CHARACTERIZATION OF VOLATILE COMPOUNDS

IN SELECTED CITRUS FRUITS FROM ASIA

JORRY DHARMAWAN

(B.Appl.Sc. (Hons.), NUS)

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF CHEMISTRY

NATIONAL UNIVERSITY OF SINGAPORE

2008

Acknowledgments

This project could not have been completed without the support of Food Science and

Technology programme of the National University of Singapore and Firmenich Asia

Pte. Ltd. for endorsing and financing the project.

I am greatly indebted to the following supervisors and consultants who graciously lent me their technical expertise and encouragement throughout the project:

• Associate Professor Stefan Kasapis for his supervision and guidance in journal

publications, and for his advices, supports and encouragements.

• Mr. Philip Curran for his supervision and guidance with his expertise and

experiences in flavour industry.

• Associate Professor Philip J Barlow and Associate Professor Conrad O Perera for

their initial supervision in this research project.

Gratitude is also expressed for the following people for their contribution to the project:

• Dr. Martin J Lear and Ms. Praveena Sriramula from Department of Chemistry,

NUS, for their assistance in the synthesis of (Z)-5-dodecenal.

• Mr. Kiki Pramudya, Ms. Chionh Hwee Khim, Ms. Alison Tan, Ms. Yukiko, Ms.

Susan Chua and Ms. Feng Peiwen from Firmenich Asia Pte. Ltd. for their advices

and participation as panellists.

• Ms. Mia Isabelle and Mr. Xu Jia for their contribution as panellists.

i

• Ms. Cynthia Lahey, Dr. Novalina Lingga and Mr. Mark Teo from Shimadzu Asia

for their technical support in GC-MS.

• Mdm. Lee Chooi Lan, Ms. Lew Huey Lee and Mr. Abdul Rahaman bin Mohd

Noor for their continuous assistance whenever I need their lending hands.

• Mdm. Frances Lim and Ms. Joanne Soong from HPLC Lab, NUS for their

assistance in GC-FID.

• Mr. Don Hendrix and staff at Firmenich Citrus Centre for their assistance and

hospitality during my visit.

• Mr. Gerald Uhde and staff at Firmenich Geneva for their assistance and

hospitality during my visit.

• Ms. Daisy Lam from Firmenich Asia Pte. Ltd. for her assistance in administrative

matters

Special thanks are owed to the following people: My parents, Mr. Hendy Dharmawan and Mdm. Phang Kim Jin, and my beloved family, together with my brothers and sisters from the Indonesian group of Hope of God Church, Singapore for their prayer support and encouragement. My gratitude is also for those whose names cannot be mentioned one by one here but have helped me in different ways throughout the duration of my postgraduate study and without them, this research will not be able to be completed. Finally and most importantly, I would like to acknowledge God’s grace and help, which has been critical to the success and completion of this project.

‘His grace is sufficient for me, for His power is made perfect in weakness.’

ii

Table of Contents

Page

ACKNOWLEDGMENT i

SUMMARY vi

LIST OF TABLES vii

LIST OF FIGURES viii

LIST OF ABBREVIATIONS ix

LIST OF PUBLICATIONS AND PRESENTATION xi

CHAPTER 1: INTRODUCTION 1

CHAPTER 2: LITERATURE REVIEW 3

2.1. Citrus Fruits 3

2.1.1. Fruit morphology 3

2.1.2. Chemical composition 5

2.1.3. Uses of citrus fruits 8

2.2. Citrus Variety 12

2.2.1. General classification 12

2.2.2. Selected citrus cultivars from Asia 14

2.3. Citrus Flavour 17

2.3.1. Important volatile compounds in citrus flavour 19

2.3.2. Factors affecting citrus flavour 21

2.4. Flavour Research 24

2.4.1. Challenges in flavour research 24

2.4.2. Systematic approach in flavour research 26

References 35

iii

CHAPTER 3: CHARACTERIZATION OF VOLATILE COMPOUNDS

IN HAND-SQUEEZED JUICES OF SELECTED CITRUS FRUITS

FROM ASIA 55

3.1. Abstract 55

3.2. Introduction 56

3.3. Materials and Methods 57

3.3.1. Materials 57

3.3.2. Chemicals 57

3.3.3. pH, brix value and titratable acidity 58

3.3.4. SPME 59

3.3.5. Continuous liquid-liquid extraction 59

3.3.6. Gas Chromatograph-Flame Ionization Detector (GC-FID) 60

3.3.7. Gas Chromatograph/Mass Spectrometry (GC/MS) 60

3.3.8. Linear Retention Index 61

3.4. Results and Discussion 61

3.4.1. Chemical composition 61

3.4.2. Volatile compounds in citrus juices 63

References 75

CHAPTER 4: CHARACTERIZATION OF VOLATILE COMPOUNDS

IN PEEL OIL OF SELECTED CITRUS FRUITS FROM ASIA 80

4.1. Abstract 80

4.2. Introduction 81

4.3. Materials and Methods 81

4.3.1. Materials 81

4.3.2. Chemicals 82

iv

4.3.3. Gas Chromatograph-Flame Ionization Detector (GC-FID) 82

4.3.4. Gas Chromatograph/Mass Spectrometry (GC/MS) 83

4.4. Results and Discussion 83

References 90

CHAPTER 5: EVALUATION OF AROMA ACTIVE COMPOUNDS IN

PONTIANAK ORANGE PEEL OIL 92

5.1. Abstract 92

5.2. Introduction 92

5.3. Materials and Methods 94

5.3.1. Materials and chemicals 94

5.3.2. Gas Chromatograph/Olfactometry (GC-O) 95

5.3.2. Aroma Extract Dilution Analysis (AEDA), Relative Flavour

Activity (RFA) and Odour Activity Value (OAV) 95

5.3.4 Aroma reconstitution and omission test 96

5.4. Results and Discussion 97

5.4.1. Aroma active compounds of Pontianak orange peel oil 97

5.4.2. Odour Activity Value (OAV) and Relative Flavour Activity

(RFA) 104

5.4.3. Aroma reconstitution 110

5.4.4. Omission experiments 112

References 113

CHAPTER 6: CONCLUSION 118

CHAPTER 7: SUGGESTION FOR FUTURE WORK 120

APPENDIX 122

v

Summary

In this research, the characterization of volatile compounds in selected citrus fruits from Asia, namely Pontianak orange from Indonesia, Mosambi from India and

Dalandan from the Philippines has been carried out for their juices and peel oils.

Continuous liquid-liquid extraction with diethyl ether and Solid Phase

Microextraction (SPME) were utilized to extract the volatiles from the juices prior to analysis with Gas Chromatography (GC), while direct injection to the GC was done for the hand-pressed peel oils. Flame Ionization Detector (FID) and Mass

Spectrometer (MS) detector were used for quantitative and qualitative analysis respectively. There was a difference between juice and peel oil in the compounds characterized as the former contained more esters. Despite some differences, the profile of volatile compounds found in Mosambi was generally comparable to typical sweet orange whereas Dalandan’s profile resembled typical mandarin. On the other hand, Pontianak orange portrayed its unique citrus flavour profile.

Consequently, further investigation has been explored to unveil the key compounds in

Pontianak orange peel oil through a systematic approach. GC-Olfactometry (GC-O) was used to screen the potent odourants by using human nose as the detectors. Aroma

Extract Dilution Analysis (AEDA) technique performed was effective in revealing 41 aroma active compounds, which were dominated by saturated and unsaturated aldehydes. Lastly, aroma reconstitution and omission test were carried out to verify the findings by sensory evaluation of aroma models. The outcome suggested that (Z)-

5-dodecenal and 1-phenyl ethyl mercaptan were the significant contributors to the flavour of Pontianak orange.

vi

List of Tables

Table 3.1. Chemical composition of various orange juice cultivars 61

Table 3.2. Volatile compounds of freshly squeezed Pontianak orange,

Mosambi and Dalandan juices 64

Table 4.1. Volatile compounds of the peel oil of Pontianak orange,

Mosambi and Dalandan 84

Table 5.1. Aroma active compounds (FD ≥ 2) in Pontianak orange peel oil 99

Table 5.2. The Odour Activity Value (OAV) and Relative Flavour

Activity (RFA) of aroma active compounds in Pontianak

orange peel oil 105

Table 5.3. Potent odourants in Pontianak orange peel oil based on their

Odour Activity Values (OAV>2000) 108

Table 5.4. Potent odourants in Pontianak orange peel oil based on their

Relative Flavour Activity (RFA>6.5) 109

Table 5.5. Sensory evaluation for the aroma model of the Pontianak

orange peel oil as affected by the omission of compounds 113

vii

List of Figures

Figure 2.1. Section of citrus fruit (Ranganna et al ., 1986) 4

Figure 2.2. Pontianak oranges 15

Figure 2.3. Mosambi 16

Figure 2.4. Dalandan 17

Figure 3.1. Diagram for the isolation of headspace flavour compounds of

orange juice by SPME (Jia et al ., 1998) 59

Figure 5.1. Chromatogram (top) and aromagram (below) of aroma active

compounds of Pontianak orange peel oil 103

Figure 5.2. Comparative flavour profile analysis of Pontianak orange peel

oil and the reconstituted aroma model solutions based on all

available compounds (Formula 1), Relative Flavour Activity

(RFA; Formula 2) and Odour Activity Value (OAV; Formula

3) 111

viii

List of Abbreviations

AEDA Aroma Extract Dilution Analysis

DVB Divinyl benzene

ECD Electron Capture Detector

EI Electron Ionization

FD Flavour Dilution

FID Flame Ionization Detector

FPD Flame Photometric Detector

GC Gas Chromatograph

GC-FID Gas Chromatograph-Flame Ionization Detector

GC/MS Gas Chromatograph-Mass Spectrometry

GC-O Gas Chromatograph-Olfactometry

LRI Linear Retention Index

MNMA Methyl-N-methyl anthranilate

MS Mass Spectrometry

NIST National Institute of Standards and Technology

NPD Nitrogen-Phosphorus Detector

OAV Odour Activity Value

PDMS Polydimethylsiloxane

PLOT Porous-Layer Open Tubular

RFA Relative Flavour Activity

SAFE Solvent-Assisted Flavour Evaporation

SBSE Stir Bar Sorptive Extraction

SCOT Support Coated Open Tubular

ix

SDE Simultaneous Distillation/Extraction

SPME Solid Phase Microextraction

WCOT Wall-Coated Open Tubular

x

List of Publications and Presentation

1. Dharmawan J, Barlow PJ and Curran P. 2006. Characterization of Volatile Compounds in Selected Citrus Fruits from Asia. In: Bredie WLP and Petersen MA (eds). Flavour Science: Recent Advances and Trends. Proceedings of the 11 th Weurman Flavour Research Symposium held in Roskilde, Denmark on 21-24 June 2005. Amsterdam: Elsevier. p 319-322.

2. Dharmawan J, Kasapis S, Curran P and Johnson JR. 2007 Characterization of Volatile Compounds in Selected Citrus Fruits from Asia Part I: Freshly-Squeezed Juice. Flavour Fragr J 22 : 228-232.

3. Dharmawan J, Kasapis S and Curran P. 2007. Aroma Active Compounds of Pontianak orange Peel Oil ( Citrus nobilis Lour. var. microcarpa Hassk.). Oral Presentation at the 5 th Singapore International Chemistry Conference held in Singapore on 16-19 December 2007.

4. Dharmawan J, Kasapis S and Curran P. 2008. Characterization of Volatile Compounds in Selected Citrus Fruits from Asia Part II: Peel Oil. J Essent Oil Res 20 : 21-24.

5. Dharmawan J, Kasapis S and Curran P. 2008. Unveiling the Volatile Compounds of Citrus Fruit from Borneo. In: Hofmann T, Meyerhof W and Schieberle P (eds). Recent Highlights in Flavour Chemistry and Biology. Proceedings of the 8 th Wartburg Symposium held in Eisenach, Germany on 27 February – 2 March 2007. Garching: Deutsche Forschungsanstalt für Lebensmittelchemie. p 265-268.

6. Dharmawan J, Kasapis S, Sriramula P, Lear MJ and Curran P. 2009. Evaluation of Aroma Active Compounds in Pontianak Orange Peel Oil ( Citrus nobilis Lour. var. microcarpa Hassk.) by Gas Chromatography/Olfactometry, Aroma Reconstitution and Omission Test. J Agric Food Chem (in press – online access DOI 10.1021/jf801070r).

xi

Chapter 1 Introduction

As the most produced fruit crop in the world, citrus fruits are largely processed for their juice, one of the most important commodities, as well for their essential oil.

Citrus essential oils are mainly utilized as flavourings by a variety of food industries, especially for beverages, ice cream, confectionery and snacks production. Despite the fact that the United States of America and Brazil are the main producers of citrus fruits, the southeastern part of Asia is believed to be the place of origin of citrus fruits.

There are many varieties of citrus fruits in the region of Asia that have distinct flavour characteristics and are only consumed locally. Some of them have great potential to be further studied and their distinct aroma profiles elucidated in order to reveal specific compounds that contribute to their uniqueness.

Ample studies have been carried out in order to investigate the flavour compounds present in countless citrus cultivars. As the massive hybridization on a range of citrus cultivars brought about the uniqueness of its flavour, the scope of the research ranged from the most famous cultivars to the native ones. Still, not many studies are reported on those from Asia. In addition to the plethora of volatile compounds reported in

Citrus species, it is the intention of this research project to unveil the aroma profiles of three selected citrus varieties from Asia:

• Pontianak Orange ( Citrus nobilis Loureiro var. microcarpa Hassk.) from

Indonesia

1 • Mosambi ( Citrus sinensis Osbeck) from India

• Dalandan ( Citrus reticulata Blanco) from the Philippines

These citrus cultivars were found to be popular and well-liked by the locals in their origin countries as they possess unique flavour profiles. The results of this study are expected to lead to the better understanding of the science of citrus fruits, particularly

Asian cultivars, and also to contribute to the innovation and development in the food ingredients industries.

To achieve this objective, a systematic approach in flavour research was undertaken.

Volatile compounds in the juices and peel oils of the three Asian citrus cultivars were extracted by various extraction methods and were characterized by gas chromatography (GC). The aroma active compounds of Pontianak orange oil were further investigated as its flavour profile was found to be more unique. For this purpose, a GC equipped with an olfactometer (GC-O) was used, and involved a number of panellists. Finally, the results were verified by reconstituting aroma models and omitting the compounds deemed to be the key contributors (i.e. omission test).

2 Chapter 2 Literature Review

2.1. Citrus fruits

Citrus fruits have been cultivated for over 4000 years (Davies and Albrigo, 1994) and are the most produced fruit crops in the world (FAOSTAT). Citrus fruits belong to the family Rutaceae, in which the leaves usually possess transparent oil glands and the flowers contain an annular disk (Kale and Adsule, 1995). The place of origin of citrus fruits is believed to be south eastern Asia and these were subsequently brought to the

Middle East and Southern Europe, and further distributed to many other countries by the assistance of travellers and missionaries following the paths of civilisation

(Samson, 1980; Ruberto, 2001; Calabrese, 2002). The production of citrus fruits, particularly the sweet oranges, continues to show a tremendous growth with Brazil being the largest producer, followed by the United States of America; both sharing more than a third of total production of sweet oranges in the world (FAOSTAT).

2.1.1. Fruit Morphology

In general, citrus fruits are composed of 3 main sections ( Figure 2.1 ): a. The outer peel

The outer peel of citrus fruits is also known as flavedo due to the presence of

flavonoid compounds (Ortiz, 2002). It consists of the cells containing the

carotenoids, which give the characteristic colour to the fruits according to the

species or cultivar. The colour ranges from deep orange or reddish to light orange,

3 yellow or greenish. The carotenoid pigments are located inside the chromoplasts

in the flavedo (Kefford, 1955). The oil glands, which contain the citrus essential

oils, are also found in the flavedo. The glands are spherical in shape and have

different sizes.

Figure 2.1. Section of citrus fruit (Ranganna et al., 1986) b. The inner peel

Also known as albedo, the inner peel is located underneath the flavedo. It is

typically a layer of spongy and white parenchyma tissue that is rich in sugars,

pectic substances, celluloses, hemicelluloses and pentosans (Ranganna et al .,

1986). The thickness of the albedo varies with the species. For example,

mandarins generally have very thin albedo while the one in citrons is very thick.

Both flavedo and albedo form the non edible part of the fruit called the pericarp,

and they are commonly known as the rind or peel. c. The endocarp

Beneath the albedo of citrus fruits is the edible portion or also known as endocarp.

It is composed of many segments or carpels, usually around 8-12 in most citrus.

Each segment is surrounded by a fairly tough, continuous membrane and covered

4 by vascular bundles that transfer nutrients for growing of the fruit. The interior of

a segment consists of 2 major components, the juice vesicles and the seeds (Soule

and Grierson, 1986). The juice vesicles are thin-walled and they constitute the

juice within the vacuole of the cell.

2.1.2. Chemical Composition

The chemical composition of citrus fruits may vary as affected by many factors such as growing conditions, maturity, rootstock, cultivar and climate (Ranganna et al .,

1986). The chemical profiles that are characteristic of particular citrus species can be used to detect the authenticity of citrus juices in quality control (Sass-Kiss et al .,

2004). Some important chemical constituents in citrus fruits are: a. Sugars

The main sugars present in citrus fruits are glucose, fructose and sucrose, which

determine the sweetness of the juices (Kefford, 1966). Maturity is the main factor

that affects the sugar content in citrus juices (Izquierdo and Sendra, 1993). The

concentration of sugars in citrus fruits may range from less than 1% in certain

limes up to 15% in some oranges. b. Polysaccharides

The main polysaccharides present in citrus fruits are cellulose, hemicelluloses and

pectic substances. Even though they are found in relatively small quantity, these

polysaccharides play a role in adding to the body of the juice and hence,

contributing to a desirable juice quality (Nagy and Shaw, 1990). Pectins present in

citrus juice are important as a colloidal stabilizer in protecting juice cloud (Croak

and Corredig, 2006)

5 c. Organic acids

The sourness of citrus fruits is imparted by the presence of organic acids, mainly

citric and malic acids (Kefford, 1955). Other organic acids found in smaller

quantities in citrus fruits are succinic, malonic, lactic, oxalic, phosphoric, tartaric,

adipic and isocitric acids (Izquierdo and Sendra, 1993). The acid concentration in

citrus fruits can be affected by maturity, storage, climate and temperatures

(Vandercook, 1977). The organic acids in citrus fruits are mainly measured as

titratable acidity, which is expressed as grams of citric acid as per 100mL of juice

(Ranganna et al ., 1986). The concentration of citric acid in oranges may decrease

with maturity and results in the decrease of acidity (Geshtain and Lifshitz, 1970). d. Lipids

The lipids present in citrus fruits include simple fatty acids in the seed,

phospholipids and complex lipids in the juice and the components of cuticle. They

constitute about 0.1% of orange juice (Moufida and Marzouk, 2003). Some major

fatty acids commonly found in citrus juices as reported by Nagy (1977a) are

palmitic, palmitoleic, oleic, linoleic and linolenic acids. As different citrus

varieties consist of different types of fatty acids, its profile can also be used as

markers for various citrus species (Nordby and Nagy, 1971). The breakdown of

lipids in citrus juices may contribute to the development of off-flavour (Nagy and

Nordby, 1970). e. Carotenoids

The colours of citrus fruits are mainly imparted by the presence of carotenoids

(Stewart, 1977). It ranges from deep orange in red tangerines to light yellow in

lemons. The complex mixture of carotenoids is located in the plastids of the

flavedo and of the internal juice vesicles. Recent study on carotenoid composition

6 of various citrus species by Agócs et al . (2007) revealed that most citrus species,

except lime, contain β-cryptoxanthin and lutein in considerable amounts. The

carotenoids present in lime are mainly β-carotene and lutein (Agócs et al ., 2007). f. Vitamins

The main vitamin present in citrus fruits is ascorbic acid. The juice typically

contains one quarter of the total ascorbic acid present in the fruit. Other vitamins

present in citrus juices in various quantities include thiamine, riboflavin, niacin,

pantothenic acid, inositol, biotin, vitamin A, vitamin K, pyridoxine, p-

aminobenzoic acid, choline and folic acid (Kefford, 1955; Ting and Attaway,

1971). g. Inorganic elements

Generally, citrus fruits are rich in potassium and nitrogen, which accounts for

about 80% of the total minerals (Izquierdo and Sendra, 1993). Other major

inorganic elements found in citrus juices are calcium, iron, phosphorus,

magnesium and chlorine (Nagy, 1977b). The concentration of these elements may

vary depending on the geographical condition, maturity, seasonal variation and

level of fertilization. Thus, the presence of these inorganic elements has been

proposed of tracing the geographic origin of the citrus fruits. h. Flavonoids

The flavonoids in citrus fruits are present in high concentrations and easily

isolated. Some of them are useful for taxonomic markers while some have distinct

taste properties and can be utilized as valuable by-products. The main 3 groups of

flavonoids are flavanones, flavones and anthocyanins (Ranganna et al ., 1986).

Generally flavanones are mainly found in higher amounts while flavones and

anthocyanins are relatively present in trace amounts. Hesperidin is the main

7 flavonoid found in sweet oranges and lemon, while naringin is the flavonoid

responsible for bitter flavour in grapefruit (Nagy and Shaw, 1990). i. Limonoids

Limonin is the only limonoid found in significant amount in citrus fruits and it

imparts bitter flavour (Kefford, 1955). Limonin is not found in fresh fruits and is

produced by acid and enzyme catalyses of limonoid acid A-ring lactone (Nagy

and Shaw, 1990). This conversion normally takes place during juice storage or

with heat treatment. j. Volatile compounds

The volatile compounds present in citrus fruits impart the flavour of the citrus

significantly. Their individual contribution and concentrations, as well as

interactions among them, give characteristic odour to individual species

(Izquierdo and Sendra, 1993). They are mainly present in the juice vesicles and in

the oil sacs of the flavedo. Limonene is the major volatile compound found in

citrus fruits.

2.1.3. Uses of Citrus Fruits

2.1.3.1. Juices

Juice is the primary product obtained from citrus fruits (Braddock, 1999) and it is also one of the most important commodities. The juices produced from the citrus fruits are either in the form of single-strength or concentrated juices (Ting and Rouseff, 1986).

The single-strength juice can be obtained directly from the fruit by adding water to the citrus concentrate, while in concentrated juice, water is removed from the juice in order to reduce the cost of transportation and storage. The citrus juices contain vitamins, minerals, carotenoids, sugars, organic acids, amino acids, phenolics,

8 nucleotides, enzymes, limonoids, lipids, proteins, pectins and other soluble and insoluble solids. The technology and choice of juice recovery methods play an important role in juice processing. Various extraction methods in juice processing are discussed profoundly by Braddock (1999). Among the citrus fruits, oranges and grapefruits are commonly extracted for their juices and they are widely consumed for their health benefits due to the content of nutrients and other bioactive compounds

(McGill et al ., 2004).

2.1.3.2. Essential oils

Another important product of citrus fruits is the essential oils extracted mainly from the peel. In order to obtain the oil, the oil-bearing sacs need to be punctured by either abrasion or scraping the surface of the peels (Redd and Hendrix, Jr., 1996). For its recovery, the oil is washed away from the peel as an aqueous emulsion and then separated from the water by centrifugation (Ohloff, 1994). Hence, expression or cold- pressing method is frequently applied in extracting the oil, and the oil is commonly known as cold-pressed oil. The oil can also be extracted from the peel by other means, such as distillation by steam or water as well as extraction with supercritical or liquid

CO 2. Cold pressed oils have finer aromas and greater stability than distilled oils due to the absence of heat during process and the inclusion of components, such as anti- oxidants (Wright, 2004). The types of citrus fruits from which their peel oils are recovered commercially are orange, grapefruit, tangerine, lemon and lime (Shaw,

1977). Some oil is also present in the juice, but in a relatively small quantity. The amount of oil in the processed juice should not exceed 0.015-0.02% by volume (Redd and Hendrix, Jr., 1996). Hence, excess oil will be removed from the juice by steam

9 distillation in order to lower the juice’s oil content for optimal citrus flavour. The essential oils contain many volatile compounds, mainly aldehydes, ketones, esters, alcohols and terpenes, which give the characteristic aromas and flavours of the citrus fruits (Kefford, 1955; Braddock, 1999). Citrus essential oils are greatly utilized as the flavourings in the food and beverage industries (Colombo et al ., 2002), and as fragrance materials in the perfumery, toiletries, fine chemicals and cosmetic products

(Buccellato, 2002; Baser and Demirci, 2007). Furthermore, citrus essential oils can also be used, to some extent, as a traditional medicine (Imbesi and De Pasquale,

2002).

