Analysis of terpenes in white wines using SPE–SPME–GC/MS approach

Analytica Chimica Acta 677 (2010) 43–49 Contents lists available at ScienceDirect Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca Analysis of terpenes in white wines using SPE–SPME–GC/MS approach Mariusz Dziadas, Henryk H. Jeleń ∗ Faculty of Food Science and Nutrition, Poznań University of Life Sciences, Wojska Polskiego 31, 60.624 Poznań, Poland a r t i c l e i n f o Article history: Received 9 December 2009 Received in revised form 31 May 2010 Accepted 26 June 2010 Available online 23 July 2010 Keywords: Wine Flavor compounds Terpenes Solid phase microextraction Solid phase extraction gas chromatography/mass spectrometry a b s t r a c t Terpenes contribute to some white wines aroma, especially these produced from Muscat grapes and others aromatic ones of high terpene contents (Gewürtztramminer, Traminer, Huxel, Sylvaner). Terpenes are present in wine in free and bound (in a form of glycosides) forms. Analyses of bound terpenes are usually performed using solid phase extraction after hydrolysis of glycosides. A new method for determination of terpenes from wine, focused on determination of terpenes released after acidic hydrolysis, based on solid phase extraction (SPE) followed by solid phase microextraction (SPME) was developed. Non-polar (free) and polar (bound terpenes) fractions were separated on 500 mg C18 cartridges. Bound terpenes were sampled using SPME immediately after acidic hydrolysis in non-equilibrium conditions. Application of combined SPE–SPME approach allowed quantification of selected terpenes in lower concentrations than in SPE approach and added a selectivity to the method, which enabled detection of compounds nondetectable in SPE extracts. Results obtained by SPE and SPE–SPME approach were correlated for free terpenes and those released after acid hydrolysis 20 white wines obtained from different grape varieties (R2 = 0.923). Although developed for wine terpenes analysis, SPE followed by SPME approach has a great potential in analysis of other bound wine flavor compounds, especially those potent odorants present in trace amounts. © 2010 Published by Elsevier B.V. 1. Introduction Wine volatiles comprise a group several hundred compounds, representing different chemical classes and present in a very diverse concentrations ranging from mg L−1 down to ng L−1 . Their role in formation of wine flavor is dependent on their abundance and odor thresholds (OT), many influence the flavor of wine, though present in very low concentrations, due to their sensory properties [1]. In the aroma of white wines monoterpenes play an important role, being a group of flavor compounds characteristic for specific grapes used for wine production. In some of white wines such as Muscat and Gewürtztramminer, they are among key odorants and their concentration can be several mg L−1 . In Riesling, Sylvaner, Traminer, Hüxel and Müller Thurgau grapes they are also relatively abundant and contribute to the flavor of wines produced from these varieties [2]. The predominant monoterpenes in white wines are linalool, geraniol, nerol, ␣-terpineol, ␤-citronellol, hotrienol and limonene. Linalool (OT = 15 ␮g L−1 ) and geraniol (OT = 30 ␮g L−1 [3]) are the most abundant monoterpene alcohols in Muscat white wines. Terpene compounds in wine are prone to changes during winemaking process. Especially monoterpene alcohols can ∗ Corresponding author. Tel.: +48 61 8487273; fax: +48 61 8487314. E-mail address: [email protected] (H.H. Jeleń). 0003-2670/$ – see front matter © 2010 Published by Elsevier B.V. doi:10.1016/j.aca.2010.06.035 undergo several reactions during wine production and storage, induced by the time of storage, relatively low pH and a presence of compounds that can interact with them. From geraniol three diols can be formed, which subsequently can undergo further transformations into hotrienol and nerol oxide. Monoterpene alcohols can easily isomerize and oxidize, forming oxides and aldehydes. Monoterpenes occur in grapes and wines in a free form, however it is known that majority of them are also bound with sugars as glycosides [2]. This phenomenon is characteristic not