate # 0

Chapter 4: Nitridation / 95

13. N.S. Corney and E.T. Turkdogan, The Effect of Alloying Elements on the Solubility of Nitrogen in Iron, J. Iron Steel Inst., Aug 1955, p 344 14. D.H. Jack and K.H. Jack, Carbides and Nitrides in Steel, Mater. Sci. Eng., Vol 11, 1973, p 1 15. H.A. Wriedt and O.D. Gonzalez, Trans. AIME, Vol 221, 1961, p 532 16. J.F. Eckel, T.P. Floridis, and B.N. Ferry, Nitrides in Type 304 Stainless Steel, Virginia J. Sci., Vol 17, 1966, p 325 17. J.F. Eckel and T.B. Cox, J. Mater., Vol 3, 1968, p 605 18. L.E. Kindlimann and G.S. Ansell, Kinetics of the Internal Nitridation of Austenitic Fe-Cr-Ni-Ti Alloys, Metall. Trans., Vol 1, 1970, p 163 19. A.J. Heckler and J.A. Peterson, The Effect of Nickel on the Activity of Nitrogen in Fe-Ni-N Austenite, Trans. Metall. Soc. AIME, Vol 245, 1969, p 2537 20. J. Litz, A. Rahmel, M. Schorr, and J. Weiss, Scale Formation on the Ni-Base Superalloys IN 939 and IN 738LC, Oxid. Met., Vol 32, 1989, p 167 21. M.A. Harper, J.E. Barnes, and G.Y. Lai, Long-Term Oxidation Behavior of Selected High Temperature Alloys, Paper No. 132, Corrosion/97, NACE International, 1997 22. G.Y. Lai, unpublished results, 2003 23. S. Han and D.J. Young, Simultaneous Internal Oxidation and Nitridation of NiCr-Al Alloys, Oxid. Met., Vol 55, 2001, p 223 24. D.L. Douglass, Anomalous Behavior During Internal Oxidation and Nitridation, JOM, Nov 1991, p 74 25. R.P. Rubly and D.L. Douglass, Oxid. Met., Vol 35, 1991, p 269 26. R.P. Rubly and D.L. Douglass, Internal Nitridation of Ni-Cr-Al Alloys, Proc. Int. Symp. On Solid-State Chemistry of Advanced Materials: High-Temperature Corrosion Workshop, 1992 27. H.J. Grabke and E.M. Peterson, Scr. Met., Vol 12, 1978, p 1111 28. J.-W. Park and C. J. Alstetter, Metall. Trans. A, Vol 18A, 1987, p 43 29. P.L. Gruzin, Y.A. Polikarpov, and G.B. Federov, Fiz. Metal. I Metalloved., Vol 4 (No. 1), 1957, p 94 30. K.G. Brickner, G.A. Ratz, and R.F. Domagala, Creep-Rupture Properties of Stainless Steels at 1600, 1800, and 2000 °F,

31. 32.

33.

34. 35.

36.

37.

38. 39. 40. 41.

Advances in the Technology of Stainless Steels and Related Alloys, STP 369, ASTM, 1965, p 99 M. Yu, R. Sandstrom, B. Lehtinen, and C. Westman, Scand. J. Metall., Vol 16, 1987, p 154 V. Guttmann and R. Burgel, CreepStructural Relationship in Steel Alloy 800H at 900–1000 °C, Met. Sci., Vol 17, 1983, p 549 M. Welker, A. Rahmel, M. Schutze, Oxidation and Nitridation of Alloy 800H at a Growing Creep Crack and for Unstressed Samples, Metall. Trans. A, Vol 20A, 1989, p 1541 J.J. Hoffman and G.Y. Lai, Paper No. 5402, Corrosion 2005, NACE International, 2005 V.P. Swaminathan and S.J. Lukezich, Degradation of Transition Duct Alloys in Gas Turbines, Advanced Materials and Coatings for Combustion Turbines, Proc. ASM 1993 Materials Congress Materials Week (Pittsburgh, PA), Oct 17–21, 1993, V.P. Swaminathan and N.S. Cheruvu, Ed., ASM International, 1994, p 99 G.Y. Lai, Nitridation of Several Combustor Alloys in a Simulated Gas Turbine Combustion Environment, Advanced Materials and Coatings for Combustion Turbines, Proc. ASM 1993 Materials Congress Materials Week (Pittsburgh, PA), Oct 17–21, 1993, V.P. Swaminathan and N.S. Cheruvu, Eds., ASM International, 1994, p 113 G.Y. Lai, Nitridation Attack in a Simulated Gas Turbine Combustion Environment, Materials for Advanced Power Engineering, Part II, D. Coutsouradis et al., Ed., Kluwer Academic Publishers, The Netherlands, 1994, p 1263 G.Y. Lai, unpublished results, Haynes International, Inc., 1995 Y.M. Park and R.E. Sonntag, Int. J. Energy Res., Vol 14, 1990, p 153 H.J. Grabke, S. Strauss, and D. Vogel, Nitridation in NH3-H2O Mixtures, Mater. Corros., Vol 54 (No. 11), 2003, p 895 J.J. Barnes and G.Y. Lai, High Temperature Nitridation of Fe-, Ni-, and Co-base Alloys, Corrosion & Particle Erosion at High Temperatures, Proc. TMS-ASM Symposium, V. Srinivasan and K. Vedula, Ed., The Minerals, Metals & Materials Society, 1989, p 617

pg 95

Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/

30/10/2007 3:57PM Plate # 0

96 / High-Temperature Corrosion and Materials Applications

42. K.M. Verma, H. Ghosh, and J.S. Rai, Brit. Corros. J., Vol 13 (No. 4), 1978, p 173 43. V. Cihal, Corrosion Mechanisms in Ammonia Synthesis Equipment, Conf. Proc., First International Congress on Metallic Corrosion (London, U.K.), April 10–15, 1961, L. Kenworthy, Ed., Butterworths, London, 1962, p 591 44. J.J. Moran, J.R. Mihalisin, and E.N. Skinner, Corrosion, Vol 17 (No. 4), 1961, p 191t 45. D.W. McDowell, Jr., Mater. Protect., Vol 1 (No. 7), 1962, p 18 46. K.H. Jack, High Temperature Gas-Metal Reactions in Mixed Environments, S.A. Jansson and Z.A. Foroulis, Ed., The Metallurgical Society of AIME, 1973, p 182 47. H. Schenck, M.G. Frohberg, and F. Reinders, Stahl Eisen, Vol 83, 1963, p 93

48. G.D. Smith and P.J. Bucklin, Some Observation on the Performance of Nickel-Containing Commercial Alloys in Nitrogen-Based Atmospheres, Paper No. 375, Corrosion/86, NACE, 1986 49. P. Ganesan and G.D. Smith, Performance of Selected Commercial Alloys in Nitrogen Based Sintering Atmospheres, Paper No. 278, Corrosion/90, NACE, 1990 50. J.J. Barnes and G.Y. Lai, Factors Affecting the Nitridation Behavior of Fe-Base, NiBase and Co-Base Alloys in Pure Nitrogen, J. Physique IV, Colloque C9, supplemental au Journal de Physique III, Vol 3, 1993, p 167 51. K. Tjokro and D.J. Young, Comparison of Internal Nitridation Reactions in Ammonia and in Nitrogen, Oxid. Met., Vol 44, 1995, p 453

pg 96

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p97-145 DOI: 10.1361/hcma2007p097

26/10/2007 12:21PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 5

Carburization and Metal Dusting 5.1 Introduction Metals and alloys are susceptible to carburization when exposed to an environment containing CO, or CH4 or other hydrocarbon gases, such as ethane (C2H6), propane (C3H8), and so forth, at elevated temperatures. Carburization attack generally results in formation of internal carbides, which often cause the alloy to suffer embrittlement as well as other mechanical property degradation. Carburization problems are quite common to heat treating equipment, particularly furnace retorts, baskets, fans, and other components used for case hardening of steels by gas carburizing. A common commercial practice for control of gas carburizing is to use an endothermic gas as a carrier enriched with one of the hydrocarbon gases, such as CH4, C3H8, and so forth (Ref 1). An endothermic gas enriched with about 10% natural gas (CH4) is a commonly used atmosphere (Ref 2). The typical endothermic gas consists of 39.8% N2, 20.7% CO, 38.7% H2, and 0.8% CH4, with a dew point of −20 to −4 °C (−5 to +25 °F) (Ref 1). Gas carburizing occurs typically at 840 to 930 °C (1550 to 1700 °F). Furnace equipment and components repeatedly subjected to these service conditions frequently suffer brittle failures as a result of carburization attack. In the petrochemical industry, carburization is one of the major modes of high-temperature corrosion for processing equipment. The pyrolysis furnace tubes for production of ethylene and oleﬁns are a good example (Ref 3–5). Ethylene is formed by cracking petroleum feedstock, such as ethane and naphtha, at temperatures up to 1150 ° C (2100 °F). This generates a strong carburizing gas stream inside the tubes. As a result, carburization was found to be a major mode of tube failure in a survey of ethylene and oleﬁn pyrolysis furnaces conducted by Moller and Warren (Ref 3).

Production of carbon ﬁbers also generates carburizing atmospheres in a furnace. As a result, the furnace’s retorts, ﬁxtures, and other components require frequent replacement because of carburization attack. Metal dusting, a form of catastrophic carburization, can occur at intermediate temperatures when a process gas stream consists of primarily H2, CO, and CO2 along with some hydrocarbons with high carbon potentials (ac > 1). Metals or alloys can suffer rapid metal wastage in a form of pitting or general thinning of the cross-sectional thickness of a metallic component. Metal dusting typically occurs at temperatures between 430 and 900 °C (800 and 1650 °F). Materials failures associated with metal dusting have been encountered in reﬁning and petrochemical processing, such as production of syngas in hydrogen, ammonia, and methanol plants, heat treating, and other industrial processes (Ref 5–10).

5.2 Carburization 5.2.1 Carburization—Thermodynamic Considerations Whether an alloy is likely to be carburized or decarburized depends on the carbon activity (ac) in the environment and that of the alloy. The thermodynamic condition that dictates either carburization or decarburization can be described simply. The alloy is likely to be carburized when: (ac )environment 4 (ac )metal

The alloy is likely to be decarburized when: (ac )environment 5 (ac )metal

Thus, in order to predict whether an alloy will be carburized, one needs to know the carbon activities of both the environment and the alloy.

pg 97

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:21PM Plate # 0

pg 98

98 / High-Temperature Corrosion and Materials Applications

Carburization can proceed by one of the following reactions when the environment contains CH4, CO, or H2 and CO: CO+H2 =C+H2 O

ðEq 5:1Þ

2CO=C+CO2

ðEq 5:2Þ

CH4 =C+2H2

ðEq 5:3Þ

Assuming that carburization follows Reaction 5.1, the carbon activity in the environment can be calculated by: DG =7RT ln

ac pH2 O pCO pH2

ðEq 5:4Þ

ac =e7DG

=RT

pCO pH2 pH2 O

DG =7RT ln

ac =e7DG

=RT

ðEq 5:5Þ

p2CO pCO2

ðEq 5:6Þ

ðEq 5:7Þ

pCH4 p2H2

!

ðEq 5:8Þ

Carbon activities as a function of ( pCH4 =p2H2 ) are plotted in Fig. 5.3. Reactions 5.1 and 5.2 have a similar characteristic, showing lower carbon activities with increasing temperature (Fig. 5.1 and 5.2).

H2 = C + H2O

Carbon activity (ac) as a function of gaseous composition in terms of (pCO pH2 =pH2 O ) ratios based on Eq 5.1 for various temperatures. Also plotted are carbon activities for carbon steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo and austenitic stainless steels (both measured ac).

ac pCO2 p2CO

Plots of carbon activities as a function of gas compositions in terms of (p2CO =pCO2 ) for various temperatures are shown in Fig. 5.2. When carburization follows Reaction 5.3, the carbon activity in the environment is: ac =e

From Eq 5.5, one can construct graphs of carbon activity as a function of gaseous composition in terms of ( pCO pH2 =pH2 O ) ratios for various temperatures, as shown in Fig. 5.1.

Fig. 5.1

7DG =RT

Rearranging the equation changes it to:

Similarly, if carburization follows Reaction 5.2, the carbon activity of the environment can also be calculated:

2CO = CO2 + C

Fig. 5.2

Carbon activity (ac) as a function of gaseous com2 position in terms of (pCO =pCO2 ) based on Eq 5.2 for various temperatures. Also plotted are carbon activities for carbon steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo and austenitic stainless steels (both measured ac).

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:21PM Plate # 0

Chapter 5:

Reaction 5.3, on the other hand, shows increased carbon activities with increasing temperature (Fig. 5.3). If the environment contains CH4, the carbon activity of the environment at higher temperatures is likely to be dominated by Reaction 5.3. When no CH4 is present in the environment, Reaction 5.1 and/or 5.2 will dictate the carbon activity. The carbon activity maps shown in Fig. 5.1 to 5.3 were previously described by Mazandarany and Lai (Ref 11) in assessing the carburizationdecarburization behavior of alloys in hightemperature gas-cooled helium environments containing H2, CO, CO2, CH4, and H2O. These activity maps provide a simple means of estimating an environment’s carbon activity for predicting whether or not the environment is thermodynamically capable of carburizing an alloy. When the gas stream contains many gaseous components, such as H2, CO, CO2, CH4, and

pg 99

Carburization and Metal Dusting / 99

H2O, and under very dynamic conditions with a high gas velocity such that the gaseous components do not have time to react to reach a thermodynamic equilibrium (i.e., nonequilibrium conditions), the gas-metal reaction can be reasonably assumed to follow the dominating reaction from one of those shown in Reaction 5.1, 5.2, or 5.3, and thus, one of the activity maps shown in Fig. 5.1, 5.2, or 5.3. Data for carbon activities of commercial alloys at temperatures below 1200 °C (2190 °F) is very limited. Natesan (Ref 12) reported that ac for 2.25Cr-1Mo steel is in the range of 1×10−1 to 10−2 from 550 to 750 °C (1020 to 1380 °F). Natesan and Kassner (Ref 13) reported the carbon activities of Fe-18Cr-8Ni alloys. These values are superimposed in Fig. 5.1 to 5.3. For carbon steels, carbon activity can be estimated by assuming that it is in equilibrium with cementite (Fe3C): ðEq 5:9Þ

3Fe+C=Fe3 C DG =7RT ln DG =7RT ln

aFe3 C (ac ) (aFe )3

1 ac

ðEq 5:10Þ

ðEq 5:11Þ

where aFe3 C and aFe are assumed to be unity. ac =eDG

CH4 = 2H2 + C

Fig. 5.3

Carbon activity (ac) as a function of gaseous com2 position in terms of (pCH4 =pH ) based on Eq 5.3 for 2 various temperatures. Also plotted are carbon activities for carbon steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo (measured ac). Carbon activities of austenitic stainless steels are below 10−2 at 800 to 1000 °C (1470 to 1830 °F).

=RT

ðEq 5:12Þ

The ac values for carbon steel based on Eq 5.12 are plotted in Fig. 5.1 to 5.3. Using Fig. 5.1, 5.2, or 5.3 one can make a quick determination as to whether an environment has a carbon potential (or activity) high enough to carburize the alloy of interest. Even though in cases where the gas mixture may not reach an equilibrium condition, it will be of great beneﬁt to better understand the gas-metal reaction in terms of the thermodynamic equilibrium condition in multicomponent gases environments. The thermodynamic equilibrium gaseous composition along with its thermodynamic potentials, such as carbon activity (ac), oxygen potential ( pO2 ), and other potentials, can be determined using a commercial software program such as “HSC Chemistry for Windows” (Ref 14) and ChemSage (Ref 15). The environment can also be characterized in terms of ac and pO2 to determine the relative severity of its carburization potential. Both carbon activity and oxygen potential can be calculated by a computer program. The environment

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:21PM Plate # 0

100 / High-Temperature Corrosion and Materials Applications

2

Log PCO /PCO

2

During carburization, the relative stabilities of these carbides can be best described by a stability diagram, such as the one shown in Fig. 5.5. If the carbon and oxygen activities of the environment are in the Cr3C2 region, conditions will favor formation of Cr3C2 on the surface and/or in the underlying metal. As carbon diffuses farther into the alloy’s interior, carbon activities will be lowered, thus favoring Cr7C3. Moving even farther into the interior, carbon activities will be further reduced, favoring the formation of Cr23C6. These chromium carbides can incorporate other alloying elements depending on alloy system. For example, in Fe-Ni-Cr system, iron with very little nickel can be incorporated into these chromium carbides. The combined metal elements in the carbide are then represented by “M,” as M3C2, M7C3, and M23C6. An example of the metallic compositions of M7C3 and M23C6 formed in carburization of Type 304L is illustrated in Fig. 5.6 (Ref 20). Both M7C3 and M23C6 contain essentially Cr and Fe with negligible amount of Ni. For many high-temperature alloys, particularly superalloys, there are other alloying elements, such as Ti, Ta, Nb (or Cb), Mo, and W, that can form carbides. The carbides of these alloying elements are important to the physical metallurgy of high-temperature alloys in that they provide an important strengthening mechanism. A general review of binary metallic Log PCH /PH 2 4 2

can then be presented in a stability diagram of a metal-carbon-oxygen system. The stability diagrams of Fe-C-O and Cr-C-O systems are shown in Fig. 5.4 and 5.5 (Ref 16). From the stability diagram, the possible phases that the alloy may form at the gas/metal interface can be predicted. As the activities of both carbon and oxygen are decreasing from the gas/metal interface to the metal interior, the possible phases that the alloy may form beneath the gas/metal interface can also be predicted. For carbon and alloy steels with low concentrations of chromium, ingress of carbon into the metal or alloy may result in the formation of iron carbides. Several forms of iron carbides have been reported (Ref 17), with compositions ranging from Fe4C to Fe2C. They are ξ phase (Fe4C), θ phase (Fe3C), χ phase (Fe2.2C), and ε phase (Fe2-3C). Fe3C (cementite) is the most stable iron carbide. Other iron carbides are less stable. Browning et al. (Ref 18) found that χ phase, which formed by carburizing αFe with butane at 275 °C (530 °F), was converted to Fe3C when heated to 500 °C (930 °F). The ε phase (Fe2-3C) is a transition phase that forms in martensite during tempering of steel (Ref 19). In ferritic and austenitic stainless steels and nickel- and cobalt-base alloys, ingress of carbon into the alloy results in the formation of mainly chromium carbides. There are three forms of chromium carbides: Cr23C6, Cr7C3, and Cr3C2.

Log PCO /PCO –18 –16 –14 –12 –10 –8 –6 –4 –2 2 Log P /P –18 –16 –14 –12 –10 –8 –6 –4 –2 H2O

H2

0

2

4

6

8

10 12

0

2

4

6

8

10 12

0 2

C(s)

0

–2 0

–2

–4

Fe3C(s)

–2

Log aC

–4

–6 –4

–6 –8 –10

–8 Fe(s)

Fe3C4(s)

–6 –10

Fe0.95O(s)

–8 –12 –10

–12

–14

–14

–16

–16

–12 –14 –50 –45 –40 –35 –30 –25 –20 –15 –10 Log PO , atm 2

Fig. 5.4

Fe2O3(s)

Stability diagram of Fe-C-O system at 870 °C (1600 °F). Source: Ref 16

–5

0

5

pg 100

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:21PM Plate # 0

pg 101

H2

Cr3C2(s)

0 –2

10 12 14

8

6

8

10 12 14

C(s)

Cr7C3(s)

–4 Cr23C6(s)

Log aC

–6 –8

Cr2O3(s)

–10

2

2

H2O

6

2

4

Log PCO /PCO –14 –12 –10 –8 –6 –4 –2 0 2 4 2 –14 –12 –10 –8 –6 –4 –2 0 2 4 Log P /P

Log PCO /PCO

Log PCH /PH2

Chapter 5: Carburization and Metal Dusting / 101

0

0

–2

–2

–4

–4

–6

–6

–8

–8

–10 –10

Cr(s)

–12

–12 –12

–14

–14 –14 –16 –16

–16 –50 –45 –40 –35 –30 –25 –20 –15 –10

–5

0

5

Log PO , atm

(a)

H2O

H2

0 –2

0

2

4

6

8 10 12

0

2

4

6

8 10 12

0 2

Cr3C2(s)

C(s)

–2 0

Cr7C3(s)

–4 –2

–4

–6

Log aC

Cr23C6(s)

–4

–6

–8 –6

–8 –10

Log PCH /PH 2

Log PCO /PCO –18 –16 –14 –12 –10 –8 –6 –4 –2 2 –18 –16 –14 –12 –10 –8 –6 –4 –2 Log P /P

4 2 2 Log PCO /PCO 2

2

–10

Cr2O3(s)

Cr(s)

–8 –12 –10

–12

–14

–14

–16

–12 –14

–16 –50 –45 –40 –35 –30 –25 –20 –15 –10 (b)

Fig. 5.5

–5

0

5

Log PO , atm 2

Stability diagrams of Cr-C-O system at (a) 620 °C (1150 °F), (b) 870 °C (1600 °F), and (c) 1090 °C (2000 °F). Source: Ref 16

carbides can be found elsewhere (Ref 21, 22). The relative stabilities of some binary carbides are shown in Fig. 5.7 (Ref 22). When the environment contains oxygen and carbon activities, temperature is an important factor in determining whether the oxide or carbide will be thermodynamically stable. Considering a chemical reaction, such as 3Cr2O3 + 4C = 2Cr3C2 + 9/2 O2, Cr3C2 will remain stable when the reaction goes from left to right. In order

to keep the reaction going from left to right (i.e., keeping Cr3C2 stable), the pO2 of the environment shall be lower than the equilibrium pO2 associated with the above reaction. On the other hand, if the pO2 of the environment is higher than the equilibrium pO2 associated with the above reaction, Cr2O3 will become stable. Temperature can be a signiﬁcant factor in determining whether chromium oxide or chromium carbide is stable, and thus signiﬁcantly affects the carburization

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:21PM Plate # 0

H2

0

Cr3C2(s)

4

6

8

10

4

6

8

10

4 –2

C(s)

2

Cr7C3(s)

–2

–4 0 –6

Cr23C-6(s)

–4 Log aC

2 2

–2 –8

–6

–4

–8

–10

Cr2O3(s)

Cr(s)

–6 –12

–10

–8 –14

–12

–10 –16

–14

–12 –18

–16 –50 –45 –40 –35 –30 –25 –20 –15 –10

Concentration, wt.%

1123 k 150 h

80

M23C6 Cr

40

Fe

Ni

M7C3

0 0.2

5

(c) 1090 °C (2000 °F). Source: Ref 16

100

0

0

2

Fig. 5.5 (continued)

20

–5

Log PO , atm

(c)

60

2

2

H2O

0 0

2

4

Log PCO /PCO –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 2 Log P /P –20 –18 –16 –14 –12 –10 –8 –6 –4 –2

Log PCO /PCO

Log PCH /PH 2

102 / High-Temperature Corrosion and Materials Applications

0.4

0.6

0.8

1

Distance to surface, mm

Fig. 5.6

Compositions of the metallic components of M7C3 and M23C6 formed in Type 304L after carburizing at 1123 K (850 °C) in H2-2.6CH4 (ac = 0.9) for 150 h. Source: Ref 20

behavior of an alloy. This can be nicely illustrated in a plot that contains the oxygen potentials of the environment (Boudouard reaction [2CO = C + CO2] is used for this example) and those in equilibrium with Cr2O3/Cr3C2 as a function of temperature. This is shown in Fig. 5.8 (Ref 23). The ﬁgure shows the equilibrium pO2 line of Cr2O3/Cr3C2 intersecting with the environment’s pO2 lines (pCO = 1.0 bar, pCO = 0.5 bar, and pCO = 0.25 bar). The intersections are between 1000 and 1200 °C. On the right side of the intersections (i.e., lower temperatures), pO2 (environment) is greater than pO2 (Cr2O3/Cr3C2).

Thus, Cr2O3 is stable. On the left side of the intersections (i.e., higher temperatures), pO2 (environment) is lower, thus favoring the formation of carbides, but not oxides. Accordingly, higher temperatures favor carburization thermodynamically and lower temperatures favor formation of oxides thus retarding carburization. Nishiyama et al. (Ref 24) examined the effect of the temperature on the stability of chromium oxide and carbides based on the ethylene pyrolysis environment, which is generated by the reaction of naphtha with steam for the production of ethylene (C2H4) and propylene (C3H6) at temperatures of approximately 900 to 1100 °C (1650 to 2012 °F). In their calculation of the oxygen potential for the reaction of naphtha with steam, three naphtha feedstocks were used with the steam/naphtha weight ratios of 0.4 and 0.5. The calculated pO2 for the pyrolysis environment as a function of temperature is plotted in Fig. 5.9. Also plotted in Fig. 5.9 are pO2 values in equilibrium with Cr3C2/Cr2O3 and those in equilibrium with Cr7C3/Cr2O3. The results in Fig. 5.9 are very similar to those shown in Fig. 5.8, where the environment was calculated from CO-CO2 reaction (Boudouard reaction). The calculation by Nishiyama et al. showed that the chromium oxide was stable up to 1030 to 1040 °C (1886 to 1904 °F) and became unstable above those temperatures. At temperatures above 1030 to

pg 102

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:21PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 103

Fig. 5.7

Standard free energies of formation for carbides. Source: Ref 22

1040 °C (1886 to 1904 °F), carbides such as Cr3C2 and Cr7C3 became stable. This oxidecarbide transition temperature can vary depending on the steam/naphtha ratio, as illustrated in Fig. 5.10 (Ref 24). As shown in the ﬁgure, when the steam/naphtha ratio is decreased to 0.1 from

the common operating ratios of 0.35 to 0.5, the oxide-carbide transition temperature is decreased to about 970 °C (1778 °F). In Reactions 5.1 to 5.3, carbon deposition (coking) can occur when the carbon activity (ac) in the environment is greater than 1.0 (ac = 1

pg 103

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

104 / High-Temperature Corrosion and Materials Applications

in equilibrium with graphite). Many laboratory carburization tests have been conducted in H2-CH4 mixtures. Figure 5.11 shows the carbon activities of H2-1%CH4 and H2-2%CH4 as a function of temperature (Ref 25). Grabke

600

800

1000

Cr3C2/Cr2O3

–10

PCO–1 bar

Cr3C2 –15

PCO–0.5 bar PCO–0.25 bar

2

Log PO , bar

1200

1800 1600 1400

T, °C

(Ref 25) recommended that the H2-1%CH4 mixture with ac <1.0 be used for laboratory testing to correspond to the industrial process that does not form coking. He also recommended the use of the H2-2%CH4 mixture with ac > 1 for testing to correspond to the processes where coking is taking place. The test temperature is thus recommended to be higher than 1000 °C (1832 °F) (see Fig. 5.11). Ethylene pyrolysis environment is known to develop coking on the internal surface of the pyrolysis furnace tubes. The deposition of carbon is the result of decomposition of ethylene (C2H4) in the reaction as described in Eq 5.13 (Ref 26). Signiﬁcant coking in the internal surface of the ethylene pyrolysis furnace tube and the repeated decoking

–20 C/CO-CO2

Cr2O3

–25 5.10–4

75.10–4

10.10–4

12.5.10–4

1/T, K–4

Fig. 5.8

The oxygen potentials of the environment based on the Boudouard reaction (2CO = C + CO2) and those in equilibrium with Cr2O3/Cr3C2 as a function of temperature. Source: Ref 23

T, °C 0,150 1,100 1,050 1,000 –17

900

850

Equilibrium pO2 of the environment based on ethylene pyrolysis of naphtha I with various steam/ naphtha weight (S/O) ratios as a function of temperature, and pO2 in equilibrium with Cr3C2/Cr2O3 and Cr7C3/Cr2O3 as a function of temperature. Source: Ref 24

Fig. 5.10

Cr-carbides + metastable Cr2O3

–18

Cr2O3+ metastable Cr-carbides –19

ac

2

Log PO .atm

950

–20

) =1 (a c 3 ) r 2O =1 /C C2 (a c 3 Cr 3 r 2O /C C3 Cr 7

II (0

Na

.5)

pht

–21 Cr-carbides –22

5

7

7.5

8

ha

and

III (

)

.4)

8.5

2% CH4

4

0.5

I (0

Carbon activity in H2-CH4 1 bar

3

9

1/T × 104, 1/K Equilibrium pO2 of the environment based on ethylene pyrolysis of naphtha with steam/naphtha ratios of 0.4 and 0.5 (in parentheses) as a function of temperature, and pO2 in equilibrium with Cr3C2/Cr2O3 and Cr7C3/Cr2O3 as a function of temperature. Naphtha I, II, and III represent different naphtha feed stocks. Source: Ref 24

1% CH4

2 1

Fig. 5.9

800

Fig. 5.11 Ref 25

850

900

950 1000 1050 1100 1150 °C

Carbon activities (ac) of H2-1%CH4 and H2-2%CH4 are plotted as a function of temperature. Source:

pg 104

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 105

operation to remove this coke deposit can signiﬁcantly degrade the tube life (Ref 27). C2 H4 =2C+2H2

ðEq 5:13Þ

5.2.2 Resistance to Carburization Carburization attack generally results in the formation of internal carbides in the alloy matrix as well as at grain boundaries. The gravimetric method has been widely used for studying carburization kinetics. This method can sometimes produce a misleading result when the environment exhibits an oxygen potential high enough to form oxides of some active alloying elements. The weight gain, in this case, is the result of both carbon ingress and oxide formation. Measurements of carburization depth have also been used by some investigators. Different alloy systems can produce signiﬁcant differences in the concentration proﬁle for the carburized layer. Thus, one alloy may exhibit a large carburization depth with only a slight concentration gradient, while another alloy may show a narrow carburization depth with a steep concentration proﬁle. Furthermore, measurements of carburization depth by the metallographic method can be difﬁcult when separating the carbides formed by carburization from those formed by thermal aging. Some investigators measured the total amount of carbon in the alloy after the exposure. The measurement of the carbon concentration proﬁle as a function of distance from the metal surface may be an excellent method for characterizing

Carbon content, %

5

A

4 B

3

2 C 1

0 ID

Distance from bore of tube, mm

OD

Wall thickness

Fig. 5.12

Three possible carbon concentration profiles of carburized alloys. Source: Ref 28

the carburized alloy. Each evaluation method has its merit. It certainly will be beneﬁcial to use as many evaluation methods as possible for characterizing the carburized alloy. With respect to the impact of carburization on an alloy’s performance, Krikke et al. (Ref 28) believed that not only the total amount of carbon absorbed but also the maximum carbon level and the maximum carbon concentration gradient are the most important factors. Figure 5.12 illustrates three possible carbon concentration proﬁles, as suggested by Krikke et al. (Ref 28). They considered proﬁle A with a steep concentration proﬁle to be the most damaging. Heubner (Ref 29) tested various commercial alloys in H2-CH4 gas mixture (ac = 0.8) at 1000 °C (1832 °F) and observed a steep carbon concentration proﬁle in Fe-Ni-Cr alloys and a low ﬂat concentration proﬁle in Ni-Cr alloys, as shown in Fig. 5.13. One Ni-Cr alloy (alloy 45TM), which contained relatively high Fe and high Si, was an exception showing a steep carbon concentration proﬁle similar to Fe-Ni-Cr alloys such as 800H, AC66, and DS. The Fe-Ni-Cr alloys were found to have suffered more room-temperature impact toughness drop in general than Ni-Cr alloys (Ref 29). However, the relative room-temperature impact toughness loss (%) was found to increase with increasing total carbon pickup (Ref 29). For carburization, the real issue is the effect of carburization on the alloy’s mechanical properties, such as creep-rupture properties and toughness or ductility. This type of data, however, is quite limited and is inadequate as a basis for making an informed materials selection. Thus, this chapter reviews mainly the carburization data in terms of mass gain, mass of carbon absorption, carburization depth, and concentration proﬁle of the carburized layer. When the environment is such that no protective oxide scale (e.g., Cr2O3 scale) is formed on the metal surface, carburization is controlled by diffusivity and solubility (Ref 30). The ingress of carbon will be greatly reduced when a chromium oxide scale is developed. Carburization kinetics in this case are then controlled by the diffusion of carbon through the oxide scale. Wolf and Grabke (Ref 31) demonstrated that there was no detectable solubility of carbon in Cr2O3 oxides by equilibrating the oxides with CO2-CO mixtures tagged with radioactive 14C at 1000 °C (1832 °F). Thus, carbon permeation is not possible through the perfectly dense Cr2O3 oxide layer, unless the oxide layer contains pores and ﬁssures (Ref 31).

pg 105

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

106 / High-Temperature Corrosion and Materials Applications

alloy compared with the as-cast surface ﬁnish (Fig. 5.17). Norton and his colleagues (Ref 35–37) at Petten Laboratories have conducted a series of studies on the effects of silicon, niobium, chromium, and iron in Fe-Ni-Cr alloys, including four commercial alloys (HK-40, HP-40Nb, Type 314 SS, and alloy 800H) and three experimental alloys. Their test environments had ﬁxed carbon activities (ac) of 0.3 and 0.8, with various oxygen potentials (pO2 ) at temperatures from 825 to 1050 °C (1520 to 1920 °F), as illustrated in Fig. 5.18. The oxygen potentials of the test environments were below that in equilibrium with Cr2O3 (Fig. 5.18). That means that no chromium oxide scales should have formed on the metal surface. However, a SiO2 scale was likely to form at 825 °C (1520 °F), but not at 1000 °C (1830 °F). The test results at 825 °C (1520 °F) showed that Type 314 stainless steel (2.04% Si) was signiﬁcantly more carburization

Resistance to carburization is an important factor in the performance of pyrolysis furnace tubes as well as pigtails for ethylene and oleﬁn plants. Furnace tubes are typically constructed of Fe-Ni-Cr cast alloys, such as HK (Fe-25Cr-20Ni or 25/20), HP (Fe-25Cr-35Ni or 25/35) and their variants. Some of these variants involved additions of niobium (or columbium), tungsten, molybdenum, silicon, and titanium. Some also involved increases in nickel and/or chromium. Some of the modiﬁed alloys are referred to as “microalloyed” castings. It has been found that these additions and increases improve carburization resistance as well as creep-rupture strengths. Figures 5.14 to 5.17 illustrate the carburization resistance of some of these modiﬁed alloys compared with HK alloy (Ref 32–34). Also shown in Fig. 5.17 is the effect of the surface ﬁnish on carburization resistance of the alloy. The machined-ﬁnished surface signiﬁcantly reduced the carbon ingress into the 3.5 AC 66

Carbon concentration, %

3 2.5

alloy 800 H

2

alloy DS

1.5 1 0.5 0 0

1

2

3

4

5

∞

5

∞

Distance to surface, mm

(a) 2.5

45 TM

Carbon concentration, %

2

1.5

1

alloy 600H

0.5 alloy 617 alloy 602 CA

0 0

1

2

3

4

Distance to surface, mm (b)

Fig. 5.13

Carbon concentration profile for alloys after testing at 1000 °C (1832 °F) for 1008 h in a H2-CH4 mixture (ac = 0.8) for (a) Fe-Ni-Cr and (b) Ni-Cr alloys. Source: Ref 29

pg 106

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 107

resistant than HK-40 (1.35% Si), HP-40Nb (1.29% Si), and alloy 800H (0.4% Si), as shown in Fig. 5.19. Also revealed by the test results was that the model alloys were signiﬁcantly less resistant to carburization than the commercial alloys with the same chromium, nickel, and iron. The model alloys had much lower silicon levels (0.13–0.28%) as well as manganese, aluminum, titanium, and so forth. The beneﬁcial effect of silicon on carburization resistance was clearly demonstrated when the data were replotted (Fig. 5.20). The presence of a SiO2 scale was conﬁrmed by Van der Biest et al. (Ref 36). When the test temperature was increased to 1000 °C (1832 °F), where the pO2 was below that in equilibrium of SiO2 (Fig. 5.18), Type 314 SS was found to be similar to HK-40, alloy 800H, and HP-40Nb (Fig. 5.21). Under these conditions, SiO2 was no longer thermodynamically stable. Thus, the silicon effect was diminished. The model alloy 50/50 (50Ni-50Cr with very low silicon level) was among the best performers in the alloys tested. The ratio of Ni to Cr + Fe is an important factor (Ref 35) in governing carburization resistance, as shown in Fig. 5.22. Decreases in Cr+ Fe in Fe-Ni-Cr alloys improved carburization resistance. Both chromium and iron are carbide

formers, constituting the major elements in M7C3 and M23C6 carbides resulting from carburization. Harrison et al. (Ref 37) found that the surface carbides removed from HK-40 sample after testing at 1000 °C (1830 °F) contained 55 to 58% Cr and 41 to 43% Fe, with very little nickel (approximately Cr4Fe3C3). Nickel reduces the diffusivity of carbon in Fe-Ni-Cr alloys, as demonstrated by Demel et al. (Ref 38) in Fig. 5.23. Nickel also decreases the solubility of carbon in Fe-Ni alloy system as shown in Fig. 5.24 (Ref 39). Decreases in carbon diffusivity and solubility can result in increases in carburization resistance. Grabke et al. (Ref 40) observed that increasing nickel improved carburization resistance in Fe-Ni-Cr alloys, with the maximum resistance achieved when the ratio of Ni to Fe was 4 to 1. This is in general agreement with the product of carbon solubility and diffusivity (Ref 41). High-nickel alloys are generally more resistant to carburization than Fe-Ni-Cr alloys. This is illustrated by the test results of Klower and Heubner (Ref 29), as shown in Fig. 5.25. The beneﬁcial effect of nickel on carburization resistance can also be clearly revealed in Fig. 5.26 when the data of Fe-Ni-Cr alloys

Fig. 5.14

Fig. 5.15

Carburization resistance of HK (25Cr-20Ni) and several HP alloys (Cr/Ni) as a function of temperature in pack carburization tests. Source: Ref 32

Carbon concentration profiles for HK (25Cr-20Ni) and several HP alloys (Cr/Ni) carburized at 1100 °C (2010 °F) for 520 h in pack carburization tests. Source: Ref 32

pg 107

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

108 / High-Temperature Corrosion and Materials Applications

and Ni-Cr alloys generated at 1000 °C for 1008 h were plotted as a function of iron content (i.e., decreasing nickel content) (Ref 29). The oxygen partial pressure (pO2 ) of the test environment—which was 10−27.92, 10−25.25, and 10−23.88 for 850, 1000, and 1100 °C, respectively—was below the pO2 in equilibrium of Cr2O3 (Ref 29). Accordingly, no chromium oxide scale would be present to provide protection for the alloys in these tests. In an extensive study undertaken by Steel and Engel (Ref 42) on the relative inﬂuence of nickel and chromium on the carburization resistance of Fe-Ni-Cr alloys, standard heats of ASTM grades varying from HC to HX cast alloys were investigated, along with many experimental cast alloys. They found nickel to be beneﬁcial, as illustrated in Fig. 5.27. The role of chromium, however, appeared to be different for different levels of nickel, as shown in Fig. 5.28. For ironbase alloys with 25% or less nickel, increasing chromium signiﬁcantly reduced carbon pickup. A slight decrease in carbon pickup with increasing chromium was noted for alloys

Fig. 5.16

containing 26 to 45% Ni. For alloys containing 46 to 70% Ni, increasing chromium resulted in an increase in carbon pickup. In this study, although there is no mention of the oxygen potential of the test environment, it is believed that the chromium oxide scale was not involved in the carburization reaction. Wolfe (Ref 43) also found that a higher Ni-containing Fe-Cr-Ni cast alloy, alloy HU (Fe-19Cr-39Ni-0.5C), was signiﬁcantly better than Type 304 (19Cr-9Ni) and Type 321 (18Cr12Ni) with similar amounts of chromium but lower nickel. Even a low Cr-containing Fe-Cr-Ni cast alloy, alloy HT (Fe-15Cr-35Ni-0.5C) was signiﬁcantly better than both Types 304 and 321 (Ref 43). These data are shown in Fig. 5.29. Small additions of some minor elements, such as titanium, niobium, tungsten, and rare earth elements, may also improve an alloy’s resistance to carburization in test environments of H2-8.6%CH4-7%H2O and H2-12%CH4-10% H2O. This is illustrated in Table 5.1 (Ref 44). The superior carburization resistance of TMA 4750 alloy to HK-40 (2% Si) can be attributed to small additions of titanium, niobium, tungsten,

Carbon concentration profiles of HK and HP alloys tested at 1050 °C (1920 °F) for 1200 h in 37%N2-40%H2-20%CO-3% CH4 (ac = 1.0, pO2 = 3:4 · 10720 atm). Source: Ref 33

pg 108

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 109

and rare earth elements. In both test environments, the oxygen potentials, although not discussed, are believed to be high enough to form oxide scales. Thus, improvements in carburization resistance may be partly due to the improved oxide scale. Silicon has also been found to be very effective in improving carburization resistance. The beneﬁcial effect of silicon on carburization resistance has been reported by Norton and Barnes (Ref 35), Steinkusch (Ref 45), Wolfe (Ref 46), and Van den Bruck and Schillmoller (Ref 47), and the data are illustrated in Fig. 5.19, and Fig. 5.30 to 5.32. Kane (Ref 48) investigated a large number of centrifugally cast tubes of HK (Fe-25Cr-20Ni) and HP (Fe-25Cr-35Ni) alloys from four producers. The tubes contained silicon varying from about 1% to more than 2%. Tests were conducted at 980 and 1090 °C (1800 and 2000 °F). A unit carbon activity (ac = 1.0) was maintained for all the test environments. Oxygen potentials of the test environments were varied by injecting different levels of H2O. With no H2O injection, the environment’s pO2 was below

that in equilibrium with SiO2 (i.e., SiO2 could not form). Thus, silicon played no role in carburization resistance. At high oxygen potentials (i.e., 1% and 10% H2O injections) where SiO2 was stable, silicon improved carburization resistance. The focus thus far has been primarily on alloys used in ethylene cracking and steam hydrocarbon reforming operations. Most of the alloys are cast alloys used for furnace tubes. A variety of wrought alloys of stainless steels, Fe-Ni-Cr

Fig. 5.18

Oxygen potentials of the test environments used by Norton and his colleagues in carburization studies at Petten Laboratories. Source: Ref 37

Fig. 5.17

Carbon concentration profiles of several centrifugally cast alloys in (a) the as-cast surface condition and (b) the machined surface condition after 1 year of field testing in an ethylene cracking furnace. Source: Ref 34

Fig. 5.19

Carburization rate constants of several Fe-Ni-Cr alloys at 825 °C (1520 °F) in the test environment with a carbon activity of 0.8 and an oxygen potential such that SiO2 is stable (but not Cr2O3), as shown in Fig. 5.18. Source: Ref 35

pg 109

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

110 / High-Temperature Corrosion and Materials Applications

Fig. 5.20

Carburization rate constants as a function of silicon content in the alloy for several Fe-Ni-Cr alloys tested at 825 °C (1520 °F) in the test environment with a carbon activity of 0.8 and an oxygen potential such that SiO2 is stable (but not Cr2O3), as shown in Fig. 5.18. Source: Ref 35

Fig. 5.22

Carburization rate constants of several Fe-Ni-Cr alloys at 1000 °C (1830 °F) in the test environment with a carbon activity of 0.8 and an oxygen potential such that SiO2 is not stable as shown in Fig. 5.18. Source: Ref 35

Carburization rate constants as a function of Ni to Cr + Fe ratio [Ni/(Cr + Fe)] for several Fe-Ni-Cr alloys tested at 1000 °C (1830 °F) in the test environment with a carbon activity of 0.8 and an oxygen potential such that SiO2 is not stable as shown in Fig. 5.18. Source: Ref 35

alloys, and Ni-Cr alloys have been widely used in various industries, including heat treating and chemical processing. Mason et al. (Ref 49) investigated various stainless steels by performing pack carburization tests. Their results are summarized in Table 5.2. Silicon was again

noted for its beneﬁcial effect, as illustrated by Type 330 (0.47% Si) versus Type 330 (1.0% Si) and Type 304 (0.39% Si) versus Type 302B (2.54% Si). Chromium was found to be beneﬁcial in Fe-Cr alloys, as shown by Type 446 (27% Cr) versus 430 (16% Cr). Small additions

Fig. 5.21

pg 110

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 111

Fig. 5.23

Effect of nickel content on the diffusion coefficient of carbon in Fe-15Cr-Ni alloys. Source: Ref 38

of titanium or niobium appeared to be beneﬁcial when comparing Type 321 and Type 347 to Type 304. Several nickel-base alloys along with HK-40 were investigated by Kane and Hosier (Ref 50). Tests were conducted in environments with unit carbon activity and various oxygen potentials. Different rankings were obtained at different oxygen potentials. Test results for two environments are summarized in Table 5.3. In the test environment of H2-12%CH4-10%H2O with 1.3 ×10−20 atm of pO2 , where SiO2 was thermodynamically stable, HK-40 (1.19% Si) performed the best. When pO2 was reduced to 1.9 × 10−24 atm in H2-0.1%C7H13OH, where SiO2 was not thermodynamically stable, and the carbon

activity was maintained at unity, alloys containing aluminum, such as alloys 601 and 617, were much more resistant to carburization than HK-40. The authors (Ref 50) attributed this to the formation of an Al2O3 scale, although alloys 601 and 617 contain only about 1.3% Al. Other investigators (Ref 24, 51) also showed beneﬁcial effect of silicon in 25Cr-30/35Ni type alloys. Nishiyama et al. (Ref 24) performed their study of carburization in a simulated ethylene pyrolysis environment (H2-15%CH4-3%CO2) for Fe-Ni-Cr alloys with high Si (1.7%) and low Si (0.3–0.5%). When tested at 1000 °C, where Cr2O3 was stable (see Fig. 5.9), chromium, not silicon, was more effective in improving carburization resistance as shown in Fig. 5.33

pg 111

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

pg 112

112 / High-Temperature Corrosion and Materials Applications

of alloy 214 compared with those of alloys X, 601, and 150 after exposure to the test environment at 980 °C (1800 °F) for 55 h (Ref 52). The

Mass change, g/m2

× alloy 602CA

100

×

50

200

0

alloy 617 ∗ alloy 601 ∗ 45TM

∗

∗

∗

0

400

(a)

Mass change, g/m2

× alloy 800H AC66

×

150

600

800

1000

1200

Time, h 350 300 250 200 150 100 50 0

× AC66 alloy 800H × × • • alloy DS • ×• • alloy 600H × × ∗ alloy 601 ∗ × ∗ 45TM ∗ ∗ alloy 602CA alloy 617

×• 0

(b)

Mass change, g/m2

(Ref 24). Alloy B (26Cr-35Ni-0.5Si) was signiﬁcantly better than alloy D (21Cr-31Ni-0.3Si), while alloy B (26Cr-35Ni-0.5Si) was similar to alloy A (25Cr-37Ni-1.8Si). At 1150 °C, where Cr2O3 oxide was not stable, high-Si alloys (alloy A: 25Cr-37Ni-1.8Si, alloy C: 32Cr-43Ni1.7Si) exhibited signiﬁcantly better carburization resistant than low-Si alloys (alloy B: 26Cr-35Ni0.5Si, alloy D: 21Cr-31Ni-0.3Si), as shown in Fig. 5.34. Aluminum is the most effective alloying element in improving an alloy’s carburization resistance at high temperatures. Lai (Ref 52) showed that when tested at 870, 930, and 980 °C (1600, 1700, and 1800 °F) alloy 214 (Ni-16Cr3Fe-4.5Al-Y) was the most carburizationresistant alloy among more than 20 commercial wrought alloys, ranging from stainless steels and Fe-Ni-Cr alloys to nickel- and cobalt-base superalloys in the test environments, which were characterized by a unit carbon activity and oxygen potentials such that Cr2O3 was not expected to form on the metal surface. Oxides of silicon, titanium, and aluminum were expected to be stable under the test conditions. The excellent carburization resistance of this alloy was attributed to the Al2O3 oxide scale formed on the metal surface. Figure 5.35 illustrates the microstructure

200

400

600

800 1000 1200 1400

Time, h

320

×

240

× ×

160 80

×

0

∗• 0

× ∗

∗

∗ • 200

•

• 400

(c)

∗ 45TM

∗

∗

• alloy DS

•

•

600

× AC66

×

alloy 601

alloy 617 alloy 602CA alloy 600H

800

1000

1200

Time, h

Fig. 5.25

Carbon pick-up after 1000 h of exposure at 1000 °C g/cm2

Results of carburization tests in H2-CH4 mixtures (ac = 0.8) at (a) 850 °C, (b) 1000 °C, and (c) 1100 °C for Fe-Ni-Cr alloys (800H, AC66, and DS) and nickel-base alloys (alloys 600H, 601, 602CA, 617, and 45TM). Source: Ref 29

400 AC66 300

•

200

alloy 600H

•

•

45TM

•

•

•

alloy 800H

•

•

alloy DS

HPM

alloy 625 alloy 601

• alloy 602CA •alloy 617

100 0

0

10

20

30

40

50

Fe, %

Fig. 5.24

Carbon solubility as a function of nickel content at different carbon activities in Fe-Ni alloy system at 1000 °C (1830 °F). Source: Ref 39

Fig. 5.26

Weight gain as a function of iron content (i.e., a function of nickel) in Fe-Ni-Cr and Ni-Cr alloys tested at 1000 °C for 1008 h in a H2-CH4 mixture (ac = 0.8 and 725:25 pO2 =10 ). Source: Ref 29

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 113

Fig. 5.27

Effect of nickel content on the carburization resistance of Fe-Ni-Cr alloys. Source: Ref 42

Fig. 5.29

Fig. 5.28

Carburization resistance of several wrought and cast Fe-Cr-Ni alloys (Type 304, 321, HT, HU, and HK) after testing in dry ethane (C2H6) for 24 h at temperatures from 880 to 1000 °C. Source: Ref 43

existence of this oxide scale was conﬁrmed by Auger analysis (Ref 52). Similar test results were also observed by Lai et al. (Ref 53) when tested in

H2-2%CH4 at 982 °C (1800 °F) for 96 h. The carbon activity for the test environment at the test temperature was greater than 1.0 (see Fig. 5.11). Their results are summarized in Fig. 5.36. Alloy

Effect of chromium on the carburization resistance of Fe-Ni-Cr alloys after testing in H2-2.5%CH4 at 1050 °C (1920 °F) for 100 h. Source: Ref 42

pg 113

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

pg 114

114 / High-Temperature Corrosion and Materials Applications

214 showed no evidence of carburization, while Type 310 and alloys 800H, 25-35NbMA, 25-35Nb, 803, and 602CA showed different degrees of carburization attack. Alloy 602CA with about 2% Al (Ni-25Cr-10Fe-2Al-Y-Zr), although not as good as alloy 214, was signiﬁcantly better than other alloys. A similar observation was also made by Kane et al. (Ref 54), showing that the alumina-forming MA956 Table 5.1 Weight gain (mg/cm2) for several cast alloys after 100 h at 1090 °C (2000 °F) in H2-CH4-H2O mixtures Weight gain, mg/cm2 Alloy(a)

HK-40 (1% Si) HK-40 (2% Si) TMA-4750 HP-45 TMA-6350

H2-8.6CH4-7H2O

H2-12CH4-10H2O

25.0 16.8 2.0 19.0 3.8

21.8 10.2 1.0 4.3 2.3

Table 5.2 Results of pack carburization tests at 980 °C (1800 °F)(a) for various stainless steels Nominal composition

Si content, %

Increase in C content(b), %

Fe-21Cr-34Ni Fe-15Cr-35Ni Fe-15Cr-35Ni-Si Fe-25Cr-20Ni Fe-25Cr-20Ni-Si Fe-25Cr-12Ni Fe-18Cr-8Ni-Nb Fe-18Cr-8Ni-Ti Fe-18Cr-8Ni Fe-18Cr-8Ni-Si Fe-28Cr Fe-16Cr

0.34 0.47 1.00 0.38 2.25 0.25 0.74 0.49 0.39 2.54 0.34 0.36

0.04 0.23 0.08 0.02 0.03 0.12 0.57 0.59 1.40 0.22 0.07 1.03

Alloy

800 330 330 310 314 309 347 321 304 302B 446 430

(a) 40 cycles of 25 h each cycle at 980 °C (1800 °F). Carburizer was renewed after each cycle. (b) Bulk analysis. Source: Ref 49

(a) HK-40 (1% Si): 0.43C-0.60Mn-0.96Si-25.4Cr-20.7Ni. HK-40 (2% Si): 0.41C0.60Mn-1.98Si-25.0Cr-20.7Ni. TMA-4750: 0.44C-0.69Mn-1.99Si-24.9Cr20.8Ni-0.11Ti-0.29Nb-0.30W-REM. HP-45: 0.51C-0.54Mn-1.65Si-25.5Cr36.1Ni. TMA-6350: 0.50C-0.70Mn-1.84Si-25.1Cr-38.4Ni-0.13Ti-0.28Nb-0.27WREM. REM denotes rare earth metals. Source: Ref 44

Fig. 5.31

Effect of silicon on the carburization resistance of cast Fe-20Ni-Cr alloys tested at 1090 °C (2000 °F) for 24 h in wet ethane (C2H6). Source: Ref 46

Fig. 5.30 Ref 45

Effect of silicon on the carburization resistance of 25Cr-20Ni and 35Cr-25Ni-Nb alloys. Source:

Fig. 5.32

Effect of silicon on the carburization resistance of HK-40 with different silicon levels tested at 1100 °C (2010 °F) for 520 h in carbon granulate (pack carburization test). Source: Ref 47

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 115

alloy (Fe-20Cr-4.5Al-0.5Y2O3) performed signiﬁcantly better than alloys 601, 800H, Type 310, and HK alloy (Table 5.4). Aluminum, which has been used for alloy addition to form external Al2O3 scale in ironand nickel-base wrought alloys, has been used in centrifugally cast alloys for oxidation or carburization resistance. Recently a commercial, centrifugally cast nickel-base alumina-forming alloy, alloy 60HT, containing approximately 25% Cr, 11% Fe, 0.4% C, and Al, was developed (Ref 55). In the paper published by Kirchheiner et al. (Ref 55), no carburization data were reported. However, the resistance to coking was studied on alloy 60HT containing three levels of aluminum (i.e., 2.35, 3.55, and 4.81%). They found signiﬁcant reduction in coking rates for the samples containing 3.55 and 4.81% Al, with the sample containing 2.35% Al exhibiting only slight reduction of coking rates compared with the conventional HP-40 alloy (4852). Coking, which is an important phenomenon in ethylene cracking, develops on the internal surface of the pyrolysis tube and reduces the heat transfer. The ethylene cracking operation has to be interrupted Table 5.3 Weight gain (mg/cm2) for several Fe-Ni-Cr and Ni-base alloys after ten 24 h cycles at 1100 °C (2010 °F) in H2-12%CH4-10%H2O and H2-0.1%C7H13OH environments Weight gain, mg/cm2 Alloy

HK-40 (1.19% Si) 800H 600 617 601 690

H2-12CH4-10H2O(a)

H2-0.1C7H13OH(b)

6.97 23.46 12.00 16.84 22.74 25.54

54.38 46.60 4.85 0.50 1.81 54.38

(a) Inlet gas mixture: ac = 1.0, pO2 ¼ 1:3 · 10720 atm at 1100 °C. (b) Inlet gas mixture: ac = 1.0, pO2 =1:9 · 10724 atm at 1100 °C. Source: Ref 50

Fig. 5.33

Carbon profiles of high-Si (HSi) and low-Si (LSi) Fe-Ni-Cr alloys after testing at 1000 °C (1832 °F) for 96 h in H2-15%CH4-3%CO. Source: Ref 24

to allow decoking, typically with steam and air to remove the coked layer. This decoking operation can have detrimental effects on the tube properties, thus reducing the tube’s service life. Other factors such as surface ﬁnish have been found to be very important in affecting carburization reactions. Machining the metal surface to improve the surface ﬁnish can signiﬁcantly increase an alloy’s carburization resistance. It is common practice to bore or hone the internal diameter of a centrifugally cast tube to remove surface shrinkage pores. Figures 5.17, 5.37, and 5.38 illustrate the signiﬁcant improvement in carburization resistance as a result of surface machining (Ref 34, 56). A cast metal surface with shrinkage pores can generate stagnant conditions in crevices, which are very conducive to carburization attack. In addition, a machined surface exhibits a cold-worked layer, which tends to accelerate the diffusion process and results in rapid formation of oxide scale or ﬁlm, thus slowing subsequent carbon ingress. Compared to the machined or ground surface, the electropolished surface exhibited accelerated carburization (Fig. 5.39). Norton and Barnes (Ref 35) considered the electropolished surface a work-free surface that failed to develop a surface oxide ﬁlm as readily as the cold-worked ground surface. Not mentioned by Norton and Barnes in their paper (Ref 35), electropolishing can produce a surface layer that may be depleted in chromium, and the surface depletion of chromium may result in the observed accelerated carburization. In their study of oxidation/ carburization of a high-temperature gas-cooled reactor (HTGR) helium environment containing parts per million (ppm) levels of H2, CO, CO2, CH4, and H2O, Mazandarany and Lai (Ref 57) observed that Type 316SS specimen (obtained

Fig. 5.34

Carbon profiles of high-Si (HSi) and low-Si (LSi) Fe-Ni-Cr alloys after testing at 1150 °C (2100 °F) for 48 h in H2-15%CH4-3%CO. Source: Ref 24

pg 115

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:22PM Plate # 0

116 / High-Temperature Corrosion and Materials Applications

from sheet material) with the as-received surface condition suffered severe carburization after testing at 649 °C (1200 °F) for 5000 h in He-1500 µatm H2-450 µatm CO-50 µatm CH4-50 µatm H2O. The 316 specimen with the as-ground specimen, on the other hand, showed no evidence of carburization tested in the same retort under the same test condition. This observation was supported by both microstructure and microhardness proﬁle results, as shown in Fig. 5.40 (Ref 57). The oxygen potentials of the test environment were reported to be high enough to form chromium oxides and too low to form iron and nickel oxides. The unusual

Fig. 5.35

observation for the 316 specimen with the asreceived surface condition after testing at 649 °C (1200 °F) in He-1500 µatm H2-450 µatm CO-50 µatm CH4-50 µatm H2O was the formation of an Fe-Ni metallic scale on the specimen surface and internal oxides underneath the metallic scale. This is shown in Fig. 5.41 (Ref 57). The 316SS specimen with the as-ground surface condition after testing under the same condition showed surface oxide scales with no outer metallic scale and internal oxides. The authors thus believed that the initial surface of the 316SS sheet was depleted in chromium at a signiﬁcant degree, such that not enough chromium was available

Optical micrographs showing the microstructures of (a) alumina-former alloy 214 (Ni-16Cr-3Fe-4.5Al-Y), and several chromia-former alloys (b) 601 (Ni-23Cr-14Fe-1.4Al), (c) X (Ni-22Cr-18Fe-9Mo), and (d) 150 (Co-27Cr-18Fe) after testing at 980 °C (1800 °F) for 55 h in Ar-5%H2-5%CO-5%CH4 (ac = 1.0,pO2 =9 · 10722 atm). Source: Ref 52

pg 116

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 117

Chapter 5: Carburization and Metal Dusting / 117

Fig. 5.36

Weight gain as a function of cycles (24 h cycle) tested at 982 °C (1800 °F) for 96 h with 24 h cycles in H2-2%CH4 for Type 310SS, 800H (N08810), 25-35Nb (25Cr-35Ni-1.3Nb-0.4C-2.0max Si), 25-35NbMA (microalloyed, 25Cr-35Ni-1.1Nb0.4C-1.4Si, Ti, REM), 803 (S35045: Fe-25Cr-35Ni), 602CA (N06025: Ni-25Cr-10Fe-2.1Al-Y-Zr), and 214 (N07214: Ni-16Cr-3Fe-4.5AlY). Note: All four alloy 800H (N08810) specimens were control samples. Source: Ref 53

Table 5.4 Carburization resistance of Fe-Ni-Cr, Ni-base, and ODS alloys in H2-2%CH4 at 1000 °C (1830 °F) for 100 h Alloy

MA956 601 800 310SS HK (1.07% Si) HK (2.54% Si)

Weight gain, mg/cm2

<0.3 10 19 36 34 29

Source: Ref 54

Fig. 5.37

Comparative carburization resistance of as-cast and machined tubes of alloys 30Cr-30Ni, 36X (Fe-0.4C25Cr-34Ni-1.2Nb), and 36XS (Fe-0.4C-25Cr-34Ni-1.5Nb-1.5W) after 3 years of field testing in an ethylene pyrolysis furnace. Source: Ref 34

to form a continuous external chromium oxide scale. Then, the nucleation of internal chromiumrich oxides resulted in a chromium-depleted

metal on the metal surface. When the surface of the test specimen from the same sheet material was ground in specimen preparation, the chromium-depleted surface layer was removed. When tested under the same condition, this ground specimen formed an external chromium oxide scale on the metal surface without the formation of an Fe-Ni metallic surface layer. The Fe-Ni metallic layer formed on the as-received surface of the 316SS specimen provided rapid absorption of carbon and a rapid diffusion path of carbon in the metal, resulting in signiﬁcant carburization. When tested at 870 °C (1600 °F) in the same environment, an external chromium oxide scale was observed to form on the

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

118 / High-Temperature Corrosion and Materials Applications

as-received surface with no evidence of carburization attack. No explanation was offered by the authors on this. It is believed that higher temperatures caused a rapid homogenization of the chromium concentration at the initially depleted surface layer due to faster diffusion rate. Since this type of environment contains only ppm levels of corrosive gaseous components, the reaction kinetics are believed to be relatively slow. This slow reaction kinetics allowed the homogenization of the chromium concentration at the metal surface when tested at 870 °C (1600 °F). Injecting sulfur compounds into the carburization gas stream can signiﬁcantly reduce the carburization kinetics. Sulfur compounds, such as H2S, can retard carburization kinetics. This is illustrated by the test results shown in Fig. 5.42, obtained by Ramanarayanan et al. (Ref 58). Beneﬁcial effect of sulfur in reducing carburization attack was demonstrated by Norton and Barnes (Ref 35) in their laboratory tests. Figure 5.43 shows that carburization of HK-40 alloy was signiﬁcantly reduced when 100 ppm H2S was injected into the test environment. The effect of different levels of H2S injection on the carburization of HK-40 is illustrated in Fig. 5.44.

Figure 5.45 shows the carburization of alloy 800H in H2-CH4 (ac=1.0) with injection of different levels of H2S (H2S/H2) (Ref 59). When this approach is used for controlling carburization, it is important to determine the optimum level of sulfur compounds needed to achieve maximum effectiveness. Too much can lead to accelerated corrosion due to sulﬁdation, which may be worse than carburization. Figure 5.46 shows that a low H2S injection failed to reduce carburization and a high H2S injection caused sulﬁdation attack for alloy 800 (Ref 59). It has been proposed by Grabke et al. (Ref 60) and Ramanarayanan and Srolovitz (Ref 61) that sulfur absorbed on the metal surface blocks the potential sites for carbon absorption from the carburizing gas, thus signiﬁcantly reducing the carbon concentration in the metal surface layer and retarding the overall carbon transfer into the metal. 5.2.3 Effect of Carburization on Mechanical Properties Carburization results in formation of internal carbides in the alloy. The room-temperature ductility or toughness can be severely degraded

Fig. 5.39 Fig. 5.38

Effect of machining on the carburization resistance of cast HK-40 alloy. Source: Ref 56

Detrimental effect of the electropolished surface condition on the carburization resistance of HK-40 alloy tested at 825 °C (1520 °F) in an environment with 0.8 ac and the oxygen potential shown in Fig. 5.18. Source: Ref 35

pg 118

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 119

Chapter 5: Carburization and Metal Dusting / 119

Oxalic etch Exposed at 649 °C (1200 °F) for 5000 h, as-received surface

(649 °C)1200 °F As-received surface 500

Microhardness, DPH

400

300 649 °C(1200 °F) and 760 °C(1400 °F) Ground surface

Oxalic etch Exposed at 649 °C (1200 °F) for 5000 h, ground surface

200 871 °C(1600 °F)

100

0 0

100

200

300

400

500

600

700

Distance from exposed surface, µm

Fig. 5.40

Microhardness profile and optical micrograph showing severe carburization attack on the 316 specimen with the original as-received surface (solid circle data point) after testing at 649 °C (1200 °F) for 5000 h in He-1500 µatm H2-450 µatm CO-50 µatm CH4-50 µatm H2O. Also shown are microhardness profile and optical micrograph of the 316 specimen with the as-ground surface condition (solid square data point) after testing under the same condition showing no sign of carburization attack. The specimens with the ground surface condition showed no carburization attack after testing at a higher temperature (i.e., 760 °C) in the same environment for the same test duration. The specimen with the as-received surface condition showed no evidence of carburization attack when tested at 871 °C (1600 °F). Source: Ref 57

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 120

120 / High-Temperature Corrosion and Materials Applications

when the alloy is heavily carburized (Ref 29). For high-temperature components that are subjected to severe carburizing environments at very high temperatures, such as ethylene pyrolysis furnace tubes, the component can be carburized through the thickness of the component with signiﬁcant amount of carbon pickup. Grabke and Jakobi

(Ref 62) performed a metallurgical analysis of a failed ethylene pyrolysis furnace tube made of HP40Nb and found the tube was carburized 4 T = 1000 °C

3.5 3

H2S Off 2.5 H2S On

2 Carburization

1.5 1

Cr2O3 Growth

Cr2O3 to Cr7C3 Conversion

0.5 0

0

20

40

60

80

100

120

Fig. 5.42

Fig. 5.41

As-received surface 316 specimen after testing at 649 °C (1200 °F) for 2000 h in He-1500 µatm H2450 µatm CO-50 µatm CH4-50 µatm H2O. Area 1: Fe-Ni metallic phase, Area 2: Cr-Mn-Si oxides, and Area 3: Fe-Ni metallic phase. Unetched condition. Source: Ref 57

Effect of sulfur on the carburization kinetics of Ni30Cr alloy. Vertical axis is mass gain per unit area, and horizontal axis is exposure time in hour. The test began with oxidation at 1000 °C (1832 °F) in a CO-CO2 environment for about 40 h, and then switched to carburization in a H2-CH4 mixture with a carbon activity of 0.9, thus prompting the conversion of oxide scale to carbides and a rapid carburization. This was then followed by injecting 100 ppm of H2S into the carburizing environment, thus immediately stopping carburization reaction until the injection of H2S was turned off, which set off a rapid carburization again. Source: Ref 58

Fig. 5.43

Carbon concentration profiles for HK-40 alloy after testing at 1000 °C (1830 °F) for 100 h in a carburizing environment (ac = 0.8, pO2 is shown in Fig. 5.18) with and without injection of 100 ppm H2S. The specimen was a flat coupon with both sides being exposed to the carburizing gas. Source: Ref 35

Fig. 5.44

Effect of H2S on the carburization behavior of HK-40 alloy. Source: Ref 35

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 121

throughout the tube wall with 3.14 to 3.3% C, as opposed to the original carbon content of about 0.5%. Furthermore, carburization can increase

the volume of the carburized zone, thus developing internal stresses. When the alloy is severely carburized, these internal stresses can be

Fig. 5.45

Weight gain of alloy 800 after 100 h in H2-CH4-H2S environment (ac = 1.0) at 900, 1000, and 1100 °C (1652, 1832, and 2012 °F) for different H2S/H2 ratios. Source: Ref 59

Fig. 5.46

Effect of H2S on the carburization behavior of alloy 800 in H2-CH4-H2S environments (ac = 1.0) at 1000 °C (1832 °F) for different H2S/H2 ratios. Also shown are corresponding microstructures of the tested specimens. Source: Ref 59

pg 121

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

122 / High-Temperature Corrosion and Materials Applications

signiﬁcant enough to develop cracking when the carburized alloy is cooled to room temperature. This is illustrated in Fig. 5.47 (Ref 25). When the alloy is in the high-temperature creep range, a thoroughly carburized alloy can retain reasonable ductility at elevated temperatures. In fact, a thoroughly carburized Fe-Ni-Cr alloy can exhibit a better creep ductility than the same alloy tested in air. Guttmann and Schonherr (Ref 63) studied the creep-rupture behavior of HK alloys in a heavily precarburized condition and found that the creep ductility was signiﬁcantly improved by carburization when creep tested at 1000 °C (1832 °F). In their study, they precarburized both HK-40 and HK-30 specimens at 1000 °C (1832 °F) in H2-CH4 to a complete saturation of M7C3 carbides (about 55 vol%) in the material with about 4% C. The creep-rupture testing was conducted at 1000 °C (1832 °F) in H2-1%CH4 (ac = 0.8) to avoid decarburization during testing. The rupture ductility of precarburized specimens was compared with that of HK-40 and HK-30 in the as-cast condition tested in air at the same temperature, as shown in Fig. 5.48. The authors also made a similar observation in comparing the rupture ductility when tested in H2-1%CH4 (ac = 0.8) at 1000 °C (1832 °F) (i.e., specimens were not precarburized prior to testing) with that tested in air, as shown in Fig. 5.49. Carburization-accelerated creep deformation at high temperatures is clearly illustrated by the two comparison creep curves tested at 1000 °C (1832 °F) in H2-1%CH4 (ac = 0.8) and air reported by Guttmann and Schonherr (Ref 63), as shown in Fig. 5.50. Because carburization accelerates creep deformation, creep strength is signiﬁcantly reduced. Carburization

as the result of either testing in a carburization environment or precarburization can signiﬁcantly reduce the 1% creep strengths of both HK-40 and HK-30. The authors attributed this to both the internal stresses developed due to

Fig. 5.47

Fig. 5.49

Cracking developed in alloy 800 due to internal stresses resulting from heavy carburization in the test coupon with no external loading during the carburization test. Source: Ref 25

Fig. 5.48

Elongation to fracture as a function of rupture life of HK-40 and HK-30 comparing the as-cast specimens tested in air and the precarburized (thoroughly carburized) specimens tested in H2-1%CH4 (ac = 0.8) (to avoid decarburization) after creep-rupture testing at 1000 °C (1832 °F). Source: Ref 63

Elongation to fracture as a function of rupture life of HK-40 and HK-30 comparing the data from tests in the carburizing environment (i.e., H2-1%CH4 [ac = 0.8]) and that from air tests at 1000 °C (1832 °F). Source: Ref 63

pg 122

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 123

volume increase in carburization and carbide coarsening during creep testing at high temperatures. This is illustrated in Fig. 5.51. Ethylene pyrolysis furnace tubes that are subjected to severe carburization during service at high temperatures typically experience signiﬁcant creep deformation as a result of carburization. Accordingly, the tube suffers stretching due to creep elongation, and frequent shortening of tube coils is required in order to avoid the tube from reaching to the furnace ﬂoor to cause tube damages (Ref 64). Jakobi and Gommans (Ref 64) also indicated brittle fracture was one of the failure modes for centrifugally cast pyrolysis furnace tubes. The brittle fracture was primarily due to the furnace trip that caused a large temperature drop of 500 to 1000 °C (932 to 1832 °F). Such a furnace trip can produce a strain of 0.75 to 1.5%, which is equivalent to about the rupture ductility of aged, carburized, and nitrided cast furnace tubes at room temperature to about 600 °C (1112 °F) (Ref 64). Thus, the furnace tubes suffer brittle fracture at low temperatures due to the large temperature drop. Taylor et al. (Ref 65) studied the effect of carburization on the creep-rupture behavior of alloy 800H at 800 °C (1472 °F) and found that carburization increased rupture strength and life. They compared the creep-rupture data of precarburized specimens (partially carburized specimens and fully carburized specimens) with that of as-received specimens as well as pre-aged specimens. The data for precarburized specimens were tested in a H2-CH4 environment (ac > 0.5) to avoid decarburization during testing, while asreceived and pre-aged specimens were tested in

air. Precarburization was carried out at 1000 °C (1832 °F) for 1200 h in a H2-CH4 mixture with ac being 0.8, which resulted in a fully carburized condition containing only M7C3 (about 30 vol %). Partially carburized specimens were carburized at 1000 °C (1832 °F) for 100 h in a H2-CH4 mixture with ac being 0.8, which produced a carburized depth of about 1 mm (equivalent to 56% of cross-sectional area being carburized). Pre-aging treatment consisted of heating the specimen at 1000 °C (1832 °F) for 100 h in air to produce a comparable structure to the uncarburized core of the partially carburized specimens. Their creep-rupture data are summarized in Fig. 5.52. Fully carburized specimens were found to exhibit signiﬁcantly higher creeprupture strengths than the as-received specimens tested in air. Partially carburized or pre-aged specimens showed strengths comparable to those of the as-received specimens. As for rupture ductility, the fully carburized specimens showed lower rupture ductility for shorter rupture times (at high stresses), but comparable to those of as-received specimens and pre-aged as well as partially carburized specimens for longer rupture times (i.e., at lower stresses). This is illustrated in Fig. 5.53. The authors also investigated the effect of the temperature on creep-rupture ductility of alloy 800H that was fully carburized. The alloy’s rupture ductility (when fully carburized) was found to decrease signiﬁcantly with decreasing temperature. The carburized specimen failed in a completely brittle manner on loading when tested at 600 °C (1112 °F). This is illustrated in Fig. 5.54.

Fig. 5.50

Fig. 5.51

Comparative creep curves of HK-40 tested at 1000 °C (1832 °F) and 15 MPa in air and H2-1% CH4 (ac = 0.8). Source: Ref 63

1% creep strengths of HK-40 and HK-30 tested at 1000 °C (1832 °F) in air, H2-1%CH4 (ac = 0.8), and for precarburized specimens tested in H2-1%CH4. Source: Ref 63

pg 123

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

124 / High-Temperature Corrosion and Materials Applications

5.3 Metal Dusting 5.3.1 Metal Dusting—Thermodynamic Considerations Metal dusting occurs in carburizing atmospheres at intermediate temperatures (approximately between 430 and 900 °C (800 and 1650 °F) with the maximum rate of attack occurring at about 600 to 700 °C (1112 to 1292 °F), depending on alloys and environments. It is now well understood that metal dusting occurs in environments that exhibit highcarbon activities (i.e., ac > 1) (Ref 66–75). As discussed in section 5.2.1, three chemical

Fig. 5.52

Stress rupture data of precarburized (fully carburized and partially carburized) specimens tested at 800 °C (1472 °F) in a carburizing environment (to prevent decarburization) is compared with that of as-received specimens and pre-aged specimens tested in air at the same temperature. Source: Ref 65

Fig. 5.53

Comparison rupture ductility of alloy 800H tested at 800 °C (1472 °F) for fully and partially carburized specimens in comparison with as-received and pre-aged specimens. Source: Ref 65

reactions (Eq 5.1–5.3) that can cause carburization can also cause metal dusting attack. Metal dusting most often occurs in syngas environments, which consist of primarily H2 and CO. In the earlier discussion of Eq 5.1: H2 + CO = C + H2O, where graphical presentation of the chemical reaction is presented as a function of carbon activity (ac), temperature, and the gas composition in terms of (pCO pH2 = pH2 O ), as shown in Fig. 5.1. Figure 5.1 illustrates that the carbon activity of the environment decreases with increasing temperature. Thus, at high temperatures, many gaseous compositions are not thermodynamically favored for causing metal dusting. Conversely, lower temperatures with increasing carbon activity are thermodynamically favored for causing metal dusting. This is why metal dusting attack diminishes as the temperature increases after passing the peak reaction temperature due to decreasing carbon activity. For example, for the gas mixture with (pCO pH2 =pH2 O ) ratio being 1.0, its carbon activity varies from more than 102 to approximately 10−1 as the temperature increases from 400 to 800 °C (752 to 1472 °F). Metal dusting attack diminishes as the temperature decreases after dropping below the peak temperature. This is due to the reaction’s kinetics that decreases with decreasing temperature. The reaction Eq 5.2, as graphically presented in Fig. 5.2, shows a similar fashion. When the environment consists of CO and CO2, metal dusting attack follows that involves H2 and CO. However, when the environment consists of H2, CO, and CO2, the Boudouard reaction (CO-CO2

Fig. 5.54

Effect of temperature on rupture ductility of fully carburized alloy 800H tested at different temperatures with stresses to cause rupture in 200 h. Source: Ref 65

pg 124

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 125

reaction) is believed to be slower in reaction kinetics than that of Eq 5.1, thus making Eq 5.1 to be the dominant reaction for the metal dusting (Ref 76, 77). The reaction 5.3 (Eq 5.3), on the other hand, increases the carbon activity with increasing temperature. The graphic presentation of this reaction is shown in Fig. 5.3. The carbon activity is generally too low for many environments containing CH4 where metal dusting occurs at 600 or 700 °C (1112 or 1292 °F). The presence of CH4 in the gas mixture containing CO and CO2 may help bring up the carbon activity at higher temperatures, thus increasing the upper temperature limit at which metal dusting can take place.

ac (gas) CO + H2 H2O + C (diss.) ac″ (Fe3C1Fe)

ac=1

(a)

Fe3C

C (diss.) + Fe

CO + H2 H2O + C (in Fe3C) (b)

CO + H2 H2O + C (graphite)

5.3.2 Mechanisms of Metal Dusting Attack Signiﬁcant advancement in understanding the mechanisms of metal dusting has been achieved in the past decade, with most of the contribution being made by Professor Grabke and his group (Ref 68–75). Two models have been proposed with one model explaining the metal dusting behavior of iron and low-alloy steels and the other model explaining the behavior of Crcontaining high-temperature alloys (i.e., Fe-Cr, Fe-Ni-Cr, and Ni-Cr alloys). These two models, which are summarized by Grabke (Ref 73), are brieﬂy described here. The model for iron and low-alloy steels is summarized in Fig. 5.55. Based on the model, the prerequisite for metal dusting to occur is a high carbon activity environment (ac > 1) that causes the oversaturated metal with the carbon activity being more than 1. This condition promotes the formation of cementite (Fe3C) on the metal surface. As shown in Fig. 5.1, the calculated carbon activities in equilibrium with cementite are higher than 1.0 at 700 °C (1292 °F) and lower temperatures. Figure 5.56 shows a metastable Fe-C-O phase diagram at 600 °C indicating that metastable cementite forms at ac > 1 (Ref 78). Once graphite deposits on the cementite with the carbon activity of the graphite at 1, cementite in contact with graphite becomes unstable and then decomposes into iron and carbon with iron migrating outward into the graphite layer. As a result, metal wastage takes place as a repeated reaction of formation of metastable cementite and decomposition of cementite into iron and carbon. For iron and low-alloy steels with no protective oxide scales, metal wastage attack in general follows uniform

(c)

CO + H2 H2O + C (graphite)

(d)

3 Fe + C

Fe3C

Fig. 5.55

Schematic showing mechanism of metal dusting for iron and low alloy steels with the following steps: (a) The metal is oversaturated with carbon (ac > 1) due to carbon transfer from a high carbon activity environment (ac > 1) to the metal, (b) thus resulting in the formation of cementite (Fe3C) at the surface, (c) and later the formation of graphite on top of cementite and then decreases in ac to one (ac = 1 for graphite) at the cementite in contact with graphite, (d) thus, cementite becomes unstable and decomposes into iron and carbon, with iron migrating into graphite layer to form iron particles (embedded in graphite), which act as catalysts for more carbon deposition and coking (d). Source: Ref 73

Fig. 5.56

Fe-C-O metastable phase stability diagram showing that metastable cementite (Fe3C) forms under high carbon activities (ac > 1) at 700 °C and lower. Source: Ref 78

pg 125

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 126

126 / High-Temperature Corrosion and Materials Applications

metal thinning (Ref 73). A layer of cementite (Fe3C) was found to occur on 1Cr-0.5Mo steel after exposure in metal dusting conditions with a layer of coke on top of the cementite (Ref 71). For chromia formers in Fe-Cr, Fe-Ni-Cr, and Ni-Cr alloys, a different mechanism was proposed, which was summarized in Fig. 5.57 (Ref 73). These high-chromium alloys form a protective chromium oxide scale. Metal dusting attack is initiated when the oxide scale develops local defects, allowing carbon transfer from a high carbon activity environment (ac > 1) to metal causing oversaturation of carbon, and subsequent formation of graphite after formation of metastable M3C carbides in low nickel alloys

Cr2O3 alloy

5.3.3 Alloy Resistance to Metal Dusting Graphite

(a) (d)

C

Internal carbides

3M+C

and C

M3C

'Coke'

(b)

(e) C

(or direct graphite formation in high nickel alloys), thus resulting in lower carbon activity (ac = 1) and decomposition of metastable carbides into metal particles and carbon. As a result, the attack starts locally and often leads to the formation of hemispherical pits (Ref 73). This type of morphology is shown in Fig. 5.58. Pippel et al. (Ref 79) found graphite on cementite that formed on Fe-5Ni alloy, when exposed to metal dusting conditions. For Fe-10Ni and Fe-25Ni to Fe-30Ni alloys, graphite was found to grow into metal with no metastable carbide formed on the alloy (Ref 79). Graphite was also observed to grow into metal directly in high-nickel Fe-Ni alloys, such as Fe-40Ni, Fe-50Ni, and Fe-80Ni (Ref 72).

Process equipment failures due to metal dusting were reported in the reﬁnery industry during the 1950s. Eberle and Wylie (Ref 80) reported metal wastage of uncooled components, such as soot blower elements, made of Types 347SS and 310SS in the waste heat boiler of a synthesis gas reactor. The synthesis gas, predominantly CO and H2 with some water vapor and carbon particles, was produced by combustion of methane with oxygen. The wastage took place at temperatures between 480 and 900 °C (900 and 1650 °F). Both 347SS and 310SS suffered severe metal wastage after only 3 weeks of service. Prange (Ref 8) reported that tubes containing a chromia-alumina catalyst at a temperature of about 590 °C (1100 °F) in a butane dehydrogenation system, where butane was converted to butene, suffered metal loss problems. The oxide

Metastable carbide

(c)

Fig. 5.57

Mechanism of metal dusting for chromia formers with the following steps: (a) Development of local defects in the oxide scale, allowing carbon transfer from the environment to the metal, (b) carbon ingress into the metal resulting in the formation of stable carbides, M23C6 and M7C3, (c) continued carbon ingress increases the carbon activity to more than one (ac > 1), resulting in the formation of metastable M3C carbides in low-nickel alloys, or resulting in direct growth of graphite in high-nickel alloys, (d) graphite deposition occurs decreasing the carbon activity to one, thus, causing the decomposition of metastable M3C into metal particles and carbon with metal particles catalyzing more coking (e). Source: Ref 73

1cm

Fig. 5.58

Metal dusting attack in alloy 800 in synthesis gas environment at 550 °C (1022 °F). Source: Ref 75

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 127

Chapter 5: Carburization and Metal Dusting / 127

dusts resulting from metal dusting contaminated the catalyst and caused undesirable side effects for the operation. The alloys reported to perform poorly were 12Cr steel, 18Cr-11Ni, and alloy 600. The alloys that performed well included Fe-27Cr, 25Cr-20Ni, and Type 302B (18Cr-8Ni with 2.4% Si). Severe metal loss in the form of pitting was also observed by Hoyt and Caughey (Ref 81) for Type 310SS equipment exposed to a gas mixture rich in H2 and CO at temperatures of 650 to 700 °C (1200 to 1300 °F) in a plant that converted coal to gasoline and other products. Metal dusting problems were also encountered in reforming plants (Ref 5), where a synthesis gas (i.e., H2 + CO) for methanol manufacturing was produced. Type 304SS and 310SS reformer outlet tubes were perforated by severe pitting attack at 650 to 725 °C (1200 to 1340 °F). Hydrodealkylation units (Ref 5), acetic acid cracking furnaces (Ref 5), and coal gasiﬁcation plants (Ref 6) were reported to suffer metal dusting problems. Perkins et al. (Ref 6) reported pitting attack of alloy 800 tubes in a preheater for gasiﬁer recycle gas rich in H2 and CO with some H2O. The tubes, with a wall thickness of 0.38 cm (0.15 in.), were perforated in a few thousand hours. The attack occurred at 540 to 870 °C (1000 to 1600 °F). Dunmore (Ref 82) reported a failure of the waste heat boiler in an ammonia plant. The alloy 800 exit ferrules suffered severe metal wastage, with large uniform circular pits penetrating through the tubes and black sooty deposits on the pitted surface. The attack occurred at a location where the metal temperature was about 600 °C (1110 °F). The environment was highly enriched in H2 and CO, along with large quantities of steam. This was unusual in that steam was considered an effective additive to mitigate the metal dusting problem (Ref 7). Grabke and Spiegel (Ref 83) discussed several industrial failure cases that were caused by metal dusting. These cases involved an alloy 800 heat exchanger for synthesis gas production, HK-40 gas heaters in a direct iron reduction plant, 5Cr and 9Cr steels in catalyst regeneration units in a reﬁnery, and an alloy 601 heat exchanger in an ammonia plant. Metal dusting has also been encountered in the heat treating industry (Ref 84). Refractory anchors, fan housing assemblies, and other components in carburizing furnaces frequently suffer metal dusting problems. Alloys typically used include stainless steels, such as Type 310, nickel-base alloys, such as alloys X and 333, and iron-base alloys, such as Multimet alloy

(or N-155). Metal dusting typically occurred at temperatures between 540 and 820 °C (1000 and 1500 °F). Severe attack frequently took place after 1 year of service and at locations where the gaseous environment became stagnant. Favorite locations included the interface with refractories, where small gaps or “dead” spaces were created. Figure 5.59 (Ref 84) shows a sample of Multimet alloy obtained from a furnace fan housing in a carburizing furnace. The fan housing suffered metal dusting at the metal

1

2

3

4

Cm

Fig. 5.59

Multimet alloy fan box suffering metal dusting attack in a carburizing furnace. (a) General view of failed the sample. Note the perforated edge of the fan box. (b) Cross section of the sample showing pitting attack and severe metal thinning. (c) Severe carburization attack beneath the pitted and wasted area. Attack was initiated in the “dead” spaces created by the refractory and the fan box. Source: Ref 84

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

128 / High-Temperature Corrosion and Materials Applications

surface that was in contact with the furnace refractory lining. The 1.9 cm (0.75 in.) thick fan housing was perforated in about a year. The nearby metal surface that was exposed to the circulating carburizing gas, on the other hand, showed no sign of metal dusting (Fig. 5.60). A chromium-rich oxide scale was observed on the metal surface. One common feature associated with metal dusting is carburization beneath the pitted area (Fig. 5.59). Manufacturing of carbon ﬁbers for carbon composite materials can also generate high temperature carburizing environments for causing metal dusting problems to some furnace components. Carbon ﬁbers are manufactured from polyacrylonitrile (PAN) (Ref 85). The last step of manufacturing, “carbonization,” is carried out at about 900 °C (1650 °F) in an environment enriched with CO, CO2, CH4, N2, HCN, H2O, and so forth. Furnace components made of nickel- and iron-base alloys were found to suffer general metal thinning and pitting attack. An example is shown in Fig. 5.61. The general characteristics of metal dusting observed in the industrial environments were reproduced in many laboratory tests involving primarily gas mixtures of H2-CO-H2O. Accordingly, laboratory testing has become an important tool for generating relevant data to allow performance ranking among engineering alloys. Laboratory data generated by various laboratories are summarized below. Maier and Norton (Ref 86) investigated 9Cr-1Mo steel (P91), 12Cr steel (Type 410),

Fig. 5.60

Oxide scale and no evidence of metal dusting on the surface of a Multimet alloy fan box exposed to flowing carburizing gas (same fan box as that shown in Fig. 5.59). The sample was plated with nickel before mounting for metallographic examination.

Model alloy (Type 410 + 2.75Si), and Type 310 (Fe-25Cr-20Ni) in H2-24.4CO-2.4H2O at 560 °C (ac > 1). The specimens were annealed in H2 at 1000 °C (1832 °F) for 1 h to eliminate surface deformation structure and produced grain coarsening. The test specimens were ground. Their results are summarized in Fig. 5.62. In this short-term test, alloys with low chromium contents (9Cr for P91 and 12Cr for 410SS) showed metal dusting attack. Type 310SS (25Cr-20Ni) showed no sign of metal dusting attack up to 200 h. Silicon was found to be a very effective alloying element in improving the resistance to metal dusting. This is illustrated by the data comparing “model” alloy (410SS + 2.75Si) with 410SS. Grabke et al. (Ref 69) investigated commercial alloys, which included Fe-Cr, Fe-Ni-Cr, Ni-base alloys, and silicon-containing alloys, at 650 °C (1200 °F) for 7 days in H2-24.7CO-1.9H2O (ac = 15). All specimens were annealed in dry hydrogen at 1000 °C (1832 °F) for 1 h to achieve large grains, followed by grinding and polishing. All test specimens were in as-ground surface conditions. The data are summarized in Fig. 5.63. The

Fig. 5.61

Type 310SS furnace component suffering metal dusting in a furnace used for manufacturing carbon fibers. (a) General view of the failed component. (b) Cross section of the sample showing pitting and thinning

pg 128

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 129

Chapter 5: Carburization and Metal Dusting / 129

hatched data show mass gain of coke deposits on the test specimen, while the solid data represent the either weight loss due to metal dusting or weight gain due to oxidation and/or carbon gain (no metal dusting attack). Sicromal (Fe-18Cr1Al-1Si) contained not very high chromium level, but showed no metal dusting. This was due

to the presence of some aluminum and silicon in the alloy. Both aluminum and silicon are effective alloying additions in improving the metal dusting resistance. X18 CrN28 (Fe-28Cr) showed signiﬁcantly better performance than Sanicro 28 (Fe-27Cr-30Ni). Both alloys contained about the same level of chromium, but one

250 560 °C, 1.5 bar

Mean metal loss, µm

200

150

100

50

0 0

50

100

150

200

Exposure time, h

Fig. 5.62

Average metal loss as a function of exposure time (hours) for P91, 410, 310, and model alloy (410 + 2.75Si). Source: Ref 86

1000 100 10

Mass gain, mg/cm

2

1 0.1 0.01 0.001 –0.01 –0.1 –1 –10

Removable deposits

Fig. 5.63

Inconel 600

AC 66

Sanicro 28

HK 40

X15 CrNiSi 2012

253 MA

12 CrMoV

Alloy 410

X18 CrN28

Sicromal

–100

Cleaned specimen

Mass gain (hatched data) due to coke deposits and metal loss of specimen (solid data) due to metal dusting as well as mass gain of specimen (solid data) due to possible oxidation/carburization after exposure at 650 °C (1200 °F) in H2-24.7CO-1.9H2O (ac = 15). X18 CrN28 (Fe-28Cr); Sanicro 28 (Fe-27Cr-30Ni); 253MA (Fe-21Cr-11Ni-2Si-0.05Ce); X15 CrNiSi 2012 (Fe-20Cr-12Ni-2Si). Source: Ref 69

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

130 / High-Temperature Corrosion and Materials Applications

Fe25Cr2.5Ni 0

Fe60Cr

Mass change, mg/mm2

–0.02

Fe25Cr

–0.04 –0.06 Fe25Cr5Ni

–0.08 –0.1 –0.12 –0.14

Fe25Cr25Ni

–0.15 Fe25Cr10Ni –0.18 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 Exposure cycles, h

Fig. 5.64

Weight loss of specimens after removing carbon deposits as a function of thermal cycles for Fe-Cr and Fe-Cr-Ni alloys in metal dusting tests at 680 °C (1256 °F) in H2-68%CO-6%H2O (ac = 2.9 and pO2 =2 · 10723 atm) cycling specimens to room temperature every 60 min in the test environment. Source: Ref 90

was a ferritic alloy (Fe-Cr) while the other was an austenitic alloy (Fe-Ni-Cr). The diffusivity of chromium in ferritic alloys is generally about two orders of magnitude higher than that in austenitic Fe-Ni-Cr alloys (Ref 87–89). As a result, ferritic stainless steels form chromium oxide scales much more readily than austenitic stainless steels, as was observed in Fig. 5.63. Both 12Cr steels (410SS and 12CrMoV) showed some metal dusting attack. Again, silicon was found to greatly improve metal dusting resistance. Both alloy 253MA (Fe-21Cr-11Ni-2Si-0.05Ce) and X15 CrNiSi 2012 (Fe-20Cr-12Ni-2Si) showed no metal dusting attack. Another siliconcontaining alloy, HK-40 (Fe-25Cr-20N-2Si), showed some metal dusting attack. The alloys that showed the worst attack in this test were Fe-Ni-Cr alloys (Sanicro 28 and AC66), and a low-chromium nickel alloy (alloy 600). The results of Toh et al. (Ref 90) further conﬁrmed that Fe-Cr alloys exhibit better metal dusting resistance than Fe-Ni-Cr alloys. They investigated Fe-25Cr alloy and Fe-25Cr with 5, 10, and 25% Ni. All alloys were annealed in Ar-10%H2 at 1050 °C (1922 °F) for 100 h. Test specimens were then cut, ground, and polished to 3 µm before being electrolytically polished. Testing was conducted at 680 °C (1256 °F) in H2-68%CO-6%H2O (ac = 2.9 and pO2 =2 · 10723 atm). Specimens were cycled by heating them to the test temperature for 60 min for each cycle. Their test results are summarized in Fig. 5.64. Fe-25Cr alloy was found to be signiﬁcantly more resistant to metal dusting than Fe-25Cr alloys with 5, 10, and 25% Ni. Fe-25Cr

with 2.5% Ni appeared to perform well. It was likely that the alloy with only 2.5% Ni remained a ferritic structure. The test also included Fe-60Cr, which showed good metal dusting resistance. Ferritic Fe-Cr alloys are more resistant to metal dusting than austenitic Fe-Ni-Cr alloys presumably due to much faster chromium diffusion rates, thus resulting in faster formation of chromium oxide scales, and then better metal dusting resistance. Rapid diffusion of chromium to the metal surface to form a protective chromium oxide scale is important in retarding metal dusting attack. Accordingly, for the same alloy, a ﬁne-grained material can form a surface oxide scale more readily than a coarse-grained material, thus better metal dusting resistance. This is illustrated in Fig. 5.65, where the ﬁne-grained Type 304SS (as-received from the supplier) with a grain size of ASTM No. 10 (average size of 10 µm) showed no metal dusting attack while the coarse-grained material with a grain size of about ASTM No. 3 to 7 (about 30–100 µm) suffered severe attack (Ref 91). The coarsegrained material was obtained by annealing the sample to 1000 °C (1832 °F) for 1 h. Metal dusting testing was conducted at 600 °C (1112 °F) in H2-24CO-2H2O. More discussion on the effects of surface conditions on the metal dusting behavior is presented later. Fe-Ni-Cr alloys, which showed less resistance to metal dusting than Fe-Cr alloys due to lower chromium diffusivity, are also much less resistance to metal dusting as compared with Ni-base alloys. This is illustrated in Fig. 5.66, which were generated on the commercial alloys without prior

pg 130

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 131

annealing treatments. Specimens were in the asreceived condition from the supplier and were ground to a 120-grit ﬁnish prior to testing in H2-90%CO for 672 days at 482, 566, 649, and 732 °C (900, 1050, 1200, and 1350 °F) (Ref 92).

10 18Cr-8Ni-Steel

Mass gain, mg/cm2

8

6 Coarse grain 4

2 Fine grain 0

0

50

100

150

200

Time, h

Fig. 5.65

Effect of grain size on the metal dusting behavior of Type 304SS tested at 600 °C (1112 °F) in H224CO-2H2O. Source: Ref 91

Fig. 5.66

Type 304, 310, and 800H—which are Fe-Ni-Cr alloys—were found to suffer metal dusting attack (i.e., the reaction rates in weight loss) while two nickel-base alloys (alloy 601 and RA333) showed no metal dusting attack (i.e., the reaction rates in weight gain). The ﬁgure also shows that alloy 85H (Fe-18.5Cr-14.5Ni-3.5Si-1Al) showed no metal dusting attack. This was due to beneﬁcial effect of silicon along with aluminum in the alloy. Beneﬁcial effects of silicon and aluminum are demonstrated by additions of these two elements to alloy 800 (Fe-20Cr-32Ni), as shown in Fig. 5.67 (Ref 93). Also in this study by Strauss and Grabke (Ref 93), other alloying elements, such as niobium, molybdenum, and tungsten, were found to have some beneﬁcial effects as well. Klower et al. (Ref 94) investigated a large number of commercial alloys including several Fe-Ni-Cr alloys, such as alloys 800H, HK-40, HP-40, and DS, and many nickel-base alloys. Test results are summarized in Table 5.5. Fe-Ni-Cr alloys were much less resistant to metal dusting than nickel-base alloys. High-siliconcontaining alloy DS (Fe-18Cr-35Ni-2.2Si) was much better than other Fe-Ni-Cr alloys. This is nicely illustrated in Fig. 5.68. In Ni-Cr-Fe

Metal dusting resistance of several Fe-Ni-Cr alloys and Ni-base alloys tested in H2-90%CO for 672 days at 482, 566, 649, and 732 °C (900, 1050, 1200, and 1350 °F). The alloys tested were Type 304 (S30403), Type 310 (S31000), alloy 85H (S30615), alloy 800H (N08810), alloy 601 (N06601), and RA333 (N06333). Source: Ref 92

pg 131

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 132

132 / High-Temperature Corrosion and Materials Applications

system, increases in chromium along with aluminum additions can signiﬁcantly improve the alloy’s metal dusting resistance, as illustrated in Fig. 5.69. Alloy 601 was more resistant than alloy 600 because of higher chromium and aluminum addition in alloy 601. Alloy 602CA was better than alloy 601 because of higher aluminum content. There were two different surface ﬁnishes involved in this test, with alloys 600 and 601 in as ground surface, and alloy 602CA in “black” surface ﬁnish. There was no discussion about the procedure for preparing the “black” surface ﬁnish in the paper. If the 602CA specimens were black annealed (heat treated in air), with about 20 +4.5%AI

+2.5%Si + +

Mass loss, mg/cm2

+3%Mo +3%Nb

–20

+3%W

Fe-32%Ni-20%Cr –60

+0.05%Ce –100

0

500

Fig. 5.67

1000 Time, h

1500

Effects of Al, Si, Nb, Mo, and W additions to a model alloy (alloy 800, Fe-20Cr-32Ni) on the metal dusting behavior of modified experimental alloys. Source: Ref 93

2.3% Al in alloy 602CA, alloy 602CA specimens might have preformed an Cr/Al-rich oxide scales prior to metal dusting testing. The effects of surface ﬁnishes are discussed further. In Table 5.5, it is surprising to ﬁnd that alloy 214 with about 4.5% Al did not perform well compared with 602CA (2.3% Al), 617 (1.3% Al), and 690 (30% Cr, no Al). It is believed that with only about 16% Cr, alloy 214 did not contain enough chromium to rapidly develop a protective chromium oxide scale at such a low temperature. And, at this low temperature, development of an exclusive Al2O3 is difﬁcult in short time. Thus, a Table 5.5 Metal wastage rates after exposure to H2-CO-H2O gas mixtures at 650 °C (1200 °F) for various commercial alloys Alloy

Surface condition

Total time, h

Wastage rates, mg/cm2h

800H HK-40 HP-40 DS 600H 601 601 601 C-4 214 HR160 45TM 602CA 617 690

Ground … … Ground Ground Black Polished Ground Ground Ground Ground Black Black Ground Ground

95 190 190 1,988 5,000 6,697 1,988 10,000 10,000 9,665 9,665 10,000 10,000 7,000 10,000

0.21 0.04 0.038 4.3 × 10−3 3.3 × 10−2 7.3 × 10−3 4.9 × 10−3 5.8 × 10−4 1.1 × 10−3 1.2 × 10−3 6.3 × 10−4 1.0 × 10−5 1.1 × 10−5 3.7 × 10−6 2.0 × 10−6

Note: Gas mixtures H2-24CO-2H2O (ac = 14 at temperature) for the ﬁrst 5000 h of testing and H2-49CO-2H2O for the subsequent 5000 h. Source: Ref 94

102

Metal wastage rate, mg/cm2h

10 1 10–1 10–2 10–3 10–4 10–5

Fig. 5.68

Metal dusting resistance of Fe-Ni-Cr alloys (800H and HP40) in comparison with Ni-base alloy 600H tested at 650 °C (1200 °F) in H2-24CO-2H2O (ac = 14). Source: Ref 94

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 133

Chapter 5: Carburization and Metal Dusting / 133

high-chromium nickel-base alloy, alloy 690 (30% Cr), performed the best in this test. It is also surprising to ﬁnd that HR-160 alloy with both high Cr (28%) and high Si (2.75%) did not perform as well as some other nickel-base alloys. Alloy 45TM, also with high Cr (27%) and high Si (2.7%) performed better than HR-160 alloy. The

alloy 45TM specimen was in a “black” surface condition. The “black” surface condition might be resulted from a preoxidation treatment that might have produced chromium- and silicon-rich oxide scales prior to metal dusting testing. In conducting metal dusting testing that involved only the specimens with as-ground

10–1

Metal wastage rate, mg/cm2h

Alloy 600 10–2 Alloy 601 10–3

10–4

Alloy 602CA

10–5

10–6 0

2,000

4,000

6,000

8,000

10,000

Exposure time, h

Fig. 5.69

Increasing Cr along with addition of aluminum improves the metal dusting resistance of the alloy in Ni-Cr-Fe alloys. Source: Ref 94

Fig. 5.70

Metal dusting behavior of various Ni-base alloys tested at 593 °C (1100 °F) in H2-18.4CO-5.7CO2-22.5H2O at 14.3 atm of pressure. Source: Ref 95

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 134

134 / High-Temperature Corrosion and Materials Applications

surface ﬁnish, Natesan and Zeng (Ref 95) performed tests at 593 °C (1100 °F) in H2-18.4CO5.7CO2-22.5H2O at 14.3 atm of pressure. Alloy 45TM suffered rapid metal dusting attack with about 59.1 mg/cm2 of weight loss after only 3300 h while HR-160 alloy exhibited only 7.3 mg/cm2 of weight loss after 9700 h. Their results are summarized in Fig. 5.70. The bestperforming alloy was alloy 693 (30% Cr, 3.3Al), followed by alloy 602CA (25% Cr, 2.3% Al), and HR-160 (28% Cr, 2.75% Si) and alloy 690 (30% Cr). The worst performer was alloy 45TM, followed by alloys 617, 214, and 601. These tests conducted in Argonne National Laboratory were under a high pressure (14.3 atm, or 210 psi), and all other data generated in other laboratories were under gas pressure close to 1 atm (atmosphere pressure). Natesan and Zeng (Ref 95) observed that high pressure could signiﬁcantly reduce the time to initiate metal dusting attack for some alloys. Tests were performed at 593 °C (1100 °F) for 246 h in 1, 14.3, and 40.8 atm pressures. Specimens were then examined for signs of metal dusting attack. Their results are summarized in Table 5.6, showing alloys 601, 690, 617, and 214 suffering metal dusting attack at high pressures but not at 1 atm pressure after 246 h. On the other hand, alloys 45TM, 602CA, and HR160 showed no metal dusting attack at both low and high pressures after 246 h. They also measured the maximum pit size and average pit depth in addition to weight loss. This is illustrated in Table 5.7. In terms of the pitting depth, HR160 was found to perform the best, while alloy 690 with little weight loss showed signiﬁcant pit depth. The data suggested that the total weight loss was not correlated well with pitting depth. Alloy 214 was found to show uniform metal wastage. This was apparently due to insufﬁcient chromium content (16%) at such a low temperature to form a continuous Cr2O3

scale. The Raman spectra of the alloy 214 tested specimen showed low intensity for the Cr2O3 band (Ref 95). In manufacturing of sheet products (or thin gage tubular products) in nickel-base alloys, the manufacturing process in the later stage typically consists of cold working (cold pilgering) and bright annealing (i.e., annealing in H2 atmosphere). The hydrogen atmosphere in bright annealing typically contains some moisture, thus it is typically characterized with a certain dew point (i.e., a certain oxygen potential, pO2 ). In general, annealing furnaces exhibit low enough dew points such that chromium oxide scales do not form on the metal surface when processing stainless steels or nickel-base alloys. Sheet products of these alloys typically look shiny. However, the dew points or oxygen potentials in hydrogen-annealing furnaces generally are high enough that Al2O3 or SiO2 will form on the metal surface of a bright-annealed sheet of a nickel alloy containing a sufﬁcient level of aluminum or silicon. Klarstrom and Grabke (Ref 96) tested bright-annealed sheet specimens of alloy 214 (alumina former) and alloy HR160 (silica former) along with alloy 230 and HR120 in the bright-annealed condition. Also included in testing were black annealed and pickled sheet specimens of alloys 800H and 601 (no brightannealed sheet samples were available for these two alloys at the time of testing). No surface grinding for bright-annealed sheet specimens (alloys 214, HR160, 230, and HR120). Surface grinding to a 120-grit surface ﬁnish was performed for black-annealed and pickled sheet specimens of alloys 800H and 601. Testing was performed at 650 °C (1200 °F) for up to 10,000 hours in H2-49CO-2H2O (ac =18.9). Test results are summarized in Table 5.8. HR160 alloy was found to perform very well, showing no evidence of metal dusting after 10,000 h

Table 5.6 Surface conditions of alloys after testing at 593 °C (1100 °F) for 246 h in H2-18.4CO-5.7CO2-22.5H2O at 1, 14.3, and 40.8 atm pressures

Table 5.7 Pit size, pit depth, and weight loss for alloys after testing at 593 °C (1100 °F) for 9700 h in H2-18.4CO-5.7CO2-22.5H2O at 14.3 atm pressure

Condition at pressure Alloy

601 690 617 602CA 214 45TM HR160 Source: Ref 95

Alloy

1 atm

14.3 atm

40.8 atm

Clean surface Clean surface Clean surface Clean surface Clean surface Clean surface Clean surface

Pits Pits Pits Clean surface Pits Clean surface Clean surface

Pits Pits Pits Clean surface Pits Clean surface Clean surface

601 690 617 602CA 214 45TM(b) HR160 693

Weight loss, mg/cm2

Pit depth, µm

Pit diameter, µm

19.5 6.5 35.1 2.1 25.6 59.1 7.3 0.1

110 147 201 96 (a) 141 13 37

450 440 887 374 (a) 600 210 99

(a) Specimen uniformly corroded. (b) Exposed for only 3300 h. Source: Ref 95

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 135

of exposure. The alloy, however, showed slight carburization underneath the metal surface. Alloy 230 showed slight metal dusting attack with a few spots 1 to 2 mm in diameter on each side of the specimen after 10,000 h of exposure. Alloy 601 suffered signiﬁcantly more metal dusting attack than alloy 230. Typical surface conditions for these three alloys after 10,000 h of testing are shown in Fig. 5.71. Alloy 214, however, did not perform well, suffering metal Table 5.8 Final metal wastage rates for various alloys tested at 650 °C (1200 °F) for up to 10,000 h in H2-49CO-2H2O (ac = 18.9) Alloy

HR120 800H 214 601 230 HR160

Total exposure time, h

Wastage rate, mg/cm2h

190 925 5,707 10,000 10,000 10,000

4.1 × 10−2 2.7 × 10−3 1.0 × 10−3 2.5 × 10−3 3.2 × 10−4 0.0(a)

dusting attack with fairly high wastage rates. Both Fe-Ni-Cr alloys, alloys 800H, and HR120, performed the worst among the alloys tested. Figure 5.72 shows typical surface conditions of alloys 214, HR120, and 800H after 5707 h, 190 h, and 925 h, respectively. Baker and Smith (Ref 97, 98) and Baker et al. (Ref 99) investigated about 20 commercial alloys in H2-80CO at 621 °C (1150 °F) for times up to 16,000 h. All materials were cold rolled and annealed sheets (i.e., bright annealed products) except alloys K-500 and 617, which were hotrolled, annealed plates (i.e., black annealed and pickled products), and alloy DS, which was extruded and annealed tubing. Nevertheless, test coupons were all ground to a 120-grit ﬁnish. For exposure times less than 10,000 h, weight changes as a function of exposure time for Fe-Ni-Cr alloys as well as 9Cr steel and Monel in comparison with some nickel-base alloys are

(a) Attack too small for analysis. Source: Ref 96

Fig. 5.71

Optical micrographs showing typical surface conditions for (a) HR160 alloy, (b) alloy 230, and (c) alloy 601 after testing at 650 °C (1200 °F) for 10,000 h in H249CO-2H2O (ac = 18.9). Magnification bar represents 20 µm for (a), 200 µm for (b) and (c). Source: Ref 96

Fig. 5.72

Typical surface conditions for (a) alloy 214 after 5707 h, (b) alloy HR120 after 190 h, and (c) alloy 800H after 925 h at 650 °C (1200 °F) in H2-49CO-2H2O (ac = 18.9). Source: Ref 96

pg 135

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:23PM Plate # 0

pg 136

136 / High-Temperature Corrosion and Materials Applications

summarized in Fig. 5.73. Nickel-base alloys except alloy 600 performed much better than Fe-Ni-Cr alloys. Nickel-base alloy 600 performed poorly because of its low chromium content (15–16%). MA956 (Fe-20Cr-4.5AlY2O3 ODS alloy) performed very well. The test results are also presented in Fig. 5.74 in terms of

maximum pitting depth as a function of exposure time. Alloys 690, 617, and MA956 showed the minimum pitting attack. Figure 5.75 shows the test results for some additional alloys and some better performing alloys tested up to 16,000 hours (Ref 99). Alloys 690, 263, and MA956 were observed to show pits with

20 263

617 0 825

864

MA 956

690 601

600 –40 9Cr-1Mo

–60

K-500

Mass change, mg/cm2

–20

DS 803

800

–80

–100 330 –120 0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

Exposure time, h

Fig. 5.73

Mass change as a function of exposure time for various alloys including 9Cr steel, Ni-Cr alloy (Monel K-500), Fe-Ni-Cr alloys, and Ni-base alloys tested at 621 °C (1150 °F) in H2-80CO. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 98

762

30 800 803

601

o -1M

20

508

330

15

381 864

10

254

5

82 5

690 617

600

127

Maximum pitting depth, µm

635

DS

9Cr

Maximum pitting depth, mils

25

MA956 0

0 K500 –5 0

1000

2000

3000

4000

5000

6000

7000

8000

–127 9000

Exposure time, h

Fig. 5.74

Maximum pit depths as a function of exposure time for various alloys including 9Cr steel,Fe-Ni-Cr alloys and Ni-base alloys tested at 621 °C (1150 °F) in H2-80CO. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 98

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

pg 137

Chapter 5: Carburization and Metal Dusting / 137

25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 –1

800

than that of alloy 693 (Fig. 5.76). It should be cautioned in interpreting pit depth data, when the specimen surface has shown general wastage without the original specimen surface surrounding the pit, pitting depth data thus measured then becomes questionable. This would particularly be the case after very long exposure when pits have spread throughout the specimen surface. In reviewing the data summarized in Fig. 5.75 and 5.76, one alloy of interest was alloy 671 (Ni-46Cr), which was found to exhibit excellent metal dusting resistance. Unfortunately, the test on this alloy was terminated at close to 10,000 h. This alloy exhibited low pit depths and very low metal loss rates (lower than alloy 693). Grabke et al. (Ref 100) also found that Ni-50Cr alloy was very resistant to metal dusting, showing a thin chromium oxide scale (about 3 µm thick) after testing in H2-49CO-2H2O at 650 °C (1200 °F) (ac = 18.9) for 10,000 h with no evidence of metal dusting or carburization. Only minute amounts of coke were observed on the specimen surface. Under this test condition, alloy 601 had already suffered metal dusting attack after only 1993 h of exposure. Similarly, a very thin chromium oxide scale was observed in

DS

610

803

559 601

508

33

0

956 263

457

690

356 305

4

254

86 5

82

203

TD 758 LCE

625

152

C–276

671

102

754

617

693

600 0

Pit depth, µm

406

602CA

400

Pit depth, mils

signiﬁcant pit depths after 14,000–16,000 h of exposure. Alloys 617 and 693, on the other hand, still exhibited pits with insigniﬁcant pit depths. Alloy 693 was a recently developed commercial alloy targeting for metal dusting environments by adding about 3% Al to alloy 690 with same amount of Cr (about 30%) but slightly lowered Fe. The alloy was found to perform much better than alloy 690. It is understandable that high Cr and high Al in nickel-base alloys should perform well in metal dusting environments. However, it was surprising to ﬁnd that alloy 617 with only 22% Cr (quite normal level chromium for hightemperature alloys) and fairly low aluminum content (about 1.3%) exhibited a pitting depth similar to that of alloy 693. In metal dusting testing by Natesan and Zeng (Ref 95) involving several nickel-base alloys including alloy 617, the test results, which were presented as weight loss, showed alloy 617 was worse than alloy 601, 602CA, and several other nickel-base alloys (Fig. 5.70). When Baker and Smith (Ref 98) presented their test results in terms of weight loss rate as a function of exposure time, alloy 617 was found to exhibit metal loss rate (in the range of alloys 263 and MA956) signiﬁcantly higher

51 0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Exposure time, h

Fig. 5.75

Maximum pit depths as a function of exposure time for various alloys including 9Cr steel, Fe-Ni-Cr alloys, and Ni-base alloys tested at 621 °C (1150 °F) in H2-80CO for exposure time up to 16,000 h. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 99

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

pg 138

138 / High-Temperature Corrosion and Materials Applications

0 80

6 0 82 00 5

10–1

33

40

0

100

1

80

3

4 75

864 DS

60

Mass loss rate, mg/cm2h

10–2

625

10–3

690

C 276

58

263

602CA

7

MA

TD

617

956

10–4

10–5 693 671 0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Exposure time, h

Fig. 5.76

Metal loss rate as a function of the exposure time for various alloys tested at 621 °C (1150 °F) in H2-80CO for exposure time up to 16,000 h. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 99

Cr-5Fe-1Y2O3 after 10,000 h of exposure in the same test, showing no evidence of metal dusting attack. 5.3.4 Effects of Surface Conditions and Finish From the discussion so far on metal dusting, formation of a good, protective chromium oxide scale is a very effective way to provide protection against metal dusting attack. Since the temperature range for metal dusting attack is quite low and chromium diffusivities are relatively low at these temperatures, rapid formation of a good, protective chromium oxide scale is critical. Accordingly, the concentration of chromium at the surface of the metal becomes important. Higher chromium concentration at the metal surface provides faster formation of a protective chromium oxide scale. Thus, alloys with higher bulk chromium concentration (higher surface chromium concentration) are more resistant to metal dusting as was discussed in the previous section. Also, for the same alloy, the ﬁne-grained structure exhibits better metal dusting resistance than the coarse-grained structure, as illustrated in

Fig. 5.65. This is because grain boundaries provide fast diffusion paths for chromium to reach to the metal surface. A ﬁne-grain-sized material will have more grain boundaries and, thus, more chromium reaching to the metal surface to form a better chromium oxide scale faster than a coarsegrained material, thus, resisting metal dusting attack much better. The surface of the metal can also be prepared by grinding or machining to produce a thin coldworked layer with high density of dislocations, which also provide fast diffusion paths for chromium to reach to the metal surface to form a protective chromium oxide scale. As a result, metal dusting is greatly improved when the surface is ground or machined. The beneﬁcial effect of the ground surface condition is illustrated Fig. 5.77 and 5.78 in a thermogravimetric study by Grabke et al. (Ref 101). For both alloy 800 and Type 310SS, the as-ground specimen was most resistant to metal dusting compared with the as-received surface (cold rolled) and electropolished surface. Electropolished surface condition was the worst because of possible surface depletion in chromium during electropolishing. Both alloy 800 and Type 310SS were

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 139

tested under the same condition; Type 310 was much more resistant to metal dusting than alloy 800. This is primarily because Type 310SS contains more chromium. Specimens with ground surface ﬁnish, alloy 601, showed better metal dusting resistance than those with electropolished surface ﬁnish, as shown in Fig. 5.79 (Ref 94). In this study, the “black” sample (believed to be a black-annealed sample) performed as poorly as the electropolished

10 Alloy 800 8 Mass gain, mg/cm2

Electropolished Cold rolled 6

4

Ground

2

0

0

5

10

15

20

25

Time, h

Fig. 5.77

Thermogravimetric testing in H2-24CO-2H2O at 600 °C (1112 °F) for alloy 800 (Fe-21Cr-32Ni) in three different surface conditions: as-ground surface (to 600-grit), cold-rolled (as-received surface), and electropolished surface. Source: Ref 101

sample. The black-annealed sample, as in electropolished specimen, exhibited surface chromium depletion due to annealing in air (i.e., black annealing). The formation of chromium oxide scales during annealing in air results in some chromium depletion in the surface layer. Similar results were obtained by Baker and Smith (Ref 98) in their study on alloy 601 with different surface conditions, as shown in Fig. 5.80 and 5.81. In this study, samples from as-received sheet manufactured from blackannealing and pickling (i.e., annealing in air followed by acid pickling to remove oxide scales), which were identiﬁed as as-produced (annealed+ pickled), were slightly better than those black-annealed samples and electropolished samples. Black annealing and acid pickling can often result in surface depletion in chromium. These processes are often involved in producing sheet products by hot rolling. When sheet products are produced via cold rolling, the sheet is typically bright annealed. Under this condition, no chromium oxide scales formed on the sheet metal; thus no chromium depletion occurred at the metal surface. For highaluminum or high-silicon alloys, bright-annealing atmospheres typically exhibit high enough oxygen potentials (or dew points) that a thin aluminum or silicon oxide ﬁlm can form, which can be exceedingly beneﬁcial to the subsequent service in metal dusting environments. This was discussed earlier for HR160. Alloy 214, unfortunately with too low a chromium level, showed no beneﬁcial effect from bright annealing in its resistance to metal dusting attack.

6

Mass gain, mg/cm2

25Cr-20Ni - Steel

5.3.5 Metal Dusting Behavior of Weldments Selection of an appropriate ﬁller metal for metal dusting environments is also critical, since

Electropolished

4

2 Cold rolled

Ground 0 0

Fig. 5.78

5

10 15 Time, h

20

25

Thermogravimetric testing in H2-24CO-2H2O at 600 °C (1112 °F) for Type 310SS (Fe-25Cr-20Ni) in three different surface conditions: as-ground surface (to 600grit), cold-rolled (as-received surface), and electropolished surface. Source: Ref 101

Fig. 5.79

Metal dusting behavior at 650 °C (1200 °F) in H2-49CO-2H2O for alloy 601 in three different surface conditions: as-ground surface, elecropolished surface, and black-annealed surface. Source: Ref 94

pg 139

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

pg 140

140 / High-Temperature Corrosion and Materials Applications

not every wrought alloy has a matching ﬁller metal. Thus, it is necessary to select a ﬁller metal that is at least as good as, but preferably better than, the wrought alloy selected for the application. The suitable ﬁller metal requires not only good resistance to metal dusting but also good

weldability. Also, in many cases, machining or grinding the weld joint may not be possible, particularly in tubular butt welds where inside diameter (ID) grinding or machining is not feasible after joining. Furthermore, alloys that resist metal dusting contain high aluminum or

Ground

Fig. 5.80

Metal dusting behavior in terms of weight change tested at 621 °C (1150 °F) in H2-80CO for alloy 601 in various surface conditions. Source: Ref 98

Fig. 5.81

Metal dusting behavior in terms of weight loss rate tested at 621 °C (1150 °F) in H2-80CO for alloy 601 in various surface conditions. Source: Ref 98

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 141

silicon along with high chromium, as discussed in previous sections. If matching ﬁller metals are available for these alloys, the weldability of these high-aluminum or high-silicon ﬁller metals can be an issue. Grabke et al. (Ref 101) investigated the metal dusting behavior of different weldments involving base metals of alloys 800H, 600H, 601, and 602CA using different ﬁller metals, such as alloy 82 (Nicrofer S 7020 or UTP 068 HH) and alloy 602CA (Nicrofer 6025 or UTP 6225 Al). Weldment specimens included alloy 800H to alloy 800H with alloy 82 ﬁller metal, alloy 600H to alloy 600H with alloy 82 ﬁller metal, alloy 601H to alloy 601H with alloy 602CA, and alloy 800H to alloy 602CA with alloy 602CA as a cap layer and alloy 82 ﬁller metal for root pass and ﬁller. Welding processes involved were listed as TIG welding (i.e., gas-tungsten arc welding) and MMAW (probably shielded metal arc welding). The surface ﬁnish for most weldment specimens 1.0 no H2S

0.1 ppm H2S 0.3 ppm H2S 0.5 ppm H2S 0.75 ppm H2S

∆m/A, mg/cm2

0.8 0.6 0.4

1 ppm H2S

0.2 0.0

0

50

100

150

200

250

were “brushed,” with one weldment being ground, one being sandblasted, and one being pickled. Tests were conducted at 600 and 650 °C (1112 and 1200 °F) in H2-24CO-2H2O. In almost all cases, metal dusting was initiated at the interface between the weld metal and the base metal (i.e., the heat-affected zone surface). Alloy 82 weld metal was found to be more resistant to metal dusting than alloy 800H and alloy 600H. The authors (Ref 101) also concluded that TIG welding led to a better resistance to metal dusting than “hand-welding,” and grinding led to a modest delay in initiation of metal dusting attack. 5.3.6 Effects of Sulfur on Metal Dusting Sulfur, which has a strong tendency to segregate to the surface and grain boundaries, can act as an inhibitor to metal dusting. Hochman (Ref 67) ﬁrst reported beneﬁcial effect of sulfur against metal dusting attack. Extensive investigations on the effect of sulfur on metal dusting behavior of iron were carried out by Grabke and Muller-Lorenz (Ref 102), Schneider et al. (Ref 103), Schneider et al. (Ref 104), and Schneider and Grabke (Ref 105). The mechanism for sulfur to inhibit metal dusting attack, as proposed by Grabke and Muller-Lorenz (Ref 102), is the absorption of sulfur on the surface of cementite suppresses the nucleation of graphite, thus inhibiting metal dusting attack. Figure 5.82 shows the beneﬁcial effect of H2S on the kinetics of metal dusting of iron at 500 °C

300

Time, h

Fig. 5.82

Effect of H2S on metal dusting behavior of iron at 500 °C (932 °F) in H2-CO-H2O-H2S gas mixture (ac = 100). Source: Ref 103

1.50

no H2S 0.7 ppm H2S 0.1 ppm H2S 1 ppm H2S 0.5 ppm H2S

∆m/A, mg/cm2

1.25 1.00 0.75

5 ppm H2S

0.50

15 ppm H2S

0.25 0.00

T=700 °C ac = 100

0

Fig. 5.83

50

100

150 200 Time, h

250

300

350

Effect of H2S on metal dusting behavior of iron at 700 °C (1292 °F) in H2-CH4-H2S gas mixture (ac = 100). Source: Ref 104

Fig. 5.84

Effect of H2S on metal dusting behavior of iron in terms of pH2 S =pH2 versus 1/T. Open data points represent that the onset of metal dusting was retarded for more than 48 h, while the solid data points represent metal dusting without retardation. The hatched region represents a transition to an iron surface saturated with sulfur. Source: Ref 74

pg 141

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

142 / High-Temperature Corrosion and Materials Applications

(932 °F) in H2-CO-H2O-H2S gas mixture (ac = 100) (Ref 103), and Fig. 5.83 shows the similar effect for iron at 700 °C (1292 °F) in H2-CH4H2S gas mixture (ac = 100) (Ref 103). Grabke (Ref 74) provides metal dusting behavior of iron in terms of pH2 S =pH2 ratio as a function of temperature (1/T) in Fig. 5.84. The ﬁgure shows the region when metal dusting was avoided by sulfur injection under certain pH2 S =pH2 ratios at different temperatures from about 500 to 1000 °C (932 to 1832 °F). Injecting a right amount of sulfur into the environment to retard metal dusting without causing accelerated sulﬁdation attack is a balancing act.

5.4 Summary Metals and alloys are generally susceptible to carburization when exposed to environments containing CO, CH4, or other hydrocarbon gases at elevated temperatures. Carburization typically results in the formation of internal carbides in the matrix as well as boundaries, causing the alloy to lose its room-temperature ductility and/or creep-rupture strengths. Fe-Ni-Cr alloys are widely used for processing equipment to resist carburization in the petrochemical industry. The cast 25Cr-20Ni, HK40, was once the workhorse of pyrolysis furnace tubes in ethylene cracking operations. Many modiﬁcations based on HK40 have been developed and used now with improved carburization resistance as well as increased creep-rupture strengths. These alloy modiﬁcations involve the use of alloying elements such as, titanium, niobium, tungsten, molybdenum, and silicon, as well as increases in nickel and/or chromium. Increasing nickel in Fe-Ni-Cr alloys improves carburization resistance. Nickel reduces the diffusivity of carbon in Fe-Ni-Cr alloys. Nickel also reduces the solubility of carbon in Fe-Ni alloys. Among these alloying elements, silicon is the most effective in improving carburization resistance. This is attributed to the formation of SiO2 scale, which is more impervious to carbon ingress than Cr2O3 scale. However, when the silicon content in the alloy is too high, the weldability of the alloy can become a serious issue. Aluminum is another alloying element that can signiﬁcantly improve carburization resistance. However, effectiveness generally requires about 4% or higher aluminum, the amount needed to form a continuous Al2O3 scale.

Alumina formers (i.e., alloys forming Al2O3 scale, such as alloy 214 and MA956) are signiﬁcantly better than chromia formers (i.e., alloys forming Cr2O3 scale). Surface ﬁnish plays an important role in carburization resistance. For cast products, machining the metal surface can signiﬁcantly reduce carburization attack. Injecting sulfur compounds (e.g., 50 to 100 ppm) into the processing gas stream is also effective in reducing carburization attack. Metal dusting is another form of carburization attack; it typically causes an alloy to suffer pitting attack and/or thinning. The metal beneath the pitted area generally shows carburization. The corrosion products typically consist of carbon soots, metal particles, carbides, and oxides. The environment in which metal dusting occurs generally contains H2, CO, CO2, and H2O with high carbon activities (i.e., aC > 1). Stagnant gas conditions can be conducive in initiating metal dusting attack. The metal temperatures at which metal dusting occurs are between 430 and 900 °C (800 and 1650 °F). Metal dusting data for various commercial alloys are presented. Nickel-base alloys containing high chromium and high aluminum (e.g., alloys 602CA and 693) or containing high chromium and high silicon (e.g., alloy HR160) showed excellent resistance to metal dusting attack. Surface conditions also play an important role in metal dusting resistance. Sulfur may also retard metal dusting attack.

REFERENCES

1. Metals Handbook, 8th ed., Vol 2, American Society for Metals, 1964, p 93 2. L.J. Haga, Heat Treat., Dec 1986, p 6 3. G.E. Moller and C.W. Warren, Paper No. 237, Corrosion/81, NACE, 1981 4. A.J. McNab, Hydrocarbon Process., Dec 1987, p 43 5. G.L. Swales, in Behavior of High Temperature Alloys in Aggressive Environments, Proc. 1979 Petten International Conference, I. Kirman et al., Ed., The Metals Society, London, 1980, p 45 6. R.A. Perkins, W.C. Coons, and F.J. Radd, in Properties of High Temperature Alloys, Proc. 1976 Fall Meeting of the Electrochemical Society, The Electrochemical Society 7. R.C. Schueler, Hydrocarbon Process., Aug 1972, p 73

pg 142

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 143

8. F.A. Prange, Corrosion, Vol 15 (No. 12), Dec 1959, p 619t 9. G.Y. Lai, M.F. Rothman, and D.E. Fluck, Paper No. 14, Corrosion/85, NACE, 1985 10. O.J. Dunmore, Proc. UK Corrosion Conf. 1982, Institution of Corrosion Science and Technology, Birmingham, UK, 1982, p 101 11. F.N. Mazandarany and G.Y. Lai, Nucl. Technol., Vol 43, 1979, p 349 12. K. Natesan, Nucl. Technol., Vol 28, 1976, p 441 13. K. Natesan and T.F. Kassner, Metall. Trans., Vol 4, 1973, p 2557 14. HSC, Chemistry for Windows, Version 6.0, A. Roine, Outokumpu Technology, Finland, www.outokumputechnology.com, accessed Dec 2006 15. ChemSage, Version 4.16, GTT-Technologies, Aachen, 1998 16. P.L. Hemmings and R.A. Perkins, “Thermodynamic Phase Stability Diagrams for the Analysis of Corrosion Reactions in Coal Gasiﬁcation/Combustion Atmospheres,” Report FP-539, EPRI, Palo Alto, CA, 1977 17. R. Hultgren, P. Desai, D. Hawkins, M. Gleiser, and K.K. Kelley, Selected Values of the Thermodynamic Properties of Elements and Binary Alloys, ASM, 1973 18. L.C. Browning, T.W. Dewitt, and P.H. Emmett, J. Am. Chem. Soc., Vol 72, 1950, p 4211 19. K.H. Jack, J. Iron Steel Inst., Vol 169, 1951, p 26 20. H.J. Christ, Experimental Characterization and Computer-Based Description of the Carburization Behaviour of the Austenitic Stainless Steel AISI 304L, Mater. Corros., Vol 49, 1998, p 258 21. R.G. Coltters, Mater. Sci. Eng., Vol 76, 1985, p 1 22. S.R. Shatynski, Oxid. Met., Vol 13 (No. 2), 1979, p 105 23. W.F. Chu and A. Rahmel, Oxid. Met., Vol 15, 1981, p 331 24. Y. Nishiyama, N. Otsuka, and T. Nishizawa, Carburization Resistance of Austenitic Alloys in CH4 -CO2 -H2 Gas Mixtures at Elevated Temperatures, Corrosion, Vol 59 (No. 8), 2003, p 688 25. H.J. Grabke, “Carburization: A High Temperature Corrosion Phenomenon,’’ MTI Publications No. 52, Materials Technology Institute of the Chemical Process Industries, St. Louis, Missouri, 1998

26. S. Forseth and P. Kofstad, Carburization of Fe-Ni-Cr Steels in CH4-H2 Mixtures at 850–1000 °C, Mater. Corros., Vol 49, 1998, p 266 27. D. Jakobi and R. Gommans, Typical Failures in Pyrolysis Coils for Ethylene Cracking, Mater. Corros., Vol 54 (No. 11), 2003, p 881 28. R.H. Krikke, J. Horing, and K. Smit, Mater. Perform., Aug 1976, p 9 29. J. Klower and U. Heubner, Carburization of Nickel-Base Alloys and its Effects on the Mechanical Properties, Mater. Corros., Vol 49, 1998, p 237 30. H.J. Grabke, Metallic Corrosion, Proc. Eighth Int. Cong. Metallic Corrosion, Vol III, Mainz, Germany, Sept 6–11, 1981 31. I. Wolf and H.J. Grabke, Solid State Commun., Vol 54, 1985, p 5 32. C.M. Schillmoller, Chem. Eng., Jan 6, 1986, p 87 33. D.J. Hall, M.K. Hossain, and J.J. Jones, Mater. Perform., Jan 1985, p 25 34. J.A. Thuillier, Mater. Perform., Nov 1976, p9 35. J.F. Norton and J. Barnes, in Corrosion in Fossil Fuel Systems, I.G. Wright, Ed., The Electrochemical Society, 1983, p 277 36. O. Van der Biest, J.M. Harrison, and J.F. Norton, in Behavior of High Temperature Alloys in Aggressive Environments, Proc. International Conference (Patten, The Netherlands), Oct 15–18, 1979, The Metal Society, London, 1980, p 681 37. J.M. Harrison, J.F. Norton, R.T. Derricott, and J.B. Marriott, Werkst. Korros., Vol 30, 1979, p 785 38. O. Demel, E. Keil, and P. Kostecki, SGAW Report No. 2538, Osterreichische, Studiengesellschaft fur Atomenergie, Lenaugasse 10, A-1082 Wien, Forschungszentrum Seibersdorf, Institut fur Metallurgie 39. R.P. Smith, Trans. Met. AIME, Vol 224, 1962, p 10 40. H.J. Grabke, U. Gravenhorst, and W. Steinkusch, Werkst. Korros., Vol 27, 1976, p 291 41. S.K. Bose and H.J. Grabke, Z. Metallkde., Vol 69, 1978, p 8 42. C. Steel and W. Engel, AFS Int. Cast Metals J., Sept 1981, p 28 43. L.H. Wolfe, Laboratory Investigation of High Temperature Alloy Failure Mechanisms, Paper No. 12, Corrosion/77, NACE, 1977 44. I.Y. Khandros, R.G. Bayer, and C.A. Smith, Paper No. 10, Corrosion/84, NACE, 1984

pg 143

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

144 / High-Temperature Corrosion and Materials Applications

45. W. Steinkrsch, Werkst. Korros., Vol 30, 1979, p 837 46. L.H. Wolfe, Mater. Perform., April 1978, p 38 47. U. Van den Bruck and C.M. Schillmoller, Paper No. 23, Corrosion/85, NACE, 1985 48. R.H. Kane, Paper No. 266, Corrosion/83, NACE, 1983 49. J.F. Mason, J.J. Moran, and E.N. Skinner, Corrosion, Vol 16, 1960, p 593t 50. R.H. Kane and J.C. Hosier, “Carburization Resistance of Some Wrought NickelContaining Alloys in Simulated Industrial Environments,” Inco Alloys International Technical Report, Inco Alloys International, Huntington, WV, 1985 51. D.R.G. Mitchell, D.J. Young, and W. Kleemann, Carburization of Heat-Resistant Steels, Mater. Corros., Vol 49, 1998, p 231 52. G.Y. Lai, in High Temperature Corrosion in Energy Systems, Proc. TMS-AIME Symposium, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 551 53. G.Y. Lai, B. Li, B. Gleeson, and H.L. Craig, Proposed Standard Carburization Test Method, Paper No. 3473, Corrosion/ 2003, NACE International, 2003 54. R.H. Kane, G.M. McColvin, T.J. Kelly, and J.M. Davison, Paper No. 12, Corrosion/84, NACE, 1984 55. R. Kirchheiner, D.J. Young, P. Becker, and R.N. Durham, Improved Oxidation and Coking Resistance of a New Alumina Forming Alloy 60HT for the Petrochemical Industry, Paper No. 5428, Corrosion/2005, 2005 56. F. Pons and M. Hugo, Paper No. 272, Corrosion/81, NACE, 1981 57. F.N. Mazandarany and G.Y. Lai, “Corrosion Behavior of Selected Structural Materials in a Simulated Steam-Cycle HTGR Helium Environment,” GA-A14446, General Atomic Company, San Diego, CA, Oct 1977 58. T.A. Ramanarayanan, R.A. Petkovic, J.D. Mumford, and A. Ozekcin, Carburization of High Chromium Alloys, Mater. Corros., Vol 49, 1998, p 226 59. H.J. Grabke, R. Moller, and A. Schnaas, Werkst. Korr., Vol 30, 1979, p 794 60. H.J. Grabke, E.M. Peterson, and S.R. Srinivasan, Surf. Sci., Vol 67, 1977, p 501 61. T.A. Ramanarayanan and D.J. Srolovitz, Carburization Mechanisms of High Chromium Alloys, J. Electrochem. Soc.: Solid-

62. 63.

