STUDIES ON FLAVOR CHANGES DURING DRYING OF DILL (Anethurn sowa Roxb.) LEAVES

STUDIES ON FLAVOR CHANGES DURING DRYING OF DILL (Anethurn sowa Roxb.) LEAVES B. RAGHAVAN,'v3 K.O. ABRAHAM,' M.L. SHANKARANARAYANA' and W.D. KOLLER' 'Plantation Products and Flavour Technology Central Food Technological Research Institute Mysore 570 013, India 'Institute of Process Engineering Federal Research Centre for Nutrition Engesserstrasse 20, D 7500 Karlsruhe I, Germany Accepted for Publication March 22, 1994 ABSTRACT Indian dill (Anethum sowa Roxb.) is a popular herb widely used in culinary preparations and is characterized by a persistent cineolic flavor. To study the flavor quality of the fresh and dried herb, it was subjected to through-flow, cross-flow, vacuum andfreeze drying procedures. The volatile oils obtained from the fresh and dried samples were examined by gas chromatography and gas chromatography - mass spectrometry and the changes in their flavor compositions determined. Seventeen compounds have been identifed in Indian dill leaf oil, of which P-cubebene is being reported for the first time. The proximate composition of the dried herb is also reported. INTRODUCTION Dill (Anethurn sowa Roxb.) is a popular herb used for its flavor in culinary preparations. It is an annual plant of the Umbelliferae family. The whole plant is regarded as a herb and is used in the same manner as other herbs such as parsley and rosemary. There are several reports available on the flavor composition of dill seed (Hendry 1982; Huopalahti et al. 1981; Koedam et al. 1979; Baslas and Baslas 1971; Blank and Grosch 1991; Schreier et al. 1981; 'To whom correspondence should be addressed. Journal of Food Quality 17 (1 994) 457-466. All Rights Reserved. Copyright 1994 by Food & Nurrition Press, Inc. Trumbull, Connecticut. 457

458 B. RAGHAVAN ETAL. Huopalahti and Link0 1983; Huopalahti 1985,1986). However, there is no study relating to the drying of the whole herb or part thereof and the changes that occur in the dried herb with reference to the flavor quality. This paper reports the results of studies on the aroma profile of the steam distilled volatile oils obtained from the fresh herb and dried herbs that have undergone different drying processes. Data on the chemical composition of dill leaves are also presented. MATERIALS AND METHODS Fresh dill herb was procured from the local market as required. The herb is generally available with the stem, leaves and roots. In all drying experiments, the roots were cut off and rejected. The leaves were dried either separately or with the stems. Drying Studies The drying of the herb was carried out using through-flow, cross-flow, vacuum and freeze dryers. Through-Flow Drying The drying experiment was carried out using a through-flow dryer fabricated in CFTRI, with the objective of determining the loss of volatile oil during the course of drying. Stems and leaves of dill herb (5.3 kg) were distributed in two wire mesh trays (73 x 71 cm) of the dryer and filled to a height of 4 cm and dried for 9 h, at a temperature of 45 f 2C. Samples were drawn at hourly intervals and the moisture and volatile oil content determined by standard procedures (ASTA 1985). In another experiment, fresh dill leaves (5.0 kg) were obtained. Half the quantity was initially blanched by dipping in boiling water for one min, followed by draining. The blanched herb was then placed in the upper wire mesh tray and dried (45 f 2C) in the through-flow dryer as before. The other half was dried as such without blanching. The moisture and the volatile oil content of the fresh and dried herb were determined. Cross-Flow Drying A 1 kg sample of fresh dill leaves was equally distributed in two stainless steel trays (80 x 40 cm) and dried in a cross-flow dryer (Armstrong Smith Pvt.

