JOURNAL

OF FOOD

COMPOSITION

AND ANALYSIS

4, 304-3 14 (199 1)

Analysis of Anisated Fermented Grape Distillates by Gas-Liquid Chromatography

A. GEAHCHAN,* C. KHALIFE,~ P. CHAMBON,*
AND R. CHAMBON*

*Laboratoire de Toxicologic, Faculte de Pharmacie, 8, avenue Rockefeller, 69373, Lyon, France; and YLaboratoire de Toxicologic, Faculte de Pharmacie, UniversitP Saint-Joseph, Bert& Lebanon Received February 26, 199 1, and in revised form June 14, 199 1 Two gas-liquid chromatographic methods have been developed for the identification and quantitative determination of methanol and other volatile compounds in anisated fermented grape distillates. Methanol, acetaldehyde, ethanol, propanol, isopropanol, ethyl acetate, butanol, isobutanol, and amyl + isoamyl alcohols were all resolved on a Porapak Q column, using tertbutanol as internal standard. A column containing 15% EGA on 80- 100 mesh Supelcoport was used for rapid quantitative estimation of anethole in the presence of anisaldehyde as an internal standard. Confirmation of the presence of different constituents in anisated spirits was investigated using a head-space injector coupled to a capillary gas liquid chromatograph. Analysis of 30 commercial and 19 artisanal samples respectively obtained for a survey of anisated grape distillates constituents in Lebanon revealed the absence of ethyl acetate, isobutanol, butanol, amyl, and isoamyl alcohols in 10 commercial samples produced by manufacturers of high proof industrial alcohol. Ten other commercial samples showed methanol levels above the maximum allowable concentration (100 mg/liter) defined by the European Economic Community. o 1991 Academic
Press. Inc.

INTRODUCTION

Alcoholic beverages obtained from fermented grape juice and flavored by adding aniseed and distilling again are largely consumed in Lebanon and to a limited extent in the neighboring countries of the middle east such as Israel, Turkey, and Greece where they are known as arak, raki, and ouzo respectively. Arak is the national alcoholic drink in Lebanon. It is exported to different European and American countries. With the continuing and rising popularity of anise-flavored grape distillates, there has been a corresponding increase in the brands available with wide price differentials. This type of situation increases the possibilities of the less expensive brands being illegally substituted for the more expensive brands by some retail alcoholic beverage dealers. Arak is subject to different falsifications, such as adding extract to industrial alcohol obtained from fermented sugar cane or sugar beet juice. Imitations include any type of distilled spirit to which aniseed extract has been added. Artisanal manufacturers have also been using methanol instead of alcohol resulting in severe intoxications and death (unpublished data). Quantitation of volatile alcoholic congeners by gas-liquid chromatography was suggested for the characterization of distilled alcoholic beverages. The distribution of volatile alcoholic congeners was used by Schoeneman et al. ( 197 1) and Schoeneman and Dyer (1973) to establish analytical profiles of classes of distilled spirits such as bourbon, scotch, vodka, gin, and tequila. Few and incomplete data are available on
To whom reprint requests should be addressed. 0889-1575/91 $3.00
Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

304

ANALYSIS

OF GRAPE DISTILLATES TABLE 1

BY GLC

305

LINEAR

REGRESSION

OF DIFFERENT COMPONENTS: INTERNAL STANDARD vs COMPONENTS CONCENTRATIONS (n = 3)

Concentration (g/L) Peak ht ratio Correlation I-

RESPONSE

RATIO

component

coefficient

Methanol

1-8

0.676

- 5.546

0.999

Acetaldehyde

1-8

0.508

- 3.953

0.993

Ethyl

acetate

1-8

0.218

- 1.700

0.998

Propanol

1-8

0.528

- 3.873

0.999

Isopropanol

l-8

0.510

- 3.998

0.999

Isobutanol

1-8

0.260

- 2.110

0.998

Butanol

l-8

0.246

- 1.782

0.998

Amy1 + Isoamyl alcohols

1-8

0.110

- 1.001

0.998

Anethole

1-6

0.840

- 4.928

0.999

Note. Peak height ratio = component peak height/internal standard peak height. Internal standard, anisaldehyde for Anethole measurement, terbutanole for all other components.

