Food Chemistry 125 (2011) Contents lists available at ScienceDirect. Food Chemistry. journal homepage:

Transcription

1 Food Chemistry 125 (211) Contents lists available at ScienceDirect Food Chemistry journal homepage: nalytical Methods Using headspace solid phase micro-extraction for analysis of aromatic compounds during alcoholic fermentation of red wine Mingxia Zhang a,b, Qiuhong Pan a, Guoliang Yan a, Changqing Duan a, a Center for Viticulture and Enology, College of Food Science & Nutritional Engineering, China griculture University, PO ox 31, Qinghua Donglu 17, eijing 183, China b Henan Institute of Science and Technology, Xinxiang 4533, Henan Province, China article info abstract rticle history: Received 27 pril 21 Received in revised form 9 July 21 ccepted 1 September 21 Keywords: HS-SPME romatic compounds Quantitative determination lcoholic fermentation Timely monitoring of changes in the type and quantity of aromatic compounds throughout the must fermentation process provides useful information for wine makers. This paper aimed to use headspace solid phase micro-extraction coupled with gas chromatography (HS-SPME/GC MS) to analyse aromatic compounds produced during must alcoholic fermentation. The results showed that both qualitatively and quantitatively, the PDMS/CR/DV fibre was suitable for extracting aromatic compounds in wine. The amount of alcohols, esters, acids and monoterpenes absorbed on the SPME fibre were influenced by the ethanol content. Calibration curves with a high correlation (R2 >.9) obtained in seven ethanol contents (2%, 4%, 6%, 8%, 1%, 12% and 14%) were established to quantitatively determine the amount of aromatic compounds during alcoholic fermentation. validated HS-SPME method for determining aromatic compounds was used to monitor aromatic compounds during Syrah alcoholic fermentation. This modified HS-SPME method was proved to be useful for controlling the oenological process. Ó 21 Elsevier Ltd. ll rights reserved. 1. Introduction The diversity of aromatic compounds in wine is immense and ranges in concentration from several mg l 1 to a few ng l 1. Some of the aromatic compounds come directly from the grapes while others are formed during fermentation and ageing (Rapp, 1998). For the neutral grape variety, the aromatic compounds formed during alcoholic fermentation play an important role in the total aroma of the finished wine (Vianna & Ebeler, 21). In this respect, oenological practices including cold maceration, enzymatic treatments and the type of yeast used may modify the final aroma. Timely monitoring of changes in these compounds throughout the must fermentation process provides useful information for the wine maker (Rapp, 1998). Thereby, it is important to select a valid method to analyse the composition of aromatic compounds in wine in order to ensure accurate monitoring of the process. HS-SPME, developed on the basis of solid extraction, is a sensitive and powerful method for the identification of organic constituents in complex matrices (Hayasaka, MaNamara, aldock, Taylor, & Pollnitz, 23). HS-SPME provides many advantages over conventional sample preparation techniques. Simplicity, speediness, solvent-free extraction and minimal sample manipulation are amongst the advantages offered by this technique (egala, Corda, Podda, Fedrigo, & Traldi, 22; osch-fustéa et al., 27; Martí, Mestres, Sala, usto, & Guasch, 23; Zhang, Xu, Duan, Qu, & Wu, Corresponding author. Tel./fax: address: chqduan@yahoo.com.cn (C. Duan). 27a). This technique permits sampling, clean-up, and concentration in the same step (osch-fustéa et al., 27). HS-SPME coupled with GC or GC/MS has been widely applied to analyse and monitor the aroma of grapes and wines (ntalick, Perello, & de Revel, 21; Howard, Mike, & Riesen, 25; Lee, Rathbone, simont, dden, & Ebeler, 24; Vianna & Ebeler, 21; Zhang et al., 27a). Vianna and Ebeler (21) and Lee et al. (24) used HS-SPME coupled with GC/MS to monitor the formation of esters during alcoholic fermentation, and they provided evidence that HS-SPME coupled with GC/ MS could provide greater insight into yeast metabolism and flavour formation. Mallouchos, Komaitts, Koutinas, and Kanellaki (22) investigated the evolution of volatile components during the alcoholic fermentation of grape must using free and immobilized cells with the help of HS-SPME. The ability of the SPME fibre to absorb aromatic compounds was affected by the coat immobilized on the fibre, extraction conditions and the composition of the wine, and different coats were suited to absorb different aromatic compounds (ntalick et al., 21; egala et al., 22; Howard et al., 25). Many authors have optimised extraction time and temperature, sample volume, sample ionic strength and other operating conditions (egala et al., 22; de la Calle García, Magnaghi, Reichenbächer, & Danzer, 1996; Whiton & Zoecklein, 2). ecause ethanol is the main component next to water in wine, the effect of ethanol on the aromatic components needs to be taken into account when analysing aromatic compounds in wine (Conner, irkmyre, Paterson, & Piggott, 1998; Rocha, Ramalheira, arros, Delgadillo, & Coimbra, 21; Whiton & Zoecklein, 2). de la Calle García /$ - see front matter Ó 21 Elsevier Ltd. ll rights reserved. doi:1.116/j.foodchem.219.8