only to grape berries but also to many other fruit such as apricots, cherries, kiwi, pineapple, strawberries [4]. It estimated that glycosidally bound volatiles are two to eight times more abundant than their free forms [5]. Some portion of these bound terpenes can be released and aglycones can enhance the flavor of wine. Knowledge on the total amount of free and bound terpenes is required to estimate the flavor potential of grapes, to foresee how much can be released in the process of winemaking using specific yeast [6,7] and assess their changes during wine ageing [8]. It can be also helpful in obtaining wines with flavor enhanced by the addition of enzymes (exogenous glycosidases) that can release monoterpenes from their nonvolatile precursors [2,9,10]. The activity of exogenous glycosidases is thoroughfully investigated and enzymes enhancing wine flavor are commercially available. On the other hand data on the role of enologically significant glycosidases in yeast used for wine production is not consistent and range from reports that indicate a low activity of yeast glycosidases to ones showing large decline 44 M. Dziadas, H.H. Jeleń / Analytica Chimica Acta 677 (2010) 43–49 in glycosides during fermentation processes. In a recent study by Ugliano and co-workers it was proven that fermentation causes a decrease (22–28% depending on a yeast strain) in the concentration of glycosides compared to 5% in non-fermented controls. Significant amounts of flavor aglycones were released. It was also observed that ␤-d-glycosides, ␣-l-arabinofuranosyl-␤-d-glucosides and ␣l-rhamnopyranosyl-␤-d-glucosides are hydrolyzed to a higher degree by yeast than ␤-d-apiofuranosyl-␤-d-glucosides [11]. Many scientific papers have been devoted to the analysis of free terpenes in wines. Due to the complexity of wine matrix and relatively low concentrations of these compounds their analyses require some isolation/preconcentration steps [12,13]. Solid phase microextraction performed from headspace phase is a method that provides good selectivity and low detection limits and have been described in several papers [14–16]. The analysis of both free and bound terpenes in wine requires fractionation of the sample and separation of volatile terpenes (non-polar fraction) from those, water soluble sugar bound (polar fraction). The fractionations of free and bound flavor compounds in wine are usually done using reversed phase SPE [17–19], HPLC [20] or Amberlite XAD-2 resins [21]. For extraction of terpenes reversed phase SPE provides is the most effective, whereas LiChrolut EN resins and XAD-2 resins perform more effectively in the extraction of acids, benzenes, phenols [22]. For the analysis of flavor compounds aglycones are of interest, so methods involving some kind of hydrolysis – enzymatic or acid have been developed and released aglycones are analyzed using GC/MS [23]. The sugar moiety is represented usually by rhamnose ((6-O-␣-l-rhamno-pyranosyl␤-d-glucopyranosyl)-terpenes) and arabinose (6-O-␣-l-arabinofuranosyl-␤-d-glucopyranosyl)-terpenes). Sometimes glycosides are analyzed directly after derivatization [11,21]. Of the described in literature methods of analysis of free and bound flavor compounds in wine, the one described by Ferreira and co-workers [22] is attractive for analytical purposes, as it requires relatively short time to perform acidic hydrolysis and is robust and simple to perform. Moreover, release and changes to flavor compounds obtained via acid hydrolysis resembles more process of wine ageing compared to enzymatic hydrolysis. Because in the analysis of both free and bound terpenes, solid phase extraction (SPE) is used for fractionation, the procedure involves extracts concentration to a relatively small volume and subsequent analysis using GC/MS systems. As there are numerous methods describing analysis of free volatile compounds, also terpenes in wines, our main goal was to focus on bound terpenes, released in a result of acidic hydrolysis. The work was aimed at combining the benefits of SPME – its selectivity and sensitivity with the standard approach to bound volatiles based on SPE, which allows sample fractionation according to fractions