Aroma Profiles of Peel Oils of Acid Citrus - American Chemical Society

Chapter 10

Aroma Profiles of Peel Oils of Acid Citrus H. Tamura, R.-H. Yang, and H. Sugisawa

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on April 1, 2016 | http://pubs.acs.org Publication Date: April 6, 1993 | doi: 10.1021/bk-1993-0525.ch010

Department of Bioresource Science, Kagawa University, Miki-cho, Kagawa, 761-07 Japan

The aroma quality of six acid citrus fruits was characterized by instrumental and objective methods such as the GC-sniff test with flavor wheel and logarithmic values of odor units. With the GC-sniff test, carvone, carveol, perillaldeyde, linalool and citronellal were found to constitute the characteristic aroma of sudachi oil. Logarithmic odor unit values of each component of lemon, lime, sudachi, kabusu, yuzu, and kabosu were found to be effective for more objective evaluation of aroma quality. Monoterpene aldehydes, alcohols and their esters, for example, geranial, geraniol, and neral contributed significantly to the characteristic aroma of lemon and lime. Citronellal, citronellol, and carvone contributed to sudachi; and citronellol, 2-decenal, geranyl acetate to kabosu. Geraniol, linalyl acetate, geranyl acetate and linalool contributed to kabusu. Unsaturated aliphatic aldehydes especially contributed to yuzu. Multivariate analysis was applied to aroma evaluation of many samples.

There are several hundred species in the Citrus family. The sour and sweet taste offers a refreshing, pleasant and special sensation. Lemon and lime are representative citrus fruits all over the world and have a particular sour taste. In this article, the name "acid citrus" refers to a group of citrus fruits having sour juice. It is defined by the sour taste of these citrus fruits. Citrus flavor plays an important role in the promotion of our appetite, and acid citruses have been used as substitutes for vinegar in many Japanese foods. Yuzu, kabosu, sudachi and kabusu, as well as lemon and lime, are very common as acid citrus in Japan. These citrus fruits are present in many Oriental flavors. To ascertain the important compounds in the fruit aroma, identification of the volatile components, quantitative analysis and sensory evaluation of citrus oils have been performed. As new analytical techniques have been developed, trace constituents of the fruits were 0097-6156/93A)525-0121$06.00/0 © 1993 American Chemical Society

Teranishi et al.; Bioactive Volatile Compounds from Plants ACS Symposium Series; American Chemical Society: Washington, DC, 1993.


