Coffee Volume 1 Chemistry

COFFEE

Volume 1: Chemistry

Volume 1: Chemistry Volume 2: Technology Volume 3: Physiology Volume 4: Agronomy Volume 5: Related Beverages

COFFEE

Volume 6: Commercial and Technico-Legal Aspects

COFFEE Volume 1: CHEMISTRY

Edited by

R.J.CLARKE Formerly of General Foods Ltd, Banbury, UK

and

R. MACRAE Department of Food Science, University of Reading, UK

ELSEVIER APPLlED SCIENCE LONDON and NEW YORK

ELSEVIER SCIENCE PUBLlSHERS LTO Crown House, Linton Road, Barking, Essex IG11 8JU, England

So/e Distributor in the USA and Canada ELSEVIER SCIENCE PUBLlSHING ca., INC.

655 Avenue of the Americas, New York, NY 10010, USA

WITH 70 TABLES ANO 38 ILLUSTRATIONS

1985 ELSEVIER SCIENCE PUBLlSHERS LTO First Edition 1985

Reprinted 1989 Softcover reprinl of Ihe hardcover 1 si edilion 1985

British Library Cataloguing in Publication Data

Coffee.

II. Macrae, R. VoI. 1: Chemistry 1. Clarke, R. J. 641.3'373 TX415

Library of Congress Catalog ing in Publication Data

Coffee.

Bibliography: p. Includes index. Contents: v. 1. Chemistry-1. Coffee. 1. Clarke, R. J. (Ronald James)

II. Macrae, R. TP645.C64 1985 641.3'373 85-6976 ISBN-13: 978-94-010-8693-6 e-ISBN-13: 978-94-009-4948-5 DOI: 1 0.1007/978-94-009-4948-5

No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein.

Special regulations for readers in the USA This publication has been registered with the Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. AII other copyright questions, including photocopying outside of the USA, should be referred to

the publisher.

AII rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the

publisher.

Preface

The term 'coffee' comprises not only the consumable beverage obtained by extracting roasted coffee with hot water, but also a whole range of intermediate products starting from the freshly harvested coffee cherries.

Green coffee beans are, however, the main item of international trade (believed second in importance only to oiI), for processing into roasted coffee, instant coffee and other coffee products, prepared for local consumers. The scientific and technical study of coffee in its entirety therefore involves a wide range of scientific disciplines and practical skills.

It is evident that green coffee is a natural product of great compositional complexity, and this is even more true for coffee products deriving from the roasting of coffee. The present volume on the chemistry of coffee seeks to provide the re ader with a full and detailed synopsis of present knowledge on the chemical aspects of green, roasted and instant coffee, in a way which has not been attempted before, that is, within the confines of a single volume solely devoted to the subject. Each chapter is directed towards a separate generic group of constituents known to be present, ranging individually over carbohydrate, nitrogenous and lipid components, not forgetting the important aroma components of roasted coffee, nor the water present and its significance, together with groups of other important components. Each chapter has been written by an expert in that particular field and this has resulted in compilations of the considerable amount of information published, but only previously available from different sources and in different languages. It is hoped that the present volume will provide a convenient and readable source of

v

Vi PREFACE

reference in the English language for alI those interested in the chemistry of cofTee. Coffee and its products have traditionalIy been substances upon which researchers have flexed their 'analytical muscles', and for this reason it has not been possible to include alI the minor components reported; some omissions are therefore inevitable.

Chemistry is only one facet of cofTee, so that the General Editors plan to provide similarly structured volumes in this series to cover its technology, physiological effects, agronomy and commercial/technico-legal aspects, and to include a volume on cofTee-related beverages, which are of consumer significance.

R. J. CLARKE R. MACRAE

Contents

Preface

List of Contributors

Abbreviations

Chapter 1 Introduction A. W. SMITH 1. Origins 2. The Coffee Plant

2.1. Species and varieties 3. Producing Countries

3.1. North/Central America 3.2. South America 3.3. Africa 3.4. Asia . 3.5. Oceania

4. Agricultural Practices 5. Processing at Origin

5. 1. Wet processing . 5.2. Dry processing . 5.3. Finishing processes

6. Roasted Coffee 7. Soluble (lnstant) Coffee 8. Decaffeination . 9. Composition

10. Physiological Effects II. Coffee Quality . 12. Coffee Substitutes

References vii

v

Xli

XIV

1 3 3 6

10 12 13 17 18 18 19 20 21 22 23 26 31 32 34 36 39 41

viii CONTENTS

Chapter 2 Water and Mineral Contents R. J. CLARKE 1. Introduction 42 2. Water Content of Green Coffee 43

2.1. Oven methods of determination 43 2.2. Entrainment distillation 48 2.3. Karl Fischer determination 49 2.4. Non-destructive methods . 50 2.5. Sorption isotherms 51

3. Water Content of Parchment Coffee 54 4. Water Content of Roasted Coffee 55

4.1. Sorption isotherms 55 5. Water Content of Instant Coffee 57

5.1. Methods of determination 57 5.2. Sorption isotherms 60 5.3. Fusion and collapse temperature 62

6. Water Content of Coffee Extracts 64 6.1. Water activity 64 6.2. Direct determination of solubles content 66 6.3. Specific gravity and refractive index of extracts 67 6.4. Viscosity 68 6.5. Diffusivity 69 6.6. Freezing point depression 70

7. Mineral Content of Green and Roasted Coffee 72 8. Mineral Content of Instant Coffee 76 9. Trace Elements in Coffees 79

References 81

Chapter 3 Carbohydrates L. C. TRUGO 1. Carbohydrates of Green Coffee 83

1.1. Low molecular weight sugars 83 1.2. Polysaccharides 85 1.3. Pectins and lignin 90

2. Carbohydrates of Roasted Coffee 91 2.1. Low molecular weight sugars 91 2.2. Polysaccharides 92 2.3. Carbohydrate conversion products 94

3. Carbohydrates of Coffee Brews, Extracts and Instant Coffee 95 3.1. Low molecular weight sugars 97 3.2. Polysaccharides 98 3.3. Carbohydrate conversion products 102

4. Some Physical Properties of Coffee Carbohydrates 106 5. Determination of Carbohydrates 109

References II 3 Chapter 4 Nitrogenous Components R. MACRAE

1. Introduction 115 2. Alkaloids (Caffeine) . . 116

2.1. Caffeine content of green, roasted and instant coffees 118 2.2. Physiological effects of caffeine 123 2.3. Determination of caffeine . 124

CONTENTS ix

3. Trigonelline 127 3.1. Determination of trigonelline 131

4. Nicotinic Acid . 132 4.1. Levels in green, roasted and instant coffee 132 4.2. Nutritional significance of nicotinic acid in coffee 134 4.3. Determination of nicotinic acid 135

5. Proteins and Free Amino Acids 137 5. 1. Proteins 138 5.2. Enzymes 143 5.3. Pigments 144 5.4. Free amino acids 147 References 149

Chapter 5 Chlorogenic Acids M. N. CLlFFQRD 1. Introduction and Brief History 153 2. Chlorogenic Acids Nomenclature 156 3. Chemical Synthesis . 158

3.1. Preparation of the protected acyl chloride 158 3.2. Preparation of the protected quinic acid 158 3.3. Esterification reactions 160 3.4. Acyl migration as a synthetic method 160

4. Physical Properties 160 4.1. Solubility and partition coefficients 160 4.2. Dissociation constant 161 4.3. Crystal form and melting points 163 4.4. Polarimetric data 165 4.5. Infrared spectra 165 4.6. Mass spectra 165 4.7. Nuclear magnetic resonance spectroscopy 165 4.8. Ultraviolet spectroscopy 167

5. Origin and Function 167 5.1. Biosynthesis 167 5.2. Function 170

6. Chlorogenic Acids Extraction and Analysis 171 6.1. Size reduction 171 6.2. Extraction 171 6.3. Possible artefacts 172 6.4. Chlorogenic acids analysis 174

7. Chlorogenic Acids Content in Green Coffee Beans 182 7.1. Normal commercial coffee beans 182 7.2. Green beans from immature fruit 188 7.3. Discoloured green beans 188 7.4. Stored green beans 189

8. Chlorogenic Acids Content of Roasted Beans and Soluble Powders . 189 8.1. Relative loss per gram dry matter loss 189 8.2. Relative loss per unit time 190 8.3. Loss in absolute terms 191 8.4. Fate of the chlorogenic acids 191

x CONTENTS

9. Organoleptic Properties 195 9.1. Model system studies and structure~activity relationships 195 9.2. Relevance to the acceptability of coffee products and coffee

beverage 196 9.3. Chlorogenic acids as predictors and determinants ofbeverage

quality 197 References 197

Chapter 6 Lipids P. FOLSTAR 1. Introduction 203 2. Coffee Oii 205

2.1. Determination of total oii content 205 2.2. Isolation of coffee oii for detailed analysis 206 2.3. Free and total fatty acids 207 2.4. Triglycerides 210 2.5. Diterpenes 211 2.6. Sterols 212 2.7. Tocopherols 215 2.8. Other compounds 215

3. Coffee Wax 217 3.1. Determination of the content ofC-5-HT in green and roasted

coffee 220 References 220

Chapter 7 Volatile Components 1. Preamble . 2. Methodology

2.1. Introduction 2.2. Headspace methods 2.3. Distillation techniques 2.4. Other techniques

S. K. DART and H. E. NURSTEN 223 224 224 224 231 234

2.5. Summary . 3. The Nature of the Volatile Components of Coffee

3.1. Introduction 3.2. Green coffee 3.3. The roasting process 3.4. Roasted coffee . 3.5. Effect of species on coffee aroma composition 3.6. Coffee processing and its effect on volatile composition 3.7. Quantitative assessment of coffee volatiles 3.8. Summary . References