2 744 M. Zhang et al. / Food Chemistry 125 (211) et al. (1996) investigated the effect of ethanol on monoterpenes. They found that as the ethanol content increased, the amount of monoterpenes extracted decreased while no significant effect was seen with alcohols and fatty acids (Lee et al., 24; Vas, lechschmidt, Kovacs, & Vekey, 1999; Vaz Freire, Costa Freitas, & Relva, 21). The effect of ethanol on the SPME fibre has mainly focused on monoterpenes and esters (de la Calle García et al., 1996; Rodríguez-encomo, Conde, Rodríguez-Delgado, García- Montelongo, & Pérez-Trujillo, 22). The impact of ethanol on the absorption by the SPME fibre of aromatic compounds is considered to result from the individual characteristics of each compound, such as its molecular weight, boiling point, molecular structure, solubility in the liquid matrix, and affinity to the fibre coating (ntalick et al., 21; Conner et al., 1998; de la Calle García et al., 1996; Rocha et al., 21). It is necessary to systematically analyse the effect of ethanol on the HS-SPME fibre for accurate quantification of aroma compounds in model wine solutions. Compared to an internal standard method, an external standard method shows good accuracy and higher confidence (Rodríguez- encomo et al., 22; Vaz Freire et al., 21). In our experiment, we used standard compounds to prepare seven sets of synthetic wines containing 7 g l 1 tartaric acid, ph adjusted to 3.3 with 1 M NaOH and their ethanol contents were %, 2%, 4%, 6%, 8%, 1%, 12% (v/v) for quantity aromatic compounds in samples during alcoholic fermentation. The goal of the present study is to establish a dependence method for quantitative determination of aromatic compounds during must alcoholic fermentation by optimising the SPME fibre and minimising the effects of ethanol on the HS-SPME. Then once optimised, the method will be tested for its ability to monitor aromatic compounds during Syrah must alcoholic fermentation. 2. Material and methods 2.1. Winemaking Syrah grapes were harvested from Changli, Hebei Province (China) on September 27, 26, with the following initial characteristics: ph 3.3, total acidity 6.5 g l 1 (tartaric acid) and 19.5 rix. The harvested grapes were quickly destemmed and crushed, and 6 mg/l SO2 were added. Cold maceration was carried out for 24 h at 5 C. The hydrated yeast was inoculated at 2 mg l 1 (Fericru VR5, DSM) after cold maceration. The fermentation temperature was maintained below 25 C. Fermentation was monitored by measuring the sample specific density from the inoculation moment and every 12 h until a constant specific density was reached. fter seven days, the wines were dry (reducing sugar <4 g l 1 ), but the wines continued to remain in contact with grape skins for seven more days. fter the alcoholic fermentation, the wines were racked. During fermentation, Syrah must in six.5 m 3 -stainless tanks was sampled every 24 h, and aromatic compound analysis was performed Preparation of model solutions Standards were purchased from ldrich and Fluka, purity P 99%. Model wine solutions were prepared using the methods reported by Rodríguez-encomo et al. (22) with minor modifications. fter each family of aromatic compounds was precisely weighed, they were completely dissolved in 1 ml ethanol and stored at 4 C as stock solutions. For analysis, all stock solutions were mixed then diluted to the appropriate concentrations and the content of ethanol was adjusted to 2%, 4%, 6%, 8%, 1%, 12%, 14% (v/ v) to constitute the model solutions. Each set of model solutions was prepared in triplicate. Seven sets of synthetic solutions were prepared and contained 7 g l 1 tartaric acid, ph adjusted to 3.3 with 1 M NaOH and their ethanol were %, 2%, 4%, 6%, 8%, 1%, 12% (v/v). The concentrations of standards in the model wine solutions are isoamyl alcohol 588 mg l 1, hexanol 316. mg l 1, E-3- hexen-1-ol 51.2 mg l 1, Z-3-hexen-1-ol 46.4 mg l 1, E-2-hexen-1- ol 42.4 mg l 1, Z-2-hexen-1-ol 46.4 mg l 1, octanol 18. mg l 1, 2,3-butandiol 118. mg l 1, methionol 116. mg l 1, decanol.9 mg l 1, benzyl alcohol 152. mg l 1, phenyl ethanol 22 mg l 1, isoamyl acetate 31.8 mg l 1, ethyl hexanoate 22.6 mg l 1, hexyl acetate.9 mg l 1, ethyl octanoate 74. mg l 1, ethyl