New In Copyright - Non-Commercial Use Permitted Rights / License: … · 2020. 3. 26. · Acknowledgements I would like to thank Prof. Dr. Renato Amado, Dr. Jean-Claude Villettaz

Research Collection

Doctoral Thesis

A methodology for assessing the quality of fruit and vegetables

Author(s): Azodanlou, Ramin

Publication Date: 2001

Permanent Link: https://doi.org/10.3929/ethz-a-004222689

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

https://doi.org/10.3929/ethz-a-004222689

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

Diss. ETHNr. 14180

A methodology for assessing the quality

of fruit and vegetables

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY

ZURICH

For the degree of

Doctor of Technical Sciences

Presented by

RAMIN AZODANLOU

Dipl. d'Ingénieur Chimiste, Université de Genève

Born December 22, 1965

Citizen of Iran

Accepted on the recommendation of

Prof. Dr. R. Amadô, examiner

Prof. Dr. F. Escher, co-examiner

Dr. J. C. Villettaz, co-examiner

Dr. C. Darbellay, co-examiner

Zurich 2001

Acknowledgements

I would like to thank Prof. Dr. Renato Amado, Dr. Jean-Claude Villettaz and Dr. Charly

Darbellay for giving me the opportunity to work in this field and for their excellent support

and guidance.

Very special thanks are extended to Dr. Jean-Claude Villettaz and Dr. Jean-Luc Luisier for

their helpful discussions, professional criticism and encouragement.

My sincerest thanks go to André Granges who was always prepared to answer my numerous

questions, to Jacques Rossier, Roland Terrettaz, Dr. Christoph Carlen and to Werner

Pfammatter for the fruitful discussions and for supplying the numerous tests samples.

For the supply of fruit samples and the organization of consumer tests I would like to thank

particularly Jean-Luc Tschabold from Migros in Bussigny-Lausanne.

I am very grateful to Stefan Willen for his expert technical assistance and for the

unforgettable moments in the analytical chemistry laboratory.

Elene Morganti, Andrea Schramm and Andrea Schürch contributed to this work through their

diploma theses and semester projects. I would like to thank them for their interest, enthusiasm

and their most valuable input.

I would also like to thank Janine Rey-Siggen, Fabienne Comby, Jeannette Salzmann, Johanna

Claude, Stephanie Berchtold and Stephane Mayor who were assisting me in the sensory and

physico-chemical analyses.

Many thanks also to the panelists who had not only the task to taste the "good" but also the

"bad" samples.

The financial support for carrying out this research work, in the frame of the COST 915 action

("Improvement of quality of fruits and vegetables according to the needs of the consumers"),

has been given by the Swiss Federal Office for Science and Education and the Canton of

Valais. Many thanks for this important support.

Special thanks to Mercedes Gonzalez for her help during this last years.

Finally, I would like to express my gratitude towards my dear dad Amin and my dear mother

Malikeh, for their unfailing support and patience.

TABLE OF CONTENTS IV

TABLE OF CONTENTS

Summary X

Résumé XII

1. Introduction and Scope of Thesis 1

2. Literature Review on Methods for Quality Assessment of Fruit 3

and Vegetables

2.1 Sensory evaluation 4

2.1.1 Hedonic approach for evaluation of food acceptance 5

2.1.2 Sensory panel 6

2.1.2.1 Sensory attributes of the panel 6

2.2 Physico-chemical methods 7

2.2.1 Aroma analysis 8

2.2.1.1 Aroma isolation techniques 8

2.2.1.2 Identification and quantification of volatile compounds 9

2.2.2 Sugar and acid content 10

2.2.3 Texture analysis 11

2.3 Statistical methods 12

2.3.1 Statistical tests for comparing samples 12

2.3.2 Multivariate analysis 12

2.4 References 13

3. A New Concept for the Measurement of Total Volatile Compounds of Food 21

3.1 Introduction 21

3.2 Materials and Methods 23

3.2.1 Analytical Procedure 23

3.2.2 SPME fiber types 23

3.2.3 Samples 24

3.2.4 Sample preparation and extraction of the volatile compounds 24

3.3 Results and Discussion 24

3.3.1 Optimization of the analytical system 24

TABLE OF CONTENTS V

3.3.2 Applications 28

3.4 Acknowledgements 30

3.5 References 31

4. Application of a New Concept for the Evaluation of the Quality of Fruits 33

4.1 Introduction 33

4.1.1 Aroma analysis 33

4.1.2 Our concept for aroma analysis 34


4.2.1 Sample preparation for extraction of the volatiles 34

4.2.2 Analytical procedure 35

4.3 Results and Discussions 35

4.3.1 Heterogeneity with respect to variety or the type of production 35

4.3.2 Determining the appropriate sample size of fruits 36

4.3.3 Maturity of strawberries 38

4.4 Conclusion 38


4.6 References 39

5. Quality Assessement of Strawberries 40

5.1 Introduction 40

5.2 Materials and methods 41

5.2.1 Fruit samples and sample preparation 41

5.2.2 Sensory evaluation 42

5.2.2.1 Consumer tests 42

5.2.2.2 Sensory panel 42

5.2.3 Instrumental analyses 43

5.2.3.1 Determination of total volatile compounds 43


5.2.3.3 Determination of the total sugar and acid contents 45

5.2.3.4 Texture analysis 45

5.2.4 Statistical evaluation 46



TABLE OF CONTENTS VI


5.3.3 Correlation between sensory and instrumental data 49

5.3.4 Hedonic classification for the assessment of strawberry quality 51

5.3.4.1 Total amount of volatile compounds 51

5.3.4.2 Aroma compounds 52

5.3.4.3 Total sugar content 55

5.3.4.4 Texture 55

5.3.5 Development of a model for the assessment of strawberry quality 55

5.4 Conclusions 56


5.6 References 57

6. Changes in Flavor and Texture During the Ripening of Strawberries 59

6.1 Introduction 59





6.2.2.2 Quantification and identification of volatile compounds in strawberries 62

6.2.2.3 Determination of total sugar and acid contents 62



6.3 Results and Discussions 63

6.3.1 Strawberry ripening stages 63

6.3.1.1 Total sugar and total acidity 63

6.3.1.2 Total volatile compounds 64



6.4 Conclusions 71


6.6 References 72

7. Objective Quality Assessment of Tomatoes and Apricots 74

7.1 Introduction 74

TABLE OF CONTENTS VII




7.2.2.1 Consumer tests 77

7.2.2.2 Sensory panel 77



7.2.3.2 Identification and quantification and of volatile compounds in tomatoes 78

and apricots

7.2.3.3 Determination of total sugar content and total acidity 78


7.2.3.5 Conductivity and mineral components analysis 79



7.3.1 Quality assessment oftomatoes 80

7.3.1.1 Sensory evaluation 80

7.3.1.2 Correlation between sensory and instrumental data 81

7.3.1.3 Hedonic classification for the assessment of tomato quality 83

7.3.1.4 Development of a model for the assessment of tomato quality 86

7.3.2 Quality assessment of apricots 87

7.3.2.1 Assessment of five apricot cultivars by consumer tests 87

7.3.2.2 Hedonic classification for the assessment of apricot quality 88

7.3.2.3 Development of a model for the assessment of apricot quality 90

7.4 Conclusions 91


7.6 References 91

8. Conclusions and Outlook 95

Annexes 98

Curriculum Vitae 103

LIST OF ABBREVIATIONS VIII

LIST OF ABBREVIATIONS

AEDA

ANOVA

C

Co

CAR/PDMS

COST

CV

CW/DVB

CW/CAR/PDMS

D

D-R

FID

FPD

G

GC

HPLC

HT

i.d.

M

Ma

MS

N

NA

NE

NIR

NS

OT

P

PA

PCA

PDMS

PDMS/DVB

Aroma extract dilution analysis

Analysis of variance

Carezza

Coop

Carboxene/polydimethylsiloxane

European Co-operation for Scientific and Technical Research

Coefficient of variation

Carbowax/divinylbenzene

Carbowax/carboxene/polydimethylsiloxane

Darselect

Dark red

Flame ionization detector

Flame photometric detector

Green

Gas chromatography

High performance liquid chromatography

Hedonic test

Internal diameter

Migros

Marmolada

Mass spectrometer

Nestlé

Not analyzed

Not evaluated

Near infrared light

Not significant

Odor threshold

Pakoba

Polyacrylate

Principal component analysis

Polydimethylsiloxane

Polydimethylsiloxane/divinylbenzene

LIST OF ABBREVIATIONS IX

PLSD Partial least square deviation

R Red

R (TA/°B) Ratio (total acidity/°Brix)

RI Retention indices

RRI Relative retention indices

%RSD Percentage of relative standard deviation

RT Retention time

SI Significant

SD Standard deviation

SPME Solid phase micro-extraction

TA Total acidity

W White

SUMMARY X

SUMMARY

When a consumer expresses himself about the quality of fruit and vegetables, he will most

likely do it in his own words. He will say "it is good", "it is not good", "it is tasty" or "it is

tasteless". This standpoint represents to the scientist a first established fact. Based on this

short statement he has then to establish a correlation between analytical values and the

consumer's judgement.

The objective evaluation of the quality of fruit and vegetables is a difficult task. This is

mainly due to the fact that every single person is not necessarily influenced by the same

sensory attributes and that the quality scale may vary strongly from one person to another. It

is therefore extremely important to choose a representative sample of consumers to carry out

the hedonic tests.

Moreover, a batch of fruit and vegetables is always heterogeneous; it is therefore important to

pick up a representative sample of the batch. These few considerations show the high degree

of complexity of this type of research work.

Among the different sensory attributes describing fruit quality, sweetness and aroma are the

ones with the strongest effects. Their contribution to the "taste" is essential. The taste of

sweetness can easily be measured by the refraction index (°Brix). We have therefore focused

our efforts on developing a new, non-destructive and rapid method to measure the aroma

intensity of fruit and vegetables. The method consists in trapping the volatile compounds of

fruits on a solid phase microextraction (SPME) and determining the total volatile compounds

without performing any separation. The obtained results are well correlated with sensory

analysis.

It was also necessary to solve the problems deriving from the heterogeneity of the batch, often

leading to very weak correlation indices between the analytical data and the sensory

judgement. This has been done by classifying the fruit and vegetables according to the

hedonic scores given by the consumers prior to instrumental analysis. By doing this, the

correlation between the consumers' appreciation and instrumental data stengthen

considerably. A quality assessement model has been proposed. It confirmed the applicability

of the evaluation of the quality of fruit and vegetables.

SUMMARY XI

Within the framework of this COST 915 project we have been able to develop a new and

rapid method for the evaluation of important quality parameters such as the flavor of fruit and

vegetables. Based on the results obtained, a quality model has been proposed, including limit

values of consumer acceptance for strawberry, tomato and apricot.

RESUME XII

RESUME

Lorsqu'un consommateur s'exprime au sujet de la qualité des fruits et des légumes, il le fait

dans un langage qui lui est propre. Il dira "c'est bon", "ce n'est pas bon", "ça a du goût" ou

"ça n'a pas de goût". Ces prises de position constituent pour le chercheur un premier constat et

un point de départ. Il s'agira donc, sur cette base, de trouver les liens de cause à effet en

analysant un certain nombre de paramètres et en déterminant les corrélations entres les valeurs

trouvées et les avis des consommateurs.

L'évaluation objective de la qualité est une opération difficile. Chaque personne n'est pas

forcément influencée par le même attribut sensoriel et l'échelle de qualité varie fortement

d'une personne à l'autre. Il s'agit donc, au travers de tests consommateurs, de choisir un

échantillon représentatif du publique cible.

En outre, un lot de fruits présente généralement une grande hétérogénéité; il s'agit donc

d'obtenir également un échantillon représentatif du lot à analyser. Ces quelques considérations

indiquent bien le haut degré de complexité de ce projet de recherche.

Parmi les différents attributs sensoriels permettant de décrire la qualité des fruits, le sucré et

l'arôme jouent un rôle essentiel. Ils constituent la notion de "goût", telle qu'exprimée par le

consommateur. Le sucré peut être quantifié au travers de l'indice de réfraction (°Brix). Pour

cette raison nous avons concentré nos travaux sur le développement d'une nouvelle méthode,

non-destructive et rapide, pour mesurer l'intensité de l'arôme des fruits et des légumes.

La méthode consiste a piéger les composés volatiles des fruits à l'aide d'une microextraction

en phase solide (SPME) et de déterminer les composés totaux volatiles sans séparation. Ses

résultats obtenus sont bien corrélés avec l'analyse sensorielle.

Fréquemment, nous avons été confrontés au problème de l'hétérogénéité des lots débouchant

sur des indices de corrélation très faibles entre les données analytiques et le jugement

sensoriel. Une solution a été trouvée en soumettant à l'analyse le même fruit que celui dégusté

par le consommateur, et en le classant préalablement en fonction de son score sensoriel. Ce

qui permet de renforcer la corrélation entre l'appréciation des consomateurs et les données

instrumentales. Ainsi, un modèle de qualité a été proposé. Ce qui confirme l'évaluation de la

qualité des fruits et légumes.

RESUME XIII

Dans le cadre de ce projet COST 915, nous avons pu développer une nouvelle méthode rapide

pour mesurer l'arôme, paramètre important pour l'évaluation de la qualité des fruits et des

légumes. Sur la base des résultats obtenus, un modèle qualitatif a été proposé, incluant les

valeurs seuils telles que ressenties par le consommateur lorsqu'il porte un jugement sur la

qualité des fraises, des tomates et des abricots.

CHAPTER 1 1

CHAPTER 1

INTRODUCTION AND SCOPE OF THESIS

The quality of fruit and vegetables is an extremely complex matter, difficult to describe

objectively. The consumer is not in a position to judge the nutritional quality of a given fruit

and vegetable, however he is able to make a statement on sensory aspects such as shape,

color, texture, juiciness, firmness, taste and aroma.

In the last decades, agronomic research has set priorities to obtain higher yield, better

resistance to diseases and to transport as well as a longer shelf life of fruit and vegetables. The

few sensory aspects taken into account were almost exclusively limited to the appearance

(shape, color) of the product. These parameters are without any doubt useful, since they are

easy to evaluate in real time on sorting lines. On top of that, they are determinant in

consumer's choice, since consumers at first "eat" with their eyes.

Consumers are more and more complaining about the quality of fruit and vegetables, which

are offered to them by commercial food distribution systems. The main complaints concern

the poor taste and sometimes the lack of it. However, when it comes to define precisely what

a consumer means by "taste", the answers obtained are not clear at all and generally are quite

diverging. In order to understand what the consumers mean by "taste", and to be able to fulfil

their wishes, a high attention has been paid to consumer tests. From this particular point of

view, our research project fits well with the main objectives of the COST 915-action

"Improvement of quality of fruit and vegetables, according to the needs of the consumers".

The first of the present research work was to find a correlation between the hedonic

judgement, as expressed by the consumers, and the values obtained from instrumental quality

assessments.

For the last few years, the fruit and vegetable industry was currently using automatic devices

(e.g. Pimprenelle) to evaluate quality by measuring firmness, juiciness, total acid and sugar

contents as well as pH. Unfortunately, main quality factors, such as odor and aroma are not

CHAPTER 1 2

measurable with the analytical devices available so far. For this reason, the development of an

objective and rapid method for measuring odor and/or aroma was the second main goal to be

achieved in our research work.

The thesis is structured in eight chapters. Chapter 1 is dealing with the introduction and

scope of the thesis. In Chapter 2 the recent literature on methods used to quality assessment

of fruit and vegetables is reviewed. Chapter 3 describes a new concept to measure the total

volatile compounds generated by food. The application of this new method for the

measurement of total volatile compounds generated by selected fruit varieties is described in

Chapter 4. Results obtained with strawberries are presented in Chapter 5, while in Chapter

6 the use of total volatile compounds to measure the ripening stage is discussed. Chapter 7 is

devoted to the evaluation of tomatoes and apricots. The conclusions and an outlook of the

research work are to be found in Chapter 8.

Chapters 3 and 4 have already been published, while chapters 5, 6 and 7 are in the process of

being published in peer reviewed journals. Therefore, these chapters were written as

independent papers with the consequence that overlapping, especially in the introductory

parts and in the materials & methods section, were unavoidable.

CHAPTER 2 3

CHAPTER 2

LITERATURE REVIEW ON METHODS FOR QUALITY ASSESSMENT OF FRUIT

AND VEGETABLES

Health and pleasure are two main reasons for eating fruits. However, the overall quality of

fruit is often criticized because of the fact that the organoleptic quality has been given little or

no priority by the agronomic research in favor of yield increase, disease- and storage

resistance and transportation tolerance.

The definition of fruit quality is an extremely complex matter. Quality is not only affected by

the fruit itself, but also by the consumer's own perception. As far as the fruit is regarded,

quality is generally dependent on the variety, the stage of ripeness and on the climatic

conditions that have prevailed during the growing period. The perception of quality is

consumer dependent and is affected by different factors such as age, knowledge and social

condition. In addition, each person has its own way and words to describe a product. These

multiple factors increase greatly the complexity of quality evaluation.

Fruit quality can be assessed by sensory and/or by instrumental measurements. The major

difficulty, and therefore the main challenge, is that both types of quality evaluation have to

reflect the consumer's judgement in terms of appearance, flavor and texture.

Close cooperation between agronomists and food scientists represents the key to progress in

the field of quality evaluation. The food scientist first has to set up discriminating quality

criteria that are measurable by sensory and/or instrumental methods. The analytical values

obtained have to reflect properly the quality judgement of consumers in order to be

considered as useful tools. Once this goal is achieved, the agronomists first have to implement

those specific criteria into their plant selection and breeding programs.

CHAPTER 2 4

Information obtained from consumer tests is a good starting point for setting up a quality

evaluation system. A well-trained sensory panel and/or instrumental analyses can be used to

measure different quality parameters more objectively. The data obtained from the sensory

panel and the instrumental measurements have finally to be treated by appropriated statistical

methods. These different tools are discussed hereafter.

2.1 SENSORY EVALUATION

Sensory analysis measures how food interacts with the senses (taste, smell, sight, touch and

hearing) and how people perceive the foods' properties. The basics of sensory analysis were

laid down in the USA during and after World War II, when the government wanted to

improve the quality of food provided for the army. It was recognized that even though food

was highly nutritious, it was often unpalatable [1]. Sensory methods were mainly developed

for economical reasons because they allow to set up values of acceptance for any given food.

Sensory analysis is a multidisciplinary topic. In recent years contributions originating from

different scientific fields such as psychology, physics, chemistry, neurosciences and statistics

have allowed to increase the potential of this analytical tool considerably.

Four basic methods of sensory evaluation are generally used:

- Analytical or discriminatory methods allow to determine differences between samples.

- Descriptive tests allow to determine the nature and the intensity of the attributes in a given

sample.

- Hedonic tests allow to obtain information on preference or acceptance.

- Sensitivity tests allow to determine the thresholds of a given stimulus or compound [2,3].

For product differentiation several sensory evaluation tests are used. The Triangle test is used

to determine differences between products. Three samples are presented and the panel is

asked to point out the sample, which is different. The Two-out-of-Five test is similar to the

Triangle test. The panelists are asked to point out those two out of five samples showing

similar characteristics. Multiple paired comparison tests, in which panelists are asked to taste

two samples and to rate attributes, such as saltiness, are used as well. In this case the panelists

are asked to mark the most or least salty sample. In all tests, the panelists are allowed to make

comments, thus they can sometimes better explain their choice.

CHAPTER 2 5

In a ranking test panelists are asked to rank samples according to a given sensory attribute.

Ranking samples of apples by judging the levels of crispiness is an example of such a test

[4,5].

The sensory properties of fruit include flavor (taste and aroma), texture and appearance. Taste

and aroma are considered to be more important than texture because they reflect the internal

sensory quality [6]. However, Janse [7] has shown that what people really taste with their

papilla and smell with their nose reflects only 20-40% of the taste perception; images or ideas

of a product are therefore determining the larger part of taste. It is believed that aroma plays a

more important role than taste in determining the overall quality appreciation of fruits. This is

easily demonstrated by the fact that it is difficult to identify flavor if the airflow through the

nose is restricted, e.g. simply by pinching the nostrils [8].

2.1.1 Hedonic approach for evaluation of food acceptance

The hedonic tests are playing a major role for the evaluation of food acceptance by

consumers. Generally, more than 100 persons participate in hedonic tests. They communicate

their feelings by giving statements such as like or dislike [9]. The obtained data are usually

qualitative.

Hedonic testing is popular because it can be performed with non-experienced and experienced

persons as well and with experts. However, a minimum amount of verbal ability is necessary

to obtain reliable results [10]. The samples are generally presented one by one to the subject

who is asked to decide how much he/she likes or dislikes the product. The judgement is given

by putting a mark on a scale [11]. By doing so the subject has the possibility to express its

own quality perception [12].

The hedonic scale is anchored verbally, often with nine different steps ranging from "like

extremely" to "dislike extremely". These words are placed on a graphic scale, either

horizontally or vertically, structured or not. Many different forms of the scale may be used

[13]. However, variations in the scale form are likely to cause marked changes in the

distribution of responses and ultimately in statistical parameters such as average values and

variances [14]. Hedonic ratings are converted to scores and treated by rank analysis or

analysis of variance. Hedonic scales are used for consumer tests and sometimes for sensory

panels as well [15,16]. The rating labels obtained on a hedonic scale may be affected by many

factors other than the quality of the test samples. Characteristics of the subjects, test

situations, attitudes or expectations of the subjects can markedly affect the results [17]. A

CHAPTER 2 6

researcher has to be cautious to draw conclusions on the bases of comparison of average

ratings obtained in different experiments.

2.1.2 Sensory panel

This panel is used for descriptive testing. It provides analytical data on appearance, flavor and

texture and is able to measure the quality of food. It is important to note that this type of panel

is not designed to give hedonic information. On the opposite, a consumer or hedonic test

cannot deliver analytical results.

There are three types of sensory panel [5]. The expert panel is typically composed of up to six

persons who have built up their skills (e. g. tea tasters) over a long period. The trained panel

consists of about 10-30 persons, selected for their ability to discriminate between particular

food characteristics and then carefully trained to use validated sensory methods. The trained

panel is used for descriptive testing, to provide analytical data about different parameters

relevant to the quality of food. Highly trained individuals can detect very small differences in

the characteristics of food, whereas in general the consumers cannot, or if they can, they are

not able to express precisely what they perceive. The semi-trained panel is similar to the

trained panel, except for the fact that the persons constituting the semi-trained panel are

trained but not selected for their ability to discriminate between particular food characteristics.

The semi-trained panel is used for the establishment of food profiles. Because errors in data

generated by the semi-trained panel are larger than those obtained by an expert or by a trained

panel, the semi-trained panel is usually larger. Its size depends on the purpose of the

investigation. The reliability of the results finally depends on the way the test is performed

and on the sensitivity of the panel.

2.1.2.1 Sensory attributes ofthe panel

Aroma and taste perception

It is believed that aroma is more important than taste in determining the overall appreciation

of food [8]. Volatile compounds that are perceived by the odors receptors either directly

through the nose (nasal reception) or indirectly through the pharynx during eating or drinking

(retro-nasal perception) are called "aroma compounds" [18,19].

The non-volatile compounds that are perceived by the tongue are called taste compounds

(sweet, sour, salty, bitter, astringent and pungent). The interaction between substances that

contribute to the taste of food, e.g. acids or salts is very important for the perception of aroma

CHAPTER 2 7

compounds. For instance, increasing the sugar or acid concentration in lemon-flavored

systems results in higher fruitiness intensity [20]. On the other hand, adding strawberry aroma

to a solution of sucrose induces an increase in perception of sweetness [21]. The appreciation

of food is very much depending on the synergy between taste and aroma. Furthermore, it has

to be emphasized that changes in stimuli occur when a food is ingested or masticated. This

affects the rate of release and concentration of both tastants and odorants [22].

Texture perception

Texture is another important parameter which affects shelf life of food and consumer

acceptance [23]. When a food generates a physical sensation in the mouth (hard, soft, crisp,

moist, dry), the consumer uses these sensory attributes as reference parameters for judging

food quality (fresh, stale, tender, ripe).

Evaluation of food texture by touch includes the use of fingers, lips, tongue, palate and teeth.

Texture is a general term related to a number of physical properties (e.g., viscosity and

elasticity), with complex relationship. Describing texture or mouth-feel with a single value

obtained from an instrument is questionable, since these parameters involve physical and

chemical interaction in the mouth from initial perception on the palate, through the first bite

and mastication, and finally to the act of swallowing [24].

2.2 PHYSICO-CHEMICAL METHODS

Sensory analysis is a time-consuming and expensive analytical tool. Moreover, several

researchers have observed that intensity attributes (aroma, taste and texture) reported by the

panelists change with time, which makes this approach subjective and therefore not always

reliable to obtain objective information. Compared to sensory analysis, instrumental results

are obtained in less time, at lower costs and are in general more objective and accurate.