polarity. Szeremley 2006, 2007), Muscats (Menada Estate, Varna 2007, Muskadet Serve et Maine, Balkania 2005, Chateau Menada) and Shell Segal. Standards of linalool, ␣-terpineol, ␤-citronellol, nerol, geraniol, cis- and trans-linalool oxides, cis- and trans-limonene epoxides, myrtenol were purchased from Sigma–Aldrich (Poznań, Poland), Standard of 2 [H]7 -geraniol for quantification purposes was synthesized in the authors laboratory according to method described by Pedersen et al. [24]. SPME fibers (2 cm carboxene/divinylbenzene/PDMS, 100 ␮m PDMS, carboxene/PDMS, divinylbenzene/PDMS, polyacrylate were purchased from Supelco (Poznań, Poland). Methanol (HPLC grade), methylene chloride (GC grade), pentane (GC grade) used for experiments were purchased from Sigma–Aldrich (Poznań, Poland). Bi-distilled in glass water was used for SPE sample preparation. Solid phase extraction was performed using 500 mg C-18 Isolute cartridges (Biotage, Sweden). 2.2. SPE fractionation and analysis of free monoterpenes A 100 mL portion of white wine, to which 13.5 ␮g of 2 [H]7 geraniol in ethanol was added as an internal standard was mixed for 5 min using magnetic stirrer, and then applied to a preconditioned 500 mg RP C18 SPE column. Preconditioning was performed by purging at 3 mL min−1 the column with 25 mL portions of methanol and water. After loading a sample onto the column it was washed with 150 mL of water. Non-polar fraction (NPF) was eluted using 25 mL of a mixture of pentane/dichloromethane (2/1, v/v). Subsequently, polar fraction (PF) was eluted using 25 mL of methanol and subjected to hydrolysis. Non-polar fraction was evaporated to approximately 500 ␮L firstly heating it at 30 ◦ C water bath without mixing or stirring for 30 min, then in a delicate stream of nitrogen and 1 ␮L of it was introduced in a splitless mode into GC/MS system. 2.3. Fast acid hydrolysis for SPE analysis – analysis of bound monterpenes Polar fraction was evaporated to dryness under a stream of nitrogen and rehydrated using 10 ml of Mc’Ilvaine citric (0.2 M) buffer (pH 2.5) according to method of Ferreira [22]. After hydrolysis, which lasted for 1 h at 100 ◦ C and was performed in a 40 mL vial, the vial was cooled down and 13.5 ␮g of 2 [H]7 -geraniol in ethanol was added as an internal standard and the sample was vortexed for 2 min. The hydrolyzate (10 mL) was loaded onto a preconditioned (with 25 mL of methanol and 25 mL of water) SPE column and freed compounds were eluted (using a flow of 3 mL min−1 ) with 25 mL of pentane/dichloromethane mixture (2/1, v/v). It was concentrated to 500 ␮L in a similar way as for free terpenes and 1 ␮L was injected splitless into GC/MS system. 2. Materials and methods 2.4. SPE–SPME analysis of free monoterpenes 2.1. Samples and reagents Moldova Sun Muscat white wine was used for the method development representing wine with relatively rich monoterpenes profile. Twenty white wines of different origin, not only Muscat types were used to compare developed method and correlate results. Ten wines were produced in Podkarpacie region in Southern Poland (Bianca 2008, Hibernal 2007, Cuvee 2007, Jutrzenka – from Golesz Winery, Bianca, Seyval Blanc, Jutrzenka – from Dwie Wieże Winery, Muscat 2007, Bianca 2007, Sibera 2007 – from Zacisze Winery) remaining ones were purchased in local wine store: Rieslings (Cave de Turcheim 2007, Peter Mertes 2007, Non-polar fraction was evaporated to dryness under a delicate stream of nitrogen. Then, 1 mL of methanol was added, vial (40 mL) vortexed and 10 mL of water was added followed by vortexing for 2 min. The vial was placed in a heating block for 20 min at 100 ◦ C, then removed from the block and, while cooling down sampled using SPME fiber for 20 min. During the extraction temperature of the vial decreased to ambient temperature. After extraction SPME fiber was placed in an injection port of GC/MS system and desorbed at 260 ◦ C for 5 min. In the evaluation of SPE–SPME method temperature of the headspace inside vials was monitored using Almemo 2390-4 (Ahlborn, Germany) thermometer with a thermocouple. M. Dziadas, H.H. Jeleń / Analytica Chimica Acta 677 (2010) 43–49 45 2.5. SPE–SPME analysis of bound monoterpenes using acid hydrolysis Polar fraction was evaporated to dryness in a stream of nitrogen and hydrolyzed using Mc’Ilvaine buffer (pH 2.5). Immediately after 1 h hydrolysis, 13.5 ␮g of 2 [H]7 -geraniol in ethanol was added through the silicon/Teflon membrane using syringe, vortexed and immediately extracted using SPME fiber similarly to the analysis of free monterpenes. After 20 min extraction fiber was transferred into a GC/MS system to desorb. 