decanoate 8 mg l 1, diethyl succinate 21. mg l 1, phenylethyl acetate 1.1 mg l 1, acetic acid mg l 1, hexanoic acid 356. mg l 1, octanoic acid 392. mg l 1, decanoic acid 39.2 mg l 1, linalool 1. mg l 1, terpinen-4-ol 1.9 mg l 1, citronellol 1.1 mg l 1, nerol 1.2 mg l 1, furfural 1.1 mg l 1, cis-oak lactone 1.1 mg l 1 and 4- ethyl-phenol 1.6 mg l 1, respectively Head space solid phase micro-extraction romatic compounds from the wine samples were extracted by HS-SPME and analysed using gas chromatography/mass spectrometry as described by Zhang, Xu, Duan, Qu, and Wu (27a). Five millilitres of wine sample and 1 g NaCl were placed in a 15-ml sample vial. The vial was tightly capped with a PTFE-silicon septum and heated at 4 C for 3 min on a heating platform with agitation at 4 rpm. The SPME (85 lm polyacrylate (P), 1 lm polydimethylsiloxane (PDMS) and 5/3 lm Divinylbenzene/ Carboxen/Polydimethylsiloxane (DV/CR/PDMS), Supelco, ellefonte, P, US), preconditioned according to manufacturer s instructions, was then inserted into the headspace, where the extraction lasted for 3 min with heating at 4 C and agitation by a magnetic stirrer. The fibre was subsequently desorbed in the GC injector for 25 min Gas chromatography/mass spectrometry (GC/MS) analysis The GC MS system used was an gilent 689 GC equipped with an gilent 5975 mass spectrometer. The column was a 6 m.25 mm HP-INNOWX capillary with.25 lm film thickness (J & W Scientific, Folsom, C, US). The carrier gas was helium at a flow rate of 1 ml min 1. Samples were injected by placing the SPME fibre at the GC inlet for 25 min in the splitless mode. The oven s starting temperature was 5 C, which was held for 1 min, then raised to 22 C at a rate of 3 C min 1 and held at 22 C for 5 min. The mass spectrometer in the electron impact mode (MS/EI) at 7 ev was scanned in the range of m/z The mass spectrometer was operated in the full scan and the selective ion mode (SIM) under autotune conditions at the same time. The area of each peak was determined by ChemStation software (gilent Technologies). nalyses were carried out in triplicate Statistical analysis One-way NOV was used to evaluate differences in the SIM peak area of each aromatic compound in solutions of various percentages of ethanol. significant difference was noted when p < 5. SPSS version 11.5 Statistical Package for Windows was used for all statistical analyses. 3. Results and discussion 3.1. Selection of HS-SPME fibres The effectiveness of HS-SPME for the quantification of aromatic compounds depends on the fibre coating (ntalick et al., 21;

3 M. Zhang et al. / Food Chemistry 125 (211) Table 1 Comparison of the peak areas for aromatic compounds absorbed by three HS-SPME fibres in model wine. Compounds PDMS/CR/DV PDMS P lcohols Isoamyl alcohol 1.15E E E + 8 Hexanol 1.91E E E + 8 E-3-Hexen-1-ol 1.77E E E + 7 Z-3-Hexen-1-ol 1.65E E E + 7 E-2-Hexen-1-ol 2.15E E E + 7 Z-2-Hexen-1-ol 1.84E E E + 7 Octanol 2.57E E E + 8 2,3-utandiol 6.12E E E + 7 Methionol 1.72E E E + 6 Decanol 5.44E E E + 6 enzyl alcohol 1.58E E E + 7 Phenyl ethanol 1.93E E E + 8 Esters Isoamyl acetate 5.2E E E + 7 Ethyl hexanoate 4.51E E E + 6 Hexyl acetate 5.44E E E + 6 Ethyl octanoate 1.85E E E + 6 Ethyl decanoate 3.98E E E + 6 Diethyl succinate 1.9E E E + 6 Phenylethyl acetate 9.85E E E + 6 cids Hexanoic acid 1.4E E E + 7 Octanoic acid 1.27E E E + 7 Decanoic acid 1.38E E + 5 Terpenes Linalool 5.3E E E + 6 Terpinen-4-ol 5.61E E E + 6 Citronellol 3.62E E E + 6 Nerol 1.76E E E + 5 Others Furfural 1.15E + 6 cis-oak lactone 2.95E Ethyl-phenol 5.4E + 5 egala et al., 22; Howard et al., 25). Therefore, it was first necessary to perform a preliminary selection of which fibre would be suitable for analysing aromatic compounds in must-wine. Twentynine standards of aromatic compounds were selected to analyse the performance of three fibres: PDMS, P and PDMS/CR/DV. These standards were dissolved in model solution with 1% ethanol. The peak areas of compounds which were absorbed by these three fibres and analysed by GC/MS with a SIM model were compared (Table 1). It was determined that P absorbed higher amounts of polar compounds such as acids, alcohols and monoterpenes than PDMS. The apolar esters bound poorly to P but more strongly to PDMS. Compared to P and PDMS, PDMS/CR/DV fibre qualitatively and quantitatively absorbed more wine aromatic compounds. PDMS/CR/DV absorbed 28 compounds while PDMS and P absorbed 25 and 24 compounds, respectively. In the experiments which extracted the same compound using three different fibres, the greatest amount of this compound was obtained in the extraction by the PDMS/CR/DV fibre. Therefore, the PDMS/ CR/DV fibre was chosen as the most suitable for this study Effects of ethanol on the HS-SPME fibre to absorb