Carbohydrates as organic raw materials III: developed from a workshop organized by the Carbohydrate Research Foundation Wageningen, The Netherlands November 28-29, 1994

Carbohydrates as Organic Raw Materials I11 Edited by Herman van Bekkum Harald Roper Fons Voragen

4b

VCH

Dihtribution: VCH, P.O. Box 10 I I 61. 0-69451 Weinheirn (Federal Republic ot' Germany) Switzerland: VCH. P. 0. Box. CH-4070 Bahel (Switzerland) United Kingdom and Ireland: VCH ( U K ) Ltd.. 8 Wellington Court. Cambridge CBI IHZ (England) USA and Canada: VCH. 770 East 73rd Street, New York, N Y 10010-4606 (USA) Japan: VCH. Eikow Building. 10-9 Hongo I-choiiie. Bunkyo-ku. Tokyo 113, (Japan) ISBN 3-577-30079- I

Carbohydrates as Organic Raw Materials I11 Edited by Herman van Bekkum Harald Roper Fons Voragen

Developed from a Workshop organized by the Carbohydrate Research Foundation

*

VCH

Wageningen, The Netherlands November 28-29, 1994

Weinheim . New York - Base1 - Cambridge Tokyo

Editors: Prof. Dr. H. van Bekkum University of Technology Delft Delft The Netherlands

Dr. H. Roper Cerestar Research & Development Vilvoorde Belgium

Prof. Dr. A.G.J. Voragen Agricultural University Wageningen Wageningen The Netherlands

This book was carefully produced. Nevertheless. authors. editors. and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements. data, illustrations, procedural details or other items may inadvertently be inaccurate.

Published jointly by VCH Verlagsgesellschaft mbH. Weinheim (Federal Republic of Germany) VCH Publishers Inc.. New York, NY (USA) Editorial Director: Dr. Hans-Joachim Kraus Production Manager: Dipl.-Wirt.-Ing. (FH) Hans-Jochen Schmitt

The cover illustration shows a MOLCAD based generation of the molecular lipophilicity potential (MLP) of the helix conformations V-amylose. It shows the hydrophilic (blue) and hydrophobic (yellow) surface areas. For further comments see p. 169 ff. of this monograph.

Library of Congress Card No. applied for.

British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library.

Die Deutsche Bibliothek

-

CIP-Einheitsaufnahme

Carbohydrates as organic raw materials : developed from a workshop / organized by the Carbohydrate Research

Foundation. - Weinheim ; New York ; Basel ; Cambridge ; Tokyo : VCH. NE: Carbohydrate Research Foundation

3. Wageningen. The Netherlands. November 28 - 29. 1994. I996 ISBN 3-527-30079-1

0 CRF Carbohydrate Research Foundation, NL-2509 JG The Hague (The Netherlands). 1996

Printed on acid-free and chlorine-free paper. All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form - by photoprinting. microfilm, or any other means - nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks. etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Printing: betz-druck gmbh. D-64291 Darmstadt Bookbinding: GroBbuchbinderei J . Schaffer. D-67269 Grunstadt Printed in the Federal Republic of Germany

Preface Following the successful first and second International Workshop on Carbohydrates as Organic Raw Materials at Darmstadt and Lyon, The Netherlands had the honour and the pleasure to act as the host for the third one (November 1994). Wageningen, internationally known as agro-center because of the presence of the Agricultural University and the Agrotechnological Research Institute (ATO), was selected as the Conference location. Thanks to the presence, ideas and advice of the former Workshop chairmen, Professor F. w . Lichtenthaler and Professor G. Descotes, and of Dr. M. Kunz of Sudzucker in the Scientific Committee we agreed on a programme which, within the framework of the Workshops’ title, contained several new topics which were not dealt with in the preceding Workshops. Thus inulin was on the agenda, a fructan material, which is being manufactured - from chicory - for several years now in Belgium and the Netherlands. We refer to the detailed and optimistic chapter written by Mrs Ir. De Leenheer. Chapters are also devoted to the (animal) disaccharide lactose and to sucrosederived lactic acid. Both compounds are enjoying an expanding tree of derivatives, and companies of the Workshop-host-country are world-market leaders. The Workshop organizers and the Book editors are really happy that many experts from industry were willing to contribute. In fact the industrial contributions amount to over sixty percent, which will further strengthen the reality sense of the text. Industry not only put able people at the Workshop’s disposal but also gave generous financial support. We gratefully acknowledge - in alphabetical order the sponsorship of AVEBE, DMV International, Eridiana BCghin-Say Vilvoorde, Pfeifer & Langen, PURAC-Biochem, Suiker Unie and Unilever. Moreover, the Dutch Innovation Oriented Research Programme on carbohydrates assisted in relieving the financial burden for organizers and participants. Last but not least we mention the efforts of the people of the Carbohydrate Research Foundation, The Hague, notably Dr. F. A. W. Koeman who acted as a central person i n the realization of this Book. The Workshop lectures were converted and extended towards the chapters of this monograph, the third of this type in the VCH series. In addition, one topic not presented at the Workshop has been included because it was considered very relevant to the theme of the Book: metal-catalyzed oxidation and hydrogenation. The delivery time of the chapters differed substantially leading to some delay in appearance of the Book.

Altogether the editors trust that this Book will be of much value to young as well as to experienced people working in the challenging field of upgrading renewable materials, particularly carbohydrates.

February 1995 Delft, Vilvoorde and Wageningen

H. van Bekkum H. Roper A. G. J. Voragen

The Carbohydrate Research Foundation wishes to thank her sponsors of the Third International Workshop on Carbohydrates as Organic Raw Materials.

t

SUIKER UNlE 0 Unileve r

4

-

. ...*

Pfeifer & Langen v

DMV I N 1 E R N A T I 0 N A 1

Cerestar

hy Eridania Bkghin-Sav Montedirnn Group

Vilvnurdc Rcsearrh & Dr\rluptnrnt Centre

Ceresucre Dutch innovation oriented research programme on carbohydrates

IOP-K

This Page Intentionally Left Blank

Contents 1

Utilization of cell wall polysaccharides from cereal by-products and beet pulp.. ...............1 A.G.J. Voragen, M.E.F. Bergmans, A. Oosteweld, H.A. Schols, and G. Beldman

2

Starch: present use and future utilization ........................................................... H. Roper

17

3

Metal-catalyzed oxidation and reduction of carbohydrates....................................... A. Abbadi and H. van Bekkum

37

4

Production and use of inulin: Industrial reality with a promising future.. .................... L. De Lxenheer

.67

5

Lactose: its manufacture and physico-chemical properties....................................... E. Timmermans

93

6

Raw materials for fermentation ................................................................... D. Wilke

.115

7

Lactic acid production and utilization .............................................................. J.A. Van Velthuijsen

129

8

Starch and dextrins in emulsion copolymerization ............................................... M. Bodiger, S. Demharter and R. Miilhaupt

141

9

Synthesis of new 'Saccharide polymers' from unsaturated monosaccharides K. Buchholz, S. Warn, B. Skeries, S. Wick and E-J. Yaacoub

10

Molecular inclusion within polymeric carbohydrate matrices ................................. S. Kubik, 0. Holler, A. Steinert, M. Tolksdor$ Y. Van der Leek, and G. Wulff

.169

11

Resistant starch ..................................................................................... M. Champ and N. Faisant

,189

12

Bulking agents: polydextrose ..................................................................... S.A.S. Craig, J.M. Anderson, J.F. Holden and P.R. Murray

.217

13

Alkyl polyglucoside, a carbohydrate-based surfactant ......................................... W. Ruback and S. Schmidt

.231

...............155

X 14

Contents

Tailor-made carbohydrate surfactants? Systematic investigations into structure-property relationships of N-Acyl N-Alkyl I-Amino-I-Deoxy-D-Glucitols. ........................... .255 H.A. Van Doren

15

Calcium sequestering agents based on carbohydrates.. ........................................ A.C. Besemer and H. Van Bekkum

.273

16

Bleach activators.. ..................................................................................

.295

R.H.F. Beck. H.Koch and J. Mentech Subject Index .........................................................................................

307

1 Utilization of cell wall polysaccharides from cereal by-products and beet pulp

A.G.J. Voragen, M.E.F. Bergmans, A. Oosterveld, H.A. Schols, and G. Beldman Department of Food Science, Section of Food Chemistry and Microbiology Wageningen Agricultural University Bomenweg 2,6703 HD Wageningen, The Netherlands.

Summary. Considerable amounts of by-products are produced annually by the beet-sugar and cereal-processing industries. Figures are given for the Netherlands. Both the current utilization of these by-products and alternative applications are reviewed. Results are presented of studies aimed at better utilization of specific polysaccharides present in wheat bran and beet pulp. Extraction conditions and yields of arabinoxylans with and without feruloyl groups attached and of arabinan-rich highly branched pectins still carrying methoxyl, acetyl and feruloyl groups and without ester groups are described. Structural characteristics, some physical properties and possibilities for enzymatically modifying the chemical structure are discussed.

Introduction Considerable amounts of by-products are produced annually by the beet-sugar and cereal processing industries. These by-products (i.e. beet pulp, cereal brans and spent grains) are rich in cell-wall polysaccharides. For the production of foods, most cereals need to be processed, leaving high amounts of fractions unfit for human consumption as a waste, e.g. wheat bran from the production of flour from wheat and spent grains from the production of beer from barley. The annual production of wheat bran in the Netherlands is approximately 450,000 tons, one third of which is used in whole-wheat products for human consumption. The annual production of rye bran is considerably lower (approximately 2,000 tons) because it is grown in lower amounts and most of the rye is used without milling to flour. Spent grains are produced annually in amounts of 600,000 tons in the Netherlands. A major share of these cereal by-products is used in animal feed. Only in recent years has human consumption of these by-products increased slightly; this can probably be ascribed to the important role of dietary fiber in human nutrition. As a result of the policy of the European Community, cereal prices have decreased, stimulating research on the valorization of these products. Knowledge about the composition of the by-products as well as about the enzymes that

2

A.G.J. Voragen et al. Table I . Composition of wheat bran, barley spent grains and beet pulp (%-w/w)'-~.

total NSP arabinoxylan cellulose P-glucan arabinan pectin other sugars starch protein lignin fat ash not analyzed

wheat bran

spent grains

beet pulp

55

47

77

35 14 4

28 14 2

2 12 15 6 5 4 2

23 21 19 14

3 9 20 4 9 4

6 5 1 12

4

are able to modify components thereof are prerequisites for valorization. Typical compositions of wheat bran',*, barley spent grain3 and beet pulp495 are shown in Table 1. The major constituents of these products are polysaccharides. Arabinoxylans make up 30 to 35% of the dry matter; cellulose represents 14-16%, and minor amounts of mixed-linkage P-glucans are present. The protein content varies between 15 and 20% and the lignin content between 4 and 6%. Potential applications of cereal by-products or components thereof are summarized in Table 26-11. They are used 'as is' in animal feed or fermentation media; after some refining treatments they are used as food ingredients or dietary fiber. Arabinoxylans are known to form viscous solutions due to their rod-like conformation in solution1*. A schematic structure of arabinoxylans with the typical glycosidic linkages is shown in Fig. 1. Arabinoxylans are also believed to play an important role in dough development and loaf properties, as a result of their water-binding capacity'. Ferulic acid esterified to some of the arabinose branches is reported to be involved in increasing viscosity or even gelation in wheat flour arabinoxylans as a result of oxidative cross-linkingl3. These physical properties make valorization of the arabinoxylans interesting as bulking, water binding, thickening, gelling, emulsifying or foam stabilizing agents. Oligomers derived from arabinoxylans might find applications as a functional ingredient in view of alleged beneficial physiological activitiesg. Feruloyl containing oligomers are being studied for anti-oxidative propertieslo. By-products can also be enzymatically saccharified to monomers and subsequently fermented to e.g. ethanolll. Ferulic acid released from brans may be converted to

Utilization of cell wall polysaccharides

3

Table 2. Alternative utilization of wheat bran and sugar beet p u l ~ ~ - " . ' ~ - * ~ .

By-product Wheat bran: Feed ingredient

Polymer

Oligomer

Monomer

Water binding

Physiological

Ethanol

ingredient Food ingredient Thickener

Anti-oxidative Vanillin ingredient

Fiber

Gelling agent

Fermentation

Emulsifier Foam stabilizing

Beet pulp

Feed ingredient

Water binding

Ethanol

Fiber

Thickener

Vanillin

Fermentation

Emulsifier

Sugars

Gelling agent Foam stabilizing Fat replacer vanillin. The economic feasibility and market opportunities for many of these products are questionable. The annual production of beet pulp in the Netherlands amounts to 1.3.106 tons (22% dry weight). The pulp obtained after extraction and pressing is either dried to ca. 89% dry matter or fermented at a dry matter content >22% to prolong its storage life. Ensilage is the classical way to

-@ 0-(1,4)-Xyl I-

A

a-GlcA linked to C(0)2 of Xyl

d

a-Ara linked to C(0)3 of Xyl

4

ferulic acid esterified to Ara

a-Ara linked to C(0)2 of Xyl

0

acetyl esterified to Xyl

Fig. 1. Hypothetical structure of wheat bran arahinoxylans.

A.G.J. Voragen et al.

4

improve both storage life and digestibility. In the initial stage of the fermentation, Bacilli will decrease the pH to below 5 and make the system anaerobic. Lactobacilli will then take over and convert residual sugar into lactic acid, further lowering the pH (<4)and thus preventing other micro-organisms from growing. The dried as well as the fermented products are almost exclusively used in the feed industry. The fermentation of sugar beet pulp with specific fungi results in a protein enriched and better digestible product suitable as feed for monogastric animaIsl4. There are small markets for beet pulp as dietary fiber and for pectins extracted from beet pulp. In addition, several other alternative applications are proposed as shown in Table 2I4-I0.The high amounts of soluble and insoluble polysaccharides make sugar beet pulp an interesting source of dietary fiber. The high water holding capacity makes the product suitable for use in bakery products, such as fibre enriched breads, soft cookies and muffins4. It has been shown that sugar beet fiber is an effective liver and serum cholesterol-lowering agent in rats21. Auffret et aL2' studied the effect of chemical treatment of sugar beet fiber on colonic fermentation and found that mild alkali and mild acid treatments resulted in increased hydration properties and fermentability. Pectins are generally extracted with mild acid treatment15~23.Phatak et al.23 obtained pectins containing 82% galacturonic acid and 10%neutral sugars of which galactose was prevailing. The resulting pectins gave low viscosity solutions from which no gels could be formed. This is attributed to the high content of acetyl groups and the relatively low molecular weight24. Dea and Madden16 isolated acetylated pectins using several extraction sequences. Many of the extracts were surface active and showed foaming and emulsifying properties. The yields, however, were quite low.

'J Galacturonic acid

0

Arabinose

0 Galactose Rharnnose

. @

Acetyl Methyl Ferulic acid

Fig. 2: Hypothetical structure of sugar beet pectin.

Utilization of cell wall polysaccharides

5

A special structural feature of sugar beet pectin is the presence of ferulic acid groups attached to arabinofuranosyl and galactopyranosyl residues in side chains of the rhamnogalacturonan backbone. A schematic structure with typical glycosidic linkages is shown in Fig. 2 . The ferulic acid groups can be cross-linked with peroxidase/H202 or ammonium persulfate. In Fig. 3 one type of ferulic acid dimer produced by oxidative coupling is shown. Recently, the formation of many different types of ferulic acid dimers by oxidative coupling has been demonstrated25. The coupling leads to an increase in viscosity or the formation of a gel. Solvent drying of these gels results in a product with a remarkable water absorbing capacity. It is suggested that this product can be used as a cloud stabilizer in drinks or as water absorbing agent in sanitary products'5. Arabinans can be extracted from beet pulp using hot alkaline conditionsl7. Enzymatic debranching of the arabinans with arabinofuranosidase results in a product that can be used as a fat replacer in for instance low fat spreads, ice cream and chilled and frozen desserts. Non-food areas in which this product can be applied are cosmetics, toiletries and pharmaceuticals. Alternative utilization of sugar beet pulp also includes saccharification of the polysaccharides and the use of the resulting monosaccharides as (feedstock for) fine chemicals. Several studies were performed on the enzymatic hydrolysis of sugar beet polysaccharides. Beldman et al.19 were able to hydrolyze the cellulose, pectin, arabinan and galactan fractions for ca. 90%. In another

0-

d

O=

AH

dH

II

CH

c

+ H202

Peroxidase

H3co

II

CH

+ 2 H20

H,CO

e, b I Pectin L O

CH

II I P-0 CH

I

P Pectin

Fig. 3: Possible mechanism for the oxidative cross-linking of sugar beet pectin

6

A.G.J. Voragen et al.

study Thibault and Rouau20 achieved ca. 80% hydrolysis of the polysaccharides present in beet pulp. The monosaccharides obtained from hydrolysis of sugar beet polysaccharides can be used as feedstock for the production of (bio)ethanol, lactic acid and galactaric acid (as sequestering agent) or galactaric and galacturonic acid based surfactants (alkyl esterslethers). Micard et al. '8 mentioned the possible use of ferulic acid from sugar beet pectins as a precursor for the production of vanillin. Sponsored by the Dutch Innovative Oriented Research Program on Carbohydrates, we are carrying out a research program aimed at better utilization of polysaccharides found in cereal byproducts and beet pulp. We are particularly interested in the relationship between chemical structure and physical properties and we want to use enzymes to customize the polysaccharide molecules. The major group of polysaccharides present in cereal by-products such as wheat bran and spent grains are arabinoxylans. Methods for the selective extraction of arabinoxylans either with or without feruloyl groups attached were developed. Structural characteristics, physical properties and possibilities to enzymatically modify the chemical structure are discussed below. For sugar beet pulp, the research mainly focusses on pectins and pectin related arabinans, which represent approximately 40% of the dry weight of the pulp. Extraction procedures for highly branched pectins still carrying methoxyl, acetyl and feruloyl groups and for pectins rich in neutral sugar side chains are described. Structural characteristics, physical parameters and possibilities for enzymatically modifying these pectins are also discussed for these fractions.

Arabinoxylans from wheat bran Extraction procedure In cereal cell walls, the polymers can be cross-linked by covalent linkages, ionic interactions and hydrogen bonds. The most prevalent covalent cross-links are ester and ether linkage+. From the presence of the different types of linkages between the cell wall polymers, it can be inferred that the extraction of a single pure component is rather difficult. However, the use of alkaline solutions for the extraction of non-cellulosic polysaccharides has been known for a long time, because alkali can disrupt hydrogen bonds and ester linkage~2~. Recently, a saturated barium hydroxide solution was introduced as a highly selective extractant for arabinoxylans from cell wall material of wheat flour28. Prior to extraction of the arabinoxylans fat, protein and adhering starch must be removed in order to obtain water-unextractable cell wall material (WUS), which consists predominantly of cell wall polysaccharides and lignin29.

Utilization of cell wall polysaccharides

7

Arabinoxylans without esterified ferulic acid Extraction yield A first aim in the investigation of the relationship between the physical and chemical properties of arabinoxylans from wheat bran was to obtain a high yield of arabinoxylans of high purity with a single e ~ t r a c t i o nSaturated ~~. barium hydroxide, with the addition of 260mM sodium borohydride, extracted 29% of the arabinoxylans originally present in the WUS of the wheat bran. Arabinoxylans represented 85% (w/w) of the total extract; the remaining 15% consisted of coextracted P-glucan, protein and lignin. By the use of delignification pretreatments of the WUS, it was attempted to increase the yield of arabinoxylans. Both a chlorite and an alkaline peroxide treatment showed a decrease of the lignin content of the WUS. As a result of a simultaneous loss of polysaccharides caused by the pretreatments, the final yields of the extracts showed no improvement compared to the yield obtained from the untreated WUS. A second attempt to increase the yield of the extract was made by changing the conditions of extraction. In a first series of extractions, only the temperature of extraction was changed. An increase in the extraction temperature from 20 "C to 95 "C resulted in an increase of the arabinoxylan yield from 29% to 50%. The purity of the extract decreased somewhat, probably due to an increase of the coextraction of P-glucan and lignin. Yields and compositions of two of these extracts are given in Table 3. In a second series of extractions, the concentration of the barium hydroxide was increased by saturating the solution at the extraction temperature. No further improvements in the yield were observed until 70°C. Above this temperature the extraction yield decreased. Substitution of barium hydroxide by calcium hydroxide resulted in lower yields and a lower selectivity (Table 3), most probably caused by its lower solubility compared to that of barium hydroxide.

Chemical and physical characteristics The intrinsic viscosities of some of the crude extracts are shown in Table 3 ; these range from 1.2-1.5 dL/g. Girhammar and Nai@ reported intrinsic viscosities of 1.7 and 5.9 dL/g for watersoluble whole wheat and rye extracts, respectively. The arabinoxylan content of those fractions was very similar to the extracts of wheat bran. The main difference appeared to be the higher molecular weight of the rye extract. The weight average molecular weights reported were 255 kDa for the wheat extract and 770,kDa for the rye extract. The weight average molecular weights determined for the wheat bran extracts were approximately 500 kDa. Other factors, such as ArdXyl-ratio, conformation and homogeneity, may also be of importance. Intrinsic viscosities of polysaccharides used in the food industry roughly range from 1.2 dL/g for gum arabic to 30 dL/g for alginate31.

A.G.J. Voragen et al.

8

Table 3. The yield and composition of arabinoxylan extracts obtained from wheat bran WUS by extraction at 20 and 70 "C with barium hydroxide saturated at room temperature (BE.20 and BE.70) or calcium hydroxide saturated at extraction temperature (CE.20 and CE.70)

wus

CE.20

CE.70

BE.20

BE.70

12 16

24 31

19 29

31 43

5.0 78.9

5.9 82.4

2.9 79.2

2.9 87.0

1.4 83.7

XYl Man Gal Glc AUA

27.1 40.7 0.2 1.1 27.0 3.3

42.8 46.9 0.4 1.5 6.7 1.8

41.8 46.0 0.1 1.6 8.2 2.4

41.1 53.2 0 0.9 0.8 3.9

39.6 48.3 0.2 1.7 5.8 4.5

Ara/Xyl

0.7

0.9

0.9

0.8

0.8

1.4

1.2

1.5

yield DMa yield AXb proteinc total sugarc molar sugar compositiond Ara

a expressed as weight percentage (dm) of WUS.

[Ara+Xyl+AUA of extract (mg)]/[Ara+Xyl+AUA of WUS (mg)] x 100%. expressed as weight percentage (dm) of each fraction. expressed as percentage (mole per 100 mole). expressed in d u g .

From further studies on the extract obtained at 70 "C by barium hydroxide saturated at room temperature, the wheat bran arabinoxylans appeared to be very heterogeneous. Fractionation by graded ethanol precipitation and anion-exchange chromatography showed that approximately 40% of the arabinoxylans in the extract have a very low degree of substitution (ArdXyl = 0.2), with only minor amounts of uronic acid. About 50% of the arabinoxylans showed a high degree of substitution (AraKyl = 1.0) with approximately 4% (w/w) uronic acid. Half of the remaining 10% of arabinoxylans was found to have an intermediate degree of substitution (Ara/Xyl = O S ) , while the other half had a high degree of substitution (Ara/Xyl = 1.3) and appeared to be associated with lignin. The physical properties of the subfractions of the arabinoxylan extract are presently under further study.

Utilization of cell wall polysaccharides

9

Arabinoxylans with esterified ferulic acid A second aim in the investigation of the relationship between physical and chemical properties of arabinoxylans from wheat bran is the extraction of arabinoxylans with the preservation of ferulic acid esterified to the arabinose branches. The objective was to investigate the gelling properties of these arabinoxylans. By alterations in the extraction conditions, such as a low concentration of alkali and a short time of extraction, an extract could be obtained that consisted for 70% (w/w) of arabinoxylans. The yield of the extract was only low: approximately 5% of the arabinoxylans originally present in the WUS. The extract, however, contained 0.2% (w/w) ferulic acid. The ArdXyl-ratio of the extract was 0.9. Gelation studies with the use of ammonium persulfate and hydrogen peroxide in combination with peroxidase are currently under investigation and the first results appear to be promising.

Enzymatic modification of arabinoxylans In order to find relationships between structure and physical properties of arabinoxylans, it would be favorable to specifically change the structure of the arabinoxylans, e.g. by removal of arabinose or glucuronic acid branches. Most of the known arabinofuranosidases do not have a high activity Type A

4

a-Ara linked to C ( 0 ) 3 of Xyl

@

0-(1,4)-Xyl

a-Ara linked to C ( 0 ) 2 of Xyl

A

a-GlcA linked to C ( 0 ) 2of Xyl

site of attack of AXH

Fig. 4 : Sites of attack of AXH on type A and type B arabinoxylan from wheat bran

10

A.G.J. Vorugen et ul.

towards arabinoxylans. However, one enzyme - arabinoxylan arabinofuranohydrolase (AXH) was reported to have high activity towards arabinose present as single substitution at the C(0)3position of xylose from wheat flour arabin0xylan3~.Until now no enzyme has been reported to be able to hydrolyze arabinoses occurring as double substitution at the C(0)2 and C(0)3-positions of a single xylose residue in the backbone. From the fractionation studies described above, the wheat bran arabinoxylans appear to consist of very lowly and very highly substituted arabinoxylans. In Fig. 4 the lowly and highly substituted arabinoxylan extracted by barium hydroxide, type A and B respectively, are shown. The arabinosyl linkages that can be hydrolyzed by the highly specific AXH are also shown. It appeared that almost all the arabinose present in the type A arabinoxylan can be removed, but only a maximum of approximately 10% of the arabinose present in type B arabinoxylan. Modification of the lowly substituted arabinoxylan is only of minor interest, as aggregation was observed to occur when arabinose is removed. Enzymatic modification of the highly substituted arabinoxylan proves to be rather difficult because no enzymes are as yet known to be able to remove arabinose from doubly substituted xylose residues.

Polysaccharides from beet pulp Extraction procedure Several extraction methods have been used for the extraction of pectins from sugar beet pulp. Acid extraction, used for the commercial extraction of pectins23.33, results in the degradation of the arabinan side chains caused by acid hydrolysis34. Arabinans or arabinan rich pectic substances are generally extracted using hot/strong alkali17.35. This results in p-eliminative degradation of the galacturonic acid backbone. Both methods, however, lead to a loss of ferulic acid and are therefore not favorable.

Alkali extracted pectic polysaccharides Since our aim was to obtain pectic polysaccharides rich in arabinose and ferulic acid, it was not possible to start with acid or strong alkali extractions. We therefore used an extraction procedure starting with a mild alkali extraction (0.15 M NaOH, 0.05 M EDTA, two times, 25 "C, 2 hours). This was followed by an extraction using strong alkali to remove residual arabinose-rich polysaccharides (4M NaOH, 0.02 M NaBH4,80 "C, 2 h0urs)3~.The yields and compositions of this extraction method (A) are shown in Table 4. The two mild alkali extractions yielded a total of 15.1 % of the dry matter. The first extract consisted mainly of galacturonic acid, although considerable amounts of neutral sugars were also present. In the second extract, arabinose,

Utilization of cell wall polysuccharides

11

Tabel 4. Yield and composition of pectic polysaccharides obtained by mild alkali (MA), strong alkali (SA) and autoclave (Auto) extraction.

Method A MA1

MA2

SA

Auto 1

Auto2

SA

3.6 37.7 0.0 0.0 5.0 0.1 53.6

5.3 59.5 0.0 0.0 7.4 0.0 27.8

4.7 53.7 5.5 2.7 0.9 5.7 6.8

2.6 29.8 0.4 0.4 3.9 1.1 61.9

3.6 60.8 0.0 0.4 6.6 0.9 27.7

6.2 57.3 6.2 2.4 11.0 5.1 11.7

Ac~ Me2 Fer A2

0.00

0.00

0.00

0.00

0.18

0.00 0.09

0.00 0.03

7.00 6.00 0.39

5.40 3.20 1.05

0.00 0.00 0.04

Sugar content2 Yield3

79.7 9.7

68.8 5.4

61.4 23.3

69.2 10.8

82.9 5.4

63.4 19.0

Rha Ara

XYl Mall Gal Glu AUA

I

Method B

: Expressed as mol percentage sugar per total sugar content : Expressed as percentage dry weight of each extract : Expressed as percentage dry weight of beet pulp

galactose and rhamnose were present in higher amounts. Strong alkali extraction yielded another 23.3 %. The extract contained a small amount of galacturonic acid. In addition to arabinose as the main sugar in this extract, xylose, mannose and glucose were also present. The amounts of acetyl and methyl esters were very low as a result of the alkali conditions. However, some ferulic acid was still present in these extracts, indicating the stronger resistance of feruloyl esters towards alkali conditions as compared to acetyl and methyl esters. High performance size exclusion chromatography (HPSEC) showed the presence of three populations in the mild alkali extracts, with molecular weights of approximately 100 kDa, 60 kDa and 25 kDa, respectively. In the literature molecular weights of sugar beet pectins between 15 kDa and 48 kDa are menti0ned16~23.37.The two high molecular weight populations prevailed in the second extract and contained almost all ferulic acid. Two populations with molecular weights of approximately 35 kDa and 15 kDa were present in the extract obtained with strong alkali, of which the first population predominated. Methylation analysis (data not shown) showed that arabinose was mainly found as terminally, 1,3-linked and 1,3,5-linked residues in all extracts. Rhamnose was 1,2-linked and 1,2,4-linked. Galactose was mainly present as terminally linked and 1,4-linked residues.

12

A.G.J. Voragen et al.

Galacturonic acid was found to be mainly 1,4-linked. These data were in agreement with the sugar linkage composition from acid and alkali extracted sugar beet pectins". DEAE chromatography showed that 'smooth' and 'hairy' regions eluted separately. These results indicate that all extracts consist of rhamnogalacturonans highly branched with arabinose side chains and of homogalacturonans, present as a separate population.

Autoclave extracted pectic polysaccharides An autoclave extraction (two times, 121 T , 40 minutes) was used to obtain pectic polysccharides rich in ferulic acid (Method B). Similar to the previous method, this was followed by a strong alkali extraction (4 M NaOH, 0.02 M NaBH4, 80 "C, 2 hours)36. The autoclave extracts yielded 16.2 % of the dry matter of beet pulp. The first extract consisted mainly of galacturonic acid. However, the amount of neutral sugars and especially arabinose was relatively high as compared to acid extracted pectins23.33. In the second extract the neutral sugars predominated. Using strong alkali a yield of 19.0 % was obtained. The sugar composition of this extract of method B was similar to the corresponding extract obtained by method A. The autoclave extracts contained considerable amounts of acetyl and methyl esters. The ferulic acid content was relatively high in these extracts, especially in the second autoclave extract. For both extracts three populations could be recognized by HPSEC with molecular weights of approximately 100 kDa, 45 kDa and 20 kDa (Fig. 5). All ferulic acid was present in the two high molecular weight populations. These

Cross-linked

20

25

30

35

40

Time (min)

pi=E] Fig. 5 : HPSEC results of the second autoclave extract before and after cross-linking

Utilization of cell wall polysaccharides

13

populations were present in higher amounts in the second extract. The HPSEC elution pattern of the strong alkali extract was similar to the corresponding extract in method A. Methylation analysis and DEAE chromatography indicated that the high molecular weight populations consisted of rhamnogalacturonans with highly branched side chains. The low molecular weight population contained mainly homogalacturonans, although a considerable amount of neutral sugars was still present. Since both autoclave extracts contained relatively high amounts of ferulic acid, further study focussed on the oxidative cross-linking of these extracts (Fig. 5). Viscosity measurements of the second extract showed a clear increase in the relative viscosity when the cross-linking reaction took place at a polysaccharide concentration higher than 0.5%. Above a concentration of 1.5% a gel was formed. Preparative size exclusion chromatography of the cross-linked extract showed the presence of two populations. The high molecular weight population consisted solely of 'hairy' regions. Almost all ferulic acid was found in this population. The 'smooth' regions remained unchanged.

Enzymatic modification of the cross-linked extracts Specific enzymes are appropriate tools for altering the structure of the pectins, in order to improve or change the physical properties of pectins. Several enzymes that can be used to modify beet pectins are discussed below. Pectin methyl esterase can remove the methyl esters from the homogalacturonan~~~. In this way high-methoxyl pectins, which are able to form acid-sugar gels, can be converted to calcium sensitive low-methoxyl pectins. The acetyl groups from sugar beet pectin can be removed using pectin acetyl esterase and rhamnogalacturonan acetyl esterase. Pectin acetyl esterase can remove the acetyl groups from the 'smooth' regions40 and may lead to an improvement of the gelation properties of the pectin, since it is known that the high acetyl content is an important factor causing its poor gelation pr0perties2~. Rhamnogalacturonan acetyl esterase can be used to specifically remove the acetyl groups from the rhamnogalacturonan backbone in the 'hairy' region#'. Arabinanases and galactanases can be used to remove the arabino(ga1actan) side chains from the 'hairy' regions. This may alter the hydration properties of the polysaccharide42. Guillon and Thibault43 suggested the use of arabinofuranosidase B to improve the accessibility of the feruloyl groups prior to oxidative cross-linking. Although many enzymes are known to be active towards sugar beet pectins, little is known about the effects of these enzymes on the physical-chemical properties. Part of the present literature also leads to contradictory conclusions4*. However, more knowledge on the subject might lead to a better understanding of and better control over the physical-chemical properties of pectins.

14

A.G.J. Voragen et al.

Conclusions and future research Using adapted extraction procedures, we were able to obtain specific polysaccharide fractions from wheat bran and beet pulp. By a single extraction step approximately 50% of the arabinoxylans originally present in the pre-purified cell wall material of wheat bran (WUS) could be isolated. Due to the alkaline conditions, the high extraction temperature and the long extraction time, the extracted arabinoxylans did not contain ferulic acid. By a short, mild alkali extraction an arabinoxylan fraction with ferulic acid esterified to the arabinose branches could be obtained, although in rather low yield. These arabinoxylans could be cross-linked by oxidative coupling, resulting in highly viscous or gelled systems. Investigations on viscosity and other physical properties such as gelation, foam and emulsion stabilization and enzymatic modification of the arabinoxylan extracts are in progress. As no enzymes are known to be able to release high amounts of arabinose from highly substituted arabinoxylans, chemical ways of modification will also be studied. Arabinose-rich pectic polysaccharides fractions could be obtained from beet pulp. Alkali extraction resulted in highly branched rhamnogalacturonans, with some ferulic acid still present, together with homogalacturonans, which did not contain any ester groups. Autoclave extraction yielded highly branched rhamnogalacturonans rich in ferulic acid. Homogalacturonans were also present, which still contained considerable amounts of methyl and acetyl groups. Oxidative crosslinking of this material at concentrations as high as 1.5% resulted in gel formation. It was shown that only the 'hairy' regions participated in the cross-linking. Current research is focussed on elucidation of the fine structure of the pectin extracts and on the effects of specific enzymatic modification on physical-chemical properties. The mechanism of oxidative cross-linking, using H 2 0 2 and peroxidase, will also be studied.

References I.

2. 3.

4.

5.

J.W. Lee, N.L. Stenvert: Conditioning studies on Australian wheat. IV. Compositional variations in the bran layers of wheat and their relation to milling. J. Sci. Food Agric. 24 (1973) 1565-1569. M.C. Ralet, J.F. Thibault, G. Della Valle: Influence of extrusion-cooking on the physico-chemical properties of wheat bran. J. Cereal Sci. 11 (1990) 249-259. R.J. Vietor, A.G.J. Voragen, S.A.G.F. Angelino: Composition of non-starch polysaccharides in wort and spent grains from brewing trials with malt from a good malting quality barley and a feed barley. J . Insf. Brew. 99 (1993) 243-248. E.H.Christensen: Characteristics of sugar beet fiber allow many food uses. Cereal Foods World 34 (1989) 54 1-544. L.F. Wen, K.C. Chang, G. Brown, D.D. Gallaher: Isolation and characterization of hemicellulose and cellulose from sugar beet pulp. J . Food Sci. 53 (1988) 826-829.

Utilization of cell wall polysaccharides 6. 7. 8. 9. 10.

II.

12. 13. 14.

IS.

16.

17. 18. 19.

20.

21.

R.M. Saunders, M.A. Connor, R.H. Edwards, G.O. Kohler: Enzymatic processing of wheat bran: effects on nutrient availability. Cereal Chem. 49 (1972) 436-442. M.S. Izydorczyk, C.G. Biliaderis; Influence of structure on the physicochemical properties of wheat arabinoxylan. Carbohydr. Polym. 17 (1992) 237-247. E. Chanliaud: Extraction, caractirisation et propriitis fonctionelles des hitiroxylanes de son de mais. Thesis, UniversitC de Paris et E.N.S.I.A. (1995). H.Yamada, K. Shiiba, H.Hara, N. Ishida, T. Sasaki: Preparation of a new arabinoxylooligosaccharide from wheat bran hemicellulose and its structure. Biosci. Biotechnol. Biochem. 58 (1994) 288-292. T. Ohta, S. Yamasaki, Y. Egashira, H.Sanada: Antioxidative activity of corn bran hemicellulose fragments. 1. Agric. Food Chem. 12 (1994) 653-656. S.L. Rosenberg: Fermentation of pentose sugars to ethanol and other neutral products by micro-organisms. Enzyme Microb. Technol. 2 (1980) 185-193. K.A. Andrewartha, D.R. Philips, B.A. Stone: Solution properties of wheat flour arabinoxylans and enzymically modified arabinoxylans. Carbohydr. Res. 77 ( I 979) 191-204. T. Geissmann, H. Neukom: On the composition of the water soluble wheat flour pentosans and their oxidative gelation. Lebensm: Wiss. 14. Technol. 6 (1973) 59-62. M. Vogel: Alternative utilization of sugar beet pulp. Zuckerind. 116 (1991) 265-270. F.M. Rombouts, J.F. Thibault: Sugar beet pectins: chemical structure and gelation through oxidative coupling. Chemist0 and function of pectins (M.L. Fishman, J.J. Jen, Eds.), ACS Symp. Ser. 310, American Chemical Society, Washington, DC ( 1986) 49-60. I.C.M. Dea, J.K. Madden: Acetylated pectic polysaccharides of sugar beet. Food Hydrocolloids 1 (1986) 7 1-88, B.V. McCleary, J.M. Cooper, E.L. Williams: Pat. Application, GB 8828380.9 (1989). V. Micard, C.M.G.C. Renard, J.F. Thibault: Studies on enzymic release of ferulic acid from sugar beet pulp. Lebensm. - Wiss. u. Technol. 27 (1994) 59-66. G. Beldman, F.M. Rombouts, A.G.J. Voragen, W. Pilnik: Application of cellulase and pectinase from fungal origin for the liquefaction and saccharification of biomass. Enzyme Microb. Technol. 6 (1984) 503-507. J.F. Thibault, X. Rouau: Studies on enzymic hydrolysis of polysaccharides in sugar beet pulp. Carbohydr. Polym. 13 (1990) 1-16. C.F. Klopfenstein: Nutritional properties of coarse and fine sugar beet fiber and hard red wheat bran. I. Effects on rat serum and liver cholesterol and triglycerides and on fecal characteristics. Cereal Chem. 67 (1990)

26.

538-54 I . A. Auffret, J.L. Barry, J.F. Thibault: Effect of chemical treatments of sugar beet fiber on their physicochemical properties and on their in-vitro fermentation. J . Sci. Food Agric. 61 (1993) 195-203. L. Phatak, K.C. Chang, G. Brown: Isolation and characterization of pectin in sugar beet Pulp. J . Food Sci. 53 (1988) 830-833. E.L. Pippen, R.M. McCready, H.S. Owens: Gelation properties of partially acetylated pectins. J . Am. Chem. Soc. 72 (1950) 813-816. J. Ralph, S. Quideau, J.H.Grabber, R.D. Hatfield: Identification and synthesis of new ferulic acid dehydrodimers present in grass cell walls. 1.Chem. Soc. Perkin Trans. I (1994) 3485-3498. K. Iiyama, T.B.T. Lam, B.A. Stone: Covalent cross-links in the cell wall. Plant Physiol. 104 (1994)

27.

3 15-320. S.C. Fry: The growing plant cell wall: chemical and metabolic analysis. Longman, Harlow, (1988).

22. 23. 24. 25.

15

16 28. 29.

30. 31. 32.

A.G.J. Voragen et al. H. Gruppen, R.J. Hamer, A.G.J. Voragen: Barium hydroxide as a tool to extract pure arabinoxylans from water-insoluble cell wall material from wheat flour. J . Cereal Sci. 13 (1991) 275-290. M.E.F. Bergmans, G. Beldman, H. Gruppen, A.G.J. Voragen: Optimization of the selective extraction of (g1ucurono)arabinoxylansfrom wheat bran: use of barium and calcium hydroxide solutions at elevated temperatures. J. Cereal Sci. accepted for publication. U. Girhammar, B.M. Nair: Certain physical properties of water soluble non-starch polysaccharides from wheat, rye, triticale, barley and oats. Food Hydrocolloids 6 (I 992) 329-343. J.R. Mitchell: Rheology of polysaccharide solutions and gels, In: folysaccharides in Food. J.M.V. Blanchard, J.R. Mitchell (Eds.), (1979) 5 1-72. Butterworths, London. F.J.M. Kormelink, H. Gruppen. A.G.J. Voragen: Mode of action of (1,4)-P-D-arabinoxylan arabinofuranohydrolase (AXH) and a-L-arabinofuranosidaseson alkali-extractable wheat flour arabinoxylan Carbohydr. Res. 249 (1993) 345-353.

33.

C.C.H. Wang, K.C. Chang: Beet pulp and isolated pectin physicochemical properties as related to freezing. J. Food Sci. 59 (1994) 1153-1 154.

34.

R.M. McCready: Polysaccharides of sugar beet pulp: a review of their chemistry. J. Amer. Soc. Sugar Beet Technol. 14 (1966) 260-270. J.K.N. Jones, Y. Tanaka: Araban from sugar beet by limewater extraction. Meth0d.s in Carbohydr. Chem. 5 (1965) 74-75. A. Oosterveld, G. Beldman, H.A. Schols, A.G.J. Voragen: Arabinose and ferulic acid rich pectic polysaccharides extracted from sugar beet pulp. Carbohydr. Res. suhmitted for publication. F. Michel, J.F. Thibault, C. Mercier, F. Heitz, F. Pouillaude: Extraction and characterization of pectins from sugar beet pulp. J . Food Sci. 50 (1985) 1499-1500. F. Guillon, J.F. Thibault: Methylation analysis and mild acid hydrolysis of the 'hairy' fragments of sugar beet pectins. Carbohydr. Res. 190 (1989) 85-96. T. Sajjaanantakul, L.A. Pitifer: Pectinesterase. The chemistq arid technology qfpecriri (H. Walter, Ed.). Academic Press, San Diego, I99 I , 135-164. G. Williamson: Purification and characterization of pectin acetylesterase from orange peel. Phytochemistv 30 (1991) 445-449. M.J.F. Searle-van Leeuwen, L.A.M. van den Broek, H.A. Schols, G. Beldman, A.G.J. Voragen: Rhamnogalacturonan acetylesterase: a novel enzyme from Aspergillus aculeatus, specific for the deacetylation of hairy (ramified) regions of pectins. Appl. Microbiol. Biotechnol. 38 (1992) 347-349. J. Hwang, Y.R. Pyun, J.L. Kokini: Sidechains of pectins: some thoughts on their role i n plant cell walls and foods. Food Hydrocolloids, 7 ( 1993) 39-53. F. Guillon, J.F. Thibault: Oxidative cross-linking of chemically and enzymatically modified sugar beet pectin. Carbohydr. Polym. 12 (1990) 353-374.

35. 36. 37. 38. 39. 40. 41.

42. 43.

2 Starch: present use and future utilization

H. Roper Cerestar, Eridania BCghin-Say Vilvoorde Research & Development Centre, Havenstraat 84, B- 1800 Vilvoorde, Belgium

Summary. In 1994 more than 6.5 million tons of starch were produced in Western Europe (EU+EFTA) from maize, wheat and potatoes. Of the 6.1 million tons of starch consumed, 55% was utilized in the food sector and 45% in the non-food sector. Starch is used as native starch, as modified starch, as starch hydrolysate or as crystalline dextrose and as their derivatives. Consumption figures per starch product type and by market sectors (foodhon-food) are presented. Main non-food sectors are: paper and cardboard, fermentation and chemical industry. The industrial raw material starch is available in sufficient amounts and in high purity. Factors influencing physico-chemical properties of native starch are explained. Starch modifications with changes of physico-chemical properties of starch lead to different application properties. General quality requirements for starch and more specifically by the food and the technical industries are summarized. In the various application areas, innovative products based on starch have entered/are entering the market, e.g. food products, detergent components, packaging materials, cosmetic ingredients, pharmaceutical excipients and fermentation feedstocks. Raw material requirements by the industry and options for further research and development are discussed.

Introduction In order to give an overview on starch properties and applications, examples of food and non-food applications have been selected to demonstrate that starch exhibits not only well-known properties such as gelatinization, viscosity development and gluing, but also other properties that can be used to solve problems in food and non-food applications. In 1994 more than 6.5 million tons of starch were produced from maize, wheat and potatoes in Western Europe (EU+EFTA). Consumption of starch products in Western Europe (EU & E R A ) in 1994 amounted to 6.1 million tons (Fig. l), divided into 3.23 million tons of hydrolysates (53%), 1.83 million tons of native starches (30%) and 1.04 million tons of modified starches (17%). Of these 6.1 million tons, 3.36 million tons (55%) was consumed in the food sector and 2.75 million tons (45%) in the nonfood sector.

18

H. Roper Sectorial consumption of starch products in Europe 1994

Consumption of starch products per product group in Europe 1994

Chamlsal

."d ,.rm.m.d

I

.(7%

\

I

Total: 6 1 mi0 t

Other nm4wd

p'odwl.

1

Fig. 1. Starch consumption by product type.

Fig. 2. Starch consumption by market sector.

Fig. 2 shows the 1994 consumption of these 6.1 million tons of starch-based products per application area, where paper ranks first: Chemical Industry & Fermentation Industry 13% (793,000 t), cardboard production 8% (488,000 t), Paper Industry 20% (1.22 mio t), other nonfood uses 4% (244,000 t), confectionery 15% (9150,000 t), beverages 13% (793,000 t), food preparations 6% (366,000 t) fruit products 6% (366,000 t) and other food uses 15% (915,000 t).

Native starch: structure and properties1?* Starch consists of nearly unbranched amylose (MW 100,000-1,000,000) and highly branched amylopectin (MW 1,000,000-10,000,000) (Fig. 3). Both starch polysaccharides are homopolymeric a-1,4-glucans with 1,6-branches and D-glucose as monomeric building unit. Amylose and amylopectin occur in different ratios according to the botanical origin of the starch. The chemical purity of starch is >99%. Native starch granules are insoluble in cold water. Upon heating the starch granules swell with viscosity increase and formation of a starch paste. Starch properties (Fig. 4) such as gelatinization behavior and gelatinization temperature, viscosity, swelling, gel formation, gel stability and retrogradation tendency, but also chemical reactivity and susceptibility towards enzymes are dependent on external factors, e.g. weather conditions, soil quality and fertilization during growth of the starch-containing plant, and also on the conditions of harvesting, drying and storage of the grain. Starch properties are also dependent on the amylose/amylopectin ratio, moisture and trace impurities of proteins, lipids and minerals after separation of the starch from the other grain components. Amylose/ amylopectin ratio, particle size distribution and the shape of starch granules are genetically determined and are characteristic

Starch: present use and future utilization

Amylose

19

Amylopectln

c

B

I

Linkage type

:

a (11*4)

Molecular weight

:

100.000 1.000.000

a (IN* 4) and a (111*6)

-

1.ooo.ooo

- 10.000.000

5 7.000

> 7.000

Morphology

:

Crystalline A, B, V structure

Amorphous - crystalline

Complexing Power

:

High (blue iodine test)

Low ( red iodine test)

Retrogradation

:

High

Low

D.P.

-

Fig 3. Amylose/amylopectin characteristics

of each starch type. Starch granules exhibit characteristic X-ray diffraction pattern: the type A for cereal starches (maize, wheat, rice) and the type B for tuber starches (potato). Type C is a mixture

of type A and B. V-amylose is an amylose-lipid inclusion complex.

Factors influencing starch properties Fertilizing Harvesting Drying Storing

-

1

Plant

I Starch Granule

I

Climate Weather Soil

Crystallinity A, 8. C, spectra

reactivity

behaviour temperature

Susceptibility to enzymes

Fig. 4. Factors influencing starch properties.

Gel formation Retrogradation

H.Roper

20

Physico-chemical properties of native starch (Fig. 5) such as swelling (viscosity, rheology), gel formation and gel strength as well as water binding are utilized in food production. In the technical industries, properties such as adhesion and binding power, gluing power, fluidity, solubility and dispersibility are used in the production of paper, cardboard, corrugated board, glues and adhesives, gypsum board, and mortar and cement, and in foundry practice. In the textiles industry properties such as film forming and degradability by amylases are used in yarn sizing. The biodegradability of starch can be used for the manufacture of biodegradable materials. In the manufacture of polymeric dispersions (latices) starch acts as protection colloid.

Starch properties Viscosity (Rheology) Gel strength

Foodstuffs

Water binding

I

1

Adhesive power

Paper/Corrugating

Film forming

Textile processing

Biodegradability Protective colloid

I

I

Biodegradable products Emulsion polymerisation

Fig.5. Starch properties and application areas.

Modified starches193 To improve basic properties, e.g. for adaptation to a specific application, starch is modified by physical (pre-gelatinization) or chemical means (Fig. 6 ) .The latter include chemical and enzymatic catalyzed degradation, substitution to starch esters and starch ethers and cross-linking with bifunctional reagents to form diesters and diethers. In this way modified starches of different molecular weight distributions and varying degree of substitution with non-ionic, anionic and cationic substituents are produced. These modifications (Fig. 7) change the basic properties of starch, such as molecular weight, molecular weight distribution, crystallinity and trace impurities, and substituents are introduced into the starch molecule. This leads to changes of key properties such as gelatinization temperature,

Starch: present use and future utilization

21

I Starch modification1 1

Physical treatment

Pregeiatinisation

Chemical treatment

r

I

Degradation

Substitution

m p & T p k iJ

I Crossbonding

Acetates

Phosphates Mono-P

Fig. 6 . Starch modification.

solubility, viscosity, gel formation, ionic charges, etc. and adaptation to specific application requirements such as rheology, water-binding ability, adhesion and binding power, gel strength, protective colloid and film formation, as well as biodegradability.

Correlation basic properties / application properties Basic Properties

Modification

Key-properties

Fig. 7. Correlation of basic properties and application properties of starch

Application properties

22

H. Roper

Acid-catalyzed degradation (hydrolysis) of starch leads to controlled viscosity reduction, good gel formation, good penetration and improved adhesion (tack). Acetylation of starch leads to improved viscosity stability, decreased gelatinization temperature and improved film formation. Fig. 8 shows application properties of commonly-used modified starches in the food and non-food (technical) sector.

Modified starches Acid-thinned starches 0 Controlled viscosity decrease

Good gelling power n Good Denetration 0 immediate tack

0 improved viscosity stability 0 Decreased gelatinisation temp. 0 Better film formation

I Di-starch phosphates I

0 improved stability a ainst low pH,

shear stress, freeze-taw cycles increased viscosity 0 Increased gelatinlsation temperature Hypochlorite oxidised starches Controlled viscosity decrease Retrogradation stability improvedcoiour/opacity 0 Increased binding power

I Cationic starches1 0 improved retention of fibres and

pigments (Inci. self-retention) (= lower COD/BOD values) Better dewatering of paper 0 Increased burst and tensile strength of papers

Fig. 8. Selected modification.

Cross-linking of starch with sodium trimetaphosphate or phosphorusoxychloride (POC1,) gives di-starch phosphates with improved acid, shear and freezehhaw stability, higher viscosity and higher gelatinization temperatures. Oxidation of starch leads to controlled decrease of viscosity, better stability towards retrogradation, improved color/opacity and increased binding power. In paper and cardboard production, cationic starches improve the retention of cellulose fibres and pigments, thus reducing the waste-water load; they improve dewatering characteristics during sheet formation and improve mechanical properties of paper, such as increased burst and tensile strength. Production and use of modified starches in food are regulated by law.

Starch: present use and future utilization

23

Food Industry Starches are traditionally used in food to improve functional properties such as texture, appearance and mouthfeel. They also facilitate processing and improve the shelf life of food products. The physiological energy content of starch is 4.5 Kcal/g. Through the carbohydrate catabolism starch provides the organism with the necessary biochemical energy to maintain important physiological functions.

UHT starches Interesting modified starches are the ultra high temperature UHT starches. These are acetylated di-starch adipates and hydroxypropylated di-starch phosphates. UHT treatment (heating to 140-142 "C for 5-6 seconds) of milk products such as desserts inactivates micro-organisms and enzymes, thus improving product storage stability. UHT starches guarantee fine-tuned and tailormade functional properties such as rheology and texture; the sensorial properties and nutritional value of the products are maintained. Products are freezehhaw and acid ~table.~.5

Emirlsifying starches Emulsifying starches are produced by reacting starch with n-octenylsuccinic acid anhydride.6 These starches are highly degraded and therefore cold-water soluble. They are used e.g. for the stabilization of aroma emulsions in soft drinks and for the production of cholesterol-free salad sauces and salad dressings.7

Trends in NutritiodFunctional Food Some years ago the main emphasis placed on carbohydrates such as starch was their nutritional value. Today there is an increasing interest in health aspects (Fig. 9). Foodstuffs can play an important role in the regulation of digestive functions and in the prevention of "civilization diseases" such as obesity, diabetes, caries, cardiovascular diseases and colon cancer. The stimulation of natural defense mechanisms including the possible stimulation of the immune system could be a very important additional role for all food products. Examples of such starch products are fat replacers and resistant starch.

24

H. Roper

Regulationof digestive functions

Prevention of "civilisationdiseases" (CVD, Obesity, Diabetes, Cancer)

Stimulation of natural defense mechanisms

-

Bifidus stimulation

Dietaw fiberintake Resistant starch

7

Reduction of calorie

/ intake

Fat replacement Maltodextrins

Non cariogenic sweetener

\ Vitamins

Fig. 9. Functional food.

Maltodextrins asfat replacers

For the prevention of cardiovascular diseases, the American Heart Association recommends a reduction in fat consumption so that not more than 30% of the total caloric intake is derived from fat allied to a corresponding increase in the intake of complex carbohydrates. Potato starch maltodextrins give with water reversible meltable gels with fat-like texture. They are therefore suitable for partial replacement of oils and fats, while retaining appearance, consistency and organoleptic properties of food products. In this way food products with up to 70% reduced calories can be produced, e.g. salad dressings, salad sauces, margarine, ice cream and meat products.8.9 Resistant starch

The important role of dietary fibres in human nutrition has been stressed for a long time. Similar health benefits are obtained with resistant starch.10,' Resistant starch is physiologically defined as the sum of starch and starch degradation products which are not digested and absorbed in the small intestine of healthy individuals. It has been proved that retrograded, insoluble amylose passes the digestive tract and reaches the colon nearly unchanged, inducing fibre-like effects: the

Starch: present use and future utilization

25

faecal bulk is increased, bile acids are bound and short-chain fatty acids (SCFA) are formed by anaerobic fermentation of resistant starch in the colon. These effects reduce the risk of colon cancer.

Resistant starch

-

DP=40 - 70

(1

a (1

0

4) linkages digestible

6) linkages

/3 - linkages

other than a (1 4) )1

and (1 11*6) - linkages indigestible "resistant"region

not digestible

region

Fig. 10 Resistant starch.

Two possible approaches to resistant starch are known that give the above-mentioned physiological effects (Fig. 10): a) insoluble products consisting of retrograded amylose with a double helical structure (DP 40-70) and b) soluble dextrin-like products with atypical internal linkages, which are formed during the dextrin production by rearrangement reactions such as transglycosylation or re~ersion.'~.'3 Neither product is digested by intestinal enzymes and thus both are only fermentatively degraded in the colon to form short-chain fatty acids (butyrate, propionate and acetate), hydrogen and methane.

Technical Industries In recent years, positive environmental aspects of the natural raw material starch have gained increasing importance, e.g. biodegradability, biocompatibility and the contribution to an equilibrated CO, balance. Selected examples will demonstrate the contribution that starch can give to reducing the environmental burden. In addition, by using starch and starch derivatives, advantages in processing and improved functional properties of products are frequently obtained.

26

H . Roper

Paper and cardboard The most important users of non-food starches are the paper and cardboard manufacturers. Basic materials for paper products are cellulose fibres (pulp), water, filler and additives. Huge amounts of starches and modified starches are used in the various stages of the paper production process (wet end, spraying, size press, coating) (Fig. 11). Not only do they function as binders but they also improve both the process, e.g. dewatering and runnability, and paper properties such as stiffness, strength, gloss, ink receptivity, printability and barrier properties of specialty papers. Of the total amount of starch used, 2/3 consists of native starches and 1/3 consists of modified starches obtained by cationization, oxidation or other derivatizations. Gluing properties of starches and dextrins are used e.g. for the production of corrugated board and paper sacks as well as for bottle labels and wallpaper glues.

Starch fibres In a precipitation and spinning process, starch can be transformed into a kind of fibrous structure that exhibits low water solubility at temperatures below 50 OC and loses its shape/morphology at temperatures above 50 "C.The fibre length is up to >150 pm (Fig. 12) and is thus far above the particle diameter of granular native maize starch (5-25 pm). This allows mechanical retention of the starch fibres in the sheet. Evaluation of starch fibres in fine papers

Application of starch in paper production

r

spraying

surface sizing

ao%

a

coating 8% 7% Fig. 1 1. Starch in paper production.

Starch: present use artd future utilization

27

Fig. 12. Starch fibres.

showed excellent retention and dry strength. The effect of these starch fibres in paper consists of a combination of improved fibre properties and binding properties.

Starch binders for dust compocting: coal (briquettes, graphite), ceramics and metals Binders based on native and modified starches are used for compacting powders and dusts. Aluminum dust, coal powder, fly ash or filter dust14,'5 are mixed with starch and water to create a

Fig. 13. Starch based binders.

28

H . Roper

dough that is pressed into shaped pieces such as pellets, bars, briquettes, cylinders and sheets (Fig. 13). The binder imparts the necessary strength to the wet pieces at ambient temperature, the "green bonding", to avoid fracture in further processing and drying. At drying temperatures of about 200 "C, the starch binder develops its maximum strength in the hot bonding step. Special modified starches are available to form a carbon skeleton by elimination of water at 250-300 "C. This effect is used in coal briquetting to impart structural stability to the briquettes during burning. Starch binders and cationic starch are also used in the production of ceramics, mineral fibre board, synthetic graphite for electrodes (for example) and in aluminum recycling. The driving force behind this utilization of starch is the reduction of corrosive air pollution associated with the use of less-expensive but problematic traditional binders such as molasses, bitumen, pith and sulfite waste waters from pulp production. Increasing legal restrictions on air contamination imply more and more cost-intensive cleaning of polluted air. Combinations of starch with synthetic polymers In many applications starch has been replaced by synthetic polymers. Today, however, because of environmental concerns, more and more combinations of starches with synthetic polymers are developed and used. In addition to improved environmental compatibility, improved functional properties are targeted. Combinations of starch with synthetic polymers can be obtained by mixing, block co-polymerization or grafting. Product examples are: co-builders for detergents and latex dispersions e.g. for tufted carpets, paper coating and special paints. Starch based co-buildersfor detergents Important ingredients of powdered detergents are the "builder/co-builder" systems. These systems complex calcium and magnesium ions, thus decreasing the water hardness and preventing precipitation of carbonates and soap salts on the fabric and scaling on the metal parts of washing machines. In Western Europe, more than 75,000 tons of polycarboxylate co-builders in combination with 850,000 tons of zeolite builders are used in detergents. The first generation of co-builders were polyacrylates, followed by acrylic acid/maleic acid co-polymers. Both have the disadvantage that they are not biodegradable. They are eliminated by adsorption on the activated sludge in the waste water treatment plant, a process which is still in

Starch: present use and future utilization

29

accordance, for example, with the German law on detergents. With the objective of inducing partial biological degradability, carbohydrates such as glucose and starch have been introduced (Fig. 14) into the copolymer. The best results have been obtained with graft-copolymers~6~~7: synthetic polycarboxylate side chains have been grafted on the starch backbone using iron and cerium ions as co-catalysts. Up to 30% of partially degraded starches can be incorporated without affecting calcium complexation and dispersing properties in comparison with the pure copolymer. Biodegradability only slightly exceeds the incorporated starch amount.

Acrylic acid / Maleic acid Copolymer DE 29 36 984

coo- coo- coo-

Acrylic acid / Maielc acid / Glucose Copolymer

US 4,963,629

Acrylic acid / Maleic acid / Starch Graft Copolymer 0-

EP 0 396 303 E P O 4 4 1 197

Fig. 14. Starch containing cobuilders.

Starch containing dispersions (latices)

Mixtures of starch with acrylatelbutadiene or styrenehtadiene copolymers are used for paper coating and for carpet back sizing. Physical mixtures of starch with polymers generally exhibit two problems: a poor wet pick strength due to the water solubility of the starch component and migration of the binder during the drying process, the "mottling effect". When the emulsion polymerization, exemplified by styreneibutadienelg, is performed in the presence of a highly degraded dextrin, there are three possibilities of interaction, as demonstrated by the enzymatic degradation of the dextrin component (Fig. 15):

30

H. Roper

Case 1: Polymer particle stabilised by a starch layer

Case 2: Polymer particle containing starch phases

Case 3: Starch particle containing polymer phases

Products of enzymatic degradation

4

0 polymer particles

.!,

0 hollow particles

0 .

fine particles

Fig. 15. Starch based emulsions.

1.

Polymer particles are stabilized by a starch layer as protective colloid. This is approximately the same situation as in a physical latedstarch mixture.

2.

Polymer particles containing starch phases.

3.

Starch particles containing polymer phases.

Cases 2 and 3 represent a very fine distribution of the starch and polymer phases in the dispersions, leading to significant improvements in paper coating a p p l i c a t i ~ n s as ~ ~compared ~*~ to starcMatex mixtures:

* * * * *

*

maximal solids contents and viscosities comparable with commonly-used, purely synthetic latices; reduced viscosity of the coating color, thus allowing the production of coating colors with higher solids content; the wet pick strength is in the range provided by synthetic latices; improved gloss in comparison to synthetic latices and physical starcNlatex mixtures; improved ink absorption; reduced mottling and blistering tendencies.

Starch: present use and future utilization

31

Biodegradable materi~ls2/-2~ The problem The driving force behind the development of biodegradable materials is the more than 100 million tons of Solid Municipal Waste (SMW) produced each year in Western Europe (Fig. 16). This contains 25-30 vol.% corresponding to 5- 10 weight% plastic materials, predominantly used for packaging. Since landfill space for these stable and non-biodegradable materials is running short and because incineration for energy generation creates additional emission problems, the development of biodegradable and compostable materials for special applications where long-term stability is not required seems to be logical. These applications include one way packaging for fast foods, containers for pralines, and ampules and materials for the agricultural sector, such as mulch foils and planting pots, and for hygiene articles such as disposable diapers and incontinence pads.

Composition of municipal solid waste by volume

EC Policy for waste disposal (priorities)

1. Reduction of waste 2. Recycling / Re-use

3. Controlled incineration 4. Landfill (if all else fails)

Fig. 16. Biodegradable materials.

EU legislation stipulates the following priority for preferences for waste management. 1.

2. 3. 4.

Prevention Reuse (recyclingkomposting) Controlled incineration Disposal in landfills.

32

H . Roper

After use, materials from renewable raw materials are transformed by composting into biomass, water and CO,, representing a closed cycle. Composting is therefore biological recycling.

Themioplastic starch Under controlled conditions (<20 % moisture, 140-170 "C, 20-70 bar) and addition of nonvolatile plasticisers, e.g. sorbitol, glycerol etc., starch can be brought into the thermoplastic state in an extruder.25.26 This thermoplastic starch can be used for the production of biodegradable films or injection-moulded articles. Various products are already available27, e.g. foamed packaging chips (loose fill) from starch. Due to the sensitivity of starch towards high humidity, development is concentrated on the improvement of the surface properties of these products.

Starch as raw material for the CosmeticsIndustry Degraded and water soluble emulsifying n-octenyl succinate starches show potential for the preparation of emulsions. Special cationic starches show potential for use in shampoos, e.g. as hair conditioners.

Starch as raw material for the Pharmaceutical Industry Starch and starch derivatives are used as excipients in tabletting, e.g. for coating, as disintegrants and as matrix for the controlled release of drugs. Using roll dried waxy maize starch, a bioadhesive tablet with miconazol nitrate as active ingredient for local application in the mouth has been developed. Controlled drug release was measured over a period of 10 hours.18

Raw materials for fermentation Due to low raw material costs, until now the fermentation industry has used molasses of fluctuating quality as carbon source with the consequences of high costs for refining and purification of final products, the need for disposal of byproducts such as gypsum, and the need for the treatment of waste waters, which are highly polluted with melanoidins, sulphates etc. Increasing environmental consciousness and corresponding legislation, and increasing capital costs linked to the necessity to keep investment costs low has increasingly forced the fermentation industry to use technologies based on pure fermentation feedstocks such as starch and starch hydrolysates. Running the fermentation at higher dry substance and achieving higher

Starch: present use and future utilization

33

spacehime yields (productivities) with pure and tailored carbon sources can lead to a reduction of production costs. The starch industry offers a broad range of carbon sources: starches, maltodextrins, dextrose, glucose syrups and maltose syrups of the highest purity and adapted composition.

Raw material requirements in industry General quality requirements for starch based raw materials are: purity, uniformity, storage stability, ease of handling, functional properties and favorable environmental characteristics at acceptable prices. The food industry emphasizes sensorial quality and nutritional and health aspects. For the technical industries, criteria such as purity, biodegradability and biocompatibility are important.

R&D options and needs The following R&D options for the starch industry have been identified: Improvement of raw materials: by means of classical plant breeding or modern genetic engineering techniques, the following improvements have been targeted: higher yields per hectare, resistance to plant vermin and diseases, better separability of starch from other grain components, and new functional properties of starches. During harvest and during treatment: less damage by heat during grain drying, by moulds and by transport. Improvements in separation and purification by using new technologies, e.g. membrane filtration, chromatography and super critical CO, extraction. Product innovation

These are new products based on starch, higher performing products and synergistic product combinations, especially for the food sector (lighthealth food ingredients), detergent components, raw materials for cosmetics, raw materials for the pharmaceutical industry, raw materials for the fermentation industry and biodegradable packaging materials.

34

H. Roper

References I.

Starch: Chemistry and Technology (R. L. Whistler, J. N. Bemiller, E. F. Paschall, Eds). Academic Press Inc., New York, 1984.

2.

H. Koch, H. Roper: Kapitel 8: StBrke. Polysaccharide (G. Franz, Ed.), Springer Verlag, Heidelberg. 1991. 177.197. H. Koch, H. Roper, R. Hopcke: New Industrial Uses of Starch. Plant Polymeric Carbohydrates (F. Meuser,

3.

4. 5.

6. 7. 8. 9. 10.

I I. 12.

13. 14. 15.

16.

17.

18.

D. J. Manners, W. Seibel, Eds.), Special Publication No. 134. The Royal Chemical Society, Cambridge. 1993, 157-179. A. Rapaille. J. Vanhemelrijk, J. Mottar: The use of starches and gums in UHT desserts. Dairy lnd. lnr. 53 (1988) 21-25. A. Rapaille, J. Vanhemelrijk: Comparative functionality study of modified starches in different UHT processing systems. Conference proceedings Food Ingredients Asia, Session 4, No 2, Bangkok ( 1992). May 13-15. Ch. N. Richards, C.D. Bauer (Anheuser-Busch, Inc.): Method of making lipophilic starch derivatives. US 4035235 (1977); Chem. Abstr. 87 (18) 137597r. V. De Coninck, M. G. Firton: n-Octenyl succinate starches. Production and application properties. Paper presented at the 43rd Starch Convention, Detmold, 28.-30.04.92. H. Haenel, F. Schierbaum: Die Verwendung von gelbildenden Maltodextrinen (SHP) zur Herstellung energiereduzierter Lebensmittel ( I . Teil). Erniihrung/Nurrition 4(7) (1980) 306-308. V. De Coninck, J. Vanhemelrijk: Carbohydrates as fat replacers in calorie reduced foods. lnternatio~alFood lngredierits 2 (1991) 27-30. H. N. Englyst, J. H. Cummings: “Resistant Starch”, a “new” food component: a classification of starch for nutritional purposes. Cereals in a European Context (I.D. Morton, Ed.), Ellis Horwood Ltd., Chichester U.K., 1987, 221-223 N. G. Asp, J. Bjorck: Resistant starch. Trends Food Sci. Technol. 3 (1992) 1 1 1 - 1 14. K. Ohkuma, 1. Matsuda, Y. Katta, Y. Hanno: Pyrolysis of starch and its digestibility by enzymes Characterisation of indiges~ibledextrin. Denpuri Kagaku 37 (1990) 107-1 14. K. Ohkurna, Y. Hanno, K. Inada, I. Matsuda, Y. Katto (Matsutani Chemical Industries Co., Ltd.): Dietary fibre manufacture from pyrodextrin. EP 477089 Al (1992); Chem. Abstr. I 17 ( I ) 6660w. W. Mitchell, B. Trevis, W.R. Wright, A. Mac Donald Hildon (Inrad Ltd.): Briquet hardening. GB 2 189 806 Al (1987); Chem. Abstr. 108 (6) 41015a. G. Detka, K.H. Tilker, W. Grafen, B.R. Oldengott, L. Messening, A. Lindic (Ruhrkohle A.G.): Verfahren zur Herstellung von Kohle- oder Koksbriketts. Ger. Offen. 3623324 Al (1988); Chem. Abstr. 108 (14) 1 15677s. S. Yamaguchi. T. Yokoi, S. Shioji, Y. Irie, T. Fujiwara (Nippon Shokubai Kagaku Kogyo Co., Ltd.): Process for producing and use of maleic acid (co-)polymer salts improved in biodegradability. EP 396303 A2 (1990); Chem. Abstr. 116 (10) 8 6 2 8 9 ~ . W. Denziger, H. Hartmann, A. Kud, R. Baur, J. Feldmann, H.J. Raubenheimer (BASF A.G.): Preparation of graft polymers from mono-, oligo-, or polysaccharides or modified polysaccharides. DE 40031 72 A I (1991); Chem. Abstr. I I5 (24) 256935f. Ch.C. Nguyen, V.J. Martin, E.P. Pauley (Pen ford Products Co.): Starch graft polymer dispersions in paper coatings. WO 9009406 Al (1990); Chem. Abstr. I 14 (6) 45303q.

Starch: present use and future utilization 19.

20. 21.

22. 23.

24. 25.

26. 21. 28.

35

K. Moller, D. Glittenberg: Novel starch containing dispersions as coating hinders. TAPPl Coating Conference ( I 990) 85-9 1. G. Rinck, K. Moller, S. Fuellert, F. Krause, H. Koch (Synthomer Chemie G.m.b.H.): Modified starches and dextrins for producing aqueous polymer dispersions. EP 408099 A l (1991); Chem. Abstr. 114 (26) 247999g. H. Roper, H. Koch: The role of starch in biodegradable thermoplastic materials. SrurcWSturke 42 (1990) 123- 130. Bundesministerium fur Forschung und Technologie, “Untersuchung zum Einsatz bioahbaubarer Kunststoffe im Verpackungshereich”, Forschungshericht Nr. 01-ZV 8904, Bonn, Aug. 1991. G. Delheye, “The role of biodegradable plastics in modern waste management”, Paper presented at the 43rd Starch Convention, Detmold, 28.-30.04.92. Institute for International Research, “Bioabhauhare Verpackungen-Chancen fur Marketing und Entsorgung”, Conference papers, Fachkonferenz Mannheim (D), 30.-3I.03.92. I . Tomka: A thermoplastically processahle starch and a process for making it. WO 9005 161 Al (1990); Chem. Ahstr. 1 13 (20) 174403a. W. Wiedmann, E. Strobel: Compounding of thermoplastic starch with twin-screw extruders. StarcldSturke 43 (1991) 138-145. Bayerisches Staatsministerium fur Ernduung, Landwirtschaft und Forsten, “Gesamtkonzept Nachwachsende Rohstoffe in Bayern”. Agra-Europe 42/9 I , 14.10.91. St. Bouckaert: Ontwikkeling en evaluatie van een huccale bioadhesieve tahlet voor lokale therapie. PhD Thesis, Ghent University, 1994.

