Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry

G Model

ARTICLE IN PRESS

CHROMA-357361; No. of Pages 11

Journal of Chromatography A, xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry T. Rodríguez-Cabo, I. Rodríguez ∗ , M. Ramil, A. Silva, R. Cela Departamento de Química Analítica, Nutrición y Bromatología, Instituto de Investigación y Análisis Alimentario (IIAA), Universidad de Santiago de Compostela, Santiago de Compostela 15782, Spain

a r t i c l e

i n f o

Article history: Received 10 February 2016 Accepted 1 March 2016 Available online xxx Keywords: Accurate time-of-flight mass spectrometry Wine analysis Pesticides Phenols Screening

a b s t r a c t The performance of gas chromatography (GC) with accurate, high resolution mass spectrometry (HRMS) for the determination of a group of 39 semi-volatile compounds related to wine quality (pesticide residues, phenolic off-flavours, phenolic pollutants and bioactive stilbenes) is investigated. Solid-phase extraction (SPE) was used as extraction technique, previously to acetylation (phenolic compounds) and dispersive liquid-liquid microextraction (DLLME) concentration. Compounds were determined by GC coupled to a quadrupole time-of-flight (QTOF) MS system through an electron ionization (EI) source. The final method attained limits of quantification (LOQs) at the very low ng mL−1 level, covering the range of expected concentrations for target compounds in red and white wines. For 38 out of 39 compounds, performance of sample preparation and determination steps were hardly affected by the wine matrix; thus, accurate recoveries were achieved by using pseudo-external calibration. Levels of target compounds in a set of 25 wine samples are reported. The capabilities of the described approach for the post-run identification of species not considered during method development, without retention time information, are illustrated and discussed with selected examples of compounds from different classes. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Wine quality is a concerning issue in a globalized market with new consumers and producers. Traditionally, quality has been associated with the sensorial properties of this foodstuff, which are correlated with volatile species displaying different odour threshold and activity values [1,2]. In addition, several groups of semi-volatile compounds must be also taken into account in order to define the quality of wines. Some examples are (1) ethyl and vinylphenols which, depending on their concentrations, exert positive or negative effects in the aroma of wine, (2) natural bioactive compounds, e.g. resveratrol, which have been correlated with healthy effects associated to a moderate wine consumption [3], (3) undesired anthropogenic species, resulting from wine contamination during elaboration and/or packaging (e.g. bisphenol A and alkylphenols from detergents degradation) [4], and (4) pesticides. The determination of pesticide residues in wine is of particular relevance since (1) several producer countries, particularly Australia and those from South and North America, have more restrictive legislations than the European ones [5]; and furthermore, (2) cer-

∗ Corresponding author. E-mail address: [email protected] (I. Rodríguez).

tain pesticides affect the formation of aromatic compounds during must fermentation [6]. Analytical chemists have developed a myriad of approaches dealing with separate groups of the above compounds, such as ethylphenols [7], stilbene derivatives [8], pesticides [9] and other anthropogenic pollutants [4] residues in wine. However, there is still a need for multiclass methodologies, covering the determination of species with different origins, concentration ranges and belonging to different chemical families. The main challenges to develop multiclass reliable methods are: (1) the optimization of compatible sample preparation (extraction and concentration) conditions for compounds with different chemical functionalities; (2) their simultaneous determination achieving suitable limits of quantification (LOQs) (taking into account the maximum residue limits, MRLs, of pesticides and the odour thresholds of off-flavours); (3) attaining linear response intervals covering the expected range of concentrations in wines with different origins; and (4) achieving enough selectivity to guarantee the unequivocal detection of target compounds in the complex chemical environment of wine extracts. High resolution mass spectrometry techniques (HRMS) dramatically reduce the risks of mass spectral interferences when quantification ions are extracted using a mass window of a few milidaltons (mDa); furthermore, Orbitrap and time-of-flight (TOF) MS analysers provide an excellent sensitivity in the scan mode,

http://dx.doi.org/10.1016/j.chroma.2016.03.005 0021-9673/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model

ARTICLE IN PRESS

CHROMA-357361; No. of Pages 11

T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

2

Normalized response (%)

Toluene

1-octanol

Chlorobenzene

Carbon tetrachloride

140 120 100 80 60 40 20 0 BPA trans-RES

EP

IP

MET

FEN

PYR

CYP

PEN

PRY

IPR

AZO

CLE

EDR

Fig. 1. Comparison of responses attained, for a selection of compounds, as function of the extractant during DLLME of SPE extracts from the same wine (red wine, Mencía variety). Normalized responses to those corresponding to toluene, n = 3 replicates. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

x102 1.1

x102 +EI EIC(176.9770) Scan Sample005.D

+EI EIC(176.9770) Scan WhiteWine_calib10ppb.D

1.0

1.0

FEN

0.9

0.8

10 ng mL-1

0.7

FEN

0.9

0.8

64 ng mL-1

0.7

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0

0 21.2 21.4 21.6 21.8 22

22.2 22.4 22.6 22.8 23

23.2 23.4 23.6 23.8 24

24.2 24.4 24.6 24.8 25

21.2 21.4 21.6 21.8 22

22.2 22.4 22.6 22.8 23

x102 1.0

23.2 23.4 23.6 23.8 24

24.2 24.4 24.6 24.8 25

Counts (%) vs. Acquision Time (min)

Counts (%) vs. Acquision Time (min)

x102 +EI EIC(134.0993) Sample005.D 1.1

+EI EIC(134.0993) Scan WhiteWine_calib10ppb.D

IP

1.0

0.9

IP

0.8

0.9

49 ng mL-1

0.8

10 ng mL-1

0.7

0.7

0.6

0.6 0.5 0.5 0.4 0.4 0.3

0.3

0.2

0.2

0.1

0.1

0

0 18.5

18.7

18.9

19.1

19.3

19.5

19.7

19.9

20.1

20.3

20.5

20.7

Counts (%) vs. Acquision Time (min)

18.5

18.7

18.9

19.1

19.3

19.5

19.7

19.9

20.1

20.3

20.5

20.7

Counts (%) vs. Acquision Time (min)

Fig. 2. EIC chromatograms for quantification (solid line) and qualification (dotted line) ions of fenhexamide (FEN) and iprovalicarb (IP) in a non-polluted wine spiked at 10 ng mL−1 (left) and in a non-spiked wine sample (right) containing 64 and 49 ng mL−1 of FEN and IP, respectively.

allowing to record data for an unlimited number of compounds. In this vein, the combination of liquid chromatography (LC) with quadrupole (Q) TOF-MS has already demonstrated an excellent performance for the analysis of phytochemical residues [9] and the investigation of resveratrol analogues [10] in wine. Replacement of electrospray (ESI) for a harder ionization source, e.g. electron ionization (EI), is expected to increase the spectral information associated to any chromatography peak. Also, the combination of HRMS with gas chromatography (GC) allows the determination of semi-volatile, small-size molecules poorly ionized in LC–MS interfaces. The qualitative applications of GC-EI-QTOF-MS have been already evaluated during the characterization of some food commodities [11], and for the identification of pesticide residues in

vegetable food samples [12] and organic pollutants in water matrices [13]. On the other hand, quantitative features of GC-QTOF-MS remain mostly unexplored. Solid-phase extraction (SPE) is regarded as the most popular technique for extraction of natural and anthropogenic compounds from wine [14,15]. The relatively low breakthrough volumes of SPE sorbents for wine matrices (when compared with less complex environmental samples) can be compensated by further combination with dispersive liquid–liquid microextraction (DLLME), aiming not only to increase the obtained enrichment factors, but also to remove some undesired species from the primary SPE extract [16,17].