Several instrumental methods for quality evaluation of food have been developed. Because

each method is essentially based on the measurement of a given physico-chemical property,

its effectiveness depends on the correlation between the measured physico-chemical property

and the quality factor of interest. Although researchers have pointed out some relationship

between physico-chemical properties and quality factors for a number of agricultural

CHAPTER 2 8

products, the inherent natural variability in composition, structure and other parameters within

the same batch, often makes it difficult to find good correlations.

The search for correlations between sensory and instrumental measurements has several

reasons: 1) the need for quality control instruments; 2) the desire to predict consumer

response; 3) the desire to understand what is being perceived in sensory assessments; 4) the

need to develop improved/optimized instrumental test methods; and finally, 5) to construct

testing equipment that will duplicate/replace sensory evaluation [24].

The following physical properties were measured using various instruments: density,

firmness, size, shape, color, internal defects, tissue breakdown, presence of unwanted objects,

external defects. Vibration characteristics, x-ray- and gamma-ray transmission, optical

reflectance and transmission as well as electrical conductivity were the most frequently

measured parameters. Chemical parameters such as acid, sugar and oil contents and aromatic

volatile emission are generally measured by titration, refraction index, GC and HPLC,

respectively.

2.2.1 Aroma analysis

Aroma is an important quality attribute for many food products. At present, the human nose is

still the best detector for identifying the aroma of food. Numerous researchers have tried for

many years to develop instruments such as sniffers or electronic noses, however with limited

success [25,26].

2.2.1.1 Aroma isolation techniques

The first step in the analysis of aroma compounds of a food is the isolation of the substances

that contribute to its aroma profile. The isolation methods commonly used are distillation,

extraction and adsorption/desorption. Distillation is classically used to obtain essential oils

[27]. For the isolation of volatiles combined techniques using simultaneously steam

distillation and extraction (SDE) are frequently applied (e. g. with a Likens-Nickerson

apparatus). The use of distillation methods at atmospheric pressure in order to isolate volatiles

does not only lead to the formation of thermally induced artifacts, but furthermore results in

an aroma profile that frequently does not correspond to the typical aroma impression of the

food sample [28]. For water-soluble compounds, liquid-liquid extraction is a better-suited

technique. The extraction can also be performed by cold trapping the headspace above the

CHAPTER 2 9

sample with liquid nitrogen. Adsorption on charcoal or the use of low temperature (cryo-

trapping) [29,30] and more recently the use of Tenax [31,32] are well suited techniques for

the isolation of volatiles.

Recently, extraction of food volatiles by the SPME (Solid Phase Micro-Extraction) method

[33-37] has been successfully used to obtain reliable quantitative results. The advantage of

this method lies in its simplicity and ease of manipulation and in the fact that no organic

solvent is needed. Up to now SPME has been used as a pre-concentration technique in

conjunction with subsequent chromatographic separation of the adsorbed substances [38-48].

However, it has to be pointed out that all these methods do not give any information about the

composition of the volatile compounds released from the food during eating [22].

2.2.1.2 Identification and quantification ofvolatile compounds

Quantitative determination of aroma compounds requires the use of techniques such as

GC/FID, GC/FPD, GC/MS, HPLC, etc. Sensory information can also be obtained from

instrumental analysis of the volatiles by sniffing the GC effluent. Such identification is called

GC-olfactometry [49]. Means for describing odor-active components in food include aroma

extract dilution analysis (AEDA) and calculation of odor values [50]. GC-olfactometry is the

most important technique used in aroma analysis [51-55].

Impact compounds are the aroma substances that are responsible for the specific character of

an aroma, e.g. vanillin for the aroma of vanilla. Nowadays about 7000 aroma compounds are

described in the literature. Only a small part of them originate from biosynthesis (primary

aroma compounds), most of them are developed during destruction of a cell either by

enzymatic reactions or by fermentative or thermal processes (secondary aroma compounds).

The aroma compounds belong to many different classes of substances such as alcohols,

aldehydes, ketones, esters, lactones, sulfides, and heterocyclic components (furanes,

pyrazines, thiazoles and thiophene, etc.). The precursors of these substances are amino acids,

fatty acids, sugars and isoprenoids, respectively [56]. The aroma of a food may consist of up

to several hundreds of substances, but only a small part of them may contribute significantly

to the sensory properties of a food. Attempts to correlate data from volatile analysis with

sensory perception have not been totally successful [57].

Recently, a retro-nasal aroma simulator system has been described [58] in which food is

sheared (to simulate mastication), hydrated with water or saliva, and the released volatiles

removed by a gas stream. Van Ruth et al. [59,60] have developed a plunger system that

CHAPTER 2 10

imparts shear to food by screw action and add artificial saliva to the food sample to simulate

hydration. Other working groups [61,62] hydrated samples to various extents and measured

the release of volatiles as a function of water content.

Various mass spectrometric methods have been proposed for breath analysis. However, some

are not sensitive enough, making them unsuitable for the analysis of volatile compounds

released during eating [63,64].

Researchers have tried for many years with limited success to develop electronic sniffers or

electronic noses [65]. Most of the electronic noses use an array of sensors, each of which is

sensitive to the concentration of one or more compounds present in the gas phase. The outputs

of the sensors are analyzed using a pattern-recognition procedure, e.g., principal-component

analysis, discriminate function analysis, or neural network. Commonly used sensors are

sintered metal oxides, conducting polymers and quartz resonators. The electronic noses have

been used to classify the flavor of various beverages or foodstuffs, such as coffee beans,

whiskeys, beers, fish and meat [66-69].

There has been an increased interest in the development of sensors for aromatic volatile

compounds to determine fruit quality. Benady et al. [70] developed a sniffer for the

determination of fruit ripeness in a non-destructive way. The sniffer uses a semiconducting

gas sensor located within a small cup to collect and measure the gases emitted by the fruit.

The authors reported that the sensor performed successfully on three muskmelon cultivars

under field and laboratory conditions, and this for two growing seasons.

2.2.2 Sugar and acid content

By measuring the total sugar content (°Brix) and total acidity of a fruit, a primary quality

characteristic can be assessed. Dull and coworkers [71,72] used NIR to determine soluble

solids in cantaloupe and honeydew melons. Kawano et al. [73] used NIR with optical fibers to

analyze the sugar content of intact peaches and found a good correlation (r = 0.97) between

NIR measurements and Brix values. Similarly, Slaughter [74] successfully used the

absorption characteristics of NIR in peaches and nectarines to predict their soluble solids

content (r = 0.92). Bellon and Sevila [75] developed an NIR system which combined

spectrophotometry and optical fibers for the determination of soluble solids in apples, whereas

acidity was measured by pH and titration with a basic solution.

CHAPTER 2 11

2.2.3 Texture analysis

Instrumental methods used to measure texture are based on the deformation and flow of

materials. The challenge consists in correlating mouth-feel with the measurable forces or the

shapes of rheological curves. In fruit quality assessment, one of the goals to be achieved is to

sort out fruits of poor quality so that they do not appear on the market.

Firmness is an important textural attribute for most foods. Many devices have been developed

for measuring firmness, including destructive tests such as puncture and compression tests

[76-78]. Non-destructive tests have also been proposed, such as deflection test, vibration

frequency measurement and ultrasonic evaluation [79,80]. Generally, firmness can be used as

a criterion for sorting out agricultural products such as e.g. fruits into different maturity

groups or for separating overripe and damaged plant tissues from intact ones. Overripe and

damaged fruits become soft due to enzymatic reactions. To determine the firmness of fruit,

different types of penetrometers are used such as a hand-held fruit-firmness tester called

Kiwifirm [81], the shear cell of Kramer [82] and the Durofel [83]. The last two equipments

are at present widely used by the fruit industry. Fruit firmness has also been evaluated by non¬

destructive methods such as force deformation. Low-pressure air, simultaneously applied to

small areas on opposite sides of peaches, was used by Perry and Perkins [84] to generate a

non-bruising maturity-indicating deformation. Delwiche et al. [85] developed a "deformater"

device for maturity detection of pears based on the measurement of deformation resulting

from pressing two steel balls with a fixed force against the opposite sides of the fruit. Mizrach

et al. [86] used a pin (3-mm 0) as a mechanical thumb to sense firmness of oranges and

tomatoes. Other force-deformation types of firmness testers were developed to assess

hardness and immaturity [87]. Bellon et al. [88] built a micro-deformater that was able to

classify peaches into three classes of firmness with 92% accuracy. Armstrong et al. [89]

developed a device which could be used to determine the firmness of small fruit samples. The

instrument used the force-deflection measurement of a whole fruit between two parallel

plates. It includes automatic data collection and can measure the firmness of a batch of 25

fruit samples within one minute.

Attempts have been made by several working groups to record the electric activity of the

facial muscles during eating, and to compare these measurements with sensory evaluation of

texture [90-92]. Studies on a variety of food have shown that sensory evaluated texture

CHAPTER 2 12

attributes correlate well with particular aspects of the chewing cycle, but up to now this

methodology has only found a limited application in the food industry.

The use of various image acquisition techniques, such as solid-state TV camera [93], line-scan

camera, impact forces [94], x-ray scanning [95], gamma-rays [96], ultrasonic scanning [97-

100], electrical properties [101] and NMR imaging [102], in conjunction with image-

processing techniques have provided new opportunities for researchers to develop many new

and improved techniques for non-destructive quality evaluation of food and agricultural

products.

2.3 STATISTICAL METHODS

2.3.1 Statistical tests for comparing samples

Several useful statistical tests for comparing samples are available. As shown in Table 2.1

different tests are used, depending on the number of samples and the questions to be

answered. Independent tests are carried out with the same panel members, whereas in paired

tests different panel members are involved. The distinction between parametric tests and non-

parametric tests is made based on the normal distribution of the Gaussian curve

(passed/failed) [103-105].

Table 2.1 Statistical tests used for comparison between samples

2 samples > 3 samples

independent paired independent paired

parametricStudent Student ANOVA* MANOVA**

test

non-parametric Mann.....

Kruskall„ . ,

t p wu-*Wilcoxon

p„. ...

Friedman

test & Whitney & Wallis

*ANOVA (one way); **MANOVA (Multiple ANOVA)

2.3.2 Multivariate analysis

The comparison between samples and correlation between instrumental and sensory data

remains difficult. Indeed, the evaluation of the results must take into account the variability of

the data and also the difficulty to define the products well. Parameters which influence each

other have to be considered. The multivariate statistical approach is helpful to resolve such

problems. Several software packages are available to treat statistically the results, e.g. Statbox

CHAPTER 2 13

(Grimmer Logiciels® 1997, Paris, France) and Statview® (Abacus Concepts 1998, Berkeley,

USA).

Descriptive methods often constitute the first stage of multivariate analysis. The visual study

is still possible with N = 2, 3 or 4 (N corresponds to the number of the initial variables) but

not when N >4. One very useful method is the PCA (Principal Component Analysis), which

replaces the initial variables by non-correlated variables (the Principal Component) and then

allows to consider only a few principal components without loosing too much information

[106].

2.4 REFERENCES

[I] Marshall, R., Kilcast, D.A. (2000). A matter of taste. Chem. & Ind. 210-213.

[2] Meilgaard, D., Civille, G.V., Carr, B.T. (1991). Sensory evaluation techniques. CRC

Press, Corp., Boca Raton.

[3] Reineccius, G., (1994). Source book of flavors. Chapman & Hall, corp. Publishers, New

York.

[4] Jellinek, G. (1985). Sensory evaluation of food: theory and practice. VCH Publishers,

Deerfield.

[5] Laming, D. (1986). Sensory analysis. Academic Press, London.

[6] Schutz, H.G., Wahl, O.L. (1981). Consumer perception of the relative importance of

appearance, flavor and texture to food acceptance. In: Solms, J., Hall, R.L. (eds.).

Criteria of food acceptance. Forster-Verlag, Zurich, 97-116.

[7] Janse, J. (1997). Aspects of flavour of tomatoes. Geisenheimer Tagung 21, 1-5.

[8] Taylor, A.J., Linforth, R.S.T. (1996). Flavor research in the mouth. Trends Food Sei.

Technol. 7, 444-447.

[9] Peryam, D.R., Pilgrim, F.J. (1957). Hedonic scale method of measuring food

preferences. Food Technol. 11, 9-14.

[10] Omahony, M. (1986). Sensory evaluation of food. Marcel Dekker, Inc. Publishers, New

York.

[II] Meiselman, H.L. (1994). A measurement scheme for developing institutional products.

In: MacFie, H.J.H., Thomson, D.M.H. (eds.). Measurement of food preferences. Blackie

Academic & Professional Publishers, London.

[12] Kelly, G.A. (1955). The psychology of personal constructs. Norton Publishers, New

York.

CHAPTER 2 14

[13

[14

[15

[16

[17

[18

[19

[20

[21

[22

[23

[24

[25

[26

[27

[28

Shepard, R.N. (1972). Multidimensional scaling: theory and applications in the

behavioral sciences. Seminar Press, New York.

Piggott, J.R. (1986). Statistical procedures in food research. Elsevier Applied Science

Publishers, Amsterdam.

Amerine, M.A., Pangborn, R.M., Roessler, E.B. (1965). Principles of sensory

evaluation of food. Academic Press, New York.

Kramer, A., Twigg, B.A. (1962). Practical food microbiology and technology. AVI

Publishing Company, Westport.

Hall, R.L. (1981). Closing remarks. In: Solms, J., Hall, R.L. (eds.). Criteria of food

acceptance. Forster-Verlag, Zurich, 449-456.

Pierce, J., Halpern, B.P. (1996). Orthonasal and retronasal identification based upon

vapor phase input from common substances. Chem. Sens. 21, 529-543.

Halpern, B.P. (1977). Psychophysics of Taste. In: Beauchamp, G.K., Bartoshuk, L.M.

(eds.). Tasting and smelling. Handbook of perception and cognition. 2nd Edition.

Academic Press, San Diego, 77-123.

Noble, A.C. (1996). Taste-Aroma Interactions. Trends Food Sei. Technol. 7, 439-443.

Frank, R.A., Ducheney, K., Mize, S.J.S. (1989). Strawberry odor, but not red color,

enhances the sweetness of sucrose solutions. Chem. Sens. 14, 371-377.

Ingham, K.E., Linforth, R.S.T., Taylor, A. J. (1995). The effect of eating on aroma

release from strawberries. J. Food Chem. 54, 283-288.

Pomeranz, Y., Meloan, CE. (1994). Food analysis: Theory and practice. Chapman &

Hall Corp., New York.

Szczesniak, A.S. (1987). Correlating sensory with instrumental texture measurements.

J. Text. Studies, 18, 1-15.

Van Geloven, P., Honore, M., Roggen, J., Leppavuori, S., Rantala, T. (1991). The

influence of relative humidity on the response of tin oxide gas sensors to carbon

monoxide. Sens. Actuators B4, 185-188.

Jonda, S., Fleischer, M., Meixner, H. (1996). Temperature control of semiconductor

metal-oxide gas sensors by means of fuzzy logic. Sens. Actuators B34, 396-400.

Parliment, T.H. (1999). Solvent extraction and distillation techniques. In: Marsili, R.

(ed.). Techniques for analyzing food aroma. Dekker Corp., New York, 1-26.

Fischer, N., Hammerschmidt, F.J. (1992). A contribution to the analysis of fresh

strawberry flavour. Chem. Mikrobiol. Technol. Lebensm. 14, 141-148.

CHAPTER 2 15

[29] Linforth, R.S.T., Taylor, A.J. (1993). Measurement of volatile release in the mouth.

Food Chem. 48, 115-120.

[30] Nassl, K., Kropf, F., Klostermeyer, H.Z. (1995). A method to mimic and to study the

release of flavour compounds from chewed food. Z. Lebensm. Unters. Forsch. 201, 62-

68.

[31] Taylor, A.J., Linforth, R.S.T. (1994). Methodology for measuring volatile profiles in the

mouth and nose during eating. In: Maarse, H., van der Heij, G. (eds.). Trends in flavor

research. Elsevier Science Publishers, Amsterdam, 3-14.

[32] Delahunty, CM., Piggott, J.R., Connor, J.M., Paterson, A. (1994). Low fat Cheddar

cheese flavor: flavor release in the mouth. In: Maarse, H., van der Heij, G. (eds.).

Trends in flavor research. Elsevier Science Publishers, Amsterdam, 47-52.

[33] Paliyath, G., Whiting, M.D., Stasiak, M.R., Murr, D.P., Clegg, B.S. (1997). Volatile

production and fruit quality during development of superficial scald in Red Delicious

apples. Food Res. Int. 30, 95-103.

[34] Song, J., Gardner, B.D., Holland, J.F., Beaudry, R.M. (1997). Rapid analysis of volatile

flavor compounds in apple fruit using SPME and GC/time-of-flight mass spectrometry.

J. Agric. Food Chem. 45, 1801-1807.

[35] Song, J., Fan, L.H., Beaudry, R.M. (1998). Application of solid phase microextraction

and gas chromatography time-of-flight mass spectrometry for rapid analysis of flavor

volatiles in tomato and strawberry fruits J. Agric. Food Chem. 46, 3721-3726.

[36] Ibanez, E., Lopez, S.S., Ramos, E., Tabera, J., Reglero, G. (1998). Analysis of volatile

fruit components by headspace solid-phase microextraction. Food Chem. 63, 281-286.

[37] Wan, X.M., Stevenson, R.J., Chen, X.D., Melton, L.D. (1999). Application of

headspace solid-phase microextraction to volatile flavour profile development during

storage and ripening of kiwifruit. Food Res. Int. 32, 175-183

[38] Arthur, C.L., Pawliszyn, J. (1990). Solid phase microextraction with thermal desorption

using fused silica optical fibers. Anal. Chem. 62, 2145-2148.

[39] Pawliszyn, J. (1997). Solid phase microextraction, Wiley-VCH. Corp., New York.

[40] Pawliszyn, J. (1999). Application of solid phase microextraction. Royal Society of

Chemistry, Hertfordshire, UK.

[41] Arthur, C, Pratt, K., Belardi, R., Motlagh, S., Pawlisczyn, J. (1992). Environmental

analysis of organic compounds in water using solid phase microextraction. J. High Res.

Chromatogr. 15, 741-744.

CHAPTER 2 16

[42] Zhang, Z., Pawliszyn, J. (1996). Sampling volatile organic compounds using a modified

solid phase microextraction device. J. High Res. Chromatogr. 19, 155-160.

[43] Valor, I., Cortada, C, Molto, J.C. (1996). Direct solid phase microextraction for the

determination of BTEX in water and wastewater. J. High Res. Chromatogr. 19, 472-

474.

[44] Santos, F.J., Sarrion, M.N., Galceran, M.T. (1997). Analysis of chlorobenzenes in soils

by headspace solid-phase microextraction and gas chromatography-ion trap mass

spectrometry. J. Chromatogr. Ill, 181-189.

[45] Steffen, A., Pawliszyn, J. (1996). Analysis of flavor volatile using headspace solid-

phase microextraction. J. Agric. Food Chem. 44, 2187-2193.

[46] Matich, A.J., Rowan D.D., Banks, N.H. (1996). Solid phase microextraction for

quantitative headspace sampling of apple volatiles. Anal. Chem. 68, 4114-4118.

[47] Field, J.A., Nickerson, G., James, D.D., Heider, C. (1996). Determination of essential

oils in hops by headspace solid- phase microextraction. J. Agric. Food Chem. 44, 1768-

1772.

[48] Yang, X.G., Peppard, T. (1994). Solid phase microextraction for flavor analysis. J.

Agric. Food Chem. 42, 1925-1930.

[49] Acree, T.E., Barnard, J. (1994). Gas chromatography-olfactometry and charm analysis.

In: Maarse, H., van der Heij, G. (eds.). Trends in flavor research. Elsevier Science

Publishers, Amsterdam, 211-220.

[50] Etievant, P.X., Moio, L., Guichard, E., Langlois, D., Leschaeve, I., Schlich, P. (1994).

Aroma extract dilution analysis (AEDA) and the representativiness of the odour of food

extracts. In: Maarse, H., van der Heij, G. (eds.). Trends in flavor research. Elsevier

Science Publishers, Amsterdam, 179-190.

[51] Acree, T.E. (1997). GC/olfactometry. Anal. Chem. News & Features 170A-175A.

[52] Egolf, L.M., Jurs, P.C. (1993). Quantitative structure-retention and structure-odor

intensity, relationships for a diverse group of odor-active compounds. Anal. Chem. 65,

3119-3126.

[53] Rossiter, K.J. (1996). Structure-odor relationships. Chem. Rev. 96, 3201-3240.

[54] Ott, A., Fay, L.B., Chaintreau, A. (1997). Determination and origin of the aroma impact

compounds of yogurt flavor. J. Agric. Food Chem. 45, 850-858.

[55] Abott, N., Etievant, P.X., Issanchou, S., Langlois, D. (1993). Critical evaluation of two

commonly used techniques for the treatment of data from extract dilution sniffing

analysis. J. Agric. Food Chem. 41, 1698-1703.

CHAPTER 2 17

[56] Fugmann, H.B. (1997). Römpp Lexikon Naturstoffe. Thieme Verlag Publishers.

Stuttgart.

[57] Cliff, M., Heymann, H. (1993). Development and use a time-intensity methodology for

sensory evaluation. Food Technol. 26, 375-385.

[58] Roberts, D.D., Acree, T.E. (1995). Simulation of retronasal aroma using a modified

headspace technique: Investigation the effects of saliva, temperature, shearing and oil or

flavor release. J. Agric. Food Chem. 43, 2179-2186.

[59] Van Ruth, S.M., Roozen, J.P., Cozijnsen, J.L. (1995). Volatile compounds of rehydrated

french beans, bell peppers and leeks. Part I. Flavor release in the mouthand in three

mouth model systems. Food Chem. 53, 15-22.

[60] Van Ruth, S.M., Roozen, J.P., Cozijnsen, J.L. (1995). Volatile compounds of rehydrated

french beans, bell peppers and leeks. Part II. Gas chromatography/sniffing port analysis

and sensory evaluation. Food Chem. 54, 1-8.

[61] Dalla Rosa, M., Pittia, P., Nicoli, M.C. (1994). Influence of water activity on headspace

concentration of volatile over food and model systems. Ital. J. Food Sei. 4, 421-432.

[62] Clawson, A.R., Linforth, R.S.T., Ingham, K.E., Taylor, A. (1996). Effect of hydration

on release of volatile from cereal foods. Lebensm.-Wiss.-Technol. 29, 158-162.

[63] Jordan, A., Hansel, A., Holzinger, R., Lindinger, W. (1995). Acetonitrile and benzene in

the breath of smokers and non-smokers investigated by proton transfer reaction mass

spectrometry. Int. J. Mass Spectrom. Ion. Process 148, L1-L3.

[64] Benoit, F.M., Davidson, W.R., Lovett, A.M., Nacson, S., Ngo, A. (1983). Breath

analysis by atmospheric mass spectrometry. Anal. Chem. 55, 805-807.

[65] Gardner, J.W., Bartlett, P.N. (1994). A brief history of electronic noses. Sensors and

Actuators B. 18, 211-220.

[66] Schaller, E., Bosset, J.O., Escher, F. (1998). Electronic noses and their application to

food. Lebensm.-Wiss.-Technol. 31, 305-316.

[67] Gardner, J.W., Shurmer, H.V., Tan, T.T. (1992). Sensors and Sensory systems for an

electronic nose. Sensors & Actuators B. 6, 71-75.

[68] Di Natale, C, Macagnano, A., Davide, F., D'Amico, A., Paolesse, R., Boschi, T.,

Faccio, M., Ferri, G. (1997). An electronic nose for food analysis. Sensors & Actuators

5.44,521-526.

[69] Gardner, J.W., Bartlett, P.N. (1992). Sensors and sensory systems for an electronic

nose. Kluwer Academic Publishers, London.

CHAPTER 2 18

[70] Benady, M., Simon, J.E., Charles, D.J., Miles, G.E. (1995). Fruit ripeness determination

by electronic sensing of aromatic volatiles. Transactions ofthe ASAE 38, 251-257.

[71] Dull, G.G., Birth, G.S. (1989). Nondestructive evaluation of fruit quality: use of near

infrared spectrophotometry to measure soluble solids in intact honeydew melons. Hort.

Science 24, 754.

[72] Dull, G.G., Leffler, R.G., Birth, G.S., Smittle, D.A. (1992). Instrument for

nondestructive measurement of soluble solids in honeydew melons. Transactions ofthe

ASAE 35, 735-737.

[73] Kawano, S., Watanabe, H., Iwamoto, M. (1992). Determination of sugar content in

intact peaches by near infrared spectroscopy with fiber optics in interactance mode. J.

Japan. Soc. Hort. Sei. 61, 445-451.

[74] Slaughter, D.C. (1995). Nondestructive determination of internal quality in peaches and

nectarines. Transactions of the ASAE 38, 617-623.

[75] Bellon, B., Sevila, F. (1993). Optimization of a non-destructive system for on-line infra¬

red measurement of fruit internal quality. International symposium on fruit, nut, and

vegetable production engineering. Valencia-Zaragoza, 317-325.

[76] Mohsenin, N.N. (1986). Physical properties of plant and animal materials. Gordon and

Breach Science publishers, New York.