aromatic compounds PDMS/CR/DV was used to extract aromatic compounds from seven sets of model solutions containing different amounts of ethanol (2 14%, v/v) in order to study the impact of ethanol on the fibre to absorb aromatic compounds. In order to correct errors produced by manual sampling and the aged fibre, 4-methyl-2- pentanol as an internal standard was added to analytical samples prior to extraction by the HS-SPME fibre (egala et al., 22). dditionally, the internal standard was added to determine whether there would be similar effects by changing the concentrations of ethanol on the extraction of HS-SPME fibre between the internal standard and the aromatic compounds lcohols The peak areas of alcohols extracted by PDMS/CR/DV in the model solutions with different ethanol contents are shown in Fig. 1. s was observed for the analogous alcohols, the effects of ethanol on the extraction of alcohols were similar. The internal standard of 4-methyl-2-pentanol gradually decreased with ethanol increase from 2% to 1%. No significant change in the amount of 4-methyl-2-pentanol was observed with the continuous rise of ethanol concentration. different change existed between internal standard and alcohols analysed with ethanol. The amount of isoamyl alcohol, a branched-chain alcohol, had no significant changes in solutions of ethanol from 2% to 6%, however, it significantly increased as the ethanol content rose from 6% to 12%, and then significantly decreased in the presence of 12% to 14% ethanol (p < 5). E-3-hexen-1-ol, Z-3-hexen-1-ol, E-2-hexen-1-ol and Z-2-hexen-1-ol, all unsaturated alcohols, followed very similar patterns of change in different ethanol contents (Fig. 1C). The extraction of these four alcohols only varied slightly as the ethanol content increased from 2% to 8%, while from 8% to 12%, they significantly increased (p < 5), then decreased as the ethanol content rose from 12% to 14% (p < 5). The extraction pattern of benzyl alcohol and phenyl ethanol was nearly identical. Their amounts gradually reduced as ethanol increased, especially in the range of ethanol from 8% to 12%. Compared to the other alcohols tested, methionol showed a low response to MS analysis and no significant change in extraction efficiency was noted with changes in ethanol content (Rocha et al., 21). The amounts of octanol and decanol decreased as the ethanol content increased, especially as ethanol rose from 6% to 14% (p < 5) (Fig. 1D). Hexanol displayed a similar extraction pattern to the four unsaturated alcohols shown in Fig. 1C, likely because they were all C6 alcohols. HS-SPME is a fast and simple sampling method for analysis of volatile compounds, but its extraction efficiency for alcohols was affected by ethanol. The extraction efficiency of the HS-SPME fibre increased as the ethanol content increased from 6% to 12% for simple and low molecular weight alcohols, while it decreased for complex and high molecular weight alcohols (such as decanol, benzyl alcohol and phenyl ethanol). The reason was that lower MW volatiles equilibrated sooner amongst the three phases present in the sampling vial: the liquid, the headspace of the vial, and the SPME fibre than higher MW volatiles (Matich, Rowan, & anks, 1996). The solubility of the low MW components decrease in the liquid as a consequence of competition with ethanol for dissolved water. Thus with an increase in ethanol content, the partitioning between gaseous and liquid phases will change. For higher MW components, ethanol acts as a co-solvent in solution, enhancing their solubility, resulting in a decreased partitioning in the gas phase of the sample headspace, thereby reducing the amount of fibre absorption. When the ethanol content was above 12%, more active sites on the fibre were occupied by ethanol molecules, which resulted in the amount of all alcohols on the fibre to be reduced (Rocha et al., 21) Esters The ability of the fibre to absorb esters depended on the amount of ethanol in the matrix and the chemical structure of the ester. The amount of diethyl succinate absorbed to the fibre steeply decreased as the ethanol content increased from 2% to 4% (p < 5), while it was only moderately influenced by 4% to 6% ethanol (Fig. 2). Isoamyl acetate, ethyl lactate and phenylethyl acetate (Fig. 2) decreased as the ethanol content increased from 2% to

