CALL (336) 669-2460 FOR SALES & SERVICE

CALL (336) 669-2460 FOR SALES & SERVICE

CALL (336) 669-2460 FOR SALES & SERVICE

Close this search box.

Espresso Coffee

The Science of Quality

This page intentionally left blank

Espresso Coffee

The Science of Quality

Second edition

Edited by




with the assistance of Furio Suggi Liverani

This book is printed on acid-free paper

First published 1995

Second edition 2005

Copyright # 1995, 2005, Illycaffe` s.p.a. All 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

Permissions may be sought directly from Elsevier’s Science & Technology Rights

Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333,

e-mail: [email protected]. You may also complete your request on-line via

the Elsevier homepage (, by selecting ‘Customer Support’

and then ‘Obtaining Permissions’

Elsevier Academic Press

525 B Street, Suite 1900, San Diego, California 92101-4495, USA

Elsevier Academic Press

84 Theobald’s Road, London WC1X 8RR, UK

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Catalog Number: 2004113615

ISBN 0-12-370371-9

Printed and bound in Italy

05 06 07 08 09 9 8 7 6 5 4 3 2 1


Contributors ix

Acknowledgements xvii

1 Quality 1

A. Illy

1.1 Origins and meanings of quality 1

1.2 Definition of quality 2

1.3 Commercial quality 3

1.4 Quality of food products 7

1.5 The experience of coffee consumption 10

1.6 The quality of espresso coffee 15

1.7 Definition of espresso 16

1.8 Conclusions 19

References 20

2 The plant 21

F. Anzueto, T.W. Baumann, G. Graziosi, C.R. Piccin,

M.R. So¨ndahl and H.A.M. van der Vossen

2.1 Origin, production and botany 21

2.2 Variety development 29

2.3 Agronomy 34

2.4 Biochemical ecology 55

2.5 Molecular genetics of coffee 67

References 76

3 The raw bean 87

S. Bee, C.H.J. Brando, G. Brumen, N. Carvalhaes,

I. Ko¨lling-Speer, K. Speer, F. Suggi Liverani, A.A. Teixeira,

R. Teixeira, R.A. Thomaziello, R. Viani and O.G. Vitzthum

3.1 Introduction 87

3.2 Harvesting 87

3.3 Processing of the harvest 91

3.4 Drying 96

3.5 Final processing for export and roasting 101

3.6 Logistics 108

3.7 Defects 116

3.8 Classification: physical and sensorial analysis 134

3.9 Blending 141

3.10 Decaffeination 142

3.11 Raw bean composition 148

References and further reading 166

4 Roasting 179

B. Bonnla¨nder, R. Eggers, U.H. Engelhardt and H.G. Maier

4.1 The process 179

4.2 Roasting techniques 184

4.3 Changes produced by roasting 191

4.4 Volatile aroma compounds 197

4.5 Melanoidins 204

4.6 Contaminants 209

References 209

5 Grinding 215

M. Petracco

5.1 Theory of fracture mechanics 216

5.2 Coffee grinders 218

5.3 Methods for measuring ground product fineness 221

5.4 Parameters influencing grinding 224

5.5 Physico-chemical modifications due to grinding 227

References 229

6 Storage and packaging 230

M.C. Nicoli and O. Savonitti

6.1 Physical and chemical changes of roasted coffee during

storage 230

6.2 Packaging of roasted coffee 245

References 255

7 Percolation 259

M. Petracco

7.1 Conceptual definitions 259

7.2 Physical and chemical characterization of the percolation

process 261

7.3 Modelling of the percolation process 266

7.4 The espresso machine 270

vi Contents

7.5 Parameters influencing percolation 274

References 287

8 The cup 290

M. Petracco

8.1 Physical and chemical characterization of the espresso

beverage 290

8.2 Organoleptic characteristics of espresso (practical aspects) 300

8.3 Espresso definition again 310

8.4 Espresso–milk mixes 311

References 313

9 Physiology of perception 316

R. Cappuccio

9.1 Introduction 316

9.2 Gustation 317

9.3 Olfaction 325

9.4 Human chemosensory psychophysics 332

References 345

10 Coffee consumption and health 352

M. Petracco and R. Viani

10.1 Consumption patterns 353

10.2 Coffee is more than caffeine 357

10.3 Coffee is beneficial to health 357

10.4 Coffee is not harmful to health 365

10.5 Conclusions 369

References 369

Closing remarks 384

E. Illy

Index 385

Contents vii

This page intentionally left blank


Francisco Anzueto graduated in 1978 in Agronomy in Guatemala and

obtained a doctorate in Coffee Plant Breeding in France in 1993. He was

regional director of Anacafe from 1978 until 1986, where he is now

research director. He holds a professorship at Landivar University in

Guatemala. He has been co-manager or director of several regional

research programs in Central America. He is the co-author of many

scientific papers.

Thomas W. Baumann received his scientific education in Medicine and

Biology at the University of Zurich, which he completed in 1969. After

training in Molecular Biology at the University of Zurich he became

Research Assistant at the Institute of Plant Biology of the same university,

where he was ‘infected’ for the first time with coffee while chemically

analysing and establishing tissue cultures of several species of Coffea.

After postdoctoral studies at the University of Southern California (Allan

Hancock Foundation) in Los Angeles in 1972 and 1973, he returned to

the Institute of Plant Biology, where he continued his phytochemical

work on coffee and other caffeine-containing plant species. In 1977 he

became Docent, and 1995 Professor, of Plant Physiology. His research

focused on the biosynthesis and ecological significance of caffeine and

related purine alkaloids. He also studied the ecological chemistry of other

plant genera, such as Physalis, Echinacea and Catharanthus, until

retirement in 2001. His hobbies are scientific writing and gardening. In

2001 he published together with his wife a voluminous biography of Henri

Pittier, the ‘Humboldt of Switzerland’ (Basle: Friedrich Reinhardt

Verlag). At his home near Zurich he houses a living collection of almost

all caffeine-containing species consumed in the world. Currently, he is

working on a book dealing with the various facets of little-known fruits of

the Neotropics.

Sarah Bee is a radiation physicist with a doctorate from University

College London, UK. She is a Chartered Physicist and Member of the

Institute of Physics. Since 1998 she has been Engineering Product

Manager within the R&D Department at Sortex. She was responsible for

the technical description of the ‘Niagara’ vegetable sorting machine

(colour and shape sorting at 40 000 objects/sec), which obtained Sortex’s

Queen’s Award for Innovation application in 2001.

Bernd Bonnla¨nder graduated in Food Chemistry at Erlangen University,

Germany. After his doctorate on the characterization of aroma precursors

in wine and spices, he joined AromaLab (R&D illycaffe`), where he

performs research on odour active and physiologically active compounds

in green and roasted coffee as well as beverages.

Carlos Henrique Jorge Brando is an engineer with graduate work at

doctoral level in Planning and Business at the Massachusetts Institute of

Technology. He was the manager and director of leading Brazilian coffee

machinery maker Pinhalense for over ten years. He is now director and

partner of a coffee consulting, marketing and trading company, P&A

International Marketing, which was a technical adviser to the Cafe´ do

Brazil Marketing Program. He has been responsible for coffee processing

and marketing projects in over 60 countries and written and spoken

frequently about the two subjects.

Gianfranco Brumen graduated in Chemistry from the University of

Trieste, and joined illycaffe` in 1972 in charge of Quality Control and

Research, becoming an expert in the raw product (characteristics of coffee

types, production countries, raw coffee selection, blend formulation, etc.),

controlling all the production cycle. As Quality Supervisor he is in charge

of blend formulation, final product conformity, quality system application

and its certification (ISO 9001), and all regulatory aspects.

Roberto Cappuccio graduated in Physics at the University of Trieste in

1997 with a thesis on Near Infrared Spectrometry. Immediately afterwards

he entered the R&D Department at illycaffe` as a researcher, where he

dealt mainly with fluid dynamics. Since 2001 his interests have moved

towards the exploration of the sensory and chemical aspects of taste,

aroma and mouthfeel perception of coffee. Since 2002 he has coordinated

illycaffe` Sensorylab. He has written several scientific publications, mainly

in coffee science and technology.

Nelson Carvalhaes is a lawyer who graduated in Agribusiness and Foreign

Trade at Mackenzie University, Sao˜ Paulo in 1981. He has been a

Director of Escrito´rio Carvalhaes Cafe´ Ltda since 1980 and Porto de

Santos C.E. Ltda since 1989.

x Contributors

Rudolf Eggers graduated and obtained his doctorate in Mechanical

Engineering/Process Techniques at the Technical University in Hanover

in 1972. He was research associate at the Technical University in

Clausthal until 1977. From 1977 until 1984 he was head of Process

Engineering at Thyssen Maschinenbau and Krupp Industrietechnik. Since

1984 he has been Professor for Heat and Mass Transfer at the Technical

University in Hamburg–Harburg. He has conducted research and

published in separation processes (extraction, adsorption, refining); coffee

processing (roasting, extraction, decaffeination); interfacial phenomena

(transport effects across phase boundaries); high pressure technique

(supercritical fluids).

Ulrich H. Engelhardt graduated as a Food Chemist and obtained his

doctorate with work on stimulant drinks at the Technical University of

Braunschweig, where he is now a teacher in food analysis, food toxicology

and chemistry and technology of foods containing polyphenols. He has

published about 70 papers, mainly on coffee and tea chemistry. He is

member of the working groups on tea and coffee of the German Institute

of Standards (DIN). In 1996 he received the Young Scientist award of the

German Chemical Society.

Giorgio Graziosi graduated in Natural Sciences in 1966 at the University

of Trieste, where he was appointed Lecturer in Genetics in 1969. He

obtained a position of Associate Professor in 1976 and of Full Professor in

1986. He was appointed Director of the Department of Biology in 1993

and President of the Faculty of Biological Sciences, University of Trieste,

in 1998. He has been visiting scientist at the Developmental Genetics

Unit, International Embryological Laboratory, ‘Hubrecht’, Utrecht,

Holland, at the Genetics Department, University College London, UK

and for some years at the Genetics Laboratory of Oxford University, UK.

His research interests 1967–86 were developmental genetics of Drosophila

melanogaster and since 1987 have focused on DNA polymorphism in

several organisms including Coffea and genomics and gene expression in


Andrea Illy graduated in Chemistry at the University of Trieste and

completed the Master Executive programme at SDA Bocconi in Milan.

He is the chief executive officer of illycaffe`. His first working experience

was in 1983, at the Nestle´ Research and Development department. In

1990 he began his career at illycaffe` as supervisor of the Quality Control

department, where he instituted a total quality control programme. He

has been president (1999) of the Association Scientifique Internationale

Contributors xi

du Cafe´ (ASIC) and organized the 19th International Colloquium on

Coffee Science in Trieste; he is presently vice-president. Since 1999, he

has been a member of the board of directors of the Italian Association of

Premium Brand Industries. In 2003 he was appointed a member of the

advisory board of SDA Bocconi in Milan.

Ernesto Illy graduated in Chemistry at the University of Bologna. He has

been chairman of illycaffe` since 1963 and an initiator active in

international bodies devoted to the scientific and professional development

of coffee. He was co-founder and a past president of Association

Scientifique Internationale du Cafe´ (ASIC), of which he is now senior

vice-president; co-founder, past president and an active member since its

foundation of the Physiological Effects of Coffee Committee (PEC); cofounder

and president of the Institute for Scientific Information on Coffee

(ISIC); President of the Promotion Committee of the Private Sector

Consultative Body (PSCB) of ICO; President of Centromarca (Italian

Association of Trademark Industries). He has been honoured with

important awards from the international coffee community and has

received from the President of Italy the title of Cavaliere del Lavoro.

Isabelle K¨olling-Speer graduated in Food Chemistry at the University of

Hamburg. She obtained her PhD in food chemistry from the same

university in 1993. Before joining the Institute for Food Chemistry at the

Technical University of Dresden in 1993, where she is Scientific

Researcher, she worked with the Hygiene Institute Hamburg in the

Department of Chemistry and Food Chemistry. She is co-author of two

books and has published over 20 scientific publications. Her research

activities are related to the processing and quality control of foods of plant


Hans Gerhard Maier studied Pharmacy and Food Chemistry at the

Universities in Freiburg (Breisgau) and Frankfurt (Main), obtaining his

doctorate in 1961. He also studied chemistry from 1963 to 1965 at the

University of Frankfurt, obtaining the doctorate in 1969. After working

from 1961 to 1963 in the Food Research Laboratory of Franck und

Kathreiner in Ludwigsburg, he became in 1963 scientific assistant at the

University of Frankfurt, lecturer in 1970, assistant professor in 1971. After

two years spent at the University of Mu¨nster as Wissenschaftlicher Rat

und Professor, he became full professor and director of the Institute of

Food Chemistry at the Technical University in Braunschweig, from where

he retired in 1998. He is the author of many important publications on

coffee, and, in particular, of a book Kaffee (Berlin: Paul Parey, 1981).

xii Contributors

Maria Cristina Nicoli graduated in Pharmaceutical Chemistry at the

University of Bologna. After a decade of post-doctoral research activity at

the University of Udine as assistant professor she moved to the University

of Sassari, where she was appointed Professor in Food Technology. One

year later she was called back to the University of Udine, where she

continues research and teaching activity. Her research activity is mainly

focused on chemical and physical factors affecting food functionality and

shelf-life with particular attention to processed fruit and vegetables and

roasted coffee. She is author of over 80 scientific papers published in

international journals of food science and technology. She has been in

charge of several Italian and European research projects.

Marino Petracco graduated in Chemical Engineering at the University of

Trieste. After 10 years spent in industrial research working in a

multinational petrochemical company, specializing on catalytic cracking,

he joined illycaffe`, working on topics ranging from plant botany to

industrial transformation processes, beverage brewing dynamics and its

effects on the consumer’s body. He has published widely, including 30

research papers on sensory chemometrics, roasted coffee grinding

dynamics, physicochemical characterization of espresso coffee, beverage

brewing hydraulics and raw coffee sampling, and is co-author of several

books. He is active in several scientific associations: in coffee science,

ASIC and PEC, where he held the chair in 1991 and 1992; in coffee

standardization, the Italian organization for standardization (UNI), the

Association Franc¸aise de Normalisation (AFNOR, commission V33A

cafe´) and the International Standardization Organization (ISO TC 34 SC

15 ‘Coffee’); in food hygiene and microbiology, the Association Africaine

de Microbiologie et d’Hygie`ne Alimentaire (AAMHA) and the Sociedad

Latino Americana de Micotoxicologı´a (SLAM).

Carlos Roberto Piccin graduated in Agronomy at the Universidade

Paulista, Jaboticabal SP, Brazil in 1981, and then attended the Graduate

School in Mineral Nutrition at the Universidade Federal, Lavras, MG,

Brazil. After a career as technical manager with various coffee companies,

in 1999 he created Agropiccin, a technical consulting firm expert in

coffee production systems, with emphasis on plant nutrition linked with

integrated pest management seeking productivity, quality product and

biological equilibrium. He is often an invited speaker at universities and

in the private sector.

Contributors xiii

Oriana Savonitti obtained a Masters in Process Chemical Engineering

from the University of Trieste and has been a researcher in the Research

and Development department of illycaffe` since 1998.

Maro R. S¨ondahl obtained in Brazil a BS degree in Agriculture

Engineering in 1968 and MS in Plant Physiology in 1974. In the USA

he obtained a PhD in Managing Technology and Innovation in 1968 and

in Cell Biology in 1978. His professional career developed in government

(13 years) and private organizations (20 years). Since 1994 he has been

President of Fitolink Corporation in the USA, after working at an

Institute of Agronomy in Brazil from 1970 to 1983, and at DNA Plant

Corp. in the USA from 1983 until 1993. He is the author of five US

patents and more than 40 scientific papers and review articles in the area

of plant physiology and plant cell genetics of tropical plants with

emphasis on coffee, cacao and oil palm. He acts as a consultant to several

private coffee companies in the USA and Europe and to international

organizations. He has been Visiting Professor and/or invited lecturer at

several universities in the USA, Brazil, Argentina, Venezuela, Costa Rica,

Colombia and Chile, and an invited speaker at scientific meetings in

more than 20 countries.

Karl Speer completed his degree in food chemistry in 1985, and his

doctorate at the University of Hamburg, Germany in 1993. He was

affiliated with the Hygiene Institute Hamburg, Department of Chemistry

and Food Chemistry, from 1984 until 1993. He has been with the

Institute for Food Chemistry in Dresden since 1993. In 1997 and 1998 he

was a visiting scholar at the Department of Food Science in West

Lafayette, Purdue University, USA. He received the Josef Schormu¨ller

Award in 1992. He is co-author of two books and published over 50

scientific publications, especially in the area of coffee, honey and the

analysis of pesticide residues and organic contaminants in foods. He is

now professor of Food Chemistry and Food Production, Department of

Food Chemistry, Technical University of Dresden.

Furio Suggi Liverani joined illycaffe` in 1991 in the role of knowledge

manager after having worked for 10 years as a computer science adviser.

Since 1995 he has been managing director of the Department of Research

and Development, directly under the CEO. He has participated in several

projects of industrial and laboratory automation and in the development

of illycaffe`’s production information system. He has developed technologies

and innovative products such as an electronic coffee sorting

machine. In 1993 he introduced the use of Internet and multimedia to

xiv Contributors

the company, actualizing the website in 1996. In 1998 on

behalf of illycaffe` he founded AromaLab, a laboratory dedicated to the

study of coffee chemistry, located in the Area Science Park of Trieste. He

has been a council member of the AI*IA (Italian Association for

Artificial Intelligence), co-ordinator of the workgroup AI*IA in industry,

and is a member of several scientific societies. Currently he is member of

the steering committee of the international conference ‘Cellular

Automata for Research and Industry’ and of the ‘Italian Congress of

Artificial Intelligence’. He is also a board member of Qualicaf and adviser

of the EDAMOK project. He is the author of 30 articles on the subjects of

coffee technology, computer science and artificial intelligence, he has

participated as spokesman in several courses, national and international

conferences and conventions. He is the co-inventor of five industrial


Aldir Alves Teixeira is an Engineer Agronomist, having graduated at Sao˜

Paulo University in 1959. In 1972 he gained a doctorate in Agronomy on

Cup Tasting from Sao˜ Paulo University. He has been president of his own

company, ASSICAFE´ , since 1992, president of the Scientific Research

Association from 1990 to 1993, president of the Evaluation Board of

Commission on the Brazilian Coffee Prize for ‘Espresso’, promoted by

illycaffe`, since 1991 and a member of the Brazilian Delegation to ISOTC34

SC 15 ‘Coffee’. He has been a scientific consultant of illycaffe` since 1991.

Ana Regina Rocha Teixeira is a Biologist who graduated at Sao˜ Paulo

University in 1988. She worked at the Phytopatologic Biochemical

Department of the Biologic Institute of Sao˜ Paulo from 1986 to 1992.

Since 1992 she has worked with ASSICAFE´ , has been a member of the

Evaluation Board of Commission on the Brazilian Coffee Prize for

‘Espresso’, promoted by illycaffe`, since 1991, she contributes to the

Brazilian Delegation to ISO TC 34 SC 15 ‘Coffee’ and has collaborated

with illycaffe` since 1992.

Roberto Antˆonio Thomaziello is an Engineer Agronomist, having

graduated at Sao˜ Paulo University in 1965. His positions have included:

assistant technician of the Coffee Office, CATI, from 1968 to 1977 and

from 1990/91 to 1998; Project Supervisor for the Coffee Agricultural

Secretariat of Sao˜ Paulo from 1977 to 1986; head of the Department of

Rural Assistance of the Agricultural Secretariat/SP from 1992 to 1995;

director of group techniques at the Agricultural Secretariat from 1995

until 1997; and since 1999 at the Coffee Research Center ‘Alcides

Carvalho’ of the Campinas Agronomic Institute.

Contributors xv

Herbert A.M. van der Vossen graduated in 1964 with an Ir (MSc) degree

in Tropical Agronomy and Plant Breeding and gained his doctorate in

1974 with a thesis on breeding and quantitative genetics in the oil palm

from the Wageningen Agricultural University. He was research officer in

charge of the Oil Palm Research Centre at Kade, Ghana, during

1964–1971 and then head of the Coffee Breeding Unit at the Coffee

Research Station near Ruiru, Kenya, from 1971 to 1981. He was research

manager then director of the breeding programmes for vegetable and

flower seed crops with Sluis & Groot Seed Company at Enkhuizen in the

Netherlands from 1981 until early retirement in 1993. From 1993 until

1996 he was seed policy adviser to the Ministry of Agriculture at Dhaka,

Bangladesh and freelance consultant in plant breeding and seed

production, including assignments in vegetable seeds, cereal crops,

cocoa, oil palm and coffee. He has contributed more than 40 scientific

papers and chapters. He is a member of the board of ASIC.

Rinantonio Viani, after graduating in Chemistry at the University of Pisa,

did postgraduate work at CalTech in Pasadena, California, and Duke

University in Durham, NC. From 1963 until 1974 he was a Research

Chemist at the Research Laboratories of Nestle´ in Vevey and Orbe. From

1975 until retirement in 1998 technical and scientific advisor on

stimulant drinks in the Nestle´ headquarters. He has been Chairman of

PEC, Administrative Secretary and President of ASIC and Chairman of

the Technical Commission of European Decaffeinators, of AFCASOLE

and of ISO TC 34 SC 15 ‘Coffee’ for many years. He has obtained several

patents and contributed many scientific publications in food science and

technology, particularly in the field of stimulant drinks. After retirement

he has worked as consultant with the UN agencies UNIDO and FAO for

coffee projects in producer countries.

Otto G. Vitzthum is a Chemist with a doctorate from the University of

Erlangen and postgraduate studies in Madrid. He was Director of

Scientific Research at HAG AG in Bremen from 1978 to 1986;

Department Manager in Central Research at Jacobs Suchard from 1986

until 1994 and Director of Coffee Chemistry worldwide at Kraft Jacobs

Suchard from 1994 until retirement in 1997; Lecturer at the Technical

University Braunscheig since 1978 and Honorary Professor since 1987.

He has been scientific secretary of ASIC and chairman of the technical

commission of European Decaffeinators for many years, and twice

Chairman of PEC. He is the author of many patents and scientific

publications in the field of coffee.

xvi Contributors


The Editors thank Ms Elaine Dentan and Mr Fabio Silizio for the

microscopy photos, Prospero S.r.l. for all figures and electronic drafts, and

Ms Andrea Appelwick for her constant editorial assistance.

This page intentionally left blank



A. Illy

The term ‘quality’ has been abused since the later 1980s, when Western

industries realized that lack of quality accounted for their diminished

competitiveness with respect to those of Japan. This tendency to speak, at

times inappropriately, about quality seldom reflects the true meaning of

the word itself, although this is perhaps understandable and justifiable if

one considers that ‘quality’ can convey just as many meanings as the

‘ability’ an object has of producing effects, as will be seen below.

Quality is highly pervasive, and this prevents anyone who decides to

‘create quality’ from clearly defining a field of action. In companies

recognized on the market for their quality, it is obvious this involves the

entire staff, with no limits in time or space. In other words, quality is


The following paragraphs will try to sketch at least an outline of this

fascinating concept ‘quality’.


The concept of quality is vast and cannot, therefore, be narrowed down to

one meaning. Aristotle distinguished four different families of qualities,

which were subsequently adopted by scholars:

1 Tendencies and aptitudes (examples of tendencies are temperance,

science and virtue, while health, illness, heat, cold, etc. are

examples of aptitudes).

2 Natural abilities and faculties, or active qualities.

3 Sentiments and passions, or passive qualities (sounds, colours, taste,


4 Shapes or geometric determination (square, circle, straight line,


These families can be classified as:

n attitudinal qualities, which include aptitudes, habits, abilities, virtues

and tendencies, or all the ‘possibilities’ of the object;

n sensorial qualities, which can be perceived by our sense organs and

include sounds, colours, smells, etc.;

n measurable qualities, which can be measured objectively, such as

speed, length, intensity, mass, etc.

The attitudinal qualities refer to the identity attributes of an object, or

what distinguishes it from another, without value hierarchies, whereas the

sensorial and measurable qualities refer to all the value attributes, which

make one thing better than another.

Identity and value are the two poles of meaning of the word ‘quality’. In

a critical analysis, the two poles must be compared with the episteme –

the cognitive dimension of the certain knowledge – and with the taste –

the subjective orientation of individual knowledge. This is of particular

importance for the understanding of quality as innate excellence.

In the modern industrial application, measurable quality prevails: ‘You

can’t have quality if you can’t measure it’ wrote Juran (1951), based on

the scientific method. Alternative and complementary approaches can be

applied based on the product, the user or the production, and the

prevailing dimensions are performance, reliability, conformity, life and

functional efficiency, each with its own units of measurement.

Two ways of considering quality will therefore be compared: the first is

philosophical, where aesthetics concerns itself with sensorial qualities,

and the second is scientific, where everything is related to the

characterization and quantification of measurable qualities.


The range of the term ‘quality’, with all its applications and facets, makes

an exact and concise definition difficult. This explains why there are so


The ‘official’ definition is provided by the International Organization

for Standardization (ISO):

The extent to which a group of intrinsic features (physical, sensorial,

behavioural, temporal, ergonomic, functional, etc.) satisfies the

2 Espresso Coffee

requirements, where requirement means need or expectation

which may be explicit, generally implicit or binding. (ISO, 2000)

There is, moreover, a series of definitions provided by specialists on the

subject. Among the best known are:

Conformance to requirements (Crosby, 1979)

Fitness for use (Juran, 1951)

The efficient production of the quality that the market expects

(Deming, 1982)

The total composite product and service characteristics of marketing,

engineering, manufacturing, and maintenance through which the

product and service in use will meet the expectations of the

customer (Feigenbaum, 1964)

Meeting or exceeding customer expectations at a cost that represents

value to them (Harrington, 1990)

Anything that can be improved (Imai, 1986)

Does not impart loss to society (Taguchi, 1987)

Degree of excellence (Webster’s, 1984)

Although the official definition provides a desirable fusion of the most

important meanings of the original definitions, it must be noted how the

interpretations of these authors make a more or less marked reference to

the various families of qualities described above. Focusing on a few of

these definitions reveals the degree of subjectivity or objectivity of the

qualitative characteristics they take into consideration.


Quality has always been of substantial importance in trade relations and

this is understandable as the value of an exchanged good is the

relationship between its quality and its price. Until a few years ago,

commercial quality was product quality. Nowadays the term total quality

is used to broaden the concept of quality to the service involved in the


1.3.1 Historic evolution

In the eighteenth and early nineteenth centuries when production was

limited exclusively to cottage industries, quality, as we know it today, did

not exist. The formal quality audits became necessary only with the

Quality 3

arrival of mass production, mostly in the military field, to such an extent

that, in 1941, the American War Department formed a committee aimed

at enabling the army to procure arms and ammunition in large quantities

without incurring problems of quality. In 1950, only one-third of the

electronic military installations worked properly, and so a working group

of the US Department of Defense was formed to deal specifically with

reliability. This is how reliability came to be one of the ‘chapters’ of


In the post-war period, with the publication in 1951 of the Quality

Control Handbook by Juran, the concept of ‘the economy of quality’ took

root and is the basis of all industrial quality management today.

In 1956, Feigenbaum introduced the concept of ‘total quality control’,

saying that it must start with the design of the product and end only when

the product has been placed in the hands of a customer who remains

satisfied. So quality control was extended to research and development of

methods and equipment, purchases, production, audit and acceptance

testing, despatch, installation and service.

At the beginning of the 1960s the Japanese quality revolution marked

the end of the previous era of ‘quality control’ and the beginning of the

new era of ‘quality management’. The 1980s saw the birth of a new

concept, that of ‘total quality’. Attention which had till then been

focused on the product alone was shifted to the customer.

Today quality is read as total quality. The concept of ‘fungibility to use’

is replaced by ‘suitability to needs’ and the focal point is no longer merely

the customer, but man, with all his (and her) physical, social and

economical needs, be they explicit or otherwise.

1.3.2 Total quality

From the producer’s point of view it can be said that total quality is

offering products and services in conformity with customers’ needs.

From a consumer’s point of view it is, however, necessary to distinguish

various types of quality:

1 Expected quality is the customer’s expectations from a particular

category of products.

2 Promised quality is the customer’s expectations from a product of one

particular brand. With respect to expected quality, promised quality

is at a higher level, as the customer, who is initially unaware, gains

awareness during his or her analysis of the offer. With the same

product, the factors that differentiate the quality promised by one

4 Espresso Coffee

producer from the other are the elements of communication such as

design, appeal and brand image, price, sales point, advertising, etc.

3 Effective quality is the quality of the product in question, in other

words, the measurable quality. If the promised quality is with regard

to the form of the product, then here we are talking about pure

substance. Effective quality is, in turn, made up of two components.

The first, which is subjective, refers to the excellence of the

characteristics of products and services. Juran’s definition of

‘fungibility to use’ mainly refers to this component. Examples of

product features relating to subjective quality are robustness,

reliability, flexibility, precision, performance, beauty, goodness,

respect for the environment, lifespan, service content, economy of

use and so on. The second component, which is objective, refers,

according to Crosby’s definition (1979), to how the product

conforms to the requirements specified of it in order to respect

the above characteristics, and therefore to the absence of defects.

4 Perceived quality is the sum of the promised and effective qualities,

and therefore both product and communication contribute to this.

This is the most important quality, as the relationship between this

and expected quality determines customer satisfaction.

Satisfaction . Perceived quality / Expected quality

5 Potential quality is how the product can be further improved.

Consumers assess all these qualities one by one. When they decide to

approach a product, they have, or they create, depending on the

information gathered, a ‘map’ corresponding to the expected quality for

that particular category of products. When they must choose between one

brand or another, they will choose the one they perceive as having the

promised quality which is most in keeping with their needs for use and/or

for price. Finally, once the product has been chosen and purchased, they

will be faced with its effective quality, which is not always in line with the

promised quality.

Psychological mechanisms also intervene in this increase in qualitative

appraisal. For instance, it is more difficult for a consumer to realize that

the effective quality of a product is poor if it has a high promised quality.

This is due to the difference between form and substance, which is not

something man is used to measure, as even in nature beauty represents

good. On the other hand, man finds it difficult to recognize even an

existing highly effective quality in a product with a poor image, and,

indeed, even more difficult if one needs to be an expert to assess the

Quality 5

quality of a product. How many people would be able to recognize a fine

Bordeaux wine served in a bottle without a label? And vice versa, how

many could recognize that a fairly good wine poured from a bottle of

premier cru classe´ is not excellent?

1.3.3 Quality certification

Total quality is part of a large movement begun at the end of the post-war

period in the West, aimed at improving the quality of life. This involved

the introduction of more and more restrictive technical regulations

regarding consumer safety and environmental protection, which gradually

reduced the freedom of the producers of industrial goods. In order to

maintain the free circulation of goods in such a restrictive market, it

became necessary to be able to certify the quality of traded goods

objectively and impartially. Therefore, especially in Europe under the

pressure of the European unification, international legislation changed

and quality certification was introduced.

There are two different types of quality certification: conformity

certification and quality system certification.

Conformity certification means the certification of the conformity of the

product to its declared characteristics. This type of certification can either

be binding, as in the case of potentially dangerous products, or consensual,

where the producers wish to attract attention to the qualitative contents

of their product. The certificate is issued by a third party, often on the

basis of tests carried out on the product in question. System certification

seeks to have an official body certifying the ‘quality assurance system’,

which includes all the structural and organizational factors put in place to

obtain and maintain a particular standard of quality. Reference is made to

the International Organization for Standardization ISO 9000 regulations,

which are recognized nearly all over the world. Its latest revision, ‘Vision

2000’, has increased senior management involvement and has also

introduced the concept of ethics in socio-economical activities.

Conformity certification is direct in that the object certified is the

effective quality of the product. System certification, on the other hand, is

indirect, as it implies that a producer who complies with ISO 9000

regulations is capable of respecting the stipulated qualitative standard. In

both cases, in order to be able to issue certificates, the certifying body

must in turn be accredited by an appointed national authority. These

bodies should actually be delegated by the public authorities, in

compliance with a specific legislation on quality, the so-called ‘Country

Quality System’.

6 Espresso Coffee


Food products are consumed both owing to man’s need to feed, but also

from the search for pleasure provided by certain foods or drinks. In the

case of unfinished products, one further aspect looked for in food products

is the service level, i.e. how easy it is to use or quick to prepare. These

three components of the expected quality for food products are listed


1.4.1 Nutritional quality

The nutritional quality of a food product depends on the nutritional

content of the food product and how safe it is for the consumer’s health.

The nutritional value implies both a quantitative aspect, in terms of the

number of calories provided by the product per unit weight, and a

qualitative aspect given by the composition in nutritive classes, or the

relative percentages of carbohydrates, proteins, lipids, vitamins and

mineral salts, plus other substances, such as fibres. Each nutritive class

can, in turn, be analysed from a qualitative point of view, by taking into

consideration the content of essential substances (such as amino acids and

fatty acids) and its digestibility.

Wholesomeness, or hygienic quality, corresponds, on the other hand, to

the complete absence of toxicity in food products. Leaving aside cases of

uncontaminated food products, which become toxic due to improper use

by the consumer, it can be said that food products only become toxic by

contamination. This contamination may be chemical or bacteriological

and can take place at any time during the life cycle of the product due to

endogenous or exogenous factors.

Generally speaking, nutritional quality is deemed the acceptable

minimum, that is, an indispensable condition for trading a product.

This means that the nutritional content of a particular product must be

consistent with its category and absolute wholesomeness. As the

acceptable minimum, nutritional quality is at a lower level than expected

quality therefore the consumer tends to take it for granted. Or consumers

only take it into consideration when deciding whether or not to try out

products in that particular category. This is why nutritional quality is

normally standardized by law, especially regarding the aspects affecting its


In some cases, however, nutritional quality is a condition for

excellence. This is the case of special products, either of registered origin

or, more generally, the products that, within their category, distinguish

themselves by offering advantages for the consumer.

Quality 7

1.4.2 Sensorial quality

Sensorial quality is a secondary quality, in that it is the effect produced by

certain primary qualities on our sense organs. The primary qualities in

question (organoleptic attributes, according to ISO 5492) are to be

tracked down to those chemical and/or physical properties of food – such

as taste, aroma, texture, aspect (e.g. colour), maybe sound (e.g.

crunchiness) – that impact on consumers’ exteroception, namely on the

senses of gustation, olfaction, haptic (tactile) sensitivity, vision and


Sensorial quality can be defined as the ability of a product to satisfy the

hedonic needs of consumers. It is, therefore, subjective as it is assessed

through an interpretation by each consumer and cannot be measured in

an absolute sense.

Despite this, products that most consumers like are considered quality

products. As a consequence, the only means of defining the sensorial

quality of a product is to bear in mind the tastes and views expressed by

consumers. These opinions are influenced both by the individual

characteristics of the sense organs of each person and each individual’s

ability to use them, and by the customs and traditions of the various

regions in the world influenced by culture, ethnic groups, religion and

social class.

Sensorial quality varies both in time and in space. Moreover, food

products are very complex, which makes it difficult to attribute a

particular sensorial quality to one or more easily analysed constituents. A

way to solve this difficulty is to define accurately the ingredients used and

the processes applied, as this is done with typical or registered products,

and to proceed to the quality control of the finished product by means of

sensorial analysis.

Consumers attach a great deal of importance to sensorial quality,

indeed it can be said that this is what sells a product. This is due to the

fact that, at least in the West, there is no longer a need to appease hunger;

the Western diet is varied and ‘complete’ foods are no longer needed.

Moreover, Western society, which takes care to guarantee consumer

safety, ensures that only products of high nutritional quality are


1.4.3 Service content

The lifespan of the packaged and/or opened product, the cooking time,

the availability, how easy it is to transport, the quantity and quality of

information that accompanies the product, the encumbrance of the

8 Espresso Coffee

packaging to be thrown away, the environmental impact of the packaging

material, packaging safety etc. are all examples of the service content a

product may or may not have.

At times the service content contrasts with other qualities. Just think

of long-life milk, which has an increased lifespan at the expense of a

reduced nutritional and sensorial quality. At other times, however, the

service content strengthens other qualities; one lucky example of this is,

indeed, coffee, which, if pressurized in inert gas, not only lasts three times

as long, but also undergoes a considerable strengthening of its aroma


1.4.4 Conformity certifications

The fact that our society is more mature regarding producer responsibility

in guaranteeing product safety has led to the development of both

horizontal and vertical regulations in the food sector:

1 European Directive 85/374, 25 July 1985 (EC Official Gazette

n. L210 – 7 August 1985) is concerned with the liability for faulty

products (which applies also to all non-food sectors).

2 The cornerstone of the European regulations on the hygiene of food

products is Regulation (EC) n. 852/2004, 29 April 2004 (EC Official

Gazette n. 139 – 30 April 2004), which focuses on prevention by

prescribing obligatory internal audits, assigning responsibility to

manufacturers by imposing the analysis of critical control points –

Hazard Analysis Critical Control Point (HACCP) – and requiring

the drawing up of a manual of correct hygiene practices as a function

of the type of products and processes and the characteristics of the


3 All the vertical regulations regarding disparate food products as a

function of their critical state are coupled up with this horizontal set

of regulations. Genetically Modified Organisms (GMOs), which so

far do not include coffee product, are regulated by numerous

directives and regulations; their presence in a food is among the

situations most critically perceived by consumers.

1.4.5 Quality and the general food law

As a consequence of the recent European Regulation 178/2002/EC

(General food law), 28 January 2002 (EC Official Gazette n. L31 – 1

February 2002), which states the principles and general requirements of

Quality 9

the food law, provides the setting up of the European Authority for Food

Safety (EFSA) and lays down procedures for food safety control, the

European standardization organizations – which refer to the ISO – are

drawing up guidelines on the systems of traceability in food companies.

According to the new models of development based on the sustainability

of the whole food chain, traceability is an instrument of health and

hygiene guarantee, which, besides inspiring trust in the consumer, can

help strengthen the identity and quality of the food product.

In addition to the safeguarding of fundamental health and hygiene

requirements, protected by the regulations in force, more restrictive

qualitative requirements, or requirements of typicality, are proposed

through the guarantee of product certifications. These can refer to

binding regulations (voluntary regulated product certifications), or to

voluntary technical regulations defined by the producer companies with

the control bodies. Organic farming production is an example.

Table 1.1 gives some examples of the validity of voluntary regulated

product certifications laid down within the European Union.

The seal of quality is attributed to farm and food products for which a

particular quality, reputation or other characteristics depend on their

geographical origin, and whose production, transformation and/or

processing take place in a certain area of production. At least one stage

of the production process must therefore take place in a particular area.

As far as wines are concerned, they are regulated by national

regulations in their respective production countries, which, in their

turn, have been harmonized with the EC Council Regulation n. 823/87,

16 March 1987 (EC Official Gazette n. L084 – 27 March 1987 – Pg. 0059-

0068 and subsequent amendments).



Ever since coffee was first consumed in the West, towards the middle of

the seventeenth century, specific local habits and traditions have

developed and overlapped, entailing great differences in coffee consumption.

There is generally a remarkable ritual component, divided into two

very distinct rites, preparation and tasting. Below is a cross-section of

contemporary coffee cultures, which clearly shows the expected and

perceived qualities in the various countries.

n Amsterdam: Essential times for drinking coffee in Holland are in the

morning, at breakfast and around 10 or 11 o’clock, often with guests.

10 Espresso Coffee

It is mostly prepared with a filter machine, even though a growing

number of people now have an espresso machine in their homes.

Much care is taken in the preparation, the presentation and how the

coffee is served. The ‘koffie verkeerd’ is their typical coffee, drunk with

plenty of milk in mugs and often served with sweets, such as apple

pie. Socializing and comfort are two aspects particularly associated

with coffee consumption in Holland and, in public, young people

want to drink coffee ‘in the right way’. Espresso has become the ideal

Quality 11

Table 1.1 Regulations related to quality in the European Union

Product Definition Regulation



A food product for which, throughout

the production cycle, the use of

chemicals (pesticides and fertilizers) is

excluded, and only the use of

environmentally friendly techniques of

cultivation and stock farming is

foreseen. Land is made fertile by crop

rotation and the use of organic manure

and natural minerals, while

environmentally friendly products and

techniques are employed to defend the

crops from parasites. Coffee can be

certified as organic

EC Regulation n. 2092/91

– Official Gazette n. L198

– 22 July 1991 and

subsequent amendments



of origin


Seal of quality attributed to food

products whose particular characteristics

essentially depend on the territory in

which they are produced. The

geographical environment includes

natural and human factors, which make

it possible to obtain a product, which

cannot be imitated outside its defined

area of production.

All stages of production, transformation

and processing must take place in a

delimited geographical area

EC Regulation n. 2081/92

– Official Gazette n. L208

– 24 July 1992 – pp. 1–8

and subsequent






Seal of quality attributed to farm and

food products for which a particular

quality, reputation or other

characteristics depend on their

geographical origin, and whose

production, transformation and/or

processing take place in a certain area

of production. At least one stage of the

production process must therefore take

place in a particular area.

EC Regulation n. 2081/92 –

Official Gazette n. L208 –

24 July 1992 – pp. 1–8 and

subsequent amendments

after-dinner coffee beverage, and possession of an espresso machine

with all its accessories has become a status symbol.

n Hamburg: Coffee in Germany is associated, first of all, with wellbeing

and euphoria, relaxation and fun, and, despite the fact that it is

consumed throughout the day, it is seldom drunk after a meal. The

day starts off with a steaming cup of coffee and a large breakfast. The

careful preparation of the brew takes place in a relaxed atmosphere,

in which the coffee is served on a perfectly laid table, every day of the

year, and drunk in quantity, followed by fresh bread rolls. Throughout

the rest of the day, plenty of coffee is drunk, both at work and out and

about, in the ‘EisKaffe’, places where beverages are served with cakes,

or in the ‘Steh-kaffee’, bars with high counters looking out onto the

street, where one can drink coffee standing up. The traditional coffee

is mostly prepared with a filter, but for a more modern consumer

espresso coffee consumption is growing. Milk is gradually replacing

cream, which traditionally used to be served with coffee. Germans

pay particular attention to coffee being at the right temperature, and

so they tend to keep it handy in a thermos. In the afternoon, around 5

o’clock, it is time for ‘Kaffeeklatsch’, an important moment for

socializing at home, which is the ideal formula for inviting someone

to the home.

n London: The main reason British people decide to drink a cup of

coffee on a weekday is for its stimulating effect to improve their

performances. Taste and aroma are of secondary importance, so much

so that some take a thermos of coffee with them from home. Lack of

time and space have contributed to the spread of extremely

functional methods of preparing coffee, such as instant coffee

prepared with the kettle that is also used for tea, which still has a

strong hold in British homes. At weekends, when people have more

time, habits change and coffee is then drunk with real pleasure,

because it no longer means only ‘getting on with something quickly’.

Some bars have become popular because they offer coffee in real

settings consumers can relate to. Coffee consumption in the UK is

therefore a modern and cosmopolitan experience, most certainly not

a substitute for the traditional rite of tea.

n Naples: Naples probably represents the city with the richest daily

idiosyncratic coffee culture in the world, having elaborated both a

philosophy and a theory on its use and consumption. It is no accident

if all expressive activities that distinguish this city, from the theatre of

de Filippo to the songs of Pino Daniele, have dealt with coffee by

creating and offering particularly original visions and interpretations

of the experience. At home, coffee is prepared with the Neapolitan

12 Espresso Coffee

coffeemaker and served in the ‘tazzulella’. At the coffee bar – an

authentic omnipresent coffee temple (approximately one for every

450 people) – a very concentrated espresso of no more than 20 ml

(called ristretto) is served, prepared by the expert hand of a

Neapolitan barman, operating on an old-fashioned lever-espresso

machine, with sugar mixed with coffee during percolation.

n New York: The first coffee of the morning, generally prepared with

a filter machine or by the infusion system, is drunk in a hurry in

the traditional mugs accompanied by the traditional muffins or

bagels, and seen as a habit, with no special rite. A common scene in

New York is that of coffee drunk in the street in a paper cup, ‘in

transit’, like street fuel, very basic and always available on every

street corner. There is an enormously wide supply of blends for the

home – besides the various brands, you can find pure origin,

flavoured coffee, organic coffee, etc. Coffee is less commonly taken

after main meals, whereas, especially on holidays, people often take

a break during the day to prepare a coffee for themselves or for

friends. There are no fixed rules about occasions for drinking coffee

as there are in other countries. Espresso, stronger tasting than the

traditional American coffee, and the beverages associated with it,

the so-called ESBAD (espresso-based), such as latte and cappuccino,

are consumed more frequently in the coffee shop chains or in

Italian restaurants and are therefore mainly considered ‘special’


n Oslo: The basic Norwegian coffee is what is known as ordinary

coffee, which is black, with no milk or sugar, mostly associated

with home consumption. Preparation with a filter is gradually taking

over from the use of a kettle. Consumption out of the home, which

mainly takes place in coffee bars, is becoming more and more

widespread and this has opened the way for new ways of drinking

coffee, such as espresso and cappuccino. Sometimes these are even

prepared at home, especially when people have guests for the

weekend or in the evening. Consumption is very high – an average

of five cups a day – which makes Norwegians, together with the rest

of the Nordic people, the highest coffee consumers per head in the


n Paris: Coffee is the heart of breakfast; most French people have a real

yearning for it. It is normally prepared using a filter and is drunk with

other typically French products, such as croissants and baguettes.

Another cup is drunk after lunch, without milk. Real coffee lovers

still go to ‘bruˆleries’, if they can find one, where one can enjoy coffee

after choosing the desired strength and aroma. ‘Cafe´’, in its two

Quality 13

accepted meanings, as a beverage and as a place of encounter,

represents an important social and collective rite that reached its

climax, thanks to the likes of Voltaire, Balzac and all the way to

Sartre. In bars and restaurants, especially in the north of the country,

coffee is generally prepared with an espresso machine, although it is

not served as concentrated as in Italy.

n San Francisco: Their first desire in this city is to start the day off with

the rite of a good cup of coffee, that is, to begin the day with an

experience of quality where its careful preparation is something to be

proud of. There are roasted coffee dealers, shops and bars that serve

coffee almost anywhere in town, and they are proud of roasting their

own coffee. San Francisco is an important centre for coffee in the

United States and played a leading role in the 1980s in the revolution

of gourmet coffee. The wide availability of different types of coffee

tempts people to try out all kinds of novelties. Milk and espressobased

beverages are like meals, so at lunch lighter drip coffee is

preferred, while ‘fun’ coffees are consumed after lunch, as a dessert.

The bars are like an extension of the sitting room at home where you

can drink a coffee as you relax, reading the newspaper or surfing on

the Internet.

n Tel Aviv: In the morning, instant coffee, quick and without much

ceremony, is a necessary provision for the working day ahead; no

great interest is taken in preparing the coffee and certainly no time is

dedicated to this. At other times of the day drinking coffee is a way of

spending time together. People in the city love sitting for hours in

roomy, high-tech cafes, where they can go to see and be seen,

especially on Friday, savouring coffee at length with little sips, often

accompanied by dry cakes, biscuits or fruits.

n Tokyo: The pace of life is very fast and people do not have much

time for themselves. For this reason, most people drink instant coffee,

of which there is a wide choice. The most widespread coffee drunk,

especially out-of-home, is American coffee. The coffee is consumed

in bars or from vending machines, mostly as an energizing drink. The

vending machines, to be found everywhere, even in city streets and

on country roads, sell various kinds of soft drinks, but above all, cans

of liquid coffee, for which Japan is at the top of the world market.

Coffee is on the menu in Western-style restaurants, which, together

with the coffee shop chains, are the main places where coffee is

drunk, especially among coffee lovers. At home, this beverage is

indeed little used, people almost only drinking instant coffee once in

the morning. Espresso coffee is very successful, even if it is thought by

many to be too bitter.

14 Espresso Coffee


Three coffee experiences can be lived throughout the day:

1 coffee for waking up is still preferred as a hot, dilute beverage;

2 espresso, as has been seen, takes the lion’s share of coffees at breaks,

especially in Latin countries and above all in Italy;

3 in Anglo-Saxon or northern European countries, espresso is

considered a speciality suitable for relaxing, where feeling and

care are involved in its preparation.

The most appreciated characteristics of espresso are its creaminess, body

and the strength of its aroma, associated with stimulating properties,

despite the fact that espresso contains less caffeine than the more dilute

coffees do (see Chapter 8). Espresso therefore represents a benchmark

universally recognized as being a great pleasure and a symbol of Italian

culture, so much so that the highest expression of coffee is that served in

bars in Italy, in restaurants in France, and restaurants and coffee shops in

the rest of the world, precisely because it is coffee prepared with a

professional machine.

As regards the nutritional content (see Chapter 10), the only

expectations one can have from the consumption of an espresso are the

intake of a moderate dose of caffeine, known above all as a stimulating

pleasure, boosting intellectual activity, improving memory and concentration,

quickening reflexes, making it easier to stay awake, improving

one’s mood, etc., and the fact of consuming a product devoid per se of


The real consumer’s expectations, however, are of a hedonistic nature

(see Chapter 10), because drinking a good espresso is genuinely wonderful.

Espresso is a true elixir, a concentrate of exquisite aromas lasting long after

it has been drunk. Even sight and touch are satisfied, thanks to the striped

hazelnut colour of the foam and to its full body. In moments of relaxation at

home this tasting rite is preceded by the eagerly awaited rite of preparation,

when it is enjoyable to play out the actions of the barman, enriching it

with your own secrets, as you anticipate the pleasure of its taste.

As far as service is concerned, consumers require the roasted and

ground coffee to retain the fragrance of the aromas developed during

roasting until the time the package is opened, and they want to know the

characteristics of the product they are consuming (see Chapter 6). It

should also go a long way, in that only a small amount is needed to make a

good cup of coffee and, once open, the aroma should last until the

contents of the package run out.

Quality 15

Grinding deserves a whole chapter (see Chapter 5). Connoisseurs

believe that in order to be excellent, the beans should only be ground

immediately prior to preparation; unfortunately, it is difficult to find and

take care of top quality household grinders. This is why 90% of packaged

coffee sold in retail shops is ground. The most promising solution for the

future is servings – ground and pressed doses of coffee, sealed in a wrapper

and ready for use. Servings offer the fastest, cleanest way of preparing

espressos, and, most importantly, they guarantee a consistency of quality

that neither coffee beans nor ground coffee can in any way guarantee due

to the serious problems in keeping all the parameters involved in

preparing a perfect espresso cup under control.

Espresso consumption is an aesthetic experience, like tasting a vintage

wine or admiring a painting. It is a search for beauty and goodness for

improving the quality of our life. As it offers such subjectively ineffable

‘goodness’, devoid of defects, the only adequate reaction to it is

astonishment – astonishment that can give birth to enthusiasm, and

therefore intellectual and spiritual enrichment.

The predominance of the experience aspect means that the official

definition of quality may be too limited for espresso. It may be more

appropriate to speak of ‘degree of excellence’. The elements characterizing

effective quality are the subject of this book and will be revealed in detail

as you read. From an organoleptic point of view, we have already seen the

importance of the aroma and the full-body – to a certain extent

represented by the visual component of the foam – while, as regards

the taste, consumers look for a slight bitterness in southern countries or a

slight acidity in northern countries, in both cases accompanied by the

characteristic sweetness of the coffee. On the opposite side, the most

common, serious defects penalizing consumption are the extreme

bitterness and foul flavour of poor quality beans.


Everyone in Italy has a clear mental picture of a cup of espresso: a small

heavy china cup with a capacity just over 50 ml, half full with a dark brew

topped by a thick layer of a reddish-brown foam of tiny bubbles, also

known by the Italian term ‘crema’. More than 50 million cups of espresso

are consumed every day in the world: its fragrance and flavour are the first

stimuli in the morning, they crown an excellent meal later in the day, and

act as frequent revivers during lengthy working sessions.

16 Espresso Coffee

1.7.1 Espresso as a lifestyle: brewed on the spur of

the moment

One of the meanings of the word espresso (express) is that it is made for a

special purpose, on the moment, on order (Marzullo, 1965; Hazon, 1981);

therefore it is made for the occasion on express request, extemporaneously

rather than fast. This concept is clarified by the saying ‘the consumer, not

the espresso must wait’. As a direct consequence, once brewed, espresso

cannot be kept and must be drunk immediately, before the foam shrinks

and collapses breaking into patches on the surface. After a while, the

surface of the liquid is completely free from foam, which has dried out on

the walls of the cup above the liquid.

If an espresso is kept waiting, smoothness of taste is lost and perceived

acidity increases with time regardless of cooling. Furthermore, if the cup

cools down, an unbalanced saltiness becomes noticeable.

n Freshness of preparation must be an integral part of the definition of

this very special brew.

1.7.2 Espresso as a brewing technique: it requires


At the beginning of the twentieth century the need for preparing a cup of

coffee within seconds of a customer’s request led to an increase in the

pressure of the extraction water. Water was heated up to its boiling point

in a sealed kettle, so that the steam in equilibrium created pressure,

accelerating extraction. A drawback of this technique was that brewing

with boiling water provokes over-extraction of astringent and bitter,

usually less soluble, substances, which give a burnt taste to the brew.

Brewing was first improved by separating the water used for brewing,

best hot but not boiling, from the heating water. Pressures as high as 10

bars could be created by a lever, multiplying the force of the arm of the

bartender, producing a thick layer of foam on the cup. The lever has now

been replaced by an electric pump, simpler and more regular to operate.

A pressure field applied within a fluid produces potential energy – what

is known as Bernoulli’s piezometric energy – which can be easily

transformed into kinetic energy, and further transformed into surface

potential energy and heat.

Quality 17

Pressure is important for the definition of espresso, making it different

from other brews. During espresso percolation (see Chapter 7), a small

amount of hot water under pressure is applied to a layer of ground roasted

coffee, the coffee cake, and this very efficiently produces a concentrated

brew, containing not only soluble solids, but also lipophilic substances,

lacking in filter and instant coffees. The foam on the top and the opaque

brew are unique to espresso, owing to the presence of a disperse phase

formed by very small oil droplets in emulsion (Petracco, 1989) (see 8.1.1),

which are perceived in the mouth as a special creamy sensation, the body.

Furthermore, the oil droplets preserve many volatile aromatic components,

which would otherwise either escape into the atmosphere or be

destroyed by contact with water as in other brewing techniques, so that

the rich coffee taste lingers in the mouth for several minutes. If coffee

were percolated under high static pressure only, the pressure would be lost

downstream and no work could be performed on the cake; while, if kinetic

energy from stirring propellers, choke nozzles, sprayers, etc., were applied

downstream from the cake, a smooth layer of foam could be produced, but

it would lack body.

The Latin etymology of the word espresso, literally meaning pressed out

(Campanini and Carboni, 1993), clearly points out the importance of

pressure in espresso brewing, making the technique an integral part of the


n Espresso is a brew obtained by percolation of hot water under pressure

through tamped/compacted roasted ground coffee, where the energy

of the water pressure is spent within the cake.

1.7.3 Italian espresso: it must be rapidly brewed

Another important feature of espresso, especially as traditionally drunk in

Italy, is the length of percolation (see 7.5.8). The diversified energy input

in espresso pressure-brewing efficiently brings into the cup both

hydrophilic and lipophilic substances. A best mix is reached within 30

seconds; if the extraction is shorter than 15 seconds a weak and

exceedingly acid unbalanced and under-extracted cup is obtained.

Conversely, if extraction lasts longer than 30 seconds, over-extraction

of substances with poor flavour will produce an ordinary harsh-tasting

cup, as can be easily seen by separately tasting the liquid fraction

percolated after the prescribed 30 seconds.

A quantitative definition can now be given:

18 Espresso Coffee

Italian espresso is a small cup of concentrated brew prepared on

request by extraction of ground roasted coffee beans, with hot

water under pressure for a defined short time.

The range of the parameters is:

Ground coffee portion 6.5   1.5 g

Water temperature 90   5



Inlet water pressure 9   2 bar

Percolation time 30   5 seconds

The requisite conditions to make a good cup of espresso will be reviewed

in detail in the following chapters.


We have seen how varied and complex the concept of quality is and how

difficult it is to use it as an objective measure, particularly in the case of

the espresso. We have, however, understood that the espresso lives almost

solely on the pleasure it gives consumers. Therefore, if the number of

lovers is to be maintained and increased, it is necessary to seek the means

of continually improving its quality, meaning its degree of excellence. In

the following chapters we will try to outline this road to improvement,

setting out everything that is known so far about the quality of the


This becomes a stage on the road to improving the quality of the

production of coffee brewed worldwide, which, as the resounding success

of gourmet coffee has shown in the Anglo-Saxon world, can lead to a

significant increase in consumption. This may result in an important

contribution to the rebalancing of the supply and demand of coffee, thus

improving the precarious financial and social situation the producer

countries find themselves in.

Wine has travelled a similar road very successfully and, besides

providing pleasure to consumers, has achieved a significant rise in the

value of overall production. This has resulted in a segmentation of the

market where no one in the world is any longer surprised by the fact that

the price of a bottle can range over several orders of magnitude.

Consequently, there has been a general increase in the well-being and

satisfaction of those along the whole chain. This road was not easy and

involved an almost ‘manic’ search for excellence, the specialization of the

distribution networks and education of the consumers.

Quality 19

Espresso can play the lead role in a similar situation in the coffee field,

thanks to the image it enjoys as a symbolic beverage and thanks to the

commitment of the whole industry to continuous improvement.


Campanini G. and Carboni G. (1993) Vocabolario latino-italiano. Milan:

Paravia, p. 524.

Crosby P. (1979) Quality Is Free. New York: McGraw–Hill.

Deming W.E. (1982) Out of Crisis. Cambridge, MA: MIT, Center for

Advanced Engineering Study.

Feigenbaum A.V. (1964) Total Quality Control. New York: McGraw–Hill.

Harrington H.J. (1990) Il costo della non-qualita`. Milan: Editoriale Itaca.

Hazon M. (1981) Grande dizionario italiano–inglese e inglese–italiano.

Milan: Garzanti, p. 1348.

Imai M.K. (1986) The Key to Japan’s Competitive Success. New York: McGraw–


ISO (1992) Sensory Analysis – Vocabulary. ISO 5492:1992. Geneva:

International Organization for Standardization.

ISO (2000) Quality Management Systems, Principles and Terminology. ISO

9000:2000. Geneva: International Organization for Standardization.

Juran J.M. (1951) Quality Control Handbook. New York: McGraw–Hill.

Marzullo A. (1965) Dizionario della lingua italiana. Milan: Fratelli Fabbri, p. 417.

Petracco M. (1989) ‘Physico-chemical characterisation of espresso coffee

brew’, Proc. 13th ASIC Coll., pp. 246–261.

Taguchi G. (1987) ‘The system of experimental design: engineering methods

to optimize quality and minimize cost’. Unpublished, White Plains, New


Webster’s (1984) New Riverside University Dictionary.

20 Espresso Coffee


The plant

F. Anzueto, T.W. Baumann, G. Graziosi,

C.R. Piccin, M.R. So¨ ndahl and

H.A.M. van der Vossen


M.R. So¨ ndahl and H.A.M. van der Vossen

2.1.1 Origin and geographic distribution

Commercial coffee production is based on two plant species, Coffea

arabica L. (arabica coffee) and C. canephora Pierre ex Froehn. (robusta

coffee). A third species, C. liberica Bull ex Hiern (liberica or excelsa

coffee) contributes less than 1% to world coffee production. All species

within the genus Coffea are of tropical African origin (Bridson and

Verdcourt, 1988). Many forms of wild C. canephora can be found in the

equatorial lowland forests from Guinea to Uganda, but natural populations

of C. arabica are restricted to the montane forests of southwestern

Ethiopia (Berthaud and Charrier, 1988).

The date of first domestication of arabica coffee in Ethiopia is uncertain

and so is the claim that Ethiopian invaders brought coffee to Arabia in

the sixth century. However, there is written evidence for extensive

cultivation of coffee in the Arabian Peninsula (Yemen) by the twelfth

century. During the following 300 years, the stimulating hot beverage

prepared from roasted and ground coffee beans, called qawha (a general

word for wine and other stimulants) by the Arabs and cahveh by the Turks,

became immensely popular in the Islamic world and from 1600 onwards

also in Europe, with Mocha as the exclusive centre of coffee trade

(Pendergrast, 1999). In Uganda and other central African countries, there

has been a centuries-old tradition of growing coffee (mainly C. canephora)

near homesteads for the purpose of chewing the dried fruits or beans for

their stimulating effect (Wrigley, 1988).

Commercial production of arabica coffee outside Yemen started in Sri

Lanka (formerly Ceylon) in the 1660s, and subsequently in Java around

1700 with coffee plants introduced by the Dutch East India Company

from Sri Lanka or the Malabar Coast of southwestern India. The arabica

coffee in India and Sri Lanka originated probably from seeds brought

directly from Yemen by Moslem pilgrims during the first decade of the

seventeenth century. A few plants taken from Java to the Botanic Garden

in Amsterdam in 1706 formed the basis of practically all the cultivars of

arabica coffee planted in the Western Hemisphere. These early

introductions of coffee plants from Yemen to Asia and Latin America

are of the variety Typica, C. arabica var. arabica. Coffee plants taken by

the French around 1715 from Yemen to the Re´union island (formerly

Bourbon) and subsequently to Latin America and East Africa are C.

arabica var. bourbon. Varieties of the Bourbon type have generally a more

compact and upright growth habit than Typica varieties, are higher

yielding and produce coffee of better quality (Carvalho, 1988).

By 1860 world coffee trade involved some 4 million bags (60 kg, the

standard unit of trade), mostly from Brazil, Indonesia and Sri Lanka. On

account of its mild and aromatic cup quality, C. arabica would have

certainly continued to be the only cultivated species, had it not been so

vulnerable to diseases, particularly to coffee leaf rust (Hemileia vastatrix).

Coffee leaf rust (CLR) was first reported in 1869 in Sri Lanka and within

20 years it had virtually wiped out coffee cultivation in Asia. Sri Lanka

switched to tea cultivation, but Indonesia has continued to be a major

coffee producer based on C. canephora, which turned out to have

resistance to CLR. Robusta was first introduced into Indonesia (Java)

from Congo in 1900 (via Belgium) and a few decades later, high-yielding

plant materials had been developed by selection, which became known as

robusta coffee and have until today remained the basis of most robusta

coffee production in the world.

Coffee cultivation is now widespread in tropical and subtropical regions,

with the bulk of arabica coffee concentrated in Latin America and robusta

coffee predominant in South-East Asia and Africa (Figure 2.1).

2.1.2 Production

World coffee production increased from 86 million bags in 1980 to 112

million bags in 2000/01 (from 10.5 million ha). However, production

graphs show large annual fluctuations, from a very low 80 million bags in

1986/87 to a record crop of some 115 million bags in 1999/2000, mainly

caused by the leading producer Brazil (ICO, 2002). The Brazilian coffee

22 Espresso Coffee

The plant 23

Figure 2.1 Coffee producing regions of the world

crop may vary from 19% to 33% of annual world production as result of

biennial fluctuations, recurrent frosts and droughts. The tremendous

expansion of robusta coffee cultivation during the past decade, in

Vietnam in particular, but also in Indonesia and Brazil, has further

exacerbated the excess in supply over demand on the world coffee market.

Consequently, the share of arabicas in the world coffee supply has also

dropped from the erstwhile 75% to 68–70%. Production data for the 10

largest coffee-producing countries are presented in Table 2.1. About 59%

of the world’s coffee in 2000/01 was produced in Latin America, 21% in

Africa and 20% in Asia. Brazil (80% arabica) alone produced 26%,

Vietnam (97% robusta) 12%, and Colombia (100% arabica) 11% of the

2000/01 coffee crop. Coffee is produced in some 60 countries, including

the coffee-exporting members of the International Coffee Organization

(40 at 30 September 2001: ITC, 2002, p. 21). In most countries,

smallholders are the main coffee producers, but medium to large coffee

plantations (30–3000 ha) can be found in countries like Brazil, El

Salvador, Guatemala, India, Indonesia, Vietnam, Kenya and Ivory Coast.

2.1.3 Taxonomy

The genus Coffea belongs to the family Rubiaceae, subfamily Ixoroideae and

tribe Coffeae. Linnaeus described the first coffee species (Coffea arabica)

24 Espresso Coffee

Table 2.1 Coffee production in the 10 largest

producing countries and world total (millions of

60 kg bags)


Year 2000/01 Annual average

Arabica Robusta Total 1996–2001

Brazil 23.7 6.4 30.1 28.9

Vietnam 0.4 14.8 15.2 8.6

Colombia 10.5   10.5 11.1

Indonesia 0.7 5.7 6.4 7.1

Mexico 5.1 0.1 5.2 5.5

India 2.4 2.6 5.0 4.2

Ivory Coast 4.0 4.0 3.9

Guatemala 4.7 0.05 4.7 4.7

Ethiopia 2.9   2.9 3.2

Uganda 0.4 2.8 3.2 3.3

Total (60 countries) 68.8 42.9 111.7 106.6

Sources: ICO, 2002; ITC, 2002

in 1753. The taxonomy of coffee has been under recent review and may

not yet be completely finalized yet. Bridson and Verdcourt (1988)

have grouped coffee species in two genera: Coffea (subgenera Coffea

and Baracoffea) and Psilanthus (subgenera Psilanthus and Afrocoffea).

The subgenus Coffea includes 80 species, 25 from continental Africa and

55 unique to Madagascar and the Mascarene Islands. The subgenus

Baracoffea was assigned only seven species (Bridson, 1994). Conventional

and molecular genetic studies confirm that five clusters of related species

can be distinguished corresponding to geographical regions (West,

Central and East Africa, Ethiopia and Madagascar), but that the process

of differentiation between these groups has not yet progressed into a stage

of strong crossing barriers (Berthaud and Charrier, 1988; Lashermes et al.,

1997). The earlier taxonomic classifications into sections and subsections

(Chevalier, 1947) or subgenera (Leroy, 1980) do not appear to be justified

any longer. Instead, a monophyletic origin of all species within the genus

Coffea has been generally accepted. Molecular evidence (Lashermes et al.,

1997; Cros et al., 1998) for fairly close genetic relations between Coffea

and the genus Psilanthus (Bridson and Verdcourt, 1988), as well as the

successful hybridization between C. arabica and a Psilanthus species

(Couturon et al., 1998) may lead to a taxonomic revision into one genus

Coffea with more than 100 species.

All Coffea species are diploid (2n . 22 chromosomes) except for the

allotetraploid C. arabica (2n . 44), which probably originated as a

spontaneous interspecific cross between two diploid species, followed by

duplication (Carvalho, 1946; Sondahl and Sharp, 1979). Recent

cytogenetic and molecular studies have indicated that indeed C.

eugenioides and C. canephora (or C. congensis) are the most likely

progenitors of C. arabica (Raina et al., 1998; Lashermes et al., 1999).

The analysis of segregating molecular markers (Lashermes et al., 2000a)

has confirmed earlier genetic and cytogenetic evidence that C. arabica is a

functional diploid. The observed disomic meiotic behaviour is probably

under the control of genes regulating pairing of chromosomes.

2.1.4 Growth, flowering and fruiting

The coffee plant is an evergreen shrub or small tree, which under free

growth may become 4–6m tall for C. arabica and 8–12m for C. canephora.

In cultivation, both species are pruned to manageable heights of less than

2m (less than 3m in mechanically harvested plantations in Brazil) with

one or more stems. The strictly dimorphic branching habit of coffee, as is

schematically presented in Figure 2.2, determines plant growth habit and

The plant 25

has important consequences for agronomic practices, such as pruning for

maintenance or rejuvenation and vegetative propagation (Cannell,

1985). Continuously growing orthotropic (vertical) stems produce a

pair of plagiotropic (horizontal) branches at each new node from the

‘head of series’ buds. The other buds on the same node remain dormant

but will grow out into orthotropic shoots called suckers, when forced by

capping of the main stem. Young seedlings produce the first pair of

plagiotropic branches on the ninth to eleventh node. Branches bear two

opposite leaves per node, with glossy dark green upper surfaces, 10–15  

5–10 cm in size, and slightly undulating in arabica. In robusta plants,

leaves are much larger and strongly corrugated. The serial buds present in

the leaf axils on plagiotropic branches may develop into inflorescences or

secondary lateral branches, but never into orthotropic shoots. In

consequence of this strictly dimorphic growth habit in coffee, only

cuttings prepared from orthotropic shoots (suckers) will, after rooting,

grow into trees with a normal upright growth habit.

At the end of the rain period, differentiation of flower buds takes place,

and this phenomenon is frequently associated with lower temperatures

and shorter daylengths. Soon after differentiation, flower buds will grow

26 Espresso Coffee

Figure 2.2 Dimorphic branching habit of coffee (Cannell, 1985, p. 113)

for about 8 weeks until reaching 5mm in length, and then become

quiescent for about 2 months. At the end of this period, they will

complete their differentiation process to form a glomerule (short

inflorescence) bearing a cluster of 2–19 flower buds in arabica

(Carvalho et al., 1969) and up to 80 buds in robusta. Later, flower buds

resume growth until they reach 8–10mm in length, and then enter into a

second quiescent period. At this point, they are quite visible, with a paleyellow

coloration. Arrested flower buds will grow again when exposed to

first showers of the new rainy period. Flower buds are then released from

dormancy, undergo a rapid expansion during 8–10 days, changing colour

from pale-yellow to white. This last growth phase will end with the

opening of the flowers (anthesis), resulting in massive flushes of

blossoming (Alvim, 1973; Cannell, 1985). Depending on the prevailing

dry season, two to five blossoming events will occur per flowering season,

leading to sequential ripening events and differentiated pickings. The

longer the dry season, the more synchronized flowering and reduced

number of blossoming events.

In equatorial regions with bimodal rainfall patterns, flowerings occur at

the beginning of both rainy seasons, and this will result in overlapping

stages of fruit development and fruit ripening extending over several

months. In unimodal rainfall regions further from the equator, there is a

single flowering season, and a much shorter harvesting period.

The white and fragrant bisexual flowers have a corolla tube 8–10mm in

length, with five lobes (five to seven in C. canephora). The five stamens

are fused to the corolla and the bilocular anthers open lengthwise. The

pistil has an inferior, bilocular ovary and a filiform style with two

stigmatic branches (Van der Vossen et al., 2000) (Figure 2.3).

Flowers wither within 2–3 days after anthesis and pollination. C. arabica

is self-fertile and less than 10%of the flowers are naturally cross-pollinated,

mainly by visiting bees (Carvalho, 1988). On the other hand C. canephora

is strictly allogamous with a gametophytic system of self-incompatibility

preventing self-pollination (Berthaud, 1980; Lashermes et al., 1996a).

Pollen tubes take about 24 hours to reach the ovary followed by

fertilization, but the first cell division of the zygote takes place not until

60–70 days later (Dedecca, 1957). Three stages of fruit development can be

distinguished (Cannell, 1985). After an initial period of dormancy with

little growth during the first 6–8 weeks after flowering (pinhead stage)

fruitlets expand in volume almost to the final size over a period of 8–10

weeks (soft-green stage) followed by a period of bean (endosperm) filling for

another 10–15 weeks (hard-green stage). Fruits mature in 7–9 (C. arabica)

to 9–11 months (C. canephora) from date of flowering. The green fruits will

turn yellow or red, depending on the variety, during ripening (Figure 2.4).

The plant 27

The fruit is a drupe, usually called a berry or cherry, containing two

seeds embedded in a fleshy pericarp (fruit skin) and a sweet tasting

mucilage layer. Sometimes the fruit contains only one round seed called

a peaberry. Fruits of C. arabica are 12–18mm and those of C. canephora

8–16mm long. The seeds (coffee beans) are plano-convex in shape and

grooved (centre cut) on the flat side. They consist mostly of endosperm

with a small embryo at the base of the seed and are enveloped in a

rudimentary integument (‘silverskin’) and fibrous endocarp (parchment).

Beans of C. canephora are generally smaller, rounder and with a tighter

centre cut than those of C. arabica.

28 Espresso Coffee

Figure 2.3 Coffee flowering

Figure 2.4 Green, ripe and overripe cherries of C. arabica: (left) yellow variety;

(right) red variety

Coffee fruits are strong assimilate-accepting sinks and the tree is unable

to regulate the crop load effectively by shedding fruits. The key issue in

coffee growing is therefore the prevention of excessive cropping that leads

to biennial bearing and also shoot and root dieback. Crop husbandry is,

therefore, aimed at maintaining enough foliage to sustain the crop (about

20 cm˙ leaf area per fruit) as well as new shoot growth by pruning and other

agronomic practices (Cannell, 1985). Copper and some other types of

fungicides improve leaf retention, particularly in ecosystems with

pronounced periods of water stress. This ‘tonic’ effect of fungicides may

result in considerably higher yields, irrespective of disease control per se

(Van der Vossen and Browning, 1978).


M.R. So¨ ndahl and H.A.M. van der Vossen

2.2.1 Breeding strategies

Breeding programmes implemented to the two coffee species have

basically the same main objective of developing new cultivars with the

potential of yielding optimum economic returns to coffee growers. Yield,

plant vigour and quality are the main selection criteria, but particularly in

arabica disease resistance and compact growth habit have also been given

much attention. Variation in climate and soil, incidence of diseases and

pests, cropping systems, socio-economic factors, market dynamics and

consumer preferences further define priorities given to selection criteria

applied in specific programmes. Breeding methods depend primarily on

the mating system and plant improvement goals (Van der Vossen, 1985,

2001; Charrier and Eskes, 1997). For the self-pollinating C. arabica the

methods applied include, in increasing order of complexity (Carvalho,


n introduction of coffee germplasm and pure line selection;

n pedigree selection after intervarietal crossing and often also backcrossing;

n F1 hybrids between genetically divergent breeding lines;

n interspecific hybridization followed by backcrossing to arabica lines

and pedigree selection.