122

BIOACTIVE VOLATILE COMPOUNDS F R O M PLANTS

identified. However, identification and quantitative analysis of all constituents in the fruits are not always essential for evaluation of flavor quality nor for the reproduction of the same flavor. A character impact compound, some character components or some composition of volatile compounds producing the unique aroma are significant. Recently a computer-aided analytical system for identification using retention index and similarity index of Mass spectral pattern has been helpful and indispensable. To determine the significant compounds in each acid citrus fruit, the following two methods were examined and compared: 1. GC-sniff test using a flavor wheel, in which intensity and quality of each flavor component can be expressed easier by a simple word, and 2. multivariate statistical analysis using logarithmic values of odor units of each component identified. The following citrus fruits grown on the island of Shikoku in Japan were used for this experiment: lemon (Gtrus limone), lime {Citrus aurantifolia), sudachi (Citrus sudachi Hort. ex Shirai), kabosu (Gtrus sphaerocarpa Tanaka), kabusu (Citrus auraruium f. kabusu), and yuzu (Citrus junos Tanaka). Isolation and identification of volatile compounds in citrus fruits Isolation of volatile compounds in 6 acid citrus fruits is the most critical step in the evaluation of aroma quality. Cold pressed oils from Valencia orange (J), grapefruit (2) and lime oil (3) were reported to have a quality very much like that obtained under mild conditions. However, the oils also included considerably higher amounts of non-volatile components. It is well-known that the characteristic aroma components in yuzu (Gtrus juno) are thermally very unstable. The aroma qualities of yuzu oil extracted by three isolation methods - simultaneous distillation extraction (SDE) at atmospheric pressure, steam distillation under reduced pressure, and solvent extraction - were compared with those of fresh peels by paired comparison test. A panel consisting of six members carried out a series of sensory evaluations by duo-trio and paired comparison tests. This panel judged that solvent extraction produced material having an aroma similar to that of fresh yuzu peels. Volatiles from the other 4 acid citrus peels (400g) with pentane and dichloromethane (7/3, v/v) were also extracted by solvent extraction and SDE methods for 2 hours, respectively, and the concentrates were shown to retain the original peel aroma. The yield of each oil by solvent extraction was 0. 26% for lemon, 0.31% for lime, 0.10% for sudachi, 0.25% for yuzu, and 0.38% for kabosu. Volatiles of kabusu were extracted only by the SDE method ( 0.61% yield). Each oil was separated into hydrocarbon and oxygenated compounds fractions with a silica gel column and was analyzed by gas chromatography (GC) equipped with a flame ionization detector and gas chromatography-mass spectrometry (GC/MS). The oxygenated compounds had an aroma similar to those of the peels. Concentration of the volatiles in each oil was calculated from GC peak areas. Ethyl decanoate was added to each extract as an internal standard. The final concentration of ethyl decanoate was adjusted to 15ppm (w/w). For the purpose of identifying the volatile compounds, MS spectral patterns and retention indices of compounds identified in 6 acid citrus fruits were matched with those in the MS data library (containing about 35000 compounds) supplied by



10. T A M U R A E T A L .

Aroma Profiles of Peel Oils of Acid Citrus

123

NBS/NIH/EPA and those in our GC retention data library (containing 3900 compounds). The volatiles identified were: 72 compounds in lemon, 76 in lime, 69 in sudachi, 71 in yuzu, 69 in kabusu, and 77 in kabosu. Oxygenated compounds identified in each citrus peel oil are shown in Table I. Terpenoid compounds composed more than 76% of the oxygenated fractions of lemon, lime, yuzu, kabusu and sudachi peel oils. Lemon and lime peel oils contained mainly terpene aldehydes (>42%). On the other hand, yuzu, kabusu and sudachi contained terpene alcohols (> 46%). About 65 % of kabosu peel oil were aliphatic aldehydes. Similarity of aroma quality was olfactorily evaluated by a panel of 6 members. As shown in Table II, lemon and lime formed one cluster and also showed most similar relationship followed by sudachi, kabosu and yuzu which had 41% similarity. GC-sniff test using a flavor wheel In 1976 a system was designed by Acree et al. (4) for sniffing effluents from gas chromatograph effluents. This method was developed as an apparatus for the measurement of human response to aroma and for finding character impact compounds. Aroma Extraction Dilution Analysis (5) and Charm Analysis (6) can be used to determine dilution values of each effluent. Nitz (7) also developed multidimensional gas chromatography in combination with sniffing-MS monitoring system. In addition, the GC sniff test with a flavor wheel was used for the characterization of important aromas in navel oranges (8). The advantage of the GC-sniff test, with the development of columns which can separate enantiomers, is that natural aroma compounds can be evaluated to determine the sensory properties of each enantiomer. In many cases, the important aroma of a terpenoid is a contribution by one enantiomer. A flavor wheel, Figure 1, was used to record odor quality and intensity of each aroma constituent in acid citrus volatiles as each material was sniffed and sensory properties described as it was emerged from a GC sniff port. Prospective panel members were selected on the basis of their ability to distinguish standard aromas with a T & T olfactometer, available from Dai-ichi Yakuhin Sangyo Co., Ltd., Tokyo, Japan. Five standard aromas are tested at 8 different concentrations. The five standard aromas are: floral-like (j3-phenylethyl alcohol), burnt-like (methyl cyclopentenolone), roten-like (iso-valeric acid), fruity-like (7-undecalactone), and fecal-like (skatole). Each individual was then trained for three months to become familiar with the compounds involved and the relationship of the sensory properties and the odor descriptions of these compounds on the odor wheel. For descriptive terms, in aroma profiles, four main terms were selected: floral-like, fruity-like, woody-like and herbaceous-like. Thirty-two terms were used to form the wheel. An odor was ranked on a 1 to 5 point scale according to its intensity: 1 = very weak aroma; 3 = medium aroma; 5 = extremely strong aroma. Two and four points were given when the intensity was judged between 1 and 3, or 3 and 5. Each aroma constituent eluted was sniffed at the sniff port under humid atmosphere. Descriptive terms and the intensity of volatile components in oxygenated fraction of sudachi peel oil (9) are shown in Table III. Carvone, carveol,