4 746 M. Zhang et al. / Food Chemistry 125 (211) e+8 7.e+8 6.e+8 5.e+8 4.e+8 3.e+8 2.e+8 1.e+8-1.e+8 4-methyl-2-pentanol 5e+8 4e+8 3e+8 2e+8 1e+8 C 2 D Gra p h 4 E-3-hexen-1-ol Z-3-hexen-1-ol E-2-hexen-1-ol Z-2-hexen-1-ol methionol benzyl alcohol phenyl ethanol 1.6e+9 1.4e+9 1.2e+9 isoamyl alcohol 1.6e+9 1.4e+9 1.2e+9 D hexanol octanol decanol 1.e+9 8.e+8 6.e+8 1.e+9 8.e+8 4.e+8 6.e+8 2.e+8 4.e+8 2.e+8 Ethanol (%,v/v) Fig. 1. Changes in the peak areas of the internal standard and alcohols absorbed by the HS-SPME fibre in the presence of various amounts of ethanol. 4%, while esters including ethyl hexanoate, ethyl octanoate, ethyl decanoate and hexyl actate increased (Fig. 2C). These results were in accordance with those of other authors, who also detected the influence of ethanol on the extraction of esters (Rocha et al., 21; Rodríguez-encomo et al., 22; Whiton & Zoecklein, 2). With increase of ethanol concentration, the ethanol molecules tend to form clusters which can dissolve more esters, leading to increase in the solubility of the esters in the liquid phase and thus decreasing the headspace concentration (Conner et al., 1998). Consequently, the distribution coefficient of esters between the fibre surface and the sample will change. Rodríguez-encomo et al. (22) suggested using an internal standard to correct the effects of ethanol on esters. However, the effects of ethanol content on esters have been resolved by using separate internal standards to target the extraction and chromatographic behaviour of analytes in the matrix (ntalick et al., 21) Fatty acids Hexanoic acid, octanoic acid and decanoic acid, all with identical chemical structures, are the main organic acids in wine (Zhang, QU, Zhang, & Duan, 27b). Effects of ethanol on the fibre to absorb these three acids were studied (Fig. 3). The amount of the three acids notably decreases as the ethanol content increases (p < 5), with the exception of decanoic acid in the presence of ethanol from 2% to 4% (Fig. 3). Whiton and Zoecklein (2) analysed the effect of ethanol (1%, 12% and 14%) on the HS-SPME fibre to absorb acetic acid. They observed that when the ethanol content varied from 1% to 14%, the amount of acetic acid was reduced. We also observed that at the same concentration, the peak areas of the three acids detected by MS were less than those of alcohols and esters. This could be due to differences in the absorption ability of the fibre for aromatic compounds as well differences in the MS response to aromatic compounds. Howard et al. (25) used GC/MS and GC/FID to analyse the aroma composition of wine and found that the limits of detection for acids was high, especially decanoic acid Monoterpenes Fig. 4 shows the impact of the presence of ethanol on the fibre to absorb the following monoterpenes: linalool, terpinen-4-ol, citronellol and nerol. The amount of compounds extracted gradually decreased with an increase in ethanol content (Fig. 4). de la Calle García et al. (1996) used SPME to sample monoterpenes in wines and model solutions. They observed that as the ethanol content increased from 2% to 12%, the measured amounts of geraniol, nerol, citronellol, linalool and a-terpineol were reduced nalysis of aromatic compounds during Syrah must alcoholic fermentation by HS-SPME/GC MS Quantitative method of aromatic compounds lcohols, esters and acids are the primary aromatic compounds in wine and are formed during the alcoholic fermentation process. The quantification of these compounds can contribute to the evaluation of optimal wine technology (Rapp, 1998). Therefore, it is important to quantitatively determine the amounts of fermented aromatic compounds. ccording to the above results, the quantitative determination of aromatic compounds in wine was influenced by ethanol, and the effects depended on the chemical properties such as boiling points and polarities of aromatic compounds analysed and the ethanol content of the matrix. The effects of ethanol on the HS-SPME fibre extraction of aromatic compounds was not changed due to the addition of the internal standard, which confirmed that the important matrix effect can not be avoided by the internal standard method. The chemical groups (alcohols, esters, acids and monoterpenes) showed specific behaviours in SPME

5 M. Zhang et al. / Food Chemistry 125 (211) e+9 1.e+9 8.e+8 6.e+8 4.e+8 3.e+7 2.5e+7 2.e+7 1.5e+7 1.e+7 isoamyl acetate ethyl lactate phenylethyl acetate 5e+8 4e+8 3e+8 2e+8 1e+8 hexanoic acid octanoic acid 5.e+6 3.e+8 2.5e+8 ethyl acetate diethyl succinate 2.5e+7 2.e+7 decanoic acid 2.e+8 1.5e+7 1.5e+8 1.e+7 1.e+8 5.e+6 5.e+7 1.e+9 8.e+8 6.e+8 4.e+8 2.e+8 1.5e+8 1.e+8 5.e+7 C ethyl octanoate ethyl decanoate ethyl hexanoate hexyl acetate Ethanol (%, v/v) Fig. 2. Changes in the peak areas of esters absorbed by the HS-SPME fibre in the presence of various amounts of ethanol. Ethanol (%, v/v) Fig. 3. Changes in the peak areas of fatty acids absorbed by the HS-SPME fibre in the presence of various amounts of ethanol. 