In case of the out-crossing C. canephora, these are (Montagnon et al.,

1998 a,b):

The plant 29

n introduction of coffee germplasm and mass selection;

n family and clonal selection;

n reciprocal recurrent selection;

n interspecific hybridization followed by family and clonal selection.

2.2.2 Traditional and modern cultivars Arabica

World coffee production is still based to a large extent on traditional

cultivars. In C. arabica cultivars were developed long ago by line selection

within the Typica and Bourbon varieties, or in offspring of crosses between

these two types. Besides, cultivars like Caturra, Pacas, San Ramon, Sumatra

and Maragogipe have their origin as single-gene mutants found in C.

arabica populations (Krug and Carvalho, 1951). The traditional cultivars

are often renowned for their excellent cup quality, but most are very

susceptible to the major coffee diseases, which makes them increasingly

obsolete for economic (costs of chemical disease control) and ecological

(chemical pollution) reasons in many coffee regions (Table 2.2).

However, as a result of considerable breeding efforts during the past 50

years, new arabica cultivars with higher yield potential and resistance to

important diseases have now started to replace traditional varieties in

several countries. These include Catimor and Sarchimor type of cultivars

in Colombia, Brazil, Central American countries and India, Icatu in

Brazil, as well as the F1 hybrid cultivars Ruiru II, generally referred as

Ruiru 11 in Kenya, and Ababuna in Ethiopia. Some of the recently

developed cultivars of arabica coffee are described in Table 2.3.

Efforts to obtain resistance to CLR, already started in India around 1920,

have had a long history of initial successes followed by disappointments

because of repeated appearance of new virulent races of the rust fungus.

However, some lines of the cultivar Catimor, which was developed from

crosses between Caturra and Hibrido de Timor (a natural interspecific

hybrid between C. arabica and C. canephora with arabica phenotype), have

shown complete resistance to all physiological races of the CLR pathogen

in most countries. These results were obtained by breeding plans normally

applied to self-pollinating crops, including recombination crosses followed

by backcrossing, inbreeding and pedigree selection (Carvalho, 1988;

Bettencourt and Rodrigues, 1988). The breeding programme in Kenya

demonstrated the advantages of F1 hybrid cultivars also for arabica coffee,

especially the simultaneous combination of compact plant type, high yield,

good quality and resistance to coffee berry disease (CBD) and CLR in one

cultivar (Van der Vossen and Walyaro, 1981). Subsequently, breeding

30 Espresso Coffee

Table 2.2 Commercially important traditional arabica cultivars (disease


Name Country Description

Typica Worldwide The original type of arabica coffee introduced from Yemen into

Asia in the early sixteenth century and after 1720 into the

Caribbean and Latin America

Bourbon Worldwide Introduced around 1715 from Yemen into Re´ union Island

(formerly Bourbon) and subsequently into Latin America and East

Africa; more compact and upright growth, higher yielding and

better bean and cup quality than Typica; red and yellow fruited

Bourbon types

Tekesik C. America High-yielding selection from Red Bourbon made in El Salvador

Kona Hawaii Selection from Typica with large beans, good quality, but low


Mundo Novo Brazil Selection from a natural cross between Sumatra (Typica) and red

Bourbon found in Sao~ Paulo in 1931; replacing much of Typica

after 1960; vigorous growth and high yielding



Jamaica A Typica selection grown in the Blue Mountains and famous for

its cup quality; susceptible to CLR, moderately resistant to CBD

Kent India Selection from Typica in 1911; some resistance to CLR (race II);

good cup quality; still grown at high altitudes

K7 Kenya Selection from Kent in 1936; some resistance to CLR (race II)

and CBD; vigorous growth and high yielding, but lower cup

quality than SL28

SL28 Kenya Selection from Bourbon ‘Tanganyika Drought Resistant II’ in

1935; high yielding and some drought resistance; excellent bean

and cup quality

SL34 Kenya Selected from ‘French Mission’ (Bourbon-type); good yield and

excellent bean size and liquor quality; adapted to highest


KP423 Tanzania Selection from Kent similar in yield and quality to K7 in Kenya

N39 Tanzania Selection from Bourbon-type coffee; excellent bean and cup

quality, but moderate yielder

Jimma Ethiopia Dry-processed forest and garden arabica coffees of many origins

and very variable quality

Harar, Gimbi Ethiopia Distinct varieties grown in Hararge and Welega provinces resp.;

dry-processed coffees highly appreciated for their ‘mocha’


Yirga Chefe,


Ethiopia Washed coffees of very good quality; names refer to place of

origin (Sidamo and Kefa provinces resp.) rather than being

distinct cultivars

Caturra Worldwide Mutant discovered in Bourbon field in Brazil in 1935; compact

growth (short internodes); medium-sized beans and moderate

cup quality

Catuai Brazil Selection from a cross (Caturra x Mundo Novo) made in 1949;

compact growth; in Brazil higher yielding and better quality than


Villa Sarchi Costa Rica Mutant with short internodes found in Bourbon coffee; syn.

Villa Lobos; similar to Caturra in plant stature and yield, but

smaller beans

Pacas El Salvador Caturra-like mutant found in Red Bourbon coffee in 1909

CLR, coffee leaf rust.

32 Espresso Coffee

Table 2.3 Commercially important modern arabica cultivars

(cultivars with disease resistance)

Name Country Description

Catimor Worldwide Selections from crosses (Caturra/Catuai x Hibrido de

Timor); compact growth like Caturra; resistant to

CLR; Oeiras (Brazil), Cauvery (India), IHCAFE90 and

CR95 (C. America) are Catimor-like cultivars

Sarchimor Worldwide Selections from cross (Villa Sarchi x H. de Timor);

similar to Catimor in compact growth and resistance

to CLR; Tupi, Obata and IAPAR59 (Brazil) are

Sarchimor-like cultivars

Colombia Colombia Synthetic variety composed of a number of Catimor

lines; large bean and good liquor quality; resistant to

CLR; some lines with resistance to CBD

S795 India Selection from a natural interspecific hybrid (C.

arabica x C. liberica) backcrossed to C. arabica;

resistant to some CLR races; high yield and good

liquor quality; most important arabica cultivar in


Ruiru II (Ruiru 11) Kenya F1 hybrids (seed by hand-pollination) between

selected Catimor lines and selected clones from

multiple crosses (tall plants); resistant to CBD and

CLR; slightly less compact than Catimor; early and

high yielding; most F1 combinations have very good

bean size and cup quality

Ababuna Ethiopia F1 hybrid (seed by hand-pollination) between

selected lines of Ethiopian germplasm with normal

(tall) plant habit; resistance to CBD

Icatu Brazil Selections developed by crossing (tetraploid) C.

canephora x C. arabica (Bourbon) followed by

backcrossing to Mundo Novo; resistance to CLR; tall

growth habit; susceptible to drought and cold; high

yielding and good cup quality. IAC3282 is an early

maturing Icatu with yellow berries

S2828 India Developed by interspecific crossing and backcrossing

similar to Icatu; tall plants; resistance to all (?) races

of CLR; high yielding and good cup quality

1. CLR, coffee leaf rust; CBD, coffee berry disease.

2. Important progenitors for disease resistance:

Hibrido de Timor: arabica-type variety cultivated in Timor; assumed to have developed from a

natural cross between C. arabica and C. canephora; CIFC (Portugal) clones 832/1, 832/2 and

1343 have resistance to all or most races of CLR; CIFC clone 1343 also carries a gene for

resistance to CBD. According to Bertrand et al. (In Press) it should be possible to find lines with

resistance genes and good beverage quality. Selection can avoid accompanying the

introgression of resistance genes with a drop in beverage quality (Bertrand et al., 2004).

Rume Sudan: semi-wild arabica variety collected in 1942 from the Boma Plateau in

southeastern Sudan; some individual plants have very high resistance to CBD.

The plant 33

strategies aiming at F1 hybrid cultivars in arabica coffee have been adopted

elsewhere, e.g. in Ethiopia (Bellachew, 1997) and in Central American

countries (Bertrand et al., 1997).

Selection for bean size and cup quality has received much attention in

arabica coffee breeding programmes in Kenya and Colombia in particular,

because the quality of new disease-resistant cultivars should be at least

equal to that of the traditional cultivars in order to uphold the country’s

high reputation and special position in the world coffee market. This was

obviously achieved in Kenya with the CBD- and CLR-resistant hybrid

cultivar Ruiru 11 (Njoroge et al., 1990) and in Colombia with the

CLR-resistant cultivar Colombia (Moreno et al., 1995), as judged by

international coffee-tasting panels. Rigorous standardization of pre- and

post-harvest practices, bean grading and cup tasting applied in these two

breeding programmes contributed to increased selection progress and

helped to overcome the initially negative effects on quality due to introgression

of disease resistances from exotic germoplasm (cupping characteristics

are, however, different from those of the parent material and

still considered by trade less suitable for espresso-type blends). Bean size

improvement is particularly noticeable in the Pacamara variety (cross

between Pacas   Maragogype), developed in El Salvador, with screen size

bean values of 19 andmire. Robusta

In case of the cross-pollinating C. canephora, most cultivars are also

propagated by seed, often from polyclonal seed gardens. Commercially

important robusta cultivars include the BP and SA series in Indonesia,

S274 and BR series in India, the IF series in the Ivory Coast and the

cultivar Apoata in Brazil (Table 2.4). Some breeding programmes in

robusta coffee (e.g. in the Ivory Coast) have recently adopted methods of

reciprocal recurrent selection with distinct sub-populations to increase

chances of producing genotypes superior in yield, quality and other

important traits (Leroy et al., 1997b). Interspecific hybrids

Interspecific hybridization has played a significant role in coffee, such as

crosses between arabica and robusta coffee with the objective of

introgressing disease resistance into arabica, e.g. the cultivars Icatu in

Brazil (Carvalho, 1988) and S2828 in India (Srinivasan et al., 2000), or

improved liquor quality into robusta coffee, such as the variety Arabusta

in the Ivory Coast (Capot, 1972). Other examples of interspecific

hybridization leading to successful robusta cultivars are Congusta in

Indonesia and the CxR variety in India.


M.R. So¨ ndahl, F. Anzueto, C.R. Piccin and H.A.M. van

der Vossen

2.3.1 Climate and soil

Arabica coffee is cultivated at medium to high altitudes (1000–2100m

a.s.l.) in equatorial regions, or at 400–1200m altitudes further from the

equator (9–24


N and S latitudes), where average daily temperatures are

about 18–22


C. In general, the cooler the climate the better will be the

cup quality of arabica coffees. However, temperatures near 0


C will kill

the leaves immediately. On the other hand, long periods of hot weather

will cause abnormal flowering (star flowers) and shoot dieback.

Depending on local evapotranspiration, it requires a minimum of 1200–

1500mm annual rainfall with no more than 3 months of less than 70mm

for good growth and production. In contrast, robusta coffee requires warm

and humid climates of tropical lowlands and foothills (100–1000m

altitude) with average daily temperatures of 22–26


C, absolute minimum

34 Espresso Coffee

Table 2.4 Some robusta cultivars

Name Country Description

BP and SA


Indonesia Important robusta cultivars developed on two research

stations in Java in the 1920s; propagated by seed or


S274, BR


India Major robusta cultivars in India; selected from plant material

of Java origin and released in the 1950s; seed propagated

IF series Ivory


Selected from robusta plant material from Java and the DR

Congo; propagated by seed and cuttings



Brazil Named after the Kioulou river in the DR Congo; the main

robusta variety (95%) in Brazil; seed propagated

Apoata Brazil Initially used as nematode-resistant rootstock for arabica

cultivars; now also used as robusta cultivar; seed


Arabusta Ivory


Developed from a (C. canephora x C. arabica) cross in the

1960s: cup quality better than robusta: limited

commercialization due to persistent low fertility; clonal


C x R


India Developed from a (C. canephora x C. congensis) cross in

1942, followed by backcrossing to S274 and selection;

released in 1976; bolder beans and better cup quality than


temperatures not below 10


C and well-distributed annual rainfall of at

least 2000 mm.

The parent material of major coffee soils include lava and tuff (e.g.

Kenya), volcanic ash (e.g. Indonesia, Central America), basalt and

granite (e.g. Brazil, West Africa, India). Soils for coffee cultivation should

be deep (at least 2 m), free-draining loams with a good water-holding

capacity, fertile and slightly acid (pH 5–6) and contain at least 2%

organic matter in the upper horizon.

2.3.2 Propagation and crop husbandry Propagation and planting

Cultivars of arabica and robusta coffee are generally propagated by seed.

Seeds germinate within 4–5 weeks after sowing in wet sand. Removal of

parchment prior to sowing reduces coffee seed germination time by half.

Vegetative propagation by rooted cuttings or grafting has found limited

application in the past to multiply high-yielding robusta clones (e.g., in

Java and Ivory Coast). Grafting of arabica cultivars on robusta rootstock is

very common in certain areas of Brazil and Central America where

nematodes are a serious problem. Recently, more efficient in-vitro

methods of clonal propagation by somatic embryogenesis have been

developed for robusta as well as arabica coffees (Sondahl and Baumann,

2001; Etienne et al., 2002).

Seedlings or clonal plants are raised in shaded nurseries, in polythene

bags filled with a soil-compost mixture, for 6–12 months before

transplanting to the field at the start of the rainy season. Seedlings are

planted in previously dug holes filled with top soil supplemented with

organic and inorganic fertilizers, or alternatively in trenches dug along the

rows by a tractor-mounted implement and refilled with a soil–compost

mixture (Mitchell, 1988). Under optimum field conditions young plants

may flower 12–15 months after planting, producing a first light crop 2.5

years after planting. Conventional plant densities vary from 1300–2600

plants/ha for tall-stature arabica varieties to 1100–1400 plants/ha for

robusta coffee. Higher densities of 3333–10 000 plants/ha are applied with

compact growing arabica cultivars like Caturra, Catuai and Catimor in

Latin America or Ruiru 11 in Kenya. Shade

In South America and East Africa coffee is mostly grown in pure stands

without permanent overstorey shade, except for widely spaced rows

The plant 35

of shade trees acting as windbreaks. In other countries (e.g. Central

America, Mexico, Cameroon, Indonesia, India), coffee is grown either

with temporary or permanent shade trees, or in association with perennial

crops (coconut palms, rubber, clove, fruit trees), or in home gardens

associated with food crops, bananas (e.g. northern Tanzania, Uganda) and

tree crops. In India and Indonesia the stems of shade trees are often

utilized to support the vines of black pepper, and so provide an additional

source of revenue to coffee growers. With intensive crop management

including high inputs (fertilizers and pest/disease control), unshaded

coffee will usually produce higher yields than coffee grown under shade.

Without high inputs or under sub-optimal ecological conditions coffee

usually produces better results under shade. Shade trees have multiple

functions, such as: reducing extreme temperatures, breaking the force of

monsoon rain and winds, controlling erosion on steep slopes, reducing the

incidence of certain pests (e.g. white stem borer), and generally

preventing overbearing and shoot dieback.

Common shade trees are Inga spp. and Grevillea robusta (Latin

America), Albizzia spp. (Africa), Erythrina indica (India) and Leucaena

sp. and Casuarina sp. (Indonesia and Papua New Guinea). In Central

America the density of shade trees varies from 156–204 trees/ha at lower

altitude to 83–100 trees/ha at higher elevations. As a general recommendation,

overstorey shade should not reduce more than 50% of total

irradiances. Pruning

Pruning is essential in coffee production to achieve the desired plant

shape, to maximize the amount of new wood for the next season crop, to

maintain a correct balance between leaf area and fruits, to prevent

overbearing and thus reduce biennial production, stimulation of root

growth, enhancement of light penetration, and also to facilitate disease

and pest control. The following main pruning systems can be distinguished

(Wrigley, 1988):

n single- or multiple-stem, capped at about 1.8m height, accompanied

by regular pruning of lateral branches and removal of suckers;

n multiple-stem with three to four uncapped stems with little

maintenance pruning; rotational system of replacement of old stems

(after three to four cropping years) by new orthotropic shoots;

n ‘agobiado’, which is a multiple-stem system on one main stem that

has been bent over at an early age;

36 Espresso Coffee

n complete stumping to 30 cm above ground level (with or without a

temporary lung branch) to encourage re-growth of suckers, which,

after thinning to two to three vertical shoots, will be brought up to

form the vertical stems for the next cycle of production.

This last pruning system is applied on a rotational basis every fourth or

fifth year in closely spaced arabica coffee blocks, usually planted with

compact-type cultivars (e.g. Caturra, Catuai), or to rejuvenate an old

block of conventionally spaced coffee after 8–12 years of production. Weed control

Weeds should be controlled, as they compete with the coffee trees for

moisture and nutrients. Broad-leaved weeds are generally less harmful

than perennial grasses (e.g. couch, star grass and lalang) and sedges (e.g.

nut- and water-grass), because the roots of the latter also produce exudates

toxic to coffee and they are more difficult to eradicate. Tillage, herbicides,

mulching or interplanting with cover crops can suppress weeds. Tillage

should be done carefully to prevent damage of the superficial feeder roots

of the coffee and also to avoid soil erosion. Herbicide applications in

combination with zero cultivation has the added advantages of not

disturbing the superficial feeder roots of the coffee trees, which can

therefore develop in the layer of decomposing organic matter from dead

weeds and so increase uptake of nutrients. Mulching produces significant

yield responses in low rainfall areas, such as in the arabica coffee regions

of East Africa and Brazil. The interplanting of leguminous cover crops is

less common with arabica than with robusta coffee, even where rainfall is

not a limiting factor. Arabica coffee is more sensitive to competition from

weeds and cover crops. During the first three years after planting, when

the young coffee trees occupy a small portion of the soil surface, betweenrow

interplanting with annual food (e.g. beans, maize, rice) and

commercial (e.g. pineapple, cassava) crops serves as an alternative

manner of controlling weeds, while producing also an additional source

of food and income (Mitchell, 1988). Fertilization

Fertilizer requirements depend on yield level and natural soil fertility.

Nutrients are removed by harvested fruits, but additional nutrients are

also required for sustaining the vegetative growth. Part of the total

nutrients is recycled back to soil by leaf fall, prunings and decaying feeder

roots. The annual nutrient uptake per hectare of a vigorously growing

The plant 37

block of arabica coffee yielding 1.0 t/ha green coffee beans is estimated at

135 kg N, 34 kg P2O5 and 145 kg K2O. Soil and leaf analyses are

important for monitoring the balance of nutrients in soil and plant tissues,

and so provide good guidance for fertilizer application. Generally,

nitrogen fertilizers at rates of 50–400 kg N/ha/year give considerable

increases in yield, especially when applied in split applications. Responses

to potassium fertilizers (up to 400 kg K2O/ha/year) vary from nil in

mulched coffee grown on volcanic soils rich in K (e.g. Kenya) to highly

significant in soils with a low K status (e.g. Brazil). Very high K

applications may induce Mg deficiency. Phosphate is often applied as

compound fertilizer (NPK: 2–1–2), but its effect is faster in foliar

applications. Calcium in the form of lime is used to correct soil acidity.

Deficiencies in Mg and minor nutrients such as boron and manganese are

also best corrected by foliar applications. Organic manures (stable

manure, cover crops, mulch and decaying coffee pulp) are alternatives

to chemical fertilizers and are often the only type available to

smallholders. Organic matter also improves physical (water retention

and aeration) and biological (microflora and soil fauna) properties of the

soil (Willson, 1985; Mitchell, 1988). Irrigation

Irrigation is essential in nurseries and during the first 1–2 years after field

planting to enable the young plants to establish well and to survive the

dry seasons. Irrigation of unshaded adult coffee can produce economic

yield responses, when annual rainfall is less than 1200mm and the dry

season(s) are extended over more than 2–3 months. Critical periods when

irrigation should be applied are at flowering time, at berry expansion stage

and at bean filling stage. Especially moisture stress at the latter (hard

green berry) stage will result in light and poor quality beans (Mitchell,

1988). However, irrigation is only applied on large estates, as it is

expensive, e.g. overhead sprinkler or drip irrigation systems. The latter is

most efficient in water use, but requires a very costly initial investment of

equipment. Organic and sustainable coffee

In several coffee-producing countries there is increasing awareness of the

need to produce sustainable and environment-friendly coffee. This is

achieved by improved land management, increased organic instead of

synthetic fertilizers, reduced pesticide use by integrated disease and pest

management (disease resistant cultivars and biological control of insect

38 Espresso Coffee

pests) and by reduced water use and pollution during post-harvest

handling. Some of these coffees may be certified as organically or

sustainably produced coffees and so attract premium prices in a gradually

developing niche market of health- and environment-conscious consumers,

particularly in the USA and Europe.

2.3.3 Diseases and pests Diseases

Coffee leaf rust (CLR), caused by the fungus Hemileia vastatrix, is the most

important disease of arabica. It reached Brazil in 1970, and is now present

worldwide in all coffee-producing countries (except Hawaii and

Australia). The use of copper and systemic fungicides and resistant

cultivars has reduced the CLR threat. Nevertheless, CLR continues to

cause considerable economic damage to arabica on a global scale due to

increased production costs by fungicide sprays (10%) and crop losses (20%

and more), because most coffee is still produced on CLR-susceptible

cultivars. There is also a problem in some countries, India in particular, of

repeated breakdowns of host resistance on CLR-resistant cultivars by the

development of new pathogenic races of the leaf rust fungus (Eskes,


The very destructive coffee berry disease (CBD) is still restricted to

arabica in Africa, although climatic conditions in certain high-altitude

coffee areas of Latin America and Asia appear to be favourable to the

fungus Colletotrichum kahawae (coffeanum). CBD epidemics can quickly

destroy 50–80% of the developing crop on susceptible arabica cultivars

during prolonged wet and cool weather conditions. Preventive control by

frequent fungicide sprays may account for 30–40% of total production

costs. New cultivars with high levels of durable CBD resistance have been

successfully developed (e.g. in Kenya), but have not yet sufficiently

replaced old susceptible cultivars (Van der Vossen, 1997). Most robusta

cultivars are resistant to both diseases.

Of the several other fungal and bacterial diseases affecting both coffee

species (Wrigley, 1988), the following four have drawn attention in

recent years. In Central America the leaf spot disease ‘Ojo de Gallo’

(Mycena citricolor) can be a serious problem in coffee grown under shade at

high altitudes, particularly affecting CLR-resistant cultivars like Catimor

and Sarchimor. Black rot (Koleroga noxia) is now the second most

important coffee disease in India after coffee leaf rust (Bhat et al., 1995).

The bacterium Xylella fastidiosa has become a major problem in some

regions in Brazil (Beretta et al., 1996). Fusarium wilt disease or

The plant 39

tracheomycosis (Fusarium xylarioides) has been seriously affecting C.

canephora in DR Congo and Uganda since the 1980s (Flood and Brayford,

1997). The traditional Nganda types are very susceptible, but some

robusta clones appear to be resistant to Fusarium wilt. Nematodes and insect pests

Nematodes (mainly Meloidogyne spp. and Pratylenchus spp.) can cause

considerable problems on arabica in Central America in particular, but

also in Brazil, India and Indonesia. Much progress has been made in

developing arabica cultivars with resistance to certain nematode species

(Anzueto et al., 2001) as well as in utilizing nematode-resistant robusta

clones as rootstock for arabica cultivars (Bertrand et al., 2000).

Among the most damaging and worldwide coffee pests are the coffee

berry borer (Hypothenemus hampei), stem borers (e.g. Xylotrechus quadripes),

scale insects (e.g. green scale, Coccus viridis) and leaf miners (e.g.

Perileucoptera coffeella). Host resistance to insect pests in natural coffee

germplasm is extremely rare, but transgenic coffee plants with resistance

to leaf miners and the coffee berry borer may become available in the near

future (Guerreiro Filho et al., 1998; Leroy et al., 2000). Integrated pest

management (IPM) has proved to be much more effective in reducing

damage to the coffee trees and crops than frequent applications of

insecticides (Bardner, 1985). IPM includes early warning systems

(monitoring of insect populations) in combination with cultural (pruning

and shading), biological (insect traps; introducing insects or microorganisms

parasitic to the pest) and chemical (carefully timed applications

with non-persistent insecticides) methods of control.

2.3.4 Harvesting and yields

In equatorial regions where the harvesting season extends over several

months, selective picking of ripe berries at 7–10 days intervals is common.

Where the harvest season is short, e.g. in Brazil, or the cost of hired labour

high, whole branches are stripped when the majority of the berries are

ripe. Harvesting costs are two to three times higher for selective hand

picking, but quality of the product is usually better (but, see also 3.2.4).

The use of self-propelled mechanical harvesters has found application in

plantation coffee of Brazil and a few other countries (see 3.2.3).

Yields may vary from 200 kg green coffee beans per ha from low-input

smallholder plots to 2 t/ha for arabica and 3.5 t/ha for robusta coffees in

well-managed plantations at conventional plant density and without

shade. Yields of 5 t/ha or higher have been obtained in some closely

40 Espresso Coffee

spaced and unshaded coffee blocks planted with compact-type arabica

cultivars, e.g. in Brazil, Colombia and Kenya.

2.3.5 Agronomic practices in selected countries

There is a great variation in production systems among the coffeeproducing

countries. Comprehensive overviews of all or most important

countries have been presented by Krug and de Poerck (1968), Wrigley

(1988) and the ITC report (2002). The four examples presented here are

Brazil, as the largest producer of natural arabica and leader in

mechanization of coffee cultivation, the Central American region, as

largest producer of washed arabicas, Kenya in East Africa, with a tradition

of producing mild arabicas of the highest quality, and India, as leader in

technology development in arabica and robusta coffees in Asia. See

Chapter 3 for details of the production techniques mentioned below. Brazil

Coffee is grown on some 2.5 million ha (5.5 billion trees) between

latitudes 10 and 24


S, mostly on gently sloping land, often with high-tech

cultivation practices. Brazil is the largest arabica producer in the world,

and its coffee plantations are composed of 73% arabica and 27% robusta

trees. Its share of robusta in the total annual crop has increased to more

than 21% in recent years (ITC, 2002). Domestic consumption accounted

for about 50% of total production of 30.1 million bags in 2000/01. In

recent years new coffee plantings have moved further north up to the states

of Bahia (arabica) and Rondonia (robusta). Recently, incidental severe

droughts have caused more impact on the Brazilian coffee crop than frost.

It is estimated that about 85% of all planted arabica coffee consists of

traditional cultivars, Mundo Novo and Catuai in particular, and the rest

are CLR-resistant cultivars like Tupi, Iapar-95, Icatu and others.

Coffee plantations in Brazil vary from relatively small (5–50 ha) and

medium-sized farms (50–200 ha) to very mechanized large plantations

(200–3000 ha). Plant densities have evolved from 1000–1900 plants/ha,

often with two plants per hole with free growth, to 3500–5700 plants/ha

in hedgerow planting with periodic pruning. The availability of more

compact cultivars has led to the adoption of even higher plant densities in

combination with regular pruning.

Preparation for planting is often entirely mechanized. Tractor-mounted

planting implements have been developed that first draw deep furrows

along the planting lines, then mix the soil with fertilizers and lime and

close the furrows all in one single passage. The bulk of planting material

The plant 41

utilized in Brazil consists of 6-month old seedlings produced in small

plastic bags, but more recently the ‘tubette’ seedlings are becoming

popular since they allow for semi-mechanized planting methods with

consequently 40% reduction on planting costs. The tubette seedlings are

of particular use for arabica cultivars grafted on nematode-resistant

robusta rootstock.

Almost all the coffee in Brazil is grown without shade. Weeds are

controlled with a combination of pre-emergent herbicides and intercropping

with native leguminous plants and monocots. A green cover of

spontaneous species is maintained by mechanized trimming and diluted

solutions of herbicides plus liquid fertilizer. Plantations under full sun

produce more, and so require higher levels of fertilizers. Recommended

fertilizer rates for each 10 bags of green coffee are (in kg/ha): N (100),

P2O5 (30), K2O (120), Mg (20), S (5), B (2) and Mn (3). Fertilization is

split in three to four applications and monitored by leaf analysis: if N is

above 3.2% and K above 3.1%, the remaining N–K dosages are

suspended. Coffee leaf rust is being effectively controlled in susceptible

cultivars by three to four rounds of sprays with copper-based and systemic

fungicides during the main growing season. Coffee berry borer, leaf miners

and in some areas also nematodes are the main coffee pests.

Yields vary from 0.6 t/ha green coffee in low-input coffee farms to 2–3

t/ha in well-maintained plantations. Very high annual yields of 4–7 t/ha

have been reported from technology-driven coffee plantations established

in the last decade in northwest Bahia with short-stature cultivars.

The method of pruning both vertical stems and lateral branches is

becoming very popular among farmers, and special machines have been

developed for these operations.

Harvest season begins in mid-April (at low-altitude farms) and extends

until August (in high lands). Stripping all cherries from branches at once is

the common method of harvesting. To maximize the percentage of ripe

cherries during the single harvesting round, it is common to grow coffee

cultivars with early, medium and late fruit ripening characteristics.

Harvesting is still predominantly a manual operation, accounting for 35–

40% of production costs. Brazil is the world leader in the testing and

development of mechanical implements to reduce costs of coffee harvesting

(see Figure 3.3). Equipment includes hand-held vibrating machines

similar to olive harvesters, tractor-pulled harvesters equipped with one or

two cylinders of vibrating rods to harvest one or both sides of a row of coffee

trees. More advanced harvesting machines include the self-propelled

models equipped with fruit collecting and cleaning systems. Harvesting is

accomplished by driving over the coffee plants and the berries are placed in

bags or directly loaded onto trucks in bulk. These vibrating harvesting

42 Espresso Coffee

machines usually operate at speeds between 0.5–1.0 km/hour. Selfpropelled

grape harvesters from Europe with a system of oscillating arms

are also being tested for ‘selective picking’ of ripe coffee berries. They can

operate at three to four times the speed of current models, but still require

some adaptations to make them suitable for coffee harvesting.

In Brazil, arabica coffee is traditionally harvested by the stripping

method and dry-processed to produce ‘natural’ coffees. Fruits from

stripping harvest include several ripening stages (green, green-ripe, ripe,

overripe and dry). By water flotation, dry and overripe fruits are isolated

for sun drying in separated lots. Certain farmers depulp ripe cherries and

dry the parchment without removing the mucilage. This semi-natural

preparation method is increasing in acceptance, since it reduces the total

patio drying time and offers reduced risk of undesirable fermentation.

Presently, it is estimated that Brazil produces about 3.5 million bags of

pulped natural coffee (see 3.3.3). The Brazilian arabica coffee is

characterized by mild acidity, aromatic notes and heavy body, but

qualitative variations do exist due to different microclimates, varieties

and processing methods (see 3.8 for details on sensorial analysis). Limited

amounts of Brazilian coffee are also processed by the fully wet method. Central America

Central America (Guatemala, El Salvador, Honduras, Nicaragua and

Costa Rica) produced altogether some 13 million bags of washed arabica

and only very little robusta coffee in 2000/01 (ITC, 2002). Coffee is

grown in mountainous regions with volcanic soils at altitudes ranging

from 400–1700 m, but the majority of fields are within 1200–1500m

altitude. In general, the rainy season extends from April/May to October,

but there is a tremendous variation in total annual rainfall (1000–

4500 mm) between regions due to topographical differences. Cultivation

under shade is common in more than 95% of the coffee areas in Central

America, except in Costa Rica, which has 25% of full-sun plantings. The

most common shade trees are Inga spp., Erythrina spp. and Grevillea spp.,

at densities ranging from 100 to 156 trees/ha. While in Costa Rica,

Nicaragua and Honduras most of the coffee is produced by smallholders,

medium to large coffee plantations have an important share of the total

crop in Guatemala and El Salvador.

The main cultivars grown at present are Caturra, Pacas, Catuai and Red

Bourbon. These have largely replaced the traditional Typica or ‘Arabigo’

variety, which was introduced from the Caribbean in the eighteenth

century. The cultivation of compact cultivars like Caturra has stimulated

important technological changes, such as high plant densities, reduced

The plant 43

shading and inputs of fertilizers and pesticides, increasing productivity.

Actually, three coffee production systems can be distinguished in Central


n technical, with compact cultivars (Caturra, Catuai) grown at high

plant densities (5000–7000 plants/ha);

n technical, with tall cultivars (Bourbon lines) at close spacing (3400–

3700 plants/ha);

n traditional, with tall Bourbon cultivars at conventional spacing and

low inputs.

In Costa Rica, the technical farming system with Caturra and Catuai is

most common, whereas in Guatemala and El Salvador, Honduras and

Nicaragua one can see all three systems of cultivation. The adoption rate

of compact growing CLR-resistant cultivars like Catimor and Sarchimor is

still rather modest (about 5%). National average yields of the Central

American countries vary from 800 to 1500 kg/ha of green coffee.

Technical-operation coffee farms can reach up to 3000 kg/ha, whereas

low-input traditional farms produce about 500 kg/ha.

Modern cultivation practices include periodic pruning by rows or by

blocs, while in traditional farming systems selective pruning is often

applied, i.e. plant-by-plant as the need arises. It is also common to bend

orthotropic stems (agobio) to induce multiple vertical branches forming a

typical plant architecture called ‘parra’. The amount and type of fertilizers

applied varies with region and level of production, but is usually given in

NPK dosages of 50–75, 100–150 and 150–300 kg/ha, in two to three split

applications during the rainy season. The economically most important

diseases are coffee leaf rust and ‘Ojo de Gallo’, whereas coffee berry borer

and nematodes are the main pests. In highly affected nematode regions in

Guatemala the control is achieved by grafting arabica cultivars on

resistant robusta rootstock.

Harvesting in low-altitude coffee farms takes place in the period

August–October, at mid-altitude during October–December and at high

altitudes, January–April. Ripe berries are picked in three to five passages

and all coffee is processed by the wet method. Mucilage is mainly removed

by fermentation, but some farms are adopting mechanical removal of

the mucilage after depulping. On-farm wet processing is common in

Guatemala, Honduras and Nicaragua, but in Costa Rica and El Salvador

most harvested cherry is processed in central coffee factories. Wet

parchment coffee is dried on patios under the sun, but mechanical dryers

are used for final drying in mid- to large processing factories. In Central

America green coffee is commercialized by growing altitude and, some-

44 Espresso Coffee

times, location in relation to the Pacific or Atlantic Ocean. For instance,

in Guatemala the type ‘strictly hard bean’ originates from altitudes above

1350 m. In Costa Rica ‘strictly hard bean North’ coffee comes from

altitudes of 1200–1600 m. In Honduras and El Salvador ‘strictly high

grown’ coffee is derived from altitudes above 1200 m. Kenya

Arabica coffee is grown in the highlands east and west of the Great Rift

Valley, just below and above the equator at 1450 to 2000m altitudes.

Kenya produced 1.1 million bags of mild arabica coffee from 150 000 ha in

2000/01, which is almost entirely exported (local consumption 4.5%).

About 60% of total Kenya coffee is produced by smallholders on 75% of

the coffee land area, and the remaining by medium to large plantations.

Coffee production in Kenya increased from 560 000 bags (30 000 ha) in

1963/64 to a peak of 2.2 million bags in 1987/88, but has been declining

since that time due to economic and socio-political factors.

The coffee flowers in March/April and again in November, under

influence of the bimodal rainfall pattern typical of equatorial climates.

Consequently, the harvest season extends over several months, with the

main crop usually maturing in October–December and an early crop in

June–August. Average total rainfall per year in the coffee zone varies from

900mm at low to 2000mm at high altitudes, but the East African climate

is notorious for its recurring cycles of wet and dry years. Larger plantations

are situated in the lower range of the coffee zone, and the water deficit is

compensated by (overhead sprinkler) irrigation to sustain economic yield

levels. Smallholders in the upper altitudes benefit from higher rainfall, but

with no irrigation facilities available their coffee suffers in years of severe

droughts. Coffee is grown without shade trees, except in very high altitude

areas for the protection against wind and extremely low temperatures.

Much of the existing coffee on old plantations is a direct descendant of

the Bourbon-type ‘French Mission’ or ‘Mocha’ coffee plot established at

the St Austin’s Catholic Mission near Nairobi around 1900 with seeds

originating from Yemen. The planting material of farms established after

1950, and all smallholder coffee fields, consists mostly of the cultivars

SL34 and SL28. These two varieties are derived from single plant

selection from French Mission and other Bourbon coffees, respectively,

and growing at high and medium altitudes. The variety K7 is a selection

from the Indian cultivar Kent, and is cultivated in the lowest coffee zones.

The cup quality of SL34, SL28 and also of French Mission is significantly

better than that of K7, but the latter became popular because it has some

resistance to CLR, which is the predominant coffee disease at lower

The plant 45

altitudes. Blue Mountain was recommended in the 1950s to smallholders

in Kissii, a high-altitude and wet coffee zone west of the Great Rift Valley,

on account of its lower susceptibility to CBD than SL34 or SL28, but its

cup quality is rather disappointing according to Kenyan standards. In

1986, the Kenyan Coffee Research Foundation (CRF) released Ruiru 11, a

seed-propagated (F1 hybrid seeds) compact growing and productive

arabica cultivar with high resistance to CBD and also to CLR. Under

good management Ruiru 11 is capable of producing a cup quality similar

to cultivars SL34 or SL28; it is, however, not looked for in top espresso

coffee blends. An estimated 20 000 ha have been planted with Ruiru 11.

Coffee is conventionally planted at a density of about 1320 trees/ha and

maintained as three-stemmed capped (plantations) or free-growth

(smallholders) trees. Plantations with irrigation equipment have often

converted to 2640 trees/ha by doubling within-row density. The diseaseresistant

and compact Ruiru 11 is best planted at 3000–5000 trees/ha,

maintained on one or two uncapped stems per mature tree and after three

full crops clean-stumped to start a new cycle of production. Mulching

(grasses, etc.) is general practice on larger plantations to control weeds

and conserve moisture, but smallholders are often unable to follow

recommended practices because the little crop residues available on the

farm may be required to feed livestock. During recent years of low coffee

prices, regular fertilizer applications are mainly restricted to coffee

plantations, as smallholders lack the financial resources for external

inputs. This also applies to chemical control of diseases (in CBD- and

CLR-susceptible coffee cultivars) and pests. Yields of small coffee farms

have declined considerably and are now often not more than 300–400

kg/ha. Well-maintained large coffee plantations produce 1–2.5 t/ha per

year. Yields of 3 t/ha and more have been obtained in high-density

coffee blocks with Ruiru 11.

All Kenyan coffee is harvested by several rounds of selective picking of

ripe berries and processed according to the wet method (see 3.3.2). A

small fraction, consisting of overripe and black cherry, is dry-processed

and sold as ‘mbuni’ on the local trade. Plantations operate their own

coffee factories, while smallholders bring harvested cherry to cooperative

factories for depulping, fermenting, washing and sun drying the parchment

coffee, usually done on raised tables with wire-mesh tops.

Some factories, mainly on large plantations, also have mechanical dryers

(see 3.4.3) to speed up the drying process during wet weather. Coffee

processing stations are situated near rivers, as large quantities of water are

required for the wet process. Regulations are now in place to prevent coffee

pulp and dirty wastewater from polluting the rivers. Dry parchment is

delivered to one of the four existing coffee mills for curing, which includes

46 Espresso Coffee

hulling, cleaning and grading by size and density. Cup quality of each lot of

green coffee is then determined by liquoring before being offered for sale at

the weekly auction in Nairobi. There are seven grades of green coffee

(from AA to T), and cup classification goes from excellent mild arabica

with high notes of pointed acidity, rich flavour and light body (class 1) and

FAQ (class 4) to a low cup type (class 7). In addition, there are another

three under-grades, including ‘mbuni’, to cup class 10. A rare Kenyan AA

coffee of class 1–2 is unmatched in quality on the European coffee market,

except occasionally by mild arabica coffee (Bourbon cultivar N39) from

the higher slopes of Mount Kilimanjaro in Tanzania. India

Coffee is cultivated mostly in the mountain ranges in the states of

Karnataka, Kerala and Tamil Nadu in southern India, between latitudes 9


and 14


N, robusta at 400–1000m and arabica areas at 700–1600m

altitudes. Small pockets of coffee cultivation also exist in east Andhra

Pradesh, Orissa (18


N) and seven northeastern states (27


N). Arabica

and robusta coffees are generally grown under shade. The upper tier

consists of natural forest trees (Ficus, Artocarpus, Terminalia and other

species) or of planted Grevillea robusta, whereas the lower tier has Erythrina

indica and orange trees. Coffee flowering occurs in March/April at the start

of the main rainy season. Arabica is harvested during the period November

to February and robusta from January to March. Harvesting is entirely

manual, by picking ripe berries in several harvesting rounds.

Coffee production increased from 1.1 million bags (73% arabica) from

92 000 ha in 1960/61 to almost 5 million bags (48% arabica) from

310 000 ha in 2000/01. The national average yield is, therefore, almost

1.0 t green coffee/ha. Robusta coffee generally produces 10–15% more

than arabica plants and large plantations yield about 25% more than

small farms. Smallholders cover about two-third of the coffee area and

contribute with 60% of the total annual coffee crop, and the remaining

40% comes from plantations. Arabica and robusta coffees are processed

according to the wet (washed) or dry methods, and subsequently cleaned

and graded in coffee curing centres. An estimated 80% of all arabica and

30% of the robusta coffees are washed. Coffee produced in large

plantations is practically all processed by the wet method. About 80%

of all the coffee produced in India is exported.

Resistance to CLR has remained the most important objective for

arabica coffee breeding in India since the release of cultivar Kent in the

1920s. The humid and relatively warm climate appears very favourable to

the pathogen and Kent (resistance to CLR race II) was soon attacked by

The plant 47

CLR. There is no other coffee-producing country in the world where CLR

has developed so many new physiological races, the latest ones apparently

capable of overcoming host resistance of some of the most advanced

cultivars. Kent is still grown at very high altitude, where CLR incidence is

relatively low, but almost 70% of all arabica grown in India at present is

S795, a cultivar released in 1946 with resistance to CLR races II and I only,

but high yielding and producing good quality Indian arabica coffee. CLR

can be controlled fairly effectively by prophylactic sprays (Bordeaux

mixture 0.5%) and by systemic fungicides (Plantvax 0.03% or Bayleton

0.02%), but that adds considerably to production costs. Even Cauvery, a

cultivar selected from Catimor (ex CIFC Portugal), showed CLR infection

within a few years after its first release in 1985 (Srinivasan et al., 2000).

There are now about 22 000 ha planted with Cauvery and it is performing

well at higher altitudes, where leaf rust epidemics are less severe. Among

the other arabica cultivars released by the Central Coffee Research

Institute (CCRI) near Chikmagalur, S2828 is most promising for its complete

resistance to all CLR races in combination with good yield, bean size

and cup quality. This cultivar was developed from an interspecific cross

between S274 (robusta) and Kent made in 1934, followed by two backcrosses

to Kent and intensive pedigree selection over several generations.

The first robusta coffee planted in India was in the Wayanad district of

Kerala State around 1905 and much of the present robusta plantings is

directly derived from those early introductions. Family selection for yield

and bean size resulted in a number of selections, of which eventually S274

and the BR series were released in the 1950s. S274 in particular has

become an important robusta cultivar. Probably inspired by favourable

reports from Java of a spontaneous hybrid between C. congensis and C.

canephora named ‘congusta’, breeders at the CCRI made interspecific

crosses between C. congensis and S274 around 1942. After backcrossing to

S274 and a number of generations of sib-mating and selection, the new

cultivar C x R was first released to some estates in 1973–6 in Kerala, but

further selection for uniformity is being pursued. The C x R cultivar has

characteristics very similar to those reported earlier for Congusta on Java,

such as bolder beans and better cup quality, somewhat lower caffeine

content, more compact growth habit, good yield and adaptation to cooler

climates. It appears to have the possibility of considerably improving the

quality of robusta coffees produced in India. Some 10 000 ha have been

planted so far, new planting as well as conversion of old plantations by

top-working. There is also in development a compact growing version of

the C x R variety, analogous to Caturra in arabica.

In general, the capped single-stem pruning system is recommended in

India for both species. Plant densities applied are 3000 plants/ha for tall

48 Espresso Coffee

arabica cultivars, 4400 plants/ha for Catimor arabica and 1100 plants/ha

for robusta plantations. Regular applications of chemical fertilizers are

common practice in large plantations to support economic yield levels

complemented with farmyard manure or composted organic wastes (e.g.

coffee pulp). Weed control is accomplished by manual slashing or

herbicides. Recently, the Indian coffee sector has started actively to

promote organic coffee farming, in which no chemical fertilizers or

herbicides are used, and with minimum use of pesticides.

Black rot is considered the second most important disease after CLR

and this disease affects arabica as well as robusta. So far, no sources of

resistance have been identified and control depends on a number of

cultural measures (e.g. thinning of shade and regular pruning) and

fungicide spray applications. The most devastating insect pests in India

are the white stem borer and coffee berry borer (since 1990). There are no

sources of genetically controlled host resistances available to these two

pests, excepting that generally the white stem borer does not attack

robusta. The CCRI has developed instead a number of effective IPM

systems, including the use of pheromone traps for white stem borer, simple

traps with alcohol mixture as attractant for coffee berry borer, and

parasitoids and natural predators against coffee berry borer, shot-hole

borer, mealy bugs and scales.

Indian coffees are classified according to a rather complex system of

grading, region of origin and specialty coffees. In addition to the regular

coffee grades (PB, AA, A, AB, B and C) for washed/unwashed arabicas

and robustas, the Indian system also distinguishes grades for ‘blacks/

browns’, removed from the regular grades by electronic and manual colour

sorting (see 3.5.4 and 3.5.5), ‘bits’ (beans smaller than C grade and broken

beans) and ‘bulk’ (ungraded and not sorted for off-types). Coffees are also

called after their region of origin, such as Mysore, Chikmagalur, Coorg

(Kodagu), Malabar, Wayanad, Biligiris, Nilgiris, Bababudan, Shevaroys

and Pulneys. Speciality coffees include the Mysore Nuggets EB (extra bold

washed arabicas), Robusta Kaapi Royale (A-grade washed robustas) and

Monsooned coffees. The latter are prepared from cherry arabica and

robusta coffees of AA and A grades, mainly for export to Scandinavian


2.3.6 Factors determining green bean quality

The following main factors determine the quality of green coffee:

environment, cultural practices, genotype of the variety/cultivar and

post-harvest handling (processing). The first three factors determine the

The plant 49

potential quality of the green coffee, which can still be easily degraded by

poor processing practices. Post-harvest handling of the cherries is carried

out by the wet, semi-wet, or dry (natural) method of processing. The wet

process favours a light body and acid beverage, whereas semi-wet and dry

methods lead to enhanced body with reduced acidity. The effect of

environmental, agronomic and genetics factors on coffee quality will be

discussed below, while coffee processing and its relation to quality is

described in 3.3 to 3.5. Environment

The influence of altitude and temperature on the quality of coffee has

been well documented. Arabica originated from the highlands of Ethiopia

and its mild and pleasant beverage is best preserved under similar growing

conditions. High altitude is critical for cultivation of arabica near the

equator. Moderate temperatures will favour a slow and uniform maturation

process of coffee berries and especially the wide amplitude between

day and night temperatures (thermoperiod) will increase flavours and

aromatic precursors in the beans. Coffee fruits developed at higher

altitudes produce more mucilage and they are richer in sugars and other

soluble solids.

The minimum annual rainfall for arabica production is about 1200mm

per year and the maximum should not exceed 2000–2400 mm. Coffee

plants grow and yield better if exposed to alternated cycles of wet and dry

seasons, and, moreover, a period of water deficit is important to

synchronize flower bud differentiation. Areas with excess precipitations,

especially during crop maturation, have a tendency to produce lower

quality coffee due to irregular cherry ripening and poor conditions for

drying the crop after harvesting. High air humidity may cause beans with

off-flavours like fermented and hardish, phenolic, medicinal and musty

‘rioy’ notes, especially in dry-processed (natural) coffees (see 3.7.2 for a

definition of these terms). Fermentation notes can be avoided if fruits are

harvested at the ripe cherry stage and wet-processed immediately. The

rioy taste seems to be associated with a biological transformation of 2,4,6-

trichlorophenol to 2,4,6-trichloroanisole, which is facilitated by excess of

moisture. In years of excessively long dry seasons, due to changes in

climate (e.g. caused by El Nin˜ o), shoot dieback and premature ripening of

the berries will result in green beans producing liquor with immature and

astringent notes. For example, a fishy off-flavour in immature beans is

attributed to 4-heptenal, an oxidation product of linolenic acid (Full

et al., 1999).

50 Espresso Coffee Cultural practices

The correct amount of nutrients, the equilibrium of elements and the

application of the right fertilizers are of prime importance for production

of high-quality coffee beans. Among the essential elements, nitrogen

potassium, calcium, zinc and boron are considered the most important.

The endogenous level of potassium will influence total sugar and citric

acid content and is a key element during bean filling. Potassium sulphate

is a better source than potassium chloride. Nitrogen is important for

amino acid and protein build-up and influences caffeine content. Nfertilization

increases the total nitrogen in beans, with a weak negative

correlation to cup quality (Amorim et al., 1973). Calcium is an essential

element for cell wall formation (Ca-pectate) providing more compact

beans and improving the resistance to pathogen attacks. Zinc influences

protein and carbohydrate metabolism, and also affects the synthesis of

auxins, which promote cell elongation during fruit formation in the ‘soft

green bean stage’ (Ferreira and Cruz, 1988). Boron influences flowering

and fruit set, consequently affecting yield.

The effective control of pests and diseases is essential for the

production of quality coffee. The coffee berry borer is a real threat to

green bean quality. Perforated beans lose weight and infected fruits are

susceptible to attack by air-borne fungi facilitating undesirable fermentations,

degrading the final quality of the whole lot. Several off-flavours

have been associated with uncontrolled fermentation in coffee, as

illustrated in the normal versus elevated values for the following

compounds: ethyl 2-methylbutyrate (2.4 in normal vs 37 mg/kg in

fermented beans), ethyl 3-methylbutyrate (22 vs 345 mg/kg) and cyclohexanoic

acid ethyl ester (Bade-Wegner et al., 1997). The infestation of

coffee berries by the fruit fly (Ceratitis capitata) is commonly associated

with a fungus infection (Fusarium concolor), which gives a pink colour to

the bean and strongly affects cup quality (Bergamin, 1963). The presence

of virus, transmitted by mites, also affects negatively the final green bean

quality through precocious fruit ripening, irregular bean formation, and in

several instances the presence of dry and immature fruits (beans with

black colour and pungent/bitter taste at cupping). Genetic aspects

The qualitative differences in cup quality among different arabica

varieties have received special attention during past decade, especially

after the release of new disease-resistant cultivars, such as Catimor,

Sarchimor, Ruiru 11 and Icatu. Comparing cup quality of arabica cultivars

The plant 51

requires rigorous standardization to eliminate confounding effects from

differences in climate, cultural practices and processing methods. New

cultivars are usually compared with traditional varieties renowned for

their excellent liquor characteristics, such as Bourbon, Blue Mountain,

Kent and SL28. Cup quality is evaluated on beverage characteristics like

aroma (flowery, ‘peasy’), taste (bitter, acid, sweet, woody-earthy), flavour

(chocolate, caramel), and body (light, medium, heavy). Data from

analytical tests of the chemical composition of green beans are poorly

correlated with the perceived cup quality as determined organoleptically.

This is an additional handicap for objectively establishing the varietal

differences, unless a panel of expert liquorers is available for judgement.

Recently, catimor and sarchimor progenies of var. Colombia have been

submitted to biochemical evaluations (caffeine, chlorogenic acids, fat,

trigonelline, sucrose) in parallel with cup tasting with respect to a

standard (Caturra): sucrose content and beverage acidity could be

correlated; no significant correlation could, however, be found between

degree of introgression and the biochemical parameters (Bertrand et al.,

unpublished results).

Differences between arabica and robusta coffee are generally very

pronounced and well documented. For instance, the musty smell of

robusta is attributed to the presence of 2-methylsoborneol (MIB)

(Vitzthum et al., 1990), and MIB ranges from 120–430 ng/kg in robusta

versus a level lower than 20 ng/kg in arabica (Bade-Wegner et al., 1993).

This knowledge led to a practical application, whereby steam treatment of

green robusta beans denatures MIB, and so is claimed to allow increased

percentage of robusta in commercial coffee blends. Other major

differences in green bean composition between arabica and robusta coffee

are summarized in Table 2.5.

Arabica has lower levels of caffeine, amino acids and chlorogenic acids

in comparison to robusta, but 60% more total oils. Chlorogenic acids

contribute to astringent notes, so the reduced amounts in arabica favours

its final cup quality (see also 3.11.6). It is known that many aromatic

volatile compounds are dissolved (trapped) in oil droplets and released

during brewing, so the oil fraction may explain some differences in cup

quality between arabica and robusta, particularly in espresso (see Chapters

8 and 9). The influence of the relative composition of the lipid fractions

in coffee on final cup quality is not yet well understood. The major lipid

fractions in arabica include (Fonseca and Gutierrez, 1971) linoleic acid

(C18:1, 47%), palmitic acid (C16:0, 41%), oleic acid (C18:2, 6.4%) and

stearic acid (C18:0, 6%) (see also 3.11.9). Most chemical data in green

arabica coffee are presented by ‘origin of sample’ (Santos, Colombia,

Kenya, etc.) without indicating the specific variety. Analytical data based

52 Espresso Coffee

on variety will provide additional information on interactions of genotype

with environment and type of processing. Sampling over successive crops

helps to understand the year-to-year fluctuations for key compounds, as

exemplified below for % total oils from Brazilian trees growing in the

same farm and under identical cultural practices (Table 2.6).

Besides fluctuations in total oil content, there are genotypic differences

in other non-volatile and volatile compounds, but these differences have

not yet been correlated to particular beverage types. It is thought that

reducing sugars contribute to bitterness upon caramelization during

roasting, so the ideal sugar content and relative fractions must be

determined. The acidity notes depend on the total acid content as well as

The plant 53

Table 2.6 Yearly fluctuation in total oil content of

Brazilian arabica coffees

Cultivar 1989 crop 1990 crop 1991 crop Average

Caturra 15.9 11.5 13.8 13.7

Red Catuai 15.8 13.6 14.3 14.6

Catimor 15.5 13.5 14.1 14.4

Mundo Novo 15.7 13.8 14.4 14.6

Laurina 12.3 13.5 14.9 13.6

Icatu 17.2 15.1 15.5 15.9

Crop Average 15.4 13.5 14.4 14.5

Source: Illy and Viani, 1995

Table 2.5 Main chemical differences between

arabica and robusta green coffees

Compound/fraction Arabica (%) Robusta (%) Reference

Caffeine 1.2 2.4 (> 4) 1

Trigonelline 1.0 0.7 1

Amino-acids 0.5 0.8 1

Chlorogenic acids 7.1 10.3 1

Lipids: total 16 (13 17) 10 (7 11) 1, 2

Oleic acid 6.7 8.2 9.7 14.2 2

Dipertenes: cafestol 0.5–0.9 0.2 1

Kahweol 0.3   1

16-0-methylcafestol   0.07 0.15 1

Sources: 1Illy and Viani, 1995; 2Speer and Kolling-Speer, 2001

on the balance among acid compounds present in green bean (Illy and

Viani, 1995). More than 30 acids are present in green coffee, the most

important being the chlorogenic acids, quinic acid, malic acid, citric acid

and phosphoric acid. Some variability seems to occur in different coffee

genotypes (Table 2.7).

The main acid fractions listed in Table 2.7 should not be considered

absolute since there are fluctuations from year to year, and differences

among varieties (Balzer, 2001). In these data, the arabica-Santos sample

is mainly represented by Mundo Novo and Catuai varieties and the

arabica-Kenya sample is an indication of Bourbon SL28 material. It will

be of interest to know how much of the acid fraction is controlled by

specific varieties, and how much is due to environmental and cultural


The coffee aroma profile is mainly controlled by genotype (species,

varieties), environmental origin and degree of roasting (see 4.4). Mayer et

al. (1999) made a comparative study among genotypes and found that

variety Typica (from Colombia) had a 52% higher content of a sulphur/

roasty compound (2-furfurylthiol) than an arabica sample from Brazil.

Holscher (1996) found similar data, reporting at least five-fold differences

of 2-furfurylthiol between different samples.

The genetic component of coffee aroma profiles may be more a

quantitative rather than a qualitative effect, i.e. the levels present in

green beans at harvest will be modulated by the interaction genotype/

environment. In other words, once an environment is fixed, it will be

always possible to select a variety that produces greater amounts of

particular aromas under a specific location. With further improvement of

analytical tools, it should become possible to differentiate specific coffee

varieties and origins by their profile of volatile compounds, which would

help tremendously in blending standards and quality control.

54 Espresso Coffee

Table 2.7 Variability in acid content among











Arabica/Santos 5.6 6.1 13.8 1.1

Arabica/Kenya 4.7 6.6 11.6 1.4

Robusta/Togo 3.1 2.5 6.7 1.6

Source: Balzer, 2001


T.W. Baumann

2.4.1 Introduction

Imagining the squadrons of bacteria, moulds, insects and grazing mammals

(not to speak of human beings and their activities), which attack, feed on

and destroy plants, we must wonder why the landscapes are not bare but

still covered with a lush vegetation. Biochemical ecology (or chemical

ecology) emerged from the discovery of the chemicals involved in the

above-mentioned interactions between plants, animals and microorganisms.

Since Fraenkel’s remarkable article (1959) regarding the raison d’eˆtre

of the so-called secondary plant substances (Hartmann, 1996), scientists

are aware that plants (and to a smaller extent also animals and bacteria)

produce a vast array of substances not only to defend themselves against

pathogens and predators but also to attract organisms for their own

benefit. Today, biochemical ecology is an established field of natural

sciences (for a comprehensive review see Harborne, 2001). It is

conceivable that a better knowledge of such chemically mediated

interactions will stimulate biocontrolled farming and assist also coffee

growers in their fight against pests in the plantations, and thus – in the

long term – will undoubtedly improve coffee quality, to which all, farmers,

manufacturers and scientists are committed (Illy, 1997).

Before passing over specifically to biochemical ecology in coffee, two

well-studied examples will serve to illustrate this rapidly emerging

discipline of science. Many flowering species attract the pollinator, such

as an insect (e.g. bee, butterfly), a bird (e.g. humming bird), or a mammal

(e.g. bat), by odours and/or pigmentation. Additionally and very

frequently, the morphology of the flower complies with the needs of a

specific pollinator, for example by providing a landing platform for bees

and bumblebees or a hanging thin throat for hummingbirds, etc.

Generally, the pollinator is rewarded by nectar and by a more or less

welcome dust of pollen. However, about one-third of all orchids do not

invest energy into nectar production but have evolved flowers simulating

various kinds of rewards. For example, several species mimic by odours,

colours, shape and texture the female of a distinct solitary bee. The

delusion is so accurate that the male tries to copulate with the flower

whereby the entire pollinia are positioned, for example, on his head to

carry them to the place of a subsequent frustrating pseudocopulation,

finally resulting in the pollination of that flower (Schiestl et al., 1999).

And moreover, after pollination the flower produces a volatile substance

The plant 55

usually emitted by non-receptive female insects to inhibit copulation

(Schiestl and Ayasse, 2001). This example nicely illustrates that

biochemistry (odours and pigmentation) of biochemical ecology cannot

be separated from physical factors including morphology, texture,

mechanical strength, nutritional energy and, last but not least, appearance

in space and time.

A second example demonstrates how the plant calls for help when

attacked by a predator. The larvae of the beet army worm, Spodoptera

exigua, are a pest of the maize plant. In a series of brilliantly designed

experiments, it has been shown that the saliva of the larvae contains a

signal compound (perhaps partly originating from the destroyed plant

membranes) that, locally, and later also systemically, induces the formation

of leaf volatiles specifically attracting parasitic wasps (Cotesia marginiventris)

which in turn lay their eggs into the larvae, and thus eliminate the

predating insect (Turlings et al., 1990; Alborn et al., 1997). Additionally,

the volatiles may protect the plant from attack by other herbivores (Kessler

and Baldwin, 2001). Today, various examples of such interactions within –

at least – three levels (tritrophic interaction: plant/predator/parasite) are

known and have been successfully studied by coupling highly sensitive

electrophysiological techniques with GC-MS (for reviews see: Dicke and

Van Loon, 2000; Pichersky and Gershenzon, 2002).

Yet, secondary compounds govern, besides the sophisticated interactions

described above, another key process of plant survival, which is

named chemical defence (reviewed in Edwards, 1992; Harborne, 2001).

From an evolutionary point of view, general strategies for optimal defence

against predation were formulated, for example that plants are expected

to accumulate protective phytochemicals in a tissue or organ in direct

proportion to the risk of predation of that unit (Rhoades, 1979).

Therefore, organs and tissues with a high nutritional value (seeds,

young leaves, pollen) have a particularly high risk of predation.

2.4.2 The main secondary metabolites in coffee

(see also 3.11)

The phytochemical catalogue of coffee is very large (for an overview of

physiologically active substances see Viani, 1988; Baumann and Seitz,

1992; Clarke and Vitzthum, 2001); however, with respect to chemical

ecology, only a few most prominent coffee compounds have been intensely

studied so far, namely caffeine, chlorogenic acids and trigonelline.

Caffeine is a purine alkaloid (PuA). PuA are divided into methylxanthines

(caffeine, theobromine, theophylline etc.) and methylated uric acids

56 Espresso Coffee

(theacrine, liberine etc). Their ecological functions in relation to plant

development will be discussed in detail below. 5-Caffeoylquinic acid (5-

CQA) is the main component among the coffee’s chlorogenic acids. It is a

metabolite of the phenylpropanoid pathway and often induced by biotic

(pathogens, herbivores) and abiotic (UV, temperature, nutrient, light)

stress. Thus, it is thought to exert specific functions in plant protection

such as defending against microbial infection and herbivores, acting like a

screen against harmful UV radiation, and scavenging free radicals and

other oxidative species (reviewed in Grace and Logan, 2000). As will be

outlined below, chlorogenic acid is the ally of caffeine, as it associates with

it in a physico-chemical complex (Sondheimer et al., 1961; Horman and

Viani, 1972; Martin et al., 1987). Lastly, trigonelline is a derivative of

nicotinic acid and influences many events in plant life (reviewed in

Minorsky, 2002): it is a hormone-like regulator of the cell cycle and arrests

the cells in the phase prior to mitosis (G2); it has a not yet fully cleared

function in nodulation of the legume alfalfa by rhizobial symbionts

(interacting individuals of different species); it is believed to be a signal

transmitter in the response to oxidative stress; it may act as an

osmoregulator; and, finally, it has been reported to induce leaf closure in

various species showing sleep movements, the so-called nyctinasty.

Additionally, trigonelline was shown to act as a reserve for the synthesis

of coenzymes (NAD) during early coffee seed germination (Shimizu and

Mazzafera, 2000). To our knowledge no report exists regarding the

chemical defence properties of trigonelline, though its considerable

concentration of around 1% (dry weight) in both seeds and leaves, and

even higher in the youngest internode (>2%), this indicating a transport

from old to young leaves (Kende, 1960).

Clearly, the chemical defence by caffeine and chlorogenic acids works

because of (a) their high concentration in the related coffee organ or

tissue, (b) the generally low body weight of the herbivore (a phytophagous

insect may be deterred after one cautious trial, while a large naive animal,

such as a mammal, may swallow one or several plant organs at a time and

thus be intoxicated, but later will avoid that organ and feed selectively on

the rest of the plant) and (c) the need of the herbivore to ingest a large

quantity of leaves due to their overall low protein content. Very

convincingly, the coffee seed, furnished with comparatively high protein

has, in addition to the chemical protection, a strong mechanical defence:

the endosperm is extremely hard (see 2.4.5) and the inner fruit wall, the

endocarp, called ‘parchment’, is tough.

If we relate chemical defence of the coffee bean to the human being

foraging on it, we can state the following: the coffee drinker’s body

weight, which nota bene does not result from the protein content of the

The plant 57

bean(!), is high compared to the biomass of ground endosperm (roast

coffee) used for the coffee brew. And still, some coffee drinkers are only

pleased with a plant extract smoothed by sugar, cream or milk. The

protecting layers around the bean (see 2.4.3 and 2.4.5) already have been

removed by the lengthy and labour-intense processing of the coffee fruit.

Finally, the target molecule (receptor) of pharmacological action in the

coffee drinker is different from that in, for example, an insect (Table 2.8).

2.4.3 From seed to plantlet

As mentioned in the introduction, biochemical ecology is also a matter of

space and time and therefore has to deal with questions such as ‘how and

to where is coffee dispersed?’, ‘how long does it take for a seed (. coffee

bean) to germinate?’, or ‘what happens during the transitions from a seed

to a whole plant?’. Most of the results cited below were obtained by

studies on arabica but, slightly modified, also apply to robusta.

The hobby gardener may seed a so-called ‘green’ (unroasted) coffee bean

possibly available at the local coffee roasting company. By doing so, he or

58 Espresso Coffee

Table 2.8 Organism-related effects of caffeine (multifunctionality)

Organism Effect


mechanism, target References


and fungi


Bacterio- and fungistatic Inhibits UV dark

repair in DNA

Kihlman, 1977,

McCready et al.,


Fungi Fungistatic, reduces

mycotoxin production

Unknown Buchanan et al.,


Plants Inhibits germination,

reduces growth

Not known; inhibits

the formation of the

cell plate, calcium?

Rizvi et al.,

1981, Gimenez-

Martin et al.,



(snails and


Molluscicidal, reduces

heart rate

Calcium release,

increases the duration

of action-potential



et al., 2002

Insects Disturbs developmental


Inhibits cyclic AMP




Mammals Activates CNS, constricts

cerebral blood vessels,

increases lipolysis, positive


Binds to adenosine


Nehlig, 1999

she has to bear in mind two things. First, the bean rapidly loses its

germinating power when harvested and processed, depending on the bean

humidity and on the temperature during storage. Generally, within 3

months after harvest and processing, the germination rate drops to zero,

but is extended to one full year or longer at 18


C, when the bean water

content never falls below 40%, achieved by storing the beans vacuumpacked

in a polyethylene bag (Valio, 1976; Couturon, 1980; Van der

Vossen, 1980). Second, under natural conditions the coffee bean will not

or very rarely get directly in touch with soil, because the seed-dispersing

animals (monkeys, elephants, large birds) attracted by the red, fleshy and

sweet coffee berries will, after having ingested them, excrete or regurgitate

seeds still covered by the hard endocarp, so to speak a parchment coffee, as

it also results technically from wet processing (see 3.3.2). Occasionally, the

entire fruit is dispersed or drops beneath the coffee tree. Then, the berry

will dry and shrink very slowly, looking like dry-processed coffee (see

3.3.1) before hulling. The seeds within the shrunk fruit are even better

protected and lose less water than in the parchment coffee. In any case, in

nature under favourable conditions (humidity) the seed will take up water

and germinate, breaking through the outer shell made up either of the

parchment or the complete husk. Finally, going back to the hobby

gardener, we have to add the following: the naked bean will germinate

earlier (by around 2 weeks) than the bean in parch, but has a higher risk of

being attacked by pests and pathogens in the soil during imbibition

(swelling) and germination. Even though caffeine is described as an

effective agent (Table 2.8) against all kinds of organism (generalists), a few

specialists have overcome the defence mechanism and not only detoxify

but possibly also metabolize the PuA for their own needs. Here, a striking

example is the coffee berry borer, Hypothenemus caffeicola, a serious

worldwide pest, which develops within the caffeine-rich endosperm.

The generalist/specialist rule is pertinent to all secondary compounds so

far studied in this respect. Additionally, one can recognize that these

phytochemicals are multifunctional, meaning that they have not only

multiple functions in the plant producing them but exert also various and

specific effects towards the target organism.

Let us return to germination: Most interestingly, during water uptake

(imbibition) caffeine remains almost completely fixed within the bean,

very likely due to an efficient caffeine barrier at its surface made out of

chlorogenic acids (Dentan, 1985). As will be outlined below, these

phenolic compounds are of crucial importance during the entire life of the

coffee plant, since they prevent autotoxicity by caffeine! Only when the

radicle (the thick primary young root) grows into the substrate, is caffeine

released at high concentrations into the environment (Baumann and

The plant 59

Gabriel, 1984). Obviously, under natural conditions the imbibed bean is

safely encapsulated – there is no need to excrete the defence compound

caffeine – whereas during the later stages of germination about one-third of

the caffeine of the young seedling, now exposed to pathogens and

predators, leaks via the root surface into the substrate, where it may inhibit

the competition by other plant species and prepare, along with other

compounds, the way for specific root colonization by microorganisms.

As nicely illustrated in Figure 2.5 (upper row), the emerging primary root

is the first visible sign of germination. However, dramatic changes occur

even earlier inside the bean: the tiny cotyledons invade the nutrient-rich

storage tissue of the bean, the so-called endosperm, and eventually occupy

the entire cavity within the bean. Simultaneously, the hypocotyl (the

region of the stem beneath the stalks of the cotyledons) stretches and

carries the head above the ground. Finally, the cotyledons unfold and shed

the seed coat. Since during invasion into the endosperm they suck up all its

constituents, cotyledons are, in simplified terms, a blot of the original coffee

bean: they contain and conserve all the caffeine of the endosperm, and

hence are most suitable for a screening to detect and select caffeine mutants

(Baumann et al., 1998). Similarly, the chlorogenic acids are also transported

into the cotyledons where, however, a large fraction is possibly used

for lignin synthesis in order to mechanically stabilize the leaf tissue (Aerts

and Baumann, 1994). Now the seedling is ready to develop into a plantlet.

2.4.4 From bud to leaf

The leaves are arranged in pairs, which alternate along the stem but are

adjusted almost in one plane along the side branches (see also Figure 2.2),

this in order to optimize light harvest. The individual leaf pair is born

from the terminal bud of the stem or of a side branch of first and second

order. In the bud, the leaf primordia (tiny pre-formed installations) are

covered by a resinous layer followed by two tough scale-like bracts

(stipules). The life of the leaf pair starts with its emergence from the bud

with the tiny leaf blades still attached to each other. Thereafter, they

separate and the individual leaf expands considerably to achieve its final

size and shape about 4–5 weeks after emergence. It is still soft, light green

and glossy. During the next 2–3 weeks, the leaf texture gets tough and,

coincidentally, the upper surface of the lamina turns from glossy to dull

dark-green, possibly resulting from a chemical change of the epicuticular

waxy coatings: long chain fatty acids are transformed into the

corresponding alkanes (Stocker, 1976). Now, 50–60 days after emergence,

the leaf is fully developed and optimally gathers solar light energy for the

60 Espresso Coffee

formation of sugars from carbon dioxide and water, or, in other words, the

net photosynthesis has reached its maximum where it remains for a long

time (Frischknecht et al., 1982; Mo¨sli Waldhauser et al., 1997). Under

natural and favourable conditions, the coffee leaf’s lifetime lasts for 10–15

months (Van der Vossen and Browning, 1978). Thereafter, senescence

starts and finally the leaf is shed.

If we throw a glance on the concomitant course of the key secondary

compounds, we find exceptionally high foliar concentrations of PuA and

The plant 61

Figure 2.5 Seed germination and seedling development. (Drawing by Yvonne

Boitel-Baur , Zurich, Switzerland). This illustration shows the various stages from

the bean in parch up to the emergence of the first foliage leaf pair

First row from left to right: Germination, the primary root, and later, the

hypocotyl emerge from the seed. Small lateral roots are formed and the

hypocotyl extends to carry the ’head’ above the ground (c.3 weeks)

Middle row from right to left: The apical hook straightens and the cotyledons (not

visible) completely invade in and dissolve the endosperm (7 weeks)

Lower row left: The cotyledons start unfolding, the remaining unresorbed layer of

the endosperm is shed, and subsequently the cotyledons fully expand (10 weeks).

The first foliage leaves appear later. In the upper left corner, coffee in parchment

(upper row) and ’naked’ beans are shown with one of the latter with the

silverskin partially removed

chlorogenic acids as soon as the very tender, nutrient-rich leaflets have left

the mechanically protective bud, what can easily be explained by their

high risk of predation by, for example, phytophagous insects (Frischknecht

et al., 1986). At this stage, both the enzymes (methyltransferases) involved

in caffeine biosynthesis and the key enzyme of phenylpropane synthesis,

phenylalanine ammonia lyase (PAL), show very high activities (Aerts and

Baumann, 1994; Mo¨sli Waldhauser et al., 1997). The leaf alkaloid

concentration increases almost to 0.1M (as related to the tissue water)

and thus has about 10 times the concentration of an espresso!

The velocity of synthesis of both caffeine and chlorogenic acids

decreases sharply during the subsequent leaf expansion. The relative

caffeine content drops as a consequence of dilution by growth. However,

the absolute amount of caffeine per leaf increases steadily because of low

enzyme activities still persisting throughout the entire period of leaf

expansion (Mo¨sli Waldhauser et al., 1997). The fully developed coffee

leaf has, on a dry weight basis, a caffeine content in the range of the bean.

Amazingly, the chlorogenic acids, even though formed in parallel to

caffeine, still continue to increase during the next six weeks (Kappeler,

1988). Shedding leaves have been reported to be caffeine-free, indicating

that the caffeine nitrogen is re-used by the plant.

The concerted formation of both the alkaloids (mainly caffeine) and

chlorogenic acids (mainly 5-CQA) has a physiological significance:

caffeine easily permeates through all kinds of biological barriers, except a

few installed by the caffeine-containing plants themselves (for example,

the coffee bean’s surface mentioned above). In order to avoid autotoxicity,

caffeine is physico-chemically complexed by 5-CQA, and

compartmented in the cell vacuole (Mo¨ sli Waldhauser and Baumann,

1996). Since chlorogenic acid is encaged in the cells where it was

synthesized, one has to assume that caffeine, due to its hydro- and

lipophilic nature, and to its complexing ability, slowly migrates within the

coffee plant towards the sites of highly accumulated chlorogenic acids. In

other words, caffeine is passively dislocated and, so to speak, collected

within the plant in proportion to the tissue concentration of chlorogenic

acid. Apparently, the coffee plant controls caffeine distribution by the

allocation of chlorogenic acids. This is nicely illustrated by the uneven

distribution of caffeine within the lamina of the coffee leaf: it is highly

accumulated at the margins and sharply decreases in concentration

towards the mid-rib (Wenger and Baumann, unpublished). Needless to

say, the chlorogenic acids show the same distribution pattern. In terms of

biochemical ecology it is important to note that this phytochemical leaf

architecture is significant: the leaf margin, a preferential site of insect

attack, is particularly well furnished with the key defence compounds.

