Article pubs.acs.org/JAFC
The Problem of 2,4,6-Trichloroanisole in Cork Planks Studied by Attenuated Total Reflection Infrared Spectroscopy: Proof of Concept Ana R. Garcia,†,§ Luís F. Lopes,† Ricardo Brito de Barros,† and Laura M. Ilharco*,† †
Centro de Quı ́mica-Fı ́sica Molecular and IN − Institute of Nanoscience and Nanotechnology, Complexo I, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal § Departamento de Quı ́mica e Farmácia, FCT, Universidade do Algarve, Campus de Gambelas, 8005-139 Faro, Portugal S Supporting Information *
ABSTRACT: Attenuated total reflection infrared spectroscopy (ATR-IR) proved to be a promising detection technique for 2,4,6-trichloroanisole (TCA), which confers organoleptic defects to bottled alcoholic beverages, allowing the proposal of a criterion for cork plank acceptance when meant for stopper production. By analysis of a significant number of samples, it was proved that the presence of TCA, even in very low concentrations, imparts subtle changes to the cork spectra, namely, the growth of two new bands at ∼1417 (νCC of TCA ring) and 1314 cm−1 (a shifted νCC of TCA) and an increase in the relative intensities of the bands at ∼1039 cm−1 (δCO of polysaccharides) and ∼813 cm−1 (τCH of suberin), the latter by overlapping with intense bands of TCA. These relative intensities were evaluated in comparison to a fingerprint of suberin (νasC−O−C), at 1161 cm−1. On the basis of those spectral variables, a multivariate statistics linear analysis (LDA) was performed to obtain a discriminant function that allows classifying the samples according to whether they contain or not TCA. The methodology proposed consists of a demanding acceptance criterion for cork planks destined for stopper production (with the guarantee of nonexistence of TCA) that results from combining the quantitative results with the absence of the two TCA correlated bands. ATR infrared spectroscopy is a nondestructive and easy to apply technique, both on cork planks and on stoppers, and has proven more restrictive than other techniques used in the cork industry that analyze the cleaning solutions. At the level of proof of concept, the method here proposed is appealing for high-value stopper applications. KEYWORDS: cork, trichloroanisole, infrared spectroscopy, ATR-IR, wine stoppers
■
INTRODUCTION Cork is a natural, renewable, and biodegradable raw material; its unique physical properties have stimulated the broadening of application fields.1,2 This is particularly true in Portugal, which is the largest worldwide producer.3,4 Cork has a very complex composition, but the main constituent is suberin, defined as a glycerolipid polyester, or a mixture of fatty acids and heavy organic alcohols that form a macromolecular network, insoluble in all common solvents.5 Another important constituent is lignin, with a less known structure than other natural and synthetic polymers, built by phenylpropane units randomly organized, having characteristic side chains with parasubstituted phenolic hydroxyl groups, either free or etherified.6,7 The 2D representations of proposed structures for suberin and lignin are schematized in Figure 1. Recent studies on the chemical composition of cork from Quercus suber L. (oak), the main natural source of cork, point out that the determining variation factor is the individual tree rather than the geographical origin.8 The average chemical composition of natural cork from Q. suber L. is approximately suberin 40%, lignin 22%, polysaccharides 18%, total extractives 15%, and ash below 1%.9 The oak cork natural variability with location is mainly reflected in the amounts of extractives and polysaccharides, whereas the relative contents in suberin and lignin are not much affected.9 Traditionally, the growth of filamentous fungi in cork was considered beneficial, but currently the detection of volatile compounds, namely, chloroanisoles and chlorophenols, is © 2014 American Chemical Society
Figure 1. Schematic 2D representations of suberin and lignin, obtained with ChemDraw Ultra 12.0.