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

behzaldehyde heptanol 6-methyl-5-hepten-2-one butyl butyrate octanal (Z)-3-hexenyl acetate hexyl acetate benzyl alcohol 1,8-cineole 2,6-dimethyl-5-heptenal octanol trans-sabinene hydrate cis-linalool oxide fcnchone trans-linalool oxide nonanal linalool heptyl acetate trans-p-mentha-2,8-dien-1 -ol trans-p-menth-2-en-1 -ol cis-limonene oxide cis-p-mentha-2,8-dien-1 -ol trans-limonene oxide camphor cis-p-menth-2-en- l-ol

hexanal (E)-2-hexenal (Z)-3-hexenol (E)-2-hexenol hexanol

No Compound

0.064 0.48 3.625 4.25 3.125 0.06 1.1 0.425 1 0.55 0.082 0.32 0.48 100 0.023 0.016 0.875 55 0.32 1.2 0.32 0.1 0.028 0.32 u u 0.25 u 0.25 4.6 u

Threshold (ppm)

0.458 15.958

0.621 6.982

34.065 24.520

1.198

3.903 13.808 0.048

0.954

0.026

14.431

0.463

2.167

Lime

1.927

0.018

1.217

0.337

Lemon

0.330

0.194

1.299 0.450

0.057 0.088

0.351 0.641 119.611 0.243

0.105 1.489

0.215

0.506 0.067 0.122

0.038

0.327

0.031

Yuzu

4.352 0.703 1.540

0.159 3.261 71.372

0.307 4.213 0.374

0.410

6.204

0.216 0.285 0.827

Sudachi

Concentration {ppm)

0.248

0.380

5.793 3.763

1.231 0.283

5.290

0.064 0.383 0.758 0.204 0.117

Kabosu

0.142 0.390

0.354

3.721 1.736 94.448 0.354 0.354 0.142

7.123

10.278

0.213

0.070

5.103

0.035 0.035

0.035 0.248

Kabusu 776 827 834 846 849 880 946 949 962 975 980 983 991 1007 1024 1036 1052 1059 1063 1072 1076 1085 1088 1090 1110 1112 1123 1124 1125 1125 1130

Found 775 827 836 845 848 882 947 950 965 977 981 986 993 1007 1030 1039 1053 1059 1068 1072 1076 1084 1088 1094 1111 1111 1122 1127 1125 1125 1128

Authentic Sample

Retention Index

Table I. Concentration and Odor Threshold of the Oxygenated Compounds in Feel Oils from Six Acid Citrus



37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

32 33 34 35 36

citronellal (E)-2-nonenol borneol nonanol 4-isopropyl-2-cyclohexen1-one (Z)-3-hexenylbutyrate terpinen-4-ol a-terpineol decanal octyl acetate cis-carveol grandisol citronellol trans-carveol nerol neral carvone piperitone geraniol (E)-2-decenal linalyl acetate geranial paraldehyde decanol thymol bornyl acetate peril! alcohol carvacrol geranyl formate undecanal (E,E)-2,4-decadienal nonyl acetate methyl geranate cis-carvyl acetate