64. 65.

66. 67.

68. 69. 70.

71. 72.

73. 74.

75. 76.

State Sci. Technol., Vol 132 (No. 9), 1985, p 2268 H.J. Grabke and D. Jakobi, High Temperature Corrosion of Cracking Tubes, Mater. Corros., Vol 53, 2002, p 494 V. Guttmann and H. Schonherr, Creep Properties of Two 25Cr-20Ni Cast Alloys in Air and under Carburizing Conditions, High Temp. Technol., Vol 3 (No. 2), 1985, p 79 D. Jakobi and R. Gommans, Typical Failures in Pyrolysis Coils for Ethylene Cracking, Mater. Corros., Vol 54, 2003, p 881 N.G. Taylor, V. Guttmann, and R.C. Hurst, The Creep Ductility and Fracture of Carburized Alloy 800H at High Temperatures, in High Temperature Alloys: Their Exploitable Potential, J.B. Marriott, M. Merz, J. Nihoul, and J. Ward, Ed., Elsevier Applied Science, 1987, p 475 R.F. Hochman, Proc. Fourth International Congress on Metal Corrosion, NACE, 1972, p 258 R.F. Hochman, Proc. Symposium on Properties of High Temperature Alloys with Emphasis on Environmental Effects, Z.A. Foroulis and F.S. Pettit, Ed., The Electrochemical Society, 1977, p 715 J.C. Nava Paz and H.J. Grabke, Metal Dusting, Oxid. Met., Vol 39, 1993, p 437 H.J. Grabke, R. Krajak, E.M. MullerLorenz, Metal Dusting of High Temperature Alloys, Werkst. Korros., Vol 44, 1993, p 89 H.J. Grabke, C.B. Bracho-Troconis, and E.M. Muller-Lorenz, Metal Dusting of Low Alloy Steels, Werkst. Korros., Vol 45, 1994, p 215 H.J. Grabke, Metal Dusting of Low- and High-Alloy Steels, Corrosion, Vol 51, 1995, p 711 H.J. Grabke, R. Krajak, E.M. MullerLorenz, and S. Straub, Metal Dusting of Nickel-Base Alloys, Mater. Corros., Vol 47, 1996, p 495 H.J. Grabke, Thermodynamics, Mechanisms and Kinetics of Metal Dusting, Mater. Corros., Vol 49, 1998, p 303 H.J. Grabke, Corrosion by Carbonaceous Gases, Carburization and Metal Dusting, and Methods of Prevention, Mater. High Temp., Vol 17 (No. 4), 2000, p 483 H.J. Grabke, Metal Dusting, Mater. Corros., Vol 54, 2003, p 736 H.J. Grabke and G. Tauber, Arch. Eisenhüttenwes., Vol 46, 1975, p 215

pg 144

Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/

26/10/2007 12:24PM Plate # 0

Chapter 5: Carburization and Metal Dusting / 145

77. S.R. Shatynski and H.J. Grabke, Arch. Eisenhüttenwes., Vol 49, 1978, p 129 78. F. Bonnet, F. Ropital, Y. Berthier, and P. Marcus, Filamentous Carbon Formation Caused by Catalytic Metal Particles from Iron Oxide, Mater. Corros., Vol 54, 2003, p 870 79. E. Pippel, J. Woltersdorf, and H.J. Grabke, Microprocesses of Metal Dusting on IronNickel Alloys and Their Dependence on the Alloy Composition, Mater. Corros., Vol 54, 2003, p 747 80. F. Eberle and R.D. Wylie, Corrosion, Vol 15 (No. 12), 1959, p 622t 81. W.B. Hoyt and R.H. Caughey, Corrosion, Vol 15 (No. 12), 1959, p 627t 82. O.J. Dunmore, A Case History of a Metal Dusting Problem Which Led to a Boiler Failure, presented at UK Corrosion/82 (London), Nov 16–18, 1982 83. H.J. Grabke and M. Spiegel, Mater. Corros., Vol 54, 2003, p 799 84. G.Y. Lai, J. Met., Vol 37 (No. 7), 1985, p 14 85. E. Fitzer, W. Frohs, and M. Heine, Carbon, Vol 24 (No. 4), 1986, p 387 86. M. Maier and J.F. Norton, Studies Concerned with the Metal Dusting of Fe-Cr-Ni Materials, Paper No. 75, Corrosion/99, NACE International, 1999 87. R.A. Perkins, R.A. Padgett, and N.K. Tunali, Met. Trans. AIME, Vol 4, 1973, p 2535 88. P.J. Albery and C.W. Haworth, Met. Sci., Vol 8, 1974, p 407 89. A.F. Smith, Met. Sci., Vol 9, 1975, p 375, 425 90. C.H. Toh, P.R. Munroe, and D.J. Young, Metal Dusting of Fe-Cr and Fe-Ni-Cr Alloys under Cyclic Conditions, Oxid. Met., Vol 58 (No. 1/2), 2002, p 1 91. H.J. Grabke, E.M. Muller-Lorenz, S. Strauss, E. Pippel, and J. Woltersdorf, Effects of Grain Size, Cold Working, and Surface Finish on the Metal-Dusting Resistance of Steels, Oxid. Met., Vol 50 (No 3/4), 1998, p 241 92. A.S. Fabiszewski, W.R. Warkins, J.J. Hoffman, and S.W. Dean, The Effect of Temperature and Gas Composition on the Metal Dusting Susceptibility of Various

Alloys, Paper No. 532, Corrosion/2000, NACE International, 2000 93. S. Strauss and H.J. Grabke, Mater. Corros., Vol 49, 1998, p 321 94. J. Klower, H.J. Grabke, E.M. MullerLorenz, and D.C. Agarwal, Metal Dusting and Carburization Resistance of NickelBase Alloys, Paper No. 139, Corrosion/97, NACE International, 1997 95. K. Natesan and Z. Zeng, Metal Dusting Performance of Structural Alloys, Paper No. 5409, Corrosion/2005, NACE International, 2005 96. D.L. Klarstrom and H.J. Grabke, The Metal Dusting Behavior of Several High Temperature Alloys, Paper No. 1379, Corrosion/ 2001, NACE International, 2001 97. B.A. Baker and G.D. Smith, Metal Dusting Behavior of High-Temperature Alloys, Paper No. 54, Corrosion/99, NACE International, 1999 98. B.A. Baker and G.D. Smith, Alloy Selection for Environments Which Promote Metal Dusting, Paper No. 257, Corrosion/2000, NACE International, 2000 99. B.A. Baker, G.D. Smith, V.W. Hartmann, L. E. Shoemaker, and S.A. McCoy, NickelBase Material Solutions to Metal Dusting Problems, Paper No. 2394, Corrosion/2002, NACE International, 2002 100. H.J. Grabke, H.P. Martinz, and E.M. Muller-Lorenz, Metal Dusting Resistance of High Chromium Alloys, Mater. Corros., Vol 54, 2003, p 860 101. H.J. Grabke, E.M. Muller-Lorenz, and M. Zinke, Metal Dusting Behaviour of Welded Ni-Base Alloys with Different Surface Finish, Mater. Corros., Vol 34, 2003, p 785 102. H.J. Grabke and E.M. Muller-Lorenz, Steel Res., Vol 66, 1995, p 252 103. A. Schneider, H. Viefhaus, G. Inden, H.J. Grabke, and E.M. Muller-Lorenz, Mater. Corros., Vol 49, 1998, p 330 104. A. Schneider, H. Viefhaus, and G. Inden, Surface Analytical Studies of Metal Dusting of Iron in CH4-H2-H2S Mixtures, Mater. Corros., Vol 51, 2000, p 338 105. A. Schneider and H.J. Grabke, Effect of H2S on Metal Dusting of Iron, Mater. Corros., Vol 54, 2003, p 793

pg 145

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:25PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p147-200 DOI: 10.1361/hcma2007p147

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 6

Corrosion by Halogen and Hydrogen Halides 6.1 Introduction Many metals react readily with halogen gases at elevated temperatures to form volatile metal halides. Some metal halides also exhibit low melting points, and some even sublime at relatively low temperatures. As a result, alloys containing elements that form volatile or lowmelting-point halides may suffer severe hightemperature corrosion. Industrial environments often contain halogen gases. Because of high vapor pressures of many metal chlorides, the chlorination process is an important step in processing metallurgical ores for production of titanium, zirconium, tantalum, niobium, and tungsten (Ref 1–3). Chlorination is also used for extraction of nickel from iron laterites (Ref 4) and for detinning of tin plate (Ref 5). Production of TiO2 and SiO2 involves processing environments containing Cl2 and/or HCl, along with O2 and other combustion products. Calcining operations in the production of (a) lanthanum, cerium, and neodymium for electronic and magnetic materials and (b) ceramic ferrites for permanent magnets, frequently generate environments contaminated with chlorine. In the chemical process industry, many processing streams also contain chlorine. Manufacturing of ethylene dichloride (EDC), which is an intermediate for the production of vinyl chloride monomer, generates chlorine-bearing environments. The reactor vessels, calciners, and other processing equipment for the above operations require alloys that are resistant to high-temperature chloridation attack. In the manufacture of ﬂuorine-containing compounds, such as ﬂuorocarbon plastics, refrigerants, and ﬁre-extinguishing agents, the processing equipment requires alloys with good resistance to corrosion by ﬂuorine and hydrogen ﬂuorides at elevated temperatures. During the

reﬁning operation in the production of uranium, UO2 is ﬂuorinated at elevated temperatures (e.g., 500 to 600 °C) with HF to produce UF4 or UF6 for separation of U235 (Ref 6). The materials of construction for this processing equipment must resist corrosion by HF at both high and low temperatures. This chapter reviews data on primarily gaseous corrosion by halogen and hydrogen halides. In many high-temperature industrial processes where fuels and/or feedstocks are often contaminated with impurities, such as alkaline metals, halogen may react readily with these metals to form halide salts. Corrosion reactions under these conditions are to be discussed in later chapters dealing with high-temperature corrosion in gas turbines, coal-ﬁred boilers, oil-ﬁred boilers, waste-to-energy boilers, black liquor recovery boilers, and so forth.

6.2 Thermodynamic Considerations The relative stabilities of various chlorides, ﬂuorides, bromides, and iodides are presented in Fig. 6.1 to 6.3, in terms of standard free energies of formation (∆G°) versus temperature (Ref 7). The ﬁgures also include the information about the melting points and boiling points of halides. Some of the halides exhibit low melting and boiling points. Tables 6.1 to 6.4 list melting and boiling points of some important halides and oxyhalides (Ref 8, 9). Since industrial environments typically have ﬁnite oxygen activities, the thermodynamic phase stability diagram in terms of the M-O-Cl (M represents metal, O represents oxygen, and Cl represents chlorine) system is quite useful in describing the possible corrosion products that may form on the metal. These diagrams can be constructed for major alloying

pg 147

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:25PM Plate # 0

148 / High-Temperature Corrosion and Materials Applications

elements of high-temperature alloys (e.g., Fe, Ni, Co, Cr, Mo, W, Al, Si, etc.). Commercial computer programs, such as HSC (Ref 10) and Chemsage (Ref 11), are available for construction of these phase-stability diagrams. Figure 6.4 shows a Ni-O-Cl stability diagram in terms of pO2 and pCl2 at 723 °C (1333 °F) (Ref 12). The

Fig. 6.1

diagram deﬁnes the boundary ( pO2 10715 atm) between Ni and NiO, the boundary (pCl2 1078 ) between Ni and NiCl2, and the boundary between NiO and NiCl2. This means that if the equilibrium gas mixture of the environment exhibits an oxygen potential ( pO2 ) higher than about 10−15 atm, NiO may form. Similarly,

Standard free energies of formation for chlorides. Source: Ref 7

pg 148

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:25PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 149

NiCl2 may form if the environment exhibits a chlorine potential ( pCl2 ) higher than about 10−8 atm when the oxygen potential ( pO2 ) is below about 10−15 atm. With phase-stability diagrams, one can predict the possible phases that are likely to form on

Fig. 6.2

a metal. Let’s consider an example where nickel is exposed to an environment consisting of air with 0.1% Cl2 at 723 °C (1333 °F). The environment would be at the location that identiﬁes pO2 being very close to 100 and pCl2 being 10−3 in the 723 °C Ni-O-Cl stability diagram (Fig. 6.4).

Standard free energies of formation for fluorides. Source: Ref 7

pg 149

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:25PM Plate # 0

150 / High-Temperature Corrosion and Materials Applications

Based on the stability diagram shown in Fig. 6.4, the environment is in the NiO regime. NiO oxide is expected to form on the surface of nickel when exposed to this environment. If NiO oxide scale formed on nickel is defect free, the pCl2 at the interface between NiO and Ni would

Fig. 6.3

be below the value needed to form NiCl2, thus preventing the formation of NiCl2. However, when defects and cracks develop in the NiO scale, Cl2 can permeate through the oxide scale and reach the nickel with high enough pCl2 to form NiCl2, initiating chloridation attack. The

Standard free energies of formation for bromides and iodides. Source: Ref 7

pg 150

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:25PM Plate # 0

pg 151

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 151

NiCl2 formed at 723 °C (1333 °F) would be a solid phase. As shown in Table 6.1, NiCl2 melts at relatively high temperature (1030 °C) compared with other metal chlorides, such as CoCl2 (740 °C), FeCl2 (676 °C), and FeCl3 (303 °C). The phase-stability diagram for Co-O-Cl at 723 °C (1333 °F) is shown in Fig. 6.5 (Ref 12), and those for Cr-O-Cl and Fe-O-Cl at 600 °C (1112 °F) are shown in Fig. 6.6 and 6.7, respectively (Ref 13). In addition to low melting points, some chlorides have low boiling points also. In Fig. 6.7, FeCl3 is in a gaseous state at 600 °C (1112 °F). Figure 6.8 shows WCl4 and WO2Cl2 in a gaseous state at 900 °C (1650 °F). Some ﬂuorides, bromides, and iodides also have either low melting points or low boiling points (see Tables 6.2–6.4). Corrosion reactions can be signiﬁcantly increased when the corrosion product is in either a liquid state or a gaseous state. Furthermore, many halides, although in a solid state, may exhibit high vapor pressures. When the corrosion product in a solid state exhibits a

Table 6.2 Melting points, temperatures at which fluoride vapor pressure reaches 10−4 atm, and boiling points of various fluorides Fluorides

Melting point, °C (°F)

Temperature at 10−4 atm, °C (°F)

Boiling point, °C (°F)

FeF2 FeF3 NiF2 CoF2 CrF2 CrF3 CuF MoF5 MoF6 WF6 TiF3 TiF4 AlF3 SiF4 MnF2 ZrF4 NbF5 HfF4 TaF5 NaF KF LiF MgF2 CaF2 BaF2 ZnF2 PbF2

1020 (1868) 1027 (1881) 1450 (2642) 1250 (2282) 894 (1641) 1404 (2559) 908 (1666) 64 (147) 17 (63) 2 (36) 1200 (2192) … … −90 (−130) 920 (1688) 932 (1710) 79 (174) … 97 (207) 992 (1818) 857 (1575) 848 (1558) 1263 (2305) 1418 (2584) 1290 (2354) 875 (1607) 822 (1512)

906 (1663) 673 (1243) 939 (1722) 962 (1764) 928 (1702) 855 (1571) … 24 (75) −82 (−116) −91 (−132) … 108 (226) 825 (1517) −160 (−256) 992 (1818) 583 (1081) … 615 (1139) 37 (99) 928 (1702) 788 (1450) 908 (1666) 1257 (2295) 1429 (2604) 1581 (2878) 806 (1483) 664 (1227)

1800 (3272) … … … … … … … 36 (97) 17 (63) … 283 (541) 1270 (2318) −95 (−139) … … 233 (451) … … 1704 (3099) 1502 (2736) 1681 (3058) 2230 (4046) 2500 (4532) 2215 (4019) 1500 (2732) 1293 (2359)

Source: Ref 8, Ref 9

Table 6.1 Melting points, temperatures at which chloride vapor pressure reaches 10−4 atm, and boiling points of various chlorides Chlorides

Melting point, °C (°F)

Temperature at 10 −4 atm, °C (°F)

Boiling point, °C (°F)

FeCl2 FeCl3 NiCl2 CoCl2 CrCl2 CrCl3 CrO2Cl2 CuCl MoCl5 WCl5 WCl6 TiCl2 TiCl3 TiCl4 AlCl3 SiCl4 MnCl2 ZrCl4 NbCl5 NbCl4 TaCl5 HfCl4 CCl4 NaCl KCl LiCl MgCl2 CaCl2 BaCl2 ZnCl2 PbCl2

676 (1249) 303 (577) 1030 (1886) 740 (1364) 820 (1508) 1150 (2102) −95 (−139) 430 (806) 194 (381) 240 (464) 280 (536) 1025 (1877) 730 (1346) −23 (−9.4) 193 (379) −70 (−94) 652 (1206) 483 (901) 205 (401) … 216 (421) 434 (813) −24 (−11) 801 (1474) 772 (1422) 610 (1130) 714 (1317) 772 (1422) 962 (1764) 318 (604) 498 (928)

536 (997) 167 (333) 607 (1125) 587 (1089) 741 (1366) 611 (1132) … 387 (729) 58 (136) 72 (162) 11 (52) 921 (1690) 454 (849) −38 (−36) 76 (169) −87 (−125) 607 (1125) 146 (295) … 239 (462) 80 (176) 132 (297) −80 (−112) 742 (1368) 706 (1303) 665 (1229) 663 (1225) 1039 (1902) … 349 (660) 484 (903)

1026 (1879) 319 (606) 987 (1809) 1025 (1877) 1300 (2372) 945 (1733) 117 (243) 1690 (3074) 268 (514) … 337 (639) … 750 (1382) 137 (279) … 58 (136) 1190 (2174) … 250 (482) 455 (851) 240 (464) … 77 (171) 1465 (2669) 1407 (2565) 1382 (2520) 1418 (2584) 2000 (3632) 1830 (3326) 732 (1350) 954 (1749)

Source: Ref 8, Ref 9

Table 6.3 Melting points, temperatures at which bromide vapor pressure reaches 10−4 atm, and boiling points of various bromides Bromides

Melting point, °C (°F)

Temperature at 10−4 atm, °C (°F)

Boiling point, °C (°F)

FeBr2 FeBr3 NiBr2 CoBr2 CrBr2 CrBr3 CrBr4 CuBr WBr5 WBr6 AlBr3 SiBr4 MnBr2 ZrBr4 NbBr5 HfBr4 TiBr4 TaBr5 NaBr KBr LiBr MgBr2 CaBr2 BaBr2 ZnBr2 PbBr2

689 (1272) … 965 (1769) 678 (1252) 842 (1548) >800 (1472) … 488 (910) 276 (529) 309 (588) 97 (207) 5 (41) 695 (1283) 450 (842) 267 (513) 424 (795) … 267 (513) 750 (1382) 740 (1364) 550 (1022) 710 (1310) 742 (1368) 854 (1569) 398 (748) 373 (703)

509 (948) 156 (313) 580 (1076) … 716 (1321) 615 (1139) 516 (961) 435 (815) … … 53 (127) … … 169 (336) … 137 (279) … 145 (293) 690 (1274) 671 (1240) 630 (1166) 626 (1159) … … 320 (608) 432 (810)

974 (1785) … 919 (1686) … … … … 1318 (2404) … … 255 (491) 153 (307) … … 361 (682) … 232 (450) 347 (657) 1393 (2539) 1383 (2521) 1310 (2390) 1230 (2246) 1800 (3272) … 650 (1202) 914 (1677)

Source: Ref 8, Ref 9

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:25PM Plate # 0

152 / High-Temperature Corrosion and Materials Applications

high vapor pressure, the corrosion reaction can increase as well. It is generally considered that when the vapor pressure of a corrosion product reaches to 10−4 atm, the corrosion reaction can become signiﬁcant. Tables 6.1 to 6.4 also include the temperature at which a halide’s vapor pressure reaches to 10−4 atm. The chlorine partial pressure (pCl2 ) needed to form a chloride corrosion product with 10−4 atm will be lower than that needed to form either solid or liquid chloride. This is illustrated in Fig. 6.9 (Ref 14). Bender and Schutze (Ref 15)

Table 6.4 Melting points, temperatures at which iodide vapor pressure reaches 10−4 atm, and boiling points of various iodides Iodides

Melting point, °C (°F)

Temperature at 10−4 atm, °C (°F)

Boiling point, °C (°F)

FeI2 NiI2 CoI2 CrI2 CrI3 CuI AlI3 SiI4 MnI2 ZrI4 NbI4 HfI4 TaI5 NaI KI LiI MgI2 CaI2 BaI2 ZnI2 PbI2

594 (1101) 780 (1436) 515 (959) 869 (1596) >600 (1112) 588 (1090) 191 (376) 122 (252) 613 (1135) 499 (930) 503 (937) 449 (840) 496 (925) 660 (1220) 685 (1265) 469 (876) 650 (1202) 740 (1364) 712 (1314) 446 (835) 412 (774)

476 (889) … … 702 (1296) … 529 (984) 144 (291) 55 (131) … 227 (441) … 244 (471) 208 (406) 651 (1204) 629 (1164) 621 (1150) 425 (797) … … 316 (601) 397 (747)

935 (1715) … … … … 1207 (2205) 385 (725) 301 (574) … … … … 545 (1013) 1304 (2379) 1330 (2426) 1170 (2138) … … … 730 (1346) 872 (1602)

Source: Ref 8, Ref 9

Fig. 6.4

Phase stability diagram for Ni-O-Cl system at 723 °C (1333 °F). Both corrosion products (NiO and NiCl2) are solid phases at this temperature. Source: Ref 12

Fig. 6.5

Phase stability diagram for Co-O-Cl system at 723 °C (1333 °F). All the corrosion products (i.e., CoO, Co3O4, and CoCl2) are solid phases at this temperature. Source: Ref 12

Fig. 6.6

Phase stability diagram for Cr-O-Cl system at 600 °C (1112 °F). All the corrosion products (i.e., Cr2O3, CrCl2, and CrCl3) are solid phases at this temperature. Source: Ref 13

Fig. 6.7

Phase stability diagram for Fe-O-Cl system at 600 °C (1112 °F). All the corrosion products are solid phases except FeCl3 at this temperature. Source: Ref 13

pg 152

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 153

Fig. 6.8

Phase stability diagram for the W-O-Cl system at 900 °C (1650 °F), showing tungsten chloride (WCl4) and tungsten oxychloride (WO2Cl2) in a gaseous state.

Fig. 6.9

Phase stability diagram for Fe-O-Cl system at 700 °C (1292 °F). The solid line represents the boundary for forming solid FeCl2, while the dotted line represents the boundary for forming FeCl2 with 10−4 atm pressure. Source: Ref 14

extensively examined the phase-stability diagrams involving the vapor pressures of chlorides reaching 10−4 atm for various alloying elements. The authors termed these phase-stability diagrams “quasi-stability diagrams.” Figure 6.10 shows the pCl2 values that are needed to form NiCl2 with 10−4 atm at various temperatures (Ref 15). Also included in the ﬁgure are two environments (air + 0.1% Cl2, and air + 2% Cl2) for illustration purpose. For both of these environments, NiO, not NiCl2 with vapor pressures of 10−4 atm or higher, is likely to form on nickel at 500 and 650 °C. However, at 800 °C

and higher, NiCl2 with vapor pressures of 10−4 atm and higher (not NiO) is to form on nickel. Molybdenum oxychlorides also exhibit high vapor pressures. Figure 6.11 shows the quasistability diagram for Mo-O-Cl system at 500 °C (Ref 15). The ﬁgure also shows that MoO2Cl2 with vapor pressures of 10−4 atm and higher is to form in the environments of air + 0.1% Cl2, and air + 2% Cl2. The metal-halogen reaction differs from other reactions, such as oxidation, in that most reaction products are characteristic of high vapor pressures and, in some cases, low melting points. The volatile halides (reaction products) formed on the metal surface can no longer provide protection against further corrosion. This is in contrast to most oxides, which generally exhibit very low vapor pressures and high melting points. Furthermore, many halides exhibit low melting points. Once the reaction products become molten, the alloy loses all protection against further corrosion, leading to rapid attack.

6.3 Corrosion in Cl2- and HCl-Bearing Environments 6.3.1 Corrosion in Cl2 Environments (No O2) This section focuses on the corrosion behavior of alloys in environments containing essentially Cl2 with no oxygen (O2) present. An excellent review of halogen corrosion data up to the mid-1970s was presented by Daniel and Rapp

pg 153

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

154 / High-Temperature Corrosion and Materials Applications

(Ref 8). They summarized the test results obtained by various authors (Ref 16–19) on chloridation of iron (Table 6.5). It should be noted that the data in Table 6.5 were obtained from short-duration tests, ranging from a few minutes to hours. Use of these test results for extrapolation to 1 year could result in signiﬁcant errors. They should not be used to estimate the service life of equipment, but should instead be used for comparison purposes. As illustrated in Table 6.5, iron exhibited little corrosion attack in Cl2 at temperatures up to 250 °C (480 °F).

0

Above 250 °C, corrosion rates abruptly increased. Iron forms two types of chlorides: FeCl2 and FeCl3. The melting and boiling points of FeCl2 are 676 and 1026 °C (1249 and 1879 °F), respectively. FeCl3, on the other hand, is extremely unstable. Its melting and boiling points are 303 and 319 °C (577 and 606 °F), respectively. Bohlken et al. (Ref 20, 21) suggested that the abrupt increase in the corrosion rate of iron in Cl2 at temperatures above 250 °C (482 °F) was related to the formation of FeCl3.

p(NiCl2) ≥ 10–4 bar

500 °C

–2 650 °C

Log p(Cl2), bar

–4 NiCl2

850 °C

–6

1000 °C

–8 –10

NiO –12

Ni

–14 –30

–25

–20

–15

–10

–5

0

Log p(O2), bar

Fig. 6.10

Quasi-stability diagram for Ni-O-Cl system for NiCl2 with vapor pressures of 10−4 atm (bar) and higher at temperatures from 500 to 1000 °C (932 to 1832 °F). Source: Ref 15

0

p(MoOxCly) ≥ 10–4 bar MoOCl4

Log p(Cl2), bar

–2 –4

MoOCl3

–6 MoCl4 –8

MoO2Cl2

–10

Mo

–12 –40

–35

–30

–25

MoO2 –20

–15

MoO3 –10

–5

0

Log p(O2), bar

Fig. 6.11

Quasi-stability diagram for Mo-O-Cl system for vapor pressures of chlorides and oxychlorides being 10−4 atm (bar) and higher at 800 °C (1472 °F). Source: Ref 15

pg 154

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

pg 155

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 155

Adding chromium and/or nickel to iron improves the alloy’s resistance to chloridation attack. Ferritic and austenitic stainless steels can resist chloridation attack at higher temperatures than cast iron and carbon steels. Brown et al. (Ref 22) reported a corrosion rate of about 600 mpy at 232 °C (450 °F) for both carbon steel and cast iron. A ferritic stainless steel (Fe-17Cr) showed a corrosion rate of about 79 mpy at 360 °C (680 °F), and a titaniumstabilized austenitic stainless steel showed a rate of 24 mpy at 418 °C (784 °F) (Ref 8). The results of studies on several stainless steels by Tseitlin and Strunkin (Ref 16) and Brown et al. (Ref 22) were summarized by Daniel and Rapp (Ref 8) and are presented in Table 6.6. Adequate aluminum when added to iron to form aluminum oxide scales is also beneﬁcial in improving the chloridation resistance. Han and Cho (Ref 23) studied corrosion behavior of Fe3Al (Fe-12.11%Al) in Ar-1%Cl2 at 750, 800, and 900 °C using a thermogravimetric method. The alloy behaved similarly at three different temperatures, showing an initial stage of an “incubation” time before the breakaway corrosion showing a drastic weight loss, as shown in Fig. 6.12. The authors observed a thin protective Al2O3 scale during the initial “incubation” stage, and nonprotective oxide scales (Al2O3, Fe2O3) Table 6.5 Temperature, °C (°F)

77 (170) 166 (330) 198 (388) 200 (392) 230 (446) 240 (464) 240 (464) 247 (477) 251 (484) 255 (491) 260 (500) 268 (514) 279 (534) 285 (545) 285 (545) 302 (576) 304 (579) 310 (590) 323 (613) 327 (621) 381 (718) 381 (718) 540 (1004) 540 (1004) 595 (1103) 599 (1110)

and small amounts of FeCl3 and FeCl2 formed at the “breakaway” stage. Also, at the breakaway stage, the specimen showed aluminum depletion at the metal/scale interface. The thin aluminum oxide scale was observed to form on the specimen during heating to the test temperature with argon gas ﬂowing through the test chamber (approximately 2 h) prior to switching to the test gas. The test gas (i.e., Ar-1%Cl2) was found to contain 1 ppm O2. Thus, under the test condition, Al2O3 could form on the metal as well as FeCl2, as shown in Fig. 6.13. Nickel and nickel-base alloys are widely used in chlorine-bearing environments. The corrosion behavior of nickel in chlorine at various temperatures was analyzed by Daniel and Rapp (Ref 8), using the test results of Downey et al. (Ref 24), Tseitlin and Strunkin (Ref 16), and McKinley and Shuler (Ref 25) (see Table 6.7). At temperatures up to 500 °C (930 °F), nickel showed relatively low corrosion rates. Corrosion rates became suddenly and signiﬁcantly higher at temperatures over 500 °C (930 °F). Nickel reacts with chlorine to form NiCl2, which exhibits relatively high melting point (1030 °C) compared to FeCl2 and FeCl3 (676 and 303 °C, respectively). This may be an important factor, making nickel much more resistant to chloridation attack than iron. Brown et al.

Corrosion of iron in chlorine Flow rate, (L/min)

100 100 100 15 100 100 15 100 100 120 15 120 120 15 15 120 120 15 120 120 120 120 15 15 120 120

pCl2, atm

1 1 1 1 1 1 1 I 1 (c) 1 (c) (c) 1 1 (c) (c) 1 (c) (c) (c) (c) (c) (c) (c) (c)

Diluent gas

None None None None None None None None None Ar None He Ar None None He Ar None Ar He Ar He N2 N2 Ar He

Test duration, min

480 0–15, 15–480 0–15, 15–480 360 0–15, 15–480 0–15, 15–480 360 0–15, 15–480 480 … 360 … … 360 60 … … 60 … … … … … … … …

Linear rate constant(a), μm/min −4

3 × 10 3.3 × 10−3, 3.8 × 10−4 5.2 × 10−3, 3.7 × 10−4 2 × 10−4 8.5 × 10−3, 2.2 × 10−4 9.5 × 10−3, 2.4 × 10−4 2 × 10-4 1 × 10−2, 1.9 ×10−4 1.55 0.94 −4 2 × 10 1.78 2.45 −4 4 × 10 20.4 4.04 4.17 3.9 8.94 11.4 40.4 65.3 6.6 1.5 187 624

Corrosion rate(b), mm/yr (mpy)

0.16 (6.3) 0.20 (7.9) 0.19 (7.5) 0.11 (4.3) 0.12 (4.7) 0.13 (5.1) 0.11 (4.3) 0.10 (3.9) 820 (32 in.) 490 (19 in.) 0.11 (4.3 mils) 940 (37 in.) 1,300 (51 in.) 0.21 (8 mils) 11,000 (433 in.) 2,100 (83 in.) 2,200 (87 in.) 1,700 (67 in.) 4,700 (185 in.) 6,000 (236 in.) 21,000 (827 in.) 34,000 (1,338 in.) 3,500 (138 in.) 830 (33 in.) 98,000 (3,858 in.) 328,000 (13,000 in.)

(a) Rate constants are given as metal loss rates, μm/min. (b) Estimated metal loss after 1 yr of exposure. (c) Tests were conducted with chlorine partial pressures up to 0.3 atm and total pressures of 1.0 atm. However, the metal loss rates were extrapolated to 1.0 atm chlorine pressure. Source: Ref 8

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

pg 156

156 / High-Temperature Corrosion and Materials Applications

Table 6.6

Corrosion of stainless steels in chlorine(a)

Alloy

Ferritic stainless (Fe-17Cr)

Austenitic stainless (Fe-18Cr-9Ni-Ti)

Austenitic stainless (Fe-18Cr-8Ni-Mo)

Austenitic stainless (Fe-18Cr-8Ni)

Temperature, °C (°F)

Flow rate, L/min

Linear rate constant(b), µm/min

Corrosion rate(c), mm/yr (mpy)

300 (572) 360 (680) 440 (824) 540 (1,004) 418 (784) 450 (842) 480 (896) 535 (995) 640 (1,184) 315 (599) 340 (644) 400 (752) 450 (842) 480 (896) 290 (554) 315 (599) 340 (644) 400 (752) 450 (842)

15 15 15 15 15 15 15 15 15 28 28 28 28 28 28 28 28 28 28

4 × 10−4 3.8 × 10−3 6.7 × 10−2 1.35 1.1 × 10−3 4.3 × 10−2 0.13 0.47 46 1.4 × 10−3 2.9 × 10−3 5.9 × 10−3 2.9 × 10−2 5.9 × 10−2 1.5 × 10−3 2.9 × 10−3 5.9 × 10−3 2.9 × 10−2 5.9 × 10−2

0.2 (7.9) 2 (79) 40 (1.6 in.) 700 (28 in.) 0.6 (24) 20 (787) 70 (2.8 in.) 200 (7.9 in.) 20,000 (787 in.) 0.8 (31) 1.5 (59) 3 (118) 15 (590) 30 (1.2 in.) 0.8 (31) 1.5 (59) 3 (118) 15 (590) 30 (1.2 in.)

(a) Chlorine pressure was approximately 1.0 atm. (b) Duration of these tests was 60–360 min for the ﬁrst two alloys and 120–1200 min for the last two alloys. (c) Estimated metal loss after one year of exposure. Source: Rcf 8

5

10 1% Cl2/Ar

0

FeCl3

0 –5

Log p Cl , g 2

Weight change, mg/cm2

750 °C

5

–10

–5

FeCl2

–10 AlCl3

–15

–15 900 °C

800 °C

750 °C

Al2O3 Fe3O4

Fe

–20

Fe2O3

Al

–20

–25 –25 0

10

20

30

40

50

60

70

–30 –60

Exposure time, h

Fig. 6.12

Thermogravimetric results for Fe3Al (Fe-12.11%Al) tested in Ar-1%Cl2 at 750, 800, and 900 °C. Source: Ref 23

(Ref 22) conducted short-term laboratory tests in chlorine on various commercial alloys. The results (see Table 6.8) suggested that, in an environment of 100% Cl2, carbon steel and cast iron are useful at temperatures up to 150 to 200 °C (300 to 400 °F) only. The 18-8 stainless steels can be used at higher temperatures—up to 320 to 430 °C (600 to 800 °F). Nickel and nickel-base alloys (e.g., Ni-Cr-Fe, Ni-Mo, and Ni-Cr-Mo alloys) were most resistant. The beneﬁcial effect of nickel on the resistance of chloridation attack in Cl2 environments is

FeO –50

–40

–30

–20

–10

0

10

Log p O , g 2

Fig. 6.13

Phase stability for Al-O-Cl system at 750 °C. The solid circle indicates the test environment. Source:

Ref 23

illustrated in Fig. 6.14 (Ref 26). This trend is also reﬂected in long-term tests (Table 6.9). Alloy 600 is the most commonly used alloy for high-temperature services in Cl2 environments. Figure 6.15 shows the corrosion rates of alloy 600 in dry chlorine gas as a function of temperature (Ref 22). MTI Publication MS-3 (Ref 27) suggests corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304, and steel in dry chlorine gas applications, as shown in Fig. 6.16. Tu et al. (Ref 28) performed phase analysis using x-ray diffraction on the external corrosion

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

pg 157

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 157

Table 6.7

Corrosion of nickel in chlorine pCl2, atm

Temperature, °C (°F)

350 (662) 375 (707) 400 (752) 407 (765) 433 (811) 465 (869) 485 (905) 500 (932) 500 (932) 525 (977) 540 (1004) 550 (1022) 550 (1022) 660 (1220) 740 (1364) 770 (1418)

~1 ~1

~1 ~1 ~1

Test duration, min

0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 atm at 16 cm3/min 0.13 atm at 16 cm3/min 0.13 0.13 atm at 16 cm3/min atm at 16 cm3/min atm at 16 cm3/min

3000–5000 … … … … … … … 3600 … 3600 … … 3600 60 60

Corrosion rate(a), mm/yr (mpy)

0.0012 (0.05)(b) 0.0014 (0.06)(b) 0.0036 (0.14)(b) 0.005 (0.2)(b) 0.0058 (0.2)(b) 0.107 (4.2) 0.142 (5.6) 0.319 (12.6) 0.492 (19.4) 2.41 (95) 2.07 (82) 6.02 (237) 5.85 (230) 40.3 (1,587) 120 (4,724) 2,150 (84,646)

(a) Estimated metal loss after 1 year of exposure to chlorine. (b) These estimates are probably low. Source: Ref 8

Table 6.8 Corrosion of selected alloys in chlorine Approximate temperature, °C (°F), at which given corrosion rate is exceeded Alloy

0.8 mm/yr (30 mpy)

1.5 mm/yr (60 mpy)

3.0 mm/yr (120 mpy)

15 mm/yr (600 mpy)

Nickel Alloy 600 Alloy B Alloy C Chromel A Alloy 400 18-8 Mo 18-8 Carbon steel Cast iron

510 (950) 510 (950) 510 (950) 480 (900) 425 (800) 400 (750) 315 (600) 288 (550) 120 (250) 93 (200)

538 (1000) 538 (1000) 538 (1000) 538 (1000) 480 (900) 455 (850) 345 (650) 315 (600) 175 (350) 120 (250)

593 (1100) 565 (1050) 593 (1100) 565 (1050) 538 (1000) 480 (900) 400 (750) 345 (650) 205 (400) 175 (350)

650 (1200) 650 (1200) 650 (1200) 650 (1200) 620 (1150) 538 (1000) 455 (850) 400 (750) 230 (450) 230 (450)

Source: Ref 22

products formed on Ni-4Cr alloy when exposed in 105 Pa (1 atm) Cl2 at 575 and 700 °C. The authors found that the scales consisted of mainly NiCl2, CrCl3, and CrCl2. The deposits on the quartz test assembly during testing of Ni-4Cr alloy were also analyzed. These deposits were mainly NiCl2 and CrCl3 with very little CrCl2, indicating both NiCl2 and CrCl3 are major vapor phases during testing Ni-4Cr alloy. 6.3.2 Corrosion in O2-Cl2 Environments Many industrial environments may contain both chlorine and oxygen. Metals generally follow a parabolic rate law by forming condensed phases of oxides, if the environment is free of chlorine. With the presence of both oxygen and chlorine, corrosion of metals then involves a combination of condensed oxides and volatile chlorides. Depending on the relative amounts of oxides and chlorides formed, corrosion can

follow either a paralinear rate law (a combination of weight gain due to oxidation and weight loss due to chlorination) or a linear rate law due to chlorination. This is illustrated by the results of Maloney and McNallan (Ref 29) on corrosion of cobalt in Ar-50O2-Cl2 mixtures (Fig. 6.17). As shown in the ﬁgure, at high Cl2 levels, the corrosion products are primarily cobalt chloride vapor, causing the weight loss to follow a linear rate law. Because of volatile corrosion products, the reaction rate can be highly dependent on the gas ﬂow rate. McNallan and Liang (Ref 30) showed that CoO specimens exhibited increased linear weight loss rates with increasing gas velocity when exposed in the Ar-O2-1Cl2 mixtures with pO2 being 0.01 and 0.15 atm pressures at 723 °C (1000 K). Furthermore, the oxygen partial pressure (pO2 ) of 0.01 atm resulted in higher weight loss rates than that of 0.15 atm. This is also illustrated in Fig. 6.18 (Ref 31), showing the corrosion of cobalt in Ar-O2-1Cl2 mixtures with three different concentrations (1, 10, and 50% O2) of oxygen at 650 °C (1200 °F). When the environment contained 10 and 50% O2, the corrosion reaction involved mainly the formation of cobalt oxide, thus following an approximate parabolic rate. When the oxygen concentration reduced to 1%, the corrosion reaction involved mainly volatile CoCl2, thus following an approximate linear weight loss with time. The ﬁgure also shows that the linear weight loss agreed very well with the volatilization of CoCl2. It should be noted that in the above test environments containing 10 and 50% O2, the corrosion reaction, which followed a parabolic rate due to formation of condensed cobalt oxides (Fig. 6.18), involved only a very

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

mg/cm2

158 / High-Temperature Corrosion and Materials Applications

690

Fig. 6.14

Effect of nickel on the corrosion resistance of alloys in Ar-30Cl2 at 704 °C (1300 °F) for 24 h. Source: Ref 26

Table 6.9 Corrosion of several alloys in Ar-30Cl2 after 500 h at 400 to 704 °C (750 to 1300 °F)(a) Descaled weight loss, mg/cm2 Alloy

400 °C (750 °F)

500 °C (930 °F)

600 °C (1110 °F)

704 °C (1300 °F)(a)

Ni-201 600 601 625 617 800 310 304 347

0.2 0.02 0.3 0.7 0.6 6 28 108 215

0.3 5 3 7 7 13 370 1100 Total

47–101 127–180 85–200 … … 200–270 … … …

97 160 215 180 190 890 820 >1000 Total

(a) 24 h test period. Source: Ref 26

short-term test (2 h). It is extremely likely that upon longer exposure times the cobalt oxide scales may eventually crack (or spall), thus allowing chlorine gas to reach the underlying metal and causing formation of volatile CoCl2 and resulting in a linear corrosion rate. Corrosion of pure nickel in O2-Cl2 environments was found to be dominated by the formation of volatile NiCl2 corrosion product, and the weight loss essentially followed a linear rate law (Ref 32, 33). Figure 6.19 shows thermogravimetric results at 927 °C in Ar-50O2 containing various amounts of Cl2 from 0 to 5% (Ref 32). The ﬁgure also shows that the experimental data were in good agreement with the

pg 158

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

/month

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 159

theoretical calculation based on vapor pressures of NiCl2. In their testing of nickel in Ar-50O2-Cl2 containing 0, 0.5, 1.0, 1.75, 3, and 5% Cl2 at 627 to 927 °C, Lee and McNallan (Ref 32) observed a very rapid reaction referred to as “ignition” at 727 and 827 °C in the environments containing higher concentrations of Cl2. McNallan (Ref 33) suggested that this rapid corrosion reaction was caused by the reaction of chlorine with nickel at the metal/oxide scale interface to form nickel chloride vapor, which then diffuses out and is converted into powdery nickel oxide. This rapid corrosion reaction causes metal temperature to increase. He further indicated that this rapid reaction (ignition) can be prevented at temperatures below 727 °C due to low vapor pressures of nickel chloride and can also be prevented at temperatures higher than 827 °C due to formation of a protective oxide scale by rapid oxidation of nickel to reduce the ingress of chlorine through the oxide scale and thus the formation of nickel chloride vapor at the metal/oxide scale interface. In an O2-Cl2 containing environment, oxidation and chloridation can take place. As discussed in section 6.2, a protective oxide scale can lower the pCl2 below the value for forming metal chlorides at the metal/oxide scale interface. However, cracking and spalling of this oxide scale resulting from, for example, thermal

Fig. 6.15

Corrosion rate of alloy 600 in Cl2 as a function of temperature. Source: Ref 22

Fig. 6.16

MTI corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304SS and steel in dry chlorine (Cl2) as a function of temperature. Source: Ref 27. Courtesy of Materials Technology Institute

pg 159

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:26PM Plate # 0

160 / High-Temperature Corrosion and Materials Applications

Fig. 6.17

Corrosion of cobalt in Ar-50O2-Cl2 at 927 °C (1700 °F). Source: Ref 29

Fig. 6.18

Thermogravimetric results of corrosion of cobalt in Ar-O2-1Cl2 with 1, 10, and 50% O2 at 650 °C (1200 °F). Source: Ref 31

cycling can allow permeation of chlorine to reach the metal underneath to initiate chloridation attack. Figure 6.20 illustrates the effect of thermal cycling on the initiation of accelerated chloridation attack for Fe-20Cr at 927 °C in Ar-20O2-0.5Cl2 (Ref 34). The ﬁgure shows the thermogravimetric results for three separate test

Fig. 6.19 Thermogravimetric results of corrosion of nickel at 927 °C in Ar-50O2-Cl2 (0, 0.5, 1.0, 1.75, 3, and 5% Cl2). Source: Ref 32 runs, showing similar behavior in initiating an accelerated chloridation attack right after a thermal cycle, which involved cooling the specimen for 30 min by lowering it from the test temperature (927 °C) to 100 °C in 5 min. As soon

pg 160

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

pg 161

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 161

as the test was resumed after the thermal cycle, specimens (three separate test runs) showed accelerated chloridation attack. A signiﬁcant amount of data have been generated for commercial alloys in environments containing both oxygen and chlorine. Table 6.10 summarizes short-term test results generated in Ar-20O2-2Cl2 at 900 °C (1650 °F) for 8 h (Ref 35). The results revealed several interesting trends. The best-performing alloy was an aluminum-containing alloy 214 (Fe-16Cr-3Fe4.5Al) with a small amount iron. Two worstperforming alloys were cobalt-base alloys (alloys 188 and 6B) containing large amounts of tungsten. Molybdenum-containing nickel-base alloys also did not perform well. Oh et al. (Ref 36) attributed this to the formation of oxychlorides of molybdenum and tungsten, which have very high vapor pressures. The partial pressures of WO2Cl2 and MoO2Cl2 in equilibrium with the oxides (WO3 and MoO3, respectively) and the test environment (Ar-20O2-2Cl2) at 900 °C (1650 °F) were 7.52 ×10−2 and 2.1× 100 atm, respectively. Accordingly, alloy 188 (14% W), alloy C-276 (16% Mo, 4% W), alloy 6B (4.5% W, 1.5% Mo), alloy X (9% Mo), and alloy S (14.5% Mo) suffered relatively high rates of corrosion attack. Simple Fe-Ni-Cr (Type 310 SS) and Ni-Cr-Fe (alloy 600) performed better than molybdenum- or tungsten-containing alloys. The 100

Mass change, mg/cm2

80 36 h 60

18 h

40

gravimetric results for representative alloys are summarized in Fig. 6.21 (Ref 36). Thermodynamic phase stability diagrams showing high vapor pressures of oxychlorides of tungsten and molybdenum were presented earlier in Fig. 6.8 and 6.11, respectively. In order to determine whether molybdenum oxychlorides would contribute to high corrosion rates in high-molybdenum-containing nickel alloys, such as alloy S (Ni-16Cr-14.5Mo) in O2-Cl2 environments, Jacobson et al. (Ref 37) used a high-pressure sampling mass spectrometer to measure volatile species produced from the preoxidized specimen of alloy S with 14.5% Mo in comparison with alloy 600 (Ni-16Cr-9Fe) with no Mo during the exposure of Ar-50O21Cl2. Thermogravimetric data for these two preoxidized alloys under the test condition are shown in Fig. 6.22, showing a signiﬁcantly higher weight loss rate for alloy S (14.5Mo) than alloy 600 (no Mo) during the exposure of the preoxidized specimens to Ar-50O2-1Cl2 at 900 °C (1650 °F). The mass spectrometer results indicated that MoO2Cl2 along with NiCl2 and CrO2Cl2 were major vapor phases in the case of alloy S. For alloy 600, NiCl2 and CrO2Cl2 were detected. Alloy R-41 (nickel-base alloy with 1.5% Al, 3% Ti and 10% Mo) suffered less chloridation attack than other nickel-base alloys containing molybdenum despite high molybdenum content in a short-term test presented in Table 6.10. However, the results of long-term tests in Ar-20O2-0.25Cl2 by Rhee et al. (Ref 38) and McNallan et al. (Ref 39) showed that these nickel-base alloys with molybdenum, such as R-41 (Ni-19Cr-11Co-10Mo-1.5Al-3Ti) and alloy 263 (Ni-20Cr-20Co-5.8Mo-0.5Al-2.2Ti), eventually suffered severe attack despite the presence of aluminum and titanium. Figure 6.23

24 h

Table 6.10 Corrosion of selected alloys in Ar-20O2-2Cl2 at 900 °C (1650 °F) for 8 h

20 Pure oxygen 0

Alloy

4

8

12

16

20

24

28

32

36

Time, h

Fig. 6.20

Thermogravimetric results of three test runs for Fe20Cr alloy tested in Ar-20O2-0.5Cl2 at 927 °C isothermally for the first 12 h, followed by a thermal cycle by cooling the specimen to 100 °C for 30 min and raising the specimen to the test temperature to resume testing. Note that the thermal cycle resulted in the initiation of an accelerated chloridation attack. Source: Ref 34

214 R-41 600 310SS S X C-276 6B 188

Metal loss, mm (mils)

Average metal affected(a), mm (mils)

0 0.004 (0.16) 0.012 (0.48) 0.012 (0.48) 0.053 (2.08) 0.020 (0.80) 0.079 (3.12) 0.014 (0.56) 0.014 (0.56)

0.012 (0.48) 0.028 (1.12) 0.035 (1.36) 0.041 (1.60) 0.063 (2.48) 0.071 (2.80) 0.079 (3.12) 0.098 (3.84) 0.116 (4.56)

(a) Metal loss + average internal penetration. Source: Ref 35

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

pg 162

162 / High-Temperature Corrosion and Materials Applications

310SS

Fig. 6.21

Gravimetric results for selected Fe-, Ni-, and Co-base alloys in Ar-20O2-2Cl2 at 900 °C (1650 °F). Source: Ref 36

Fig. 6.22

Thermogravimetric data showing weight loss of two preoxidized specimens during the exposure in Ar-50O2-1Cl2 at 900 °C (1650 °F). Source: Ref 37

shows corrosion test results for aluminumcontaining nickel-base alloys with and without molybdenum, such as 214, 601, R-41, and 263,

tested at 900 °C (1650 °F) (Ref 39). Test results for all the alloys tested at 900 and 1000 °C (1650 and 1830 °F) are summarized in Tables 6.11 and 6.12 (Ref 39). The beneﬁcial effect of aluminum, as well as the detrimental effect of molybdenum and tungsten, on the resistance to chloridation attack in oxidizing environments is further substantiated by the results of long-term tests in another environment with a higher concentration of Cl2, as shown in Fig. 6.24 (Ref 35). Similar results were obtained by Elliott et al. (Ref 40) from tests conducted in air-2Cl2 at 900 °C (1650 °F) for 50 h (Fig. 6.25). Chloridation attack in O2-Cl2 environments at these high temperatures primarily consisted of metal wastage and internal penetration for most alloys, with the exception of Ni-Cr-Mo alloys containing high levels of molybdenum, such as alloys S and C-276, which showed mainly metal wastage with little or no internal penetration. In general, the scales formed on the alloy surface were loose when the test specimens were cooled to room temperature after the exposure test, as illustrated in Fig. 6.26. Scales were basically

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

pg 163

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 163

oxides, as shown in Fig. 6.27 and 6.28. Fe-Cr oxides were found to form on Type 310 SS, while aluminum-rich oxides were the major oxide phase along with little nickel-rich oxides that formed on alloy 214. Internal attack consisted of voids for some alloys and of oxides for other alloys. Some alloys appeared to contain both internal voids and oxides. Figure 6.29 shows internal void formation in alloy R-41 (Ni-Cr-CoMo-Al-Ti) and alloy 25 (Co-Cr-Ni-W) after exposure for 50 h at 900 °C (1650 °F) in air2Cl2. It was suggested that the internal penetration may involve halogen-carbide reactions and simultaneous void formation (Ref 25). Figure 6.29 shows some evidence of carbides being converted into voids during chloridation attack. Some alloys, however, showed internal oxides instead of voids, as illustrated in Fig. 6.30. Elliott et al. (Ref 40) have identiﬁed volatile species of the condensed products removed from the exit end of the test apparatus during their investigation in air-2Cl2. Their results are shown in Table 6.13. No oxychlorides were detected. Since the analysis of the volatile species was performed on the condensed phases collected at the exit end of the test apparatus, oxychlorides were likely to be in a gaseous state at the exit end,

and thus were not collected. As discussed earlier, Jacobson et al. (Ref 37) used a high-pressure sampling mass spectrometer to measure volatile species from alloy S with 14.5% Mo and alloy 600 (Ni-16Cr-9Fe) with no Mo during the exposure of Ar-50O2-1Cl2 at 900 °C (1650 °F) and found that MoO2Cl2 along with NiCl2 and CrO2Cl2 were major vapor phases for alloy S and NiCl2 and CrO2Cl2 for alloy 600. McNallan et al. (Ref 39) reported corrosion behavior in Ar-20O2-0.25Cl2 at 900 and 1000 °C (1650 and 1830 °F). This was followed by a study (Ref 41) using the same environment to investigate the same alloys at lower temperatures (i.e., 700, 800, and 850 °C). The results of the tests at 700, 800, and 850 °C (1290, 1470, and 1560 °F) are summarized in Table 6.14 Table 6.11 Corrosion of various alloys in Ar-20O2-0.25Cl2 for 400 h at 900 and 1000 °C (1650 and 1830 °F) Weight loss, mg/cm2 Alloy

214 601 600 800H 310SS 556 X 625 R-41 263 188 S C-276

900 °C (1650 °F)

1000 °C (1830 °F)

4.28 20.67 72.08 26.91 47.15 40.29 54.41 99.07 63.83 82.57 139.77 228.21 132.05

9.05 124.99 254.96 87.05 97.40 82.74 153.49 220.09 207.32 229.53 156.30 248.98 298.85

Source: Ref 39

Table 6.12 Depth of attack after 400 h at 900 and 1000 °C (1650 and 1830 °F) in Ar-20O2-0.25Cl2 900 °C (1650 °F) Alloy

Fig. 6.23

Corrosion of several aluminum-containing nickelbase alloys with and without molybdenum in Ar-20O2-0.25Cl2 at 900 °C (1650 °F). Source: Ref 39

214 601 600 800H 310SS 556 X 625 R-41 263 188 S C-276

Metal loss, mm (mils)

Total depth(a), mm (mils)

0.023 (0.9) 0.150 (5.9) 0.061 (2.4) 0.264 (10.4) 0.127 (5.0) 0.252 (9.9) 0.043 (1.7) 0.191 (7.5) 0.086 (3.4) 0.152 (6.0) 0.046 (1.8) 0.152 (6.0) 0.099 (3.9) 0.218 (8.6) 0.208 (8.2) 0.272 (10.7) 0.114 (4.5) 0.244 (9.6) 0.130 (5.1) 0.193 (7.6) 0.216 (8.5) >0.356 (14.0) 0.315 (12.4) 0.353 (13.9) 0.300 (11.8) 0.320 (12.6)

1000 °C (1830 °F) Metal loss, mm (mils)

Total depth(a), mm (mils)

0.013 (0.5) 0.203 (8.0) 0.330 (13.0) 0.203 (8.0) 0.191 (7.5) 0.152 (6.0) 0.318 (12.5) 0.356 (14.0) 0.381 (15.0) 0.368 (14.5) 0.254 (10.0) 0.419 (16.5) 0.419 (16.5)

0.051 (2.0) 0.295 (11.6) 0.386 (15.2) 0.424 (16.7) 0.246 (9.7) 0.300 (11.8) 0.434 (17.1) 0.437 (17.2) 0.457 (18.0) 0.424 (16.7) 0.417 (16.4) 0.472 (18.6) 0.450 (17.7)

(a) Metal loss + internal penetration. Source: Ref 39

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

164 / High-Temperature Corrosion and Materials Applications

Fig. 6.24

Corrosion of several nickel- and cobalt-base alloys in Ar-20O2-1Cl2 at 900 °C (1650 °F). Source: Ref 35

Fig. 6.25

Corrosion of several iron- and nickel-base alloys in air-2Cl2 at 900 and 1000 °C (1650 and 1830 °F) for 50 h. Source: Ref 40

(Ref 41). The corrosion behavior of alloys as a function of temperature from 700 to 1000 °C (1290 to 1830 °F) can best be summarized in Fig. 6.31 using three different alloy systems (Ni-Cr-Mo alloys S, Fe-Ni-Cr alloy 800H, and Ni-Cr-Al alloy 214). As discussed earlier, refractory metals, such as molybdenum and tungsten, are detrimental to chloridation resistance

in oxidizing environments at high temperatures. Alloy S was found to be less corrosion resistant than alloy 800H. However, both alloys S and 800H suffered increasing corrosion attack with increasing temperatures. This represents a typical trend for most alloys in oxidizing environments. One exception is the Ni-Cr-Al system. As illustrated in Fig. 6.31, alloy 214 showed a sudden decrease in corrosion attack as the test temperature was increased from 900 to 1000 °C (1650 to 1830 °F). This sharp reduction in corrosion attack at 1000 °C (1830 °F) was attributed to the formation of a protective Al2O3 scale. At lower temperatures, such as 900 °C or less, the kinetics of Al2O3 formation was not fast enough to form a protective oxide scale in O2-Cl2 environments. Formation of a protective Al2O3 scale is favored at higher temperatures (e.g., 1000 °C or higher). Most of the data presented so far were generated at fairly high temperatures (i.e., 900 °C and higher). Chloridation was quite aggressive at those high temperatures. Schwalm and Schutze (Ref 42–44) investigated a large number of commercial alloys at signiﬁcantly lower temperatures, varying from 300 to 800 °C (572 to 1472 °F) for times up to 300 h in air-2Cl2. The corrosion behavior of various alloys were generated in terms of the decrease in specimen crosssection thickness (i.e., metal loss) and the depth of internal attack. Extensive characterization of the corrosion products formed on the alloys was

pg 164

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 165

investigated using SEM/EDX analyses, which showed distribution of elements including oxygen and chlorine in the corrosion products. Their results are brieﬂy summarized below. Figure 6.32(a) shows the decrease in metal cross-section thickness for 2.25Cr-1Mo steel (10CrMo9 10), alloy 800H, alloy AC66, alloy 45TM, and alloy 690 after testing for 300 h as

Fig. 6.26

a function of temperature. The data on the depth of internal corrosion attack are presented in Fig. 6.32(b). The corrosion attack at 300 °C (572 °F) was quite negligible after 300 h for the alloys tested including 2.25Cr-1Mo steel. Nevertheless, the Cr-Mo steel was found to exhibit a fragile oxide scale, which contained Fe, O, and Cl. All other alloys including two

Loose scales on samples of several nickel-base alloys after testing at 900 °C (1650 °F) in Ar-20O2-1Cl2 for 100 h

Semiquantitative EDX analysis, wt% Area

1 2

Fig. 6.27

Fe

Cr

Si

85.1 69.5

14.9 29.9

… 0.6

Scanning electron micrograph showing oxide scale formed on Type 310SS sample exposed at 900 °C (1650 °F) for 400 h in Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1 and No. 2, are tabulated. Magnification bar represents 33.3 µm

pg 165

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

166 / High-Temperature Corrosion and Materials Applications

nickel alloys (alloys 45TM and 690) showed pitting type of attack. The phase in the pit was heavily enriched in Cl and O with some Cr, Fe, and Ni. Figure 6.33 shows the morphology of the pit along with an x-ray map for Cl. At 500 °C (932 °F), 2.25Cr-1Mo steel suffered both signiﬁcant thickness reduction and internal attack, while alloys AC66, 800H, 45TM, and 690 showed little attack. At 650 °C, alloy AC66 suffered signiﬁcantly more thickness loss than alloys 800H, 45TM, and 690. The elemental distribution in the corrosion products for alloy AC66 (worst alloy in this group) is shown in Fig. 6.34. The chromium oxides with a layer of iron-rich oxide that formed on AC66 became convoluted. Chlorine was detected at the metal/ oxide scale interface and in the metal underneath the metal/oxide scale interface, where internal attack was observed. Alloy 690, on the other hand, showed a continuous chromium oxide scale. No chlorine was detected. Some internal

attack was observed underneath the oxide scale, and these internal particles are believed to be aluminum oxides. The morphology of the corrosion products formed in alloy 690 is shown in Fig. 6.35. Corrosion behavior for alloys 59, C-2000, and HR160 after testing for 300 h as a function of temperature is shown in Fig. 6.36. All three alloys exhibited little corrosion attack at 300 and 500 °C. At 650 °C, both alloys 59 and HR160 continued to exhibit little corrosion attack, while alloy C-2000 suffered much more corrosion attack. Both C-2000 (Ni-23Cr-16Mo-1.6Cu) and 59 (Ni-23Cr-16Mo-0.3Al) exhibit similar chemical compositions except C-2000 contains additional 1.6% Cu and alloy 59 contains additional 0.3Al. It is not clear whether Cu in alloy C-2000 was responsible. At 650 °C, both alloys 59 and HR160 were found to perform better than alloys 690 and 45TM (Fig. 6.32). Alloy 59 was found to exhibit a thin, continuous chromium-rich oxide

Semiquantitative EDX analysis, wt%

Fig. 6.28

Area

Ni

Cr

Al

Fe

1 2

72 6

20 5

4 88

4 1

Scanning electron micrograph showing oxide scale formed on alloy 214 sample tested at 900 °C (1650 °F) for 400 h in Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1 and No. 2, are tabulated. Magnification bar represents 33.3 µm

pg 166

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 167

scale after exposure at 800 °C for 300 h in air2Cl2. as shown in Fig. 6.37. Also observed was an outer Fe-, Ni-, and Cr-rich oxide scale. Also observed were a small amount of chlorine and slight molybdenum enrichment at the metal/ chromium-oxide scale interface. At 800 °C, all three alloys showed higher corrosion attack. Alloy 59 continued to perform the best. The

Fig. 6.29

high concentration of molybdenum (16%) showed no detrimental effect on the alloy’s corrosion resistance in this oxidizing environment containing 2% Cl2. Thermodynamically, this environment at 800 °C, molybdenum oxychloride (MoO2Cl2) would be stable, as shown in Fig. 6.11. Molybdenum oxychloride was believed to contribute to high corrosion rates for

Scanning electron micrographs showing internal void formation in (a) alloy R-41 and (b) alloy 25 after exposure for 50 h at 900 °C (1650 °F) in air-2Cl2

pg 167

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:27PM Plate # 0

pg 168

168 / High-Temperature Corrosion and Materials Applications

high-molybdenum-containing alloys, such as alloys C-276, S, and so forth when exposed to O2-Cl2 environments at 900 °C (1650 °F) and higher (Ref 35, 39, 40). The chloride phases exhibiting 10−4 atm pressure that are predicted thermodynamically under this test condition are NiCl2, CoCl2, CrO2Cl2, FeCl3, and MoO2Cl2 (Table 6.1), while the oxides are Al2O3, SiO2, Cr2O3, and Fe2O3. HR160 alloy showed an inner

layer of silicon-rich oxide scale underneath the chromium-rich oxide scale with slight internal silicon oxides after exposure at 800 °C for 300 h in air-2Cl2, as shown in Fig. 6.38. The ﬁgure also shows that silicon oxides contained what appeared to be tiny internal voids. Schwalm and Schutze (Ref 44) also tested alumina former alloy 214 (Ni-16Cr-3Fe-4.5AlY-Zr), intermetallic Fe3Al (Fe-5.5Cr-15.9Al0.2Zr), and intermetallic TiAl (Ti-36Al). The corrosion data are shown in terms of crosssection thickness loss (Fig. 6.39a) and internal attack (Fig. 6.39b) after exposure for 300 h at temperatures from 300 to 800 °C (572 to 1472 °F) in air-2Cl2. Alumina former alloy 214 performed very well, comparable to alloy 59 (Fig. 6.36). Fe3Al, while showing little corrosion attack at 300 and 500 °C, suffered severe corrosion at 650 °C. At 800 °C, Fe3Al showed very little corrosion again. The oxide scales formed on Fe3Al at 650 °C consisted of nonprotective, multilayers of Cr-containing Fe2O3, Al2O3, and Fe2O3. At 800 °C, a thin, continuous aluminum-rich oxide scale was found to form on Fe3Al when tested at 800 °C. It is clear that

Semiquantitative EDX analysis, wt%

Table 6.13 Volatile species of condensed products(a) after testing at 900 °C (1650 °F) for 50 h in air-2Cl2

Area

Ni

Cr

Fe

Al

Ti

Alloy

1 2 3 4 5

1 19 25 29 56

62 66 60 45 26

1 7 12 13 18

27 6 2 10 …

9 2 1 3 …

Alloy 214 Alloy 601(b) Type 310SS Alloy 800H(b) Alloy 25 Alloy 625 Alloy 617 Alloy 263 Alloy C-276

Fig. 6.30

Scanning electron micrograph showing oxide scales and internal oxides for alloy 601 exposed at 900 °C (1650 °F) for 400 h in Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1, No. 2, No. 3, No. 4, and No. 5, are listed.

Major constituents

… NiCl2, AlCl3 FeCl3·2H·2O, NiCl2·6H2O FeCl3·2H2O, NiCl2·6H2O CoCl2, NiCl2·6H2O, WCl6 NiCl2·6H2O, FeCl3·2H2O, MoCl5 NiCl2·6H2O, FeCl3·2H2O, CoCl2 NiCl2·6H2O, CoCl2, MoCl5, FeCl3·2H2O NiCl2·6H2O, MoCl5

(a) Collected at the downstream, cooler section of the test apparatus. (b) Tested at 1000 °C (1830 °F) for 50 h. Source: Ref 40

Table 6.14 Depth of attack for various alloys after 400 h at 700, 800, and 850 °C (1290, 1470, and 1560 °F) in Ar-20O2-0.25Cl2 700 °C (1290 °F) Alloy

214 600 800H 310SS 556 S C-276 188 Source: Ref 41

800 °C (1470 °F)

850 °C (1560 °F)

Metal loss, mm (mils)

Total depth, mm (mils)

Metal loss, mm (mils)

Total depth, mm (mils)

Metal loss. mm (mils)

Total depth, mm (mils)

0.010 (0.4) … 0.025 (1.0) … … 0.079 (3.1) 0.033 (1.3) …

0.010 (0.4) … 0.033 (1.3) … … 0.081 (3.2) 0.046 (1.8) …

0.018 (0.7) 0.020 (0.8) 0.023 (0.9) 0.036 (1.4) 0.020 (0.8) 0.145 (5.7) 0.066 (2.6) 0.058 (2.3)

0.061 (2.4) 0.086 (3.4) 0.046 (1.8) 0.053 (2.1) 0.051 (2.0) 0.150 (5.9) 0.071 (2.8) 0.074 (2.9)

0.018 (0.7) 0.038 (1.5) 0.031 (1.2) 0.031 (1.2) 0.020 (0.8) 0.224 (8.8) 0.163 (6.4) 0.025 (1.0)

0.066 (2.6) 0.132 (5.2) 0.097 (3.8) 0.061 (2.4) 0.079 (3.1) 0.257 (10.1) 0.175 (6.9) 0.264 (10.4)

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 169

higher temperatures are favored for forming an aluminum oxide scale. Unlike Fe3Al, TiAl suffered severe corrosion attack at 500, 650, and 800 °C. In many combustion processes involving burning fuels or feedstock containing hydrocarbon, combustion environments under substoichiometric combustion conditions most often would contain CO2. McNallan et al. (Ref 45) examined the effect of CO2 on the chloridation resistance of Fe-Cr and Fe-Ni-Cr alloys and found that CO2 signiﬁcantly increased the alloy’s metal wastage and internal attack. The authors compared the environments between Ar-20O2-2500 ppm (0.25%) Cl2 and Ar-20CO22500 ppm Cl2 for Fe-20Cr and 800H at 927 °C (1700 °F). Figure 6.40 shows metal loss data for both environments as a function of exposure time. For both alloys, the CO2-Cl2 environment caused more metal wastage than the O2-Cl2 environment with Fe-20Cr being more severely affected than alloy 800H. The authors did not offer an explanation in the paper. It was likely that when the environment was switched from Ar-20O2-Cl2 to Ar-20CO2-Cl2 the environment changed from oxidizing to reducing with its oxygen potential being reduced to closer to, or in, the CrCl3 and FeCl2 regimes in the Cr(Fe)-O-C diagrams, resulting in formation of more volatile

chlorides, thus more metal loss. In addition to metal loss, the internal attack for alloy 800H was more severe in the CO2-Cl2 environment than the O2-Cl2 environment (Fig. 6.41). The authors hypothesized the mechanism of the internal corrosion by forming internal chromium carbides, which are then converted to chromium chlorides with carbon reacting with chlorine to form more chromium carbides. Thus, internal corrosion is the result of formation of internal voids and pores in the metal. Fe-20Cr alloy, on the other hand, suffered no internal corrosion. The authors attributed this lack of internal corrosion to the alloy’s low carbon content and ferritic structure. In another paper by McNallan et al. (Ref 46), the CO2-containing environments were further examined. The authors observed internal carburization of Type 310SS in Ar-20CO2 at 800 °C for 24 h. When the test was conducted in Ar-20CO2-Cl2 at 800 °C for 24 h, Type 310SS suffered internal corrosion attack in forms of voids and carbides. Similar internal attack was observed for alloy 800. The corrosion, suggested by the authors, proceeded in two stages with carburization preceding the chlorine-accelerated oxidation. In combustion environments, H2O is invariably present among the combustion products produced. The oxygen potential is dictated by partial pressures of oxygen and hydrogen. Hydrogen reacts with Cl2 to form more stable HCl molecule. The chloridation behavior in HClcontaining environments is discussed in the next two sections. 6.3.3 Corrosion in O2-HCl Environments In some combustion environments, chlorine is present as HCl instead of Cl2. This section covers the corrosion data in oxidizing environments containing HCl. In an oxidizing environment containing HCl, chlorine partial pressure, pCl2 , can be calculated from the equilibrium condition of the reaction below. The oxide and chloride phases that are likely to form thermodynamically can then be determined from the thermodynamic phase stability diagrams discussed earlier in section 6.2. 4HCl(g)+O2 (g)=2Cl2 (g)+2H2 O(g)

Fig. 6.31

Corrosion behavior of alloy 214 (Ni-Cr-Al-Y), alloy S (Ni-Cr-Mo) and alloy 800H (Fe-Ni-Cr) in Ar20O2-0.25Cl2 for 400 h at 700–1000 °C (1290–1830 °F). Source: Ref 39 and Ref 41

The effects of adding O2 to HCl environment were extensively studied by Ihara et al. (Ref 47) on iron, by Ihara et al. (Ref 48) on nickel, and by Ihara et al. (Ref 49) on chromium. Adding O2

pg 169

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

170 / High-Temperature Corrosion and Materials Applications

to HCl signiﬁcantly accelerates the corrosion of iron (Ref 47), as shown in Fig. 6.42. The increased corrosion was attributed to the formation of highly volatile FeCl3 (Ref 47). For nickel, adding O2 to HCl does not signiﬁcantly affect the metal’s corrosion rate, as shown in Fig. 6.43 (Ref 48). This is due to the formation of primarily NiCl2 whether it is 100% HCl or O2-HCl mixtures (Ref 48). Adding O2 to HCl suppresses the corrosion rate of chromium at lower temperatures (400–600 °C) by forming Cr2O3 and accelerates the corrosion rate at higher temperatures (700 and 800 °C) by forming highly volatile CrCl3 (Ref 49). The corrosion rate of chromium as a function of O2-HCl

mixtures at different temperatures is shown in Fig. 6.44 (Ref 49). Devisme et al. (Ref 50) investigated the effect of O2 and CO2 in HCl-bearing environments on the corrosion behavior of four nickel-base alloys, alloys C-276, 600, 601, and 214. Corrosion data in terms of metal loss, which was determined by weight change and metallographic examination of the test specimen, are summarized in Tables 6.15 and 6.16. Adding 2% O2 to Ar20HCl signiﬁcantly increased corrosion attack for alloy C-276 (Ni-Cr-Mo alloy), but signiﬁcantly decreased corrosion attack for alloy 214 (Ni-Cr-Al alloy). Both alloys 600 and 601 were not signiﬁcantly affected. Nevertheless, adding

1600 1575

Alloy AC66

Decrease, µm

500 400

Alloy 800H

10CrMo 9 10*

300 Alloy 45TM 200 Alloy 690

100 0 300

350

400

450

500

550

600

650

700

750

800

Temperature, °C

(a) 225

Alloy AC66

200

10CrMo 9 10*

175

Depth, µm

150 125 Alloy 800H

100 75 50

Alloy 45TM Alloy 690

25 0 300 (b)

350

400

450

500

550

600

650

700

750

800

Temperature, °C

Fig. 6.32 Corrosion behavior of 2.25Cr-1Mo steel (10CrMo9 10), alloy 800H, alloy AC66, alloy 45TM and alloy 690 tested for 300 h at temperatures from 300 to 800 °C (572 to 1472 °F) in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and (b) depth of internal corrosion attack as a function of temperature. Source: Ref 42

pg 170

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 171

10O2 to Ar-5HCl signiﬁcantly increased corrosion attack for alloys C-276, 600, and 601. For Ar-5HCl, adding 0.5% H2O to the environment caused increased corrosion for all four alloys, particularly severe for alloys 600 and 601. The test results obtained by Devisme et al. (Ref 50) showed that overall, alloy C-276 (Ni-Cr-Mo) performed better than alumina former alloy 214 in reducing environments, such Ar-HCl and ArHCl-H2 (Table 6.15). Addition of 10% CO2 was found have less effect on corrosion attack for four alloys in Ar-20HCl environment at 600 and 700 °C, as shown in Table 6.16. Ganesan et al. (Ref 51) investigated the corrosion behavior of nickel-base alloys (625, 825, and 600) and iron-base alloys (800HT, 316SS, and 347SS) in a combustion environment consisting of N2, 4 and 9O2, 12CO2, 500 ppm SO2 with two levels of HCl (1% HCl and 4% HCl). For both levels of HCl, three nickelbase alloys were signiﬁcantly more resistant to chloridation attack than iron-base alloys. For three nickel-base alloys, alloys 625 and 600

Fig. 6.33

Scanning electron backscattered image (a) and an x-ray map for Cl (b), showing a typical pit on alloy 800H tested for 300 h at 300 °C (572 °F) in air-2Cl2. Source: Ref 42

were better than alloy 825. Their test results are shown in Fig. 6.45 to 6.47. Smith and Ganesan (Ref 52) conducted further extensive studies on the corrosion behavior of iron-base alloys (Type 316SS, Type 347SS, and alloy 800HT) and nickel-base alloys (alloys 825, 600, and 625) in simulated combustion environments consisting of N2, O2, SO2 and various amounts of HCl at 426, 593, and 704 °C (800, 1100, and 1300 °F). Also included in their studies was the effect of H2O in the environment on the alloys’ corrosion behavior. Table 6.17 summarizes the test results generated from tests conducted in N2-10O2-50 ppm SO2-500 ppm HCl at 426, 482, and 593 °C (800, 900, and 1100 °F) for 1008 h. HCl is known to be more corrosive than SO2 in high-temperature corrosion. Thus, in this environment, corrosion attack is primarily from HCl. The test results show that the environment that contained about 500 ppm HCl was not corrosive at all for Type 316SS, Type 347SS, 800HT, 825, 600, and 625 at temperatures up to 593 °C (1100 °F). When the HCl content was increased to 4%, both Types 316 and 347 showed much higher corrosion rates at 593 °C (1100 °F), while alloys 800HT, 825, 600, and 625 showed little corrosion attack at 593 °C (1100 °F) after 1008 h, as shown in Table 6.18. However, when the temperature was increased to 704 °C (1300 °F), only high-nickel alloys, such as alloys 600 and 625 exhibited good corrosion resistance in the environment containing 4% HCl (Table 6.19). The environment containing 10% HCl became very corrosive to nickel-base alloys 825, 600, and 625 even at 593 °C (1100 °F) (Table 6.18). At high temperatures, nickel- and cobalt-base alloys, while exhibiting low metal loss, suffered more internal attack in O2-HCl environments. Elliott et al. (Ref 53) examined the corrosion attack in terms of the metal loss and internal penetration for nickel-base alloys (alloys 214, 600, and 601) and cobalt-base alloys (alloys 25 and 188) along with Fe-Ni-Co-Cr alloy (alloy 556), Fe-Ni-Cr alloy (alloy 800H), and Type 310SS after testing in Ar-5.5O2-1HCl-1SO2 at 900 °C (1650 °F) for 800 h with thermal cycling to 200 °C (390 °F) every 100 h. Although the environment contained SO2, hydrogen chloride (HCl), which is known to be more corrosive than SO2, was the primary corrodent causing the corrosion attack. All alloys, while exhibiting low metal losses except Type 310SS, suffered signiﬁcant internal penetration attack. These test results are shown in Fig. 6.48.

pg 171

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

172 / High-Temperature Corrosion and Materials Applications

6.3.4 Corrosion in HCl and HCl-Bearing Reducing Environments Brown et al. (Ref 22) reported corrosion rates of various commercial alloys in a dry HCl environment (Table 6.20). Test duration varied

from 2 to 20 h. Thus, extrapolation to a year would yield an unreliable corrosion rate. Hossain et al. (Ref 54) performed long-term tests in HCl on several nickel alloys and one stainless steel (Type 310SS). Their results are summarized in Table 6.21 and Fig. 6.49. Type 310SS was found

(a)

(d)

(b)

(e)

(c)

(f)

Fig. 6.34

(a) Scanning electron backscattered image of the corrosion products formed on alloy AC66 tested at 800 °C for 300 h in air2Cl2 and the x-ray maps showing elemental distribution for (b) chlorine, (c) chromium, (d) oxygen, (e) iron, and (f) nickel. Source: Ref 42

pg 172

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 173

to be the worst among the alloys tested. All nickel alloys (Ni-201, 601, 625, C-4, B-2, and 600) were much better than Type 310SS. In this HCl test environment containing no O2, the molybdenum-containing nickel-base alloys, such as alloys 625 (Ni-22Cr-9Mo-3.5Nb) and C-4 (Ni-16Cr-15.5Mo), were found to be the

best performers. In a study by Hossain et al. (Ref 54), nickel performed reasonably well in HCl until the temperature reached 700 °C (1290 °F). At 700 °C, nickel was inferior to many nickel-base alloys, such as alloys 600, 625, and C-4 (Table 6.21). Alloy 400 was found to be very susceptible to chloridation

(a)

(d)

(b)

(e)

(c)

(f)

Fig. 6.35

(a) Scanning electron backscattered image of the corrosion products formed on alloy 690 tested at 800 °C for 300 h in air2Cl2 and the x-ray maps showing elemental distribution for (b) chlorine, (c) chromium, (d) iron, (e) oxygen, and (f ) aluminum. Source: Ref 42

pg 173

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

174 / High-Temperature Corrosion and Materials Applications

attack in HCl (Ref 54). Alloy 400 specimens (6 mm diam × 12 mm length) were completely destroyed after exposure for 100 h at 400 °C (750 °F) in HCl. Alloy 600 is commonly used for high-temperature service in HCl environments. Figure 6.50 shows the corrosion rates of alloy 600 in HCl gas as a function of temperature (Ref 22). MTI Publication MS-3 (Ref 27) suggests corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304, and steel in HCl environments, as shown in Fig. 6.51. In reducing environments, such as Ar-4H24HCl, investigated by Baranow et al. (Ref 35), Ni-Cr-Mo alloys, such as alloys C-276 and S,

were signiﬁcantly better than alloys 600, 625, 188, and X. Their test results generated from the 8 h tests at 900 °C (1650 °F) in Ar-4H2-4HCl are shown in Fig. 6.52. In general, alloys suffered very little metal losses, but suffered signiﬁcant internal penetration attack. Figure 6.53 shows the cross sections of the specimens in various iron- and nickel-base alloys after testing in Ar4H2-4HCl at 900 °C (1650 °F) for 8 h (Ref 55). A study was performed by Brill et al. (Ref 56) investigating the chlorination resistance of nickel and nickel-base alloys in H2-10HCl. Pure nickel, Ni-Mo, and Ni-Cr-Mo alloys containing little or no iron were found to be more resistant

Alloy C-2000

500

Decrease, µm

400 Alloy HR160

300

40 30 20

Alloy 59

10 0 300

350

400

450

500

550

600

650

700

750

800

Temperature, °C

(a) 22 20

Alloy C-2000 Alloy 59

18 16

Depth, µm

14

Alloy HR160

12 10 8 6 4 2 0 300

(b)

Fig. 6.36

350

400

450

500

550

600

650

700

750

800

Temperature, °C

Corrosion behavior of alloys 59, C-2000, and HR160; tested for 300 h at temperatures from 300 to 800 °C (572 to 1472 °F) in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and (b) depth of internal corrosion attack as a function of temperature. Source: Ref 43

pg 174

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

Fig. 6.37

26/10/2007 12:28PM Plate # 0

(a) Scanning electron backscattered image of the corrosion products formed on alloy 59 tested at 650 °C for 300 h in air-2Cl2 and the x-ray maps showing elemental distribution for (b) chromium, (c) chlorine, (d) molybdenum, (e) iron, (f ) nickel, and (g) oxygen. Source: Ref 43

pg 175

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

176 / High-Temperature Corrosion and Materials Applications

than some nickel-base alloys with higher iron contents. This is illustrated in Fig. 6.54, showing alloy 205 (pure nickel), B-2 (Ni-28Mo), C-4 (Ni-16Cr-16Mo), and 59 (Ni-23Cr-16Mo) being much more resistant than alloys 625 (Ni-22Cr9Mo-3Fe) and 600H (Ni-16Cr-9Fe) (Ref 56).

Fig. 6.38

In another study by Devisme et al. (Ref 50) on chlorination of Ni-Cr-Mo alloy C-276, NiCr-Fe alloy 600, Ni-Cr-Fe-Al alloy 601 (1.4Al), and Ni-Cr-Al-Fe alloy 214 in Ar-HCl environments. Their test results conducted at 600 °C in Ar-5HCl, Ar-10HCl, and Ar-20HCl are

(a) Scanning electron backscattered image of the corrosion products formed on alloy HR160 tested at 800 °C for 300 h in air-2Cl2, and the x-ray maps showing elemental distribution for (b) chromium, (c) chlorine, (d) silicon, (e) oxygen, and (f ) nickel. Source: Ref 43

pg 176

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 177

shown in Fig. 6.55. Barnes (Ref 57) investigated nickel-base alloys in Ar-HCl mixtures with concentrations of HCl varying from 13 to 100% and found that Ni201 (pure nickel) was slightly more resistant than alloy 600 (8% Fe) for all the concentrations tested except 100% HCl, as shown in Table 6.22. Other nickel-base alloys were also included in his test program, with the test results summarized in Table 6.23. Except for pure nickel (Ni201), nickel-base alloys suffered internal attack by formation of internal voids. This surface zone with extensive internal voids was found to be highly depleted in chromium in the matrix near voids. For example, alloy 600 showed about 1.7% Cr and 4.7% Fe in the matrix in the porous zone after testing at 735 °C (1355 °F) for 100 h in Ar-33HCl (Ref 57). Barnes’s test environments were strictly HCl (inlet gas), thus making pCl2 in the environment signiﬁcantly higher than that if the initial inlet gas contain both H2 and HCl. This is shown in Fig. 6.56, where pCl2 was plotted as a function of temperature for 100% HCl, Ar-33HCl, and H2-30HCl along with several metal chlorides

(Ref 57). The ﬁgure shows that there is no signiﬁcant difference in pCl2 between the 100HCl and Ar-33HCl environments. However, with the

Decrease, µm

400 350 300 250 200 150 Fe3Al 50 TiAl 40 30 20 10 Alloy 214 0 300 350 400 450 500 550 600 650 700 750 800 Temperature, °C

(a) 105 100

Depth, µm

95 30 25 20 15 10 5 0

Fe3Al

Alloy 214

TiAl 300 350 400 450 500 550 600 650 700 750 800

(b)

Temperature, °C

Fig. 6.39

Corrosion behavior of alumina former alloy 214 and intermetallics Fe3Al and TiAl; tested for 300 h at temperatures from 300 to 800 °C (572 to 1472 °F) in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and (b) depth of internal corrosion attack as a function of temperature. Source: Ref 44

Fig. 6.40

Metal loss as a function of time for alloys Fe-20Cr and 800H tested at 927 °C (1700 °F) in (a) Ar20O2-2500 ppm (0.25%) Cl2 and (b) Ar-20CO2-2500 ppm Cl2. Source: Ref 45

pg 177

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

178 / High-Temperature Corrosion and Materials Applications

Fig. 6.41

Depth of internal penetration as a function of time (h0.5) for alloy 800H tested at 927 °C (1700 °F) in (a) argon 20% O2, 0.25% Cl2 and (b) argon 20% CO2, 0.25% Cl2. Source: Ref 45

introduction of H2 in the H2-HCl gas mixture, pCl2 becomes signiﬁcantly lower, as shown in H2-30HCl. Even though chlorine partial pressure (pCl2 ) stays relatively unchanged (Fig. 6.56) thermodynamically, the kinetic of the HCl-metal reaction (corrosion rate) in Ar-HCl mixtures was found to increase with increasing HCl concentration (HCl partial pressure). This is shown in Fig. 6.57 for both Ni201 and alloy 600 tested at 735 °C (1355 °F). It should also be noted that Barnes’s test environments involving mainly HCl were extremely corrosive for nickel alloys tested. In addition to high pCl2 values, the environments were so reducing that no oxides were thermodynamically stable. Thus, in those test environments, the gas-metal reaction mainly involved formation of chlorides (highly volatile corrosion products) and no oxides (nonvolatile solid phases). Figure 6.58 shows nickel chlorides that formed on a Ni201 coupon after testing at 735 °C (1355 °F) in Ar-33HCl (Ref 57). In a simulated waste incineration environment consisting of N2, 12% CO2, 500 ppm SO2, and

1

1

Fig. 6.42

Effect of oxygen in O2-HCl mixtures on the corrosion rate of iron at 300–700 °C (570–1290 °F). Source: Ref 47

Fig. 6.43

Effect of oxygen in O2-HCl mixtures on the corrosion rate of nickel at 400–700 °C (750–1290 °F). Source: Ref 48

pg 178

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:28PM Plate # 0

pg 179

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 179

1% HCl, Ganesan et al. (Ref 51) observed that nickel-base alloys, particularly alloy 625, performed signiﬁcantly better than iron-base alloys, such as alloy 800HT, Type 316SS, and Type 347SS, as shown in Fig. 6.59 and 6.60. Although not discussed in the paper, it is believed that the oxygen potential (pO2 ) was high enough to form some oxides, such as Cr2O3. Thus, little corrosion attack was detected for nickel-base alloys even when the test temperature was increased to 704 °C (1300 °F) (Ref 51). Formation of chromium oxides can signiﬁcantly reduce the corrosion attack in reducing, HCl-bearing

1

800HT

316SS 347SS

Fig. 6.44

Effect of oxygen in O2-HCl mixtures on the corrosion rate of chromium at 400–800 °C (750–1470 °F). Source: Ref 49

Table 6.15 Corrosion of nickel-base alloys in terms of metal loss µm (mils) after 500 h at 600 °C (1112 °F) in the indicated test environments

Fig. 6.45

Weight change as a function of exposure time for nickel-base alloys (alloys 625, 600, and 825) and iron-base alloys (alloy 800HT, 316SS, and 347SS) in N24O2-12CO2-1HCl-500 ppm SO2. Testing was initially performed at 649 °C, then increased to 704 °C, and finally to 760 °C as indicated. Source: Ref 51

Metal loss, μm (mils) Environment

Ar-20HCl Ar-20HCl-2O2 Ar-5HCl Ar-5HCl-10O2 Ar-5HCl-0.5H2O Ar-5HCl-3H2

C-276

600

601

214

60 (2.4) 330 (13.0) 35 (1.4) 120 (4.7) 90 (3.5) 5 (0.2)

150 (5.9) 185 (7.3) 50 (2.0) 140 (5.5) 240 (9.5) 15 (0.6)

150 (5.9) 120 (4.7) 90 (3.5) 160 (6.3) 255 (10.0) 15 (0.6)

260 (10.2) 65 (2.6) 30 (1.2) 55 (2.2) 80 (3.2) 20 (0.8)

800HT

Source: Ref 50

347SS

Table 6.16 Metal loss after 500 h at 600 and 700 °C in the indicated test environments

316SS

Metal loss, μm Environment

Ar-20HCl Ar-20HCl-10CO2 Ar-20HCl Ar-20HCl-10CO2 Source: Ref 50

Temperature, °C

C-276

600

601

214

600 600 700 700

60 110 200 280

150 100 280 195

150 120 225 250

260 230 360 190

Fig. 6.46

Weight change as a function of exposure time for nickel-base alloys (alloys 625, 600, and 825) and iron-base alloys (alloy 800HT, 316SS, and 347SS) in N29O2-12CO2-1HCl-500 ppm SO2. Testing was initially performed at 649 °C, then increased to 704 °C, and finally to 760 °C as indicated. Source: Ref 51

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 180

180 / High-Temperature Corrosion and Materials Applications

800HT 316SS

347SS

704

Fig. 6.48

Fig. 6.47

Weight change as a function of exposure time for nickel-base alloys (alloys 625, 600, and 825) and iron-base alloys (alloys 800HT, 316SS, and 347SS) in N29O2-12CO2-4HCl-100 ppm SO2. Testing was initially performed at 593 °C, then increased to 704 °C, and to 816 °C, and finally to 927 °C as indicated. Source: Ref 51

Table 6.17 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at 426, 482 and 593, 900 °C (800, 900, and 1300 °F) for 1008 h in N2-10O2-50 ppm SO2-500 ppm HCl

The metal loss and internal penetration for nickelbase alloys (alloys 214, 600, and 601) and cobaltbase alloys (alloys 25 and 188) along with Fe-Ni-Co-Cr alloy (alloy 556), Fe-Ni-Cr alloy (alloy 800H), and Type 310SS tested in Ar-5.5O2-1HCl-1SO2 at 900 °C (1650 °F) for 800 h with thermal cycling to 200 °C (390 °F) every 100 h. Source: Ref 53

Table 6.19 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at 593 and 704 °C (1100 and 1300 °F) in N2-10O250 ppm SO2-4HCl Metal loss 593 °C(a)

Metal loss 426 °C

482 °C

593 °C

Alloy

µm/yr

mpy

µm/yr

mpy

µm/yr

mpy

Type 316 Type 347 Alloy 800HT Alloy 825 Alloy 600 Alloy 625

0.07 0.09 0.07 0.02 0.04 0.02

0.003 0.004 0.003 0.0008 0.002 0.0008

2.54 2.29 1.02 1.27 2.03 1.78

0.1 0.09 0.04 0.05 0.08 0.07

5.08 7.11 5.33 2.54 3.05 2.79

0.2 0.28 0.21 0.1 0.12 0.11

704 °C(b)

Alloy

µm/yr

mpy

µm/yr

mpy

Type 316 Type 347 Alloy 800HT Alloy 825 Alloy 600 Alloy 625

914 1245 74 20 25 16

36 49 2.9 0.8 1.0 0.6

3,810 … 12,039 2,083 41 152

149 … 470 81 1.6 5.9

(a) Test duration: 72 h. (b) Test duration: 192 h. 1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 52

1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 52

Table 6.20 Corrosion of alloys in dry HCl(a) Approximate temperature, °C (°F), at which given corrosion rate is exceeded

Table 6.18 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at 593 °C (1100 °F) for 72 h in N2-9O2-12CO2100 ppm SO2-4 and 10HCl Metal loss 4% HCl

10% HCl

Alloy

µm/yr

mpy

µm/yr

mpy

Type 316 Type 347 Alloy 800HT Alloy 825 Alloy 600 Alloy 625

914 1245 74 20 25 16

36 49 2.9 0.8 1.0 0.6

… … … 1066 1219 1549

… … … 42 48 60

1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 52

Alloy

0.8 mm/yr (30 mpy)

1.5 mm/yr (60 mpy)

3.0 mm/yr (120 mpy)

15 mm/yr (600 mpy)

Nickel 600 B C D l8-8Mo 25-12Cb 18-8 Carbon steel Ni-resist 400 Cast iron Copper

455 (850) 425 (800) 370 (700) 370 (700) 288 (550) 370 (700) 345 (650) 345 (650) 260 (500) 260 (500) 230 (450) 205 (400) 93 (200)

510 (950) 480 (900) 425 (800) 425 (800) 370 (700) 370 (700) 400 (750) 400 (750) 315 (600) 315 (600) 260 (500) 260 (500) 148 (300)

565 (1050) 538 (1000) 480 (900) 480 (900) 455 (850) 480 (900) 455 (850) 455 (850) 400 (750) 370 (700) 345 (650) 315 (600) 205 (400)

675 (1250) 675 (1250) 650 (1200) 620 (1150) 650 (1200) 593 (1100) 565 (1050) 593 (1100) 565 (1050) 538 (1000) 480 (900) 455 (850) 315 (600)

(a) Based on short-term laboratory tests. Source: Ref 22

environments. This is illustrated by the test results generated by Strafford et al. (Ref 58). The authors tested Fe-Cr alloys containing 2, 5, 9, 14, and 25% Cr at 1000 °C (1832 °F) in a

H2-H2O-HCl mixture giving pO2 value of 1.5 × 10−16 atm and pCl2 value of 10−8 atm at the test temperature. Based on the M-O-Cl stability

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 181

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 181

diagram at the test temperature, the environment was in the location where Cr2O3, but not FeO, was to form, and FeCl2 was to form. Test results show that Fe-25Cr alloy exhibited no weight loss for the exposure time up to 100 h. This suggests that once the alloy contains sufﬁcient chromium to form a continuous Cr2O3 scale, as in the case of Fe-25Cr, corrosion can be signiﬁcantly

reduced or prevented. The results also indicate that Fe-Cr alloys with low chromium contents, such as 2%, 5%, and 9%, suffered signiﬁcant corrosion attack (Fig. 6.61). In an investigation into possible candidate alloys for a process developed by the Bureau of Mines for extracting alumina from Kaolinitic clay, Carter et al. (Ref 59) tested various

Table 6.21 Corrosion of selected alloys in HCl at 400, 500, 600, and 700 °C (750, 930, 1110, and 1290 °F) Metal loss mg/cm2 400 °C (750 °F)

500 °C (930 °F)

600 °C (1110 °F)

700 °C (1290 °F)

Alloy

300 h

1000 h

100 h

300 h

1000 h

100 h

300 h

96 h

Ni-201 601 310 625 C-4 B-2 600

1.19 1.58 3.26 0.74 0.55 0.75 0.93

0.91 1.47 5.16 1.1 1.12 0.76 0.81

1.60 2.57 6.74 2.42 2.09 2.10 1.69

2.89 4.14 13.65 3.78 3.36 2.65 3.31

4.86 9.38 46.60 8.64 7.24 5.87 7.81

11.46 9.01 15.65 6.79 7.31 12.93 7.67

37.7 19.46 32.6 14.6 19.14 62.3 17.3

377 102.5 1025 26.5 34.9 126.4 49.6

Source: Ref 54

Fig. 6.49

Corrosion rates of several iron- and nickel-base alloys in HCl at 400 to 700 °C (750 to 1290 °F). Source: Ref 54

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

182 / High-Temperature Corrosion and Materials Applications

alloy. Long-term tests (1584 h) at 260 to 380 °C (500 to 720 °F) showed corrosion rates of 0.3 to 8 µm/yr for alloy 625 and 4 to 10 µm/yr for alloy R-41. The results were in agreement with those obtained from short-term tests. Titanium and Ti-0.2Pd were also tested in the same environment for 15 days. The corrosion rates for titanium were found to be 246, 28, and 13 µm/yr at 400, 300, and 200 °C, respectively. Ti-0.2Pd was found to corrode at 1900, 108, 0, and 2 µm/yr at 500, 400, 300, and 200 °C, respectively. In a study by Reeve (Ref 60) involving 80% HCl and 20% H2O, corrosion rates were obtained for steel and stainless steels (Fig. 6.62). Mild steel was found to suffer corrosion rates of less than 0.76 mm/yr (30 mpy) up to 400 °C (752 °F), and the 18-8 stainless steel, less than 1.0 mm/yr (40 mpy) up to 500 °C (930 °F). These values were signiﬁcantly lower than those predicted by Brown et al. (Ref 22) in dry HCl (Table 6.20). It appears that addition of signiﬁcant amount of H2O into HCl environment makes HCl less corrosive.

month

commercial alloys in an environment of 40% HCl and 60% H2O at various temperatures up to 500 °C (930 °F). The results of iron- and nickelbase alloys are summarized in Tables 6.24 and 6.25. Corrosion data at temperatures below 200 °C (390 °F) are not included because of dew point corrosion. Stainless steels and nickel alloys tested showed low corrosion rates at all test temperatures in this environment. A cobalt-base alloy (alloy 188) and an Fe-Ni-Co-Cr alloy (Multimet alloy) were also tested in the temperature range of 315 to 375 °C (600 to 710 °F), showing no measurable attack for alloy 188 and a corrosion rate of only 0 to 3 µm/yr for Multimet

6.4 Corrosion in F2- and HF-Bearing Environments 6.4.1 Corrosion in F2 Environments Fig. 6.50

Early studies of corrosion in ﬂuorine gas for various metals and alloys were carried out by

Corrosion rate of alloy 600 in HCl as a function of temperature. Source: Ref 22

Corrosion rate, mm/yr 0.025 700

0.05

0.10

0.25

Tubes / Internals

0.50

1.00

2.5 1290

Vessels / Pipes 1110

600 Nickel 200 and Alloy 600

400

1020 930

Type 304

300

750

Alloy 400

572

Carbon steel

200

Temperature, °F

Temperature, °C

500

100

390

0 1

2

4

6

10

20

40

60

100

Corrosion rate, mpy

Fig. 6.51

MTI corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304SS and steel in HCl as a function of temperature. Source: Ref 27. Courtesy of Materials Technology Institute

pg 182

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 183

Metal loss

214 Continuous internal penetration

C-276

S

601

R-41

600

625

188

X 0

40

80

120 160 200 240 280 320

360

Corrosion rate, mils/month

Fig. 6.52

Corrosion rates in terms of metal loss and internal penetration for nickel- and cobalt-base alloys at 900 °C (1650 °F) in Ar-4HCl-4H2. Data was based on 8 h tests. Source: Ref 35

Myers and DeLong (Ref 61). Their test duration was short, normally 4 h. The longest test run was 15 h. The corrosion rates they obtained are summarized in Table 6.26. These rates, extrapolated from short-term tests, are not recommended for estimating the service life of process equipment. Rather, the data are useful for making performance comparisons among various alloys. The results obtained by Myers and DeLong (Ref 61) indicated that nickel had good corrosion resistance in ﬂuorine gas at temperatures up to 500 °C (930 °F). Even at temperatures higher than 500 °C (930 °F), corrosion rates for nickel were signiﬁcantly lower than those of other alloys. Nickel is commonly used for plant equipment handling ﬂuorine at temperatures up to 500 °C (930 °F) (Ref 62). The resistance of nickel to ﬂuorine gas was attributed to the formation of an adherent nickel ﬂuoride scale (Ref 62). The reaction of nickel and ﬂuorine was found to follow a parabolic rate law (Ref 63).

Hauffe (Ref 64) summarized the results of corrosion tests by Myers and DeLong (Ref 61), Jarry et al. (Ref 62), and Lukyanchev et al. (Ref 65) in Table 6.27. The tests were very short in duration, from 30 min to 32 h. The corrosion rates obtained by Myers and DeLong were two orders of magnitude higher than those observed by Jarry et al. and Lukyanchev et al. Based on the corrosion rates observed by Jarry et al. and Lukyanchev et al., nickel exhibited good resistance in ﬂuorine at temperatures up to 810 °C (1490 °F) (about 12 mpy). Hale et al. (Ref 66) tested six different types of nickel with various degrees of purity at 590 °C (1100 °F) for 95 h in ﬂuorine and found no measurable corrosion for all the samples, except one with a slightly higher silicon content (about 8 mils of general corrosion). The results by Hale et al. (Ref 66) are shown in Table 6.28. They also found that the material’s purity played an important role when tested at 700 °C (1290 °F) for 210 h (Ref 66). Both high-purity vacuum-melted nickel and electrolytic nickel showed negligible attack. Low-carbon nickel I (about 0.015% Si), carbonyl nickel, and “A” nickel (commercial grade pure nickel) began to show intergranular void formation at 700 °C (1290 °F). At the same temperature, low-carbon nickel II (about 0.05% Si) suffered signiﬁcant general corrosion and intergranular attack (void formation). Silicon appears to be a detrimental impurity in pure nickel in resisting corrosion attack by ﬂuorine. Results obtained by Steindler and Vogel (Ref 67) showed that “A” nickel (commercial pure nickel) suffered signiﬁcant corrosion at 750 °C (1380 °F), as shown in Table 6.29. A general review on the kinetic aspects of nickel-ﬂuorine reactions can be found in Ref 8. Additions of alloying elements to nickel generally are detrimental to ﬂuorine corrosion resistance. Many nickel-base alloys were found to be signiﬁcantly more susceptible than nickel to ﬂuorine corrosion (Ref 61, 66, 67). Alloy 600 (Ni-15.5Cr-8Fe), Inco 61 weld wire (Ni-1.5Al-3Ti), Ni-O-Nel (Ni-21Cr-31Fe-3 Mo-1.75Cu), INOR-1 (Ni-20Mo), INOR-2 (Ni5Cr-16Mo), INOR-3 (Ni-16Mo-1Al-1.6Ti), INOR-4 (Ni-17Mo-1.7Fe-2Al-1.7Ti), INOR-5 (Ni-13Mo-2.7W-2.2Cb+Ta-1Mn), Hastelloy W (Ni-25Mo-5.5Fe-2.5Co), Hastelloy B (Ni25Mo-6Fe), HyMa 80 (Ni-16Fe-4Mo), and Monel (Ni-30Cu) suffered signiﬁcantly more corrosion than nickel in ﬂuorine at 590 °C (1100 °F), as shown in Table 6.28 (Ref 66).

pg 183

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 184

184 / High-Temperature Corrosion and Materials Applications

Fig. 6.53

Internal penetration in terms of voids for various iron- and nickel-base alloys after testing in 900 °C (1650 °F) for 8 h in Ar-4H2-4HCl. Source: Ref 55

0

Metal loss, µm

–50

–100

–150

–200

Fig. 6.54

Corrosion of nickel-base alloys in H2-10HCl at 850 °C with 24 h cycles. Source: Ref 56

–250

5% HCl 10% HCl 20% HCl

–300 C276

Dura-Nickel (Ni-3Ti-1.5Al), 70Cu-30Ni, and Ni-10Co suffered slightly more corrosion than nickel at 590 °C (1095 °F) (Ref 66). Fluorides of molybdenum, tungsten, titanium, and other elements have low melting points and/or high vapor pressures (Table 6.2). Iron is signiﬁcantly less resistant to ﬂuorine attack than nickel. Myers and DeLong (Ref 61) observed that Armco Iron is resistant to ﬂuorine

600

601

214

Alloy

Fig. 6.55

Corrosion in terms of loss of sound metal (µm) of alloys C-276, 600, 601, and 214 in Ar-5HCl, Ar10HCl, and Ar-20HCl at 600 °C for 500 h. Source: Ref 50

at temperatures up to 250 °C (480 °F). Steels were found to be less resistant than Armco Iron (Ref 61). The concentration of silicon appears

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 185

Table 6.22 Metal loss rate(a) in Ar-HCl and 100HCl environments at 735 °C (1355 °F)(b) Metal loss Alloy

Ni201 Alloy 600 Ni201 Alloy 600 Ni201 Alloy 600 Ni201 Alloy 600

Environment

mm/yr

mpy

Ar-13HCl Ar-13HCl Ar-33HCl Ar-33HCl Ar-52HCl Ar-52HCl 100HCl 100HCl

2.8 3.5 5.3 7.9 9.6 11.4 18.4 14.2

110 138 209 311 378 449 724 559

(a) Metal loss rate did not include internal void penetration. Internal void penetration was a small portion of the total metal loss. (b) Test durations from 15–100 h. Source: Ref 57

to be important in affecting the steel’s ﬂuoridation resistance, as shown in Table 6.26 (Ref 61). SiF4 has a very low melting point and high vapor pressure (Table 6.2). Ferritic and austenitic stainless steels, except Type 347SS, showed negligible corrosion at 200 and 250 °C (390 and 480 °F) (Table 6.26). At higher temperatures, corrosion of these alloys became signiﬁcant. Jackson (Ref 68) also reported signiﬁcant corrosion rates for several austenitic stainless steels at 370 °C (700 °F), as

Table 6.23 Estimated metal loss rates for pure nickel and nickel-base alloys tested at 685, 735, and 785 °C (1265, 1355, and 1445 °F) in 100HCl Metal loss Alloy

Ni201 Alloy 600 Alloy HR160 Alloy 214 Alloy 602CA Alloy HR160 Ni201 Alloy 600 Alloy HR160 Source: Ref 57

Temperature, °C (°F)

mm/yr

mpy

685 (1265) 685 (1265) 685 (1265) 686 (1265) 686 (1265) 735 (1355) 785 (1445) 785 (1445) 785 (1445)

4.2 7.5 7.0 5.7 28.6 45.6 16.7 26.4 96.0

165 295 276 224 1130 1800 657 1040 3780

Fig. 6.57

Metal loss rate (mm/yr) as a function of HCl concentrations (pHCl) in Ar-HCl mixtures at 735 °C (1355 °F). Source: Ref 57

H2-30% HCI

Fig. 6.56

Thermodynamic equilibrium chlorine partial pressure (pCl2 ) as a function of temperature for several environments (100HCl, Ar-33HCl, and H2-30HCl) and several chlorides. Source: Ref 57

pg 185

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 186

186 / High-Temperature Corrosion and Materials Applications

Fig. 6.58

Nickel chlorides formed on Ni201 after testing at 735 °C (1355 °F) for 15 h in Ar-33HCl. Source: Ref 57

347SS

316SS 316SS 347SS

Fig. 6.59

Mass change as a function of time for nickeland iron-base alloys tested initially at 593 °C (1100 °F), then increased to 649 °C (1200 °F), and finally to 704 °C (1300 °F) in N2-12%CO2-500ppm SO2-1%HCl. Source: Ref 51

Mass change as a function of time for nickel- and iron-base alloys tested at 593 °C (1100 °F) in N212%CO2-500ppmSO2-1%HCl. Source: Ref 51

shown in Table 6.30. Cobalt and cobalt-base alloys are not as resistant to ﬂuorine as nickel (Ref 66) (Table 6.28). Limited data for other metals, such as copper, aluminum, and magnesium, are shown in Tables 6.26, 6.29, and 6.31. Aluminum was resistant to ﬂuorine at temperatures up to approximately 500 °C (930 °F) (Table 6.26 and 6.31). AlF3 has a relatively high melting point (1197 °C or 2187 °F) (Ref 64). Corrosion of nickel, alloy 400, and alloy 600 by various volatile metallic ﬂuorides was investigated by Vogel et al. (Ref 69) in their studies on

volatile ﬁssion product ﬂuorides, which were associated with the development of a process to recover uranium and plutonium from partially spent nuclear reactor fuels. Most of these ﬁssion product ﬂuorides were much less corrosive than ﬂuorine gas. However, whenever ﬂuorine gas was present along with the ﬂuoride, the corrosion rate was generally more aggressive. Nickel and alloy 400 exhibited relatively low corrosion rates for all the volatile ﬂuorides at 500 °C (930 °F), except TeF6, as shown in Table 6.32. Alloy 600, on the other hand, suffered extremely high corrosion rates.

Fig. 6.60

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 187

6.4.2 Corrosion in HF Environments Hydrogen ﬂuoride is generally less corrosive than ﬂuorine for most metals and alloys. However, it is still very corrosive at elevated temperatures. Corrosion rates of various metals and

Fig. 6.61

Weight change as a function of exposure time at 1000 °C (1832 °F) in a H2-H2O-HCl mixture giving pO2 value of 1.5 × 10−16 atm and pCl2 value of 10−8 atm for Fe-Cr alloys containing various amounts of chromium. Source: Ref 58

Table 6.24 Corrosion of iron-base alloys in 40HCl-60H2O Corrosion rate(a) μm/yr Temperature, °C (°F)

500 (930) 400 (750) 300 (570) 210 (410) 200 (390)

316LSS

29-4SS

430SS

E-Brite 26-1

1020 steel

4130 steel

483 15 3 … 7

… … 10 3 …

… … … 8 …

363 5 2 … 3

2100 326 700 … 46

1700 406 870 … 51

(a) Linearly extrapolated from 15 d (360 h) laboratory tests. Note: mpy = (µm) × 0.0394. Source: Ref 59

Table 6.25 Corrosion of nickel-base alloys in 40HCl-60H2O

alloys in HF based on short-term tests were reported by Myers and DeLong (Ref 61), as shown in Table 6.33. Nickel was the most resistant among the alloys tested. Copper, alloy 400, and alloy 600 were slightly worse than nickel. Carbon steels and stainless steels showed poor resistance. Tyreman and Elliott (Ref 70) found that nickel, cobalt, copper, and molybdenum were more resistant to HF corrosion than iron, chromium, niobium (columbium), and tantalum (Fig. 6.63). Chromium was found to be detrimental to resistance to corrosion by HF for Ni-Cr alloys (Fig. 6.64) (Ref 71). Chromium ﬂuorides were the major corrosion products for Ni-Cr alloys tested in HF (Ref 71). Marsh (Ref 72) did an extensive investigation for his Ph.D. thesis on the resistance of several nickel-base alloys to corrosion attack in HF environments at elevated temperatures. In NiCr alloys tested in HF at 650 °C (1200 °F), chromium ﬂuorides were found to form on the metal surface as well as internally as internal phases. This is illustrated in Fig. 6.65, showing elemental distribution for Ni-40Cr alloy tested at 650 °C (1200 °F) in anhydrous HF environment (Ref 72). Surface corrosion products were found to be enriched in chromium and ﬂuorine, but depleted in nickel, and similar elemental distribution was observed for the internal phases. Several commercial nickel-base alloys were also tested. Figure 6.66 shows corrosion attack of alloy 600 in HF after 92 h at 650 °C (1200 °F). In this case, both the corrosion products formed on the metal surface and the internal phases were found to consist of both CrF3 and FeF2, as determined by x-ray diffraction analysis (Ref 72). In Ni-Cr alloys, it appears that NiCr-Mo alloy 625 was more resistant to HF corrosion attack than Ni-Cr alloys X750 and

Corrosion rate(a), μm/yr Temperature, °C (°F)

500 (930) 380 (720) 380 (720) 380 (720) 375 (710) 375 (710) 365 (690) 350 (660) 350 (660) 315 (600) 310 (590) 290 (550) 260 (500)

Ni-200

600

625

R-41

B-2

… 13 15 15 25 13 20 3 18 13 … … 30

… 58 46 58 48 48 46 30 33 15 … … 13

132 … 13 23 10 15 13 8 15 8 0 … 8

… 3 3 3 5 0 3 3 5 0 0 … 3

… 15 18 20 13 28 33 … 36 … … … 28

(a) Linearly extrapolated from 15 d (360 h) laboratory tests. Note: mpy = (µm) × 0.0394. Source: Ref 59

Fig. 6.62

Corrosion rates of mild steel, Type 304SS, and alloy 800 in 80HCl-20H2O at 300 to 600 °C (570 to 1110 °F). Source: Ref 60

pg 187

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 188

188 / High-Temperature Corrosion and Materials Applications

Table 6.26

Corrosion of various metals and alloys in fluorine Corrosion rates(a), mpy

Materials

200 °C

250 °C

300 °C

350 °C

400 °C

450 °C

500 °C

600 °C

650 °C

700 °C

Nickel 400 600 Copper Aluminum Magnesium 430SS 347SS 309CbSS 310SS Armco Fe Steel (0.007% Si) 1020 (0.22% Si) 1030 (trace Si) 1030 (0.18% Si) 1015 (0.07% Si) Music wire (0.13% Si)

… … … … … Nil 8.4 Nil Nil Nil Nil Nil 456 24 … … …

… … … … … Nil Nil 1,740 Nil Nil 24 192 5,760 96 … … …

… … … … … Nil 3,060 2,556 900 372 108 48 7,920 108 9,000 9,960 4,800

… … … … … … 936 6,204 5,544 4,248 96 2.4 1,764 Nil … … …

8 6 456 1,920 Nil … 936 9,540 7,980 6,732 288 144 6,480 180 … … …

23 18 1,152 … Nil … … … … … 3600 … 18 in. 6,480 … … …

61 24 744 1,440 156 … … … … … 139 in. … … 238 in. … … …

348 720 2,040 11,880 216 … … … … … … … … … … … …

192 960 1,560 … … … … … … … … … … … … … …

408 1,800 6,120 … … … … … … … … … … … … … …

(a) Note: mpy × 0.0254 = mm/yr. Source: Ref 61

Table 6.27 Corrosion of nickel in fluorine at various temperatures Temperature, °C (°F)

300 (570) 300 (570) 400 (750) 400 (750) 400 (750) 500 (930) 500 (930) 500 (930) 550 (1020) 600 (1110) 660 (1220) 720 (1330) 810 (1490)

Table 6.28 Results of corrosion tests in fluorine at 590 °C (1100 °F) for 95 h

Pressure, atm

Test duration, h

Corrosion rate, mm/yr (mpy)

Ref

0.9 0.9 0.9 0.9 0.99 0.9 0.9 0.99 0.99 0.9 0.99 0.99 0.99

8 32 0.5 28 … 0.25 30 … 2 32 1.5 2 1.7

0.005 (0.2) 0.00073 (0.03) 0.015 (0.59) 0.0012 (0.05) 0.21 (8.3) 0.018 (0.7) 0.003 (0.12) 1.5 (59.1) 0.036 (1.4) 0.017 (0.7) 0.150 (5.9) 0.240 (9.4) 0.310 (12.2)

62 62 62 62 61 62 62 61 65 62 65 65 65

Source: Ref 64

601 containing no signiﬁcant amounts of molybdenum, as illustrated in Fig. 6.67 (Ref 71). Corrosion attack in HF for alloys 601, 625, and N is shown in Fig. 6.68 to 6.71. Alloy N (Ni-5Cr16Mo) was found to be signiﬁcantly more resistant to HF corrosion than nickel-base alloys 600, 601, and 625. This suggests that nickel-base alloys containing low chromium and high molybdenum would be more resistant to HF corrosion attack than Ni-Cr alloys. Molybdenum showed little corrosion attack in HF after 10 h at 850 °C (1560 °F), as shown in Fig. 6.63 (Ref 70). The results of Zotikov and Semenyuk (Ref 73) indicated that both molybdenum and tungsten were quite resistant to corrosion attack by HF at temperatures from 300 to 800 °C (up to 700 °C for tungsten), as illustrated in Table 6.34. Molybdenum appears

Alloy

Depth of corrosion attack(a), mm (mils)

High-purity nickel sheet High-purity nickel rod Low-carbon nickel (II) Low-carbon nickel (I) Carbonyl nickel Electrolytic nickel “A” nickel 70-30 Cupronickel Inconel (low carbon) Inco “61” weld wire Duranickel Ni-O-NEL

NM NM 0.20 (8) NM NM NM NM 0.06 (2.5) 0.64 (25) 0.64 (25) 0.10 (4) >0.42 (16.5)

INOR-1

>0.83 (32.5)

INOR-2

>0.83 (32.5)

INOR-3

>0.83 (32.5)

INOR-4

>0.81 (32)

INOR-5 Hastelloy alloy B

0.76 (30) >0.48 (19)

HyMa 80 90Ni-10Co 80Ni-20Co Monel

0.37 (14.5) 0.06 (2.5) >0.13 (5) >0.80 (31.5)

Hastelloy alloy W

>0.79 (31)

310SS

>0.60 (23.5)

Carpenter 20

>0.66 (26)

Haynes alloy No. 25

>0.34 (13.5)

Cobalt

>1.52 (60)

(a) NM, not measurable. Source: Ref 66

Comments

… … General attack … … … … General attack General attack General attack General attack Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides General attack Completely converted to ﬂuorides General attack General attack General attack Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides Completely converted to ﬂuorides

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 189

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 189

to improve the corrosion resistance of nickelbase alloys in HF. The results published by International Nickel Company (Ref 74) showed that alloy N (Ni-16Mo-5Cr) and alloy B (Ni25Mo) were slightly better than nickel in HF (Table 6.35). Field testing in an HF-bearing environment showed that alloy N and alloy S (Ni-16Cr-15Mo) were better than alloys 625 and C-22 at 900 °C (1650 °F) (Ref 75). Barnes (Ref 76) investigated the corrosion behavior of a wide range of materials in HF environments at 1000 °C (1832 °F). Tests were Table 6.29 Corrosion of several metals and alloys in fluorine(a) Metal

Test temperature, °C (°F)

Time, h

Weight gain, mg/cm2

Corrosion rate, mm/y (mils/y)

Ni(b) 600 Cu Ni(b) 400 600 Cu Ni(b) 400 600 600 Cu

550 (1,020) 550 (1,020) 550 (1,020) 650 (1,200) 650 (1,200) 650 (1,200) 650 (1,200) 750 (1,380) 750 (1,380) 750 (1,380) 750 (1,380) 750 (1,380)

6.2 5.0 2.42 5.28 6.0 5.6 4.8 4.1 6.1 4.7 4.7 5.8

1.9 706 17.4 21.5 16.0 1,743 134 100 −1, 831 4, 907 −12, 220 278

0.11 (4.4) 81 (3,200) 2.8 (110) 1.5 (59) 1.0 (41) 180 (7,100) 10.9 (430) 9.0 (353) 74 (2,900) 610 (24,000) 660 (26,000) 19 (750)

(a) Tests were conducted in ﬂowing gas (30 to 130 cc/min). (b) Nickel “A” (commercial grade pure nickel). Source: Ref 67

Table 6.30 Corrosion of several alloys in fluorine(a) Corrosion rate, mm/yr (mpy) Alloy

Exposure time, h

400

Ni-200

304 304L 347 Illium “R” 600

5 24 120 5 24 120 5 24 120 5 5 5

200 °C (400 °F)

370 °C (700 °F)

540 °C (1000 °F)

0.013 (0.5) 0.048 (1.9) 0.76 (29.8) 0.013 (0.5) 0.043 (1.7) 0.29 (11.3) 0.003 (0.1) 0.031 (1.2) 0.18 (7.2) 0.084 (3.3) 0.043 (1.7) 0.62 (24.5) 0.013 (0.5) 0.031 (1.2) 0.41 (16.1) 0.003 (0.1) 0.010 (0.4) 0.35 (13.8) 0.155 (6.1) 40 (1565) … 0.191 (7.5) 153 (6018) … 0.65 (25.4) … … 0.102 (4.0) 108 (4248) … 0.152 (6.0) 0.32 (12.7) 103 (4038) 0.015 (0.6) 2.0 (78.0) 88 (3451)

(a) Tests were conducted in ﬂowing ﬂuorine. Source: Ref 68

Table 6.31

Corrosion of aluminum in fluorine

Test temperature, °C (°F)

26 (79) 201 (394) 356 (673) 543 (1009) Source: Ref 68

Corrosion rate, mm/yr (mpy)

0.00087 (0.03) 0.0003 (0.01) 0.57 (22) 2.2 (87)

conducted in Ar-5HF, Ar-15HF, and Ar-35HF mixtures. In nickel-base alloys, alloys B-3 (Ni28Mo-1.5Cr) and 242 (Ni-25Cr-8Mo) were found to be signiﬁcantly better than alloys 600, 617, and 602CA. His test results are summarized in Table 6.36 and Fig. 6.72. Both Ni-Mo alloys containing low Cr formed thin surface scales with very little internal ﬂuoride penetration. Surprisingly, alloy 188, a cobalt-base alloy with high chromium (22%) and high tungsten (14%), also exhibited good corrosion resistance with a thin surface scale and little internal ﬂuoride penetration. Nickel aluminide intermetallic (Ni-8Al-8Cr-1.4Mo-1.7Zr) was found to exhibit good corrosion resistance in Ar-5HF. However, when tested in Ar-35HF, the nickel-aluminide Table 6.32 Corrosion of nickel and nickel-base alloys by various volatile fluorides at 500 °C (930 °F) Corrosive environment

Test duration, h

GeF4 + F2 AsF5 AsF5 + F2 SeF6 SeF6 SeF6 + F2 MoF6 MoF6 + F2 MoF6 + F2 TeF6 TeF6 TeF6 + F2 SF6 UF6 F2 F2

8.4 7.0 7.2 7.0 29.5 6.6 9.2 6.0 7.0 18.9 5.7 7.8 28.7 28.8 7.0 8.6

Corrosion rate(a), mm/yr (mpy) Ni-200

Alloy 400

Nil(b) 0.22 (8.8) 0.22 (8.8) Nil 0.44 (17.5) 0.67 (26.3) 0.22 (8.8) 0.22 (8.8) (c) (c) 0.22 (8.8) 0.22 (8.8) 0.22 (8.8) Nil 0.22 (8.8) Nil (c) (c) 8.9 (350) 1.8 (70) 135 (5326) 56 (2190) 0.22 (8.8) 0.22 (8.8) Nil Nil 0.22 (8.8) … 0.66 (26) 0.22 (8.8) … …

Alloy 600

29 (1139) Nil 45 (1770) 0.44 (17.5) … 22 (850) 0.22 (8.8) 44 (1726) 63 (2488) 2.4 (96) 5.6 (219) 40 (1568) Nil … 9.8 (385) 31 (1209)

(a) Calculated from weight loss after descaling. (b) Rates reported as nil are less than 0.001 mils/h. (c) Scale was not completely removed by descaling. Source: Ref 69

Table 6.33 Corrosion of various metals and alloys in anhydrous HF Corrosion rate, mm/yr (mpy) Material

Nickel 400 600 Copper Aluminum Magnesium Carbon steel (1020) 304 347 309Cb 310 430 Source: Ref 61

500 °C (930 °F)

550 °C (1020 °F)

600 °C (1110 °F)

0.9 (36) 1.2 (48) 1.5 (60) 1.5 (60) 4.9 (192) 12.8 (504) 15.5 (612) … 183 (7,200) 5.8 (228) 12. 2 (480) 1.5 (60)

… 1.2 (48) … … … … 14.6 (576) … 457 (18,000) 43 (1,680) 100 (3,960) 9.1 (360)

0.9 (36) 1.8 (72) 1.5 (60) 1.2 (48) 14.6 (576) … 7.6 (300) 13.4 (528) 177 (6,960) 168 (6,600) 305 (12,000) 11.6 (456)

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

190 / High-Temperature Corrosion and Materials Applications

suffered extensive corrosion attack, while alloy 242 remained resistant under the same test condition. The results indicated that Ni-Al system was not as good as Ni-Mo system in resisting HF corrosion attack. Barnes (Ref 76) also tested various pure metals, which included gold (Au), platinum (Pt), palladium (Pd), chromium, Ni200, Ni201,

Fig. 6.63

Corrosion of various metals in HF and HF-50H2O at 850 °C (1562 °F) for 10 h. Source: Ref 70

Fig. 6.64

Effect of chromium on resistance to HF at 650 °C (1200 °F) for Ni-Cr alloys. Source: Ref 71

Ni270, and copper. The test results on pure metals are summarized in Table 6.37 (Ref 76). Precious metals (Au, Pt, and Pd) showed no weight changes and little changes in specimen surface appearance after testing. However, both gold and platinum specimens were found to exhibit surface etching on the specimen surface. When tested in a higher concentration of HF (Ar-35HF), palladium was heavily corroded with some molten corrosion product formed on the specimen surface. There was no discussion in the paper about the performance of gold and platinum in Ar-35HF mixture. Copper was also found to exhibit few changes in specimen appearance after testing in Ar-5HF. In fact, the copper specimen retained its original metallic luster after testing in Ar-5HF. Chromium was found to be quite susceptible to ﬂuoridation attack when tested in Ar-5HF. The metal suffered extensive internal chromium ﬂuoride penetration and extensive weight loss due to vaporization of chromium ﬂuorides. Nickel was found to be extremely resistant to HF corrosion attack. Three different types of nickel were tested: Ni200 and Ni201 (99% pure nickel), and Ni270 (99.9% pure nickel) (Ref 76). Ni201 is a low carbon nickel (0.02% C), while Ni200 is a high carbon nickel (0.15% C). The impurities and minor elements in these three nickel specimens in the test program are: 0.15C, 0.4Fe, 0.35Si, 0.35Mn, and 0.25Cu for Ni200; 0.02C, 0.4Fe, 0.35Si, 0.35Mn, and 0.25Cu for Ni201; and 0.02C, 0.001Cu, and 0.05 max Fe for Ni270. The test results after 15 h in Ar-5HF are summarized in Table 6.37 (Ref 76). All nickel specimens retained metallic luster appearance after testing. A nickel wire that had been used for holding test specimens during testing of specimens in the Ar-5HF test gas at 1000 °C for 1600 h was sectioned for metallographic examination, showing only a penetration of about 20 µm (0.8 mils) of internal nickel ﬂuoride precipitates. These nickel ﬂuorides, which were randomly distributed, were found to be less than 2 µm in diameter. Some nickel ﬂuorides were also detected on the metal surface. Ceramic materials and graphite were also tested by Barnes (Ref 76) under the same test conditions. The results are summarized in Table 6.38. Graphite was found to be extremely resistant to HF corrosion. Ceramic materials, such as alumina (polycrystalline or single crystal), sintered silicon carbide, chemical vapor deposited (CVD) silicon carbide, and silicon nitride, suffered high corrosion rates.

pg 190

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 191

Fig. 6.65

Scanning electron backscattered image (a) and x-ray maps for Cr (b), Ni (c), and F (d) for Ni-40Cr alloy tested at 650 °C (1200 °F) for 22 h in anhydrous HF environment. Source: Ref 72. Courtesy of Glyn Marsh

Fig. 6.66

Optical micrograph showing corrosion attack of alloy 600 after testing in HF for 92 h at 650 °C (1200 °F). Source: Ref 72. Courtesy of Glyn Marsh

pg 191

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

192 / High-Temperature Corrosion and Materials Applications

Adding H2O to HF appears to make HF less corrosive. Figure 6.63 shows the corrosion data generated by Tyreman and Elliott (Ref 70) by

Fig. 6.69

Optical micrograph showing corrosion attack on alloy 601 after testing in HF for 16 h at 650 °C (1200 °F). Source: Ref 72. Courtesy of Glyn Marsh

Fig. 6.67

Corrosion kinetics of alloys 625, X750, and 601 as a function of time in HF at 650 °C (1200 °F). Source:

Ref 71

Fig. 6.70

Optical micrograph showing very little corrosion attack (about 1 to 2 µm in depth) on alloy N (Ni-5Cr-16Mo) after testing in HF for 46 h at 650 °C (1200 °F). Source: Ref 72. Courtesy of Glyn Marsh

Fig. 6.71

Fig. 6.68

Optical micrograph showing corrosion attack on alloy 625 after testing in HF for 142 h at 650 °C (1200 °F). Source: Ref 72. Courtesy of Glyn Marsh

Scanning electron micrograph showing the corrosion scale (mainly NiF2 with some CrF3, as determined by x-ray diffraction analysis) formed on alloy N after testing in HF for 115 h at 650 °C (1200 °F). Source: Ref 72. Courtesy of Glyn Marsh

pg 192

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 193

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 193

comparing the 100% HF and HF-50%H2O environments for nickel, copper, cobalt, iron, chromium, molybdenum, niobium, and tantalum. All metals except molybdenum showed more corrosion attack in HF than in HF-50H2O. Corrosion data generated in HF-10H2O and HF50%H2O are summarized in Tables 6.39 and 6.40, respectively. 6.4.3 Corrosion in O2-HF Environments Oxygen appears to make HF more corrosive. During nuclear fuel reprocessing, stainless steel fuel cladding was being chemically removed by reacting it with an O2-HF mixture (40–60% HF) at 377 to 627 °C (710 to 1160 °F) (Ref 77). The Table 6.34 Corrosion of molybdenum and tungsten in HF containing 0.6% H2O Corrosion rate, mm/yr (mpy) Test temperature, °C (°F)

Molybdenum

Tungsten

300 (570) 400 (750) 500 (930) 600 (1110) 700 (1290) 800 (1470)

0.003 (0.1) 0.011 (0.4) 0.014 (0.6) 0.023 (0.9) 0.144 (5.7) 1.3 (51)

0.003 (0.1) 0.009 (0.4) 0.017 (0.7) 0.022 (0.9) 0.91 (36) …

removal rates (or corrosion rates) were more than 1 mm/h (40 mils per h (Ref 77). Macheteau et al. (Ref 78) also found that oxygen contamination accelerated ﬂuoridation attack of iron. Marsh and Elliott (Ref 79) observed that cobalt exhibited a protective CoF2 ﬁlm with a very low corrosion rate when exposed to HF at 650 °C (1200 °F). However, once air was introduced to mix with the test gas of HF (i.e., HF-O2-N2 mixture), the corrosion rate increased signiﬁcantly. This is illustrated in Fig. 6.73 (Ref 79). The rapid corrosion attack of cobalt when air was mixed with HF was associated with the formation of numerous Co3O4 oxide protrusions, as shown in Fig. 6.74. At the Co3O4 oxide and the cobalt metal interface, no protective ﬂuoride ﬁlms were observed except remnant segments of CoF2 ﬁlms were still present as indicated the lower line in Fig. 6.74. It is believed that as soon as the air was introduced to mix with HF gas stream, a protective CoF2 ﬁlm was “broken,” prompting the formation of Co3O4 oxide protrusions. Schutze and Simon (Ref 80) tested a number of alloys in N2-5O2-3HF at 1100 °C for 75 h in

Source: Ref 73

Table 6.35 Corrosion of various nickel-base alloys in HF at 500 to 600 °C (930 to 1110 °F) Alloy

Corrosion rate, mm/yr (mpy)

Alloy C Alloy 600 Alloy B Ni-200 Ni-201 Alloy 400 Alloy K-500 70Cu-30Ni

0.008 (0.3) 0.018 (0.7) 0.051 (2) 0.229 (9) 0.356 (14) 0.330 (13) 0.406 (16) 0.406 (16)

Tests were performed in HF (7 lb/h) for 36 h. Source: Ref 74

Fig. 6.72

Depth of internal fluoride penetration for alloys tested in Ar-5HF at 1000 °C (1832 °F) for 15 h.