[77] Lee, C.H., Rha, C. (1979). Rheological properties of proteins in solution. In: Sherman,

P. (ed.). Food texture and rheology. Academic Press, London, 245-263.

[78] Hermansson, A.M. (1979). Aspects of protein structure, rheology and texturization. In:

Sherman, P. (ed.). Food texture and rheology. Academic Press, London, 265-282.

[79] Tollner, E.W., Brecht, J.K., Upchurch, B.L. (1993). Nondestructive evaluation:

Detection of damage. In: Shewfelt R.L., Prussia, S.E. (eds.). Postharvest Handling: A

Systems Approach. Florida Academic Press, Orlando, 225-255.

[80] Abott, J.A., Bachman, G.S., Childers, N.F., Fitzgerald, J.V., Matuski, F.J. (1968). Sonic

techniques for measuring texture of fruits and vegetables. Food Technol. 22, 101-112.

[81] Sharrah, N., Kunze, M.S., Pangborn, R.M. (1965). Beef tenderness: comparison of

sensory methods with the Warner-Bratzler and L.E.E.-Kramer shear presses. Food

Technol. 19, 136-143.

[82] Szczesniak, A.S. (1971). Effect of mode of rehydration on textural parameters of

precooked freeze-dried sliced beef. J. Texture Studies 2, 18-30.

[83] www.copa-technologie.com

CHAPTER 2 19

[84] Perry, R.L., Perkins, R.M. (1968). Separators for frost damaged oranges. California

Citrograph 53, 304-305, 307-308, 310, 312.

[85] Delwiche, M.J., Tang, S., Mehlschau, J.J. (1989). An impact force response fruit

firmness sorter. Transactions of the ASAE 32, 321-326.

[86] Mizrach, A., Nahir, D., Ronen, B. (1992). Mechanical thumb sensor for fruit and

vegetable sorting. Transactions ofthe ASAE 35, 247-250.

[87] Takao, H., Ohmori, S. (1994). Development of device for nondestructive evaluation of

fruit firmness. Japan Agric. Res. Quarterly 28, 36-43.

[88] Bellon, V., Vigneau, J.L., Crochon, M. (1994). Nondestructive sensing of peach

firmness. International symposium on fruit, nut and vegetable production engineering.

Valencia-Zaragoza, 291-297.

[89] Armstrong, P.R., Brown, G.K., Timm, E.J. (1995). Non-destructive firmness

measurement of soft fruit for comparative studies and quality control. ASAE Paper No.

95-6172. St. Joseph.

[90] Brown, W.E., Dauchel, C, Wakeling, I.J. (1996). Non-destructive firmness

measurement of soft fruit for comparative studies and quality control. J. Texture Studies

27, 433-435.

[91] Brown, W.E., Eves, D., Ellison, M., Braxton, D.J. (1998). Use of combined

electromyography and kinesthesiology during mastication to chart the oral breakdown

of foodstuffs: Relevance to measurement of food texture. J. Texture Studies 29, 145-

167.

[92] Tuorila-Ollikainen, H., Alanko, T., Hirvi, T. (1983). Facial representation of sensory

profiling data. Lebensm.-Wiss.-Technol. 16, 376-377.

[93] Park, B., Chen, Y.R. (1996). Multispectral image co-occurrence matrix analysis for

poultry carcasses inspection. Transactions of the ASAE 39, 1485-1491.

[94] Nahir, D., Schmilovitch, Z., Ronen, B. (1986). Tomato grading by impact force

response. ASAE Paper. No. 86-3028. St. Joseph.

[95] Morita, K., Tanaka, S., Ogawa, Y., Thai, C.N. (1996). .Detection of non-metallic

foreign materials in food by soft X-ray system with CdTe sensor. ASAE Paper No. 96-

6059. St. Joseph.

[96] Garrett, R.E, Talley, W.K. (1970). Use of gamma ray transmission in selecting lettuce

for harvest. Transactions ofthe ASAE. 13, 820-823.

CHAPTER 2 20

[97] Abbott, J.A., Bachman, G.S., Childers, N.F., Fitzgerald, J.V., Matuski, F.J. (1968).

Sonic techniques for measuring texture of fruits and vegetables. Food Technol. 22,

101-112.

[98] Shmulevich, I., Galili, N., Rosenfeld, D. (1996). Detection of fruit firmness by

frequency analysis. Transactions ofthe ASAE 39, 1047-1055.

[99] Galili, N., Rosenhouse, G., Mizrach, A. (1994). Ultrasonic technique for fruit and

vegetable quality evaluation. International symposium on fruit, nut, and vegetable

Production Engineering., Valencia-Zaragoza, 281-289.

[100]Haugh, CG. (1994). Detecting hollow hearts in potatoes using non-invasive acoustic

techniques. Int. Agrophysics 8, 509-518.

[101] Nelson, S.O., Forbus, W.R., Lawrence, K.C. (1995). Assessment of microwave

permittivity for sensing peach maturity. Transactions ofthe ASAE 38, 579-585.

[102] Ray, J.A., Stroshine, R.L, Krutz, G.W., Wai, W.K. (1993). Quality sorting of sweet

cherries using magnetic resonance. ASAE Paper. No. 93-6071. St. Joseph.

[103]Danzart, M. (1998). Comparaison de produits. In: Depledt, F. (ed.). Evaluation

sensorielle, Manuel méthodologique. Lavoisier, Tec & Doc, Paris, 259-272.

[104] Lea, P., Robotten, M., Naes, T. (1997). Analysis of variance for sensory data. John

Willey & Son Ltd., New York.

[105] Moskowitz, H. (1988). Applied sensory analysis of foods. CRC Press, Boca Raton.

[106] Danzart, M. (1998). Cartographic In: Depledt, F. (ed.). Evaluation sensorielle, Manuel

méthodologique. Lavoisier, Tec & Doc, Paris, 290-296.

CHAPTER 3 21

CHAPTER 3

A NEW CONCEPT FOR THE MEASUREMENT OF TOTAL VOLATILE

COMPOUNDS OF FOOD*

ABSTRACT

The aim of our work was the development of a rapid and reliable methodology for the

evaluation of the total volatile fraction of fruits (strawberries, raspberries, tomatoes and

apples). Our method consists in trapping the volatile compounds of fruits on a solid phase

micro-extraction fiber (SPME) and determining the total amount of the adsorbed substances

after desorption in a GC system, without performing any separation. The patterns obtained by

using several SPME fiber types permitted us to differentiate the total volatile compounds

present in the sample upon their chemical nature.

Using strawberries as a model, we could show that our method 1) leads to easily reproducible

results, 2) allows for the differentiation between six varieties in a way which is consistent

with a hedonic evaluation of these varieties, 3) shows the variation in total volatile

compounds between individual fruits. The technique is rapid, practical, cheap, and promising

for an objective evaluation of the volatile fraction of fruits.

3.1 INTRODUCTION

The search for a reliable method to easily assess the aroma of food is a challenge for the

analytical chemist, but also a necessity for the food producer, who needs an extensive and

well-trained sensory panel to judge the quality of his/her products. On the one hand,

chromatographic methods have developed tremendously since the introduction of capillary

columns, but they are time-consuming, require highly trained personnel and the results are

* Azodanlou, R., Darbellay, C, Luisier, J. L., Villettaz, J. C, Amadö, R. (1999). A new

concept for the measurement of total volatile compounds of food. Z Lebensm. Unters. Forsch

A. 208, 254-258.

CHAPTER 3 22

often difficult to interpret. On the other hand, "the electronic nose" is a new tool which

requires an intensive research; at present, information about this method is scarce.

The measurement, the differentiation, and the identification of aroma compounds requires the

utilization of techniques such as GC, GC-MS, HPLC, or sensorimetry (e.g. olfactometry,

sensory analysis by a panel). The electronic nose is an interesting technique under

development which uses gas sensors such as metallic oxides, polymers and piezoelectrical

quartz as bases for the determination. The main limitation to the use of the electronic nose is

its high sensitivity to compounds such as ethanol, carbon dioxide, and to humidity, which

masks the desired responses. An additional problem is the poor selectivity of the electronic

nose, which needs the combination of multiple sensor arrays to increase its accuracy (4-32)

[1,2].

In flavor analysis, sample preparations usually involve concentrating the analytes of interest

using headspace, purge and trap, liquid-liquid extraction, solid phase extraction, or

simultaneous distillation/extraction techniques. These methods have various drawbacks,

including long preparation times and the use of organic solvents.

The newly developed "solid-phase micro extraction" (SPME) technique eliminates most of

these drawbacks [3]. SPME [4] is a headspace sampling technique which concentrates

analytes by adsorption onto a polymer-coated silica fiber prior to thermal desorption in the

injection port of a gas Chromatograph. SPME was originally developed for sampling organic

contaminants (chlorinated hydrocarbons) in water by direct immersion of the fiber into the

sample [5], but has more recently been applied to headspace sampling of solid and liquid

samples [6-11]. The applicability of SPME-GC in food analysis is well developed [12-15].

In the context of our research work on the evaluation of the quality of fruits, we were looking

for a rapid method to measure aromas. By using SPME, we tried an oversimplification of

chromatography which led us to a new method: adsorption by SPME and direct desorption of

the adsorbed substances into a chromatographic detector, without any separation. The signal

obtained depends on the quantity of the adsorbed substances, and its shape is influenced by

the chemical properties of the volatile compounds. By using several SPME fiber, more than

one signal is obtained, leading to a differentiation of the volatile substances present in the

sample.

CHAPTER 3 23

The novelty of the method consists in trapping the food volatile fraction on SPME fibers and

determining directly the total amount of adsorbed substances by a GC detector. The first

results showed that our method gives easily reproducible results and allows differentiation

between varieties, as well as the measurement of the variation in volatile compounds between

individual fruits. The measurement time is short and permits the analysis of significant

number of samples. The applicability of our method to other foods was also tested.

3.2 MATERIALS AND METHODS

3.2.1 Analytical Procedure

SPME was performed, as is usual for GC separations. The volatile compounds were collected

by inserting a needle through a Teflon-coated silicone septum into the headspace of the

sample flask. After a defined sampling time (see below) the adsorbed substances were

desorbed in the injection port of a GC (Carlo Erba HRGC 5300, Carlo Erba S.p.A., Milano,

Italy).

The splitless injector port was coupled directly to the flame ionization detector using a

transfer tubing (20 cm in length, 0.1 mm i.d., N°160-2630, J&W, New Brighton, USA), with

a helium pressure of 150 kPa, at a flow rate of approximately 5 mL/min. The oven

temperature was set at 250°C. The injection port and detector temperatures were 200 and

250°C, respectively.

For quantification, the area under the peak was measured by a GC integration program

(Chrom Card; Fisons Inst., Milan, Italy) in mVmin or by Borwin software (JMBS

Développements, Grenoble, France) in jiV-min. The total analysis time was approximately 40

min, including 30 min sampling time. The sampling time can easily be reduced to 2 min, thus

permitting rapid analysis. Each sample was analyzed in triplicate.

3.2.2 SPME fiber types

The following types of SPME fiber were used: poly(dimethylsiloxane) (PDMS) with different

thicknesses (100, 30, 7um nos. 5-7300-U, 5-7308, 5-7302); polyacrylate (85um no. 5-7304);

porous fibers [65um Carbowax/divinylbenzene (CW/DVB); no. 5-7312]; bi-polar fibers

(65(im PDMS/DVB no. 5-7310-U) and the new Carboxen-PDMS 75um fiber (no. 5-7318), all

available from Supelco Co., Bellefonte, USA.

CHAPTER 3 24

3.2.3 Samples

Six varieties of strawberries (Evita, Don, Ciref 777, Mara des bois, Seascape, and Tango),

from the same field and with the same degree of maturity were used. The fruits were obtained

from the Swiss Federal Research Station for Plant Production (Conthey, Switzerland). Ground

pepper and clove samples were purchased from local food stores [Coop (C) and Migros (M)],

and at Pakoba (P) in the case of Sawarak Pepper. Coffee samples, two Arabica (Colombia)

and two Robusta (Togo) varieties of two different roasting grades (CTN 90 and 110) were

supplied by Nestlé (N; R&D Center, Orbe, Switzerland). The coffee beans were ground in a

household mill for 30 s at 20 000 rpm to yield a homogeneous coffee powder (particle size 0.3

- 0.4 mm).

3.2.4 Sample preparation and extraction of the volatile compounds

Ca. 200 g of strawberries (Evita variety) were placed in a tightly closed IL round-necked

flask, fitted with a Teflon-coated silicone septum, and maintained at 25°C in a water bath for

30 min to allow for headspace equilibrium. SPME was performed with a PDMS, 100 urn

fiber. A sampling time of exactly 5 min from the headspace of the flask was chosen.

For the measurement of the heterogeneity of strawberries, individual fruits (Tango cv.) were

used. Fruits of approximately 10 g were placed in a 150-mL round-necked flask with a wide

opening (NS/45), and fitted with a Teflon-coated silicone septum, at 25 °C. The time allowed

for sample equilibration in the flask was 15 min, and the sampling time with SPME was 5

min.

For all other experiments, 0.5 g of the food sample was placed in a 100 mL round-necked

flask fitted with a Teflon coated silicone septum. The time to reach headspace equilibrium

was 30 min and the SPME sampling time was 5 min; no regulation of the temperature was

made. Two types of SPME fiber were used: PDMS 100 um and CW/DVB 65 urn.

3.3 RESULTS AND DISCUSSION

3.3.1 Optimization of the analytical system

In order to show the suitability of the system for the quantification of volatile compounds of

food, we first had to optimize the experimental conditions and to investigate the

reproducibility of the system (Table 3.1).

CHAPTER 3 25

Table 3.1 Reproducibility of the measurements. CFCoefficient of variation.

No. of sample peak area (mV.min)1 26859.4

2 27043.5

3 26731.4

4 28503.2

5 27595.5

6 28051.5

7 27051.3

8 29753.6

9 27016.5

10 27060.7

average 27566.7

SD 954.0

CV 3.46 %

We chose a sampling temperature of 25 °C, slightly higher than room temperature, in order to

minimize biodégradation of the samples by enzymatic reactions. These conditions were close

to those at which these products are normally consumed. Rapid and total desorption of the

adsorbed substances from SPME fibers is essential for quantification purposes. As shown in

Fig. 3.1, desorption from the fiber was rapid, the peak maximum for the total volatile fraction

was usually detected after less than 0.5 min and it took less than 2 min to regain the baseline

value. This allowed us to set the desorption time at 2 min.

3000t

0 1 2

time (min)

Fig. 3.1 Signal obtained by direct desorption of compounds adsorbed by solid-phasemicroextraction (SPME) fibers

The time during which the SPME fiber is exposed to the headspace of the sample ("sampling

time") was determined by exposing the fiber for various times to the headspace of the sample.

CHAPTER 3 26

Strawberry volatile compounds reached an almost steady-state concentration within 5 min.

The equilibration time represents the time needed for saturation of the headspace.

Equilibration was usually reached within 30 min (Fig. 3.2).

3.0E+05

~ 2.0E+051

1>

ro 1.0E+05cd

as

co0a.

0.0E+00

0 20 40 60

equilibration time (min)

Fig. 3.2 Effect of sampling time on peak area of the total volatile fraction of strawberries

With optimized experimental conditions, measurements of the total volatile fraction gave a

coefficient of variation of about 5%. The amount of extracted volatile compounds depended

linearly on sample weight in a certain weight range. By increasing the sample weight,

adsorption on all tested SPME fibers increased rapidly up to a certain level and then remained

almost constant (Fig. 3.3).

3.0E+05

E

> 2.0E+05

co

Ëco

.*:

coCDQ.

1.0E+05

0.0E+00

50 50 50 75 75 75 100 100 100 125 125 125 150 150 150

weight (g)

Fig. 3.3 Effect of sample weight on the signal of the total volatile fraction of strawberries

CHAPTER 3 27

Each type of fiber adsorbed different amounts of total volatile compounds, thus yielding a

representative "fingerprint" of each sample (Fig. 3.4). It is well known that for SPME, the

area of the measured signal and the selectivity ofthe adsorbed volatile compounds depend on

the coating phase. These SPME properties are based on polarity, thickness and porosity. The

use of several SPME fibers permits to achieve different profiles, which, contrary to profiles

obtained with the electronic nose, are not influenced by humidity and carbon dioxide.

PDMS 100 urn

3.00E+06,.

CAR/DVB 65 urn ,c2.00E+06--

CAR/PDMS 75 urn

PDMS/DVB 65 urn

PDMS 30 urn

PDMS 7 um

PA 85 urn

Fig. 3.4 Dependence of the peak area (uV.min) signal on the type of SPME fiber

It has been shown that it is not necessary to reach full equilibrium to obtain reliable results,

but a consistent sampling time and other sampling parameters (e.g. sampling temperature, air

circulation) have to be fixed in order to obtain good reproducibility. Furthermore, it is

important that the flask size, the sample volume and the location of the fiber in the headspace

of the flask are kept constant. Adsorption of the volatiles onto the SPME fiber has been

shown to be dependent on the temperature and air circulation in the extraction flask.

Especially air circulation increased the recovery of the volatiles. This could be confirmed by

additional experiments (not shown).

CHAPTER 3 28

3.3.2 Applications

The amount of total volatile compounds measured with the PDMS fiber (100 urn) differed for

each variety of strawberries (Fig. 3.5). This measurement introduces a new element when

estimating the quality of fruits and complements the measurements made by the sensory panel

and those of classic parameters such as pH, °Brix, acidity, etc.

7.0E+05

5.0E+05

CD

ro 3 0E+05

COCDQ.

1.0E+05

j4 #>

Evita Selua Mara des bois Seascape

name of varieties

Fig. 3.5 Effect of the variety of strawberry on the size of the total volatile fraction

Interesting results were obtained when individual strawberries were analyzed. For this we

used fruits grown in fields picked at the end of the season. As shown in Table 3.2, large

differences between the response of individual strawberries were found. These results indicate

a high heterogeneity among strawberries of the same variety (Tango), although their weight

and size were approximately equal.

This led to the following conclusions: 1) a correct quantification of the total volatile fraction

of strawberries may require a large sample size which has to be determined by using sampling

statistics; 2) since each strawberry releases its own aroma, the heterogeneity of individual

fruits has to be taken into account in the sensory analysis, either by measuring the total

volatile compounds of the tested sample, or by selecting, using the method described, a

restricted number of fruits for analysis. As a consequence, correlations between the results of

the sensory analysis and those obtained via laboratory instruments should be improved.

CHAPTER 3 29

Table 3.2 Aroma released by individual strawberries

strawberry peak area

weight (g) (mV.min)10.1 70016

10.3 46320

9.1 91711

10.5 21152

10.1 49966

9.7 49727

average 54815.3

SD 23862.7

CV 43.5%

Using strawberries as a model, we could show that the method fulfils classic scientific criteria

and gives easily reproducible results. It allows the differentiation between strawberry varieties

and clearly demonstrates the heterogeneity between individual fruits. Once the differences

between samples have been detected, it is possible to investigate the individual substances

responsible for these by standard SPME GC or GC/MS techniques (see chapter 5).

In order to show the applicability of our method to other types of food, coffee and spices

samples were analyzed under the same experimental conditions using PDMS (100 (im) and

CW/DVB (65 |im) fibers (Table 3.3). The level of reproducibility proved to be satisfactory.

The amount of total volatile compounds measured differed for each type of coffee. One of the

standard methods used for the determination of the quality of spices is based on the chromic

acid oxidation of the essential oil released by the spice. The results obtained with pepper

showed the method to be a promising way for the estimation of this quality parameter of

spices. However, it is interesting to note that even though the total volatile fraction measured

at 25 °C permitted easy comparison of the same types of spice, the areas measured did not

correspond to the differences in their essential oil contents. It is also interesting to note that

the Sarawak sample, which was older than samples of the other peppers, gave a smaller peak

area of total volatile compounds (values not shown).

CHAPTER 3 30

Table 3.3 Peak area of the total volatile compounds of coffee and spice samples. PDMS

Poly(dimethylsiloxane), CW/DVB Carbowax/divnylbenzene, NNestlé, MMigros, Co Coop, P

Pakoba

type and OriginPDMS 100 u.m (mV-min) CW/DVB 65 urn (mV-min)

average CV (%) average CV (%)

coffee Arabica 90 (N) 117.2 4.7 232.5 2.8

coffee Arabica 110 (N) 85.7 4.6 190.4 1.0

coffee Robusta 90 (N) 104.7 2.9 188.9 3.6

coffee Robustal 10 (N) 83.4 3.3 148.1 2.1

clove (M) 1461.1 3.5 2044.3 8.4

black pepper (Co) 7167.6 4.9 2667.7 2.9

black pepper (M) 4016.2 2.9 1826.1 3.5

black pepper (P) 784.3 3.7 540.4 4.6

white pepper (Co) 7423.9 3.7 3133.5 5.4

The system described here appears to be highly appropriate for the measurement of the

concentration of total volatile compounds, which are related to aroma. Research is being

carried out to correlate sensory (panel) test results with the amounts of total volatile

compounds, so that the criteria for the measurement of strawberry aroma can be clarified.

The rapidity of the method described overcomes the limitation of traditional methods used to

analyze volatile compounds, and permits the collection of on-line data. The daily sampling

capacity of the method is in the range of 10-20 samples and can be increased. The system

developed proved to be a convenient and appropriate methodology for the rapid quantitative

analysis of total volatile compounds of foods. At present, the applicability of the method for

different types of food is under investigation.

Applications should be possible in different fields: the detection of threshold values (e.g. gas

reaction control in chemistry and biotechnology) or quantitative measurements (e. g. quality

control, research and development: in food industry, flavor industry, cosmetics and

environmental chemistry).

3.4 ACKNOWLEDGEMENTS

The authors thank the Swiss Federal Office for Science and Education and the Canton of

Valais for their financial support in the context of the COST 915 action ("Improvement of

CHAPTER 3 31

quality of fruits and vegetables, according to the needs of the consumers"). We also thank

Fabienne Comby for her technical assistance.

3.5 REFERENCES

[I] Van Geloven, P., Honore, M., Roggen, J., Leppavuori, S., Rantala, T. (1991). The

influence of relative humidity on the response of tin oxide gas sensors to carbon

monoxide. Sens. Actuators B4, 185-188.

[2] Jonda, S., Fleischer, M, Meixner, H. (1996). Temperature control of semiconductor

metal-oxide gas sensors by means of fuzzy logic. Sens. Actuators B34, 396-400.

[3] Arthur, C.L., Potter, D.W., Buchholz, K.D., Motlagh, S., Pawliszyn, J. (1992). Solid

phase microextraction for the direct analysis of water: theory and practice. LC/GC 10,

656-661.

[4] Belardi, R.P., Pawliszyn, J. (1989). The application of chemically modified fused silica

fibers in the extraction of organics from water matrix samples and their rapid transfer to

capillary columns. Water Pollut. Res. J. Can. 24, 179-191.

[5] Arthur, C.L., Pawliszyn, J. (1990). Solid phase microextraction with thermal desorption

using fused silica optical fibers. Anal. Chem. 62, 2145-2148.

[6] Page, B.D., Lacroix, G. (1993). Application of solid phase microextraction to the

headspace gas chromatographic analysis of halogenated volatiles in selected foods. J.


[7] Arthur, CL., Killam, L.M., Motlagh, S, Lim, M., Potter, D.W., Pawliszyn, J. (1992).

Analysis of substituted benzene compounds in ground-water using SPME. Environ. Sei.

Technol. 26, 979-983.

[8] Arthur, C.L., Pratt, K., Motlagh, S., Pawliszyn, J. (1992). Environmental analysis of

organic compounds in water using solid phase micro-extraction. J. High Res.


[9] Arthur, C.L., Killam, L.M., Buchholz, K.D., Pawliszyn, J., Berg, J.R. (1992).

Automation and optimization of solid-phase microextraction. Anal. Chem. 64, 1960-

1966.

[10] Yang, X., Peppard, T. (1994). Solid Phase Microextraction for Flavor Analysis. JAgric.

Food Chem. 42, 1925-1930.

[II] Neubeller, J., Buchloch, G. (1978). Aromen und andere Fruchtinhaltsstoffe

verschiedener Erdbeersorten. Erwerbsobstbau 20, 189-192.

CHAPTER 3 32

[12] Penton, Z. (1996). Flavor volatiles in a fruit beverage with automated SPME. Food

Test. Anal. 2, 16-18.

[13] Pelusio, F., Nilsson T., Montanarella, L., Tilio, R., Larsen, B., Facchetti, S., Madsen,

J.O. (1995). Headspace solid phase microextraction analysis of volatile organic sulfur

compounds in black and white truffle aroma. J. Agric. Food Chem. 43, 2138-2143.

[14] Yang, X., Peppard, T. (1995). Solid-phase microextraction of flavor compounds - a

comparison of two fiber coatings and a discussion of the rule of tumb for adsorption

LC-GC 13, 882-886.

[15] Garcia, D., Maghaghi, S., Reichenbacher, M., Danzer, K. (1996). Systematic

optimization of the analysis of wine bouquet components by solid phase micro-

extraction. J. High Resol. Chromatogr. 19, 257-262.