1e+8 8e+7 6e+7 4e+7 2e+7 linalool terpinen-4-ol citronellol nerol Ethanol (%, v/v) Fig. 4. Changes in the peak areas of monoterpenes absorbed by the HS-SPME fibre in the presence of various amounts of ethanol. analysis. Therefore, calibration curves in the presence of seven different percentages of ethanol (2%, 4%, 6%, 8%, 1%, 12% and 14%) were established to quantitatively determine the amounts of aromatic compounds during alcoholic fermentation. The concentration ranges of aromatic compounds were selected according to the concentrations of these compounds in must-wine. The correlation coefficients of the calibration curves were above.9 with a RSD of less than 1%. lthough the detection limit of the analytes changed as the percentages of ethanol varied, the calibration curves were satisfactory for the quantitative analysis of samples. Calibration in a model wine is useful because it can be used as a screening protocol for monitoring overall changes in a wide variety of matrices. For the more accurate quantification, calibration curves in matrices similar to wine samples should be used (e.g., standard addition calibrations) for all analytes (Canuti et al., 29). ccording to the content of ethanol in samples detected, the corresponding calibration curve was used for quantification of aromatic compounds lcohols During Syrah must alcoholic fermentation, the major alcohols produced are isoamyl alcohol, phenyl ethanol, hexanol, Z-3-hexen-1-ol, E-2-hexen-1-ol, 2,3-butanediol and methionol. The level of isoamyl alcohol is the highest, varying from to 34.5 mg l 1, followed by phenyl ethanol at 7.6 to 13.3 mg l 1. It was determined that isoamyl alcohol, phenyl ethanol, 2,3-butanediol and methionol followed a similar production pattern during the course of fermentation, increasing quickly after inoculation, rising to their highest concentration at the peak of fermentation, followed by a

6 748 M. Zhang et al. / Food Chemistry 125 (211) Concentration (mg/l) isoamyl alcohol hexanol Z-3-hexen-1-ol E-2-hexen-1-ol 2,3-butanediol methionol phenylethanol C acetic acid hexanoic acid octanoic acid decanoic acid Concentration (mg/l) 5 ethyl acetate isoamyl acetate ethyl hexanoate hexyl acetate 4 ethyl octanoate ethyl decanoate phenylethyl acetate D linalool terpinen-4-ol citronellol nerol Fermentation time (h) Fermentation time (h) Fig. 5. Changes of alcohols (), esters (), fatty acids (C) and monoterpenes (D) concentration over time during Syrah must alcoholic fermentation. decrease (Fig. 5). Z-3-hexen-1-ol and E-2-hexen-1-ol reached their highest concentration 24 to 48 h after inoculation and then decreased sharply to a low concentration. In fact, E-2-hexen-1-ol was depleted within the first few days of cell growth in all the cases. These results were in agreement with previous reports (Mauricio, Moreno, Zea, Ortega, & Medina, 1997; Zhang et al., 27b). The formation and disappearance of various alcohols likely resulted from their origin (Mauricio et al., 1997). lcohols are synthesised from a keto acid resulting from the oxidative deamination of an amino acid or fatty acid. Isoamyl alcohol, phenyl ethanol are synthesised from leucine (and isoleucine), valine and phenylalanine, respectively via their ketoacids, -ketoisocaproate (and - keto-methylvalerate), -ketoisovalerate and phenylpyruvate, the production of which depends on cellular growth (Mauricio et al., 1997). portion of Z-3-hexen-1-ol, E-2-hexen-1-ol and hexanol comes directly from grapes while another portion is formed by means of an oxidation reaction of high fatty acids. fter the grapes are crushed, the lipids are catalysed by lipoxygenase and transformed to Z-3-hexen-1-ol, E-2-hexen-1-ol and hexanol so that their concentrations are increased in the pre-fermentative maceration process with a peak 24 h after inoculation (Salinas, Garijo, Pardo, Zalacain, & lonso, 23). The reduction of alcohols may be due to the formation of the corresponding esters which are products of yeast metabolism. lcohols and acyl-coenzyme are substrates for ester synthesis Esters Fig. 5 shows the changes in the esters during Syrah must alcoholic fermentation. s shown, seven esters showed very similar patterns of formation. There was a slow production of esters at the beginning of pre-fermentative period. t the peak of fermentation, the maximum concentrations of esters appeared in the winemust. The amount began to reduce approximately 148 h after inoculation, then increased followed by another decrease. s a consequence, there were two peak levels of esters during alcoholic fermentation, probably as a result of their hydrolysis under the action of cellular esterases, the activity of which increases at the end of fermentation (Mauricio et al., 1997). The result was similar to the literature (Lee et al., 24; Vianna & Ebeler, 21). Ethyl acetate and isoamyl acetate were detected at the highest concentration throughout fermentation. Ethyl