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model CHROMA-357361; No. of Pages 11

ARTICLE IN PRESS T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

3

Fig. 3. Total ion current (TIC) chromatogram (A), EIC chromatogram (342.0321 ± 0.005 Da) (B), experimental (C) and library (D) EI-HRMS spectra for a compound identified as boscalid in a non-spiked wine sample.

The aim of this research is to develop a multiclass method for the quantification of a significant set of semi-volatile compounds related to wine quality. The proposed approach involves sequential combination of SPE, as extraction technique, with DLLME concentration and final selective determination of target compounds by GC-EI-QTOF-MS. The list of selected compounds comprises a total of 27 pesticides (including fungicides, insecticides and herbicides) and 12 phenolic species considered either as natural components in wine (ethyl and vinylphenols, the ethyl ester of 4-hydroxy benzoic acid EB, and resveratrol derivatives) or as undesired anthropogenic species (bisphenol A and alkylphenols). Their expected concentrations range from the very low ng mL−1 up to a few ␮g mL−1 , depending on the compound. Also, different approaches for the post-run search of additional compounds in the recorded GC-EIHRMS chromatograms are discussed.

These four species were used as internal surrogates (IS) during sample preparation and determination steps. The list of target compounds including their CAS numbers, abbreviated names, family and octanol-water partition values (log Kow ) are summarized in Table 1. Individual solutions of each compound and the IS were prepared in methanol and maintained at −20 ◦ C. Diluted mixtures were made in the same solvent and stored at 4 ◦ C. Methanol and acetonitrile (HPLC grade) were obtained from Merck (Darmstadt, Germany). Pesticide grade acetone, toluene, 1octanol, chloroform, carbon tetrachloride and chlorobenzene were purchased from Sigma-Aldrich. Acetic anhydride was also from Sigma-Aldrich. Ultrapure water was obtained from a Milli-Q Gradient A-10 system (Millipore, Bedford, MA, USA). Solid-phase extraction (SPE) cartridges, containing 60 mg of the mixed-mode (reversed-phase and anionic exchanger) OASIS MAX sorbent, were provided by Waters (Milford, MA, USA).

2. Experimental 2.2. Samples and sample preparation 2.1. Standards, solvents and sorbents Pterostilbene (PTES, 98% purity) and cis-resveratrol (99%) were purchased from TCI Europe (Zwijndrecht, Belgium) and Cayman Chemicals (Ann Arbour, MI, USA), respectively; the rest of standards were provided by Sigma-Aldrich (St. Louis, MO, USA). Isotopically labelled compounds: d4 -methyl-p-hydroxybenzoate (d4 -MB), d4 4-n-nonylphenol (d4 -NP), 13 C6 -trans-resveratrol (13 C6 -transRES) and 13 C6 -metalaxyl (13 C6 -MET) were also obtained from Aldrich.

Red and white wine samples, proceeding from different geographic areas in Spain, were purchased in local markets and stored at room temperature for a maximum of 1 month before analysis. Sample preparation conditions were adapted from previous works, dealing with DLLME of pesticides and off-flavours from wine [16,17], and finely tuned to obtain the best overall performance for the range of compounds involved in this study. Unless different values are specified, sample preparation was optimized with aliquots

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

Family

Log Kow

Retention time (min)

b,c

Selected ions (m/z)

Mass errors (mDa, SD)

4-ethylphenol 4-ethylcatechol 4-ethylguaiacol 4-vinylphenol 4-vinylguaiacol 4-ethyl-p-hydroxybenzoate 4-n-octylphenol 4-n-nonylphenol Bisphenol A trans-pterostilbene cis-resveratrol trans-resveratrol Pyraclostrobin Iprovalicarb Benalaxyl Metalaxyl Triadimenol Fenhexamid Myclobutanil Pyrimethanil Triadimefon Cyprodinil Flusilazole Penconazole Tebuconazole Propiconazole Diniconazole Procymidone Prochloraz Iprodione Difenoconazole Azoxystrobin Chlorpyrifos ethyl Teflubenzuron Chlorpyrifos methyl ␣-endosulfan ␤-endosulfan Endrin Trifluralin d4 -methyl-p-hydroxybenzoate d4 -4-n-nonylphenol 13 C6 -trans-resveratrol 13 C6 -metalaxyl

123-07-9 1124-39-6 2785-89-9 2628-17-3 7786-61-0 120-47-8 27193-28-8 25154-52-3 80-05-7 537-42-8 61434-67-1 501-36-0 175013-18-0 140923-17-7 71626-11-4 70630-17-0 55219-65-3 126833-17-8 88671-89-0 53112-28-0 43121-43-3 121552-61-2 85509-19-9 66246-88-6 107534-96-3 60207-90-1 83657-24-3 32809-16-8 67747-09-5 36734-19-7 119446-68-3 131860-33-8 2921-88-3 83121-18-0 5598-13-0 959-98-8 33213-65-9 72-20-8 1582-09-8 362049-51-2 1173019-62-9 1185247-70-4 1356199-69-3

EP EC EG VP VG EB OP NP BPA PTES cis-RES trans-RES PY IP BEN MET TR FEN MYC PYR TRI CYP FLU PEN TEB PR DI PRY PRC IPR DIF AZO CLE a TEF CLM ␣-END ␤-END EDR TRF d4 -MB d4 -NP 13 C6 -transRES 13 C6 -MET

Aroma

2.58 1.84 2.43 2.61 2.57 2.39 5.30 5.74 3.64 4.06 3.02 3.02 3.13 3.56 3.87 2.15 2.97 3.85 3.07 2.55 2.73 4.00 3.83 3.66 3.58 3.64 4.20 2.93 4.59 3.10 4.00 5.10 5.00 5.01 3.98 2.60 2.31 4.48 4.55 – – – –

6.86 11.42 9.71 7.23 10.09 11.26 15.47 16.74 21.22 24.11 23.28 26.42 26.73 19.44, 19.67 20.94 16.24 18.18 22.75 19.51 14.61 17.18 17.69 19.57 17.91 21.44 20.72, 20.87 20.21 18.30 24.78 22.04 26.82, 26.89 27.89 17.10 6.75 15.89 18.74 20.08 19.85 13.07 10.07 16.74 26.27 16.24