CHAPTER 4 33

CHAPTER 4

APPLICATION OF A NEW CONCEPT FOR THE EVALUATION OF THE

QUALITY OF FRUITS*

4.1 INTRODUCTION

The aim of our work within the project COST 915 is an evaluation of the quality of fruit

(strawberries, tomatoes and apples). This is traditionally done by using different criteria such

as acidity, degree Brix, color and texture. In addition to these physico-chemical

measurements, the aroma (taste and odor) is an important sensory property of food, which is

perceived as quality parameter. Aroma is caused by the interaction of sensory organs with

volatiles and semi-volatiles associated with the food matrix. It is a challenge for analytical

chemists to study and define the complex mixtures of volatile food aroma compounds. In the

course of our work, we developed a simple and suitable method for the evaluation of total

volatile compounds in food [1]. This method gives us the possibility to investigate the

heterogeneity of fruit (tomatoes, apples and strawberries) with regard to their volatiles and to

determine the ripeness of strawberries by measuring the quantity of their volatiles.

4.1.1 Aroma analysis

In aroma analysis, sample preparation usually involves concentrating the analytes of interest,

using headspace, purge and trap, liquid-liquid extraction, solid-phase extraction, or

simultaneous distillation/extraction techniques. These methods have various drawbacks,

including long preparation times and the use of organic solvents. The newly developed "solid-

phase micro-extraction", SPME, eliminates most of the drawbacks of sample preparation.

* Azodanlou, R., Darbellay, C, Luisier, J.L., Villettaz, J.C., Amadö, R. (1999). Applicationof a new concept for the evaluation of the quality of fruits. In: Hägg, M., Ahvenainen, R.,

Evers, A.M., Tiililkkala, K. (eds.). Agri-food quality management of fruit and vegetables. The

Royal Society of Chemistry, Cambridge, 266-270.

CHAPTER 4 34

SPME is a technique of headspace sampling which concentrates analytes by adsorption onto a

polymer-coated silica fiber prior to thermal desorption in the injection port of a GC [2-5].

4.1.2 Our concept for aroma analysis

The system we propose is a hybrid between electronic nose and SPME-GC. The idea consists

of trapping the volatiles on a SPME fiber and determining the total amount of the adsorbed

substances after desorption without separation in a gas chromatography detector. The

feasibility of the method was proved by several applications with strawberries. For this work

the optimized sampling parameters described previously by Azodanlou et al. [1] were used.


4.2.1 Sample preparation for extraction of the volatiles

Two varieties of strawberries {Tango and Evita) were obtained from the Swiss Federal

Research Station for Plant Production (Conthey, Switzerland) in autumn 1997. Marmolada

and Elvira, Italian varieties, were purchased at the local market (Placette, Sion, Switzerland)

in spring 1998.

The tomatoes {Sombrero) were also purchased from a local food store (Placette) and the

apples {Maigold) were obtained from the Valais Cantonal Station for Fruit Production

(Chateauneuf, Switzerland).

For the estimation of the heterogeneity of strawberries {Tango and Elvira), intact individual

fruit of an approximate weight of 10 g were placed in a 150 mL round-necked flask with a

wide opening (NS/45) at 25 °C. The sample equilibration time was 15 min and the sampling

time by SPME (PDMS, 100 um) was 5 min.

In order to investigate heterogeneity the number of fruits was increased steadily. Strawberries

were placed in a tightly closed 2 L round-necked flask, apples and tomatoes were put in a 6 1

flask. Sample equilibration time was 5 min. The sampling time for adsorption was 5 min, each

sample was analysed five times. All the fibers described below were used.

For the measurements of maturity, 200 g of strawberries {Evita) were placed in a tightly

closed 1 L round-necked flask. The same analytical conditions as above were applied, except

for sample equilibration time, which was 30 min.

CHAPTER 4 35

4.2.2 Analytical procedure

SPME was performed as described for GC separations. The volatiles were collected by

inserting the needle through a Teflon-coated silicone septum into the headspace of the sample

flask. After a defined sampling time (see above) the adsorbed substances were desorbed in the

injection port of a GC Carlo Erba HRGC 5300 (Carlo Erba S.p.A., Milano, Italy). The

splitless injector port was coupled directly to the flame ionization detector using a transfer

tubing (20 cm in length, 0.1 mm i.d., no. 160-2630, J&W, New Brighton, USA), with a

helium pressure of 150 kPa, at a flow rate of approximately 5 mL/min. The oven temperature

was 250°C. The injection port and detector temperatures were 200 and 250°C, respectively.

For quantification, the area under the peak was measured by a GC integration program Chrom

Card (Fisons Inst., Milan, Italy) in u.Vs.

The following SPME fiber types were used: poly(dimethylsiloxane) lOOum (PDMS) (Cat No.

5-7300-U); polyacrylate 85îm (Cat. No. 5-7304); porous fibers Carbowax/divinylbenzene

65um (CW/DVB; Cat. No. 5-7312); bi-polar fibers: PDMS/DVB 65um (Cat. No. 5-7310-U)

and the new fiber Carboxen-PDMS 75um (Cat. No. 5-7318), all available from Supelco Co.,

Bellefonte, USA.


From our preceding work [1] we knew that the measurements were reproducible. So we could

investigate the differences between individual fruits. As a measure for the heterogeneity of a

batch we used the CV of five batches.

4.3.1 Heterogeneity with respect to variety or the type of production

Interesting results were obtained when individual strawberries were analysed. As shown in

Table 4.1, larger differences between individual strawberries of Tango variety were found

than for the fruits of the Elvira variety, although the sample weights and sizes were

approximately equal. The ratio of signal area to weight is strongly different and no correlation

between the weight of fruit and signal area could be found. These differences can be

explained by the type of production and the harvest season of the strawberries. One variety of

strawberries {Tango) was produced in a field crop in autumn, whereas the other {Elvira) was

grown in a glass-house at the end of winter. Even if these large differences between fruits and

between varieties can be influenced other parameters such as °Brix or pH, the measurement of

those differences will be invaluable for sensory analysis.

CHAPTER 4 36

Table 4.1 Aroma releases of individual strawberries

Tango variety Elvira variety

weight (g) peak area ratio weight (g) peak area ratio

10.1 70016 6932.3 11.4 16728 1467.4

10.3 46320 4484.0 10.5 16820 1598.9

9.1 91711 10045.0 10.9 15751 1434.5

10.5 21152 2012.6 11.3 20276 1791.2

10.1 49966 4937.4 10.0 15857 1584.1

9.7 49727 5115.9 10.1 18865 1869.7

average 9.9 5485.3 5587.9 10.7 17382.7 1624.3

SD 0.5 23862.7 2695.2 0.6 1805.6 173.8

CV 4.9 43.5 48.2 5.6 10.4 10.7

4.3.2 Determining the appropriate sample size of fruits

The first results on the heterogeneity of batches led us to investigate the appropriate sample

size to characterize a batch. Enlarging sample size improves the accuracy of quantitative

measurements. Using only different fruits, we made five measurements of each batch with an

increasing number of fruit from one to ten. The measurements of total volatile compounds

were performed with all available SPME fiber types. As a measure of the accuracy of the

measurements we used the variance or CV as shown for strawberries as an example in Table

4.2.

Table 4.2 Batch size and accuracy: strawberries on CW-PDMS fiber. CV: coefficient of

variation

number of

fruitsweight (g) peak area specific area CV on area (%) CV on weight (%)

1 31.6 13486 427.2 25.6 3.3

2 59.6 20145 337.8 15.5 2.8

3 94.9 24235 255.4 9.4 1.3

4 134.9 23799 176.4 13.1 1.3

5 164.6 30168 183.3 13.3 1.1

6 182.0 28971 159.1 10.8 0.5

7 201.3 31916 158.6 10.2 1.2

8 214.9 31648 147.3 10.5 0.7

9 249.9 33513 134.1 6.6 0.5

10 284.0 38354 135.0 5.2 0.9

With strawberries, we obtained a variance of about 20% with individual fruits per batch and

5.2% with 10 fruits per batch, respectively. Fig.4.1 shows the development of the CV with

CHAPTER 4 37

increasing the number of fruits measured. As we can see, the minimum batch for a correct

measurement should be about ten fruits for strawberries. With other fiber types we obtained

very similar results.

40

>O

30-

20-

10-

vP

4 8

number of fruits

—i

12

Fig. 4.1 CV (%) as a function of batch size; strawberries; CW-PDMS fiber

Another interesting result of these measurements is the fact that the specific area per weight is

decreasing with a higher number of fruits, being almost constant at/above six fruits (Fig. 4.2).

This could arise from geometric factors either of the fruits or of the measuring device.

750.0-

-c 500.0-

,|(0

m 250.0-

0.0 —i 1 1 1 1 1

0 2 4 6 8 10 12

number of fruits

Fig. 4.2 Evolution of the specific area with the number of fruits

CHAPTER 4 38

For apples {Maigold variety) or tomatoes {Sombrero), we obtained very similar results, but

with a much smaller variability. 3 apples are necessary to reach a CV of about 4%, whereas

with one fruit it was 15%; tomatoes exhibit the smallest heterogeneity between individual

fruit, giving a CV of about 11% for one, and 3.25% for 3 fruits, respectively.

4.3.3 Maturity of strawberries

With strawberries, we were able to document the differences in total volatile compounds

between ripe and unripe fruit. We selected one batch of ripe fruits and one of unripe fruits by

their colour. The total volatile compounds measured by the peak area is notably higher when

the strawberries are ripe. This technique permits to follow the ripening of strawberries (Fig.

4.3).

4.0E+05

J 3.0E+05

S 2.0E+05TO

TO<DQ.

1.0E+05

0.0E+00

*

Dada d = ao d

p

—i 1 1 1 1

0 20 40 60 80 100

equilibration time (min)

Fig. 4.3 Effect of maturity on the signal of total volatile compounds of strawberries. The

symbols denote values measured with SPME (PDMS): : ripe fruits, U: unripe fruits

4.4 CONCLUSION

This method is convenient to measure the intensity of aroma in fruit samples. It does not only

allow to differentiate between fruits but gives also the possibilities to describe some

properties such as maturity, variety, and harvest time.

For a correct quantification of the total volatile compounds the minimal number of fruits has

to be determined for each fruit species. The use of several SPME fibers permits to obtain

CHAPTER 4 39

different profiles. This concept is a convenient and appropriate sampling technology for rapid

quantitative analysis of aroma of food products.

4.5 ACKNOWLEDGEMENTS

The authors are grateful to the Swiss Federal Office for Science and Education, and to the

Canton of Valais for financial support in the framework of the COST 915 project

("Improvement of quality of fruit and vegetables, according to the needs of the consumers").

4.6 REFERENCES

[1] Azodanlou, R., Darbellay, C, Luisier, J. L., Villettaz, J. C, Amadö, R. (1999). A new

concept for the measurement of total volatile compounds of food. Z. Lebensm. Unters.

Forsch A. 208, 254-258.

[2] Belardi, R., Pawliszyn, J. (1989). The application of chemically modified fused silica

fibres in the extraction of organics from water matrix samples and their rapid transfer to

capillary columns. Water Pollut. Res. J. Can. 24, 179-191.

[3] Arthur, C.L., Pawliszyn, J. (1990). Solid Phase Microextraction with Thermal

Desorption Using Fused Silica Optical Fibers. Anal. Chem. 62, 2145-2148.

[4] Pawliszyn, J. (1995). New directions in sample preparation for analysis of organic

compounds. Trends Anal. Chem. 14, 113-122.

[5] Page, B.D., Lacroix, G. (1993). Application of solid phase microextraction to the

Headspace Gas Chromatographic Analysis of Halogenated Volatiles in Selected Foods.

J. Chromatogr. 648, 199-211.

CHAPTER 5 40

CHAPTER 5

QUALITY ASSESSMENT OF STRAWBERRIES*

ABSTRACT

Several cultivars of strawberries, grown under different conditions were analyzed both by

sensory and instrumental methods.

The overall appreciation, as expressed by consumers, was mainly reflected by attributes such

as sweetness and aroma. No strong correlation was obtained with odor, acidity, juiciness and

firmness. The sensory quality of strawberries can be assessed with a good level of confidence

by measuring the total sugar level (°Brix) and the total amount of volatile compounds.

Sorting out samples using the score obtained with a hedonic test (named hedonic

classification method), enabled us to strengthen considerably the correlation between

consumer's appreciation and instrumental data. Based on the results obtained, a quality model

was proposed.

Quantitative GC-FID analyses were performed to determine the major aroma components of

strawberries. Methyl butanoate, ethyl butanoate, methyl hexanoate, cz's-3-hexenyl acetate and

linalool were identified to be the most important compounds for the taste and aroma of

strawberries.

5.1 INTRODUCTION

Consumers are often criticizing the organoleptic quality of strawberries. According to the

information obtained from a large Swiss food retailer (Federation of Migros Cooperatives,

Bussigny, Switzerland), 26% of the consumers are often disappointed and 33% sometimes

disappointed with the quality of strawberries. Agronomic research have so far set the

* Azodanlou, R., Darbellay, C, Luisier, J.L.,Villettaz, J.C., Amadô, R. (in preparation).

CHAPTER 5 41

priorities on appearance, storage and transport resistance as well as on yield increase.

Therefore it is not surprising that the sensory properties only partly satisfy the expectations of

the consumers.

Compounds contributing to the flavor of strawberries, especially the volatile ones, have been

extensively studied. Nijssen [1] identified more than 360 volatile compounds. About 15-20 of

them are believed to be essential for the sensory quality of strawberries, together with the

non-volatile sugars and organic acids [2-7]. Flavor intensity and fruitiness persistence are

influenced by the concentrations of sugars and acids [8,9]. On the other hand, adding

strawberry flavor compounds to a sucrose solution induces an increase in the perception of

sweetness [10]. Alavoine and Crochon have shown that the total sugar content is correlated

with strawberry taste [11].

In spite of the extensive research done on strawberry flavor, the responsible substances for

aromatic distinction between cultivars have not yet been fully characterized [2,12]. The

differences in flavor of the three strawberry cultivars described by Ulrich et al. [13] is due to

different concentrations and ratios of the key flavor compounds: Wood strawberry, with a

spicy odor derived from anthranilic acid methyl ester; Fragaria Virginia, with a fruity aroma

characterized by esters; and Fragaria ananassa, characterized by furaneol and mesifurane

[14]. The typical strawberry aroma is not due to a single compound, but is rather the result of

a complex multi-component relation between many aromatic constituents [15]. The

interactive effects of these compounds are still poorly understood.

The aim of the present work was to assess for the quality of strawberries. Consumer tests and

sensory evaluations by a semi-trained panel were performed to establish quality criteria. In

addition a newly developed concept [16-18] was used to determine the amount of total

volatile compounds in strawberries. Sensory evaluation and physico-chemical analyses were

used as complementary tools to determine and to set quality acceptance limits.


5.2.1 Fruit samples and sample preparation

During three growing seasons (1997, 1998, and 1999), 80 samples representing 24 strawberry

cultivars, grown on field and/or under plastic tunnels, were harvested at the ripe stage and

CHAPTER 5 42

used immediately for sensory evaluation and instrumental analyses. The following cultivars

were analyzed by consumer tests as well: "Mara des bois", "Carezza", "Pegasus",

"Madeleine", "Elsanta" and "Marmolada". The samples were obtained either from the

Swiss Federal Research Station for Plant Production in Conthey (Switzerland) or from a large

Swiss food retailer (Federation of Migros Cooperatives).

Intact fruits were used for sensory evaluation and for determination of total volatile

compounds.

Strawberries classified by sensory evaluation were homogenized at high speed in a

professional blender (Kenwood professional, Kenwood, USA) for approximately 30 s to

produce a homogeneous puree which was directly used for instrumental analyses. To

inactivate the endogeneous enzymes, 50 g of a saturated ammonium sulfate (purum, Fluka

AG Buchs, Switzerland) solution was added to 50 g of fruit, directly into the blender. Finally

2-methyl-l-pentanol (Fluka purum; 1 mg/100 g of homogenate) was added as internal

standard.

5.2.2 Sensory evaluation

5.2.2.1 Consumer tests

Standard hedonic consumer tests with an average of 120 participants were carried out in

supermarkets (Federation of Migros Cooperatives) in different Swiss cities (La Chaux de

Fonds, Pully, Sion and Bienne). The test persons were asked to give an overall appreciation of

strawberry quality on a 1 to 9 scale, (1 = extremely bad to 9 = extremely good (annex 1). In

the 1999 campaign a modified procedure was adopted. Each fruit was divided into halfs; one

half was used to assess the sensory quality, while the other half was assigned to different

baskets according to the score obtained (1 to 9). The pooled samples were homogenized as

described above and used for instrumental analyses. This way of classifying samples is

hereafter called hedonic classification.

5.2.2.2 Sensorypanel

The sensory panel consisted of 10 to 15 semi-trained subjects. The subjects were asked to rate

the following sensory attributes: odor, aroma, sweetness, acidity, firmness, juiciness and to

give their overall appreciation (annex 2). The volatile compounds were evaluated in two

ways: first through the nose (odor) and then by the retro-nasal way through the pharynx after

masticating the sample (aroma). The panel rated the different parameters on a 1 to 9 scale

CHAPTER 5 43

(e.g. 1 = very weak aroma intensity and 9 = very strong aroma intensity). The same scale was

used for the overall appreciation (extremely bad to extremely good). Panelists were given

water (Volvic, Puy-de-Dome, France) as neutralizing beverage between sample testing. The

evaluation was carried out in a standard sensory laboratory under well-controlled conditions

using red light to mask any color differences.

5.2.3 Instrumental analyses

5.2.3.1 Determination oftotal volatile compounds

Fresh intact strawberries (400 ± 1 g) were carefully placed in a 2 L headspace flask with wide

opening (NS/160/100) as shown in Fig. 5.1. The flask was sealed with a teflon lid allowing

the simultaneous recovery of the volatile compounds by several SPME fibers. Different types

of SPME fibers were used: polydimethylsiloxane (PDMS) with lOOum thickness (Cat No. 5-

7300-U); polyacrylate (PA) 85um (Cat. No. 5-7304); porous fibers Carbowax/divinylbenzene

(CW/DVB), 65um (Cat. No. 5-7312); bi-polar fibers: PDMS/DVB 65um (Cat. No. 5-7310-

U), Carboxen-PDMS (CAR/PDMS) 75um (Cat. No. 5-7318) and CW/CAR/PDMS 50/30um

(5-7328-U) all obtained from Supelco Co. (Bellefonte, USA).

The fruits were left for 5 min at 25°C to obtain the necessary gas equilibrium in the

headspace. Aliquots of the volatile compounds were then collected by inserting the SPME

needle through a teflon-coated silicone septum into the headspace of the flask. After 5 min

(sampling time) the adsorbed substances were desorbed into a gas Chromatograph HRGC-

5300 (Carlo Erba S.p.A., Milano, Italy) equipped with a splitless injector port, directly

coupled to the flame ionization detector, using a transfer tube (20 cm in length, 0.1 mm i.d.,

N°160-2630, J&W, New Brighton, USA). The following GC conditions were used:

- helium carrier gas pressure 150 kPa at a flow rate of approx. 5 mL/min;

- hydrogen and air pressure for the FID: 50 kPa and 80 kPa, respectively.

- the oven temperature was set at 250°C.

- the injection port and the detector temperatures were set at 200 and 250°C, respectively.

CHAPTER 5 44

Fig. 5.1 Equipment for the determination of total volatile compounds according to Azodanlou

etal. [17]

A mixture (3 mg/kg) of 1 -methoxy-2-propyl-acetate (Merck, for synthesis), 2-methyl ethyl

ketone (Fluka, purum), and butanol (Fluka, puriss.) was used as external standard. The total

volatile peak (^V-min) was measured with a Borwin integrator (JMBS Développements,

Grenoble, France). Between each analysis, the headspace flask was cleaned by purging with

filtrated air that had previously passed through a charcoal trap (Supelpure-HC Trap, Supelco

Co).

Measurements on strawberry puree were carried out by spreading the sample into a

crystallizing dish (10cm diameter, 3cm height) which was then placed in the 6L headspace

flask. Total analysis time was approximately 15 min, including 5 min for both equilibration

and sampling. Each sample was analyzed in triplicate.


The volatile compounds of strawberries were extracted by SPME and identified and

quantified by GC. The volatiles were extracted as described in 5.2.3.1 using a 2 L headspace

flask and adsorbed on a CAR/PDMS fiber. Desorption was carried out directly into the

injector of the GC at 250°C.

CHAPTER 5 45

Identification procedure: A GC (HP 5890, series II, Hewlett Packard, Palo Alto, USA) linked

to a mass selective detector (HP 5971 A) and to an ionization gauge controller (HP 59822 B)

was used. Separation was achieved on a glass column (25m x 0.2mm i.d.) coated with a 0.33

urn film of diphenyl (5%)/dimethylsiloxane (95%) copolymer (HP-5), using the following

conditions:

- carrier gas helium at a flow 2 mL/min, 100 kPa; temperature program: 40°C for 2 min,

linear temperature gradient 40°C to 190°C at 4°C/min, 190°C for 5 min;

- injector and detector temperature, 250°C and 280°C, respectively;

- interconnecting line temperature: 300°C; MS settings: ion source pressure 10" Torr;

filament voltage 70eV, scan speed 1.9 scan/s.

Identification was performed by a combination of Kovats retention indices and a GC-MS

library (Flavornet, Geneva, USA). Some components were identified by comparison of

retention time and mass spectra with authentic substances. The following reference substances

were used: hexanal, butyl acetate, trans-2-hoxeno\, propyl butanoate, butyl butanoate, hexyl

acetate, isopropyl hexanoate, 1-octanol, linalool (Fluka); isoamyl acetate, 3-methylbutyl

butanoate, 3-phenyl-l-propanol, bornyl acetate (Aldrich, Milwaukee, USA); dimethyl

disulfide, ethyl butanoate, /ra«s-2-hexenal, ethyl 2-methylbutanoate, 2-methyl butanoic acid,

hexyl hexanoate, 4-hydroxy-2,5-dimethyl-3(2H)-furanone (Givaudan-Roure, Dübendorf,

Switzerland).

Quantification procedure: A GC (HP 6890) equipped with a FID was used for separation at

the same conditions as described for the GC-MS procedure. Hydrogen and air flow for the

FID were set at 40 mL/min and 450 mL/min respectively.

Quantification was performed by electronic integration of the peaks (HP-Chem-Station) using

2-methyl-1-pentanol (0.2 mg/kg) as internal standard.

5.2.3.3 Determination ofthe total sugar and acid contents

200 g of strawberries were homogenized to a puree as described in 5.2.1. Total sugar content

(°Brix) was determined using a refractometer (Atago, PR-1, Atago, Tokyo, Japan). pH and

total acidity were measured with a titrator (Mettler DL 25, Mettler-Toledo, Greifensee,

Switzerland). For determination of total acidity 10 ± 0.1 g of sample were titrated to pH 8.0

using 0.1 M NaOH. The titrated volume (mL) corresponds directly to total acidity expressed

as g/L citric acid.

CHAPTER 5 46

5.2.3.4 Texture analysis

The firmness of the strawberries (100 ± 1 g) was determined using a Kramer's shear cell

operated by a shear test machine (Versa Test + Advanced Forces Gauge, Memesin, Briitsch &

Riiegger, Zurich, Switzerland). The device speed was set at 250 mm/min. The fruits were

divided in two parts prior to measurements, which were performed in triplicate, at ambient

temperature.

5.2.4 Statistical evaluation

The Statview® program (Abacus Concepts Inc., Berkeley, USA) was used for the analysis of

variance (ANOVA). Significant differences in instrumental measurements between samples

were determined by PLSD (protected least significant difference) with P <0.05. Where the

test of normality failed, the non-parametric test was applied to the individual panel scores for

the investigated intensity criteria and then transformed into ranking numbers. The non-

parametric test was processed using the Kruskall & Wallis with (P <0.05). Pearson's

correlation analysis was carried out to identify the interdependence between different

variables of sensory, instrumental and chemical data (P <0.05).


Consumer tests were carried out to establish an overall appreciation of the quality of

strawberries. In addition, a sensory panel was employed to identify quality attributes. The

established sensory parameters allowed to distinguish between different cultivars and, most

important, between different quality attributes. A relationship between data obtained by

sensory evaluation and instrumental analyses allowed to develop a model for the appreciation

of strawberry quality.


A sensory panel was used to set up quality descriptors as outlined in 5.2.2.2. Eighty

strawberry samples out of twenty-four cultivars were used for the identification of the most

important quality attributes. Because the normality test failed (large variance of results), the

non-parametric test was used for statistical evaluation of the results. In the first year (1997),

the panels' objective was to define sensory descriptors such as odor, aroma, sweetness,

CHAPTER 5 47

acidity, firmness, juiciness, fondant and crunchiness (Table 5.1). Attributes such as fermented

odor and fermented taste, bitterness, herbaceous taste were not retained, because of their low

significance (P > 0.05). Some attributes such as fondant and crunchiness could not be

measured instrumentally and were therefore not further retained.