62 Espresso Coffee

Whether phytochemical leaf architecture is genetically controlled and

can be influenced by breeding remains to be investigated.

Thus, caffeine is an ideal defence compound: the organism feeding on

caffeine-containing tissues is unable to hinder it from rapid distribution

within and action on its body. However, from the plant’s view the

problem of autotoxicity had to be solved. The solution was, as already

mentioned, complexation by phenolics: all PuA-containing plants

allocate high concentrations of either chlorogenic acids (coffee, mate´),

catechins (cocoa, cola, guarana´), or both (tea). By the use of suspensioncultured

coffee cells as a model system, it has been shown that up to

77.4% of the caffeine is fixed in the complex at 25


C (Mo¨sli Waldhauser

and Baumann, 1996). Amazingly, in 1964 Sondheimer calculated a

complexation degree of 78% for caffeine in the coffee bean at the same

temperature (1964). Obviously, the concentration achieved by the

remaining fraction (approximately 20–25%) is beyond the autotoxicity

level. A temperature increase lowers the degree of complexation and vice

versa. In this context, several intriguing questions regarding the

interdependence between phenolics and PuA in terms of metabolic

regulation, their final concentrations and organ growth rate are not yet

fully answered but are well considered (e.g. Ky et al., 1999) in the work on

interspecific crosses by the group of Michel Noirot in Montpellier, France


2.4.5 From flower to fruit

The time period between flower opening (anthesis) and the fully ripe fruit

is species-specific and varies considerably among the coffee species. It

depends further on the genotype and on climatic and cultural conditions.

The economically important species C. arabica and C. canephora require

6–8 and 9–11 months for maturation, respectively (Guerreiro Filho, 1992).

As might be expected, all flower organs contain the purine alkaloid

caffeine, with highest concentrations in the stamens. Amazingly, the

latter accumulate, besides traces of theobromine, easily detectable

amounts of theophylline, indicating an alternative biosynthetic pathway

in the male part of the flower with theophylline as the direct precursor of

caffeine. In leaves and seeds, caffeine biosynthesis was shown to proceed

via theobromine. In analogy to citrus plants, where the highest caffeine

concentration has been found in the protein-rich pollen (Kretschmar and

Baumann, 1999), we may assume a preferential PuA allocation also to

coffee pollen grains. However, the related analyses have not yet been

done. Bees are, in contrast to many insects, not only amazingly tolerant

The plant 63

against caffeine and other phytochemicals (Detzel and Wink, 1993), but

rather, after caffeine uptake, have an improved performance such as a

boost in ovipostion by the young queen, an enhanced activity of the bees

outside the hive, and an improved defence by bees against hornets at the

hive entrance (reviewed in Kretschmar and Baumann, 1999).

When the blossom falls from the coffee tree the persisting ovary

develops into the young green coffee fruit (Figure 2.6 and Figure 2.7).

Fruits always signify high investment costs and, therefore, to defend them

against predators the coffee plant pursues several linked strategies.

First, the very young and green fruits are not showy but rather

inapparently arranged in clusters in the leaf axil (Figure 2.6). Secondly,

both chlorogenic acids and purine alkaloids are highly concentrated, and,

thirdly, the development of the real endosperm is postponed until

mechanical protection works. This last point is a most remarkable feature

of the coffee fruit development. Within 3–4 months after anthesis the still

green fruit reaches a size suggesting readiness for maturation. When cut

across, two greenish beans, already typically rolled up, can be recognized.

However, the appearances are deceptive: the fruit is far from being mature,

the (generally) two beans are false perisperm beans made up of mother

tissue (Carvalho et al., 1969). At the adaxial pole (towards the fruit stalk)

of each bean one can see the beginnings of endosperm development: a

whitish, very soft tissue (also called liquid endosperm) starts to invade into

and resorb the perisperm bean. Recent studies show that perisperm

metabolites such as sugars and organic acids are most likely acquired by the

endosperm (Rogers et al., 1999b). We may assume that this process is

similar to the invasion of the cotyledons into the endosperm during

germination described above: the metabolites shift from one tissue to the

other, whereby they have to pass through the so-called apoplast, i.e. the

extracellular space between peri- and endosperm during seed development,

and between endosperm and cotyledons during germination. However, the

endosperm is more than a simple blot of the perisperm, since it owns high

biosynthetic activities. In conclusion, and philosophically speaking, in

coffee the way to the next generation is characterized by transitions in

which the metabolites are shuffled around twice.

During this invasion the inner layer (endocarp) of the fruit wall

(pericarp) noticeably and increasingly solidifies and later results in the

parch layer described above (see 2.4.2 and 2.4.3). The mechanical

defence of the endosperm itself is remarkably increased by the formation

of thick cell walls containing, besides cellulose, the so-called hemicelluloses,

i.e. arabinogalactan and galactomannan (Bradbury, 2001).

Hemicelluloses are highly complex polysaccharides (see 3.11.3) primarily

renowned for giving an amazing degree of hardness to palm seeds (cf. date,

64 Espresso Coffee

Phoenix dactylifera; vegetable ivory, Phytelephas macrocarpa). It remains to

be mentioned that the coffee perisperm finally atrophies into the thin

seed coat, the silverskin, that falls off during roasting. Very soon after

anthesis the pericarp contains an absolute amount of caffeine kept

unchanged until ripeness. However, the initially high (>2%) caffeine

The plant 65

Figure 2.6 Coffee plant (C. arabica) flowering and fruiting. (Drawing by Beatrice

Ha¨ sler , Uster, Switzerland). The drawing shows a side branch with flowers at

various stages: wilting and falling off, in fresh bloom, or buds (from the base to the

apex). Above, flowers are illustrated in detail and enlarged. The yellow stamens are

inserted in the throat of the corolla. Similarly, various fruit stages are shown: like the

flowers they are arranged in composite clusters at the leaf axil

concentration drops by dilution to around 0.2% during the further growth

and maturation processes, culminating in the transformation of the fruit

wall (pericarp) into three distinct layers which serve for fruit dispersal: the

tough endocarp protects the seed from digesting enzyme activities in the

gut of the frugivores such as birds or mammals; the fleshy, sugarcontaining

(Urbaneja et al., 1996) middle layer (mesocarp) softened by

enzymes (Golden et al., 1993) acts as a reward, while the vivid coloration

by anthocyanins (Barboza and Ramirez-Martinez, 1991) of the outermost

layer (exocarp) is to attract the dispersing animal.

We should not close this section without relating our thoughts about

biochemical ecology to a practical question of our daily life: how does the

espresso bean get its caffeine? Though numerous publications on caffeine

biosynthesis exist (for a comprehensive review see Ashihara and Crozier,

1999), this problem has never been addressed and therefore we can only

speculate about it. Clearly, the endosperm has a certain biosynthetic

capacity for caffeine. But is this all? Are there contributions of other

sources? The perisperm provides around one-third of the seed caffeine as

66 Espresso Coffee

Figure 2.7 Fruit development (C. arabica)

Above: Within about 4 months the green fruit grows to a considerable size.

Comparatively late, i.e. between stage 5 and 7, the endosperm starts to develop

(2–3 months)

Below: When the fruit turns olive, the endosperm is already hard (4–5 months).

Now the mesocarp gets fleshy and the exocarp partially red (5–6 months). After 6

months the exocarp is bright red and the mesocarp very fleshy. Later (7–8

months), the fruit colour turns to dark (dull) red and the mesocarp starts to dry

out. Finally, the fruit starts to shrink and the exocarp gets dark and darker

estimated from the caffeine content of the perisperm bean (see Figure

10.5 in Sondahl and Baumann, 2001). The leaves are not directly

contributing to seed caffeine, but the pericarp may be a valuable source, as

studies with labelled caffeine have shown (Keller et al., 1972; Sondahl

and Baumann, 2001). Obviously, caffeine migrates from the fruit wall into

the developing seed, most likely due to the high concentration of

chlorogenic acids allocated to the perisperm/endosperm. Unfortunately,

the extent of this caffeine transport is unknown. Conceivably, this

fraction depends on both the fruit developmental time and the

chlorogenic acids allocations, and is correspondingly larger in a slowripening

species with a high ratio of seed to pericarp chlorogenic acids.

Again, synthesis, transport and accumulation of chlorogenic acids

eventually determine where and how much caffeine is to be allocated

in the seed. In conclusion, perisperm and pericarp are certainly important

sources of the seed caffeine, whereas the leaves, the pericarp and perhaps

also the greenish perisperm may provide most of the chlorogenic acids

crucial to gather and firmly fix the caffeine to the coffee bean. However,

the degree of contribution from each side (maternal tissues versus

endosperm and embryo) is not yet known. Additional studies on the

developmental biology of the coffee seed (Marraccini et al., 2001a, 2001b)

as well as reciprocal crosses between coffee species differing in their

caffeine and chlorogenic acids content (see 2.4.4, From bud to leaf) will

cast some light into the espresso’s darkness!


G. Graziosi*

2.5.1 Introduction

Research on the main arabica cultivars has two main complications: one

historical and the other biological. As reported by Berthaud and Charrier

(1988), most of the cultivars were derived from a small number of plants

that survived transport from Yemen to Southern Asia and to Central and

South America via Europe. Consequently, the genetic base of the main

The plant 67

*I am indebted to Maro Sondahl, Philippe Lashermes, Francois Anthony and Alexander de Kochko for

providing samples of beans and leaves on many occasions. I wish to thank Maro Sondahl for critical reading

of the manuscript and Alberto Pallavicini, Barbara de Nardi, Paola Rovelli and Elisa Asquini for preparing

the figures. The results reported here were supported in part by the European Community (Microsatellites:

Contract ERBIC18CT970181 and ICA4-CT-2001-10070) and in part by illycaffe` SpA (ESTs).

cultivars is rather restricted and the overall diversity is poor. To overcome

this problem and to introduce new valuable genes, a number of

interspecific crosses has been carried out; Arabusta, Icatu` and Catimor

are well known hybrids (Capot, 1972; Carvalho, 1988). The biological

complication is brought about by the self-fertilizing capability of C.

arabica, which enhances genetic homogeneity.

The relatively small amount of fundamental research on coffee has

been somewhat expanded by recent technical developments and by a

moderate increase in scientific studies. New momentum has been gained

through the advent of genomics. DNA sequencing techniques are now

commonly available in most laboratories, and the Coffea arabica genome

(see Glossary at the end of this section for technical words) has been

considered for large-scale analysis.

2.5.2 The genome

The genome of C. arabica consists of 44 small chromosomes (2n . 4x),

twice the number of C. canephora chromosomes and all other Coffea species

(Krug and Mendes, 1940; Bouharmount, 1959; Kammacher and Capot,

1972; Charrier, 1978; de Kochko et al., 2001). C. arabica is an allotetraploid

plant formed by the spontaneous fusion between the genome of C.

canephora, as a male parent, and the genome of a plant of the C. eugenioides

group as a female parent (Lashermes et al., 1999). Indeed the total amount

of DNAper nucleus in arabica (2.6 pg, about 2.4 109 base pairs) is twice as

much as any other Coffea (Cros et al., 1995). The interspecific cross

occurred considerably less than one million years ago and therefore this

species should be considered, in evolutionary terms, as recent.

If C. arabica is a relatively young species, then the two ancestral

genomes should have maintained their individuality with little redistribution

of genes among the chromosomes of different origin (similar

chromosomes between these two genomes are said to be homoeologous).

Evidence to this effect has been provided by Lashermes et al. (2000a), who

studied the inheritance of single genes in this allotetraploid organism.

They found a normal disomic inheritance, as in any other diploid species.

Nevertheless, arabica must have at least four copies of each gene, two

derived from the canephora ancestor and two derived from the eugenioides

ancestor. It is also reasonable to assume that the pairs of orthologous loci

(corresponding genes on the homoeologous chromosomes) could be

somewhat different. Thus, the homozygosity generated by the selffertilizing

property of this organism is presumably compensated by the

constitutive heterozygosity brought about by the two ancestral genomes.

68 Espresso Coffee

As a final note, it is important to mention that a genetic map of C.

arabica does not exist. Nevertheless, low-density maps of C. canephora and

of interspecific crosses are available (Paillard et al., 1996; Ky et al., 2000;

Lashermes et al., 2001; Herrera et al., 2002). Apart from the relatively low

level of research mentioned above, the lack of an arabica map is due to a

number of reasons: the complexity of a tetraploid genome of recent origin,

the high level of homozygosity and last, but not least, the low number of

available polymorphisms.

2.5.3 DNA polymorphisms and molecular diversity

DNA polymorphisms are an important tool for studying genomes and

they have allowed for the identification and mapping of expressed genes

in many organisms. Whilst it is difficult to identify genes responsible for a

given trait, polymorphisms associated with an interesting gene can be

easily followed through various generations and crosses to assist breeders

in their introgressive and selective protocols. In addition, in the past few

years the C. arabica genome has undergone screening projects aimed to

identify polymorphisms. Some of these studies and the main techniques

are reported here.

Cros et al. (1993) and Lashermes et al. (1996b) reported the

identification of many restriction fragment length polymorphisms

(RFLPs) in various Coffea species. This approach detects sequence

variations, frequently single base mutations, and involves the digestion

of DNA by a specific restriction enzyme. Unfortunately, this classical

technique requires Southern blots, which are rather complex and costly.

Moreover, these polymorphisms are frequently not very informative

because they do not have more than two alleles. A new generation of

RFLP is required, possibly based on the enzymatic amplification (PCR:

polymerase chain reaction) of expressed genes.

Several publications have reported a random amplified polymorphic

DNA (RAPD) approach (Cros et al., 1993; Lashermes et al., 1993, 1996c;

Orozco-Castillo et al., 1994; Zezlina et al., 1999; Ruas et al., 2000). This

technique is very useful since it allows for the identification of DNA

polymorphisms in almost any organism even when there is no other

information available on the DNA under study. It is based on the

enzymatic amplification of short DNA stretches (100–1000 base pairs)

lying between two inverted repeats. Agwanda et al. (1997) reported the

first case of association between RAPD polymorphisms and the T gene for

resistance to coffee berry disease (CBD) in Coffea.

The plant 69

The AFLP technique (amplified fragment length polymorphism) has

proved to be very informative: many electrophoretic bands (30–90 bands)

can be obtained in a single PCR and the probability of finding a

polymorphic band is relatively high. This technique is based on the

digestion of the DNA by two restriction enzymes, adapters are added at

both ends of the DNA fragments and some of the DNA fragments are PCR

amplified. AFLP has been successfully used for characterizing varieties

(Anthony et al., 2002) as well as in evaluating the introgression of genes in

hybrids (Lashermes et al., 2000b). Nevertheless, AFLP is relatively

complex and it does not allow for the identification of heterozygotes.

Undoubtedly, microsatellites are the most useful source of polymorphisms

and they are widely used in plant, animal and human research.

Microsatellites are DNA sequences repeated in tandem and the number of

repeats can vary in different organism of the same species. Usually, they

are very informative because one locus can show many alleles (high

heterozygosity) and because they display codominance, thus allowing for

the identification of heterozygotes. Moreover, analyses are easily carried

out by PCR and a large number of samples can be analysed at low cost and

in a short time. However, the development of assays for these

polymorphisms is sometimes difficult and expensive. The first description

of microsatellites in C. arabica was reported by Mettulio et al. (1999),

Rovelli et al. (2000) and Combes et al. (2000). Since then, microsatellites

have been used in a variety of studies, as reported by Lashermes et al.

(2000b), Prakash et al. (2001) and Anthony et al. (2002). As expected,

microsatellites can show different alleles as well as individual differences

in C. canephora (Taylor et al., 2002) but they can show some

complications in C. arabica. In the cross reported in Figure 2.8, the

microsatellite GTG52 shows normal disomic inheritance, but it is not

unusual to find individual plants displaying more than two alleles (Figure

2.9). Most probably, the primers used in these experiments recognized

more than one locus, as would be expected for a polyploid organism.

There are other classes of polymorphic sequences, for instance the

Internally Transcribed Sequence (ITS) of the ribosomal region, but their

use is limited to specific purposes since they can display sequence

variations within the same plant (Zezlina et al., 1999).

The general picture that can be drawn from polymorphic DNA studies

is that C. arabica is a peculiar crop with a very low level of heterozygosity.

This is not surprising because of the restricted genetic base and the high

incidence of autofertility. There is little doubt that this low heterozygosity

will hamper both the development of a genetic map and the use of

molecular genetics in breeding programmes until the development of a

reasonable number of informative polymorphisms.

70 Espresso Coffee

The plant 71

Figure 2.8 Genescan analysis of the cross Sarcimor x ET6 with the 9TG52


Line 1: The Sarcimor parent was homozygote for the 219 bp allele

Line 2: The ET6 parent was homozygote for the 235 bp allele

Lines 3 and 4: The two F1 samples presented both alleles (heterozygote) as the

other 15 plants of the same progeny. (Cross performed by F. Anthony at CATIE)

Figure 2.9 Samples as in Figure 2.8 analysed with the 7TG46 microsatellite. The

two F1 samples showed three alleles: 131 bp, 140 bp and 151 bp. Most probably

the primers amplified two loci, one of which was homozygote for the 131 bp allele

in both parents

2.5.4 Expressed genes

Completion of the model plant Arabidopsis thaliana genome sequence

(The Arabidopsis Genome Initiative, 2000) has given new insights into

the complexity of the plant genome. About 25 000 genes are sufficient for

supporting the life cycle of a flowering plant and there is no reason to

believe that the coffee plant requires more genes. This estimate could be

valid for C. canephora but it is certainly an underrepresentation for C.

arabica, which has twice as many chromosomes, and the gene copies on

the homoeologous chromosomes might have significantly diverged in

sequence and function throughout evolution. None the less, the size of

the arabica genome renders it amenable to mass analysis.

The first catalogue of expressed genes was reported by Pallavicini et al.

(2001). They partially sequenced a cDNA library prepared from root

meristem mRNA and obtained about 1200 ESTs (expressed sequence tag)

which eventually were reduced to 901 clusters. Sequence homology

analysis identified a provisional function for about 70% of the clusters

(Figure 2.10). The complete list of clusters can be obtained from the

home page of the Department of Biology, University of Trieste.

Recently, the public DNA sequence databases reported new C. arabica

ESTs provided by M.A. Cristancho and S.R. McCouch (Genebank

Accession Numbers: BQ4488720-BQ449166) but a reference publication

is unavailable.

72 Espresso Coffee

Figure 2.10 Functional classification of 1200 ESTs

The public databases report a total of 712 sequences of C. arabica and

nine sequences of C. canephora, mainly microsatellites or ESTs. The

remaining sequences can be classified as reported in Table 2.9.

Among the enzyme sequences, it is worth mentioning the complete

sequence of N-methyltransferase, one of the enzymes involved in caffeine

metabolism, which was provided by Kretschmar and Baumann

(Genebank Accession Numbers: AF94411-AF94420). Also important

are the studies on sugar metabolism (Marraccini et al., 2001b), which,

together with the aromatic compounds, are relevant for coffee quality.

The most studied gene encoded the 11S storage protein, the most

abundant protein in the coffee bean (Acun˘a et al., 1999; Marraccini et al.,

1999; Rogers et al., 1999a). The authors identified the complete coding

sequence as well as the promoter region by expressing the protein in

transformed tobacco plants.

Notably, a number of nucleotide binding site–leucine rich repeat

(NBS–LRR) sequences, which are believed to confer resistance to a wide

spectrum of pests, have been identified (Noir et al., 2001). This is the first

step in the possible use of these genes in breeding programmes or


2.5.5 Introducing new genes

Classical breeding has been reviewed by a number of authors, the article

by Van der Vossen (2001) being the most recent. Unfortunately, the

coffee plant is perennial and breeding programs take 20–30 years. Such a

long breeding time coupled with the peculiar genetic structure of C.

arabica (tetraploidy, low heterozygosity and self-fertility) strongly suggests

the use of an alternative route as genetic transformation.

The plant 73

Table 2.9 Categories of

sequences of C. arabica and

C. canephora other than

microsatellite or EST categories

Category No.

Enzymes 19

Structural proteins 5

Ribosomal DNA 4

Chloroplast 6

Successful transformation experiments have been reported by Leroy et

al. (1997a, 1999, 2000) and Spiral et al. (1999). They introduced the

cryIAc gene of Bacillus thuringiensis into arabica and canephora plants

through a disarmed strain of Agrobacterium tumefaciens. The analysis of 60

transformed plants showed a correlation between expression levels of the

cryIAc gene and the newly acquired resistance to leaf miner Perileucoptera

spp. The same technique has been used by Ribas et al. (2001), who

introduced two sequences for the resistance to the herbicide gluphosinate.

An elegant transformation experiment has been reported by Moisyadi

et al. (1998, 1999). They transformed coffee plants with an antisense

sequence of an enzyme apparently involved in the biosynthetic pathway

of caffeine and they obtained plants producing little or no caffeine.

Even if the introduction of genes by transformation must be regarded as

an important tool for improving coffee quality and production, possible

drawbacks must be taken into account. There is little doubt that the

introduction of a gene conferring resistance to one of the many coffee

plant pests would enhance production and be a considerable help to

farmers. Nevertheless, coffee is a perennial plant and a field of plants

resistant to a specific insect, for instance, eventually would select for an

insect insensitive to the toxin. This problem could be partially overcome

by the introduction of more than one toxin for the same pest. In addition,

an important consideration is the attitude of consumers who might be

unwilling to consume a beverage derived from transformed plants.

2.5.6 Prospects

There is little doubt that in the near future more ESTs will be produced

and more polymorphism assays will be developed. In fact, some public

Brazilian funding agencies have launched a research programme for the

production of 200 000 ESTs. Presumably, a number of ‘important’ genes

will be identified, as those involved in flowering and ripening control, or

those responsible for resistance to diseases and pests. These molecular

tools should allow for the amelioration of the coffee plant through

classical breeding. A good example is the introgression of resistance to the

nematode Meloidogyne exigua reported by Bertrand et al. (2001).

Far more complex is the genetic approach to improving the quality of

the coffee beverage. Taste and aroma are presumably under the control of

a large number of genes and affected by environmental conditions.

Nevertheless there is no possibility of having a good quality coffee

without a good genetic base. To approach this problem we have to wait

for the development of new tools. Genome-wide analysis and microarray

74 Espresso Coffee

analysis will allow the study of gene expression and gene–environment

interactions and could become a diagnostic approach to assess quality.

2.5.7 Glossary

AFLP: Amplified fragment length polymorphism. This technique is

based on the digestion of DNA by two restriction enzymes; adapters

are added at both ends of the DNA fragments and some of the DNA

fragments are polymerase chain reaction amplified (see after).

Alleles: Different variants of the same gene.

Allotetraploid: Organism with four sets of chromosomes, two sets

inherited from one ancestral progenitor of a given species and the

remaining two sets inherited from a different species.

Codominance: Two gene variants (alleles) are codominant when they

are simultaneously expressed and visible.

Disomic inheritance: Normal way of inheriting chromosomes or genes.

EST: Expressed sequence tag. Partial sequence of an expressed gene.

Genome: The totality of all the chromosomes and genes of an organism.

Homoeologous chromosomes: Chromosomes carrying similar genes but

originated from two ancestral plants of different species.

Homozygosity: Uniformity of alleles.

ITS: Internally transcribed spacer. Specific portion of DNA located

between the genes coding for the ribosomal RNA.

Microarray: Hybridization technique to visualize the genome expression.

Microsatellite: DNA sequences repeated in tandem.

NBS LRR: Nucleotide binding site leucine rich repeat. Type of gene

conferring resistance to diseases and pests.

PCR: Polymerase chain reaction. In vitro reaction allowing for the

amplification of specific genes or DNA fragments.

Polyploid organism: Organism that contains more than two sets of


Primers: Short fragments of DNA allowing for the initiation of the

PCR (see above).

RAPD: Random amplified polymorphic DNA. Enzymatic amplification

of short DNA stretches lying between two inverted repeats.

RFLP: Restriction fragment length polymorphism. Variation of the

recognition site of a restriction enzyme (see below).

Restriction enzyme: Enzyme able to cut DNA at or near specific


Southern blots: Chromatographic technique for isolating and identifying

specific DNA fragments.

The plant 75

76 Espresso Coffee


Acun˘a R., Bassu¨ner R., Beilinson V., Cortina H., Cadena G., Montes V. and

Nielsen N.C. (1999) Coffee seeds contains 11S storage proteins. Physiol.

Plant. 105, 122–131.

Aerts R.J. and Baumann T.W. (1994) Distribution and utilization of

chlorogenic acid in Coffea seedlings. J. Experiment. Botany 45, 497–503.

Agwanda C.O., Lashermes P., Trouslot P., Combes M.C. and Charrier A.

(1997) Identification of RAPD markers for resistance to coffee berry

disease, Colletotricum kahawae, in arabica coffee. Euphytica 97, 241–248.

Alborn T., Turlings T.C.J., Jones T.H., Stenhagen G. and Loughrin J.H.

(1997) An elicitor of plant volatiles from beet army worm oral secretion.

Science 276, 945–949.

Alvim P. de T. (1973) Factors affecting flowering of coffee. J. Plant. Crops 1,


Amorim H.V., Teixeira A.A., Moraes R.S., Reis A.J., Pimentel-Gomes F. and

Malavolta E. (1973) Studies on the mineral nutrition of coffee. XXVII.

Effect of N, P and K fertilization on the macro- and micronutrients

of coffee fruits and on beverage quality. Ann. Esc. Sup. Agric. LQ

(Piracicaba), 20, 323–333 (in Portuguese).

Anthony F., Combes M.C., Astorga C., Bertrand B., Graziosi G. and

Lashermes P. (2002) The origin of cultivated Coffea arabica L. varieties

revealed by AFLP and SSR markers. Theor. Appl. Genet. 104, 894–900.

Anzueto F., Bertrand B., Sarah J.L., Eskes A.B. and Decazy B. (2001)

Resistance to Meloidogyne incognita in Ethiopian Coffea arabica: detection

and study of resistance transmission. Euphytica 118, 1–8.

Ashihara H. and Crozier A. (1999) Biosynthesis and metabolism of caffeine

and related purine alkaloids in plants. Adv Botanical Res. 30, 117–205.

Bade-Wegner H., Bending I., Holscher W. and Wollmann R. (1997) Volatile

compounds associated with the over-fermented flavour defects. Proc.

17th ASIC Coll., pp. 176–182.

Bade-Wegner H., Holscher W. and Vitzhum O.G. (1993) Quantification of 2-

methylisoborneol in roasted coffee by GC-MS. Proc. 15th ASIC Coll.,

pp. 537–544.

Balzer H.H (2001) Acids in coffee. In R.J. Clarke and O.G.Vitzthum (eds),

Coffee: Recent Developments. Oxford: Blackwell Science, pp. 18–32.

Barboza C.A. and Ramirez-Martinez J.R. (1991) Antocianinas en pulpa de

cafe´ del cultivar Bourbon rojo. Proc. 14th ASIC Coll., pp. 272–276.

Bardner R. (1985) Pest control. In M.N. Clifford and K.C. Willson (eds),

Coffee: Botany, Biochemistry and Production of Beans and Beverage.

London: Croom Helm, pp. 208–218.

Baumann T.W. and Gabriel H. (1984) Metabolism and excretion of caffeine

during germination of Coffea arabica L. Plant Cell Physiol. 25, 1431–1436.

Baumann T.W. and Seitz R. (1992) Coffea. In R. Ha¨nsel, K. Keller, H.

Rimpler and G. Schneider (eds), Hagers Handbuch der Pharmazeutischen

Praxis, Volume 4. Berlin: Springer, pp. 926–940.

The plant 77

Baumann T.W., Sondahl M.R., Mo¨sli Waldhauser S.S. and Kretschmar J.A.

(1998) Non-destructive analysis of natural variability in bean caffeine

content of Laurina coffee. Phytochem. 49, 1569–1573.

Bellachew B. (1997) Arabica coffee breeding in Ethiopia: a review. Proc. 17th

ASIC Coll., pp. 406–414.

Beretta M.J.G., Harakawa R., Chagas C.M., Derrick K.S., Barthe G.A.,

Ceccardi T.L., Lee R.F., Paradela-F. O., Sugimori M. and Ribeiro I.A.

(1996) First report of Xylella fastidiosa in coffee. Plant Dis. (St Paul) 80,


Bergamin J. (1963) Pests and diseases in coffee. In C.A. Krug et al. (eds),

Cultivation and Fertilization of Coffee. Sao˜ Paulo: Inst. Bras. Potassa, pp.

127–141 (in Portuguese).

Berthaud J. (1980) L’incompatibilite´ chez Coffea canephora: me´thode de test

et de´te´rminisme ge´ne´tique. Cafe´ Cacao The´ 24, 267–274.

Berthaud J. and Charrier A. (1988) Genetic resources of Coffea. In R.J. Clarke

and R. Macrae (eds), Coffee: Volume 4 – Agronomy. London: Elsevier

Applied Science, pp. 1–42.

Bertrand B., Aguilar G., Santacreo R., Anthony F., Etienne H., Eskes A.B.

and Charrier A. (1997) Comportement d’hybrides F1 de Coffea arabica

pour la vigueur, la production et la fertilite´ en Ame´rique Centrale. Proc.

17th ASIC Coll., pp. 415–423.

Bertrand B., Anthony F. and Lashermes P. (2001) Breeding for resistance to

Meloidogyne exigua of Coffea arabica by introgression of resistance genes

of Coffea canephora. Plant Pathol. 50, 637–643.

Bertrand B., Guyot B., Anthony F., Selva J.C., Alpizar J.M., Etienne H. and

Lashermes P. (2004) Selection can avoid accompanying the introgression

of Coffea canephora gene of resistance with a drop in beverage

quality, 20th ASIC Coll. In press.

Bertrand B., Pen˜a Duran M.X., Anzueto F., Cilas C., Etienne H., Anthony F.

and Eskes A.B. (2000) Genetic study of Coffea canephora coffee tree

resistance to Meloidogyne incognita nematodes in Guatemala and

Meloidogyne sp. nematodes in El Salvador for selection of rootstock

varieties in Central America. Euphytica 113, 79–86.

Bettencourt A.J. and Rodrigues C.J. (1988) Principles and practice of coffee

breeding for resistance to rust and other diseases. In R.J. Clarke and R.

Macrae (eds), Coffee: Volume 4 – Agronomy. London: Elsevier Applied

Science, pp. 199–234.

Bhat S.S., Daivasikamani S. and Naidu R. (1995) Tips on effective

management of black rot disease in coffee. Indian Coffee 59 (8), 3–5.

Bouharmount, J. (1959) Recherche sur les affinite´s chromosomiques dans le

genre Coffea. INEAC Se´ries Sci. 77, Brussels.

Bradbury A.G.W. (2001) Chemistry I: Non-volatile compounds, 1A –

carbohydrates. In R.J. Clarke and O.G. Vitzthum (eds), Coffee –

Recent Developments. Oxford: Blackwell Science, pp. 1–17.

Bridson D.M. (1994) Additional notes on Coffea (Rubiaceae) from Tropical

East Africa. Kew Bulletin 49 (2), 331–342.

78 Espresso Coffee

Bridson D.M. and Verdcourt B. (1988) Rubiaceae (Part 2). In R.M. Polhill

(ed.), Flora of Tropical East Africa. Rotterdam: Balkema, pp. 703–723.

Buchanan R.L., Tice G. and Marino D. (1981) Caffeine inhibition of

ochratoxin A production. J. Food Sci. 47, 319–321.

Cannell M.G.R. (1985) Physiology of the coffee crop. In M.N. Clifford and

K.C. Willson (eds), Coffee: Botany, Biochemistry and Production of Beans

and Beverage. London: Croom Helm, pp. 108–124.

Capot J. (1972) L’ame´lioration du cafe´ier robusta en Coˆte d’Ivoire: Les

hybrides Arabusta. Cafe´ Cacao The´ 16, 3–18 and 114–126.

Carvalho A. (1946) Geographic distribution and botanical classification of

genus Coffea with special reference to arabica species. Bull. Superint.

Serv. Cafe´ 23, 3–33 (in Portuguese).

Carvalho A. (1988) Principles and practice of coffee plant breeding for

productivity and quality factors: Coffea arabica. In R.J. Clarke and R.

Macrae (eds), Coffee: Volume 4 – Agronomy. London: Elsevier Applied

Science, pp. 129–165.

Carvalho A.C., Ferwerda F.P. et al. (1969). Coffee. In F.P. Ferwerda and F.

Wit (eds), ‘Outlines of perennial crop breeding in the tropics’,

Miscellaneous papers No. 4, Agricultural University, Wageningen,

The Netherlands, pp. 189–241.

Charrier A. (1978) La structure ge´ne´tiquedes cafe´iers spontane´s de la re´gion

malgache (Mascarocoffea). Me´moires ORSTOM 87, Orstom Ed., Paris.

Charrier A. and Eskes A.B. (1997) Les Cafe´iers. In A. Charrier, M. Jacquot, S.

Hamon and D. Nicolas (eds), ‘L’Ame´lioration des plantes tropicales’, pp.

171–196. Repe`res CIRAD and ORSTOM Montpellier (English edition,


Chevalier A. (1947) Les cafe´iers du globe. III. Syste´matique des cafe´iers et

faux cafe´iers, maladies et insectes nuisibles. In Encyclope´die biologique no.

28. Paris: Paul Le Chevalier.

Clarke R.J. and Vitzthum O.G. (2001) Coffee – Recent Developments. Oxford:

Blackwell Science.

Combes M.C., Andrzejewski S., Anthony F., Bertrand B., Rovelli P., Graziosi

G. and Lashermes P. (2000) Characterisation of micro-satellite loci in

Coffea arabica and related coffee species. Mol. Ecol. 9, 1178–1180.

Couturon E. (1980) Le maintien de la viabilite´ des graines de cafe´iers par le

controˆle de leur teneur en eau et de la tempe´rature de stockage. Cafe´

Cacao The´ 24, 27–31.

Couturon E., Lashermes P. and Charrier A. (1998) First intergeneric hybrids

(Psilanthus ebracteolatus Hiern x Coffea arabica L.) in coffee trees.

Canadian J. Botany 76, 542–546.

Cros J., Combes M.C., Chabrillange N., Duperray C., Monnot des Angles A.

and Hamon S. (1995) Nuclear DNA content in the subgenus Coffea

(Rubiacee): inter- and intra-specific variation in African species.

Canadian J. Botany 73, 14–20.

Cros J., Lashermes P., Marmey F., Anthony F., Hamon S. and Charrier A.

(1993) Molecular analysis of genetic diversity and phylogenetic

relationship in Coffea. Proc. 14th ASIC Coll., pp. 41–46.

The plant 79

Cros J., Trouslot P., Anthony F., Hamon S. and Charrier A. (1998)

Phylogenetic analysis of chloroplast DNA variation in Coffea L. Mol.

Phylogenet. Evol. 9, 109–117.

de Kochko A., Louarn J., Hamon P., Hamon S. and Noirot M. (2001) Coffea

genome structure and relationship with evolution. Proc. 19th ASIC Coll.,


Dedecca D.M. (1957) Anatomy and antogenetic development of Coffea

arabica L. var. Typica Cramer. Bragantia 16, 315–366.

Dentan E. (1985) The microscopic structure of the coffee bean. In M.N.

Clifford and K.C. Willson (eds), Coffee: Botany, Biochemistry, and

Production of Beans and Beverage. London: Croom Helm, pp. 284–304.

Detzel A. and Wink M. (1993) Attraction, deterrence or intoxication of bees.

Chemoecology 4, 8–18.

Dicke M. and Van Loon J.J.A. (2000) Multitrophic effects of herbivoreinduced

plant volatiles in an evolutionary context. Entomol.

Experimentalis Applicata 97, 237–249.

Edwards P.J. (1992) Resistance and defence: the role of secondary plant

substances. In P.G. Ayres (ed.), Pest and Pathogens. Plant Responses to

Foliar Attack. Environmental Plant Biology Series, Abingdon: Bios

Scientific Publishers, pp. 69–84.

Eskes A.B. (1989) Resistance. In A.C. Kushalappa and A.B. Eskes (eds),

Coffee Rust: Epidemiology, Resistance and Management. Boca Raton, FL:

CRC Press, pp. 171–291.

Etienne H., Anthony F., Dussert S., Fernandez D., Lashermes P. and Bertrand

B. (2002) Biotechnological applications for the improvement of Coffee

(Coffea arabica L.). Vitro Cell. Dev. Biol. – Plant, 38, 129–138.

Ferreira M.E. and Cruz M.C.P. (1988) Symposium on Micronutrients in

Agriculture: I. Jaboticabal. Sao˜ Paulo, Brazil (in Portuguese).

Flood J. and Brayford D. (1997) Re-emergence of Fusarium wilt of coffee in

Africa. Proc. 17th ASIC Coll., pp. 621–628.

Fonseca H. and Gutierrez L.E. (1971) Study on content and composition of

oil from some coffee varieties. Ann. Esc. Sup. LQ, Piracicaba 18, 313–322

(in Portuguese).

Fraenkel G.S. (1959) The raison d’eˆtre of secondary plant substances. Science

129, 1466–1470.

Frischknecht P.M., Eller B.M. and Baumann T.W. (1982) Purine alkaloid

formation and CO2 gas exchange in dependence of development and of

environmental factors in leaves of Coffea arabica. Planta 156, 295–301.

Frischknecht P.M., Ulmer-Dufek J. and Baumann T.W. (1986) Purine

alkaloid formation in buds and developing leaflets of Coffea arabica:

expression of an optimal defence strategy? Phytochem. 25, 613–616.

Full G., Lonzarich V. and Suggi Liverani F. (1999) Differences in chemical

composition of electronically sorted green coffee beans. Proc. 17th ASIC

Coll., pp. 35–42.

Gimenez-Martin G., Lopez-Saez J.F., Moreno P. and Gonzalez-Fernandez A.

(1968) On the triggering of mitosis and the division cycle of

polynucleate cells. Chromosoma 25, 282–296.

Golden K.D., John M.A. and Kean E.A. (1993) ß-Galactosidase from Coffea

arabica and its role in fruit ripening. Phytochemistry 34, 355–360.

Grace S.C. and Logan B.A. (2000) Energy dissipation and radical scavenging

by the plant phenylpropanoid pathway. Phil. Trans. R. Soc. Lond. B 355,


Guerreiro Filho O. (1992) Coffea racemosa Lour – une revue. Cafe´ Cacao The´

36, 171–186.

Guerreiro Filho O., Denoit P., Pefercen M., Decazy B., Eskes A.B. and

Frutos R. (1998) Susceptibility to the coffee leaf miner (Perileucoptera

spp.) to Bacillus thuringiensis d-endotoxins: a model for transgenic

perennial crops resistant to endocarpic insects. Curr. Microbiol. 36,


Harborne J.B. (2001) Twenty-five years of chemical ecology. Natl Prod. Rep.

18, 361–379.

Hartmann T. (1996) Diversity and variability of plant secondary metabolism:

a mechanistic view. Entomol. Experimentalis Applicata 80, 177–188.

Herrera J.C., Combes M.C., Anthony F., Charrier A. and Lashermes P.

(2002) Introgression into the allotetraploid coffee (Coffea arabica L.):

segregation and recombination of the C. canephora genome in the

tetraploid interspecific hybrid (C. arabica x C. canephora). Theor. Appl.

Genet. 104, 661–668.

Hollingsworth R.G., Armstrong J.W. and Campbell E. (2002) Caffeine as a

repellent for slugs and snails. Nature 417, 915–916.

Holscher W. (1996) Comparison of some aroma impact compounds in roasted

coffee and coffee surrogates. In A.J. Taylor and D.S. Mottram (eds),

Flavour Science, Recent Developments. Cambridge: The Royal Society of

Chemistry, pp. 239–244.

Horman I. and Viani R. (1972) The nature and conformation of the caffeinechlorogenate

complex of coffee. J. Food Sci. 37, 925–927.

ICO (2002) Annual Statistics on Coffee Production and International Trade.

London: International Coffee Organization.

Illy E. (1997) How science can help to improve coffee quality. Proc. 17th

ASIC Coll., pp. 29–33.

Illy A. and Viani R. (eds) (1995) The plant composition. In A. Illy and R.

Viani (eds), Espresso Coffee. The Chemistry of Quality. London:

Academic Press, pp. 23–38.

ITC (2002) Coffee, an Exporter’s Guide. International Trade Centre. Product

and marketing development. UNCTAD CNUCED, WTO OMC.


Kammacher P. and Capot J. (1972) Sur les relations caryologiques entre

Coffea arabica et C. canephora. Cafe` Cacao The 16, 289–294.

Kappeler A.W. (1988) Der Purinalkaloid-Chlorogensa¨ure-Komplex. Seine

physikalisch-chemische Natur und seine Bedeutung in der Gewebekultur

und in den Bla¨ttern von Coffea arabica L. Inaugural Dissertation. Institut

fu¨ r Pflanzenbiologie, Universita¨t Zu¨rich.

80 Espresso Coffee

Keller H., Wanner H. and Baumann T.W. (1972) Kaffeinsynthese

in Fru¨chten und Gewebekulturen von Coffea arabica. Planta 108,


Kende H. (1960) Untersuchungen u¨ber die Biosynthese und Bedeutung von

Trigonellin in Coffea arabica. Berichte Schweiz. Botanisch. Ges. 70,


Kessler A. and Baldwin T. (2001) Defensive function of herbivore-induced

plant volatile emissions in nature. Science 291, 2141–2144.

Kihlman B.A. (1977) Caffeine and Chromosomes. Amsterdam: Elsevier.

Kretschmar J. and Baumann T. (1999) Caffeine in citrus flowers. Phytochem.

52, 19–23.

Krug C.A. and Carvalho A. (1951) The genetics of Coffea. Adv. Genet. 4,


Krug C.A. and Mendes A.J.T. (1940) Cytological observation in Coffea. J.

Genet. 39, 189–203.

Krug C.A. and de Poerck R.A. (1968) World Coffee Survey. FAO

Agricultural Studies no. 76, Rome.

Ky C.L., Barre P., Lorieux M., Trouslot P., Akaffou S., Louarn J., Charrier A.,

Hamon S. and Noirot M. (2000) Interspecific genetic linkage map,

segregation distortion and genetic conversion in coffee (Coffea sp.).

Theor. Appl. Genet. 101, 669–676.

Ky C.L., Louarn J., Guyot B., Charrier A., Hamon S. and Noirot M. (1999)

Relations between and inheritance of chlorogenic acid contents in an

interspecific cross between Coffea pseudozanguebariae and Coffea liberica

var. ‘dewevrei’. Theor. Appl. Genet. 98, 628–637.

Lashermes P., Cros J., Marmey P. and Charrier A. (1993) Use of random

amplified DNA markers to analyse genetic variability and relationship of

Coffea species. Genet. Resour. Crop Evol. 40, 91–99.

Lashermes P., Couturon E., Moreau N., Pailard M. and Louarn J. (1996a)

Inheritance and genetic mapping of self-incompatibility in Coffea

canephora. Theoret. Appl. Genet. 93, 458–462.

Lashermes P., Cros J., Combes M.C., Trouslot P., Anthony F., Hamon S. and

Charrier A. (1996b) Inheritance and restriction fragment length

polymorphism of chloroplast DNA in the genus Coffea L. Theor. Appl.

Genet. 93, 626–632.

Lashermes P., Trouslot P., Combes M.C. and Charrier A. (1996c) Genetic

diversity for RAPD markers between cultivated and wild accessions of

Coffea arabica. Euphitica 87, 59–64.

Lashermes P., Combes M.C., Trouslot P. and Charrier A. (1997) Phylogenetic

relationships of coffee tree species (Coffea L.) as inferred from ITS

sequences of nuclear ribosomal DNA. Theoret. Appl. Genet. 94,


Lashermes P., Combes M.C., Robert J., Trouslot P, D’Hont A., Anthony F.

and Charrier, A. (1999) Molecular characterization and origin of the

Coffea arabica L. genome. Mol. Gen. Genet. 261, 259–266.

The plant 81

82 Espresso Coffee

Lashermes P., Paczek V., Trouslot P., Combes M.C., Couturon F. and Charrier

A. (2000a) Single-locus inheritance in the allotetraploid Coffea arabica L.

and interspecific hybrid C. arabica x C. canephora. J. Heredity 91, 81–85.

Lashermes P., Andrzejewki S., Bertrand B., Combes M.C., Dussert S., Graziosi

G., Trouslot P. and Anthony F. (2000b) Molecular analysis of

introgressive breeding in coffee (Coffea arabica L.). Theor. Appl. Genet.

100, 139–146.

Lashermes P., Combes M.C., Prakash N.S., Trouslot P., Lorieux M. and

Charrier A. (2001) Genetic linkage map of Coffea canephora: effect of

segregation distortion and analysis of recombination rate in male and

female meioses. Genome 44, 589–596.

Leroy J.F. (1980) Evolution et taxoge´nese chez les cafe´iers. Hypothese sur leur

origine. CR Acad. Sci. Paris 291, 593–596.

Leroy T., Henry A.M., Royer M., Altosaar I., Frutos R., Duris D. and Philippe

R. (2000) Genetically modified coffee plants expressing the Bacillus

thuringiensis cry1Ac gene for resistance to leaf miner. Plant Cell Reports

19, 382–389.

Leroy T., Montagnon C., Cilas C., Yapo A., Charrier A. and Eskes A.B.

(1997b) Reciprocal recurrent selection applied to Coffea canephora

Pierre. III. Genetic gains and results of first cycle intergroup crosses.

Euphytica 95, 347–354.

Leroy T., Philippe R., Royer M., Frutos R., Duris D., Dufour M., Jourdan I.,

Lacombe C. and Fenouillet C. (1999) Genetically modified coffee trees

for resistance to coffee leaf miner. Analysis of gene expression,

resistance to insects and agronomic value. Proc. 18th ASIC Coll., pp.


Leroy T., Royer M., Paillard M., Berthouly M., Spiral J., Tessereau S., Legavre

T. and Altosaar I. (1997a) Introduction de ge`nes d’inte´reˆt agronomique

dans l’espe`ce Coffea canephora Pierre par transformation avec

Agrobacterium sp. Proc. 17th ASIC Coll., pp. 439–446.

Marraccini P., Allard C., Andre´ M-L., Courjault C., Gaborit C., LaCoste N.,

Meunier A., Michaux S., Petit V., Priyono P., Rogers J.W. and Deshayes

A. (2001a) Update on coffee biochemical compounds, protein and gene

expression during bean maturation and in other tissues. Proc. 19th ASIC

Coll., CD-ROM.

Marraccini P., Deshayes A., Pe´tiard V. and Rogers W.J. (1999) Molecular

cloning of the complete 11S seed storage protein gene of Coffea arabica

and promoter analysis in transgenic tobacco plants. Plant Physiol.

Biochem. 37, 273–282.

Marraccini P., Rogers W.J., Allard C., Andre M.L., Caillet V., Lacoste N.,

Lausanne F., Michaux S. (2001b) Molecular and biochemical characterization

of endo-beta-mannanases from germinating coffee (Coffea

arabica) grains. Planta 213, 296–308.

Martin R., Lilley T.H., Falshaw P., Haslam E., Begley M.J. and Magnolato D.

(1987) The caffeine-potassium chlorogenate molecular complex.

Phytochemistry 26, 273–279.

The plant 83

Mayer F., Czerny M. and Grosch W. (1999) Influence of provenance and

roast degree on the composition of potent odorantes in arabica coffees.

Eur. Food Res. Technol. 209, 242–250.

McCready S.J., Osman F. and Yasui A. (2000) Repair of UV damage in the

fission yeast Schizosaccharomyces pombe. Mutat. Res. 451, 197–210.

Mettulio R., Rovelli P., Antony F., Anzueto F., Lashermes P. and Graziosi G.

(1999) Polymorphic microsatellites in Coffea arabica. Proc. 18th ASIC

Coll., pp. 344–347.

Minorsky P.V. (2002) The hot and the classic. Trigonelline: A diverse

regulator in plants. Plant Physiol. 128, 7–8.

Mitchell H.W. (1988) Cultivation and harvesting of the arabica coffee tree.

In R.J. Clarke and R. Macrae (eds), Coffee: Volume 4 – Agronomy.

London: Elsevier Applied Science, pp. 43–90.

Moisyadi S., Neupane K.R. and Stiles J.I. (1998) Cloning and characterisation

of a cDNA encoding xantosine-N7-methyltransferase from coffee (Coffea

arabica). Acta Hort. 461, 367–377.

Moisyadi S., Neupane K.R. and Stiles J.I. (1999) Cloning and characterisation

of xantosine-N7-methyltransferase, the first enzyme of the caffeine

biosynthetic pathway. Proc. 18th ASIC Coll., pp. 327–331.

Montagnon C., Leroy T. and Eskes A.B. (1998a) Ame´lioration varie´tale de

Coffea canephora. I. Crite`res et me´thodes de se´lection. Plantations,

Recherche, De´v. 5 (1), 18–33.

Montagnon C., Leroy T. and Eskes A.B. (1998b) Ame´lioration varie´tale de

Coffea canephora. II. Les programmes de se´lection et leurs re´sultats.

Plantations, Recherche, De´v. 5 (2), 89–98.

Moreno G., Moreno E. and Cadena G. (1995) Bean characteristics and cup

quality of the Colombia variety (Coffea arabica) as judged by international

tasting panels. Proc. 16th ASIC Coll., pp. 574–578.

Mo¨ sli Waldhauser S.S. and Baumann T.W. (1996) Compartmentation of

caffeine and related purine alkaloids depends exclusively on the physical

chemistry of their vacuolar complex formation with chlorogenic acids.

Phytochem. 42, 985–996.

Mo¨ sli Waldhauser S.S., Kretschmar J.A. and Baumann T.W. (1997)

N-methyltransferase activity in caffeine biosynthesis: biochemical

characterisation and time course during leaf development of Coffea

arabica. Phytochem. 44, 853–859.

Nathanson J.A. (1984) Caffeine and related methylxanthines: possible

naturally occurring pesticides. Science 226, 184–187.

Nehlig A. (1999) Are we dependent upon coffee and caffeine? A review on

human and animal data. Neurosci Biobehav. Rev. 23, 563–576.

Njoroge S.M., Morales A.F., Kari P.E. and Owuor J.B.O. (1990) Comparative

evaluation of the flavour qualities of Ruiru II and SL28 cultivars of

Kenya arabica coffee. Kenya Coffee 55 (643), 843–849.

Noir S., Combes M.C., Anthony F. and Lashermes P. (2001) Origin, diversity

and evolution of NBS disease-resistance gene homologues in coffee trees

(Coffea L.). Mol. Gen. Genomics 265, 654–662.

84 Espresso Coffee

Orozco-Castillo C., Chalmes K.J., Waugh R. and Powell W. (1994) Detection

of genetic diversity and selective gene introgression in coffee using

RAPD markers. Theor. Appl. Genet. 87, 934–940.

Paillard M., Lashermes P. and Pe´tiard V. (1996) Construction of a molecular

linkage map in coffee. Theor. Appl. Genet. 93, 41–47.

Pallavicini A., Del Terra L., De Nardi B., Rovelli P. and Graziosi G. (2001) A

catalogue of genes expressed in Coffea arabica L. Proc. 19th ASIC Coll.,


Pendergrast M. (1999) Uncommon Grounds: the History of Coffee and How It

Transformed Our World. New York: Basic Books, Perseus Books Groups.

Pichersky E. and Gershenzon J. (2002) The formation and function of plant

volatiles: perfumes for pollinator attraction and defense. Curr. Opin.

Plant Biol. 5, 237–243.

Prakash S.N., Combes M.C., Naveen S.K., Graziosi G. and Lashermes P.

(2001) Application of DNA marker technologies in characterizing

genome diversity of selected coffee varieties and accessions from India.

Proc. 19th ASIC Coll., CD-ROM.

Raina S.N., Mukai Y. and Yamamoto M. (1998) In situ hybridization identifies

the diploid progenitor species of Coffea arabica (Rubiaceae). Theoret.

Appl. Genet. 97, 1204–1209.

Rhoades D.F. (1979) Evolution of plant chemical defense against herbivores.

In Herbivores. New York: Academic Press, pp. 3–54.

Ribas A.F., Kobayashi A.K., Bespalhok Pilho J.C., Galva˜o R.M., Pereira L.F.P.

and Vieira L.G.E. (2001) Trasformac¸a˜o gene`tica de cafe` mediada por

Agrobacterium tumefaciens. II Simposio de Pesquisa dos Cafes do Brasil,

Victoria ES, pp. 420–428.

Rizvi S.J.H., Mukerji D. and Mathur S.N. (1981) Selective phyto-toxicity of

1,3,7-trimethylxanthine between Phaseolus mungo and some weeds.

Agric. Biol. Chem. 45, 1255–1256.

Rogers W.J., Be´zard G., Deshayes A., Pe´tiard V. and Marraccini P. (1999a)

Biochemical and molecular characterization and expression of the 11Stype

storage protein from Coffea arabica endosperm. Plant Physiol.

Biochem. 37, 261–272.

Rogers W.J., Michaux S., Bastin M. and Bucheli P. (1999b) Changes to the

content of sugars, sugar alcohols, myo-inositol, carboxylic acids and

inorganic anions in developing grains from different varieties of robusta

(Coffea canephora) and arabica (C. arabica) coffees. Plant Sci. 149,


Rovelli P., Mettulio R., Anthony F., Anzueto F., Lashermes P. and Graziosi G.

(2000) Microsatellites in Coffea arabica L. In T. Sera, C. R. Soccol, A.

Pandey and S. Roussos (eds), Coffee Biotechnology and Quality.

Dordrecht: Kluwer Academic, pp. 123–133.

Ruas P.M., Diniz L.E.C., Ruas C.F. and Sera T. (2000) Genetic polymorphism

in species and hybrids of Coffea revealed by RAPD. In T. Sera, C. R.

Soccol, A. Pandey and S. Roussos (eds), Coffee Biotechnology and Quality.

Dordrecht: Kluwer Academic, pp. 187–195.

The plant 85

Schiestl F.P. and Ayasse M. (2001) Post-pollination emission of a repellent

compound in a sexually deceptive orchid: a new mechanism for

maximising reproductive success? Oecologia 126, 531–534.

Schiestl F.P., Ayasse M., Paulus H.F., Lofstedt C., Hansson B.S., Ibarra F. and

Francke W. (1999) Orchid pollination by sexual swindle. Nature 399,


Shimizu M.M. and Mazzafera P. (2000) A role for trigonelline during

imbibition and germination of coffee seeds. Plant Biol. 2, 605–611.

Sondahl M.R. and Baumann T.S. (2001) Agronomy II: development and cell

biology. In R.J. Clarke and O.G. Vitzthum (eds), Coffee – Recent

Developments. Oxford: Blackwell Science, pp. 202–223.

Sondahl M.R. and Sharp W.R. (1979) Research in Coffea spp. and

applications of tissue culture methods. In W.R. Sharp et al. (eds),

Plant Cell and Tissue Culture. Principles and Applications. Columbus, OH:

Ohio State University Press, pp. 527–584.

Sondheimer E. (1964) Chlorogenic acid and related depsides. Botanical Rev.

30, 667–712.

Sondheimer E., Covitz F. and Marquise´e M.J. (1961) Association of naturally

occurring compounds, the chlorogenic acid-caffeine complex. Arch.

Biochem. Biophys. 93, 63–71.

Speer K. and Kolling-Speer I. (2001) Lipids. In R.J. Clarke and O.G.

Vitzthum (eds), Coffee – Recent Developments. Oxford: Blackwell

Science, pp. 33–49.

Spiral J., Leroy T., Paillard M. and Pe´tiard V. (1999) Transgenic coffee

(Coffea species). In Y.P.S. Bajaj (eds), Biotechnology in Agriculture and

Forestry. Berlin: Springer, pp. 55–76.

Srinivasan C.S., Prakash N.S., Padma Jyothi D., Sureshkumar V.B. and

Subbalakshmi V. (2000) Coffee cultivation in India. In T. Sera, C.R.

Soccol, A. Pandey and S. Roussos (eds), Coffee Biotechnology and Quality.

Dordrecht: Kluwer Academic Publishers, pp. 17–26.

Stocker H. (1976) Epikutikula¨re Blattwachse bei Coffea. Vera¨nderungen

wa¨hrend der Blattentwicklung; Blattwachse als chemosystematisches

Merkmal. Inaugural Dissertation, Institu¨t fu¨r Pflanzenbiologie,

Universita¨t Zu¨rich.

Taylor E., McGirr R., Bates J., Lee D., Rovelli P., Graziosi G. and Donini P.

(2002) Modern technology for traceability and authenticity of coffee

throughout food processing. Plant, Animal and Microbe Genomes X

Conference, 12–16 January, San Diego, p. 174.

The Arabidopsis Genome Initiative (2000) Analysis of the genome sequence

of the flowering plant Arabidopsis thaliana. Nature 408, 796–815.

Turlings T.C.J., Tumlinson J.H. and Lewis W.J. (1990) Exploitation of

herbivore-induced plants odors by host-seeking parasitic wasps. Science

250, 1251–1253.

Urbaneja G., Ferrer J., Paez G., Arenas L. and Colina G. (1996) Acid

hydrolysis and carbohydrates characterization of coffee pulp. Renewable

Energy 9, 1041–1044.

Valio I.F.M. (1976) Germination of coffee seeds (Coffea arabica L. cv. Mundo

Novo). J. Exp. Bot. 27, 983–991.

Van der Vossen H.A.M. (1980) Methods of preserving the viability of coffee

seeds in storage. Kenya Coffee 45, 31–35.

Van der Vossen H.A.M. (1985) Coffee selection and breeding. In M.N.

Clifford and K.C. Willson (eds), Coffee: Botany, Biochemistry and

Production of Beans and Beverage. London: Croom Helm, pp. 48–96.

Van der Vossen H.A.M. (1997) Quality aspects in arabica coffee breeding

programmes in Africa. Proc. 17th ASIC Coll., pp. 430–438.

Van der Vossen H.A.M. (2001) Coffee breeding practices. In R.J Clarke and

O.G. Vitzthum (eds), Coffee – Recent Developments. Oxford: Blackwell

Science, pp. 184–201.

Van der Vossen H.A.M. and Browning G. (1978) Prospects of selecting

genotypes of Coffea arabica which do not require tonic sprays of fungicide

for increased leaf retention and yield. J. Horticult. Sci. 53, 225–233.

Van der Vossen H.A.M. and Walyaro D.J. (1981) The coffee breeding

programme in Kenya: a review of progress made since 1971 and plan of

action for the coming years. Kenya Coffee 46, 113–130.

Van der Vossen H.A.M., Soenaryo and Mawardi S. (2000) Coffea L. In

H.A.M. van der Vossen and M. Wessel (eds), Plant Resources of South-

East Asia no.16. Stimulants. Leiden: Backhuys Publishers, pp. 66–74.

Viani R. (1988) Physiologically active substances in coffee. In R.J. Clarke and

R. Macrae (eds), Coffee: Volume 3 – Physiology. London: Elsevier Applied

Science, pp. 1–31.

Vitzthum O.G., Weisemann C., Becker R. and Kohler H.S. (1990)

Identification of an aroma key compound in robusta coffees. Cafe´

Cacao The´ 34, 27–33.

Willson K.C. (1985) Mineral nutrition and fertilizer needs. In M.N. Clifford

and K.C. Willson (eds), Coffee: Botany, Biochemistry and Production of

Beans and Beverage. London: Croom Helm, pp. 135–156.

Wrigley G. (1988) Coffee (Tropical Agriculture Series). Harlow: Longman

Scientific and Technical.

Zezlina S., Soranzio M., Rovelli P., Krieger M.A., Sondhal M.R. and Graziosi

G. (1999) Molecular characterisation of the cultivar Bourbon LC. Proc.

18th ASIC Coll., CD-ROM.

86 Espresso Coffee


The raw bean

S. Bee, C.H.J. Brando, G. Brumen,

N. Carvalhaes, I. Ko¨ lling-Speer, K. Speer,

F. Suggi Liverani, A.A. Teixeira,

R. Teixeira, R.A. Thomaziello, R. Viani

and O.G. Vitzthum


A.A. Teixeira

As seen in the previous chapter, the genotype, the macro- and

microclimates, as well as agricultural practices, play a fundamental role

in the quality of coffee. One of the main characteristics of quality coffee is

the possession of good organoleptic properties. Ripe cherry coffee from the

Coffea arabica species is the ideal raw material for obtaining fine quality

coffee (Figure 3.1). In order to maintain the bean’s inherent qualities,

special care is required during processing.


A.A. Teixeira, C.H.J. Brando, R.A. Thomaziello

and R. Teixeira

Preparation for harvesting should be made in advance in order to avoid

last minute surprises and damage to the quality of the coffee. The cleaning

Figure 3.1 Coffee plantation (left) and cherry coffee on the tree (right)

and maintenance of the entire infrastructure and the checking of

equipment to be used should be accomplished at least one month before

the start of the harvest.

Harvesting should only start after a very careful examination of the

level of maturation, when most of the cherries are ripe, with a minimum

(5%) presence of unripe cherries. Unripe cherries lead to light-green

beans, which, when dried at a temperature above 30


C, become blackgreen;

both types weigh less than a normal bean (Teixeira et al., 1991).

The extended stay of the cherries on the tree or on the ground should be

avoided in order not to increase the quantity of sour and black beans,

which also weigh less than normal beans. Light-green, black-green, sour

and black beans are the defects (see 3.7.2) that have the greatest negative

effect on overall coffee quality, damaging type, appearance, roast, yield

and beverage (Teixeira, 1978).

Harvesting can be accomplished in different ways: by stripping onto the

ground (not recommended) or onto sheets, by selective hand picking or

by mechanical means.

3.2.1 Stripping (Figure 3.2a)

This operation is accomplished by stripping all the cherries from the

branches at the same time. The biggest problem is to evaluate the right

time to perform this operation, therefore the need to collect samples of

100 cherries to determine the correct percentage at each stage of

maturation. At the time of harvest, depending on the number and spacing

between one flowering and the other, one can find on a tree immature,

greenish, ripe, overripe and dry cherries.

In the preparation of natural coffee, cherries are dried whole, that is, the

beans are dried while surrounded by the pulp (mucilage and husk). In this

operation, the homogeneous maturation of the cherries is very important

and, therefore, the percentage of green cherries collected should be as low

as possible. In regions with a dry climate, stripping can begin when the

majority of the cherries are in the overripe or partially dry stage. In this

case, drying will be quite uniform and the risk of undesirable fermentation

very low, in contrast to stripping with a high percentage of ripe cherries.

Harvesting by stripping should be done directly onto sheets placed

under the tree beforehand, to avoid contamination and contact with

cherries that have spontaneously fallen and which may have already

deteriorated (Carvalho et al., 1972). For this reason, stripping onto the

ground must be avoided. In general, the sheets are laid under the coffee

trees, in the same direction as the rows of trees, and normally reach

88 Espresso Coffee

beneath three or four plants. Once the stripping is complete, the cherries

should be collected and screen tossed (‘winnowed’) before being bagged

and taken to the processing infrastructure on the day of harvest.

Mechanical winnowing can be used in the coffee plantation or at the

processing installation.

Cherries that have fallen directly onto the ground before harvesting,

called sweepings (varric¸a˜o in Brazil), and those stripped directly onto the

ground should always be processed separately in all the phases of

preparation and should never be mixed with cherries stripped onto sheets.

3.2.2 Hand picking (Figure 3.2b)

In this operation only ripe cherries are picked, normally in baskets or bags.

This type of harvesting is generally employed in regions with constant

rain, where many flowerings occur throughout the year. Harvesting of the

cherries should be performed as many times as necessary. The labour force

should be sufficient to avoid the fall of cherries and their consequent

deterioration. To avoid undesirable fermentation, the ripe cherries should

be taken to the preparation infrastructure and processed on the day of


3.2.3 Mechanical harvesting (Figure 3.3)

This technique is mainly applied in Brazil and Hawaii. The operation can

be done by different systems, all based on the vibration of the branches of

the coffee tree:

n self-propelled machines that remove the cherries from the tree,

collect and winnow them and put the product into bags, hoppers,

carts or trucks;

The raw bean 89

Figure 3.2 Harvesting (a) by stripping and (b) by hand picking

(a) (b)

n mechanical stripping machines, pulled by tractors, that drop the

cherries on the ground;

n portable stripping machines (derric¸adeiras in Brazil) carried by the

picker, using pneumatic or internal combustion power, that drop the

cherries on the ground, whether covered (recommended) or not by

cloths or plastic sheets.

All these systems are being constantly improved in order to increase

selectivity and give priority to the harvest of ripe cherries. An

increasingly common practice in mechanical harvesting is to conduct

the stripping operation at two or three different times, concentrating each

time on the part of the tree with the most ripe cherries.

Self-propelled stripping machines and those pulled by tractors can only

be used on plantations that are flat or slightly undulated, from 10 to 20%

of slope depending on the technology. Portable stripping machines are

compatible with any topography and represent an important alternative

to increase the efficiency of the harvest in mountainous areas.

Mechanical harvesting enables a substantial reduction in the operating

costs that comprise one of the main components of the total cost of coffee


3.2.4 The myth of selective picking

It is widely believed that quality coffee can only be obtained if selective

hand picking is used. This is certainly valid for small plantations, but

becomes a long-standing myth on modern medium-sized to large estates.

Quality coffee can be produced using a variety of harvesting systems,

manual or mechanical. Selective coffee picking – the careful hand picking

of only ripe coffee cherries – is indeed one way to ensure that quality on

the tree is transferred to the cup. However, it is not the only way. In fact,

90 Espresso Coffee

Figure 3.3 Mechanical harvesting

selective picking is nothing more than an indicator that only sound, fresh,

ripe coffee cherries are used as raw material to produce the finest beans

from which a perfect cup is brewed. Sound, fresh, ripe cherries may be

obtained from a variety of picking practices combined with processing

techniques. Top-quality coffee can be produced regardless of the

harvesting technique employed.

When unwanted products are picked, quality must be maintained by

post-harvest separation techniques so that high-quality coffee may still be

produced from the remaining ripe cherries. Coffee growers must therefore

be prepared to cope with harvesting that brings in mixed cherries. Their

ability to separate sound, fresh, ripe cherries from immature, over-ripe,

semi-dry and dry cherries is becoming a crucial factor in the production of

top-quality coffees at a reasonable cost. Modern post-harvest processing

equipment now available can handle such mixed cherries efficiently while

using less water and causing less pollution.


A.A. Teixeira, C.H.J. Brando, R.A. Thomaziello and

R. Teixeira

Processing must start on the same day as the harvest to avoid undesirable

fermentation and reduce the risk of mould contamination, starting with

the pulp of the fruit, rich in nutrients and moisture (Figure 3.4). This is a

critical point in the preparation of coffee and can often jeopardize an

entire year of work caring for the plants.

In all three types of processing (natural, pulped natural or washed),

whenever possible, it is important that the cherries pass through washer–

separators. In this equipment, rocks and impurities are eliminated and the

cherries are separated by density: on one side the lighter cherries (floaters

– the dry and over-ripe cherries) and on the other side the heavier ones

The raw bean 91

Figure 3.4 Stereo microscope view of a coffee cherry: (left) entire; (centre) pulp

(mucilage and husk) partially removed; (right) section

(immature and ripe cherries), enabling the separation of cherries with

different humidity levels, facilitating drying and making the lot of coffee

more homogeneous (Figure 3.5).

3.3.1 The natural (dry) process (Figure 3.6)

Harvesting must begin when most of the cherries are at the over-ripe and

dry stages, especially in regions characterized by a dry harvest season. The

harvesting of a high percentage of ripe cherries can cause undesirable

fermentation, especially if the drying process is not begun on the same day

as the harvest (see also for the risk of mould formation).

In the processing of natural coffee, the whole cherries (bean, mucilage

and pulp) are dried on patios or racks under the sun or in mechanical

dryers. By this process, if well conducted, it is possible to produce a good

coffee with ‘body’ and a pleasant ‘aroma’ (see 3.8.8 for an explanation

of these terms), greatly appreciated in the preparation of espresso


92 Espresso Coffee

Figure 3.5 Cherries passing through washer–separators

Figure 3.6 Patio drying of natural coffee: (left) just harvested; (right) almost dry

3.3.2 The washed (wet) process (Figure 3.7)

The so-called washed or wet process requires a raw material composed of

only ripe cherries that have been selectively picked or are mechanically

separated in the process itself. After passing through the washer–separators

and before removal of the pulp, the separation of the green immature

cherries from the ripe ones can be performed, using differences in pressure,

in a separator of green cherries. This machine has a specially designed

screen with long slotted holes forming a cylinder containing a rotor that

forces the cherries against the stationary screen and towards the edge(s) of

the cylinder. The soft, ripe cherries pass through the holes of the screen.

The hard, unripe cherries, which cannot pass through the holes, go to the

edge(s) of the cylinder where a counter-weight controls their outflow.

The pulping process consists in the removal of the pulp by a pulper,

followed by the removal of the mucilage from the parchment, which can

be accomplished either mechanically, by the use of chemical products, or

by fermentation. The mucilage that adheres to the parchment, with

thickness of 0.5–2.0 mm, is very slimy and composed of pectins and sugars.

Fermentation to remove it is carried out in tanks at ambient

temperature in the presence of microorganisms for between 12 and 36

hours, depending on local temperature. The mass of coffee can be either

immersed in water (wet fermentation) or not (dry fermentation). The

latter type is increasingly employed since it reduces fermentation time. In

wet fermentation, the water must just cover the coffee mass. This process,

however, takes more time and, if longer than 72 hours, can increase the

formation of fluorescent beans, which easily generate ‘stinker’ beans

during storage (see for a definition of the term ‘stinker’).

Fermentation ends when the parchment loses the slimy feel of the

mucilage. This stage can be recognized by rubbing the beans in one’s

hand: when they no longer slip against each other and produce a

characteristic noise, fermentation is concluded.

The raw bean 93

Figure 3.7 Steps of pulping process

Warm water rich in microorganisms is generally used in order to

decrease fermentation time. The use of chemicals is less common and

more risky, and is applied only in research laboratories. In the

fermentation tanks, minerals and sugars are freed from the seeds (coffee

bean with parchment plus mucilage) and can then become concentrated

if the washing water is not changed regularly. The water can easily be

contaminated by all types of microorganisms, leading to uncontrolled

fermentation and a deterioration in cup quality: in this case a slight

‘onion’ flavour, which can reach a very bad ‘stinker’ taste, and can easily

be detected in cup tasting.

Too high pH levels and the presence of ferric ions above 5 mg/l also

lead to the production of off-flavours (Vincent, 1987). Volatile acidity,

mostly owing to the formation of acetic acid, increases with fermentation,

especially if the process lasts longer than 20 hours. After the conclusion of

the pulping operation, the coffee must be very carefully washed so as to

ensure the removal of any trace of mucilage. Incomplete washing can

cause undesirable fermentation, with detrimental consequences to quality.

Removal of mucilage can be carried out mechanically by friction. In

general the mucilage is not completely removed. There has been a strong

trend in the adoption of mechanical removal of the mucilage after the

advent of modern machines, which consume little water and energy and

substitute the traditional aquapulpas. Modern mucilage remover machines

avoid environmental pollution since they consume and contaminate less

water than wet or dry fermentation.

3.3.3 The pulped natural process

A process called pulped natural (cereja descascado in Brazil), intermediate

between the natural and washed processes, began to be used in Brazil in

the early 1990s. In this process, the cherries are pulped and the beans in

parchment dried while surrounded by the mucilage. Fermentation for the

removal of the mucilage is not used in this process. As early as 1960, an

experiment on good cup quality and increased germinative power

demonstrated the validity of pulping cherries without submitting them

to fermentation (Ferraz et al., 1960). The quality of pulped natural coffee,

when well processed, has been shown to be excellent, with the advantage

of producing coffee with greater body than coffee produced using the wet


In this case, harvesting can be conducted by stripping when the

majority of the cherries are ripe. Processing must take place on the same

day, starting with the washer–separators. The pulpers used for the

94 Espresso Coffee

preparation of the pulped natural coffees are equipped with separators of

green (immature) cherries, to separate by pressure immature cherries from

mature ones. In the preparation of pulped natural coffees, the pulp and

remains of mucilage must be disposed of in a location far from the

preparation area. After initial decomposition, it can be used on the farm

as a fertilizer. The immature cherries must be dried separately and coffee

in parchment with unremoved mucilage must be immediately put to dry.

3.3.4 Environmental impact

The processing of the harvest can produce liquid and solid residues that

are aggressive to the environment in case they are not correctly disposed

of or treated.

The natural process is the only one that does not attack the

environment, since it does not require the use of water and does not

produce solid and liquid residues rich in organic substances. Even when

light and heavy cherries are separated by water, contamination is low

(only suspended solids) or inexistent, because the contact between water

and cherries is rapid and the whole cherry does not lose contaminating

substances in the water. The natural process is, therefore, environmentally

friendly by definition.

The pulped natural process produces solid residue (the pulp) and

contaminates the water that enters into contact with the pulp and the

parchment. Finally, the washed (wet) process is the most pollutant of all,

since it requires, in addition to pulping, the removal of the mucilage,

which is extremely rich in organic substances and corresponds to about

half of the polluting load of the process. The washed process is particularly

aggressive when the removal of the mucilage occurs by natural

fermentation, which requires great volumes of water that consequently

become contaminated.