Source: Ref 76

Table 6.36 Corrosion behavior of nickel-base alloys along with one cobalt-base alloy (alloy 188) and nickel aluminide Intermetallic (IC221M) in Ar-5HF at 1000 °C (1832 °F) for 15 h Alloy

Alloy 242 Alloy B-3 IC221M Alloy 188 Alloy 617 Alloy 600 Alloy 602CA

Maximum attack(a), mm/side (mils/side)

Specimen appearance

Comments

0.015 (0.6) 0.02 (0.8) 0.025 (1.0) 0.051 (2.0) 0.13 (5.3) 0.15 (5.9) 0.18 (7.1)

Surface scale Surface scale Surface scale Surface scale Nodules on surface Nodules on surface Molten corrosion product

Good resistance at 35HF … Extensive corrosion at 35HF(b) … Extensive internal ﬂuorides Extensive internal ﬂuorides Extensive internal ﬂuorides

(a) Surface metal loss + internal ﬂuoride penetration. (b) IC221M exhibited good resistance in Ar-5HF, but poor resistance in Ar-35HF. The specimen was almost completely converted to ﬂuorides. Source: Ref 76

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:29PM Plate # 0

pg 194

194 / High-Temperature Corrosion and Materials Applications

Table 6.37 Materials

Corrosion behavior of pure metals in Ar-5HF at 1000 °C (1832 °F) for 15 h Weight change, mg/cm2

Corrosion rate, mpy

Specimen appearance

0 −0.2 0.2 −35.25 −0.1 −0.1 0.0

0.0 1.1 … 1454(b) 1.4 1.4 0.0

Unchanged Matte grey Unchanged Green Unchanged Unchanged Unchanged

Gold Platinum Palladium Chromium(a) Ni200 Ni201 Ni270

Comments

… … Heavily corroded in 35HF Total depth of attack = 8.3 mils No internal attack No internal attack No internal attack

(a)Tested for 50 h at 1000 °C in Ar-5HF. (b) Based on total depth of attack (maximum metal affected) of 8.3 mils/side in a 50 h test. Source: Ref 76

Table 6.38 Corrosion behavior of ceramic materials and graphite in Ar-5HF at 1000 °C (1832 °F) for 15 h Materials

Graphite (CZR-1) Graphite (DFP-2) Alumina (polycrystalline) Alumina (sapphire) CVD SiC Sintered SiC Si3N4

Weight change, mg/cm2

Corrosion rate, mpy

Comments

0 0 −2.4 −0.9 −5.7 −6.4 −2.2

0.0 0.0 141 52 411 458 155

Unchanged Unchanged General wastage Frosted appearance General wastage General wastage General wastage

Source: Ref 76

Table 6.39 Corrosion of several alloys in HF-10H2O at 850 °C (1560 °F) Alloy

Corrosion rate, mm/yr (mpy)

304LSS 310SS 800H 600 625

100–130 (4000–5000) 100–130 (4000–5000) 100–130 (4000–5000) 23 (900) 6.7 (265)

Source: Ref 70

Table 6.40 Corrosion of nickel and Alloy 400 in 50HF-50H2O at various temperatures Corrosion rate, mm/yr (mpy)

their search for a candidate alloy for air nozzles in a circulating ﬂuidized-bed combustor for reprocessing spent pot lining material in aluminum production. Among the metallic materials tested, alloy 242 was found to exhibit the smallest metal loss (50 µm). Although the alloy suffered extensive nitridation attack, the authors considered that nitridation attack would not affect the alloy’s performance as air nozzles. In a lowtemperature test (450 °C) in N2-8O2-10CO215H2O-5HF, Crum et al. (Ref 81) observed no measurable corrosion attack for many nickelbase alloys and a superaustenitic stainless steel after 155 h of exposure. Their test results are summarized in Table 6.41. Stress-corrosion cracking (SCC) resistance of these alloys was also included in the test program using the same test environment with U-bend test specimens. All alloys showed no cracking after 100 and 155 h except alloy 600. Authors did not explain why alloy 600 suffered SCC while other alloys including many nickel-base alloys and one superaustenitic stainless steel (alloy 25-6MO) showed no cracking.

6.5 Corrosion in Bromine and Iodine Environments Very little data have been reported on the performance of metals and alloys in bromine and

550 °C (1020 °F)

600 °C (1110 °F)

650 °C (1200 °F)

700 °C (1290 °F)

750 °C (1380 °F)

Nickel 0.79 (31) 1.83 (72) 2.74 (108) 3.66 (144) 3.05 (120) Alloy 400 … 0.61( 24) 1.52 (60) 3.96 (156) 5.18 (204) Source: Ref 61

iodine environments at elevated temperatures. Miller et al. (Ref 82) investigated the corrosion behavior of copper, nickel, and nickel-base alloys in bromine at 300 and 500 °C (470 and 930 °F) (Table 6.42). Copper suffered rapid attack by bromine at 300 °C (570 °F). The CuBr compound exhibits very high vapor pressures. The vapor pressure of CuBr reaches 1× 10−4 atm (a value considered high enough to cause hightemperature corrosion) when the temperature is as low as 435 °C (815 °F) (Table 6.3). Nickel exhibited good corrosion resistance in Br2 at 300 and 500 °C (470 and 930 °F). Bromine reacts with nickel to form NiBr2, which melts at 965 °C (1769 °F), signiﬁcantly higher than the melting point of CuBr. Duranickel 301 (Ni-5Al) had a corrosion resistance similar to nickel. Alloy 400 was less resistant than nickel and Duranickel 301. Smith and Ganesan (Ref 83) investigated the corrosion behavior of various commercial alloys in a simulated combustion environment containing a very high level of HBr (about 4%) at 593 and 927 °C (1100 and 1700 °F). The test

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:30PM Plate # 0

pg 195

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 195

gas (the inlet gas) consisted of 9% O2, 12% CO2, 4% HBr, 100 ppm SO2, and balance N2. The equilibrium partial pressures of the test environment at 593 and 927 °C (1100 and 1700 °F) are listed in Table 6.43. The test results, which are summarized in Table 6.44, showed surprisingly very little corrosion attack at 593 °C (1100 °F) for all the alloys tested, which included austenitic stainless steels (Type 309, 316, and

347), alloy 800, and nickel-base alloys. At 927 °C (1700 °F), nickel-base alloys showed little weight loss, while austenitic stainless steels, Fe-Ni-Cr alloy 800, and high-iron-containing nickel alloy 825 suffered high weight losses. The authors (Ref 83) analyzed the corrosion products spalled from test specimens as well as Table 6.41 Corrosion rates generated from tests in N2-8O2-10CO2-15H2O-5HF at 450 °C (842 °F) for 100 and 155 h Alloy

Alloy 600 Alloy 400 Alloy 825 Alloy C-276 Alloy 686 Alloy 622 Alloy 25-6MO

Corrosion rate, mm/yr (mpy) from 100 h tests

Corrosion rate, mm/yr (mpy) from 155 h tests

0.01 (0.4) 0 0.01 (0.4) 0 0 0 0

0 0.002 (0.1) 0 0 0 0 0

Source: Ref 81

Table 6.42 Corrosion of several alloys in bromine at 300 and 500 °C (570 and 930 °F) Alloy

Ni-201

Fig. 6.73

Weight change of cobalt at 650 °C (1200 °F) when exposed to HF during the first 12 h, showing a very low corrosion rate. At time “X,” air was introduced into the test gas to mix with HF, rapid corrosion attack was observed. Source: Ref 79

Temperature, °C (°F)

300 (570) 300 (570) 300 (570) Duranickel 301 300 (570) 300 (570) 400 300 (570) 300 (570) Copper 300 (570)(a) 300 (570)(a) Ni-201 500 (930) 500 (930) Duranickel 301 500 (930) 500 (930)

Time, days

Weight loss, mg/cm2

Corrosion rate, mm/yr (mpy)

11 11 11 10 10 2 2 <1 <1 10 10 10 10

0.28 0.16 0.14 0.17 0.07 7.07 6.65 Rapid attack Rapid attack 3.83 1.40 2.14 3.64

<0.03 (1.2) <0.03 (1.2) <0.03 (1.2) <0.03 (1.2) <0.03 (1.2) 1.5 (60) 1.5 (60) … … 0.15 (6.0) 0.06 (2.4) 0.09 (3.6) 0.15 (6.0)

(a) Tests were terminated before 300 °C was reached because of rapid corrosion attack. Source: Ref 82

Table 6.43 Thermodynamic equilibria in terms of partial pressures (atm) for the inlet gas of N2-9O2-12CO2-4HBr-100 ppm SO2 at 593 and 927 °C (1100 and 1700 °F) Gas component

Fig. 6.74

A Co3O4 oxide protrusion formed on cobalt immediately after test gas was switched from HF to air-HF mixture at 650 °C. The upper line indicates the Co3O4 oxide protrusion, and the lower line indicates the CoF2 phase formed at the interface between the Co3O4 oxide and the substrate cobalt. Source: Ref 72. Courtesy of Glyn Marsh

N2 CO2 O2 Br2 H 2O HBr SO3 SO2 H2 Source: Ref 83

Partial pressure at 593 °C, atm

Partial pressure at 927 °C, atm

0.78 0.12 0.81 × 10−1 0.20 × 10−1 0.20 × 10−1 0.11 × 10−3 0.90 × 10−4 0.63 × 10−6 0.39 × 10−14

0.78 0.12 0.81 × 10−1 0.19 × 10−1 0.19 × 10−1 0.22 × 10−2 0.82 × 10−5 0.82 × 10−4 0.82 × 10−9

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:30PM Plate # 0

196 / High-Temperature Corrosion and Materials Applications

the extraction residue obtained from the test specimens using x-ray diffraction technique to determine the phases present. The results of the x-ray diffraction analysis are summarized in Table 6.45 for the corrosion scales spalled from tested specimens and in Table 6.46 for the extraction residue obtained from tested specimens by electrolytic extraction of phases. In both cases, the phases found were oxides. No bromide

Table 6.44 Mass change data for alloys tested in N2-9O2-12CO2-4HBr-100 ppm SO2 at 593 and 927 °C (1100 and 1700 °F) for 300 h Mass change, mg/cm2 Alloy

593 °C (1100 °F)

927 °C (1700 °F)

309 316 347 800 825 625 601 600 617 690

0.19 0.75 0.59 0.55 −0.11 0.10 0.56 0.55 0.15 0.50

−218.70 −83.35 −67.57 −53.25 −28.57 −8.77 −7.78 −6.04 −5.57 −3.88

phases were detected. The results by Smith and Ganesan (Ref 83) suggest that nickel-base alloys appear to be suitable for applications in HBrcontaining environments up to about 4%. It is also suggested that HBr is not as corrosive as HCl at elevated temperatures. More tests, particularly long-term tests, are needed to conﬁrm the current ﬁndings. Corrosion of various alloys in I2 was investigated by Shapiro (Ref 84) when determining the most suitable alloy for construction of vessels for iodine zirconium processing in the production of zirconium. Tests were conducted at 300 and 450 °C (570 and 840 °F) in iodine vapor with 400 mm Hg (0.53 atm) pressure. Results are shown in Table 6.47 and 6.48. Platinum, gold, tungsten, and molybdenum were very resistant to corrosion attack in I2. Alloy B (Ni-Mo), alloy C (Ni-Cr-Mo), alloy 600 (Ni-Cr-Fe), and nickel were quite resistant to I2. Stainless steels and alloy 400 were less resistant. In the production of zirconium using the iodide process, closure gaskets made of gold are used for handling dry iodine vapors at 500 °C (939 °F) (Ref 85).

Source: Ref 83

Table 6.45 Results of x-ray diffraction analysis on the spalled corrosion products for the test specimens after testing in N2-9O2-12CO2-4HBr100 ppm SO2 at 927 °C (1700 °F) for 300 h Alloy

Major phases

Minor phases

316 800 825 600 601 617 625 690

Fe2O3 NiFe2O4, Cr2O4 Cr2O3, NiCr2O4 Cr2O3, NiFe2O4 Cr2O3 Cr2O3 Cr2O3 Cr2O3, Fe2O3

NiCr2O4, M3O4 Fe2O3 … … M 3O 4 M 3O 4 M 3O 4 M 3O 4

Table 6.47 Corrosion rates of several metals and alloys in iodine at 300 °C (570 °F)(a) Alloy

Platinum Gold Tungsten Molybdenum Tantalum Alloy B Alloy C Alloy 600 Nickel Alloy 400

Corrosion rate, mm/yr (mpy)

0 0 0 0.003 (0.11) 0.004 (0.16) 0.044 (1.7) 0.057 (2.2) 0.107 (4.2) 0.27 (10.6) 0.56 (22.0)

(a)400 mm Hg (0.53 atm) iodine pressure. Source: Ref 84

Source: Ref 83

Table 6.48 Corrosion rates of several metals and alloys in iodine at 450 °C (840 °F)(a) Table 6.46 Results of x-ray diffraction analysis on the extracted residue from the test specimens after testing in N2-9O2-12CO2-4HBr100 ppm SO2 at 927 °C (1700 °F) for 300 h Alloy

Major phases

Minor phases

316 800 825 600 601 617 625 690

Cr2O3, M3O4 NiFe2O4, Cr2O3 Cr2O3 NiCrO3 Al2O3, Cr2O3 Cr2O3 Cr2O3, NiCr2O4 Cr2O3

… NiO NiFe2O4 NiO … M3O4 … NiFe2O4

Source: Ref 83

Alloy

Platinum Tungsten Gold Molybdenum Alloy B Alloy 600 Tantalum Nickel Type 310SS Type 316SS Type 347SS Alloy 400 Type 304SS

Corrosion rate, mm/yr (mpy)

0.0055 (0.2) 0.008 (0.3) 0.024 (0.9) 0.033 (1.3) 0.46 (18) 0.54 (21) 0.88 (35) 1.2 (47) 1.8 (71) 2.1 (83) 2.1 (83) 2.6 (102) 3.2 (126)

(a)400 mm Hg (0.53 atm) iodine pressure. Source: Ref 84

pg 196

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:30PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 197

6.6 Summary High-temperature corrosion behavior of metals and alloys in gaseous environments containing (a) chlorine and hydrogen chloride, (b) ﬂuorine and hydrogen ﬂuoride, (c) bromine and hydrogen bromide, and (d) iodine and hydrogen iodide is reviewed. Metals or alloys in these environments form metal halides as corrosion products. Many metal halides exhibit high vapor pressures; the corrosion products can thus be in a gaseous phase. Some metal halides also exhibit low melting points; the corrosion products can thus be in a liquid phase. As a result, the metal-halogen or metal-hydrogen-halide reactions can proceed at a rapid rate at elevated temperatures. The corrosion behavior of alloys can be signiﬁcantly different between “oxidizing” environments that contain O2 and reducing environments that contain no O2. Also, corrosion by halogen or by hydrogen halide can proceed at different rates. Accordingly, the corrosion data are categorized in different environment types. The arrangement of the corrosion data in this way may also help readers to determine the type of data that may be more pertinent to their intended applications or conditions. For the corrosion behavior of alloys in environments containing Cl2 and HCl, the data are presented in (a) Cl2 environments containing no O2, (b) O2-Cl2 mixtures, (c) O2-HCl mixtures, and (d) HCl and HCl-bearing, reducing environments. For corrosion in F2- and HF-bearing environments, the data are presented in (a) F2 environments, (b) HF environments, and (c) O2-HF mixtures. For bromine- and iodine-containing environments the corrosion data are very limited, and a brief review is presented.

REFERENCES

1. W.J. Kroll, Method of Manufacturing Ti and Alloys Thereof, U.S. Patent 2205854, Jan 1940 2. W.A. Henderson, J. Met., Vol 16, 1964, p 155 3. S.M. Shelton, The Metallurgy of Zirconium, B. Lustman and F. Kerze, Jr., Ed., McGrawHill, 1955, p 59 4. I. Iwasaki, Y. Takahasi, and H. Kahata, Trans. SME AIME, Vol 243, 1966, p 308 5. C.L. Mantell, Tin, Reinhold, 1949 6. W.E. Berry, Corrosion in Nuclear Applications, John Wiley & Sons, 1971

7. T. Rosenquist, Principles of Extractive Metallurgy, McGraw-Hill, 1974 8. P.L. Daniel and R.A. Rapp, Advances in Corrosion Science and Technology, Vol 5, M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1970, p 55 9. O. Kubaschewski and E. Evans, Metallurgical Thermochemistry, Pergamon Press, 1958 10. HSC, Chemistry for Windows, Version 6.0, A. Roine, Outokumpu Technology, Finland, www.outokumpotechnology.com, accessed Dec 2006 11. ChemSage, Version 4.16, GTT-Technologies, Aachen, 1998 12. M.J. McNallan, W.W. Liang, J.M. Oh, and C.T. Kang, Morphology of Corrosion Products Formed on Cobalt and Nickel in Argon-Oxygen-Chlorine Mixtures at 1000 °K, Oxid. Met., Vol 17, 1982, p 371 13. A. Zahs, M. Spiegel, and H.J. Grabke, The Inﬂuence of Alloying Elements on the Chlorine-Induced High Temperature Corrosion of Fe-Cr Alloys in Oxidizing Atmospheres, Mater. Corros., Vol 50, 1999, p 561 14. F.H. Stott and C.Y. Shih, The Inﬂuence of HCl on the Oxidation of Iron at Elevated Temperatures, Mater. Corros., Vol 51, 2000, p 277 15. R. Bender and M. Schutze, The Role of Alloying Elements in Commercial Alloys for Corrosion Resistance in OxidizingChloridizing Atmospheres—Part I: Literature Evaluation and Thermodynamic Calculations on Phase Stabilities, Mater. Corros., Vol 54, 2003, p 567 16. K.L. Tseitlin and V.A. Strunkin, J. Appl. Chem. USSR, Vol 31, 1958, p 1832 17. K.L. Tseitlin, J. Appl. Chem. USSR, Vol 28 (No. 5), 1955, p 467 18. G. Heinemann, F.G. Garrison, and P.A. Haber, Ind. Eng. Chem., Vol 38, 1946, p 497 19. R.J. Fruehan, Metall. Trans., Vol 3 (No. 10), 1972, p 2585 20. S.F. Bohlken, A. Klinkenberg, and H.W. Nicolai, Ind. Chim. Belge., Vol 20, 1955, p 579 21. S.F. Bohlken, A. Klinkenberg, and H.W. Nicolai, C. R. Cong. Int. Chim. Ind., Vol 27, 1954, p 2 22. M.H. Brown, W.B. DeLong, and J.R. Auld, Ind. Eng. Chem., Vol 39 (No. 7), 1947, p 839 23. G. Han and W.D. Cho, High-Temperature Corrosion of Fe3Al in 1% Cl2/Ar, Oxid. Met., Vol 58, 2002, p 390

pg 197

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:30PM Plate # 0

198 / High-Temperature Corrosion and Materials Applications

24. B.J. Downey, J.C. Bermel, and P.J. Zimmer, Corrosion, Vol 25 (No. 12), 1969, p 502 25. J.D. McKinley, Jr. and K.E. Shuler, J. Chem. Phys., Vol 28, 1958, p 1207 26. R.H. Kane, Process Industries Corrosion, B.J. Moritz and W.I. Pollock, Ed., NACE, 1986, p 45 27. MTI Publication MS-3, Materials Selector for Hazardous Chemicals, Vol 3, Hydrochloric Acid, Hydrogen Chloride and Chlorine, Materials Technology Institute of the Chemical Process Industries, Inc., C.P. Dillon and W.I. Pollock, Ed., 1999 28. J.P. Tu, Z.Z. Li, and Z.Y. Mao, Internal Chlorination of Ni-Based Alloys and Its Relation to Volatilization Corrosion, Mater. Corros., Vol 48, 1997, p 441 29. M.J. Maloney and M.J. McNallan, Met. Trans. B, Vol 16, 1985, p 751 30. M.J. McNallan and W.W. Liang, GaseousTransport-Controlled Chlorination of CoO in Flowing Oxygen-Chlorine-Argon Mixtures at 1000 °K, J. Am. Ceram. Soc., Vol 64 (No. 5), 1981, p 302 31. N.S. Jacobson, M.J. McNallan, and Y.Y. Lee, The Formation of Volatile Corrosion Products During the Mixed OxidationChlorination of Cobalt at 650 °C, Metall. Trans., Vol 17A, 1986, p 1223 32. Y.Y. Lee and M.J. McNallan, Ignition of Nickel in Environments Containing Oxygen and Chlorine, Metall. Trans., Vol 18A, 1987, p 1099 33. M.J. McNallan, High-Temperature Corrosion in Halogen Environments, MP, Sept. 1994, p 54 34. J.C. Liu and M.J. McNallan, Effects of Temperature Variations on Oxidation of Iron-20% Chromium Alloys at 1200 °K in Ar-20% O2-Cl2 Gas Mixtures, Mater. Corros., Vol 50, 1999, p 253 35. S. Baranow, G.Y. Lai, M.F. Rothman, J.M. Oh, M.J. McNallan, and M.H. Rhee, Paper No. 16, Corrosion/84, NACE, 1984 36. J.M. Oh, M.J. McNallan, G.Y. Lai, and M.F. Rothman, Metall. Trans. A, Vol 17A, 1986, p 1087 37. N.S. Jacobson, M.J. McNallan, and Y.Y. Lee, Mass Spectrometric Observations of Metal Oxychlorides Produced by OxidationChlorination Reactions, Metall. Trans. A, Vol 20A, 1989, p 1566 38. M.H. Rhee, M.J. McNallan, and M.F. Rothman, High Temperature Corrosion in

39. 40. 41. 42.

43.

44.

45.

46.

47. 48. 49. 50.

51.

Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 483 M.J. McNallan, M.H. Rhee, S. Thongtem, and T. Hensler, Paper No. 11, Corrosion/85, NACE, 1985 P. Elliott, A.A. Ansari, R. Prescott, and M.F. Rothman, Paper No. 13, Corrosion/85, NACE, 1985 S. Thongtem, M.J. McNallan, and G.Y. Lai, Paper No. 372, Corrosion/86, NACE, 1986 C. Schwalm and M. Schutze, The Corrosion Behavior of Several Heat Resistant Materials in Air + 2 vol-% Cl2 at 300 to 800 °C: Part 1—Fe-Base and Fe-Containing Alloys, Mater. Corros., Vol 51, 2000, p 34 C. Schwalm and M. Schutze, The Corrosion Behavior of Several Heat Resistant Materials in Air +2% Cl2 at 300 to 800 °C: Part 2—Nickel Base Alloys, Mater. Corros., Vol 51, 2000, p 73 C. Schwalm and M. Schutze, The Corrosion Behavior of Several Heat Resistant Materials in Air + 2% Cl2 at 300 to 800 °C: Part 3—Alumina Formers and Intermetallics, Mater. Corros., Vol 51, 2000, p 161 M.J. McNallan, S. Thongtem, J.C. Liu, Y.S. Park, and P. Shyu, Corrosion of Chromium Containing Alloys in Non-steady State Environments Containing Oxygen, Carbon, and Chlorine, J. Phys. IV, Coll. C9, Suppl. J. Phys. III, Vol 3, Dec 1993, p 143 M.J. McNallan, Z. Niemczura, H.H. Lu, and Y.S. Park, Carbide Formation and Oxidation of Austenitic Alloys in Oxidizing/ Carburizing Environments Contaminated with Chlorine, Paper No. 252, Corrosion/93, NACE, 1993 Y. Ihara, H. Ohgame, K. Sakiyama, and K. Hashimoto, Corros. Sci., Vol 21 (No. 12), 1981, p 805 Y. Ihara, H. Ohgame, and K. Sakiyama, Corros. Sci., Vol 22 (No. 10), 1982, p 901 Y. Ihara, H. Ohgame, and K. Sakiyama, Corros. Sci., Vol 23 (No. 2), 1983, p 167 F. Devisme, P. Falgoux, F. Lefebvre, and T. Flament, “High Temperature Corrosion in Atmospheres Containing Hydrogen Chloride,” 11th International Incineration Conference (Albuquerque, NM), May 11–15, 1992 P. Ganesan, G.D. Smith, and L.E. Shoemaker, The Effects of Excursions of Oxygen and Water Vapor Contents on Nickel-Containing Alloy Performance in

pg 198

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:30PM Plate # 0

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 199

52.

53.

54. 55.

56.

57.

58.

59. 60. 61. 62. 63.

A Waste Incineration Environment, Paper No. 248, Corrosion/91, NACE, 1991 G.D. Smith and P. Ganesan, Metallic Corrosion in Waste Incineration: A Look at Selected Environmental and Alloy Fundamentals, Heat-Resistant Materials II, Conf. Proc. Second International Conference on Heat-Resistant Materials, K. Natesan, P. Ganesan, G. Lai, Ed., ASM International, 1995, p 631 P. Elliott, R. Prescott, and C.J. Tyreman, Aspects of Halogen Corrosion at High Temperatures, Paper No. 373, Corrosion/86, NACE, 1986 M.K. Hossain, J.E. Rhoades-Brown, S.R.J. Saunders, and K. Ball, Proc. UK Corrosion/ 83, 1983, p 61 G.Y. Lai, M.F. Rothman, S. Baranow, and R.B. Herchenroeder, Recuperator Alloys for High-Temperature Waste Heat Recovery, J. Met., Vol 35 (No. 7), 1983, p 24 U. Brill, J. Klower, and D.C. Agarwal, Inﬂuence of Alloying Elements on the Chlorination Behavior of Nickel- and Iron-Based Alloys, Paper No. 439, Corrosion/96, NACE International, 1996 J.J. Barnes, Materials Limitations in High Temperature Chlorination Environments, Paper No. 446, Corrosion/96, NACE International, 1996 K.N. Strafford, P.K. Datta, and G. Forster, High-Temperature Chloridation of Binary Fe-Cr Alloys at 1000 °C, High Temperature Corrosion 2: Advanced Materials and Coatings, Proc. Second International Symposium on High Temperature Corrosion of Advanced Materials and Coatings (Les Embiez, France), May 22–26, 1989, R. Streiff, J. Stringer, R.C. Krutenat, and M. Caillet, Ed., Elsevier Science Publishers, London, 1989, p 61 J.P. Carter, B.S. Covino, Jr., T.J. Driscoll, W.D. Riley, and M. Rosen, Corrosion, Vol 40 (No. 5), 1984, p 205 L. Reeve, J. Iron Steel Inst., Sept 1955, p 26 W.R. Myers and W.B. DeLong, Chem. Eng. Prog., Vol 44 (No. 5), 1948, p 359 R.L. Jarry, J. Fischer, and W.H. Gunther, J. Electrochem. Soc., Vol 110 (No. 4), 1963, p 346 R.L. Jarry, W.H. Gunther, and J. Fischer, “The Mechanism and Kinetics of the Reaction Between Nickel and Fluorine,” ANL-6684, Argonne National Laboratory, Argonne, IL, Aug 1963

64. K. Hauffe, Z. Werkstofftech., Vol 15, 1984, p 427 65. Y.A. Lukyanchev, I.I. Astakhov, and N.S. Nikolaev, Mechanism of the Formation of Fluoride Films on Nickel and Their Properties, Bull. Acad. SSSR. Div. Chem. Sci., 1965, p 577 66. C.F. Hale, E.J. Barber, H.A. Bernhardt, and K.E. Rapp, “High Temperature Corrosion of Some Metals and Ceramics in Fluorinating Atmospheres,” Report K-1459, Union Carbide Nuclear Co., Sept 1960 67. M.J. Steindler and R.C. Vogel, “Corrosion of Materials in the Presence of Fluorine at Elevated Temperatures,” ANL-5662, Argonne, IL, Jan 1957 68. R.B. Jackson, “Corrosion of Metals and Alloys by Fluorine,” NP-8845, Allied Chemical Corp., Morristown, PA, 1960 69. R.C. Vogel, J.H. Schraidt, and J. Royal, “Chemical Engineering Division SemiAnnual Report,” ANL-7125, Argonne National Laboratory, Argonne, IL, May 1966 70. C.J. Tyreman and P. Elliott, Paper No. 135, Corrosion/88, NACE, 1988 71. G. Marsh and P. Elliott, High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 467 72. G. Marsh, Ph.D. thesis, UMIST, Manchester, U.K., Dec 1982 73. V.S. Zotikov and E.Ya, Semenyuk, Metallschutz Moscow, Vol 6, 1970, p 218 74. “Corrosion Resistance of Nickel-Containing Alloys in Hydroﬂuoric Acid, Hydrogen Fluoride and Fluorine,” CEB-5, International Nickel Co., 1968 75. Haynes International, Inc., unpublished results, 1988 76. J.J. Barnes, Materials Behavior in High Temperature HF-Containing Environments, Paper No. 518, Corrosion/2000, NACE International, 2000 77. G. Dumont and W. Goossens, Chemical Decladding of Fuel Elements with HF/O2 Mixture, Chem. Eng. Technol., Vol 43, 1971, p 800 78. Y. Macheteau, J. Gillardeau, P. Plurien, and J. Oudar, Oxid. Met., Vol 4 (No. 3), 1972, p 141 79. G. Marsh and P. Elliott, Aspects of High Temperature Hydroﬂuorination of Cobalt, High Temp. Technol., Vol 3 (No. 4), 1985, p 215

pg 199

Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/

26/10/2007 12:30PM Plate # 0

200 / High-Temperature Corrosion and Materials Applications

80. M. Schutze and P. Simon, Materials for Use in Atmospheres Containing HF in a Pyrohydrolysis Plant for Re-processing Spent Pot Lining from Aluminum Production, Werkst. Korros., Vol 45, 1994, p 435 81. J.R. Crum, G.D. Smith, M.J. McNallan, and S. Himyj, Characterization of Corrosion Resistant Materials in Low and High Temperature HF Environments, Paper No. 382, Corrosion/99, NACE International, 1999 82. P.D. Miller, E.F. Stephan, W.E. Berry, and W.K. Boyd, “Corrosion Resistance of Nickel and Two Nickel Alloys to Gaseous

Bromine,” BMI-X-489, Battelle Memorial Institute, Columbus, Ohio, Jan 1968 83. G.D. Smith and P. Ganesan, Performance and Scale Formation of Selected High Temperature Alloys in Simulated Waste Incineration Environments Containing Gaseous Bromides and Chlorides, Paper No. 406, Corrosion/87, NACE, 1987 84. Z.M. Shapiro, Metallurgy of Zirconium, B. Lustman and F. Kerze, Jr., Ed., McGrawHill, 1955, p 135 85. Handbook of Corrosion Data, B.D. Craig, Ed., ASM International, 1989

pg 200

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p201-234 DOI: 10.1361/hcma2007p201

26/10/2007 11:13AM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 7

Sulfidation 7.1 Introduction Sulfur is one of the most common corrosive contaminants in high-temperature industrial environments. Sulfur is generally present as an impurity in fuels or feedstocks. Typically, fuel oils are contaminated with sulfur varying from fractions of 1% (No. 1 or No. 2 fuel oil) to about 3% (No. 6 fuel oil). United States coal may contain from 0.5 to 5% S, depending on where it is mined. Feedstocks for calcining operations in mineral and chemical processing are frequently contaminated with various amounts of sulfur. Sulfur is one of the corrosive species in petroleum-reﬁning operations. When combustion takes place with excess air to ensure complete combustion of fuel for generating heat in many industrial processes, sulfur in the fuel reacts with oxygen to form SO2 and SO3. A combustion atmosphere of this type is oxidizing in nature. Oxidizing environments are usually much less corrosive than reducing environments, where sulfur is in form H2S. However, sulﬁdation in oxidizing environments (as well as in reducing environments) is frequently accelerated by other fuel impurities, such as sodium, potassium, and chlorine, which may react among themselves and/or with sulfur during combustion to form salt vapors. These salt vapors may then deposit at lower temperatures on metal surfaces, resulting in accelerated sulﬁdation attack. This accelerated sulﬁdation attack due to salt deposits is frequently referred to as “hot corrosion” in gas turbines (Chapter 9), “coal ash corrosion” in coal-ﬁred boilers (Chapter 10), and “oil ash corrosion” in oil-ﬁred boilers (Chapter 11). They are discussed in these chapters. Oxidizing environments that contain high concentrations of SO2 can be produced in the chemical process used to manufacture sulfuric acid. Sulfur, in this case, is used as a feedstock. Combustion of sulfur with excess air takes place in a sulfur furnace at about 1150 to 1200 °C

(2100 to 2200 °F). The product gas typically contains about 10 to 15% SO2 along with 5 to 10% O2 (balance N2), which is then converted to SO3 for sulfuric acid. In many industrial processes, combustion is carried out under stoichiometric or substoichiometric conditions to convert feedstock to process gases consisting of H2, CO, CH4, and other hydrocarbons. Sulfur is converted to H2S. The environment, in this case, is reducing and is characterized by low oxygen potentials. Coal gasiﬁcation, which converts coal to substitute natural gas or medium- and low-BTU fuel gases, is an example of a process that generates this type of atmosphere. Reducing conditions may also prevail in localized areas, in some cases, even when combustion is taking place with excess air. A coal-ﬁred boiler equipped with low NOx combustion burners is an example. More discussion on this issue is presented in Chapter 10 on high-temperature corrosion and other materials issues in coal-ﬁred boilers. Furthermore, ash deposits on the metal surface can sometimes turn an oxidizing condition in the gaseous environment into a reducing condition beneath the deposits. In most cases, alloys rely on oxide scales to resist sulﬁdation attack; most high-temperature alloys rely on chromium oxide scales. In oxidizing environments, oxide scales form readily because of high oxygen activities. Thus, oxidation dominates the corrosion reaction. When the environment is reducing (i.e., low oxygen potentials), the corrosion reaction becomes a competition between oxidation and sulﬁdation. Thus, lowering the oxygen activity tends to make the environment more sulﬁdizing, resulting in increased domination by suﬁdation. Conversely, increasing the oxygen activity generally results in a less-sulﬁdizing environment with increased domination by oxidation. Sulﬁdation reaction is thus controlled by both sulfur and oxygen activities (also referred to as “potentials”). When corrosion involves more than one mode,

pg 201

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

202 / High-Temperature Corrosion and Materials Applications

including sulﬁdation, sulﬁdation generally dictates materials selection. Sulﬁdation attack, in many cases, can be quite localized. Figure 7.1 shows a high-nickel alloy that suffered sulﬁdation attack at about 930 °C (1700 °F) in a furnace ﬁring ceramic tiles. The cross section at the corroded area showed sulﬁdes through the cross section of the component. The breakdown of a protective oxide scale (i.e., Cr2O3 scale for most high-temperature alloys) usually signiﬁes the initiation of breakaway corrosion, which is generally followed by rapid corrosion attack. Another example of component failure due to sulﬁdation attack is illustrated in Fig. 7.2. Alloy 800H suffered breakaway corrosion. The chromium oxide scale broke down and was replaced by iron-rich oxides, nickel-rich sulﬁdes, and chromium-rich sulﬁdes, as illustrated in Fig. 7.2(b). The alloy matrix beneath the corrosion products was found to be severely depleted of chromium, as shown in Fig. 7.2(b). Materials problems related to sulﬁdation attack have been encountered in various industries. Some of these problems have been reported in calcining of mineral and chemical feedstock (Ref 1), petrochemical processing (Ref 2), fossilﬁred boilers (Ref 2), petroleum reﬁning (Ref 3), coal gasiﬁcation (Ref 4, 5), waste incineration (Ref 6, 7), ﬂuidized-bed coal combustion (Ref 8–10), and oil-ﬁred boilers. This chapter reviews corrosion data involving mainly gaseous environments. The data are grouped into three different types typical of industrial environments: (1) H2/H2S mixtures and sulfur vapor with extremely low oxygen activities such that a Cr2O3 scale is not thermodynamically stable, (2) reducing, mixed-gas environments with sufﬁciently high oxygen activities such that Cr2O3 is thermodynamically stable, and (3) SO2-bearing environments. Sulﬁdation accelerated by salt deposits (hot corrosion in gas turbines, coal ash corrosion, and

oil ash corrosion) are covered in Chapters 9 to 11, respectively.

7.2 Thermodynamic Considerations Iron, nickel, and cobalt are the alloy bases for a majority of high-temperature alloys. Most hightemperature alloys rely on chromium to form a protective chromium oxide scale to resist oxidation and other high-temperature corrosion attack. When the sulfur potential is high enough, sulﬁdes of iron, nickel, cobalt, and chromium are likely to form under sulﬁdation attack. An Ellingham diagram such as the one shown in

Fig. 7.2

Fig. 7.1

Alloy 601 tube suffering localized sulfidation attack. The tube was in service at about 930 °C (1700 °F) in a natural gas-fired furnace making ceramic tiles. Sulfur was believed to come from the ceramic feedstock.

Alloy 800H recuperator suffering severe sulfidation attack in a nonferrous metal scrap melting furnace. The 9.5 mm (0.4 in) thick recuperator was perforated in less than 2 years at metal temperatures of about 650 to 760 °C (1200 to 1400 °F). (a) General view of a corroded sample. (b) Cross section of the sample showing corrosion products. 1, Fe-rich oxide (57Fe18Ni-16Cr-5S-4K); 2, Fe-rich oxide (80Fe-8Ni-12Cr); 3, Ni sulfide (light gray phase) (72Ni-2Cr-27S); 4, Cr depleted zone (72Ni28Fe); 5, Cr-rich sulfide (internal phase) (40Cr-10 Fe-50S).

pg 202

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

Chapter 7:

Fig. 7.3 (Ref 11) can help determine whether an environment has a sulfur potential high enough to form sulﬁdes. The sulfur potential is represented by either pS2 or pH2 S =pH2 . The sulfur partial pressure (pS2 ) in equilibrium with a sulﬁde

Sulfidation / 203

can be read from Fig. 7.3 by drawing a straight line from point “S” through the free-energy line of the sulﬁde phase through the temperature of interest, and intersecting with the pS2 scale. The intersection at the pS2 scale gives the sulfur

H2S/H2 ratio

p

Fig. 7.3

pg 203

Standard free energies of formation of selected sulfides. Source: Ref 11

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

pg 204

204 / High-Temperature Corrosion and Materials Applications

Metal-sulfur-oxygen stability diagrams at 870 °C (1600 °F) for iron, nickel, cobalt, and chromium are shown in Fig. 7.4 to 7.7 (Ref 15). The diagrams map out the regions within which different phases will be in equilibrium at the gas/ metal interface. Gordon and Nagarajan (Ref 16) constructed stability diagrams for the Fe-Cr-Ni system at 870 °C (1600 °F) as shown in Fig. 7.8. The spinel phases were deleted from the original ﬁgure for better clarity. In an environment marked “X,” for example, phases of Cr2O3, FeS, and NiS are likely to form at the gas/metal interface for the Fe-Cr-Ni system. Perkins (Ref 14) proposed that in the environment within the upper right region of the CrS-O diagram (Fig. 7.9) both CrS and Cr2O3 will form initially on the metal surface of chromium and high-chromium alloys. The Cr2O3 will grow and overtake the sulﬁde because it is a stable phase in this phase stability region. The alloying elements that form stable sulﬁdes can diffuse through the oxide scale and eventually form sulﬁdes on the surface of the oxide scale, leading to breakaway corrosion. It is particularly serious when the sulﬁdes become liquid. Some metalmetal sulﬁde eutectics have low melting

partial pressure in equilibrium with the sulﬁde phase. The pH2 S =pH2 value in equilibrium with a sulﬁde can be obtained similarly by using the starting point “H” for the straight line; the intersection with H2S/H2 scale gives the equilibrium H2S/H2 value. It is clear from Fig. 7.3 that the sulﬁde of chromium is more stable than those of iron, nickel, and cobalt. Most sulﬁdizing environments exhibit both sulfur and oxygen activities. It is thus better to characterize the environment in terms of pS2 and pO2 . Both sulfur and oxygen potentials of an environment containing various gaseous components in an equilibrium condition can be calculated using commercial software, such as HSC (Ref 12) and Chemsage (Ref 13). The environment in terms of pS2 and pO2 can be presented in a stability diagram of a metal-sulfur-oxygen system. Perkins (Ref 14) and Hemmings and Perkins (Ref 15) discuss brieﬂy how stability diagrams can be used to aid understanding corrosion behavior. Various metal-sulfur-oxygen stability diagrams at different temperatures can be found in Ref 15. Stability diagrams allow one to predict the phases that are likely to form on pure metals.

Log pSO /pSO 3 2 Log pCO /pCO

–26 –24

–22 –20 –18

–18 –16 –14

–16 –14

–12 –10 –8

–6

–4

–2

0

2

4

–12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

–16 –14 –12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

2

Log pH O /pH 2 2

–18

S(l )

2

0 Fe2(SO4)3(s)

FeS1+x (s)

–5

0 –2

FeSO4 (s)

–6

–15 –20

–8

Fe3O4 (s)

–10

Fe(s)

–25

Fe0.950(s)

–12

–30

–14 Fe2O3(s) –16

–35

–18 –50

–45

–40

–35

–30

–25

–20

–15

Log pO , atm 2

Fig. 7.4

Stability diagram of the Fe-S-O system at 870 °C (1600 °F). Source: Ref 15

–10

–5

0

5

Log pH S/pH 2 2

–4

2

Log pS , atm

–10

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

pg 205

Chapter 7:

Log pCO /pCO 2

–26 –24

–22 –20 –18

–18 –16 –14

Log pH O /pH 2 2

–18

–16 –14

–12 –10 –8

–6

–4

–2

0

2

4

–12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

–16 –14 –12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

S(l )

2

0

0

NiSy (l)

–5

–2

NiSO4 (s)

–4 –6

–15

–8

2

Log pS , atm

–10

–20

–10

Ni(s)

–25

Log pH S/pH 2 2

Log pSO /pSO 3 2

Sulfidation / 205

–12

NiO(s) –30

–14

–35

–16 –18 –50

–45

–40

–35

–30

–25

–20

–15

–10

–5

0

5

Log pO , atm 2

Fig. 7.5

Stability diagram of the Ni-S-O system at 870 °C (1600 °F). Source: Ref 15

–26 –24

Log pSO /pSO 3

–22 –20 –18

–16 –14

–12 –10 –8

–6

–4

–2

0

2

4

2

Log pCO /pCO 2

–18 –16 –14

Log pH O /pH 2 2

–18

–12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

–16 –14 –12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

S(l )

0

2

CoS1+x (s)

0

–5 –10

–2

–6

–15

–8

–20

–10

CoO(s)

Co(s)

–25

–12 –30

–14 Co3O4 (s)

–35

–16 –18

–50

–45

–40

–35

–30

–25

–20

–15

–10

Log pO , atm 2

Fig. 7.6

Stability diagram of the Co-S-O system at 870 °C (1600 °F). Source: Ref 15

–5

0

5

Log pH S/pH 2 2

–4

2

Log pS , atm

CoSO4 (s)

Co4S3+x (s)

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

pg 206

206 / High-Temperature Corrosion and Materials Applications

chromium oxide scale for protection against sulﬁdation. Most industrial environments have sufﬁcient oxygen activities to form a chromium oxide scale. The scale may eventually break

points: 635 °C (1175 °F) for Ni-Ni3S2, 880 °C (1616 °F) for Co-Co4S3, and 985 °C (1805 °F) for Fe-FeS (Ref 17). Most high-temperature alloys are chromia formers, which rely on

–26 –24

Log pSO /pSO 3

–22 –20 –18

–16 –14

–12 –10 –8

–6

–4

–2

0

2

4

2

Log pCO /pCO 2

–18 –16 –14

Log pH O /pH 2 2

–18

–12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

–16 –14 –12 –10

–8

–6

–4

–2

0

2

4

6

8

10

12

S(l )

0

2 0

–5

CrS (s) –2

–6

–15 Cr2O3 (s)

–20

–8

Log pH S/pH 2 2

–4

2

Log pS , atm

–10

–10 –25

Cr (s)

–12

–30

–14

–35

–16 –18 –50

–45

–40

–35

–30

–25

–20

–15

–10

–5

0

5

Log pO , atm 2

Fig. 7.7

Stability diagram of the Cr-S-O system at 870 °C (1600 °F). Source: Ref 15

0 Cr2 S3

Cr2 O3

FeS NiS

–4

FeO

Log pS

2

NiO –8 Fe

Ni

NiO

CrS FeO

–12 Fe Cr Ni

Cr2 O3

–16

–36

–32

–28

–24

–16

–20

–12

–8

Log pO

2

Fig. 7.8

Stability diagram for the M-S-O systems at 870 °C (1600 °F), where M stands for metal. Source: Ref 16

–4

0

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

Chapter 7:

down, leading to breakaway corrosion. Perkins (Ref 14) referred to this region as “limited life”; see the upper right region of the Cr-S-O diagram shown schematically in Fig. 7.9. The environments within this region typically contain H2, H2O, CO, CO2, H2S, and so forth, referred to in this chapter as “reducing, mixed-gas environments,” which are characterized by low oxygen and high sulfur potentials. Another type of sulﬁdizing environment with common characteristics for the sulﬁdation of various alloys is the upper left region in Fig. 7.9. In this region, sulﬁdes are stable phases. Alloys in these environments form sulﬁde scales. The environments include sulfur vapor and H2/H2S mixtures with extremely low oxygen activities such that Cr2O3 is not thermodynamically stable.

7.3 Sulfidation in Sulfur Vapor, Hydrocarbon Streams (No H2), H2S, and H2-H2S Mixtures

p

Sulfur vapor, hydrocarbon streams with no hydrogen (H2), H2S, and H2-H2S mixtures have a common feature in sulﬁdation reactions. This common feature is that the corrosion products in these environments are sulﬁdes. The data presented in this section are not applicable to other environments where oxides and sulﬁdes coexist. Sulﬁdation behavior of metals and alloys in sulfur vapor environments has been studied generally at sulfur pressures higher than 10−3 atm. Studies of H2-H2S mixtures have been typically carried out at sulfur partial pressures less than 10−2 atm. Mrowec and Przybylski (Ref 18) and Young (Ref 19) gave excellent reviews on sulﬁdation of metals and alloys in sulfur vapor and

p

Fig. 7.9

Stability diagram for the Cr-S-O system. Source: Ref 14

pg 207

Sulfidation / 207

H2-H2S environments. Sulﬁdation of pure metals, such as iron, nickel, and cobalt, and iron-, nickel-, and cobalt-base alloys was found to proceed with the formation of sulﬁdes (Ref 18, 19) in these environments. High sulfur potentials combined with extremely low oxygen potentials have resulted in formation of sulﬁdes. Thus, data generated in these types of environments are applicable only to industrial environments where sulﬁdes, but not oxides, are thermodynamically stable. Various investigations of sulﬁdation of iron (247–977 °C, or 475–1790 °F), nickel (400–640 °C, or 750–1185 °F), and cobalt (497– 1000 °C, 925–1830 °F) in sulfur vapor and H2-H2S mixtures were summarized by Mrowec and Przybylski (Ref 18). Sulﬁdation of these metals was found to follow a parabolic rate law, with sulﬁde scales forming on metal surfaces in all cases. Chromium also sulﬁdizes at a parabolic rate in sulfur vapor and H2-H2S mixtures with sulfur potentials of 10−7 to 1.0 atm at 700 to 1000 °C (1290 to 1830 °F), with only sulﬁde scales forming on metal surfaces (Ref 18). Data for Fe-Cr, Ni-Cr, and Co-Cr alloys were generated by several studies (Ref 20–26). Mrowec and Przybylski (Ref 18) summarized these data in Fig. 7.10, which shows that at sulfur potentials greater than 10−2 atm, all three alloy systems follow a similar trend of decreasing sulﬁdation rate with increasing chromium content. Sulﬁdation rates were expressed in terms of parabolic rate constants since all three alloy systems were found to follow a parabolic rate law. It is surprising to ﬁnd that there is no signiﬁcant difference in sulﬁdation resistance among the three alloy systems. It appears that the binary alloys with less than 15% Cr showed more scattering. Improved sulﬁdation resistance for alloys with less than 40% Cr (at.%) was attributed to the formation of an inner sulﬁde layer: Fe(Fe2−xCrx)S4 for Fe-Cr alloys, chromium sulﬁdes with nickel for Ni-Cr alloys, and chromium sulﬁdes with cobalt for Co-Cr alloys (Ref 18). In alloys with higher chromium (40% at.), a single layer of chromium sulﬁde (Cr2S3) was observed (Ref 18). Davin and Coutsouradis (Ref 20) investigated the relative sulﬁdation resistance of Fe-Cr, Ni-Cr, and Co-Cr alloys with 20 and 35% Cr in H2S at 800 °C (1470 °F). For 20% Cr alloys, the Co-Cr alloy was most resistant, followed by the Fe-Cr alloy. The Ni-Cr alloy was least resistant. When chromium was increased to 35%, however, all three alloys showed similar resistance.

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

208 / High-Temperature Corrosion and Materials Applications

Aluminum is a beneﬁcial alloying element in Fe-Cr and Co-Cr systems in both sulfur vapor and H2-H2S environments. The effect of aluminum on the parabolic rate constant of the

Fe-Cr-Al system was summarized in Fig. 7.11 by Mrowec and Przybylski (Ref 18) using data generated by various studies (Ref 27–30). The improved sulﬁdation resistance of Fe-Cr alloys

p

p

p p

Fig. 7.10

Parabolic rate constants for Fe-Cr, Ni-Cr, and Co-Cr alloys as a function of chromium content. Source: Ref 18

10–6 800 °C (1470 °F)

--Fe-19 Cr-AI, MROWEC,WEDRYCHOWSKA (1979) --Fe-17 Cr-AI, JALLOULI et al. (1979) --Co-25 Cr-AI, BIEGUN, BRÜCKMAN (1980) --Fe-25 Cr-AI, SMELTZER et al. (1982) pS = 1 atm 2

" g2cm4 s–1 kp,

10–7

10–8

pS = 1 atm 2

pS = 8.10–5atm 2 10–9

pS = 10–7atm

Fe-Cr-AI Co-Cr-AI

2

10

20

30

% at AI

Fig. 7.11

Parabolic rate constants for Fe-Cr-Al and Co-Cr-Al alloys as a function of aluminum content. Source: Ref 18

pg 208

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

Chapter 7:

due to aluminum additions was attributed to the formation of an inner sulﬁde layer of presumably Fe(FexAlyCrz-x-y)S4 (Ref 27). This sulﬁde decomposed at room temperature (Ref 27). The outer layer consisted of Fe1−yS (Ref 27). Biegun and Bruckman (Ref 29) reported a similar beneﬁcial effect in Co-25Cr alloys when aluminum content was varied 1 to 22% (at.%) at temperatures of 700 to 1100 °C (1290 to 2010 °F) in sulfur vapor at 1 atm. However, the inﬂuence of aluminum in the Co-Cr-Al system was not as effective as in the Fe-Cr-Al system (Fig. 7.11). The Ni-Cr-Al system behaved differently from both the Fe-Cr-Al and Co-Cr-Al systems (Ref 31). Biegun and Bruckman (Ref 31) investigated Ni-25Cr alloys containing 1 to 10 at.% aluminum in sulfur vapor (1 atm) at 580 to 950 °C (1080 to 1740 °F) with erratic results. The kinetics rates observed ranged from parabolic to linear. Corrosion rates of several austenitic stainless steels in sulfur vapor at 571 °C (1060 °F) are shown in Table 7.1 (Ref 32). Higher-chromium alloys, such as 314SS (24Cr, 2Si), 310SS (25Cr), and 309SS (23Cr), were found to perform the best. Silicon may not be very effective at this relatively low temperature. In addition to Type 314SS, which contains about 2% Si, there was another silicon-containing stainless steel, Type 302B (18Cr, 2.5Si). Type 302B performed quite similarly to Type 304SS (19Cr, no Si). For comparison, the corrosion rates of austenitic

Table 7.1 Corrosion rates of several austenitic stainless steels in sulfur vapor at 571 °C (1060 °F) for 1295 h Alloy

Corrosion rates, mm/yr (mpy)

314 310 309 304 302B 316 321

0.43 (16.9) 0.48 (18.9) 0.57 (22.3) 0.69 (27.0) 0.76 (29.8) 0.79 (31.1) 1.39 (54.8)

Source: Ref 32

Table 7.2 Alloy

Type 310 Type 304 Type 316 Source: Ref 32

pg 209

Sulfidation / 209

stainless steels in H2-50%H2S at 500 to 700 °C (932 to 1292 °F) is shown in Table 7.2 (Ref 32). Corrosion becomes much more aggressive in 100% H2S. Malinowski et al. (Ref 33) reported corrosion rates of iron and iron-aluminum alloys in 100 H2S at 400, 600, and 700 °C (752, 1112, and 1292 °F) (Fig. 7.12). Although aluminum was found to be beneﬁcial in corrosion rates of Fe-Al system, the content of aluminum that was needed to signiﬁcantly reduce the alloy’s corrosion rates was found to be about 10% at 600 and 700 °C (1112 and 1292 °F), and slightly lower at 400 °C (752 °F). Bruns (Ref 34) tested many commercial alloys in 100% H2S at 593 °C (1100 °F). Also included in the tests were experimental alloys based on Fe-32Ni-20Cr with various levels of aluminum (Ref 34). His test results are summarized in Fig. 7.13 in terms of the percent of either chromium or aluminum content in the alloy (Ref 34). The data generated in 100% H2S by Bruns indicate that increasing chromium signiﬁcantly improves the alloy’s sulﬁdation resistance. It also indicated that adding 8.5% Mn to Fe-18Cr alloy failed to improve the alloy’s corrosion resistance. For Fe-32Ni20Cr alloy (the base composition is similar to alloy 800/800H) with various aluminum contents, Fig. 7.13 shows strong beneﬁcial effect for aluminum in increasing the alloy’s sulﬁdation resistance. Furthermore, the aluminum content that signiﬁcantly increased the sulﬁdation resistance of Fe-32Ni-20Cr alloy was found to be in a range of only 2 to 4%. In reﬁneries, sulﬁdation, which is commonly referred to as “sulﬁdic corrosion” in reﬁnery industry, is a common materials problem. The temperatures of the reﬁnery equipment having sulﬁdation problems are typically between about 260 and 540 °C (500 and 1000 °F). Sulfur compounds originating from crude oils include polysulﬁdes, hydrogen sulﬁde, mercaptans, aliphatic sulﬁdes, disulﬁdes, and so forth (Ref 35). Sulﬁdation occurs in some processing units where no hydrogen is present in the system, such as crude distillation units. The crude distillation units that process mostly sweet crude oils (less

Corrosion rates of Types 310, 304, and 316 in H2-50%H2S at 500–700 °C (932–1292 °F) Corrosion rate 500 °C (932 °F), mm/yr (mpy)

Corrosion rate 550 °C (1022 °F), mm/yr (mpy)

Corrosion rate 600 °C (1112 °F), mm/yr (mpy)

Corrosion rate 700 °C (1292 °F), mm/yr (mpy)

0.91 (36) 1.12 (44) 1.50 (59)

1.45 (57) 1.57 (62) 2.44 (96)

2.79 (110) 2.95 (116) 4.45 (175)

8.94 (352) 10.2 (400) 10.8 (424)

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

210 / High-Temperature Corrosion and Materials Applications

than 0.6% total sulfur, with essentially no hydrogen sulﬁde) experience relatively less sulﬁdation problems. More sulﬁdation problems are encountered in the distillation units that process mostly sour crudes (Ref 36). The main fractionation tower is usually made of carbon steel with the lower part lined with a stainless steel containing about 12% Cr, such as Type 405 (Ref 37). Type 309L weld overlay cladding is most commonly used for replacing the ferritic stainless steel lining or cladding that suffered sulﬁdation attack. The corrosion rates of the crudes and their liquid fractions can be predicted by the so-called modiﬁed McConomy curves. Figure 7.14 shows the modiﬁed McConomy curves for liquid hydrocarbon streams containing 0.6 wt.% S (Ref 38). For hydrocarbon streams containing more than 0.6% sulfur, the corrosion rate multiplier as shown in Fig. 7.15 can be used in conjunction with the data shown in Fig. 7.14 for making the prediction of the corrosion rate for various steels at different temperatures (Ref 38). Hydrogen in hydrotreating, hydrocracking, and hydrodesulfurizing processes is used to remove sulfur (to convert it to hydrogen sulﬁde) and nitrogen (to ammonia) for separation from

Fig. 7.12

Corrosion rates of iron and iron-aluminum alloys in 100% H2S at 400, 600, and 700 °C (752, 1112, and 1292 °F). Source: Ref 33

the hydrocarbon streams (Ref 39). Hydrocracking processes combine desulfurization and cracking operation that can convert hydrocarbon feedstocks into various products (Ref 39). Sulﬁdation in these processing units is dictated by the H2S concentration in the H2-H2S environment. Sulﬁdation in H2-H2S environments is severe since the corrosion products are sulﬁdes with no protection by oxide scales in these environments. Austenitic stainless steels are generally used as a cladding or weld overlay for sulﬁdation resistance. Because of high hydrogen temperatures and pressures in the system, the vessels generally are made of low-alloy steels for resisting hydrogen attack, and the selection of materials to avoid hydrogen attack problem should follow the recommendations of the API Publication 941 (Ref 40). Materials issues related to hydrogen attack are discussed in Chapter 17: Hydrogen Attack. The reactors in hydrotreating and hydrocracking units are normally made of low-alloy steels with Type 347 cladding or weld overlay (Ref 41). The use of Type 347 (a stabilized grade of austenitic stainless steel) for cladding or weld overlay is to avoid intergranular stress-corrosion cracking by polythionic acid during downtime (Ref 39, 41). Corrosion rates of metals and alloys in H2-H2S environments are dependent on the H2S concentration. Corrosion of austenitic stainless steels was very aggressive in H2-50%H2S and 100% H2S environments, as shown in Table 7.2 and Fig. 7.13. However, the concentrations of hydrogen sulﬁde (H2S) in H2-H2S environments in hydrotreating, hydrocracking, hydrodesulfurizing, and catalytic reforming processes are signiﬁcantly lower, thus making carbon steels, low-alloy steels and stainless steels capable of resisting sulﬁdation attack in different temperature ranges. For example, Neumair and Schillmoller (Ref 42) reported the corrosion rates of steels and austenitic stainless steels as a function of hydrogen sulﬁde concentration for a hydrodesulfurization equipment at 415 °C (780 °F) and 27.2 atm (400 psig) (Fig. 7.16) and for a catalytic-reformer equipment at 510 °C (950 °F) and 27.2 atm (400 psig) (Fig. 7.17). The corrosion data in catalytic reforming units generated by different reﬁnery companies was summarized by Sorell (Ref 43) and are shown in Table 7.3 to illustrate the sulﬁdation of various steels in H2-H2S mixtures. Catalytic reforming is used in petroleum reﬁneries to upgrade the octane number of gasoline (Ref 44). Severe corrosion attack on processing equipment by

pg 210

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

Chapter 7:

hydrogen sulﬁde has been encountered in several catalytic reforming and desulfurizing units (Ref 43, 45, 46). In Table 7.3, Sorell (Ref 43) reported hydrogen sulﬁde corrosion data generated by various reﬁnery operators in laboratory tests, pilot plant testing, ﬁeld testing in commercial units, and inspection of commercial operating equipment. Austenitic stainless steels (18Cr-8Ni) were most resistant, followed by straight chromium stainless steels (12–16% Cr), with low-chromium steels (0–9% Cr) being worst.

Fig. 7.13

pg 211

Sulfidation / 211

Sorell and Hoyt (Ref 45) reported the sulﬁdation resistance of Fe-Cr and Fe-Cr-Ni alloys in H2-16H2S at 315 to 480 °C (600 to 900 °F) (Fig. 7.18). Slight to moderate addition of nickel (e.g., 8 or 40%) to Fe-20Cr improved the alloy sulﬁdation resistance. Greater nickel addition (e.g., 65% or more) reduced sulﬁdation resistance. Figure 7.18 also surprisingly revealed that Fe-20Cr-65Ni exhibited a sulﬁdation resistance similar to that of Fe-20Cr (Ref 45). Sulﬁdation of alloys in H2-H2S mixtures has been described by isocorrosion rate curves, which show corrosion

Corrosion rates of commercial alloys as well as experimental Fe-30Ni-20Cr with various aluminum contents in 100% H2S at 593 °C (1100 °F). Note the upper curve shows the corrosion rates as a function of chromium content in the alloy, while the lower curve show the corrosion rates as function of Al content in the alloy. Note: alloy 202 (Fe-18Cr-8.5Mn), CL (Fe-32Ni-20Cr-0.25C), CI (Fe-32Ni-21Cr-0.025C), 804 (Fe-41Ni-29Cr-0.97Ti), 2.24% Si alloy (Fe-32Ni-21Cr-2.24Si). Source: Ref 34

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:13AM Plate # 0

212 / High-Temperature Corrosion and Materials Applications

rate as a function of H2S concentration and temperature. Backensto and Sjoberg (Ref 48) determined isocorrosion rate curves for chromium steels (0 to 5% Cr) and austenitic stainless steels, shown in Fig. 7.19 and 7.20, respectively. “Couper-Gorman” curves are generally believed to provide the most practical correlation among corrosion rates, temperature and

concentration of H2S for various steels (Ref 38). These curves were based on a survey conducted by NACE Committee T-8 on Reﬁning Industry Corrosion (Ref 49). The modiﬁed Couper-Gorman curves (i.e., the original curves were extended to higher concentrations of H2S with dashed lines) for 5Cr-0.5Mo steels (also applicable to carbon steels) and 18Cr-8Ni steel are shown in Fig. 7.21 and 7.22, respectively. It was found that total pressure between 1 and 18 MPa (150 and 2650 psig) was not a signiﬁcant variable. The authors also observed that no sulﬁdation occurred at very low H2S concentrations and at temperatures above 315 °C (600 °F) because formation of iron sulﬁdes in that regime is not feasible thermodynamically.

7.4 Sulfidation in Environments with Low-Oxygen and High-Sulfur Potentials

Fig. 7.14

Modified McConomy curves providing predicted corrosion rates as a function of temperatures for various steels in a liquid hydrocarbon stream containing 0.6% S. Source: Ref 38

Fig. 7.15

Effect of sulfur content in hydrocarbon streams on corrosion rates predicted by the modified McConomy curves in the 290–400 °C (550–750 °F) range. Source: Ref 38

The environments covered in this section typically contain H2, CO, CO2, H2O, H2S, and other gaseous components with oxygen and sulfur potentials high enough to form oxides and sulﬁdes for most high-temperature alloys. These types of reducing environments are generated by combustion under substoichiometric conditions with insufﬁcient air or oxygen. The corrosion reactions in these environments generally involve oxidation and sulﬁdation. Sulﬁdation is usually the determining factor in the alloy’s performance capability. In most cases, the alloy remains protected during the oxidation period until the initiation of breakaway corrosion, which is then followed by rapid sulﬁdation attack by forming primarily sulﬁdes. Sulﬁdation behavior of alloys in this type of environments, where oxidation and sulﬁdation reactions are involved, has been extensively studied during the 1970s, 1980s, and early 1990s when coal gasiﬁcation programs had been actively pursued. Coal gasiﬁcation atmospheres typically have sulfur potentials (pS2 ) of 10−5 to 10−10 atm and oxygen potentials (pO2 ) of 10−15 to 10−20 atm (Ref 50). Signiﬁcant understanding of corrosion reactions in these environments has been achieved (Ref 51–67). In addition, a large engineering database on commercial alloys has been generated. Major contributions to this database came from Metal Properties Council (MPC) programs from 1972 to 1985 that evaluated more than 80 commercial alloys and coatings for coal gasiﬁcation. The tests were

pg 212

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

Chapter 7:

carried out mostly by IIT Research Institute, and results were documented in MPC annual reports (Ref 51–59). All these results were summarized

pg 213

Sulfidation / 213

by Howes (Ref 60) and Verma (Ref 63). The test gas mixtures used are tabulated in Table 7.4, and their thermodynamic potentials (pS2 and

Fig. 7.16

Maximum anticipated corrosion rates for alloys in hydrodesulfurization equipment at 415 °C (780 °F) and 27.2 atm (400 psig). Source: Ref 42

Fig. 7.17

Maximum anticipated corrosion rates for alloys in catalytic reforming equipment at 510 °C (950 °F) and 27.2 (400 psig). Source: Ref 42

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

pg 214

214 / High-Temperature Corrosion and Materials Applications

Table 7.3 Summary of corrosion data in H2-H2S mixtures typical of catalytic reforming generated by various petroleum refining companies Type of corrosion data(a)

Source

Temperature, °C (°F)

Test duration, days

Hydrogen pressure, atm (psi)

H2S concentration, vol%

Materials

Corrosion rate, mm/yr (mpy)

1.5 (59) 1.5 (59) 0.76 (30) 0.33 (13) 0.1–2.5 (5–100) 0.03–1.5 (1–60) 0.03–0.2 (1–8) 1.0 (38) 0.18 (7) 1.0 (40) 0.4 (16) 0.2 (8) 0.06 (2.5) 1.9–10 (73–398) 0.1–1.1 (5–42) 0.03–10 (1–400) 0.01–1.5 (0.4–60) 2.3 (90) 5.1 (200) 0.3–1.0 (13.5–41.5)

American Oil

I C

510 (950) 510–550 (950–1025)

36–89 89

18–27 (265–400) 18–27 (265–400)

0.03–0.07 0.03–0.07

Atlantic

L, P, C, I

480 (900)

1–365

34 (500)

I

480 (900)

180

34 (500)

D-X Sunray Canadian Petroﬁna

I C

465 (870) 480–490 (890–920)

180 90

27 (400) 21–22 (310–330)

0.013–0.074 0.011–0.13 0.015–0.27 0.036 0.016 0.04 0.008

Humble

C

480–510 (900–956)

1–4

20 (300)

0.013–0.14

P

540 (1000)

4

20 (300)

0.007–0.15

I

450–520 (850–975) 480–550 (900–1025) 470–520 (875–960)

127

20 (300)

0.035

139–577

34–36 (500–535)

0.035–0.09

1¼Cr, 2¼Cr 0–9Cr 12Cr 18Cr–8Ni 0–5Cr 11½–13½Cr 18Cr–8Ni 2¼Cr 12Cr 5Cr 0–9Cr 12Cr 18Cr–8Ni 0–12Cr 18Cr–8Ni 0–12Cr 18Cr–8Ni 1¼Cr CS, 2¼Cr 2¼Cr, 5Cr

0.015

12Cr 18Cr–8Ni 0–5Cr

0.28–0.7 (11–27) 0.1 (3.7) 0.5 (18.9)

12Cr 18Cr–8Ni 2Cr 0–9Cr 12Cr 18Cr–8Ni 0–9Cr 12Cr 5Cr 9Cr 0–9Cr 12Cr 18Cr–8Ni 0–7Cr 13Cr 18Cr–8Ni 0–7Cr 13Cr 18Cr–8Ni 5Cr 12Cr 0–5Cr 7–16Cr Cr–Ni 5Cr 0–9Cr 12–16Cr Cr–Ni 0–9Cr 12Cr Cr-Ni 0–9Cr 12Cr 18Cr–8Ni CS

0.2 (8.3) 0.1 (4.5) 0.8 (33) 0.4–7.9 (15–310) 0.3–6.1 (10–240) 0.1–0.5 (3.5–20) 0.5 (18) 0.2 (8) 1.4–3.4 (56–135) 1.7 (68) 0.9–1.2 (35–47) 0.4–0.6 (15–22) 0.1–0.2 (4.3–7.2) 2.0–11 (80–450) 0.5–5.3 (20–210) 0.1–1.2 (3–48) 0.1–0.3 (2.4–10) 0.05–0.2 (1.9–6.5) 0.01–0.02 (0.55–0.6) 0.2 (90). 0.2 (6.5) 0.03–7.2 (1–285) 0.03–2.6 (1–102) 0.03–0.2 (1–8.7) 1.1 (44) 2.7 (108) 0.5 (21) 0.1 (5) 0.5–1.6 (18–62) 0.3–0.6 (13–23) 0.02–0.1 (0.7–4.3) 0.3–13 (10–500) 0.1–5 (4–200) 0.03–1.6 (1.3–62.5) 1.8 (70) 1.1 (45) 1.0 (37.5) 0.1 (3.3) 0.2–0.8 (6.1–29.7) 0.1–0.3 (5.8–11.7) 0.02–0.05 (0.7–1.9) 0.07 (2.9) 0.05 (1.9) 0.01 (0.3)

Major U.S. Ref. Comp. Major West Coast Ref. Comp.

C

C

480–490 (900–920)

72

37–38 (540–565)

Pure Oil Richﬁeld

I L, P, C, I

480–540 (900–1000) 510 (950)

240 16.7

34 (500) 27 (400)

Shell Oil

C

510 (950)

71

43 (625)

0.026 0.011–1.0 0.014–1.0 0.025–1.0 0.015

I C

510 (950) 510–610 (955–1135) 500 (925)

148–758 758 148–251

46 (680) 46 (670) 46 (680)

0.05–0.08 0.08 0.05–0.08

L

510 (950)

4.2

34 (500)

C

490–500 (910–940)

212–382

34 (500)

0.08–1.0 0.1–1.0 0.1–1.0 0.001–0.0089

I

120–270

34 (500)

0.005

L

480–550 (900–1025) 480–510 (900–950) 530 (985)

6.3–20.8

33 (485)

L C

490 (905) 470–500 (885–935)

21 55

31 (460) 12 (175)

0.009–0.2 0.017–0.2 0.06–0.2 0.085 0.5

C

470–490 (885–920)

50.5–213

37–39 (550–575)

0.018–0.04

L

480 (900)

0.7–11.7

19 (285)

0.013–0.44

C

16.5–46

20 (300)

C

510 (950) 510–540 (950–1000) 550 (1030)

47

23 (340)

0.023 0.032 0.029

Sun Oil

C

500–520 (925–960)

188–395

20–41 (300–600)

0.017–0.04

Texas

C

500 (940)

133–668

41 (600)

0.0016

Sinclair

Socony Mobil

Standard Oil (Ind.)

0–9Cr 18Cr–8Ni 0–5Cr 12 Cr 18Cr–8Ni 0–5Cr 12Cr 18Cr–8Ni

(a) L, laboratory corrosion tests; P, pilot plant corrosion tests; C, commercial unit corrosion tests; I, inspection of commercial operating equipment. Source: Ref 43

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

Chapter 7:

pO2 ) at various test temperatures and pressures are shown in Table 7.5. The test environments in terms of pO2 and pS2 in the phase-stability diagrams for two test temperatures—650 and 980 °C (1200 and 1800 °F)—are shown in Fig. 7.23 and 7.24 (Ref 63). Most alloys suffered breakaway corrosion in several thousands of hours or less. Breakaway corrosion is illustrated in Fig. 7.25 (Ref 60). Both Type 310SS and alloy 800H followed a parabolic reaction rate prior to rapid corrosion attack. However, a few alloys did not exhibit breakaway corrosion within the test duration of up to 10,000 h, including the Co-Cr-W alloy 6B as shown in Fig. 7.25. Verma (Ref 63) discussed in detail the performance of various alloys in the MPC environments. Chromium was the most important alloying element in resisting sulﬁdation attack. Most alloys were protected by chromium oxide scales. The scale may eventually break down, leading to breakaway corrosion (Fig. 7.26) (Ref 63). Breakdown of the protective chromium oxide scale results from the sulﬁdes formed over it (Ref 14, 61). Formation of liquid

Fig. 7.18

Corrosion rates of Fe-Cr and Fe-Cr-Ni alloys in H216H2S at 315 to 480 °C (600 to 900 °F) and 75 atm (1100 psig) pressure. Source: Ref 45, summarizing results of Ref 47

pg 215

Sulfidation / 215

sulﬁde slag over the chromium oxide scale is most damaging. Alloying elements, such as manganese, iron, cobalt, nickel, and so forth, diffuse through the chromium oxide scale and react with the environment on top of the oxide scale to form external sulﬁdes. This was proposed by Perkins (Ref 14) as a possible mechanism for breakaway corrosion, schematically illustrated in Fig. 7.27. Manganese is the fastest diffusing element, followed by iron, cobalt, nickel, and chromium (Ref 14). In addition to outward diffusion of alloying elements to form external sulﬁdes, the corrosion reaction also involves sulfur penetrating through the oxide scale to form discrete particles of sulﬁdes in the matrix (Ref 14). Natesan (Ref 62) showed that for a given sulfur potential, there exists a threshold value for oxygen potential beyond which a continuous protective oxide scale is developed. This threshold oxygen partial pressure (kinetic boundary) is about 103 times the oxygen partial pressure for chromium oxide and

Fig. 7.19

Corrosion rates of chromium steels (0–5% Cr) generated from laboratory tests in H2-H2S at hydrogen pressures of 12 to 34 atm (175 to 500 psig) as a function of H2S concentration and temperature. IPY, inch per year. Source: Ref 48

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

216 / High-Temperature Corrosion and Materials Applications

chromium sulﬁde equilibrium (thermodynamic boundary) (Ref 62). The sulﬁdation resistance of various commercial alloys generated in the MPC programs is summarized in Fig. 7.28 to 7.31 (Ref 60) and Fig. 7.32 and 7.33 (Ref 67). For nickel-base alloys, increasing nickel generally increases susceptibility to sulﬁdation attack (Fig. 7.28). High-nickel alloys, such as alloys 600 and 601, are particularly susceptible to sulﬁdation attack, as illustrated in Fig. 7.32 (Ref 67). The Ni-Ni3S2 eutectic melts at 635 °C (1175 °F). Molten sulﬁde slag can easily destroy the chromium oxide scale and cause catastrophic sulﬁdation attack. Increasing chromium generally improves sulﬁdation resistance in Fe-Cr-Ni alloys (Fig. 7.29) (Ref 60). This also holds true for iron- and nickelbase alloys. Type 446SS (Fe-27Cr) and 50Ni50Cr alloys (alloys 671 and 657) are good examples (Fig. 7.30) (Ref 60). Despite its high nickel content, the 50Ni-50Cr alloy exhibited good sulﬁdation resistance due to its extremely high chromium content.

In the MPC data summarized in Figs. 7.28 to 7.31, many stainless steels and nickel alloys, while showing very high wastage rates at high temperatures (e.g., 816 °C and higher), showed low corrosion rates at 482 and 650 °C (900 and 1200 °F). Review of the test data generated at 650 °C (1200 °F) for 10,000 h in the MPC test programs summarized in Howes’s report (Ref 60) showed that alloys containing at least 18% Cr, which included 304, 309, 310, 312, 329, 22-13-5, 29-4-2, Ebrite 26-1, 446, 20Cb-3, 800, 253MA, Crump 25, 825, RA333, 617, and Multimet, exhibited corrosion rates of 0.02 mm/yr or less (1 mpy or less) when tested in a 500 psig gas pressure and 1% H2S in the MPC coal gasiﬁcation environment. Type 410 (12Cr) showed the corrosion rate of 0.05 mm/yr (2 mpy), and 9Cr-1Mo steel suffered the corrosion rate of 1.01 mm/yr (40 mpy) (Ref 60). John et al. (Ref 68) reported the corrosion data generated at 700 °C (1292 °F) and lower (1 atm pressure) for various commercial alloys in syngas environments that were rich in CO and H2 with low water. The corrosion data for carbon

Fig. 7.20

Fig. 7.21

Corrosion rates of Cr-Ni austenitic stainless steels generated from laboratory tests in H2-H2S at hydrogen pressures of 12 to 34 atm (175 to 500 psig) as a function of H2S concentration and temperature. IPY, inch per year. Source: Ref 48

Modified Couper-Gorman curves showing corrosion rates as a function of H2S concentration (mol. %) and temperatures for 5Cr-0.5Mo steel. The data also is applicable to carbon steels and alloy steels with less than 5% Cr (naphtha desulfurizers). Source: Ref 38

pg 216

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

pg 217

Chapter 7:

steel and 1Cr-0.5Mo steel tested in CO-51.2H20.8CO2-1.7H2S-0.2HCl at 330, 400, and 500 °C (626, 752, and 932 °F) are summarized in Fig. 7.34 (Ref 68). The corrosion behavior of alloy 625 in the same test environment at 330 to 700 °C (626 to 1292 °F) is summarized in Fig. 7.35 (Ref 68). The data for various Fe-, Ni-, and Co-base alloys at 700 °C (1292 °F) and lower in CO-26H2-1.0CO2-0.9 H2S-2H2O-10N2 is summarized in Fig. 7.36 (Ref 68). It was surprising that Co-base alloys, such as alloys 25

Sulfidation / 217

Table 7.5 Oxygen and sulfur potentials ( pO2 and pS2 ) in the MPC-CGA test gas mixtures at 68 atm (1000 psig) Temperature, °C (°F)

H2S in test gas, vol%

pO2 (atm)

pS2 (atm)

0.1 0.1 0.1 0.1 0.1 0.5 0.5 0.5 0.5 0.5 1.0 1.0 1.0 1.0 1.0

1.2 × 10−23 7.0 × 10−21 3.0 × 10−18 6.4 × 10−17 1.3 × 10−15 1.2 × 10−23 7.0 × 10−21 3.0 × 10−18 6.4 × 10−17 1.3 × 10−15 1.2 × 10−23 7.0 × 10−21 3.0 × 10−18 6.4 × 10−17 1.3 × 10−15

1.5 × 10−10 8.8 × 10−10 5.4 × 10−9 1.3 × 10−8 3.1 × 10−8 3.5 × 10−9 2.2 × 10−8 1.4 × 10−7 3.1 × 10−7 7.6 × 10−7 1.5 × 10−8 8.8 × 10−9 5.5 × 10−7 1.3 × 10−6 3.1 × 10−6

480 (900) 650 (1200) 816 (1500) 900 (1650) 980 (1800) 480 (900) 650 (1200) 816 (1500) 900 (1650) 980 (1800) 480 (900) 650 (1200) 816 (1500) 900 (1650) 980 (1800) Source: Ref 60

p

NiO Co/CoO

p

Fig. 7.23

Fig. 7.22 Modified Couper-Gorman curves showing corrosion rates as a function of H2S concentration (mol. %) and temperatures for 18Cr-8Ni steel. Source: Ref 38

Table 7.4 mixtures

The phase-stability diagram and the test environment (star) in terms of equilibrium pO2 and pS2 in the MPC coal gasification test programs at 650 °C (1200 °F). Source: Ref 63

Inlet and equilibrium gas compositions at different temperatures for the MPC-CGA test gas Equilibrium gas composition(a), vol%

Gaseous component

H2 CO CO2 CH4 NH3(b) H2S H2O

Inlet gas, vol%

480 °C (900 °F)

650 °C (1200 °F)

816 °C (1500 °F)

900 °C (1650 °F)

980 °C (1800 °F)

24 18 12 5 1 0–1.5 bal

4 5 25 19 1 0.5 bal

11 8 22 14 1 1.5 bal

23 11 19 9 1 0.1–1.5 bal

27 14 17 6 1 0.5–1.0 bal

31 17 15 3 1 0–1.0 bal

(a) Computer calculations at 6.9 MPa (1000 psig). (b) NH3 will decompose to N2 and H2 on increasing temperature. Source: Ref 60

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

218 / High-Temperature Corrosion and Materials Applications

(Co-20Cr-10Ni-115W) and 150 (Co-27Cr18Fe), were not signiﬁcantly better than Type 310, and but not as good as Type 446. The sulﬁdation behavior of Co-base alloys at higher temperatures is discussed later. Bakker (Ref 69) and Bakker and Bonvallet (Ref 70) investigated the effect of H2O concentration on the corrosion behavior of Type 310 and alloy 800 in coal gasiﬁcation environments at 540 °C (1004 °F) for 600 h. The test gas environments with different levels of H2O from 0 to 15% along with oxygen partial pressures (pO2 ) at 540 °C (1004 °F) are tabulated in Table 7.6. The test results are summarized in Fig. 7.37. Both oxides and sulﬁdes were observed to have formed on both alloys in all test environments. Type 310 was found to suffer signiﬁcantly less metal loss than alloy 800, and its corrosion was relatively unaffected by H2O concentrations. Alloy 800 suffered much higher metal loss, with highest metal loss surprisingly at about 3% H2O. Long-term tests are needed to conﬁrm this “abnormal” observation. Cobalt-base alloys as well as cobaltcontaining alloys (e.g., Fe-Ni-Co-Cr) generally have better sulﬁdation resistance than nickelbase alloys and Fe-Cr-Ni alloys (Fig. 7.33) (Ref 67). Alloys 6B and 188 (cobalt-base alloys) and alloy N155 (Fe-Ni-Co-Cr alloy) are better than alloys RA333 and X (nickel-base alloys) and alloys 800H, HK-40, and Type 310SS

(Fe-Ni-Cr alloys). In general, cobalt-base alloys and cobalt-containing alloys have higher temperature capabilities and are more resistant to breakaway corrosion. Results generated by Lai (Ref 71) at 760, 870, and 980 °C (1400, 1600, and 1800 °F) also revealed that cobalt-base

Fig. 7.25

Corrosion behavior of Type 310 stainless steel, alloy 800, and alloy 6B at 980 °C (1800 °F) in the MPC coal gasification atmosphere, showing breakaway corrosion for Type 310SS and alloy 800. Inlet gas: 24H2-18CO-12CO25CH4-1NH3-0.5H2S (balance H2O) at 6.9 MPa (100 psig). pO2 = 1.3 × 10−15 atm. pS2 = 7.6 × 10−7 atm. Source: Ref 60

Fe/FeS –5 Ni/NiS Cr/CrS

–40

–35

–30

–25

–20

FeO/Fe

Al2O3/Al

–25

–30

Cr2O3/Cr

–20

–15

Ni/NiO

Al/AlS –15

FeO/Fe3O4 Fe3O4/Fe2O3

2

Log pS , atm

–10

–10

–5

Log pO , atm 2

Fig. 7.24

The phase stability diagram and the test environment (star) in terms of equilibrium pO2 and pS2 in the MPC coal gasification test programs at 980 °C (1800 °F). Source: Ref 63

pg 218

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

Chapter 7:

Fig. 7.26

pg 219

Sulfidation / 219

Corrosion behavior of alloy 800 in the MPC coal gasification atmosphere (0.5% H2S at 900 °C or 1650 °F, and 6.9 MPa or 1000 psig) showing oxide scales during the protective stage and oxides/sulfides after breakaway corrosion. Source:

Ref 63

alloys were more sulﬁdation resistant than nickel-base and Fe-Ni-Cr alloys with similar chromium contents (Fig. 7.38). Higher chromium (preferably more than 25%) is needed for an alloy to form a protective oxide scale in coal gasiﬁcation environments. Verma (Ref 63) suggested that chromium oxide scale would be more protective if doped with cobalt, as was the case in alloy 6B (Co-Cr-W) that formed a (Cr,Co)2O3 scale. Bradshaw et al. (Ref 65) believed that titanium modiﬁed the Cr2O3 scale, thus improving the alloy’s sulﬁdation resistance. They compared 3% Ti-modiﬁed Type 310SS with Type 310SS after exposure in an MPC gas mixture with 1% H2S at about 1.0 atm pressure for 100 h at 980 °C (1800 °F).

Fig. 7.27

Formation of external sulfides on top of the chromium oxide scale and the formation of internal sulfides. Source: Ref 14

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

220 / High-Temperature Corrosion and Materials Applications

Fig. 7.28

Corrosion rates of high-nickel alloys in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Table 7.4 and 7.5 for gas composition). Source: Ref 60

330

Fig. 7.29

Corrosion of Fe-Cr-Ni alloys in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas composition). Source: Ref 60

pg 220

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

Chapter 7:

Type 310 sample showed some sulﬁde nodules as well as some spalled oxides, while Ti-modiﬁed Type 310 sample showed an adherent oxide scale with no sulﬁde nodules. Their analysis indicated a signiﬁcant amount of titanium in the Cr2O3 scale (Ref 65). They also tested alloys 800 and 801 in the same environment. Alloy 800 sample was totally corroded after 100 h at 980 °C (1800 °F), while the alloy 801 sample showed an adherent oxide scale with only about 1.4 mg/cm2 weight gain. The only apparent difference in the chemical composition between the two alloys is titanium (about 0.4% for alloy 800 and about 1.1% for alloy 801). Additional data indicating the beneﬁcial effect of titanium are shown in Table 7.7 (Ref 64). Results obtained by

Matl 446 302 304 309 310 314 316 800 793 600 601 671 310 (Al)

Sound metal loss, mils/yr

482

Temperature, °C 650 816 900

800 (Al)

982

310 (Cr)

Completely corroded

800 (Cr)

0.1 0.5 1.0 0.1 0.5 0.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0 0.1 0.5 1.0

NT

Scaling loss NT 0.5% H2S

NT

671 657

<2

446 671 657 900

1800

1650 1500 1200 Temperature, °F

Internal penetration Not tested Welded specimens

Completely corroded

125

NT 125 75.1

Completely corroded 43.8

NT

NT 0

2

4

6

0

50

100

150

20–50 2–20

Sulfidation / 221

% H2S

446

>50

pg 221

12 8 10 Total corrosion, mil

14

16

18

20

200 250 300 350 Total corrosion, µ m

400

450

500

Fig. 7.32

Corrosion of stainless steels and nickel-base alloys at 816 °C (1500 °F) for 100 h in the MPC coal gasification atmosphere with 0.1, 0.5, and 1.0% H2S. Also included were aluminized Type 310 and alloy 800 [310 (Al) and 800 (Al], and chromized Type 310 and alloy 800 [310 (Cr) and 800 (Cr)] (see Tables 7.4 and 7.5 for gas composition). Source: Ref 67

Fig. 7.30

Corrosion of Fe-27Cr (Type 446) and 50Ni-50Cr alloys (alloy 671 and 657) in the MPC coal gasification stmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas composition). Source: Ref 60

Alloy 310 HK-40 Cru-25

Sound metal loss, mils/yr

482

Temperature, °C 650 816 900

N155

982 T63wC

Completely corroded

188-Breakway occurs in 2000–10,000 h

>50

Alloy X

20–50

556

2–50

N155

6B 188

6B

6B

556 188 <2 N155 RV-18 & 19 900

Fig. 7.31

RA333

800

RV-18 & 19

1200 1500 1650 Temperature, °F

671

1800

Corrosion of cobalt-base alloys (alloys 188 and 6B) and cobalt-containing alloys (alloys 556, N155, RV-18, and RV-19) in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas composition). Source: Ref 60

Time, h 1000 2000 1000 2000 5000 10,000 1000 2000 5000 10,000 1000 2000 5000 10,000 1000 2000 5000 10,000 1000 2000 5000 7278 1000 2000 5000 10,000 1000 2000 5000 10,000 1000 2000 5000 10,000 1000 2000 5000 7000 8000 1000 2000 5000 10,000

2.0

Legend Scaling Penetration

0.67

0

2

0

50

Fig. 7.33

1.5

0.9

4

6

8 10 12 14 Total corrosion, mil

16

18

20

100 150 200 250 300 350 400 450 500 Total corrosion, µm

Corrosion of Fe-Cr-Ni, nickel-base, and cobalt-base alloys at 980 °C (1800 °F) in the MPC coal gasification atmosphere with 0.5% H2S (see Tables 7.4 and 7.5 for gas composition). Source: Ref 67

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:14AM Plate # 0

pg 222

222 / High-Temperature Corrosion and Materials Applications

Lai (Ref 71) also revealed the beneﬁcial effect of titanium in resisting sulﬁdation attack. As shown in Fig. 7.38, alloy R-41 and Waspaloy alloy (both contain about 3% Ti) were the best of the

nickel-base alloys tested. In fact, they approached some cobalt-base alloys. Alloy 263 (2.5% Ti), while performing well at 760 °C (1400 °F), suffered severe sulﬁdation attack at 870 and 980 °C (1600 and 1800 °F) (Ref 71). Bradshaw et al. (Ref 64) also examined the effects of molybdenum and aluminum. Addition of 6% Mo to high-purity Type 310SS Temperature, °C 600

500 Alloy 600

1

Alloy 625 Alloy 25 10

Alloy DS AISI 310

0.1

AISI 446

Corrosion after 1 year, mm

Corrosion after 1 year, mils

100

700

1 1.00 1.04 1.08 1.12 1.16 1.20 1.24 1.28 1000/(Temperature), 1/K

Fig. 7.34

Total surface recession including internal attack for carbon steel and 1Cr-0.5Mo steel at 330, 400, and 500 °C (626, 752, and 932 °F) in CO-51.2H2-0.8CO2-1.7H2S0.2HCl. Source: Ref 68

Corrosion after 1 year, mils

100

Temperature, °C 600

700

500

HK4M Cr30A

1 Alloy 556

10

Alloy 188 Alloy 150 0.1

Cr35A

Corrosion after 1 year, mm

5000

1 1.00 1.04 1.08 1.12 1.16 1.20 1.24 1.28 1000/(Temperature), 1/K

Fig. 7.36

Total surface recession including internal attack for Fe-, Ni-, and Co-base alloys at 700 °C (1292 °F) and lower in CO-26H2-1.0CO2-0.9H2S-2H2O-10N2. Source: Ref 68

Table 7.6 Test gas environments (nonequilibrium) used for investigating the effect of water content on corrosion behavior of Type 310 and alloy 800 at 540 °C (1004 °F) for 600 h Gas composition, vol %

Fig. 7.35

Total surface recession including internal attack for alloy 625 at 330 to 700 °C (626 to 1292 °F) in CO-51.2H2-0.8CO2-1.7H2S-0.2HCl. Source: Ref 68

Component

No. 1

H 2O H2 CO CO2 H 2S HCl pO2 at 540 °C

0.0 1.0 2.0 3.0 5.0 10.0 15.0 32.0 30.2 29.2 28.8 28.2 28.2 30.0 64.0 64.0 64.0 64.0 60.0 52.0 45.0 3.2 4.0 4.0 4.0 6.0 9.0 9.2 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.04 0.04 0.04 0.04 0.04 0.04 0.04 10−29 10−28.4 10−27.8 10−27.4 10−7 10−26.4 10−26

Source: Ref 70

No. 2

No. 3

No. 4

No. 5

No. 6

No. 7

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

Chapter 7:

signiﬁcantly improved sulﬁdation resistance (Table 7.7). However, a layer of (Fe, Ni} sulﬁdes was observed on top of the chromium oxide scale. Thus, formation of molten sulﬁde slag is expected upon further exposure. No explanation was given in the paper by Bradshaw et al. (Ref 64) on the beneﬁcial effect of molybdenum. The test temperature used by Bradshaw et al. (Ref 64) was excessively high (i.e., 1000 °C). It would be of interest if the tests were conducted at much lower temperatures. Natesan (Ref 72) found that molybdenum and TZM (molybdenum with 0.5Ti and 0.04Zr) formed very thin adherent sulﬁde scales when tested at 871 °C (1600 °F) in a sulﬁdizing environment with low oxygen and high sulfur potentials. Test results are shown in Fig. 7.39 (Ref 72). He et al. (Ref 73) showed that Ni-10Mo was more resistant to sulﬁdation than nickel (Fig. 7.40) in short-term tests. The beneﬁcial effect of molybdenum in improving the alloy sulﬁdation resistance in the environments with low oxygen, high sulfur potentials would be of interest, since molybdenum has been used in many high-temperature nickel-base alloys to provide solid solution strengthening for the alloy. Some of these wrought alloys include alloys X, S, 617, 625, and R-41. Furthermore, molybdenum is also a major alloying element in Ni-Cr-Mo corrosion-resistant alloys, such as C-276, C-22, (622), 59, C-2000, and 686. These Ni-Cr-Mo alloys contain molybdenum in a range of 13 to 16%. High levels of molybdenum in these Ni-Cr-Mo alloys can cause a thermal

pg 223

Sulfidation / 223

stability issue when used at high temperatures because of formation of intermetallic phases during the long-term, high temperature exposures. However, application of these alloys at temperatures below 650 °C (1200 °F) may not raise this thermal stability issue. Addition of 3% Al to high-purity Type 310SS signiﬁcantly improved sulﬁdation resistance, as shown in Table 7.7. However, this composition was not considered worthwhile for further development because of expected fabrication problems (Ref 64). Manganese has a deleterious effect on an alloy’s sulﬁdation resistance. In their tests in Ar-40H2-15H2O-1.4H2S at 1000 °C (1830 °F) for 24 h, Bradshaw et al. (Ref 64) found that addition of 2% Mn adversely affected sulﬁdation resistance of Ni-30Cr and Type 310SS, particularly a high-purity 310SS (Table 7.8). Manganese diffuses the quickest through chromium oxide scale to form external sulﬁdes on top of it, thus causing accelerating corrosion (Ref 14). The effect of silicon was investigated by Nagarajan et al. (Ref 74) in Fe-18Cr alloys in 24H2-39H2O-18CO-12CO2-5CH4-1H2S-1NH3. The Fe-18Cr-2Si alloy exhibited sulﬁdation resistance signiﬁcantly better than that of Fe-18Cr-0.5Si alloy (i.e., 20 mg/cm2 versus slightly more than 200 mg/cm2) after 120 h at 980 °C (1800 °F) (pO2 =9:9 · 10716 atm, pS2 = 2:4· 1076 atm, and ac = 0.3 at the test temperature). A nickel-base wrought alloy (HR160 alloy)

Metal loss (µm) 200

Alloy 800 100

SS 310 20-35-3.25 0 0

5

10

15 % H2O

Fig. 7.37

Effect of H2O on the metal loss of Type 310, alloy 800, and experimental alloy Fe-20Cr-35Ni-3.25Si (20-35-3.25) tested at 540 °C (1004 °F) for 600 h. The test gas environments are shown in Table 7.6. Source: Ref 70

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

224 / High-Temperature Corrosion and Materials Applications

was developed by Lai (Ref 75) using a combination of silicon, chromium, and cobalt to maximize the alloy’s sulﬁdation resistance as well as its metallurgical stability, creep-rupture properties, and weldability. With high chromium (28%) and high silicon (2.75%), a protective oxide scale consisting of a chromium oxide scale with underlying silicon enrichment formed on the alloy surface (Ref 76, 77). In addition, cobalt provided further improvement in the alloy’s sulﬁdation resistance. Figure 7.41 shows the sulﬁdation resistance of alloy HR160 compared to some familiar alloys, such as alloys 556, 800H, and 600 in Ar-5H2-5CO-1CO2-0.15H2S

Fig. 7.38

at 870 °C (1600 °F) (Ref 77). The sulﬁdation resistance of alloy HR160 was quite comparable to that of cobalt-base alloy 6B and signiﬁcantly better than other cobalt-base alloys, such as alloys 188, 25, and 150 (Table 7.9). Norton et al. (Ref 78) conducted corrosion tests on several Fe-Ni-Cr and Ni-base alloys in comparison with alloy HR160 at 700 °C (1290 °F) for up to 1000 h in H2-7CO-1.5H2O-0.6H2S (pO2 =10723 atm, pS2 =1079 atm, and ac = 0.3 to 0.4). Their test environment, as located in the phase stability diagram at the test temperature, is shown in Fig. 7.42 (Ref 78), and their test results are summarized in Fig. 7.43 (Ref 78).

Corrosion of iron-, nickel-, and cobalt-base alloys after 215 h at (a) 760 °C (1400 °F), (b) 870 °C (1600 °F), and (c) 980 °C (1800 °F) in Ar-5H2-5CO-1CO2-0.15H2S. Source: Ref 71

pg 224

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

Chapter 7:

Norton et al. (Ref 79) also conducted corrosion tests on an alumina-forming alloy MA956 (Fe-20Cr-4.5Al-0.5Y2O3), which is an oxide-dispersion-strengthened alloy. Also included in the tests were HR160, 45TM (Ni-Cr-FeSi), HR 3C (Fe-25Cr-20Ni-0.5Nb-0.2N), and an experimental alloy (Fe-26Cr-39Ni-3.2V). Tests were conducted in CO-32H2-3.8CO2-0.2H2S at 600 °C (1112 °F) and the test environment in terms of pO2 and pS2 potentials in the phasestability diagram as shown in Fig. 7.44. Their test results are summarized in Fig. 7.45 (Ref 80). MA956 was found to be slightly less resistant compared with HR160 and 45TM. Norton and Levi (Ref 80) reported that a mixed Cr-Al oxide scale with Cr-rich oxide on the outer layer and Al-rich oxide on the inner layer formed on MA956, and Fe-rich sulﬁdes grew on the top of

pg 225

Sulfidation / 225

the oxide scale. On the other hand, a thin, continuous and adherent Si-rich oxide scale enriched in titanium in the outer region of the oxide scale formed on HR160, and Cr-rich sulﬁdes grew on the top of the oxide scale (Ref 80). The corrosion

(a)

Table 7.7 Results of corrosion tests at 1000 °C (1832 °F) for 100 h in Ar-30H2-30H2O-1H2S(a) Total affected depth, μm (mils)

Alloy

310SS

310HP(b) 310HP + 2% Ti

310HP + 3% Ti 310HP + 6% Mo

92 (3.6) 810 (31.9) 605 (23.8) >1500 (59.1) 650 (25.6) 330 (13.0) 62 (2.4) 38 (1.5) 38 (1.5) 34 (1.3) 12 (0.5)

310HP + 3% Al

Comments

Internal sulﬁdation Liquid sulﬁdes Liquid sulﬁdes Liquid sulﬁdes Sulﬁde penetration Sulﬁde penetration Adherent oxide Adherent oxide Adherent oxide Adherent oxide Spalling oxides and (Fe.Ni) sulﬁdes over Cr2O3 Spalling oxides Spalling oxides

5 (0.2) 5 (0.2)

(a) Test gas was at 1 atm; pO2 =3 · 10715 atm and pS2 =3 · 1076 atm: (b) HP indicates high-purity material. Source: Ref 64

(b)

Fig. 7.40

Weight change data comparing nickel (a) with Ni10Mo alloy (b) when tested at 550 and 600 °C (1022 and 1112 °F) in Ar-13.56H2-0.6H2O-1.89H2S (pO2 = 2.7 × −27 10 , pS2 = 4.1 × 10−8 at 600 °C; pO2 = 4.5 × 10−29, pS2 = 1 × 10−8 at 550 °C). Source: Ref 73

Weight change, mg/mm2

2.0 1.6

Alloy 800

Test temperature 871 °C p = 4.1 x 10–18 atm O p 2 = 9.4 x 10–7 atm

Table 7.8 Results of corrosion tests at 1000 °C (1832 °F) for 24 h in Ar-40H2-15H2O-1.4H2S(a)

S2

1.2

Alloy

0.6

Total affected depth, µm (mils)

310SS

56 (2.2)

310SS + 2% Mn

72 (2.8)

310 SS 0.4

Ta

TZM

Nb 0

0

30

60

90

120

Mo v 150

Exposure time, h

Fig. 7.39

Thermogravimetric data for refractory metals, such as, Mo, TZM, Nb, Ta, and V in comparison with conventional alloys in a sulfidizing environment with high sulfur and low oxygen potentials tested at 871 °C (1600 °F). Source: Ref 72

310HP(b) + 2% Mn Ni-30Cr Ni-30Cr-2Mn

420 (16.5) 38 (1.5) 1220 (48.0)

Comments

Sulﬁde layer Oxide layer Intergranular sulﬁdes Internal sulﬁdes Sulﬁde layer Oxide layer Internal sulﬁdes Liquid sulﬁdes Oxides Internal sulﬁdes Liquid sulﬁdes

(a) Test gas was gas at 1 atm; pO2 =3 · 10716 and pS2 =3 · 1076 atm: (b) HP indicates high-purity material. Source: Ref 64

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

pg 226

226 / High-Temperature Corrosion and Materials Applications

morphologies for both alloys after exposure for 2000 h are schematically illustrated in Fig. 7.46 (Ref 80). Additional low temperature sulﬁdation data generated in reducing environments can be found in Tables 10.5 and 10.6 and Figures 10.17 and 10.18.

7.5 Corrosion in SO2-Bearing Environments

binary alloys in SO2 and SO2-O2 environments. Metals and alloys exposed to these environments generally form oxides and/or sulﬁdes as corrosion products. Corrosion products and corrosion rates depend strongly on temperature. The

Table 7.9 Sulfidation resistance of Alloy HR160 compared to cobalt-base alloys at 870 °C (1600 °F) for 500 h(a) Alloy

Most corrosion studies of SO2-bearing environments have been conducted in either SO2 or O2-SO2 mixtures. Materials investigated include nickel (Ref 81–84), iron (Ref 85–87), cobalt (Ref 88–89), chromium (Ref 90), Co-Cr alloys (Ref 91), and Ni-Cr alloys (Ref 92–97). Kofstad (Ref 98) reviews the corrosion of pure metals and

Fig. 7.41

HR160 556 188 25 150 6B

Weight change, mg/cm2

Metal loss, mm (mils)

Maximum metal affected(b), mm (mils)

−0.5 183.6 40.6 33.0 131.7 10.7

0.01 (0.2) 0.52 (20.6) 0.19 (7.6) 0.10 (4.1) 0.26 (10.3) 0.01 (0.3)

0.13 (5.2) 0.90 (35.6) 0.60 (23.6) 0.37 (14.6) 0.72 (28.3) 0.08 (3.3)

(a) Ar-5H2-5CO-lCO2-0.15H2S; pO2 =3 · 10719 atm, pS2 =0:9 · 1076 atm: (b) Metal loss + maximum internal penetration. Source: Ref 76

Sulfidation resistance of alloy HR160 compared to those of alloys 556, 800H, and 600 after 215 h at 870 °C (1600 °F) in Ar-5H2-5CO-1CO2-0.15H2S (pO2 = 3 × 10−19 atm, pS2 = 0.9 × 10−6 atm). Samples were cathodically descaled before being mounted for metallographic examination. Source: Ref 77

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

pg 227

Chapter 7:

0 FeS2

reaction between SO2 and O2 leads to formation of SO3. The kinetics for this reaction are relatively slow without catalysts such as platinum. Most tests were conducted in conjunction with platinum catalysts to obtain an equilibrium SO3 partial pressure. SO3 is an important reactant in the corrosion reaction involving SO2 (Ref 83, 93). Corrosion rate was found to be dependent on the ratio of SO2:O2. When that ratio is 2:1, which gives the highest partial pressure of SO3, the corrosion rate is generally fastest, as illustrated in Fig. 7.47 (Ref 83). The corrosion reaction is also temperature dependent. Corrosion rate increases with increasing temperature until a maximum is reached. Further increases in temperature result in a reduced corrosion rate (Ref 83). With an SO2:O2 ratio of 2:1, nickel corroded fastest at 700 to 800 °C (1290 to 1470 °F) (Ref 83). 250

∆W, mg/cm2

150 Type 347 SS 100 Alloy 800H 50 0 0

200

(a)

600 400 Exposure, h

800

1000

100 80

Alloy 625

60

Alloy X

40

Alloy 617

20

NiS

0

–5

0 Cr2S3

Test gas

Co9S8

200

(b)

oO

400 600 Exposure, h

800

1000

C

2

Log pS , bar

Type 321 SS

200

∆W, mg/cm2

reaction rate generally peaks at a certain temperature, then decreases with increasing temperature. The highest corrosion rate is normally related to the formation of sulﬁdes. Sulﬁdes provide paths for rapid outward diffusion of metals, such as nickel, iron, and chromium, and so forth, resulting in rapid corrosion attack (Ref 98). The rate is particularly rapid when liquid sulﬁde eutectics are formed. The Ni-Ni3S2 eutectic melts at 635 °C (1175 °F), the Fe-FeS eutectic at 985 °C (1805 °F), and the Co-Co4S3 eutectic at 880 °C (1616 °F) (Ref 17). The temperature at which the corrosion rate is highest varies from metal to metal. For nickel, it is around 600 °C (1110 °F). At temperatures above 800 °C (1470 °F), the rate decreases with increasing temperature (Ref 98). Sulﬁdes become less and less stable as temperature increases. Eventually, NiO becomes the only corrosion product (Ref 98). For cobalt, the corrosion rate is highest at 920 °C (1688 °F), and decreases above that temperature (Ref 98). The corrosion rate for iron is high above 940 °C (1724 °F) (Ref 98), which corresponds approximately to the iron sulﬁde eutectic temperature. Chromium, on the other hand, does not form sulﬁdes in SO2 at temperatures from 700 to 1000 °C (1290 to 1830 °F), but instead forms a Cr2O3 scale (Ref 90). Thus, chromium as an alloying element improves sulﬁdation resistance in Co-Cr (Ref 91) and Ni-Cr alloys (Ref 92). Corrosion tests in environments containing both O2 and SO2 were conducted primarily in SO2-O2 mixtures with various ratios. The

Sulfidation / 227

Ni3S2

–10

FeS

Oo

10

Ni

Type 347 SS

Fe

∆W, mg/cm2

CrS –15

8

FeO Fe3O4

Cr Cr2O3

NiO

Alloy 800H Alloy HR-120

6

Alloy 556

4 2

–20 –35

–30

–25

–20

–15

Log pO , bar

Alloy HR-160

0 0

2

(c)

Fig. 7.42 Ref 78

Test environment in terms of pO2 and pS2 is plotted in the stability diagram at 700 °C (1202 °F) Source:

Fig. 7.43

200

400 600 Exposure, h

800

1000

Resuts of corrosion tests in H2-7CO-1.5H2O0.6H2S at 700 °C (1292 °F). Source: Ref 78

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

p

228 / High-Temperature Corrosion and Materials Applications

p The test environment in terms of pO2 and pS2 potentials at 600 °C in CO-32H2-3.8CO2-0.2H2S is plotted as an equilibrium condition (identified as “E”) and a nonequilibrium (NE) is plotted in the phase-stability diagram. Source: Ref 79

weight gain, mg/cm2

Fig. 7.44

8 7 6 5 4 3 2 1 0

26/37/3 HR3C

MA960 (ave)

0

500

1000

1500

45 TM MA956 (low) 2000 HR160

time, h

Fig. 7.45

Corrosion behavior of MA956 in comparison with HR160, 45TM, HR3C, and an experimental alloy 26Cr/37Ni/3V (26/37/3) in CO-32H2-3.8CO2-0.2H2S at 600 °C (1112 °F). Source: Ref 79

Fig. 7.46

Corrosion scales formed on MA956 (left picture) and HR160 (right picture) after exposure at 600 °C (1112 °F) for 2000 h in CO-32H2-3.8CO2-0.2H2S. For MA956, the oxide scale (in black) was a mixed Cr-Al-rich oxide scale with the outer layer being chromium-rich oxide scale and the inner layer being aluminum-rich oxide scale. For HR160, the oxide scale (in black) was silicon-rich oxide enriched with titanium on the outer layer of the oxide scale. Source: Ref 80

Nickel forms nickel oxides, nickel sulﬁdes, and NiSO4 when the corrosion rate is highest (Ref 98). The sulﬁdes provide paths for outward

diffusion of nickel. At sufﬁciently high temperatures, NiO becomes the only stable corrosion product (Ref 98). Then the corrosion rate is lowered. Ni-20Cr alloy in SO2-O2 mixture behaved similarly to nickel (Ref 98). Vasantasree and Hocking (Ref 93) investigated Ni-Cr alloys with various chromium contents. Their results are shown in Fig. 7.48. Rates were highest at 700 °C (1290 °F) for Ni-Cr alloys. Formation of oxides was favored at high temperatures, particularly 900 and 1000 °C (1650 and 1830 °F). Therefore, corrosion rates became much lower at these temperatures. Very few investigators have tested commercial alloys in SO2-bearing environments. Barnes and Lai (Ref 99) examined the oxidation behavior of two iron-base commercial alloys, Type 304SS and alloy 556, in Ar-5O2-5CO2 with and without 10% SO2 at 980 °C (1800 °F). The equilibrium partial pressures for the Ar-5O2-5CO2-10SO2 at

pg 228

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

p

Chapter 7:

p

Fig. 7.47

Initial linear rate constants of nickel in different SO2-O2 mixtures at 1 atm and 800 °C (1470 °F). Also shown are equilibrium SO3 pressures at different SO2-O2 mixtures. Source: Ref 83

Parabolic rate constant, mg2/cm4/h

103

700°

SO2:O2::2:1 Temp. °C 600 700 800 900 1000 L=Linear

800°

102

600° 101

900° 100 1000°

10–1

10–2 100 Cr

Fig. 7.48

50

20 10 at.%, Cr

1

0.1 0 Ni

Parabolic rate constants of Ni-Cr alloys as a function of chromium concentration in SO2:O2 mixture (2:1 ratio) at various temperatures. Source: Ref 93

pg 229

Sulfidation / 229

the test temperature were 4.9 × 10−2 atm for pO2 , 9.7 × 10−2 atm for pSO2 , 3.6 × 10−3 atm for pSO3 , and 1 × 10−22 atm for pS2 . In the SO2-bearing test environment, Type 304 sample suffered severe oxidation/sulﬁdation attack, with some areas being completed consumed by corrosion attack. On the other hand, alloy 556 formed a thin, compact chromium-rich oxide scale with some internal chromium-rich sulﬁdes underneath the oxide scale. Table 7.10 summarizes the corrosion attack for both alloys in both environments. The results suggest that an alloy that forms a protective chromium oxide scale in a purely oxidizing environment is likely to form a similar scale in SO2-O2 environment. Many industrial processes that generate SO2-bearing environments are generally at much lower temperatures. Yates et al. (Ref 100) tested alloys X (identiﬁed as HX Ni-22Cr-9Mo-18Fe), 617 (Ni-22Cr12Co-9Mo-1.2Al), 230 (Ni-22Cr-14W-La), 188 (Co-22Cr-20Ni-14W-La), and 214 (Ni-16Cr3Fe-4.5Al-Y) at 704 °C (1300 °F) in O2- 4% SO2 for more than 40 days. Their test results are summarized in Fig. 7.49. Based on these test results, Ni-Cr and Co-Cr alloys containing about 22% Cr are considered to have adequate corrosion resistance in O2-4%SO2 at 704 °C (1300 °F). All four 22Cr alloys (X, 617, 188, and 230) exhibited a parabolic reaction kinetics with low mass changes over more than 40 days of exposure, indicating formation of protective chromium-rich oxide scales. On the other hand, Ni-Cr-Al alloy 214 with only about 16% Cr showed some indication of breakaway corrosion. The test temperature of 704 °C (1300 °F) was likely too low for rapid formation of Al2O3 in O2-SO2 mixtures. The amount of chromium (about 16%) is considered to be inadequate in forming protective Cr2O3 scales under the test condition. Field test data, which was generated in a chemical plant with the environment containing about 18% SO2 in the temperature range of 260 to 371 °C (500 to 700 °F), showed minimal corrosion rates (about 0.3 mpy and less) for Type 316, 317, alloy 20, and alloy 825 (Ref 101). The levels of SO2 used for most corrosion studies in either SO2 environments or SO2-O2 mixtures are signiﬁcantly higher than the amounts expected in the combustion of sulfurbearing fuels such as coal or oil. Typical combustion ﬂue gas produced in a coal-ﬁred boiler, for example, contains approximately 0.25% SO2 (Ref 102). Very few investigators have studied SO2 corrosion at this low level (i.e., less than

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

pg 230

230 / High-Temperature Corrosion and Materials Applications

Table 7.10 Corrosion of Type 304SS and alloy 556 at 980 °C (1800 °F) for 550 h in an oxidizing environment with and without SO2 Type 304SS Test environment

Ar-5O2-5CO2 Ar-5O2-5CO2-10SO2

Metal loss, mm (mils)

0.31 (12.2) > 0.61 (24)

Alloy 556

Maximum depth of attack, mm (mils)(a)

0.44 (17.2) > 0.61 (24)

Metal loss, mm (mils)

Maximum depth of attack, mm (mils)(a)

0.005 (0.2) 0.06 (2.5)

0.056 (2.2) 0.10 (4.0)

(a) Metal loss + maximum internal penetration. Source: Ref 99

Table 7.11 Corrosion of Ni-15Cr alloy in SO2-bearing environments at 870 °C (1600 °F) for 30 h Test environment

Weight loss, mg/cm2

Depth of attack, mm (mils)

Pure SO2 N2-1.35SO2 N2-1SO2 N2-0.2SO2 N2-0.05SO2 N2-0.2SO2-0.01O2 N2-0.2SO2-0.1O2 N2-0.2SO2-2O2 N2-0.2SO2-3O2

0.6 1.8 2.0 86.0 13.7 6.5 0.3 0.3 …

0.008 (0.3) … … 0.52 (20.5) … … … … 0.008 (0.3)

Source: Ref 103

Fig. 7.49

Mass change as a function of exposure time for alloys X (identified as HX) 617, 230, 188 and 214 at 704 °C (1300 °F) in O2-4% SO2. Source: Ref 100

1%). However, studies have been conducted of corrosion in environments containing low concentrations of SO2 in conjunction with ash/salt deposits in order to simulate fuel ash corrosion in combustion systems, particularly fossil-ﬁred boilers. These data are discussed in later chapters covering speciﬁcally on coal-ﬁred boilers and oil-ﬁred boilers. Viswanathan and Spengler (Ref 103) observed that Ni-15Cr alloy suffered signiﬁcantly more corrosion attack in 0.2% SO2 (balance N2) than in pure SO2 at 870 °C (1600 °F). The sample exposed to N2-0.2SO2 at 870 °C (1600 °F) for 30 h showed about 0.5 mm (20 mils) of attack with internal globular sulﬁde phases, while the sample exposed to 100% SO2 for same temperature and time exhibited only about 0.008 mm (0.3 mil) of attack. Addition of oxygen signiﬁcantly reduced the alloy’s corrosion rates. The weight loss for the sample after 30 h fell from 86 mg/cm2 in N20.2SO2 to about 0.3 mg/cm2 in N2-0.2SO20.1O2. The results of their tests (Ref 103) are summarized in Table 7.11. Low concentrations

of SO2 in the environment with no “free” oxygen can cause Ni-Cr alloys to suffer severe sulﬁdation attack even without ash/salt deposits.

7.6 Summary The sulﬁdation behavior of a wide variety of alloys was reviewed. The data presented are primarily related to gaseous environments; sulfate-accelerated sulﬁdation (hot corrosion) is covered in Chapter 9. The corrosion data are grouped into three different types of environments: (1) sulfur vapor, hydrocarbon streams containing no hydrogen gas, H2S, and H2-H2S, (2) reduced, mixed gas environments with low oxygen and high sulfur potentials, and (3) SO2bearing environments. Sulfur vapor, hydrocarbon streams (no H2), H2S, and H2-H2S environments have a common feature in that the corrosion products formed in these environments are essentially sulﬁdes. The reduced, mixed gas environments with low oxygen and high sulfur potentials typically contain H2, CO, CO2, H2O, H2S, and other gaseous components where oxides and sulﬁdes can form on most high temperature alloys. In these environments, alloys are protected by a protective

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

Chapter 7:

oxide scale before the initiation of breakaway corrosion due to formation of fast-growing, nonprotective sulﬁdes. The SO2-bearing environments are generated by combustion of a sulfurcontaining fuel or feedstock with excess air or oxygen. These environments are oxidizing, and are generally less corrosive than reducing environments.

REFERENCES

1. G.Y. Lai, J. Met., July 1985, p 14 2. G.L. Swales, in Behavior of High Temperature Alloys in Aggressive Environments, I. Kirman et al., Ed., Proc. Petten International Conference, Oct 15–18, 1979, The Metals Society, London, 1980, p 45 3. G. Sorell, M.J. Humphries, E. Bullock, and M. Van de Voorde, Int. Met. Rev., Vol 31 (No. 5), 1986, p 216 4. J.F. Norton, Ed., High Temperature Materials Corrosion in Coal Gasiﬁcation Atmospheres, Elsevier, Amsterdam, 1984 5. K.J. Barton, V.L. Hill, and R. Yurkewycz, in The Properties and Performance of Materials in the Coal Gasiﬁcation Environments, V.L. Hill and H.L. Black, Ed., American Society For Metals, 1981, p 65 6. S.K. Srivastave, G.Y. Lai, and D.E. Fluck, Paper No. 398, Corrosion/87, NACE, 1987 7. J.A. Harris, W.G. Lipscomb, and G.D. Smith, Paper No. 402, Corrosion/87, NACE, 1987 8. J. Stringer, in High Temperature Corrosion, R.A. Rapp, Ed., Conference Proceedings (San Diego, CA) March 2–6, 1981, NACE, 1981, p 389 9. A.J. Minchener, D.M. Lloyd, and P.T. Sutcliffe, “Materials Evaluation for Fluidized Bed Combustion Systems,” CS3511, Final Report to EPRI on Research Project RP979-11, Electric Power Research Institute, Palo Alto, CA, 1984 10. J. Stringer, Paper No. 90, Corrosion/86, NACE, 1986 11. S.R. Shatynski, Oxid. Met., Vol 11 (No. 6), 1977, p 307 12. HSC, Chemistry for Windows, Version 6.0, A. Roine, Outokumpu Technology, Finland, www.outokumputechnology.com, accessed Dec 2006 13. ChemSage, Version 4.16, GTT-Technologies, Aachen (1998)

pg 231

Sulfidation / 231

14. R.A. Perkins, in Environmental Degradation of High Temperature Materials, Series 3, Vol 2 (No. 13), 1980, p 5/1 15. P.L. Hemmings and R.A. Perkins, “Thermodynamic Phase Stability Diagrams for the Analysis of Corrosion Reactions in Coal Gasiﬁcation/Combustion Atmospheres,” EPRI Report FP-539, Lockheed Palo Alto Research Laboratories, Palo Alto, CA, 1977 16. B.A. Gordon and V. Nagarajan, Oxid. Met., Vol 13 (No. 2), 1979, p 197 17. M. Hansen and K. Anderko, Constitution of Binary Alloys, McGraw-Hill, 1958 18. S. Mrowec and K. Przybylski, High Temp. Mater. Proc., Vol 6 (No. 1 and 2), 1984, p 1 19. D.J. Young, Rev. High Temp. Mater., Vol 4 (No. 4), 1980, p 299 20. A. Davin and D. Coutsouradis, Cobalt, Vol 17, 1962, p 23 21. S. Mrowec, T. Walec, and T. Werber, Oxid. Met., Vol 1, 1969, p 93 22. T. Narita, W.W. Smeltzer, and K. Nihida, Oxid. Met., Vol 17, 1982, p 299 23. T. Narita and K. Nihida, Oxid. Met., Vol 6, 1973, p 157 and 181 24. S.K. Mrowec, T. Werber, and M. Zastawnik, Corros. Sci., Vol 6, 1966, p 47 25. D.P. Whittle, S.K. Verma, and J. Stringer, Corros. Sci., Vol 13, 1973, p 247 26. T. Biegun, A. Bruckman, and S. Mrowec, Oxid. Met., Vol 12, 1978, p 157 27. S. Mrowec and M. Wedrychowska, Oxid. Met., Vol 13, 1979, p 481 28. E.M. Jallouli, J.P. Larpin, M. Lambertin, and J.C. Colson, J. Electrochem. Soc., Vol 126, 1979, p 2254 29. T. Biegun and A. Bruckman, Bull. Acad. Polon. Ser. Sci. Chim., Vol 28, 1980, p 377; Vol 29, 1981, p 69 30. W.W. Smeltzer, T. Narita, and K. Przybylski, in Proc. Corrosion-Erosion, Wear of Materials in Emerging Fossil Energy Systems, A.V. Levy, Ed., NACE, 1982, p 860 31. T. Biegun and A. Bruckman, “Sulﬁdation of Ni-Cr-Al Alloys,” Report No. 2.34264, Institute of Physical Chemistry, Polish Academy of Science, Warsaw, 1980 32. L.A. Morris, Chapter 17: Resistance to Corrosion in Gaseous Atmospheres, in Handbook of Stainless Steels, D. Peckner and I.M. Bernstein, McGraw-Hill, 1977, p 17.1

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

232 / High-Temperature Corrosion and Materials Applications

33. E. Malinowski, R. Bigot, and E. Herzog, Le Metallurgic, Vol 94 (No. 4), 1962 34. F.J. Bruns, Corrosion of Ni-Cr-Al-Fe Alloys by Hydrogen Sulﬁde at 1100 to 1800 °F, Corrosion, Vol 25 (No. 3), 1969, p 119 35. Z.A. Foroulis, High Temperature Degradation of Structural Materials in Environments Encountered in the Petroleum and Petrochemical Industries: Some Mechanistic Observations, Anti-Corrosion, Vol 32 (No. 11), 1985, p 4–9 36. J. Gutzeit, R.D. Merrick, and L.R. Scharfstein, Corrosion in Petroleum Reﬁning and Petrochemical Operations, in Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 1262 37. F.A. Hendershot and H.L. Valentine, Materials for Catalytic Cracking Equipment (Survey), Mater. Prot., Vol 6 (No. 10), 1967, p 43 38. J. Gutzeit, High Temperature Sulﬁdic Corrosion of Steels, in Process Industries Corrosion—The Theory and Practice, NACE, 1986 39. The Role of Stainless Steels in Petroleum Reﬁning, Originally published by the Committee of Stainless Steel Producers, AISI (1977), Nickel Development Institute, Toronto, Ontario, Canada, April 1996 40. Steels for Hydrogen Service at Elevated Temperatures and Pressures in Petroleum Reﬁneries and Petrochemical Plants, Publication 941, 3rd ed., American Petroleum Institute, 1983 41. R.A. White and E.F. Ehmke, Materials for Reﬁneries and Associated Facilities, NACE, 1991, p 51 42. B.W. Neumaier and C.M. Schillmoller, “How Richﬁeld Plans to Combat HighTemperature Sulﬁde Corrosion in Its New Catalytic Reformer,” presented at the 21st Midyear Meeting of the American Petroleum Institute’s Division of Reﬁning (Montreal, Canada), May 14, 1956 43. G. Sorell, “Compilation and Correlation of High Temperature Catalytic Reformer Corrosion Data,” Technical Committee Report, Publication 58-2, NACE, Houston, TX, 1957 44. E.B. Backensto, R.E. Drew, J.E. Prior, and J.W. Sjoberg, “High-Temperature Hydrogen Sulﬁde Corrosion of Stainless Steels,” Technical Committee Report, Publication 58-3, NACE, Houston, TX, 1957

45. G. Sorell and W.B. Hoyt, “Collection and Correlation of High Temperature Hydrogen Sulﬁde Corrosion Data,” NACE Technical Committee Report, Publication 56-7, NACE, Houston, TX, 1956 46. E.B. Backensto, “Corrosion in Catalytic Reforming and Associated Processes, Summary Report of the Panel on Reformer Corrosion to the Subcommittee on Corrosion,” presented at the 22nd Midyear Meeting of API’s Division of Reﬁning (Philadelphia, PA), May 13, 1957 47. E. Dittrich, Chem. Fab., Vol 10 (No. 13/ 14), 1947, p 145 48. E.B. Backensto and J.W. Sjoberg, “Iso-Corrosion Rate Curves for High Temperature Hydrogen-Hydrogen Sulﬁde,” Technical Committee Report, Publication 59-10, NACE, Houston, TX, 1958 49. A.S. Couper and J.W. Gorman, Computer Correlations to Estimate High Temperature H2S Corrosion in Reﬁnery Streams, Mater. Prot. Perform., Vol 10 (No. 4), 1971, p 31 50. V.L. Hill and H.S. Meyer, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 29 51. A.O. Schaefer, C.H. Samans, M.A. Howes, S. Bhattacharyya, E.R. Bangs, V.L. Hill, and F.C. Chang, “A Program to Discover Materials Suitable for Service under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1975 Annual Report, The Metal Properties Council, New York, 1976 52. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1976 Annual Report, The Metal Properties Council, New York, 1977 53. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1977 Annual Report, The Metal Properties Council, New York, 1978 54. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1978 Annual Report, The Metal Properties Council, New York, 1979

pg 232

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

Chapter 7:

55. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” Supplement to 1978 Annual Report, The Metal Properties Council, New York, 1979 56. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1979 Annual Report, The Metal Properties Council, New York, 1980 57. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1980 Annual Report, The Metal Properties Council, New York, 1981 58. A.O. Schaefer, “A Program to Discover Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1981 Annual Report, The Metal Properties Council, New York, 1982 59. A. Humphreys and A.O. Schaefer, “A Program to Discover Materials Suitable for Service under Hostile Conditions Obtaining in Equipment for the Gasiﬁcation of Coal and Other Solid Fuels,” 1982 Annual Report, The Metal Properties Council, New York, 1983 60. M.A.H. Howes, “High Temperature Corrosion in Coal Gasiﬁcation Systems,” Final Report GRI-8710152, Gas Research Institute, Chicago, Aug 1987 61. R.A. Perkins and S.J. Vonk, EPRI Report FP-1280, Electric Power Research Institute, Palo Alto, CA, Dec 1979 62. K. Natesan, in High Temperature Corrosion, R.A. Rapp, Ed., NACE, 1983, p 336 63. S.K. Verma, Corrosion of Commercial Alloys in A Laboratory-Simulated Medium-BTU Coal Gasiﬁcation Environment, Paper No. 336, Corrosion/85, NACE, 1985 64. R.W. Bradshaw, R.E. Stoltz, and D.R. Adolphson, Report SAND 77-8277, Sandia Laboratories, Livermore, CA 1977 65. R.W. Bradshaw, A.N. Nagelberg, R.E. Stoltz, and D.R. Adolphson, Report SAND 78-8260, Sandia Laboratories, Livermore, CA, 1978 66. J.A. Kneeshaw, I.A. Menzies, and J.F. Norton, Werkst. Korros., Vol 38, 1987, p 473

pg 233

Sulfidation / 233

67. J.L. Blough, V.L. Hill, and B.A. Humphreys, in The Properties and Performance of Materials in the Coal Gasiﬁcation Environment, V.L. Hill and H.L. Black, Ed., American Society For Metals, 1981, p 225 68. R.C. John, W.C. Fort III, and R.A. Tait, Prediction of Alloy Corrosion in the Shell Coal Gasiﬁcation Process, Mater. High Temp., Vol 11 (No. 1–4), 1993, p 124 69. W.T. Bakker, Effect of Gasiﬁer Environment on Materials Performance, Mater. High Temp., Vol 11 (No. 1–4), 1993, p 81 70. W.T. Bakker and J.A. Bonvallet, Corrosion of Stainless Steels on the Wrong Side of the Kinetic Boundary, in Heat-Resistant Materials II, Conf. Proc. Second International Conference on Heat-Resistant Materials, K. Natesan, P. Ganesan, G. Lai, Ed., ASM International, 1995, p 121 71. G.Y. Lai, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 227 72. K. Natesan, Surface Modiﬁcation for Corrosion Resistance, Mater. High Temp., Vol 11 (No. 1–4), 1993, p 36 73. Y.-R. He, D.L. Douglass, and F. Gesmundo, The Corrosion Behavior of Ni-Mo Alloys in a H2/H2O/H2S Gas Mixture, Oxid. Met., Vol 37 (No. 5/6), 1992, p 413 74. V. Nagarajan, R.G. Miner, and A.V. Levy, J. Electrochem. Soc., Vol 129 (No. 4), 1982, p 782 75. G.Y. Lai, Sulﬁdation-Resistant Co-Cr-Ni Alloy With Critical Contents of Silicon and Cobalt, U.S. Patent No. 4711763, Dec 1987 76. G.Y. Lai, Paper No. 209, Corrosion/89, NACE, 1989 77. G.Y. Lai, J. Met., Vol 41 (No. 7), 1989, p 21 78. J.F. Norton, F.G. Hodge, and G.Y. Lai, A Study of the Corrosion Behavior of Some Fe-Cr-Ni and Advanced Ni-Based Alloys Exposed to a Sulphidising/Oxidising/Carburising Atmosphere at 700 °C, in High Temperature Materials for Power Engineering (Part I), Conf. Proc., E. Bachelet et al., Ed., Kluwer Academic Publishers, Dordrecht, The Netherlands, 1990, p 167 79. J.F. Norton, T.P. Levi, and W.T. Bakker, High Temperature Corrosion of Candidate

Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/

26/10/2007 11:15AM Plate # 0

234 / High-Temperature Corrosion and Materials Applications

80.

81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91.

Heat Exchanger Alloys in a Dry-Feed Entrained Slagging Gasiﬁer Atmosphere, in High Temperature Materials for Power Engineering (Part II), Conf. Proc., E. Bachelet et al., Ed., Kluwer Academic Publishers, Dordrecht, The Netherlands, 1994, p 1617 J.F. Norton and T.P. Levi, A Laboratory Study of the Corrosion Behaviour of Alloys Exposed in a Non-Equilibrated Coal-Gasiﬁcation Atmosphere at 600 °C, Mater. Corros., Vol 46, 1995, p 286 M. Seiersten and P. Kofstad, Corros. Sci., Vol 22 (No. 5), 1982, p 487 M.R. Wootton and N. Birks, Corros. Sci., Vol 12, p 829 B. Haﬂan and P. Kofstad, Corros. Sci., Vol 23 (No. 12), 1983, p 1333 A. Andersen, B. Haﬂan, P. Kofstad, and P.K. Lillerud, Mater. Sci. Eng., Vol 87, 1987, p 45 J. Gilewicz-Wolter, Oxid. Met., Vol 11, 1977, p 81 A. Rahmel, Oxid. Met., Vol 9, 1975, p 491 A. Rahmel, Corros. Sci., Vol 13, 1975, p 125 P. Singh and N. Birks, Oxid. Met., Vol 12, 1978, p 1 P. Singh and N. Birks, Oxid. Met., Vol 12, 1978, p 22 C.D. Asmundis, F. Gesmundo, and C. Bottino, Oxid. Met., Vol 14 (No. 4), 1980, p 351 P. Singh and N. Birks, Oxid. Met., Vol 13 (No. 5), 1979, p 457

92. H. Lewis and J.E. Whittle, in Proc. Fourth Int. Cong. Met. Corros., NACE, TX, 1972 93. V. Vasantasree and M.G. Hocking, Corros. Sci., Vol 16, 1976, p 261 94. M.G. Hocking and V. Vasantasree, Corros. Sci., Vol 16, 1976, p 279 95. K.N. Strafford, P.K. Datta, A.F. Hampton, and P. Mistry, Corros. Sci., Vol 29 (No. 6), 1989, p 673 96. P.S. Sidky and M.G. Hocking, Corros. Sci., Vol 27 (No. 2), 1987, p 183 97. M.G. Hocking and P.S. Sidky, Corros. Sci., Vol 27 (No. 2), 1987, p 205 98. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science, 1988 99. J.J. Barnes and G.Y. Lai, Paper No. 90276, Corrosion/90, NACE, 1990 100. D.H. Yates, P. Ganesan, and G.D. Smith, Recent Advances in the Enhancement of Inconel Alloy 617 Properties to Meet the Needs of the Land Based Gas Turbine Industry, in Advanced Materials and Coatings for Combustion Turbines Conference Proceedings, V.P. Swaminathan and N.S. Cheruvu, ASM International, 1994, p 89 101. A Guide to Corrosion Resistance, Climax Molybdenum Company, Greenwich, CT, 1981 102. R.W. Borio, A.L. Plumley, and W.R. Sylvester, in Ash Deposits and Corrosion Due to Impurities in Combustion Gases, R.W. Bryers, Ed., Hemisphere Publishing, 1978, p 163 103. R. Viswanathan and C.J. Spengler, Corrosion, Vol 26 (No. 1), 1970, p 29

pg 234

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p235-248 DOI: 10.1361/hcma2007p235

26/10/2007 12:33PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 8

Erosion and Erosion-Corrosion 8.1 Introduction Industrial plants are often involved in processes where gas streams are laden with particles. Metallic components can suffer severe metal wastage when subjected to constant impingement by this particle-laden gas stream under high velocity and high particle loading. The continual removal of the material from a component by this process is referred to as “erosion.” When the component is exposed to elevated temperatures where oxidation and/or other modes of high-temperature corrosion are involved, it is frequently referred to as “erosion-corrosion.” Materials problems that are caused by erosion or erosion-corrosion in industries are numerous. In the petrochemical industry, the pyrolysis furnace tubes for the production of ethylene are a good example. Ethylene is produced by cracking petroleum feedstocks, such as ethane and naphtha, at temperatures up to 1150 °C (2100 °F), thus making the process gas stream inside the tube highly carburizing in nature. The furnace tubes suffer both carburization and coking on the internal surface of the tube. In order to maintain the process efﬁciency, the coke deposits have to be regularly removed from the tube inner diameter (ID) surface by a process referred to as “decoking,” which involves injecting a mixture of steam and air into the furnace tube. Thus, during the decoking operation, the return bends of the pyrolysis furnace tubes can suffer erosion or erosion-corrosion due to the coke particle-laden gas stream (Ref 1, 2), with average tube skin temperatures varying from 800 to 1120 °C (1475 to 2050 °F) and gas velocities greater than 200 m/s (656 ft/s). In coal-ﬁred boilers, it is well known that ﬂyash erosion can be a serious problem for boiler tubes. When coal (which contains ash) is combusted in the lower furnace of the boiler, some of the ash drops out of the furnace from the bottom with remaining ash being carried by the combustion ﬂue gas stream to the top of the furnace

and through the convection pass. The ash carried by the ﬂue gas stream is referred to as “ﬂy-ash.” With pulverized coal ﬁring, about 70 to 90% of the ash in the coal is carried by the ﬂue gas stream, while only about 40% of the ash is carried by the ﬂue gas in a stoker-ﬁred furnace (Ref 3). The cyclone-ﬁred boiler generates about 15 to 30% of the ash from the coal in the ﬂue gas. Major constitutes of ﬂy-ash are SiO2, Al2O3, and Fe2O3. (More information about ash constituents in coal is available in Chapter 10 “Coal-Fired Boilers.”) This ﬂy-ash-laden ﬂue gas stream can pose ﬂy-ash erosion problems to convection pass tubes, such as superheater, reheater, boiler bank, and economizer tubes. To reduce ﬂy-ash erosion problems for the convection pass tubes, boiler designers typically have set maximum ﬂue gas velocities in these areas. For example, Babcock & Wilcox typically limits ﬂue gas velocity to 19.8 m/s (65 ft/s) or less for relatively nonabrasive low ash coal, and 13.7 m/s (45 ft/s) or less for coals with high ash quantities and/or abrasive ash (Ref 4). Combustion Engineering (now Alstom Power) typically set the design velocity in the range of 12 to 18 m/s (40 to 60 ft/ s) (Ref 5). Lower velocities would be used for a boiler burning coals that yield heavy loading of erosive ash, which is usually indicated by high silica content (Ref 5). For ﬂuidized-bed coalﬁred boilers, ﬂue gas streams are laden with not only ﬂy-ash but also sands from the bed. The convection pass tubes in these boilers are subject to erosion attack. Waste-to-energy (WTE) boilers burning municipal and industrial waste can also experience ﬂyash erosion problems for their convection pass tubes (e.g., superheater and economizer). Because of much more corrosive environments in WTE boilers than coal-ﬁred boilers, the boiler designers have used lower maximum design velocities in WTE boilers. Combustion Engineering, for example, generally limits the ﬂue gas velocities entering into the superheater or economizer to 6 to 7.5 m/s (20 to 25 ft/s) (Ref 6).

pg 235

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

236 / High-Temperature Corrosion and Materials Applications

Babcock & Wilcox sets the maximum design velocity for superheaters and generating bank at 9.1 m/s (30 ft/s) (Ref 7). In practice, lower velocities (3 to 4.6 m/s, or 10 to 15 ft/s) are used (Ref 7). Local ﬂow disturbances can sometimes create conditions that are more conducive to erosion attack. This is illustrated in a case of the ﬂy-ash erosion in the backpass area of the economizer when the ﬂue gas stream makes a 90° directional turn and creates nonuniform ﬂow of gas stream and uneven distribution of ﬂy-ash caused by centrifugal force (Ref 8). Other systems where erosion can present a serious material issue include coal gasiﬁcation, combined cycles, and gas turbine. Gas turbines may involve extremely high gas velocities, up to 250 m/s (820 ft/s), but with small particles (typically 5 µm) (Ref 9). The erosion problems that have been encountered in the aforementioned systems all occur at elevated temperatures where the metal surface not only experiences mechanical damage caused by erosion attack but also experiences oxidation or other high-temperature corrosion reactions at the same time. Although it is often referred to as “erosion” in industry, the damage reactions, in many cases, may involve a combination of erosion and corrosion, and are more appropriately referred to as “erosion-corrosion.” Since industry continues to use a trial-and-error approach to solve this practical erosion-related materials issue, the question is whether the

Fig. 8.1

research data generated can provide industry with some practical guidance for materials selection based on either the mechanical properties of the alloy (to select a more erosionresistant alloy) or based on the resistance of the alloy to oxidation or high-temperature corrosion (to select a more erosion-corrosion resistant alloy). The objective of this chapter is to answer this question by reviewing the relevant erosion and erosion-corrosion data that are mainly related to a particle-laden gas stream “jet” impacting on the metal surface. The erosion or erosioncorrosion that is related to (a) in-bed components of bubbling ﬂuidized-bed boilers and (b) waterwalls of circulating ﬂuidized-bed boilers is not covered in this chapter. Discussion of erosion in ﬂuidized-bed coal-ﬁred boilers is presented in Chapter 10 “Coal-Fired Boilers.”