For the 1998 and 1999 harvests, the large number of fruit samples needed to evaluate the

quality by the consumer test, prompted us to use the hedonic test by the sensory panel. The

descriptors retained were again significant and could be used to explain differences between

fruit samples. Descriptors such as aroma, sweetness, firmness and juiciness turned out to be

significant quality attributes to describe the overall quality of strawberries.

Table 5.1 Comparison of strawberry samples by the sensory panel

harvest year

quality descriptors 1997 1998 1999

(16 samples) (31 samples) (30 samples)

oodor NS 0.002 SI 0.001 SI

o fermented NS NE NE

aroma 0.004 SI 0.001 SI 0.001 SI

sweetness 0.001 SI 0.001 SI 0.001 SI

wacidity 0.015 SI 0.001 SI 0.001 SI

§ bitter NS NE NE*"

herbaceous NS NE NE

fermented NS NE NE

persistence 0.005 SI NE NE

firmness NE 0.001 SI 0.001 SI

•s juiciness 0.001 SI 0.001 SI 0.001 SI

& fondant 0.001 SI NE NE

crunchiness 0.001 SI NE NE

E-i overall^ appreciation

NE 0.001 SI 0.001 SI

P values of Kruskall & Wallis, SI: Significant >95% and NS: Not Significant <95% level,

NE: Not Evaluated. HT: Hedonic Test

The relationship between the overall appreciation and some of the sensory descriptors such as

odor, aroma, sweetness, acidity, firmness and juiciness has been confirmed (P <0.05) in 1998

and 1999. As shown in Table 5.2, aroma and sweetness are two descriptors that correlates

well with the overall appreciation.

CHAPTER 5 48

Table 5.2 Correlation between overall appreciation and some sensory descriptors defined bythe sensory panel

sensory descriptors 1998 harvest 1999 harvest

aroma 0.86 0.94

sweetness 0.86 0.87

odor 0.56 0.55

juiciness 0.55 0.49

Consumer tests were performed on six strawberry cultivars as described in 5.2.2.1, to obtain a

preference score (Table 5.3). Although the sensory score obtained at different days varied for

a given cultivar, it was interesting to note that the ranking of the cultivars was always the

same. The comparison between strawberry cultivars was significantly different (P <0.05).

"Mara des bois" was always judged to have the best quality.

Table 5.3 Overall appreciation of 6 strawberry cultivars by consumers

dates of consumer test

18.5.1999 25.5.1999 2.6.1999

"iT Mara des bois (6.4) Mara des bois (7.0) Mara des bois (7.3)

c3 8 Carezza(ô.l) Carezza (6.8) Carezza (6.7)

2 Pegasus (5.6) Pegasus (6.2) Pegasus (6.3)

o ^ Madeleine (5.5) Madeleine (6.0) Madeleine (5.4)8> Elsanta(4.9) Elsanta(5.9) Elsanta (4.6)w Marmolada (4.2) Marmolada (5.5) Marmolada (3.6)

score: 1 = extremely bad, 9 = extremely good; (mean values)

To identify the most relevant quality attributes, sensory panel tests were performed on the

same cultivars used in the consumer tests (Table 5.3). Sweetness and aroma were the only

two attributes that correlated well several times (Table 5.4).

Table 5.4 Correlation between the overall appreciation by consumers and the qualityattributes of the sensory panel

„ i coefficient of correlation

attributes 18.5.1999 25.5.1999 2.6.1999

odor NS NS NS

aroma NS NS 0.89*

sweetness 0.94* 0.71* 0.89*

acidity NS NS NS

juiciness NS NS NS

firmness NS NS NS

*significant with P <0.05, NS = Not Significant

CHAPTER 5 49

The results indicate that strawberry is always highly appreciated if its sweetness and aroma

intensity are high.


Several chemical and physico-chemical parameters were established by instrumental methods.

The total sugar content (°Brix), pH and total acidity data were obtained for strawberries of all

harvests (1997, 1998 and 1999). The total volatile compounds were determined for the fruits

of the 1998 and 1999 seasons, and texture measurements were carried out on strawberries of

the 1999 harvest only.

5.3.3 Correlation between sensory and instrumental data

The relationships between data obtained by the sensory panel and instrumental methods have

been established. In Fig. 5.2 A the relationship between the total amount of volatile

compounds (P <0.05, r = 0.54) and the overall appreciation is shown. Because of the weak

correlation (P <0.05, r = 0.14) in the 1999 harvest, only the values for 1998 were presented

here. The total sugar content (Fig. 5.2 B) correlated significantly (P <0.05) with the overall

appreciation by the sensory panel for both the 1998 and the 1999 harvest (r = 0.68 and r =

0.67 respectively).

A)B)

</>

•oc

3 .—

O O)

Q- =*E o)

o E

0 -5-

p1 D.

CO*-*

O

1.6

1.3

1 0

0.7

0.4

12

* 11i_

COo

r ioc

CD

1 9

Ü

ni 8

I 7CO

U.O

2 3 4 5 6overall appreciation

2 3 4 5 6 7

overall appreciation

Fig. 5.2 Relationship between (A) total volatile compounds (PDMS fiber) and overall

appreciation given by the sensory panel for the 1998 harvest, and (B) °Brix and overall

appreciation given by sensory panel; 1998 (O) and 1999 ()

CHAPTER 5 50

It was interesting to note that for a given °Brix (e.g. 8.0), the overall appreciation score varied

considerably between the 1998 (2.5) and the 1999 (4.8) harvest season. The °Brix therefore

seems to be more robust than the total volatile compounds in terms of quality evaluation.

However it is clearly not appropriate to set the same °Brix as quality attribute to obtain

customer satisfaction every year.

A good correlation (P <0.05, r - 0.81) was found in 1998 between total volatile compounds

and the overall appreciation by consumers (Fig. 5.3A). In the 1999 trials, the correlation was

less evident (r = 0.14) probably because of the higher heterogeneity of the strawberries

obtained during the 1999 season (values not shown). The total sugar content (Fig. 5.3B)

correlated significantly (P <0.05) with the consumer ratings both for the 1998 and the 1999

harvest (r = 0.79 and r = 0.74 respectively). Total sugar content and amount of total volatile

compounds seem to be quality indicators for strawberries.

A)B)

-oc

E &

o -5,v CO

11So

\.l -

51.4 -

2/Ï1.1 - ly

a S

0.8 -

0.5- i i

12

11X

1»

CO 102

c qa)

r

oo

8

k.

Ol 7-jin

ni 6

o

5 6 7


-1 1 1 1 1

4 5 6 7 8


Fig. 5.3 Relationship between A) total volatile compounds (PDMS fiber) and overall

appreciation by the consumers in 1998 (O), and B) °Brix and overall appreciation by the

consumers in 1998 (O) and in 1999 ()

The results obtained by the sensory panel and by the consumer appreciation gave the same

relationship with the instrumental data used (°Brix and total volatile compounds). These

parameters as well seem to be appropriate for quality assessment of strawberries. Table 5.5

summarizes the results of the comparison between the consumers' appreciation and the

instrumental data.

CHAPTER 5 51

Table 5.5 Correlation between the overall appreciation by consumers and instrumental

analyses

instrumental coefficient of correlation

measurements 3.6.1998 18.5.1999 25.5.1999 2.6.1999

total sugar content 0.79* NS 0.89* 0.92*

total acidity NS NS NS NS

pH NS NS NS NS

penetrometer NS NA NA NA

Kramer's shear cell NA NS NS NS

conductivity NS NS NS NS

total volatiles

compoundsCAR/PDMS 0.77* 0.80* NS NS

PDMS/DVB 0.68* NS NS NS

CW/DVB NS NS 0.75* NS

PDMS 0.81* NS 0.91* NS

PA NS NS NS NS

CAR/PDMS/DVB NA NS NS NS

* significant at P <0.05, NA = Not Analyzed, NS = Not Significant

The high correlation between the total sugar content and the overall appreciation on the one

side and between the amount of total volatile compounds (measured with some of the SPME

fibers) and the overall appreciation on the other hand, led to the conclusion that the two

attributes "sweetness" and "aroma" are determinant for the quality of strawberries.

5.3.4 Hedonic classification for the assessment of strawberry quality

The main problem in the development of a model for the assessment of the quality of fruits

was the heterogeneity of the fruit samples, as it has been demonstrated in a previous work (see

chapter 4). Introduction of the hedonic classification (see 5.2.2.1) successfully solved this

problem. Indeed, the same fruit sample could be analyzed by instrumental methods and by the

consumer, what made a direct comparison of the results possible.

5.3.4.1 Total amount ofvolatile compounds

Using the hedonic classification a good correlation between the total amount of volatile

compounds and the consumers' overall appreciation was found for nearly all SPME fibers

used (P <0.05, r = 0.73-0.94), compared to correlations of r = 0.04-0.44 prior to classification.

As examples, the relationship between consumer appreciation and total volatile compounds

CHAPTER 5 52

(mg/kg) extracted by PDMS (r = 0.94) and CAR/PDMS (r = 0.90) fibers using hedonically

classified strawberries, are shown in Fig. 5.4A (PMDS fiber) and 5.4B (CAR/PDMS fiber).

4 5 6 7 8


t r

56

7 8


T

6

Fig. 5.4 Relationship between total volatile compounds extracted by a PDMS fiber (A) and a

CAR/PDMS fiber (B) of hedonically classified strawberry samples and consumer

appreciation.

The values for total volatile compounds obtained with the CAR/PDMS fiber showed a strong

correlation to the consumer ratings (r = 0.90) and confirmed the results obtained in 1998 (r =

0.77). With the CAR/PDMS fiber, which is porous and slightly more polar than the PDMS

fiber, a different behavior was observed. At first, the total amount of extracted volatile

compounds was smaller and the curve exhibited a maximum at an appreciation score of 7.

The reasons for the observed differences remain to be established, the porosity and polarity of

the SPME fibers are thought to play a key role.

5.3.4.2 Aroma compounds

The aroma of the strawberry is composed of a large number of substances belonging to

different classes of chemicals such as esters, alcohols and carbonyl compounds [1,6,13-14].

These substances contribute to the fruity and green notes (herbaceous odor) of strawberries

and were identified and quantified by GC-MS and GC-FID as described in 5.2.3.2. Taking

into account that the results obtained for total volatile compounds had shown that the different

types of SPME fibers adsorbed the same volatiles, however in different amounts, GC-analyses

were carried out with one SPME fiber type only. The CAR/PDMS fiber was chosen because

CHAPTER 5 53

of its good differentiation ability between the scores 3 to 6 in the overall appreciation (Fig.

5.4B). The results of the aroma analysis are presented in Table 5.6.

Table 5.6 Volatile compounds in strawberries, sampled with SPME (CAR/PDMS), identified

by GC-FID and GC-MS

identification RT RRI RI(Lit.)"'e OT(mg/kg)bc odor characteristics"

propyl acetate 4.65 682 716"

methyl butanoate3 5.41 714 723' 10"3-10-2b fruity, cheese,

ethereal

dimethyl disulfide 5.45 718 744", 742e onion

methyl 6.28 742 776' sweet

2-methylbutanotehexanal 7.89 793 801e 10-2-10"lb green, sour,cut grass

ethyl butanoate3 8.05 778 771", 804' 10-6-10"5b fruity, sweet,

cheese, apple

butyl acetate 8.10 800 816", 817e 10"2-10-lb apple, glue, pear

isopropyl butanoate 8.50 834 847' 10"2-10lb pungent

trans-2-hexenal 9.57 837 857",854e 0.17e fatty, green, fatty

trans-2-hexeno\ 10.54 862 862", 887' io-'-ib green leaves, fruity,burnt

ethyl 2-methylbutanoate 10.55 861 846"

isoamyl acetate 10.60 865 876e banana

propyl butanoate 11.70 893 898' pineaple

methyl hexanoate3 12.42 909 934e 10"2-10lb fruity, pineapple

2-methyl butanoic acid 13.07 924 838",873' 10"2-10-lbfruity,sourish,sweetybutylbutanoate14.03947isobutylbutanoate15.88990cw-3-hexenylacetate316.199981009egreenbananahexylacetate316.239991008"10-2-10"lbbanana,apple,pearisopropylhexanoate16.7010191040'fresh2-methlbutylbutanoate17.9810413-methylbutylbutanoate17.8910391093"mesifurane18.3910511-octanol18.4810541075",1072'chemicallinalool319.9810861101d,1100e10"4-10'3blemonpeel,flowershexylbutanoate22.2511471185'applepeel3-phenyl-1-propanol24.341200bornylacetate25.9212421289"hexylhexanoate29.2713341390eapplepeelHDF*31.181389io-3-io_2bburnt,sweet,caramel*4-hydroxy-2,5-dimethyl-3(2H)-furanone;RT:retentiontime;RRI:relativeretentionindices;RI:retentionindices;OT:odorthreshold.aConsideredasanimportantcontributortofreshstrawberryaromaquality;Odorsthresholds:publishedbyLarsenetal.[3,4]andbyUlrichetal.c[6,7];Kovatsindicesandodorcharacteristics"[19]andinflavornet'[20].

CHAPTER 5 54

Quantification of the most relevant aroma components such as: methyl butanoate (0.12-0.98

mg/kg), ethyl butanoate (0.006-0.7 mg/kg), methyl hexanoate (0.01-0.1 mg/kg), hexyl acetate

(0.01-0.7 mg/kg) and linalool (0.006-0.06 mg/kg) clearly indicated that all relevant

compounds were present in amounts above their threshold limit taken from the literature

(Table 5.6).

As expected, summing up the peak areas measured by GC-FID, the concentration of the

volatile compounds increased up to an appreciation score of 7 and then remained constant. A

good correlation (r*: polynomial correlation) with the consumer appreciation (r* = 0.88) has

been obtained (Fig. 5.5A). A similar correlation (r* = 0.97) was obtained between the total

amount of volatile compounds and the consumer appreciation. As shown in Fig 5.5B, the

esters (r* = 0.91) and the alcohols (r* = 0.74) seems to be major contributors to the

appreciation of the strawberries.

A) B)IU.0 -

2? 9.0 -

O)

Ü. 7.5- Sw

aj _ _

= 6.0 -

ro

> 4.5 -

«—

o

E 3.0-

CO

/

1.5-

0.0 -

1"

i— i i i i i

-ac

ro

<" -ä

Tr °><D E

= tn

jj o

° -S> o

o ro

E

ID

Ô.0 -

3.0-

O

2.5-Â

2.0-

1.5-

1.0-

0.5- V

f*.

0.0- 1

4 5 6 7 8


4 5 6 7 8 9


Fig. 5.5 Relationship between overall appreciation and (A) sum ofthe volatile compounds ()and (B) Sum of the volatile esters (0) and sum of volatile alcohols (A)

Based on the correlation coefficients, it can be stated that esters contribute essentially to the

overall appreciation of strawberry quality (r* = 0.91). Among this group of substances,

methyl butanoate (r* = 0.81) ethyl butanoate (r* = 0.84), methyl hexanoate (r* = 0.63), cis-3-

hexenyl acetate (r* = 0.73) and hexyl acetate (r* = 0.44) play a major role for the aroma of

strawberries.

Linalool also showed a good correlation with the consumer appreciation (r* = 0.57). The sum

of esters and linalool correlated strongly with the consumer appreciation (r = 0.90).

CHAPTER 5 55

5.3.4.3 Total sugar content

A very strong relationship (P <0.05, r = 0.94) between total sugar content (°Brix) after

hedonic classification of the samples and consumer appreciation was established as shown in

Fig. 5.6. Here again the advantage of the hedonic classification was evident; without hedonic

classification the correlation was poor (r = 0.16). These results reflect well the heterogeneity

of fruits, which was critical in the development of the quality model.

90 -,

X

CO 8.5 -

0

*^

•4-»

c 80 -

(1)

c

O 7.5 -

CJ

L_

CO7.0 -

Z3

(/}

~m 6.5 -

£6.0 H 1 1 1 1 1 1 1

23456789


Fig. 5.6 Relationship between total sugar content (°Brix) and consumer appreciation after

hedonic classification

5.3.4.4 Texture

The correlation between texture data as measured by the Kramer's shear cell and the overall

appreciation improved considerably with the hedonic classification method (before and after

hedonic classification r = 0.30 and -0.65, respectively). Nevertheless, the results were not

significant for grading the strawberries' quality (values not shown).

5.3.5 Development of a model for the assessment of strawberry quality

Based on these results a model containing three quality levels: "bad", "medium" and "good"

was developed. The distribution of the samples into the three quality classes was

approximately 1:1:1 (33.1%: 35.7%: 31.1%). The average appreciation for "bad" samples was

4.5 (range: 4-5), for "medium" samples 6 (range: 5-7), and for "good" samples 8.5 (range: 8-

9). In Table 5.7 intervals and limit values for the different quality attributes are given.

CHAPTER 5 56

Table 5.7 Quality model with limit values (and intervals) for total volatile compounds

(mg/kg), total sugar content (°Brix) and firmness (Fmax in N)

quality class

instrumental data "bad" "medium" "good"

total volatile

compounds (mg/kg)CAR/PDMS 0.21a (0.16-0.28) 0.52b (0.38-0.62) 0.58ab (0.56-0.60)PDMS/DVB 0.40a (0.31-0.49) 0.60b (0.50-0.66) 0.73b (0.53-0.87)CW/DVB 0.39a (0.31-0.45) 0.49b (0.36-0.61) 0.50b (0.46-0.55)PDMS 0.37a (0.32-0.39) 0.56b (0.32-0.81) 0.96c (0.88-1.03)DVB/CAR/PDMS 0.62a (0.41-0.84) 0.90b (0.65-1.09) 1.09e (1.01-1.18)

total sugar content 6.7a (6.5-7.0) 7.4b (7.1-7.7) 8.3e (8.3-8.4)

(°Brix)

firmness (F maxin N) 245.7ab (227.6-264.4) 240.7a (173.6-299.7) 226.1b (190.8-258.5)

level of significance 5%, different letters in a line mean significant differences

Two out of the six fiber types used, DVB/CAR/PDMS and PDMS, allowed to distinguish

between the 3 quality classes. The total sugar content (°Brix) was shown to be a very good

parameter to distinguish between the 3 quality levels as well, whereas texture measurements

(firmness) did not allow a clear cut distinction in the present study.

For the other instrumental measurements the average values for bad, medium or good quality

samples were always overlapping (values not shown).

Recent work performed by Carlen (not published), who used the hedonic classification

method for experiments with the strawberry harvest of the year 2000, confirmed the strong

relationship between instrumental methods (total volatile compounds using the PDMS fiber,

and total sugar content) and consumer appreciation.

5.4 CONCLUSIONS

The present study has clearly demonstrated aroma and sweetness to be the most important

quality attributes for strawberries. Hedonic classification of samples allowed to significantly

improve the correlation between instrumental data and consumer appreciation, and enabled us

to develop a quality evaluation methodology based on three quality levels. Multiple variable

analysis enabled to discriminate between the quality levels. Measurement of the total volatile

compounds using the best performing fibers (DVB/CAR/PDMS and PDMS) and

CHAPTER 5 57

determination of total sugar content (° Brix) were shown to be generally sufficient for

assessing the quality of strawberries.

The compounds that contributed significantly to the peak area, measured by the total volatile

compound analysis, were quantified and identified. Some of the individual compounds were

correlated to the consumer appreciation. In particular, esters were found to greatly contribute

to the sensory quality of strawberries.

5.5 ACKNOWLEDGMENTS


Valais for their financial support in the context of the COST 915 action ("Improvement of

quality of fruit and vegetables, according to the needs of the consumers"). We also thank Mr.

Roland Terrettaz and the cooperative Migros in Bussigny for supplying strawberry samples

and for organizing and setting up sensory tests.

5.6 REFERENCES

[1] Nijssen, L.M. (1996). Volatile compounds in food: Qualitative and quantitative data.

TNO Nutrition and Food Research Institute, Zeist, Netherlands.

[2] Fischer, N., Hammerschmidt, F.J. (1992). A contribution to the analysis of fresh

strawberry flavour. Chem. Mikrobiol. Technol. Lebensm. 14, 141-148.

[3] Larsen, M., Poll, L. (1992) Odor thresholds of some important aroma compounds in

strawberries. Z. Lebensm. Unters. Forsch. 195, 120-123.

[4] Larsen, M., Poll, L., Olsen, CE. (1992). Evaluation of the aroma composition of some

strawberry (Fragaria-Ananassa Duch) cultivars by use of odors threshold values. Z.

Lebensm. Unters. Forsch. 195, 536-539.

[5] Schieberle, P. (1994). Heat-induced changes in the most odor-active volatiles of

strawberries. In: Maarse, H., Van der Heij, D.G. (eds.). Trends in flavor research.

Elsevier Science Publishers, Amsterdam, 345-351.

[6] Ulrich, D., Eunert, S., Hoberg, E., Rapp, A. (1995). Analysis of strawberry aroma with

solid-phase microextraction. Dew?. Lebensm.Rundsch. 91, 349-351.

[7] Ulrich, D. Krumbein, A., Rapp, A. (1997). Analysis of aroma compounds in strawberry,

sour cherry and tomato by gas chromatography after solid phase micro-extraction. Deut.

Lebensm. Rundsch. 93, 311-316.

CHAPTER 5 58

[8] Bonnans, S., Noble, A.C. (1993). Effect of sweetener type and of sweetener and acid

levels on temporal perception of sweetness, sourness and fruitiness. Chem. Senses 18,

273-283.

[9] Noble, A.C. (1996). Taste-aroma interactions. Trends Food Sei. Technol. 7, 439-443.

[10] Reineccius, G.A. (1996). Instrumental means of monitoring the flavor quality of foods.

In : Tung-ching Lee, Hie-joon Kim (eds.). Chemical markers for processed and stores

fruit. ACS, Sym. Ser. 631, 241-252.

[11] Alavoine, F., Crochon, M. (1989). Taste quality of strawberry. Acta Horticulturae 265,

449-452.

[12] Hirvi, T. (1983). Mass fragmentographic and sensory analyses in the evaluation of the

aroma of some strawberry. Lebensm. Wiss. Technol. 16, 157-161.

[13] Ulrich, D. Rapp, A. Hoberg, E. (1995). Analysis of strawberry flavor-quantification of

the volatile components of cultivars of cultivated and wild strawberries. Z Lebensm.

Unters. Forsch. 200, 217-220.

[14] Zabetakis, I., Holden, M.A. (1997). Strawberry flavour: analysis and biosynthesis. J.

Sei. FoodAgri.74, 421-434.

[15] Nikirov, A., Jirovetz, L., Woidich, A. (1994). Evaluation of combined GC/FTIR data

sets of strawberry aroma. Food Quality and Preference 5, 135-137.

[16] Azodanlou, R., Darbellay, C, Luisier, J.L., Villettaz, J.C., Amadô, R. (1999).

Application of a new concept for the evaluation of the quality of fruits. In: Hägg, M.,

Ahvenainen, R., Evers, A.M., Tiililkkala, K. (eds.). Agri-food quality management of

fruit and vegetables. RSC, Cambridge, 266-270.

[17] Azodanlou, R., Darbellay, C, Luisier, J.L., Villettaz, J.C., Amadô, R. (1999). A new

concept for the measurement of total volatile compounds of food. Z Lebensm. Unters.

Forsch.. 208, 254-258.

[18] Azodanlou, R., Luisier, J.L., Villettaz, J.C., Amado, R. (1998). Development of new

aroma measurement quantitative device for analysis in fruits. Travaux de Chimie

Alimentaire et d'Hygiène. 89, 650.

[19] Kondjoyan, N.; Berdagué, J. L. (1996). A compilation of relative retention indices for

analysis of aromatic compounds. Saint Genes, Champanelle, France.

[20] http://www.nysaes.cornell.edu/flavornet/chem.html

CHAPTER 6 59

CHAPTER 6

CHANGES IN FLAVOR AND TEXTURE DURING THE RIPENING OF

STRAWBERRIES*

ABSTRACT

The amount of total volatile compounds, total acidity, total sugar content (°Brix) and fruit

firmness were used to characterize the degree of ripeness of three strawberry varieties

("Carezza", "Darselect" and "Marmolada"). The present chapter describes the application

of a novel concept using the measurement of total volatile compounds to distinguish between

various stages of strawberry ripeness. The CAR/PDMS SPME fiber was found to be best

suited to differentiate between the stages of ripeness. The amount of total volatile compounds

rapidly increased near maturity (between the % red stage and dark-red stage). Most of the

identified compounds were esters, followed by aldehydes, and alcohols. The most abundant

compounds were propyl butanoate, 3-phenyl-l-propanol, butyl butanoate, isobutyl butanoate,

3-methyl butyl butanoate and isopropyl hexanoate. The concentration of green aroma

components such as hexanal, trans-2-\\Qxeno\ and cz's-3-hexenyl acetate progressively

decreased during the maturation process until they became minor components in mature

strawberries.

6.1 INTRODUCTION

The quality of strawberries is defined by sensory attributes such as flavor and texture. Plant

selection, growing conditions (soil, climate, etc.), harvest date, and post-harvest treatments are

all influencing these quality attributes [1].