The washed process is, therefore, attracting increasingly great attention

among environmentalists whose pressure is provoking changes to protect

the environment. The pulped natural process was conceived as a less

polluting alternative since it does not require fermentation and was

developed at a time when a greater concern for the environment already


The tendency in wet processing is to use as little water as possible,

transporting the products in a dry state and not in water channels, using

modern machines that consume less water, substituting fermentation by

mechanical removal of the mucilage and re-circulating used water after

The raw bean 95

the removal of solids. The residual waters can be used for irrigation,

infiltrated into the ground, or be treated before disposal in waterways.

The major recent changes to make coffee mills more environmentally

responsible, which are still ongoing, can be grouped in three areas:

n design or layout of the mills;

n water-saving equipment;

n recycling and safe disposal of wastewater.

Old coffee mills had many channels to convey coffee – and water that

became contaminated in the process – from one machine to another.

Modern wet mill design brings machines close together to enable gravity

feeding without water (dry feeding) when slope is available. If slope is

unavailable dry conveyors are used (e.g., bucket elevators, inclined

conveyors, etc. that convey coffee without water) instead of pumps that

can only move coffee that is mixed with water. New water-saving washer–

separators, replacing water-intensive siphon tanks, pulpers and mechanical

mucilage removers, consume much less water than conventional

machines; this is particularly the case of mucilage removers used to

replace traditional fermentation (the most water-intensive stage of the

process and the most contaminating one). Finally, recycling of liquid and

solid wastes produced by wet milling is becoming a required practice, as is

the safe disposal of such water in a manner that does not harm the

environment. Since wastewater treatment and disposal is cheaper and

more convenient for smaller volumes with high pollution loads than for

large volume with smaller pollution loads, the use of as little water as

possible becomes a very critical issue.

Attempts to make the pulping process totally dry have been frustrated

by a fall in the quality of the beans and have subsequently been



A.A. Teixeira, C.H.J. Brando, R.A. Thomaziello and

R. Teixeira

The more homogeneous the mass of coffee, the better and more uniform

will be the drying process. In green immature cherries, humidity reaches

70%, while in mature cherries it varies from 50 to 70%, in over-ripe

cherries from 35 to 50% and in dry cherries from 16 to 30% (Tosello,

1946; Rigitano et al., 1963).

96 Espresso Coffee

The drying of the coffee should be slow. The withdrawal of the bound

water of the bean is difficult and, therefore, a drawn-out process. The

slower the drying operation, including periods of rest, the more

homogeneous will be the final product.

The drying can be conducted under the sun only, on patios or

suspended tables; in mechanical dryers only; or initiated under the sun

and concluded in mechanical dryers.

3.4.1 Patio drying of natural coffee

It is important that natural coffees, normally harvested by stripping,

consisting of immature, greenish, mature, over-ripe and partially dry

cherries, be dried separately. Therefore, the floaters (over-ripe and

partially dry cherries) must be separated from the immature and mature

cherries to form groups of similar humidity to achieve a uniform product,

of good appearance and homogeneous dryness.

The patio area must be well calculated and large enough to avoid the

coffee remaining in thick layers, especially at the beginning of the drying

process, and so damaging its final quality. The drying patios must be

constructed in areas without accumulations of cold air, with great

exposure to sunlight, with a slope between 0.5 to 1.5% to facilitate water

drainage and be coated with bricks, tiles, concrete or even asphalt (see

Figure 3.6).

At the beginning of the drying process, the coffee must be spread in

thin layers with a height of 2–3 cm, and be constantly turned, always

observing the direction of the sun. If the sun is in front of the worker his

shadow should be behind him or if the sun is behind the worker his

shadow should be in front of him.The volume or kg/m2 of coffee required

to obtain such a thickness varies with the stage of maturation and the

percentage of humidity: unripe, ripe, over-ripe and dry cherries (natural

process) or coffee in parchment (wet process). An indicative correlation

indicates that a layer of 5–6 cm of thickness takes 30–40 kg of fruits per m2

of yard: 2–3 kg of cherry coffee (at 60–65% moisture) or 6–9 kg of

parchment coffee per m2 can be dried per day (Matiello et al., 2002).

The coffee must be constantly turned, 15–17 times per day, in both

directions, to speed up the removal of the easily eliminated external water

and to avoid the appearance of fermented or mouldy beans, easily

identified during classification and rejected by the market.

Every evening during the initial days of drying the coffee should be

heaped in thin rows, of 5–10 cm height, in the direction of the patio’s

The raw bean 97

slope. With the passing of the days, the rows can be thickened until the

coffee reaches a stage of semi-dryness, or 20–30% of humidity. From now

on, every evening, the coffee should be heaped and covered with cotton

or waxed fabric that will facilitate the equalization and uniformity of


3.4.2 Patio drying of pulped natural and washed


Fully washed parchment coffee (parchment devoid of mucilage) is dried in

drying patios or suspended tables (Figure 3.8). In the initial phase, the

layer should be very thin, 2–3 cm, with constant turning. Only after skin

drying (the elimination of the external water) can the coffee be dried in

thicker layers. Once it is half dry, the coffee can be heaped and covered

during the night.

It has been observed that the light of the sun, especially radiation

between 400 and 480 nm, is important for the maintenance of coffee

quality, improving aroma and acidity, and reducing mould and bacterial

load, above all in the final phase of the intermediate stage of drying, when

the beans have reached a moisture content below 40% and a dark

translucid aspect (Northmore, 1969).

The care of pulped natural coffees must be greater than that of washed

coffees due to the presence of the (sticky) mucilage. The layers should be

2–3 cm high and constantly turned until the coffee is skin-dry, that is, it

has lost all external humidity. If this drying operation is well performed, it

can be concluded in a single day. Later, the operation will be the same as

the one for coffee in parchment without the mucilage, in patios or

mechanical dryers.

98 Espresso Coffee

Figure 3.8 Sun drying of pulped natural or washed coffee: (left) patio drying;

(right) suspended table drying

3.4.3 Mechanical drying of coffee (Figure 3.9)

Some types of dryer can directly receive the coffee parchment without

mucilage, while others require that the coffee be pre-dried in the sun.

Drying can be accomplished solely by mechanical means in regions with

adverse climate, but normally mechanical drying is used after partial

drying in patios, when the coffee attains a moisture level between 20 and

30%. The dryers should be equipped with heat exchangers (indirect fire

furnaces or steam or water radiators) or gas burner to avoid direct contact

between combustion fumes and coffee and the smell of smoke in coffee.

The use of coffee husks or parchment as fuel can reduce the cost of drying

and the cutting down of trees. Firewood should always be dry, to enable a

higher yield, and to avoid smoke and pollution of the environment.

Mechanical dryers should be loaded with beans of the same moisture to

improve uniformity of drying. All types and brands of dryers are designed

to work with a full load and to avoid heat loss, increase in drying time and

unnecessary energy and labour expenses. The temperature in the coffee

mass should not exceed 40–45


C, to avoid death of the embryo, and

unpleasant smells. When immature cherries are present, the temperature

should not exceed 30


C to avoid their transformation into dark-green and

black-green defects, of very low quality (Teixeira et al., 1982). The 12%

humidity level is reached after the coffee is removed from the dryer,

which it must leave with a slightly higher (around 1%) humidity.

3.4.4 Comparison between natural and washed


The biggest difference between the chemical composition of washed and

natural coffee is in the soluble solids content, which is higher in the case of

natural coffee. Pulped natural coffee presents an intermediate position.

The raw bean 99

Figure 3.9 Mechanical drying of coffee

The difference can be explained by osmotic phenomena. During cherry

drying a gradient of concentration is established between the mucilage

(mesocarp) and the bean (endosperm) with a migration to the bean of

some sugars present in the mucilage. On the other hand, in the wet process,

there occurs a loss of 0.7% of soluble solids during fermentation and 0.7%

during washing, by the diffusion to the water with the mucilage of aqueous

solids from the internal part of the bean (Wootton, 1971). Sugars consist of

one-third of the total loss of soluble solids during fermentation with a

reduction ranging between 0.34 and 0.51%. Studies show that soluble

solids in arabica varieties cultivated in Campinas, Brazil, ranged from 26.7

to 30.5%; the highest values were measured with the Mokka variety. Dryprocessed

samples yielded 0.4–0.9% more solids than wet-processed ones,

where leaching of water-soluble matter occurs during fermentation (Table

3.1) (Carvalho, 1988). The soluble solid content measured by the

Navellier method (Navellier and Brunin, 1963; Wilbaux, 1967) ranged

from 26.1 to 30.6%, with an average of 28.3% in robusta, while in arabica

coffee it ranged from 23.8 to 27.3%, with an average of 25.6%. In Icatu, an

interspecies hybrid between arabica and robusta coffee, this index reached

levels of 26.8–28.6% (Moraes et al., 1973). The greater amount of soluble

solids present in natural coffees is undoubtedly responsible for their

increased body in comparison with washed ones (Table 3.1).

The mineral content is slightly reduced in wet-processed beans

compared to dry-processed ones (Table 3.2).

The natural process does not require high volumes of water and is less

polluting. However, in the wet process, about 60% of the cherry (pulp)

volume is removed, significantly reducing the area and the time required

for drying. In the wet process, where the pulp and mucilage are removed,

the risk of undesirable fermentation is significantly reduced. A further

advantage that must be considered is the possibility of obtaining a product

with a more homogeneous appearance and drying in the cases of pulped

natural and washed coffees. The body produced during the process of

100 Espresso Coffee

Table 3.1 Soluble solids in washed and

natural arabica coffees (% dry matter)

Variety Washed Natural Difference

Caturra 26.71 27.12 +0.41

Mundo Novo 27.91 28.33 +0.42

Yellow Bourbon 27.81 27.85 +0.04

Maragogype 27.90 28.84 +0.94

Mokka 29.69 30.47 +0.83

drying in natural coffees is, however, basic to the sensorial quality of an

espresso coffee, from which derives the recommendation for their use in

this type of preparation. Pulped natural coffees present a body

intermediate between the washed ones, with low body, and the natural

ones, with good body. These coffees are also strongly appreciated in

blends for espresso coffee.



C.H.J. Brando and A.A. Teixeira

Dry parchment coffee or dry cherry coffee must be stored at moisture

levels below 12% in order to avoid the development of musty, earthy or

fermented flavours. Dry coffee can be stored in bulk (silos) or in bags in a

dry atmosphere, preferably at mild temperatures and in the dark, before

final processing for export or roasting. Dry cherry and parchment coffee is

stored in Brazil in large wooden cells, the ‘tulhas’, for up to several weeks

to ‘mature’, i.e., to equilibrate the moisture before final processing

(Figure 3.10).

The raw bean 101

Figure 3.10 Tulhas for maturation of coffee used in Brazil: (left) empty; (right) full

Table 3.2 Mineral content of green

coffee (% dry matter)

Type of coffee Minerals Potassium

Dry-processed robusta 4.14–4.39 1.84–2.00

Dry-processed arabica 4.11–4.27 1.77–1.88

Wet-processed arabica 3.58–3.95 1.63–1.70

In areas where coffee is dry-processed, it is common for the farmer to

deliver green (hulled) coffee to a mill. There it is submitted to rigorous

processing which may include size grading, density separation and colour

sorting. Small farmers may deliver semi-finished product in the form of

dry cherries to cooperatives for further processing.

In areas where coffee is wet-processed, parchment coffee is delivered to

a mill. Hulling and further processing of washed or pulped natural coffees

should only take place shortly before coffee is exported.

3.5.1 Cleaning

Final dry coffee processing starts with the cleaning of the parchment or

cherries, performed in two stages: pre-cleaning and destoning. Precleaning

is accomplished by the suction of light impurities and dust

combined with the sieving of impurities larger or smaller than coffee.

Destoning employs flotation of coffee to separate it from the heavier

stones. A plate or rotary magnet may be used at either stage to separate

iron particles.

3.5.2 Hulling and polishing

Hulling is the process by which the outer shell (husk) of parchment or dry

cherry coffee is removed. Several systems are used depending on the

product and on local practices.

Parchment coffee is often hulled by friction as it is forced to travel in

the space between a screw or multi-facet rotor and its static case. These

machines often remove the parchment husk as well as the ‘silverskin’ (the

perisperm) in a process that is also called polishing or hulling-polishing.

In either case, heat results from the friction and care must be taken to

avoid overheating, which may have negative impact on quality.

Arabica cherry coffee is usually hulled in machines with rotating

blades. These ‘beaters’ force coffee through the openings of a cylindrical

case that encloses the rotor. Since little or no heat is generated in the

process, the silverskin is not removed. Friction hullers for cherry coffee are

used primarily for robusta coffee. They have a rotating cylinder with cleats

and a static adjustable blade. In this case heat is developed and some

polishing may take place.

Removal of silverskin (polishing) is commonly required for washed

coffees but seldom for natural coffees. The cross-beater huller used for

cherry coffee may be also used for parchment coffee if polishing is not


102 Espresso Coffee

All types of hullers include an air column, the ‘catador’ that separates

parchment or cherry husk from the hulled green coffee beans. A fan blows

or aspirates husk and discharges it away from the machine, outside the

building or into a husk cyclone and/or silo. Some hullers have a built-in

larger catador, or may be coupled to one, that separates hulled coffee in

different density fractions, e.g., the ‘heavier’, sound beans and the

‘lighter’, often defective beans. A built-in repassing device is often found

in some types of hullers, especially those for cherry coffee.

The proportion between husk and green coffee is around 20% for

parchment coffee and 50% for cherry coffee.

3.5.3 Grading

Green coffee may be separated by size not only for marketing purposes but

also to enable better density and colour separation. Machines that grade

coffee by size, i.e., coffee graders, can also be used to separate coffee beans

by shape. Flat (regular) beans are the ones with one flat and one concave

surface. Round beans from cherries holding just one seed, called

‘peaberries’, ‘mocas’ or ‘caracoles’, occur in much smaller proportions.

Coffee graders use sieves of different sizes and shapes to separate green

beans according to size and shape. Though most countries rely on sieves

with holes measured in multiples of 1=64 s of one inch (e.g., beans size 18

are retained by a sieve 18=64 in.), millimetres are also used. Very large

beans are usually retained by sieve 19 (7.54 mm) and medium to large

beans by sieve 16 (6.35 mm). Peaberries are separated by slotted (oblong)

screens of several sizes (Figure 3.11).

3.5.4 Mechanical sorting

Densimetric separation aims at removing defects associated with less

dense beans, such as malformed beans, insect-damaged beans, fermented

The raw bean 103

Figure 3.11 Cleaning and grading of coffee (courtesy of Sortex Ltd)

beans, some types of black beans, etc. Traditional catadors, that remove

such defects by airflow, are being increasingly replaced by densimetric

tables, also called gravity separators, using flotation for a more efficient


Air flotation created by powerful fans located below the deck of a

gravity separator causes the light, defective beans to float whereas the

heavier sound beans lie at the bottom, in contact with the deck.

Vibration causes these two fractions to leave the deck separately.

3.5.5 Electronic sorting

S. Bee and F. Suggi Liverani

Differences in bean colour are usually associated with defects (Illy et al.,

1982; Clarke, 1988). This allows the distinguishing of good beans from

defective ones. Coffee sorting by hand is still practised in regions where

labour rates remain low. However, when consumers began demanding

increased quality, cost and complexity of this labour increased.

Automated techniques were introduced since machines can maintain

greater levels of consistency than hand sorting (Anon., 1987), thus

providing a premium quality product at increased margins.

The widespread use of natural coffees in espresso blends makes

electronic sorting a prerequisite for the production of high-quality blends.

The size, cost and complexity of electronic sorting machines varies,

depending on the output and on the complexity of optical measurement

and processing unit. Colour sorters (Vincent, 1987) generally consist of

four principal subsystems (Figure 3.12): feeding system, optical system,

ejection devices and processing unit. Equipment subsystems

Feeding system: In the feeding system, bulk coffee is fed from a vibrating

hopper onto a flat, or channelled, gravity chute. This method separates

the product into a uniform ‘curtain’, or monolayer. The feeding systems

commonly employed are: inclined gravity chute, a flat belt, inclined belt,

contra-rotating rollers and a narrow grooved belt.

Optical system: The optical system measures the magnitude of light

reflected from each single bean. Two or three cameras are used to view

the product from different angles as it leaves the end of the chute,

increasing the defect detection efficiency.

Ejection devices: The ejection process typically takes place while the

product is in free fall. Accepted particles fall along their normal trajectory

while rejected ones are deflected into a receptacle with a blast of

104 Espresso Coffee

compressed air from a high-speed solenoid or piezoelectric valve,

connected to a nozzle. Pneumatic ejector valves must have rapid action

(on/off time of less than 0.25 msec, a typical duty cycle of 150–300 Hz,

firing a pulse of air for 1–3 msec), reliability, a long lifetime (at least a

billion cycles or more) and mechanical strength.

Processing unit: The processing unit manages the control of the

machine and classifies particles as either ‘acceptable’ or ‘rejected’ on the

basis of colour, or both colour and shape. The electronic processing

systems in sorting machines have progressed from the simple analogue

circuits of the early machines to the advanced digital microprocessorbased

ones, as in the present generation of machines. Most of the setting

up of the sorting parameters can be done by the machine itself including,

in some cases, the ability of the machine to ‘learn’ the difference between

a good and a bad product, this is usually represented as a two-dimensional

The raw bean 105

Figure 3.12 Schematic layout of an electronic sorting machine (courtesy of Sortex


reflectivity ‘colour map’ (Figure 3.13), where the boundary curve is the

contour, outlining the acceptable product. A sophisticated control system

will track the average colour of the product so that, even though the

average product colour may change with time, the machine will continue

to remove only the predefined abnormal particles (Figure 3.14).

Advanced sorting machines have the capability of providing information

about the product, for example the number of rejects, or information

about any drifts in colour in a certain batch of product. Sorting techniques

Monochromatic sorting: Monochromatic sorting is based on the

measurement of reflectance at a single isolated band of wavelengths.

The removal of beans lighter or darker in hue than the average is a typical

application of this system.

106 Espresso Coffee

Figure 3.13 Bichromatic colour sorting map (courtesy of Sortex Ltd)

Bichromatic sorting: Bichromatic sorting is only used when a simple

monochromatic measurement is not adequate for effective optical sorting.

Usually green and red filters are used for arabica beans and red or near

infrared for robusta beans, allowing the elimination of unripe, waxy,

chipped, insect-damaged or broken beans as well as of white and black

beans (see 3.7) (Figure 3.15).

Trichromatic sorting: Trichromatic sorting enhances bichromatic

sorting, adding the capability to sort beans by size or shape or for the

detection of gross defects such as the presence of foreign material like

glass, stones and insects. In this way, objects of the same colour but

different shapes, or with holes or cracks, can be effectively removed.

UV fluorescence sorting: It has been found that certain non-visible

defects, such as moulds or bacteria, fluoresce when irradiated with long-

The raw bean 107

Figure 3.15 Spectral reflectance curves obtained from green arabica coffee. The

solid lines represent acceptable beans; the dotted lines represent discoloured beans

(courtesy of Sortex Ltd)

Figure 3.14 An example of coffee sorting (courtesy of Sortex Ltd)

wave ultraviolet light (360 nm). This property may be used as a basis for

sorting. This technique was originally developed for removing ‘stinkers’

from green arabica coffee beans, but has found applications in sorting

other materials as well. However, the fluorescence effects can be best

sorted on freshly harvested coffee. In the case of light beans after a long

storage period following the harvest, this technique faces more difficulties.

Infrared sorting: Over the last decade, the wavelength range used by

sorting machines has been extended from the visible into the infrared

region. Here, both water absorption and other chemical effects play an

important part in determining the reflectivity characteristics of beans.

Optical sorting with lasers: The use of lasers is a technique that is still

in its relative infancy. A laser beam is used to illuminate the product and

the magnitude of reflected light is affected by the amount of laser light

that is either scattered from the surface, or diffused within an object.

Since a laser produces narrow beams of coherent light at a single

wavelength, there is no need to use optical band pass filters. Future trends

Computer vision systems are increasingly being used in the food industry,

pushing the development of new devices and software algorithms, offering

many benefits over a conventional colour sorter, as the ability to simultaneously

sort objects on the basis of multi-criteria. Furthermore, improved

computer technology (Suggi Liverani, 1991) has enabled classification of

green coffee lots automatically, using colour mapping together with

specialized algorithm based on fuzzy logic (Suggi Liverani, 1995). This

improved technology has overcome the limitations of man-made classification

by operators who, for all their experience, may be subject to

variations and imprecision in their perception in detecting defects.


N. Carvalhaes, R. Teixeira, A.A. Teixeira and

C.H.J. Brando

3.6.1 Storage

Correct storage is the maintenance of the product for a certain period

while preserving its original characteristics. Storage can serve the interests

of a nation’s economy by compensating for cyclical and other variations of

agricultural production.

108 Espresso Coffee

Raw coffee, when stored under appropriate conditions of ambient

moisture, temperature and humidity, is quite stable and deterioration is

slow (see 3.11.7 for the changes in amino acid content during storage).

The warehouse must be constructed in a well-ventilated area not subject

to accumulations of cold air. A temperature close to 20


C, and a relative

humidity of the air never exceeding 60%, are optimal. There should be as

little natural light as possible and artificial lighting must be controlled,

preferably located in passageways and corridors. The warehouse must have

good ventilation, without draughts, especially from the south, with welllocated

doors. To avoid development of a woody taste and colour

bleaching of the beans (already observed at moisture levels above 12%)

the coffee must not contain more than 11 0.5% humidity when prepared

by the dry process and 12 0.5%, by the wet process, because excess

humidity in the beans facilitates attacks of mould and bacteria (see also The increase in the humidity level of the beans is on average

0.18% for each percentage unit of increase in the relative humidity of the

air. The correct waterproofing of the floor prevents the deterioration of the

lowest bags in the piles. The coffee bags can be placed on wooden pallets,

or be protected by plastic sheets, to prevent direct contact with the floor

and should never be stacked against walls to prevent any type of humidity.

Coffee warehouses should not contain other agricultural goods or

chemicals that can transmit foreign flavours.

General warehouses with flow for commerce and exportation are

substituting the traditional 60.5 kg bags by so-called ‘big bags’ (with

capacity equivalent to 20 bags), which are made of crossed polypropylene

and are odourless and long-lasting. These serve to store coffee, for

transport to mills, to stuff containers in bulk and for delivery to those

industries that are structured for receiving the product in silos. The factors

that stimulate use of the big bag are: economy of space, ease in internal

transport with forklifts, greater speed and improved stock control.

Some of the changes that take place in the early stages of storage, about

30–40 days of warehousing, may improve quality with the disappearance

of the slightly green taste of fresh recently harvested coffee.

During prolonged storage of hulled coffee, colour changes in the bean

from green to white are common. This discoloration presents a serious

problem since, in addition to depreciating the product in terms of

appearance; such beans can present an inferior cup quality, with

corresponding losses for producers and exporters. It has been verified

that beans proceeding from washed coffee are more affected than those

coming from the dried cherry (pod). The first report published on this

subject (Bacchi, 1962) indicated that bruising of the beans, such as

normally caused by mechanical hulling, is the indirect cause of the

The raw bean 109

whitening of the coffee. Among the extrinsic factors studied, the relative

humidity of the air was the most important. The higher this factor is,

especially at levels above 80%, the faster and the more intense is the

discoloration of the beans.

Another experiment carried out to verify the behaviour of stored beans

in different types of packing and storage time disclosed the occurrence,

after the sixth month of storage, of an irreversible whitening and an

increase in the volume of beans stored in burlap, cotton and paper bags,

while beans stored in cans and plastic bags maintained their green colour

and a normal volume. The whitening of the beans of stored coffees was

more influenced by type of packing than by storage time. This bean

discoloration was accompanied by physical (humidity level, level of dry

matter, absolute density and bean weight) and chemical (soluble nitrogen

content) changes and depreciation in cup quality. The alterations in the

physical and chemical properties are attributed to enzymatic degradation

processes by the polyphenoloxidases and proteases present in the coffee

bean, and are possibly stimulated by atmospheric oxygen and variations in

the humidity of the air, according to the results obtained in paper, cotton

and burlap packing (Mello et al., 1980).

Several factors affect whitening, such as temperature and relative

humidity of the air, ambient light and bean humidity. A study of these

variables indicated that, under constant conditions, only at 10


C of

temperature and 52–67% of air relative humidity did the stored coffee

beans not lose their original colour after 192 days. In some conditions,

even maintaining the bean humidity below 13%, discoloration occurred

during the period. The interaction between temperature and relative

humidity at higher levels makes the beans start to lose colour within a few

days from the beginning of storage (Vilela et al., 2000).

The beans stored in cells of a silo or in bags in warehouses are presented

as a porous mass, constituted by the beans and the interstitial, also called

inter-granulate, space. The oxygen present in the inter-granulate space is

used in the beans’ respiratory process. After harvesting, the beans continue

to live and, like all living organisms, to breathe. The breathing process is

accompanied by the decay of the product’s nutritional substances. The rate

of decay of the stored beans is accelerated as temperature increases.

Beans stored in dry conditions, with humidity between 11 and 13%,

maintain a discreet breathing process. However, if the humidity level is

increased, the breathing and, consequently, the deterioration are

considerably accelerated (Puzzi, 1973).

In Bacchi’s type graphs (Figure 3.16) one can observe that the

hygroscopicity curves are of the sigmoid type, whose final portion,

corresponding to high relative humidity, presents a sharp rise.

110 Espresso Coffee

Temperature, storage time and, most important, moisture are the

critical parameters to control for the prevention of spoilage. About 1% of

the moisture is present as bound water, a further 4% is weakly bound and

the remainder is present as free water. Water availability is measured in

terms of activity (aw), i.e., the ratio between the partial pressure of water

in the bean (P) and the partial pressure of pure water (P0) at the same

temperature (aw.P/P0), and ranges between 0 and 1. Figure 3.17 shows

The raw bean 111

Figure 3.16 Isotherm of moisture content

Figure 3.17 Relationship between water activity and main causes of degradation

the dependence of lipid oxidation (1), non-enzymatic browning (2),

hydrolytic reactions (3), ascorbic acid loss (4), enzymatic activity (5) and

the growth of moulds (6), yeasts (7) and bacteria (8) on aw (Labuza et al.,

1972). The most serious damage to green coffee derives from mould

contamination, the most common moulds belonging to the Penicillium and

Aspergillus species. The principal fungal species found in the storage of

green coffee are: Penicillium spp., Aspergillus ochraceus, A. niger, A.

carbonarius, A. fumigatus, A. tamarii, A. flavus, Wallemia sebi and Eurotium


Humidity, temperature and storage time are dependent variables. The

shorter the storage time, the better tolerated are extreme conditions of

humidity and temperature. Such interdependence can be well described

by the isochronous storage diagram (Figure 3.18). For instance, for a

storage time of 100 days, the relative humidity must be below 75% and

temperature below 30



Relatively minor variations in storage conditions may lead to marked

differences: for example, at 35


C, coffee changes colour after just one

month even when it is stored under the best possible conditions, whereas

at 30


C it is still unaltered after four months. After one year, quality is

generally reduced, and the beans change colour, becoming white–yellow

112 Espresso Coffee

Figure 3.18 Isochronous storage diagram

and fluorescent, their aroma becomes woody and stale while sourness

increases. Such decay is matched by a chemical degradation of the free

amino acids and sugars, and by oxidation of the lipids; it has also been

linked to attack by Streptococci and oxidation (Multon et al., 1973).

Laboratory tests have shown that coffee placed under inert gas

maintains colour and cup quality for years.

Apart from keeping humidity as low as possible in the storage room,

another way to prevent the microbial contamination that damages coffee

could be irradiation with gamma rays before the beginning of storage; this

method, however, has not been implemented, as it may lead to rejection

by consumers. Tests carried out showed that doses ranging between 7.5

and 10 kGy can be used to sterilize coffee in parchment and effectively

reduce moulds, and at the same time delay, even eliminate recontamination,

without producing structural or organoleptic alterations

(Lopez Garay et al., 1987).

Coffees of different qualities must be kept apart in the warehouse;

otherwise defective coffees may rapidly contaminate the good ones. For

example, after three and a half months good coffees stored near stinker

beans pick up the defect. ‘Rio’ coffee (see can also contaminate

sound coffees.

Green coffee at 11–12% moisture is stored in jute or sisal bags (Figure

3.19). These fibres are treated with paraffin and emulsifying agents to

soften and facilitate spinning and weaving into fabrics. The batching oil

content of the bag is normally between 2 and 4%, and, if unclean oil is

used, it can impart a taint to the coffee beans, with loss of cup quality (Grob

et al., 1991a, 1991b). The replacement of mineral oil by neutral vegetable

oils is now common, and the risk of contamination has disappeared.

The density of green beans is about 0.7 gl–1, and a lot of one hundred

60.5 kg bags requires 8.3m3 of warehouse space (Rigitano et al., 1963).

The market now demands current crop coffees and no longer accepts

old crop coffees for blending. There exists, however, a small market

The raw bean 113

Figure 3.19 Storage

segment that purchases such old crop coffees, where price, not quality, is

the most relevant factor. Depending on climatic conditions and warehousing,

the washed coffees and the pulped natural coffees must be

commercialized at most 6 months after processing, because the loss in

quality is greater than that of natural coffees.

3.6.2 Transport

In recent years, with the delivery of merchandise to consuming countries

at the moment of consumption (just-in-time), the operations of transport

in the ports and container terminals have become very efficient and

shorter in duration. The logistical and transportation conditions vary in

the diverse producing origins, from highly developed to rudimentary.

Most of the producing countries have their coffee transported in 20 feet

containers, usually filled with 60.5 kg bags, sometimes in bulk, rarely in

big bags of 1200 kg, while shipment of loose bags in the hold of ships is

disappearing due to the high risk of damage to the load.

For shipment, coffee is normally transferred to new bags, with a

capacity of 60.5 kg, as used in Brazil (in most countries the bag is 60 kg,

while in Colombia it weighs 70 kg). Arabica coffee is still traded

according to the number of defects (see 3.7) and to cup quality.

Robusta coffee is rarely cup tested by the trade, but the exported quality

must contain few defects and foreign materials.

Up to 320 bags can be shipped in 20 feet containers. After arrival

at destination, they are stored at roasting installations for a period of 1 to 3

months. Coffee can be transported in bulk in 20 feet containers that can

hold the equivalent of 360 bags of 60.5 kg. The use of this system is only

possible when the industry at the receiving end has adequate installations.

High-quality (speciality or gourmet) coffee should always be transported

in 60.5 kg bags in containers, for enhanced quality control.

Containers must be well ventilated and equipped with an appropriate

ventilation system, drainage and moisture-collecting system, free from

smells or strange odours.

The critical parameters that need to be taken into consideration for

transport are once again the relative air humidity, the humidity of the

coffee at the time of loading, the changes in the environmental

conditions from the port of loading to the port of discharge and the

duration of the journey. Controls over the loading and unloading

operations are very important. Below follow the most important

conditions in the transportation.

114 Espresso Coffee

n During container shipment, relative humidity should not exceed

70%; in such conditions the coffee travels safely and the risk of

damage is reduced; in case of high relative humidity, the container

must be loaded with a smaller load (maximum of 285 bags) and/or

stowed in well-ventilated areas.

n The shorter the duration of the journey, the lower the risk of


n The more extreme the climatic conditions at loading and unloading

(for example: usual tropical conditions when loading and European

cold season when unloading), the shorter the duration of the journey

should be (Jouanjan, 1980).

Transport in bulk offers several advantages, such as:

n Enabling the loading of the container with 10–15% more coffee (up

to 360 bags).

n The moisture in the mass of coffee is more constant.

n There is a saving of US$1 to US$1.50 of the cost of each bag,

depending on the origin of the shipment.

n Loading and unloading are simplified.

n General costs (insurance, freight, port costs and others) are reduced.

When a container is exposed to the sun in the daytime, water evaporates

and then condenses at night in cold spots, giving rise to pockets of moist

coffee, which can go mouldy and contaminate the remaining portion of

the lot during unloading. This problem can be avoided by completely

lining the container with special paper that protects the coffee mass in

bulk. The same problem can occur with 60.5 kg bags, but in this case the

outer surface of the contaminated bags would present blackish or whitish

spots that can easily be recognized and separated. It is important to

observe the quality of the container. Before loading, the coffee must be

checked, weighed and a phytosanitary certificate issued. This procedure is

very important because phytosanitary controls in consuming countries are

becoming stricter and more demanding. Some producing countries, such

as Brazil, are adopting the use of baits in the interior of the container to

control insects.

Once unloaded, the coffee is again weighed, the figures compared with

those indicated on the certificate of loading and a phytosanitary

certificate is issued. Any resulting difference should be within the limits

of the humidity of coffee respiration. Moreover, samples are taken to

verify if the lot complies with local import regulations.

Figure 3.20 gives the flow-chart of coffee processing.

The raw bean 115


R. Teixeira and A.A. Teixeira

Espresso coffee is rich in lipids, which are excellent aroma carriers,

making espresso an optimal way to increase the aroma of the cup.

Unfortunately, taints coming from defects already present in the green

bean can also be enhanced, so that defects may constitute a major

problem for the quality of an espresso brew, since all aromas, both good

116 Espresso Coffee

Figure 3.20 Flow-chart of coffee processing

and poor, are obvious before drinking, by smelling the foam. Indeed, it

takes about 50 coffee beans to make a cup of espresso coffee, therefore, a

low percentage of defects of the order of 1–2%, common in fair quality

lots on the market, implies the presence of at least one defective bean

among those used to prepare the cup. Furthermore, volatile substances

contributing to the defect are often perceived at extremely low levels,

masking the pleasant aroma of the sound beans (Table 3.3). For this

reason, a brew, still acceptable in spite of defects when prepared by

another technique, can become undrinkable with the espresso system.

The dimension of the problem has led the coffee scientific community

to devote much effort to the characterization of defects from a chemical,

physical and morphological point of view, to the definition of their causes

and to the development of suitable techniques that can prevent the

appearance of defects. This section will provide a detailed characterization

of defects and then discuss means of eliminating them. In addition to

the defects that lead to loss of organoleptic quality of an espresso, the

physical and morphological defects that hamper good roasting and proper

cup preparation will also be discussed.

3.7.1 Characterization of defects

Coffee beans are considered normal when they produce a drink capable of

satisfying the consumer. Common market usage is that coffee beans

without defects make a sound lot. A defect is anything that diverges from

a normal bean inside the lot and that can be produced in the field or

during the harvest, processing, transport and storage. In commercial use,

it is defined as the number of defective beans and foreign matter present

in specific green coffee samples, normally in 300–500 g, or 1000 beans.

Most producing countries have their own criteria for the classification of

coffee, based on the sum of specific defects, each of which is evaluated in

The raw bean 117

Table 3.3 Threshold level in water of the

most odorous chemicals present in defective


Odorant Odour threshold in water (lgl 1)

2,4,6-trichloroanisole 0.001

2-methylisoborneol 0.0025

geosmin 0.005

Source: R. Viani, 2003

accordance with established commercial criteria. In an effort to cover all

coffee types in a general norm, the International Organization for

Standardization (ISO) has produced in the past Standard 10470-1993 –

Green Coffee – Defect Reference Chart (ISO, 1993), which met with

little utilization. This standard is currently in the final stage of a revision

process that aims at a simple and more precise enforcement. Along with a

reorganized chart of defects, now also shown in photographs, the main

new concepts introduced will be:

n Defects are grouped in five classes, namely

n non-coffee defects (foreign matter) (see and Table 3.4);

n defects of non-bean origin (e.g., husks/hulls) (see and

Table 3.5);

n irregularly formed beans (e.g., ears/shells/broken beans/nipped

beans) (see and Table 3.6);

n beans with an irregular visual appearance, with risk of influencing

cup taste) (see and Table 3.7);

118 Espresso Coffee

Table 3.4 Defects associated with foreign matter



definition Causes

Brew flavour/

roasting Origin

Stone* Stone of any diameter

found in a green

coffee lot



Effect mainly



Stick* Twig of any diameter

found in a green

coffee lot






Clod* Granulated lump of

soil particles



Effect mainly



*Found in WPA/WPR/DPA/DPR.

Abbreviations: WPA, wet-processed arabica; WPR, wet-processed robusta; DPA, dry-processed

arabica; DPR, Dry-processed robusta.

H, Harvest-damaged beans: inadequate crop management (picking cherries before or after

maturation, cherries from the ground, etc.)

P, Process-damaged beans: defects due to imperfect processing operations (pulping, washing,

drying, cleaning, hulling, etc.).

The raw bean 119

Table 3.5 Defects associated with fruit parts



definition Causes



roasting Origin

Bean in


Coffee bean entirely or

partially enclosed in its

parchment (endocarp)

Faulty hulling or husking

of the dry parchment




Piece of


Fragment of dried

endocarp (parchment)

Inadequate separation of

the parchment after

hulling or husking








Dried cherry of the

coffee tree, comprising

its external envelopes

and one or more beans

Incorrect dehusking,

allowing whole dried

cherries through, not

removed subsequently

Foul flavour P




Fragment of the dried

external envelope


Inadequate separation

after dehusking

Foul flavour P

*Found in WPA/WPR/DPA/DPR.


Found in DPA/DPR.

P, Process-damaged beans: defects due to imperfect processing operations (pulping, washing,

drying, cleaning, hulling, etc.).

n off-tastes (defects identifiable on cup testing only) (see

and Table 3.8).

n Quantifying defects by weight, to allow for exact calculations to be

applied to any contract of purchase of green coffee that may be

negotiated between provider and client.

Defects can modify cup quality leading to unpleasant flavours, the loss of

product due to the presence of foreign matter or the burning of fragments or

small beans in the roaster. Some defects, such as mouldiness, which can

produce toxins, can also affect consumers’ health. Many defects may appear

equally in wet-processed arabica (WPA), wet-processed robusta (WPR),

dry-processed arabica (DPA) and dry-processed robusta (DPR). Other

120 Espresso Coffee

Table 3.6 Defects associated with regularity/integrity of bean shape

Name Characteristics/definition Causes

Brew flavour/

roasting Origin

Shell or ear*;

shell core*

Part of bean originated from the elephant

bean: shell or ear – external part of the

elephant bean presenting a cavity; shell core

– internal part of the elephant bean

Splitting of the elephant bean (growth

defect) generally through handling (dehulling

or dehusking), producing the separation of

the inner and outer parts

Uneven roast;

slight bitterness;

less acidity


Bean fragment* Fragment of a coffee bean of volume

less than half a bean

Formed mainly during faulty dehulling,

dehusking or pulping operations. Over-drying

making beans easily breakable during


Uneven roast;

bitterness; less



Broken bean* Fragment of a coffee bean of volume

equal to or greater than half a bean

Formed mainly during faulty dehulling,

dehusking or pulping operations.

Over-drying making beans easily breakable

during handling

Uneven roast;

bitterness; less



Insect-damaged bean* Coffee bean damaged internally

or externally by insect attack

Attack on cherries by Hypothenemus

haempei (coffee berry borer)


bitterness; less



The raw bean 121



Coffee bean containing one or more live or

dead insects at any stage of development

Bean attacked by storage pests, generally

Araecerus fasciculatus insect

Loss of flavour;

slight bitterness



pulper-cut beany

Wet-processed coffee bean cut or bruised

during pulping, often with brown or blackish


Faulty adjustment of pulping machine or

feeding with under-ripe cherries or

malformed beans


downgrading to

slightly putrid or



Crushed bean* Crushed beans often partly split and faded

with centre-cut largely open

Treading on the beans during drying.

Hulling or polishing of soft under-dried



downgrading to

rancid, slightly



*Found in WPA/WPR/DPA/DPR.

yFound in WPA/WPR.

Abbreviations: WPA, wet-processed arabica; WPR, wet-processed robusta; DPA, dry-processed arabica; DPR, Dry-processed robusta.

F, Field-damaged bean: defects originating in the field, the coffee tree (genetic problems), the environment (climate, soil, water and nutrient stress), attacks by

pests and diseases.

P, Process-damaged beans: defects due to imperfect processing operations (pulping, washing, drying, cleaning, hulling, etc.).

S, Storage damaged beans: defects due to deficient storage (faulty storage practices and storage pests).

122 Espresso Coffee

Table 3.7 Defects associated with beans irregular in visual appearance (colour and surface texture)

Name Characteristics/definition Causes Brew flavour/roasting Origin

Black bean* Coffee bean whose interior is

partly (partly black bean) or

totally black (endosperm).

Bean from over-ripe cherry, fallen on the ground.

Due to Colletotrichum coffeeanum attack, other

fungi species and pests. Carbohydrate deficiency

bean due to poor cultural practices. Mature cherries

subjected to over-fermentation by moulds/yeasts and

subsequent drying

Slow to roast; beans tend

to be yellowish; loss of

acidity; harsh; ashy flavour


Black-green bean* Unripe coffee bean often with

a wrinkled surface, with darkgreen

(dark-green bean) to

black-green silverskin colour

High temperatures affecting immature beans,

causing chemical transformation (like fermentation)

with the silverskin becoming dark or black-green

Slow roast; rotten fish




immature bean*

Unripe coffee bean often with

a wrinkled surface. The bean

has a greenish or metallicgreen

silverskin colour. Cell

walls and internal structure are

not fully developed

Beans from cherries picked before ripening Slow roast;

bitterness; less acidity;

astringent; metallic;

sometimes fermented


Sour (ardido)


Coffee bean deteriorated by

excess fermentation, with a

range of colours: light to dark

brown-reddish, dark brown or

yellowish green internally

(endosperm). In some cases it

has a waxy appearance or a

brown silverskin colour

Bean from immature, mature or overripe cherries

that have been in adverse conditions, becoming

fermented by bacteria or xerophilic moulds, with

embryo death. More frequent in fruits fallen to the

ground, can also appear in fruits on the tree, and

when there is an excessive time between harvesting

and drying or pulping

Sour; fermented; acetic;

fruity; sulphurous; vinegar



The raw bean 123

Withered bean* Coffee bean wrinkled and light

in weight

Underdeveloped fruit due to genetic and

physiological problems, nutritional lack, drought or

heavily stressed tree

Slight loss of acidity F

Spongy bean* Coffee bean of consistency like

cork that may be verified by

pressure of fingernail. Whitish

in colour

Moisture absorption during storage or transportation

leading to deterioration by enzymatic activity

Roasts rapidly; lack of

acidity; woody


Whitish bean* Coffee bean with whitish

colour with tissue of normal


Discoloration of surface due to bacteria of the genus

Coccus, during storage or transportation. Associated

with old crop. Faulty drying of the bean

Woody; stale; bitterness



Mouldy bean* Coffee bean with mould

growth and evidence of attack

by mould visible to the naked


Temperature and humidity conditions favourable to

mould growth

Musty flavour P/S



Coffee bean with spotted

silverskin like a ‘spotted quail


Frost damaged coffee bean Rancid flavour F

*Found in WPA/WPR/DPA/DPR.

Abbreviations: WPA, wet-processed arabica; WPR, wet-processed robusta; DPA, dry-processed arabica; DPR, Dry-processed robusta.

F, Field damaged bean: defects originating in the field, the coffee tree (genetic problems), the environment (climate, soil, water and nutrient stress), attacks by

pests and diseases.

H, Harvest-damaged beans: inadequate crop management (picking cherries before or after maturation, cherries from the ground, etc.).

P, Process-damaged beans: defects due to imperfect processing operations (pulping, washing, drying, cleaning, hulling, etc.).

S, Storage damaged beans: defects due to deficient storage (faulty storage practices and storage pests).

124 Espresso Coffee

Table 3.8 Defects associated with beans irregular in cup taste after proper roasting and brewing (off-taste coffee)

Name Characteristics/definition Causes



roasting Origin




The bean presents a normal appearance.

Unpleasant foul flavour is detected in

the cup like earthy, woody, musty or


Unfavourable conditions of temperature, humidity and

time in processing, storage or transportation. Use of

bad quality jute-bag

Musty; earthy,

woody, jutebag





The bean presents usually a normal appearance. A

very unpleasant flavour is detected in the cup like

stinker, over-fermented or rotten. Stinker smell when

cut or scratched

Delay in pulping; too long period of fermentation;

wild fermentation in beans trapped in the pulping

machines or tanks. Abrasion of beans during pulping,

losing the superficial protective layer, leaving them

susceptible to microorganism attack. Contamination

with recycled polluted water

Stinker; overfermented,

rotten flavour




The bean presents a normal appearance. Medicinal

smell when cut or scratched.

Overripe cherry left on branch, possibly contaminated

by microorganisms. The drying patio soil is heavily

contaminated by microorganisms and/or TCA




flavour of



*Found in WPA/WPR/DPA/DPR.

Abbreviations: WPA, wet-processed arabica; WPR, wet-processed robusta; DPA, dry-processed arabica; DPR, dry-processed robusta.

F, Field-damaged bean: defects originating in the field, the coffee tree (genetic problems), the environment (climate, soil, water and nutrient stress), attacks by

pests and diseases.

H, Harvest-damaged beans: inadequate crop management (picking cherries before or after maturation, cherries from the ground, etc.).

P, Process-damaged beans: defects due to imperfect processing operations (pulping, washing, drying, cleaning, hulling, etc.).

S, Storage-damaged beans: defects due to deficient storage (faulty storage practices and storage pests).

defects are characteristic of a certain type of processing. The different types

of defects can be divided into categories, which are discussed below. Defects associated with foreign matter

These include defects such as stones, sticks, clods (agglomerations of

earth), metal and foreign matter. They must be removed at an appropriate

stage, for example during the cleaning of green coffee, by sieving,

classifying or by removal of metals. Defects associated with coffee fruit parts

These include defects such as beans in parchment, pieces of parchment,

husk fragments and dried cherries (pods). In general, these are removed by

sieving or by air classifying, leading to a loss of physical volume. Defects associated with regularity/integrity of bean


These include defects like malformed beans (shell/ear and shell core/body

of the elephant), bean fragments, broken beans, insect-damaged beans,

pulper-nipped/pulper-cut beans and crushed beans. They are in general

removed by densimetric sorting and in some cases by optical sorting. Defects associated with beans irregular in visual

appearance (colour and surface texture)

These include defects like black beans, dark- and black-green immature

beans, light-green immature beans, sour beans etc. Some of these can be

removed by hand or by optical sorting techniques. Defects associated with beans irregular in cup taste

after proper roasting and brewing (off-taste coffee)

Defects of sensory concern, to be identified after sample roasting and cup

testing, also entailing a further risk of contamination of other beans. These

include defects like ‘stinker’ beans, foul/dirty beans and ‘rio’ beans. Offtastes

are hard to remove by sorting. They can be identified after cupping a

sample of roast and ground coffee, following proper roasting and brewing.

The raw bean 125

3.7.2 Definition of defects

Many defects have been studied from the physical, chemical, microbiological

and flavour quality points of view. Light-green immature beans

The presence of immature beans, characterized by a light-metallic-green

silverskin colour, is due to the harvest of unripe cherries. The flavour of

the defect, described as an increased bitterness in both arabica and robusta

beans (ISO, 1993), is probably better characterized as metallic and

astringent. The evolution of the phenolic constituents during ripening

(Clifford et al., 1987; Guyot et al., 1988a, 1988b) may give some

indication on the nature of this defect, which has been explained as being

due to an excess of dichlorogenic acids with respect to monochlorogenic

acids, particularly important before the last five weeks of ripening

(Clifford and Ohiokpehai, 1983). This may not be the full explanation

since recent studies have shown that 5-caffeoylquinic acid, the major

component of the family present in coffee, has an astringent taste of its

own (Naish et al., 1993) and so the question remains open. The

microscopic examination of immature beans confirms their incomplete

ripeness (Dentan and Illy, 1985).

The green grassy flavour perceived in some beans has been shown to be

due to incompletely mature beans, which have already reached the size of

ripe beans (Dentan, 1991). These beans appear to be normal, but are still

unripe, even though they are riper than immature beans.

A concentration of light-green immature beans greater than 15%

causes noticeable harm to the brew (Teixeira et al., 1969).

Unripe beans ferment very easily when exposed to temperatures around



C, producing dark-green immature beans or, worse, black-green

immature beans (Teixeira et al., 1982). It has been observed that lightgreen

immature beans are sometimes also fermented, which can be

perceived when the beans are scratched or during cupping.

Light-green immature beans are characterized by:

n light-metallic-green silverskin colour;

n metallic and astringent taste, reminiscent of the taste of some amino

acids, perceivable when the proportion of unripe beans exceeds 1% in


n vivid silverskin, adhering well to the bean surface;

n cell-walls much thinner than in ripe beans, with a lower cellulose


n absence of saccharose, a constituent of sound ripe beans;

126 Espresso Coffee

n presence of arabinose, absent in ripe beans;

n complete absence of serotonin, formed from tryptophane during


n greater than average content of chlorogenic acids (up to 17% in

unripe beans as opposed to 7% in ripe ones);

n reduced lipid content;

n reduced level of oleic acid and increased level of linoleic acid in total


n oleic/stearic acids ratio less than 1;

n rather marked minimum at 672 nm in the reflectance spectrum

attributable to the presence of chlorophyll. Black-green/dark-green immature beans

Black-green and dark-green immature beans are formed when unripe

beans (light-green immature beans) are dried at or exposed to an

excessive temperature, above 40


C (Teixeira et al., 1982). They have

black to dark-green silverskin colour and are characterized by astringency

and a taste reminiscent of rotten fish, attributed to high levels of cis-4-

The raw bean 127

Figure 3.21 Arabica beans

Above: Optical microscope images: (left) immature – many cells are still empty,

the cell-walls are thin, the lipids (red) surround proteins (yellow); (right) mature –

the cells are full; the lipids (red) have migrated towards the walls

Below: SEM images: (left) immature; (right) mature

heptenal (Full et al., 1999), which makes this a very serious defect. This

compound is a product of the auto-oxidation of linoleic acid (Grosch,

1998). An experiment was made of removing the silverskin from beans

with these defects and they were classified as sour beans (Teixeira et al.,


Black-green immature beans can be distinguished under the microscope

(Dentan, 1989) by:

n dark to black-green silverskin colour and wrinkled appearance;

n no microbial attack is apparent, and the content of the cell is

congealed into an undistinguished mass;

n total degradation of proteins in the cells of the surface layers, smaller

in the central area;

n absence of lipids in the peripheral cells, owing to lipase activity;

n very low reflectance; the minimum present in unripe beans is almost


n density well below that of a healthy bean, which facilitates

densimetric separation. Black beans

Coffee beans whose interior (endosperm) is partially (partially black

beans) or totally black. These beans are more frequently found in overripe

cherries on the tree or fallen on the ground, but can also be

encountered at other phases of maturation (Carvalho et al., 1972) or

during processing. More than 10% of black beans cause noticeable harm

to the brew (Gomes et al., 1967). They have a harsh and ashy flavour.

Black beans can be distinguished by microscopy (Dentan, 1989) by the

following characteristics:

n black beans, where the interior of the bean is also more or less

completely black depending on the severity of the attack, are beans

having undergone a yeast fermentation starting at the epidermis; the

surface of the bean is covered with minute holes surrounded by

mineral micro crystals, left after enzymatic degradation of cellulose.

The more serious the damage, the blacker is the interior of the bean.

When the fermentation has progressed long enough, the black beans

are just a light empty shell, which can be separated densimetrically

from normal beans;

n particularly in Africa, black beans, characterized by more or less

homogeneously black surface, are due to fungal attack by

128 Espresso Coffee

Colletotrichum kahawae (coffeanum); in this case the beans are often

infected by Aspergilli, Penicillia or other moulds;

n easy peeling off of the silverskin. Sour beans

Sour (ardido) beans are deteriorated by excess fermentation, with a sour

taste. They have a range of colours: light to dark reddish-brown, dark

brown or yellowish-green internally (endosperm) and sometimes have a

waxy appearance, typical of dead beans. This defect has different causes. A

study was made removing the silverskin from black-green immature beans

and they were classified as sour beans (Teixeira et al., 1971). Further

studies indicated that sour beans constitute one phase of coffee deterioration,

which ends by reaching a black colour (Gomes et al., 1967; Carvalho

et al., 1972). Microscopic examination indicates that they are ripe beans

probably killed by overheating during processing, or by infection with

xerophilic moulds, such as Aspergilli and Eurotia (Dentan, 1989, 1991).

They are characterized by:

n yellow-green to dark reddish-brown colour, sometimes with waxy


n sour and fermented smell and taste. Stinker beans

Stinkers are over-fermented beans, usually with normal appearance but a

rotten smell and flavour. Morphological analyses show that some stinkers

have lost the embryo. Microscopic examination often reveals multiple

contaminations with bacteria and moulds (Dentan, 1991). In one specific

case only one bacterium, Bacillus brevis, was identified. The defect has

been associated with the presence of excessive amounts of dimethylsulphide,

>0.60 mg/kg instead of <0.30 mg/kg found in healthy arabica

coffee, and of dimethyldisulphide, >0.40 instead of 0.15 (Guyot et al.,


Short chain aliphatic esters and acids, formed during lactic fermentation,

have also been associated with this defect, and a count of lactic acid

bacteria above 106/g is often associated with a fermented, rotten flavour.

A useful indicator of the presence of stinker beans is the development

of red colour by litmus paper suspended above a sample for an hour due to

the presence of acetic acid. This reaction is specific to stinker beans and

does not occur with other defects or with healthy beans, the faster the

reaction, the more serious the defect.

The raw bean 129

Under examination using ultraviolet light, stinker beans show a white

or white-blue fluorescence, probably associated with larger quantities of

free caffeic acid compared with healthy beans. Nevertheless, other defects

and old crop beans are also fluorescent, making it difficult to sort stinker

beans solely by this method.

Stinker beans can contaminate a large batch, even when present at

very low levels, turning healthy beans into stinkers at a lower level of

intensity (Gibson and Butty, 1975). Beans that have become stinkers only

by contact are not fluorescent. A further indicator is headspace

gaschromatography, which shows peaks that do not appear in healthy


This defect has been attributed to alcoholic fermentation by

Saccaromyces cerevisae.

Possible causes of stinkers are the following:

n too long fermentation;

n wild fermentation of beans trapped in the interstices of pulping

machines, fermentation tanks, etc.;

n wild fermentation with recycled fermentation water, contamination

with polluted water;

n abrasion during pulping, with loss of the superficial protective layer,

which renders the bean easily susceptible to attack by microorganisms;

n contamination of healthy beans by stinkers;

n delay in pulping;

n over-heating during processing: at 55


C, stinkers develop in 4 hours,

whereas at 30


C 4 days are necessary; the most critical processing

steps are the storage of cherries before processing, fermentation, and

the white stage of drying;

n exposure to over-heating in the sun of the heaps of coffee covered by

sheets during drying;

n presence of Ceratitis capitata (as larva) in the cherries. Rioy beans

These beans have a flavour described as medicinal and iodine-like. Rioy

beans were first detected in Brazilian coffee grown in the Rio de Janeiro

area and subsequently in other areas and countries. The substance

responsible for this important defect is 2,4,6-trichloroanisole (TCA),

which in coffee has an odour threshold of 8/ng/l and a flavour threshold of

1–2/ng/l. By direct olfaction it has been described as dusty, musty, earthy,

woody, corky, cereal, iodine-like, phenolic and an oral/retronasal

130 Espresso Coffee

perception of bitterness, burned, rubbery, phenolic, acrid, pungent,

earthy, stale and medicinal (Spadone et al., 1990). Rio-tainted beans

are heavily infested with moulds (Aspergilli, Fusaria, Penicillia, Rhizopus),

and bacteria (Lactobacilli, Streptococci), accompanied by a degradation of

the cell structure (Dentan, 1987). The levels of TCA found in Rioy coffee

are usually well above the threshold. 2,4,6-trichlorophenol (TCP) appears

to be the direct precursor of TCA in coffee, probably by microbial

degradation. The origin of TCP itself is still unclear: some authors

maintain that it derives from the chlorophenols used as fungicides, others

that it is biosynthesized by a yet unknown microorganism. Of the two

hypotheses, the latter is more likely, for only one TCP isomer has been

found in Rioy beans, whereas synthetic products should contain all the

possible isomers (Spadone et al., 1990). Furthermore, moulds with

chloroperoxidase activity are known. Aspergillus fumigatus, found in

Rioy beans but absent in sound beans, could be the source of the defect;

yet, studies carried out on wine, whose corky smell is also due to the

presence of TCA, have shown that Penicillia can synthesize TCA in the

presence of chlorine and methionine. Rioy beans are characterized

(Amorim et al., 1977, Dentan, 1987) as follows (Table 3.9):

n external appearance similar to that of a normal bean;

n full or excessive ripeness; the defect is never present in unripe beans;

n contamination with moulds (Aspergilli, Fusaria, Penicillia, Rhyzopus)

and bacteria (Lactobacilli, Streptococci) degrading the cellular structure;

n adjacent cells no longer attached to one another, and both the

volume and the thickness of the cell-walls well below those of

healthy beans;

n density lower than that of normal beans;

n different microbial populations present inside (Fusaria, Pseudomonas

and yeasts) and outside (Aspergillus versicolor, Wallemia sebi);

n low polyphenoloxidase activity;

The raw bean 131

Table 3.9 Morphological differences between sound and

rioy beans

Sample Bean weight (mg) Density Cell-wall thickness (lm)

Strictly soft 73.03 1.085 6.2

Soft 57.69 1.014 5.6

Rioy 58.21 0.967 5.0

Strong rio flavour 49.24 0.761 4.3

n lower than average content of hydrolysable phenols;

n lower than average content of hydrolysable proteins;

n absence of high and medium molecular weight (150 000 and 64 000

Daltons) proteins, usually present in sound coffees;

n level of low molecular weight proteins higher than in sound coffees;

n presence of some low molecular weight (9000 Daltons) proteins that

usually do not appear in sound coffees;

n UV-VIS reflectance spectrum identical to that of a normal bean.

Figure 3.22(a) shows that the cell-walls have been opened by moulds

forming channels (the white spots are moulds).

The origin of the defect could be due to several factors:

n cherries that are contaminated by moulds and/or specific bacteria

while still on the plant;

n the drying stage after harvesting is too slow;

n the drying patio soil is heavily contaminated by microorganisms and/

or TCA. Whitish beans

The surface discoloration of whitish beans is due to fermentation by

Streptococcus bacteria that in the most severe cases may also reach the

cells under the epidermis (Dentan, 1991). The attack can occur if storage

132 Espresso Coffee

Figure 3.22 (a) Degraded cell structure of a rioy bean; (b) electron microscope

section of a mould-invaded coffee bean

(a) (b)

is too long or in conditions of excessive humidity. White beans are not

very aromatic and give slightly bitter and woody cups. Under ultraviolet

radiation they show a blue fluorescence probably attributable to

chlorogenic acids and caffeic acid, which makes them difficult to sort

out from stinkers because of the fluorescent backgrounds they produce. Mouldy beans

A mould/yeast level above 105/g is always associated with mustiness in

flavour. Geosmin, identified in a heavily rioy and musty tasting sample of

Portorican coffee, is probably the substance responsible for mouldiness in

beans (Spadone et al., 1990).

These beans have a greyish colour, and under microscopic examination

the epidermis, silverskin and central cut appear covered with moulds,

mostly Aspergilli (A. tamarii, A. niger, A. ochraceus and A. Flavus,

essentially). Pseudomonas bacteria living in symbiosis are also visible.

Although the cell-walls are intact, only the peripheral lipids remain inside

(Dentan, 1991). In Figure 3.22b hyphes formed by Aspergillus moulds

are clearly visible in the central split. The presence of moulds or yeasts

(more than 105 units per gram) always produces a putrid smell due to the

presence of geosmin. The cause of contamination can be either inadequate

drying or storage in overly humid and poorly ventilated conditions. Earthy beans

The presence of 2-methylisoborneol, a secondary metabolite of

Actinomycetes, Cyanobacteria and moulds, has been associated with the

earthy flavour of robusta coffee, with a threshold level in water of

1–100 ng/l, and a flavour described as earthy, musty, robusta-like.

The levels present in robusta coffee are at least three times as high as in

arabica (Vitzthum et al., 1990). These data indicate that robusta taste

results, at least partially, from contamination by microorganisms rather

than from specific aroma components.

Defective beans may have colours varying from clear brown to almost

black; the silverskin, but neither the epidermis nor the central cut, is always

infected by various microorganisms, bacteria, yeasts and moulds, Fusarium,

Geotrichum, Eurotium, are often the major populations present, as well as

different Aspergilli (A. flavus, A. fumigatus, A. niger). Although the cell

walls are intact, only the peripheral lipids remain inside (Dentan, 1991).

A patent has been registered for a method of eliminating 2-methylisoborneol

from robusta. By this method, coffee is treated with saturated

The raw bean 133

vapour at 138


C, at a pressure of 3.8–3.9 bar, for 75–90 minutes

(Vitzthum et al., 1990). Peasy beans

This defect, encountered only in Central African arabica coffees, has been

identified as 2-isopropyl-3-methoxypyrazine with a perception threshold of

about 300 ppb in air and 0.1 ppb in water (Becker et al., 1987). The defect

is due to a contamination of the cherry by a bacterium of the

entherobacteriaceae, probably transmitted by Antestiopsis orbitalis ghesquierei

feeding on the cherry (Bouyjou et al., 1993). In the most seriously defective

beans, the concentration of 2-isopropyl-methoxypyrazine can reach 2500

ppb as against 70–90 ppb in non-defective beans (a ratio of 35:1).

Peasy beans have a smell strongly reminiscent of fresh green peas.

Morphologically, peasy beans appear normal, though they show some




R. Teixeira, A.A. Teixeira and C.H.J. Brando

The classification of coffee, which includes physical and sensorial analysis,

is a very important phase in the commercialization process. Price

quotations, as well as national regulations governing importation into

consuming countries, are established on the basis of such classifications.

In order to discuss the quality and production of fine coffees, there is a

fundamental need for technicians, producers and purchasers to be aware

of the factors that enter into the evaluation of the final item, and to be

able to judge its merits and demerits. Unfortunately, the existence of a

variety of classification systems means that each country adopts a different

classification, requiring equivalency norms for use at the international

level, such as ISO 10470-1993 – Green Coffee – Defect Reference Chart

(ISO, 1993) and ISO 4149-1980 – Green Coffee – Olfactory and Visual

Examination and Determination of Foreign Matter and Defects (ISO,

1980), which are currently being revised.

In Brazil, it is customary to classify by type or defects as well as by cup

quality. In Colombia, coffee is classified by the characteristics of the

plantation, with regard to altitude, by bean size and by region of origin.

Countries of Central America (El Salvador, Honduras, Mexico,

Nicaragua, Guatemala, Costa Rica and Panama) have altitude as the

main criterion, but also take into account appearance, bean size and cup

134 Espresso Coffee

quality. In Africa, the classification varies according to the area of

production. Angola, Cameroon and the Ivory Coast have adopted systems

similar to that of Brazil, which uses defects to establish types. In Kenya

and Tanzania, coffee is classified by ranking in classes, in accordance with

bean size and cup quality, in some of which defects are permissible.

Indonesia classifies its coffees in accordance with the species, method of

preparation, origin and number of defects (Jobin, 1982).

Overall, classification is based on the evaluation of all of the

parameters related to the coffee bean and each one of them must be

examined, separately, by physical methods and by sensory analysis

(Table 3.10).

3.8.1 Classification by species and varieties

Coffee is classified by species and variety of origin. Only two easily

distinguished species are important in economic terms: arabica and robusta.

Arabica fetches a higher price in the commercial market, is more delicate,

demands greater care in its culture, and possesses superior organoleptic

characteristics in comparison with robusta, which is more resistant, has a

lower cost of production, but results in an inferior cup quality.

3.8.2 Classification by screen (size and shape)

Different producing countries use different classifications of bean size. In

Brazil, beans are classified according to size and shape. With regard to

shape, the coffees receive the denomination of chatos or flats (long shape),

and peaberries (round shape), also called mocas or caracoles. The size or

The raw bean 135

Table 3.10 Parameters used in commercial classification

Parameters Description

Species and varieties Arabica /robusta

Classification by screen size Size and shape of the beans

Classification by type (number of defects) Defective beans and foreign matter

Density Specific weight of the beans

Appearance and dryness Bean uniformity

Colour Coloration of the beans

Processing Natural (dry), pulped natural and washed


Roast Roast regularity, smell, etc.

Cup quality Characteristic aroma and flavours

screen of the beans is measured by the dimension of the holes in the

screen that holds them back, designated by numbers that, divided by 64,

express the size of the holes in fractions of an inch (see 3.5.3 above).

Screens with round holes are used for chatos (varying from 19 to 10) and

ones with elongated holes are used for peaberries (13 to 8). Some

countries use screens with holes measured in millimetres.

Separation by screens is important, since it allows the selection of beans

by size, separating them into groups adequate for uniform roasting. In

Colombia, coffee is classified as ‘Supremo’, ‘Excelso’ and ‘Pasilla’ (the

latter not exportable); in Kenya the designations used are AA, AB, PB, C,

E, TT, T, UG, which indicate size and shape (Jobin, 1982).

3.8.3 Classification by type (Figure 3.23)

Most producing or importing countries have their own classification

system for defects. Commercial classifications are established in accordance

with the number of visual defects in a certain sample. At the

international level, the ‘New York Coffee and Sugar Exchange’

introduced the concept of black bean equivalent. This system uses the

black bean as a standard of measurement for all defects (for example: 1/2

means 2 defects are equivalent to 1 black bean). Other bodies have

introduced similar scales for counting green coffee defects (Table 3.11).

The International Standardisation Organisation has established several

norms regarding olfactory and visual examinations, determination of

defects and foreign matter (ISO, 1980), determination of the proportion of

beans damaged by insects (ISO, 1985), and the defect reference table

(ISO, 1993), that deal with the majority of defects. Currently, several

norms are being revised, including the defect reference table now in the

draft stage, where each defect is defined and characterized according to its

degree of influence (minimum, medium or maximum) in relation to the

136 Espresso Coffee

Figure 3.23 Classification of coffee: (left) size determination; (right) olfactory


The raw bean 137

Table 3.11 Equivalency ratings of green coffee defects in different countries (ISO, 1993)

Defect USA UK France Brazil Tanzania Indonesia Ethiopia

Brazils Centrals Colombian Liffe 1965 1992 1982 1973

Dried cherry (pod) 1 1 1 1 1 1 1 1 2

Black bean 1 1 1 1 1 1 1 1 2

Semi-black bean – 1/2–1/5 1/2–1/5 – 1/2 – 1/2 1/2 1/2

Sour bean 1 1 1 1/2 1 1/2 1/8 – –

Insect-damaged 1/5–1/10 – – 1/2–1/5 1/10 1/2–1/5 1 1/5–1/10 1/2

Immature bean 1/5 – – 1/5 1/5 1/5 1/2 1/5 1/5

Floater (white light) 1/5 1/5 1/5 1/5 1/3 – – – 1/5

Bean in parchment – – – – 1/2 1/2 1/5 1/2 –

Broken bean

More than half 1/5 1/5 1/5 1/5 – 1/5 1/5 1/5 1/10

Less than half 1/5 1/5 1/5 1/5 1/5 1/5 – 1/5 1/5

Shell 1/3 1/3 1/5 1/5 1/5 1/3 1/5 – –

Husk fragments

Large – – – 1/2 1 1 1 1 3

Medium 1/2 1/2 1/3 – – 1/2 1/2 1/2 1

Small – – – 1/5 – 1/3 – 1/5 1

Parchment fragments

Large – – – 1/2 – – – 1/2 –

Medium 1/2 1/2 1/3 – – – – 1/3 1

Small – – – – 1/3 – – 1/10 –


Large 2–3 2–3 2–3 5 2 5 5 5 10

Medium 1 1 1 2 1 2 2 2 5

Small 1/2–1/3 1/3 1/3 1/2 1/3 1 1 1 3


Large 2–3 2–3 2–3 5 – 5 5 5 10

Medium 1 1 1 2 – 2 2 2 5

Small 1/2–1/3 1/3 1/3 1/2 – 1 1 1 3

Figures in columns give defect values (i.e. 1/2 means: 2 defects equivalent to 1 black bean as the reference).

loss of mass and to organoleptic properties (sensorial concern). A formula

is being introduced that intends to measure the total impact on quality, to

be calculated by multiplying the percentage of defects in a sample by the

coefficient of sensorial concern, according to the values found in the table.

Brazilian type classification, widely used in the international trade,

provides for seven types with decreasing values from 2 to 8 resulting from

the analysis of a 300 g sample of milled coffee, according to the norms

established in the ‘Brazilian Official Classification Table’ (Table 3.12).

Each type has a greater or lesser number of defects consisting of beans that

have been altered in the field, harvesting, processing or storage (e.g.,

physiological or genetic origin and presence of foreign matter).

3.8.4 Appearance and humidity

Appearance and humidity are fundamental factors in the evaluation of

quality, since they serve to predict a good or bad roast and other indications

of the quality of the final product. Appearance is good when the majority of

the beans are perfectly formed, uniform in size, colour and humidity, and

bad when the majority of the beans are not uniform and defective beans are

found. In coffee that has been correctly dried, the humidity level should be

11 0.5% for natural coffees, washed and pulped natural coffees. Incorrect

drying can be identified by spotted or humid beans.

3.8.5 Colour

Colour is always associated with quality. A bluish-green colour is very

desirable in washed coffee, being considered a sign of high quality and

freshness, while a yellowish colour is a sign of old coffee and low quality.

138 Espresso Coffee

Table 3.12 Official Brazilian classification of

green coffee (Classificac¸a˜o Oficial Brasileira – COB)

Grade Number of defects in 300 g of coffee

COB 2 4

COB 3 12

COB 4 26

COB 5 46

COB 6 86

COB 7 160

COB 8 360

Unsuitable for export >380

Factors that contribute to colour variations are: degree of dryness, time of

exposure to light, processing method, storage conditions, bruising,

polishing, etc. The polishing of the bean improves its appearance but

hides defects, due to the removal of the silverskin (Teixeira et al., 1985).

The ISO recommends the following colour classification: blue, greenish,

whitish, yellowish and brownish. The Brazilian classification adopted for

export purposes is: green, greenish, pale, yellow and old (Teixeira et al.,


3.8.6 Processing

Coffee can be classified according to processing method as natural (dryprocessed),

pulped natural, and washed (wet-processed) coffee. The

processing system can be recognized by the bean colour and by the

appearance of the silverskin. Washed coffees possess a characteristic

shiny, translucent and green-bluish colour. Natural coffees have a semiopaque

colour and a yellowish or even brown skin. Pulped natural coffee

has an intermediate aspect.

3.8.7 Roast

The analysis of the roast is an important step, and an important gauge of

quality. Defects that are not observed in the raw beans come out in the

roast. The immature and sour beans become yellowish while black beans

appear to be burnt. Broken beans, shells and shell core beans, due to their

reduced volume in relation to perfectly formed beans, become darker. In

Brazil, the roast is classified in accordance with appearance and, for

natural coffees, it can be measured as: fine, good, regular and bad

according to the degree of uniformity. The roast of washed and fermented

coffee is considered to be characteristic when the majority of the beans

present a clear and distinct silverskin in the ventral ridge of the bean

(Teixeira and Ferraz, 1963). The roasting of natural coffee presents a

brown silverskin. Pulped natural coffees and those with mechanically

removed mucilage present a roast intermediate between those of natural

and washed coffees.

3.8.8 Cup quality (Figure 3.24)

Even today, greater importance is attached to physical rather than

organoleptic classification. A possible reason is the fact that when these

criteria were introduced well-selected washed arabicas prevailed in the

The raw bean 139

market and such a classification corresponded closely to the quality of the

coffee in the cup. Another reason is that classification systems were

introduced many years ago when knowledge about the subject of cup

quality was limited. The obsolescence of classification systems can

sometimes lead to paradoxical situations, where good coffees are classified

as low grade, due to bean size, while mediocre coffees are well evaluated

and expensive, even if they are poor from an organoleptic point of view.

According to the International Trade Centre (ITC, 2002) the cup or

liquor can be characterized by the following terms:

n ‘Acidy – A desirable flavour that is sharp and pleasing but not biting.

The term ‘acid’ as used by the coffee trade refers to coffee that is

smooth and rich, and has verve, snap and life as against heavy, old

and mellow taste notes.’

n ‘Aroma – Usually, pleasant-smelling substances with the characteristic

odour of coffee.’

n ‘Body – A taste sensation or mouth feeling . . . of a drink

corresponding to a certain consistency or an apparent viscosity . . .

Sought after in most if not all coffees.’

Organoleptic characteristics vary according to the producing country and

this parameter must be considered as specific to each commercial origin.

In general, all washed arabica coffees from Eastern Africa (such as Kenya,

Tanzania and Ethiopia) as well as from Central and South America (such

as Guatemala, Costa Rica and Colombia), are marked by a degree of

acidity and an intense aroma, although each origin has different

characteristics. Natural dry-processed arabica coffees, especially those

from Brazil, are less acid and have a less marked aroma but are endowed

with a richer body ideal for espresso coffee. Finally, robusta coffees are

characterized by a wooden and earthy flavour.

Classification systems that take organoleptic characteristics into

account are used in some producing countries; unfortunately, this type

140 Espresso Coffee

Figure 3.24 Cup tasting

of classification has only a small influence on the price. Brazilian coffees

are classified as to cup quality according to the following ranking (from

best to worst): strictly soft, soft, softish, hard (or hardish), rioy and rio.

Kenyan coffees are graded by size and density into seven grades, divided

into 10 classes by liquoring (fine, fair, fair to good, etc.). Colombia and

Central American countries take altitude into consideration (Jobin,

1982). In addition to this lack of uniformity in classification systems, each

producing country weighs organoleptic defects in a different way, which

may lead to discrepancies in sensorial evaluation between producing and

consuming countries.


G. Brumen

The blending of coffee is as old as coffee itself. Although the techniques

vary, blending is used to optimize aroma, body and flavour: the goal is to

make a coffee that is higher in cup quality than any of the ingredients

individually, and, extremely important, maintain consistency in the final

roasted product.