108.0547, 107.0508, 122.0729 139.0731, 138.0698, 123.0418 153.0885, 152.0857, 137.0617 121.0618, 120.0587, 91.0553 151.0748, 150.0700, 135.0461 167.0681, 166.0651, 121.0311 206.1702, 107.0508 220.1840, 107.0495 228.1150, 213.0915 256.1121, 298.1226 230.0870, 228.0808, 312.1019 230.0870, 228.0808, 312.1019 132.0444, 283.0419 134.0993, 116.0714 148.1132, 176.1081 160.1128 206.1195 168.1129, 128.0028 176.9770, 266.0942 179.0367 150.0104 198.1034 183.0787 208.0278, 128.0019 224.1188, 210.1031 233.0627, 206.0543 248.0954, 158.9760 250.0740, 125.0149 259.0283, 172.9565 268.0044, 232.0278 283.0167, 96.0568 308.0012, 180.1137 314.0099, 186.9581 323.0227, 264.9823 344.1035, 388.0944 196.9202, 313.9572 222.9407, 159.976 285.9255, 124.9826 236.8408, 338.8720 236.8408, 338.8719 262.8549,244.9507 264.0222, 306.0703 126.0568 224.2072 234.0993 166.1323

1.5, 1.3, 1.0 1.7, 1.6, 9.8 1.7, 1.7, 1.6 1.4, 1.4, 1.2 1.6, 1.7, 1.6 1.8, 1.8, 1.5 2.0, 1.3 2.1, 1.3 2.1, 2.0 2.3, 2.7 1.4, 2.4, 2.7 1.7, 2.1, 2.5 1.4, 2.5 1.0, 1.4 1.6, 1.8 2.2, 2.3 2.0, 1.4 1.8, 2.4 1.7, 1.4 1.9, 1.7 2.2, 1.6 2.1, 2.2 2.3, 1.7 1.9, 1.8 2.4,1.4 2.0, 1.6 2.5, 2.2 2.5, 1.2 2.7, 3.0 2.6, 3.7 4.6, 2.0 3.6, 3.8 3.5, 2.3 2.1, 1.6 2.6, 1.4 1.6, 1.7 2.1, 1.7 2.4, 2.1 2.4, 2.7 1.4 2.1 2.0 1.0

a b c

Retention time and masses for the isocyanate degradation product. Data in bold type correspond to quantification ions. Data in italics correspond to the base peak.

Anthropogenic pollutants Stilbenes

Fungicides

Insecticides

Herbicide Internal surrogates

ARTICLE IN PRESS

Abbreviation

G Model

CAS number

T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

Compound

CHROMA-357361; No. of Pages 11

4

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

Table 1 Summary of compounds considered in this study.

G Model CHROMA-357361; No. of Pages 11

ARTICLE IN PRESS T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

5

Fig. 4. Results obtained with the Find by Formula search of dihydro-resveratrol (C14 H14 O3 ) in a red wine (Tempranillo variety) processed using the developed method. A, EIC chromatogram for the molecular ion (mass window 5 mDa). B, experimental and predicted (boxes in red) spectra for the molecular ion (M.+ ) of the candidate peak. C, Experimental EI-HRMS spectrum of the candidate peak with mass errors. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

of a red wine (Mencía variety) spiked with analytes at 50 ng mL−1 , except resveratrol isomers and 4-ethylphenol derivatives already present, at significant levels, in this matrix. The mixture of the four IS was added to wine samples at a concentration of 50 ng mL−1 . Under final working conditions, 2 mL of unfiltered wine were diluted with the same volume of ultrapure water and passed through a SPE cartridge. The cartridge was rinsed with 2 mL of a 12% (w/v) ethanolic solution and dried for 15 min with a stream of nitrogen. Compounds were recovered with 2 mL of methanol. A fraction of this extract (1 mL) was transferred to conical bottom, 10 mL volume glass vessel containing 8 mL of a 5% (w/v) K2 HPO4 aqueous solution. Acetylation of phenolic compounds (including the fungicide fenhexamide, FEN) was achieved by addition of 0.050 mL of acetic anhydride and manual shaking for 5 min [17]. Thereafter, a binary mixture of acetone: toluene (0.5 mL of acetone plus 0.06 mL of toluene) was added using a gas tight syringe. The organic phase of toluene (0.045 mL) was recovered after manual shaking (1 min) and centrifugation (5 min, 3500 rpm).

2.3. Determination conditions Compounds, as native species or acetylated derivatives, were determined using a GC-QTOF-MS instrument comprised of a 7890A gas chromatograph and a 7200 QTOF spectrometer, both acquired from Agilent (Wilmington, DE, USA). The MS system was operated in the single MS mode, using EI at 70 eV. Full scan MS spectra, in the range of m/z values from 70 to 650 Da, were recorded at a frequency of 2.5 Hz (4 spectra s−1 ). The TOF mass analyser was operated in the 2-GHz extended dynamic range resolution mode. Mass resolution

varied from 4900 at m/z 131–7500 at m/z 414. Automated recalibration of the mass axis was carried out every 5 injections, by infusion of PFTBA in the EI source. Compounds were separated in BP-5MS type capillary column (30 m × 0.25 mm i.d., 0.25 ␮m film thickness) acquired from Agilent. Helium was used as carrier gas at a constant flow of 1.2 mL min−1 . The column temperature was programed as follows: 90 ◦ C (1 min) rated at 8 ◦ C min−1 to 300 ◦ C (15 min). The temperatures of the transfer line and the EI source were set at 280 and 230 ◦ C, respectively. Injections (1 ␮L) were made in the pulsed splitless mode (25 Psi, 1.1 min) with the injection chamber set at 300 ◦ C. The splitless time and the split flow were 1 min and 60 mL min−1 , respectively. The solvent delay was fixed at 6 min.

2.4. Performance of the quantitative method Efficiencies of the two main steps (SPE extraction and DLLME concentration) involved in the sample preparation process were assessed with spiked samples of different wines. Selective extracted ion current (EIC) chromatograms for target compounds were obtained using a mass window of 5 mDa centred in the m/z values of selected ions from their experimental EI-HRMS spectra. The yield of SPE extraction was calculated as the ratio between responses obtained for samples spiked before and after the SPE step, submitted to identical derivatization and DLLME concentration conditions. The extraction efficiencies (EEs, %) of the DLLME step were evaluated following an indirect method using liquidliquid extraction (LLE) of aqueous acetylated solutions (spiked SPE extracts plus 8 mL of a 5% K2 HPO4 solution and 0.05 mL of

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model

ARTICLE IN PRESS

CHROMA-357361; No. of Pages 11

T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

6

x101 +EI TIC Scan Sample007.D 4.5

A

4.0

TIC

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 8.0

8.5

9.0

9.5

10.0

10.5

11.0

Counts (%) vs. Acquision Time (min)

11.5

12.0

Find by Formula Cpd 3: C8 H10 O2: +EI EIC(138.0675) Scan Sample007.D x102 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

8 7 6 Cpd 4: C8 H10 O2: + FBF Spectrum (rt: 7.991-8.031 min) Sample007.D Subtract