associated with the incidence of some fungi.10,11 Those volatiles can also be produced by different sources of phenol and Received: Revised: Accepted: Published: 128
July 11, 2014 December 1, 2014 December 9, 2014 December 9, 2014 dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
background. The spectra were transformed to Kubelka−Munk units using FIRST software. The external surfaces of the stoppers were analyzed by infrared attenuated total reflection (IR-ATR) spectra, with a horizontal Pike Technologies (HATR) accessory. A slim slice of the stopper external surface was cut and placed on a SeZn ATR crystal (4 mm thick and 80 mm long) with 10 reflections and an incidence angle of 45°. Conditions for an “ideal contact” between the sample and the crystal were guaranteed, so that the band intensities remained constant as contact slightly changed.26 The spectra were also obtained as a ratio of 500 single-beam scans of the sample to the same number of background (clean ATR crystal) scans. The spectra were transformed to log10(1/R) using FIRST software. A Bruker Hyperion 2000 microscope coupled to a Vertex 70 FTIR spectrometer was used to map the distribution of TCA in some of the samples. The spectra of very thin cork slices were acquired in transmission mode, with a spatial resolution of 15 μm.
chlorine, such as some cleaning products and sanitizers (that contaminate surface waters). The honeycomb-type structure can provide adequate reservoirs for holding these volatile compounds in the cork core.1,12 The presence of volatiles is a critical problem in the food industry, because chloroanisoles and chlorophenols are associated with “musty” off-flavors and odors. Because cork materials are widely used to seal bottles of alcoholic beverages, the migration of chloroanisoles from the stoppers may result in those undesirable odors or tastes.13 In fact, the production of chloroanisoles by fungi in cork slabs is considered the most frequent cause of organoleptic defects of wines.14,15 The strongest chloroanisole is 2,4,6-trichloroanisole (TCA),16 formed mainly due to the microflora associated with corks.15,17 It is usually considered to be primarily responsible for economical losses in the wine industry, due to the increase of cork taint in bottled wine.18 Several analytical techniques can now be employed to detect the presence of TCA, and numerous processes have been developed to reduce its quantity in cork.19−24 Bleaching is the most used, but it does not remove all of the TCA. Moreover, when chlorinated products are used in the process, the formation of chlorophenols can be induced.3 The sensitivity of the olfactory receptor cells to TCA is extremely high: the threshold in wine can be low as 1.4−4.6 ng/L (or parts per trillion, ppt),25 which is a detection limit difficult to match by analytical methods. Given the dimension of this problem, it is now settled that a preventive action is a main priority, starting with the detection of these compounds in the cork planks before stoppers are produced. In the present work, a systematic analysis of attenuated total reflection infrared spectroscopy (ATR-IR) spectra of a substantial number of natural cork stoppers was performed to establish a safe criterion for TCA absence that may eventually be used for acceptance of cork planks intended for stoppers production.
■
■
RESULTS AND DISCUSSION Infrared Spectrum of TCA. The infrared spectrum of 2,4,6-trichloroanisole is known, but the detailed analysis has not been published so far. It is shown in Figure 2, in DRIFT.
EXPERIMENTAL PROCEDURES
Samples. Natural cork stoppers for wines that need to age in the bottle, from the Q. suber L. variety, were obtained from one manufacturer (the name is undisclosed for confidentiality reasons). The information provided consisted of whether the corks came from batches containing detectable traces of TCA in the last cleaning solution or not. No particular information was specified about the detection method, except that it was by chromatography. The stoppers, cut out directly from natural cork planks, were cylindrical single pieces, a total of 50. Two series of samples were provided: series A (15), from batches in which TCA was not detected in the last cleaning solution and, therefore, allegedly without TCA; and series R (30 five), from batches in which TCA was detected in the last cleaning solution and therefore expected to have TCA. A 0.01 M solution of 2,4,6-trichloroanisol (Sigma-Aldrich, 99%) in ethanol (absolute, ≥99.8%, Sigma-Aldrich) was prepared to artificially embed TCA onto clean cork. Different amounts were deposited onto the cork surface to obtain a variety of low TCA concentrations upon solvent evaporation. Infrared Characterization. All of the infrared spectra were scanned in a Mattson FTIR spectrometer RS1, equipped with a wide band MCT detector (400−4000 cm−1). The spectra, with 4 cm−1 resolution, are presented without baseline or smooth corrections. For analyzing pure TCA and cork powder from different regions of the same stopper, diffuse reflectance infrared Fourier transform (DRIFT) spectra were recorded using a Graseby/Specac selector. The samples were diluted in KBr (from Aldrich, FTIR grade) in a weight proportion of 1:4. The spectra were the ratio of 500 single-beam scans for the sample to the same number of scans for pure KBr, used as
Figure 2. DRIFT spectrum of pure 2,4,6-trichloroanisole.