Chapter 8 Carboxylic Acids J. S. WOODMAN 1. The Role of Acids in Infusions

1.1. The importance of acidity to taste and fiavour 1.2. Relationship between pH and acid content

2. The Acid Content of Green Coffee .

235 236 236 236 239 246 251 254 259 261 262

266 267 268 271

CONTENTS Xl

3. The Acid Content of Roasted Coffee 271 3.1. Identified acids . 271 3.2. Quantitative data 272 3.3. Changes on roasting 277 3.4. Changes on storage 278 3.5. Re!ationships to perceived acidity 278

4. The Acid Content of Oried Coffee Extracts (Instant Coffees) 279 5. Determination of Acids 281

5.1. pH and titratable acidity . 281 5.2. Individual acids 282

6. The Origins of Acids Found in Coffee Infusions 286 References 287

Index 291

List of Contributors

R. J. CLARKE

Ashby Cottagc, Donnington, Chichcster, Wcst Susscx P020 7PW, UK

M. N. CLIFFORD

Dcpartment of Biochcmistry, Unircrsity 01 Surrcy, Guildford, Surrcy GU25XH, UK

S. K. DART

Dcpartmcnt of Food Scicncc, Unl'crsity 01 Rcading, Whitcknights, PO Box 226, Rcading RG6 2AP, U K

P. FOLSTAR

Laboratory 01 Food Chcmistry, Agricultural Unl'crsity, Sahcrdaplcin 10, PB 9101, 6700 HB Wageningcn, The Ncthcrlands. Present address: H. 1. Hein::. BV, Postbus 6, 6660 AA Elst (Gld), Thc Nctherlands

R. MACRAE

Dcpartmcnt 01 Food Scicncc, Unt'crsity 01 Rcading, Whitcknights, PO Box 226, Rcading RG6 2AP, UK

H. E. NURSTEN

Dcpartmcnt 01 Food Scicncc, Unil'crsity 01 Rcading, Whiteknights, PO Box 226, Reading RG6 2AP, UK

Xll

LIST OF CONTRIBUTORS xiii

A. W. SMITH

Nestle Co. Ltd, St George's House, Croydon, Surrey CR9 I NR, UK

L. C. TRUGO

Universidade Federal do Rio de Janeiro, Centro de Ciencias da Saude, Instituto de Nutric;ao, Rio de Janeiro, Brazii

J. S. WOODMAN

Department of Hotel and Catering Studies, SheJficld City Polytechnic, Pond Street, SheJfield SI I WB, UK

Abbreviations

The following abbreviations have been used throughout the book:

as IS

ASIC

db ECD FID FPD HMF HPLC IR MS mg % NMR NPD ppb ppm RH SCOT UV WCOT

Composition ba sed on total weight of sample (i.e. no correction for water content) Association Scientifique Internationale du Cafe (aII references to ASIC Colloquia use the date of the meeting, nof the date of publication of the proceedings). For address see p. 82. Dry basis (i.e. corrected for water content) Electron capture detector Flame ionisation detector Flame photometric detector 5-Hydroxymethylfurfural High performance liquid chromatography Infrared Mass spectrometer/spectrometry mg per IOOg Nuclear magnetic resonance Nitrogen/phosphorus detector Parts per billion (i.e. Ilg/kg) Parts per million (i.e .. mg/kg) Relative humidity Support-coated open tubular Ultraviolet Wall-coated open tubular

XIV

Chapter 1

Introduction A. W. SMITH

Nest/e Co. Ltd, Croydon, Surrey, UK

1. ORIGINS

Nobody, today, would dispute that cotfee is big business; in fact, after oiI it is reckoned to be the most widely traded commodity in the world and provides employment for some twenty million people. Its gradual spread across the world, both as a crop and as a beverage, has been remarkable. The cotfee tree is indigenous to Ethiopia but the early history of its cultivation and the use of cotfee as a beverage, as we know it, is centred on Arabia. It is likely that, long before their use as the basis of a drink, the cotfee fruit and beans were chewed and found to be stimulating; later the people of Ethiopia were discovered using crushed dried cotfee beans mixed with fat and rolled into balls, as food to sustain them on their joumeys. In these early days the juice of the fruit may also well have been fermented and used as a beverage. A wealth of legends and anecdotes exists about the discovery of cotfee and its early use by man.1.2

The date at which cotfee was introduced into Arabia from Ethiopia has been given as some time in the fifteenth century.3 In the early l500s cotfee was being cultivated in the Yemen and the practice of infusing the ground roasted beans was well established in many parts of the Islamic world. Mohammedan religious leaders prohibited its consumption, believing it to be an intoxicant, and, later, condemned the popularity of cotfee houses which was atfecting attendance at mosques.

The drink was introduced to Europe by the Turks around the year 1600 and soon became popular in many countries. It was reported to be on sale

1

2 COFFEE: VOLUME l-CHEMISTRY

in Rome in 1625 and the first English coffee house was opened in Oxford in 1650. By 1675 there were nearly three thousand coffee houses in England, and King Charles II, denouncing them as seditious meeting places, issued a proclamat ion rescinding their licences. However, this created such opposition that it was hurriedly withdrawn. By the middle of the next century the habit of coffee drinking was weU established throughout Europe and North America.

The history of the spread of coffee cultivation is no less spectacular. InitiaUy the Arabians, as sole providers of coffee to the world, were highly secretive as to the origin of their coffee beans and were said not only to have prohibited foreigners from visiting their plantations but also to have insisted that aU exported coffee was steeped in boiling water to prevent germination. 2 Tradition has it that the first plants were smuggled out ofthe country by a Moslem pilgrim from India in the year 1600, but it was the enterprising Dutch who began the large-scale cultivation of coffee in Sri Lanka (then Ceylon) in 1658 from these Indian plants. 2 Further planting in Java followed in 1696, resulting in the first smaU commercial shipment of Java coffee reaching Amsterdam in 1711. 3

Meanwhile the Amsterdam Botanical Gardens had managed to produce seeds and seedlings from these plants, later classified as Coffea arabica var. arabica (otherwise var. typica), and one of them, presented to King Louis XIV of France in 1714, became the progenitor of most of the billions of trees now growing in South and Central America, the Caribbean, and ultimately many other coffee-growing countries. Quite separately, the French had established a plantation on the Indian Ocean island of Reunion, then known as Bourbon, in the early eighteenth century, from seedlings obtained direct from Arabia. This coffee, which was also destined to spread across the tropics, turned out to be a different variety now known as Coffea arabica var. bourbon. 3 This and the arabica (typica) variety remain the most important strains of arabica coffee in the world today.

The other commercially important species, Coifea canephora (always known in the trade as robusta), is one of several other species native to Africa, from where much of it is now obtained. It was discovered growing wild in U ganda, in the Congo Basin and ne ar the coast of West Africa only during the last hundred years, and now accounts for some 20 % of world exports.

The origin of the name 'coffee' perhaps deserves a mention. Although Kaffa is the name of an Ethiopian province where coffee can still be found growing wild, the root name is believed to be the Arabic word 'qahwah' (applicable to wine and other beverages besides coffee), reaching Europe as

INTRODUCTION 3

the Turkish word 'kahweh'. The Kiswahili word used in East Africa, and again derived from the Arabic, is 'kahawa'.

2. THE COFFEE PLANT

As has already been indicated, the two important coffee species of commerce are CojJea arabica and C. canephora.

The coffee tree belongs to the Rubiaceae family. Rothfos4 quotes the subdivision of the CojJea genus by Chevalier into four groups, of which EucojJea is of interest here. Of the tive subdivisions of EucojJea, the tirst (ErythrocojJea) contains both the arabica and the canephora species.

The shrub is a perennial evergreen dicotyledon which can reach a height of 10 m in the wild state, but plantation coffee is pruned to a maximum of about 3 m to facilitate harvesting and to maintain optimum tree shape. The primary branches are opposed, horizontally or drooping, and the leaves grow in pairs on short stalks. They are about 15 cm in length in C. arabica and longer in C. canephora, oval or lanceolate, and shiny dark green in appearance.

The tirst fiowers are produced at an age of 3 to 4 years, creamy white and sweetly scented, appearing in clusters in the axils of the leaves. The corolla is about 20 mm in length, the upper part dividing usually into tive petals. For fertilisation, C. canephora depends on cross-pollination; self-pollination usually occurs in C. arabica. After the fiowers fade, the ovaries slowly develop into oval drupes up to 18mm in length and 10-15mm in diameter, at tirst green, ripening to a bright red (referred to as 'cherries'), at which stage they are ready for harvesting. It is common to tind blossoms, green fruit and red cherries fiourishing on the same branch, especially in regions where there is an even annual rainfall distribution.

The coffee 'beans' are of course the seeds, of which two are normally found in each fruit, lying with their fiat sides facing (Fig. 1). Each bean is covered with a thin closely titting tegument known as the silverskin, outside of which is a looser, yellowish skin called the parchment, the whole being encased in a mucilaginous pulp which forms the fiesh of the 'cherry'. Should one of the two seeds faii to deveIop, the other becomes round in cross-section and is known as a 'peaberry'; other abnormal shapes are occasionally encountered.

2.1. Species and Varieties As already mentioned, coffee beans are the seeds of an evergreen shrub belonging to the family Rubiaceae and the genus CojJea. Two species are of


Outer skin - ____ ., Bean

Pulp Silverskin

--If.+-- Parchment

Fig. 1. Sectian af caffee cherry.

importance commercially, Coffea arabica Linn. and Coffea canephora Pierre ex Froehner. These are known in the trade, respectively, as arabica and robusta, and will be referred to by these names in the remainder ofthis chapter. Two other species, Coffea liberica Bull ex Hiern (known as 'liberica'), and Coi/ea dewevrei De Wild and Durand var. excelsa Chevalier (known as 'excelsa'), may be encountered, but their commercial importance nowadays is not significant.