2.6. GC/MS analysis of samples prepared by SPE and SPE–SPME approaches GC/MS analyses were performed using Agilent 7890GC A with a split/splitless injector coupled to Agilent 5975C VL MSD quadrupole mass spectrometer equipped with triple axis detector (TAD) and a diffusion pump. Liquid samples were injected using Agilent 7683B autosampler. All samples were injected in the splitless mode: liquid samples at 220 ◦ C using double gooseneck liner, SPME was performed at 260 ◦ C using 0.75 mm SPME liner. Analyzed compounds were resolved on a Supelcowax-10 column (30 m × 0.25 mm × 0.25 ␮m, Supelco, Bellefonte, PA). Analyses were performed at programmed temperature starting from 40 ◦ C for 1 min, then increased at 5 ◦ C min−1 to 200 ◦ C, followed by an increase of 20 ◦ C min−1 to 230 ◦ C, at which the temperature was kept constant for 9.5 min. The total run time was 44 min. GC/MS transfer line temperature was 240 ◦ C, ion source temperature was 230 ◦ C and quadrupole – 150 ◦ C. Spectra were acquired using electron impact ionization (EI, 70 eV) in a full scan mode, using scan range of m/z 33–333 Da. Detector was run at a gain factor (GF = 1) mode at EMV voltage of 1760 V. Helium flow was 0.8 mL min−1 (32.4 cm s−1 ). Identification of terpenes of interest was performed based on the comparison of retention time and mass spectra of that with authentic standards or tentatively, when standards were not available. Quantitation of all monoterpenes was performed using deuterated geraniol isotopomer – 2 [H]7 -geraniol. AMDIS (v. 2.65 build 116,66) software was used for compounds identification, the quantitative purposes and s/n determination. To use AMDIS features target compound library (.msl) and internal standard library (.isl) were created for the compounds of interest. Fig. 1. Influence of water/headspace ratio on the amount of main monoterpene alcohols adsorbed on PDMS fiber. bath at 30 ◦ C for 20 min, then a delicate stream of nitrogen was applied to remove the solvent. This approach guaranteed minimal loses of target compounds in the evaporation process. After evaporation of pentane/dichloromethane the residue was re-dissolved in methanol, which was subsequently diluted with water (1:10), to facilitate SPME sampling and improve partition of terpenes from liquid matrix into headspace. To check whether post-SPE solvent evaporation induces losses of terpenes, evaporation of SPE extracts of increasing concentrations and their subsequent SPME sampling was tested on a standard solution of selected monoterpene alcohols. For concentration between 20 ␮g L−1 and 1000 ␮g L−1 a linearity above 0.9751 was achieved for SPME of reconstituted extracts (for linalool R = 0.9919; ␣terpineol R = 0.9751; ␤-citronellol R = 0.9961; nerol R = 0.9884; 3. Results and discussion 3.1. Development of SPE–SPME method The combination of SPE–SPME method relied on the SPE method described by Ibarz and co-workers. Solid phase microextraction parameters however had to be evaluated. Different phase ratios, fiber coatings, extraction times and mode of extraction were tested. In the initial experiment for SPME sampling, 40 mL vials were tested filled with 11, 22 and 33 mL of 10% methanol solution with added standards of main monoterpene alcohols. Their amount adsorbed at 50 ◦ C on PDMS SPME fiber was inversely proportional to the amount of liquid phase in the vial (Fig. 1). In all subsequent experiments fractions after SPE were evaporated to dryness and then reconstituted in 1 mL of methanol to which, subsequently 10 mL of water was added. Such solution was sampled with SPME fiber. In the SPE–SPME approach evaporation of SPE extracts to dryness is necessary to eliminate especially methylene chloride solvent, which can have a detrimental influence on SPME fiber and adsorption. To minimize losses of volatile compounds, SPE extract containing free terpenes was placed in a water Fig. 2. Comparison of SPME fibers used for extraction of wine volatile compounds after acidic hydrolysis of Moldova Sun wine. Total peak area of all volatiles (total) as well as ions typical for terpene compounds (EIC; m/z 136, 152, 154, 204) are compared. Volatile compounds were sampled using SPE–SPME approach. C/D/P – carboxene/divinylbenzene/PDMS; C/P – carboxene/PDMS; D/P – divinylbenzene/PDMS; P – PDMS (polydimethylsiloxane); PA – polyacrylate. 