FLAVOR LOSS DURING DRYING OF DILL LEAVES 4.59 Ltd., India) for a period of 5 h at 40 k 2C. The drying was repeated for a blanched sample (1 kg) at the same temperature and for the same period. The moisture and volatile oil content in the two samples were determined. Vacuum Shelf-Drying Drying was carried out using a vacuum shelf dryer (Stokes Process Equipments, F.J. Stokes Machine Co., Philadelphia, PA) for the fresh and blanched dill herb (1.525 kg each). The fresh material was spread on 6 aluminium trays (58 x 30 cm), loaded into the dryer and the vacuum applied (25 in.). The drying time and temperature were 8.75 h and 45C, respectively. The experiment was repeated for the blanched sample under the same conditions, the drying time being 7 h. The moisture and volatile oil content in both the samples were determined. Freeze Drying The fresh dill leaf (1.8 kg) was subjected to freeze drying (Leybold Heraus, Germany) by distributing the material in 8 trays (58 X 30 cm) and drying for 9.5 h. The freeze drying was carried out under a pressure of 300-150 microns and a shelf temperature of 40C. The experiment was repeated for the blanched sample (2.2 kg) by spreading it in 9 trays and drying for 7.5 h. The moisture and volatile oil content were determined in both the samples. Proximate Composition of Dill Leaves The dried dill leaves were analyzed for ash, acid-insoluble ash, protein, starch, crude fiber and chlorophyll content according to standard procedures (Ranganna 1986). Gas Chromatographic Examination of Dill Oils Fresh dill procured from a local market, was thoroughly washed under running water to get rid of the adhering earthy matter and the root portion was eliminated. The remaining part of the herb along with the stem (2.0 kg) were cut into smaller pieces and subjected to hydrodistillation as follows: the herb was transferred into a 20 L round bottomed flask, mounted on a heating mantle; 8.0 L water was added to the flask connected to a laboratory-fabricated glass trap

460 B. RAGHAVAN ET AL. closely akin to the conventional Clevenger apparatus. The flask was heated and distillation carried out for 6 h. The yield of oil was 4.2 ml. On-column injection of the volatile oil diluted with hexane (1:200) was carried out on a Perkin-Elmer Sigma 3B gas chromatograph equipped with a flame ionization detector. The instrument was fitted with a fused silica capillary column (25 m x 0.2 mm i.d.) bonded with HP-Ultra 2 (5% diphenyl and 95% dimethyl polysiloxane). Column temperature was programmed at 50C for 2 min followed by an increment of 3C per min up to 300C and the injector and detector temperatures were maintained at 220 and 260C, respectively. Column pressure was 14 psi and the split ratio, 1:lOO. Helium was used as the carrier gas (1.O mllmin); hydrogen flow, 30 ml/min; air flow, 300 ml/min and sample injection volume 0.2 p1. The GC profile was recorded and integration of peak areas obtained through a computer (Trilab 2000, Trivector Scientific Ltd., Nieder-Olm/Mainz, FRG). Other samples of volatile oils obtained from fresh stem and dried dill were analyzed using a Shimadzu 15 APF gas chromatograph hooked to a Shimadzu C-R3A Chromatopac. The GC was fitted with a SS column (10 ft x 0.125 in. i.d.) packed with 5% SE 52 on chromosorb. The column temperature was programmed from 75C to 215C at the rate of 5C per min. The injector block and the detector (FID) temperatures were 220 and 230C, respectively. Nitrogen was the carrier gas. The sample injection volume after 1:20 dilution with acetone was 1 pl. Gas Chromatography - Mass Spectrometry A Hewlett Packard 5985 GC-MS Data System in which the GC was fitted with a fused silica capillary column (25m x 0.2 mm i.d.) bonded with HP-Ultra 2 (5 % diphenyl and 95 % dimethyl polysiloxane) and directly inserted into the ion source of the MS was used for the analysis of the volatile oil obtained from the fresh dill herb. The conditions under which the GC-MS analysis was carried out were the same as described earlier. RESULTS AND DISCUSSION Although dill is a delicate herb, it has a strong and persistent flavor. Drying of this herb is a time-consuming process. Blanching was found to help in the reduction of drying time. Without blanching and using a flow-through dryer over a period of 9 h, the volatile oil content dropped from 1.62% to 0.34%. About 75% of the volatile oil was lost in the first 3 h of drying, when the moisture