Comprehensive descriptions of the varieties of coffee have been given by Haarer. 3 Coi/ea arabica, which accounts for some 80 % of world production, and which originated in the high mountains in the south of Ethiopia where it may stiH be found growing wild, exists in a number of forms, most of which have been purpose-bred or have originated from mutants in cultivated fields. However, the two 'original' varieties are usually acknowledged to be C. arabica var. arabica (syn. var. typica) and C. arabica var. bourbon. The former is believed to be the variety which fiourished in the Amsterdam Botanical Gardens, was taken to the Caribbean by the French, and is now grown widely in South and Central America, also in East Africa where it is known as 'Nyasa' coffee.

C. arabica var. bour bon as mentioned earlier takes its name from the island of Bourbon or Reunion where it was cultivated by the French. It is probably a mutant form of the typica variety and is said to give higher yields and, according to some reports, better liquoring qualities. Like typica it is also widely planted in the Americas and East Africa.

C. arabica var. maragogype is occasionally encountered. This arose as a mutation in a Brazilian plantation in 1870 and could be described as a gigantic form of C. arabica, with larger leaves, fruit and seeds, but is no longer popular due to uncert,.in yields and poor liquoring qualities.

INTRODUCTION 5

C. arabica var. amarella is a yellow fruiting variety, again not widely encountered.

Rothfos4 mentions a large number of other varieties/cultivars of C. arabica amongst which are the following:

Caturra, a bourbon variety, previously popular especially in BraziI for its high yield;

Mundo Navo, a hybrid between the bourbon and sumatra varieties, with good yield and disease resistance;

Catuai, often encountered in South and Central America, known for its rapid and high yield;

Kent, said to have originated from South India and widely grown in East Africa for its yield and resistance to leaf rust;

Blue Mountain, a Jamaican variety, again grown in East Africa, highly resistant to coffee berry disease and able to thrive at high altitudes.

The other important species, Coffea canephora, includes the various forms known in the trade as robusta coffee. The latter name is derived from its former specific name C. robusta, but nowadays C. robusta is regarded as synonymous with the original variety of C. canephora.

Three other varieties of robusta are worth noting. One is C. canephara var. kouilouensis, discovered in French Equatorial Africa in the 1880s and later planted in the Dutch East Indies. This variety may also be found today in West Africa, Madagascar and BraziI, where it is known as Conillon. Another is C. canephara var. nganda, similar to the above in many respects, but with a shrub-like growing habit. It is found in Uganda and other parts of Equatorial Africa. Rothfos also lists niaouli as an important variety.4

Some hybrids ofthe arabica and canephora species have been developed. The best known is probably the arabusta, developed in the Ivory Coast in an attempt to combine good cup characteristics with high disease resistance. It is not yet to be found in significant quantities, however, on the world market.

A few differences between the two main commercial species should perhaps be recorded. Robusta coffee will grow at relatively low altitudes, will tolerate higher temperatures and heavier rainfall, and demands a higher soil humus content than arabica. Generally it is also much more resistant to disease. Whilst the arabica bean is green to paIe green in colour and oval in shape, the robusta tends to be rounder and may be brownish rather than green. The arabica species produces those coffees most appreciated by the discerning coffee drinker, and is further subdivided in the trade according to whethe.r the coffee has been processed at origm by


the 'wet' or the 'dry' method, both of which will be described later. The better quality coffees, and usually the most expensive, are considered to be arabicas prepared by the 'wet' process and known as 'milds'; the best will have an 'acidy' cup character, a fine, perhaps aromatic flavour, and a good fuU body. Arabica beans prepared by the 'dry' process are mainly represented by BraziI, the world's largest producer. Whilst very useful blending coffees, Brazils tend to lack the fine flavour characteristics of the milds.

Most robusta coffees today are prepared by the 'dry' process. Although they can assist in providing a blend with a full-bodied base, they do not usually contribute much in the way of fine coffee flavour.

3. PRODUCING COUNTRIES

The areas of the world in which coffee can be grown commercially are limited, primarily, by temperature, since the plant is easily damaged by frost. Thus, latitudes north and south, respectively, of the Tropics of Cancer and Capricorn may, generally speaking, be dismissed as unsuitable. Within the Tropics, altitude then becomes a factor, since the nearer the Equator, the higher the altitude at which there is a risk offrost, so the plant may be found growing at 2500 m on the Equator, but not above about 100 m at latitudes of 25 N and 25 os. There is also, in respect of temperature, an upper limiting factor, since the plant will not tolerate temperatures approaching 30 ac, especially in conditions of low humidity.

Rainfall also has to be taken into account. Ideally for coffee, annual rainfall should not be less than about 150 cm, and in some areas producing excellent coffee it may exceed 250 cm.

The type of soil suitable for coffee growing is generally friable and loamy, often of lateritic or volcanic origin, and very deep (Rothfos quotes 3 m mmimum). The robusta species additionally requires a soil rich in humus.

The locations of today's principal coffee-growing and exporting countries are shown in Fig. 2 and these are also listed in Table 1, which will give some idea of their relative importance in the world markets. Coffee is also grown in a few other countries which consume their own production and will not therefore be usually encountered commercially.

A great deal of information on the coffee business in individual producing countries is to be found elsewhere 5 - 7 and only brief details are given here.

Trop

lCol

..

. tin

""

'H

aw

aII

0-

T,o

p.co

' Ca

-pr

lcoJn

Ma_

le;l -<

CARBOHYDRA TES 97

extracts once water-extracting temperatures are rai sed above 130 cC (as for green cotTee shown by the work of Pictet and Moreau, 11 already described) and generally so up to 180 ce. Again robustas at the same roast level and conditions of extraction will produce somewhat higher yields of soluble solids, includ ing carbohydrates.

The saccharide composition of cotTee extracts, comparable with commercial extracts of the instant cotTee industry, has been examined by Thaler 24 ,25 and by Kroplien. 5 ,20 In addition, various actual commercial instant cotTees, with the additional step of drying, have been examined by various workers, though the information as to their blend of cotTee, roast level and extraction yield taken is not always available.

3.1. Low Molecular Weight Sugars Commercial samples of instant cotTee generally contain small quantities of arabinose, galactose and mannose, together with traces of other sugars such as sucrose, ribose and xylose. Glucose and fructose are also sometimes found, though the presence of the latter may well be due to the addition of instant chicory. The main data which have been published are compiled in Table 8, with results obtained using modern analytical techniques. The wide range of values observed for ali the sugars is a refiection of both ditTerences in samples studied and inherent ditTerences in the analytical methods. The TLC method 20 involves removal ofthe components from the plates and subsequent colorimetry, whilst the GLC method 26 involves derivatisation; both of these procedures may lead to losses. In theory the HPLC methods 2 27 should be more straightforward and hence less subject to error and the ditTerences here are primarily due to changes in the extraction procedures. The preliminary data published in the ASIC Proceedings 2 involved a warm water extraction (50 CC) and subsequent lead acetate precipitation; when this method was further scrutinised and the precipitation step avoided by the use of Sep-Pak cartridges, significantly higher values were obtained. On this basis, the higher values are likely to be more accurate and further work is in progress. The 7 samples in the preliminary study were also included in the subsequent work but the range of samples was extended, and this also contributed to the higher values.

These figures for amounts of arabinose, galactose and mannose are relatively high compared with the negligible levels in roast cotTee. In terms of glucose and fructose, Kroplien 20 showed that the instant cotTees from Germany had substantially lower contents of each as compared with instant cotTees of lighter roast. To allow a meaningful comparison of these


instant coffee values with those for roasted coffee it is necessary to multiply the former figures by the fractional extraction yield of total soluble solids. Kroplien 20 provides a calculation based on a determination of the amounts of polysaccharides present from Thaler's data in a medium roast arabica coffee (see Tables 6 and 7), and an assumed average yield from roasted coffee of 50 %, which may not be correct. This suggests that, of the original 'araban', though present in small quantity, some 61 % is hydrolysed to arabinose during extraction but only 3 % of the 'galactan', 12 % of the mannan and O 3 % of the 'glucan' to the corresponding monosaccharides.

If, however, in fact arabinose units are only part of the actual polysaccharide present, e.g. as a side-chain in arabinogalactan (see p. 86), these araban hydrolysis figures suggest a stripping of arabinose units during processing, leaving the galactan part relatively untouched, as also mannan and cellulose.

Kroplien in the same study followed up experimentally the formation of monosaccharides in instant coffees by succes si ve batch extractions of roast and ground coffees at increasing water temperatures from 120 to 200 ec in 20 ec steps using a laboratory autoclave for 1 h at each temperature. He demonstrated a gradual increase in the amounts of arabinose, galactose and mannose in each of the extracts at increasing extraction yields for each ofColombian arabica, Salvador arabica and a Congo (ZaYre) robusta. For example, the cumulative amounts after pressure water extraction at 160 ec of the Salvador arabica, at a soluble solids yield of 46 % on roasted coffee, of arabinose, galactose and mannose based on dry weight of extract were 1,28,0,38 and O' 31 % respectively, whilst for the robustacoffee for the same temperature, but at a yield of 60 %, the corresponding figures were 1,50, 025 and 017 % respectively. Though this type of processing only very crudely simulates large-scale coffee extraction in counter-current batteries, the values are somewhat similar to those reported for commercial instant coffees. At an extraction temperature of 120 ec, arabinose was still present in the order of 02 %, but no galactose or mannosewas found. These results were obtained on liquid coffee extracts which would then have been subsequent1y spray- or freeze-dried in commercial practice. No infor-mat ion is available on any changes that may subsequently take place on drying; some diminution by Maillard reaction might be expected if the spray-drying conditions are poor (i.e. long contact time at high temperatures), as can be found sometimes in liquid milk drying.