5

138.0689 [C8H10O2]+

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0

4 3 2

B

10.673

Cpd 3: C8 H10 O2: + FBF Spectrum (rt: 10.641, 10.701 min) Sample007.D Subtract

138.0705 [C8H10O2]+

Score 94.7% 101 139.0712 [C8H10O2]+ 137.5

138.0

138.5

139.0

139.5

140.0704 [C8H10O2]+

2.0

140.0

1.5

140.5

Counts (%) vs. Mass-to-Charge (m/z)

Score 94.6%

Cpd 2: C8 H10 O2: + FBF Spectrum (rt: 11.284-11.337 min) Sample007.D Subtract

138.0700 [C8H10O2]+

2.5

Score 94.8%

1.0 139.0747 [C8H10O2]+

0.5 139.0670 [C8H10O2]+ 137.5

138.0

138.5

139.0

139.5

140.0670 [C8H10O2]+ 140.0

140.0757 [C8H10O2]+

0 137.5

138.0

138.5

139.0

139.5

140.0

140.5

Counts (%) vs. Mass-to-Charge (m/z)

140.5

Counts (%) vs. Mass-to-Charge (m/z)

1

11.411 8.011

0 8.0

8.5

9.0

9.5

10.0

10.5

11.0

11.5

12.0

Counts (%) vs. Acquision Time (min)

Spectral deconvoluon

C x102 +EI Compound Spectrum (rt: 11.250-11.456 min) Sample007.D

x101 Cpd 7: 8.011: +EI Compound Spectrum (rt: 7.977-8.058 min) Sample007.D 95.0508

9

109.0672

7 6

138.0701

1.0

Unidenfied

8

EC

0.8 123.0470

83.0502

0.6

5

138.0689

4 165.1300

124.0893

3 2

152.0845

1

0.4 193.1245 0.2

180.1533

100

120

140

160

180

200

109.0770

0

0 80

180.0810

79.0550

208.1478

80

220

90 100 110 120 130 140 150 160 170 180 190 200 Counts (%) vs. Mass-to-Charge (m/z)

Counts (%) vs. Mass-to-Charge (m/z)

x102

+EI Compound Spectrum (rt: 10.614-10.735 min) Sample007 138.0708

1.0 0.8

p-tyrosol

0.6 108.0561

0.4 0.2 77.1178

120.0590

0 80

180.0832

90 100 110 120 130 140 150 160 170 180 190 200 Counts (%) vs. Mass-to-Charge (m/z)

Fig. 5. A, detail of the TIC chromatogram for a red wine. B, results obtained using the FBF function when searching for species C8 H10 O2 . C, Full scan, deconvoluted EI-HRMS spectra of found peaks with assigned identities. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

acetic anhydride) with 5 mL of toluene. EEs were calculated as: EE(%) = (1 − Aes /Ans )*100, being Aes and Ans the responses for each compound in the LLE extracts from solutions submitted (Aes ) and not submitted (Ans ) to DLLME concentration, respectively. The enrichment factors (EFs) provided by the DLLME step were defined as: EF = (1/0.045) xEE, with 1 and 0.045 mL representing the vol-

umes of SPE extract, used in the DLLME step, and the recovered volume of the toluene phase. Linearity was evaluated with samples spiked at 8 different concentrations, containing a fixed level (50 ng mL−1 ) of the IS mixture. The ratios between peak areas for each compound and the corresponding IS were plotted versus the added concentration and fitted

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model CHROMA-357361; No. of Pages 11

ARTICLE IN PRESS T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

to a linear model. Procedural blanks were prepared with aliquots of synthetic wine (2 mL) submitted to the whole sample preparation process. Identification of target compounds in non-spiked wines was established on the basis of (1) retention time match with standards (allowed difference ± 0.05 min) and (2) presence of at least two characteristic ions (mass window 5 mDa) in their spectra. Their concentration levels were established by pseudo-external calibration, by comparison with responses obtained for spiked aliquots of a white wine sample. Overall recoveries were calculated as the ratio between concentrations measured for spiked a non-spiked aliquots (n = 3 replicates) of each sample divided by the added concentration and multiplied by 100. 2.5. Screening of novel compounds The post-run screening capabilities of the GC-QTOF-MS system were evaluated using two different strategies. The first involved the use of a EI-HRMS spectral library of 650 pesticides provided by Agilent. A group of phytochemicals, amenable to GC analysis and approved to be used in vineyards, was selected from this data base. The m/z value for a characteristic ion (usually the base peak or the molecular ion) in their spectra was extracted (mass window 5 mDa) in the chromatograms of the processed wines. Then, spectra for peaks in the EIC chromatograms were compared with those in the EI-HRMS library. A minimum of three common ions (maximum differences 5 mDa), displaying similar intensity ratios, was required for a tentative identification. Final confirmation was attained by injection of authentic standards. A second strategy was tested for the search of phenolic species not considered in the list of target compounds (Table 1), although known to be potentially present in some wines. Given that the typical EI-MS fragmentation pattern of acetylated phenols corresponds to replacement of the CH3 CO moiety by hydrogen, chromatograms were explored for the molecular formula of native phenols (that is • the cluster of ions for the M + ion) using the Find by Formula (FBF) function implemented in the Mass Hunter software. Again, the EIHRMS spectra of candidate peaks were examined for compatibility with the expected for search compounds using, when available, the fragmentation pattern contained in the low resolution EI-MS NIST database. 3. Results and discussion 3.1. Compounds identification Table 1 summarizes retention times, representative ions and an evaluation of the variability of their m/z values for target analytes in consecutive injections (n = 5 replicates) of the DLLME extract from a spiked (20 ng mL−1 ) red wine sample. A single peak was noticed for the acetylated derivatives of the 12 phenolic species and the fungicide FEN. The rest of pesticides rendered also a single peak, with the exceptions of iprovalicarb (IP), propiconazole (PR), difenoconazole (DIF) and triadimenol (TR). The chromatograms for the first 3 species showed two peaks, corresponding to their diastereomers, with similar intensities. The sum of their peak areas was used as response variable. In the case of TR only the response for the most intense peak was considered for quantification purposes. The peak obtained for teflubenzuron (TEF) corresponds to the isocyanate generated in the injector of the GC due to thermal degradation of the precursor insecticide [18]. In case of anthropogenic species, whose concentrations in wine remain in the low ng mL−1 level, the ions selected for quantitative and qualitative purposes were the base peak and the 2nd most intense signal in their EI-MS spectra, Table 1. For natural origin compounds with expected concentrations in the medium, or even