46 M. Dziadas, H.H. Jeleń / Analytica Chimica Acta 677 (2010) 43–49 Fig. 3. Comparison of extraction times used for isolation of wine volatile compounds after acidic hydrolysis. Ions typical for terpene compounds (EIC; m/z 136, 152, 154, 204) are shown. Volatile compounds were sampled using SPE–SPME approach with carboxene/divinylbenzene/PDMS fiber. geraniol R = 0.9921). Due to the differences in volatility between terpenes and pentane/dichloromethane mixture sample evaporation in aforementioned experiment did not contribute to terpenes losses. To check the particular SPME fiber suitability for SPE–SPME method development five fibers were tested. Muscat wine was used for fiber comparison, without spiking it with standards. Using wine instead of standard mixture gives conditions and a matrix more similar to the natural ones. Fig. 2 shows a comparison of five fibers used in these experiments. Both total peak areas (total) and selected ions (EIC; m/z 136, 152, 154, 204, characteristic for terpenes) of compounds extracted from Muscat wine after acidic hydrolysis were compared. Extraction was performed in the described below manner, starting immediately after termination of hydrolysis process and lasted for 20 min. Carboxene/divilylbenze/PDMS revealed the highest peak responses, followed by carboxene/PDMS divinylbenzene/PDMS and PDMS fibers. Extraction times were also tested for extraction performed immediately after hydrolysis – extraction lasted 20, 40 or 60 min. Results shown on Fig. 3 indicate minor differences for the areas of compounds of interest (terpenes), the highest differences being observed for ion m/z 136 (mainly monoterpene hydrocarbons). Acidic hydrolysis used to release wine bound terpenes required high temperature (in the case of method used it was 100 ◦ C). The hydrolysate was then cooled, and fractionated on a SPE with released monoterpene aglycones eluted with pentane/dichloromethane, concentrated and analyzed by GC/MS. As the temperature of hydrolysis is high terpene compounds will be present mainly in the headspace immediately after completing of this process. Assuming this, using SPME allows transfer of volatiles into the fiber with additional preconcentration, offering also a selectivity of SPME coating. Our idea was to perform SPME sampling immediately after hydrolysis termination, when majority of terpenes are still in the headspace. To sample volatile compounds released in the hydrolysis process the vial was removed from the heating block after 1 h heating at 100 ◦ C and sampled using SPME fiber. Sampling at high temperature has a benefit of increased transfer of compounds into a headspace phase from the liquid phase. However, at the high temperature desorption Fig. 4. Upper graph: peak areas of compounds extracted using SPME (PDMS) for 20 min (in triplicates); headspace A – sampling started after removing vial from 100 ◦ C heating block and letting it cool down; headspace B – sampling from vial placed in a 50 ◦ C heating block and heated throughout sampling; compounds codes: (a) cis-limonene oxide; (b) trans-limonene oxide; (c) linalool; (d) ␣-terpineol; (e) ␤citronellol; (f) myrtenol; (g) nerol; (h) deuterated geraniol; (i) geraniol. Lower graph: temperatures measured in the headspace of vials from which SPME sampling was performed. from the fiber can overcome the benefits of compound partition into the headspace. To minimize the influence of desorption from the fiber extraction of volatiles from hydrolysate was performed immediately after hydrolysis, but after removal the vial from the heating block, so the temperature was