the composition of anisated grape distillates (Dagher and Ruhayem, 1975). Grape brandy distillates which are manufactured in a similar way to arak except for the manner in which the flavor is added were reported to contain 70-300 ppm (w/v) of propyl alcohol, lo-250 ppm (w/v) of isobutyl alcohol, and 220-880 ppm (w/v) of amyl alcohol (Guymon, 1970). A number of brandy samples from various parts of the world have been analyzed by Askew and Lisle (197 1) who found propyl alcohol to be present in the range of 20-l 3,320 ppm, isobutyl alcohol 10-1080 ppm, amyl alcohol 40-3000 ppm, and methanol in the range of 20-8750 ppm (w/v). Bamett and Einsmann (1977) reported the presence of isobutanol, butanol, amylol, and 1-hexanol in grape brandy distillates in the range of 3-36, 3-12, 10-40, and 7-57 ppm (w/v), respectively. Other distilled beverages have also been analyzed for their fuse1 oil content (Kahn and Blessinger, 1972; Martin and Caress, 197 1, Onishi et al., 1978). No data are available on the anethole content of anise-flavored grape distillates. Anethole, a primary constituent of anise oil, contributes largely as fuse1 oil constituents to the consistency in flavor and quality of the final product which requires a suitable and constant level of these constituents.

306

GEAHCHAN

ET

AL.

8
MIN

12

16

20

24

FIG. 1. Gas chromatogram of ethanolic standard solution on Porapak Q containing 2 mg/ml of the following constituents: (1) methanol, (2) acetaldehyde, (3) ethanol, (4) isopropanol, (5) propanol, (6) internal standard let?-butanol, (7) ethyl acetate, (8) isobutanol, (9) butanol, and (10) amyl + isoamyl alcohols.

Various methods have been reported for analysis and separation of fuse1 oils in brandy distillates (Kahn and Blessinger, 1972; Martin and Caress, 197 1; Schoeneman and Dyer, 1973) and volatile compounds in wines (Reglero et al., 1986; Loyola et al., 1988; Herraiz et al., 1989), and anethole in anisated liqueurs (Parker, 1974), using gas-liquid chromatography and different stationary phases. This paper reports the development of a new and rapid method for determining methanol, acetaldehyde, ethyl acetate, fuse1 oils, and anethole in anise-flavored brandy distillates. The method was used to analyze a large number of samples and resulted in a full range of data concerning commercial, and to a lessextent artisanal manufactured, anisated distillates. METHOD Apparatus Equipment specified is not restrictive; other suitable equipment may be used. Gas chromatographs. (1) Pet-kin-Elmer Model Sigma 2B equipped with flame ionization detector. GC operating conditions: injection port, 19OC; detector, 250C.

ANALYSIS

OF GRAPE DISTILLATES

BY GLC

307

8M,N12

16

20

24

FIG. 2. Gas chromatogram of an&ted distilled spirit produced from fermented grape juice on Porapak Q. (1) methanol, (2) acetaldehyde, (3) ethanol, (5) propanol, (6) internal standard terf-butanol, (7) ethyl acetate, (8) isobutanol, and (10) amyl + isoamyl alcohols.