2.1.3.3. Essence oil and aroma

During the process of juice concentration, some of the natural flavour compounds are also removed together with the water, including the small amounts of peel oil remaining in the juice. The volatiles recovered during the production of juice concentrates are called essence (Redd and Hendrix, Jr., 1996). The water-soluble portion of the essence is known as aqueous essence or aroma while essence oil or oil phase essence refers to the oil-soluble portion. Aroma and essence oil are commonly used as natural flavourings for citrus juice products as they contain many volatile compounds found in cold-pressed oil (Shaw, 1977).

2.1.3.4. Other citrus by-products

The main by-products of citrus processing are the peel, pulp and seeds, which account for 40-60% of the weight of the raw material (Licandro and Odio, 2002). These residues can be further processed into 3 main categories: animal feed, raw material used for further extraction of marketable products and food products. Although most

10 of the citrus by-products are used for animal feed (Ting and Rouseff, 1986), there are many useful by-products made from different portions of the citrus fruits, such as pectin, dried pulp, molasses, marmalades, candied peel, peel seasoning, purees, beverage bases, citrus alcohol, bland syrup, citric acid, seed oil, flavonoids and other products (Kesterson and Hendrickson, 1958; Braddock and Cadwallader, 1992;

Braddock, 1995; Hendrix, Jr. and Hendrix, 1996; Braddock, 1999; Licandro and

Odio, 2002). In the past, by-products became the source of additional revenue for many citrus processors with low juice values (Braddock, 1995). Hence, the utilization of citrus by-products to produce more valuable products is getting increasingly important as future world citrus production increases and then surpasses the demand for citrus juices and beverage products. Furthermore, the future uses of citrus by- products will also need to expand beyond the current major use as low-value animal feed.

On the whole, the current rapid growth of the citrus industry is largely due to population increase and improved economic conditions in the consuming nations of the world, together with the rapid advance of agricultural sciences and technology for the production of by-products. The fact that citrus fruits is a rich source of essential minerals, vitamins and dietary fibres with its distinctive natural flavour and that the consumers are nowadays more nutrition-conscious, have also contributed to the increased demand for citrus fruits and their by-products.

11

2.2. Citrus variety

2.2.1. General classification

As a result of massive hybridisation, there are literally thousands of citrus cultivars in the world. Consequently, the taxonomic classification of citrus becomes quite complex with many diversities and is not universally agreed upon (Young, 1986).

However, in general, citrus can be categorized into five major groups that are significant economically: a. Sweet oranges ( Citrus sinensis Osbeck)

Sweet orange is grown throughout the world and provides the greatest fresh fruit

production of any citrus groups (Young, 1986). It is round to oval in shape, orange

coloured, tight skinned and has a juice and sweet flesh. It can be eaten out-of-hand

easily and is used as fresh ingredients in salads, in fresh juice and for juice

concentrate. It can be sub-divided into four categories – round or common

oranges, navel oranges, acidless oranges and blood oranges (Ortiz, 2002). Some

popular cultivars of sweet oranges are Valencia, Jaffa, Mosambi, Pineapple,

Hamlin, Washington navel and Shamouti. b. Mandarins ( Citrus reticulata Blanco)

Mandarin ranks second in the citrus production worldwide and China is the largest

producer of mandarins (FAOSTAT). Although the name tangerine is used

interchangeably with mandarin, tangerine usually refers to those varieties

producing deep orange coloured fruits (Webber, 1948). Mandarin is round in

shape, sweet in taste, loose skinned and orange in colour. Its segments are easily

separable. It is used primarily for eating out-of-hand, in fresh juice, and to a

limited extent for processing. It can be sub-divided into four classes – Satsuma

12 group, Mediterranean mandarin, Tangerine or Clementine group and other

mandarins, such as King mandarin (Ortiz, 2002). Some important commercial

cultivars of mandarin groups are Dancy, Ponkan, Mikan, Owari and Temple. c. Grapefruits ( Citrus paradisi Macfadyen )

Grapefruit is probably a hybrid between the pummelo and the sweet orange

(Morley-Bunker, 1999). It is sweet, juicy, medium to large in size and has thick

and spongy rind. It has few cultivars – white-fleshed, pink-fleshed and red-fleshed

(Young, 1986). The commercial cultivars are prized as breakfast fruit and for

salads and juice due to their refreshing flavour and mild bitterness. Examples of

popular grapefruit cultivars are Marsh, Star Ruby, Ruby Red and Foster. d. Lemons ( Citrus limon Burmann)

Lemon constitutes an important fresh fruit group even though it is not eaten fresh

as mandarins and oranges. They usually have high acid content although acidless

cultivars also exist (Ortiz, 2002). It is used primarily for drinks and fresh juice or

lemonade, cooking and flavouring, especially in the making of lemon pies, lemon

cakes, candies, jams and marmalades, and also for medicinal purposes due to its

high content of vitamins (Webber, 1948). The fruit is generally oval to elliptical

with characteristic necks and nipples. The peel is yellow at maturity and has

prominent oil glands. The flesh is pale yellow in colour and very sour. There are

three major groups of lemons: the Femminello, the Verna and the Sicilian groups

(Morley-Bunker, 1999). e. Limes ( Citrus aurantifolia Swingle)

Lime is commonly used in limeade and carbonated beverages, and as a constituent

of alcoholic drinks. They can also be used for pickling; for culinary purposes,

such as flavouring for jellies, jams and marmalades; as a garnish, especially with

13 meats and fish; for medicinal purposes, especially in the treatment and prevention

of scurvy; as well as a source of lime oil (Webber, 1948; Young, 1986). It is

greenish-yellow in colour and thin skinned. The juice is highly acidic. The two

major groups include the acid and acidless limes of which the acid limes are of

commercial importance (Davies and Albrigo, 1994). Two popular acid lime

cultivars are Tahiti and Key (Mexican) limes.

On top of these 5 major groups, there are other citrus groups that are widely cultivated and important for various purposes, such as sour or bitter oranges ( Citrus aurantium

Linnaeus), pummelos ( Citrus grandis Osbeck), citrons ( Citrus medica Linnaeus), calamondins ( Citrus mitis Blanco), bergamot ( Citrus bergamia Risso), Kaffir lime

(Citrus hystrix DC.) and kumquats ( Fortunella sp. Swingle). Moreover, the feasibility of hybridization across various groups of citrus results in the emergence of many novel cultivars, and in some cases are difficult to identify (Ortiz, 2002). Some of these hybrids are tangelos (hybrids of grapefruits and mandarins), tangors (hybrids of mandarins and sweet oranges), orangelos (hybrids of sweet oranges and grapefruits), citranges (hybrids of trifoliate oranges and sweet oranges), citrangors (hybrids of citranges and sweet oranges), limequats (hybrids of limes and kumquats) and other hybrid varieties.

2.2.2. Selected citrus cultivars from Asia

2.2.2.1. Pontianak orange (Citrus nobilis Loureiro var. microcarpa Hassk.)

Pontianak orange ( Figure 2.2 ) belongs to the mandarin group. According to Morton

(1987), Citrus nobilis is suggested to be the possible hybrid between sweet orange

(Citrus sinensis ) and mandarin ( Citrus reticulata ). Pontianak orange is the most

14 cultivated citrus cultivar in Indonesia due to its high yield, ease of cultivation and it is also well-liked by the locals (Sarwono, 1986). Pontianak orange has thin, fairly shiny and yellowish green-coloured skin. The diameter of the fruit ranges from 5.5 to 5.9 cm. The flesh is orange in colour, contains a lot of juices and has a very sweet taste.

The history of Pontianak oranges in Indonesia is believed to originate in 1936, when they were firstly grown in surrounding towns nearby Pontianak of West Kalimantan province in Indonesia by local farmers (Sarwono, 1986). However, there was a major damage of this cultivar in Indonesia due to government policy and disease infection over the last decade. Recently, the local government begins to encourage the cultivation of Pontianak orange across the country (Syaifullah and Harijono, 2004).

Figure 2.2. Pontianak oranges

2.2.2.2. Mosambi (Citrus sinensis Osbeck)

Mosambi ( Figure 2.3 ) is particularly popular in central India and is probably the most important orange cultivar of India (Hodgson, 1967). The fruit is round in shape and moderately seedy. The colour of the skin is light yellow to pale orange at maturity.

15 The skin is relatively thick and the surface is moderately to roughly pebbled. The flesh is somewhat firm and juicy whereby its flavour is fairly bland due to its very low acid content. Mosambi grown in India subcontinent has virtually very low acid content due to the climate while it has some acidity when it is grown elsewhere.

Hence, Mosambi is often mistakenly regarded as the acidless type of sweet orange

(Saunt, 2000). This very distinctive cultivar is of unknown origin, but the name, of which there are numerous spellings, suggests that it was taken from Mozambique,

East Africa, to India, presumably by the Portuguese (Hodgson, 1967). Mosambi can be grown under both subtropical and tropical conditions.

Figure 2.3. Mosambi

2.2.2.3. Dalandan (Citrus reticulata Blanco)

Dalandan ( Figure 2.4 ) belongs to the common mandarin group and they are characterized by the loose skin. Batangas province is the place in the Philippines

16 where dalandans are widely cultivated on a large scale. Locally, dalandans are also known as naranjita , dalanghita or sintones . Dalandans are green in colour and turn greenish yellow as they mature. They are relatively sour, especially when it has not fully ripened. There are many varieties of dalandans varying in sizes. The varieties commonly grown in the Philippines are Ladu, Szinkom, Batangas, Ponkan, Taikat and

King (DOST Region X). Szinkom and Ladu are the popular cultivars in the

Philippines as the trees are early maturing. However, they are not the native citrus of the Philippines but were introduced to the country in early 1900s from India (Wells et al ., 1925). In general, Szinkom cultivar is smaller in size and sweeter than Ladu.

Figure 2.4. Dalandan

2.3. Citrus Flavour

Flavour is one of the important attributes, other than texture and appearance, in the selection and acceptance of a particular food (Fisher and Scott, 1997). According to

Hall (1968), flavour can be defined as the sensation produced by a material taken in

17 the mouth, perceived principally by the senses of taste and smell, and also by the general pain, tactile, and temperature receptors in the mouth. Thus, flavour refers to the overall sensation that results from the impact of the food consumed as it is composed of 3 components, namely odour, taste and the tactile sensation in the mouth, such as pungency, heat and cooling (IFT, 1989). Ohloff (1972) classified food flavours into 9 groups, namely fruit, vegetable, spice, beverage, meat, fat, cooked, empyreumatic (smoked or roasted flavour) and stench (e.g. cheese).

Belonging to the fruit flavour group, citrus flavour is among the most popular fruit flavours for beverages and other sweet products, such as cookies, confectionery and desserts (Colombo et al ., 2002). It is therefore not surprising if citrus flavour has been extensively investigated and reviewed. Most of the knowledge of citrus flavour is obtained from the studies of volatile compounds present in the juices, essential oils or the fresh fruits (Shaw, 1991). Ample literatures have been published on the characterization of the volatile compounds and evaluation of aroma active compounds in various species and cultivars of citrus fruits, from the famous cultivars such as

Valencia and Navel oranges (Buettner and Schieberle, 2001), blood and blond oranges (Näf et al ., 1996), mandarin and Dancy tangerine (Shaw and Moshonas,

1993; Naef and Velluz, 2001), Clementine (Buettner et al ., 2003; Chisholm et al .,

2003), bergamot (Sawamura et al ., 2006), Satsuma mandarin (Choi and Sawamura,

2002) and ponkan (Sawamura et al ., 2004), to the local cultivars such as Italian blood and blond oranges (Maccarone et al ., 1998), Ethiopian sweet oranges (Mitiku et al .,

2000), Japanese Yuzu (Song et al ., 2000a), Venezuelan sweet oranges (Ojeda de

Rodriguez et al ., 2003), Libyan oranges (MacLeod et al ., 1988), Vietnamese citrus

(Minh Tu et al ., 2002a), Philippine Calamondin (Takeuchi et al ., 2005), Chinese

18 sweet oranges (Sawamura et al ., 2005), Korean Hallabong (Choi, 2003) and Turkish

Kozan (Selli et al ., 2004).

Among all the citrus fruits, orange is the most popular flavour of the fruit-flavoured beverages (Shaw, 1996a) and the flavour of orange juice has been studied more than that of any other type of citrus fruit (Selli et al ., 2004). This may be due to the appreciation of orange juice as the most popular fruit beverage worldwide and its great demand as a result of its nutritional and sensory properties (Maccarone et al .,

1998).

2.3.1. Important volatile compounds in citrus flavour

There are differences in the flavour of various citrus species due to the difference in the profile of volatile compounds among them (Shaw, 1996b). Limonene is the major volatile compound found in most citrus oils but it is not the most important volatile contributor to the citrus flavour (McGorrin, 2002). Among the citrus fruits, orange has the most complex flavour profile followed by mandarin, grapefruit, lime and lemon. a. Orange

The volatile compounds important in contributing to fresh orange flavour as

reported by Ahmed et al. (1978a) are d-limonene, alpha-pinene, valencene,

acetaldehyde, octanal, nonanal, citronellal, neral, geranial, ethyl butyrate and

linalool. It is believed that no single or two compounds is solely responsible for

orange flavour, as it is rather the result of a combination of volatile compounds in

given proportions (Shaw, 1991).

19 b. Mandarin

In mandarin flavour, methyl-N-methylanthranilate and thymol are believed to be

the important contributors as stated by Kugler and Kovats (1963). Furthermore,

together with those two compounds, beta-pinene and gamma-terpinene are also

considered to be important for a full and complete mandarin flavour (Shaw,

1996b). Some aldehydes that are deemed important to mandarin flavour are alpha-

sinensal, octanal and decanal (Shaw, 1979). Yet, just with orange flavour, a

mixture of several compounds in the specific proportions also seems essential for

creating a mandarin flavour (Wilson and Shaw, 1981). c. Grapefruit

Grapefruit flavour is characterized by its harsh and bitter notes, which may also be

contributed by the right proportions of sugar and acid ratio, and the composition

of the volatile compounds (Shaw, 1986). The study on cold-pressed grapefruit oil

by Lin and Rouseff (2001) revealed that over 200 volatiles reported in grapefruit

oil, only 22 were essential in contributing to grapefruit flavour. Nootkatone and

1-p-menthene-8-thiol are the commonly known flavour impact compounds in

grapefruits (McGorrin, 2002). Besides, other volatile compounds considered

important to grapefruit flavour are decanal, acetaldehyde, methyl butyrate,

limonene, ethyl acetate, ethyl butyrate, and 2,8-epithio-cis-p-menthane (Shaw,

1996b). There are also some sulphur-compounds found to be odour-active in

grapefruit juice, such as 4-mercapto-4-methyl-2-pentanone, methional, 3-

mercaptohexyl acetate and 3-mercaptohexan-1-ol (Buettner and Schieberle, 1999;

Lin et al ., 2002).

20 d. Lemon

Citral, a mixture of neral and geranial isomers, is a well-known character impact

compound for lemon flavour (McGorrin, 2002). Study by Cotroneo et al . (1986)

revealed that higher quality lemon peel oil contained higher levels of citral,

linalool, nerol, citronellol, alpha-terpineol and terpinen-4-ol. Other compounds

found to be important in contributing to lemon flavour are methyl epijasmonate

and its isomer, methyl jasmonate (Nishida and Acree, 1984). e. Lime

Citral is also found to be an important compound in contributing to lime flavour,

together with alpha-terpineol (McGorrin, 2002). Clark Jr. et al . (1987) reported

that a sesquiterpene, germacrene B, is important in imparting the fresh lime peel

character. Extensive study on the characterization of the volatile compounds in

cold-pressed Key and Persian lime oils was done by Dugo et al . (1997).

2.3.2. Factors affecting citrus flavour

The flavour quality of citrus fruits can be affected by various factors, namely agricultural practice, which includes fertilization, climate, rootstock and maturity; the ratio of total sugar and acid present in the juice; the presence of pectin, high molecular weight carbohydrates, bitter compounds like limonin and naringin; as well as presence of volatile compounds (Nagy and Shaw, 1990). Yet, it is well known that the fresh and uniquely delicate flavour of citrus fruits is mainly contributed by the complex combinations of many volatile compounds blended in the proper proportions (Shaw,

1991; Moshonas and Shaw, 1995). Other factors that may influence the flavour include taste threshold of volatiles, synergistic effect between volatiles and the interaction of non-volatile with volatile flavour compounds (Nisperos-Carriedo and

21 Shaw, 1990). The volatile compounds present in the citrus fruits can be divided into two broad categories, those present in the oil and those in the juices.

2.3.2.1. Volatile compounds in the citrus oil

The oil-soluble compounds of citrus fruits are present in peel oil and in juice oil. The peel oil is located in small, ductless glands present in the outer portion of the peel or flavedo. The peel oil from each citrus variety affects the characteristic aroma and flavour of that variety. Some peel oil is mechanically transferred to the fruit sections or juice as the fruit is either peeled or juiced (Nagy, 1996). The peel oil contains mostly terpenoids (Ahmed et al ., 1978b). Shaw (1991) reported that juice sacs in most citrus species also contain an oil that is somewhat different in composition from that of peel oil. The oil is also known as juice oil. Peel oil and juice oil have different characteristics as their chemical compositions vary (Wolford et al ., 1971). In general, the juice oil had an aroma more like that of fresh citrus juice and contained fewer aldehydes and more esters than peel oil. Hunter and Brogden (1965) reported that orange juice oil generally contained higher percentage of sesquiterpenes, particularly valencene than the peel oil, while grapefruit juice oil contained higher percentage of valencene but lower percentages of other sesquiterpenes, such as cadinene, alpha- copaene and beta-cubebene, than the corresponding peel oils.

2.3.2.2. Volatile compounds in the citrus juices

Generally, the characteristic flavour of fresh citrus fruits is contributed by the volatile compounds present in the juices (Huet, 1969). There are more than 200 volatile compounds that have been identified in orange juice (Johnson et al ., 1996) and the important contributors to orange juice flavour include esters, aldehydes, ketones,

22 terpenes and alcohols (Selli et al ., 2004). However, freshly squeezed orange juice has a full, fruity flavour quality that has not been completely reproduced in any orange juice product (Shaw, 1986). Mandarins contain a number of species, and are more diverse than other citrus groups (Hodgson, 1967). As a result, various cultivars of mandarins have distinct profiles of volatile compounds even though all of them belong to the mandarin group (Moshonas and Shaw, 1997). While generally a mixture of thymol, methyl-N-methyl anthranilate and several monoterpene hydrocarbons in the proper proportions were suggested to be important flavour contributors to mandarin (Shaw and Wilson, 1980), methyl-N-methyl anthranilate and thymol were not found in Satsuma mandarin (Yajima et al ., 1979) and the former was not found to be the aroma active compound in Dancy and Sunburst cultivars (Evans and Rouseff,

2001). Grapefruit juice has a unique flavour quality and unlike other citrus juices, a tinge of bitterness is desirable in grapefruit juice. More than 60 volatiles were reported in grapefruit juice (Núñez et al ., 1985) and the volatile compounds that have been shown to be important in grapefruit juice flavour include aldehydes, esters and nootkatone (Shaw and Wilson, 1980). More than 300 volatile compounds were detected in lemon juice (Mussinan et al ., 1981). They reported that the ether, p- cymen-8-yl ethyl ether was found to possess a lemon-juice like flavour, and citral, the most important compound in lemon flavour, was also found in the juice. Some sulphur compounds have also been found to play important role in the flavour citrus juices, especially hydrogen sulphide in most of citrus juices and 1-p-menthene-8-thiol in grapefruit juices (Shaw, 1991). The delicate flavour of citrus juices is easily changed by heat treatment during processing or storage due to the degradation of some volatile compounds (Pérez-López and Carbonell-Barrachina, 2006). The compounds found to be responsible for the quality loss of the citrus juices after

23 excessive heat treatment are 1-ethyl-2-formylpirrole, 2-hydroxy-3-methyl-2- cyclopenten-1-one, 5-methyl-2-furaldehyde, 2,5-dimethyl-4-hydroxy-3(2H)-furanone,

2-methyl-4-ethylphenol and alpha-terpineol (Pérez-López et al ., 2006).

2.4. Flavour Research

A key step towards understanding what constitutes the flavour of any food product is to establish the chemical nature of the volatile compounds that act, either independently or in combination, to produce a highly characteristic aroma response for that particular product (Cronin, 1990). Hence, it is necessary to isolate the volatile aroma compounds from the non-volatile bulk of the food in order to study the chemical basis of flavour. Subsequently, the appropriate techniques of physical and chemical analysis need to be applied in order to obtain the maximum qualitative and quantitative information on all compounds of possible sensory importance.

2.4.1. Challenges in flavour research

There are many challenges faced in flavour research mainly due to the complex nature of the foods and the limitation of the analytical techniques. Some of the common problems faced in carrying out the flavour research are (Flath et al ., 1981; Parliment,

2002): a. Concentration level

The concentration of aromatic compounds is generally low, typically in the ppm,

ppb or ppt range. Thus, it is difficult to isolate those compounds at a very low

concentration. Moreover, besides isolating the compounds, it is also necessary to

concentrate them for further analysis.

24 b. Matrix

The volatiles are usually intracellular and must be liberated by disruption. The

sample may also contain non-volatile compounds that may complicate the

isolation process. c. Aroma diversity and complexity

The aromatic composition of foods is generally very complex and may consist of

hundreds of compounds with various classes of functional groups. Furthermore,

different classes of compounds have different polarities, solubilities, volatility and

pH values. d. Instability

Many volatile compounds contain various functional groups that are unstable and

may be oxidized by air or degraded by heat or extremes of pH. e. Small fraction of potent odourants

Not all volatile compounds present in foods are contributors to the food flavour

and there are only a small fraction that is of significant importance (Grosch,

1993). Hence, potent odourants need to be distinguished from the less odour

active or odourless compounds in foods.

As a result, no single technique has proven to be the most suitable for all samples and that is able to isolate all aroma compounds in food that represent the actual aroma profile of the food. It is important to ensure that the methods applied in flavour research do not cause the decomposition and loss of the desired compounds.

25 2.4.2. Systematic approach in flavour research

To effectively identify the key flavour compounds that contribute significantly to flavour of foods, a systematic approach is necessary. It starts from the isolation of flavour compounds and the identification of odour active compounds to the confirmation of the findings by aroma reconstitution and omission tests. Each step is crucial in determining the final results and is described below.