* Azodanlou, R., Darbellay, C, Luisier, J.L.,Villettaz, J.C., Amadô, R. (in preparation)

CHAPTER 6 60

Fruit ripening is a very complex process. It is influenced by the synthesis and action of

hormones responsible for the rate of ripening, the biosynthesis of pigments (carotenoids,

anthocyanins, etc.), the metabolism of sugars, acids and volatile compounds involved in

flavor development. In addition the modifications of the structure and composition of the cell

wall are known to affect the texture ofthe fruit [2,3].

Glucose, fructose and sucrose are by far the most abundant soluble components in

strawberries. The sugars constitute precursors of flavor compounds (e.g. furanones), and are

primarily used as energy source in the ripening process [4]. Acids can affect flavor as well,

but they are more important in processing [5]. Further it has been shown that volatile fatty

acids play an important role in ester formation in strawberries [6].

A small number of researchers have focused their work on changes in the profile of volatile

compounds during fruit ripening. Yamashita et al. have demonstrated that during ripening the

capacity to esterify 1-pentanol is developed by the fruit [6]. These reactions are presumably

enzymatic, although the responsible enzymes have not been characterized yet. Alternatively,

some flavor components could be the products of non-enzymatic reactions (e.g. the reaction

of an alcohol with an acid), where the formation of precursors is enzymatically controlled.

The authors pointed out that esters and furanones are key compounds of the strawberry flavor.

Ito et al. reported an increase in the total amount of volatile compounds, mainly esters. During

ripening particularly large amounts of methyl acetate and methyl butanoate were observed

[7]. Pérez et al. found that the concentration of volatile compounds increased during ripening

of "Chandler" strawberries and that ethyl hexanoate, ethyl butanoate and methyl butanoate

were the predominant components present in the fully ripe fruit [8]. Miszczak et al. studied

the changes and the correlation between volatile compounds and color of "Kent" strawberries

[9]. Gomes and Chaves de Neves confirmed the main results obtained by the above mentioned

authors [10]. The relationship between the degree of maturity and the concentration of 2,5-

dimethyl-4-hydroxy-3(2/f)-furanone and its derivatives has been investigated in 7 strawberry

varieties by Sanz et al. [11].

Parliment [12] discussed the influence of sampling on flavor analysis and stated that liquid-

liquid extraction and simultaneous distillation may lead to the production of artifacts, whereas

SPME allowed a more reliable flavor analysis [13].

CHAPTER 6 61

In the present study a quantitative analysis of flavor compounds and an assessment of texture

changes during strawberry ripening were carried out. Three strawberry varieties were

selected. Flavor analysis was performed by SPME, measurement of total volatile compounds,

GC-FID and GC-MS. For each strawberry variety the chromatographic aroma profiles at six

different ripening stages were compared. The results were processed by principal component

analysis (PCA). Total acidity, total sugar content (°Brix) and fruit firmness (using Kramer

shear cell) were determined as well.



Three commercial varieties of strawberries ("Carezza", "Darselect" and "Marmolada")

grown under plastic tunnels were obtained from the Swiss Federal Research Station for plant

production in Conthey (Switzerland). The fruits were harvested at six different ripening stages

(green, white, half red,% red, red-mature and dark-red).

Preparation offruit puree: Strawberries were homogenized at high speed in a professional

blender (Kenwood professional, Kenwood, USA) for approximately 30 s. The endogenous

enzymes were inactivated by adding 50 g of a saturated ammonium sulfate (purum, Fluka

AG, Buchs, Switzerland) solution to 50 g of fruit directly into the blender. Finally 2-methyl-

1-pentanol (1 mg/100 g of puree, Fluka, purum) was added as internal standard. The puree

was directly used for instrumental analysis.



Total volatile compounds of strawberries were analyzed as described in chapter 5. The

determinations on fresh intact strawberries (400 ± lg) were carried out in a 2 L headspace

flask. For the measurements on puree, 100 ± lg of the homogenate were spread out into a

crystallizing dish (10cm diameter, 3cm high) which was placed in the headspace flask.

The analyses were carried out with the same types of SPME fibers, and the experimental

procedure described in chapter 5 was strictly followed.

CHAPTER 6 62

6.2.2.2 Quantification and identification ofvolatile compounds in strawberries

The concentration of the volatile compounds present in strawberries was quantified by GC-

FID. The volatiles present in the headspace (250 mL) were analyzed using the same procedure

and the SPME fiber (CAR/PDMS) as described in chapter 5.

The volatile components were identified using reference compounds, Kovats retention indices

data and GC-MS library data. The same procedure was adopted as described in chapter 5.

6.2.2.3 Determination oftotal sugar and acid contents

For the determination of total sugar content (°Brix) and total acidity 200 ± 1 g of strawberry

puree was used to ensure sample homogeneity. Total sugar content was measured by

refractometry (Atago, PRl, Atago, Tokyo, Japan). Total acidity was determined by titrating a

10 ± 0.1 g sample to pH 8.0 using 0.1 M NaOH in a Mettler DL25 titrator (Mettler-Toledo,

Greifensee, Switzerland). The titrated volume (mL) expressed directly the total acidity in g/L

citric acid.


The firmness of strawberries (100 + lg) was measured with a shear test equipment

(VersaTest+Advanced Forces Gauge, Memesin, Brütsch & Rüegger, Zurich, Switzerland)

fitted with a Kramer shear cell. Strawberries were split in two parts prior to measurements.

The fruits were sheared at a speed of 250 mm/min at ambient temperature. The measurements

were performed in triplicate.


The Statview® program (Abacus Concepts Inc., Berkeley, USA) was used for analysis of

variance (ANOVA). Significant differences in instrumental measurements between samples

were determined by PLSD (protected least significant difference) with P > 0.05. The

Statbox program (Grimmer Logiciels Corp., Paris, France) was used for the Pearson's

correlation (P <0.05) and the principal component analysis (PCA) to identify the

interdependence between different stages of ripeness of the fruits and the instrumental data (P

<0.05).

CHAPTER 6 63

6.3 RESULTS AND DISCUSSIONS

6.3.1 Strawberry ripening stages

6.3.1.1 Total sugar and total acidity

The flavor of strawberries is greatly influenced by the presence of sugars and acids. It was

shown by several authors that the total sugar content increases during ripening and stimulates

the formation of secondary metabolites such as anthocyanins and furanones [14-16]. On the

other hand, a decrease of acidity with maturity was observed, particularly citric acid, malic

acid and quinic acid decreased considerably [14,17]. In the present study the two first stages

of ripeness were not analyzed because the blender could not produce a puree with green and

white strawberries. In agreement with previous studies, we observed a general increase in the

total sugar content (°Brix) for the three strawberry varieties investigated at the more advanced

stages of ripeness (Table 6.1). In addition, a significant decrease in total acidity (TA) during

different stages of ripening was observed. The "Darselect" variety contained more sugar than

the two other varieties.

Table 6.1 Total sugar content (°Brix) and total acidity (TA) of three strawberry varieties at

different stages of ripeness. Values are means of three replicates

strawberry ripening stage and analytical measurements

'AR 3AR R D-R

Varieties TA °Brix TA °Brix TA °Brix TA °Brix

Carezza

Darselect

Marmolada

9.7 7.7

8.9 7.5

7.3 6.8

8.9 8.3

8.4 8.4

6.5 6.2

8.0

6.7

5.7

8.2

9.3

8.4

8.2 8.4

6.4 9.3

4.9 8.7

TA: total acidity in g citric acid/L

1/2 R: half-red, 3/4 R: Va red, R: red-mature D-R: dark-red.

The ratio TA /°Brix decreased during the ripening process (Fig. 6.1). At the red stage

TA/°Brix ratios of 0.97, 0.72 and 0.68 were measured for the three strawberry varieties

"Carezza", "Darselect" and "Marmolada", respectively. Surprisingly there was no change in

this ratio for "Carezza" between the two most advanced stages of ripeness (red and dark-red).

Further investigations are necessary to find out if this result is dependent on the variety or due

to growing conditions.

CHAPTER 6 64

Cd

1.4

1.2

D

0.8

0.6

0.4 1 1 1 1 1 1

G W 1/2R 3/4R R D-R

stages of ripeness

Fig. 6.1 Ratio of total acidity/total sugar content (TA/°Brix) for () "Carezza",

(A) "Darselect" and (O) "Marmolada"

strawberries at different stages of ripeness.

Stages of ripeness: G: green, W: white, 1/2 R: half red, 3/4 R: 3/t red, R: red-mature D-R:

dark-red. Values are means of three replications

6.3.1.2 Total volatile compounds

The amount of total volatile compounds present in the headspace of strawberries adsorbed on

two different SPME fiber types increased rapidly with increasing degree of ripeness of the

fruits. The analyses performed with the PDMS and the CAR/PDMS fibers (Fig. 6.2) gave

correlation values of r = 0.65-0.97. However, very weak correlations were obtained using the

PA fiber (r = 0.08-0.55), probably because of poor adsorption.

A)

•oc

2 cn

8 E.u CO

aiw

o

1.4-,

1.2-

1.0-

0.8

0.6

0.4

0.2

B)

co

S o>

EÉS w

j= a

P.2 oo

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

W 1/2 3/4 R D-R

R R

stages of ripeness

W 1/2 3/4 I

R R

stages of ripeness

D-R

Fig. 6.2 Amounts of total volatile compounds of ()"Carezza", (A)"Darselect" and

(0)"Marmolada" strawberries at different stages of ripeness. A) PDMS fiber; B) CAR/PDMS

fiber. Stages of ripeness: G: green, W: white, 1/2 R: half red, 3/4 R: % red, R: red-mature D-

R: dark-red. Values are means of three replications.

CHAPTER 6 65


Using the methods described in 6.2.2.2, 29 volatile compounds were identified and quantified

in the headspace of strawberries. Most of them showed a statistically significant increase in

concentration during maturation (Table 6.2). Many of the volatile compounds described in

Table 6.2 clearly allowed to discriminate between unripe and ripe strawberries.

Seite Leer /

Blank leaf

CHAPTER

667

Table

6.2.

Identificationandquantification

ofvolatilecompounds

inthest

rawberry

varieties"Carezza","Darselect"and"Marmolada"by

GC-MS

andGC-FID,

usin

gtheCAR/PDMSSPME

fiber

concentration(mg/kg)

odor

'Carezza"

"Darselect"

"Marmolada"

volatilecompounds

threshold

(mg/kg)

odor

unripe

ripe

unripe

ripe

unripe

ripe

esters

butylacetate

ÎO^-IO

"1*

apple,

glue

*0.00a

1.67b

0.00a

0.30b

0.00a

0.85b

isoa

mylacetate

banana**

0.03a

0.54b

0.00a

0.44b

0.00a

0.14b

hexylacetate

KrMO"1*

bana

na,apple,

pear*

1.03a

8.11b

0.09a

3.48b

0.27a

4.10b

c/s-

3-he

xeny

lacetate

greenbanana**

4.29a

1.72a

5.43a

1.25b

3.78a

2.10a

born

ylacetate

0.04a

0.16a

0.07a

0.19a

0.19a

0.59b

methylbutanoate

l(r3-l(v2*

fruity,cheese

*0.23a

4.75b

0.00a

0.96b

0.16a

5.96b

meth

yl2-methylbutanote

sweet**

0.00a

0.17b

0.00a

0.26b

0.00a

0.47b

ethy

lbutanoate

i(v6-i(r5*

fruity,sweet,cheese

*0.00a

2.87b

0.00a

1.58b

0.00a

3.29b

propyl

butanoate

pine

appl

e**

0.00a

0.08b

0.00a

0.06b

0.00a

0.10b

isop

ropy

lbutanoate

0.01-0.1*

pungent**

0.00a

0.92b

0.00a

0.13b

0.00a

0.80b

ethy

l2-

methylbutanoate

0.31a

0.39a

0.00a

0.00a

0.00a

0.00a

butylbutanoate

0.35a

2.65b

0.21a

3.36b

0.05a

3.33b

isob

utyl

butanoate

0.00a

2.69b

0.15a

2.69b

0.00a

4.56b

2-me

thylbutylbutanoate

0.00a

0.08b

0.00a

0.00a

0.00a

0.23b

3-me

thyl

butylbutanoate

0.00a

0.27b

0.00a

0.13b

0.00a

0.66b

hexylbutanoate

applepe

el**

0.30a

1.13a

0.23a

2.83b

0.33a

1.95b

methylhexanoate

io'Mo'1*

pineappl

e**

0.00a3.12b0.00a9.02b0.11a4.58bethylhexanoateio-5-io-4*applepeal**1.09a5.25b0.25a2.55b0.20a2.90bisopropylhexanoatefresh**0.00a0.40b0.00a0.19b0.00a0.47bhexylhexanoateapplepeel**0.45a0.53a0.36a0.55a0.11b0.24

b

CHAPTER

668

Table6.2(c

ont.

)

concentration(mg/kg)

volatilecompounds

odor

threshold

(mg/kg)

odor

"Carezzct"

unripe

ripe

"Darselect"

unripe

ripe

"Marmolada"

unripe

unripe

aldehydes

hexanal

10'MO"1*

green,sour

*0.63a

0.00b

0.54a

0.16b

0.65a

0.29b

trans-2-\\Qxena\

0.17**

fatt

y**

0.04a

1.01b

0.22a

0.00b

0.00a

0.20b

trans-2-hexena.\

0.17**

fatt

y**

0.04a

1.01b

0.22a

0.00b

0.00a

0.20b

alcohols

/rora-2-hexenol

10"'

-1*

green,

fruity,burnt*

2.25a

0.20b

0.78a

0.00b

0.56a

0.29a

1-octanol

chemical**

0.00a

0.00a

0.00a

0.00a

0.00a

0.12b

linalool

10"4-10"3*

lemon

peel

,flowers*

0.18a

1.56b

0.09a

1.81b

0.07a

1.07b

3-ph

enyl-1

-propanol

0.00a

0.10b

0.00a

0.10b

0.00a

0.18b

acid

2-methylbutanoicacid

10"2-10-'*

sweet**

0.00a

0.46b

0.00a

0.00a

0.00a

0.09b

furans

4-hy

droxy-2,5-dimethyl-

10"j-1

0"2*

burn

t,sweet,caramel*

0.00a

2.96b

0.00a

0.42b

0.00a

0.84b

3(2H)-fura

none

(furaneol)

4-methox

y-2,

5-di

meth

yl-

0.00a

2.96b

0.00a

0.42b

0.00a

0.84b

3(2H)-furanone(m

esif

uran

e)

valuesgi

ven

inthecolumn

"unr

ipe"

representtheaverageofthede

gree

sofripeness

"green",

"white"and

"Vired".Valuesgiven

inthecolumn

"ripe"

represent

the

average

of

the

degr

ees

of

ripeness

"%red",

"red",

and

"dark-red".

Values

which

aremarked

by

different

lett

erare

significan

tlydifferentbasedonthePLSD-test

atP

<0.05.*Larsenand

Poll

,1992

[18];**Acree

et

al.

[19]

.

CHAPTER 6 69

After sorting out non-discriminant volatile compounds, principle component analysis (PCA)

was applied to cluster ripe and unripe fruits. 76% of the variability could be explained by the

two principal components (Fig. 6.3). From this figure it can be clearly seen that the first

principal component discriminated the unripe samples according to the contents of hexanal,

propyl butanoate, 3-phenyl-l-propanol, butyl butanoate, isobutyl butanoate, 3-methyl butyl

butanoate and isopropyl hexanoate. Hexanal was the predominant component in the

chromatograms of the green fruit samples in all varieties. The concentration of green aroma

components such as hexanal, trans-2-hexenol and c/s-3-hexenyl acetate progressively

decreased during the ripening process until they became minor components in the mature

samples.

CM

CM

O(0LL

19-8-

yhexyl butanoate u *> ""

X methyl hexanoate

X methyl 2- r>R (C)methylbutanoate

3-phenyi-1-propanol _ „ „ /ra

XJjutyf butanoate D-R (D) R^W(D)

X isobutyl butanoate

X j3-methyl butyl butanoate 0 2 -

X llnal0°x ethyl butanoate

/ W(C)

3/4R (C)/ W(M)

propyl butanoate ^ 2-mBttwl butyl butanoateR (D)

,X isopropyl hexanoafe , n

B1/2R(C) /G P)tf/fR (D)

2 -1 -0 8 -0 6 -0 4 -0 2 (^t(2R (D)

) G(M)»2v '

0 4 0 6 A8

Xisoamyl acetate 1/2R (M) '"G(C)

hexanalX methylbutanoate X trans-2-hexenol

isopropyl butanoate X cis-3-hexenyl acetate

>X ethyl hexanoate

Xhexyl acetate3/4R(M)

X butyl acetateR

-0 4-

X mesifurane

-0 6 -

X 2-methyl butanoic acid

X trans-2 hexenal

I -6-8-I

Factor 1 (48 5%)

Fig. 6.3 Principal components of the strawberry varieties "Carezza" (C), "Darselect" (D) and

"Marmolada" (M). 24 variables were evaluated, (x) volatile components and () stages of

ripeness: G: green, W: white, 1/2 R: half-red, 3/4 R:3Ared,R:red-matureD-R:dark-red.Theresultssuggestedthattheripeningprocessinstrawberriescanbefollowedbyanalyzingthecontentofspecificvolatilecompounds.Itwasthuspossibletocharacterizethedegreeofripenessofthetwovarieties"Carezza"and"Marmolada"bytheircontentsin2-methyl

CHAPTER 6 70

butanoic acid and 2-methyl butyl butanoate. Ethyl 2-methyl butanoate proved to be typical

for the "CarezzcT variety.

It has been reported that furaneol (2,5-dimethyl-4-hydroxy-furane) is the most important

component for the strawberry aroma [11,20]. Our results indicated the furaneol concentration

to be quite fluctuating throughout the ripening process, probably due to its metabolism. It is in

fact known that this compound is not very stable [8].

According to Table 6.2 the flavor components of strawberries belong predominantly to the

chemical classes of esters, aldehydes, alcohols, acids and furans. It was interesting to note that

the sum of the esters considerably increased during the last stages of ripening (Fig. 6.4).

70i

j? 6CH

a> 30|

| 2«

0

G W 1/2R 3/4 R R D-R

stages of ripeness

Fig. 6.4 Sum of volatile esters in (M)"Carezza", (À)''Darselect" and (0)"Marmolada"strawberries at different stages of ripeness

Stages of ripeness: G: green, W: white, 1/2 R: half red, 3/4 R: % red, R: red-mature D-R:

dark-red. Values are means of three replications.

Butyl acetate and ethyl butanoate seem to play an important role. Their concentrations

correlated well with the overall appreciation expressed by consumers [21].


The firmness of the strawberries as measured by the Kramer's shear cell decreased

dramatically during the ripening process (Fig. 6.5). At the ripening stage red-mature (R) the

maximal force for the varieties "Carezza", "Darselect", and "Marmolada" were 167N, 200N,

330N, respectively.

CHAPTER 6 71

1600 - 4

g 1300 - \

% 1000 - \

| 700- \W400- ^^<?\^-îoo -l—i—i—i—r~^~~j i

G W1/2R 3/4R R D-R

stages of ripeness

Fig. 6.5 Firmness (Fmax in N) of the strawberry varieties (M)uCarezza", (A)"Darselect" and

(0)"Marmolada" at different stages of ripeness

Stages of ripeness: G: green, W: white, 1/2 R: half red, 3/4 R: 3A red, R: red-mature D-R:

dark-red. Values are means of three replications.

Strawberry softening is a consequence of changes in physical and mechanical properties of

the tissue based on changes in the chemical structure of the cell wall polysaccharides

(cellulose, hemicelluloses and pectins). It was shown [22] that pectins play a key role in fruit

softening, since they are partially solubilized by endogenous pectin degrading enzymes

(polygalacturonases and pectin methyl esterases) during ripening, leading to tissue softening.

6.4 CONCLUSIONS

The quantitative determination of total volatile compounds in the headspace by the method

developed by Azodanlou et al. [23] has been shown to be a powerful tool for the

characterization of the ripeness of strawberries. The use of PCA turned out to be very useful

to visualize the correlations between different strawberry varieties ("Carezza", "Darselect"

and "Marmolada "). Important volatile components and classes of compounds were identified

by GC-FID and GC-MS. It was possible to correlate some of the individual components to the

stages of ripeness. Esters represent key aroma components in strawberries. Their total amount

increased during ripening.

CHAPTER 6 72

6.5 ACKNOWLEDGMENTS


Valais for their financial support of this project which was carried out in the frame of the

COST 915 action ("Improvement of quality of fruit and vegetables, according to the needs of

the consumers").

6.6 REFERENCES

[1] Dirinck, P., De Pooter, H., Schamp, N. (1989). Aroma development in ripening fruits. In:

Teranishi, R., Butterey, R.G., Shahidi, F., (eds.). Flavor Chemistry: Trends and

Developments. ACS, Washington D.C, 23-34.

[2] Abeles, F.B., Takeda, F. (1990). Cellulase activity and ethylene in ripening strawberry

and apple fruits. Sei. Hortic. Amsterdam 42, 269-275.

[3] Mitchell, W.C., Jelenkovic, G. (1995). Characterizing NAD- and NADP-dependent

alcohol dehydrogenase enzymes of strawberries. J. Amer. Soc. Hort. Sei. 120, 798-801.

[4] Makinen, K.K., Séderling, E. (1980). A quantitative study of mannitol, sorbitol, xylitol

and xylose in wild berries and commercial fruits. J. Food Sei. 45,37-371.

[5] Coultate, T.P. (1996). Flavours. In: Coultate, T.P. (ed.). Food: The chemistry of its

components. Royal Society of Chemistry Publishers, Cambridge. 169-207.

[6] Yamashita, I., lino, K., Nemoto, Y., Yoshikawa, S. (1977). Studies on flavor

development in strawberries. 4. Biosynthesis of volatile alcohols and esters from

aldehydes during ripening. J. Agric. Food Chem. 25, 1165-1168.

[7] Ito, O., Sakakibara, H., Yajima, I., Hayashi, K. (1990). The changes in the volatile

components of strawberries with maturation. Flavour Sei. Technol. Weurman Symp. 6th.

69-72.

[8] Pérez, A.G., Rios, J. J., Sanz, C, Olias, J.M. (1992). Aroma components and free amino

acids in the strawberry variety Chandler during ripening. J. Agric. Food Chem. 40, 2232-

2235.

[9] Miszczak, A., Forney, C. F., Prange, R. (1995). Development of aroma volatiles and

color during post-harvest ripening of Kent strawberries. J. Amer. Soc. Hort. Sei. 120,

650-655.

CHAPTER 6 73

[10] Gomes da Silva, M.D.R., Chaves das Neves, H.J. (1997). Differentiation of strawberry

varieties through purge-and-trap HRGC-MS-FTIR and principal component analysis. J.

High Resol. Chromatogr. 20, 275-283.

[11] Sanz, C, Richardson, D.G., Pérez, A.G. (1995). 2,5-Dimethyl-4-hydroxy-3(2H) furanone

and derivatives in strawberries during ripening. ACS, Symp. Ser. 268-275.

[12] Parliment, T. H. (1997). Solvent extraction and distillation techniques. In: Marsili, R.

(ed.). Techniques for Analyzing, Food Aroma. Dekker Corp., New York, 1-26.

[13] Parliment, T. H. (1997). Solid-phase extraction for the analysis of flavors. In: Marsili, R.

(eds.). Techniques for analyzing, food aroma. Dekker Corp., New York, 81-112.

[14] Woodward, J. R. (1972). Physical and chemical changes in developing strawberry fruits.

J. Sei. FoodAgric. 23, 465-473.

[15] Forney, CF., Breen, P.J. (1986). Sugar content and uptake in the strawberry fruit. J.

Amer. Soc. Hort. Sei. Ill, 241-247.

[16] Pisarnitskii, A.F., Demechenko, A. G., Egorov, I.A., Gvelesiani, R.K., (1992).

Methylpentoses are probable precursors of furanonesin fruits. Appl. Biochem. Microbiol.

28,97-100.

[17] Sistrunk, W.A., Cash, J.N. (1973). Nonvolatile acids of strawberries. J. Food Sei. 38,

807-809.

[18] Larsen, M., Poll, L., Olsen, C.E.Z. (1992). Evaluation of the aroma composition of some

strawberry (Fragaria ananssa Dutch) cultivars by use of odour threshold values. Z

Lebensm. Unters. Forsch. 195, 536-539.

[19] http://www.nysaes.cornell.edu/flavornet/chem.html

[20] Ulrich, D. Rapp, A. Hoberg, E. (1995). Analysis of strawberry flavor-quantification of

the volatile components of varieties of cultivated and wild strawberries. Z. Lebensm.

Unters. Forsch. 200, 217-220.

[21] Azodanlou, R., Darbellay, C, Luisier, J.L., Villettaz, J.C., Amadö, R.. (in preparation).

[22] Fischer, R.L., Bennett, A.B., (1991). Role of cell wall hydrolases in fruit ripening. Annu.

Rev. Plant Physiol. 42, 675-703.