3.2. Polysaccharides The study of individual carbohydrate fractions may also be achieved by fractionation of instant coffee by extraction with different solvents and

CARBOHYDRA TES 99

water. According to Pictet,28 two distinct fractions may be obtained by extraction with aqueous acetone (90 % v/v), one soluble in water (A), representing 29-37 % ofthe total solids, and another insoluble in water (B), representing only 27-75 %. A third fraction (C), which accounts for some 32-41 % of total solids, may then be extracted with ethanol (70 % v/v) and the fourth fraction is the residue ofthe organic extraction, soluble in water, and representing 22-31 % of total solids. Fraction A contains only small amounts of sugars, being mainly composed of caffeine, chlorogenic acids, proteins and minerals. Fraction B contains mainly polymers of galactose, arabinose and glucose. Oligosaccharides consisting of galactose, mannose and arabinose are the major components of fraction C. The water-soluble residue (D) shows the highest amount of polysaccharides which are formed of galactose, mannose and arabinose, with a degree of polymerisation in the range 15-25. Some minerals, proteins and traces of chlorogenic acids are also present in this fraction.

The polysaccharides of coffee brews and extracts prepared by hot water extraction at temperatures of not more than 100 ac can be assessed from the data of Thaler for his hot/cold water fractions as shown in Table 6. For a medium roast arabica, with exhaustive extraction, we could expect to find some 35 % polysaccharide with respect to green coffee weight (somewhat higher on roasted coffee weight); this polysaccharide is based primarily on arabinose and galactose, and some mannose.

The main work on the polysaccharides of higher yield coffee extracts and instant coffee has again been carried out by Thaler and his colleagues, and has been reported in two main papers already cited. 23.25 In the second (in English) they reported the polysaccharide composition, in the same araban-galactan-mannan-cellulose terms as before, of higher yield extracts obtained by 'technical pilot plan!' extraction of roasted coffees: one a Colombian arabica roasted to a medium degree (roasting loss of 170 % as is, i.e. includ ing unstated moisture content) and the other an Angolan robusta coffee (roasting loss 187 % as is). The extraction conditions are not given, but included samples extracted at 100 ac with stated extraction yields based on roasted coffee weight of 380 % (arabica) and 395 % (robusta), which seem to be rather high values at this temperature.

Three other samples for each coffee were available, giving the highest yields of 53 % (arabica) and 58 % (robusta). The extracts were first hydrolysed with sulphuric acid, and the monosaccharides quantified by paper chromatography and then eventually by enzymatic analysis for the expected sugars mannose and galactose, in the largest amounts, and glucose and arabinose.


Samples were also included, prepared in the laboratory using boiling water in a 1 :25 ratio in a Melitta filter, when the high extraction yields of 364 % (arabica) and 388 % (robust a) were stated as being obtained, though apparently23 defatted roasted coffee was used. The arabica 'home' brew was reported to have a content of 745 %db carbohydrate (which presumably includes simple sugars already existing) with respect to roasted coffee weight, though only approximately 48 % is reported in the hot/cold water fraction of roasted Colombian coffee (even at the darker roast-see Table 6) in the previous study. 15 It is apparent that this yield figure must be incorrect, or that another explanation is required.

The results found were first presented graphically by plotting the determined carbohydrate contents (as total monosaccharides) in the extracts as percentages of dry roasted coffee weight, against the cited corresponding extraction yield of soluble solids (Fig. 1). It was evident that the arabica coffee delivered increasing amounts of carbohydrate (i.e. some 18 %) up to a yield of 436 %, which then remained constant up to a yield of 53 %. On the other hand, with robusta coffee, polysaccharides are released continuously even at the highest yield (of 58 %) sample examined. The somewhat anomalous nature of the laboratory extract is again apparent from the graphs, since if these results were omitted a reasonably straight line would be produced (for robusta).

A ;! 2 O lf} w >--< c:: a >-I 10 o ID c:: -< u

O 35 45 55

EXTRACTION YIELD (./.)

Fig.1. Extraction yields of roasted coffee: (A) robusta; (8) arabica. * Carbohydrates (%) on a roasted coffee basis. (Reproduced from Thaler. 26 )

CARBOHYDRA TES 101

With respect to the individual polysaccharides, for arabica coffee, somewhat more 'galactan' than 'mannan' was found in the extract (calculated as a direct percentage) up to a yield of 40 %; from then on more 'mannan' is present, until at 53 % yield these are approximately similar. For robusta coffee the results are similar, but up to 58 % there is always more 'galactan' than 'mannan'. In each coffee the quantities of'glucan' present are always very low, whilst the 'araban' also remains at a low level, though increasing slightly with increasing yield, e.g. to 14 % in the arabica coffee extract. These 'araban' values wilI, however, include arabinose already in the monosaccharide state, which was not separately reported. A further anomalous feature is that the brew extracted in the laboratory at 100 0 e contained a relatively high level of 'araban', whereas that prepared in the pilot plant at a similar yield, also at 100 aC, contained a negligible quantity.

Thaler clearly considered that part (if not alI) of the polysaccharides themselves extracted were also bound in some way to Maillard products, particularly those polysaccharides of 'high' molecular weight. In this study he therefore used chlorine dioxide to treat the coffee extracts, which were only then dialysed (membrane cut-off 5000 daltons) and nine volumes of ethanol added, in order to precipitate 'high polymer' polysaccharides which were washed and dried, and weighed prior to hydrolysis for quantitative estimation. Such precipitates apparentIy still contained the converted polysaccharides (with no monosaccharide content). These polysaccharides, or high polymer carbohydrates, were found present in arabica coffee extracts and generally made up more than half the total polysaccharide content; whilst with robusta extracts, at low yields, content was about a half but at higher yields the proportion of the 'low' polymers increased rapidly. The composition of the two polymeric types was, however, similar in each coffee.

Further separation of polysaccharides in this study was achieved, following chlorine dioxide treatment, by quantitative estimation of precipitates after addition of Fehling's solution and treatment with methanol-acetic acid. These precipitates were found to be primarily mannan (though with 5 % galactan in the case of arabica coffee, and 12 % with robusta) without the presence of 'araban'. Their quantitative interest lies in the observation that whilst arabica extracts showed a constant release of this polysaccharide, robusta extracts showed increasing release with higher yields, and in amounts close to that of the 'high' polymers. SubsequentIy Ara and Thaler 29 noted that this mannan increased in quantity with roasting degree.

In the absence of certainty about the exact structures of the


polysaccharides actually present, interpretation of these results may be misleading, though re-examination in terms of arabinogalactan--(galacto)-mannan-glucan of Wolfrom and Anderson could be revealing. They also reported 30 isolation of actual polysaccharides from an instant coffee, of unknown source, when they showed their procedure to obtain an arabinogalactan, but only at about 5 % yield. The ratio of the saccharides was found to be 2: 25 (with a degree of polymerisation of 15), which is very different from that found in green coffee (2: 5 ratio), presumably due to the removal of arabinose units from the chain during the instant coffee processing as already discussed. This removal could well be due to the slightly acid conditions prevailing during extraction. Mannan was also isolated but at even smaller yield (about 1 %) compared with the expected 10-15 %, with negligible galactan associated with it, with a possible degree of polymerisation at 10-13. Thaler makes specific references to his belief that hydrolysis is not occurring during extraction, though clearly arabinose in particular is being released.

3.3. Carbohydrate Conversion Products During extraction, not only will polysaccharides be extracted, whether free Of complexed, but also the carbohydrate conversion products (colouredjpigmented or uncoloured) already discussed in Section 2.3. The aqueous extraction ofthese substances at temperatures of 100 De and above has been little studied, though no doubt it is also progressive in amount with increasing extraction yield. The amount could be deduced from the difference between the extraction yield and the carbohydrate yield, both taken with respect to roasted coffee weight, together with information about 'protein' that is extracted, since the other solubilised substances will be totally extractable with water at 100 De.

Aqueous extracts and brews taken at 100 De extracting temperatures clearly contain such substances, especially the brown pigmentedjcoloured material. Sucrose caramelised products are expected to be readily soluble.

Gel-filtration techniques providing a separation to varying degrees by molecular size (and therefore indirectly a molecular weight profile) are useful tools in helping to unravel the nature of polymeric soluble substances in coffee extracts.

Streuli31 in 1962 separated three fractions of a roasted coffee extract using Sephadex G-25, together with ultraviolet absorption as a monitor. The first fraction consisted of high molecular weight material (in excess of 4000) and was strongly coloured; the second fraction contained, together with 'melanoidin' substances, trigonelline and caffeine; whilst the third, located

CARBOHYDRA TES 103

in the low molecular weight region, contained relatively high levels of 'chlorogenic acids'.

Feldman et al. 32 in 1969 reported results of chromatographic separations using Sephadex G-25, similarly to Streuli, but using a refractive index monitor, when again the first fraction or high molecular weight region showed a range of material, including the brown compounds, though without much resolution. A typical chromatogram is shown in Fig. 2.

By the use of the same technique on Sephadex G-50 but with a preliminary separation of the extract into two fractions by means of column chromatography with nylon powder, Maier et al. 33 separated a Colombian coffee extract further into seven fractions, and similarly a Santos extract 34 prepared from the defatted roast coffees (roast loss 19 5 %) using al: 9 coffeejwater ratio at 90 ec to give a 298 % extraction yield of solubles (fat-free basis ?). In the fractions free of phenols and low molecular weight substances, the presence of a galactomannan (MW 5000-10000) together with brown compounds (from which the two could be physically separated) and polymerised material (MW 5000-50000) was shown. The latter material on hydrolysis was composed of mannose, galactose and arabinose, and approximately 6-12 amino acid residues per molecule. A peptide was also isolated, but the nature of the chromogenic

2 3 w If) z o a.. If) w. c:: ~ ci

30 60 90 TUBE NUMBER

Fig. 2. Chromatographic separation of roasted Santos coffee water extract on Sephadex G-25 column. Fraction 1, high molecular weight region; fraction 2, diffusible substances (trigonelline and caffeine); fraction 3, adsorbed sub-stances (chlorogenic acids). (Reproduced from Feldman et al. 32 by permission

of the American Chemical Society.)


group could not be discovered, nor was the chromatographic positioning of converted carbohydrate (eported.

