Volatile composition of tequila. Evaluation of three extraction methods Composición volátil del tequila. Evaluación de tres métodos extractivos

This article was downloaded by: [ITESM - Campus Monterrey], [Sandra Martín del Campo] On: 28 June 2011, At: 10:05 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

CyTA - Journal of Food Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tcyt20

Volatile composition of tequila. Evaluation of three extraction methods Composición volátil del tequila. Evaluación de tres métodos extractivos a

Sandra T. Martín-del-Campo Present address: Departamento de Industrias Alimentarias y Biotecnología, Instituto Tecnológico y des Estudios Superiores de Monterrey, Campus Querátaro. Epigmenio González 500, Fracc. San Pablo, Querétaro Qro., CP 76130, b

b

a

México , Héctor E. Gómez-Hernández , Humberto Gutiérrez , Héctor Escalona , a

Mirna Estarrón & Ricardo Cosío-Ramírez

a

a

Centro de Investigación y Asistencia en Tecnología y Diseño del Estado de Jalisco (CIATEJ, A.C.), Normalistas 800 Colinas de la Normal, Guadalajara, Jalisco, CP, 44270, México b

Centro Universitario de Ciencias Exactas e Ingeniería. Universidad de Guadalajara, Blvd. Marcelino García Barragán 1421, Guadalajara, Jalisco, CP, 44430, México Available online: 28 Jun 2011

To cite this article: Sandra T. Martín-del-Campo Present address: Departamento de Industrias Alimentarias y Biotecnología, Instituto Tecnológico y des Estudios Superiores de Monterrey, Campus Querátaro. Epigmenio González 500, Fracc. San Pablo, Querétaro Qro., CP 76130, México , Héctor E. Gómez-Hernández, Humberto Gutiérrez, Héctor Escalona, Mirna Estarrón & Ricardo Cosío-Ramírez (2011): Volatile composition of tequila. Evaluation of three extraction methods Composición volátil del tequila. Evaluación de tres métodos extractivos, CyTA - Journal of Food, 9:2, 152-159 To link to this article: http://dx.doi.org/10.1080/19476337.2010.499569

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan, sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

CyTA – Journal of Food Vol. 9, No. 2, August 2011, 152–159

Volatile composition of tequila. Evaluation of three extraction methods Composicio´n vola´til del tequila. Evaluacio´n de tres me´todos extractivos Sandra T. Martı´ n-del-Campoa*{, He´ctor E. Go´mez-Herna´ndezb, Humberto Gutie´rrezb, He´ctor Escalonaa, Mirna Estarro´na and Ricardo Cosı´ o-Ramı´ reza

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

a

Centro de Investigacio´n y Asistencia en Tecnologı´a y Disen˜o del Estado de Jalisco (CIATEJ, A.C.), Normalistas 800 Colinas de la Normal, Guadalajara, Jalisco CP 44270, Me´xico; bCentro Universitario de Ciencias Exactas e Ingenierı´a. Universidad de Guadalajara, Blvd. Marcelino Garcı´a Barraga´n 1421, Guadalajara, Jalisco CP 44430, Me´xico (Received 7 February 2010; final version received 28 May 2010) Extraction performance of the liquid–liquid batch (LLB) and continuous extraction, as well as simultaneous distillation–extraction were evaluated in order to analyze volatile compounds in tequila. Initially, the best extraction conditions were obtained for each method using tequila samples. Conditions tested were solvent type, extraction time, sample’s alcoholic concentration and, for the LLB, the number of successive extractions. Response variables were chromatograms, number of peaks, and total area. In a second phase, its extraction performance was evaluated with a model solution of 38 compounds. In both steps, obtained extracts were analyzed by gas chromatography. The best results were obtained with the LLB method since it makes possible to recover a good number of volatile compounds in few minutes. Additionally, this method showed the best performance after extraction and concentration, and the lower CV. Keywords: aroma; flavor; batch liquid–liquid; continuous liquid–liquid; simultaneous extraction–distillation; tequila Se evaluo´ el desempen˜o de tres me´todos extractivos (extraccio´n lı´ quido-lı´ quido por lotes y en continuo y extraccio´ndestilacio´n simulta´nea) seleccionados para analizar vola´tiles de tequila. Inicialmente se obtuvieron las mejores condiciones de extraccio´n para cada me´todo utilizando muestras de tequila. Las condiciones evaluadas fueron el tipo de solvente, tiempo de extraccio´n, concentracio´n alcoho´lica de la muestra y para el me´todo lı´ quido-lı´ quido por lote, nu´mero de extracciones sucesivas. Las variables de respuesta fueron el nu´mero de picos y el a´rea total en los cromatogramas. En una segunda etapa, se evaluo´ su desempen˜o de extraccio´n con una solucio´n modelo de 38 compuestos. En ambas fases, los extractos fueron analizados por cromatografı´ a de gases. Los mejores resultados se obtuvieron con el me´todo lı´ quido-lı´ quido por lote, ya que e´ste permite recuperar un buen nu´mero de compuestos en pocos minutos. Adicionalmente, este me´todo mostro´ el mejor desempen˜o despue´s de extraccio´n y concentracio´n ası´ como el menor CV. Palabras clave: aroma; sabor; extraccio´n lı´ quido-lı´ quido por lote; extraccio´n lı´ quido-lı´ quido en continuo; extraccio´ndestilacio´n simulta´nea; tequila