0.1 0.062 0.775 0.79 1.38 7 2.29 0.2 0.04 0.01 0.6 u u

I

0.32 6.4 15 0.07 0.45 4 u 0.062 4 0.68 0.1 0.067 u 0.01 0.017

0.046 0.13 0.14 1 u

31.993 184.512 0.477 42.280

23.975 116.840 0.332 26.538

0.161 0.371 0.057

0.144 0.403 1.344 0.243

290.535 4.644 1.101 0.143 0.755 0.569 0.221 0.009 1.208 0.135

6.800

4.747

193.071 3.836 0.035 0.323 0.323

6.807 71.880 7.515 0.207

1.890

1.611 0.770

2.554 27.583 1.685 0.145

8.747

4.597

0.516 0.228

0.317 0.284

0.046

0.202

0.672

0.224

0.141

0.123

2.821

0.422

0.557 0.837 0.348

0.295

0.277 0.233

0.186 0.714

0.429 1.875 27.982 0.410

0.244 0.188

0.744 0.343 0.310 0.353

0.038

2.016 10.287 2.203 0.041 0.138

0.085

0.140 0.224 8.819

1.190 1.531

5.300 0.721

5.535

2.501

1.679 29.230 4.599

1.004

5.812 0.145

1135 1152 1155 1156 1165 1167 1169 1177 1185 1193 1198 1200 1209 1211 1211 1219 1224 1236 1236 1242 1247 1247 1252 1259 1267 1275 1276 1278 1279 1287 1292 1292 1302 1315

1132 1153 1153 1156 1165 1166 1167 1178 1184 1190 1198 1204 1209 1211 1212 1218 1220 1235 1236 1240 1242 1246 1253 1262 1268 1274 1275 1277 1277 1287 1289 1290 1299 1311

Continued on next page

0.957 0.142 1.453

1.240

3.473 5.635

57.377

17.578

8.541 0.248 3.579

1.950 31.223 9.108 8.718 0.815

2.410

0.248



o co vo CN CN 00 Tj- CN Tf co co CO vo 00 O N CM VO co co co co co co co co - vo CN d

ON

in o

o

CM 00 o

^

vo o oo in —« CN oo ^ oo' — co C N

^

vo vo d

TJ-

0 must impart the individual character in each given aroma. In these studies, detection thresholds of aroma were used. Perhaps recognition thresholds as well as detection thresholds of each compound should be used. The concentration at which panel members can no longer recognize individual characteristic flavor is defined as the recognition threshold. At the recognition threshold level of each citrus oil, Uo of each compound has a special important value. Thus, Uo at the recognition threshold of each oil is expressed by the following equation (2): Tr Limited odor unit (Lo)= (2) Tc where:

Lo is the limited odor intensity unit of the component, and Tr is the concentration of individual components at the recognition threshold of volatile oils in ppm, and Tc is the threshold of the components in ppm. It is an objective indication that compounds having Lo values more than 1 or nearly equal to 1 make considerable contributions to characteristic aromas. Recognition thresholds of sudachi and lemon oils were measured by the six members of the sensory panel and were found to be 0.033 and 1 ppm, respectively. The Lo values of individual compounds found in lemon oil, which has a recognition threshold of 1 ppm, were calculated. Neral, geranial, and geraniol (Uo>3.1) had Lo values of approximately 1. However, the components of sudachi oil had Lo values considerably below 1. Key compounds must be found to explain the recognition threshold of 0.033 ppm of sudachi oil. If aroma recognition thresholds in foods can be determined, compounds important in imparting the characteristic flavors can be easily determined because such compounds will have log [Lo] values greater than zero. Multivariate statistical analysis using logarithmic values of odor units of each component of 6 acid citrus fruits. Similarity of odor quality of 6 acid citrus fruits was calculated by cluster analysis with the log [Uo] of each component in the oils as shown in Fig. 3. Among the dendrogram patterns of cluster analysis from peak area %, odor unit and log [Uo], dendrogram of log [Uo] (Fig.3) were reflected in the result of sensory evaluation (Table II). Thus, lemon and lime were shown to be most similar, and sudachi, kabosu, yuzu formed another cluster. Kabusu does not seem to be similar to any of the others. By applying factor analysis and subsequently the VARIMAX procedure for