7

the high, ng per mL range (ethyl and vinylphenols, EB and resveratrol isomers), a 3rd ion, displaying a relatively low intensity, was used for quantification. This ion is the M + 1 (due to natural abundance of 13 C), or even M + 2, species in the cluster of the base peak. In most cases, mass errors remained below 3 mDa for all the considered ions. Thus, a mass window of 5 mDa was used for extraction of selective EIC chromatograms and re-calibration of the m/z axis was accomplished every 5 injections. 3.2. Sample preparation optimization 3.2.1. SPE conditions In order to prevent phase separation problems during DLLME of wine adjusted at basic pH [17], samples were passed through an OASIS MAX sorbent aiming to remove very polar and acidic compounds, e.g. monosaccharides, salts, poly-hydroxylated species and carboxylic acids, while analytes were recovered in the extract [19]. Elution solvents were selected on the basis of their selectivity, recoveries of target analytes and compatibility with the further DLLME step. Acetonitrile provided colourless extracts for red and white wines; however, it failed to recover most of the phenolic species, data not given. Acetone and methanol rendered reddish, although transparent, extracts in case of red wine samples; nevertheless, phase separation problems were not noticed during the further DLLME step. Recovery data indicate that acetone failed to extract resveratrol isomers, also it provided lower recoveries for some compounds (e.g. 4-ethylcatechol, EC; bisphenol A, BPA; and 4-nonylphenol, NP) than methanol, see Table S1. Thus, methanol was used for cartridges elution. 3.2.2. DLLME parameters Previous applications have reported the combination of acetylation and DLLME for phenolic compounds concentration in wine matrices [17,20]; thus, variables related to the derivatization reaction were set at values indicated in Section 2.2, and only those parameters controlling the efficiency of DLLME concentration were optimized. Five extractant solvents were tested. Chloroform, carbon tetrachloride and chlorobenzene settled at the bottom of the conical shaped DLLME tube after centrifugation. On the other hand, 1octanol and toluene remained floating above the aqueous phase. Fig. S1 shows a picture of the DLLME extracts obtained with the above solvents (0.1 mL in all cases combined with 0.5 mL of acetone as dispersant) for a red wine (Mencía variety) after SPE. Chloroform extracts presented an intense reddish colour; thus, this solvent was not further considered. The normalized responses, for a group of selected compounds, obtained with the other 4 solvents are plotted in Fig. 1. Depicted data correspond to peak areas normalized to those measured for toluene, without any correction for differences between the volumes of the final extract (from 74 to 85 ␮L, depending on the extractant). The rest of analytes followed the same trend as those included in Fig. 1. Toluene rendered the highest responses for most compounds, thus, it was selected as extractant. A reduction in the volume of toluene from 0.1 to 0.06 mL improved the responses of all compounds (data not shown), providing enough extract volume (0.045 mL) to be handle with the autosampler of the GC–MS system. The type of dispersant (acetone, acetonitrile and methanol) played a non significant influence on the yield of the DLLME; however, acetone was preferred since it allowed a better separation between the aqueous phase and the floating drop of toluene. Fig. S2 shows the effects of the volume of acetone (from 0.5 to 2 mL) in the responses for a selection of analytes. Increasing the volume of dispersant resulted in a diminution of responses measured for polar compounds (e.g. metalaxyl (MET) and pyrimethanil (PYR) with logKow values of 2.15 and 2.55, respectively); on the other

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model CHROMA-357361; No. of Pages 11

ARTICLE IN PRESS T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

8

hand, highly lipophilic species (such as chlorpyrifos ethyl, CLE, logKow 5.00), displayed increasing responses with the volume of dispersant. For most compounds, this variable exerted a negligible effect in the performance of the DLLME concentration. The adopted decision was set the volume of acetone at 0.5 mL. Table S2 summarizes the EEs (%) and the average EFs provided by the optimized DLLME procedure for two different wine matrices. For most compounds, EEs (%) remained above 70% resulting in EFs higher than 15 times. The lowest efficiency during the DLLME step (c.a. 30%) was observed for MET, which is the most polar within the group of compounds (excluding those undergoing acetylation previously to DLLME concentration) compiled in Table 1. 3.3. Performance of the method Mass accuracy, sensitivity and linear response range of TOF MS analysers depend on the number of transients per spectrum. In this study, spectra were recorded at 2.5 Hz (4 spectra s−1 , with 3400 transients per spectrum) as a compromise between sensitivity and number of spectral points per chromatographic peak. Under these conditions, the response of analog to digital converter (ADC) detector was saturated for signals in the range between 107 –108 counts. Signal saturation turns in poor mass accuracy and non-linear responses, preventing the quantification of compounds whose concentrations stay in the high ng mL−1 range. These drawbacks were overcome by selecting a low intensity ion from their EI-HRMS spectra. Fig. S3 illustrates this situation for a white wine spiked with EP in the range from 5 to 1000 ng mL−1 . A poor linearity was observed using responses measured for the base peak (107.0508 ± 0.005 Da) in the scan spectrum of EP; however, when selecting the minor intensity 108.0547 ± 0.005 Da ion, from same chromatographic injections, an excellent linearity was noticed, Fig. S3. The determination coefficients (R2 ) obtained for red (Tempranillo variety) and white wines (mixture of Treixadura and Torrontés varieties) spiked at eight increasing concentration levels (1–200 ng mL−1 for anthropogenic compounds and PTES; 10–2000 ng mL−1 resveratrol isomers and 5–1000 ng mL−1 rest of species) are compiled in Table 2. In most cases, R2 values stayed above 0.99. The ratios between slopes of addition curves for red and white wines varied between 0.86 and 1.18, which points out to small changes in the efficiency of the overall method (including sample preparation and determination steps) for both matrices. A clear exception of the above behaviour was noticed for TEF with a slope ratio of 0.46. Likely, this data indicates that the yield of TEF degradation is affected by the sample matrix. The accuracy of the method was investigated with aliquots of the same wines used for linearity assessment (Table 2), spiked at two different concentrations. The low spiked level represented 10 ng mL−1 for anthropogenic species, 50 ng mL−1 for ethyl, vinylphenols and EB; and 200 ng mL−1 for stilbenes. The high level was 5-fold above concentrations. Concentrations for spiked and non-spiked (n = 3 replicates) fractions of each wine sample were established by comparison with a calibration curve obtained for another white wine (Ribera de Duero, Viura variety) selected on the basis of its low contents of anthropogenic compounds. Most recoveries were in the range from 70 to 120%, with standard deviation values below 10%, Table 2. In the case of the low level spiked samples, recoveries could not be calculated for some species such as MET and IP in the white wine, and resveratrol isomers in red wine. The reason is that added levels represent only around 10% of native concentrations of these species. For TEF, the pseudo-external calibration method provided unacceptable recoveries, particularly for red wines, Table 2. The LOQs of the method, defined for a signal to noise (S/N) of 10, were estimated from the lowest addition level in the linearity