8.2 Erosion and Erosion-Corrosion Finne (Ref 10) described the response of a material to erosion at room temperature in two distinctively different modes in terms of the particle incident angles (or particle impingement angle). For a ductile material, such as aluminum, the maximum erosion attack occurs at low incident angles (less than 30°) with respect to the particle impingement direction and the minimum attack at about 90°, as shown in Fig. 8.1 (Ref 10). On the other hand, the maximum erosion attack on a brittle material, such as glass, occurs at close

Effect of the particle incidence angle on the room-temperature erosion wear for a ductile material (aluminum) and a brittle material (glass). Source: Ref 10

pg 236

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

Chapter 8: Erosion and Erosion-Corrosion / 237

to 90° (Ref 10). In both cases, no erosion occurs when the incident angle is 0° indicating the surface of the object is in parallel with the particle impingement direction. The erosion model described by Finne (Ref 10) is related to the room-temperature response of materials. In industrial environments, erosion problems are most often related to elevated service temperature. Levy (Ref 11) examined the erosion response of Type 310 as a function of temperature at incident angles of 30° and 90°. The test environment consisted of nitrogen gas (N2), thus eliminating the effect of oxidation. The data indicated that there was signiﬁcantly less erosion attack at a 90° impingement angle than 30° particularly at high temperatures, as shown in Fig. 8.2 (Ref 11). The erosion behavior of Type 310 in N2 with no oxidation in the erosion reaction is quite similar to the room-temperature erosion model of a ductile material proposed by Finne. Figure 8.2 also shows that erosion attack increased with advancing temperature at 30° impingement angle. Similar temperature behavior was also observed for Type 304 tested in N2 environment, as shown in Fig. 8.3 (Ref 11).

Carbon steel was also observed to behave similarly in air, showing increasing erosion (or erosion-corrosion) with advancing temperature, as shown in Fig. 8.4 (Ref 12). Shida and Fujikawa (Ref 13) examined the effect of the impingement angle on the erosion rate for carbon steel, 1.25Cr-1Mo-0.3V steel, and Type 304 at 300 °C (570 °F) in an argon environment (i.e., elimination of oxidation participation). Their test results are shown in Fig. 8.5 (Ref 13). The maximum erosion attack was found to occur at about a 30° impingement angle for all three alloys with very little erosion attack at a 90° impingement angle. The test was carried out in an inert environment—argon. Type 304 suffered much more erosion attack than both carbon steel and Cr-Mo-V steel in an inert environment. This suggests that austenitic stainless steels may not

Fig. 8.2

Fig. 8.3

Erosion rate as a function of temperature in N2 for Type 310 steel for 30° and 90° impingement angles (α). Source: Ref 11

N2

Ref 11

Erosion rate as a function of temperature in N2 for Type 304 steel for 30° impingement angle. Source:

pg 237

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

pg 238

238 / High-Temperature Corrosion and Materials Applications

be as resistant to erosion attack as ferritic steels when oxidation is not involved. This is further substantiated by the data that included carbon steel, 2.25Cr-1Mo steel, and 12Cr steel compared with Type 304 and alloy 800 (Fig. 8.6). Alloy 800 suffered the worst erosion attack at all temperatures from room temperature to 650 °C (1200 °F) when oxidation was not involved. Nagarajan and Wright (Ref 14) investigated iron-, nickel-, and cobalt-base alloys and the effect of an oxidizing environment on the erosion resistance of the alloys. Their test environments included argon (an inert atmosphere) and a

Test conditions: 1018 steel, erodent: 180 µm angular Al2O3 v = 10 m/s, α = 30°, t = 8 h, loading: 600 g, air

11

Thickness loss, µm

10

8

6

4

simulated ﬂuidized-bed combustion atmosphere (N2-3%O2-15%CO2-0.026%SO2), which was referred to as “FBC” environment in the paper. For all three alloy systems, erosion rates were found to be signiﬁcantly higher in the oxidizing environment than in the inert environment. Figure 8.7 shows the behavior of Stellite No. 1 (HS 1) and Stellite 6B (HS 6B) at 760 °C (1400 °F) at both 30° and 90° impingement angles. Both alloys No. 1 and No. 6B are wearresistant alloys. With much more carbon and tungsten, alloy No. 1 is generally more resistant to wear than alloy 6B. The data, however, show that both alloys exhibited similar resistance to erosion attack under the test condition in an inert environment (Ar) at both 30° and 90° impingement angles. When tested in an oxidizing environment (FBC), both alloys also exhibited similar erosion rates, although alloy 6B showed only slightly higher erosion rates at the 90° impingement angle. However, both alloys showed higher erosion rates in an oxidizing environment (FBC) than in an inert environment (argon) for both impingement angles. As for the effect of the impingement angle, a 90° impingement angle produced slightly higher erosion rates than did 30° in the oxidizing environment (FBC) while

2

25

100

200

300

400

500

600

Test temperature, °C

Fig. 8.4

200 304

150 Max thickness loss, µm/h

Max thickness loss, µm/h

Effect of temperature on erosion (or erosioncorrosion) of carbon steel in air at 30° impingement angle under the particle velocity of 10 m/s (32.8 ft/s) with 180 μm alumina particles. Source: Ref 12

Alloy 800

100 C-steel 304 2.25Cr-1Mo

50

C-steel

12Cr-1Mo-V

100

1.25Cr-1Mo-0.3V 0

0

Fig. 8.5

10

20

30 40 50 60 70 80 Angle of impingement, degree

0 RT

90

Effect of impingement angle on the erosion of Type 304, carbon steel and Cr-Mo-V steel at 300 °C (570 °F) in argon with 120 m/s (394 ft/s) particle velocity, 3 120 g/m particle concentration, and silica particles of 120 µm average particle size. Source: Ref 13

300

500

650

Temperature, °C

Fig. 8.6

Effect of temperature on the erosion rates of various alloys in argon at the impingement angle of 20° with 120 m/s (394 ft/s) particle velocity, 120 g/m3 particle concentration, and silica particles of 120 µm average particle size. Source: Ref 13

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

Chapter 8: Erosion and Erosion-Corrosion / 239

both 30° and 90° angles showed similar erosion rates in an inert environment (argon). In the same study by Nagarajan and Wright (Ref 14), ferritic FeCrAlY (Fe-25Cr-4Al-1Y) and high-strength nickel-base superalloy IN-100 (Ni-15Co-10Cr-5.5Al-4.7Ti-3Mo) were tested. For the ferritic FeCrAlY alloy, no signiﬁcant difference in erosion rates was observed between argon and the FBC oxidizing environment. The alloy also exhibited similar erosion rates at both 30° and 90° impingement angles in both

HS 1, argon HS 6B, argon

HS 1, FBC

(a)

HS 6B, argon HS 1, argon

(b)

Fig. 8.7

Effect of the test environments (argon and fluidizedbed combustion) on the erosion behavior of HS 1 (Co-30Cr-12W-2.5C) and HS 6B (Co-30Cr-4W-1C) at 760 °C (1400 °F), 42.7 m/s (140 ft/s), with 12 µm alumina (Al2O3) as erodent (a) 30° impingement angle and (b) 90° impingement angle. The argon test environment contained 1% H2, where H2 was used to remove O2, while fluidized-bed combustion (FBC) environment was a simulated test gas consisting of N2, 3% O2, 15% CO2, and 0.026% SO2. Source: Ref 14

environments (Fig. 8.8). For the nickel-base superalloy IN-100, signiﬁcantly higher erosion rates were observed in an oxidizing environment (FBC) than in an inert environment (argon) at both impingement angles. The most interesting ﬁnding from the study by Nagarajan and Wright (Ref 14) (which included high particle velocity of 42.7 m/s, or, 140 ft/s test conditions) was that the low-strength ferritic FeCrAlY alloy was signiﬁcantly more resistant to erosion than the high-strength IN-100 superalloy in an oxidizing environment (Fig. 8.8). FeCrAlY alloy was also found to be much more erosion resistant than wear-resistant cobalt-base alloys No. 1 and 6B in an oxidizing environment. There was no direct comparison of tensile strength data between FeCrAlY and IN-100. However, FeCrAlY (Fe-25Cr-4Al-1Y) is believed to exhibit tensile strengths similar to Type 446 (Fe-25Cr). For example, the ultimate tensile strength of Kanthal D (Fe-22Cr-4.8Al) at 900 °C (1650 °F) is 34.5 MPa (5 ksi) (Ref 15), while that of Type 446 is 29 MPa (4.2 ksi) (Ref 16). At 700 °C (1300 °F), Type 446 exhibits ultimate tensile strength of about 9.0 MPa (1.3 ksi) (Ref 16) as opposed to 1270 MPa (184 ksi) for IN-100 (Ref 17). Thus, IN-100 would be expected to exhibit signiﬁcantly higher tensile strength than FeCrAlY at 700 °C (1300 °F) and would likely be the same at the test temperature of 760 °C (1400 °F) in the erosion study conducted by Nagarajan and Wright. It is thus surprising to see that IN-100 alloy with its tensile strength approximately more than 30 times higher than that of FeCrAlY alloy suffered signiﬁcantly higher erosion rates than FeCrAlY under a particle velocity of about 42.7 m/s (140 ft/s) in an oxidizing environment. It is also surprising to see that the ferritic FeCrAlY alloy was much more resistant to erosion than Stellite alloys No. 1 and No. 6B, which are considered to be excellent wear-resistant alloys. Both alloys No. 1 and No. 6B also exhibit much higher tensile strengths compared with FeCrAlY. For example, at 675 °C (1250 °F), the ultimate tensile strength of Stellite 6B is about 793 MPa (115 ksi) (Ref 18). Stellite No. 1 and No. 6B also show much higher hardness than ferritic Type 446, as shown in Fig. 8.9 (Ref 19). Pettit and Birks along with their research group have investigated erosion behavior of nickel and cobalt (Ref 20, 21), and Cr2O3- and Al2O3-forming alloys (Ref 22, 23) at elevated temperatures under very high particle velocities with 20 µm alumina particles. Kang et al.

pg 239

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

240 / High-Temperature Corrosion and Materials Applications

(Ref 20) indicated that air caused much higher erosion rates than N2 for nickel under their test conditions involving particle velocities of 90 and 1 FeCrAlY, FBC

0

FeCrAlY (2541), argon

Specimen weight change, mg/cm2

–1 –2

IN-100, argon

–3 –4 –5 –6 –7 IN-100, FBC

–8 –9 –10 0

5

10

15

20

25

30

35

40

Erodent impacted, g/cm2 (a) 2

140 m/s (295 and 459 ft/s). They observed that at 800 °C (1470 °F) there was essentially no erosion in N2 at both 90 and 140 m/s (295 and 459 ft/s) (Ref 20). Erosion rates for nickel were signiﬁcantly increased when the test environment was switched from N2 (an inert environment) to air (oxidizing environment). This is illustrated in Fig. 8.10 (Ref 20). Similar results were also observed for cobalt (Ref 20). The effect of the impingement angle on erosion under the same test condition for both nickel and cobalt was reported Chang et al. (Ref 21). The data for nickel are summarized in Fig. 8.11 and for cobalt in Fig. 8.12 (Ref 21). Erosion tests under the same conditions also included Cr2O3- and Al2O3-forming alloys (Ref 22, 23). Figure 8.13 shows the effect of the impingement angle on erosion rates of MA754 (Ni-20Cr-0.6Y2O3, an oxide-dispersion-strengthened alloy) at 780 °C (1435 °F) under 140 m/s (459 ft/s) particle velocity (Ref 22). Erosion test results for nickel, cobalt, MA754, Ni30Cr, CoCrAlY (Co-22Cr11Al-0.17Y), and Ni20Al (Ni-20Al) are summarized in Fig. 8.14 (Ref 22). Both nickel and cobalt were found to suffer high erosion rates due to formation of nickel and cobalt oxides, respectively. Both MA754 and Ni30Cr alloys that formed Cr2O3 scales showed better resistance to erosion attack, and Al2O3-forming

0 Specimen weight change, mg/cm2

IN-100, argon –2 –4

FeCrAlY, argon FeCrAlY, FBC

–6 –8 –10 –12 –14

IN-100, FBC

–16 0

10

20 30 40 50 60 Erodent impacted, g/cm2

70

80

(b)

Fig. 8.8

Effect of the test environments (argon and fluidizedbed combustion) on the erosion behavior of FeCrAlY (Fe-25Cr-4Al-1Y) and IN-100 (Ni-15Co-10Cr-5.5Al-4.7Ti-3Mo) at 760 °C (1400 °F), 42.7 m/s (140 ft/s), with 12 µm alumina (Al2O3) as erodent (a) 30° impingement angle and (b) 90° impingement angle. The argon test environment contained 1% H2, where H2 was used to remove residual O2, while fluidized-bed combustion (FBC) environment was a simulated test gas consisting of N2, 3% O2, 15% CO2, and 0.026% SO2. Source: Ref 14

Temperature, °C (°F)

Fig. 8.9

Hot hardness data for various alloys as a function of temperature. Source: Ref 19

pg 240

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

pg 241

Chapter 8: Erosion and Erosion-Corrosion / 241

alloys (CoCrAlY and Ni20Al) were much more resistant to erosion attack presumably due to formation of aluminum oxide scales. Ives

(Ref 24) conducted erosion tests in a combustion gas stream using a propane-ﬁred abrasive jet rig at 975 °C (1790 °F) with 55 m/s (123 mph) particle velocity of about 135 µm (5 mil) SiC

2.5 780 °C, 140 m/s

Erosion rate, ×10–5 g/cm2 s

2

1.5 600 °C, 140 m/s

1

0.5 780 °C, 70 m/s 0 0

30 60 Impact angle, degrees

90

Fig. 8.10

Effect of the test environments (N2 and air) on the erosion behavior of nickel with particle velocities of 90 and 140 m/s (295 and 459 ft/s) at 90° impingement angle. Source: Ref 20

Fig. 8.11

Effect of the impingement angle on erosion rate of nickel in air at 780 °C (1435 °F) and 140 m/s (459 ft/s) particle velocity. Source: Ref 21

Fig. 8.12

Fig. 8.13

Effect of the impingement angle on erosion rate of cobalt in air. Source: Ref 21

Effect of the impingement angle on erosion rate of MA754 (Ni-20Cr-0.6Y2O3, an oxide-dispersionstrengthened alloy) at 780 °C (1435 °F) in air under 140 m/s (459 ft/s) particle velocity. Source: Ref 22

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

242 / High-Temperature Corrosion and Materials Applications

particles. The erosion test results in terms of the specimen thickness loss (µm/h) are summarized in Table 8.1. The data indicate that nickel-base alloys were better than Fe-Ni-Cr alloys, which were better than Fe-Cr alloys. Tabakoff et al. (Ref 25) found that aluminized coatings were capable of providing a signiﬁcant reduction in erosion rates under very high particle velocities and high temperatures in the conditions that were similar to gas turbines. This is illustrated in Fig. 8.15 and 8.16 (Ref 25). In the

erosion tests conducted by Tabakoff et al., coalash particles consisting of primarily SiO2, Al2O3, and Fe2O3 with a mean particle diameter of about 15 µm were used. MAR-M246 (Ni-10Co9Cr-10W-2.4Mo-1.5Ta-1.5Ti-5.5Al) and X40 (Co-25Cr-10Ni-7.5W) were tested. Aluminized coatings tested were “C” coating on X40 and “N” coating on M246. Also included were platinum-modiﬁed aluminized coating, RT22, and rhodium/platinum-modiﬁed aluminide coating, RT22B. Both RT22 and RT22B

(a)

Fig. 8.14

Erosion behavior of nickel, cobalt, Cr2O3-forming alloys (MA754 and Ni30Cr alloys), and Al2O3forming alloys (CoCrAlY and Ni20Al alloys) in air at 600 and 780 °C (1110 and 1435 °F) under 140 m/s (459 ft/s) particle velocity and 30° impingement angle. Source: Ref 22

(b)

Table 8.1 Metal loss rates for alloys tested in a combustion atmosphere using a propane-fired abrasive jet rig at 975 °C (1790 °F), 55 m/s (123 mph) particle velocity of about 135 µm (5 mils) SiC abrasive particles Metal loss rate, µm/h Alloy

45° impinging angle

90° impinging angle

671 601 800 310 446 316 304

31 62 80 80 130 160 190

19 29 70 73 90 150 195

Source: Ref 24

(c)

Fig. 8.15

Erosion behavior of M246 (a), X40 (b), and RT22 coating (c) under hot, oxidizing combustion gas stream at 815 °C (1500 °F) and 366 m/s (1200 ft/s) with fly-ash as erodent. Source: Ref 25

pg 242

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

Chapter 8: Erosion and Erosion-Corrosion / 243

were deposited on M246. The thickness was 0.076 mm (3 mils) for aluminized coating “C,” “N,” and RT22B, and 0.127 mm (5 mils) for RT22. Test results showed that aluminized coatings signiﬁcantly reduced the erosion rates of the alloy with both platinum- and rhodium/platinum-modiﬁed aluminide coatings providing the most resistance. The erosion and erosion-corrosion (oxidation) data presented so far involve very high particle velocities. Erosion or erosion-corrosion rates of materials can be a strong function of particle velocities. Many of the industrial applications involve particle velocities that are much lower than the data presented earlier. Wright et al. (Ref 26) investigated the erosion-corrosion behavior of 10 commercial alloys as a function of particle velocities (27 to 52 m/s, or 88 to 170 ft/s) at 760 °C (1400 °F) in an oxidizing environment (N2-15CO2-3O2-0.03SO2), which was referred to as “FBC” gas. Alumina particles (15 µm size) with a loading in the gas stream of about 15,000 ppm (by wt) were used in the tests with each test being about 4 to 6 h. Test results showed a sharp transition from low metal-loss rates to a regime of rapid metal-loss rates at particle velocities in the range of about 27 to 34 m/s (90 to 110 ft/s) (Ref 26). The authors termed this regime of high metal-loss rates as erosion dominated. The test results on some of the alloys are summarized in Fig. 8.17 (Ref 26). Under the particle velocities of 90 and 140 m/s (295 and 459 ft/s), Kang et al. (Ref 20) and Chang et al. (Ref 21) observed that signiﬁcant 4

X40

Erosion rate, mgm/gm

M246 3

2 C 1 N RT22B

RT22

0

Fig. 8.16

Erosion behavior of alloys and coatings under hot, oxidizing combustion gas stream at 815 °C (1500 °F) and 366 m/s (1200 ft/s) with fly-ash as erodent. Source: Ref 25

plastic deformation developed on the metal surface. This was caused by the impinging particles for nickel and cobalt tested in air at 800 °C (1470 °F). In addition, thin, discontinuous oxide scales were observed to form on the eroded surface (Ref 20, 21). The impinging particles not only removed the oxide scales, but also caused signiﬁcant plastic deformation on the metal surface, developing a rippled surface with mounds and valleys (Ref 20 to 23). However, the repeated removal and reformation of oxide scales was attributed to the accelerated erosion-corrosion rates in oxidizing environments (Ref 20 to 23). Since signiﬁcant plastic deformation was observed under high particle velocities, the mechanical properties, hardness, and the microstructure of the alloy may play an important role in affecting the erosion-dominated erosioncorrosion behavior of the alloy. Various mechanisms for erosion have been proposed in the literature, such as, cutting by Finne (Ref 10), cutting and ploughing by Hutchings and Winter (Ref 27), ﬂake formation by Brown et al.(Ref 28), platelet formation by Bellman and Levy (Ref 29), and deformation wear by Bitter (Ref 30). Hardness has often been used in industry as a key material property for making materials selection for resisting wear. Hardness increases can arise from different hardening mechanisms, such as cold working, formation of ﬁne-coherent precipitates, martensitic transformation, and second-phase hard particles (e.g., eutectic carbides, tungsten carbides, etc.). In industrial trialand-error tests, hardening that resulted from cold working, coherent precipitation, and martensitic transformation is somewhat beneﬁcial in resisting sliding wear but not erosion or erosioncorrosion. Hardness increases by second-phase, hard particles appear to be beneﬁcial in increasing the erosion resistance of the alloy. Under erosion-corrosion conditions at high particle velocities and elevated temperatures, alloys that exhibit good oxidation resistance (or hightemperature corrosion resistance) and are hardened by second-phase hard particles are likely to be promising candidates for the application. Increasing hardness can adversely affect alloy toughness. Thus, application of a suitable coating or weld overlay hardfacing material appears to be a more viable practical approach to the problem. At lower particle velocities, the erosioncorrosion behavior can be strongly dependent on the type of oxide scales formed on the alloy. Steels and low-alloy steels that form iron oxides can suffer much more erosion-corrosion attack

pg 243

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

244 / High-Temperature Corrosion and Materials Applications

than stainless steels with adequate chromium for forming chromium oxide scales. This is illustrated by the test results generated by Levy and Man (Ref 31) in Fe-Cr alloys with various chromium contents at 850 °C (1560 °F) in air under the particle velocity of 35 m/s (115 ft/s) with alumina particles (130 µm particle size). The results are summarized in Fig. 8.18 (Ref 31). Levy and Man (Ref 31) also performed static oxidation tests on the same alloys in air at the same temperature (850 °C, or 1560 °F). Their oxidation tests showed decreasing weight gain with increasing chromium content in the alloy, as shown in Fig. 8.19. This behavior is quite similar to that generated under erosion-corrosion conditions (Fig. 8.18). Erosion-corrosion behavior of alloys in a simulated coal gasiﬁcation environment was studied by Agarwal and Howes (Ref 32). The alloy that was less resistant to sulﬁdation attack

in the coal gasiﬁcation environment was also found to be less resistant to erosion-corrosion attack in the same environment. This is illustrated in Fig. 8.20 (Ref 32), showing Type 310 suffering signiﬁcantly more erosion-corrosion attack than alloy 6B in a MPC coal gasiﬁcation environment (24H2-18CO-12CO2-39H2O-5CH41NH3-1H2S) tested at 815 °C (1500 °F) with a particle velocity of 15.3 m/s (50 ft/s) and using metallurgical coke as the erodent (300 to 600 µm particle size). [MPC, Materials Property Council coal-gasiﬁcation test program (See Chapter 7 “Sulﬁdation”).] Hard particles, such as alumina, can cause more erosion-corrosion attack than soft particles, such as coke. This is illustrated in Fig. 8.21 (Ref 33). In general, the alloys that are more resistant to sulﬁdation are better at resisting erosion-corrosion in sulﬁdizing environments. Figure 8.22 shows the results of erosion-corrosion tests at 980 °C (1800 °F) in

(a)

(b)

(c)

(d)

Fig. 8.17

Effect of particle velocity on the erosion-corrosion rate for (a) Type 446, (b) alloy 671, (c) alloy 188, and (d) alloy 6B tested at 760 °C (1400 °F) in a simulated combustion gas stream (designated as FBC gas, N2-15CO2-3O2-0.03SO2) containing 15 µm alumina particles with a particle loading of 15,000 ppm (by wt). 1 ft = 0.305 m. Source: Ref 26

pg 244

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

pg 245

Chapter 8: Erosion and Erosion-Corrosion / 245

a coal-gasiﬁcation environment eroded with metallurgical coke at 30.5 m/s (100 ft/s) particle velocity as a function of exposure time up to 200 h (Ref 33). Some less-resistant alloys were found to suffer “breakaway” erosion-corrosion. This “breakaway” erosion-corrosion is quite similar to “breakaway” corrosion in reducing sulﬁdizing environments. After 200 h, cobaltbase alloy No. 1 (Co-30Cr-12W-2.5C) was found to be most resistant among alloys tested, showing no breakaway erosion-corrosion. Alloy 671 (Ni-48Cr-0.5Ti) also showed no breakaway erosion-corrosion. Both of these were highchromium alloys, which are known to be highly

120

Corrosion-erosion

110

resistant to sulﬁdation attack. (For more information about the sulﬁdation behavior of alloys in coal-gasiﬁcation environments, readers are referred to Chapter 7.) In the regime where particle velocities are lower, the erosion-corrosion behavior is likely to be dominated by corrosion. The alloys that are more resistant to corrosion (i.e., high-temperature corrosion) are generally better at resisting erosion-corrosion in the same environment. This may provide a practical guide to materials selection for applications that are under erosioncorrosion conditions. In this regime, it is believed that impinging particles cause damage only to oxides or corrosion products without signiﬁcantly affecting the underlying metal. The most likely scenario in this regime involves the impinging particles removing oxide scales or corrosion products and allowing the fresh metal to be exposed to the corrosive environment. Oxide scales or corrosion products form and are then removed by subsequent impinging particles. This process involves repeated removal of the corrosion products by eroding particles and reformation of the fresh corrosion products, thus resulting in erosion-induced accelerated corrosion. This is best described graphically by a schematic showing a corrosion mode where

60 Alloy 310

Fig. 8.18

Effect of chromium in Fe-Cr alloys on the erosioncorrosion resistance of the alloys at 850 °C (1560 °F) in air with 35 m/s (115 ft/s) particle velocity (130 µm alumina particles). Source: Ref 31

Maximum thickness loss, mils

50

40

30

20

Alloy 6B

10

0 20

40

60

80

100

Impingement angle, degrees

Fig. 8.20 Fig. 8.19

Effect of chromium in Fe-Cr alloys on the oxidation resistance of the alloys at 850 °C (1560 °F) in air. Source: Ref 31

Erosion-corrosion behavior of Type 310 and alloy 6B in MPC coal-gasification environment (24H218CO-12CO2-39H2O-5CH4-1NH3-1H2S) tested for 100 h at 250 psig, 1500 °F (815 °C), and 50 ft/s, with metallurgical coke as erodent (300 to 600 µm). Source: Ref 32

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

246 / High-Temperature Corrosion and Materials Applications

Fig. 8.21

Erosion-corrosion behavior of various alloys tested at 980 °C (1800 °F) for 50 h in MPC coal-gasification environment at 1000 psig with 30.5 m/s (100 ft/s) at 45° impingement angle, using coke and alumina as erodents. S-1: Stellite No. 1: Co30Cr-12W-2.5C; Cru 25: Fe-25Cr-25Ni; LM 1866: Fe-18Cr-6Al-0.6Hf; Alloy 671: Ni-48Cr-0.5Ti; RA333: Ni-25Cr-18Fe-3Mo-3W; Alloy 188: Co-22Cr-22Ni-14W-0.04La; 800AL (aluminized alloy 800); and 310AL (aluminized 310). Source: Ref 33

the corrosion scale reaches a steady state and an erosion-corrosion mode where the corrosion scale is continuously reformed after it is repeatedly removed by eroding particles, as shown in Fig. 8.23 (Ref 34).

8.3 Summary Erosion and erosion-corrosion behavior of alloys are reviewed. The data examined are generated mainly in laboratory tests using a “jet”type apparatus where the metal is impacted by a particle-laden gas stream. Under conditions involving very high particle velocities, such as 100 m/s (328 ft/s) or higher, at elevated

temperatures, oxidizing environments (e.g., air) signiﬁcantly accelerate the erosion-corrosion rates compared with an inert environment, such as N2 or argon. Under such high particle velocities, thin, disconnected oxide scales (instead of a continuous oxide scale) were observed to form in addition to severe plastic deformation on the underlying metal that caused the formation of a rippled surface with mounds and valleys. Some limited data suggest alloys that form Cr2O3 or Al2O3 showing less scaling are better than those having high scaling rates for resisting erosion-corrosion attack. The data also suggest that aluminized coatings are capable of reducing erosion-corrosion rates in oxidizing environments. Hardness has often been used as an important material property in resisting wear.

pg 246

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:33PM Plate # 0

Chapter 8: Erosion and Erosion-Corrosion / 247

Fig. 8.22

Erosion-corrosion behavior of various alloys tested at 980 °C (1800 °F) in MPC coal-gasification environment, 500 psig, 100 ft/s, coke as erodent. S-1: Stellite No. 1: Co-30Cr-12W-2.5C; Cru 25: Fe-25Cr-25Ni; LM 1866: Fe-18Cr6Al-0.6Hf; Alloy 671: Ni-48Cr-0.5Ti; RA333: Ni-25Cr-18Fe3Mo-3W; Type 310: Fe-25Cr-20Ni. Source: Ref 33

Fig. 8.23

A corrosion mode where the corrosion rate is diminishing in time and an erosion-corrosion (E-C) mode where the corrosion products are repeatedly removed by impinging particles and reformed subsequently. Source: Ref 34

REFERENCES

Limited laboratory data generated so far have failed to provide practical guidance on the hardness of the metal in resisting erosioncorrosion attack at elevated temperatures. In industrial trial-and-error tests, hardening of the alloy by second-phase, hard particles appears to be beneﬁcial in improving resistance to erosioncorrosion attack at elevated temperatures. Erosion-corrosion studies conducted by Wright et al. (Ref 26) suggest that an erosiondominated erosion-corrosion regime occurs when the particle velocities are in excess of the range of about 27 to 34 m/s (90 to 110 ft/s), and below that particle velocity range is a corrosiondominated erosion-corrosion regime. At lower particle velocities, that is, in a corrosiondominated erosion-corrosion regime, in both oxidizing and sulﬁdizing environments, the data suggest that the alloys that provide better oxidation or sulﬁdation resistance are likely to provide better erosion-corrosion resistance in the same environment. This may provide a general practical guide to materials selection for application in the regime where the corrosion dominates the erosion-corrosion reactions.

1. G.E. Moller and C.W. Warren, Survey of Tube Experience in Ethylene and Oleﬁns Pyrolysis Furnaces: T-5B-6 Task Group Report, Corrosion/81, NACE 2. D. Jakobi and R. Gommans, Typical Failures in Pyrolysis Coils for Ethylene Cracking, Mater. Corros., Vol 54 (No. 11), 2003, p 881 3. S.C. Stultz and J.B. Kitto, Ed., Steam: Its Generation and Use, Babcock & Wilcox, 1992, p 33-1 4. S.C. Stultz and J.B. Kitto, Ed., Steam: Its Generation and Use, Babcock & Wilcox, 1992, p 20-16 5. J.G. Singer, Ed., Combustion Fossil Power, Combustion Engineering, Inc. (now Alstom Power), 1991, p 7–8 6. J.G. Singer, Ed., Combustion Fossil Power, Combustion Engineering, Inc. (now Alstom Power), 1991, p 8–14 7. S.C. Stultz and J.B. Kitto, Ed., Steam: Its Generation and Use, Babcock & Wilcox, 1992, p 27-1 8. J.G. Singer, Ed., Combustion Fossil Power, Combustion Engineering, Inc. (now Alstom Power), 1991, p 23–21

pg 247

Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/

26/10/2007 12:34PM Plate # 0

248 / High-Temperature Corrosion and Materials Applications

9. M.M. Stack, F.H. Stott, and G.C. Wood, Mater. High Temp., Vol 9 (No. 3), 1991, p 153 10. I. Finne, Wear, 1960, p 87 11. A.V. Levy, Chapter 5: Erosion and ErosionCorrosion of Steels at Elevated Temperatures, Solid Particle Erosion and Erosion-Corrosion of Materials, ASM International, 1995 12. A.V. Levy, Erosion-Corrosion of Tubing Steels in Combustion Boiler Environments, Paper No. 236, Corrosion 93, NACE, 1993 13. Y. Shida and H. Fujikawa, Particle Erosion Behaviour of Boiler Tube Materials at Elevated Temperature, Wear, Vol 103, 1985, p 281 14. V. Nagarajan and I.G. Wright, Inﬂuence of Oxide Scales on High Temperature Corrosion Erosion Behavior of Alloys, High Temperature Corrosion, R.A. Rapp, Ed., Conf. Proc. (San Diego, CA), March 2–6, 1981, NACE, 1981, p 398 15. Kanthal Handbook: Resistance Heating Alloys and Elements for Industrial Furnaces, Kanthal Heating Systems, Hallstahammar, Sweden, p 7 16. AISI Type 446, Alloy Digest, Engineering Alloys Digest, Inc., May 1982 17. M.J. Donachie and S.J. Donachie, Superalloys: A Technical Guide, 2nd ed., ASM International, 2002, p 249 18. Haynes Stellite Alloy No. 6B, Alloy Digest, Engineering Alloys Digest, Inc., Sept 1960 19. I.G. Wright, V. Nagarajan, and R.B. Herchenroeder, Some Factors Affecting Solid Particle Erosion-Corrosion of Metals and Alloys, Corrosion-Erosion Behavior of Materials, K. Natesan, Ed., Conf. Proc., Fall Meeting of The Metallurgical Society of AIME (St. Louis, MO), October 17–18, 1978 20. C.T. Kang, F.S. Pettit, and N. Birks, Mechanisms in the Simultaneous ErosionOxidation Attack of Nickel and Cobalt at High Temperatures, Metall. Trans. A, Vol 18, 1987, p 1785 21. S.L. Chang, F.S. Pettit, and N. Birks, Effect of Angle of Incidence on the Combined Erosion-Oxidation Attack of Nickel and Cobalt, Oxid. Met., Vol 34 (No. 1 & 2), 1990, p 47

22. S.L. Chang, F.S. Pettit, and N. Birks, Some Interactions in the Erosion-Oxidation of Alloys, Oxid. Met., Vol 34 (No. 1 & 2), 1990, p 71 23. D.M. Rishel, F.S. Pettit, and N. Birks, Some Principle Mechanisms in the Simultaneous Erosion and Corrosion Attack of Metals at High Temperatures, Mater. Sci. Eng., Vol A143, 1991, p 197 24. L.K. Ives, Erosion of 310 Stainless Steel at 975 °C in Combustion Gas Atmospheres, Trans. ASME, April 1977, p 126 25. W. Tabakoff, A. Hamed, M. Metwally, and M. Pasin, High-Temperature Erosion Resistance of Coatings for Gas Turbine, Trans. ASME, Vol 114, 1992, p 242 26. I.G. Wright, V. Nagarajan, W.E. Merz, and J. Stringer, The Kinetics of HighTemperature Erosion-Corrosion of Oxidation-Resistant Alloys, Corrosion/81, NACE, 1981 27. I.M. Hutchings and R.E. Winter, Wear, Vol 27, 1974, p 121 28. R. Brown, E.J. Jun, and J.W. Edington, Wear, Vol 70, 1981, p 347 29. R. Bellman and A.V. Levy, Wear, Vol 70, 1981, p 1 30. J.G.A. Bitter, Wear, Vol 6, 1963, p 5 31. A.V. Levy and Y.F. Man, Erosion-Corrosion Mechanisms and Rates in Fe-Cr Steels, Wear, Vol 131 (No. 1), 1989, p 39 32. S.C. Agarwal and M.A.H. Howes, Erosion-Corrosion Materials in HighTemperature Environments: Impingement Angle Effects in Alloys 310 and 6B under Simulated Coal Gasiﬁcation Atmosphere, J. Mater. Energy Syst., Vol 7 (No. 4), 1986, p 370 33. M.A.H. Howes, Elevated Temperature Erosion-Corrosion of Alloys in Sulﬁdizing Gas/Solid Streams: Mechanistic Studies, Proc. Conf., Corrosion-Erosion-Wear of Materials at Elevated Temperatures, NACE, 1986, p 230 34. V.K. Sethi and I.G. Wright, Observations on the Erosion-Oxidation Behavior of Alloys, in Proc. TMS Conf. on Corrosion and Particle Erosion, V. Srinivasan and K. Vedula, Ed., 1986, p 245

pg 248

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p249-258 DOI: 10.1361/hcma2007p249

pg 249

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 9

Hot Corrosion in Gas Turbines 9.1 Introduction During combustion in the gas turbine, sulfur from the fuel reacts with sodium chloride from ingested air at elevated temperatures to form sodium sulfate. The sodium sulfate then deposits on the hot-section components, such as nozzle guide vanes and rotor blades, resulting in accelerated oxidation (or sulﬁdation) attack. This is commonly referred to as “hot corrosion.” Sulfur in the fuel is generally limited to 0.3% for commercial jet engines and to 1.0% for marine gas turbines (Ref 1). Sodium chloride comes from seawater (see Table 9.1) (Ref 2). Seawater is also a source of sulfur. For aircraft engines, Tschinkel (Ref 1) suggested that runway dust may be a source of salts. Gas turbines generally use large amounts of excess air for combustion (a large fraction of air, in fact, is also used to cool the combustor), with a typical air-to-fuel ratio from about 40 to 1 (during takeoff) to 100 to 1 (at cruising speed) for aircraft gas turbine engines (Ref 2). These air-to-fuel ratios correspond to about 0.12 to 0.18 mole fractions of oxygen in the combustion zone (Ref 2). Thus, the combustion gas atmosphere is highly oxidizing. The sulfur partial pressure in the atmosphere can be extremely low, varying from 10−40 to 10−26 atm over the range from 330 to 1230 °C (620 to 2240 °F) (Ref 2). These sulfur partial pressures are well below those necessary to form chromium sulﬁdes, which are frequently observed in alloys suffering hot corrosion attack. High-temperature alloys that suffered hot corrosion attack were generally found to exhibit both oxidation and sulﬁdation. The hot corrosion morphology is typically characterized by a thick, porous layer of oxides with the underlying alloy matrix depleted in chromium, followed by internal chromium-rich sulﬁdes. It is generally believed that the molten sodium sulfate deposit is required to initiate hot corrosion attack. The temperature range for hot corrosion attack,

although dependent on alloy composition, is generally 800 to 950 °C (1470 to 1740 °F). The lower threshold temperature is believed to be the melting temperature of the salt deposit, and the upper temperature is the salt dew point (Ref 3). This type of corrosion process is sometimes referred to as Type I hot corrosion to differentiate it from Type II hot corrosion, which occurs at lower temperatures (typically 670 to 750 °C, or 1240 to 1380 °F) (Ref 4). Type II hot corrosion is characterized by pitting attack with little or no internal attack underneath the pit (Ref 4). Type II hot corrosion is rarely observed in aeroengines because the blades are generally operated at higher temperatures (Ref 5). However, marine and industrial gas turbines, which operate at lower temperatures, can experience low-temperature Type II hot corrosion. Type I hot corrosion generally proceeds in two stages: an incubation period exhibiting a low corrosion rate, followed by accelerated corrosion attack. The incubation period is related to the formation of a protective oxide scale. Initiation of accelerated corrosion attack is believed to be related to the breakdown of the protective oxide scale. Many mechanisms have been proposed to explain accelerated corrosion attack; the salt ﬂuxing model is probably the most widely accepted. Oxides can dissolve in Na2SO4 as anionic species (basic ﬂuxing) or cationic species (acid ﬂuxing), depending on the salt composition (Ref 6). Salt is acidic when it is high in SO3, and basic when low in SO3. The hot corrosion Table 9.1 Element

Chlorine Sodium Magnesium Sulfur Calcium Potassium Beryllium Source: Ref 2

Composition of seawater Composition, ppm

18,980 10,561 1,272 884 400 380 65

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

pg 250

250 / High-Temperature Corrosion and Materials Applications

mechanism by salt ﬂuxing has been discussed in detail in Ref 7, 8, and 9. The topic of hot corrosion has been extensively covered in several reports and conference proceedings (Ref 2, 10–13).

9.2 Alloys Resistant to Hot Corrosion Various test methods have been used to study hot corrosion. Immersion testing (or crucible testing), which was the ﬁrst laboratory test method, is not considered reliable for simulating the gas turbine environment (Ref 14, 15). The salt-coated method is quite popular in academia for studying corrosion mechanisms. Engine manufacturers, however, use the burner rig test system to determine relative alloy performance ranking. The rig burns fuel with excess air to produce combustion gases with continuous injection of a synthetic sea-salt solution. This type of test system represents the best laboratory apparatus for simulating the gas turbine environment. A special issue of High Temperature Technology published in 1989 contained a number of papers discussing burner rig test procedures (Ref 16). The data reviewed here are limited to those generated by burner rig test systems. 9.2.1 High Temperature or Type I Hot Corrosion Bergman et al. (Ref 17) studied hot corrosion resistance of various nickel- and cobalt-base alloys at temperatures from 870 to 1040 °C (1600 to 1900 °F) with 5 ppm sea-salt injection. Their results are tabulated in Table 9.2. The Table 9.2

data show a good correlation between alloy performance and chromium content. Increasing chromium in the alloy signiﬁcantly improves resistance to hot corrosion. Alloys with 15% Cr or less are very susceptible to hot corrosion attack. Cobalt-base alloys are generally better than nickel-base alloys. This may simply be due to higher chromium contents in cobalt-base alloys. One nickel-base alloy (Hastelloy X) with a chromium level similar to those of cobalt-base alloys was found to behave similarly to cobaltbase alloys. Among the alloys tested (Ref 17), alloy X-40 (Co-25Cr-10Ni-7.5W) performed best. This is in good agreement with the operating experience obtained by Royal Navy Ship (U.K.), which has demonstrated the superior hot corrosion resistance of alloy X-40 in a marine environment (Ref 18). Alloy X-40 was also found to be signiﬁcantly better than nickel-base alloys (Ref 19), such as B-1900, U-700, U-500, and IN738 (Table 9.3). After 240 h, alloy X-40 showed hardly any corrosion attack, while alloy B-1900 (Ni-10Co-8Cr-6Mo-4.3Ta-6Al-1Ti) suffered severe attack. Alloy U-500 (Ni-18Co-19Cr4Mo-2.9Al-2.9Ti) and IN738 (Ni-8.5Co-16Cr1.7Mo-2.6W-1.7Ta-0.9Nb-3.4Al-3.4Ti) were similar, suffering only mild attack. Surprisingly, alloy U-700 (15% Cr) was found to be slightly worse than alloy B-1900 (8% Cr). Alloy B-1900 along with IN100 (10% Cr) and Nimonic 100 (11% Cr) were considered to be poor in hot corrosion and suggested that they not be considered for use without coatings, even in mildly corrosive environments (Ref 20). Burner rig tests were conducted (Ref 21) using residual oil, containing 3% S and 325 ppm

Results of burner rig hot corrosion tests on nickel- and cobalt-base alloys Loss in sample diameter, mm (mils)

Alloy

SM-200 IN100 SEL-15 IN713 U-700 SEL U-500 Rene 41 Hastelloy alloy X L-605 (alloy 25) WI-52 MM-509 SM-302 X-40

Chromium content in alloy, %

870 °C (1600 °F) 500 h

950 °C (1750 °F) 1000 h

980 °C (1800 °F) 1000 h

1040 °C (1900 °F) 1000 h

9.0 10.0 11.0 13.0 14.8 15.0 18.5 19.0 22.0 20.0 21.0 21.5 21.5 25.0

1.6 (64.4) 3.3+ (130+) 3.3+ (130+) 3.3+ (130+) 1.7+ (66+) 1.2 (45.8) 0.2 (7.6) 0.3 (10.3) … … 0.5 (21.4) … 0.14 (5.4) 0.11 (4.2)

3.3+ (130+) 3.3+ (130+) 3.3+ (130+) 2.0+ (77+) 1.6 (63.9) 1.3 (51.8) 0.8 (31.7) … 0.3 (12.0) 0.4 (15.3) 0.5 (18.2) 0.3 (10.9) 0.3 (10.0) 0.3 (11.6)

… … … … … 0.3 (11.4) 0.7 (29.3) 0.8 (30.8) 0.4 (15.2) 0.3 (11.3) … … … …

… … … … … … …

Note: 5 ppm sea salt injection. Source: Ref 17

… 1.1 (41.9) 1.9 (73.9) 0.8 (31.8) 0.6 (23.1) 0.5 (18.5)

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

Chapter 9:

NaCl (equivalent to 5 ppm NaCl in air), at 870 °C (1600 °F) for 600 h on several cobaltbase alloys, which were X-45 (Co-25Cr-10Ni7.5W), MAR-M302 (Co-21.5Cr-10W-9Ta-0.2Zr), MAR-M509 (Co-21.5Cr-10Ni-7W-3.5Ta-0.2T0.5Zr), S-816 (Co-20Cr-20Ni-4W-4Nb-4Mo), FSX-418 (Co-30Cr-10Ni-7W-0.15Y), and FSX414 (Co-30Cr-10Ni-7W). The test results are shown in Fig. 9.1. All six cobalt-base alloys with chromium varying from 20 to 30% suffered little corrosion attack (about 0.04 to 0.12 mm, or 0.002 to 0.005 in., or 2 to 5 mils). Under the same test condition, Udimet 700 (Ni-15Cr18.5Co-5.2Mo-5.3Al-3.5T) suffered about 0.76 mm (0.03 in., or 30 mils) of attack (Ref 21). Figure 9.2 summarizes the data for a group of nickel- and cobalt-base alloys at 870 to 1040 °C (1600 to 1900 °F) (Ref 22). Alloy U-700 was found to be inferior to IN713 at 950 °C (1750 °F). This was contrary to ﬁeld experience. Alloy U-700 has served most reliably in aircraft jet engines, whereas IN713 has suffered severe hot corrosion in many applications (Ref 22). The authors attributed the high corrosion rate of alloy U-700 to the low chromium content in this heat (i.e., 13.6% versus 15.0% for regular heats). A systematic study was conducted (Ref 22) to determine the effects of alloying elements on hot corrosion resistance. In the Ni-10Co-15Cr4Al-2Ti system, decreasing chromium from 25 to 10% resulted in increases in hot corrosion attack (Fig. 9.3). The data also suggest that decreasing aluminum while increasing titanium improves hot corrosion resistance. Furthermore, addition of 8% W to Ni-10Co-15Cr-4Al-2Ti alloy resulted in no apparent change in hot corrosion resistance. In binary and ternary alloy systems of nickel- and cobalt-base alloys, these authors further observed the effectiveness of chromium in improving hot corrosion resistance at 910, 950, and 1040 °C (1675, 1750, and 1900 °F) (Ref 22). Results are shown in Fig. 9.4.

pg 251

Hot Corrosion in Gas Turbines / 251

In examining the effect of the third alloying element in Ni-Cr alloys, they found that (Ref 22):

Tungsten (8%) showed little effect at 910 and 950 °C (1675 and 1750 °F), but a slightly detrimental effect at 1040 °C (1900 °F). Cobalt was only slightly beneﬁcial. Molybdenum was detrimental at 1040 °C (1900 °F), but had little effect at lower temperatures. Titanium (5%) showed signiﬁcant improvement at 1040 °C (1900 °F), but little effect at lower temperatures. Aluminum (8%) was detrimental, causing severe hot corrosion attack at 1040 °C (1900 °F). For Co-25Cr alloys, the effect of the third alloying element was summarized as (Ref 22):

Tungsten (8%) was detrimental at 1040 °C (1900 °F), with little effect at lower temperatures.

ATTACK, mm per side 0

0.04

0.08

0.12

FSX– 414

FSX–418

S–816

MAR–M 509

MAR–M 302

X–45

Table 9.3 Results of burner rig tests at 874 °C (1605 °F) for nickel- and cobalt-base alloys

0

1

2

3

4

5

6

ATTACK, mils per side

Penetration depth, mm/1000 h (mils/1000 h) Exposure time, h

100 170 240

B-1900

U-700

U-500

IN738

X-40

2.8 (111) 2.5 (97) 2.1 (83)

… 3.3 (129) …

… 0.7 (29) 0.5 (20)

… … 0.9 (35)

… … Slight

Note: Diesel fuel containing 1.0% S, 125 ppm Na, 15 ppm Mg, 4.8 ppm Ca, 4.1 ppm K, and 225 ppm Cl, air-to-fuel ratio was 50 : 1, and 100 h cycle. Source: Ref 19

Surface loss Maximum penetration

Fig. 9.1

Relative hot corrosion resistance of cobalt-base alloys obtained from burner rig tests using 3% S residual oil and 325 ppm NaCl in fuel (equivalent to 5 ppm NaCl in air) at 870 °C (1600 °F) for 600 h. Source: Beltran (Ref 21)

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

pg 252

252 / High-Temperature Corrosion and Materials Applications

Molybdenum (6%) was detrimental at 950 and 1040 °C (1750 and 1900 °F), with little effect at 910 °C (1675 °F). Tantalum (7%) and nickel (10%) showed little effect. Burner rig tests were conducted (Ref 23) at 900 °C (1650 °F) on several wrought superalloys and nickel aluminides. The combustion gas stream was generated by using No. 2 fuel oil containing about 0.4 wt% S with an air-to-fuel ratio of 35 to 1 and injection of either 5 or 50 ppm sea salt into the combustion gas stream. The specimens were loaded in a carousel, that rotated

at 30 rpm during testing to ensure that all the specimens were subjected to the same test condition. The specimens were cycled out of the combustion gas stream once every hour for 2 min, during which time the specimens were cooled by forced air (fan cool) to less than 205 °C (400 °F). Superalloys tested were alloy X (Ni-22Cr-18.5Fe-9Mo-0.5W), alloy S (Ni-15.5Cr-14.5Mo-0.05La), alloy 230 (Ni-22Cr14W-2Mo-0.02La), alloy 625 (Ni-21.5Cr-9Mo3.6Nb), alloy 188 (Co-22Cr-22Ni-14W-0.04La), alloy 25 (Co-20Cr-10Ni-15W), and alloy 150 (Co-27Cr-18Fe). Two nickel aluminides, IC-50 (Ni-11.3Al-0.6Zr-0.02B) and IC-218

Surface loss Maximum penetration

1600 °F (870 °C) 1750 °F (950 °C) 1900 °F (1040 °C) SL- Surface loss MP- Maximum penetration Alloy U–700

75.6 SL 92.4 MP 80.2

SM–200 SEL

IN713

U–500 X

WI–52 109.2 SM–302

X–45

0

10 (250)

20 (500)

30 (750)

40 (1000)

50 (1250)

60 (1500)

70 (1750)

Loss in diameter, mils (µm)

Fig. 9.2

Relative hot corrosion resistance of nickel- and cobalt-base alloys obtained from burner rig tests at 870, 950, and 1040 °C (1600, 1750, and 1900 °F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea-salt injection. Source: Bergman et al. (Ref 22)

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

Chapter 9:

(Ni-7.8Cr-8.5Al-0.8Zr-0.02B), which were developed by Oak Ridge National Laboratory, were included in the test program. The results of tests at 900 °C (1650 °F) for 200 h with 50 ppm sea salt are summarized in Table 9.4 (Ref 23). Both IC-50 and IC-218 nickel aluminides suffered severe hot corrosion attack after 200 h at 900 °C (1650 °F) with 50 ppm sea salt being injected into the combustion gas stream. Scanning electron microscopy with energy-dispersive x-ray spectroscopy (SEM/ EDX) analysis showed that both nickel aluminides exhibited porous nickel or nickel-rich oxides with nickel sulﬁde penetrating through the remaining metal (Ref 23). Figure 9.5 shows the cross section of a corroded IC-218 specimen after hot corrosion burner rig testing at 900 °C

pg 253

Hot Corrosion in Gas Turbines / 253

(1650 °F) for 200 h with 50 ppm sea salt, revealing the formation of nickel oxides and nickel sulﬁdes. SEM/EDX analysis showed that a thin, protective chromium-rich oxide scale formed on alloys X, 230, and 188. Alloy 25, although exhibiting little weight change (Table 9.4), showed evidence of initial breakdown of the chromium-rich oxide scale. SEM/EDX analysis revealed the formation of cobalt-rich oxide nodules on the outer oxide scale on alloy 25. This indicated the initiation of the breakaway corrosion for alloy 25 after 200 h at 900 °C (1650 °F) with 50 ppm sea salt. Longterm test results under the same test condition clearly showed that alloy 25 suffered severe hot corrosion in excess of 200 h of testing, as shown in Fig. 9.6 (Ref 23).

Loss in diameters, µm 0

Alloy

250

500

750

1000

1250

1500

1750

2000

Ni-10Co-0.5Nb-4Al-2Ti-25Cr TEL-1 Ni-10Co-0.5Nb-4Al-2Ti-20Cr TEL-2 Ni-10Co-0.5Nb-4Al-2Ti-15Cr TEL-3 140

Ni-10Co-0.5Nb-4Al-2Ti-10Cr TEL-4 Ni-10Co-0.5Nb-6Al-0Ti-15Cr TEL-5 Ni-10Co-0.5Nb-5Al-3Ti-15Cr TEL-6 Ni-10Co-0.5Nb-2Al-4Ti-15Cr TEL-7 Ni-10Co-0.5Nb-8W-4Ai-2Ti-15Cr TEL-8 Ni-10Co-0.5Nb-4Mo-15Cr TEL-9 Ni-0Co-0.5Nb-2Mo-4W-15Cr TEL-10 Ni-25Co-0.5Nb-15Cr TEL-11 Ni-0Co-0.5Nb-6Al-4Ti-15Cr TEL-12 Ni-15Co-4Mo-3.7Al-1.8Ti-19 Cr MELINI-1 Ni-15Co-3.5Al-1.8Ti-4.5 Re-19Cr MELINI-2 Ni-16Co-3.6Al-2.1Ti-21Cr MELINI-3 Surface loss

Ni-15Co-4Mo-2.9Al-1.8Ti-0.2Y-19Cr MELNI-4 Ni-25Co-4Mo-3.5Al-2.2Ti-0.2Y-19Cr MELNI-5

Maximum penetration

950 °C (1750 °F) 1040 °C (1900 °F)

Ni-15Co-4Mo-3.7Al-1.8Ti-0.08Y-19Cr MELNI-6

0

10

20

30

40

50

60

70

80

Loss in diameters, mils

Fig. 9.3 (Ref 22)

Relative hot corrosion resistance of experimental alloys obtained from burner rig tests at 950 and 1040 °C (1750 and 1900 °F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea-salt injection. Source: Bergman et al.

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

254 / High-Temperature Corrosion and Materials Applications

Figure 9.6 also shows alloys 230 and 188 exhibiting very little weight change for exposure time up to 1000 h. The results of the 1000 h tests Maximum penetration measurements 1675 °F (910 °C) 1750 °F (950 °C) 1900 °F (1040 °C)

CP-Complete penetration through the specimen

Alloy RL-1 RL-2 RL-3 RL-4 RL-5 RL-6

Ni-15Cr

Ni-25Cr

Ni-15Cr-8W

Ni-15Cr-24Co

130CP

Ni-15Cr-6Mo

9.2.2 Low-Temperature or Type II Hot Corrosion

Ni-15Cr-5Ti

RL-9 RL-10

130CP

Ni-10Cr

RL-7 RL-8

}

UNALLOYED NICKEL

90

Ni-15Cr-8Al

UNALLOYED COBALT

RL-11

}

130CP

Co-25Cr

RL-12

Co-25Cr-8W

RL-13

130CP

Co-25Cr-6Mo

RL-14

Co-25Cr-7Ta

RL-15

Co-25Cr-10Ni

0

20 30 40 50 60 70 10 (250) (500) (750) (1000) (1250) (1500) (1750) Loss in diameter, mils (µm)

Fig. 9.4

are summarized in Table 9.5 (Ref 23). Three nickel-base alloys (alloys S, X, and 625) and one cobalt-base alloy (alloy 25) were completely corroded before the test reached 1000 h, while nickel-base alloy 230 and cobalt-base alloys 150 and 188 exhibited little corrosion attack after 1000 h. Under the same test conditions in the same burner rig, nickel-base alloy HR-160 with 29% Co, 28% Cr and 2.75% Si was found to perform as well as alloys 230, 150, and 188. Figure 9.7 shows the conditions of the test specimens comparing HR-160 with other wrought alloys (Ref 24). Another burner rig test was conducted (Ref 23) with 5 ppm sea salt at 900 °C (1650 °F) under the same combustion conditions (i.e., No. 2 fuel oil with 0.4% S and 35-to-1 airto-fuel ratio). The results are summarized in Table 9.6. Even at this low level of sea salt (5 ppm) in the combustion gas stream, cobaltbase alloy 25 continued to exhibit very poor hot corrosion resistance compared with some nickelbase alloys.

Relative hot corrosion resistance of experimental alloys obtained from burner rig tests at 910, 950, and 1040 °C (1675, 1750, and 1900 °F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea salt injection. Source: Bergman et al. (Ref 22)

“Low-temperature” or Type II hot corrosion has been observed at temperatures lower than the temperature range where Type I hot corrosion has been encountered. Severe hot corrosion of alloy S590 and Nimonic 80A after several thousand hours of operation in a gas turbine that burned blast-furnace gas with the 700 to 730 °C (1290 to 1345 °F) turbine entry temperature was reported in Ref 25. In 1976, a new form of hot corrosion attack of gas turbine airfoil materials in a marine gas turbine was reported (Ref 26). The ﬁrst-stage turbine blades coated with a CoCrAlY coating, which had exhibited satisfactory hot corrosion resistance for metal temperatures in the range 800 to 1000 °C (1470 to 1830 °F), were found to suffer corrosion attack for metal temperatures at about 600 to 730 °C (1110 to 1345 °F) (Ref 26). The corrosion

Table 9.4 Results of burner rig hot corrosion tests at 900 °C (1650 °F) for 200 h with 50 ppm sea salt with specimens being cycled once every hour Alloy

X 230 25 188 IC-50 IC-218

Weight change, mg/cm2

Metal loss, mm (mils)

Total depth of attack(a), mm (mils)

−0.76 −1.35 −1.62 0.93 72 83

0.02 (0.6) 0.02 (0.8) 0.02 (0.9) 0.02 (0.7) >0.72 (28.3), completely corroded >0.75 (29.5), completely corroded

0.07 (2.8) 0.06 (2.4) 0.07 (2.8) 0.04 (1.6) >0.72 (28.3), completely corroded >0.75 (29.5), completely corroded

(a) Metal loss + internal penetration. Source: Ref 23

pg 254

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

Chapter 9:

Hot Corrosion in Gas Turbines / 255

4

5

3 2 1

100 µm

Fig. 9.5

Scanning electron backscattered image showing the cross section of a corroded IC-218 nickel aluminide specimen after hot corrosion burner rig testing at 900 °C (1650 °F) for 200 h with 50 ppm sea salt using No. 2 fuel oil (0.4% S) for combustion at 35:1 air-to-fuel ratio. The results (wt%) of EDX analysis are: 1: 100% Ni; 2: 74% Ni, 26% S; 3: 100% Ni; 4: 88% Ni, 8% Cr, 4% Al; and 5: 98% Ni, 2% Al. Areas 1, 4, and 5 were essentially nickel oxides, area 2 was nickel sulfide, and area 3 was pure nickel. Courtesy of Haynes International, Inc.

products formed on the CoCrAlY coating were found to contain CoSO4 and NiSO4 (Ref 26). It was proposed (Ref 27) that the mechanism of low-temperature hot corrosion attack of a CoCrAlY coating involved the formation of the low melting Na2SO4-CoSO4 eutectic (melting point of 565 °C, or 1045 °F). Nickel-base alloys were, in general, more resistant to Type II hot corrosion than cobalt-base alloys as found in Ref 28. It was also found (Ref 29) that NiCrAlY and NiCoCrAlY coatings were, in general, more resistant than CoCrAlY coatings. Increases in chromium content in superalloys and coatings provided signiﬁcant increases in low-temperature hot corrosion resistance (750 °C, or 1380 °F) for these materials (Ref 30). Both IN939 (23% Cr) and NiCrAlY coating (39% Cr) were found to exhibit good low-temperature hot corrosion resistance. NiCrAlY coatings with 26, 34, and 42% Cr were tested and found signiﬁcant resistance to lowtemperature hot corrosion for coatings with only 34 and 42% Cr (Ref 31). MCrAlY coatings (M = Co and/or Ni) with 30 to 35% Cr were tested (Ref 32). All of the coatings showed improved resistance to low-temperature hot

corrosion (705 °C, or 1300 °F) compared with CoCrAlY coating with about 20% Cr. A critical chromium content of no less than 37% was required for cobalt-base coatings to provide resistance to both low- and high-temperature hot corrosion (Ref 33).

9.3 Summary High-temperature or Type I hot corrosion generally occurs in the temperature range of 800 to 950 °C (1470 to 1740 °F). It is believed that the molten sodium sulfate deposit is required to initiate hot corrosion attack. The Type I hot corrosion morphology is typically characterized by a thick, porous layer of oxides with the underlying alloy matrix depleted in chromium, followed by internal chromium-rich sulﬁdes. Low-temperature or Type II hot corrosion generally occurs in the temperature range of 670 to 750 °C (1238 to 1382 °F). Type II hot corrosion is characterized by pitting attack with little or no internal attack underneath the pit. Cobalt-base alloys are more susceptible to Type II hot corrosion, which generally involves

pg 255

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

256 / High-Temperature Corrosion and Materials Applications

20 Haynes alloy 230

10

Haynes alloy 188

0

Table 9.5 Results of burner rig hot corrosion tests at 900 °C (1650 °F) for 1000 h with 50 ppm sea salt with specimens being cycled once every hour

–10 Alloy

–20

Weight change, mg/cm2

…

X

…

625

…

230 25

11.5 …

–80

150 188

10.0 9.9

–90

(a) Metal loss + internal penetration. Source: Ref 23

–30 –40 –50 Weight change, mg/cm2

Total depth of attack(a), mm (mils)

S

–60 –70

Completely corroded in 350 h (1.3 mm, or 49.2 mil thick specimen) Completely corroded in 500 h (1.2 mm, or 45.9 mil thick specimen) Completely corroded in 940 h (1.6 mm, or 63.7 mil thick specimen) 0.11 (4.4) Completely corroded in 476 h (1.0 mm, or 37.8 mil thick specimen) 0.13 (5.2) 0.08 (3.2)

–100 Haynes alloy 25

–110 –120 –130 –140 –150 –160 –170 –180

0 100 200 300 400 500 600 700 800 9001000 Exposure time, h

Fig. 9.6

Results of burner rig tests at 900 °C (1650 °F) with 50 ppm sea salt using No. 2 fuel oil (0.4% S) for combustion at 35:1 air-to-fuel ratio for alloys 230, 188, and 25. Source: Lai et al. (Ref 23)

Na2SO4 and CoSO4. Increasing chromium in alloys or coatings will improve the resistance of the material to both Type I and Type II hot corrosion attack.

REFERENCES

1. J.G. Tschinkel, Corrosion, Vol 28 (No. 5), 1972, p 161 2. J. Stringer, “Hot Corrosion in Gas Turbines,” Report MCIC-72-08, Battelle Columbus Laboratories, Columbus, OH, 1972 3. J. Stringer, High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 3 4. J. Stringer, Coatings in the Electricity Supply Industry: Past, Present, and Opportunities for the Future, Surf. Coat. Technol., Vol 108/109, 1998, p 1

5. A.K. Koul, J.P. Immarigeon, R.V. Dainty, and P.C. Patnaik, Degradation of High Performance Aero-Engine Turbine Blades, Advanced Materials and Coatings for Combustion Turbines, V.P. Swaminathan and N.S. Cheruvu, Ed., ASM International, 1994, p 69 6. G.H. Meier, High Temperature Corrosion 2 —Advanced Materials and Coatings (Les Embiez, France), May 22–26, 1989, Elsevier Science, 1989, p 1 7. J.A. Goebel, F.S. Pettit, and G.W. Goward, Metall. Trans., Vol 4, 1973, p 261 8. J. Stringer, Am. Rev. Mater. Sci., Vol 7, 1977, p 477 9. R.A. Rapp, Corrosion, Vol 42 (No. 10), 1986, p 568 10. J. Stringer, R.I. Jaffee, and T.F. Kearns, Ed., High Temperature Corrosion of Aerospace Alloys, Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organizations, AGARD-CP-120, Harford House, London, 1973 11. J.W. Fairbanks and I. Machlin, Ed., Proc. 1974 Gas Turbine Materials in The Marine Environment Conf., MCIC-75-27, Battelle Columbus Laboratories, 1974 12. Hot Corrosion Problems Associated with Gas Turbines, STP 421, ASTM, 1967 13. A.B. Hart and A.J.B. Cutler, Ed., Deposition and Corrosion in Gas Turbines, Applied Science, London, 1973 14. J.F.G. Conde and G.C. Booth, Deposition and Corrosion in Gas Turbines, A.B. Hart and A.J.B. Cutler, Ed., John Wiley & Sons, 1973, p 278

pg 256

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

Chapter 9:

pg 257

Hot Corrosion in Gas Turbines / 257

310SS

RA330 alloy

Alloy 800H

Alloy 625

HR160® alloy

1000 h

476 h

1000 h

940 h

1000 h

Fig. 9.7

Test specimens alloy HR-160, 625, 800H, RA330, and Type 310 at 900 °C (1650 °F) in the combustion gas stream generated by a burner rig using No. 2 fuel oil (0.4% S) for combustion at 35:1 air-to-fuel ratio and with injection of 50 ppm sea salt into the combustion gas stream. During testing, specimens were cycled once every hour. Source: Ref 24

Table 9.6 Results of burner rig hot corrosion tests at 900 °C (1650 °F) for 1000 h with 5 ppm sea salt with specimens being cycled once every hour Alloy

Weight change, mg/cm2

Metal loss, mm (mils)

Total depth of attack(a), mm (mils)

X 230 625 25

−0.24 −0.79 5.87 …

0.04 (1.6) 0.03 (1.2) 0.05 (1.9) …

188

1.09

0.02 (0.8)

0.14 (5.5) 0.13 (5.1) 0.14 (5.3) Completely corroded (1.09 mm, or 43 mils) 0.07 (2.8)

(a) Metal loss + internal penetration. Source: Ref 23

15. M.J. Donachie, R.A. Sprague, R.N. Russell, K.G. Boll, and E.F. Bradley, Hot Corrosion Problems Associated with Gas Turbines, STP 421, ASTM, 1967, p 85 16. High Temperature Technology, Special Issue on Hot-Salt Corrosion Standards Test Procedures and Performance, Vol 7 (No. 4), Nov 1989 17. P.A. Bergman, A.M. Beltran, and C.T. Sims, Development of Hot Corrosion-Resistant Alloys for Marine Gas Turbine Service, Final Summary Report to Marine Engineering Lab., Contract N600 (61533) 65661, Navy Ship R&D Center, Annapolis, MD, Oct 1, 1967 18. J. Clelland, A.F. Taylor, and L. Wortley, Proc. 1974 Gas Turbine Materials in the Marine Environment Conf., MCIC-75-27, J.W. Fairbanks and I. Machlin, Ed., Battelle Columbus Laboratories, 1974, p 397

19. M.J. Zetlmeisl, D.F. Laurence, and K.J. McCarthy, Mater. Perform., June 1984, p 41 20. J. Stringer, Proc. Symp. Properties of High Temperature Alloys with Emphasis on Environmental Effects, Z.A. Foroulis and F.S. Pettit, Ed., The Electrochemical Society, 1976, p 513 21. A.M. Beltran, Cobalt, Vol 46, 1970, p 3 22. P.A. Bergman, C.T. Sims, and A.M. Beltran, Hot Corrosion Problems Associated with Gas Turbines, STP 421, ASTM, 1967, p 38 23. G.Y. Lai, J.J. Barnes, and J.E. Barnes, “A Burner Rig Investigation of the Hot Corrosion Behavior of Several Wrought Superalloys and Intermetallics,” Paper 91-GT-21, International Gas Turbine and Aeroengine Congress and Exposition (Orlando, FL), June 3–6, 1991 24. “Haynes HR-160 Alloy,” H-3129A, Haynes International, Inc., Kokomo, IN 25. W. Moller, Deposition and Corrosion in Gas Turbines, A.B. Hart and A.J.B. Cutler, Ed., John Wiley and Sons, 1973, p 1 26. D.J. Wortman, R.E. Fryxell, and I.I. Bessen, A Theory for Accelerated Turbine Corrosion at Intermediate Temperatures, Proc. Third Conference on Gas Turbine Materials in a Marine Environment, Session V, Paper 11 (Bath, England), 1976 27. K.L. Luthra, Metall. Trans. A, Vol 13, 1982, p 1853 28. D.J. Wortman, R.E. Fryxell, K.L. Luthra, and P.A. Bergman, Proc. Fourth US/UK Conf. on Gas Turbine Materials in a Marine

Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/

26/10/2007 2:02PM Plate # 0

258 / High-Temperature Corrosion and Materials Applications

Environment (Annapolis, MD), U.S. Naval Academy, 1979, p 317 29. L.F. Aprigliano, Proc. Fourth US/UK Conf. on Gas Turbine Materials in a Marine Environment (Annapolis, MD), U.S. Naval Academy, 1979, p 151 30. A.R. Taylor, B.A. Wareham, G.C. Booth, and J.F. Conde, Low and High Pressure Rig Evaluation of Materials and Coatings, Proc. Third Conference on Gas Turbine Materials in a Marine Environment, Session III, Paper No. 3 (Bath, England), 1976 31. J.F.G. Conde, G.C. Booth, A.F. Taylor, and C.G. McGreath, Hot Corrosion in

Marine Gas Turbines, Proc. Conf. High Temperature Alloys for Gas Turbines, COST 50, 1982, p 237 32. J.A. Goebel, Advanced Coating Development for Industrial/Utility Gas Turbine Engines, Proc. First Conf. on Advanced Materials for Alternative Fuel Capable Directly Fired Heat Engines (Castine, ME), Department of Energy–Electric Power Research Institute, 1979, p 473 33. K.L. Luthra and J.H. Wood, High Chromium Cobalt-Base Coatings for Low Temperature Hot Corrosion, Thin Solid Films, Vol 119, 1984, p 271

pg 258

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p259-320 DOI: 10.1361/hcma2007p259

30/10/2007 4:01PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 10

Coal-Fired Boilers 10.1 Introduction A coal-ﬁred electric power generating station is a highly complex plant. A major piece of the equipment in the power plant is the boiler, which generates steam that is then delivered to the turbine for generation of electricity. This chapter reviews high-temperature corrosion problems associated with utility boilers. A brief description of boilers is included to help readers understand some of the basics in the boiler. A brief discussion of coal is also presented since different types of coal may present different issues in combustion, thus resulting in different types of materials issues and corrosion. More detailed discussion and coverage of boilers and the fuel can be found in Steam—Its Generation and Use (Babcock & Wilcox, Ref 1) and Combustion Fossil Power (Combustion Engineering, Ref 2).

10.2 Boiler Description The primary function of the boiler is to generate steam for the turbine to convert to electricity. The steam is converted from water in a furnace where heat is generated by combustion of coal. The modern furnace in a square or rectangular horizontal cross section is enclosed with water-cooled tubes with a narrow plate (commonly referred to as a “membrane”) connected between two adjacent tubes. When the furnace is heated, the water under high pressure inside these water-cooled tubes (also referred to as waterwalls) is converted to steam. Some boilers use a tangent tube design for waterwalls where no membranes are used to connect adjacent tubes. In terms of the operating pressure of the water/ steam in the waterwalls, there are two classiﬁcations of boilers: subcritical and supercritical units. The subcritical unit is the boiler that operates at pressures below the 218.2 atm

(3208 psig) critical point, while the supercritical unit operates at pressures higher than 218.2 atm (3208 psig). When heated at atmosphere pressure, water boils at 100 °C (212 °F). As the pressure of the water increases, the boiling (saturation) temperature also increases. The saturation temperature becomes constant once the water is boiling at a given pressure. Accordingly, the saturation temperature increases as the pressure increases. Thus, for subcritical units, the saturation temperature is established by a given operating waterwall water/steam pressure. For example, for the pressure of 170 atm (2500 psig), the saturated temperature is 353.4 °C (668.11 °F) (Ref 3). In a subcritical unit, the water (slightly undercooled) in the waterwall tubes rises from the bottom of the waterwall. As the water absorbs heat, bubbles form. More bubbles form as the water continues to rise in the waterwall tubes as shown in Figs. 10.1 and 10.2 (Ref 1). At the top of the waterwall tubes, a mixture of steam and water leaves the waterwall tubes and enters the steam drum where steam is separated to be heated further in the convection section for higher temperatures prior to delivery to the turbine for electricity generation. Water at the bottom of the steam drum is then mixed with the replacement water (feedwater) and is returned to the bottom of the boiler through an unheated downcomer to reenter the waterwall tubes. In a supercritical unit, water at pressures above 218.2 atm (3208 psig) does not boil. The temperature rises steadily, and the water changes completely to steam. There is no saturation temperature and no need for a steam drum. The supercritical unit operates on a once-through ﬂow system with water ﬂowing in the furnace waterwall tubes from the bottom of the furnace and changing to 100% steam with a continuous temperature increase. The steam then leaves the furnace waterwall tubes and is further heated in superheaters before entering into the steam turbines.

pg 259

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:01PM Plate # 0

260 / High-Temperature Corrosion and Materials Applications

There are several methods for burning coal. Firing with pulverized coal has been the most dominant method for utility boilers. Coal is burned as ﬁne powers suspended in the furnace, and almost all types of coal from anthracite to lignite can be burned by pulverized ﬁring (Ref 1). Pulverized coal in particle sizes of 50 µm diameter or smaller can be completely combusted in a matter of 1 to 2 s (Ref 1). Babcock & Wilcox developed the Cyclone furnace for ﬁring coal grades that have a low fusion temperature and are not suitable for pulverized coal ﬁring because of potentially forming a molten slag, thus developing a severe slagging problem in superheaters (Ref 1). In Cyclone boilers, the Cyclone barrels burn coal in such a way that most of the coal ash is captured to form a molten slag that coats the inside surface of the Cyclone barrels (Ref 1). The combustion ﬂue gas from the Cyclone barrels then enters the main furnace to generate steam. Stoker-ﬁring boilers are very versatile for burning a wide range of solid fuels including various types of coals, municipal waste, wood waste, and other types of biomass fuels. In a stoker-ﬁring boiler, a mechanical stoker at the

Fig. 10.1

Water enters the waterwall tubes at the furnace bottom and turns into a mixture of water and steam that leaves the waterwall tubes at the top (C) and enters the steam drum where steam and water is separated. Water mixed with the replacement water (A) is returned to the waterwall tubes at the furnace bottom (B). Source: Ref 1. Courtesy of Babcock & Wilcox

bottom of the furnace is used to feed fuel, such as coal, onto a grate where coal is burnt. Fluidizedbed combustion (FBC) technology is considered to be an emerging technology for power generation. In a FBC boiler, combustion of coal (or other type of fuel) takes place in a ﬂuidized bed. There are two types of bed: bubbling bed and circulating bed. The bed temperature is typically maintained in a range of 816 to 900 °C (1500 to 1650 °F) for maximum efﬁciency of both combustion and sulfur capture. The bed consists of essentially sand and limestone or dolomite. Air enters the bed from the air distributor plate at the bottom and renders the bed into a ﬂuidization condition. Limestone or dolomite is added to the bed as a sorbent to capture SO2 from the ﬂue gas by forming calcium sulfate (CaSO4). In a bubbling bed, an evaporator tube bundle is typically immersed in the bed to help maintain the designed bed temperature. Figure 10.3 shows a schematic of a bubbling ﬂuidized-bed boiler (Ref 1).

10.3 Coal and Coal Ash Coal is generally classiﬁed by rank, which indicates the geological formation history of the coal. There are four different types of coal. The youngest, or lowest-rank coal, is lignite, which is then followed by an older, or higher-rank subbituminous coal, then bituminous coal, and then anthracitic coal. Heating values, moisture content, volatile matter content, ash content, and sulfur content can be different among these different types of coals. Bituminous coal is the most commonly used coal for utility boilers in the United States (Ref 1). Subbituminous coal in the United States generally contains very low sulfur, with many deposits containing less than 1%. Table 10.1 shows the properties of some U.S. coal (Ref 1). There is a big variation in moisture, ash content, ash softening temperatures, and sulfur content from various grades of coals. The material factors associated with coal can directly or indirectly affect the boiler tube material performance at different locations in the boiler. Sulfur is the most important impurity in coal for causing high-temperature corrosion in the boiler. Sulfur is present in coal in forms of organic sulfur, pyritic sulfur (i.e., pyrite), and iron sulfate. High-sulfur coals also cause SOx emission problems and require expensive air pollution control equipment. As a result of the

pg 260

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:01PM Plate # 0

Chapter 10:

U.S. Federal Clean Air Act emissions issues, the low-sulfur, Powder River Basin (PRB) coal (a subbituminous coal) has become extremely popular for the past few years (Ref 4). The PRB coal, which is from mines in southern Montana through northern Wyoming, contains less than 1.2 lb of sulfur per million btu, making it compliant with Clean Air Act emissions limits without air pollution control equipment (Ref 4). Ash from coal after combustion can be entrained in ﬂue gas to cause ﬂy-ash erosion on heat-absorbing surfaces in the lower furnace, such as waterwalls, and in the convection pass, such as superheater, reheater, generating bank, and economizer. Ash can also deposit on the furnace wall, causing a slagging problem, and on superheater and reheater bundles, causing a fouling problem. Ash deposits can cause heat transfer problems, and they are required to be removed regularly using devices such as soot

Fig. 10.2

Coal-Fired Boilers / 261

blowers (with steam), waterlances, or water cannons (with room-temperature water), depending on the nature of the ash deposits. One of the important characteristics of ash is its fusion temperature. Table 10.1 lists the ash fusion temperatures of several coals under reducing conditions (Ref 1). These fusion temperatures can affect the nature of the ash deposits, whether in the form of “dust” or a tenacious slag. If ash reaches the heat-absorbing surface at a temperature near its softening temperature, the resulting deposits are likely to be porous and can be easily removed by sootblowing. Also, if such a deposit is subjected to high gas temperature, the ash deposit can reach its melting point (due to the thermal insulating properties of the ash) and run down the furnace wall surface (Ref 1). This solidiﬁed slag is tightly bonded and is difﬁcult to remove. This slag may require water lances or water cannons to create thermal shock for the

Same water and water/steam circulation in the furnace waterwall tubes as in Fig. 10.1 with illustration of the furnace waterwall tubes. Source: Ref 1. Courtesy of Babcock & Wilcox

pg 261

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:01PM Plate # 0

262 / High-Temperature Corrosion and Materials Applications

removal of this slag deposit. The furnace walls that are subject to radiant heat are likely locations for developing this slagging problem. The analyses of ash for several types of U.S. coal are shown in Table 10.1. The compositions shown in Table 10.1 are presented as oxides. However, most ash constituents in coal are minerals. Typical minerals found in coal are shown in Table 10.2 (Ref 2). When these minerals are exposed to oxidizing environments at appropriate high temperatures, oxides are likely to be the stable phases. Most of these oxides typically exhibit very high melting points. However, some minerals may themselves react at combustion temperatures to form reaction products, most likely complex salts, that may exhibit much lower melting temperatures. This is illustrated in Table 10.3 (Ref 2). If the ash constituents are present as oxides, the ratio of basic oxides

Distributor Plate

Fig. 10.3

(e.g., Fe2O3, CaO, MgO, Na2O, and K2O) to acidic oxides (e.g., SiO2, Al2O3, and TiO2) may determine the fusion (fusibility) temperature of the reaction product. It was reported that the ash may exhibit low fusibility temperature with higher slagging potential when its base/acid ratio is in a range of 0.4 to 0.7 (Ref 2). Many other parameters, such as SiO2/Al2O3 ratio, Fe2O3/ CaO ratio, Fe2O3/(CaO+MgO) ratio, (Na2O + K2O), and so forth are also used for predicting the fusibility temperature of the coal ash. It has been suggested that SiO2 is more likely than Al2O3 to form lower melting species (Ref 2). The fusibility temperature of coal ash will be lowered when the Fe2O3/CaO ratios are in a range of 0.2 to 10. Alkalies are important in affecting the fusibility of coal ash and the furnace slagging potential (Ref 2). Many sodium compounds melt at temperatures below 900 °C (1650 °F) (Ref 2).

Windbox

Bubbling fluidized-bed boiler. Source: Ref 1. Courtesy of Babcock & Wilcox

pg 262

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

Table 10.1

pg 263

Coal-Fired Boilers / 263

Properties of several types of U.S. coal Coal type

Property

Anthracite

Illinois #6 Bituminous Illinois

Spring Creek Subbituminous Wyoming

Bryan Lignite Texas

Proximate composition, % Moisture Volatile matter, dry Fixed carbon, dry Ash, dry

7.7 6.4 83.1 10.5

17.6 44.2 45.0 10.8

24.1 43.1 51.2 5.7

34.1 31.5 18.1 50.4

Composition, % Carbon Hydrogen Nitrogen Sulfur Oxygen Ash

83.7 1.9 0.9 0.7 2.3 10.5

69.0 4.9 1.0 4.3 10.0 10.8

70.3 5.0 1.0 0.4 17.7 5.7

33.8 3.3 0.4 1.0 11.1 50.4

Thermal properties Heating value (as received), Btu/lb

11,890

10,300

9,190

3,930

… … …

1930 2040 2700

2100 2160 2700

2370 2580 2900+

51.0 34.0 3.5 2.4 0.6 0.3 0.7 2.6 … 1.4

41.7 20.0 19.0 0.8 8.0 0.8 1.6 1.6 … 4.4

32.6 13.4 7.5 1.6 15.1 4.3 7.4 0.9 0.4 14.6

62.4 21.5 3.0 0.5 3.0 1.2 0.6 0.9 … 3.5

Ash fusion temperatures (reducing atmosphere), °F Initial deformation Softening Fluid Ash analysis(a), % SiO2 Al2O3 Fe2O3 TiO2 CaO MgO Na2O K 2O P 2O 5 SO3

(a) Elements present in the ash are determined and reported as oxides. Source: Ref 1

More detailed discussion about these parameters can be found in Ref 1 and 2. Chlorine content is also an important indication for fouling potential. When chlorine in coal is greater than 0.3%, fouling potential becomes high (Ref 2). Alkali metals and chlorine can also play a signiﬁcant role in high-temperature corrosion in boilers.

10.4 Combustion Environments Stoichiometric combustion is a complete combustion of a fuel with a theoretically calculated amount of oxygen in the combustion air. Complete combustion of all combustible constituents in coal requires excess air beyond the theoretically calculated value needed in the combustion air. Excess air is typically expressed in terms of percentage above the theoretically calculated value. For pulverized coal, typically 15 to 30% excess air is required to enssure adequate combustion (Ref 2). This will make the combustion environment an oxidizing atmosphere.

Table 10.2 coal Mineral species

Kaolinite Illite Biotite Orthoclase Albite Calcite Dolomite Siderite Pyrite Gypsum Quartz Hematite Magnetite Rutile Halite Sylvite

Typical mineral species found in Formula

Al2O3·2SiO2·H2O K2O·3Al2O3·6SiO2·2H2O K2O·MgO·Al2O3·3SiO2·H2O K2O·Al2O3·6SiO2 Na2O·Al2O3·6SiO2 CaCO3 CaCO3·MgCO3 FeCO3 FeS2 CaSO4·2H2O SiO2 Fe2O3 Fe3O4 TiO2 NaCl KCl

Source: Ref 2

Nitrogen oxides (NOx), which form during combustion under oxidizing conditions, are an undesirable pollutant from the boiler and contribute to acid rain and ozone formation. In order to reduce this NOx (i.e., NO and NO2) emission from the coal-ﬁred boiler, low NOx burner

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

pg 264

264 / High-Temperature Corrosion and Materials Applications

systems and staged combustion techniques have been developed and employed with great success in signiﬁcantly reducting NOx emission. The majority of NOx is formed by high-temperature oxidation of N2 in the combustion air under oxidizing conditions. Nitrogen in coal can also lead to formation of NOx during combustion. Reducing the availability of air during initial combustion can signiﬁcantly reduce the formation of NOx. This results in reducing environments. For staged combustion, the lower furnace is under reducing conditions, and additional air is then introduced at a higher elevation in the furnace to complete the combustion process. More detailed discussion of NOx control in combustion is available in Ref 1 and 2. As a result of low NOx combustion techniques, the furnace environment changes from an oxidizing to a reducing condition. The location for this change is variable and will ﬂuctuate as the load changes for the boiler operation especially under non-base-load operation. Sulfur in the coal forms H2S in reducing conditions instead of SO2 in oxidizing conditions. High wastage rates for waterwall tubes under reducing conditions and oxidizing/reducing alternating conditions can become a very serious materials issue. Ash released from coal during combustion can cause signiﬁcant problems to the behavior of the boiler components. During combustion, signiﬁcant amounts of ash are carried by the ﬂue gas stream and cause ﬂy-ash erosion problems particularly for superheaters and reheaters in the convection path. In addition, it can produce slagging on the furnace walls and fouling for superheaters and reheaters. The amount of ash being carried by the ﬂue gas stream depends on the type of ﬁring, for example, approximately 70 to 90% for pulverized coal units, approximately 40% for stoker-ﬁring units, and all the ash along with some ﬂuidized-bed material for circulating ﬂuidized-bed boilers (Ref 1). During combustion, the gas temperature can reach a range of 1370 to 1650 °C (2500 to Table 10.3

3000 °F), some of the mineral constituents and compounds can be in a molten or plastic state (see Tables 10.1 and 10.3). These ash constituents can then deposit on the heat-absorbing surfaces in the boiler. Two general types of ash deposition, slagging and fouling, can take place. Slagging is the deposition of molten, partially fused deposits on the furnace walls and the upper furnace radiant superheaters exposed to radiant heat (Ref 1). Fouling is the deposition of more loosely bonded deposits on the heat-absorbing surfaces in the convection pass, such as superheater and reheater, that are not exposed to radiant heat (Ref 1). These ash deposits, which are impeding the heat transfer, require their regular removal from the tube surfaces. Soot blowers using steam are generally adequate for removing the ash deposits on the fouled tube surfaces, while waterlances or water cannons using thermal shock by water may be needed to clean the slagged tube surfaces. Material behavior under the inﬂuence of ash deposits is discussed in later sections.

10.5 Fireside Corrosion and Other Materials Problems in Boilers This section discusses the materials problems related to the heat-absorbing surfaces in the furnace combustion area (i.e., waterwalls) and also in the convection pass, such as superheaters and reheaters. A schematic of a coal-ﬁred boiler showing these two areas is shown in Fig. 10.4. Accelerated tube wall wastage attack resulting from the change of the combustion environment from oxidizing to reducing due to installation of low-NOx combustion technology intended to reduce NOx emissions is discussed. Also discussed in detail are circumferential cracking of the furnace waterwalls, soot blower erosion/corrosion, thermal fatigue cracking induced by waterlances and water cannons for deslagging the furnace

Melting points of coal ash constituents

Element

Oxide

Melting point, °C (°F)

Compound

Melting point, °C (°F)

Si Al Ti Fe Ca Mg Na K

SiO2 Al2O3 TiO2 Fe2O3 CaO MgO Na2O K2O

1715 (3120) 2043 (3710) 1838 (3340) 1565 (2850) 2521 (4570) 2799 (5070) Sublimes at 1277 (2330) Decomposes at 349 (660)

Na2SiO3 K2SiO3 Al2O3·Na2O·6SiO2 Al2O3·K2O·6SiO2 FeSiO3 CaO·Fe2O3 CaO·MgO·2SiO2 CaSiO3

877 (1610) 977 (1790) 1099 (2010) 1149 (2100) 1143 (2090) 1249 (2280) 1391 (2535) 1540 (2804)

Source: Ref 2

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

waterwalls, and superheater/reheater corrosion. Limited discussion includes erosion issues in ﬂuidized-bed boilers. Hydrogen attack encountered in subcritical units as a result of water corrosion in the internal diameter of the waterwall tubes is discussed in “Hydrogen Attack” in Chapter 17. 10.5.1 Fireside Corrosion of Furnace Waterwalls The furnace is enclosed by four walls (or waterwalls) with front, rear, and two side walls in a square or rectangular form for the horizontal cross section of the furnace. The waterwall consists of tubes with membranes (i.e., steel plates) connecting adjacent tubes (a tube-membranetube construction), or of tubes without membranes connecting adjacent tubes (a tangent tube construction). These waterwalls, which enclose the furnace combustion zone, are heat-absorbing surfaces that heat the rising water in the tubes (from bottom) to become steam. The gas temperature in the combustion inside the furnace can reach a range of 1370 to 1650 °C (2500 to 3000 °F). As a result, these waterwalls, particularly the crown location of the tube on the ﬁreside

Fig. 10.4 Source: Ref 5

Coal-Fired Boilers / 265

(i.e., the side facing the combustion), are subject to extreme radiant heat from the combustion. The tube wall exhibits temperature gradient across the tube wall thickness. The outer metal temperature depends on the temperature of the water and/or steam inside the tube along with many other factors. For subcritical units of drum-type boilers, the temperature of water/steam depends on the pressure. For example, for the pressure of 170 atm (2500 psig), the saturated temperature is 353.4 ° C (668.11 °F) (Ref 3). In a supercritical unit with pressures above 218.2 atm (3208 psig), the water inside the tube changes to 100% steam with a continuous temperature increase. The tube metal temperature for the supercritical unit is generally higher than that of a subcritical unit. Other factors that can affect tube metal temperature include oxide scales formed on the tube inside diameter (ID) surface, slag deposits on the tube outside diameter (OD) surface, ﬂame impingement, and so forth. Typical temperature range for the waterwall tube on the ﬁreside has been given as 380 to 450 °C (716–842 °F) by Lees and Whitehead (Ref 6), below about 450 °C (842 °F) by Hay (Ref 7), 427 to 482 °C (800 to 900 °F) by Cunningham and Webster (Ref 8), and 400 to 500 °C (752 to 932 °F) by Stringer

Coal-fired boiler, showing two main areas for discussion of materials problems in a boiler: the furnace combustion area (i.e., furnace walls) and the heat-absorbing surfaces in the convection path, such as superheaters and reheaters.

pg 265

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

266 / High-Temperature Corrosion and Materials Applications

(Ref 9). Eurich and Kramer (Ref 10) indicated that the design waterwall tube surface metal temperature was between 482 and 510 °C (900 and 950 °F) for each of the ﬁve 676 MW(e) opposed-ﬁred supercritical boilers. Yeager (Ref 11) indicated that the design waterwall surface temperature for the two supercritical, tangentially ﬁred boilers (each with 750 MWe) at Montour Station was between 482 and 538 °C (900 and 1000 °F). Figure 10.5 (Ref 12) shows a schematic of the temperature gradient across the tube wall and different oxide layers and ash/slag deposits. Wastage. Excess air is normally used for combustion to ensure adequate combustion of coal. For example, typically 15 to 30% excess air is required for ﬁring pulverized coal (Ref 2). This produces an oxidizing condition with major combustion gaseous products being N2, O2, CO2, and H2O. Under this normal oxidizing condition, waterwalls made of carbon steels and low-alloy steels have been found to show metal wastage rates of approximately 40 nm/h (13 mpy) (Ref 12). A typical tube wastage proﬁle for the corroded waterwall tube is illustrated in Fig. 10.6 (Ref 13). The maximum wastage typically occurred at the crown location (12 o’clock position). Figure 10.7 shows a cross section of a carbon steel waterwall tube removed from a subcritical unit after 21 years of service with an observed wastage rate of about 0.19 mm/y (7.5 mpy) (Ref 14). The front face of the tube at the crown location facing the combustion radiant

heat suffers the worst metal wastage. One boiler designer suggests the maximum tube metal temperature limits based on oxidation for typical ferritic steels used for waterwalls, as shown in Table 10.4 (Ref 2). For example, carbon steel is limited to 454 °C (850 °F), but without mentioning the design service life. However, due to

Fig. 10.6

Fig. 10.7

Fig. 10.5

Temperature gradients through the inner oxide scale, tube wall, outer oxide scale, and ash/slag deposits. Source: Ref 12

Typical wastage profile of a corroded waterwall tube. Source: Ref 13

Cross section of a carbon steel waterwall tube from a subcritical unit in the United States after 21 years of service, showing the maximum wastage at the crown location with about 1.9 mm/yr (7.5 mpy) of wastage rate. Source: Ref 14. Courtesy of Welding Services Inc.

pg 266

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

the complexity of the combustion in a large coalﬁred boiler it is almost impossible to establish a normal oxidizing condition uniformly in the ﬂue gas throughout the entire furnace cross section in a dynamic condition. As a result, reducing conditions are developed in some localized areas near the waterwall. Chemical analysis of ash deposits formed on tube surfaces revealed an appreciable amount of free carbon (Ref 15). The presence of free carbon indicates the establishment of reducing conditions at the furnace wall surface (Ref 16). Localized reducing conditions with up to 10% CO in the furnace atmosphere in the vicinity of the furnace walls have been observed (Ref 17, 18). Sulﬁdation. One of the most corrosive impurities in coals is sulfur, which is present in coals as pyrite (FeS) and organic sulfur. Under an Table 10.4 Maximum outer tube metal temperature limits based on oxidation for common ferritic steels used in waterwalls, as suggested by a boiler designer Steel designation

Type

SA210 A1 T1 T11 T22

Carbon Steel Carbon-0.5Mo 1.25Cr-0.5Mo 2.25Cr-1Mo

Source: Ref 2

Fig. 10.8

Oxidation limit, °C (°F)

454 (850) 482 (900) 552 (1025) 593 (1100)

Coal-Fired Boilers / 267

oxidizing condition, sulfur oxidizes to form SO2 during combustion. The presence of SO2 in an oxidizing combustion atmosphere does not generally cause accelerated corrosion. However, under reducing conditions, sulfur reacts to form H2S. This can signiﬁcantly increase corrosion rates by sulﬁdation. Lees and Whitehead (Ref 6) examined several corroded carbon steel waterwall tube samples as well as Type 310 coextruded tube samples removed from several power plants in the United Kingdom that suffered severe wastage problems. They had observed sulﬁdation in some severely corroded carbon steel tubes in addition to a chlorine-rich phase at the boundary between the oxide/sulﬁde phases and the metal. (Chlorine issues are discussed next in this section). In one carbon steel sample that had suffered a wastage rate of about 530 nm/h (170 mpy), they found the corrosion product to be predominantly sulﬁdes. They even observed a substantial layer of oxides/sulﬁdes on the corroded Type 310 cladding in a Type 310/CS composite tube that suffered more than 50 nm/h (16 mpy) (Ref 6). Figure 10.8 shows a waterwall carbon steel tube after only 1 year of service in a subcritical unit in the United States, showing pitting attack. The boiler was in the United States, burning a coal containing about 3.0 to 3.5% S and about 300 to 400 ppm chlorine (Ref 14). Metallurgical examination of the

Close-up view of a waterwall carbon steel tube showing pitting attack after 1 year of service in a subcritical unit in the United States, burning coal containing about 3.0 to 3.5% S and about 300 to 400 ppm chlorine. Source: Ref 14. Courtesy of Welding Services Inc.

pg 267

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

pg 268

268 / High-Temperature Corrosion and Materials Applications

sample indicated that pitting was associated with sulﬁdation attack, as shown in Fig. 10.9, which shows initiation of internal sulﬁdation attack (Ref 14). The tube was in the initiation stage of the accelerated corrosion by sulﬁdation. There were also areas in the same tube that showed only oxidation attack, as illustrated in Fig. 10.10 (Ref 14). Sulﬁdation of the waterwall tubes in the lower furnace causes a serious wastage issue when the furnace is ﬁred intentionally under reducing conditions (i.e., substoichiometric conditions) when retroﬁtted with low-NOx burners to reduce the formation of NOx. This subject is discussed in section 10.5.2. Chlorine. Another important impurity in coal that can signiﬁcantly affect the corrosion of metals in boilers is chlorine. Most coals from the U.S. contain low chlorine (<0.1% Cl) (Ref 19). Thus, there is very little experience and few studies have been conducted on the effects of high-chlorine coals on corrosion of metals in U.S. boilers. British coals contain much higher chlorine. The average chlorine content in coals used in British power plants is about 0.25% with some as much as 0.8% (Ref 20). Chlorine in coal

converts mostly to HCl during combustion. It is believed that every 0.1% Cl in coal produces approximately 80 ppm HCl in the ﬂue gas (Ref 21). Clarke and Morris (Ref 18) performed gas analysis on the gas samples obtained from the vicinity of the furnace wall at various locations in a 120 MWe boiler that historically suffered high wastage rates (excess of 150 nm/h, or 48 mpy, or 1.2 mm/yr). This boiler burned high-chlorine coal containing about 0.4 to 0.6% Cl. The authors observed that reducing conditions occurred in the areas of high wastage, while oxidizing conditions were observed in the areas that were outside of these high-wastage areas. In addition, the concentration of HCl was found to be about 400 ppm in the high-wastage areas. The level of H2S was found to be about 300 to 400 ppm only when CO concentration exceeded 3%, but in some areas no H2S was observed under those reducing conditions. In laboratory tests in a simulated combustion gas environment

19 µm

Fig. 10.9

Fig. 10.10

1: 1.1% S, 0.7% Al, 0.8% Si, 0.6% Mn, 95% Fe, and trace elements 2: 0.8% S, 0.3% Al, 0.6% Si, 0.5% Cl, 0.5% Ca, 1.9% Zn, 94% Fe, and trace elements 3: 13.4% S, 0.8% Al, 0.4% Si, 0.8% Mn, 84.1% Fe, and trace elements 4: 9.2% S, 0.3% Al, 0.4% Si, 0.6% Mn, 89.3% Fe, and trace elements 5: 1.7% S, 0.8% Al, 1.0% Si, 1.2% Cl, 0.2% Ca, 0.7% Mn, 91.8% Fe, and trace elements

1: 8.9% Si, 1.5% Al, 86.9% Fe, and trace elements 2: 44.3% Si, 22.9% Al, 2.2% Mg, 6.7% Ca, 5.6% K, 14.7% Fe, and trace elements 3: 1.7% Si, 1.0% Al, 1.0% S, 1.5% Zn, 93.7% Fe, and trace elements 4: 1.3% Si, 1.0% Cu, 1.3% Zn, 1.2% S, 91.7% Fe, and trace elements 5: 2.9% Cu, 93.8% Fe, and trace elementsa 6: 1.0% Si, 1.3% Zn, 93.1% Fe, and trace elements

Scanning electron microscopy backscattered electrons image of the corrosion products showing initiation of sulfidation attack on the tube, where pitting attack was observed on the tube surface as shown in Fig. 10.8 Chemical compositions of the phases of the corrosion products at different locations were analyzed by energy dispersive x-ray spectroscopy (EDX) with the results summarized as:

Scanning electron microscopy backscattered electrons image of the corrosion products showing ash deposits and iron oxides with no evidence of sulfidation attack on other area of the tube that did not suffer pitting attack (Fig. 10.8). Chemical compositions of the phases of the corrosion products at different locations were analyzed by energy dispersive x-ray spectroscopy (EDX) with the results summarized as:

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

However, after 27,000 h, both co-extruded tube alloys showed a maximum corrosion rate of about 60 nm/h (19 mpy). Both Type 304 and 310 alloys exhibited signiﬁcant performance improvement compared with carbon steel. Examination of corroded 304 and 310 samples revealed substantial amounts of sulﬁdes along with some chlorine. The effect of chlorine in coals on the furnace wall corrosion was summarized by Daniel et al. (Ref 19) based on CEGB data generated in boilers in the United Kingdom. The effect is illustrated in Fig. 10.13 and 10.14. Coals with more than 0.2% Cl may result in accelerated ﬁreside

Fig. 10.12

Metal loss data for carbon steel in a simulated combustion atmosphere consisting of N2-10CO10H2O-0.5SO2 containing different levels of HCl (0, 400, and 2000 ppm) at 500 °C. Source: Ref 22

Corrosion rate, mpy

Metal loss, µm

(N2-10%CO-10%H2O-0.5%SO2), Brooks and Meadowcroft (Ref 22) showed that carbon steel corroded at linear rates of about 130 and 550 nm/h (42 and 176 mpy) at 400 and 500 °C (752 and 932 °F), respectively, with 400 ppm HCl added to the test environment. Their test data are shown in Fig. 10.11 and 10.12. The laboratory 400 °C (752 °F) test data of Brooks and Meadowcroft (Ref 22) and the ﬁeld data of Clarke and Morris obtained from a boiler (Ref 18) were in the same order of magnitude. Figures 10.10 and 10.11 also show that the corrosion of carbon steel followed a parabolic reaction rate when the environment contained no chlorine. In the examination of waterwall tube samples from several plants by Lees and Whitehead (Ref 6), corrosion rates were found to vary from about 200 to 530 nm/h (64 to 170 mpy) in different plants. The corrosion products the authors observed were mainly sulﬁdes with a chlorinerich phase along the metal/scale interface for some samples. Other samples contained mainly sulﬁdes with no chlorine-rich phases. If iron chlorides formed at the metal/scale interface, the phases that were highly volatile with high vapor pressures could easily escape into the environment (see Chapter 6). In a ﬁeld testing of coextruded tubes with Type 304 and Type 310 stainless steel claddings in a Central Electricity Generating Board (CEGB) boiler (500 MW) that had suffered aggressive furnace wall corrosion problems, Flatley et al. (Ref 23) found that the corrosion rates were about 77 nm/h (25 mpy) for Type 304, 55 nm/h (17 mpy) for Type 310, and 180 to 250 nm/h (58 to 80 mpy) for carbon steel after 15,000 h of exposure.

Coal-Fired Boilers / 269

10

Fig. 10.11

Metal loss data for carbon steel in a simulated combustion atmosphere consisting of N2-10CO10H2O-0.5SO2 containing different levels of HCl (0, 400, and 2000 ppm) at 400 °C. Source: Ref 22

Fig. 10.13

Effect of chlorine in coals on the corrosion rate of carbon steel and low-alloy steels under reducing conditions, based on CEGB laboratory data (Ref 16). Source: Ref 19

pg 269

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

270 / High-Temperature Corrosion and Materials Applications

corrosion for furnace wall tubes under reducing conditions (Ref 19). In 1997, James and Pinder (Ref 24) provided a different perspective on the effect of coal chlorine on the furnace wall corrosion. They indicated that one source of data that contributed to the correlation of furnace wall tube corrosion and the chlorine concentrations in coals was obtained from a power station with ﬁve front-ﬁred boilers (each with 120 MW). Between 1967 and 1977, the chlorine content in coals for this station in United Kingdom was increased from 0.35 to 0.65%. The maximum corrosion rates over this period showed an increase from 170 to 550 nm/h (58.5 to 189.2 mpy) when chlorine was increased from 0.35 to 0.65%, independent of sulfur content and load factor (Ref 24). James and Pinder summarize the data in Fig. 10.15 (Ref 24). Also included are data from the other plant. The data show a strong dependence of furnace wall corrosion on the chlorine content in coal. Figure 10.16 shows the data from many plants plotted together (Ref 25), showing a signiﬁcant scattering. This may not be unexpected, since there were so many plants that were probably of different

Fig. 10.14

boiler designs and were under various operating parameters, along with many other variables that were involved. The correlation would be best established when the data were obtained from the same boiler under constant operating parameters with the chlorine content in the coal as the only variable. 10.5.2 Fireside Corrosion of Furnace Walls under Low NOx Combustion Conditions In order to comply with the Clean Air Act Amendments of 1990, power plants are required to reduce NOx emissions from boilers. Nitrogen oxides (NOx) are produced during combustion under oxidizing conditions, and they are undesirable pollutants from the boiler because they contribute to acid rain and ozone formation. In order to reduce NOx emissions, staged ﬁring is used to produce a substoichiometric combustion (i.e., combustion with insufﬁcient oxygen or a reducing condition) in the lower furnace, which is followed by introduction of adequate air at the overﬁre air ports at a higher elevation to complete the combustion process. As a result, the lower

Corrosion rate as a function of temperature for carbon steel and austenitic stainless steels obtained from laboratory testing as well as field testing at Eggborough Power Station. Source: Ref 19

pg 270

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

content in the alloy produces a more protective chromium oxide scale with less probability of forming sulﬁdes. Similar results (Table 10.6) were reported by Gilroy (Ref 27) using a simulated test environment with much lower H2S level, showing that Type 310, Type 446, alloy 671 and chromized coatings were signiﬁcantly more resistant to sulﬁdation than carbon steel. The simulated environment used by Gilroy contained about 600 ppm HCl. Where chloridation attack was involved in this testing, chromium was also effective in increasing the alloy’s resistance to chloridation attack. Kung and Eckhart (Ref 28) conducted tests in H2S-containing environments for carbon steel, low-alloy steel, and several stainless steels. Figure 10.17 shows the corrosion rates for carbon steel and 2.25Cr-1Mo steel tested at 370 °C (700 °F) in N2-5.1CO-16.7CO24.6H2O-0.55H2 containing several levels of H2S (0.05, 0.25, 0.5, 4.8% H2S) for 1000 h. Similar tests were conducted for Types 304L and 310 (Fig. 10.18). Both 304L and 310 showed signiﬁcant reduction in corrosion rates compared with carbon and 2.25Cr-1Mo steels. At 480 °C (900 °F) in the gas mixture with 0.05% H2S, the corrosion rate was about 0.05 mm/yr (2 mpy) for

Corrosion rate (max), nm/h

furnace is under reducing conditions and/or shifting alternatively from oxidizing to reducing conditions. Under reducing conditions, sulfur from coal forms H2S instead of SO2. H2S can cause severe sulﬁdation attack on carbon steel furnace walls. Chou and Daniel (Ref 26) conducted a simulated test with a relatively high level of H2S for a variety of materials. Their results are summarized in Table 10.5 (Ref 26). Carbon and lowalloy steels were found to suffer corrosion rates in a 1 to 1.3 mm/yr (40 to 50 mpy) range. Type 304, which is signiﬁcantly more resistant than carbon and low-alloy steels, showed corrosion rates of 0.2 to 0.3 mm/yr (8 to 12 mpy). This is because the corrosion products were changed from iron oxides/sulﬁdes on steels to chromium oxides (or possibly with chromium sulﬁdes) on Type 304. Both chromium oxides and/or chromium sulﬁdes exhibit much slower growth rates than iron oxides/sulﬁdes; thus much less metal is consumed and accordingly metal wastage rates are much lower. Type 309 showed a very low corrosion rate. Alloy 671 and chromized coatings showed extremely low corrosion rates. Increases in corrosion resistance are mainly related to the chromium content in the alloy. Higher chromium

Coal-Fired Boilers / 271

Weighted mean chlorine content, %

Fig. 10.15

Corrosion rates of furnace walls (carbon steel) as a function of chlorine content in coal from boilers in two plants. The equation of the line is Rc =1380 (%Cl)-208, where Rc is corrosion rate (nm/h). Source: Ref 24 from original data (Ref 25)

pg 271

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

pg 272

Corrosion rate, nm/h

272 / High-Temperature Corrosion and Materials Applications

Weighted mean coal chlorine content

Fig. 10.16

Corrosion rates of furnace walls (carbon steel) as a function of chlorine content in coal from boilers in many plants. Source: Ref 24 from original data (Ref 25)

Table 10.5 Corrosion of various materials in N2-5CO-16CO2-10H2O-0.5H2-2H2S at 482 °C (900 °F) for 4000 h Alloy

Carbon steel 2.25Cr-1Mo Type 304 Type 304L Type 309 Alloy 800 Alloy 671 Chromized carbon steel Chromized carbon steel Chromized 2.25Cr-1Mo

Corrosion rate, mm/yr (mpy)

1.04 (41) 1.32 (52) 0.2 (8.2) 0.3 (12.0) 0.04 (1.6) 0.33 (13.0) 0.005 (0.18) 0.006 (0.25) 0.008 (0.32) 0.007 (0.28)

Source: Ref 26

Type 310 and about 0.2 mm/yr (8 mpy) for Type 304L (Ref 28). Dooley et al. (Ref 29) indicated that most boilers experienced waterwall wastage rates of 29 nm/h (10 mpy) or less prior to installation of low NOx burners. The wastage rates have been dramatically increased to a range of 145 to 350 nm/h (50 to 120 mpy) for some boilers after installing low NOx burners (Ref 29). In some boilers, waterwall wastage rates, which were in a range of 49 to 87 nm/h (17 to 30 mpy), were found to increase to a range of 96 to 226 nm/h (33 to 78 mpy) after retroﬁtting with low-NOx burners (Ref 30). It was also reported that the corrosion scales were found to consist of iron sulﬁdes and oxides. Furthermore, the ﬂue gas in

Table 10.6 Weight gain (mg/cm2) of several alloys after testing in N2-10CO-5CO2-10H2O0.1H2S-600 ppm HCl at 400 and 500 °C (752 and 932 °F) for 3000 h Alloy

Carbon steel 50Ni/50Cr, sprayed coating Aluminized (A) Aluminized (B) Fe-27Cr-6Al-2Mo FAL (Fe-13Cr-4Al) Ferralium (Fe-25Cr-5Ni-4Mo) 44-LN (Fe-26Cr-5Ni-1.5Mo) Monit (Fe-25Cr-5Ni-4Mo) 29-4-C (Fe-28Cr-4Mo) 29-4-2 (Fe-29Cr-4Mo-2Ni) E Brite (Fe-26Cr-1Mo) Fecralloy (Fe-16Cr-5Al-0.35Y) Fecralloy (Fe-19Cr-5Al-0.32Y) Fecralloy (Fe-20Cr-5Al-0.34Y) Fecralloy A (Fe-16Cr-4.5Al-0.26Al) GE2541 (Fe-26Cr-5Al-0.45Y) Type 310 Type 310Nb (0.8Nb) Alloy 800H Type 446 Alloy 671 Chromized 2.25Cr-1Mo Chromized carbon steel

400 °C (752 °F)

25.0 … 8.0 5.5 … 3.1 0.7 0.8 0.3 0.2 0.2 0.7 0.6 0.7 0.7 0.2 0.6 0.7 0.4 0.3 (1500 h exposure) 0.5 0.6 0.3 0.2

500 °C (932 °F)

90.0 9.5 … 8.8 3.8 (2000 h exposure) 0.1 2.1 0.5 0.7 1.1 0.9 0.8 0.6 0.7 0.4 0.2 0.4 0.6 0.3 0.3 (1500 h exposure) 0.3 0.5 0.2 0.3

Source: Ref 27

the vicinity of the furnace walls that suffered accelerated wastage rates was observed to have levels of CO greater than 5% and of H2S up to

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

Coal-Fired Boilers / 273

500 ppm (Ref 30). Also observed in the high wastage areas were carbon-rich deposits and pyrites. Urich and Kramer (Ref 10) reported that four supercritical boilers (each with 676 MWe, opposed-ﬁred) at Gibson Generating Station were retroﬁtted with low NOx burners and overﬁre air. Prior to the installation of low NOx burners, tube wastage problem was only limited to a small area on the side walls at locations where H2S concentrations were measured to be as high as 1000 ppm. Since the retroﬁt, a marked increase in waterwall tube wastage was observed for all four boilers. Severe corrosion was found to occur on the side walls, with the most severe corrosion occurring in the top burner elevation up to the level of the overﬁre air ports (Ref 10). In some cases, wastage rates were on the order of 285 nm/h (100 mpy). Furthermore, the corrosion scales in these areas were mainly iron sulﬁdes. Gas sampling from the side wall in one of the boilers showed the levels of CO as high as 11 or 12% and those of H2S as high as 1000 ppm (Ref 10). It was reported that the service life of the waterwall panels has been reduced from 12 to 15 years to 4 years after installation of the low NOx burners in these boilers at Gibson Station (Ref 31). Two supercritical units with tangentially ﬁring (582 MWe each) were also reported to have a more serious waterwall wastage issue after installation of low NOx burners with overﬁre air (Ref 31). Yeager (Ref 11) reported that a supercritical boiler, Unit No. 2, at Montour Station was retroﬁtted with low NOx burners in 1994. In

1996, an inspection revealed approximately 370 m2 (4000 ft2) of the waterwall had suffered severe wastage with maximum wastage rates being in excess of 2.2 mm/yr (85 mpy). The boiler, a tangentially ﬁred unit (750 MWe), burned Eastern bituminous coal with 1.9 to 2.2% S, had a waterwall wastage problem prior to installation of low NOx burners with the average wastage rates in the range of 0.25 to 0.38 mm/yr (10 to 15 mpy). Yeager indicated that the ﬁreball had tendency to be closer to certain burner corners, thus causing higher wastage rates. The furnace wall gas sampling tests revealed high total reduced sulfurs (TRS) and CO levels at the burner corners that suffered high wastage rates and an oxidizing atmosphere at the burner corners that showed little wastage (Ref 11). Clark and Morris (Ref 18) reported that a level of H2S in a range of 300 to 400 ppm and of CO in a range of greater than 3.5% were observed in the vicinity of the furnace walls with accelerated wastage rates in U.K. boilers. The level of H2S observed in U.K. boilers was similar to what was observed in U.S. boilers. The areas that suffer severe waterwall wastage can be highly dependent on boiler design and ﬁring conﬁguration, among other factors. James and Pinder (Ref 32) indicated that the side and/or rear walls were most vulnerable in a front-ﬁred boiler, while in a tangentially ﬁred boiler, the wastage was invariably highest on the front wall. Bakker and Kung (Ref 33) observed that FeS deposits can signiﬁcantly increase the corrosion rates of carbon and low-alloy steels. They found that FeS deposits showed no effects in reducing environments, but accelerated the corrosion in alternating oxidizing and reducing environments or oxidizing environments. This is shown in

Fig. 10.17

Fig. 10.18

Corrosion rates of carbon steel and 2.25Cr-1Mo steel (T-22) as a function of H2S in the N2-5.1CO16.7CO2-4.6H2O-0.55H2 gas mixture at 370 °C (700 °F) for 1000 h. Source: Ref 28

Corrosion rates of Types 304L and 310 as a function of H2S in the N2-5.1CO-16.7CO24.6H2O-0.55H2 gas mixture at 370 °C (700 °F) for 1000 h. Source: Ref 28

pg 273

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

274 / High-Temperature Corrosion and Materials Applications

Tables 10.7 and 10.8. The authors believe that FeS under oxidizing conditions can produce volatile reduced sulfur species including sulfur vapor, H2S, COS, and so forth, thus making the environment more aggressive and causing accelerated wastage rates. Since FeS is stable only under reducing conditions and causes accelerated corrosion to carbon and low-alloy steels only under oxidizing conditions, the waterwall area that is likely to suffer accelerated wastage rates appears to be subject to alternating reducing and oxidizing conditions. These alternating reducing and oxidizing conditions may exist in staged ﬁring boilers where the overﬁre air mixes with substoichiometric ﬂue gas from the burner zone (Ref 33). In one boiler where two locations at the waterwall made of T2 steel were monitored for their wastage rates as well as the ash deposits, the maximum wastage rate of 1.5 mm/yr (60 mpy) at one location and of 0.37 mm/yr (15 mpy) at the other location were observed after 12,633 h of service (Ref 34). The ﬂue gas measurements near the waterwall at those two locations showed highly reducing conditions with about 7% CO and 500 ppm H2S during full-load operation and slightly oxidizing conditions during low-load conditions. However, there are signiﬁcant differences in the ash deposits between the two locations; the area suffering high wastage rates had an ash deposit consisting of 90% FeS and 10% ﬂy ash, while the low wastage rate area had an ash deposit of mainly Fe3O4 and ﬂy ash (Ref 34). This ﬁeld experience was consistent with the laboratory test results observed by Bakker and Kung (Ref 33). In the same boiler as discussed previously in this paragraph, another panel testing was conducted after 17,155 h of exposure, the FeS content in the ash deposits was in the 40 to 90% range, but with chlorine content in the range of 500 to 1800 ppm. The maximum wastage rate in the area was about 3.5 mm/yr (140 mpy), a signiﬁcantly higher wastage rate than the earlier test panel with about 90% FeS

Table 10.7 Corrosion rates of T2 in N2-7CO-12CO2-6H2O-0.13H2S-0.008HCl with various ash deposits at 427 °C (800 °F) for 600 h Ash deposit

No deposit 30% FeS, 70% ﬂy ash 60% FeS, 40% ﬂy ash 60% FeS, 20% carbon, 20% ﬂy ash 90% FeS, 10% ﬂy ash Source: Ref 33

Corrosion rate, mm/yr (mpy)

0.42–0.57 (17–23) 0.1–0.3 (5–12) 0.1–0.17 (5–7) 0.2–0.6 (9–22) 0.1–0.27 (5–11)

and only 90 ppm chlorine in the ash deposits (maximum wastage rate of 1.5 mm/yr or 60 mpy) (Ref 34). Higher chlorine content in the ash deposits might play a role in causing higher wastage rates (Ref 34). Bakker (Ref 34) reported preliminary results of the laboratory test involving the effect of the KCl-NaCl (equal amounts) in the ash deposits of char, FeS, and ﬂy ash. The test involved ﬂue gas temperature of 1000 °C (1830 °F), maximum ash deposit temperature of 650 °C (1200 °F), metal tube sample temperature of 450 to 470 °C (850 to 900 °F), and the temperature of the cooling air (inside the tube sample to produce the temperature gradient) of 320 °C (610 °F) at the exit end. The metal loss was found to be about 20 to 22 µm in 100 h when the ash deposits contained no chloride and about 44 to 50 µm in 100 h when 0.2% alkali chloride (KCl-NaCl in equal amounts) was added to the ash deposits. Lees and Whitehead (Ref 6) observed sulﬁdes as well as a chlorine-containing phase at the sulﬁde/metal interface when they examined the corroded carbon steel waterwall samples using SEM/EDX analyses. Similar observation was also found by Lai (Ref 14). Figure 10.19 shows a corroded carbon steel waterwall sample from a boiler ﬁred with low NOx burners (Ref 14). The boiler, a subcritical unit, had been burning Western Kentucky bituminous coal containing about 3.5% S. The waterwalls started to experience accelerated wastage rates after the boiler began stage ﬁring with low NOx burners and overﬁre air. SEM/EDX analysis of the corrosion products formed on the sample revealed iron sulﬁde phases and a chlorine-containing phase at the sulﬁde/metal interface, as shown in Fig. 10.20 (Ref 14). Coating and Weld Overlay. As more boilers experienced accelerated wastage rates for carbon and low-alloy steels at waterwalls in boilers that had been retroﬁtted with low-NOx burners, Table 10.8 Corrosion rates of T2 in alternating reducing(a) and oxidizing environments(b) with various ash deposits at 427 °C (800 °F) for 400 h Ash deposit

No deposit 30% FeS, 70% ﬂy ash 60% FeS, 40% ﬂy ash 60% FeS, 20% carbon, 20% ﬂy ash 90% FeS, 10% ﬂy ash 50% carbon, 50% ﬂy ash

Corrosion rate, mm/yr (mpy)

0.37–0.5 (15–20) 1.0 (41) 0.82 (33) 1.2 (49) 1.2 (49) 0.5 (20)

(a) 16 h in N2-7CO-12CO2-6H2O-0.13H2S-0.008HCl. (b) 8 h in N2-1O2-17CO20.13SO2-0.008HCl. Source: Ref 33

pg 274

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

industry looked for a viable coating or weld overlay to extend the waterwall life. Bakker et al. (Ref 35) reported the results of in-plant tests of thermal sprayed coatings (arc sprayed coatings and HVOF coatings), diffusion coatings (chromized coating and Cr/Si diffused coating), and weld overlays (stainless steels 410, 309, and 312, and alloy 625). Waterwall test panels, which consisted of uncoated T2 tubes, coated and weld overlay tubes, were tested in No. 2 boiler at Hatﬁeld’s Ferry Station. The boiler, a supercritical unit, was retroﬁtted with a low NOx cell burner system. Test panels were installed at North side wall at the burner elevation. Side walls typically suffered the worst corrosion attack for an opposed wall ﬁring boiler such as this one. During testing, the boiler burned Eastern bituminous coals with about 2.2% S with no measurements or control of chlorine content. However, the Pennsylvania coals typically contained about 0.05 to 0.15% Cl. The test results generated by Bakker et al. (Ref 35) are summarized:

After 12,683 h of exposure, T2 waterwall tubes were found to suffer a wastage rate of about 1.6 mm/yr (62 mpy) at the burner elevation. The arc sprayed coating of Ni44Cr alloy wire suffered minor spallation. The high-velocity oxyfuel (HVOF) coating

Coal-Fired Boilers / 275

of Ni-42Cr-2Si alloy also suffered minor spallation. The arc sprayed coating of Ni45Cr-2Al was completely spalled off from the tube. Metallurgical evaluation of all the coated tubes showed some sulfur penetration and corrosion attack at the coating/metal interface. Furthermore, a thickness loss of 20 to 50% of the original thickness (typically about 0.5 mm, or 20 mils) was observed at the areas where no spallation occurred after 12,683 h of exposure. After additional 10,072 h of exposure, most coatings had lost 35 to 75% of the original coating thickness. The authors concluded that thermal sprayed coatings, which might require replacement after 21,000 h exposure, were not adequate in providing corrosion protection for waterwalls under low NOx combustion conditions. The thickness of the coating in the as-coated condition for both diffusion coatings was about 0.25 mm (10 mils). The chromized coating exhibited a surface layer containing about 20 to 30% Cr in the as-coated condition. The Cr/Si diffusion coating contained about 10% Cr, 14% Si, and balance Fe. In a

Fig. 10.20

Fig. 10.19

A corroded carbon steel tube sample from the waterwall of a boiler (subcritical unit) retrofitted with a low NOx burner system with overfire air ports. The waterwall tube suffered accelerated wastage after the furnace was retrofitted with NOx burner system. Courtesy of Welding Services Inc.

Scanning electron micrograph (backscattered electron image) showing the corrosion scales formed on the fireside of the tube sample (shown in Fig. 10.19). The chemical compositions of the corrosion scales at different locations were analyzed semiquantitatively by energy dispersive x-ray spectroscopy (EDX) analysis. The results are summarized as: 1: 66.9% Fe, 26.6% S, 1.9% Al, 2.6% Si, and minor elements 2: 69.5% Fe, 30.0% S, and minor elements 3: 64.3% Fe, 35.2% S, and minor elements 4: 74.6% Fe, 23.6% S, and minor elements 5: 91.0% Fe, 3.2% S, 4.4% Cl, and minor elements

pg 275

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

276 / High-Temperature Corrosion and Materials Applications

second test panel in the same boiler after exposure for about 17,155 h, both chromized coating and Cr/Si diffusion coating had completely failed, with corrosion penetrating into the substrate steel. Four weld overlays were also included in the second test panel; they were Type 410 (12% Cr), Type 309 (23.5% Cr), Type 312 (31% Cr), and alloy 625 (21% Cr). The chromium content in the weld wire, as reported in the paper, is indicated in the parentheses. The overlays were reported to exhibit about 15% dilution. Thus, the chromium content in the overlay was, then, about 10% for Type 410 overlay, 20.0% for Type 309 overlay, 26.4% for Type 312 overlay, and 17.9% for alloy 625 overlay. After exposure for about 17,155 h, Type 410 overlay with only about 10% Cr suffered general wastage and sulﬁdation at a wastage rate equivalent to half of that of carbon steel, thus providing no protection. Alloy 625 with about 18% Cr showed wastage rates of about 0.125 to 0.25 mm/yr (5 to 10 mpy). Both 309 and 312 overlays with 20% or higher Cr showed no measurable losses in the overlay thickness.

Fig. 10.21

Weld overlay using the composition that matches the boiler tube chemistry had often been used in the past to make temporary repair of the boiler tube in the ﬁeld by manual welding techniques. These manually applied weld overlays were often referred to as “pad welds.” In mid1980s, an automated weld overlay technology using gas metal arc welding (GMAW) process was developed to apply a corrosion-resistant overlay alloy onto a large area of waterwall in a waste-to-energy boiler that burned municipal solid waste (Ref 36). This modern, automated weld overlay technology has been expanded from relatively small waste-to-energy boilers to large coal-ﬁred boilers (Ref 37–40). A schematic showing an example of overlay weld beads covering the waterwall by a GMAW overlay welding process is shown in Fig. 10.21. Figure 10.22 shows an example of the cross section of a corrosion-resistant overlay on a waterwall sample that was cut from an overlaid waterwall. The weld overlay was applied in the ﬁeld using automated overlay welding process. The weld overlay was characterized with consistency in weld bead sequence and overlay thickness of each weld bead. This automated

Overlay weld beads applied to cover the waterwall during GMAW overlay welding process. The indicated bead sequence number is for illustration purposes. The actual overlay welding of the waterwall may not follow the bead sequence number indicated here. The weld bead typically progresses in a vertical down mode from top (left side of the schematic) to the bottom (right side of the schematic). Courtesy of Welding Services Inc.

pg 276

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

Chapter 10:

GMAW overlay welding process can be carried out on-site in the boiler or in-shop for overlaid panels (Ref 37–40). Figure 10.23 shows a partial view of a waterwall that was overlaid in the ﬁeld using a nickel-base alloy. The overlay cladding can also be accomplished using a laser technology, but the laser cladding can only be performed in-shop (Ref 41–43). For the weld overlay to provide adequate resistance to ﬁreside corrosion and erosion/corrosion, dilution in chemistry must be low when overlay welding is applied onto the waterwall. In overlay welding, the surface layer of the substrate steel must be melted to establish a metallurgical bond between the overlay and the substrate steel. If too much substrate steel is melted and mixed with the molten overlay weld wire, the overlay chemistry will thus be signiﬁcantly diluted. This can result in lowering the concentration of chromium in the overlay too low to form a protective chromium oxide scale, thus losing the protective nature of a weld overlay. For example, one utility company selected Type 430 stainless steel (18% Cr) for ﬁeld application of the overlay on the waterwall of a boiler using a submerged arc welding process (Ref 14). The “430” overlay was found to experience severe wastage problems due to signiﬁcant dilution in overlay chemistry (Fig. 10.24). The overlay was found to contain only about 11% Cr, thus making the overlay vulnerable to oxidation/sulﬁdation attack. This overlay was, thus, not capable of forming protective chromium oxide scales, instead forming heavy scales in Fe-Cr oxides and sulfurrich Fe-Cr phases at the scale/metal interface.

Coal-Fired Boilers / 277

Figure 10.25 shows the heavy Fe-Cr oxide scales formed on this overlay during service. The most common overlay alloys that have been applied in both subcritical and supercritical units to protect the waterwalls from severe ﬁreside corrosion attack under the low NOx combustion conditions are austentic stainless steel Type 309 and two nickel-base alloys 625 and 622 (Ref 39–41). Chromium is the most important alloying element in the alloy (or overlay) to allow

Fig. 10.23

Partial view of a waterwall that was overlaid with a nickel-base alloy using automated GMAW process on-site. Courtesy of Welding Services Inc.

Fig. 10.22

An example of an overlaid waterwall sample that was cut from the overlaid waterwall applied by automatic overlay welding in the field, showing consistency in weld bead sequence and overlay thickness of each weld bead. The waterwall shown in the figure is of a tangent tube design (no membranes). Courtesy of Welding Services Inc.

Fig. 10.24

A Type 430 stainless steel overlaid waterwall sample, showing severe overlay wastage and cracking. Courtesy of Welding Services Inc.

pg 277

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:02PM Plate # 0

278 / High-Temperature Corrosion and Materials Applications

the overlay to form chromium oxide scales to provide adequate protection by preventing the formation of sulﬁdes and/or iron-rich oxides. The speciﬁcation range of chromium is 23.0 to 25.0% for Type 309 weld wire (ER309), 20 to 23% for alloy 625 weld wire (ERNiCrMo-3), and 20.0 to 22.5% for alloy 622 (ERNiCrMo-10). The automated GMAW process typically aims to achieve a dilution in the overlay of approximately 10% or less. Figure 10.26 shows the chronology of the total square feet of waterwall overlays from early 1990s to 2001 by a leading applicator of weld overlay cladding in terms of the year of application for three major overlay alloys (Ref 14). Beginning in 2000 and 2001, the industry began switching from alloy 625 to alloy 622 in waterwall overlay cladding applications in coalﬁred boilers, particularly in supercritical units. This switch was related to circumferential grooving of the alloy 625 overlay on the waterwalls of several supercritical units. (The issue of circumferential grooving and cracking is discussed in section 10.5.3.) In ﬁreside corrosion and erosion/corrosion, it is desirable to select an overlay alloy with adequate chromium to form protective chromium oxide scales. Another factor that is viewed by some boiler operators to be important is the thermal expansion coefﬁcient mismatch between the overlay and the substrate ferritic steel. Nickel-base alloys, such as alloys 625 and 622, generally have thermal expansion coefﬁcients that match quite well with carbon and Cr-Mo steels. Austenitic stainless steels, such as Type 309, exhibit thermal expansion coefﬁcients that are higher than those of carbon and Cr-Mo steels. Mean coefﬁcients of thermal expansion for

carbon and Cr-Mo steels (Ref 44), Type 309 (Ref 14), alloy 625 (Ref 45), and alloy 22 (or 622) (Ref 46) at temperatures of interest for waterwalls are shown in Table 10.9. There exists some thermal expansion coefﬁcient mismatch between the 309 overlay and T-11 substrate steel. Nevertheless, no cracks have been found to occur at the interface (or the fusion boundary) between the 309 overlay and the substrate steel waterwall tube due to this thermal expansion mismatch. Figure 10.27 shows typical interface structure between the 309 overlay and the T11 substrate steel waterwall tube in a supercritical unit after 10 years of service. The stresses developed due to this thermal expansion coefﬁcient mismatch appear to be too small to initiate cracking. Furthermore, both the 309 overlay and the substrate carbon and Cr-Mo steels exhibit excellent ductility and toughness to prevent any development of cracks at the interface. The concern of high stresses that can lead to cracking at the overlay/substrate interface during operation in the 309 overlay because of its higher thermal expansion coefﬁcients prompted Coleman and Gandy (Ref 47) to examine Type 312 (a duplex stainless steel) as an alternative stainless steel overlay alloy. Type 312 exhibits thermal expansion coefﬁcients that are much lower than those of Type 309, and the alloy contains higher chromium (about 30% Cr in the weld wire), making it a potentially good candidate overlay alloy. However, this alloy is a duplex stainless steel, which consists of austenite and ferrite phases, and can suffer severe embrittlement when exposed to a temperature range of

Total overlay area, ft2

35,000 Type 309 Alloy 625 Alloy 622

30,000 25,000 20,000 15,000 10,000 5,000 0

1993 1994 1995 1996 1997 1998 1999 2000 2001 Year overlay performed

Fig. 10.26

Fig. 10.25

Optical micrograph showing nonprotective Fe-Cr oxide scales formed on the “430SS” weld overlay. Courtesy of Welding Services Inc.