This Page Intentionally Left Blank

3 Metal-catalyzed oxidation and reduction of carbohydrates

A. Abbadi and H. van Bekkum Laboratory of Organic Chemistry and Catalysis, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands.

Summary. Metal-catalyzed oxidation and hydrogenation of carbohydrates are reviewrd. Bismuth-promoted palladium is presently the catalyst of choice for the oxidation of aldoses towards aldonic acids. Selective oxidation at C-2 of aldonic acids towards 2-keto acids can he achieved by applying similarly promoted platinum in weakly acidic medium. The noble metal catalyzed oxidation of primary alcohol groups still needs improvement; hest results are obtained in a continuous process using supported platinum and operating at low oxygen concentration. Carbonyl group hydrogenation is generally performed over nickel catalysts. Here, also ruthenium catalysts come to the fore. Reductive amination of carbohydrates has received much attention in recent years. Mainly nickel and platinum catalysts are applied. Finally attention is paid to carbohydrate hydrogenolysis under severe conditions towards glycerol and 1.2-propanediol.

1. Introduction Oxidation and reduction are among the oldest conversion methods applied to carbohydrates. The early analytical techniques in the carbohydrate field were based on oxidationheduction reactions. Industrial oxidation products from carbohydrates include gluconic acid and its 2-keto derivative, lactobionic acid, oxidized starch, and 6-carboxy-cellulose. Examples of industrial carbohydrate reduction products are sorbitol, mannitol, lactitol and isomaltitol. Sometimes a synthetic route involves both a reduction and an oxidation step, such as in the manufacture of vitamin C. Oxidationheduction of carbohydrates can be achieved by applying stoichiometric chemical methods as well as bio- and chemo-catalytic methods. In carbohydrate reduction chemo-catalytic methods prevail whereas the oxidation of carbohydrates is still dominated by bio-catalytic procedures using enzymes or microorganisms to perform the desired transformation. In some cases, a biochemical oxidation cannot be approached by a direct chemo-catalytic procedure. Examples include the bacterial conversion of sorbitol to L-sorbose and the oxidation of D-glucose to 5-keto-D-gluconic acid.

38

A. Abhadi and H. van Bekkum

In this section we will focus on the metal-catalyzed oxidation and hydrogenation of carbohydrates. The first part will discuss the catalytic oxidation of monosaccharides and polysaccharides. Much effort has been devoted to this area in recent years. The state of the art in selective oxidation will be discussed. Some industrial applications of the final products will also be briefly mentioned. The second part is devoted to the hydrogenation of carbohydrates to polyols as well as to reductive amination. Finally a brief report on carbohydrate hydrogenolysis is included.

2. Oxidation The oxidation of carbohydrates is an important way in which new compounds and materials with interesting physical-chemical properties can be provided. Bio-catalytic, stoichiometric chemical as well as homogeneous and heterogeneous chemo-catalytic methods have been applied for the oxidation of carbohydrate molecules. 1.2 The application of noble metal catalysts in the liquid-phase oxidation of carbohydrates introduces several problems. The fragility of the substrates and their multi-functionality raise selectivity problem^.^-^ Moreover, many parameters must be controlled, such as pH, temperature, surface oxygen coverage, and the nature and stability of the catalyst in order to obtain the desired products. Considerable effort has been devoted, particularly by the Eindhoven group, towards understanding the behavior of supported platinum catalysts in carbohydrate oxidation. Recently, leaching of platinum from the catalyst followed by re-deposition has been reported.6-7 In general, heterogeneously catalytic oxidations are performed with air or oxygen in weakly alkaline aqueous medium (pH = 8 - 9) and in the presence of a supported noble metal catalyst. The first example was reported by Von Gorup-Besanz in 1861, who oxidized D-mannose using platinum black as the catalyst.* Heyns et al. have reviewed the reactivity toward oxidation of different functional groups of the carbohydrate molecules when using catalysts based on platinum in weakly alkaline medi~m.5-~-' I The following sequence of the reactivity toward oxidation was reported: HCO > COCH20H > CH20H > CHOH,, > CHOH,, The use of other metal catalysts for the oxidation of carbohydrates was studied by several research groups.'-* We will report here on metal and bi-metal catalyzed oxidation of carbohydrate molecules. Electrochemical oxidations which might be considered as indirect metal-catalyzed reaction are also discussed. This section will be divided according to the functional group involved in the oxidation reaction.

Metal-catalyzed oxidation and reduction of carbohydrates

39

2.1 Selective oxidation of the anomeric carbon The oxidation of the anomeric carbon (C-1) of aldoses leads to the formation of aldonic acids. By far the most important aldonic acid from an industrial perspective is D-gluconic acid, with a global annual production of 50,000 tons. It is marketed as 50% solution, as the sodium, potassium, or calcium salt, or as D-ghcono-6-lactone. The most-commonly available D-ghconate is the sodium salt which finds application in de-rusting and paint-stripping of metals, and as a concrete additive for extending setting time and enhancing compressive strength. It is also used as a sequestering agent to prevent the formation of insoluble soap films in alkaline cleaning of glassware and foodprocessing equipment. Calcium D-gluconate and ferrous D-ghconate are used to treat calcium deficiency and iron deficiency, respectively, while magnesium D-gluconate is an a n t i s p a ~ m o d i c . ' The ~ , ~ ~free acid and D-glucono-8lactone find application as latent acid catalysts in textile printing. D-Glucono-lactone also serves as starting chemical in the preparation of glucono-N-alkyl amides, a group of new specialty surfactants.14 The early industrial production of D-gluconic acid was based on the indirect electrochemical oxidation of D-ghCOSe.'5 The oxidizing species is the free bromine which is generated in situ from hydrobromic acid by a graphite anode. Other electrode materials such as copper, nickel or iron were also used. Details concerning the mechanism of bromine oxidation of aldoses were reported by Isbell and Pigman.I6 The oxidation reaction is presented below. 2 HBr + Br2 + 2 H+ + 2eR-CHO + Br2 + H20 + R-COOH + 2 HBr The use of a buffer during the oxidation is necessary to maintain a neutral pH of the reaction mixture; otherwise inhibition of the oxidation process will occur.17 Undesired side reactions are negligible and the yield is almost quantitative with respect to the amount of current passed. The same oxidation process was applied successfully to produce lactobionic acid.'* Further development in the field of catalysis led to several patents dealing with the heterogeneous catalytic oxidation of D-glucose to D-gluconic acid under mildly alkaline conditions.19-22Supported noble metals such as palladium and platinum were used as catalysts in the presence of oxygen or air as terminal oxidant. The use of other metal cations to carry out the oxidation of aldoses to aldonic acids was also investigated. Mercuric acetate was reported to provide about 50% of aldonic acids during the oxidation of aldoses in aqueous solution.23 Morgenlie reported on the use of silver carbonate supported on Celite (diatomaceous earth) in the conversion of aldoses into ald~no-lactones.~~ Other metal cations such as copper, iron and cobalt were also studied.l.2 However, these oxidation procedures have only found limited use as

40

A. Abbadi and H . van Bekkuin

analytical methods due to their lack of selectivity and the absence of sufficiently effective metal reoxidation techniques. The latter is a pre-requisite for a catalytic procedure. Among noble metal catalysts, palladium was found to be superior in the selective oxidation of aldoses to aldonic acids.25-26However, this catalyst suffered from de-activation by oxygen and the maximum conversion reached was about 92%. Recently, bi-metallic catalysts were developed for the C- 1 oxidation of carbohydrates. Thus bismuth-promoted palladium showed increased activity in D-glucose oxidation.27-29 The presence of the promotor seemed to suppress the poisoning of the catalyst and conversion >97% was achieved with high selectivity toward D-ghconic acid. Recently, Besson et al.30 showed that the promotor 'bismuth' probably exists as Bi2O3 and participates actively in the oxidation cycle.

Similarly, Despeyroux er af.31 developed a tri-metallic catalyst which is composed of palladium 4%, platinum 196, bismuth 5% on active carbon to carry out the oxidation of D-glucose to D-gluconic acid. The additional presence of platinum in the formulation of the catalyst seemed to enhance the catalytic activity while the selectivity was comparable to that of the Pd-Bi catalyst. The Pd-Bi catalyst was also used successfully to oxidize the anomeric carbon of the reducing unit of disaccharides and polysaccharides.32,33 Hendriks et ~ 1 . 3 4 2 3 5have reported on in-situ bismuth-promoted palladium in the oxidation of lactose to lactobionic acid. The same results were achieved as with a pre-manufactured Pd-Bi heterogeneous catalyst. We must also mention the work of the Delft gr0up3~on the facile dehydrogenation of D-glucose and related mono- and disaccharides such as D-galactose and lactose on platinum or rhodium catalyst under alkaline conditions (pH = 13.5). During this reaction, hydrogen was evolved while D-glucose was transformed to D-gluconic acid. Other metals, such as nickel, palladium and ruthenium, were less active and showed much lower selectivity.

Pt/C

HHO ofio-

pH HO

D-Glucose

or Rh/C

q~,%'c!

+H,

coo -

HO HO D-Gluconate

Metal-catalyzed oxidation and reduction of carbohydrates

41

The dehydrogenation is first order in catalyst, in aldose and in hydroxide. Clearly the reaction proceeds via the aldose anion (apparent pKa= 13.5 under these conditions37-38),whereas rupture of the IC-H bond is the rate-determining step. The reaction was recently applied by Solvay-workers to prepare lactobionic acid from lactose. Furthermore, the hydrogen-donor properties of reducing aldoses were used in hydrogentransfer reactions. Platinum- (or rhodium-)catalyzed oxidatiordreduction of D-glucose/D-fructose under alkaline conditions (pH = 13) was thus a ~ h i e v e d . 3Essentially ~ all the hydrogen evolved from D-glucose dehydrogenation was used in the reduction of D-fructose.

D-Glucose

5 % Pt /C or

D-Gluconic acid

TI\ o 5 M KOH

D-Fructose

03 M

N,

Mannitol Sorbitol

We were interested in developing a single-step heterogeneous catalytic process for the conversion of starch to D-gluconic acid. This process is characterized by simultaneous hydrolysis of starch and oxidation of liberated D - g h c ~ s eSolid . ~ ~ acids (zeolitic materials, ion exchange resin in H+-form) were used as catalyst to carry out the hydrolysis reaction while supported noble metal catalysts (Pd, Pt) catalyzed the oxidation reaction with oxygen. When maltose was used as model compound, the yield of D-gluconic acid amounted to -40% and inhibition of the oxidation reaction was observed. We have investigated separately the nature of the poisoning of palladium41 and platinum42 catalysts during the oxidation of D-glucose to D-ghconic acid in acidic medium. The oxidation reactions were carried out in batch. Cyclic voltammetry was applied to understand the behavior of the catalysts. It seemed that free D-ghconic acid is the main poisoning species. The free carboxyl group is apparently involved in strong interaction with the active sites of the catalyst leading to inhibition of its catalytic activity. However, the presence of a promotor such as Bi or Pb suppressed the poisoning of the platinum catalyst and D-glucose was oxidized to 2-keto-D-gluconic acid (see Section 2.3).

2.2 Selective oxidation of primary hydroxyl groups Natural C-6 oxidized polysaccharides such as alginate and pectin constitute an important and valuable class of industrial compounds which find application in food and non-food products. The

42

A . Abbadi and H. van Bekkiim

C-6 oxidation of inexpensive polysaccharides such as starch seems to be an interesting approach to

this type of valuable materials. However - not unexpectedly - the metal-catalyzed oxidation of the primary hydroxyl group of polysaccharides suffers from steric hindrance. Long reaction times are needed with platinum as catalyst, while the uronic acid content does not exceed = 2O%.43,44 Edye et al. have obtained 7% uronic acid during the oxidation of inulin over Pt/C after a reaction time of 8.5 hours.45 10%Pt/C was used as catalyst in neutral medium (pH = 7) at 100 "C. Better results in Pt-catalyzed inulin oxidation under weakly alkaline conditions (pH = 9) were obtained in our laboratory (=20% uronic acid) but degradative oxidation appeared to be a major side reaction.46 Recently, Kerkenaar and Besemet'7 achieved an oxidation degree of maltodextrins of =70% when the oligosaccharide was brought into contact with a nickel electrode. The oxidation reaction was carried out at pH = 13 in the presence of fatty acid and exhibited high selectivity towards C-6 oxidation. The oxidizing species is nickel peroxide which is electrochemically generated at the surface of the electrode. The fatty acid present in the reaction mixture seemed to be involved in the formation of an inclusion complex with maltodextrins. The latter adopted a coil conformation around the fatty acid with the primary hydroxyl groups of the D-glucose units directed to the outer surface. Formation of a complex of this type enhances the accessibility of C-6 hydroxyl groups to the catalyst.

CH,OH

CH,OH

In our work, the oxidation of maltodextrins with nickel peroxide, generated in-situ from nickel(I1) salt (catalytic amount) and hypochlorite, exhibited the same selectivity towards C-6 oxidation, albeit that the oxidation degree was lower (= 15%).4* Similarly, the C-6 oxidation of monosaccharides is also of industrial importance. The selective oxidation of aldoses to uronic acids can provide products with detoxicant activity such as D-ghcuronic acid. The latter might be used as starting material for the synthesis of vitamin C

(40,000 tons per year). However, the preparation of uronic acids from aldoses by catalytic oxidation requires the protection of the anomeric carbon. For this purpose, alkyl glycosides in which C-1 is protected are considered as suitable starting materials, mainly because of their easy preparati0n.~9-52Supported platinum is the catalyst of choice to carry out this oxidation reaction.

Metal-catalyzed oxidation and reduction of carbohydrates

43

The oxidation of methyl D-glucoside on platinum was extensively studied.7 The catalyst exhibited a decrease of its selectivity (70%)at higher conversion. This problem is overcome in the new Siidzucker continuous oxidation process, which uses a fixed Pt/C bed reactor coupled to an electrodialysis system (see sucrose oxidation). The modification of platinum by a promotor such as bismuth or lead in the batch oxidation of methyl D-glucoside does not influence the selectivity of the catalyst.s3 When long alkyl chains (C8 and C12) were attached to C-1 instead of the methyl group, improvement of the selectivity toward C-6 oxidation of the alkyl glucoside over platinum was observed (C8 = 90%).54955 The enhancement of the selectivity may originate from (i) association of glucosides leading to protection of the pyranose ring against oxidation, and/or (ii) hydrophobic effects exerted by the alkyl chains in analogy to the effects of 3,4- and 3,s-alkyl substituents on acidity and reactivity of benzoic acids.56 The oxidation of I-0-alkyl glucosides provides a new group of surfactants combining nonionic and anionic properties in analogy to a class of ethylene oxide-derived surface-active compounds. In addition to the selectivity problems during the oxidation of alkyl D-glucosides, their conversion to free D-glucuronic acid by acid-catalyzed removal of the protecting group at the anomeric carbon results in serious degradation of the product. The over-all yield of D-glucuronic acid was reported not to exceed = 2O%?9-50 The use of other C-1 protecting groups was also investigated. Thus a-D-glucopyranose 1 -phosphate was considered as a substrate to prepare D-glucuronic acid.57 a-D-Ghcopyranose 1-phosphate is obtained from the action of phosphorylase on starch in the presence of inorganic phosphate. The oxidation reaction catalyzed by platinum yields 70% D-glucuronic acid 1 -phosphate.

I

Ho OP0,2'

I

Ho OP0,2'

Van Dam et al.57-58 have examined the influence of the particle size and the Pt distribution over the active carbon on the activity and the selectivity of the catalyst in this reaction. Van Dam used a diffusion-stabilized catalyst which is characterized by large uniform particles ( e . g . extrudates) in which the occurrence of oxygen diffusion limitation leads at a certain depth in the particle to proper tuning of the reaction. This catalyst exhibited a high steady-state reactivity while the selectivity was the same as with a powder catalyst. On the other hand, the selectivity towards

44

A. Abbadi and H. van Bekkum

D-ghcuronic acid I-phosphate was influenced by the way of preparation of the carrier and reached 87% with higher oxygen content of the active carbon. Once again, the conditions applied to remove the phosphate group can decrease the total yield of D-glucuronic acid substantially (= 12%).57 Oxidation of 1,2-O-isopropylidene-a-D-glucofuranose over Pt/C catalyst at pH = 8-9 was carried out to prepare D-glucuronic acid.59-61 The intermediate 1,2-0-isopropylidene-a-Dglucofuranuronic acid was obtained in 60% yield and hydrolyzed almost quantitatively to D-ghcuronic acid, which was isolated as the crystalline D-glucofuranurono-3,6-lactone.

coo -

When this reaction was performed under slightly acidic conditions, simultaneous oxidation of C-6 and C-5 took place leading to the formation of 1,2-O-isopropylidene-a-D-5-keto-

glucofuranurono-3.6-lactone,62~63a synthetic intermediate in the preparation of ascorbic acid. The oxidation of aryl and/or alkyl disaccharides on Pt/C was also studied. Benzyl P-cellobioside64 as well as benzyl P-maltoside65-67 yielded the corresponding uronic acid. It is noteworthy that only the primary hydroxyl group of the D-glucose residue remote from the benzyl group is oxidized. In contrast, the oxidation of methyl P-maltoside yielded all three possible glycuronate disaccharides, indicating random oxidation of both primary hydroxyl groups.68 The substitution of the methyl group by a long alkyl chain ranging from 8 to 12 carbon atoms showed an exclusive oxidation of the C-6 of the terminal D-glucose ~ n i t . 5As~ was mentioned earlier for alkyl glucosides, we believe that this selectivity is probably due to association in a micelle structure of the alkyl or aryl maltosides making of the terminal D-glucose residue the most accessible unit to the platinum catalyst. The catalytic oxidation of sucrose in which the anomeric carbon of D-glucose is involved in the glycosidic linkage was also described in several patents.52.69-7' The oxidation products (oxidation of the primary hydroxyl groups) are mono-, di-, or tricarboxylic acids which can serve various applications. Tricarboxysucrose, in which all three primary hydroxyl groups are oxidized, was obtained in 35% yield when an 8% solution of sucrose was oxidized over a Pt/C catalyst at 80 "C and pH = 9.70Recently, a detailed study on the nature of the oxidation products of sucrose in neutral medium (pH = 7) was rep0rted.~5The reaction was carried out at 100 "C using 10% Pt/C as the catalyst. The oxidation was found to be highly specific for the primary hydroxyl

Metal-catalyzed oxidation and reduction of carbohydrates

45

groups at the 6 and 6' position of sucrose while no evidence for C-1' oxidation was found. The reluctance of the C-1' hydroxyl group to undergo oxidation is probably due to the difficult accessibility of this hydroxymethyl group to the active sites of the catalyst. Selective oxidation of sucrose towards sucrose monocarboxylic acid is obtained in a new sucrose oxidation process developed by S i i d z ~ c k e rIn . ~this ~ process an oxygen-saturated aqueous sucrose solution is continuously fed to a fixed bed of 5% Pt/C catalyst. The product stream is separated in an electrodialysis system into oxidized products and a recycle stream. A selectivity of 96% towards sucrose monocarboxylic acid was reached. The three primary hydroxyl groups of sucrose were converted in a ratio 47:44:5 (C-6':C-6:C- 1'). The oxidation of isomaltulose [6-~-(a-D-glucopyranosy~)-D-fructose], which is obtained by 3 was used as catalyst and both bacterial conversion of sucrose, was described by K u n ~ . ~Pt/C hydroxymethyl groups, at 1 and 6' position, are oxidized. When the Pt-catalyzed oxidation was applied to ketoses, the primary hydroxyl group at C-l reacted readily. The adjacent carbonyl group at C-2 seemed to increase the reactivity of this hydroxyl group while the second primary hydroxyl group at C-6 was slowly oxidized. According to Heyns, the catalytic oxidation of L-sorbose and D-fructose provided 2-keto-~-gulonicacid and 2-keto-D-gluconic acid, respectively, in 60% ~ i e l d . ~ jRecently, -~* it was reported that modification of Pt catalyst with tertiary amines under neutral conditions enhances both the rate of the reaction and the initial selectivity toward the oxidation of the C- 1 primary hydroxyl group of ~-sorbose.79 Furthermore, the addition of a promotor such as Bi and/or Pb to Pt and/or Pd catalysts in the oxidation of ketoses increased the yield of the corresponding 2-keto-aldonic acid to 85%.80-83 The behavior of Bi-modified Pt catalysts was recently studied by cyclic v ~ l t a m m e t r y .We ~ ~ have prepared these 2-keto-acids in high yield (>97%) under weakly acidic conditions using the corresponding aldonic acid as substrate and bismuth- or lead-modified platinum as the catalyst85.86 (see Section 2.3). The catalytic oxidation of 2,3:4,6 di-0-isopropylidene-L-sorboseover Pt/C gave an essentially quantitative yield of the corresponding 2-keto-acid.5 The electrochemical oxidation of this compound using a nickel electrode was also reported to provide the corresponding acid in high yield (93%).87 Indirect electrochemical oxidation was also applied to carry out this reaction in aqueous alkaline solution in the presence of a surfactant and added oxometal anions such as chromate or permanganate as catalyst.88 On the other hand, the oxidation of the partially protected over PtK at higher temperature yielded the L-sorbose (2,3-0-isopropylidene-L-sorbofuranose) diacid.89.90 Methyl a-L-sorbopyranoside was oxidized selectively under controlled conditions at C- 1 over Pt/C as the catalyst.9' The primary hydroxyl group of aldonic acids was oxidized with a moderate selectivity on Pt/C catalyst to yield aldaric acid. Dirkx el al.92 reported a 50-55% yield of D-glucaric acid when using D-gluconic acid as starting material. The reaction was carried out under weakly alkaline

46

A. Abbadi and H. van Bekkum

conditions (pH = 8-9). When using D-glucose as substrate, simultaneous oxidation of C-1 and C-6 took place and D-glucaric acid was obtained in 50-55% yield.92 Separately, Venema et al. have reported the same yield (50-55%) of aldaric acids in the platinum-catalyzed oxidation of aldopento~es.~3 In the presence of an equimolar amount of boric acid, D-ghcaric acid showed good Ca-sequestering proper tie^.^ The I ,4-lactone possesses specific anti-b-D-glucosiduronase activity and has a wide range of applications as a biochemical reagent.94 It is also of therapeutic interest and is reported to be effective i n the treatment of bladder cancer.95 Treatment of D-giucaric acid with strong mineral acids affords furan-2,5-dicarboxylic acid. This compound can also be prepared from 5-hydroxymethylfurfural by catalytic oxidation in weakly alkaline medium over platinum.96 The starting material (HMF) is obtained by dehydration of D-fructose in the presence of an acid catalyst?’ and finds several appli~ations.~8 Under controlled conditions, the oxidation of HMF leads with good selectivity towards 5-carboxyfurfural.99~~~~ The use of a combined silver oxidekopper oxide catalyst, however, afforded 5-carboxyfurfuryl alcohol. Oxidation at the C-6 position of partially protected monosaccharides using other metals was also studied. Ruthenium tetroxide was applied in combination with sodium periodate in aqueous acetone to catalyze the oxidation of the C-6 position to a carboxyl group.lOl Recently, the oxidation of methyl and octyl a-D-glucopyranoside to their corresponding glucuronic acid, catalyzed by high-valent ruthenium species using NaBr03 as oxidant, was described.’O* In aqueous alkaline media, potassium ferrate (VI) ion has been reported as selective reagent for the oxidation of the hydroxymethyl group of methyl a-D-hexopyranoside to an aldehyde group.103 The oxidation product was not isolated but oxidized in situ with sodium halites to give the corresponding uronic acid. Theander applied a two-phase system in the oxidation of 1,2-0isopropylidene-a-D-glucofuranosewith chromic acid as reagent.Io4 1,2-0-Isopropylidene-a-Dgluco-hexo- and -D-xylo-pento-dialdo- I ,4-furanose were identified among the reaction products. Also, partially acetylated derivatives of maltosel05 and cellobiosel06 were treated stoichiometrically with potassium permanganate in acetic acid to offer the corresponding biouronic and pseudobiouronic acids.

2.3 Selective oxidation of the secondary hydroxyl groups Selective oxidation of secondary OH functions can be achieved when both the aldehyde and the primary hydroxyl group of aldoses are protected. Hydroxyl groups axially attached to the pyranose ring are preferentially oxidized over Pt-based catalysts. Thus the oxidation of methyl 4,6-0-ethylidene-a-D-mannopyranoside in which the 2 OH is axially attached to the pyranose ring

Metal-catalyzed oxidation and reduction of carbohydrates

47

offered 82% yield of the 2 - k e t o - d e r i v a t i ~ eOther . ~ ~ ~ ~aldohexoses ~~~ exhibited the same selectivity towards the oxidation of the axially attached secondary hydroxyl groups.Io9 Heyns et al. have oxidized a series of benzyl P-D-pentopyranosides in which the primary hydroxyl group is involved

in the formation of the pyranose ring. Arabinose and ribose were oxidized over Adams catalyst (reduced PtO2) in neutral medium to yield the 4- and 3-glyculoside, respectively.110 Interestingly, showed higher reactivity of the the oxidation of 2,3-O-isopropylidene-~-D-fructopyranose secondary hydroxyl groups than the C-l primary hydroxyl group toward oxidation.111 The main reaction product is the diacid which is formed by overoxidation of the two secondary hydroxyl groups and cleavage of the 4C-5C bond (glycolic cleavage).

I

OH

CH,OH

When subjecting 1,6-anhydro-P-D-hexopyranoses to the oxidation reaction, the following order of reactivity of the different secondary hydroxyl groups is to be expected' 12: 3-CHOHaX > 4-CHOHaX > 2-CHOHaX > 4-CHOHeq > 2-CHOHeq > 3-CHOHeq. The selective catalytic oxidation of the 2-CHOH of aldonic acids was also reported. 1,4-D-glucono-~actonewas oxidized in methanol and in the presence of phosphoric acid, on vanadium pentaoxide using chloric acid as oxidant to yield methyl 2-keto-D-gl~conate.~ 13.1 14

0

V,06,HCI0, MeOH

OH

COOMe

I I HO-CI

C= 0

H-C-

I

H OH

H-C-OH

I

CH,OH 2-Keto-D-gluconate can also be prepared from D-gluconate by oxidation on lead- or bismuthmodified platinum catalyst using air or oxygen as terminal oxidant.' I5 The selectivity toward the

48

A . Abbadi und H . van Bekkum

oxidation of 2-CHOH originates from the formation of a complex between the promotor (Pb or Bi) and D-gluconate. The 2-hydroxyl and the carboxyl group of D-gluconic acid serve as coordinating groups in the formation of this complex which leads to the activation of 2-CHOH toward oxidation.l16*117

H -+-OH H--G-OH

I CH,OH

-

H -+-OH H4-OH

I

CH,OH

The nature and the stability of the complex formed between D-gluconic acid and Pb or Bi are strongly dependent on the pH.'lg.l19 The performance of the catalyst is therefore influenced by the pH of the reaction mixture. We have found that under weakly acidic conditions, when the poisoning of the catalyst is avoided, conversions >97% of aldonic acids to the corresponding 2-keto-aldonates are achieved with a selectivity >98%.85q86 Under alkaline conditions (pH = 8-9, as applied in the earlier work' 15) coordination e.g. at C-3 C-4 is also expected, and the selectivity of oxidation will decrease.

0

The oxidation of acetylated methyl P-D-hexopyranosides and of the corresponding a-or P-D-furanosides with chromium trioxide yielded 5-hex- and 4-hex-ul0sonates.'~~ Methyl 2,3,4,6-tetra-acetyl-5-keto-D-gluconate was obtained in 76% yield in this way. An efficient chemo-catalytic method for preparing this component is as yet unavailable.

Metal-catalyzed oxidation and reduction of carbohydrates

49

COOMe OAc

I I AcO -CH I H-C-OAc I c=o I H -C-OAc

AcO

Me

AcO

AcOH

CH,OAc

5-Keto-D-gluconic acid, which can be obtained by fermentation together with some 2-keto-D-gluconic acid, serves as intermediate for the preparation of the important fragrance

4-hydroxy-5-methyl-3(2H)-furanone.

COOH

I I H0-C-H I H-C-OH I

H-C-OH

D-glucose - -

c=o

CH,OH 5-keto-D-gluconic acid 2.4 Diol-cleavage oxidation and selective oxidative degradation The cleavage of glycol moieties of polysaccharides yields new products which have a broad spectrum of application (see the contribution by Besemer and Van Bekkum in this book). For example: dicarboxy-starch and dicarboxy-inulin, which show good sequestering properties, can be used in detergent formulati~ns.'~l Another interesting derivative of starch is the dialdehyde which is prepared by periodate oxidation in aqueous medium.122-125 This reagent exhibits a high selectivity toward diol cleavage to form dialdehyde. Lead tetra-acetate was also reported to be an efficient oxidant in performing oxidative scission of diols.126 The oxidative behavior of these two reagents is mainly attributed,to the ability of the central atom (Pb or I) to complex with a 1,2-diol group and effect a two-electron transfer.'?' Details on the mechanism and the kinetics of this reaction were reported elsewhere.I28

50

A. Abbadi and H. van Bekkuni

During the oxidation of starch by periodate to form the dialdehyde-starch, the oxidant can be regenerated in sifu by electrochemical oxidation at a lead anode.l29-l3l This indirect electrochemical process was recently reviewed and its economical feasibility has been improved.I3*

0%

w m

The use of the tungstate-H202 system to prepare dicarboxy-starch was studied by the Delft group.133 The catalytic system which operates under acidic conditions (pH = 2) was found to be unsuitable for this purpose. In addition to the glycol cleavage of the 2C-3C diol moieties in the internal D-glucose units, undesired hydrolysis of the glycosidic linkage and step-wise decarboxylation occurred. Recently, a vanadium catalyst was reported to provide dicarboxy-starch in the presence of nitric and sulfuric acids.134 Oxidation of the 6-CH20H was also observed. Careful control of the temperature and the pH of the reaction mixture is necessary to limit acid-catalyzed hydrolysis of the glycosidic linkage. The ability of vanadium salt to perform oxidative scission of 1,2-diol systems was demonstrated earlier.135,136 We have studied the oxidative properties of nickel peroxide generated in situ by hyp0chlorite.~8This catalyst is able to perform oxidative scission of diol groups as well as oxidation of the primary hydroxyl group.137 The use of a nickel electrode as a source of nickel peroxide in the oxidation of carbohydrates showed a selectivity towards the 6-CH20H.47.879138 We have observed the same selectivity when using maltodextrins as substrate. However, the application of this catalytic system in the oxidation of methyl a-D-glucopyranoside as model compound yielded a dicarboxy-derivative with removal of 3-C as formic acid, while the primary hydroxyl group was resistant to oxidation. We believe that steric hindrance plays a major role in determining the selectivity with this system.

H*

Ni (I1),NaOCJ

HO

-oocLY~j+

HCoo-

pH =lo

OMe

OMe

Another promising metal catalyst that can be applied in diol-cleavage of carbohydrates is ruthenium tetroxide. A cheap oxidant such as hypochlorite can be used as terminal 0xidant.13~

Metal-catalyzed oxidation and reduction of curbohydmtes

51

This system was applied successfully in the preparation of optically pure D-glyceric acid by cleavage of 1,2-5,6-di-O-i~opropylidene-D-mannitol.~~~

HO-C-H I

I

H-C-OH

I

-

Ru/C or RuCI, NaOCl pH = 8

0

0-

qc/ I

c-0

c-0

The selective oxidative degradation of monosaccharides is also an important transformation. It can provide other carbohydrate molecules which are difficult to obtain from natural sources. One example is the Celv catalyzed degradation of D-glucose and/or D-gluconic acid to D-arabinose. 1 4 1 , 1 4 2The oxidizing species is generated in siti4 under acidic conditions by a platinum electrode.

COOCal/, H - L H H0-J-H H - L H

HC= 0

F-?

Ce'"

H - L H

I

CH,OH Calcium D-gluconate

CeIII

I I H---OH I

HO-C-

H

H-C-OH

I

CH,OH D-arabinose

D-Arabinose can also be prepared from D-ghconic acid by selective decarboxylation in the presence of Fell1 salt using aqueous H202 as the oxidant (Ruff degradation).I4'-145 Morgenlie

et al. have reported the degradation of aldoses and ketoses into lower sugars by cleavage of the endocyclic carbon-carbon bond next to the anomeric center when using silver carbonate suspended on Celite as reagent. 146-148 Recently, the Delft group developed a degradation method for aldoses to the next lower aldoses using hydrogen peroxide as the oxidant and boric acid as both a catalyst and a product-protecting agent.l49vl5o Thus lactose is degraded to yield 76% galarose. In this case, borate ester stabilities determine whether the method will be efficient. When the starting material is 2-keto-D-aldonic acid, D-pentonic acid is directly obtained by the action of FeWH202.l5l The same result can be achieved by indirect electrochemical oxidation with bromine. 2-keto-D-gluconic acid was thus degraded to D-arabinonic acid i n 65% yield.152.153

52

A . Abbadi and H. van Bekkum

The use of vanadium salt as catalyst in the presence of nitric acid to carry out the oxidative degradation of D-gulono-lactone leads to the cleavage of the 2C-3C bond and the simultaneous oxidation of the primary hydroxyl group at the C-6 position with formation of oxalic acid and D-threaric acid (dextro-tartaric

136

3. Reduction The reduction of the carbonyl group of carbohydrates is an important industrial process. Various catalytic and stoichiometric systems have been employed to carry out this reaction.154 However, the use of metal catalysts such as Pt, Pd, Ni and Ru acting in conjunction with molecular hydrogen is the preferred method for industrial application. When the monosaccharide is brought into contact with a metal catalyst of this type under a hydrogen atmosphere, polyols are formed. Polyol compounds are widely used in the food industry as low caloric and non-cariogenic sweeteners. They also serve as starting material for further chemical and biochemical modifications such as the conversion of sorbitol to dianhydrosorbitol and to L-sorbose. 155 When severe conditions are applied during hydrogenation, hydrogenolysis prevails and the carbohydrate molecule is broken down into smaller polyhydroxy compounds. The additional presence of an amine in the reaction mixture during the reduction of the carbonyl group leads to the formation of I-(a1kyl)amino-Ideoxypolyols by reductive amination. Amines with long alkyl chain yield non-ionic surface active compounds. In this section, we will first review the reduction of carbohydrates to polyols without scission of C-C bonds. The second part will focus on reductive amination of carbohydrates. Finally, carbohydrate hydrogenolysis will be briefly discussed.

3.1 Hydrogenation of Carbonyl Groups The most important carbohydrate reduction product from an industrial perspective is sorbitol (D-glucitol), which is obtained by hydrogenation of D - g l ~ c o s e . World-wide ~ ~ ~ - ~ ~ ~production is estimated at 650,000 tons per year. Sorbitol is applied in numerous cosmetic, food and beverage formulations and as a starting material, e.g. in the manufacture of ascorbic acid (vitamin C). Sorbitol was originally prepared by electrochemical reduction of D-glucose,160 but nowadays its manufacture is based almost completely on catalytic hydrogenation using nickel as catalyst. Other metals such as platinum and ruthenium were also reported to be effective catalysts for the hydrogenation of D-glucose to sorbit01.~5~ Homogeneous ruthenium complexes were also described as good catalysts for the hydrogenation of D-glucose. 161.162

Metal-catalyzed oxidation and reduction of carbohydrates

53

In a batch process, a 45 - 50% (w/v) aqueous solution of D-glucose is hydrogenated in the presence of Raney nickel (3 - 6% on D-glucose) at pH = 5-6 and 120 - 150 T , under 30 - 70 bar hydrogen. In a continuous process, higher hydrogen,pressures such as 170 bar with supported nickel - for example nickel on silica - are applied. When corn starch hydrolyzate is used as starting material instead of D-glucose, C12 and C18 polyols are formed as byproducts. Pure C12 polyols can also be prepared by catalytic hydrogenation of the corresponding disaccharide on a nickel ~ a t a 1 y s t . lAccording ~~ to this procedure, lactitol and maltitol are obtained from lactose and maltose , respectively. Lactitol is currently produced on a commercial scale by PURAC in the Netherlands. Another industriallyimportant C12 polyol is isomalt (palatinit) which is produced (Sudzucker - Germany = 10.000 tons each year) by catalytic hydrogenation of isomaltulose (palatinose) on Raney nickel catalyst.164.165 Isomalt is an isomer mixture of two polyols as a result of the hydrogenation of the keto group of isomaltulose. The starting material, isomaltulose, is obtained by bacterial conversion of sucrose.

bov OH

HO

OH

HO CH ?OH

Ni/HL

0 + O H

HO

OH OH

OH

OH

lsomalt "Palatinit"

OH

lsomaltulose OH

OH

OH

By contrast, a mixture of high polyols was prepared by the simultaneous action of a- or P-amylase and nickel catalyst on starch (maize) under hydrogen pressure.166 A syrup with DE 1 1 and 82% solubles was obtained. Higher DE level syrups cannot be obtained by this combined hydrolysis-hydrogenation process since inhibition of the enzyme by the nickel was found to occur. Recently, an efficient single-step catalytic process was developed for the conversion of polysaccharides of the glucan type, especially starch, to sorbitol.167 This process is characterized by the simultaneous hydrolysis of the polysaccharide and hydrogenation of the liberated monosaccharide. The catalyst used is Ru-loaded H-USY zeolite (3% wt Ru) in which the zeolitic material fulfills the role of metal carrier (Ru) and solid acid catalyst. The zeolite seemed to provide the BrGnsted acidity required for the hydrolysis reaction while the Ru catalyzes the hydrogenation of D-glucose to sorbitol.

54

A. Abbadi and H. van Bekkum

CH,OH

I

H-C-OH

Ru-H-USY

I I H -C-

______)

H,(55 bar)

H

HO -C-

, 180°C

OH

I

I

H-C-OH

I

CH,OH

Sorbitol Typical reaction conditions for the conversion of starch are: batch autoclave, 180 "C, 55 bar concentration 30% wt, Ru/starch ratio 0.002 (w/w). Under these conditions essentially quantitative conversion is reached within 1 hour. The selectivity to sorbitol is >95% and the catalyst can be re-used many times. Similar results are obtained by combining a 5% Ru/C catalyst with a heterogeneous acidic zeolite catalyst (H-USY, H-mordenite or H-ZSM-5). It has been also reported in the literature that Ru/C was the catalyst preferred for carrying out the hydrogenation step in a combined process in the presence of a homogeneous inorganic acid.168 The use of other metal catalysts such as Pt/C leads to degradation of the reaction products and hence moderate selectivity toward sorbitol. The second most important polyol is rnannitol. It is used in chewing gum, for example, and in pharmaceutical preparations.169Mannitol can be prepared by catalytic hydrogenation of the keto group of D-fructose followed by purification by crystallization since sorbitol is formed as we11.154 The water solubility of mannitol is much lower than that of sorbitol, enabling selective crystallization. H2, starch

CH,OH

CH,OH

I c= 0 I HO-C-H I H-C-OH I H-C--OH I

CH,OH

D-Fructose

CH,OH

I

I I

H-C-OH Cat.

H,

____)

I I H-C-OH I

HO-C-H

H-C--OH

I

CH,OH Sorbltol

HO-C-H

+

HO-C-H

I

H-C-OH

I I

H-C--OH CH,OH Mannitol

The selectivity of the hydrogenation towards mannitol depends on the metal catalyst used and I ~ ~ a small amount of borate is added to varies from 40% on Ru/C to 70% on Cu on s i 1 i ~ a . When the reaction mixture in the presence of Cu catalyst, the selectivity towards mannitol is increased and reaches 90%.170-172

Metal-catalyzed oxidation and reduction of carbohydrates

55

Extensive studies were carried out to understand the mechanism of D-fructose hydrogenation on copper catalysts. Experiments performed with deuterium clearly showed that the enediol form is not involved in hydrogenation and D-fructose is assumed to be preferentially hydrogenated via its furanose form by attack of a copper hydride-like species at the anorneric carbon with inversion of configuration.I73 The proposed mechanism could explain the diastereo selectivity obtained in the hydrogenation of the other ketoses according to the a l p ratios of their furanose forms present in solution. For instance, D-xylulose which is present in solution in a ratio of 62.3:18.1:19.6 of p-furanose:a-furanose:openform174.17s yields 65% arabitol under reductive conditions. The hydrogenation of invert sugar ( 1/1 D-glucose/D-fructose mixture), as starting material for mannitol, over nickel catalyst yields a mixture of 7/3 sorbitol-mannitol from which the mannitol is obtained by crystallization. I76 Mannitol can also be obtained from D-glucose when the hydrogenation is carried out under conditions that allow its isomerization to D - f r ~ c t o s e . The ~ ~ ~use - ~of~ calcium ~ hydroxide or sodium bicarbonate-sodium hydroxide as alkaline agent for the isomerization of D-glucose, in the presence of Raney nickel as the hydrogenation catalyst, yielded 27% mannitol.I*o Another approach by the Delft group for the preparation of mannitol from D-glucose is the use of a bi-catalytic system based on the cooperation of glucose isomerase and copper catalyst.l70.'71 The role of glucose isomerase in this bio-chemo-catalytic system is to perform the isomerization of D-ghcose to D-fructose while copper simultaneously catalyzes the hydrogenation step. Moreover, the use of enzyme to carry out the isomerization step offers the advantage of avoiding alkaline degradation reactions that take place when the isomerization is catalyzed by alkaline agents such as calcium hydroxide. Typical operation conditions are: 60 g invert sugar in 200 mL water, 5 g 20% copper on silica, 0.1 g Na2B407.1OH20 (as selectivity enhancer of the copper catalyst), 3 g Optisweet 22 (8% glucose isomerase immobilized on silica), 0.3 g MgS04 (enzyme-stabilizing cation), 0.05 g EDTA (to protect the enzyme against traces of copper ions), 0.5 g CaC03 (as buffering agent), pH = 7.1-7.6, 70 "C and 50 bar hydrogen. Under these conditions, the yield of mannitol exceeds 60%. However, long reaction times are required (60 to 80 hours) due to poisoning of the copper catalyst, probably by adsorption of mobile fragments originating from the immobilized enzyme system. Ruddlesden and Stewart181 reported on the combination of glucose isomerase with Ru-loaded zeolite Y as bi-catalytic system for the preparation of mannitol from D-glucose. The advantage in using zeolite as carrier for the hydrogenation catalyst is to build a barrier through the pore system of the zeolite in order to avoid the accessibility of the metal catalyst to the inhibiting species which are assumed to be large molecules. A yield of 29% mannitol was achieved with this system.

56

A. Abbadi and H.van Bekkum

Also, a multi-step process was developed for the preparation of mannitol from D-glucose. l x 2 , 1 8 3 First D-glucose is epimerized to D-mannose by molybdate. The remaining D-glucose in the mixture is submitted to the action of glucose isomerase to reach equilibrium with D-fructose. Finally, the glucose-mannose-fructose mixture is hydrogenated to yield 40% mannitol. We can also mention the oxidative dehydrogenation of D-glucose to D-gluconic acid over a Pt or Rh catalyst at high pH and simultaneous reduction of D-fructose to a mixture of mannitolsorbitol (see oxidation section 2.1). Nowadays, Inulin, which is glucose-(fructose),, is available in large quantities. It seems to be a logical raw material for the preparation of mannitol. Presumably, hydrolysis and hydrogenation can be performed consecutively or in a combined process. Another polyol with promising applications is xylitol (3,000 tons per year). It is prepared by hydrogenation of ~-xylose.Is4Xylitol is as sweet as sucrose and is applied in a number of food applications, including chewing gum. The non-catalytic reduction of aldonolactones by sodium borohydride or lithium aluminum hydride to the corresponding aldoses and/or alditols was also studied.IS4 Due to the stoichiometric character of these procedures, however, they are limited to laboratory use.

3.2 Reductive amination The hydrogenation of carbohydrates over Raney nickel or platinum in the presence of amines leads to the formation of amino derivatives.I84.185 For instance, D-glucose and D-galactose in liquid ammonia are hydrogenated in the presence of Raney nickel under 50-100 bar hydrogen and 40- 120 "C to yield glucitylamine and galactitylamine, respectively.186 Firstly, the aldose reacts with ammonia (or alkylamine) to form (alky1)-glycosylamine, which is hydrogenated in-situ to yield the corresponding 1-(alky1)-amino-1-deoxyalditol. Recently, a process using a fixed-bed nickel catalyst for the preparation of I-amino- 1-deoxyD-glucitol (glucitylamine) from D-glucose was developed.lx7 Catalytic reductive amination was also applied to various disaccharides such as lactose, maltose and isomaltulose to yield (alky1)aminodeoxypolyols.188 Other metal catalysts such as Pt and Pd were also used to carry out the reductive amination of carbohydrates.I89-191 1-Benzylamino- 1-deoxypolyols were also prepared which, upon the removal of the benzyl group by hydrogenolysis, give the corresponding free 1-amino- 1-deoxyalditols. 192.193 On the other hand, the addition of ammonium chloride to methanolic ammonia containing D-glucose catalyzes the formation of diglucosylamine which upon hydrogenation gives bis( I-deoxy-Dglucit- 1-yl)amine (dialditylamines).194

Metal-catalyzed oxidation and reduction of carbohydrates

57

Hydrogenation of aldose oximes (a1doses:D-arabinose,D-mannose and D-galactose) over 5% Pt/C catalyst, under 100 bar hydrogen and 50 "C gave a quantitative conversion of the oximes into a mixture of mono- and dialditylamines, the latter of which were isolated in fairly good yield (26-80%).195 The preparation of N-(4,5-dimethyl-2-nitrophenyl)-D-ribosylamine followed by hydrogenation to yield 1-(2-amino-4,5-dimethylanilino)l-deoxy-~-ribitolwas also achieved. This compound is condensed with alloxan to form vitamin B2 (lactoflavin).l96 CH,OH

HO--C-H

Lactoflavin

I

I

HO-C-H

Recently, the Delft g r o ~ p has l ~ reported ~ on the reductive amination of aldohexoses over Pt/C with mono- and bi-functional alkylamines. The chemical nature of the species present i n solution was studied by IH and 13C NMR. When the reductive amination was carried out in the presence of ethylenediamine, the resulting product was submitted to carboxymethylation to form an EDTA-like (ethylenediaminetetraacetate)complexing agent aiming at improved biodegradability. The sequestering capacity of such compounds was also investigated. - O o c 7

N H,C

nr

J-

coo -

N

L coo-

I I HO-C-H I HO-C-H I H-C-OH I CH,OH H-C-OH

N-(1 -deoxy-D-galactitol-l-yl)ethylenedlaminetriacetate

58

A. Abbadi arid H. van Bekkum

Amino sugars obtained by the reductive amination with alkylamine instead of ammonia can find various applications. 1 -Deoxy- 1-methylamino-D-glucitol, for instance, is used as hydrophilic component in particular to obtain water-soluble salts of X-ray contrast materials as 2,4,6-triiodobenzoic acid. By using long chain alkylamines amino sugars with application as surfactant, liquid crystalline material in polymers, cosmetics and i n pharmaceuticals are prepared. 19*,199

3.3 Hydrogenolysis Hydrogenation of carbohydrates under severe conditions (high temperature and pressure) leads to the formation of polyols followed by isomerization as well as scission of C-C bonds. Isomerization at chiral centers has been demonstrated in the case of D-glucose when the setting time during hydrogenation was prolonged. Sorbitol, the primary hydrogenation product of D-glucose, leads to the formation of a mixture of hexitols when heated under alkaline conditions at 170 "C and H2 pressure (130 bar) in the presence of nickel catalyst. The composition of the polyols mixture after 3 to 4 hours is as follows: 41% sorbitol/31% mannitol/26% idito1.200 Dehydrogenation at C-1 or C-6 followed by alkaline isomerization through the enediol and hydrogenation back to a hexitol was suggested as a mechanism.201 Dehydrogenation at C-2 followed by hydrogenation might also account for the mannitol formation. It was also reported that under the conditions described above alditols and methyl glycosides exhibited deuterium exchange at carbon atoms bound to free hydroxyl groups with retention of configuration (deuterium oxide 100 "C in the presence of Raney nickel). This reaction could be considered as a route to 2H-labelled sugars.202 On the other hand, when carbohydrates are submitted to hydrogen pressures >I00 bar and temperatures >200 "C in the presence of a metal catalyst, hydrogenolysis takes place and the carbohydrate molecule is fragmentated to lower polyhydroxy compounds. Sorbitol, for example, (obtained by hydrogenation of D-glucose) was converted over a nickel catalyst at 215 "C into 40% gIyceroI.203 Glycerol can also be obtained together with 1,2-propanediol and ethylene glycol from hydrogenolysis of inulin.*o4.*05 A typical reaction procedure is as follows: 100 g inulin, 10 g catalyst (CuO/Ce02/Si02:95/5/100), 1 g Ca(OH)2, 150 mL methanol/water (3/1, v/v), 200 "C and H2 pressure 2 100 bar. Higher polyols such as pentitols and hexitols were also formed. The cleavage selectivity defined as EC3/xC1.5 amounts to 88% for inulin and 72% for sucrose. Hydrogenolysis of sucrose in a slurry type reactor in the presence of 5% Ru/C was recently reported.206An adsorbed complex was proposed to explain the selectivity toward the formation of 1,2-propanediol.

Metal-catalyzed oxidation and reduction of carbohydrates

59

A continuous process for the production of glycerol by catalytic high-pressure

hydrogenolysis of sucrose was also developed.207 Here, glycerol selectivity is almost 50%. Unfortunately, these processes have lost much of their initial attractiveness since glycerol, which has a market of about 0.6 Mt per year, is a byproduct in the growing production of free fatty acids (and bio-diesel) from fats and oils. Recently, Schuster et al. have reported on the selective hydrogenolysis of sucrose on a Co/Cu/Mn catalyst towards 1,2-pr0panediol.208~~~~ The reaction was carried out at 250 "C and a H2 pressure of 250 bar for 4 hours and exhibited a selectivity of 60% towards 1,2-propanediol. A decrease of the reaction temperature (200 - 220 "C) leads to the formation of a polyols mixture rich in 1,2,5,6-hexanetetroI and lower content of ethylene glycol and 1,2-propanediol. A mixture of polyols of this type was used after removal of 1,2-propanediol and ethylene glycol (by distillation)

in the manufacture of rigid polyurethane foam.2'0

4. Conclusion Oxidation and reduction are logical means for upgrading carbohydrates. Metal catalysis plays a dominant role in hydrogenation and reductive amination. Elaborated chemo-catalytic oxidation processes were developed, but bio-catalysis still prevails in this type of transformation of carbohydrates. The challenge for the chemist is to develop clean and economic chemo-catalytic systems that can compete with the bio-transformations.

References I

2

3 4

0. Theander, in W. Pigman and D. Horton (eds.), The Carbohydrates, vol. 1 B, Academic Press, New York, (1980) 1013. J.W. Green, in W. Pigman and D. Horton (eds.), The Carbohydrates, vol. IB, Academic Press, New York. (1980) 1101. H. van Bekkum, in F.W. Lichtenthaler (ed.). Carbohydrates as Organic Raw Materials. VCH, Weinheim, (1991) 289. H. Roper, in F.W. Lichtenthaler (ed.), Carbohydrates as Organic Raw Materials, VCH, Weinheim, (1991) 267. K. Heyns and H . Paulsen, Adv. Carbohydr. Cheni. Biochem., 17 (1962) 169. F.R. Venema, J.A Peters, and H. van Bekkum, J. Mol. Catal., 77 (1992) 75. Y.J.B.M. Schuurman, Thesis, Eindhoven University o f Technology (1992). E. von Gorup-Besanz, Ann., 118 (1861) 257. K. Heyns and R. Hcinemann, Ann., 558 (1947) 187.

60 10

II 12 13 14 15 16

17 18

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

A. Abbadi and H. van Bekkum K. Heyns and H. Paulsen. in W. Foerst (ed.), Newer Methods of Preparative Organic Chemistry, vol. II, Academic Press, New York, (1963) 303. K. Heyns, H. Paulsen, G. Ruediger and J. Weyer, Fortschr. Chem. Forsch., 11 (1969) 285. The Merck Index-, eleventh edition, Merck & Co, Inc., New York, (1989) 699. E.P. Yufera, C.A. Easas and A.A. Cnrrasco, Rev. Agroquirn. Technol. Alimentos, 1 (1961) 40, Chem. Absrr. 57 (1963) 163848. M. Monsigny, French Pat., 2,523,962. H.S. Ishell, US. Pat., 1,976,731 (1934). H.S. Ishell and W.W. Pigman, Bur. Stand. J. Res., 10 (1933) 337. C.S. Hudson and H.S. Ishell. J. Amer. Cheni. SOC., 51 (1929) 2225. H.L. Frush and H.S. Isbell, Methods Carbohyd. Cheni., 2 (1963) 14. G.J.K. Acresand A.E.R. Budd, Brir. Pat., 1,208,101 (1970). T. Kimura, T . Sugiura and T . Kiyoura, Jpn. Kokai, 76 63, I2 I (1 976). K. Hattori, 9 . Miya, M. Matsuda. M. Ishii, H. Saito, H. Watanahe and H. Takizawa, Jpn. Kokai, 78 40,713 (1978). H. Saito and M. Nozue, Jpn. Kokui TokkyoKoho, 80 07,230 (1980). G.N. Dorofeenko and E.I. Antsupova, Nauchn. Zap. Luganskogo Sel'skokhoz. Inst., 7 (1960) 208, Chem. Ahsrr., 58 (1963) 4633g. S. Morgenlie, Acra Cheni. Scand., 26 (1972) 2518. T. Kiyoura, T. Kimura and T. Sigiura, Jpn. Kukai, 76 52,121, (1976). M. Nakayama, A. Kimura. H. Eguchi and T. Matsui, Eur. Par., 48,974, (1982). H. Saito, S. Ohnaka and S. Fukuda. Eur. Pat., 142 725, (1985). P. Fuertes and G. Flkche, Eur. Pat., 233 816 (1987). M. Besson, P. Gallezot. F. Lahmer. G. Fleche and P. Fuertes, in J.R. Kosak and T.A. Johnson (eds.), Catalysis of Organic Reactions, Marcel Dekker Inc., New York, (1994) 169. M. Besson, F. Lahmer. P. Gallezot, P. Fuertes and G. Fliche, J. Catal., 152 (1995) 116. B.M. Despeyroux, K. Deller and E. Peldszus, in G. Centi and F. Trifiro (eds.), New Developments in Selectice Ox-idafion(Stud. Surf. Sci. Catal., vol. 59), Elsevier, Amsterdam. (1990) 159. P. Fuertes and G. Fleche, Eur. Par., 232 202 (1986). H. Kimura , A. Kimura and Y. Mitsuta, Jpn. fat., 62/269 748 (1987). H.E.J. Hendriks, Thesis, Eindhoven University of Technology ( 1 991). H.E.J. Hendriks. B.F.M. Kuster and G . 9 Marin, J. Mol. Catal., 93 (1994) 317. G. de Wit, J.J. de Vlieger, A.C. Kock-van Dalen. R. Heus, R. Laroy, A.J. van Hengstum, A.P.G. Kiehoom and H. van Bekkum, Curhohydr. Res., 91 (1981) 125. G. de Wit, A.P.G. Kiehoom and H. van Bekkum, Tetrahedron Lett.. (1975) 3943. G. de Wit, A.P.G. Kiehoom and H. van Bekkum, R e d . Trav. Chim. Pays-Bas, 98 (1979) 355. A.J. van Hengstum, A.P.G. Kiehoom and H. van Bekkum, Starch, 36 (1984) 317. A. Abbadi and H. van Bekkum. in preparation. A. Ahhadi. M. Makkee, W. Visscher, J.A.R. van Veen and H. van Bekkum. J . Carhohydr. Cheni.. 12 (1993) 573. A. Ahhadi and H. van Bekkum, J . Mol. Carol., 97 (1995) I I I , G.O. Aspinall and A. Nicolson, J . Chem. Soc., (1960) 2503. (3.0.Aspinall and A. Nicolson, J . Cheni. Soc., (1960) 3998. L.A. Edye, G.V. Meehan and G.N. Richards, J. Curhohydr. Cheni., 10 (1991) I I .

Metal-catalyzed oxidation and reduction of carbohydrates 46

D.L. Verraest, J.A. Peters and H. van Bekkum, unpublished resulrs.

47 48 49 50 51 52 53

A. Kerkenaar and A.C. Besemer, PCT., WO 91/04988 (1991). E.J.M. Mombarg, A. Abbadi and H. van Bekkum, in prepararion.

54 55

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88

S.A. Barker, E.J. Bourne and M. Stacy, Chem. Ind., (1951) 970. D.B. Easty, J . Org. Chem., 27 (1962) 2102. D. Keglevic, Adv. Carbohydr. Chem. Biochem., 36 (1979) 70. R.A. Reiners, U.S. f a r . . 2,845,439 (1958). A. Abbadi and H. van Bekkum, unpubLished results. P. Vinke, Thesis, Delft University of Technology (1991). T. Boecker and J. Thiem, Tens. Sutf Det., 26 (1989) 318. A.J. Hoefnagel and B.M. Wepster, Collect. Czech. Chem. Commun., 55 (1990) 119. H.E. van Dam, Thesis, Delft University of Technology (1989). H.E. van Dam, A.P.G. Kieboom and H. van Bekkum, Reel. Truv. Chirn. fays-Bas, 108 (1989) 404. C.L. Mehltretter, US. Pat., 2,559,652 (1951). C.L. Mehltretter, B.H. Alexander, R.L. Mellies and C.E. Rist, J. Anier. Cheni. Soc.,73 (1951) 2424. A.D. Voitenko, D. Feldmane, M.V. Shimanskaya and M. Lidaks, USSR. far., 883054 (1981). 0. Theander, Swed. Par., 325.409 (1970).

J. Bakke and 0. Theander. Cheni. Commirn., (1971) 175. G. Jayme and W. Demmig, Ber., 93 (1960) 356. Y. Hirasaka, Yakugaku Zasshi, 83 (1963) 960, Chem. Abstr.. 60 (1964) 4232c. Y. Hirasaka, Yakugaku Zasshi, 83 (1963) 971, Cheni. Absrr., 60 (1964) 4233d. G.G.S. Dutton and K.N. Slessor, Can. J. Chem.. 42 (1964) 1 1 10. D. Abbott and H. Weigel, J . Cht-m. Soc., (1965) 5157. Ferntan Pat., 886,305 (1953). W. Frirsche-Lang, I. Leopold and M. Schlingmann, Gernlan. far.,3.535.720 (1985). E.G. Skibida, I.P.A.M. Sakharov and A.M. Sakharov, Eur. f a t . , 0 548 399 (1991 1. M. Kunz et al., private conirnunication. M. Kunz, in F.W. Lichtenthaler (ed.), Carbohydrates as Organic Raw Materials, VCH, Weinheim, (1991) 127. K. Heyns, German Par.*692.897 (1936). 0. Dalmer and K. Heyns, Canadian Par.. 38 1,575 (1939). 0. Dalmer and K. Heyns, U.S. far., 2,189.778 (1940). 0. Dalmer and K. Heyns, U.S.Pat., 2,190,377 (1940). K. Heyns, Ann., 558 (1947) 177. C. Bronnimann, T. Mallat and A. Baiker. J. Chem. Soc.. Chem. Coniniun., (1995) 1377. Mitsui Toatsu Chemicals Inc., Jpn. Kokai Tokhyo Koho, 57 163 340, (1982). Mitsui Toatsu Chemicals Inc., Jpn. Kokai Tokkyo Koho, 60 54 338, (198.5). L. Zengmin and Z. Manzheng, Cuihua Xuebao. 10 (1989) 5 5 , Chem. Absrr., I I I (1989) 99276. L. Zengmin and 2. Manzheng, Cuihua Xuebao, I I (1990) 106, Chem. Abstr.. I13 (1990) 25837. C. Bronnimann, 2. Bodnar, P. Hug. T. Mallat and A. Baiker, J . Cafal., 150 (1994) 199. A. Abbadi and H. van Bekkum J. Appl. Caral.. 124 (1995) 409. A. Abbadi, K.F. Gotlieb and H. van Bekkum, Durch Par., 93,02127 (1993). H.J. Schafer and R. Schneider, Tetrahedron, 47 (1991) 715. R. Wittman, U.S. Par., 4,008.1 32 (1977).

61

62 89 90 91 92 93 94 95 96 97 98 99 100 101

I02 103 104

I05 106 I07 I08 109 I10 111

I12 I I3 I I4 I 15 1 I6 I17 I18 I19 120 121

122 123 I24 I25 I26 127 128 129

A. Abbadi and H. van Bekkum N.R. Trenner, U.S. Pat.. 2,428,438 (1947). N.R. Trenner, U.S. Pat., 2,483,251 (1949). D.F. Hinkley and A.M. Hoinowski, U.S. Pat., 3,721,663 (1973). J.M.H. Dirkx, H.S. van der Baan and J.M.A.J.J. van den Broek, Carbohydr. R e x , 59 (1977) 63. F.R. Venema, J.A. Peters and H. van Bekkum, J . Mol. Catal., 77 (1992) 75. G.A. Levy, Biochem. J . , 52 (1952) 464. E. Boyland and D.C. Williams, Biochem. J . , 64 (1956) 578. B.W. Lew, U.S. Pat., 3,326,944 (1967). B.F.M. Kuster, Starch, 42 (1990) 314. M. Kunz, in A. Fuchs (ed.), Iriulin and Irrulin-containing Crops, Elsevier, Amsterdam, (1993) 149. E.I. Leupold, M. Wiessner, M. Schlingmann and K. Rapp, German Pat., 3,826,073 (1988). P. Vinke, H.E. vam Dam and H. van Bekkum, in G. Centi and F. Trifiro (eds.), New Developments in Selective Oxidation (Stud. Surf. Sci. Catal., vol. 59), Elsevier, Amsterdam, (1990) 147. J. Smejkal and L. Kalvoda, Collect. Czech. Chern. Comrnun., 38 (1973) 1981. A.E.M. Boelrijk and J. Reedijk, J . Mol. Catal., 89 (1994) 63. J.N. BeMiller and S.D. Darling, U.S. Par., 3,632,802 (1972). 0. Theander, Acta Cliem. Scand., 17 (1963) 1751. Y. Hirasaka and 1. Matsunaga, Chem. Pharm. Bull., 13 (1965) 176. I. Johansson, B. Lindberg and 0. Theander, Acta Chem. Scand., 17 (1963) 2019. B. Lindberg, I. Svensson and 0. Theander, Acta Chem. Scand., 17 (1963) 930. J.S. Brimacombe and M.C. Cook, J. Chem. Soc., (1964) 2663. J.S. Brimacombe, M.C. Cook and L.C.N. Tucker, J . Chem. Soc., (1965) 2292. K. Heyns and J. Lenz, Angew. Clieni., 73 (1961) 299. K. Heyns, W.-D. Soldat and P. KOII, Chem. Ber., 108 (1975) 3619.

K. Heyns. J. Weyer and H. Paulsen, Ber., 100 (1967) 2317. P.P. Regna and B.P. Caldwell, J . Amer. Chem. Soc., 66 (1944) 243. H.S. Isbell, J. Res. Nut. Bur. Stand., 33 (1944) 45. P.C.C. Smits, Eur. Par., 151,498 (1985). P.C.C. Smits, Thesis, Eindhoven University o f Technology, 1984. P.C.C. Smits, B.F.M. Kuster, K. van der Wiele and H.S. van der Baan, Appl. Catal., 33 (1987) 83. D.T. Sawyer and J.R. Brannan, Inorg. Chem., 5 (1966) 65. D.T. Sawyer, Chem. Rev., 64 (1964) 633. S.J. Angyal and K. James, Aust. J. Chem., 23 (1970) 1209. A.C. Besemer, Thesis, Delft University o f Technology (1993). E.L. Jackson and C.S. Hudson, J . Anier. Chem. Soc., 59 (1937) 2049. J.W. Sloan, B.T. Hofreiter, R.L. Millies and LA. Wolff, Ind. Eng. Chem., 49 (1957) 350. G . Tegge, Starch. 12 (1960) 321. C.L. Mehltretter, Srarch, 15 (1963) 313. R. Criegee, Ber., 64 (I93 I ) 260. L.J. Heidt, E.K. Gladding and C.B. Purves, Paper Trade J., 121 (1945) 81. A S . Perlin, in W. Pigman and D. Horton (eds.), The Carbohydrates, vol. IB, Academic Press, New York, (1980) 1167. C.L Mehltretter, U.S. Par.. 2,830,941 (1958).

Metal-catalyzed oxidation and reduction of carbohydrates 130 131 132 133 I34 I35 136 I37 138 139 140 141 142 143 144 145 I46 I47 148 149 150

151

I52 153 I54 155

I56 157 I58 I59 160 161 I62 163 I64 I65 166 167 I68

63

V.F. Pfeifer, V.E.Sohns, H.F. Conway, E.B. Lancaster, S.Dabic and E.L. Griffin Jr, lnd. Eng. Cheni., 52 (1960) 201. J.A. Radley ed., Starch Production Technology, Appl. Sci., Barking, (1976) 423. S . Veelaert, D. de Wit and H . Tournois, P C T / N L 94/00272. 0 M. Floor, K.M. Schenk, A.P.G. Kieboom and H. van Bekkum, Starch, 41 (1989) 303. S.W. Heinzman and J.S. Dupont, Eur. Par., 542,496 (1993). A.F. Odell, U.S. Par., 1,425,605 (1922). J.K. Dale and W.F. Rice Jr, J. Amer. Chem. SOC.,55 (1 933) 4984. M.V. George and K.S. Balachandran, Chem. Rev., 75 (1975) 491. D. Court and D. Pletcher, Spring Meeting ofrhe Elecrrocheniisrry Soc., Abstr. Nr.334 (1984) Cincinnati Ohio. C.L. Stevens and C.P. Bryant, Methods Carbohydr. Chem., 6 (1972) 337. E.C. Emons, Thesis, Eindhoven University of Technology, 1991. Noguchi Institute, Jpn. Pat., 57 128,697 (1982). Noguchi Institute, Jpn. Par., 58 39,695 (1983). H.G. Fletcher Jr, Methods Carbohydr. Chem., 1 (1962) 77. R.L. Whistler and J.N. BeMiller, Methods Carbohydr. Chem., 1 (1962) 79. O.N. Richards, Methods Carbohydr. Chem., 1 (1962) 180. S. Morgenlie. Acta Chem. Scand., 26 (1972) 1709. S . Morgenlie, Acta Chem. Scand., 27 (1973) 1557. S. Morgenlie, Acta Chem. Scand., 27 (1973) 2607. R. van den Berg, J.A. Peters and H. van Bekkum, J. Chem. Soc. Perkin Trans.. (1994) I 1 17. R. van den Berg, J.A. Peters and H. van Bekkum, Carbohydr. Res., 267 (1995) 65. T.S. Gardner and E. Wenis, J. Amer. Chem. Soc., 73 (1951) 1855. C.L. Mehltretter, W. Dvonch and C.E. Rist, J. Amer. Chem. SOC.,72 (1950) 2294. C.L. Mehltretter and W. Dvonch, U.S. Pat., 2,502.472 (1950). J.W. Green, in W. Pigman and D. Horton (eds.), The Carbohydrates, vol. IB, Academic Press, New York, (1980) 989. P. Stoss and R. Hemmer. Adv. Carbohydr. Chem. Biochem., 49 (1991) 93. L.W. Wright, Chem. Technol., 42 (1974). P.H. Brahme and L.K. Doraiswamy, lnd. Eng. Chem. Process Des. Dev.. 15 (1975) 130. J . Wisniak and R. Simon, Ind. Eng. Chem. Prod. Res. Dev.. 18 (1979) 50. R. Albert, A. Stratz and G. Vollheim, Chem. lng. Technol., 52 (1980) 582. M. Fedoronko, Adv. Carbohydr. Chem. Biochem., 29 (1974) 107. U.S. Par., 3,935,284 (1976). W.M. Kruse and L.W. Wright, Carbohydr, Res., 64 (1978) 293. M.W. Kearsley and G.G. Birch, in G.G. Birch and K.J. Parker (eds.), Sugar: Science and Technology. Applied Science Publishers, London (1979) 287. H. Schiweck, G. Steinle and H. Haberl, German Pat., 2,217,628 (1972). G. Darsow and W. Biedermann, German Pat.. 3,403,973 (1984). M.W. Kearsley, S.H. Satti and I . Tregaskis, Starch. 32 (1980) 169. P.A. Jacobs and H. Hinnekens. Eur. Pat., 329,923 (1989). A.A Balandin, N.A. Vasyunina, S.V. Chepigo and G.S. Barysheva, Doklady Akad. Nauk. S.S.S.R., 128 (1959) 941, Chem. Absrr., 54 (1959) 7140f.

64 I69 I70

171 172 I73 I74 175 I76 177 I78 179 I80 181 182 183 I 84 185

186 I87 I88 I89 190 191 192 193 194 195 I96 I97 198 I99 200 20 I 202 203 204 205 206

A. Abbadi and H. van Bekkum Chem. Eng. NeK's, 17 (1972) 13. M. Makkee, Thesis, Delft University of Technology, 1984. M. Makkee, A.P.G. Kiehoom and H. van Bekkum, Carbohydr. Res., 138 (1985) 225, 237. J. Dewing, J.F. Ruddlesden, A. Stewart and D.J. Thomson, Eur. f a t . , 6313 (1980). J.F. Ruddlesden, A. Stewart, D.J. Thomson and R. Whelan, Faraday Discuss. Chem. Soc., 72 (1981) 397. J. Wu, T. Vuorinen and A S . Serianni, Carbohydr. Res.. 206 (1990) I . S. Angyal, Adv. Carbohydr. Chem. Eiocheni., 49 (1991) 19. A.J. de Berardinis, U.S. f a t . , 3,763,246 (1973). A.J. de Berardinis. Gertnan Pat., 2,025.3 14 (1971). L.W. Wright and J.D. Brander, Gernran Pat., 1,268,606 (1968). N.J. Gerharz, Dutch Pat., 6,515,786 (1966). H.C.M. Pijnenburg, B.F.M. Kuster and H.S. van der Baan, Starch. 30 (1978) 199. J.F. Ruddlesden and A. Stewart, J. Chem. Res., 378 (1981). T. Motohiro, 1. Mochihiro, T. Yoshiaki, 0. Yuji and M. Hiroyuki, U.S. Pat., 4,083,881 (1977). W.M. Kruse, U.S. Pat., 4,173,514 (1979). H. Paulsen and K.-W. Pflughaupt, in W. Pigman and D. Horton (eds.), The Carbohydrates, vol. IB, Academic Press, New York, (1980) 881. J.W. Long and G.N. Bollenback, in R.L. Whistler and M.L. Wolfrom (eds), Methods in Carbohydrate Chemistry, vol. 11, Academic Press, New York, (1963) 79. Kirk-Orhmer Etqclopedia of Chemical Technology, vol. 4, 4th ed., Interscience, New York, (1992) 925. H. Kelkenherg, Tens. Su$ Det., 25 (1988) 8. J. Klein, W. Behrens and M. Kunz, Eur. Pat., 225,033 (1987). R.B. Flint and P.L. Salzberg, U.S. Pal., 2,016,962 (1974). J.M.J. Tronchet, B. Baehler and J.-B. Zumwald, Helv. Chim. Acta, 60 (1977) 1932. J.M. Larkin, E.L. Yeakey and Jr. L.W. Watts, U.S.Pat., 4,540,821 (1985). F. Kagan. M.A. Rehenstorf and R.V. Heinzelmann, J. Amer. Chem. Soc., 79 (1957) 3541. 1. Christiansen-Brams, M. Meldal and K. Bock, J. Carbohydr. Chem., 11 (1992) 813. J.E. Hodge and B.F. Moy, J. Org. Chem., 28 (1963) 2784. J. van Haveren, H. Lammers, J.A. Peters and H. van Bekkum, Carbohydr. Res., 243 (1993) 259. R. Kuhn and R. Strobele, Ber., 70 (1937) 773. H. Lammers, J.A. Peters and H. van Bekkuni, Tetrahedron, 50 (1994) 8103. H. Koch, R. Beck and H. Roper, Starch, 45 (1993) 2. G.A. Jeffrey and L.M. Wingert, Liq. Crystals, 12 (1992) 179. L. Wright and L. Hartmann, J . Org. Chem.. 26 (1961) 1588. A.S. Perlin E. v. Rudloffand A.P. Tulloch, Can. J. Chem.. 36 (1958) 921. H.J. Koch and R.S. Stuart, Carbohydr. Res., 67 (1978) 341. I.T. Clark, hid. Dig. Chem., 50 (1958) I 125. A. Fuchs, in D. Schliephake and P. Kramer (eds), Agricultural Surpluses, Dechema, Frankfurt / Main, (1986) 171. G. van Ling, Thesis, Twente University of Technology 1967. P. Miiller. P. Rimnielin, J.P. Hindermann, R. Kieffer, A. Kiennemann and J. Carrt, in M. Guisnet et al. (eds.), Heterogeneous Catalysis and Fine Chemicals / I (Stud. Surf. Sci. Catal., vol 59). Elsevier, Amsterdam,

(1991) 237.