2.4.2.1. Isolation of flavour compounds

Isolation of flavour compounds from the food matrix is the initial step in flavour research and is vital because no analytical method will be valid unless the isolates represent the food materials being studied (Teranishi, 1998). In order to obtain a volatile mixture that represents the true characteristics of the food, several isolation techniques have been developed. Some of them are: a. Solvent extraction

Solvent extraction is one of the simplest and most efficient approaches in isolation

of flavour compounds (Reineccius, 2006). Also known as liquid-liquid extraction,

the volatile compounds are extracted by the organic solvent from the aqueous

phase (Da Costa and Eri, 2005). It can be carried out in batches by using

separatory funnel, or continuously by using liquid-liquid extractor. The solvents

are usually selected based on its selectivity and boiling point, and must be of high

purity (Sugisawa, 1981). There are many kinds of solvents readily available for

solvent extraction as listed by Weurman (1969), but the solvents commonly used

today in flavour research are diethyl ether, diethyl ether/pentane mixtures,

hydrocarbons, Freon and methylene chloride (Parliment, 2002). Diethyl ether is

commonly used due to its high extraction efficiency while hydrocarbons are

26 relatively non-selective but have low extraction efficiency. Methylene chloride is

found to be a satisfactory general purpose solvent, particularly for flavour

compounds with enolone structure, such as Maltol and Furaneol. However, it is

relatively toxic and is an animal carcinogen. b. Distillation

The main principle of distillation is the separation of volatiles from the non

volatiles by applying heat and collecting the vapours by removal of heat (Fisher

and Scott, 1997). It is one of common techniques employed today for flavour

research due to its simplicity of operation and apparatus required, reproducibility,

rapidity and capability of handling a wide range of samples. Distillation, as an

isolation method, can be carried out by directly heating the sample (simple

distillation) or by using steam (steam distillation) under atmospheric or reduced

pressure (Teranishi et al ., 1971). In general, steam distillation is more rapid and

results in less decomposition of the sample. To some extent, distillation can also

be used for concentration of volatiles by reflux stripping using vigreux column

(fractional distillation). A combination of distillation and extraction steps with

single apparatus was developed by Likens and Nickerson (1964) and it is also

known as simultaneous distillation/extraction (SDE). A good review of this

method has been reported by Chaintreau (2001). c. Headspace sampling

In headspace sampling, the volatile analytes from the sample are extracted by

investigation of the atmosphere adjacent to the sample, leaving the actual sample

material behind (Wampler, 2002). In general, headspace sampling techniques can

be divided into 2 categories: static headspace and dynamic headspace or purge-

and-trap (Da Costa and Eri, 2005). In static headspace, the volatiles in the

27 atmosphere around the sample is directly injected onto the gas chromatograph

(GC) column while in dynamic headspace, the volatiles from larger samples of the

headspace are swept away by carrier gas and concentrated onto a trap prior to

injection into the GC. The most commonly trap used in dynamic headspace is the

Tenax® trap (Reineccius, 2002). The main advantage of headspace sampling is

that the analytes are removed from the sample matrix without the use of an

organic solvent which may interfere with the extraction process. However, the

results obtained from headspace sampling may not be truly representative of the

compounds present in the sample as it does not permit the determination of

higher-boiling point compounds that may be significant to the aroma (Sugisawa,

1981). d. Solid phase microextraction (SPME)

SPME is a solvent-free method developed by Pawliszyn and his group (Zhang and

Pawliszyn, 1993). It involves extracting volatile compounds from their matrices

by partitioning them from a liquid, gaseous or solid sample into an immobilized

stationary phase, i.e. fused silica fibre in the SPME apparatus. The analytes are

adsorbed by the fibre phase until equilibrium is reached in the system. The amount

of an analyte extracted is determined by the magnitude of the partition coefficient

(distribution ratio) of the analyte between the sample matrix and coating material

(Pawliszyn, 1999). Thus, in SPME, analytes are not extracted quantitatively from

the matrix. After adsorption of the volatile compounds into the SPME fibre, the

desorption can then be performed thermally by direct injection into GC column.

Some advantages of this technique are that it is rapid, requires no solvent, can be

applied for liquids, solids or gases, and can be performed without heating the

sample (Harmon, 2002). Various factors that need to be taken into account in

28 using SPME for isolation of volatile compounds include sample and headspace

volume, extraction time and temperature, salt addition and one of the most

important factors – the selection of SPME fibre. The selection of SPME fibre is

based primarily on polarity and volatility characteristics of the analyte (Pawliszyn,

2001). More about this method and fibre selection was recently reviewed by

Wardencki et al . (2004). e. Emerging isolation methods

Apart from the common techniques mentioned above, there are emerging isolation

techniques that have been applied recently in flavour research. Some examples of

these isolation methods include high-pressure extraction with supercritical CO 2

(Werkhoff et al ., 2002), solvent-assisted flavour evaporation (SAFE) developed

by Engel et al . (1999) and stir bar sorptive extraction (SBSE) invented by

Baltussen et al . (1999).

2.4.2.2. Separation of flavour compounds

Upon isolation of the volatile compounds by any of the above mentioned techniques, the compounds need to be separated and identified for characterization. Gas chromatograph (GC) is an indispensable instrument and a premier choice of separation technique, particularly for analysis of volatile compounds (Reineccius,

2006). In GC, samples are volatilized and transported by the mobile phase (carrier gas) to the column, where separation takes place (Martín-Hernández and Juárez,

1993). The carrier gas should be inert and commonly used gases are Helium, Nitrogen and Hydrogen. Upon separation at the end of the column, the compounds are then detected by a detector. Basic components of a GC system are injector, column and detector.

29

a. Injector

In general, the process of injection into GC system must be capable of accepting

the sample, vapourizing the sample and transferring the sample to the column

without decomposition of the analyte or discrimination between different

components of the analyte (Currell, 2000). For vapourization, the sample is

rapidly heated at a certain injection temperature and is then carried into the

column by the carrier gas. b. Column

GC columns can be classified into 2 groups: packed and capillary columns.

Packed columns are tubes made of copper, glass, stainless steel or other materials

formed in any shape that will fit the GC oven. They are filled with finely divided,

inert, solid support material (mostly prepared from diatomaceous earth) coated

with liquid stationary phase. On the other hand, capillary column, or also known

as open-tubular column, can be sub-divided into wall-coated open tubular

(WCOT) type, porous-layer open tubular (PLOT) type and support coated open

tubular (SCOT) type. The most important and widely used type is WCOT (Wang

and Paré, 1997). In general, capillary columns tend to give higher resolution than

the packed columns while the latter are able to separate much more materials and

they require longer time to complete the separation. The choice of column is

therefore a critical factor in successful separation of flavour compounds. c. Detector

A detector detects the effluents from the column and provides a record of the

chromatography in the form of a chromatogram. The detector signals can be used

for quantitative analysis since they are proportional to the quantity of each

30 analyte. The general detector usually coupled with GC is flame ionization detector

(FID). Other selective detectors for specific compounds are flame photometric

detector (FPD) for sulphur compounds, electron capture detector (ECD) for

sulphur- and halogen-containing compounds, and nitrogen-phosphorus detector

(NPD) for nitrogen- and phosphorus-containing compounds. Wang and Paré

(1997) provide a good review on these detectors.

2.4.2.3. Characterization and identification of flavour compounds

For identification of volatile compounds, the most commonly used and effective detector is mass spectrometer (MS). It is able to provide qualitative information for identification and characterization of compounds, which is lacking in other GC detectors (Chaintreau, 1999). In a GC/MS system, after the compounds are separated in the GC capillary column, the analytes are then ionized by either electron or chemical ionization methods. Subsequently, the ions are separated according to their mass-to-charge (m/z) ratio in the mass analyzer prior to their detection. The most widely used mass analyzer and ion detector in GC/MS system are linear quadrupole analyzer and electron multiplier, respectively (Niessen, 2001). As a result, apart from the chromatogram showing the peaks of various compounds; the mass spectrum, a plot of the relative abundance of the generated ions as a function of the m/z ratio, for each compound is also generated. Eventually, the identification of volatile compounds detected can be carried out by matching the mass spectrum of the compounds to that of the mass spectral libraries. However, since the positional isomers are difficult to discriminate by MS, GC retention time of the standard compounds is often used for confirmation. Thus, the proof of identity can be achieved by matching both mass spectra and GC retention time. Due to its reliability and efficiency, GC/MS has been

31 used extensively as a powerful tool for flavour research in the analysis of volatile compounds in various food products as reviewed by Careri and Mangia (2001).

Moreover, Orav (2001) also reviewed the identification of terpenes, the compounds mainly present in fruits and essential oil of plants, by using GC/MS.

2.4.2.4. Determination of aroma active compounds

As mentioned previously, not all volatile compounds present in foods contribute significantly to the food flavour, but often only a small fraction does (Guth and

Grosch, 1999). These potent odourants are also known as aroma active or odour active compounds. Even though GC/MS is able to separate and identify the volatile compounds, it does not give any illustration on their significance in contributing to the flavour of the products. This is because the peak profile obtained by GC does not necessarily reflect the aroma profile of the food. Moreover, many aroma active compounds are present at very low concentration due to their low odour thresholds

(van Ruth, 2001). Hence, there is a need of an approach that can provide information on the aroma activity of each volatile compound in food. This led to the invention of a technique in flavour analysis known as gas chromatography-olfactometry (GC-O), initially proposed by Fuller et al . (1964). It is the combination of GC separation technique and olfactometry, i.e. the use of human detectors (the nose) to assess odour activity in defined air streams. It can also be coupled with other detectors such as MS or FID for identification of the compounds. The main principle of GC-O is the involvement of the human nose as the detector to sniff the GC effluent that will be channelled to the GC sniffing port (Blank, 2002). Several techniques have been developed to collect and process GC-O data and to estimate the sensory contribution

32 of single aroma components. They can be classified into 3 categories (van Ruth, 2004;

Delahunty et al ., 2006): a. Dilution to threshold

In this method, an extract is diluted stepwise and each dilution is sniffed until no

odour is perceived. The odour potency of an odourant is determined by the last

dilution at which that compound is detected (van Ruth, 2001a). Two common

techniques based on dilution method have been developed: CharmAnalysis

(combined hedonic response measurement) by Acree et al . (1984) and aroma

extract dilution analysis (AEDA) by Ullrich and Grosch (1987). The results are

commonly expressed as Charm value and Flavour Dilution (FD) factor,

respectively. b. Detection frequency

In this method, a number of assessors sniff the same extract and the number of

assessors who can detect an odour is used as an estimate of the odour’s intensity

of that compound (van Ruth, 2001b). Initially developed by Linssen et al . (1993),

this method is simple and does not require much training of the assessors

(Delahunty et al ., 2006). c. Direct intensity

Using this method, assessors are required to use a scale to measure the perceived

intensity of the compound as it elutes. This group of methods include posterior

intensity (Casimir and Whitfield, 1978), finger-span (Étiévant et al ., 1999) and

time-intensity methods (Sanchez et al ., 1992).

The two commonly used GC-O methods are dilution techniques and time-intensity measurements (Blank, 2002). In flavour analysis, GC-O can also be applied for other purposes, such as off-flavour analysis (Lee, 2003) and flavour creation (Iwabuchi et

33 al ., 2001). A good review on GC-O methods was published recently by d'Acampora

Zellner et al . (2008). Nevertheless, there are some drawbacks in using GC-O for flavour analysis, many of which are directly related to inconsistency and subjectivity of humans as detectors (Friedrich and Acree, 2000).

2.4.2.5. Odour Activity Values (OAV) and Relative Flavour Activity (RFA)

The importance of a flavour compound in contributing to a food flavour can also be estimated theoretically by calculating the ratio of its concentration in food to its threshold in a medium that predominates in the food, such as water, oil and starch

(Grosch, 1993). This value is known as ‘aroma value’, ‘flavour unit’, ‘odour unit’,

‘odour value’ or most commonly known as ‘odour activity value’ (OAV). It is hypothesized and accepted that the higher the OAV of a compound, the more potent that compound in contributing to the overall flavour of a particular food (Guth and

Grosch, 1999). However, there is still no guarantee that all flavour compounds with high OAV are the aroma impact compounds. The limitation of OAV concept was criticized by Frijters (1978) has been presented by Audouin et al . (2001). Still, OAV has been found to be useful in practice and is frequently used as a screening method

(Buttery, 1999). Therefore, OAV is commonly analyzed in conjunction with GC-O results, particularly FD factor as their values are often proportional (Grosch, 1994).

In citrus fruits, limonene is the major compound present that can account up to 90% and hence, its OAV is often the highest. However, limonene is not the most important compound in citrus fruits (Jia et al ., 1998). In order to compensate for the inability of

OAV to determine the important aroma active compounds in citrus fruits, the concept of relative flavour activity (RFA) was developed. It was obtained by equation RFA =

34 log 2 n/S 0.5 (Song et al ., 2000a), where 2 n is the FD factor and S is the weight percent of a compound. This concept has been used to determine the important character compounds of yuzu ( Citrus junos ) peel oil (Song et al ., 2000a), daidai ( Citrus aurantium L. var. cyathifera Y. Tanaka) peel oil (Song et al ., 2000b) Citrus

Hyuganatsu (Choi et al ., 2001), Tosa-buntan ( Citrus grandis Osbeck forma Tosa ) fruit (Sawamura et al ., 2001), Kabosu ( Citrus sphaerocarpa Tanaka) peel oil (Min Tu et al ., 2002), Citrus Hallabong (Choi, 2003) and ponkan oil (Sawamura et al ., 2004).

Still, RFA does not always reflect the primary contributors of the food and GC-O sniffing plays an effective role in determining the character impact odourants

(Sawamura et al ., 2001).

2.4.2.6. Aroma reconstitution and omission test

In order to validate the significance of aroma active compounds identified, aroma reconstitution and omission test are often carried out. For aroma reconstitution, a synthetic blend of odourants is prepared based on analytical result obtained and the aroma of this aroma model is then compared to that of the original food (Grosch,

2001). While aroma reconstitution does not reveal specific compounds that actually contribute to the aroma, omission test can be performed to address this matter. In this test, one or more volatile compounds considered to be the potent odourants are omitted from the aroma models and its aroma is then compared to the original food aroma. Only when there is a significant difference in the aroma between the original food and omitted aroma model that the compounds are said to be the key odourants of that food (Engel et al ., 2002).

35 References Acree TE, Barnard J and Cummingham DG. 1984. A Procedure for the Sensory Analysis of Gas Chromatographic Effluents. Food Chem 14 : 273-286.

Agócs A, Nagy V, Szabó Z, Márk L, Ohmacht R and Deli J. 2007. Comparative Study on the Carotenoid Composition of the Peel and the Pulp of Different Citrus Species. Innovat Food Sci Emerg Tech 8: 390-394.

Ahmed EM, Dennison RA, Dougherty RH and Shaw PE. 1978a. Flavour and Odour Thresholds in Water of Selected Orange Juice Components. J Agric Food Chem 26 : 187-191.

Ahmed EM, Dennison RA and Shaw PE. 1978b. Effect of Selected Oil and Essence Volatile Components on Flavour Quality of Pumpout Orange Juice. J Agric Food Chem 26 : 368-372.

Audouin V, Bonnet F, Vickers ZM and Reineccius GA. 2001. Limitations in the Use of Odor Activity Values to Determine the Important Odorants in Foods. In: Leland JV, Schieberle P, Buettner A and Acree TE (eds). Gas Chromatography- Olfactometry: The State of the Art. Washington, DC: American Chemical Society. p 156-171.

Baltussen E, Sandra P, David F and Cramers CA. 1999. Stir Bar Sorptive Extraction (SBSE), a Novel Extraction Technique for Aqueous Samples: Theory and Principles. J Microcolumn Sep 11 : 737-747.

Baser KHC and Demirci F. 2007. Chemistry of Essential Oils. In: Berger RG (ed). Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability. Berlin: Springer-Verlag. p 43-86.

Blank I. 2002. Gas-Chromatography-Olfactometry in Food Aroma Analysis. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 297-331.

36

Braddock RJ. 1995. By-Products of Citrus Fruit. Food Technol . 49 (9): 74-77.

Braddock RJ. 1999. Handbook of Citrus By-Products and Processing Technology. New York: Wiley-Interscience. 247p.

Braddock RJ and Cadwallader KR. 1992. Citrus By-Products Manufacture for Food Use. Food Technol 46 (2): 105-110.

Buccellato F. 2002. Citrus Oils in Perfumery and Cosmetic Products. In: Dugo G and Di Giacomo A (eds). Citrus: The Genus Citrus. London; New York: Taylor and Francis. p 557-567.

Buettner A and Schieberle P. 1999. Characterization of The Most Odour-Active Volatiles in Fresh, Hand-Squeezed Juice of Grapefruit ( Citrus paradisi Macfayden). J Agric Food Chem 47 : 5189-5193.

Buettner A and Schieberle P. 2001. Evaluation of Aroma Differences between Hand- Squeezed Juices from Valencia Late and Navel Oranges by Quantitation of Key Odourants and Flavour Reconstitution Experiments. J Agric Food Chem 49 : 2387- 2394.

Buettner A, Mestres M, Fischer A, Guasch J and Schieberle P. 2003. Evaluation of the Most Odour-Active Compounds in the Peel Oil of Clementines ( Citrus reticulata

Blanco cv. Clementine). Eur Food Res Technol 216: 11-14.

Buttery RG. 1999. Flavor Chemistry and Odor Thresholds. In: Teranishi R, Wick EL and Hornstein I (eds). Flavour Chemistry: Thirty Years of Progress. New York: Kluwer Academic/Plenum. p 353-365.

Calabrese F. 2002. Origin and History. In: Dugo G and Di Giacomo A (eds). Citrus: The Genus Citrus. London; New York: Taylor and Francis. p 1-15.

37

Careri M and Mangia A. 2001. Gas Chromatography-Mass Spectrometry Analysis of Flavours and Fragrances. In: Niessen WMA (ed). Current Practice of Gas Chromatography-Mass Spectrometry. New York: Marcel Dekker Inc. p 409-440.

Casimir DJ and Whitfield FB. 1978. Flavour Impact Values: A New Concept for Assigning Numerical Values for Potency of Individual Flavour Components and Their Contribution to Overall Flavour Profile. Ber Int Fruchtsaftunion 15 : 325-347.

Chaintreau A. 1999. Analysis Technology. In: Swift KAD (ed). Current Topics in Flavours and Fragrances: Towards a New Millennium of Discovery. Dordrecht: Kluwer Academic Publishers. p 97-122.

Chaintreau A. 2001. Simultaneous Distillation-Extraction: From Birth to Maturity – Review. Flavour Fragr J 16 : 136-148.

Chisholm MG, Jell JA and Cass Jr DM. 2003. Characterization of the Major Odourants Found in the Peel Oil of Citrus reticulata Blanco cv. Clementine Using Gas Chromatography-Olfactometry. Flavour Fragr J 18 : 275-281.

Choi HS. 2003. Character Impact Odourants of Citrus Hallabong [( C. unshiu Marcov x C. sinensis Osbeck) x C. reticulata Blanco] Cold-Pressed Peel Oil. J Agric Food Chem 51 : 2687-2692.

Choi HS, Kondo Y and Sawamura M. 2001. Characterization of the Odour-Active Volatiles in Citrus Hyuganatsu ( Citrus tamurana Hort. Ex Tanaka). J Agric Food Chem 49 : 2404-2408.

Choi HS and Sawamura M. 2002. Comparison of the Cold-Pressed Peel Oil Composition between Korean and Japanese Satsuma Mandarin ( Citrus unshiu Marcov. forma Miyagawa-wase) by GC, GC/MS and GC-O. J Food Sci Nutr 7(1): 5- 11.

38 Clark, Jr. BC, Chamblee TS and Iacobucci GA. 1987. HPLC Isolation of the Sesquiterpene Hydrocarbon Germacrene B from Lime Oil and Its Characterization as a Flavour Impact Constituent. J Agric Food Chem 35 : 514-518.

Colombo E, Ghizzoni C and Cagni D. 2002. Citrus Oils in Food and Beverages: Uses and Analyses. In: Dugo G and Di Giacomo A (eds). Citrus: The Genus Citrus. London; New York: Taylor and Francis. p 539-556.

Cotroneo A, Dugo G, Licandro G, Ragonese C and Di Giacomo G. 1986. On the Genuineness of Citrus Essential Oils. Part XII. Characteristics of Sicilian Lemon Essential Oil Produced with the FMC Extractor. Flavour Fragr J 1: 125-134.

Croak S and Corredig M. 2006. The Role of Pectin in Orange Juice Stabilization: Effect of Pectin Methylesterase and Pectinase Activity on the Size of Cloud Particles. Food Hydrocolloids 20 : 961-965.

Cronin DA. 1990. Techniques of Analysis of Flavours: Chemical Methods Including Sample Preparation. In: Morton ID and MacLeod AJ (eds). Food Flavours Part A: Introduction. Amsterdam: Elsevier. p 15-48.

Currell G. 2000. Analytical Instrumentation: Performance Characteristics and Quality. Chichester: John Wiley and Sons Ltd. 307p. d'Acampora Zellner B, Dugo P, Dugo G and Mondello L. 2008. Gas Chromatography-Olfactometry in Food Flavour Analysis. J Chromatogr A 186 : 123- 143.

Da Costa NC and Eri S. 2005. Identification of Aroma Chemicals. In: Rowe DJ (ed). Chemistry and Technology of Flavours and Fragrances. Oxford: Blackwell. p 12-34.

Davies FS and Albrigo LG. 1994. Citrus. Wallingford: CAB International. 254p.

Delahunty CM, Eyres G and Dufour JP. 2006. Gas Chromatography – Olfactometry. J Sep Sci 29 : 2107-2125.

39

DOST Region X. http://region10.dost.gov.ph/technologies/English/citrus.htm (last accessed 30 June 2008)

Dugo P, Mondello L, Lamonica G and Dugo G. 1997. Characterization of Cold- Pressed Key and Persian Lime Oils by Gas Chromatography, Gas Chromatography- Mass Spectroscopy, High-Performance Liquid Chromatography and Physicochemical Indices. J Agric Food Chem 45 : 3608-3616.

Engel W, Bahr W and Schieberle P. 1999. Solvent Assisted Flavour Evaporation – A New and Versatile Technique for the Careful and Direct Isolation of Aroma Compounds from Complex Food Matrices. Eur Food Res Technol 209 : 237-241.

Engel E, Salles C, Nicklaus S, Septier C, Leconte N and Le Quere JL. 2002. Use of Omission Tests to Evaluate Taste-Active Compounds in Food: Application to Cheese and Tomato. In: Reineccius GA and Reineccius TA (eds). Heteroatomic Aroma Compounds. Washington DC: American Chemical Society. p 312-327.

Étiévant PX, Callement G, Langlois D, Issanchou S and Coquibus N. 1999. Odor Intensity Evaluation in Gas Chromatography-Olfactometry by Finger Span Method. J Agric Food Chem 47 : 1673-1680.

Evans KC and Rouseff RL. 2001. Characterization of Aroma Active Compounds in Mandarin ( Citrus reticulata Blanco) Juice Using Gas Chromatography-Olfactometry. Abstracts of Papers, 222 nd ACS National Meeting, Chicago, IL, USA, August 26-30, 2001. Washington DC: American Chemical Society. AGFD-143.

FAOSTAT Agricultural Data. http://faostat.fao.org/ (last accessed 30 June 2008)

Fisher C and Scott TR. 1997. Food Flavours: Biology and Chemistry. Cambridge: Royal Society of Chemistry. 165p.

40 Flath RA, Sugisawa H and Teranishi R. 1981. Problems in Flavour Research. In: Teranishi R, Flath RA and Sugisawa H (eds). Flavour Research: Recent Advances. New York: Marcel Dekker Inc. p 1-10.

Friedrich JE and Acree TE. 2000. Issues in Gas Chromatography-Olfactometry Methodologies. In: Risch SJ and Ho CT (eds). Flavour Chemistry: Industrial and Academic Research. Washington, DC: American Chemical Society. p 124-132.

Frijters JER. 1978. A Critical Analysis of the Odour Unit Number and Its Use. Chem Senses Flavour 3: 227-233.

Fuller GH, Steltenkamp R and Tisserand GA. 1964. The Gas Chromatograph with Human Sensor: Perfumer Model. Ann NY Acad Sci 116 : 711-724.

Geshtain A and Lifshitz A. 1970. Organic Acids in Orange Juice. Lebensm Wiss Technol 3: 115.

Grosch W. 1993. Detection of Potent Odorants in Foods by Aroma Extract Dilution Analysis . Trends Food Sci Technol 4: 68-73.

Grosch W. 1994. Determination of Potent Odourants in Foods by Aroma Extract Dilution Analysis (AEDA) and Calculation of Odour Activity Values (OAV). Flavour Fragr J 9: 147-158.

Guth H and Grosch W. 1999. Evaluation of Important Odourants in Foods by Dilution Techniques. In: Teranishi R, Wick EL and Hornstein I (eds). Flavour Chemistry: Thirty Years of Progress. New York: Kluwer Academic/Plenum. p 377-386.

Hall RL. 1968. Food Flavours: Benefits and Problems. Food Technol 22 : 1388-1392.

Harmon AD. 2002. Solid-Phase Microextraction for the Analysis of Aromas and Flavours. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 75-106.