One ofthe three fractions commonly obtained by gel-filtration of roasted cotTee extracts with Sephadex G-25 has been demonstrated to be mainly formed by the caramelisation of sucrose, similar to that obtained with model systems. 35

In a series of papers 36 - 38 from the University of Rostock, Mucke and his colleagues claimed the quantitative estimation and isolation of coffee humic acids. Fractionation by Sephadex G-25 of a coffee extract, followed by oxidation with permanganate, showed 16 3 % of this ill-defined substance in a roast coffee and 145 % in an instant coffee. Four properties for the substance were given: formation of insoluble lead salt from which an ammonium salt could be isolated, capacity for chelation with ferrous and ferric ions, with reduction of the latter, and release of phenols on alkaline hydrolysis. The humic acid from the instant coffee contained some 3'5-4'2 % nitrogen, ofwhich about one-third was in amino acid form after hydrolysis. It was also found associated with a number of breakdown products of chlorogenic acid, in particular caffeic acid, but also some ten other phenolic substances, which could be released by alkaline pressure hydrolysis.

Modern column-packing materials have been developed recently for gel-filtration, which are mechanically stable and therefore can be used as small particles under high pressure, as discussed in Section 5. Results can be obtained in a shorter time, and in most cases provide considerable improvement in resolution. TSK PW 4000 columns were used by Trugo 2 7 to study changes in the molecular weight profile of aqueous extracts of cotTee (SOC, 1: 10 coffee/water ratio) as a function of the degree of roasting. When a refractive index detector was used to monitor the column eluent, the chromatograms obtained, as expected, showed major ditTerences between the green and roasted coffee,samples (Fig. 3) both in the high molecular weight region (> 25 x 10 5 ) and in the low region (~103). There was an obvious increase in amount in the former for the roasted cotTee, but this amount declined with the severity of roasting. In the intermediate region a more complex pattern appeared in the roasted coffee samples. By use of a visible detector at 420 nm on the same eluates (Fig. 4), pigmented material can be identified. Whilst little pigmented material appears in the green coffee, as expected, it seems to be distributed across the whole high molecular weight range for the roasted coffee. When detection at 325 nm is used, relatively simple patterns are observed, and in contrast to detection by refractive index or at 420 nm, no peaks are detected above a

CARBOHYDRA TES 105

a

b

d

.

O 20

25 7 0:1 (x 10 4 ) molecular weight

60

Fig.3. Gel-filtration chromatography of green and roasted robusta coffee water extract on TSK PW4000 column, with refractive index detection. Chro-matograms of (a) very dark roasted, (b) dark roasted, (c) medium roasted, (d) light roasted and (e) green coffee.

(Reproduced from Trugo. 27 )

a

b

d

i O

o

20 m,n

25 7 0:1 (x 10 4 ) molecular weight

o

60

Fig.4. Gel-filtration chromatography of green and roasted robusta coffee water extract on TSK PW 4000 column, with detection at 420 nm. Chro-matograms of (a) very dark roasted, (b) dark roasted, (c) medium roasted, (d) light roasted and (e) green coffee.

(Reproduced from Trugo.27 )

molecular weight of 25 x 10 5 , which appears only after roasting. This would suggest that phenolic compounds (chlorogenic acids) participate in the formation of high polymers during the roasting process. The high molecular weight material (i.e. above 25 x 10 5) was seen to increase with the degree of roasting, but decreased again with severe roasting. There was some evidence that the high molecular weight material is more thermally stable in robusta than arabica. Most of the peaks which appear with 325 nm detection are also detected at 280 nm with similar shapes and distribution, though generally smaller. Caffeine, which absorbs at 280 nm, is strongly retained on the column, being eluted only after 60 min, showing that absorption as weB as gel-filtration mechanisms are operative. The appearance of high molecular weight material detected at 280 nm in the water extract of roasted coffee may be a further indication that components which absorb at this wavelength (which includes proteins and chlorogenic acids) take part in polymerisation reactions with the soluble converted or


unchanged polysaccharides, or loosely linked or physically attached to them.

Gel-filtration has also been used for the study of molecular weight profiles of instant cotTees. 2.2 7 Samples used in this study show approximately the same amount of high molecular weight material (i.e. about 25 x 105) as water extracts of roasted cotTee. The main ditTerence, apart from those associated with ditTerences in amount of such substances as chlorogenic acids and catTeine, is in the small but variable amount of monosaccharides, such as arabinose, mannose and galactose. These amounts can be precisely quantified by related HPLC techniques, as already discussed. The amount of high molecular weight polysaccharide-type material has not yet been quantified by these methods, since the reference substances are not generally available except for dextrans of specific molecular weight ranges. The very high molecular weights that ha ve been discussed in these studies are much gre ater than those likely for the pure polysaccharides that have been isolated from green cotTee, and are still present in roasted cotTee. The molecular weights of polysaccharides, whether free or combined within green cotTee, are not known, nor of those polysaccharides which can be extracted into aqueous solution by the use of water-extracting temperatures up to say 180 aC. Further information as to the nature of the bonding between these polymeric materials could be obtained by repeating the gel-filtration in dissociating solvents, such as urea, where aggregation due to hydrogen bonding would be removed.

4. SOME PHYSICAL PROPERTIES OF COFFEE CARBOHYDRATES

Carbohydrates, like many other substances, have the capacity for binding volatile compounds at adsorptive sites, as demonstrated by Maier 39.40 with particular reference to carbohydrates and other components of cotTee. The binding of known volatile cotTee aroma compounds by various cotTee-sorptive substrates is of interest. Sorption of volatiles can be assessed on a short-term basis by a gas-chromatographic method,41 but the equilibrium or maximum level has to be obtained by gravimetric methods using a closed container or desiccator. A measure of the amount of irreversibly held volatile component can be obtained after headspace evacuation of the previously equilibrated sample by application of a high vacuum at ambient Of other tempera ture.

Table 9 gives an indication of the kind of results obtained by Maier for

CARBOHYDRA TES

Table 9 Sorption of ethanol by coffee and other food products

(Data from Maier39 )

107

Produce Amount adsorbed in mg/g dry matter

Roasted coffee Coffee extract Cellulose Starch Pectin Casein Skimmed milk (powder) Full cream milk (powder) Ovalbumin Strawberry powder

Alter 1 h

16'7 31 92 10 23 52 4-4 62 26 52

At Desorption equilibrium at room

tempera ture

50 29 3 00

98 245 208 115 00 290 35-4 216 97 202 92 230 376

19 00

ethanol sorbed on to various solid food components including a roasted coffee and dry coffee extract. Data were also obtained for volatiles such as hexane and acetone. Coffee extract adsorbs significant1y lower amounts of volatile compounds than roasted coffee. However, roasted coffee contains significant amounts ( '" 15 %) of coffee oiI, in which the sorption is now bulk absorption by virtue of solubility. Coffee oiI itselfwas shown by Maier to give high sorptive amounts of these same volatiles. The level of coffee oiI in immediately dried instant coffees is low but variable. In practice, commercial instant coffees have subsequent surface applications of coffee oiI. Adsorption is, however, also highly dependent upon the condition of the powdered material, that is, its porosity and particle size, which is dependent upon the way in which it is prepared. This phenomenon is demonstrated by different water sorption isotherms for instant coffees dried by different methods, and in the high ethanol adsorption of specially prepared dry corn syrup solids hydrolysates.

Maier also determined the equilibrium sorption amounts and partition coefficients offive selected volatile compounds, pyridine, ethanol, diacetyl, ethyl acetate and acetone, for three spray-dried coffee extracts at high extraction yield (36 %, 42 % and 53 %) obtained from Thaler (p. 99). The increase in sorption amount (expres sed as nmoles per g substrate) with increase of extraction yield was evident with both pyridine and ethanol, the former volatile compound showing the higher level of sorption. Results


were also provided, by the gas-chromatographic method, for partition coefficients of these same volatiles, in respect of the content of total polysaccharides, and of the 'high polymer' polysaccharides separately. It was evident that the higher the quantity of the latter, the greater was the sorptive capacity for the volatile compounds. In addition, in this study, the sorption amounts of both pyridine and acetone were determined for selected components of coffee, i.e. mannan, 'Maillard produci' (from glycine and mannose), caffeic acid, quinic acid and caffeine. The two coffee acids, present to the extent of about 1-2 % in instant coffees, and the Maillard product were shown to have quite high binding power for pyridine, whilst mannan adsorbed both these volatile compounds to the same high level.