Metal-catalyzed oxidation and reduction of carbohydrates 207

G. van Ling, A.J. Driessen, A.C. Piet and J.C. Vlugter, Ind. Eng. Chern. Prod. Res. Develop., 9 (1970)

2 10. 208 209 210

LSchuster and W. Himmele, Eur. far., 344,561 (1988). L.Schuster, Eur. Pat., 415,202 (1989). R. H e m , W. Hinz, L. Schuster and K. Vorspohl, Eur. Par., 458,095 (1990).

65

This Page Intentionally Left Blank

4 Production and use of inulin: Industrial reality with a promising future

L. De Leenheer Raffinerie Tirlemontoise, Aandorenstraat 1, B-3300 Tienen, Belgium

Summary. Inulin used to be a mere scientific topic. Today it has become an industrial reality. The chemistry of inulin. its agriculture, biochemistry, production, nutritional and functional properties, as well as its applications and future developments will be discussed. Inulin is a polydisperse very slightly p (2+6) branched p (2+1) fructan. The average DP (degree of polymerization) is dependent on both the plant source and the time of harvest. A small fraction (< 1 % ) of the 'native' inulin does not contain any glucose units at all (Fm). In our temperate climate, chicory is the most appropriate crop for inulin production . The average inulin content of industrially harvested crops varied only slightly from 16 to 17% these last 5 years. A yield of 40-45ton roots h a ' is considered to be normal though yields can vary from 25 to 75 tons. Two fructosyltransferases (SST and FFT) are currently considered to be the key-enzymes for the synthesis of the inulin chain. Breakdown is caused by the fructan exohydrolyse enzyme, producing fructose - a process which seems to be significantly influenced by (freezing) temperature: the later the harvest, the shorter the inulin chain will be. Inulin production goes through two phases. First we have extraction and epuration via lime carbonatation, similar to sugar beet processing, The second phase is more akin to starch processing and conists of demineralisation through ion exchange resins and decolorisation with active carbon. lnulin is a food ingredient, not an additive, with most interesting nutritional and health promoting characteristics: it is a soluble dietary fibre and a Bifidus promotor; it has only 1/4 of the calories of sugar and is suited for consumption by diabetics. It has a neutral flavour without any off-flavour or aftertaste, and it improves mouthfeel. Moreover, a creamy texture can be obtained by simply shearing an aqueous solution of inulin, which makes it an ideal fat-replacer. Future developments in the non-food area are eagerly expected. The derivatization of inulin through oxidation, esterification, etherification, cross linking etc ... will open the door to yet undreamed of applications.

1. Introduction

Inulin used to be a topic of purely scientific interest, with papers even dating back to 18 18.1 In the recent years, this topic has become more current and vibrant than ever because of its industrial applications.

68

L. De Leenheer

In this paper, different aspects concerning the production of inulin and its applications will be discussed, covering inulin's analytical aspects, agriculture, biochemistry, production, functionality, nutritional properties and applications as well as some future developments.

2. Chemistry and analytical aspects Inulin has been defined2 as carbohydrate material consisting mainly, if not exclusively, of p (2+1) fructosyl - fructose links. A starting glucose moiety can be present but is not necessary. Levan, on the other hand, consists mainly or exclusively of p (2+6) fructosyl - fructose links. As is the case with inulin, glucose is allowed but not necessary. Fructan is the more general name used for any compound in which one or more fructosyl - fructose links constitute the majority of linkages. The term 'Fructan' therefore covers both inulin and levan. Going back to the definition of inulin then, both GF, and F, compounds are considered to be included under that same nomenclature. In chicory inulin, n, the number of fructose units linked to a terminal glucose can vary from 2 to more than 70 units.3 The presence of F, compounds in native inulin extracts, however, has been demonstrated regularly. The analytical techniques backing up this statement are based on gas chromatography and HPAEC - PAD chromatography. With HPLC, only a crude separation into fructose, glucose, DFA (Difructose-Dianhydride), GF, F2, F3, DP3, DP4 and 2 DP5 is obtained. (Table 1). Gas chromatography on an A1 Clad OV1 capillary column (Alltech) temperature programmed from 105 "C to 440 "C, however, permits us to separate compounds quantitatively up to DPlO; it also enables us to differentiate between GF, and F, compounds (Table 2). Table 1. HPLC analysis of native chicory inulin (extract November '93). ~

% on carbohydrates

native chicory inulin

Fructose Glucose DP2 (DFA + F2) Saccharose F3 DP3 DP4 DP5+

0.3 0.0 0.1 6.4 0.3 4.3 4.6 84.1

Production and use of inulin

69

Table 2. Gas chromatographic analysis of native chicory inulin (extract November '93).

in % on total carbohydrates

native chicory inulin

Fructose Glucose Saccharose DFA

0.2 0.0 6.5 0.0

F3 GF3 F4 GF4 F5 GF5 F6 GF6 F7 GF7 F8

0.0 4.1 0.2 4.4 0.2 4.8 0.2 4.6 0.2 4.1 0.0 3.6 0.0

GF8 F9 DPlO

3.1 0.0 2.2

> DPlO sum GFn (in range DP2-DP9) sum Fm (in range DP2-DP9)

61.6 28.7

F2 GF2

0.8

Using this technique, we found that the 'DP3' of HPLC corresponded in reality to 'GF2' and 'F4' as determined by gas chromatography. Another technique which also differentiates between GF, and F, compounds, and which also provides a fingerprint of the Molecular Weight distribution of inulin, is High Pressure Anion Exchange Chromatography (HPAEC). This analytical technique uses a Dionex series 4000 ion chromatograph coupled with a Pulsed Amperometric Detector (PAD). During the analysis the carbohydrates are eluted under alkaline conditions; the high pH (13-14) of the eluant converts the hydroxyl groups into oxyanions. The degree of oxy-anion interaction with the anionic exchange resin then determines the carbohydrate retention times. To reduce the retention times, a competing ion such as acetate is

70

L. De Leenheer

added to the eluant. The PAD system oxidizes and detects the now separated carbohydrates as they pass through the detector. Fig. 1 shows a Dionex of this type for native chicory inulin and the refined product RAFTILINEB.

!I Y

3

\

I

'

Native inulin

n u P

c

0 4

Fig. 1. Dionex chromatograms for native chicory and RAFTALINE.

Production and use of inulin

71

The major drawback of HPAEC-PAD is that as yet it is impossible to quantify high DPoligomers, due on the one hand to the lack of appropriate standards and on the other to the reduced sensitivity of the PAD detector for high DP polymers. The detector actually measures the electrons released during the oxidation of the carbohydrates at the gold electrode. Chatterton et al. suggest that as carbohydrates become larger, proportionally fewer electrons are released per fructosyl unit and so the PAD output per pg sugar decreases as the DP is increased.4 Researchers at the A T 0 (Instituut voor Agro-Technologisch 0nderzoek)j also worked with HPAE chromatography, now coupled with a Pulsed Electrochemical Detector (PED). The sensitivity of the PED detector decreased clearly from DP2 to DPS, but they noted only a slow decrease for DP 10 - 17. No data were available for DP5 to DPIO. From this they calculated the PED responses for inulin oligomers with different DP's relative to sucrose. This enabled them to quantify oligomers up to DP17. Based on these relative responses they were then able to calculate the number-average and the weight-average DP and hence the dispersion of inulin as expressed in -

D=

-

DPn Current methods based on the hydrolysis of inulin, however, simply cannot provide us with a measure of dispersion F (DPn= -+ 1). G And yet isolating well-characterized and pure DP fractions from inulin that can serve as standards remains a challenge. Because the fractions obtained through preparative RP - 18 HPLC chromatography are not pure,3 we attempted to achieve a pure separation via preparative Dionex itself. Unfortunately, this process is tedious. The degree of polymerization (DP) of inulin as well as the presence of branches are important properties because they influence inulin's functionality in a highly strikingly manner. The DP of inulin varies according to plant origin, weather conditions (during growth) and the physiological age of the plant - in other words the time of its harvest (see Agricultural aspects). Native chicory inulin has an average DP, of 10 to 14. For native Dahlia and native Jerusalem Artichoke an average DP, of 20 and 6, respectively, has been reported.6 By 'native' - in this particular context - we mean inulin that is extracted from fresh roots prior to its analysis, taking precautions to inhibit the plant's own inulinase activity as well as acid hydrolysis. Moreover, no fractionation procedure is applied to eliminate smaller oligosaccharides and monomers. This is also the reason why we do not consider the commercially-available inulin from Sigma derived from Dahlia, Jerusalem Artichoke and Chicory to be 'native'.3,25 These

12

L. De Leenheer

products hardly represent the inulin so typical for the plants from which they are extracted. Not only is their average DP 27-29, which is very high, but chains smaller than 10 units are absent (Fig. 2). m N

a 0

0 0

4 0

:

2

Sigma's Dahlia i n u l l n

4

n

0

1 4

0

0

O

t

5

2

"

0

Fig. 2. Dionex chromatograms for Sigma's inulins.

Sigma's Chlcory I n u l l n

Sigma's J.rUslla.

Artichok.

inulln

Production and use of inulin

73

Until recently inulin was considered to be a linear molecule. By optimized permethylation analysis it was possible to demonstrate that even native chicory inulin (DP 12) has a very small degree of branching (approximately 1%); analysis of Dahlia inulin from Sigma with a DP of 27 showed that 5% of the fructose molecules present were branched. Regarding the structural chemistry of inulin, we refer to the work of French.7.8 However, no clear model of what inulin looks like is available to date. Computer models and crystal structures were established for Inulobiose (F2), 1-kestose (GF2) and nystose (GF3). From these, models of inulin were built based on the dominant values for the inter-residue linkage conformation and the geometry of the fructo-furanose rings. It is commonly known that the various fructose molecules in the GF, forms of inulin are all present in the furanose form. Only in the F, forms is the reducing fructose in the pyranose form9 (Fig. 3).

5cx2

(-A

HOHsC

HO

HO

p!+I

HOHpC

*OH

Fig. 3. In Fm forms of inulin the reducing fructose is in the pyranose form.

In trying to predict the likely shapes of the furanose rings, the energy of various model rings was calculated. The most likely furanose ring form is 4T3. T-forms have three atoms in a plane and the other two adjacent atoms placed in opposite directions above or below that plane. However, a continuum of ring shapes seems to occur, requiring little energy for conversion. Thus the fructo-furanose ring seems very flexible.

14

L. D e Leenheer

From the study of the two F-F rings in the above-mentioned compounds, French proposes 3 models; the respective energies of the models are quite similar and none has intramolecular hydrogen bonds! The numbers of residues per turn and the rise-per-residue values are about -4.5,1.3 8,; 2, 2.75 8, and 4,2.15 8, (Fig. 4).

Fig. 4.3 Inulin models as proposed by French.

3. Agricultural aspects After starch, fructans are the most-abundant non-structural polysaccharides found in nature. They are present in a wide variety of plants and in some bacteria and fungi. Most literature on the natural occurrence of fructans, however, does not differentiate between levan and inulin. Strictly speaking, plants containing inulin primarily belong to either the Liliales, e.g. leek, onion, garlic and asparagus, or the Compositae, e.g. Jerusalem Artichoke, Dahlia, Chicory and Yacon (Table 3). Given their high inulin content (> 15%) Dahlia, Jerusalem Artichoke (Helianthus tuberoses) and Chicory (Cichorium intybus) could initially be considered to be good candidates for industrial production in our temperate regions. Many Dahlia cultivars (cultivated varieties) are available, but they have all been selected for their flowers rather than for their inulin production. The tuberous roots have no buds and can be propagated only if attached to a piece of stem tissue, which hampers crop establishment of tubers.

Production and use of inulin

75

Table 3. Inulin content (as % on fresh weight) in plants that are commonly used for human nutrition.

source

edible parts

dry solids content

inulin content

Onion Jerusalem Artichoke Chicory Leek Garlic Artichoke Banana

bulb tuber root bulb bulb leaves-heart fruit

Rye Barley Dandelion Burdock Camas Murnong Yacon Salsify

cereal cereal leaves root bulb root root root

6-12 19-24 20-25 15-20* 40-45* 14-16 24-26 88-90 NA 50-55* 21-25 31-50 25-28 13-31 20-22

2-6 14-18 15-20 3-10 9-16 3-10 0.3-0.7 0.5- 1 * 0.5- 1.5* 12-15 3.5-4.0 12-22 8-13 3-19 4-1 1

NA : figures not available,

* estimated

When propagated from seed, sowing must be delayed until late spring, given the Dahlia's extreme sensitivity to frost.10 Mechanical harvesting of the tubers is feasible only on sandy grounds. Although the mean DP of Dahlia inulin is higher than that of the chicory-inulin, its yield is only half that of chicory. (Table 4). These considerations make Dahlia a less-interesting inulin crop. Jerusalem Artichoke is another candidate with a rather high inulin content (14 - 18%). This plant has an annual life cycle. Two types of plants exist: early cultivars and late cultivars. Although the date of planting is the same (late March - early April), early cultivars attain their final height at the end of July (+ 140 cm), which coincides with the start of flowering, whereas the growth of late cultivars does not cease until the middle of October (k280 cm)." A relatively large fraction of the total dry matter production in Jerusalem Artichoke, however, is diverted to structural stem dry matter (approximately 4 to 9 t ha-1). Only a smaller part is devoted to inulin storage (4 to 8 t ha-l). At tuber harvest, these stems contain mainly cellulose, hemicellulose and lignin, resulting in a rather poor feed and fibre quality.10 Although early cultivars yield about 8 t inulin ha-1, only

76

L. De Leenheer

60% of this inulin has a DP 2 4. For late cultivars, the transport of photosynthates from the tops into the tubers is slow, yielding only 5 t inulin ha-1 with 70% as 2 DP4 fraction. In addition, the extract of Jerusalem Artichoke tubers provides not only glucose, fructose, and inulin oligomers but also galactose.'* Table 4. Yields and composition for Dahlia, Jerusalem Artichoke and Chicory.

Rootsflubers;

tonha variation

DM %

average variation

Inul. %

Inul. tonha Mean DP,

Dahlia

Jerus. Artichoke

Chicory

25

45 35-60

43* 25-75*

19-24

22,3* 20-25*

14-18

14.9- 18.3*

2.5-3 20

4.5-8.5 6

5-1 1 10-14

39 21 40

52 22 20 6

31 24 28 17

average variation

12

DP distribution DP<10 } DP 10-19 ] DP 20-40 DP > 40 (Ref : 6 , 10, 1 1 , 12 and own results;

* industrial figures 1993)

The production of Jerusalem Artichoke tubers is comparable to potato production. It is not recommended to cultivate Jerusalem Artichoke in clay-soil because the tubers are small and irregular, hence bringing with them a lot of soil attached to the tubers. To their advantage, Jerusalem Artichokes are frost-tolerant. In theory they could be harvested during winter and even early spring, so that a very long processing period would be possible. A big disadvantage, however, is that during this long frosty period the inulin chain length would continue to decrease significantly.

Production and use of inulin

77

Chicory, then, is the third candidate for industrial inulin production and, as we will see, the favored one! Chicory is a biennial plant. During the first season the chicory plants remain in the vegetative phase and make only leaves, tap-roots and fibrous roots. This makes them very efficient plants as they store their assimilates in the tap roots. These roots look like small oblong sugar beets. The inulin content is high and fairly constant from year to year for a given region. During the period 1990-1994 the mean inulin content of the industrially-harvested chicory crops varied from 16 to 18%. This variation does not seem to be dependent on the type or quantity of fertilizer used. High N-dosages (> 80 U) do not improve either the inulin concentration or average dry matter yield. Yields in ton roots ha-1, however, show a wider variation. The figures reported in Table 4 are industrial yields, i.e. ton chicory roots received at the extraction plant per ha harvested. These yields are 15 to 20% less than theoretical yields, because during the harvest some of the small ends of the roots and tiny side-roots are broken off and left in the ground. The harvest is done with modified sugar beet machines because the chicory roots are somewhat thinner. Although the production of chicory is comparable to that of sugar beet, the culture is more drought-resistant. After the dry July 1994 month, sugar beet fields could be clearly distinguished from chicory fields; the leaves on the former looked wilted, limp and sometimes yellowish green while the leaves on the latter were mostly fresh, strong and verdant. Analysis of root yields and inulin content on samples taken in September 1994 showed no significant difference from 1993 samples. For sugar beets, the yields expressed as white sugar ha-1 were 15 to 20% lower in September 1994 as compared to September 1993. The chicory used for inulin production (Cichorium intybus) is the same variety as the one used for the production of the coffee substitute. The limited but steady breeding of the 'coffee type' chicory has led to cultivars highly suited to inulin production. The aims were similar: large root yields with high dry matter content. Recent efforts towards the upgrading of the culture focus on high inulin content, high germination capacity, greater compactness of the root for ease of harvest, high and constant degree of polymerization, etc ... Over the last couple of years ORAFTI elaborated a successful and efficient herbicide control program, making the culture of chicory almost as convenient as the culture of sugar beet. Only during the post-emergency period is some manual labor necessary. The interest from growers is high.

78

L. De Leenheer

4. Biochemistry

The enzymology of the inulin metabolism has not yet been clearly resolved. The hypothesis of Edelman and Jeffordl3 is currently quoted most often even though recent enzymological data reveal that their model is both incomplete and flawed.14 Edelman and Jefford suggest that in Jerusalem Artichoke, synthesis proceeds via the concerted action of two fructosyltransferases, with sucrose as the primary fructosyl donor. SST : sucrose : sucrose fructosyltransferase (EC 2.4.1.99). G-F + G-F + G-F-F" kestose) + G. FFT : Fructan : fructan fructosyltransferase (EC 2.4.1.100). G-F-(F)" + G-F-(F)m t)G-F-(F)n+I + G-F-(F)m-I In this reaction, sucrose acts only as an acceptor, not as a donor. This model postulates that the continuous production of the trisaccharide 1-kestose by SST generates fructosyl donors for FFT and thus permits the progressive elongation of acceptor fructan chains. FEH : Fructan exohydrolase (EC.2.4.1.80) is the third important enzyme in this model. G-F-(F)n + G-F-(F),.l + F This enzyme catalyses fructan breakdown through cleavage between the terminal fructosyl group and its adjacent fructose residue, under the assumption that this does not catalyze sucrose hydrolysis. The problem with this hypothesis is that it was formulated without in vitro evidence for the synthesis of inulin beyond a degree of polymerization greater than 3, and with no published evidence for the existence of the key enzyme SST. Moreover, later experiments reporting de nuvu enzymatic syntheses of fructan DP > 3 were based on measurements using only partially purified extracts. Given only e.g. an invertase-activity present in the extract, could in itself explain how the syntheses of oligofructans of DP > 3 are catalyzed. Where functional separation of SST on FFT activities was achieved, the necessary reconstitution experiments are lacking to demonstrate de n o w inulin synthesis from the recombined activities. Recent results obtained by Lusher el al. I5 using dandelion (Taraxacum offercinale Weber) demonstrated that on incubation with a purified FFT fraction, sucrose was neither degraded ncr transformed to trisaccharides.

Production and use of inulin

79

Incubation with single oligofructans of the inulin series of DP 3-7 yielded polymers of higher DP than the substrates used. The chain, however, was only elongated by 1, maximum 2 units and no oligomers with DP > 9 were detected. This same hypothesis of Edelman and Jefford, based on Jerusalem Artichoke metabolism, is also commonly used to explain inulin formation in chicory. The role of SST, however, appears to be different in both types of plants. In Jerusalem Artichoke SST disappears rapidly from the tissue when tubers stop growing, whereas in chicory SST was detectable in the root tissue throughout and beyond growth.'6a Now for several years in succession, as soon as the first frost had occurred, we observed a significant increase of fructose and sucrose in chicory extracts while glucose levels remained unchanged. This would imply that, through the action of fructan hydrolases, a surplus of fructose residues - not necessary for metabolic or osmoregulatory reasons - are released and reconverted to glucose, from which more sucrose can then be produced.16bAs SST and FFT remain active, new inulin molecules can then be formed, albeit of a smaller molecular size than the original ones. The fact that it is mainly the DP that changes as a function of harvesting time, and not so much the concentration of inulin, confirms this theory (Table 5).

m),

Table 5 . Evaluation of average DP ( maximal DP and inulin concentration (% roots) of native chicory inulin as a function of harvesting time in 1993 (*1989).

Date

End August End September Mid October End October Half November End December*

-

DP extract 11.5 11.7 10.6 9.7 8.5 6

-

DP inulin sensu strict0 13.6 14.0

DP max

72

-

11.7 10.1

66 60 51

% inulin concentrate

17.4 17.8 17.3 17.1 16.5 15.0

The presence of fructose as a product of exohydrolase activity is widely accepted. The presence of fructose-oligomers, however, mainly F3, is more controversial. In our observations, the concentration of these fructose oligomers increases slightly from the end of September onwards to be doubled by the end of November. We can therefore hardly consider their presence to be an artefact or a result of erroneous extraction procedures. This is just one more observation that needs to be elucidated biochemically!

80

L. De Leenheer

5. Production

The production of inulin goes through two phases. First we have the extraction phase accompanied with a first purification step leading to a semi-refined syrup; the second step is the refining phase that results in a commercial finished product that is more than 99.5%.pure . The first phase of the ORAFTI chicory process is rather similar to the sugar beet process. Roots are harvested and stored in piles on the field which are preferably not too high. To minimize losses we also insist that no more than 7 days elapse between harvest and delivery date. The roots are then transported to the factory by truck, then weighed and carefully stored on the court. Chicory's metabolic losses during storage are twice as high as the losses observed in sugar beets. During a storage trial under controlled atmosphere at 10 'C, C02 escape is some 250 L day-1 t-l for chicory versus 130 L C02 day-1 t-I for sugar beet, with bruised chicory roots producing 300 to 320 L CO2 day-1 t-1. Losses are highest from the third to the fifth day of storage. In practice we account for losses of 0.05 to 0.1 wt% roots day-1. From the court the roots are then transported on a stream of water to the factory, where they are washed and sliced. Raw inulin is then extracted from the resulting 'chips' with hot water in a diffuser. The leached chips are then dried and sold as feed. A first epuration step is applied to the extraction juice by liming and carbonatation at high pH. The resulting CaC03 precipitates and peptides, anions e.g. phosphates, degraded proteins, and colloids are trapped in the flocks. This generates a foam-type product which is eagerly used by the farmers to improve their soil structure, as it is rich in Calcium and organic matter. However, processing chicory is more difficult than processing sugar beets. One has to consider the chicory process as one big compromise, balancing the degradation of the inulin chain against color formation, Maillard reaction, infection, or even improper removal of peptides and colloids. In the second step, the raw juice is further refined using cationic and anionic ion exchangers for demineralization and active carbon for decolorization. The technology used is comparable to that of starch processing. Major points of concern are again pH and temperature leading to the same conclusion as for the first phase of the process: compromise. After demineralization and decoloration the juice is passed over a 0.2 p filter to be sterilized, then evaporated and spray-dried. For inulin, spray-drying is the most convenient technology for converting the refined product into a storable, microbiological stable and commercial end-product.

Production and use of inulin

81

The use of ion exchange processes for demineralization purposes has often been criticized as environmentally unacceptable. The ORAFTI process proves that the opposite can be true provided the necessary investments are made. The chemicals used for the regeneration of the ion exchangers are not NaOH and HCI as is usual, but NH3 and H2SO4. The price of regeneration is higher with these products but the advantage is that the effluents can be converted into reusable by-products. For this reason all effluents are carefully collected and evaporated. At high concentrations easily crystallized salts such as (NH&S04 and K2SO4 precipitate; they are separated from the mother liquor by centrifugation and sold commercially as fertilizer (containing N and K). The mother liquor is further evaporated into a stable and storable end-product and sold as feed based on its high organic matter content. The condensates generated through this evaporation process are re-used as process waters. This way the circle between processing industry and agriculture is fully closed. The process described above is not only used for the production of inulin, but also for the production of fructose syrups (RAFTISWEETB) and oligofructose syrups (RAFTILOSEB). The only additional process needed is a hydrolysis-step. At ORAFTI we use enzymes for this hydrolysis. The fructose syrup is produced using a mold inulinase from NOVO (Fructozyme), which has both endo- and exo-activity. Working conditions are pH 4.5 and 60 "C. The resulting fructose syrup has a high fructose content, on average 85% or higher, low oligosaccharide content (< 4%) and no aftertaste. By contrast, with an acid hydrolysis process, these criteria cannot be met without on the one hand degrading fructose further into HMF (hydroxymethylfurfural), and on the other hand Table 6 : Composition of a refined MITISWEET@ and MITILOSE@ L85

% carbohydrates

Fructose Glucose DFA + F2 Saccharose F3 DP3 DP4 DP5+ % Ash

Color: Icumsa

RAFTISWEETB

RAFTLOSE@

85

6

13 <1 <1 <1 0 0

1 3 6 24 21 14 25

0.01 10

0.01 17

<1

~ 8 5

82

L. De Leenheer

forming a 'burned' aftertaste. For the production of RAFTILOSEB a mold endo-inulinase is used because the aim is to produce inulin oligomers with the least possible formation of monomers. However, small' quantities of glucose, fructose and sucrose cannot be avoided as they are present in the crude chicory extract from the start. Table 6 gives a typical composition of refined RAFTISWEET@ and RAFTILOSEB,. Today there are three industrial chicory processors: ORAFTI, the name of the brand new business unit of the RAFFINERIE TIRLEMONTOISE responsible for the chicory products (Belgium); WARCOING, Belgium's oldest inulin producer, and BENULINE, a joint venture between WARCOING and SUIKER UNIE (The Netherlands). A fourth producer, CHAMTOR (France), designed a factory for the treatment of both chicory (for fructose) and wheat (for glucose). Today PFEIFER UND LANGEN own a majority interest in this enterprise. The fact that chicory processing has evolved from a laboratory or pilot-plant scale to a full industrial scale is amply proven with the data in Table 7. These figures are based on the quota assigned to each of these producers by the EU and are expressed on a fructose basis (quota in sugar-equivalents divided by 1.9). Industrial reality dictates that existing chicory factories continue to produce mainly fructose: the market for fructose syrups is well-established while the inulin market is still in its building phase. Table 7 can therefore be considered a relatively accurate representation of industrial reality. Table 7.Attributed quota (1994) in fructose equivalents (sugar equivalent divided by 1.9). attributed to the 4 industrial chicory producers.

Quota in fructose equivalent; Ton ds ORAFrI WARCOING CHAMTOR BENULINE

76.803 45.822 12.984 42.914

TOTAL

178.522

Prodi~tioriand iise of iniilin

83

6. Properties 6.1. Nutritional properties Inulin is a dietary fibre, a soluble dietary fibre to be more precise. Dietary fibre is defined as "the remnants of plant cells resistant to hydrolysis by the alimentary enzymes of man. It is composed of cellulose, hemicellulose, oligosaccharides, pectin, gums, waxes and lignin". 17 According to this definition inulin and/or oligofructose are dietary fibers and have typical dietary-fibre properties. They are only hydrolysed to a negligible extent while passing the rather acidic conditions of the stomach; neither are they hydrolysed by the human carbohydrase enzymes that are abundantly present in the small intestine.18 As no monosaccharides are formed, the consumption of inulin will increase neither the glycemy nor the insulin level of the blood. This is why inulin can be used as a diabetic food.lga As inulin enters the large intestine it is readily fermented by the intestinal microflora. The carbohydrates are metabolized by the bacteria into short-chain fatty acids (SCFA; mainly acetic, propionic, butyric) lactic acid and gas. As a consequence of that fermentation, a considerable amount (f40% on inulin DS) of bacterial mass is produced, which in itself increases stool weight - a typical dietary fibre effect.lgb

The SCFA that are produced are absorbed by the large intestine and are further metabolized in the liver. In this way some energy becomes available for the body, but this pathway is far less efficient than when carbohydrates are absorbed in the small intestine and directly metabolized in the liver. Thus the caloric value of inulin is calculated to be 1 .O Kcal/g, in comparison to 4 Kcal/g for sucrose.2" By means of rat experiments, it was shown that feeding inulin or oligofructose reduces the level of triglycerides in the blood. The consumption of inulin did not alter the total cholesterol content in the serum of the rats, but significantly improved (increased) the HDL to LDL ratio, in itself a positive health effect. This way inulin interacts with both the lipid and cholesterol metabolism; an effect which is also considered to be a typical dietary fibre property." Inulin is a Bifidus promoter. As explained previously, inulin and oligofructose are fermented by the colonic microflora. In vitro studies have shown that they are mainly used as a source of energy by Bifidobacteria.** These Bifidobacteria are also able to inhibit the growth of several pathogenic intestinal bacteria, such as C1. perfringens, E. coli, Shigella, Listeria, Vibrio cholerae, etc. This effect is only partly caused by a reduction of the pH; it is mostly caused by the production of a selective antibacterial agent.

84

L. De Leenheer

This agent was obtained through methanol-acetone extraction of lyophilized Bifidus cultures and further purified by passage through a Sephadex G15 The Bifidus promoting effects of oligofructose and inulin were also confirmed in vivo. The faecal Bifidus content of human volunteers who were served 15 g oligofructose per day significantly increased (p < 0.01) as compared to a preceding standard feeding period based on sucrose. It even became the most important bacterial population (17%

* 82%). Clostridia were

significantly reduced (p < 0.05). Other important groups (Fusobacteria, other 'aerobes') were not affected by the feeding. Lactobacilli increased in all experiments. The observations were not statistically significant, however, due to the limited experimental s e t - ~ p . ~ ~ The same Bifidus promoting effect was observed using inulin at 15 g per day. The Bifidobacteria increased significantly (p < 0.01) and the Clostridia decreased qualitatively. The Bifidobacteria also became the most important bacterial population (20% 7 1%). We are beginning to understand the beneficial effects of Bifidobacteria in the following series of properties: they induce a reduction of the colonic pH by the conversion of various sugars into L-(+)-lactic acid and acetic acid, they inhibit the growth of numerous harmful bacteria, they produce substances that stimulate the host's immune system, they synthesize vitamins of the B-group, and they help to restore the normal intestinal flora after antibiotic treatment Consequently, it is possible to conclude that promotion of Bifidobacteria, and thus intake of oligofructose and inulin, contributes to man's health.

6.2. Functional properties of RAFTILINEB, the inulin produced by ORAFTI RAFTILINEB is a white, odorless, spray-dried powder. It has a neutral flavor, slightly sweet, without any off-flavor or aftertaste. It combines easily with other ingredients; it does not change but can even suitably enhance delicate aromas and flavors. It gives a pleasant 'full bodied' feeling in the mouth. Its solubility is moderate. The viscosity of the solution is rather low: 100 m Pas-1 at 30% DM and 10 "C. RAFTILINEB ST takes up water at lower humidity levels than sucrose. It also has a waterbinding capacity of about 1 g water/ 1 g RAFTILINEB. The influence on freezing point depression and boiling point elevation are significantly smaller than that of sugar (Table 8). The sensitivity of inulin to acid hydrolysis has already been mentioned. This effect is more pronounced under high temperature and low dry substance conditions. RAFTILINEB is stable in applications with a pH higher than 4. At lower pH values, hydrolysis is limited to less than lo%,

Production and use of iniilin

85

Table 8. Freezing point depression and Boiling Point elevation effect of RAFTILINE@ ST.

RAFHLINEB (% DM)

Freezing Point "C

5

- 0.1

10

- 0.3

15

-

Boiling Point "C

0.5

30 35

101.2

102.0

even over longer periods of storage (months) if goods are stored at low temperatures (I10 "C)or have a high dry substance content (2 70%).

7. Applications A clear distinction can made between food and non-food applications.

7.1. Food applications Inulin has always been a natural compound of our diet, since it is present in many vegetables and foods: onions, garlic, leek, bananas, and even in flour or wheat and rye. As early as 1912, Lewis evaluated the value of inulin as a foodstuff.25 Van Loo26 calculated that the daily intake of inulin per capita varies between 3 and 10 g in the Benelux countries and between 6 and 17 g in Spain. Inulin is considered to he a food ingredient and not an additive. This was already specifically confirmed by the authorities in all EU countries, Australia, Japan, Sweden and Switzerland. In the USA a committee of experts has determined the GRAS (Generally Recognized As Safe) Status of inulin. In addition to the application of inulin as a soluble fibre, Bifidus promoting compound, or taste enhancer, a unique property is that it can be used as a fat-replacer. To this end, RAFTILINEB is added to an aqueous solution and high energy mixing (shear) is applied. After a few hours a smooth creamy structure is obtained, ideal for replacing fats in spreads, dressings, baked goods, dairy products and even in ice-cream. From RAFTILINEB ST, an improved product, RAFTILINEB ST GEL, was developed to bypass the need for high shear forces. Gentle mixing is sufficient to obtain a stable creme.

86

L. De Leenheer Inulin is an ingredient that fulfills the requirements of the current marketplace. The 'light'

trend is almost over: the consumer expects to receive healthy products, with a nutritional 'plus'. They are tired of consuming less, with less pleasure for more money. New nutritional and marketing concepts will grow and inulin will play an important role in the realization of these concepts ...

7.2. Diagnostic and therapeutic use The non-toxicity of inulin was dramatically demonstrated by Shannon and Smith27 when one of the authors injected himself 160 g inulin intravenously and "no disturbances nor ill effects were observed". This test was necessary before inulin could be used as a substrate to test renal clearance. For GFR (Glomerular filtration rate) studies, inulin is considered to be the standard.28The GFR rate is an indication of the ability of the kidneys to remove a substance from the blood and to deliver it

into the urine. Inulin is an ideal product for assessing this GFR because it is non-metabolizable. non-toxic and freely filtered through the kidneys. Inulin permeation is also used as a standard test applied to artificial kidneys. Here a preparation consisting of mainly high DP molecules ~

(DPn> 25) is generally applied. In the future, inulin might also have applications as a drug carrier. Some drugs are very active when assessed in vitro, but lack efficiency in vivo due to poor chemical or biological stability and lack of organ tropism. Coupling an effective drug to an 'activated' inulin is an elegant solution for overcoming this problem. Schacht et nl.29 and Molteni'o 'activated' inulin through periodate oxidation, epichlorohydrine reaction or reaction with succinic anhydride. These derivatized inulins were then coupled to amino group containing drugs, e.g. procainamide. The inulin conjugates were rapidly and completely excreted by the kidneys without undergoing any metabolism and the pharmacological activity was directly exerted in the urinary tract.

7.3. Other non-food applications of inulin All the applications mentioned hereafter require transformation of native inulin either by fermentation, enzymatic treatment or chemical modification. Classical examples using fermentation are the production of ethanol, acetone-butanol and 2, 3-butanediol; even lactate and succinate can be produced by fermentation starting from inulin.3 I

Production and use of inulin

87

Synthesis of hydroxymethylfurfural (HMF) from fructose, its use and possible applications were also extensively ~tudied.3~ All these examples, however, are still in a laboratory or pilot-scale phase or are not yet industrially applied due to economical feasibility. 'Cyclofructan' is produced by the Mitsubishi Kasei company, using an extracellular enzyme of Bacillus circulans. This enzyme forms mainly cycloinulo-hectaose (CFR-6) in addition to small amounts of cyclinulo-heptaose and -octaose from inulin by an intramolecular transfructosylation reaction.33J5 The cavity in the hexamer is not large enough to accommodate a guest as cyclodextrins can; it is perhaps the octamer that will be able to form a c0mplex.3~Recently, methylated cyclofructans have also been produced. Me-CFR6 has properties33 resembling the cation-captive abilities of crown ethers but is also soluble in organic solvents. Whether these products will really be used in industrial applications is as yet unclear. They can actually be classified as high-price 'specialty' chemicals. Other products from inulin made by intramolecular (depolymerizing) fructose transferases are DFAs (Difructose Dianhydrides), cyclic forms of difructose. Two types of DFAs can be formed when inulin is the substrate: DFAI is di-D-fructofuranose 2',1; 2,l' - dianhydride and DFAIII is di-D-fructofuranose 2', 1; 2,3'-dianhydride. The corresponding fructotransferases were purified from Arthrobacter globiformis and Arthrobacter urefaciens or Pseudomonas.35J6 DFAs have half the sweetness of sucrose, are difficult to digest and therefore well-suited for use in low calorie foods.37 A bitter aftertaste is present. Their industrial application status is as yet unknown to the author. A whole range of alternative possibilities in the area of non-food applications of inulin will become available if chemical modification is considered. A promising development in this field is certainly the work of Besemer and Van Bekkum, who produced Dicarboxyinulin.38,39A one-step oxidation method using NaOCl as the oxidant and bromide as the catalyst is applied. The resulting fully-oxidized ring-opened polycarboxylate has excellent Calcium complexing properties, and might thus be an interesting candidate for replacing tripolyphosphate or polyacylates as builder or co-builder in detergent formulations. Even partially oxidized material shows reasonable calcium binding properties, which may be advantageous in view of the required bi~degradability.~~ This product might become the first industrial inulin application in the non-food sector ...

88

L. De Leenheer

Other methods for converting inulin into polycarboxylates are oxidation and carboxymethylation. The resulting products could have possible applications as metal-ion binding molecules, co-builders in detergents or emulsifiers. Carboxymethylation of the hydroxyl groups of inulin with mono-chloroacetate in an alkaline medium is one of the methods for introducing carboxyl groups into the polymer. The resulting products inhibit the growth of CaC03 crystals. The longer the DP of the original inulin used for modification, the better the inhibitory properties.jO Platinum-catalyzed oxidation of the primary hydroxyl at C6 to a carboxyl group is another way to produce carboxylates .40 Selective oxidation of the primary alcohol group is possible via a TEMPO (2,2,6,6Tetramethyl- 1-piperidiniloxy) mediated reaction in the presence of hypobromite. An aldehyde group is first formed which then reacts further with hypobromite to form a carboxylate group.jl In addition to oxidation and carboxymethylation reactions, esterifications were also tried. Berghofer et al.42 report on a derivatization of this type with a mixed acetichdipic acid anhydride and with sodium trimetaphosphate. Both esterifications were feasible and both derivatives showed increased viscosity and reduced solubility. The authors remark that “the low stability of the inulin molecules” was found to be disadvantageous.

8. Future developments Fuchs4’ noticed a rapid increase i n the number of scientific papers concerning fructan research over the last decade. He also observed a shift from papers concerning ‘crops’ to publications dealing with ‘products’. This trend will certainly continue throughout the next decade. Moreover, we are seeing a third type of publication emerging concerning a new kind of inulin research: the effects the consumption of this food ingredient has on human health. Under the auspices of the European Union, ORAFTI started a European Research project called E N D 0 (European Non Digestible Oligosaccharides), which it coordinates. The University of Louvain (Belgium), the MRC Dunn Clinical Nutrition Centre of Cambridge (United Kingdom) and the TNO - Zeist (Nutrition and Food Research Institute in The Netherlands) are participants. Fundamental biochemical mechanisms governing the altered lipid metabolism will be investigated; also the impact of inulin on the composition of the colonic flora and the influence of the altered bacterial interaction on the host’s health status will be the subject of thorough research; finally the

Production and use of inulin

89

impact of NDO on lipid and cholesterol metabolism, mineral absorption and intestinal function will be checked on human volunteers. In crop improvement the use of genetic engineering will become more and more important, a case of (Science?) fiction becoming reality. The first transgenic sugar beets transformed with levan are now being field-tested by Vanderhave, a seed producing firm! The group around S m e e k e n ~ ~ "has ~ 5 done a lot of interesting work on the transformation of potato and tobacco plants with the levansucrase gene of Bacillus subtilis, producing levan, and the inulosucrase (ftf, of Streptococcus mutans producing an 'inulin type' molecule. The transformations with the levansucrase gene were most successful: tobacco produced up to 8% of the dry leaf weight as levans, and for the potato even 30% was shown to be levan. Moreover, these levans were not degraded (hydrolysed) during senescence. Within the framework of the EC Biotech program, three industrial and three university partners are collaborating in search of 'tailor-made fructans'. The aim is to have fructans of high DP expressed in chicory or sugar beets. To obtain high expression of the product and to have it in the right place, i.e. the vacuole of the roots, specific expression and targeting signals must be inserted together with the gene of interest. There is no indication, however, that the DP is an important factor for the nutritional properties mentioned above. For functionality, in contrast, we observed a marked difference between standard RAFTLINE@ with mean DP of 10 and RAFTILINEB HP, a fractionated inulin with mean DP of 23. Only half of the quantity of the latter material is needed to have the same fat replacement abilities or mouth-feel qualities. That is also why the product was called HP: 'High Performance'. Research in crop and culture improvement will mainly make use of classical breeding techniques, using hybrids for instance. Desired properties are: good germinal force independent of

soil and weather conditions, fast closing of the canopy so that photosynthates are stored in the roots as soon as possible, high inulin to dry matter ratio, high and constant chain length of the inulin, and more thick-set roots so that less earth sticks to it and so that losses via snapped-off root-ends are minimized. More research is to be expected in the area of chemical derivatization. Our actual knowledge of inulin can be compared to the state of knowledge that existed for starch 20 years ago. At that

time, no one had insight into the broad applications modified starch could be used for. However, a major drawback in comparing inulin to starch could be the price and possibly the instability of the fructose chains during processing. Moreover, actual environmental requirements demand

90

L. De Leenheer

'efficient' and 'environmentally-friendly' processes. Researchers must be fully aware, even when performing lab trials, that only industrially feasible and environmentally friendly processes will have a future! Derivatized inulin could be headed for a bright future: think in this respect of applications such as gelling substances, lubricants, sequestering agents and even drug carriers.

Conclusion It has been demonstrated that inulin evolved from a mere scientific topic into industrial reality. Production on a large scale is not only feasible but also offers a resulting product that is - in the true sense of the word - an active food ingredient. Further research must bring additional insights into the health claims already being made. New developments will generate new products. The extent of their possible applications in both food and non-food areas is as yet unpredictable.

Ackowledgement. The author thanks Dr. Johan De Rycker for professionally reviewing this paper, Dr. Jan Van Loo and Paul Robyns for additional information, and Bea Stouthuyzen for the careful typing.

References I. 2. 3. 4.

5.

6. 7. 8.

Thomson T.: A System of Chemistry, Vol 4, from 5th London edition, Abraham Small, Philadelphia, 1818,

65. Waterhouse A.L.; Chatterton N.J.: 'Glossary o f fructan terms, in 'Science and Technology of fructans', Suzuki M.; Chatterton N.J., Eds., CRC Press, Florida, 1993, 2-7. De Leenheer L.;Hoebregs H.: 'Progress in the elucidation of the Composition of Chicory inulin', Srarch 46 (1994) 193-196. Chatterton N.J.; Harrison P.A.; Thornley W.R.; Bennet J.H.: 'Separation and quantification of fructan (inulin) oligomers by anion exchange chromatography', in 'Inulin and Inulin-containing crops', Studies in Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science Publishers.1993, 93-99. Timmermans J.W.; Van Leeuwen M.B.; Tournois H.; de Wit D.; Vliegenthart J.F.G.: Proceedings of the fourth seminar on inulin. Wageningen, 1993, 12-15. Praznik W.; Beck R.H.F.: Nitsch E.:'Determination of fructan oligomers of degree of polymerization 2-30, hy high performance liquid chromatography'. J . Chromarogr. 303 ( 1 984) 4 17-421. French A.D.: 'Recent Advances in the structural chemistry of inulin', in 'Inulin and Inulin-containing crops', Studies in Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science Publishers, 1993, 121-127. French A.D.; Waterhouse A.L.: 'Chemical structure and characteristics', in 'Science and Technology of Fructans', Suzuki M.; Chatterton N.J., Eds., CRC Press, Florida, 1993, 41-81.

Production and use of inulin 9.

10

I I. 12.

13. 14. 15.

I6a.

91

Debruyn A.; Alvarez A.P.; Sandra P.; De Leenheer L.: 'Isolation and identification of P-D-fructofuranosyl(2+ I ) - ~ - D - f r u c t o f u r a n o s y ~ -I()-D-fructose, 2~ a product of the enzymatic hydrolysis of the inulin from Cichorium intyhus', Carbohydr. Res. 235 (1992) 303-308. Meyer W.J.M.; Matthysen E.W.J.M.; Borm G.E.L.: 'Crop characteristics and inulin production of Jerusalem Artichoke and Chicory'. in 'Inulin and inulin-containing crops', Studies in Plant Science, Vol. 3. Fuchs A, Ed.. Elsevier Science Publishers, 1993. 29-44. Van den Berg er al.:, 'Inulinewinning uit landbouwgewassen' Rapport vakgroep Proceskunde, Landhouwhogeschool, Wageningen, 1985.6-8. Zuhr J.; Pedersen H.S.: Characteristics of growth and development of different Jerusalem Artichoke cultivars', in 'Inulin and inulin-containing crops', Studies in Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science Publishers, 1993, 11-19, Edelman J.; Jefford T.G.: 'The mechanism of fructan metabolism in higher plants as exemplified in Helianthus tuherosus L', New Pliyrologisr 67 (1968) 5 17-53I . Cairns A.J.: 'Evidence for the de novo synthesis of fructan by enzymes from higher plants: a reappraisal of the SSF/FFT model', New Phytologisr 123 (1993) 15-24. Nisher M.; Frehner M.; Niisherger J.: 'Purification and some properties of fructan : fructan fructosyl transferase from dandelion (Taraxacum officinale Weher)' New Phyrologisr 123 (1993) 437-442. Limami A,; Fiala V.: 'Fructan polymerization during the growth of chicory (chicorium intybus L.) plants', in 'Inulin and inulin-containing crops', Studies in Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science

Puhlishers. 1993, 191-197. 16b. Pontis H.G.; Del Campillo E.: 'Fructans' Chapter 5 in 'Biochemistry of storage carhohydrates in green plants'. Academic Press, London. 1985, 22 I . 17. Trowell H.; Burkitt D.: 'Physiological role of dietary fiher : a ten-year review', Bol. Asoc. Med. P. Rico, 1986, 54 1-544. Knudsen K.E.B.; Hessov 1.: 'Recovery of inulin from Jerusalem artichoke (Heliantus tuherosus) in the small 18. intestine of man', Brirish Journal ofNirrririon 74 (1995) 101-1 13. 19a. Sanno T.; Ishikawa M.; Nozawa Y.; Hoshi K.. Someya K.: 'Application of Neosugar P for diabetic suhjects. The effect of Neosugar P on blood glucose, Proc. 2nd NeosugarB Research Conference, Tokyo 1984. 19h. Gibson G.R.; Roherfroid M.B.: 'Dietary modulation of the human colonic microbiota-introducing the concept 20. 21 22 23. 24. 25. 26.

of prehiotics', Journal ofNicfririon 125 (1995) 1401-1412. Roherfroid M.; Gibson G.R.; Delzenne N.: , 'Biochemistry of oligofructose, a non-digestihle fructooligosaccharide : an approach to calculate its caloric value', Nurr. Rev. 51 ( 5 ) (1993) 137-146. Delzenne N.M. ef al.: 'Dietary fructooligosaccharides modify lipid metabolism in rats', Am. J. Clin. Nurr.

57 - (Suppl.) (1993) 820 s. Wada K.: 'In vitro fermentability of oligofructose and Inulin by some species of human intestinal flora', Internal report for the Tiense Suikerraffinaderij, 1990. Gibson G.R.; Wang X.: 'Regularity effects of Bifidobacteria on the growth of other colonic bacteria', J. Appl. Bart. 77 (1994) 412-420. Gihson G.R.; Macfarlane S;Cummings J.H.; Wang X.; Macfarlane G.T. 'In vivo effects of oligofructose on the composition of the human colonic microflora' AGA abstracts - Gastroenterology 104 (4) (1993) 2/AS12. Lewis H.B.: 'The value of inulin as foodstuff, J. Am. Med. Assoc. 58 (1912) 1176-1 177. Van Loo J. et af.:'On the presence of inulin and oligofructose as natural ingredients in the Western diet', Accepted for publication in Critical reviews for food science and nutrition.

92 27. 28.

29. 30. 31.

32

33. 34. 35.

36.

37. 38. 39.

40.

41

42.

43. 44. 45.

L. De Leenheer Shannon J.A.; Smith H.W. , 'The excretion of inulin, xylose and urea by normal and phlorizinized man', J . of Clin. Investig. 14 ( 1935) 393-40 I . Preuss H.G.; Podlasek S.J., Henry J.B.: 'Evaluation of renal function and water electrolyte, and acid - base balance'. Book : Clinical Diagnosis and Management, Ed. JB. Henry Publishers WB Saunders, 1991, 119-121. Schacht E. et al.: 'Use ofpolysaccharides as drug carriers. Dextran and inulin derivatives of procainamide', Ann. N . Y. Acad. Sci. 446 (1985) 199. Molteni L., 'Dextran and inulin conjugates as drug carriers', Merhods Enzyrnol. 112 (1985) 285. Fuchs A,: 'Production and utilization of inulin. Part I1 : Utilization of inulin' in Science and Technology of fructans'. Suzuki M.; Chatterton N.J.. Eds., CRC Press, Florida, 1993, 319-352. Kunz M.: 'Hydroxymethylfurfural, a possible basic chemical for industrial intermediates', in 'Inulin and inulin-containing crops', Studies i n Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science Publishers, 1993. 149- 160. Mitsuhishi Kasei Corporation , documentation file. French A.D.: 'Recent Advances in the structural chemistry of inulin', in 'Inulin and Inulin-containing crops'. Studies in Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science Publishers. 1993, 121-127. Ychiyama T.: 'Metabolism i n microorganisms. Part I1 : Biosynthesis and degradation of fructans by microbial enzymes other than levansucrase', in 'Science and Technology of fructans', Suzuki M . ; Chatterton N.J., Eds., CRC Press, Florida, 1993, 169-190. Tamura K.; Kuramoto T.; Kitahata S. (1988), 'Enzymatic manufacture of di-D-fructosylfuranose 1.2'; 2,3' dianhydride', Jap. Pat. 63219389. Norinsho; Nihon Denpun Kogyokk , 'Di-fructose di-anhydride, obtd by heating inulin (plant extract) with inulin fructo-transferase fixed on solid support ofchitosan and silica', (1989) Jap. Pat. 501285195. Besemer A.C.: 'The bromide-catalyzed hypochlorite oxidation of starch and inulin. Calcium complexation of oxidized fructans', Ph. D. thesis, 1993, Delft University of Technology. Besemer A.C.; van Bekkurn H.: 'Dicarboxy-starch vs dicarboxy-inulin; preparation and characteristics', Proceedings of the fourth seminar on Inulin, Wageningen, 1994, 24-36. Van Bekkum H.; Verraest D.L.; de Nooy A.E.J.; Besemer A.C.; Peters J.A. , 'Oxidation and carboxymethylation of sucrose and inulins', paper presented at the 'Workshop on products of sugarbeet and sugarcane', Helsinki, 1994. De Nooy A.E.J.; Besemer A.C.; van Bekkum H.: 'Selective oxidation of the primary hydroxyl fructans of starch and inulin by a TEMPO-mediated reaction', Proceedings of the fourth seminar on Inulin. Wageningen, 1994, 42-48. Berghofer E.; Cramer A,; Schiesser E.: 'Chemical modification of chicory root inulin', in 'Inulin and Inulincontaining crops', Studies in Plant Science, Vol. 3, Fuchs A, Ed., Elsevier Science Publishers B.V.. 1993, 135-142. Fuchs A.: 'Preface', Proceedings of the fourth seminar on Inulin, Wageningen, 1994. i-vii. Smeekens J.C.M.; Ehskamp M.E.; Van der Meer I.M.; Weisbeek P.J.: 'Fructan accumulating transgenic tohacco and potato plants', Proceedings of the fourth seminar on Inulin, Wageningen, 1994, 90-94. Ehskamp M.J.: 'Fructan Accumulation in transgenic plants', Ph. D. thesis, 1994, Universiteit Utrecht.

5 Lactose: its manufacture and physico-chemical properties

E. Timmermans Manager Research & Development, Borculo Whey Products P.O. Box 46,7270 AA Borculo, The Netherlands

Summary. Lactose is a unique carbohydrate only found in mammalian milk. It is a disaccharide made up of one molecule each of D-glucose and D-galactose (0- 1.4-glycosidiclinkage). Lactose exists in two isomeric forms, alpha and beta, differing in configuration of the hydroxyl group. Each form exists in a crystalline state, the alpha usually as monohydrate and beta as anhydrous, in addition to which lactose can exist in a glassy amorphous form which is a mixture of a and p lactose. The principal source of lactose is whey, the remaining liquid in the production of cheese, quark and casein. Lactose, being the major constituent of skimmed milk powder, has been part of many food recipees, for a long time. From the moment lactose was made available in a purified form, the number of applications within the food industry has increased tremendously, due to its unique physico-chemical properties. For some time already, lactose has been used as a starting material for the synthesis of lactose derivatives, the most well-known being lactulose, lactitol, lactobionic acid and galactooligosaccharides either by chemical or biotechnological conversion methods. The range of commercially attractive lactose derivatives will with certainty increase in the near future. In this contribution the production of lactose, some of its physico-chemical properties and major food applications will be highlighted. Attention will be paid to lactose being a raw material for lactose derived carbohydrates.

1. Introduction Lactose is a unique carbohydrate only found in mammalian milk. Lactose, also called milk sugar, is a disaccharide made up of one molecule each of D-glucose and D-galactose linked through the aldehyde group of D-galactose (p- 1,4-glycosidic linkage). Lactose exists in two isomeric forms, alpha and beta, differing in configuration of the hydroxyl moiety (Fig. 1.) Each form exists in a crystalline state, the alpha usually as monohydrate and beta as anhydrous, in addition to which lactose can exist in a glassy amorphous form which is a mixture of alpha and beta lactose.

94

E. Timmermans

H

H

HO

H

OH

a-lactose

OH

p-lactose

Fig. 1. a-lactose and p-lactose.

2. Lactose manufacture The principle source of lactose is whey, the remaining liquid in the production of cheese, quark and casein. The various types of whey can be divided into two groups: sweet cheese whey having a pH above 5.5 and acid casein whey, having a pH below 5.5. Table 1 shows the composition of two types of whey. Table 1. Composition of different types of whey.

Water Total dry matter Fat Protein Lactose Minerals

%

%

cheese whey

casein whey

94.0 6.0

93.7

0.8 4.4 0.8

0.7 4.4 1.2

%

% % %

Of these types of whey, sweet cheese whey is the most common source of lactose. The process route for the production of lactose is as follows (Fig. 2.)

Q thickened whey storage

te storage

J

levaPoratio; 2nd slagel

-

I wet lactose

drying

Q giinding / sieving

I

bela laclose

pharmaceutical sugar slurry I

qgrinding- sieving 7 /

-

1 I

I

""'t'i""] 1

pharmaceutical laclose

grinding / sieving

I Fig. 2 . Production scheme.

pharmaceutical beta lactose

I

96

E. Tirnmermans

Sweet cheese whey, drained off from the cheese vat, is centrifuged to remove remaining fat and curd particles (defatting and clarification). Then the whey is evaporated in two stages, first from 6% dry matter to 28-30% dry matter; subsequently from 28-30% to 60% dry matter. This process takes place in multi-stage falling film evaporators in which a high vacuum is used to concentrate the whey with a minimal denaturation of the whey proteins. Temperatures during evaporation range from 70 "C to 45 "C. The concentrated whey is cooled down to 15 "C, seeded with very fine lactose crystals to promote crystallization, and left to crystallize. a-Lactose monohydrate crystals are formed, which are separated from the mother liquor in centrifuges (decanters). The crude lactose is washed several times and centrifuged again to yield a lactose with a higher purity: a creamy white edible lactose. The wet lactose containing about 4% free moisture is dried with hot air on a fluid bed drier to yield edible a-lactose. This edible a-lactose can be sieved and milled to give products differing in particle size distribution for special applications. The slightly yellow color of edible lactose is caused by the presence of riboflavin (vitamin B2) which adsorbs at the crystal surface.

For the production of refined edible and/or pharmaceutical lactose, the wet lactose is dissolved in hot water, heated for 30-40 minutes at 95-98 "C and filtrated once or twice to remove the remaining impurities such as proteins, minerals and colorants (riboflavin) to yield a product with a very high purity which will meet all the requirements of the pharmaceutical industry. After the filtration step, processing of pharmaceutical lactose is comparable to processing of edible lactose. If the wet lactose is dissolved again at a temperature above 93.5 "C and dried above this temperature by means of a roller drier, steam-heated from the inside, then the beta-form will crystallize. P-Lactose crystallizes without crystal water and is also called anhydrous lactose. Depending on the presence or absence of different filtration steps in the process, edible P-lactose or pharmaceutical p-lactose will be produced. Spray-dried lactose (Zeparox) is made by spray-drying a lactose slurry containing a large amount of a-lactose monohydrate crystals (approximately 85% of the total amount of lactose is present as crystals). Upon spray-drying, the a-lactose monohydrate crystals are coated with a thin layer of amorphous lactose (lactose glass). Spray-dried lactose particles are almost spherical in shape.

Lactose: its manufactureand physico-chemical properties

91

3. Lactose properties In some respects lactose differs significantly from other carbohydrates (disaccharides); the most important aspects of lactose will be dealt with below.

3.1. Crystal forms a-Lactose usually crystallizes as a monohydrate. The crystals are very hard and non-hygroscopic. Anhydrous p-lactose crystallizes above 93.5 "C.These crystals are usually smaller with a larger surface area (Fig. 3). Amorphous lactose is formed when a solution is dried very rapidly. Amorphous lactose is in fact a very concentrated solution and quickly dilutes on addition of water. If the water content of the amorphous lactose is low, 3% for example, crystallization may be postponed almost indefinitely; nucleation rate is negligible because of the extremely high viscosity of the 'solution'. The product is, however, very hygroscopic, and when moisture content rises to about 8%, a-lactose monohydrate starts to crystallize. Crystals of a-lactose monohydrate can have very different shapes (which is a matter of variation in morphology: the crystal lattice is always the same). The most common shape is the 'tomahawk' (Fig. 4).In this figure the Miller indices of the various faces are given. Lactose in solution a=P Rapid drying

C

._ c

e.!

-b

[p] [a]= 1.64 - 0.0027 T Rapid freezing

Amorphous lactose [p]/[a]= 1.25

C

.-c ez l my 8 "

2 "

T = 100 presence of water vapor

4.................................................? Dissolve, T < 93.5

a

hydrate (lactose=LHpO)

T = 100,vacuum

Water uptake T > 93.5

.....................................................

(.

Dissolve,T < 93.5

Modifications of lactose (T = temperature in "C)

Fig. 3. Modifications of lactose (T is temperature in "C).

Fig. 4. A common shape of an a-lactose monohydrate crystal. The crystallographic axes and the Miller indices of the faces are given.

98

E. Tirnmermans Table 2. The rate of growth of some faces (as indicated by their Miller indices) of a-lactose monohydrate crystals as a function of liquid composition.

Super saturation %

55 55 55 55 120 55 55 55 55 55

Growth (v m . h-1 ) of Face

Remarks on composition 010 3.8 1.2 2.7 0.0 43 3.2 6.6 0.2 19.1 0.0

+ 10 ppm gelatin + 100 ppm riboflavin + 10 ppm TMODAC* own pH (=4) PH 7 3 x recrystallized Nonionic** Nonionic + LMP***

110 3.3 1.o 0.0 0.5 34 2.7 5.0 0.7 9.1

110 1.3 1.o 0.0 0.6 21 1.6 2.7 1.3 3.1 0.9

0.0

--

1l o 0.3 0.4 0.0 0.0 12 0.4 1.2 0.5 1.2 0.5

oi i 0

0 7 0

Source: After T. Michaels and A. Van Krefeld. Nerh. Milk Dairy J. 20 (1966) 163; and R. Visser. Ibid. 34 (1980) 255. Note: Unless stated otherwise, solutions are analytical grade a-lactose monohydrate dissolved in water and brought to mutarotation equilibrium. * Trimethyloctadecyl ammonium chloride ** Solution passed through an anion exchanger *** Lactose monophosphate

growth (rnrn/h)

0.7

0.3

0.2 0.1

0

10

20

30

40

p-lactose (g/loo g water)

Fig. 5 . Effect of the concentration of p-lactose on the growth rate of some faces of an a-lactose monohydrate crystal. Supersaturation of a-lactose is 170%. After A. v. Krefeld, Neth. Milk Dairy J . 23 (1969) 259.

Lactose: its manufacture and physico-chemical properties

99

Several substances inhibit the growth. As the inhibition differs widely between different faces, the great variety in crystal forms is easily explained. The most notorious inhibitors are riboflavine, lactose monophosphate and especially p-lactose. It is now clear why lactose crystallizes so slowly even if growth is not retarded by the slowness of mutarotation, because p-lactose is present in a high concentration under most conditions. If the p-lactose content were to be lowered considerably, needle crystals would form because of very rapid growth of the faces 01 1 and 017 Fig. 5 shows the effect of p-lactose on the growth rate of some faces of an a-lactose monohydrate crystal.

3.2. Mutarotation Sugars are optically active. Each sugar has its characteristic specific rotation which is the angle, either positive or negative, by which the plane of polarization is rotated per unit concentration and per unit path length. Specific rotation depends on temperature and wave length. a - And p-lactose differ in specific rotation (Table 3). A freshly made solution will change in rotation, and the conversion of one form to the other in solution is called mutarotation. The equilibrium is temperature-dependent, so a change in temperature causes mutarotation. The speed with which the equilibrium is reached increases with increasing temperature (Fig. 6 and Table 4).

Tahle 3. Differences between a- and (j-lactose.

Characteristics Specific optical rotation ([a]:,O,) Soluhility 20°C (@I00g water)

a

Balance

p

+91. I

+55.5

+33.5

Tahle 4. Soluhility and solution. Temperature

p/a ratio

0

1.63 1.62 1.59 1.58 1.57 1.55 1.52 1.50 1.48 1.45 1.42

10

7.4

19.1

4 8

Relative sweetness (sucrose = 100)

16

32

Melting point "C

202

252

Molecular weight (Dalton)

360

342

p/a ratio of lactose in aqueous

20 25 30 40 50 60 70 80 90

Soluhility in water % g/IOOg HzO

10.6 13.0 16.0 17.9 19.9 24.7 30.3 36.9 43.9 51.1 58.1

11.9 14.9 19.1 21.8 24.8 32.8 43.5 58.4 78.3 104.6 138.6

100

E. Tirnrnerrnans

A

%O

B

K

100 80 60

40 20

0.5

1

1.5

2 t(h)

2

4

6

8

PH

Fig. 6 . Mutarotation in lactose solutions. (A) Course of the reaction (9% equilibrated) as a function of time t; (B) effect of pH on mutarotation rate constant (K in h-I). After Troy and Sharp, J. Dairj Sci. 13 (1930): 140.

At room temperature, it takes many hours before equilibrium is reached; at 70 "C a few minutes. Several substances affect the rate of mutarotation. The salts in whey increase the mutarotation rate, whereas high sugar concentrations will have an adverse effect. In high sugar syrups the dissolution rate of lactose may be influenced by the decreased rate of mutarotation.

3.3. Solubility Lactose differs from other kinds of sugars due to its low solubility (Fig. 7). The solubility differences between the two forms of lactose, p-lactose being much more soluble in water than a-lactose, give lactose some typical characteristics. Lactose solutions can be supersaturated easily, and nucleation does not occur quickly. At a relative supersaturation below about 1.6, seeding with lactose crystals is usually needed to induce crystallization; the solution is metastable (Fig. 8). If a lactose solution is brought to crystallize at a low temperature, crystal growth is retarded because removal of alpha-lactose from the solution is followed by conversion of p to a-lactose. Hence the solution can stay at a low supersaturation for a long time. This mutarotation can play an important role in crystallization processes.

Lnctose: its rnanufnctltre and physico-chemicalproperties

101

g lactosell 009 water 200

100

40

20

5 0

20

40

60

80

100

T ("C) Fig. 7. Solubility of various sugars in water as a function of temperature.

Fig. 8. The soluhilities of a and p-lactose, the final solubility of lactose (line I ), and supersaturation by a factor of 1.6 and 2. I respectively (a-lactose excluding water of crystallization) from various sources.

3.4. Sweetness Lactose has a clean, sweet taste without any aftertaste. While the sweetness character of lactose closely resembles that of sucrose, its intensity is relatively low compared to sucrose (Table 5). The sweetness of lactose is concentration-dependent: the more concentrated the solution, the less lactose is needed to reach equivalent sweetness (Table 6). The sweetness is also temperaturedependent. The higher the temperature, the sweeter a lactose solution. Synergistic effects of lactose with other food ingredients are known, including sweetness enhancement and sweetness suppression. Solutions of lactose and fructose and of lactose and glucose, respectively, are slightly less sweet than might be expected from the contribution of the individual carbohydrates. Tryptophan enhances the sweetness of lactose and lactose prolongs the duration of sweetness sensation of xylitol. An effect of this type could be advantageous in the manufacture of chewing gum. Another desirable synergistic effect is the masking of the bitter aftertaste of saccharin by lactose. If we compare the two lactose-isomers, then p-lactose is considered to be much sweeter than a-lactose. The main reason for this is the higher dissolution rate of p-lactose which causes a higher sweetness sensation. The higher sweetness level of p-lactose has some advantages in

102

E. Timmermans

Table 5. Relative sweetness of some carbohydrates.

Carbohydrates Sucrose Fructose Xylitol Maltitol Glucose Mannitol Sorbitol Maltose Isomaltitol Lactitol Lactose

relative sweetness (wt base) 1 .O (standard)

1.3 1 .o

0.8 0.7 0.7 0.5 0.5 0.5 0.4 0.3

Table 6. Equivalent sweetness concentration (in % lactose and other sugars).

Lactose

Sucrose

Dextrose

Fructose

I .90

0.5

0.89

0.42

3.46

1 .o

I .84

0.76

6.54

2.0

3.57

1.66

15.74

5.0

8.25

4.19 8.62

25.92

10.0

13.86

34.60

15.0

20.00

products which contain no or hardly any water, such as chocolate, or which have a very high viscosity. In aqueous solutions mutarotation will take place and after some time the same equilibrium will be reached as when dissolving a-lactose under the same conditions.

3.5. Moisture sorption Moisture sorption affects the stability and quality attributes of all dehydrated foods. The control of moisture sorption and water activity is very important and influences many chemical reactions. Low molecular weight sugars like lactose may be present in one of several states: -

crystalline solids (a-monohydrate, P-anhydrous)

-

amorphous solids (spray-dried lactose)

-

aqueous solutions The tightly packed lactose molecules in the pure crystalline state adsorb very little water and

only on the crystal surface when the water activity is high (aw > 0.8) (Fig. 9). When amorphous lactose is present, exposure to increasing relative humidities results in water up-take, reaching sorption levels far higher than that of the crystalline form. This difference is due to the larger internal area available for water sorption in amorphous, less-regular material. Some hydrogen bonds are disrupted, and at a given water activity the sugar molecules are somewhat diluted and acquire sufficient mobility to undergo transformation from the amorphous state to the crystalline state, the a-lactose monohydrate. In this process, lactose loses water and the crystals become tightly packed. Visually, the amorphous solid appears to undergo actual viscous flow immediately prior to crystallization (Fig. 10).Crystallization time is dependent on a, and temperature. Thus, if moisture gain is measured as a function of time, a discontinuity in the isotherm occurs at constant aw. The

6%

-

11<101~

7

mhrdrout

.w monoh,dra#. 11<101.

5%

4%-

1%-

i 1%-

1%-

Fig. 9. Moisture sorption isotherm of milk sugar.

Fig. 10. Water sorption isotherm for lactose initially in the amorphous form at 25’C.

break in the curve occurs between a, 0.43 and 0.52. Beyond this point, the lactose becomes nonhygroscopic and very little water sorption occurs. The water that is released on crystallization is available to interact with other components and effects chemical and physical deterioration of other constituents that may be associated with the lactose e.g. in whey powder and milk powders. 4. Utilization Lactose, being the major constituent of skimmed milk powder, has been part of many food recipes for a long time. From the moment lactose was made available in a purified form, the number of applications within the food industry has increased tremendously, because of its unique physicochemical properties as compared to other commercially available mono- and disaccharides. The main fields of application of lactose are shown in Table 7. Tahle 7. Use of lactose.

Use of lactose 1.

2. 3. 4.

5.

Pharmaceutical industry Infant nutrition Food Application Fermentation Synthesis of derivatives

The largest market for lactose is the food industry, which uses about 40% of the total world production. The second-largest is the infant food industry. In the Tables 8 up to and including 12 the main uses and functional properties of lactose in the various industries are shown. As is apparent from these tables, i n all of its forms lactose is used for a wide variety of applications.

104

E. Timniermaris

Table 8. Pharmaceutical uses.

Tahle I I . Food Applications.

*

*

*

* *

tablets wet granulation direct compression capsules inhalation products injectables: pyrogen free

* * *

* *

Table 9. Pharmaceutical Industry. Lactose shows a number of very favorable characteristics for use in pharmaceutical products: * I t is a natural product. * It is a stable product from a chemical, physical and microbiological point of view. * It has high compatibility with other excipients/active ingredients. * It is available in different forms.

Table 12. Functionalities of lactose in food applications.

* * * *

* * * *

Table 10. Lactose in infant food.

* * * * * *

Sugar confectionery Chocolate confectionery Bakery products Meat products Dairy products Powdered products

Sweetness reduction Flavor canierknhancer Texture modification Moisture regulation Browning agent: maillard reaction Crystallization regulation Fat substitution Shelf-life prolongation

* ... . . . . . . ..

Source of energy Source of glucose Source of galactose Osmotic value Calcium Intestinal flora

In the next section, special attention is devoted to the synthesis of lactose derivatives.

5. Lactose derivatives Lactose is used more and more often as a raw material for the synthesis of novel carbohydrates. The following figure shows derivatives from lactose that are produced on a commercial scale (Fig. 11).