41 Hendrix, Jr. CM and Hendrix DL. 1996. Citrus Specialty and By-Products. In: Redd JB, Shaw PE, Hendrix, Jr. CM and Hendrix DL (eds). Quality Control Manual for Citrus Processing Plants, Volume III: Flavour: General, Systems, Important Volatiles, Shelf-Life, Specialty and By-Products, Relationships – Raw to the Processed Product, Miscellaneous Conversion Charts and Tables. Auburndale: Agscience Inc. p 203-230.

Hodgson RW. 1967. Horticultural Varieties of Citrus. In: Reuther W, Webber HJ and Batchelor LD (eds). In: The Citrus Industry, Volume I: History, World Distribution, Botany, and Varieties; Revised Edition. Berkeley and Los Angeles: University of California. p 431-591.

Huet R. 1969. The Aroma of Citrus Juices (in French). Fruits 23 : 453-471.

Hunter GLK and Brogden WB. 1965. Analysis of the Terpene and Sesquiterpene Hydrocarbons in Some Citrus Oils. J Food Sci 30 : 383-387.

IFT. 1989. Food Flavours. A scientific status summary by the IFT expert panel on food safety and nutrition. Food Technol 43 : 99-106.

Imbesi A and De Pasquale A. 2002. Citrus Species and Their Essential Oils in Traditional Medicine. In: Dugo G and Di Giacomo A (eds). Citrus: The Genus Citrus. London; New York: Taylor and Francis. p 577-601.

Iwabuchi H, Imayoshi Y, Yoshida Y and Saeki H. 2001. Application of Gas Chromatography-Olfactometry to Flavour Creation. In: Leland JV, Schieberle P, Buettner A and Acree TE (eds). Gas Chromatography-Olfactometry: The State of the Art. Washington DC: American Chemical Society. p 11-22.

Izquierdo L and Sendra JM. 1993. Citrus Fruits: Composition and Characterization. In: Macrae R, Robinson RK and Sadler MJ. Encyclopaedia of Food Science, Food Technology and Nutrition. London; San Diego: Academic Press. p 999-1006.

42 Jia M, Zhang QH and Min DB. 1998. Optimization of Solid-Phase Microextraction Analysis for Headspace Flavour Compounds of Orange Juice. J Agric Food Chem 46 : 2744-2747.

Johnson JR, Braddock RJ and Chen CS. 1996. Flavour Losses in Orange Juice during Ultrafiltration and Subsequent Evaporation. J Food Sci 61 : 540-543.

Kale PN and Adsule PG. 1995. Citrus. In: Salunkhe DK and Kadam SS (eds). Handbook of Fruit Science and Technology: Production, Composition, Storage and Processing. New York: Marcel Dekker, Inc. p 39-65.

Kefford JF. 1955. Recent Additions to Knowledge of the Chemistry of Citrus Fruits. Revs Pure Appl Chem 5: 77-98.

Kefford JF. 1966. Citrus Fruits and Processed Citrus Products in Human Nutrition. World Rev Nutr Diet 6: 197-249.

Kesterson JW and Hendrickson R. 1958. Utilization of Citrus By-Products. Econ Bot 12 (2): 164-185.

Kugler E and Kovats E. 1963. Information on Mandarin Peel Oil (In German). Helv Chim Acta 46 : 1480-1513.

Lee SJ. 2003. Finding Key Odourants in Foods: Gas Chromatography Olfactometry (GC-O). Food Sci Biotechnol 12 : 597-602.

Licandro G and Odio CE. 2002. Citrus By-products. In: Dugo G and Di Giacomo A (eds). Citrus: The Genus Citrus. London; New York: Taylor and Francis. p 159-178.

Likens ST and Nickerson GB. 1964. Detection of Certain Hop Oil Constituents in Brewing Products. Am. Soc. Brewing Chemists Proc : 5-13.

43 Lin J and Rouseff RL. 2001. Characterization of Aroma-Impact Compounds in Cold- Presed Grapefruit Oil Using Time-Intensity GC-Olfactometry and GC/MS. Flavour Fragr J 16 : 457-463.

Lin J, Jella P and Rouseff RL. 2002. Gas Chromatography-Olfactometry and Chemiluminescence Characterization of Grapefruit Juice Volatile Sulphur Compounds. In: Reineccius GA and Reineccius TA (eds). Heteroatomic Aroma Compounds. Washington DC: American Chemical Society. p 102-112.

Linssen JPH, Janssens JLGM, Roozen JP and Posthumus MA. 1993. Combined Gas Chromatography and Sniffing Port Analysis of Volatile Compounds of Mineral Water Packed in Polyethylene Laminated Packages. Food Chem 46 : 367-371.

Maccarone E, Campisi S, Fallico B, Rapisarda P and Sgarlata R. 1998. Flavour Components of Italian Orange Juice. J Agric Food Chem 46 : 2293-2298.

MacLeod AJ, MacLeod G and Subramanian G. 1998. Volatile Aroma Constituents of Orange. Phytochemistry 27 (7): 2185-2188.

Martín-Hernández MC and Juárez M. 1993. In: Macrae R, Robinson RK and Sadler MJ. Encyclopaedia of Food Science, Food Technology and Nutrition. London; San Diego: Academic Press. p 968-972.

McGill CR, Wilson AMR and Papanikolaou Y. 2004. Health Benefits of Citrus Juices. In: Wilson T and Temple NJ. Beverages in Nutrition and Health. Totowa: Humana Press Inc. p 63-78.

McGorrin RJ. 2002. Character Impact Compounds: Flavours and Off-Flavours in Foods. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 375-413.

Minh Tu NT, Thanh LX, Une A, Ukeda H and Sawamura M. 2002a. Volatile Constituents of Vietnamese Pummelo, Orange, Tangerine and Lime Peel Oils. Flavour Fragr J 17 : 169-174.

44

Minh Tu NT, Onishi Y, Choi HS, Kondo Y, Mitiku Bassore S, Ukeda H and Sawamura M. 2002b. Characteristic Odour Components of Citrus sphaerocarpa Tanaka (Kabosu) Cold-Pressed Peel Oil. J Agric Food Chem 50 : 2908-2913.

Mitiku SB, Sawamura M, Itoh T and Ukeda H. 2000. Volatile Components of Peel Cold-Pressed Oils of Two Cultivars of Sweet Orange ( Citrus sinensis (L). Osbeck) from Ethiopia. Flavour Frag J 15 : 240-244.

Morley-Bunker, M. 1999. Citrus. In: Jackson DI and Looney NE (eds). Temperate and Subtropical Fruit Production, 2 nd Edition. New York, Wallingford: CABI Pubshing. p 229-239.

Morton JF. 1987. Fruits of Warm Climates. Miami: Julia F. Morton. 505p.

Moshonas MG and Shaw PE. 1995. Fresh Orange Juice Flavour: A Quantitative and Qualitative Determination of the Volatile Constituents. In: Charalambous G (ed). Food Flavours: Generation, Analysis and Process Influence. Amsterdam: Elsevier Science. p 1479-1492.

Moshonas MG and Shaw PE. 1997. Quantitation of Volatile Constituents in Mandarin Juices and Its Use for Comparison with Orange Juices by Multivariate Analysis. J Agric Food Chem 45 : 3968-3972.

Moufida S and Marzouk B. 2003. Biochemical Characterization of Blood Orange, Sweet Orange, Lemon, Bergamot and Bitter Orange. Phytochemistry 62 : 1283-1289.

Mussinan CJ, Mookherjee BD and Malcolm GI. 1981. Isolation and Identification of Fresh Lemon Juice. In: Mokherjee BD and Mussinan CJ (eds). Essential Oils. Wheaton: Allured Publishing Corp. p 199-228.

Näf R, Velluz A and Meyer AP. 1996. Volatile Constituents of Blood and Blond Orange Juices: A Comparison. J Essent Oil Res 8: 587-595.

45 Naef R and Velluz A. 2001. Volatile Constituents in Extracts of Mandarin and Tangerine Peel. J Essent Oil Res 13 : 154-157.

Nagy S. 1977a. Lipids: Identification, Distribution and Importance. In: Nagy S, Shaw PE and Veldhuis MK (eds). Citrus Science and Technology, Volume I: Nutrition, Anatomy, Chemical Composition and Bioregulation. Westport: AVI Publishing Co. p 266-301

Nagy S. 1977b. Inorganic Elements. In: Nagy S, Shaw PE and Veldhuis MK (eds). Citrus Science and Technology, Volume I: Nutrition, Anatomy, Chemical Composition and Bioregulation. Westport: AVI Publishing Co. p 479-495.

Nagy S. 1996. Factors Affecting the Flavour of Citrus Fruits and Their Juice Products. In: In: Redd JB, Shaw PE, Hendrix, Jr. CM and Hendrix DL (eds). Quality Control Manual for Citrus Processing Plants, Volume III: Flavour: General, Systems, Important Volatiles, Shelf-Life, Specialty and By-Products, Relationships – Raw to the Processed Product, Miscellaneous Conversion Charts and Tables. Auburndale: Agscience Inc. p 102-133.

Nagy S and Nordby HE. 1970. The Effects of Storage Conditions on the Lipid Composition of Commercially Prepared Orange Juice. J Agric Food Chem 18 :593- 597.

Nagy S and Shaw PE. 1990. Factors Affecting the Flavour of Citrus Fruit. In: Morton ID and MacLeod AJ (eds). Food Flavours Part C: The Flavour of Fruits. Amsterdam: Elsevier. p 93-124.

Niessen WMA. 2001. Principles and Instrumentation of Gas Chromatography-Mass Spectrometry. In: Niessen WMA (Ed). Current Practice of Gas Chromatography- Mass Spectrometry. New York: Marcel Dekker Inc. p1-29.

Nishida R and Acree TE. 1984. Isolation and Characterization of Methyl Epijasmonate from Lemon ( Citrus limon Burm.). J Agric Food Chem 32 : 1001-1003.

46 Nisperos-Carriedo MO and Shaw PE. 1990. Comparison of Volatile Flavour Components in Fresh and Processed Orange Juices. J Agric Food Chem 38 : 1048- 1052.

Nordby HE and Nagy S. 1971. Comparative Citrus Fatty Acid Profiles of Triglycerides, Monogalactosyl Diglycerides, Steryl Esters, and Esterified Steryl Glucosides. Lipids 6: 554-561.

Nordby HE and Nagy S. 1980. Processing of Oranges and Tangerines. In: Nelson PE and Tressler DK. Fruit and Vegetable Juice Processing Technology, 3 rd Edition. Westport: AVI Publishing Co. p 35-96.

Núñez AJ, Maarse H and Bemelmans H. 1985. Volatile Flavour Components of Grapefruit Juice ( Citrus paradisi Macfadyen). J Sci Food Agric 36 : 757-763.

Ohloff G. 1972. Classification and Genesis of Food Flavours. Flavour Ind 3: 501-508.

Ohloff G. 1994. Scent and Fragrances: The Fascination of Odours and Their Chemical Perspectives. Berlin; New York: Springer-Verlag. p 127-198.

Ojeda de Rodriguez G, Ysambertt F, Sulbaran de Ferrer B and Cabrera L. 2003. Volatile Fraction Composition of Venezuelan Sweet Orange Essential Oil ( Citrus sinensis (L.) Osbeck). Ciencia 11 : 55-60.

Orav A. 2001. Identification of Terpenes by Gas Chromatography-Mass Spectrometry. In: Niessen WMA (ed). Current Practice of Gas Chromatography-Mass Spectrometry. New York: Marcel Dekker Inc. p 483-494.

Ortiz JM. 2002. Botany: Taxonomy, Morphology and Physiology of Fruits, Leaves and Flowers. In: Dugo G and Di Giacomo A (eds). Citrus: The Genus Citrus. London; New York: Taylor and Francis. p 16-35.

Parliment T. 2002. Solvent Extraction and Distillation Techniques. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 1-23.

47

Pawliszyn J. 1999. Quantitative Aspects of SPME. In: Pawliszyn J (ed). Applications of Solid Phase Microextraction. Cambridge: The Royal Society of Chemistry. p 3-21.

Pawliszyn J. 2001. Solid Phase Microextraction. In: Rouseff RL and Cadwallader KR (eds). Headspace Analysis of Foods and Flavours: Theory and Practice. New York: Kluwer Academic/Plenum Publishers. p 73-87.

Pérez-López AJ and Carbonell-Barrachina ÁA. 2006. Volatile Odour Components and Sensory Quality of Fresh and Processed Mandarin Juices. J Sci Food Agric 86 : 2404-2411.

Pérez-López AJ, Saura D, Lorente J and Carbonell-Barrachina ÁA. 2006. Limonene, Linalool, α-terpineol, and Terpinen-4-ol as Quality Control Parameters in Mandarin Juice Processing. Eur Food Res Technol 222 : 281-285.

Ranganna S, Govindarajan VS and Ramana KVR. 1986. Citrus Fruits: Varieties, Chemistry, Technology, and Quality Evaluation. Part II, Chemistry, Technology, and Quality Evaluation. A. Chemistry. Crit Rev Food Sci Nutr 18 : 313-386.

Redd JB and Hendrix, Jr. CM. 1996. Flavour Concentration Systems. In: Redd JB, Shaw PE, Hendrix, Jr. CM and Hendrix DL (eds). Quality Control Manual for Citrus Processing Plants, Volume III: Flavour: General, Systems, Important Volatiles, Shelf- Life, Specialty and By-Products, Relationships – Raw to the Processed Product, Miscellaneous Conversion Charts and Tables. Auburndale: Agscience Inc. p 62-101.

Reineccius G. 2002. Instrumental Methods of Analysis. In: Taylor AJ (ed). Food Flavour Technology. Sheffield: Sheffield Academic Press. p 210-251.

Reineccius G. 2006. Flavour Chemistry and Technology. Boca Raton: Taylor and Francis. p 33-72.

48 Ruberto G. 2001. Analysis of Volatile Components of Citrus Fruit Essential Oils. In: Jackson JF and Linskens HF (eds). Analysis of Taste and Aroma. New York: Springer. p 123-157.

Samson JA. 1990. Tropical fruits. New York: Longman Inc. 250p.

Sanchez NB, Lederer CL, Nickerson GB, Libbey LM and McDaniel MR. 1992. Sensory and Analytical Evaluation of Beers Brewed with Three Varieties of Hops and an Unhopped Beer. In: Charalambous G (ed). Food Science and Human Nutrition. Amsterdam: Elsevier. p 403-426.

Sarwono B. 1986. Jeruk dan Kerabatnya (Citrus and Its Family). Jakarta: Penebar Swadaya. p 1-2.

Sass-Kiss A, Toth-Markus M and Sass M. 2004. Chemical Composition of Citrus Fruits (Orange, Lemon and Grapefruit) with Respect to Quality Control of Juice Products. In: Shahidi F (ed). Nutraceutical Beverages: Chemistry, Nutrition, and Health Effects. Washington DC: American Chemical Society. p 24-34.

Saunt J. 2000. Citrus Varieties of the World. Norwich: Sinclair International Limited. 156p.

Sawamura M, Song HS, Choi HS, Sagawa K, Ukeda H. 2001. Characteristic aroma component of Tosa-buntan ( Citrus grandis Osbeck forma Tosa ) fruit. Food Sci Technol Res 7: 45-49.

Sawamura M, Minh Tu NT, Onishi Y, Ogawa E and Choi HS. 2004. Characteristic Odour Components of Citrus reticulata Blanco (Ponkan) Cold-Pressed Oil. Biosci Biotechnol Biochem 68 : 1690-1697.

Sawamura M, Minh Tu NT, Yu X and Xu B. 2005. Volatile Constituents of the Peel Oils of Several Sweet Oranges in China. J Essent Oil Res 17 : 2-6.

49 Sawamura M, Onishi Y, Ikemoto J, Minh Tu NT and Lan Phi NT. 2006. Characteristic Odour Components of Bergamot ( Citrus bergamia Risso) Essential Oil. Flavour Fragr J 21 : 609-615.

Schieberle P. 1995. New Developments in Methods for Analysis of Volatile Flavour Compounds and Their Precursors. In: Gaonkar AG (ed). Characterization of Food: Emerging Methods. Amsterdam; New York: Elsevier. p 403-431.

Selli S, Cabaroglu T, Canbas A. 2004. Volatile Flavour Components of Orange Juice Obtained from the cv. Kozan of Turkey. J Food Composition Anal 17 : 789-796.

Shaw PE. 1977. Essential Oils. In: Nagy S, Shaw PE and Veldhuis MK (eds). Citrus Science and Technology, Volume I: Nutrition, Anatomy, Chemical Composition and Bioregulation. Westport: AVI Publishing Co. p 427-478.

Shaw PE. 1979. Review of Quantitative Analyses of Citrus Essential Oils. J Agric Food Chem 27 : 246-257.

Shaw PE. 1986. The Flavour of Non-Alcoholic Fruit Beverages. In: Morton ID and MacLeod AJ (eds). Food Flavours Part B: The Flavour of Beverages. Amsterdam: Elsevier. p 337-368.

Shaw PE. 1991. Fruits II. In: Maarse H (ed). Volatile Compounds in Foods and Beverages. New York: Marcel Dekker Inc. p 305-327.

Shaw PE. 1996a. Shelf Life and Ageing of Citrus Juices, Juice Drinks and Related Soft Drinks. In: Redd JB, Shaw PE, Hendrix, Jr. CM and Hendrix DL (eds). Quality Control Manual for Citrus Processing Plants, Volume III: Flavour: General, Systems, Important Volatiles, Shelf-Life, Specialty and By-Products, Relationships – Raw to the Processed Product, Miscellaneous Conversion Charts and Tables. Auburndale: Agscience Inc. p 173-199.

Shaw PE. 1996b. Volatile Components Important to Citrus Flavours. In: Redd JB, Shaw PE, Hendrix, Jr. CM and Hendrix DL (eds). Quality Control Manual for Citrus

50 Processing Plants, Volume III: Flavour: General, Systems, Important Volatiles, Shelf- Life, Specialty and By-Products, Relationships – Raw to the Processed Product, Miscellaneous Conversion Charts and Tables.Auburndale: Agscience Inc. p 134-172.

Shaw PE and Moshonas MG. 1993. Volatile Components in Juice from Mandarin and Mandarin Hybrid Fruit. J Essent Oil Res 5: 101-104.

Shaw PE and Wilson CW, III. 1980. Importance of Selected Volatile Components to Natural Orange, Grapefruit, Tangerine and Mandarin Flavours. In: Nagy S and Attaway JA (eds). Citrus Nutrition and Quality. Washington DC: American Chemical Society. p 167-190.

Song HS, Sawamura M, Ito T, Kawashimo K and Ukeda H. 2000a. Quantitative Determination and Characteristic Flavour of Citrus junos (Yuzu) Peel Oil. Flavour Frag J 15 : 245-250.

Song HS, Sawamura M, Ito T, Ido A and Ukeda H. 2000b. Quantitative Determination and Characteristic Flavour of Daidai ( Citrus aurantium L. var. cyathifera Y. Tanaka) Peel Oil. Flavour Frag J 15 : 323-328.

Soule J and Grierson W. 1986. Anatomy and Physiology. In: Wardowski WF, Nagy S and Grierson W (eds). Fresh Citrus Fruits. Westport: The AVI Publishing Company Inc. p 1-22.

Stewart I. 1977. Provitamin A and Carotenoid Content of Citrus Juices. J Agric Food Chem 25 : 1132-1137.

Sugisawa H. 1981. Sample Preparation: Isolation and Concentration. In: Teranisi R, Flath R and Sugisawa H, eds. Flavour Research, Recent Advances. New York: Marcel Dekker Inc. p11-51.

Syaifullah M and Harijono T. 2004. Petani Jeruk Tersenyum Kembali ( Orange Farmers Smile Again). Kompas Daily, 15 June 2004.

51 Takeuchi H, Ubukata Y, Hanafusa M, Hayashi S and Hashimoto S. 2005. Volatile Constituents of Calamondin Peel and Juice ( Citrus madurensis Lour.) Cultivated in the Philippines. J Essent Oil Res 17 : 23-26.

Teranishi R. 1998. Challenges in Flavour Chemistry: An Overview. In: Mussinan CJ and Morello MJ (eds). Flavour Analysis: Developments in Isolation and Characterization. Washington DC: American Chemical Society. p 1-6.

Teranishi R, Issenberg P, Hornstein I and Wick EL. 1971. Flavour Research: Principles and Techniques. New York: Marcel Dekker Inc. 315p.

Ting SV and Attaway JA. 1971. Citrus Fruits. In: Hulme AC (ed.). The Biochemistry of Fruits and Their Products, Vol. 2. London: Academic Press. p 107-169.

Ting SV and Rouseff RL. 1986. Citrus Fruits and Their Products: Analysis and Technology. New York: Marcel Dekker Inc. 293p.

Ullrich F and Grosch W. 1987. Identification of the Most Intense Odour Compounds Formed during Autoxidation of Linoleic Acid. Z Lebensm Unters Forsch 184 : 277- 282. van Ruth SM. 2001a. Methods for Gas Chromatography-Olfactometry: A Review. Biomol Eng 17 : 121-128. van Ruth SM. 2001b. Evaluation of Three Gas Chromatography-Olfactory Methods: Comparison of Odour Intensity-Concentration Relationships of Eight Volatile Compounds with Sensory Headspace Data. Food Chem 74 : 341-347.

Van Ruth SM. 2004. Evaluation of Two Gas-Chromatography-Olfactory Methods: The Detection Frequency and Perceived Intensity Method. J Chromatogr A 1054 : 33- 37.

Vandercook CE. 1977. Organic Acids. In: Nagy S, Shaw PE and Veldhuis MK (eds).

52 Citrus Science and Technology, Volume I: Nutrition, Anatomy, Chemical Composition and Bioregulation. Westport: AVI Publishing Co. p 208.

Wampler TP. 2002. Analysis of Food Volatiles Using Headspace-Gas Chromatographic Techniques. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 25-54.

Wang Z and Paré JRJ. 1997. Gas Chromatography (GC): Principles and Applications. In: Paré JRJ And Bélanger JMR (Eds). Instrumental Methods in Food Analysis. Amsterdam: Elsevier. p 61-91.

Wardencki W, Michulec M and Curylo J. 2004. A Review of Theoretical and Practical Aspects of Solid-Phase Microextraction in Food Analysis. Int J Food Sci Tech 39 : 703-717.

Webber HJ. 1948. Cultivated Varieties of Citrus. In: Webber HJ and Batchelor LD (eds). The Citrus Industry, Volume I: History, Botany and Breeding. Berkeley: University of California Press. p 475-668.

Wells AH, Agcaoili F and Orosa MY. 1925. Philippine Citrus Fruits. Philip Journ Sci 28 (4): 453-527.

Werkhoff P, Brennecke S, Bretschneider W and Bertram HJ. 2002. Modern Methods for Isolating and Quantifying Volatile Flavour and Fragrance Compounds. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 139-204.

Weurman C. 1969. Isolation and Concentration of Volatiles in Food Odour Research. J Agric Food Chem 17 : 370-384.

Wilson CW, III and Shaw PE. 1981. Importance of Thymol, Methyl-N- Methylanthranilate and Monoterpene Hydrocarbons to the Aroma and Flavour of Mandarin Cold-Pressed Oils. J Agric Food Chem 29 : 494-496.

53 Wolford RW, Kesterson JW and Attaway JA. 1971. Physicochemical Properties of Citrus Essential Oils from Florida. J Agric Food Chem 19 : 1097-1105.

Wright J. 2004. Flavour creation. Carol Stream: Allured Publishing Corp. 247p.

Yajima I, Yanai T, Nakamura M, Sakakibara H and Hayashi K. 1979. Compositions of the Volatiles of Peel Oil and Juice from Citrus unshiu. Agric Biol Chem 43 : 259- 264.

Young RH. 1986. Fresh Fruit Cultivars. In: Wardowski WF, Nagy S and Grierson W (eds). Fresh Citrus Fruits. Westport: The AVI Publishing Company Inc. p 102-126.

Zhang Z and Pawliszyn J. 1993. Headspace Solid-Phase Microextraction. Anal Chem 65 : 1843-1852.