Columns: (a) stainless steel, 1.8 X 2-mm i.d. packed with Porapak Q; carrier gas flow, nitrogen, 30 ml/min, oven temperature, 165C. (b) Glass column, 2 X 2-mm i.d. packed with 15% ethylene glycol adipate (EGA) coated on 80- 100 mesh Supelcoport; carrier gas flow, nitrogen, 25 ml/mm, oven temperature, 180C. Computing integrator, Perkin-Elmer Model LCI 100. Operating conditions: chart speed, 10 mm/min, attenuation, 1. (2) Perkin-Elmer Model Sigma 2000 capillary gas chromatograph. GC column: fused silica 50 X 0.32-mm i.d. CP Sil8B with 1.2~pm Iilm thickness. GC operating conditions: temperature programmed as follows: initial, 45C; initial hold time, 3 min; programming rate, SC/min (45-250C); hold 2 min at 250C; carrier gas flow, helium, 1 ml/ min. Detector: flame ionization, temperature, 300C. Head-space sampler: PerkinElmer Model HS- 100 with the following operating conditions: sample temperature, 55C; transfer temperature, 60C; thermostating time, 60 min, pressurization time, 2 min; injection time, 8 min; withdrawal time, 0.2 min; ventilation and split mode

308

GEAHCHAN

ET AL.

2
MIN

ii

FIG. 3. Gas chromatogram of anisated distilled spirit on glass column packed with EGA 15% coated on Supelcoport. (1) anethole, (2) anisaldehyde.

injection. Computing integrator Hitachi-Merck speed, 10 mm/min; attenuation, 2. Reagents

D 2000. Operating conditions: chart

Unless otherwise noted, all the chemicals used in this work were analytical grade, commercially available reagents (from Merck). Anethole and anisaldehyde were obtained from Fluka Chemical Corp. They are supplied as pure grade chemicals with a purity > 98% GC. Water was distilled and deionized. Ethanol + water solution (50

ANALYSIS

OF GRAPE

DISTILLATES

BY GLC

309

TABLE 2
ACETALDEHYDE, ETHYL ACETATE, AND VARIOUS ALCOHOLS CONTENT MANLJFAC~URED BY COMMERCIAL DISTILLERS OF

30 ARAK SAMPLES

20 ND ND

Note. ND = not detected. y The limits of detection were as follows: 5 mg/liter for methanol, acetaldehyde, isopropanol, 11, 8, 10, and 22 mg/liter for ethyl acetate, isobutanol, butanol, and amyl + isoamyl alcohols,

and propanol;
respectively.

+ 50 v/v) was used for the preparation of internal and standard solutions. Tert-butanol and anisaldehyde internal standard solutions were prepared at 400 and 800 mg/l, respectively. Alcohols, aldehyde, and ester standard mixture were prepared and adjusted to give a final solution containing 2 mg of each component per milliliter. The final reference solution was stored in refrigerator and was stable for 15 days. Anethole standard solution was prepared at 2 mg/ml. The solution was kept in the dark and made fresh every 24 h.

Standardand SampleAnalysis (a) Methanol,acetaldehyde, ethyl acetate, andfuse1oils determination.Quantitatively transfer 1 ml aliquot of standard solution to a 5-ml glass stoppered test tube. Add 5 ~1 of tert-butanol internal standard solution. Vortex and inject 2 ~1 into gas chromatograph using Porapak Q as the stationary phase. To a l-ml an&ted distillate sample in a 5-ml glass stoppered test tube add 5 ~1 of tert-butanol internal standard solution and continue as above. (b) Anetholedetermination.Quantitatively transfer a l-ml aliquot of standard solution to a 5-ml glass stoppered test tube and add 5 ~1 of anisaldehyde internal standard solution. Vortex and inject 1 ~1 into the gas chromatograph using EGA as the stationary phase and relative operating conditions described above. Repeat the same procedure for sample analysis substituting standard solution with an an&ted brandy distillate sample. (c) Calculation.Determine methanol, acetaldehyde, ethyl acetate, and fuse1 oils concentrations by measuring the area ratio of each analyte to tert-butanol of sample

310

GEAHCHAN TABLE ACETALDEHYDE,

ET 3

AL.