Chronology of the total area of the weld overlay applied to the waterwalls of coal-fired boilers (both subcritical and critical units) from early 1990s to 2001 as a function of the year of application for three major overlay alloys— Type 309, alloy 625, and alloy 622—by a weld overlay application company. The data covers only field application overlays. 1.0 ft2 = 0.093 m2. Source: Ref 40

pg 278

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

Table 10.9

pg 279

Coal-Fired Boilers / 279

Mean coefficient of thermal expansion Coefﬁcient of thermal expansion, 10−6 in/in/°F

Temperature

Carbon steel(a)

1.25Cr-0.5Mo(a)

309 SS(b)

Alloy 625(c)

Alloy 22(d)

70–600 °F 70–800 °F 70–1000 °F

7.2 7.6 8.0

… 7.7 7.8

9.3 9.5 9.7

7.4 7.6 7.8

7.0 7.4 7.7

(a) EPRI data Ref 44. (b) Weld overlay data Ref 14. (c) Wrought alloy data Ref 45. (d) Wrought alloy data Ref 46. Alloy 22 and alloy 622 meet UNS N06022 speciﬁcation. Unit conversions: 1.8 ×10−6 in/in/°F = 10−6/K = 10−6 m/m/°C; 70 °F = 21 °C; 600 °F =316 °C, 800 °F = 427 °C, 1000 °F = 538 °C

Overlay

Substrate

0.010 in.

Fig. 10.27

Optical micrograph showing typical interface structure in longitudinal orientation between the 309 overlay and T11 substrate steel (waterwall tube) after 10 years of service in a supercritical boiler, revealing no cracking at the interface. Substrate steel was etched with nital to show the interface. Courtesy of Welding Services Inc.

approximately 371 to 593 °C (700 to 1100 °F) due to formation of ordered α‫ ׳‬phases. This embrittlement phenomenon is commonly referred to as 475 °C (885 °F). This embrittling temperature range is within the waterwall metal temperature range. Furthermore, Type 312 weld overlay is susceptible to weld solidiﬁcation cracking when a weld overlay having a low dilution is attempted (Ref 48). There have been several boilers where Type 312 overlay was applied to protect the waterwalls against ﬁreside corrosion. In one supercritical unit where Type 312 waterwall overlay was inspected recently after about 6.5 years of service, and an overlay sample was obtained from the waterwall for metallurgical examination. Surprisingly, the overlay was found to exhibit excellent condition, showing no evidence of cracking or embrittlement, or circumferential grooving/cracking during service (Ref 49). Typical macro cross sections of the overlay waterwall tubes after about 6.5 years of service in a supercritical unit are shown in Fig. 10.28. Figure 10.29 shows typical

Fig. 10.28

Macro cross sections cut from a Type 312 overlay waterwall panel sample obtained from a supercritical unit showing the overlay after 6.5 years of service in (a) the transverse cross section and (b) the longitudinal cross section. The macro samples were etched in nital to reveal the overlay. The scale is in inches. 1.0 in. = 25.4 mm. Source: Ref 49

metallographic cross section of the 312 overlay after about 6.5 years of service, showing essentially no corrosion and no cracking. Bonnington and Brennan (Ref 43) reported installation of laser-clad 312 overlay panels in supercritical boilers at the Mirant Mid-Atlantic station in 2000. Mirant also tested a laser-clad

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

280 / High-Temperature Corrosion and Materials Applications

Type 446SS overlay in a high-wastage area of a supercritical unit in 1991. The test results showed excellent performance of Type 446 overlay in the boiler (Ref 43). However, the cost of producing a laser-clad panel with Type 446 was considered to be prohibitive at that time (Ref 43). Type 446 is a Fe-25Cr ferritic stainless steel, which is also susceptible to the 475 °C (885 °F) embrittlement. Ferritic steels, such as carbon and low-alloy steels, typically exhibit higher thermal conductivities than austenitic stainless steels and nickel-base alloys. As a result, the outer skin metal temperature of the waterwall tube will be increased somewhat during service when a weld overlay cladding of an austenitic stainless steel (e.g., Type 309) or a nickel-base alloy (e.g., alloys 625 and 622) is applied onto a carbon steel or low-alloy steel. Blough (Ref 50) performed heat transfer calculation using the thermal conductivities of wrought alloys. His results are summarized in Fig. 10.30, which compares outer skin overlay temperatures for Type 309, alloy 625, and alloy 622 overlays with T2 (0.5 Mo

steel) weld metal buildup to the same thickness of about 0.090 in. (2.3 mm) on the bare steel tube (Ref 50). For example, for a bare tube with an outer metal skin temperature of 482 °C (900 °F), the application of an overlay of T2 of about 90 mils (2.3 mm) thick to the bare tube increases the skin temperature by approximately 33 °C (60 °F). In comparison with T2 overlay, Type 309 overlay would increase the skin temperature by about 22 °C (40 °F) higher than that of T2 overlay. Similarly, alloy 622 would increase the skin temperature by about 28 °C (50 °F) higher than that of T2 overlay, and alloy 625 by about 36 °C (65 °F) higher than that of T2 overlay. Type 312 overlay is expected to show lower outer skin metal temperature than Type 309 (or alloys 625 and 622) because of its duplex microstructure containing a mixture of austenite and ferrite. Since the application of a weld overlay using Type 309, Type 312, or nickel-base alloy has essentially eliminated the general wastage problem, it becomes desirable to apply a thinner overlay to lower the outer skin overlay metal temperature. The overlays of Type 309, alloy 625, alloy 622, and Type 312 have so far accumulated various service times with 309 and 625 overlays staying in service the longest (about 9 to 10 years) in some boilers. These overlays have essentially eliminated the ﬁreside wastage problems caused by staged combustion with low NOx burners and overﬁre air (Ref 35, 37–43). However, several supercritical boilers have been found to experience circumferential grooving and cracking problem. The circumferential grooving/cracking problem is discussed in section 10.5.3.

09 in.

Fig. 10.29

Optical micrograph showing the transverse cross section of Type 312 overlay on the waterwall in a supercritical unit after about 6.5 years of service, revealing essentially no corrosion and no cracking

Fig. 10.30

Calculated overlay metal temperatures for T2 bare tube with 0.090 in weld metal build-up of T2 in comparison with 0.090 in overlay of Type 309, alloy 625, and alloy 622. Source: Ref 50

pg 280

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

10.5.3 Circumferential Grooving of Furnace Waterwalls Fireside corrosion of the furnace waterwalls particularly in supercritical units has caused carbon and low-alloy steel waterwall tubes to suffer circumferential (transverse) grooves that resemble an alligator hide or an elephant hide. An example of a waterwall tube exhibiting this type of circumferential grooving appearance is shown in Fig. 10.31. This type of waterwall corrosion has been discussed by Wright (Ref 51) and French (Ref 15). Wright (Ref 51) indicated that circumferential grooving has been found to occur more frequently in supercritical boilers than subcritical boilers. He also indicated that the waterwall areas that suffered this type of attack received the highest heat ﬂux with superimposed thermal stresses. French (Ref 15) suggested that the circumferential grooving was caused by a corrosion-fatigue mechanism. He indicated that the corrosion products in the circumferential groove contained sulfur and the ash deposits on the waterwall tube surface contained carbon (Ref 15). He thus suggested that the reducing atmospheres were an essential part of the grooving-corrosion mechanism. The circumferential grooves in the later stage exhibit a thermalfatigue crack appearance. Some supercritical boilers historically had circumferential grooving/ cracking problems even prior to the installation of weld overlay claddings, such as the ones in

pg 281

Coal-Fired Boilers / 281

Conemaugh Station (Ref 52), Morgantown, and Chalk Point Generating Stations (Ref 53). In order to determine the root cause of circumferential grooving, it is best to examine the initial stage of the groove formation. This is illustrated by several examples for carbon and low-alloy steel tubes. Figure 10.32 shows the initiation of several circumferential grooves on a T22 tube in a supercritical boiler. Scanning electron microscopy with energy dispersive x-ray spectroscopy (SEM/EDX) was used to determine semiquantitatively the chemical compositions of the corrosion products inside the groove. One of the grooves shown in Fig. 10.32 was analyzed by SEM/EDX with the results shown in Fig. 10.33. The corrosion products were mainly enriched in iron and sulfur particularly in light grayish “stringers” or “channels” in the interior region containing higher sulfur. Areas 2 and 3 are located in the light grayish channels (or stringers), as shown in Fig. 10.33. Similar observation was made for a T2 tube in another supercritical boiler, as shown in Fig. 10.34, revealing largely iron sulﬁdes in the corrosion products. Again, there were light grayish phases in a form of stringers or “channels” containing higher sulfur (Area 2, 3, and 4 in Fig. 10.34). These stringers or channels, which tended to form in the interior region, extended from the surface of the corrosion products and penetrated along with the overall corrosion penetration. The one in the center tends to be much wider and longer. These stringers, which were found to be more enriched in sulfur than the surrounding phases, appear to be gas paths

0.0010 in.

Fig. 10.31

An example of the circumferential grooving encountered in the T2 steel waterwall. Courtesty of Welding Services Inc.

Fig. 10.32

Optical micrograph showing the initial development of several circumferential grooves on a T22 waterwall tube in a supercritical unit. Courtesy of Welding Services Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

pg 282

282 / High-Temperature Corrosion and Materials Applications

formed during the corrosion process. In some cases, these stringers were quite complex and do not appear to be the cracks which are subsequently ﬁlled with another corrosion product or ash deposit. Figure 10.35 shows the case of a very dense group of stringers or channels formed in the circumferential groove on T11 tube in a supercritical boiler. Several supercritical boilers have been found to experience circumferential grooving for Type 309 overlay (Ref 39) and alloy 625 overlays (Ref 39, 54) on waterwalls. Figure 10.36 shows a longitudinal cross section of an alloy 625 overlay waterwall tube at the crown location after 1 year of service in a supercritical boiler. Two tiny circumferential grooves are visible. One of the grooves was analyzed by SEM/EDX, showing that the alloy 625 overlay had suffered essentially sulﬁdation penetration. Figure 10.37 shows the corrosion products to be essentially chromium sulﬁdes. There was a centerline channel which

was depleted in chromium (Fig. 10.37b), but enriched in sulfur, as shown in Fig. 10.37(c), and also in Ni (the Ni dot map was not shown). Thus,

(a)

140 µm

60 µm

Fig. 10.33

Scanning electron micrograph (backscattered electron image) showing one of the circumferential grooves formed on a T22 waterwall tube (2.25Cr-1Mo) as shown in Fig. 10.32. Semiquantative energy dispersive x-ray spectroscopy (EDX) analysis in terms of weight percent at different locations of the corrosion products inside the groove is summarized below. Courtesy of Welding Services Inc. 1: 83.9% Fe, 7.1% S, 4.6% Cr, 2.1% Mo, 0.5% Al, 0.7% Si, 0.9% Mn, and trace elements 2: 72.2% Fe, 21.2% S, 3.6% Cr, 1.3% Mo, 0.5% Al, 0.9% Mn, and trace elements 3: 73.3% Fe, 22.8% S, 2.2% Cr, 0.6% Mn, and trace elements 4: 80.6% Fe, 9.8% S, 5.1% Cr, 2.5% Mo, 0.9% Mn, 0.5% Al, and trace elements 5: 87.6% Fe, 1.3% S, 6.7% Cr, 2.6% Mo, 1.2% Al, 0.8% Si, and trace elements 6: 87.0% Fe, 0.8% S, 7.5% Cr, 2.8% Mo, 0.8% Si, 0.7% Mn, and trace element

(b)

60 µm

Fig. 10.34

(a) Scanning electron micrograph (backscattered electron image) showing typical morphology of the circumferential groove formed on a T2 tube (0.5Cr-0.5Mo) in a supercritical boiler. The results (in wt%) obtained from the semiquantitative analysis using EDX on the corrosion products are summarized as: 1: 86.8% Fe, 10.5% S, 1.4% Cr, and minor elements 2: 79.7% Fe, 18.2% S, 0.64% Cr, and minor elements 3: 69.0% Fe, 30.4% S, and minor elements 4: 67.3% Fe, 32.0% S, and minor elements 5: 93.7% Fe, 0.5% S, 2.5% Cr, 1.9% Mo, 1.2% Si, and minor elements 6: 91.8% Fe, 3.7% S, 1.3% Cr, 1.2% Mo, and minor elements (b) Detail of regions 1 to 4

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

this centerline channel appeared to be nickel sulﬁde. Similar observation of the corrosion products in the circumferential groove formed in alloy 625 overlay cladding was also reported by Luer et al. (Ref 54). Figure 10.38 shows a circumferential groove that formed on alloy 625 overlay on a waterwall tube after 2 years of service in another supercritical boiler. The corrosion products were found to consist of essentially chromium sulﬁdes. The alloy 625 was etched to show the dendritic microstructure of the weld overlay. Figure 10.39 shows the formation of the initial preferential sulﬁdation penetration at a high magniﬁcation, revealing in detail the corrosion phases, which consisted of essentially chromium sulﬁdes. Type 309 overlay on the waterwall of a supercritical unit was found to suffer circumferential grooving after 10 years of service. Figure 10.40 shows circumferential grooves at the early stage. The circumferential groove consisted of several branches of preferential penetrations. These several branches of preferential penetrations tended to converge into a groovelike attack. The morphology of the circumferential groove (at the early stage of development) formed in Type 309 overlay (an austenitic stainless steel) is somewhat different from that formed in nickelbase alloy 625. The corrosion products in one of the branches of preferential penetrations, as shown in Fig. 10.40, were analyzed by SEM/ EDX with the results summarized in Fig. 10.41. The SEM/EDX results indicate the corrosion

pg 283

Coal-Fired Boilers / 283

products were enriched in chromium and iron with some sulfur. Sulfur was particularly high within the phases inside the light grayish stringers that were clustered around the inner portion of the corrosion penetration attack. However, the levels of sulfur were signiﬁcantly lower than those observed in alloy 625 overlay. It is thus believed that corrosion penetration was essentially combined oxidation and sulﬁdation in the 309 overlay. From the previous discussion of the initial development of the circumferential grooving in both alloy 625 and 309 overlays, it is concluded that the circumferential grooves were initially developed by way of preferential sulﬁdation penetration. For nickel-base alloy 625 overlay, sulﬁdation was the major mode of preferential attack, while for Type 309 overlay (an austenitic stainless steel), sulﬁdation along with oxidation was involved in the preferential attack. As the preferential attack continued, the circumferential groove, which penetrated farther into the overlay with increasing stresses at the penetration tip, eventually developed into a crack at a later stage. This is illustrated in Fig. 10.42. In another power station, Type 309 overlay has provided protection for the waterwalls of two supercritical units for 8 years in one unit and 10 years for the other unit. The 309 overlay in one unit was inspected using nondestructive testing including dye-penetrant testing in 2006 after 8 years of service. The inspection revealed

0.5 mm

0.0010 in.

Fig. 10.35

Optical micrograph showing a circumferential groove containing a complex “network” of stringers formed on T11 tubes (1.25Cr-0.5Mo) in a supercritical boiler. Courtesy of Welding Services Inc.

Fig. 10.36

Optical micrograph showing the initial development of two tiny circumferential grooves formed on the alloy 625 overlay at the crown bead of a waterwall tube in a supercritical unit after 1 year of service. The metallographic mount was in the longitudinal cross section. One of the grooves formed on the overlay (left side of the micrograph) was analyzed by SEM/EDX with the results shown in Fig. 10.37. Source: Ref 40

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

pg 284

284 / High-Temperature Corrosion and Materials Applications

no sign of tube wastage and circumferential cracking (Ref 14). Typical weld overlay is shown in Fig. 10.43. The unit was equipped with low NOx burners and overﬁre air, burning bituminous coal containing 3.0 to 3.5% sulfur. Type 309 waterwall overlay, which has accumulated so far for 10 years of service, was to be inspected during the next maintenance shutdown. In laboratory testing of various alloys in sulﬁdizing environments, several authors (Ref 55–58) have found that applied tensile stresses during exposure to a sulﬁdizing environment at

elevated temperatures can produce preferential sulﬁdation penetration on wrought alloys. Guttmann et al. (Ref 56) showed that several alloys were suffering preferential sulﬁdation penetration under tensile strains in a sulﬁdizing environment. Their tests were conducted at 600 °C (1112 °F) in CO-32H2-4CO2-0.2H2S (inlet gas mixture) with equilibrium sulfur and oxygen potentials at the test temperature being about 10−11 and 10−28 bar (or atm), respectively. Figure 10.44 shows alloy 45TM (Ni-27Cr-23Fe2.8Si) after exposure to the environment at

(a)

(b)

Fig. 10.37

0.0010 in.

50 µm

(c)

50 µm

(a) Optical micrograph of a circumferential groove in Fig. 10.36. (b) Chromium x-ray dot map for the corrosion products inside the groove. (c) Sulfur x-ray dot map for the corrosion products inside the groove. Source: Ref 40

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

temperature for 2000 h under (a) no external strain and (b) a 2% strain. The specimen under no external strain during exposure showed essentially uniform corrosion attack, while the one under a 2% strain showed preferential penetration attack. The morphology of the preferential penetration attack is similar to the circumferential groove observed in nickel-base alloy 625 overlay on the waterwall. The preferential penetration was found to be essentially oxidation/sulﬁdation. Also observed in this preferential penetration attack were channels that were similar to those observed in circumferential groove. The channel and stringers contain phases with much lighter color, similar to what was observed in circumferential grooves formed in Cr-Mo steel waterwall tubes or alloy 625 overlay in few supercritical boilers. Coze et al. (Ref 59) observed similar phenomena in testing Fe-12Cr-3Al-3Ti experimental alloy in H2-34.3H2O-18.5CO2-3.8CH4-7.9CO-1.3H2S at 600 °C (1112 °F) 615 h under no external stresses and external stresses, as shown in Fig. 10.45. The internal penetration attack under stresses (upper micrograph of Fig. 10.45) is very similar to circumferential grooves formed in Type 430 overlay on the waterwall tube in a boiler, as shown in Fig. 10.46.

Fig. 10.38

Coal-Fired Boilers / 285

For Fe-Cr-Ni alloys, such as austenitic stainless steels, the stress-enhanced sulﬁdation penetration attack takes a different morphology. Guttmann et al. (Ref 56) found that alloy HR3C (Fe-25Cr-20Ni-0.5Nb-0.2N), which is a wrought alloy, suffered stress-enhanced preferential attack along grain boundaries instead of a ﬁngerlike protrusion, as was observed in alloy 45TM (a nickel-base alloy). Figure 10.47 shows alloy HR3C after testing in CO-32H2-4CO20.2H2S (inlet gas mixture) at 600 °C (1112 °F) under (a) no external stresses showing general uniform corrosion attack after 2100 h and (b) a 1.3% strain after 250 h of exposure showing grain-boundary attack. For Type 309 overlay (approximately Fe-21Cr-12Ni) on the waterwall, preferential sulﬁdation penetration attack followed individual branches of continuous penetrations during the initial stage of development in the circumferential grooving, as shown in Fig. 10.40. These individual branches of sulﬁdation penetrations were most likely to be interdendritic boundaries. From the discussions in this section, it can be concluded that circumferential grooving is a preferential sulﬁdation penetration attack under tensile stresses. Tensile stresses can be varying. However, thermal cycling is not a prerequisite for

A circumferential goove formed on alloy 625 overlay applied to the waterwall of a supercritical boiler after 2 years of service. The overlay was etched to show the dendritic microstructure of the alloy 625 overlay. Courtesy of Welding Services

pg 285

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

286 / High-Temperature Corrosion and Materials Applications

initiating circumferential grooving, even though the circumferential groove at a later stage can develop into a crack. Wright (Ref 51) indicated that circumferential grooving typically occurs in the areas of the waterwall receiving the highest heat ﬂux and is apparently a result of superimposed thermal stress. It is believed that the preferential sulﬁdation penetration that initiates the circumferential grooving at certain locations of the waterwall is due to overheating under sulﬁdizing or alternating sulﬁdizing/oxidizing conditions. Although a temperature range for waterwall tubes has been mentioned to be 400 to 500 °C (752 to 932 °F) (Ref 9) and the design temperature of some

(b)

Fig. 10.39

supercritical units was mentioned to be 482 to 510 °C (900 to 950 °F) (Ref 10) and 482 to 538 °C (900 to 1000 °F) (Ref 11), it is most likely that the overheating at certain locations of the waterwall is due to ﬂame impingement. Since alloy 625 overlay has been found to suffer circumferential grooving in several supercritical units, the outer overlay metal temperature of the overlay can be estimated based on the aging characteristics of alloy 625. Alloy 625 is known to exhibit age hardening at intermediate temperatures of 538 to 760 °C (1000 to 1400 °F). The alloy age hardens by forming γ″ (Ni3Nb) precipitates in this temperature range. At 538 to 593 °C (1000 to 1100 °F)

(c)

Scanning electron micrograph (a) showing a very early stage of the circumferential grooving formed in alloy 625 overlay on a waterwall tube in the same supercritical boiler, as shown in Fig. 10.38. EDX spectra (b) and (c) show the alloying elements associated with the “light” phases (b) and grayish phases (c), indicating the corrosion phases were essentially chromium sulfides. The sample surface was plated prior to mounting to protect the corrosion products for SEM examination. Courtesy of Welding Services Inc.

pg 286

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

0.0010 in.

Fig. 10.40 Optical micrograph showing circumferential grooves that formed in Type 309 overlay applied to the waterwall of a supercritical boiler after 10 years of service. Courtesy of Welding Services Inc.

Coal-Fired Boilers / 287

for 1 to 2 years of aging, the alloy forms ﬁne, coherent precipitates causing signiﬁcant hardness increases particularly at 593 °C (1100 °F). With aging at 650 °C (1200 °F), initial precipitates formed in the alloy are ﬁne, coherent γ″, but later become incoherent long, platelet Ni3Nb precipitates, as shown in Fig. 10.48 (Ref 61). The microhardness proﬁle was determined on the alloy 625 overlay that was found to show initiation of circumferential grooving after about 1 year in service (shown in Fig. 10.36). The results, which are shown in Fig. 10.49, indicate age hardening to 35 to 45 HRC at the outermost surface layer of the overlay (Ref 40). The interior part of the overlay exhibited a hardness level (23 to 25 HRC) that essentially corresponded to the hardness of the as-overlay alloy 625. Based on the hardness data, the metal temperature of the overlay surface layer is believed to have increased to above 538 °C (1000 °F), most likely to about 593 °C (1100 °F), the temperature that causes signiﬁcant hardening for alloy 625. In another supercritical boiler where alloy 625 overlay on the waterwall suffered circumferential

50 µm

Fig. 10.41

Scanning electron micrograph (backscattered electron image) showing the corrosion products in a circumferential groove formed in Type 309 overlay on the waterwall of a supercritical boiler after 10 years of service. The results of the semiquantative EDX analysis of the corrosion products at different locations are as summarized below. Courtesy of Welding Services Inc. 1: 62.2% Cr, 29.2% Fe, 4.6% Mn, 1.7% Ni, 1.4% S, and trace elements 2: 40.7% Cr, 39.5% Fe, 4.2% Mn, 4.1% Ni, 8.4% S, 2.4% Si, and trace elements 3: 28.0% Cr, 59.8% Fe, 2.6% Mn, 4.5% Ni, 2.9% S, 1.5% Si, and trace elements 4: 18.1% Cr, 62.4% Fe, 1.5% Mn, 4.7% Ni, 11.6% S, 1.0% Si, and trace elements 5: 46.8% Cr, 40.2% Fe, 2.3% Mn, 2.2% Ni, 5.8% S, 2.2% Si, and trace elements 6: 45.8% Cr, 43.2% Fe, 3.2% Mn, 2.3% Ni, 3.5% S, 1.5% Si, and trace elements

Fig. 10.42

Some well-developed circumferential grooves, which developed into cracks at a later stage, were observed in the 309 overlay on the waterwall after 10 years of service. There was no evidence of cracking developed at the interface between the 309 overlay and the T11 substrate even with some thermal expansion coefficient mismatch. The substrate steel was etched with nital to reveal the fusion boundary. Original magnification: 25×

pg 287

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

288 / High-Temperature Corrosion and Materials Applications

grooving after 2 years of operation, the overlay surface layer was also found to exhibit age hardening to about 40 to 42 HRC. Unlike alloy 625, Type 309 does not exhibit age-hardening characteristics. As a result, the possible overheated overlay surface temperature could not be estimated using microhardness data. Chromium carbides have been found to precipitate along grain boundaries in a Type 309 waterwall overlay after 10 years of service in a supercritical unit, as shown in Fig. 10.50. Precipitation of these chromium carbides may be an indication that the overlay was overheated to above 538 °C (1000 °F). In performing metallurgical evaluation of a 309 overlay sample that was removed from the same boiler in a waterwall area (probably from a different location as the area discussed in Fig. 10.50) after 7 years of service, the microstructure of the overlay showed no chromium carbide precipitates formed along grain boundaries (Fig. 10.51) (Ref 37). The temperature of the steam in this area was estimated by a plant engineer to be 371 to 427 °C (700 to 800 °F) (Ref 14). Without the evidence of

chromium carbides along grain boundaries, the overlay metal temperature was likely to be below 538 °C (1000 °F). Microhardness measurements made across the overlay that corresponds to the microstructure containing carbides, as shown in Fig. 10.50, indicated about 250 HV (about 100 HRB) as opposed to 200 HV (about 90 HRB) for the overlay that corresponds to the microstructure with no carbides, as shown in Fig. 10.51. The 309 overlay is typically applied using a low-carbon Type 309 weld wire (ER309LSi). Thus, age-hardening response is generally minimal. Accordingly, the development of preferential grooving and later cracking was not related to the aging behavior of the 309 weld overlay, which is expected to exhibit good ductility, after exposure to 538 to 593 °C (1000 to 1100 °F) for 10 years.

Fig. 10.43

Fig. 10.44

Type 309 overlay on the waterwall in another supercritical unit equipped with low NOx burners and overfire air after service for 8 years, showing no sign of metal wastage or circumferential cracking. The boiler reportedly burned bituminous coal containing 3 to 3.5% S. Courtesy of Welding Services Inc.

Alloy 45TM (Ni-27Cr-23Fe-2.8Si) after exposure to the test gas of CO-32H2-4CO2-0.2H2S (inlet gas mixture) at 600 °C (1112 °F) for 2000 h under (a) no external strain and (b) 2% strain. The equilibrium sulfur and oxygen potentials for the test environment at the test temperature were about 10−11 and 10−28 bar (or atm), respectively. Source: Ref 56

pg 288

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

Tensile ductility after long-term aging may not be an important factor in the development of circumferential cracking. An example is illustrated by Type 312 waterwall overlay that has been in service in a supercritical unit for about 6.5 years with no evidence of circumferential cracking being revealed by dye-penetrant testing during a recent inspection (Ref 14). Samples obtained from the overlaid waterwall for metallurgical examination also showed similar results. Figure 10.28(b) shows a longitudinal macro cross section of the overlay sample, revealing no circumferential grooving or cracking. Figure 10.52 shows typical overlay surface condition in a longitudinal metallographic cross section, again, revealing no

Coal-Fired Boilers / 289

circumferential grooving or cracking. The microhardness proﬁle across the 312 overlay at the crown bead from two samples is shown in Fig. 10.53. Both samples showed some hardening at the region close to the overlay surface with one sample showing signiﬁcantly more hardening (close to 50 HRC). Type 312 overlay is a duplex stainless steel containing a mixture of austenite and ferrite phases. When exposed to a temperature range of approximately 343 to 593 °C (650 to 1100 °F), the ferrite phase is subject to ordering reaction forming α′ ordered precipitates, resulting in signiﬁcant hardening and causing the alloy to become brittle. This phenomenon is commonly referred to as the 475 °C (885 °F) embrittlement. Even under this brittle condition, the 312 overlay showed no evidence of circumferential grooving or cracking. The overlay was found to contain about 29% Cr. The corrosion products observed on the

Fig. 10.45

Fe-12Cr-3Al-3Ti experimental alloy after testing in H2-34.3H2O-18.5CO2-3.8CH4-7.9CO-1.3H2S at 600 °C (1112 °F) for 615 h under no external stresses (lower micrograph) and external stresses (upper micrograph). The external strain applied during test was unknown. Source: Ref 59

Fig. 10.47

Fig. 10.46

Circumferential grooves formed in Type 430 weld overlay on the waterwall tube in a coal-fired boiler. Source: Ref 60

Alloy HR3C after testing in CO-32H2-4CO20.2H2S (inlet gas mixture) at 600 °C (1112 °F) under (a) no external stresses showing general uniform corrosion attack after 2100 h and (b) a 1.3% strain after 250 h of exposure showing grain-boundary attack. Source: Ref 56

pg 289

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

290 / High-Temperature Corrosion and Materials Applications

overlay were found to be chromium-rich oxides, thus providing adequate protection against preferential sulﬁdation penetrations, a precursor of circumferential grooving and cracking. In the laboratory tests discussed above, applications of about 1 to 2% tensile strains on the test specimens were capable of causing preferential sulﬁdation penetration at about 593 °C (1100 °F) for less than a couple of thousands of hours of test duration for various alloys including nickel-base alloys and Fe-Cr-Ni alloys. It would appear that under lower tensile strains (e.g., 0.5

Fig. 10.48

to 1%) the alloys could develop preferential sulﬁdation penetration after longer exposure times, such as a year or longer). For circumferential grooving to develop, the axial stresses would have to be large enough to develop tensile strains of approximately 0.5 to 1%. Axial stresses generated on the waterwall tubes may come from the dead load of the waterwall panels, thermal stresses, pressures of water/steam inside the tube, and bending stresses due to localized ﬂame impingement (overheating) due to constraints from the surrounding cooler waterwall areas as

Aging behavior of alloy 625 (wrought alloy samples) at 650, 760, and 870 °C (1200, 1400, and 1600 °F) for 16,000 h in terms of microstructure, losses in impact toughness and elongation, and increases in strength. Source: Ref 61

pg 290

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

Chapter 10:

scales in a sulﬁdizing environment, thus allowing preferential sulﬁdation attack to take place. Three conditions are required to be met for developing circumferential grooving for the weld overlaid waterwall: (a) sulﬁdizing environment, (b) adequate tensile strains (e.g., 0.5 to 1%), and (c) overheating of tube outer skin metal to approximately 593 °C (1100 °F). For widely used weld overlays of Type 309, alloy 625 and alloy 622, the chromium content of the weld overlay at the crown location (i.e., the location subjected to the highest heat ﬂux) was typically 20%. This level of chromium in the

Rockwell C, HRC

well as buckstays. (Buckstays are structural shapes or trusses that encircle and restrain the movement of the furnace waterwalls caused by ﬂuctuation in furnace pressure, or transient internal or external loads.) The combination of these stresses may be adequate to lead to the development of preferential sulﬁdation penetrations and circumferential grooving when the overlay outer layer is heated to approximately 593 °C (1100 °F) under localized sulﬁdizing conditions. The development of preferential sulﬁdation penetrations are believed to result from local breakdown of the protective oxide

Coal-Fired Boilers / 291

Fig. 10.49

Microhardness profile measured using Vickers hardness tester with a 500 g load as a function of the distance from the overlay surface for alloy 625 weld overlay on the waterwall of a supercritical boiler after 1 year of operation when circumferential grooves, as shown in Fig. 10.36, were initiated. Vickers hardness values (HV) were converted to Rockwell C (HRC) values. Hardening of the overlay surface layer (within 0.5 mm, or 20 mils) is believed to result from age hardening of alloy 625 due to formation of fine, coherent γ″ (Ni3Nb) precipitates when heated to probably 593 °C (1100 °F). Source: Ref 40

Fig. 10.50

Optical micrograph showing typical microstructure of the 309 overlay on the waterwall that suffered circumferential grooving (Fig. 10.40) in a supercritical boiler after 10 years of service. Original magnification: 200×

pg 291

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:03PM Plate # 0

292 / High-Temperature Corrosion and Materials Applications

weld overlay may not be adequate when the overlaid waterwall is subject to the aforementioned three conditions. To avoid the local breakdown of the chromium oxide scales under these conditions, an overlay containing a higher level of chromium is required to ensure the adequate chromium content to maintain the protective chromium oxide scales over a prolong service duration. Alloy 52 (AWS ERNiCrFe-7 or UNS N06052) with 30Cr, 9Fe and balance Ni provides a good candidate overlay alloy with potentially improved resistance to circumferential grooving

and cracking. Alloy 52 overlay, which was applied to the waterwall of a supercritical unit, has now been in service for 3 years with no reported degradation problems (Ref 14). Paul et al. (Ref 62) indicated some ﬁeld testing of alloy 33 (UNS R20033: Fe-33Cr-32Ni) overlay test panels in two supercritical boilers, showing no cracking after slightly less than 2 years of testing. Both boilers were tangential ﬁring with low NOx and overﬁre air. Alloy 33 test panels were located in the overﬁre air region in both boilers. In both cases, alloy 622 overlay panels were installed at the same time in the same area. After slightly less than 2 years of exposure, both alloy 33 and alloy 622 overlays tube samples with similar exposure times were removed for metallurgical evaluation, showing no evidence of circumferential cracking for both overlay alloys (Ref 62). 10.5.4 Sootblowing

0.10 mm

Fig. 10.51

Microstructure of Type 309 overlay on the waterwall after 7 years of service from the same boiler, but likely from different area in the boiler. No chromium carbides along grain boundaries were detected in this area. The overlay showed no circumferential grooving or cracking. Source: Ref 37

One important by-product of the combustion of coal is ash. During combustion, some of the mineral constituents and compounds can be in a molten or plastic state. These ash constituents can then deposit on the heat-absorbing surfaces causing slagging and fouling. If ash reaches the heat-absorbing surface at a temperature near its softening temperature, the resulting deposits are likely to be porous and can be removed by sootblowing. Slagging is the deposition of molten, partially fused deposits on the furnace walls and the upper furnace radiant superheaters exposed to radiant heat (Ref 1). Fouling is the deposition of more loosely bonded deposits on the heat-absorbing surfaces in the convection path, such as superheater and reheater, that

0.0010 in.

Fig. 10.52

Optical micrograph showing typical longitudinal cross section at the crown bead of the 312 waterwall overlay after about 6.5 years of service in a supercritical unit. The corrosion scales were chromium-rich oxides as identified by SEM/EDX analysis. Courtesy of Welding Services Inc.

Fig. 10.53

Microhardness profile across the overlay from the overlay surface measured using Vickers hardness tester with a 500 g load. Vickers hardness values (HV) are converted to Rockwell C (HRC) values. Data were obtained from two different overlay samples (Series 1 and 2). 1 in. = 25.4 mm. Source: Ref 49

pg 292

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

are not exposed to radiant heat (Ref 1). Soot blowers using steam are generally adequate for removing the ash deposits on the fouled tube surfaces, while waterlances or water cannons using thermal shock by room-temperature water may be needed to clean the slagged tube surfaces. The damage that causes the tube surface due to steam sootblowers is generally referred to as sootblower erosion. Under normal operations, steam sootblowing is in the erosion/corrosion (i.e., erosion/oxidation) regime. An example is given here to illustrate this phenomenon. A 905 MW(e) supercritical unit had experienced severe waterwall tube wastage problem with its SA213 T11 tubes (1.25 in. outside diameter × 0.220 in. minimum wall thickness) subjected to steam sootblowing by wall blowers. These waterwall tubes had to be replaced every 18 months because of severe tube wastage. The boiler, which was not equipped with low NOx burners, burned the Illinois basin coal with about 2.5% S. The corrosion products were most likely iron oxides with possibly some iron sulﬁdes (if sulﬁdation also took place). No analysis was done on the corrosion products formed on T11 waterwall tube. However, when the soot blower affected area was weld overlaid with Type 309 using automatic GMAW process, the wastage problem was essentially eliminated. A Type 309 overlay tube sample was removed after 7 year of service for metallurgical evaluation. During this 7 year period, the boiler had burned the Illinois basin coal (2.5% S) for about 4 years and an Eastern Appalachian low sulfur coal for 3 years with low NOx burners and separated overﬁred air during these 3 years. The overlay showed no evidence of metal wastage and erosion/corrosion attack by steam sootblowing (Fig. 10.54 and 10.55). The reason for signiﬁcant life improvement offered by the 309 overlay is the erosion/corrosion resistance of the 309 overlay. The 309 overlay forms chromium oxide scales, which are much thinner and grow signiﬁcantly more slowly than iron oxides that form on T11 (1.25Cr-0.5Mo steel). Steam impingement on the steel removes iron oxide scales, which are thick, nonprotective, and fast growing. The iron oxide scales reformed repeatedly after they were removed by steam impingement, thus resulting in accelerated wastage. Once the tube was protected by the 309 overlay, metal wastage due to erosion/corrosion was signiﬁcantly reduced due to the formation of thin, protective chromium oxide scales.

Coal-Fired Boilers / 293

Nickel-base alloy overlays, such as alloy 622 overlay, that form protective chromium oxide scales will have similar resistance to sootblower erosion-corrosion. More discussion on erosion/ corrosion is covered in Chapter 8: Erosion and Erosion-Corrosion.

Fig. 10.54

Field-applied Type 309 overlay on the waterwall tube after 7 years of service in a 905 MW(e) supercritical boiler: (a) close-up view of the crown and side beads of the overlay, and (b) cross section of the crown bead of the overlay showing both the overlay surface and the overlay/substrate interface (as-polished condition, original magnification: 25×). Courtesy of Welding Services Inc.

pg 293

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

pg 294

294 / High-Temperature Corrosion and Materials Applications

10.5.5 Deslagging by Waterlances and Water Cannons One of the important characteristics of ash is its ash fusion temperatures. If the ash fusion temperatures are low enough, the ash deposit reaches its melting point (due to the ash’s thermal insulating properties) and the molten slag runs down the furnace wall surface (Ref 1). When solidiﬁed, this slag is tightly bonded and is difﬁcult to remove. This slag may require waterlances or water cannons to create thermal shock for the removal of this slag deposit. The furnace walls are likely locations for developing this slagging problem when coal with low ash fusion temperatures is combusted. When room-temperature water is used for deslagging the waterwall through either waterlances or water cannons, the waterwall tube is subjected to thermal shock. Repeated use of waterlances or water cannons for deslagging can cause thermal fatigue cracking of the waterwall tubes. Thermal fatigue cracking of waterwall tubes due to deslagging by waterlances or water cannons has been reported by Kessler (Ref 63), Carlisle et al. (Ref 64), Ray et al. (Ref 65), and Blinka (Ref 66). Kessler (Ref 63) reported ﬁeld measurement data that indicated a rapid temperature drop from approximately 379 to 317 °C (715 to 603 °F) for the waterwall tube right after water spraying from water cannons. He indicated in Ref 63, that this temperature drop measured

from the ﬁeld was signiﬁcantly lower than the EPRI’s 288 °C (550 °F) drop in temperature measured by a waterlance simulator used to test the waterwall panels when sprayed without a slag (no thermal insulating by a slag). The boiler at Plant Miller (Alabama Power), which burned Powder River Basin (PRB) coal, used waterlances for continuous waterwall cleaning with a 3 hour cycle time. This resulted in reduced waterwall tube life due to quench cracking (Ref 64). Ray et al. (Ref 65) reported that a boiler suffered waterwall tube damage (a waterwall blowout) after 2½ years of “unrestricted” waterlance usage. The morphology of thermal fatigue cracking on carbon steel waterwall tube as a result of water spraying from water cannons is illustrated in Fig. 10.56. Surface appearance of thermal fatigue cracks on a waterwall tube due to water spraying from waterlances is shown in Fig. 10.57. Figure 10.58 shows typical morphology of the circumferential cracks in a longitudinal cross section from the sample shown in Fig. 10.57. Figure 10.59 shows an SEM of a circumferential crack along with the energy dispersive x-ray spectroscopy (EDX) analysis on the corrosion product inside the crack. The EDX analysis showed the corrosion product inside the crack to be essentially iron oxides. One boiler operator decided to repair the cracked waterwall tubes, which were made of SA210 A1 steel, due to water spraying from waterlances by applying alloy 622 weld overlay cladding in the ﬁeld after the cracks were ground

0.1 mm

Fig. 10.55

Optical micrograph showing a very thin oxide scale on the surface of the 309 overlay on a waterwall tube (T11) after 7 years of service subjected to steam sootblowing in a supercritical boiler, revealing no evidence of erosion/corrosion damage or cracking. The 309 overlay was not etched showing white portion of the micrograph with a magnification marker at the bottom of the micrograph. Source: Ref 40

0.5 mm

Fig. 10.56

Optical micrograph showing typical morphology of circumferential thermal fatige cracking on a carbon steel waterwall tube due to water spraying from water cannons. The tube OD (fireside) is on the left side of the micrograph and the ID (water/steam side) is on the right side. Courtesy of Welding Services Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

off from the tubes. The repaired waterwall area was more than 420 m2 (4500 ft2). After 4 years of service, dye penetrant testing (PT) of the whole overlaid waterwall area showed no evidence of cracking. Figure 10.60 shows typical overlay area after PT testing. 10.5.6 Superheater and Reheater Corrosion Steam from either the steam drum in a subcritical unit or from the furnace wall in a supercritical unit ﬁrst passes through a primary

Fig. 10.57

Appearance of thermal fatigue cracks occurred on a carbon steel waterwall tube (viewed from 12 o’clock crown position) due to water spraying from waterlances. Source: Ref 40

0.5 mm

Fig. 10.58

Optical micrograph showing circumferential thermal fatigue cracks that developed on a carbon steel waterwall tube (shown in Fig. 10.57) due to water spraying from waterlances. Source: Ref 40

Coal-Fired Boilers / 295

superheater, typically a horizontal heat exchanger located above the economizer, and then through a secondary or ﬁnishing platen superheater (i.e., tubes are in a ﬂat arrangement and hung in the furnace roof) with a number of platens in parallel. In some boilers, this ﬁnishing superheater is heated by radiant heat from the furnace combustion. The steam leaving the secondary or ﬁnishing superheater passes through a high-pressure steam turbine. After expanding through the high-pressure steam turbine, the steam is returned to the boiler to be reheated in a reheater. Steam from the reheater then passes through an intermediate-pressure turbine, followed by passing through a low-pressure turbine. Typical superheated steam temperature in most utility boilers in the United States is about 538 °C (1000 °F). The metal temperatures of superheaters and reheaters may be up to 650 °C (1200 °F). Superheater and reheater tubes suffer oxidation attack at lower temperatures. Oxidation attack generally results in lower wastage rates. When ﬂue gas stream is entrained with ﬂy ash, the tubes can also suffer erosion/corrosion attack. The wastage rates under erosion/corrosion conditions can be much higher. When the metal temperature is approaching 650 °C (1200 °F) or higher, superheater and reheater tubes can suffer accelerated wastage rates due to coal ash corrosion. The accelerated wastage rates due to coal ash corrosion are the result of molten salt (sulfate) corrosion. French (Ref 15) suggests that the corrosion follows a “hockey stick” type behavior with two distinctive corrosion regimes: low metal wastage rate by oxidation and accelerated wastage rates by molten salt corrosion, as shown in Fig. 10.61. The molten salt corrosion behavior exhibits a “bell-shaped” curve with respect to temperature for austenitic stainless steels. The rate increases with temperature to a maximum, then decreases with increasing temperature. The accelerated corrosion associated with this bell-shaped curve is related to the formation of molten alkali metal-iron-trisulfate [(Na,K)3Fe (SO4)3] (Ref 67, 68). Variation of the sodium-topotassium ratio greatly affects the melting point of the complex sulfate in an ash deposit (Ref 68). Figure 10.62 illustrates a wide variation of melting points of mixtures of sodium iron trisulfate and potassium iron trisulfate (Ref 68). Corrosion rate was also affected by the sodiumto-potassium ratio in the complex sulfate (Ref 68). Nelson and Cain (Ref 69) conducted a series

pg 295

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

296 / High-Temperature Corrosion and Materials Applications

Fe

10,000 9,000 8,000

Counts

7,000 6,000

EDS-1 corrosion products

5,000 4,000 3,000 2,000 1,000

(a)

120 µm

O Fe C Mn

Fe Si S

Cr Mn 0 0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.000 11.000 12.000 13.000 14.000 15.000 keV

(b)

Fig. 10.59

(a) Scanning electron micrograph (backscattered electron image) showing a circumferential thermal fatigue crack (from the sample shown in Fig. 10.57) along with (b) an EDX spectrum showing the corrosion product inside the crack to be essentially iron oxides. Source: Ref 40

Fig. 10.61

Fig. 10.60

Tube wastage as a function of tube metal temperature after 125,000 h, with low wastage rate by oxidation and accelerated wastage rates by molten salt corrosion. Source: Ref 15

of laboratory tests with ﬂowing synthetic ﬂue gas (N2-15CO2-3.6O2-0.25SO2) over a mixture of potassium sulfate, sodium sulfate, and iron oxide (molecular ratio 1.5:1.5:1.0) that covered the test coupons. Samples were exposed at different temperatures for 5 days. Results are summarized in Fig. 10.63 (Ref 69). The corrosion rate was greatly enhanced when the sulfate was molten. Corrosion rate also depends on SO2 concentration in the ﬂue gas and Na2SO4 + K2SO4 concentration in the ash deposit, as illustrated in Fig. 10.64 and 10.65 (Ref 70). For alkali sulfates, only the acid-soluble ones are of concern in coalash corrosion (Ref 71).

The typical corroded superheater or reheater tube is characterized by two wastage ﬂats at about the 2 o’clock and 10 o’clock positions, as shown schematically in Fig. 10.66 (Ref 13). At these two locations, where the ash was thin and the heat ﬂux was high, a molten sulfate layer formed, resulting in severe corrosion attack (Ref 13). On the other hand, the front face of the tube, with a heavy ash deposit providing sufﬁcient insulation to keep the metal surface temperature below the melting point of the sulfate, suffered signiﬁcantly less corrosion attack (Ref 13). Figure 10.67 shows the cross section of a Type 304H reheater tube after service for about 4 years with the reheated steam at about 538 °C (1000 °F) suffering coal-ash corrosion attack in a subcritical boiler. The locations that suffered the

Alloy 622 overlay (dye penetrant tested) after 4 years of service involving the use of waterlances for deslagging. The overlay was applied onto the carbon steel waterwall after thermal fatigue cracks caused by waterlances were ground off. The dye penetrant testing showed no cracking. Courtesy of Welding Services Inc.

pg 296

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

Coal-Fired Boilers / 297

worst coal-ash corrosion attack were at approximately the 2 and 10 o’clock positions (when the ﬂue gas impinges at 12 o’clock position). The boiler used staged combustion and overﬁre air for NOx reduction control. The surface appearance of the maximum wastage area is shown in Fig. 10.68. SEM/EDX analysis was performed on the corrosion scale formed on the maximum wastage area with the results being summarized in Fig. 10.69 (Ref 14). The analysis shows sulﬁdation attack along with oxidation.

Fig. 10.64

Effect of SO2 content in flue gas on the corrosion of several superheater/reheater materials exposed to synthetic ash containing 5 wt% (Na2SO4 + K2SO4) at 650 °C (1200 °F). Source: Ref 70

Fig. 10.62

Melting point curve for the K3Fe(SO4)3-Na3Fe (SO4)3 system. Source: Ref 68

Fig. 10.65

Effect of Na2SO4 + K2SO4 content in synthetic ash on the corrosion of several superheater/reheater materials at 650 °C (1200 °F) in flue gas containing 0.25% SO2. Source: Ref 70

Fig. 10.63

Results of laboratory tests with flowing synthetic flue gas (N2-15CO2-3.6O2-0.25SO2) over a synthetic coal ash (K2SO4, Na2SO4, and Fe2O3 with a molecular ratio of 1.5:1.5:1.0) that covered the test coupons. Source: Ref 69

Fig. 10.66 Ref 13

Typical wastage feature of a corroded superheater and reheater tube from a coal-fired boiler. Source:

pg 297

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

298 / High-Temperature Corrosion and Materials Applications

It was further observed that carburization occurred at the locations (i.e., about 2 and 10 o’clock positions) where coal-ash corrosion took place, as shown in Fig. 10.70. However, the

Fig. 10.67

Cross section of a Type 304H reheater tube showing two wastage flats with the maximum wastage. Note the wastage flats on both sides of the tube surface where the flue gas impinging at the 90° location (i.e., facing the ruler in the photo). Courtesy of Welding Services Inc.

Fig. 10.68

surface that did not suffer coal-ash corrosion attack, such as at the 3 or 9 o’clock position, showed no carburization (Fig. 10.71). It is believed that carburization was not the cause that resulted in coal ash corrosion because of the formation of chromium carbides that reduced the chromium level in the metal matrix. On the contrary, coal-ash corrosion (or sulﬁdation) caused the metal surface to form sulﬁdes along with unprotective chromium oxide scales, thus resulting in carburization when the localized environment was under reducing conditions with the presence of CO and/or unburned carbon soot. A superheater tube made of Type 304 suffering coal-ash corrosion attack is shown in Fig. 10.72. Similar to the reheater tube discussed earlier (Fig. 10.67), the worst attack occurred at the 2 and 10 o’clock positions (when the ﬂue gas was impinging at the 12 o’clock position). The morphology of corrosion attack at the worst corrosion attack area is shown in Fig. 10.73, showing general wastage and some internal attack in the matrix as well as intergranular attack. Figure 10.74 shows a scanning electron micrograph (backscattered electron image), showing both general corrosion scales and internal attack. The EDX analysis of the corrosion products indicated sulﬁdes (marked as 1) and internal oxide phases (marked as 2). Area 1 area was found to contain primarily 45.5%

Surface appearance of one of the wastage flats with the maximum wastage of the 304H reheater tube shown in Fig. 10.67. Courtesy of Welding Services Inc.

pg 298

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

Cr, 33.5% Fe, 9.2% Ni, and 8.6% S (in wt%) and area 2 primarily 56.8% Fe, 27.5% Cr, and 11.7% Ni. In the same boiler, Type 304H reheater tubes suffered maximum wastage rates at about 30° locations on either side of the direct ﬂue gas impingement position, as shown in Fig. 10.75. The wastage pattern in this case is believed to be caused by ﬂy-ash induced erosion/corrosion instead of coal-ash corrosion. Figure 10.76 shows the cross section of the maximum wastage area where ﬂy-ash deposits were in contact with the metal surface. The EDX analysis of the ﬂyash deposits, which are marked as 1, showed primarily 46.4% Si, 21.6% Al, and 20.7% Fe. Impinging ﬂy-ash particles removed oxide scales from the metal surface that prompted the fresh metal surface to form oxides again, which were

21 µm

Fig. 10.69

Scanning electron micrograph (backscattered electron image) showing the corrosion products formed on the maximum wastage area of Type 304H reheater shown in Fig. 10.67. Semiquantative EDX analysis shows the compositions (wt%) at different locations as indicated below. Courtesy of Welding Services Inc. 1: 34.4% Fe, 49.1% Cr, 5.6% Ni, 4.9% S, 3.7% Mn, 1.3% Si, and trace elements 2: 34.7% Fe, 17.6% Cr, 22.8% Ni, 22.3% S, 1.6% Mn, 0.6% Si, and trace elements 3: 32.1% Fe, 45.5% Cr, 6.8% Ni, 7.6% S, 4.5% Mn, 0.9% Si, 0.7% Na, and trace elements 4: 28.9% Fe, 49.0% Cr, 6.3% Ni, 6.6% S, 5.0% Mn, and trace elements 5: 33.7% Fe, 17.6% Cr, 22.4% Ni, 23.1% S, and trace elements 6: 32.4% Fe, 49.1% Cr, 5.0% Ni, 5.7% S, 6.1% Mn, and trace elements 7: 37.7% Fe, 48.3% Cr, 1.8% Ni, 6.8% S, and trace elements 8: 47.2% Fe, 40.4% Cr, 4.0% Ni, 4.8% S, 2.4% Mn, and trace elements

Coal-Fired Boilers / 299

again removed by subsequent impinging ﬂy-ash particles. This repeated oxide removal and reformation process caused the wastage rate to be increased due to erosion-enhanced corrosion reaction. (Detailed discussion of erosion/corrosion is covered in Chapter 8) Some internal attack was observed underneath the ﬂyash deposits. Internal attack was essentially in form of oxides along grain boundaries (marked 2, 3, and 4). Trace of chlorine was detected in the phase marked by 4. In adjacent area, which was still on the ﬂat wasted face (but near the area shown in Fig. 10.76) relatively thin oxide scales were observed. These oxide scales appeared to be fragmented and fractured possibly by impinging ﬂyash (Fig. 10.77). The EDX analysis showed presence of sulfur in these oxides. These oxides were likely the subsequently “reformed” oxides that formed after preceding oxides were removed by impinging ﬂy-ash particles. At the position where the ﬂue gas directly impinging on the tube in the same 304H reheater tube sample, signiﬁcant amount of ash deposits was observed, as shown in Fig. 10.78. At this location, the tube wall wastage was signiﬁcantly less. The EDX analysis was performed in the areas in (a) the top portion of the deposits (Fig. 10.79), (b) midsection of the deposits/ corrosion products (Fig. 10.80), and (c) the deposits/corrosion products near and at the interface (Fig. 10.81). The top layer was found to be essentially ash deposits enriched in aluminum, silicon, iron, sulfur and arsenic, and calcium. The midsection also consisted of ash deposits enriched primarily in silicon, arsenic and iron. Corrosion products formed on the tube were found to consist of essentially Cr-Fe-rich sulﬁdes and Fe-Ni-Cr-rich sulﬁdes. Some internal sulﬁdes were also observed. Many investigators have been using laboratory simulation tests involving synthetic ash (typically Na2SO4, K2SO4, and Fe2O3) to cover the test specimens with a ﬂowing synthetic ﬂue gas (typically N2-O2-CO2-H2O-SO2) in a test retort to rank alloy performance for resistance to coal-ash corrosion. Figure 10.82 shows the results of laboratory coal-ash corrosion tests involving various austenitic stainless steels, Fe-Ni-Cr alloy 800H, and “50Ni-50Cr” alloy 671 (Ref 72). The best performer was the alloy with the highest chromium content (alloy 671). Resistance to coal-ash corrosion has been found to increase with increasing chromium content in the alloy. Castello et al. (Ref 73) summarized the coal-ash corrosion data generated by various

pg 299

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

300 / High-Temperature Corrosion and Materials Applications

investigators to show the beneﬁcial effect of chromium in resisting coal-ash corrosion in Fig. 10.83. Kihara et al. (Ref 78) also presented their laboratory simulated coal-ash corrosion tests as a function of chromium contents in the alloys in Fig. 10.84. Synthetic ash used in laboratory testing consisted of 2.5, 5, and 20% Na2SO4 + K2SO4 (equal amounts) with Fe2O3, Al2O3 and SiO2 (equal parts). Tests were performed at 600, 650, and 700 °C (1112, 1200, and 1292 °F) for times up to 300 h in a synthetic ﬂue gas containing 0.05, 0.1, 0.25, and 1.0% SO2. Plant exposure data were also included in the ﬁgure. Plant exposure data were obtained using corrosion probes inserted into the operating boiler at Tennessee Valley Authority’s Gallatin Station Unit No. 2. Alloys tested included 17-14 CuMo (17Cr-14Ni austenitic stainless steel), Type 347, 25Cr-20Ni-Nb, 21Cr-11Ni-Si-Ce austenitic stainless steel, 20Cr-18Ni-Si-Al austenitic stainless steel, alloy 800, 30Cr-45Ni-2Mo, and

Fig. 10.70 Services Inc.

chromized T91. During the 16,000 h exposure, the boiler burned coal containing 3.08 to 3.26% S, 8 to 11% ash. Field data showed signiﬁcant scattering with alloys containing about 20% Cr. Alloys containing about 25% Cr and higher were found to perform signiﬁcantly better with less data scattering. At TVA’s Gallatin Station, previous 304 and 321 stainless steel reheaters in Unit No. 2 (a subcritical boiler) typically required replacement after 7 years of service (Ref 79). The reheater made of higher chromium-containing alloy HR3C (25Cr-20Ni-0.5Nb-0.25N) had been in service for about 45,004 h (approximately 5 years) without causing any forced outage due to corrosion-induced tube wastage problem (Ref 79). In order to develop coal-ash corrosion data for superheaters and reheaters at higher steam temperatures in ultra-supercritical boilers, a ﬁeld exposure test program was initiated that involved using tube sleeves to increase metal temperatures. Samples were exposed at 521 to

Optical micrograph showing Type 304H reheater (the same one shown in Fig. 10.67) that suffered carburization on the surface where severe coal-ash corrosion took place (about 2 and 10 o’clock positions). Courtesy of Welding

pg 300

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

Fig. 10.71

Coal-Fired Boilers / 301

Type 304H reheater (the same one shown in Fig 10.67) at 3 or 9 o’clock positions, where no severe coal-ash corrosion attack occurred, showing no carburization attack. Courtesy of Welding Services Inc.

685 °C (970 to 1265 °F) for about 45,004 h. Blough et al. (Ref 79) summarized the test results on the tube shields (higher metal temperatures) in Fig. 10.85, showing metal wastage rates as a function of metal temperature. Two highest chromium-containing alloys, HR3C (25Cr), and CR30A (30 Cr), performed the best. The data in Fig. 10.85 show what appeared to be low wastage rates for chromized T22 at 585 °C (1085 °F). In fact, the chromizing layer with about 0.33 mm (0.013 in.) in thickness was found to be completely corroded at certain locations during the test duration, and the corrosion had penetrated into the substrate T22 steel (Ref 79). Thus, the authors concluded that chromizing layer can only offer temporary protection to the substrate steel (Ref 79). A long-term plant exposure program to determine suitable superheater and reheater alloys against coal-ash corrosion for applications in ultra-supercritical boilers was conducted by Babcock & Wilcox (B&W), the U.S. Department of Energy (DOE), and the Ohio Coal Development Ofﬁce (OCDO) under the “Coal Ash

Fig. 10.72

Cross section of a superheater tube made of Type 304SS suffering coal-ash corrosion attack. Note the wastage flats on both sides of the tube surface where the flue gas impinged at the 90° location (i.e., facing the ruler in the photo). Courtesy of Welding Services Inc.

pg 301

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

302 / High-Temperature Corrosion and Materials Applications

Corrosion Resistant Materials Testing Program.” The results of the test program were published by McDonald (Ref 80) and McDonald and Robitz (Ref 81). Plant exposure tests were conducted in a B&W 120 MW(e) cyclone-ﬁred boiler (a subcritical unit burning a 3 to 3.5% S Ohio coal) at Reliant Energy’s plant in Niles, Ohio. Test sections, which were cooled by 600 °F/315 psi reheat steam, were located within the superheater bank. Figure 10.86 shows cross sections of SAVE 25 (Sumitomo’s 25Cr-20Ni steel) tube specimens tested at three different temperatures, showing increased wastage rates with increasing exposure temperature. The results summarizing the performance of various alloys are presented in Fig. 10.87. The alloys that exhibit acceptable wastage rates were found to be alloy 671 cladding and alloy 72 weld overlay. Both alloys 671 and 72 were nickel alloys with very high chromium contents (47% Cr for former and 44% Cr for the latter). They were both used as a cladding. Alloy 671 clad tubing was produced by coextrusion method, while alloy 72 weld

Fig. 10.73

overlay tube was produced by spiral overlay welding method using gas metal arc welding (GMAW) process. The Central Electricity Generating Board (CEGB) in the United Kingdom tested and tried co-extruded tubing with a high chromium alloy as a cladding for superheaters and reheaters in 1970s and 1980s. Latham et al. (Ref 82) reported test trials of Type 310/Esshete 1250 co-extruded tubes and 671/800H co-extruded tubes in CEGB boilers. One test trial was conducted in a 550 MW(e) boiler burning coal with 0.45% Cl. The test tubes were welded to the hottest section of the reheater with maximum tube metal temperature of approximately 650 °C (1200 °F), and ﬂue gas temperature of about 1150 °C (2100 °F). After 15,400 h of exposure, 671/800H coextruded tubes along with Type 316 and 347 tubes were removed for evaluation. In addition, Type 310/E1250 and 671/800H coextruded tubes were removed after 18,000 h of exposure, and 671/800H co-extruded tubes were removed after 34,000 h of exposure. The results

Optical micrograph showing the morphology of the corrosion attack on the 304H superheater tube at the severely wasted area. Courtesy of Welding Services Inc.

pg 302

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

in terms of wastage rates (nm/h or mpy) are summarized in Table 10.10. In another test conducted in a 275 MW(e) boiler burning coal with less than 0.15% Cl, a complete ﬁnal reheater was retubed with

pg 303

Coal-Fired Boilers / 303

Type 310/E1250 co-extruded tubes. After 17,500 and 29,000 h of exposure, Type 310/E1250 co-extruded tubes showed a wastage rate of only 6 nm/h (1.9 mpy) versus a wastage rate of about

30 µm 11 µm

Fig. 10.76 Fig. 10.74

Scanning electron micrograph showing the corrosion products (general corrosion scales and internal corrosion phases) on Type 304H superheater at the severely wasted area. The EDX analysis showed that the corrosion product (area 1) was Cr-Fe-rich sulfide (45.5Cr-33.5Fe-9.2Ni8.6S), and the internal corrosion phase (area 2) was Fe-Cr-rich oxide (56.8Fe-27.5Cr-11.7Ni). Area 3 is the alloy matrix. Courtesy of Welding Services Inc.

Scanning electron micrograph (backscattered electron image) showing fly-ash deposits (46.4 Si-21.6Al-20.7Fe) (marked 1) on the surface of Type 304H reheater (Fig. 10.75) that suffered the maximum wastage at location 30° away from the direct flue gas impingement point. The 304H also suffered intergranular corrosion attack (areas 2–4). Courtesy of Welding Services Inc.

50 µm

Fig. 10.75

Cross section of a Type 304H reheater tube suffering fly-ash erosion/corrosion damage. Courtesy of Welding Services Inc.

Fig. 10.77

Scanning electron micrograph showing the oxide scales formed on the nearby location of the one shown in Fig. 10.76 on the severely wasted area for Type 304H reheater tube. Courtesy of Welding Services Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

pg 304

304 / High-Temperature Corrosion and Materials Applications

50 nm/h (16 mpy) for previous Type 347 tubes. A ﬁnal superheater was retubed with Type 310/ E1250 co-extruded tubes for a 500 MW(e) boiler burning a 0.3% Cl coal, Type 310/E1250 coextruded tubes showed maximum corrosion rate of about 50 to 60 nm/h (16 to 19 mpy) versus maximum corrosion rates of about 100 nm/h (32 mpy) for the original Type 316 tubes. Experiences in the United Kingdom with coextruded tubes for superheaters and reheaters were summarized in a 1987 paper by Flatley et al. (Ref 83). Flatley et al. (Ref 83) indicated that boilers in several major power stations retubed the hottest sections of superheaters and reheaters with Type 310/E1250 co-extruded tubes with excellent results. In one installation, however, some of the hotter leading tubes were less satisfactory in performance advantage of the 310/ E1250 co-extruded tubes over the original Type 316 tubes with wastage rates being about 50 nm/ h (16 mpy) for Type 310 cladding versus about 70 to 100 nm/h (22 to 32 mpy) for the original Type 316 tubes. For alloy 671/800H co-extruded tubes installed in a complete reheater in a 500 MW(e) boiler burning high chlorine (0.5%) coal, the alloy 671 cladding showed a wastage rate of less than 0.1 mm/yr (4 mpy) after 60,000 h of operation as opposed to a wastage

11 µm

Fig. 10.79

Scanning electron micrograph showing only the top portion of the ash deposits/corrosion scale shown in Fig. 10.78 at higher magnification. The gray and white phases were analyzed by EDX with the results (in wt%) summarized below. Courtesy of Welding Services Inc. 1: 30.2% Fe, 21.0% As, 14.1% Si, 11.0% Al, 8.7% Ca, 9.7% S, 2.3% P, and residual elements 2: 80.7% Fe, 11.1% Si, 2.1% Al, 1.5% As, and residual elements

11 µm

30 µm

Fig. 10.78

Scanning electron micrograph showing ash deposits that formed at the 12 o’clock position where the flue gas made a direct impingement on the Type 304H reheater tube (Fig. 10.75). This location showed significantly less tube wall wastage. Courtesy of Welding Services Inc.

Fig. 10.80

Scanning electron micrograph showing the middle section of the ash deposits/corrosion scale shown in Fig. 10.78 at higher magnification. The dark gray (3) and light gray (4) phases were analyzed by EDX with the results (in wt %) summarized below. Courtesy of Welding Services Inc. 3: 32.8% Si, 26.1% As, 22.6% Fe, 9.6% Al, 2.4% P, 1.5% S, and residual elements 4: 93.8% Fe, 2.4% Cr, and residual elements

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

rate of about 2 mm/yr (79 mpy) for the original Type 347 tubes. Some of the performance experiences for this 671/800H co-extruded tubing have been reported by Fahrmann and Smith (Ref 84) and Kiser and Orsini (Ref 85). Fahrmann and Smith (Ref 84) reported the results of the metallurgical evaluation of two superheater tubes and one reheater tube after 18 years of service in a utility coal-ﬁred boiler. The alloy 671 cladding in the reheater tube was found to exhibit very little corrosion attack, with the cladding thickness very close to that of the original as-fabricated tube. For two superheater tubes, the alloy 671 cladding generally showed good conditions except at few locations where pitting attack was observed. In one case, the pitting attack was almost penetrated to the entire cladding thickness. Even at this location, the corrosion rate would still be very low (less than 0.1 mm/yr, or 4 mpy, assuming the original cladding thickness being 2 mm, or 80 mils). In late 1990s, a spiral overlay welding technology involving a tandem gas metal arc and

pg 305

Coal-Fired Boilers / 305

gas tungsten arc welding method was developed for manufacture of weld overlay bimetallic tubing for superheaters and reheaters (Ref 86). Two spiral weld overlay tubes tested in the “Coal

-

Fig. 10.82

Results of laboratory tests conducted in synthetic flue gas (80N2-15CO2-4O2-1SO2, saturated with H2O) with synthetic ash (37.5 mol% Na2SO4, 37.5 mol% K2SO4, and 25 mol% Fe2O3) covering the samples. Exposure time was 50 h. Source: Ref 72

11 µm

Fig. 10.81

Scanning electron micrograph showing the bottom portion of the corrosion products shown in Fig. 10.78 at higher magnification. The phases (5–8) were analyzed by EDX with the results (in wt%) summarized below. Courtesy of Welding Services Inc. 5: 47.2% Cr, 35.9% Fe, 13.0% S, 2.5% Mn, and residual elements 6: 54.2% Cr, 31.3% Fe, 5.7% S, 3.7% Mn, 2.8% Ni, and residual elements 7: 45.9% Fe, 27.5% Ni, 15.9% Cr, 6.4% S, 3.7% Si, and residual elements 8: 50.9% Fe, 31.9% Ni, 10.3% Cr, 5.4% S, and residual elements

Fig. 10.83

Metal loss as a function of chromium contents in the alloys generated by various investigators in laboratory coal-ash corrosion tests as well as plant exposure. Source: Ref 73. Note: “This work” by Castello et al. in Lab tests: 10% alkali, 1% SO2, 700 °C (Ref 73); Plumley et al. (Ref 74): Plant exposure, 677–727 °C, 1.7% S in coal, 6% alkali in ash; Kihara et al. (Ref 75): Lab test, 7.3% alkali, 0.3% SO2, 700 °C; Van Weele et al. (Ref 76): Lab test, 10% alkali, 1% SO2, 10% H2O, 700 °C; Blough (Ref 77): Plant exposure, 538–871 °C, 0.6–3.6% S, 0.4– 0.9% Na in coal.

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

306 / High-Temperature Corrosion and Materials Applications

Ash Corrosion Resistant Materials Testing Program” (Ref 80, 81) were identiﬁed as IN72WO and IN52WO in Fig. 10.87. For the sample tube

Fig. 10.84

Metal loss as a function of chromium contents in the alloys in laboratory coal-ash corrosion tests (solid line) and plant exposure using corrosion probes inserted into the operating boiler at Tennessee Valley Authority’s Gallatin Station Unit No. 2. The vertical axis on the right-hand side is in mils. Source: Ref 78

Fig. 10.85

identiﬁed as IN72WO, the weld overlay was produced using alloy 72 weld wire (AWS ERNiCr4: Ni-44Cr), while the alloy 52 weld overlay for IN52WO was produced using alloy 52 weld wire (AWS ERNiCrFe-7: Ni-30Cr-9Fe). In Fig. 10.87, alloy 72 overlay was found to perform similarly to alloy 671 cladding. Alloy 52 overlay, although not performing as well as alloy 72 overlay and alloy 671 cladding, performed signiﬁcantly better than many austenitic alloys particularly at higher temperatures. In one subcritical boiler (255 MWe) burning high-chlorine coal (about 0.3% Cl), Type 304H reheater tubes typically lasted for about 4 years. Test trials were performed for alloy 72/304H and alloy 52/304H weld overlay tubes as part of the reheater (538 °C/1000 °F outlet steam) for 3½ years. Typical tube cross sections are shown in Fig. 10.88 for Alloy 72/304H tube and Fig. 10.89 for alloy 52/304H tube. Very little tube thinning was observed for alloy 72 overlay. Slight corrosion was observed for alloy 52 overlay. The maximum wastage rate was estimated to be less than 0.05 mm/yr (2 mpy) for alloy 72 and about 0.2 mm/yr (8 mpy) for alloy 52. Kiser et al. (Ref 87) indicated that, in one boiler with low NOx burners, alloy 72 overlay reheater tubes were found to show negligible wastage of about 0.076 mm (3 mils) after service for about 6 years at the metal temperature of about 649 to 677 °C (1200 to 1250 °F). The authors (Ref 87) also

Metal wastage rates as a function of metal temperature for 304HSS, 347SS, alloy 800H, alloy NF709 (20Cr-25Ni1.5Mo-0.2Nb-Fe), alloy HR3C (25Cr-20Ni-0.5Nb-0.25N-Fe), alloy CR30A (30Cr-48Ni-2Mo-0.25Al-0.25TiFe), and chromized T22 (T22Cr). The data were generated from plant exposure tests at TVA’s Gallatin Station Unit No. 2 boiler. Source: Ref 79

pg 306

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

Chapter 10:

indicated that, in another boiler, alloy 72 overlay superheater tubes showed essentially no sign of corrosion or cracking after 6 years of service in sootblower lanes (Fig. 10.90).

10.6 Erosion in Fluidized-Bed Boilers In ﬁxed-bed, ﬂuidized-bed boilers, the major high-temperature materials issues are essentially corrosion and erosion of in-bed heat-exchanger components. Corrosion is mainly sulﬁdation/ oxidation. In general, the 300 series austenitic

pg 307

Coal-Fired Boilers / 307

stainless steels have adequate resistance, since the metal temperatures for these heat exchangers normally do not exceed 650 °C (1200 °F) (Ref 88–90). Materials wastage due to erosion is a major issue for the in-bed heat exchangers. Stringer (Ref 88, 89, 91), Stallings and Stringer (Ref 92), and Rademakers et al. (Ref 90) provided excellent reviews on erosion related to inbed heat exchangers. The wastage problem for the in-bed heat exchanger appears to be a low-temperature phenomenon. Rademakers et al. (Ref 90) showed that the tube wastage rates were high at low temperatures and decreased with increasing temperature, as shown in Fig. 10.91. Stringer

Table 10.10 Results of field tests on austenitic stainless steel tubes and co-extruded tubes in the reheater of a 500 MW (e) boiler Wastage rate Material

Fig. 10.86

Tube cross-section wastage profile for alloy Save 25 (Sumitomo’s 25Cr-20Ni steel, similar to ASME SA213TP310H or SA213TP310HCbN) after exposure to the temperature for about 15.5 months. Wastage rates in each case are also indicated. Source: Ref 81

Fig. 10.87

316 347 310/E 1250 671/800H

Exposure, h

15,400 15,400 18,000 34,000

nm/h

mpy

160 200 60

51 64 19 Negligable

Source: Ref 82

Tube metal wastage rates as a function of surface metal temperature for various alloys tested at Reliant Energy’s Niles plant under the “Coal Ash Corrosion Resistant Materials Testing Program” conducted jointly by B&W, DOE, and the Ohio Coal Development Office. Source: Ref 80

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:04PM Plate # 0

308 / High-Temperature Corrosion and Materials Applications

(Ref 89) indicated that this tube wastage does not appear to be a problem above approximately 500 °C (932 °F). Wastage has taken place on the in-bed evaporator tubes, but seldom occurred

on the in-bed superheater tubes (Ref 92). This may suggest a temperature effect. The wastage damage is commonly referred to be that of erosion. Stringer and Wright (Ref 88), Stringer (Ref 89), and Stallings and Stringer (Ref 92) considered that mechanical processes, such as erosion and abrasion wear, are involved in the damage mechanisms. Stringer (Ref 91) deﬁnes erosion and abrasion: erosion is material removal by the impact of particles moving freely before and after the impact; abrasion wear is material removal by particles that are loaded onto the surface and remain in contact for a period of time. Possible wastage mechanisms are discussed in

Fig. 10.88

The cross section of alloy 72/304H overlay tube after testing as part of the reheater (1000 °F steam) for 3½ years in a 255 MW(e) subcritical boiler, which burned high chlorine coal (about 0.3%). The 304H reheater tubes exhibited a typical 4 year life. Courtesy of Welding Services Inc.

Fig. 10.89

The cross section of alloy 52/304H overlay tube after testing as part of the reheater (538 °C, or 1000 °F steam) for 3½ years in a 255 MW(e) subcritical boiler, which burned high chlorine coal (about 0.3%), showing slight pitting attack. The 304H reheater tubes exhibited a typical 4 year life. Courtesy of Welding Services Inc.

Fig. 10.90

Alloy 72 overlay superheater tubes in a sootblower lane after 6 years of service. Original weld bead ripples are still clearly visible. Source: Ref 87

pg 308

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

Chapter 10:

detail by these authors in Ref 88, 89, 92, and 93. Stringer and Wright (Ref 88) brieﬂy summarize possible wastage mechanisms: (a) erosion by in-bed moving particles, (b) abrasion wear by “loaded” particles, (c) erosion by particles in the wake of bubbles, (d) bubbles track along vertical or sloping tubes, (e) erosion by particles thrown into the metal surface by bubble collapse, (f) erosion in the splash zone by bubble collapse at the bed surface, (g) erosion induced by in-bed jets, associated with coal or acceptor injection ports, particle recirculation ports, and so forth, and (h) erosion by “gulf stream” (long-range ﬂow patterns) in the bed. Materials wastage issues in ﬂuidized-bed boilers have been extensively reviewed in Ref 88 to 93. This section discusses the major erosion and/or abrasion problem (a) for in-bed evaporator tubes in bubbling ﬂuidized-bed boilers and (b) at the refractory/waterwall interface region in circulating ﬂuidized-bed boilers. 10.6.1 In-Bed Evaporator Tubes in Bubbling Fluidized-Bed Boilers The evaporator tubes in a bubbling ﬂuidizedbed boiler (TVA 20 MWe boiler) experienced severe wastage problems with carbon steel tubes during operation in 1986. The average wastage rates for the carbon steel (SA210 A1) tubes were about 13 mils/1000 h with some tubes as high as 35 mils/1000 h (Ref 94). A test program was subsequently conducted to evaluate several cladding and coating materials that included

Fig. 10.91

Coal-Fired Boilers / 309

Type 304/SA210 A1 co-extruded tubes, two proprietary thermal sprayed coatings, sprayed and fused WC-based coated (Extendalloy coating) tubes, and chromized T11 tubes. The thickness of Extendalloy coating was about 25 mils. The in-bed evaporator tubes were operated at approximately 343 to 371 °C (650 to 700 °F). Test tubes were removed for metallurgical evaluation after exposure of about 6500 h. The ﬁndings, as reported by Lewis et al. (Ref 94), indicated that the maximum wastage rate was 8 mils/1000 h (70 mpy) for carbon steel tubes (SA210 A1 control samples) and 95 mils/1000 h (83 mpy) for Type 304 cladding (co-extruded tubes). Extendalloy, chromized, and thermal sprayed coatings, however, showed no measurable wastage. Nevertheless, the authors observed that oxide phases had formed at the coating/ substrate interface for the thermal sprayed coatings and concluded that this could lead to coating spallation (Ref 94). As for a chromized coating, it is questionable whether such a thin coating (about 400 µm thick) could sustain a long-term performance under ﬂuidized-bed erosive conditions. In examining evaporator tube wastage experienced by the TVA 20-MW(e) pilot plant (atmospheric bubbling ﬂuidized-bed boiler) ﬁring Pyro coal during operation in 1986, Stallings and Stringer (Ref 92) concluded that the tube wastage occurred mainly at the bottom of the tube. This was because the larger, harder, more angular bed materials tend to migrate toward the bottom region of the bed where they

Tube wall wastage rates as a function of tube wall temperature from the in-bed tube tests in a 4 MW atmospheric fluidizedbed combustor (AFBC). St. 35.8 is a carbon steel. Source: Ref 90

pg 309

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

310 / High-Temperature Corrosion and Materials Applications

were picked up in the wakes of bubbles and rose upward. The tube wastage was primarily caused by the impact and motion of particles entrained in bubble wakes on the underside surface of the tubes. In a 130 MW(e) atmospheric bubbling ﬂuidized-bed boiler, the in-bed evaporator was constructed out of sprayed and fused WC-based hardfacing coated tubes. These coated tubes were in general satisfactory in performance. Nevertheless, occasional local failures did occur according to the plant operator. Repair of the localized failures could be difﬁcult for these sprayed and fused hardfacing coating. The plant operator was interested in weld overlay hardfacing materials that were weld repairable when local wear took place. A test tube consisting of test sections of different weld overlay alloys was tested as part of an evaporator tube. The steam in the evaporator tube element was about 336 °C (637 °F). The bed temperature was typically 788 to 870 °C (1450 to 1600 °F). Test duration was 14,021 operating hours. The overlay alloys tested included Type 309, Type 312, alloy 82, alloy 625, a proprietary WC-based hardfacing alloy (HF60), and a thermal sprayed coating on alloy 625 overlay (butter layer). Evaluation of test specimens after 14,021 operating hours showed that thermal sprayed coating was gone (worn away). Figure 10.92 shows spallation of the coating. Type 309, 312, alloy

0.1 mm

Fig. 10.92

Spallation of the thermal sprayed coating on the bottom surface of the test tube sample after exposure for 14,021 h as part of the in-bed evaporator tubes in a 130 MW(e) bubbling fluidized-bed boiler. The coating was applied to alloy 625 overlay, which acted as a butter layer between the coating and the substrate steel tube. Courtesy of Welding Services Inc.

82, and alloy 625 overlays were completely worn away at the underside of the tube samples. The WC-base hardfacing overlay (HF60) showed no sign of erosion or abrasive wear. The cross sections of all the test specimens are shown in Fig. 10.93 (Ref 14). The wastage was found to be the greatest at the bottom of the tube surface (6 o’clock position). Traditional stainless steels and nickel alloys are not adequate in providing erosion or abrasive wear protection. A proprietary WC-based hardfacing weld overlay showed no evidence of wastage. The overlay surface showed a polished appearance. The microstructure of this hardfacing weld overlay and the analyses of the phases are shown in Fig. 10.94. The hardness of this weld overlay was found to be in the range of 50 to 60 HRC (converted from Vickers hardness values) across the overlay. A proprietary hardfacing weld overlay (HF35), based on chromium eutectic carbides, was developed and tested in the same bubbling ﬂuidized-bed boiler. Testing was conducted on tubes in the loop section, which was historically a high wear area. Since it was a loop section of the tube, the tube was weld overlaid using a manual technique. Figure 10.95 shows one of the overlaid tubes in the loop section after exposure for close to 3 years as part of the in-bed evaporator tube bundle. The manually applied longitudinal weld beads are still visible. Two tubes in the loop section were later removed from the boiler for metallurgical evaluation. The weld overlay overall remained protective under the erosive conditions for about 3 years. There was, however, a localized wear area near the pin studded carbon steel tube end. Both tubes showed the similar locations that suffered localized wear. An in-bed superheater tube bundle located above this in-bed evaporator bundle reportedly showed no erosion or abrasion problems. The superheater was made of an austenitic stainless steel. This appears to conﬁrm earlier observation by Rademakers et al. (Ref 90) that the erosion or abrasion wear of the in-bed tubes were signiﬁcantly reduced at higher temperatures and by Stringer (Ref 89) that the in-bed tube wastage problem does not appear to be a problem above approximately 500 °C (932 °F). The plant engineer indicated that the in-bed superheater tube bundle was more tightly arranged than the in-bed evaporator tubes (Ref 14). It is not clear whether this could be a factor also.

pg 310

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

Chapter 10:

10.6.2 Waterwalls in Circulating Fluidized-Bed Boilers One serious materials issue for circulating ﬂuidized-bed boilers is the erosion or abrasion wear that occurs on the waterwall at the refractory lining-waterwall transition region (Ref 90,

95–97). It is generally believed that in a circulating ﬂuidized-bed boiler the bed particles ﬂow upward in the center of the boiler and migrate downward along the waterwall (Ref 98, 99). This downward ﬂow of particles is believed to be responsible for the erosion wear above the refractory lining. The authors suggested that the

(a)

(b)

(c)

(d)

Fig. 10.93

Coal-Fired Boilers / 311

Cross sections of test tube samples showing wear profiles for (a) Type 309 overlay tube, (b) alloy 625 overlay tube, (c) Type 312 overlay tube, and (d) HF60 hardfacing overlay tube after exposure for 14,021 h as part of the in-bed evaporator tubes in a 130 MWe bubbling fluidized-bed boiler. Courtesy of Welding Services Inc.

pg 311

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

312 / High-Temperature Corrosion and Materials Applications

particles ﬂowing downward along the waterwall were forced to change ﬂow direction at a tapered (or sloped) refractory lining forcing the particles to ﬂow from the membrane (or web) outward against the tube walls on the tapered refractory

19 µm

Fig. 10.94

Scanning electron micrograph (backscattered electron image) showing various hardface particles in the proprietary tungsten carbide based hardfacing weld overlay, HF60. The results (wt%) of semiquantative EDX analyses of various phases are summarized as: Light color phases (A, B, C, E, and F) are tungsten carbides (60–70W, 14–16Ni, 8Cr, 5Si), grayish phase (H) is also tungsten carbide ( 36W, 39Ni, 18Cr, 2.5Si), and dark phases (I and G) are matrix (70–80Ni, 10Cr, 5–10W). Courtesy of Welding Services Inc.

Fig. 10.95

surface. Slusser et al. (Ref 95) reported the waterwall erosion wear problem at the transition region in a 50 MW(e) boiler. Figure 10.96 illustrates schematically the relationship between the waterwall and the refractory lining at the transition region (Ref 95). The ﬁgure also indicates that the potentially high wear area was protected by a weld overlay. No technical information about the weld overlay alloy used was discussed in the paper. However, the weld overlay was reported to have suffered severe metal loss (Ref 95). Some of the protective methods suggested by the authors (Ref 95–97), although somewhat helpful, may not be long-term solutions. For example, installing a shelf above the refractory lining by interrupting the downward particle ﬂow was found to be helpful in reducing wastage rates at the waterwall-refractory transition region, but induced wear at the shelf location (Ref 95). Because of the interruption of the particle ﬂow direction by the tapered refractory surface, particles are then forced to ﬂow against the tube walls on the tapered refractory surface, thus resulting in high waterwall tube wear at the transition region. This is illustrated schematically in Fig. 10.97. Nevertheless, Slusser et al. (Ref 96) reported in their 1992 paper indicating that shelves (7 in. wide sheets) installed perpendicular to the waterwalls around the entire perimeter of the boiler at 2.2 m (7 ft) and 5.2 m (17 ft) above the refractory transition area were capable of reducing wastage rates from >25 mm/ yr (>1000 mpy) to a manageable rate, thus allowing reasonable operating periods between

One tube with a proprietary HF35 hardfacing weld overlay showing manually applied weld overlay beads along the tube that are still visible after exposure for approximately 3 years as part of an in-bed evaporator tube bundle in a 130 MWe bubbling fluidized-bed boiler. Courtesy of Welding Services Inc.

pg 312

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

Chapter 10:

inspections. Wastage rates at the shelves were found to be accelerated. For example, the waterwalls at the 17 ft shelf location suffered wastage rates of about 0.76 to 1.78 mm/yr (30 to 70 mpy) (Ref 96). Some other protection methods including cast erosion blocks made of Type 310 and thermal sprayed coatings of various alloys were mentioned under trials in a 1997 paper by Solomon (Ref 97). Some coatings such as thermal sprayed coatings of 50Ni-50Cr alloy and Cr2O3-base coating were found to be unacceptable in performance (Ref 97). Some success was attained in an approach involving bending the waterwall tubes out of the vertical plane and into the wall at the interface with the refractory lining making a smooth, vertical transition from the waterwall tubes to the refractory (Ref 93). 10.6.3 Erosion in the Convection Pass Tubes in Fluidized-Bed Boilers In ﬂuidized-bed boilers, ﬂue gas stream leaving the combustor and enters a cyclone, where particles are removed, before exiting to the convection pass. Typically, a superheater and an economizer are in the convection pass in the downstream of the cyclone. The ﬂue gas stream is entrained with ﬁne particles that are not removed by the cyclone. High tube metal wastages could occur at superheater tube banks and economizer tube banks (Ref 93). Erosion by ﬂy ash can be serious when the uneven distribution of ﬂy ash in the

Fig. 10.96

Waterwall in the transition region that suffers severe erosion wear immediately above the tapered refractory surface. Source: Ref 95

Coal-Fired Boilers / 313

ﬂue gas stream occurs under certain conditions. For example, excessive ﬂy-ash erosion of the economizer can occur in the backpass area when the ﬂue gas stream making the direction change and creating uneven distribution of ﬂy-ash particles due to centrifugal force (Ref 1). An example is given below on erosion or erosion/corrosion of the hanger tubes in the backpass area for the economizer of a circulating ﬂuidized-bed boiler (90 MWe) burning Eastern bituminous coal (Ref 14). The current hanger tubes are protected with tube shields, which typically last for a year. A Type 309 overlay tube and an HF35 hardfacing overlay tube were tested as part of the hanger tubes at top of the backpass. The temperature of the steam inside the tubes was reportedly about 315 °C (600 °F). The tubes were removed for metallurgical evaluation after 2 years of exposure. For both tubes, the leeward side typically exhibited some thin scales, while the windward side showed no scales. Figure 10.98 shows typical cross sections of these tube samples after 2 years of exposure. The Type 309 overlay tube showed more thinning on the overlay on the windward side of the tube (top side of the cross section in the photograph). This can be seen in tube cross sections as shown in Fig. 10.98. The metallographic cross sections showing the overlay in comparing the windward side with the leeward side of the tube for both 309 and HF35 overlay tubes, as shown in Fig. 10.99 and 10.100. The surface condition as well as the microstructure of the weld overlay at the windward side for both overlays is shown in Figs. 10.101 and 10.102. The 309 overlay exhibits austenitic structure with some delta ferrite phases, with a hardness range of about 91 to 92 HRB (converted from Vickers hardness values). The HF35 overlay exhibits

Fig. 10.97

Particles flowing outward against the waterwall tube wall from the membrane due to interruption of the downward flow of particles on a shelf. Source: Ref 96

pg 313

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

pg 314

314 / High-Temperature Corrosion and Materials Applications

chromium eutectic carbides formed along interdendritic boundaries, with a hardness range of about 35 to 40 HRC (converted from Vickers hardness values). The wastage rate was calculated to be about 0.38 mm/yr (15 mpy) for Type 309 overlay and 0.2 mm/yr (8 mpy) for HF35 hardfacing overlay. Both overlays contained similar chromium contents, about 21% Cr

for the 309 overlay and about 20% Cr for the HF35 overlay, and both should form chromium oxide scales. The reduced wastage rate exhibited by the HF35 overlay over the 309 overlay is believed to be due to its erosion resistance resulting from hardfacing particles of chromium eutectic carbides.

10.7 Summary A general description of coal-ﬁred boilers and their combustion conditions is presented. Fireside corrosion of the furnace waterwalls is discussed with particular emphasis on the materials problems associated with a staged ﬁring by substoichiometric combustion (i.e., combustion with insufﬁcient oxygen) in the lower furnace for reducing NOx emissions. As a result of the staged

Fig. 10.98

Cross sections of Type 309 overlay tube (a) and HF35 overlay tube (b) showing the wastage profile of the weld overlay after exposure of 2 years as part of the hanger tube for the economizer in a circulating fluidized-bed boiler. The top of the tube cross section was the windward side while the bottom (i.e., the ruler side) was the leeward side. Courtesy of Welding Services Inc.

(a)

0.010 in.

(b)

0.010 in.

Fig. 10.99

Type 309 overlay tube in the unaffected leeward side (a) and the windward side (b) after testing as part of hanger tubes for the economizer in a circulating fluidizedbed boiler for 2 years. The substrate steel was etched with nital to reveal the fusion boundary. Courtesy of Welding Services Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

Chapter 10:

ﬁring, sulﬁdation dominates the corrosion process for waterwall materials. The wastage rates for carbon and low alloy steels have been found to increase to an unacceptable rate for many boilers, particularly those of supercritical units. The current most widely used waterwall protection method against tube wall wastage problems involves weld overlay cladding in the boiler on the affected waterwall area with a corrosion-resistant alloy using automatic gas-metalarc welding (GMAW) process. Cladding of waterwall panels can also be applied in shop using either GMAW or laser cladding techniques, and the cladded panels are then installed in the boiler. The use of a corrosion-resistant cladding has essentially eliminated the tube wastage problems. For some supercritical units, circumferential grooving and cracking has been encountered on overlaid waterwall tubes

pg 315

Coal-Fired Boilers / 315

presumably due to overheating and preferential sulﬁdation penetration. The possible causes to circumferential grooving and cracking are discussed. Also, included in the discussion are erosioncorrosion caused by steam sootblowing and thermal fatigue cracking due to the use of water lances or water canons for deslagging. Superheater/reheater wastage problems as well as erosion-corrosion issues in convection pass tubes are discussed. Erosion, erosion-corrosion, or abrasion wear issues encountered in ﬂuidized-bed boilers including both bubbling beds and circulating beds are also discussed.

0.0010 in.

Fig. 10.101

(a)

0.010 in.

(b)

0.010 in.

Fig. 10.100

HF35 overlay tube in the unaffected leeward side (a) and the windward side (b) after testing as part of the hanger tube for the economizer in a circulating fluidized-bed boiler for 2 years. The substrate steel was etched with nital to reveal the fusion boundary. Courtesy of Welding Services Inc.

The surface of Type 309 overlay tube at the windward side after exposure for 2 years as part of the hanger tube for the economizer in a circulating fluidizedbed boiler. Microstructure of Type 309 overlay consists of delta ferrite (dark phases) in austenite. Courtesy of Welding Services Inc.

0.0010 in.

Fig. 10.102

The surface of HF35 weld overlay tube at the windward side after exposure for 2 years as part of the hanger tube for the economizer in a circulating fluidizedbed boiler. The overlay contained hardfacing chromium eutectic carbides (dark phases) formed along interdendritic boundaries. Courtesy of Welding Services Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

316 / High-Temperature Corrosion and Materials Applications

REFERENCES

1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, Barberton, OH, 1992 2. J.G. Singer, Ed., Combustion Fossil Power, 4th ed., Combustion Engineering, Inc., Windsor, CT, 1991 3. C.A. Meyer, et al., ASME Steam Tables, 5th ed., The American Society of Mechanical Engineers, 1983 4. R. Peltier and K. Wicker, PRB Coal Makes the Grade, Power, Oct 2003, p 28 5. J. Stringer, High Temperature Corrosion Problems in the Electric Power Industry and Their Solution, in High Temperature Corrosion, R.A. Rapp, Ed., Proc. Conf. (San Diego, CA, March 2–6, 1981, NACE, 1981 6. D.J. Lees and M.E. Whitehead, Microanalysis of Scales and Deposits Formed on Corroding Furnace Tubes in Coal-Fired Boilers, in Corrosion Resistant Materials for Coal Conversion Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science, London, 1982, p 63 7. K.A. Hay, High Temperature Environmental Degradation of Materials and Future Requirements in The Power Generation Industry, in Environmental Degradation of High Temperature Materials, Vol 2 (No. 3), Series 3, 1980, p 4/1 8. G.T. Cunningham and T. Webster, “Design, Manufacturing, and Testing of a Boiler Metal Surface Temperature Probe,” Paper No. IJPCG 2000-15075, presented at the 2000 International Joint Power Generation Conference (Miami Beach, FL), July 23–26, 2000 9. J. Stringer, Materials for Fossil Energy Systems—Past, Present, and Future, Paper No. 225, Corrosion/93, NACE, 1993 10. J.A. Urich and E. Kramer, Designing Solutions for Low NOx-Related Water Wall Corrosion, Conf. Proc., 1996 ASME Joint Power Generation Conference, S.M. Smouse and A. Gupta, Ed., ASME, 1996, p 25 11. M.D. Yeager, “The Effect of Low NOx Staging on Waterwall Wastage,” presented at the EPRI International Conference on Boiler Tube Failures in Fossil Plants (Nashville, TN), Nov 11–13, 1997 12. A.J.B. Cutler and E. Raask, External Corrosion in Coal-Fired Boilers: Assessment

13. 14. 15. 16. 17.

18.

19.

20.

21.

22.

23.

24.

from Laboratory Data, Corros. Sci., Vol 21 (No. 12), 1981, p 789 A.J.B. Cutler, T. Flatley, and K.A. Hay, Metall. Mater. Technol., Feb 1981, p 69 Welding Services Inc., Norcross, GA D.N. French, Metallurgical Failures in Fossil Fired Boilers, John Wiley & Sons, 1993 D.N. French, Paper No. 133, Corrosion/88, NACE, 1988 T. Flatley, E.P. Latham, and C.W. Morris, in Inst. Metallurgists Spring Residential Conference: Environmental Degradation of High Temperature Materials, Vol 2 (No. 13), Series 3, 1980 F. Clarke and C.W. Morris, Combustion Aspects of Furnace Wall Corrosion, in Corrosion Resistant Materials for Coal Conversion Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science Publishers, London, 1982, p 47 P.L. Daniel, L.D. Paul, J.M. Tanzosh, and R. Hubinger, Estimating Effects of Chlorine on Fireside Corrosion of Furnace Walls in Coal-Fired Boilers, Paper No. 138, Corrosion/88, NACE, 1988 E. Latham, D.B. Meadowcroft, and L. Pinder, The Effects of Coal Chlorine on Fireside Corrosion, Chlorine in Coal, J. Stringer and D.D. Barnerjee, Ed., Elsevier Science Publishers B.V., Amsterdam, 1991, p 225 S. Brooks and D.B. Meadowcroft, The Role of Chlorine in the Corrosion of Mild and Low Alloy Steels in Low pO2 and High pS2 Environments, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 515 S. Brooks and D.B. Meadowcroft, The Inﬂuence of Chlorine on the Corrosion of Mild and Low Alloy Steels in Sub-stoichiometric Combustion Gases, in Corrosion Resistant Materials for Coal Conversion Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science, London, 1982, p 105 T. Flatley, E.P. Latham, and C.W. Morris, Co-Extruded Tubes Improve Resistance to Fuel-Ash Corrosion in U.K. Utility Boilers, Paper No. 62, Corrosion/80, NACE, 1980 P.J. James and L.W. Pinder, The Impact of Coal Chlorine on the Fireside Corrosion Behavior of Boiler Tubing: A UK Perspective, Corrosion/97, NACE International, 1997

pg 316

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

Chapter 10:

25. D.J. Lees, “A Summary of Observations Relating Furnace Wall Fireside Corrosion to Chlorine Content of Coal,” SSD/MID/M26/ 79, 1979 26. S.F. Chou and P.L. Daniel, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985 27. K.S. Gilroy, Laboratory Evaluation of Candidate Materials for Furnace Wall Applications, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 345 28. S.C. Kung and C.F. Eckhart, Corrosion of Iron-Base Alloys in Reducing Combustion Gases, Paper No. 242, Corrosion/93, NACE, 1993 29. B. Dooley, R. Tilley, T.P. Sherlock, and C.H. Wells, “State of Knowledge Assessment for Waterwall Wastage,” presented at EPRI International Conference on Boiler Tube Failures in Fossil Plants (Nashville, TN), Nov 11–13, 1997 30. Workshop on Materials Issues Associated with Low-NOx Combustion in Fossil-Fired Boilers, Summary of Workshop held during Advanced Research and Technology Development’s Tenth Annual Conference on Fossil Energy Materials (Knoxville, TN), May 14–16, 1996, Materials & Components in Fossil Energy Applications, Department of Energy and Electric Power Research Institute, No. 123, August 1, 1996 31. C. Jones, Maladies of Low-NOx Firing Come Home to Roost, Power, Jan/Feb, 1997, p 54 32. P.J. James and L.W. Pinder, “Furnace Wall Fireside Corrosion: Taming the Beast Within,” presented at EPRI International Conference on Boiler Tube Failures in Fossil Plants (Nashville, TN), Nov 11–13, 1997 33. W.T. Bakker and S.C. Kung, Waterwall Corrosion in Coal-Fired Boilers—A New Culprit: FeS, Paper No. 246, Corrosion/ 2000, NACE International, 2000 34. W. Bakker, “Root Causes of Accelerated Waterwall Wastage in Coal-Fired Boilers,” presented at EPRI International Conference on Materials and Corrosion Experience for Fossil Power Plants (Isle of Palms, SC), Nov 18–21, 2003 35. W.T. Bakker, J.L. Blough, S.C. Kung, T.L. Banﬁeld, and P. Cunningham, Long

36. 37.

38.

39.

40.

41.

42.

43.

44.

45. 46.

Coal-Fired Boilers / 317

Term Testing of Protective Coatings and Weld Overlays in a Supercritical Boiler, Retroﬁtted with Low NOx Burners, Paper No. 2384, Corrosion/2002, NACE International, 2002 P. Hulsizer, Paper No. 246, Corrosion/91, NACE, 1991 G. Lai, P. Hulsizer, and R. Lee, “Waterwall Wastage Mitigation for Coal-Fired Boilers Using Automatic Pulse Spray GMAW Overlay Technology,” presented at 1999 EPRI Fossil Plant Maintenance Conference (Atlanta, GA), 1999 G. Lai and P. Hulsizer, Corrosion & Erosion/ Corrosion Protection by Modern Weld Overlays in Low NOx Coal-Fired Boilers, Paper No. 258, Corrosion/2000, NACE International, 2000 G.Y. Lai, “Performance of Automatic GMAW Overlays for Waterwall Protection in Coal-Fired Boilers,” presented at EPRI ﬁfth International Conference on Welding and Repair Technology for Power Plants (Point Clear, AL), June 26–28, 2002 G.Y. Lai, Fireside Corrosion and Erosion/ Corrosion Protection in Coal-Fired Boilers, Paper No. 4522, Corrosion/2004, NACE International, 2004 M.S. Brennan and R.C. Gassmann, “Laser Cladding of Nickel- and Iron-Base Alloys on Boiler Pressure Panels and Tubes,” presented at the EPRI Third International Conference on Welding and Repair Technology for Power Plants (Scottsdale, AZ), June 9–12, 1998 M.S. Brennan and R.C. Gassmann, Laser Cladding of Nickel and Iron Base Alloys on Boiler Waterwall Panels and Tubes, Paper No. 235, Corrosion/2000, NACE International, 2000 A.J. Bonnington and M.S. Brennan, “Type 312 Stainless Steel Laser Cladding for Waterwalls in Supercritical Units,” presented at the EPRI/DOE Conference on Advances in Life Assessment and Optimization of Fossil Power Plants (Orlando, FL) March 11–13, 2002 “Boiler Tube Failure Metallurgical Guide, Volume 2: Appendices,” EPRI TR-102433V2, Electric Power Research Institute, Oct 1993 “INCONEL alloy 625,” Inco Alloys International Huntington, WV “HASTELLOY C-22 Alloy,” H-2019D, Haynes International Inc., Kokomo, IN

pg 317

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

318 / High-Temperature Corrosion and Materials Applications

47. K. Coleman and D. Gandy, “Corrosion Resistant Waterwall Overlays Selection of Alternative Materials,” presented at International Conference on Boiler Tube Failures and HRSG Tube Failures, and Inspections (Phoenix, AZ), Nov 6–8, 2001 48. G. Lai, The Microstructure, “Properties and Corrosion Resistance of Type 312 Overlay in Batch Digesters,” Conf. Proc., 2002 TAPPI Fall Technical Conference (San Diego, CA), Sept 8–11, 2002 49. G.Y. Lai, presented at WSI Boiler Tube Overlay Forum (Atlanta, GA) July 10–12, 2006 50. J.L. Blough, private communication, First Energy, Mayﬁeld Village, OH, Sept 29, 2006 51. I.G. Wright, Hot Corrosion in Coal- and OilFired Boilers, Metals Handbook, Vol 13, 9th ed., Corrosion, ASM International, 1987, p 995 52. S. French, K. Rumbaugh, and P.N. Hulsizer, Fireside Corrosion-Erosion Mitigation via the Application of Weld Metal Overlay, Proceedings: Welding and Repair Technology for Power Plants, EPRI, Charlotte, NC 1997, p 477 53. A.J. Bonnington and T.M. Cullen, “Performance of Chromized Waterwall Panels in Supercritical Units,” presented at EPRI International Conference on Boiler Tube Failures in Fossil Plants, (Nashville, TN), Nov 11–13, 1997 54. K. Luer, J. DuPont, A. Marder, and C. Skelonis, Corrosion Fatigue of Alloy 625 Weld Claddings in Combustion Environments, Mater. High Temp., Vol 18 (No. 1), 2001, p 11 55. K. Stein, V. Guttmann, and W.T. Bakker, The Inﬂuence of Deformation on High Temperature Corrosion of CRONIFER 45TM, in Heat-Resistant Materials II, Conf. Proc. of the Second International Conference on Heat-Resistant Materials, K. Natesan, P. Ganesan, G. Lai, Ed., ASM International, 1995, p 367 56. V. Guttmann, K. Stein, and W.T. Bakker, Deformation-Corrosion Interactions in Selected Advanced High Temperature Alloys, Mater. High Temp., Vol 14 (No. 2/3), 1997, p 61 57. P. Castello, V. Guttmann, N. Farr, and G. Smith, Simulated Coal Ash Corrosion of Ni-Based Alloys, Mater. High Temp., Vol 19 (No. 1), 2002, p 29

58. M.F. Stroosnijder, V. Guttmann, and R.J.N. Gommans, Inﬂuence of Creep Deformation on the Corrosion Behavior of a CeO2 Surface-Modiﬁed Alloy 800H in a Sulﬁdizing-Oxidizing-Carburizing Environment, Mater. Sci. Eng., Vol A121, 1989, p 581 59. J.L. Coze, et al., The Development of HighTemperature Corrosion-Resistant Aluminum-Containing Ferritic Steels, Mater. Sci. Eng., Vol A120, 1989, p 293 60. E.C. Lewis and A.L. Plumley, Chromizing for Combating Fireside Corrosion, in Advances in Materials Technology for Fossil Power Plants, R. Viswanathan and R.I. Jaffee, Ed., ASM International, 1987, p 291 61. H.M. Tawancy, Structure & Properties of High Temperature Alloys: Applications of Analytical Electron Microscopy, King Fahd University of Petroleum & Minerals, Dhahran, Saudi Arabia, 1993 62. L. Paul, G. Clark, and A. Ossenberg-Engels, “Protection of Waterwall Tubes from Corrosion in Low NOx Coal Fired Boilers,” presented at Coal-Gen Conference (Cincinnati, OH), Aug 16–18, 2006 63. R.E. Kessler, “Thermal Fatigue Cracking of Waterwall Tubes from Waterlances and Water Cannons,” presented at EPRI Intelligent Sootblowing Workshop (Houston, TX), March 19–21, 2002 64. M. Carlisle, J. Sorge, and B. Mead, “Water Cannon Application at Alabama Power’s Plant Miller,” presented at EPRI Intelligent Sootblowing Workshop (Houston, TX), March 19–21, 2002 65. B. Ray, R. Hemperley, and R. Courtney, “W.A. Parish Units 7 & 8 ISB Project,” presented at EPRI Intelligent Sootblowing Workshop (Houston, TX), March 19–21, 2002 66. H.S. Blinka, “W. A. Parish Unit 5 ISB Project,” presented at EPRI Intelligent Sootblowing Workshop (Houston, TX), March 19–21, 2002 67. R.W. Borio, A.L. Plumley, and W.R. Sylvester, in Ash Deposits and Corrosion Due to Impurities in Combustion Gases, R.W. Bryers, Ed., Hemisphere Publishing/ McGraw-Hill, 1978, p 163 68. C. Cain, Jr. and W. Nelson, J. Eng. Power, Trans. ASME, Oct 1961, p 468 69. W. Nelson and C. Cain, Jr., J. Eng. Power, Trans. ASME, July 1960, p 194

pg 318

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

Chapter 10:

70. J.L. Blough and S. Kihara, Paper No. 129, Corrosion/88, NACE, 1988 71. R.W. Borio and R.P. Hensel, J. Eng. Power, Trans. ASME, Vol 94, 1972, p 142 72. J. Stringer, in, Corrosion, Vol 13, 9th ed., ASM International, 1987, p 998 73. P. Castello, V. Guttmann, N. Farr and G. Smith, Laboratory-Simulated Fuel-Ash Corrosion of Superheater Tubes in CoalFired Ultra-Supercrtical-Boilers, Mater. Corros., Vol 51, 2000, p 786 74. A.L. Plumley, J.I. Accort, and W.R. Roczniak, NACE-DOE Conference, Berkeley, CA, 1979 75. S. Kihara, K. Nakagawa, A. Ohtomo, H. Aoki, and S. Ando, Simulated Test Results for Fireside Corrosion of Superheater and Reheater Tubes Operated at Advanced Steam Condition in CoalFired Boilers, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 361 76. S. Van Weele, J.L. Blough, and J.H. DeVan, Paper No. 182, Corrosion/94, NACE International, 1994 77. J.L. Blough, ORNL/Sub/93-SM401/ 01, Foster Wheeler Corp., Livingston, NJ, 1996 78. S. Kihara, J.L. Blough, W. Wolowodiuk, and W.T. Bakker, “Prediction of Corrosion Rate of Superheater Tube in Boilers Burning Various Kinds of Coals,” presented at the NACE International Conference on Life Prediction of Corrodible Structures (Kauai, HA), 1991 79. J.L. Blough, G.J. Stanko, W.T. Bakker, and J.B. Brooks, “Superheater Corrosion in Ultra-Supercritical Power Plants,” Paper No. 250, Corrosion/2000, NACE International, 2000 80. D.K. McDonald, “Coal Ash Corrosion Resistant Materials Testing Program: Evaluation of the First Section Removed in November 2001,” Babcock & Wilcox Report, Barberton, OH 81. D.K. McDonald and E.S. Robitz, “Coal Ash Corrosion Resistant Materials Testing Program: Evaluation of the Second Section Removed in August 2003,” Babcock & Wilcox Report, Barberton, OH 82. E.P. Latham, T. Flatley, and C.W. Morris, Comparative Performance of Superheated Steam Tube Materials in Pulverized Fuel Fired Plant Environments, in Corrosion

83.

84.

85.

86. 87.

88. 89.

90.

91.

92.

Coal-Fired Boilers / 319

Resistant Materials for Coal Conversion Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science Publishers, London, U.K., 1982, p 137 T. Flatley, E.P. Latham, and C.W. Morris, CEGB Experience with Co-Extruded Tubes for Superheated and Evaporative Sections of PF Fired Boilers, in Advances in Materials Technology for Fossil Power Plants, Conf. Proc., R. Viswanathan and R.I. Jaffee, Ed., ASM International, 1987, p 219 M.G. Fahrmann and G.D. Smith, Evaluation of Clad Tubing after 18 Years of Service in a Coal-Fired Utility Boiler, Paper No. 232, Corrosion/2000, NACE International, 2000 S.D. Kiser and T. Orsini, Beneﬁts of Chromium in Nickel for Corrosion Protection of Superheater and Reheater Tubes in Coal Fired Boilers, Paper No. 5455, Corrosion/ 2005, NACE International, 2005 P.N. Hulsizer, Dual Pass Weld Overlay Method and Apparatus, U.S. Patent No. 6,013,890, Jan 11, 2000 S.D. Kiser, E.B. Hinshaw, and T. Orsini, Extending the Life of Fossil Fired Boiler Tubing with Cladding of Nickel Based Alloy Materials, Paper No. 06474, Corrosion/2006, NACE International, 2006 J. Stringer and I.G. Wright, Materials Issues in Fluidized Bed Combustion, J. Mater. Energy Systems, Vol 8 (No. 3), 1986, p 319 J. Stringer, Alloys for Advanced Power Systems, Heat-Resistant Materials, Proc. of the First International Conference, (Fontana, WI), Sept 23–26, 1991, K. Natesan and D.J. Tillack, Ed., ASM International, 1991 P.L.F. Rademakers, D.M. Lloyd, and V. Regis, AFBC’s: Bubbling, Circulating and Shallow Beds, in High Temperature Materials for Power Engineering 1990, Part I, Conf. Proc. (Liege, Belgium), Sept 24–27, 1990, E. Bachelet, et al., Ed., Kluwer Academic Publishers, Dordrecht, Netherlands, 1990, p 43 J. Stringer, Erosion/Corrosion in Fluidized Bed Combustion Boilers, in Advances in Materials Technology for Fossil Power Plants, Conf. Proc., R. Viswanathan and R.I. Jaffee, Ed., ASM International, 1987, p 319 J.W. Stallings and J. Stringer, Mechanisms of In-Bed Tube Wastage in Fluidized-Bed Combustors, in Corrosion-Erosion-Wear of Materials at Elevated Temperatures, Conf.

pg 319

Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/

30/10/2007 4:05PM Plate # 0

320 / High-Temperature Corrosion and Materials Applications

93.

94.

95.

96.

Proc., A.V. Levy, Ed., NACE International, 1991, p 19-1 I.G. Wright, A Review of Experience of Wastage in Fluidized-Bed Boilers, Mater. High Temp., Vol 14 (No. 2/3), 1997, p 207 E.C. Lewis, D.A. Canonico, and R.Q. Vincent, Metal Wastage Experiences in BFBC Environments, in Corrosion-Erosion-Wear of Materials at Elevated Temperatures, Conf. Proc., A.V. Levy, Ed., NACE International, 1991, p 20-1 J.W. Slusser, A.D. Bixler, and S.P. Bartlett, Materials Experience from A Circulating Fluidized Bed Coal Combustor, Paper No. 285, Corrosion/90, NACE, 1990 J.W. Slusser, A.D. Bixler, and D.E. Thompson, Four Years of Field Performance

of A Circulating Fluidized Bed Combustor, Paper No. 138, Corrosion/92, NACE, 1992 97. N.G. Solomon, Erosion-Resistant Coatings for Fluidized-Bed Boilers, Paper No. 135, Corrosion/97, NACE International, 1997 98. M.D. Mirolli and W.P. Bauer, “A Summary of Combustion Engineering’s Programs to Control CFB Material Wastage,” presented at the EPRI-Argonne Workshop on Materials Issues in Circulating Fluidized Bed Combustors, Argonne, IL, 1989 99. J. Zhao, R. Wu, R. Senior, R. Legros, C. Brereton, J. Grace, and C. Lim, “Spatial Variation Inside a Pilot Scale Circulating Fluidized Bed Combustion Unit,” presented at the EPRI-Argonne Workshop on Materials Issues in Circulating ﬂuidized Bed Combustors, Argonne, IL, 1989

pg 320

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:44PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p321-334 DOI: 10.1361/hcma2007p321

pg 321

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 11

Oil-Fired Boilers and Furnaces 11.1 Introduction Fuel oils primarily consist of residues from distillation of crude oil in petroleum reﬁning. Sulfur in fuel oils can vary from a fraction of a percent for lighter oils to 3.5% for some residual oils (Ref 1). Two other corrosive constituents in fuel oils are vanadium and sodium. Vanadium exists in certain crude oils as an oil-soluble porphyrin complex (Ref 2). Fuel oils from some crudes contain low levels of vanadium, while certain Middle Eastern crudes and Venezuelan crudes contain high levels of vanadium with up to several hundred parts per million (ppm) (Ref 2). Sodium can originate from crudes, the neutralizer used in crude distillation during the reﬁning process, and contamination with seawater in transportation and storage (Ref 2). During combustion, compounds formed by these three constituents make up a major part of the ash that deposits on the metallic components. For example, superheater and reheater tubes in the boiler, and tube hangers and other uncooled parts in boilers or reﬁnery and petrochemical furnaces. When the metal temperature of the component reaches 540 °C (1000 °F) or higher, compounds in the ash may become molten and corrosion attack can become severe. This mode of corrosion is frequently referred to as “oil-ash corrosion” for boilers or furnaces ﬁred with fuel oils.

11.2 Oil-Ash Corrosion During combustion of the fuel oils that contain vanadium, sulfur, and sodium, low-melting compounds can form. These compounds are mixtures of vanadium pentoxide (V2O5), sodium oxide (Na2O), or sodium sulfate (Na2SO4) (Ref 3). Signiﬁcant amounts of vanadium, sodium, and sulfur were observed in the deposits collected from superheater/reheater tubes, or from the uncooled tube hangers for the superheater/ reheater tubes in boilers. The analysis of the

deposits collected from corrosion probes at metal temperatures of 475 to 675 °C (890 to 1250 °F) from a utility boiler at Marchwood Power Station (United Kingdom) is shown in Table 11.1 (Ref 4). Table 11.2 shows the results of the analysis of the deposits from another utility boiler burning Buncker “C” fuel oil (Ref 5). Ash deposits formed on metallic components whether cooled (e.g., superheaters/reheaters) or uncooled (e.g., tube hangers) can exhibit low melting points. Table 11.3 lists a number of oil-ash constituents and their melting points (Ref 6). It is clear from the table that the melting point of the ash salt

Table 11.1 Analysis of deposits in terms of percent water soluble collected from the corrosion probe at metal temperatures of 475–675 °C (890–1250 °F) at 0.4% boiler oxygen in a utility boiler fired with fuel oil containing 2.65% S, 49 ppm V, and 44 ppm Na Content, % Constituent

S as water soluble SO3 Na as water soluble Na2O V as water soluble V2O5 V as water insoluble V2O5 Fe as water insoluble Fe2O3 Ni as water insoluble NiO Other metal oxides

1-h exposure

5-h exposure

43.4 31.8 10.9 1.0 2.6 1.3 9.0

44.1 32.6 6.8 4.1 4.5 2.0 5.9

Source: Ref 4

Table 11.2 Analysis of deposits/scales removed from the test specimens exposed in the superheater area at temperatures from 565 to 850 °C (1050 to 1560 °F) in a utility boiler Constituent

Si as silicon oxide Al as aluminum oxide V as vanadium pentoxide S as sulfuric anhydride Ni as nickel oxide Na as sodium oxide Mg as magnesium oxide P as phosphate anhydride Source: Ref 5

Content, wt %

7.8 2.0 36.37 32.80 2.16 5.12 2.35 0.5

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:44PM Plate # 0

322 / High-Temperature Corrosion and Materials Applications

deposit can vary widely, depending on composition. The phase diagram for V2O5-Na2O system showing a series of low-melting eutectic compounds is shown in Fig. 11.1 (Ref 7). Lowmelting compounds in the V2O5-Na2SO4 system are shown in Fig. 11.2 (Ref 7). Some of the eutectics become molten at 538 °C (1000 °F) or even lower. Table 11.3 Melting points of some oil-ash constituents Compound

Melting point, °C (°F)

Ferric oxide, Fe2O3 Ferric sulfate, Fe2(SO4)3

1565 (2850) Decomposes at 480 (895) to Fe2O3 2500 (4530) Decomposes at 1125 (2060) to MgO 2090 (3795) Decomposes at 840 (1545) to NiO 1720 (3130) 880 (1615) 250 (480) 400 (750) 1970 (3580) 1970 (3570) 675 (1250) 630 (1165) 640 (1185) 850 (1560) >900 (>1650) >900 (>1650) 860 (1580) 855 (1570) 625 (1160)

Magnesium oxide, MgO Magnesium sulfate, MgSO4 Nickel oxide, NiO Nickel sulfate, NiSO4 Silicon oxide, SiO2 Sodium sulfate, Na2SO4 Sodium bisulfate, NaHSO4 Sodium pyrosulfate, Na2S2O7 Vanadium trioxide, V2O3 Vanadium tetraoxide, V2O4 Vanadium pentoxide, V2O5 Sodium metavanadate, Na2O.V2O5 Sodium pyrovanadate, 2Na2O.V2O5 Sodium orthovanadate, 3Na2O.V2O5 Nickel pyrovanadate, 2NiO.V2O5 Nickel orthovanadate, 3NiO.V2O5 Ferric metavanadate, Fe2O3.V2O5 Ferric vanadate, Fe2O3.2V2O5 Sodium vanadic vanadate, Na2O.V2O4.V2O5 Sodium vanadic vanadate, 5Na2O.V2O4.11V2O5

535 (995)

Source: Ref 6

Fig. 11.1 Ref 7

Phase diagram for V2O5-Na2O system showing a series of low melting eutectic compounds. Source:

Kawamura and Harada (Ref 8) observed that when the value of the (Na +S)/V in at.% ratio was 20 or higher, the melting point of the deposits on the superheater tubes was more than 800 °C (1470 °F). However, this melting point could be reduced to about 500 °C (930 °F) when the ratio became less than 3 to 4. This is illustrated in Fig. 11.3 (Ref 8). The analysis of the chemical compositions (wt%) of the deposits that formed on the superheater tubes when the boiler was ﬁred with fuel oils with various ratios of (Na + S)/V showed that the amounts of S as SO3, V as V2O5, and Na as Na2O varied as a function of the (Na +S)/V ratios in the fuel, as shown in Table 11.4 (Ref 8). It has been well accepted that the oil-ash corrosion is the result of the formation of molten vanadate compounds involving V2O5-Na2O and/ or V2O5-Na2SO4 systems. Molten vanadate compounds ﬂux away the oxide scales formed on the metal, thus causing rapid corrosion attack. The corrosion mechanism is essentially hot corrosion involving a fused salt in an oxidizing environment. The hot corrosion mechanism involving fused Na2SO4 has been extensively studied in high-temperature corrosion of gas turbine components (see Chapter 9). The solubility of an oxide in a fused salt can vary as the melt chemistry changes. A salt can dissociate into a base (Na2O) and acid (SO3). The melt is generally characterized by Na2O concentration (melt basicity). When the melt exhibits high solubility of an oxide formed on an alloy, a high corrosion rate will occur. Rapp (Ref 9) provides

Fig. 11.2 Source: Ref 7

Melting points of low melting compounds in the V2O5-Na2SO4 system during heating and cooling.

pg 322

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:44PM Plate # 0

Chapter 11:

Fig. 11.3

pg 323

Oil-Fired Boilers and Furnaces / 323

Melting of the deposits formed on the superheater tubes as a function of the value of (Na + S)/ V ratio (in atomic percent) in the fuel oils used in firing a boiler (375 MW) producing superheated steam of 570 °C (1060 °F). Source: Ref 8

Table 11.4 Chemical compositions (wt %) of the deposits formed on the superheater tubes in a boiler when fired with fuel oils with different concentrations of vanadium Fuel oil(a)

S as SO3 V as V2O5 Na as Na2O Fe as Fe2O3 Ni as NiO Ca as CaO Mg as MgO SO3+V2O5+Na2O

0.2–0.3% S 1–3 ppm V

2.7–2.8% S 45–65 ppm V

1.6–1.8% S 130–150 ppm V

2.4–2.5% S 200–250 ppm V

51.8 0.85 34.4 4.70 3.38 2.06 1.92 87.1

24.4 30.0 17.6 13.0 6.42 2.25 1.41 72.0

21.6 49.7 17.8 11.2 2.24 1.17 0.88 72.0

0.89 83.0 2.69 6.48 7.45 0.22 0.20 86.6

(a) Sodium in fuel oils was in a range of 8–15 ppm. Source: Ref 8

a general review on the subject of the hot corrosion mechanism in fused salt. Figure 11.4 shows the solubilities of various oxides in fused Na2SO4 at 927 °C (1700 °F) and 1.0 atm O2 as a function of the activity of Na2O (aNa2O) (Ref 9). Oil-ash corrosion may involve fused salt of sodium vanadate and/or sodium sulfate. The partial pressure of SO3 plays an important role in the stability of the fused salt, such as sodium vanadate, or sodium vanadate/sodium sulfate mixture, or sulfate, formed on the metal. Seiersten and Kofstad (Ref 10) discuss the effect of SO3 on vanadate-induced hot corrosion. Fig. 11.4

11.3 Corrosion Problems in Oil-Fired Boilers and Furnaces Oil-ash corrosion problems in boilers are typically associated with superheaters and reheaters because the tube metal temperature may reach the melting point of the ash deposit. The

The solubilities of various oxides in fused Na2SO4 at 927 °C (1700 °F) and 1.0 atm O2 (The solubility of SiO2 is at 900 °C, or 1650 °F). Source: Ref 9

uncooled components, such as tube hangers, in oil-ﬁred boilers are subject to much higher temperatures and thus can suffer more serious corrosion problems. Many furnaces in reﬁnery and petrochemical plants often use residual fuel oils for heating high-temperature processing

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

324 / High-Temperature Corrosion and Materials Applications

equipment. The uncooled components in these furnaces can suffer severe oil-ash corrosion attacks. This section discusses oil-ash corrosion problems in both power boilers and reﬁnery furnaces. 11.3.1 Waterwall Corrosion of Oil-Fired Boilers Because of much lower metal temperatures, furnace waterwalls generally do not suffer serious corrosion problems in oil-ﬁred boilers. The temperature of the ﬁreside metal surface for the furnace waterwalls for fossil-ﬁred boilers including oil-ﬁred boilers in the United Kingdom is generally limited to about 450 °C (840 °F) (Ref 11). Reichel (Ref 12) indicates the metal temperatures for the furnace waterwalls were in the range of 300 to 460 °C (570 to 860 °F) for fossil-ﬁred boilers including oil-ﬁred boilers in Germany. However, in case of ﬂame impingement, the furnace waterwall may be subjected to higher-temperature exposure. Formation of oxide scales or corrosion products on the internal diameter (ID) of the waterwall tubes can also signiﬁcantly increase the outer tube metal temperature. French (Ref 13) indicated that a thin internal deposit can raise the tube metal temperature into the ash-corrosion range, into the creep-failure range, or into the rapid-oxidation range, leading to serious furnace-tube problems. In some cases, oil-ﬁred boilers are used as a “peaking” unit.* As a result, the load in these units is often cycled daily or weekly, thus subjecting the boiler tubes to thermal cycling. The thermal cycling can lead to circumferential grooving and cracking of the waterwall tubes under corrosive conditions such as under the oilash deposits and higher tube metal temperatures. The development of circumferential grooving/ cracking on the waterwall tubes in coal-ﬁred boilers is discussed in Chapter 10 “Coal-Fired Boilers.” The example that is given below shows the circumferential grooving/cracking of the waterwall tubes that is believed to be the result of thermal cycling, higher metal temperatures, and oil-ash corrosive conditions (Ref 14). A supercritical oil-ﬁred boiler (590 MWe) required replacement of waterwall panels, which were made of 2.25Cr-1Mo steel (T22), every 2 to 3 years because of severe corrosion. The * The peaking unit is a boiler that is used mainly during highpower demand seasons, such as summer.

waterwall tube had the dimensions of 11/8 in. outside diameter (OD) by 0.220 in. MWT. Several panels that were aluminized by a commercial company were tried in the boiler. A tube leakage developed after about 2.5 years of service. Thus, the aluminized coating failed to extend the life of the waterwall tube. A close-up view of the waterwall tube showing circumferential grooves at the location near the localized rupture area is shown in Fig. 11.5. The aluminized coating was completely destroyed. Oxidation attack penetrated through the aluminized coating and into the steel tube wall (Fig. 11.6).

Fig. 11.5

Close-up view of the tube surface near the rupture area showing numerous circumferential grooves and cracks. Courtesy of Welding Services Inc.

Fig. 11.6

Optical micrograph showing the oxidation penetrated through the aluminized coating and into the tube wall. Note the tube inside diameter surface is visible at the right side bottom of the micrograph and the fire side is on the leftside of the micrograph. Courtesy of Welding Services Inc.

pg 324

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

Chapter 11:

The remnant material of the aluminized coating can be seen in Fig. 11.7 (area No. 5 in Fig. 11.7). The deposits and corrosion phases were analyzed by scanning electron microscopy with energy dispersive x-ray spectroscopy (SEM/EDX) with the results summarized in Fig. 11.7. A signiﬁcant amount of vanadium along with magnesium, phosphorus, sodium, zinc, and other elements was observed in the ash deposit and corrosion products. The preferential corrosion penetration that forms a groove into the metal was essentially iron with little sulfur and is believed to consist of iron oxides. Thermal cycling of the waterwall caused by load changes and oil-ash corrosion with possible higher waterwall tube temperatures caused by ﬂame impingement are believed to have caused preferential oxidation penetrations, which developed into circumferential grooves and cracks. Kawamura and Harada (Ref 8) also observed circumferential grooving or cracking on furnace waterwall tubes made of 2.25Cr-1Mo steel in an oil-ﬁred supercritical unit.

Fig. 11.7

Scanning backscattered electron image showing the oxidation attack of the diffusion coating and the substrate steel. The chemical analysis at different locations was performed using EDX. The results of the analysis (wt %) are summarized as: No. 1: 23% V, 20% Ni, 17% Si, 10% Fe, 8% Mg, 9% Al, 5% P, 3% Na, 2% Zn, and trace elements. No. 2: 53% Fe, 19% Ni, 12% Zn, 11% Al, 2% Mg, 2% V, and trace elements. No. 3: 34% Si, 24% Al, 12% V, 8% Fe, 7% Mg, 5% Na, 4% Ca, 2% Ni, and trace elements. No. 4: 54% Si, 20% Al, 8% Na, 6% Fe, 3% P, 4% Ca, and trace elements. No. 5: 70% Fe, 28% Al, 1% Cr, and trace elements. No. 6: 74% Fe, 14% Al, 10% V, and trace elements. No. 7: 90% Fe, 4% Cr, 2% Mo, 1% S, and trace elements. No. 8: 90% Fe, 4% Cr, 2% Mo, 2% S, and trace elements. No. 9: 97% Fe and trace elements.

Oil-Fired Boilers and Furnaces / 325

11.3.2 Superheater and Reheater Corrosion in Oil-Fired Boilers To avoid the formation of molten vanadate compounds in the ash deposits formed on superheaters or reheaters in boilers, European practice in general has limited the steam temperatures to 540 °C (1000 °F) (Ref 15). This steam-outlet temperature limit implies that the outer tube metal temperature is not greater than 580 °C (1080 °F) (Ref 11). In the United States, oil-ﬁred boilers ordered after 1965 have design steam-exit temperatures limited to 540 °C (1005 °F) to prevent oil-ash corrosion problems (Ref 16). The changes in steam-exit temperatures for oil-ﬁred boilers constructed during the ﬁrst eight decades of the 20th century are summarized in Fig. 11.8 by Paul and Seeley (Ref 16). In general, oil-ash corrosion problems tend to occur at the secondary superheaters and reheaters where the metal temperatures are the highest. For austenitic stainless steels, corrosion rates were found to be strongly dependent on the ﬂue gas temperature. Figure 11.9 shows the corrosion rates of austenitic stainless steels as a function of the tube surface temperature with ﬂue gas temperatures of 800 and 1150 °C (1470 and 2100 °F) (Ref 11). Austenitic stainless steels suffered a rapid increase in wastage rates with increasing tube surface temperature when the ﬂue gas was 1150 °C (2100 °F) compared with the 800 °C (1470 °F) ﬂue gas. Ferritic steels, on the other hand, were less sensitive to the ﬂue gas temperature. Holland et al. (Ref 15) conducted corrosion probe tests in an oil-ﬁred boiler at Marchwood Power Station for three austenitic stainless steels (Type 316, 321, and 347) and a 12Cr ferritic steel (12Cr, 0.5Mo, 0.25V). The boiler was operated under base-load conditions with 0.5% O2 in ﬂue gas. The analysis of the fuel oil used during testing showed 3.54% S, 91 ppm V, and 71 ppm Na. The test results are summarized in Fig. 11.10 and 11.11 (Ref 15). The results conﬁrmed that the 12Cr ferritic steel performed much better than all three austenitic stainless steels. Among three austenitic stainless steels, Type 347 was found to be much better than Type 316 and 321. The authors did not offer any explanation on the performance ranking of the three austenitic stainless steels. However, the results of the chemical analyses of the three alloys showed 18.2% Cr for Type 347, 17.1% Cr for Type 321, and 16.7% Cr for Type 316. It is, thus, believed that the performance of three austenitic stainless

pg 325

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

Fig. 11.8

Steam exit temperature, °C

Steam exit temperature, °F

326 / High-Temperature Corrosion and Materials Applications

Summary of the design steam-exit temperatures for oil-fired boilers built during the 20th century. Source: Ref 16

Fig. 11.9

Corrosion rates of austenitic stainless steels and ferritic steels as a function of metal temperature and flue gas temperatures. Source: Ref 11

steels followed the chromium concentration levels in the alloys. The metal-loss data appear to follow a linear kinetic when plotted as a function of exposure time, as shown in Fig. 11.12 (Ref 15). This ﬁgure clearly shows the superior performance of the 12Cr ferritic steel compared with austenitic stainless steels containing 16 to 18% Cr. Alexander et al. (Ref 17) found the similar test results showing ferritic steels with less chromium being signiﬁcantly more resistant to oil-ash corrosion than austenitic stainless steels with more

chromium. The authors compared several ferritic steels with several austenitic stainless steels by welding the tube samples together to form a test tube. This test tube was then installed in the secondary superheater tube bank in a boiler ﬁred with fuel oil containing 4.0% S, 0.007% Na, 0.007% V, and 0.017% Cl at Bromborough Power Station (United Kingdom). During the exposure test, the metal temperatures ranged 315 to 680 °C (600 to 1225 °F) for about 20,026 h. The tube was exposed to 590 to 620 °C (1100 to 1150 °F) for about 8000 h. Test tube samples were measured for maximum metal loss caused by the loss in thickness by corrosion (measured after the corrosion scales were removed by descaling) and for the total metal wastage that included maximum metal loss and intergranular penetration (measured by metallography). Ferritic steels showed very little internal penetration, while austenitic stainless steels suffered some internal penetration. The data were plotted in terms of the chromium concentration in alloys for both ferritic steels and austenitic stainless steels and are summarized in Fig. 11.13 (Ref 17). More data comparing ferritic steels with austenitic stainless steels were generated by Parker et al. (Ref 4) using corrosion probes tested in a boiler ﬁred with fuel oil containing 2.65% S, 49 ppm V, and 44 ppm Na at Marchwood Power Station (United Kingdom). Tube metal loss data for ferritic steels and austenitic stainless steels are summarized in Fig. 11.14 and 11.15 (Ref 4). The test results showed both 9Cr and 12Cr ferritic steels were signiﬁcantly more corrosion resistant

pg 326

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

Chapter 11:

than austenitic stainless steels including Type 310 (25Cr-20Ni). It is somewhat surprising to ﬁnd that 2.25Cr-1Mo steel was much more resistant than austenitic stainless steels (Fig. 11.14). It is more surprising to ﬁnd that 2.25Cr1Mo steel was comparable to 9Cr and 12Cr steels at approximately 550 to 625 °C (1020 to 1155 °F). This was in contrast with test results that were generated by Alexander et al. (Ref 17)

Oil-Fired Boilers and Furnaces / 327

in a boiler at Bromborough Power Station, showing that increasing chromium concentration signiﬁcantly increased the corrosion resistance for ferritic test steels 2.25Cr-1Mo, 9Cr, and 12Cr steels (Fig. 11.13). Among the four austenitic stainless steels, Type 310 (25Cr-20Ni) was more resistant than Type 316, 321, and 347 due to higher chromium content. The test results generated at the upper furnace are summarized

Fig. 11.10

Results of corrosion probe tests for 1000 h as a function of the tube metal temperature in a boiler fired with fuel oil containing 3.54% S, 91 ppm V, and 71 ppm Na. Source: Ref 15

Fig. 11.11

Results of corrosion probe tests for 2000 h as a function of the tube metal temperature in a boiler fired with fuel oil containing 3.54% S, 91 ppm V, and 71 ppm Na. Source: Ref 15

pg 327

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

328 / High-Temperature Corrosion and Materials Applications

in Fig. 11.15 (Ref 4), again showing both ferritic steels were more resistant than austenitic stainless steels. In this upper furnace test, 2.25Cr1Mo steel was not included.