FLAVOR LOSS DURING DRYING OF DILL LEAVES 46 1 content was reduced by almost 50%. The stem isolated from the herb was found to contain substantial amount of volatile oil (0.46-0.68). In the first experiment, the herb was dried as such (including the stem), which took more than 9 h to dry, resulting in a heterogenous product consisting of over-dried leaf with dark brown color and dried stem. Therefore, in subsequent experiments, the stem was removed before drying. When blanching was used, dill leaves dried in 4.5 h compared with 7.5 h for the unblanched product. The loss of volatile oil was higher, 86.7%, for the blanched product compared with 73.8% for the unblanched sample. When cross flow drying was used, the time taken for drying was only 5 h. This was mainly due to the fact that the quantity employed for drying was only one kilogram, which was distributed equally in two trays. For the unblanched and blanched samples the losses were 16 and 55 %, respectively. Vacuum drying took longer. The loss of volatile oil for the blanched sample was 96.5% compared with 74.7% for the unblanched sample. Freeze drying resulted in a product of better color and appearance as examined visually; the loss of volatiles was 57.8% for the blanched and 69.7% for the unblanched samples. However, for a product of this type, freeze drying may be untenable, being too expensive. Information on the chemical composition of dill leaves is not readily available. Table 1 presents data on the chemical composition of dill leaves. This will be useful in establishing standards and specifications for the product. It can be seen that dill leaves are a rich source of protein apart from its value as a flavoring agent. The chlorophyll contents ranged from 0.58 to 1.00%. TABLE 1. PROXIMATE COMPOSITION OF DILL LEAVES Constituent Compostion X Ash 8.00-9.95 Acid insoluble ash 1.20-1.63 Protein (N x 6.25) 26.61-33.31 Starch 5.06-10.01 Crude fiber 10.01-13.21 Chlorophyll, total 0.56-1.04 Flavor Quality of the Dried Herb Gas chromatographic analysis was carried out on the volatile oils distilled from the fresh herb (Fig. l), stem and from the dried herbs obtained by

462 B. RAGHAVAN ETAL..

FLAVOR LOSS DURING DRYING OF DILL LEAVES 463 FIG. 2. CHROMATOGRAPHIC PROFILE OF FRESH DILL OIL OBTAINED USING GC-MS (1) a-pinene, (2) camphene, (3) sabinene, (4) 0-pinene (5) Myrcene. (6) a-phellandrene, (7) a- Terpinene, (8) p-cymene, (9) @-Phellandrene, (10) y-terpinene (1 1) Terpinolene, (12) Undecane, (13) @-Cubebene (14) Myristicin, (15) dillapiole, (16) Apiole. TABLE 2. FLAVOR CONSTITUENTS IDENTIFIED IN FRESH DILL OIL Component Peak no. RTCHin) us Nonane 3 9.60 + a-pinene 5 11.19 + Camphene 6 11.87 + Sab inene 7 13.09 + 8-P i nene 8 13.24 + Hyrcene 9 13.96 + a-phel landrene 10 14.98 + a-terpinene 11 15.30 + p-cymene 12 15.74 + 8-Phel landrene 13 16.05 + I--Terpinene 16 17.31.( Terpinolene 17 18.95 + Undecane D-Cubebene Hyristicin 18 33 34 19.55 38.20 39.81 Dillapiole 35 44.70 + Apiole 36 45.68 +

TABLE 3. RELATIVE CONCENTRATIONS OF VOLATILE CONSTITUENTS IN FRESH AND DRIED DILL HERB OILS a-phellandrene 60.3843 50.4081 40.6102 20.3973 41.7672 28.6600 P 5 a-terpinene 11.1430 14.2726 23.5740 15.9580 18.5266 7.3395 2. J3-Phellandrene 1.6664 1.6409 1.7598 0.4149 0.6238 0.9564 Undecane 0.2275 0.5551 0.9866 0.4187 0.9168 0.7752 D-Cubebene 2.4149 0.9549 1.7141 4.5051 2.2402 10.2880 Myristicin 0.3077 0.2410 0.1670 1.1215 0.3135 0.9668 Dillapiole 16.7208 22.1077 23.2892 40.0143 21.5577 38.8850 Apiole 0.1856 0.2180 0.2465 0.3021 0.2715 0.4439 QI P Q & Compound Percentage concentration of volatile constituents in - Fresh Stem oil Oil from Oil from Oil from 1 eaf cross flow through vacuum Oil from f reeze oi 1 dried herb flow dried shelf dried herb herb dried herb a-pinene 1.6214 0.2490 I.4217 3.0832 1.9459 0.1209 m B-Pinene 1.4595 2.0838 1,1296 0.5648 1.1375 0.5685 c 8