hexanoate, ethyl octanoate and ethyl decanoate were detected in moderate amounts, and hexyl acetate and phenylethyl acetate were detected in the lowest amounts. Isoamyl acetate reached a maximum concentration of about 45.1 up to 17.9 mg l 1. This result was in disagreement with that of Vianna and Ebeler (21). They reported isoamyl acetate was no more than 3.2 mg l 1 during fermentation. possible reason for this difference may be that they analysed esters in the gas phase of the fermentor headspace, and the effect of ethanol on SPME to absorb esters was not considered. The amount of esters in the gas or liquid phase is determined by their partition coefficient and can be impacted by ethanol (Conner et al., 1998). In order to avoid the effects of ethanol, calibration curves of matrices containing seven different ethanol contents were used to quantify analytes in our experiment, so that a more accurate measurement of the amounts of esters was obtained, resulting in values which were higher than those of previous reports Fatty acids cetic acid is the major fatty acid found during alcoholic fermentation. It is produced by an ethanol oxidation reaction during alcoholic and malolactic fermentation. The amount of acetic acid increased gradually throughout fermentation (Fig. 5C). t low levels this compound improves wine flavours, however, at high levels it is detrimental to the taste of wine due to a sour taste and therefore, its amount should be tightly controlled (Martí et al., 23). s shown in Fig. 5C, hexanoic acid, octanoic acid and decanoic acid followed an identical pattern of formation. They reached a maximum concentration at 96 h after inoculation, and then sharply de-

7 M. Zhang et al. / Food Chemistry 125 (211) clined from 96 to 168 h. The final concentrations of hexanoic acid and decanoic acid were lower than that of octanoic acid, only.27 and.54 mg l 1, respectively. Hexanoic acid, octanoic acid and decanoic acid are the products of lipid oxidation. During alcoholic fermentation, hexanoic acid, octanoic acid and decanoic acid take part in fatty acid ethyl synthesis (Lee et al., 24; Vianna & Ebeler, 21), which leads to their reduction Monoterpenes During the alcoholic fermentation process, not only are alcohols, esters and acids synthesised but many changes occur on terpenes. They come directly from grapes and are not metabolized by yeast during alcoholic fermentation (Rapp, 1998). s shown in Fig. 5D, the monoterpenes were at low concentrations in fermented Syrah must, especially terpinen-4-ol and nerol, both of which were not detected at the anaphase of alcoholic fermentation. The concentration of linalool was near to its detection limit, so it could not be consistently measured as shown in Fig. 5D. In general, these low levels of monoterpenes may result from esterification. Some monoterpenes increased at the beginning of fermentation, particularly citronellol, which increased sharply from 3 to.15 mg l 1. This was likely because most monoterpene are found linked to sugar moieties in the grape juice and wines and they could be liberated by acids or glycosidases found in grapes and yeast, which eventually leads to an increase in monoterpene concentration (Rapp, 1998). Yeast is capable of modifying the terpenic profile of the wine and thus, citronellol could be produced from geraniol and nerol (Vaudano, Morunol, & di Stefano, 24). This may be the reason for why the concentration of citronellol is higher than other monoterpenes in Syrah must-wine. 4. Conclusions HS-SPME is a fast and simple sampling method for the analysis of volatile compounds. lthough its ability to extract alcohols was impacted by the amount of ethanol in the matrix, it is feasible to minimise this using calibration curves developed by analysing an external standard contained in the corresponding amount of ethanol. This method has proved useful for monitoring aromatic compounds during must alcoholic fermentation. cknowledgement This research was supported by the Special Funds of Modern Industrial Technology System for griculture (nycytx-3). References ntalick, G., Perello, M. C., & de Revel, G. (21). Development, validation and application of a specific method for the quantitative determination of wine esters by headspace-solid-phase microextraction gas chromatography mass spectrometry. Food Chemistry, 121, egala, M., Corda, L., Podda, G., Fedrigo, M.., & Traldi, P. (22). Headspace solidphase microextraction gas chromatography/mass spectrometry in the analysis of the aroma constituents of Cannonau of Jerzu wine. Rapid Communications in Mass Spectrometry, 16, osch-fustéa, J., Riu-umatell, M., Guadayol, J. M., Caixach, J., López-Tamames, E., & uxaderas, S. (27). Volatile profiles of sparkling wines obtained by three extraction