Introduction Mexico produces a wide array of fermented beverages from diverse agave species, such as tequila, mezcal, bacanora, and pulque, among others. Particularly, tequila has achieved high acceptance both in Mexico and the rest of the world. Tequila’s flavor and aroma are determined by their great diversity of volatile compounds. Some of these compounds show high concentration (major compounds), but most of the aroma compounds are present in very small concentrations (minor compounds). Tequila’s major volatile compounds are evaluated according to the Mexican law (NOM-006-SCFI-2005,

2006) and by some authors (Arrizon, Fiore, Acosta, Romano, & Gschaedler, 2006; Lachenmeier, Sohnius, Attig, & Lopez, 2006) by direct injection of tequila samples in a gas chromatographer. This method makes it possible to evaluate major volatile compounds such acetaldehyde, ethyl acetate, methanol, and higher alcohols. However, minor volatile compounds could not been evaluated by this method since their concentration is usually under the method’s detection limit. In order to evaluate these compounds, and to make an adequate characterization of them, it is necessary to increase their concentration by extracting them from their hydro alcoholic matrix. The election of an extractive method to separate these compounds has a

*Corresponding author. Email: [email protected] { Present address: Departamento de Industrias Alimentarias y Biotecnologı´ a, Instituto Tecnolo´gico y des Estudios Superiores de Monterrey, Campus Quera´taro. Epigmenio Gonza´lez 500, Fracc. San Pablo, Quere´taro Qro., CP 76130, Me´xico ISSN 1947-6337 print/ISSN 1947-6345 online Ó 2011 Taylor & Francis DOI: 10.1080/19476337.2010.499569 http://www.informaworld.com

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

CyTA – Journal of Food special importance since the profile of compounds obtained is directly correlated with the extractive method used (Bosch-Fuste et al., 2007; Caldeira, Rodrigues, Perestrelo, Marques, & Camara, 2007). There is a considerable amount of extractive methods that have been used for the characterization of the volatile fraction of alcoholic drinks, such as liquid–liquid in batch (Benn & Peppard, 1996; Pino, Villarreal, & Roncal, 1994; Sefton, Francis, & Williams, 1993) or continuous (Gonzalez-Vinas, PerezCoello, Salvador, Cabezudo, & Martin-Alvarez, 1996; Pino et al., 1994), adsorption–desorption (Girard, Kopp, Reynolds, & Cliff, 1997), headspace (Stashenko, Macku, & Shibamoto, 1992), microwave (Razungles et al., 1994), simultaneous distillation–extraction (Blanch, Reglero, & Herraiz, 1996; Blanch, Tabera, Herraiz, & Reglero, 1993; Schultz, Flath, Mon, Eggling, & Teranishi, 1977) and solid phase microextraction (SPME) (Caldeira et al., 2007; De LeonRodriguez, Gonzalez-Hernandez, De la Rosa, Escalante-Minakata, & Lopez, 2006; Pen˜a-Alvarez, Capella, Juarez, & Labastida, 2006). According to the method, several operational conditions are used, such as the solvent or fiber type, as well as the sample’s volume, and extraction time, etc. However, even though there are a wide variety of methods, the profile of volatile compounds obtained from each method is different since it depends not only just on the type of extraction (Caldeira et al., 2007; Jennings & Filsoof, 1977; Razungles, et al., 1994) but also on the selected solvents (Caldeira et al., 2007) and the extraction time and conditions. Additionally, the concentration method used to concentrate the obtained extract influences the final profile of volatile compounds (Hubert, Brunerie, Le Quere, & Drilleau, 1990; Langlois, Malterre, & Etievant, 1997). All these methods have delivered good results for the characterization of alcoholic drinks such as wines, whisky, cognac, and champagne, among others. However, there are few studies carried out for tequila’s volatile compounds evaluation (Bauer-Christoph et al., 2003; Benn & Peppard, 1996; Bluhm, 1983; De Leo´nRodrı´ guez et al., 2008; Manjarrez & Llama, 1969; Pen˜a-Alvarez et al., 2006) but those studies did not evaluate the effect of extractive methods in the volatile compounds profile obtained. The selection of the best extractive conditions for this beverage will make possible to characterize simultaneously compounds owing high, medium or low volatility. The high ethanol concentration in tequila, as it occurs in other spirits, interferes with the extraction of the volatile fraction (Pino et al., 1994); therefore, the choice of extractive methods have to be centered on those that allow a high recovery of the volatile fraction but a poor recovery of ethanol. Characterization of tequila volatile compounds profile requires an extractive method which represents as much as possible the original profile of the beverage. Modern methods such