FLAVOR LOSS DURING DRYING OF DILL LEAVES 465 different drying methods. The identification of the flavoring components was achieved by the use of GC-MS (Fig. 2). Based on the GC retention time and also matching with the reference mass spectra, 17 components have been identified (Table 2). Whereas Indian dill is reported to contain carvone and dillapiole as the major constituents, the present study has revealed the presence of a-phellandrene as the major constituent, followed by dillapiole. 6-Cubebene, a sesquiterpene, has been identified for the first time in dill oil. The relative concentration of the major components, in the fresh leaf oil, stem oil, and then oils obtained from dried dill herb are presented in Table 3. The presence of a-phellandrene to the extent of about 60% lends a fresh herbal character to the leaf oil, and this is in consonance with the observation that the higher the content of phellandrene, the more the oil resembles the fresh herb (Virmani and Datta 1970). a-phellandrene is reported to be the character-impact compound of the dill flavor, which is rounded off by dillapiole (Blank ef al.). The combined concentrations of a-phellandrene and dillapiole in fresh leaf oil is almost retained in all the dried samples obtained by the different drying methods. However, a-phellandrene decreased considerably in its concentration, in through-flow and freeze dried products. The ratio of a-phellandrene to dillapiole in the samples of cross-flow and vacuum-shelf-dried products was nearer to that of the fresh leaf oil, whereas the relative Concentration of dillapiole was much higher in case of through-flow and freeze dried samples. Considering the time taken for drying and flavor quality of the dried herb, cross-flow drying is the preferred method for obtaining a quality dried herb. ACKNOWLEDGMENTS One of the authors (BR) is thankful to Council of Scientific and Industrial Research, New Delhi and German Academic Exchange Service (DAAD), Bonn for providing him an opportunity for a study visit to Germany where part of the work was carried out. The authors are also thankful to Dr. S.R. Bhowmrk, Director, CFTRI for his keen interest in this work. REFERENCES ASTA. 1985. official Analytical Methoa's of the American Spice Trade Assodation, 3rd Ed., 68 pp. American Spice Trade Assoc., Englewood Cliffs, NJ.

466 B. RAGHAVAN ETAL. BASLAS, B.K. and BASLAS, R.K. 1971. Chemical studies of the essential oils from the plants of Anethum graveolens and Anethum SOW (dill oils). Indian Perfum. 15, 27-29. BLANK, I. and GROSCH, W. 1991. Evaluation of potent odorants in dill seed and dill herb (Anefhum graveolens L.) by aroma dilution analysis. J. Food Sci. 56, 63-67. BLANK, I., SEN, A. and GROSCH, W. 1992. Sensory study on the characterimpact flavour compounds of dill herb (Anefhum gruveolens L.). Food Chem. 43, 337-343. HENDRY, B.S. 1982. Composition and characteristics of dill: A review. Perf. Flav. 7, 39-44. HUOPALAHTI, R. 1985. The content and composition of aroma compounds in three different cultivars of dill (Anethum graveolens L.). Lebensm. Wiss. Technol. 19, 27-30. HUOPALAHTI, R., KALLIO, H., KARPPA, P. and LINKO, R.R. 1981. Comparison of two isolation procedures for aroma compounds of dill. In Flavour 81, (Peter Schreier, ed.) pp. 369-376, Walter de Gruyter & Co., Berlin, NY. HUOPALAHTI, R. and LINKO, R.R. 1983. Composition and content of aroma compounds in dill at three different growth stages. J. Agric. Food Chem. 31, 331-333. KOEDAM, A., SCHEFFER, J.J.C. and BAERHEIM SVENDSEN, A. 1979. Comparison of isolation procedures for essential oils. I. Dill (Anethum graveolens L.). Chem. Mikrobiol. Technol. Lebensm. 6, 1-7. RANGANNA, S. 1986. Handbook ofanalysis and Qualify Control forfruit and Vegetable Products, 2nd Ed., 112 pp, Tata McGraw-Hill Pub. Co. Ltd., New Delhi. SCHREIER, P., DRAWERT, F. and HEINDZE, I. 1981. The quantitative composition of natural and technologically changed aromas of plant. VIII. Volatile constituents of fresh dill herb, Anethum graveolens L. (Umbelliferae). Lebensm. Wiss. Technol. 14, 150-152. VIRMANI, O.P. and DATTA, S.C. 1970. Essential oil of Anefhum graveolens L. Flav. Ind. I, 856-862.