Two groups of very strong bands are observed, in the ∼1600−1370 and ∼1000−600 cm−1 regions, respectively, plus a number of strong ones in the 2900−3200 cm−1 region. A detailed assignment was made mainly by comparison with the parent molecules 2,4,6-trichlorophenol and anisole27,28 and is presented in Table 1. Cork Infrared Characterization. In general, the cork stoppers have a light color and, after bleaching, the external surface becomes more whitish. Both the heterogeneities of natural cork and the bleaching process may lead to differences in the infrared absorption bands. To assess these possibilities, several regions of a bleached cork stopper from a batch where no TCA was detected in the cleaning solution (sample A1) were characterized by DRIFTS. The spectra are compared in Figure 3. The spectra in Figure 3, normalized to the strongest suberin band (νasCOC, at 1161 cm−1), show only changes in the bands’ relative intensities, confirming some variability in the chemical components’ proportions. The differences associated with the presence or absence of spots (sites a and b) are very small, whereas those related to the influence of bleaching (mostly site d) are clearer, pointing to bleaching as the main cause for adjustments in chemical proportions: the intensities of the 129
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
Table 1. Band Assignment of the DRIFT Spectrum of 2,4,6Trichloroanisolea ν̃ (cm−1) 3127(sh,m) 3080(m) 3017(sh,m) 2987(sh,m) 2952(m) 2936(sh,m) 2897(m)
assignment νCH νasCH3 νasCH3
ν̃ (cm−1) 1331(w) 1304(m) 1265/1249(db,S) 1227(sh,m) 1192(m) 1140(m) 1080(m)
assignment νCC δCH δCH
2863(m) 2831(m)
νsCH3 νsCH3
1032(m) 988(VS)
2800−1700
overtone and combination bands νCC (ring) νCC (ring) δasCH3 νCC (ring) δsCH3
861/853(db,VS)
δCH νring (breathing) νC−O−C νring (breathing) γCH
815/801(db,VS) 753(VS) 693(VS) 576(VS) 446(m)
νCCl γCH (oop) γCH (oop) δring(oop) δCOC
1570(sh,m) 1552(S) 1469(S) 1418(S) 1386(S)/1372(sh)
Figure 4. ATR infrared spectra of slides of two cork stoppers (sample A1, black line; and sample A4, red line) after being submitted to the bleaching process.
Table 2. Band Assignment of the Infrared ATR Spectrum of Bleached Corka
a
w, weak; m, medium; S, strong; VS, very strong; sh, shoulder; db, doublet; oop, out of plane.
ν̃ (cm−1)
assignment
cork component
∼3350(S) 2924(S) 2853(S) 1739(VS) 1714(sh) 1635(m) 1606(m) 1509(m) 1457(S) 1420(sh) 1364(m) 1236(VS) 1161(VS) 1094(VS) 1039(VS) 852(m) 818(m) 723(m)
νO−H νasCH2 νasCH2 νCO νCO νCC νCC νC...C δsCH2 δs(O)CH3 δsCH3 νCO νasCOC δCH δCO τCH (oop) ρCH3 τCH (oop)
suberin, lignin and adsorbed water suberin alkyl chains suberin alkyl chains suberin ester groups lignin suberin and lignin aliphatic groups suberin aliphatic groups lignin aromatic rings (fingerprint) suberin aliphatic groups (fingerprint) lignin suberin (fingerprint) suberin (fingerprint) suberin (fingerprint) polysaccharides polysaccharides suberin suberin/lignin suberin
Figure 3. DRIFT spectra of different regions of a bleached cork stopper: inner surface without defects (a), with spots (b), and close to the edge (c); outer surface (d). The spectra are normalized to the maximum.