54 Chapter 3 Characterization of Volatile Compounds in Hand-Squeezed Juices of Selected Citrus Fruits from Asia

3.1. Abstract

Volatile compounds in the hand-squeezed juices of selected citrus fruits from Asia, namely Pontianak orange ( Citrus nobilis Lour. var. microcarpa Hassk.) from

Indonesia, Mosambi (Citrus sinensis Osbeck) from India and Dalandan ( Citrus reticulata Blanco) from the Philippines, were characterized. Compounds from the headspace of the juices were isolated by solid-phase microextraction (SPME) technique prior to separation with gas chromatograph (GC) and identification by the mass spectrometry (MS). Extracts of the juices were obtained by performing continuous liquid-liquid extraction with diethyl ether and the volatile compounds were quantified by using GC - flame ionization detector (FID). A total of 51 volatile compounds were detected in Pontianak orange, 50 in Mosambi and 41 in Dalandan juices. In general, they comprise terpenes, carbonyls, alcohols, esters and hydrocarbons, with limonene as the main compound. The juice of each citrus cultivar studied contained compounds not frequently reported in other citrus fruits, such as beta-chamigrene in Mosambi as well as tentatively identified 2,6-dimethyl-1,3,5,7- octatetraene and isopiperitenone in Pontianak orange and in Dalandan juices respectively. The characterization results showed that Mosambi and Dalandan juices

55 portray the typical profile of sweet orange and mandarin correspondingly, while

Pontianak orange demonstrates its own unique traits.

3.2. Introduction

The production of citrus fruits across the world shows a significant figure as indicated by its total global production that reached 110 millions metric tonnes in 2004

(FAOSTAT). Citrus fruits are largely processed for their juice, one of the most important commodities (Braddock, 1999). Despite the fact that the United States of

America and Brazil are the main producers of citrus fruits, Southeast Asia is believed to be the place of origin of citrus fruits (Davies and Albrigo, 1994). Ample studies have been carried out in order to investigate the volatile compounds present in many citrus fruits. The scope of the research ranges from the famous cultivar like Valencia and Navel oranges (Buettner and Schieberle, 2001), blood and blond oranges (Näf et al ., 1996), mandarin and Dancy tangerine (Naef and Velluz, 2001) and ponkan

(Sawamura et al ., 2004) to the local cultivars such as Italian blood and blond oranges

(Maccarone et al ., 1998), Japanese yuzu (Song et al ., 2000) and Turkish kozan (Selli et al ., 2004). Research was also extended from the citrus essential oils (Lota et al .,

2000; Ruberto, 2001; Högnadóttir and Rouseff, 2003) to the juices (Nisperos-Carriedo and Shaw, 1990; Brat et al ., 2003; Mahattanatawee et al ., 2005) as they have different aroma profiles.

There are many varieties of citrus fruits in the region of Asia that have distinct characteristics and are only consumed locally. Some of these varieties have great potential to be further explored, such as: Pontianak Orange ( Citrus nobilis Lour. var.

56 microcarpa Hassk.) from Indonesia, Mosambi ( Citrus sinensis Osbeck) from India and Dalandan ( Citrus reticulata Blanco) from the Philippines. According to the botanical classification, Pontianak orange is a mandarin. It has thin, fairly shiny and yellowish green-coloured skin. The flesh is orange in colour and it has a very sweet taste. Mosambi is one of the varieties of sweet oranges most widely cultivated in

India. The fruit is round in shape and moderately seedy. The colour of the skin is light yellow to pale orange at maturity. The flesh is somewhat firm and juicy although its flavour is fairly bland because of very low acidity. Dalandan belongs to mandarin family. It has green-coloured skin and is generally sour. The present study was undertaken to investigate the volatile components of the freshly-squeezed juices of selected citrus fruits from Asia: Pontianak orange, Mosambi and Dalandan.

3.3. Materials and Methods

3.3.1. Materials

Pontianak oranges, Mosambi and Dalandans were harvested in the month of August,

November and October 2004 respectively. They were acquired from a local fruit farm in Indonesia, India and the Philippines correspondingly. The freshly squeezed juices of citrus samples were obtained by careful hand-squeezing of citrus fruits with a citrus juicer (Citromatic MPZ 9; Braun, Germany).

3.3.2. Chemicals

Calcium chloride dihydrate, sodium chloride and sodium hydroxide were purchased from Merck (Darmstadt, Germany); diethyl ether was from J.T. Baker (Phillipsburg,

NJ, USA). Aroma standards were obtained from the following sources: Ethanol was

57 from Hayman (Witham, Essex, England); pentane was from Tedia (Fairfield, OH,

USA); ethyl acetate, ethyl butyrate, decanal, alpha-copaene and beta-chamigrene were from Fluka (Buchs, Switzerland); beta-elemene was from Apin Chemicals (Abingdon,

Oxfordshire, UK); sabinene, beta-pinene, dihydrocarvone and aromadendrene were from ChromaDex (Santa Ana, CA, USA); dehydro-p-cymene, ethyl-3- hydroxyhexanoate, carveol, alpha-humulene, alpha-farnesene and alkane standards

(C6-C19) were from Aldrich (St. Louis, MO, USA); acetaldehyde, methyl acetate, ethyl propanoate, ethyl isobutyrate, hexanal, (E)-2-hexenal, heptanal, alpha-pinene, myrcene, butyl butyrate, ethyl hexanoate, alpha-phellandrene, delta-3-carene, octanal, limonene, ocimene, gamma-terpinene, cis- and trans- linalool oxide, 1-octanol, terpinolene, linalool, nonanal, 1-terpineol, citronellal, 4-terpineol, p-cymen-8-ol, alpha-terpineol, citronellol, neral, L-carvone, geranial, perillaldehyde, thymol, undecanal, p-vinyl guaiacol, neryl acetate, geranyl acetate, beta-damascenone, dodecanal, beta-caryophyllene, valencene, alpha-sinensal and nootkatone were from

Firmenich Asia Pte. Ltd. (Singapore).

3.3.3. pH, Brix value and Titratable acidity

The pH of freshly squeezed citrus juices was measured by a pH-meter (410; Thermo

Orion, USA), while their total soluble solids were reported as oBrix and measured by refractometer (3T; Atago, Japan). The amount of acid present, mainly citric acid, in the juices was determined by titration method using 0.1M NaOH and phenolphthalein as the indicator until it gave a persistent pink colour (end point pH ≈ 8).

58 3.3.4. SPME

Five grams of freshly squeezed orange juice in 2M CaCl 2 (1:1) were transferred into a

10-mL vial containing a magnetic stirring bar. NaCl (26%) was added to the sample matrix to decrease the solubility of volatile compounds in the water phase and give the ‘salting out’ effect. The sample vial was air-tightly sealed by a Teflon septum and an aluminium cap. SPME fibre coated with 2cm-50/30 µm DVB

(divinylbenzene)/Carboxen/PDMS (polydimethylsiloxane) by Supelco (Bellefonte,

PA, USA) was manually inserted into the headspace of the sample vial. The extraction was done at 50 oC for 30 minutes prior to injection into GC/MS (Jia et al .,

1998). The diagram for the extraction of volatile compounds by SPME is shown in

Figure 3.1 .

Figure 3.1. Diagram for the isolation of headspace flavour compounds of orange juice by SPME (Jia et al., 1998)

3.3.5. Continuous liquid-liquid extraction

Freshly squeezed orange juice (500mL) was immediately poured into aqueous saturated CaCl 2 solution (400mL) in order to inhibit the enzymatic reactions

(Hinterholzer and Schieberle, 1998). The aqueous mixture was extracted with diethyl

59 ether (500mL) for 6 hours in a continuous liquid-liquid extractor (Normag, Ilmenau,

Germany). The extract was dried over Na 2SO 4 overnight and was finally concentrated with Vigreux column at 38±1 oC (Buettner and Schieberle, 1999).

3.3.6. Gas Chromatograph–Flame Ionization Detector (GC-FID)

The amount of volatile compounds present in the citrus juices was analyzed by using an Agilent 6890A gas chromatograph equipped with flame ionization detector.

Capillary column used was BPX-5 (5% phenyl/95% methyl polysilphenylene- siloxane – 30m x 0.25mm, 0.25 µm film thickness; SGE, Ringwood, Australia). The temperature was programmed from 35 oC to 200 oC at a rate of 4 oC/min and then ramped to 300 oC at 30 oC/min, with a 3-min final temperature hold. The injector and detector temperatures were set at 270 oC and 320 oC respectively. The carrier gas used was helium at a flow rate of 1.2mL/min. The analyses were done in triplicates and the retention indices of the compounds were determined based on a homologous series of n-alkanes (C 5-C19 ).

3.3.7. Gas Chromatograph/Mass Spectrometry (GC/MS)

The volatile compounds were characterized and identified with Shimadzu’s GC/MS

QP5000. The column and temperature program used was similar to those indicated above. Both injector and MS transfer line temperatures were set at 270 oC. The electron ionization (EI) method was used for the MS at the ionization energy of 70eV with the scan range of 40-300 m/z. The compounds were identified by comparison of mass spectra of the target compounds with those of the NIST (National Institute of

Standards and Technology) library and verified by the retention indices of pure standard compounds.

60 3.3.8. Linear Retention Index

For confirmation of volatile compounds identified, Linear Retention Index (LRI) was used. LRI value of each compound was obtained by comparing the retention times of the particular compound with that of closely eluting alkane standards, i.e. the alkanes that elute just before and just after the compound peak. The LRI was calculated by applying the following formula (van den Dool and Kratz, 1963):

t R(x) – t R(n) RI = —————— 100 + 100n tR(n+1) – t R(n)

tR(x) = retention time of compound x

tR(n) = retention time of closest alkane that elute before the compound x

tR(n+1) = retention time of closest alkane that elute after the compound x n = number of Carbon atom of closest alkane that elute before compound x

3.4. Results and Discussion

3.4.1. Chemical composition

Table 3.1. Chemical composition of various orange juice cultivars pH* oBrix* % Acidity*

Pontianak orange 4.78 ± 0.07 13.58 ± 0.63 0.24 ± 0.04

Mosambi 4.60 ± 0.08 10.05 ± 0.44 0.19 ± 0.02

Dalandan 3.55 ± 0.12 10.15 ± 0.44 1.18 ± 0.26

Valencia (1) 3.5 11.8 0.98

Tangerine (1) 3.2 14.33 1.26

Kozan (2) 3.4 12 0.93

*Average of 15 replicates (refer to Appendices A.1 to A.3) (1) Roberts and Gaddum (1937) (2) Selli et al . (2004)

61 The concentration of sugar and citric acid in citrus juices may contribute to the distinct flavour of each citrus variety by imparting the sweet and sour tastes to the juices (Tressler et al ., 1939). The pH, Brix value as well as acidity of each juice sample were, therefore, determined and also compared with the values of other orange juice cultivars as shown by literature (Roberts and Gaddum, 1937; Selli et al ., 2004).

The above results showed that each citrus cultivar has its characteristic chemical composition with regards to pH, Brix value and acidity. Brix value is commonly used to measure the total soluble solids content, i.e. total sugars, in the juice (Nagy and

Shaw, 1990). They also mentioned that in citrus juices, the only sugars found in significant quantities are sucrose, glucose and fructose, which usually occur in proximate ratio of 2:1:1 respectively. As shown above, Pontianak orange juice had the highest Brix value among the three citrus samples. Comparing with other varieties,

Pontianak orange juice had fairly similar Brix value to Tangerine while Mosambi and

Dalandan had the lowest Brix values.

In citrus juices acidity is mainly affected by the presence of citric acid; hence, the acidity usually refers to citric acid concentration. As citrus fruits mature, the total acidity drops while the total sugars gradually increase. At the same time, other changes that take place include the increase of juice content, colour change, decrease in bitter and astringent juice components, and development of desirable flavour compounds (Nagy and Shaw, 1990). The acidity of Dalandan was five times greater than that of Pontianak orange and Mosambi, and was comparable to other citrus cultivars, namely Tangerine, Kozan and Valencia. The pH values of the juices confirmed the high acid content of Dalandan but low in Pontianak orange and

Mosambi.

62

3.4.2. Volatile compounds in citrus juices

The list of volatile compounds present in the juice of Pontianak orange, Mosambi and

Dalandan was shown in Table 3.2 . There were 51 volatile compounds that could be characterized in freshly-squeezed Pontianak orange juice, 50 compounds in Mosambi juice and 41 compounds in Dalandan juice. The results confirmed that SPME method was able to extract more volatile compounds than solvent extraction method due to higher sensitivity of the former technique towards analytes (Cai et al ., 2001). In this study, the SPME fibre used was a mixture of Carboxen, polydimethylsiloxane

(PDMS), and divinylbenzene (DVB), the best rated fibre for SPME as it is ideal for a broad range of analyte polarities and molecular weights (Rega et al ., 2003).

Nevertheless, there were some compounds extracted by diethyl ether but not by the

SPME fibre. This might be due to low concentration or less volatility of the compounds in the headspace.

As in the case of other citrus cultivars, limonene was also the major compound present in the citrus varieties studied. It comprised 80-90% of the total volatile compounds in Pontianak orange and Dalandan juices while it was around 50% in

Mosambi juice. Even so, limonene is not the most important compound in oranges

(Jia et al ., 1998). With the exception of few compounds, the concentrations of most compounds were less than 1%.

63 e d Ident MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, RI MS, RI MS, RI MS, RI MS, RI MS, MS, RI MS, MS, RI MS (931) MS c ― ― ― ― ― ― ― ― ― ― 0.86 1.43 Sol Ext Sol % Conc b Dalandan + + + + + + ― ― ― ― ― ― SPME c ― ― ― ― ― ― ― ― ― ― 0.50 1.58 Sol Ext % Conc b Mosambi + + + + + + + + + + k orange, Dalandan k orange, and Mosambi juices ― ― SPME c ― ― ― ― ― ― ― ― ― ― 0.65 0.21 Sol Ext % % Conc b + + + + + + + + + + + ― PontianakOrange SPME a ― RI 934 926 504 613 712 758 806 859 910 804 536 Volatile ofcompoundsfreshly-squeezed the Pontiana Table 3.2. Table alpha-pinene alpha-thujene ethanol ethyl acetate ethyl propanoate ethyl isobutyrate hexanal (E)-2-hexenal heptanal ethyl ethyl butyrate methyl acetate Volatile compounds acetaldehyde 2. 4. 5. 6. 8. 9. 7. 3. 1. 12. 11. 10. No.

64 d MS Ident MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI c ― ― ― ― ― ― 1.01 0.69 0.42 0.50 0.47 86.86 Sol Ext Sol %Conc b Dalandan + + + + + + + + ― ― ― ― SPME c ― ― ― ― 0.37 0.44 0.60 0.47 0.33 0.29 0.68 52.23 Sol Ext % Conc b Mosambi + + + + + + + + + + ― ― SPME c ― ― ― ― ― ― 0.25 0.12 0.14 0.46 0.39 89.65 Sol Ext % % Conc b + + + + + + + + ― ― ― ― PontianakOrange SPME a RI 975 980 989 998 1061 1029 1012 1039 1033 1007 1006 1001 Volatile compounds gamma-terpinene 7-endo-ethenyl-bicyclo[4,2,0]- oct-1-ene octanal ocimene D-limonene delta-3-carene alpha-phellandrene ethyl ethyl hexanoate sabinene beta-pinene myrcene butyl butyrate butyl 24. 21. 20. 23. 22. 19. 18. 17. 13. 14. 15. 16. No.

65 e d MS Ident MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS (1163) MS c ― ― ― ― ― ― 0.71 0.70 0.80 0.37 0.41 0.41 Sol Ext Sol %Conc b Dalandan + + + + + + + + ― ― ― ― SPME c ― ― ― ― ― ― ― ― ― 1.28 0.78 0.83 Sol Ext % Conc b Mosambi + + + + + + ― ― ― ― ― ― SPME c ― ― ― ― ― ― 0.33 0.60 0.59 0.17 0.28 0.48 Sol Ext % % Conc b + + + + + + + + + ― ― ― PontianakOrange SPME a RI 1074 1084 1088 1090 1096 1104 1117 1133 1142 1137 1160 1161 cis-linalooloxide 1-octanol terpinolene trans-linalool oxide dehydro-p-cymene linalool nonanal ethyl-3-hydroxyhexanoate 1-terpineol 2,6-dimethyl-1,3,5,7-octatetraene citronellal trans beta-terpineol Volatile compounds 25. 26. 27. 28. 29. 30. 31. 32. 34. 33. 35. 36. No.

66 d MS Ident MS, MS, RI MS, RI MS, RI MS, RI MS, MS, RI MS, MS, RI MS, RI MS, RI MS, MS, RI MS, RI MS, RI c ― ― ― ― ― ― ― 0.65 1.61 0.21 0.52 0.29 Sol Ext Sol % Conc b Dalandan + + + + + + + + + ― ― ― SPME c ― ― ― ― ― ― 0.66 0.46 0.35 0.23 0.31 0.31 Sol Ext % Conc b Mosambi + + + + + + + + + + ― ― SPME c ― ― ― ― 0.97 0.49 0.20 0.20 0.38 0.84 0.49 0.24 Sol Ext % % Conc b + + + + + + + + + + ― ― PontianakOrange SPME a RI 1199 1210 1214 1217 1227 1193 1199 1241 1248 1237 1257 1280 p-cymen-8-ol alpha-terpineol decanal dihydrocarvone trans-carveol 4-terpineol 3,9-epoxy-p-mentha-1,8(10)- diene cis-carveol neral citronellol L-carvone geranial Volatile compounds 44 39. 40. 41. 42. 43. 37. 38. 45. 46. 47. 48. No.

67 e d MS Ident MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS (1339) MS c ― ― ― ― ― ― ― ― ― ― 0.41 0.59 Sol Ext Sol %Conc b Dalandan + + + + + + + ― ― ― ― ― SPME c ― ― ― ― ― ― ― ― ― ― ― 0.49 Sol Ext % Conc b Mosambi + + + ― ― ― ― ― ― ― ― ― SPME c ― ― ― ― ― ― ― 0.35 0.14 0.33 0.35 0.14 Sol Ext % % Conc b + + + + + + + ― ― ― ― ― PontianakOrange SPME a RI 1340 1328 1415 1397 1314 1394 1303 1294 1285 1385 1380 1367 Volatile compounds delta-elemene p-vinylguaiacol dodecanal beta-elemene undecanal beta-damascenone thymol perillaldehyde isopiperitenone geranyl geranyl acetate alpha-copaene neryl neryl acetate 54. 53. 60. 59. 52. 58. 51. 50. 49. 57. 56. 55. No.

68 e e e d MS MS MS Ident MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS (1485) MS MS (1524) MS MS (1480) MS c ― ― ― ― ― ― ― ― ― ― ― ― Sol Ext Sol %Conc b Dalandan + + ― ― ― ― ― ― ― ― ― ― SPME c ― ― ― ― ― ― ― ― 0.91 0.68 1.79 32.35 Sol Ext % Conc b Mosambi + + + + + + + + ― ― ― ― SPME c ― ― ― ― ― ― ― ― ― 0.11 0.31 0.13 Sol Ext % % Conc b + + + ― ― ― ― ― ― ― ― ― PontianakOrange SPME a RI 1486 1481 1490 1528 1514 1467 1489 1509 1628 1451 1498 1432 Volatile compounds beta-selinene gamma-selinene germacrene d delta-cadinene eremophilene alpha-humulene beta-chamigrene alpha-farnesene hexadecanal aromadendrene valencene beta-caryophyllene 65. 64. 67. 71. 70. 63. 66. 69. 72. 62. 68. 61. No.

69 d Ident MS,RI MS,RI MS,RI c ― ― Sol Ext Sol % Conc b Dalandan + ― SPME c ure compounds. standard acted juices. acted This the based peakis on area ― 1.09 Sol Ext % Conc b ) ) indicatesthatthe were compoundsnot detected. Mosambi ― + ― SPME c ― ― Sol Ext % Conc % b ― ― PontianakOrange SPME ; RI = verifiedRI ;with= theretention indices the p of eadspace methodSPME ( while ation) theofation) compounds present in the solvent extr a RI 1764 1837 alpha-sinensal nootkatone Volatile compounds from theGC-FID analysis. from 73. 74. (+) indicates(+) that the compoundsdetected were h by = by identifiedMS librarymatching MS withspectra No. relative concentration (mean triplicateof determin value RI The the based literatureon (Adams, 2007) Experimental RetentionLinear Indices. a a b c d e

70 McGorrin (2002) reported that there was no single flavour impact compound known for oranges. It is believed that citrus flavour is the result of a complex combination of terpene, aldehyde, ester and alcohol volatiles in specific proportions (Shaw, 1991).

Among these compounds, the major contributors to orange juice flavour, as reported by Ahmed et al . (1978a), were octanal, decanal, linalool, ethyl butyrate, acetaldehyde, citral (neral and geranial), citronellal, dodecanal, limonene, myrcene, nonanal, alpha- pinene and (E)-2-hexenal. All of these compounds were detected in Pontianak orange juice. Dodecanal was not found in Mosambi juice while in Dalandan juice, citronellal and (E)-2-hexenal were not detected.

Apart from limonene, other terpenes dominated as major compounds in the three citrus juice samples. Nevertheless, not all terpenes displayed aroma activity

(Högnadóttir and Rouseff, 2003). Valencene (32.4%) was the second most abundant compound found in Mosambi juice. Shaw (1991) mentioned that valencene might contribute positively to orange flavour but Elston et al. (2005) reported that it had no direct contribution, either positive or negative, to orange aroma. In Dalandan juice, gamma-terpinene accounted for 1% of the total compounds and it was reported to be one of the important contributors for mandarin flavour (Shaw, 1991). Myrcene was the second most abundant terpene found in Pontianak orange juice (2.5%). Its taste is

‘almost citrusy’ and ‘sweet-balsamic-herbaceous’ at concentration below 10 ppm while higher concentration tends to give pungent, bitter and ‘gassy’ notes (Arctander,

1994).

Aldehydes are found to play a major role in orange flavour (Ahmed et al ., 1978b). A total of 14 aldehydes ( Table 3.2 ) were found in the juices. Pontianak orange juice

71 contained the highest amount of aldehydes while Dalandan juice was the least. Those found in all three cultivars were acetaldehyde, octanal, nonanal, decanal, neral, geranial and perillaldehyde. Citral, a mixture of neral and geranial isomers, is the basic flavour-impact compound of lemon (McGorrin, 2002). It was reported by Shaw

(1986) that citronellal, octanal and nonanal were generally considered important contributors to orange flavour while decanal and dodecanal made negative contribution to orange juice flavour. He also added that the relative contribution of individual compound to overall juice flavour was only significant when its concentration in the juice reached its flavour threshold without neglecting the interaction with other volatile and non-volatile compounds. Esters also play important role in contributing to orange flavour. Ethyl acetate and ethyl butyrate were found in all citrus juice samples. Together with ethyl propanoate and ethyl isobutyrate, they are the important contributors to a desirable orange flavour (Ahmed et al., 1978b;

Hinterholzer and Schieberle, 1998). These four esters were detected in Pontianak orange juice.

Alpha-terpineol, a degradation product of d-limonene, and p-vinylguaiacol (PVG), a degradation product of ferulic acid, were detected in Pontianak orange juice. Both compounds are known to be off-flavour contributors to orange juice when they exceed certain concentrations (Nisperos-Carriedo and Shaw, 1990; McGorrin, 2002). Alpha terpineol, however, is commonly found in almost all citrus fruits, including Mosambi and Dalandan juice, and is frequently used in synthetic citrus flavourings at the right concentration (Arctander, 1994).

72 Shaw (1991) reported that thymol and methyl-N-methyl anthranilate (MNMA) were the main contributors to mandarin flavour, with additional contributions from beta- pinene, gamma-terpinene and alpha-sinensal. All of these compounds, except

MNMA, were identified in Dalandan juice. However, Evans and Rouseff (2001) indicated that MNMA was not found to be major aroma active compound in mandarin. Even though Pontianak orange is a mandarin, only beta-pinene and gamma-terpinene were found in its juice while the rest of the important flavour contributors of mandarin were not.