study (from 1 to 10 ng mL−1 ), considering also potential contamination problems evaluated through the analysis of procedural blanks. BPA was the only species noticed in blanks. Taking into account blanks stability, a LOQ of 0.3 ng mL−1 was set for this compound. LOQs of pesticides ranged from 0.2 to 1.4 ng mL−1 . For EB, phenolic off-flavours and stilbenes, LOQs were estimated from responses obtained for spiked aliquots of synthetic wine. The obtained values ranged from 1 to 5 ng mL−1 , Table 2. 3.4. Levels of target compounds in wine samples A total of 25 wines, produced between years 2011 and 2014, were processed in duplicate. Table 3 summarizes the range of values and the average concentrations for natural origin phenolic species in these samples. For this group of analytes, large differences were noticed between red and white wines. In the first case, ethylphenols (EP, EC and EG) were found at much higher concentrations than vinylphenols (VP and VG), whereas the opposite situation was noticed for white wines. In several samples, ethyland vinylphenol levels stayed above their odour thresholds [21]. Resveratrol isomers were also present at much higher concentrations in red wines versus white ones. It is also worth noting the ubiquity of EB in wine samples. This compound, usually referred as ethyl paraben, is added as preservative and antioxidant to canned foods, personal care products and pharmaceuticals; however, its presence in wine has a natural origin [22]. Table 4 compiles the concentrations of anthropogenic species in the same wines. Alkylphenols, insecticides and herbicides remained undetected in all samples. On the other hand, several fungicides could be quantified with some anti-mildium (IP and MET) and several anti-botrytis agents (FEN, PYR, CYP and IPR) displaying occurrence frequencies around, or even above, 50%. Fig. 2 shows the EICs chromatograms for FEN and IP in a non-polluted wine, spiked with both compounds at 10 ng mL−1 (left), and in a highly polluted, non-spiked wine (right). In both cases, similar ratios were observed between quantification and qualification ions. Also, the selectivity resulting from using a mass extraction window of 5 mDa can be appreciated in this figure. BPA was also often found in wine samples; however, the quantified levels usually remained around or below 10 ng mL−1 , Table 4. Despite the apparent high residues of some fungicides compiled in Table 4, they remained below 10% of the MRLs established for wine grapes in the European Union (e.g. 1 ␮g g−1 for MET) [23]. 3.5. Post-run screening of additional targets Table 5 compiles a list of characteristic ions obtained from EIHRMS library spectra for a selection of 18 pesticides not considered during method development. These ions were extracted (mass window 5 mDa) from TIC chromatograms of the processed samples and spectra for peaks in the selective EIC chromatograms were compared with those existing in the EI-HRMS library. Fig. 3 shows the chromatograms and spectrum obtained during the post-run target search of boscalid in a non-spiked white wine. In this case, the EIC plot (m/z 342.0321 ± 0.005 Da) showed a single peak; moreover, its spectrum contained several ions matching those existing in the EI-HRMS library for the candidate compound. When coelution problems are suspected, spectral deconvolution [12] is advisable before library search. Four out of the 18 pesticides included in Table 6 were first tentatively identified, and further confirmed with authentic standards, in at least one of the 25 processed wine samples. Fig. 4 shows the results obtained applying the FBF function to dihydro-resveratrol (empirical formula C14 H14 O3 ) search. This stilbene has been identified in wine samples using different tech-

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model

ARTICLE IN PRESS

CHROMA-357361; No. of Pages 11

T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

9

Table 2 Performance of the developed method for wine samples. Compound

IS

2

EP EC EG VP VG EB OP NP BPA PTES cis-RES trans-RES PY IP BEN MET TR FEN MYC PYR TRI CYP FLU PEN TEB PR DI PRY PRC IPR DIF AZO CLE TEF CLM ␣-END ␤ −END EDR TRF

d4 -MB

d4 -NP

13

C6 -transRES

13

C6 -MET

2

LOQs (ng mL−1 )

Recoveries (%) ± SD (n = 3 replicates)

Linearity evaluation R , red wine

R , white wine

a

0.982 0.989 1.000 0.995 0.984 1.000 0.999 0.998 0.998 0.999 0.997 0.999 0.985 1.000 0.996 0.999 0.999 1.000 1.000 0.995 1.000 0.997 1.000 0.999 1.000 0.999 0.999 1.000 0.994 0.999 0.997 1.000 1.000 0.999 1.000 1.000 0.998 0.999 0.999

0.989 0.994 0.992 0.982 0.993 0.999 0.999 1.000 0.998 1.000 0.999 0.998 0.995 0.996 1.000 0.996 0.999 0.998 0.998 1.000 0.996 1.000 0.998 0.997 0.996 0.995 0.994 0.997 0.983 0.997 0.988 0.996 0.999 0.992 0.999 0.999 0.998 0.998 0.997

1.04 1.02 1.18 1.15 1.12 1.04 0.88 0.94 0.88 1.03 0.90 0.90 0.88 0.97 0.96 0.88 0.91 1.03 0.98 0.98 1.00 0.97 0.97 0.95 1.00 1.01 0.96 0.97 0.95 0.99 1.05 0.93 0.93 0.46 0.93 0.86 0.92 0.86 0.99

Slopes ratio

White wine (Low level)

White wine (high level)

Red wine (Low level)

Red wine (High Level)

99 ± 9 70 ± 4 90 ± 5 n.e. 89 ± 3 100 ± 2 100 ± 3 93 ± 4 98 ± 10 106 ± 7 100 ± 2 110 ± 8 94 ± 6 n.e. 104 ± 8 n.e. 100 ± 6 103 ± 12 105 ± 8 100 ± 8 92 ± 9 95 ± 8 109 ± 10 95 ± 9 92 ± 9 98 ± 9 96 ± 12 99 ± 8 95 ± 11 90 ± 8 105 ± 16 95 ± 14 95 ± 11 139 ± 13 96 ± 11 116± 9 109 ± 9 104 ± 12 102 ± 16

105 ± 4 100 ± 1 97 ± 6 96 ± 3 115 ± 1 110 ± 1 110 ± 5 101 ± 5 121 ± 8 111 ± 5 90 ± 2 101 ± 2 95 ± 1 112 ± 4 104 ± 5 110 ± 2 114 ± 7 97 ± 4 107 ± 3 86 ± 10 82 ± 5 95 ± 3 104 ± 4 94 ± 2 97 ± 1 95 ± 5 96 ± 3 98 ± 2 80 ± 7 88 ± 4 81 ± 9 86 ± 9 93 ± 6 124 ± 21 96 ± 2 116 ± 6 104 ± 6 107 ± 8 91 ± 9

100 ± 2 95 ± 4 120 ± 4 80 ± 3 86 ± 2 105 ± 3 92 ± 3 100 ± 3 114 ± 4 130± 3 n.e. n.e. 107 ± 4 90 ± 2 104 ± 1 106 ± 2 83 ± 2 104 ± 3 91 ± 3 103 ± 3 97 ± 2 92 ± 1 94 ± 2 90 ± 1 113 ± 1 105 ± 1 96 ± 1 88 ± 1 77 ± 3 89 ± 3 97 ± 3 84 ± 2 94 ± 8 39 ± 4 101 ± 1 95 ± 8 96 ± 2 90 ± 5 89 ± 6