gradually dropping to near ambient. To test the extraction efficiency performed in such a way, sampling starting at 100 ◦ C after removal of the vial from the heating block was compared with sampling at a constant temperature of 50 ◦ C set on the heating block. Samples were stirred with magnetic stir-bar at 600 rpm. Results of this comparison are shown in Fig. 4. Temperature graph indicates that starting temperature in the headspace after removal of the vial from the block goes down from 76 ◦ C in the time 0–28 ◦ C after 20 min. The relatively high differences between heating block and headspace temperature was a result of using 40 mL vials, which were placed in a heating block leaving almost half of it above the surface of the block. When peak M. Dziadas, H.H. Jeleń / Analytica Chimica Acta 677 (2010) 43–49 47 Fig. 5. Chromatograms of volatile compounds after acidic hydrolysis from Moldova Sun wine obtained using SPE (A) and SPE–SPME (B) approach. areas were compared in these two setups tested (upper graph) for all tested compounds their peak areas were higher for sampling performed at “non-equilibrium” conditions. When errors of the sampling for test compounds in “non-equilibrium” conditions are compared with those for sampling at constant temperature it can be observed that non-equilibrium sampling does not provide worse precision. Because routinely SPE extracts are concentrated to 500 ␮L prior to GC/MS analysis, to increase concentration entering the column concentration of SPE extracts to 50 ␮L was tested, but no comparable results to SPE–SPME approach were obtained. Fig. 5 shows chromatograms of volatile compounds obtained after acid hydrolysis using “classical” sample preparation with SPE (upper chromatogram) and the same sample obtained using SPE–SPME approach. Chromatograms were adjusted to the same time and intensity scale. The number of peaks is much higher, so are the areas, however the baseline for TIC is also more unstable. 3.2. Verification of SPE and SPE–SPME method parameters The SPE method used for comparison with authors’ novel approach was that of Ibarz [22] using acidic hydrolysis for bound terpenes release. Recoveries of SPE method for linalool oxides, limonene 1,2-epoxides, linalool, ␣-terpineol, ␤-citronellol, myrtenol, nerol and geraniol determined using standards ranged from 58% (for trans-linalool oxide) to 102% for geraniol. Limits of detection for aforementioned group of compounds ranged from 1.1 to 1.9 ␮g L−1 . Limit of detection determined for the instrument used were in a range of 1–3 ng ␮L−1 in full scan (33–333 amu) mode. Because of the complexity of chromatograms and the baseline, for the quantitation and determination of limits of quantitation AMDIS deconvolution software supplied with NIST mass spectral library was applied. Amounts of terpenes of interest in Moldova Sun Muscat wine and also the limits of detection and reproducibility for SPE–SPME method was calculated using Moldova Sun wine. It was based on compounds present in analyzed wine and liberated after hydrolysis. Sample after hydrolysis was taken out from the heating block, internal standard solution was added by syringe through the septum, sample was mixed and compounds extracted using SPME in conditions described in Section 2. Terpenes were quantified based on TIC for compounds of interest comparing it with the TIC area of internal standard. For the compounds used for SPE parameters evaluation mentioned earlier in this section the response factors were ranging from 0.99 to 1.20, so response factors in quantita- M. Dziadas, H.H. Jeleń / Analytica Chimica Acta 677 (2010) 43–49 48 Table 1 Terpene compounds detected in Moldova Sun using SPE–SPME approach. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Compound Rt. [min] Conc. ± SD [␮g L−1 ]] LOD1 [␮g L−1 ] RSD1 [%] SPE Geranic oxide 1 Geranic oxide 2 1,3,5,5-Tetramethyl 1,3-cyclohexadiene Isoterpinolene ␤-Myrcene ␣-Terpinene Limonene ␤-cis-ocimene Tetrahydro-2,2-dimethyl-5-(1-methyl-1-propenylfuran) ␣-Pinene ␤-trans-ocimene m-Cymene Terpinolene cis-Rose oxide cis-Linalool oxide 3,6-Dihydro-4-methyl-2-(2-methyl-1-propenyl 0 2H pyran Linalool Hotrienol Ocimenol 1 Ocimenol 2 ␣-Terpineol ␤-Citronellol 4,5,9,10-Dehydro isolongifolene 7.50 7.69 8.29 8.72 8.98 9.44 9.95 10.85 11.05 11.19 11.28 11.76 12.19 14.16 16.79 16.95 18.65 20.18 21.19 21.74 22.26 23.69 27.26 41 103 3 49 7 13 42 16 10 10 16 15 47 10 20 79 172 97 34 41 179 18 82 0.35 0.52 0.09 0.38 0.12 0.14 0.30 0.18 0.10 0.17 0.19 0.16 0.48 0.04 0.21 0.50 0.69 0.45 0.24 0.24 0.61 0.17 0.34 8 11 14 9 16 14 6 15 12 9 7 15 13 16 15 7 12 17 11 11 8 12 14 − − − − − − − − − − − − − − + − + + + + + + − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3.5 12.2 0.4 4.5 1.0 1.9 2.4 2.1 1.3 1.3 1.1 2.2 6.2 0.2 0.3 5.6 20.5 16.2 3.7 4.3 14.3 2.1 13.7 Rt. – retention time [min]; Conc. – concentration of compounds in Moldova Sun wine [␮g L−1 ± standard deviation]; LOD1 – limit of detection for SPE–SPME method [␮g L−1 ]; RSD1 – repeatability of the SPE–SPME method [relative standard deviation, %, n = 3]; SPE – solid phase extraction; column shows if the compounds of interest were detectable using SPE method. tion of compounds in Moldova wine, when standards were not available were assumed to be 1. Limits of detection and repeatability of SPE–SPME method were estimated based on the analysis of diluted samples from the Moldova wine. The dilutions were 50%, 25%, 10%, 5% of the initial one, prepared on separate SPE cartridges. Results summarizing terpenes detected in Moldova Sun wine together with data on SPE–SPME method performance are shown in Table 1. Twenty three compounds were identified in Moldova Sun wine using SPE–SPME method in the concentrations from 3 to 179 ␮g L−1 . LODs for the method ranged from 0.04 ␮g L−1 for cis-rose oxide to 0.69 for linalool. The benefit of using SPE–SPME method for extraction of terpenes released after hydrolysis can be observed when results obtained by SPE are taken into consideration. The last column of Table 1 shows the number of compounds identified using traditional SPE only approach. 3.3. SPE and SPE–SPME methods for the analysis of free and bound terpenes in white wines To check the applicability of SPE–SPME method for the determination of analyzed monoterpenes in different white wines, 20 wines of different character were analyzed using SPE–SPME and compared to SPE approach. For the determination of bound terpenes in these wines acid hydrolysis was applied. In analyzed wines bound terpenes were up to 61.3 ␮g L−1 , whereas free terpenes ranged up to 392.7 ␮g L−1 . Correlation between results obtained using both methods (Fig. 6) can be expressed with R2 = 0.9203. Table 2 shows main monoterpenes detected in free forms and after acid hydrolysis in examined white wines. Data on terpenes in monovarietal Polish wines (from Bianca, Hibernal, Seyval Blanc, Sibera and Jutrzenka grapes) is reported for the first time. The dom- Table 2 Ranges of free and bound monoterpenes in monovarietal Polish wines and Muscat and Riesling wines compared in the experiments. Wine Linalool ␣-Terpineol ␤-Citronellol Nerol Geraniol cis-Linalol oxide cis-Limonene epoxide trans-Linalool oxide Free [␮g L−1 ] Bound [␮g L−1 ] Bianca 16.8–24.1 0.6–2.5 8.3–10.9 0.3–4.6 3.1–8.8 nd 0.8–1.6 nd–0.3 4.5–11.5 nd nd nd nd nd nd nd Hibernal 19.5 1.6 7.9 6.2 1.8 nd 1.0 0.4 5.9 nd nd nd nd nd nd nd Seyval Blanc 10.3 nd 4.7 0.3 2.9 nd 1.1 nd 4.2 nd nd nd nd nd nd nd Jutrzenka 44.3–58.0 nd–2.8 30.2–126.6 3.8–5.0 21.8–26.1 nd–1.0 7.3–11.5 nd–8.9 26.6–30.3 nd–8.7 nd nd nd nd nd nd Muscat 29.9–355.8 0.4–24.7 77.6–392.7 6.6–61.3 3.1–25.1 nd–1.9 0.5–13.8 nd–0.9 5.5–61.4 nd–1.9 nd nd–1.1 nd nd – 17.0 nd nd–1.0 Sibera 6.1 0.4 4.9 1.1 3.4 nd 0.4 nd 2.9 nd nd nd nd nd nd nd Riesling 16.6–54.4 0.6–1.4 21.9–53.8 2.7–11.6 1.3–5.8 nd 0.1–2.4 nd 3.9–10.3 nd nd nd nd nd–4.2 nd nd M. Dziadas, H.H. Jeleń / Analytica Chimica Acta 677 (2010) 43–49 49 be observed on SPE–SPME chromatograms, it provides much more information than the “classical” SPE approach and can be probably used to analyze other volatile compounds present in either free or bound forms in wines. To obtain a full information on volatile compounds in wine, apart from terpenes norisoprenoids, volatile phenols, vanillin derivatives, benzenes and lactones should be taken into consideration [26], many of them can appear in bound form. The use