The plant exposure tests have shown that 9Cr and 12Cr ferritic steels are much more resistant to oil-ash corrosion than austenitic stainless steels, such as Type 316, 321, and 347, typically used

Exposure time, h

Fig. 11.12

Fig. 11.13

Metal loss as a function of exposure time for 12Cr ferritic steel and three austenitic stainless steels. Source: Ref 15

Tube wastage data, which included both metal loss (tube thickness loss) and total wastage (metal loss+intergranular penetration), for ferritic steels and austenitic stainless steels in terms of chromium concentration in alloys. The data were generated in a boiler at Bromborough Power Station (United Kingdom) fired with fuel oil containing 4.0% S, 0.007% Na, 0.007% V, and 0.017% Cl. Open data points: total wastage (maximum metal loss+intergranular penetration). Solid data points: Maximum metal loss. Source: Ref 17

pg 328

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

Chapter 11:

for superheaters and reheaters in utility boilers. Nevertheless, the ferritic steels may not have adequate creep strengths at higher temperatures. Composite tubes with austenitic stainless steels for providing materials strength and outer tube

Oil-Fired Boilers and Furnaces / 329

cladding for providing corrosion resistance offer better alternatives for higher-temperature applications. Bolt (Ref 18) evaluated some of the composite tubes in an oil-ﬁred experimental boiler heated with fuel oil containing 2.2% S, 200 ppm V, and 50 ppm Na at metal temperatures from 500 to 700 °C (930 to 1290 °F) and two ﬂue gas temperatures 1000 and 1125 °C (1830 and 2060 °F). The tests were conducted in corrosion probes for 2000 h of exposure with a continuous load (i.e., constant metal temperature) and a discontinuous load (varying temperatures—8 h at higher and 16 h at lower levels). Four coextruded composite tubes tested were Esshete 1250/Type 310, 800H/Type 446, 1714 CuMo/35CrA, and 800H/671; Type 310, Type 446, 35CrA, and 671 were cladding alloys. The 35CrA cladding was 35Cr-45Ni-Fe alloy, and the 671 cladding was 46Cr-Ni alloy. The test results are summarized in Fig. 11.16 and 11.17. A bell-shaped curve with the maximum attack at about 650 °C was observed for Type 347 for both continuous and discontinuous loads. Type 347 and 310 showed unacceptable corrosion rates (>1 mm/yr, or >39 mpy) at 630 to 675 °C (1165 to 1250 °F). Type 446 cladding appeared to be

Fig. 11.14

Maximum metal loss as a function of metal temperature tested for 10,000 h in the superheater inlet zone in a boiler fired with fuel oil containing 2.65% S, 49 ppm V, and 44 ppm Na. Source: Ref 4

Type 347H

Fig. 11.15

Maximum metal loss as a function of metal temperature tested for 10,000 h in the combustion zone in a boiler fired with fuel oil containing 2.65% S, 49 ppm V, and 44 ppm Na. Source: Ref 4

Fig. 11.16

Corrosion rates of Type 347 in comparison with 310, 446, 35CrA, and 671 claddings as a function of temperature (°C) with flue gas temperatures of 1000 and 1125 °C (1830 and 2060 °F) in a constant thermal load. Source: Ref 18

pg 329

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

330 / High-Temperature Corrosion and Materials Applications

sufﬁciently corrosion resistant. Both 35CrA and 671 claddings performed extremely well even though alloy 35CrA contained about 45% Ni and alloy 671 contained about 54% Ni with 35% Cr in the former and 46% Cr in the latter. Carburization has been observed at times in superheaters and reheaters. Lopez-Lopez et al. (Ref 19) found severe carburization in a heavily corroded Type 321H superheater tube after 107,000 h of service in a boiler that had been used as a “peaking” unit. The tube had suffered severe metal wastage. However, underneath the corroded surface was a shallow, severely carburized layer about 80 µm deep and a peak hardness of about 400 HV (Vickers hardness number). In this study, the authors failed to report (or analyze) the chemistry of the ash deposits including carbon collected from the superheater tube. In a separate study, Wong-Moreno et al. (Ref 20) analyzed nine different ash deposits collected from superheaters in different oil-ﬁred boilers ranging from 84 to 350 MW. Most deposits contained about 0.03 to 0.06% C, with only two deposits showing about 0.14% C. This level of carbon in the deposits is unlikely to cause severe carburization in austenitic stainless steels at the superheater tube temperatures. Therefore, Discontinuous load,

Discontinuous load,

Type 347H

Fe

it is believed that the observed carburization was not caused by oil-ash deposits. Internal carburization can occur in CO2-containing environments. McNallan et al. (Ref 21) observed internal carburization of Type 310 in Ar-20CO2 at 800 °C (1470 °F) for 24 h. Furthermore, it is believed that the observed carburization was not the major factor causing the severe superheater wastage problem. Carburization was often observed to take place after the alloy suffered severe wastage with no protective oxide scales on the corroded surface. In addition to selecting a better-performing alloy to resist oil-ash corrosion, methods that could reduce superheater and reheater corrosion include (a) decreasing excess air for combustion and (b) the use of additives to raise the oil-ash melting point. Excess air promotes the formation of SO3 and of V2O5, which melts at 675 °C (1250 °F), instead of V2O3 or V2O4, which melts at 1970 °C (3580 °F) (Ref 6). However, keeping very low excess air for combustion can probably be difﬁcult to maintain in practice. The Central Electricity Generating Board (CEGB) in the United Kingdom claimed to have made signiﬁcant achievements in combustion in oil-ﬁred boilers with low excess air (Ref 22). Boilers at Merksem Power Station were reported to have operated at 1% excess air (0.2% O2) as a regular operating practice (Ref 23). The use of additives to increase the melting point of the vanadate or sulfate is likely to be a more practical approach. Useful additives include (Ref 8):

Magnesium compounds: MgO, Mg(OH)2, MgCO3, MgSO4, MgCO3·CaCO3, organic magnesium compounds Calcium compounds: CaO, Ca(OH)2, CaCO3 Barium compounds: BaO, Ba(OH)2 Aluminum and silicon compounds: Al2O3, SiO2, 3Al2O3·2SiO2 The additive reacts with vanadium compounds to form reaction products with higher melting points. When magnesium compounds are used, some of the reaction products and their melting points are:

Fig. 11.17

Corrosion rates of Type 347 in comparison with 310, 446, 35CrA, and 671 claddings as a function of temperature (°C) with flue gas temperatures of 1000 and 1125 °C (1830 and 2060 °F) in varying loads. Source: Ref 18

MgO·V2O5: 671 °C (1240 °F) 2MgO·V2O5: 835 °C (1535 °F) 3MgO·V2O5: 1191 °C (2175 °F) With injection of magnesium compounds into the fuel, the higher Mg/V ratios increase the

pg 330

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

Chapter 11:

melting point of the oil-ash deposits (Fig. 11.18) (Ref 6). Increasing the melting point of oil-ash deposits would result in lowering the corrosion rates. Figure 11.19 shows the effectiveness of the Mg(OH)2 additive injection in reducing the corrosion rate of Type 321 superheater tubes (Ref 8). Disadvantages of the additive injection approach include additional operating costs and a substantial increase in ash volume that may require additional furnace downtime for tube cleaning (Ref 24). 11.3.3 Tube Supports or Hangers in Boilers and Refinery/Petrochemical Furnaces Many reﬁnery and petrochemical furnaces are ﬁred with residual fuel oils containing signiﬁcant amounts of sulfur, vanadium, and sodium. Structural supports and tube hangers for the

Fig. 11.18

Effect of MgO addition on the melting point of oilash deposit on superheater tubing of an oil-fired boiler. Source: Ref 6

Fig. 11.19 Source: Ref 8

Corrosion of Type 321 superheater tubes with and without Mg(OH)2 injection in an oil-fired boiler.

Oil-Fired Boilers and Furnaces / 331

heater tubes used to process hydrocarbon ﬂuid in the furnace are uncooled and thus can be subjected to temperatures as high as 900 °C (1650 °F) or higher. Severe materials problems due to oil-ash corrosion were illustrated by numerous case histories presented in a 1958 NACE Technical Committee Report (Ref 25). For example, Type 309 tube supports suffered a corrosion rate of about 12.7 mm/yr (500 mpy) in an oil-ﬁred heater in a reﬁnery plant (Ref 25). This heater increased the temperature of the oil charge to 350 to 470 °C (670 to 880 °F) in a furnace operating at 940 to 980 °C (1700 to 1800 °F) ﬁred with residual fuel containing about 4.1% S and 0.05% ash composed of about 16.2% V2O5. Another example involving uncooled superheater spacers made of cast HH stainless steel (25Cr-12Ni) failed after only 7 months of service in a boiler ﬁred with a Buncker “C” oil, producing 565 °C (1050 °F) steam (Ref 25). The corrosion of the HH tube spacers was found to proceed by sulﬁdation/ oxidation attack (Ref 25). McDowell et al. (Ref 5) conducted ﬁeld test racks in a utility boiler ﬁred with Buncker “C” fuel oil. The test racks were exposed to the ﬂue gas in the second bank of the superheater. During testing, the boiler operated at full load during the day for 5 days a week and at half and low loads during early morning hours and weekends. The temperature of the ﬂue gas was about 850 °C (1560 °F) during full load of power, 660 °C (1220 °F) during half load, and 565 to 590 °C (1050 to 1100 °F) during minimum loads. No fuel analysis was reported in the paper. Analysis of the deposits/corrosion scales showed a signiﬁcant amount of V2O5 (Table 11.2). The test results, which were extrapolated from three test racks exposed for 504, 648, and 707 h, respectively, are summarized in Table 11.5. None of the alloys tested showed acceptable wastage rates. More rack tests in boilers ﬁred with Buncker “C” oils containing high concentrations of vanadium (150 to 450 ppm) were reported by McDowell and Mihalisin (Ref 26). Test racks were exposed to the ﬂue gas in the superheater section. Alloys ranging from low-alloy steels to iron- and nickel-base alloys suffered severe corrosion attack. The results of one test rack are shown in Table 11.6 (Ref 26). Specimens included 5Cr steel, stainless steels (both 400 and 300 series), Fe-Ni-Cr alloys, Ni-Cr-Fe alloy, and 50Ni-50Cr alloy, along with two cast stainless steels (HE and HH alloys). All the alloys tested

pg 331

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

332 / High-Temperature Corrosion and Materials Applications

exhibited unacceptable corrosion rates. Even the best performer (50Ni-50Cr alloy) suffered a corrosion rate of 3.1 mm/yr (121 mpy) (Ref 26). Some investigators have reported excellent performance of “50Cr-50Ni” alloy as tube supports in furnaces or boilers ﬁred with fuel oils containing sulfur and vanadium. Spafford (Ref 27) reported good performance of the 50Ni-50Cr alloy in reﬁnery heaters for coking and catalytic reformer units. The heaters were ﬁred with heavy fuel oil containing 2.5 to 4% S and 50 to 70 ppm V (occasionally up to 150 ppm). The original Table 11.5 Results of field rack tests conducted in the second bank of the high-temperature superheater at the location about 3 ft from the brick side wall in a utility boiler fired with Buncker “C” fuel oil with flue gas stream fluctuating from 565 to 850 °C (1050 to 1560 °F) while the boiler changed from half to full loads during the exposure test Alloy

Carbon steel 2.25Cr-1Mo 502 (5Cr steel) 410 (12Cr steel) 406 (12Cr-3Al steel) 430 446 302 321 309 800 HW (12Cr-60Ni) HF (21Cr-9Ni) HE (28Cr-10Ni) Inconel 600 (Ni-15Cr-7Fe)

Wastage rate, mm/yr (mpy)

17.7 (695) 17.0–19.1 (670–750) 6.2–14.2 (244–557) 14.1 (555) 2.7 (108) 9.5–26.9 (374–1060) 4.8 (189) 10.3 (406) 17.5 (690) 6.2–7.7 (244–305) 5.5 (217) 5.8–15.3 (230–601) 10.7 (422) 3.1 (124) 2.9 (113)

hangers and tube supports made of cast HH alloy (25Cr-12Ni steel) suffered severe corrosion attack. Metal temperatures were in the range of 730 to 890 °C (1350 to 1630 °F). The highest corrosion rates of cast HH alloy were 6.4 to 9.5 mm/yr (250 to 375 mpy). Replacements with alloy 657 (a cast 50Ni-50Cr alloy) were reported to perform very well, with minimal maintenance and repair (Ref 27). Swales and Ward (Ref 2) reported the results of a ﬁeld test that showed alloy 657 performed 10 times better than HH and HK alloys, as illustrated in Fig. 11.20. There are additional cases where alloy 657 tube supports provided excellent performance; two such examples (Ref 2) are given below. The authors also concluded that at temperatures higher than 900 °C (1650 °F), alloy 657 often suffered severe corrosion attack, but provided good protection up to that temperature (Ref 2). Catalytic Reformer Heaters in a Refinery. The furnace was ﬁred with a mixture of fuel oil (about 50%) and gas. The ash analysis showed 48% V2O5 and 2% Na2O. Tube supports were exposed to temperatures up to 900 °C (1650 °F). The tube supports made of cast HH stainless steel (25Cr-12Ni) suffered severe corrosion after 2 years of service. Tube supports were replaced with cast alloy 657 (Ni-48Cr) showing no noteable corrosion after 7 years of service.

Source: Ref 5

Table 11.6 Results of a field test (uncooled specimens) exposed in the superheater section at 815 °C (1500 °F) in a boiler fired with high vanadium (150–450 ppm) Buncker “C” fuel Alloy

5Cr steel 406 431 446 302B 309 321 347 310 Incoloy(a) Incoloy 804 Inconel(a) 50Cr-50Ni HE HH

Wastage rate, mm/yr (mpy)

32–45 (1270–1775) 5.7 (224) 17–23 (655–925) 3.8 (149) 14–16 (533–644) 5.0 (196) 9–13 (346–505) 6.2 (243) 4.7 (187) 9–12 (364–458) 13–18 (495–710) 5.0 (196) 3.1 (121) 4.8 (187) 12–16 (467–645)

(a) Incoloy may refer to Incoloy alloy 800, while Inconel may refer to Inconel alloy 600. Source: Ref 26

Fig. 11.20

Results of a field test at 700 °C (1290 °F) for 6 months in a refinery (crude oil) heater comparing alloy 657 to HH and HK alloys. Source: Ref 2

pg 332

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

Chapter 11:

Catalytic Reformer Heater in a Refinery. The furnace was ﬁred with fuel oil containing 3 to 4% S, 40 to 50 ppm V (90 to 150 ppm occasionally for short periods). Tube hangers were exposed to 700 to 900 °C (1290 to 1650 °F). The original HH cast roof tube hangers suffered severe wastage problems after 1 to 2 years of service. All tube hangers were then replaced with alloy 657. No failures occurred after more than 4 years of service for alloy 657 hangers.

11.4 Summary Fireside corrosion can present a serious problem in oil-ﬁred boilers or reﬁnery/petrochemical furnaces ﬁred with low-grade fuels with high concentrations of vanadium, sulfur, and sodium. This corrosion is frequently referred to as “oil-ash corrosion.” Accelerated attack by oil-ash corrosion is related to the formation of lowmelting point molten vanadium pentoxide and sodium sulfate eutectics, which ﬂux the protective oxide scale from the metal surface. In boilers, superheater and reheater tubes are susceptible to oil-ash corrosion attack. Uncooled components in the boilers, such as tube supports and spacers, can suffer severe corrosion attack because of higher temperatures. Oil-ash corrosion can also occur in reﬁnery and petrochemical furnaces burning low-grade fuels. The resistance of oil-ash corrosion for various alloys in both boilers and reﬁnery/petrochemical furnaces is reviewed.

REFERENCE

1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, Barberton, OH, 1992 2. G.L. Swales and D.M. Ward, Strengthened 50% Chromium, 50% Nickel Alloy (IN657) Reﬁnery Heater Tube Supports to Combat Fuel Ash Corrosion—A Review of Service Caswe Histories, Paper No. 126, Corrosion/ 79, NACE, 1979 3. D.N. French, Corrosion of Superheaters and Reheaters in Fossil Fired Boilers, Environmental Degradation of Engineering Materials in Aggressive Environments, Conf. Proc., Virginia Polytechnic Institute, Blacksburg, VA, 1981, p 407

Oil-Fired Boilers and Furnaces / 333

4. J.C. Parker, D.F. Rosborough, and M.J. Virr, High Temperature Corrosion Trials at Marchwood Power Station—10,000 Hour Corrosion Probe Trials, J. Inst. Fuel, Feb 1972, p 95 5. D.W. McDowell, R.J. Raudebaugh, and W.E. Somers, High-Temperature Corrosion of Alloys Exposed in the Superheater of an Oil-Fired Boiler, Trans. ASME, Feb 1957, p 319 6. M. Fichera, R. Leonardi, and C.A. Farina, Fuel Ash Corrosion and Its Prevention with MgO Addition, Electrochim. Acta, Vol 32 (No. 6), 1987, p 955 7. W.T. Reid, External Corrosion and Deposits —Boiler and Gas Turbines, Elsevier Publishing, 1971 8. T. Kawamura and Y. Harada, “Control of Gas Side Corrosion in Oil Fired Boilers,” Mitsubishi Technical Bulletin No. 139, Mitsubishi Heavy Industries, Ltd., May 1980 9. R.A. Rapp, Hot Corrosion of Materials, Pure Appl. Chem., Vol 62 (No. 1), 1990, p 113 10. M. Seiersten and P. Kofstad, The Effect of SO3 on Vanadate-Induced Hot Corrosion, High Temp. Technol., Vol 5 (No. 3), 1987, p 115 11. A.J.B. Cutler, T. Flatley, and K.A. Hay, FireSide Corrosion in Power Station Boilers, Metall. Mater. Technol., Feb 1981, p 69 12. H.H. Reichel, Fireside Corrosion in German Fossil-Fuel Fired Power Plants: Appearance, Mechanism and Causes, Werkst. Korros., Vol 39, 1988, p 54 13. D.N. French, Metallurgical Failures in Fossil Fired Boilers, 2nd ed., John Wiley & Sons, 1993, p 370 14. Welding Services Inc. unpublished data 15. N.H. Holland, D.F. O’Dwyer, D.F. Rosborough, and W. Wright, High-Temperature Corrosion Investigations on an Oil-Fired Boiler at Marchwood Power Station, J. Inst. Fuel, May 1968, p 206 16. L.D. Paul and R.R. Seeley, Oil Ash Corrosion—A Review of Utility Boiler Experience, Paper No. 267, Corrosion/90, NACE International, 1990 17. P.A. Alexander, R.A. Marsden, J.M. NelsonAllen, and W.A. Stewart, Operational Trials of Superheater Steels in a C.E.G.B. OilFired Boiler at Bromborough Power Station, J. Inst. Fuel, Feb 1964, p 59 18. N. Bolt, Fireside Corrosion Phenomena in Superheaters of Coextruded Materials with

pg 333

Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/

26/10/2007 12:45PM Plate # 0

334 / High-Temperature Corrosion and Materials Applications

19.

20.

21.

22.

18, 25, 35 and 50% Cr at Metal Temperatures of up to 700 °C, International Congress on Metallic Corrosion, Conf. Proc., Vol IV, Oxford & IBH Publishing Co. Pvt. Ltd, 1987, p 3593 D. Lopez-Lopez, A. Wong-Noreno, and L. Martinez, Unusual Superheater Tube Wastage Associated with Carburization, Mater. Perform., Dec 1994, p 45 A. Wong-Moreno, Y.M. Martinez, and L. Martinez, High Temperature Corrosion Enhanced by Residual Fuel Oil Ash Deposits, Paper No. 185, Corrosion/94, NACE International, 1994 M.J. McNallan, S. Thongtem, J.C. Liu, Y.S. Park, and P. Shyu, Corrosion of Chromium Containing Alloys in Non-steady State Environments Containing Oxygen, Carbon, and Chlorine, J. Phys. IV, Coll. C9, Suppl. J. Phys. III, Vol 3, Dec 1993, p 143 E.M. Hamilton and G.R. Stern, “Operation of Oil Fired Boilers with Very Low Excess

23.

24. 25.

26.

27.

Air,” CEGB Report RD/H/MI, Central Electricity Generating Board J. Remeysen, “Operation of Large Boilers at Very Low Excess Air Levels,” Proceedings of Brussels Conference of Institute of Fuel, Paper 1, Sept 1964 J.R. Wilson, Understanding and Preventing Fuel Ash Corrosion, Paper No. 12, Corrosion/76, NACE, 1976 The Present Status of the Oil Ash Corrosion Problem, NACE Technical Committee Report, Pub. 58-11, Corrosion, 1958, p 369t D.W. McDowell, Jr. and J.R. Mihalisin, Paper No. 60-WA-260, presented at ASME Winter Annual Meeting, Nov 27–Dec 2, 1960 B.F. Spafford, in UK Corrosion ’82, Conf. Proc. (Birmingham, U.K.), Nov 15–17 1982, Institution of Corrosion Science & Technology, Birmingham, U.K., 1982, p 67

pg 334

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p335-358 DOI: 10.1361/hcma2007p335

30/10/2007 4:09PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 12

Waste-to-Energy Boilers and Waste Incinerators 12.1 Introduction Municipal solid waste (MSW) is a combination of residential, commercial, and industrial refuse. Historically, the typical means of disposal of this refuse has been landﬁlling. A novel approach in recovering the energy from this “resource” in boilers for electricity generation and minimizing the landﬁll requirements began in the 1960s in Europe. It became apparent from these early European boilers that corrosion of the boiler tube materials was a major obstacle for operating these so-called wasteto-energy (WTE) boilers (Ref 1, 2). Managing the corrosion problems continues to be a big challenge for operators of modern boilers worldwide. This chapter focuses on the corrosion issues related to boilers burning municipal solid waste.

12.2 Fuels, Combustion Environments, and Boilers Municipal solid waste typically contains paper and paperboard, plastics, rubber, textile, leather, batteries, food waste, yard waste, metal, glass, and other miscellaneous materials. This type of fuel is very heterogeneous and varies greatly by geographic location, country, and “culture.” In North America, typical constituents of the MSW in 1990 were paper and paperboard (34.2–40.0%), plastics (7.2–9.2%), food waste (7.3–8.5%), yard waste (17.6–19.9%), metal (8.1–9.6%), glass (7.7–9.7%), and other (10.3– 13.2%) (Ref 3). Polyvinyl chloride (PVC) plastic is a dominant source of chlorine, which is a major corrosive constituent making the combustion environment corrosive to boiler tube materials. Polyvinyl chloride contains about 36 wt%

chlorine (Ref 4). Rubbers such as Hypalon, chloroprene, and neoprene also contain high concentrations of chlorine (Ref 4). Other waste constituents that can contribute to corrosive environments under combustion are batteries (automobile batteries, button-type batteries in watches and calculators, alkaline, zinc, nickel, and cadmium batteries) and consumer electronics. Batteries contribute lead and cadmium; household batteries contribute cadmium; and consumer electronics contribute both lead and cadmium (Ref 5). Rademakers et al. (Ref 6) reported their calculation of the heavy metal composition of Dutch solid waste based on 1994 data in g/kg waste: 0.01 Cd, 0.0004 to 0.0006 Hg, 0.22 to 0.56 Pb, 0.02 to 0.04 Sb, 0.63 to 1.04 Zn, 0.01 As, 3.96 to 6.89 Cl, 0.11 to 0.19 Br, 0.1 to 0.2 F, and 1.48 to 2.96 S. During combustion, these heavy metals can react with chlorine to form some low-melting-point chlorides, thus causing severe boiler tube material corrosion problems. Corrosion mechanisms are discussed in the next section. There are primarily two types of combustion technologies, namely, mass-burning and refusederived fuel (RDF) burning. For mass-burning units, the fuel, which is not segregated except for appliances, furniture, and other large articles, burns in the as-received condition. In RDF units, on the other hand, the fuel is segregated, classiﬁed, and shredded to size. Metals and glass are typically removed from RDF fuel prior to combustion. Thus, RDF fuel has a higher heating value than that of a mass-burning unit. A small number of boilers in use today are based on the ﬂuidized-bed technology. The mass-burning unit uses a mechanical stoker to feed the fuel onto a grate where the fuel is combusted. For most RDF units, the fuel is blown into the furnace sidewalls through an RDF burner with a fuel-distribution impeller (Ref 3). Most RDF units burn the fuel in

pg 335

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:09PM Plate # 0

pg 336

336 / High-Temperature Corrosion and Materials Applications

boilers developed in the 1960s were designed to operate at higher steam pressures and temperatures, particularly those in Germany, with one in Mannheim operating at 12.5 MPa/530 °C (1800 psig/980 °F) steam and another one in Munich at 18 MPa/540 °C (2650 psig/1000 °F) steam (Ref 1). Higher steam temperatures result in higher tube metal temperatures and thus more serious corrosion problems. Combustion of the MSW fuel is signiﬁcantly different from that of coal or oil because the solid waste is a heterogeneous fuel. In addition, the fuel contains numerous impurities that include chlorine, sulfur, sodium, potassium, cadmium, zinc, lead, and other heavy metals. Because of these impurities, combustion of this fuel generates a very corrosive environment that causes serious corrosion problems for boiler tube materials. Gaseous combustion products include N2, O2, CO2, H2O, SO2, HCl, HF, and other gaseous impurities such as CO and HBr. These gaseous constituents are often measured by plant operators. Examples of these ﬂue gas compositions measured in mass-burning units at different plants are shown in Table 12.1 (Ref 7, 11–13). However, vapors of metal chlorides and sulfates are also produced during combustion. These compounds are normally not quantiﬁed by plant personnel. Many of these metal chlorides exhibit high vapor pressures and/or low melting points. Some physical properties of many metal chlorides can be found in Chapter 6 “Corrosion by Halogen and Halides.” As is discussed in this chapter, some of these metal chlorides are primarily responsible for the corrosion of the boiler tube materials. The furnace is typically enclosed by four walls (rear, front, and two side walls) using a tubemembrane construction (i.e., individual tubes are connected by narrow plates). These tubemembrane walls, which are commonly referred to as “waterwalls,” provide heat-absorbing

suspension (Ref 3). RDF units also employ moving grates. Combustion gas temperatures are generally at or below approximately 1090 °C (2000 °F) in mass-burning units and approximately 1315 to 1370 °C (2400 to 2500 °F) in RDF units (Ref 7). The guidelines for solid waste combustion introduced by the European authorities require that the temperature should be above 850 °C (1560 °F) for a 2 s residence time at the level 1 m above the secondary or tertiary air (Ref 6). The MSW boilers operate at much lower steam pressures and temperatures compared with coalﬁred utility boilers, as discussed in Chapter 10. Kubin (Ref 7) reported that Ogden Martin boilers in the United States typically operated at steam pressures of 4 to 6 MPa (615 to 880 psig) and temperatures of 370 to 440 °C (700 to 830 °F) for mass-burning units, and steam pressures of 4.5 to 6 MPa (665 to 900 psig) and temperatures of 370 to 440 °C (700 to 830 °F) for RDF units. These pressures and temperatures appear to be representative of other WTE boilers in the United States. For example, two boilers at the Wheelabrator Concord facility operate at a steam pressure of 650 psig (4.5 MPa) and a temperature of 400 °C (750 °F), and two at the Wheelabrator Spokane facility operate at 6.2 MPa (900 psig) and 440 °C (830 °F) steam (Ref 8), with some other Wheelabrator boilers operating at higher pressures and temperatures (Ref 9). Some of the European boilers operate at higher steam pressures and temperatures. For example, the boilers at Ivry Paris (France) operate at 7.5 MPa (1103 psig, or 75 bar) and 480 °C (900 °F) steam, at HKW Mannheim (Germany) at 8 MPa (1176 psig, or 80 bar) and 500 °C (930 °F) steam, at BSR Berlin (Germany) at 7.5 MPa (1088 psig, or 74 bar) and 420 °C (790 °F) steam, and at EVO Oberhausen (Germany) at 7 MPa (1029 psig, or 70 bar) and 480 °C (896 °F) steam (Ref 10). Early European WTE

Table 12.1

Examples of flue gas compositions generated by different WTE mass-burning boilers Composition, vol%

Gas

Ogden-Martin (US)(a)

Saint Ouen (France)(b)

Richtlinien (Germany)(c)

Kuririn (Japan)(d)

O2 CO2 SO2 HCl HF CO H 2O N2

8.5–9.5 8–9 100–200 ppm 400–600 ppm 5–20 ppm … 13–15 bal

9–10 9–12 90–130 ppm 600–1200 ppm 20–40 ppm <20 ppm NR bal

NR NR 100–2000 ppm 560–2240 ppm … 64–640 ppm NR NR

9.1 11.1 40 ppm 810 ppm … 42 ppm 21.4 bal

NR, not reported. (a) Source: Ref 7. (b) Source: Ref 11. (c) Source: Ref 12. (d) Source: Ref 13

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:09PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 337

surfaces for converting the water inside the tubes to steam as the water rises from the bottom of the waterwall. At the top of the waterwall tubes, a mixture of steam and water leaves the waterwall tubes and enters the steam drum where steam is separated from water. The furnace waterwall typically operates with water at a saturation temperature. Thus, the temperature of the water/ steam in the waterwall tubes, which depends on the pressure of the water/steam, is generally at or below 277 °C (530 °F) (Ref 14). The ranges of the waterwall tube metal temperatures were cited to be 260 to 293 °C (500 to 560 °F) (Ref 7), 260 to 290 °C (500 to 550 °F) (Ref 15), and 260 to 315 °C (500 to 600 °F) (Ref 16). For higherpressure boilers, the temperature of the steam will be higher. The steam from the steam drum after separating from water is then further heated in superheaters (typically two superheaters— primary and secondary or ﬁnal superheaters) to higher temperature in the convection path before it is delivered to turbines for electricity generation. The combustion ﬂue gas exits from the furnace at the top and then enters into the convection path, ﬂowing typically through the superheater, then the boiler bank, and the economizer. The ﬂue gas temperature entering into the superheater may vary from approximately 650 to 900 °C (1200 to 1650 °F).

Figure 12.1 shows a schematic of a massburning unit. For some mass-burning units, screen tubes are installed in front of the superheater to lower the temperature of the ﬂue gas entering into the superheater section. In some other mass-burning units, there are additional gas passes (one or more additional walls to allow ﬂue gas to pass through in the convection path) to allow the ﬂue gas stream to lower its temperature before entering into the superheater section. This type of design is shown schematically in Fig. 12.2 (Ref 17). Licata et al. (Ref 17) indicated that passing ﬂue gas through two 180° turns and one 90° turn prior to entering the superheaters can not only lower the ﬂue gas temperature but also remove particles from the ﬂue gas stream. In addition, the ﬂue gas stream can achieve better mixing to minimize local reducing conditions (Ref 17). Figure 12.3 shows a schematic of an RDF unit. Carbon steels and sometimes lowalloy steels are typical construction materials for waterwalls, screen tubes, the boiler bank, and economizer. For superheaters, carbon and low-alloy steels are typically the materials of construction. This chapter focuses mainly on waterwalls and superheaters that have experienced severe corrosion problems. Also discussed are the latest protection methods against corrosion and

Refuse feed hopper

Boiler

Refuse receiving area

Refuse pit

Plunger ash extractor Ash removal

Fig. 12.1

Mass-burning unit, with a grate where the fuel burns and the superheater right above the arch (or bull nose) in the upper furnace, followed by a boiler bank and an economizer in the convection pass downstream of the superheater. Source: Ref 3

pg 337

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

pg 338

338 / High-Temperature Corrosion and Materials Applications

13

Drum

Radiant pass

1

2ND pass

SH. 2

3RD pass

Evap.

SH. 1

Econ.

14

6 10 2

5

4

4 12 3 11

A

B C

15

D

E

7 9 8

Fig. 12.2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Refuse feed hopper Refuse chute Refuse incineration grate Secondary air supply Furnace Auxiliary burner Primary air hopper Ash extractor Scraper conveyor Fly-ash hopper Fly-ash conveyor Boiler ash Superheater steam outlet Boiler feed water inlet Primary air supply

Mass-burning unit of different design involving multiple passes for the flue gas stream before entering the superheaters. Also shown is the grate where the fuel is combusted. Source: Ref 17

erosion/corrosion at these two critical areas in the boiler.

12.3 Boiler Tube Fireside Corrosion Problems

Prepared fuel storage Boiler

Ash removal system

Fig. 12.3

Refuse-derived fuel unit, with superheaters (right above the nose arch or bull nose) and a boiler (generating) bank in the convection pass. The bottom of the furnace is a traveling grate. Source: Ref 3

Materials problems and high wastage rates experienced by furnace waterwalls and superheaters in numerous waste-to-energy plants worldwide have been extensively reviewed by Krause (Ref 1), Sorell (Ref 2), and Krause and Wright (Ref 18). Many of the early boilers, particularly European boilers, suffered severe boiler tube failure problems. A boiler (410 °C, or 770 °F, and 5 MPa, or 725 psig steam) at the Issy plant in Paris suffered tube failures in the side wall after only about 5000 h of operation, and the same boiler suffered serious superheater corrosion problems from the start of the operation (Ref 1). In a boiler (525 °C, or 980 °F, and 12 MPa, or 1800 psig steam) in Mannheim, Germany, the furnace wall tubes failed only after

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 339

3000 h of operation, and superheater tube failed after only 2000 h of operation (Ref 1). The ﬁrst U.S. WTE boiler with relatively low steam temperature (320 °C, or 610 °F) in Nashville, TN, suffered both waterwall and superheater failure within the ﬁrst year of operation (Ref 18). Many other tube failures have been cited in Ref 1, 2, 18–21). All these initial boiler tubes materials were carbon and low-alloy steels with no corrosion protection. Modern WTE boilers have increasingly used various corrosion protection methods to protect waterwalls, superheaters, and boiler banks against corrosion and erosion/corrosion. Many boilers with no corrosion protection for their waterwalls and superheaters continue to suffer premature failures. Figure 12.4 shows an example of a tube failure at the waterwall in one boiler after only 8 months of operation (Ref 22). One effective corrosion protection method for the furnace waterwalls was developed in the 1985 to 1986 period by using alloy 625 overlay cladding, which was applied on-site in an RDF boiler in Lawrence, MA using an automatic gas metal arc welding (GMAW) process (Ref 23). A total of about 21,000 lb of alloy 625 weld overlay metal was applied to this boiler. The performance of the waterwall overlay of alloy 625 in this ﬁrst overlaid waterwall proved to be very successful during the next 3 to 4 years of boiler operation. Between 1989 and 1990, a total of about 260,000 lb of alloy 625 weld overlay metal was applied to the waterwalls of 29 boilers (Ref 23).

Fig. 12.4

Alloy 625 overlay has been found to provide dramatic reduction in metal loss in areas where carbon or low-alloy steels have suffered unacceptable wastage rates (Ref 7, 16, 22–25). For mass-burning units, the waterwalls are typically protected by alloy 625 weld overlay above the refractory lining at the lower, high-radiant section of the boiler (Ref 9). The waterwalls in an RDF boiler are generally fully protected with alloy 625 weld overlay (Ref 9). Kubin (Ref 7) indicated that all RDF boiler waterwalls for Ogden plants were virtually fully covered with alloy 625 weld overlay. In many plants, alloy 625 as an outer tube cladding for corrosion protection of superheaters (with higher metal temperatures) has also proved to provide signiﬁcant reduction in tube-wall wastage rates compared to carbon and low-alloy steels (Ref 10, 22, 24). Excellent performance was observed for alloy 625 overlay superheater tubes in a boiler in Europe, as shown in Fig. 12.5. The cladding can either be in a form of weld overlay (Ref 24) or coextruded tube cladding (Ref 10). In some plants, however, alloy 625 cladding in superheaters was found to experience high wastage rates (Ref 10, 12, 22, 24). This is illustrated in Fig. 12.6 (Ref 10, 22). The wastage rates ranged from less than 10 mpy to hundreds of mpy. Superheater corrosion appears to be highly dependent on the individual boiler. Possible corrosion mechanisms involved in WTE environments are discussed in the next section.

Carbon steel waterwall suffering blown tubes due to high wastage rates resulting in significant tube-wall thinning after 8 months of service in a WTE boiler. Courtesy of Welding Services Inc.

pg 339

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

340 / High-Temperature Corrosion and Materials Applications

12.4 Corrosion Mechanisms As described in Section 12.3, the furnace waterwalls and superheaters made of carbon or low-alloy steels suffered severe wastage problems in numerous boilers, experiencing tube failures after only thousands of hours of operation. Corrosion was the result of exposure to the combustion products generated during the combustion of MSW. Typical gaseous components generated during combustion are O2, CO2, H2O, SO2, and HCl with HF in some units. In some cases, local areas at the waterwall may be in reducing conditions. In those local furnace wall areas, there could be some CO present. Among these gaseous components, HCl is considered to be the most corrosive. In general, the level of HF,

Fig. 12.5

if present, is signiﬁcantly lower than that of HCl (see Table 12.1). Thus, HF plays no signiﬁcant role in corrosion of boiler tubes. Carbon and low-alloy steels are very susceptible to corrosion by HCl at elevated temperatures. The level of HCl generated in combustion of MSW fuel is typically in the range of 100 to 2000 ppm. The tube metal temperatures of the waterwall, although depending on the operating pressure of the boiler along with other factors, are generally in the range of 260 to 315 °C (500 to 600 °F). The superheater tube metal temperatures are generally in the range of 370 to 540 °C (700 to 1000 °F), depending on the steam temperature, along with other factors. A brief review of the corrosion data in terms of the direct HCl corrosion is discussed below to see whether HCl

Alloy 625 overlay superheater tubes (on 15Mo3 steel substrate) after 4.5 years of service in a superheater producing 405 °C (760 °F)/42 bar (609 psi) superheated steam, showing no evidence of corrosion or erosion/corrosion. Source: Ref 24

Steam temperature, °C –18

95

205

315

425

535

350

Series 1 Series 2

Wastage rate, mpy

300 250 200 150 100 50 0 0

200

400

600

800

1000

Steam temperature, °F

Fig. 12.6

Wastage rates as a function of steam temperature for alloy 625 cladding in weld overlay tubes and coextruded tubes tested as part of superheater tube bundles at various WTE boilers. Source: Ref 10, 22

pg 340

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

pg 341

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 341

Table 12.2 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at indicated temperatures for 1008 h in N2-10O2-50ppmSO2-500ppm HCl Metal loss at 425 °C (800 °F) Alloy

µm/yr

Type 316 Type 347 Alloy 800HT Alloy 825 Alloy 600 Alloy 625

0.07 0.09 0.07 0.02 0.04 0.02

mpy

0.003 0.004 0.003 0.0008 0.002 0.0008

at 480 °C (900 °F)

at 590 °C (1100 °F)

µm/yr

mpy

µm/yr

mpy

2.54 2.29 1.02 1.27 2.03 1.78

0.1 0.09 0.04 0.05 0.08 0.07

5.08 7.11 5.33 2.54 3.05 2.79

0.2 0.28 0.21 0.1 0.12 0.11

1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 27

corrosion can be responsible for severe waterwall and superheater corrosion. Paul and Daniel (Ref 26) conducted laboratory tests at 315 °C (600 °F) for 720 h in simulated ﬂue gas environments with one being an oxidizing atmosphere (N2-2.6O2-15.0CO2-11.9H2O1000ppmSO2-1000ppmHCl) and the other a reducing atmosphere (N2-11.4CO2-7.0CO4.4H2-11.9H2O-1000ppmHCl). Specimens were not coated with salts nor covered with ash deposits. Both environments, containing about 1000 ppm HCl, showed that the corrosion rates for both carbon steel (SA178C) and Type 304 were less than 0.25 mm/yr (<10 mpy). Smith and Ganesan (Ref 27) conducted extensive tests in simulated ﬂue gas environments with HCl concentrations varying from 500 ppm to 10% and temperatures from 425 °C (800 °F) to 590 °C (1100 °F). Corrosion rates were found to be quite low for austenitic stainless steels and nickel-base alloys at temperatures up to 590 °C (1100 °F) in ﬂue gas environment containing 500 ppm HCl (Table 12.2). Even for the environment containing 4% HCl, the nickel-base alloys (alloys 825, 600, and 625) continued to show low corrosion rates. However, austenitic stainless steels exhibited high corrosion rates. This is illustrated in Table 12.3. A level of 4% HCl was more than an order of magnitude higher than what can be expected in a refuse-ﬁred combustion environment, and the temperature of 590 °C (1100 °F) was much higher than the superheater metal temperature of current WTE boilers. Nevertheless, alloy 625 as either coextruded cladding or weld overlay was observed to suffer wastage rates of more than 1.3 mm/yr (>50 mpy) in some aggressive refuse-ﬁred boilers (see Fig. 12.6). Table 12.3 shows that alloys 625 suffered a corrosion rate of about 1.53 mm/yr (60.4 mpy) when HCl was increased to 10% in the environment. Additional data for nickel-base alloys are shown in Table 12.4 (Ref 27). Devisme

Table 12.3 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at 590 °C (1100 °F) for 72 h in N2-9O2-12CO2100ppmSO2-4 and 10HCl Metal loss 4% HCl Alloy

Type 316 Type 347 Alloy 800HT Alloy 825 Alloy 600 Alloy 625

10% HCl

µm/yr

mpy

µm/yr

mpy

914.4 1244.6 73.66 20.32 25.4 15.75

35.7 48.5 2.9 0.8 1.0 0.6

… … … 1066 1219 1549

… … … 42.0 47.5 60.4

1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 27

et al. (Ref 28) conducted tests in Ar-5HCl and Ar-5HCl-10O2 environments at 600 °C (1110 °F), showing much higher corrosion rates in an oxidizing environment. However, the test tempered was much higher than the current superheater metal temperature. Their test results are summarized in Table 12.5. More corrosion data in halogen and halide environments are available in Chapter 6. It is generally agreed that chlorine in the fuel is an important cause of corrosion problems of boiler tubes. However, from the previous discussion, the direct corrosion of HCl, a gaseous component that formed when the chlorine in the fuel is combusted in the boiler, is not likely to be primarily responsible for the boiler tube corrosion. In Section 12.2, it was explained that there are a number of heavy metals, such as lead (Pb), zinc (Zn), cadmium (Cd), arsenic (As), and tin (Sb), along with alkali metals (i.e., Na, K) that are present in MSW. These metallic elements can react with chlorine to form various chlorides that have very low melting points. Many investigators (Ref 12, 14, 16, 20, 24, 29, 30) have considered that corrosion attack by low melting point chlorides is a principal mode of corrosion for boiler tubes in WTE plants.

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

pg 342

342 / High-Temperature Corrosion and Materials Applications

Table 12.4 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at 590 °C (1100 °F) for 72 h in N2-9O2-12CO2100ppmSO2-4HCl with 0% and 10% H2O Metal loss 0% H2O

10% H2O

Alloy

µm/yr

mpy

Type 316 Alloy 825 Alloy 600 Alloy 625

914.4 20.32 25.4 15.75

35.7 0.8 1.0 0.6

µm/yr

mpy

1.85 0.07 3.81 0.15 8.85 0.35 No measurable corrosion

1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 27

Table 12.5 Corrosion of nickel-base alloys in terms of metal loss after 500 h at 600 °C (1110 °F) in the indicated test environments

Ar-5HCl Ar-5HCl-10O2 Source: Ref 28

: Field test data 400 h, gas temp.: 735 °C (1355 °F)

0.30

: Laboratory test data 40-fold 100 h thickness loss 0.25 Melting point of deposits on field test tube 301 °C (570 °F)

0.20

0.15

0.10

0.05

0 0

Metal loss, μm (mils) Environment

exhibited melting points from approximately 330 to 650 °C (625 to 1200 °F), as shown in Fig. 12.8 (Ref 32). The ﬁgure also indicates that larger amounts of salt deposits melt at around 380 and 500 °C (715 and 930 °F). In the deposits collected from the 3000 h exposure tests, sulfur (as SO3) as high as about 40% and chlorine as high as 9%, along with Na, K, Pb, and Zn were detected. The authors observed that the corrosion attack increased with higher fused salt content in the deposits. This is illustrated in Fig. 12.9 (Ref 32) for Type 347 and alloy 625 in the corrosion probe tests. The ﬁgure shows the corrosion attack increases as the sum of the heats of fusion (or the amount of fused salt) of the deposit increases. The salts in the deposits are mainly chlorides and sulfates. Chlorides exhibit much lower melting points than sulfates. Under the operating conditions of WTE boilers, many metal chlorides can be in a molten state in the deposits on waterwall/screen tubes (evaporator tubes) and superheaters. This is illustrated in Fig. 12.10, showing many low melting point chlorides. Approximate steam temperatures at the waterwall and superheater are shown for general comparison (Ref 30). In Fig. 12.9, it is shown that increasing the amount of salts in the deposit could increase

Average corrosion thickness loss, mm/400 h

Kawahara and Kira (Ref 31) observed that the corrosion rate of carbon steel changed to a much higher rate above approximately 300 °C (150 °F), which corresponded to the melting point of the deposits collected from the tube samples in ﬁeld testing. This is illustrated in Fig. 12.7. They also performed laboratory tests, which involved actual deposits collected from the boiler to which ZnCl2 was added as 8% ZnO. The laboratory data conformed well with the ﬁeld data. The temperature at which carbon steel suffered accelerated corrosion attack was found to coincide with the melting point of the boiler tube deposits. The data strongly suggest that accelerated corrosion attack in the WTE boiler combustion environment was caused by the formation of molten phases in the tube deposits. Otsuka et al. (Ref 32) performed an extensive analysis on the salt deposits collected from three commercial WTE boilers (stoker type furnaces). A total of 23 salt deposits were collected from three boilers using corrosion probes at 550 °C (1020 °F) (probe metal temperature) for exposure of 700 and 3000 h, respectively. Each collected deposit was analyzed for the heat of fusion and the melting point by differential scanning calorimeter measurements. The heat of fusion is related to the amount of the fused salt in the deposit, with higher heat of fusion indicating a larger amount of fused salt. The deposits

C-276

600

601

214

35 (1.4) 120 (4.7)

50 (2.0) 140 (5.5)

90 (3.5) 160 (6.3)

30 (1.2) 55 (2.2)

200

250

300

350

400

Metal temperature, °C

Fig. 12.7

Corrosion in terms of thickness loss for carbon steel (SA178) as a function of the metal temperature in field exposure tests. Also superimposed are data generated from laboratory tests. Source: Ref 31

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

pg 343

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 343

the corrosion attack on the boiler tube, and Fig. 12.10 shows that there are many lowmelting salts, particularly metal chlorides, that can become molten at the operating temperatures of furnace walls and superheaters. The type of

35

30

Heat of fusion, J/g

25

20

15 10

chloride that forms on the deposit, as well as the type of the corrosion products formed between the chloride deposit and the underlying tube metal, can signiﬁcantly affect the corrosion attack. For example, some of the eutectic chloride salts involving FeCl3 can become molten at 150 to 205 °C (300 to 400 °F) as shown in Fig. 12.10. At higher metal temperatures, some chloride salts exhibit high vapor pressures. Figure 12.11 shows vapor pressures of some alkali metal salts (KCl, NaCl, KOH, and NaOH) and heavy metal chlorides (ZnCl2 and PbCl2) (Ref 26). K, Na, Zn, and Pb are among the most frequently detected elements in the tube deposits. It is generally believed that vapor pressures of about 10−4 atm are adequate to cause signiﬁcant high-temperature corrosion attack. The temperature at which the vapor pressure of ZnCl2 is at 10−4 atm is about 350 °C (660 °F) (see Chapter 6). At temperatures of 350 °C (660 °F)

5 0 300

400

500

600

PbCl2

700

Melting point temperature, °C

900 (480)

Fig. 12.8

Melting point temperatures versus the heat of fusion for the deposits collected at the corrosion probes (with the probe metal temperature of about 550 °C, or 1020 °F) in three commercial WTE boilers. The open data points are from the outer portion of the deposit, and the solid data points are from the inner portion of the deposit. Source: Ref 32

ZnSO4 Na2S2O7

800 (425)

PbCl2/FeCl2

KCl/PbCl2

NaCl/PbCl2

700 K2SO4·Na2SO4·ZnSO4 (370) (30–50%:10–30%:40–60%)

2

NaCl/FeCl2 KCl/FeCl2

600 (315)

1.5 K2S2O7 Max. waterwall fluid temperature

1

Temperature °F (°C)

Maximum corrosion thickness loss, mm/3000 h

Max. superheater steam temperature

NaCl/CaCl2 PbCl2/CaCl2 PbCl2/MgCl2

0.5

0 0

2

4

6

8

10

12

ZnCl2 500 (260)

400 (205)

NaCl/ZnCl2 SnCl2 ZnCl2/KCl ZnCl2/FeCl3 PbCl2/FeCl3

300 (150)

SnCl2/ZnCl2 NaCl/SnCl2 KCl/SnCl2

NaCl/FeCl3

Total sum of heat of fusion up to 550 °C, J/S

Fig. 12.9

Maximum thickness loss of corrosion probes exposed at metal temperature 550 °C (1020 °F) for 3000 h in WTE boilers as a function of the sum of heat of fusion (corresponding to the amount of fused salt) of the deposits. The open data points are Type 347H, and the solid data points are alloy 625. Source: Ref 32

Fig. 12.10

Melting temperatures of various salts that are likely to form in WTE boilers. The temperatures of waterwall saturated fluid and superheater steam are also indicated for comparison. Source: Ref 30

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

344 / High-Temperature Corrosion and Materials Applications

and higher, vapor-phase corrosion by ZnCl2 can be signiﬁcant. Below this temperature, ZnCl2 condenses, thus causing its vapor pressure below 10−4 atm and making vapor-phase corrosion less likely. Accordingly, waterwall corrosion is less likely to be caused by vapor-phase corrosion involving ZnCl2. The temperature at which the vapor pressure of PbCl2 is at 10−4 atm is about 485 °C (900 °F) (see Chapter 6). If the vapor phase of PbCl2 is involved in the corrosion attack, it is likely to be for superheaters. Both KCl and NaCl will be at much higher temperatures when these chlorides reach the vapor pressure of 10−4 atm. Thus, both KCl and NaCl are most likely to be involved in molten salt corrosion not vapor corrosion. Gleeson et al. (Ref 33) examined the effect of ZnCl2 on the corrosion of several iron- and nickel-base alloys using a modiﬁed Dean test with a three-zone furnace arrangement. The test involved a ﬂowing ﬂue gas (N2-3.6O2-14CO20.25SO2) passing ﬁrst through a crucible containing molten ZnCl2 salt at 425 to 540 °C (800 to 1000 °F) to pick up ZnCl2 vapor (1.3×10−3 atm at 425 °C, or 800 °F, and 2.8×10−2 at 540 °C, or 1000 °F), followed by the second zone heated to about 815 °C (1500 °F) with a catalyst to equilibrate the gas mixture, and to the third zone where specimens were exposed to a 540 °C (1000 °F) gas mixture containing ZnCl2 vapor. Because the gas mixture drops in temperature from 815 °C (1500 °F) in the

second zone to 540 and 510 °C (1000 and 950 °F) in the third zone where specimens were tested, the salt deposition rate onto the test specimen was found to be about 0.07 mg/cm2 h of molten ZnCl2. The test specimens were thus coated with molten ZnCl2 during the testing. Their test results are summarized in Table 12.6 (Ref 33). The test temperature of 510 °C (950 °F) is very close to some superheaters with steam temperatures of about 455 to 480 °C (850 to 900 °F). The corrosion rates of alloy 625 were on the same order of magnitude as those of alloy 625 (as a weld overlay or coextruded cladding)

Table 12.6 Corrosion rates after testing for 190 h in N2-3.6O2-14CO2-0.25SO2 containing ZnCl2 vapor, which condensed onto the test specimens as molten ZnCl2 at a rate of 0.07 mg/cm2 h Corrosion rate, mm/yr (mpy) Material

540 °C (1000 °F) front row

C-2000 625 C-22 214 230 G-30 825 HR-160 HR-120 556 Type 310

510 °C (950 °F) back row

… 2.97 (117) 2.11 (83) 2.18 (86) 2.87 (113) 4.17 (164), 4.42 (174) 3.15 (124) 3.45 (136) 5.26 (207) 3.63 (143) …

1.83 (72) 3.07 (121) 2.9 (114) 1.42 (56) 4.24 (167) … 4.72 (186) 2.18 (86) 8.94 (352) 7.11 (280) 7.77 (306)

Source: Ref 33

Temperature, °C 200

400

600

800

1000

1200

1 ZnCI2 10–1

KOH

Condensation occurs

Vapor pressure, atm

PbCI2

KCI

10–2

10–3

10–4

Species remains as a vapor NaOH

10–5 NaCI 10–6 600

800

1000

1200

1400

1600

Temperature, K

Fig. 12.11

Vapor pressures of some alkali and heavy metal salts as a function of temperatures. Source: Ref 26

pg 344

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 345

and corrosion products. Wright et al. (Ref 16) reported that the deposits (on a severely corroded waterwall tube) were found to contain 10 to 30% chloride, 33 to 47% Pb, 3 to 13% Zn, 3 to 14% Na, and 2 to 11% K. Spiegel (Ref 12) analyzed two overlay superheater tube samples that were severely corroded in German boilers. One tube sample was overlaid with Ni-16.7Cr-8Fe-4.6Si alloy, while the other overlaid with alloy 625. In both cases, the overlays were severely corroded. Analysis of the deposits in both cases showed salt melts were sulfates in contact with the ﬂue gas and chlorides at the deposit/scale phase boundary (Ref 12). The sulfate smelt was a CaSO4-K2SO4-Na2SO4 system including ZnSO4 and PbSO4. The chloride melt was KClZnCl2 and also NaCl. Heavy metals, such as Zn and Pb, along with Cl, have been detected in the corrosion front

in some very aggressive boilers, as shown in Fig. 12.6. It is thus believed that the severe corrosion of the alloy 625 overlay (or cladding) superheater tubes was most likely due to molten ZnCl2 and/or molten PbCl2. Because of considerable scattering in the data, the test results were not adequate for making an alloy ranking among nickel-base alloys. X-ray diffraction analysis of the scale and corrosion products was performed on alloys C-22, 625, G-30, 825, HR120, and Type 310, showing mainly Cr2O3 and NiCr2O4 for all alloys, with NiSO4 being detected for C-22, 625, 825, and Type 310 (Ref 33). From analysis of severely corroded tube samples obtained from operating boilers, it was learned that Cl, Zn, Pb, Cd, Na, K, and S, along with major alloying elements from the tube (e.g., Fe if the tube was steel, or Cr and Ni if alloy 625 overlay) were often detected in the deposit

4

60 µm

Fig. 12.12

Scanning electron micrograph (backscattered electron image) showing the deposits and corrosion scales formed on a carbon steel (SA178A) superheater tube suffering severe tube-wall wastage. Chemical compositions at different locations were analyzed by energy-dispersive x-ray spectroscopy (EDX) analysis (trace elements not reported here): 1: 31% Ca, 29% Si, 14% Mg, 15% Fe, 9% S, and 2% Zn 5: 72% Fe, 6% Cl, 7% Zn, 4% S, 4% Na, and 2% K 2: 63% Fe, 16% Cl, 9% Zn, 4% Pb, and 2% S 6: 88% Fe, 2% Cl, 2% Zn, 2% Cd, and 1% Na 3: 20% Fe, 13% Cl, 3% Zn, 41% Pb, 11% S, 4% Na, 3% K, and 2% Ca 7: 76% Fe, 8% Cl, 5% Zn, 2% Cd, 2% Na, and 2% S 4: 67% Fe, 12% Cl, 7% Zn, 4% S, and 6% Na

pg 345

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

346 / High-Temperature Corrosion and Materials Applications

of superheater tubes from operating boilers (Ref 22). This is illustrated in Fig. 12.12 (Ref 22) for a carbon steel superheater tube that suffered severe corrosion attack. The steam temperature and pressure of the superheater were reported to be 400 °C (750 °F) and 4.5 MPa (625 psig), respectively. The tube wall was reportedly reduced from the original 5.6 mm (0.220 in.) to about 1.02 mm (0.040 in.) after 11 months of service, with a wastage rate of approximately 5 mm/yr (196 mpy). Figure 12.12 shows the deposit and corrosion products. The corrosion front (location No. 5, 6, and 7 in Fig. 12.12) was found to contain Cl, Zn, Cd, Na, S, and Fe. The compounds are believed to contain iron chlorides with Zn, Cd, and Na.

The composition of the deposit and corrosion products can vary signiﬁcantly from plant to plant. This is shown in another boiler described in Fig. 12.13 and 12.14 (Ref 22), which involved an alloy 625 overlay superheater tube (410 °C, or 770 °F, steam) after service for about 6.5 months. Figure 12.13(a) shows the full cross section of the weld overlay with slight surface pitting attack, and Fig. 12.13(b) shows one of the corrosion pits at a higher magniﬁcation. The chemical compositions of the deposit and the corrosion products, which were analyzed by SEM/EDX, are shown in Fig. 12.14. Signiﬁcant amounts of lead were found throughout the deposit and corrosion products. It is also signiﬁcant that Pb, Zn, and Cl, along with Cr, Ni, and

(a)

0.025 mm

(b)

Fig. 12.13

(a) Optical micrograph showing slight pitting corrosion attack on alloy 625 overlay in an alloy 625 spiral overlay superheater tube after about 6.5 months of service. Micrograph (a) also shows the fusion boundary and substrate carbon steel. (b) Higher-magnification view of one of the corrosion pits. SEM/EDX analysis on the corrosion products in one of the surface pits is summarized in Fig. 12.14.

pg 346

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 347

Mo (major alloying elements from the weld overlay) were detected in the corrosion front, as indicated in locations No. 10 and 12 in Fig. 12.14. Both the morphology and the corrosion front that contained heavy metals, such as Zn and/or Pb, as were observed in the above two cases, suggest that the corrosion mechanism is ﬂuxing by molten chloride salts. In some cases, the corrosion morphology for a weld overlay of a Ni-Cr-Mo alloy involves general metal wastage followed by internal corrosion penetration along dendrites of the weld overlay. This internal dendrite corrosion attack (or

penetration) is in a way similar to internal oxidation attack or intergranular oxidation attack following general oxidation (or general metal wastage by oxidation) for wrought alloys. Similarly, internal corrosion attack (or intergranular corrosion attack) also takes place in wrought alloys under gaseous chloridation attack. Figure 12.15, shows general material wastage followed by internal dendrite corrosion penetration in the overlay for an alloy 622 overlay superheater tube after 225 days of exposure. This weld overlay tube was still in excellent condition, as shown in Fig. 12.16, showing the entire cross section of this alloy 622 overlay (on

30 µm

Fig. 12.14

Scanning electron micrograph (backscattered electron image) showing a localized corrosion pit on alloy 625 overlay of a superheater tube. SEM examination using energy-dispersive x-ray (EDX) spectroscopy showed the chemical compositions (wt%) at different phases, marked as No. 1 through No. 12. The chemical compositions (wt%) at different phases are: 1: 68% Pb, 11% Mo, 6% Cr, 3% Fe, 3% Ni, 4% S, 3% Cl, and trace elements 2: 63% Pb, 9% S, 6% Cl, 7% Cr, 5% Mo, 2% Fe, 3% Ni, 2% Na, and trace elements 3: 31% Cr, 24% Ni, 2% Fe, 27% Pb, 6% Zn, 5% S, 1% Cl, and trace elements 4: 32% Pb, 14% Cr, 8% Ni, 5% Fe, 10% Mo, 7% Zn, 9% Cl, 5% K, 3% Na, 3% S, and trace elements 5: 52% Pb, 11% K, 13% S, 7% Cl, 6% Mo, 4% Na, 4% Cr, 2% Ni, and trace elements 6: 34% Ni, 23% Cr, 3% Fe, 23% Pb, 7% Zn, 6% Mo, 2% Cl, and trace elements 7: 49% Pb, 12% K, 3% Na, 13% S, 7% Cl, 5% Cr, 6% Mo, 3% Ni, and trace elements 8: 30% Cr, 28% Pb, 13% Ni, 11% Zn, 7% Mo, 6% Cl, 2% K, and trace elements 9: 26% Pb, 22% Ni, 16% Cr, 11% Zn, 7% Cl, 5% Mo, 7% Na, 3% Fe, and trace elements 10: 27% Pb, 19% Cr, 14% Ni, 13% Zn, 9% Mo, 5% Cl, 4% Na, 2% K, and trace elements 11: 54% Ni, 24% Cr, 6% Pb, 5% S, 6% Fe, and trace elements 12: 27% Cr, 19% Ni, 19% Pb, 15% Zn, 9% Mo, 5% Cl, 2% K, 2% Fe, and trace elements

pg 347

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

348 / High-Temperature Corrosion and Materials Applications

a carbon steel tube). The morphology of this internal dendritic corrosion penetration is believed to be the result of gaseous corrosion reactions, but not by the molten salt ﬂuxing mechanism. In the same exposure test, an alloy 625 overlay tube was installed at the next platen on the same row as the alloy 622 overlay tube discussed in Fig. 12.15 and 12.16. After 152 days of exposure, the tube was removed for metallurgical examination. The alloy 625 overlay showed slight surface corrosion pitting attack (similar to alloy 622 overlay), but no internal dendritic corrosion attack, as illustrated in Fig. 12.17 (Ref 22). In another boiler where alloy 625 overlay superheater tube suffered severe corrosion attack, the overlay showed only general metal wastage with no internal dendritic corrosion attack (Ref 22). Montgomery and Larsen (Ref 34) also found that alloy 622 weld overlay, which was applied to the waterwall (the rear wall of the boiler in the ﬁrst pass) in a WTE boiler (Haderslev plant) in Denmark, showed internal dendrite corrosion penetration after 8000 h of exposure, while alloy 625 weld overlay at a similar location tested during the same time period showed no internal dendrite corrosion penetration after 8000 h of exposure. Both overlays showed only narrow pitting corrosion attack. Nevertheless,

Spiegel (Ref 12) observed similar internal dendrite corrosion attack on an alloy 625 overlay superheater tube where alloy 625 overlay suffered severe corrosion attack. He found that those internal dendrite corrosion phases were chlorides of mainly iron, chromium, and nickel, along with the corresponding oxides. He suggested that the molten chlorides reacted with oxides (formed on the metal) to produce chlorine, which then reacted with the metal to form metal chlorides. The internal dendrite corrosion penetration is similar to internal corrosion attack within the alloy matrix or internal grain-boundary corrosion attack in wrought alloys under high-temperature gaseous corrosion, such as oxidation and chloridation. In high-temperature gaseous corrosion, the extent of internal corrosion attack can vary from alloy to alloy. Some alloys may exhibit more internal attack than others. It is important to note, however, that the alloys with less internal oxidation attack are not necessarily more corrosion resistant. There is no clear indication that the weld overlay with internal dendrite corrosion penetration would be less resistant in terms of the alloy’s overall corrosion attack. It is believed that the corrosion rate due to general metal wastage by molten salt ﬂuxing is the rate-controlling factor.

0.025 mm

Fig. 12.15

Internal dendrite corrosion attack that followed general wastage was observed in the overlay of an alloy 622 weld overlay in an alloy 622 overlay superheater tube. Lower-magnification micrographs showing the through-thickness overlay is shown in Fig. 12.16(a) and general surface corrosion morphology is shown in Fig. 12.16(b).

pg 348

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 349

12.5 Corrosion Protection for Furnace Waterwalls In mass-burning units, the lower part of the waterwall is generally protected by refractories (e.g., SiC) (Ref 3, 7, 8, 16). The refractories provide protection for the waterwall against mechanical damage from abrasion of sliding large articles on the moving grate and also against corrosion from the combustion products. Relatively lengthy installation and upkeep time as well as reducing the heat-absorbing surface of the waterwalls can be an issue for refractory linings (Ref 20). Nevertheless, Licata et al. (Ref 17) indicated that refractory linings are needed to provide thermal insulation for ensuring that the designed combustion ﬂue gas

temperature is reached. For corrosion protection of the waterwall above the refractory, the current prevailing method is the application of alloy 625 weld overlay using automatic gas metal arc welding (GMAW) on site for existing boilers. The waterwall can also be constructed with shopfabricated weld overlay panels with alloy 625 overlay or with coextruded composite tubes with alloy 625 cladding. In RDF units, the same refractory design used for mass-burning units was tried initially by a major boiler designer resulting in slagging problems for the refractory wall surface (Ref 3). This was caused by the insulation of the lower furnace waterwalls by the refractory that resulted in higher ﬂame temperatures and caused signiﬁcant refractory wall slagging (Ref 3).

(a)

0.010 mm

(b)

Fig. 12.16

Alloy 622 overlay superheater tube after 225 days of exposure in a boiler. (a) Cross section of the overlay showing slight pitting attack. Micrograph (a) also shows the fusion boundary and substrate carbon steel. (b) Higher-magnification micrograph showing surface corrosion morphology with general metal wastage and internal dendrite corrosion penetration

pg 349

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

350 / High-Temperature Corrosion and Materials Applications

Following rapid corrosion attack of the bare carbon steel waterwalls in an RDF unit in Lawrence, MA, the lower furnace waterwall was weld overlaid with Inconel material in 1986, and this overlay protection proved to be very effective (Ref 3). The overlay alloy applied in the Lawrence unit was alloy 625 (Ref 23). The furnace waterwalls of modern RDF units are protected by alloy 625 cladding either as weld overlay applied on-site, or shop-applied panels, or as coextruded composite tubes with no refractory. Kubin (Ref 7) indicated in 1990 that the waterwalls of all Ogden RDF boilers were virtually fully covered with alloy 625 weld overlay. Figure 12.18 shows a general view of an automatic GMAW-applied alloy 625 weld overlay on the waterwall in a boiler. Alloy 625 weld

overlay on the waterwalls of the boiler was found to perform well in both mass-burning and RDF units. Generally, the wastage rates of alloy 625 overlay on waterwalls in U.S. boilers were quite low, typically approximately 125 μm/yr (5 mpy) (Ref 22, 25). The boiler tube metal temperatures are in a range of 260 to 315 °C (500 to 600 °F) for most U.S. boilers. For some European boilers running higher water/steam pressures, the waterwall tube metal temperatures could be higher. Figure 12.19 shows the cross section of an overlaid waterwall tube sample obtained from a RDF boiler in Lawrence, MA after 16 years of service. The metallographic cross section of the alloy 625 weld overlay of that overlay waterwall tube is shown in Fig. 12.20, revealing the full

(a)

0.010 mm

(b)

Fig. 12.17

Alloy 625 overlay superheater tube after 152 days of exposure in the superheater platen next to the alloy 622 superheater tube, which is shown in Fig. 12.16. (a) Cross section of the overlay showing slight pitting attack. Micrograph (a) also shows the fusion boundary and substrate steel. (b) Higher-magnification micrograph showing surface corrosion morphology with general metal wastage and no internal dendrite corrosion penetration.

pg 350

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 351

cross section of the overlay with little evidence of corrosion attack. Figure 12.21 shows a closeup view of alloy 625 overlaid waterwall after about 10 years of service in another RDF boiler. The common mode of corrosion attack on alloy 625 weld overlay on the waterwall has been found to be pitting attack. This is illustrated in Fig. 12.22. When corrosion or pitting attack becomes extensive, the corroded waterwall area can be repaired by ﬁrst grinding off the corroded metal prior to performing overlay welding, as shown in Fig. 12.23. It has been reported by Vrchota (Ref 25) that the bare carbon steel waterwall above the weld overlaid waterwall had experienced higher wastage rates as the boiler continued to increase its service duration. As a result, the weld overlay area has been “creeping” upward gradually, leading to application of weld overlay at increasingly higher elevation in the boiler. This situation has also been experienced in other plants (Ref 22). However, there is no consensus on a technical explanation about this “phenomenon.” Thermal sprayed coatings have not yet been used on a large scale in WTE boilers. One of the major issues is related to the intrinsic characteristics of the sprayed coating in terms of interconnecting pores that allow the corrosive to permeate through the coating and cause corrosion attack at the coating/substrate interface, thus leading to spallation. Another issue is lack of automatic application system that can cover a large waterwall area in achieving a consistent quality over the large coating area. In the 1990s, some tests on sprayed coatings were performed in boilers without satisfactory results (Ref 7, 16,

35). DeVincentis et al. (Ref 35) conducted corrosion probe tests on sprayed coatings of three nickel-base alloys (Ni-Cr-Co-Si alloy HF-160,

Fig. 12.19

Cross section of an overlaid waterwall tube showing alloy 625 overlay after 16 years of service in a RDF unit in Lawrence, MA. Courtesy of Welding Services Inc.

Fig. 12.20

Fig. 12.18 Services Inc.

General view of an overlaid waterwall with alloy 625 overlay in a WTE boiler. Courtesy of Welding

Optical micrograph showing the cross section of the alloy 625 overlay of an overlaid waterwall sample (Fig. 12.19) obtained from a RDF boiler (Lawrence, MA) after 16 years of service. Micrograph also shows the fusion boundary and substrate carbon steel. Courtesy of Welding Services Inc.

pg 351

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

352 / High-Temperature Corrosion and Materials Applications

Ni-Cr-Fe-Al-Y alloy 214, and Ni-Cr-Mo alloy C22). The corrosion probe tests were conducted in the ﬁrst pass of the front wall in a boiler at Hempstead plant for 76 days of exposure with the metal temperature maintained between 180 and 230 °C (350 and 450 °F). Coatings were applied by the inert gas electric arc spraying method. Disbonding of the coating during exposure was found to be a major problem. Alloy C-22 and 214 coatings were found to suffer disbonding after exposure. Kubin (Ref 7) indicated that a sprayed coating of alloy 50Ni-50Cr was tested in the upper furnace of a mass-burning unit and failed after 2.2 years of service. Furthermore, the removal of all the coating material in preparation for overlay welding with alloy 625 later was

Fig. 12.21

Close-up view of alloy 625 overlay on the waterwall of another RDF boiler after 10 years of service. Shown in the photograph are two tubes and a membrane. Courtesy of Welding Services Inc.

found to be an arduous and time-consuming operation (Ref 7).

12.6 Corrosion Protection for Superheaters Metal temperatures for superheater tubes can be 370 to 480 °C (700 to 900 °F) or higher. In this temperature range, there will be more chloride salts that become molten, as shown in Fig. 12.10. This may also result in a higher concentration of molten chloride salts in the ash deposits forming on the tube surface. Presence of more chloride salts will result in more corrosion attack (Fig. 12.9). At higher temperatures, some of the heavy metal chlorides, such as ZnCl2 and PbCl2, exhibit higher vapor pressures, as shown in Fig. 12.11. This can make the environment more corrosive for superheater tube materials. As discussed in Section 12.3, premature failures of superheater tubes made of carbon or low-alloy steels were quite common. Figure 12.24 shows the cross section of a carbon steel (SA178A) superheater tube removed after 11 months of service in a mass-burning unit (Ref 22). The superheated steam temperature and pressure for this boiler were 400 °C (750 °F) and 4.5 MPa (625 psig), respectively. Blough et al. (Ref 36) conducted corrosion probe tests in a mass-burning unit at Charleston Resource Recovery. A wide variety of commercial alloys were tested, including carbon and low-alloy steels, austenitic stainless steels, FeNi-Cr alloys, and nickel-base alloys. Tests were

Fig. 12.23 Fig. 12.22

Pitting attack is the common mode of corrosion of alloy 625 weld overlay on the waterwall. Shown in the photograph are three tubes and three membranes.

Repair welding can be performed on the corroded waterwall overlay by grinding followed by overlay welding. An alloy 625 weld overlay on the waterwall after 10 years of service in a RDF unit. Shown in the photograph are two tubes and three membranes.

pg 352

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:10PM Plate # 0

pg 353

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 353

Fig. 12.24

Carbon steel (SA178A) superheater tube after 11 months of service in a mass-burning unit. The superheated steam temperature and pressure were 400 °C (750 °F) and 4.5 MPa (625 psig), respectively. Courtesy of Welding Services Inc.

conducted for 4492 h. The results are summarized in Fig. 12.25 (Ref 36). Alloy 625 was found to be the best at all test temperatures. Austenitic stainless steels and Fe-Ni-Cr alloys and some nickel-base alloys except alloy 625 form a big scattering band between alloy 625 and ferritic steels. In examining the wastage proﬁles of the probe sample cross sections, some alloys showed the maximum wastage at the 90° position, where the ﬂue gas stream impinged upon, while others showed the maximum wastage at about 45° from either side from the 90° position. In another ﬁeld test conducted in an RDF unit at Elk River Station, Blough et al. (Ref 37) reported that alloy 625 was found to perform signiﬁcantly better than T-22, chromized T-22, Type 304H, HR3C, and 825 at both 470 and 500 °C (880 and 935 °F). Tests were conducted with tube samples of different alloys welded together as part of the lead superheater tubes after about 1180 h of exposure. The results are summarized in Fig. 12.26 and 12.27. The chromized layer was

Temperature, °C 205

260

315

370

425

480

535

0.160 Ferrite band

Ferritic steels 0.140

Austenitic steels 625

0.120

825

0.100

600

Max. wastage, in.

800H

Austenite band

HR160 0.080

0.060

0.040

625

0.020

0.000 400

450

500

550

600

650

700

750

800

850

900

950

1000

Temperature, °F

Fig. 12.25

Results of corrosion probe tests for various alloys in a boiler at Charleston Resource Recovery. The corrosion probes were installed in the convection path with the exposure time of 4492 h. Ferritic steels included carbon steel, T-22, and T-91. Austenitic steels included Type 304, 347, 310, and 27Cr-31Ni-3.5Mo (N08028). Source: Ref 36

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:11PM Plate # 0

pg 354

354 / High-Temperature Corrosion and Materials Applications

cladding in coextruded composite tubes has also been widely used in WTE boilers (Ref 10). For superheaters, the corrosiveness of the environment varies greatly from boiler to boiler. In some boilers, alloy 625 has performed well as an overlay or cladding in superheaters. One example is shown in Fig. 12.5. In some other boilers, the environment was so corrosive that alloy 625 cladding lasted for only about 1 year. For example, in boilers with superheated steam temperatures between 400 and 455 °C (750 and 850 °F), the alloy 625 cladding was found to exhibit a wide range of wastage rates from a negligible rate to 7.5 mm/yr (300 mpy), as shown in Fig. 12.6. The factors that affect the corrosiveness of the environment can be very complex. In his corrosion probe testing in an operating boiler, Krause (Ref 29) found that the ﬂue gas temperature could be important in affecting the corrosion rate of carbon steel. His data are shown in Fig. 12.28. The ﬁgure shows that when the ﬂue gas 2

80 315° wastage

60

1.5

40

1

20

0.5

Corrosion, mm

45° wastage

Corrosion, mils

found to be consumed in large areas and thus provided no protection. Kubin (Ref 7) had tested chromized coatings for superheater applications with mixed results, showing some success at one facility, but not at another facility. Furthermore, chromizing was not found to be cost effective (Ref 7). Other diffusion coatings, such as aluminizing and aluminum-silicon codiffusion coatings, were tested and found to be inadequate in performance (Ref 7). Alloy 625 as a weld overlay in spiral overlay tubing or a cladding in coextruded tubing offers a viable solution to the superheater corrosion problems. Table 12.7 summarizes the comparative performance between a bare carbon steel tube and an alloy 625 overlay tube in a ﬁnishing superheater in a side-by-side ﬁeld test in a boiler. The wastage rate was found to be about 2.8 mm/ yr (110 mpy) for carbon steel and 0.46 mm/yr (18.3 mpy) for alloy 625 overlay. Based on these data, the expected replacement interval for carbon steel tubes and alloy 625 overlay tubes are 1.4 and 5.8 years, respectively (Table 12.7). Figure 12.5 shows alloy 625 overlay superheater tubes (405 °C, or 760 °F, and 42 bar, or 609 psi, superheated steam) exhibiting excellent overlay condition with no sign of corrosion or erosion/ corrosion after 4.5 years of service in a boiler in the Netherlands (Ref 24). Alloy 625 overlay tubing has been widely used for superheater applications in WTE boilers. Alloy 625 as a

10

400 45° wastage

315° wastage

0

8 300

0 304

HR-3C T-22Cr 825

T-22

625

200 4

100

0

Corrosion, mm

Corrosion, mils

Alloys 6

Fig. 12.27

Tube metal loss of various alloys tested at 470 °C (880 °F) for 1180 h in an RDF unit at Elk River Station. T-22CR represents the chromized T-22 tube sample. The chromized layer (on T-22) was found to be consumed in a large area. Source: Ref 37

2

304

HR-3C T-22CR

825

T-22

625

0

Table 12.7 Wastage rates for the carbon steel tube and the alloy 625 overlay tube in a side-by-side test as part of a finishing superheater in a boiler in New York

Alloys

Fig. 12.26

Tube metal loss of various alloys tested at 500 °C (935 °F) for 1180 h in an RDF unit at Elk River Station. T-22CR represents the chromized T-22 tube sample. The chromized layer (on T-22) was found to be consumed in a large area. Source: Ref 37

Wastage rate Tube construction

Carbon steel 625 overlay

mm/yr

mpy

Expected tube life, years

2.8 0.46

110 18.3

1.4 5.8(a)

(a) Overlay (0.080″ thick) + steel tube. Source: Ref 24

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:11PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 355

temperature increased from 760 to 845 °C (1400 to 1550 °F), the corrosion rate of carbon steel was found to accelerate as the metal temperature exceeded 430 °C (800 °F). The author (Ref 29) believed that the increased corrosion rate was the result of the formation of volatile FeCl3, since below 430 °C (800 °F) the corrosion product was primarily FeCl2, which would not volatilize at these temperatures. This ﬂue gas temperature increase is not likely to signiﬁcantly affect the corrosion rate of alloy 625, which is a nickel-base alloy and not likely to form either FeCl2 or FeCl3. Metal temperature can signiﬁcantly affect the corrosion rate of the alloy. The corrosion reaction is a thermally activated process, thus increasing metal temperature can result in a higher corrosion rate. Furthermore, increasing metal temperature can increase the range of various chloride salts that become molten (Fig. 12.10). Increased amounts of molten chloride salts can also increase corrosion rates. Furthermore, increasing temperature can increase vapor pressures of chloride salts, thus resulting in increased corrosion by chloride vapors. The concentration of the chloride deposits and the type of chlorides can vary from boiler to boiler. Both the concentration of chloride salts and the type of chlorides are not normally monitored in plants. Furthermore, ﬂue gas velocity can be an important factor in the superheater wastage. In some cases, the superheater wastage can be the result of erosion/corrosion. The only

way to judge the severity of the superheater wastage issue in a particular boiler is to examine the historical data of that particular boiler. For “aggressive” boilers, efforts have been underway in the industry to identify an alloy that can outperform alloy 625. The results, however, have not been encouraging thus far. The ﬁndings of some of the comparison tests are summarized in the next paragraph. In a side-by-side ﬁeld test on alloys 625 and 622 overlay tubes (ﬁve-tube platen panel each)

(a)

Metal temperature, °C 200

0.6

600

760 °C (1400 °F) gas temperature 855 °C (1550 °F) gas temperature

0.5

13 11

0.4 9 0.3

7

0.2

5

Corrosion rate, µm/h

Corrosion rate, mils/h

400

3

0.1

(b) 1 0

0

200

400

600

800

1000

1200

Metal temperature, °F

Fig. 12.28

Effect of flue gas temperature on the corrosion rate of carbon steel in short-time corrosion probe tests (10 h exposure) in an operating boiler. Source: Ref 29

Fig. 12.29

Alloys 72 overlay superheater tube (a) and alloy C276 overlay superheater tube (b) after 7200 operating hours (10 months) in a RDF unit. The windward side of the tube, where the flue gas impinged upon the tube surface was the top side of the tube cross section as shown in the figure. The tube cross section was polished and etched with nital to reveal alloy 625 overlay (white portion of the metal) which was not etched by nital. Courtesy of Welding Services Inc.

pg 355

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:11PM Plate # 0

356 / High-Temperature Corrosion and Materials Applications

with the steam temperature of about 340 °C (640 °F) inside the tube and about 815 °C (1500 °F) ﬂue gas temperature, alloy 625 overlay tubes were exposed for 152 days and alloy 622 tubes for 225 days. The wastage rate was estimated to be about 175 μm/yr (7 mpy) for alloy 625 and about 355 μm/yr (14 mpy) for alloy 622. Alloy 622 containing 13% Mo and 3% W was found to be not as good as alloy 625 with about 9% Mo and no W but about 3.5% Nb. Both alloys contain about 21 to 22% Cr. Overlay tubes made of C-276 (Ni-16Cr-16Mo) and alloy 72 (Ni-44Cr) were also tested in an RDF unit, where alloy 625 overlay lasted for about 11 months. Both overlays were found to be not as good as alloy 625 overlay. The overlays of both alloys were consumed after about 10 months as shown in Fig. 12.29 (Ref 22). Alloy 686 (Ni-21Cr16Mo-4W) was compared with alloy 625 in a side-by-side test as overlays applied to the waterwall (72 bar and 290 °C, or 555 °F) of a boiler in Denmark (Ref 34). Evaluation of the overlay waterwall samples removed from the boiler after 8000 h of exposure showed both 625 and 686 overlays exhibited shallow pitting with pitting depth of about 50 µm. An alternate protection method for superheaters is the use of tube shields, which are typically made of Type 309, 310, and 253MA. Figure 12.30 shows carbon steel superheater tubes protected by metallic tube shields awaiting installation at a WTE plant. The shields are typically attached to the tube by mechanical straps and clamps with ﬁllet welds. The tube shields generally do not receive adequate heat transfer through the air gap between the shield

Fig. 12.30

and the tube. As a result, the shields can experience temperatures as high as those of ﬂue gas streams. The shields can thus suffer warping, distortion, and creep damage in addition to hightemperature corrosion. Accumulation of ash/salt deposits in the crevice behind the shields can further accelerate the corrosion attack. Loosened or fallen shields can cause problems by impeding the gas ﬂow. Tube shields are generally considered to be a “sacriﬁcial” part and are replaced regularly during the plant maintenance shutdown. Tube shields are sometimes made of cast stainless steels. A cast tube shield can be made much thicker than a wrought alloy sheet, thus enabling the tube shield to last until the next annual maintenance shutdown. Vrchota (Ref 25) reported that a 7.5 mm (0.3 in.) thick cast tube shield made of HD stainless steel (Fe-27Cr-5Ni) lasted for 12 months, which coincided with the maintenance shutdown cycle in an RDF boiler. To improve the heat transfer of the tube shield, silicon carbide cement was used to ﬁll the gap between the shield and the superheater tube. This allows for the reinstallation of new tube shields every 12 months.

12.7 Summary Materials issues related to waste-to-energy boilers for burning municipal solid waste (MSW) for electricity generation are presented. Combustion of MSW generates a very hostile environment for waterwall tubes and superheater tubes. The wastage rates of waterwall tubes made

Carbon steel superheater tubes protected by metallic tube shields awaiting installation at one WTE plant.

pg 356

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:11PM Plate # 0

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 357

of carbon or low-alloy steels have been found to be unacceptably high if no protection method is used. The corrosion is believed to result from molten chloride salts. The current, widely used method for protecting the waterwalls is the use of alloy 625 overlay cladding applied by automatic gas metal arc welding process. The waterwalls can also be constructed out of alloy 625/carbon steel coextruded tubes. Superheater tubes also require some methods of protection against corrosion attack. Alloy 625 overlay tubes and coextruded tubes have been used for superheaters successfully in many boilers. However, for some boilers with higher steam temperatures and/or more corrosive environments, alloy 625 overlay or cladding was found to be inadequate. An alternate corrosion protection method is the use of metallic tube shields or refractories. Tube shields are generally considered to be a “sacriﬁcial” part and are replaced regularly during the plant maintenance shutdown. Refractories require regular repair or replacement.

9. 10.

11.

12. 13.

14. REFERENCES

1. H.H. Krause, Historical Perspective of Fireside Corrosion Problems in RefuseFired Boilers, Paper No. 200, Corrosion/93, NACE, 1993 2. G. Sorell, The Role of Chlorine in High Temperature Corrosion in Waste-To-Energy Plants, Mater. High Temp., Vol 14 (No. 2/3), 1997, p 137 3. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, 1992 4. C.F. Knights, I.W. Cavell, and B.A. Phillips, Corrosion During Incineration of a Sulfur and Chlorine Bearing Mixture of Rubbers and Plastics, Werkst. Korros., Vol 40, 1989, p 163 5. E.A. Bretz, Energy from Wastes, Power, March 1990, p S-1 6. P. Rademakers, W. Hesseling, L.A. Tange, and R. Montaigne, “Review on Corrosion in Waste Incinerators, and Possible Effect of Bromine,” CEF-12, Laan van Westenenk, The Netherlands, 2002 7. P.Z. Kubin, Materials Performance and Corrosion Control in Modern Waste-ToEnergy Boilers Applications and Experience, Paper No. 90, Corrosion/99, NACE International, 1999 8. L. Strach and D.T. Wasyluk, Experience with Silicon-Carbide Tiles in Mass-Fired

15. 16.

17.

18.

19.

20.

21.

Refuse Boilers, Paper No. 219, Corrosion/ 93, NACE, 1993 R.L. Anderson, Wheelabrator Technologies Inc., private communication, 2006 A. Wilson, U. Forsberg, M. Lundberg, and L. Nylof, Composite Tubes in Waste Incineration Boilers, Stainless Steel World 99 Conference on Corrosion-Resistant Alloys (Conf. Proc.), Book 2, KCl Publishing BV, The Netherlands, 1999, p 669 F. Soutrel, C. Rapin, P. Steinmetz, and G. Pierotti, Corrosion of Fe, Ni, Cr and Their Alloys in Simulated Municipal Waste Incineration Conditions, Paper No. 428, Corrosion/98, NACE International, 1998 M. Spiegel, Salt Melt Induced Corrosion of Metallic Materials in Waste Incineration Plants, Mater. Corros., Vol 50, 1999, p 373 M. Noguchi et al., Experience of Superheater Tubes in Municipal Waste Incineration Plant, Mater. Corros., Vol 51, 2000, p 774 P.L. Daniel, L.D. Paul, and J. Barna, FireSide Corrosion in Refuse-Fired Boilers, Mater. Perform., May 1988, p 20 J.G. Singer, Ed., Combustion Fossil Power, 4th ed., Combustion Engineering, Inc., Windsor, CT, 1991 I.G. Wright, H.H. Krause, and R.B. Dooley, A Review of Materials Problems and Solutions in U.S. Waste-Fired Steam Boilers, Paper No. 562, Corrosion/95, NACE International, 1995 A.J. Licata, L.A. Terracciano, R.W. Herbert, and U. Kaiser, Design Features for Superheater Corrosion Control in Municipal Waste Combustors, Materials Performance in Waste Incineration Systems, G.Y. Lai and G. Sorell, Ed., NACE, 1992, p 5-1 H.H. Krause and I.G. Wright, Boiler Tube Failures in Municipal Waste-To-Energy Plants: Case Histories, Paper No. 561, Corrosion/95, 1995 A. Pourbaix, Corrosion of a Waste Incinerator: Effects of Design and Operating Conditions, Werkst. Korros., Vol 40, 1989, p 157 A.L. Plumley, W.R. Roczniak, and E.C. Lewis, Materials Performance of Heat Transfer Surfaces in A MSW-Fired Incinerator, Materials Performance in Waste Incineration Systems, G.Y. Lai and G. Sorell, Ed., NACE, 1992, p 7-1 W.G. Schuetzenduebel, I.E. Johnson, and C.W. Clemons, Accelerated Tube Metal

pg 357

Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/

30/10/2007 4:11PM Plate # 0

358 / High-Temperature Corrosion and Materials Applications

22. 23.

24.

25.

26.

27.

28.

29.

Wastage in Municipal Solid Waste Fired Furnaces, Materials Performance in Waste Incineration Systems, G.Y. Lai and G. Sorell, Ed., NACE, 1992, p 10-1 Welding Services Inc., unpublished data, Norcross, Georgia P.N. Hulsizer, Problems and Solutions in Applying Weld Overlay to Waste Boiler Incinerators, Materials Performance in Waste Incineration Systems, G.Y. Lai and G. Sorell, Ed., NACE, 1992, p 11-1 G.Y. Lai, “Corrosion Mechanisms and Alloy Performance in Waste-To-Energy Boiler Combustion Environments,” presented at the 12th North American Waste To Energy Conference (NAWTEC 12) (Savannah, GA), May 17–19, 2004 S. Vrchota, Fireside Corrosion Management in RDF Waste-To-Energy Boilers, Paper No. 5317, Corrosion/2005, NACE International, 2005 L.D. Paul and P.L. Daniel, Corrosion Mechanisms in Oxidizing, Reducing, and Alternating Combustion Gases in RefuseFired Boiler Environments, Paper No. 216, Corrosion/93, NACE, 1993 G.D. Smith and P. Ganesan, Metallic Corrosion in Waste Incineration: A Look at Selected Environmental and Alloy Fundamentals, Heat-Resistant Materials II (Conf. Proc.), Second International Conference on Heat-Resistant Materials, K. Natesan, P. Ganesan, and G. Lai, Ed., ASM International, 1995, p 631 F. Devisme, P. Falgoux, F. Lefebvre, and T. Flament, “High Temperature Corrosion in Atmospheres Containing Hydrogen Chloride,” presented at the 11th International Incineration Conference (Albuquerque, NM), May 11–15, 1992 H.H. Krause, Chlorine Corrosion in Waste Incineration, Materials Performance in

30. 31.

32.

33.

34.

35.

36.

37.

Waste Incineration Systems, G.Y. Lai and G. Sorell, Ed., NACE, 1992, p 1–1 I.G. Wright, V. Nagarajan, and H.H. Krause, Paper No. 201, Corrosion/93, NACE, 1993 Y. Kawahara and M. Kira, Corrosion Prevention of Waterwall Tube by Field Metal Spraying in Municipal Waste Incineration Plants, Corrosion, Vol 53 (No. 3), 1997, p 241 N. Otsuka, Y. Tsukaue, K. Nakagawa, Y. Kawahara, and K. Yukawa, A Corrosion Mechanism for the Fireside Wastage of Superheater Materials in Waste Incinerators, Paper No. 157, Corrosion/97, NACE International, 1997 B. Gleeson, J.E. Barnes, and M.A. Harper, Corrosion Behavior of Various Commercial Alloys in a Simulated Combustion Environment Containing ZnCl2, Paper No. 196, Corrosion/98, NACE International, 1998 M. Montgomery and O.H. Larsen, Field Investigation of Various Weld Overlays in a Waste Incineration Plant, Paper No. 5309, Corrosion/2005, NACE International, 2005 D.M. DeVincentis, S.P. Goff, J.W. Slusser, Z. Zurecki, and J.T. Rooney, Solving Fireside Corrosion in MSW Incinerators with Thermal Spray Coatings, Paper No. 198, Corrosion/93, 1993 J.L. Blough, G.J. Stanko, and M.T. Krawchuk, In Situ Materials Testing in a WasteTo-Energy Power Plant, Mater. High Temp., Vol 14 (No. 2/3), 1997, p 181 J.L. Blough, G.J. Stanko, W.T. Bakker, and T. Steinbeck, Superheater Corrosion in a Boiler Fired with Refuse-Derived Fuel, Heat-Resistant Materials II (Conf. Proc.), Second International Conference on HeatResistant Materials, K. Natesan, P. Ganesan, and G. Lai, Ed., ASM International, 1995, p 645

pg 358

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p359-377 DOI: 10.1361/hcma2007p359

31/10/2007 12:47PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 13

Black Liquor Recovery Boilers in the Pulp and Paper Industry 13.1 Introduction The black liquor recovery boiler is an integral part of the kraft pulping process in the pulp and paper industry. In the kraft process, wood chips are cooked in an aqueous solution, known as “white liquor,” containing primarily sodium hydroxide (NaOH) and sodium sulﬁde (Na2S) under pressure in a digester for manufacturing pulp. After cooking, the pulp is separated from residual liquor for paper manufacturing. The residual liquor, which contains wood lignins, organic materials, and inorganic compounds, then undergoes a series of evaporation processes to increase the concentration of solids from approximately 15 to 70 wt% or higher. This liquor with high concentrations of solid is known as “black liquor.” The black liquor is burned in a boiler to produce molten inorganic material, known as molten smelt, which is rich in sodium carbonate (Na2CO3) and sodium sulﬁde (Na2S). The molten smelt forms in a bed on the furnace ﬂoor and is discharged from the furnace ﬂoor into a water tank outside the furnace to form “green liquor” containing primarily sodium carbonate (Na2CO3) and sodium sulﬁde (Na2S). This green liquor undergoes the causticizing reaction to convert it to white liquor consisting of sodium hydroxide (NaOH) and sodium sulﬁde (Na2S) for use in cooking wood chips to manufacture pulp again. This chemical recovery process in the kraft pulping is described in detail in Ref 1 and 2. Because of its function of recovering cooking chemicals, the boiler is often referred to as a “chemical recovery boiler” or simply “recovery boiler.” In addition to recovering cooking chemicals for the pulping process, the boiler also recovers energy by producing process steam. Grant (Ref 3) reported in 1997 that about 216 recovery boilers which were operated by 130 mills in the United States produced approximately 41% of the total energy used by these mills.

The boiler construction is similar to that of coal-ﬁred boilers (Chapter 10), oil-ﬁred boilers (Chapter 11), and waste-to-energy (WTE) boilers (Chapter 12) using the tube-membrane waterwalls for the furnace enclosure, with the exception that the furnace ﬂoor is also enclosed with a tube-membrane construction. However, the combustion characteristics and problems are uniquely different from those of fossil-ﬁred boilers and WTE boilers. The ﬁrst and most signiﬁcant issue is the formation of molten smelt (or molten salts) on the furnace ﬂoor, and thus a boiler tube leak (under certain conditions) may lead to a smelt-water explosion (Ref 4). Recovery boilers have the potential for devastating explosions causing fatalities and injuries (Ref 4). Combustion of the black liquor in the boiler generates a sulfur-containing, reducing furnace environment in the lower furnace and a pile of char bed with molten smelt (or molten salts) on the ﬂoor tube bed. The major materials problems involved tube-wall thinning due to sulﬁdation when carbon steels were used and subsequent severe cracking problems for Type 304L/carbon steel coextruded tubes when the furnace ﬂoor tubes and waterwall tubes were upgraded with these coextruded tubes. A review of the current understanding on the tube cracking issues and improved alloys is presented in this chapter.

13.2 Fuel, Combustion, and Boilers Black liquor contains dissolved wood components, inorganic constituents (NaOH, Na2S, Na2CO3, Na2SO4, Na2S2O3, and NaCl), and water. An example of the elemental compositions of black liquor from North American wood species is shown in Table 13.1 (Ref 5). Black liquor is injected by spraying into the boiler through liquor guns at about 4.6 m (15 ft) above

pg 359

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:47PM Plate # 0

360 / High-Temperature Corrosion and Materials Applications

the boiler at the location above the secondary air ports. The liquor droplets injected into the boiler are generally 0.5 to 5 mm in size (Ref 7). The droplets should be small enough so they become dry or nearly dry before reaching the char bed to aviod smelt-water contact. However, they should not be too small, otherwise the droplets can be entrained in the ﬂue gas stream resulting in fouling and plugging in the convection path. (Ref 7). Large droplets may travel directly to the walls (Ref 7). In general, black liquor burns in approximately four stages, which are drying, devolatilization (pyrolysis), char burning, and smelt coalescence and reactions (Ref 7). The furnace can typically be divided into three zones: the reducing zone at the bottom, the drying zone where liquor is ﬁred, and the oxidizing zone in the upper furnace, as schematically illustrated in Fig. 13.2 (Ref 8). The ﬁgure also shows the char bed that forms on the furnace hearth during combustion of black liquor.

the hearth for some boilers (Ref 6). Air for combustion is injected into the furnace separately from primary and secondary air ports for substoichiometric combustion and also from tertiary air ports to complete the combustion. Figure 13.1 shows a schematic diagram of a recovery boiler with primary, secondary, and tertiary air ports (Ref 6). The liquor is shown to be sprayed into Table 13.1 Elemental compositions of black liquor from North American wood species Element

Concentration, wt%

Carbon Hydrogen Oxygen Sodium Sulfur Potassium Chlorine Nitrogen Others

34–39 3–5 33–38 17–25 3–7 0.1–2 0.2–2 0.04–0.2 0.1–0.3

Source: Ref 5

Steam flow to mill Stack gas Feedwater Sootblowing steam Induced draft fan

Boiler bank Superheater

Economizer

Waterwalls Bullnose

Electrostatic precipitator Dust Recycle

Blowdown

Forced draft fan Furnace Steam coil air heater

Liquor guns Black liquor mix tank

Tertiary air

Liquor heater

Secondary air Primary air Smelt spouts

Strong black liquor from concentrator

Fig. 13.1

Makeup salt cake

Recovery boiler. Source: Ref 6

Direct heating steam

Smelt to dissolving tank

pg 360

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 361

Flue gas

Oxidizing

Secondary air

Drying

Black liquor

Primary air

Reducing

Smelt

Fig. 13.2

Recovery boiler illustrating three zones of reactions in the furnace. Source: Ref 8

Formation of a char bed on the furnace hearth during combustion of black liquor makes the recovery boiler uniquely different from fossilﬁred boilers and WTE boilers discussed in previous chapters. The char bed generally consists of carbon, partially pyrolyzed black liquor solids, and molten and frozen smelt (Ref 9). The characteristics of char beds are different for different furnace ﬂoor designs. The sloped (or inclined) furnace ﬂoor, such as for B&W boilers, and the decanting ﬂoor, such as for Alstom CE boilers, exhibit different characteristics of char beds, and their characteristics are schematically illustrated in Fig. 13.3 and 13.4 for the sloped ﬂoor and decanted ﬂoor, respectively (Ref 9). For the sloped ﬂoor design, molten smelt ﬂows from the active char bed and is collected in troughs around the perimeter of the bed and then discharged through smelt spouts (Ref 9). For the decanting ﬂoor design, molten smelt is collected and contained by the decanting bottom of the boiler (Ref 9). Since the ﬂoor tubes and lower furnace waterwall tubes are cooled with water, a layer of the frozen (or solidiﬁed) smelt is formed on the tube surface, thus preventing rapid corrosion of tube metal by the molten salts. With studded carbon steel tubes, a thicker frozen layer is formed (Ref 10). The frozen smelt is not corrosive as long as it remains solid on the tube surface. However, once this frozen smelt layer suffers localized disturbances or instability,

Active layer • Pyrolysis, combustion, reduction • 10–25 cm (4–10 in.) thick • 800–1200 °C (1500–2200 °F)

Secondary air

Inactive core • Solidified smelt • Carbon • <760 °C (1400 °F) Primary air Hearth 230–290 °C (450–550 °F) Solid smelt Smelt spout

Fig. 13.3

Liquid smelt

Char beds formed in the sloped floor boiler. Source: Ref 9

pg 361

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

pg 362

362 / High-Temperature Corrosion and Materials Applications

Secondary air

Chemically active layer • Pyrolysis, combustion, reduction • 800–1200 °C (1500–2200 °F)

Physically active layer

Primary air

Liquid smelt Solid smelt Smelt spout

Fig. 13.4

Hearth 260–340 °C (500–650 °F)

Char beds formed in the decanting hearth boiler. Source: Ref 9

molten smelt may come within close proximity to, and possibly in direct contact with, the ﬂoor tubes, prompting severe tube overheating and thus premature tube failures (Ref 11). Hogan (Ref 11) reported several ﬂoor tube overheating incidents; in one case the tubes were overheated in excess of 675 °C (1250 °F) and in another case in excess of 720 °C (1330 °F). The author (Ref 11) attributed the overheating to localized disturbances or instability of the frozen melt on the tube surface. The furnace combustion conditions can be greatly affected by many boiler operating factors. For example, plugging of air ports by unburned material or unreacted smelt as well as plugging of liquor gun nozzles can greatly inﬂuence the furnace combustion conditions (Ref 12). Above the char bed up to the tertiary air ports, the gaseous environment remains under reducing conditions. Singh et al. (Ref 13) analyzed gas samples collected from the areas near the waterwall between the secondary and tertiary air ports in an operating boiler. The samples collected were analyzed using the gas chromatograph technique. The samples were collected from the area of the carbon steel waterwall that had experienced high corrosion. Their results are summarized in Table 13.2 (Ref 13). The amount of water vapor (% H2O) in the environment was

Table 13.2 The compositions of major gas species detected in gas samples collected from the different locations of the lower furnace waterwall between the secondary and tertiary air ports Composition, % Gas species

N2 CO2 CO H2 O2 CH4 H 2S SO2/COS Methyl mercaptan

At waterwall surface

2.5 cm (1 in.) away

24.8 29.4 7.5 2.4 2.9 4.0 18 0.2 0.17

32.9 23.9 9.1 2.3 1.0 2.0 24.2 0.27 0.51

30 cm (12 in.) away

47.6 18.3 16.8 1.5 1.1 0.9 0.6 0.04 0.002

Source: Ref 13

excluded from the data in the table. In addition, the authors indicated that during the present study, the black liquor was intentionally sprayed on the sidewalls for drying, thus leading to higher concentrations of H2S. It should be noted that the samples collected were not in equilibrium; thus, oxygen was detected. The results show that the concentration of H2S was signiﬁcantly reduced as the probe moved away from the waterwall. Hydrogen sulﬁde (H2S) is the major sulfur species of the gaseous sulfur compounds generated in the lower furnace. Other gaseous sulfur

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 363

compounds include methyl mercaptan, dimethyl sulﬁde, and dimethyl disulﬁde. These compounds are frequently referred to as “total reduced sulfur” (TRS) gases. In the lower furnace, the concentration of TRS gases is typically in a range of few hundred to a thousand ppm (Ref 14). As the ﬂue gas stream rises to the upper furnace and enters into the superheater bundles, the oxidizing atmosphere converts H2S to SO2. The ﬂue gas stream is also entrained with ﬂy-ash particles, which include carryover of black liquor droplets and fume that is formed by condensation of volatilized inorganic salts (Ref 15). These ﬂyash particles deposit on the cooler surfaces of the heat-exchanger tubes, such as superheaters and generating banks, in the convection path. The deposits consist mainly of sodium sulfate (Na2SO4) and sodium carbonate (Na2CO3) with small amounts of sodium sulﬁde (Na2S), sodium chloride (NaCl), and potassium salts. The ﬁrst melting temperature (FMT) of the deposits typically varies between 520 and 580 °C (970 and 1075 °F), depending on composition (Ref 16). The superheater tube can suffer accelerated wastage when the deposit in contact with the tube surface exceeds its FMT (Ref 17).

13.3 Materials Problems in Lower Furnace and Superheaters The ﬁrst Tomlinson recovery boiler built in 1929 had refractory furnace walls that were too costly to maintain (Ref 1). The furnace design was later changed to a completely water-cooled furnace enclosure including a furnace ﬂoor made of carbon steel, with the ﬁrst such unit built in 1934 (Ref 1). Lower furnace waterwalls and ﬂoor tubes were protected by pin studs that held frozen smelt providing protection against molten smelt (Ref 1). There have been several incidents of pin stud ﬂoor tube wastage and failure involving both sloped ﬂoor and decanting ﬂoor boilers (Ref 18). Clement and Blue (Ref 18) characterized the failure of the pin stud carbon steel ﬂoor tubes to be stud wastage, or burn-back, and tube-wall thinning. Since the 1970s, an increasing number of mills, ﬁrst in Scandinavia and then in North America, gradually upgraded carbon steel waterwalls and ﬂoors in the lower furnace to coextruded tubes with Type 304L outer cladding to provide corrosion protection for the inner carbon steel tube (Ref 19). Singbeil et al. (Ref 19) reported in 1997 that most of the 56 recovery

boilers in Scandinavia (34 in Sweden and Norway and 22 in Finland) had coextruded tube walls and more than half had coextruded tube ﬂoors, and in North America about 65 recovery boilers out of approximately 340 kraft recovery boilers had coextruded tube ﬂoors and many others had coextruded tube walls. One major U.S. boiler manufacturer (Babcock & Wilcox) installed coextruded tubes in the smelt ﬂow areas adjacent to the sidewalls with carbon steel studded tubes in the center of the ﬂoor in 22 boilers and in the complete ﬂoor for 26 boilers (Ref 18). In the 1990s, cracking of the 304L cladding of the composite tubes had become a major materials issue in recovery boilers, although some cracking was reported in the 1980s (Ref 19, 20). In 1995, a United States Department of Energy (DOE) program was established to determine the cause of the tube cracking and to identify alternative materials or process changes to prevent this type of cracking. This project, coordinated by J.R. Keiser of the Oak Ridge National Laboratory (ORNL), was carried out by researchers at ORNL, the Pulp and Paper Research Institute of Canada (PAPRICAN), the Institute of Paper Science and Technology (IPST) with support of more than a dozen paper companies, and boiler and tubing manufacturers (Ref 20). The results of the research projects under this program, which were presented at regular ORNL review meetings and published in various technical journals, provided a signiﬁcant understanding on the root causes of the 304L composite tube cracking problems. This chapter reviews (a) the cracking problems of the 304L cladding in composite tubes used as ﬂoor tubes, smelt openings, and primary air port openings and (b) alternative materials. Superheater corrosion is also included in the discussion. 13.3.1 Furnace Floor Tubes Cracking of the 304L cladding has been found to occur in (a) both high-pressure (8.3 to 10.3 MPa, or 1200 to 1500 psi) and low-pressure (4 to 6.2 MPa, or 600 to 900 psi) boilers, (b) both sloped ﬂoor and decanting ﬂoor designs, and (c) boilers by different boiler manufacturers (Ref 19). Cracks typically initiated from the outer diameter of the clad tube and propagated inward to the substrate steel; in most cases, the cracks terminated at the cladding/steel interface with some cracks changing the direction of propagation along the cladding/steel interface (Ref 19–21). This is illustrated in Fig. 13.5 for the

pg 363

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

pg 364

364 / High-Temperature Corrosion and Materials Applications

1250 µm

500 µm

Fig. 13.5

Fig. 13.6

clad tube and Fig. 13.6 for the clad membrane. These cracks were found to be essentially transgranular in nature (Ref 19, 20). In order to determine the cracking mechanisms, the DOE program included the following studies: (a) measurement and modeling of residual stresses and the stress states of the composite tubes, (b) tube metal temperature measurements during operation, (c) thermal fatigue testing of tube materials, and (d) the effect of waterwashing during shutdowns (Ref 20). To determine whether thermal fatigue was a possible cause of ﬂoor tube cracking, ﬂoor tube metal temperature was measured in a boiler at the Weyerhaeuser mill in Prince Albert, Saskatchewan, Canada, during September 1, 1998 through February 28, 1999 (Ref 22). This boiler with a sloped ﬂoor had experienced cracking of Type 304L cladding in coextruded ﬂoor tubes. In many boilers, cracking had occurred within 2 m of the rear (spout) wall. Keiser et al. (Ref 22) installed 25 thermocouples on the ﬂoor and recorded the temperature spikes over a 6-month period, showing an average of about one thermal spike per thermocouple per day. A thermal spike was deﬁned as a temperature spike of more than 50 °C (90 °F) above approximately 275 °C (530 °F). The authors observed that the thermocouples located near the spout wall that suffered tube cracking showed the fewest thermal spikes, while the areas farther from the spout wall that exhibited little or no cracking experienced the greatest numbers of thermal spikes (Ref 22). The authors concluded that the frequency of thermal spikes was far too low for thermal fatigue to be the cause for the cracking (Ref 22).

Thermal fatigue data for Type 304H and 304L cycling to 500, 550, and 600 °C (930, 1020, and 1110 °F), respectively, were generated and compared with the ASME Sec. III Subsec. NH design curve for Type 304H at 430 °C (800 °F) in isothermal fatigue. All the thermal fatigue data fell near or above the ASME design curve (Ref 20, 23). Assuming thermal cycling to 450 °C (840 °F) is considered, the cyclic strains would be about 0.25% and the fatigue life for the stainless steel cladding is expected to be in excess of 100,000 cycles (Ref 23). There is some thermal expansion mismatch between the 304L cladding and the carbon steel substrate in the composite tube due to differences in mean coefﬁcients of thermal expansion between these two alloys (e.g., 7.6× 10−6 and 10.0 × 10−6 in./in. · °F from 70 to 800 °F for carbon steel and Type 304, respectively). The residual stresses formed in the cladding in both the as-fabricated condition and after service were studied under the DOE program. Also investigated in the program was ﬁnite element modeling of stresses in the cladding. Keiser et al. (Ref 20) summarized the results of both residual stress measurements and ﬁnite element modeling of the 304L/SA210 coextruded composite tubes. Residual stress measurements using x-ray and neutron diffraction showed compressive axial and tangential stresses on the outer surface of the cladding of the as-fabricated tube, and tensile axial and hoop stresses (up to a maximum of 300 MPa) on the crown location of the coextruded tube after service in the boiler ﬂoor (Ref 20). Finite element modeling studies showed that the 304L cladding developed tensile stresses when cooled down

Cracks initiated on the outer diameter of the 304L clad tube, propagated inward to the substrate steel and terminated at the cladding-steel interface. Courtesy of Oak Ridge National Laboratory.

Cracks initiated on the outer surface of the 304L cladding, propagated inward to the substrate steel of the membrane and terminated at the cladding-steel interface. Courtesy of Oak Ridge National Laboratory.

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 365

from the “operating” temperature (e.g., 300 °C, or 570 °F) to room temperature (Ref 20). Furthermore, the 304 cladding developed tensile stresses when cooled down from the thermal spike to the “operating” temperature (e.g., cooling down from 550 to 300 °C, or 1020 to 570 °F) (Ref 20). As a result, tensile stresses can be present in the 304 cladding at “operating” temperature after a thermal spike as well as when ﬂoor tubes are cooled to room temperature. The presence of tensile stresses in the 304L cladding when the composite ﬂoor tubes are at low temperatures is signiﬁcant in that it can make stress-corrosion cracking (SCC) a potential cracking mechanism since the morphology of cracking resembles that of SCC. Prescott and Singbeil (Ref 24) reported that Type 304L suffered stress-corrosion cracking under tensile stresses when exposed to concentrated solutions of alkali compounds containing Na2S at temperatures from 160 to 200 °C (320 to 390 °F). In C-ring tests of Type 304L/SA210A1 coextruded tubes in a salt consisting of 90% Na2S·9H2O and 10% NaOH at about 170 °C (340 °F) for 48 h, cracks developed at the outer diameter of the cladding and propagated inward and terminated at the cladding/substrate interface (Ref 20). Stress-corrosion cracking of Type 304L takes place in this type of environment within a certain temperature range. For example, in constant load tests with 275 MPa (40 ksi) in Na2S-10%NaOH, stress-corrosion cracking of Type 304L readily occurred at 150 to 250 °C (300 to 480 °F) (Ref 23). It is generally believed that waterwashing of the boiler when there are signiﬁcant remnants of smelt on the ﬂoor results in aqueous solutions that contain Na2S along with NaOH or Na2CO3. These aqueous solutions are believed (Ref 20, 23, 25, 26) to be an essential requirement for stress-corrosion cracking of the 304L cladding either during the boiler shutdown or startup when the ﬂoor tubes are covered with these salt solutions and subject to this SCC temperature range. The critical temperature range is believed to be 150 to 200 °C (300 to 400 °F) (Ref 23). C-ring tests were conducted in a salt consisting of 90% Na2S·9H2O and 10% NaOH at about 170 °C (340 °F) for 48 h for Type 304L/SA210A1 coextruded tube, Sanicro 38*

* Sanicro 38 meets UNS N08825 (alloy 825) according to Sandvik Steel's technical bulletin on Sanicro 38/4L7 coextruded tubes (Ref 27), thus the cladding should be considered alloy 825, not modiﬁed alloy 825.

(modiﬁed alloy 825) coextruded tube (on carbon steel), Type 309 overlay tube, alloy 625 overlay tube, Type 310 coextruded tube, and carbon steel (Ref 20). The test results of these tubes are summarized: (a) Type 309 overlay tube was found to be as bad as the 304L coextruded tube, (b) Type 310 coextruded tube also suffered SCC cracking with the longest crack being about 325 µm, (c) Sanicro 38 cladding (modiﬁed alloy 825) exhibited cracks, but less than 50 µm deep, (d) alloy 625 overlay showed no cracking, and (e) carbon steel showed no cracking (Ref 20). One of the conditions required for developing stress-corrosion cracking is the presence of tensile stress. In this regard, alloys 825 and 625 have advantages over Type 304L in that both exhibit coefﬁcients of thermal expansion much closer to those of carbon steels. Instead of developing axial tensile stresses in Type 304L when cooled from the operating temperature to room temperature, both alloys 825 and 625 develop axial compressive stresses (Ref 23). This is important in that cracking predominantly occurred in the transverse direction (Ref 23). The alloy 625 overlay in the as-overlaid tube typically exhibits residual tensile stresses that can be eliminated by annealing the tube at 900 °C (1650 °F) for 20 min (Ref 28). The overlay tubing is produced by a spiral overlay welding mode using gas metal arc and gas tungsten arc welding processes (Ref 29), and its structure and properties are reported elsewhere (Ref 30, 31). Keiser et al. (Ref 26) reported the results of the tests comparing austenitic stainless steel claddings of Type 304L coextruded tubes and Type 309L overlay tubes with nickel-base alloy claddings of alloy 825 coextruded tubes, alloy 625 coextruded tubes, and alloy 625 overlay tubes in a recovery boiler in North America. Both 304L coextruded tubes and 309L overlay tubes showed cracking during the second year of service, while alloy 825 coextruded, 625 coextruded, and 625 overlay tubes showed no cracking after 5 years of service. Barna and Rivers (Ref 21) reported in their 1999 paper that several ﬂoors have been constructed with alloy 825 coextruded tubes since 1996. Wilson et al. (Ref 32) reported good performance results of Sanicro 38 coextruded tubes as ﬂoor tubes in recovery boilers. The authors (Ref 32) indicated that the outer cladding alloy in Sanicro 38 coextruded tube was a modiﬁed alloy 825. Nevertheless, the chemical composition of

pg 365

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

366 / High-Temperature Corrosion and Materials Applications

the outer cladding published in the paper was within the speciﬁcation of UNS N08825 (alloy 825). Sandvik Steel’s technical bulletin (Ref 27) on Sanicro 38/4L7 composite tubes also indicates that the cladding meets UNS N08825. Lai and Wensley (Ref 31) reported the performance experience of alloy 625 overlay ﬂoor tubes in several boilers where good results were observed. One boiler (a Babcock & Wilcox design) operated reportedly at 1420 psig and produced 883,400 lb/h steam. After severe cracking of Type 304L coextruded tubes in 1995, alloy 625 overlay tubes were installed in the smelt runs in 1997 replacing the 304L coextruded tubes. Figure 13.7 shows a general view of the smelt run ﬂoor tubes in the boiler (Ref 31). The alloy 625 overlay tubes had been in service for about 7 years with no reported material problems when the paper was published in 2005. Barna and Rivers (Ref 21) also reported the installation of alloy 625 overlay tubes for boiler ﬂoors in several boilers. One installation involved covering the entire width of about onehalf the boiler ﬂoor in 1996, and three other boilers were ﬁtted with partial ﬂoor installations of alloy 625 overlay tubes in 1995 and 1997. No cracks or other damage were reported when the paper was presented in 1999 (Ref 21).

Fig. 13.7

13.3.2 Smelt Spout Openings Cracking of Type 304L coextruded tubes was initially observed at smelt openings in the early 1980s before the observation of cracking of Type 304L coextruded ﬂoor tubes (Ref 21). Cracks can occur on adjacent tubes that form the spout wall (Ref 21). The circumferential cracks that occur in the 304L coextruded tubes at the smelt openings have been found to propagate from the outer stainless steel cladding through the interface and into the carbon steel substrate (Ref 19). Smelt openings are subject to thermal ﬂuctuations due to intermittent exposure of the ﬂowing smelt through the spout (Ref 21, 33). The wall tubes including the tubes adjacent to spout openings are also subject to thermal ﬂuctuations due to rising and falling of the surface of the smelt pool (Ref 18). It is believed that thermal fatigue cracking may be a contributing factor for cracking of smelt opening tubes made of Type 304L coextruded tubes (Ref 18). Dykstra et al. (Ref 34) attributed the cracking of Type 304L coextruded smelt spout opening tubes to thermal fatigue. The authors (Ref 34) indicated that fractographic examination of the crack surfaces showed evidence of cyclic crack progression. In addition to the thermal fatigue type cracks, craze cracks were also observed in Type 304L

General view of the alloy 625 overlay smelt run floor tubes in the boiler. Source: Ref 31

pg 366

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 367

coextruded spout opening tubes (Ref 34). These craze cracks (typically branched) were found to terminate at the cladding/steel interface (Ref 34). In order to investigate alternate cladding alloys, Clement and Blue (Ref 18) conducted thermal fatigue testing of alloy 825 coextruded tube and alloy 625 overlay tube compared with Type 304L coextruded tube and Type 304L monolithic tube. The substrate tubes for the above three composite tubes were all SA210 A1 steel. The test involved a water-cooled test tube that was rotated to make the tube-wall temperature ﬂuctuate from the cladding temperature at the point of ﬂame impingement to a low of about 40 °C (105 °F). The temperature of the cladding at the point of ﬂame impingement was measured inside the cladding at the point about 0.25 mm (10 mils) from the cladding surface. The cladding metal temperature cycled from about 650 °C (1200 °F) (high) to 40 °C (105 °F) (low). Solid 304L tubes failed in less than 5000 cycles, and the 304L clad tubes failed in slightly over 10,000 cycles to less than 10,000 cycles. Alloy 825 coextruded tubes cycled for 25,000 cycles showing no cracks, and alloy 625 overlay tubes cycled for 30,000 cycles showing no cracks. Both 825 coextruded tubes and 625 overlay tubes were then cycled from 815 °C (1500 °F) (high) to 40 °C (105 °F) (low) for

Fig. 13.8

30,000 cycles and showed no cracking. The test results showed that both alloy 825 coextruded tubes and alloy 625 overlay tubes were very resistant to thermal fatigue cracking. Barna and Rivers (Ref 21) reported in 1999 that there were about 60 smelt openings that were made of alloy 825 coextruded tubes. However, it has been reported that cracking of alloy 825 cladding was encountered in some boilers (Ref 21). Furthermore, the cracks were found to terminate at the fusion boundary or turn at 90° and proceed along the fusion boundary (Ref 21). The authors (Ref 21) also reported that alloy 625 overlay tubing was ﬁrst installed in a smelt opening in 1987 in a boiler where Type 304L coextruded tube smelt openings required replacement approximately every 6 months as well as several other installations including one boiler where alloy 625 overlay smelt opening tubes were installed in 1994. Lai and Wensley (Ref 31) reported that a smelt spout wall (including opening tubes) made of alloy 625 overlay tubes was installed in a B&W 1050 psi boiler in 2000 replacing Type 304L coextruded tubes that had experienced cracking problems. At the time the authors (Ref 31) presented their paper in 2005, no cracking problems had been reported. Figure 13.8 shows that the alloy 625 smelt spout wall including opening tubes revealed no

Alloy 625 overlay smelt spout wall including opening tubes after 2 years of service showing no indication of cracking by liquid dye penetrant testing. Source: Ref 31

pg 367

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

368 / High-Temperature Corrosion and Materials Applications

cracking in dye penetrant testing after 2 years of service (Ref 31).

13.3.3 Air Port Openings Cracking in both craze (or mosaic) and circumferential patterns has been observed in Type 304L coextruded tubes at the air port openings (Ref 35–38). These authors observed that some of the circumferential cracks penetrated into the carbon steel tube. Most cracks, however, did not progress into the carbon steel tube (Ref 36). Almost all the cracking was found to occur at the bottom half of the air port openings (Ref 36, 38). Temperature measurements of the air port opening tubes in several mills have indicated that those air port opening tubes that had historical cracking problems showed the greatest temperature ﬂuctuations in frequency and amplitudes (Ref 36). The air port opening tubes with no historical cracking problems exhibited no signiﬁcant thermal ﬂuctuations (Ref 36). In one example, based on the 5-day temperature measurements, the air port opening tubes with no history of cracking problems in one mill showed temperature ﬂuctuations between 300 and 400 °C (570 and 750 °F), while the other mill with a history of cracking problems showed extensive thermal ﬂuctuations from about 350 °C (660 °F) to temperatures exceeding 500 °C (930 °F) (Ref 36). In making mill visits to inspect the air port openings, the authors (Ref 36) observed that the air port opening tubes with a shorter, wider design appeared to suffer more frequent cracking problems than the ones with a longer, narrow opening design. This certainly suggests that the cold-worked conditions of the tube bends and/or the increased stresses due to increased constraint might be a factor. The results of the residual stress measurements using x-ray and neutron diffraction methods showed that the 304L coextruded tube typically exhibits compressive axial stresses on the surface of the cladding at the unbent section and tensile axial stresses in some areas of the bent section (Ref 37). One interesting observation was made with a video camera used to record the air port opening activities in operating boilers. What appeared to be molten smelt, which was found on the opening tube surfaces, ﬂowed down the tube surfaces (Ref 37). This suggests that molten smelt could likely be on air port opening tube surfaces during operation, thus resulting in possible local

temperature excursions (spikes) and corrosion reactions with the tube surface. Shenassa et al. (Ref 39) indicated that the air port design and fabrication may be a factor in causing primary air port cracking problems. The authors (Ref 39) discussed two air port opening designs (the casting design and the nonwelded insert design) provided by a major European boiler designer. The opening is formed by bending one or two composite tubes out of the plane of the waterwall toward the cold side of the boiler. It was observed that the casting design, which involved a cast iron casting being bolted to the bent tubes from the cold side, appeared to show better resistance to cracking because of a low bending angle and the optimum bending radius of the opening tubes (Ref 39). This observation was consistent with what was observed by Keiser et al. (Ref 36), who observed that opening tubes with a shorter, wider design appeared to suffer more frequent cracking problems than the ones with a longer, narrow opening design. In a search for alternate alloys in coextruded tubes to replace Type 304L coextruded tubes in primary air port openings with the same air port design, alloy 625 coextruded tubes were installed. After 1 year of operation, cracking was observed on many of the primary air port opening tubes (Ref 39). Alloy 825 coextruded tubes have been installed in primary air port openings in several boilers including one unit with alloy 825 coextruded primary air port openings having been in operation for 3 years; no cracks had been observed on these air port opening tubes (Ref 39). Keiser et al. (Ref 36) reported that the inspection of a mill revealed extensive thinning at the ﬁreside of alloy 825 coextruded air port opening tubes with no evidence of cracking. Lai and Wensley (Ref 31) reported severe tube thinning for primary air port openings fabricated from spiral overlay Fe-20Cr-1Nb composite tubes after 1 year of service in a mill in South America. A close-up view of the ferritic alloy overlay tubes is shown in Fig. 13.9. Keiser et al. (Ref 37) reported a case where the air port opening made of alloy 625 overlay tubes was found to suffer overlay thinning of about 1 mm (0.04 in.) after the ﬁrst 6 months of service. However, no further additional wastage was observed during subsequent exposure (Ref 37). In fact, Wensley (Ref 40) reported that there was no further overlay thinning after an additional 5 years of operation for this boiler. It is believed that the initial corrosion was likely to result from

pg 368

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 369

changes in liquor ﬁring practices (Ref 40). Furthermore, no cracking has been encountered for these air port opening tubes (Ref 40). Lai and Wensley (Ref 31) summarized the performance experience of alloy 625 overlays in primary air port openings. In a B&W boiler operating at 1420 psig and producing 883,400 lb/h steam, about 29 air port openings were replaced with alloy 625 overlay tubes in 1999. These tubes have been inspected annually using dye penetrant testing (PT). No cracking was reported at the time the authors presented their paper (Ref 31). Figure 13.10 shows some of these alloy 625 overlay air port opening tubes. In another boiler, alloy 625 overlay air port opening tubes were installed in 2000, and the inspection of these tubes in 2002 did not reveal any cracks or corrosion attack (Ref 31). However, inspection a few years later showed cracks were developed in few of these overlay tubes, and these cracks had advanced at least to the carbon steel (Ref 41). The performance experience of alloy 625 coextruded tubes used in air port openings has been reported by several authors (Ref 36, 37, 39). Shenassa et al. (Ref 39) reported that air port openings made of alloy 625 coextruded

tubes (Sanicro 63) in a mill suffered cracking after only 1 year of operation. Cracking was observed in many of the primary air port openings (Ref 39). Keiser et al. (Ref 37) observed that the circumferential cracks that occurred in alloy 625 coextruded tubes were intergranular in nature (i.e., cracking along grain boundaries) (Fig. 13.11). Intergranular cracking was also

Fig. 13.10

Primary air port openings fabricated from alloy 625 overlay tubes have been in service since 1999. No cracking or corrosion has been observed. The photo was taken after liquid dye penetrant testing. Source: Ref 31

Fig. 13.9

Primary air port opening made of Fe-20Cr-1Nb spiral overlay tubes suffered severe fireside wastage after 1 year of service in a mill in South America. Source: Ref 31

Fig. 13.11

Intergranular cracking that occurred in the alloy 625 coextruded air port opening tube. Courtesy of Oak Ridge National Lab.

pg 369

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

pg 370

370 / High-Temperature Corrosion and Materials Applications

observed in alloy 625 coextruded ﬂoor tubes and smelt opening tubes (Ref 42). Primary air port openings made of alloy 625 coextruded tubes in a boiler in the Southeast (U.S.) were also found to suffer primarily circumferential cracking after 1 year of service (Ref 43). The cracking was also found to be intergranular (Fig. 13.12). However, these circumferential cracks were found to terminate at the cladding/ steel interface and to change the direction of propagation to follow the cladding/steel interface, as shown in Fig. 13.13 (Ref 43). Corrosion products formed inside the crack at the top portion (Fig. 13.14) and at the crack tip (Fig. 13.13b) were analyzed using energy-dispersive x-ray spectroscopy (EDX). In the top portion of the crack, the corrosion products appeared to be primarily Ni-rich oxides (Fig. 13.14). The oxides formed at the crack tip (Fig. 13.13b) were Ni-Cr rich (area No.3) before the crack reached the cladding/steel interface and Fe-Ni rich (area No.1 and 2) when the crack propagated along the cladding/steel interface. It is more important to note that both Na and K were detected in these corrosion products, although it is not clear what role these alkali metals played in cracking. Very little data have been reported on the coldwork condition of the alloy 625 coextruded tube that was formed for the air port opening. A primary air port opening tube made of alloy 625 coextruded tube that suffered circumferential cracking after 1 year of service in a boiler in

Southeast (U.S.) was removed from the boiler and examined for the cold-work condition at different locations of the tube. The examination was carried out by conducting microhardness

600 µm

(a)

3 2

1 (b)

Fig. 13.13

Fig. 13.12

Intergranular cracking of the alloy 625 coextruded air port opening tube in a mill in Southeast (U.S.) after 1 year of service. The microstructure was not etched, but cracking could be traced to follow austenitic grain boundaries. Courtesy of Welding Services Inc.