The significance of adsorption of volatile compounds is doubtful in the drying of coffee extracts, especially during freeze-drying. Rey and Bastien 42 considered that adsorption of volatile compounds was important, and that these compounds would be found in the outer layers of a freeze-dried particle or slab. Flink and Karel,43.44 however, demonstrated in model experiments with maltose solutions and propan-2-01, that after freeze-drying, the volatile component was retained in the are a ofthe layer in which it had originally been located, and not in the other layers. Again according to these workers, the nature of the matrix material enabling 'micro regions' for the volatile compounds is important in freeze-drying. It is necessary that the matrix material be kept amorphous and not capable of viscous fiow, which is dependent upon the tempera ture and its moisture content; 'collapse' temperatures were defined, above which temperature volatile compounds from any given layer in the slab or granule are liable to escape, maybe accompanied by 'bubbling' or 'foaming'. These collapse temperatures are noted to be quite low for simple monosaccharides, such as glucose and fructose, as contained in fruit juices, e.g. - 40 e at 25 % solids concentration, whilst polysaccharides of different kinds had much higher collapse temperature, e.g. - 20 e at 25 % solids concentration for coffee extract, and proportionately higher temperatures at lower water con-centrations in each case. The higher the content of polysaccharides, and probably of converted polysaccharides, the greater is the ease of freeze-drying for good volatile retention since the higher relatively is the temperature at which the frozen layers need to be maintained.4S

In respect of spray-drying but also of freeze-drying and the air-drying of slabs, Thijssen el al. 46 have developed the concept of selective diffusion, together with supportive technical experimental evidence in numerous papers. In simplistic terms, as the concentration of the outer layers

CARBOHYDRA TES 109

increases, so the volatile loss with respect to that of water evaporating is determined not by volatility factors, but by diffusion factors. At high solids concentrations, the diffusion of volatile substances is very much less than that ofwater, so that according to the rate at which evaporation ofwater is occurring, the retention of volatile compounds will be high in a fully dried product. These relative diffusion rates are also determined by molecular weight factors which in turn influence solution viscosity at different concentrations. The higher the molecular weight, the higher is the viscosity. The presence of high molecular weight substances in coffee extracts, especially ofhigher yield, is therefore favourable to the retention of volatile substances. Polysaccharides, at least in their pure form, are known to have high viscosities at quite low concentrations as, for example, solutions of various polysaccharide gums. Information on the viscosity of coffee extracts at various concentrations is available (Chapter 2), though little on the particular individual polysaccharides in coffee extracts or their complexed derivatives. The beneficial effect in retention of volatile compounds in the freeze-drying of substrates based on homologous polymeric compounds of increasing molecular weight has been dem-onstrated by Voilley and Simatos. 4 7 Adsorption and absorption effects may play some role in the later stages of drying, which would be difficult to disentangle from simple diffusion rate changes. In subsequent storage such effects would, however, be important in the loss of volatile compounds, though in practice controlled by the use of sealed containers.

5. DETERMINATION OF CARBOHYDRATES

The determination of carbohydrates in complex materials such as coffee and its products presents considerable difficulties, and different attempts to overcome these problems may give rise to variable results.

Components have first to be obtained generally in aqueous solution, unless liquid coffee extracts are being examined. For accurate quantifi-cation complete recovery is essential and again this is often a difficult task. General-purpose methods for the determination of reduc ing sugars such as Fehling's re agent lack specificity, and the titration end-point is difficult to determine precisely when strongly coloured solutions are used. Ethanolic extracts, cleared by addition of lead acetate (to remove proteinaceous material, etc.), have been studied by paper chromatography followed by visualisation using aniline-oxalate, benzidine-trichloracetic acid or urea-hydrochloric acid reagents. 48 49 Similarly, paper chromatography

110 COFFEE: VOLUME I-CHEMISTRY

after sample purification with a mixture of FIorex and Ce li te column chromatography has also been applied. 4 Thin-Iayer chromatography (TLC) was a further development used by Kroplien 5 in his determinations of monosaccharides; similar techniques were also used by Thaler. 1 2 An extensive and time-consuming sample purification procedure prior to chromatography is nevertheless needed.

The procedure consisted essentially of column chromatography, using a combination of charcoal-polyamide, cation-exchange and anion-exchange columns. The sugars are then eluted from the successive columns and analysed by TLC followed by densitometry of the sugar derivatives formed with 4-aminobenzoic acid. More rapid enzyme oxidation methods have been used, e.g. various dehydrogenases specific to particular sugars by Thaler 25 and Pictet and Moreau. 11

In the use of gas-liquid chromatography (GLC) for the quantification of sugar content, the sugars have first to be derivatised by silylating agents. 26 The extracts may be obtained by aqueous ethanol extraction and then cleared by ion-exchange chromatography. More recently, high perfor-mance liquid chromatography methods (HPLC) have been described for use with coffee products. HPLC has the advantage of not requiring derivatisation. However, some limitations are encountered, in the use of this technique, particularly when extracts containing polymeric material (coloured or not) yet containing low free sugar content are examined. The most commonly employed column material for HPLC determinations of sugars is a polar bonded phase, such as Spherisorb Amino, with a mobile phase of aqueous acetonitrile. Under these conditions any proteinaceous material will bind to the column (it is also a weak anion-exchanger) and any polysaccharides may be precipitated in the mobile phase. Either of these occurrences will seriously reduce the efficiency of the column. The major limitation with HPLC for sugar determination is, however, the relatively poor sensitivity of the refractive index (RI) detector, which in turn means that concentrated coffee extracts must be prepared, which exacerbates the problems of interfering compounds. Ultraviolet detectors can be used at low wavelengths (e.g. 200 nm) but problems still exist with poor sensitivity and interferences.

However, with the use of the mass detector improved determinations have been established. The operation of the mass detector is based 50 upon the detection of solute molecules by light scattering, after nebulisation and evaporation of the chromatographic solvent (e.g. aqueous acetonitrile). It has two significant advantages over the RI detector: it is more sensitive and also allows gradient elution which is not otherwise feasible. Applications of

CARBOHYDRA TES 111

this detector have been described for sugar determinations 51 and for lipids S2 in different food products.

The use of gradient elution allows, for example, glucose and sucrose to be separated simultaneously with the more highly retained oligosaccharides such as raffinose and stachyose, and in a shorter time in comparison with isocratic elution (Fig. 5). Examples of the separation that can be achieved for sugar!'. (in standard mixtures and instant coffees) by HPLC using the mass detector are shown (Fig. 6). Due also to its higher sensitivity, the mass detector allows the use of more dilute sample extracts for the chromatography and therefore easier sample clean-up. Sample purifi-cation, particularly in sugar determinations, has been greatly improved with the advent of clean-up cartridges packed with chromatographic phases (e.g. silica or reversed-phase packing material). More efficient sample clean-up, with a considerable reduction in time of analysis, is then possible by passing the sample extract through Sep-Pak C 18 cartridges. 2 7

Gel-filtration chromatography, which is useful for studying molecular weight profiles, and also for isolating compounds, was first carried out using Sephadex (G-25 or G-50) or Biogel. More recently, mechanically stable column-packing materials have been commercially developed to be

2

a

2 c

3 4

2 b

min min

Fig. 5. Separation of oligosaccharides by HPLC. Column: Spherisorb-5-amine (150 x 5 mm i.d.). Solvent: acetonitrile and gradient of water at 15 ml/min. Detection conditions: mass detector set at attenuation x 1, photomultiplier x 2 and evaporation temperature of 90C. Chromatograms of (a) green arabica coffee, (b) green robusta coftee and (c) standard mixture.

1, Glucose; 2; sucrose; 3, raffinose; 4, stachyose. (From Trugo. 27 )


Fig. 6. Separation of monosaccharides by HPLC. Column: Spherisorb-5-amino (250 x 5 mm i.d.). Solvent: acetonitrile/water (84: 16 v/v) at 15 ml/min. Detection conditions: mass detector set at attenuation x 1 , photo-multiplier x2 and evaporation temperature of 70C. Chromatograms of (a) standard mixture and (b) instant coffee. 1, Ribose; 2, xylose; 3, arabinose; 4, fructose; 5,

glucose; 6, galactose. (From Trugo.27 )

a

b

i O

4

5 1 2

3

4

5

min

used as small particles under pressure, i.e. with HPLC. The TSK PW 4000 column is one example, as used in the previously described studies of green and roasted cotfee aqueous extracts, and instant cotfee solution. 2,2 7 Such extracts merely require filtration through a Millipore filter (e.g. 0-45 flm). The actual chromatography may be performed using one column (say 300 or 600 mm length x 8 mm internal diameter), though more in series can improve resolution. The resolution achieved is better than with Sephadex, and furthermore time of analysis is greatly reduced (30-60 min being normally sufficient). HPLC detectors may also be used without difficulty, since adequate solvent f1ows, free of air bubbles, are easily obtained.

In the quantitative determination of polysaccharides, similar procedures may be used provided they have first been hydrolysed to the constituent monosaccharides. The conventional reagent for this purpose is 72 % (v Iv) aqueous sulphuric acid solution. The polysaccharides present cannot of course be specifically identified. There is, however, no reported work on the use of modern GLC or HPLC techniques on these hydrolysates. In particular, it is envisaged that the application of high resolution gel-filtration could lead to the isolation of individual, or at least defined groups, of polysaccharides. These could then be characterised further in terms of their sugar composition, after hydrolysis, again using instrumental chromatographic techniques. This could also be extended by derivatisation to provide structural information as to how the sugar units are linked. Such

CARBOHYDRA TES 113

constituent sugar determinations can also be carried out on the whole green or roasted cotTee beans, and not merely on their solvent or aqueous extracts. The general procedure is tirst to grind these beans very tinely, and then to defat by Soxhlet extraction using such solvents as petroleum ether. It is usual to remove low molecular weight sugars, together with other constituents such as chlorogenic acids, by extraction with 70-80 % (v Iv) alcohol, prior to the acid hydrolysis.

Enzymatic methods may also tind application both for selective cleavage of polysaccharides and also for constituent sugar determinations. The detailed characterisation of the polysaccharides in coffee products clearly requires much further work.

ACKNOWLEDGEM ENT

The author would like to thank the Editors for assistance in the preparation of this chapter.