There were few volatile compounds not previously reported in other orange juices but detected and tentatively identified in Pontianak orange juice. The pure standards of these compounds are not available commercially. They were 2,6-dimethyl-1,3,5,7- octatetraene, 7-endo-ethenyl-bicyclo[4,2,0]-oct-1-ene and 3,9-epoxy-p-mentha-

1,8(10)-diene. The latter two compounds were also found in Mosambi juice. There were also some compounds detected in Mosambi orange juice not commonly reported in other orange juices, such as methyl acetate, beta-chamigrene and tentatively identified eremophilene. The aroma activities of these compounds in citrus flavour are not well documented. Few volatile compounds were found in Dalandan juices but not in Pontianak orange or in Mosambi juices, namely alpha-phellandrene, 1-terpineol, trans beta-terpineol, p-cymen-8-ol, beta-damascenone and alpha-farnesene. These compounds were reported previously in mandarin juices (Araki and Sakakibara, 1991;

Chisholm et al ., 2004). Isopiperitenone was the only compound found and identified tentatively in Dalandan orange juice not previously reported in mandarin or other citrus juices. It was reported to be found in ponkan oil and have oily aroma upon sniffing with GC-O (Sawamura et al ., 2004).

73

Each citrus cultivar investigated in this research possessed a characteristic profile of volatile compounds that contributed to its distinct flavour attributes. The presence of valencene and nootkatone in Mosambi juice and the absence of thymol and alpha- sinensal implied its resemblance to the typical sweet orange cultivars. However, the presence of terpenes such as beta-chamigrene and aromadendrene, and tentative identification of eremophilene, beta-selinene and gamma-selinene denotes its uniqueness since these compounds are not commonly found in the sweet oranges. In

Dalandan juice, the presence of gamma-terpinene, alpha-sinensal, beta-pinene and thymol signified its likeness to the typical mandarin cultivars. Nonetheless, the tentatively identified isopiperitenone and rare presence of beta-damascenone in mandarin being found in Dalandan juice marks the characteristic feature of this cultivar. Despite the fact that Pontianak orange is a mandarin, its profile of volatile compounds does not reflect that fully, particularly with the absence of thymol and alpha-sinensal. While some of the terpenes present were quite typical of sweet orange, the absence of valencene shows that Pontianak orange has its own notable traits and merits further investigation. Apart from the presence of various volatile compounds, the combination of chemical properties, such as acidity and sugar content, might also play a role behind the sweetness of Pontianak orange, the mild-taste of Mosambi as well as the sourness of Dalandan.

The characterization of volatile compounds in each citrus cultivar by GC/MS provided limited information on which compounds were the major contributors to the flavour quality of that particular cultivar. It is well known that the aroma activity of each compound in particular food product does not depend merely on its

74 concentration in the product but on its flavour threshold as well as its interaction with other volatile and non-volatile compounds. What complicates the matter is that many aroma-active compounds in orange juice are potent low-level volatiles difficult to detect using typical FID or MS detectors. On the contrary, major volatiles in orange juice had little or no aroma activity (Mahattanatawee et al., 2005).

References

Adams RP. 2007. Identification of Essential Oil Components by Gas Chromatography/Mass Spectrometry. Illinois: Allured Publishing Corp

Ahmed EM, Dennison RA and Shaw PE. 1978a. Effect of Selected Oil and Essence Volatile Components on Flavour Quality of Pumpout Orange Juice. J Agric Food Chem 26 : 368-372.

Ahmed EM, Dennison RA, Dougherty RH and Shaw PE. 1978b. Flavour and Odour Thresholds in Water of Selected Orange Juice Components. J Agric Food Chem 26 : 187-191.

Araki C and Sakakibara H. 1991. Changes in the Volatile Flavour Compounds by Heating Satsuma Mandarin ( Citrus unshiu Marcov.) Juice. Agric Biol Chem 55 : 1421- 1423.

Arctander S. 1994. Perfume and Flavour Chemicals (Aroma Chemicals), volume I and II. Illinois: Allured Publishing Corp.

Braddock RJ. 1999. Handbook of Citrus By-Products and Processing Technology. New York: Wiley-Interscience. 247p.

75 Brat P, Rega B, Alter P, Reynes M and Brillouet JM. 2003. Distribution of Volatile Compounds in the Pulp, Cloud and Serum of Freshly Squeezed Orange Juice. J Agric Food Chem 51 : 3442-3447.

Buettner A and Schieberle P. 1999. Characterization of The Most Odour-Active Volatiles in Fresh, Hand-Squeezed Juice of Grapefruit ( Citrus paradisi Macfayden). J Agric Food Chem 47 : 5189-5193.

Buettner A and Schieberle P. 2001. Evaluation of Aroma Differences between Hand- Squeezed Juices from Valencia Late and Navel Oranges by Quantitation of Key Odourants and Flavour Reconstitution Experiments. J Agric Food Chem 49 : 2387- 2394.

Cai J, Liu B and Su Q. 2001. Comparison of Simultaneous Distillation Extraction and Solid-Phase Microextraction for the Determination of Volatile Flavour Components. J Chromatogr A 930 : 1-7.

Chisholm MG, Jell JA, Cass Jr DM and Morrison ML. 2004. Identification and Comparison of the Most Intense Odourants Found in Citrus reticulata Blanco, cv. Clementine and cv. Murcott (Honey Tangerine) Using Gas Chromatography- Olfactometry. Abstracts of Papers, 228 th ACS National Meeting, Philadelphia, PA, USA, August 22-26, 2004. Washington DC: American Chemical Society. AGFD-104.

Davies FS and Albrigo LG. 1994. Citrus. Wallingford: CAB International. 254p.

Elston A, Lin J and Rouseff RL. 2005. Determination of the Role of Valencene in Orange Oil as a Direct Contributor to Aroma Quality. Flavour Fragr J 20 : 381-386.

Evans KC and Rouseff RL. 2001. Characterization of Aroma Active Compounds in Mandarin ( Citrus reticulata Blanco) Juice Using Gas Chromatography-Olfactometry. Abstracts of Papers, 222 nd ACS National Meeting, Chicago, IL, USA, August 26-30, 2001. Washington DC: American Chemical Society. AGFD-143.

FAOSTAT Agricultural Data. http://faostat.fao.org/ (last accessed 30 June 2008)

76

Hinterholzer A and Schieberle P. 1998. Identification of the Most Odour-Active Volatiles in Fresh, Hand-Extracted Juice of Valencia Late Oranges by Odour Dilution Techniques. Flavour Fragr J 13 : 49-55.

Högnadóttir A and Rouseff RL. 2003. Identification of Aroma Active Compounds in Orange Essence Oil Using Gas Chromatography-Olfactometry and Gas Chromatography-Masss Spectrometry. J Chromatogr A 998 : 201-211.

Jia M, Zhang QH and Min DB. 1998. Optimization of Solid-Phase Microextraction Analysis for Headspace Flavour Compounds of Orange Juice. J Agric Food Chem 46 : 2744-2747.

Lota M-L, Serra DR, Tomi F and Casanova J. 2000. Chemical Variability of Peel and Leaf Essential Oils of Mandarins from Citrus reticulata Blanco. Biochem Syst Ecol 28 : 61-78.

Maccarone E, Campisi S, Fallico B, Rapisarda P and Sgarlata R. 1998. Flavour Components of Italian Orange Juice. J Agric Food Chem 46 : 2293-2298.

Mahattanatawee K, Rouseff R, Valim MF and Naim M. 2005. Identification and Aroma Impact of Norisoprenoids in Orange Juice. J Agric Food Chem 53 : 393-397.

McGorrin RJ. 2002. Character Impact Compounds: Flavours and Off-Flavours in Foods. In: Marsili R (ed). Flavour, Fragrance and Odour Analysis. New York: Marcel Dekker, Inc. p 375-413.

Näf R, Velluz A and Meyer AP. 1996. Volatile Constituents of Blood and Blond Orange Juices: A Comparison. J Essent Oil Res 8: 587-595.

Naef R and Velluz A. 2001. Volatile Constituents in Extracts of Mandarin and Tangerine Peel. J Essent Oil Res 13 : 154-157.

77 Nagy S and Shaw PE. 1990. Factors Affecting the Flavour of Citrus Fruit. In: Morton ID and MacLeod AJ (eds). Food Flavours Part C: The Flavour of Fruits. Amsterdam: Elsevier. p 93-124.

Nisperos-Carriedo MO and Shaw PE. 1990. Comparison of Volatile Flavour Components in Fresh and Processed Orange Juices. J Agric Food Chem 38 : 1048- 1052.

Rega B, Fournier N and Guichard E. 2003. Solid Phase Microextraction (SPME) of Orange Juice Flavour: Odour Representativeness by Direct Gas Chromatography Olfactometry (D-GC-O). J Agric Food Chem 51 : 7092-7099.

Roberts and Gaddum. 1937. Composition of Citrus Fruit Juices. Ind Eng Chem 29 : 574-575.

Ruberto G. 2001. Analysis of Volatile Components of Citrus Fruit Essential Oils. In: Jackson JF and Linskens HF (eds). Analysis of Taste and Aroma. New York: Springer. p 123-157.

Sawamura M, Minh Tu NT, Onishi Y, Ogawa E and Choi HS. 2004. Characteristic Odour Components of Citrus reticulata Blanco (Ponkan) Cold-Pressed Oil. Biosci Biotechnol Biochem 68 : 1690-1697.

Selli S, Cabaroglu T, Canbas A. 2004. Volatile Flavour Components of Orange Juice Obtained from the cv. Kozan of Turkey. J Food Composition Anal 17 : 789-796.

Shaw PE. 1986. The Flavour of Non-Alcoholic Fruit Beverages. In: Morton ID and MacLeod AJ (eds). Food Flavours Part B: The Flavour of Beverages. Amsterdam: Elsevier. p 337-368.

Shaw PE. 1991. Fruits II. In: Maarse H (ed). Volatile Compounds in Foods and Beverages. New York: Marcel Dekker Inc. p 305-327.

78 Song HS, Sawamura M, Ito T, Kawashimo K and Ukeda H. 2000. Quantitative Determination and Characteristic Flavour of Citrus junos (Yuzu) Peel Oil. Flavour Frag J 15 : 245-250.

Tressler DK, Joslyn MA and Marsh GL. 1939. Fruit and Vegetable Juice. Westport: AVI Publishing Co. van den Dool H and Kratz PD. 1963. A Generalization of the Retention Index System Including Linear Temperature Programmed Gas—Liquid Partition Chromatography. J Chromatogr 11 : 463-471.

79 Chapter 4 Characterization of Volatile Compounds in Peel Oils of Selected Citrus Fruits from Asia

4.1. Abstract

Volatile compounds of the hand-pressed peel oil from the selected citrus fruits from

Asia, namely Indonesian Pontianak oranges ( Citrus nobilis Lour. var. microcarpa

Hassk.), Indian Mosambi ( Citrus sinensis Osbeck) and Philippine Dalandans ( Citrus reticulata Blanco), were characterized by GC-FID and GC/MS. A total of 32 volatile compounds were found in Pontianak orange, 29 in Mosambi and 37 in Dalandan peel oils. Limonene dominated the composition of each citrus oil and most of the volatile compounds were present in the concentrations less than 0.1%. The characteristic compound of sweet orange oil, delta-3-carene, was found in Mosambi peel oil while some characteristic compounds of mandarin oil, such as thymol, alpha-sinensal, beta- pinene and gamma-terpinene, were found in Dalandan oil. On the other hand, the absence of some important contributor compounds to mandarin family in Pontianak orange showed that it was a mandarin that has unique characteristic flavour. There was a difference of profiles of volatile compounds between the freshly-squeezed juices and hand-pressed peel oils for similar citrus cultivars, where generally the juices were richer in ester derivatives than the oil.

80 4.2. Introduction

Citrus fruits belong to the family Rutaceae, in which the leaves usually possess transparent oil glands and the flowers contain an annular disk (Kale and Adsule,

1995). The fruits are largely produced for their juices and essential oils as the by- products. Citrus oils are mainly obtained from the peel and cuticles of the fruits by cold-pressing method. The oils are greatly utilized as the flavourings in the food industries (Colombo et al ., 2002) and also for its fragrance in perfume and aromatherapy (Baser and Demirci, 2007). Ample studies have been carried out to investigate the volatile compounds in the various citrus oils (Lota et al., 2000; Naef and Velluz, 2001; Ruberto, 2001; Högnadóttir and Rouseff, 2003). Asia region is believed to be the place of origin of citrus fruits (Davies and Albrigo, 1994) and due to massive hybridization, there are many varieties of citrus fruits in this region that have distinct flavour and are only known by the locals. The cultivars selected for this research are Pontianak Orange ( Citrus nobilis Lour. var. microcarpa Hassk.) from

Indonesia, Mosambi ( Citrus sinensis Osbeck) from India and Dalandan ( Citrus reticulata Blanco) from the Philippines. The present study was undertaken to investigate the volatile components of the peel oils of these three citrus fruit varieties.

4.3. Materials and Methods

4.3.1. Materials

Pontianak oranges, Mosambi and Dalandans were harvested in the month of August,

November and October 2004 respectively. They were acquired from a local fruit farm in Indonesia, India and the Philippines correspondingly. The hand-pressed peel oil was obtained by careful hand-squeezing of the citrus peels.

81

4.3.2. Chemicals

The standards for aroma compounds were obtained from the following sources:

Nonanal, decanal, citral (a mixture of neral and geranial) and alpha-copaene were from Fluka (Buchs, Switzerland); beta-elemene was from Apin Chemicals

(Oxfordshire, UK); sabinene, beta-pinene and beta-farnesene were from ChromaDex

(CA, USA); limonene oxide, alpha-humulene, alpha-farnesene and alkane standards

(C9-C18) were from Aldrich (MO, USA); the rest of aroma standards, namely alpha- pinene, camphene, myrcene, alpha-phellandrene, delta-3-carene, octanal, limonene, gamma-terpinene, terpinolene, linalool, citronellal, alpha-terpineol, carvone, perillaldehyde, thymol, undecanal, neryl acetate, dodecanal, valencene, beta- caryophyllene and alpha-sinensal were obtained from Firmenich Asia Pte. Ltd.

(Singapore).

4.3.3. Gas chromatograph-Flame Ionization Detector (GC-FID)

The quantitative analysis of the citrus peel oil was performed using an Agilent 6890A gas chromatograph equipped with flame ionization detector using a BPX-5 capillary column (5% phenyl/95% methyl polysilphenylene-siloxane – 30m x 0.25mm, 0.25 µm film thickness; SGE, Australia). The temperature program used was 35 oC to 200 oC at a rate of 4 oC/min and then ramped to 300 oC at 30 oC/min, with a 3-min final temperature hold. The injector and detector temperatures were set at 270 oC and 320 oC respectively. Oil samples of 1 µL were injected at the split ratio of 100:1. The carrier gas was helium at a flow rate of 1.2mL/min. The analyses were done in triplicates and the retention indices of the compounds were determined based on a homologous series of n-alkanes (C 9-C18 ).

82

4.3.4. Gas Chromatograph/Mass Spectrometry (GC/MS)

Analysis of the citrus peel oil was carried out using Shimadzu’s GC/MS QP5000 system. The column and temperature program used was similar to those indicated above. Both injector and MS transfer line temperatures were set at 270 oC. The electron ionization (EI) method was used for the MS at the ionization energy of 70eV with the scan range of 40-300 m/z for 50 minutes. The compounds were identified by comparison of mass spectra of the target compounds with those of the NIST (National

Institute of Standards and Technology) library and by retention indices.

4.4. Results and Discussion

Table 4.1 listed the volatile compounds found in the peel oil of Pontianak orange,

Mosambi and Dalandan. Thirty-two compounds were detected in Pontianak orange peel oil and they were mainly terpenes, esters, alcohols and carbonyl compounds. The major compound was limonene (96.3%), followed by myrcene (2.0%). The rest of the compounds were present in the concentrations of less than 1%. Preliminary work of an undergraduate project on cold pressed Pontianak orange peel oil reported that limonene and myrcene comprised around 95% and 2.4% of the volatile compounds, respectively (Nugroho, 1995). Nevertheless, compounds listed in Table 4.1 , like alpha-thujene, camphene, alpha-phellandrene, gamma-terpinene, terpinolene, limonene oxide, carvone, perillaldehyde and alpha-copaene, were not reported in that study. This discrepancy may be due to factors such as source of oranges, method of oil isolation and GC condition.

83 c ― MS Ident MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, tr ― 0.18 0.14 0.95 0.01 0.15 0.37 1.89 4.30 0.09 90.84 Dalandan b nge, nge, Mosambi andDalandan tr tr ― ― ― 0.44 0.54 0.33 0.04 2.01 0.41 95.57 Mosambi % Concentration % tr tr ― ― ― 0.11 0.56 0.14 0.01 0.19 2.02 96.27 Pontianak Orange Pontianak a RI 927 935 953 977 982 994 1015 1013 1010 1071 1052 1024 Volatilecompounds oilpeelthe of of Pontianakora ) 16 Table4.1. H 10 (C d alpha-thujene alpha-pinene Compounds camphene octanal sabinene beta-pinene delta-3-carene myrcene alpha-phellandrene gamma-terpinene limonene unknown 1. 2. 10. No. 3. 9. 4. 5. 8. 6. 7. 12. 11.

84 c ― MS Ident MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, tr tr tr ― 0.33 0.28 0.01 0.01 0.05 0.09 0.01 0.09 Dalandan b tr tr tr tr ― 0.08 0.15 0.08 0.01 0.09 0.01 0.05 Mosambi % Concentration % tr tr ― ― 0.05 0.20 0.01 0.01 0.14 0.01 0.01 0.06 Pontianak Orange Pontianak a RI 1094 1115 1089 1145 1164 1199 1218 1245 1255 1286 1293 1086 ) 16 H 10 (C e linalool nonanal unknown trans-limoneneoxide citronellal alpha-terpineol decanal neral carvone geranial isopiperitenone Compounds terpinolene 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. No. 13. 14.

85 c MS MS Ident MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, tr tr tr ― 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.03 Dalandan b tr ― ― ― ― ― ― ― 0.01 0.01 0.03 0.02 Mosambi % Concentration % ― ― 0.03 0.01 0.01 0.02 0.01 0.03 0.02 0.01 0.02 0.02 Pontianak Orange Pontianak a RI 1494 1468 1430 1459 1421 1383 1398 1340 1366 1299 1321 1320 germacrene D alpha-humulene beta-caryophyllene Compounds beta-farnesene* dodecanal alpha-copaene beta-elemene delta-elemene neryl acetate perillaldehyde undecanal thymol 36. 35. 33. No. 34. 32. 30. 31. 28. 29. 25. 27. 26.

86 c MS MS MS MS MS MS Ident Ident MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, MS, RI MS, tr tr ― ― ― ― 0.02 0.01 0.06 0.02 0.01 0.06 dardcompounds Dalandan Dalandan b ― ― ― ― ― ― 0.04 0.02 0.01 0.04 0.02 0.01 Mosambi Mosambi % Concentration % ), 93 (100), 93 (47), 79 ), 41 (22), 53 (55). ), 93 ), 41 (14), 65(100), (49), 77 (46). tr tr ― ― ― ― ― ― 0.01 0.01 0.01 0.01 ; RI = ;= RI confirmedwith retentionindices of the stan ean triplicateean of determinations Pontianak Orange Pontianak Pontianak Orange Pontianak a RI 1504 1508 1510 1529 1574 1771 1504 1508 1510 1529 1574 1771 C, C, m/z 136 (rel. 105 int.): (74), [M]+ (23 121 (66), C,m/z 136 (rel.105 int.): (99), [M]+ (22 121 (46), o o valencene germacrene B alpha-farnesene* Compounds delta-cadinene gamma-elemene alpha-sinensal valencene germacrene B alpha-farnesene* delta-cadinene gamma-elemene alpha-sinensal = not detected;= tr =trace (<0.01%) MS, 70 eV,70270 MS, MS, 70270 MS, eV, Experimental Retention Linear Indices relative relative concentration theon based GC peak area,m MSby identified matching = MS librarywith spectra 37. 38. 39. No. a b c d e 40. 41. 42. ― Correctisomer * not identified 37. 38. 39. 40. 41. 42.

87 Some compounds identified in the Pontianak orange peel oil were not detected in the juice (Dharmawan et al ., 2007). These compounds were camphene, alpha- phellandrene, alpha-copaene and beta-caryophyllene. Similar to the juice, the important flavour contributors of mandarin, thymol and methyl-N-methyl anthranilate

(MNMA), as described by Shaw (1991), were not found in Pontianak orange peel oil either. This result indicates that Pontianak orange is a type of mandarin that has a flavour profile distinct from typical mandarins. This might be due to the fact that

Pontianak orange is a possible hybrid between sweet orange and mandarin (Morton,

1987).

The amount of compounds found in the peel oil of Mosambi was around half of those found in the juice (Dharmawan et al ., 2007). Out of 29 volatile compounds that could be identified from the peel oil of this sweet orange ( Table 4.1 ), limonene constituted

95.6% of the volatile compounds. Other major compounds found were myrcene

(2.0%), alpha-pinene (0.5%) and sabinene (0.3%). All of these compounds were reported by Boelens (1991) to be the dominant monoterpenes found in the peel oil of sweet orange. While Mosambi juice contained 32.4% of valencene, its concentration in the peel oil was less than 0.1%. Among the three citrus varieties, delta-3-carene was only discovered in Mosambi. It was found to be characteristic for Brazilian sweet orange oil (Boelens, 1991). The compounds that were not found in the juice but only in the peel oil were dodecanal, beta-cubebene, beta-farnesene, germacrene D and alpha-farnesene.

Similar to Pontianak orange and Mosambi, limonene (90.8%) was also the most abundant compound found in Dalandan peel oil ( Table 4.1 ). The other compounds

88 found at a concentration higher than 1% were gamma-terpinene (4.3%) and myrcene

(1.9%). Being a typical mandarin, the major contributors to mandarin flavour namely thymol, alpha-sinensal, beta-pinene and gamma-terpinene (Shaw, 1991) were also detected in the Dalandan peel oil while MNMA was not found. The tentatively identified compound, isopiperitenone, was the only compound found in Dalandan peel oil but not commonly reported in other citrus oils. It was found in the oil of Seville orange ( Citrus aurantium ) as characterized by Yang et al . (1992). As it was also found in Dalandan juice, this might point to its significance in contributing to

Dalandan flavour.

Two unknown compounds were found in Dalandan and Mosambi oils. The linear retention indices (LRI) of the compounds were 1021 and 1086, respectively. Both compounds had fairly similar spectra and most likely are isomers of C 10 H16 compounds. It seems reasonable that the compounds are in p-menthadiene series with isopropylidene residues due to high abundance of m/z 93. These unknown compounds could be the potential contributors to Dalandan and Mosambi flavour. The identification of these compounds and their significance in contributing to Dalandan and Mosambi flavours are subject to further verification.

On the whole, the peel oil of each citrus cultivar contained a distinct profile of volatile compounds ( Table 4.1 ). It provided a database on which compounds may play a major role in contributing to the flavour profile of each citrus cultivar. There was also a discrepancy of profiles of flavour compounds between the freshly-squeezed juices and hand-pressed peel oils for similar citrus cultivars, in terms of the volatile compounds present and their composition. The juices were generally richer in ester

89 derivatives and hence, imparted particular fruity-notes as indicated by Nagy and Shaw

(1990). The result of present study illustrates that the amount of volatile compounds present in the juices of the three citrus species was generally more than that in the peel oils.

References

Baser KHC and Demirci F. 2007. Chemistry of Essential Oils. In: Berger RG (ed). Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability. Berlin: Springer-Verlag. p 43-86.

Boelens MH. 1991. A Critical Review on the Chemical Composition of Citrus Oils. Perfum Flav 16 : 17-34.

Colombo E, Ghizzoni C and Cagni D. 2002. Citrus oils in food and beverages: Uses and analyses. In: Dugo G and Di Giacomo A (eds). Citrus: The genus Citrus. London; New York: Taylor and Francis. p 539-556.

Davies FS and Albrigo LG. 1994. Citrus. Wallingford: CAB International. 254p.

Dharmawan J, Kasapis S, Curran P and Johnson JR. 2007. Characterization of Volatile Compounds in Selected Citrus Fruits from Asia – Part I: Freshly-Squeezed Juice. Flav Fragr J 22 : 228-232.

Högnadóttir A and Rouseff RL. 2003. Identification of Aroma Active Compounds in Orange Essence Oil Using Gas Chromatography-Olfactometry and Gas Chromatography-Masss Spectrometry. J Chromatogr A 998 : 201-211.