5.1. Lactose hydrolysis Lactose consists of one molecule each of D-glucose and D-galactose linked through the aldehyde group of D-galactose (p- 1,4-glycosidic linkage). The chemical bond between the glucose and the galactose molecule is comparatively stable, i.e. more stable than the bond between glucose and fructose in sucrose. This chemical bond can be split by the use of strong mineral acids, but in dairy

Lactose: its manufacture and physico-chemicalproperties CH20H

H

105

CH20H

OH

H

OH

Lactobionic acid

Oligosaccharides

Lactitol

Fig. 1 I . Lactose derivatives.

products such as milk and whey, and even when hydrolyzing pure lactose, enzymatic hydrolysis is preferred (Fig. 12). Hydrolysis of lactose is performed in milk and whey, in part to overcome the problem of lactoseintolerance. Some very small quantities galactose are produced after hydrolysis by crystallization. The application of galactose is unknown. CH9OH CH20H

H,O

HO H

H H +

H C OH&

H

H

OH

D-galactose Fig. I ? . Hydrolysis of lactose to galactose and glucose

F

? H

+ Ho H

HO

Bglucose

106

E. Tirnmermans Compared to lactose, hydrolyzed lactose can be characterized by the following properties

(Table 13). Table 13. Hydrolyzed lactose: Properties.

increased solubility increased sweetness easier digestibility more intensive browning action increased flavor enhancement increased osmotic pressure easier fermentability Hydrolyzed lactose is applied in various product areas (see Table 14). The ice cream industry, when using lactose-rich products, may occasionally find that as a result of the low solubility of lactose, crystallization takes place with the consequent rise of sandiness. This can be reduced by either seeding with microfine lactose just before freezing, so that a large amount of very fine crystals will be formed, or by replacing part of the lactose by hydrolyzed lactose; at the same time, the increased sweetness would allow reduction of the amount of added sucrose. This sweetness of hydrolyzed lactose is also advantageous in other dairy products such as flavored, pasteurized and fermented milks, when in addition to increased sweetness there is also benefit to be gained from the increase in flavor-enhancing properties of hydrolyzed lactose. Table 14. Application of hydrolyzed lactose products.

Application of hydrolyzed lactose products Bakery products Sugar confectionery Ice cream Dairy products Soft drinks

Lactose: its manufacture and physico-chemicalproperties

107

5.2. Oligosaccharides

Oligosaccharides can be produced by the same type of enzymes that are used to hydrolyze lactose into glucose and galactose. Oligosaccharides are carbohydrates consisting of two to eight monosaccharides that are not actively digested by the enzymes of the mucosa.

Production These oligosaccharides are produced from lactose by P-galactosidases by their transgalactosylic action. In contrast with the hydrolysis action, where water is used to hydrolyze, lactose itself is used, leading to the formation of oligosaccharides (Fig. 13).

Lactose (Gal (1-4) Glu) Gal

1 hydrolysis + Glu 1

internal rearrangement

Gal (1-2) Glu Gal (1-3) Glu Gal (1-6) Glu (allolactose)

1 Gal (1-6)Gal (1-4) Glu (6' galactosyl lactose) Gal (1.3) Gal (1-4) Glu (3' galactosyl lactose)

1 Gal (1-6) Gal (1-6) Glu

1 Gal (1 -6) Gal (1-6) Gal

1 tetrasaccharides

1 pentasaccharides

1 hexasaccharides Fig. 13. Possible reaction products from the reaction of lactose with P-galactosidase.

Fig. 14. Hydrolysis of oligosaccharides by human intestinal mucosa.

E. Timmermans

108

Physiological properties

Table 15 presents some properties of galacto-oligosaccharides. Galacto-oligosaccharides (from lactose) are not actively -.gested by any digestive enzyme when orally applied and reach the human large intestine. Galacto-oligosaccharides are therefore considered to be low-calorie carbohydrates. When they reach the colon, they can be digested by the bacterial flora. All bifidobacteria are capable of utilizing galacto-oligosaccharides; most E.coli or Salmonella strains are not, thus promoting the growth of the beneficial bifidobacteria. Consumption of galacto-oligosaccharides will increase the number of bifidobacteria in the large intestine. Moreover, galacto-oligosaccharides are not utilized by oral Streptococci that produce plaque leading to dental caries. Thus galacto-oligosaccharidescan be considered non-cariogenic (Fig. 15). Table 15. Properties of galacto-oligosaccharides.

Conditions for caries development

: 7

Properties of galacto-oligosaccharides

microorganisms

-

non-cariogenic

- low calory carbohydrates

I

- growth promotor of bifidobacteria - lower pH of faeces -

suppress growth of pathogenic bacteria

-

stimulate immune system

- synthesize vitamins

galactooligosaccharides

Fig. 15. Schematic representation of caries etiology.

5.3. Lactitol Lactose can be converted into the corresponding sugar alcohol, lactitol, by catalytic hydrogenation. In this reaction the glucose part of the molecule is reduced through hydrogenation. The molecular structures of both lactose and lactitol are shown in Fig. 16. Lactitol is chemically described as 4-O-~-D-galactopyranosyl-D-glucitol. Like a-lactose monohydrate, the solid lactitol also contains water of crystallization. It appears either as a monohydrate or a dihydrate.

Lnctose: its rnanirfacfure arid physico-chemical properties

109

Production The reduction of lactose to lactitol is performed in a 30-40% lactose solution in water at about 100 "Cand 40 bar hydrogen pressure in the presence of a Raney Nickel catalyst. After the reaction the solution is filtrated, decolorized by activated carbon, demineralized by ion exchange, concentrated, and cooled to promote crystallization of lactitol, which is separated in centrifuges to yield a crystalline lactitol (Fig. 17).

Lactose solution 30-40 % + Raney nickel

Lactose tKJ

I

CHz OH

+ Hz hydrogenation I00 "C. 40 bar

OH

sedimentation

ao&"'"I""" Lac t i to1

tKJ tKJ

CHz OH

OH

OH

residue spent catalyst

I

+ active cabon decoloiization

ion exchange

I

,

residue spent carbon

evaporation

mother liquor

Fig. 16. Molecular structure lactose and lactitol.

crystals of lactito1

Fig. 17. Manufacturing process of lactitol

By varying concentration and temperature of crystallization, either the monohydrate or the dihydrate of lactitol is formed.

Physiological properties The metabolism of disaccharides like sucrose and lactose normally starts in the small intestine with the enzymatic hydrolysis of the glycosidic linkage, after which the monosaccharides are quite easily absorbed. Lactitol cannot be absorbed as a whole molecule, so a crucial question is whether the small intestine contains enzymes like P-galactosidase that are able to hydrolyze lactitol.

110

E. Timmermans

Modification of the glucose moiety of the lactose molecule always leads to slower hydrolysis, and it has been shown by various studies, both in vitro and in vivo, that hydrolysis and therefore absorption of lactitol in the small intestine is negligible. This makes lactitol a bulk sweetener that can be used in food for diabetics. Tests showed that in diabetics and in 'normal people' the blood-glucose content was not influenced by the consumption of 24g/day of lactitol. When 50 g of sucrose was combined with 50 g of lactitol, the blood-glucose content was even clearly lower than after consumption of 50 g of sucrose alone. Studies at the Agricultural University of Wageningen showed that after consumption of 50 g lactitol per day during 18 days, faeces of people did not contain significant amounts of lactitol. This means that lactitol is almost completely metabolized in the large intestine by the microflora present, leading to fatty acids, hydrogen and methane gas. Studies with volunteers at Wageningen showed that the metabolic energy value of anhydrous lactitol is maximally 50% of that of sucrose. Table 16. Properties of lactitol

- non-cariogenic - low-calorie carbohydrate - constipation regulator

5.4. Lactulose

Lactulose is prepared by base-catalyzed isomerization of lactose by the Lobry de Bruyn-Alberda van Ekenstein transformation (Fig. 18). Generally, sodium hydroxide is applied as the base and only relatively low conversions are allowed to prevent degradation. Lactulose is obtained as a syrup. A recent process applies borate to shift the equilibrium and to protect the lactulose. Lactulose forms, in its pyranose and especially in its furanose form, strong borate esters, whereas lactose only weakly interacts with borate. This process became recently operative and leads to crystalline lactulose. Obviously, complete recycling of borate is required. Anhydrous crystalline lactulose is a mixture of two furanose and a pyranose forms (85: 15). The trihydrate is in the p-furanose form. (G.A. Jeffrey et al. Curbohydr. Res. 226 ( 1992) 20-42) The main application of lactulose is in medicine, based on the fact that it is not, or is only at a low rate, hydrolyzed by P-galactosidase in the small intestine. It passes down into the colon where it is metabolized by the microflora. Its metabolic products are low-molecular weight organic acids,

Lactose: its manufacture and physico-chemicalproperties

111

hydrogen gas and methane gas. The osmotic properties of the acids attract water, which makes the stool softer.

H

CH20H

HO-

Isomerization

H

Hou $E>o H

CH20H

H OH

H OH Lactose 4-0- PD-galactopyranosyl- D-glucose

OH Lactulose H 4-0- ED-galactopyranosyl- D-fructose

Fig. 1 R. Schematic process of laciulose formation

As the bacteria now have a rich environment, they grow well and increase the amount of

biomass in the faeces. Taken in moderate quantities, lactulose prevents constipation and as such it has found a wide application (Table 17). The stimulation of the micro-organisms of the gut and their growth in a medium rich in carbohydrates has a lowering effect on the concentration of ammonia and other nitrogenous compounds in the blood. This is important when the liver cannot remove ammonia due to cirrhosis. Serious cirrhosis may be caused by alcohol intoxication or by jaundice. This gives rise to neuro-psychiatric symptoms called partalsystemic encephalopathy. These symptoms are effectively controlled by the administration of lactulose. Lactulose does not raise blood sugar levels of diabetic patients and healthy human adults. Table 17. Applications of lactulose.

- constipation regulator - growth promotor of bifidobacteria - treatment of PSE - low calory carbohydrate - diabetic foods

I12

E. Tirnmermcins

5.5. Lactobionic acid Introduction Lactobionic acid (4-0-p-D-galactopyranosyl-D-gh~conicacid) is obtained by oxidation of lactose. The free aldehyde group of the glucose part of the lactose is transferred into a carboxyl group. From the literature it appears that lactobionic acid can be synthesized according to four different production methods: microbiologically with Pseudomonas bacteria, electrochemically with bromine water, and chemically with oxygen and Pd-Bi-C as catalyst, or at high pH with Pt-

or Rh-catalyst upon which hydrogen evolves. Dependent on the conditions the acid, the salt or the lactone is isolated. The structures are as depicted in Fig. 19.

CH9OH

CH20H

0

CH20H

H

OH

Lactobionic acid

Fig. 19. Schematic process of lactohionic acid formation.

Applications for lactobionic acid are shown in Table 18. Table 18. Applications of lactobionic acid.

- carrier for antibiotics - preservation of organs for transplantation

* *

- mineral suppletion - growth promotor for bifidobacteria - food acidulant

Finally we mention the possibilities for degrading lactose by oxidation towards 3 - 0 - p - D galactopyranosyl-D-arabinose (using hydrogen peroxidelborate) or towards 3 - 0 - p - D galactopyranosyl-D-arabinonicacid (applying oxygen at pH > 13.5).

Lactose: its nianicfacture and physico-chemical properties

1 13

References Berg, R. Van den, The structure and (local) stability constants of borate; Peters, J.A., Esters of mono-and disaccharides as studied hy 1 IB and Bekkum, H. Van , 13C NMR spectroscopy, Carbohydr. Res., 253 (1994) 1-12 Hendriks, H.E.J., Selective catalytic oxidations of lactose and related carbohydrates, Thesis, University of Eindhoven, 1991, Kreveld A. Van, Growth rates of lactose crystals in solutions of stable anhydrous a-lactose, Nerh. Milk D a i p J. 23 (1969) 258. Michaelis AS., Influences of additives on growth rates in lactose crystals, A. Van Kreveld, Nerh. Milk D a i n J. 20 ( 1 966) 263. Modler H.W., Proceedings of the IDF-workshop on lactose hydrolysis, IDF Bulletin 289, Munich 1993. Morrissey P.A.. Lactose: chemical and physico-chemical properties, in Developments in Dairy Chemistry 3, ed. P.F. Fox . Nickerson T.A., Lactose sources and recovery, Developments in Food Carhohydrates I(77-90) ed. G.G. Birch, R.S. S ha1len herger. Nijpels H.H.. Lactases, i n Developments in Food Carhohydrates 3 (23-48), ed. C.K. Lee. M.G. Lindley Paige D.M., Lactose digestion, clinical and nutritional implications, Th.M. Bayless, London, 198 I . Rajah K.K., The ALM Guide to Lactose properties & uses, London, D.E. Blenford 1988. Thelwall L.A.W.. Developments in the chemistry and chemical modification of lactose, in Developments in Dairy Chemistry 3, ed. P.F. Fox. Timrnermans H.J.A.R., Lactose: A different sugar in confectionery, in Lactose as a food ingredient, ed. E.H. Reimerdes 1990. Visser R.A., A natural crystal growth retarder in lactose. Nerh. Milk D a i p J. 34 (1980) 255. Visser R.A., Crystal growth kinetics of a-lactose hydrate, Thesis, University of Nijmegen. 1983. Visser R.A., Lactose and its chemical derivatives, Trends in Utilization of Whey and Whey Derivatives (33-44). IDF-Bulletin 233, 1988. Walstra P. and R. Jenness, Dairy Chemistry and Physics, New York. 1984. Williams C.A.. Lactose Hydrolysis syrups: physiological and metaholical effects, Developments i n sweeteners 2, (27-50), ed. F.H. Grenhy. K.J. Parker, M.G. Lindley Wit, de G., Catalytic dehydrogenation of reducing sugars i n alkaline Solution, Carbohydr. Rex. 91 (1981 ) 125.138.

This Page Intentionally Left Blank

6 Raw materials for fermentation

D. Wilke

Dr. Wilke & Partner Biotech Consulting GmbH D 30967 Wennigsen, Germany

Summary. Carbohydrate substrate choice, quality requirements and acceptable raw material prices strongly differ for specialty products, such as pharmaceutical ingredients, and for bulk fermentation products such as ethanol, citric- and amino acids. For specialty products, the choice of carbohydrate for fermentation processes depends on the genetic constitution of the production strain - its oligo- and polysaccharide growth spectrum. All further decisions directly pertain to process economics and technical feasibility, e.g. carbohydrate price. specific fermentor yield and yield on carbohydrate substrate, or product recovery and purification aspects. For cost reasons the use of non-refined carbohydrate substrates often becomes mandatory i n bulk product fermentations. However, the cost advantages of non-refined crops must be carefully evaluated and compared to any negative impacts on the fermentation and recovery process as well as the by-product and waste-mass streams. For large volume products such as bio-ethanol and citric acid, the carbohydrate sourcing conditions and logistics have already induced a forward and backward integration of the fermentation and the agro-industry. This trend is certain to be continued for new bulk fermentation products such as lactic acid and other potential organic intermediates.

Organic chemicals through biotechnology From a theoretical point of view, biochemistry and industrial biotechnology are the preferred tools for the conversion of highly oxidized natural feedstocks such as carbohydrates into defined chemical products. The limited success of organic chemistry in converting renewable resources, except for fats and oils, clearly attests to this statement. However, while the number of biotechnological conversion products in the field of low volume high priced specialties is rather impressive, biotechnology's contribution to the bulk market is still limited. About one hundred pharmaceutical and nutritional specialties are manufactured by fermentation, microbial transformation and enzymatic synthesis, yet only a few large-scale products are derived from biotechnology. These are the fermentation products bioethanol, citric acid, monosodium glutamate and lysine, as well as glucose from enzymatic starch hydrolysis (Table 1). It is, however, precisely the bulk market that is of significance with respect

116

D. Wilke

to crude-oil substitution and environmental impacts, or as an alternative outlet for agricultural

surpluses. Table 1. Biotechnological production of organic chemicals.

Specialities & fine chemicals pharmaceutical ingredients pharmaceutical intermediates food & feed additives Bulk chemicals bio-ethanol

12 m ton p.a.

citric acid monosodium L-glutamate L-lysine

650 t ton p.a. 400 t ton p.a. 150 t ton p.a.

L-lactic acid

60 t ton p.a.

gluconic acid

40 t ton p.a.

glucose

12 m ton p.a.

The reason for biotechnology's limited contribution to the conversion of carbohydrates into industrial products is the lack not of raw materials or market demand, but of products and product ideas and of the relevant manufacturing technologies. Technological shortcomings thereby pertain to a serious lack of processing feasibility and of concomitant cost competitiveness in comparison to established petrochemical products. Nevertheless, biotechnology is a well-accepted industrial tool for those who are familiar with it, and a number of critical success factors can be derived from existing processes in order to

evaluate how carbohydrates and other agricultural raw materials can effectively be converted into existing or new chemical compounds.

Microbial metabolites by fermentation Among the three basic biotechnological processing concepts, fermentation stands for a microbiological production process with living, growing and actively metabolizing cells. Microbial

Raw materials for fermentation

117

transformation refers to a product synthesis or substrate conversion with resting cells, and enzymatic synthesis implies a substrate conversion using the isolated biochemical catalyst (Table 2). Table 2. Manufacturing tools for industrial biotechnology.

Fermentation

growing, actively memulti-step synthesis tabolizing microbial cells

Microbial transformation free or carrier fixed resting cells

1-3 step synthesis / conversion

Enzymatic synthesis

1-2 step synthesis / conversion

isolated free or carrier fixed enzyme catalysis

Except for a few examples, including the conversion of glucose into gluconic acid, sorbitol into sorbose and sucrose into dextrane or laevane, which indeed resemble microbial transformation more than fermentation, the structure of the carbohydrate substrate of fermentation processes is no longer visible in the final product. In a typical fermentation process the c6 moiety is converted via ten to twenty enzymatic steps into primary metabolites, e.g. ethanol, acetic acid, citric acid and amino acids or into secondary metabolites such as antibiotics, alkaloids or other heterocyclic compounds (Fig. 1). Similarly, the conversion of the c6 moiety into microbial polysaccharides oligo & polysaccharides

I C6 sugars

I fructose-6-phosphate t xanthan, polysaccharides

I

glycerol 4

lactic acid 4

aminoacids

ethanol, acetic acid, propanol, butyric acid, acetone, PHB,

bbutanol,

4

succinicacid 4 Fig. I . Metabolic pathway to fermentation products.

L-glutamic acid, amino acids,

118

D. Wilke

such as xanthan or into non-polysaccharide type polymers such as poly-p-hydroxybutyrate involves a considerable number of enzymatic steps.

Yield on substrate and volume productivity In a typical fermentation process, the carbohydrate serves as the main carbon and energy source of the organism, while nitrogen is derived from ammonia, nitrate, amino acids or proteins. A certain amount of carbohydrate substrate is therefore inevitably converted into energy and CO;?, which necessarily results in a molar product yield per substrate below 1.0. In anaerobic fermentations, the carbohydrate substrate is predominantly used as energy source. In the absence of oxygen the carbohydrate is incompletely oxidized and fermentation end products such as ethanol and lactic acid are massively secreted into the fermentation medium as a kind of metabolic waste. However, in aerobic microorganisms the carbohydrate consumption is strictly regulated in order to obtain maximum resource management. The carbohydrate demand for growth and energy metabolism strongly competes with product synthesis and accumulation. As a condition for metabolite and end product oversupply, the genetic and biochemical control of the particular synthetic pathway must be overcome. The product yield on carbohydrate feedstock is further reduced by the number of enzymatic steps that are necessary for its synthesis, and by the number of alternative metabolic outlets, which compete for carbon precursors provided by the central metabolism (Fig. 2 ) .

substrate

co2 product Fig 2. Over-production of microbial metabolites.

Raw materials,forferrneritatiori

1 19

The most important task in the development of industrial fermentation processes is therefore increasing the product yield per substrate and the over-all or volume productivity of the process. This must be achieved in a coordinated approach by improving the microbial production strain and by over-all process engineering. Carbohydrate uptake as well as the regulatory pattern and the absolute activity/ concentration of all necessary enzymes of the respective synthetic route can be optimized by strain selection, conventional mutagenesis and genetic engineering (Fig. 3). Moreover, the synthesis of undesired by-products can be diminished through strain selection and mutation. Similarly, the over-all fermentation conditions - substrate choice, substrate concentration and substrate feeding, pH and temperature profile, agitation and aeration - strongly influence the yield of product on carbohydrate substrate and the ratio of product to by-products, and also the over-all productivity of the system. Contrary to organic chemistry, biotechnology generally depends on water as the reaction medium. Volume productivity, commonly referred to as fermentor yield, is therefore the major parameter for process feasibility in order to cope with product separation from biomass and water soluble by-products as well as product recovery from aqueous solution. At least in theory any microbial synthesis can now be shifted to industrial yields by strain improvement through genetic engineering. Evidently this is most straightforward for pharmaceutical proteins and industrial enzymes, and also very efficient for amino acid fermentations and for many secondary metabolites. Genetic engineering, however, cannot overcome physicochemical limitations of fermentation processes, for example product inhibition or cell membrane damage caused by high concentrations of alcohols or acetic acid as fermentation Strain improvement

improved synthetic activity/ enzyme activities

- strain selection from nature - conventional mutagenesis - genetic engineering

improved yield on substrate improved ratio o f products vs. by-products

increased substrate dosage shortened fermentation time/ batcb cycle improved fermentor yield/ volume productivity

Process engineering

- fermentation conditions - substrate/ carbohydrate choice & dosage I

- downstream processing

I

I

Fig 3 . Fermentation products - strain and process optimization

+

120

D. Wilke

products. Process engineering and the development of new or improved recovery technologies is therefore the other condition for an extended application of fermentation and biotechnology.

Yield-dependent manufacturing costs The yield per carbohydrate substrate is the key factor for low variable costs of an optimized fermentation process, while the fermentor yield and the fermentation timehatch cycle determine the fixed costs of the process - labor costs and depreciation on plant facilities per kg or ton of product. Any improvement of the recovery yield affects both the variable and the fixed costs. The influence of the fermentor yield on the total manufacturing costs of fermentation products is shown in Fig. 4.In an early research stage using wild-type strains grown under laboratory conditions, the product yield per 1 culture broth is often in the mg or even pg range. Strain improvement and process engineering can improve the fermentor yield step by step into the hundred mg-to-one g range that is typical for many pharmaceutical specialties, and into the 100 g-and-above range represented by bulk fermentation products such as L-lysine, L-glutamic

0

50

100

150

200

fermentor yield ( g/I ) Fig. 4. Yield impact on manufacturing costs.

Strain improvement and process development programs have an immediate impact not only on the total manufacturing costs but also on the cost structure. At low volume productivities, the fixed costs predominate, and although the absolute expenses for the carbohydrate substrate per kg of product are high its relative share of the total manufacturing costs is small (Fig. 5). As a

Raw materials for fermentation

Speciality

Bulk product

low yield Fermentation: 1 g/l - 5 days

high yield fermentation: 120 gh- 3 days

other var. & util. 10%

121

-w

Labour, maintenance & depreciation 80%

other variables & utilities 15%

total manufacturing costs: $800/kg

total manufacturing costs: $1.20/kg

Fig. 5. Relative impact of raw material costs.

consequence, for low yield fermentations the fermentor yield is far more decisive than the yield on carbohydrate substrate, and process and yield standardization prevails over relative and absolute carbohydrate cost savings. More precisely, the best-performing carbohydrate will be applied regardless of its price ranking compared to other available substrates, and similarly an expensive pharma-grade sugar that provides reproducible processing characteristics will be preferred over a cheap food- or feed-grade carbohydrate of varying specification range.

Raw material cost impacts on bulk products However, the carbohydrate substrate predominates the cost structure of bulk fermentation products such as citric acid and ethanol. For bulk products under present competitive circumstances, the manufacturer must therefore search for a cost optimum that integrates yield and processing conditions on the one hand and actual carbohydrate prices, qualities, sourcing reliability and waste disposal logistics on the other hand. From an industrial point of view, the weight per weight (w/w) product yield on substrate under large-scale production conditions is the most important economic parameter for bulk product fermentations. In this respect, citric and lactic acid are positive examples because they can be manufactured with high w/w yields of 85% and 98%, respectively. Other fermentation products

122

D. Wilke

such as ethanol, butanol or acetic acid suffer from the loss of two moles of C02 per mole of c6carbohydrate. Their w/w yields on an industrial scale are therefore below 50%, which further enhances the cost impact of the carbohydrate substrate. Bio-ethanol can effectively compete against its synthetic counterpart, although a certain amount of administrative support must be admitted. Acetic acid from low-yield fermentation processes and at a w/w yield of 50% can by no means achieve the present low cost figures of an ethylene or methanol chemistry. The cost competitiveness of fermentative acetic acid, however, would immediately improve if an efficient anaerobic homo-acetic acid fermentation could be developed, i.e. the formation of three moles of acetic acid per mole of c6- carbohydrate avoiding COz as a major by-product.

Substrate choice The chemical nature of the carbohydrate substrate, whether and which mono-, di- or polysaccharide can be used for a particular fermentation process, depends on the genetic constitution of the microbe. Most organisms grow on glucose; many strains can also hydrolyze sucrose. Lactose utilization depends on the induction of a P-galactosidase and a transport system for lactose through the cell membrane. Similarly, many important fermentation organisms cannot grow on starch, which is the least-expensive defined carbohydrate substrate available, because it lacks amylases for starch liquefaction and saccharification. Enzymatically liquefied starch is a preferred substrate for those strains that have sufficient own P-amylase or glucoamylase activity for converting dextrins into maltose and glucose, respectively.

As mentioned above, there are only few examples, such as gluconic acid production from glucose, where the straightforward formation of the metabolite depends on a particular carbohydrate substrate. In all other cases the genetic characteristics of the production strain determine the substrate choice rather than the biochemical pathway leading to the relevant endproduct. Simplified, any carbohydrate is finally directed into the central metabolism via fructose as the central entry door. I t is an empirical observation that the sugar or carbohydrate with the best growth characteristics is not necessarily the substrate for the best product formation. Growth-limiting carbohydrates often provide higher substrate yields and improved volume productivity. Similarly, carbohydrates that are metabolized via different catabolic shunts can strongly influence the over-all product yield. This may further narrow the range of substrates available for the particular production process and may often give preference to comparatively expensive carbohydrates. Within the range of carbohydrates that is accepted by the microbial strain depending on its genetic constitution, the final substrate choice for commercial production is directed by an

Raw materialsfor fermentation

123

interactive ranking of quality, price, and technological and processing features. The special properties of major carbohydrates as fermentation substrates are listed in Table 3. Pure glucose, sucrose or lactose of food- or even pharma-grade is used in expensive cellculture media for monoclonal antibody production, or in bacterial fermentations for recombinant human proteins that depend on excessive downstream processing and purification. For all other fermentation systems, regardless of the relative impact of raw material costs, sooner or later the price of the carbohydrate becomes an important issue in the development process and is under continuous surveillance in current production. Sucrose, the most expensive food-grade sugar, is therefore a rare fermentation substrate in contrast to the inexpensive byproducts of sugar refining, cane and beet sugar molasses. Cane sugar molasses is an important substrate for ethanol fermentation in Brazil; beet sugar molasses for yeast fermentation, and both types of molasses are still in use for citric acid fermentation. Table 3. Carbohydrate substrates for fermentation.

Substrate

sucrose cane, beet molasses

Fermentable

Price*

carbhydrate

($/kg C6 sugar)

pros

100

0.30-0.90

non-reducing sugar

47-53

0.15-0.25

growth factors

cons

brown colour, high nitrogen and ash content, storageproblems reducing sugar

lactose, edible

94

0.50-0.90

whey powder

75

0.50-1.30

protein, growth factors

whey permeate

79

0.45-0.70

growth factors

lactose crystal. liquors

48

0.15-0.20

high ash content high ash content, storage problems

dextose monohydrate

91

0.60-0.95

reducing sugar

glucose syrup, 70%

70

0.45-0.80

storage problems

potato starch

81

0.50-0.60

non-reducing carbohydrate viscosity (native starch) reservoir substrate

corn starch

85

0.25-0.60

shredded corn

70

0.10-0.45

non-reducing carbohydrate viscosity (native starch) reservoir substrate protein content

fibers, insolubles

* Strong price differences with respect to actual E.U.,U.S. and world market sourcing conditions

124

D. Wilke

Technology dependent substrate replacements Citric acid is a good example in reference to the over-all dynamics of even mature fermentation processes. High substrate yields of citric acid depend on low concentrations of trace metal ions in the fermentation medium in order to favor the accumulation of citric acid rather than its further oxidation to COz. This is accomplished by pre-treatment of the common raw material molasses with ferrocyanide. This pre-treatment is necessary in surface as well as in submerged fermentation; the former, however, has now broadly been abandoned due to over-all equipment and labor cost constraints. In conventional citric acid recovery, precipitated calcium citrate is treated with diluted sulfuric acid, resulting in free citric acid and concomitant gypsum formation. Gypsum deposition in landfills is increasingly difficult; its discharge to gypsum processors is strongly aggravated by the blue color of the material derived from the ferrocyanide treatment. The switch from molasses to glucose is now offering either a more-marketable white gypsum by-product or a complete substitution of gypsum formation by a new recovery process based on solvent extraction or ion-exchange chromatography.

Non-refined raw materials, by-products and waste materials As in case of molasses, the price advantages of less-refined agricultural raw materials and by-

products from food processing, such as whey or lactose crystallization residues, are often jeopardized by the inhibitory effects of the organic and inorganic impurities or by an aggravated product recovery and purification. Especially in bulk fermentations, the use of agricultural and food-processing by-products forwards the waste disposal problems, solid waste, BOD, COD and salt freight, from the primary processor to the fermentation industry. The use of by-products and waste materials must therefore be carefully evaluated with respect to over-all cost savings, keeping in mind the additional complications in processing and logistics. However, there are many examples in which the complex impurities of corn steep liquor, molasses or whey fractions can supplement amino acid, nucleotide, vitamin or salt requirements of the production strain which must otherwise be satisfied by the pure compounds or by still expensive protein hydrolysates and yeast extracts.

Raw materialsfor fermentation

125

Reducing sugars and medium sterilization The inhibitory effects of by-product impurities are often based on undefined Maillard products that are derived from the reaction of reducing sugars with nitrogen compounds. In addition to the potential yield reduction, the brownish color of the Maillard products which provokes additional efforts in the purification process must be considered. The formation of Maillard products is also a problem in medium sterilization. Joint sterilization of glucose or lactose with nitrogen containing medium components results in a strong color formation and often in deterioration of the fermentation course. In principle, this can be avoided by separate sterilization of the reducing sugar or by filter sterilization, neither of which is very feasible for large-scale fermentation processes. Its better sterilization stability would clearly favor sucrose, but only at competitive prices. Native starch is not an alternative because of its gelatinization upon heat exposure.

Starch as carbohydrate substrate The strong viscosity of starch-containing media impedes the necessary oxygen input in agitated aerobic fermentations. However, liquified starch that is treated with normal or heat-stable a-amylase before or during the sterilization step is an ideal carbohydrate substrate for all microorganisms that are able to utilize dextrines. The application of liquified starch or dextrines provides an additional advantage. It can be considered as a storage or reservoir carbohydrate that is gradually released during the fermentation process. It exhibits a low osmotic value and diminishes the substrate inhibition and peak oxygen demands which are otherwise observed at high mono- and di-saccharide concentrations of industrial production media.

Integrated production complexes At a certain production volume per manufacturer and plant site, the access to a feedstock source will definitely render a cost advantage. A strong trend of forward and backward integration can already be seen in the starch and the fermentation industries with respect to citric acid production and the ambitious lactic-acid development projects. There is an even stronger need for backward integration at a production volume and plant size of the fuel ethanol manufacturers, in order to cope with feedstock, by-product and waste disposal logistics. The cane sugar industry in Brazil is in a position to channel various amounts of crude sugar syrup, refining by-products and molasses into ethanol fermentation depending on the

126

D. Wilke

actual market conditions for methanol, ethanol and refined sugar including the sugar export terms. As far as the waste products are concerned, bagasse is burned for energy generation and the

distillation residue, vinasse, is recycled by irrigation, which is said to have been practiced for over hundred years without any ecological damage.

All major US. fuel ethanol manufacturers are back-integrated into the corn wet-milling industry, or (more precisely) are offsprings of the agro-industry. This business structure is a clear prerequisite for cost competitive feedstock sourcing as part of integrated management of all corn refining products. At a certain scale of operation, and provided that an additional million ton bulk product may

show up, the delivery balance of starch, gluten, corn oil and steep water to its established markets may no longer be sustainable. In addition to the price advantage per kg of fermentable carbohydrate, this will favor the transition to the respective unrefined feedstocks cane sugar juice, whole or shredded corn and grain. Naturally, the mass streams of feedstock, by-products and liquid and solid waste materials are then directly influenced. As stated above, many problems are then shifted from the feedstock level to the processing sector, albeit with the advantage that in an ideal set-up the bulk fermentation plant can be integrated into the ago-industrial complex as part of a closed-cycle operation.

For the time being, only bio-ethanol is manufactured in a scale justifying the above considerations. For all other present bulk fermentation products, including citric acid, at least the waste discharge problems can still be handled based on established concepts: liquid wastes to the aerobic or anaerobic sewage plant, landfill or burning for solid wastes, and valorization or discharge as feed additives and fertilizers as alternative outlets for the biomass.

New products from carbohydrates During the past decade, many new fermentation products have been introduced or acquired a commercial market size. Most products belong to the group of high-priced pharmaceutical specialties, either natural microbial secondary metabolites or recombinant mammalian proteins. Without a doubt, the unique therapeutic properties of these products and the unique manufacturing tool of recombinant gene expression will nourish an ongoing development and introduction of new specialty fermentation products - without the necessity of administrative interference, without a major influence on resource management and without any significant impact on the valorization of agricultural surpluses. Only a few new products have been launched into technical markets with varying success. Xanthan, the microbial thickener that competes with plant gums, and poly-P-hydroxybutyric acid as a biodegradable polymer both suffer in large-scale applications from their comparatively high

Raw materials forfemientation

121

price and manufacturing costs. For both products it is not the fermentor yield that creates the cost constraints but the processing hurdles related to the viscosity of the bacterial thickener and the intracellular synthesis of the bacterial polymer, respectively. Similar processing constraints are also envisaged with fermentative surfactants. They emphasize the limits of fermentation technology in the synthesis of functional polymers and point to emerging alternatives such as enzymatic synthesis from monomers or production in transgenic crops. Lactic acid is a good example of the two basic product strategies for organic chemicals from fermentation - performance versus bulk. D-Lactic acid has developed a specialty application in a technical market as an inexpensive chiral precursor for enantiopure herbicides, and considerable potential exists for the many optically-active fermentation products to serve as chiral intermediates

in the synthesis of modern low-dosage therapeutics and pesticides. L-Lactic acid is the favored candidate of several U S . starch manufacturers as a new bulk chemical based on agricultural raw materials. Lactic acid's growth in the food market is broadly saturated; however, lactic acid ethers and polymers are expected to create new technical markets for the traditional compound. A prerequisite for market access is the competitive pricing of lactic acid and its derivatives. This again pertains to downstream processing because substrate and fermentor yields of lactic acid in batch or continuous culture are fairly high. Various groups have worked on lactic acid separation by electrodialysis and on direct esterification and distillation as inexpensive and technically feasible recovery technologies that could replace the old-fashioned calcium lactate/gypsum process. In general, anaerobic fermentation provides the most candidates for new bulk chemicals from renewable resources. Traditional alcohols and acids such as propanol, butanol, propionic acid and butyric acid, but also propanediol and butanediol, may serve as C2- C4- intermediates fitting into an established chemical reaction technology. At changing price and sourcing conditions for common petrochemical feedstocks, these anaerobic fermentation products may experience a revival, provided, however, that their production technologies can be significantly improved by advanced tools such as genetic engineering and metabolic design and by new processing concepts. In the future, chemicals from renewable resources will certainly also be assessed according to their ecobalance and life-cycle analysis. Energy net-gain and all direct and indirect pollution aspects will undoubtedly be considered and applied as competitive arguments. However, based on today's product policies and development rationales, any technological and cost structure improvement will concomitantly lead to an improved ecobalance of these old or new fermentation products.

128

D. Wilke

Who will benefit from the growth potential of new fermentation products? The pharmaceutical and chemical industries dominate the specialty business of fermentation products. The traditional "independent fermentation industry" is rather small; the few existing examples are found among the enzyme and citric acid manufacturers and the biopharmaceutical companies insofar as they have not yet been acquired by the big international players. With respect to bio-ethanol and the lactic acid projects, the situation is a different one. The agro-industry and the U.S. corn wet-millers in particular have invested in these products and are claiming this business as their original field of diversification. By contrast, the chemical industry is more reluctant, insisting on their favored raw materials and established petrochemical processing techniques. In any case, an extended contribution of fermentation to the conversion of carbohydrates into organic chemicals depends on significant R&D and conceptual efforts, but the vision is highly promising. Here again, the U.S. is in a leading position when we consider the Alternative Feedstocks Program of the Department of Energy, which inter aha supports the development of new fermentation processes for a total of twelve alcohols and carboxylic acids. It is my perception that those who start now may enjoy a rather comfortable business beyond the year 2000 in a progressively changing economic, ecological and administrative environment.

7 Lactic acid production and utilization

J.A. Van Velthuijsen PURAC, P.O.Box 2 1, Gorinchem, The Netherlands

Summary. Although racemic lactic acid is also produced by chemical synthesis, most of the lactic acid produced in the world is by fermentation. The fermentation and downstream process will be described. Presently, the large scale fermentation product is the natural L(+)-lactic enantiomer. D(-)-Lactic acid is produced by fermentation only on a small scale. In utilization, L(+)lactic acid and (L)-lactates are preferred in food, cosmetics and pharmaceutical products, as it is the natural enantiomer existing in the body. Commercial derivatives of lactic acid are their salts, esters and polymers. Applications will be mentioned in food, pharmaceuticals, cosmetics, technical industry and as a huilding block in chiral synthesis.

Introduction The PURAC group of companies, the world's largest and most experienced manufacturer of natural lactic acid and lactates, is part of CSM nv, a Dutch multinational company that manufactures and markets a wide range of food products and ingredients. PURAC utilizes a fermentation process, converting pure sugar or dextrose into lactic acid. Lactic acid is the acid formed in the souring of milk, hence the name lactic, from the Latin name for milk. It was discovered in 1780 by the Swedish chemist Scheele. It is 2-hydroxy propionic acid, the simplest hydroxy acid with an asymmetric carbon atom, existing in a racemic form and in two optically active forms with opposite rotations of polarized light. The lactic acid found in muscle tissue was re-examined by Liebig in 1847 and he discovered that it was not identical to the one isolated from souring milk. Later, in 1869, Wislicenus determined the optical activity of the lactic acid isolated from meat extract. In 1874, Van 't Hoff and Le Be1 found an explanation in the conception of different placings of the atoms in space. The industrial manufacture of lactic acid was first established in 1883 in a factory in Avery bear Boston, Massachusetts, followed by a factory of Boehringer in Ingelheim, Germany in 1895.

130

J.A. Van Velthuijsen

In 1936 the Schiedamse Melkzuurfabriek was the first lactic acid factory in Holland, and the production know-how acquired by that company forms the basis of PURAC's technological development.

Lactic acid production Although racemic lactic acid can also be produced by synthesis (via lactonitrile), most of the lactic acid produced in the world is made by fermentation. Because the theme of this part of the workshop is "Fermentation Feedstocks", only the fermentation process will be discussed here. For a Lactic acid fermentation you need: Substratedraw materials sucrose, dextrose, glucose syrups with high DE molasses, whey (lactose) Nutrientsfor the growth of bacteria yeast extract, cornsteep liquor (amino acids and peptides) B-vitamins ammoniumphosphate Neutralizing agents chalk (Ca-carbonate), lime (Ca-hydroxide) caustic soda ammonia Micro-organisms e.g. thermotolerant lactic acid bacteria

Homofermentative L(+)-lactic acid bacteria * Lactobacillus casei' (42 "C) * Lactobacillus helveticus (37 "C) * Lactobacillus acidophilus (40 "C) * Bacillus coagulans (54 "C)

: * * *

*

Lactobacillus delbrueckii (45 "C) ssp. delbrueckii ssp. bulgaricus ssp lactis

Lnctic acid production and iitilization

13 1

Heterofermentative D(-)-lactic acid bacteria

*

Sporolactobacillus sp

Lactic Acid fermentation

; I ..

Milk of Lime

Nutrients

diammonium OhosDhate )sucrose 60% solution M Y Y deni-water M bruden water

h4

8 steam

coolin water

1

!a-

inoculum

7 decantation

Fig. I . Fermenter.

132

J.A. Van Velthuijsen

Fermentation conditions

.

Aqueous solution of substrate/product with suspension of bacteria (and CaC03). In the past, medium was at pH 5.3-5.5 by neutralizing with excess of chalk (CaCO,). Now most of the fermentation processes are at pH 6.0-6.5 by additions of milk of lime (Ca(OH)*) during the acid formation. Temperature: 45-55 "C, depending on the strain. . In most cases the fermentation product is a calcium lactate solution, but now new plants in the United States are neutralizing with caustic soda or ammonia, forming sodium lactate or ammonium lactate as fermentation product. The fermentation process is explained in Fig. 1, which shows the fermenter with periphery.

Downstream process After the fermentation process the aqueous broth must be converted into a final purified lactic acid.

.

Killing and removal of the biomass, now mostly by heating at 80 "Cand increasing pH by adding lime, followed by decantation and filtration. Acidification with sulfuric acid, with the formation of a gypsum precipitate, which can be . removed by filtration or centrifugation. When the byproduct gypsum represents a waste problem, alternative processes can be used. The flow diagram of the filtration and acidification process is shown in Fig. 2.

Purification process A dilute solution of crude lactic acid, containing residual sugars and polysaccharides, amino aciddpeptides and colored substances (Maillard products), must be purified from: cationic impurities : Ca2+, NH4+, Fe*+, etc. anionic impurities : S042-, CIother impurities : acetic-, succinic acid, etc. by passing over a battery of columns filled with several ion exchange resins and active carbon, resulting in a purified solution. Evaporatiodconcentration to a 50-90% food grade (edible) lactic acid.

Lactic acid production and utilization

c crude dilute acid tank

Fig. 2. Biomass removal and acidification.

133

134

J.A. Van Veltliuijsen In order to obtain a "pharmaceutical" grade, this edible acid is esterified to methyl lactate or ethyl lactate, which is distilled and hydrolyzed with water, to give very high purity lactic acid. The entire flow diagram of the lactic acid process is given in Fig. 3. F l o w Lactic Acid Production

Fcrm

Aka1

Acid

Emterificntion Hydrolysation

Pharmaceutical

Fig. 3. Block diagram production process.

T Edible

Lmtic acid production and utilization

135

Utilization The use of racemic lactic acid and lactates or one of the two optically active forms depends on the local availability and the application. In the past only racemic lactic acid was available, because the fermentation and its bacteriological control were not capable of producing a pure enantiomer, thus ending with a fermentation product containing 40-60% of both enantiomers. Later synthetic lactic acid came on the market as

50/50 racemic material. Now, most of the fermentation product is the natural L(+)-enantiomer. This is the preferred form as it is the only enantiomer formed by metabolism in humans and animals. L(+)-Lactic acid and the @)-lactates are preferred in food, cosmetics and pharmaceutical products, as this is the natural form existing in the human body, even without physical exertion, where it is produced and metabolized at an amount of 125 g per day. The D(-)-lactic acid is formed by the bacteria in the gut of animals, so there is a low level of D(-)-lactic acid detectable in blood and urine. D(-)-lactic acid is produced by fermentation on a small commercial scale, mostly for captive use, and is used only for chiral syntheses to prepare specific enantiomerically pure intermediates for the synthesis of pharmaceuticals or herbicides. Only in Japan and the United States, where synthetic (racemic) lactic acid is produced, this acid is widely used, but the total volume is much smaller than that of the present world-wide production of L(+)-lactic acid by fermentation.

Production of lactic acid derlvafives

Fig. 4.Derivatives production.

136

J.A. Van Velthuijsen

Lactic acid is used as the free acid in industry, and is also converted to derivatives such as salts and esters. The diagram (Fig. 4)shows that esters can be prepared starting from edible lactic acid and alcohols. Furthermore, salts such as sodium lactate, and metal lactates can be produced, as well as lactide and polylactic acid. The product line of existing commercial products is shown in Fig. 5. With regard to the properties of these products, in general the salts are very soluble in water and are used as mineral supplements in diets; the esters are good solvents and are starting materials for chiral syntheses.

Product h e

Fig. 5 . Lactic products.

Biodegradable polymers Like all 2-hydroxycarboxylic acids, lactic acid is subjected to inter-esterification. When it is concentrated and heated, a polycondensation starts (Fig. 6). Via condensation reaction only polyesters with a low molecular weight (MW: 2000 D) can be produced, because the removal of water from the viscous product becomes limited. For high molecular weight polymers, a purified dilactide, such as the L(-) dilactide (Fig. 7) has to be produced; this is polymerized via ring opening polymerization using special catalysts (Fig. 8).

Lnctic acid production and utilization

137

0

Lactic acid

Trirner

Dirner

Fig. 6. Lactic acid polycondensation. Dilactide occurs in four different forms: the L-form (S,S), the D-form (R,R), the meso-form

(R,S) and the racemic DL-dilactide, the mixture of D- and ~-dilactide. Homopolymer P(L)-LA is crystalline and mechanically strong; the copolymer P(DL)-LA is amorphous. Copolymers with glycolide can also be made. The polymers are biodegradable and also biocompatible in medical- and pharmaceutical applications (surgery and controlled drug release).

Dilactide

Fig. 7. L (-)-Dilactide (S,S).

Poly-lactic acid (P-LA)

Fig. 8. Ring opening polymerization

The preparation of biodegradable plastics on an industrial scale is now nearly reality as several large multinational companies are entering this field with products based on L-lactic acid. These are thermoplastics that can be processed with the equipment normally used in the plastic industry. (Bio)degradation takes place by hydrolysis with water to lactic acid. Moreover, it is compostable and the material can be recycled by dissolving the polymer in alkaline water (pH >lo,

60 "C) making a lactate salt, or by acid hydrolysis at 1.50 "C.

138

J.A. Van Velthuijsen

Applications of lactic acid and lactates We can classify 3 types of uses: food, pharmaceuticskosmetics and industrial applications.

Food The general functions of lactic acid in food products are flavoring (mild acid taste), pH-regulation and preservation. As a food acid, it is the preferred acid in dairy products and dry sausages because of its natural flavor. As a preservative, lactic acid occurs naturally in a number of food products (sauerkraut, pickles, olives) and it can be added to enhance this effect. The inhibition of microbial growth depends on pH, concentration of undissociated acid and a specific acid effect. Both lactic acid and acetic acid are weak acids, occurring in dissociated and undissociated forms. In foods with pH 4-6, a significant part is undissociated, which is hundreds of times more effective in growth inhibition than the dissociated forms. Organic acids therefore have a stronger antimicrobial effect than inorganic acids. We can divide the food applications into:

Bakery Lactic acid powder in rye bread . Calcium lactate in whipped bakery products . Beverages Lactic acid in wine, cider and acidified beverages . Lactic acid in beer (mashing, wortcooking, final pH) . Confectionery (candy industry) . lactic acid in wine gums buffered lactic acid in high-boiled sweets . Dairy products cottage cheese, imitation Mozzarella cheese . . butter (acidification of sweet cream) . calciumlactate in pudding powders Animalfeed . lactic acid: digestible by young animals . calcium lactate for prawns

Lactic acid production and utilization

+ + + +

Flavor andjizgrances . savory flavors (meat components) . marinades Preserves and sauces . Jams and marmalades . Mayonnaise and salad dressing Meat andfish . casings for sausages (acid) . decontaminant in slaughtering lines (acid) . sodium lactate in processed meat, poultry and fish Others semi-bulk storage of vegetables (cover brines) . . calcium lactate as firming agent (fruitshegetables)

Pharmaceuticals/cosmetics We can divide these applications into:

+

+

MedicaVphnnaceutical . Ca-lactate in calcium tablets Co-polymers in controlled drug release . Sodium lactate in dialysis solutions . Skin and hair care Lactic acid (skin renewal process) . . Sodium-, ammonium-lactate (skin moisturizer) . Hair conditioners

Industrial applications These can be divided into:

+

Electronics Lactate esters in solvents photoresist formulations .

.

Solder flux remover

139

140

+

+

+

J.A. Van Velthuijsen Cleaning . Replacing ozone-depleting solvents . Degreasing/cleaning of metal surfaces Coating and ink . Cataphoretic electro-deposition coating (acid) Solvent for coating and ink (esters) . Others . Electroless plating of metal . Electroplatingkomplexing Acid in mordant dyeing of textiles .

The chiral pool of lactic acid Enantiomerically pure lactates are relatively inexpensive raw materials for the chiral synthesis of optically active pharmaceuticals and herbicides, for the replacement of existing racemic products, and for the development of new active substances. Halogenation of lactate esters for enantiomeric pure 2-chloro-propionates by nucleophilic substitution (Walden inversion). Acylation of OH-group to esters using acid anhydridedacid chlorides (retention of stereochemical configuration). Reduction after protection of OH-group to propanediol-derivatives Mesylation/Tosylation of esters with methanesulphonyl chloride or with p-toluene sulphonylchloride (retention of configuration). AlkylatiodArylation under ether formation of sec. OH-group (herbicides). Amidation and conversion of lactate esters to amides (retention of chiral center).

8 Starch and dextrins in emulsion copolymerization

M. Bodiger, S. Demharter and R. Mulhaupt Freiburger Materialforschungszentrum und Institut fur Makromolekulare Chemie der Albert-Ludwigs Universitat, Stefan-Meier-Strasse 3 1, D-79104 Freiburg i.Br., Germany

Summary. Dextrins and their hydrophobic, anionic and cationic derivatives were applied in methylmethacrylate/butylacrylatc ( S O / S O ) emulsion polymerization in absence and presence of ionic sodium dodecylsulfate emulsifier. The effects of process variation and dextrin type on dispersion properties, in particular dextrin incorporation, latex morphology, viscosity, reaction kinetics, and paper coating application, were studied. An analytical method involving sedimentation was dcvcloped to identify insolubilized dextrin resulting from adsorption and compatibilization. Compatibilization of synthetic polymers and dextrins was achieved via grafting, salt formation, and dextrin hydrophobization. Core/shell- and multiphase-latex were ohtained in a starved-feed-emulsion polymerization. A new class of carhohydrate-based amphiphiles useful as non-ionic emulsifiers and gel-forming agents consisted of semicarbazone derivatives of carbohydrates, e.g. maltose and maltotriose.

Introduction For many years starch and dextrins have been used as carbohydrate-based binders together with dispersions of synthetic polymers, e.g. poly( styrene-co-butadiene) and polyacrylates. Dispersions of this type are applied as coatings for paper or textiles and as components of adhesives, cosmetics, and pharmaceutical products. 1-4 Starch components are known to improve paper coating properties such as dry pick strength, stiffness, film formation, rheology, and economics. However, several drawbacks are associated with high molecular weight and water-solubility of starch, such as reduced solids contents, low wet pick strength, binder migration and leaching. In order to achieve property synergisms of synthetic polymers and carbohydrates, starch molecular weights must be reduced and both mutually immiscible components must be made compatible, thus improving interfacial adhesion and simultaneously reducing water-solubility of the carbohydrate component. An important research objective is therefore to tailor dextrins as starch substitutes for use in dispersion applications. In principle, two strategies appear feasible for overcoming compatibility problems: (A) using dextrin derivatives or sugars that are effective as emulsifiers, and (B) using compatibilized dextrins. In the traditional approach (A), dextrins form

142

M. Bodiger, S.Demharter and R. Miilhaupt

protective colloids where dextrins adhere to the latex surface and provide steric and ionic stabilization. In contrast to passive carbohydrate binders, which are added after completing emulsion polymerization, modified dextrins claim an active role as emulsifiers and give control on both latex particle formation and stabilization. In approach (B), dextrins are tailor-made to match their compatibility with synthetic polymers, thus serving as a blend component. In addition to hydrophobization, numerous patents and publications, e.g. reviewed by Berlin and Kislenkos, claim compatibility improvement via grafting synthetic polymers such as polyacrylates, polystyrene, poly(styrene-co-butadiene)onto carbohydrates. Grafting has been achieved by means of several processes, illustrated with selected examples below. ( I ) Chain transfer with polysaccharides to form polymeric free radicals that intitiate free radical polymerization.5.6

CH,OH

C,H,OH

Q

Q(

OH

OH

---

(2) Chain ransfer with xanthates7.loor thiols.' 1.12

Na' (3) redox initiated grafting involving carbohydrate metal complexes, e.g. Cea+ cornplexesl~-19,Mn" complexes20.2l or V5+ complexes.20

(4) Carbohydrate-macroinitiators based upon anthranilate22, a2023 or xanthate7-10derivatives

Starch and dextrins in emulsion copolymerization

143

( 5 ) Acrylate-functional carhohydrates.24

Although xanthate and thiol derivatives of dextrins are much more efficient chain transfer agents with respect to sulfur-free carbohydrates, odor and color formation and the use of toxic and flammable carbon disulfide have hampered application. Macroinitiators, which are prepared in a separate step, give excellent grafting yield accompanied by extensive homopolymerization of the graft comonomers. More recently, Gallot and Marchin24 have converted glucamine into monofunctional acrylic glucamide. Conversion of dextrins, for example by reacting them with diacrylicacid anhydrides, would afford polyfunctional acrylates which cross-link upon exposure to free radicals. Most favorable with respect to the formation of transition-metal-free graft copolymers in emulsion copolymerization processes is chain transfer involving carbohydrate components and synthetic polymeric free radicals. Depending on both the compatibility difference between synthetic polymer and dextrin and the mixing mode, either corekhell latexes with synthetic polymer core and dextrin shell or

Fig. I . Latex morphology of synthetic polymer (dark) modified with dextrins (grey): corekhell (left) and multiphase latex (right).

144

M . Bodiger, S. Demharter and R. Mulhaupt

multiphase latexes with dextrin being dispersed in the continuous matrix of the synthetic polymer are feasible (Fig. 1). Core/shell-type morphologies are typical for compatibilized dextrins, dextrin emulsifiers as well as passive binders. In the case of dextrin binders there exists an equilibrium between dextrin loosely attached to latex surface and dextrin in solution. Our research is aimed at understanding the influence of compatibilized dextrins on emulsion copolymerization of methylmethacrylate (MMA)/butylacrylate (BAc) (50wt.-%/50wt.-%), especially with respect to dextrin insolubilization, latex particle sizes, latex morphology, influence of process variation, such as batch process versus starved-feed process, and properties of compatibilized dispersion blends of carbohydrates and synthetic polymers in paper coating applications. An important objective is to reduce the amount of water-soluble dextrin via compatibilization. Moreover, blends of dextrin binders with latex are compared with compatibilized dispersion blends in paper coating application.

Experiment Details Materials: Maltose, non-functional dextrin (7380), carboxylic-acid-functional dextrin (5582), cationic 2-hydroxypropyltrimethylammoniumchloride-functional dextrin (CWD 3 1.3, 2-hydroxypropyldimethyllaurylammoniumchloride-functional dextrin (HCDI), hydrophobic heptanoatemodified dextrin (HD 1) and hydrophobic n-octenylsuccinic-halfester-modifieddextrin (3nOSD, 5nOSD) were obtained from Eridania Beghin-Say in Vilvoorde/Belgium and used without further purification after drying at 90 "C by means of a Sartorius Moisture Analyzer MA30. Properties such as molecular weight, polydispersity and degree of substitution are listed in Table 3. Starved-feed emulsion polymerization: a solution of lOOg dry dextrin in 500ml distilled water was charged into the 2L reaction vessel of the METTLER RC1 reaction calorimeter. In order to calibrate the reaction calorimeter the temperature was increased to 77 "C and then heated to 80 "C with 0.2 "C/min heating rate to determine the system's heat capacity. Then log 0.1 molar aqueous K2S208 solution was added. Next 300g monomer solution (MMA/BAc 50 wt.-%/50 wt.-%)) containing 0.5wt.-% tert.-dodecyl- I-thiol and an initiator consisting of 90g 0.1 molar aqueous K2S208 solution were fed simultaneously by means of two feed pumps (lg/min monomer and 0.3g/min initiator feed rate). Polymerization was carried out for the duration of seven hours at 80 "C and 100 r.p.m. stirrer speed. The resulting dispersions contained approximately 40wt.-% solids. When the polymerization reaction was completed, calibration was performed again to compensate heat flow changes caused by viscosity build-up. Then concentrated aqueous ammonia solution was added in order to reach pH6. A small amount of sodium azide was added to stabilize

Starch and dextrins in emulsion copolymerization

145

the dispersion against degradation caused by microorganisms. Filtration over a coarse fritte containing glass wool gave less than 1wt.-% coagulate.

Starved-feed emulsion polymerization feeding preemulsijkd monomer: monomer solution and dextrin solution were emulsified in a small reactor using Ultra-Turrax (IKA T25, S25 N-25F) dispersing tool at 8000 r.p.m. This pre-emulsion process was carried out continuously with two minutes hold-up time prior to feeding the emulsion into the reaction calorimeter containing distilled water. Particle size measurement: particle sizes were determined by means of photon correlation spectroscopy (Malvern Instruments, Zetasizer 3) using highly diluted latex at a 90" angle and 20 "C.

Particle tnorphologies were determined by means of transmission electron microscopy (Zeiss CEM 902) on lOOnm sections of concentrated dispersions embedded in melamine resins. In an alternative method, a carbon-coated copper grid obtained from Serva was dipped in diluted latex. For contrast enhancement, thin slices or sprayed particles were exposed to uranylacetate or phosphoroustungsten acid. As selective staining reagent for imaging of starch, gaseous Os04 was used as previously described by Moller and Glittenberg.4 Typically both sample types were exposed to OsO4 vapors for the duration of thirty minutes. Determination of insoliibilized dextrin: The dextrin content in aqueous phase prior to and after emulsion polymerization was determined by means of centrifugation and lyophilization of the aqueous phase. It is important to take into account limited water-solubility of some dextrins and water-solubility of certain components of the synthetic polymer. The percentage of insolubilized dextrin D, resulting from adsorption as well as grafting, was determined according to the following equation:

with SD= water soluble dextrin fraction prior to polymerization (16.7wt.-% dextrin in water was heated at 80 "C for the duration of five hours; then the water-insoluble dextrin fraction was separated in five hours at 10 "C and 23000 r.p.m. followed by lyophilization to determine water soluble dextrin). Ss,p is the water-soluble fraction of the synthetic polymer (with respect to dextrin), e . g . oligomers and KHS04, which are found in the aqueous phase and cannot be separated from soluble dextrins by means of centrifugation. Ss,p is determined by polymerization of the monomers in the absence of dextrin and subsequent centrifugation/lyophilizationprocedures

146

M. Bodiger, S. Demharter and R. Miilhaupt

as described above to identify the total content of water-soluble components remaining after polymerization. S D ,is~ the total amount of both water-soluble dextrin and water-soluble synthetic polymer, which are found in aqueous phase after completing emulsion polymerization in the presence of dextrin. Since SDcan vary dramatically as a function of dextrin type, temperature and duration of heating, this method, which takes actual dextrin solubility and content of water-soluble oligomers into account, represents a reliable measure for insolubilization of dextrin resulting from adsorption and grafting. It should be noted that this method cannot distinguish between adsorption and grafting.

Paper Coatings. Filter paper, 20cm/10cm (Schleicher & Schull, type 1450),was immersed in the dispersion for the duration of 60 seconds. Then the impregnated papers were sandwiched between steel rolls at Ikg pressure. After drying ten minutes at 110 "C and three minutes at 140 "C, tensile bars ( 15mm/60mm/0.28mm) were cut perpendicular to fiber axis. Usually the impregnation consisted of 0.5 to 0.6g/100cm2. Strength measurements on 12 tensile bars were performed on an Instron 4204 with lOON and 5mmlmin crosshead speed. For dry strength measurements, paper samples were stored 24 hours at 50% humidity and 23 "C. For wet strength filter papers were stored for one hour in water prior to measurement. Water-uptake was recorded according to procedures described by Cobb and DIN 53 132.

Results and Discussion Interaction of Latex and Dextrin Binders Emulsion copolymerization of methylmethacrylate (MMA) and butyacrylate (BAc) with M M A B A c (50/50) was performed using K2S20g initiator at 80 "C for the duration of seven hours at 80 "C in absence and presence of sodium doceylsulfate (SDS). As a result of the thermal decomposition of the peroxydisulfate, forming suloxy radicals which initiate polymerization, or SDS addition, respectively, the resulting latexes were ionically stabilized by sulfate ions located at the latex surface. Dextrins and dextrin derivatives were added either as conventional binders after copolymerization was completed or as additives during copolymerization. For evaluation of the dextrin rendered insoluble as a result of adsorption or compatibilization, several analytical methods were evaluated. Solvent extraction and dialysis proved to be of very limited value. Sedimentation by means of centrifuge followed by lyophilization of the aqueous phase, thus seperating latex and soluble fractions, gave reproducible degrees of insolubilization when water soluble dextrin content prior to and after polymerization were compared. It should be noted that dextrin solubility and formation of water-soluble oligo(methylmethacrylate-co-butylacrylate)and KHSO4 must be taken

Starch and dextrins in emulsion copolymerization

147

into account. Dextrin solubility in water also depends markedly on dextrin type as well as time and temperature and would cause unacceptable reproducibility problems. This sedimentation technique is described in more detail in the experiment section. Table I . Dextrins as binders and additives of Poly(MMA-co-BAc) Latex

dextrin

dextrin type@

7380

non-modified

latexb)

modec)

insolubilized dextrin (wt.-%)

K2S208

binder binder

7 15 7

in situ

27

binder

19

in situ

27

~

in situ 5nOSD

anionic & hydrophobic

HCD 1

cationic &

K2S208/SDS

hydrophobic Properties are listed in Table 3, b, starved-feed polymerization is reported in experimental part: lOOg dextrin, 150g MMA, 15Og BAc, lOOg O.IN K2S20s,80 "C, 100 r.p.m., 7h, in the case of cationic dextrin 7g SDS were added to dextrin solution, c , dextrin added as binder after polymerization was completed. in siru: dextrin was present during copolymerization.

Three dextrins of similar molecular weights were added either before or after MMA/BAc copolymerization: non-modified dextrin (7380), hydrophobic anionic dextrin obtained by modification of dextrin with octenyl-succinic-anhydride (5nOSD) and cationic hydroxy-

propyldimethyllaurylammoniumchloride-functionaldextrin (HCDI ):

OH DEX-OH

0

rr

DEX-CI CH~FHCOOH

Interestingly, when poly(ethyleneoxide)-block-poly(methylmethacrylate)with Mn(PEO) = 3000g/mol and Mn(PMMA)= 1000g/mol was present during polymerization, no adsorption and insolubilization of dextrin was detected. In Table 1 adsorption of dextrin binders and insolubilization resulting from adsorption as well as in-situ compatibilization via chain-transferinduced grafting are compared as a function of dextrin type. Cationic dextrin addition caused coagulation when only sulfate groups originating from K2S208 initiator were present as ionic stabilizers, whereas SDS/K2S208 gave stable dispersions with 40% solids also in the presence of

148

M. Bodiger, S. Demharter and R. Miilhaupt

cationic dextrin HCDI. As a rule, dextrin insolubilization was much higher when dextrins were added as additives during polymerization. High incorporation of dextrin was obtained when anionic latex was combined with cationic dextrin HCDl or anionic 5nOSD (see also Table 1). In spite of small amounts of cationic functionality, the resulting dispersions contained much less cationic dextrin in the aqueous phase. As illustrated in Scheme 1, this can be attributed to formation of quarternary ammonium salts involving ammonium groups of dextrins and sulfate groups at the latex surface. SDS sulfate groups can also form hydrophobic ammonium salts which contribute to amphiphilic properties of dextrins.

p-aso;

K+ +

-

.+ CI NM~,R-DEX

+ -

P-OSO; ~ , R - D E X + K CI

Scheme I . (P: latex or n-doceyl, respectively)

Influence of polymerization process conditions Incorporation of dextrin also depended upon process condition, especially monomer and initiator feed rates. In fact, batch processes gave poor process control and instable dispersions with broad particle size distribution and very high viscosity. In Table 2 rapid feed of monomer (300g/h) is compared with starved feed (50g/h), where polymerization rate is much faster than feed rate. In both cases initiator solution was fed simultaneously at constant feed rate. In the presence of excess monomer, similar to batch polymerization, copolymerization was difficult to control and afforded viscous dispersions with bimodal particle size distributions. In contrast to the rapid feed process, starved-feed copolymerization was the method of choice, circumventing all drawbacks encountered in the case of batch-like processes. As is apparent from Fig. 2, polymerization-rate time functions Table 2. Comparison of starved-feed and rapid-feed MMAlBAc copolymerization.")

process

starved feedb)

rapid feedexcess monomelc]

good

poor

dispersion viscosity (cP)d)

102

104

dextrin incorporation (9%)

28

16

monomodal

bimodal

200

170/2500

Properties dispersion s ta bi 1ity

particle size distribution particle size (nm) standard deviation

75 > 1000 lOOg dextrin 5nOSD in 500 mL water, 300 g MMA/BAc feed, lOOg 0.1Naq. K2S2O8, 80 "C, 100 r.p.m., 7h. b, monomer feed within 6 hours (50g/h), c, monomer feed in 60 min (300g/h), dl Brookfield (T=25 'C, 10 r.p.m.),

Starch and dextrins in eniulsion copolymerization

149

are very similar to those expected according to Smith-Ewart kinetics. Moreover, dextrin incorporation almost doubled in the starved-feed process.

heat flow (W) heat flow

7 0-

i / y 1 L . ,

50

0

1

feed

r

feed (d

feed

: -0

.-, 300

Fig. 2. Polymerization rate. expressed as heat flow as a function of time for rapid and starved-feed MMA/BAc copolymerization (METTLER RC 1 reaction calorimeter with simultaneous feed of monomers and initiator solution).

In rapid and starved-feed emulsion copolymerization as well as postpolymerization binder addition, similar core/shell particle morphologies were detected by transmission electron microscope (TEM) imaging using uranylacetate and Os04 staining. As is apparent from Fig. 3 (left) the synthetic polymer formed the core which was embedded i n a dextrin shell. Few microphases were visible within the poly(MMA-co-BAc) core. However, when the comonomers were emulsified in aqueous dextrin solution prior to polymerization and then fed into the reactor according to starved-feed conditions, the resulting particles exhibited multiphase morphologies with subinclusions clearly visible in the polymer core. Most likely, this multiphase latex was not formed by phase transfer and phase dispersion processes outlined i n Fig. 1. One possible explanation for the formation of these multiphase latexes is the formation of very fine core/shell latexes which agglomerate to afford the observed particle morphologies.

M.Bodiger, S. Deniharter and R. Miilhaupt

150

Fig. 3. TEM images of poly(MMA-co-BAc) latexes prepared by starved-feed polymerization feeding monomer solution (left) and monomer emulsions in aqueous dextrin solution (right).

Emulsifier-free Emulsion Copolymerization in the Presence of Various Dextrins In the presence of 2wt.-% (with respect to total amount of monomers) methacrylic acid, emulsion copolymerization of MMA/BAc (50150)did not require emulsifiers. The results of emulsifier-free copolymerization are listed in Table 3. While cationic dextrins gave the best performance in Tahle 3. Emulsifier-free MMAh3Ac copolymerization.a1

dextrin

modification

DS of OHh) (9% of OH)

MnC)

MwC)

MwmnC)

dextrin incorporation I%)

maltose 5582 CWD3 15 HCD 1 7380 HD1 3nOSD 5nOSD

none COOH HFTMAe) HPDLA~ none heptanoate 0%)

0s)

0.72 1.80 1.33 1.67 1.03 1.47

n.d.d) 3270 2840 3240 3530 2830 3020 3260

n.d.d) 11610 5950 6630 15680 13760 16280 15160

n.d.d) 3.6 2.1 2.1 4.4 4.9 5.4 4.7

3 9 10

12 14 16

18 23

a) l00g dextrin in 500ml water, simultaneous feed of 300g MMA/BAc containing 2wt.-B methacrylic acid (MAA) and l00ml 0.IN K,S20, within 6h, reaction time 7h, 80 "C, 100r.p.m., b, DS: degree of substitution in percent, c , siLe exclusion chromatography (SEC) in 0.05N aq. NaOH, dl n.d. not determined, HPTMA: 2-hydroxypropyItrimethylarnmoniumchloride, n HPDLA: 2-hydroxypropyIdimethylaurylammoniumchloride,g) 0s: octenyl-succinic anhydride which forms halfester when reating with dextrin hydroxy groups.

Starch and dextrins in emulsion copolymerization

15 1

ionically stabilized dispersions, cationic dextrins did not perform as well in emulsifier-free copolymerization. The best results with respect to dextrin incorporation were found for octenylsuccinic-anhydride-modifieddextrins (3nOSD and 5nOSD) with 18 or 23% dextrin incorporation, respectively. The heptanoate-derivative also gave 16% dextrin incorporation. This can be attributed to the amphiphilic character of heptanoate and succinate-modified dextrins.

Semicarbazone-based carbohydrate amphiphiles as non-ionic emulsifiers Most modified dextrins are rather ill-defined compounds due to the hetereogeneous nature of the chemical polymeranalogue reactions. Therefore, novel carbohydrate amphiphiles were prepared via semicarbazone coupling of N-alkyl- and N-alkyl-hP-methyl-semicarbazides and aldehyde groups of sugars and oligo- and polysaccharides.25.26In a typical reaction sequence displayed in Scheme

2, long-chain alkylamines were converted into semicarbazides via activated carbamoyl derivatives. Then semicarbazides were reacted with aldehyde groups of maltose or maltotriose, respectively, to yield semicarbazone amphiphiles in high yields. Similar to glucamides described by Pfannemuller27 and Fuhrhop28, semicarbazone amphiphiles aggregate to form highly organized supramolecular architectures. In fact, at concentrations
F

&H

O

H CH=N-NH-CO-NH-R'

CH=O

-0

-0 OH

OH

I

+ YN-NH-CO-NH-R'

OH

Scheme 2. Sernicarhazone amphiphiles (saccharide: maltose. maltotriose, R: n-alkyl chain).

152

M. Bodiger, S. Demharter and R. Miilhaupt

semicarbazone concentration was raised well above CMC, the resulting particle sizes were found to be very similar, ranging from 120 to 160nm when the solids content was approximately 16wt.% polystyrene.

800

0,16 0,14

.?

600

0,12 0

2 E

400

&l

0,1 0,08 0,06 0,04 0,02 0

mi 2oo

P

0 1

2

3

4

5

,o

-

3 u

5

6

number n

Fig. 4.Polystyrene particle size (bars) and CMC (line) as a function of number n of C atoms of CnH2"+,alkyl-chain in semicarhazone amphiphile derived from maltose.

Paper Coatings In order to investigate whether compatibilized dextrin blends with synthetic poly(MMA-co-BAc) offered any advantages over conventional dispersions and conventional dextrin binders, which were added to the synthetic latex after polymerization was completed, filter paper was coated with

stress (A)

1

dextrid

40-BAc)

dispersion blend

P(MMAIBAc)

Fig. 5 . Wet strength (shaded bars) and dry strength (dark bars) of filter paper coated with dextrin solution, dextrin/poly(MMAco-BAc) mixture, poly(MMA-co.BAc) latex, and cornpatibilized dextrin/poly(MMA-co-BAc) dispersion hlend.

Starch and dextrins in emulsion copol.ymerization

153

dextrin, synthetic latex, compatibilized blend of synthetic latex and dextrin, and the corresponding mixtures of dextrin with synthetic latex. Dextrin content of dispersion blend and mixture were identical. The results are summarized in Fig. 5. As expected for dextrin-containing dispersions, wet strength performance of mixture and compatibilized dispersion blends did not exceed that of dextrin-free synthetic latex. However, blends of compatibilized dextrin and synthetic latex gave very similar performance with repsect to pure synthetic latex. Moreover, wet strength of the compatibilized blend was markedly better than that of the mixture. Dry strength of the compatibilized dispersion blend was much better than dry strengths of synthetic latex, noncompatibilized mixture and pure dextrin. This is a good indication that compatibility and incorporation of dextrin in a compatibilized blend of dextrin and synthetic polymer can lead to unexpected property synergisms which are not predictable on the basis of mixing ratios and properties of the individual blend components.

Conclusion The comparison of non-compatibilized dextrin mixtures with synthetic latex and the corresponding compatibilized dispersion blend of dextrin with synthetic latex revealed that blend concepts based

on improved compatibility were successfully applied to dispersions useful in paper coating applications. Compatibility between synthetic and dextrin blend components was achieved using different strategies, for example starved-feed emulsion copolymerization and chemical modification of dextrins. In addition to ionic interactions between cationic dextrin derivatives and anionically stabilized synthetic dispersion, efficient in-situ grafting of the synthetic polymer onto dextrin gave substantial improvements, especially with respect to viscosity, particle size, narrow particle distribution and improved properties of coated filter papers. This blend technology promotes applications of polymers from renewable resources in dispersion applications.