153

as SPME has been used to evaluate selected volatile compounds in tequila samples (Pen˜a-Alvarez et al., 2006). This method is rapid, selective, and ecology friendly, but it only made it possible to evaluate a limited number of compounds. On the other hand, potent odorant compounds have been reported in tequila in very low concentration (Benn & Peppard, 1996). The aim of this work was to select a method that make it possible to characterize tequila’s volatile profile in order to evaluate differences due to process that may affect the aromatic quality of the product. Three methods were evaluated to obtain the operation conditions and to evaluate the performance of three extractive methods and select the one with the best extraction yield and efficiency in terms of extraction time. The methods evaluated were: liquid–liquid batch (LLB) extraction, liquid–liquid continuous (LLC) extraction, and the method described by Likens and Nickerson (1964) known as ‘‘simultaneous distillation– extraction’’ (SDE). Both LLB and LLC have been reported for the analysis of volatiles compounds in tequila (Aguilera Rojo, Martı´ n del Campo, Cosı´ o Ramirez, Escalona Buendia, & Estarro´n Espinosa, 2003; Benn & Peppard, 1996), while SDE has been used for other alcoholic beverages (Blanch et al., 1996). Materials and methods For this study, we selected three extractive methods which were evaluated in two steps. Initially, white tequila samples were extracted using different operation conditions in order to select the best conditions in terms of high extraction yields in shorter extraction times. Then, we evaluated the extraction methods performance by measuring the extraction yield and calculating the coefficient of variation (CV) by analyzing a model solution. Tequila samples In order to select the best extraction conditions, 44 subsample units from a 20 l homogeneous batch of white 100% agave tequila containing 55% (v/v) ethanol were obtained. The whole batch was distilled at a pilot-size still pot and preserved at 20 8C + 2 8C during the duration of the experiment. According to the experimental planning, the samples’ alcohol content was adjusted to 20% or 30% (v/v) ethanol by adding distilled water and were verified by using the GayLussac scale at 15 8C with calibrated alcoholmeters (Dujardin-Salleron, Paris). Model solution and solvents In order to evaluate the extraction yield of the three selected methods, the same model solution was used. This model solution was prepared with 38 reference

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

154

S.T. Martı´n-del-Campo et al.

standards with a minimum purity of 90% provided by Sigma-Aldrich (St. Louis, MO, USA). The reference compounds were selected among those previously identified in tequila (Benn & Peppard, 1996) that belonged to different chemical families and showing retention index all along the chromatogram according to the identification made in the first part of this work (Supplementary Table 1). The major volatile compounds evaluated by Mexican law (NOM-006-SCFI2005, 2006) were not chosen since they are evaluated by direct injection without extraction. Model solution contained 10 mg/ml of each one of the compounds in a 30% v/v ethanol/water solution (Lichrosolv, Merck, Darmstadt, Germany). This solution was kept under refrigeration at 0–5 8C until it was used. In order to evaluate the concentration procedure, 10 mg/l of ethyl butyrate (Sigma-Aldrich) was added as reference. Liquid–liquid batch (LLB) extraction For the initial evaluation of this method, four factors were considered: alcoholic content (20% and 30% v/v), extraction time (3 min and 5 min), number of extractions (1 6 45 ml or 3 6 15 ml) and the type of solvent (Pentane (Fisher, Leicester, U.K), Dichloromethane (Fischer) and a mixture of Pentane/Dichloromethane 3:1 v/v). Form this, the extraction of the 24 combination of the factor levels were randomly carried out. A sample of 325 ml, previously adjusted to the ethanol concentration, was extracted with the respective solvent during the selected time according to the experimental design. Extracts were dried with sodium sulfate (Mallinckrodt, Paris, USA) and preserved in amber flasks at 0–5 8C for subsequent concentration. In the second part, five replicates were carried out for this method. The model solution was extracted using the optimized conditions determined before. The resulting extracts were dehydrated with anhydrous sodium sulfate. An aliquot of 0.4 ml was taken from each of the extracts and were stored in tight closed vials under freezing at –40 8C without concentration until the time of analysis. The rest of the volume was kept in amber flasks at –40 8C until the time of its concentration. Liquid–liquid continuous extraction (LLC) For this method, three factors were evaluated: alcoholic content (20% and 30% v/v), extraction time (3 h and 6 h), and the type of solvent (pentane and pentane/ dichloromethane 3:1 v/v). A continuous liquid–liquid extraction apparatus for lighter-than-water solvents was used to perform the study. The eight level-factor combinations were randomly carried out. Samples of 1.250 l, previously adjusted for ethanol content, were extracted using 160 ml of solvent for 3 h or 6 h according to the experimental design. A coolant tempered to 0 8C was circulated through the condenser

during the whole extraction time using a recirculating bath. The extracts were dehydrated with anhydrous sodium sulfate and preserved in amber flasks at 0–5 8C for subsequent stages. In the second step for the evaluation of this method, three replicates were carried out. The model solution was extracted using the operation conditions previously selected. The resulting extracts were dehydrated with anhydrous sodium sulfate. From the three obtained extracts, an aliquot of 0.4 ml was not concentrated and stored in tightly closed vials under freezing at –40 8C until the time of analysis. The rest of the volume was kept in closed amber flasks at –40 8C until the time of its concentration. Simultaneous distillation–extraction (SDE) For this method, three factors were evaluated: alcoholic content (20% and 30% v/v), extraction time (1 h and 2 h), and type of solvent (pentane, dichloromethane or pentane/dichloromethane 3:1 v/v). The device for SDE described by Schultz et al. (1977) was used to carry out the eight factor-level combinations. The method is described by Blanch et al. (1996). Briefly, 50 ml samples, previously adjusted to certain alcohol content, were extracted with 50 ml of solvent for 1 h or 2 h. The extracts were dried with anhydrous ammonium sulfate and preserved at 0–5 8C for subsequent analyses. For this method, five replicates were carried out. The model solution was extracted using the optimized conditions selected before. The resulting extracts were dehydrated with anhydrous sodium sulfate. Nonconcentrated aliquots (0.4 ml) from three of the extracts were stored in tightly closed vials under frozen storage (740 8C) until the time of analysis. The rest of the volume was kept in the same amber flasks at –40 8C until the time of its concentration. Concentration All extracts were concentrated using an apparatus Kuderna-Danish. The equipment containing the whole extract (45, 160, and 50 ml for LLB, LLC, and SDE, respectively) was placed in a water bath at 40 8C. The final volume was adjusted to 0.4 ml with a nitrogen gas flow. The concentrated extracts were first placed in micro-vials and preserved in tight closed vials at 740 8C until their chromatographic analysis. Quantification Quantification was performed in the second part of this work. The external standard technique was used to quantify selected volatile compounds. Standard calibration curves for the same compounds of the model solution were obtained using successive dilutions to obtain concentrations that ranged from 50 to 500 mg/l.