Kale PN and Adsule PG. 1995. Citrus. In: Salunkhe DK and Kadam SS (eds). Handbook of Fruit Science and Technology: Production, Composition, Storage and Processing. New York: Marcel Dekker, Inc. p 39-65.

90

Lota M-L, Serra DR, Tomi F and Casanova J. 2000. Chemical Variability of Peel and Leaf Essential Oils of Mandarins from Citrus reticulata Blanco. Biochem Syst Ecol 28 : 61-78.

Morton JF. 1987. Fruits of Warm Climates. Miami: Julia F. Morton. 505p.

Naef R and Velluz A. 2001. Volatile Constituents in Extracts of Mandarin and Tangerine Peel. J Essent Oil Res 13 : 154-157.

Nagy S and Shaw PE. 1990. Factors Affecting the Flavour of Citrus Fruit. In: Morton ID and MacLeod AJ (eds). Food Flavours Part C: The Flavour of Fruits. Amsterdam: Elsevier. p 93-124.

Nugroho SA. 1995. Ekstraksi dan Karakterisasi Minyak Kulit Jeruk Pontianak ( Citrus nobilis var microcarpa), Navel ( Citrus sinensis ) Serta Valencia ( Citrus sinensis ). (Extraction and Characterization of the Peel Oils of Pontianak, Navel and Valencia Oranges). Undergraduate final year thesis, Faculty of Agriculture, Bogor Institute of Agriculture, Indonesia. 109p.

Ruberto G. 2001. Analysis of Volatile Components of Citrus Fruit Essential Oils. In: Jackson JF and Linskens HF (eds). Analysis of Taste and Aroma. New York: Springer. p 123-157.

Shaw PE. 1991. Fruits II. In: Maarse H (ed). Volatile Compounds in Foods and Beverages. New York: Marcel Dekker Inc. p 305-327.

Yang SK, Wang Q, Zhang DR, Huang MB and Zhang DX. 1992. Constituents of Essential Oil Extracted from Seville Orange ( Citrus aurantium ) and Its Processed Product. Zhongcaoyao 23 : 14-15.

91 Chapter 5 Evaluation of Aroma Active Compounds in Pontianak Orange Peel Oil

5.1. Abstract

The aroma active compounds of Pontianak orange peel oil ( Citrus nobilis Lour. var. microcarpa Hassk.) were characterized by using Gas Chromatography/Olfactometry

(GC-O) and Aroma Extract Dilution Analysis (AEDA) techniques. Forty one compounds were found to be aroma active, which were mainly dominated by saturated and unsaturated aldehydes. The flavour dilution (FD) factor was within the range of 2 to 2048, and compounds having the highest FD factor were alpha-pinene, beta-pinene, linalool and 2-methoxy-3-(2-methylpropyl) pyrazine, including few unknown compounds. Based on GC-O results, Odour Activity Value (OAV) and

Relative Flavour Activity (RFA) were determined for aroma model reconstitution.

These resembled the original aroma of the peel oil for the green, fatty, fresh, peely, floral and tarry attributes, with the model solution derived from OAV being the closest to Pontianak oil. Omission tests were carried out to reveal the significance of

(Z)-5-dodecenal and 1-phenylethyl mercaptan as key compounds in the aroma of

Pontianak orange peel oil.

5.2. Introduction

Citrus fruits are widely produced and processed for their fruit juice and the essential oils extracted from the peel (Shaw, 1991; Ohloff, 1994). The latter is utilized as a

92 flavouring in the food industry and in perfume or aromatherapy applications (Baser and Demirci, 2007). The history of citrus can be traced back to more than 4000 years ago and it is believed that the fruit is native to the Southeast Asia from where it spread worldwide (Young, 1986). There are at least 160 cultivars of citrus cultivated throughout Indonesia (Maryoto, 2004), and Pontianak orange ( Citrus nobilis Lour. var. microcarpa Hassk.) is a preferred variety due to its high yield and pleasant organoleptic properties (Sarwono, 1986). The fruit has thin, fairly shiny, yellowish green-coloured skin, and the juice possesses a distinct sweet taste with a slight sulfurous note. Characterization studies of the volatile compounds in Pontianak orange juice and peel oil have been undertaken (Dharmawan et al ., 2006; 2007;

2008), but results did not provide information on the aroma active compounds. A recent publication by Fischer et al . (2008) initiated the quest for identification of aroma active compounds in Pontianak orange.

Gas chromatography/Olfactometry (GC-O) is commonly used for the identification of aroma active compounds (Delahunty et al ., 2006). Using the human nose as detector, a GC-O technique known as Aroma Extract Dilution Analysis (AEDA) can be employed for screening of aroma active compounds (Grosch, 2001). The result of

AEDA is expressed as the flavour dilution (FD) factor, which is the last dilution that the aroma of a volatile compound is detectable. The significant contribution of each odourant to the characteristic flavour can be determined by 2 possible ways, namely the odour activity value (OAV) and the relative flavour activity (RFA). OAV is the ratio of concentration to the odour threshold of the compound, and it is proportional to the FD factor (Schieberle, 1995). It is well accepted that compounds with higher OAV contribute more to the aroma of the food (Guth and Grosch, 1999). Even though the

93 use of this value has been criticized (Frijters, 1978), OAV has been used widely in determining potent odourants in foods (Grosch, 1994; Schieberle, 2000; Buettner and

Schieberle, 2001; Ferreira, 2002; Qian and Wang, 2005). Alternative to OAV, RFA is obtained by the ratio of log FD factor to the square root of weight percentage of the compound. Through the years, RFA has also been used to determine the significant contribution of aroma active compounds in various citrus cultivars (Song et al ., 2000;

Choi et al ., 2001; Choi, 2003; Sawamura et al ., 2004).

In order to verify the significance of aroma active compounds, aroma reconstitution and omission experiments are normally carried out (Grosch, 2001). For aroma reconstitution, flavour compounds are mixed according to the analytical data obtained and their aroma attributes are compared with the original aroma. The importance of potent odourants is subsequently confirmed by omitting those compounds from the aroma models (Choi et al ., 2001). In this case, the aroma active compounds in

Pontianak orange peel oil that contributed to its aroma were investigated stepwise from GC-O sniffing to the calculation of their OAV and RFA values. Results were validated by carrying out aroma reconstitution and omission experiments.

5.3. Materials and Methods

5.3.1. Materials and chemicals

Fresh Pontianak oranges ( Citrus nobilis Lour. var. microcarpa Hassk.) grown in a fruit farm in West Kalimantan of Indonesia were harvested in August 2006. The peel oil was obtained by careful hand-squeezing of the peels of the fruits. The standard chemicals were obtained from the following sources: beta-pinene (ChromaDex,

94 Irvine, CA), R-(+)-limonene, citral (mixture of neral and geranial), nonanal, decanal

(Fluka, Buchs, Switzerland), carveol (Aldrich, St. Louis, MO), 1-phenylethyl mercaptan (Endeavour, Northamptonshire, UK), alpha-pinene, beta-myrcene, linalool, citronellal, (E,Z)-2,6-nonadien-1-ol, (E)-2-nonenal, 1-nonanol, camphor, 2-methoxy-

3-(2-methylpropyl) pyrazine, 4-terpineol, citronellol, nerol, geraniol, L-carvone, (E)-

2-decenal, 1-decanol, perillaldehyde, undecanal, (E,E)-2,4-decadienal, neryl acetate, geranyl acetate, dodecanal, (E,Z)-2,6-dodecadienal and (E)-2-dodecenal were obtained from Firmenich Asia Pte. Ltd. (Singapore). (Z)-5-dodecenal was synthesized at the Department of Chemistry, NUS (Dharmawan et al ., 2009).

5.3.2. Gas Chromatography-Olfactometry (GC-O)

The GC-O instrument comprises of Shimadzu’s GC/MS QP5000 with an olfactometer

ODO II (SGE, Ringwood, Australia) attached to it. The column used was DB-5MS

(5% phenyl/95% methyl polysiloxane – 60m x 0.32mm, 1 µm film thickness; J&W

Scientific, Folsom, CA) while both injector and MS interface temperatures were set at

270 oC. The electron ionization (EI) method was used for the MS at the ionization energy of 70eV with the scan range of 40-300 m/z. The compounds were identified by comparison of mass spectra of the target compounds with those of the NIST (National

Institute of Standards and Technology) library and verified by the retention indices of pure standard compounds. Two µL of Pontianak orange peel oil were injected and the temperature program was set from 120 oC to 240 oC at the rate of 2 oC/min, and increased to 270 oC at 10 oC/min with a 2-min final temperature hold. The flow rate of helium carrier gas was 2.3 mL/min and humid air was constantly added to the effluent at the sniffing port. Four flavourists (two females and two males) from Firmenich

Singapore were the panellists for sniffing the oil. Sniffing of the compounds eluted

95 from the sniffing port was divided into 4 sessions of 15 minutes with a break of 15 minutes in between. The panellists were asked to describe the odour perceived and the detection of an odourous compound at the sniffing port by fewer than three assessors was considered to be noise.

5.3.3. Aroma Extract Dilution Analysis (AEDA), Relative Flavour Activity

(RFA), and Odour Activity Value (OAV)

For aroma extract dilution analysis (AEDA), the peel oil was diluted stepwise 2-fold with diethyl ether by volume to obtain dilutions of 1:2, 1:4, 1:8, 1:16 and so on, with dilutions being injected into the GC-O. The highest dilution in which an aroma active compound was detectable is then referred as the flavour dilution (FD) factor of that compound. On the basis of AEDA results, relative flavour activity (RFA) of each aroma-active compound was obtained by using the equation RFA = log 2 n/S 0.5 (Song et al ., 2000), where 2 n is the FD factor and S is the weight percent of a compound. In order to obtain the Odour Activity Value (OAV), the absolute concentration of the compound was derived from its GC/MS calibration curve while its odour threshold in water was obtained from the literature (Tamura et al ., 1993; 1996; 2001; Arena et al .,

2001; Mihara and Masuda, 1988). The threshold of each compound was determined by using a two-out-of-five sensory evaluation test, whereby each aqueous solution was diluted by a factor of two until the solution was judged to be odorless. The concentration at which the odor of compounds could not be detected by panelists was defined as its odor threshold (Tamura et al ., 1993).

96 5.3.4. Aroma reconstitution and omission test

The synthetic blends of odourants (aroma models) were prepared based on the analytical data of the concentration of aroma active compounds in the original peel oil extract. Three different sets of aroma models were prepared. The first model contained the thirty-three compounds identified by GC/MS. The second and third models were prepared based on the upper range of RFA and OAV obtained from the original set of compounds as shown in Table 5.3 and 5.4, respectively. Sensory evaluation was carried out by 7 trained assessors (2 males and 5 females). They were asked to rate given odour qualities (green, fatty, fresh, peely, floral and tarry) of the original Pontianak orange peel oil and the three aroma models using a 7-point intensity scale ranging from 0.0 to 3.0 at intervals of 0.5. Omission tests were carried out to verify the findings by omitting from the aroma models (Z)-5-dodecenal and 1- phenylethyl mercaptan, which were deemed to be important contributors of Pontianak orange oil aroma. Assessors were asked to rate the degree of similarity between the original Pontianak orange oil and the aroma models with omitted compounds. Score assigned was from 1 being extremely different to 9 being extremely similar, and statistical analysis of the result was performed using t-test.

5.4. Results and Discussion

5.4.1. Aroma active compounds of Pontianak orange peel oil

Forty one aroma active volatiles were detected from the GC-O analysis of Pontianak orange peel oil ( Table 5.1. ). AEDA method was performed to categorize the compounds according to their odour potency. The FD factors of the compounds were detected to fall within the range of 2 to 2048 and Figure 5.1. displayed the

97 chromatogram and aromagram of the aroma active compounds. The chromatogram was obtained from the GC/MS detector while the aromagram was plotted based on

FD factors from AEDA results. Compounds having the highest FD factor (2048) were alpha-pinene, beta-pinene, linalool, 2-methoxy-3-(2-methylpropyl) pyrazine and some unknown compounds that exhibited woody, metallic, green, earthy, sulfury and citrus- like odour quality. Limonene, which is the most abundant compound of Pontianak orange peel oil, exhibited an FD factor of 1024.

The aroma active compounds found in Pontianak orange peel oil were dominated by saturated and unsaturated aldehydes (FD factor from 16 to 512), such as nonanal, decanal, undecanal, dodecanal, (E)-2-nonenal, (E)-2-decenal, (E)-2-dodecenal, (Z)-5- dodecenal, (E,E)-2,4-decadienal and (E,Z)-2,6-dodecadienal. From these compounds,

(Z)-5-dodecenal has rarely been reported in citrus fruits (Chisholm et al ., 2003), and no significance in contributing to the aroma of the oil was assigned. The compound has been found to be one of the major volatiles in insect pheromones (Bestmann et al .,

1993). In addition, 1-phenylethyl mercaptan, which was also detected in the present study with an FD factor of 256, was reported to be part of the composition of the

Asian Pontianak orange peel oil Grab et al ., 2002; Fischer et al ., 2008). Its odor was described as sulfurous and resinous resembling that of the whole fruit (Fischer et al .,

2008). Both (Z)-5-dodecanal and 1-phenylethyl mercaptan were documented presently to have unique characteristics contributing to the flavor of Pontianak orange peel oil.

98 d 4 64 128 256 128 128 128 256 FD 2048 1024 2048 2048 c fruity, green, grassy fruity, Odour quality fresh, green fresh, melon-like, fatty, cucumber-like metallic, citrusy, woody cucumber-like orange-like, orange-like, fruity floral, green soapy, aldehydic soapy, woody, piney, citrusy woody, sulfury, mango-like, metallic sulfury, lemongrass-like tarry, sulfury tarry, 2) in 2) in orange peelPontianak oil ≥ e b . active Aroma (FD compounds Table 5.1 camphor Compound 1-nonanol (E)-2-nonenal beta-pinene (E,Z)-2,6-nonadien-1-ol limonene linalool nonanal alpha-pinene beta-myrcene citronellal 1-phenylethyl mercaptan1-phenylethyl a RI 963 1187 1160 1181 1174 1014 1170 1059 1114 1120 1003 1166 7 3 9 4 5 6 1 2 8 12 11 10 No.

99 d 8 8 2 2 64 512 512 128 128 256 FD 2048 2048 2048 c citrusy, soapy citrusy, earthy, tarry, sulfury earthy, chilli-like, peppery floral, fresh soapy, aldehydic soapy, sweet, floral sweet, citrusy, rose-like citrusy, citrusy, fruity citrusy, earthy sulfury, fruity sulfury, floral, green fruity, citrusy fruity, Odour quality minty e b unknown unknown unknown 2-methoxy-3-(2-methylpropyl) pyrazine 4-terpineol decanal unknown nerol citronellol unknown unknown trans carveol trans neral Compound geraniol L-carvone a RI 1208 1203 1211 1218 1228 1240 1236 1229 1246 1255 1193 1260 1271 15 14 16 17 18 21 20 19 22 23 13 24 25 No.

100 d 2 8 4 2 16 64 64 128 256 256 128 128 FD 2048 c fatty,aldehydic fresh, floral fresh, lemon-like almond-like, floral soapy, soapy, aldehydic fatty,oily earthy floral, floral, green green, citrusy floral citrusy soapy, soapy, citrusy Odour quality soapy, soapy, aldehydic b (E)-2-decenal 1-decanol geranial perillaldehyde undecanal undecanal (E,E)-2,4-decadienal (E,E)-2,4-decadienal unknown unknown unknown unknown neryl acetate neryl unknown geranyl geranyl acetate (Z)-5-dodecenal dodecanal Compound a RI 1276 1279 1284 1309 1319 1335 1344 1353 1365 1379 1393 1405 1421 26 27 28 29 30 31 32 33 34 35 36 37 38 No.

101 d 256 512 128 FD erceived at the sniffing d c Odour quality Odour fruity, soapy fruity, fatty, citrusy fatty, oily, creamy oily, n ion timeion andwith thespectra mass reference standar compound wascompound identifiedon based odour the quality p mpound b Compound (E,Z)-2,6-dodecadienal (E)-2-dodecenal unknown unknown a RI 1465 1482 1490 39 40 41 Experimental retention linear index colum on DB5-MS identified was compound by The retentcomparing its (FD)odour-activeDilution factor theof Flavour co Odour quality perceivedOdour sniffingthrough port the interpretation. too was for MS signal The weak The No. a b c d e port and the retention port and the detection time itsof odour

102 ctive incompounds Pontianak orange oil peel Chromatogram (top) and aromagram a of (below) aroma (top) aromagram and Chromatogram Figure5.1.

103 5.4.2. Odour Activity Value (OAV) and Relative Flavour Activity (RFA)

In order to determine the relative contribution of each compound to the aroma of

Pontianak orange peel oil, OAV and RFA have been employed ( Table 5.2 ). OAV was obtained by taking into account the concentration and odour threshold of each compound while RFA utilized the FD factor and weight percentage of the compound.

Table 5.3 displayed 18 compounds that have the highest OAV in descending order.

Due to unavailability of odour threshold data in the literature or the below-detection peak intensity, the OAV of 1-phenylethyl mercaptan, 2-methoxy-3-(2-methylpropyl) pyrazine, (Z)-5-dodecenal and (E,Z)-2,6-dodecadienal was not determined. It showed that limonene has the highest OAV followed by (E)-2-nonenal, linalool, (E)-2- dodecenal, (E,Z)-2,6-nonadien-1-ol and myrcene. Compounds like camphor, 4- terpineol, trans-carveol and neryl acetate were among those with the lowest OAV

(below 100). In general, compounds that had high FD factor also had high OAV, which confirms the positive relationship between FD factor and OAV (Grosch, 1994).

Since OAV often depends on concentration and may not always reveal the characteristic odourants, the concept of relative flavour activity (RFA) could be used to identify potent aroma-active compounds (Choi, 2003). RFA is calculated by using

FD factors instead of the odour threshold values, and the concept of RFA was created to compensate for the inability of OAV to be invariably correct in judging the important aroma active compounds. To demonstrate this, a compound like limonene may have high OAV but a relatively low RFA value and it is not the most important contributor to aroma of citrus fruits. Table 5.4 reproduces 17 compounds with the highest RFA. This was not calculated for 1-phenylethyl mercaptan and 2-methoxy-3-

(2-methylpropyl) pyrazine because their concentrations could not be determined.

104 e — 5.5 5.4 1.4 7.0 5.3 5.0 0.3 5.1 26.9 84.2 28.0 RFA d — 0.008 0.012 0.001 0.376 2.149 0.089 0.386 0.004 0.445 0.174 95.694 % Weight tianak orange tianakpeel orange oil. c 1 — 109 2290 29310 OAV 17770 18560 38010 15870 182570 145240 4372500 b ) ) 1 ) ) ( ) ) (3 1 2 g ( ( 1 1 ( ( 1.0 4.6 0.1 1.5 1.5 0.046 0.028 0.2 0.2 N/A N/A 0.67 0.67 0.19 0.19 0.001 0.0004 0.0004 in Water in Water (ppm) Odour Threshold Odour Threshold a f our Activityour active aroma of (RFA) compounds Ponin nd 5.72 38.01 73.03 108.58 817.47 3525.84 3439.92 1586.98 4066.78 19640.32 874495.60 Conc. (ppm) Conc. 1-nonanol camphor citronellal limonene alpha-pinene beta-pinene (E,Z)-2,6-nonadien-1-ol (E)-2-nonenal beta-myrcene nonanal linalool Compound mercaptan 1-phenylethyl . The Odour .ActivityOdour The Relative Values and (OAV) Flav RI 963 1181 1187 1166 1059 1014 1170 1174 1003 1120 1114 1160 Table 5.2 Table 8 4 3 1 9 2 6 5 7 11 12 10 No. 105 e — 6.3 2.8 8.1 3.0 2.5 3.2 6.6 13.8 27.0 22.1 10.0 11.8 RFA d — 0.004 0.183 0.011 0.050 0.089 0.015 0.006 0.012 0.015 0.009 0.019 0.023 % Weight c 6 — 26 102 7400 1200 1330 5570 1620 7860 1710 3460 23890 OAV b ) 4 ( 4 6.4 0.1 0.1 0.07 0.68 0.01 0.062 0.067 0.017 0.775 0.062 0.000045 0.000045 in Water in Water (ppm) Odour Threshold Odour Threshold a f nd 39.02 55.74 78.74 458.54 132.54 815.32 104.42 108.31 133.65 170.63 214.38 1672.48 Conc. (ppm) Conc. 4-terpineol decanal citronellol neral nerol trans carveol trans geraniol L-carvone (E)-2-decenal 1-decanol geranial perillaldehyde Compound 2-methoxy-3-(2-methylpropyl) pyrazine RI 1211 1218 1236 1255 1240 1246 1260 1271 1276 1279 1284 1309 1193 14 15 16 19 17 18 20 21 22 23 24 25 13 No. 106 e 3.0 7.9 5.7 8.0 18.3 16.5 19.2 25.3 RFA d 0.022 0.021 0.010 0.071 0.011 0.012 0.009 0.070 % Weight c — 51 — the the bracket: 623 8300 ompound eported eported thresholds in water OAV 11750 15900 ubstances 143000 f all f all compounds b g g 2 0.15 0.01 0.04 0.055 N/A N/A N/A N/A 0.0014 in Water Water (ppm) in Odour Threshold Odour a . (2001);Mihara . (4) and Masuda (1998) 93.49 82.97 200.08 109.69 195.30 646.46 101.17 635.99 et al et Conc. (ppm) Conc. ., 1993), stated theunless number1993), ., by otherwise in is the FD factor and percent a the FD Sis ofc the weight is ding theding concentrations by the odourants of r their centrationinrelative to the total concentration o n ed by plotting the by ed standard curve theof references et al

, where 2 , where 0.5 . (1996);Arena . (3) /S n et al et .(2) (2001);Tamura (E)-2-dodecenal (Z)-5-dodecenal (E,Z)-2,6-dodecadienal Compound geranyl geranyl acetate dodecanal neryl acetate (E,E)-2,4-decadienal undecanal undecanal et al RI 1482 1405 1465 1393 1421 1365 1335 1319 33 30 32 29 31 28 27 26 Weight percentageWeight basedeachcompound of its on con available not Data The Odour Activity divi obtained Odour was by The Value (OAV) logFlavour Relative 2 Activity = (RFA) (1) Tamura (1) The concentrationThe aroma-active of obtain compounds, in reported the literatureOdour thresholds (Tamura nd, notnd, determined No. a b c d e f g 107 Table 5.3 . Potent odourants in Pontianak orange peel oil based on their Odour Activity Values (OAV>2000)

No. Compound OAV a

1 limonene 4372500

2 (E)-2-nonenal 182570

3 linalool 145240

4 (E)-2-dodecenal 143000

5 (E,Z)-2,6-nonadien-1-ol 38010

6 beta-myrcene 29310

7 decanal 23890

8 alpha-pinene 18560

9 citronellal 17770

10 undecanal 15900

11 nonanal 15870

12 dodecanal 11750

13 (E,E)-2,4-decadienal 8300

14 (E)-2-decenal 7860

15 citronellol 7400

16 geraniol 5570

17 perillaldehyde 3460

18 beta-pinene 2290 a The Odour Activity Value (OAV) was obtained by dividing the concentrations of the odourants by their reported thresholds in water

108 Table 5.4 . Potent odourants in Pontianak orange peel oil based on their Relative Flavour Activity (RFA>6.5)

No. Compound RFA a

1 camphor 84.2

2 (E,Z)-2,6-nonadien-1-ol 28.0

3 geraniol 27.0

4 (E)-2-nonenal 26.9

5 (E,E)-2,4-decadienal 25.3

6 L-carvone 22.1

7 (Z)-5-dodecenal 19.2

8 (E)-2-dodecenal 18.3

9 (E,Z)-2,6-dodecadienal 16.5

10 4-terpineol 13.8

11 perillaldehyde 11.8

12 (E)-2-decenal 10.0

13 citronellol 8.1

14 undecanal 8.0

15 dodecanal 7.9

16 citronellal 7.0

17 geranial 6.6 a Relative Flavour Activity (RFA) = log 2 n/S 0.5 , where 2 n is the FD factor and S is the weight percent of a compound.

