Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
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
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 Materials and Methods 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
4.5 Acknowledgements 39
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
5.2.3.2 Identification and quantification of volatile compounds 44
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
5.3 Results and Discussion 46
5.3.1 Sensory evaluation 46
TABLE OF CONTENTS VI
5.3.2 Instrumental analyses 49
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
5.5 Acknowledgements 57
5.6 References 57
6. Changes in Flavor and Texture During the Ripening of Strawberries 59
6.1 Introduction 59
6.2 Materials and Methods 61
6.2.1 Fruit samples and sample preparation 61
6.2.2 Instrumental analyses 61
6.2.2.1 Determination of total volatile compounds 61
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
6.2.2.4 Texture analysis 62
6.2.3 Statistical evaluation 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
6.3.1.3 Identification and quantification of volatile compounds 65
6.3.1.4 Texture analysis 70
6.4 Conclusions 71
6.5 Acknowledgements 72
6.6 References 72
7. Objective Quality Assessment of Tomatoes and Apricots 74
7.1 Introduction 74
TABLE OF CONTENTS VII
7.2 Materials and Methods 76
7.2.1 Fruit samples and sample preparation 76
7.2.2 Sensory evaluation 77
7.2.2.1 Consumer tests 77
7.2.2.2 Sensory panel 77
7.2.3 Instrumental analyses 77
7.2.3.1 Determination of total volatile compounds 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
7.2.3.4 Texture analysis 78
7.2.3.5 Conductivity and mineral components analysis 79
7.2.4 Statistical evaluation 79
7.3 Results and Discussion 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
7.5 Acknowledgements 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.
Chromatogr. 648, 199-211.
[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.
Chromatogr. 15, 741-744.
[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 MATERIALS AND METHODS
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 85^im (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.
4.3 RESULTS AND DISCUSSION
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 MATERIALS AND METHODS
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.
5.2.3.2 Identification and quantification ofvolatile compounds
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).
5.3 RESULTS AND DISCUSSION
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.
5.3.1 Sensory evaluation
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.
5.3.2 Instrumental analyses
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
overall appreciation
-1 1 1 1 1
4 5 6 7 8
overall appreciation
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
overall appreciation
t r
56
7 8
overall appreciation
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-^A
2.0-
1.5-
1.0-
0.5- V
f*.
0.0- 1
4 5 6 7 8
overall appreciation
4 5 6 7 8 9
overall appreciation
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
overall appreciation
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
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
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.
6.2 MATERIALS AND METHODS
6.2.1 Fruit samples and sample preparation
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.
6.2.2 Instrumental analyses
6.2.2.1 Determination oftotal volatile compounds
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.
6.2.2.4 Texture analysis
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.
6.2.3 Statistical evaluation
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
6.3.1.3 Identification and quantification ofvolatile compounds
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].
6.3.1.4 Texture analysis
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- ^^<?\^-^ioo -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
The authors thank the Swiss Federal Office for Science and Education and the Canton of
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.
7.2 MATERIALS AND METHODS
7.2.1 Fruit samples and sample preparation
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 Sensory evaluation
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).
7.2.3 Instrumental analyses
7.2.3.1 Determination oftotal volatile compounds
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.
7.2.3.4 Texture analysis
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).
7.2.4 Statistical evaluation
The Statview® program (Abacus concepts Inc., Berkeley, USA) was used for the analysis of
variance (ANOVA). Significant differences in instrumental measurements 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.
7.3 RESULTS AND DISCUSSION
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
overall appreciation
5 6
overall appreciation
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
overall appreciation
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
overall appreciation
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
overall appreciation
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
overall appreciation
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
overall appreciation
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
The authors thank the Swiss Federal Office for Science and Education and the Canton of
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
very weak medium very strong
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
very weak medium very strong
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
very weak medium very strong
odor D D D D D D D D D
2. Taste intensity
very weak medium very strong
tomato aroma D D D D D D D D D
very weak medium very strong
sweetness D D D D D D D D D
very weak medium very strong
acidity D D D D D D D D D
3. Texture characteristics
very weak medium very strong
skin hardness D D D D D D D D D
very weak medium very strong
flesh firmness D D D D D D D D D
very weak medium very strong
juiciness D D D D D D D D
very weak medium very strong
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.