11 µm

Scanning electron micrographs (backscattered electron image) showing a crack penetrated through the 625 cladding (of a coextruded tube) and then terminated at the cladding/steel interface at low magnification (original micrograph 35×) (a), and at high magnification showing the crack changed direction and followed the cladding-steel interface (b). Semiquantitative EDX analysis (wt%) of the corrosion products on the crack tip is summarized below. Courtesy of Welding Services Inc. 1: 52.0% Fe, 29.4% Ni, 9.5% Cr, 3.4% K, 1.4% Na, 0.8% S, and trace elements 2: 61.3% Fe, 25.8% Ni, 7.6% Cr, 2.1% K, and trace elements 3: 14.0% Fe, 50.3% Ni, 24.4% Cr, 4.0% K, 2.2% Na, and trace elements

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

pg 371

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 371

measurements across the cross section of the cladding at 0.25, 0.50, 0.75, 1.00, and 1.25 mm (0.010, 0.020, 0.030, 0.040, and 0.050 in.) from the cladding surface (Ref 43). Vickers hardness tester with a 500 g load was used for microhardness measurements. Vickers hardness values (HV) were then converted to Rockwell C scale (HRC). The schematic of the air port opening tube sample along with measured hardness at different locations is shown in Fig. 13.15. According to the plant personnel, the air port opening tubes were reportedly not stress relieved or annealed after cold forming and prior to installation (Ref 43). Circumferential cracks were found to develop at the crown location (i.e., area with the highest heat ﬂux) where the tube section was bent in the bottom portion of the air port opening. This area showed hardness of about 36 to 39 HRC near the cladding surface (about 0.25 mm, or 0.01 in., from the cladding surface), which were slightly lower than either the extrados (the exterior bend section) (40 HRC) or the intrados (the inner bend section) (42 HRC) of the bend. The straight section was found to be about 30 HRC (about 0.25 mm, or 0.01 in., from the cladding surface). The microhardness measurements of this air port opening sample showed that the area that suffered cracking was in a coldworked condition. From the above discussion, it appears that cracking of the primary air port openings tends to occur at the locations where the tube is bent. The cladding in these locations is under a coldworked condition. The material at these locations is characterized by high residual stresses. Some alloys with high residual stresses (e.g., in a highly cold-worked condition) can suffer brittle, intergranular cracking when heated to or in service at temperatures of 425 to 650 °C (800 to 1200 °F), depending on the alloy. The morphology of this type of intergranular cracking, often referred to as “stress-relaxation cracking,” is similar to the morphology of cracking observed in Type 304L and alloy 625 claddings of the coextruded primary air port tubes. This cracking mechanism, which has not been discussed by any of the authors, may play a role in the air port opening cracking. The process of stress-relaxation cracking can be described in a simple way: when a heavily cold-worked alloy with high residual stresses (i.e., the residual stress in the alloy in the cold-worked condition is essentially its roomtemperature yield strength) is heated to 540 °C (1000 °F), for example, the yield strength of the alloy at that temperature will be lower than the

residual stress of the alloy. This condition prompts the alloy to undergo local plastic deformation to relieve the stresses. However, when the grain matrix is strengthened by ﬁne precipitates, it cannot allow plastic deformation to occur to relieve those stresses, thus causing the alloy to develop cracking at grain boundaries, which are often the weakest locations. Austenitic stainless steels and nickel-base alloys that form homogeneous, coherent precipitates, such as γ 0 (Ni3Al) or γ 00 (Ni3Nb), are more susceptible to this type of brittle cracking. Detailed discussion on this subject is covered in Chapter 14 “StressAssisted Corrosion and Cracking.” 13.3.4 Waterwalls above “Cut Line” The waterwalls above the butt weld joints with composite tubes typically are bare carbon steel. Corrosion of these waterwall carbon steel tubes is primarily caused by sulﬁdation by reduced sulfur gases, primarily H2S. The corrosion rate is typically 0.2 mm/yr (8 mpy), but can be as high as

1 2

3

4

21 µm

Fig. 13.14

Scanning electron micrographs (backscattered electron image) showing the top portion of the crack (shown in Fig. 13.13a) that penetrated through the 625 cladding (of a coextruded tube) and then terminated at the cladding/steel interface and then changed direction and followed the cladding/steel interface. Semiquantitative EDX analysis (wt%) of the corrosion products on the top portion of the crack is summarized below. Courtesy of Welding Services Inc. 1: 55.6% Ni, 20.9% Cr, 4.7% Fe, 4.6% S, 2.6% Na, 7.1% Si, 1.9% Ca, 1.5% Al, and trace elements 2: 84.6% Ni, 5.2% Cr, 3.1% Fe, 2.8% Na, 1.6% Si, 1.0% S, and trace elements 3: 85.6% Ni, 5.7% Cr, 3.1% Fe, 2.2% Na, 1.0% Si, 0.7% S, and trace elements 4: 71.1% Ni, 18.3% Cr, 3.5% Fe, 2.3% Na, 1.2% Si, 1.6% S, and trace elements

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

372 / High-Temperature Corrosion and Materials Applications

0.8 mm/yr (32 mpy) (Ref 10). When sulﬁdation becomes excessive, one effective approach is to apply Type 309 overlay using automatic overlay welding process. Instead of forming FeS on unprotected carbon steel, protective Cr2O3 scales form on the 309 overlay containing typically 20% Cr. Numerous papers have been published on the beneﬁcial effects of chromium on sulﬁdation resistance of alloys (see Chapter 7 “Sulﬁdation”). Moberg et al. (Ref 44) illustrates the beneﬁcial effect of chromium on the sulﬁdation resistance of steels containing chromium in a sulﬁdizing environment at 400 °C (750 °F) (Fig. 13.16). Lai and Hulsizer (Ref 45) reported that Type 309 overlay was applied on the waterwalls of a recovery boiler in a Midwest mill (U.S.) in 1987, 1989, and 1992. Figure 13.17 shows the cross section of a Type 309 overlay sample obtained from the overlaid waterwall after 13 years of service, showing essentially the original overlay thickness with only tiny surface corrosion pits (Ref 45). A close-up view of the 309 overlay on the front wall after 8 years of boiler operation in the same boiler is shown in Fig. 13.18. 13.3.5 Superheater Tube Corrosion As the combustion ﬂue gas rises to the upper furnace, black liquor droplets entrained in the ﬂue gas stream as carryover particles can

33 HRC at 0.01 in. (31–33 HRC)

potentially contribute to the deposits on superheater platens. The deposits form primarily on the windward side of the tube. When the tube metal temperature reaches the melting temperature of the deposits, corrosion of the superheater can become serious. Corrosion is more serious on the windward side than the leeward side of the superheater tube. This is illustrated Fig. 13.19 (Ref 46). Sulﬁdation/oxidation is the major corrosion mechanism for superheaters (Ref 10). Common superheater alloys are T-11 or T-22. When T-11 or T-22 superheater tubes suffer high wastage rates, one cost-effective solution will be to switch to austenitic stainless steels that are capable of forming protective chromium oxide scales instead of formation of iron oxides and/or sulﬁdes on unprotected T-11 or T-22 tubes. An example was given by Lai and Wensley (Ref 31) for a boiler in South America, where T-11 superheater tubes suffered severe wastage at the bend sections and required annual replacement. A T-11 superheater tube bend sample was removed for metallurgical evaluation after 6 months of service. The sample is shown in Fig. 13.20. The wastage rate was estimated to be approximately 3.9 mm/yr (154 mpy). Superheater steam temperature and pressure were reportedly 435 °C (815 °F) and 6.4 MPa (63.5 bars, or 930 psig), respectively. Examination of the sample revealed that the corrosion products

39 HRC at 0.01 in. (36–39 HRC) 36 HRC at 0.01 in. (35–37 HRC)

37 HRC at 0.01 in. (37–39 HRC)

42 HRC at 0.01 in. (40–42 HRC)

Cracks 40 HRC at 0.01 in. (40–42 HRC) 40 HRC at 0.01 in. (38–40 HRC)

Fig. 13.15

30 HRC at 0.01 in. (28–30 HRC)

Air port opening made of alloy 625 coextruded tube that suffered cracking after 1 year of service in a boiler in Southeast (U.S.) with hardness data at different locations of the tube. Cracking occurred at the bottom half of the opening (right-hand side of the tube). Rockwell C hardness values (converted from Vickers microhardness values) were reported in a range at different locations of the tube as well as the hardness at 0.25 mm (0.01 in.) from the cladding surface. Courtesy of Welding Services Inc.

pg 372

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:48PM Plate # 0

pg 373

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 373

14

400 °C (750 °F)

Rate of weight gain, g/m2/h

12

10

8

6

4

2

0

0

5

10

15

20

25

30

Chromium, %

Fig. 13.16

Effect of chromium on sulfidation resistance of steels containing various amounts of chromium tested at 400 °C (750 °F) in N2-15H2O-10CO2-10H2-0.1O2-0.1H2S. Source: Ref 44

were essentially iron oxides but with the presence of chlorine. Figure 13.21 and 13.22 show the corrosion products and the energy-dispersive x-ray spectroscopy (EDX) analysis of the corrosion products. The results of the EDX analysis indicated that the corrosion products were essentially iron oxides with no sulfur. However, a relatively high chlorine (Cl) peak was detected in the corrosion scales at the scale/metal interface. It is, thus, believed the corrosion was caused by chlorine-accelerated oxidation. Type 310 overlay tubes were thus selected to replace numerous T-11 tube bends in this boiler in 2000, 2001, and 2003. So far, no corrosion or other degradation has been encountered with these overlay tubes. Figure 13.23 shows the 310 overlay superheater tubes after 2 years of operation in the boiler, revealing no evidence of corrosion attack.

13.4 Summary

0.5 mm

Fig. 13.17

Optical micrograph showing the cross section of Type 309 overlay at the crown location of the rear waterwall tube after 13 years of boiler operation in a boiler in a Midwest mill (U.S.). The metallographic mount was etched with nital to reveal the substrate steel including the fusion boundary and the heat-affected zone (HAZ). Source: Ref 45

A brief description of a black liquor recovery boiler along with its fuel and combustion conditions is presented. Materials problems with ﬂoor tubes, smelt spout openings, and air port openings in the lower furnace are discussed. The materials that have been tested and tried with good results are reported. The corrosion issue for the waterwalls above the “cut line” (i.e., the butt weld joints where carbon steel waterwall

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:49PM Plate # 0

374 / High-Temperature Corrosion and Materials Applications

tubes join with the composite tubes at the lower part of the furnace) is discussed. In addition, the corrosion of superheater tubes is also discussed. REFERENCES

1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, (1992), p 26–1 2. G.A. Smook, Handbook for Pulp & Paper Technologists, 2nd ed., Angus Wilde Publications, Vancouver, Canada, (1992), p 74 3. T.J. Grant, Update of the American Forest & Paper Association’s Recovery Boiler Program, TAPPI Engineering & Papermakers Conference (Conf. Proc.), Book 2, TAPPI, (1997), p 589

Fig. 13.18

A general view of the 309 overlay on the front wall after about 8 years of boiler operation in Midwest mill (U.S.). Source: Ref 45

4. J. Gommi, Root Causes of Recovery Boiler Leaks, TAPPI Engineering & Papermakers Conference (Conf. Proc.), Book 2, TAPPI, (1997), p 509 5. W.J. Frederick, Chapter 3, Black Liquor Properties, Kraft Recovery Boilers, T.N. Adams, Ed., TAPPI Press, (1997), p 61 6. T.M. Grace, Chapter 5, Chemical Recovery Process Chemistry, Chemical Recovery in the Alkaline Pulping Processes, 3rd ed., R.P. Green and G. Hough, Ed., TAPPI Press, (1992), p 57 7. W.J. Frederick and M. Hupa, Chapter 5, Black Liquor Droplet Burning Processes, Kraft Recovery Boilers, T.N. Adams, Ed., TAPPI Press, (1997), p 131 8. G.A. Smook, Handbook for Pulp & Paper Technologists, 2nd ed., Angus Wilde Publications, Vancouver, Canada, (1992), p 133 9. T. Grace and W.J. Frederick, Chapter 6, Char Bed Processes, Kraft Recovery Boilers, T.N. Adams, Ed., TAPPI Press, (1997), p 163 10. H. Tran, Chapter 10, Recovery Boiler Corrosion, Kraft Recovery Boilers, T.N. Adams, Ed., TAPPI Press, (1997), p 285 11. E.F. Hogan, Investigation of Chemical Recovery Unit Floor Tube Overheating Failures, TAPPI Engineering & Papermakers Conference (Conf. Proc.), Book 2, TAPPI, (1997), p 567 12. C.M. Wells, Chapter VIII, Chemical Recovery Area, Pulp and Paper Manufacturing, Vol 10, Mill-Wide Process Control & Information Systems, D.B. Brewster and M.J. Kocurek, Ed., Joint Textbook

5 T-22 probe

Metal loss, mm

4

Windward

3

2 Leeward 1

0 440

460

480

500

520

540

560

580

600

620

640

660

Temperature, °C

Fig. 13.19

Results of corrosion probe tests for T-22 in the lower superheater region for 840 h in a boiler. Source: Ref 46

pg 374

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:49PM Plate # 0

pg 375

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 375

A

C B

Fig. 13.20

T-11 superheater tube sample showing severe corrosion attack after 6 months of operation.

Source: Ref 31

13.

14. 15.

16.

17.

18.

19.

20.

Committee of the Paper Industry, Canada, (1993), p 124 P.M. Singh, S.J. Al-Hassan, S. Stalder, and G. Fonder, Corrosion in Kraft Recovery Boilers—In-Situ Characterization of Corrosive Environments, 1999 TAPPI Engineering/Process and Product Quality Conference (Conf. Proc.), Vol 3, TAPPI, (1999), p 1047 A. Borg, A. Teder, and B. Warnqvist, TAPPI, Vol 57 (No. 1), (1974), p 126 H. Tran, Chapter 9, Upper Furnace Deposition and Plugging, Kraft Recovery Boilers, T.N. Adams, Ed., TAPPI Press, (1997), p 247 H. Tran, M. Gonsko, and X. Mao, Effect of Composition on the First Melting Temperature of Fireside Deposits in Recovery Boilers, TAPPI J., Vol 82 (No. 9), (1999), p 93 H. Tran, Recovery Boiler Plugging and Prevention, TAPPI Kraft Recovery Operations Short Course Notes, TAPPI Press, 1992, p 209–218 J.L. Clement and J.D. Blue, Recovery Furnace Floor Design and Alternative Materials, presented at Tenth Latin American Recovery Congress (Concepcion, Chile), Aug 26–30, 1996 D. Singbeil, R. Prescott, J. Keiser, and R. Swindeman, Composite Tube Cracking in Kraft Recovery Boilers—A State-OfThe-Art Review, 1997 Engineering & Papermakers Conference (TAPPI Conf. Proc.), Book 3, TAPPI Press, (1997), p 1001 J.R. Keiser et al., Analysis of Cracking of Co-Extruded Recovery Boiler Floor

(a)

(b)

160 µm

24 µm

Fig. 13.21

Scanning electron micrographs showing (a) the corrosion products formed on the T-11 superheater tube sample at the tube bend (Fig. 13.20) and (b) the corrosion products on area C at high magnification. EDX analysis showed essentially iron with tiny chromium peak. Area C showed presence of chlorine (Cl) in addition to iron. The x-ray spectra from area C are shown in Fig. 13.22. Source: Ref 31

Tubes, 1997 Engineering & Papermakers Conference (TAPPI Conf. Proc.), Book 3, TAPPI Press, (1997), p 1025 21. J.L. Barna and K.B. Rivers, Improving Recovery Boiler Furnace Reliability with Advanced Materials and Application Methods, presented at Canadian Pulp and Paper Association Meeting, (Montreal, Quebec, Canada), Jan 25–29, 1999

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:49PM Plate # 0

376 / High-Temperature Corrosion and Materials Applications

8000

Fe

7000 6000

Counts

5000 4000 3000 2000 1000 C

0 Fe Cr

C1 Fe Si

Cr Cr Mn

S

0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

keV

Fig. 13.22

X-ray spectra from area C, as shown in Fig. 13.21, showing a relatively high chlorine (Cl) peak. Source: Ref 31

(a)

Fig. 13.23

(b) Type 310 overlay superheater tubes (a) and in close-up (b) after 2 years of operation in the boiler at a mill in South America. Source: Ref 31

22. J.R. Keiser, L.M. Hall, K.A. Choudhury, G.B. Sarma, J.P. Gorog, and R.E. Baker, Thermal Behavior of Floor Tubes in a Kraft Recovery Boiler, 1999 TAPPI Engineering/ Process and Product Quality Conference (TAPPI Conf. Proc.), TAPPI Press, (1999), p 1109

23. J.R. Keiser et al., Status Report on Studies of Recovery Boiler Composite Floor Tube Cracking, 1999 TAPPI Engineering/Process and Product Quality Conference (TAPPI Conf. Proc.), TAPPI Press, 1999, p 1099 24. R. Prescott and D.L. Singbeil, Stress Corrosion Cracking of Type 304L Stainless

pg 376

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/

31/10/2007 12:49PM Plate # 0

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 377

25.

26.

27. 28.

29. 30.

31.

32.

33.

34.

Steel in Kraft Recovery Boiler Environments, Ninth International Symposium on Corrosion in the Pulp and Paper Industry (Conf. Proc.), CPPA (Montreal, Quebec, Canada), (1998), p 185 H. Tran, B. Habibi, and C. Jia, Drying Behavior of Waterwash Solution and the Effect on Composite Floor Tube Cracking in Recovery Boilers, 1999 TAPPI Engineering/ Process and Product Quality Conference (TAPPI Conf. Proc.), TAPPI Press, (1999), p 1061 J.R. Keiser et al., Why Do Recovery Boiler Composite Floor Tubes Crack? 2000 TAPPI Engineering Conference (TAPPI Conf. Proc.), TAPPI Press, 2000 “Sandvik Sanicro 38/4L7,” S-12126-Eng, Sandvik Steel, Sweden, June 1996 X.L. Wang, E.A. Payzant, B. Taljat, C.R. Hubbard, J.R. Keiser, and M.J. Jirinec, Experimental Determination of the Residual Stresses in a Spiral Weld Overlay Tube, Mater. Sci. Eng. Vol A232, (1997), p 31 P.N. Hulsizer, Dual Pass Weld Overlay Method and Apparatus, U.S. Patent No. 6013890, Jan 2000 G. Lai, M. Jirinec, and P. Hulsizer, The Properties and Characteristics of Unifuse 625 Overlay Tubing for Recovery Boiler Applications, 1998 TAPPI Engineering Conference (Conf. Proc.), Book 2, TAPPI Press, 1998, p 417 G.Y. Lai and A. Wensley, Metallurgical Characteristics and Performance Experience of Spiral Overlay Tubes in Black Liquor Recovery Boilers, 2005 TAPPI Engineering, Pulping and Environmental Conference (Conf. Proc.), TAPPI Press, 2005 A. Wilson, M. Lundberg, and U. Forsberg, Alloy 825 Mod/SA210-A1 Composite Tube for Black Liquor Recovery Boiler Floors, 1997 Engineering & Papermakers Conference (TAPPI Conf. Proc.), Book 3, TAPPI Press, (1997), p 1043 D. Singbeil, Inspection for Cracking of Composite Tubes in Black Liquor Recovery Boilers, 2002 TAPPI Fall Conference (Conf. Proc.), TAPPI Press, 2002 H. Dykstra, N. Risebrough, and A. Wensley, Corrosion and Cracking of Lower Furnace Wall Tubes in Recovery Boilers, 1999 TAPPI Engineering/Process and Product

35.

36.

37.

38.

39.

40. 41. 42.

43. 44.

45.

46.

Quality Conference (TAPPI Conf. Proc.), TAPPI Press, (1999), p 1071 A. Wensley, Alternative Materials for Floor and Lower Waterwall Tubes in Black Liquor Recovery Boilers, 2000 TAPPI Engineering Conference (Conf. Proc.), TAPPI Press, 2000 J.R. Keiser et al., Recent Observations of Recovery Boiler Primary Air Port Cracking and Characterization of Environmental Conditions, 2001 TAPPI Engineering/Finishing & Converting Conference (Conf. Proc.), TAPPI Press, 2001 J.R. Keiser et al., Relationship of Recovery Boiler Parameters and Primary Air Port Cracking, 2002 TAPPI Fall Conference (Conf. Proc.), TAPPI Press, 2002 A. Wensley and B. Woit, Inspection of Primary Air Port Opening Tubes in Recovery Boilers, 2001 TAPPI Engineering/ Finishing & Converting Conference (Conf. Proc.), TAPPI Press, 2001 R. Shenassa, K. Haaga, and J. Tuiremo, Primary Air Port Tube Integrity—A Critical Review of Primary Air Port Design and the Effect of Boiler Design Parameters, 2002 TAPPI Fall Conference (Conf. Proc.), TAPPI Press, 2002 A. Wensley, presented at the International Symposium on Corrosion in the Pulp and Paper Industry, Charleston, SC, 2004 J.R. Keiser, private communication, 2006 J.R. Keiser et al., Causes and Solutions for Cracking of Co-extruded and Weld Overlay Floor Tubes in Black Liquor Recovery Boilers, Ninth International Symposium on Corrosion in the Pulp and Paper Industry (Conf. Proc.), TAPPI Press, 1998 Welding Services Inc., unpublished data, Norcross, GA O. Moberg, P.E. Ahlers, and L. Dahl, Recovery Boiler Corrosion, Pulp and Paper Industry Corrosion Problems, NACE, (1974), p 125 G. Lai and P. Hulsizer, Performance of Type 309 SS Overlay in the Lower Furnace of A Black Liquor Recovery Boiler, 2001 TAPPI Engineering Conference (Conf. Proc.), TAPPI Press, 2001 H.N. Tran, D.C. Pryke, and D. Barham, Local Reducing Atmosphere—A Cause of Superheater Corrosion in Kraft Recovery Units, TAPPI, Vol 68 (No. 6), 1985, p 102

pg 377

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p379-408 DOI: 10.1361/hcma2007p379

31/10/2007 4:06PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 14

Stress-Assisted Corrosion and Cracking 14.1 Introduction Most high-temperature components are under stress during service. The stress can be residual, resulting from welding or forming operations prior to service. When the temperature is not high enough, these stresses can remain in the component during service. The stress can also come from externally imposed thermal stresses or mechanical loading to which the component is subjected during service. The stresses imposed upon the component during service can affect the behavior of high-temperature corrosion attack or cause intergranular cracking. One important subject area is the effect of tensile stresses (or strains) on high-temperature corrosion attack. Tensile stresses that are imposed on the component during service can cause preferential corrosion penetration. As this preferential corrosion penetration grows into the metal interior, it can eventually develop into a crack. One important industrial example involving this phenomenon is the circumferential cracking that occurs on the waterwall tubes of some supercritical coal-ﬁred boilers, which are ﬁred under low NOx combustion conditions. The low NOx combustion produces sulﬁdizing environments at or near the waterwall of the boiler. The waterwall tubes, which are made of a carbon or low-alloy steel, or even with a stainless steel or nickel-base alloy weld overlay that protects the waterwall steel tubes, can suffer circumferential cracking in some boilers. This particular phenomenon (i.e., circumferential cracking of the waterwall) is covered in Chapter 10 “Coal-Fired Boilers.’’ In that chapter, the correlation between the circumferential cracking and the stress-assisted preferential corrosion penetration is discussed. The current chapter discusses the phenomenon of the preferential high-temperature corrosion penetration that forms under the combination of stresses and corrosive conditions, primarily in sulﬁdizing environments.

The other subject area discussed in this chapter is stress-induced cracking. This is a brittle fracture with cracks propagating along grain boundaries. This is often referred to as “stressrelaxation cracking,” “reheat cracking,” or “strain-age cracking” in the literature. Typically, a highly constrained component, such as a heavy wall construction, or a welded component, or a cold-worked structure, can be susceptible to this type of intergranular, brittle cracking for some alloys when exposed to lower end of the intermediate temperature range, such as 480 to 700 °C (900 to 1290 °F). Alloys that are susceptible to this type of intergranular fracture include ferritic steels, austenitic stainless steels, Fe-Ni-Cr alloys, and nickel-base alloys. Some γ′ [Ni3Al] strengthened nickel-base alloys can be susceptible to this type of embrittlement at temperatures higher than 700 °C (1290 °F). This type of cracking can also occur during reheating of the component in a postweld heat treatment or annealing/stress-relieving. Weldments or weld joints are particularly susceptible to this type of cracking. Some components can be subject to thermal cycling during service. The components that are subject to thermal cycling are generally more easily understood from the operating standpoint. When both stress magnitudes (or strain magnitudes) and frequencies are high enough, the metal can eventually fail by thermal fatigue cracking. The morphology of thermal fatigue cracking is generally very well deﬁned and easily identiﬁed. The subject of thermal fatigue cracking has been extensively discussed in the literature and is not covered in this chapter.

14.2 Stress-Assisted Preferential Corrosion Penetration In most industrial environments, alloys rely on the formation of a continuous oxide scale

pg 379

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

380 / High-Temperature Corrosion and Materials Applications

for protection against accelerated corrosion attack during service at elevated temperatures. Most commercial alloys rely on chromium oxide (Cr2O3) scales. The majority of the corrosion data have been generated from test specimens that were not subjected to external stresses during exposure. However, as discussed earlier, most high-temperature components are subject to external stresses during plant operation. When a metal with a protective oxide scale is under tensile stress, the oxide scale may crack as the tensile stresses or strains become high enough. Figure 14.1 shows cracking of the oxide scale on alloy 800 after testing at 800 °C (1470 °F) in air under the strain rate of 10−6 s−1 (Ref 1). Under certain conditions, these cracks can be rehealed. This is illustrated in Fig. 14.2, showing a healed oxide crack on alloy 800 at 800 °C (1470 °F) in air under the strain rate of 10−8 s−1 (Ref 2). Schütze (Ref 2) observed that cracking of the oxide scale on a deforming metal can cause signiﬁcant internal corrosion penetration, as shown in Fig. 14.3 for Fe-18Cr steel (with 0.8Al and 1.5Si) deforming at 800 °C (1470 °F) in air under a strain rate of 10−8 s−1. The ﬁgure shows that once the oxide scale reached the initial “threshold” strains for developing cracking, further deforming of the steel caused signiﬁcant internal oxidation penetration. On the other hand, the nondeformed specimen (i.e., the specimen that was not under creep deformation during the exposure) showed no changes in the depth of the internal oxidation

penetration during the exposure for times up to 1000 h. The internal oxidation penetration in the deforming Fe-18Cr-0.8Al alloy specimen was found to be essentially in the form of nitridation attack involving aluminum nitrides (Ref 1). (Nitridation attack in air or combustion environments is discussed in Chapter 4 “Nitridation.”) Oxidation in air or oxidation in general is considered to be the least corrosive among various high-temperature corrosion modes. As the corrosivity of the environment increases, the effect of the external tensile stresses on preferential corrosion penetration becomes increasingly more signiﬁcant. This section focuses on the stress-induced preferential corrosion penetration in sulﬁdizing environments. Smolik and Flinn (Ref 3) examined the effects of stresses on sulﬁdation of alloy 800H in coal gasiﬁcation environments, which were characterized with high sulfur and low oxygen potentials ( pS2 and pO2 ). Tests involved exposing the outer diameter of the alloy 800H tube to three test environments, which were Ar-23%H2O10%H2 with 0.1, 0.2, and 0.4% H2S, respectively. The corresponding partial pressures of oxygen ( pO2 ) and sulfur ( pS2 ) at the test temperature of 870 °C (1600 °F) were 2 ×10−18 and 1 × 10−8 atm for the 0.1 H2S test environment, 2× 10−18 and 5 × 10−8 atm for the 0.2 H2S environment, and 2 ×10−18 and 2 × 10−7 atm for the 0.4 H2S environment. The circumferential strain was created by pressurizing the test tube internally with argon. Preferential corrosion penetration was observed to take place along

Fig. 14.1

Fig. 14.2

Cracking of the oxide scale on alloy 800 during deformation at the strain rate of 10−6 s−1 at 800 °C (1470 °F) in air. Source: Ref 1

Rehealed oxide-scale crack on alloy 800H during deformation at the strain rate of 10−8 s−1 at 800 °C (1470 °F) in air. Source: Ref 2

pg 380

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

grain boundaries. Both oxides and sulﬁdes were found to form along grain boundaries. This intergranular oxidation/sulﬁdation penetration eventually affected the circumferential strains to failure for the test tubes. The higher H2S concentration produced deeper penetrations, thus resulting in smaller strains to failure. The circumferential strains to failure were found to be about 8 to 10% for the 0.1% H2S, about 6 to 7% for the 0.2% H2S, and 1 to 4% for the 0.4% H2S. Researchers at the Joint Research Centre, Petten Establishment, The Netherlands, have conducted a series of extensive studies on the effects of deformation on preferential corrosion in reducing, sulﬁdizing environments (i.e., simulated coal gasiﬁcation environments) (Ref 4–8). Some of their results are summarized in this section. In their study of the effect of the creep deformation on the corrosion of alloy 800H in

pg 381

Stress-Assisted Corrosion and Cracking / 381

coal gasiﬁcation environments, Guttmann and Timm (Ref 4) observed that creep deformation markedly accelerated preferential corrosion penetration. Creep tests were conducted at 800 °C (1470 °F) in H2-1.2H2O-7CO-0.4H2S (5 ×10−22 atm pO2 , 5×10−9 atm pS2 , and 0.3 ac) and H2-1.2H2O-0.4H2S (5×10−22 atm pO2 , 4 ×10−9 atm pS2 ). Exposure tests were also conducted in the same test environments with no applied stresses for comparison of the corrosion behavior. Under the stressed condition preferential corrosion penetration was found to be deepest at grain boundaries that were normal to the stress direction. Samples in the stress-free condition showed shallow corrosion depths following grain or twin boundaries without preferential penetration. Preferential intergranular corrosion penetration was found to consist of essentially sulﬁdation/oxidation. It was found that the sulﬁdation/oxidation penetration in the

ε = 10–8s–1 Without deformation

150

Scale cracking

x i, µm

ε = 0.7%

100

50

0

0

500

1000

Time, h

Fig. 14.3

Depth of internal corrosion penetration for Fe-18Cr-0.8Al-1.5Si between the undeformed specimens (open data points) and the deforming specimens (solid data points) under the strain rate of 10−8 s−1 at 800 °C (1470 °F) in air. Source: Ref 2

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

382 / High-Temperature Corrosion and Materials Applications

Depth of corrosion, µm

stressed specimens was about 2 to 3 times more than that of the unstressed specimens. This is illustrated in Fig. 14.4 (Ref 4). For the COcontaining environment, specimens in both stressed and unstressed conditions suffered carburization as well as sulﬁdation/oxidation. Under these conditions, carburization dominated the corrosion penetration depth; the authors found no signiﬁcant effect on the corrosion depth when the corrosion attack involved carburization (Fig. 14.4). In examination of the surface cracks of the specimens tested in air and those tested in the sulﬁdizing environments, Guttmann and Timm (Ref 4) found that the surface cracks, which became blunted in air, were deeply penetrating and long and sharp with severe corrosion at the crack tip and in the crack vicinity. Stroosnijder et al. (Ref 5) made similar observations when comparing the creep deformation behavior of alloy 800 at 700 °C (1290 °F) between air and H2-1.2H2O-7CO-0.2H2S (2 ×10−25 atm pO2 , 2 ×10−10 atm pS2 ). These authors observed blunted surface cracks when tested in air and deeply penetrating sharp cracks along grain boundaries with deep corrosion paths in front of cracks when tested in the sulﬁdizing environment. The corrosion products formed at the grain boundary in the sulﬁdizing environment were found to consist of oxides and sulﬁdes. The above discussion focuses on the effects of relatively large deformation (i.e., large strain or high strain rate) during creep testing. For many

103

102 No load Load Sulfidation Oxidation Carburization 10 10

102

103

104

Exposure time, h

Fig. 14.4

Corrosion depth for unstressed and stressed specimens of alloy 800H tested at 800 °C (1470 °F) in H2-1.2H2O-0.4H2S (5 × 10−22 atm pO2 , 4 × 10−9 atm pS2). Also included was the data generated from H2-1.2H2O-7CO-0.4H2S (5 × 10−22 atm pO2 , 5 × 10−9 atm pS2 , and 0.3 ac). The latter environment caused carburization in addition to sulfidation/ oxidation. Source: Ref 4

structural components during service at high temperatures, the strains imposed on the component are likely to be much smaller. The design strain rate (or creep rate) of ASME Boiler and Pressure Vessel Codes is 10−5% h−1 (2.8 × 10−11 s−1) (Ref 9). It is therefore important to examine the effects of low tensile stress (or strain) on the preferential corrosion penetration. Stroosnijder et al. (Ref 6) investigated the effect of strain on preferential corrosion of alloy 800H when tested at 800 °C (1470 °F) in H2-1.2H2O-7CO-0.4H2S (5 × 10−22 atm pO2 , 5 × 10−9 atm pS2 , and 0.3 carbon activity, ac) with (a) no external stress (or strain) during testing and (b) strains up to 4% and strain rates down to 10−9 s−1. In this study, alloy 800H specimens were surface treated with CeO2 using cerium sol-gel techniques. The authors observed that the unstressed specimens showed some external and internal corrosion with no evidence of preferential corrosion penetration after 663 h of exposure (Fig. 14.5a). For stressed specimens after 663 h and subjected to about 2.3% strain under a strain rate of 10−8 s−1, preferential corrosion penetration along the grain boundary began to take place (Fig. 14.5b). After testing for 1247 h with about 5% strain (10−8 s−1 strain rate), intergranular corrosion attack along grain boundaries was found to penetrate farther into the metal interior (Fig. 14.5c). Guttmann et al. (Ref 7) examined the effect of a much lower strain range (typically 1 to 2%) on the preferential corrosion penetration in a sulﬁdizing environment. The strains the authors examined were much closer to the strains that might be experienced by operating hightemperature components. Tests were conducted at 600 °C (1110 °F) in CO-32H2-4CO2-0.2H2S. The strain rates during creep were 10−10 s−1 to 10−8 s−1. Similar to what was observed earlier in alloy 800H (Fe-33Ni-2Cr-0.4Al-0.4Ti) (Ref 4–6), Guttmann et al. (Ref 7) observed that alloy HR3C (Fe-20Ni-25Cr-0.5Nb-0.2N) showed stress-assisted intergranular corrosion attack under the test conditions. This is illustrated in Fig. 14.6 (Ref 7). The corrosion products formed at grain boundaries were enriched in chromium, oxygen, and sulfur, but depleted in nickel and iron. The intergranular corrosion attack in alloy HR3C tested after 1810 h with 2.2% strain is shown in Fig. 14.7 for the scanning electron microscopy (SEM) backscattered electron image of the intergranular corrosion attack and the x-ray maps for chromium, oxygen, and sulfur. The intergranular corrosion attack was

pg 382

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

(a)

(b)

pg 383

Stress-Assisted Corrosion and Cracking / 383

(c)

Corrosion morphology of alloy 800H tested at 800 °C (1470 °F) in H2-1.2H2O-7CO-0.4H2S (5 × 10−22 atm pO2 , 5 × 10−9 atm pS2 , and 0.3 ac) (a) after 663 h under no external stresses, (b) after 663 h at about 2.3% strain under the strain rate of 10−8 s−1, and (c) after 1247 h under strain rate of 10−8 s−1. Note: the specimen surface was plated prior to the mounting of the sample to retain the corrosion products. Source: Ref 6

Fig. 14.5

(a)

25 µm

(b)

25 µm

Fig. 14.6

Strain-assisted intergranular corrosion attack in alloy HR3C after testing at 600 °C (1110 °F) for 250 h in CO-32H2-4CO20.2H2S with (a) 1.3% strain and (b) 2% strain. Corrosion products formed on the metal surface were also observed. Note: the tested specimen surface was plated prior to the mounting of the metallographic sample to retain the surface corrosion products. Source: Ref 7

oxidation/sulﬁdation, similar to the earlier observations in alloy 800H. For nonstressed specimens, the alloy suffered external sulﬁdation with Fe- or Ni-rich sulﬁdes ([Ni,Fe]9S8) and Ni3S2. Underneath the surface sulﬁde scales were ﬁne internal corrosion phases randomly distributed within the grain matrix with no deep penetrating intergranular corrosion attack, as shown in Fig. 14.8. Both alloys 800H and HR3C are Fe-Ni-Cr alloys with austenitic microstructure. Both alloys were found to suffer preferential corrosion penetration along grain boundaries. It is interesting to note that nickel-base alloys with the same austenitic microstructure would follow a

completely different morphology in preferential corrosion penetration under stress. In fact, the morphology of stress-assisted preferential corrosion attack for nickel-base alloys tends to be similar to that for iron-base alloys having a ferritic structure. Guttmann et al. (Ref 7) tested nickel-base alloy 45TM (Ni-27Cr-23Fe-2.76Si) under the same test condition as for alloy HR3C. Alloy 45TM exhibited a ﬁngerlike preferential corrosion penetration, as shown in Fig. 14.9 (Ref 7). The corrosion products formed in the preferential corrosion penetration were found to be enriched in chromium, oxygen, sulfur, and silicon, as shown in Fig. 14.10. By contrast, the unstressed specimens were found to exhibit a

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

384 / High-Temperature Corrosion and Materials Applications

Fig. 14.7

Scanning electron backscattered images showing intergranular corrosion penetration along with the x-ray maps for Cr, O, and S for alloy HR3C after testing at 600 °C (1110 °F) for 1810 h in CO-32H2-4CO2-0.2H2S with 2.2% strain. Source: Ref 7

Fig. 14.8

Corrosion products formed on the metal surface as well as internal corrosion products formed in alloy HR3C with no external strain imposed on the test specimen when tested at 600 °C (1110 °F) for 2100 h in CO-32H2-4CO2-0.2H2S. Note: the tested specimen surface was plated prior to the mounting of the metallographic sample to retain the surface corrosion products. Source: Ref 7

fairly uniform surface corrosion attack with no preferential corrosion penetration, as shown in Fig. 14.11. In the same study, the authors tested an iron-base ferritic oxide-dispersion-strengthened

Fig. 14.9

Stress-assisted preferential corrosion penetration for alloy 45TM tested to 2% strain at 600 °C (1110 °F) for 1820 h in CO-32H2-4CO2-0.2H2S. Note: the tested specimen surface was plated prior to the mounting of the metallographic sample to retain the surface corrosion products. Source: Ref 7

(ODS) alloy, MA956 (Fe-20Cr-4.6Al-0.5Y2O3), under the same condition and found that this iron-base alloy with ferritic structure exhibited

pg 384

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

a ﬁngerlike preferential corrosion penetration under stresses, as shown in Fig. 14.12. The result of x-ray elemental mapping (Fig. 14.12) reveals that (a) the external surface layer was ironrich sulﬁdes, (b) the inner surface layer was chromium- and aluminum-rich oxides along with some sulﬁdes, and (c) the inner surface layer (Cr-Al rich oxides) also extended to the preferential corrosion penetration as an outer layer with a center core of iron-rich sulﬁdes that were extended from the external surface layer discussed in (a). When tested under a similar

Fig. 14.10

Stress-Assisted Corrosion and Cracking / 385

condition but with no external stresses, alloy MA956 showed only uniform corrosion attack, as shown in Fig. 14.13. For the unstressed specimen, the external corrosion layer was found to be iron-rich sulﬁdes and inner layer consisted of (Fe,Cr)1−xS, Cr2O3, and Al2O3. Ferritic stainless steels appeared to follow the similar ﬁngerlike preferential corrosion penetration when the specimen was under creep deformation, as shown in Fig. 14.14 (Ref 8). This ﬁgure also shows that the specimen suffered essentially uniform attack when under no external stress.

Scanning electron micrograph for the stress-assisted preferential corrosion penetration in alloy 45TM tested to 4% strain at 600 °C (1110 °F) for 2000 h in CO-32H2-4CO2-0.2H2S and elemental x-ray maps for the corrosion products in chromium, oxygen, sulfur, silicon, and iron. Source: Ref 7

pg 385

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

386 / High-Temperature Corrosion and Materials Applications

The authors did not report the percent of strain when the test was terminated and the specimen was examined; however, the test duration was only 615 h, and the strain was likely to be high. Upon closer examination of Fig. 14.14, it appears that a surface crack formed in the center of each corrosion penetration, further indicating that the strain might be large. Preferential corrosion penetration under strain (or stress) was also observed in other aggressive environments. The stress-assisted preferential corrosion attack is quite sensitive to the environment. Le Calvar et al. (Ref 10) performed low strain rate tests in air + 4CO2 + 8H2O at 610 °C (1130 °F). Figure 14.15 shows Type 304H suffering preferential oxidation penetration under this test condition at a strain rate of 3×10−8/s with total strain of about 2%. The authors referred to this preferential oxidation attack as “cracking” in their paper (Ref 10). Similar tests were also performed in a vacuum environment, and the results showed no “cracking.” The ﬁgure suggests that the phenomenon Le Calvar et al. observed was more like stress-assisted preferential corrosion penetration than stress-assisted cracking followed by formation of the corrosion products inside the crack. Rorbo (Ref 11) reported that the external stresses in Type 304 caused cracks to develop in the nitrided layer, and the cracks resulted in increased nitridation attack in front of the crack. This is illustrated in Fig. 14.16. The nitrided layer is believed to be very brittle, and very little strain is required to develop cracking in the nitrided layer. In this case, cracking could have developed in the nitrided layer and caused preferential nitridation penetration in front of the crack.

Fig. 14.11

Corrosion morphology for alloy 45TM tested under no external strains at 600 °C (1110 °F) for 2000 h in CO-32H2-4CO2-0.2H2S. Source: Ref 7

The waterwalls in some coal-ﬁred boilers, particularly some supercritical units, have been found to encounter preferential corrosion penetration. These supercritical boilers are typically equipped with low NOx burners and overﬁre air, thus creating localized reducing, sulﬁdizing environments at or near the waterwall in the lower furnace. Furthermore, the waterwalls can also be subjected to high heat ﬂux and possibly ﬂame impingement, thus resulting in higher metal temperatures and higher thermal stresses. The combination of these conditions leads to the development of stress-assisted preferential sulﬁdation penetration. As the preferential sulﬁdation penetration continues to grow into the metal interior, it eventually develops into a crack. At later stages, these cracks resemble thermal fatigue cracks. These cracks, which are in the transverse direction with respect to the tube axis, are commonly referred to as “circumferential cracks.” Typical alloys of construction have experienced circumferential cracking. The commonly used weld overlay cladding alloys including Type 309 and alloy 625 have also been observed to suffer circumferential cracking. This phenomenon is covered in detail in Chapter 10 “Coal-Fired Boilers” (Section 10.5.3). The morphologies of the preferential sulﬁdation penetration are strikingly similar between what has been observed in the waterwall of coalﬁred boilers and that observed in test specimens under tensile stresses (or low strain rate creep tests) in laboratory sulﬁdizing environments. Figure 10.32 in Chapter 10 shows an example of preferential sulﬁdation penetrations observed on T-22 (2.25Cr-1Mo) in a supercritical coalﬁred boiler equipped with low NOx burners and overﬁre air. These preferential sulﬁdation penetrations are considered to be precursors to circumferential cracks observed from tubes that failed or were about to fail. The preferential sulﬁdation penetration often exhibits “channels” in its core, as shown in Fig. 14.17. These channels, which generally exhibited lighter color (Fig. 14.17), were found to consist of sulﬁdes (see Fig. 10.33 and 10.34 in Chapter 10). In some cases, these channels were much more pronounced and numerous, such as the one shown in Fig. 14.18. The preferential sulﬁdation penetration formed in steels was found to consist of essentially iron oxides in the outer region of the penetration with the inner region (core) including channels being essentially iron sulﬁdes

pg 386

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

(see the SEM/EDX analysis in Fig. 10.33 and 10.34 in Chapter 10). This is illustrated schematically in Fig. 14.19. Proposed reaction steps involved in developing preferential sulﬁdation penetration in carbon or low-alloy steel waterwall tubes are: 1. Iron oxide scales initially form on the steel surface.

Fig. 14.12

Stress-Assisted Corrosion and Cracking / 387

2. Tensile stresses cause the breakdown of the iron oxide scales. 3. The surface oxide scale breakdown initiates the development of a preferential oxidation penetration. 4. As the preferential oxidation penetration continues to grow inward into the metal interior, the inner region of the oxide penetration (i.e., penetration core) becomes more

Scanning electron micrograph in a backscattered electron image along with x-ray maps for Cr, O, S, Al and Fe showing the stress-assisted preferential corrosion penetration (fingerlike penetration attack) on alloy MA956 tested to 5% strain at 600 °C for 1830 h in CO-32H2-4CO2-0.2H2S. Source: Ref 7

pg 387

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 388

388 / High-Temperature Corrosion and Materials Applications

deﬁcient in oxygen, thus lowering oxygen potentials in the core region. 5. Iron sulﬁdes begin to form in the core region due to insufﬁcient oxygen. 6. As the penetration continues to grow farther into the metal interior, oxidation/sulﬁdation processes continue to penetrate inward into the metal interior with oxidation forming in the outer region and sulﬁdation in the inner region of the corrosion penetration. 7. Channels are likely created during the growth of sulﬁde phases in the core region under the tensile stresses. These channels may also serve as gas passages as the oxidation/

sulﬁdation penetration grows deeper into the metal interior. For a Ni-Cr-Fe alloy, the preferential sulﬁdation penetration under large strains could develop a large core of iron sulﬁdes in the inner region with chromium oxides formed in the region next to the unaffected metal, such as in the case for alloy 45TM tested to 4% strain at 600 °C (1110 °F) for 2000 h in CO-32H2-4CO2-0.2H2S, as shown in Fig. 14.10 (Ref 7). Because of a large tensile strain (4%), the entire center core became a one big channel. However, when the strain was reduced to about 2%, alloy 45TM showed one big channel through the external surface scale followed by a number of ﬁne channels in the preferential sulﬁdation penetration, as shown in Fig. 14.9. For MA956

Fig. 14.13

Alloy MA956 tested under no external strain (or stress) at 600 °C (1110 °F) for 2000 h in CO-32H2-4CO2-0.2H2S, showing uniform corrosion with no preferential corrosion penetration. Source: Ref 7

10 µm

Fig. 14.14

Fe-12Cr-3Al-3Ti showing preferential corrosion penetration for the specimen under creep deformation at 600 °C (1110 °F) in H2-34.3H2O-18.5CO2-3.8CH47.9CO-1.3H2S for 615 h (upper figure) and uniform corrosion attack for the specimen under no external stress (or strain) after exposure to the same test environment and duration (lower figure). Source: Ref 8

Fig. 14.15

Type 304H tested in air + 4CO2 + 8H2O at 610 °C (1130 °F) under the strain rate of 3 × 10−8/s with about 2% strain, showing preferential oxidation penetration (the authors referred to as “cracking”). SEM/EDX analysis showed the oxide in area A was Fe-3Cr-4Si-0.5Mo, in area B was Fe-29Cr-5Ni-0.8Mo-0.8Si, and in area C was Fe-37Cr-21Ni1.5Si. Source: Ref 10

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

pg 389

Stress-Assisted Corrosion and Cracking / 389

(a Fe-Cr-Al alloy), the entire core became a one big channel consisting of iron sulﬁdes with the outer region consisting of a mixture of chromium oxides and aluminum oxides at 5% strain, as shown in Fig. 14.12 (Ref 7). For operating components, such as waterwalls in boilers, the strains imposed onto the component during operation are likely to be much lower. As shown in Fig. 14.20, alloy 625 weld overlay on the waterwall of a supercritical unit suffered preferential sulﬁdation penetration, which exhibited ﬁne channels. The ﬁgure also shows no evidence of the formation of a crack near or at the tip Fig. 14.16

Cracks developed in the nitrided layer formed in Type 304 due to external stresses and accelerated nitridation attack in front of the crack. Source: Ref 11

Iron sulfides (higher sulfur) Iron oxides

Iron sulfides Iron oxides

Carbon or low-alloy steel

Fig. 14.19 Fig. 14.17 Preferential sulfidation penetration (a precursor of the circumferential cracking) observed on a T-22 (2.25Cr-1Mo) wingwall tube of a supercritical boiler, showing “channels” (light color stringers) in the core of the penetration. Courtesy of Welding Services Inc.

Optical micrograph showing numerous “channels” in a circumferential groove formed on a T-11 (1.25Cr-0.5Mo) tube in a supercritical coal-fired boiler. Courtesy of Welding Services Inc.

Fig. 14.18

Fig. 14.20

Preferential sulfidation penetration formed in waterwall steel tubes

Scanning electron (backscattered electron) image showing a circumferential groove formed in alloy 625 weld overlay on a waterwall tube of a supercritical coal-fired boiler. Courtesy of Welding Services Inc.

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

390 / High-Temperature Corrosion and Materials Applications

of the preferential sulﬁdation penetration. The corrosion products were essentially chromium sulﬁdes with nickel-rich sulﬁdes forming the channels. The results of the SEM/EDX analysis of the corrosion products are shown in Fig. 10.37 in Chapter 10. Figure 14.21 shows another example of a circumferential groove that formed in the alloy 625 weld overlay on the waterwall tube after 6 years of service in another supercritical coal-ﬁred boiler. The circumferential groove was found to contain essentially nickeland chromium-rich sulﬁdes with numerous light grayish channels. Some light grayish channels were nickel-rich sulﬁdes. Sulﬁdes in the groove were also found to contain lead (Pb), arsenic (As), and zinc (Zn), which are common impurities in coal ash. Fine channels were also observed in the preferential sulﬁdation penetrations formed in the 309SS weld overlay on the waterwall, as shown in Fig. 10.40 and 10.41 in Chapter 10. In the waterwall, the 309SS overlay (Fe-Cr-Ni alloy) suffered preferential sulﬁdation

80 µm

Fig. 14.21

Scanning electron (backscattered electron) image showing a circumferential groove formed in alloy 625 weld overlay on a waterwall tube after 6 years of service in a supercritical coal-fired boiler under low NOx combustion. The compositions (wt%) of the phases at different locations were analyzed by EDX, showing either nickel-rich or chromium-rich sulfides. Some sulfides were also found to contain Pb, As, and Zn that are commonly found in coal ash. Note numerous light grayish channels. Some of these channels (No. 3 and 4) were nickel-rich sulfides. Courtesy of Welding Services Inc. 1: 56% Ni, 13.9% Cr, 6.2% Fe, 18.7% S, 2.5% Zn, 1.4% As 2: 53.8% Cr, 10.2% Ni, 7.0% Fe, 14.6% S, 9.4% Pb, 2.6% Zn 3: 58.9% Ni, 5.8% Cr, 4.6% Fe, 26.1% S, 1.9% As, 1.0% Pb 4: 56.4% Ni, 5.1% Fe, 3.3% Cr, 6.6% Mo, 24.4% S, 1.9% Pb 5: 52.5% Cr, 18.4% Mo, 7.5% Fe, 3.8% Ni, 9.5% Mo, 5.5% S 6: 18.7% Ni, 17.8% Cr, 13.7% Nb, 13.7% Mo, 21.3% Pb, 22.1% S, 4.6% Fe

penetration attack along dendritic boundaries, as shown in Fig. 10.40 in Chapter 10. This is quite similar to the wrought alloys, such as HR3C (Fe-Ni-Cr alloy), that suffered preferential sulﬁdation penetrations along grain boundaries when tested in laboratory environments (Fig. 14.6).

14.3 Stress-Assisted Intergranular Cracking When external stress is applied to a metallic component, the metal ﬁrst undergoes elastic deformation. With increasing applied stress, the metal then deforms plastically and eventually reaches the limit of its plastic deformation. At this point, with further increase of the applied stress, the metal can no longer deform to relieve the stresses built up in the metal, thus resulting in the initiation of cracks at the weakest locations to relieve those stresses. An engineering alloy can normally undergo a large plastic deformation before developing cracks. When the metallic component is subject to external stresses at elevated temperatures, the metal undergoes creep deformation, which is a time-dependent deformation governed by diffusion. An engineering alloy typically exhibits a large creep deformation prior to developing cracks and ﬁnal rupture. However, as the temperature decreases, the creep strain at rupture decreases. Under certain conditions, a metal can only sustain very little deformation (plastic and/or creep deformation) before developing brittle fracture. The fracture typically takes the form of intergranular cracking (i.e., cracking along grain boundaries). This brittle, intergranular fracture has occurred during heat treatments or service for some engineering alloys, including ferritic steels, austenitic stainless steels, Fe-Ni-Cr alloys, and nickel-base alloys. This type of brittle, intergranular cracking phenomenon has been described by different names, such as “reheat cracking,” “stress-relaxation cracking,” “strain-age cracking,” and “gamma-prime embrittlement.” This phenomenon typically occurs at the lower end of the intermediate temperature range, such as 480 to 700 °C (900 to 1290 °F). 14.3.1 General Conditions That Cause Stress-Assisted Cracking Embrittlement Several material conditions can increase the susceptibility of the stress-assisted cracking

pg 390

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

embrittlement for an alloy. An alloy in a coldworked condition, which exhibits higher yield strength and higher hardness, and lower ductility, is much more susceptible to this type of brittle intergranular cracking. Figure 14.22 shows an example of the effect of cold work on hardness and yield strength of several nickel-base alloys and the corresponding effect on its tensile elongation (Ref 12). The hardness and yield strength of an alloy can be signiﬁcantly increased by increasing the amount of cold work. The yield strength of a cold-worked metal also represents the residual stress in the cold-worked structure. An alloy with a larger amount of cold work exhibits a higher residual stress. Higher residual stresses in a structural component can increase the susceptibility to stress-relaxation cracking. After the alloy is strengthened by cold working, the capability of the grain matrix to plastically deform to relieve additional external stresses imposed on the metal during service at intermediate temperatures is signiﬁcantly reduced. When these external stresses cannot be relieved through deformation, cracking then develops at grain boundaries to relieve those stresses. The matrix of an alloy can also be strengthened by precipitates, particularly ﬁne, coherent precipitates (e.g., γ′ [Ni3(Al,Ti)], γ″ [Ni3Nb], Ni2Mo ordered phases). Figure 14.23 shows ﬁne, coherent γ″ [Ni3Nb] precipitates formed in the alloy matrix (i.e., grain interior) of alloy 625 (Ni-22Cr-9Mo-3.5Nb) at 650 °C (1200 °F) for 24 h (Ref 13). Examples of ﬁne precipitates formed in other alloy systems are shown in Fig. 14.24 for Ni2(Cr,Mo) ordered phases formed in Ni-16Cr-15Mo-3Fe alloy, Fig. 14.25 for γ′ (Ni3Al) precipitates formed in alloy 214, and Fig. 14.26 for γ′ (Ni3Al) precipitates in alloy 601. Precipitation of these ﬁne, coherent precipitates not only strengthens the grain matrix but also causes volume contraction that causes additional internal stresses in the alloy, making it more susceptible to stress-relaxation cracking. Carbides, such as TiC, NbC, and Cr23C6, also produce some alloy strengthening and can increase the susceptibility to stress-relaxation cracking. The combination of a cold-worked structure and ﬁne precipitates in the matrix can cause the material to be extremely susceptible to stress-relaxation cracking. Formation of precipitates, such as Cr23C6, at grain boundaries can further promote grain-boundary cracking, thus increasing the susceptibility of stress-relaxation cracking.

Stress-Assisted Corrosion and Cracking / 391

Another factor that is important in increasing the susceptibility of stress-assisted cracking embrittlement is the surface geometry of the component, which can lead to stress risers such as an abrupt thickness change in a component. A component that is under a highly constrained condition, such as a heavy section part, is most susceptible to this type of cracking embrittlement since the additional stresses imposed on the component render it incapable of accomodating any dimensional change or deformation, and it develops cracks to relieve those stresses. A majority of cracking embrittlement cases occur under two different conditions. The cracking can occur during heat treatment, such as a postweld heat treatment (PWHT) for a weldment, which involves a short duration. The cracking can also occur during service, which involves a relatively longer duration, but still relatively short in terms of the design life of the component, with most failures occurring after 1 to 2 years of service or less. In both cases, the metal (or the metallic component) has undergone very little deformation prior to intergranular cracking fracture. An example of heat-treatment-induced cracking embrittlement is described as follows. When welding is performed on a metal, residual stresses are developed in the heat-affected zone (HAZ) of the base metal due to the volume shrinkage of adjacent weld metal changing from the molten state to the solid state. The maximum residual stress is approximately the roomtemperature yield strength of the metal. When this weldment receives a PWHT, the yield strength of the metal at this particular PWHT temperature would be lower than the residual stresses (i.e., the room-temperature yield strength of the metal), thus causing the metal to undergo deformation to accommodate this additional stress. However, when the metal cannot deform to “relax” or “relieve” this additional stress through deformation, cracks develop at the weakest locations, typically grain boundaries, to relieve this additional stress, thus resulting in this intergranular cracking embrittlement. The service-related cracking embrittlement typically occurs when the service temperatures are relatively low such that creep deformation is too low to relieve the applied stresses, and the alloy matrix is signiﬁcantly strengthened such that these stresses are not able to be “relieved” through deformation, thus resulting in cracking for the relief of this applied stress. Detailed discussion on stress-relaxation cracking is presented

pg 391

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

392 / High-Temperature Corrosion and Materials Applications

Hardness, HRC

50

0

10

20

30

40

50

60 50

40

40

30

30

20

20 Alloy 25 Alloy 188 Alloy 625 230 Alloy Alloy X

10

0

0

10

20

30

10

40

0 60

50

Cold work, %

(a) 0

10

20

30

40

50

60 200

1350

1100

150 125

850

350

0

10

20

30

40

75 50 60

50

Cold work, %

(b)

70

Elongation, %

100

Alloy 25 Alloy 188 Alloy 625 2230 Alloy Alloy X

600

0

10

20

30

40

50

60

50

50 Alloy 25 Alloy 188 Alloy 625 230 Alloy Alloy X

40 30

40 30

20

20

10

10

0

Fig. 14.22

60 70

60

0 (c)

0.2 % yield strength, ksi

0.2 % yield strength, MPa

175

10

20

30

40

50

0 60

Cold work, %

Effects of cold work on hardness (a) room-temperature yield strength (b) and the corresponding tensile elongation (c) for several nickel-base alloys. Source: Ref 12

pg 392

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

in the following sections based on the alloy systems. 14.3.2 Cracking in Ferritic Steels Cracking of CrMo and CrMoV steel weldments in steam pipework and valve assemblies during stress-relief heat treatment (e.g., postweld heat treatment) and high-temperature service have been encountered in power-generating equipment (Ref 17). Similar reheat cracking has also been reported in the heat-affected zone (HAZ) of low-alloy steels for pressure vessels during hydrostatic pressure tests (Ref 17). The phenomenon of reheat cracking also received much attention when small cracks in the HAZ under the austenitic alloy cladding, which were

Stress-Assisted Corrosion and Cracking / 393

referred to as “underclad cracking,” were encountered in nuclear pressure vessels in the 1970s (Ref 17). Other publications discussing reheat and underclad cracking of ferritic steels in the 1970s can be found in Ref 18 to 21. Dhooge and Vinckier (Ref 17) provided a detailed review of the factors that might be responsible for causing reheat cracking of ferritic steels. These factors included intragranular precipitation hardening that strengthens the grain matrix and segregation of impurities to grain boundaries that weakens grain boundaries. Segregation of impurities to grain boundaries may play a signiﬁcant role in reheat cracking

Fig. 14.25 Fig. 14.23

Transmission electron micrograph showing fine, coherent c 00 (Ni3Nb) precipitates formed in the grain matrix of alloy 625 at 650 °C (1200 °F) for 24 h. Source: Ref 13

Fig. 14.24

Transmission electron micrograph showing a dark-field image of fine, Ni2(Cr,Mo) ordered phases formed in the grain matrix of Ni-16Cr-15Mo-3Fe alloy at 540 °C (1005 °F) for 16,000 h. Source: Ref 14

Transmission electron micrograph showing a dark-field image of fine, coherent c 0 (Ni3Al) precipitates formed in the grain matrix of alloy 214 (Ni-16Cr4.5Al-3Fe-Y) at 800 °C (1470 °F) for 8 h. Source: Ref 15

Fig. 14.26

Transmission electron micrograph showing a dark-field image of fine, coherent c 0 (Ni3Al) precipitates formed in the grain matrix of alloy 601 at about 590 °C (1100 °F) for 2.5 years. Original magnification: 97,000×. Source: Ref 16

pg 393

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 394

394 / High-Temperature Corrosion and Materials Applications

50 2.25Cr-1Mo, 2 kJ/mm 2.25Cr-1Mo, 3 kJ/mm 40

2.25Cr-1Mo, 4 kJ/mm

Reduction in area, %

HCM2S, 2 kJ/mm HCM2S, 3 kJ/mm 30

HCM2S, 4 kJ/mm

20

10

0 550

600

650

700

750

PWHT, °C

Fig. 14.27

Reduction in area for 2.25Cr-1Mo and HCM2S at various postweld heat treatment (PWHT) temperatures with an initial applied tensile stress of 325 MPa (47 ksi). Source: Ref 22

in ferritic steels. However, reheat cracking (or stress-relief cracking or strain-age cracking) also occurs in austenitic stainless steels, alloy 800H, and gamma prime (γ′) strengthened alloys; however, the impurity segregation is not likely to be as critical a factor in these austenitic alloys. Reheat cracking (or stress-relief cracking) of ferritic steels can be described using the test results generated by Nawrocki et al. (Ref 22, 23) when comparing 2.25Cr-1Mo steel with a modiﬁed 2.25Cr steel (HCM2S). HCM2S contains similar chromium amounts, but lower carbon contents (to reduce hardness of HAZ) along with the substitution of molybdenum (Mo) with tungsten (W), vanadium (V), and niobium (Nb) to increase creep strength, compared with 2.25Cr-1Mo steel (Ref 22). Comparing these two alloy steels in terms of their respective resistance to stress-relief cracking will allow a better understanding of the important metallurgical factors that contribute to reheat cracking. Nawrocki et al. (Ref 22, 23) used Gleeble simulation techniques to evaluate stress-relief cracking. Specimens were subjected to singlepass weld thermal simulation cycles with a peak temperature of 1315 °C (2400 °F) followed by cooling to room temperature. A tensile stress was imposed on the specimen during cooling and held for the duration of the test to simulate the residual stresses present in an actual weldment. After cooling to room temperature, the

20 µm

Fig. 14.28

Scanning electron micrograph of fracture surface of HCM2S ruptured air at 675 °C (1250 °F) under an initial stress of 325 MPa (47 ksi). Source: Ref 22

specimen was then subjected to a simulated, programmed postweld heat treatment temperature and held at constant temperature and load until failure (Ref 22). The rupture ductility of HCM2S was found to be signiﬁcantly lower than that of 2.25Cr-1Mo, as shown in Fig. 14.27 (Ref 22). HCM2S was found to be very brittle, as shown in Fig. 14.28 which reveals essentially complete intergranular fracture (Ref 22). The authors (Ref 22) concluded that HCM2S was more susceptible to stress-relief cracking than 2.25Cr-1Mo steel. Nawrocki et al. (Ref 23) attributed the brittle, intergranular fracture of

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

HCM2S to the formation of ﬁne (5 to 40 µm) carbides (W/Fe-rich carbides) in the grain matrix that were resistant to plastic deformation and the formation of grain-boundary Fe-rich M3C carbides that nucleated microcavities. With more carbide-forming alloying elements, such as tungsten, vanadium, and niobium, HCM2S is expected to form more matrix-strengthening carbides, thus resulting in brittle, intergranular fracture due to signiﬁcantly reduced deformation capability of the grain matrix. 14.3.3 Cracking in Austenitic Stainless Steels and Fe-Ni-Cr Alloys Among austenitic stainless steels, Type 321 and 347 have been frequently found to suffer stress-relaxation cracking (or reheat cracking). Titanium is used in Type 321 to “stabilize” the alloy by forming titanium carbides (TiCs) in the grain interior to prevent sensitization (i.e., formation of chromium carbides along grain boundaries, thus creating a chromium-denuded zone along grain boundaries) during heat treatment or cooling following welding. Type 347, on the other hand, uses niobium (referred to as columbium until 1968) to form NbC in the grain interior to avoid sensitization during heat treatment or cooling following welding. Stressrelaxation cracking has been reported for these two grades of austenitic stainless steels during service at temperatures between approximately 500 and 750 °C (930 and 1380 °F) (Ref 24). De Santis et al. (Ref 24) summarized a number of failure cases involving these two grades of stainless steels. These authors indicated that the failures were the result of intergranular cracking, and the components that failed were either in a cold-worked structure or were associated with welding. Also, all failures were found to have occurred after relatively short service times. Cases of service failures related to Type 321 and 347 reported by De Santis et al. (Ref 24) are summarized:

A straight Type 347 tube, which was thermally insulated externally and carried a gaseous hydrocarbon stream with traces of hydrogen sulﬁde at 650 to 680 °C (1200 to 1255 °F), failed after only 96 h of service. A similar failure also occurred in the same plant after only 15 days of service involving a heavier wall tube made of Type 347 at the bend near the weld. Failures of Type 321 tubes occurred after service of about 1 month at the bend with the

Stress-Assisted Corrosion and Cracking / 395

tube carrying a synthesis gas stream at about 600 °C (1110 °F) with the outer diameter being thermally insulated. The failure occurred at the intrados of the tube bend. Type 321 tubes imbedded in a catalytic bed of aluminum silicate of a cracking reactor and carrying saturated steam failed at the bend section after 3 months of operation. The tube metal temperatures were between 650 and 750 °C (1200 and 1380 °F). Type 321 serpentine heat-exchanger tubes, which were exposed to synthesis gases at 700 to 750 °C (1290 to 1380 °F) with internal temperature of about 300 °C (570 °F), failed at the bend section after 3 months of service.

The authors (Ref 24) observed that these failures exhibited a common characteristic in which intergranular cracks were oxidized, leaving a metallic ﬁlm in the center of the crack. The metallic ﬁlm was found to be essentially iron with nickel varying from 0 to about 15% and little chromium. In addition to oxides observed in the crack, the authors also observed in most cases iron and nickel sulﬁdes. Both Type 321 and 347 are known to be susceptible to intergranular cracking in the heataffected zone (HAZ) during services at a certain temperature range (Ref 18). It is generally believed (Ref 25–30) that titanium carbides (in Type 321) and niobium carbides (in Type 347) in the region next to the fusion are put back into solution during welding and subsequent reprecipitation of these ﬁne carbides (during cooling) that strengthens the grain matrix in the HAZ (particularly the coarse-grain HAZ) during service or reheating during PWHT. Stress relaxation in the matrix-strengthened HAZ during service or PWHT causes cracking to develop at grain boundaries, thus resulting in brittle, intergranular fracture at the HAZ. Alloy 800H (Fe-32Ni-21Cr-0.4Al-0.4Ti) is the most widely used Fe-Ni-Cr alloy for hightemperature applications in the petrochemical industry. Stress-relaxation cracking has also been observed in this alloy under certain conditions. Kohut (Ref 31) reported intergranular cracking failures in a transfer line made of alloy 800H after several months of service in a temperature range of 540 to 705 °C (1000 to 1300 °F). The transfer line ran between a pressure vessel and a feedefﬂuent exchanger in a petrochemical unit. The pipe was made from solution-annealed alloy 800H plate. Fabrication of pipe involved press breaking the plate into two pipe halves followed

pg 395

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 396

396 / High-Temperature Corrosion and Materials Applications

by seam welding using shielded metal arc welding. Press breaking produced cold-worked structure in the pipe halves. The 30 cm (12 in.) (diameter) transfer line failed at about 1 m (3 ft) from the ﬂange of the feed-efﬂuent exchanger. Four months later, a second failure occurred at about 10 m (30 ft) from the ﬁrst failure. Both failures were found to occur in the areas with some deformation, which was believed to be the result of the local repair. A third failure occurred in the heat-affected zone of a weld initiating from the inside diameter of tube surface after 1 year of service. Electron microprobe analysis showed no chlorine, sulfur, or other contaminants that would cause embrittlement of the alloy. The material near the fracture still exhibited good room-temperature tensile ductility but much higher yield and tensile strengths. Table 14.1 shows the room-temperature tensile test results of the alloy 800H material obtained from the 30 cm (12 in.) transfer line near the fracture area compared with the tensile test results of the material that was obtained from the same area but was re-solution annealed. The stress-rupture tests at 595 °C (1100 °F), however, revealed that the material near the fracture was very brittle, as shown in Table 14.2 (Ref 31). The sample that was obtained from the material near the fracture exhibited extremely low elongation under the test condition, while the sample obtained from the same area, but was given a re-solution annealing treatment, was found to exhibit excellent ductility under the same test condition. Korkhaus (Ref 32) observed similar brittle, intergranular cracking in an alloy 800H pipe that was part of an outlet header of a process air preheater coil operated at about 600 °C (1110 °F). The leakage of the header was detected after 2 years of service. Cracks were observed to occur primarily near the pipe-to-tube weld and in the vicinity of the longitudinal Table 14.1 Results of room-temperature tensile tests on alloy 800H samples obtained from the area near the fracture as well as the material from the same area but was re-solution annealed Material

Material near fracture Re-solution annealed Source: Ref 31

Yield strength, MPa (ksi)

Ultimate tensile strength, MPa (ksi)

Elongation, %

Reduction in area, %

460 (68)

765 (111)

23.5

46.5

205 (30)

506 (74)

55.0

72.2

pipe seam weld. The alloy 800H header was 508 mm in diameter, had 34.5 mm wall thickness, and was 4720 mm long. The header was fabricated by cold bending the plate material into pipe halves that were then seam welded using a nickel-base alloy ﬁller metal. No heat treatment of this cold-formed and seam-welded pipe was performed prior to installation as a header. A joint industry sponsored project addressing stress-relaxation cracking failures in austenitic welded joints was undertaken by Van Wortel (Ref 33) at TNO Institute of Industrial Technology in Apeldoorn, The Netherlands. Van Wortel reported major observations and conclusions derived from the test program. Since the brittle, intergranular cracking failures were mainly associated with cold working and/or welding, the program mainly focused on the effect of cold working and welding on the susceptibility of stress-relaxation cracking. A specially developed three-point bending test rig, which was capable of producing a cracking failure with a similar crack morphology to the service-induced failure, was used to determine the susceptibility of the alloy to stress-relaxation cracking. Testing procedures were not described in the paper. The test results generated by Van Wortel are summarized in Table 14.3 (Ref 33). Type 321 is generally considered to be susceptible to stress-relaxation cracking in the industry. In Van Wortel’s test program, however, Type 321 was found to suffer no cracking under the test condition. It is believed that the 321H material used in Van Wortel’s test program was a ﬁne-grained material with ASTM grain size No. 7. Fine-grained materials are less susceptible to stress-relaxation cracking. All other materials tested were coarse-grained materials with ASTM grain sizes of No. 2 and coarser. Van Wortel (Ref 33) reported that several heat treatments were beneﬁcial in improving the resistance of the alloy to stress-relaxation cracking based on

Table 14.2 Results of stress-rupture tests performed at 595 °C (1100 °F) and 276 MPa (40 ksi) on alloy 800H samples obtained from the area near the fracture as well as the material from the same area but was re-solution annealed Material

Material near fracture Re-solution annealed Source: Ref 31

Rupture life, h

Elongation, %

Reduction in area, %

385 261

1.2 15.8

12.5 18.6

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

Table 14.3

pg 397

Stress-Assisted Corrosion and Cracking / 397

Results of relaxation tests on cold-worked plates and welded joints 304H

Condition

at 575 °C

Weldment PWHT(a) Aged(b) CW(c) HT(1)(d) HT(2)(e) HT(3)(f)

Cracks No cracks No cracks Cracks No cracks No cracks …

321H

AC66

at 650 °C

at 575 °C

at 650 °C

at 575 °C

800H at 650 °C

at 650 °C

617

No cracks No cracks … … … … …

No cracks … … No cracks … … …

No cracks … No cracks No cracks … … …

No cracks No cracks … … … … …

Cracks No cracks Cracks Cracks No cracks No cracks No cracks

Cracks No cracks Cracks Cracks … … …

Note: Grain sizes for the alloys under testing were ASTM No. 2 for 304H, No. 7 for 321H, No. 2 for AC66, No. 0 for 800H, and No. 1 and No. 0 for 617. (a) PWHT at 875 °C/ 3 h for 304H and 800H; 980 °C/3 h for 617 after welding and before testing. (b) Material aged at 650 °C for 16,000 h before testing. (c) Cold worked (CW) the metal from 0 to 15% before testing. (d) Heat treated the metal at 875 °C/3 h for 304H and 980 °C/3 h for 800H before cold working and testing. (e) Heat treated the metal after cold working and before testing at 875 °C/3 h for 304H and 980 °C/3 h for 800H. (f ) Heat treated the metal from in-service cracked material at 980 °C/3 h before testing. Source: Ref 33

Testing temperature, °C 100

50

200

300

400

500

600

700

Solution annealed Aged 4000 h at 540 °C (1000 °F) Aged 4000 h at 595 °C (1100 °F) Aged 4000 h at 650 °C (1200 °F)

45

300

250

35 30

200

25 150

15 10 5

Room temperature

20 100

0.2% offset yield strength, MPa

40 0.2% offset yield strength, ksi

his three-point bending testing method. However, no examples were given to illustrate the effectiveness of these heat treatments in actual applications. Van Wortel (Ref 33) indicated that the alloy is most susceptible to stress-relaxation cracking when the alloy matrix is in an age-hardened condition with a high density of ﬁne precipitates within the grain. This condition can signiﬁcantly reduce the deformation capability of the alloy. Alloys under this condition are expected to suffer cracking after only 0.1 to 0.2% relaxation strain (Ref 33). Cold working can accelerate the precipitation process within the grain, making the alloy more susceptible to stress-relaxation cracking. Van Wortel (Ref 33) proposed that the microstructure in alloy 800H susceptible to stress-relaxation cracking consisted of very ﬁne matrix carbides, which strengthen the grain matrix, and grain-boundary carbides along with a denuded carbide zone along the grain boundaries that weaken the grain boundaries. Van Wortel, however, did not mention the formation of γ′ [Ni3 (Al,Ti)] precipitates in the grain matrix in the age-hardened condition for Alloy 800 or 800H. The major strengthening mechanism for alloy 800H at temperatures at which the alloy is susceptible to stress-relaxation cracking is, in fact, the precipitation of ﬁne, coherent γ′ [Ni3(Al, Ti)] precipitates (Ref 34–36). Lai and Kimball (Ref 37) observed that the maximum age hardening for alloy 800H was at 595 and 650 °C (1100 and 1200 °F). These authors further indicated that the strengthening was more pronounced at the elevated temperatures than at room temperature. This is illustrated in Fig. 14.29 (Ref 37). The retained tensile elongation at both room-temperature and aging temperatures after 4000 h of aging was found to be quite good (approximately 35 to 45% tested at

50

0 200

0 400 600 800 1000 1200 1400 Testing temperature, °F

Fig. 14.29

Strengthening at room temperature compared with strengthening at the aging temperatures after aging at 540, 595, and 650 °C (1000, 1100, and 1200 °F) for 4000 h for alloy 800H containing 0.39% Al and 0.44% Ti. Source: Ref 37

room temperature and approximately 25 to 35% tested at respective aging temperatures) (Ref 37). It is well known that age hardening in alloy 800 or 800H is a strong function of combined Al+Ti content in the alloy. Alloys with higher content of Al + Ti exhibit higher degree of age hardening due to formation of more ﬁne γ′ [Ni3(Al,Ti)] precipitates. An example is given in Fig. 14.30, which shows the aging behavior of two heats of alloy 800H with Heat A containing higher combined Al + Ti content (0.83%) than Heat E

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 398

398 / High-Temperature Corrosion and Materials Applications

100

95 650 °C (1200 °F)

Heat A Heat E

90

595 °C (1100 °F)

540 °C (1000 °F)

Hardness, HRB

85

540 °C (1000 °F) 595 °C (1100 °F)

80

650 °C (1200 °F) 75

760 °C (1400 °F)

70 760 °C (1400 °F)

815 °C (1500 °F)

65

60 1 Unaged

103

104

105

Aging time, h

Aging behavior of two heats of alloy 800H with Heat A containing higher combined Al + Ti (0.83%) showing significantly more age hardening than Heat E with much lower combined Al + Ti (0.58%). Source: Ref 37

(0.58% of combined Al + Ti) (Ref 37). Heat A showed signiﬁcantly more age hardening than Heat E with both materials having similar grain sizes. Increasing the combined Al + Ti content in alloy 800H causes signiﬁcant strengthening in creep-rupture strength. This is illustrated in Fig. 14.31 (Ref 38). Increasing the combined Al +Ti content in alloy 800H increases the 1% creep strength at 650 °C (1200 °F), which was the maximum age-hardening temperature. At 850 °C, γ′ [Ni3(Al,Ti)] precipitates coarsen in alloy 800H; thus the effect of the combined Al +Ti content signiﬁcantly diminishes, as shown in Fig. 14.31 (Ref 38). Accompanying the strengthening of the alloy produced by γ′ [Ni3(Al,Ti)] precipitates, alloy 800H exhibits creep embrittlement at the maximum age-hardening temperature, particularly for the heats containing higher levels of combined Al +Ti content. This is illustrated in Fig. 14.32 (Ref 38). Figure 14.33 also shows the similar effect of drastic reduction of stressrupture ductility for alloys containing higher levels of Al+Ti in alloy 800H when tested at 650 °C (1200 °F)—the maximum agehardening temperature (Ref 39). The data suggest that the low rupture ductility in alloy 800 at 600 and 650 °C (1110 and 1200 °F) is the result of γ′ [Ni3(Al,Ti)] precipitates with higher combined Al +Ti content producing a larger

150

RP1/10,000 650 °C (1200 °F) Stress (N/mm2)

Fig. 14.30

102

10

815 °C (1500 °F)

RP1/30,000

100

50 850 °C (1560 °F)

RP1/10,000 RP1/30,000

0 0.4

0.6

0.8

1.0

1.2

1.4

Σ Ti + AI, %

Fig. 14.31

Effect of combined Al + Ti content in alloy 800H on the creep strength (1% in 10,000 h and 30,000 h) at 650 and 850 °C (1200 and 1560 °F). Source: Ref 38

volume of these ﬁne, coherent precipitates, thus resulting in signiﬁcant reduction in stress-rupture ductility at these temperatures. The rupture ductility of alloy 800 can be further reduced to close to nil with prior cold work when tested at 600 or 650 °C (1110 or 1200 °F). Stone et al. (Ref 39) showed the effect of cold work on stress-rupture ductility when tested at 600 °C (1110 °F) for alloy 800. Their test results showed that the

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

rupture elongation was reduced from about 10% in the solution-annealed condition to almost nil with 10 to 20% prior cold work, as shown in Fig. 14.34 (Ref 39). Similar results were observed by Smith (Ref 40) when he conducted stress-rupture tests at temperatures from 540 to 650 °C (1000 to 1200 °F). Smith’s test results are summarized in Table 14.4 (Ref 40). All the prestrained specimens were quite brittle under the test conditions, exhibiting rupture elongation of less than 1%. The elongation for the

60

50

Elongation, %

0.1 h 40

30

1000 h

20 10,000 h 10

0

30,000 h

0.4

0.6

0.8

1.0

1.2

1.4

Σ Ti + AI, %

Fig. 14.32

Creep rupture ductility of alloy 800H as a function of combined Al + Ti content in the alloy tested at 650 °C (1200 °F). Source: Ref 38

300

Stress, MPa

0.4% AI 0.6% Ti 0.2% AI 0.4% Ti 100

50

Elongation, %

30 50 40 30 20 10 0 10

Fig. 14.33

0.2% AI 0.4% Ti 0.4% AI 0.6% Ti

102

103 Time, h

104

105

Effects of combined Al + Ti content in alloy 800 on stress rupture strength and rupture ductility at 650 °C (1200 °F). Source: Ref 39

Stress-Assisted Corrosion and Cracking / 399

prestrained specimens, which was so low it could not be obtained by the traditional method of ﬁtting the ends of the ruptured specimen together for determination of the percent of elongation, was taken from the creep curves. 14.3.4 Cracking in Nickel-Base Alloys Some wrought nickel-base chromia-forming alloys (i.e., nickel-base alloys forming chromium oxide scales for high-temperature corrosion resistance) that are designed for high-temperature applications contain a relatively low level of aluminum, typically 1 to 2%, to further improve the oxidation resistance of the alloy. Some of the more familiar alloys in this group includes alloys 617 (1.2% Al) and 601 (1.4% Al). As discussed earlier, Van Wortel (Ref 33) reported that alloy 617 was susceptible to stress-relaxation cracking at 650 °C (1200 °F), as shown in Table 14.3. The author, however, provided little information about the microstructure and aging characteristics of alloy 617. Alloy 617 also is known to age harden at 600 and 650 °C (1110 and 1200 °F), as shown in Fig. 14.35 (Ref 41). Alloy 617, similar to alloy 800H, exhibits excellent room-temperature ductility after longterm aging, as shown in Fig. 14.36 (Ref 41). Nevertheless, alloy 617, similar to alloy 800H, was susceptible to stress-relaxation cracking at 650 °C (Table 14.3). Bassford and Schill (Ref 42) believed that γ′ precipitates are a major contributor to age hardening in alloy 617 at 650 °C as well as 600 °C. Both alloys 800H and 617 are strengthened primarily by γ′ precipitates at temperatures when stress-relaxation cracking has been observed. For nickel-base alloys, this type of cracking is generally referred to as “strain-age cracking.” Stahl et al. (Ref 43) reported a reformer tube failure in a steam reformer for hydrogen production. The reformer tube was made of alloy 601. The product gas cooled from 900 to 600 °C (1650 to 1110 °F) in the tube and then ﬂowed through the outlet manifold system. Cracking was observed to occur near the outlet header in the heat-affected zone parallel to the weld fusion line after about 1 year of operation. Their repair method involved cutting out all the problem connections between the reformer tube and the outlet header, which were replaced with new outlet manifold welded connections. These welded connections were then heat treated at 950 °C (1740 °F) for 1 h before they were connected to the reformer tubes. The postfabrication

pg 399

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 400

400 / High-Temperature Corrosion and Materials Applications

Stress, MPa

500 5% CW 10% CW 20% CW 25% CW

300

200

ST. 1120 °C (2050 °F)

100

Elongation, %

40 30 20 ST. 1120 °C (2050 °F)

10 0 10

102

103

104

105

Time, h

Fig. 14.34

Effects of cold work (cw) on the rupture ductility of alloy 800 at 600 °C (1110 °F). Solid lines represent the data generated from the specimens in the as-solution heat treated condition. The specimens were solution heat treated (ST) at 1120 °C (2050 °F). Source: Ref 39

540 (1000) 565 (1050) 595 (1100)

650 (1200)

Applied stress, MPa (ksi)

345 (50) 310 (45) 310 (45) 275 (40) 241 (35) 207 (30) 207(a) (30(a)) 172 (25) 155 (22.5) 138 (20)

Rupture life, h SA

1214 3350 615 1953 772 2670 … 546 1152 2352

PS

129 319 92 546 247 721 7342(a) 861 1565 2589

Rupture elongation, % SA

PS

19 13 15 11 13 14 … 16 20 22

0.26 0.20 0.2 0.3 0.2 0.28 0.3(a) 0.45 0.4 0.33

Exposure temperature, °C 650 705

760

90

600

Exposure time, h 48,000 h

80 12,000 h

70

500

4,000 h 1,000 h

60

400 100 h

50

Not exposed

300

Yield strength, MPa

Test temperature, °C (°F)

595

Yield strength, ksi

Table 14.4 Effect of 20% prestrain in compression on stress rupture ductility of alloy 800H tested at 540, 565, 595, and 650 °C (1000, 1050, 1100, and 1200 °F) using bar specimens in both solution-annealed (SA) and prestrained (PS) conditions

40 30

1100

1200 1300 Exposure temperature, °F

1400

(a) Prestrained 20% in tension; broke at extensomer weld. Source: Ref 40

Fig. 14.35

heat treatment, according to the authors (Ref 43), was (a) to eliminate the effect of retained coldwork and welding stresses and (b) to coarsen both the grain-boundary precipitates and grainmatrix precipitates to reduce age-hardening effects. These authors did not mention in their paper about γ′ (Ni3Al) precipitates being a possible strengthening mechanism in alloy 601,

which contains about 1.4% Al and is known to form γ′ (Ni3Al) precipitates. The heat treatment at 950 °C is believed to mainly coarsen both grain-boundary and intragranular carbides in addition to relieving residual stresses resulting from cold working and welding. It should provide some improvement to the resistance of the alloy to strain-age cracking. Nevertheless, the

Room-temperature yield strengths of alloy 617 after aging at different temperatures for various times up to 48,000 h. Source: Ref 41

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

temperature of 950 °C (1740 °F) in the heat treatment would be above the solvus temperature of the γ′ (Ni3Al) precipitates, and γ′ (Ni3Al) precipitates can form again during the service at strain-age cracking temperatures of 600 and 650 °C (1110 and 1200 °F). Lai (Ref 16) reported a case of strain-age cracking involving a recuperator shell made of alloy 601 that was used for preheating air for combustion. The recuperator shell suffered through-thickness cracking after 2.5 years of service. The failure occurred at the location where the metal temperature was approximately 590 °C (1100 °F). No cracking was observed at other locations with temperatures either higher or lower than 590 °C (1100 °F). Both the air side and ﬂue gas side of the recuperator shell showed little oxidation or corrosion attack. The recuperator shell (7.9 mm, or 5/16 in. thick) failed by intergranular fracture, as shown in Fig. 14.37. Figure 14.38 shows a crack propagating along grain boundaries. The ﬁgure shows some evidence of cold work that is likely to result from the original fabrication of the recuperator shell. The tensile property of alloy 601 from the recuperator shell near the main fracture was evaluated. Tensile blanks were obtained and machined into round tensile specimens with adequate material being machined off from both internal and external surfaces (ID and OD) of the shell. Table 14.5 shows the tensile properties of the material from the area that was near the main fracture (Ref 16). The material from the recuperator shell exhibited good tensile ductility at

Stress-Assisted Corrosion and Cracking / 401

room temperature. At 590 °C (1100 °F), however, the ductility was found to be much lower and the yield strength was very high, which was close to that at room temperature. Transmission electron microscopy examination of the recuperator shell showed the presence of ﬁne, γ′ precipitates, as shown in Fig. 14.26. Tensile blanks were obtained from the recuperator shell and re-solution annealed at 1150 °C (2100 °F) for 30 min followed by water quenching. Tensile properties of these re-solution annealed specimens at both room temperature and 590 °C (1100 °F) are included in Table 14.5 for comparison. The tensile properties of re-solution annealed material were similar to those of solution-annealed material reported in the literature (Ref 44). Assuming that the re-solution annealed material has the properties of the original plate material prior to service, the data indicated that the alloy 601 recuperator shell increased its yield strength at 590 °C (1100 °F) by almost three times (but only two times at room temperature) as a result of the service exposure at about 590 °C (1100 °F). In order to determine that

(a)

600

Exposure temperature, °C 650 700

750

Not exposed

70

Elongation, %

60 50 Exposure 40

100 h 1000 h

time, h 4,000 h 12,000 h

30 48,000 h

8,000 h

20 1100

Fig. 14.36

1200 1300 Exposure temperature, °F

1400

Room-temperature tensile elongation of alloy 617 after aging at different temperatures for various times up to 48,000 h. Source: Ref 41

(b)

Fig. 14.37

Scanning electron micrographs, (a) low magnification and (b) high magnification, showing intergranular fracture surface of the alloy 601 recuperator shell. Original magnification: (a) 13× and (b) 100×. Source: Ref 16

pg 401

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 402

402 / High-Temperature Corrosion and Materials Applications

the recuperator shell material was in a brittle condition (extremely low creep ductility) under creep condition, stress rupture testing was performed at 590 °C (1100 °F) involving smoothnotch specimens obtained from the recuperator

Fig. 14.38

Optical micrograph showing intergranular cracking near the main fracture. Original magnification: 100×. Source: Ref 16

shell material and the shell material after resolution annealing at 1150 °C (2100 °F). The test results are shown in Table 14.6 (Ref 16). The brittle nature of the alloy under the creep test condition at 590 °C (1100 °F) was substantiated from these tests. Furthermore, the test results indicated that a full recovery of the original ductile material condition can be obtained by re-solution annealing the material. Lai (Ref 16) presented additional test data to further substantiate that alloy 601 creep embrittlement was related to the age-hardened condition produced during service at 590 °C (1100 °F). An annealed alloy 601 plate of different heat number was obtained from an alloy supplier for evaluation. Specimen blanks were aged at 590 °C (1100 °F) for 10,000 h. Stressrupture tests were performed at 590 °C (1100 °F) and 310 MPa (45 ksi) on both annealed and aged specimens using smooth-notch specimens. Results of these tests are summarized in Table 14.7. The aged specimen, although fractured at the smooth section, clearly showed an embrittled condition with very low creep ductility and intergranular fracture. The as-received specimen showed much better creep ductility with dimple rupture, which is characteristic of a ductile rupture. Many high-temperature, nickel-base alloys are solid-solution strengthened using primarily molybdenum and/or tungsten to increase their tensile and creep rupture strengths. These alloys

Table 14.5 Tensile properties of alloy 601 specimens obtained from the area near the main fracture in the alloy 601 recuperator shell that suffered intergranular through-wall cracking after 2.5 years of service with the metal temperature at about 590 °C (1100 °F) during service. For each test condition, two tests were run and both values are reported. Material

From recuperator RSA(a)

Test temperature, °C (°F)

0.2% yield strength, MPa (ksi)

Ultimate tensile strength, MPa (ksi)

Elongation, %

Reduction in area, %

Room temperature 590 (1100) Room temperature 590 (1100)

426, 414 (61.8, 60.0) 372, 386 (54.0, 56.0) 232, 224 (33.7, 32.5) 154, 142 (22.4, 20.6)

883, 865 (128.0, 125.5) 606, 600 (87.9, 87.0) 616, 614 (89.3, 89.1) 481, 478 (69.7, 69.3)

36, 34 18, 13 56, 53 63, 62

32, 34 15, 15 66, 53 55, 54

(a) Test blanks from the recuperator shell near the main fracture were re-solution annealed (RSA) at 1150 °C (2100 °F) for 30 min followed by water quenching. Source: Ref 16

Table 14.6 Results of stress rupture tests performed at 590 °C (1100 °F) and 310 MPa (45 ksi) on smooth-notch round specimens obtained from the 601 recuperator shell and from the re-solution-annealed recuperator shell material Specimen condition

From recuperator shell From recuperator shell + 1150 °C (2100 °F) for 30 min and water quench Source: Ref 16

Failure location

Fracture mode

Notch, rupture life: 3110 h Smooth section, rupture life: 175 h, 22% elongation, 14% reduction in area

Intergranular fracture Dimple rupture

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

Table 14.7 Results of stress rupture tests at 590 °C (1100 °F) and 310 MPa (45 ksi) on smooth-notch specimens of as-received and aged alloy 601 specimens from alloy 601 plate of different heat number obtained from an alloy supplier Specimen

Failure Rupture Elongation, Reduction location life, h % in area, %

As-received Smooth Aged(a) Smooth

963 5348

19 4

21 4

Fracture mode

Dimple Intergranular

(a) Aged at 590 °C (1100 °F) for 10,000 h. Source: Ref 16

Table 14.8 Results of stress rupture tests at 590 °C (1100 °F) and 310 MPa (45 ksi) on smooth-notch specimens of as-received and aged alloy X specimens Specimen

Failure location

Rupture life, h

Elongation, %

Reduction in area, %

Fracture mode

As-received Aged(a)

Smooth Smooth

1574 1919

42 55

44 63

Dimple Dimple

(a) Aged at 590 °C (1100 °F) for 10,000 h. Source: Ref 16

Stress-Assisted Corrosion and Cracking / 403

is actually strengthened by ﬁne, coherent γ″ (Ni3Nb) precipitates, and alloy X-750 containing Al, Ti, and Nb is strengthened by both γ′ and γ″ precipitates. Many of the applications for these alloys are in sheet products. As a result, fabricated products are under much less constraint than are heavy section components. Furthermore, the grain sizes for sheet products are typically ASTM No. 4 or ﬁner. Accordingly, components made of sheet products are less susceptible to strain-age cracking. Nevertheless, in some cases components made of sheet products for these alloys can be heavily formed with severe cold-worked conditions, thus making them susceptible to strain-age cracking even during postfabrication heat treatments. Rowe (Ref 45) studied strain-age cracking of wrought γ′- or γ″-strengthened sheet alloys by using a controlled heating rate testing (CHRT) method that was developed in 1960s (Ref 46). Testing involves heating a tensile specimen at a controlled heating rate to a test temperature in the γ′ or γ″ precipitation temperature range, and then pulling the specimen in tension to failure. The heating rate of the postweld annealing heat treatment of a fabricated component could be used for the CHRT in the test. The minimum tensile elongation from the test results is then used as a guide for the susceptibility to strain-age cracking for the alloy. Figure 14.39 summarizes the test results generated by Rowe (Ref 45). The results indicated that increasing the combined Al + Ti + Nb content in the alloy decreases the minimum elongation, thus increasing the susceptibility to strain-age cracking.

20 Minimum CHRT elongation, %

are also strengthened by chromium-rich carbides. Among the most widely used solid-solutionstrengthened alloys for high-temperature applications is alloy X (Ni-22Cr-18Fe-9Mo), which among other applications is widely used for gas turbine combustion chambers in both landbased gas turbine power generation and aircraft engines. In smooth-notch stress rupture tests to investigate the susceptibility of strain-age cracking of alloy 601, Lai (Ref 16) included alloy X to examine this solid-solutionstrengthened alloy, which is also strengthened by carbides (primarily M6C carbides) but with no γ′ precipitates. The results of his tests on alloy X are summarized in Table 14.8 (Ref 16). Alloy X aged for 10,000 h at 590 °C (1100 °F), under the same test conditions as were used for alloy 601, was found to exhibit excellent creep ductility (about 55% rupture elongation) with ductile dimple rupture. Some nickel-base alloys use γ′ [Ni3(Al,Ti)] precipitates for providing strengthening in addition to solid-solution strengthening using elements such as molybdenum and/or tungsten. Some of the widely known wrought alloys in this group include alloys 263 (Ni-20Cr-20Co6Mo-0.5Al-2.2Ti), 718 (Ni-18Cr-19Fe-3Mo5Nb), X-750 (Ni-16Cr-7Fe-0.7Al-2.5Ti-1Nb), Waspaloy (Ni-19Cr-14Co-4Mo-1.5Al-3Ti), and R-41 (Ni-19Cr-11Co-10Mo-1.5Al-3Ti). Alloy 718, which contains Nb instead of Al and Ti,

pg 403

263

18 16 14 12

X-750

10

718

8 6

Waspaloy

4

PK33

2

R-41

0 2

3

4

5

6

7

8

Al + Ti + Nb, at.%

Fig. 14.39

Results of controlled heating rate tests (CHRT) in terms of minimum elongation (%) as a function of combined Al + Ti + Nb content in atomic percent for 0 selected c - and/or c 00 -strengthened wrought sheet alloys. Source: Ref 45

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

pg 404

404 / High-Temperature Corrosion and Materials Applications

Nickel-base alloys containing high levels of aluminum or aluminum plus titanium can be highly susceptible to strain-age cracking, particularly in plate or thick-wall tubing products. Some of these alloys include alloy 214 (4.5% Al), which forms a tenacious Al2O3 oxide scale to resist oxidation at very high temperatures, and alloy 693 (3% Al), which forms Cr2O3/ Al2O3 oxide scales to resist metal dusting in the temperature range of 480 to 700 °C (900 to 1300 °F). Ni-Cr-Mo and Ni-Mo alloys have the potential for developing a long-range ordered Ni2(Cr,Mo) phase after long aging times at a certain temperature range. Tawancy (Ref 47) observed the formation of [Ni2(Cr,Mo)] phases in alloy S (Ni-15.5Cr-14.5Mo) after aging for 8000 h at 540 °C (1000 °F), as shown in Fig. 14.40 (Ref 47). The ordered phase caused the alloy to signiﬁcantly increase its room-temperature tensile strengths, particularly yield strength, as shown in Table 14.9 (Ref 47). Tawancy and Asphahani (Ref 48) reported that C-276 alloy (Ni-16Cr-16Mo-4W) exhibited signiﬁcant roomtemperature strengthening when the alloy was aged to develop a fully ordered structure, as shown in Table 14.10. The Ni2(Cr,Mo) ordered phase exhibits a morphology similar to γ′ and γ″ precipitates and produces similar strengthening effects as γ′- and γ″-strengthened alloys.

It is quite likely that a long-range ordered NiCr-Mo alloy may also be susceptible to strain-age cracking. Alloy C-22 (Ni-22Cr-13Mo-3W) is selected to be a candidate nuclear waste container material for the Yucca Mountain site in Nevada (Ref 49). Studies on the changes in microstructure, mechanical properties, and corrosion resistance of C-22 alloy after long-term aging have been carried out in recent years (Ref 50– 52). Rebak et al. (Ref 52) reported the observation of long-range ordered phases, [Ni2(Cr, Mo)], in C-22 alloy after aging at 595 °C (1100 °F) for 1000 and 16,000 h, at 540 °C (1000 °F) for 1000 h, and at 430 °C (800 °F) for 30,000 and 40,000 h. Room-temperature tensile reduction in area of C-22 was reported to have been about 80% in annealed condition to 75% after aging for 40,000 h at 430 °C (800 °F), which was apparently in the early stage of longrange ordered condition (Ref 52). In a similar study on aging behavior of C-22 alloy, the alloy after aging at 595 °C (1100 °F) for 16,000 h showed no strengthening for one heat but slight strengthening for the other heat, as shown in Table 14.11 (Ref 51). The authors (Ref 51) observed continuous grain-boundary carbides in the aged samples and mentioned about the formation of Ni2(Cr,Mo) ordered phases, but without presenting positive identiﬁcation of the ordered phases using transmission electron microscopy. With either no strengthening or slight strengthening as presented in Table 14.11,

Table 14.9 Room-temperature tensile properties of alloy S comparing annealed material with the material aged for 8000 h at 540 °C (1000 °F) Condition

Annealed specimen Aged specimen(a)

0.2% yield strength, Ultimate tensile Elongation, MPa (ksi) strength, MPa (ksi) %

436 (63) 824 (119)

889 (129) 1284 (186)

58 42

(a) Aged at 540 °C (1000 °F) for 8000 h. Source: Ref 47

Table 14.10 Room-temperature tensile properties of C-276 alloy comparing annealed material (disordered structure) with aged material with a long-range ordered structure Condition

Fig. 14.40

Transmission electron micrograph showing longrange ordered phases [Ni2(Cr,Mo)] in a dark field image using a h220i reflection in alloy S after 8000 h at 540 °C (1000 °F). Source: Ref 47

Annealed Ordered structure(a)

0.2% yield strength, MPa (ksi)

Ultimate tensile strength, MPa (ksi)

Elongation, %

356 (51.6) 770 (111.7)

792 (114.8) 1240 (179.8)

61 28

(a) Aging condition was not reported in the paper. Source: Ref 48

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

it is unlikely the ordered phases were fully developed. The slight reduction in ductility as shown in Table 14.11 is most likely caused by the formation of continuous grain-boundary carbides. The mechanical properties that were under investigation in these studies (Ref 50–52) included tensile tests and impact toughness testing. When long-range ordered phases are fully developed, it may be imperative to perform stress-relaxation tests or smooth-notch stress rupture tests at the ordering temperatures to determine the susceptibility of the alloy to stressrelaxation cracking (or strain-age cracking). A relatively new age-hardenable Ni-Cr-Mo alloy (C-22HS) with 21Cr and 17Mo in nickel was developed using long-range ordered phases [Ni2(Cr,Mo)] to produce strengthening through a heat treatment (Ref 53). The heat treatment produces a fully developed ordered structure that produces signiﬁcant strengthening, particularly at the ordering temperature (i.e., 595 °C, or 1100 °F). Tensile properties of Alloy C-22HS in both as-annealed and as-heat-treated conditions are summarized in Table 14.12 (Ref 53). This behavior is quite similar to the alloys strengthened by γ′ precipitates, such as Waspaloy, R-41, and so forth. Thus, the alloys that are strengthened by the ordered Ni2(Cr,Mo) phase may be equally susceptible to strain-age cracking as γ′ alloys.

Table 14.11 Room temperature tensile properties of C-22 alloy comparing annealed material with the material aged for 16,000 h at 595 °C (1100 °F)

Heat

Condition

No. 1 Annealed Aged(a) No. 2 Annealed Aged(a)

0.2% yield strength, MPa (ksi)

Ultimate tensile strength, MPa (ksi)

340 (49.3) 381 (55.2) 342 (49.6) 485 (70.3)

768 (111.3) 806 (116.9) 757 (109.7) 931 (135.0)

Elongation, Reduction % in area, %

66.2 55.4 70.0 39.0

76.1 50.7 77.0 31.1

(a) Aged at 595 °C (1100 °F) for 16,000 h. Source: Ref 51

pg 405

Stress-Assisted Corrosion and Cracking / 405

14.3.5 Mitigation of Reheat Cracking and Stress-Relaxation Cracking or Strain-Age Cracking As was discussed in previous sections, the reheat cracking due to heat treatment or the stress-relaxation cracking (or strain-age cracking) that occurs during service is primarily related to strengthening of the grain matrix by ﬁne precipitates that formed at the intermediate temperatures. Thus, to minimize the chances of developing reheat cracking during postweld heat treatment of a welded component or annealing/ stress-relieving of a cold-formed component, it is recommended that the component be heated through that temperature range quickly to minimize the formation of those detrimental precipitates. As for the stress-relaxation cracking (or strain-age cracking) that occurs during service, one effective method for mitigating the cracking is to remove the residual stresses in the component by performing a postweld heat treatment for a welded component or an annealing/stress-relieving heat treatment for a coldformed component (e.g., pipe U-bend section) prior to service. Other factors that may be beneﬁcial in reducing the cracking tendency during service include (a) minimizing stress risers, (b) the use of a ﬁne-grained material, and (c) a heat treatment involving heating the component to a temperature high enough to coarsen precipitates (e.g., carbides) to reduce the matrix strengthening.

14.4 Summary Most high-temperature components are under stress during service. The tensile stresses (or strains) that are imposed on the component can cause the alloy to suffer preferential corrosion penetration in high-temperature, corrosive environments, particularly low-oxygen and highsulfur partial pressure conditions. The effects

Table 14.12 Tensile properties of C-22HS (a heat treatable alloy) comparing the annealed material and heat treated material tested at both room temperature and 595 °C (1100 °F) Condition

Test temperature, °C (°F)

0.2% yield strength, MPa (ksi)

Ultimate tensile strength, MPa (ksi)

Elongation, %

Reduction in area, %

Annealed HT(a) Annealed HT(a)

Room temperature Room temperature 595 (1100) 595 (1100)

406 (59) 742 (108) 222 (32) 542 (79)

822 (119) 1232 (179) 640 (93) 934 (135)

61 40 72 48

75 50 67 66

(a) Heat treated at 705 °C (1300 °F)/16 h/furnace cool to 605 °C (1120 °F)/32 h followed by air cool. Source: Ref 53

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

406 / High-Temperature Corrosion and Materials Applications

of the tensile stresses and strains on the preferential sulﬁdation penetration attack on various alloys in sulﬁdizing environments are reviewed. The preferential sulﬁdation penetration is considered to be a precursor to the circumferential cracking that has been observed in the waterwall tubes for some supercritical coal-ﬁred boilers ﬁred under low NOx combustion conditions. A detailed discussion on the preferential sulﬁdation penetration and the circumferential cracking of the waterwall tubes observed in some supercritical coal-ﬁred boilers is presented. Additional discussion on preferential sulﬁdation penetration and circumferential cracking observed in waterwall tubes is also presented in Chapter 10 “Coal-Fired Boilers.” Stresses that are either residual stresses in the component or externally applied stresses can cause brittle, intergranular cracking when the component is under service at the low end of the intermediate temperature range. This cracking phenomenon is frequently referred to as “reheat cracking,” or “stress-relaxation cracking,” or “strain-age cracking.” Reheat cracking occurs when a welded component is subjected to a postweld heat treatment (PWHT) or when a cold-formed component is subjected to an annealing or stress-relieving heat treatment. Stress-relaxation cracking (or strain-age cracking) occurs when the component is under service. This stress-induced cracking behavior of ferritic steels, stainless steels, Fe-Ni-Cr alloys, and nickel-base alloys is reviewed.

REFERENCES

1. M. Schütze, Deformation and Cracking Behavior of Protective Oxide Scales on Heat-Resistant Steels under Tensile Strain, Oxid. Met., Vol 24 (No. 3/4), 1985, p 199 2. M. Schütze, The Healing Behavior of Protective Oxide Scales on Heat-Resistant Steels After Cracking under Tensile Strain, Oxid. Met., Vol 25 (No. 5/6), 1986, p 409 3. G.R. Smolik and J.E. Flinn, Stress and Environmental Interactions for INCOLOY 800H in Coal Gasiﬁcation Environments, J. Mater. Energy Systems, Vol 8 (No. 3), 1986, p 297 4. V. Guttmann and J. Timm, Corrosion and Creep of Alloy 800H under Simulated Coal Gasiﬁcation Conditions, Werkst. Korros., Vol 39, 1988, p 322

5. M.F. Stroosnijder, V. Guttmann, and J.H.W. de Wit, Corrosion and Creep Behaviour of Alloy 800H in Sulphidizing/Oxidizing/ Carburizing Environments at 700 °C—Part II: Creep Behaviour, Werkst. Korros., Vol 41, 1990, p 508 6. M.F. Stroosnijder, V. Guttmann, and R.J.N. Gommans, Inﬂuence of Creep Deformation on the Corrosion Behaviour of a CeO2 Surface-Modiﬁed Alloy 800H in a Sulphidising-Oxidixing-Carburising Environment, Mater. Sci. Eng., Vol A121, 1989, p 581 7. V. Guttmann, K. Stein, and W.T. Bakker, Deformation-Corrosion Interactions in Selected Advanced High Temperature Alloys, Mater. High Temp., Vol 14 (No. 2/3), 1997, p 61 8. J. Le Coze, U. Franzoni, O. Cayla, F. Devisme, and A. Lefort, The Development of High-Temperature Corrosion-Resistant Aluminium-Containing Ferritic Steels, Mater. Sci. Eng., Vol A120, 1989, p 293 9. Materials, Part D—Properties 2005 Addenda, ASME Boiler and Pressure Vessel Code, ASME, New York, July 1, 2005 10. M. Le Calvar, P.M. Scott, T. Magnin, and P. Rieus, Strain Oxidation Cracking of Austenitic Stainless Steels at 610 °C, Corrosion, Vol 54 (No. 2), 1998, p 101 11. K. Rorbo, Experience with Nitriding of Austenitic Stainless Steel and Inconel in Ammonia Environments, Environmental Degradation of High Temperature Materials, Series 3, No. 13, Vol 2, The Institution of Metallurgists, London, 1980, p 147 12. D.L. Klarstrom, Metallurgical Factors that Promote Cracking During the Heat Treatment of High Performance Alloys, HeatResistant Materials II (Conf. Proc.) Second International Conference on HeatResistant Materials, K. Natesan, P. Ganesan, and G. Lai, Ed., ASM International, 1995, p 487 13. H.M. Tawancy, Structure & Properties of High Temperature Alloys: Applications of Analytical Electron Microscopy, King Fahd University of Petroleum & Minerals, Dhahran, Saudi Arabia, 1993, p 144 14. H.M. Tawancy, Structure & Properties of High Temperature Alloys: Applications of Analytical Electron Microscopy, King Fahd University of Petroleum & Minerals, Dhahran, Saudi Arabia, 1993, p 176 15. R.B. Herchenroeder, G.Y. Lai, and K.V. Rao, A New, Wrought, Heat-Resistant

pg 406

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

Chapter 14:

16. 17.

18. 19.

20.

21.

22.

23.

24.

25.

26. 27. 28. 29.

Ni-Cr-Al-Fe-Y Alloy, J. Met., Vol 35 (No. 11), 1983, p 16 G.Y. Lai, presented at ASM Materials Week (Cincinnati, OH), Oct 20–24, 1991 A. Dhooge and A. Vinckier, Reheating Cracking—A Review of Recent Studies, Int. J. Pressure Vessels Piping, Vol 27, 1987, p 239 C.F. Meitzner, Stress-Relief Cracking in Steel Weldments, WRC Bull, Vol 211, Nov 1975 A. Dhooge, R.E. Dolby, J. Sebille, R. Steinmetz, and A.G. Vinckier, A Review of Work Related to Reheat Cracking in Nuclear Reactor Pressure Vessel Steels, Int. J. Pressure Vessels Piping, Vol 6, 1978, p 329 A.G. Vinckier and A.W. Pense, “A Review of Underclad Cracking in Pressure Vessel Components,” WRC Bulletin No. 197, Welding Research Council, Aug 1974 R.E. Dolby and G.G. Saunders, “A Review of the Problem of Reheat Cracking in Nuclear Vessel Steels,” Report 3453/1/75, The Welding Institute, April 1975 J.G. Nawrocki, J.N. DuPont, C.V. Robino, and A.R. Marder, The Stress-Relief Cracking Susceptibility of A New Ferritic Steel— Part 1: Single-Pass Heat-Affected Zone Simulations, Weld. J., Dec 2000, p 355s J.G. Nawrocki, J.N. DuPont, C.V. Robins, J.D. Puskar, and A.R. Marder, The Mechanism of Stress-Relief Cracking in a Ferritic Alloy Steel, Weld. J., Feb 2003, p 25s R. De Santis, W. Dumini, and G. Casarini, “Service Failure of Stressed Stabilized Stainless Steels at Elevated Temperatures,” WRC Bulletin No. 421, Welding Research Council, 1997, p 17 M. Chabaud, L. Allais, P. Dubuisson, C. Hogrel, and A. Pineau, Ageing and Cracking of Simulated Heat Affected Zones in 321 Stainless Steel, Mater. High Temp., Vol 15 (No. 3/4), 1998, p 395 L. Li and R.W. Messler, Stress Relaxation Study of HAZ Reheat Cracking in Type 347 Stainless Steel, Weld. J., 2000, p 137s R.J. Christoffel, Notch-Rupture Strength of Type 347 HAZ, Weld. J., July 1960, p 315 R.J. Truman and H.W. Kirkby, Some Ductility Aspects of 18-12-1Nb Steel, JISI, 1960, p 180 N.E. Moore and J.A. Grifﬁths, Microstructural Causes of HAZ Cracking in

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

Stress-Assisted Corrosion and Cracking / 407

Heavy Section18-12-Nb Austenitic Stainless Steel Welded Joints, JISI, 1961, p 29 R.N. Younger and R.G. Baker, Heat Affected Zone Cracking in Welded Austenitic Steels During Treatment, Brit. Weld. J., 1961, p 579 G.B. Kohut, A Case History of Cracking Failures in Deformed Austenitic Alloy Pipe Operating in the Temperature Range of 1000–1300 Deg F (538–705 Deg C), Trans. ASME, 1975, p 316 J. Korkhaus, Application of CorrosionResistant Steels in the Chemical Industry, Stainless Steels World 1999 Conference, Conf. Proc., Book 1, KCI Publishing BV, The Netherlands, 1999, p 27 J.C. Van Wortel, Relaxation Cracking in the Process Industry, an Underestimated Problem, BALTICA IV: Plant Maintenance for Managing Life & Performance, Vol 2, S. Hietanen and P. Auerkari, Ed., VTT Technical Research Centre of Finland, Espoo, Finland, 1998, p 637 P.G. Stone, J. Orr, and J.C. Guest, Status Review of Alloy 800, Proc. British Nuclear Energy Society Conference, S.F. Pugh, Ed., AERE Harwell, 1975, p 15 J. Orr, “A Review of the Structural Characteristics of Alloy 800,” Paper No. 2-1, Proc. Petten Conference, North-Holland Publishing, 1978 N. Persson, “Mechanical Properties of Alloy 800 Above 600 C,” Paper No. 3.2-1, Proc. Petten Conference, North-Holland Publishing, 1978 G.Y. Lai and O.F. Kimball, “Aging Behavior of Alloy 800H and Associated Mechanical Property Changes,” Report GA-A15194, General Atomic Company, San Diego, CA, Nov 1978 V. Coppolecchia, J. Bryant, F. Hofmann, and K. Drefahl, Loss of Creep Ductility in Alloy 800H with High Levels of Titanium and Aluminum, Performance of High Temperature Materials in Fluidized Bed Combustion Systems and Process Industries, Conf. Proc., P. Ganesan and R.A. Bradley, Ed., ASM International, 1987, p 201 P.G. Stone, J. Orr, and J.C. Guest, A Status Review of Alloy 800, Proc. British Nuclear Energy Society Conference, University of Reading, Sept 25–26, 1974, S.F. Pugh, Ed., AERE Harwell, UK, 1975, p 15 A.B. Smith, Characterization of Materials for Service at Elevated Temperatures,

pg 407

Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/

31/10/2007 4:07PM Plate # 0

408 / High-Temperature Corrosion and Materials Applications

41.

42.

43.

44. 45. 46.

47. 48.

MPC-7, G.V. Smith, Ed., ASME, 1978, p 159 W.G. Lipscomb, J.R. Crum, and P. Ganesan, Mechanical Properties and Corrosion Resistance of INCONEL Alloy 617 for Reﬁnery Services, Paper No. 259, Corrosion/89, NACE, 1989 T.H. Bassford and T.V. Schill, “A Review of INCONEL Alloy 617 and Its Properties after Long-Time Exposure to Intermediate Temperatures,” Special Metals, Inc., Huntington, WV H. Stahl, G. Smith, and S. Wastiaux, StrainAge Cracking of Alloy 601 Tubes at 600 °C, Practical Failure Analysis, Vol 1 (No. 1), Feb 2001 “INCONEL Alloy 601,” Special Metals Product Literature, Special Metals Company, Huntington, WV M.D. Rowe, Ranking the Resistance of Wrought Superalloys to Strain-Age Cracking, Weld. J., Feb 2006, p 27s R.W. Fawley, M. Prager, J.B. Carlton, and G. Sines, “Recent Studies of Cracking During Postwelding Heat Treatment of Nickel-Base Alloys,” WRC Bulletin No. 150, Welding Research Council, 1970 H.M. Tawancy, Order-Strengthening in a Nickel-Base Superalloy (Hastelloy Alloy S), Metall. Trans., Vol 11A, 1980, p 1764 H.M. Tawancy and A.I. Asphahani, Ordering Behavior and Corrosion Properties of Ni-Mo and Ni-Mo-Cr Alloys, HighTemperature Ordered Intermetallic Alloys (Symposium Proc.), Vol 39, C.C. Koch,

49.

50.

51.

52.

53.

C.T. Liu, and N.S. Stoloff, Ed., Materials Research Society, 1985, p 485 “1997 Findings and Recommendations,” Report to The U.S. Congress and The Secretary of Energy, U.S. Nuclear Waste Technical Review Board, Arlington, VA, April 1998 R.B. Rebak and N.E. Koon, Localized Corrosion Resistance of High Nickel Alloys as Candidate Materials for Nuclear Waste Repository—Effect of Alloy and Weldment Aging at 427 °C for up to 40,000 H, Paper No. 153, Corrosion/98, NACE International, 1998 T.S.E. Summers, M.A. Wall, M. Kumar, S.J. Matthews, and R.B. Rebak, Phase Stability and Mechanical Properties of C-22 Alloy Aged in the Temperature Range 590 to 760 C for 16,000 Hours, Scientiﬁc Basis for Nuclear Waste Management XXII, (Symposium Proc.), Vol 556, D.J. Wronkiewicz and J.H. Lee, Ed., Materials Research Society, 1999, p 919 R.B. Bebak, T.S.E. Summers, and R.M. Carranza, Mechanical Properties, Microstructure and Corrosion Performance of C-22 Alloy Aged at 260 to 800 °C, Scientiﬁc Basis for Nuclear Waste Management XXI (Symposium Proc.), Vol 608, R.W. Smith and D.W. Shoesmith, Ed., Materials Research Society, 2000, p 109 L.M. Pike and D.L. Klarstrom, A New Corrosion-Resistant Ni-Cr-Mo Alloy with High Strength, Paper No. 04239, Corrosion/ 2004, NACE International, 2004

pg 408

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:42PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p409-421 DOI: 10.1361/hcma2007p409

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 15

Molten Salt Corrosion 15.1 Introduction Molten salt technology plays an important role in various industries. In the heat treating industry, molten salts are commonly used as a medium for heat treatment of metals and alloys (e.g., annealing, tempering, hardening, quenching, and cleaning) as well as for surface treatment (e.g., case hardening). In nuclear and solar energy systems, they have been used as a medium for heat transfer and energy storage. Other applications include extraction of aluminum, magnesium, sodium, and other reactive metals; reﬁning of refractory metals; and high-temperature batteries and fuel cells. Table 15.1 summarizes the general applications of molten salt technology in several industries (Ref 1). The containment material, which is in contact with the molten salt, is subject to molten salt corrosion. This chapter reviews the data relevant to the corrosion of containment materials. Although the literature related to studies of corrosion in molten salts is extensive, as can be seen in an annotated bibliography prepared by Janz and Tomkins (Ref 2), corrosion data useful in selecting materials are rather limited and fragmented.

15.2 Corrosion Process Molten salts generally are a good ﬂuxing agent, effectively removing oxide scales from a Table 15.1

metal surface. The corrosion reaction proceeds primarily by oxidation, which is then followed by dissolution of metal oxides in the melt. Oxygen and water vapor in the molten salt thus often accelerate molten salt corrosion. Corrosion can also take place through mass transfer due to thermal gradient in the melt. This mode of corrosion involves dissolution of an alloying element at hot spots and deposition of that alloying element at cooler spots. This can result in severe fouling and plugging in a circulating system. Corrosion is also strongly dependent on temperature and velocity of the salt. Corrosion can take the form of uniform thinning, pitting, or internal or intergranular attack. In general, molten salt corrosion is quite similar to aqueous corrosion. More complete discussion on the mechanisms of molten salt corrosion can be found in Ref 3 to 5.

15.3 Corrosion in Molten Chlorides Chloride salts are widely used in the heat treating industry for annealing and normalizing of steels. These salts are commonly referred to as neutral salt baths. The most common neutral salt baths are barium, sodium, and potassium chlorides, used separately or in combination in the temperature range of 760 to 980 °C (1400 to

General applications of molten salt technology in several industries

Power

Metals/materials

Chemicals

Solar/thermal: collection, storage, transfer Nuclear: homogeneous reactors, reprocessing Batteries

Extraction: refractory metals, actinides, lanthanides, transition, and light metals Processing: heat treatment, annealing, quenching, cleaning, cementation, electroforming Surface ﬁnishing: anodizing, plating

Fuels: cracking, catalysts

Fuel cells

Joining: ﬂuxes and slags for welding, brazing, soldering, and electroslag reﬁning Composites: glasses, ceramics, slags Recycling

Source: Ref 1

Plastics: curing, etching, vulcanizing Pyrolysis: recycling, scrap treatment, hazardous materials disposal Synthesis: organics, gases Special applications: liquid crystals, single-crystal growing, matrix

pg 409

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:42PM Plate # 0

410 / High-Temperature Corrosion and Materials Applications

1800 °F). Compositions of some common neutral salt baths are (Ref 6):

50NaCl-50KCl 50KCl-50Na2CO3 20NaCl-25KCl-55BaCl2 25NaCl-75BaCl2 21NaCl-31BaCl2-48CaCl2

Jackson and LaChance (Ref 6) performed an extensive study on the corrosion of cast Fe-Ni-Cr alloys in the NaCl-KCl-BaCl2 salt bath. They found that alloys suffered intergranular attack more than metal loss. Corrosion data in terms of metal loss and intergranular attack are shown in Fig. 15.1 and 15.2, respectively. The ﬁgures also indicate that resistance to the molten salt (NaClKCl-BaCl2) increases with decreasing chromium and increasing nickel in Fe-Ni-Cr alloys. HW alloy (Fe-12Cr-60Ni) was consistently the best performer among the four commercial cast alloys (HW, HT, HK, and HH alloys) studied. These authors further noted that intergranular attack generally followed grain-boundary carbides.

Fig. 15.1

Thus, lowering carbon from 0.4% to about 0.07% resulted in a threefold improvement. Decreasing grain size also improved alloy resistance to intergranular attack. Five different neutral salt baths were compared for HW, HT, and three Fe-Cr alloys, as shown in Fig. 15.3. In general, the four chloride salt baths were quite similar. The KCl-Na2CO3 salt bath was signiﬁcantly less aggressive than pure chloride baths. It is also interesting to note that Fe-17Cr alloy was better than HW (Fe-12Cr-60Ni) and HT (Fe-15Cr-35Ni) alloys in NaCl-KCl, NaClKCl-BaCl2, and NaCl-BaCl2-CaCl2 salt baths. Lai et al. (Ref 7) evaluated various wrought iron-, nickel-, and cobalt-base alloys in a NaClKCl-BaCl2 salt bath at 840 °C (1550 °F) for 1 month (Fig. 15.4). Surprisingly, two highnickel alloys (alloys 600 and 601) suffered more corrosion attack than stainless steels such as Types 304 and 310. Co-Ni-Cr-W, Fe-Ni-Co-Cr, and Ni-Cr-Fe-Mo alloys performed best. Laboratory testing in a simple salt bath failed to reveal the correlation between alloying elements

Corrosion rates in terms of metal loss for four commercial cast Fe-Ni-Cr alloys in a 20NaCl-25KCl-55BaCl2 salt bath under different conditions of rectification at 870 °C (1600 °F) for 60 h. Source: Ref 6

pg 410

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

Chapter 15:

Molten Salt Corrosion / 411

Fig. 15.2

Corrosion rates in terms of intergranular attack for four commercial cast Fe-Ni-Cr alloys in a 20NaCl-25KCl-55BaCl2 salt bath under different conditions of rectification at 870 °C (1600 °F) for 60 h. Source: Ref 6

Fig. 15.3

Comparison of different neutral salt baths for HW, HT, and Fe-Cr alloys at 870 °C (1600 °F) for 60 h. Source: Ref 6

pg 411

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

pg 412

412 / High-Temperature Corrosion and Materials Applications

and performance. Tests were conducted at 840 °C (1550 °F) for 100 h in a NaCl salt bath with fresh salt for each test run. Results are tabulated in Table 15.2 (Ref 7, 8). Similar to the ﬁeld test results, Co-Ni-Cr-W and Fe-Ni-Co-Cr alloys performed best. Evaluating a eutectic sodium-potassium-magnesium chloride (33NaCl-21.5KCl-45.5MgCl2, mol%) as a possible heat-transfer and energystorage medium in solar thermal energy systems for power generation, Coyle et al. (Ref 9) conducted corrosion tests on various commercial alloys at 900 °C (1650 °F) for 144 and 456 h

(Table 15.3). Fifteen alloys, including iron-, nickel-, and cobalt-base alloys, were evaluated. After 144 h of exposure, specimens of eight alloys were consumed. The remaining seven alloys disintegrated after 456 h of exposure. The authors concluded that the chloride salt was too aggressive to be used at 900 °C (1650 °F). At lower temperatures, molten salts generally become less aggressive. Susskind et al. (Ref 10) conducted corrosion tests at 450 to 500 °C (840 to 930 °F) in molten NaCl-KCl-MgCl2 eutectic and found many alloys resistant to molten salt corrosion (Table 15.4). Low corrosion rates may also be attributed to the vacuum environment used in these tests. Investigating the corrosion behavior of alloys at 400 and 500 °C Table 15.3 Results of corrosion tests in molten eutectic NaCl-KCl-MgCl2 salt at 900 °C (1650 °F) Weight change, mg/cm2 Alloy

Fig. 15.4

Results of a field rack test in a NaCl-KCl-BaCl2 salt bath at 840 °C (1550 °F) for 1 month. Source: Ref 7

Table 15.2 Results of laboratory tests in a NaCl salt bath at 840 °C (1550 °F) for 100 h Alloy

188 25 556 601 Multimet 150 214 304 446 316 X 310 800H 625 RA330 617 230 S RA330 600

304 316 800 800H 556 Nickel 600 214 X N S 230 X-750 R-41 188

144 h

456 h

Disintegrated Disintegrated Disintegrated −310 −250 Disintegrated −280 −120 Disintegrated Disintegrated −400 −300 Disintegrated −150 Disintegrated

… … … Disintegrated Disintegrated … Disintegrated Disintegrated … … Disintegrated Disintegrated … Disintegrated …

N2-(0.1–1H2O)-(1–10O2) was used for the cover gas. Source: Ref 9

Total depth of attack(a), mm (mils)

0.051 (2.0) 0.064 (2.5) 0.066 (2.6) 0.066 (2.6) 0.069 (2.7) 0.076 (3-0) 0.079 (3.1) 0.081 (3.2) 0.081 (3.2) 0.081 (3.2) 0.097 (3.8) 0.107 (4.2) 0.109 (4.3) 0.112 (4.4) 0.117 (4.6) 0.122 (4.8) 0.140 (5.5) 0.168 (6.6) 0.191 (7.5) 0.196 (7.7)

A fresh salt bath was used for each test run; air was used for the cover gas. (a) Mainly intergranular attack; no metal wastage. Source: Ref 7 and 8

Table 15.4 Corrosion of alloys in molten eutectic NaCl-KCl-MgCl2 salt at 450 to 500 °C (840 to 930 °F) with 50 °C temperature differential under vacuum for 1000 h Alloy

1020 2.25Cr-lMo 304 310 316 347 410 430 446 600 N Molybdenum Tantalum Source: Ref 10

Maximum penetration, mm/yr (mpy)

0 0.08 (3) <0.01 (0.4) 0 <0.01 (0.4) 0.12 (4.7) 0.03 (1.0) 0.05 (2) <0.01 (0.4) 0.05 (1.8) 0.05 (2) 0 0.07 (2.9)

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

Chapter 15:

Molten Salt Corrosion / 413

1100 °F). It was not clear what type of cover gas was involved in these tests. The results from both salt mixtures are summarized in Fig. 15.5 and 15.6. Steels and Fe-Cr alloys suffered severe corrosion in both types of salts. Chromium in Fe-Cr alloys and nickel in Fe-Ni alloys improved performance. Fe-Cr-Ni alloys performed signiﬁcantly better than steels and Fe-Cr alloys. Intergranular corrosion is the major corrosion morphology by molten chloride salts. Figures 15.7 and 15.8 show typical intergranular corrosion by molten chloride salt. Figure 15.7 shows the intergranular attack of a Ni-Cr-Fe alloy (alloy 600) coupon welded to a heat treat basket that underwent heat treat cycles involving

(750 and 930 °F) in the molten LiCl-KCl eutectic, which was being considered as an electrolyte for lithium-sulfur fuel cells, Battles et al. (Ref 11) also found many alloys resistant to molten salt (Table 15.5). Tests were conducted in closed quartz crucibles. All the alloys tested showed negligible corrosion rates. Aluminum in the aluminum-clad Type 434 SS sample corroded at a higher rate due to the galvanic couple between aluminum and stainless steel (Ref 11). Takehara and Ueshiba (Ref 12) investigated the corrosion behavior of steel, Fe-Cr, Fe-Ni, and Fe-Cr-Ni alloys in molten 20NaCl-30BaCl250CaCl2 and molten 25LiCl-25ZnCl2-16BaCl224CaCl2-10NaCl at 500 and 600 °C (930 and Table 15.5

pg 413

Corrosion rates of several metals and alloys in molten LiCl-KCl eutectic Corrosion rate, µm/yr (mpy) Annealed samples

Material

304 316 347 430 E-Brite Al-Clad Type 434 Iron (99.999% Fe) Armco electromagnet iron

Sensitized samples(a)

400 °C (750 °F)

500 °C (930 °F)

400 °C (750 °F)

500 °C (930 °F)

2 (0.08) 2 (0.08) … 2 (0.08) 8 (0.32) 130 (5.1) 41 (1.6) 12 (0.47)

6 (0.24) … 2 (0.08) … 6 (0.24) … … …

3 (0.12) … 1 (0.04) … 4 (0.16) … … …

5 (0.2) … 1 (0.04) … 5 (0.2) … … …

Tests were conducted in closed quartz crucibles. (a) Samples were sensitized at 650 °C (1200 °F) for 120 h. Source: Ref 11

Fig. 15.5

Corrosion rates of steel, Fe-Cr, Fe-Ni, and Fe-Cr-Ni alloys in molten 20NaCl-30BaCl2-50CaCl2 at 500 and 600 °C (930 and 1110 °F). Source: Ref 12

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

414 / High-Temperature Corrosion and Materials Applications

a molten KCl salt bath at 870 °C (1600 °F) and a quenching salt bath of molten sodium nitrate and sodium nitrite at 430 °C (800 °F) for 1 month (Ref 13). Figure 15.8 shows the intergranular attack of a heat treat basket made of the same alloy after service for 6 months in the same heat treat cycling operation (Ref 13). Another frequently observed corrosion morphology is internal attack by void formation (Fig. 15.9) (Ref 13). Voids tend to form at grain boundaries as well as in the grain interior (Ref 4). The continuing formation and growth of chromium compounds at the metal surface causes outward migration of chromium and inward migration of vacancies, thus leading to internal void formation (Ref 4).

Cold work may signiﬁcantly affect corrosion of alloys in molten chloride salts. Lai (Ref 14) discovered severe intergranular corrosion at the sheared edge of a Ni-Cr-Fe-Mo alloy (alloy X) coupon after exposure in a molten CaCl2-NaCl salt bath at 570 °C (1050 °F). The sample edge

Fig. 15.6

Corrosion rates of steel, Fe-Cr, Fe-Ni, and Fe-CrNi alloys in molten 25LiCl-25ZnCl2-16BaCl224CaCl2-10NaCl at 500 and 600 °C (930 and 1110 °F). Source: Ref 12

Fig. 15.7

Intergranular attack of a Ni-Cr-Fe alloy coupon welded to a heat treat basket after service for 1 month in a heat treat operation cycling between a molten KCl bath at 870 °C (1600 °F) and a quenching salt bath of molten sodium nitrate-nitrite at 430 °C (800 °F). Source: Ref 13

Fig. 15.8

Intergranular attack of a Ni-Cr-Fe alloy heat treat basket after service for 6 months in the same heat treat cycling operation described in Fig. 15.7. Source: Ref 13

pg 414

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

Chapter 15:

was cold worked during sample shearing prior to exposure. There was no evidence of intergranular corrosion in the area away from the sheared edge (Fig. 15.10). The prior cold work also resulted in a thicker oxide scale during exposure in the molten salt. A Co-Ni-Cr-W alloy (alloy 188) exhibited similar behavior.