Results indicated that camphor is the highest RFA compound followed by (E,Z)-2,6- nonadien-1-ol, geraniol, (E)-2-nonenal, (E,E)-2,4-decadienal and L-carvone.

Compounds having the lowest RFA were limonene, myrcene and neral. In general,

109 compounds that have high OAV are low in their RFA and vice versa, with limonene obeying this trend and camphor exhibiting low OAV/high RFA. Notable exceptions were (E)-2-nonenal and (E,Z)-2,6-nonadien-1-ol with high OAV and RFA owing to very low concentrations in peel oil and similarly low odour thresholds that relate to relatively high FD factors.

5.4.3. Aroma reconstitution

In order to verify the contribution of aroma-active compounds to the flavour profile of

Pontianak orange peel oil, synthetic blends were made based on the aforementioned findings. Three formulas were prepared according to the results shown in Tables 5.2 ,

5.3 and 5.4 , and were described in more details in Appendices A.4 to A.6.

Furthermore, 1-phenylethyl mercaptan, 2-methoxy-3-(2-methylpropyl) pyrazine, (Z)-

5-dodecenal and (E,Z)-2,6-dodecadienal were included in each aroma model, since these are relatively rare compounds in citrus but found in Pontianak oil hence may contribute considerably to its overall flavour.

As the concentrations of 1-phenylethyl mercaptan and 2-methoxy-3-(2-methylpropyl) pyrazine could not be determined due to a weak MS signal, trials were carried out to determine optimum reconstitution levels. Results indicated that additions of 0.001% w/w from each compound produced a reconstituted aroma blend that matched best that of the natural material. Thus, 3 model solutions were prepared as follows: Model solution 1 included all compounds in Table 5.2 (33 compounds), model solution 2 included all compounds in Table 5.3 (18 compounds plus 1-phenylethyl mercaptan,

2-methoxy-3-(2-methylpropyl) pyrazine, (Z)-5-dodecenal and (E,Z )-2,6-dodecadienal that the OAV could not be obtained ), and model solution 3 included all compounds in

110 Table 5.4 (17 compounds plus 1-phenylethyl mercaptan and 2-methoxy-3-(2- methylpropyl) pyrazine that the RFA could not be obtained, and limonene).

Six sensory properties, namely green, fatty, fresh, peely, floral and tarry were selected to be the major characteristics of Pontianak orange oil following sensory trials and consensus among the flavourists. The aroma of the Pontianak peel oil was then compared with the sensory characteristics of the aroma models. Figure 5.2 showed that the intensities of floral, peely and fresh were rated slightly higher in the peel oil than the models, whereas the green attribute of the oil was rated slightly lower than the models. Details of the results were described in Appendices A.7 to A.10.

Green 2.5

2.0

1.5 Tarry Fatty 1.0 Pontianak Oil 0.5 Formula1 0.0 Formula2

Floral Fresh Formula3

Peely

Figure 5.2. Comparative flavour profile analysis of Pontianak orange peel oil and the reconstituted aroma model solutions based on all available compounds (Formula 1), Relative Flavour Activity (RFA; Formula 2) and Odour Activity Value (OAV; Formula 3).

111 Overall for the examined attributes, the aroma of all models was found to be comparable to the original Pontianak oil, as there was no significant statistical difference (p < 0.5). Nevertheless, the panellists felt that the OAV based aroma model was closer to the aroma of Pontianak orange. This is in agreement with previous findings that OAV shortlists effectively potent aroma compounds while RFA may not always correlate directly to the characteristic aroma compounds in food (Choi, 2001;

Grosch, 2001).

5.4.4. Omission experiments

Results of the various aspects of this work (GC-O, aroma reconstitution and sensory trials) on all available compounds indicated that (Z)-5-dodecenal and 1-phenylethyl mercaptan played a major role in the aroma of Pontianak orange peel oil, and work in this section was carried out to confirm this. Individual and binary omissions of those two compounds from the OAV based model system were prepared and evaluated orthonasally by the panellists. There was no significant difference (p < 0.05) in the average score of ratings in model solutions when either compound was omitted and compared to the complete aroma model ( Table 5.5 ). However, the binary omission resulted in significant difference (p < 0.05) in relation to the complete model mixture, an outcome which suggests the importance of both (Z)-5-dodecenal and 1-phenylethyl mercaptan to the aroma of Pontianak peel oil. Details of the results can be viewed in

Appendix A.11.

112 Table 5.5 . Sensory evaluation for the aroma model of the Pontianak orange peel oil as affected by the omission of compounds

No. Compound(s) omitted Average score*

1. None (Pontianak orange aroma model) 6.8 a 2. (Z)-5-dodecenal 5.4 a 3. 1-phenylethyl mercaptan 5.4 a 4. (Z)-5-dodecenal and 1-phenylethyl mercaptan 4.3 b

* The average score of 7 panellists with a scale of 1 (extremely different from) to 9 (extremely similar to) Pontianak orange peel oil. The difference between levels with same letter is not significant (p < 0.05)

From this work, it was concluded that extensive experimentation combining instrumental GC-O with sensory evaluation of reconstituted aroma formulations and omission tests are essential for the evaluation of the aroma active compounds of

Pontianak orange peel oil. Aroma extract dilution analysis was utilized to obtain the flavour dilution factor, which was instrumental in describing the OAV and RFA characteristics of the peel-oil compounds. The approach was successful in the identification of potent odourants and shortlisted (Z)-5-dodecenal and 1-phenylethyl mercaptan as essential contributors to the aroma of Pontianak orange peel oil.

References

Arena E, Campisi S, Fallico B, Lanza MC and Maccarone E. 2001. Aroma Value of Volatile Compounds of Prickly Pear ( Opuntia ficus indica (L.) mill. Cactaceae). Ital J Food Sci 13 : 311-319.

113 Baser KHC and Demirci F. 2007. Chemistry of Essential Oils. In: Berger RG (ed). Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability. Berlin: Springer-Verlag. p 43-86.

Bestmann HJ, Attygalle AB, Garbe W, Kern F, Martichonok V, Schafer D, Vostrowsky O and Hasenfuß I. 1993. Chemical structure and final steps of biosynthesis of the female sex pheromone of Gastropacha quercifolia (Lepidoptera: Lasiocampidae). Insect Biochem Molec Biol 23 : 793-799.

Buettner A and Schieberle P. 2001. Evaluation of Aroma Differences between Hand- Squeezed Juices from Valencia Late and Navel Oranges by Quantitation of Key Odourants and Flavour Reconstitution Experiments. J Agric Food Chem 49 : 2387- 2394.

Chisholm MG, Jell JA and Cass Jr DM. 2003. Characterization of the Major Odourants Found in the Peel Oil of Citrus reticulata Blanco cv. Clementine Using Gas Chromatography-Olfactometry. Flavour Fragr J 18 : 275-281.

Choi HS. 2003. Character Impact Odourants of Citrus Hallabong [( C. unshiu Marcov x C. sinensis Osbeck) x C. reticulata Blanco] Cold-Pressed Peel Oil. J Agric Food Chem 51 : 2687-2692.

Choi HS, Kondo Y and Sawamura M. 2001. Characterization of the Odour-Active Volatiles in Citrus Hyuganatsu ( Citrus tamurana Hort. Ex Tanaka). J Agric Food Chem 49 : 2404-2408.

Delahunty CM, Eyres G and Dufour JP. 2006. Gas Chromatography – Olfactometry. J Sep Sci 29 : 2107-2125.

Dharmawan J, Barlow PJ and Curran P. 2006. Characterization of Volatile Compounds in Selected Citrus Fruits from Asia. In: Bredie WLP and Petersen MA (eds). Flavour Science: Recent Advances and Trends. Amsterdam: Elsevier. p 319- 322.

114 Dharmawan J, Kasapis S, Curran P and Johnson JR. 2007 Characterization of Volatile Compounds in Selected Citrus Fruits from Asia Part I: Freshly-Squeezed Juice. Flavour Fragr J 22 : 228-232.

Dharmawan J, Kasapis S and Curran P. 2008. Characterization of Volatile Compounds in Selected Citrus Fruits from Asia Part II: Peel Oil. J Essent Oil Res 20 : 21-24.

Dharmawan J, Kasapis S, Sriramula P, Lear MJ and Curran P. 2009. Evaluation of Aroma Active Compounds in Pontianak Orange Peel Oil ( Citrus nobilis Lour. var. microcarpa Hassk.) by Gas Chromatography/Olfactometry, Aroma Reconstitution and Omission Test. Submitted for publication to J Agric Food Chem (in press – online access: DOI 10.1021/jf801070r).

Ferreira V, Ortín N, Escudero A, López R and Cacho J. 2002. Chemical Characterization of the Aroma of Grenache Rosé Wines: Aroma Extract Dilution Analysis, Quantitative Determination, and Sensory Reconstitution Studies. J Agric Food Chem 50 : 4048-4054.

Fischer A, Grab W and Schieberle P. 2008. Characterization of The Most Odour- Active Compounds in a Peel Oil Extract from Pontianak Oranges ( Citrus nobilis var. Lour. microcarpa Hassk.). Eur Food Res Technol 227 : 735-744.

Frijters JER. 1978. A Critical Analysis of the Odour Unit Number and Its Use. Chem Senses Flavour 3: 227-233.

Grab W, Ratcliff DJ and Furrer S. 2001. Flavor and Fragrance Compositions. EP Patent 1,264,547A1.

Grosch W. 1994. Determination of Potent Odourants in Foods by Aroma Extract Dilution Analysis (AEDA) and Calculation of Odour Activity Values (OAV). Flavour Fragr J 9: 147-158.

115 Grosch W. 2001. Evaluation of the Key Odourants of Foods by Dilution Experiments, Aroma Models and Omission. Chem Senses 26 : 533-545.

Guth H and Grosch W. 1999. Evaluation of Important Odourants in Foods by Dilution Techniques. In: Teranishi R, Wick EL and Hornstein I (eds). Flavour Chemistry: Thirty Years of Progress. New York: Kluwer Academic/Plenum. p 377-386.

Maryoto A. 2004. Kita Memiliki Ratusan Varietas Jeruk (We Have Hundreds of Citrus Varieties). Kompas Daily , October 26, 2004. http://air.bappenas.go.id/doc/pdf/kliping/Kita%20Memiliki%20Ratusan%20Varietas%20Jeru k.pdf (last accessed 30 June 2008)

Mihara S and Masuda H. 1988. Structure-Odour Relationships for Disubstituted Pyrazines. J Agric Food Chem 36 : 1242-1247.

Ohloff G. 1994. Scent and Fragrances: The Fascination of Odours and Their Chemical Perspectives. Berlin; New York: Springer-Verlag. p 127-198.

Qian MC and Wang YY. 2005. Seasonal Variation of Volatile Composition and Odour Activity Value of ‘Marion’ ( Rubus spp. hyb ) and ‘Thornless Evergreen’ ( R. laciniatus L.) Blackberries. J Food Sci 70 : C13-C20.

Sarwono B. 1986. Jeruk dan Kerabatnya (Citrus and Its Family). Jakarta: Penebar Swadaya. p 1-2.

Sawamura M, Nguyen TMT, Onishi Y, Ogawa E and Choi HS. 2004. Characteristic Odour Components of Citrus reticulata Blanco (Ponkan) Cold-Pressed Oil. Biosci Biotechnol Biochem 68 : 1690-1697.

Schieberle P. 1995. New Developments in Methods for Analysis of Volatile Flavour Compounds and Their Precursors. In: Gaonkar AG (ed). Characterization of Food: Emerging Methods. Amsterdam; New York: Elsevier. p 403-431.

116 Schieberle P, Hofmann T and Münch P. 2000. Studies on Potent Aroma Compounds Generated in Maillard-Type Reactions Using the Odour-Activity-Value Concept. In: Risch SJ and Ho CT (eds). Flavour Chemistry: Industrial and Academic Research. Washington DC: American Chemical Society. p 133-150.

Shaw PE. 1991. Fruits II. In: Maarse H (ed). Volatile Compounds in Foods and Beverages. New York: Marcel Dekker Inc. p 305-327.

Song HS, Sawamura M, Ito T, Kawashimo K and Ukeda H. 2000. Quantitative Determination and Characteristic Flavour of Citrus junos (Yuzu) Peel Oil. Flavour Fragr J 15 : 245-250.

Tamura H, Yang RH and Sugisawa H. 1993. Aroma Profiles of Peel Oils of Acid Citrus. In: Teranishi R and Sugisawa H (eds). Bioactive Volatile Compounds from Plants. Washington DC: American Chemical Society. p 121-136.

Tamura H, Boonbumrung S, Yoshizawa T and Varanyanond W. 2001. The Volatile Constituents in the Peel and Pulp of Green Thai Mango, Khieo Sawoei Cultivar (Mangifera indica L.). Food Sci Technol Res 7: 72-77.

Tamura H, Fukuda Y and Padrayuttawat A. 1996. Characterization of Citrus Aroma Quality by Odour Threshold Values. In: Takeoka GR, Teranishi R, Williams PJ and Kobayashi A (eds). Biotechnology for Improved Foods and Flavours. Washington DC: American Chemical Society. p 282-294.

Young RH. 1986. Fresh Fruit Cultivars. In: Wardowski WF, Nagy S and Grierson W (eds). Fresh Citrus Fruits. Westport: The AVI Publishing Company Inc. p 102-126.

117 Chapter 6 Conclusion

Characterization of volatile compounds in selected citrus cultivars from Asia, i.e.

Pontianak orange, Mosambi and Dalandan revealed their unique profiles with the presence of volatile compounds well identified in citrus fruits as well as those not previously reported in other citrus varieties. In general, Mosambi showed its resemblance to typical sweet orange while Dalandan to mandarin. On the other hand,

Pontianak orange had distinct characteristics as some important contributor compounds to mandarin were not found. There was a discrepancy of profiles of flavour compounds between the freshly-squeezed juices and hand-pressed peel oils for similar citrus cultivars, where the juices were generally richer in ester derivatives than the oil.

As Pontianak orange possess outstanding flavour profile with slight sulfurous note, further investigation was carried out to identify its contributors by combining GC-O analysis with sensory evaluation of reconstituted aroma formulations and omission tests. This stepwise approach was effective in revealing 41 potent odourants and shortlisting both (Z)-5-dodecenal and 1-phenylethyl mercaptan as key contributors to

Pontianak orange flavour. In conclusion, results of this research demonstrated the feasibility of the approach taken in unveilling the aroma active compounds and key odourants in Pontianak orange peel oil. Results of this study also lead to a better understanding of volatile compounds present in citrus cultivars of Asia, particularly

118 Pontianak orange, Mosambi and Dalandan. This knowledge could be utilized subsequently by food ingredients industries for various applications and innovation, specifically related to flavour compounds.

119 Chapter 7 Suggestion for Future Work

It has been recognized that some compounds found in nature have enantiomeric properties, or also known as chirality. Chiral compounds are isomers that are mirror image to each other and cannot be superimposed. Enantiomeric compounds found in food naturally may not be present in equal proportion and they may have different organoleptic properties. For example, R-(+)-limonene has fresh citrus, orange-like odour while S-(-)-limonene has harsh, turpentine-like and lemon-like aroma. Thus, it is important to find out the actual distribution of compounds with chiral properties in

Pontianak orange so that it can provide more accurate information of the volatile compounds present and create a database to detect its authenticity, quality and geographical origin for future reference. The enantiomeric ratio of volatile compounds can be determined by using high resolution gas chromatograph (HRGC) and the most widely used chiral stationary phase in HRGC is modified cyclodextrins

(CD), either pure or diluted in polysiloxanes.

Furthermore, current research focused on 3 selected citrus cultivars amid numerous varieties in Asia that have superb and unique flavour characteristics. While their flavour profiles are not widely investigated, characterization of their volatile compounds is therefore necessary for screening the outstanding cultivars whose aroma profiles can be investigated further. A continual discovery of novel compounds

120 will add value to existing collection of flavour compounds that can be applied in various food products.

As different types of food products have distinct physical and chemical properties, the application of flavour volatiles is not the same for a variety of food systems. It is therefore crucial to look into the application of flavour compounds formulated in various types of food products, such as ice cream, beverages and confectionery in which citrus flavour is commonly applied.

121 Appendix

A.1. Chemical composition of Pontianak orange Trial No. pH oBrix % Acidity 1 4.86 12.4 0.26 2 4.65 14.3 0.22 3 4.8 14.25 0.23 4 4.83 13.4 0.23 5 4.8 12.9 0.25 6 4.77 13.4 0.27 7 4.85 12.9 0.19 8 4.85 14.1 0.18 9 4.8 14.3 0.22 10 4.85 13.4 0.21 11 4.71 12.9 0.23 12 4.79 13.6 0.31 13 4.71 13.4 0.28 14 4.78 14.4 0.3 15 4.67 14 0.2 Average 4.78 13.58 0.24 Std Dev 0.07 0.63 0.04

A.2. Chemical composition of Mosambi Trial No. pH oBrix % Acidity 1 3.4 10.1 1.65 2 3.55 9.7 1.13 3 3.4 10.2 1.21 4 3.43 9.6 1.44 5 3.67 10.3 1.26 6 3.62 11 0.95 7 3.47 10.05 1.11

122 Trial No. pH oBrix % Acidity 8 3.53 9.8 0.73 9 3.58 9.5 1.44 10 3.42 10.2 1.18 11 3.57 10.8 1.12 12 3.56 10.45 1.54 13 3.7 10.2 1.01 14 3.63 9.8 1.07 15 3.78 10.6 0.83 Average 3.55 10.15 1.18 Std Dev 0.12 0.44 0.26

A.3. Chemical composition of Dalandan Trial No. pH oBrix % Acidity 1 4.53 9.8 0.17 2 4.58 10.2 0.22 3 4.6 10.4 0.2 4 4.72 10.5 0.23 5 4.6 10.1 0.16 6 4.45 9.7 0.2 7 4.56 9.6 0.19 8 4.68 10.6 0.18 9 4.57 10.4 0.21 10 4.6 9.7 0.2 11 4.58 9.3 0.2 12 4.64 10.4 0.17 13 4.77 10.4 0.15 14 4.57 9.3 0.22 15 4.59 10.3 0.2 Average 4.60 10.05 0.19 Std Dev 0.08 0.44 0.02

123 A.4. Aroma reconstitution: Formula 1 (ALL) No. Compound % Concentration 1 alpha-pinene 0.386 2 beta-myrcene 2.149 3 beta-pinene 0.376 4 limonene 95.692 5 linalool 0.445 6 nonanal 0.174 7 1-phenylethyl mercaptan 0.001 8 citronellal 0.089 9 (E,Z)-2,6-nonadien-1-ol 0.004 10 (E)-2-nonenal 0.008 11 1-nonanol 0.012 12 camphor 0.001 13 2-methoxy-3-(2-methylpropyl) pyrazine 0.001 14 4-terpineol 0.004 15 decanal 0.183 16 citronellol 0.050 17 nerol 0.089 18 trans carveol 0.011 19 neral 0.015 20 geraniol 0.006 21 L-carvone 0.012 22 (E)-2-decenal 0.015 23 1-decanol 0.009 24 geranial 0.019 25 perillaldehyde 0.023 26 undecanal 0.070 27 (E,E)-2,4-decadienal 0.009 28 neryl acetate 0.011 29 geranyl acetate 0.010 30 (Z)-5-dodecenal 0.012 31 dodecanal 0.071

124 No. Compound % Concentration 32 (E,Z)-2,6-dodecadienal 0.021 33 (E)-2-dodecenal 0.022

A.5. Aroma reconstitution: Formula 2 (RFA) No. Compound % Concentration 1 (E)-2-nonenal 0.008 2 camphor 0.001 3 geraniol 0.006 4 (E,Z)-2,6-nonadien-1-ol 0.004 5 (E,E)-2,4-decadienal 0.009 6 (E,Z)-2,6-dodecadienal 0.022 7 L-carvone 0.012 8 (Z)-5-dodecenal 0.012 9 4-terpineol 0.004 10 (E)-2-dodecenal 0.023 11 undecanal 0.072 12 perillaldehyde 0.024 13 dodecanal 0.073 14 (E)-2-decenal 0.015 15 citronellal 0.093 16 citronellol 0.052 17 geranial 0.019 18 linalool 0.461 19 limonene 99.088 20 1-phenylethyl mercaptan 0.001 21 2-methoxy-3-(2-methylpropyl) pyrazine 0.001

125

A.6. Aroma reconstitution: Formula 3 (OAV) No. Compound % Concentration 1 limonene 95.878 2 (E)-2-nonenal 0.008 3 linalool 0.446 4 (E)-2-dodecenal 0.022 5 (E,Z)-2,6-nonadien-1-ol 0.004 6 beta-myrcene 2.153 7 decanal 0.183 8 alpha-pinene 0.387 9 citronellal 0.090 10 undecanal 0.070 11 nonanal 0.174 12 dodecanal 0.071 13 (E,E)-2,4-decadienal 0.009 14 (E)-2-decenal 0.015 15 citronellol 0.050 16 geraniol 0.006 17 perillaldehyde 0.024 18 beta-pinene 0.377 19 1-phenylethyl mercaptan 0.001 20 2-methoxy-3-(2-methylpropyl) pyrazine 0.001 21 (Z)-5-dodecenal 0.012 22 (E,Z)-2,6-dodecadienal 0.021

126 A.7. Sensory evaluation of Pontianak orange oil Attribute Panel1 Panel2 Panel3 Panel4 Panel5 Panel6 Panel7 Average Green 1 1 0.5 2 1.5 1 2 1.29 Fatty 2 2 1 2.5 3 2 1.5 2.00 Fresh 2 2 0.5 1.5 2 2 2 1.71 Peely 1.5 2 2 2.5 2 2.5 2.5 2.14 Floral 2 2 2 1 0.5 1 0.5 1.29 Tarry 1 1 2 1 2 2 3 1.71

A.8. Sensory evaluation of reconstituted Pontianak orange oil: Formula 1 Attribute Panel1 Panel2 Panel3 Panel4 Panel5 Panel6 Panel7 Average Green 1.5 1.5 2 2.5 2.5 2.5 2.5 2.14 Fatty 1 1.5 0.5 1.5 2.5 2 2.5 1.64 Fresh 1 2 1 2.5 2.5 2.5 2.5 2.00 Peely 1.5 0.5 0.5 1 1 1.5 0.5 0.93 Floral 2.5 2 1.5 2 2 2 2 2.00 Tarry 1.5 1.5 2 2.5 2.5 2.5 2.5 2.14

A.9. Sensory evaluation of reconstituted Pontianak orange oil: Formula 2 Attribute Panel1 Panel2 Panel3 Panel4 Panel5 Panel6 Panel7 Average Green 1.5 2 2 2 2 2 2.5 2.00 Fatty 1 1.5 0.5 2 2 2 1.5 1.50 Fresh 2 2 0.5 1.5 1.5 1.5 2 1.57 Peely 2 1.5 1 2 1.5 1.5 1.5 1.57 Floral 2 0.5 0.5 1 1.5 1 0.5 1.00 Tarry 2.5 1 1 1 1.5 1.5 0.5 1.29

127 A.10. Sensory evaluation of reconstituted Pontianak orange oil: Formula 3 Attribute Panel1 Panel2 Panel3 Panel4 Panel5 Panel6 Panel7 Average Green 1.5 1.5 2.5 2.5 0.5 0.5 2 1.57 Fatty 1 1.5 2 3 1.5 1.5 2 1.79 Fresh 1.5 2 0.5 1 1.5 1.5 1.5 1.36 Peely 2 2 2 3 1 2 2 2.00 Floral 1.5 1.5 0.5 1 0.5 0.5 1 0.93 Tarry 2 1.5 3 1 1 1 1 1.50

A.11. Omission test Compound (s) omitted P1 P2 P3 P4 P5 P6 P7 Avg

None (Pontianak orange 5 2 5 4 5 4 5 4.3 aroma model)

(Z)-5-dodecenal 6 6 5.5 3 6 7 4 5.4

1-phenylethyl 5 5 5.5 7 4 5 6 5.4 mercaptan

(Z)-5-dodecenal and 1- 7 7 6.5 8 7 8 4 6.8 phenylethyl mercaptan *P = Panel; Avg = Average

128