CyTA – Journal of Food Calibration curves were considered adequate when they showed a high correlation (r 4 0.990 and r2 4 0.980).

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

Chromatographic analysis Tequila’s concentrated extracts were analyzed using a gas chromatography system (Hewlett Packard 6890, Palo Alto, USA) with flame ionization detection (FID). Volatile compounds were separated in a polar capillary column HP-20M (50 m 6 200 mm ID 6 0.10 mm) using helium as carrier gas (1.2 ml/min). The temperatures of the injector and detector were 230 8C and 250 8C, respectively. The injection volume was 0.5 ml and the split ratio was 60:1. Oven temperature was programmed starting at 50 8C for 5 min, then increasing to 210 8C at a rate of 2.5 8C/min and holding this maximum temperature during 40 additional minutes. In addition, the extracts obtained with the best operation conditions were analyzed by gas chromatography coupled with a mass selective detector (GC– MS). The chromatograms were obtained by using a gas chromatograph (Hewlett Packard 5890 Series II) coupled to a mass selective detector (HP 5972). Volatile compounds were separated in the same column used for GC–FID analysis. The injector temperature, sample volume, split relation, and oven program were also the same. The total ion chromatograms (TIC) and the mass spectra corresponding for each compound were acquired at an impact electron-ion (EI) at 70 eV, at 1.6 scans/s and using a mass range of m/z 30–350. The identification of the compounds was done by comparison of the compound spectrum with the Wiley 138 spectra library software installed in the computer of the chromatograph. The compound’s identity was confirmed in some cases by comparison of the mass spectra and retention times of reference standards and/ or by the compound retention index reported in the literature by Jennings and Shibamoto (1980) and Kovats.org (http://kovats.org/database/kovats/kovats-index.php). The model solution extracts and concentrated extracts were also analyzed by GC–MS with the same conditions described before in order to confirm the retention times and the identification of the compounds. Compounds were quantified via GC–FID with the conditions described before. Percentages of recovering and variation coefficients for each compound were calculated using the quantification data obtained for each one of the extracts. Statistical analysis The STATGRAPHICS software (Manugistics Inc., Rockville, USA) was used for statistical analysis. In the first part of this work, analysis of variance (ANOVA) was carried out in order to identify the best extraction

155

conditions of each evaluated method. Data of the total area and the total number of integrated peaks (after subtracting the solvent peaks); both computed when all chromatograms were obtained for each tequila extract, were used as response variables. Either a three- or fourfactor ANOVA (depending of the method) was carried out on the resulting data and followed by least significant difference (LSD) tests for each significant factor. LSD confidence intervals were calculated using the error mean square of each ANOVA. Significant differences were considered at p 5 0.05. Since these were screening experiments, just only one replicate was carried out. Results and discussion Selection of operation conditions Differences in the extracted compound profile were observed between the three extractive methods evaluated. Chromatograms obtained under the best extractive conditions are shown in Supplementary Figure 1. LLB (Supplementary Figure 1A) and LLC (Supplementary Figure 1B) methods made it possible to obtain more and higher peaks that SDE (Supplementary Figure 1C). For the LLB extraction method, the ANOVA showed that the effects of extraction time, number of extractions, and ethanol content were not significant (p 4 0.05) neither for the total area nor for the number of peaks. On the other hand, both solvent type and the interaction between solvent type and ethanol content showed significant effects (p 5 0.05) for both response variables. These results are similar to those obtained by Pino et al. (1994) who found same ethanol recoveries in model solutions containing between 10% and 30% (v/v) ethanol. Furthermore, they reported differences on compounds recovered according to the type of solvent. The LSD test for the solvent effect showed that dichloromethane alone or mixed with pentane had the best yields considering total areas under the peaks (Supplementary Table 1). Regarding the solvent effect on the number of peaks, the solvent mixture showed the best results (Supplementary Table 1). Additionally, the interaction solvent vs. ethanol content showed that the best results were obtained with the mixture of pentane/dichloromethane and 30% ethanol (v/v) for both, total area, and peak number (Supplementary Figure 2). Considering all these results, in addition to the observed non-significant differences for the rest of the factors, the selection of the optimal levels for this extractive method was based on the shorter time and the lower number of successive extractions. Therefore, the best extraction conditions were: one extraction during 3 min with a solvent mixture pentane/dichloromethane 3:1 (v/v), all this for a tequila sample adjusted to an ethanol content of 30% (v/v). These conditions made it possible the recovery of an average of 145 compounds eluted

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

156

S.T. Martı´n-del-Campo et al.