15.4 Corrosion in Molten Nitrates/ Nitrides Molten nitrates or nitrate-nitride mixtures are widely used for heat treat salt baths, typically operating from 160 to 590 °C (325 to 1100 °F). They are also used as a medium for heat transfer or energy storage. Molten drawsalt (NaNO3KNO3) is being considered as a heat-transfer and energy-storage medium for a solar central receiver for power generation from solar energy. Numerous studies (Ref 15–21) have been carried out to determine potential candidate containment materials for handling molten drawsalt. Bradshaw and Carling (Ref 22) recently summarized these studies as well as the results of their study (Table 15.6) (Ref 22). The data suggest that, for temperatures up to 630 °C (1170 °F), many alloys are adequate for handling

molten NaNO3-KNO3 salt. Carbon steel and 2.25Cr-1Mo steel exhibited low corrosion rates (<0.13 mm/yr, or <5 mpy) at 460 °C (860 °F). At 500 °C (932 °F), 2.25Cr-1Mo steel exhibited a corrosion rate of about 0.026 mm/yr (1 mpy). Aluminized Cr-Mo steel showed higher resistance, with a corrosion rate of less than 0.004 mm/yr (<0.2 mpy) at 600 °C (1110 °F). Austenitic stainless steels, alloy 800, and alloy 600 were more resistant than carbon steel and Cr-Mo steels. Nickel, however, suffered high corrosion rates. Slusser et al. (Ref 23) evaluated the corrosion behavior of a variety of alloys in molten NaNO3-KNO3 (equimolar volume) salt with an equilibrium nitrite concentration (about 6 to 12 wt%) at 675 °C (1250 °F) for 336 h. A

Fig. 15.10 Fig. 15.9

Corrosion attack consisting of voids in a nickel-base alloy after 2 months at 870 °C (1600 °F) in a molten BaCl2 salt bath. Source: Ref 13

Molten Salt Corrosion / 415

Corrosion of Ni-Cr-Fe-Mo alloy in a molten CaCl2-NaCl salt bath at 570 °C (1050 °F) for 6 months. (a) Little corrosion at the area not cold worked (away from the sheared edge). (b) Intergranular attack at the cold worked area (specimen sheared edge). Source: Ref 14

pg 415

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

pg 416

416 / High-Temperature Corrosion and Materials Applications

constant purge of air in the melt was maintained during testing. Nickel-base alloys were generally much more resistant than iron-base alloys. Increasing nickel content improved alloy corrosion resistance to molten nitratenitrite salt. However, pure nickel suffered rapid corrosion attack. Figure 15.11 shows the corrosion rates of various alloys as a function of nickel content (Ref 23). Silicon-containing alloys, such as RA330 and Nicrofer 3718, performed poorly. A long-term test (1920 h) at 675 °C (1250 °F) was performed on selected alloys, showing corrosion rates similar to those obtained from 336 h exposure tests (Table 15.7). Alloy 800, however, exhibited a higher corrosion rate in the 1920 h test than in the 336 h test. As the temperature was increased to 700 °C (1300 °F), corrosion rates became much higher, particularly for iron-base alloy 800, which suffered an unacceptably high corrosion rates (Table 15.7). Boehme and Bradshaw (Ref 24) attributed the increased corrosion rate with increasing temperature to higher alkali oxide concentration. Slusser et al. (Ref 23) found that adding sodium peroxide (Na2O2) to the salt increased the salt corrosivity.

resistant to molten NaOH (Ref 26–29), particularly low-carbon nickel such as Ni 201 (Ref 30). Gregory et al. (Ref 29) reported corrosion rates of several nickel-base alloys obtained from static tests at 400 to 680 °C (750 to 1256 °F) (Table 15.8). Molybdenum and silicon appear to be detrimental alloying elements in molten NaOH salt. Iron may also be detrimental. Molybdenum and iron were found to be selectively removed from nickel-base alloys with less than 90% nickel, leading to the formation of internal voids (Ref 31). Molten sodium hydroxide becomes increasingly aggressive with increasing temperature. Coyle et al. (Ref 9) evaluated a variety of alloys for a possible containment material for molten sodium hydroxide operating at 900 °C (1650 °F) for a solar power generation system. Test results are tabulated in Table 15.9. Many

15.5 Corrosion in Molten Sodium Hydroxide (Caustic Soda) The reaction of metals with molten sodium hydroxide (NaOH) leads to metal oxide, sodium oxide, and hydrogen (Ref 25). Nickel is most Table 15.6 Corrosion rates of selected metals and alloys in molten NaNO3-KNO3 Alloy

Carbon steel 2.25Cr-lMo 9Cr-lMo Aluminized Cr-Mo steel 12Cr steel 304SS 316SS 800

600 Nickel Titanium Aluminum Source: Ref 22

Temperature °C (°F)

Corrosion rate, mm/yr (mpy)

460 (860) 460 (860) 500 (932) 550 (1020) 600 (1110) 600 (1110) 600 (1110) 600 (1110) 600 (1110) 630 (1170) 565 (1050) 600 (1110) 630 (1170) 600 (1110) 630 (1170) 565 (1050) 565 (1050) 565 (1050)

0.120 (4.7) 0.101 (4.0) 0.026 (1.0) 0.006 (0.2) 0.023 (0.9) <0.004 (0.2) 0.022 (0.9) 0.012 (0.5) 0.007–0.010 (0.3–0.4) 0.106 (4.2) 0.005 (0.2) 0.006–0.01 (0.2–0.4) 0.075 (3.0) 0.007–0.01 (0.3–0.4) 0.106 (4.2) >0.5 (20) 0.04 (1.6) <0.004 (0.2)

Fig. 15.11

Corrosion rates of various alloys as a function of nickel content in the alloy tested in molten NaNO3-KNO3 salt at 675 °C (1250 °F). Source: Ref 23

Table 15.7 Corrosion rates of selected alloys at 675 and 700 °C (1250 and 1300 °F) in sodium-potassium nitrate-nitrite salt Corrosion rate, mm/yr (mpy)

Alloy 214 600 N 601 800 Source: Ref 23

675 °C (1250 °F) 1920 h

700 °C (1300 °F) 720 h

0.41 (16) 0.25 (10) 0.23 (9.1) 0.48 (19) 1.85 (73)

0.53 (21) 0.99 (39) 1.22 (48) 1.25 (49) 6.6 (259)

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

Chapter 15:

pg 417

Molten Salt Corrosion / 417

conditions to about 8 mm/yr (320 mpy) at a rotational speed of 600 rpm (Ref 32). Metals or alloys in molten sodium hydroxide are susceptible to mass transfer due to thermal gradients in the melt. This causes corrosion in the hot zone, and potential tube plugging in the cold zone, of a circulating system. For example, 6.35 mm (0.25 in.) nickel tubing was plugged after 5000 h at 440 to 480 °C (830 to 900 °F), and after 50 h at 690 to 730 °C (1280 to 1350 °F) (Ref 27).

iron-, nickel-, and cobalt-base alloys disintegrated in 84 h. Samples of the alloys that survived the 84 h exposure test were severely corroded. Scales that formed on these samples were reportedly cracked and spalled. The weightgain or weight-loss data of surviving samples were no longer indicative of alloy performance ranking. No metallographic examination was performed on these samples. The authors concluded that no further studies on molten sodium hydroxide were necessary, because the salt was too aggressive to metallic materials operating at 900 °C (1650 °F). The marked inﬂuence of temperature on the corrosiveness of molten sodium hydroxide is also demonstrated by the results shown in Table 15.10 (Ref 26). Corrosion of metals and alloys in molten NaOH depends strongly on the velocity of the salt. Gregory et al. (Ref 32) showed that corrosion of nickel under dynamic conditions was enhanced by as much as several times at 540 °C (1000 °F) and higher. The corrosion rate for nickel at 680 °C (1250 °F), for example, varied from about 1 mm/yr (40 mpy) under static

15.6 Corrosion in Molten Fluorides Corrosion of alloys in molten ﬂuoride salts has been extensively studied for nuclear reactor applications. The molten salt nuclear reactor uses a LiF-BeF2 base salt as a fuel salt, containing various amounts of UF4, ThF4, and ZrF4 (Ref 33). The reactor coolant salt is a NaBF4-NaF mixture (Ref 33). A nickel-base alloy, Hastelloy alloy N, has proved to be the most corrosion resistant in molten ﬂuoride salts (Ref 34). The

Table 15.8 Corrosion rates of selected nickel-base alloys obtained from static tests in molten sodium hydroxide Corrosion rate, mm/yr (mpy) Alloy

Ni-201 C D 400 600 301SS 75

400 °C (750 °F)

500 °C (930 °F)

580 °C (1080 °F)

680 °C (1260 °F)

0.023 (0.9) … 0.018 (0.7) 0.046 (1.8) 0.028 (1.1) 0.043 (1.7) 0.028 (1.1)

0.033 (1.3) 2.54 (100) 0.056 (2.2) 0.13 (5.1) 0.06 (2.4) 0.08 (3.2) 0.36 (14.3)

0.06 (2.5) (a) 0.25 (9.9) 0.45 (17.6) 0.13 (5.1) 0.26 (10.4) 0.53 (20.8)

0.96 (37.8) … (a) … 1.69 (66.4) 1.03 (40.7) 1.21 (47.6)

(a) Severe corrosion. Source: Ref 29

Table 15.10 Corrosion of various metals and alloys in molten sodium hydroxide Table 15.9 Results of corrosion tests in molten sodium hydroxide at 900 °C (1650 °F) for 84 h Alloy

Weight change, mg/cm2

304 316 800 800H 556 Nickel 600 214 X N S 230 X-750 R-41 188

Disintegrated Disintegrated +60 +65 Disintegrated −50 −27 +160 +22 Fractured 4 Disintegrated +35 Disintegrated Disintegrated

N2-(0.1–1H2O)-(1–10O2) was used for the cover gas. Source: Ref 9

Weight change rate(a), mg/cm2-day Material

Ni-201 Copper Chromium Aluminum Silver Gold Platinum 70Au-30Pt Palladium Colmonoy No. 5 Colmonoy No. 6 Colmonoy No. 9 Chromel P Zirconium

538 °C (1000 °F)

816 °C (1500 °F)

+0.12 +1.54 −0.12 −0.34 +0.02 −0.33 −0.32 +0.36 +6.59 +0.63 −0.48 −0.23 +0.09 +0.56

+1.7 −2.2 −40 … +2.9 Broke Broke Broke +175 +70 −72 Dissolved in 24 h Dissolved in 24 h …

(a) Based on 24 h exposure tests. Source: Ref 26

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

pg 418

418 / High-Temperature Corrosion and Materials Applications

alloy was the primary containment material for a molten salt test reactor successfully operated from 1965 to 1969 (Ref 35). Koger (Ref 33) reported a corrosion rate of less than 0.0025 mm/yr (0.1 mpy) at 704 °C (1300 °F) in the LiF-BeF2 base salt (fuel salt) and about 0.015 mm/yr (0.6 mpy) at 607 °C (1125 °F) in the NaBF4-NaF coolant salt for alloy N. Iwamoto et al. (Ref 36) performed corrosion tests in eutectic LiF-NaF-KF salt in a test loop with a 750 °C (1380 °F) hot leg and a 685 °C (1265 °F) cold leg for 500 h. Alloy N exhibited only 2.06 mg/cm2 of maximum weight loss at the hot leg. At lower temperatures, austenitic stainless steels showed good performance. Type 316 exhibited 0.015 mm/yr (0.59 mpy) at 650 °C (1200 °F) in 66LiF-34BeF2 (mol%), and about 0.002 mm/yr (0.08 mpy) at 530 °C (986 °F) in 22LiF-31LiCl-47LiBr (mol%) (Ref 37). In a LiFBeF2 fuel salt containing UF4, ThF4, and ZrF4, Type 304 suffered a corrosion rate of only about 0.028 mm/yr (1.1 mpy) at 690 °C (1270 °F) (Ref 33). Corrosion can become more aggressive as temperature increases. It is particularly severe for stainless steels because of tube-plugging problems due to mass transfer. Adamson et al. (Ref 38) conducted corrosion tests in a thermal convection loop involving 43.5KF-10.9NaF44.5LiF-1.1UF4 (mol%) with an 815 °C (1500 °F) hot leg and a 704 °C (1300 °F) cold leg. Types 410, 430, 316, 310, and 347 suffered severe tube-plugging problems at the cold leg within short test durations. Nickel and nickelbase alloys, on the other hand, showed no plugging even after 500 h of testing. However, these alloys suffered corrosion at the hot leg after 500 h of exposure. Alloy 600 suffered internal attack consisting of voids about 0.30 to 0.38 mm (12 to 15 mils) deep. Nimonic alloy 75 suffered intergranular pitting about 0.20 to 0.33 mm (8 to 13 mils) deep, and nickel suffered even metal removal of about 0.23 mm (9 mils). Misra and Whittenberger (Ref 39) reported corrosion data for a variety of commercial alloys in molten LiF-19.5CaF2, which was being considered for a heat-storage medium in an advanced solar space power system, at 797 °C (1467 °F) for 500 h. The tests were conducted in alumina crucibles with argon as a cover gas. Results are tabulated in Table 15.11. For nickel-base alloys, chromium was detrimental. No inﬂuence of chromium, however, was noted on iron-base alloys.

Molten ﬂuorides are generally used in a closed system under vacuum or an inert atmosphere. However, hydrogen ﬂuoride may be present in the system, resulting in increased corrosion rates. Moisture, a common impurity in ﬂuoride salts, can react with ﬂuorides to produce gaseous HF (Ref 39), some of which may dissolve in the melt. Corrosion by HF will also be involved, leading to production of hydrogen (Ref 39). Therefore, it is important to reduce the moisture level in the salt to reduce corrosion attack (Ref 39).

15.7 Corrosion in Molten Carbonates Molten carbonates are generally less corrosive than molten chlorides or hydroxides. Coyle et al. (Ref 9) evaluated three different molten salts for a possible heat-transfer and energy-storage medium capable of operating 900 °C (1650 °F) for a solar power generation system. Both the eutectic sodium-potassium-magnesium chloride (33NaCl-21.5KCl-45.5MgCl2, mol%) and the sodium hydroxide were found to be too corrosive for many commercial alloys. The eutectic sodium-potassium carbonate (58Na2CO342K2CO3, mol%), on the other hand, showed promise because of the much lower corrosion rates exhibited by many commercial alloys. Corrosion data generated in molten carbonate salt at 900 °C (1650 °F) for 504 h are summarized in Table 15.12. The best performer was Ni-Cr-Fe-Mo alloy (alloy X), followed by Ni-Cr-Fe-Al alloy (alloy 214) and CoNi-Cr-W alloy (alloy 188). The Ni-Cr-Mo Table 15.11 Results of corrosion tests in LiF-19.5CaF2 at 797 °C (1467 °F) for 500 h Depth of attack, µm (mils) Alloy

Mild steel 304 310 316 RA330 B N S X 600 718 75 25 188

General(a)

Grain boundary(b)

… … … … … 30 (1.2) 15 (0.6) 90 (3.5) … 90 (3.5) 45 (1.8) 30 (1.2) … …

155 (6.1) 185 (7.3) 130 (5.1) 165 (6.5) 270 (10.6) … 15 (0.6) … 140 (5.5) 30 (1.2) 120 (4.7) 135 (5.3) 95 (3.7) 105 (4.1)

Tests were conducted in alumina crucibles under argon. (a) Intragranular voids near surface. (b) Intergranular voids. Source: Ref 39

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

Chapter 15:

alloy (alloy S) was severely corroded. No evidence of a systematic trend for the correlation between alloying elements and performance is noted. A signiﬁcant difference in corrosion was observed between two samples of alloy 800 obtained from different suppliers. However, two samples of alloy 600 showed good agreement. At lower temperatures, molten carbonate corrosion generally becomes less severe. In molten eutectic alkali metal carbonate, Grantham et al. (Ref 40) found that many commercial alloys, including Types 304L, 310, and 347, and alloys 600, C, N, X, and 25, exhibited low corrosion rates (about 0.01mg/cm2/h or less) at 600 °C (1110 °F). Nonoxidizing N2 was used for the cover gas in their corrosion tests, which may also have contributed to low corrosion rates. The temperature dependence of corrosion rate can also be seen in the results generated by Grantham and Ferry (Ref 41) in the eutectic Li2CO3-Na2CO3-K2CO3 (about equal weight) at 500, 600, and 700 °C (930, 1110, and 1290 °F). The cover gas in this case was not reported. Corrosion rates were found to be less than 0.025, 0.025, and 2.54 mm/yr (1, 1, and 100 mpy) at 500, 600, and 700 °C (930, 1110, and 1290 °F), respectively.

15.8 Summary The corrosion behavior of alloys in molten chlorides, nitrates/nitrites, sodium hydroxide, ﬂuorides, and carbonates was reviewed. In Table 15.12 Results of corrosion tests in molten eutectic sodium-potassium carbonate at 900 °C (1650 °F) for 504 h Alloy

X 214 188 556 X-750 600(b) 600(b) R-41 N 304SS 316SS 230 Nickel 800(b) 800(b) S

Total depth of attack(a), mm (mils)

0.12 (4.7) 0.19 (7.5) 0.22 (8.7) 0.26 (10.2) 0.27 (10.6) 0.34 (13.4) 0.44 (17.3) 0.42 (16.5) 0.51 (20.1) 0.54 (21.3) 0.63 (24.8) 0.77 (30.3) >0.30 (11.8) 0.25 (9.8) >0.8 (31.5) >1.43 (56.3)

N2-0.1CO2-(1–10O2) was used for the cover gas. (a) All alloys showed metal loss only except nickel, which suffered 0.2 mm (7.9 mils) metal loss and more than 0.11 mm (4.3 mils) intergranular attack. (b) Two samples from two different suppliers. Source: Ref 9

Molten Salt Corrosion / 419

general, corrosion rates of metals and alloys are strongly dependent on temperature and can generally be reduced by decreasing the temperature. Reducing oxidizing impurities, such as oxygen and water vapor, in the melt can also signiﬁcantly reduce the corrosiveness of the molten salt. Thermal gradients in the melt, in the case of circulating systems, may cause dissolution of an alloying element at the hot leg and deposition of that element at the cold leg, leading to potential tube-plugging problems. Corrosion data generated from different chloride salts at temperatures ranging from 400 to 900 °C (750 to 1650 °F) for various cast and wrought alloys were presented. The data are rather limited and fail to show a deﬁnitive correlation between alloying elements and performance. Intergranular attack is the major corrosion morphology in molten chlorides. Cold work may signiﬁcantly accelerate intergranular attack. If the cold work effect is conﬁrmed to be a general phenomenon in molten chloride salts, annealing of fabricated components prior to service is recommended. The corrosion behavior of commercial alloys in molten NaNO3-KNO3 is reasonably well understood. For applications at temperatures up to 630 °C (1170 °F), many commercial alloys, including austenitic stainless steels, are capable of handling the molten salt. At higher temperatures, nickel-base alloys are preferred because of the increased salt corrosivity. The resistance of alloys to molten NaNO3-KNO3 improves with increasing nickel content. Nickel, however, performed poorly. Corrosion data related to molten sodium hydroxide (caustic soda) are rather limited. At 900 °C (1650 °F), the salt is probably too corrosive for metallic materials to handle. Metals and alloys may have adequate resistance to molten sodium hydroxide at 680 °C (1260 °F) and lower temperatures. Nickel, particularly low-carbon nickel such as Ni201, is most resistant to the molten salt. Corrosion rates can be signiﬁcantly increased under dynamic conditions. Corrosion of alloys in molten ﬂuorides has been extensively studied for nuclear reactor applications. A nickel-base alloy, Hastelloy alloy N, has been found to be most suitable in molten ﬂuoride salts. Since molten ﬂuorides are used in a circulating system with temperature differentials, mass transfer induced corrosion due to thermal gradients in the loop becomes an important issue. It creates the potential for tube fouling

pg 419

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

420 / High-Temperature Corrosion and Materials Applications

and plugging at the cold leg of the loop. Stainless steels are generally worse than nickel and nickel-base alloys in terms of tube-plugging problems. Molten carbonates are generally less corrosive than molten chlorides or hydroxides. Some of the alloys that perform best in molten eutectic Na2CO3-K2CO3 at high temperatures such as 900 °C (1650 °F) include Ni-Cr-Fe-Mo alloy (alloy X), Ni-Cr-Fe-Al alloy (alloy 214), and Co-Ni-Cr-W alloy (alloy 188).

16.

17.

18. REFERENCES

1. D.G. Lovering, in Molten Salt Technology, D.G. Lovering, Ed., Plenum Press, 1982, p 1 2. G.J. Janz and R.P.T. Tomkins, Corrosion, Vol 35 (No. 11), 1979, p 485 3. J.W. Koger, in Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 51 4. J.W. Koger and S.L. Pohlman, in Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 88 5. A. Rahmel, in Molten Salt Technology, D.G. Lovering, Ed., Plenum Press, 1982, p 265 6. J.H. Jackson and M.H. LaChance, Trans. ASM, Vol 46, 1954, p 157 7. G.Y. Lai, M.F. Rothman, and D.E. Fluck, Paper No. 14, Corrosion/85, NACE, 1985 8. G.Y. Lai, unpublished results, Haynes International, Inc., 1986 9. R.T. Coyle, T.M. Thomas, and G.Y. Lai, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 627 10. H. Susskind, F.B. Hill, L. Green, S. Kalish, L. Kukacka, W.E. McNulty, and E. Wirsing, Chem. Eng. Prog., Vol 56 (No. 3), 1960, p 57 11. J.E. Battles, F.C. Mrazek, W.D. Tuohig, and K.M. Myles, in Corrosion Problems in Energy Conversion and Generation, C.S. Tedmon, Jr., Ed., The Electrochemical Society, 1974, p 20 12. K. Takehara and T. Ueshiba, J. Soc. Mater. Sci. Jpn., Vol 179, 1968, p 755 13. S.K. Srivastava, unpublished results, Haynes International, Inc., 1989 14. G.Y. Lai, unpublished results, Haynes International, Inc., 1989 15. R.W. Bradshaw, “Thermal Convection Loop Corrosion Tests of Type 316SS and Alloy 800 in Molten Nitrate Salts,” SAND

19.

20. 21.

22.

23. 24. 25. 26.

27.

28.

81-8210, Sandia Laboratory, Livermore, CA, Feb 1982 P.F. Tortorelli and J.E. DeVan, “Thermal Convection Loop Study of the Corrosion of Fe-Ni-Cr Alloys by Molten NaNO3-KNO3,” ORNL TM-8298, Oak Ridge National Laboratory, Oak Ridge, TN, Dec 1982 R.W. Bradshaw, “A Thermal Convection Loop Study of Corrosion of Alloy 800 in Molten NaNO3-KNO3,” SAND 82-8911, Sandia Laboratory, Livermore, CA, Jan 1983 R.W. Bradshaw, “Kinetic Oxidation and Elemental Depletion of Austenitic and Ferritic Steels in Molten Nitrate Salt,” SAND 87-8011, Sandia Laboratory, Livermore, CA, 1987 R.W. Bradshaw, “Oxidation of ChromiumMolybdenum Steels by Molten Sodium Nitrate-Potassium Nitrate,” SAND 87-8012, Sandia Laboratory, Livermore, CA 1987 R.W. Carling, R.W. Bradshaw, and R.W. Mar, J. Mater. Energy Sys., Vol 4 (No. 4), 1983, p 229 R.W. Bradshaw, “Oxidation and Chromium Depletion of Alloy 800 and Type 316SS in Molten NaNO3-KNO3 at Temperatures above 600 °C,” SAND 86-9009, Sandia Laboratory, Livermore, CA, Jan 1987 R.W. Bradshaw and R.W. Carling, “A Review of the Chemical and Physical Properties of Molten Alkali Nitrate Salts and Their Effect on Materials Used for Solar Central Receivers,” SAND 87-8005, Sandia Laboratory, Livermore, CA, April 1987 J.W. Slusser, J.B. Titcomb, M.T. Heffelﬁnger, and B.R. Dunbobbin, J. Met., July 1985, p 24 D.R. Boehme and R.W. Bradshaw, High Temp. Sci., Vol 18, 1984, p 39 G.P. Smith, “Corrosion of Materials in Fused Hydroxides,” USAEC Report ORNL-2048, March 1956 C.M. Craighead, L.A. Smith, and R.I. Jaffee, “Screening Tests on Metals and Alloys in Contact with Sodium Hydroxide at 1000 and 1500 °F,” USAEC Report BMI-705, Nov 1951 E.M. Simmons, N.E. Miller, J.H. Stang, and C. Weaver, “Corrosion and Components Studies on Systems Containing Fused NaOH,” USAEC Report BMI-1118, July 1956 R.A. Lad and S.L. Simon, A Study of Corrosion and Mass Transfer of Nickel by

pg 420

Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/

26/10/2007 12:43PM Plate # 0

Chapter 15:

29.

30.

31. 32.

33. 34. 35.

Molten Sodium Hydroxide, Corrosion, Vol 10 (No. 12), 1954, p 435 J.N. Gregory, N. Hodge, and J.V.G. Iredale, “The Static Corrosion of Nickel and Other Materials in Molten Caustic Soda,” AEREC/M-272, Atomic Energy Research Establishment, Harwell, U.K., March 1956 R.R. Miller, “Thermal Properties of Sodium Hydroxide and Lithium Metal,” Quarterly Progress Report May 1–Aug 1, 1952, NRL3230-201/52, 1952 G.P. Smith and E.E. Hoffman, Corrosion, Vol 13, 1957, p 627t J.N. Gregory, N. Hodge, and J.V.G. Iredale, “The Corrosion and Erosion of Nickel by Molten Caustic Soda and Sodium Uranate Suspensions under Dynamic Conditions,” AERE Report C/M 273, Atomic Energy Research Establishment, Harwell, U.K., March 1956 J.W. Koger, Corrosion, Vol 29 (No. 3), 1973, p 115 J.W. Koger, Corrosion, Vol 30 (No. 4), 1974, p 125 J.R. Keiser, D.L. Manning, and R.E. Clausing, in Proc. Int. Symp. Molten Salts,

36. 37. 38.

39.

40.

41.

Molten Salt Corrosion / 421

J.P. Pemsler et al., Ed., The Electrochemical Society, 1976, p 315 N. Iwamoto, Y. Makino, K. Furukawa, Y. Katoh, and H. Katsuta, Trans. JWRI, Vol 9 (No. 2), 1980, p 117 J.R. Keiser, J.H. DeVan, and E.J. Lawrence, J. Nucl. Mater., Vol 85/86, 1979, p 295 G.M. Adamson, R.S. Crouse, and W.D. Manly, “Interim Report on Corrosion by Alkali-Metal Fluorides,” ORNL-2337, Oak Ridge National Laboratory, Oak Ridge, TN, 1959 A.K. Misra and J.D. Whittenberger, “Fluoride Salts and Container materials for Thermal Energy Storage Applications in Temperature Range 973 to 1400 °K,” NASA Tech. Memo. 89913, NASA Lewis Research Center, Cleveland, OH, 1987 L.F. Grantham, P.H. Shaw, and R.D. Oldenkamp, in High Temperature Metallic Corrosion of Sulfur and Its Compounds, Z.A. Foroulis, Ed., The Electrochemical Society, 1970, p 253 L.F. Grantham and P.B. Ferry, in Proc. Int. Symp. Molten Salts, J.P. Pemsler et al. Eds., The Electrochemical Society, 1976, p 270

pg 421

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p423-435 DOI: 10.1361/hcma2007p423

31/10/2007 12:52PM Plate # 0

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 16

Liquid Metal Corrosion and Embrittlement

Liquid metals are frequently used as a heattransfer medium because of their excellent heat-transfer properties, such as high thermal conductivities, high heat capacities, and low vapor pressures. Various metals have been investigated for use as coolants in nuclear reactors. Sodium, for example, has been used as a coolant in fast breeder nuclear reactors. Most corrosion studies of molten metals have been carried out in conjunction with nuclear reactor applications (Ref 1–4). Other applications of liquid metals include heat treat baths (e.g., molten lead) and power generation (Ref 5). The containment material, which is in contact with the molten metal, is subject to molten metal corrosion. This chapter reviews data relevant to the corrosion of containment materials. Discussion of liquid metal embrittlement is also included.

major alloying elements) with the molten metal to form an intermetallic compound. This requires some solubility of the liquid metal in the containment metal. When a molten metal is used as a heat-transfer medium in a loop, mass transfer due to thermal gradients in the melt can be a critical issue (Ref 6). It involves dissolution of an alloying element in hot zones and deposition of that alloying element in cold zones, resulting in tube plugging. The corrosion can also involve impurity and interstitial interactions, which are generally associated with light elements such as carbon, nitrogen, and oxygen. This type of reaction dictates the corrosion process when the solubilities of the major alloying elements in the molten metal are low. An example is carburization-decarburization behavior of alloys in a sodium loop. More complete discussions on the mechanisms of molten metal corrosion can be found in articles by numerous authors (Ref 1, 2, 7–10).

16.2 Corrosion Process

16.3 Corrosion in Molten Aluminum

Molten metal corrosion of a containment metal is most often related to its solubility in the molten metal. This type of corrosion is simply a dissolution-type attack. A containment metal with higher solubility in the molten metal generally exhibits a higher corrosion rate. In the case of an alloy, the solubilities of the major alloying elements could dictate the corrosion rate. The solubility of an alloying element in a molten metal typically increases with increasing temperature. As the temperature increases, the diffusion rate also increases. Thus, the alloy corrosion rate increases with increasing temperature. Corrosion by molten metal can also proceed by alloying of the containment metal (or its

Aluminum melts at 660 °C (1220 °F). Iron, nickel, and cobalt, along with their alloys, are readily attacked by molten aluminum. Extremely high corrosion rates of iron-, nickel-, and cobaltbase alloys in molten aluminum are illustrated by the laboratory test results shown in Table 16.1 (Ref 11). Samples of carbon steel and ironand nickel-base alloys were consumed in 4 h at 760 °C (1400 °F). Cobalt-base alloys, which appeared to be better than iron- and nickel-base alloys, were corroding at rates too high to be considered for containment materials. In addition, titanium, although exhibiting a corrosion rate lower than iron-, nickel-, and cobalt-base alloys, should not be considered for use as a containment material because of its rapid corrosion rate.

16.1 Introduction

pg 423

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

pg 424

424 / High-Temperature Corrosion and Materials Applications

16.4 Corrosion in Molten Zinc Zinc melts at 420 °C (790 °F). Molten zinc is widely used in the hot dip galvanizing process to coat steel for corrosion protection. Galvanizing tanks, along with baskets, ﬁxtures, and other accessories, require materials resistant to molten zinc corrosion. Nickel and high-nickel alloys react readily with molten zinc by direct alloying. This is illustrated in Fig. 16.1, which shows a nickel-base alloy coupon after immersion in molten zinc at 455 °C (850 °F). Iron- and cobalt-base alloys are generally corroded by dissolution, even those

containing up to 33% Ni, such as alloy 800H (Fig. 16.2). The results of static tests in molten zinc for selected iron-, nickel-, and cobalt-base alloys are summarized in Table 16.2. Nickel-base alloys suffered the worst attack, followed by austenitic stainless steels, Fe-Ni-Cr alloys, and Fe-Cr alloys. Cobalt-base alloys generally performed better. However, an Fe-Ni-Co-Cr alloy

Table 16.1 Results of static immersion tests in molten aluminum at 760 °C (1400 °F) for 4 h Maximum depth of corrosion attack, mm (mils)

Alloy

Titanium 6B 188 150 556 X 671 Carbon steel

0.22 (8.5) 0.43 (16.8) 0.51 (20.2) 0.73 (28.9) >0.5 (20.6)(a) >0.6 (23.8)(a) >0.7 (26.3)(a) >1.6 (63.1)(a)

(a)

200 µm

(b)

200 µm

(c)

200 µm

(a) Sample was consumed. Source: Ref 11

Fig. 16.2 Fig. 16.1 Ref 11

Nickel-base alloy coupon after immersion testing in molten zinc at 455 °C (850 °F) for 50 h. Source:

Sample cross sections for (a) alloy 556, (b) alloy 800H, and (c) Type 446 after immersion testing in molten zinc at 455 °C (850 °F) for 50 h. The top edge of each photograph represents the original surface of the sample prior to immersion testing. Uniform dissolution of metal from the sample surface is noted for all three alloys. Source: Ref 12

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

Chapter 16:

(556 alloy) performed as well as cobalt-base alloys. Intergranular cracking was observed for some alloys in contact with molten zinc. Field testing in a galvanizing tank, with test coupons immersed in molten zinc at approximately 455 °C (850 °F) for 19 runs (8 h immersion per run) for a total of 152 h, revealed intergranular cracking for coupons of alloy 25 (Co-Ni-Cr-W alloy) and Type 316, as shown in Fig. 16.3 (Ref 12). The other two alloys in the same ﬁeld test (cobaltbase alloy 188 and Fe-Ni-Co-Cr alloy 556) did not suffer cracking. The metallurgical factors responsible for the cracking of alloy 25 and Type 316 are unclear. Liquid metal embrittlement (LME) is discussed in Section 16.9. Molten zinc corrosion is expected to become more severe at higher temperatures. However, very little data, particularly at 500 °C (930 °F) and higher, have been reported in the literature.

31/10/2007 12:52PM Plate # 0

pg 425

Liquid Metal Corrosion and Embrittlement / 425

(Ref 15). For example, cast iron centrifugal pumps are used to pump liquid lead (Ref 15). Ali-Khan (Ref 16) performed extensive corrosion tests in molten lead. The results of his static corrosion tests at temperatures from 575 to 750 °C (1070 to 1380 °F) for chromium steels (7 to 17Cr) and austenitic stainless steels (18Cr-8 to 16Ni) are shown in Tables 16.3 and 16.4, respectively. Both exhibited similar performance, with corrosion rates of less than 0.15 mm (6 mils) after 3250 h of exposure. Wilkinson et al. (Ref 17) investigated molten lead corrosion at a much higher temperature. These authors performed static tests at 1000 °C (1830 °F) for up to 408 h. Their test results (Table 16.5) showed that the 17Cr steel (Type 430) was the most resistant. A sample of cast iron, on the other hand, completely dissolved in 408 h.

16.5 Corrosion in Molten Lead Lead melts at 327 °C (620 °F). Nickel and nickel-base alloys generally have poor resistance to molten lead corrosion (Ref 13, 14). The solubility of nickel in molten lead is higher than that of iron. Cast iron, steels, and stainless steels are commonly used for handling molten lead Table 16.2 Results of static immersion tests in molten zinc at 455 °C (850 °F) for 50 h Alloy

Depth of corrosion attack, mm (mils)

556 25 188 446SS 800H 304SS 625 X

0.04 (1.6) 0.06 (2.3) 0.06 (2.5) 0.24 (9.3) 0.28 (11.0) 0.36 (14.1) >0.61 (24.0)(a) >0.61 (24.0)(a)

(a)

100 µm

(b)

100 µm

(a) Complete alloying. Source: Ref 11

Table 16.3 Results of static corrosion tests at various temperatures for 3250 h in molten lead for chromium steels Depth of attack, μm (mils) Alloy

7Cr-0.8Al 13Cr 13Cr-1.0Al 13Cr-1.0Mo 17Cr

575 °C (1067 °F)

650 °C (1200 °F)

750 °C (1382 °F)

25 (1.0) 10 (0.4) 10 (0.4) 10 (0.4) (a)

(a) 15 (0.6) 15 (0.6) 30 (1.2) (a)

30 (1.2) 100 (3.9) … 150 (5.9) 35 (1.4)

(a) Little or no corrosion. Source: Ref 16

Fig. 16.3

Intergranular cracking for (a) alloy 25 and (b) Type 316 coupons after field testing in a hot dip galvanizing tank at 455 °C (850 °F) for a total of 152 h (19 runs). Source: Ref 12

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

pg 426

426 / High-Temperature Corrosion and Materials Applications

The corrosion behavior of both chromium steels and austenitic stainless steels in a thermal convection loop was studied by Ali-Khan (Ref 16). The results are shown in Tables 16.6 and 16.7. Compared with tests under static conditions (Tables 16.3 and 16.4), alloys under thermal convection conditions corroded at faster rates. Tolson and Taboada (Ref 18) also reported corrosion data generated from a thermal convection loop (Table 16.8). The results of the Brookhaven loop, which used magnesium and zirconium as inhibitors in the melt, are included in the table for comparison. It appears that the inhibitors signiﬁcantly reduced corrosion by

Table 16.6 Results of corrosion tests in a molten lead thermal convection loop for 1002 h at 600 °C (1112 °F, hot leg) with a temperature differential of 232 °C (450 °F) for chromium steels

Table 16.4 Results of static corrosion tests at various temperatures for 3250 h in molten lead for austenitic stainless steels

Alloy

Depth of attack, μm (mils) Alloy

18Cr-9Ni 17Cr-8Ni 18Cr-8Ni 18Cr-10Ni 18Cr-11Ni-2Mo 18Cr-12Ni-2.5Mo 18Cr-12Ni-2Mo 18Cr-16Ni-3.5Mo 18Cr-10Ni

575 °C (1067 °F)

650 °C (1200 °F)

750 °C (1382 °F)

40 (1.6) 60–90 (2.4–3.5) 10 (0.4) 40 (1.6) 55 (2.2) 25 (1.0) 40 (1.6) 80 (3.2) 110 (4.3)

10 (0.4) 15 (0.6) 20 (0.8) 25 (1.0) 15 (0.6) (a) 10 (0.4) 35 (1.4) 35 (1.4)

25 (1.0) 10 (0.4) (a) 150 (5.9) 10 (0.4) 70 (2.8) 15 (0.6) 20 (0.8) 20 (0.8)

(a) Little or no corrosion. Source: Ref 16

Table 16.5 Corrosion of several alloys in molten lead at 1000 °C (1830 °F) for up to 408 h Alloy

Corrosion rate, mm/yr (mpy)

Cast iron ARMCO iron 1020 steel 430SS 302SS 347SS

Complete dissolution in 408 h 4.0 (157) 2.5 (100) 0.7 (27) 10.0 (393) 9.8 (387)

Source: Ref 17

Table 16.8

molten lead. Croloy 2.25Cr steel revealed no detectable corrosion attack in the Brookhaven loop after 27,765 h of exposure, while the same alloy showed 0.2 mm (8 mils) of attack in the ORNL loop containing no inhibitors, although the ORNL loop was operated 26 °C (47 °F) higher. The beneﬁcial effect of inhibitors has been conﬁrmed by Asher et al. (Ref 19). It is believed that inhibitors are instrumental in the formation of a protective ﬁlm, which reduces the solubility of the containment metal in molten lead (Ref 20).

Depth of attack, µm (mils)

7Cr-0.8Al 13Cr steel 13Cr-1.0Al 13Cr-1.0Mo 17Cr

130–180 (5.1–7.1) 273–300 (11–12) 187–200 (7.4–7.9) 143–240 (5.6–9.5) 130–140 (5.1–5.5)

Source: Ref 16

Table 16.7 Results of corrosion tests in a molten lead thermal convection loop for 1008 h at 600 °C (1112 °F, hot leg) with a temperature differential of 151 °C (304 °F) for austenitic stainless steels Alloy

Depth of attack, µm (mils)

18Cr-9Ni 17Cr-8Ni 18Cr-8Ni 18Cr-10Ni 18Cr-11Ni-2Mo 18Cr-12Ni-2.5Mo 18Cr-12Ni-2Mo 18Cr-16Ni-3.5Mo 18Cr-10Ni

80–140 (3.2–5.5) 53–80 (2.1–3.2) 53–80 (2.1–3.2) 62–100 (2.4–3.9) 80–140 (3.2–5.5) 80 (3.2) 130–200 (5.1–7.9) 120–140 (4.7–5.5) 55–70 (2.2–2.8)

Source: Ref 16

Results of thermal-convection loop tests in molten lead

Loop material

Maximum temperature, °C (°F)

DT, °C (°F)

Operated time, (h)

Maximum depth of attack, µm (mils)

Inhibitors

Croloy 2.25Cr steel 410 Croloy 2.25Cr steel Carbon steel (A-106) Nb-1Zr (cladded on type 446) Croloy 2.25Cr steel(a)

654 (1210) 654 (1210) 593 (1100) 593 (1100) 760 (1400) 550 (1022)

149 (300) 149 (300) 93 (200) 93 (200) 204 (400) 105 (221)

266 1,346 5,156 5,064 5,280 27,765

25 (1.0) 50 (2.0) 203 (8.0) 254 (10) None None

None None None None None Mg+Zr

(a) Brookhaven loop. Source: Ref 18

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

Chapter 16:

16.6 Corrosion in Molten Lithium Liquid lithium is an attractive candidate blanket material for fusion reactors. Numerous studies (Ref 21–27) have evaluated the compatibility of various containment materials with liquid lithium. This section gives a brief summary of the comparative performance of alloys in molten lithium. The solubilities of some important alloying elements in liquid lithium are shown in Fig. 16.4 (Ref 25). Nickel has the highest solubility among the metals investigated by Cleary et al. (Ref 25). Thus, nickel and nickel-base alloys are generally not considered good candidates for handling liquid lithium. Cobalt-base alloys are only slightly more resistant than nickel-base alloys (Ref 21, 22, 28).

pg 427

Liquid Metal Corrosion and Embrittlement / 427

Figure 16.5 illustrates the performance of several iron-, nickel-, and cobalt-base alloys in molten lithium at 890 °C (1635 °F) (Ref 28). Two nickel-base alloys (RA333 and Hastelloy alloy X) suffered the worst corrosion attack in terms of weight loss, followed by a cobalt-base alloy (Airesist 213) with two iron-base alloys (E-Brite alloy and Type 304L) being the best performers. However, in terms of penetration depths, Airesist 213, E-Brite, and Type 304L behaved similarly (Fig. 16.6). TZM (Mo-0.5Ti0.1Zr) was also included in the investigation, revealing no corrosion attack at all. Freed and Kelly (Ref 21) performed corrosion tests in circulating molten lithium at 705 to 815 °C (1300 to 1500 °F) for iron and stainless steels (Table 16.9). Both suffered considerable metal loss due to dissolution. For austenitic

Temperature, °F 10–1

1700

1600

1500

1400

Nickel

1300

1200

Nickel

Chromium Iron 10–2

Titanium

Solubility, atomic percent solute

Titanium results reported as being <10 ppm Molybdenum Columbium

10–3

Chromium

Iron

10–4

Titanium

Columb

ium

Molybdenum

10–5 0.82

0.86

0.90

0.94

0.98

1.02

1.03

1.10

1000 / T , K

Fig. 16.4

Solubilities of some metals in molten lithium. Note: Columbium is the former (pre-1968) name of niobium. Source: Ref 25

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

pg 428

428 / High-Temperature Corrosion and Materials Applications

stainless steels, intergranular attack was also observed. The stabilized stainless steels, such as Types 321 and 347, were more resistant to intergranular attack. Also observed were signiﬁcant amounts of deposits due to mass transfer. This can result in ﬂow restrictions, or eventual tube plugging, in a circulating system. Chopra et al. (Ref 24) summarized the data generated from different circulating loops (forced circulation loop, FCL, and thermal convection loop, TCL) at different laboratories. The data, illustrated in

Fig. 16.7, are presented in terms of an Arrhenius plot of dissolution rates for ferritic steels (HT-9 and 9Cr-1Mo) and austenitic stainless steels (Type 316, cold-worked Type 316, and Type 304). HT-9 exhibits excellent compatibility with molten lithium.

16.7 Corrosion in Sodium Liquid sodium is used as a coolant for fast breeder reactors. Extensive studies (Ref 29–35) on the corrosion of various alloys in liquid

1 10

RA-333 Hastelloy x 1 Airesist-213

10–2 E-Brite 26-1

Penetration, mm

Weight loss, gm/cm2

10–1

RA-333 Hastelloy x

Airesist-213

10–1 304L s.s

E-Brite 26-1

304L stainless steel 10–3 TZM (no attack) 104

105

TZM (no attack) 106

10–2 104

Weight loss as a function of time for selected alloys in molten lithium (Ti-gettered) at 890 °C (1635 °F). Source: Ref 28

106

Time, s

Time, s

Fig. 16.5

105

Fig. 16.6

Total penetration as a function of time for selected alloys in molten lithium (Ti-gettered) at 890 °C (1635 °F). Source: Ref 28

Table 16.9 Results of corrosion tests in molten lithium at 705 to 815 °C (1300 to 1500 °F) in a forced-convection loop Alloy

Exposure time, h

Maximum depth of attack(a), mm (mils)

Maximum thickness of deposits, mm (mils)

Iron(b)

108–138

0.32–0.38 (12.5–15.0)

Iron

138–187

304

105–138

310

64–96

321

69–200

347

82–160

IG 0.0 ML 0.05–0.11 (2.0–4.5) IG 0.0–0.02 (0.0–0.6) ML 0.13–0.17 (5.0–6.6) IG 0.03–0.13 (1.0–5.0) ML 0.10–0.15 (3.8–6.0) IG 0.08–0.10 (3.0–4.0) ML 0.06–0.12 (2.2–4.7) IG 0.0–0.05 (0.0–2.0) ML 0.15–0.16 (6.2–6.4) IG 0.01–0.02 (0.5–0.6) ML 0.11–0.12 (4.3–4.9)

0.36–0.46 (14.0–18.0) 0.47–0.51 (18.5–20.0) 0.49–0.61 (19.5–24.0) 0.64–0.81 (25.0–32.0) 0.84–1.02 (33.0–40.0)

(a) ML, metal loss due to apparent solution attack, decrease in wall thickness; IG, intergranular attack. (b) Titanium getter in lithium ﬂow stream. Source: Ref 21

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

Chapter 16:

pg 429

Liquid Metal Corrosion and Embrittlement / 429

corrosion database for a wide variety of alloys has been published in Nuclear Systems Materials Handbook (Ref 36) to provide guidance to corrosion allowance for design calculation. This

sodium have been carried out in support of sodium breeder reactor programs. In addition to numerous papers and reports discussing the behavior of alloys in liquid sodium, a large

Temperature, °C 650

103

600

550

500

Type 316 stainless steel TCL and HT-9 TCL and FCL

450

400

350

Weight loss in lithium

ORNL ANL WARD UW 102

SU

Dissolution rate, mg/m2.h

JAERI

10

Type 316 stainless steel FCL

Scatter band type 316 stainless steel FCL

TCL Type 316 stainless steel PCA HT-9

1

FCL Type 316 stainless steel Type 316 CW stainless steel Type 304 stainless steel PCA HT-9 9 Cr-1Mo

10–1

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1000 /T, K

Fig. 16.7

Corrosion rates of Types 304 and 316, PCA (primary candidate alloy), and HT-9 and 9Cr-1Mo steels in flowing lithium. CW, cold worked; FCL, forced-convection loop; TCL, thermal-convection loop. Source: Ref 24

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

pg 430

430 / High-Temperature Corrosion and Materials Applications

section gives a very brief summary of the comparative performance of alloys in liquid sodium. Berry (Ref 1) summarized the data generated by numerous authors (Ref 29–34). The data were generated in both static and ﬂowing systems at temperatures from 550 to 595 °C (1025 to 1100 °F), with some data generated at 1000 °C (1830 °F). Carbon steel, Cr-Mo steels, and ferritic and austenitic stainless steels exhibited low corrosion rates at temperatures up to 595 °C (1100 °F) (Table 16.10). The major problem for low-chromium alloy steels in sodium is decarburization and resultant loss of strength (Ref 35). For austenitic stainless steels and nickel-base alloys, the reaction between the alloy and the sodium leads to carburization (Ref 35). Corrosion of alloys in liquid sodium can be severe at higher temperatures. Ferritic and austenitic stainless steels suffered rapid corrosion attack at 1000 °C (1830 °F) (Table 16.10). A nickel-base alloy (alloy 600) also exhibited a rapid corrosion rate (Table 16.10). Borgstedt et al. (Ref 37) investigated the corrosion behavior of several Fe-Ni-Cr and nickel-base alloys in liquid sodium at 1000 °C (1830 °F). Their results are summarized in Table 16.11. Nickelbase alloys suffered more attack than Fe-Ni-Cr alloys.

16.8 Corrosion in Other Molten Metals Corrosion data for other liquid metals, such as magnesium (melting point of 651 °C, or 1205 °F); cadmium (melting point of 321 °C, or 610 °F); mercury (melting point of −39 °C, or −38 °F); tin (melting point of 232 °C, or 450 °F); antimony (melting point of 631 °C, or 1165 °F); and bismuth (melting point of 271 °C,

Table 16.10 Corrosion of carbon steel, chromium-molybdenum steels, and stainless steels in liquid sodium under isothermal conditions

Materials

1010 steel 2.25Cr-1Mo

5Cr-0.5Mo

9Cr-1Mo

410 420 304 310 316 347 410 430 446 304 316 310 347 600

Test temperature, °C (°F)

Exposure time, h

Test system

Weight change rate, mg/cm2 per month

593 (1100) 593 (1100) 552 (1026) 556 (1033) 593 (1100) 593 (1100) 552 (1026) 566 (1033) 593 (1100) 593 (1100) 552 (1026) 566 (1033) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 593 (1100) 1000 (1830) 1000 (1830) 1000 (1830) 1000 (1830) 1000 (1830) 1000 (1830) 1000 (1830) 1000 (1830)

1000 1000 943 902 1000 1000 943 1913 500 500 943 902 500 500 1000 1000 1000 1000 1000 1000 500 500 1000 1000 500 500 400 400 400 400 400 400 400 400

Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Flowing Static Static Static Static Static Static Static Static Static

−0.49 −0.37 −0.12 −0.12 −0.14 −0.09 +0.22 -0.06 +0.23 −0.08 +0.35 −0.05 +0.70 +0.29 +0.38 +0.35 +0.33 +0.31 +0.17 +0.15 +0.75 +0.27 +0.10 +0.13 +1.46 +0.22 +29.8 +46.8 +28.2 +25.5 +29.6 +28.2 +44.2 +18.7

Note: Sodium contained a maximum of 100 ppm oxygen. Source: Ref 1, based on Ref 29–34

Table 16.11 Corrosion of several iron-nickel-chromium and nickel-base alloys in liquid sodium at 1000 °C (1830 °F) Test

Run No. 1, 1000 h

Run No. 2, 1100 h

Source: Ref 37

Alloy

Cr

Ni

Fe

Thermon 617 X AC-66 ASL71 Pyrotherm 800 625 625 617 X Pyrotherm 253-MA

22 21 21 27 20 20 20 22 22 21 21 20 21

bal bal bal 32 20 33 33 bal bal bal bal 33 11

30 1.5 18 bal bal bal bal 2.5 2.5 1.5 18 bal bal

Others

W, Nb Co, Mo, Al Mo Nb, Ce Co, W Nb Al, Ti Mo, Nb Mo, Nb Co, Mo, Al Mo Nb Si

Weight change, mg/cm2

−13.0 −13.0 −4.5 −2.35 −1.94 −2.0 −0.7 −36.01 −35.35 −25.11 −16.32 +1.43 +9.95

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

Chapter 16:

or 520 °F); are rather limited. Frequently, it is possible to use phase diagrams to determine which metals may react readily with the liquid metal. It is clear from the Ni-Mg phase diagram that magnesium reacts readily with nickel to form a low-melting-point eutectic, with a large liquid phase ﬁeld extending to low temperatures. More than 50% Ni, for example, can be in solution with magnesium in liquid at 800 °C (1470 °F). Nickel and nickel-base alloys are therefore not suitable for use as a containment material for molten magnesium. Cobalt also forms a low-melting-point eutectic, but with a much smaller liquid-phase ﬁeld. The solubility of iron in liquid magnesium is very low. It is thus reasonable to assume that iron-base alloys with low nickel are more suitable than cobalt- and nickel-base alloys to handle molten magnesium. The solubility of iron increases with increasing temperature. Thus, the corrosion rate is expected to increase with temperature as well. The solubility of iron in molten cadmium is low. Daniels (Ref 38) indicated the inactivity of steel in molten cadmium. Tammann and Oelsen (Ref 39) reported a solubility of 2 to 3× 10−4 wt% Fe in molten cadmium at 400 and 700 °C (750 and 1290 °F). Jackson and Adams Temperature, °C 780 1000

727

679

636

596

560

527

496

Ti Zr Ni

Ni

Zr

100 Ti 10 ppm in Hg

Cr

31/10/2007 12:52PM Plate # 0

pg 431

Liquid Metal Corrosion and Embrittlement / 431

(Ref 15) indicated that cast iron has frequently been used for handling molten cadmium. Nickel and nickel-base alloys are not suitable for handling molten tin because of relatively high solubilities of nickel in molten tin. Cast iron is often used in the laboratory for handling molten tin (Ref 15). Cobalt, nickel, and iron have high solubilities in molten antimony. These metals, as well as their alloys, are expected to have poor corrosion resistance in molten antimony. The solubilities of a number of metals in liquid mercury are illustrated in Fig. 16.8 (Ref 40). Both iron and cobalt have low solubilities in liquid mercury. Table 16.12 illustrates corrosion rates of carbon steels and stainless steels in liquid mercury (Ref 8).

16.9 Liquid Metal Embrittlement Metallic components may suffer embrittlement when in contact with liquid metal. Components are more susceptible to liquid metal embrittlement (LME) when under stresses. McDonalt (Ref 41) indicated that stainless steels, nickel alloys and Cu-Ni alloys may suffer LME by the molten brazing alloy while the component is stressed under brazing. Since the brazing operation involves contact of the metallic component with a molten brazing alloy, LME can occur during brazing operation. Heiple et al. (Ref 42) reported that austenitic stainless steels can be severely embrittled by copper and high copper braze alloys. However, austenitic stainless steels are not affected by silver-base braze alloys (Ref 42). It was also found that Type 430 (a ferritic stainless steel) was not embrittled by copper-base braze alloys, silver-base braze alloys, aluminum, or gold (Ref 42). Small

Co Fe

1.0

0.10

Table 16.12 Corrosion of metals and alloys in flowing liquid mercury

V

Nb Ta < 0.002

0.01 0.95

1.0

1.05

1.1

Maximum temperature, °C (°F)

Test duration, h

Corrosion rate, mm/yr (mpy)

Mild steel

482 (900) 538 (1000) 593 (1100) 649 (1200) 482 (900) 538 (1000) 593 (1100) 649 (1200) 652 (1205) 650 (1200)

(a) (a) (a) (a) (a) (a) (a) (a) 460 400–500

0.10 (4) 0.23 (9) 0.56 (22) 1.35 (53) 0.05 (2) 0.10 (4) 0.25 (10) 0.64 (25) 0.51 (20) 1.19 (47)

5Cr steel

1.15

1.2

1.25

1.3

1000/ T, K –1

Fig. 16.8

Material

Solubilities of some metals in liquid mercury. Note: Columbium is the former (pre-1968) name of niobium. Source: Ref 40

304 310

(a) Average of a large number of laboratory tests as well as samples from largescale boiler operations; exposures up to 10,000 h. Source: Ref 8

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

432 / High-Temperature Corrosion and Materials Applications

amounts of copper contaminant on the surface of the alloy that is under welding can cause LME in the heat-affected zone of cobalt-base alloys (Ref 43, 44). Savage and Mushala (Ref 44) believed that copper and cobalt are essentially insoluble in each other and thus form a classic LME couple. Korb (Ref 45) summarized some of the LME issues of various alloy combinations involved in the construction of manned spacecraft as follows. Many structural alloys are embrittled by lowmelting-point metals. Aluminum is embrittled by mercury, indium, tin, and zinc; steel by tin, cadmium, zinc, lead, copper, and lithium; stainless steels by cadmium, aluminum, lead, and copper; titanium by cadmium and mercury; and nickel by zinc, cadmium, and mercury. Many of these embrittling elements are used in low-temperature solders for electrical and avionics applications, for example, lead-tin for electrical soldered connections, indium-base solders for

(a)

(b)

Fig. 16.9

(a) Fractured studs due to cadmium-induced LME during service at 315 °C (600 °F), and (b) optical micrograph showing intergranular cracks at the cross section of a fractured stud. Source: Ref 46

glass-metal seals. Most silver solders (for higherstrength applications) melt in a temperature range of 610 to 800 °C (1125 to 1475 °F) and may contain copper, cadmium, zinc, lithium, and tin. Alloys are more susceptible to LME under stress. Many brazing alloys that are used for many stainless steel plumbing systems contain copper. For example, brazed manifold tube joints using Nicoro 80 braze alloy (81.5Au-16.5Cu2Ni) to join 21-6-9 stainless steels for the shuttle orbiter were found to suffer cracking. The cracking also was observed for brazed joints for 304L and alloy 718. Ebert (Ref 46) indicated that cadmium-plated Cr-Mo steel (ASTM A 193 grade B) studs from a steam line connector associated with a power turbine fractured during service at 315 °C (600 °F) by cadmium-induced LME. Figure 16.9(a) shows the fractured studs, and Fig. 16.9(b) shows intergranular cracks in the cross section of a fractured stud. Zinc has been widely used in the industry as a corrosion-resistant coating for carbon steels (e.g., hot dip galvanizing, electroplating, and spray painting) (Ref 47). Korbrin (Ref 47) indicated (a) carbon steels are susceptible to LME by zinc particularly under stresses or cold-worked conditions, (b) austenitic stainless steels and nickel alloys can suffer LME when in contact with molten zinc, or when welded to galvanized steels or parts contaminated with zinc. Intergranular cracking was observed in Type 316 and alloy 25 (Co-20Cr-10Ni-15W) coupons after exposure to molten zinc in a hot dip galvanizing tank at 455 °C (850 °F) for 50 h, as shown in Fig. 16.3 (Ref 12). Dillon (Ref 48) found that a Type 321 nozzle suffered LME due to molten zinc contamination of welds from zinc-pigmented painting overspray during initial fabrication. Dillon (Ref 49) observed the failure of ASTM A193 2H nuts during service at 370 °C (700 °F) and attributed the failure to LME by zinc due to cadmiumplated/zinc phosphate coated nuts. The Cd-Zn eutectic melts at 270 °C (515 °F). Gutzeit et al. (Ref 50) indicated that austenitic stainless steels are susceptible to LME by zinc when welding or during heat treatment of stainless steel components contaminated with zinc-rich paint. Figure 16.10 shows the formation of intergranular cracking at the heat-affected zone of a Type 304 pipe weld joint when the area was contaminated with zinc-rich paint during welding. Zinc-rich paints containing only metallic zinc powders as a principal component can cause

pg 432

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

Chapter 16:

100 µm

Fig. 16.10

Intergranular cracking in the heat-affected zone of a weld joint for Type 304 pipe that was contaminated with zinc-rich paint when welding was performed. Source: Ref 50

LME, while paints containing zinc oxide or zinc chromates do not cause LME (Ref 50).

31/10/2007 12:52PM Plate # 0

Liquid Metal Corrosion and Embrittlement / 433

Cast iron, steels, and stainless steels are commonly used for handling molten lead. Nickel and nickel-base alloys have poor resistance to molten lead corrosion because of the high solubility of nickel in molten lead. Nickel also has relatively higher solubility than iron in molten lithium, as well as in molten sodium. Thus, nickel-base alloys are not good candidates for handling either metal. Iron-base alloys are more suitable. Nickel reacts more readily with molten magnesium. The solubility of iron in molten magnesium is low. Thus, iron-base alloys with low nickel contents are preferred as containment materials. Due to low solubilities in the molten metal, iron and iron-base alloys may be suitable for molten cadmium, while iron, cobalt, and iron- and cobalt-base alloys may be suitable for molten mercury. Iron, nickel, and cobalt exhibit high solubilities in molten antimony. These metals and their alloys are thus not recommended for use as containment materials for molten antimony. Liquid metal embrittlement of various alloys by low-melting-point metals, such as copper, zinc, cadmium, mercury, lead, tin, and so forth, is reviewed. LME cases are often associated with welding, brazing, and soldering. Most brazing and soldering alloys contain alloying elements that can induce LME. The components are more susceptible to LME when they are under stress during welding or soldering.

16.10 Summary The corrosion behavior of alloys in molten aluminum, zinc, lead, lithium, sodium, magnesium, mercury, and other molten metals is reviewed. Corrosion data useful in assisting selection of materials are presented. Also presented is the information about the LME caused by a wide variety of low-melting-point metals during welding or heat treatment. Molten aluminum is extremely aggressive. Iron-, nickel-, and cobalt-base alloys are readily attacked by molten aluminum. Molten zinc is less aggressive. Nickel and nickel-base alloys, however, react readily with molten zinc and are not recommended for use. Cast iron, steels, and iron- and cobalt-base alloys are generally suitable for containment applications. Cobaltbase alloys are generally more resistant than ironbase alloys. However, under some conditions, various metals or alloys may be susceptible to LME by molten zinc.

REFERENCES

1. W.E. Berry, Corrosion in Nuclear Applications, John Wiley & Sons, 1971 2. H.U. Borgstedt, Ed., Materials Behavior and Physical Chemistry in Liquid Metal Systems, Plenum Press, 1982 3. J.E. Draley and J.R. Weeks, Ed., Corrosion by Liquid Metals, Plenum Press, 1970 4. C. Bagnall and W.F. Brehm, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 91 5. D.L. Katz, Liquid-Metals Handbook, R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.), U.S. Government Printing Ofﬁce, Washington, D.C., 1952, p 1 6. J.V. Cathcart and W.D. Manly, Corrosion, Vol 12, 1956, p 87t 7. P.F. Tortorelli, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 56

pg 433

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

31/10/2007 12:52PM Plate # 0

434 / High-Temperature Corrosion and Materials Applications

8. E.C. Miller, Liquid-Metals Handbook, R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.), U.S. Government Printing Ofﬁce, Washington, D.C., 1952 9. W.D. Manly, Corrosion, Vol 12 (No. 7), 1956, p 336t 10. A. Brasunas, Corrosion, Vol 9, 1953, p 78 11. G.Y. Lai, unpublished results, Haynes International, Inc., 1985 12. S.K. Srivastava, Proc. Conf. Performance of High Temperature Materials in Fluidized Bed Combustion Systems and Process Industries, P. Ganesan and R.A. Bradley, Ed., ASM International, 1987, p 161 13. F.R. Morrall, Wire Wire Prod., Vol 23, 1948, p 484, 571 14. L.R. Kelman, W.D. Wilkinson, and F.L. Yaggee, “Resistance of Materials to Attack by Liquid Metals,” Report ANL-4417, Argonne National Laboratory, 1950 15. C.B. Jackson and R.M. Adams, LiquidMetals Handbook, R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.), U.S. Government Printing Ofﬁce, Washington, D.C., 1952 16. I. Ali-Khan, Materials Behavior and Physical Chemistry in Liquid Metal Systems, H.U. Borgstedt, Ed., Plenum Press, 1982, p 243 17. W.D. Wilkinson, E.W. Hoyt, and H.V. Rhude, “Attack on Materials by Lead at 1000 °C,” USAEC Report ANL-5449, Argonne National Laboratory, 1955 18. G.M. Tolson and A. Taboada, “A Study of Lead and Lead-Salt Corrosion in Thermal Convection Loops,” ORNL-TM-1437, Oak Ridge National Laboratory, 1966 19. R.C. Asher, D. Davies, and S.A. Beetham, Corros. Sci., Vol 17, 1977, p 545 20. O.F. Kammerer et al., Trans. AIME, Vol 212, 1958, p 20 21. M.S. Freed and K.J. Kelly, “Corrosion of Columbium Base and Other Structural Alloys in High Temperature Lithium,” Report No. PWAC-355, Pratt and Whitney Aircraft—CANEL, Division of United Aircraft Corp., June 1961 (declassiﬁed in June 1965) 22. E.E. Hoffman, “Corrosion of Materials by Lithium at Elevated Temperatures,” USAEC Report ORNL-2924, Oak Ridge National Laboratory, Oak Ridge, TN, 1960 23. J.R. DiStefano, “Corrosion of Refractory Metals by Lithium,” USAEC Report ORNL3551, Oak Ridge National Laboratory, Oak Ridge, TN, 1964

24. O.K. Chopra, D.L. Smith, P.F. Tortorelli, J.H. DeVan, and D.K. Sze, Fusion Technol., Vol 8, 1985, p 1956 25. R.E. Cleary, S.S. Blecherman, and J.E. Corliss, “Solubility of Refractory Metals in Lithium and Potassium,” USAEC Report TIM-850, Nov 1965 26. D.A. Bates, G.R. Edwards, and D.L. Olson, An Evaluation of Engineering Alloys for High Temperature Lithium Containment, Mater. Perform., March 1980, p 41 27. V. Coen, H. Kolbe, L. Orecchia, and T. Sasaki, Materials Behavior and Physical Chemistry in Liquid Metal Systems, H.U. Borgstedt, Ed., Plenum Press, 1982, p 121 28. P.A. Steinmeyer, D.L. Olson, G.R. Edwards, and D.K. Matlock, Rev. Coatings Corros., Vol 4 (No. 4), 1981, p 349 29. R.F. Koening and E.G. Brush, Mater. Methods, Vol 42, 1955, p 112 30. A. Brasunas, “Interim Report on StaticLiquid Metal Corrosion,” USAEC Report ORNL-1647, Oak Ridge National Laboratory, Oak Ridge, TN, 1954 31. R.F. Dudek and K.M. Ferguson, “The Corrosion Testing of Various Materials in Sodium: Part I and II,” USAEC Report BW-7020, Babcock & Wilcox, April 1957 32. V.W. Eldred, “Interactions Between Solid and Liquid Metals and Alloys,” British Report AERE-X/R-1806, Nov 1955 33. W. Markert, Jr., “The Corrosion Testing of Various Materials in Sodium,” USAEC Report BW-3792, Babcock & Wilcox, Aug 1954 34. W.C. Hayes and O.C. Shepard, “Corrosion and Decarburization of the Ferritic Chromium-Molybdenum Steels in Sodium Coolant Systems,” USAEC Report NAASR-2973, North American Aviation, Dec 1958 35. J.H. Stang, E.M. Simons, J.A. DeMastry, and J.M. Genco, “Compatibility of Liquid and Vapor Alkali Metals with Construction Materials,” DMIC Report 227, Defense Metals Information Center, Battelle Memorial Institute, April 1966 36. Nuclear Systems Materials Handbook, Hanford Engineering Laboratory, Richland, WA, 1976 37. H.U. Borgstedt, G. Frees, and H. Jesper, Werkst. Korros., Vol 40, 1989, p 525

pg 434

Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/

Chapter 16:

38. E.J. Daniels, J. Inst. Met., Vol 46, 1931, p 87 39. G. Tammann and W. Oelsen, Z. Anorg. Chem., Vol 186, 1930, p 277 40. J.R. Weeks, Corrosion, April 1967, p 98 41. M.M. McDonalt, Corrosion of Brazed Joints, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 876 42. C. Heiple, W. Bennett, and T. Rising, Embrittlement of Several Stainless Steels by Liquid Copper and Liquid Braze Alloys, Mater. Sci. Eng., Vol 52, 1982, p 177 43. S.J. Matthews, M.O. Maddock, and W.F. Savage, How Copper Surface Contamination Affects Weldability of Cobalt Superalloys, Weld. J., May 1972 44. W.F. Savage and M. Mushala, Copper Contamination Cracking in the Weld Heat Affected Zone, Weld. J., May 1978, p 145 45. L.J. Korb, Corrosion of Manned Spacecraft, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, Metals Park, Ohio, 1987, p 1059

31/10/2007 12:52PM Plate # 0

Liquid Metal Corrosion and Embrittlement / 435

46. H.E. Ebert, Liquid Metal Embrittlement of Flange Connector Studs in Contact with Cadmium, Handbook of Case Histories in Failure Analysis, Vol 1, R.C. Uhl, et al., Ed., ASM International, 1992, p 335 47. G. Korbrin, Materials Selection, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 321 48. C.P. Dillon, Unusual Corrosion Problems in the Chemical Industry, MTI Publication No. 54, Materials Technology Institute of the Chemical Process Industries, Inc., 2000, p 161 49. C.P. Dillon, Unusual Corrosion Problems in the Chemical Industry, MTI Publication No. 54, Materials Technology Institute of the Chemical Process Industries, Inc., 2000, p 204 50. J. Gutzeit, R.D. Merrick, and L.R. Scharfstein, Corrosion in Petroleum Reﬁning and Petrochemical Operations, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 1263

pg 435

Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p437-441 DOI: 10.1361/hcma2007p437

31/10/2007 4:06PM Plate # 0

pg 437

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

CHAPTER 17

Hydrogen Attack 17.1 Introduction Hydrogen attack can result in brittle fracture of a steel component during high-temperature service. It has also been referred to as “hydrogen damage.” Carbon steels or low-alloy steels can suffer severe hydrogen attack during hightemperature service, resulting in brittle failure due to severe loss in tensile and rupture strengths as well as ductility. Hydrogen attack should not be confused with hydrogen embrittlement. Hydrogen attack occurs only when steel is in service at elevated temperatures, while hydrogen embrittlement occurs near or at room temperature. Carbon and Cr-Mo steels derive strength primarily from carbon in the metal. In carbon steels, carbon in the metal combines with iron to form iron carbides, primarily cementite (Fe3C), which along with ferrite lamellae, constitutes to pearlite phases in the matrix of ferrite, as shown in Fig. 17.1. For low-alloy steels containing chromium and molybdenum, these two alloying elements can form carbides, such as M3C, M2C, and M23C6 where M can represent Fe, Cr, and Mo.

When steel is in contact with hydrogen molecules (H2) at elevated temperatures, hydrogen atoms can be absorbed at the steel surface and then diffuse into the metal. Hydrogen atoms in the metal then react with iron carbide (Fe3C) to form methane (CH4): Fe3 C+4H ! 3Fe+CH4

(17:1)

As Reaction 17.1 continues with ongoing ingress of hydrogen atoms into the steel, methane gas is increasingly generated in the steel. With its low diffusivity in steel, methane gas is accumulating at grain boundaries and other interfaces, causing the methane gas pressure to build up in the steel. This eventually results in development of microcracks and microﬁssures. At the same time, carbon is being removed from iron carbide (Fe3C) as a result of this chemical reaction, leading to decarburization in the affected region. The formation of microcracks and microﬁssures as well as decarburization causes the steel to lose its tensile and rupture strengths as well as ductility, thus resulting in rupture of the steel component during service. There are two sources of hydrogen that contribute to hydrogen attack of steels in industries. One source of hydrogen is from the corrosion of steel by boiler water in the waterwall tube for coal-ﬁred boilers, while the other source of hydrogen is from high-pressure, hightemperature hydrogen-containing atmospheres used in petroleum reﬁning. Hydrogen attack problems related to these two separate areas are discussed in the following sections.

17.2 Hydrogen Attack in Coal-Fired Boilers Fig. 17.1

Typical microstructure of carbon steel, consisting of ferrite (white grains) and pearlite (dark regions). The pearlite consists of cementite (Fe3C) and ferrite lamellae.

Hydrogen attack occurs more often in subcritical drum boilers that use a recirculating steam-generating system. Adequate control of

Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/

31/10/2007 4:06PM Plate # 0

438 / High-Temperature Corrosion and Materials Applications

water chemistry is critical to avoid internal deposition and corrosion of boiler tubes. The water chemistry is typically controlled by (a) water puriﬁcation to remove impurities using makeup water, polishing of returned condensate, deaeration, and blowdown, and (b) chemical treatments to control pH, electrochemical potential, and dissolved oxygen concentration (Ref 1). In the waterwall tubes, as steam forms, dissolved solids concentrate in the boiler water; once the solubility limit of an impurity is exceeded, deposits can precipitate out (Ref 1). Typical boiler deposits are hardness deposits of calcium and magnesium salts and metal oxides (Ref 1). These deposits formed on the surface of the waterwall tube can provide a location for accumulation of corrosive boiler water contaminants and/or chemical additives and their reaction compounds (Ref 2). The deposits may lead to low pH water chemistry in the local area, resulting in acidic corrosion attack (Ref 3). With adequate water chemistry control, the waterwall steel tube forms a protective magnetite (Fe3O4) scale when steel is corroded by water under normal operating conditions: 3Fe+4H2 O ! Fe3 O4 +4H2

17.1, the affected area is decarburized. The microstructure of the steel suffering hydrogen attack is shown in Fig. 17.2. Hydrogen attack typically is associated with heavy scales (Fig. 17.2). Figure 17.3 shows microcracks and microﬁssures that formed along grain boundaries for the waterwall steel tube (ASME SA192) suffering hydrogen attack in a subcritical coalﬁred boiler. Also shown in the ﬁgure is the complete decarburization of the steel in the affected area. Figure 17.4 shows the typical pearlite-ferrite microstructure in the unaffected waterwall tube in the same boiler as discussed in Fig. 17.2 and 17.3. Large cracks developed in the waterwall steel tube through the linking of numerous microcracks and microﬁssures. This is illustrated in Fig. 17.5. Large cracks were often found to show oxides formed on the surface of these cracks (Fig. 17.5). The tube eventually ruptured from the internal ﬂuid pressure. The locations that are often susceptible to hydrogen attack in the boiler are the burner zone and the bull nose area, as illustrated in Fig. 17.6 (Ref 4). Monitoring and controlling boiler water

(17:2)

The by-product of the reaction is hydrogen. The growth of the magnetite (Fe3O4) scale that forms on the steel surface follows a parabolic rate law with the scale growth rate diminishing with time. Accordingly, the generation of hydrogen likewise diminishes with time once a steady state is reached. Atomic hydrogen produced combines with other hydrogen atoms to form molecular hydrogen that is then dispersed in the boiler water as a gas or in solution (Ref 2). However, when water chemistry is poorly controlled, heavy internal tube deposits may develop. This condition may lead to acidic corrosion attack on the tube under the deposits, resulting in large amounts of hydrogen atoms being absorbed by the steel. Hydrogen atoms then react with iron carbide (Fe3C) in the steel to form methane gas (CH4) according to Reaction 17.1. Methane gas then accumulates at grain boundaries and other interfaces due to its low diffusivity. Microcracks and microﬁssures are eventually developed by increasing local gas pressure created by the increasing amount of methane gas produced by the iron-carbide/ hydrogen reaction. Furthermore, because of iron carbides being reduced to iron by Reaction

Fig. 17.2

Low-magnification optical micrograph showing numerous microcracks and microfissures formed in the carbon steel (ASME SA192) of the waterwall tube that suffered hydrogen attack in a subcritical coal-fired boiler. Note numerous microcracks and microfissures. Also shown are heavy corrosion scales formed on the tube internal surface. Unetched

pg 438

Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/

31/10/2007 4:06PM Plate # 0

Chapter 17:

chemistry is critical in preventing internal tube deposits and hydrogen attack (Ref 3).

17.3 Hydrogen Attack in Petroleum Refining In petroleum reﬁning, some reﬁnery equipment, such as reactors in hydrotreating, reforming, and hydrocracking units, is exposed to a high-temperature, high-pressure hydrogen atmosphere. Hydrogen attack is potentially a very serious materials issue in the design and operation of this type of equipment in hydrogen service (Ref 5, 6). Steels are susceptible to hydrogen attack as functions of temperature and

Fig. 17.3

Optical micrograph showing the formation of microcracks and microfissures along grain boundaries and decarburization of carbon steel (ASME SA192) in a waterwall tube that suffered hydrogen attack in a subcritical coal-fired boiler

Fig. 17.4

Optical micrograph showing typical microstructure (pearlite and ferrite) in the unaffected carbon steel waterwall tube in the same boiler as discussed in Fig. 17.2 and 17.3

Hydrogen Attack / 439

partial pressure of hydrogen in the system. For hydrotreating, reforming, and hydrocracking units, steel can suffer hydrogen attack at temperatures above approximately 260 °C (500 °F) and hydrogen partial pressures above 0.689 MPa (100 psig) (Ref 7). The mechanism for hydrogen attack of steels in petroleum reﬁning is essentially the same as that described in Section 17.2. Atomic hydrogen is ﬁrst absorbed by the steel at the surface of the reﬁnery equipment that is in contact with the high-temperature, high-pressure hydrogen gas stream. Hydrogen atoms then diffuse into the steel and react with iron carbide (Fe3C) in the steel to form methane gas (CH4) according to Reaction 17.1. Methane gas then accumulates at grain boundaries and other interfaces due to its low diffusivity. Continuous ingress of hydrogen atoms into the steel results in an increasing amount of methane gas being generated by the iron-carbide/hydrogen reaction, eventually causing microcracks and microﬁssures in the steel. Cracks then develop by linking microcracks and microﬁssures. Furthermore, because iron carbides are being reduced to iron by Reaction 17.1, the affected area is decarburized. As a result, the steel eventually ruptures from the internal pressure of the reactor or vessel. The microstructure of the steel that suffers hydrogen attack is characterized by numerous microcracks and microﬁssures (or cracks) and decarburized microstructure, which is essentially the same as that described in the waterwall tube that suffers hydrogen attack in the boiler. Typical optical microstructure of the hydrogen-attacked steel is similar to that shown in Fig. 17.3.

Fig. 17.5

Optical micrograph showing a large crack that was developed by linking microcracks and microfissures in the steel (ASME SA192) of the waterwall tube that suffered hydrogen attack. Oxides were often observed to form around the surface of the crack.

pg 439

Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/

31/10/2007 4:06PM Plate # 0

440 / High-Temperature Corrosion and Materials Applications

Adding chromium and/or molybdenum to the steel to increase the stability of iron carbides can increase the resistance of the steel to hydrogen attack. Some concerns have been raised for the long-term performance of C-0.5Mo steel in hightemperature hydrogen environments (Ref 8). Thus, Cr-Mo steels are much more resistant to hydrogen attack than carbon and C-0.5Mo steels. The conditions under which carbon and Cr-Mo steels can be used in high-temperature hydrogen service are described in detail in API 941 (Ref 9). The behavior of carbon and Cr-Mo steels with respect to their resistance to hydrogen attack is summarized in Nelson curves in Fig. 17.7 (Ref 9).

17.4 Summary During service at elevated temperatures, carbon steel can react with atomic hydrogen and result in brittle fracture. This phenomenon is often referred to as “hydrogen attack” or

“hydrogen damage.” The chemical reaction involves atomic hydrogen reacting with iron carbide in the steel to form methane gas. Continuing ingress of atomic hydrogen into the metal causes an increasing amount of methane gas to be generated and accumulated at grain boundaries and other interfaces, resulting in the formation of microcracks and microﬁssures in the steel. In addition, the steel is decarburized. Continuing growth of microcracks and microﬁssures along with decarburization of steel eventually result in rupture of the steel component during service at elevated temperatures. The source of atomic hydrogen can be from the rapid waterside corrosion at the internal diameter of the waterwall tubes in a coal-ﬁred boiler when water chemistry is not properly controlled, thus resulting in hydrogen attack. Hydrogen attack can also occur in some reﬁnery equipment, such as reactors in hydrotreating, reforming, and hydrocracking units, which is exposed to a high-temperature, high-pressure hydrogen atmosphere. Hydrogen attack in both of these systems is discussed.

Steam Risers 1

3 2 Downcomer

Heat flux profile

1 1

Furnace tubes Steamwater mixture

1

1 2

Burners

1 1

Heat flux Steam-free subcooled water in

Fig. 17.6 Wilcox

Locations, marked as 1, in the boiler that are susceptible to hydrogen attack. The area marked as 2 shows other modes of waterside corrosion that are outside of the current discussion topic. Source: Stultz and Kitto (Ref 4) Courtesy of Babcock and

pg 440

Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/

31/10/2007 4:06PM Plate # 0

Chapter 17:

Hydrogen partial pressure, MPa absolute 1500

3.45

6.90

10.34

13.79

17.24

pg 441

Hydrogen Attack / 441

34.5 62.1 20.7 48.3 75.8 800

1400 1300

700

1200 600

1000

6.0Cr-0.5Mo steel 1.25Cr-0.5Mo steel

500

3.0Cr-0.5Mo steel

900

2.25Cr-1.0Mo steel

1.0Cr-0.5Mo steel

800

2.0Cr-0.5Mo steel

700 600

1.25Cr-0.5Mo or 1.0Cr-0.5Mo steel

Temperature, °C

Temperature, °F

1100

400

300

500 Carbon Steel

400 300 Legend: Surface decarburization Internal decarburization (Hydrogen attack)

Fig. 17.7

200 500

1000

1500

2000

Hydrogen partial pressure, 1b/psia

2500

3000 7000 11,000 5000 9000 13,000 Scale Change

Nelson curves showing the temperature and hydrogen partial pressure conditions under which carbon and Cr-Mo steels are susceptible to hydrogen attack. Source: API 421 (Ref 9). Courtesy of American Petroleum Institute.

REFERENCES

1. P. Daniel, Corrosion of Steam- and WaterSide of Boilers, Corrosion: Environments and Industries, Vol 13C, ASM Handbook, ASM International, 2006, p 466 2. H.A. Grabowski, Chapter 17, Management of Cycle Chemistry, ASME Handbook on Water Technology for Thermal Power Systems, P. Cohen, Ed., ASME, 1989, p 1379 3. R.B. Dooley, Corrosion of Steam/Waterside Boilers, Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, 1987, p 990 4. S.C. Stultz and J.B. Kitto, Ed., Steam and Its Generation and Use, 40th ed., Babcock & Wilcox, 1992, p 42-22 5. G. Sorell and M.J. Humphries, High Temperature Hydrogen Damage in Petroleum

6. 7.

8.

9.

Reﬁnery Equipment, Mater. Perform., Vol 17 (No. 8), 1978, p 33 A.R. Ciuffreda and W.R. Rowland, Hydrogen Attack of Steel in Reformer Service, Proc. API, Vol 37 (No. III), 1957, p 116 R.D. Merrick and A.R. Ciuffreda, Hydrogen Attack of Carbon-0.5 Molybdenum Steels, Proc. API, Vol 61 (No. III), 1982, p 101 R. Chiba, K. Ohnishi, K. Ishii, and K. Maeda, Effect of Heat Treatment on Hydrogen Attack Resistance of C-0.5Mo Steels for Pressure Vessels, Heat Exchangers, and Piping, Corrosion, Vol 41 (No. 7), 1985, p 415 Steels for Hydrogen Service at Elevated Temperatures and Pressures in Petroleum Reﬁneries and Petrochemical Plants, Publication 941, 3rd ed., American Petroleum Institute, 1983

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

High-Temperature Corrosion And Materials Applications George Y. Lai, editor, p445-453 DOI: 10.1361/hcma2007p445

pg 445

Copyright © 2007 ASM International® All rights reserved. www.asminternational.org

APPENDIX 2

Chemical Compositions of Alloys and Filler Metals Table 1 Chemical compositions (wt.%) of wrought ferritic steels Grade*

C

Cr

Mo

Mn

Si

Others

T2 T11 T12 T17 T21 T22 T23

0.10–0.20 0.05–0.15 0.05–0.15 0.15–0.25 0.05–0.15 0.05–0.15 0.04–0.10

0.50–0.81 1.00–1.50 0.80–1.25 0.80–1.25 2.65–3.35 1.90–2.60 1.90–2.60

0.44–0.65 0.44–0.65 0.44–0.65 … 0.80–1.06 0.87–1.13 0.05–0.30

0.30–0.61 0.30–0.60 0.30–0.61 0.30–0.61 0.30–0.60 0.30–0.60 0.10–0.60

0.10–0.30 0.50–1.00 0.50(a) 0.15–0.35 0.50(a) 0.50(a) 0.50(a)

T5 T9 T91 T92

0.15(a) 0.15(a) 0.08–0.12 0.07–0.13

4.00–6.00 8.00–10.00 8.00–9.50 8.50–9.50

0.45–0.65 0.90–1.10 0.85–1.05 0.30–0.60

0.30–0.60 0.30–0.60 0.30–0.60 0.30–0.60

0.50(a) 0.25–1.00 0.20–0.50 0.50(a)

T122

0.07–0.14

10.00–12.50

0.25–0.60

0.70(a)

0.50(a)

… … … V: 0.15 min … … W: 1.45–1.75, V: 0.20–0.30, Nb: 0.02–0.08, B: 0.0005–0.006 … … V: 0.18–0.25, Nb: 0.06–0.10, N: 0.03–0.07 V: 0.15–0.25, W: 1.5–2.0, Nb: 0.04–0.09, B: 0.001–0.006, N: 0.03–0.07 V: 0.15–0.30, W: 1.50–2.50, Cu: 0.30–1.70, Nb: 0.04–0.10, B: 0.0005–0.005, N: 0.04–0.10

(a) Maximum, Nb = Cb. * For tubes

Table 2 Chemical compositions (wt.%) of wrought stainless steels Alloy

UNS No.

C

Cr

Ni

Fe

403 410 414 416 416 (Se) 420 431 440A 440B 440C GREEK ASCOLOY 154CM 405 430 430F 430F (Se) 446 409 439 444

S40300 S41000 S41400 S41600 S41623 S42000 S43100 S44002 S44003 S44004 S41880 … S40500 S43000 S43020 S43023 S44600 S40900 S43035 S44400

0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.20(a) 0.60–0.75 0.75–0.95 0.95–1.20 0.12 1.05 0.08(a) 0.12(a) 0.12(a) 0.12(a) 0.20(a) 0.08(a) 0.07(a) 0.025(a)

11.5–13.0 11.5–13.5 11.5–13.5 12.0–14.0 12.0–14.0 12.0–14.0 15.0–17.0 16.0–18.0 16.0–18.0 16.0–18.0 12.6 14.0 11.5–14.5 14.0–18.0 14.0–18.0 14.0–18.0 23.0–27.0 10.5–11.75 17.0–19.0 17.5–19.5

… … 1.25–2.50 … … … 1.25–2.50 … … … 2.0 … … … … … … 0.5(a) 0.5(a) 1.0(a)

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

E-BRITE MONIT

S44627 S44635

0.01(a) 0.025(a)

25.0–27.5 24.5–26.0

0.5(a) 3.5–4.5

Bal Bal

SEA-CURE

S44660

0.03(a)

25.0–28.0

1.0–3.5

Bal

AL 29-4C

S44735

0.03(a)

28.0–30.0

1.0

Bal

AL 29-4-2 18SR 12SR

S44800 … …

0.01(a) 0.03 0.02

28.0–30.0 18.0 12.0

2.0–2.5 … … (continued)

Bal Bal Bal

(a) Maximum. Cb = Nb

Others

… … … S: 0.15 min Se: 0.15 min … … Mo: 0.75(a) Mo: 0.75(a) Mo: 0.75(a) W: 3.0 Mo: 4.0 At: 0.10–0.30 … S: 0.15 min Se: 0.15 min N: 0.25(a) Ti: (6 × C) min, 0.75(a) Ti: 0.2 + 4(C + N) min, 1.10(a) Mo: 1.75–2.50, N: 0.035(a), Ti + Cb: 0.2 + 4(C + N) min, 0.8(a) Mo: 0.75–1.50, N: 0.015(a), Cb: 0.05–0.20 Mo: 3.4–4.5, N: 0.035(a), Ti + Cb: 0.2 + 4(C + N) min, 0.80(a) Mo: 3.0–3.5, N: 0.04(a), Ti + Cb: 0.2–1.0, 6(C + N) min Mo: 3.6–4.2, N: 0.045(a), Ti + Cb: 0.2–1.0, 6(C + N) min Mo: 3.5–4.2, N: 0.02(a), C + N: 0.025(a) Al: 1.8, Ti: 0.4 Al: 1.2, Cb: 0.6, Ti: 0.3

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

446 / High-Temperature Corrosion and Materials Applications

Table 2

(Continued)

Alloy

UNS No.

C

Cr

Ni

Fe

Others

ALFA-I ALFA-II 329 URANUS 50 CD-4MCu 44LN DP-3 3RE60 2205 FERRALIUM 255 7-Mo: PLUS SUPEER-FERRIT 201 202 301 302 302B 303 303(Se) 304 304L 304H 305 308 309 309S 310 310S 314 316 316L 316H 317 321 321H 347 347H 348 253MA RA85H 17-4PH 17-7PH PH15-7Mo A286 AM 350 AM 355 16-18 17-14CuMo 20-29CuMo 17-10P HMN TENELON 254SMO 19-9 DL 904L 16-25-6 15-5PH CUSTOM 450 CUSTOM 455 AL-6XN MVMA 22-4-9 NITRONIC 60 NITRONIC 50 NITRONIC 40 (21-6-9) SNR-4 317LM 17-14-4LM JS700 CRUTEMP: 25

… … S32900 S32404 J93370 S31200 S31260 S31500 S31803 S32550 S32950 … S20100 S20200 S30100 S30200 S30215 S30300 … S30400 S30403 S30409 S30500 S30800 S30900 S30908 S31000 S31008 S31400 S31600 S31603 S31609 S31700 S32100 S32109 S34700 S34709 S34800 S30815 S30615 S17400 S17700 S15700 K66286 S35000 S35000 … … … … … S21400 S31254 K63198 N08904 … S15500 S45000 S45500 N08367 S30415 … S21800 S20910 S21900

0.025 0.025 0.08(a) 0.04(a) 0.04(a) 0.03(a) 0.03(a) 0.03(a) 0.03(a) 0.04(a) 0.03(a) … 0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.15(a) 0.08(a) 0.03(a) 0.04–0.10 0.12(a) 0.08(a) 0.2(a) 0.08(a) 0.25(a) 0.08(a) 0.25(a) 0.08(a) 0.03(a) 0.04–0.10 0.08(a) 0.08(a) 0.04–0.10 0.08(a) 0.04–0.10 0.08(a) 0.08 0.2 0.04 0.07 0.07 0.08 0.1 0.13 0.05 0.12 0.05 0.12 0.30 0.08 0.02 0.3 0.02 0.06 0.07 0.05(a) 0.03 0.03 0.05 0.5 0.05 0.03 0.05

13.0 13.0 23.0–28.0 20.5–22.5 24.5–26.5 24.0–26.0 24.0–26.0 18.0–19.0 21.0–23.0 24.0–27.0 26.0–29.0 28.0 16.0–18.0 17.0–19.0 16.0–18.0 17.0–19.0 17.0–19.0 17.0–19.0 17.0–19.0 18.0–20.0 18.0–20.0 18.0–20.0 17.0–19.0 19.0–21.0 22.0–24.0 22.0–24.0 24.0–26.0 24.0–26.0 23.0–26.0 16.0–18.0 16.0–18.0 16.0–18.0 18.0–20.0 17.0–19.0 17.0–19.0 17.0–19.0 17.0–19.0 17.0–19.0 21.0 18.5 16.5 17.0 15.0 13.5–16.0 16.5 15.5 16.0 16.0 20.0 17.0 18.5 17.0 19.5–20.5 19.0 19.0–23.0 16.0 15.0 15.5 11.75 20.0–22.0 18.5 21.5 17.0 22.0 21.0

… … 2.5–5.0 5.8–8.5 4.75–6.0 5.5–6.5 5.5–7.5 4.25–5.25 4.5–6.5 4.5–6.5 3.5–5.2 3.2 3.5–5.5 4.0–6.0 6.0–8.0 8.0–10.0 8.0–10.0 8.0–10.0 8.0–10.0 8.0–10.5 8.0–12.0 8.0–10.5 10.0–13.0 10.0–12.0 12.0–15.0 12.0–15.0 19.0–22.0 19.0–22.0 19.0–22.0 10.0–14.0 10.0–14.0 10.0–14.0 11.0–15.0 9.0–12.0 9.0–12.0 9.0–13.0 9.0–13.0 9.0–13.0 11.0 14.5 4.25 7.0 7.0 24.0–27.0 4.25 4.25 19.0 14.0 29.0 10.5 9.5 … 17.5–18.5 9.0 23.0–28.0 25.0 4.5 6.0 8.5 23.5–25.5 9.5 4.0 8.5 12.5 6.0

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

Al: 3.0, Ti: 0.4 Al: 4.0, Ti: 0.4 Mo: 1.0–2.0 Mo: 2.0–3.0, Cu: 1.5, N: 0.2(a) Mo: 1.75–2.25, Cu: 3.0 Mo: 1.2–2.0, N: 0.14–0.20 Mo: 2.5–3.5, Cu: 0.2–0.8, N: 0.1–0.3, W: 0.1–0.4 Mo: 2.5–3.0, N: 0.08–0.15, Si: 1.4–2.0 Mo: 2.5–3.5, N: 0.08–0.2 Mo: 2.0–4.0, Cu: 1.4–2.5, N: 0.1–0.25 Mo: 1.0–2.5, N: 0.15–0.35 Mo: 2.1 Mn: 5.5–7.5, N: 0.25(a) Mn: 7.5–10.0, N: 0.25(a) … … Si: 2.0–3.0 S: 0.15 min Se: 0.15 min … … … … … … … … … Si: 1.5–3.0 Mo: 2.0–3.0 Mo: 2.0–3.0 Mo: 2.0–3.0 Mo: 3.0–4.0 Ti: 5 × C min Ti: 5 × C min Cb + Ta: 10 × C min Cb + Ta: 10 × C min Cb + Ta: 10 × C min, Ta: 0.1(a) Si: 1.7, N: 0.17, Ce: 0.04 Si: 3.5, Al: 1.0 Cb: 0.25, Cu: 3.6 Al: 1.15 Mo: 2.25, A1: 1.15 Mo: 1.0–1.5, Ti: 1.9–2.35, V: 0.1–0.5 Mo: 2.75, N: 0.1 Mo: 2.75, N: 0.12 … Mo: 2.5, Cb: 0.4, Ti: 0.3, Cu: 3.0 Mo: 2.20, Cu: 3.20 P: 0.28 Mn: 3.5, P: 0.25 Mn: 14.5, N: 0.4 Mo: 6.0–6.5, Cu: 0.5–1.0, N: 0.18–0.22 Mo: 1.25, W: 1.25, Cb: 0.4 Mo: 4.0–5.0, Cu: 1.0–2.0, V: 0.1–0.5 Mo: 6.0 Cb: 0.3, Cu: 3.5 Mo: 0.75, Cb: 8 × C min, Cu: 1.5 Cb: 0.3, Ti: 1.2, Cu: 2.25 Mo: 6.0–7.0, N: 0.18–0.25 Si: 1.3, N: 0.15, Ce: 0.04 N: 0.4, S: 0.1 Si: 4.0, Mn: 8.0, N: 0.13 Mo: 2.0, Mn: 5.0, N: 0.3, Cb: 0.2, V: 0.2 Mn: 9.0

S31753 S31725 S31726 N08700 …

0.03 0.03 0.03 0.03 0.05

18.5 18.0 17.0 21.0 25.0

13.5 15.0 13.0 25.0 25.0

Bal Bal Bal Bal Bal

Mo: 3.6 Mo: 4.1 Mo: 4.2, N: 0.15 Mo: 4.5, Mn: 1.7 …

(a) Maximum. Cb = Nb

pg 446

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

pg 447

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 447

Table 3 Chemical compositions (wt.%) of cast corrosion- and heat-resistant alloys Alloy

UNS No.

C

Cr

Ni

Fe

Others

CA-15 CA-40 CB-30 CC-50 CE-30 CF-3 CF-3M CF-8 CF-20 CF-8M CF-12M CF-8C CF-16F CG-8M CH-20 CK-20 CN-7M HA HB HC HD HE HF HH HI HK HL HN HT HU HW HX HP THERMAX 40B THERMAX 47 THERMAX 63 THERMAX 63WC THERMALLOY 63W THERMALLOY 63WC (SUPERTHERM) MANAURITE XU MANAURITE 36X MANAURITE 36XS MANAURITE 900 MANAURITE XT MO-RE 1 HOM-3 22-H (NA 22H or HOM-5) SUPER 22-H IN-657

… … … … … … … … … … … … … … … … … … … J92605 J93005 J93403 J92603 J93503 J94003 J94224 J94604 J94213 J94605 J95405 N08001 N06006 J95705 … … … … … …

0.15(a) 0.40(a) 0.30(a) 0.50(a) 0.30(a) 0.03(a) 0.03(a) 0.08(a) 0.20(a) 0.08(a) 0.12(a) 0.08(a) 0.16(a) 0.08(a) 0.20(a) 0.20(a) 0.07 0.2(a) 0.3(a) 0.5(a) 0.5(a) 0.2–0.5 0.2–0.4 0.2–0.5 0.2–0.5 0.2–0.6 0.2–0.6 0.2–0.5 0.35–0.75 0.35–0.75 0.35–0.75 0.35–0.75 0.4 0.4 0.45 0.45 0.45 0.45 0.5

11.5–14.0 11.5–14.0 18–22 26–30 26–30 18–21 18–21 18–21 18–21 18–21 18–21 18–21 18–21 18–21 22–26 23–27 19–22 8–10 18–22 26–30 26–30 26–30 19–23 24–28 26–30 24–28 28–32 19–23 13–17 17–21 10–14 15–19 25 25 25 23 25 25 28

1.0(a) 1.0(a) 2.0(a) 4.0(a) 8–11 8–11 9–12 8–11 8–11 9–12 9–12 9–12 9–12 8–11 12–15 19–22 28–31 … 2(a) 4(a) 4–7 8–11 9–12 11–14 14–18 18–22 18–22 23–27 33–37 37–41 58–62 64–68 35 13 20 35 35 36 36

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

Mo: 0.5(a) Mo: 0.5(a) … … … … Mo: 2.0–3.0 … … Mo: 2.0–3.0 Mo: 2.0–3.0 Cb: 8 × C min, 1.0(a) Mo: 1.5(a), Se: 0.2–0.35 Mo: 3.0 min … … Mo: 2.0–3.0, Cu: 3.0–4.0 … … … … … … … … … … … … … … … … W: 0.5 W: 0.5 W: 0.5 Co: 15.0, W: 5.0 W: 5.0 Co: 15.0, W: 5.0

… … … … … … … … … …

0.35–0.45 0.4 0.4 0.13 0.4 0.45 0.45 0.5 0.5 0.06

24–27 25 25 21 35 26 26 28 28 48

32–36 34 34 33 44 Bal Bal Bal Bal Bal

Bal Bal Bal Bal Bal 33 16 16 13 …

Nb: 0.5, W: 0.5 Cb: 1.2 Cb: 1.5, W: 1.5 Cb: 1.2 Cb: 1.5, W: 1.5 W: 1.6 Mo: 3.0, W: 3.0, Co: 3.0 W: 5.0 Co: 3.0, W: 5.0 Cb: 1.5

(a) Maximum. Cb = Nb

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

448 / High-Temperature Corrosion and Materials Applications

Table 4

Chemical compositions (wt.%) of wrought iron-, nickel-, and cobalt-base alloys

Alloy

UNS No.

INCOLOY 800 INCOLOY 800H INCOLOY 800HT INCOLOY 802 INCOLOY 903 INCOLOY 904 INCOLOY 907 INCOLOY 909 INCOLOY DS RA 330 RA 330HC AC66 SANICRO 28 20Cb-3 20Mo-4 20Mo-6 Nicrofer 3033 (alloy 33) HAYNES HR-120 HAYNES 556 MULTIMENT alloy (N-155) V-57 W-545 DISCALOY PYROMET CTX-1 CHROMEL D KANTHAL Al KANTHAL AF FECRALLOY A Aluchrom S Aluchrom ISE Aluchrom Y Aluchrom O Aluchrom PSI Ni 200 Ni 201 Ni 270 MONEL 400 MONEL 401 MONEL R-405 MONEL K-500 MONEL 450 FERRY alloy CUPRO 107 INCONEL 600 INCONEL 601 Nicrofer 6025HT (602CA) INCONEL 617 INCONEL 625 INCONEL 690 INCONEL 693 INCOLOY 825 INCOLOY 890 INCOLOY 925 INCONEL 706 INCONEL 718 INCONEL X-750 INCONEL 751 INCONEL 671 INCONEL 686 Nicrofer 5923 (59) HAYNES 214 HAYNES 230 HAYNES 242 HAYNES HR-160 Nicrofer 45TM HASTELLOY X HASTELLOY W HASTELLOY S HASTELLOY N HASTELLOY C-22

N08800 N08810 N08811 N08802 N19903 … N19907 N19909 … N08330 … N33228 N08028 N08020 N08024 N08026 R20033 … R30556 R30155

(a) Maximum. Cb = Nb

C

Cr

Ni

0.06 0.05 0.4 0.05 0.01 0.02 0.02 0.02 0.015(a) 0.05 0.1 0.1

21 21 21 21 … … … … 17 19 19 28 27 20 23.5 24 33 25 22 21

32.5 32.5 32.5 32.5 38 32.5 38 38 35 35 35 32 31 33 37 36 32 37 20 20

… K66545 K66220 … … K92500 … … … … … … …N02200 N02201 N02270 N04400 N04401 N04405 N05500 C71500 … … N06600 N06601 N06025

0.8(a) 0.08 0.06 0.03 … … … 0.03 0.08(a) 0.10(a) 0.01–0.1 0.08(a) 0.015–0.03 0.08 0.02(a) 0.01 … … … … … … … 0.08(a) 0.10(a) 0.2

14.8 13.5 14.0 … 18.5 22.0 22.0 15.8 20 20 21 22 22.5 … … … … … … … … … … 15.5 23.0 25

N06617 N06625 N06690 N06693 N08825 N08890 N09925 N09706 N07718 N07750 N07751 … N06686 N06059 … N06230 … N12160 N06045 N06002 N10004 N06635 N10003 N06022

0.07 0.10(a) 0.02 0.2 0.03 0.1 0.01 0.03 0.04 0.04 0.05 0.05 0.01(a) 0.01(a) 0.04 0.1 0.03(a) 0.05 0.1 0.1 0.12(a) 0.02(a) 0.06 0.01(a)

22.0 Bal 12.5 1.5 21.5 Bal … 2.5 29.0 Bal … 9.0 29.0 Bal … 4.0 21.5 Bal … 30.0 25 42.5 … Bal 21.0 Bal … 28.0 16.0 Bal … 37.0 18.0 Bal … 18.5 15.5 Bal … 7.0 15.0 Bal … 7.0 48.0 Bal … … 21.0 Bal … 5.0(a) 23 Bal … 1.5(a) 16.0 Bal … 3.0 22.0 Bal … 3.0(a) 8.0 Bal … … 28.0 Bal 29.0 1.5 27 Bal … 23 22.0 Bal 1.5 18.5 5.0 Bal 2.5 6.0 15.5 Bal … 3.0(a) 7.0 Bal … 5.0(a) 22.0 Bal 2.5(a) 3.0 (continued)

0.05 0.08 0.08 0.4

… … … …

Co

Fe

Mo

W

… … … … 15 14.5 13 13 … … … … … … … … … … 18 20

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

… … … … … … … … … … … … 3.5 2.2 3.8 5.6 1.25 … 3.0 3.0

… … … … … … … … … … … … … … … … … … 2.5 2.5

Al: 0.3, Ti: 0.3 Al: 0.4, Ti: 0.4 Al + Ti: 1.0 … Ti: 1.4, Al: 0.9, Cb: 3.0 Ti: 1.6 Ti: 1.5, Cb: 4.7, Si: 0.15 Ti: 1.5, Cb: 4.7, Si: 0.4 Si: 2.3 Si: 1.2 Si: 1.2 Nb: 0.8, Ce: 0.07 Cu: 1.0 Cu: 3.3, Cb: 0.5 Cu: 1.0, Cb: 0.25 Cu: 3.0 Cu: 0.75, N: 0.5 Cb: 0.7, N: 0.2 Ta: 0.6, La: 0.02, N: 0.2, Zr: 0.02 Cb + Ta: 1.0, N: 0.15

27.0 … 26.0 … 26.0 … 37.7 16.0 36.0 … … … … … … … … … … … … … … … … … 99.6 … 99.6 … 99.98 … Bal … Bal … Bal … Bal … Bal … Bal … Bal … Bal … Bal … Bal …

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal 1.2 0.3 1.2 1.0 0.7 … 0.8 8.0 14.4 10

1.25 1.5 3.0 … … … … … … … … … … … … … … … … … … … … … … …

… … … … … … … … … … … … … … … … … … … … … … … … … …

Al: 0.25, Ti: 3.0, V: 0.5(a), B: 0.01 Al: 0.2, Ti: 2.85, B: 0.05 Al: 0.25, Ti: 1.7 Cb: 3.0, Al: 1.0, Ti: 1.7 Si: 1.5 Al: 5.8 Al: 5.3, Y Al: 4.8, Y: 0.3 Al: 4.5, Zr: 0.3(a) Al: 5.0 Al: 5.5, Zr: 0.01–0.1, Y: 0.05–0.15 Al: 5.5, Zr: 0.3(a) Al: 5.6, Hf: 0.3 … … … Cu: 31.5, Mn: 1.1 Cu: 55.5, Mn: 1.63 Cu: 31.5, Mn: 1.1 Cu: 29.5, Ti: 0.6, Al: 2.7 Cu: 68.0, Mn: 0.7 Cu: 55.0 Cu: 68.0, Mn: 1.1 … Al: 1.4 Al: 2.1, Ti: 0.15, Y: 0.05–0.12, Zr: 0.01–0.1 A1: 1.2 Cb: 3.6 … Al: 2.5–4.0, Nb: 0.5–2.5 Cu: 2.2 Ta: 0.2 Cu: 1.8, Ti: 2.1, Al: 0.3 Ti: 1.8, Al: 0.2, Cb: 2.9 Cb: 5.1 Ti: 2.5, Al: 0.7, Cb: 1.0 Ti: 2.5, Al: 1.1, Cb: 1.0 Ti: 0.35 Ti: 0.02–0.25 … Al: 4.5, Y La: 0.02, B: 0.015(a) … Si: 2.75, Ti: 0.5, Nb: 1.0(a) Si: 2.5–3.0, Ce: 0.03–0.09 … … La: 0.05, B: 0.015(a) … …

9.0 9.0 … … 3.0 1.5 3.0 … 3.0 … … … 16.0 15.5 … 2.0 25.0 … … 9.0 24.0 14.5 16.5 13.0

… … … … … … … … … … … … 4.0 … … 14.0 … … … 0.6 … … … 3.0

Others

pg 448

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

pg 449

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 449

Table 4 (Continued) Alloy

UNS No.

Cr

Ni

Co

Fe

Mo

W

HASTELLOY C-276 HASTELLOY C-2000 HASTELLOY C-4 HASTELLOY C HASTELLOY B HASTELLOY B-2 HASTELLOY B-3 NICROFER 6224 (B-10) HASTELLOY G HASTELLOY G-3 HASTELLOY G-30 HASTELLOY G-35 HASTELLOY G-50 HASTELLOY H-9M RA 333 CHROMEL A (or NICHROME 80) NA 224 NIMONIC 70 NIMONIC 75 NIMONIC 80A NIMONIC 81 NIMONIC 86 NIMONIC 90 NIMONIC 91 NIMONIC 105 NIMONIC 115 NIMONIC 901 NIMONIC AP 1 NIMONIC PE 11 NIMONIC PE 16 NIMONIC PK 31 NIMONIC PK 33 NIMONIC PK 50 NIMONIC PK 37 WASPALOY alloy 263 HAYNES 282 RENÉ 41 RENÉ 95 RENÉ 100 UDIMET 400 UDIMET 500 UDIMET 520 UDIMET 630 UDIMET 700 UDIMET 710 UDIMET 720 UNITEMP AF2-IDA UNITEMP AF2-ID6 ASTROLOY D-979 IN 100 IN 102 IN 587 IN 597 M 252 PYROMET 31 PYROMET 860 REFRACTORY 26 625 PLUS IN 100 GATORIZE HAYNES 188 HAYNES 25 (L-605) HAYNES 150 (UMCo-50) HAYNES 6B S-816 MAR-M 918 MP 35N MP 159 AR 213 ULTIMET alloy

N10276 N06200 N06455 N10002 N10001 N10665 N10675 … N06007 N06985 N06030 N06035 N06950 … N06333 …

0.01(a) 0.01(a) 0.01(a) 0.08(a) 0.05(a) 0.01(a) 0.01(a) 0.01(a) 0.05(a) 0.015(a) 0.03(a) 0.05(a) 0.02(a) 0.03(a) 0.05 …

15.5 23 16.0 15.5 … … 2 8 22.0 22.0 29.5 33 20 22.0 25.0 20.0

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

2.5(a) … 2.0(a) 2.5(a) 2.5(a) … … … … 5.0(a) 5.0(a) … 2.5(a) 5.0(a) 3.0 …

5.5 3.0(a) 3.0(a) 6.0 5.0 2.0(a) 2 7 19.5 19.5 15.0 2.0(a) 17 19.0 18.0 …

16.0 16 15.5 17.0 28.0 28.0 27–32 23 6.5 7.0 5.0 8 9 9.0 3.0 …

4.0 … … 4.0 … … … … … 1.5(a) 2.5 … … 2.0 3.0 …

… … … N07080 … … N07090 … … … … … … … … … … … N07001 … … N07041 … … … … … … … … … … … … N09979 N13100 N06102 … … N07252 N07031 … … N07716 … R30188 R30605 … … R30816 … R30035 … … R31233

0.5

27.0 20.0 19.5 19.5 30.0 25.0 19.5 28.5 15.0 15.0 12.5 15.0 18.0 16.5 20.0 18.0 19.5 19.5 19.0 20.0 19.5 19.0 14.0 9.5 17.5 18.0 19.0 18.0 15.0 18.0 17.9 12.0 12.0 15.0 15.0 10.0 15.0 28.5 24.5 20.0 22.5 13.0 18.0 20.0 12.4 22.0 20.0 27.0 30.0 20.0 20.0 20.0 19.0 19.0 26.0

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal 22.0 10.0 … … 20.0 20.0 35.0 25.5 … 9.0

18.5 25.0 … … … … … … … … 36.0 … 34.0 34.0 … … … … … … 1.5(a) 5.0(a) … … … … … 18.0 … … … … … … 27.0 … 7.0 … … … 15.0 28.9 16.0 5.0 … 3.0(a) 3.0(a) 18.0 … 4.0 … … 9.0 … 3.0

… 6.0 … … … … … … … … 10.0 … … … … … 5.0 … 4.0 … 5.8 … 5.0 … 5.2 … 3.3 … 4.5 … 7.0 … 4.25 … … … 4.3 … 5.8 … 8.5 … 10.0 … 3.5 3.5 3.0 … 4.0 … 4.0 … 6.0 1.0 3.0 3.0 5.2 … 3.0 1.5 3.0 1.3 3.0 6.0 2.7 6.5 5.3 … 4.0 4.0 3.0 … 3.0 3.0 … … 1.5 … 10.0 … 2.0 … 6.0 … 3.2 … 9.0 … 3.2 … … 14.0 … 15.0 … … 1.5(a) 4.5 4.0 4.0 … … 10.0 … 7.0 … … 4.5 5.0 2.0

(a) Maximum. Cb = Nb

C

0.10 0.06 0.03 0.07

…

… …

0.08 0.15

… …

0.05 0.05 0.04

… … …

0.08 0.06 0.06 0.09 0.15 0.16 0.06 0.08 0.05 0.03 0.03 0.07 0.03 0.35 0.04 0.06 0.05 0.15 0.06 0.05 0.05 0.15 0.04 0.05 0.03 0.02 0.07 0.10 0.10 0.06 1.20 0.38 0.05

0.17 0.06

… …

… … … … … … 16.5 20.0 20.0 15.0 … 17.0 … … 14.0 14.0 13.5 16.5 14.0 20.0 10 11.0 8.0 15.0 14.0 18.5 12.0 … 18.5 15.0 14.7 10.0 10.0 17.0 … 15.0 … 20.0 20.0 10.0 … 4.0 20.0 … 18.5 Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

Others

… Cu: 1.6 … … V: 0.03 … … … Cb + Ta: 2.0, Cu: 2.0 Cb + Ta: 0.3, Cu: 2.0 Cu: 2.0 … … … … Si: 1.0 … Al: 1.0, Ti: 1.25, Cb: 1.5 … Al: 1.4, Ti: 2.4 Al: 0.9, Ti: 1.8 Ce: 0.03 Al: 1.5, Ti: 2.5 Al: 1.2, Ti: 2.3 Al: 4.7, Ti: 1.3, B: 0.005 Al: 5.0, Ti: 4.0 Ti: 2.9 Al: 4.0, Ti: 3.5 Al: 0.8, Ti: 2.3 Al: 1.2, Ti: 1.2 Al: 0.4, Ti: 2.35, Cb: 5.0 Al: 2.1, Ti: 2.4 Al: 1.4, Ti: 3.0 Al: 1.5, Ti: 2.5 Al: 1.5, Ti: 3.0, Zr: 0.05, B: 0.006 Al: 0.5, Ti: 2.2 Al: 1.5, Ti: 2.1, B: 0.005 Al: 1.5, Ti: 3.0, B: 0.006 Cb: 3.5, Al: 3.5, Ti: 2.5, Zr: 0.05 Al: 5.5, Ti: 4.2, Zr: 0.06, B: 0.015 Cb: 0.5, Al: 1.5, Ti: 2.5, Zr: 0.06, B: 0.008 Al: 2.9, Ti: 2.9, Zr: 0.05, B: 0.006 Al: 2.0, Ti: 3.0, B: 0.005 Cb: 6.5, Al: 0.5, Ti: 1.0 Al: 5.3, Ti: 3.5, B: 0.03 Al: 2.5, Ti: 5.0 Al: 2.5, Ti: 5.0, Zr: 0.03, B: 0.033 Ta: 1.5, Al: 4.6, Ti: 3.5, Zr: 0.1 Ta: 1.5, Al: 4.0, Ti: 2.8, Zr: 0.1, B: 0.015 Al: 4.0, Ti: 3.5, B: 0.03 Al: 1.0, Ti: 3.0 Al: 5.5, Ti: 4.7, Zr: 0.06, V: 1.0, B: 0.015 Cb: 3.0, Al: 0.4, Ti: 0.6, Mg: 0.02, Zr: 0.03 Cb: 0.7, Al: 1.2, Ti: 2.3, Zr: 0.5 Cb: 1.0, Al: 1.5, Ti: 3.0, Zr: 0.5 Al: 1.0, Ti: 2.6, B: 0.005 Al: 1.4, Ti: 2.3, Cu: 0.9, B: 0.005 Al: 1.0, Ti: 3.0, B: 0.01 Al: 0.2, Ti: 2.6, B: 0.015 Cb: 3.1, Al: 0.2, Ti: 1.3 Al: 5.0, Ti: 4.3, Zr: 0.06, B: 0.02, V: 0.8 La: 0.04 … … … Cb: 4.0 Ta: 7.5, Zr: 0.10 … Cb: 0.6, Al: 0.2, Ti: 3.0 Al: 3.5, Ta: 6.5, Zr: 0.15, Y: 0.1 N: 0.08

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

450 / High-Temperature Corrosion and Materials Applications

Table 5

Chemical compositions (wt.%) of cast nickel- and cobalt-base alloys

Alloy

Cr

Ni

B-1900 IN 100 IN 162 IN 731 IN 738 IN 792 IN 713C IN 713LC M 21 M 22 M 252 MAR-M 200 MAR-M 246 MAR-M 247 MAR-M 421 MAR-M 432 MC 102 RENÉ 77 RENÉ 80 TAZ-8A TAZ-8B: (DS) TRW-NASA-VIA

0.10 0.18 0.12 0.18 0.17 0.12 0.12 0.05 0.13 0.13 0.15 0.15 0.15 0.15 0.15 0.15 0.04 0.07 0.17 0.12 0.12 0.13

C

8.0 10.0 10.0 9.5 16.0 12.4 12.5 12.0 5.7 5.7 20.0 9.0 9.0 8.25 15.8 15.5 20.0 15.0 14.0 6.0 6.0 6.0

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

WAZ-20 (DS) DELORO 40 DELORO 50 DELORO 60 AIRESIST 13 AIRESIST 213 AIRESIST 215 FSX 414 J-1650 MAR-M 302 MAR-M 322 MAR-M 509 MAR-M 918 NASA Co-W-Re S-816 V-36 WI-52 X-40 (HAYNES 31) Stellite 1 Stellite 6 Stellite 12 Stellite 21 Stellite 6K Stellite 704 Stellite 706 Stellite 712 Stellite 706K Tribaloy T-400C Tribaloy T-401

0.20 0.2 0.45 0.7 0.45 0.20 0.35 0.25 0.20 0.85 1.00 0.60 0.05 0.40 0.40 0.27 0.45 0.50 2.45 1.2 1.6 0.25 1.6 1 1.2 1.6 1.6 0.1(a) 0.2

… 7.5 10.5 14.5 21.0 20.0 19.0 29.0 19.0 21.5 21.5 23.5 20.0 3.0 20.0 25.0 21.0 22.0 31 28 29.5 27 31 30 29 29 31 14 17

Bal Bal Bal Bal … 0.5 0.5 10.0 27.0 … … 10.0 20.0 … 20.0 20.0 … 10.0 3(a) 3(a) 3(a) 2.5 3.0(a) 2(a) 3(a) 3(a) 3(a) 1(a) 0.8(a)

Co

Fe

Mo

W

Ta

Zr

Others

10.0 15.0 … 10.0 8.5 9.0 … … … … 10.0 10.0 10.0 10.0 9.5 20.0 … 15.0 9.5 … 5.0 7.5

… … … … … … … … … … … 1.0 … 0.5 … … … … … … … …

6.0 3.0 4.0 2.5 1.7 1.9 4.2 4.5 2.0 2.0 10.0 … 2.5 0.7 2.0 … 6.0 4.2 4.0 4.0 4.0 2.0

… … 2.0 … 2.6 3.8 … … 11.0 11.0 … 12.5 10.0 10.0 3.8 3.0 2.5 … 4.0 4.0 4.0 6.0

4.0 … 2.0 … 1.7 3.9 … … … 3.0 … … 1.5 3.0 … 2.0 0.6 … … 8.0 8.0 9.0

0.10 0.06 0.10 0.06 0.10 0.10 0.10 0.10 0.12 0.60 … 0.05 0.05 0.05 0.05 0.05 … 0.04 0.03 1.0 1.0 0.13

… 2.5 3.5 4.0 … 0.5 0.5 1.0 … 0.5 0.5 … … … 4.0 3.0 2.0 1.5 2.5(a) 3(a) 2.5(a) 3(a) 3(a) 2(a) 3(a) 3(a) 3(a) 1(a) 0.8(a)

… … … … … … … … … … … … … … 4.0 4.0 … … … … … 5.5 … 14 4.5 8.5 4.5 27 22

20.0 … … … 11.0 4.5 4.5 7.5 12.0 10.0 9.0 7.0 … 25.0 4.0 2.0 11.0 7.5 13 4.5 8.5 … 4.5 … … … … … …

… … … … 2.0 6.5 7.5 … 2.0 9.0 4.5 3.5 7.5 … … … … … … … … … … … … … … … …

1.5 … … … … 0.1 0.1 … … 0.2 2.0 0.5 0.1 1.0 … … … … … … … … … … … … … … …

Al: 6.0, Ti: 1.0, B: 0.015 Al: 5.5, Ti: 5.0, B: 0.01, V: 1.0 Al: 6.5, Ti: 1.0, Cb: 1.0, B: 0.02 Al: 5.5, Ti: 4.6, B: 0.015, V: 1.0 Al: 3.4, Ti: 3.4, Cb: 0.9, B: 0.01 Al: 3.1, Ti: 4.5, B: 0.02 Al: 6.0, Ti: 0.8, Cb: 2.0, B: 0.012 Al: 6.0, Ti: 0.6, Cb: 2.0, B: 0.01 Al: 6.0, Cb: 1.5, B: 0.02 Al: 6.3 Al: 1.0, Ti: 2.6, B: 0.005 Al: 5.0, Ti: 2.0, B: 0.015, Cb: 1.0 Al: 5.5, Ti: 1.5, B: 0.015 Al: 5.5, Ti: 1.0, Hf: 1.5, B: 0.015 Al: 4.3, Ti: 1.8, Cb: 2.0, B: 0.015 Al: 2.8, Ti: 4.3, Cb: 2.0, B: 0.015 Cb: 6.0 Al: 4.3, Ti: 3.3, B: 0.015 Al: 3.0, Ti: 5.0, B: 0.015 Al: 6.0, Cb: 2.5, B: 0.004 Al: 6.0, Cb: 1.5, B: 0.004 Al: 5.5, Ti: 1.0, Cb: 0.5, Hf: 0.4, Re: 0.5, B: 0.02 Al: 6.5 Si: 3.5 Si: 4.0, B: 2.0 Si: 2.0–4.5, B: 3.3 Al: 3.4, Y: 0.1 Al: 3.5, Y: 0.1 Al: 4.3, Y: 0.1 B: 0.01 Ti: 3.8, B: 0.02 B: 0.005 Ti: 0.75 Ti: 0.2 … Re: 2.0, Ti: 1.0 Cb: 4.0 Cb: 2.0 Cb + Ta: 2.0 … … … … … … … … … … Si: 2.6 Si: 1.3

… 1.5(a) … … Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

(a) Maximum. Cb = Nb

Table 6

Chemical compositions (wt.%) of oxide-dispersion-strengthened (ODS) alloys

Alloy

INCOLOY MA 956 INCONEL MA 754 INCONEL MA 758 INCONEL MA 6000

C

Ni

Cr

Fe

Ti

Al

Others

… 0.05 0.05 0.05

… Bal Bal Bal

20.0 20.0 30.0 15.0

Bal 1.0 … …

0.5 0.5 0.5 2.5

4.5 0.3 0.3 4.5

Y2O3: 0.5 Y2O3: 0.6 Y2O3: 0.6 Y2O3: 1.1, Mo: 2.0, W: 4.0, Ta: 2.0

pg 450

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

pg 451

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 451

Table 7 Chemical compositions of iron-, nickel- and cobalt-base filler metals Filler Metal

UNS

C

Cr

Ni

Fe

Co

409 409Nb 410 410NiMo 420 430 26-1 630

ER409 ER409Nb ER410 ER410NiMo ER420 ER430 ER26-1 ER630

AWS

S40900 S40940 S41080 S41086 … … … …

0.08(a) 0.08(a) 0.12(a) 0.06(a) 0.25–0.40 0.1(a) 0.01(a) 0.05(a)

10.5–13.5 10.5–13.5 11.5–13.5 11.0–12.5 12.0–14.0 15.5–17.0 25.0–27.5 16.0–16.75

0.6(a) 0.6 (a) 0.6 (a) 4.0–5.0 0.6(a) 0.6(a) … 4.5–5.0

Bal Bal Bal Bal Bal Bal Bal Bal

… … … … … … … …

16-8-2 Nitronic 50W

ER16-8-2 ER209

… S20980

0.1(a) 0.05(a)

14.5–16.5 20.5–24.0

7.5–9.5 9.5–12.0

Bal Bal

… …

Nitronic 60W

ER218

S21880

0.1(a)

16.0–18.0

8.0–9.0

Bal

…

219 240 307 308 308H 308L 308Mo 308LMo 308Si 308LSi 309 309L 309Mo 309LMo 309Si 309LSi 310 312 316 316H 316L 316Si 316LSi 317 317L 321 347 347Si

ER219 ER240 ER307 ER308 ER308H ER308L ER308Mo ER308LMo ER308Si ER308LSi ER309 ER309L ER309Mo ER309LMo ER309Si ER309LSi ER310 ER312 ER316 ER316H ER316L ER316Si ER316LSi ER317 ER317L ER321 ER347 ER347Si

S21980 S24080 S30780 S30880 S30880 S30883 S30882 S30886 S30881 S30888 S30980 S30983 S30982 S30986 S30981 S30988 S31080 S31280 S31680 S31680 S31683 S31681 S31688 S31780 S31783 S32180 S34780 S34788

0.05(a) 0.05(a) 0.04–0.14 0.08(a) 0.04–0.08 0.03(a) 0.08(a) 0.04(a) 0.08(a) 0.03(a) 0.12(a) 0.03(a) 0.12(a) 0.03(a) 0.12(a) 0.03(a) 0.08–0.15 0.15(a) 0.08(a) 0.04–0.08 0.03(a) 0.08(a) 0.03(a) 0.08(a) 0.03(a) 0.08(a) 0.08(a) 0.08(a)

19.0–21.5 17.0–19.0 19.5–22.0 19.5–22.0 19.5–22.0 19.5–22.0 18.0–21.0 18.0–21.0 19.5–22.0 19.5–22.0 23.0–25.0 23.0–25.0 23.0–25.0 23.0–25.0 23.0–25.0 23.0–25.0 25.0–28.0 28.0–32.0 18.0–20.0 18.0–20.0 18.0–20.0 18.0–20.0 18.0–20.0 18.5–20.5 18.5–20.5 18.5–20.5 19.0–21.5 19.0–21.5

5.5–7.0 4.0–6.0 8.0–10.7 9.0–11.0 9.0–11.0 9.0–11.0 9.0–12.0 9.0–12.0 9.0–11.0 9.0–11.0 12.0–14.0 12.0–14.0 12.0–14.0 12.0–14.0 12.0–14.0 12.0–14.0 20.0–22.5 8.0–10.5 11.0–14.0 11.0–14.0 11.0–14.0 11.0–14.0 11.0–14.0 13.0–15.0 13.0–15.0 9.0–10.5 9.0–11.0 9.0–11.0

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

… … … … … … … … … … … … … … … … … … … … … … … … … … … …

318

ER318

S31880

0.08(a)

18.0–20.0

11.0–14.0

Bal

…

320

ER320

N08021 0.07(a)

19.0–21.0

32.0–36.0

Bal

…

320LR

ER320LR

N08022 0.025(a)

19.0–21.0

32.0–36.0

Bal

…

383 904L 2209 2507 253MA RA85H RA330 Marathon 21/33 353MA 33

ER383 ER385 ER2209 … … … … …

N08028 N08904 S39209 S32750 S30815 S30615 N08334 …

0.025(a) 0.025(a) 0.02(a) 0.02(a) 0.07 0.2 0.25 0.12

26.5–28.5 19.5–21.5 22.5 25 21 18.5 19 21.5

30.0–33.0 24.0–26.0 8 10 10 14.5 35 32

Bal Bal Bal Bal Bal Bal Bal Bal

… … … … … … … …

… 0.03 R20033 0.015(a)

28 31.0–35.0

34 30.0–33.0

Bal Bal

… …

556

ER3556

R30556 0.1

22

20

Bal

18

R30155 0.12

21

20

Bal

20

… …

…

MULTIMET (N-155) NOREM 02A NOREM 03A 61 67

… … ERNi-1 ERCuNi

(a) Maximum. Nb = Cb

… …

1.10–1.35 23.0–26.0 0.9–1.2 20.0–23.5

N02061 0.15(a) C71581 0.04(a)

3.7–5.0 4.0–5.0

… 93.0 min … 29.0–32.0 (continued)

Bal Bal

… …

1(a) 0.40–0.70

… …

Others

Ti: 10 × C min/1.5 max Nb: 10 × C min/0.75 max … … … … N: 0.015 Cu: 3.25–4.0, Nb + Ta: 0.15–0.30, Mo: 1.0–2.0 Mo: 1.5–3.0, Mn: 4.0–7.0, N: 0.2 Mn: 8.0, Si: 3.5–4.5, N: 0.08–0.18 Mn: 9.0, N: 0.2 Mn: 10.5–13.5 Mo: 0.5–1.5, Mn: 3.3–4.75 … … … Mo: 2.0–3.0 Mo: 2.0–3.0 Si: 0.65–1.0 Si: 0.65–1.0 … … Mo: 2.0–3.0 Mo: 2.0–3.0 Si: 0.65–1.0 Si: 0.65–1.0 … … Mo: 2.0–3.0 Mo: 2.0–3.0 Mo: 2.0–3.0 Mo: 2.0–3.0 Mo: 2.0–3.0 Mo: 3.0–4.0 Mo: 3.0–4.0 Ti: 9 × C min/1.0 max Nb: 10 × C min/1.0 max Nb: 10 × C min/1.0 max, Si: 0.65–1.0 Mo: 2.0–3.0, Nb: 8 × C min/1.0 max Mo: 2.0–3.0, Cu: 3.0–4.0, Nb: 8 × C min/1.0 max Mo: 2.0–3.0, Cu: 3.0–4.0, Nb: 8 × C min/0.4 max Mo: 3.2–4.2, Cu: 0.7–1.5 Mo: 4.2–5.2, Cu: 1.2–2.0 Mo: 3, N: 0.14 Mo: 4, N: 0.25 Si: 1.6, N: 0.16, Ce: 0.05 Si: 3.7, Al: 1.0 Mn: 5.25 Nb: 1.7 N: 0.15, Ce: 0.03 Mo: 0.5–2.0, Cu: 0.3–1.2, N: 0.35–0.6 Mo: 3, W: 2.5, Ta: 0.6, N: 0.2, Zr: 0.02, La: 0.02 Mo: 3, W: 2.5, Nb + Ta: 1, N: 0.15 Mn: 4.5, Si: 3.3, Mo: 2.0 Mn: 4.0, Si: 2.8, Mo: 1.7–2.2 Ti: 2.0–3.5, Al: 1.5(a) Cu: Bal, Ti: 0.20–0.50

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

452 / High-Temperature Corrosion and Materials Applications

Table 7

(Continued)

Filler Metal

Cr

Ni

Fe

Co

60

ERNiCu-7

AWS

N04060 0.15(a)

…

62.0–69.0

2.5(a)

…

…

ERNiCu-8

N05504 0.25(a)

…

63.0–70.0

2(a)

82 72 76 62 92 52 52M

ERNiCr-3 ERNiCr-4 ERNiCr-6 ERNiCrFe-5 ERNiCrFe-6 ERNiCrFe-7 ERNiCrFe-7A

N06082 N06072 N06076 N06062 N07092 N06052 N06054

18.0–22.0 42.0–46.0 19.0–21.0 14.0–17.0 14.0–17.0 28.0–31.5 28.0–31.5

67.0 min Bal 75.0 min 70.0 min 67.0 min Bal Bal

3(a) 0.5(a) 2(a) 6.0–10.0 8(a) 7.0–11.0 7.0–11.0

…

ERNiCrFe-8

N07069 0.08(a)

14.0–17.0

70.0 min

5.0–9.0

53MD

ERNiCrFeAl-1

N06693 0.15(a)

27.0–31.0

Bal

2.5–6.0

601 602CA

ERNiCrFe-11 …

N06601 0.1(a) N06025 0.2

21.0–25.0 25

58.0–63.0 14 Bal 10

214 65

… ERNiFeCr-1

… 0.05(a) N08065 0.05(a)

16 19.5–23.5

Bal 3 38.0–46.0 22.0 min

718

ERNiFeCr-2

N07718 0.08(a)

17.0–21.0

B N W 242 B-2 … … B-3 G

ERNiMo-1 ERNiMo-2 ERNiMo-3 … ERNiMo-7 ERNiMo-8 ERNiMo-9 ERNiMo-10 ERNiCrMo-1

N10001 N10003 N10004 … N10665 N10008 N10009 N10675 N06007

0.08(a) 1(a) 0.04–0.08 6.0–8.0 0.12(a) 4.0–6.0 0.03(a) 8 0.02(a) 1(a) 0.1(a) 0.5–3.5 0.1(a) … 0.01(a) 1.0–3.0 0.05(a) 21.0–23.5

Bal Bal Bal Bal Bal 60.0 min 65.0 min 65.0 min Bal

4.0–7.0 5(a) 4.0–7.0 2(a) 2(a) 10(a) 5(a) 1.0–3.0 18.0–21.0

X 625 C-276 C-4 S

ERNiCrMo-2 ERNiCrMo-3 ERNiCrMo-4 ERNiCrMo-7 …

N06002 N06625 N10276 N06455 …

0.05–0.15 0.1(a) 0.02(a) 0.015(a) 0.02(a)

20.5–23.0 20.0–23.0 14.5–16.5 14.0–18.0 14.5–17.0

Bal 58.0 min Bal Bal Bal

17.0–20.0 5(a) 4.0–7.0 3.0(a) 3.0(a)

…

ERNiCrMo-8

N06975 0.03(a)

23.0–26.0

47.0–52.0 16

G-3 G-30

ERNiCrMo-9 ERNiCrMo-11

N06985 0.015(a) N06030 0.03(a)

21.0–23.5 28.0–31.5

G-35 RA333 C-22/622 C-22HS 59 686CPT C-2000 725

… … ERNiCrMo-10 … ERNiCrMo-13 ERNiCrMo-14 ERNiCrMo-17 ERNiCrMo-15

N06035 N06333 N06022 … N06059 N06686 N06200 N07725

33 25 20.0–22.5 21 22.0–24.0 19.0–23.0 23 19.0–22.5

617 230-W

ERNiCrCoMo-1 N06617 0.05–0.15 20.0–24.0 ERNiCrWMo-1 N06231 0.05–0.15 20.0–24.0

Bal Bal

3.0(a) 3.0(a)

HR-160 R-41 WASPALOY

ERNiCoCrSi-1 … …

Bal Bal Bal

2.0(a) 5.0(a) 2.0(a)

DELORO 40 DELORO 50 DELORO 60 Colmonoy 88 188 25 (L-605)

… … … … … …

Bal Bal Bal Bal

2.5 3.5 4 3 3.0(a) 3.0(a)

(a) Maximum. Nb = Cb

UNS

C

0.1(a) 0.01–0.10 0.08–0.15 0.08(a) 0.08(a) 0.04(a) 0.04(a)

0.05(a) 0.05 0.015(a) 0.01(a) 0.01(a) 0.01(a) 0.01(a) 0.03(a)

N12160 0.05 28 … 0.05–0.12 19 … 0.08 19 … … … … … …

0.2 0.45 0.7 0.8 0.05–0.15 0.1

7.5 10.5 14.5 15 22 20

Bal

Bal Bal

18.5

18.0–21.0 15

Bal 2.0(a) Bal 17 Bal 2.0–6.0 Bal 2.0(a) Bal 1.5(a) Bal 5.0(a) Bal 3.0(a) 55.0–59.0 7

22 10 (continued)

Others

Cu: Bal, Mn: 4.0(a), Ti: 1.5–3.0, Al: 1.25(a) … Cu: Bal, Ti: 0.25–1.00, Al: 2.0–4.0 … Nb + Ta: 2.5, Mn: 3.0 … Ti: 0.3–1.0 … Ti: 0.15–0.50 … Nb + Ta: 1.5–3.0 … Ti: 3.0, Mn: 2.0–2.7 … Ti: 1.0(a), Al: 1.1(a) … Ti: 1.0(a), Al: 1.1(a), Nb + Ta: 0.50–1.0 … Ti: 2.0–2.75, Al: 0.4–1.0, Nb + Ta: 0.7–1.2 … Al: 2.5–4.0, Nb + Ta: 0.5–2.5, Ti: 1.0(a) … Al: 1.0–1.7 … Al: 2.1, Ti: 0.15, Y: 0.05–0.12, Zr: 0.01–0.1 … Al: 4.5, Y: 0.01 … Mo: 3, Cu: 1.5–3.0, Ti: 0.6–1.2 … Ti: 0.65–1.15, Al: 0.20–0.80, Nb + Ta: 4.75–5.50 2.5(a) Mo: 26.0–30.0, V: 0.3 … Mo: 15.0–18.0 … Mo: 23.0–26.0 … Mo: 25.0 … Mo: 26.0–30.0 … Mo: 18.0–21.0, W: 3.0 … Mo: 19.0–22.0, W: 3.0 … Mo: 27.0–32.0, Mn: 3.0(a) 2.5(a) Mo: 6.5, Cu: 2.0, Nb + Ta: 1.75–2.5, Mn: 1.5 0.5–2.5 Mo: 9.0, W: 0.2–1.0 … Mo: 9.0, Nb + Ta: 3.15–4.15 2.5(a) Mo: 16.0, W: 3.0–4.5 2.0(a) Mo: 16.0 2.0(a) Mo: 14.0–16.5, Al: 0.1–0.5, La: 0.01–0.1 … Mo: 6.0, Ti: 0.7–1.5, Cu: 1.0 5.0(a) Mo: 7.0, W: 1.5(a), Cu: 2.0 5.0(a) Mo: 5.0, W: 1.5–4.0, Nb + Ta: 0.3–1.5, Cu:1.0–2.4 … Mo: 8.0 3 Mo: 3.0, W: 3.0, Mn: 2.5 2.5(a) Mo: 12.5–14.5, W: 2.5–3.5 … Mo: 17.0 … Mo: 16.0, Al: 0.1–0.4 … Mo: 16.0, W: 3.0–4.4 … Mo: 16.0, Cu: 1.6 … Mo: 7.0–9.5, Nb + Ta: 2.75–4.0, Ti: 1.0–1.7 10.0–15.0 Mo: 9.0, Al: 0.8–1.5 5.0(a) W: 14.0, Mo: 2.0, Al: 0.2–0.5 29 Si: 2.75, Ti: 0.5, Nb: 1.0(a) 11 Mo: 10.0, Ti: 3.1, Al: 1.5 13.5 Mo: 4.3, Al: 1.5, Ti: 3.0, Zr: 0.05 1.5(a) Si: 3.5 … Si: 4.0, B: 2.0 … Si: 2.0–4.5, B: 3.3 … W: 17.0, Si: 4.0, B: 3.0 Bal W: 14.0, La: 0.02–0.12 Bal W: 15.0

pg 452

Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/

26/10/2007 1:55PM Plate # 0

pg 453

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 453

Table 7 (Continued) Filler Metal

ULTIMET Stellite 1 Stellite 6 Stellite 12 Stellite 21 Stellite 6K Stellite 704 Stellite 706 Stellite 712 Stellite 706K Tribaloy T-400C Tribaloy T-401 (a) Maximum. Nb = Cb

AWS

UNS

… … … … … … … … … … … …

R31233 … … … … … … … … … … …

C

0.06 2.45 1.2 1.6 0.25 1.6 1 1.2 1.6 1.6 0.1(a) 0.2

Cr

26 31 28 29.5 27 31 30 29 29 31 14 17

Ni

9 3.0(a) 3.0(a) 3.0(a) 2.5(a) 3.0(a) 2.0(a) 3.0(a) 3.0(a) 3.0(a) 1.0(a) 0.8(a)

Fe

Co

Others

3 2.5(a) 3.0(a) 2.5(a) 3.0(a) 3.0(a) 2.0(a) 3.0(a) 3.0(a) 3.0(a) 1.0(a) 0.8(a)

Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal Bal

Mo: 5.0, W: 2.0 W: 13.0 W: 4.5 W: 8.5 W: 5.5 W: 4.5 Mo: 14.0 Mo: 4.5 Mo: 8.5 Mo: 4.5 Mo: 27.0, Si: 2.6 Mo: 22.0, Si: 1.3