Each batch has it own personality in terms of taste, smell, body,

chemical resilience to the hydrolytic action of water, etc., and blending

can complete it and round it up or level it off.

Most espresso blends are based on high quality Brazil arabicas, some

washed, some dry-processed. They often involve some African coffees for

winey acidity or flowery fruitiness, or a high-grown Central American for

a clean acidity. Some roasters add a little robusta to increase body.

Dry-processed coffees are responsible for the attractive ‘crema’ on the cup,

among other mechanical factors in the extraction process (see 8.1.1). Wetprocessed

Central Americans add positive aromatic qualities. Robustas

are used in cheaper blends to increase body and produce more foam.

Besides subjective quality (see Chapter 1), blending also assists in

maintaining objective quality, because the more complex a blend, the

easier it is to maintain constant quality when some ingredients change.

With the exception of a few countries that pay considerable attention

to quality, the majority of producer countries often add up small batches

produced by different growers to form larger ones of a size required by

roasters. Although care is taken so that only batches of equivalent quality

are blended, the result of this deplorable practice is often a quality

downgrading to a level below that of the best fractions.

Coffee history records a number of popular blends that are published

and available for public consumption. Other ‘proprietary’ blends tend to

The raw bean 141

be closely guarded, with the information staying within a company

structure. Proprietary or signature blend leads consumers to equate a

particular coffee profile with a particular brand image. Blending requires

the expert skill of knowing each ingredient coffee, having in mind a clear

cup profile as the goal, and knowing how to achieve it.

Blending may be done before or after roasting. Blending before roasting

is traditionally used by retail and institutional roasters. In this method

coffees with similar characteristics are combined and roasted to the same

development. Generally, professional in-house ‘cuppers’ evaluate the

results of the blend, adjusting components if necessary to satisfy taste

requirements and standards.

n Advantage: Consistency of product.

n Disadvantage: Inability to optimize the character of each coffee.

Blending after roasting is the method traditionally used by many speciality

coffee roasters. The flavour profile development requires that each

individual coffee used in the blend be roasted separately to optimize

flavour. In other words, each coffee will have a different time and

temperature setting. Consequently, the final roast development will be

different for each coffee used in the blend. After roasting, each

component of the blend is individually tasted (cupped), as is the final

blend composition.

n Advantage: Ability to optimize the character of each coffee.

n Disadvantage: Inconsistency of product.


O.G. Vitzthum

A major impact compound, as well as the pleasant aroma and flavour

components in roast coffee, is caffeine; it is the physiologically most

active substance in coffee.

Caffeine stimulates the central nervous system, and is responsible for

the vigilant effect of the coffee beverage. However, some people can suffer

from insomnia and restlessness after drinking coffee in the evening.

People who habitually drink coffee may develop a certain tolerance to the

drink, whereas non-coffee drinkers experience a stimulating effect even

from a single cup of coffee (for a discussion of the physiological effects of

caffeine see 10.3.2).

142 Espresso Coffee

It was the German writer J.W. von Goethe who, in 1820, gave some

coffee beans to the chemist Runge (1820) requesting him to isolate the

pharmacologically active compound therein. He found the alkaloid

caffeine, a colourless, slightly bitter-tasting substance. The name was

derived from the botanical name Coffea for coffee.

The idea of reducing or eliminating this physiologically active ingredient

in coffee originated at the end of the nineteenth century. It was after the

frustrating experiments of his chemist Meyer, who tried to extract the

caffeine out of green coffee beans with caffeine-dissolving solvents, that

Bremen coffee merchant L. Roselius had the successful idea of using raw

coffee beans pre-swollen with water for decaffeination with solvents

(Roselius, 1937). He knew that by using a steam treatment the volume of

green coffee could be increased by about 100%. Consequently the solvent

was able to penetrate more easily into the swollen beans and dissolve the

caffeine. Water-immiscible solvents are necessary to avoid the extraction

of the flavour precursor water-soluble components. The hydrophilic

caffeine fortunately dissolves well both in water and in various solvents.

The first commercially decaffeinated coffee was produced in Bremen as

Kaffee HAG at the beginning of the twentieth century. Due to the patent

protection for this invention, many attempts were made in the early 1920s

and 1930s to apply other solvents and processes in search of a legal

circumvention of this technique. Until 1970 two solvents were successfully

in use for the decaffeination of raw coffee beans or aqueous extracts

thereof: methylene chloride or dichloromethane (DCM) and ethylacetate

(EA). A new approach for the decaffeination of coffee was initiated by

Zosel (1971) when, in 1970, he suggested the use of supercritical carbon

dioxide. This activity spurned the development of other processes such as

water extraction and the use of adsorbents or fats and oils.

Nowadays conventional processes such as extraction with DCM and

EA are in use as well as various technologies based on decaffeination with

carbon dioxide or water and certain extraction aids.

3.10.1 Processing

A comprehensive description of decaffeination principles with technical

details has recently been published by W. Heilmann (2001). Here, only

the various decaffeination techniques will be described. Conventional decaffeination

The procedures for decaffeination by the solvents dichloromethane

CH2Cl2 (DCM) and ethylacetate CH3COOC2H5 (EA) are similar,

The raw bean 143

although more coffee solids are lost during EA decaffeination. There are

direct and indirect solvent-based decaffeination processes.

In the direct solvent decaffeination process (Figure 3.25), after

cleaning, green beans are swollen by addition of water or steam into

the extraction vessel. Thereafter solvent is added and decaffeination

started by recycling of the solvent under elevated temperature conditions

(70–100 oC). Processing continues for 8–12 hours, whereby the solvent is

continually replaced by fresh re-distilled material. In a separate vessel the

caffeine-rich solvent is distilled off and led back to the decaffeination

vessel. After a corresponding refining process, the crude caffeine is pure

enough for use in caffeinated beverages. Subsequent to completion of the

decaffeination process, the residual solvents are eliminated by steaming.

The beans are dried and are then ready for roasting. Decaffeinated beans

are usually lighter in colour than the original green raw coffee beans.

In the indirect solvent decaffeination process (Figure 3.26), beans do

not come into direct contact with DCM (Berry and Walter, 1943). A

saturated aqueous extract of green beans, containing all water-soluble

components including caffeine, is decaffeinated by the solvent in a

liquid–liquid extraction process in a separate vessel. The caffeine-free

144 Espresso Coffee

Figure 3.25 Direct solvent decaffeination

extract after decaffeination is returned to the green beans and the process

repeated using a new batch of green beans. Any traces of solvents present

in the returned caffeine-free extract must be removed by subsequent

steaming. After drying, the beans are ready for roasting. Modern decaffeination

A revolution in decaffeination evolved with the invention of caffeine

removal by supercritical carbon dioxide (Figure 3.27) (Martin, 1982;

Lack and Seidlitz, 1993). This process uses the specific solvation power of

gaseous carbon dioxide under elevated pressure conditions for caffeine.

CO2 is in a supercritical physico-chemical state above its critical

temperature of 31 oC and, although still a gas, it has nearly the density

of a liquid if pressures of about 200 atm are applied.

In comparison with conventional solvents, the solubility of caffeine in

supercritical CO2 is low; but multiple recycling of the gas through the

green beans eventually leads to complete caffeine extraction. Caffeine is

removed in every cycle from the gas stream by an effective adsorber such

as active carbon or an ion-exchange resin. The advantage of the process

lies in the fact that CO2 is highly selective in extracting the caffeine out

of the beans so that components that contribute to the aroma-yielding

fraction in coffee are not solubilized.

The raw bean 145

Figure 3.26 Indirect solvent decaffeination

The indirect adsorbent decaffeination process (Figure 3.28) (Fischer

and Kummer, 1979; Blanc and Margolis, 1981) is similar to indirect

solvent decaffeination, except that the solvent is replaced by an extract

that is saturated with the water-soluble compounds of the coffee material.

Such a heated aqueous extract extracts caffeine from the fresh beans;

thereafter the caffeine is selectively removed from the solution in a

separate vessel by a ‘caffeine-selective’ adsorbent such as coated active

carbon or non-ionic microporous resins. These serve to more or less

selectively extract all the caffeine but not the aroma precursor containing

water-soluble substances. This extract, which now contains no caffeine, is

recycled back to the extraction vessel and caffeine is extracted again. The

treated active carbon may be pH-adjusted or coated by carbohydrates to

block the hydrophilic sites on the adsorbent. After use the adsorbents

must be regenerated accordingly. Organoleptic comparison of the different techniques

Although slight differences in taste are noticeable between the same

blend, decaffeinated by MC, supercritical CO2 or water, if the

decaffeination process and the roasting are carried out correctly the

difference in the cup with the corresponding non-decaffeinated blend is


146 Espresso Coffee

Figure 3.27 Supercritical CO2 decaffeination

If the process is not carried out under optimal conditions, typical defects

appear at cupping:

n MC-decaffeinated coffee has the ‘cooked’ flavour, associated with

decaffeinated coffee.

n CO2-decaffeinated coffee is flat.

n Water-decaffeinated coffee is thin, due to the loss and/or poor reincorporation

of the soluble solids co-extracted with caffeine (this

defect is often masked by increasing roasting level).

n EA-decaffeinated coffee has always the typical ‘cooked’ flavour

associated with (poorly) decaffeinated coffee.

3.10.2 Perspectives for caffeine elimination

Whereas no major developments are expected for decaffeination

processing in the near future, the application of molecular biology and

biotechnology to coffee may yield new naturally caffeine-free coffee plants

within 5–10 years. The deciphering of sequences of coffee genes is

progressing and biochemical pathways for the formation of caffeine in the

plant are being elucidated (Stiles et al., 2000).

The possibility of cultivating coffee plants without caffeine by

recombinant DNA technology is approaching. Ogawa et al. (2001)

The raw bean 147

Figure 3.28 Indirect adsorbent decaffeination

recently reported on the successful cloning of a corresponding methyltransferase,

which is an important regulator enzyme in the biochemical

pathway of caffeine formation in the plant.

3.10.3 Regulatory aspects Caffeine residue

In European countries the maximum allowed caffeine content in coffee

beans is 0.1%, in soluble coffees, 0.3%. In the USA at least 97%

decaffeination (equivalent to 99.7% caffeine-free coffee) is common in

decaffeinated coffees. Solvent residue

No regulations exist for indirect decaffeination via coated activated

carbon or fats and oils. No restrictions are necessary for ubiquitous liquid

or supercritical carbon dioxide.

The following regulations regarding solvent residues in roasted coffee

beans exist for dichloromethane:

n 2–5 ppm in Europe, 10 ppm in South America. Practical observations

under the provision of good manufacturing practices (GMP) yielded

residues of <1 ppm in European DCM decaffeinated coffees.

n GMP have been requested specifically for EA; in Italy the residues for

EA are restricted to 15 ppm in roasted coffee.

3.10.4 Decaffeination capacity

The approximate (best estimate) decaffeination capacities from 1998 are

given in Table 3.13 (opposite). Consumption of decaffeinated coffees is

highest in Germany, Switzerland and the USA (12–18%).


I. Ko¨ lling-Speer and K. Speer

The beans of Coffea arabica (arabica coffee) and Coffea canephora (robusta

coffee) vary in mass from 100 mg to over 200 mg, with some evidence of

variation associated with the geographical origin. Both species show

quantitative and qualitative differences in their chemical compositions.

Arabica contains more lipids and trigonelline, while robusta contains

more caffeine and chlorogenic acids. Minor constituents specific to one

148 Espresso Coffee

species have also been identified. The composition of raw coffee beans is

indicated in Table 3.14.

3.11.1 Water content

The water content of green beans, from dry or wet processing, influences the

water activity and stability during storage and varies in Europe from 9 to 13%.

Dry-processed robusta coffees (Vietnam) sometimes show higher values. For

The raw bean 149

Table 3.13 Estimated decaffeination capacity by country

Country Solvent used



Germany DCM; supercritical CO2; liquid CO2; EA 160 000

France DCM; water 45 000

Italy DCM; EA 16 000

Portugal DCM 5 000

Spain EA; water 18 000

Switzerland DCM; EA; water 8 000

Total Europe 252 000

Canada DCM; water 12 000

USA DCM; EA; water; supercritical CO2 120 000

South America DCM; EA; water; fats and oils 50 000

Asia DCM; EA 6 000

World total 440 000

Table 3.14 Composition of raw coffee

beans (% dry matter)

Class Arabica Robusta

Caffeine 1.2 2.2

Lipids 16 10

Chlorogenic acids 6.5 10

Aliphatic acids 1.7 1.6

Total aminoacids 10.3

  free 0.5 0.8

Trigonelline 1.0 0.7

Glycosides 0.2 Traces

Minerals 4.2 4.4

  potassium 1.7 1.8

Carbohydrates (by difference) 58.9 60.8

storage and transportation the moisture level proposed by the European

Coffee Federation (ITC, 2002, p. 254) is 12.5%. Below 10% the germinating

power decreases and, furthermore, the beans show crack formation. At water

contents >12.5% there is a significant risk of microbiological spoilage. An

international reference method for moisture analysis (ISO, 1983) should be

used to standardize the rapid methods now available.

3.11.2 Ash and minerals

Coffee contains about 4% of mineral constituents (arabicas: 3.6%–4.5%;

robustas 3.6%–4.8%), with potassium amounting to 40% of the total. The

level is slightly higher in dry-processed robusta and arabica than in wetprocessed

arabica (see Table 3.2). The main minerals of raw coffee beans

are given in Table 3.15; fertilizing may change the contents (Amorim

et al., 1973).

Coffee has the highest content of rubidium, which has been analysed in

food (arabica: 25.5–182 mg/kg dry matter; robusta: 6.6–95.2 mg/kg dry

matter). Quijano Rico and Spettel (1975) determined higher content of

copper, strontium und barium in robustas than in arabicas.

3.11.3 Carbohydrates

The total amount of carbohydrates represents about 50% dry base of green

coffee. The composition is complex with a range of different poly-, oligoand

monosaccharides (Table 3.16).

Lu¨llmann and Silwar (1989) analysed the low molecular weight

carbohydrate profile of 20 green coffee samples from 13 different producer

countries. Sucrose content varied from 6.25% to 8.45% for the arabica

samples, and from 0.9% to 4.85% for the robusta samples. Besides sucrose,

150 Espresso Coffee

Table 3.15 Mineral content

of raw coffee (dry base)

Mineral Percent

Potassium 1.63–2.00

Calcium 0.07–0.035

Magnesium 0.16–0.31

Phosphate 0.13–0.22

Sulphate 0.13

small amounts of the monosaccharides fructose, glucose, mannose,

arabinose and rhamnose have been identified, traces of reducing sugars

being higher in robustas.

There was no evidence in green coffees of other simple oligosaccharides

such as raffinose or stachyose. Maltose was determined only in one robusta


It would be expected that sucrose content would increase with degree

of ripening and this is apparent with defective coffee beans, where for

both immature-black and immature-green Brazilian robusta beans,

sucrose levels were found to be one-third and one-fifth of the level of

normal beans, respectively (Mazzafera, 1999). This was also observed

for Vietnamese robustas, where black beans sorted from a sample

contained 0.9% sucrose; normal beans from the same batch contained

4.0% sucrose.

Glucose content has been negatively correlated with aroma level and

positively with cup sweetness, while fructose content has been negatively

correlated with sweetness (Illy and Viani, 1995). Franz and Maier (1993,

1994) identified inositol tri-, tetra-, penta- and hexaphosphate (phytic

acid). Total inositol phosphate content in robusta tended to be higher

than in arabica, ranging from 0.34 to 0.40% and 0.28 to 0.32%,

respectively. Polyalcohols have also been found in green coffee. Noyes

and Chu (1993) found low levels of mannitol in blends of Brazilian

robustas and arabicas (average content 0.027%).

The chemical structure of green coffee polysaccharides has been

extensively studied by Bradbury and Halliday (1987, 1990) and Fischer et

al. (2000). The soluble fraction is composed of sucrose and of polymers of

galactose, arabinose and mannose (galactomannans and arabinogalac-

The raw bean 151

Table 3.16 Carbohydrates in

raw coffee (% dry base)

Constituent Arabica Robusta

Monosaccharides 0.2–0.5

Sucrose 6–9 3–7

Polysaccharides 43.0–45.0 46.9–48.3

Arabinose 3.4–4.0 3.8–4.1

Mannose 21.3–22.5 21.7–22.4

Glucose 6.7–7.8 7.8–8.7

Galactose 10.4–11.9 12.4–14.0

Rhamnose 0.3

Xylose 0–0.2

tans). The insoluble cell constituents include ‘holocelluloses’, which

contain, in addition to some cellulose, the very hard  -1,4-mannan, and a

small amount of ‘hemicellulose’, mainly arabinogalactan. Other polysaccharides

common to the plant kingdom such as starch or pectin are

present, if at all, at only low levels in mature coffee beans. No lignine,

characterized as polymers of xylose, is present in the seed.

3.11.4 Glycosides

A series of diterpene glycosides, the atractylglycosides, was identified in

green and roasted coffee in the 1970s by Spiteller’s group (Ludwig et al.,

1975; Obermann and Spiteller, 1976). The contents in coffee of the

principal components of this compound class, termed KAI, KAII and

KAIII were determined by Maier and Wewetzer (1978) and Aeschbach et

al. (1982). Bradbury and Balzer (1999) have shown that these glycosides

in green coffee actually contain two carboxyl groups at C-4, and

consequently, the compounds were identified as carboxyatractylglycosides.

The carboxyl group is labile, thus explaining why earlier studies

only observed the atractyl forms. Arabica beans contain significantly

more of the glycosides than robusta beans.

3.11.5 Carboxylic acids

Data on acids in green coffee (Table 3.17) came from Kampmann and

Maier (1982), Scholze and Maier (1983, 1984), Engelhardt and Maier

(1984, 1985) and Wo¨hrmann et al. (1997). The total content is similar in

arabica and robusta coffees (average 1.7%) according to Van der Stegen

and Van Duijn (1987).

Quinic acid, the aliphatic acid moiety in the chlorogenic acids, is also

present in the free state. Its content is presumably affected by factors such

as processing, fermentation and age. Free quinic acid content can increase

to 1.5% in old beans. Besides quinic acid, the major acids in green coffee

are malic and citric acids and phosphoric acid. Minor coffee acids were

identified by Ba¨hre (1997) and Ba¨hre and Maier (1999).

The acidity of brewed coffees is an important organoleptic character

and is associated with the best high grown arabicas, such as Kenyans.

3.11.6 Chlorogenic acids

The chlorogenic acids are a group of phenolic acids esterified to quinic

acid (Clifford, 1999; Parliment, 2000). This class of acidic compounds

152 Espresso Coffee

accounts for up to 10% of the weight of green coffee. Mono- and dicaffeoylquinic

acids have been identified in coffee with substitution at the

3-, 4- and 5-position of quinic acid. The phenolic acid fraction of coffee is

composed of caffeoyl-, p-coumaroyl- and feruloyl-acids (Figure 3.29, see

page 154).

In coffee these components are essentially mono- and diesters. The

monoesters decrease during the unripe to semi-ripe stage then increase to

the ripe stage. This is followed by a decrease in the slightly overripe stage,

more important with wet than with dry processing (Castel de Menezes

and Clifford, 1988).

3.11.7 Amino acids, peptides and proteins

Free amino acids are present in raw coffee beans at levels about 1%

(Trautwein, 1987). Post-harvest treatments influence the contents of

individual free amino acids. This has been estimated for eight samples of

arabica by Arnold (1995): The total free amino acids do not show clear

changes after drying at 40


C, but the individual contents change for some

acids, especially glutamic acid, which shows an increase of about 50%,

and aspartic acid, which mainly decreases; the hydrophobic acids (valine,

The raw bean 153

Table 3.17 Contents (g/kg) of quinic, malic, citric

and phosphoric acids of different raw coffees

Species Origin









Arabica Santos 5.6 6.1 13.8 1.1

Burundi 5.7 5.1 13.0 1.1

Kenya 4.7 6.6 11.6 1.4

Colombia 5.5

Mocca 4.6 10.5 1.5

Robusta Burundi 3.5 3.8 10.0 1.4

Angola 2.8 9.2 2.2

Togo 3.1 2.5 6.7 1.6

Guinea 3.9

Arabusta Ivory Coast 4.5 9.1 1.7

Excelsa Ivory Coast 5.2 10.9 1.5

Stenophylla Ivory Coast 2.9 10.0 2.2

Liberica Unknown 3.9 10.6 1.8

Source: Kampmann and Maier, 1982; Scholze and Maier, 1983, 1984

phenylalanine, leucine, isoleucine) generally increase. Lipke (1999)

isolated and characterized several coffee peptides.

During storage of the green beans, particularly at elevated temperatures,

there are changes due to proteolysis (e.g. increases in alanine,

isoleucine and tyrosine) and to losses of free amino acids by nonenzymatic

browning reactions (Pokorny et al., 1975).

The proteins consist of a water-soluble (albumin) and a water-insoluble

fraction, present in approximately equal amounts. The majority of the

proteins have molecular weights above 150 000 Daltons. Crude protein

content, calculated from total nitrogen content, must be corrected for

caffeine and ideally also for trigonelline nitrogen. If such corrections are

made the protein content in green coffee is close to 10% with little

quantitative and qualitative differences between species and there is no

significant effect that can be attributed to green bean processing.

Mazzafera (1999) found a higher protein content in mature beans than

in the immature ones.

Several enzymes have been characterized in green coffee:  -galactosidase,

malate dehydrogenase, acid phosphatase, peroxidase and more

154 Espresso Coffee

Figure 3.29 Phenolic acids in raw coffee

extensively polyphenol oxidase/tyrosinase/catechol oxidase as indicators

of green coffee bean quality (Clifford, 1985). Polyphenol oxidase is

responsible for the discoloration of defective beans, which is caused by

catalysing the oxidation of the chlorogenic acids.

3.11.8 Non-protein nitrogen Trigonelline

Raw coffee contains trigonelline (arabica: 0.6–1.3%, robusta: 0.3–0.9%).

Dewaxing, decaffeination procedure and steaming (Stennert and Maier,

1994) do not lead to appreciable changes in the trigonelline content

(Figure 3.30). Caffeine and trace purine alkaloids (Figure 3.31)

Caffeine is the major purine in green coffee, where it is probably

associated with the chlorogenic acids by a  -electron complex (Horman

and Viani, 1972). The average content of caffeine is 1.2–1.3% in arabica

and 2.2–2.4% in robusta, showing a marked inter-specific difference as

well as high intra-specific variability. Caffeine is absent or present only in

traces in some wild non-commercial species.

In addition to caffeine, coffee also contains the three dimethylxanthines

– paraxanthine, theobromine and theophylline – and other

trace purines, present in larger quantities in robusta – particularly in

immature beans – than in arabica. Kappeler and Baumann (1986) found

that unripe beans have higher contents of theophylline, theobromine,

liberine and theacrine than ripe ones. Weidner and Maier (1999) also

identified paraxanthine and theacrine. Prodolliet et al. (1998) tried to

determine the geographic origin of coffee using the isotopic ratios (C, N,

H) in samples of caffeine extracted from green arabicas and robustas from

16 countries. Both univariate and multivariate analysis allowed neither

the determination of the species nor of the country of origin.

The raw bean 155

Figure 3.30 Trigonelline, a precursor of many aroma compounds

3.11.9 Lipids

The lipid content of green arabica coffee beans averages some 15%, whilst

robusta coffees contain much less, around 10%. Most of the lipids – the

coffee oil – are located in the endosperm; only a small amount, the coffee

wax, is on the outer layer of the bean. The yield of crude lipids depends

not only on bean composition, but also on extraction conditions,

particularly particle size and surface area, choice of solvent and duration

of extraction.

Coffee oil is composed mainly of triglycerides with fatty acids in

proportions similar to those found in common edible vegetable oils. The

relatively large unsaponifiable fraction is rich in diterpenes of the kaurane

family, mainly cafestol, kahweol and 16-O-methylcafestol, which have

received increasing attention in recent years because of their physiological

effects (see 10.3). Furthermore, 16-O-methylcafestol can be used as a

reliable indicator for robusta content in coffee blends. Among the sterols,

that are also part of the unsaponifiable matter, various desmethyl-,

methyl- and dimethylsterols have been identified. The composition of the

lipid fraction of green coffee is given in Table 3.18. Fatty acids

For the most part, the fatty acids are present as glycerol esters in the

triglycerides, some 20% are esterified with diterpenes, and a small

proportion as sterol esters. Folstar et al. (1975) and Speer et al. (1993)

investigated the fatty acids of the triglycerides and the diterpene esters in

detail. The fatty acids in sterol esters were determined by Picard et al.


156 Espresso Coffee

Figure 3.31 Purine alkaloids in raw coffee

The presence of free fatty acids (FFA) in coffee has been described by

various authors (Kaufmann and Hamsagar, 1962b; Calzolari and Cerma,

1963; Carisano and Gariboldi, 1964; Wajda and Walczyk, 1978). Speer et

al. (1993) developed a method for the direct determination of free fatty

acids: Nine different free fatty acids were detected, uniformly distributed

in both robusta and arabica coffees. In both species the main fatty acids

are C18:2 and C16. It was also possible to detect large proportions of C18,

C18:1, C20 and C22, whereas there were no more than traces of C14, C18:3

and C24. While the proportion of stearic acid is noticeably smaller than

that of oleic acid in robustas, the percentages of these two acids in arabica

coffees are almost equal. The ratio stearic acid/oleic acid may give a first

indication of robusta content in coffee blends. Diterpenes

The main diterpenes in coffee are pentacyclic diterpene alcohols based on

the kaurane skeleton. Working groups under Bengis and Anderson

(1932), Chakravorty et al. (1943a, 1943b), Wettstein et al. (1945),

Haworth and Johnstone (1957) and Finnegan and Djerassi (1960)

elucidated the structure of two of the coffee diterpenes, namely kahweol

and cafestol. In 1989, 16-O-methylcafestol was isolated from robusta

coffee beans (Speer and Mischnick, 1989; Speer and Mischnick-

Lu¨bbecke, 1989). 16-O-methylkahweol, another O-methyl diterpene,

has recently also been found in robusta coffee beans (Speer et al., 2000).

The structural formulae of these diterpenes are shown in Figure 3.32.

The raw bean 157

Table 3.18 Composition of the lipid

fraction of raw coffee (percentages of

total lipids, average)

Component Percent

Triglycerides 75.2

Esters of diterpene alcohols and fatty acids 18.5

Diterpene alcohols 0.4

Esters of sterols and fatty acids 3.2

Sterols 2.2

Tocopherols 0.04–0.06

Phosphatides 0.1–0.5

Tryptamine derivatives 0.6–1.0

Source: Maier, 1981

Arabica coffee contains cafestol and kahweol. Robusta coffee contains

cafestol, small amounts of kahweol and, additionally, 16-O-methylcafestol

(16-OMC) and traces of 16-O-methylkahweol, which were found

only in robusta coffee beans (Speer and Mischnick-Lu¨bbecke, 1989;

Speer and Montag, 1989; Speer et al., 1991b; Ko¨lling-Speer et al., 2001).

Absence of 16-OMC in arabica coffee beans has been confirmed by

Frega et al. (1994), White (1995) and Trouche et al. (1997). Because of

its stability during the roasting process, 16-OMC can be used to detect

the presence of down to less than 2% robusta in arabica blends (Speer et

al., 1991b).

The diterpenes cafestol, kahweol and 16-OMC are mainly esterified

with various fatty acids. In their free form, they occur only as minor

components in coffee oil (Speer et al., 1991a; Ko¨lling-Speer et al., 1999).

Both free cafestol and kahweol are present in arabica coffees, cafestol

being the major component. In robusta coffee, free cafestol content is

slightly higher than 16-OMC content, and free kahweol is either present

only in traces or absent. Comparison of free diterpenes with total

diterpenes after saponification shows that in arabicas proportions ranged

from 0.7 to 2.5%; in robustas the proportions of free diterpenes were

slightly higher with 1.1 to 3.5%; experiments to verify if the difference is

due to the processing technique or to the species are ongoing.

158 Espresso Coffee

Figure 3.32 Structural formulae of raw coffee diterpenes

Fatty acid esters have been reported (Kaufmann and Hamsagar, 1962a;

Folstar et al., 1975; Folstar, 1985; Pettitt, 1987; Speer, 1991, 1995;

Kurzrock and Speer, 1997a, 1997b): Cafestol, 16-OMC and kahweol

esters with fatty acids such as C14, C16, C18, C18:1, C18:2, C18:3, C20, C22,

C24 were identified, as well as esters with the fatty acid C20:1 and some

odd-numbered fatty acids such as C17, C19, C21 and C23 (Kurzrock and

Speer, 2001a, 2001b). The individual diterpene esters were irregularly

present in the coffee oil. The odd-numbered fatty acid esters were minor

components, whereas the diterpenes esterified with palmitic, linoleic,

oleic, stearic, arachidic, and behenic acid, were present in larger amounts.

Table 3.19 indicates the distribution of the six main esters (summing up

to nearly 98% of the respective diterpenes) for arabica coffee.

The total content of these six cafestol esters in sum ranged between 9.4

and 21.2 g/kg dry weight, corresponding to 5.2 to 11.8 g/kg cafestol in

different arabica coffees, and between 2.2 and 7.6 g/kg dry weight in

robusta, corresponding to 1.2 to 4.2 g/kg cafestol, notably less than in

arabica. Sterols

Coffee contains a number of sterols that are common in other seed oils as

well (Figure 3.33, see page 160). In addition to 4-desmethylsterols, various

4-methyl- and 4,4-dimethylsterols have been identified. Coffee sterols

have been found both in the free (around 40%) and in the esterified form

(around 60%) (Nagasampagi et al., 1971; Itoh et al., 1973a, 1973b;

Tiscornia et al., 1979; Duplatre et al., 1984; Picard et al., 1984; Mariani

The raw bean 159

Table 3.19 Distribution (%) of the diterpene esters in raw

arabica coffee

Cafestol + kahweol1

(n2 = 1)

Cafestol + kahweol2

(n = 1)


(n = 10)


(n = 10)

C16 42.5 51.4 40–49 46–50

C18 17.5 9.1 9–11 8–11

C18:1 11.0 7.4 9–15 8–12

C18:2 20.5 26.4 24–30 25–29

C20 6.0 4.6 3–6 3–6

C22 2.5 1.1 0.6–1.2 0.7–1.3

Kahweol esters calculated as cafestol esters.

n = Number of samples analysed.

Sources: 1Kaufmann and Hams, 1962a; 2Folstar, 1985; 3Kurzrock and Speer, 1997a;

4Kurzrock and Speer, 2001a,b

and Fedeli, 1991; Frega et al., 1994). The desmethylsterols represent

90% of the total sterol fraction, which ranges from 1.5% to 2.4% of the

lipids (Picard et al., 1984). Nagasampagi et al. (1971) found higher

proportions (5.4%).

Table 3.20 presents the distribution of the main desmethylsterols in

different robusta and arabica coffee samples. The main sterol is

 -sitosterol, with about 50%. 24-methylenecholesterol and  5-avenasterol,

occurring in larger amounts in robusta than in arabica coffee beans,

are suitable for coffee blend studies (Duplatre et al., 1984; Frega et al.,

1994). However, because of their varying natural contents, they can be

only used for determining proportions of robusta in a mixture with arabica

from 20% onward. Tocopherols (Figure 3.34)

The presence of tocopherols in coffee oil was described for the first time

by Folstar et al. (1977). They found concentrations of  -tocopherol of

160 Espresso Coffee

Figure 3.33 Sterols of raw coffee beans

89 to 188 mg/kg oil, for  - .  -tocopherol 252–53 mg/kg oil. The

predominance of  -tocopherol is a prominent feature of coffee beans, in

contrast to other vegetables and fruits (Aoyama et al. 1988; Ogawa et al.

1989). Coffee wax

A thin wax layer, corresponding to 2–3% of the total lipids, covers the

surface of green coffee beans. The wax content is generally defined as the

The raw bean 161

Table 3.20 Mean distribution (%)

of desmethylsterols in raw arabica

and robusta coffees (30 samples)

Sterols Arabica Robusta

Cholesterol 0.3 0.2

Campesterol 15.8 16.9

Stigmasterol 21.9 23.1

 -Sitosterol 51.6 45.4

 5-Avenasterol 2.7 9.1

Campestanol 0.4 0.2

24-Methylenecholesterol 0.2 1.9

Sitostanol 2.0 0.8

 7-Stigmastenol 2.2 0.2

 7-Avenasterol 1.5 0.4

 7-Campesterol 0.6 0.2

 5.23-Stigmastadienol 0.3 0.5

 5.24-Stigmastadienol 0.1 Traces

Clerosterol 0.5 0.7

Source: Mariani and Fedeli, 1991

Figure 3.34 Tocopherols of raw coffee beans

material obtained from the intact beans by extraction with chlorinated

solvents such as dichloromethane. The first investigations about wax

composition of green arabica coffee beans were done by Wurziger and coworkers

(Dickhaut, 1966; Harms, 1968). They isolated and identified

three carbonic acid 5-hydroxytryptamides (C-5-HT) (Figure 3.35).

Arachidic acid (n . 18), behenic acid (n . 20) and lignoceric acid

(n . 22) are combined with the primary amino group of 5-hydroxytryptamine.

In addition, Folstar et al. (1980) reported the presence of

stearic acid 5-hydroxytryptamide (n . 16), later !-hydroxyarachidic acid

5-hydroxytryptamide (n . 18), and !-hydroxybehenic acid 5-hydroxytryptamide

(n . 20). Arachidic acid and behenic acid 5-hydroxytryptamide

are dominant, but the other amides are only minor constituents.

The content ranges between 500 and 2370 mg/kg in green arabica, and

from 565 to 1120 mg/kg in green robusta (Maier, 1981). Polishing,

dewaxing and decaffeination of the beans leads to a substantial reduction

in C-5-HT (Harms and Wurziger, 1968; Van der Stegen and Noomen,

1977; Hunziker and Miserez, 1979; Folstar et al., 1980). The C-5-HT have

antioxidant properties (Lehmann et al., 1968; Bertholet and Hirsbrunner,


3.11.10 Volatile constituents

Some 180 volatile substances have been identified in green coffee beans

(Boosfeld and Vitzthum, 1995; Holscher and Steinhart, 1995; Cantergiani

et al., 1999). Arabicas are similar to robustas, but arabicas are

distinguished by a large content of terpenes and fewer aromatic

compounds. The range is affected by green bean processing, and could

be used for evaluating green bean quality. The odour of defective green

beans has been linked with specific components (see 3.7.2).

162 Espresso Coffee

Figure 3.35 Structural formulae of carbonic acid 5-hydroxytryptamides (C-5-HT)

3.11.11 Contaminants Mycotoxins

R. Viani

The occasional occurrence of several mycotoxins that may contaminate

green coffee beans has been known since the early 1970s: they are

sterigmatocystine, aflatoxin B and ochratoxinA(OTA) (Levi et al., 1974).

Sterigmatocystine was only found in extremely contaminated beans,

clearly unfit for consumption, and is of no practical relevance in coffee.

The aflatoxins, aflatoxin B in particular, have sometimes been detected

in green coffee beans, and – in extremely small amounts, apparently of no

practical concern for public health – also in the beverage (Nakajima et al.,


OTA contamination of coffee has been the subject of intensive

research during the past 15 years, after it was realized, thanks to improved

analytical techniques, that not all the OTA present in green beans was

destroyed during roasting (Tsubouchi et al., 1988). Much information is

available through the three workshops, organized in Nairobi (Anon.,

1997), Helsinki (Van der Stegen and Blanc, 1999) and Trieste (Van der

Stegen et al., 2001) by the Association Scientifique Internationale du Cafe´

(ASIC). A comprehensive review has been published by Bucheli and

Taniwaki (2002).

The mould species, which are known to produce (OTA) in coffee, have

been identified as Aspergillus ochraceus, A. carbonarius and, occasionally,

A. niger (Frank, 2001; Ismayadi and Zaenudin, 2001; Joosten et al., 2001;

Paneer et al., 2001; Pitt et al., 2001). The specific time and place of

contamination have not yet been clearly identified. Both natural and wetprocessed

coffees appear to be at risk of attack by the mould, although

wet-processed coffees are rarely found contaminated by OTA.

The risk of OTA contamination can clearly be minimized by following

good agricultural and good manufacturing practices at the harvest, postharvest

and storage-transportation stages (Teixeira et al., 2001; Viani,


n Storage of fresh cherries – ripe and overripe cherries, stored over

days piled up or, worse, in plastic bags, often in the sun, are at

increased risk of OTA contamination (Bucheli et al., 2000);

therefore, avoiding storage of fresh cherries before processing is

important not only for coffee organoleptic quality (see 3.2), but also

to reduce the risk of OTA contamination (Figure 3.36).

n Drying – after drying begins, the length of time spent at a water

activity above 0.80 at any time until roasting determines the risk of

The raw bean 163

mould growth and OTA production, which may explain the

increased risk for sun drying with respect to mechanical drying, and

of cherry sun drying with respect to the shorter parchment sun drying.

Organoleptic quality considerations require a regular long drying

process (see 3.3), while at the same time reduction of the risk of OTA

contamination asks for quick drying. The drying process therefore

needs to be optimized to adjust to both requirements (Ismayadi et al.,

2001; McGaw et al., 2001).

n Husks – contamination appears to start within the husk, and, in the

‘appropriate’ moisture conditions, it can reach the bean in 3–4 days;

therefore, freeing green coffee beans from all husk material during

cleaning and grading is important to reduce the OTA load (Pittet et

al., 1996; Blanc et al., 1998; Bucheli et al., 2000).

n Re-wetting – storage in the tropics of green robusta coffee for up to

one year at 70% relative humidity and at temperatures up to 30



has not produced any increase in the OTA load (Bucheli et al.,

1998). Transportation trials from the producer countries have

shown that risk of condensation and wetting of coffee occur mainly

during transport overland to the harbour for shipping, and on arrival

at destination in the winter season; this may occur even with

properly dried coffee if too large temperature variations cause

evaporation followed by water condensation on the upper layer of

coffee (Blanc et al., 2001). OTA contamination can also occur

during the decaffeination process (see 3.9), when decaffeinated wet

beans are not properly and rapidly dried (unpublished data).

Avoiding re-wetting at any stage before roasting is therefore very

important (Figure 3.37).

164 Espresso Coffee

Figure 3.36 Storage of fresh cherries must be short! Polycyclic aromatic hydrocarbons (PAH)

I. Ko¨ lling-Speer and K. Speer

Different PAH, e.g. fluoranthene, pyrene, benz[a]anthracene, chrysene,

triphenylene, benzo[b]fluoranthene, benzo[j]fluoranthene, benzo[k]fluoranthene,

benzo[a]fluoranthene, benzo[e]pyrene, benzo[a]pyrene, perylene,

indeno[1,2,3-cd]pyrene, dibenz[ah]anthracene, dibenz[ac]anthracene and

benzo[ghi]perylene were identified and quantified in green and roasted

coffees. The individual PAH can be formed both by the fuel and by the

The raw bean 165

Figure 3.37 Drying coffee – cherry or parchment – must be protected from rewetting

coffee bean under sub-optimal roasting conditions. On comparing the

PAH contents of the unroasted coffee with those of the roasted product it

is seen that it is not so much the roasting process after its optimization but

the PAH content of the raw material which determines the content in

the roasted product (Klein et al., 1993). The contamination of the green

coffee may be due to exhaust gases of cars or to transportation of the

coffee beans in contaminated jute and sisal sacks. Their fibres are treated

before spinning with batching oil. Before bean contamination was

discovered the batching oil commonly consisted of an uncontrolled

mineral oil fraction, which could even be discarded motor oil (Grob et al.,

1991a, 1991b; Moret et al., 1997). The contamination was identified by

analysing the PAH profile (Speer et al., 1990; Moret et al., 1997). Strict

rules have been suscribed and are now adhered by bag manufacturers, who

generally mark by a tag sacks prepared according to GMP. Pesticide residues

I. Ko¨ lling-Speer and K. Speer

Due to the protective layers covering the beans in the fruit (skin, pulp and

parch), the level of pesticides present in green coffee beans is very low.

Cetinkaya (1988) could not detect either organochlorine or organophosphorus

pesticide residues in 50 green coffee samples imported from 11

different countries.


Aeschbach R., Kusy A. and Maier H.G. (1982) Diterpenoide in Kaffee. Z.

Lebensm. Unters. Forsch. 175, 337–341.

Amorim H.V., Cruz A.R., St Angelo A.G. et al. (1977) Biochemical, physical

and organoleptic changes during raw coffee quality deterioration. Proc.

8th ASIC Coll., pp. 183–186.

Amorim H.V., Teixeira A.A., Moraes R.S., Reis A.J., Pimentel-Gomes F. and

Malavolta E. (1973) Studies on the mineral nutrition of coffee. XXVII.

Effect of N, P and K fertilization on the macro and micro nutrients of

coffee fruits and on the beverage quality. An. Esc. Sup. Agr. LQ

(Piracicaba) 20, 323–333.

Anon. (1987) Electronic sorting reduces labour costs. Food Technol. NZ, p. 47.

Anon. (1997) Special workshop on the enhancement of coffee quality by

reduction of mould growth. Proc. 17th ASIC Coll., pp. 367–369, and

following articles.

166 Espresso Coffee

Aoyama M., Maruyama T., Kanematsu H., Niiya I., Tsukamoto M., Tokairin

S. and Matsumoto T. (1988) Studies on the improvement of antioxidant

effect of tocopherols. XVII. Synergistic effect of extracted components

from coffee beans. Yukagaku 37, 606–612.

Arnold U. (1995) Nachweis von Aminosa¨urevera¨nderungen und Bestimmung

freier Aminosa¨uren in Rohkaffee, Beitra¨ge zur Charakterisierung von

Rohkaffeepeptiden. Thesis, University of Dresden.

Bacchi O. (1962) O branqueamento dos gra˜os de cafe´. Bragantia (Campinas)

21 (28), 467–484.

Ba¨hre F. (1997) Neue nichtflu¨chtige Sa¨uren im Kaffee. Thesis, Technical

University of Braunschweig.

Ba¨hre F. and Maier H.G. (1999) New non-volatile acids in coffee. Dtsche

Lebensm.-Rundsch. 95, 399–402.

Becker R., Do¨hla B., Nitz S. and Vitzthum O.G. (1987) Identification of the

‘peasy’ off-flavour note in central African coffees. Proc. 12th ASIC Coll.,

pp. 203–215.

Bee S.C. (2001) Optical sorting for the coffee industry. Proc. 19th ASIC Coll.,


Bengis R.O. and Anderson R.J. (1932) The chemistry of the coffee bean. I.

Concerning the unsaponifiable matter of the coffee bean oil. Extraction

and properties of kahweol. J. Biol. Chem. 97, 99–113.

Berry N.E. and Walter R.H. (1943) Process of decaffeinating coffee. US Patent


Bertholet R. and Hirsbrunner P. (1984) Preparation of 5-hydroxytryptamine

from coffee wax. European Patent 1984–104 696.

Blanc M. and Margolis G. (1981) Caffeine extraction. European Patent 0049


Blanc M., Pittet A., Mun˜oz-Box R. and Viani R. (1998) Behavior of

ochratoxin A during green coffee roasting and soluble coffee manufacture.

J. Agric. Food Chem. 46, 673–675.

Blanc M., Vuataz G. and Hickmann L. (2001) Green coffee transport trials.

Proc. 19th ASIC Coll., CD-ROM.

Boosfeld J. and Vitzthum O.G. (1995) Unsaturated aldehydes identification

from green coffee. J. Food Sci. 60, 1092–1096.

Bouyjou R., Fourny G. and Perreaux D. (1993) Le gouˆt de pomme de terre du

cafe´ arabica au Burundi. Proc. 15th ASIC Coll., pp. 357–369.

Bradbury A.G.W. and Balzer H.H. (1999) Carboxyatractyligenin and

atractyligenin glycosides in coffee. Proc. 18th ASIC Coll., pp. 71–77.

Bradbury A.G.W. and Halliday D.J. (1987) Polysaccharides in green coffee

beans. Proc. 12th ASIC Coll., pp. 265–269.

Bradbury A.G.W. and Halliday D.J. (1990) Chemical structures of green

coffee bean polysaccharides. J. Agric. Food Chem. 38, 389–392.

Bucheli P. and Taniwaki M.H. (2002) Research on the origin, and the impact

of postharvest handling and manufacturing on the presence of

ochratoxin A in coffee. Food Addit. Contam. 19 (7), 655–665.

The raw bean 167

Bucheli P., Kanchanomai C., Meyer I. and Pittet A. (2000) Development of

ochratoxin A during Robusta (Coffea canephora) coffee cherry drying. J.

Agric. Food Chem. 48, 1358–1362.

Bucheli P., Meyer I., Pittet A., Vuataz G. and Viani R. (1998) Industrial

storage of green robusta coffee under tropical conditions and its impact

on raw material quality and ochratoxin A content. J. Agric. Food Chem.

46, 4507–4511.

Calzolari C. and Cerma E. (1963) Sulle sostanze grasse del caffe`. Riv. Ital.

Sostanze Grasse 40, 176–180.

Cantergiani E., Brevard H., Amado R., Krebs Y., Feria-Morales A. and

Yeretzian C. (1999) Characterisation of mouldy/earthy defect in green

Mexican coffee. Proc. 18th ASIC Coll., pp. 43–49.

Carisano A. and Gariboldi L. (1964) Gaschromatographic examination of the

fatty acids of coffee oil. J. Sci. Food Agric. 15, 619–622.

Carvalho A. (1988) Principles and practice of coffee plant breeding for

productivity and quality factors: Coffea arabica. In R.J. Clarke and

R. Macrae (eds), Coffee Volume 4 – Agronomy. London: Elsevier Applied

Science, pp. 129–165.

Carvalho A., Garruti R.S., Teixeira A.A., Pupo L.M. and Moˆnaco L.C.

(1972) Ocorreˆncia dos principais defeitos do cafe´ em va´rias fases de

maturac¸a˜o dos frutos. Bragantia (Campinas) 29 (20), 207–220.

Castel de Menezes H. and Clifford M.N. (1988) The influence of stage of

maturity and processing method on the relation between the different

isomers of caffeoilquinic acid in green coffee beans. Proc. 12th ASIC

Coll., pp. 377–381.

Cetinkaya M. (1988) Organophosphor- und Organochlorpestizidru¨cksta¨nde

in Rohkaffee. Dtsch. Lebensm. Rundsch. 84, 189–190.

Chakravorty P.N., Levin R.H., Wesner M.M. and Reed. G. (1943b)

Cafesterol III. J. Am. Chem. Soc. 65, 1325–1328.

Chakravorty P.N., Wesner M.M. and Levin R.H. (1943a) Cafesterol II. J.

Am. Chem. Soc. 65, 929–932.

Clarke R.J. (1988) International standardization. In R.J. Clarke and R.

Macrae (eds), Coffee: Volume 6 – Commercial and Technological Aspects.

London: Elsevier Applied Science, pp. 112–121.

Clarke R.J. and Vitzthum O.G. (eds) (2001) Coffee – Recent Developments.

Oxford: Blackwell Science.

Clifford M.N. (1985) Chemical and physical aspects of green coffee and

coffee products. In M.N. Clifford and K.C. Willson (eds), Coffee:

Botany, Biochemistry and Production of Beans and Beverage. Westport, CT:

AVI, pp. 314–315.

Clifford M.N. (1999) Chlorogenic acids and other cinnamates. Nature,

occurrence and dietary burden. J. Sci. Food Agric. 79, 362–372.

Clifford M.N. and Ohiokpehai O. (1983) Coffee astringency. Analyt. Proc.

20, 83–86.

168 Espresso Coffee

Clifford M.N., Kazi T. and Crawford S. (1987) The content and washout

kinetics of chlorogenic acids in normal and abnormal green coffee beans.

Proc. 12th ASIC Coll., pp. 221–228.

Dentan E. (1987) Examen microscopique de grains de cafe´ riote´s. Proc. 12th

ASIC Coll., pp. 335–352.

Dentan E. (1989) Etude microscopique de quelques types de cafe´ de´fectueux:

grains noirs, blanchaˆtres, cireux et ‘ardidos’. Proc. 13th ASIC Coll.,

pp. 283–301.

Dentan E. (1991) Etude microscopique de quelques types de cafe´ de´fectueux.

II: grains a` gouˆt d’herbe, de terre, de moisi; grains puants, endommage´s

par les insectes. Proc. 14th ASIC Coll., pp. 293–311.

Dentan E and Illy A. (1985) Etude microscopique de grains de cafe´ matures,

immatures et immatures fermente´s arabica Santos. Proc. 11th ASIC

Coll., pp. 341–368.

Dickhaut G. (1966) U ¨ ber phenolische Substanzen in Kaffee und deren

analytische Auswertbarkeit zur Kaffeewachsbestimmung. Thesis,

University of Hamburg.

Duplatre A., Tisse C. and Estienne J. (1984) Contribution a` l’identification

des espe`ces arabica et robusta par e´tude de la fraction ste´rolique. Ann.

Fals. Exp. Chim. 828, 259–270.

Engelhardt U.H. and Maier H.G. (1984) Nichtflu¨chtige Sa¨uren im Kaffee.

Thesis, Technical University of Braunschweig.

Engelhardt U.H. and Maier H.G. (1985) Sa¨uren des Kaffees. XII. Anteil

einzelner Sa¨uren an der titrierbaren Gesamtsa¨ure. Z. Lebensm. Unters.

Forsch. 181, 206–209.

Ferraz M.B. and Veiga A.A. (1960) Melhor bebida e maior poder germinativo

do cafe´. Boletim de Superintendeˆncia dos Servic¸os do Cafe´ 05–18, 398–399.

Finnegan R.A. and Djerassi C. (1960) Terpenoids XLV. Further studies on the

structure and absolute configuration of cafestol. J. Am. Chem. Soc. 82,


Fischer A. and Kummer P. (1979) Verfahren zum Entcoffeinieren von

Rohkaffee. European Patent 008 398.

Fischer M., Reimann S., Trovato V. and Redghwell R.J. (2000) Structural

aspects of polysaccharides from arabica coffee. Proc 18th ASIC Coll.,

pp. 91–94.

Flament I. (2002) Coffee Flavour Chemistry. New York: John Wiley.

Folstar P. (1985) Lipids. In R.J. Clarke and R. Macrae (eds), Coffee: Volume 1

– Chemistry. London: Elsevier Applied Science, pp. 203–222.

Folstar P., Pilnik W., de Heus J. G. and Van der Plas H. C. (1975) The composition

of fatty acids in coffee oil and wax. Lebensm. Technol. 8,


Folstar P., Schols H.A., Van der Plas H.C., Pilnik W., Landherr C.A. and Van

Vildhuisen A. (1980) New tryptamine derivatives isolated from wax of

green coffee beans. J. Agric. Food Chem. 28, 872–874.

The raw bean 169

Folstar P., Van der Plas H.C., Pilnik W. and de Heus J.G. (1977) Tocopherols

in the unsaponifiable matter of coffee bean oil. J. Agric. Food Chem. 25,


Frank J.M. (2001) On the activity of fungi in coffee in relation to ochratoxin

A production. Proc. 19th ASIC Coll., CD-ROM.

Franz H. and Maier H.G. (1993) Inositol phosphates in coffee and related

beverages. I. Identification and methods of determination. Dtsche

Lebensm.-Rundsch. 89, 276–282.

Franz H. and Maier H.G. (1994) Inositol phosphates in coffee and related

beverages. II. Coffee beans. Dtsche Lebensm.-Rundsch. 90, 345–349.

Frega N., Bocci F. and Lercker G. (1994) High resolution gas chromatographic

method for determination of Robusta coffee in commercial

blends. J. High Resolution Chromatogr. 17, 303–307.

Full G., Lonzarich V. and Suggi L.F. (1999) Differences in chemical

composition of electronically sorted green coffee beans. Proc. 18th

ASIC Coll., pp. 35–42.

Gibson A. and Butty M. (1975) Overfermented coffee beans (‘stinkers’)

a method for their detection and elimination. Proc. 7th ASIC Coll.,

pp. 141–152.

Gomes F.P., Cruz V.F., Castilho A., Teixeira A.A. and Pereira L.S.P. (1967)

A influeˆncia de gra˜os pretos em ligas com cafe´s de bebida mole. Anais da

E.S.A. ‘Luiz de Queiroz’ (Piracicaba) 24, 71–81.

Grob K., Biedermann M., Artho A. and Egli J. (1991a) Food contamination

by hydrocarbons from packaging materials determined by coupled

LC–GC. Z. Lebensm. Unters. Forsch. 193, 213–219.

Grob K., Lanfranchi M., Egli J. and Artho A. (1991b) Determination of food

contamination by mineral oil from jute sacks using coupled LC-GC. J.

Assoc. Off. Anal. Chem. 74, 506–512.

Grosch W. (1998) Welche Verbindungen bevorzugt der Geruchssinn bei

erhitzten Lebensmitteln? Lebensmittelchemie 52, 143–146.

Guyot B., Petnga E. and Vincent J.C. (1988a) Analyse qualitative d’un cafe´

Coffea canephora var. robusta. I. Evolution des caracte`res physiques et

organoleptiques. Cafe´ Cacao The´ 32, 127–140.

Guyot B., Petnga E., Lotode´ R. and Vincent J.C. (1988b) Analyse qualitative

d’un cafe´ Coffea canephora var. Robusta en fonction de la maturite´. II.

Application de l’analyse statistique multidimensionnelle. Cafe´ Cacao

The´ 32, 229–242.

Guyot B., Cochard B. and Vincent J.C. (1991) De´termination quantitative

du dime´thylsulfure et du dime´thyldisulfure dans l’aroˆme de cafe´. Cafe´

Cacao The´ 35, 49–56.

Harms U. (1968) Beitra¨ge zum Vorkommen und zur Bestimmung von

Carbonsa¨ure-5-hydroxy-tryptamiden in Kaffeebohnen. Thesis, University

of Hamburg.

Harms U. and Wurziger J. (1968) Carboxylic acid 5-hydroxytryptamides in

coffee beans. Z. Lebensm. Unters. Forsch. 138, 75–80.

170 Espresso Coffee

Haworth R.D. and Johnstone R.A.W. (1957) Cafestol Part. II. J. Chem. Soc.

(Lond.), pp. 1492–1496.

Heilmann W. (2001) Decaffeination of coffee. In R.J. Clarke and O.G.

Vitzthum (eds), Coffee – Recent Developments. Oxford: Blackwell

Science, pp. 108–124.

Holscher W. and Steinhart H. (1995) Aroma compounds in green coffee. In

G. Charalambous (ed.), Food Flavours: Generation, Analysis and Process

Influence. 37 A. Amsterdam: Elsevier Science, pp. 785–803.

Horman I. and Viani R. (1972) The nature and conformation of the caffeinechlorogenate

complex of coffee. J. Food Sci. 37, 925–927.

Hunziker H.R. and Miserez A. (1979) Bestimmung der 5-Hydroxytryptamide

in Kaffee mittels Hochdru¨ck-Flu¨ssigkeitschromatographie. Mitt. Geb.

Lebensum. Unters. Hyg. 70, 142–152.

Illy A. and Viani R. (eds) (1995) Espresso Coffee. The Chemistry of Quality.

London: Academic Press, p. 29.

Illy E., Brumen G., Mastropasqua L. and Maughan W. (1982) Study on the

characteristics and the industrial sorting of defective beans in green

coffee lots. Proc. 10th ASIC Coll., pp. 99–128.

Ismayadi C. and Zaenudin. (2001) Toxigenic mould species infestation in

coffee beans taken from different levels of production and trading in

Lampung – Indonesia (2001). Proc. 19th ASIC Coll., CD-ROM.

Ismayadi C., Zaenudin and Priyono S. (2001) Mould species infestation

during sun drying of sound and split coffee cherries. Proc. 19th ASIC

Coll., CD-ROM.

ISO (1980) Green Coffee – Olfactory and Visual Examination and

Determination of Foreign Matter and Defects. ISO 4149-1980.

Geneva: International Organization for Standardization.

ISO (1983) Green Coffee – Determination of Loss in Mass at 105



6673-1983. Geneva: International Organization for Standardization.

ISO (1985) Green Coffee – Determination of Proportion of Insect-damaged

Beans. ISO 6667-1985. Geneva: International Organization for


ISO (1993) Green Coffee – Defect Reference Chart. ISO 10470-1993.

Geneva: International Organization for Standardization.

ITC (2002) Coffee: an Exporter’s Guide. Geneva: International Trade Centre,


Itoh T., Tamura T. and Matsumoto T. (1973a) Sterol composition of 19

vegetable oils. J. Am. Oil Chem. Soc. 50, 122–125.

Itoh T., Tamura T. and Matsumoto T. (1973b) Methylsterol compositions of

19 vegetable oils. J. Am. Oil Chem. Soc. 50, 300–303.

Jobin P. (1982) Les Cafe´s produits dans le monde. Le Havre: Jobin.

Joosten H.M.L.J., Goetz J., Pittet A., Schellenberg M. and Bucheli P. (2001)

Production of ochratoxin A by Aspergillus carbonarius on coffee cherries.

Int. J. Food Microb. 65, 39–44.

The raw bean 171

Jouanjan F. (1980) Transport maritime du cafe´ vert et conte´neurisation. Proc.

9th ASIC Coll., pp. 177–188.

Kampmann B. and Maier H.G. (1982) Sa¨uren des Kaffees. I. Chinasa¨ure. Z.

Lebensm. Unters. Forsch. 175, 333–336.

Kappeler A.W. and Baumann T.W. (1986) Purine alkaloid pattern in coffee

beans. Proc. 11th ASIC Coll., pp. 273–279.

Kaufmann H.P. and Hamsagar R.S. (1962a) Zur Kenntnis der Lipoide der

Kaffeebohne. I. U ¨ ber Fettsa¨ure-Ester des Cafestols. Fette Seifen

Anstrichmittel 64, 206–213.

Kaufmann H.P. and Hamsagar R.S. (1962b) Zur Kenntnis der Lipoide der

Kaffeebohne. II. Die Vera¨nderung der Lipoide bei der Kaffee-Ro¨stung.

Fette Seifen Anstrichmittel 64, 734–738.

Klein H., Speer K. and Schmidt E.H.F. (1993) Polycyclic aromatic

hydrocarbons (PAH) in raw and roasted coffee. Bundesgesundheitsblatt

36, 98–100.

Ko¨lling-Speer I. and Speer K. (1997) Diterpenes in coffee leaves. Proc. 17th

ASIC Coll., pp. 150–154.

Ko¨lling-Speer I., Kurzrock T. and Speer K. (2001) Contents of diterpenes in

green coffees. Proc. 19th ASIC Coll., CD-ROM.

Ko¨lling-Speer I., Strohschneider S. and Speer K. (1999) Determination of free

diterpenes in green and roasted coffees. J. High Resolution Chromatogr.

22, 43–46.

Kurzrock T. and Speer K. (1997a) Fatty acid esters of cafestol. Proc. 17th

ASIC Coll., pp. 133–140.

Kurzrock T. and Speer K. (1997b) Identification of cafestol fatty acid esters.

In R. Amado` and R. Battaglia (eds), Proc. Euro Food Chem. IX

(Interlaken), Volume 3, pp. 659–663.

Kurzrock T. and Speer K. (2001a) Diterpenes and diterpene esters in coffee.

Food Rev. Int. 17, 433–450.

Kurzrock T. and Speer K. (2001b) Identification of kahweol fatty acid esters

in arabica coffee by means of LC/MS. J. Sep. Sci. 24, 843–848.

Labuza T.P., McNally L., Gallager D. et al. (1972) Stability of intermediate

moisture foods. 1. Lipid oxidation. J. Food Sci. 37, 154–159.

Lack E. and Seidlitz H. (1993) Commercial scale decaffeination of coffee and

tea using supercritical CO2. In M.B. King and T.R. Bott (eds), Extraction

of Nature Products using Near Critical Solvents. Glasgow: Blackie, pp.


Lehmann G., Neunhoeffer O., Roselius W. and Vitzthum O. (1968)

Antioxidants made from green coffee beans and their use for protecting

autoxidizable foods. German Patent 1,668,236.

Levi C.P., Trenk H.L. and Mohr H.K. (1974) Study of the occurrence of

ochratoxin A in green coffee beans. J. Assoc. Official Analyt. Chemists 57,


Lipke U. (1999) Untersuchungen zur Charakterisierung von

Rohkaffeepeptiden. Thesis, University of Dresden.

172 Espresso Coffee

Lopez Garay C., Bautista Romero E. and Moreno Gonzales E. (1987) Use of

gamma radiation for the preservation of coffee quality during storage.

Proc. 12th ASIC Coll., pp. 771–782.

Ludwig H., Obermann H. and Spiteller G. (1975) New diterpenes found in

coffee. Proc. 7th ASIC Coll., pp. 205–210.

Lu¨llmann C. and Silwar R. (1989) Investigation of mono- and disaccharide

content of arabica and robusta green coffee using HPLC. Lebensm.

Gericht. Chem. 43, 42–43.

Maier H.G. (1981) Kaffee. Berlin and Hamburg: Paul Parey.

Maier H.G. and Wewetzer H. (1978) Bestimmung von Diterpen-Glykosiden

im Bohnenkaffee. Z. Lebensm. Unters. Forsch. 167, 105–107.

Mariani C. and Fedeli E. (1991) Sterols of coffee grain of arabica and robusta

species. Rivista Ital. Sostanze Grasse 68, 111–115.

Martin H. (1982) Selective extraction of caffeine from green coffee beans and

application of similar processes on other natural products. Proc. 10th

ASIC Coll., pp. 21–28.

Matiello J.B., Santinato R., Garcia A.W.R., Almeida S.R. and Fernandez D.R.

(2002) Coltura do cafe´ no Brasil. Novo manual de recomendac¸oes

MAPA – SARC/Procafe´ – SPC/Decaf. Sa˜o Paulo (Brazil), p. 357.

Mazzafera P. (1999) Chemical composition of defective coffee beans. Food

Chem. 64, 547–554.

McGaw D., Comissiong E., Tripathi K., Maharaja A. and Paltoo V. (2001)

The drying characteristics of coffee beans. Proc. 19th ASIC Coll., CDROM.

Mello M., Fazuoli L.C., Teixeira A.A. and Amorim H.V. (1980) Alterac¸o˜es

fı´sicas, quı´micas e organole´pticas em gra˜os de cafe´ armazenados. Cieˆncia e

Cultura (Sa˜o Paulo) 32 (4), 467–472.

Moraes R.M., Angelucci E., Shirose T. and Medina J.C. (1973) Soluble solids

determination in arabica and robusta coffees. Coll. Inst. Techn. Alim.

(Campinas), 5, 199–221.

Moret S., Grob K. and Conte L.S. (1997) Mineral oil polyaromatic

hydrocarbons in foods, e.g. from jute bags, by on-line LC-solvent

evaporation (SE)-LC-GC-FID. Z. Lebensm. Unters. Forsch. 204,


Multon J.L., Poisson J., Cachagnier B. et al. (1973) Evolution de plusieurs

caracte´ristiques d’un cafe´ arabica au cours d’un stockage expe´rimental

effectue´ a` 5 humidite´s relatives et 4 tempe´ratures diffe´rentes. Proc. 6th

ASIC Coll., pp. 268–277.

Nagasampagi B.A., Rowe J.W., Simpson R. and Goad L.J. (1971) Sterols of

coffee. Phytochemistry 10, 1101–1107.

Naish M., Clifford M.N. and Birch G.G. (1993) Sensory astringency of 5-Ocaffeoylquinic

acid, tannic acid and grapeseed tannin by a time-intensity

procedure (1993). J. Sci. Food Agric. 61, 57–64.

Nakajima M., Tsubouchi H., Miyabe M. and Ueno Y. (1997) Survey of

aflatoxin B1 and ochratoxin A in commercial green coffee beans by

The raw bean 173

high-performance liquid chromatography linked with immunoaffinity

chromatography. Food Agric. Immunol. 9, 77–83.

Navellier P. and Brunin R. (1963) Suggestions pour l’expression des re´sultats

des analyses de cafe´. Proc. 1st ASIC Coll., pp. 317–320.

Northmore J.M. (1969) Overfermented beans and stinkers as defectives of

arabica coffee. Proc. 4th ASIC Coll., pp. 47–54.

Noyes R.M. and Chu C.M. (1993) Material balance on free sugars in the

production of instant coffee. Proc. 14th ASIC Coll., pp. 202–210.

Obermann H. and Spiteller G. (1976) Die Strukturen der Kaffee-

Atractyloside. Chem. Ber. 109, 3450–3461.

Ogawa M., Herai Y., Koizumi N., Kusano T. and Sano H. (2001)

7-Methylxanthine methyltransferase of coffee plants – gene isolation

and enzymatic properties. J. Biol. Chem. 276, 8213–8218.

Ogawa M., Kamiya C. and Iida Y. (1989) Contents of tocopherols in coffee

beans, coffee infusions and instant coffee. Nippon Shokuhin Kogyo

Gakkaishi 36, 490–494.

Paneer S., Velmourougane K., Shanmukhappa J.R. and Naidu R. (2001)

Studies of microflora association during harvesting and on-farm

processing of coffee in India. Proc. 19th ASIC Coll., CD-ROM.

Parliment T.H. (2000) An overview of coffee roasting. In T.H. Parliment, CT.

Ho, P. Schieberle (eds), Caffeinated Beverages, Health Benefits,

Physiological Effects, and Chemistry. ACS symposium series No. 754,

pp. 188–201.

Pettitt B.C. Jr. (1987) Identification of the diterpene esters in arabica and

canephora coffees. J. Agric. Food Chem. 35, 549–551.

Picard H., Guyot B. and Vincent J.-C. (1984) E ´ tude des compose´s ste´roliques

de l’huile de cafe´ Coffea canephora. Cafe´ Cacao The´ 28, 47–62.

Pitt J.I., Taniwaki M.H., Teixeira A.A. and Iamanaka B.T. (2001)

Distribution of Aspergillus ochraceus, A. niger and A. carbonarius in

coffee in four regions of Brazil. Proc. 19th ASIC Coll., CD-ROM.

Pittet A., Tornare D., Huggett A. and Viani R. (1996) Liquid chromatographic

determination of ochratoxin A in pure and adulterated soluble

coffee using an immunoaffinity column cleanup procedure. J. Agric. Food

Chem. 44, 3564–3569.

Pokorny J., Nguyen-Huy C., Smidralova E. and Janicek G. (1975)

Nonenzymic browning. XII. Maillard reactions in green coffee beans

on storage. Z. Lebensm. Unters. Forsch. 158, 87–92.

Prodolliet J., Baumgartner M., Martin Y.L. and Remaud G. (1998)

Determination of the geographic origin of green coffee by stable isotope

techniques. Proc. 17th ASIC Coll., pp. 197–200.

Puzzi D. (1973) Conservac¸a˜o dos Gra˜os Armazenados. Armaze´ns e Silos. S.

Paulo: Editora Agronoˆmica Ceres.

Quijano Rico M. and Spettel B. (1975) Determinacion del contenido en

varios elementos en muestras de cafes de diferentes variedades. Proc. 7th

ASIC Coll., pp. 165–173.

174 Espresso Coffee

Rigitano A., Souza O.F. and Fava J.F.M. (1963) Coffee processing. In C.A.

Krug (ed.), Agricultural Practices and Fertilization of Coffee. Instituto

Brasilero Potassa (S. Paulo) (in Portuguese), pp. 215–259.

Roselius L. (1937) Die Erfindung des coffeinfreien Kaffes. Chemiker Zeitung 61

(1), 13.

Runge F. (1820) Neueste phytochemische Entdeckungen 1, 144–159.

Scholze A. and Maier H.G. (1983) Die Sa¨uren des Kaffees. VII. Ameisen,

A ¨

pfel-, Citronen- und Essigsa¨ure. Kaffee Tee Markt 33, 3–6.

Scholze A. and Maier H.G. (1984) Sa¨uren des Kaffees. VIII. Glycol- und

Phosphorsa¨ure. Z. Lebensm. Unters. Forsch. 178, 5–8.

Spadone J.C., Takeoka G. and Liardon R. (1990) Analytical investigation of

rio off-flavor in green coffee. J. Agric. Food Chem. 38, 226–233.

Speer K. (1989) 16-O-Methylcafestol – ein neues Diterpen im Kaffee –

Methoden zur Bestimmung des 16-O-Methylcafestols in Rohkaffee und

in behandelten Kaffees. Z. Lebensm. Unters. Forsch. 189, 326–330.

Speer K. (1991) 16-O-methylcafestol – a new diterpene in coffee; the fatty

acid esters of 16-O-methylcafestol. In W. Baltes, T. Eklund, R. Fenwick,

W. Pfannhauser, A. Ruiter and H.-P. Thier (eds), Proc. Euro Food Chem.

VI Hamburg, Germany, Volume 1. Hamburg: Behr’s Verlag, pp. 338–342.

Speer K. (1995) Fatty acid esters of 16-O-methylcafestol. Proc. 16th ASIC

Coll., pp. 224–231.

Speer K. and Mischnick P. (1989) 16-O-Methylcafestol – ein neues Diterpen

im Kaffee – Entdeckung und Identifizierung. Z. Lebensm. Unters. Forsch.

189, 219–222.

Speer K. and Mischnick-Lu¨bbecke P. (1989) 16-O-Methylcafestol – ein neues

Diterpen im Kaffee. Lebensmittelchemie 43, 43.

Speer K. and Montag A. (1989) 16-O-Methylcafestol – ein neues Diterpen im

Kaffee – Erste Ergebnisse: Gehalte in Roh- und Ro¨stkaffees. Dtsch.

Lebensm.-Rundsch. 85, 381–384.

Speer K., Hruschka A., Kurzrock T. and Ko¨lling-Speer I. (2000) Diterpenes in

coffee. In T.H. Parliment, C-T. Ho and P. Schieberle (eds), Caffeinated

Beverages, Health Benefits, Physiological Effects, and Chemistry. ACS

symposium series No. 754, pp. 241–251.

Speer K., Sehat N. and Montag A. (1993) Fatty acids in coffee. Proc. 15th

ASIC Coll., pp. 583–592.

Speer K., Steeg E., Horstmann P., Ku¨hn T. and Montag A. (1990)

Determination and distribution of polycyclic aromatic hydrocarbons

in native vegetable oils, smoked fish products, mussels and oysters,

and bream from the river Elbe. J. High Resolution Chromatogr. 13,


Speer K., Tewis R. and Montag A. (1991a) 16-O-Methylcafestol – ein neues

Diterpen im Kaffee – Freies und gebundenes 16-O-Methylcafestol.

Z. Lebensm. Unters. Forsch. 192, 451–454.

Speer K., Tewis R. and Montag, A. (1991b) 16-O-methylcafestol – a quality

indicator for coffee. Proc. 14th ASIC Coll., pp. 237–244.

The raw bean 175

Speer K., Tewis R. and Montag A. (1991c) A new roasting component in

coffee. Proc. 14th ASIC Coll., pp. 615–621.

Stennert A. and Maier H.G. (1994) Trigonelline in coffee. II. Content of

green, roasted and instant coffee. Z. Lebensm. Unters. Forsch. 199,


Stiles J.I., Moisyadi I. and Neupane K.R. (2000) Purified proteins,

recombinant DNA sequences and processes for producing caffeine free

beverages. US Patent 6,075,184.

Suggi Liverani F. (1991) A tool for the classification of green coffee samples.

Proc. 14th ASIC Coll., pp. 657–665.

Suggi Liverani F. (1995) Green coffee grading using fuzzy classification. In G.

Della Riccia, R. Kruse and R. Viertl (eds), Mathematical and Statistical

Methods in Artificial Intelligence. New York: Springer, pp. 237–245.

Teixeira A.A. (1978) Estudo preliminar sobre a qualidade do cafe´ no estado

de Sa˜o Paulo safra 78/79. 6


Congresso Brasileiro de Pesquisas Cafeeiras

(Ribeira˜o Preto, SP), pp. 316–322.

Teixeira A.A. and Ferraz M.B. (1963) A caracterizac¸a˜o da membrana

prateada nos cafe´s despolpados. A Rural (Sa˜o Paulo) 43, 28–29.

Teixeira A.A. and Figuereido J.P. (1985) Efeito do brunimento sobre a

qualidade do cafe´. Biolo´gico (Sa˜o Paulo) 51 (9), 233–237.

Teixeira A.A., Carvalho A., Moˆnaco L.C. and Fazuoli L.C. (1971) Grao˜s

defeituosos em cafe´ colhido verde. Bragantia (Campinas) 30(8), 77–90.

Teixeira A.A., Gomes F.P., Pereira L.S.P., Moraes R.S. and Castilho A.

(1969) A influeˆncia de gra˜os verdes em ligas com cafe´s de bebida mole.

Cieˆncia e Cultura 21, 355–356.

Teixeira A.A., Hashizume H., Nobre G.W., Cortez J.G. and Fazuoli L.C.

(1982) Efeito da temperatura de secagem na caracterizac¸a˜o dos efeitos

provenientes de frutos colhidos verdes. Proc. 10th ASIC Coll., pp. 73–


Teixeira A.A., Pereira L.S.P. and Pinto J.C.A. (1970) Classificac¸a˜o de cafe´ –

noc¸o˜es gerais. Ministe´rio da Indu´stria e do Come´rcio – Instituto

Brasileiro do Cafe´.

Teixeira A.A., Taniwaki M.H., Pitt J.I., Iamanaka B.T. and Martin C.P.

(2001) The presence of ochratoxin A in coffee due to local conditions

and processing in four regions in Brazil. Proc. 19th ASIC Coll., CDROM.

Teixeira A.A., Toledo A.C.D., Toledo J.L.B., Inskava J.M. and Azevedo

W.O. (1991) O prejuı´zo causado pelos gra˜os de cafe´ denominados

defeitos verdes e preto verdes. 17


Congresso de Pesquisas Cafeeiras

(Varginha, MG), pp. 25–27.