Acknowledgement. This research was sponsored hy the Bundesminister fur Forschung und Technologie as part of the BMFT project "Legierungen von synthetischen und nativen Polymcrcn in Dispersion" (Fiirderkcnnzeichen 03 10020A) carried out jointly with the Freihurger Matcrialforschungszentrum in Frcihurg, Synthomer Chemie GmhH in Frankfurt. and Cerestnr (Eridnnia Beghin-Say group) in Vilvoorde/Belgium. The authors thank Dr.K. Miiller, Dr. N. Eidam at Synthorncr Chemie GmhH, Dr. H. Koch, Dr. H. Roper, and Dr. K.-H- Bahr at Ceresiar for thcir assistance.

154

M. Bodiger, S. Demhurter und R. Miilhaupt

References I. 2.

H. Koch, H. Roper, Starch 40 (1988) 121. H. Koch, H. Roper, Starch 40 (1988) 453.

3. 4.

N.-.O. Bergh, W. Luckenhaus, Wochetiblarr fur Papierfabrikufion, 114 (1986). 75. K. Moller, D. Glittenherg, "Novel starch containing polymer dispersions as coating hinders", 1990 Coating Conference, May 1990, TAPPI Press, Atlanta 1990. A.A. Berlin, V.N. Kislenko, Prog. Polym. Sci. 17 (1992) 765. V.N. Kislenko, A.D. Berlin, Cell. Chem. Techn. 26 (1992) 387.

5. 6. 7. 8. 9. 10.

I I. 12. 13. 14.

15. 16. 17. 18. 19 20 21 22 23 24 25 26 27 28

A. Higazy, A. Bayazeed, A. Hebeish, Angew. Makromol. Chenr. 169 (1989) 101. A. Bayazeed, Farag, A. Hebeish, Srarch 38 (1986) 268. O.Y. Mansour, A . Nagaty. Prog. Polym. Sci. 1 1 (1985) 91. B.V. Kotka, C. Deneault, ACS Symp. Ser. 187 (1982) 269. H. Baczkowicz, H. Palasinski, P. Tomasik, Starch 34 (1982) 413. D.K. Chaudhuri, J.J. Hermans. J . Polym. Sci. 48 (1960) 159. B.K. Patel, V.K. Sinha, C.P. Patel, H.C. Trivedi. Starch45 (1993) 178. F.E. Okieimen, F. Egharevha, A. Jideonowo, Angew. Makroniol. Chenl. 184 (1991) 1 . F.E. Okieimen, O.B. Said, Acra Po/yni. 40 (1989) 708. S.H.0 Eghoh, B. Jinadu, Atigew. Makroniol. Chem. 163 (1988) 93. 1. Goni, M. Gurruchaga, M. Valero, G.M. Guzman, J.Po1ytn. Sci,, Polym. Chem. Ed. 21 (1983) 2573. M. Vera-Pacheco, H. Vazquez-Torres, G. Canche-Escamilla, J . Appl. Polym. Sci. 47 (1993) 53. A. Bazuaya, J . Polym. Sci., Purt C: Polym. Lett 27 (1989) 433. D. Takahisa, R. Cenita, R. Bengt, Macrottiolecirks 17 (1984) 25 12. R. Mehrotra, B. Ranhy, J . Appl. Polyni. Sci. 21 (1977) 1647. R. Kniewske, F. Krause, Ger. OfSeen.DE 3430676 (1986) assigned to Maizena. R. Kniewske, Eur. Pat. Appl. EP175517 (1985) assigned to Maizena. B. Gallot, B. Marchin, Liquid Crysrais 5 (1989) 1729. S. Demharter. W., Richtering, R. Miilhaupt, Polymer Bull. 34 (1995) 271. S . Demharter, H. Frey. M. Drechsler, R . Miilhaupt. Colloid Polym. Sci. 273 (1995) 661. B. Pfannemiiller, Srarch 40 (1988) 476. J.-H. Fuhrhop, W. Helfrich, Chenr. Rev. 93 (1993) 1565.

9 Synthesis of new 'Saccharide polymers' from unsaturated monosaccharides

K. Buchholz, S. Warn, B. Skeries, S. Wick and E-J. Yaacoub Institut fur Technologie der Kohlenhydrate an der Technischen Universitat Braunschweig - Zuckerinstitut

D-38 106 Braunschweig, Germany

Summary. The synthesis of unsaturated monosaccharides is described; these are used as monomers i n radical polymerization reactions. Glucose, fructose and ribose were used as sources and derivatives of glycal-, pseudoglycal- and exoglycal-type were (co)polymerized. The preparation and some characteristics of the new type of linear 'saccharide polymers' are presented both for homo- and copolymerization. Depending on the comonomers and the polymerization procedure, a wide range of molecular weights were obtained, varying from 4,500 to 106,000. The structures and compositions of the soluble copolymers were established by elemental analysis, 'H and I3C NMR, and FT-IR spectroscopy. Due to the difference of monomer rertctivities, the copolymers contain 5 to SO mole-% sugar.

Introduction The synthesis of vinylsaccharides as a potential source for the preparation of polymers mainly through radical polymerization reactions has been thoroughly investigated. In recent decades many papers have been published about the synthesis and properties of the 'Polyvinyl~accharides''-~. The synthesis of a new kind of saccharide based polymers'o, called 'saccharide polymers', in which the sugar moieties are an integral part of the polymer backbone and not supported on it (as a side group), presents an alternative to the polyvinylsaccharides because they have different structures and, consequently, other properties. For this purpose, unsaturated monosaccharides have been synthesized and used as sources for polymerization by radical reactions. The unsaturated carbohydrates have a structure that is different from the vinylsaccharides, because the polymerizable olefinic double bound is in 'endo'- or 'exo'-cyclic position of the

K. Buchholz et al.

156

/OR

fOPOR

& fO> RO

RO

Type A : " Glucal "

Type B :

@)OR

RO

"

Pseudoglucal 'I

P- -GOR

RORO

RO OR

OR

-

Type C : " exo Glucal "

Type D : " exo - Fructal "

with R = Ac, Bz, Bn, Alkyl, (iopropylidene), H

and

Type E : " ex0 - Ribene " R' = Ac, Bz, Me, Et

Fig. I . General structures of unsaturated saccharides as monomers for polymerization.

pyranoid or furanoid structure of the saccharide derivatives. Some examples of the general pyranoid and furanoid structure of modified saccharide monomers are presented in Fig. 1. The sugar monomers of type A and B, which have a double bound in 'endo' position, are generally called 'glucal' and 'pseudoglucal'. For the sake of simplicity, C D and E types with double bound in 'exo' position will be called 'exo-glucal', 'exo-fructal' and 'exo-ribene'. The A, C, D and E Type of monomers are classified as 'pyranoid or furanoid vinyl ethers'. Monomers of the type B are called 'pyranoid ally1 ethers'. Polyvinylsaccharide structures are generally composed of a C-C polymer chain with side groups of saccharides. In the saccharide polymers, monosaccharides are inserted in the polymeric chain. The difference between polyvinylsaccharides and saccharide polymers is shown in Fig. 2. Concerning the novel synthetic polymers from unsaturated sugars, a few papers involving the copolymerization of glucal-derivatives (monomer type A) and maleic anhydride have been publishedl l-14. Alternating copolymers have been obtained; all are oligomers (Mw < 6,000)

Synthesis of new 'Saccharide polymers'

157

Polyvinylsaccharide with pendent sugar residues

Saccharide copolymer with sugar monomer type A or type 6

Sac

Sac

Saccharide copolymers with sugar monomer type C, D or E

Fig. 2: Structure comparison between poly vinylsaccharide and saccharide polymers.

Synthesis of unsaturated sugars Mono- and disaccharides such as glucose, fructose and leucrose have been used as source because of their low cost and large-scale production. In the series of unsaturated monosaccharides, water soluble monomers (with R = H) and water-insoluble saccharide derivatives have been prepared by protecting the hydroxyl functions either with acetyl-, benzoyl-, benzyl- and methyl groups or with a classical protecting group commonly used in organic chemistry, i.e. isopropylidene or isobenzylidene. Most of the saccharide monomers used in our work have already been synthesized and published14-24. Some investigations were undertaken to improve the over-all yield of the monomers by modifying the product isolation (through column chromatography) and purification

158

K. Buchholz et al.

procedures by an appropriate choice of the recrystallization solvent. Another aim was to reduce the synthesis steps as much as possible (see for example the synthesis of ‘exo-ribene’). All of the unsaturated monosaccharide derivatives have also been synthesized in our laboratory on a preparatory scale (50-80 g) in the same way. The conversion of D-glucose into pseudoglucal derivativesIs--” such as I-methoxy4,6-di-O-acetyl-pseudoglucal (l),1,4,6-tri-O-acetyl-pseudoglucal(2) and 1-ethoxy-4,6-di-Oacetyl-pseudoglucal (3) was amply demonstrated. De-acetylation of the latter results in a watersoluble monomer, 1-ethoxy-pseudoglucal(4).The chemical formulae are presented in Fig. 3.

/OAC

AcO

Me

F

OOEt

AcO

AcOO

OAc

b..

HO

Fig.3 . Chemical formulae of pseudoglucal-derivatives(monomer type B).

According to published procedures and starting from glucose, the following ‘exo-glucal’ monomers (type C) such as 1,2,3,4-tetra-~-acetyl-6-deoxy-~-D-xylo-hex-5-enopy~dnos~de (5), 1,2,3,4- tetra-O-benzoyl-6-deoxy-~-D-xylo-hex-5-enopyranoside ( 6 )and 1,2,3,4-tetra-0-methyl6-deoxy-a-D-xylo-hex-5-enopyranoside (7)have been synthesizedIx-2O.The chemical structures are given in Fig. 4.

Synthesis of new 'Saccharide polymers'

159

Qr AcO

I

OAc

&yBz

BZO

OBZ

Fig. 4.Chemical formulae of exo-glucal derivatives (monomer type C).

The 3,4,5-tri-O-benzoyl-2-deoxy-D-fructo-hexI-enopyranose (8) (Fig. 5 ) , monomer (type D), called 'exo-fructal', was synthesized in three steps starting from D-fructose, in a high over-all yield ( 8 3 % ) as described by F.W. Lichtenthaler2'.

(8) Fig. 5 . Structure of 3,4.5-tri-O-henzoyl-'exo-fructal' (monomer type D).

160

K. Buchholz et al.

Methyl 5-deoxy-2,3-0-isopropylidene-~-D-erythro-pent-4-eno-furanoside (Fig. 6), shortly called 'exo-ribene' (monomer type E), has also been synthesized in three steps from ribose, with a very high over-all yield (90%)22-25. This sugar monomer was previously synthesized in two more steps by substituting the 5-p-toluenesulfonyl derivative with NaI in acetone in a sealed tube at 100 "C;then the elimination reaction occurred in pyridine with an heavy metal (e.g. anhydrous silver fluoride).

a°CR

TsO-H~C n

muOK I USuOH

0 0

X

0 0

X -

( exo ribene )

Fig. 6 . Synthesis of 'exo-ribene'.

Synthesis of new Saccharide Polymers The homopolymerization of the unsaturated monosaccharides and their copolymerization with commercially-available comonomers have been investigated. The monosaccharide monomers were polymerized by means of radical reactions, in bulk and/or in solution. For the water-insoluble saccharide monomers, the solution polymerization took place in neutral organic solvent; for the unprotected sugar monomers, the chain reaction was performed in water. All of the monosaccharide monomers were freshly purified just before use. The crystal monomers (3), ( 6 ) and (8) were dried again at 60 "C under vacuum for 3 hours to weight constant. For the oily exo-ribene no further purification took place. All commercial, stabilized comonomers were also freshly purified from radical inhibitors by vacuum destillation before use. Hydrophobic as well as hydrophilic commercial comonomers were selected and tested for the copolymerization with the protected monosaccharide monomers. Some examples are presented in Tables 1 through 4.

Synthesis of new 'Saccharide polymers'

161

The organic solvents (toluene, di-tert-butylbenzene) were dried, destilled, de-gassed and filtered just before use. Depending on the reaction temperature, the bulk and/or solution polymerizations were achieved using appropriate radical initiators such as azobisisobutyronitrile (AIBN), dibenzoylperoxide (BOP) or di-tert-butylperoxide (DTBP). A general procedure for the solution polymerization is described: the crystal monomers were dissolved in an appropriate solvent, the comonomer and initiator were added at room temperature. Then the reaction mixture was heated for 24 or 48 hours. The polymerizations were carried out under vacuum or under argon atmosphere. The copolymerization reactions of exo-ribene were carried out in bulk under vacuum, in sealed glass tubes using the well-known 'freeze-thaw' procedure. The polymerization reactions were terminated by air supply and cooling. Pure copolymers were isolated by repeated precipitations in a large amount of a non-solvent at room temperature and then dried overnight at 60 "C under vacuum. White, powdery substances were usually obtained. Yields of the polymerization reactions were in the range of 19 to 63 Wt%. The polymer compositions (the ratio between saccharide and comonomer) were determined on the basis of elemental analysis and from 1H-NMR. The weight average molecular weights of saccharide copolymers were in the range from 4,500 to 106,000 under various polymerization conditions. The weight average molecular weights (Mw) of polymers were determined by Gel Permeation Chromatography (GPC) calibrated with polystyrene standard curve. Some copolymers were analyzed with GPC coupled with a laser lightscattering detector as absolute method for determining Mw. The results were in agreement with the GPC analysis. A typical example of solution copolymerization of unsaturated saccharide monomers (type (B)) with hydrophilic comonomer such as vinylpyrrolidone (VPy), as well as hydrophobic styrene (St) and methyl methycrylate (MMA), is summarized in Table 1. Polymers with high molecular weight (e.g. >20 000) were obtained with VPy and MMA. The first copolymer contained 15 mol-% monosaccharide and the second 13 mol-%. With the high reactive styrene, the copolymer contained only 5 mol-% sugar. The random saccharide copolymer is schematically presented in Fig. 7.

K . Buchholz et al.

162

Table 1. Copolymerization of 1-ethoxy-4,6-diacetyl-pseudoglucal(3) and comonomers ( M ) under following reaction conditions: Solvent: Toluene; Initiator: AIBN ( I %); Temperature: 60 "C;Monomers ratio in feed:(3)=(M): Reaction time: 48 hours.

(MI

Yield (Wt%)

Mw GPC

Copolymer compositions) (3):W) (mol%)

VPY

35

78,000

15:85

(75,000b))

MMA

st ~

20

32,000

13:87

19

17.000

5:95

~~

a)determined on the basis of elemental analysis. b)GPC-multiangle laser light-scattering detector (Wyatt technology).

Radical copolymerization of benzoylated exo-glucal (monomer type C) with various commercial comonomers such as maleic anhydride (MAh), methacrylate (MA) and methyl methacrylate (MMA) were also carried out; the results are summarized in Table 2. With equimolar amount of monomers in feed, saccharide copolymers have been obtained with varying sugar content. The polymer compositions show that only MAh copolymerized alternatively with exo-glucal (the ratio between saccharide and MAh was found to be 50:50).The alternation might be caused by some attractive interaction between both monomers. This result suggests that the alternating copolymer structure can be schematized as shown in Fig. 8. R'OOC ROOC

>

R'OOC ROOC

,,,,,,,1.1..

R'OOC.

-

ROOC

k

RO

OR

WOOC,

>

ROOC

>

ROOC

I

Fig. 7. Schematic representation of a random saccharide copolymer containing a pseudoglucal derivative.

Synthesis of new 'Saccharide polymers'

163

Table 2: Solution copolymerization of benzoylated exo-glucal (6) and cornonomers (M) under the following reaction conditions: Solvent: Toluene; Initiator: BOP( 1%); Temperature: 80 'C ; Monomer ratio in feed:(6)=(M); Reaction time: 48 hours.

(M)

Yield

Mw

Wt%)

GPC

Copolymer compositiona)

(6):(M) (mol%)

MAh

41

47,600

50:50

MA

63

1 1,000

19:81

MMA

32

12,500

35:65

a) determined on the basis of elemental analysis.

Surprisingly, benzoylated exo-fructal (monomer type D) has successfully been homopolymerized. The homopolymer exhibits a high molecular weight (Table 3). No ring opening

0

I OR

MAh (Electron-Acceptor )

exo-Glucal derivative (Electron-Donor)

I

initiator

L

Fig. 8. Alternating saccharide copolymer.

J m

with R = Ac, 02, Me

164

K. Buchholz et al.

had been detected by IH- and I3C-NMR analysis of the homopolymer. Copolymerization with MAh and MMA were tested. As expected, the copolymerization with MAh also leads to an alternating copolymer with a molecular weight of about 11,000. Table 3. Polymerization of henzoylated exo-fructal (8) and coinnnomers (M). Reaction conditions: Solvent: ter-1.-hutylhenzene;Initiator: DTBP(I 5%); Monomer ratio in feed:(8)=(M): Reaction time: 48 hours.

(MI

T("C)

Yield (Wt%)

Mw GPC

polymer compositiona) (8):(M) (mol%)

120

30

6 1,000

100%(of 8)

MAh

120

59

10,600

5050

MMA

80

65

4,500

955

a) determined on the hasis of elemental analysis.

It is to be expected that 'exo-ribene', monomer type E, such as furanoid vinyl ether will show different behavior because of its furanoid structure. Copolymers with the same comonomers, as used for the other unsaturated monosaccharides, were synthesized. Because exoribene is liquid and thermally stable (no degradation has been observed up to 130 "C), the radical polymerization reaction was achieved in bulk under strong mechanical stirring. Polymerization conditions and results are presented in Table 4. Table 4: Copolymerizafion of 'exo-rihene' and cnmonomers (M) under the following reacfinn conditions: Solvent: none; Initiator: DTBP (1%); Temperature: 120 "C; Monomers ratio in feed: (exo-rihene)=(M);Reaction time: 6 hours.

(M)

Yield (Wt%)

Mw (GPC)

Copolymer compositiona) exo-ribene:(M) (mol%)

VPY

34

106,000

33:67

MMA

54

52,000

2575

( 84,O0Oh))

MAh

28

29.000

Determined from elemental analysis. h, GPC on-line with a inultiangle laser light-scattering detector (MALLS).

a)

5050

Synthesis of new 'Saccharide polymers'

165

High molecular weight products have been obtained especially by using VPy and MMA. The mole fraction of sugar in the first two copolymers is higher than 20%. Alternating copolymers were obtained with MAh. It can be assumed that in this case as well, the radical polymerization takes place under a charge transfer complex mechanism. Although the polymerization temperature is relatively high (120 "C) no side reactions have been observed, checked by the spectroscopic analysis of the saccharide polymers. The FT-IR-spectrum of (exo-ribene:VPy) copolymer was compared with those of polyvinylpyrrolidone and (exo-ribene) homopolymer as references (Fig. 9). Analysis of these polymers shows the characteristic bands of sugar and of comonomer. The intense band at 1676 cm-' characteristic for the stretching vibration of the carbonyl group of the N-substituted pyrrolidone is present in both spectra b) and c). The carbon double bond stretching vibration of the vinyl group of the sugar monomer is not found at 1670 cm-1 in the copolymer nor in the homopolymer. The deformation mode at 1424 cm-1 of the pyrrolidone ring of the comonomer is present in the copolymer spectrum. The C-N stretching vibration band appears in the copolymer at 1289 cm-1. The absorption at 1105 cm-1 which is typical for cyclic ether stretching vibration (furanoid ring of sugar) is present in the copolymer. The band at 1375 cm-1 of the copolymer is related to cyclic ether of sugar (out of plane deformation vibration).

I

1700

I

1600

I

I

I

1500 1400 IjOO Wavenumber (em-1)

I

1200

Fig. 9. m-IR spectra of exo-ribene homopolymera), (exo-ribene:VPy) copolymerb) and polyvinylpyrrolidonec).

I

I100

166

K . Buchlzolz et al.

The '3C-NMR spectrum of the (exo-ribene:VPy) copolymer confirms also the presence of the sugar moiety in the copolymer (Fig. 10). The chemical shift of the 'exo-ribene' and the comonomer are indicated as follows:

6 (CDC13, ppm)= 18 (C-4', VPy), 25-27 (C-8,9 of isopropylidene protecting group of sugar), 31 (C-3', VPy), 40-46 (C-7'of VPY and C-5 of sugar), 55 (C-6, OCH3 of sugar), 78 (C-3, sugar), 82 (C-2, sugar), 83-86 (C-6',VPy), 86-90 (C-4 of sugar), 108 (C-1 of sugar), 112 (C-7 of isopropylidene protecting group of sugar), 175 (C-2', carbonyl group of VPy).

I

I

I

10U

160

146

12.

PPb

10.

I

I

80

6r

I 41

4..

Fig. 10. 13C-NMR spectrum of (exo4bene:VPy) copolymer in CDCll solution.

Conclusion It has been shown that unsaturated sugar derivatives polymerized well with several kinds of commercial comonomers under radical conditions.

Both pyranoid and furanoid sugar vinyl ethers were successfully copolymerized and polymers of high molecular weight were obtained in several cases. Water soluble polymers can be prepared by hydrolysis of protected saccharide polymers and in some cases of comonomers, for example with maleic anhydride. The position of the sugar in the main chain provides for a new type of polymers. The characterization and properties of the 'saccharide polymers' will be further investigated.

Synthesis of new 'Saccharide polymers'

167

References I.

2.

R.L. Whistler and J. L. Goatley: Copolymerisation of I -acrylamido-1-deoxy-D-glucitol and of I-deoxy-lmethacrylamido-D-glucitol with various vinyl monomers. J. Pol. Sci. I (1961) 127-132. W.A.P. Black, E.T. Dewar, and D. Rutherford: Polymerization of unsaturated derivatives of 1,2:5,6-di-0-

3.

isopropylidene-D-glucofuranose. J. Cheni. SOC.(1963) 4433-4439. T. P. Bird, W.A.P. Black, J.A. Colquhoun, E.T. Dewar, and D. Rutherford: Preparation and derivatives of

4.

5. 6.

7. 8. 9. 10.

11. 12. 13. 14.

15.

16. 17. 18. 19.

poly-(6-0-methacryloyl-D-galactose)and poly-(6-0-acryloyl-D-galactose).J. Chem. Soc.(Cj ( 1966) 1913-1918. K. Kobayashi and H. Sumitomo: Speciality polymers having sugar as the pendant group: synthesis, characterization, and binding of organic solute in water. Macromolecules 13 (1981) 234-239. J. Klein, D. Herzog, A. Haji Begli: Poly(viny1 saccharide)s, I : Emulsion polymerization of poly(methacryloylglucose). Makromol. Chem Rap. Comm. 6 (1985) 675-678. J. Klein, K. Blumenberg: Synthesis and polymerization of 6-0-methylallyl-galactosederivatives. Makromol. Chem. 189 (1988) 805-813. J. Klein, M. Kunz, J. Kowalczyk: New surfactant polymers based on carbohydrates. Makromol. Chem. 191 (1990) 517-528. K. Nakamae, T. Miyata, N. Ootsuki: Surface studies on copolymers having pendant monosaccharides. Macromol. Cheni. Phys. 195 (1994) 2663-2675. G. Wulff and G. Clarkson: New type of polyvinylsaccharides with N,N-dimethylbarbituric acid as a linker between sugar and styrene residue. Macromol. Chem. Phjs. 195 (1994) 2603-2610. K. Buchholz, E. Yaacoub, S. Warn, B. Skeries, S. Wick, M. Boker: Polymerisate aus ungesattigten Sacchariden und deren Derivaten sowie deren Copolymerisate mit ethylenisch ungesattigten Verbindungen und Verfahren zu ihrer Herstellung. Ger. Pat. Appl. 44 08 391 (1995). M. J. Han, K. B. Choi, K. H. Kim and B. S. Hahn: Biologically active polymer-targeting polymeric antitumor agents. Makromol. chem.. Macroniol. Symp. 33 (1990) 301-309. M. J. Han, C. W. Lee, K. H. Kim and W. Y. Lee: Synthesis and biological activity of poly((tri-0-acetyl-Dgluca1)-alt-(maleic anhydride)) derivatives. Bull. Korean Clienz. Soc. 12 (1991) 85-87. Y. Koyama, M. Kawata and K. Kurita: Polymerization of unsaturated sugar I. Radical copolymerization of D-glucal derivatives and maleic anhydride. Polymer J. 19 (1987) 687-693. Y. Koyama, M. Kawata and K. Kurita: Polymerization of unsaturated sugar II. Radical copolymerization of a 1-enofuranose. Polymer J . 19 furanoid glucal, 3-0-benzyl- 1,2-dideoxy-5,6-0-isopropylidene-D-arabino-hex(1987) 695-700. R.J. Ferrier and N. Prasad: Unsaturated carbohydrates. Part IX. Synthesis of 2,3-dideoxy-ol-D-erythro-hex-2eno-pyranosides from tri-0-acetyl-D-glucal. J . Chem. Soc. (Cj (1969) 570. J. Habus and V. Sunjic: Preparation of the chiral diol(2R. 3R)-2-hydroxymethyl-3-hydroxy-tetahydropyran from D-glucose via reductive rearrangement of pseudo-D-glucal triacetate. Croar. Cheni. Acta 58 (1985) 321. G.Descotes and J. C. Martin: Sur I'isonierisation du I ,5 anhydro-3,4,6-1ri-O-henzyl-l,2-didesoxy-D-arabinohex-I-enitol en presence d'acides de Lewis. Carbohydr. Res. 56 (1977) 168. R. Blattner and R. J. Ferrier: Photobromination of carbohydrate derivatives. Part 2. J. Chem. Soc. Perkin Trans. I (1980) 1523. R. Blattner, R. J. Ferrier and P. C. Tyller: Unsaturated carbohydrates. Part 22. Alkenes from 5-bromo hexopyranose derivatives. J . Chem. Soc. Perkin Trans. I(1980) 1535.

168 20. 2I.

22. 23. 24. 25.

K. Buchholz et al. D. Semeria. M. Philippe, J. M. Delaumeny, A-M. Sepulchre and S. D. Gero: A general synthesis of cyclitols and aminocyclitols from carbohydrates. Synthesis (1983) 710. F.W. Lichtenthaler: Enantiopure building block from sugars and their utilization in natural product synthesis. Modern Synthetic Methods, Ed.: R. Scheffold, VCH Publ., 6 (1992) 348. L. Hough and B. Otter: Furanoid vinyl ethers. Chenr. Cornm. (1966) 173. P. A. Levene and E.T. Stiller: Acetone derivatives of D-ribose. J . Eiol. Chem. 104 (1933) 187. P. A. Levene and E.T. Stiller: Acetones derivatives of D-rihose. 11. J . Biol. Chenr. 106 (1934) 421. J. J. Rabelo and T. van Es: Derivatives of 5-seleno-D-ribose. Curbohydr. Rex 30 (1973) 381.

10 Molecular inclusion within polymeric carbohydrate matrices

S. Kubik, 0. Holler, A. Steinert, M. Tolksdorf, Y. Van der Leek, and G. Wulff

Institut fur Organische Chemie und Makromolekulare Chemie, Heinrich-Heine-Universitat, Universitatsstr. 1, D-40225 Dusseldorf

Summary. The possible industrial utilization of amylose for the encapsulation of organic guest molecules was investigated. By introducing different substituents into the amylose chain, soluble amylose complexes with a variety of guest molecules could be ohtained. The dependence of the type of substituent and its degree of substitution on the complexing behavior of amylose was systematically studied. In the case of slightly hydroxypropylated amylose, methods such as c.d. spectroscopy and microcalorimetry were used for the characterization of the complexes. Our results give new indications about the structure of amylose complexes in solution and the mechanism of their formation. It was concluded that amylose can adjust its helix conformation to the shape and structure of the guest molecule included, i.e. the selectivity of complex formation of amylose is relatively low. However, this selectivity can be manipulated by intra- or inter-molecular cross-linking of the amylose chains. Results on the complexation of synthetic polymers with amylose are also presented.

Introduction The macroscopic properties of polysaccharides are important criteria for their industrial utilization. These properties depend on the polysaccharide type, e.g. cellulose or starch, and also on its origin. For example: starches from different plants often show significant differences in their gelatinization behavior'. This is partly a result of the ratio of amylose and amylopectin - the two starch components - in the granule. Other factors that influence the properties of starch are the fine structure of the starch components and their arrangement in the granules, the size of the granules, and the presence of traces of non-carbohydrate compounds such as lipids, proteins, and phosphates. The given properties of these natural products can be further manipulated by chemical modifications. For example: a slight cross-linking of the polysaccharide chains in the starch granule strongly influences their gelatinization, and hence the viscosity of a starch suspension. There are various other ways for the chemical conversion of starch, e.g. hydrolysis, oxidation, reduction, modification of the hydroxyl groups, etc. and a large number of different products can be derived from this renewable resource. The resulting products are especially interesting when they can compete with products from petrochemical processes.

170

S. Kubik et al.

The properties of a material based on starch are determined mainly by the structure and by possible inter- or intra-molecular interactions of the polysaccharide chains as well as the presence of additional non-carbohydrate compounds'. But the properties of these additives are also influenced by the starch matrix; e.g. it has been known for some time that the extraction of lipids from native starch is relatively difficult2. With some solvents such as chloroform, an extraction in not possible at all, whereas methanol is relatively well-suited. These results were interpreted in terms of a special interaction of the lipid molecules with the polysaccharide chains. In fact, today we know that lipids form stable complexes with the amylose fraction of starch3 and complexes between lipids and amylopectin are also discussed4.

Inclusion complexes of amylose and cyclodextrins Amylose is a linear polysaccharide composed of 1,4-a-linked glucose units. As a result of this structure, amylose is able to adopt a helical conformation with the hydroxyl groups of the glucose units arranged on the surface of the helix. In contrast to this polar surface, the cavity of the helix is relatively non-polar. The closely related amylopectin is a highly-branched polysaccharide with short amylose chains linked by 1,6-a-glycosidic bonds. Amylose crystallizes in four different modifications: A, B, C, and V. The A, B, and C modifications consist of double-stranded amylose helices with different amounts of water present in the crystal lattice. In contrast, in the V modification the amylose chains have a single helical conformation.' This modification is formed when amylose crystallizes in the presence of organic compounds such as fatty acids, long chain alcohols, etc. These compounds are included into the non-polar cavity of the amylose helix and are often very tightly bound. The helix conformations of A- and V-amylose are presented in Fig. 1. These pictures resulted from computer simulations of the molecular structures of these amylose modifications.4 A MOLCAD based generation of the molecular lipophilicity potential (MLP) of the helices shows their hydrophilic (blue) and hydrophobic (yellow) surface areas. The half-opened model of V-amylose also reveals its hydrophobic center channel. A variety of different guest molecules can be included into the amylose helix.3~5Electron diffraction investigations of solid amylose complexes showed that, depending on the size of the guest molecule included, amylose forms helices with 6 , 7, or 8 glucose units per turn.6 In all cases, the pitch of the helices is about 8A but the inner diameter varies from 9.0 to 12.5A (Table 1). The amylose chain is obviously relatively flexible. These amylose complexes are closely related to the inclusion complexes formed by cyclodextrins. Cyclodextrins are cyclic maltooligosaccharides. The more common derivatives have 6 (a),7 (p), and 8 (y) glucose units in the ring (Fig. 2).

Molecular inclusion within polymeric ccirbohydrate matrices

Fig. I .Computer-generated representations of the conformations of V-amylose (upper) and A-arnylose (below) helices.

1I 1

172

S. Kubik et al. Table I , Structure of various crystalline arnylose complexes as determined by electron diffraction.T guest

glucose units per helix turn

outer cavity diameter [A]

inner cavity diameter [A]

n-butanol, Iz, fatty-acids

6

13.2

9.0

tert-butanol

7

14.7

10.5

a - n a p hthol

8

16.2

12.5

As in amylose, the cavity of the cyclodextrins is non-polar and organic guest molecules can

be included. However, the conformation of the cyclodextrins is much more rigid and larger guests molecules can often not be included into the smallest a-cyclodextrin. In recent years, cyclodextrins have gained considerable attention in the field of host-guest chemistry or supramolecular chemistry.7 Cyclodextrins have been used for the complexation of a variety of compounds. In many cases, the structure and properties of these complexes have been investigated by means of spectroscopic, crystallographic, or computational methods. Cyclodextrins themselves, and especially some chemically modified derivatives, have served as powerful enzyme models. The cross-linking of cyclodextrins with various reagents leads to new polymeric materials which are useful as stationary phases for chromatography. Their separation properties can be explained in terms of a combination of common Sephadex gels with the specific inclusion phenomena of cyclodextrins. In this context, it is interesting to note that a Japanese group recently succeeded in the synthesis of a linear polymeric cyclodextrin derivative with the cyclodextrin rings arranged like

A

6H

---\

CH~OH

Fig. 2. Structure o f a-cyclodextrin

Molecular iticlirsioti within polymeric carbohydrate matrices

11 3

pearls on a 'molecular necklace' (Fig. 3).8 These nanotubes may have interesting properties for the transport of organic molecules through membranes. Their structure is somewhat similar to an amylose helix.

3 OH

OH

OH

OH

OH

OH

Fig. 3 . Schematic representation of cyclodextrin polymers and a nanotube built up from cyclodextrins

However, cyclodextrins are not only interesting from the scientific point of view. Their inclusion properties also make them interesting for industrial applications7 and the number of patents dealing with potential industrial uses of cyclodextrins and cyclodextrin polymers is steadily increasing. A variety of compounds can be encapsulated by inclusion into a cyclodextrin ring. This often changes their properties significantly: -

Their stability against light, oxygen, or reaction with other compounds is increased.

-

The storage and handling of instable or toxic compounds is much easier when they are complexed with cyclodextrins.

-

-

Volatile compounds can be bound with cyclodextrins. The activity of drugs or other bioactive compounds, such as herbicides, pesticides, or fungicides, is increased because of a molecular dispersion of these compounds in the complexed state. They may also be active for a longer period of time because they are only slowly released from the complexes.

-

This slow release is also advantageous for the encapsulation of flavors and aromas.

-

The solubility of non-polar compounds in water can be increased significantly.

-

It is possible to remove some compounds selectively from mixtures, e.g. cholesterol from eggs, or nicotine from cigarette smoke, by complexation with cyclodextrins.

For a long time, the high price of cyclodextrins prevented the utilization of their complexes in industry. Only recently has the price of P-cyclodextrin permit a wider industrial use. Today P-cyclodextrin is mainly used in drugs and cosmetics. In some countries its utilization in the food industry is also allowed, e.g. in the United States for the separation of cholesterol from yolk. However, the prices for a-and y-cyclodextrins are still relatively high and their uses are restricted

174

S. Kubik et al.

to specialty chemicals. It may be speculated that the considerably less-expensive amylose or even starch, which are also able to form inclusion complexes with organic molecules, can replace the cyclodextrins in some of their applications. The following discussion outlines various possible approaches.

Microencapsulation and molecular encapsulation with starch A group at the University of Marburg is investigating the microencapsulation of various types of

solids with starch.9 Fine suspensions of drugs or other biorelevant compounds, in some cases even pollen, charcoal, metal oxides, or glass, are coated with a thin layer of retrograded amylose (Fig. 4).

retrogradation

Fig. 4. Schematic representation of the microencapsulation of solids with starch.

Because of this treatment the resulting products can easily be dispersed in water. These products may have promising pharmaceutical applications because it may be possible to aid the resorption of drugs by coating them with starch. However, after degradation of the starch layer, an organism still has to cope with relatively large clusters of molecules, hence their activity may not be optimal. In contrast, the molecular encapsulation of drugs in cyclodextrins seems to be more advantageous in this respect. It was our goal to transfer the principle of molecular encapsulation to the inclusion complexes of amylose. In contrast to cyclodextrin complexes, at the beginning of our work, little was known about the structure and behavior of amylose complexes with organic guest molecules. This may be due to the low solubility of most of the amylose complexes. In fact, originally, complex formation was often only detected by the formation of a precipitate after mixing an amylose solution with a solution of an organic compound.5 Only the soluble amylose-iodine complex had been studied in detail.10

Molecular inclusion within polymeric carbohydrate matrices

175

However, some information about their structure can even be obtained from the insoluble amylose complexes. After their isolation the amount of guest molecules included into the helix can be determined quantitatively by extraction with methanol. From the ratio of the amount of amylose and the amount of guest one can calculate how many glucose units are necessary for the complexation of certain guest molecules. Some results are summarized in Table 2 . Tahle 2. Stoichiornetry of arnylose complexes with various guest molecules.

guest : glucose

amount guest [g] :

guest

units

amount amylose [g]

fenchone

1:6

1 : 6.4

indanone

1 :6

1 : 7.4

vitamin-A acetate

1 : 20

1 : 9.9

a-naphthol

1 :8

1 : 9.0

P-naphthol

1 :8

1 : 9.0

p-ionone

1 : 38

1 : 32.0

In general, the presented values may be underestimated and slightly more glucose units could be required for the complexation of these guest molecules. This is a result of the fact that guests which are only adsorbed on the surface of the amylose helix are also detected by this method. Still, these results show interesting tendencies: guest molecules that are known to form stable complexes with amylose can occupy almost the entire space inside the helix. Only about 7 to 10 glucose units of the amylose chains are required for the complexation of one guest molecule. In this respect, the amylose has no disadvantage compared to cyclodextrins. Almost the same amount of amylose or cyclodextrin is necessary for the complexation of a certain amount of guest. It is reasonable to assume that more glucose units are necessary for the complexation of the large vitamin-A acetate molecule. Compounds such as P-ionone form relatively instable complexes with amylose and only few molecules are complexed along the chain.

In this context, it is interesting to note that relatively large molecules such as vitamin-A can also be included into the amylose helix. We are currently investigating the complexation of other biorelevant compounds. It is expected that the amylose is able to form complexes with a variety of interesting molecules. It is not possible to derive any information about the structure of amylose complexes from the results of this method alone. Because of the good water solubility of most of the cyclodextrin complexes, common spectroscopic methods such as u.v., n.m.r., or circular dichroism (c.d.) have been used for their characterization, and a considerable amount of information is available about

176

S. Kubik et al.

their structure.7 W e therefore tried to modify the structure of amylose in a way that it is able to form soluble complexes, and to employ the same methods for their characterization which have been used for the cyclodextrin complexes before.

Modification of amylose The solubility of amylose can be increased by modification of the hydroxyl groups of the glucose units. Sometimes only a small degree of substitution (DS) is sufficient to significantly increase the water solubility. The question is how this modification influences the complexing behavior of amylose. To study this, we systematically introduced five different substituents into the arnylose helix, varied their DS, and determined the complexing ability and solubility of the products. The results are presented in Table 3. Table 3 . Cornplexing ability and solubility of potato arnyloses with various residues and degrees of substitution (DS) (n.d. = not determined).

BV

Llax

13.4

[nml 640

solubility (5% aqueous solution) not stable

5.6 0.2

595 n.d.

warm cold, 2h

1:8 1:5

11.4 6.8 3.4 0.4 0.1

620 600 570 n.d n.d.

warm warm warm cold, 2h cold, 2h

0.06 0.20 0.30

1 : 50 1 : 15 1 : 10

11.2 4.6 0.9

620 590 560

warm warm cold, l h

-C(O)-CH3

0.06 0.16 0.26 0.43

1 :SO 1 : 19 1 : 12 1:7

8.9 4.0 0.9 0

600 585 540 n.d.

warm warm warm cold, 2h

-CH;?-COOH

0.09 0.20 0.28 0.67

1 : 33 I : 15 1 : 11

8.9 2.2 0.4 0

620 570 530 n.d.

warm warm cold, Ih cold, 0.5h

R

DS

-H

0

-CHzCH;?OH

0.15 0.50

1 :20

0.06 0.13 0.2 1 0.38 0.56

1 : 50 1 : 23 1 : 14

-CH2CHOH I CH3

-CH 2FHOH

CH 20H

Wglucose units 0

1:6

I :4

Moleciilar inchuion within polynieric carbohydrate matrices

177

The DS is the number of residues per glucose unit. The blue value (BV) is a measure for the iodine binding capacity of the amylose: the lower the BV, the lower the ability of amylose to complex iodine.

A,,

is the wavelength of the

U.V.

absorbance of the blue amylose iodine

complex. It is proportional to the length of iodine chains in the amylose helix;" the lower the shorter the iodine chains. In general, BV and La,show the same tendency.

A,,,,

Hydroxyethyl, 2-hydroxypropyl, and 2,3-dihydroxypropyl groups have been introduced by reaction of the amylose with the corresponding epoxides in aqueous 1 N NaOH. Acetylation was achieved by reaction with acetic anhydride and the reaction with chloroacetic acid gave

,

carboxymethylated amylose (Fig. 5).

0 / \ HZC-CHZ

Am-0-CH2-CH2-OH

qH3 Am-0-CH2-CH-CH20H

CHj-CH-CHz

Am-OH

yH2OH Am-0-CH2-CH-CH20H

CI-CHZ-COOH

>

Am-0-CH2-COOH

Fig. 5 . Modification of amylose.

In general, the water solubility of all products increased as expected with increasing DS. But the ability to bind iodine decreased in the same direction. In this respect, the acetylated and carboxymethylated products showed a slightly faster decrease of their BV with increasing DS. No significant differences of the dependence of their BV on the DS could be observed for the three hydroxypropylated products. These amyloses still possess a good iodine binding capacity when the DS is not too large. After addition of organic guest molecules to aqueous solutions of these amylose derivatives, however, no precipitates could be observed for a DS > 0.06, although with native amylose usually insoluble complexes are formed under the same conditions. This either

178

S. Kubik et al.

shows that hydroxypropylated amyloses completely loose their ability to form inclusion complexes or that their complexes are indeed water soluble. This question was answered during the following c.d-spectroscopic investigations. Hydroxypropylated amylose with a DS of approximately 0.8 was used for the measurements because of the ease of synthesis and the reproducibility of the DS.

C.d.-spectroscopic investigations C.d.-spectroscopy has proven to be a useful method for the investigation of cyclodextrin complexes. When an achiral guest molecule having a suitable chromophore is included into the chiral cavity of a cyclodextrin, one can usually observe an induced Cotton-effect.I2 Hence, a Cotton-effect definitely proves the existence of a complex because neither the uncomplexed guest nor the empty cyclodextrin themselves exhibit Cotton-effects. Moreover, the geometry with which a guest is included into the cyclodextrin can be deduced from the sign of the Cotton-effect.l3 It is reasonable to assume that if a suitable guest is included into an amylose helix, one should also be able to observe induced Cotton-effects. In fact, a negative Cotton-effect was observed after addition of 2-hexanone to a solution of hydroxyproylated amylose (Fig. 6). This proves that the hydroxypropylated amylose is indeed able to form soluble complexes and that c.d-spectroscopy can be used for their investigation.

205

300

350

X/nm

-5

205

350

300

X/nm

Fig. 6. C.d.-spectra of amylose (a), a-(b), p- (c), and y-cyclodextrin (d) complexes with 2-hexanone (c(amy1ose) = 20 mglmL H20; c(a-, p-, and y-cyclodextrin) = 5 mg/mL; pH 7; c(2-hexanone) = 9.0 . lo-*mol . L-I).

Moleciilar inclusion within polymeric carbohydrate matrices

179

The 2-hexanone complex with or-cyclodextrin also has a negative Cotton-effect, whereas those of p- and y-cyclodextrin have positive ones. From this it can be concluded that 2-hexanone is included into the amylose with a comparable geometry as in a-cyclodextrin. This, however, is only possible if the size of the amylose helix is of the same order as that of a-cyclodextrin. In other words, the amylose must have a conformation with six glucose units per turn. Hence, by comparing the c.d-spectra of the amylose complexes with those of the corresponding cyclodextrin complexes, the conformation of the amylose helix in the complex can be determined. When aromatic compounds such as 4-terr-butylphenol are used as guests, the c.d-spectrum of the amylose complex is analogous to that of the corresponding

p- (and y-)

cyclodextrin

complexes (Fig. 7).

-2

L

- 1 o ! ' " ' " ~ " "

230

300

350

230

X/nm

3 '0

300

X/nm

Fig. 7. C.d.-spectra of amylose (a), a- (h), p- (c), and y-cyclodextrin (d) complexes with 4-!err-butylphenol (c(amy1ose) = 20 mg/mL H,O; c(a-, p-, and y-cyclodextrin) = 5 mg/mL; mol . L-I), pH 7; c(4-terf-hutylphenol) = 1.9 .

From this, it can be concluded that in the complex with tert-butylphenol the amylose helix has a conformation with seven glucose units per turn. The flexible amylose can obviously adjusts its conformation in solution to the structure of the included guest molecule. By the same methodology, the structure of some other arnylose complexes with organic molecules have been detennined.I4 Relatively large molecules such as phenolphthalein also exhibit strong induced Cotton-effects after complexation with amylose. The special structure of this complex has been studied in detail.I4 We even have reason to believe that chlorophyll interacts with the amylose helix. This complex is currently under investigation and results will be published in due course.

180

S. Kubik et al.

C.d.-spectroscopy is obviously not only a powerful method for detecting complex formation but also yields information about the structures of the complexes. The amylose derivatives presented in Table 4 show induced Cotton-effects with typical guest molecules such as 4-methyl-2pentanone and 4-tert-butylphenol. As expected, the absolute intensity of the c.d.-bands decrease with increasing DS (Table 4). Table 4.Dependence of the intensity o f induced Cotton-effects of amylose complexes on the type and degree of substitution (c(amylore) = 20 mg/mL H 2 0 ; pH 7; c(4-inethyl-2-pentanone)= 9.0. lo-* mol L-’; c(4-ferf-hutylphenol)= 1.9 . 10’ mol L-’)

R

DS

-CH2CH20H

0.15 0.50

-CH2CHOH I CH3

0.06 0.2 1 0.56

- 1.45

-2.30

-1.15 -0.85

- 1.40

-CH 1CHOH

0.06 0.20 0.30

- 1S O -1.15 -0.85

-

-C(O)-CH3

0.16 0.43

- 1.59 - 1.49

- I .s9 - I .46

-CH2-COOH

0.09 0.20 0.67

- 1.40

-2.20

-0.60 -0.20

-1

&I

CH 20H

C.d. value [mdeg] 4-methyl-2-pentanone - 1.20 -0.65

C.d. value [mdeg] 4-trrt--butylphenol - 1.50 -0.75

-0.60 -2.05 1.35 -1.00

.so

-0.47

Relatively large Cotton-effects are observed for the acetylated amyloses, even for high values of DS. We are currently investigating whether this amylose may be better suited for c.d.spectroscopy .

Microcalorimetric investigations Another method we employed for the characterization of amylose complexes was microcalorimetry.15 In general, this method makes it possible to measure heats of complex formation.

For our investigations, enzymatically-synthesized amyloses with chain lengths ranging from 9 to 1000 glucose units were used to study the dependence of complex formation on the polymerization degree (P,). From the results, the mechanism of complex formation could be deduced for certain guest molecules.

Molecular inclusion within polymeric ccirbohJldrate rmtrices

18 I

Guest molecules that are included into an amylose helix with six glucose units per turn can even be complexed by amyloses with a P, as low as 9 i n a fast reaction. For the complexation of larger molecules such as 4-terr-butylphenol which require a larger cavity, a conformational reorientation of the amylose chain leads to a helix with seven glucose units per turn. In this case, complex formation is considerably slower and no reaction was observed for amyloses with less than 250 glucose units. Either these amyloses are not able to form complexes with guests which require a larger helix or complex formation is too slow to be followed calorimetrically. With this method, we could demonstrate that amylopectin is also able to form inclusion complexes. As expected, only guest molecules with long alkyl chains such as surfactants are complexed which can be included into the short amylose segments of the amylopectin molecules. Various new results about the structure and behavior of amylose complexes in solution have thus been obtained from our spectroscopic and calorimetric investigations. All investigations were carried out with native or modified potato amylose. Its relatively high price may prohibit the use of pure amylose for industrial applications. However, we recently showed that, i n principle, highamylose starches from peas exhibit analogous properties. As a result of the amylopectin content, their complexing ability is somewhat reduced as compared to pure amylose, but after hydroxypropylation water soluble products can also be obtained which still exhibit relatively good complexing properties. By removing impurities such as proteins and lipids before derivatization, the properties of the products are slightly improved.

Crosslinking of amylose

In our previous investigations, we were able to show that amylose is a rather flexible host molecule that can adjust the inner diameter of its helix to the size and shape of the guest molecules included. In other words, the selectivity i n binding of amylose is low. However, it can be speculated that the complexing behavior of amylose can be manipulated when the conformational interchange between different helix conformations is prevented, e.g. if an amylose helix with six glucose units per turn could be stabilized, it should no longer be able to incorporate molecules that are usually included into a larger helix. In principle, there are two possible ways for the stabilization of amylose helices, both based on cross-linking reactions (Fig. 8). By intramolecular cross-linking between adjacent helix turns, the flexibility of the amylose should be reduced and the helix should no longer be able to change its conformation. However, the fixation of amylose in a three-dimensional network by a combination of inter- and intramolecular cross-linking should also stabilize the helix conformations present during the react ion.

182

S.Kubik et al. 4

Fig. 8. Schematic representation of intramolecularly- and intermolecularly-crow linked amylose helices.

We investigated both aIternatives.'b The intramolecular cross-linking reaction gave a soluble product whereas after the intermolecular reaction, completely insoluble materials were obtained.

For the intramolecular reaction, epichlorohydrin and divinyl sulfone were used as cross-linking reagents. The reactions were carried out in dilute solutions to suppress intermolecular side reactions. Under these conditions, no significant intermolecular reaction was observed since the molecular weights of the amyloses before and after the reaction showed practically no differences. Differences in their complexing abilities, however, resulted from the cross-linking, which depend on whether the reaction was carried out in the presence or absence of guest molecules. While amyloses cross-linked in the absence of guests completely lost their ability to bind iodine, those cross-linked in the presence of one kept their complexing properties to a certain degree. This was interpreted in terms of a stabilization of some helical segments in the amylose chain. Quantitative results about the influence of cross-linking on the complexing behavior of amylose could be obtained from intermolecularly-cross-linked products. In this case, the reactions were carried out in the presence or absence of various guest molecules; Cyanuric chloride and divinyl sulfone were used as cross-linking agents. The selectivity of these materials regarding the complexation of different guests was studied with batch experiments. For this, a certain amount of cross-linked amylose was equilibrated with an aqueous solution of a guest molecule of known concentration. The amount o'f guest bound by the amylose can be detected by measuring the decrease of concentration of the batch solution by means of u.v.-spectroscopy. Some results are summarized in Table 5. Obviously, the selectivity of the cross-linked amyloses can be influenced by the guest molecules present during the cross-linking reaction. At present, however, it is difficult to predict the complexing abilities of the products. For example: it is not easy to understand why 4-tert-

Molecular irichision within polymeric carbohydrate matrices

183

Table 5,Complexation of sodium dodecylphenylsulfonate (DPS) and 4-reri-hutylphenol (BP) with amyloses intermolecularly crosslinked with cyanuric chloride in the presence of various guest molecules.

guest

complexation [mmol guest. lO4/mg amylose DPS

BP

2-octanone

4.27

1.90

cyclohexanone

3.91

1.42

fenchone

4.83

1:19

sodium dodecylsulfate

3.23

0.28

crosslinked without guest

1.42

0.27

butylphenol is bound best by an amylose cross-linked in the presence of 2-octanone. It may be that, for the explanation of the complexing behavior of these materials, the effects of an adsorption of the guest molecules on the surface of the amylose helix must also be taken into account. However, we could demonstrate that the complexing properties of amylose can be manipulated by cross-linking reactions. We are currently trying to systematically vary the cross-linking reagents, reaction conditions, and guest molecules. The rate of release of the complexed molecules from the polysaccharide network is also being studied. These materials may be interesting for the slow release of biorelevant compounds such as vitamins, flavors or pesticides.

Starch and polymers An important application in which starch is already being used in industry is in the production of polymer blends with synthetic polymers to obtain biodegradable thermoplastic materials.” A variety of products that contain different amounts of starch are already available. These are primarily used as packaging materials. In some cases, inclusion complexes between the two different components are discussed to explain their strong interaction and the good mechanical properties of the products. X-ray diffraction experiments did in fact show that the interaction of starch with a copolymer of ethylene and acrylic acid is caused by an incorporation of the copolymer into helical polysaccharide chains.’* In this case, a conformation of the helix with six glucose units per turn analogous to the complexes with fatty acids was found. We recently realized that polymers such as polyethyleneglycol (PEG) and polyvinylalcohol (PVA) form typical insoluble inclusion complexes with amylose. We also tried to determine the structure of these special complexes in solution. In this case, however, an investigation using c.d.-spectroscopy is not possible since neither amylose nor the polymers PEG or PVA have suitable chromophores.

184

S. Kubik et al.

Thus in a first approach, we introduced a small number of naphthyl groups (DS = 0.003) into the helix by reaction of hydroxypropylated amylose with naphthyl isocyanate (Fig. 9).

Am-OH

+

0

OCN

Fig. 9. Chemical modification of amylose with naphthyl isocyanate

Because of their chiral environment, it should be possible to detect these chromophores by means of c.d.-spectroscopy. It can also be expected that the shape of the induced Cotton-effect should respond to conformational reorientations of the amylose helix. We did in fact find different spectra for the uncomplexed form of this amylose and its complexes with cyclohexanol and sodium dodecylsulfate (SDS) (Fig. 10).

25

I

200

,

I

I

I

I

I

I

300

I

I

I

X/nm

Fig. 10. C.d.-spectra of the naphthylcarbamate substituted amylose (a) and its complexes with cyclohexanol (b) and sodium dodecylsulfate (c(amy1ose)= 1 mg/mL HzO;pH 7: c(cyclohexano1) = 6.6 mol . L'; c(SDS) = 2.4 , rnol , L-').

I

3! 0

Moleciilar inclusion within polymeric carbohydrate matrices

185

In the case of the SDS complex, the sign of the Cotton-effect changed from minus for the uncomplexcd state to plus for the complexed one. These two spectra most probably represent two different helix conformations. Similarly, the addition of PEG to the naphthyl substituted amylose resulted in a change of the shape of the c.d.-spectrum (Fig. 11).

30

0

-10

I

200

I

1

I

1

I

l

l

I

I

I

300

I

I

I

X/nm

I

3 D

Fig. 1 1 C.d.-spectra of the naphthylcarbamate substituted amylose (a) and its complex with polyethylene glycol (b) (c(arnylose) = 0.75 mg/ml H 2 0 ; pH 7; c(PEG) = 4.0 (pL/mL).

The resulting spectrum can be interpreted in terms of a superimposition of the spectrum of uncomplexed amylose with that of the SDS complex. Presumably, PEG is only incorporated into some regions of the amylose chain. In principle we could demonstrate the interaction of PEG with amylose by c.d.-spectroscopy. However, definite information about the structure of these complexes could not yet be obtained. In the case of PVA, we tried to introduce chromophores into the polymer chains by oxidation of some hydroxyl groups to carbonyl groups or attachment of a 2-oxopropyl residue to some OH-groups of the polymer (Fig. 12). But no significant c.d.-spectra could be observed when hydroxypropyl amylose was mixed with these derivatives. We are now trying to apply other methods such as DSC for the characterization of these complexes. We hope that an exact knowledge about the different interactions of amylose with synthetic polymers will lead to interesting new materials.

186

S.Kubik et al.

Oxidation

)-OH

I *

Fig. 12. Chemical modification of polyvinyl alcohol.

In general, we believe our investigations showed that there are a variety of possible applications for inclusion complexes of amylose. In some cases, amylose can be an interesting alternative to cyclodextrins. Especially in the food industry, the use of starch as complexing agent should be less problematic than that of cyclodextrins. Not only from the industrial point of view but also scientifically, the investigation of the structure and behavior of amylose inclusion complexes is a fascinating area of research.

Acknowledgment. The authors would like to thank the Bundesminesterium fur Forschung und Technologie (registry number 03 I9057 A) and Cerestar, Vilvoorde (Belgium) for their financial support. We also thank Prof. F.W. Lichtenthaler, Darmstadt for the color representations of the amylose structures.

References I.

2. 3. 4. 5. 6. 7. 8.

9.

R.L. Whistler, J.N. Bemiller, E.F. Paschall, Starch - Chemistry and Technology, 2nd edition, Academic Press, New York (1984). R.L. Whistler, G.E. Hilhert, J . Ani. Cheni. Soc. 66 (1944) I72 I - 1722. W. Banks, C.T. Greenwood, Starch and its Components, Edinburgh University Press, Edinhurgh ( 1975). J. Langendijk, H.J. Pennings, Cereal Sci. Today 15 (1970) 354.356; H.J. van Lonkhuysen, J . Blankenstijn, StarcWStarke 28 (1976) 221-233. F.W. Lichtenthaler, S. Immel, Int. Sugar J . 97 (1995) 12-22: F.W. Lichtenthaler, S. Ininiel. Tetrahedron: Asyrnnierry 5 (1994) 2045-2060. T. Kuge, K . Takeo, Agr. B i d . Chem. 32 (1968) 1232-1238. Y. Yamashita, J . P o l p . Sci. Purr A 3 (1965) 3251-3260; Y. Yamashita, N. Hirai, J . Polyni. Sci. Part A - 2 2 (1966) 161-171; Y. Yamashita, K. Monobe,J. Pofyni. Sci. Part A-29 (1971) 1471-1481. W. Saenger, Aiigew. Cliem. 92 (1980) 343-361; Attgew. Cheni. 1nr. Ed. Engl. 19 (1980) 344-362; G . Wenz, Angew. Cheni. 106 (1994) 851-870; Angew. Cliern. In!. Ed. E n g l 3 3 (1994) 803-822. A. Harada, J. Li, M. Kamachi, Nature364 (1993) 516-518.

Molecidur iriclusion within polymeric carbohydrate matrices 10.

I I. 12. 13. 14

15.

16.

17.

18. 19.

1 87

H. Rein, K.J. Steffens, Mikroverkapselung schwerl"s1icher Arzneistoffe mit Starke; lecture given by H. Rein at the 45th Starch Convention of the Arheitsgenieinschaft Gerreideforschung at Detmold during April 20th to E n d , 1994; H. Rein, Dissertation. Philipps-Universitat Marhurg (1993). B. Pfannem,ller, G. Ziegast, In[. J. B i d . Mucromol. 4 (1982) 9-17. B. Pfannem,ller. H. Mayerh-fer, R.C. Schulz, Biopolymers 10 (1971) 243-261. K. Sensse, F. Cramer. Chem. Ber. 102 (1969) 509-521: K. Takeo, T. Kuge, Sfurrh/S/sturke 24 (1972)

28 1-284. M.C. Kajtar, E. Horvath-Toro. E. Kuthi, J. Szejtli. Acru Chim. Arud. Sci. Hung. 110 (1982) 327-355: K. Harata, H. Uedaira, Bull. Chem. Soc. Jpn. 48 (1975) 375-378; H. Shimizu. A. Kaito, M. Hatano. Bull. Cheni. Soc. Jpn. 52 (1979) 2678-2684. G. Wulff, S. Kuhik. Curbohydr. Res. 237 (1992) 1-10. G. Wulff, S. Kuhik, Mukromol. Chem. 193 (1992) 1071-1080. S. Kuhik, G. Wulff, SrurcWSrarke 45 (1993) 220-225; G. Wulff, S. Kuhik, H.G. Breitinger, Nuclnzuchsende Rohsfoffe- Perspektirienfiir die Cheniie. p. 31 1-322, M. Eggersdorfer, S. Warwel. G. Wulff (eds.). VCH. Weinheim ( 1 993). H. Riiper, H. Koch, Sturch/Starke 42 (1990) 123-130. R.L. Shogren, A.R. Thompson, R.V. Greene, J. Appl. Polym. Sci. 47 (1991) 2279-2286; R.L. Shogren. A.R. Thompson, F.C. Felkner, R.E. Harry-OiKuru, S.H. Gordon, R.V. Greene. J.M. Gould. J . Appl. Polyni. Sci. 44 (1992) 1971-1978.

This Page Intentionally Left Blank

11 Resistant starch

M. Champ and N. Faisant Laboratory of Nutrition and Applied Technology I.N.R.A., rue de la GCraudikre, B.P. 1627,44316 Nantes cedex 03, France

Summary. Resistant starch (RS) had been defined as "the sum of starch and products of starch degradation not absorbed in the small intestine of healthy individuals". It is thus composed of potentially digestible starch, oligosaccharides and crystalline fractions highly resistant to digestion. None of the analytical methods presently available are ahle to take into account all these fractions, but several methods provided similar data on most of the samples that had been analyzed. The fate of RS appears to be very similar to that of some of the non-starchy polysaccharides. Indeed it mostly disappears in the large intestine with a fermentation pattern similar to wheat bran. There is some evidence of its positive effect on both fecal bulking and lipid metabolism. Fermenting in the large intestine, its energetic value is lower than that of digestible starch. RS consumption seems to be rather low in most European countries (4 g/day); however, it should be possible to increase its intake if the positive effects of RS on health are confirmed in the future.

Introduction Numerous publications have shown that some raw or even cooked starches were partly unavailable for digestion by most animals species, including human~',~.3. Even if resistant starch is a rather new concept, the resistance of part of the starches has been recognized for a number of years. The term 'resistant starch' was first introduced by Englyst and co-workers4 at the MRC, Dunn Clinical Centre, Cambridge (UK) in 1982. These researchers found that many processed foods had higher apparent non-starch polysaccharides than the corresponding raw products. A detailed analysis revealed that the apparent increase was caused by a glucan that could be dispersed in potassium hydroxide. Thus they first defined Resistant Starch as starch resistant to dispersion in boiling water and hydrolysis with pancreatic amylase and pullulanase4. This fraction primarily consisted of retrograded amylose which appeared to be highly resistant to in vivo digestion. This definition was enlarged when in vivo experiments demonstrated that other fractions of starch could escape small intestine digestion in humans. In 1990, a European Flair concerted action was initiated to study the physiological implications of the consumption of resistant starch in humans, named EURESTA (acronym for

190

M.Champ and N . Faisant

European Resistant Starch research group). In 1992, resistant starch (RS) was defined within EURESTA as "the sum of starch and products of starch degradation not absorbed in the small intestine of healthy individuaIs"5. It is therefore the fraction of the starch which will not provide glucose to the organism but which will ferment in the large intestine to produce mainly gases and short-chain fatty acids. Resistant starch is thus comparable to dietary fibres and is even often considered a dietary fibre6.7.

Classification of RS; formation and origin of the resistance of starch; physical characteristics of RS RS cannot be properly defined chemically due to the fact that the resistance of starch is related to hydrolysis conditions (nature of the enzyme(s), ratio starch/enzyme, characteristics of the hydrolysis...). There is probably no starch fraction undigestible by any enzyme if the concentration is not limiting and the duration of the hydrolysis long enough. Thus a chemical definition should be based on starch not digested by amylolytic enzymes in standard conditions optimized on the basis of in vivo determinations (preferably healthy human subjects). The physiological definition is simpler, but one has to keep in mind that RS, in vivo, will depend on: - the physiological state of the subject,

- the environment of the starch when it is ingested, - and the other food components of the meal. Thus, there could not be any absolute value of RS content of a food. In 1987, Englyst and Cummingsg put forward a number of reasons explaining the incomplete digestion of starch in the small intestine: ( 1) Physically inaccessible starch (occurring mainly in partly milled grains and seeds), later identified as RS type I; (2) Native resistant starch granules (present in raw potato and banana), presently identified as RS type 11; (3) Retrograded amylopectin, now excluded from RS classification; and (4)Retrograded amylose (found, for instance, in cooled, cooked potato, bread and corn flakes). This fourth category of RS, now identified as RS Type 111, was initially the only one to be (partly) quantified by the analytical methods (Englyst et al., 19824 or methods derived from the AOAC method for determination of dietary fibresg). This classification proposed by Englyst is still valid10 (Table 1). However, it should be noted that types I and I1 or I and 111 can coexist in the same foodI1. Physically-inaccessible starch (RS 1) occurs when treatment of foods or chewing is not efficient enough to disrupt plant cell walls. Indeed in unrefined foods such as beans, starch

Resistant starch

191

Table 1 . I n virro nutritional classification of starch.I0

Type of starch

Example of Occurence

Rapidly digestible starch (RDS) Slowly digestible starch (SW Resistant starch (RS) I . Physically inaccessible starch 2. Resistant starch granules 3. Retrograded starch

Freshly cooked starch food

Probable digestion in small intestine Rapid

Most raw cereals

Slow but complete

Partly milled grain and seeds Resistant Raw potato and banana Cooled cooked potato, bread, and corn flakes

Resistant Resistant

granules are entrapped in cells and remain inaccessible to salivary and pancreatic a - m y lases. except when a mechanical treatment (chewing, stomachal grinding or preparation of the t'ood) has previously destroyed the structure of the cells. Native starch granules exhibiting a B-type X-ray diffraction pattern (RS2) have long been known as highly resistant to enzymatic hydrolysis. They are characteristic of many tubers, but also of banana and some varieties of corn and rice (high amylose corn and rice). However, most of the starchy foods eaten by humans are cooked. The banana when eaten before complete ripening contains substantial amounts of starch (B-type); the content of starch decreases during ripening while the sucrose, fructose and glucose content increases'z. When starch granules are fully gelatinized and dispersed, the starch becomes digestible. However, as the gel cools and ages, the polymers may again form a partially crystalline structure. Recrystallization or retrogradation depends on the formation of inter-chain hydrogen bonds and occurs more rapidly for the linear amylosel3. Retrogradation of amylopectin is limited by its branched structure and the polymers of retrograded amylopectin are less firmly bound than those of retrograded amylose. Retrograded starch characteristically forms the B-type pattern. Differential scanning calorimetry is a technique often used complementary to X-ray diffraction to characterize the crystalline structure of starch. It measures and records the amount of heat involved in the gelatinization process. These characteristics in addition to some others are listed in Table 2 for some of the samples studied during the EURESTA project.

Analysis of RS The analysis of resistant starch implies the performance of an enzymatic hydrolysis (or-amylase in most cases) which is usually supposed to mimic the hydrolysis of starch by endogenous enzymes

Roquette, Lestrem (F) Roquette, Lestrem (F) Nestle, Vevey (CH) INRA, Nantes (F) INRA, Nantes (F)

National Starch Cerestar, Vilvoorde (B)

Cerestar, Vilvoorde (B)

Raw potato starch Pregelatinized potato starch starch

Cooked green banana (flour)

High amylose corn starch (HACS)

Retrograded HACS

Purified resistant starch

pancreatin treatment

milled, 4'C 48h, dried and milled Retrograded HACS purified by triple

pealed, cooked freezefried and milled (Hylon 7), no specification Hylon 7, extruded

no specification no specification see reference 14 peeled, freeze-fried and milled

Preparation

* EURESTA document. a) Tmax "C : Melting temperature (at the maximum of the endotherm).

Bean flakes Raw green banana (flour)

Origin

product

Table 2. Chemical and physical characteristics of some starchy samples.* Resistant starch

96.2

98.6

96.3

12.6

30.0

68.8

4.2

54.4

73.0 71.9

66.4 5.0

47.1

B

B

84.3

B

148.6

144.9

108.3

140.4

50.9

amorphous

69. I

62.2

B amorphous amorphous B 0

91.1

DSC T max OC

64.9

panern

x-ray diffraction

97.5

(%, dry maner basis)

Total starch

Resistant starch

193

in the upper part of the digestive tract (mouth, stomach and small intestine). The quantification of 1 6 ~by1 7 RS can be made by a direct analysis of the residual starch after the h y d r o l y ~ i s 9 ~ ~ 5 ~ or subtracting the amount of starch that had been digested from the total starch content of the sample 10. The first attempt to analyze RS was performed by Englyst et d4. Their method was only able to analyze retrograded 'enzyme resistant starch'. Indeed the grinding of the sample and the subsequent thermic treatment at 100°C made the quantification of RSI and RS2 impossible. Englyst later identified the fraction quantified by this method as retrograded amylose. The main modifications introduced by Berry et a1.15 and then by the collaborators of the EURESTA interlaboratory study16 concerned the elimination of the gelatinization step and of the pullulanase hydrolysis. Consequently, both RS3 and RS2 could be quantified using this new method. Independently, Englyst el al. 10 developed a more sophisticated methodology set up to analyze rapidly digestible starch (RDS), slowly digestible starch (SDS) and resistant starch (RS). RS could be further divided in three categories (RSI, RS2 and RS3). Asp and collaborators provided a different method consisting of the measurement of 'residual starch' and 'total starch' in a fibre residue9,'s. However, their method was only able to analyze part of RS3 but neither RSI nor RS2. The only advantage was that in this case, RS is quantified as dietary fibre in one single analysis. A first collaborative study was conducted to compare methods of determining resistant starch

(RS) in various foods and food productsl6. Principally, two methods were used to analyze samples containing various forms of RS (Fig. 1): method A derived from Berry's methodls and method B derived from the method used by Bjorck et al.9. Several methods were also used to determine total starch. The RS yield by method A compared to method B was higher in all samples except unmilled bean flakes. When the starch was a raw B-type (potato starch), method B failed to detect any RS because of its initial heating step, whereas method A yielded 48% RS on an 'as is' basisl6. A comparison of these results with data obtained by Englyst et (11.l o in a separate ring test revealed that for most of the samples, the results were similar between the modified Berry method and the Englyst method. An exception was noticed for potato starch (75% RS according to Englyst method and only 56% according to Berry method). These results were confirmed in a ring test carried out by Hoffem and DysseIerl9. Minor modifications to method A were then proposed as described by Champ16 by SauraCalixto and colleagues (unpublished data) and by Faisant et These modifications were undertaken to improve the slight underestimation of RS. One of the modifications is the use of sodium azide to prevent bacterial proliferation during amylase hydrolysis. Saura-Calixto proposed

194

M.Champ and N . Faisant I100 mg fibre residue (+ celite)*/ piiiGzq Add I m l H 2 0 Add 10 ml enzyme solution in buffer, pH 6.9 I00 mg fibre residue (+ c e m pancreatic a-amylase (500 U) 0.1 M tris maleate buffer solution Add l n h H 2 0 Add 1 ml KOH 4M (Calcium chloride 4 mM) /Shake30 min. room temp.] then cool to room temp. Add lml Na' Acet. buffer pH 4.75,0.4M & about I .5 ml HCI 2M [Leave I h. centrifuael (to obtain pH 4.75) Add 0.5ml Na Acet. buffer Wash electrode with pH 4.15,O.ZM I Wash residue twice with 80% EtOH. drv at 60°C I 30pl amyloglucosidase 1.5 ml0.1M Acet. buffer Add 1.5'rnl water I ( l o mglml) then I .5 m/ 4 M KOH Add 60 pl amyloglucosidase (10 mglml) Mix 0.5 h at room temp. Shake 30 min.. 60°C. I centrifu e Shake 30 min.. 60°C. Add I2 rnl water entrifu e Wash pellet with htwi-

I

I

I

add about 0.65 ml 2 M acetic acid (to obtain pH 4.5)

(2OU/O.l &ml0. 0.1 ImlMamyloglucosidase Na acetate buffer pH 4.5

I

[Shake 90 y i n at 65°C

I

\T

)

1

Determination of glucose Glucose oxidase assay => Resistant starch

/

centrifuge Supernatant Adjust 10

with wafer

1 ml H20. centrifuge

Adjust to lOml with water

I

~

I

~of glucose ~ Determination~ of glucose ~ Glucose Oxidare assay Glucose oxidase assay => Residual starch => "Total" starch RS = "Total starch" Residual starch

-

Fig. I . Determintion of resistant starch by methods A and B.I6 * The fibre residdcelite was first isolated by the AOAC procedure.

to introduce a de-proteinization step with pepsin. Both proposed the elimination of the drying step before the solubilization with KOH.