ETHYLACETATE,ANDVARIOUSALCOHOLS MANUFACTURED BY ARTISANAL

CONTENT DISTILLERS a

OF 19 AUKSAMPLES

Note. ND = not detected.

(1The limits of detection 1 I, 8, 10, and 22 mg/liter were as follows: 5 mgJliter for methanol, acetaldehyde, isopropanol, and propanol; for ethyl acetate, isobutanol, butanol, and amyl + isoamyl alcohols, respectively.

and standard peaks. Determine anethole concentration by measurement of the area ratio of anethole to anisaldehyde of sample and standard peaks. Proof and Total Acidity Determination Proof was determined by the AOAC hydrometer method (957.03), and total acidity by the AOAC titrimetric method (945.08) (Dyer, 1990).
RESULTS AND DISCUSSION

Linearity In order to study the linearity of the above method, concentrations of different analytes in anisated fermented grape distillates were assayed in concentrations from 1 to 8 mg/ml. Anethole levels were varied from 0 to 6 mg/ml. For each component, response ratios were plotted vs analyte concentrations. Results of the least-squares linear regression analysis (Table 1) indicated good linearity. Nonzero intercepts were negligible. Reproducibility Within-day precision of the method was assessed by analyzing standard solution samples containing all different constituents (including anethole) at 2 mg/ml of each.

ANALYSIS

OF

GRAPE TABLE

DISTILLATES 4

BY

GLC

311

PROOF,

ACIDITY, AND ANETHOLE CONTENT MANUFACTURED BY COMMERCIAL

name Proof ( mg/lOO

OF 30 ARAK DISTILLERS

SAMPLES

Commercial

Acidity mL acetic

acid

Anethole ( g/L 1

Abi - chaker Abi - haila Abou - akl Al - arz Ayoub Balaban Chabat Dakach Dawal i Excellent Fakra Gergi Abi-raad Ghantous & Abou-raad Jeita Kef raya
KE3PSJXl

101 102
140 96 113 90

110
108 103 100 122 96 101 88 102 110 122 100 103 106 104 97 90 107 101 96 107 98 83 92

Khazen Kortbawi Ksarak Lebrun Lobos Marechal Mechaalani Musar Nakad Rayes Abou-kalil Rif Sibheli Stephan TOLUI~

601 715 1080 2210 1907 476 1669 1311 1320 1072 1192 1490 476 541 1549 1503 1847 2861 476 2026 360 26 661 3338 834 1625 2503 2980 830 59

2.6 2.6 2.8 2.4 2.4 1.8 2.5 2.7 3.0 2.4 3.5 2.2 2.4 1.2 2.4 3.8 3.0 2.5 2.1 3.0 2.3 1.6 2.0 2.7 2.3 2.1 2.9 2.3 1.3 1.9

At this concentration, the method yielded for methanol, acetaldehyde, propanol, ethyl acetate, isobutanol, butanol, amyl + isoamyl alcohols, isopropanol, and anethole coefficients of variation (n = 6) of 6.7,5.7, 1.9,3.6, 1.3,0.93, 1.2, 1.5, and 3.0%, respectively. Day-to-day precision was determined by six repeated analyses of the same samples in three different days, The coefficients of variation (n = 18) were 9.2, 8.1, 3.1, 6.4, 1.3, 3.0, 3.0, 3.1, and 6.1%, respectively. Limits of Detection The limit of detection was calculated by assuming arbitrarily that a detector response equivalent to 5 times the baseline noise was the minimum peak height that could be appreciated with a reasonable confidence. Limits of detection were 5 mg/liter for methanol, acetaldehyde, isopropanol, and propanol, 11, 8, 10, and 22 mg/liter for ethyl acetate, isobutanol, butanol, and amyl + isoamyl alcohols, respectively. Resolution of D@erent Constituents on Column Chromatography Different column packings and stationary phases for the separation and quantitative estimation of different fuse1 oil constituents of various brands of distilled alcoholic beverages were reported in the literature (Guymon, 1970; Barnett and Einsmann, 1977; Onishi et al., 1978). The use of these stationary phases resulted in an incomplete

312

GEAHCHAN TABLE
ANETHOLE CONTENT OF 11 ARAK number SAMPLES

ET 5

AL.