methods and gas chromatography mass spectrometry (GC MS) analysis. Food Chemistry, 15, Canuti, V., Conversano, M., Calzi, M. L., Heymann, H., Matthews, M.., & Ebeler, S. E. (29). Headspace solid-phase microextraction gas chromatography mass spectrometry for profiling free volatile compounds in Cabernet Sauvignon grapes and wines. Journal of Chromatography, 1216, Conner, J. M., irkmyre, L., Paterson,., & Piggott, J. R. (1998). Headspace concentrations of ethyl esters at different alcoholic strengths. Journal of the Science of Food and griculture, 77, de la Calle García, D., Magnaghi, S., Reichenbächer, M., & Danzer, K. (1996). Systematic optimization of the analysis of wine bouquet components by solidphase microextraction. Journal of High Resolution Chromatography, 19, Hayasaka, Y., MaNamara, K., aldock, G.., Taylor, R. L., & Pollnitz,. P. (23). pplication of stir bar sorptive extraction for wine analysis. nalytical and ioanalytical Chemistry, 375, Howard, K. L., Mike, J. H., & Riesen, R. (25). Validation of a solid-phase microextraction method for headspace analysis of wine aroma components. merican Journal of Enology and Viticulture, 56, Lee, S. J., Rathbone, D., simont, S., dden, R., & Ebeler, S. E. (24). Dynamic changes in ester formation during chardonnay juice fermentations with different yeast inoculation and initial rix conditions. merican Journal of Enology and Viticulture, 55, Mallouchos,., Komaitts, M., Koutinas,., & Kanellaki, M. (22). Investigation of volatiles evolution during the alcoholic fermentation of grape must using free and immobilized cells with the help of solid phase microextraction (SPME) headspace sampling. Journal of gricultural and Food Chemistry, 5, Martí, M. P., Mestres, M., Sala, C., usto, O., & Guasch, J. (23). Solid-phase microextraction and gas chromatography olfactometry analysis of successively diluted samples. new approach of the aroma extract dilution analysis applied to the characterization of wine aroma. Journal of gricultural and Food Chemistry, 51, Matich,. J., Rowan, D. D., & anks, N. H. (1996). Solid phase microextraction for quantitative headspace sampling of apple volatiles. nalytical Chemistry, 68, Mauricio, J. C., Moreno, J., Zea, L., Ortega, J. M., & Medina, M. (1997). The Effects of grape must fermentation conditions on volatile alcohols and esters formed by Saccharomyces cerevisiae. Journal of the Science of Food and griculture, 75, Rapp,. (1998). Volatile flavour of wine: Correlation between instrumental analysis and sensory perception. Nahrung, 6, Rocha, S., Ramalheira, V., arros,., Delgadillo, I., & Coimbra, M.. (21). Headspace solid phase microextraction (SPME) analysis of flavor compounds in wines. Effect of the matrix volatile composition in the relative response factors in a wine model. Journal of gricultural and Food Chemistry, 49, Rodríguez-encomo, J. J., Conde, J. E., Rodríguez-Delgado, M.., García-Montelongo, F., & Pérez-Trujillo, J. P. (22). Determination of esters in dry and sweet white wines by headspace solid-phase microextraction and gas chromatography. Journal of Chromatography, 963, Salinas, R. M., Garijo, J., Pardo, F., Zalacain,., & lonso, G. L. (23). Color, polyphenol, and aroma compounds in Rosé wines after prefermentative maceration and enzymatic treatments. merican Journal of Enology and Viticulture, 54, Vas, G., lechschmidt, I., Kovacs, T., & Vekey, K. (1999). Examination of aroma production kinetics of different commercial wine yeast in fermenting Muscat Ottonel wines with the help of SPME headspace sampling and fast GC analysis. cta limentaria, 28, Vaudano, E., Morunol, E. G., & di Stefano, R. (24). Modulation of geraniol metabolism during alcohol fermentation. Journal of the Institute of rewing, 11, Vaz Freire, L. M. T., Costa Freitas,. M., & Relva,. M. (21). Optimization of solid phase microextraction analysis of aroma compounds in a Portuguese Muscatel wine must. Journal of Microcolumn Separations, 13, Vianna, E., & Ebeler, S. E. (21). Monitoring ester formation in grape juice fermentations using solid phase microextraction coupled with gas chromatography mass spectrometry. Journal of gricultural and Food Chemistry, 49, Whiton, R. S., & Zoecklein,. W. (2). Optimization of headspace solid-phase microextraction for analysis of wine aroma compounds. merican Journal of Enology and Viticulture, 51, Zhang, M., QU, W., Zhang, H., & Duan, C. (27b). Effect of maceration enzymes on the formation of aroma compounds during Cabernet Sauvignon alcohol fermentation. GRO Food Industry Hi-tech., 18, 5 7. Zhang, M., Xu, Q., Duan, C., Qu, W., & Wu, Y. (27a). Comparative study of aromatic compounds in young red wines from Cabernet Sauvignon, Cabernet Franc, and Cabernet Gernischet varieties in China. Journal of Food Science, 72,