REFERENCES

1. Tressl, R., Holzer, M. and Kamperschroer, H., Proc. lOth Coli. ASlc. 1982, 279-92.

2. Trugo, L. C. and Macrae, R., Proc. lO/h Col!. AS1C, 1982, 187-92. 3. Barbirolli, O., Res. Chim., 1965, 17,261-3. 4. Wolfrom, M. L., Plunkett, R. A. and Laver, M. L., J. agric. Fd Chem., 1960,8,

58-65. 5. Kroplien, U., Proc. 5th Coli. ASIC, 1971,217-23. 6. Pokorny, J., Ouyen-Huy Con, J., Bulantova, H. and Janicek, O., Nahrung,

1974, 18, 799-805. 7. Southgate, D. A. T., Determination of Food Carbohydrates, Applied Science

Publishers, London, 1976. 8. Wolfrom, M. L., Laver, M. L. and Patin, D. L., J. org. Chem., 1961, 26,

4533-5. 9. Wolfrom, M. L. and Patin, D. L., J. agric. Fd Chem., 1964, 12, 376-7.

10. Wolfrom, M. L. and Patin, D. L., J. org. Chem., 1965,30,4060-3. Il. Pictet, O. and Moreau, A., Proc. 4/h Coli. ASIC, 1969,75-84. 12. Thaler, H. and Arneth, W., Proc. 3rd Coli. AS1C, 1967, 127-35. 13. Thaler, H. and Arneth, W., Z. Lebensm. Unters. Forsch., 1968,138,26-35. 14. Thaler, H. and Arneth, W., Z. Lebensm. Unters. Forsch., 1968, 138, 137--45. 15. Thaler, H. and Arneth, W., Z. Lebensm. Un/ers. Forsch., 1969,140,101-9. 16. Thaler, H., Z. Lebensm. Un/ers. Forsch., 1970, 143,342-8. 17. Thaler, H., Proc. 7th Coli. ASIC, 1975, 175-85. 18. Clifford, M. N .. Proc. Biochem., 1975 (March), 20-9.


19. Maier, H. G., Kaffee, Paul Parey, Hamburg, 1982. 20. Kroplien, U., J. agric. Fd Chem., 1974, 22, 110-15. 21. Coulson, J. N., in Developments in Food Colours-l, Ed. J. Walford, Applied

Science Publishers, London, 1980,201. 20. Lee, F. A., Basic FoodChemistry,AVI PublishingCo., Westport,Conn., 1983,

Chap. 12. 23. Asante, H. and Thaler, H., Z. Lebensm. Unters. Forsch., 1975, 159, 93-6. 24. Thaler, H., Chem. Mikrobiol. Technol. Lebens., 1974,3, 1-7. 25. Thaler, H., Fd Chem., 1979,4, 13-22. 26. Sabbagh, N. K., Faria, J. B. and Yokohizo, Y., Coletanea do lnstituto de

Technologia de Alimentos, 1977,8, 55-73. 27. J'rugo, L. C., PhD thesis, University of Reading, 1984. 28. Pictet, G. A., Proc. 7th Col!. ASIC, 1975, 189-200. 29. Ara, A. and Thaler, H., Z. Lebensm. Unters. Forsch., 1977, 164, 8-10. 30. Wolfrom, M. L. and Anderson, L. E., J. agric. Fd Chem., 1967, 15,685-7. 31. Streuli, H., Chimia, 1962,16,371-2. 32. Feldman, W. S., Ryder, W. S. and Kung, J. T., J. agric. Fd Chem., 1969, 17,

733. 33. Maier, H. G., Diemair, W. and Gansmann, J., Z. Lebensm. Unters. Forsch.,

1968, 137, 282-92. 34. Maier, H. G. and Buttle, H., Z. Lebensm. Unters. Forsch., 1973, 150,331--4. 35. Nakabayashi, T. and Watanabe, c., Nippon Shokuhin Kogyo Gakkaishi, 1977,

24, 124-9. 36. Klockung, R., Hofmann, R. and Mucke, D., Z. Lebensm. Unters. Forsch.,

1967, 135, 1-9. 37. Aurich, H., Hofmann, R., Klockung, R. and Mucke, D., Z. Lebensm. Unters.

Forsch., 1967, 135, 59-64. 38. Klockung, R., Hofmann, R. and Mucke, D., Z. Lebensm. Unters. Forsch.,

1970, 146, 79-83. 39. Maier, H. G., Z. Lebensm. Unters. Forsch., 1969,141,332-8. 40. Maier, H. G., Proc. 7th Col!. ASIC, 1975,211-19. 41. Maier, H. G., Z. Lebensm. Unters. Forsch., 1970,143.24-31. 42. Rey, L. and Bastien, M. c., in Freeze-Drying of Foods, Ed. F. R. Fisher,

National Academy of Sciences, Washington, DC, 1962,25. 43. Flink, J. and Karel, M., J. agric. Fd Chem., 1970, 18,295-7. 44. Flink, J. and Karel, M., J. Fd Sci., 1970,35,444-7. 45. Karel, M. and Flink, J., J. agric. Fd Chem., 1973,21,16-21. 46. Thijssen, H. A. C., Bomben, J. L. and Bruin, S., Advances in Food Research,

Academic Press, New York, 1973,20,2-111. 47. Voilley, A. and Simatos, D., in Food Process Engineering, Ed. P. Linko et al.,

Applied Science Publishers, London, 1980, 1, 371-84. 48. Courtois, J. E., Percheron, F. and Glomaud, J. c., Caje Cacao The, 1963,7,

231-6. 49. Shadaksharaswamy, M. and Ramachandra, G., Phytochem., 1968,7,715-19. 50. Charlesworth, J. M., Anal. Chem., 1978, 50, 1414-20. 51. Macrae, R. and Dick, J., J. Chromatogr., 1981,210, 138--45. 52. Macrae, R., Trugo, L. C. and Dick, J., Chromatographia, 1982, 15,476-8.

Chapter 4

Nitrogenous Components R. MACRAE

Department of Food Science, University of Reading, UK

1. INTRODUCTION

Cotfee is undoubtedly one of the most complex of the more commonly encountered food commodities from the point ofview ofits chemistry. Not so much because the green cotfee bean contains a wide range of different chemical compounds, but rather that these compounds react and interact at aU stages of coffee processing to produce a final product (acup of coffee) with an even greater diversity and complexity of structures. Our understanding of this complex mixture has been exacerbated by two major analytical problems, namely the extremely low levels at which some compounds are present, and yet wiU have sensory significance, and secondly that many of the interactions lead to the formation of high molecular weight polymeric material which is often very difficult to characterise structurally. Nonetheless, polymeric material may well contribute up to 50 % of a cotfee brew and is important for aroma retention as weU as physical properties. These analytical problems become readily apparent in a consideration of the nitrogenous components. Thus, for example, volatile nitrogen heterocyclic compounds may be present at the sub-ppb level, and the brown pigments or coloured matter present in coffee brew are mainly high molecular weight products of the browning reaction (melanoidins).

The term 'nitrogenous component' should strictly apply to aU those components which contain inorganic or organic nitrogen. However, in this chapter emphasis wiU be placed on three main groups of compounds:

115


alkaloids, trigonelline together with nicotinic acid, and amino acids and proteins. The other major nitrogenous components will be discussed under different headings. Thus, the volatile nitrogenous components will receive attention in Chapter 7 dealing with aroma components, and fat-soluble nitrogenous compounds, e.g. hydroxytryptamides, will be considered in Chapter 6 which is concerned with lipids. Any such divisions amongst the nitrogenous components, and indeed between other groups of compounds, are purely arbitrary and in many instances far from ideal. It is of ten necessary in discussing interactions between components to include material from more than one section, as for example in the case of reactions of amino acids and sugars.

2. AlKAlOIDS (CAFFEINE)

The puri ne ring system is widely distributed in nature, but purine itself is not encountered:

Purine

A series of methylated dioxypurines do, however, exist, of which 1,3,7-trimethyl-2,6-dioxopurine or 1,3, 7-trimethylxanthine, more commonly known as caffeine, is the most important. In fact this is the only significant xanthine alkaloid present in coffee, although traces of 1,3-dimethyl-xanthine (theophylline) and 3,7-dimethylxanthine (theobromine) have been reported:

Caffeine

Caffeine is a white compound which melts at 236 cC, although this is relatively unimportant as it sublimes at the much lower temperature of

NITROGENOUS COMPONENTS 117

178C. When obtained by crystallisation from aqueous solution, caffeine is obtained as a hydrate, which until recently was thought to contain one molecule ofwater to each caffeine molecule. However, more recent studies have indicated that it is in fact a 4/5 hydratewith 695 % water. 1 At elevated temperature (ca. 140C) there is a phase transformation [rom (X- to f3-caffeine which is revealed by X-ray diffraction. Caffeine is moderately soluble in water (ca. 46 %(w/w) at 40C), but the exact figure obtained depends on the conditions of measurement. Above 52C anhydrous caffeine is stable in contact with aqueous solution and conversely below 52C only the hydrate is stable. However, the interconversion is not very rapid and thus to determine solubilities over a range oftemperatures which includes this transition temperature (52 aC), it is necessary to commence the experiment with the correct [orm, hydrate for values below 52C and anhydrous above. It has been suggested that the wide range of solubilities reported may be explained by this hitherto unappreciated inter-conversion. 1

Caffeine has a considerable aqueous solubility at higher temperatures due to the formation of aggregates by 'basestacking'. This effect is illustrated by the solubility data shown in Table 1. Caffeine is moderately soluble in a wide range of organic solvents even at relatively low temperatures, as shown in Table 2. Of these solvents many have been used for decaffeination, although in recent years dichloromethane (methylene chloride) is the most widely used. It is also soluble in supercritical carbon dioxide and this solvent has been more recently used for decaffeination, on account of possible toxic effects of methylene chloride residues. The actual solubility of caffeine in these chlorinated solvents, and presumably also in

Table 1 Solubility of caffeine in water 2

emperature ro c)

o 15 20 25 30 40 50 60 70 80

Solubility (g per 100g H 2 0)

060 100 1-46 213 280 464 675 970

1350 1923

118 COFFEE: VOLUME }-CHEMISTRY

Table 2 Solubility of caffeine in organic solvents 2 ,3

Solvent

95% aqueous ethanol Ethanol Ethyl acetate Methanol Acetone Benzene Carbon tetrachloride Chloroform Ether Petroleum ether Trichloroethylene D ic h loroethylene Toluene

Dichloromethane 4 (methylene chloride)

Tempera ture tG) 25 25 18 25 30,5 18 18 17 18

15-17 15 15 25 33

Solubility (g per 100g solvent)

1 ,32 1 ,88 0,73 1 ,14 2,32 0,91 0,09

12,9 0,12 0,03 0,76 1,82 0,58

9

supercritical carbon dioxide, depends on the crystal form and water content of the cafTeine,

Caffeine is a very weak base forming unstable salts, e,g. acetate from which the acetic acid may be readily volatilised. It is relatively stable in dilute acids and alkalis but can [orm a series of complexes with other cofTee components, for example chlorogenic acids or polynuclear aromatics. In fact this property is also used in the selective extraction of polynuclear aromatics from other foods.