MANUFACTURED Anethole

BY ARTISANAL ( g/L )

DISTILLERS

Sample

1 2 3 4 5 6 7 8 9 10 11

4.2 3.9 1.8 3.1 1.9 3.1 2.1 0.5 0.9 1.8 3.5

a Anethole level was not artillery shelling of Beirut.

measured

in all artisanal

samples.

Remaining

samples

have

been destroyed

by

separation of methanol and acetaldehyde in the presence of excess of ethanol. Better results were obtained with Porapak Q for which high resolution of all constituents was observed (see Figs. 1 and 2). Identity of compounds was confirmed using head-space capillary gas-liquid chromatography. Ten milliliter standard and sample solutions were run, respectively, and different peaks retention times were compared. Anethole was rapidly separated and estimated using a glass column packed with 15% ethylene glycol adipate coated on Supelcoport. No interferences were observed from volatile constituents occurring in brandy distillates purchased from artisanal distillers before the final anisation step (Fig. 3). Methanol, Acetaldehyde, Ethyl Acetate, and Fuse1 Oil Components in Commercial and Artisanal Arak Samples The samples analyzed in this survey were randomly drawn from anisated brands which were commercially available in Lebanon from 1986 to 1989. One sample of each brand was analyzed. Nineteen other samples were purchased directly from local artisanal distillers producing anisated distilled spirits for their own use. The results of the quantitative determination of the various alcohols, ethyl acetate, and acetaldehyde in arak are presented in Tables 2 and 3. Methanol, acetaldehyde, ethyl acetate, and fuse1 oil constituents are absent in 10 commercial samples produced by manufacturers suspected of having used high proof alcohol for the anisation step. Methanol, acetaldehyde, and propanol are present in all remaining commercial and artisanal samples. Their levels vary from 9 to 540 mg/liter, 6 to 220 mg/liter, and 17 to 132 mg/liter, respectively, in commercial samples, and 141 to 1754 mg/liter, 34 to 4 14 mg/liter, and 16 to 2 14 mg/liter, respectively, in artisanal distillates. Ten com-