In addition to the practical aspects, the fundamental problem is the definition of what should be analyzed with the method. None of the methods, including 'Englyst method', takes into account the whole amount of RS defined as "starch and products of starch degradation not absorbed in the small intestine of healthy individuals"l7. On the one hand, RS collected in vivo at the end of the small intestine of humans (intubation technique or ileostomates, see 5 'In vivo determination of RS') seems to consist of three more or less distinct fractions: - oligosaccharides (including glucose), - crystallites (linear chains of a-glucans), - long chains or starch granules damaged to a greater or lesser degree. On the other hand, the in vitro RS consists of linear chains of a-glucansl7 (crystallites as observed in vivo)and starch granules in the case of native B-type starch".

i

~

~

~

Rrsistaiit starch

195

In conclusion, the definition of resistant starch adopted in 1991 is a physiological one: "resistant starch is the sum of starch and products of starch degradation not absorbed in the small intestine of healthy individuals". As a result, RS as defined is not strictly resistant to the amylases and the term RS is often regarded as confusing by biochemists. None of the methods presently available are able to analyze RS as defined because they only quantify enzyme resistant starch whereas potentially digestible starch reaches the end of the small intestine. The method described by C h a m p t 6 and recently improved by Saura-Calixto (unpublished) and Faisant et

is quicker and more accurately reproducible than Englyst's

method, primarily because of the direct quantification of RS. Table 3 lists the RS content of some common foods according to Englystlo, who has provided the largest list of foods analyzed, Tahle 3. Total (TS) and resistant starch (RS) content of some common

Sample White bread Corn flakes Instant potato Boiled potato (hot) Banana flour (unripe) Boiled potato (cold) Spaghetti Peas (frozen, boiled 5 min) Bean flakes

Dry matter (9%)

g/100 g dry matter RS

54.5

1

95.8 16.7 22.8 99.1 23.8 28.3 18.3 93.6

3 I 5 57 10

5 5 6

TS 77 78 73 74 75 75 79 20 49

Industrial production of RS; functional properties of RS In order to increase RS consumption, several approaches can be proposed to the consumer and the food industry. Dietary recommendations can suggest the increase of the consumption of nonrefined food such as legumes, which, when cooked in normal conditions, can provide substantial amounts of RS. However, the evolution of the dietary habits in most European countries would impose a diversification of the 'ready-to-eat' foods that will provide a variety of meals with a high RS content. The trend in multi-grain breads has no doubt increased resistant starch through encapsulation in plant cell walls. A second way to increase RS consumption would be to provide consumers with foods enriched in RS by adding concentrates of RS (high amylose corn starch, for

196

M. Champ and N. Faisant

instance) or by promoting starch retrogradation during the cooking process and post-cooking treatment. Food technologists and processors were quick to realize that processing techniques that increase the amount of resistant starch in foods would have potential nutritional and commercial value. As autoclaved cereal starches were the first to be characterized, this led to the use of the expression 'man-made fibre' to designate retrograded amylose. Two starches with a high content in resistant starch are already commercially available (NoveloseTM, National Starch and CrystaleanR, Opta Food Ingredients, Inc.): these are probably both obtained from high amylose corn starch treated to promote retrogradation. These products can easily be incorporated in a number of foods such as breads or cookies. Optimal conditions for retrogradation depend on the nature of the starch but they appear in most cases to be around 4°C with an hydration level above 70%. Furthermore, there is a need to explore the new food products such as the various kinds of ready-to-eat products kept at 4°C or -20°C in order to know if there is a possibility of increasing RS consumption without completely modifying food habits of populations. Another obvious method of increasing resistance of starch is by chemical modification of starch. Modified starches are already widely used in the food industry for their functional properties. The enzymatic hydrolysis depends on the degree of cross-linking or substitution as well as the types of cross-linking agent, esterification or etherification groups. The nutritional contribution of such starches in foods and their behavior during analytical procedures require systematic studies. Several groups are already trying to produce breads or biscuits enriched in high amylose starch (native or thermically treated)21.22. Final concentration of the RS in the products could reach 14 g/100 g DM. Moreover, one recent study23 shows that RS can be used as texture agent. The authors claimed that the starch preparation forming a particle gel network could replace part or all of the oil in dressings.

In vivo determination of RS In order to be digested and absorbed in the small intestine, starch must be completely hydrolysed into glucose. This depolymerization is done by several digestive enzymes that cleave the a( 1-4) and a(1-6) glucosidic bonds. In humans, the a-amylases (a(1-4) glucan hydrolase E.C. 3.2.1.1.) are present in salivary and pancreatic secretions. They act on both amylose and amylopectin in an endo fashion, releasing glucose, maltose and higher dextrins. While glucose is absorbed directly through the small intestinal mucosa, the oligosaccharides are taken up by membrane-bound glucosidases which cleave a( 1-4) and a( 1-6) bonds and finally release glucose.

Resistant starch

197

The complete depolymerization of starch to glucose was considered totally efficient until the 1980s. The resistance of some starch to digestion has become more and more obvious since. Several methods are available to assess physiologically resistant starch in vivo (i.e. starch fraction escaping small intestine digestion). RS can be quantified directly by collecting ileal samples in humans or animals, or indirectly by estimating the amount of starch fermented in the colon.

In humans, one direct method consists of working with ileostomized patients (persons who have undergone a colonic resection). The advantage of this technique is the direct and total collection of the ileal contents. However, such patients cannot be considered physiologically 'normal'. Indeed, a bacterial proliferation takes place in the terminal part of the gut. While several authors consider that the activity of this flora is sma11'2~24~2~~26, the fermentation cannot be Table 4. Starch malabsorption determined with ileostomized suhjects.

Product ingested

Amount of starch ingested

Malabsorption

0 Rice +bread Rice + bread Bread Wheat + potatoes Potatoes White bread Whole wheat bread Lentils White bread Raw oat flakes Corn flakes Ripe banana Green banana Potatoes cooked cooled reheated Potato flakes + rice +bread Bean flakes + rice + bread Wheat finely ground Coarse wheat

98-127 107 107 157 25 100 100 100 61.9 57.8 74.2 2.1 20.1 45.4 47.2 47.2 161.3 154 52.4 52.4

Ref.

(% of ingested starch)

2.1 0.7 4.0 14.3 26.4 2.4 7.0 5.7

(a) the calculation does not take into account the oligosaccharides. (b) the calculation takes into account all the a glucans.

(b) 0.5 0.9 0.8 2.4 2.0 13.0 11.0 21.0 2.5 2.2 5.0 66.7 95.0 3.3 12.9 7.6 1.8 8.6 0.7 1.1

27 28 29 30

24

12 25

26 31

198

M . Champ arid N . Faisarit

neglected and could result in underestimation of the amount of undigested starch. Moreover, an adaptation consisting of an increase of the water and electrolytes absorption capacity has been observed27. This phenomenon may have an incidence on nutrients absorption. Several authors used the ileostomy model to quantify the amount of starch escaping small intestine digestion depending on the food ingested. Results are reported in Table 4. A second direct method applicable to humans consists of the intubation of healthy volunteers

with a triple lumen tube passing through the nose down to the terminal ileum. One lumen is used to sample ileal content 5 cm above the ileocecal junction. The second one, 25 cm proximal to the aspiration port, allows the infusion of a recovery marker used to estimate water flow rate through the distal ileum. A third one is used to inflate a bag containing mercury at the tip to facilitate passage into the gut. Once the bag reaches the caecum (confirmed by fluoroscopy), it is deflated and the subject is required to remain in a semi-recumbent position. Each experiment consists in a continuous aspiration of the ileal effluents during fourteen hours after the ingestion of the experimental meal; the total amount of starch recovered over this period is the amount of undigested starch. The main advantage of this technique is that it is performed on healthy normal humans. It also allows the measurement of the total digestibility of starch by analyzing the faeces. The main drawback, however, is the presence of the tube along the small intestine and its possible Table 5. Starch malahsorplion determined by the intubation technique in humans

Product ingested

Amount of starch ingested

Malabsorption (% of ingested starch)

(g)

(a)

(b)

300

9.5

10.5

beans

61

5.9

8.0

Bread -tpasta + potatoes

100

4.I

5.2

300

3.2

4.1

Banana + rice

Ref.

33

Banana + rice + potatoes +

I,

I,

!I

34

Retrograded high amlylose corn starch

33

49.4

35

Complexed high amylose corn starch

33

Green banana flour

23.1

White beans

68.5

(a) and (b) : same as in table 4.

20.6 74. I

83.7

11

11.3

36

Resistant starch

199

influence on the transit. Read et al.32 observed that the residence time of food in the intestine was decreased from 6.5 to 5.3 hours (p
Product ingested

Amount of starch ingested (g/day) 8.4 5.3 5.0 5.0

Wheat starch + gluten Wheat flour Wheat bread crust crumb Wheat starch + autoclave wheat starch Wheat starch (30 %) + potato starch (5 %) + precooked bean flour (38 %) Preecooked bean flour Precooked lentil flour Cow peas Flat bread with normal corn Flat bread with high amylose corn Complexed high amylose corn starch (HAS) Retrograded HAS

Malabsorption

Ref.

(% of ingested starch)

0.05 0.06 0.2 0.3

9

6.0

38

6.3

39

2.5 2.5 5.0 5.3 5.3

10 11 10 4.2 32.5

37

10

13.8*

10

6.9*

8.2

40 41

42

200

M. Champ and N. Faisant

the small intestine. This technique allows the comparison between total digestibility (untreated rats) and ileal digestibility (treated rats). However, bacterial activity is never totally suppressed by antibiotics9. So undigested starch could be underestimated37. Results obtained with this technique are reported in Table 6. Pigs are often used as models for human nutrition. The ingraft of an ileo-cecal fistula allows the direct collection of ileal effluents. The collection can be total if a balloon is placed past the canula to prevent the effluents from passing through the colon, or only partial if not. In the last case, the use of a flow rate marker such as chromic oxide is needed43. In pigs, it is known that the small intestine inhabits an active bacterial flora that could induce an overestimation of starch digestion in the small intestines. In humans, hydrogen breath test has been used to quantify the amount of starch entering the colon. This indirect technique is based on the measurement of hydrogen in expired air. Indeed, the fermentation of carbohydrates entering the colon produces short chain fatty acids and gases such as CO,, CH, and H,. Part of these gases is absorbed and transported by the blood and excreted through the lungs in expired a@. For each subject, the total amount of H, measured in breath after ingestion of an experimental meal is compared to the amount of H2 expired after ingestion of Table 7. Starch malabsorption estimated by the hydrogen breath test.

Product ingested White bread Whole bread Lentils Wheat flour Whole wheat flour Potatoes Rice Cooked beans Potatoes cooked cooled reheated Raw oat flakes Cooked oat flakes Green banana Conventional rice Precooked rice Banana + rice nm : non mentioned

Amount of starch ingested 120 120 120 50 50 50 100 100 60 60 60 35 35 100 50 50 nm

Malabsorption

Ref.

(% of ingested starch)

10.7 8.3 17.6 2.8 13.6 10 1.8 38.0 6.7 22.8 8.0 7.8 5.8 38.6 0.7 0.8 0.0

30

46

47

48 49 50 33

Resistant starch

201

a known amount (usually 10 g) of lactulose. This carbohydrate is not digested in the small intestine and is totally fermented in the colon. The 'calibration' of the subjects with lactulose allows the quantification of carbohydrate fermented in the colon. Many authors have used this technique because i t is easy and non-invasive (Table 7), but several drawbacks must be noted before discussing the results. When complex foods are studied, different sources of fermentable carbohydrates can reach the colon and it becomes impossible to distinguish H, coming from starch or dietary fibre fermentation. Moreover, fermentative pathways are different from one substrate to another and the comparison between the production of H, by starch and lactulose fermentation is difficult". The proportion of H, absorbed by the colon and expired through the lungs depends on the use of H, by microflora and the kinetics of its production49. Therefore, the breath test cannot however, be useful in detecting really be used to quantify starch m a l a b ~ o r p t i o n ~It~ can, ~~'~ ~~. some clinical malabsorption or in screening the digestibility of different products. In all cases, it is important to work in highly standardized experimental conditions. Regarding all these methods used to assess starch digestibility in the upper part of the gut, it seems very difficult to arrive at a precise determination of RS in vivo. All have drawbacks but are still useful in estimating (and studying) RS in various foods.

Composition and characterization of in vivo RS The quantitative results obtained by the various methods show a large variability between one technique and another, and also between one food and another. By analyzing the ileal effluents, some authors observed that free glucose could be found in the terminal ileum33.34 and that part of the undigested starch in vivo could still be digested by amylasesl2.24.25.26. A physico-chemical characterization of ileal starch collected after ingestion of different starchy foods showed that in vivo RS consisted of 3 populations of a-glucansl7.5~:the first one was composed of oligosaccharides (DPn15)which were products of starch hydrolysis; the second consisted of crystalline fraction (DPn=15) composed of retrograded starch formed after cooking, known to be intrinsically very resistant to amylase hydrolysis; and the third one was composed of high molecular weight a-glucans attributed to starch fragments (DP,>100) that had escaped digestion because of inaccessibility or insufficient time of contact with enzymes (Table 8). In the case of uncooked starches that are intrinsically resistant to enzymic digestion (raw green banana, for example), most of the undigested starch was made of starch granules that were hardly altered superficially and the rest were oligosaccharidesl 1 . No intermediate population could be observed (Table 8).

202

M.Champ and N. Faisant

The respective proportions of these fractions varied from one starch sample or starchy food to another. Native B-type starches (extracted or within food) were poorly digested in the small intestine resulting in a large fraction of starch granules slightly eroded but mainly undamaged. The RS fraction of retrograded high-amylose corn starch mainly showed the characteristics of retrograded amylose with a large proportion of crystallized residues with an approximate DPn of 15. 'Normal foods' such as bean flakes or potato flakes were extensively degraded during their passage in the mouth, stomach and small intestine; the RS was constituted of the 3 fractions mentioned above, indicating a partial retrogradation of starch during cooling of the flakes. In all the cases, oligosaccharides were present in the RS, probably revealing limiting activities of the enzymes of the brush border oligosaccharidases (including isomaltase and glucoamylase) at the end of the small intestine or limiting capacities of absorption. Botham er al.54 also found that starch arriving at the end of the small intestine of humans after ingestion of a meal containing a retrograded starch gel contained molecules with a wide range of molecular weights. A major fraction consisted of fragments of amylose with a weight average Table 8. Chararacteristics of in vivo RS.

Origin of the RS

Retrograded high amylose com starch Complexed high amylose corn starch Bean flakes Potato flakes Green banana flour

in vivo RS (% of TS) 49

Repartition of the sizes DP& 8

D P p 15 79

DPn> 100 13

17,53

21

6

83

11

17.53

19 3

13 58 13

53 17 0

34 25 87

53 53 11

DPw 70-80 = main fraction

Retrograded starch gel

Drum dried corn starch Autoclaved high amylose corn starch

Ref

4 44

DP120 5 5

DPwa4O main fraction main fraction

54

55 55

DP : degree of polymerisation at the maximum of the peak of the gel permeation chromatography ; DPn : number average degree of polymerisation ; DPw : weight average degree of polymerisation ; (DPn = ECi/X(Ci/DPi) ; DPw = XCiDPi/ECi ; Ci & DPi : a-glucans concentration and degree of polymerisation of each fraction). The DPi was calculated from Kav using lintnerised wrinkled pea starch as reference. a Determined from Kav using dextran standards.

Resistant starch

203

Table 9. Degree of polymerisation (DP, DPn and DPw) of the crystalline fraction of in vivo RS.1J.53a

DP

DPn

DPW

Retrograded high amylose corn starch

17

17

38

Bean flakes

16

14

26

DP : degree of polymerisation at the maximum of the peak of the gel permeation chromatography; DPn : number average degree of polymerisation; DPw : weight average degree of polymerisation; (DPn = XilC(Ci/DPi) ; DPw = XCiDPiECi ; Ci & DPi : a-glucans concentration and degree of polymerisation of each fraction) There is a mistake in the calculation of DPn and DPw in this paper.

degree of polymerization (DP,) of 70-80 (Table 8). The authors concluded that the resistant product obtained in vivo was comparable to the in vitro resistant products. However, they did not consider the 'potentially digestible' fraction of RS. Ekwall et al.55 analyzed the ileostomy effluents after ingestion of drum-dried 'snow-flakes' (25% amylose corn starch) and autoclaved Hylon VII (70% amylose corn starch). Five percent of the starch recovered in the effluents during both test periods was extractable in 60% ethanol (DP120). The main starch fraction identified had an average DP of about 80 (Table 8). Table 9 illustrates the importance of the mode of expression of the degree of polymerization. The use of DP (without n or w as an indice) usually means degree of polymerization at the maximum of the peak of the gel permeation chromatography. The definition of DP, and DPn are indicated in the legend of Tables 8 and 9. These studies highlight the fact that, in addition to intrinsic structure of starch, several factors such as transit time, extent of chewing, physical inaccessibility, amylase concentration, and the presence of other food components influence the digestion of starch and contribute to the amount of a-glucans that reach the terminal ileum.

Energy value of RS The malabsorption of starch in the small intestine consequently decreases the amount of available glucose from a food and then represents a loss of energy. Part of this loss can be recovered through fermentations in the colon. RS entering the large intestine is fermented to a greater or lesser degree with the production of SCFA, gases (CO,, H2 and CH, (in part of the population)), heat and bacterial cells. But only absorbed SCFA make a significant contribution to energy

204

M . ChampandN. Faisant

salvage. The major determinant on energy salvage is the amount of RS fermented. It can be estimated by the difference between ‘total’(food-to-faeces disappearance) and ‘ileal’(food-to-end of the ileum disappearance) digestibilities. Since some of this organic matter (OM) will be incorporated into bacterial cells, a means of predicting net bacterial cell yield per unit RS fermented is required. Such information is not yet available, but considering information regarding nonstarch polysaccharides (NSP) fermentation, bacterial OM yield would be approximately 263 mg/g fermented RS56. The energetic contribution of unabsorbed carbohydrates in the small intestine depends on the methods of calculation and on the substrates. In his calculations of energy salvage from NSP digestion, Mathers56 considered the fermentation pattern with molar proportions (mmol/mol) of 650, 200 and 150 for acetate, propionate and butyrate. He also assumed that NSP fermentation follows conventional anaerobic stoichiometry, i.e. each hexose molecule can give rise to two molecules of acetate or two molecules of propionate or one molecule of butyrate and that 0.95 of net SCFA production is absorbed. Based on these assumptions, over-all energy absorption (using heats of combustion of cellulose and of individual SCFA) per g NSP fermented was estimated to be 9.0 kJ/g fermented NSP. Few data are available on the patterns of SCFA resulting from fermentation of various RS products in viv042.”. More information has been obtained within in vitro fermentation studies. They seem to indicate a relatively high production of butyrate. Increasing the molar proportion of butyrate from 150 to 250 mmol/mol at the expense of acetate (whilst propionate is held constant at 200 mmol/mol) increased the energy salvage by RS fermentation only slightly from 9.0 to 9.2 KJ/g. If acetogenesis (synthesis of acetate from CO, and H2) is very efficient, which is not known, the energy value of RS could be increased by up to 33%58. The energy value of RS could then be considered to be between 9.0 and 12.2 KJ/g. However, using a different calculation, Livesey et ( ~ 1 . 5estimated ~ the energetic value of RS in the rat to be 12.4 kJ/g and 15 kl/g for a-amylase-resistant, retrograded pea and maize starches.

RS and the glycemic index The extent and kinetics of digestion of starch in the small intestine influence both the amount of starch reaching the terminal ileum (i.e. RS) and the glycemic response to the food containing the starch (i.e. the passage of glucose into the blood). RS is an absolute value that corresponds to the total amount of malabsorbed starch at the end of the small intestine after the ingestion of ;1 meal including a starchy food. The glycemic response reflects the early digestion of starch and absorption of glucose. Although some studies have demonstrated a good correlation between in vitro amylolysis rates and in vivo glycemia responses60,61,62, no correlation exists between RS content and

Resistant starch

205

Table 10. Relationship between glycemic index of some starchy foods and their RS content.6+55 ~

Food

Glycemic index (glucose = 100)

RS content (% of total starch)

70-98

7 4 1

Bread Banana Spaghetti

80 72 69 62 50

76 6

Legumes

29-40

3-15

Potatoes Corn flakes Rice

1

glycemic index62963 (Table 10).For example, Liljeberg et a1.62 observed very different glycemic indexes for breads all containing about 1% RS. Only some foods, such as legumes and pasta, yield low glycemic index in agreement with their relatively high RS content. In other respects, RS is considered part of the dietary fibres. It is known that some (soluble) fibres can influence the glycemic index. This effect is mainly related to their viscosity, which delays gastric emptying. In the case of RS this effect cannot be expected but several authors explored both acute and long-term effects of RS on glycemia and insulinemia. By comparing the effect of ingestion of the same amount (30 g) of an RS (lintner from high amylose maize starch) and cellulose on energy expenditure and blood glucose and insulin, Ranganathan et a1.66 observed that this RS had an effect similar to the insoluble fibre cellulose. Raben el al.67 made the same observation by comparing raw potato starch to cellulose. In fact, the glycemic index reflects the way the digestible fraction of starch is digested and absorbed while RS content is a quantification of the unabsorbed fraction without taking the kinetic phenomena into account.

RS and lipid metabolism The incidences of RS consumption on cholesterol and triglyceride metabolism are uncertain. By substituting part of the digestible starch by a large amount of RS (up to 40% of the food), different authors observed a decrease of plasma cholesterol and triglycerides in normal rats after several weeks adaptation68-69-70.However, Gee et aL71 used lower amounts of substitution and did not observe any effect of retrograded amylose on cholesterolemia and triglyceridemia in rats over two weeks. In humans, Behall et al.72compared the influence of two starches, one containing a high

206

M . Chump and N. Faisunt

level of amylose, the other a high level of amylopectin: they observed a lower basal plasma triglyceride and cholesterol with the high amylose starch than with the other. RS could, therefore, have long-term effects similar to dietary fibres. As dietary fibres were shown to have an effect on postprandial metabolism when added as a supplementation to a we tried to test a similar effect with RS. We compared postprandial metabolism after a normal meal to the same meal supplemented with 30 g raw potato starch on six healthy individuals. No significant differences could be observed, but two of the six subjects had lower cholesterol and triglyceride postprandial responses with the supplemented mea174. This must considered as a case study, but the influence of RS on the postprandial metabolism should be studied further. The mechanisms by which RS could interfere with lipid metabolism are still hypothetical. One could mention the role of a lower insulinic response with some food, a possible influence of RS on bile acids secretion68.75. or the role of metabolites produced by colonic fermentation on liver metabolism68.

RS, colonic fermentation and physiology of the large bowel The total digestibility of starches in most of the cases is close to 100%; an exception is high amylose maize starch which is only 90% digestible51.76. Therefore, almost all the starch reaching the colon is fermented. Fermentative metabolism of starch was studied in in vitro experiments using rats or human fecal inocula. Several authors showed that a number of sources of starch were fermented in vitro, some very slowly77, and that fermentation products included H,, CO,, acetate, propionate, butyrate and lactate78.79. Animal studies, principally in rats, have shown RS to be fermented to different extents in vivo depending on source, and studies using antibiotics showed that the microflora played an essential role in RS fermentation80.81. A ring trial study organized within EURESTA program on in vitro fermentation of RS with human flora showed that rates of fermentation were variable depending on sources. Retrograded amylose was slowly fermented while pre-gelatinized potato starch was fermented the most rapidly82. While Weaver et aLg3 observed differences in kinetic and fermentative profiles between methane-producers and nonmethane producers, the collaborative study did not show any differences. But comparisons between one study and another are difficult since experimental conditions play an important role on final results whatever the substrate. For example, rat cecal flora is different from human fecal fl0ra77. Despite these variabilities, several authors observed that starch fermentation induced the production of a high proportion of butyrate (25 to 30% of total SCFA) compared to other rapidly

Resistant starch

207

fermented substrates83. When acarbose was used to inhibit small bowel digestion of starch, Scheppach et al.84observed an increase of 50% of fecal butyrate excretion. Colonic flora is also able to adapt to RS ingestion over a period of several days: amylolytic enzymes are induced and metabolic pathways seem to be modified during the adaptation79.81. Faulks et ~ 1 . 8 1observed a four-day adaptation in rats to maize starch and an eight-day adaptation to pea starch. All these phenomena tend to increase the efficiency of fermentations. Any fermentative event can influence colonic content and intestinal morphology. Several studies showed that RS feeding induced cecal pH fall (around pH5) associated to lactate accumulation8~~86~87. In parallel, caecum size and content increased. But here again, results were different from one substrate to an0ther57.8~.While little starch can be detected in faeces, RS tends to increase fecal excretion. Shetty and Kurpads9 showed that a supplementation by 100 g per day of raw maize starch increased faecal mass by 30% without modifying the transit time. By inhibiting small bowel digestion with acarbose, Scheppach et ~ l . observed 8 ~ an increase of fecal mass by 68%. These observations were caused by the large development of bacterial mass with the available substrate. However, such laxative properties could hardly be observed in more realistic experimental conditions of feedinggo. Cummings et ~ 1 . 9 1showed that RS laxative effect was much less than that of wheat bran. Several epidemiological studies have shown the role of colonic fermentations of dietary fibres on the prevention of colonic diseases92. Several indices lead to the conclusion that RS could participate in this prevention. RS induced a decrease of the level of ammonia, branched chain fatty acids and fecal enzymes involved in detoxification87. Moreover, the reduced pH that accompanies fermentation of starches and other carbohydrates may, in part, be responsible for the reduced transformation of primary bile acids to the mutagenic secondary metabolites86 and the reduced activities of other bacterial specific biotransformations in the large bowe188. Feeding rats with Hylon-VII (a high amylose maize starch) or the amylolysis residue of retrograded high amylomaize starch induced changes in crypt cell proliferation93. In a human feeding study with fourteen healthy volunteers consuming 45 g Hylon VII daily, significant changes in colonic function and crypt cell proliferation were seen including an increase in stool output and in the relative amount of butyric acid in faeces (Munster et al. cited by Cummings et ~ 1 . 8 2 ) .The concentration of deoxycholic acid in fecal water fell as did fecal water toxicity. Crypt cell production rates obtained from rectal biopsies showed a reduced labelling index, indicative of suppressed cell proliferation. In conclusion, a number of forms of RS can reach the colon and can have different effects. RS is largely fermented, producing mainly SCFA, and may be important in determining colonic

208

M.Chump and N. Faisarit

epithelial cell health through effects on bile acids, butyrate production and moderation of nitrogen metabolism. Table 1 1 summarizes these different effects. Table 1 I . Effects of RS in the large intestine.82

Property of RS

Present evidence

Physical effects

None

Fermentation

Mostly complete

Short chain fatty acids

Some retrograded more resistant Butyrate production increased

C02, H2 production Bile acids Cell proliferatiodgenetic effects

Higher yields compared to NSP Occurs. Yields less than lactulose Reduced deoxycholate Stimulated in proximal colon Suppressed in small intestine and distal

Bowel habit

large bowel Stool output increased

Transit time

No effect

N metabolism Minerals

Increased biomass and increased N

Diseases

Colo-rectal cancer; constipation

No effect

Epidemiological study on RS consumption in Europe The consumption of starchy foods and resistant starch in Europe has recently been e~aluated9~. Most of these data are based on national surveys determined on the basis of consumption at home

or on the amount of food sold at a national level. Numerous analyses have been performed on food by the participants according to the method described by Englyst et ~ 1 . 1 01992 or Champlh. The average European consumption of RS would thus be around 4.1 g per day94. The values presented in Table 12 are indicative of the respective consumption of RS in each European country. As evaluations are made on different bases, no real conclusions on the dietary habits of each country can be drawn from these data. The differences cannot be attributed to the nature of the analytical procedure by itself, but under- or overestimations in the analysis of some starchy foods might have been made by some of the countries. In any case, the data should be completed by using more complete and detailed National Surveys and foods should be analyzed as eaten in each country.

1.43 2.8 1.1 1.96 1.45 2.52 0.96 2.6 1.05 1.79

Belgium

Denmark

England

France

Germany

Netherlands

Norway

Spain

Sweden

Switzerland

Mean

Breads

Country

1.3

0.54

0.17

0.3

1 .o

1.9

0.03

0.29

0.03

0.17

0.2

0.09

0.2

0.24

0.6

0.72

0.65

0.04

0.6 1

4.1 1

4.38

3.36

5.74

3.22

5.29

3.73

3.97

3.67

3.99

Total

0.36

0.36

Vegetables

3.75 0.5

0.64

0.6

0.56

Banana

1.3

0.3 1

0.12

0.1 1

0.16

0.53

& breakfast cereals

Biscuits

0.73

0.16

0.15

Rice

0.04

0.0 1

0.14

Pasta

0.91

1.17

0.97

0.59

1.5

0.54

1.18

Potatoes

Table 12. Consumption of resistant starch in Europe ( g R S / d a ~ ) ? ~

2

2 J

t

B

2'

210

M . Champ and N . Faisant

Conclusions The definition of resistant starch adopted in 1992 is a physiological one: "resistant starch is the sum of starch and products of starch degradation not absorbed in the small intestine of healthy individuals". As a result, RS as defined is not strictly resistant to the amylases and the term RS is often found by the biochemists to be confusing. RS as defined by EURESTA in 19925 is composed of oligosaccharides (including glucose),

a-glucans of high molecular weight (mainly starch granules) and a crystalline fraction whose size depends on the origin and the treatment of the starch. An analytical method of RS should then take into account the three fractions identified and defined above, whatever the source of starch. Even if there is some underestimation of the RS content in some of the food, it seems that the amount of RS actually eaten by most Europeans is very small (4.1 g/day/person). That observation raises several questions. Is the increase of RS consumption necessary and useful for the health of most of the population? Is RS more favorable than non-starch polysaccharides-rich materials or foods in improving some metabolic disorders such as hyperlipidemia or diabetes? Even if there is some evidence to suggest that RS is beneficial to the health, more studies should be performed to investigate the physiological consequences of the consumption of the different forms of RS.

References 1.

2. 3. 4.

5.

6.

7. 8.

C.F. Langworthy, H.J. Deuel Jr.: Digestibility of raw rice, arrowroot. canna. cassava, taro, tree-fern, and potato starches. J. B i d . Cheni. 52 ( I 922) 25 1-26 I . M.J. Wolf, U . Khoo: Partial digestibility of cooked amylomaize starch in humans and mice. Sfarch 12 (1977) 401-405. M . Champ: Digestion des glucides chez le monogastrique. Reprod. Nictr. Divelop. 25 (4B) (1985) 819-842. H.N. Englyst, H.S. Wiggins, J.H. Cummings: Determination of the non-starch polysaccharides in plant foods by gas-liquid chromatography of constituent sugars as alditol acetates. Analjsr 107 (1982) 307-3 18. N.G. Asp: Preface 'Resistant starch - Proceedings from the second plenary meeting of EURESTA: European FLAIR Concerted Action NO1 I on physiological implications of the consumption of resistant starch in man'. Eur. J. Clin. Nirtr. 46 (Suppl.2) (1992) SI. A.B. Blakeney: The occurrence and chemistry of resistant starch, in: Dietary Fibre and Beyond - Australian Perspectives. S. Santman and G. Annison (Eds). Nutrition Society of Australia. Occasional Publications Volume I , 1993. pp. 37-46. N.G. Asp: Nutritional classification and analysis of food carbohydrates. Am. J . ClOr. Nurr. 59(Suppl) ( 1994) 679s-68 IS. H.N. Englyst, J.H. Cummings: Digestion of polysaccharides of potato in the small intestine of man. Am. J . C h . Nurr. 45 [ 1987) 423-43 1.

Resistrrrit starch 9.

211

1. Bjorck, M. Nyman, B. Pedersen, M. Siljestrom, N.G. Asp, B.O. Eggum: On the digestibility of starch in

12.

wheat bread - studies in virro and in rivo. J. Cereal Sci. 4 ( 1986) I - I I , H.N. Englyst, S.M. Kingman. J.H. Cummings: Classification and measurement of nutritionally important resistant starch fractions. Eur. J . Clin.Nufr.. 46 (Suppl.2) (1992) S33-SS0. N. Faisant, A. BulCon, P. Colonna, C. Molis, S. Lartigue. J.P. Galmiche, M. Champ: Digestion of raw banana starch in the small intestine of healthy humans: structural features of resistant starch. Br. J. Nurr. 73 (1995) I 1 1-123. H.N. Englyst, J.H. Cunimings: Digestion of the carhohydratcs of banana (Musu paradisiacu .sapiemtrn) in the

13.

human small intestine. Am. J. Clin. Nurr. 44 (1986) 42-50. P. Colonna, A. BulCon, V. Leloup: Limiting factors of starch hydrolysis. Eur. J. Clin. Nutr. 46 (1992)

10.

I I.

14. 15.

16.

s 17-32. L. Tappy, P. Wiirsch, J.P. Randin, J.P. Felher, E. JCquier: Metabolic effect of precooked instant preparations of bean and potato in normal and diabetic subjects. Am. J. Cliti. Nirrr. 43 (1986) 30-36. C.S. Berry: Resistant starch: formation and nieasurenient of starch that survivcs exhaustive digestion with amylolytic enzymes during the determination of dietary fibre. J. Cereal Sci. 4 (1986) 301 -3 14. M. Champ: Determination of resistant starch in foods and food products: Inter-laboratory study. Eur. J. Clin. Nutr. 46 (Suppl.2) (1992) S51-62.

17. 18.

19.

20. 21.

22.

23. 24. 25. 26.

27.

N. Faisant, M. Champ, P. Colonna, A. BulCon: Structural discrepancies in resistant starch obtained in L1ii.o in humans and in virro. Carhoh. Polyrn. 21 (1993) 205-209. M. Siljestrom, N.G. Asp: Resistant starch formation during baking - Effect of baking time and temperature and variations in the recipe. 2. Leberisrn Unters Forsch. 181 (1985) 4-8. D. Hoffem, P. Dysseler: I . Results of the ring tests (1992-1993 and 1993-1994) on the determination on the resistant starch. Proceedings of the EURESTA Summing-up meeting, La Londe - Les Maures (F), (1994)

p. 21. N. Faisant, V. Planchot, F. Kozlowski, M.P. Pacouret, P. Colonna, M. Champ: Adaptation of the Berry modified method to high RS products. Sci. Alinz. 15 (1995) 83-89. C. Hoebler, A. Karinthi, F. Guillon, H. Chiron, A. Buleon, C. Melegari, J.L. Barry: Low metabolic responses of cereal foods and structure of starch in the food matrix. JournCes de la Chimie et de la Biochimie des Glucides, Le Croisic, France. June 1994. Abstract. R.C. Eerlingen. I.P. van Haesendonck, C. de Paepe, J.A. Delcour: Enzyme-resistant starch. 111. The quality of straight-dough bread containing varying levels of enzyme-resistant starch. Cereal Clzern. 71 i1994) 165170. P. Wiirsch, G. Gumy, D. Sievert: RS containing starch as texture agent. Proceedings of the EURESTA Summing-up meeting. La Londe - Les Maures (F), April 1994. Abstract p. 39. H.N. Englyst, J.H. Cummings: Digestion of the polysaccharides of some cereal foods in the human small intestine. Am. J. Clin. Nutr. 42 (1985) 778-787. H.N. Englyst, J.H. Cummings: Digestion of polysaccharides of potato in the small intestine of man. A t ~ i J. . Clin. Nurr. 45 ( 1987) 423-43 I . T.F. Schweizer, H. Andersson,,A.M. Langkilde, S. Reirnann, 1. Torsdottir: Nutrients excreted in ileostomy effluents after consumption of mixed diets with beans or potatoes. 11. Starch, dietary fihre and sugars. Eur. J . Clin. Nutr. 44 (1990) 567-575. A.S. Sandberg, H. Anderson, B. Hallgren. K. Hasselblad. I . Isaksson: Experimental model for iti viw determination of dietary fihre and its effect on the absorption of nutriments in the small intestine. J. Nurr. 45 (1981) 283-294.

212 28.

29. 30.

31. 32 33. 34.

35.

M.Chump and N . Faisant A S . Sandberg, H. Andersson. B. Kivisto, B. Sandstrom: Extrusion cooking of a high-fibre cereal product: I - Effects on digestibility and absorption of protein, fat, starch, dietary fibre and phytate in the small intestine. Br. J . Nutr. 55 (1986) 245-254. R.W. Chapman, J.K. Sillery, M.M. Graham, D.R. Saunders: Absorption of starch by healthy ileostomates: effects of transit time and carbohydrate load. Am. J. Clin. Nutr. 41 (1985) 1244-1248. T.M.S. Wolever, Z. Cohen, L.U. Thompson, M.J. Thorne, M.J.A. Jenkins, E.J. Prokipchuk, D.J.A. Jenkins: Ileal loss of available carbohydrate in man: comparison of a breath hydrogen method with direct measurement using a human ileostomy model. Am. J. Gastroenterol. 81 (1986) 115-122. L.J.D. ODonnell, P.M. Emmett, K.W. Heaton: Size of flour particles and its relation to glycaemia, insulinaemia, and colonic disease. Br. Med. J . 298 (1989) 1616-1617. N.W. Read, M.N. Al janabi, T.E. Bates, D.C. Barber: Effect of gastrointestinal intubation on the passage of a solid meal through the stomach and small intestine in humans. Gasrroenterol. 84 (1983) 1568-1572. A.M. Stephen, A.C. Haddad, S.F. Phillips: Passage of carbohydrate into the colon; Direct measurements in humans. Gastroenterol. 85 (1983) 589-595. B. FlouriC, A. Leblond, C. Florent, M. Rautureau, A. Bissalli, J-C. Rambaud: Starch malabsorption and breath gas excretion in healthy humans consuming low- and high-starch diets. Gastroenterol. 95 (1988) 356-363. C. Molis, M. Champ, B. FlouriC, P. Pellier, F. Bornet, P. Colonna, F. Kozlowski, J.C. Rambaud, J.P. Galmiche: Small intestinal digestibility of processed corn starches in healthy human subjects. Eur. J. Clin. Nutr.46 (1992) S131-SI32.

36.

L. Noah, F. Guillon, B. Bouchet, A. BulCon, D.J. Gallant, P. Colonna, C. Molis, N. Faisant, J.P.

44.

Galmiche, M. Champ: Digestion of carbohydrate components of dry beans (Phaseolus vulgaris) in healthy humans. European Association for Grain Legume Research Conference (1995), July 9- 13, Abstract. J. Tovar, I.M. Bjorck, N.G. Asp: Incomplete digestion of legume starches in rats: a study of precooked flours containing retrograded and physically inaccessible starch fractions. J. Nutr. 122 (1992) 1500-1507. I. Bjorck, M. Nyman, B. Pedersen, M. Siljestrom, N.-G. Asp, B.O. Eggum: Formation of enzyme resistant starch during autoclaving of wheat starch: studies in vitro and in vivo. J. Cereal Sci. 6 (1987) 159-172. J. Tovar, R.D. Duan, C. Erlanson-Albertsson, I.M. Bjorck: Starch digestibility in the diabetic rat. Nutrition research 11 ( 1991 ) 1329-1334. Y.H. Tuan, R.D. Phillips: Effect of the hard-to-cook defect and processing on protein and starch digestibility of cowpeas. Cereal Chem. 68 (1991) 413-418. Y.E. Grandfeldt, A.W. Drews, I.M.E. Bjorck: Starch biovailability in arepas made from ordinary or high amylose corn: concentration and gastrointestinal fate of resistant starch in rats. J . Nutr. 123 (1993) 1676-1684. M. Champ, F. Kozlowski, G. Dellavalle, J. Delort Laval: The rat as a model to evaluate RS digestion i n humans. XV International Congress of Nutrition, Adelaide, Australia, 9 October 1993. Workshop W540. H. Graham, W. Lowgren, P. Aman: An in vitro method for studying digestion in the pig. 2- Comparison with in vivo ileal and faecal digestibilities. Br. J. Nutr. 61 (1989) 689-698. K.E. Bach Knudsen: Methodological aspects of in vivo methods for measuring of starch digestibility: animal

45

models. In: 'Methodological aspects of in vivo methods for measurement of starch digestibility', Ed. E. Gudmand-Hoyer, Report of the European Flair Concerted Action Workshop (EURESTA, Contract no. AGRF/0027), Velbaek, Copenhagen, 10-12 November 1991,40-54. M.D. Levitt: Production and excretion of hydrogen gas in man. N. Engl. J. Med. 281 (1969) 122.127.

37. 38. 39. 40. 41.

42. 43.

Resistant starch 46. 47. 48

2 13

M.D. Levitt, P. Hirsch, C.A. Fetzer, M. Sheahan, A.S. Levine: H2 excretion after ingestion of complex carbohydrates. Gastroenferol. 92 (1987) 383-389. W. Scheppach, M. Bach, P. Bartram, S. Christl, W. Bergthaller, H. Kasper: Colonic fermentation of potato starch after a freeze-thaw cycle. Dig. Dis. Sci. 36 ( 199I ) I601 - 1605. E.K. Lund, I.T. Johnson: Fermentable carbohydrate reaching the colon after ingestion of oats in humans. J. Nirrr. 121 (1991) 311-317.

49.

50. 51.

52.

S.U. Christl, P.R. Murgatroyd, G.R. Gibson, J.H. Cummings: Production, metabolism. and excretion of hydrogen in the large intestine. Gastroenrerol. 102 (1992) 1269-1277. M.S. Casiraghi, F. Brigenthi, N. Pellegrini, E. Leopardi, G. Testolin: Effect of processing on rice starch digestibility evaluated by in vivo and in riirro methods. J. Cereal Sci. 17 (1993) 147-156. J.H. Cummings, H.N. Englyst: Measurement of starch fermentation in the human large intestine. Can. J. Physiol. Pharntacol. 69 (1991) 121-129. G. Livesey, T. I . Johnson, J.M. Gee, T. Smith, W.E. Lee, K.A. Hillan, J. Meyer, S.C. Turner: 'Determination' of sugar alcohol and polydextrose absorption in humans by the breath hydrogen (H2) technique: the stoichiometry of hydrogen production and the interaction between carbohydrates assessed in

53

54.

55.

56. 57

58. 59.

60.

61

62. 63.

vivo and in Litro. Eur. J. Clin. Nutr. 47 ( 1993) 4 19-430. N. Faisant, M. Champ, P. Colonna. A. Buleon, C. Molis, A,-M. Langkilde, T. Schweizer, B. Flourii, J.-P. Galmiche: Structural features of resistant starch at the end of the human small intestine. Eur. J. Clin. Nutr.

47 (1993) 285-296. R.L. Botham, P. Cairns, V.J. Morris, S.G. Ring: Physicochemical characterization of resistant starch in ileostomy effluent. Proceedings of the EURESTA Summing-up meeting, La Londe - Les Maures (F), April 1994, p. 34. H. Ekwall, A.M. Langkilde, I. Bjorck, N.G. Asp, H. Anderson: Digestion of starch - a comparison of amount and composition of RS recovered in vitro and in vivo from balance experiments in ileostomy patients. Proceedings of the EURESTA Summing-up meeting, La Londe - Les Maures (F). April 1994. p. 13. J.C. Mathers: Digestion of non-starch polysaccharides by non-ruminant omnivores. Proc. Nurr. Soc. 50 (1991) 161-172. J.C. Mathers, L.D. Dawson: Large bowel fermentation i n rats eating processed potatoes. British J . Nurr. 66 (1991) 313-329. J.C. Mathers: Energy value of resistant starch. Eur. J . Clin. Nutr. 46 (1992) S 129-S 130. G. Livesey, I.R. Davies, J.C. Brown, R.M. Faulks, S. Southon: Energy balance and energy values of a-amylase (EC 3.2.1. I.) resistant maize and pea (Pisum sarivum) starches in the rat. Er. J. Nurr. 63 (1990) 467-480. D.J.A. Jenkins, H. Ghofari, T.M.S. Wolever, R.H. Taylor, A.L. Jenkins, H.M. Barker, H. Fielden. A.C. Bowling: Relationship between rate of digestion of foods and post-prandial glycaemia. Diabetol. 22 (1982) 450-455. F. Bornet, A.M. Fontvieille, S. Rizkalla. P. Colonna. A. Blayo. C. Mercier. G. Slama: Insulin and glycemic responses in healthy humans to native starches processed in different ways: correlation with i n iirro a-amylase hydrolysis. Am. J. Clin. Nurr. 50 (1989) 315-323. H. Liljeberg, Y. Granfelt, I . BjGrck: Metaholic responses to starch in bread containing intact kernels wrsirs milled flour. Eirr. J. Clin. Nurr. 46 ( 1 992) 56 1-575 A.S. Truswell: Dietary and plasma lipids - A review. Ahstract in 'Topics in dietary fibre research', 5-7 May 1992, Rome, Italy, p. 14.

214 64.

6.5. 66.

67.

68. 69.

70. 71. 72. 73.

M. Champ arid N. Fuisant D.J.A. Jenkins, T.M.S. Wolever, R.H. Taylor, H.M. Barker, H. Fielden, J.M. Baldwin, A.C. Bowling, H.C. Newman, A.L. Jenkins, D.V. Goff Glycaemic index of foods: a physical hasis for carbohydrate exchange. Am. J. Clin. Nurr. 34 ( I 98 I ) 362-366. J.H. Cummings, H.N. Englyst: Measurement of starch fermentation in the human large intestine. Cau. J. Physiol. Pharmacol. 69 (1991) 121-129. S. Ranganathan, M. Champ, C. Pechard, P. Blanchard, M. N'Guyen, P. Colonna, M. Krempf Comparative study of the acute effects of resistant starch and dietary fibers on metabolic indexes in men. Am. J. Clin. Nutr. 59 (1994) 879-883. A. Raben, A. Tagliabue, N.J. Christensen, J. Madsen, J.J. Holst, A . Astrup: Resistant starch: the effect on postprandial glycemia, hormonal response and satiety. Proceedings of the EURESTA Summing-up meeting, La Londe -Les Maures (F), April 1994, p. 33. E. Sacquet, C. Leprince, M. Riottot: Effect ofamylomaize starch on cholesterol and hile acid metabolisms in germfree (axenic) and conventional (holoxenic) rats. Reprod. Nutr. Divelop. 23 (1993) 783-792. M. Champ, M. Riottot, F. Bornet, P. Colonna: Amylomai's traitis par cuisson-extrusion: digestihilitks totales et ilCales chez le rat; effet de ces amidons sur la holestCrolCmie. Gasrroeriterol. Clin. B i d . 14 (1990) A I SO. C. Morand, C. Remesy. M.A. Levrat, C. DemignC: Replacement of digestible wheat starch by resistant cornstarch alters splanchnic metaholism in rats. J. Nutr. 122 (1992) 345.354. J.M. Gee, R.M. Faulks, I.T. Johnson: Physiological effects of retrograded, a-amylase-resistant cornstarch in rats. J. Nutr. 121 (1991) 44-49. K.M. Behall, D.J. Scholfield, J. Canary: Effect of starch structure on glucose and insulin responses in adults. Am. J . Clin. 47 (1988) 428-432 L. Cara. C . Dubois, P. Borel, M. Armand, M. Senft, H. Portugal, A.M. Pauli, P.M. Bernard, D. Lairon: Effect of oat bran, wheat fiber and wheat germ on postprandial lipemia in healthy adults. Am. J. C h . Nufr. 55 (1992) 81-88.

74.

N. Faisant, M. Champ. S. Ranganathan, C. Azoulay, M.F. Kergueris, M. Krempf Effects of resistant starch supplementation on postprandial metabolism in healthy subjects. Symposium of human and animal nutrition, 13-14 January 1994, Tours, France. Abstract.

7s.

K.K. Caroll, R.M.G. Hamilton, M.W. Huff, A.D. Fallonner: Dietary fibre and cholesterol metabolism in rabbits and rats. Am. J. Clin. Nutr. 31 (1978) S203-S207. M.J. Wolf, U. Khoo, G.E. Inglctt: Partial digestibility of cooked amylomaize starch in humans and mice. Starch, 29 ( 1977) 40 1-40s.

76. 77. 78. 79. 80.

81.

G.M. Wyatt, N. Horn: Fermentation of resistant food starches by human and rat intestinal bacteria. J. Sci. Food Agric. 44 (1988) 281-288. H.N. Englyst, G.T. Macfarlane: Breakdown of resistant and readily digestihle starch hy human gut hacteria J. Sci. Food Agric. 37 (1986) 699-706. G.T. Macfarlane, H.N. Englyst: Starch utilization by the human large intestinal microflora. J. Applied Nutr. 60 ( 1986) 195-20 I I. Bjorck, M. Nyman, B. Pedersen. M. Siljestrom, N . 4 . Asp, B.O. Egguni: Formation of enzyme resistant starch during autoclaving of wheat starch: studies in vilro and in vivo. J. Cereal Sci. 6 (1987) 159.172. R.M. Faulks. S. Southon, G. Livesey: Utilization of a-amylase (EC 3.2. I . 1 .) resistant maize and pea (Pisum sarivum) starch in the rat. Br. J. Nutr. 61 (1989) 291-300.

Resistant starch 82

83.

84. 85. 86. 87. 88. 89. 90. 91. 92. 93.

94

2 15

J.H. Cummings, Ch. Edwards, J.M. Gee, F. Nagengast, J. Mathers: Physiological effects of resistant starch i n the large bowel. Final report of European FLAIR Concerted action on resistant starch (1995): Working group IIIB. G.H. Weaver. J.A. Krause, T.L. Miller, M.J. Wolin: Cornstarch fermentation by the colonic microbial community yields more butyrate than does cabbage fiber fermentation: cornstarch fermentation rates correlate negatively with methanogenesis. Am. J. Clin. Nutr. 55 (1992) 70-77. W. Scheppach, C. Fabian, F. Ahrens, M. Spengler, H. Kasper: Effect of starch rnalabsorption on colonic function and metabolism in humans. Gasrroenterol. 95 (1988) 1549-1555. M.A. Levrat, C. Remesy. C. Demignk: Very acidic fermentations in the rat cecum during adaptation to a diet rich in amylase-resistant starch (crude potato starch). J. Nurr. Biochem. 2 (1991 ) 31-36. C. Andrieux, D. Gadelle, C. Leprince, E. Sacquet: Effects of some poorly digestible carbohydrates on bile acid bacterial transformations in the rat. Br. J . Nutr. 62 (1989) 103-1 19. J.M. Gee, R.M. Faulks, I.T. Johnson: Physiological effects of retrograded. a-amylase-resistant cornstarch in rats. J. Nutr. 121 (1991) 44-49. A.K. Mallet, C.A. Bearne. P.J. Young, I.R. Rowland, C. Berry: Influence of starches of low digestibility on the rat caecal microflora. Br. J. Nutr. 60 (1988) 597-604. P.S. Shetty, A.V. Kurpad: Increasing starch intake in the human diet increases fecal hulking. Arm J. Clirr. Nurr. 43 ( 1986) 2 10-2 12. J.Tomlin, W. Read: The effect of resistant starch on colon function in humans. British J. Nictr. 64 (1990) 589-595. J.H. Cummings, E.R.Beatty, S. Kingman, S.A. Bingham, H.N. Englyst: Laxative properties of resistant starches. Gastroenterol. 102 (1992) A548.71 S.A. Bingham: Mechanisms and experimental and epidemiological evidence relating dietary fibre (non-starch polysaccharides) and starch to protection against large bowel cancer. Pror. Nurr. Soc. 49 (1990) 153-171. J.M. Gee, G.M. Wortley: Effects of resistant starch on intestinal structure and function. In: Intestinal Cell proliferation with Emphasis on Dietary Manipulation (Gee. JM & Nagengast F.M., eds.). EURESTA Physiological implication of the consumption of resistant starch in Man - Contract No. AGRF/0027; (1992) 32-34. P. Dysseler, D. Hoffem: Survey on resistant starch consumption in Europe. Proceedings of the EURESTA Summing-up meeting, La Londe - Les Maures (F) April 1994. Abstract p. 10.

This Page Intentionally Left Blank

12 Bulking agents: polydextrose

S.A.S. Craig', J.M. Andersonl, J.F. Holden' and P.R. Murray2 'Cultor Food Science R&D, Eastern Point Rd., Groton, CT 06340, USA 2Xyrofin (UK) Ltd., 41-51 Brighton Rd., Redhill, Surrey, RHI 6YS, UK

Summary. Polydextrose is a 1 kcal/g carbohydrate made by melt polymerization of glucose, sorhitol and citric acid. LitesseB and Litesse 1 1 8 are improved versions of polydextrose with reduced bitterness and also reduced acidity. Polydextrose has a variety of useful properties, including high water solubility, high glass transition temperature and good stability. Functionally, it can reduce calories by replacing sugars and partially replacing fats in food. Polydextrose can also act as a humectant and texturizer. The development and properties of the family of polydextrose products are discussed, including some new variations.

1. Introduction Interest in the relationship between diet and health continues to grow throughout the world. Reduction of calories and fat in food is a major driving force for development of bulking agents and fat-replacers. Today, over 90% of American consumers are eating lite foods and beverages.' Similar trends exist in most developed nations.2 Interest in fat, sugar, calories and fiber varies from country to country, driven by differences in culture and food regulations. The Cultor Food Science (formerly Pfizer Food Science) commitment to lite ingredients (bulking agents, high intensity sweeteners, and fat substitutes) has evolved over thirty years. The effort began in the 1960s with the development of polydextrose as a sugar substitute for incorporation in diet foods. Since commercialization in the early 1980s. polydextrose has demonstrated versatility as both a sugar- and partial fat-replacer in many foods. Caloric reduction occurs because polydextrose has an energy value of 1 kcal/g compared with fat at 9 kcal/g and sugar at 4 kcal/g. Polydextrose is a good example of the value-added technology derived from renewable organic raw materials.

218

S.A.S. Craig et al.

2. The development of polydextrose and Litesse 2.1 Discovery and Structure Polydextrose was invented by Dr. Hans Rennhard at Pfizer Central Research in the late 1960s while investigating the potential of polysaccharides as non-caloric and reduced-calorie replacements for sugar as well as partial replacements for fat, flour and starch.3 Oligomers were prepared by vacuum thermal polymerization of glucose using various catalysts and plasticizers. In an extensive program, optimum properties were exhibited by a polydextrose formed using citric acid catalyst, and sorbitol as plasticizer and chain terminator. The starting ratio of g1ucose:sorbitol:citric acid is 89: 10:1. Polydextrose is composed of randomly cross-linked glucose with all types of glucosidic bonds (1,6 bonds predominate).l.4 A representative structure is shown in Fig. 1. The R-groups may be hydrogen, glucose, sorbitol, citric acid or a continuation of the polydextrose polymer. In addition, small amounts of starting material (glucose, sorbitol, citric acid) remain. Levoglucosan and hydroxymethyl-furfural are produced as trace impurities via caramelization of glucose (Table 1). The representative structure is derived from characterizations that include methylation and periodate oxidation for degree of branching; Smith periodate degradation, methylation and acetolysis for linkage positions; Smith degradation and acetolysis for anomeric linkage configuration and "C-NMR.4 Polydextrose is more highly branched than other natural carbohydrates, such as the amylopectin component of starch, which contains mainly a-1,4

Fig. I . Representative structure for polydextrose.

Bulking agents: polydextrose

2 19

linkages with about 4-5% of ~r-1,6linkages as branch points. The structural compactness and complexity of polydextrose prevents mammalian digestive enzymes from readily hydrolyzing the molecule, imparting reduced caloric content. Thus, radio-tracer studies in both rats and humans found that polydextrose was utilized at about one-fourth of that of sucrose (i.e. 1 kcal/g).5 Recent studies have confirmed this value.6.7 Tahle 1: Polydextrose composition.

%

Polydextrose

>90a

Glucose Sorbitol

<2a

Levoglucosan

<4a

Water

<4

<4a

a Anhydrous, ash-free basis

Following extensive safety studies, U.S. FDA approval was obtained in 1981 for polydextrose FCC (Food Chemicals Codex).* Polydextrose is now approved in over 40 countries.

2.2 Litesse forms The original polydextrose is acidic (pH 2.5-3.5, 10% s o h ) and imparts a slightly tart flavor and bitterness in some foods. In order to improve polydextrose for food categories where these characteristics are undesirable, several improved versions have been developed. Litesse is produced from polydextrose using additional processing to reduce bitterness by removal of citrate esters, thereby improving the flavor profile. Litesse I1 is a further refinement. Table 2 shows a property comparison of polydextrose, Litesse and Litesse 11. The Litesse product line allows higher levels to be used, particularly in foods with delicate flavor. Flavor profiles are shown in Fig. 2 (polydextrose versus Litesse) and Fig. 3 (Litesse versus Litesse 11). Litesse is significantly less bitter, astringent, acidic and drying than polydextrose. Litesse I1 has a clean, mildly sweet flavor that increases the perception of richness and creaminess in many foods. Neutralization curves are shown in Fig. 4.In most food systems, Litesse will not require neutralization. Litesse I1 is recommended where sugar inversion or fat rancidity is a concern. Additional versions of Litesse in development include reduced and fat-coated counterparts. Reduced polydextrose (Litesse 111) is a pure white powder or a water-clear solution with excellent

220

S.A.S. Craig et al. Table 2. Comparison of polydextrose, Litesse and Litesse 11.

Taste Acidity (meq/g) pH (10%soh.) Typical usage levels

Polydextrose

Litesse

Litesse I1

Bitter, Tart, Acidic 0. I

Clean, Mildly Sweet 0.003

Strong

Bland, Neutral 0.03 >3.0 Medium 5-10% Manageable

High

Modest

Minimal

Yes

Sometimes

Rarely

>2.5

Low <5%

Extent of sugar inversion Extent of fat rancidity Neutralization

>4.0

High

>lo% Negligible

required

stability. It has substantially no reducing glucose groups on the molecule, has a bland flavor and a greatly decreased reactivity towards food ingredients having amine functionality. With residual glucose reduced to sorbitol, sugar-free claims are possible with this product. Litesse I11 is currently being marketed in Japan. Fat-coated Litesse is also being developed, particularly as a fat substitute to help reduce calories. We have found that fat-coating allows a number of functional advantages in food. icing sugar (0)

icing Sugar (F) Astringency (M)

0ilylp iastlc (F)

Bltter (F)

Fig. 2 . Flavor profiles of polydextrose vs. Litesse.

Bulking agents: polydextrose

22 1

lclng sugar (0) Oilylplastlc (0)

I

------L

-

Llterro

0 Odor

--

LHerre II

W

F Flavor

Mouthfeel

Fig. 3. Flavor profiles of Litesse vs. Litesse 11.

14 12 10 Lltuu

8

I,

Polydmxtroso

6

-

/

I *

0

5

10

15

20

25

mls 0.1N NaOH Fig. 4.Neutralization curves of LitessB 11, LitesseB and polydaexrose (20 grams to 200 grams with DI water).

30

222

S.A.S. Craig et al.

3. Properties

3.1 Solubility Polydextrose has higher water-solubility than most carbohydrates, allowing 80%+ solutions at

25 "C (Fig. 5 ) . This can provide a significant advantage in certain foods. It is insoluble in ethanol, but is partially soluble in glycerin and propylene glycol. Aqueous solutions are readily prepared, and concentrated solutions (>SO%) are best prepared by slowly adding polydextrose to heated water with efficient shear. Pre-blending with other water-soluble ingredients facilitates dispersion. Litesse is available in the following forms: powdered, granulated (for quicker dissolution and reduced dusting) and solution (70% and 80%).

100 - ...... ......

... ....... ....... ............. .......

1 II II iI .....

......

.............

.......

Fig. 5. Water solubility of bulking agents at 25 "C.

3.2 Molecular Weight Sorbitol limits the upper molecular weight during polymerization, thereby preventing the formation of water-insoluble materials. Recent studies with HPLC and light-scattering detectors gave a

Bulkirig agents: pol-vdextrose

223

molecular weight distribution shown in Fig. 6. The weight average molecular weight is -2000 (degree of polymerization -12) and the range of molecular weight is 162 to about 20,000.

1.6

1.2

2 0

- 0.8

2 U

0.4

0 100

1,000

10,000

100,000

M (g/mol) Fig. 6. Differential molecular weight distribution of polydextrose (Courtesy of Wyatt Technology).

3.3 Viscosity Polydextrose solutions behave as Newtonian fluids. Polydextrose viscosity is higher than that of sucrose or sorbitol at equivalent concentrations and conditions. For example, at 25 "C, 70% solutions of polydextrose, sucrose and sorbitol are about 1800 cP, 400 CP and 200 CP respectively. A contour plot (Fig. 7) relates concentration, temperature and viscosity for polydextrose solutions.

224

S.A.S. Craig et al.

50

55

60 65 70 75 Concentration (%)

80

85

Fig. 7. Contour plot of polydextrose viscosity (cP) at various concentrations and temperatures.

3.4 Humectancy

Polydextrose can act as a humectant in foods. Fig. 8 shows the moisture absorption of polydextrose powder. A moisture absorption isotherm (Fig. 9) shows the manner in which polydextrose controls water in foods. At low concentrations (below 60%), sucrose lowers water activity (Aw) slightly more than polydextrose. At higher concentrations (i.e., the water content found in most foods), polydextrose is more effective in reducing Aw. This is because sucrose can crystallize at high concentrations and these crystals do not interact with the water to lower Aw. Therefore, polydextrose can be used effectively to control Aw in foods.

Bulking agents: polydextrose

225

30

-

25

75%RH-

h

5 20 mln

u)

$ 15

.-

E

P, ._

2 10 5 23% R H

0

0

2

4 Time (weeks)

a

6

Fig. 8. Hygroscopicity of polydextrose at 25 "C

3.5 Melting properties and glass transition Polydextrose powder is an amorphous glass with an anhydrous glass transition temperature (Tg) of - 1 10°C.This is significantly greater than that of most other sugars9 (Fig. 10) and is partly a function of molecular weight. Heating above the Tg leads to a flowable melt which, after cooling, 1

0.8

N,

3

a

0.4

0.2

0

50

70

60

80

Carbohydrate (%) in water Fig. 9. Moisture adsoption isotherms.

90

100

S.A.S. Craig et al.

226

150

i . -YDEXTROSE ...

100

*

..... .

50 L

GLUCOSE

0

2 d

0

-0

e

DISACCHARIDES

I

'

FRUCTOSE

I - 0

IU L

-50

-100

100

200

500

1,000

2,000

Molecular Weight Fig. 10. Molecular weight (mw) vs. Tg (anhydrous) for low mw carhohydrates.

produces a clear glass with a brittle texture. The high Tg of polydextrose can be useful to help raise the composite Tg of a food. A food is in a 'rubbery' state when it's Tg is less than the storage temperature, and quality changes will occur over a short time (e.g. structural collapse, water migration, stickiness). When the Tg is above storage temperature, the food is in a 'glassy' state and is much more stable. Polydextrose is a mixture of branched molecules of varying molecular weight, so it will not crystallize. In fact, polydextrose can be used to stabilize foods by preventing sugar crystallization (e.g. hard candies).

3.6 Cryoprotectant

Polydextrose can protect the structure of frozen and thawed materials. This can be useful in applications such as frozen dairy desserts, frozen dough and surimi. Products stored in a freezer can undergo deleterious changes in texture (e.g. ice- and solute-crystallization, starch retrogradation), structure ( e . g . collapse, shrinkage, stickiness), and chemical composition (e.g. enzymatic activity, oxidation, flavorkolor degradation). Polydextrose can act by interrupting sugar recrystallization and starch retrogradation, providing structure, and raising the composite Tg' (the glass transition temperature of a maximally freeze concentrated solution). The Tg' values (where

Bulkirig agetits: plydextrose

227

ice can no longer form) of lactose (-28 "C), sucrose (-32 "C), fructose (-42 "C), glucose (-43 "C) rind sorbitol (-43.5 "C) are all lower than polydextrose (-24 "C).lo This means that replacement of these sugars with polydextrose raises the composite Tg' of a food. Freezer storage stability improves when the difference between Tg' and storage temperature (typically -18 "C for a home freezer) is minimized or if Tg' exceeds storage temperature.10

3.7 Freezing point depression This is an important factor in frozen products (e.g. ice cream and frozen dairy desserts). If the freezing point is too high, a product will be hard. If the freezing point is too low, the product will be too soft. Each product has its distinct requirements. Freezing point depression (FPD) is a colligative property (smaller molecular weight sugars have a greater depressing effect). Fig. 1 1 compares five carbohydrates and demonstrates how the larger molecular size of polydextrose has the least effect on freezing point.l1Il2Note that the sugars with the greatest FPD have the lowest Tg'.

0

-0.5

e

h

-f

2 3

-1

E

P

-1.5

0

2

4

6 8 10 Concentration (%)

Fig. I I . Freezing points of sugar solu~ions.

12

14

16

228

S.A.S. Craig et al.

3.8 Heat of solution The heat of solution of polydextrose is 8 cal/g, as measured by solution calorimetry (Fig. 12). Blends of polydextrose with polyols can produce an over-all cooling effect similar to sucrose.

20 1

lfi

Fig. 12. Heat of solution for hulking agents.

3.9 Stability Polydextrose powder and solutions are stable under normal storage and processing conditions. A recent study showed that dilute polydextrose solutions (5%) were stable from pH 4.5 to 6.0 at 70-100 "C for 5 hours followed by storage at -20 to 40 "C for 30 weeks.13 Severe treatment at low pH and high temperatures can cause slight depolymerization of polydextrose, producing glucose, but without significant lowering of viscosity.14

Bulking agents: polydextrose

229

3.10 Sweetness Polydextrose provides the bulk of sugars with little or no sweetness. Polydextrose is useful for applications where sweetness is undesirable and, when utilized with high intensity sweeteners, provides a sugar substitute with diminished caloric value.

3.11 Metabolism and caloric value Polydextrose is partially fermented by intestinal micro-organisms producing volatile fatty acids, which are absorbed and utilized. This utilization accounts for most of the 1 kcal/g caloric value for polydextrose. Since polydextrose is digested like soluble dietary fiber, it is used in fiber-fortified foods in a number of countries. Human toleration studies have shown that polydextrose is well tolerated. The mean laxative dose for polydextrose is 90 @day and a practical no-effect level with respect to toleration in most humans is 50 g/day.

4. Applications Polydextrose has been successfully incorporated into a wide variety of foods including baked goods, beverages, confections and frozen dairy desserts. Polydextrose is a functional ingredient which can play a variety of roles in formulations. Each ingredient in a food provides unique technical contributions and interacts with other ingredients. Properties such as freezing-point depression may be important in some foods but not in others. Identifying and understanding these important contributions is the first step to developing new and improved foods with polydextrose.

References I, 2. 3.

4. 5.

Consumer survey on the use of lite foods and beverages, conducted by Gallup for the Calorie Control Council, 1994. Pfizer Food Science Studies; Europe, Australia, Canada (1993) H.H. Rennhard. (to Pfizer Inc.): Polysaccharides and their preparation, US patent 3 766 165 (October 16, 1973). R.P. Allingham: Polydextrose - a new food ingredient: technical aspects. Chemistv of Foods and Beverages: Recent Developments. Academic Press, New York (1982) 293-303. S.K. Figdor and H.H. Rennhard: J . Agric. Food Chem. 29 (1981) 1181-1 189.

230 6. 7 8. 9. 10.

II 12.

13. 14

S.A.S. Craig et al. N. Juhr and J. Franke: A method for estimating the available energy of incompletely digested carbohydrates in rats. J. Nutrition. 122 (1992) 1425-1433. L. Achour, B. Flourie, F. Briet, P. Pellier, P. Marteau and J. Rambaud: Gastrointestinal effects and energy value of polydextrose in healthy non-obese men. Am. J . Clin. Nurr. 59 (1994) 1362.1368. 21 CFR 172.841, 46 FR 30080, June 5, 1981, Food Additive Petition No. 9A3441. L. Slade and H. Levine: Glass transitions and water-food interactions. Advances in Food arid Nurririori Research. Academic Press, San Diego (1994). L. Slade and H. Levine: Beyond water activity: recent advances on an alternative approach to the assessment of food quality and safety. Crir. Rev Food Sci Nurr. 30 (1991) I 15-360. R.J. Baer and K.A. Baldwin: Freezing points of bulking agents used in manufacture of low-calorie frozen desserts. J . Dairy Science 67 (1984) 2860-2862. R.J. Baer and K.A. Baldwin: Bulking agents can alter freezing. Dairy Field (1985) February. M . Beer, E. Arrigoni, D. Uhlmann, D. Wechsler and R. Amado: Stability of Polydextrose solutions to heat treatment and storage under acid conditions. Lebensm. Wiss. 11. Technol. 24 (1991) 245-251. S.A.S. Craig and F.M Kopchik: Unpublished results.

13 Alkyl polyglucoside, a carbohydrate-based surfactant

W. Ruback and S . Schmidt

Huls AG, D-45764 Marl, Germany

Summary. Alkyl polyglucosides comprise a new class of surfactants obtained by reaction of glucose with a fatty alcohol which is also a natural product. Continuous preparation processes are based in particular on the two-stage synthesis with butyl glucoside as intermediate. The best design of this two-stage synthesis and the important factors in process optimization are discussed in detail. The effects of the synthesis parameters on the composition and quality of the products play an important part in process optimization. Some examples of tests are quoted to illustrate the exceptionally favorable ecological and toxicological properties of the alkyl polyglucosides. Comparison of the surfactant properties with conventional surfactants shows that alkyl polyglucosides have a wide range of uses as surfactants. The synergism with conventional surfactants is considerable in some cases. The fact that derivatization of the alkyl polyglucosides leads to novel surfactants and emulsifiers is shown by the three examples of alkylglucuronic acids, APG sulphates and APG fatty-acid esters. Of these three derivatives, the APG fatty-acid esters have the best commercial prospects as potential emulsifiers.

Introduction

The annual consumption of synthetic surfactants (excluding soaps) in Western Europe is about 1.4 million tons'. The distribution over the various areas of use is shown in Fig. 1. The main use of surfactants is for cleaning purposes, i.e. in domestic and commercial detergents (49%) and industrial cleaners (5%). The cosmetics and pharmaceutical sectors require 1 1 % of the total. Two other large sectors are textiles and fibers and the chemical industry with 10% and 8%, of total surfactant consumption, respectively. The remaining 17% is distributed over uses in areas such as the leather and paper industries, paints, surface coatings and plastics, flotation, oil production and crop protection and pesticides. Given the fact that at least 54% of the surfactants used in Germany enter the domestic sewage system directly via the end user, it is evident that their environmental compatibility should be given high priority. Thus the use of natural-based raw materials for the production of detergents has become increasingly important in recent years2J.

232

W . Ruback and S. Schrnidl

Uses of surfactants (not including soaDs) Total annual consumption in western Europe about 1.4 million tone

V8rloua Crop prot8ctlon. p8atICld8a 0. flot8tlon. o i i f l d d chmnic8ia Pminta. coatings. Phatica % L88th.r. C 8 l l U l O l 8 and paP8r Induatry Bullding Induatry. m8t.l proc8aaing 3% Foodatuff Induatry 10% TOXtll8. 8nd fibre.

comm8rcl8i 8%

C h w n l c d induatry

lnduatrl8i cleaning

Fig. I . Uses of surfactants

In order to obtain the typical surfactant structure consisting of a hydrophilic and a hydrophobic part from natural-based raw materials, carbohydrates as the hydrophilic group are coupled with fatty-acid derivatives as the hydrophobic group. The most-abundantly available carbohydrates that can be used in surfactant syntheses are listed in Fig. 24.5. Because starch, sucrose, glucose and sorbitol are widely used as food applications, there is no doubt about their toxicological safety and environmental compatibility.

H

OH

H

L Starch

H ~

CH ,OH

CH ,OH

CH ,OH

HOH

OH

Glucose 5 000 000 t

OH o

~

c

H

OH

H

Sucrose 100 000 000 t

Fig. 2. Carbohydrates for surfactant syntheses.

2

0

$;;i HH

CH ,OH Sorbitol 650 000 t

Alkyl polyglucoside, a carbohydrate-based surfactant

233

To prepare surfactants from carbohydrates and fatty-acid derivatives, chemical conversions must be developed that are applicable on an industrial scale. These chemical conversions suffer from the multi-functionality of carbohydrates, their degradation under drastic reaction conditions, and their low solubility in organic solvents. Fig. 3 shows some carbohydrate-based surfactants that have been produced and used on an industrial ~cale~q-5.