[23] Azodanlou, R., Darbellay, C, Luisier, J. L., Villettaz, J. C, Amadö, R. (1999). A new

concept for the measurement of total volatile compounds of food. Z. Lebensm. Unters.

Forsch A. 208,254-258.

CHAPTER 7 74

CHAPTER 7

OBJECTIVE QUALITY ASSESSMENT OF TOMATOES AND APRICOTS*

ABSTRACT

The quality of tomato and apricot cultivars was assessed by sensory evaluation and

instrumental methods.

The overall sensory appreciation of tomatoes was mainly reflected by attributes such as

"sweetness", "aroma", "juiciness" and "firmness". Instrumental measurements were focused

on total sugar content (°Brix) and total amount of volatile compounds. For apricots the

instrumental measurements were focused on total sugar content (°Brix), total amount of

volatile compounds and texture (firmness).

The results obtained with tomatoes and apricots confirmed the applicability of the quality

assessment model developed for evaluation ofthe quality of strawberries.

7.1 INTRODUCTION

Attributes such as color, size, shape and external defects of fruit and vegetables

predominantly determine the choice made by the consumers. However, these parameters

alone do not guarantee the flavor and texture quality of a product. The sum of sugars, organic

acids as well as the amount and type of volatile compounds determine the sensory properties

of fruit and vegetables, e.g. tomatoes and apricots [1-3].

The flavor of tomatoes and apricots can be characterized by nearly the entire set of their

constituents. Indeed, the flavor is not only directly reflected by the sum of the volatile and

non-volatile components, but also depends on their interactions [3-5]. Although taste and odor

* Azodanlou, R., Darbellay, C, Luisier, J.L.,Villettaz, J.C., Amadô, R. (in preparation).

CHAPTER 7 75

are perceived by different senses, the proximity of the sensory organs and their connection

through the pharynx render a separate analyses of taste and odor difficult.

Sugars and acids reflect the overall taste preference for a fruit. Over the past decades research

has been carried out to enhance the sugar and acid contents, thereby acting on the pleasant

sweet-sour taste of the fruit. The sugar fraction of tomatoes is essentially composed of glucose

and fructose. Taste character and intensity are greatly affected by salts and by the buffering

effect of the cations and anions present. The incidence of the sugar/acid ratio has been shown

to be of little/no importance for tomatoes [6,7].

So far, only little attention has been paid to the contribution of the volatile compounds to the

quality of tomatoes. Although more than 400 volatile compounds have been identified, only

few of them such as hexanal, trans-2-hexenal, c/s-3-hexenal, c/s-3-hexenol, trans-2-trcms-A-

decadienal, 2-isobutylthiazole, 6-methyl-5-hepten-2-one, l-penten-3-one and ß-ionone seem

to play an important role for the flavor of tomatoes [4, 8-12]. 2-isobutylthiazole is present in

very small amounts, however its concentration increases rapidly during mastication of the

tomato in the mouth. On the other hand, hexanal is produced in mashed tissue by an

enzymatic lipid oxidation. Hexenol is formed by the action of alcohol dehydrogenase on

hexenal [13]. Butterey et al. [9,14] provide a scientific basis for the subjective observation

that cold storage is deleterious to fresh tomato flavor. The same authors have proposed a

mixture of substances to match the typical aroma of a tomato paste.

Only few reports with analytical data thorough enough to be useful for quality control of

apricots are available [15-18]. The first studies on apricot flavor were performed by Tang and

Jennings [19]; a number of terpenes and alcohols were identified to be present in the

"Blenheim" cultivar. Several constituents, such as lactones, terpene alcohols and

benzaldehyde were identified in different apricot cultivars [20-22]. Guichard and Souty

compared the relative concentrations of various volatiles in six different apricot cultivars [23],

and showed the Cg lipid degradation products, lactones, terpenes and ketones to be the most

abundant constituents. Odor unit values and odor threshold data indicated thatvolatilecompoundssuchas/7-ionone,linalool,^decalactone,hexanal,(ïj-2-hexenal,(E,E)-decadienal,f£)-2-nonenalandj^dodecalactonerepresentedthemajorcontributorstotheapricotaroma[24].

CHAPTER 7 76

On the other hand the quantitative composition of organic acids and soluble sugars have often

been used as indicators for the quality of apricots [25,26]. Recently, GC-MS was used to

evaluate sugars, sugar alcohols, acids and amino acid derivatives, based both on the total ion

current (TIC) and selective fragment ion (SFI) values [27].

The study of polyphenols is also important because of their contribution to the sensory quality

of fruits (color, astringency, bitterness, and flavor) [28]. Analysis of phenolic constituents

including the flavonoids allowed to characterize and differentiate apricot cultivars [29-31].

The present work describes the application of a new concept developed for quality assessment

of strawberries (see chapter 5) for the evaluation of the quality of tomatoes and apricots. The

total amounts of volatile compounds of tomatoes and apricots were determined by the method

of Azodanlou et al. [32-34]. Sensory evaluation and physico-chemical instrumental methods

have been used as complementary tools to determine and to set quality acceptance limits.



Tomatoes: During three growing seasons (1997, 1998, 1999) 149 samples representing 28

tomato cultivars, grown on field, in glasshouses or in plastic tunnels were harvested at the ripe

stage and used immediately for sensory evaluation and instrumental analyses. The fruits were

obtained either from the Swiss Federal Research Station for Plant Production in Conthey

(Switzerland) or from a large food retailer (Federation of Migros Cooperatives, Bussigny,

Switzerland). The tomatoes were harvested at different periods in June and July.

Apricots: Consumer tests and instrumental analyses with apricots were carried out in 1999 on

the cultivars "Jumbo", "Luiset", "Bergeron", "Fantasme" and "Tardif de Tain". The

apricots were obtained from the local Research Station for Fruit Production (Châteauneuf,

Switzerland), and were harvested at two or three different stages of ripeness: pre-ripe (near to

ripe), ripe and over-ripe.

Intact fruits were used for sensory evaluation and for determination of total volatile

compounds.

CHAPTER 7 77

Tomatoes and apricots classified by sensory evaluation were homogenized at high speed

either in a Solemio blender (Fiseldem, Cinisello, Italy) or in a professional blender (Kenwood

professional, Kenwood, USA) for approximately 30 s to produce a homogeneous puree

which was directly used for instrumental analyses. To inactivate the endogeneous enzymes,

50 g of a saturated ammonium sulfate (purum, Fluka AG Buchs, Switzerland) solution were

added to 50 g of fruit, directly into the blender. Finally 2-methyl-l-pentanol (purum, Fluka; 1

mg/100 g of homogenate) was added as internal standard.


7.2.2.1 Consumer tests

Approximately 120 consumers participated in a hedonic test performed in supermarkets

(Federation of Migros Cooperatives) in different Swiss cities (Bern, Lausanne, St. Gallen, San

Antonino, Sion, Zug). The test persons were asked to give an overall appreciation of tomatoes

and apricots on a 1 to 9 scale: 1 = extremely bad to 9 = extremely good (annex 1) To improve

quality assessment, the hedonic classification described in chapter 5 has been applied for the

1999 harvest; the experimental procedure described in chapter 5 was strictly followed.

7.2.2.2 Sensory panel

The sensory panel consisted of 10 to 15 semi-trained subjects. The panel rated the different

parameters on a 1 to 9 scale (e.g. 1 =

very weak aroma intensity and 9 =

very strong aroma

intensity). The same procedure as described in chapter 5 was used.

The subjects were asked to rate the following sensory attributes: odor, aroma, sweetness,

acidity, skin hardness, flesh firmness, juiciness, mealiness and to give their overall

appreciation (annex 3).



Fresh intact tomatoes and apricots (400 ± lg) were carefully placed in a 6L headspace flask

with wide opening (NS 160/100). For the measurements on puree, 100 ± lg of the

homogenate were spread out into a crystallizing dish (10cm diameter, 3cm high) which was

placed in the headspace flask.

CHAPTER 7 78

The analyses were carried out using the same types of SPME fibers as described in chapter 5

and the experimental procedure was strictly followed.

7.2.3.2 Identification and quantification ofvolatile compounds in tomatoes and apricots

The volatile compounds of tomatoes and apricots were extracted by SPME (CAR/PDMS) and

identified and quantified by GC. The volatiles present in the headspace (250 mL) were

analyzed using the same procedure as described in chapter 5.

Identification was performed by a combination of Kovats retention indices and a GC-MS

library (Flavornet, Geneva, USA). Some components were identified by comparison of

retention time and mass spectra with authentic substances. The following reference substances

were used: hexanal, butyl acetate, cw-3-hexenol, 3-methyl-l-pentanol, /r<ms-2-hexenol,

hexanol, 2-heptanone, butyl butanoate, hexenyl acetate, 2,6-dimethyl-6-hepten-2-ol, 1-

octanol, propyl hexanoate, linalool, isobutyl hexanoate (Fluka); isoamyl acetate, trans-

2,/r<ms,-4-decadienal, (Aldrich, Milwaukee, USA); dimethyl disulfide, /rara-2-hexenal,

methional, 6-methyl-5-hepten-2-one, n-hexenyl propanoate, butyl hexanoate, pentyl

hexanoate, hexyl hexanoate, cc-ionone, y-decalactone, (Givaudan-Roure, Dübendorf,

Switzerland). The same procedure was adopted as described in chapter 5.

7.2.3.3 Determination oftotal sugar content and total acidity

200 ± 1 g of tomato and apricot puree were used for these analyses. Total sugar content

(°Brix) was determined using a refractometer (Atago, PR-1, Tokyo, Japan). pH and total

acidity were measured with a titrator (Mettler DL 25, Mettler-Toledo, Greifensee,

Switzerland). For determination of total acidity 10 ± 0.1 g of sample were titrated to pH 8.0

using 0.1 M NaOH. The titrated volume (mL) corresponds directly to total acidity expressed

as g/L citric acid.


The firmness of the fruits (100± lg) was determined either using a penetrometer (PNR 20

Benchtop Loud & Tensile Tester, Petrolab Co., Latham, USA) fitted with a round or a conic

head or a Kramer's shear cell operated by a shear test machine (VersaTest+Advanced Forces

Gauge, Memesin, Briitsch & Riiegger, Zurich, Switzerland). ). The device speed was set at

CHAPTER 7 79

250 mm/min. The fruits were divided into four parts prior to measurements which were

performed in triplicate, at ambient temperature.

7.2.3.5 Conductivity and mineral components analysis

The conductivity was measured directly on the tomato and apricot paste using a Metrohm 660

conductimeter (Metrohm AG, Herisau, Switzerland). For the quantitative analysis of Na, K,

Mg, and Ca, the puree was first centrifuged at 30'000xg for 10 min at room temperature.

Aliquots of the supernatant were analyzed by inductively coupled plasma emission

spectroscopy (ICP-ES 400, Perkin Elmer, Palo Alto, USA).


The Statview® program (Abacus concepts Inc., Berkeley, USA) was used for the analysis of

variance (ANOVA). Significant differences in instrumental measurements among samples

were determined by PLSD (protected least significant difference) with P <0.05. Where the

test of normality failed, the non-parametric test was applied to the individual panel scores for

every investigated intensity criteria and then transformed into ranking numbers. The non-

parametric test was processed by the Kruskall & Wallis test (P <0.05). The Statbox program

(Grimmer Logiciels Corp., Paris, France) was used for the Pearson's correlation (P <D.05)

and the principal component analysis (PCA) was carried out to identify the interdependence

between different variables of sensory, instrumental and chemical data.


The aim of this study was to evaluate the applicability of the quality assessment system

developed for strawberries (chapter 5) on tomatoes and apricots. The quality of tomatoes was

investigated through the years 1997, 1998 and 1999, whereas the experiments with apricots

were limited to 1999 only. In a first step, the sensory panel was used to define quality

attributes. Then, samples were judged by consumers and the relationship between sensory and

instrumental data was investigated. Finally, a model for quality assessment of tomatoes and

apricots was proposed.

CHAPTER 7 80

7.3.1 Quality assessment of tomatoes

7.3.1.1 Sensory evaluation

A sensory panel was used to set up quality descriptors as outlined in 7.2.2.2. 149 tomato

samples out of 28 cultivars were used for the identification of the most important quality

attributes. Because the normality test failed (large variance of the results), the non-parametric

test was used. In 1997 the objective of the panel was to define descriptors for the sensory

quality of tomatoes (Table 7.1). Aroma, sweetness, skin hardness as well as flesh firmness,

juiciness and mealiness were shown to be statistically relevant, whereas the attributes

herbaceous odor and salty taste were not relevant (low significance: P >0.05). Although

mealiness turned out to be an important quality critérium, its determination was unfortunately

not possible because of the lack of an instrumental device.

Table 7.1 Comparison of tomato samples by the sensory panel

harvest year

quality descriptors 1997 1998 1999

(49 samples) (60 samples) (40 samples)

£ odor NS NS 0.021 SI

o herbaceous NS NE NE

aroma 0.003 SI 0.004 SI 0.001 SI

-SJ sweetness 0.001 SI 0.001 SI 0.001 SI

S acidity NS 0.001 SI 0.001 SI

saltiness NS NE NE

flesh firmness 0.001 SI 0.001 SI 0.001 SI

S skin hardness 0.001 SI 0.001 SI 0.001 SI

& juiciness 0.001 SI 0.028 SI 0.001 SI

mealiness 0.029 SI NS 0.043 SI

|^ overall

^ appreciationNE 0.001 SI 0.001 SI

P values of Kruskall & Wallis, SI: Significant > 95% and NS: Not Significant < 95%, NE:

Not Evaluated. *HT: Hedonic Test

Most of the descriptors found to be significant in 1997 have been confirmed during the 1998

and 1999 campaigns (Table 7.1). The results allowed to define "aroma", "sweetness", "skin

hardness", "flesh firmness" and "juiciness" to be significant attributes to describe the quality

of tomatoes.

Consumer tests carried out in 1998 and 1999 made clear that the overall appreciation by a

hedonic test was highly significant (Table 7.1). Because of the large number of fruits needed

CHAPTER 7 81

for the consumer tests, we decided to abstain from a complete covering of this type of sensory

evaluation. Instead, the sensory panel was asked to give an overall appreciation of the fruits.

Out of the defined sensory descriptors, only "juiciness" correlated significantly with the

overall appreciation over a two-years' period (P <0.01, r = 0.54 for 1998 and r = 0.68 for

1999, respectively).

The overall appreciation by the consumers correlated well with a few of the quality attributes

established by the sensory panel, but for most of the attributes no significant correlation was

found (Table 7.2). The heterogeneity of the fruit samples was thought to be responsible for

these results. Nevertheless, consumer tests have been regarded to be appropriated to give

reliable information on the quality of tomatoes.

Table 7.2 Correlation between overall appreciation by the consumers and the quality

descriptors defined by the sensory panel

correlation coefficient (30 samples)

descriptors 17.6.98 29.7.98 22.6.99 13.7.99 13.8.99 31.8.99

odor 0.78* NS NS NS NS 0.90*

aroma NS NS NS NS NS NS

sweetness NS 0.98* 0.84* NS NS NS

acidity NS NS NS NS NS NS

hardness NS NS NS NS NS NS

mealiness NS NS -0.74* NS NS NS

firmness NS -0.78* NS NS NS NS

juiciness NS NS 0.85* NS NS NS

* significant with P <0.05, NS = Not Significant

Although the reproducibility of the measurements was low, odor and sweetness were judged

to be important quality descriptors for tomatoes. It is of particular interest to point out that

both descriptors can be determined by instrumental methods (total sugar content and amount

of total volatile compounds).

7.3.1.2 Correlation between sensory and instrumental data

Several chemical and physico-chemical parameters were measured by instrumental methods.

The sugar content (°Brix), pH, total acidity, cations (Na, K, Ca, Mg) and firmness data were

obtained for tomatoes of all harvests (1997, 1998 and 1999). The amounts of total volatile

compounds were determined for the fruits of the 1998 and 1999 seasons.

CHAPTER 7 82

Relationships between data obtained by consumer tests and instrumental methods have been

established. A good correlation was found between total volatile compounds (TV) and the

overall appreciation (Fig. 7.1A). Using the SPME fibers CW/DVB and PDMS correlation

factors of r = 0.82 and r = 0.65 were determined at the P <0.05 significance level.

Experiments carried out with fruits from the 1999 harvest gave a much lower correlation (r =

0.11 and r = 0.25 respectively, P <0.05). On the other hand, the total sugar content correlated

significantly (P <0.05) with the overall appreciation by the consumers (r = 0.45) in the 1999

campaign (Fig. 7.IB). In the 1998 harvest, the correlation was less evident (r = 0.22). The

difficulty to reproduce the correlation was probably due to a higher heterogeneity of fruit

samples in total volatile compounds and total sugar content (°Brix).

A) B)

O)

"O O)

i£oQ. CDE >O Q° 50) >

5 Ü

fi-°

> CO

CO CO

O S~

a0-

1.00

0.80

0.60

0.40

0.20

5.8

g 5.5

o

~ 5.2

o 4.9o

§> 4.6 -

CO

S 4.3-Io

4.0

5 6


5 6


7

Fig. 7.1 Relationship between (A) total volatile compounds and overall appreciation, (O)CW/DVB and () PDMS given by the consumer test for the 1998 harvest, and (B) total sugar

content (°Brix) and overall appreciation given by the consumer test in 1999 ()

Table 7.3 summarizes the results of the comparison between the consumers appreciation and

the instrumental data.

CHAPTER 7 83

Table 7.3 Correlation between the overall appreciation by consumers and instrumental

analysescorrelation coefficient

Instrumental1761998 29.7.1998 22.6.1999 13.7.1999 3.8.1999 31.8.1999

data

total sugar

content

total aciditypH

penetrometerKramer's

shear cell

conductivityK content

Na content

Mg content

Ca content

total volatiles

compoundsCAR/PDMS

PDMS/DVB

CW/DVB

PDMS

PA

CAR/PDMS/

DVB

NS

NS

NS

NS

NA

NS

NS

NS

NS

NS

NS

NS

0.91*

0.81*

NS

NA

0.89"

NS

NS

NS

NA

NS

NS

NS

NS

NS

NS

NS

NS

NS

NS

NA

0.68" 0.7611 0.78* 0.68*

NS NS NS 0.67*

NS NS NS NS

NA NA NA NA

NS NS NS NS

0.72* NS NS NS

NS NS NS NS

0.81* NS NS 0.67*

NS NS NS NS

NS NS NS NS

NS NS NS NS

NS NS NS NS

0.90* NS NS NS

NS NS NS NS

NS NS NS NS

NA 0.63* NS NS

*significant at P <0.05, NA = Not Analyzed, NS = Not Significant

The high correlation between the total sugar content and the overall appreciation on the one

side and between the amount of total volatile compounds (measured with some of the SPME

fibers) and the overall appreciation on the other hand, led to the conclusion that the two

attributes "sweetness" and "aroma" are determinant for the quality of tomatoes.

7.3.1.3 Hedonic classificationfor the assessment oftomato quality

The main problem in the development of a model for the assessment of the quality of fruits

was the heterogeneity of the fruit samples, as demonstrated in a previous work (see chapter 4).

Introduction of the hedonic classification (see chapter 5) successfully solved this problem.

Indeed, the same fruit sample could be analyzed by instrumental methods and by the

consumer, what made a direct comparison of the results possible.

CHAPTER 7 84

Total amount ofvolatile compounds

The hedonic classification showed a good correlation between the total amount of volatile

compounds and the consumers' overall appreciation for nearly all SPME fibers used (P <0.05,

r = 0.87-0.98), except for the PA fiber (P <D.05, r = 0.10) when compared to correlations of r

= 0.10-0.25 prior to classification. Only the CAR/PDMS fiber showed a linear relationship

between the consumer appreciation and total volatile compounds (r = 0.98) using hedonically

classified tomatoes (Fig. 7.2).

1.00

(/>

TJ •-—v

r O)~i .*:

o mu.

EE

ot » C/3

n>:>

m

Q0_

o r?-><

re oo

0.80

0.60

0.40

0.20

0.00

123456789


Fig. 7.2 Relationship between total volatile compounds extracted by a CAR/PDMS fiber of

hedonically classified tomato samples and consumer appreciation for four harvest dates

In spite of good correlations obtained with other fibers, the large variance of the results

increased the difficulty to differentiate between tomatoes of different classes of quality

(values not shown).

Tomato aroma compounds

The aroma of the tomato is composed of a large amount of substances belonging to different

classes of chemicals such as esters, alcohols and carbonyl compounds [4, 6, 9-11]. These

substances contribute to the fruity and green notes (herbaceous odor) of tomatoes and were

identified and quantified by GC-MS and GC-FID as described in 7.2.3.2. Taking into account

the results obtained for total volatile compounds, where it was shown that the different types

of SPME fibers adsorbed the same volatiles, however in different amounts, the GC-analyses

were carried out with one SPME fiber type only. The CAR/PDMS fiber was chosen because

CHAPTER 7 85

of its good differentiation ability between the scores 1 to 3 and 7 to 9 in the overall

appreciation (Fig. 7.2).

As expected, summing up the peak areas measured by GC-FID gave a higher total amount of

volatile compounds when compared to the results obtained with the method used to determine

the total volatile compounds. GC-FID was clearly more sensitive than the measurement of

total volatile compounds.

A weak correlation was obtained between the sum of volatile compounds determined by GC-

FID and the consumer overall appreciation (r = 0.46). However, a much better correlation (r =

0.98) was obtained between the total amount of volatile compounds and the consumers'

appreciation.

Nevertheless, it can be stated that esters determined by GC-FID contribute essentially to the

overall appreciation of tomatoes (r = 0.88). Among the ester group, butyl hexanoate (r = 0.71)

was shown to play a major role for the aroma of tomatoes. In the group of carbonyl

compounds /raws-2-hexenal (r = 0.74), 6-methyl-5-hepten-2-one (r = 0.58), and geranyl

acetone (r = 0.84) seem to be important. Finally the sulfur compound 2-isobutylthiazole (r =

0.83) also showed a good correlation with the consumers' appreciation.

Total sugar content

A very strong relationship (r = 0.98; P <0.05) between total sugar content (°Brix) after

hedonic classification of the samples and consumer appreciation was established as shown in

Fig. 7.3. Here again the advantage of the hedonic classification was evident. Without hedonic

classification the correlation was much lower (r = 0.43).

5.6 -i

—.

X

1_

CÛ 5.3 -

•s

^—'

c

CD 5-c

oÜ

1_

CO4.7 -

enZ3

to

44 -

m-•—»

o

123456789


Fig. 7.3 Relationship between total sugar content (°Brix) and consumer appreciation after

hedonic classification

CHAPTER 7 86

Total acidity

A good correlation was also found between the consumer ratings and total acidity (r = -0.90)

when the tomatoes were hedonically classified. Without hedonic classification r = 0.20 was

obtained.

On the whole, the hedonic classification of the samples allowed to enhance substantially the

correlation between instrumental data and the consumers' overall appreciation of tomatoes.

7.3.1.4 Development ofa modelfor the assessment oftomato quality

Based on these results we decided to apply the model developed for assessment of strawberry

quality (chapter 5) to tomatoes as well. Three quality levels, "bad", "medium" and "good"

were fixed. The average appreciation for "bad" samples was 2 (range: 1-3), for "medium"

samples 5 (range: 4-6), and for "good" samples 8 (range: 7-9). In Table 7.4 intervals and limit

values for the different quality attributes as well as sample distribution are given.

Table 7.4 Sample distribution, limit values (and intervals) for total volatile compounds

(mg/kg) and total sugar content (°Brix) for the three classes used for assessment of the qualityof tomatoes

quality class

instrumenta data "bad" "medium" "good"VI Vi 23.6.1999 30.3 39.4 30.3

73-« e

14.7.1999 8.3 47.7 44.0

to C £U 03 M

5.8.1999 20.0 53.7 26.3

ci° 2.9.1999 17.1 45.2 37.7

•Ö £ 1999 average 18.8 46.2 35.0

uM M 23.6.1999 0.22a (0.18-0.26) 0.38ab(0.21-0.66) 0.53d (0.35-0.74)

cä ci M_ -tri Li

14.7.1999 0.09a (0.00-0.31) 0.46b (0.38-0.57) 0.62c (0.58-0.68)

tal

voompo AR/P] (mg/1 5.8.1999 0.40a (0.36-0.55) 0.44ab (0.39-0.52) 0.72c (0.56-1.05)

2.9.1999 0.27a (0.00-0.49) 0.44ab (0.34-0.54) 0.52b (0.46-0.60)a ° u 1999 average 0.26a (0.00-0.55) 0.43b (0.21-0.66) 0.60c (0.35-1.05)

—23.6.1999 4.4a(4.1-4.7) 4.9b (4.6-5.1) 5.4C(5.1-5.6)

SP c 14.7.1999 4.6a (4.3-5.0) 5.0b (4.9-5.1) 5.1b(5.1-5.2)CO -4-»

—c 5.8.1999 4.5a (4.4-4.7) 4.9b (4.7-5.1) 5.2C(5.1-5.3)

c3 O

o °2.9.19994.3a(4.2-4.4)4.6b(4.4-4.7)5.Ie(4.9-5.3)1999average4.5a(4.1-5.0)4.9b(4.4-5.1)5.2e(4.9-5.6)levelofsignificance5%;differentlettersinalinemeansignificantdifferencesDeterminationsoftheamountoftotalvolatilecompoundsusingtheCAR/PDMSfiberandofthetotalsugarcontentwouldallowtopredictthequalityoftomatoesatharvest.Bymeasuringtheamountoftotalvolatilecompoundsandthetotalsugarcontentitwaspossibletopredict

CHAPTER 7 87

the quality of tomatoes at harvest. Overlapping between average values for "bad", "medium"

and "good", samples were observed for all other instrumental parameters.