ANALYSIS

OF GRAPE DISTILLATES

BY GLC

313

mercial as well as nineteen artisanal samples showed methanol levels above the maximum allowable concentration (100 mg/liter) defined by the European Economic Community. Methanol and acetaldehyde levels are much higher in artisanal samples. Highly efficient rectifying columns used by commercial distillers reduce their levels appreciably as they are the most volatile components. Propanol is almost always present in commercial samples. One commercial sample showed a high propanol level (760 mg/liter) and absence of all other constituents, due probably to the use of highly contaminated industrial alcohol for the anisation step. Ethyl acetate was present in 15 commercial samples and was almost always accompanied by fuse1 oil components of grape distillates. Ethyl acetate, isobutanol, and amyl alcohols levels varied considerably between samples analyzed ranging from 18 to 870 mg/liter, 13 to 523 mg/liter, and 7 1 to 1692 mg/liter, respectively, in commercial samples, and 85 to 62 1 mg/liter, 40 to 1687 mg/liter, and 204 to 363 1 mg/liter, respectively, in artisanal samples. The differences in the fermentation conditions such as temperature and yeast strain, the distillation technique, and the grape varieties are the most important factors that control the levels of these alcohols. Artisanal and commercial samples obtained from fermented grape juice show the same analytical profile. One aspect of these useful parameters is their consistent presence in anisated distilled spirits obtained from fermented grape juice. The applications of these parameters are obvious in quality control and forensic situations. Tables 2 and 3 show that butanol has little application as a discriminating parameter. It is significant that it is found in only two commercial and one artisanal beverage at 111, 400, and 412 mg/ liter, respectively. The results of the quantitative determination of anethole in 30 commercial and 11 artisanal samples are presented in Tables 4 and 5. Anethole levels vary from 1.2 to 3.8 g/liter and 0.5 to 4.2 g/liter, respectively. Higher variation levels are observed in artisanal samples. It has been demonstrated in an earlier study that level variations of anethole are in direct relation with the amounts of aniseed used in the anisation step (unpublished data). Total Acidity Determination of total acidity was performed on commercial samples. Total acidity varied from 26 to 3400 mg/lOO ml expressed as acetic acid. No significant variations were observed between commercial samples produced from fermented grape juice and high proof industrial alcohol. No relationship was observed between proof and individual volatile components levels in beverage distillates. REFERENCES
ASKEW, B., AND LISLE, D. B. (197 1). Variation in the concentrations of higher alcohols, methanol and ethyl acetate in brandies. J. Sci. Food Agric. 22, 102-104. BARNETT, J. H., AND EINSMANN, J. R. (1977). Occurrence and distribution of 2-butanol, I-butanol, lpentanol and I-hexanol in distilled alcoholic beverages. J. Assoc. Ofl Anal. Chem. 60, 297-301. DAGHER, S. M., AND RUHAYEM, I. G. (1975). Fuse1 oil and methanol content of Lebanese arak. J. Food Sci. 40, 9 17-9 18. DYER, R. H. (1990). Beverages, distilled liquors. In Oficial Methods ofAnalysis (K. Hehich, Ed.), 15th ed. pp. 690-707. AOAC, Arlington, VA. GUYMON, J. F. (1970). Composition of California commercial brandy distillates. Am. J. Enol. Viti. 21,6 I69.

314
HERRAIZ, (1989). yeasts.

GEAHCHAN T., MARTIN-ALVAREZ, P., PEDRO, Differences between white fermented

ET

AL. M., AND CABEZUDO, dioxide using various gas-liquid M. D. selected

J., REGLERO, G., HERRAIZ, wine with and without sulfur

J. Sci. Food Agric. 49, 249-258.

study of the quantitative in whisky. J. Assoc. 08 M. D. (1988). chromato-

KAHN, J. H., AND BLESSINGER, E. T. (1972). Collaborative graphic determination of fuse1 oil and other components

Anal. Chem. 55, 549of Chilean Pisco. by from

556.
LOYOLA, E., ALMY, J., MARTIN-ALVAREZ, P., AND CABEZUDO, The flavor

Dev. Food Sci. 17, 729-742.

MARTIN, G. E., AND CARESS, E. A. (197 1). Ethyl acetate and fuse1 oil determination gas liquid chromatography and confirmation by mass spectrometry. J. Sci. Food ONISHI, M., Thompson PARKER, REGLERO, designed CROWELL, E. A., AND GUYMON, J. F. (1978). seedless and three white wine grape varieties. Gas-liquid chromatographic Comparative in distilled spirits

Agric. 22,587-589.
of brandies

composition

Am. J. Enol. Viti. 29, 54-59.

of anethole in liqueurs. micropacked

R. G. (1974).

determination

J. Assoc. Off:
columns

Anal. Chem. 51,954-956.

G., HERRAIZ, T., HERRAIZ, M., AND CABEZUDO, for selective separation of fermentation products. R. L., AND DYER, R. H. (1973). Analytical M. D. (1986). Mixed

SCHOENEMAN, SCHOENEMAN,

Chromatographia 22, 358-362. profile of Scotch whiskies. J. Assoc. 08 Anal.

profile of straight Bourbon whiskies.

Chem. 56, l-10.

R. L., DYER, R. H., AND EARL, E. M. (197 1). Analytical

J. Assoc. OfI Anal. Chem. 54, 1247-1261.