H

CH,0CO(CH,)nCH3

Sorbitan fatty acid esters

OH

OH

H

Sucrose fatty acid esters

R - NCO(CHJ,CH, H 0

q

o

o

HW H z hCH3

HO H. OH CH,OH

Acyl glucamides

OH

n Alkyl polyglucosides

Fig. 3. Carbohydrate based surfactants.

Sorbitan esters5.6 are obtained by esterification of sorbitol with fatty acids. The annual world production is >10,000 tons. Sorbitan esters are lipophilic by nature but the hydrophilicity and water solubility can be increased by subsequent reaction with ethylene oxide. They are mainly used as emulsifiers and solubilizers with applications in the cosmetics, pharmaceutical and food industries. Sucrose fatty-acid esters7 are produced by transestrerification from sucrose and fatty-acid methyl ester. The obtained reaction products are complex mixtures, but since only the monoesters show favorable surface active properties, elaborate purification steps are necessary. The annual world production is only about 2,000 tons. Sucrose fatty-acid esters are used as specialty surfactants in the food, cosmetics and pharmaceutical industries. Fatty acyl glucamidesg are obtained by reductive amination of glucose in the presence of ammonia or short-chain primary amines followed by acylation of the glucamine with fatty-acid methyl esters. N-Methyllauroylglucamide has recently experienced a great increase in interest and is mainly used in dishwashing compositions. The environmental compatibility is reported to be very goodg. However, to date doubts about the toxicological safety have not been satisfactorily

234

W. Ruback and S. Schrnidr

disposed of because N-methylglucamine may form a nitrosamine which shows carcinogenic potential in experiments on ratslo. Alkyl polyglucosides are obtained by an acid-catalyzed reaction from glucose and (fatty) alcohol with elimination of water. This class of substances has been known for more than a century' I but economic syntheses suitable for production have only been available for a relatively short time. Like all acetals, the APGs are alkali-resistant and they hydrolyse under acidic conditions into glucose and alcohol. APG is not a single substance (Fig. 4) but a very complex mixture1*. In addition to the reaction of glucose with alcohol forming alkyl monoglucosides, free glucose can react further with these. Oligomers with up to twenty glucose units can thus be formed. The name "alkyl oligoglucosides" would therefore be more correct than the name commonly used: "alkyl polyglucosides". In addition, every single glycosidic bond can be in the a- or p-configuration. Fig. 5 shows the a-and p- isomers of dodecyl monoglucoside with some physico-chemical characteristics. The a/p isomerism, the pyranose and furanose configuration of the glucose unit, and the oligomerization of glucose units mainly via 1,4- and 1,6-linkages lead to a rapid increase in the number of isomers (Fig. 4).An alkyl maltotetrose can form as many as 1216 isomers. Glucose can also oligomerize under acidic conditions to give non-alkylated and non-surface active oligoglucoses.

OR

H, 0 OH

OH

R-OH: fatty, Ziegler-, 0 x 0 alcohol (C4- C p ) X 2 0,preferably 0, I , 2, 3 glycosidic linkage: a,p G-G-linkage: 1,4 ; 1.6 ; ....

X

Number of isomers

0 1

2 16

2 3

1216

I40

Fig 4. Alkyl polyglucosides: definition.

An important variable for characterization of an APG is the degree of glucosidationl3. It is defined as the ratio of the total molar amount of glucose incorporated in the APG to the total molar amount of incorporated fatty alcohol (Fig. 6), using the average molecular weight for fatty alcohol mixtures. By this definition, the degree of glucosidation of pure alkyl monoglucoside is 1 and that of pure alkyl maltoside is 2. The lowest possible average DP (= degree of polymerization) of an APG is 1. Most of the industrial prepared APGs have an average DP of 1.2-1.6.

Alkyl polyglucoside, a carbohydrate-based su factant DodecyI- a-D-glucoside

235

Dodecy 1- PD-glucoside

OH

M [a]? p

: 74.5 1 5 1 - 1 5 2 ' j

MP : 125127°C [a]:: : -23.5 CMC : 0.080molL

CMC : 0.072 mol/L

Fig. 5 . Comparison of dodecyl-a- and P-glucoside.

Fig. 7 shows a typical oligomer distribution for a C12/CI,-APG. The alkyl monoglucosides are the major component. The higher the degree of glucosidation of the individual components is, the lower their weight content in the mixture, approaching zero asymptotically. The weight distributions of the oligomers can be determined by HPLC and high-temperature gas chromatography. By means of TOF-MALDI-MS (time of flight matrix assisted laser desorption ionization mass spectroscopy). even the higher glucoside oligomers present in an APG mixture are seen. However, no quantitative information can be obtained since the response factors of the -

DP

=1c

moles of monoglucose in the APG Average molecular weight of the fatty alcohol : 196

moles of fatty alcohol in the APG 36

% by weight

When fatty alcohol mixtures are used : Average molecular weight taken into account.

pure alkyl monoglucoside pure alkyl maltoside pure alkyl maltotrioside

:

: :

DP = 1 DP.2 DP = 3

The lowest possible DP of an alkyl glucoside is 1 !

Fig. 6. Definition of the average degree of glycosidation.

1

2

3

4

6

DP of the individual components

Overall-DP

:

I23

Fig. 7. Oligomer distrihution of a typical APG (50 % active matter)

236

W. Ruback and S. Schmidt

individual constituents differ widely. Fig. 8 shows such a spectrum of a typical APG, containing components with a DP of up to 14.

L

,116

TOF-MALDI : lime of flight matrix assisted laser desorption ionization Fig 8. TOF-MALDI mass spectrum of an alkyl polyglucoside.

APG synthesis Fig. 9 presents industriallyLapplied synthetic routes for APGslS. Many other synthetic routes published thus far involve the chemistry of protection groups; these are of no industrial interest. Starting from starch it is possible to carry out simultaneous hydrolysis and glucosidation with butanol in the presence of an acid catalyst at about 150 "C (under elevated pressure). The result is the intermediate called butyl glucoside (abbreviated to "BG") which has no surfactant properties. BG can also be synthesized from glucose syrup and butanol with acid catalysis at lower

Alkyl polyghcoside, a carbohydrate-based surfactant

237

0

L

Starch (n = 100 - 1500)

Butyl oligoglucoside p 2 @ H

Iy&.-

H

O

OH

OH

m

Glucose

m = O - 10

Fig 9. Synthetic routes for alkyl polyglucosides.

temperatures under atmospheric pressure to give products that are considerably more pure and paler in color. The glucosidation is a typical equilibrium reaction, comparable to an esterification. Rapid removal of the reaction water, and a large excess of alcohol favors the formation of the glucoside. The butyl group is replaced by the longer fatty-alkyl chain during the reaction of BG with fatty alcohol. This "transglucosidation" is also an equilibrium reaction. A large excess of fatty alcohol and rapid removal of the liberated butanol by vacuum distillation improve the APG-yield. Besides via this two-stage synthesis, it is also possible to prepare APG in one stage from glucose and fatty alcohol. However, the glucose must be substantially anhydrous on addition of the acid catalyst, so that either relatively costly anhydrous glucose must be used or the water must be removed from the glucose in a preliminary drying step. The reaction is normally carried out in vacuo (which makes it easier to remove the water produced in the reaction) at 100-120 "C using a marked excess of fatty alcohol. The flow diagram (Fig. 10) illustrates the two-stage process used for preparing APG at Hiils. The glucose syrup used has a dextrose equivalent (DE) above 90. In the first stage butyl glucoside is prepared by reacting glucose with butanol. The water introduced with the glucose syrup and the reaction water are removed by azeotropic distilIation15. The second stage is the transglucosidation of BG with fatty alcohol, preferably a Cl2/CI, cut. The liberated butanol is removed by vacuum distillation and returned to the first stage16,17.18.The resulting APG, which comprises about 25% solid matter in fatty alcohol, is subsequently treated with base to neutralize the acid catalystlg. The excess fatty alcohol is removed by distillation20 and re-used in the transglucosidation. The resulting hot APG melt is mixed with water under pressure in the ratio 1: 1 and subsequently bleached. The final product is a pale yellowish, viscous paste.

238

W. Ruback and S. Schmidt

Bu

glucoalde

APQlFA

Synthesis

APQ

Distillat ion

I

1

H20

Final product

Bleaching

Fig. 10. Process diagram for alkyl polyglucosides (APG).

To produce butyl glucoside continuously, it proved advantageous to use a stirred reactor with an external heat exchanger followed by a reaction column'5 (Fig. 11). The glucose syrup is dosed directly into the reactor, which is equipped with a distillation column to continuously remove the azeotropic butanouwater mixture. The condensate is separated into a butanol phase and an aqueous phase. The butanol is returned to the column, and the water is discharged to the wastewater system. The mixture from the stirred vessel is passed to a reaction column where it is brought into contact with butanol vapor introduced at the bottom of the column. The wet butanol

I Fig. 1 1 . Contiuous butyl glucoside synthesis.

Alkyl polyglucoside, a carbohydrate-based surfactant

239

vapor leaving the top of the column is used to feed butanol into the stirred reactor. The product collected at the bottom of the column contains approximately 35% butyl glucoside and is passed on for the following transglucosidation. What are the advantages of butyl glucoside synthesis in a reaction column as compared to synthesis in a stirred vessel (Fig. 12)? The more-efficient removal of water shortens the residence time while lower reaction temperatures can be applied. Consequently the product obtained is considerably less colored, thus resulting in a more convenient bleaching step. The narrow distribution of residence times reduces the formation of undesired polyglucoses as a result of the minimal reaction time. Last but not least, a reaction column is less costly to set up and takes up less space than a stirred reactor with comparable throughput. These advantages of a reaction column become increasingly evident as the scale of a continuous BG synthesis increases. - More efficient removal of water at comparatively low

temperatures (countercurrent stripping with anhydrous butanol vapour) - Hence fast reaction with little thermal stress + pale-coloured

product - Narrower range of residence timescompared to a stirred vessel + less polyglucoses, less forming of coloured by-products - Can be implemented at reasonable cost and with space-saving Fig. 12. Advantages of the BG synthesis in a reaction column.

The properties of the final product can be controlled to a reasonable extent, especially in the transglucosidation stage, by optimization of the residence time and the ratio of starting materials. Increasing the excess of fatty alcohol results in higher monoglucoside content and consequently a lower degree of glucosidation (Fig. 13). At too-low fatty-alcohol excess - below about 4 mol of fatty alcohol per mol of butyl glucoside - the products have a very poor color quality and operation problems arise due to precipitation of polyglucoses. However, a very large excess of fatty alcohol - more than 10 rnol of fatty alcohol per mol of butyl glucoside - makes the synthesis less economic and causes a negligible change in the product composition. Fig. 14 shows the significant points in the optimization of the residence time during the transglucosidation. There is a clear maximum for the alkyl rnonoglucoside content in the product at a definite residence time. However, apart from the alkyl monogluco side content, attention must be paid to the residual BG and polyglucose contents. Polyglucoses are produced partly by breakdown of the alkyl monoglucoside if the residence time is too long. The ideal residence time for the

240

W. Ruback and S. Schmidt

DP 3,5

Alkyl monoglucoside

I

60

3 --

50

2 3 --

40

2 -1 3 --

+ I

1

3

4

+ I

I

-- 30

+ I

I

I

I

- 20

5 6 7 8 9 10 moles of fatty alcohol per mole of butyl glucoside

Fig. 13. Dependence of the alkyl monoglucoside content and the DP on the butyl glucoside : fatty alcohol ratio.

transglucosidation is thus inevitably a compromise between maximizing the alkyl monoglucoside content and minimizing the butyl glucoside and polyglucose contents. The product composition is greatly influenced not only by the starting material ratio and

Content (YO by weight)

irreversible for ma t ion of polyglucoses

-.- _

- - _ _- - - - - . . _ _ . . _ .

Residence time Fig. 14. Starting material, product and by-products in the APG reaction mixture as a function of the residence time.

Alkyl polyglucoside,a carbohydrate-basedsudactant

24 1

residence time but also by the catalyst concentration and temperature. At high catalyst concentrations and temperatures, the conversions are high with short residence times but the products have a high oligomer content - high DP - and are dark. In contrast, when the catalyst concentrations and temperatures are too low, an adequate conversion cannot be achieved with acceptable residence times. Fig. 15 illustrates the points which have been found to be particularly important in the development of the two-stage process for preparing APG. When attempting to optimize the continuous two-stage APG process it is particularly important that the thermal stress on the reaction mixtures in both stages be minimized. This results in less-colored products. In addition, effective prevention of excessive production of polyglucoses is necessary because they may interfere considerably with the process by precipitating. There has been extensive discussion on how these two conditions can be met. -

Reduction in thermal stress (less forming of colour, less bleaching agent needed)

- Reduction in olioglucose formation - Fatty alcohol removal both as effective as possible and not

deleterious for the product (+ very good vacuum) - Improvement of bleaching:

* * *

best possible utilization of the bleaching agent best possible colour stability on storage avoidance of decomposition products and odorous components

Fig. 15. Aims of the process optimization in contimuous APG synthesis.

The distillation to remove the fatty alcohol must also be carried out in a way which is not deleterious for the product. Advance neutralization of the acid catalyst in the solution of the alkyl polyglucoside in the fatty alcoholl9 is necessary in order to avoid decomposition and cleavage of the APG on removal of the fatty alcohol by distillation. The pH must be sufficiently high to rule out decomposition. However, an unnecessarily high pH itself adversely affects the color of the product so that experiments are necessary to find the best compromise. The subsequent multi-stage distillation20 involves initial removal of low boiling constituents - including water from neutralization - at moderate temperature and under relatively weak vacuum.

In the subsequent stages, the pressure is reduced and the temperature is increased. Most of the fatty alcohol is removed under very low pressure so that this step can be carried out at a comparatively moderate temperature which does not adversely affect the product. The resulting

242

W. Ruback and S. Schmidt

APG still loses quality if it remains in the molten state, above 130 "C, for too long. This is why it is mixed with water as quickly as possible for cooling down. Although the product is relatively pale in color, bleaching is necessary for stringent product specifications. The final peroxide bleaching of the crude aqueous product is an astonishingly complex step. The over-all product quality - and not just the color of the product - depends greatly on the conditions chosen for the bleaching. The particular aim of this is to obtain pale-colored products using the minimum amount of bleach. However, the color of the APG must remain unchanged even on lengthy storage. Effective prevention of the formation of decomposition products and malodorous components is also important. The one-stage process in which APG is synthesized directly from glucose and fatty alcohol, avoiding the butyl glucoside intermediate, is also suitable for the industrial scalel4. In this case, the actual synthesis normally takes place in batch, while working up and bleaching are usually continuous. The main disadvantage of the one-stage APG synthesis is that the starting materials are more costly. Instead of low-cost glucose syrups it is necessary to use distinctly more costly powdered glucoses, for example anhydrous dextrose or dextrose monohydrate, with only a low water content. In addition, it is difficult to obtain products with a low DP from the one-stage procedure. Although this is theoretically possible by using an even larger excess of fatty alcohol, it considerably reduces the space-time yield, which adversely affects the economics of the synthesis. 100

100

100

Sulphate. ash F a t t y alcohols

w

BUtYl mOf10~lUCOIlde8*o

Glucose

0H i g h e r

glucoildes

60

Alkyl msltosides Alkyl rnonoglucoaidea

40

20

0

O n e - i t a g e prsparstlon

Two- a tags p reparation

1.9 1,3

1.6 1,5 0.8 0,1 18.8 15,6 60,4

Sulphete. 8.h F a t t y alcohols B u t y l rnonoalucoaidea G~UCOIE H i g h e r glucosidea Alkyl maltorides Alkyl monoglucosldes

Fig. 16. Comparison of APG compositions.

0.1 27

13.2 56,5

Alkyl polyglucoside, a carbohydrate-based surfactant

243

Fig. 16 compares the compositions of APGs prepared in two stages and in one stage. The differences in composition tend to be small except for a higher oligomer content in the product from the one-stage preparation (27.0% rather than 19.9%).

Selected application properties of the alkyl polyglu~osides~~-25 The following table (Fig. 17) summarizes the principal surfactant parameters of APGs compared with conventional surfactants. As expected, APG greatly resembles the non-ionic fatty-alcohol ethoxylate in many respects such as CMC, surface tension, and wetting and foaming capacities in deionized water. This is particularly evident on inspection of the critical micelle concentration (i.e. CMC 1 for APG) as a function of the chain length (Fig. 18), where the plots for alkyl ethoxylate and APG are virtually identical.

C 12C14APG

C I ~ C ~ ~ ( E O ) ~ LAS H

C12C14(E0)2S04Na

33

550 35

180 34

200

5

10

2

Deionized water

53

45

8

42

13" GHa

110

45

13

55

Deionized water

240

230

560

620

13" GH4

90

230

330

580

cmc (ppm)

20

20

Surface tension

30

(mN/m) Viscosity (mPas)2 Wetting time (sec),

Foam ( m L ) 5

Fig. 17. Comparison of surfactant properties'. I@;

*)

100 g/L:

'1

Cotton discs: 4, German Hardness:

DIN 53902/1.

Now some of the properties of the APGs which are particularly important for their use as detergents and cleaners will be dealt with. Wetting capacity is a particularly important criterion for the efficiency of a surfactant. It is determined by measuring the time from introduction of a piece of test fabric into a surfactant solution to its sinking below the surface. Fig. 19 shows that the wetting times for fatty-alcohol ethoxylate and C,2/C14-APGin soft water are virtually identical. In hard water the wetting times increase with APG, unlike ethoxylate. This behavior is typical of anionic

244 100

,

W. Ruback and S. Schmidt 140

*

120 -

C,,,,APG (13"GH) ri;

100 -

80

-

Alkyl glucoslde 60 -

Alkyl ethoxylate

0.1 i 7

1 0'01 0,001

I

i

40

20

-c

1

6

8

10

12

14

0

-

' ~

0

03

1

1.5

2

Surfactant concentration (g/L)

Number of C atoma in the alkyl chain

Fig. 19. Wetting capacity as a function of the surfactant concentration (25 "C).

Fig. 18. Critical micelle concentration of surfactant solutions.

surfactants. For most applications this effect is not disadvantageous, as builder systems incorporated in detergents soften the water. Remission

60

(%I

60 40

30 20 10

0

Polyester

Test fabric Cc&14(EO)s-H + APG

C1,Clsaikylbenzenesulphonate + APG

Fig. 20. Washing effects of a C12Cld-APGLAS mixture (1:l) compared with C&14-APG/fatty alcohol ethoxylate (1:l). (Domestic drum washing machine, 30 "C, 7.5 g L ) .

Alkyl polyglucoside, a carbohydrate-based surjactant

245

Fig. 20 shows the results of textile washing tests on APG-containing surfactant mixtures. Various types of fabrics - cotton, polyester, and a blend of both - with standardized soiling were washed in a domestic washing machine at 30 "C with 1:l mixtures of C12/C14-APGand CI1/C13LAS, and of C ~ ~ / C I ~ - Aand P GC12/C14 fatty-alcohol ethoxylate. The reflectance of the test fabric was then measured to determine its whiteness. For all tested fabrics, the APG/fatty-alcohol ethoxylate mixture gave significantly better results than the APGLAS mixture. In this respect too, APG tends to behave like an anionic and not like a non-ionic surfactant. The advantageous properties of the alkyl polyglucosides, such as being tolerated well by skin and being highly synergistic on combination with an anionic surfactant, are particularly beneficial for use in manual dishwashing detergents. The synergism achieved by a mixture of APG N u m b e r o f plates

Commercial LDLS

0

20 80

40 60

60 40

8 0 FAES 100 20 C o s u r f a c t a n t

Anstett and Schuck test (34)

Fig. 21. Dish washing effect as a function of the composition. (Miniplate test 0.075 g/L, 13' German hardness at 50 "C).

246

W. Ruback and S. Schmidt

with fatty-alcohol ether sulphate in the mini-plate test is distinctly superior to the effect of commercial dishwashing detergents (Fig. 21). A good foaming capacity is of great interest for many applications of surfactants, especially in household and personal-care products. In this respect APG is as good as fatty-alcohol ethoxylate in soft water (Fig. 22). As the hardness of the water increases, there is no change in the foaming capacity of fatty-alcohol ethoxylates, while that of APG decreases significantly. In this respect too, APG behaves more like an anionic surfactant. The diminution in foaming can be instead of C12/C14. compensated for by using a short-chain fatty alkyl group, for example C~o/C12 So it is possible to formulate products with APG that show good foaming even with hard water.

Foam (mL)

1

700

600

500

400

300

200

100

0 0

I

03

1

13

Surfactant concentration (g/L) Fig. 22. Foaming capacity as a function of the surfactant concentration (25 *C).

2

AIkylpolyglucoside, a carbohydrate-based surjkctant

247

The selected examples show that APG has properties leading to numerous advantages for use in the suggested applications.

Environmental compatibility and toxicological safety of

APG22J6J7

Fig. 23 shows a number of requirements which must now be met by a surfactant, especially one being newly developed. In addition to economic criteria such as sufficient availability of low-cost raw materials and the existence of a suitable industrial process, other crucial factors are related to performance. High surface activity, synergism on combination with other surfactants, and suitability for a wide variety of uses are all important, as mentioned previously. However, nowadays particularly important factors are guaranteed environmental compatibility and toxicological safety. How does APG compare with other surfactants in this respect? - Sufficient supply of low-cost raw materials - Economic synthesis - High surface activity - Synergism on combination with other surfactants - Wide variety of possible uses - Confirmed environmental compatibility (biodegradability) - Toxicoloeical safetv. well tolerated bv skin

Fig. 23. Crucial criteria for a surfactant.

Since the Detergents Act requires that every organic ingredient in detergents and cleaners be completely biodegradable, novel surfactants must undergo thorough testing for biodegradability. Fig. 24 shows that APG is among the best of the known surfactants in this respect. The DOC decrease of 9597% means that alkyl polyglucosides can be said to be completely degradable. The

G to the approximately 15% lower figure for the degradation of C ~ ~ / C I ~ - AisP attributable branching in the synthetic alcohol. As do all surfactants by nature, alkyl polyglucosides also have a certain toxicity for aquatic organisms. This is comparable to that of other surfactants (Fig. 25). The lack of toxicity for non-marine organisms of the alkyl polyglucosides is evident from tests on rats (Fig. 26). Feeding tests on conventional surfactants show a 50% mortality at a dosage of 1 to 4 g per kg body weight. All the rats fed with APG survived, even with doses above 10 g k g .

W. Ruback and S. Schmidt

248

Surfactant C C

12/14 -alkyl

polyglucoside (nat. alcohol basis) 12/13 -alkyl polyglucoside (synth.. alcohol basis)

Alcohol ethoxylates Alkylphenol ethoxylates Lineary alkylbenzenesulphonate Fatty alcohol ether sulphates Paraffinsulphonate

DOC* - decrease (%) 95 - 97 91 - 9 4 80 - 85 75 80 67 - 95 90 - 95

Fig. 24. Biodegradability (OECD - Coupled Units Test, DOC' = Dissolved Organic Carbon)

Surfactant C 12/14 -alkyl polyglucoside Fatty alcohol ethoxylate (7 EO) Lineary alkylbenzenesulphonate Tallow fatty alcohol ether sulphate (3 EO)

LCso(mgW 5 3

4 9

Fig. 25. Toxicity for fish. (DIN 38412/15 Golden orfe test).

Surfactant -alkyl polyglucoside C 12/14 -alkyl polyglucoside C

10112

C 12/14 -fatty alcohol ethoxylate (7 EO) Lineary alkylbenzenesulphonate

Fatty alcohol ether sulphate

LD5grat (mg/kg) >I0 000

>I0 000 4150

650 - 2500 1000- 2000

Fig. 26. Oral toxicity. (OECD - Guidelines for testing of chemicals sect. 4, Method No. 401(1981).

The dermatological properties of the APGs can also be regarded as favorable. The skin irritation is usually less than with conventional surfactants (Fig. 27). The same is true for mucous membrane irritation as well. Because APGs are well tolerated by skin and mucous membranes and are non-toxic, they are distinctly more favorable than most conventional surfactants, so that they definitely have potential for use in cosmetics. It also appears justified to test the suitability of APGs as emulsifiers in food and pharmaceutical applications.

Alkyl polyglucoside,a carbohydrate-basedsurfactant

249

~

Surfactant type C C

10/1 I

C

12/14 -alcohol

-alkyl polyglucoside' 12/14 -alkyl polyglucoside] ethoxylate2 * 7 EO

C 1 1 -1ineary alkylbenzenesulphonate3 Fatty alcohol ether sulphate * 3 EO

Cutaneous irritation index

Evaluation

1.55 1.92

slightly irritant slightly irritant

3.1 - 5.0

moderately irritant moderately-

5.0 - 6.0 1.6 - 6.5

strongly irritant slightlystrongly irritant

Fig. 27. Dermatological properties. (Active matter content: 1 50%; 2 100%; 3 50%; 4 60%.). (* Test on rabbit skin; OECD-Guidelines for testing of chemicals Sec. 4;Method No. 405).

Derivatization of alkyl polyglucosides The polyol structure of the oligosaccharide moiety of the alkyl polyglucosides appears to be highly suitable for functionalization to modify the properties of the APGs in a specific manner. A wide variety of reactions is available to obtain a wide range of novel non-ionic and ionic surfactants (Fig. 28). The reactivities of the hydroxyl groups differ only slightly so that the selectivity of modification is usually low. Only the primary OH group in position 6 of the glucose is known to be more reactive than the secondary OH groups.

T

I = CH20(CH;?CH20)m-H CH ,OCH ,COO0 NaO

H

OH

R = alkyl radical n = O - 10

CH20P032-2Na0 CH20(CHdm-NH2 0 0 COO Na CH20S0: NaO CH 20-CO-R'

Fig. 28. APG derivatizations.

The only reaction in which a clear discrimination can be observed is the catalytic oxidation of alkyl polyglucosides28~29which takes place selectively at position 6 and leads to the formation of alkylglucuronic acids (Fig. 29). The reaction runs smoothly when metallic platinum is used with

250

W. Ruback and S. Schrnidl

oxygen or air as oxidizing agent and the pH is maintained in the region of 9 by adding sodium hydroxide. The extent of the conversion can be deduced directly from the amount of sodium hydroxide used.

Hr& - H<xR COONa

O2 NaOH I cat.

HO

H

HO

OH

Reaction conditions: - Dilute aqueous APG solution (5 - 20 %) - Pt-Catalyst (e.g. 5% Pt/C) - Oxidation with air or 0, - Constant pH 9,50 "C

H

OH

Results:

- Advantageous use of properties as for APG - Operating life and efficiency of the catalyst

inneed of optimization

Fig. 29. Catalytic oxidation of APG.

At first sight, a reaction of this type appears to be ideal. The reaction conditions are exceptionally mild, and no oxidizing agents which might lead to residues ( e . g inorganic salts) in the product are used. However, there are also difficulties that prevent the reaction from being carried out on a large scale. The space-time yields are usually unsatisfactory. The catalysts are efficient only with dilute solutions of the APG, about 10% strength. The considerable decrease in conversion at higher APG concentrations cannot be compensated for by raising the temperature, prolonging the reaction time or increasing the amount of catalyst. In addition, concentrating the dilute aqueous product solutions is not a trivial matter in view of the foam problems. The surfactant properties of the alkylglucuronic acids can be categorized as good. They substantially correspond to the properties of the original APG. However, there are no evident performance advantages over APGs so that it remains doubtful whether the expense of catalytic APG oxidation as an industrial process would be justified. Sulphation of alkyl polyglucosides30~3~ (Fig. 30) contrasts with catalytic oxidation in that it is very unselective, meaning that virtually all the available hydroxyl groups react. The reaction is normally carried out in a falling film reactor with APG in solution using an S03lnitrogen mixture. As APG melts above 120°C it is not possible to use it solvent-free, because this temperature is too high for the sulphation reaction. It is, of course, necessary to work water-free to avoid hydrolysis of the APG and forming of sulfuric acid. Only two possibilities remain. Either the APG is introduced into the reactor in an inert solvent which subsequently must be removed again. The solvent must also be sufficiently polar, resistant to SO3 and relatively volatile. These conditions are

Alkyl polyglucoside, u curbohydruie-basedsufuctunt

25 1

really met only by halogenated hydrocarbons which are now completely discredited. Alternatively, the APG can be sulphated in a 20-30% strength solution in the fatty alcohol, although this results in a mixture of a little APG sulphate and a lot of fatty-alcohol sulphate, which has no distinct performance advantages over pure fatty-alcohol sulphate. Additional factors are the fact that the synthesis is rather difficult to control and decomposition can occur easily. The possible difficulties become evident when one remembers what happens when a sugar comes into contact with sulphuric acid. It is therefore essential to minimize the time the reactive SO3 is in contact with the APG and to neutralize the mixture immediately after the reaction. Use of less-active sulphating agents such as S03/amine complexes give good results on a laboratory scale. This method is hardly suitable for an industrial process, however, because complete removal and recycling of the amine is very complicated. The biodegradability of APG sulphates is good, and they give good

HcxR HrjR

results particularly in textile washing tests. However, to date the synthetic difficulties as mentioned have deterred people from taking more interest in this class of substances.

1.

so,

2. NaOH

HO

H

OH

Reaction conditions: - Reaction of dissolved APG with S03/N2 in a falling film reactor - Immediate subsequent neutralization with NaOH - Possible: use of a APG in fatty alcohol

HO

H

OH

Results: - Good applicational properties - Synthesis not straightforward - Fatty alcohol as the preferred solvent (leads to FAS in the product mixture)

Fig. 30. Sulphation of APG.

APG fatty-acid ested2.33 are obtained by base-catalyzed transesterification of fatty-acid methyl esters with solid APG, which is substantially free of water and fatty alcohols, by removing the liberated methanol by distillation (Fig. 3 1). The reaction has low selectivity, although the primary OH group in position 6 of the glucose ring reacts somewhat more readily than the secondary OH groups. Since the original APG comprises a large number of components, the resultant mixture of substances is very complex. The hydrophilicity, or HLB, of the resulting products can be controlled within wide limits by the molar ratio of the fatty-acid methyl esters and APG starting materials. It is also possible to vary the fatty-acid residue. Optimization of the ratio of starting materials and the chain lengths makes it possible to obtain APG fatty-acid esters with good emulsifier properties particularly suitable for use in cosmetics. The products are compatible with

252

W. Ruback and S. Schmidt

skin and toxicologically safe, and they are based on completely natural raw materials. Thus, in contrast to many other emulsifiers, their synthesis does not involve any ethylene oxide. The waterin-oil cream formulations of the products have an advantageous consistency and a pleasant feeling on the skin. Many of the formulations are stable on long-term storage even without addition of a co-emulsifier.

HrxR - HrJR Fatty acid

methyl Base ester - CH 3 0 H

HO

H

HO

OH

H

R =C / C

6.

OCOR'

RCOOH = fatty acid

Properties: - Cosmetic emulsifiers produced without EO for W/O

systems based on renewable raw materials - Compatible with skin, toxicologically safe - Formulation with excellent consistency and

pleasant feeling on the skin - Coemulsifier unnecessary - Exceptionally high long-term stability of the emulsions

Fig. 3 I . Alkyl polyglucoside fatty acid esters.

Of the APG derivatives discussed, those with the best prospects are therefore the APG fattyacid esters. These substances have the potential of becoming a new class of widely-used cosmetic emulsifiers.

References I. 2.

B. Fahry; SOFW 120, 377 (1994) W. Ruback; Chiniica oggi 1994, 15

3. 4. 5. 6.

F. Lichtenthaler; Narhr. Chon. Tech. Lab. 38, 860 (1990) H. Koch, R. Beck, H. Roper; Srarke 45,2 (1993) M. Biermann, K. Schrnid. P. Schulz; Stiirke 45. 281 (1993) M. Biermann, F. Lange, R. Piorr, U. Ploog, H . Rutzen, J . Schindler, R. Schniid; Synthesis of surfactants, in: Surfactants in Consumer Products. Ed. J. Falbe. Springer Verlag, Heidelherg 1987, pp. 23-1 32 N.B. Desai, N. Lowicki; Parjiini Kosniet. 64, 463 (1983) H. Kelkenberg; Tenside Su$. Det. 25, 8 (1988) M. Stalmanns. E. Matthijs, E. Weeg, S. Morris; S O F W 119, 749 (1993)

7.

8. 9.

Alkyl polyglucoside, a carbohydrate-based surjiuctunt 10. I I. 12. 13.

ACS Monogruph 182 - Chemical Carcitiogens. Washington, DC, 1984, Vol. 2, p. 755 E. Fischer, Ber. 26, 2400 ( I 893) H. Luders, D. Balzer; Congress Reports of the 2nd World Su~$actuntCongress, Paris 1988. Vol. 1 p. 8 I P.A. Siracusa; HAPPI 24, 100 (1992)

13.

P. Schulz; Chimicu oggi 1992, 33 A. Oherholz. J . Kahsnitz, S. Schmidt; EP 0482 325 (1991) H. Luders; EP 0249 013 (1987) N. Ripke; EP 0501 032 ( I 99 I ) D. Balzer, N. Ripke; EP 056Y 682 ( 1993) N. Ripke; EP 0521 258 (1993) B. Muller. N. Ripke; EP 0531 647 (1993) D. Balzer, Tenside Surf Der. 28, 4 19 ( I 99 1 )

15. 16. 17. 18. 19.

20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

J. Knaut. G. Kreienfeld, Chimica oggi 1993, 41 D. Nickel, C . Nitsch, C.-P. Kurzendorfer, W. von Rybinski; Henkel-Referare 28. I I (1992)

9.-D. Holdt, P. Jeschke. R. Menke, H.-D. Soldanski; SaFW 120, 42 (1994) P. Busch, H. Hensen. M. Tesmann; Tenside Surf: Der. 30, 116 (1993) H. Kelkenberg. lecture ("Carbohydrate-based surfactants") at the 3671h DECHEMA colloquium on 22.03.1990 H. Andree, 9. Middelhauve; Tenside Surf Det. 28, 413 (1991) T. Bocker. diploma thesis, Munster University, Prof. J. Thiem's group, pp. 27-33 (1987) N. Ripke; E P 0326 67.? (1988) N. Ripke; EP 0363 601 (1989) T. Bocker, T. Lindhorst, J. Thiem. V. Vill; Curbohydr. Res. 230, 245 (1992) B.W. Lew, DE 2360367 (1973) C.C. Akoh, B.G. Swanson; JAOCS60, 1295 (1989)

253

This Page Intentionally Left Blank

14 Tailor-made carbohydrate surfactants? Systematic investigations into structure-property relationships of N-Acyl N-Alkyl 1-Amino-1-Deoxy-D-Glucitols.

H.A. van Doren Netherlands Institute for Carbohydrate Research TNO (NIKO TNO) Rouaanstraat 27,9723 CC Groningen, The Netherlands

Summary. This paper describes the research carried out at NIKO TNO aimed at influencing and understanding several important parameters of nonionic carbohydrate-derived surfactants and the construction of models for predicting the behavior of novel derivatives. Our research has focused on N-acylated alkylamino-D-glycopyranosidesand (N-acyl) N-alkyl I-amino-I-deoxy-Dalditols. Parameters that have been studied include:

+ + + +

the solubility of surfactants in water (Krufr temperature) the type of aggregate formed in aqueous solution the critical aggregate or micelle concentration (c.a.c. or c.m.c.) the biodegradabi1ity/toxicity of compounds.

The molecular structure of surfactants can he manipulated in a number of ways in order to lower the Krufr temperature. A novel technique for the quantitative determination of the Krufr temperature is described. The type of aggregate formed in dilute aqueous solutionlsuspension of nonionic carbohydrate-derived surfactants can he predicted accurately with the aid of a straightforward model. The total number of carbon atoms in the alkyl chain(s) of the surfactant can he plotted against its number of hydroxyl and carbonyl groups. The position in the plot that results for a specific compound provides the type of aggregate most likely formed in aqueous solution. The critical aggregation concentration can he lowered by lengthening the alkyl chain. Generally, the c.a.c. is lowered by a factor of 10 when the alkyl chain length is increased by two methylene units. All compounds studied are at least partially biodegradable (within the limited time span of the tests that were used). Preliminary results indicate that the toxicity of surfactants increases strongly with increasing alkyl chain length, in some cases reaching a maximum around alkyl chain lengths of C,2-C,4.

1. Introduction Currently, most detergent components are manufactured from non-renewable sources. Assuming that detergents will continue to be used in the same fashion (recycling or catalytic processes will

256

H.A. van Dorm

not be considered here) and will therefore end up in the environment after use, one can make a list of requirements for components in future detergent generations:

+ + + + +

inexpensive materials excellent price/performance ratio efficient, 'clean' manufacturing processes all materials must be based on renewable sources (=biomass), e.g. carbohydrates, vegetable oils all materials must be biodegradable, non-allergenic and nontoxic.

The suitability of carbohydrate-derived materials for replacing a number of detergent ingredients has been the subject of numerous studies over the past decades. Several promising results have been obtained, some of which will be discussed in considerable detail in other chapters of this book. Rijper e f al. 1 have recently reviewed the potential for carbohydrate-derived materials as ingredients for laundry detergents. According to these authors, it is feasible to substitute SO-75% of all detergent components for carbohydrate derivatives. Co-builders are added to laundry detergents to assist zeolites in binding calcium and magnesium ions and to prevent precipitation of insoluble salts thereof on clothes. They usually consist of polycarboxylates of the polyacrylate type. Carbohydrates can be oxidized quite readily to yield derivatives with moderate to excellent builder performance. Examples' include glucaric acid, citric acid (obtained microbiologically from glucose), tricarboxysucrose, maltobionic acid, dicarboxy starch2 and dicarboxy inulin3. Bleaching activators usually consist of polyacetylated amines, such as tetra-acetyl ethylenediamine (TAED) or nitrilotriacetic acid (NTA). Polyacetylated mono- and oligosaccharide derivatives are interesting alternatives'. Compounds such as D-glucose pentaacetate, hexa-0-acetyl D-glucitol, nona-0-acetyl D-maltitol and 'SUPA'4, which is a mixture of poly-0-acetyl sucrose compounds, have been evaluated. The replacement of surface-active components of detergents by carbohydrate derivatives has already been implemented on a modest scale. Compounds which are already in use include alkylpolyglucosides5 ('APG's'), (ethoxylated) sorbitan fatty esters6 ('Tweens' and 'Spans') and sucrose fatty esters'. The introduction of other products, such as 6-0-acyl 1 -0-alkyl-a-Dglucopyranosides8, N-acyl N-methyl 1-amino- 1-deoxy-D-glucitols9 ('MEGA'S') and N-alkyl lactobionamides1° is still under consideration. At NIKO we have focused on the synthesis and characterization of new carbohydratederived surfactants. The most commonly-encountered strategy for research into new surfactants in industrial laboratories is, by necessity, that of trial and error. Research at our institute is performed

Tailor-made carbohydrate suqactants?

257

on a precompetitive level. Therefore, we have been able to carry out systematic investigations into structure-property relationships by variation of individual structural parameters. A thorough evaluation of the results obtained should enable us to devise models for the prediction of

+ + + +

the nature of the aggregate formed in aqueous solution the solubility of new products (Kraffr temperature) the critical aggregate concentration the biodegradability/(eco)toxicity. These parameters will all be discussed in the next sections, preceded by a section on the

general synthesis of the types of aminosurfactants used in this study.

2. General synthesis of aminosurfactants Carbohydrate-derived surfactants with amino or arnide functions are well documented’ I . There have been some reservations with respect to large-scale application of surfactants containing

OH OH

-1

NaBH4 or H2, cat.

*

HO+

“(CH&,CH3 6H

Scheme I . General synthesis of glucose-hased aminosurfactants.

H

3

258

H.A. van Doren

nitrogen. Compounds containing primary or secondary amino functions are susceptible to nitrosylation, thus yielding N-nitroso amines, which are suspected mutagens. This problem is eliminated when nitrogen is present in the form of an amide. Several amide-containing carbohydrate surfactants are currently under consideration for large-scale application (e.g. the N-alkyl lactobionamides and the MEGA series, mentioned earlier). Scheme 1 shows the general synthetic methodology with glucose as the starting material. The types of compounds we have concentrated on are the (acylated) alkyl 1-amino-D-glyco-pyranosides(1 and 2) and the (acylated) I -alkylamino- I-deoxy-D-alditols (3 and 4). Suitable starting materials for the synthesis of these products are readily available reducing mono- and disaccharides, such as glucose, maltose and lactose. All reactions were carried out in ethanol or methanol, usually at room temperature'*ql3. Yields are generally good to excellent. Compounds 1.are not hydrolytically and thermally stable, the other products can be stored at ambient temperature over long periods of time.

3. Model for the prediction of the type of aggregate in dilute aqueous solution Surface-active agents (surfactants) are amphiphilic species, i.e. the molecules consist of a hydrophilic part (usually called headgroup) and a lipophilic part (in most cases an alkyl chain or a fatty acid residue, the so-called tail). The behavior of the molecule is very much dependent on the relative sizes of the headgroup and the tail, the so-called hydrophilic-lipophilic balance or HLB. There is a clear correlation between the HLB-value of an amphiphile, its solubility in water and its potential applicati~n'~. This correlation is shown in Table 1. Table 1 . Relationship between solubility in water, type of application and HLB-value.

Solubility in water

Type of application

HLB-value

poorly soluble

water-in-oil emulsifier

3-6

milky dispersion

wetting agent

6-8

stable dispersion

wetting agent and

8- 10

oil-in-water emulsifier translucent/clear solution

oil-in-water emulsifier

10-13

clear solution

detergent

13-15

Tailor-made carbohydrate surfactants?

259

In principle, the HLB value of carbohydrate-derived amphiphiles can be calculated directly from the chemical formulae. According to Davies" the following equation may be used: HLB = 7 +

(hydrophylic group numbers) - x(lipophi1ic group numbers)

in which the group numbers have been determined empirically. A number of relevant group numbers, taken from Davies and Rideallsb can be found in Table 2. The values thus obtained for sorbitan monoesters were found to be in good agreement with those determined experimentally. Miethchen and Peters have calculated the HLB values of a number of alkyl and acyl glycosides with this equation16. To our knowledge, the group numbers for amine and amide functions have not yet been determined. Table 2. HLB group numbers.'5b

Hydrophilic

group number

group

Lipophilic

group number

group

-COO- K+

21.1

-CH-

.475

-COO- Na+

19. I

-CH,-

,475

-COOH

2.1

-CH,

.475

free ester

2.4

=CH-

.475

ring ester

6.8

phenyll6

-OH (free)

1.9

-OH (ring)

0.5

-0-

1.3

1.663

Our experience with the contact preparation method" prompted us to evaluate the following, more practical approach. When a small quantity of surfactant is melted on a microscope slide (with cover slip) and then allowed to resolidify, a smooth surface is obtained. Water is subsequently added near the edge of the cover slip and the sample is immersed in water through capillary action. Above the Kraff? temperature (see section 4) water penetrates the sample and usually one or more lyotropic liquid-crystalline phases are formed as a function of increasing surfactant concentration. It is assumed that lyotropic liquid-crystalline phases consist of two aggregation levels' 8 . Amphiphilic molecules aggregate into micelles/vesicles above the critical aggregation concentration. When the concentration of the aqueous solution is increased, the micelles/vesicles can aggregate into structures with significantly higher order, i.e. lyotropic mesophases. Therefore, the nature of the mesophase at the interface between the aqueous solution and the surfactant

260

H.A. van Doren

provides a strong indication for the nature of the aggregates found in dilute solution. A normal hexagonal mesophase (HI) at the interface with water is indicative of (worm-like) micelles in solution, a fluid lamellar mesophase (La) corresponds with vesicular aggregates and an inverted hexagonal mesophase (HI,) consists of worm-like inverted micelles. We applied this principle to several dozens of amphiphilic polyhydroxy compounds and plotted the results in a graph (Fig. 1). The X-axis depicts the total number of carbon atoms in the alkyl chain(s) of the surfactant and the Y-axis depicts the total number of hydroxyl and carbonyl groups in the headgroupI9.

0 II

+

7 -

A

6 -

AAAAAAA

A

00'

0

I 5 0 4 -

A 00'0 0

.*

0

000

AAAL),~O 0

3 -

0

2 -

,

,,boo

0

0

000000

0°

0

0

.OJ-VV bv .

0

0 .*

.-.-

0 .

.-

V

.We* .I

c O ) O

1 -

0

A

A

+ 0

z

,

0 0 0

0 .

'

~

~

'

'

'

1

'

'

"

'

'

'

1

~

'

'

'

'

~

'

~

~

~

'

'

'

1

1

'

V = normal hexagonal (HI) phase, (worm-like) micelles in solution. 0 = fluid lamellar (La) phase, vesicles in solution.

A = inverted hexagonal (HII)phase, corresponding with inverted micelles. 7 = La at first; HI, at higher T. Fig. 1. Correlation between the total number of hydroxyl/carbonyl groups, the total number of carbon atoms in the alkyl chain(s) and the lyotropic mesophase at the interface between the amphiphile and hulk water.

Fig. 1 shows that these two readily available parameters, i.e. alkyl chain length and number of hydrophilic groups, allow us to predict fairly accurately what type of aqueous aggregates a new compound will form and therefore what its potential application profile will be. Obviously, our approach is qualitative. Surfactants which form cubic mesophases at the interface with water were not included, as we were unsure of the nature of the aggregates from which they are formed. Moreover, the stereochemistry of the headgroups, the nature of the hetero-atom, and a-or

~

'

'

Tailor-made carbohydrate surfactants?

26 1

p-orientation at the anomeric center have not been taken into account and may have some influence. A striking example of the importance of stereochemistry was observed for octyl I-thioa-D-talopyranoside20. This compound forms vesicles in aqueous solution, whereas the corresponding glucose and galactose derivatives form micellar solutions, as would be expected from Fig. 1. This means that the effective size of the headgroup in the talopyranoside is much smaller (less hydrated) than in the gluco- or galacto-pyranosides. This is presumably due to the presence of an intramolecular hydrogen bond, which persists in aqueous solution21 (fewer hydroxyl groups available for hydration).

4. The Krafft temperature. Introduction Surfactants are only effective when their solubility is sufficient to form aggregates (micelles, vesicles) in aqueous solution. The temperature at which the solubility of a surfactant equals the critical aggregate concentration is called the Krafft temperature (TKrafft).It has been observed that the solubility of surfactants increases dramatically above TKrafft. The Krufft phenomenon can be interpreted in different ways. Essentially, there are two theories: 1.

2.

Below TKrafftthe crystalline surfactant is in equilibrium with 'monomeric' surfactant molecules in solution. When the solubility of a surfactant reaches the critical micelle concentration (TKrafft) a new situation arises, because the surfactant molecules can now aggregate into micelles, thus providing a much greater solubility (Moroi22, see Fig. 2a). TKrafft signifies the melting point of a crystalline hydrate of a surfactant. This means that a phase transition occurs from (hydrated) solid + H 2 0 to the aqueous solution in which the liquefied surfactant is dissolved/dispersed (Shinoda et ~ f . ~ For ~ ) .micelle-forming surfactants, the curve of the c.m.c. intersects the very steep solubility curve just above TKrafft(see Fig. 2b).

We studied a large number of nonionic and ionic surfactants with TKraffthigher than room temperature by means of differential scanning calorimetry (DSC)24.Invariably, a phase transition was found at a temperature, which corresponds very closely with TKrafftobserved by the classic solubility method. Fig. 3 shows a DSC-curve, the surfactant in this specific case being N-butanoyl N-dodecyl 1-amino- 1-deoxy-D-glucitol (4, m=2,n= 1 1). A phase transition is also found for surfactants that form lamellar or inverted aggregates. In this case, the solubility does not necessarily increase dramatically above the phase transition, however, water does penetrate the

262

H.A. van noren

- solubility

----

c.m.c.

I

Temperature ("C)

- solubility

----

c.m.c.

Temperature ("C) Fig. 2. The solubility and cmc-curves of a surfactant in water as a function of temperature; (a) The Kraflphenomenon according to Moroi, (b) The KrafSr phenomenon according to Shinoda.

Tailor-made carbohydrate sudactants ?

0

20

40

60

80

263

100

Temperature ("C) Fig. 3. The DSC-curve of a sample of N-butanoyl N-dodecyl I-amino-I-deoxy-D-glucitol (heating, rate 5 K m i n - I )

sample (cf. Tpenas defined by Lawrence25). Moreover, the temperature at which the phase transition takes place was found to depend on the aggregation state (crystal, crystalline hydrate, supercooled mesophase) of the bulk of the surfactantZ6.Consequently, our observations strongly favor Shinoda's approach. We have found that micellar solutions will eventually precipitate in all cases studied so far when stored at a temperature below TKrafft; this observation is also explained more easily from Theory 2 . Therefore the following slightly modified definition for TKrafftis proposed: 'ITKrafft is the melting point of an amphiphile immersed in excess water"27.

Influencing TKrafftin carbohydrate-derived surfactants TKrafft is highly dependent on the molecular structure of a surfactant; it is primarily determined by the strength of the crystal lattice of the solid surfactant (anhydrous or crystalline hydrate). Monosaccharide derivatives with one alkyl chain are usually rod-like molecules. They tend to crystallize in bimolecular layers with head-to-head carbohydrate moieties and interdigitized alkyl chains28. The resulting crystal lattice is fairly strong, which results in a rather high TKrafft. Within one homologous series, TKrafft usually increases with increasing chain length; as a general rule the

264

H.A. van Dorm

increase will be ca. 10 "Cfor every two additional methylene units in the alkyl chain29. A rather strong odd-even effect may be observed in some cases", similar to the odd-even effect which is sometimes observed in the melting points of homologous series31 of surfactants. TKrafftcan be lowered by:

+ +

forcing the molecules in a non-linear conformation or structure (by manipulating the headgroup or the location and mode of attachment of the alkyl chain) synthesizing branched molecules (alkyl chain branching, lateral substituents).

The following examples may serve to illustrate these principles: We have synthesized a series of N-decyl 1 -amino- I-deoxy-alditols, These products were prepared by coupling the appropriate monosaccharide to 1-aminodecane through reductive amination (NaBH4). The melting points, enthalpy of melting and TKrafft were determined for these compounds and were found to vary significantly as a function of the stereochemistry of the headgroup. The differences were especially apparent for the pentitol derivatives (see Scheme 2) and can be explained as follows: the energetically most favorable conformation of these acyclic derivatives is an extended 'ullrruns' conformation. This leads to rod-like molecules, which can conveniently pack into dense crystals possessing a relatively high TKrafft.In the case of the D-ribitol and the D-xylitol derivatives, the hydroxyl groups on C-2 and C-4 are both on the same side of the molecule. In a fully extended conformation, this will lead to an energetically unfavorable steric hindrance (l,3-syninteraction). Therefore, the molecule will twist out of plane, towards a sickle-shaped headgroup. As a consequence, the crystal packing is less than optimal, leading to a lower melting point, a lower enthalpy of melting and a lower TKrafft. A straightforward example of the influence of changing the location of attachment of the alkyl chain was described by Dahlhojf32, who compared the 4-0-alkyl-D-glucitols with the I-0-alkyl-D-glucitols and found that the former have substantially lower melting points. TKrafft was not determined for these compounds, but in recent studies33 it was found that the structurally related 3-0-alkyl-D-glucitols do have relatively low Krafft temperatures. The introduction of branched alkyl chains is a classic method for the lowering of TKrafft,but it has also been found that surfactants with branched alkyl chains are significantly more resistant to biodegradation than their linear counterparts. However, in the case of the alkyl 1-amino-1-deoxy-D-glycitols, there is another way of introducing branching in the surfactant molecule, by simply attaching a second substituent on the nitrogen atom. Scheme 3 shows the synthetic pathways for two N-methyl derivatives. The influence of such a lateral substituent on the melting points and on TKrafft is rather dramatic, as is shown in Table 3.

Tailor-made carbohydrate sutjktants ? n-decyl 1-amino- 1-deoxy- D-lyxitol m.p. 102 "c (AHmelt = 52 kJ.mol-1)

QH H o y r 0 ( C H & C H 3 6H

6H

265

k a f i ? 57

H H

"c

n-decyl 1-amino-I-deoxy- D-arabinitol m.p. 135 "C (AHmelt = 69 kJ.mol-1)

H HI

TKrafi? 98°C

Scheme 2. The influence of 1.3-syn-interactions on the melting points and on the KruD temperatures of four N-decyl 1-amino-I-deoxy-D-alditols.

1

CH3(CH2)&(0)OMeb MeOH, 'OCH?, A -.

H

0 OH

MEGA-10

Scheme 3. Syntheses of N-methyl N-decyl I-amino- I-deoxy-D-glucitol and MEGA-I0

266

H.A. van Doren Table 3. The influence of small lateral substituents on the melting points and KrafSl temperatures of 1-amino- 1-deoxy-D-glucitol derivatives.

Compound

Melting Point ("C)

TKraffi

("'1

3 (n=9)

124

83

5

82

44

4 (m=O, n=9)

44

< 20

MEGA-10 (4, m=8, n=O)

80

35

These results prompted us to study three homologous series of (N-acyl) N-alkyl 1-amino-l-deoxyD-glucitols (see Scheme 1) with alkyl chain lengths ranging from n-heptyl to n-hexadecyl and, respectively, no lateral substituent (3),an N-acetyl group (4, m=O) and an N-propanoyl group (4, m=l). T Kwas~determined ~ ~ for ~ all~these derivatives and the results are shown in Fig. 4.

-+100 80

1

H

- . A - . COCH3

- 0-

COCH2CH3

/+-+-+

+ + I + +

60

40 20 0

-20

6

8

10

12

14

16

18

Alkyl chain length (n)

Fig. 4. The Kraffr temperatures of three homologous series of N-alkyl I-amino-1-deoxy-D-glucitols with different lateral substituents ( & 4 m=O, 4 m=l).

Tailor-madecarbohydrate su$actants?

267

It is clear that the N-acetyl derivatives are quite soluble in water at low temperatures.

Unfortunately, these derivatives solidify very slowly at ambient temperature and are therefore difficult to purify through recrystallization. The N-propionyl derivatives are much easier to handle. Longer N-acyl substituents up to N-heptanoyl were introduced in N-dodecyl 1-amino- 1-deoxy-Dglucitol (3, n=l 1). The resulting products were all solids, TKrafft varied in an irregular fashion, ranging from 15 to 45 “C (Fig. 5).

100

80

-

I”’..

60 -

20 40 0 --

v

0

1

2

3

4

5

Acyl c h a i n l e n g t h

6

7

8

(n)

Fig. 5. The dependence of TKrafft on the size of the lateral substituent in a series of (N-acyl) N-dodecyl I-amino-I-deoxy-D-glucitols (4. m=0-5, n=l1).

5. The critical aggregation concentration The critical aggregation or micelle concentration (c.a.c. or c.m.c.) of a surfactant is of vital importance for the performance of surfactants. Surfactants are only effective (for cleaning purposes) above the c.m.c.. The lower the c.m.c., the smaller the amount of surfactant required in the detergent composition. However, if one wishes to use these compounds for the solubilization34, reconstitution’5 and crystallization36 of membrane proteins, a high c.m.c. is desirable to facilitate the removal of the surfactant during dialysis.

268

H.A. van Doren

The main factor influencing the c.m.c. of a given type of surfactant is the alkyl chain length. As a general approximation, the c.m.c. is lowered by a factor of ten when the chain length is increased by two methylene units. We have determined the c.m.c. of a homologous series of N-propionyl N-alkyl 1-amino- 1-deoxy-D-glucitols (4,m=l ; n=6- 13). The results are shown in Fig. 6.

0.1

0.01

E

v

0.0001

0.00001

6

8

10

12

14

16

Alkyl chain length (n) Fig. 6. The critical micelle concentrations at 40 "C (tetradecyl derivative at 60 "C) of a homologous series of N-propionyl N-alkyl 1-amino-I-deoxy-D-glucitols (4, m=2, n=6- 13).

With the exception of the tetradecyl derivative, which had to be measured at 60 "C due to solubility problems, the expected pattern was found. The anomalous behavior of the tetradecyl derivative can probably be explained by the fact that it does not form true micelles in aqueous solution. We have also investigated the influence of the stereochemistry of ten starting pentoseshexoses on the ~ . m . c . ~of' the resulting N-acetyl N-decyl l-amino-l-deoxy-D-alditols13. All measured c.m.c.'s were in the range of 0.6 to 1.4 mmol/L. This means that the differences as a result of headgroup variation did not exceed the experimental error boundaries. Straath0f3~ compared a number of homologous series of nonionic monosaccharide- and disaccharide-derived surfactants. The maximum difference in c.m.c. observed for compounds with the same alkyl chain length was well within one order of magnitude. Apparently, the alkyl chain length is a far more

Tailor-made carbohydrate siirfnctants?

269

determining factor for the c.m.c.-value than the type of carbohydrate and the nature of the group which links the headgroup and the alkyl chain.

octyl P-D-glucopyranoside 0 ,-" HO

OH

(CH3$H3

octyl I-thio-

-P-D-glucopyranoside

H0-SHO ,(cH2)1CH3 OH

~o*y~~3 HO OH

N\(CH,)j€H3

OH O H O ~ ' ~ ~ HO+N,(CHIIICH, 6H H

N-acetyl N-octyl 1-amino-a$-D-glucopyranoside

N-Acetyl N-octyl 1-amino-1 -deoxy-D-glucitol

Scheme 4. The structural formulae of five surfactants. all derived from D-glucose or D-glucitol and all containing an n-octyl chain, hut with different hetero atoms.

6. Toxicity and biodegradability Three series of compounds were evaluated in a number of standard biodegradability tests39 (modified Sturm test, BOD-test, Closed Bottle test) and a toxicity test4O. The toxicity tests were run against four selected micro-organisms by determining the maximum surfactant concentration at which uninhibited growth occurred (microtiter plates, optical density measurements in a Bioscreenapparatus). Initially, we studied the compounds shown in Scheme 4.The alkyl chain length was the same in all five compounds, but the type of hetero atom linking the hydrophilic and lipophilic moieties was varied. A D-glucitol moiety was present instead of a D-glucopyranoside in two of the compounds.

270

H.A. van Dorm

Subsequently, two homologous series of N-acyl N-alkyl I-amino-P-D-glucopyranosides (2,m=O, n=7- 15 and 2,m= 1, n=7- 15) were evaluated. We have not yet run the tests in duplicate, so only preliminary conclusions can be given: compounds with an intact glucose ring tend to be less toxic than glucitol derivatives compounds containing only C, H and 0 tend to be less toxic than derivatives with additional hetero atoms the toxicity for E. coli and B. subtilus increases dramatically with increasing alkyl chain length (from > I g/L to < I mg/L) the toxicity for S. cerevisiae and A. niger appears to show a maximum for the N-dodecyl or N-tetradecyl derivatives all compounds are at least partially biodegradable in all tests when using sludge from the local communal waste water treatment plant controls such as sodium dodecyl sulfate and dodecyl hexaethylene glycol were consumed more rapidly than the test substances. It would be interesting to repeat the biodegradation experiments after allowing adaptation of the sludge to the carbohydrate-derived surfactants. The data which were obtained are in general agreement with the results of Matsumuru et al.4t on a number of structurally similar glucosamine derivatives.

7. Conclusions We have found that it is possible to predict several important characteristics of carbohydratederived surfactants, such as type of aggregate formed in aqueous solution, Krufft temperature, c.m.c. and toxicityhiodegradability on the basis of the molecular structure. Obviously, further research is necessary to determine the molecular basis for such practical criteria as foam formation and foam stability, calcium tolerance, storage stability, etc., but the usefulness of a systematic approach has been clearly demonstrated.

Acknowledgements. The results described in this paper would not have been obtained without the (experi)mental efforts of Mr. Titus Rinia, Mr. Ronald Terpstra, Mr. Marc Benerink and Mr. Marco van d e r Toorn, all of whom spent the final part of their training 10 become laboratory technicians at our Institute, of Dr. Saskia Galema, who devoted a significant part of her PhD-studies (University of Groningen, The Netherlands) to carbohydrate-derived surfactants and of Dr. Lavinia Wingert, together with whom the author devised the predictive model for the type of aggregation as a function of alkyl chain length and number of hydroxykarbonyl groups during his stay at the University of Pittsburgh (Pennsylvania, USA).

Tailor-madecarbohydrate su$actants?

27 1

9. References I 2

H. Koch. R. Beck and H. Roper, StarcWStarke, 45 (1993) 2-7. M. Floor, A.P.G. Kieboom and H. van Bekkum, StarcWStarke, 41 (1989) 348-354; S. Matsumura, K. Aoki and K. Toshirna, J. Am. Oil. Chem. Soc., 71 (1994) 749-755.

3 4

A.C. Besemer and H. van Bekkum, SrarcWSturke, 46 (1994) 419-422. J. Mentech, R. Beck and F. Burzio, in Carbohydrates as Organic Raw Materials I/ (G. Descotes, Ed.), 1993, VCH Publishers, Weinheim, pp. 185-201.

5

H. Andree and B. Middelhauve, Tenside Sut$ Deterg., 28 (1991) 413-418; D. Balzer, ibid., 28 (1991) 419427. McCutcheon's Emulsi$ers and Detergents. Inlernational Ed., 1986, The Manufacturing Confectioner Publishing Co., Glen Rock, New Jersey.

6 7 8 9 10

J.R. Hurford, in Developments in Food Carbohydrate 2, (C.K. Lee, Ed.), 1980, Applied Science Publishers, London, pp. 327-350. K. Adelhorst, F. Bjorkling, S.E. Godtfredsen and 0. Kirk, Synthesis, (1990) 112. J.E.K. Hildreth, Biochem. J., 207 (1982) 363; M. Okawauchi, M. Hagio, Y. Ikawa. G . Sugihara, Y. Murata and M. Tanaka, Bull. Chem. SOC.Jpn., 60 (1987) 2719-2725. T.J. Williams. N.R. Plessas, I.J. Goldstein and J. Lonngren, Arch. Biochem. Biophys., 195 (1979) 145151.

I1 12 13 14 15

16 17 18 19 20 21 22 23 24 25 26 27 28

H. Kelkenberg, Tenside Surf: Dererg., 25 (1988) 8-13. H.A. van Doren, R. van der Geest, C.F. de Ruijter, R.M. Kellogg and H. Wynberg, Liq. Cryst., 8 (1990) 109- 12 I . H.A. van Doren and T.C.J. Rinia, submitted for publication in J. Carbohydr. Chem.. E. Penzel and K. Oppenlander, Tenside Deterg., 11 (1974) 129. J.T. Davies, Proc. 2nd Congr. Surface Act., 1 (1957) 426, J.T. Davies and E.K. Rideal, in Interfacial Phenomena, 2nd ed., 1963, Academic Press, New York and London, pp. 372-374. R. Miethchen and D. Peters, Wissenschafrliche Zeitschrifr der Universitat Rostock, 36 (1987) 55. H.A. van Doren and L.M. Wingert, Liq. Cryst., 9 (1991) 41-45. G.J.T. Tiddy and M.F. Walsh, in Aggregation Processes in Solution (E. Wyn-Jones and J. Gormally, Eds.), 1983, Elsevier Scientific Publishers, Amsterdam, pp. 151-185. H.A. van Doren and L.M. Wingert, Recl. Trav. Chim. fays-Bas, 113 (1994) 260-265. H.A. van Doren, S.A. Galema and J.B.F.N. Engberts, Langmuir, 11 (1995) 687-688. S.A. Galema, E. Howard, J.B.F.N. Engberts, J.R. Grigera, Carbohydr. Res., 265 (1994) 215-225. Y. Moroi and R. Matuura, Bull. Chem. SOC.Jpn., 61 (1988) 333; Y. Moroi, in Micelles, Theoretical and Applied Aspects, 1982, Plenum Press, New York and London, pp. 113-129. K. Shinoda and E. Hutchinson, J . Phys. Chem., 66 (1962) 577; K. Shinoda, J. Phys. Chem., 85 (1981) 3311. Typically, 1-3 mg of surfactant i s heated from 0-100 "C i n 50-60 pI of water in a sealed aluminum sample pan. A.C.S. Lawrence, in Liquid Crystals, vol. 2, Part 1 (G.H. Brown, Ed.), 1969, Gordon &Breach, London, p.1. H.A. van Doren and L.M. Wingert, Liq. Cryst., 9, 41-45. H.A. van Doren, S.A. Galema and J.B.F.N. Engberts, in preparation. G.A. Jeffrey and L.M. Wingert, Liq. Cryst., 12 (1992) 179-202.

212 29 30 31 32 33

34 35 36 37 38 39 40 41

H.A. van Doren T. Gu and J. Sjoblom, Acra Chem. Scand., 45 (1991) 762-765. J.K. Weil, F.S. Smith, A.J. Stirton and R.G. Bistline Jr., J . Am. Oil Chem. Soc., 40 (1963) 538. H.A. van Doren, R. van der Geest, R.M. Kellogg and H. Wynberg, Carbohydr. Res., 194 (1989) 71-77. W.V. Dahlhoff, Liebigs Ann. Chem., (1991) 463-467. H.W.C. Raaymakers, Ph.D-Thesis, University of Nijmegen, The Netherlands: H.W.C. Raaymakers, E.G. Arnouts, B. Zwanenburg, G.J.F. Chittenden and H.A. van Doren, Red. Trau. Chim. Pays-Bas, 114 (1995) 301-3 10. P. Rosevear, T. van Aken, J. Baxter and S. Ferguson-Miller, Biochemistry, 19 (1980) 4108-41 15. A. Walter, Molec. Cell. Biochem., 99 (1990) 117-123. F. Reiss-Husson in Crystallization of Nucleic Acids and Proteins - A Practical Approach, (A. Ducruix and R GiegC, Eds.), 1993, Oxford University Press, Oxford, pp. 175-193. The c.m.c.s were determined at 40°C with a manually operated Kriiss K-l tensiometer equipped with a duNouy-ring. A.J.J. Straathof, Carbohydrates i n fhe Nerherlands, (1988) 27-30. OECD guidelines for Testing Chemicals, 301D, Closed-bottle test, Organization for Economic Cooperation and Development (OECD), 1981. R.G. Bristline Jr,, E.W. Maurer, F.D. Smith and W.M. Linfield, J . Am. Oil Chem. Soc., 57 (1980) 98. S. Matsumura, Y. Kawamura, S. Yoshikawa, K. Kawada and T. Uchibori, J . Am. Oil Chern. Soc., 70 (1993) 17-22.

15 Calcium sequestering agents based on carbohydrates

A.C. Besemerl and H. van Bekkum2 ITNO Nutrition, P.O. Box 360, 3700 AJ Zeist, The Netherlands. *Laboratory of Organic Chemistry and Catalysis, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands.

Summary. Present-day West-European builder systems in detergent formulations often consist of zeolite NaA together with a synthetic polycarboxylate. The latter materials are nonbiodegradable and substitutes should be considered. Carbohydrates seem attractive raw materials in this respect, since carboxylate groups that are essential for calcium binding in organic molecules can be introduced rather easily. In this paper three types of carbohydrates suitable for this purpose will be discussed. (i) Glucose can be converted into gluconic and glucaric acid. The calcium binding capacity of glucarate is moderate, but can be greatly enhanced in the presence of borate at pH > 9. A new option is to oxidize starch to polyglucuronic acid. Subsequent hydrolysis and oxidation of the glucuronic acid yields glucaric acid. (ii) Naturally-occurring carboxylated carbohydrates such as alginic acid and pectin exhibit modest calcium binding ability, which is not strong enough for application as a co-builder. Moreover, the price is (too) high. (iii) The conversion of polysaccharides (starch, cellulose and inulin) into carboxylated derivatives appears to be the most attractive approach. Two methods will be dealt with: - introduction of carboxymethyl- and dicarboxymethyl groups. The latter type in particular has strong calcium binding strength. However, there is strong evidence that these materials are poorly biodegradable. Specific oxidation. Oxidation at the 6-position will lead to polyuronic acids, which have modest calcium binding ability. So far, the best and most promising method for the preparation of builders from carbohydrates is by oxidative cleavage of the glycolic groups. The resulting dicarboxyderivatives from e.g. starch and inulin have calcium-binding capacity comparable to that of sodium tripolyphosphate and nitrilotriacetic acid and only slightly less than that of the synthetic polycarboxyates. In view of their high biodegradability and low cost, starch and inulin are the most promising raw materials. The most feasible oxidation method is the bromide-catalyzed hypochlorite oxidation.

274

A.C. Besetnerand H. Van Bekkum

Introduction During the last forty years, the consumption of detergents has increased considerably. The worldwide consumption of detergents is currently estimated at thirty million tons per year.' Laundry detergents, which form a substantial part of this consumption, are applied in households, industries, hospitals and laundries. Many substances are included in a detergent formulation, each with a specific function. Four classes of substances are generally distinguished: surfactants -

builders and co-builders bleaching agents and bleaching boosters additives

Table 1 gives a survey of the composition of a typical powdered detergent. Table I . Composition of a Western European laundry detergent.a

CLASS surfactants builders bleaching agentstboosters additives

SUBSTANCES nonionics, soaps a1kylbenzenesulfonates phosphates, zeolite, polycarboxylates, NTA, soda perborate, percarbonatemAED enzymes, carboxymethyl cellulose, anticorrosion agents, optical brighteners

CONTENT(%) 20-25

30-50

20-25 5-10

the past fillers e.g. Na2S0, have heen used.

In order to improve the detergent performance, considerable research effort is being devoted to improving the individual components. As seen in Table 1, a major constituent (sometimes up to 50%) is represented by the builder system. The main function of a builder is to lower the calcium and magnesium ion concentration in the washing water to prevent incrustation by calcium and magnesium salts. Incrustation will render serious negative effects such as greying and hardening of the fabrics, resulting in damage. For the sequestering of calcium and magnesium ions two principles are applied (i) homogeneous coordination (ii)

ion exchange with suspended solids

Calcium sequestering agents based on carbohydrates

275

Sodium tripolyphosphate (STPP) has been used for many years and is still in use as builder. This compound is inexpensive and non-toxic and has an excellent performance. However, the consumption of this material is decreasing. The estimated consumption in 1985 amounted to lo7 tons, whereas an estimate for 1995 gives 1.5.106 tons. The reason for this dramatic decrease is the fact that STPP has a distinct drawback: STPP is a nutrient for algae and eutrophication may occur, especially in stagnant surface waters where the material will persist. This has in fact been observed at many places.2 Table 2. Some calcium binding compounds."

Structure

Compound

Citrate

-0OC-

Log KCa

H2? - cooC - OH

Biodegradability

3.6

Readily

3.8

Good

4.4

Good

5.8

Bad

5.1

Bad

11.8

Bad

6.8

Good (at T >lO°C)

I

H2C - COO-

H2C-O-CH2

Ox -diacetate

l

I

coo-

-0oc

( O h

H

Carboxymethyloxy succinate (CMOS)

H2C -0

I

-0oc

-coo

H H2C-C-

I

Oxydisuccinate

-0oc

(ODs)

- C - CH2 I I

1

coo-

H 0-C-

I

coo-

CH2

boo-

cooCarboxymethyloxy rnaionate (CMOM)

H2C

I

-0oc -0OC

Ethylenediamine .etra-acetic acid -0OC

Vitrilotriacetic acid

-0 -

H

- CH2 N - CH2 -CH2 -N /

-CH2

/

CH2CooN - CH2COO-

\ CHzCOOLog

Kca = 6.0.'

- coo-

l coo-

CH2 -COO/ \

CH2 --COO-

216

A.C. Besemerand H . Van Bekkum

Nowadays STPP can be removed from water by precipitation in a 'third-step sewage sludge installation'. This is practiced in Sweden. It is also noteworthy that the contribution of detergent STPP to the total phosphate burden is often relatively small. Despite this there is a ban on STPP in many countries. A suitable substitute for STPP is needed and over many years considerable effort has been devoted to finding suitable alternative sequestering agents that meet the most important demands e.g. performance, price, biodegradability, etc.. Numerous compounds have been synthesized and tested for their calcium and magnesium ion binding ability and for their biodegradability (an extensive survey is given in references3 and4). A few representative examples are shown in Table 2. We see that compounds with good performance combine two or more carboxylate groups with preferably an ether oxygen or an amine function. Nitrogen-free compounds are generally preferred. It is striking that materials with satisfactory calciudmagnesium binding properties are generally poorly biodegradable whereas compounds with lower performance such as citric acid are readily biodegradable. Despite the extensive investigations, so far only a few suitable substitutes have been found, e.g. nitrilotriacetic acid (NTA)S, which is used in Canada. A solid suspended inorganic ion exchanger, zeolite NaA (see Fig. 1) is now applied in large quantities in Europe (> lo6 tons per year).