Tiscornia E., Centi-Grossi M., Tassi-Micco C. and Evangelisti F. (1979)

Sterol fractions of coffee seeds oil (Coffea arabica L.). Rivista Ital. Sostanze

Grasse 56, 283–292.

Tosello A. (1946) Studies on the drying of agricultural products. Bragantia 6

(2), 39–107 (in Portuguese).

176 Espresso Coffee

Trautwein E. (1987) Untersuchungen u¨ber den Gehalt an freien und

gebundenen Aminosa¨uren in verschiedenen Kaffeesorten sowie u¨ber

deren Verhalten wa¨hrend des Ro¨stens.Thesis, University of Kiel.

Trouche M.-D., Derbesy M. and Estienne J. (1997) Identification of robusta

and arabica species on the basis of 16-O-methylcafestol. Ann. Fals. Exp.

Chim. 90, 121–132.

Tsubouchi H., Terada H., Yamamoto K., Hisada K. and Sakabe Y. (1988)

Ochratoxin A found in commercial roast coffee. J. Agric. Food Chem. 36,


Van der Stegen G. and Blanc M. (1999) Report on the workshop

‘Enhancement of coffee quality by reduction of mould growth’. Proc.

18th ASIC Coll., pp. 219–222, and following papers.

Van der Stegen G.H.D. and Noomen P.J. (1977) Mass-balance of carboxy-5-

hydroxytryptamindes (C-5-HT) in regular and treated coffee.

Lebensmittelwiss. Technol. 10, 321–323.

Van der Stegen G.H.D. and Van Duijn, J. (1987) Analysis of normal organic

acids in coffee. Proc. 12th ASIC Coll., pp. 238–246.

Van der Stegen G., Blanc M. and Viani R. (2001) Highlights of the workshop

– Moisture management for mould prevention. Proc. 19th ASIC Coll.,


Viani R. (1993) The composition of coffee. In S. Garattini (ed.), Caffeine,

Coffee, and Health. New York: Raven Press, pp. 17–41.

Viani R. (2002) Effect of processing on ochratoxin A (OTA) content of

coffee. Adv. Med. Biol. 504, 189–193.

Viani R. (2003) Coffee physiology. In Encyclopedia of Food Science and

Technology, 2nd edn., Volume 3. London: Elsevier Science,

pp. 1511–1516.

Vilela A.R.E., Chandra P.K. and Oliveira G.A. (2000) Efeito da temperatura

e umidade relativa no branqueamento de gra˜os de cafe´. Ver. Bra´s. Vic¸osa

Especial 1, 31–37.

Vincent J.C. (1987) International standardization. In R.J. Clarke and R.

Macrae (eds), Coffee: Volume 1 – Technology. London: Elsevier Applied

Science, pp. 28–30.

Vitzthum O.G., Weisemann C., Becker R. and Ko¨hler H.S. (1990)

Identification of an aroma key compound in robusta coffees. Cafe´

Cacao The´ 34, 27–33.

Wajda P. and Walczyk D. (1978) Relationship between acid value of

extracted fatty matter and age of green coffee beans. J. Sci. Food Agric.

29, 377–380.

Weidner M. and Maier H.G. (1999) Seltene Purinalkaloide in Ro¨stkaffee.

Lebensmittelchemie 53, 58.

Wettstein A., Spillmann M. and Miescher K. (1945) Zur Konstitution des

Cafesterols 6. Mitt. Helv. chim. Acta 28, 1004–1013.

White D.R. (1995) Coffee adulteration and a multivariate approach to quality

control. Proc. 16th ASIC Coll., pp. 259–266.

The raw bean 177

Wilbaux R. (1967) Rapport sur les recherches en collaboration relatives aux

me´thodes de dosage de l’extrait soluble dans l’eau du cafe´ torre´fie´. Proc.

3rd ASIC Coll., pp. 77–85.

Wo¨hrmann R., Hojabr-Kalali B. and Maier H.G. (1997) Volatile minor acids

in coffee. I. Contents of green and roasted coffee. Dtsche Lebensm.-

Rundsch. 93, 191–194.

Wootton A.E. (1971) The dry matter loss from parchment coffee during filed

processing. Proc. 5th ASIC Coll., pp. 316–324.

Zosel K. (1971) Verfahren zur Entcoffeinierung von Rohkaffee. German Patent


178 Espresso Coffee



B. Bonnla¨ nder, R. Eggers,

U.H. Engelhardt and H.G. Maier


R. Eggers

At first sight, roasting of coffee seems to be a well-known and simple

process: It is just the application of heat to raw coffee beans. What is

important is to generate and control the correct temperatures at the right

moment, then stop the process when the aroma has fully developed and

the colour of the coffee is homogeneous throughout the whole bean.

However, on closer inspection, questions arise that have not yet received

an answer: the dependency of the instationary (changing with the time)

temperature distribution in the coffee bean on the parameters governing

the process, such as roast gas temperature, fluid flow conditions and

material properties of the coffee bean. The difficulty in seizing the whole

process comes from the dramatic changes in nearly all the parameters

related to the process: the temperatures, the material properties and

the geometry of the beans. Figure 4.1 demonstrates the meaning of

temperature distribution in a coffee bean during roasting.

The bean has a finite geometry of complicated shape; its inner structure

is heterogeneous; by heat admission its volume swells and the inner

structure changes. Mathematically, temperature becomes a three-dimensional

function of non-steady-state character with unknown moving

boundaries. From a chemical-engineering point of view the roasting

process consists of a combined heat and mass transport superposed by

endothermic and exothermic reactions. Thus the application of heat to the

coffee beans not only generates a temperature field, it also causes inner

pressures and a re-distribution of moisture depending on time and location.

These effects are illustrated in Figure 4.2: Heat energy is admitted to

the surface of the whole green bean, mainly by external hot gas flow, with

additional radiation and contact heat transfer depending on the type of


The temperature of the bean surface increases – with heat conduction

into the porous material – due to the temperature gradient. When the

local temperature reaches the evaporation temperature of the bean

moisture, a front of evaporation starts moving towards the centre of the

bean. During this first part of the roasting process the walls of the whole

bean are still relatively firm. Thus the vapour that has been generated

cannot permeate and the pressure build-up makes the bean volume

expand. Evaporation of the bean moisture, being an endothermic

operation, needs latent energy leading to a slowed down kinetic in the

temperature rise within the bean. The swelling and drying result in a

strong decrease in heat conductivity within the section between the

vaporization front and the outer surface of the bean. As a consequence the

temperature gradient is steeper in the dried region of the bean because of

the enhanced resistance to heat transfer. Mechanical and thermal stresses

moving towards the centre of the bean are created, which make the beans

crack or even burst if the superposed stresses overcome the tensile

strength of the bean.

Roasting reactions – browning with formation of flavour compounds at

elevated inner pressures – begin at higher temperatures (T>160


C) –

180 Espresso Coffee

Figure 4.1 Roasting of coffee beans – main aspects

starting at the bean surface and moving towards the inner dry preexpanded

structure of the bean. This second front of moving latent heat is

exothermic. Gaseous reaction products – mainly carbon dioxide – are

generated and entrapped within the cell structure, increasing inner

pressure until they permeate through the walls that are weakened and

partly destructed by the high temperatures as well (Perren et al., 2001).

The roasting process is then a counter-current process with heat transport

inside and mass outside. The phenomenological temperature profile

(Figure 4.2) shows that the distance between the two moving zones of

evaporation and roasting depends on the outer heat transfer, given by the

roasting technique applied, and on the structure of the bean. The latter is

linked with the geographical origin of the coffee to be roasted (Figure 4.3).

In order to achieve a roasting profile as homogeneous as possible, the

process must be precisely controlled, aiming at small temperature

gradients throughout the bean; on the contrary, fast roasting leads to an

overlapping of the evaporation and roasting steps and to an inhomogeneous


When the desired degree of roast is reached, the beans have to be

cooled down rapidly by water quenching or cold air in order to stop

further changes in colour, flavour and volume. The roasting loss is either

Roasting 181

Figure 4.2 Assumed profiles in the coffee bean during roasting

measured by the dry mass loss (the weight loss based on dry green beans)

or by the total mass loss (moisture plus organic matter loss). Table 4.1

shows the correspondence between roasting degree and dry mass loss

(Clarke, 1987). During roasting the density of the beans decreases to

nearly half its initial value.

The shape of a coffee bean can be defined as half an ellipsoidal body.

Attempts at modelling the temperature field of a coffee bean during

roasting have been made. The influence of the endothermic vaporization

enthalpy as well as of the exothermic reaction enthalpy is shown in

Figure 4.4. In the comparison the temperature increase has been

182 Espresso Coffee

Figure 4.3 Volume expansion for coffees of different origin

Table 4.1 Material data of arabica coffee beans

Coffee Mass Moisture




loss Density Volume Radius* Porosity

(g) (wt%) (wt%) (wt%) (g/ml) (ml) (mm) ( )

Green 0.15 10–12 0 0 1.2–1.4 0.11–0.13 3 <0.1


to dark


0.13 2–3 15–18 5–8 0.7–0.8 0.16–0.19 3.5 0.5

*Of a sphere with the same volume.

calculated without taking into consideration latent energies. The results,

when latent energies are introduced, indicate a delayed temperature rise

up to 150


C caused by vaporization followed by an accelerated

exothermic temperature increase above 180


C. Additional measurements

of three temperatures – at the surface, between core and surface and in the

core itself – show some remarkable aspects of roasting in a roasting

experiment lasting 600 seconds (Figure 4.5).

Starting at ambient conditions all three temperatures increase with

declining slope until nearly constant temperature has been reached.

During the first period (0–50 sec) of heating up the bean the difference

between core and surface temperatures passes a maximum of at least 70

degrees, whereas the temperature difference between core and half

distance to the surface of the bean slowly increases up to a maximum of

approximately 10 degrees throughout nearly the whole roasting time.

Obviously, the roasting process is much faster in the outer sections of the

bean compared to the core, probably leading to high thermal stresses in

the outer parts of the bean.

A small temperature drop occurs suddenly after 200 seconds’ roasting

time. At this time the inner temperature differences are clearly higher in

comparison to the outer temperature differences. However, the temperature

course drops slightly simultaneously throughout the whole bean.

After this unsteady phenomenon, temperatures increase more rapidly

than before! This could be explained by a pressure built up in the centre

Roasting 183

Figure 4.4 Temperature calculation of a roasting bean

of the bean due to water evaporation starting from the surface towards the

central region, but proceeding with decreasing velocity because the

driving force of heat transfer – the temperature difference to the ambient

– is diminishing all the time. So the pressure inside the bean increases

until the temperature overcomes the vaporization temperature, which is

higher at elevated pressure. Thus the remaining water in the centre of the

bean evaporates spontaneously and an endothermic flash creates the

slight temperature drop. Immediately after flashing the full exothermic

roast reaction is enabled to run to the core of the bean.

In order to roast the bean as homogeneously as possible, a stepwise

process with a slow increase of the temperature of the heating gas seems to

be advantageous. Further, a build-up of pressure in the bean is important

for the generation of sufficient aroma. Thus, temperature control of the

heating gas, permitting not only temperature profiles with moderate

differences, but also a sufficient pressure built up inside the bean, has to be

the objective of roasting.


R. Eggers

Conventional roasting techniques have been described by Sivetz and

Desrosier (1979), Rothfos (1984) and Clarke (1987). Recently, newer

184 Espresso Coffee

Figure 4.5 Experimental temperature profiles in the bean

developments have been summarized by Clarke and Vitzthum (2001).

Documentation is also available from manufacturers of roasting equipment,

e.g. Probat Werke at Emmerich and Neuhaus Neotec at Reinbek,

both in Germany.

From an engineering standpoint, the principles of roasting can be

described from a mechanical, thermal and operational point of view

(Figure 4.6).

In the case of traditional roasters, a tendency towards a larger variety in

coffee products has led to smaller uniform charges and therefore the

demand for large continuous roasters is decreasing in the roast and ground

market. Due to modern control techniques, very consistent products can

be obtained with batch roasters, where the heat input can be varied over

time. Roasters operating under pressure and steam roasters are unavailable

commercially, in spite of ongoing research.

Table 4.2 summarizes the basic principles of modern roasting

technology: a forced convective flow of hot gases passes through a

moving bed of coffee beans. The motion of the beans is either produced

by rotation or by the flow of roasting gases. Hot gas is either produced by

gas or by oil burners.

Due to increasing cost of energy and environmental considerations,

modern roasting equipment usually includes re-circulation of the exhaust

gas, after removal of the solid particles it carries (chaff, dust) by retention

on a cyclone. The gas is either brought back to the burner, or to a thermal

afterburner, operating at temperatures between 400 and 600


C, to burn

Roasting 185

Figure 4.6 Principal roasting techniques

186 Espresso Coffee

Table 4.2 Basic principles in modern roasting technology

Type Principle Characteristics




With/without perforated walls

Direct heating by convective flow of

hot gases

Indirect heating by hot drum walls


Continuously operated by an inner


Gas temperatures: 400–550



Roasting times: 8.5–20 min

Bowl Direct heating by convective flow of

hot gases

Continuously operated across the gas

stream; rotating

Gas temperatures: 480–550



Roasting times: 3–6 min



Direct heating by convective flow of

hot gases

Batch operated

Gas temperatures: 400–450



Roasting times: 3–6 min



Direct heating by fluidizing gas

Batch operated

Gas temperatures: 240–270



Roasting times: 5 min



Direct heating by fluidizing gas

Batch operated

Fast roasting:

Gas temperatures: 310–360



Roasting times: 1.5–6 min

Slow roasting:

Gas temperatures: 230–275



Roasting times: 10–20 min

off both the residual particles from above 1600 to well below 50 mg/Nm3

and the volatile organic matter (aerosols, condensate, etc.) present, as

required by law in some countries. Its residual thermal energy can be used

to preheat incoming fresh air (Illy and Viani, 1995, p. 94).

4.2.1 Energy balance and heat transfer

The investigation of the modes of heat transport from the heating gas to

the surface and inside the bean is rather complex, hence only approximate

calculations are possible. Nevertheless, the heat transfer theory helps one

to understand the roasting process. Assuming spherical coffee beans with

an average and constant diameter of 6mm, the tendency of the

convective heat transfer can be calculated for a given gas temperature.

The heat transfer coefficient   is shown versus the Reynolds number Re

in Figure 4.7. In a motionless system (vrel . 0), the heat transfer

coefficient is around 14W/m2K and a minimum fluidization velocity

(e.g. Re . 300) causes  -values in the range of 75W/m2K.

The heat transfer to single beans is different from the heat transfer to a

bed of beans. The calculated values are higher than those for single

spheres at the same superficial velocity and the actual roasting process is

expected to show heat transfer values in the region between the two

curves. Two important consequences arise: on the one hand, only a slight

increase in heat transfer is possible at higher velocities; on the other hand,

the improvement of the outer heat transfer is limited due to the heat

transfer resistance of the inner bean structure.

The amount of heat energy Q (KJ) transferred to the bean can be

calculated with the following equation:

Q .  


Acs.t. Tg



dt .4:1.


Acs(t) surface area of the coffee bean

(increasing during roasting)

Roasting 187



Tangential gas inlet

Spiral upward motion of the beans

Direct heat transfer of a moved packed


Gas temperatures: 280



Roasting times: 1.5–3 min

Tcs (t) surface temperature of the coffee bean

(increasing during roasting)

  convective heat transfer coefficient

Although the packed bed of coffee beans reveals the highest heat

transfer coefficient, the advantage of roasting using fluidizing gases

becomes obvious from equation 4.1. The surface area Acs of the fluidized

beans is in intimate contact with the hot gases and, as a result, the heat

transfer is more effective and homogeneous.

Figure 4.8 shows the balance of energy flow of a roasting section

operated continuously, where:


M mass flow

c heat capacity

T temperature

Index g gas

Index c coffee

Index in input

Index out output

The enthalpies of vaporization,  hv, and exothermic reaction,  hex,

have to be considered respectively as heat sinks and sources, in balancing

188 Espresso Coffee

Figure 4.7 Heat transfer of coffee beans during roasting

the energy flows. There is a further heat loss _QQloss from the roaster to the






. _MMc;in



.  hex



. _MMg;out



. _MMc;out


. _QQloss


The figures for latent energy are not yet very well known.

By neglecting the difference between _MMg;in and _MMg;out that is due to the

development of roast gases and flavour, and keeping the specific heat of

gas constant, the specific energy demand q.kJ=kg roastcoffee  becomes:

q .



  cg Tg;in



. _QQloss

h i




In the case of batch roasting, the temperature of the gas at the outlet of

the roaster increases with time and the specific roasting energy needed

can be calculated by integration:

q .








. _QQloss

h i





A comparison of the equation representing the energy demand with

the heat transfer equation indicates that the roasting process to be

energetically optimized must be operated with small temperature

differences and high heat transfer coefficients, meaning low temperatures

and high velocities of the hot gases.

Roasting 189

Figure 4.8 Balance of energy flow

4.2.2 High yield or fast roasting

Although processing at low temperatures has the advantage of a

homogeneous roast in favourable energetic conditions, there is an

alternative for enhanced heat transfer related to equations (4.1) and

(4.4): the increase of the gas temperature Tg and the hot gas to coffee


Fast roasting, i.e., providing the thermal energy required for roasting in

a shortened period of time, down to 90 seconds or even less (60 seconds

has been indicated as the shortest roasting time achievable now), has

been made possible by the development of forced convection roasting at

temperatures of 300–400


C. Because of the poor thermal conductivity of

the bean, there is a roasting gradient within the bean (particularly

noticeable at low roast temperatures), bean volume is increased with a

characteristic puffing by 10–15% with respect to conventional roasting,

bulk density is reduced below 300 g/l and extraction yield at brewing is

increased by 20%.

Summarizing, the profile of bean temperature variation during roasting

is an important parameter.

Both bean temperature and roasting time depend strongly on heat

transfer and therefore on the technology applied. The results of the LTLT

(Low Temperature Low Time) and HTST (High Temperature Short

Time) processes integrate into a slightly increasing function. Personal

measurements (Clarke and Vitzthum, 2001, p. 93) are in good agreement

with the literature data. Data on the variation of bean temperature during

roasting have been reported by Da Porto et al. (1991) and by Schenker et

al. (1999). The latter authors roasted 100 g of arabica beans as a fast

fluidized bed in a hot air flow of 0.01885m3/s in two ways: first the socalled

LTLT roasting with a hot gas temperature of 220


C and a roasting

time between 9 and 12 minutes, then HTST roasting at a gas temperature

of 260


C and roasting time between 2.6 and 3 minutes. In the case of

LTLT roasting, the bean temperature and other parameters were given as

a function of roasting time: the temperature rose continuously from 20 to



C in 2 minutes; this increase corresponds to 90% of the final increase



C after 14 minutes). The bean temperature data published by Da

Porto et al. (1991) were obtained with a laboratory roaster (HTLT) and

Brazilian Santos coffee. It can be seen that, in comparison with

this traditional laboratory roaster, a fluidized bed roaster produces a

significantly faster increase in bean temperature. However, because of the

poor thermal conductivity of the bean, there is a gradient of roasting level

within the bean, particularly noticeable at low roast; the volume is

increased by 10–15% with respect to conventional roasting with a

190 Espresso Coffee

characteristic puffing of the beans and a bulk density lower than below

300 g/l, and extraction yield at brewing 20% higher.

High yield roasting is not considered optimal for Italian espresso, owing

to the high residual content of chlorogenic acids, which carries into the

cup an astringent sour note (Maier, 1985; Illy and Viani, 1995, p. 96).


H. G. Maier

4.3.1 Physical changes

The main transformations occurring in the bean with increasing

temperature are given in Table 4.3, while the most important macroscopic

changes produced by roasting are indicated in Table 4.4 (Pittia et

al., 2001). The roasted coffee bean becomes brown to black, very brittle,

with an increase in volume up to 100% (dark roasting) and a corresponding

decrease of density (300–450 g/l in contrast to 550–700 g/l in

raw beans) (see Figure 4.10), with many macro- and micropores (Schenker

et al., 2000), the cellular structure ruptured, especially upon espresso

roasting. The water content should be near 1%, if no water quench is


Roasting 191

Figure 4.9 Roasting loss versus bean temperature – different processes

4.3.2 Chemical changes Overall changes

As shown in Table 4.3, during roasting, water (the moisture water of the

green beans and that generated by reactions) and carbon dioxide escape.

They are accompanied by some carbon monoxide and organic volatiles.

Water and carbon dioxide are generated by the very important Maillard

reaction, which leads to the coloured products, the melanoidins, and to

the main part of the organic volatiles. In addition, water and carbon

dioxide are produced by numerous other pyrolytic reactions. Table 4.4

summarizes the macroscopic changes occurring throughout the roasting

operation, and Figure 4.11 gives an idea of the effect on the contents of

the main groups of coffee constituents. The values for roasted coffee are

calculated on green coffee dry basis, in order to make a better comparison

of the changes. The changes within the main groups will be dealt with in

the following paragraphs. Carbohydrates

Of the mono- and disaccharides of green coffee, after roasting only traces

of free sugars remain. Sucrose is supposed to be partially hydrolysed, the

rest being pyrolysed (caramelized). From the reducing sugars or their

192 Espresso Coffee

Table 4.3 Macroscopic changes during roasting

Temperature change within

the bean (


C) Effect

20–130 Liquid-vapour transition of water (bean drying). Colour


130–140 First endothermic maximum. Yellow colouring and

swelling of bean with beginning of non-enzymatic

browning. Roast gases are formed and begin to evaporate

140–160 Complex series of endothermic and exothermic peaks.

Colour changes to light brown. Large increase in bean

volume and micropores. Rests of silverskin are removed.

Bean is very brittle. Some little fissures at the surface

occur. Aroma formation starts

160–190 Roasting reactions move towards the inner dry structure

of the bean (see also Figures 4.2 and 4.4)

190–220 Micro fissures inside the bean. Smoke escapes. Large

volumes of carbon dioxide escape and leave the bean very

porous. Typical flavour of roasted coffee appears

fragmentation products, many volatiles (aroma compounds, volatile acids)

and non-volatiles (melanoidins and their precursors, acids) are formed by

the Maillard reaction and, to a lesser extend, by caramelization. The

Maillard reaction has lower activation energy and is therefore favoured if

reactive nitrogen compounds (amino acids, free amino groups in proteins

and peptides) are present.

Roasting 193

Table 4.4 Macroscopic changes brought about by roasting

Parameter Comment

Colour of bean It fades to whitish-yellow at the beginning of roasting, it darkens

regularly with increasing temperature (see Figure 4.10); natural

arabicas and robustas need a higher final temperature to reach the

same colour as washed arabicas. There is a dark to light colour

gradient from the surface within the bean, particularly noticeable

with low-roasted fast-roasted coffees

Surface of bean Oil sweats to the surface rendering it brilliant, particularly at high

roasting levels

Structure of the


The large volumes of carbon dioxide freed render the bean porous

(see Figure 4.10c)

Brittleness It increases with roasting to a maximum, with an important

modification of the internal texture of the bean, which under the

microscope looks like a volcanic land covered with an amalgam of

the original constituents (see Figure 5.1)

Density Decreases regularly from 550 to 700 g l

 1 in raw beans to

300–450 g l

 1 in roasted beans (the lowest figures are attained with

fast-roasted coffee)

Hot water


It decreases slightly from green to light-roasted beans, then

increases again slightly with increasing roasting degree (highest with

fast-roasted coffee)

Moisture Both free water, which decreases regularly all through roasting, and

chemically bound water, liberated above 100


C, decrease to below

1% unless water quenching is applied (moisture release is less

efficient with fast-roasted coffee)

Organic losses Destruction of carbohydrates, chlorogenic acids, trigonelline, amino

acids becomes important above 160


C, varying between 1 and 5%

at the lightest commercial roasts, 5–8% for medium-roasted

coffees, to above 12% for dark-roasted ones (lower for fast-roasted


Release of CO2 continues for several days



Aroma content reaches a maximum at low roasting, while

destruction becomes more important than generation at medium

roasting losses (aroma generation is higher in fast-roasted coffees,

where a maximum is reached at medium roast)

pH of beverage In washed arabicas it increases from 4.9 for low-roasted coffees to

5.4 for dark-roasted one; higher for dry-processed coffees (lower

values for fast-roasted coffee)

The polysaccharides except cellulose are partially solubilized (Bradbury,

2001; Redgwell et al., 2002). Nevertheless the sum of soluble carbohydrates

is lower in roasted coffee. In espresso, the foam stability is related to

the amount of galactomannan and arabinogalactan and a function of the

degree of roast (Nunes et al., 1997). The carboxyatractyloglycosides are

decarboxylated quantitatively and the so generated atractyloglycosides

are degraded about 50% under ‘normal’ roasting conditions (Bradbury,

2001). Non-volatile lipids

Overall, there are only slight changes in the lipid fraction upon roasting.

The sterols and most of the triglycerides remain unchanged. The level of

194 Espresso Coffee

Figure 4.11 Main families of coffee constituents (average of arabica coffees)

Figure 4.10 Stereo microscope section of a bean: (a) green; (b) toasted to c.70



(c) roasted, showing the porous structure

trans fatty acids increases, especially the contents of C18:2ct and C18:2tc.

The linoleic acid content decreases slightly with roasting temperature.

The diterpenes, cafestol and kahweol, are decomposed to some extent.

Increasing with roasting temperature, dehydrocafestol, dehydrokahweol,

cafestal and kahweal are formed (0.5–2.5% of each educt). 16-Omethylcafestol

(in robusta beans) is less affected. Up to 20% of the

tocopherols and 25–50% of the carbonic acid 5-hydroxytryptamides are

destroyed (Wurziger and Harms, 1969; Wurziger, 1972; Speer and

Ko¨lling-Speer, 2001; Kurt and Speer, 2002). Proteins, peptides and amino acids

Crude protein content (Kjeldahl nitrogen multiplied by 6.25) changes

only very slightly upon roasting, but the nitrogenous components are

thoroughly changed. There is a loss of amino nitrogen of 20–40%, in dark

roasted espresso, about 50% (Thaler and Gaigl, 1963; Macrae, 1985).

Essentially all of the protein in the green coffee is denatured (Macrae,

1985). Some cross-bonds between the amino acid residues of the proteins

are formed (Homma, 2001). Parts of the L-amino acids are isomerized to

form D-amino acids, in commercial blends 7–70%, depending on the

individual acid (Nehring, 1991). The amino acid composition is changed

too: the more stable acids, e.g. glutamic acid, survive while others, e.g.

cysteine or arginine, decrease or are completely destroyed (Macrae, 1985).

In espresso, the foamability depends on the amount of protein in the

infusion and on the degree of roast (Nunes et al., 1997). Of the free amino

acids, only traces are left after roasting. The reaction products are

Maillard products (melanoidins, their precursors and volatiles) and

dioxopiperazines, of which the proline-containing ones are bitter (Ginz,

2001). Some intact amino acid residues are incorporated into the

melanoidins (Maier and Buttle, 1973). Chlorogenic acids

Since chlorogenic acids are the most important group of acids in green

coffee, they will be treated separately. These acids are for the most part

destroyed during roasting (Figure 4.12). In addition, some intact

chlorogenic acids may be bound by melanoidins (Heinrich and Baltes,

1987). It is assumed that the greatest part of chlorogenic acid is

hydrolysed. In addition, isomerization and, to a smaller extent, lactonization

takes place. The 6 lactones of the mono-hydroxycinnamoyl-quinic

acids are assumed to contribute to the bitter taste of roasted coffee (Ginz,

Roasting 195

2001). The 3 di-hydroxycinnamoyl-quinides are supposed to modulate the

effect of caffeine (Figure 4.12) (Martin et al., 2001) (see Chapter 10).

The free quinic acid is also isomerized (5 isomers), forming 4  -lactones

and 3  -lactones (Scholz-Bo¨ttcher et al., 1991; Homma, 2001). A little

part of this acid is decomposed to form simple phenols like hydroquinone.

The greatest part, however, remains unchanged, so that in dark roasts,

quinic acid is the most prevailing of the individual low molecular acids.

Little is known about the fate of caffeic acid. A minor part degrades to

form simple phenols. Two tetra-oxygenated phenylindans have been

found in model roasting of caffeic acid as well as in coffee at levels of

10–15 mg/kg, and neolignans (caffeicins) only in model roastings

(Clifford, 1997). The rest is assumed to be incorporated into the

melanoidins (see 4.5). Other acids

Most of the organic acids present in green coffee behave like the

chlorogenic acids: they are partially decomposed, e.g. citric acid to form

citraconic, itaconic, mesaconic, succinic, glutaric acids and eventually

some minor acids, malic acid to form fumaric and maleic acids (Scholz-

Bo¨ttcher, 1991; Ba¨hre and Maier, 1999). It is likely that esters of these

acids are formed upon roasting, because the acids are liberated on heating

the brew (Maier et al., 1984). Scholz-Bo¨ttcher (1991) detected a fumaric

acid/malic acid ester and another malic acid monoester after roasting of

196 Espresso Coffee

Figure 4.12 Quinides formed during roasting

malic acid. Phosphoric acid is stable, and its content increases by

hydrolysis of the inositol phosphates (Franz and Maier, 1994). A great

number of acids is generated by Maillard reaction or caramelization. The

most prominent are formic and acetic acids. Their contents reach a

maximum at medium roast; upon darker roasting volatilization prevails

over formation. Moreover (in decreasing order of the contents in

espresso), glycolic, lactic, metasaccharinic, threo-and erythro-3-deoxypentonic,

glyceric, 2-furanoic and several minor acids are formed (Ba¨hre

and Maier, 1999). Minerals

With the exception of phosphoric acid (see, the mineral content

does not change upon roasting. Alkaloids

Caffeine is stable upon roasting, but a small part is lost by sublimation.

Nevertheless this is often overcompensated by the roasting loss.

Trigonelline is partially decomposed. The remaining content is about

50% in a light roast, traces in a very dark roast (Viani and Horman, 1975;

Stennert and Maier, 1996). The products of model roastings are nicotinic

acid, N-methyl nicotinamide, methyl nicotinate and numerous volatile

nitrogen compounds, e.g. 46% pyridines and 3% pyrroles (Viani and

Horman, 1974). Most are found in coffee, especially nicotinic acid, which

constitutes only 1.5% of the degraded trigonelline, but nevertheless acts

as an important source of vitamin. The trigonelline/nicotinic acid ratio

can be used for a rough estimation of the degree of roast (Stennert and

Maier, 1996).


B. Bonnla¨ nder

Green coffee contains about 300 volatiles (Flament, 2001). The content

of some of them, e.g. 3-isobutyl-2-methoxypyrazine, is not changed by

roasting. The content of others, e.g. ethyl-3-methylbutyrate, is diminished,

but most of the volatiles increase upon roasting (Czerny and

Grosch, 2000; Flament, 2001). In addition, about 650 new volatiles have

been identified, bringing to more than 850 the number of volatiles

identified in roasted coffee.

Roasting 197

4.4.1 Generation of roast aroma

Most of the delightful aromatic character of coffee is the result of the

roasting process. Green coffee shows a typical green bell pepper-like

aroma, where isobutylmethoxypyrazine (MIBP) could be identified as the

character impact compound with a ‘peasy’ smell (Vitzthum et al., 1976).

Only part of the more than 300 additional green coffee aroma compounds

identified so far (Flament, 2001) survive the roasting process. The high

temperature (usually 170–230


C for 10–15 min) and the elevated

pressure inside the bean (up to 25 atm) trigger a vast number of chemical

reactions leading to dark colour and more than 1000 volatile and nonvolatile

compounds identified in roasted coffee (Stadler et al., 2002b). In

contrast with green coffee, no single compound but a mixture of approx.

25 very potent compounds represents the significant impression of coffee.

The aroma compounds (approx. 1 g/kg) are concentrated in the coffee oil.

The concentrations of the most potent ones are in the lower ppm (part

per million) or even ppt (part per trillion) range.

4.4.2 Precursors of aroma compounds

The non-volatile constituents play an important role as precursors in the

formation of the volatile compounds in roast coffee. Along with the rising

temperature, and the subsequent drying and swelling of the bean, the

following reactions take place under elevated pressure conditions:

n caramelization and degradation of carbohydrates forming mainly

aldehydes and volatile acids (Yeretzian et al., 2002);

n denaturation of proteins and reaction of free amino acids with

carbohydrates and their degradation products (Maillard reactions);

n production of phenols and taste active compounds from chlorogenic

acids (Pypker and Brouwer, 1970);

n degradation of trigonelline (Stadler et al., 2002a).

According to the different precursor compositions of the two main coffee

species, also a different aroma is observed. The aroma of Coffea canephora

(robusta) is characterized by higher amounts of phenolic compounds

(guaiacol, vinylguaiacol), perceived as harsh earthy notes, originating

from the chlorogenic acids (Vitzthum et al., 1990). Another interesting

class of aroma precursors is that of the glycosides. They can liberate the

bound aroma compounds enzymatically, during post-harvest treatment or

roasting (Weckerle et al., 2002), and are of special interest also in other

foods as aroma storage forms. The common precursor prenylalcohol is a

198 Espresso Coffee

source for sulphur odorants as shown by model reactions with sulphurcontaining

amino acids (Holscher and Steinhart, 1992).

4.4.3 Identification and characterization of aroma


Pioneer research on coffee aroma was already performed before 1926

when the Nobel Prize winners Reichstein and Staudinger succeeded in

identifying the first aroma compounds like furfurylthiol and the guaiacols

well before the discovery of gas chromatography (GC). After a period of

identifying more and more compounds by GC-mass spectrometry (GCMS)

the current trend in research is to identify the active-smelling ones

by GC-olfactometry (GC-O) in combination with the chemical structural

information obtained from GC-MS. Generally, the procedure requires

extraction and enrichment of the aroma compounds from the natural

material, followed by chromatographic fractionation, and finally the

authentic identification. Extraction procedures

The extraction procedure has to separate the volatile compounds from the

coffee matrix. A very important question is how representative is the

extract, or, in other words, how to avoid artefact formation during

extraction and enrichment (Sarrazin et al., 2000). Several techniques are

widely known in literature, such as simultaneous distillation extraction

(SDE) and high vacuum distillation (HVD) or solvent-assisted flavour

extraction (SAFE, Engel et al., 1999). In addition, the volatile compounds

in the headspace above ground or brewed coffee can be analysed by

headspace (HS) techniques either statically or in dynamic mode. Solid

phase micro extraction (SPME) can be operated without solvent in fully

automated manner either from the brew (IS–SPME) or from the

headspace (HS–SPME). Increasing the amount of stationary phase

dramatically improves the detection limits and is used as stir bar sorptive

extraction (SBSE) in the brew or in the headspace (HSSE), reported by

Bicchi et al. (2002). As the aroma compounds tend to be very volatile,

extraction losses are compensated by adding stable isotope-labelled

reference compounds. Instrumental sensory analysis

GC in combination with olfactometric detection (GC-O) helps to detect

potent odorants without knowing their chemical structure. The effluent

Roasting 199

of a gas chromatograph is split to a conventional detector and a heated

sniffing port, where trained people evaluate and record the sensory

impression of individual compounds (Holscher et al., 1990). Dilution

techniques are used to determine the so-called flavour dilution (FD)

factors. By a stepwise dilution of the aroma extract (1:1 by volume)

followed by GC-O analysis the most important contributors can be

smelled at the highest dilution, thus getting the highest FD factors. This

technique, developed by Ullrich and Grosch (1987), is known as aroma

extract dilution analysis (AEDA).The odour activity value (OAV) can be

expressed as the concentration divided by its threshold only if the

structure and the odour threshold of a substance are known. A method for

the prediction of OAV from FD factors is reported as well as the precision

and the optimal design of AEDA (Ferreira et al., 2001).

4.4.4 Aroma impact compounds in roasted coffee

In arabica coffee the most important contributors to the aroma of roast

and ground coffee are determined by the techniques mentioned above.

3-Mercapto-3-methylbutylformate (MMBF), 2-furfurylthiol, methional,

 -damascenone as well as two pyrazines and furanones show the highest

FD factors (Holscher et al., 1992). A comparison of the results of the

major research groups in coffee aroma is given in Table 4.5, together with

the aroma impressions of the substance.

The absolute amounts of the aroma compounds shown in the last

column differ by two orders of magnitude. The potency, especially that

of the sulphur compounds, is demonstrated by their low concentrations.

Because of a very low odour threshold (in air), only 130 mg/kg of MMBF

are sufficient to generate the strongest odour impression, which is more

than 10 000 times over its threshold of 0.003 ng/l of air (OAV of

37 000).

4.4.5 Effects on cup impression

Everybody who has ever had a really bad cup of coffee in direct

comparison to a perfect one knows how big the difference can be. The

perceived quality depends on objective criteria such as quality of green

coffee, but also on subjective or cultural preferences like type of


200 Espresso Coffee

Roasting 201

Table 4.5 Aroma impact compounds in roasted coffee powder

Odorant Odour impression Holscher1 (FD) Grosch2 (FD) Schenker3 (FD) Mayer4 (lg/kg)

Methanethiol Putrid, cabbage-like 25 – 4500

2-Methylpropanal Pungent, malty 100 – 24 000

2-Methylbutanal Pungent, fermented 100 16–128 28 600

2,3-Butanedione Buttery 200 – 256–1024 55 700

2,3-Pentanedione Buttery 100 – 4–128 28 300

3-Methyl-2-buten-1-thiol Animal-like, skunky 200 – 64–256 13

2-Methyl-3-furanthiol Roasted meat-like 500 128 32 60

Mercaptopentanone Sweaty, catty 100 – – –

2,3,5-Trimethylpyrazine Roasty, musty 200 64 16–32 6000

2-Furfurylthiol Roasty, coffee-like 500 256 1024 1350

2-Isopropyl-3-methoxy-pyrazine Peasy 100 128 – –

Acetic acid Vinegar-like 100 – – –

Methional Cooked potato 500 128 1024 148

2-Ethyl-3,5-dimethyl-pyrazine Roasty, musty 200 2048 1024 55

(E)-2-Nonenal Fatty 5 64 – 100

2-Vinyl-5-methyl-pyrazine Roasty, musty 200 – 53

3-Mercapto-3-methyl-butylformate Catty, roasted coffee-like 500 2048 1024 130

2-Isobutyl-3-methoxy-pyrazine Paprika-like 500 512 4–64 84

5-Methyl-5H-6,7-dihydro-cyclopentapyrazine Peanut-like 50 – –

2-Phenylacetaldehyde Honey-like 25 64 – 2500

3-Mercapto-3-methylbutanol Soup-like 100 32 – –

2/3-Methylbutanoic acid Sweaty 500 64 1024 25 000

(E)- -Damascenone Honey-like, fruity 500 1024 16–128 258

Guaiacol Phenolic, burnt 200 – 512–1024 3420

4-Ethylguaiacol Clove-like 25 256 – 1780

4-Vinylguaiacol Clove-like 200 512 256 45 100

Vanilline Vanilla-like – 32 – 4050

Peaks sorted by retention time on DB-wax column. Compounds with FD  32 are reported.

Sources: 1Holscher et al. (1990); 2Czerny and Grosch (2000); 3Schenker et al. (2002), range according to roasting isothermal high/low temperature; 4Mayer

and Grosch (2001) Green coffee quality

Correlations of geographical origins to the cup impression can be

recognized by a trained cup taster, but chemical mapping is also possible.

Modern statistical methods (e.g. analysis of variance, ANOVA) help to

map the composition of the aroma against the growing region (Freitas and

Mosca, 1999). Even a mapping against different growing conditions (e.g.

shade or sun) is possible (Bonnlaender, Unpublished results). Next in the

chain of production are the harvest and the post-harvest treatments of the

cherries. Errors in handling can lead to mould growth or other defects (see

2.3.6 and 3.7). Roasting

Older studies compared the composition of the volatiles extracted at

different degrees of roasting (Gretsch et al., 2000). The use of modern fast

analysis tools, such as resonance enhanced multiphoton ionization timeof-

flight mass spectrometry (REMPI/TOFMS), makes online monitoring

of the roasting process possible; a continuous monitoring with sampling

rates up to one per second can be performed showing the evolution

characteristics of specific compounds (Yeretzian et al., 2002). Generally,

most of the aroma compounds are generated at medium roast. Some

aroma compounds degrade at higher temperatures, whereas others like the

guaiacols, furfurylthiol and pyridine, show an increase up to very high

roasts; they contribute particularly to the aroma of dark roasted coffees

(Mayer et al., 1999). Preparation

During the preparation of the beverage the flavour compounds are

extracted from roast and ground coffee as a function of their solubility in

water. Whereas large amounts of 2,3-butandione, 2,3-pentandione and

the furanones get extracted from coffee, the yield of less polar compounds

like furfurylthiol is reduced (Mayer and Grosch, 2001). Figure 4.13 shows

the headspace stir bar sorptive extraction (HSSE) GC-MS analysis of an

espresso versus a drip cup to illustrate differences between the important

Italian beverage espresso and the drip preparation used worldwide.

Addition of milk to the beverage generally reduces the perceived coffee

aroma, increasing the creamy impression. Bucking and Steinhart (2002)

found a characteristic change in sensorial impression, depending on the

type of milk used (skim milk, coffee whitener).

202 Espresso Coffee Staling of coffee (see also 6.1.2 and 6.1.3)

The delicious aroma of fresh roast and ground coffee only lasts for a short

time. Soon after grinding the very fresh, mild, roasty notes diminish and

strong spicy notes appear in the flavour (Mayer and Grosch, 2001). In

particular, the very potent sulphur compounds deteriorate in contact with

air and the ground coffee starts smelling stale after about 10 days. The

structure of the whole bean provides protection to a certain degree by

keeping CO2 gas in its pores. Analytically the ratio of 2-methylfuran to

2-butanone (M/B ratio) can be used as a good freshness indicator, before

lipid oxidation leads to rancid products after several weeks (Arackal and

Lehmann, 1979). Packaging under inert gas (CO2 or N2) with a minimum

of residual oxygen guarantees long shelf life. Freezing also helps, slowing

down staling reactions. After the package has been opened, it is

recommended to store coffee in cool and dry conditions in the dark.

The staling reactions have not been completely explained. Hofmann and

Schieberle (2002) found reduction of sulphur compounds (MMBF,

furfurylthiol) in roast and ground and in liquid coffees by binding to

melanoidins. A radical mechanism is suggested by Blank et al. (2002).

Hydrolysis of chlorogenic quinic lactone leads to liberation in the brew of

the organoleptically undesirable quinic acid.

Roasting 203

Total lon Chromatogram (TIC): HSSE

Total lon Chromatogram (TIC): HSSE













Time 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00

Time 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00






From stir bar

From stir bar












Figure 4.13 Differences in preparation techniques. HSSE of an espresso (upper

total ion chromatogram) versus a drip coffee (lower chromatogram) analysed by

thermo desorption gas chromatography–mass spectrometry (TDS GC–MS EI)


U.H. Engelhardt

Melanoidins are coloured (brown) pigments developed during thermal

treatment of foods via the Maillard reaction or by dehydration

(caramelization) reactions of carbohydrates followed by polymerization

(Bradbury, 2001). Basically, melanoidins are formed by interactions

between carbohydrates and compounds, which possess a free amino group,

such as amino acids or peptides. Recently, melanoidins – not only those

from coffee – have attracted lots of interest as regards their occurrence in

foods and the corresponding impact on human health. This is

documented by the fact that the European Community set up a COST

action (COST Action 919: ‘Melanoidins in food and health’) starting in

1999 and ending in 2004. This action includes research on the separation

and characterization of melanoidins and related macromolecules, the

flavour binding, colour, texture and antioxidant properties of melanoidins

and investigation of the physiological effects and fate of melanoidins

(COST, 2002). The foods of relevance specified are coffee, malt, beer,

breakfast cereals and bread.

4.5.1 Chemistry

It can be anticipated that melanoidins are a very heterogeneous group of

compounds as regards molecular mass as well as the chemical and

biological properties. Taking this into account, one may argue that there

is no justification for referring to all those compounds as melanoidins. On

the other hand, is there any alternative? This situation is comparable with

other polymers in foods and beverages, e.g. the thearubigins in tea or

phlobaphenes in cocoa. It is agreed that those are constituents of the

beverage, and contents can be found in tables, but nobody really knows

what the properties and the structures of the compounds are.

What do we know about the formation of melanoidins? Most of the free

mono- and disaccharides are lost during the roasting process; the same is

true for free amino acids (Trautwein, 1987; Bradbury, 2001). They are at

least in part converted into melanoidins. Another part of the sugars

undergoes caramelization reactions, also yielding melanoidin-like pigments

(Bradbury, 2001) or is degraded yielding acidic compounds (Ginz et

al., 2000). A small proportion of the amino acids could be converted into

diketopiperazines (cyclic dehydration products of dipeptides), but,

according to Ginz (2001), it seems to be more likely that the

diketopiperazines are formed via protein degradation. Not much is yet

204 Espresso Coffee

known about the structures of melanoidins. As already mentioned above,

melanoidins are formed during the Maillard reaction along with a variety

of flavour compounds. The Maillard reaction in a food matrix is a

complex process because of the multitude of compounds present.

Consequently, information on the Maillard reaction is best generated

by model experiments with only a limited number of educts present;

which, however, still give rise to a considerable number of volatile and

non-volatile products. Reviews of the Maillard reaction can be found in

Flament (2001, p. 39).

Coffee melanoidins contain phenolic moieties (Heinrich and Baltes,

1987; Homma, 2001), as indicated by experiments using Curie point

pyrolysis high-resolution gas chromatography/mass spectrometry (HRGC/

MS), which gave 97 products, 33 of which were phenols (Heinrich and

Baltes, 1987). The molecular mass of the melanoidin fractions is between

3000 and 100 000 Dalton according to studies using gel chromatography

or HPLC as analytical tools (Homma, 2001). There seems to be more

high molecular weight material in robusta coffee compared to arabica

(Steinhart et al., 1989). Contents of melanoidins in coffee up to 30%

(Bradbury, 2001) or 29.4% (Belitz et al., 2001) are reported in the

literature – in the latter referred to as unknown constituents (colorants,

bitter compounds), while other authors give 15% for brewed coffee and

about 23% for roasted coffee on a dry weight base (Parliment, 2000).

However, it should be stressed that this is just a determination of the

melanoidins by difference and not at all a determination of the

compounds themselves. Regardless what the figures might be, melanoidins

make a relevant contribution in the cup.

Gel filtration chromatography was usually employed to purify the

melanoidin fractions according to the procedure of Maier et al. (1968)

and Maier and Buttle (1973). This fractionation includes extraction with

diethyl ether and adsorption of the water-soluble compounds on

polyamide. The compounds not adsorbed on polyamide were separated

into three fractions using Sephadex  G-25. The amount adsorbed gave

two fractions, one of which gave three subfractions. Details can be found

in papers by Maier et al. (1968), Maier (1981) and Macrae (1985). More

recent work usually also employs gel chromatography, but in most cases

without the polyamide step (Steinhart et al., 1989; Nunes and Coimbra,

2001). Basically, coffee is prepared by hot water extraction and defatted

using dichloromethane. The freeze-dried extract is either treated by

ultrafiltration or separated by gel chromatography (Sephadex  G-25;

eluent: water) and again freeze-dried (Hofmann et al., 2001; Steinhart et

al., 2001). Hydrolysis was often employed to detect information about the

moieties present in the polymers and sugars, amino acids and phenols.

Roasting 205

An interesting intermediate involved in melanoidin formation has

been recently detected by Hofmann et al. (1999) using EPR and LC-MS.

The compounds are named CROSSPY – (1,4-bis-5-amino-5-carboxy-1-

pentyl)pyrazinium radical cations – which are oxidized yielding diquaternary

pyrazinium ions. Figure 4.14 shows the reaction pathway proposed

by Hofmann and Schieberle (2002), and Hofmann et al. (2001).

Even though the formation of melanoidins is far from being fully

understood, these findings are an important step in gaining knowledge of

206 Espresso Coffee

Figure 4.14 Reaction scheme of melanoidin formation: (a) via CROSSPY radical;

(b) via diquaternary pyrazinium ions

Roasting 207

the mechanisms involved. The amounts of high-molecular mass melanoidins

increased with increasing roasting loss (Ottinger and Hofmann,


In a study on the interaction of isolated coffee melanoidins and selected

aroma compounds in model experiments Hofmann et al. (2001) detected

a decrease in free furfurylthiol, 3-methyl-2-butene-1-thiol and mercapto-

3-methylbutylformate, which was most significant when the 1500–3000

Da coffee melanoidins were used. In another experiment by Hofmann and

Schieberle (2002) with (2H2)-2-furfurylthiol the spectroscopic data

(2H-NMR, LC-MS) indicated that furfurylthiol reacts with oxidation

products of CROSSPY. The authors proposed that a covalent bond is

formed between these compounds leading to a decrease of the odour

quality of coffee right after brewing (especially as regards the sulphuryroasty

odour quality). Other useful techniques used in melanoidin/

Maillard research are capillary electrophoresis (Borelli et al., 2002; del

Castillo et al., 2002; Re et al., 2002) and matrix-assisted laser desorption

ionization time-of-flight mass spectrometry (MALDI/TOFMS) (Borelli

et al., 2002; Kislinger et al., 2002). Capillary electrophoresis (after a

fractionation by ultrafiltration with a 5000 Da cut-off dialysis membrane)

has been used by Ames et al. (2000) to monitor the development of

coloured compounds in coffee. It can be expected that both techniques

will be used in the characterization of coffee melanoidins.

4.5.2 Physiological effects of melanoidins (see also

Chapter 10)

Maillard reaction products in general may have negative and positive

(among those antioxidant activity) effects (Friedman, 1996). Claims have

been made that the melanoidins do contribute to the antioxidant activity

of coffee brews. Daglia et al. (2000) studied the in vitro antioxidant activity

and the ex vivo protective activity of coffee constituents and stated that the

higher molecular mass fractions had an antioxidant activity while the

lower molecular mass fractions had a protective activity. The antioxidant

properties are related to the degree of roast. Medium roasted coffee had the

most pronounced antioxidative effect (Steinhart et al., 2001). Del Castillo

et al. (2002) used the ABTS . assay to assess the contribution of high and

low molecular compounds to antioxidant activity of coffee brews. The

ABTS test is a spectrophotometric method which is based on the different

colour of 2,20-azino-di-(3-ethylbenzthiazoline-6-sulfonic acid and the

corresponding radical cation, by monitoring at 734 nm the decrease

in ABTS .concentration after addition of the antioxidant; Trolox

(6-hydroxy-2,5,7,8-tetramethylchroman-2-carbon acid) serves as reference.

Again, it was observed that medium-roasted coffee had the highest

antioxidant activity in vitro and that the low molecular mass compounds

had a greater contribution to the antioxidant activity compared to the

high molecular mass compounds. The same group analysed the antioxidant

activity of a coffee model system (consisting of chlorogenic acid, N-acetyl-

1-arginine, sucrose, and cellulose) and found that there was no relationship

between antioxidant activity and colour generation (Charurin et al.,

2002). Morales and Babbel (2002) also used an in vitro system (bleaching of

DMPD .) and found a significant lower antioxidant activity of melanoidins

compared with classical antioxidants such as caffeic or gallic acid. As

with other polymers in foods, there is practically no information available

on the bioavailability of the melanoidins. It seems reasonable to anticipate

that the fractions with different molecular mass also will have a different


Claims have been made that melanoidins suppress the formation of

n-nitroso amines (Wuerzner et al., 1989). Among other components of

coffee, such as trigonelline, nicotinic and chlorogenic acids, a low

molecular weight coffee melanoidin fraction (1000–3500 Da) contributes

to anti-adhesive properties against Streptococcus mutans adhesive activity

on saliva-coated hydroxyapatite beads, which might contribute to an

anticaries effect (Daglia et al., 2002). As animal data are missing

the authors recommend a careful interpretation of their findings.

Wijewickreme and Kitts (1998) stated a modulation of metal-induced

cytotoxicity by Maillard reaction products from coffee brews.

4.5.3 Summary and outlook

Our knowledge on the chemistry and physiology of coffee melanoidins

and of their impact on the taste of coffee are still fragmentary and far away

from a real structural elucidation. Recently, new information has become

available on one type of melanoidins containing CROSSPY and on their

interaction with sulphur-containing odorants. It might be expected that

our knowledge on melanoidins in general and coffee melanoidins in

particular will increase within the next few years as concentrated efforts

are made in this field of research (e.g. COST action). As soon as

individual structures of melanoidins are known, biological testing of the

compounds becomes possible as well as a more precise evaluation of the

impact of the melanoidins on coffee taste. These results might then

enable roasters to control the formation of melanoidins, which is very

important for the aroma and the overall taste of the coffee beverage.

208 Espresso Coffee


H.G. Maier

4.6.1 Mycotoxins (see also

The mycotoxins are largely destroyed upon roasting (Maier, 1991). For

instance, the destruction of aflatoxin B1 in artificially contaminated

green coffee beans during roasting ranges from 90 to 100% (Micco et al.,

1991). Beyond that, in non-mouldy beans, only ochratoxin A has been

found (Maier, 1989). In all the experiments where naturally contaminated

beans, a sampling procedure adapted to mycotoxin inhomogeinity

and roasting conditions within the range of actual practice were

employed, half to almost all of the ochratoxin A disappeared during

roasting (Blanc et al., 1998; Van der Stegen et al., 2001; Viani, 2002).

4.6.2 Polycyclic aromatic hydrocarbons (see also

In most cases, benzo[a]pyrene (BP) has been determined as a leading

substance for all polycyclic aromatic hydrocarbons. After normal roast the

content is often reduced in respect to green coffee. No differences

between roasting by direct firing and by indirect air heating could be

detected. More BP is formed by heat transfer to the beans from contact

with hot surfaces than by transfer from hot gases. The transfer of BP to the

coffee brew depends on the concentration of benzo[a]pyrene in the

roasted coffee and on the water-to-coffee ratio, amounting on average to

approximately 5%, so that the content in coffee brews and extracts is

insignificant (Maier, 1991).


Ames J.M., Royle L. and Bradbury A.G.W. (2000) Capillary electrophoresis

of roasted coffees. In T.H. Parliment, C.-T. Ho, P. Schieberle (eds),

Caffeinated Beverages. ACS Symposium Series 754, pp. 364–373.

Arackal A. and Lehmann G. (1979) Messung des Quotienten 2-Methylfuran/

2-Butanon von ungemahlenem Ro¨stkaffee wa¨hrend der Lagerung unter

Luftausschluß. Chem. Mikrobiol. Technol. Lebensm. 6, 43–47.

Ba¨hre F. and Maier H.G. (1999) New non-volatile acids in coffee. Dtsch.

Lebensm.-Rundsch. 95, 399–402.

Belitz H.D., Grosch W. and Schieberle P. (2001) Lehrbuch der

Lebensmittelchemie, 5th edn. Heidelberg: Springer, p. 934.

Roasting 209

210 Espresso Coffee

Bicchi C., Iori C., Rubiolo P. and Sandra P. (2002) Headspace sorptive

extraction (HSSE), stir bar sorptive extraction (SBSE), and solid phase

microextraction (SPME) applied to the analysis of roasted arabica coffee

and coffee brew. J. Agric. Food Chem. 50, 449–459.

Blanc M., Pittet A., Mun˜oz-Box R. and Viani R. (1998) Behavior of

ochratoxin A during green coffee roasting and soluble coffee manufacture.

J. Agric. Food Chem. 46, 673–675.

Blank I., Pascaul E., Devaud S., Fay L., Stadler R., Yeretzian C. and Goodman

B. (2002) Degradation of the coffee flavor compound furfuryl mercaptan

in model fenton-type reaction systems. J. Agric. Food Chem. 50,


Borelli R.C., Fogliano V., Monti S.M. and Ames J.M. (2002) Characterization

of melanoidins from a glucose-glycine model system (DOI 10.1007/

s00217-002-0531-0, 26 June 2002).

Bradbury A.G.W. (2001) Carbohydrates. In R. J. Clarke and O. G. Vitzthum

(eds), Coffee – Recent Developments. Oxford: Blackwell Science, pp. 1–17.

Bucking M. and Steinhart H. (2002) Headspace GC and sensory analysis

characterization of the influence of different milk additives on the flavor

release of coffee beverages. J. Agric. Food Chem. 50, 1529–1534.

Charurin P., Ames J.M. and del Castillo M.D. (2002) Antioxidant activity of

coffee model systems. J. Agric. Food. Chem. 50 (13), 3751–3756.

Clarke R. (1987) Roasting and grinding. In R.J. Clarke and R. Macrae (eds),

Coffee: Volume 2 – Technology. London: Elsevier.

Clarke R.J. and Vitzthum O. (2001) Coffee – Recent Developments. London:

Blackwell Science.

Clifford M.N. (1997) The nature of chlorogenic acids. Are they advantageous

compounds in coffee? Proc. 17th ASIC Coll., pp. 79–91.

COST (2002) see under

Czerny M. and Grosch W. (2000) Potent odorants of raw arabica coffee. Their

changes during roasting. J. Agric. Food Chem. 48, 868–872.

Da Porto C., Nicoli M.C., Severini C., Sensidoni A. and Lerici C.R. (1991)

Study on physical and physiochemical changes in coffee beans during

roasting, Note 2. Ital. J. Food Sci., pp. 197–207.

Daglia M., Papetti A., Gregotti C., Berte` F. and Gazzani G. (2000) In vitro

antioxidant and ex vivo protective activities of green and roasted coffee.

J. Agric. Food Chem. 48, 1449–1454.

Daglia M., Tarsi R., Papetti A., Grisoli P., Dacarro C., Pruzzo C. and Gazzani

G. (2002) Antiadhesive effect of green and roasted coffee on

Streptococcus mutans’ adhesive properties on saliva-coated hydroxyapatite

beads. J. Agric. Food Chem. 50, 1225–1229.

del Castillo M.D., Ames J.M. and Gordon M.H. (2002) Effect of roasting on

the antioxidant activity of coffee brews. J. Agric. Food Chem. 50,


Engel W., Bahr W. and Schieberle P. (1999) Solvent assisted flavour

evaporation – a new and versatile technique for the careful and direct

isolation of aroma compounds from complex food matrices. Eur. Food

Res. Technol. 209, 237–241.

Roasting 211

Ferreira V., Pet’ka J. and Aznar M. (2001) Aroma extract dilution analysis.

Precision and optimal experimental design. J. Agric. Food Chem. 50,


Flament I. (2001) The volatile compounds identified in green coffee beans. In

Coffee Flavor Chemistry. Chichester: J. Wiley & Sons, pp. 29–34.

Franz H. and Maier H.G. (1994) Inositolphosphate in Kaffee und

Kaffeemitteln. II. Bohnenkaffee. Dtsch. Lebensm.-Rundsch. 90, 345–349.

Freitas A. and Mosca A. (1999) Coffee geographic origin – an aid to coffee

different. Food Res. Int. 32, 565–573.

Friedman M (1996) Food browning and its prevention: an overview. J. Agric.

Food Chem. 44, 631–653.

Ginz M. (2001) Bittere Diketopiperazine und Chlorogensa¨urederivate in

Ro¨stkaffee. Thesis, TU Braunschweig, Germany.

Ginz M., Balzer H.H., Bradbury A.G.W. and Maier H.G. (2000) Formation of

aliphatic acids by carbohydrate degradation during the roasting of coffee.

Eur. Food Res. Technol. 211, 404–410.

Gretsch C., Sarrazin C. and Liardon R. (2000) Evolution of coffee aroma

characteristics during roasting. Proc. 18th ASIC Coll., pp. 27–34.

Heinrich L. and Baltes W. (1987) Vorkommen von Phenolen in Kaffee-

Melanoidinen. Z. Lebensm. Unters. Forsch. 185, 366–370.

Hofmann T. and Schieberle P. (2002) Chemical interactions between odoractive

thiols and melanoidins involved in the aroma staling of coffee

beverages. J. Agric. Food Chem. 50, 319–326.

Hofmann T., Bors W. and Stettmaier K. (1999) Radical-assisted melanoidin

formation during thermal processing of foods as well as under

physiological conditions. J. Agric. Food Chem. 47, 391–396.

Hofmann T., Czerny M., Calligaris S. and Schieberle P. (2001) Model studies

on the influence of coffee melanoidins on flavor volatiles of coffee

beverages. J. Agric. Food Chem. 49, 2382–2386.

Holscher W. and Steinhart H. (1992) New sulfur-containing aroma-impact

compounds in roasted coffee. Proc. 14th ASIC Coll., pp. 130–136.

Holscher W., Vitzthum O. and Steinhart H. (1990) Identification and

sensorial evaluation of aroma-impact-compounds in roasted Colombian

coffee. Cafe´ Cacao The´ 34, 205–212.

Holscher W., Vitzthum O. and Steinhart H. (1992) Prenyl alcohol – source

for odorants in roasted coffee. J. Agric. Food Chem. 40, 655–658.

Homma S. (2001) Non-volatile compounds, Part II. In R.J. Clarke and O.G.

Vitzthum (eds), Coffee – Recent Developments. Oxford: Blackwell

Science, pp. 50–67.

Illy A. and Viani R. (eds) (1995) Espresso Coffee. London: Academic Press.

Kislinger T., Humeny A., Seeber S., Becker C.-M. and Pischetsrieder M.

(2002) Qualitative determination of early Maillard-products by MALDITOF

mass spectrometry peptide mapping. Eur. Food Res. Technol. 215,


Kurt A. and Speer K. (2002) Untersuchungen zum Einfluss der

Da¨mpfungsparameter auf die Diterpengehalte von Arabica Roh- und

Ro¨stkaffees. Dtsch. Lebensm. Rdsch. 98, 1–4.

212 Espresso Coffee

Macrae R. (1985) Nitrogenous compounds. In R. J. Clarke and R. Macrae

(eds), Coffee: Volume 1 – Chemistry. Barking: Elsevier Applied Science,

pp. 115–152.

Maier H.G. (1981) Kaffee. Berlin: Paul Parey, pp. 64–67.

Maier H.G. (1985) Zur Zusammensetzung kurzzeitgero¨steter Kaffees.

Lebensmittelchem. Gerichtl. Chem. 39, 25–29.

Maier H.G. (1989) Zum Stand der Forschungen u¨ber Kaffee. Lebensmittelchem.

Gerichtl. Chem. 43, 25–33.

Maier H.G. (1991) Teneur en compose´s cance´rige`nes du cafe´ en grains. Cafe´

Cacao The´ 35, 133–142.

Maier H.G. and Buttle H. (1973) Zur Isolierung und Charakterisierung der

braunen Kaffeero¨ststoffe. II. Mitteilung. Z. Lebensm. Unters. Forsch. 150,


Maier H.G., Diemair W. and Ganssmann J. (1968) Zur Isolierung und

Charakterisierung der braunen Kaffeero¨ststoffe. Z. Lebensm. Unters.-

Forsch. 137, 282–292.

Maier H.G., Engelhardt U.H. and Scholze A. (1984) Sa¨uren des Kaffees. IX.

Zunahme beim Warmhalten des Getra¨nks. Dtsch. Lebensm. Rdschau 80,


Martin P.R., Depaulis T. and Lovinger D.M. (2001) Non-caffeine dicinnamoylquinide

constituents of roasted coffee inhibit the human

adenosine transporter. Proc. 19th ASIC Coll., CD-ROM.

Mayer F. and Grosch W. (2001) Aroma simulation on the basis of the

odorant composition of roasted coffee headspace. Flavour Fragr. J. 16,


Mayer F., Czerny M. and Grosch W. (1999) Influence of the provenance and

roast degree on the composition of potent odorants in Arabica coffees.

Eur. Food Res. Technol. 209, 242–250.

Micco C., Miraglia M., Brera C., Desiderio C. and Masci V. (1991) The effect

of roasting on the fate of aflatoxin B1 in artificially contaminated green

coffee beans. Proc. 14th ASIC Coll., 183–189.

Morales F.J. and Babbel M.B. (2002) Melanoidins exert a weak antiradical

activity in watery fluids. J. Agric. Food Chem. 50, 4657–4661.

Nehring U. (1991) D-Aminosa¨uren in Ro¨stkaffee. Thesis, TU Braunschweig,


Nunes F.M. and Coimbra M.A. (2001) Chemical characterization of the high

molecular weight material extracted with hot water from green and

roasted arabica coffee. J. Agric. Food Chem. 49, 17773–1782.

Nunes F.M., Coimbra M.A., Duarte A.C. and Delgadillo I. (1997)

Foamability, foam stability, and chemical composition of espresso coffee

as affected by the degree of roast. J. Agric. Food Chem. 45, 3238–3243.

Ottinger H. and Hofmann T. (2001) Influence of roasting on the melanoidin

spectrum in coffee beans and instant coffee. In: Proceedings of the COST

Action 919, Volume 2, pp. 119–125.

Parliment T.H. (2000) An overview of coffee roasting. In T.H. Parliment,

C.-T. Ho and P. Schieberle (eds), Caffeinated Beverages. ACS

Symposium Series 754, pp. 188–201.

Roasting 213

Perren R., Geiger R. and Escher F. (2001) Mechanism of volume expansion in

coffee beans during roasting. Proc. 19th ASIC Coll., CD-ROM.

Pittia P., Manzocco L. and Nicoli M.C. (2001) Thermophysical properties of

coffee as affected by processing. Proc. 19th ASIC Coll., CD-ROM.

Re R., Pellegrini N., Proteggente A., Pannala A., Yang M. and Rice-Evans C.

(1999) Antioxidant activity applying an improved ABTS radical cation

decolorization assay. Free Radic. Biol. Med. 26, 1231–1237.

Redgwell R.J., Trovato V., Curti D. and Fischer M. (2002) Effect of roasting

on degradation and structural features of polysaccharides in arabica

coffee beans. Carbohydrate Research 337, 421–431.

Rothfos B. (1984) Kaffee. Hamburg: Gordian.

Sarrazin C., Le Quere J.-L., Gretsch C. and Liardon R. (2000)

Representativeness of coffee aroma extracts – a comparison of different

extraction methods. Food Chem. 70, 99–106.

Schenker S., Handschin S., Frey B., Perren R. and Escher F. (1999) Structural

properties of coffee beans as influenced by roasting conditions. Proc. 18th

ASIC Coll., pp. 127–135.

Schenker S., Handschin S., Frey B., Perren R. and Escher F. (2000) Pore

structure of coffee beans affected by roasting conditions. J. Food Sci. 65,


Schenker S., Heinemann C., Huber M., Pompizzi R., Perren R. and Escher F.

(2002) Impact of roasting conditions on the formation of aroma

compounds in coffee beans. J. Food Sci. 67, 60–66.

Scholz-Bo¨ttcher B. (1991) Bildung von Sa¨uren und Lactonen, insbesondere

aus Chlorogensa¨uren, beim Ro¨sten von Kaffee. Thesis, TU

Braunschweig, Germany.

Scholz-Bo¨ttcher B.M., Ernst L. and Maier H.G. (1991) New stereoisomers of

quinic acid and their lactones. Liebigs Ann. Chem. 1991, 1029–1036.

Sivetz M. and Desrosier N.W. (1979) Coffee Technology. Westport, CT: AVI.

Speer K. and Ko¨lling-Speer I. (2001) Lipids. In R.J. Clarke and O.G.

Vitzthum (eds), Coffee – Recent Developments. Oxford: Blackwell

Science, pp. 33–49.

Stadler R., Varga N., Hau J., Vera F. and Welti D. (2002a) Alkylpyridiniums.

1. Formation in model systems via thermal degradation of trigonelline. J.

Agric. Food Chem. 50, 1192–1199.

Stadler R., Varga N., Milo C., Schilter B., Vera F. and Welti D. (2002b)

Alkylpyridiniums. 2. Isolation and quantification in roasted and ground

coffees. J. Agric. Food Chem. 50, 1200–1206.

Steinhart H., Luger A. and Piost J. (2001) Antioxidative effect of coffee

melanoidins. Proc. 19th ASIC Coll., CD-ROM.

Steinhart H., Mo¨ller A. and Kletschkus H. (1989) New aspects in the analysis

of melanoidins in coffee with liquid chromatography. Proc. 13th ASIC

Coll., pp. 197–205.

Stennert A. and Maier H.G. (1996) Trigonelline in coffee. Part 3.

Calculation of the degree of roast by the trigonelline/nicotinic acid

ratio. New gas chromatographic method for nicotinic acid. Z. Lebensm.

Unters. Forsch. 202, 45–47.

Thaler H. and Gaigl R. (1963) Untersuchungen an Kaffee und Kaffee-Ersatz.

VIII. Das Verhalten der Stickstoffsubstanzen beim Ro¨sten von Kaffee. Z.

Lebensm. Unters. Forsch. 120, pp. 357–363.

Trautwein E. (1987) Untersuchungen u¨ber den Gehalt an freien und

gebundenen Aminosa¨uren in verschiedenen Kaffee-Sorten sowie u¨ber

das Verhalten wa¨hrend des Ro¨stens. Thesis University of Kiel, Germany.

Ullrich F. and Grosch W. (1987) Identification of the most intensive volatile

flavor compounds formed during autoxidation of linoleic acid. Z.

Lebensm. Unters. Forsch. 184, 277–282.

Van der Stegen G.H.D., Essens P.J.M. and van der Lijn J. (2001) Effect of

roasting conditions on reduction of ochratoxin A in coffee. J. Agric. Food

Chem. 49, 4713–4715.

Viani R. (2002) Effect of processing on ochratoxin A (OTA) content of

coffee. Adv. Exp. Med. Biol. 504, 89–93.

Viani R. and Horman I. (1974) Thermal behaviour of trigonelline. J. Food Sci.

39, 1216–1217.

Viani R. and Horman I. (1975) Determination of trigonelline in coffee. Proc.

7th ASIC Coll., 273–278.

Vitzthum O., Weissmann C., Becker R. and Ko¨hler H. (1990) Identification

of an aroma key compound in robusta coffees. Cafe´ Cacao The´ 34, 27–36.

Vitzthum O., Werkhoff P. and Ablanque E. (1976) Fluechtige Inhaltsstoffe

des Rohkaffees. Proc. 7th ASIC Coll., pp. 115–123.

Weckerle B., Gati T., Toth G. and Schreier P. (2002) 3-Methylbutanoyl and

3-methylbut-2-enoyl disaccharides from green coffee beans (Coffea

arabica). Phytochem. 60, 409–414.

Wijewickreme A.N. and Kitts D.D. (1998) Modulation of metal-induced

cytotoxicity by Maillard reaction products from coffee brew. Toxicol.

Environm. Health, Part A 55, 151–168.

Wuerzner H.P., Jaccaud E. and Aeschbacher U. (1989) In vivo inhibition of

nitrosamide formation by coffee and coffee constituents. Proc. 13th ASIC

Coll., pp. 73–81.

Wurziger J. (1972) Carbonsa¨urehydroxytrpyptamide oder a¨therlo¨sliche

Extraktstoffe zum Nachweis und zur Beurteilung von bearbeiteten

beko¨mmlichen Ro¨stkaffees. Kaffee- und Tee-Markt 22 (14), 3–11.

Wurziger J. and Harms U. (1969) Beitra¨ge zum Genußwert und zur

Beko¨mmlichkeit von Rohkaffee. IV. Hydroxytryptamide in Ro¨stkaffees

aus normalen Handelssorten sowie aus in verschiedener Weise bearbeiteten

Rohkaffees. Kaffee- und Tee-Markt 19 (18), 26–29.

Yeretzian C., Jordan A., Badoud R. and Lindinger W. (2002) From the green

bean to the cup of coffee: investigating coffee roasting by on-line

monitoring of volatiles. Eur. Food Res. Technol. 214, 92–104.

214 Espresso Coffee



M. Petracco

The classical representation of coffee usually pictures it in beans, either

green or roasted. But for coffee to be consumed as food, it is necessary to

destroy the beautiful form nature has given the seed by transforming it

into a powder. The operation which effects this conversion is called, in

engineering language ‘comminution’, which means breaking down

particles into smaller fragments, while in coffee industry jargon it is

commonly known as grinding.

Grinding is performed by a grinder, or coffee-mill, and the resulting

product is ground coffee. The term ‘grind’ applies to an important variable

in espresso coffee preparation, as we shall see later: it refers to the degree

of comminution associated with the concept of fineness of ground coffee.

The main objective of grinding is to increase the specific extraction

surface, or rather, to increase the extent of the interface between water

and the solid per unit weight of coffee, so as to facilitate the transfer of

soluble and emulsifiable substances into the brew. Each time a solid body

is broken up, additional surface is generated that comes into contact with

the surrounding environment, in this case the extraction water.

Two apparently contradictory needs must be satisfied to prepare a good

cup of espresso: on the one hand a short percolation time is required,

while, on the other hand, high concentration of soluble solids must be

reached. Both requirements can only be attained if a close contact

between solid particles and extraction water can be achieved. Thus,

espresso percolation needs a plurimodal particle size distribution, where

the finer particles enhance the exposed extraction surface (chemical

need) and the coarser ones allow the water flow (physical need).


Before dealing with the various types of devices employed for grinding, it

is useful to give a qualitative notion of the fundamental laws of physics

acting in comminution. To reduce solid blocks into smaller sizes, it is

necessary to produce fractures by applying a set of forces. A quantitative

theoretical analysis of this process would only be possible for very simple

geometric forms; nevertheless, a general outline may be useful to

understand the machine’s application of complex stress configurations

on the material (Fayed and Otten, 1984).

Stresses exerted on an elementary object fall into two categories:

compressive stresses and shear stresses. Both produce strains that may

follow either simple laws, like Hooke’s law in the case of elastic materials

(” .  E, where ” stands for strain caused by stress  , and E represents the

modulus of elasticity or Young’s modulus), or more complex laws for

materials displaying a plastic or viscous-plastic behaviour, which is not

the case of roasted coffee. A fracture results when the stress exceeds the

limit given by Hooke’s law, that is, when it reaches the so-called ‘breaking

load’ (Pittia and Lerici, 2001).