Nevertheless, it can be stated that the model developed for the assessment of strawberry

quality can be used for tomatoes as well, although the results were not as convicing as for

strawberries. Further work remains to be done to improve the system.

7.3.2 Quality assessment of apricots

The hedonic classification was also applied on apricot, sorted out in three quality classes

"bad", "medium" and "good".

7.3.2.1 Assessment offive apricot cultivars by consumer tests

Five cultivars "Jumbo", "Fantasme", "Bergeron", "Luiset", and "Tardifde Tain" harvested at

different stages of maturity were analyzed in 1999 by consumer preference tests. Results

obtained were treated by PCA. It was interesting to note that 77% of the variability could be

explained by the two principal components (Fig. 7.4). From this figure it can be clearly seen

that the first principal component discriminated the consumer scores and cultivars tasted at a

specific ripening stage.

4-

08

0note4

06

Jumbo (ripe)

nnote9 Jumbo (over ripe)

0

note8° Fantasme (ripe) 02

0

note7_

Bergeron (rip

note2

Luiset (over ripe) °

Luiset (pre-ripe)

e) Jumbo(pre-ripe) onote3

5 -1 -0 5 I 05 1 1

Fantasme (pre-ripe)B Tardif detain (half ripe)

Bergeron (pre-ripe)

Tardif detain (ripe)-0 6

Luiset (ripe)

O note6

-0 8

1_

0note5

Factor 1 (53.8%)

Fig. 7.4 Principal component analysis of different apricot cultivars tasted by consumers at

different ripening stages (pre-ripe, ripe and over-ripe) in the 1999 season. (O) consumer

scores and () cultivars tasted at specific ripening stage

CHAPTER 7

The results clearly demonstrated that the first principal component analysis made it possible

to discriminate between different cultivars. "Jumbo" "Fantasme" and "Bergeron" obtained

the best-appreciated score. The consumers disliked the non-fully ripe fruits.

7.3.2.2 Hedonic classification for the assessment ofapricot quality

Total amount ofvolatile compounds

A good and statistically significant (P <0.05) correlation was found between the total amount

of volatile compounds and the consumers' overall appreciation for the SPME fibers PDMS (r

= 0.93), CAR/PDMS/DVB (r = 0.87) and CAR/PDMS (r = 0.80). As an example, the

relationship between total volatile compounds (mg/kg) extracted by the PDMS (r = 0.93) fiber

and the overall appreciation by the consumers using hedonically classified apricots is shown

in Fig. 7.5.

1.50-1

8 1.40-

o d5 130- i.

X

CL _*

E "B> 1.20-R E

« «1-1o- 1 \ y^\

J> w *»<A\ 1is^ 10°-*sfT

5 cl 0.90 - J: X /1%.

5 0.80 -

X t/-4—*

i Hf

0.70 - -lir—r^n .... .

1 1 —1—....,,, .

123456789


Fig. 7.5 Relationship between total volatile compounds extracted by a PDMS fiber of

hedonically classified apricots samples and consumer appreciation for 2 harvest dates:

22.7.1999 (x) and 10.8.1999 (D)

Two curve shapes were observed for the two sample lots obtained at different harvest dates;

this could be explained by the presence of more pre-ripe apricots in the 22.7.1999 harvest

date.

Total sugar content

A very strong relationship was established between total sugar content (°Brix) with hedonic

classification of the samples and consumer appreciation as shown in Fig. 7.6 (r = 0.90; P

CHAPTER 7 89

<0.05). In this case, the differences between the two harvest dates were not significant and

had no consequence on the shape of the curves.

14.0 -i

X

m 13.5 -

?_•

g 130 "

y./rc

8 12.5 -

i— aj*g> 12.0 -

P3

X

« 11-5 - u

-4-« Yo*"'

11 n .

I I.U 1 1 i i i i i

1 2 3 4 5 6 7 8 9


Fig. 7.6 Relationship between total sugar content (°Brix) of hedonically classified apricots

samples and consumer appreciation for 2 harvest dates: 22.7.1999 (x) and 10.8.1999 (D)

Total acidity play a minor role compared to the sugar content. The correlation was r = 0.53

between the consumer ratings and total acidity and become stronger with the ratio total

acidity/°Brix (r = -0.92).

Texture analysis

A good correlation (r = -0.78) between texture data, as measured by the Kramer's shear cell,

and the overall appreciation after hedonic classification was obtained as shown in Fig. 7.7.

2000

1600

z

^1200v>

(ü

| 800

v=

400

—i—i—i—i—i—i—i—i—i

123456789


Fig. 7.7 Relationship between firmness (Fmax) of hedonically classified apricot samples and

consumer appreciation for 2 harvest dates: 22.7.1999 (x) and 10.8.1999 (D)

CHAPTER 7 90

In contrast to the results obtained with tomatoes, firmness played an important role in the

appreciation of the quality of apricots.

7.3.2.3 Development ofa modelfor the assessment ofapricot quality

The hedonic classification allowed us to obtain a clearer distinction between the 3 different

quality classes. A model for assessment of apricot quality can therefore be proposed (Table

7.5).

Table 7.5 Sample distribution, limit values (and intervals) for total volatile compounds

(mg/kg), total sugar content (°Brix) and firmness (N) for the three classes used for assessment

of the quality of apricots

quality class

instrumental data "bad" "medium" "good"

»! T3 -=:

22.7.1999 20.1 34.9 45.0

arve s

an ;amp 11.8.1999 4.4 37.6 58.0

A- 4-> Cj_

•go 1999 average 11.6 36.4 52.0

vs -a„^

« CM DO

22.7.1999 0.87a (0.72-1.04) 1.12b (0.97-1.38) 1.23b (0.95-1.38)

totalvol

compou PDM (mg/k 11.8.1999

1999 average

0.25a (0.00-0.82)

0.84a (0.00-1.04)

1.04b (0.81-1.22)

1.08b (0.81-1.38)

1.17b (1.06-1.29)

1.21e (0.95-1.38)

£3^

22.7.1999 11.7a (11.3-11.9) 12.1b (11.7-12.4) 13.2e (12.8-13.6)3 <5CO *^

_C

03 O

11.8.1999 11.7a (11.5-11.9) 12.2b (11.9-12.3) 12.9e (12.6-13.4)

3 °

1999 average 11.7a (11.3-11.9) 12.2b (11.7-12.4) 13.1e (12.6-13.6)

22.7.1999

1446.63

(1037.0-1838.1)

828.4b

(670.2-1026.1)

532.2e

(363.4-704.1)

CO

OC

e

IS

11.8.1999

905.2a

(800.0-1053.7)

600.3b

(545.4-684.3)

660.5b

(542.5-795.1)

1999 average

1311.23

(800.0-1838.1)

707.7b

(545.5-1026.1)

596.4b

(363.4-795.1)

level of signi Icance 5%; different letters in a line mean significant differences.

With respect to the determination of total volatiles only the PDMS fiber allowed to

distinguish between the 3 quality classes. Nevertheless, the total sugar content was shown to

be a very good parameter to distinguish between the 3 quality levels, whereas firmness allow

a distinction between "bad" and "medium-good" quality.

CHAPTER 7 91

For the others,data obtained by instrumental methods overlapping between average values for

"bad", "medium" and "good" quality samples was always observed (values not shown).

7.4 CONCLUSIONS

Aroma and sweetness were shown to be the most important quality attributes for tomatoes and

apricots. In contrast to the results obtained with tomatoes, firmness allowed to reject the "bad"

apricots.

A hedonic classification of the samples allowed us to significantly improve the correlation

between instrumental data and consumer overall appreciation and enabled us finally to

propose a model for the assessment of the quality of tomatoes and apricots. Multiple variable

analysis enabled us to discriminate between quality classes. For the determination of the

amount of total volatile compounds the appropriated SPME fiber had to be used. For tomatoes

the CAR/PDMS fiber and for apricots the PDMS fiber allowed to distinguish between the 3

quality classes.

The total sugar content was shown to be a very good parameter to distinguish between the 3

quality levels of both fruits, whereas texture measurements (firmness) allowed a distinction

between the "bad" samples and the other classes of apricots.

7.5 ACKNOWLEDGMENT


Valais for their financial support in the frame of this COST 915 action ("Improvement of

quality of fruit and vegetables, according to the needs of the consumers"). We also thank Mr.

André Granges, Mr. Jacques Rossier and the Federation of Migros Cooperatives in Bussigny

for supplying fruit samples and setting up consumer tests.

7.6 REFERENCES

[1] Resurreccion, A.V.A., Shewfelt, R.L. (1985). Relationships between sensory attributes

and objective measurements of postharvest quality of tomatoes. J.Food Sei. 50, 1242-

1245.

CHAPTER 7 92

[2] Porretta, S. (1993). Analysis of sensory and physicochemical data on commercial

tomato puree with pattern recongnition techniques. Z Lebensm. Unters. Forsch. 197,

531-536.

[3] Parolari, R., Virgili, R., Bolzoni, L., Careri, M., Mangia, A. (1992). Analysis of sensory

and instrumental data on apricot purees with pattern recognition techniques. Anal.

Chim. Acta 259, 257'-265.

[4] Petro-Turza, M. (1987). Flavor of tomato and tomato products. Food Rev. Int. 2, 311-

353.

[5] Kavanagh, E.E., McGlasson, W.B. (1983). Determination of sensory quality in fresh

market tomatoes. CSIRO Food Res. Qual. 43, 81-89.

[6] Kader, A.A., Moris, L.L., Stevens, M.A., Albright-Holten, M.A. (1978). Composition

and flavor quality of fresh market tomatoes as influenced by some postharvest handling

procedures. J. Amer. Soc. Hort. Sei. 103, 6-13.

[7] Stevens, M.A., Kader, A.A., Albright-Holten, M.A., Algazi, M. (1977). Effect of fruit

ripeness when picked on flavor and composition in fresh market tomatoes. J. Amer. Soc.

Hort. Sei. 102,724-731.

[8] Buttery, R.G., Seifert, R.M., Guadagni, D.G.; Ling, L.C. (1971). Characterization of

additional volatile components oftomato. J. Agric. Food Chem., 19, 524-529.

[9] Buttery, R.G., Teranishi, R., Ling, L.C. (1987). Fresh tomato aroma volatiles - A

quantitative study. J. Agric. Food Chem. 35, 540-544.

[10] Buttery, R.G., Teranishi, R., Ling, L.C, Flath, R.A. (1988). Quantitative studies on

origins of fresh tomato aroma volatiles. J. Agric. Food Chem., 36, 1247-1250.

[11] Ho, CT., Ichimura, N. (1982). Identification of heterocyclic-compounds in the volatile

flavor of fresh tomato. Lebensm. Wiss. Technol. 15,340-342.

[12] Ulrich, D., Krumbein, A., Rapp, A. (1997). Analysis of aroma compounds in

strawberry, sour cherry and tomato by gas chromatography after solid phase micro-

extraction. Dtsch. Lebensm. Rundsch. 93, 311-316.

[13] Linforth, R.S.T., Ingham ,K.E., Taylor, A.J. (1996). Time course profiling of volatile

release from foods during the eating process. Taylor, A.J., Mottram, D.S. (eds.). Flavour

science: recent developments. Elsevier Science Publishers, Cambridge, p. 361-368.

[14] Buttery, R.G., Teranishi, R., Ling, L.C, Turabaugh, J.G. (1990). Quantitative and

sensory studies on tomato paste volatiles. J. Agric. Food Chem. 38, 336-340.

[15] Lo Voi, A., Impembo, M., Fasanaro, G, Castaldo, D. (1995). Chemical characterization

of apricot puree. J. Food Comp. Anal. 8, 78-85.

CHAPTER 7 93

[16] Gherardi, S., Poli, M., Bigliardi, D., (1978). Caractteristiche analitiche di cultivar di

albicocca. Pear espesca di diversa provenienza. Nota I. Ind. Conserve 53, 288-290.

[17] Blandau, R. Burkhard, A., Faethe, W., Götz, M., Greiner, G., Kleinnau, HJ.,

Olschimke, D., Reinhard, C, Wallrauch, S. (1985). Richtwerte und

Schwankungsbreiten bestimmter Kennzahlen (RSK-werte) für Aprikosenmark (-saft).

Flüss. Obst. 52, 644-646.

[18] Monastero, E., Deutsch, W., Ellero, M. (1992). Caratterizzazione di albicocche con il

contributo dell'analisi degli amminoacidi. IndAliment. 30, 817-824.

[19] Tang, C. S., Jenninigs, W.G. (1968). Lactonic compounds of apricot. J. Agric. Food

Chem. 16, 252-254.

[20] Rodriguez, F., Seek, S., Crouzet, J. (1980). Constituants volatils de l'abricot variété

Rouge du Roussillon. Lebensm. Wiss. Technol. 13, 152-155.

[21] Chairote, G., Rodriguez, F. Crouzet, G. (1981) Components of apricots. J. Food Sei. 46,

1898-1901.

[22] Bernreuther, A., Schreier, P (1991). Multidimensional gas chromatography/mass

spectroscopy: a powerful tool for the direct chiral evaluation of aroma compounds in

plant tissues. II. Linalool in essential oils and fruits. Phytochem. Anal. 2, 167-170.

[23] Guichard, E., Souty, M. Comparison of the relative quantities of aroma compounds

found in fresh apricot (Prunus armeniacd) from six different cultivars. Z. Lebensm.

Unters. Forsch A. 186, 301-307.

[24] Takeoka, G.R., Flath, R.A, Thomas, R.M., Teranishi, R., Guentert, M. (1990). Volatile

constituents of apricot [Prunus armeniacd). J. Agric. Food Chem. 38, 471-477.

[25] Parolari, G., Virgili, R., Bolzoni, L. (1992). Analysis of sensory and instrumental data

on apricot purees with pattern recognition techniques. Anal. Chim. Acta 259, 257-265.

[26] Bartolozzi, F., Bertazza, G., Bassi, D., Cristoferi, G. (1997). Simultaneous

determination of soluble sugars and organic acids as their trimethylsilyl derivatives in

apricot fruits by gas-liquid chromatography. J. Chromatogr. A 758, 99-107.

[27] Katona, Z. F., Sass, P., Molnar-Perl, I. (1999). Simultaneous determination of sugars,

sugar alcohols, acids and amino acids in apricots by gas chromatography-mass

spectrometry. J. Chromatogr. A, 887, 91-102.

[28] Herrmann, K. (1990). Significance of hydroxycinnamic acid compounds in food. II.

Effects on the activity of food enzymes, alleged antithiamine effect, bitter taste,

pharmacologicaleffectivenessandtherapeuticuses,andsignificanceasantimutagenesandanticarcinogens.Chem.Mikrobiol.Technol.Lebensm.12,161-167.

CHAPTER 7 94

[29] Fernandez de Simon, B.F., Pérez-Ilzarbe, J, Hernandez, T., Gomez-Cordoves, C,

Estrella, I (1992). Importance of phenolic compounds for the characterization of fruit

juices. J. Agric. Food Chem. 40, 1531-1535.

[30] Garcia-Viguera, C, Bridle, P., Ferreres, F., Tomas-Barberan, A.T. (1994). Influence of

cultivar, maturity and processing on phenolic compounds of apricot juices and jams. Z

Lebensm. Unters. Forsch A. 199, 433-436.

[31] Möller, B., Herrmann, K. (1983). Quinic acid esters of hydroxycinnamic acids in stone

pome fruit. Phytochemistry 22, 477-481.

[32] Azodanlou, R., Darbellay, C, Luisier, J.L., Villettaz, J.C, Amadö, R. (1999).

Application of a new concept for the evaluation of the quality of fruit. In: Hägg, M.,

Ahvenainen, R., Evers, A.M., Tiililkkala, K. (eds.). Agri-food quality management of

fruit and vegetables. RSC, Cambridge, 266-270.

[33] Azodanlou, R., Darbellay, C, Luisier, J.L., Villettaz, J.C, Amadö, R. (1999). A new

concept for the measurement of total volatile compounds of food. Z Lebensm. Unters.

Forsch.. 208, 254-258.

[34] Azodanlou, R., Luisier, J.L., Villettaz, J.C, Amadô, R. (1998). Development of new

aroma measurement quantitative device for analysis in fruits. Travaux de Chimie

Alimentaire et d'Hygiène. 89, 650.

CHAPTER 8 95

CHAPTER 8

CONCLUSIONS AND OUTLOOK

The objective of this research work was to develop a methodology for evaluating fruit quality.

In order to achieve this goal we focused our efforts on sensory and instrumental analysis. As

far as the sensory approach is concerned, two different paths were followed. At first, sensory

attributes that are discriminant for quality were elaborated with a sensory panel. Secondly,

consumer tests were run to obtain hedonic information on the overall quality. With respect to

the instrumental approach, a new method to evaluate the amount of total volatile compounds

was developed. In addition, several parameters, such as total sugar content, total acidity,

identity and amounts of several groups of volatile compounds, and texture attributes such as

firmness were measured. Finally we tried to establish correlations between analytical and

sensory data in order to propose quality models for the different fruits analyzed.

Several cultivars of strawberries, tomatoes and apricots grown under different conditions were

analyzed both by sensory and instrumental methods.

The study clearly demonstrated aroma and sweetness to be the most important quality

attributes for the investigated fruits. Sweetness can easily be measured by determination of

the total sugar content by refractometry (°Brix), while aroma represents a complex fraction of

volatile compounds, interacting with each other and therefore very difficult to analyze. Within

this framework we put considerable efforts in developing a new concept to measure the total

volatile compounds. The system described proved to be highly appropriate to measure the

total volatile intensity and, by doing this, the quality of fruits.

In a second step, research was carried out to correlate sensory and instrumental data in order

to objectively measure the fruit quality. Unfortunately, the heterogeneity of the fruit batch and

the variability of the sensory judgements generally result in weak correlations.

CHAPTER 8 96

This problem was successfully solved by introducing the so called hedonic classification of

the fruit samples.

Hedonic classification of samples allowed to significantly improve the correlation between

instrumental data and consumer appreciation, and enabled us to develop a quality evaluation

system. The method is based on three quality classes (bad, medium, good) that have been

characterized by the limit values. Measurement of the total volatile compounds using the best

performing fiber (chapter 5 and 7), and determination of total sugar content was schown to be

generally sufficient to assess the quality of the fruits.

In the case of strawberries, the compounds that contributed significantly to the peak area

measured by the total volatile compound analysis were quantified and identified. Some of the

individual compounds were correlated to the consumer appreciation. In particular, esters were

found to contribute greatly to the sensory quality of strawberries.

The process of ripeness plays an important role in fruit quality. The amount of total volatile

compounds, total acidity, total sugar content and texture measurements (firmness) were used

to characterize the degree of ripeness of three strawberry varieties (chapter 6).

The new concept using the measurement of total volatile compounds enabled us to distinguish

between various stages of strawberry maturation. The CAR/PDMS SPME fiber was found to

be best suited to differentiate between the stages of ripeness. Most of the identified

compounds were esters, followed by aldehydes, and alcohols. The most abundant compounds

increased, while the concentration of certain aroma components known to be responsible for

the "green" notes, decreased during the maturation process until they became minor

components in mature strawberries.

In summary, the amount of total volatile compounds and the total sugar content (°Brix) have

proven to be very useful tools for objectively determining the quality of fruits. Sorting out

samples using the score obtained with the hedonic classification method, enabled us to

considerably strengthen the correlation between consumers' appreciation and instrumental

data.

In order to enhance the accuracy and the reliability of the quality model further work will be

required. This problem may be solved by changing some parameters such as headspace

equilibration time, sampling time, extraction temperature, fibers used for adsorption, etc.

CHAPTER 8 97

The results obtained by the sensory panel are subject to a wide variability, even with well-

trained assessors. To improve this inconvenience, regular, intensive and long lasting training

sessions have to be carried out, what is cost-intensive.

The consumer tests could be differently organized in terms of its composition. Setting up

homogeneous consumer groups in terms of age, sex, living standard, etc. would certainly give

interesting and probably more consistent results.

Finally, the homogeneity of the fruits to be investigated is of outstanding importance for the

assessment of their quality. Special attention has therefore to be paid to this very important

point. Research has to be carried out by hedonic classification of fruit samples, so that the

limit values of consumer acceptance validate the quality model.

The rapidity of the described method for total volatile analysis allows on-line analysis and

increases the daily sampling capacity. The applicability of the method for different types of

food could also be investigated.

ANNEXES

98

Annex

1

Hedonictest(t

rans

late

dfromFrench)

Tastethe

fruitandmarkyouroverallap

prec

iati

on

This

frui

tis...

Note

916*

525*

238*

246*

850*

467*

extremelygood

9D

DD

DD

D

verygood

8D

DD

DD

D

good

7D

DD

DD

D

rela

tive

lygood

6D

DD

DD

D

medium

5D

DD

DD

D

rela

tive

lybad

4D

DD

DD

D

bad

3D

DD

DD

D

verybad

2D

DD

DD

D

extremelybad

1D

DD

DD

D

Isthis

fruitaromatic?

916*

525*

238*

246*

850*

467*

yes

DD

DD

DD

no

DD

DD

DD

*samplecode

age:

15-25

D

25-40

D

40-60

D

60andmoreD

sex:

Male

DFemale

D

ANNEXES 99

Annex 2

Sensory panel form: Strawberry (translated from French)

Name : First name

Date:

Sample code:

Instructions: You are receiving fruit samples. The evaluation of odor, taste and

texture is required. Please judge the intensity of each sample without comparing it to

other samples.

I.Odor intensity

very weak medium very strong

odor D D D D D D D D D

2. Taste intensity


aroma D D

very weak

D D D

medium

D D D D

very strong

sweetness D D

very weak

D D D

medium

D D D D

very strong

acidity D D D D D D D D D

3. Texture characteristics


firmness D D

very weak

D D D

medium

D D D D

very strong

juiciness D D D D D D D D D

4. Please indicate your overall appreciation (hedonic test):

bad medium excellent

overall D D D D D D D D D

appreciation

ANNEXES 100

Annex 2 (cont.)

5. Please indicate the criterion that influenced most your decision (indicate one or

more criteria)

Odor Taste Texture characteristics

strawberry odor D aroma D firmness

fermented odor D sweetness D juiciness

acidity D fondant ^

bitter D flesh smoothness E

herbaceous D

fermented taste D

Please indicate here any perception, which were not indicated above:

Comments:

ANNEXES 101

Annex 3

Sensory panel: Tomato (translated from French)

Name : First name

Date:

Sample code:

Instructions: You are receiving fruit samples. The evaluation of odor, taste and

texture is required. Please judge the intensity of each sample without comparing it to

samples.

1 .Odor intensity


odor D D D D D D D D D

2. Taste intensity


tomato aroma D D D D D D D D D


sweetness D D D D D D D D D


acidity D D D D D D D D D

3. Texture characteristics


skin hardness D D D D D D D D D


flesh firmness D D D D D D D D D


juiciness D D D D D D D D


mealiness D D D D D D D D D

ANNEXES 102

Annex 3 (cont.)

4. Please indicate the criterion which influenced most your decision (indicate one

or more criteria)

bad medium excellent

overall D D D D D D D D D

appreciation

5. Please indicate the criterion that influenced most your decision (indicate one

or more criteria)

Odor Taste Texture characteristics

tomato odor D tomato D flesh firmness D

herbaceous aroma juiciness

sweetness D mealiness O

acidity D skin hardness ^

saltiness D

Please indicate here any perception, which are not indicated above:

Comments:

CURRICULUM VITAE 103

Curriculum Vitae

1965 Born in Tehran, Iran

1973-1977 Primary school in Tehran, Iran

1977-1986 High school in Tehran, Iran.

1986-1988 High school in Annemasse, France, graduating with Baccalauréat in

science

1989-1994 Bachelor of Science in Chemistry at the University of Geneva,

Switzerland.

1994-1996 Master of Science in Chemical Engineering at the University of

Geneva, Switzerland.

1997-2001 Ph.D. thesis at the Institute of Food Science, Swiss Federal Institute of

Technology, Zurich, Switzerland, in collaboration with the

Department of Food Technology/Chemistry and Biotechnology at

Ecole d'Ingénieurs du Valais, Sion, Switzerland and with the Swiss

Federal Research Station for Plant Production, Conthey, Switzerland.

Documents

New In Copyright - Non-Commercial Use Permitted Rights / License: … · 2020. 3. 26. · Acknowledgements I would like to thank Prof. Dr. Renato Amado, Dr. Jean-Claude Villettaz