0

0

Fig. 1. Structure of zeolite NaA.

Its use and eco-toxicological properties are discussed in numerous papers.6-20The substance has attractive properties and is safe in many respects. A significant increase in the consumption of zeolite is seen in North America (from 1.1.105 tons in 1985 to 3.6.1OS tons in 1992).9 Zeolite NaA is a microporous ahminosilicate which consists of small cubic crystals (1-2 pm). Given the

Calcium sequestering agents based on carbohydrates

277

Si/A1 ratio of 1, zeolite NaA possesses a maximum CdMg exchange capacity. However, zeolite NaA applied alone falls somewhat short of the ideal. The reason is that the compound has good calcium binding properties, but magnesium ions (also responsible for incrustation) are exchanged too slowly, especially at lower washing temperatures. Moreover, being a solid, zeolite NaA cannot reach incrustations consisting of calcium and magnesium deposits. To overcome these problems a soluble co-builder is added, generally synthetic polycarboxylates. These materials will bind magnesium and may remove incrustations formed during the washing process. Polycarboxylates may also act as a crystallization inhibitor, thereby preventing incrustation of calciudmagnesium compounds, especially carbonates. Their use and properties are described in a review article.2' An important aspect with regard to the respective detergent constituents is their (bi0)degradability. Most of the constituents are biodegradable, but the zeolite/polycarboxylate builder system does not meet this requirement. Zeolite NaA, being an inorganic material, cannot be converted into carbon dioxide and water, but belongs to a natural class of materials. Most of the substance will be removed in sewage sludge installations, but especially the small particles will pass through these systems and consequently will be released into surface water.' Hydrolysis takes place slowly and results in the release of aluminium ions. In addition to zeolite NaA, a new type, zeolite P, with exchange capacity similar to zeolite NaA but with an improved exchange rate, was recently introduced by Unilever.22 The polymeric material (e.g. polyacrylate) is poorly degradable.21Most of the material will adsorb onto particles in the sewage sludge solution, but part will pass through and will consequently be released into the environment.

In view of their structure and reactive character, carbohydrates seem to be suitable as raw material for the preparation of materials that can bind calcium and magnesium ions.2"*4 Carboxylic groups are present or can easily be introduced by oxidation or substitution reactions. In most cases acetal oxygen atoms are present that will contribute to the calciudmagnesium coordination by the carboxylate groups. An advantage of carbohydrates is the fact that they are renewable and their derivatives are probably biodegradable or can be made to be so. In this chapter a survey of the most important options will be given. Characterization of the products with respect

to their calcium binding properties involves several methods and definitions; these will be briefly compared and discussed.

218

A.C. Besemerand H. Van Bekkuin

Calcium binding capacity When calcium ions are bound with a single coordinating agent(L) according to the simplified equation Ca(I1) + L

Ca(I1)L

the calcium binding strength is expressed as the association equilibrium constant K, defined as

Values for some compounds are presented in Table 2 . In polymeric materials with many identical binding sites, the above approach will not suffice, because upon successive coordination of these ions the binding strength will change (decrease). In this situation a ‘Tanford plot”5-27 can be used to characterize the polymer. An intrinsic complex constant is defined as follows: K = Z/Cc,.( 1-Z) = KinteCZ in which Z is the fraction of calcium-ion coordinated sites. The intrinsic complex constant is now defined at Z = 0. Generally, agents with K > I 0 5 or polymeric substances with Kin, > los at Z = 0.5 are satisfactory as co-builders. It is recognized that for a good performance of an anionic laundry detergent the calcium ion concentration in the washing liquid should be lower than 10-5 M.27 The calcium sequestering capacity has been defined as the number of mmol Ca(I1) that can be added to I g of complexant until the concentration of non-coordinated Ca(I1) reaches 10-5 M.26 Some authors use a similar definition with, for instance, 5.10-5 M Ca(I1) as the upper limit28, accounting only for the precipitation of calcium carbonate. Two methods are used for determining the sequestering capacity: I.

calcium oxalate is sparingly soluble in water. In the presence of a coordinating agent the formation of oxalate precipitate is retarded. The point at which precipitation in a solution containing the coordinating agent commences (at C,, = 10-5 M) is determined by turbidimetry. It should be noted that crystallization inhibition may lead to erroneous results.

Crrlci~insequestering agents based O H carboliydrates 2.

279

measurement of the free (hydrated) Ca(I1) concentration in solution may be carried out with an ion specific electrode. The calcium sequestering capacity is determined by titration i.r. the Ca(I1) concentration is measured upon introduction of small amounts of Ca(I1) into a solution (pH 10) containing a known amount of the coordinating species. This method gives particularly reliable results. Moreover, with this method the complex constant or intrinsic complex constant is measured. The values obtained using the first method are higher than those obtained using the second

method, presumably because crystallization inhibition plays a role in the oxalate method. The utmost caution should therefore be applied when comparing the performance of the complexing products.

Monosaccharide-based builder systems Many natural organic acids exhibit modest calcium binding properties. Both Mehltretter et ~il.23,?4 and Van Duin et cil.29-31measured the calcium binding properties of lactic, tartaric, gluconic and glucaric acid. A survey of the properties is presented in Table 3. Gluconic acid exhibits calcium binding properties, but only at pH > 13, which is too high for most washing processes. A better 3 that glucaric acid performance is given by glucaric acid. However, Mehltretter el ( ~ 1 . ~established

alone does not suffice in washing experiments. Table 3. Calcium binding of some organic acids

Substance

Author a+b

SC (mmol C d g ) SC with borate (mmol C d g )

citrate

1

0.9

citrate

2

I .4

1.4

tartrate

2

0.6

2.1

gluconate

1

0.1

gluconate

2

0.7

glucarate

1

0.3

glucarate

2

0.4

2-ketogluconate

1

0.3

STPP

2

1.4

Mehltretter el u L ? ~ Van Duin et al.?9.3'

1.7 2.7 1.4

280

A.C. Besemerand H . Van Bekkurn As claimed in two early Dutch patents, the performance of glucarate can be greatly improved

by adding boric acid or borate salts.32-33Borate acts as a bridging agent and above pH 9 esters will be formed whose structure is represented in Fig. 2.

Fig. 2. Calcium(I1) coordination in the diasiereomeric (R)-and (S)-borate diesters of D-glucarate in water at pH 10. The horium atom is central in the complex and is surrounded by four oxygen atoms

Van Duin et ~ l . 2 ~ - studied " the structure and characteristics of these compounds. From their investigations it could be concluded that the calcium binding ability of borate esters of glucaric and idaric acid holds promising perspectives for their use as a co-builder in detergents.

Possibilities for the preparation of glucaric acid Since glucose is an attractive starting material, it is interesting to consider the glucose-glucaric acid and also the gluconic acid-glucaric acid conversion. In spite of continuous efforts during a period of twenty years to find suitable preparation methods, no satisfactory process could be developed. The classical way to oxidize glucose is with concentrated nitric acid at high temperature, but this method is not attractive for large scale preparation. The Pt/O2 oxidation of glucose was found to be insufficiently specific.34 The formation of oxalic acid is considered a major problem. The yield is too low to make the process economically feasible, and purification difficulties are foreseen. Glucaric acid can in principle also be prepared from gluconic acid. In the literature nitric acid oxidation of 2.4-3,sdimethylenegluconic acid to 2,4-dimethyleneglucaric acid is described.3s It should be noted that

Calciuni sequestering agents based on carboliydrates

28 1

despite the high temperature and low pH during the reaction, one of the protecting groups is not removed. As mentioned, nitric acid as oxidant is not attractive. Another possibility for preparing aldaric acids is to oxidize the corresponding uronic acid. Uronic acids can be obtained by hydrolysis of certain polysaccharides (e.g. pectic acid, alginic acid). However, the price of these polysaccharides together with the cost of subsequent hydrolysis and oxidation processing steps does not make this a favourable route. T o overcome the non-selectivity of the oxidation, partially protected glucose or gluconic acidderivatives can be used. An attractive route for the preparation of glucaric acid is shown in the following diagram. H

c=

H e

H

TEMPO

H e

~

H

HOBr/OBr HO

OCH3

HO

OH a-methylglucopyranoside

I

I CI

COOH

CH20H

I OH 0

c=

0

-

I

Pd/Bi/C

-C

C-

OCH~

C-

I CI

OH

a-methylglucuronic acid

C

C-

> O H ‘” D-glucuronic acid

I CI C 0”

>H

D-glucaric acid

Scheme I . Preparation of glucaric acid

De Nooy et ~ 1 . 3 6 . 3 7recently described the selective and quantitative oxidation of starch and methyl glucoside towards a-polyglucuronic acid and 1-0-methylglucuronic acid, respectively, by a TEMPO-mediated? hypochlorite/hypobromite oxidation of a- and P-methyl glucoside. Both products can be converted by hydrolysis to yield glucuronic acid, which can be readily oxidized with oxygen using Pd/Bi/C38 as the catalyst to give glucaric acid. It is essential here to develop an efficient hydrolysis method. The new Sudzucker method involving continuous oxidation operation applying a fixed bed of Pt-catalyst can also be considered. High selectivity of 6-oxidation of methyl glucoside is

claimed.

Natural polysaccharides Many polysaccharides found in nature contain carboxylate groups e.g. pectin, alginic acid and various gums, such as the branched bacterial xanthan gum and gellan gum. These materials have

‘ TEMPO = 2.2.6.6-tetramethylpiperidin-

1 -oxyl

282

A.C. Beseiner- trrid H . Vrrn B e k k i m

found various applications, especially i n food. Some of these substances have modest calcium binding ability; however, because these materials are relatively expensive, no application in detergent formulations is foreseen.

Introduction of carboxylate groups in polysaccharides Various methods are available for the introduction of carboxylate groups in polysaccharides. The most important are: -

carboxymethylation

-

oxidation of the primary hydroxyl group

- oxidative cleavage of the glycolic group - combinations of these methods.

Carboxymethylation of polysaccharides It is a well-known fact that carboxylate groups may be introduced in a polysaccharide by reaction of chloroacetic acid in alkaline medium (reaction of Williamson). An example of a material

prepared in this way is carboxymethyl cellulose (CMC), which is used in detergent formulations as Table 4. Calcium sequestering capacity (SC) of various polysaccharide derivatives obtained via oxidation or substitution.j"

Substance

sc

B iodegradabilityb

(mmol Cdg)" slow

Carboxymethyl cellulose

0.7 0.7

Dicarboxymethyl starch (DS=2.1)

2.1

readily(initia1 phase); slow

Dicarboxymethyl starch (DS=I .6) Dicarboxymethyl starch (DS=I .3)

1.9 1.2

readily(initia1 phase); slow slow

Dicarboxymethyl cellulose

1.6

slow

Polycarboxylate' from bamboo

1.4

no data

Dicarboxy-starch' Polycarboxylatec from reed

1.9 I .4

slow no data

Carboxymethyl starch

slow

The calcium sequestering capacity (SC) has been defined as the number of rnmol Ca(I1) that can be added to 1 g of coniplexing agent until the concentration of non-coordinated Ca(I1) M [2h]. reaches Biodegradability depends somewhat on molecular weight and degree of suhstitution. Obtained by oxidation with sodium hypochlorite.

Calciirin seqiresteririg agents based on carbohydrates

283

an anti-redeposition agent. Polysaccharides other than cellulose have been tested and patented.39 For the preparation of calcium binding agents, materials with an average degree of substitution of 2 have been mostly considered. Eldib39 claims that carboxymethylated starches with DS > 1 are good calcium binders. Carboxymethyl starch and carboxymethyl cellulose were also prepared and tested by Diamantoglou et al. .40 These authors measured the calcium binding capacity using the calcium ion selective electrode method and found only a modest calcium binding ability. Some dicarboxymethyl derivatives of starch and cellulose were also prepared. These materials have a much better calcium binding ability (see Table 4 and Fig. 3). The starch derived materials in particular have a much better performance than the corresponding carboxymethyl derivatives. It was established that the calcium binding ability of dicarboxymethyl starch is proportional to the degree of substitution (see Table 4). Probably because of several drawbacks of these substances with regard to price, performance and biodegradability, no further efforts have been undertaken to develop materials of this kind.

10

mM Ca(ll)

t 1

0.1

0.01

1 = Sodium tripolyphosphate 2 = Dicarboxy-starch 3 = Dicarboxymethyl starch 4 = Carboxymethyl starch

Fig. 3. Calcium sequestering ahility of carboxymethyl starch, dicarboxy starch and sodium tripol yphosphate.

A.C. Besenierand H . Van Bekkiim

284

One of the most recent investigations concerning carboxymethylation is described by These authors prepared carboxymethyl inulin with different DS but the calcium Verraest et a1..3134? binding ability was only very modest. It was found, however, that the materials act as a crystallization inhibitor (see Fig. 4).Efficiency approaches that of the high molecular weight polyacrylates. No biodegradability information is known to date.

t

a

r

3

2

0

c

a 0

0

60

120

180

240

300

time (min)

Fig. 4. Crystallization inhibition by carhoxymethyl inulin. Growth curves of calcite crystals (a) in the ahsence of an inhibitor. (b) i n the presence of 5ppm carboxymethyl inulin (degree of substitution 0.36), (c) in the presence of 5ppm carboxymethyl inulin (degree of substitution 0.68). (d) in the presence of 5ppm carhoxymethyl inulin (degree of substitution I.05).

C6-Oxidation and oxidative cleavage of the glycolic group Another possibility for introducing carboxylate functions in polysaccharides is by specific oxidation.36.37.43-6’The respective hydroxyl functions in polysaccharides can be oxidized to yield materials with one or more carboxylate functions per monomer unit. Only a few oxidation methods lead to specific conversion of carbohydrates. As raw materials, starch, cellulose and inulin should be considered in view of their price and availability. However, it has been established that other starch products, such as maltose, maltotriose, dextrins and cyclodextrins, also yield products that perform we11.26.28

Calcium sequestering agents based on carbohydrates

285

The C6-oxidation of starch and cellulose was discussed above. It has been established that the calcium binding property of a-polyglucuronate is apparently less than that of 2,3-dicarboxystarch. The C6-oxidized glucans have various prospects, but not primarily as a builder in detergents. In the glucans the glucose groups should be oxidized for this purpose, preferably at the C2 and C, positions. The structure of the product is presented in Fig. 5a.

Fig. 5 . Structure of dicarboxy-starch: a. single oxidized glucose unit b. ODA-moieties in two oxidized units.

In addition to their preparation, the characteristics of glycol-oxidized polysaccharides have been described by the present authors. In two patents43.44 a comparison is made between the performance of dicarboxy-starch and dicarboxy-cellulose and that of STPP. From these investigations it must be concluded that oxidized carbohydrates behave quite well as a builder in detergent formulations. The use of dicarboxy-starch as a co-builder was foreseen by Nieuwenhuizen et al..4s This was confirmed by Koch et aL6*, who compared the performance of SOKOLAN CP5 (a synthetic polycarboxylate) and dicarboxy-starch in washing trials. structure is The oxydiacetate structure favours calcium binding a b i l i t ~(Fig. ~ ~ ,6). ~ This ~ found in starch when two neighbouring glucose groups are oxidized (Fig. 5b).

Fig. 6 . Structure of oxydiacetic acid and its calcium complex

Various methods are available for the desired conversion: I . a two-step process; in the first step a glucose unit is transformed into the corresponding dialdehyde by oxidation of C, and C, hydroxyl groups (see Fig. 7).

286

A.C. Besemer and H. Vun Bekkutn

Fig. 7. Structure of dialdehyde starch.

This reaction was investigated for the first time more than twenty-five years ago. Since then, several groups have investigated the process and tested the products. The reaction can be carried out with sodium periodate or lead (IV) tetraacetate. With the former reagent, high yields are possible. The reaction can be carried out in an economicallyfeasible manner by recycling the reagent or via an indirect electrochemical process. The second step consists of oxidation with sodium chlorite: Dialdehyde starch + 2 NaC102 + Dicarboxy-starch + 2 NaOCI. A rather large amount of oxidant is needed because sodium hypochlorite - a reaction product

-decomposes sodium chlorite.

2 C10,

+ HOCl + 2 C102 + C1- + OH-

Floor e f al.26.49 improved the process by scavenging NaOCl with hydrogen peroxide, according to the equation NaOCl + H20,

+ NaCl + H 2 0 + 0,.

The products obtained in this way have excellent calcium binding properties (see Table 5). 2. The second way to obtain dicarboxy derivatives is a one-step process and can be achieved by

sodium hypochlorite2h.28,"-46.48.55-58 or sodium hypobromite.s0-s"s~"-6'This process is well-known and is usually applied for slight modification of starches (alteration of solubility, viscosity etc.). Besides oxidation, some depolyrnerization occurs. The products obtained via hypochlorite oxidation have a lower calcium binding capacity than the products obtained via the two-step process (see Table 5). Recently, Santacesaria et ~ 1 . 2 8stated

Calcium sequestering agents based on carbohydrates

287

that the hypochlorite process yields better products than the sodium periodatekodium chlorite process. Table 5 . Calcium sequestering agents prepared from glucans.

Material

Methoda Overall Yield (%)

COONa sc Content(%) (mmol Cdg)

Dicarboxy-starch

1

90

86

2.39

Dicarboxy-starch

2

93

79

2.5 I

Dicarboxy-starch

3a.

86

63

1.04

Dicarboxy-starch

3b.

91

69

1.18

Dicarboxy-starch

3c.

38

81

1.77

Dicarboxy-starch

4

97

76

1.38

Dicarboxy-amylose

2

64

74

2.39

Dicarboxy-arnylopectin

2

40

69

2.58

Dicarboxy-cyclodextrin

2

no data

95

2.09

Dicarboxy-maltodextrins

2

52

84

2.52

Dicarboxy-dextran

2

82

68

1.80

method 1 NaI04/6 NaCIO, [26,49] method 2 NaIO4/2 NaC102-2 H,O, [26,49] method 3a. 3 NaOCl at T = 293 K [48] method 3b. 3 NaOCl at T = 273 K [48] method 3c. as 3a. with repeated oxidation of the product with NaOCl method 4. 3 NaOCl and NaBr as the ca~alyst.’~

The hypochlorite oxidation can be improved by using bromide as a catalyst.so.52-53Starch is then in fact oxidized by sodium hypobromite. OCI’

Tr-y

CI- &Br-

pol Dicarbox ysacckir ide

/HOBr

Polysaccharide

Scheme 2. Catalytic effect of bromide in the hypochlorite oxidation of carbohydrates.

The result is that the reaction proceeds faster and that products with better performance are obtained. Also, the use of excess oxidant appears to improve the sequestering capacity of the products.48

288

A.C. Besemerand H . Van Bekkurn

A polysaccharide that will be available in large quantities in the near future is inulin (see Fig. 8). It is to be expected that this material will afford a good calcium binding agent because the ODA-

structure is obtained immediately upon glycolic oxidation (see Fig. 9), whereas substantial oxidation is needed in starch.

CH20H

0

Fig. 8. Structure of inulin.

Fig. 9. Structure of dicarboxy-inulin.

This proved to be true (see Table 6). It is remarkable that the best but more expensive method for the oxidation of starch, the two step procedure, seems to fail when applied to i n u h 4 5 However, excellent calcium binding products are obtained with sodium hypochlorite. Moreover, the use of bromide is not necessary (see Table 6). An important difference between starch and inulin is seen when partial oxidation is applied (see Fig. 10 and 11). The relation between the degree of oxidation and the calcium sequestering capacity of inulin is linear while for starch this relation is an S-shaped curve. From these results it follows that inulin can be a more attractive starting material. At a low degree of oxidation, the calcium binding capacity is higher than that of the corresponding dicarboxy-starch. Because of the low degree of oxidation, the perspectives for biodegradability are better.65 Table 6 . Oxidation of inulin to dicarbo~y-inulin.~~

Methoda A

B B C C D I

sc (mmol Cdg) 2.4 2.0 2.1 2.3 2.5 2.3

log Kin, 9.6 9.0 9. I 8.6 9.6 7.8

Yield(%) 82 95 86 73 72 79

Method A: hypochlorite, pH 9 Method B: hypochlorite, bromide as a catalyst, pH 10, gradual addition of the oxidant Method C: hypochlorite, bromide as a catalyst, pH 10. addition of the oxidant at once Method D: excess of hypochlorite (35%). pH 10, addition of the oxidant at once, bromide as a catalyst.

Calcium sequestering agents based on carbohydrates

289

sc

2.5

mmde C a l ~

2.0

I I I

I I

I I II I

1.5

10

0.5

0

20

I 40

I €0

I 80

I 100

Fig. 10. Relation between Ca-sequestering capacity and degree of oxidation (OD) of dicarboxy-starch. Line I represents the resuits of a product obtained via hypochloritebromide oxidation. Line 2 represents the results of a product obtained via periodatekodium chlorite-hydrogen peroxide oxidation.

0

I

I

I

I

25

50

75

1M)

Fig. 1 1. Relation hetween Ca-sequestering capacity and degree of oxidation (OD) of dicarboxy-inulin.

Nieuwenhuizen et a1.45 used NMR spectroscopy to characterize the complexes of dicarboxystarch and calcium. They established that in the coordination of calcium(I1) with oxidized starch, at least seven non-water oxygen atoms participate. Similar conclusions follow from the study by Floor et af.26, who investigated the complexes of dicarboxy-P-cyclodextrin with calcium and lanthanides (lanthanides are often isostructural to calcium complexes and behave in a similar way). In dicarboxy-inulin, one calcium ion is also coordinated by seven oxygen atoms. Since an oxidized glucose or fructose moiety cannot provide seven sites, it seems likely that two adjacent oxidized groups coordinate one calcium(I1) ion. An example of a Ca-dicarboxy-inulin complex is presented in Fig. 12 (proposed structure).

Conclusions Of the possibilities discussed above, three saccharide-type builders should be mentioned: a.

Glucaric acid, which can be prepared by oxidation of glucuronic acid. The latter compound can be prepared from methyl glucoside by oxidation at the 6-position followed by hydrolysis

290

A.C. Brserner and H . Van Bekkitm

F"0X - -

Figure 12. Structure of a 1 :2 Ca-DCI complex. Two oxidized fructose units participate in the coordination of one calcium ion.

and oxidation at C , . Glucuronic acid can also be approached from starch through TEMPOcatalyzed oxidation at the 6-position followed by hydrolysis. b. Dicarboxy-starch, which - in view of the price of the chemicals involved - should be obtained from starch by sodium hypochlorite oxidation, preferably with bromide as the catalyst. c. Dicarboxy-inulin, which should also be prepared from the corresponding oligosaccharide with sodium hypochlorite.

Perspectives Four aspects of the application should be considered: - pe$onnance in washing tests

Washing experiments have been reported with dicarboxy-starch (DCS) and dicarboxy-cellulose (DCC). It was found that both DCS and DCC behave rather well in comparison to sodium tripolyphosphate. DCS has also been tested as co-builder and the substance proved to work well.

Calcium sequestering agents based on carbohydrates

29 1

From our experiments and also from other reports it can be concluded that polycarboxylatebased polysaccharides are suitable as phosphate and/or polyacrylate replacement. - price The expected price of dicarboxy-starch and dicarboxy-inulin is about NLG 4.00 - 10.00/kg, which is of the same order as the cost of the synthetic polycarboxylates. The performance of the latter may be somewhat better. Since the amounts needed per washing are similar, no significant price difference is to be expected and the 'green label' might favour the dicarboxypoly saccharides. - feasibility of large-scale production

Large-scale application of dicarboxy-polysaccharides as co-builder in Europe in combination with zeolite would require a production capacity of about 50,000 tons per year. It is necessary that a process of this type can be carried out safely and inexpensively, using chemicals that are available in large quantities. - biodegradability

From a few studies and unpublished data, it can be deduced that high molecular weight glycol oxidized products are poorly biodegradable. Results presented by Matsumura et u1.65 indicate that the degree of oxidation plays an important role, i.e. the higher the degree of oxidation, the lower the rate of biodegradation. Here, partially-oxidized inulin constitutes an interesting option.

References 1.

2. 3. 4. 5.

6. 7. 8. 9. 10. 1I. 12. 13. 14. 15. 16.

T. Egli, Microbiological Scierices 5 , 36 (1988). M. Dwyer, S. Yeoman, J.F. Lester and R. Perry, Environriienral Technology 11 (1990) 263. M.M. Crutchfield, J. Am. Oil C/iernists'Soc. 55 (1978) 58. E.A. Matzner, M.M. Crutchfield, R.P. Langguth and R.D. Swisher, Tenside Deterg. 10 (1973) 239. R. Perry, P.W.W. Kirk, T. Stephenson and J.N. Lester, Water Res. 18 (1984) 255. D.W. Breck, W.G. Eversole, R.M. Milton, T.B. Reed and T.L. Thomas, J. Am. Clieni. Soc. 78 (1956) 5963. T.B. Reed and D.W. Breck, J. Am. Cliem. Soc. 78 (1956) 5972. M.S. Rigutto. in: lntroducliori to Zeolire Science arid Practice. Ed. H. van Bekkum, E.M. Flanigen and J.C. Jansen, Elsevier, Amsterdam (1991). 731. S. Rock, PQ Corporation, personal communication (1992). M.J. Schwuger and H.G. Smolka, Colloid & Po/yn7. Sci. 254 (1976) 1062. M.J. Schwuger, H.G. Smolka and C.P. Kurzendorfer, Tenside Deterg. 13 (1976) 305. H.G. Smolka and M.J. Schwuger, Tenside Dererg. 14 (1977) 222. P. Berth, J. Am. Oil Chemists'Soc. 55 (1978) 52. P. Berth, W.K. Fischer and R. Schmid, Tenside Dererg. 14 (1977) 1. P. Berth, Tenside Deterg. 15 (1978) 176. M. Ettlinger and H. Ferch, Manufacturing Chemist & Aerosol News (1978) 5 I .

292 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

A.C. Besemerand H . Van Bekkum H. Krueszmann. P. Vogel, H. Hloch and H. Carlhoff, Seifen-Ole-Fette-Wachse 105 (1979) 3. P. Berth, M. Berg and K. Hachmann, Tenside Deterg. 20 (1983) 276. H. Upadek, P. Krings and E.J. Smulders, Chimicaoggi (1990) 61. K. Henning, J. Kandler and H.D. Nielen, Seifen-Ole-Fetre-Wachse 103 (1977) 571. M. Hunter. D.M.L. da Motto Marques, J.N. Lester and R. Perry, Env. Technol. Letters 9 (1988) 1. E. Vogt, Chemisch Magazine (1995)148. C.L. Mehltretter, B.H. Alexander and C.E. Rist, Ind. Eng. Chem. 45 (1953) 2782. C.L. Mehltretter and C.E. Rist, J. Agr. Food Chem. 1 (3).(1953) 779. C. Tanford, Physical Chemisrry of Macromolecules, Wiley, New York, 1963. M. Floor, Thesis, Delft University of Technology (1989). F.L.M. Smeets, Naruur en Techniek 58 (1990) 1. E. Santacesaria, F. Trulli, G.F. Brussani, D. Gelosa and M. DiSerio, Carbohydrate Polymers 23 (1994) 35. M. van Duin, J.A. Peters, A.P.G. Kieboom and H. van Bekkum, J. Chem. SOC.Dalron Trans. (1987) 205 I . M. van Duin, J.A. Peters, A.P.G. Kieboom and H. van Bekkum, J. Chern. Soc. Perkin Trans. I1 (1987) 473. M. van Duin. J.A. Peters, A.P.G. Kieboom and H. van Bekkum, Carbohydr. Res. 162 (1987) 65. Neth. Pat. 215.202 (1961); Chemical Absrracfs 56 (1961) 12682. Neth. Pat. 2 15. I80 ( 1 972); Chemical Abstracts 8 I (1974) 176040. P.J.M. Dijkgraaf. Thesis, Eindhoven University of Technology (1989). LA. Colon, R. Fernindez-Garcia, L. Amoros and H. Blay, J. Am. Chem. Soc. 71 (1949) 4131. A.E.J. D e Nooy, A.C. Besemer and H. van Bekkum, Recl. Trav. Chim. Pays-Bas 113 (1994) 165. A.E.J. D e Nooy, A.C. Besemer and H. van Bekkum, Carbohydr. Res. 269 (1994) 89. H.E.J. Hendriks, B.F.M. Kuster and G.B. Marin, Carbohydr. Res. 204 (1990) 121. US patent 3,629,121 (1971) M. Diamantoglou, H. Magerlein and R. Zielke. Tenside Deterg. 14 (1977) 250. D.L. Verraest, J.A. Peters, H.van Bekkum and G.M. van Rosmalen, paper submitted to J. Am. Oil Chemists' Soc. D.L. Verraest. J.A. Peters, J.G. Batelaan and H.van Bekkum, Carbohydr. Res. 271 (1995) 101. Brit. Pat. 1,330,122 (1973). Brit. Pat. 1,330,123 (1973). M.S. Nieuwenhuizen, A.P.G. Kieboom and H. van Bekkum, StarcWSrarke 37 (1985) 192. R. Kohn and K. Thilarik, Collectioti Czechoslovak Chem. Commun. 49 (1984) 21 16. M. Floor, K.M. Schenk, A.P.G. Kieboom and H. van Bekkum, StarcWSrarke 41 (1989) 303. M. Floor, A.P.G. Kieboorn and H. van Bekkum, StarcWStarke 4 1 (1989) 348. M. Floor, A.P.G. Kieboom and H. van Bekkum, Recl. Trav. Chim. Pays-Bas 108 (1989) 384. Patent application. EP-A1427349 (1990). Patent application, W.O. 17189 (1991). A.C. Besemer and H. van Bekkum, SrarcWStarke 46 (1994) 95. A.C. Besemer and H. van Bekkum, StarcWSrarke 46 (1994) 101. A.C. Besemer and H. van Bekkum, Recl. Trav. Chim. Pays-Bas 113 (1994) 398. C.H. Hullinger and R.L. Whistler, Cereal Chemisrq 28 (1951) 153. R.L. Whistler. E.G. Linke and S. Kazeniac. J. Am. Chem. Soc. 78 (1956) 4704. R.L. Whistler and R. Schweiger. J . Am. Chem. Soc. 79 (1957) 6460. R.L. Whistler and R. Schweiger, .I. Am. Chem. Soc. 80 (1958) 5701.

Calciurn seqiiestering agents based on ccirbohydrates 59. 60. 6I. 62 63.

J . Potze and P. Hiemstra, SrarcWStarke 15 (1963) 217. W.M. D o m e and R.L. Whistler, StarcldStarke 16 (1964) 177. 1. Ziderman and J. Bel-Ayche, Carbohydr. Res. 27 ( 1973) 34 1 . V.A. Uchtman and R.P. Oertel, J . Am. Chem. SOC.95 (1973) 1802. M.S. Nieuwenhuizen, A.H.E.F. Ebaid, M . van Duin, A.P.G. Kiehoom and H. van Bekkum. Tenside Dererg. 21 (1984) 221.

64. 65.

293

M. Koch, H Roper and H . Beck, SrarcWSfurke 45 (1993) 7. S. Matsumura: R. Kohayashi and S . Yoshikawa. Oil Chemistry, 55 (1991) 73.

This Page Intentionally Left Blank

16 Bleach activators

R.H.F. Beck’, H. Koch2 and J. Mentech-’ I

Eridania BCghin-Say, Vilvoorde Research and Development Centre - Cerestar, Havenstraat 84, B- 1800 Vilvoorde, Belgium 2 Eridania

BCghin-Say, Vilvoorde Research and Development Centre, Havenstraat 84, B- 1800 Vilvoorde, Belgium

3 Eridania BCghin-Say - Ceresucre, 27 bd du 1 1 Novembre 1918, PO Box 2 132, F-69603 Villeurbanne Cedex

Summary. Heavy-duty laundry detergents are mainly composed of the water softening builder/co-builder system, surfactants, the bleach/bleach activator system and some minor ingredients, such as enzymes, optical brighteners and stabilizers. The bleach activator is usually incorporated in modern heavy-duty compact powders at a level of 5.6% weight basis. The bleach activator supports the action of the bleaching agent, typically perborate or percarbonate, which releases hydrogen peroxide upon dissolution by reducing the temperature necessary to perform the bleaching reaction. The bleach system is designed to work on stains that are either not removed by surfactant action or chemically bound to the tissue. Bleaching reaction proceeds via the oxidative degradation of the chromophores. Typical examples of such bleachable stains are red wine and tea. The performance of the bleach system of a detergent is dependent on factors such as the dosage, pH-value, and washing temperature. With these variables in mind, a model is developed based on the principal physico-chemical data of the active bleaching species. Carbohydrates can easily be transformed into bleach activators by serving as acyl carriers. Carbohydrate esters of this type undergo perhydrolysis on contact with hydrogen peroxide under washing conditions. In addition to their bleach activator property, some carbohydrate-based bleach activators have anti-incrustation properties, meaning that they reduce the build-up of inorganic deposits on the fabrics.

Introduction Modern, heavy-duty compact powdered detergents contain a dozen or so ingredients‘, each of which contribute in its own way to the washing performance. When divided according to function, the ingredients can be classified in two major groups: the first group covers all those that remove stains such as surfactants, bleaching agents and enzymes; the second group consists of compounds

296

R.H.F. Beck, H. Koch and J. Mentech

that prevent organic and inorganic materials from being deposited on the fabric, such as zeolites, alkalis, silicates, and polycarboxylates. Focusing more on the specific mode of action of the first group: this group can be subdivided into products that remove stains by applying physical principles and products that make use of chemical reactions. The latter include, for example, enzymes and bleaching agents. This contribution covers some basic aspects of the chemistry of oxygen-based bleaches, the use of bleach activators with special focus on carbohydrate-based products, and the additional benefits carbohydrate-based bleach activators might have in comparison with petrochemicallyderived activators.

Bleaching systems

-

General aspects2

When studying the composition of a heavy duty compact powder detergent (Fig. l), the bleaching system comprises about a quarter of the total weight. The bleaching system must remove stains that cannot be removed by surfactant action by means of oxidative degradation.

Heavy duty compact powder detergent / Bleach activator

Builder / Co-builder Alkali /Corrosion inhibitors

Enzymes, Stabilizers Optical brighteners

Fig. 1. Heavy duty compact powder detergent.

The first oxygen-based bleaching agent374 that was introduced in heavy duty detergents was sodium perborate (Fig. 2 ) . Upon dissolution the perborate releases hydrogen peroxide, which only reacts with stains at temperatures above 80 "C.

Blench nctivutors

\ \

297

Perborate Inorganic Persalts + Activator (e.g. TAED; SNOBS)

\

Inorganic Persalts + Activator (e.g. TAED) + Accelerator

\

Organic Peracids (Salts)

Fig. 2. Bleach / bleach activator systems.

As second-generation bleaching system perborate or percarbonate in combination with the bleach activator TAED5 (tetraacetyl ethylene diamine) was introduced in Europe in the late 1970s. In 1994 the combination of percarbonate and TAED in combination with an 'accelerator'

-a manganese complex- was launched in a heavy duty compact powdered detergent' i n the UK, the Netherlands and France. Whether this combination will conquer a share of the market or whether alternative bleaching systems such as peracids will take the place the traditional bleaching systems, is a question only time can answer.

Bleach activation

- General aspects6

How is bleach activation perceived by the consumer? The bleach activator is put into the detergent formulation to make the bleaching system more effective. This higher effectiveness of the bleaching system allows the consumer to use lower washing temperatures while still obtaining the same efficiency in stain removal they obtained at 90 "C. Lowering the washing temperature is an on-going trend, which is driven on the one hand by the environmental consciousness of consumers and on the other hand by the dictate of fashion (Fig. 3). In the consumer's perception, the following benefits are related to lowering the washing temperature: -

energy savings: less energy is needed to heat the washing liquid. time savings: less time is required for heating, resulting in shorter wash cycles.

298 -

R.H.F. Beck, H . Koch and J . Mentech fabric and colour protection: colourful and delicate fashion fabrics require lower washing temperatures to retain their appeal.

Product development in the area of bleach activators must take these basic requirements into account. However, these general statements must first be translated into a technical, productrelated language. The translation in chemical terms would be something like: 'Develop a bleach system by means of which the fabric-bound chromophores are quicker and more easily oxidatively destroyed than by the sole action of hydrogen peroxide. Optimize according to temperature, time, concentration, redox potential, pH and so on.'

\ Same cleaning effect, but at lower temperature \ Energy saving \ Time saving \ Textile and colour protection \ Oxidative decomposition of tissue-bound chromophores, in function of temperature, concentration, redox potential, ...

Fig. 3. Bleaching activation: Consumer perception and chemist's translation

The chemistry of hydrogen peroxide and bleach activators7 As hydrogen peroxide is the parent molecule for oxygen-based bleaching systems, it is useful to take a closer look at its chemistry under practical washing conditions. Under the alkaline conditions of the washing liquid, hydrogen peroxide can react in three different ways (Fig. 4). It can react as a strong oxidizing agent or as a weak reducing agent, and it can exothermally decompose.

Bleach activators

HO;+ H,O + 2 e-

30H-

E,

HO;+ OH-

0, + H,O + 2 e-

E,

299

= 0.88 (strong oxidizing agent)

-

= -0.076 (weak reducing agent)

2H,O + 0, + 46.9 Kcal

2HP2

Fig. 4. Chemistry of alkaline hydrogen peroxide

Of these three possible reactions, only the first, in which hydrogen peroxide reacts as an oxidant, is desired for performing the bleaching reaction. The redox equation strongly suggests that the hydroperoxide anion is involved in the oxidative bleaching reaction. How then is the activation of hydrogen peroxide solved in practice? The hydroperoxide anion reacts with an ester or amide sufficiently activated to allow perhydrolysis (Fig. 5). This reaction releases an organic peracid, or to be more precise the anion of an organic peracid. Peroxocarboxylic acids are known to be more reactive than hydrogen peroxide in performing the bleaching reaction.

e.g. TAED 0

0 0 CH,-C\

C-CH,

N-CH,-CH,-N CH,-C'

;

, II

I1 II

'

+ 2 ~0;-

0

0

II

II

CH,-C-NH-CH,-CH,-NH-C-CH,

0 II

+ PCH,-C-0-0

C-CH, II

0

Fig. 5 . Chemistry of bleach activators. Molar conversion of hydroperoxide anion into peracetate anions.

300

R.H.F. Beck, H. Koch and J . Mentech The best-known bleach activator is certainly TAED. TAED stands for tetraacetyl

ethylenediamine, even though its correct chemical name is N,N'-ethanediylbis(N-acetylacetamide). TAED's 'secondary amide' structure strongly activates the acetyl substituents, so that two peroxoacetate anions are formed on contact with hydroperoxide anions. Noteworthy is the fact that the bleach activation is a process that converts hydrogen peroxide on a molar basis into peroxocarboxylic acids. The basic chemistry of the bleaching reaction, which is a type of oxidative attack on the chromophores of the stain, remains unchanged. The above-mentioned bleaching system that relies on an accelerator seems to work on the basis of another principle. Most likely the accelerator acts as a catalyst that changes the active bleaching species by liberating free radicals, such as the hydroxyl or the superoxide radical. lt is common knowledge that these free radicals are more reactive than either hydrogen peroxide or peroxycarboxylic acids. But let us return to some basic aspects in comparing the reactivity of hydrogen peroxide and peracetic acid. As shown earlier (Fig. 4), the hydroperoxide anion appears as the oxidizing species in the redox equation. From this it follows that the higher the amount of dissociated hydrogen peroxide, the higher the concentration of active bleaching species and the more pronounced the bleaching effect should be. Being a very weak acid, it is clear that the dissociation of hydrogen peroxide is strongly influenced by the pH-value and the pH buffering capacity of its environment. In principle i t should be possible to increase the dissociation of hydrogen peroxide and thus increase the amount of the active oxidizing species simply by raising the pH-value. Under practical conditions,

however, this possibility is highly limited, on the one hand due to the lack of stability of the hydrogen peroxide under very alkaline conditions and on the other hand due to the pHrequirements of some of the other detergent ingredients. But not only variation of the pH-value results in an increase in the dissociation of a weak electrolyte such as hydrogen peroxide. It is a well-known fact that the temperature has a pronounced effect on the dissociation of weak electrolytes. The neutral point of water at 100 "C, for example, is shifted by one pH unit from 7.0 to 6.0, which is equivalent to an increase in the dissociation constant by two orders of magnitude. In Fig. 6 the dissociation constants for hydrogen peroxide as a function of temperature are given. The dissociation constants were calculated using the reaction isobars, giving a quantitative correlation between the dissociation constant and the absolute temperature. As can clearly be seen, the dissociation of hydrogen peroxide increases significantly when the temperature increases from 60 "C to 90 O C . Using these temperature corrected dissociation constants in a real-life, heavy-duty compact powdered detergent, however, renders unambiguous results. At 25 "C only about 2% of

Bleach activators

30 1

the hydrogen peroxide is dissociated into the bleach-active hydroperoxide anion. At 60 "C the amount of dissociation rises to 5%, but is still very low. Only at 90 "C is a substantial amount of hydroperoxide anion, about 30%, formed. This sharp rise in the amount of bleach-active hydroperoxide anion, as predicted by the model calculation, has been confirmed in application tests. Hydrogen peroxide-based bleaches only exhibit a significant activity at temperatures around

90 "C.

25°C

60°C

90°C

0

I1

"202

H202 CH,C-OOH

H20,

pKa

11.62

11.0 8.2

9.9

pH Detergent

10.0

9.7

9.7

9.6

2 Yo

5%

97%

30%

o/o

Dissociation

Fig. 6. Dissociation of hydrogen peroxide and peracetic acid in function of temperature and pH-value in a formulated heavy duty compact powder detergent.

Peracetic acid, even though it is also a weak acid, is about three orders of magnitude more strongly dissociated than hydrogen peroxide. The effect on the dissociation of washing liquid pH is thus enormous. Practically all the peracetic acid is dissociated into the reactive peroxoacetate anion at temperatures around 60 "C. Peracetic acid therefore exhibits its maximum bleaching activity around 60 "C. A continued increase in the temperature obviously will not produce a significant increase in bleaching power. A relevant practical recommendation for the optimal amount of bleach activator to be used in

detergents is derived by comparing this set of data. Once again stressing the fact that only about one third of the hydrogen peroxide is dissociated at 90 "C and that peracetic acid is almost quantitatively dissociated at 60 "C, in principle one third of the amount of peracetic acid as compared to the amount of hydrogen peroxide is able to perform the same job. The optimal dosage of a bleach activator, expressed as a molar ratio of the hydrogen peroxide derived from perborate or percarbonate, to bleach activator should therefore be around 1:3. Analyzing some commercial

heavy-duty compact powdered detergents proved that this level of bleach activator (25-30 mol%) is indeed used.

302

R.H.F. Beck, H . Kocli and J. Mentech

Further requirements of a bleach activator In addition its function as an activator, which is a matter of course, bleach activators must fulfil a series of additional requirements (Fig. 7). By definition, bleach activators are reactive chemicals, but they must not react with any other ingredients during the detergent manufacturing process or during storage in either supermarkets or the home. Chemical stability and compatibility of a bleach activator is therefore of the utmost importance.

Stability Cornpatibility Spot test Solubility Dispersibility Solid Granulometry Bulk density Fig. 7. Additional physical and chemical requirements of a bleach activator.

When using a powdered detergent, obviously the detergent ingredients must pass from the solid state into solution in order to perform. This solubilization, however, is not an instantaneous process; some ingredients dissolve faster than the others. Bleach activators are inherently hydrophobic compounds, meaning that their solubility in water is low. By increasing the specific surface area that can be reached by finely dispersing the product, the speed of dissolution will be enhanced. A good dispersibility will thus largely overcome the problem of the low solubility of a bleach activator. Local over-concentration of bleach activator, which might cause some damage to the colours of the fabrics (known as spotting) can also be avoided by a good dispersibility. Last but not least, we must take the physical appearance of a bleach activator into account. Because bleach activators are used in powdered detergents, they should be solids or at least be

Bleach activators

303

adsorbable onto a solid carrier. For psychological reasons, bleach activators should also be as white in colour as possible. Moreover, the granulometry of a bleach activator is important since large differences in the granule size compared to the other granulated detergent ingredients could cause demixing of the detergent powder. In the age of compact powdered detergents, the bulk density is, of course, also of interest.

Carbohydrate-based bleach activators To arrive at the desired properties as outlined above, there are several possible approaches to designing carbohydrate-based bleach activators. Peracetylation is one possibility, as shown in Fig. 8. Typical representatives are pentaacetyl glucose (PAG)8 and a product to which we gave the name SORMAN9. SORMAN is a mixture of peracetylated sorbitol and mannitol, derived from about 75% sorbitol and 25% mannitol. This particular mixture was chosen based on its better performance as compared to the individual, pure, peracetylated hexitols. Most probably this mixture forms a less-perfect crystalline structure, causing a quicker dispersion into very small crystallites upon contact with water, which then undergo perhydrolysis at a higher rate. CH,-

a

OAC

(pO& H-C

I

AcO-C

AcO

I

H-C

I

I H-C I

6Ac

-0Ac

Sucrose

-H

-0Ac -0Ac

CH,-

OAc SUPA

SORMAN

PAG

II

AcO-C

I

I

I H-C I

H-C

I

I

I

C - N -

H-C

R

Ac

0

CH,

-0Ac

-H -0Ac -0Ac

CH,-

CH,-N

OAc

N-Methyl gluconamide

H-C

I I

- CH,

/""

-0Ac

I

AcO-C

-H I I

H-C H-C

I I

NH-

-0Ac -0Ac

CH,- OAC N-Methyl glucamine

Fig. 8. Examples of carbohydrate-hased bleach activators

AcO OAc Glucosylamine

Ac

304

R.H.F. Beck, H . Koch and J. Mentech

A second possibility for increasing the reactivity and overcoming the low solubility is partial acetylation of a carbohydrate. A representative of this class of compounds is a product that we called SUPA'O.1 I . SUPA stands for sucrose polyacetate. Its degree of acetylation amounts to about 60-75% of the theoretically-possible peracetylated structure. SUPA has no defined chemical structure; most of the possible isomers are formed during the partial acetylation of sucrose and are actually present in the product. A third strategy for increasing the effectiveness and reactivity of carbohydrate-based bleach activators is to introduce nitrogen atoms into the basic carbohydrate skeleton. This functionalization yields products which in their peracetylated form have a significantly higher aqueous solubility than their peracetylated carbohydrate analogues. Three industrially-feasible possibilities are shown in Fig. 8, i.e. the peracetylated forms of N-methyl gluconamideI*, N-methyl glucaminel3, and glucopyranosyl amineI4.

Additional benefits of some carbohydrate-based bleach activators Surprisingly, we have found that one of the evaluated carbohydrate-based bleach activators had an additional property that can be exploited for detergent applicationsg. In addition to its bleach activator function, the product SORMAN shows some co-builder properties. Co-builders were introduced into detergent formulations when phosphates were substituted by zeolites. To retain the same performance level, a co-builder must be added into zeolite-built detergent formulations to compensate for the lower calcium and magnesium binding kinetics. Cobuilders proved to have two different effects in the detergent, i.e. the inhibition of the growth of calcium carbonate crystals and preventing deposits from forming on the fabrics. In Fig. 9 the build-up of inorganic incrustation during 25 washing cycles - which is considered a measure for the performance of a co-builder - is shown. A formulation without cobuilder and one with addition of the standard co-builder (a synthetic polycarboxylate) were used as reference points. In two test series, 100% and 75% of the polycarboxylate co-builder were substituted by SORMAN. As is evident in Fig. 9, the addition of SORMAN has a strong effect on reducing the formation of inorganic incrustation. At a substitution level of 75% of the polycarboxylate by SORMAN, the performance level is virtually unchanged. Full or partial substitution of polycarboxylates is of high interest, because this class of polymers is not biodegraded but only bio-eliminated.

Bleach activators

305

inorganic incrustation %' . 2

5

10 Without cobuilder +4%SORMAN

P

15 /////

25 washing cycles

+ 4% Polycarboxylate + 3% SORMAN + 1 % Polycarboxylate

Fig. 9. Anti-incrustation properties of SORMAN 60 "C, 45 "F, test in washing machines

This additional benefit of using a carbohydrate-based bleach activator is a potential incentive for the detergent industry for switching from their traditional TAED-based formulations and undertaking the expense and the risk of reformulating their detergents. Carbohydrates as multifunctional industrial raw materials is certainly a theme15.16, which many in addition to this forum would like to see developing.

Outlook and future trends in the bleach activator market Market studies17 in general foresee a positive future for bleach activators (Fig. 10).New markets, such as automatic dish-washing detergents, will also contribute to a prosperous future for bleach activators. European consumption at this time is estimated at about 75,000 tons per year. The expected growth rate for bleach activators is 6% per year until the year 2000. Whether or not part of the European bleach activator market that is currently solely held by TAED can be substituted by alternatives - hopefully carbohydrate-based bleach activators - only the future knows.

Acknowledgement. A multinational task force 'carbohydrate-based bleach activators' has contributed to this work. Special thanks are due for the contributions by Dr. M. Elseviers (Cerestar), Dr. F. Burzio (Ausimont), Dr. E. Wong (BCghin-Say), Dr. I. Janicot (E.S.C.I.L.) and Prof. G. Descotes (E.S.C.I.L.)

306

R.H.F. Beck, H. Koch and J. Mentech

\

West European consumption of about 75.000 t/year

\

Favourable growth rate of about 6% 1year predicted

\

Application in new market segments, e.g., automatic dish washing detergents broadens the use of bleach activators

Fig. 10. Market prospects for bleach activators.

REFERENCES I.

2. 3. 4. 5. 6. 7. 8. 9. 10. 1 I.

12. 13. 14. 15. 16. 17.

Perner J., in Proceedings of the 3rd World Conference on Detergents: Global Perspectives, edited by A. Cahn, AOAC Press, Champaign, Illinois (1994) 168. Kuzel P. and Lieser Th., Tenside S u e Det.. 27 (1990) 23. Basset F., lnfo Chiniie 361 (1994) 204. Jiirges P., in Proceedings of the 3rd World Conference on Detergents: Global Perspectives, edited by A. Cahn, AOAC Press, Champaign, Illinois (1994) 178. Reinhardt G . , Schuler W. and Quack J.M., Corn. Jorn. Corn. Esd. Deterg., 20 (1989) 165. Hauthal H.G., Schmidt H., Scholz H.J., Hoffmann J. and Pritzkow W., Tenside Sud Det., 27 (1990) 187. James A.P. and Mackirdy I S . , Chetnistry & Industry, (1990) 641. The Procter and Gamble Company, EP-B 0 095 904. Ausimont S.p.A.. EP-A 0 525 239. Eridania Bkghin-Say, PTC, WO 93/01200. Ausimont S.p.A., EP-A 0 492 000. Ausimont S.p.A., EP-A 0 5 17 969. BASF AG, DE 4308123. Ausimont S.p.A.. EP-A 0 600 359. Koch H., Beck R. and Roper H., StarcWStarke. 45 (1993) 2. Koch H. and R6per H., StarcWStdrke, 40 (1988) 121. Colin A. Houston & Associates Inc., West European Household Detergent Ingredients - Surjactants to 2000 with Builder and Bleach Systems, 1992.

Subject Index A A. Niger 270 Acarbose 207 Acetic acid 117, 122, 132, 185 Acetogenesis 204 Acrylatehtadiene 29 Acrylic acid 28, 183 Acrylic glucamide 143 Additives 274 Adhesives 20 Adipates 23 Aerobic microorganisms I 18 Aerobic sewage plant 126 Agro-industry 126 Alcohol ethoxylates 248 Alcohol intoxication I I 1 Aldaric acids 281 Aldonic acids 39 Aldose oximes 57 Aldoses 39 Alginate 7 Alginic acid 28 1 Alkaloids 1 17 Alkylbenzenesulphonate 248 Alkyl glucuronic acids 249 Alkyl glycosides 42 Alkylphenol ethoxylates 248 Alkylpolyglucoside 248, 252, 256 Allolactose 107 Aluminosilicate 276 Aminodeoxypolyols (alky1)- 56 Ammonia 81, I l l , 118, 130, 144 Ammonium: chloride 56 phosphate 130 Amphiphilic molecules 259 a-Amylase 53, 125, 196 P-Amylase 53, 122 Amylolysis 204 Amylopectin 18, 169, 190, 196 Amylose 18. 169, 174, 190, 196 helix 170, 175, 178 Amylose-iodine complex 174 Anaerobic fermentations I 18 Anaerobic sewage plant 126 Antibiotics 112, 117, 199 Arabinanases 13 Arabinans 5 Arabinofuranohydrolase I0 Arabinofuranosidase 5 , 9 , 13

1

Arabinose I1,47 Arabinoxylans 2.9 Arabitol 55 Aromas 173 Arthobacter Globiformis 87 Arthobacter Urefacienr 87 Artificial kidneys 86 Ascorbic acid (Vitamin C) 37, 42.44, 52 Asparagus 74 Azobisisobutyronitrile (AIBN) 161

B B-Vitamins 130 Bacillus Circulans 87 Bacillus Coagulans 130 Bacillus Subtilis 89, 270 Bagasse 126 Banana 75, 85, 190, 195, 197, 200, 205, 209 Barium hydroxide 7 Barley 1.75 Bean flakes 192 Beer I , 138 Beet sugar 123 I-Benzylamino-I-deoxypolyols56 Beverages 18,2I7 Bimetal catalyzed oxidation 38 Bifidobacteria 83, 108. 112 Bifidus 83 Binders 27 Bio- and chemo-catalytic 37, 55 Bio-diesel 59 Bio-ethanol 122, 126 Biocompatibility 25, 33 Biodegradability 21, 25, 33, 57, 87, 247, 251, 269, 284 Biodegradable 28, 137,247,276 Biotechnology 1 15 Biscuits 196 Bismuth 43 Bismuth-promoted palladium 40 Bitter aftertaste 101 Bitumen 28 Bleaching 24 I , 274, 295 Blood 135 Blood-glucose 110 Boric acid 5 1,280 Bread 190, 195, 197,200, 205,209 Briquettes 27 Bromine 39 Br~nstedacidity 53

308

Subject index

Builders 274 Burdock 75 2.3-Butanediol 86 Butanol 122, 172 rerr-Butanol 172 Butter 138 Butyl acrylate 144, 146 4-tert-Butylphenol 179, 183

C Calcium hydroxide 55 Calcium oxalate 278 Calories 220 Calorimetry 228 Cane sugar 123 Carbon disulfide 143 6-Carboxy-cellulose 37 5-Carboxyfurfural 46 5-Carboxyfurfuryl alcohol 46 Carboxymethyl: cellulose 282 inulin 284 starch 282 Carboxymethylated amylose 177 Carboxymethylation 88 Cardboard 18.26 Cardiovascular diseases 23 Caries 23, 108 Casein 94 Caustic soda 130 CelV catalyzed 5 1 Celite (Diatomaceous earth) 39 Cellobiose 46 P-benzyl 44 Cellulose 2, 26, 75, 83, 169, 284 Cement 20 Ceramics 27 Cereal bran 1 Chalk 130 Cheese 94 Chewing gum 56, 101 Chicory 68, 77 Chlorophyll 179 Chocolate 102 Cholesterol 83, 173, 205 Cholesterol-free 23 Cholesterolemia 205 Chromic acid 46 Chromic oxide 200 Chromium trioxide 48

Cider 138 Circular dichroism 175, 178 Cirrhosis I 1 I Citric acid 115, 117, 120, 124, 218, 256, 276 Cleaners 231 Clostridiu Perfringens 83 Coal 27 Coatings 141 Cobalt 39 Coffee substitute 77 Colloids 80 Colon 110 Colon cancer 23 Compositae 74 Compostable materials 3 1 Constipation I I 1 Controlled drug release 32, 137 Copolymer I37 Copolymerization 161 Copper 39 Copper oxide 46 Corn 123 Corn flakes 190, 195, 197, 205 Corn starch 53 Cornsteep liquor 130 Corrugated board 20 Cosmetics 5, 138, 173, 231 Cotton 245 Cotton-effect 178, 184 Critical Aggregation Concentration (C.A.C.) 267 Critical Micelle Concentration (C.M.C.) 15 I , 26 I , 267 Cross-linking 172, 181, 196 Crown ethers 87 Cryoprotectant 226 Crypt cell proliferation 207 Crystal lattice 263 Crystallinity 20 Crystallization 267 Crystallization inhibitor 284 Cyclodextrin 87, 170, 178, 284 Cyclofructan 87 Cycloinulo-hectaose 87

D Dahlia 71 Dandelion (Tururucum Offerciriule Weber) 75, 78 De-rusting 39 Degradable 277 Degradation 51 Dehydrogenation 40, 56.58

Subject index Deoxycholic acid 207 Depolymerization 197, 228, 286 Detergent 23 1, 255, 274, 295 Detoxification 207 Dextran I17 Dextrin 25, 125, 141, 151, 196, 284 Dextrose 33, 129 Di-tert-butylperoxide (DTBP) 161, 164 Diabetes 23 Diabetic food 83 Dianhydrosorbitol 52 Dibenzoylperoxide (BOP) 161 Dicarboxy : inulin 49, 87. 256, 288 starch 49,256 -P-cyclodextrin 289 Dicarboxymethyl: cellulose 282 starch 282 Dietary fibre 83 Differential Scanning Calorimetry (DSC) 26 1 Diglucosylamine 56 Dilactide 136 Dimethyl sulfoxide 185 Diol-cleavage 49 Dishwashing 233, 245 Disposable diapers 3 I Dressings 85 Drug carrier 86, 89 Drugs 173

E

Exo-glucal 156 Exo-ribene 156 Exohydrolase 78 Extruder 32

F

i

E. Coli 83, 108, 270 Early cultivars 75 EDTA (Ethylenediamine tetraacetate) 10 EDTA-like 57 Electrochemical oxidations 38 Electrodes 28 Electrodialysis 45 Emulsifiers 233 Enzymatic synthesis 115 Epichlorohydrine 86 Esterification 196 Etherification 196 Ethylene 183 Ethylene glycol _ . 58 EURESTA (European Resistant Starch Research Group) 189, 206 Eutrophication 275 Exo-fructal 156

I

Fat replacer 23, 85, 217 Fatty alcohol ethoxylate 248 Fenchone 175 Fermentation 115, 123, 129, 204 Ferrocyanide 124 Fertilizers 126 Ferulic acid 2 Flavors 173, 183, 219 Flour 85 Foaming capacity 246 Freeze-thaw procedure 161 Freezing point depression 227 Fructan 68, 74, 78, 88 Fructan hydrolases 79 Fructo-hex- I -enopyranose: 3,4,5-Tri-O-benzoyl-2-deoxy-DI 59 Fructose 68, 102, 104, 12, 191, 227 D-Fructose 4 I , 45 Fructosyltransferases 78. 87 Fungicides 173 Furan-2,5-dicarboxylic acid 46 Furanoid vinyl ethers 156 Furanoside: methyl 5-deoxy-2,3-0-isopropylidene0-D-erythro-pent-4-eno- 160 Fusobacteria 84

G Galactanases 13 Galactaric acid 6 Galactitylamine 56 4-O-~-D-Galactopyranosyl-D-glucitol. 108 4-O-~-D-Galactopyranosyl-D-gluconic acid (lactobionic acid) 37, 112 Galactose 4, I I , 76, 105 D-Galactose 40.93 P-Galactosidase 107, 109, 122 Galacturonic acid 4, I 1 Galarose 51 Garlic 74, 85 Gel formation 18 Gel stability 18 Gelatinization 17 Gelatinization behavior 169

309

3 10

Subject index

Gelation 2, 9, 13 Gellan gum 281 Gelling substances 89 Genetic engineering 33, 119 German hardness 243,245,246 Glucal 156 Glucamides 15 1, 233 Glucamine 143 N-methyl 304 Glucan 53 a-Glucans 194,201 P-Glucans 2 Glucaric acid 45,256, 279 2,4-di-methylene- 280 D-Glucitol 52, 269 hexa-0-acetyl D- 256 I-amino-I-deoxy-D- (Glucitylamine) 56 Glucitols 264 I-0-alkyl-D- 264 3-0-alkyl-D- 264 4-0-alkyl-D- 264 Glucitylamine 56 Glucoamylase 122, 202 Glucofuranose: 1.2-0-isopropylidene-a-D-44 Glucofuranuronicacid: 1,2-0-isopropylidene-a-D44 Glucofuranurono-3,6-lactone: 1,2-0-isopropylidene-a-D-5-keto44 Gluconamide: N-methyl 304 Gluconate: methyl 2,3,4,6-tetra-acetyl-5-keto-D48 D-Gluconic acid 37, 39, 45, 1 16, 122, 279 2-keto- 45, 49 2,4-3,5-di-methylene- 280 5-keto-- 37.49 D-Glucono-&lactone 39 Glucono-N-alkyl amides 39 1-phosphate a-D-Glucopyranose 43 D-Glucopyranoside 269 Glucopyranosyl amine 304 Glucose 11, 18. 29, 33, 37, 68, 93, 102, 115, 122, 130, 175, 191, 196, 218, 227, 232 D-Glucose 37,93 Glucose isomerase 55 Glucoside: butyl 236, 238 Glucoside: (continued) dodecyl-a- 235 dodecyl-p- 235

Glucuronic acid 9,43 Glues 20 Gluing 17, 26 L-Glutamic acid 120 Gluten 126 Glycemia 204 Glycemic index 205 D-Glycenc acid 5 I Glycerol 32, 58, 222 I-amino-I -deoxy-D-Glycitols 264 Graft-copolymers 29 2-keto-L-Gulonicacid 45 Gums 83,281 Gut I l l , 135 Gypsum (board) 20, 32,124

H Hair conditioners 32, 139 Hemicellulose 75, 83 Herbicides 135, 140, 173 Herbicides 173 1,2,5,6-Hexanetetrol 59 2-Hexanone 179 Hexitols, peracetylated 303 High Pressure Anion Exchange Chromatography (HPAEC) 69 Homogalacturonans 13 Hurnectancy 224 Humectant 224 Hydrobromic acid 39 Hydrogen peroxide 298 Hydrogenation 38, 108 Hydrogenolysis 52 Hydrophilic-lipophilic balance 258 Hydrophilicity 251 2-Hydroxypropionic acid (lactic acid) 6, 116, 118, 127, 129, 279 Hydroxy-5-methyl-3(2H)-furanone4- 49 poly-P-Hydroxybutyrate I 18, 126 Hydroxymethylfurfural (HMF) 46, 87, 21 8 Hydroxypropylated di-starch phosphates 23 Hylon-VII 207 Hypobromite oxidation 281 Hypochlorite 42 oxidation 286

1 Ice-cream 24, 85, 106, 227 Ileostomates 194 Incontinence pads 3 I

L

reaction 80

3 12

Subject index

Marmalades 139 Mayonnaise I39 Membrane filtration 33 Mercuric acetate 39 Metal-catalyzed oxidation 38 Methacrylic acid 150 Methane gas 1 10 N-Methyllauroylglucamide 233 Methyl methacrylate 144, 146, 161, 162 Micelles 259 Microbial transformation 1 15 Microcalorimetry 180 Microencapsulation 174 Milk 23, 105, 129 mammalian 93 Milksugar 93 Miller indices 97 Modified starches 17, 196 Modified Srurm test 269 Moisture sorption 102 Molasses 28, 32, 123, 130 MOLCAD 170 Molecular encapsulation of drugs 174 Molecular Lipophilicity Potential (MLP) 170 Molybdate 56 Monoclonal antibody production 123 Monosodium glutamate I15 Morphology 26 Mortar 20 Mottling effect 29 Moulds 33 Mouth 193, 202 Mouth-feel 23, 89 Mucosa 107, 196 Mulch foils 31 Multiangle Laser Light-Scattering (MALLS) 164 Murnong 75 Mutarotation 99 I

N a-Naphthol 172, 175 P-Naphthol 175 Naphthyl isocyanate I84 Native starches 17 Nickel 39 peroxide 42.50 Nicotine 173 Nitric acid 50 oxidation 280 Nitrilotriacetic acid (NTA) 256.276

Noble metal catalysts 38 Nutritional value 23 Nystose 73

I

0 ~

Obesity 23 Olives 138 Onion 74, 85 Optical activity 129 Optically active 99 Optisweet 22 55 Organoleptic properties 24 Oxalic acid 52 Oxidation 37 C-6 Oxidation 42 Oxidized starch 37

P Paint-stripping 39 Paints 28, 231 Palladium 39 Pnlatinit (Isomalt) 53 Palatinose (Isomaltulose) 45, 53 Pancreatic amylase 189 Paper 26, 141,231 coating 28, 152 Paraffinsulphonate 248 Partalsystemic encephalopathy 1 1 1 Pasta 209 Peas 195 Pectic acid 28 1 Pectin 4, 83, 281 acetyl esterase 13 methyl esterase I3 4-methyl-2-Pentanone I80 Pepsin 194 Peptides 80 Peracetic acid 300, 301 Peracetylated hexitols 303 mannitol 303 sorhitol 303 Periodate oxidation 49, 86,218 Peroxidase 5 . 9 Peroxocarboxylic acids 299 Pesticides 173, 183, 231 Pharmaceuticals 5, 135, 138, 141 Phenolphthalein 179 Phosphorus oxychloride 22 Photosynthates 89

Subject index Pickles 138 Pigments 22 Pith 28 Plaque 108 Plasticizers 2 18 Plastics 231 Platinum 39, 249, 280 black 38 catalysts 38 catalyzed oxidation 88 Polyacrylates 28, 141, 284 Polycarboxylates 277, 296 Polyester 136, 245 Polyethyleneglycol (PEG) I83 Polyglucuronate 285 Polylactic acid 136 Polymerization 21 8 Polystyrene 142 Polyurethane foam 59 Polyvinylalcohol (PVA) 183 Polyvinylsaccharides 155 Potato 17, 76, 89, 123, 190, 195, 197, 200, 205, 209 amyloses 176 Preservative 138 Procainamide 86 1.2-Propanediol 58 Propylene glycol 222 Proteins 169 Pseudoglucal 156 I-ethoxy- 158 I -ethoxy-4,6-di-O-acetyl- 158 I-methoxy-4,6-di-O-acetyl158 1,4,6-tri-O-acetyl- 158 Pseudonionas 87, 1 I2 Pullulanase 189 Pulsed Amperometric Detector (PAD) 69 Pulsed Electrochemical Detector (PED) 71 Pyranoid ally1 ethers 156 Pyranoid vinyl ethers 156 Pyrrolidone 165

Quark 94

R Radical reactions 155 Raftiline 70 Raftilose 81 Raftisweet 81 Raney Nickel 53,55, 109

Recombinant human proteins 123 Reconstitution 267 Reduction 37 Reductive amination 38, 57, 264 Residual starch 193 Resistant starch 23 Retrogradation 226 tendency 18 Retrograded amylose 206 Rhamnogalacturonan 5, 13 acetyl esterase 13 Rhamnose I I Rheology 20, 241 Rhodium 40 D-Ribitol 264 Riboflavin (Vitamin B2) 96 Ribose 47, 160 Rice 19, 191, 197, 200, 209 Ruff degradation 51 Ruthenium 40 tetroxide 46 Rye 75, 85 bran 1 bread 138

S S. Cerevisiae 270

Saccharin 101 Saccharose 69 Salad dressing 23, 139 Salad sauces 23 Salivary a-amylases 191 Salmonella 108 Salsify 75 Sauerkraut 138 Sausages 138 Semicarbazides 15I Semicarbazone: amphiphiles 151 coupling 151 Sequestering: agents 89 capacity 57 Sewage 126, 275,277 Shigella 83 Shredded corn 123 Silicates 296 Silver: carbonate 39 fluoride 160

3 13

3 14

Subject index

Silver: (conrinued) oxide 46 Skimmed milk powder 103 Skin moisturizer 139 Small intestine 24, 83, 110, 190, 193, 196, 199, 202 Smith degradation 218 Sodium: azide 144 bicarbonate 5 5 borohydride 7.56.264 dodecylphenylsulfonate (DPS) 183 dodecylsulfate (SDS) 146, 184 hydroxide 55. 1 10 hypobromite 286 hypochlorite 286 perborate 296 periodate 46,286 trimetaphosphate 22 tripolyphosphate (STPP) 275 Solubilization 267 Solubilizers 233 Sorbitan esters 233 Sorbitan fatty esters 256 Sorbitol 32, 37, 52, 58, 102, 117, 218, 222, 227, 232 peracetylated 303 Sorbofuranose: 2,3-O-isopropylidene-L- 45 Sorbopyranoside: methyl a-L-45 Sorbose 37, 45, 52, I17 2.3:4,6 di-0-isopropylidene-L- 45 Sorman 303 Spaghetti 195, 205 Specific rotation 99 Spent grain 1 SporoluctobucillusSp 13 I Spray-drying 80 Spreads 85 Starch (maize) 53 Starch 17. 29, 33, 42, 74, 122, 125, 141, 169, 232, 284 hydrolysis 1 15 Sterilization 125 Stomach 193,202 Styrene 161 Styrenelbutadiene 29 Succinic acid 132 anhydride 86 n-octenyl 23, 32, 147, 151 Sucrose 101, 110. 117. 122, 130, 191, 223, 227, 232 monocarboxylic acid 45

Sucrose (conrinued) polyacetate 304 Sugar beet 77 Sulfur trioxide 250 Sulfuric acid 50, 81, 250 Sulphation 250 Supersaturation 98, 100 Surfactant 39, 231, 245, 257, 274, 295 Surgery 137 Sweeteners 217 Sweetness 101, 229 Swelling 18 Synergisms I41 Synergistic effects 101 Synthetic polymers 28

T TAED (Tetraacetyl ethylenediamine) 256,297,300 Talopyranoside: I-thio-a-D- 261 Tanford plot 278 Taraxacum Offercinale Weber 75.18 Tartaric acid 279 Taste enhancer 85 TEMPO (2,2,6,6-Tetramethyl-l-piperidiniloxy)88, 28 1 Tensile strength 22 Terminal ileum 204 Tetraacetyl ethylenediamine (TAED) 256, 297, 300 Thermoplasticstarch 32 D-Threaric acid (dextro-tmaricacid) 52 Tobacco 89 TOF-MALDI-MS (Time Of Flight Matrix Assisted Laser Desorption Ionization Mass Spectroscopy) 235 Toxicity 269 Transesterification 25 1 Transgalactosylic action 107 Transglucosidation 237,240 Transglycosylation 25 Transmission Electron Microscopy (TEM) 145, 149 Transplantation I12 Tricarboxysucrose 44,256 Triglyceridemia 205 2,4,6-Triiodobenzoic acid 58 Tryptophan I0 1 Tuftedcarpets 28 Tungstate-H202 50

sugars 157 Urine 86,135

X

Vanadium pentaoxide 47 Vanillin 3 Vegetable oils 255 Vesicles 259 Vibrio Cholerae 83 Vinasse 126 Vinylpyrrolidone 161 Vinylsaccharides 155 223,228,243 Viscosity 2, 4,7, 13,17, Vitamin A acetate 175 Vitamin B2 (Lactoflavin, Riboflavin) 57,96 Vitamin C 37,42,44,52 Vitamins 183

w

Xylitol 56,101, 264 Xylo-hex-5-enopyranoside: I ,2,3,4-letra-0-acelyI-6-deoxy-fi-D158

1,2,3,4-tetra-O-benzoyl-6-deoxy-fi-D158 I ,2,3,4-tetra-O-methyI-6-deoxy-a-DI58 Xylose 10, I 1 D-Xylulose 55

Y Yacon 74

Z

This Page Intentionally Left Blank