In reality, the stress is not distributed uniformly inside an object, not

even in the ideal case of an object having a regular form and consisting of

a homogeneous and isotropic material. As a matter of fact, the interior of

an object always displays slight discontinuities, or flaws, like the defects in

the crystal lattice of a mineral, or inhomogeneities in animal or plant


Lines of force are concentrated along the rim of discontinuities, and

may easily cause the stresses to reach very high values, which lead to

local micro-fractures that propagate rapidly. In the case of shear stresses,

the relevant flaws tend to lie on the plane of shear, and cause the

formation of a roughly uniform surface of fracture (depending on the

nature of the material), which generally splits the object into two. In the

case of compression forces, the level of stresses is more uniform, and the

flaws may be randomly scattered within the object: a concurrent

formation of fracture lines in various points causes the object to burst,

creating fragments of different random sizes. In an object that does not

possess an elementary form, but has an articulated or compound

geometry, the most stressed points susceptible to failure can be

determined by the use of the design techniques employed in construction


The geometry of a coffee bean is anything but simple, as is often the

case with the works of nature. The coffee bean displays two distinct levels

of complexity:

216 Espresso Coffee

n At a macroscopic level, the bean is made of a layer of tissue folded

together (interposed with a thin layer of a different tissue, the

perisperm, also called ‘silverskin’), forming a solid resembling an

ellipsoid of revolution, as was shown in Figure 4.10c.

n At a microscopic level, the tissue is composed of cells of different

shapes filled with a mass of sugars, proteins and lipids, enclosed by

walls made of polysaccharides (see Figure 3.21b). This content can be

observed only in green coffee, while it apparently disappears after

roasting due to transformation into polymeric material and volatile

compounds (Figure 5.1).

A study of the cellular geometry of roasted coffee, performed by

computer simulation employing the ‘finite element method’, shows that

the application of shear to a reticular structure tends to produce a

comminution into many small particles and a few larger ones of different

size, as shown in Figure 5.2 (Petracco and Marega, 1991).

Thus, two phases of comminution must be taken into consideration in

the design of a grinder:

Grinding 217

Figure 5.1 SEM structure of roasted coffee bean, where the cell walls are no

longer visible, as they are covered by oil

n a crushing phase, where the convolute structure of the brittle roasted

bean is broken down into fragments of approximately one millimetre;


n a second phase, properly called grinding, in which such fragments are

submitted to shearing forces.

This grinding process produces a particle distribution that, as proposed in

section 7.5, has a favourable if not indispensable effect on espresso

brewing, though theoretically it does not contribute to maximization of

the surface.


Roasted coffee possesses a particular feature that excludes from the outset

the use of a major class of grinders widely employed in the mining

industry, naming ball mills. The solubility of coffee in water, even in cold

water, does not allow the use of wet-processing equipment. This is

unfortunate, because wet-processing equipment is most appropriate when

a specific distribution spectrum is sought for, since it operates with

adjusted recycling of material. This only leaves the use of dry-processing

equipment, operating according to two commonly employed operating

principles: ‘impact’ grinding and ‘gap’ grinding.

218 Espresso Coffee

Figure 5.2 Simulation of sub-particle formation

5.2.1 Impact grinding

Impact grinding is accomplished by a series of blades, rotating at high

speed, which exert a shock impulse on the particles encountered in their

trajectory. This effect resembles grinding by compression, with the

difference that, in the case under consideration, the reaction is given by

the inertial force of the particle suddenly removed from its state of

relative rest. This technique is utilized only in small-size grinders in the

home, because of one major defect: it does not easily allow the regulation

of particle size, since the ground product remains between the blades even

after being ground and so undergoes further impacts, making powder

fineness also dependent on the length of the operation.

5.2.2 Gap grinding

Gap grinding is based on the passage of the beans, usually by dropping,

through a gap between moving tools, called ‘cutting tools’. The geometry

of the cutters causes a gradual reduction of the width of the gap during

rotation, forcing the particles to come into contact with both cutters,

which apply a force couple on each particle. According to the shape of

the cutting surfaces, the force exerted may either be of a compressive or of

a shear type. The most common styles of cutter pairs used in the coffee

industry are (Figure 5.3):

n Roller cutters: a couple of rifled cylinders with parallel axes, counterrotating

and juxtaposed so as almost to touch themselves at their

generating lines. Beans are fed into the device by timed dropping, so

as not to obstruct the gap between cutters. These tools are mainly

used at industrial level.

Grinding 219

Figure 5.3 Cutters: (left) roller cutters; (centre) conical cutters; (right) flat cutters

n Conical cutters: a male conical wheel rotating coaxially in a static

truncated-cone-shape cavity. Again, feeding is performed by gravity,

often by rejection, namely a column weight of the material to be

ground. Excellent size versus throughput ratio.

n Flat cutters: a couple of disks with truncated-cone cavity, juxtaposed

so as to almost touch their bases. One of them, usually the lower one,

rotates coaxially. Feeding, by rejection or by forcing with worm screw

system, is accomplished from the upper to the lower part. Ground

coffee powder does not come out by gravity, but by the effect of the

centrifugal force of rotation. These cutters are used both for industrial

and professional grinding.

As already stated, the design of a grinding device must take into

consideration the existence of two phases of comminution: the crushing

of the bean (also called pre-breaking) and the grinding of the fragments to

the desired fineness (also called finishing). These requirements can be

met by two systems:

n A rather inexpensive system widely used in professional (coffee-bar

usage) or home grinders fitted with conical or flat cutters, where

cutters with unequal teeth are used. They are capable of forming gaps

of different size, and of seizing whole beans as well as fragments.

n A system mostly reserved to industrial equipment because of its

manufacturing complexity uses a series of grinding stages, each

optimized for a specific particle size. In this configuration, each stage

may also function as a regulator for the feeding of the following stage:

in this case, rejection feeding is utilized only in the first stage, that is,

in the stage that receives whole beans.

Obviously, a greater complexity of configuration is matched not only by

higher equipment costs, but also by a greater difficulty in controlling

stability. For this reason, industrial grinders are manufactured mainly

according to two basic models:

n A grinder with one or two pairs of toothed rollers, in series, preceded

by another pair of pre-breaking rollers, with capacities between

100 kg/h and several tons per hour.

n A grinder with a pair of flat cutters preceded by a conical prebreaking

stage, used for small to medium throughput (50–500 kg/h) of

ground product.

220 Espresso Coffee

5.2.3 Homogenization step

Besides the cutting tools, grinding equipment usually includes a

homogenization and blending step for the ground product, consisting of

rotating units with blades, worm screws, etc. Once the product has been

ground, it is transported to the packaging section of the plant by means of

conveyor belts, cups or pneumatic conveyors, sometimes under inert




The objective of grinding is to obtain a distribution of particles suitable

for a specific purpose. This statement is demonstrated by a practical

method applied by espresso bartenders for checking the fineness of

ground coffee before introducing it into the espresso machine: the

operator draws a single portion of coffee powder from the grinder, and uses

it to prepare a cup of coffee. The quantity of liquid percolated in a given

time is used as a measurement unit for regulating the primary variable of

the grinder: the so-called ‘notch’, namely the cutters’ position determining

the shear gap.

This method has the clear advantage of allowing an immediate

feedback, and an intuitive understanding of the cause–effect relationship.

However, it may be prone to significant interference, because the

quantity of beverage in a cup constitutes the final physical variable of

the entire process, from grinding to extraction. As such, it may be

influenced not only by the adjustment of the grinder, but also by other

variables such as the ground coffee’s waiting time before undergoing

extraction, the compacting force exerted on the cake, and other

parameters in adjusting the espresso machine (pressure, temperature,

etc.; see also Chapter 7).

In order to relieve the evaluation of grinding effectiveness from any

percolation tests, it is necessary to find a method that can be applied

directly on the coffee powder. One such example is the simple, classical

tactile examination performed by master grinding-operators for decades:

it consists in assessing the coarseness of ground coffee by rubbing a

pinch of powder between finger tips. However, such a subjective,

individual analysis obviously pertains more to the domain of art than to

that of science. Consequently, a major concern of the coffee industry

is to devise an objective method for measuring ground coffee particle


Grinding 221

5.3.1 Sifting methods

One of the earliest methods of analysis adopted (Allen, 1968) consists in

shaking the ground product through a stack of graded-mesh sieves, with

incrementally smaller-sized openings. The net weight retained on each

sieve is recorded on linear, semilogarithmic or logarithmic charts

(Prasher, 1987), showing a curve that characterizes size distribution.

Various standardization proposals have been put forward for the shape of

the sieve as well as for automated or manual shaking procedures (DIN,

1997). Among the many ingenious proposals, some consider sieving aided

by vibrating chains, counter air-current as in DIN Norm 10765–1975, or

solvent counterflow.

The sieving system has not met with much favour in the specific field of

coffee grinding, specially when espresso is concerned, due to two basic


n All methods have a poor repeatability, due to a substantial amount of

oil coming up to the powder surface after grinding. The oil acts as an

adhesive on the particles, forming random yet stable aggregates

unable to pass through meshes. Moreover, particles passing through

one sieve may commingle with other aggregates on the following

sieves, thus distorting the objective results of the analysis. This

drawback is more evident in espresso than in other brewing

techniques, because of its darker roasting and finer grinding needs.

n Sieving methods are long and laborious, and yield a characterization

of the ground material which is often inadequate, since it is drawn

from a too-limited series of sieves.

5.3.2 Imaging and sensing zone methods

Another intuitive approach to measuring ground coffee fineness consists

in placing a powder sample on a microscope slide marked with a grid, and

directly determining the size of a significant number of particles. This

method is quite labour-intensive, even if eased nowadays by photographic

microscopy techniques assisted by an automated specimen holder, and by

computerized image analysis that provides histograms by particle size

groups. As with all other micro-determination techniques, this measuring

method also suffers from technical problems in sampling, and from the

issue of representativeness of samples having the size of a few cubic

millimetres, instead of tens or hundreds of grams used in sieving. A further

difficulty lies in spreading out the particles in a thin layer, so as to allow

222 Espresso Coffee

their counting and measuring. Other methods (Svarovski, 1990) propose

dispersing the powder in a solvent in order to determine physical

characteristics such as electric conductivity (sensing zone counter),

sedimentation rate, and density.

5.3.3 Laser diffractometry

Size distribution analyses based on optical principles appear much more

rapid and promising. The most widely used method is laser diffractometry,

suitable for ground coffee particles sized from 5 mm to 2000 mm. This

technique uses the exact parallelism of coherent light emitted by laser

sources in the visible or near-infrared. By means of very high-quality lens

systems, it is possible to concentrate a beam of such light on an extremely

small punctiform target (<0.1 mm) surrounded by concentric annular

sensors. Particles passing through the beam produce a scattering, which

disperses a little amount of light into a conical geometry. The smaller the

particle, the greater the angle of the diffraction cone. The quantity of

light striking each of the sensors displays a trend characterizing the

particle sizes. Computer processing of this information produces a chart

that associates size ranges with the concentration of particles contained in

that range.

A difficulty set by this method lies in sample presentation: if performed

dry, by pneumatic dispersion devices (Illy, 1994), it allows a short analysis

time (a few minutes), which renders this method much more interesting

than the more accurate but time-consuming wet presentation. A

pneumatic dispersor is a tube to which a considerable quantity of coffee

powder (up to 100 g) is continuously fed by vibrating conveyors.

Compressed air is injected in various sites at different angles, exerting

shear forces on particle clusters, breaking them up. The amount of feed

and of air used must be calibrated to obtain the prescribed beam

obscuration. Wet presenters require a much smaller quantity of powder

(less than one gram), to be suspended in a convenient liquid (hexane, hot

water) sometimes with the help of ultrasonication and surfactants. The

suspension is then circulated through a shallow transparent cell placed

across the beam’s path. Typical analysis times range from a couple of

minutes for dry presentation up to 20 minutes for wet presentation.

A picture of the size distribution of a powder sample is given by a

histogram or, when many measure points are used, by a chart showing a

grading curve (Figure 5.4).

The curve is plotted on a diagram in which the axis of the abscissa,

usually on a logarithmic scale, carries the ‘equivalent diameter’ variable:

Grinding 223

this variable can be defined differently according to the technique

adopted, but it is somehow linked with particle size. The percentages of

particles belonging to each size group are plotted on the ordinate axis in

differential (frequency chart) or integral form (cumulative curve). The

percentage may refer to the number of particles, to their volume or mass

or weight, or even to their surface. It is possible to pass from one chart

form to the other by calculation, once the relationship between particle

shape and density factors has been determined. The comparison of ground

coffee batches obtained from different grinding equipments shows

differences that characterize each ground coffee by its own grading

curve, just as fingerprints identify a person.

Several companies sell turnkey computerized analysers, designed for

general powders. They differ by bench length (that determines the

maximum measurable particle diameter) and presentation accessories. An

effort to adapt sample presentation devices to ground coffee is always

needed, due to its sticky character and fouling tendency. Some brands

worth mentioning are Coulter, Malvern, Fritsch and Sympatech.


The unit operation ‘comminution’ involves two types of input variables:

those linked with raw material and those linked with grinding machinery.

224 Espresso Coffee

Figure 5.4 Particle size frequency distributions: (a) percentage at 0.205mm (max.

count); (b) percentage at 1.097mm (max. surface); (c) percentage at 517.2mm

(max. volume)

5.4.1 Variability of the coffee blend

The first type of variable includes, chiefly, the variability in composition

of the coffee to be ground. As explained in Chapter 4 on roasting, espresso

needs a balance of flavours and aromas, obtained from a blend composed

of different origins. Thus the grinder is confronted with beans that not

only come from different botanical varieties and producing countries, but

are also dried and processed by diverse methods. This often results in a

lack of homogeneity in hardness: in addition, hardness is constantly

susceptible to daily changes owing to dissimilarities in the batches, since

it must be remembered that coffee is an agricultural commodity subject to

the laws of climate and of natural evolution.

5.4.2 Roasting degree

Pyrolysis reactions forming volatile gases and aromas produce a considerable

expansion of the bean, depending on final roasting temperature and

rate of roasting (see 4.4). This expansion occurs at the expense of the

elasticity of the cell walls, which become brittle and lose tenacity,

changing from a nearly plastic behaviour in the green beans to an elastic

behaviour, and eventually to a brittle failure in the dark roasted beans. If

coffee is air-cooled after roasting, the high quantity of gas trapped in the

cells may affect the grinding process stability. A way to induce degassing

naturally, performed in industrial practice, is to allow the roasted beans to

rest in a silo for a suitable number of hours.

5.4.3 Moisture of the roasted beans

As seen in 4.1, water quenching may result in a substantial residual

moisture, which increases the bean’s tenacity and hence the grinder’s

consumption of energy with a resulting over-heating of the coffee.

Furthermore, contrary to air-cooling, quenching provokes cell-wall

cracking from the sudden volume contractions resulting from the brutal

temperature drop. This causes a larger degassing of the beans before

grinding and modifies the resistance to comminution of the cell structure.

5.4.4 Cutters’ distance

Machinery-related variables capable of affecting grinding results may

be grouped in control variables and in disturbances. The major control

Grinding 225

variable, closely related to the ‘notch’ of the coffee-bar grinder, is the

mechanical adjustment that allows regulation of the shear gap (or gaps) to

obtain a finer powder. In flat or conical cutters this adjustment consists in

drawing together coaxially the pair of cutting elements. The operation

requires a good degree of precision and low working tolerance, and must

be free from hysteresis; it is, however, easy to perform. In roller cutters,

the two cutting cylinders are drawn together by translating the axis of the

mobile cutter parallel to that of the fixed cutter: this delicate operation

requires complex and costly mechanisms. In both cases, an important role

is played by the mechanism controlling the separation between the

cutters, which places the pair of rotating cutters almost into contact.

5.4.5 Other parameters

Equally important is the adjustment of the connected units, primarily the

feeding and pre-breaking mechanism; such operation has a considerable

effect on hourly throughput and on ground product quality, by

contributing to the avoidance of flooding and overheating. Less

important, and therefore less commonly performed, is the regulation of

the cutters’ rotation speed; such operation could either be performed by

complex and delicate kinematic mechanisms or through more modern but

equally expensive means, such as three-phase power supply frequency


5.4.6 Disturbances

Disturbances are those hard-to-control variables causing equipment to

perform irregularly, which industrial practice aims to minimize.

A major cause of loss of control in a grinder undoubtedly comes from

energy dissipation during the application of mechanical power to break

down the product. As in all mechanical operations, there is an inevitable

element of friction, which adds to the heat generated by fracturing and

causes differential expansion of metal parts. This is especially critical

where tenths of a millimetre play an important role in controlling particle

size distribution. Even though various systems for minimizing these

harmful effects (thermostatation jacket with circulating liquid, injection

of boiling-point liquid nitrogen between the cutters) have been proposed,

the golden rule of grinding is still to wait for the machine to reach

thermal stability before considering it ‘operational’, that is, before

considering the ground product optimal.

226 Espresso Coffee

As in all machinery, grinders are prone to wear and tear, and need

proper maintenance. The parts that must be kept under control most

carefully are of course the cutting tools, whose blades need periodical

sharpening to recover their original performance. The result of poorly

sharpened tools is an increase in adverse thermal effects, due to the higher

compression-to-shear ratio applied to the coffee particles by blunt cutting

edges. Worn cutters impart to the beverage a ‘burnt’, bitter taste by

overheating the coffee.

Also worth mentioning are the disturbances caused by the electric

power supply, or by foreign matter. Power supply disorders may affect

rotation speed for a few seconds and destabilize a possibly already unstable

equilibrium. Foreign matter that may be present in spite of the utmost

attention and in spite of all the equipment installed for its elimination

(pebble removers, magnets, etc.), may suddenly block the cutters and

even damage them permanently.



The main influence of the grinding process on its feed, the roasted beans,

is of course fragmentation. Its efficacy, often evaluated by the empirical

method of rubbing a pinch of powder between the fingers to feel its

coarseness, should be quantified by analysing the product by means of an

objective measurement method, as seen in 5.3.

Grinding operations, besides the described effect on the physical

structure, also impart to coffee some chemical modifications due to the

temperature reached during grinding. It is quite common reading

temperatures up to 80


C on the ground coffee thermocouple, which

probably correspond to real temperatures of over 100


C. These levels are

well below those reached in the roasting process, and the coffee beans are

exposed to them for a shorter period than during roasting (only a few

seconds, instead of minutes). Nevertheless, such temperatures speed up

Maillard and oxidation reactions, with the detrimental effects on coffee

taste that can be judged by comparative organoleptic tests.

Temperature acts also on a macroscopic phenomenon that is important

for espresso beverage: the release of volatile aroma. When broken down

by grinding, coffee cells release their content of pyrolytic gas formed at

roasting. This gas is mostly composed of CO2 and CO, accompanied by

small amounts of hundreds of volatile chemical species (see 4.4).

Although present only in low concentrations, volatiles are essential in

Grinding 227

bringing forth the typical coffee aroma. Most industrial grinding

equipment is designed to handle the release of gas by conveying it

externally through a filter. Techniques for aroma recovery have been

proposed, such as cryogenic condensation, absorption in oil, or adsorption

on porous material.

However, the unfractured cells still present in the larger-size particles

(over 50 mm) keep their content of high pressure roasting gas, which will

contribute to the formation of ‘crema’, the typical espresso foam (see

8.1.1). Inevitably, the gas will slowly escape during storage through microcracks,

or through the natural porosity of the cell walls. This evolution is

accelerated by high grinding temperature, which increases the pressure

within the cell. While the pros and cons of degassing will be discussed in, it is certain that the worst method to degas ground coffee is by

overheating it.

One last macroscopic effect induced by grinding consists in the

appearance of coffee oil on the surface of the powder, resulting from the

release of the lipidic content of so far undamaged cells. As already

explained in 3.11.9 and, the major part of the coffee lipid fraction

is composed of triglycerides of palmitic and linoleic acids, and of stearic

and oleic acids in a lower percentage. This blend is very viscous at room

temperature, but starts becoming fluid around 40


C (Table 5.1). At

higher temperatures, fluidity is such that oil can easily flow through micro

cracks and cover the outer surface of the particles with a layer that

regains its sticky, semi-solid state at room temperature. This phenomenon

affects the cohesive behaviour of ground coffee particles when compacting

(see, hence influencing the hydraulic resistance of the cake.

Overheating on grinding may thus be responsible for erratic percolation

behaviour when brewing espresso. It must be stressed therefore that

temperature control in grinding equipment is one of the main

requirements to obtain a powder fit for preparing consistently perfect

espresso cups.

228 Espresso Coffee

Table 5.1 Coffee

oil viscosity

T (


C) Viscosity (mPa s)

0 Solid

20 70.6

50 22.4

92 7.8


Allen T. (1968) Particle Size Measurement. London: Chapman and Hall,

pp. 165, 606.

DIN (1997) DIN-Taschenbuch 133: Partikelmesstechnik Deutsches Institut

fu¨r Normung, Berlin.

Fayed M.E. and Otten L. (1984) Handbook of Powder Science and Technology.

New York: Van Nostrand, pp. 562–606.

Illy E. (1994) System for controlling the grinding of roast coffee. US patent


Petracco M. and Marega G. (1991) Coffee grinding dynamics: a new approach

by computer simulation. Proc. 14th ASIC Coll., pp. 319–330.

Pittia P. and Lerici C.R. (2001) Textural changes of coffee beans as affected

by roasting conditions. Lebensmitt. Wiss. Technol. 34, 168–175.

Prasher C.L. (1987) Crushing and Grinding Process Handbook. Chichester: John

Wiley & Sons, pp. 87–102.

Svarovski L. (1990) Characterization of powders. In M.J. Rhodes (eds),

Principles of Powder Technology. Chichester: John Wiley & Sons,

pp. 51–61.

Grinding 229


Storage and


M.C. Nicoli and O. Savonitti



M.C. Nicoli

Roasted coffee is a shelf-stable product. In fact, thanks to the high

temperatures attained in the roasting process and to its low water activity

(aw), no enzymatic and microbial spoilage occurs. However, during storage,

coffee undergoes important chemical and physical changes, which greatly

affect quality and acceptability of the brew. These events are responsible

for staling of roasted coffee but they are also believed to be involved in the

so-called ageing process. The border existing between ageing and staling is

unclear, and the reactions which prevail in ageing with respect to staling

are still unknown. Ageing is a short length of time, just after roasting, in

which coffee is allowed to rest under proper technological conditions for

improving its sensorial properties. Although many of the physical and

chemical changes occurring in roasted coffee during storage are considered

unavoidable, the rate at which they occur mostly depends on some

environmental and processing variables such as oxygen and moisture

availability, exposed surface as well as packaging conditions. The adoption

of proper grinding and packaging conditions can greatly slow down staling

reactions. However, they can be accelerated again, proceeding at a

considerable rate, after the consumer opens the packaging. In fact, home

storage conditions, mainly in terms of oxygen availability, relative

humidity and temperature, are the key factors in determining the socalled

secondary shelf life of the product (Full et al., 2001).

Table 6.1 summarizes the most important physical and chemical events

that take place in roasted coffee during storage.

Solubilization and adsorption represent the main mechanisms through

which aroma compounds are retained into the bean structure. In fact,

volatiles, as well as carbon dioxide, can exist entrapped in the pores,

dissolved in coffee oils and moisture and probably adsorbed into active

sites (Labuza et al., 2001). This represents an important requisite for

obtaining a high quality coffee cup. However, volatile compounds are

easily released in the vapour phase due to diffusion mechanisms. The

latter, together with oxidation reactions, are considered the main causes

of coffee staling. The evolution of carbon dioxide, which is the most

important non-aromatic volatile found in fresh roasted coffee, represents

an additional problem during storage, mainly from a technological point

of view. In fact, in the case of use of flexible pouches, it causes a rapid loss

in the package integrity due to swelling and bursting. The amount of CO2

formed depends on the degree of roasting and can be up to 10 ml/g (SPT),

corresponding to 1–2% of roasted coffee weight (Sivetz and Desrosier,

1979). For this reason, different procedures such as degassing or

pressurization, together with the adoption of suitable packaging solutions,

have been developed in order to allow carbon dioxide to be released,

minimizing aroma losses and oxidations (see 6.2).

Since roasted coffee is a dehydrated product, the major factors

controlling its stability during storage are oxygen concentration (O2),

water activity (aw) – particularly if roasting had been quenched by adding

water – and temperature. Technological operations such as grinding,

degassing and packaging may accelerate the above-mentioned changes

due to the increase in the exposed surface, as well as in oxygen and

moisture availability. In addition, since the pressure in the bean pores

should be greater than atmospheric pressure, the external pressure applied

to the coffee beans represents an additional variable to be considered for

enhancing coffee stability during storage. In fact, as the external pressure

becomes higher than the partial pressure of volatiles present in the beans,

the degassing rate is reduced allowing a larger quantity of volatiles to be

dissolved in the lipid fraction or adsorbed on the active sites (Clarke,

1987a; Labuza et al., 2001).

Storage and packaging 231

Table 6.1 Physical and

chemical changes occurring

in roasted coffee during


Volatile solubilization CO2 release

Volatile adsorption Oxidations

Volatile release Oil migration

6.1.1 The role of Maillard reaction products

on roasted coffee stability during storage

(see also 4.5)

Coffee staling takes its origin from roasting. Maillard reaction and

pyrolysis, which are the main chemical events occurring in coffee during

roasting, transform the beans into a very unstable and reactive system. In

broad terms, the most macroscopic changes occurring in coffee during

roasting and storage can be attributed to the formation of Maillard

reaction products (MRPs) and carbon dioxide. The latter is a reaction

product from both Maillard reaction and pyrolysis (Hodge, 1953;

Yaylayan and Huyghues-Despointes, 1994). The formation of volatiles

and CO2 during roasting causes the expansion of the beans due to the

internal build-up of gases which, along with the high temperatures, allows

internal pores and pockets to be formed. However, the increased CO2

formation in the advanced stages of the roasting process corresponds to a

progressive reduction of the ability of the bean to entrap and retain CO2

and other volatiles, as the bean loses its elasticity and surface cracks are

formed under the push of the increased internal pressure (Massini et al.,

1990). A comparison between the amount of CO2 formed during roasting

and retained into the bean during storage is shown in Figure 6.1.

232 Espresso Coffee

Figure 6.1 Comparison between amount of CO2 produced during roasting and

retained into the coffee beans (data elaborated from Barbera, 1967; Massini et al.,

1990). Values of CO2 retained are referred to the percentage of CO2 present in the

headspace in equilibrium with the sample after 5 minutes’ storage after roasting

It can be observed that, as the bean temperature rises during roasting,

increased amounts of CO2 are formed. However, this corresponds to a

progressive loss of the CO2 retention capacity of the bean. Among the

hundreds of compounds generally considered as MRPs, volatiles and

melanoidins play a key role in determining the aroma and body of the

coffee brew. Also, the development of the Maillard reaction during

roasting is responsible for other important changes affecting coffee

stability during storage. The most important effects exerted by MRPs in

roasted coffee are summarized in Table 6.2.

Many of the low molecular weight MRPs are responsible for the

formation of the typical coffee flavour. It is well known that the aroma

profile of roasted coffee is given by several hundreds of compounds

covering many different chemical classes. The development of the typical

stale flavour of old coffee is the sum of two different actions: the partition

of volatiles in the vapour phase and the development of off-flavours as the

result of their oxidation. It has been observed that methanethiol, Strecker

aldehydes and  -dicarbonyls are the most important low-boiling

compounds responsible for the freshness of coffee aroma (Steinhardt

and Holscher, 1991). These molecules, in addition to their high volatility,

are particularly prone to oxidation. However, it must be pointed out that

the aroma perception could also be affected by the ability of melanoidins

to bind specific volatile molecules, as recently observed for some thiol

compounds by Hofmann et al. (2001).

Most of the effects of MRPs on the shelf life of roasted coffee trace their

origin to the ability of these compounds to act as both pro-oxidants and

antioxidants. These contradictory functions may be explained by the

complex and different chemical structures of MRPs. In broad terms, while

melanoidins are generally recognized to be strong antioxidants, uncoloured

compounds, such as those formed during the early and intermediate

stages of Maillard reaction, may exert pro-oxidant activity, being

Storage and packaging 233

Table 6.2 Significance of MRPs in coffee

during storage

Flavour Shelf life

Volatile formation Oxidation reactions:

Degradation of aroma compounds

Lipid oxidation slow-down

Microbial growth inhibition

Volatile binding

in most cases radicals (Namiki, 1990; Manzocco et al., 2001, 2002).

During storage, radical reactions may occur leading to the formation of

non-radical forms as well as to the generation of radical species

chemically different from those initially present (Goodman et al., 2001).

Among melanoidins, recent findings suggest that the fraction of low

molecular weight melanoidins (MW<1000–3000 Da) shows higher antioxidant

efficiency than that of high molecular weight (MW up to

150 000) (Hofmann, 2000). These compounds seem to act both as

primary and as secondary antioxidants since they have been found to

behave as chain-breakers, reducing compounds and metal chelators

(Homma et al., 1986; Nicoli et al., 1997). The antioxidant capacity of

roasted coffee can greatly vary depending on the roasting degree. In Table

6.3 data of chain-breaking activity and redox potential with reference to

light, medium and dark coffee samples are reported.

While the assessment of the chain-breaking activity allows estimation

of the quenching rate of coffee compounds towards reference radicals, the

redox potential gives indications on the effective oxidation/reduction

efficiency of all the antioxidants present, including the ‘slow’ ones, which

cannot be detected by the traditional chain-breaking kinetic assays. ‘Slow’

antioxidants are expected to play an important role in determining the

shelf life of roasted coffee during storage (Nicoli et al., 2004). While slight

changes in the chain-breaking activity can be observed from light to dark

roasted coffees, indicating the existence of a balance between the antiradical

efficiency of naturally occurring antioxidants, which are degraded

during roasting, and the heat-induced ones, the redox potential

significantly decreases. This indicates that MRPs show higher reducing

properties than natural antioxidants. In broad terms, this means that dark

234 Espresso Coffee

Table 6.3 Chain-breaking activity and redox potential

values of coffee samples having different roasting degree


Weight loss


Chain-breaking activity

( Abs 3 min 1mgss


Redox potential


Raw 0 0.320 0.13* 109

Light 14.5 0.273 0.03* 45

Medium 16.2 0.283 0.13*  2

Dark 18.9 0.390 0.04  35

*Data not significantly different.

The chain-breaking activity assay was performed using the DPPH test (Nicoli

et al., 2001).

roasted coffee is expected to have a greater antioxidant reservoir

potentially available during storage.

The presence of such reactive species in roasted coffee can explain the

huge variety of redox and radical reactions occurring in coffee during

storage and even, at an increased rate, after beverage extraction (Nicoli

et al., 2001; Anese and Nicoli, 2003). The oxidation of these reactive

compounds has a double relation with roasted coffee shelf life: they are

the main cause of aroma compounds degradation, but they also represent

an efficient barrier against the oxidation of the lipid fraction. In fact,

despite the very low aw value, lipids in coffee show an extraordinary

stability against oxidation thanks to the presence of lipid-soluble

MR antioxidants (Whitfield, 1992). Figure 6.2 shows the influence of

melanoidins on the stability of coffee lipids. It is interesting to note that

samples previously freed from melanoidins showed lower oxidation

stability as compared to that measured for lipids containing melanoidins.

The antioxidant efficiency of melanoidins has been found to be strictly

related to the intensity of the heat treatment used to allow their

formation. The higher the intensity of heating, the higher their intrinsic

antioxidant capacity (Anese et al., 2000). For dark roasted coffee, the

stability of the lipid fraction was observed to range from 4 to 10 months at



C storage depending on oxygen availability and packaging conditions

(Nicoli et al., 1993).

Storage and packaging 235

Figure 6.2 Influence of lipid-soluble melanoidins on oxidation stability of coffee

lipids (from Anese et al., 2000). The uncoloured samples were obtained by

removing melanoidins with active carbon. Data are expressed as induction time

(hours) prior to oxidation detected by Rancimat test

The ability of MRPs to act as antioxidants can in part explain their

inhibiting action against the growth of some food-poisoning microorganisms.

However, the extent of MRPs antimicrobial activity appears to be

linked to their concentration and to the bacteria tested, indicating that

different mechanisms other than radical scavenging activity may be

involved, since diluted coffee shows a weaker antibacterial activity than

concentrated coffees. In addition, the extent of the antimicrobial activity

strictly depends on the species tested (Einarson, 1987; Sheikh-Zeinoddin

et al., 2000; Daglia et al., 2000).

An increasing interest towards coffee MRPs is justified by the fact that,

by virtue of their antioxidant activity, they are expected to exert a range

of positive health effects on the human organism (O’Brian and Morissey,

1989; Faist and Erbesdobler, 1999). Thus, the preservation of the intrinsic

antioxidant properties of roasted coffee during storage could be of a

certain importance to obtain coffee products with high functional

properties. Unfortunately, up to now no data are available about the

influence of different environmental and processing conditions on the

evolution of the overall antioxidant properties of roasted coffee during


6.1.2 Staling kinetics Degassing

Carbon dioxide formed during roasting is trapped in the cellular structure

of the bean and is only released over a period of weeks following roasting,

resulting in a 1.5–1.7% weight loss. The amount of gas released can be

estimated at 6–10 litres per kilogram of beans depending on roasting

degree, the higher figures being valid for dark roasted blends (Clarke,

1987b), recently confirmed by the results of Shimoni and Labuza (2000).

Degassing rate is inversely related to time from roasting. The massive

degassing that takes place in the early hours after roasting slows down

gradually, and it may take months for all the CO2 to be released from the

beans. The process is slow because much of the CO2 is bound to the bean

structure. From a physico-chemical point of view, carbon dioxide in the

coffee bean can be solubilized partly in the lipid phase, partly in the water

(moisture). Some will be adsorbed on the apolar sites, and, finally, a part

will be blocked within the pores, which have collapsed because of the

phase changes in the structure of the bean due to the roasting process.

The amount of CO2 adsorbed in the lipid and aqueous fractions can be

roughly estimated at 15   10 5 mg of CO2 per gram of fresh roasted

coffee in the lipid phase, and 2.13   10 5 mg of CO2 per g of fresh roasted

236 Espresso Coffee

and ground coffee in water, considering coffee in air at normal CO2

pressure (Shimoni and Labuza, 2000; Labuza et al., 2001).

The driving force at the basis of carbon dioxide and volatile release

from roasted coffee is given by a diffusion flow due both to concentration

and pressure gradients. The release begins at a considerable rate just after

the end of roasting; then it gradually slows down as the CO2 and volatile

concentration in the headspace increases, reducing the mass transfer

driving force.

Figure 6.3 shows an example of CO2 and volatiles released from roasted

coffee beans and roasted ground coffee.

Recent studies carried out by Labuza et al. (2001) suggest that CO2

release from ground coffee is controlled by two different mechanisms: at

the beginning the diffusion is regulated by pressure gradients between the

outside and the inside of the coffee particles, then by molecular diffusion.

The time at which the changeover in mechanism occurs in STP is about

120 min. The few data available about volatile partition kinetics from

roasted coffee packed in air indicate that the rates of CO2 and volatiles

release are of the same magnitude. The volatiles present in the headspace

at equilibrium with the product after three days’ storage were 50%, 35%

and 18% at 4, 24 and 40


C respectively. As regards CO2 evolution, the

residual gas percentage in the headspace after the same storage time and

temperature was 63%, 37% and 18% (Nicoli et al., 1993).

As mentioned above, the rate of volatile compounds and gas release

from coffee is strongly affected by temperature. Additional factors are

pressure, particle size distribution, and during the degassing process,

packing, bed depth and entrainment gas composition.

Storage and packaging 237

Figure 6.3 Changes in headspace volatile and CO2 concentration, expressed as

percentage, in (left) roasted coffee beans and (right) roasted ground coffee stored

at 25


C in air (Nicoli et al., 1993) Effect of temperature

The temperature dependence of CO2 and volatile release from roasted

coffee can be well described by the Arrhenius equation. An example of an

Arrhenius plot, referring to volatile release from roasted coffee beans, is

shown in Figure 6.4. The corresponding activation energy was 28.1 kJ/

mol, which gives a temperature sensitivity, expressed in terms of Q10, of

around 1.5. It means that, in the experimental conditions adopted, for any



C increase, the rate of volatile release increased 1.5-fold. This result

agrees with data on the temperature dependence of diffusion coefficients

for CO2, recently reported by Labuza et al. (2001). Moreover, the Q10 of

CO2 diffusion coefficients calculated for ground coffee were shown to be

approximately twice those of the beans.

The importance of the temperature variations due to day–night

fluctuations and to transport is difficult to assess. It may be assumed

that, because of temperature changes, competing degradation reactions

occur at different times, so that the resulting deterioration is likely to

occur faster than at constant temperature. Besides temperature, additional

factors affecting staling reaction rates of coffee during storage are aw and

O2 availability. Effect of moisture

Water in roasted coffee may have different origins: despite the strong

dehydration occurring during the first stage of the roasting process, new

238 Espresso Coffee

Figure 6.4 Arrhenius plot of volatile release from roasted coffee beans. Data are

expressed as ln volatile release rate vs. temperature–1 (K). (Elaborated from Nicoli

et al., 1993)

water molecules are formed as a consequence of the Maillard reaction.

Additional water may come by the use of water quenching instead of aircooling.

Furthermore, the eventual sorption of ambient moisture from

roasted coffee during storage represents an additional cause of moisture

and hence of aw increase. Moisture sorption isotherms at 22


C of roasted

coffee beans and ground coffee are shown in Figure 6.5. Due to the higher

exposed surface, which affects concentration and availability of polar

active sites, ground coffee shows a considerable higher capacity to bind

water molecules compared to beans.

In addition, depending on aw, the bean structure can be brittle or

elastic as shown in Figure 6.6.

It is well known that at normal storage temperatures the bean structure

is brittle and easy to grind, as shown by the low fracture force measured for

air-cooled roasted coffee samples with aw values ranging at 0.1–0.2. This is

due to the fact that roasted coffee beans are greatly below the glass

transition temperature, which was recently observed to range from 130

to 170


C (Karmas and Karel, 1994; Eggers and Pietsch, 2001). However,

increasing aw, the fracture force was shown to double up to aw 0.85. This is

due to the plasticizing effect of water responsible for a decrease in the glass

transition temperature (Tg), making the coffee structure rubbery and

more elastic. For aw values higher than 0.85, solubilization of coffee

compounds and a sharp decrease in the fracture force are detected.

Storage and packaging 239

Figure 6.5 Moisture sorption isotherm at 22


C of roasted coffee beans and

ground coffee. (Data from Labuza et al., 2001; Pittia, 2002) Oxygen uptake

As mentioned above, oxidation reactions represent an additional cause of

quality loss of roasted coffee during storage. In general terms the rate of all

the oxidation reactions occurring in coffee, including those of aroma

compounds, increases with increasing oxygen pressure, temperature and

aw. A study of the influence of these three variables on the overall

oxidation rate of ground roasted coffee was recently reported by Cardelli

and Labuza (2001). The oxygen uptake and sensorial evaluations were

used as indicators of product acceptability. By matching the map of rate of

food deterioration reactions as a function of water activity with the water

sorption isotherm of roasted and ground coffee at room temperature it

could be possible to understand what type of chemical reactions are

involved in coffee deterioration. In coffee, moisture content ranges from

2% to 5%, which corresponds to 0.2–0.4 water activity. Using these data,

it is evident that the oxidation of the coffee lipids is the prevalent


The influence of aw and temperature on the shelf life of roasted ground

coffee, calculated on the basis of the oxygen uptake index, is illustrated in

Figure 6.7. On the basis of these results it can be stated that temperature

has a smaller effect than moisture on the shelf life of the product.

The increase in the rate of degradation can be quantified, as a function

of oxygen, by introducing a QO2 index, which represents the relative

variation of the rate of degradation caused by an increase in oxygen

content of 1%. According to Cardelli and Labuza (2001), QO2 is 10.5,

240 Espresso Coffee

Figure 6.6 Changes in fracture force of dark roasted coffee beans as a function of

water activity (Pittia, 2002)

when oxygen concentration is low (0.1–1.1%). It is 1.1 when oxygen

increases up to 5%. Further increase of oxygen, up to 21%, does not affect

the index, which remains 1.1. Summarizing, it can be stated that the rate

of oxygen uptake increases 10-fold, from 0.1 to 1.1%; above 1.1% of

oxygen concentration, for each 1% increase of oxygen there is a

consequent increase for the rate of degradation of 10%.

6.1.3 Other physico-chemical changes

O. Savonitti Volatile compounds

The aroma complex formed during roasting consists of several hundred

compounds covering many different chemical classes. Some of them are

stable and undergo no change during storage other than loss due to

evaporation. Apparently, such compounds do not contribute much to the

quality of the perceived aroma. Aroma-impact components are mainly

unstable substances particularly prone to oxidation; some, aldehydes and

ketones, are present in intermediate states of oxidation, whilst many more,

alcohols, phenols, thiols, etc., are in a reduced form. There is then a significant

risk of aroma changes occurring as a result of oxidation, particularly

of compounds such as the thiols, which have very low threshold values.

The mechanisms responsible for the deterioration of aroma constituents

have not yet been fully elucidated; immediately after roasting it may

be simply due to evaporation and physical loss of the most volatile

compounds; conversion to other volatile or non-volatile compounds,

retained in the coffee matrix, also takes place; intermolecular reactions

are also likely to occur – for instance,  -dicarbonyls and aldehydes can

Storage and packaging 241

Figure 6.7 Water activity and temperature dependence of the shelf life of ground

and roasted coffee (Labuza et al., 2001)

bind to compounds containing free amino groups. These reactions are

apparently influenced by the presence of atmospheric oxygen, suggesting

an oxidative route to degradation. The change in aroma profile cannot be

simply explained by the loss of the compounds affected, as the reaction

products themselves may make a sensory contribution. In the case of

thiols, for instance, the corresponding condensation products, the

disulphides, have a major sensory impact, different from that of the

fresh product. Afterwards, oxidation reactions become predominant, as

indicated by the correlation between deterioration and oxygen uptake;

and oxidation of the aroma complex will lead to a further dramatic shift

in aroma quality.

Oxidation, responsible for the typical stale flavour of old coffee, is the

sum of two actions: the loss of pleasant aroma components accompanied

by the formation of off-flavours. A study has concentrated on identifying

the compounds responsible for the freshness of coffee aroma, defined as

the fine and pleasant smell arising from freshly roasted beans (Steinhart

and Holscher, 1991). The conclusion was that low-boiling compounds –

such as methanethiol, Strecker aldehydes and  -dicarbonyls – were the

most important; in particular, loss of methanethiol was considered as the

most important indicator. All of these compounds are volatile and

extremely reactive species, therefore, easily oxidized. Figure 6.8 shows

that their disappearance, particularly noticeable among those with the

lowest boiling point, occurs mainly over the first three weeks of storage,

and is principally due to evaporation.

Successive research during the 1990s led to a set of 28 compounds in

roasted and ground coffee as the most odour-active components associated

to fresh coffee aroma (Grosch, 1999, 2001; Czerny and Schieberle, 2001).

242 Espresso Coffee

Figure 6.8 Loss of volatile compounds during storage

One of these, namely methanethiol, is frequently used in many

freshness indices for coffee samples stored in different conditions (Sanz

et al., 2001). Many of these compounds can be used, in a definite linear

combination, as indicators of the change of freshness for a roasted and

ground coffee exposed to air at different temperatures (Cappuccio et al.,


The loss of volatile substances from coffee is not only related to their

volatility, but also to the way they are trapped in the cells. In fact, two

additional mechanisms of aroma retention play an important role. The

cell wall consists of polar polysaccharides and melanoidins, which can

adsorb and retain a wide range of volatile substances, particularly polar

ones. Oil, on the contrary, preferentially retains lipophilic substances,

such as alkylpyrazines, etc. In order to evaporate, volatile particles

‘trapped’ in the cell structure of the beans must first migrate across the cell

walls. In ground coffee this is made easy by the small particle size, so that

aroma loss occurs at a higher rate in ground coffee than in coffee beans

and is inversely proportional to particle size. Hence, finely ground

espresso blends lose their aroma just a few hours after being exposed to air,

and start smelling stale due to oxidation, if not stored in packages.

However, even in this case coffee aroma can undergo modifications

through time, if stored in packages somehow permeable to oxygen. Coffee

staling is associated with a degradation of some particular odorants, like

odour-active aldehydes,  -dicarbonyls and thiols (2-furfurylthiol has the

main impact), rather than to the creation of off-odorants. This is due

again to the presence of oxygen, because of the non complete

impermeability of the package (Czerny and Schieberle, 2001). Non-volatile compounds

Production of carbon dioxide during roasting forms an effective barrier,

which excludes most of the atmospheric oxygen from the cellular

structure so delaying the onset of oxidation. As degassing goes on, lipid

oxidation can no longer be stopped by the antioxidant activity of the

Maillard reaction products formed during roasting (Homma, 2001;

Steinhart et al., 2001), and in a few months, at the end of degassing,

the beans are rancid.

Ground coffee is particularly sensitive to oxidation because grinding

removes the CO2 barrier almost completely, while at the same time the

oil–air contact surface dramatically increases, promoting oxygen uptake.

Lipids, an important fraction in roasted coffee, are easily oxidized

during storage (Huynh-ba et al., 2001). The major unsaturated fatty acid

in coffee is linoleic acid (C18:2), which contains two double bonds and is,

Storage and packaging 243

therefore, rather susceptible to oxidation. Figure 6.9 shows the reduction

in linoleic acid content of rancid beans with respect to fresh beans; while

the major saturated fatty acid, stearic acid (C16:0), remains practically

unchanged (Fourny et al., 1982).

An important contribution to oxidation off-flavours is the formation of

volatile aldehydes, such as trans-2-nonenal, the main breakdown product

of lipid oxidation, which has an odour threshold level of 0.08 mg/l in water

(Flament, 2002). Oil migration

Oil migration starts during roasting and goes on during degassing because

carbon dioxide tends to push oil outwards through the cell pores. The

increase in oil viscosity at lower temperatures slows down the process.

Oil migration to the surface of the beans, where the risk of oxidation is

maximal, is particularly important in fine ground dark espresso blends,

since dark roasting leads both to fast degassing and to increased porosity,

from disruption of cell walls. A further problem linked with oil exudation

is the increase in stickiness of the particles, which tend to aggregate into

lumps, making brewing irregular. The aggregation of particles on storage is

worsened by the absorption of moisture, as discussed in Effect of light

Light plays a catalytic role in many chemical reactions; in the case of

espresso (arabica) blends, particularly rich in unsaturated fatty acids, light

244 Espresso Coffee

Figure 6.9 Fatty acid content of (left) fresh and (right) rancid beans

catalyses the prime trigger of their auto-oxidation reaction, i.e. the

formation of H., R. (alkyl) and ROO. (peroxide) free radicals which then

cause the reaction to propagate.


O. Savonitti

All the potential causes of deterioration highlight the importance of

packaging. Indeed, it may be stated that, because of its instability, roasted

coffee should either be consumed straight away or packaged in a waterand

oxygen-tight container. Special packaging materials and techniques,

with careful process monitoring, are required not to jeopardize the final

quality by poor manufacturing practices.

6.2.1 Packaging materials

In order to meet total quality requirements (see Chapter 1) the package


n act as a barrier against water and moisture;

n act as a barrier against atmospheric oxygen;

n preserve coffee aroma and keep out foreign odours;

n be grease-proof;

n be light-proof;

n allow the carbon dioxide released during degassing to escape;

n be chemically inert;

n be hygienically safe and suitable for foodstuffs;

n be long-lasting;

n be sturdy and withstand pressure variations;

n be cheap;

n be practical;

n be environmentally friendly;

n be consumer-friendly.

Most of these requirements are aimed at preventing spoilage; some are

exclusively intended to add value to consumers. Packaging materials

should be grease-proof to prevent air oxidation of coffee oil leaking

through the package, which would then spoil the beans. Chemical inertia

of the material is also important, and metals or other substances, which

could act as oxidation catalysts, should be avoided. Table 6.4 lists the

materials meeting most of these packaging requirements.

Storage and packaging 245

Commonly used materials are the inexpensive flexible polymeraluminium

multi-ply, ensuring an efficient barrier and an optimal use of

space all along the life cycle, and tinplate, particularly useful for positive

pressure-resistant packages. The aluminium contained in multi-ply

materials constitutes the only completely effective barrier; the inner

layer is a film of waterproof weldable material; the centre layer is an

246 Espresso Coffee

Table 6.4 Materials commonly used for packaging roast coffee

Material Advantages Disadvantages

Tinplate Total barrier

Resistant to pressurse

Satisfactory ecobalance*




Its rigidity prevents an

optimal use of space

Aluminium Total barrier

Resistant to pressure



Its rigidity prevents an

optimal use of space

Poor ecobalance

Difficult to sort from other


Glass Total barrier

Good ecobalance


Resistant to pressure


Heavy material


Its rigidity prevents an optimal

use of space

Difficult to sort from other


Combined flexible

multi-ply polymer



Simple manufacturing technology

Large flexibility and optimal use of


Satisfactory ecobalance

Non-tight barrier

Resistant only to negative


Poor strength






Total barrier

Satisfactory ecobalance

Difficult to sort from

other wastes

Non resistant to pressure

Poor strength

Complex manufacturing


Flexible combined

multi-ply polymer


Simple manufacturing technology

Satisfactory CO2 permeability

No need to sort for other wastes

*Ecobalance is defined as the total pollution resulting from energy consumption, and that

resulting from the product, during manufacture, use and disposal.

airtight aluminium film and the outer layer consists of a material that

makes the structure more rigid and resistant to mechanical stress.

None of the materials listed in Table 6.4 ensures the level of

permeability to carbon dioxide required to keep a stable internal pressure

during degassing; other means must be applied.

Blow-moulded rigid plastic materials (HD-PE or PET) have lately been

introduced for roast and ground coffee packaging: one-way valves on easypeel-

off membranes or applied directly on the packaging allow degassing

and avoid deformations. This type of container is light, resistant to

breakage, can be easily piled up, and recycled. Production technology is

well known and the form can be chosen at will. However, the material

does not provide a perfect barrier to gases.

6.2.2 Packaging techniques

The techniques used for packaging roast coffee are listed in order of

increasing product protection:

n air packaging

n vacuum packaging

n inert gas packaging

n pressurization

n combined use of one of the previous techniques with active

packaging. Air packaging

Air packaging consists in simply filling and hermetically sealing the

package; coffee is protected against humidity, off-flavours and light, but

the presence of air inside the package means high oxygen levels and

consequently shortened shelf life. This technique can only be used with

degassed coffee to prevent swelling and possible explosion of the airtight

package. If the coffee, particularly coffee beans, is not degassed, swelling

can be prevented by fitting a one-way safety valve on the package

allowing carbon dioxide to escape without letting air in. As CO2 is

heavier than air, it tends to stratify at the bottom expelling most of the

oxygen gas; the flushing effect reduces residual oxygen and increases

product shelf life. Air packaging using a one-way safety valve is an

acceptable technique for air-cooled coffee beans since they contain large

quantities of gas; its main disadvantage is that most of the aroma escapes

together with CO2, so dulling cup taste.

Storage and packaging 247 Vacuum packaging

Vacuum packaging meets two different objectives: air extraction with

lowering of oxygen level, and use of flexible materials. The technique,

which can also be used with rigid materials such as tinplate, is commonly

applied to flexible materials to make the coffee ‘bricks’ sold in

supermarkets. The box shape is obtained by an intimate contact between

the material and the coffee, which must be completely degassed to

avoid a decrease in compactness of the package provoked by a release of

carbon dioxide making it floppy or swollen. This is why bricks

are generally used only for packaging water-cooled ground coffee.

Disadvantages of vacuum packaging in flexible containers are linked

with water-cooling, which shortens product shelf life and may decrease

yield. On the other hand, in rigid containers the internal vacuum

widens the difference between the partial pressure of volatile aroma and

ambient, so that more aroma volatilizes to saturate the headspace,

dulling the cup. Furthermore, even if rigid containers can be

strengthened by ribbing to prevent collapse, the vacuum level still

remains less than in flexible bags. Inert gas packaging

In inert gas packaging the air inside the container is replaced by inert gas

either through the compensated vacuum technique or by flushing the

inside of the package with inert gas. In the former case, first a vacuum is

created in the package, then enough inert gas to balance the internal and

external pressures is admitted. In the latter case, a drop of liquefied inert

gas is placed on the bottom of the package, which evaporates pushing the

air out. This process generally uses nitrogen or carbon dioxide which,

although not an inert gas, behaves as such in a moisture-free environment

and, moreover, is naturally present in roast coffee. The use of an inert gas

for coffee packaging increases shelf life three-fold (see Table 6.5) with

respect to vacuum packaging. Even with the same type of package – and

therefore with the same permeability to oxygen and water vapour – the

type of conditioning process will impact on the product’s shelf life, and

consequently on the cup (Alves et al., 2001).

The pressure in the package is in equilibrium with the atmosphere at

the moment of sealing. Just like in air packaging, to prevent internal

pressure rising due to degassing, coffee must be either packed after

degassing or the package fitted with a one-way valve.

From a legal standpoint the gas added to the package is considered a

processing aid rather than an additive, since it dissipates on opening.

248 Espresso Coffee Pressurization

Pressurization is the same as inert gas packaging except that the internal

pressure is higher than atmospheric pressure. If coffee is immediately

packaged after air-cooling, the pressure normally rises due to degassing.

The packaging technique is the same as in compensated vacuum

packaging, but in order to withstand pressure containers must be made

of rigid materials, normally tinplate or aluminium; they must also be

equipped with a safety valve opening when pressure increases by 0.5

atmospheres due to the large amount of gas released (see and

Figure 6.1).

Pressurization has an ‘ageing’ effect with quality improvement after

10–15 days. Indeed, an espresso cup prepared from an aged product has an

improved body and aroma. The ageing may be explained by a binding of

aroma constituents to the oil trapped within the cell structure. During

degassing, pressure rises in the container and the reduced pressure

gradient between the packaging environment and the bean structure

produces two combined effects underlying the ageing mechanism: first, a

decrease in degassing rate reduces oil migration so that more of it remains

in the beans (Clarke, 1987b, p. 204); second, the pressure in the package

becomes higher than the partial pressure of most volatile compounds

Storage and packaging 249

Table 6.5 Barrier properties of packaging

materials for roasted coffee

Material Permeability* Barrier

Tinplate, glass aluminium Complete

Three-ply with aluminium <0.5 Very high

Three-ply with metal 0.5–3.0 High

Two-ply with metal 3.0–30 Moderate

Two-ply 30–150 Low

Single film, treated paper >150 Very low

*Values expressed in:

cm3/m2 24 h bar at 23


C and 0% R.H. for gases: gas

volume (cm3 of gas) that passes through 1m2 of surface

in one day (24 hours), with one bar of pressure difference

between the two layers at 23


C and 0% of relative

humidity g/cm2 24 h bar at 38


C and 90% R.H. for water

values gas mass (grams of water vapour) that passes

through 1m2 of surface in one day (24 hours), with one

bar of pressure difference between the two layers at 38



and 90% of relative humidity.

present in the cells, allowing for a larger quantity to be dissolved in the

lipid phase or to bind to melanoidins.

Pressurization, by creating high pressure within the bean, also has the

effect of spreading the oil on the cell walls to form an effective barrier

against oxygen; this effect is clearly visible in Figure 6.10, which shows

the cross-sections of, on the left, a non-pressurized roast bean, and, on the

right, a pressurized one; caramelized material, containing lipids (dissolving

volatile aroma), is shown in orange-yellow, while cell walls are reddish.


Due to the difference between internal bean pressure (which at the end

of the degassing process reaches equilibrium with the pressure in the

container) and atmospheric pressure, when the package is opened, there is

still a residual amount of CO2 in the beans to be released. Moreover,

when oxygen enters the beans at the end of the degassing phase, oxidation

is reduced because the aroma constituents are dissolved in the compacted

lipid aggregates, so that, even once the package is open, a pressurized

blend retains its fragrance longer than a blend packaged by other

techniques. Active packaging

The need to stress the performance of existing passive packages and

related packaging techniques on one hand, and the need to guarantee an

adequate product freshness on the other has brought the concept of

‘active packaging’ into consideration for coffee as well. Rooney (1995)

defined active packaging as ‘packaging that performs a role other than an

inert barrier to the outside environment’. A more recent definition of

active packaging is ‘packaging that changes the condition of the packaged

food to extend shelf life or improve food safety or sensory properties, while

maintaining the quality of the packaged food’. This second definition has

250 Espresso Coffee

Figure 6.10 Cross-section of cells from (left) non-pressurized and (right)

pressurized beans

been chosen for the European FAIR project Ct.98-4170. (De Kruijf et al.,


Active packaging systems can be classified into active scavenging

systems (absorbers) and active-releasing systems (emitters). In the specific

case of active coffee packaging, it includes oxygen and carbon dioxide

scavengers. The oxygen scavenger has the precise task to eliminate

residual oxygen present within the headspace of the packaging in the case

of high barrier packages and conditioning in inert gas, or to eliminate the

oxygen present, in the case of air packaging. In both cases, the

elimination of oxygen contributes to extending the stability of the

product during the storing phase. Moreover, independently from the

conditioning used, in the case of permeable packaging, the oxygen

absorber has the task of capturing oxygen entering through the walls of

the package until it becomes saturated.

In general, the existing oxygen scavenger technologies are based on one

or more of the following principles: (1) iron powder oxidation, (2)

ascorbic acid oxidation, (3) photosensitive dye oxidation, (4) enzymatic

oxidation, (5) unsaturated fatty acids and (6) yeast immobilized on solid

material. They may be used as sachets inserted into the package, or be

bonded by a label to the wall of the package. They can be used as linings

in the closure of the package or can be dissolved, dispersed or immobilized

in the polymer of the package material. The most commonly applied

scavengers in coffee packaging are based on iron powder oxidation; their

advantage is that they are in a position to reduce the level of oxygen

below 0.01%, a value much lower than 0.3–3%, the residual values

typically reached when applying a modified atmosphere packaging


Carbon dioxide scavengers have the task of removing a proportion of

the carbon dioxide released from the coffee in order to prevent the

package from bursting when the coffee has not been previously degassed.

They can represent an alternative use to the one-way valves or to

materials permeable to carbon dioxide (which are in general also very

permeable to oxygen and water vapour) (Vermeiren et al., 1999).

A combined solution, such as dual action oxygen and carbon dioxide

scavenger systems in the form of sachets or labels, is common in canned

and foil-pouched coffees in Japan and the USA. This dual action system

typically contains iron powder as oxygen scavenger and calcium

hydroxide that scavenges carbon dioxide in the presence of high moisture

levels ( ‘A fresh

approach’, Brian Day Campden and Chorleywood Food Research

Association). Some oxygen scavengers are subject to competition

between oxygen and carbon dioxide; for this reason the dual action

Storage and packaging 251

system contains alkaline hydroxide specific for carbon dioxide

(Matsushima et al., 1995).

6.2.3 Control parameters

The packaging process starts from the selection of the packaging materials

and techniques, through conditioning, when coffee is filled into the packs

and sealed, to the final phase when bags and cans are stored and taken to

the place of consumption. Each step of this long chain must be carefully

monitored to prevent or at least correct all possible defects so that the

quality (see 1.3) of the ready-to-use coffee blend complies with

regulations and is up to consumers’ expectations. Various parameters

are monitored throughout the different phases of the packaging process. Packaging material

The packaging material must be adapted to the packaging technique and

to the intended shelf life. As mentioned in 6.2.1, the most important

packaging requirements concern the barrier properties of the chosen

material, i.e., its permeability to gas and moisture. Such barrier properties

can be classified as shown above in Table 6.5.

Glass and metals are air- and watertight; the choice of packaging

materials based on their barrier properties, therefore, only applies to

multi-ply materials, which can be prepared from different polymers. Their

permeability is a function of the type of polymer, of possible surface

treatments on the polymer, their thickness and the combination chosen.

In addition to the packaging material, particularly in the case of tight

containers, the packaging shape must also be decided. The main

parameters to be considered in this choice are volume of headspace,

which should be as small as possible to limit the amount of aroma

volatilized to reach saturation equilibrium, and the ratio of volume to

surface area, which should also be as low as possible, particularly when

incomplete barrier materials are used, since permeability depends on

surface area. Choice of the packaging technique

As seen in 6.1.1, oxygen consumption is the main cause of deterioration

of roast coffee, which should therefore be protected from entering into

contact with oxygen as soon as roasted. The lower the level of residual

252 Espresso Coffee

oxygen in the package, the longer the shelf life of the product. But the

lower the amount of residual oxygen the more effective the package

should be as a barrier to keep oxygen at the same level throughout the

shelf life of the product. The best techniques to improve shelf life are, in

decreasing order, pressurization, inert gas and vacuum packaging; the

technique must be chosen as a function of the desired shelf life.

The choice of the most suitable material is also influenced by the

degree to which it can withstand positive internal pressures, when coffee

is pressurized, or negative ones, when it is vacuum packaged. Table 6.6

shows the implications of these two parameters on shelf life and packaging

requirements. Conditioning

When coffee is air-packed no conditioning is needed; packages are just

filled up and sealed, the only parameter to be monitored being air tightness,

as the package needs only to protect coffee from moisture pick-up.

In vacuum packaging, conditioning consists in creating a 300–500 mbar

vacuum inside the pack before sealing it. The vacuum level must then be

monitored, particularly in rigid cans, which might otherwise collapse. If

the vacuum is replaced by inert gas, the applied gas pressure and/or the

internal pressure of the package must be checked. Pressure levels near

atmospheric pressure are normally applied, while valveless packages are

conditioned under a slight vacuum.

In pressurization, conditioning consists in exactly the same process, but

the cans are pressure-resistant and coffee is not degassed. In this case, in

Storage and packaging 253

Table 6.6 Packaging parameters




Shelf life


Absolute Pint

(Atm) Material

In air:

tight 16–18 1 nRT/V Rigid

with valve 10–12 3 1.01* Indifferent

Under vacuum 4–6 4–6 0.3 Better


Under inert atmosphere 1–2 6–8 1.01* Indifferent

Under pressure <1 >18 Up to 2.2 Rigid

*Pressure at which the valve opens.

addition to residual oxygen, the pressure in the package should also be

monitored once degassing has made the package swell.

The airtight seal, already important in air packaging, becomes critical

with all the other types of packages because any leak in the packaging

material would allow oxygen in, drastically reducing shelf life.

Pressure is measured by standard pressure gauges; the airtight seal is

checked by measuring residual oxygen and/or, in the case of positive

internal pressures, by plunging the package in water and checking for

leaks. Various sensor methods are available to measure residual oxygen

levels. Logistics

In logistics, pressure, temperature and time are the key parameters.

The pressure level must be monitored to avoid the risk of swelling or

even explosion from a build-up of pressure in valveless packs. This is why

air transportation of coffee should be avoided, unless pressurized

compartments are available.

The spoilage rate of roasted coffee increases seven-fold with every



C rise in temperature. Coffee should therefore be stored at low

temperature. Furthermore, low temperatures slow down evaporation of

volatile compounds, so that coffee kept at low temperature retains more

aroma than at normal temperature, as recently confirmed on roasted and

ground coffee in air by Cappuccio et al. (2001). Cooled coffee beans

should, however, be left at room temperature for a few hours before

grinding to trap the volatile aroma present in the cells in the oil and

melanoidins; a large aroma loss at grinding might otherwise dull the

flavour of the cup. In pressurized coffee the binding of aroma by oil is

hindered if the oil becomes too viscous, as is the case at temperatures

approaching 0


C; pressurized coffee should therefore be kept at room

temperature for a few days after packaging. Temperature fluctuations

should be avoided as they accelerate spoilage.

The last parameter to be monitored is time, i.e. the total storage period,

which should always be kept shorter than the actual shelf life of the

product. On average, coffee is consumed within three months from

production which makes time a critical factor only for the least

sophisticated packaging techniques. Testing of finished products

Although lipid oxidation plays a major role in coffee deterioration, the

peroxide test, commonly performed to assess rancidity in foods, is not

254 Espresso Coffee

sensitive enough for coffee. Peroxides reach significant levels only in the

very last stages of spoilage, when coffee has already become undrinkable.

Linoleic acid content is another parameter of little use to analysts, as it

requires reference values measured on the same blend when fresh.

Electron spin resonance (ESR) measurements have also been carried

out to detect the presence of free radicals; the intensity of the ESR signal

has been found to be inversely related to the level of perceived acidity,

which rises with deterioration; however, even this technique is not

sensitive enough.

The concentration of specific volatile compounds in the headspace of

the package or of the cup, measured by gas chromatography, has been used

to determine the degree of spoilage. For instance, the ratio of

2-methylfuran to butanone, the ‘aroma index’ M/B, drops from 3.2 in

fresh coffee to 2.3 in spoilt coffee. However, such indices depend on blend

composition and on roasting technique, normally unknown to the analyst,

so that these parameters can only be relied on to a very limited extent.

None of the test methods available can yet replace sensory evaluation,

and tasting remains the most reliable of all test methods.


Alves R.M.V., Mori E.E., Milanez C.R. and Padula M. (2001) Roasted and

ground coffee in nitrogen gas flushing packages. Proc. 19th ASIC Coll.,


Anese M. and Nicoli M.C. (2003) Antioxidant properties of ready-to-drink

coffee brews. J. Agric. Food Chem. 51, 942–946.

Anese M., De Pilli T., Massini R. and Lerici C.R. (2000) Oxidative stability

of the lipid fraction in roasted coffee. Ital. J. Food Sci. 12, 457–463.

Anese M., Manzocco L. and Maltini E. (2001) Determination of the glass

transition temperatures of solution ‘A’ and HMW melanoidins and

estimation of viscosities by WLF equation: a preliminary study. In J.

Ames (ed.), Melanoidins in Food and Health Volume 2, Proceedings of

COST Action 919 workshop held at the Technical University of Berlin,

7–8 April 2000 and the Institute of Chemical Technology, Prague 22–23

September 2000, pp. 137–141; EUR 19684 EN, Belgium.

Barbera C.E. (1967) Gas-volumetric method for the determination of the

internal non odorous atmosphere of coffee beans. Proc. 3rd ASIC Coll.,

pp. 436–442.

Cappuccio R., Full G., Lonzaric V. and Savonitti O. (2001) Staling of roasted

and ground cofffee at different temperatures: combining sensory and GC

analysis. Proc. 19th ASIC Coll., CD-ROM.

Cardelli C. and Labuza T.P. (2001) Application of Weibull hazard analysis to

the determination of the shelf life of roasted and ground coffee. Lebensm.

Wiss. Technol. 34, 273–278.

Storage and packaging 255

Clarke R.J. (1987a) Packing of roast and instant coffee. In R.J. Clarke and

R. Macrae (eds), Coffee: Volume 2 – Technology. London: Elsevier

Applied Science, pp. 201–215.

Clarke R.J. (1987b) Roasting and grinding. In R.J. Clarke and R. Macrae

(eds), Coffee: Volume 2 – Technology. London: Elsevier Applied Science,

pp. 73–197.

Clarke R.J. and Macrae R. (eds) Coffee: Volume 2 – Technology. London:

Elsevier Applied Science.

Czerny M. and Schieberle P. (2001) Changes in roasted coffee aroma during

storage – influence of the packaging. Proc. 19th ASIC Coll., CD-ROM.

Daglia M., Papetti A., Gregotti C., Berte` F. and Gazzani G. (2000) In vitro

antioxidant and ex vivo protective activities of green and roasted coffee.

J. Agric. Food Chem. 48, 1449–1454.

De Krujf N., Van Beest M., Rijk R. J., Sipilainen-Malm T., Paseiro Losada P.

and De Meulenaer B. (2002) Active and intelligent packaging:

applications and regulatory aspects. Food Addit. Contam. 19, 144–162.

Eggers R. and Pietsch A. (2001) Technology I: Roasting. In R.J. Clarke and

O.G. Vitzhum (ed.), Coffee: Recent Developments. Oxford: Blackwell

Science, pp. 90–107.

Einarson H. (1987) The effect of time, temperature, pH and reactants on the

formation of antibacterial compounds in the Maillard reaction. Lebensm.

Wiss. Technol. 20, 51–55.

Faist V. and Erbesdobler H.F. (1999) In vivo effects of melanoidins. In J. Ames

(ed.), Melanoidins in Food and Health Volume 1. Proceedings of COST

Action 919. Workshop held at the University of Reading, UK, 2–4

December 1999, pp. 79–88; EUR 19684 EN Belgium.

Flament I. (2002) Coffee Flavor Chemistry. New York: Wiley & Sons, p. 118.

Fourny G., Cross E. and Vincent J.C. (1982) Etude pre´liminaire de l’oxidation

de l’huile de cafe´. Proc. 10th ASIC Coll., pp. 235–246.

Full G., Savonitti O. and Cappuccio R. (2001) Staling of roasted and ground

coffee at different temperatures: combining sensory and GC analysis.

Proc. 19th ASIC Coll., CD-ROM.

Goodman B.A., Pascual E.C. and Yeretzian C. (2001) Free radicals and other

paramagnetic ions in soluble coffee. Proc. 19th ASIC Coll., CD-ROM.

Grosch W. (1999) Key odorants of roasted coffee: evaluation, release,

formation. Proc. 18th ASIC Coll., pp. 17–26.

Grosch W. (2001) Chemistry III. Volatile compounds. In R.J. Clarke and

O.G. Vitzthum (eds), Coffee: Recent Developments. Oxford: Blackwell

Science, pp. 68–90.

Hodge J.E. (1953) Chemistry of browning reaction in model systems. J. Agric.

Food Chem. 1, 928–943.

Hofmann T. (2000) Isolation, separation and structure determination of

melanoidins. In J. Ames (ed.), Melanoidins in Food and Health – Volume 1,

Proceedings of COST Action 919, Workshop held at the University of

Reading, UK, 2–4 December 1999, pp. 31–43; EUR 19684 EN.

256 Espresso Coffee

Hofmann T., Czerny M., Calligaris S. and Schieberle P. (2001) Model studies

on the influence of coffee melanoidins on flavour volatiles of coffee

beverages. J. Agric. Food Chem. 49, 2382–2386.

Homma S. (2001) In R.J. Clarke and O.G. Vitzthum (eds), Coffee: Recent

Developments. Oxford: Blackwell Science, pp. 50–68.

Homma S., Aida K. and Fujumaki M. (1986) Chelation of metals with brown

pigments in coffee. In M. Fujimaki, H. Kato and M. Namiki (eds),

Amino-Carbonyl Reactions in Food and Biological Systems.

Amsterdam: Elsevier, pp. 165–172.

Huynh-Ba T., Courtet-Compondu M.C., Fumeaux R. and Pollien Ph. (2001)

Early lipid oxidation in roasted and ground coffee. Proc. 19th ASIC Coll.,


Karmas R. and Karel M. (1994) The effect of glass transition on Maillard

browning in food models. In T.P. Labuza, G.A. Reineccius, V. Monnier,

J. O’Brien and J. Baynes (eds), Maillard Reaction in Chemistry, Food and

Health. Cambridge: The Royal Society of Chemistry, pp. 164–169.

Labuza T.P., Cardelli C., Andersen B. and Shimoni E. (2001) Physical

chemistry of carbon dioxide equilibrium and diffusion in tempering and

effect on shelf life of fresh roasted ground coffee. Proc. 19th ASIC Coll.,


Manzocco L., Calligaris S., Mastrocola D., Nicoli M.C. and Lerici C.R. (2001)

Review of non-enzymatic browning and antioxidant capacity of

processed foods. Trends Food Sci. Technol. 11, 340–346.

Manzocco L., Calligaris S. and Nicoli M.C. (2002) Assessment of pro-oxidant

activity of foods by kinetic analysis of crocin bleaching. J. Agric. Food

Chem. 50, 10.

Massini R., Nicoli M.C., Cassara` A. and Lerici C.R. (1990) Study on physicoand

physico-chemical changes in coffee beans during storage. Note 1.

Ital. J. Food Sci, 2, 123–130.

Matsushima T., Oguro N. and Ichiyanagi S. (1995) Stability improvement of

roasted and ground coffee by oxygen absorbent. Proc. 16th ASIC Coll.,

pp. 426–434.

Namiki M. (1990) Antioxidants/antimutagens in food. Crit. Rev. Food Sci.

Nutr. 29, 273–300.

Nicoli M.C., Anese M., Manzocco L. and Lerici C.R. (1997) Antioxidant

properties of coffee brews in relation to the roasting coffee. Lebensm.

Wiss Technol. 30, 292–297.

Nicoli M.C., Anese M. and Calligaris S. (2001) Antioxidant properties of

ready-to drink coffee beverages. Proc. 19th ASIC Coll., CD-ROM.

Nicoli M.C., Innocente N., Pittia P. and Lerici C.R. (1993) Staling of roasted

coffee: volatile release and oxidation reactions during storage. Proc. 15th

ASIC Coll., pp. 557–566.

Nicoli M.C., Toniolo R. Anese M. (2004) Relationship between redox

potential and chain-breaking activity of model systems and foods. Food

Chem. 88, 79–83.

Storage and packaging 257

O’Brian J. and Morissey P.A. (1989) Nutritional and toxicological aspects of

Maillard browning reaction in food. Crit. Rev. Food Sci. Nutr. 28 (3),


Pittia P. (2002) Personal communication.

Radtke R. (1979) Zur kenntnis des sauerstoffverbrauchs von rostkaffee und

seiner auswirkung auf die sensoriesch ermitelte qualitata des kaffeegetranks.

Chem. Mikrobiol. Technol. Lebensm. 6, 36–42.

Rooney M.L. (1995) Overview of active food packaging. In M.L. Rooney

(ed.), Active Food Packaging. New York: Blackie Academic and

Professional/Chapman and Hall, pp. 1–37.

Sanz C., Pascual L., Zapelena M.J. and Cid M.C. (2001) A new ‘aroma Index’

to determine the aroma quality of a blend of roasted coffee beans. Proc.

19th ASIC Coll., CD-ROM.