(Supplementary Figure 1A). On the other hand, there were some discrepancies with Pino et al. (1994) as they reported higher yields using pentane in comparison with dichloromethane. In this work, even though dichloromethane was not the best solvent, it showed higher yields than pentane considering both for total area and number of peaks. On the other hand, for the LLC extraction method, ANOVA showed that the effects of time and ethanol content were not statistically significant neither for the total area nor the number of peaks. On the other hand, the solvent type showed significant effects for both variables (p 5 0.05). Results are in accordance to Pino et al. (1994) who concluded that there were no differences on the recovery of compounds assayed in samples containing relatively lower ethanol contents (10–30% v/v). The non-significant difference between the two extraction times (3 h or 6 h) observed in this study differ from findings of Pino et al. (1994) who reported that increasing extraction times with pentane enhanced the recovery. The same authors determined 6 h as the optimum extraction time. The LSD test showed that the solvent mixture (Supplementary Table 1) was more effective both for total area and number of peaks. Interestingly, in the present work a high number of recovered compounds (Supplementary Figure 1B) were obtained, 146 in average. For the interaction between solvent type and alcoholic content (Supplementary Figure 3), the mixture pentane/dichloromethane showed the best results at 30% v/v ethanol for both response variables. According to these results, the best operation conditions were extracting a tequila sample adjusted to 30% (v/v) ethanol during 3 h using a 3:1 (v/ v) solvent mixture of pentane/dichloromethane. Finally, for the SDE method ANOVA showed that the extraction time and the ethanol content were not statistically significant neither for total area nor number of peaks. On the other hand, the effect of the type of solvent was significant different (p 5 0.05) for both variables. These results are in accordance to Cosı´ o and Rene´ (1996) who used the same extractive method for a model solution and obtained a similar recovery of compounds extracted for 1 h or 2 h. The LSD test for solvent type showed that dichloromethane and the solvent mixture yielded similar results both for total area and number of peaks (Supplementary Table 1). Also, the two solvent systems proved to be more efficient compared to pentane alone (Supplementary Table 1). The selection of the optimum operation conditions was based on the shorter extraction time and the less sample dilution (neither of both factors were significant) for the more efficient solvents. The behaviors of the interaction plots (Supplementary Figures 4 and 5) provided information to obtain these parameters. Even though both solvents showed similar results, the best conditions were using 3:1 pentane/ dichloromethane (v/v) as it required lower operation

time (1 h) and a less diluted test sample (30% v/v ethanol). These conditions made it possible the recovery of an average of 63 compounds eluted (Supplementary Figure 1C). It is worth mentioning that in the chromatograms obtained both from pentane/dichloromethane and dichloromethane alone there was the presence of a large peak identified as ethanol. Similar observations were previously reported by Blanch et al. (1996) who analyzed wine samples with the same method. As has been mentioned before, differences in the extracted compound profile were observed between the three extractive methods evaluated (Supplementary Figure 1). The quantity of identified compounds in extracts obtained under the best conditions was different for each method. Globally 129 compounds (Supplementary Table 2) were identified belonging to different chemical families. The extracts obtained with the LLB method showed the higher number of identified compounds (119 different compounds; Supplementary Figure 1A), followed by the LLC (Supplementary Figure 1B) method (113 compounds identified). The SDE method showed the lowest quantity of compounds (Supplementary Figure 1C), since only 58 compounds were detected. Most of the compounds extracted and identified in this work were reported before in tequila (Benn & Peppard, 1996; Pen˜a-Alvarez, et al., 2006; VallejoCordoba, Gonzalez-Cordova, & Estrada-Montoya, 2004). Benn and Peppard (1996) identified 187 different compounds by using liquid–liquid extraction of 1750 ml of tequila with 400 ml of dichloromethane. In this work, we identified a smaller amount of compounds by LLB but we also used five times the sample’s volume and almost nine times the solvent volume. Pen˜a-Alvarez et al. (2006) and VallejoCordoba, et al. (2004) evaluated different conditions for SPME methods in order to evaluate selected compounds in tequila. Pen˜a-Alvarez et al. (2006) evaluated terpenes by headspace SPME, while Vallejo-Cordoba et al. (2004) evaluated ethyl esters by SPME direct extraction. In those works only a chemical family could be evaluated each time so it only allows to obtain a partial volatile profile. The methods evaluated in our work made it possible to evaluate terpenes and esters simultaneously, as well as compound from other chemical families. For a more objective comparison of the three methods, the peak identified as ethanol was removed from the total area of the chromatograms obtained after SDE extraction so only the areas corresponding to volatiles were taken into consideration. Moreover, the total area obtained per each chromatogram, for the three extractive methods, was normalized dividing it by the sample volume in ml used for each extraction enabling a yield comparison among methods. Supplementary Table 3 shows a comparison of the parameters. The method LLC was considered as reference.