83 ± 6 88 ± 2 92 ± 3 71 ± 7 92 ± 4 101 ± 2 102 ± 3 96 ± 1 106 ± 1 108 ± 3 119 ± 2 108 ± 5 97 ± 7 105± 6 93 ± 6 93 ± 7 101 ± 7 106 ± 5 93 ± 5 112 ± 8 101 ± 4 90 ± 6 99 ± 6 88 ± 4 114 ± 6 105 ± 6 114 ± 5 83 ± 7 117 ± 6 84 ± 5 96 ± 6 119 ± 5 108± 4 41 ± 10 100 ± 6 93 ± 4 81 ± 4 87 ± 5 91 ± 7

1 5 5 5 5 2 0.4 0.4 0.3 0.3 5 5 0.4 1 0.3 1 0.4 0.4 0.5 0.5 0.7 0.2 0.5 1.2 0.7 0.5 0.7 0.3 1.4 0.7 1.0 1.0 0.6 0.5 0.3 0.6 0.6 0.5 0.4

n.e., not evaluated. a Ratios of red wine/white wine slopes. Table 3 Average concentrations (ng mL−1 ) and range of values of phenolic compounds in wine samples. Odour threshold (ng mL−1 )

Compound

Red wine (n = 11 samples)

White wine (n = 14 samples)

Averagea (positive samples)

Range

Averagea (positive samples)

Range

EP EC EG VP VG EB PTES cis-RES trans-RES

150 (10) 35 (6) 35 (6) 22 (4) n.d. 68 (11) 2 (6) 1222 (11) 937 (11)

n.d.−493 n.d.−51 n.d.−55 n.d.−50 n.d. 21–171 n.d.−5 123–2300 25–2074

2 (6) n.d. 7 (6) 208 (14) 85 (14) 19 (13) n.d. 84 (12) 52 (11)

n.d.−16 n.d. n.d.−10 15–533 6–243 n.d.−52 n.d. n.d.−346 n.d.−374

440 440 33 180 40 n.a. n.a. n.a. n.a.

n.d., not detected. n.a., not available. a Average value for samples above the LOQ.

niques, including LC-QTOF-MS [10]. The FBF function explores the TIC chromatogram for cluster of peaks compatible with the molec• ular ion (M + ) of the selected formula, extracting its EIC plot and displaying the corresponding FBF spectrum. Despite several peaks appear in the EIC chromatogram (Fig. 4A), only that at 23.63 min present a cluster of ions compatible with the above formula. Fig. 4B shows the experimental FBF spectrum and the predicted one (boxes in red) for the M+• cluster. A normalized score of 85% (from 0 to 100)

was obtained considering mass accuracy, isotopic pattern and spacing between ions in this cluster. Dihydro-resveratrol is not included in the NIST low resolution library; thus, the experimental full scan spectrum of the peak at 23.63 min was examined for compatibility with the proposed structure. Differences between calculated and experimental masses of fragment ions remained below 5 mDa, Fig. 4C.

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model

ARTICLE IN PRESS

CHROMA-357361; No. of Pages 11

T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

10

Table 4 Number of positive samples for anthropogenic compounds and range of concentrations in wine. Compound

BPA IP BEN MET TR FEN MYC PYR CYP TEB PRY IPR AZO a

Red wine (n = 11 samples)

White wine (n = 14 samples)

Positive samples

<10 ng mL−1

10–30 ng mL−1

>30 ng mL−1

Positive samples

<10 ng mL−1

10–30 ng mL−1

>30 ng mL−1

7 3 3 5 4 2 0 3 5 4 1 9 5

5 3 3 3 4 1 0 3 5 4 1 6 5

2 0 0 2 0 0 0 0 0 0 0 1 0

0 0 0 0 0 1 (35)a 0 0 0 0 0 2 (52)a 0

10 10 3 12 6 9 3 10 10 3 4 11 1

10 4 3 3 6 6 3 6 9 3 4 6 1

0 2 0 2 0 0 0 1 1 0 0 2 0

0 4 (49)a 0 7 (97)a 0 3 (64)a 0 3 (82)a 0 0 0 3 (73)a 0

Maximum measured concentrations of each compound are given within parenthesis, when above 30 ng mL−1 .

Table 5 Database of accurate masses (m/z values) screened in the 25 processed wine samples. Compound

CAS number

EI-HRMS Library m/z values

Positive samples

Retention time (min)

Ametoctradin Boscalid Captan Dicofol Dimethomorph Famoxadone Fenamidone Fenarimol Fludioxonil Fluquinconazol Folpet Kresoxim-methyl Metrafenone Quinoxyfen Thiabendazole Trifloxystrobin Vinclozolin Zoxamide

865318-97-4 188425-85-6 133−06-2 115-32-2 110488-70-5 131807-57-3 161326-34-7 60168-88-9 131341-86-1 136426-54-5 133-07-3 143390-89-0 220899-03-6 124495-18-7 148-79-8 141517-21-7 50471-44-8 156052-68-5

246.17133 342.03214 78.94038 138.99452 301.06259 330.13629 238.11006 138.99450 248.03918 340.03958 259.93344 206.08118 395.03174 237.05844 201.03552 172.03687 284.99539 186.97168

0 7 0 0 5 0 0 0 11 0 0 0 0 0 0 0 0 1

– 25.45 – – 27.79, 28.16 – – – 19.16 – – – – – – – – 21.41

Application of the FBF function for post-run identification of phenolic species in wines has also some limitations. The lower the molecular weight of the searched formula, the higher the possibility of finding a fragment ion in the EI-HRMS spectrum of a larger molecule fitting that formula. Also, phenolic compounds with different structures and the same empirical formula exist in wine. Thus, inspection of the EI-HRMS scan spectra of candidate peaks is mandatory. Fig. 5 illustrates the above possibilities with the results obtained when searching for compounds with a formula C8 H10 O2 (e.g. EC). The FBF function extracted three compounds from the relative complex TIC chromatogram (Fig. 5A) of a red wine. The FBF spectra of these three peaks matched that expected for the formula C8 H10 O2 , with normalized scores above 90%. Upon inspection of their deconvoluted full scan spectra (Fig. 5C), it became obvious that the peak at 8.01 min does not correspond to the searched phenol but to a higher molecular weight compound. Peaks at 10.67 and 11.41 min are compatible with the searched empirical formula (C8 H10 O2 ). Using the information provided by fragment ions, the peak at 10.67 was tentatively identified, and finally confirmed, as p-tyrosol (CAS number 501-94-0), whereas the peak at 11.41 was EC.

mL were attained for pesticides. Linear responses for natural origin compounds, with expected concentrations in the medium and high ng per mL range, are guaranteed by selecting a low intensity ion from their EI-HRMS spectra. The combination of narrow-mass extraction windows (typically 5 mDa) with retention times data permits the high selective detection of target compounds. Accurate EI-HRMS, scan spectra allow the post-run screening of novel compounds, without retention time information, when rendering the molecular ion in the EI source and/or when exact m/z ratios for fragment ions are known. Final comparison of spectra for candidate peaks with those in HR libraries, or detailed interpretation when libraries are not available, increase the probability of positive identifications.