132

BIOACTIVE VOLATILE COMPOUNDS FROM PLANTS


Cone, (ppb) I Odor Unit (Uo)

1

1

I

-3 1

3 2

15 11

18 17

21 19

23 22

29 27

33 32

39 42 48 54 59 68 79 82 38 40 44 53 56 61 72 80

Peak number Figure 2. Concentration and odor unit o f the oxygenated compounds in sudachi peel o i l .

100

n

Lemon

Lime

Sudachi

Kabosu

Yuzu

Figure 3. Cluster analysis of peel oils from six acid citrus fruits.


Kabusu


10.

TAMURA ET AL.


133

eigenvector rotation, the first two eigenvectors explained 62% of the total variance. Factor analysis of aroma contribution using logarithmic values of odor units of components in 6 acid citrus fruits is shown in Fig. 4. The resultant factor score for the selected compounds (41 compounds) is added on to the result of Factor analysis. The loadings for 6 acid citrus fruits are shown as eigenvectors, the lengths of which roughly indicate their relative importance. The vectors between lemon and lime, and between sudachi and kabosu, indicate a similar direction. Compounds which exist in the same direction of each eigenvector of 6 acid citrus fruits were selected as the most contributing components to the aroma quality. Compounds contributing to individual aroma in 6 acid citrus are listed in Table V. Compounds in sudachi and kabosu selected from factor analysis (Fig. 4) were not always found in both oils. Therefore, compounds contributing to sudachi and kabosu were different. Compounds for lemon and lime were purported to be quite similar, and the balance of the composition between both fruits contributes to the variability in both oils. On the other hand, the difference of the aroma quality of sudachi, yuzu, kabosu and kabusu can be explained by the fact that there are different contributing compounds in each of these fruits. The differences of aroma quality among 6 acid citrus were attributed to geranial, neryl acetate and neral for lemon and lime; hexanal, cis-limonene oxide and carvone for sudachi; and carvone, 2-nonenal, and cis-limonene oxide for kabosu. Octanal, linalool and thymol were the compounds responsible for the yuzu aroma, and nonanol, linalyl acetate, 2,4-decadienal, and decyl acetate were the contributing compounds to kabusu. The model sample mixture prepared from compounds having values Log [Uo]>0 represented the individual aromas of 6 acid citrus fruits. However, we could not find any character impact compounds. Each citrus aroma seems be a combination of a mixture of these compounds identified, or new unknown compounds having large log [Uo] which have not been found. If all compounds having large log [Uo] values were found, each characteristic aroma should be reproduced by a mixture of such compounds. It was concluded that the GC-sniff method can directly evaluate the aroma quality of one citrus fruit. Log [Uo] values can also indicate contributing components from one sample, and it was not necessary to judge the aroma quality subjectively. Only subjective analysis must be the determination of odor threshold of the components. Multivariate analysis using log [Uo] can deal with several citrus species to evaluate each sample at the same time. From log [Uo] values it was shown that monoterpene aldehydes, alcohols and their esters, for example, geranial, geraniol, and neral significantly contributed to the characteristic aroma of lemon and lime. Citronellal, citronellol and carvone contributed to sudachi, and citronellol, 2-decenal, geranyl acetate to kabosu. Geraniol, linalyl acetate, geranyl acetate and linalool contributed to kabusu. Unsaturated aliphatic aldehydes contributed especially to yuzu.