CyTA – Journal of Food

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

The method with the best yield expressed as area/ml of sample was SDE followed by LLB. The LLC method showed the lowest yield. On the other hand, the LLB method enabled the highest number of recovered compounds, followed by the LLC method. Therefore, the SDE method had the lowest recovery of volatile compounds. Extraction methods performances Once the model solution extracted compounds were identified, their retention times were coupled with the signal obtained by GC–FID. When some compounds coeluted using the selected GC-column, their recovery was considered altogether as the sum of their concentration. Concerning the linearity of the system, the minimum correlation was for propionic acid (r ¼ 0.997 and r2 ¼ 0.994) accomplishing the linearity criteria previously specified (r 4 0.990 and r2 4 0.980). The method that showed the lowest recovery percentage was LLC (Supplementary Table 4). None of the compounds showed recoveries above 40%. On the other hand, LLB and SDE showed recoveries up to 100% for most of the studied compounds. It is worth mentioning that none of the methods produced a concentration profile similar to the model solution (Supplementary Figure 6). Regarding the extraction for LLB, it was observed that nine compounds showed yields below 38% (X  1s); however, variation coefficients (CV) were lower than 10% for all the cases (Supplementary Table 4). The rest of compounds showed higher yields reaching up to 100% with CV below 10%. It is important to mention that ethyl propionate showed a higher CV (19.1%) probably because eluted at the same time as the solvent; therefore affecting the proper peak integration (Supplementary Table 4). On the other hand, there was a general decrease on the yield obtained after extraction (Supplementary Table 4). From the whole group of compounds, eight had yields lower than 24% (X  1s). In despite of this decrease, variation coefficients were low and similar to those of the extraction (CV 5 12%). Bornyl acetate and nerolidol were the compounds with the highest variation. Concentration profiles differed after concentrating extracts because compounds showed different recoveries when comparing with the profile of the direct extract. Some compounds showed an important reduction after concentration. Tetradecanol and carvacrol were not evaluated in concentrated extracts because they were not adequately separated, while tetradecanoic acid did not show confident results in the concentrated fractions. Recovery results differed from those previously reported by Pino et al. (1994) probably because the solvent mixture used herein was different. In general, higher yields were observed for 1butanol (10.1% with pentane, 31.2% with the mixture

157

pentane-ethyl ether 2:1 v/v and 6.4% with dichloromethane) and eugenol (49.3% with pentane, 50.3% with the mixture pentane-ethyl ether 2:1 v/v, and 44.7% with dichloromethane). On the other hand, lower yields than those reported by Pino et al. (1994) were observed for phenyl ethanol (46.3% with pentane, 43.4% with the mixture pentane-ethyl ether 2:1 v/v and 45.3% with dichloromethane) and ethyl propionate (60.3% with pentane, 50.3% with pentane-ethyl ether 2:1 v/v and 46.2% with dichloromethane). It is important to remark that yields of isoamyl alcohol varied according to the solvent used. Pino et al. (1994) obtained higher yields in some cases (2.0% with the mixture of pentane-ethyl ether 2:1 v/v and 41.3% with dichloromethane). Finally, the yields obtained for benzaldehyde were similar to those reported by Pino et al. (1994). Regarding the coefficients of variation obtained in the extraction, there were some differences to those reported by Ortega-Heras, Gonzalez-SanJose, and Beltran (2002), who reported less variation for most of the compounds except for butanol (5.94%), and ethyl decanoate (1.84%) which showed similar variation, and for benzyl alcohol (10.35%), and Eugenol (8.62%) which showed higher variation. The lower variation could be attributed to the cooled system used by the authors to perform the extraction. On the other hand, for LLC none of the recovered compounds showed yields higher than 40% for the non-concentrated extracts (Supplementary Table 4). Only seven compounds had yields below 19%. Regarding the variation coefficient, 21 compounds had CV higher than 10%. Tetradecanoic acid with a CV of 31% had the highest variation. It is important to mention that this recovering was not evaluated after concentration because the chromatographic method did not provide a good resolution due to the high quantities of concentrated compounds. For this reason, only the results of the non-concentrated extracts are shown. Results on LLC showed different yields to those reported by Pino et al. (1994), who used pentane as solvent. Our results showed higher yields for 1-butanol (14–22.4%), eugenol (46.3–60.0%), and benzaldehyde (51.0–65.2%) and lower yields for isoamyl alcohol (47.5–60.3%), phenyl ethanol (48.3– 61%) and ethyl propionate (55.2–65%). It is important to mention that these compounds had a similar behavior when extracted with the LLB procedure. As observed with the other extraction procedures, for SDE, each compound had a different yield after SDE extraction (Supplementary Table 4). Four of the compounds were not evaluated because their concentrations were either below the detection limit or showed inconclusive results. Considering compounds that were successfully evaluated (Supplementary Table 4), only seven showed recoveries below 56% (X  1s), the rest showed recoveries up to 100%. On the other hand, six of the compounds showed high variation coefficients (CV 4 10%), from which ethyl decanaoate and phenyl

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

158

S.T. Martı´n-del-Campo et al.