of SPE followed by SPME can provide additional information also for these groups of compounds. Acknowledgement Polish Ministry of Science and Higher Education is acknowledged for financing the project under project number N312 052 32/2781. References Fig. 6. Correlation obtained for results of determination of free and bound (after acidic hydrolysis) monoterpene alcohols in 20 different white wines using SPE and SPE–SPME approaches. inating monoterpene alcohols are linalool and ␣-terpineol in all of them, and Jutrzenka contained the highest amounts of monitored monoterpenes among wines obtained from Polish grape varieties. Jutrzenka is the most aromatic grape variety with fruity/flowery notes used for wine production in Podkarpacie region. Along with monovarietal wines, ranges of analyzed terpenes in Rieslings and Muscats were included. In Muscat wines linalool and ␣-terpineol reached 355.8 and 392.7 ␮g L−1 respectively, with lower concentrations of remaining main terpene alcohols. In Muscat varieties low amounts of linalool oxides and limonene epoxide were also detected. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] 4. Conclusions Presented method utilizes described in the literature fractionation of polar (bound) and non-polar (free) terpenes in wines using SPE followed up with a SPME analysis of SPE extracts. It adds the preconcentration abilities and selectivity of SPME to that of SPE. It is especially important for the analysis of compounds released in the hydrolysis. The most common hydrolysis process is the acidic one. The differences in the profiles of volatile compounds released after harsh acid and enzymatic hydrolyses were recently discussed in a detailed study by Loscos et al. [25]. They concluded the acid hydrolysis to be more useful for predicting aroma potential of grape varieties, than the enzymatic one, resembling more the rearrangements that undergo in acidic environment during wine production. Although described for monoterpene compounds analysis, as can [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] S. Ebeler, Food Rev. Int. 17 (2001) 45. J.J. Mateo, J. Jimenez, J. Chromatogr. A 881 (2000) 557. H. Guth, J. Agric Food Chem. 45 (1997) 3027. S. Macais, J.J. Mateo, Appl. Biotechnol. Microbiol. 67 (2005) 322. G. Krammer, P. Winterhalter, M. Schwab, P. Schreier, J. Agric. Food Chem. 39 (1991) 778. M. Fernandez-Gonzalez, R. Di Stefano, A. Briones, Food Microbiol. 20 (2003) 35. B.W. Zoecklein, J.E. Marcy, J.M. Williams, Y. Jasinski, J. Food Comp. Anal. 10 (1997) 55. B.W. Zoecklein, C.H. Hackney, S.E. Duncan, J.E. Marcy, J. Ind. Microbiol. Biotechnol. 22 (1999) 100. J.-E. Sarry, Z. Günata, Food Chem. 87 (2004) 509. T. Cabaroglu, S. Selli, A. Canbas, J.-P. Lepoutre, Z. Günata, Enzyme Microbial. Technol. 33 (2003) 581. M. Ugliano, E.J. Bartkowsky, J. McCarthy, L. Moio, P.A. Henschke, J. Agric. Food Chem. 54 (2006) 6322. M. Ortega-Heras, Gonzales-SanJosè Beltrán, Anal. Chim. Acta 458 (2002) 85. Z. Piñeiro, M. Palma, C.G. Barroso, Anal. Chim. Acta 513 (2004) 209. J.S. Câmara, M.A. Alves, J.C. Marques, Anal. Chim. Acta 555 (2006) 191. R.M. Pena, J. Barciela, C. Herrero, S. Garcia-Martin, J. Sci. Food Agric. 85 (2005) 1227. E. Sánchez-Palomo, M.C. Diaz-Maroto, M.S. Pèrez-Coello, Talanta 66 (2005) 1152. P.J. Williams, W. Cynkar, I.L. Francis, J.D. Gray, P.G. Iland, B.G. Coombe, J. Agric. Food Chem. 43 (1995) 121. J.J. Mateo, N. Gentilini, T. Huerta, M. Jimenez, R. Di Stefano, J. Chromatogr. A. 778 (1997) 219. M. Esti, P. Tamborra, Anal. Chim. Acta 563 (2006) 173. S. Bitteur, Z. Gunata, J.M. Brilloquet, C. Bayonove, R. Cordonnier, J. Sci. Food Agric. 47 (1989) 341. N. Carro, E. Lopez, Z.Y. Günata, R.L. Baumes, C.L. Bayonove, Analysis 24 (1996) 254. M.J. Ibarz, V. Ferreira, P. Hernandez-Orte, N. Loscos, J. Cacho, J. Chromatogr. A 1116 (2006) 217. R. Schneider, A. Razungles, C. Augier, R. Baumes, J. Chromatogr. A 936 (2001) 145. D.S. Pedersen, D.L. Capone, G.K. Skoroumounis, A.P. Pollnitz, M.A. Sefton, Anal. Bioanal. Chem. 375 (2003) 517. N. Loscos, P. Hernandez-Orte, J. Cacho, V. Ferreira, J. Agric. Food Chem. 57 (2009) 2468. N. Loscos, P. Hernandez-Orte, J. Cacho, V. Ferreira, J. Agric. Food Chem. 55 (2007) 6674.