134

BIOACTTVE VOLATILE COMPOUNDS F R O M PLANTS

Factor 1 (37.6%)

A B C D E F

-i 1.0

48

- Lemon - Lime - Sudachi - Yuzu - Kabosu - Kabusu

9

23 22

-£2_ 1.0

-1.0

Factor 2 (24.9%)

72

60 61 44

88

39

%

82 15 46

32

77 53 3447

Figure 4. Factor analysis of 41 oxygenated compounds o f the peel oils from six acid citrus.


10.

135


TAMURAETAL.

Table V. Compounds Selected by Factor Analysis


No. Compound 1 11 15 17 19 22 23 27 29 33 34 35 41 46 47 48 52 53 54 55 56 60 61 62 63 68 69 73 75 77 79 82

hexanal octanal 1,8-cineol octanol cis-linalool oxide nonanal linalool cis-limonene oxide trans-limonene oxide (E)-2-nonenol borneol nonanol octyl acetate nerol neral carvone linalyl acetate gexanial perillaldehyde decanol thymol geranyl formate undecanal (EJE)-2,4-decadienal nonyl acetate citxonellyl acetate neryl acetate decyl acetate (E)-2-dodecenal tridecanal elemol tetradecanal

Lemon

lime

Sudachi

Yuzu

Kabosu

Kabusu

O O O

O O O

O

O O O O

O O

O

O O O

O

O O O

O °

O O O O O O O O O O

O O O O

O

O

O

O

O


136


Acknowledgements


The authors are very grateful to Mr. Kenji Nakanishi and Nobuo Takagi for supplying the acid citrus fruits. We also thank Mr. Hiroyuki Nakatani for his technical assistance. This work was partially supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan. Literature Cited 1. Shaw, P. E.; Coleman, R. L. J. Agric. Food Chem. 1974, 22, 785-787. 2. Wilson, C. W.; Shaw, P. E. J. Agric. Food Chem. 1978, 26, 1432-1434. 3. Shaw, P. E. J. Agric. Food Chem. 1979, 27, 246-257. 4. Acree, T. E.; Butts, R. M . ; Nelson, R. R.; Lee, C. Y. Anal. Chem. 1976, 48, 1821-1822. 5. Ullrich, F.; Grosch, W. Z. Lebensm. Unters. Forsch. 1987, 184, 277-282. 6. Acree, T. E.; Barnard, J.; Cunningham, D. G. Food Chem. 1984, 14, 273-286. 7. Nitz, S.; In Topics in FlavourResearch;Berger, R. G.; Nitz, S.; Schreier, P.; Eds.; H . Eichhorn, Marzling-Hangenham, 1985, pp 43-58. 8. Yang, R-H.; Otsuki, H . ; Tamura, H . ; Sugisawa, H . Nippon Nogeikagaku Kaishi; 1987, 61, 1435-1439. 9. Yang, R-H.; Sugisawa, H . Nippon Shokuhin Kogyo Gakkaishi, 1990, 37, 946-952. 10. Guadagni, D. G.; Okano, S.; Buttery, R. G.; Burr, H . K. Food Technol. 1966. 20, 166-169. 11. Rothe, M.; Thomas, B. Z. Lebensm. Unters. Forsch. 1963, 119, 302-310. 12. Shibamoto, T. J. Food Sci., 1986, 51, 1098-1099. 13. Buttery, R. G.; Teranishi, R.; Ling, L. C.; Turnbaugh, J. G. J. Agric. Food Chem., 1990, 38, 336-340. 14. Sugisawa, H . ; Takeda, M . ; Yang, R-H., Takagi, N . Nippon Shokuhin Kogyo Gakkaishi, 1991, 38, 668-674. 15. Yang, R-H.; Sugisawa, H.; Nakatani, K.; Tamura, H . ; Takagi, N . Nippon Shokuhin Kogyo Gakkaishi, 1992, 39, 16-24. RECEIVED August 19, 1992


Aroma Profiles of Peel Oils of Acid Citrus - American Chemical Society

Recommend Documents