Acknowledgements This study has been financially supported by the Spanish Government, Xunta de Galicia and E.U. FEDER funds (projects CTQ2015-68660-P and GRC2013-020). T.R.C. acknowledges a FPI fellowship to the Spanish Government.

4. Conclusions Appendix A. Supplementary data For the first time, quantitative and screening possibilities of GC-EI-HRMS for the multiclass determination of semi-volatile compounds in wine have been evaluated. Following previous concentration and acetylation steps, LOQs at the very low ng per

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2016.03. 005.

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005

G Model CHROMA-357361; No. of Pages 11

ARTICLE IN PRESS T. Rodríguez-Cabo et al. / J. Chromatogr. A xxx (2016) xxx–xxx

References ˜ J. Cacho, Volatile composition and sensory [1] M. Vilanova, A. Escudero, M. Grana, properties of North West Spain white wines, Food Res. Int. 54 (2013) 562–568. [2] H. Hopfer, J. Nelson, S.E. Ebeler, H. Heymann, Correlating wine quality indicators to chemical and sensory measurements, Molecules 20 (2015) 8453–8483. [3] R. Zamora-Ros, M. Urpi-Sarda, R.M. Lamuela-Raventos, M.A. Martinez-Gonzalez, J. Salas-Salvado, F. Aros, M. Fito, J. Lapetra, R. Estruch, C. Andres-Lacueva, High urinary levels of resveratrol metabolites are associated with a reduction in the prevalence of cardiovascular risk factors in high- risk patients, Pharmacol. Res. 65 (2012) 615–620. [4] Z. Breen-Struckhofova, M. Cichna-Markl, Determination of bisphenol A in wine by sol- gel immunoaffinity chromatography, HPLC and fluorescence detection, Food Addit. Contam. 23 (2006) 1227–1235. [5] http://www.foodsafety.govt.nz/industry/sectors/plant-products/pesticidemrl/worldwide.htm. [6] M. Gónzalez-Álvarez, C. González-Barreiro, B. Cancho-Grande, J. Simal-Gándara, Impact of phytosanitary treatments with fungicides (cyazofamid famoxadone, mandipropamid and valifenalate) on aroma compounds of Godello white wines, Food Chem. 131 (2012) 826–836. [7] J.D. Carrillo, M.T. Tena, Determination of ethylphenols in wine by in situ derivatization and headspace solid-phase microextraction-gas chromatography-mass spectrometry, Anal. Bioanal. Chem. 387 (2007) 2547–2558. ˜ [8] J.I. Cacho, N. Campillo, P. Vinas, M. Hernández-Córdoba, Stir bar sorptive extraction with gas chromatography–mass spectrometry for the determination of resveratrol, piceatannol and oxyresveratrol isomers in wines, J. Chromatogr. A 1315 (2013) 21–27. [9] A.R. Fontana, I. Rodríguez, M. Ramil, J.C. Altamirano, R. Cela, Solid-phase extraction followed by liquid chromatography quadrupole time-of-flight tandem mass spectrometry for the selective determination of fungicides in wine samples, J. Chromatogr. A 1218 (2011) 2165–2175. [10] T. Rodríguez-Cabo, I. Rodríguez, P. López, M. Ramil, R. Cela, Investigation of liquid chromatography quadrupole time-of-flight mass spectrometry performance for identification and determination of hydroxylated stilbene antioxidants in wine, J. Chromatogr. A 1337 (2014) 162–170. [11] M. Moniruzzaman, I. Rodríguez, M. Ramil, R. Cela, S.A. Sulaiman, S.H. Gan, Assessment of gas chromatography time-of-flight accurate mass spectrometry for identification of volatile and semi-volatile compounds in honey, Talanta 129 (2014) 505–515.

11

[12] F. Zhang, H. Wang, L. Zhang, J. Zhang, R. Fan, C. Yu, W. Wang, Y. Guo, Suspected-target pesticide screening using gas chromatography-quadrupole time-of-flight mass spectrometry with high resolution deconvolution and retention index/mass spectrum library, Talanta 128 (2014) 156–163. [13] T. Portolés, J.G.J. Mol, J.V. Sancho, F. Hernández, Use of electron ionization and atmospheric pressure chemical ionization in gas chromatography coupled to time-of-flight mass spectrometry for screening and identification of organic pollutants in waters, J. Chromatogr. A 1339 (2014) 145–153. [14] R. Castro, R. Natera, E. Duran, C. García-Barroso, Application of solid phase extraction techniques to analyse volatile compounds in wines and other enological products, Eur. Food Res. Technol. 228 (2008) 1–18. [15] L. Lagunas-Allue, J. Sanz-Asensio, M.T. Martínez-Soria, Optimization and validation of a simple and fast method for the determination of fungicides in must and wine samples by SPE and GC/MS, J. AOAC Int. 95 (2012) 1511–1519. [16] R. Montes, I. Rodriguez, M. Ramil, E. Rubí, R. Cela, Solid-phase extraction followed by dispersive liquid-liquidliquid–liquid microextraction for the sensitive determination of selected fungicides in wine, J. Chromatogr. A 1216 (2009) 5459–5466. [17] I. Carpinteiro, B. Abuín, I. Rodríguez, M. Ramil, R. Cela, Mixed-mode solid-phase extraction followed by dispersive liquid-liquidliquid–liquid microextraction for the sensitive determination of ethylphenols in red wines, J. Chromatogr. A 1229 (2012) 79–85. [18] H.G.J. Mol, A. Rooseboom, R. Van Dam, M. Roding, K. Arondeus, S. Sunarto, Modification and re-validation of the ethyl acetate-based multi-residue method for pesticides in produce, Anal. Bioanal. Chem. 389 (2007) 1715–1754. [19] I. Carpinteiro, M. Ramil, I. Rodríguez, R. Cela, Determination of fungicides in wine by mixed-mode solid phase extraction and liquid chromatography coupled to tandem mass spectrometry, J. Chromatogr. A 1217 (2010) 7484–7492. [20] T. Rodríguez-Cabo, I. Rodríguez, R. Cela, Determination of hydroxylated stilbenes in wine by dispersive liquid–liquid microextraction followed by gas chromatography mass spectrometry, J. Chromatogr. A 1258 (2012) 21–29. [21] V. Ferreira, R. Lopez, J.F. Cacho, Quantitative determination of the odorants of young red wines from different grape varieties, J. Sci. Food Agric. 80 (2000) 1659–1667. [22] N. Loscos, P. Hernández-Orte, J. Cacho, V. Ferreira, Release and formation of varietal aroma compounds during alcoholic fermentation from nonfloral grape odorless flavor precursor fractions, J. Agric. Food Chem. 55 (2007) 6674–6684. [23] http://ec.europa.eu/food/plant/pesticides/eu-pesticides-database/public.

Please cite this article in press as: T. Rodríguez-Cabo, et al., Multiclass semi-volatile compounds determination in wine by gas chromatography accurate time-of-flight mass spectrometry, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.03.005