2.1. Caffeine Content of Green. Roasted and Instant Coffees The cafTeine content of green cofTees varies widely. with differences in species being the most important factor. However, even within a species there is a verywide range ofvalues, The data shown in Table 3 give some idea ofthis wide range of results. Clearly robusta cofTees in general have a higher caffeine content with an overall mean value of 22 % db, whilst the value for arabicas is about 1 ,2 % db. Intermediate values have also been reported for commercially less important species such as liberica (mean value 1, 35 % db) and the arabusta hybrid (mean value 1, 72 % db), The availability of cofTees of the Paracojlea genus in Africa and Asia with very low cafTeine contents raises the possibility of forming new genetic combinations by cross-

NITROGENOUS COMPONENTS 119

breeding, and hybrids with very 10w caffeine contents (02 %) are available (Java and Ivory Coast). However, to date these hybrids have had little impact on the commercial markets, mainly due to poor coffee quality and also to the fact that the caffeine removed in the normal decatTeination process has commercial value. Plant-breeding trials of this kind demand ace ura te and precise analytical methods for the determination of catTeine contents, and it is only in recent years that the methods available have been adequate, particularly in terms of specificity, as different cotTee types may contain different interfering components. 1 7 Environmental and agricul-tural factors are considerabl~ less important than genetic variations in controlling the caffeine contents of green beans, and it has been reported that fertilisers, in particular potassium, phosphate, magnesium and calcium, do not have a significant effect on caffeine or chlorogenic acid content and also do not affect the colour of the final ground cotTee. 6

During the roasting of green coffee the bean temperature wiU be raised, by a combination of external heating and exothermic chemical reactions, to above 200 ae. This tempera ture is weU in excess of the sublimation point of caffeine and thus it would be expected that considerable losses would occur. However, it is found in practice that the losses are relatively modest and, unless severe roasting conditions are employed, rarely amounts to more than a few per cent. Indeed, as the weight ofthe green beans wiU be reduced by u p to 20 % or more (sa y 10 % water, 10 % dry ma tter) during roasting, the actual percentage amount of caffeine may increase by up to 10% on a dry roasted basis. The reasons for this modest loss of caffeine are complex, but the two major contributing factors are probably an increase in the sublimation point of caffeine as a result of pressure build-up within the bean and a poor rate of diffusion of vapour through its outer layers. Caffeine wiU also form salts as a result of the mildly acidic conditions which prevail within the bean, and which increase during roasting, but as these salts are relatively weak they wiU decompose and thus should have little effect on the sublimation process.

The solubility of caffeine in water increases rapidly with temperature, with the result that, under conditions used for the industrial preparation of instant coffee, virtuaUy aU the caffeine is extracted to yield instant coffee powders with levels in the range 28-46 % 7 as determined by HPLC on 13 commercial samples in the UK. 18 The values wiU depend on both the soluble solids yield from the roasted coffee and also the blend. Brazilian instant coffee powders generaUy contain somewhat lower levels. 19 In theory, ifthe caffeine content of the roasted beans to be extracted is known and its content in the final powder can be determined, it is possible to

Tabl

e 3

Caffe

ine

con

tent

of g

reen

co

ffees

(%db

) R

efer

ence

M

etho

d o

f Co

ffees

R

esul

ts

dete

rmin

atio

n e

xam

ined

Ty

pe

No.

of

Mea

n R

ange

SD

sa

mpl

es

valu

e

Mai

er5

Kum

-Tat

t14

Co

mpi

latio

n,

Arab

ica

Nar

row

1

2 0

9-1

4 u

sing

Cha

rrier

's ra

nge

figur

es

Extre

me

06-

19

rang

e Ro

bust

a N

arro

w

20

15-

26

Extre

me

12-

40

Cliff

ord

6 Co

mpi

latio

n Ar

abica

0

9-1

2 (C

harrie

r Ro

bust

a 1

6-2

4 n

ot c

ited)

Kr

oplie

n 7

Koga

n e

t al. 1

5

AII c

om

me

rcia

l Ar

abica

36

1

19

096

-14

0 sa

mpl

es,

Robu

sta

7 1

88

156

-21

6 w

orld

wid

e St

reul

i 8

Com

pila

tion

(7)

Arab

ica

120

020

Ro

bust

a 1

90

0

20

Libe

rica

135

R

offi

9 Ku

m-T

att1

6

AII A

ngol

an

Arab

ica

8 1

32

121

-14

5 (D

'Orn

ano

coffe

es

Robu

sta

26

242

2

18-2

72

mo

difie

d)

Chas

seve

nt e

t al. 1

0 Ku

m-T

att1

6

Cam

eroo

n Ro

bust

a 10

2

21

Smilll

(D

'Orn

ano

robu

sta

(dry p

roce

ss)

mo

difie

d)

(5 di

ffere

nt

Robu

sta

30

219

Sm

all

loca

tions

) (w

et pr

oces

s) Ch

arrie

r and

Ku

m-T

att1

6 Co

ffees

in

Arab

ica I.

C.

144

122

0

84-1

52

Berth

aud1

1 (D

'Orn

ano

geno

type

76

0,

72-1

,57

C: la

rrier

1 2

mo

difie

d)

colle

ctio

ns in

38

3 1

20

0,77

-1,9

0 th

e Iv

ory

Coas

t, M

. 13

0 1

16

058

-16

9

0'18

M

adag

asca

r and

(ga

ussia

n) Ca

mer

oon

C.

34

135

0,

90-1

,89

Braz

ilian

culti

vars

0,

80-1

,20

(Bou

rbon

, Cat

urra

an

d M

ondo

Nov

o)

Robu

sta

(Can

epho

ra)

I.C.

(clon

es)

163

251

}

r41 va

r. Q

uillo

u 2

76

1.16

--4.0

va

r. Ro

bust

a 2

44

M.

(clon

es)

681

214

0'32

W

urzi

ger1

3 Ku

m-T

att1

4 Iv

ory

Coas

t Ar

abus

ta

7 1

72

1,47

-1,8

3 ar

abus

ta

Char

rier1

2 Ku

m-T

att1

4 Lo

w c

affe

ine

Coffe

e co

ffee

in

C. e

uge

noid

es

023

-0,5

1 ge

nera

or

C. ra

cem

osa

0,50

-1,2

0 su

b-ge

nera

M

asca

roco

ffea

Gen

eral

ly a

bsen

t Pa

raco

ffea

Abse

nt


estimate the yield of soluble solids from the roasted coffee. Such an approach assumes complete extraction of the caffeine followed by progressive extraction ofthe less soluble fractions, such as polysaccharides and other polymeric material. Such estimates have also been attempted with other soluble fractions, e.g. mineral components, 20 but here again the wide variability of the levels in the starting material means that unless this is actually analysed, the determination of the analyte in the final product has !ittle predictive value (see Chapter 2).

Concern over the possible toxic effects and undesirable physiological effects of caffeine has led to an increasing demand for decaffeinated green beans to the extent that some 8-10% of the world export market is now treated, with the majority being sold in the USA and Western Europe. The most commonly employed method 9 uses chlorinated hydrocarbons (e.g. dichloromethane) or other organic solvents (e.g. ethyl acetate) to extract the caffeine directly from green beans which have been previously steamed to give a moisture content of about 40 %. The extraction is a slow process and may be semi-continuous or batch-wise; in the latter case several extractions will be made, each taking 1-2 h. The solvent is finally drained from the green beans and final traces removed by extensive steaming for several hours to provide a finallevel well below 15 ppm. The beans are then dried with warm air in rotary, fluid bed or vacuum driers. The decaffeinated beans may discolour rapidly, due to oxidation, as the protective wax layer has been removed. Only a small proportion of the beans is sold for roast and ground coffee, mainly in the USA, the majority being processed directly for decaffeinated instant coffee. Alternative methods of de-caffeination use hot water directly, which is recycled after removal of the caffeine with organic solvents 21 or supercritical carbon dioxide. 22 Irrespective of the process employed, the final caffeine content of 'decaffeinated' instant coffee must be generally below O 3 % (e.g. legislation in EEC countries), although in many commercial products it is within the range 01-02%.23

Caffeine is odourless but has a marked bitter taste. Several attempts ha ve been made to correlate the sensory characteristic of bitterness of coffee brews with their caffeine contents, but with little success. In fact, it has been shown that caffeine contributes only a relatively small proportion (ca. 10%) of the perceived bitterness. 2425 The nature of the other components contributing to bitterness is not fully understood and in particular the sensory properties of high molecular weight browning products are poorly defined. Correlations of this kind between sensory characteristics

Coffee Volume 1 Chemistry

Documents

chemistry of coffee

coffee products

roasted coffee

instant coffee

roasting of coffee

term coffee

publication data coffee

chemistry volume