Sheikh-Zeinoddin M., Perehinec T.M., Hill S.E. and Rees C.E.D. (2000)

Maillard reaction causes suppression of virulence gene expression in

Listeria monocytogenes. Int. J. Food Micr. 61 (1), 41–49.

Shimoni E. and Labuza T.P. (2000) Degassing kinetics and sorption

equilibrium of carbon dioxide in fresh roasted and ground coffee. J.

Food Process Eng. 23, 419–436.

Sivetz M. and Desrosier N. (1972) Coffee Technology. Westport, CT: AVI


Steinhardt H. and Holscher W. (1991) Storage related changes of low-boiling

volatiles in whole beans. Proc. 14th ASIC Coll., pp. 156–174.

Steinhart G., Luger A. and Piost J. (2001) Antioxidative effect of coffee

melanoidins. Proc. 19th ASIC Coll., CD-ROM.

Vermeiren L., Devlieghere F., Van Beest M., de Krujf N. and Debevere J.

(1999) Developments in the active packaging of foods. Trends in Food

Sci. Technol. 10, 77–86.

Whitfield F.B. (1992) Volatiles from interactions of Maillard reactions and

lipids. Crit. Rev. Food Sci. Nutr. 31(1/2), 1–58.

Yaylayan V.A. and Huyghues-Despointes A. (1994) Chemistry of Amadori

rearrangement products: analysis, synthesis, kinetics, reactions and

spectroscopic properties. Crit. Rev. Food Sci. Nutr. 34 (4), 321–369.

258 Espresso Coffee



M. Petracco

For all types of coffee beverages, brewing is intrinsic to the relevant

definition (Petracco, 2001). This holds especially in the case of espresso,

where, among all subsequent technological processes through which

coffee must go before finally getting into the cup, the last one –

percolation – is the stage crucially contributing to the beverage’s qualities.

On the one hand, it is possible to sort, roast, grind and package coffee

following all the established rules for an espresso coffee, as described in

the preceding chapters, and finally prepare a cup by infusion or drip

filtering. The beverage so obtained certainly cannot be called an espresso.

On the other hand, the inverse operation could be reasonably well

performed: it could be possible to prepare an espresso cup (even if not

the best one) by using a low quality raw material, which has been

inadequately roasted and improperly ground, provided that it is

transferred into the cup by an extraction process following the espresso

percolation process, as will be discussed in 7.4 below. With the

importance of extraction as an integral part of the definition of espresso

ascertained, we have to face the necessity of giving a scientific definition

of percolation, and establishing the conditions for a correct espresso



7.1.1 Definition of percolation

Percolation takes place each time a fluid flows through a porous medium.

In hydrodynamics, this concept materializes in many aspects of our

everyday life: rainwater percolates through sand in the soil; the engine

lubricant percolates through the oil filter. While such phenomena are

cited in common language as filtration, this term is used improperly. In

fact, filtration strictly refers to the process of separation undergone by a

suspension (solid particles dispersed in a liquid) flowing through a

filtering baffle: the baffle may consist of a bed of packed particles, a fibrous

tangle or simply a perforated plate. Unlike other forms of liquid/solid

separation (sedimentation, centrifugation, electro-deposition, etc.), filtration

cannot take place without percolation; conversely, a wide range of

percolation phenomena, where there is removal of mass from the solid bed

instead of accumulation, bear no relationship to filtration but rather to


The term percolation is, in a more general meaning, used also in the

language of topology, the branch of geometry that studies the characteristics

of forms. Under this meaning, an abstract property of space can

spread from one area to another according to fixed mathematical laws

(Stauffer, 1985). A classical example is Bernoulli’s percolation through a

lattice randomly composed of conducting or insulating segments. If at

least one uninterrupted series of conducting segments stretches from one

end of the lattice to the other, electric current may flow through the

lattice, which is thus called ‘percolating’ (Mandelbrot, 1977). This

reference is pertinent to our case, as the application of topological, or

more generally, mathematical methods may contribute to the understanding

and modelling of hydrodynamic phenomena and, in particular,

of espresso coffee brewing.

7.1.2 Definition of coffee cake

Back to hydrodynamics, it is useful to explain what is meant by ‘porous

medium’. It is an aliquot of space, possibly bounded by walls impermeable

to fluids, and partially filled with a solid material leaving empty spaces,

filled in their turn with a pre-existing fluid that can be displaced by the

percolating fluid (Bear and Verrujt, 1987). A simple example: a bucket

full of gravel has interstices between pebbles, which are filled up by air. By

pouring water into the bucket, the liquid takes up the space previously

held by air and fills up the interstitial spaces. The measure of porosity is

called ‘fraction void’, “, which is defined (Kay, 1963) as

” . empty spaces volume

total volume of the porous body

A porous medium may be composed of a connected structure of

interstices, where the spaces are intercommunicating. It can therefore

260 Espresso Coffee

be filled up by a fluid, a phenomenon called invasion. Conversely, the

interstices may form a non-connected structure (e.g. insulating foam,

made up of many gas cells or bubbles entrained in a solid polymer matrix).

The first case concerns us most, even if ground coffee does not have the

same porosity structure of other typical percolating media with connected

pores, such as tangles of fibres (the most common example is felt) or

sponges (where solid fibres have grown together to form a single body).

Actually, ground coffee behaves as a bed of particles, namely an aggregate

of solid granules in mutual contact under the force of their own weight,

which do not penetrate each other owing to their solid state. Such a bed

may be unstable and prone to rearrangement by vibration, or remain

stable under gentle vibrations but get reshuffled under stronger forces (e.g.

overturning). Particles may also stick together, forming a mass having its

own shape and capable of being handled as one whole solid (e.g. sintered

metals, made up of metal pellets fused together by mutual friction).

Apart from an ideal case, all particles are formed by materials that are

to a certain extent elastic or plastic; therefore, particles are subject to the

force of their own weight and to other external forces causing reversible or

irreversible deformations. Particles form beds that can be compacted by

compression, with reduction of the total apparent volume of the bed.

Owing to this compaction and to a certain degree of adhesion of the

external particle layer, beds can take up a shape stable enough to

withstand handling (with due care in the case of coffee).

This last description provides an adequate definition of a ground coffee

bed, also known as a ‘cake’, ready to be extracted in the espresso machine.

By contrast, ground coffee used in other brewing methods, as for instance

drip filter coffee, lets percolation occur through particle beds that are

much less tightly aggregated than espresso cakes.




7.2.1 Physical description

From a physical point of view, preparing a cup of espresso coffee means

percolating a definite quantity of liquid through a compact bed of ground

coffee particles, the cake. The latter is usually contained in an impermeable

metal holder, generally made of stainless steel for its inalterability

and easy cleaning. The holder is popularly called a ‘filter’, in spite of the

Percolation 261

inappropriateness of this term as pointed out above: it is approximately

cylindrical in shape, with a perforated bottom. Percolation takes place

when liquid penetrates the interstices between the particles at a certain

spatial velocity, displacing the pre-existing air. The concept of spatial

velocity, borrowed from chemical reactor theory, refers to the flow of

liquid per unit section of the bed. When dealing with espresso, the liquid

involved in this process is water, heated to a temperature just above 90



by means of direct or indirect heating.

The driving force that generates the flow is produced by the pressure

drop undergone by the hot water that enters the cake with positive

hydrostatic pressure. Pressure, a well-known notion of general physics,

defines a force exerted perpendicularly over a unit of area. In the domain

of fluid dynamics, pressure refers to an intensive-type property, defined

instant by instant in any point of space taken up by a fluid; it can be

represented by means of a scalar field. The existence of a pressure field

within a fluid yields a potential energy (Bernoulli’s piezometric energy)

that can be easily transformed into kinetic energy, thus giving speed to

elementary masses of fluid.

Pressure energy is partly dissipated as heat by the friction generated

when water flows through the particles, and partly gradually transformed

into kinetic energy of the beverage which flows into the cup, according to

the general equation (Bullo and Illy, 1963):

Pe . Pr . Pc . Pa


Pe . driving pressure (water pressure at the filter inlet)

Pr . pressure drop from bed resistance

Pc . kinetic pressure drop

Pa . atmospheric pressure

The average water velocity must remain constant through the cake in

order to provide a continuous jet, while pressure decreases along the

percolation axis until it reaches atmospheric level. At this point, the

‘liquor’ obtained by percolation flows into the cup underneath by


It is beyond dispute that every percolation method requires some

pressure (at least several millimetres of water column height) to overcome

the head loss required for obtaining a flow through ground coffee beds;

nevertheless, a higher relative pressure (one order of magnitude, or two) is

needed when dealing with finer grinds. Even higher pressure, up to 10 atm

(corresponding to 103 327mmH2O: four orders of magnitude higher),

262 Espresso Coffee

must be used when trying to pass water through a compacted cake of

finely ground coffee, as happens in espresso brewing. The energy

expended during this kind of operation produces interesting effects, like

driving of micron-size solid particles and oil droplets into the cup. This

may influence the beverage’s properties dramatically.

Hydrostatic pressure is generally generated by a device constituted of a

rotary or reciprocating pump: some set-ups nevertheless utilize compressed

air or steam pressure to pressurize a water tank, or exploit the centrifugal

force of a rotary system to push water through the cake. Even a head

formed by a water column (tens of metres high, anyway!) might serve the

task. A description of technological devices that have been developed in

nearly a century, since the birth of espresso coffee, to produce this

characteristic beverage will be discussed in 7.4 below. However, it must be

stressed that an accurate design and a precise mechanical manufacture of

espresso machines are most essential because these machines, even the

more economical models for home usage, are far from being simple (see

later, Figure 7.6).

7.2.2 Physico-chemical description

The intimate contact of water with roasted coffee is the cardinal

requirement for producing a coffee beverage. From a chemical point of

view, percolation induces two processes: the extraction of the watersoluble

substances and the emulsification of the insoluble coffee oils. Both

phenomena together are responsible for the considerable differences

between pure water and the espresso brew (Table 7.1).

Solid-liquid extraction produces the removal of a soluble fraction – in

the form of a solution – from an insoluble permeable solid phase where it

belongs. Two steps are always involved in extraction: contact of solid and

solvent to effect mass transfer of solubles to the solvent, and separation of

Percolation 263

Table 7.1 Espresso brew contrasted with pure water

Colour Brown-black, due to partially defined pigments (melanoidins)

Refractive index Increased by the presence of solutes

Surface tension Reduced by surface-active substances present

pH Lowered by organic acids and by phosphoric acid

Electric conductivity Higher, owing to the ions presents

Viscosity Higher, mainly due to the oils in emulsion

Density Higher, due to the high solute content

the resulting solution from the residual solid, this latter step is usually

done by filtration. The mechanism of the first step is favoured by

increased surface per unit volume of solids to be extracted, and decreased

radial distances that must be traversed within the solids, both of which are

favoured by decreased particle size as it occurs in coffee ground for


The main dependent variable that can be objectively measured (along

with the obviously all-important sensory ones, which are largely based on

subjective evaluation) is brewing yield, namely the ratio between the

mass of the coffee material that passes into the cup and the total coffee

material used (the balance to be disposed as spent grounds). Brewing

temperature exerts a preponderant role on yield variability, which may

range between 20% and 30%, and darker roasting procures higher yields.

It is worth mentioning that yield is a different concept from beverage

concentration – or strength, measured in grams of extracted matter per

litre of beverage – and can be set independently, if a different brewing

formula (coffee/water ratio) is chosen.

The general principle ruling extraction is Fick’s law of diffusion, which

gives the quantity s of solute diffusing from a solid particle surrounded by

a liquid. It can be written in the following form (Barbanti and Nicoli,


s . k T/  A/x (C c) 


k . constant, depending on molecular factors

T . absolute temperature

  . viscosity of the liquid, f (T)

A . layer cross section around the particle

x . layer depth

C . solute concentration in the solid

c . solute concentration in the liquid

  . contact time

In espresso percolation, the short extraction time does not allow

equilibrium to be reached. This calls for consideration of reaction kinetics

too, especially the branch dealing with the diffusion of water entering,

and of molecules of various solutes leaving, a cellular structure. Optical

and electronic microscopic examinations of the cell structure of coffee

show the presence of soluble substances both outside and inside the cell

wall already in the green bean (Dentan, 1977), and no evidence of pores

or channels running through the cell wall has yet been produced.

264 Espresso Coffee

Therefore, it may be assumed that percolation provokes an extraction

mainly by washing out the outer surface of the particle, rather than by

diffusing from the interior of the particle to the solution.

A naive way of thinking could suggest that the more one extracts from

one’s purchased material, the better: nothing is less true, quality-wise.

Since the many chemical species identified in roasted coffee (more than

1800 so far) exhibit different extraction rates (Lee et al., 1992), it is

logical that extraction yield relates to sensory quality too, leading to the

two celebrated converse brewing errors: under- and over-extraction. The

former is due to lower yields (deriving from whatsoever reason, like low

water temperature, short contact time or too coarse grind), when the most

soluble substances – the acidic and sweet ones – are predominantly driven

into the cup, producing the so-called under-extracted beverage. The latter

happens when higher yields are obtained, by forcing more substances –

the bitter and astringent ones – into the cup. It is therefore evident that

brewing yield should be adjusted according to personal taste, in agreement

with the nature of the beans used (origin, blend, roasting).

Only few authors have tackled so far the study of coffee extraction

kinetics (Voilley and Clo, 1984; Nicoli et al., 1987; Cammenga et al.,

1997). The only substance that has been studied to a considerable extent

is caffeine, the constituent producing the most remarkable physiological

effect (Spiro and Selwood, 1984; Spiro and Page, 1984; Spiro and Hunter,

1985; Zanoni et al., 1991; Spiro, 1993). Caffeine exhibits high solubility

in hot water (Macrae, 1985; Cammenga and Eligehausen, 1993);

nevertheless, its quantitative extraction may be obtained only by coffee

preparation methods that permit a long contact period between water and

ground coffee (Peters, 1991). In espresso percolation, caffeine extraction

yield is usually within the range 75–85% (Petracco, 1989); this relatively

low figure may be explained by the fact that caffeine is extracted from

within the cell, by diffusion, and not by simple washout kinetics.

The other important chemical phenomenon occurring during coffee

percolation is the emulsification in water of non-soluble matter, mostly

composed of lipids, present in coffee in quantities ranging from 8 to 16%

dry basis (see, that are mainly contained inside the cells of roasted

coffee. By means of the energy transferred to the bed by the water pressure

drop, the lipid fraction in broken cells may easily form an emulsion.

Optical microscopy of emulsified lipids in an espresso brew (see Figure

8.2) (Heathcock, 1988) shows that they are present as small droplets

varying from a minimum of 0.5 mm to a maximum of 10 mm in diameter.

As will be seen in 8.1.2, the stable emulsion formed during espresso

percolation does not coalesce spontaneously, even after weeks of rest.

None of the artifices commonly used to break up emulsions (filtration,

Percolation 265

dilution, addition of electrolytes, etc.) overcomes the reluctance of

espresso coffee to be separated into two distinct liquid phases. The small

diameter of droplets is a factor that helps in precluding coalescence;

moreover, the density of the disperse phase is not very different from that

of the solution – this prevents droplets from buoying up and concentrating

on the free surface. The key factor of emulsion’s stability, whose

importance in the sensory context will be discussed in 9.5, is constituted

by a layer of surface-active substances surrounding each droplet and thus

generating the chemical stability (Navarini et al., 2004).



The first attempt to set up a mathematical model for the espresso brewing

process dates back to the early 1960s (Bullo and Illy, 1963), when the key

role of the pressure drop due to the hydraulic resistance of the ground

coffee cake was first recognized. As a starting approximation, it can be

assumed that the percolation process can be hydraulically defined by an

equation associating five variables, sufficient to characterize the process in

its macroscopic physical aspect:

V .




R.t; T.t.  dt


V . volume of beverage in the cup

T . temperature inside the coffee cake

t . percolation time

 p . pressure difference above and under the cake

R . hydraulic resistance of the cake.

This choice of variables is not the only possible one: for instance, the

volume of beverage in the cup, V, might be replaced by its derivative in

relation to time, the instantaneous flow Q. Moreover, the hydraulic

resistance R depends on sub-variables as the dose and particle size of the

ground coffee, as well as the compacting of the bed. Anyway, five

variables are sufficient to characterize the process in its macroscopic

physical aspect, in agreement with the traditional rule of the barista (the

espresso bartender):

266 Espresso Coffee

To prepare an espresso cup correctly, it is necessary to set the right

temperature and pressure, then to adjust the hydraulic resistance (by

grinding and compacting), until the right volume of beverage is obtained

at the right time.

To model the phenomenon theoretically, that is to forecast the value of

a dependent variable once the values of the independent ones have been

set, a semi-empirical approach uses flow, pressure and temperature

continual sensors to obtain flow curves as a function of time (Petracco

and Suggi Liverani, 1993). Classical hydraulics would apply Darcy’s law –

‘The loss of pressure in a pipe is proportional to the flow rate of the liquid

through it’ – or:

 p . RQ


Q . flow

R . hydraulic resistance of the system

But two observations that lie in clear contradiction with this equation

emerge from the experimental curves (Figure 7.1):

Percolation 267

Figure 7.1 Pressure dependence of flow during percolation

n flow is not constant for a constant pressure drop, but after an initial

transient peak it decreases in time until it reaches an apparently

asymptotic value (dependent on temperature);

n the mean flow is not proportional to the pressure applied, but increases

with it up to a certain value, then remains constant or even decreases.

The anomalous behaviour of the pressure is not linked to the extraction

reactions, started by high water temperatures not far from 100


C. It is

actually feasible to percolate ground coffee with as cold water as possible

(in practice not lower than 4


C to prevent freezing) so as to minimize any

influence of solubilization on the physical phenomenon. Data show that

the asymptotic non-monotonic behaviour persists also under such

conditions, even if higher flows are needed (Figure 7.2).

However, the flow’s quasi-exponential decay suggests that the bed of

ground coffee particles should present a time-dependent geometry

(Baldini, 1992). An indication that some form of modification does

occur inside the bed derives from electron microscopy of the bed structure

(Figure 7.3), which is made up of coarse particles containing fragments of

cell walls, called fines. The latter are assumed to migrate to and eventually

to concentrate at the bottom of the bed. Indirect confirmation comes from

268 Espresso Coffee

Figure 7.2 Temperature dependence of flow during percolation

tests where the direction of percolation was reversed, using an extraction

chamber that could be turned upside-down.

The direct flow behaves normally, remaining unchanged even when

the pump is stopped for a brief moment. Surprisingly, flow increases when

the percolation chamber is inverted (Figure 7.4). This effect can be

explained by assuming that the finest particles, which concentrate in a

lower section of the cake causing an increase in the hydraulic resistance,

now counter-migrate, with an increase in hydraulic conductivity in the

opposite direction, until the system reaches a new steady state.

Attempts have been made (Baldini and Petracco, 1993) to formulate a

law providing an explanation for this experimental behaviour, but finding

Percolation 269

Figure 7.3 Microstructure of ground coffee particles

Figure 7.4 Direct/inverse discharge curve

an exact analytical solution to the consequent system of differential

equations still remains a difficult, if not impossible, task. Computer-aided

calculation of the numerical solution to these equations becomes feasible

only by fixing the boundary conditions; the coefficients, however, must

always be determined experimentally. The theory has yet to be expanded

to include not only the physical, but also the chemical aspect of

percolation and possibly to allow also forecasting sensory features like

foam, body and flavour.

Another method, alternative to differential equations, is offered by

‘molecular ontology’, a branch of ‘naive physics’ that studies physical

phenomena starting from experimental and qualitative observation via a

heuristic approach (Bandini and Cattaneo, 1988). A model is constructed

by breaking down a physical phenomenon into its constituent elements

(in the present case, coffee particles and elemental volumes of water), and

by describing their local interactions, whose synergism builds up the

system’s overall behaviour.

Simulations of espresso percolation obtained by powerful computers

(Figure 7.5) demonstrate that particles with a given possibility of

movement tend to migrate and to accumulate in a specific critical

section simulating a metallic filter (Bandini et al., 1997). This technique

shows micro-vortexes and allows the trajectories of the liquid and the

particles to be followed (Cappuccio and Suggi Liverani, 1999).


As can be seen from what has been said so far, preparing an espresso coffee

requires equipment able to:

n keep a portion of ground coffee in the form of a compact cake inside a

container that allows the brew to drip out while holding back the

particles (spent grounds);

270 Espresso Coffee

Figure 7.5 Stages of cake percolation simulation

n impart to the water a temperature close to 90


C and a pressure

customarily fixed at 9 relative atmospheres.

Such equipment is called an ‘espresso coffee machine’, often shortened

into ‘espresso machine’ or, in Italy, ‘coffee machine’. Coffee machines are

available on the market in many models, ranging from compact light ones

for occasional home usage to the professional computerized units. The

main parts of the machine are the pump, the heat exchanger and the

extraction chamber, as sketched in Figure 7.6.

The origin of espresso coffee machines dates back to the beginning of

the 1900s, when the pioneer Bezzera patented a machine in which

pressure was applied to water by simple boiling in an autoclave. The

drawback of that design is that the attainment of the desired pressure

levels (around 1.5 atm) required a very high temperature, well above



C. These conditions cause the extraction of substances normally

insoluble, imparting an unpalatable bitter flavour to the brew. With the

knowledge of the espresso technique available now, a brew obtained with

such a machine could no longer be called an espresso, due to the much

lower pressure applied; this principle is still widely employed today in

Italian home coffee brewing by the commonly used inexpensive machine

called a ‘Moka’ (Petracco, 2001).

This drawback can be avoided by separating the generation of pressure

from that of temperature. After the compressed-air machine, whose

technology was probably ahead of its time (Illy, 1935), a lever-system was

developed in 1945 by Achille Gaggia. It consisted in a spring which,

wound up by the muscular force of the barista, released energy – via a

Percolation 271

Figure 7.6 Schematic flow diagram of an espresso machine

piston – to the water contained in a pressurization chamber. This system is

still sometimes favoured in traditionalist coffee houses; its performance

depends on the operator’s manual ability in properly dosing and applying

force and time, rendering espresso preparation an ‘art’ with uncertain

results. The system has the advantage of allowing the manufacture of

machines that do not require electric energy, as heating can be

accomplished by a gas burner, and no pump is needed to pressurize the


7.4.1 The pump

The spreading of electricity allowed the automation of espresso machines,

mainly through the introduction of a pump as the source of pressure.

Professional pumping devices consist in an impeller with fins, driven by

an electric motor, and in an adjustable bypass valve. The purpose of the

latter is to withdraw part of the water delivered by the rotor at the

discharge outlet of the pump, and to recycle it to the suction inlet: this

power dissipation is legitimate because it enables the pump head (back

pressure) to remain constant, independent from the variability of the

hydraulic resistance (coffee cake) downstream. Also, the repeatability of

percolation has considerably benefited from the use of such pumps,

because an electric motor is far less prone to imprecise behaviour than

muscular force.

Machines for home usage, where neither the encumbrance nor the cost

of a centrifugal pump is justified, utilize preferably a vibration pump. This

is a volumetric pump where a small piston is set in a vibrating

reciprocating motion by an electromagnet coupled with a return spring.

These pumps exhibit a very high head (up to 30 atm), heavily depending

on the hydraulic resistance downstream. As a consequence, the quality of

the beverage in the cup is strongly influenced by the amount of ground

coffee used, and by the degree of grinding and of compacting. A less

common alternative, used to generate water pressure in commercial

domestic machines, is centrifugal force.

7.4.2 The heat exchanger

The invention of waterproof electric resistors, which permitted the

manufacture of direct heat exchangers, was another major innovation. In

modern espresso machines (see Figure 7.6), the heating of water is based

on a low-pressure boiler of a capacity of 3–10 litres, heated by a

waterproof electric resistor immersed in it (seldom by a gas burner). This

272 Espresso Coffee

reservoir supplies hot water for tea and other infusions, and saturated

steam for cappuccino steamed milk. In addition, it works as heat storage to

warm up the espresso percolation water, which runs through a completely

different circuit. This water comes directly from the tap and is driven by

the pump through a coil immersed inside the boiler, where it reaches

dynamically the desired temperature (many degrees lower than the

temperature in the boiler itself). This system, particularly cumbersome

and expensive, is applied in all professional machines and only in some of

the classiest home-use machines, where it permits an ample production of

steam for cappuccino preparation.

In smaller machines, the boiler is replaced by a direct heat exchanger,

consisting of a metal block with an enclosed electrical resistance fed by a

bimetallic thermostat or, in more modern models, by an electronic power

supply driven by a thermocouple sensor. Cold water is forced by the pump

to pass through a sinuous circuit inside the block, and then passes directly

through the coffee cake beneath. In professional machines, conversely,

temperature is regulated by varying the boiler pressure (see also 7.5.6).

7.4.3 The extraction chamber

The extraction chamber consists in an upper block, on which is inserted

(like a bayonet) a filter-holder cup. The ground coffee portion is poured

and compacted into this filter. Since the block protrudes from the body of

the machine, it is recommended to heat it by means of a small calibrated

flow of boiler water passing through a built-in tube. When the pump is

switched off, water is no longer able to flow through the cake and remains

under pressure within the chamber, due to the increased bed resistance

caused by hydration of the spent coffee grounds. If not overcome, this

residual pressure would cause troublesome splashes when removing the

filter-holder from its block. Therefore, the hydraulic circuit of all espresso

machines is equipped with some kind of device that allows depressurization

of the extraction chamber at the end of percolation, and this is

usually achieved by calibrated check valves, or servo-controlled valves

synchronized with the pump power switch.

7.4.4 Water pre-treatment

Another all-important piece of equipment is the pre-treatment unit for

the water, as the latter usually comes directly from the mains water supply

system with substantial calcium and magnesium content. Any system of

heating hard water must forestall the formation of alkaline-earth

Percolation 273

carbonate deposits on heated surfaces, and the consequent harmful

decline in heat exchange efficiency (Cammenga and Zielasko, 1997). An

operation most commonly implemented to prevent malfunctioning in

professional machines is water softening. It consists in passing the water

through a bed of ion-exchange resins, which capture calcium ions present

in the water and replace them with sodium ions, whose salts are

completely soluble. Water softeners need regular regeneration by washing

with sodium chloride: this is a vital operation for ensuring efficient water

conditioning and a long service life of the espresso machine. Finally, the

effect of water acidity and hardness on espresso cake may affect extraction

(Fond, 1995). More details about water influence on espresso will be

found in 7.5.5 below.


The three main features that lead to the definition of espresso are:

n ‘on the spur of the moment’: the brew must be prepared just before

serving it;

n ‘in a short time’: brewing must be fast;

n ‘under pressure’: mechanical energy must be spent within the coffee


Whereas the feature ‘on the spur of the moment’ (or, more colourfully,

‘extemporaneously’) refers not so much to the coffee brewing method as

to the ‘espresso lifestyle’, the other two features are strictly related to


These general features are necessary for the identification of an

espresso. Nevertheless, they are insufficient to establish whether that cup

has been prepared correctly, that is, whether it falls within the interval of

the n-dimensional space of the fundamental preparation variables that

allows the consumer to expect a sensory experience that is up to the wellestablished

fame of espresso. In this interval, the ideal espresso is

represented by a point, and probably not always by the same point nor by

one single point. Reaching this point is an art, and is so related to the

subjective sphere of the individual that it cannot be completely clarified

in a technical book. The present work can, however, strive to fix the

boundaries of the n-dimensional interval. They could be imagined as the

limits of an ‘excellent espresso’, with each one being a necessary, but not a

sufficient condition (Swartz, 1997).

274 Espresso Coffee

The first difficulty encountered is establishing the number of dimensions

needed to describe an ‘excellent espresso’. Certainly, they are more

than those five variables mentioned in 7.3, because each of the latter is

the aggregation of several sub-variables. Moreover, some variables have

not as yet been mentioned, inasmuch as, if we wished to tackle every

aspect of coffee preparation, a myriad of variables should be considered

(Illy, 2002). For practical reasons, we shall limit ourselves to discussing

the following ten:

n ground coffee portion

n particle size distribution

n cake porosity

n cake shape

n water quality

n temperature

n pressure

n percolation time

n cake moistening

n machine cleanliness

The first five variables refer to the raw materials: four of these are related

to the product to be extracted (the ground coffee cake) and can be seen as

factors influencing its hydraulic resistance; an additional one is dedicated

to the second main raw material of espresso coffee, water, which is often

neglected because of its inexpensiveness. The last five variables refer to

the machine. Three of them, temperature, pressure and percolation time,

are, together with hydraulic resistance, the primary independent variables

of the percolation equation (see 7.3). The last two, cake moistening

methods and machine cleanliness, are also important, yet neglected,


7.5.1 Ground coffee portion

A ‘portion’ is the weight of roast and ground coffee required for preparing

one cup. The acceptable extreme range of values for a portion may be

from a minimum of 5 g to a maximum of 8 g. Minimum values are

sometimes used in coffee houses that wish to ‘save’ on coffee: these values

are only barely tolerable when dark roasted coffee is used, for it has a

higher content of soluble substances. Maximum values are often

encountered in home-brewed coffee, as the host inclines to overdose

the serving in order to offer the guest the best possible coffee cup. This

Percolation 275

practice is risky, however, because, as will be seen in 7.5.4, an excessive

amount of ground coffee does not permit sufficient expansion during cake

wetting. This causes over-compacting (deriving from the imbibition

forces), which disturbs percolation and may cause the deposit of solids

into the cup.

In professional practice, a double portion is generally used because

coffee-bar espresso machines are designed to meet consumer demand in

peak hours. This need is fulfilled in two ways: the simultaneous use of

more than one percolation head (up to five, or most often three per

machine) or by preparing two cups per head at the same time. The social

role of an espresso, particularly in Italy, is such that a person wishing to

enjoy a cup of espresso hardly ever enters a coffee house alone, therefore

espresso machines are usually optimized for a double portion, equal to 13 g

of ground coffee. Accordingly, the optimal portion used for one cup

(namely the single serving asked for by an individual customer) is of 6.5 g

in the case of arabica blends, something less for robusta blends because of

higher extraction yield. Also roasting degree influences yield, up to a

point: therefore darker roasts allow for slightly reduced dosages.

Unlike in quality-control laboratories, where the portion is determined

by weighing on analytical balances (with a precision of 0.1 g), in espresso

bars it is measured volumetrically using the so-called ‘grinder-doser’. It is

an additional appliance fitted on the professional grinding machine (with

a capacity of at least 2 kg of ground coffee per hour), consisting in a device

with rotary paddles into which ground coffee powder drops from the

grinding tools. The height of the pallets can be adjusted, so that the

portion released at each ‘stroke’ corresponds to a standard weight. At

home, the easiest way to determine a portion is by using a measuring

spoon, which allows the right volume of coffee powder to be drawn from

the package, or the right volume of beans to be ground each time. The

volume-to-weight equivalence is not absolute, as it depends on the

density of the powder, which in turn depends on the degree of compacting

in the package; variability is even larger for whole beans. In practical

terms, however, the error is negligible.

7.5.2 Particle size distribution

Undoubtedly, coffee granulometry is a crucial factor for espresso, but it is

also one of the least investigated, and probably least understood, variables

of percolation. Particle size refers to the measurement of the geometric

characteristics of a population of particles constituting a powder, and to

its representation. The most studied characteristic is size distribution,

276 Espresso Coffee

while particle shape is also a matter of great interest. Common ways of

measuring it have been described in 5.3.

Size distribution of a powder sample may be represented by a histogram

or, when many measure points are used, by a chart showing a grading

curve. The curve is plotted on a diagram (see Figure 5.4) in which the

axis of the abscissa, usually on a logarithmic scale, carries the ‘equivalent

diameter’ variable: this variable can be defined differently according to

the technique adopted, but it is generally linked with particle size. The

percentages of particles belonging to each size group are plotted on the

axis of the ordinates in differential (frequency chart) or integral form

(cumulative curve). The percentage may refer to the number of particles,

to their volume or to their surface, or even to their mass or weight. By

calculation, it is possible to pass from one form of chart to the other,

once the relationships between particle shape and density factors have

been determined. The comparison of ground coffee batches obtained

from different grinding equipment shows differences that characterize

each ground coffee by its own grading curve, as fingerprints identify a


A fundamental parameter to obtain a correct percolation, the so-called

‘fineness’, refers to the average particle diameter and depends on the

adjustment of the grinder. The most suitable particle size distribution for

espresso percolation is not, however, a single-size grading; consequently,

neither the average diameter nor the mode of the curve are sufficient to

describe the correct size distribution.

An empirical compromise reached by grinder manufacturers resorts to

an intermediate distribution, which can be modelled either by power-law

or log-normal distribution (Vicsek, 1991), occasionally with a typical

bimodality or even trimodality. Such a complex characteristic of particle

size is believed to produce a double effect: on the one hand, it forms a

coarse fixed structure, which allows the correct flow through the cake; on

the other hand, it forms a large quantity of fines of high specific surface,

which permit the extraction of a large amount of soluble and emulsifiable

material. Furthermore, the coffee cake appears to behave as a selffiltering

structure, retaining the fines that have been displaced by the


7.5.3 Cake porosity

Theoretically, the most suitable particle size distribution is the one that

exposes the maximum surface area to the action of water. However, such

a distribution would not permit any porosity of the cake capable of

Percolation 277

maintaining the hydraulic resistance within limits that allow a correct

flow. While in aggregations of particles having the same shape the

fraction void (see 7.1.2) does not vary with different powder fineness, it

can be profoundly modified when size distribution is changed (for

example, by transforming a monomodal into a bimodal distribution: as

is the case when sand is added to gravel in order to render it less

permeable). The same statement of bimodality holds true for the shape of

particles, because irregular particles (such as fines made up of fragments of

fractured cells) may easily sneak through regular-shaped particles and

increase specific surface.

Another factor affecting cake porosity is compacting. The bed of loose

ground coffee poured into the filter by dropping must be compacted,

usually by hand. The compacting force may vary from a few kilos, for a

vertical upward thrust (a tamping plate is usually built-in with the

grinder), to approximately 20 kgf for downward compacting by a hand

tool, the tamper: in the latter case, the force of the thrust may be

increased by part of the operator’s own weight.

Compacting influences percolation even when exerted weakly, that is,

the difference in hydraulic resistance between a loose bed and a weakly

compacted one is large, but there is only a minor variation between

weakly or more forcefully compacted beds. A plausible explanation for

this phenomenon could be the formation of small localized channels in

the upper layer of a loose bed which, when struck by a jet of boiling

water, does not resist to impacting water as firmly as a compacted cake


Compacting forces much larger than those just mentioned (40 kgf or

more) are employed in the industrial production of a patented system of

single portions of ground coffee, sealed between two layers of filter paper

(Illy, 1982). These portions (also called pods or servings) are percolated in

especially designed semi-automatic coffee machines aimed at the

preparation of an espresso at home, where they free the operator from

manual dosing of ground coffee and from most possible errors in


7.5.4 Cake shape

The shape of the coffee cake, determined by the filter’s shape, is

important when considering peculiar forms that deviate from the

traditional ones, universally accepted by current coffee machines. The

classical shape of the double dose is cylindrical, around 12mm in height

by 60mm in width, the exact dimensions depending on the machine. If

278 Espresso Coffee

this height-to-diameter ratio of about 0.2 is lowered, very fine grounds are

required for coffee to percolate within a standard time; this choice would

degrade the reproducibility of percolation, because of the creation of

localized channels by water. A larger height-to-diameter ratio would give

the cake the shape of a column; in this case, optimal extraction would

demand an excessively high pressure (as in liquid-phase chromatographic

columns). Conversely, for normal pressure values the grind would be so

coarse as not to offer a sufficient number of fractured cells, resulting in a

low extraction. Consequently, only minor variations from the optimal

form of the established filter can be recommended.

An important phenomenon to be taken into account when considering

the shape of the cake is the expansion of the bed due to the swelling of

coffee particles when wetted. This is a phenomenon of colloidal

imbibition, where a chemical hydration of the organic materials (such

as the polysaccharides) takes place. The water involved in wetting the

coffee grounds is bound, therefore, in an aggregation state differing not

only from the ordinary liquid state, but also from that of interstitial water,

which is subject to the laws of capillary phenomena. This is shown by the

fact that imbibition water cannot be eliminated by oven heating at



C for a protracted length of time, but only by raising the oven

temperature by several degrees. During expansion, wet coffee grounds

exert a pressure comparable to that of the wooden wedge used in the past

to cleave marble blocks, whose force of expansion, once it is driven into

the block and wetted, prevails over the cohesive forces of the stone and

splits the block into slabs. On account of this behaviour, an empty space is

left over the ground coffee cake inside the extraction chamber. The actual

expansion of the cake varies with blend, roasting degree and dose, and

determines the exact headspace (around 5–6 mm) needed to prevent


A final shape characteristic worth mentioning is the filter perforation.

In 7.2.1, the term filter was mentioned as inappropriate for indicating the

perforated plate that constitutes the base of the cup containing the

ground coffee bed. Actually, this plate does not act as a filter but as a

supporting structure for the compacted coffee cake. The diameter of the

holes (usually 0.25 mm) is designed to prevent the coarse particles that

break away from the lower layer of the cake from passing into the cup. In

order to prevent solidified caramelized residue from obstructing the holes,

they are triangular in vertical cross-section: the coffee powder comes into

contact with the smaller sized opening of the hole, beneath which there is

an open truncated-conical cavity. The filter contribution to the overall

hydraulic resistance is small, and becomes negligible when the holes get


Percolation 279

7.5.5 Water quality

The second most important ingredient of espresso – after the coffee itself

– is water, constituting more than 95% of any coffee beverage. However

self-evident this statement may seem, the role of water in coffee

preparation must be taken into due consideration. It would be otherwise

senseless to devote utmost care and attention to only one of the two

ingredients, while neglecting the other. Yet, this regularly occurs, partly

because the choice of the proper water is left to the end user and partly

because that person (espresso bartender, home user) has an extremely

limited influence over the choice of the water. In fact, very few people

brew coffee using mineral water, whose characteristics are stated on the

label and are supposed to be constant. Most consumers trust tap water

quality, which they rarely purify in any way.

It must be stressed that the quality of treated water coming from public

waterworks is perfectly suitable for human consumption (WHO, 1993)

and further purification or disinfecting treatments are unnecessary.

However, treatment of drinking water may turn out to be beneficial in

two critical cases; one is linked to sensory perception, and the other

concerns the use of the espresso machine.

In the first case, the beverage must be free from any unpleasant foreign

flavours left over in the water by generalized disinfecting treatments, such

as chlorination. Salts that release free chlorine (hypochlorites) lend the

beverage their peculiar taste, which is perceptible when the initial

concentration of Cl2 exceeds 0.5 mg/l. A system most commonly used for

removing chlorine taste involves passing the liquid through a bed of

granules of activated carbon, which has the ability to adsorb many odourbearing

substances. A drawback in the use of activated carbon is that

saturation cannot be easily detected, so that it is necessary to regenerate

or replace the bed regularly.

The second critical case is related to water hardness, namely to the

water’s calcium and magnesium content. On heating, these cations

produce insoluble salts (mainly carbonates, but also sulphates and

silicates). These salts tend to precipitate in the form of compact plaques

(scaling), particularly on the heated surfaces. Such scales form a coating

on the surfaces, which detrimentally affect the heat exchange coefficient.

This may lead to serious technical troubles with the espresso machine, in

particular causing a reduction in heat transfer effectiveness with a possible

consequent widening of the fluctiations in temperature (too hot/too cold)

and risk of failure in electrical resistance due to overheating.

Potable water may be nearly always defined as hard, that is, having a

total hardness that exceeds 15 French degrees, corresponding to a

280 Espresso Coffee

concentration of 150 mg/l of CaCO3. In some waterworks in Italy, for

instance, water hardness may reach a peak of 37 French degrees. This is

not a reason for health concern, because the hardness for drinking water

recommended by Italian law (DPR 236, 24/5/88) ranges from 15 to 50

French degrees. It is also worth remembering that mineral waters may

exceed 100 French degrees.

Direct use of hard mains supply water in both professional and home

espresso machines produces unacceptable deposits in 1–3 months.

Manufacturers of home-use machines advise removing these deposits by

periodic washing (see 7.5.10) with weak acid solutions: citric or formic

acid, even vinegar, will do. In professional machines, scaling is prevented

by fitting the machine upstream of a cleanup system specifically active on

calcium and magnesium ions, but nearly as active on other metal ions.

Two types of treatments are commonly available: softeners or demineralizers.

A softener consists of a bed of ion exchange resins that traps calcium

ions and replaces them with sodium ions (which do not affect hardness,

and do not form deposits) to a hardness depending on the saturation of

the resin. Optimal water hardness to avoid scales and have good

percolation is around 9 French degrees, as recommended by manufacturers.

The resins must be periodically regenerated by washing with NaCl

so as to restore their effectiveness. In the case of treatment of very hard

water, the amount of sodium released (in combination with other ionic

species present) may become perceptible to taste and increase water pH

on heating. Two troubles may be encountered with softeners:

n A functional problem, inasmuch as localized channels form in the

resin bed after some time, diminishing the effect of ion exchange.

This can be avoided by stirring the resins from time to time.

n A hygiene risk, namely the danger of a possible proliferation of

microorganisms in the bed. This can be averted by using beds

containing a bacteriostatic additive.

Demineralizers eliminate Ca.. ions without introducing any foreign

ions. Ion exchange of the demineralizer employs either separate beds of

resins, fitted in series, or one single bed in which resins have been mixed

but that can be separated by floating. One resin is of the so-called cationic

type and captures Ca.. and Mg.. by replacing them with H.; the other

resin is of anionic type and captures the negative ions (Cl, HCO3



– – and others) by replacing them with hydroxyls, which neutralize

hydrogen ions forming H2O. These types of resins must be regenerated

separately with HCl and with NaOH, respectively. A more modern type

Percolation 281

282 Espresso Coffee

of demineralizer is based on the principle of reverse osmosis. Pressure

energy is employed to filter water through a semi-permeable membrane

capable of retaining ions, which can be discarded as concentrated

solution, while letting the smaller H2O molecules through. With the use

of a suitable membrane or of a graded bypass, water can be obtained at the

desired degree of hardness.

The main influence of hardness on beverage quality is an increase in

percolation time, compared to soft water below 2 French degrees.

However, only minor differences occur when water hardness is raised

above 8 French degrees to typical mains water values (Rivetti et al.,

2001). Hence, it is not recommended to soften or demineralize water

below this value, otherwise a coarser grind should be used to compensate

for the increase in percolation time. Water pH also affects percolation

time even if in a less predictable way, probably in association with

hardness. The influence of hardness and pH on percolation might be

explained by the change, due to Ca and H ions, in foam-producing and

emulsifying properties of the natural surfactants present in roasted coffee

(see It is unclear, though, whether the increase in percolation

time is to be ascribed to increased foam and emulsion viscosity, with

respect to pure water, or to the interstitial formation and precipitation of

insoluble calcareous soaps or other pH-dependent material.

7.5.6 Temperature

Quantitatively, this relationship between the rate a reaction proceeds and

its temperature is determined by the Arrhenius equation:

k . A e–E/RT


k . reaction rate

A . constant

E . activation energy

R . gas constant (0.082 litre atm/mole



T . temperature

Extraction of coffee, like any dissolution reaction of a chemical species in

water, depends on the constant of equilibrium and on the rate. Higher

temperatures are matched by increased percentage of extracted material.

It is disputable whether the additional fraction of substances extracted

at higher temperature improves the taste of the brew, or enhances any

desirable property. The appeal of an espresso prepared under optimal

Percolation 283

hydraulic conditions is known to increase with temperature, up to a

maximum at 92


C. The extraction of bitter and astringent substances at

higher temperatures reduces the overall sensory appreciation.

Unfortunately, in practice, temperature cannot be easily singled out

from the other percolation variables, because hydraulic resistance of the

cake depends on extraction temperature. For instance, reference has been

made in 7.3 to the fact that cold water flows are larger than hot water

ones, when obtained through the same cakes and under the same

conditions. This difference becomes noticeable when trimming the

canonical temperature of 92


C down to 70


C; the effect is slighter, yet

significant, at temperatures around the optimal value (Andueza et al.,

2001). This makes temperature regulation one of the most important

operations in designing an espresso machine (see 7.4.2).

The control of espresso water temperature in professional machines is

established by varying steam pressure in the boiler, where water is heated

passing through an immersed coil acting as a heat exchange surface. The

temperature of steady water in equilibrium with its own steam is univocally

dependent on pressure. Espresso water heated this way has a reasonably

constant temperature, thanks to over-dimensioned exchange surfaces.

Home-use machines, when not equipped with a boiler, are fitted with a

thermostat controlling the temperature of the metallic block, which acts

as a heat exchanger. This thermostat is often built in on the unit, and

fixed; in models where it is adjustable its fine-tuning can usually be done

using a small screwdriver until the temperature of the water dripping from

the empty extraction chamber reaches 90   2



The temperature of the water is certainly the main thermal factor in

percolation, but not the only one. Also important is the cake temperature,

induced by both the mobile and the fixed equipment constituting the

extraction chamber. In order to homogenize temperature in professional

machines, a small quantity of circulating boiler water constantly heats up

the fixed equipment (the block). When using home espresso machines, it

is recommended to pre-heat the extraction group by letting a small

quantity of hot water drip through before inserting the cake. The same

operation may also be used to pre-heat the empty cup, filling it up with

hot water before percolating the coffee; an excellent espresso is always

served in a warm cup. To this purpose, professional machines are equipped

with a cup-warming grid.

7.5.7 Pressure

The role of water pressure on percolation is widely acknowledged, but the

influence of the method of exerting pressure on the dry cake has not been

adequately studied and deserves more attention. Pressure does not remain

constant during the process of percolation (as often assumed), but varies

as a function of the characteristics of the hydraulic circuit above the cake.

This can be verified by continual pressure sensors: the recordings yield a

function p . p(t), which exhibits an initial transient rise, increasing more

or less slowly. The dry coffee bed still lacks adequate cohesion and is

susceptible to resettlement; therefore the initial phase appears to be

decisive in attaining the definitive microstructure of the cake. After

wetting, ground coffee particles swell up, interpenetrating each other

firmly, and percolation can follow its stationary course. This effect may

explain why, as already mentioned in 7.3, the flow is not constant during

percolation, rather showing a non-monotonic dependence on the average

pressure and an asymptotic decay.

The pressure value universally applied in professional espresso

preparations, where a centrifugal pump is employed, is 9 relative

atmospheres or, technically, 9 kg/cm2 (equivalent to 10 ata, i.e. absolute

technical atmospheres). It is easily adjusted by regulating the bypass valve

located downstream of the pump, which, as already mentioned in 7.4.1,

can supply a flow exceeding percolation needs.

This value results from a series of ‘trial and error’ attempts in the early

years of espresso machine technology, whose results were measured by

consumer satisfaction. It is interesting to point out how an alert espresso

operator is in a good position to test any slight variation in coffee

preparation procedures, judging immediately from the consumer’s face (a

genuine sensor) and drawing useful indications for further trials. In the

course of time, such a method has established the optimal pressure level

that minimizes failure possibilities in obtaining a perfect espresso.

Hydrodynamic experiments (Baldini and Petracco, 1993) demonstrate

that the average flow ceases to depend linearly on pressure in the

proximity of 9 atm, and that an increase in pressure actually causes a

decrease in the average flow despite the already discussed law of Darcy.

This suggests that the combined machine/cake system is capable of selfadjusting

to overcome pressure variations. Such a phenomenon may lead

to cups of coffee that are apparently (hydraulically) similar, though

originating from completely different extraction conditions, and thus

potentially having different brew qualities.

In home espresso machines, which are driven by a small vibration pump

without bypass valve, pressure cannot be adjusted and is settled by the

hydraulic resistance of the cake. This makes the grind even more

important: it should be adapted to each machine model, because the

wrong particle size does not only influence percolation time, but

extraction pressure as well.

284 Espresso Coffee

7.5.8 Percolation time

Percolation time is the variable that most depends on the person who

prepares the espresso. Nevertheless, in consumers’ perception, extraction

time is not an independent variable, but rather the result of the other

preset parameters (pressure and temperature, grind and tamping/compacting).

The real variable allowed to consumer’s choice is actually the

volume of brew in the cup, as it is the easiest thing to check: once the cup

is full, percolation is stopped without minding the prescribed percolation


The volume in the cup is therefore left to the personal taste of the

consumer, who chooses it according to the local food tradition and

personal eating habits. Optimal results are attained with approximately

25–30 ml volume reached in 25–30 seconds, common in Southern Europe

(Italy, Spain, Greece).

Very short extraction times (less than 10 s) would produce a brew that

is under-extracted, tasting dull and diluted.

Local tastes may ask for a volume up to 70, 90 or even 120 ml or more,

particularly in northern Europe (Germany, Holland, Scandinavia) and in

the USA. However respectable this criterion may be, it appears quite

insidious because it does not set any rule for the duration of percolation,

which can become too long (over 60 s). The resulting brew is then overextracted,

tasting bitter and astringent.

It is therefore recommended that the volume in the cup be considered

as a consequence of percolation, and that the extraction time be

controlled as the main variable. Long expertise produces a consensus on

a percolation time in the optimal range of 25–35 seconds, obtained by

commanding other variables, particularly grinding. If a larger brew

volume is desired, a better cup is prepared by increasing the particle

size of the grounds instead of prolonging extraction time. However, its

body and flavour cannot be compared to those of a full-fledged Italian

espresso, prepared following the full set of rules (see 8.3).

7.5.9 Cake moistening

The moistening of the ground coffee cake is a term that indicates a

specific operation, also called pre-infusion, for which every espresso

bartender has their own personal method. It consists of pouring a certain

quantity of hot water into the extraction chamber so as to wet the ground

coffee, starting the actual percolation a few seconds later. The resulting

beverage is claimed to have a richer body and better flavour, which could

Percolation 285

derive from an improved extraction caused by the swelling of the bed

during the pre-infusion period.

Dynamics similar to pre-infusion can also be observed in the common

straightforward preparation. Water never enters the extraction chamber

instantaneously, not even when the hydraulic circuit consists simply of a

tube linking the pump to the chamber; this is because the start-up

transient of the pump produces a wave of pressure that reaches the ground

coffee surface in a finite time. Moreover, a number of espresso machines

are equipped with an empirical device – a choke nozzle (also known by its

French term ‘gicleur’ or jet), which somewhat renders the pre-infusion

operation automatic by slowing down the water entering into the

chamber. The opening of the device, inserted in series on the hydraulic

circuit just above the chamber, is usually about 0.1mm wide.

7.5.10 Machine cleanliness

Let us conclude the list of variables influencing percolation by dealing

with cleanliness – truly irksome and yet also a delight for baristas and

home users alike. Following the same hygiene rules recommended for

other kitchenware, percolation equipment must be kept clean to prevent

foreign matter from being dragged into the cup and to eliminate any

possibility of proliferation of microorganisms. Nevertheless, it must be

stressed that the organoleptic qualities of the brew may also be negatively

influenced by excessive cleanliness. An established tradition dictates that

coffee pots (and espresso machines too) must be washed with warm

running water, avoiding soap or common household detergents to avert

the risk of off-flavours from residues.

This rule stems from two considerations, the first of which is easy to

explain: the presence of metal filters densely arranged along the water and

coffee path may hold back some cleansing agent in spite of the most careful

rinsing. The residual detergent is then gradually released, spoiling the

flavour of the cups brewed soon after washing. The second reason to forbid

using common detergents is subtler and concerns the surface state of the

metallic parts in contact with coffee. After the first few percolations, a thin

layer of coffee oil and extremely fine solid particles deposits on the metallic

parts, acting as a passivator of the metal surface. Excessively clean coffee

machines produce a beverage having the typical metallic flavour of new

material, probably produced by the catalytic activity of the metal (often

copper or its alloys, but also aluminium, steel, chrome) or, possibly, by

metal ions released into the water. This activity is apparently inhibited by

the formation of a protective layer of coffee matter, which can be washed

286 Espresso Coffee

away if strong or abrasive detergents are used. Mild detergents, devoid of

scents or foaming additives, have been developed specifically for coffee

machines; their regular use is recommended to prevent the accumulation

of sediments in critical sections of the equipment.

A particular aspect of cleanliness regards the removal of calcium

deposits formed by hard water, as detailed in 7.5.5. This is a common

occurrence in home-use espresso machines, where periodical decalcification

is recommended in order to prevent serious damage to the valves and

heating resistor. It is performed by re-circulating for a few minutes a water

solution of a weak acid (acetic, formic, citric), or any decalcification

liquid available on the market, through the looped hydraulic circuit of the

machine. Thorough rinsing must always follow washing.


Andueza S., Maeztu L., De Pen˜a M.P. and Cid C. (2001) Influence of

extraction temperature in the final quality of the Colombian coffee cups.

Proc. 19th ASIC Coll., CD-ROM.

Baldini G. (1992) Filtrazione non lineare di un fluido attraverso un mezzo

poroso deformabile. Thesis, University of Florence.

Baldini G. and Petracco M. (1993) Models for water percolation during the

preparation of espresso coffee. 7th Conference of the European

Consortium for Mathematics in Industry – Montecatini (Italy),

pp. 21–22.

Bandini S. and Cattaneo G. (1988) A Theory for Molecule Structures: the

Molecular Ontology Theory. 2nd International Workshop on

Qualitative Physics, Paris.

Bandini S., Casati R., Illy E., Simone C., Suggi Liverani F. and Tisato F.

(1997) A reaction-diffusion computational model to simulate coffee

percolation. Proc. 17th ASIC Coll., pp. 227–234.

Barbanti D. and Nicoli M.C. (1996) Estrazione e stabilita` della bevanda caffe`:

aspetti chimici e tecnologici. Tecnologie Alimentari, 1/96, 62–67.

Bear J. and Verrujt A. (1987) Modelling Groundwater Flow and Pollution.

Dordrecht: Reidel, pp. 27–43.

Bullo T. and Illy, E. (1963) Conside´rations sur le proce´de´ d’extraction. Cafe´

Cacao The´ 7 (4), 395–399.

Cammenga H.K., Eggers R., Hinz T., Steer A. and Waldmann C. (1997)

Extraction in coffee-processing and brewing. Proc. 17th ASIC Coll.,

pp. 219–226.

Cammenga H.K. and Eligehausen S. (1993) Solubilities of caffeine, theophylline

and theobromine in water, and the density of caffeine solutions.

Proc. 15th ASIC Coll., p. 734.

Cammenga H.K. and Zielasko B. (1997) Kinetics and development of boiler

scale formation in commercial coffee brewing machines. Proc. 17th ASIC

Coll., pp. 284–289.

Percolation 287

Cappuccio R. and Suggi Liverani F. (1999) Computer simulation as a tool to

model coffee brewing cellular automata for percolation processes; 2D and

3D techniques for fluid-dynamic simulations. Proc. 18th ASIC Coll.,

pp. 173–178.

Dentan E. (1977) Structure fine du grain de cafe´ vert. Proc. 8th ASIC Coll.,

pp. 59–72.

Fond O. (1995) Effect of water and coffee acidity on extraction. Dynamics of

coffee bed compaction in espresso type extraction. Proc. 16th ASIC Coll.,

pp. 413–421.

Heathcock J. (1988) Espresso microscopy and image analysis. Private


Illy E. (1982) Coffee machine which brews coffee beverages from pods of

ground coffee. European Patent 0 006 175.

Illy E. (2002) The complexity of coffee. Sci. Am. 286 (6), 86–91.

Illy F. (1935) Apparecchio per la rapida ed automatica preparazione

dell’infusione di caffe`. Italian Patent 333293/26 December 1935.

Kay J.M. (1963) Fluid Mechanics and Heat Transfer. Cambridge: Cambridge

University Press, p. 253.

Lee T., Kempthorne R. and Hardy J. (1992) Compositional changes in brewed

coffee as a function of brewing time. J. Food Sci. 57, 1417–1421.

Macrae R. (1985) Nitrogenous components. In R.J. Clarke and R. Macrae

(eds), Coffee: Volume 1 – Chemistry. London: Elsevier, p. 117.

Mandelbrot B. (1977) The Fractal Geometry of Nature. New York: Freeman &

Co, p. 126.

Navarini L., Ferrari M., Suggi Liverani F., Liggieri L. and Ravera F. (2004)

Dynamic tensiometric characterization of espresso coffee beverage. Food

Hydrocolloids 18, 387–393.

Nicoli M.C., Dalla Rosa M. and Lerici C.R. (1987) Caratteristiche chimiche

dell’estratto di caffe`: Nota 1. Cinetica di estrazione della caffeina e delle

sostanze solide. Industrie Alimentari 5/87, 467–471.

Peters A. (1991) Brewing makes the difference. Proc. 14th ASIC Coll.,

pp. 97–106.

Petracco M. (1989) Physico-chemical and structural characterisation of

espresso coffee brew. Proc. 13th ASIC Coll., pp. 246–261.

Petracco M. (2001) Beverage preparation: brewing trends for the new

millennium. In R.J. Clarke and O.G. Vitzthum (eds), Coffee: Recent

Advances. Oxford: Blackwell Science, pp. 140–164.

Petracco M. and Suggi Liverani F. (1993) Espresso coffee brewing dynamics:

development of mathematical and computational models. Proc. 15th

ASIC Coll., pp. 702–711.

Rivetti D., Navarini L., Cappuccio R., Abatangelo A., Petracco M. and Suggi

Liverani F. (2001) Effects of water composition and water treatment on

espresso coffee percolation. Proc. 19th ASIC Coll., CD-ROM.

Spiro M. (1993) Modelling the acqueous extraction of soluble substances from

ground roasted coffee. J. Sci. Food Agric. 61, 371–373.

288 Espresso Coffee

Spiro M. and Hunter J.E. (1985) The kinetics and mechanism of caffeine

infusion from coffee: the effect of roasting. J. Sci. Food Agric. 36,


Spiro M. and Page C.M. (1984) The kinetics and mechanism of caffeine

infusion from coffee: hydrodynamic aspect. J. Sci. Food Agric. 35,


Spiro M. and Selwood R.M. (1984) The kinetics and mechanism of caffeine

infusion from coffee: the effect of particle size. J. Sci. Food Agric. 35,


Stauffer D. (1985) Introduction to Percolation Theory. London: Taylor &


Swartz N. (1997) The concepts of necessary and sufficient conditions. Class


Vicsek T. (1991) Modelling of the structure of coffee layers by random

packing of spheres. Personal communication.

Voilley A. and Clo G. (1984) Diffusion of soluble substances during brewing

of coffee. In B.M. McKenna (ed.), Engineering and Food, Volume 1.

London: Elsevier, pp. 127–137.

WHO (1993) Guidelines for Drinking-water Quality, 2nd edn. Volume 1 –

Recommendations. Geneva: World Health Organization, pp. 122–130.

Zanoni B., Pagliarini E. and Peri C. (1991) Modelling the aqueous extraction

of soluble substances from ground roasted coffee. J. Sci. Food Agric. 58,


Percolation 289


The cup

M. Petracco

At the beginning of this book, a first portrait of an espresso coffee serving

was presented as ‘a small heavy china cup with a capacity just over 50 ml,

half full with a dark brew topped by a thick layer of a reddish-brown foam

of tiny bubbles’. A chemist would not like such a description and prefer a

less poetic but more explicit wording. Nevertheless, this definition

characterizes the espresso beverage quite well, at least from a macroscopic

point of view, pinpointing its main features:

n a small volume, indicating that the brew is more concentrated than


n a composite physical nature, where the foam is as important as the


These features are, of course, the result of a very special percolation

method used to prepare the cup, but they could be recognized and

characterized even if nothing were known about the history of the cup

presented to the analyst’s examination. In the following sections espresso

will be described in detail and characterized both from a physico-chemical

and an organoleptic point of view.




Espresso is unstable: all the parameters that are prone to fast modification

after brewing must be determined as soon as possible. The most rapidly

varying property is gas content. Gas release occurs owing to the sudden

pressure drop undergone by the beverage when issuing out from the

espresso machine (see 7.5.7).

Dissolved gases – mainly CO2 – quickly effervesce in the cup, and

bubble up to build a layer of froth. This makes espresso a composite

beverage where two distinct constituents are present: supernatant foam

and underlying liquid.

8.1.1 Foam

Espresso foam – also known by the Italian term of crema – is by itself a

biphasic system composed of gas globules framed within liquid films (called

lamellae) constituted by a water solution of surfactant. These films tend to

set in a configuration of two layers of surface-active molecules facing the

gas, with water molecules between them. The high molecular force in the

film allows its peculiar geometry: a bubble if isolated, or a honeycomb-like

structure of many bubbles growing close together (Figure 8.1).

An important characteristic of the crema is persistence, and it must

survive at least a couple of minutes before breaking up and starting to

show an uncovered spot on the dark surface of the beverage. The

disappearance can be imputed to the phenomenon of drainage, which

causes entrapped water to flow out of the films, leaving only surfactant

molecules to bear the stress of the structure. The films eventually fail due

to their limited elasticity, thus releasing the entrapped gas. Persistence

can be easily measured by watching the surface of a standard cup at

constant temperature and recording the first sign of weakening of the

foam’s layer, revealed by formation of a hole.

As espresso foam is fairly short-lived, its quantity must be measured

soon after percolation. This is done by transferring the whole beverage

The cup 291

Figure 8.1 Pattern of foam in espresso: (left) at beginning of brewing; (right) after

resting overnight, proteins have agglomerated

from the cup to a graduated cylinder, allowing it to stabilize and recording

the volumes of both foam and liquid. Their ratio constitutes a ‘foam

index’ that in well-prepared cups should be at least 10%.

Solidity is a noteworthy rheological property of foam, denoting its

capability to bear for some time the weight of a spoonful of sugar. It can be

quantified in the laboratory by loading the foam with a calibrated wire

mesh ring of a standard weight and measuring the time it takes to


8.1.2 Liquid

The liquid part of the espresso beverage, where the foam stays afloat, is an

even more complex system. Along with the properties of the liquid as a

bulk, the presence of several dispersed phases must be taken into account

to explain the peculiar characteristics of espresso. Physical and chemical

parameters of the beverage are detailed in Table 8.1, which shows typical

analyses of correctly prepared espresso cups (30 ml/30 s).

As already pointed out in Chapters 1 and 7, the polyphasic nature of

the liquid is what makes espresso stand out in comparison with all other

brews, both from a physico-chemical and from a sensory standpoint. The

peculiarity of espresso beverage is the simultaneous presence of three

292 Espresso Coffee

Table 8.1 Typical analysis of an espresso brew

Parameter Pure arabica Pure robusta

Density at 20


C (g/ml) 1.0198 1.0219

Viscosity at 45


C (mPa s) 1.70 1.74

Refractive index at 20


C 1.341 1.339

Refractive index of filtrate at 20


C 1.339 1.339

Surface tension at 20


C (mN/m) 46 48

Total solids (mg/ml) 52.5 58.2

Total solids of filtrate (mg/ml) 47.3 55.6

Total lipids (mg/ml) 2.5 1.0

Unsaponifiable lipids (mg/ml) 0.4 0.2

pH 5.2 5.2

Chlorogenic acids (mg/ml) 4.3 5.0

Soluble carbohydrates (mg/ml) 8.0 10.0

Total elemental nitrogen (mg/ml) 1.8 2.3

Caffeine (mg/ml) 2.6 3.8

Ash (mg/ml) 7.2 7.0

Elemental potassium (mg/ml) 3.2 2.9

dispersed phases coexisting within a matrix, namely a concentrated

solution of salts, acids, sugars, caffeine and many other hydrophilic

substances. These phases are:

1 an emulsion of oil droplets;

2 a suspension of solid particles;

3 an effervescence of gas bubbles, which evolves into a foam.

While the presence of dispersed phases is evident, for the beverage is

opaque, some microscopic investigation is needed to enlighten this

feature. Light microscopy, scanning electron microscopy (SEM) and

transmission electron microscopy (TEM) have been employed to observe

the discontinuities of the matrix. To optimize light microscopy, some

form of highlighting of the particles is needed: interference techniques are

quite useful to reveal bubbles, while staining media specific for lipids (e.g.

osmium tetroxide or ‘Oil Red O’) have proved useful to examine oil

droplets (Heathcock, 1989).

The study of the dispersed phases requires special care, because they

differ in their stability. The effervescence, which can be observed for no

longer than one or two minutes, must be tackled as soon as possible by

some form of ‘freezing’, while the suspension remains stable for days, and

the emulsion even longer. Furthermore, the last is difficult to break in

order to collect the oil phase for analysis.

The relative amounts of stable particles, namely lipid droplets and solid

particles, are strongly skewed in favour of droplets. An optical microscope

picture of a perforated filtering sheet, retaining particles larger than

0.2 mm, shows a few cell wall fragments along with dozens of droplets of

various sizes, up to 10 mm (Figure 8.2).

Their size distribution (Figure 8.3) was evaluated by digital image processing

of light micrographies (Heathcock, 1989) and shows dependence

from the percolation variables used to produce the beverage (see Figure

8.3) (Heathcock, 1988).

Distribution is then one factor affecting the organoleptic appreciation

of the cup. Particle size can be measured by laser scattering analysis, and

average values of 2 to 5 mm have been reported (Petracco, 1989). Density

Density is little influenced by coffee solids. Differences with respect to

pure water are limited to the second decimal digit; this could be explained

by the presence of coffee lipids dispersed as an emulsion, which are

less dense than water and decrease the overall density. Addition of sugar

The cup 293

294 Espresso Coffee

Figure 8.2 SEM of coffee lipid droplets and solid particles filtered out from their

liquid matrix

Figure 8.3 Droplet size distribution

can raise density to around 1.08 g/ml, while synthetic sweeteners like

saccharin or aspartame add sweetness with little or no density increase.

This changes the perception of the sensory variable ‘body’ between

naturally and artificially sweetened espresso. Viscosity

Viscosity is considerably higher than that of pure water, with reported

values approximately double at usual consumption temperatures (Table

8.2). This property is influenced by the presence of dispersed phases,

related to the amount of lipid droplets in emulsion. Increased viscosity has

been associated with improved body, as shown by the comparison of a

correctly prepared cup with a poorly prepared one (Figure 8.3), where

viscosity differences as high as 33% have been recorded (Table 8.2)

(Petracco, 1989). This may be explained by the fact that emulsions

usually display higher viscosity than pure solutions. Refractive index

The refractive index is usually determined to evaluate solute concentrations

in transparent liquids. Espresso’s polyphasic nature makes this

method difficult to apply, and scarcely convenient to predict the body

character. In comparison with the index of pure water (1.333 at 20



espresso beverages display values ranging from 1.339 to 1.341. Refractive

index is influenced mainly by the solute concentration, and the dispersed

phases are less important as indicated by the minor differences found after

filtering aliquots of espresso. Commercial refractometers specially

graduated for coffee are available and correlate with total solids content. Surface tension

Surface tension is a property of interfaces, due to the tendency of

molecules to diffuse from one material to another: pure water–air

The cup 295

Table 8.2 Viscosity of high- and

low-body espresso, compared with

water (mPa s)


C Pure water Low body High body

45 0.61 1.32 1.64

30 0.81 1.51 2.01

interfaces have a tension of 73 10 3N/m at 20


C. The lower values found

in coffee beverages, down to 46 10 3N/m, are due to the presence of

surface-active agents, the surfactants, which accumulate at the interface

between two phases building up monomolecular layers. These molecules,

composed of an apolar chain with a polar head, help to form and stabilize

foam and emulsion, and may also influence the fluid’s behaviour during

percolation through the narrow interstices of the packed coffee cake,

where capillarity plays an important role. The chemical nature of natural

coffee surfactants has not yet been fully clarified: several classes of

complex molecules, like glycolipids or glycoproteins, might exhibit such

behaviour (Nunes et al., 1997). Apparently, in espresso beverages there

are two classes of compounds responsible for two different properties:

foamability seems to be influenced by a proteic fraction, while foam

persistency depends more on a polysaccharidic fraction (Petracco et al.,

1999). An additional interesting effect of coffee surfactants, besides foam

and emulsion promotion, is that they help the beverage to come into close

contact with taste buds in the tongue, thus enhancing the sensory

perception. Total solids

Total solids content – the beverage’s overall concentration – is the most

important characteristic in the chemical composition of espresso, often

perceived by lay consumers as ‘strength’. This property is determined by

drying the brew in an oven to constant weight without filtering.

Accordingly, suspended solid material, emulsified lipids and solutes are

included. Total solids concentration in the beverage is dictated by the

brewing formula, namely the coffee/water ratio, but varies greatly also

depending on both roasting degree and percolation temperature: a darker

roasting produces more solubles, and a higher temperature extracts them

more efficiently (Nicoli et al., 1991). Values as low as 20 mg/ml up to

60 mg/ml may be encountered. The true soluble fraction, typically 90% of

the total solids, can be determined by oven drying after filtering the

liquid. Lipids

The emulsified lipid fraction can be determined by liquid–liquid solvent

extraction. Different amounts have been reported in both espresso and

non-espresso coffee beverages (Petracco, 1989; Peters, 1991; Ratnayake et

al., 1993; Sehat et al., 1993). For an espresso made from a 100% arabica

blend, a figure of the order of 5% of the total solids may be considered as

296 Espresso Coffee

typical (Table 8.1). This is a little less than one-tenth of the total lipids

present in the roast and ground coffee portion, and represents only a

minor contribution to the dietary energy intake: less than 1 kcal per cup.

Within the unsaponifiable fraction of coffee lipids, two compounds

belonging to the diterpene family – cafestol and kahweol – are of

relevance to blood cholesterol level (see 10.3.3); present in roasted

arabica beans at levels around 0.6%, they are transferred to the espresso

brew in small amounts (Urgert, 1997). Acids

Acidity of coffee beverages is an important organoleptic parameter

(Woodman, 1985). The presence of acetic, formic, malic, citric and lactic

acids has been detected in the brew, along with quinic and chlorogenic

acids (Peters, 1991; Severini et al., 1993). The content of the last ones is

of course smaller when using dark roasted coffee, in which they have

largely disappeared (Clifford, 1985). A mineral acid – phosphoric – has

also been found in brews (Maier, 1987), deriving from thermal

degradation of phytic acid, the phosphoric ester of inositol. As already

mentioned in, espresso acidity cannot be described just by pH.

Measurements of this variable obtained using electrode-type instruments

report values ranging from 5.2 to 5.8, showing an increase with roasting

degree (Dalla Rosa et al., 1986a; 1986b) and depending on extraction

time (Nicoli et al., 1987). Carbohydrates

This broad class of compounds is present in coffee seeds, as in all plant

materials, both as simple sugars and as polysaccharides (Trugo, 1985).

Upon roasting, heavy reactions and transformations happen, changing the

balance between soluble and insoluble carbohydrates. While only the

former are relevant to the beverage, the brewing method may influence

the hydrolytic degradation of insoluble carbohydrate polymers, adding to

the solubles content. Typically, only small amounts of monosaccharides

can be found in the brew, but no sucrose is present, because it has already

been transformed on roasting into bitter tasting Maillard compounds

(Trugo and Macrae, 1985). Much research has been done on coffee

carbohydrates (Leloup et al., 1997), but few data on their content in

coffee beverages are available: in espresso a typical figure for soluble

carbohydrates is 8 g/l, corresponding to some 15% of total solids

(Petracco, 1989).

The cup 297 Nitrogen compounds

Nitrogen-containing compounds are present in coffee beverages as

transformed proteic material – grouped under the broad name of

melanoidins – and caffeine, while trigonelline has largely disappeared

during roasting to yield volatile aroma compounds (Macrae, 1985). A way

to analyse this complex class of compounds is by determining total

elemental nitrogen, subtracting the caffeine nitrogen and multiplying for

the standard factor 6.25 to supply a conventional protein content value.

Typical figures are in the neighbourhood of 1.4 mg/ml total N,

corresponding to approximately 4 mg/ml of proteic material.

Caffeine is the most studied coffee component, due to its physiological

activity (see 10.2). Its solubility in water displays a marked dependence

on temperature, with a more than 30-fold increase from standard

conditions to atmospheric boiling temperature, where almost quantitative

extraction takes place at brewing provided enough time is allowed to cope

with the kinetics (Spiro, 1993). Also water hardness seems to influence

caffeine extraction to a certain extent (Cammenga and Eligehausen,


Caffeine presence in espresso beverages has not been much studied.

Data suggest that caffeine extraction from coffee grounds is incomplete

in espresso percolation and, as already indicated in 7.2.2, the yield is

usually within the range 75–85%. This is due to the short time available

with espresso brewing to extract caffeine from the cellular structure, as

shown by fractionated extraction curves (Figure 8.4) (Petracco, 1989).

Therefore, caffeine concentration in the brew ranges between 1.2 mg/ml

up to 4 mg/ml, depending on cup size and blend composition.

Corresponding cup contents range between 60 mg for pure arabica

blends, and 130 mg for pure robusta blends. Minerals

The presence of minerals in the espresso brew is low, as can be seen from

ash content scarcely exceeding 7 mg/ml. Nevertheless, they constitute

about 15% of total solids and, since potassium forms a substantial part of

them, they might be considered as a beneficial contribution to the diet.

8.1.3 Volatile aroma

The meaning of the substantive ‘aroma’ is traditionally considered to be ‘a

strong pleasant smell, usually from food or drink’ (Procter, 1995). In the

298 Espresso Coffee

domain of food science, however, it is better addressed as ‘the quality or

principle of plants or other substances which constitutes their fragrance;

agreeable odor; as, the aroma of coffee’ (Webster, 1913). The following,

more general, definition applies even better to our chemical context: ‘any

property detected by the olfactory system’ (WordNet, 2003), with the

additional specification that by ‘aroma’ we shall designate any set of

‘odorants’, i.e. chemical substances capable of reaching our

olfactory organs.

There is little doubt that the aroma of any coffee beverage is dissimilar

from that of the roast and ground coffee used to brew it, inasmuch as the

composition of its odorants is different (Grosch, 2001). This is the result

of extraction dynamics, which itemizes fast and slow extractors among the

several hundreds of volatile compounds