ethanol had the highest variation (42% and 30%, respectively). During the concentration, there was an important reduction in yields and an increase on the variation. Additionally, some of the non-evaluated compounds after extraction were susceptible of evaluation after concentration since their concentration was higher than the detection limit. Considering all the evaluated compounds, four showed yields below 13.8% (X  1s) and ethyl propionate could not be detected. Regarding the variation, only toluene had a CV 5 10%, and the rest showed CV’s in between 10% and 25%. Results showed higher extraction yields for ethyl hexanoate, similar to those obtained by Schultz et al. (1977) who reported a 100% recovery using hexane or pentane. Blanch et al. (1993) also reported yields of 100% and 90.7% with dichloromethane and pentane, respectively. None of these works reported coefficients of variation. However, after concentration yields were strongly reduced (21.5%) increasing the CV (from CV ¼ 7.66 to 22.79%). Regarding hexanol, the yield was 78.9% below the yield reported by Schultz et al. (1977), who reported a 100% of recovery using either hexane or pentane. Furthermore, Blanch et al. (1993) obtained yields of 100% and 73% with dichloromethane and pentane, respectively. For benzaldehyde, yields (46.98%) were lower than those obtained by Blanch et al. (1993), who recovered 100% with dichloromethane and 87% with pentane. Regarding phenyl ethanol, results showed an average yield of 38.0%, which is an intermediate value among the yields previously reported by Blanch et al. (1993) using dichloromethane (55%) and pentane (11.26%). Furthermore, it was not possible to evaluate linalool because coeluted with 5-methyl furaldehyde. Nevertheless, the sum of both showed extraction yields of 92%, similar to those reported by Schultz et al. (1977) just for linalool (73–100%) and by Blanch et al. (1993) who reported a recovery 100% with dichloromethane and 90% with pentane. It is worth mentioning that, in general for this work, results showed yields around 100%, similar to those previously reported by other authors (Blanch et al., 1993; Likens & Nickerson, 1964; Schultz et al., 1977). The three methods (LLB, LLC, and SDE) were successfully compared in the extraction stage. However, after concentration this comparison was possible only for LLB and SDE. The results of LLC were not reliable. Table 3 shows a general comparative among the three methods. Even though the three methods showed similar coefficients of variation after extraction, the method LLC had much lower yields compared to the other methods (value of 26.9%). After concentration, the LLB method showed higher yields with lower CV compared to the SDE procedure. Conclusions The best operation conditions, under the selected experimental conditions, were obtained for three

extractive methods of the volatile fraction of tequila. Thus, the methods with the best recovery per ml of sample were LLB and SDE. Additionally, the yields of three extractive methods were evaluated using a model solution. The methods that allowed the best yields were LLB (69.00% average) and SDE (78.27% average). Nevertheless, SDE showed a higher total recovery than LLB, it showed the highest variability (6.46% CV for LLB and 9.09% CV for SDE). After concentration, the method with higher yields (55.26% average) and less variability (CV 9.95%) was the LLB. Furthermore, this method was the one that required less extraction time and had also the advantage of requiring a smaller sample volume. However, further studies are needed in order to evaluate the sensitivity of this method to detect differences in real tequila’s volatile profile due to process differences. Supplementary material The supplementary material for this article is available online at http://dx.doi.org/10.1080/19476337.2010. 499569 Acknowledgment The authors thank CONACyT-Mexico through the ‘‘Sistema Jose´ Ma. Morelos (SIMORELOS)’’ for the financial support to the project Ref. 96-05-004 carried out at CIATEJ A.C.

References Aguilera Rojo, D.K., Martı´ n del Campo, S.T., Cosı´ o Ramirez, R., Escalona Buendia, H., & Estarro´n Espinosa, M. (2003). Identification of distinctive parameters between Tequila mixto and Tequila 100% agave by gas chromatography. In J.L. LeQuere & P.X. Etievant (Eds.), Flavour research at the dawn of the twenty-first century (pp. 761–764). Paris, France: Lavoisier Tec. & Doc/ Intercept. Arrizon, J., Fiore, C., Acosta, G., Romano, P., & Gschaedler, A. (2006). Fermentation behaviour and volatile compound production by agave and grape must yeasts in high sugar Agave tequilana and grape must fermentations. Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology, 89(1), 181–189. doi: 10.1007/s10482-005-9022-1. Bauer-Christoph, C., Christoph, N., Aguilar-Cisneros, B.O., Lopez, M.G., Richling, E., Rossmann, A., & Schreier, P. (2003). Authentication of tequila by gas chromatography and stable isotope ratio analyses. European Food Research and Technology, 217(5), 438–443. Benn, S.M., & Peppard, T.L. (1996). Characterization of tequila flavor by instrumental and sensory analysis. Journal of Agricultural and Food Chemistry, 44(2), 557– 566. doi: 10.1021/jf9504172. Blanch, G.P., Reglero, G., & Herraiz, M. (1996). Rapid extraction of wine aroma compounds using a new simultaneous distillation-solvent extraction device. Food Chemistry, 56(4), 439–444. doi: 10.1016/0308-8146(95) 00214-6. Blanch, G.P., Tabera, J., Herraiz, M., & Reglero, G. (1993). Preconcentration of volatile components of foods: optimization of the steam distillation-solvent extraction at normal pressure. Journal of Chromatography A, 628(2), 261–268. doi: 10.1016/0021-9673(93)80009-W.

Downloaded by [ITESM - Campus Monterrey], [Sandra Martín del Campo] at 10:05 28 June 2011

CyTA – Journal of Food Bluhm, L. (1983). Distilled beverages. In G. R (Ed.), Biotechnology, vol. 5 – food and feed production with microorganisms (pp. 447–476). Weinheim, Germany: Verlag Chemie. Bosch-Fuste, J., Riu-Aumatell, M., Guadayol, J.M., Calxach, J., Lopez-Tamamaes, E., & Buxaderas, S. (2007). Volatile profiles of sparkling wines obtained by three extraction methods and gas chromatography mass spectrometry (GC-MS) analysis. Food Chemistry, 105(1), 428–435. Caldeira, M., Rodrigues, F., Perestrelo, R., Marques, J.C., & Camara, J.S. (2007). Comparison of two extraction methods for evaluation of volatile constituents patterns in commercial whiskeys Elucidation of the main odouractive compounds. Talanta, 74(1), 78–90. doi: 10.1016/ j.talanta.2007.05.029. Cosı´ o, R., & Rene´, F. (1996). Volatile compounds from banana. I. Comparative study of two extraction methods. Sciences des aliments, 16(4), 383–392. De Leo´n-Rodrı´ guez, A., Escalante-Minakata, P., Jime´nezGarcı´ a, M.I., Ordon˜ez-Acevedo, L.G., Flores Flores, J.L., & Barba de la Rosa, A.P. (2008). Characterization of volatile compounds from ethnic agave alcoholic beverages by gas chromatography-mass spectrometry. Food Technology and Biotechnology, 46(4), 448–455. De Leon-Rodriguez, A., Gonzalez-Hernandez, L., De la Rosa, A.P.B., Escalante-Minakata, P., & Lopez, M.G. (2006). Characterization of volatile compounds of mezcal, an ethnic alcoholic beverage obtained from Agave salmiana. Journal of Agricultural and Food Chemistry, 54(4), 1337– 1341. Girard, B., Kopp, T.G., Reynolds, A.G., & Cliff, M. (1997). Influence of vinification treatments on aroma constituents and sensory descriptors of Pinot noir wines. American Journal of Enology and Viticulture, 48(2), 198–206. Gonzalez-Vinas, M.A., Perez-Coello, M.S., Salvador, M.D., Cabezudo, M.D., & Martin-Alvarez, P.J. (1996). Changes in gas-chromatographic volatiles of young Airen wines during bottle storage. Food Chemistry, 56(4), 399– 403. doi: 10.1016/0308-8146(95)00207-3. Hubert, C., Brunerie, P., Le Quere, J.M., & Drilleau, J.F. (1990). Volatile aroma compounds of ciders: rapid extraction and analysis. Sciences des aliments, 10(3), 603–618. Jennings, W., & Shibamoto, T. (1980). Quantitative analysis of flavor and fragance volatiles by glass capillary gas chromatography. New York: Academic Press. Jennings, W.G., & Filsoof, M. (1977). Comparison of sample preparation techniques for gas chromatographic analysis. Journal of Agricultural and Food Chemistry, 25(3), 440– 445. doi: 10.1021/jf60211a027. Lachenmeier, D.W., Sohnius, E.M., Attig, R., & Lopez, M.G. (2006). Quantification of selected volatile constituents and anions in Mexican Agave spirits (Tequila, Mezcal, Sotol, Bacanora). Journal of Agricultural and Food Chemistry, 54(11), 3911–3915.

159

Langlois, D., Malterre, M., & Etievant, P. (1997). Cryoconcentration of volatile compounds in water: a study on a model aqueous solution. Sciences des aliments, 17(1), 69– 78. Likens, S.T., & Nickerson, G.B. (1964). Detection of certain hop oil constituents in brewing products. American Society of Brewing Chemist, Proceedings, 22, 5–13. Manjarrez, A., & Llama, M. (1969). Quantification of the volatile components of tequilas and mescals by Gas Chromatography. Revista de la Sociedad Quı´mica de Me´xico, 13, 1–5. NOM-006-SCFI-2005, 2006. Norma oficial mexicana. Bebidas alcoho´licas-tequila-especificaciones. Diaro Oficial de la Federacio´n (pp. 6–23). Me´xico. Ortega-Heras, M., Gonzalez-SanJose, M.L., & Beltran, S. (2002). Aroma composition of wine studied by different extraction methods. Analytica Chimica Acta, 458(1), 85– 93. doi: 10.1016/S0003-2670(01)01526-4. Pen˜a-Alvarez, A., Capella, S., Juarez, R., & Labastida, C. (2006). Determination of terpenes in tequila by solid phase microextraction-gas chromatography-mass spectrometry. Journal of Chromatography A, 1134(1–2), 291– 297. Pino, J., Villarreal, A., & Roncal, E. (1994). Comparison of extraction methods for the isolation of volatile compounds from distilled alcoholic beverages. Food/Nahrung, 38(3), 307–310. doi: 10.1002/food.19940380310. Razungles, A., Tarhi, E.H., Baumes, R., Gunata, Z., Tapiero, C., & Bayonove, C. (1994). Rapid analysis of volatile compounds of grapes and wines by microwave extraction. Sciences des aliments, 14, 725–739. Schultz, T.H., Flath, R.A., Mon, T.R., Eggling, S.B., & Teranishi, R. (1977). Isolation of volatile components from a model system. Journal of Agricultural and Food Chemistry, 25(3), 446–449. doi: 10.1021/jf60211a038. Sefton, M.A., Francis, I.L., & Williams, P.J. (1993). The volatile composition of chardonnay juices: A study by flavor precursor analysis. American Journal of Enology and Viticulture, 44(4), 359–370. Stashenko, H., Macku, C., & Shibamoto, T. (1992). Monitoring volatile chemicals formed from must during yeast fermentation. Journal of Agricultural and Food Chemistry, 40(11), 2257–2259. doi: 10.1021/jf00023a040. Vallejo-Cordoba, B., Gonzalez-Cordova, A.F., & EstradaMontoya, M.d.C. (2004). Tequila volatile characterization and ethyl ester determination by solid phase microextraction gas chromatography/mass spectrometry analysis. Journal of Agricultural and Food Chemistry, 52(18), 5567–5571.