a
bands at 2924, 1739, 1464, and 1240 cm−1 decrease in comparison to the suberin fingerprint, at 1161 cm−1, whereas those at 1093 and 1039 cm−1 appear to increase. For the systematic study that follows, the inner surfaces of cork stoppers will be analyzed. The ATR spectra of two slides of stoppers from a bleached batch (A1 and A4) were used to establish the standard “clean” cork and are shown in Figure 4. The spectra in Figure 4 were also normalized to the characteristic suberin band νasCOC, at 1161 cm−1, and correlate well with ATR infrared spectra of cork stoppers in the literature. 29−31 The assignment of the main bands is summarized in Table 2.6,32−34 The band shifts observed between the two samples are within the spectral resolution, and the changes in relative intensities are probably due to heterogeneities, as expected in natural samples. In fact, the amount of lignin in natural cork can vary from 19 to 24%, probably being the chemical component with
less variability. In the case of suberin, its content varies from ∼30 to almost 50%.8 Cork with TCA Embedded in Situ. The objective of characterizing cork samples with TCA embedded in the laboratory was to establish spectral regions unambiguously assignable to TCA on cork. The ATR infrared spectra were obtained only after complete evaporation of the solvent. The spectra of the same stopper, clean and with embedded TCA, are compared in Figure 5. The presence of TCA induces the appearance of new bands at 3078, 1550, 1417, 1383, 987, 796, and 752 cm−1, which can be considered as fingerprints of TCA in cork. Other strong modes of TCA can alter the band shape of the cork spectrum due to overlapping, as in the case of the bands at 1138, 1080, 854, and 692 cm−1. Moreover, in the region between 1470 and 1415 cm−1, the shoulder at 1468 cm−1 and the maximum at 1452 cm−1 invert their relative intensities, and the cork band at 1236 cm−1 splits into two components (1259 and 1248 cm−1),
w, weak; m, medium; S, strong; VS, very strong; sh, shoulder; oop, out of plane.
130
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
Figure 5. ATR infrared spectra of a clean cork stopper (black line) and the same sample after incorporation of TCA (blue line). The spectra were not baseline corrected and were normalized to the suberin band at 1161 cm−1.
the first shifted from the maximum observed for the pure TCA doublet (1265 and 1249 cm−1). Evaluation of the Presence of TCA in Stoppers from Natural Cork. Taking into account that the natural amount of TCA expected in stoppers is very low, we are aware that only the very strong vibrational bands of TCA will probably be observed in the tainted samples. On the other hand, because TCA is formed within natural cork, it is possible that the interactions with the medium lead to conformational changes and thus to spectral differences regarding pure or artificially added TCA and also that a heterogeneous spatial distribution throughout the sample occurs. This effect is well illustrated in Figure 6A, which refers to one stopper of series R (from a batch in which TCA was detected in the last cleaning solution): the honeycomb-like cork structure is observed over an area of 0.02 mm2, as well as the corresponding 3D mapping of the integrated intensity of the band at 1417 cm−1 (the strong νC C of TCA). The ATR spectra of seven cork stopper slices of series R are presented in Figure 6B. Only the region 1800−700 cm−1 is shown, because the main differences are observed in this wavenumber region.
Figure 7. Details of the ATR infrared spectra of cork stopper slices in which TCA was detected in the cleaning solution, compared to a “clean” sample (A1). The spectra are baseline corrected and normalized at 1161 cm−1.
As expected, the changes in the overall spectral shape are much more subtle than when TCA was embedded in situ. However, some effects are clear, and the corresponding regions are labeled (1−4) in Figure 6B. These spectral regions are enhanced in Figure 7, where the same samples are compared with one from a “TCA clean” stopper (A1). Analysis of Figure 7 establishes spectral characteristics inherent to corks with TCA:
Figure 6. (A) Image of a 0.02 mm2 area of a stopper slice in which TCA was detected in the cleaning solution (sample R2) and corresponding 3D distribution of the integrated intensity of the TCA mode at 1417 cm−1. (B) ATR infrared spectra of cork stoppers in which TCA was detected in the cleaning solution (series R): highlighted bands at 1417 cm−1 (1), ∼1314 cm−1 (2), and ∼1039 cm−1 (3) and the spectral region between 950 and 800 cm−1 (4). The spectra are baseline corrected and normalized at 1161 cm−1. 131
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
∼1039 cm−1 in comparison to the suberin fingerprint at 1161 cm−1 can be an indirect sign. The observation of these four effects, although subtle, attests to the presence of TCA in cork, even in low concentrations, and ATR appears as a good candidate method for TCA detection. However, the purpose of the present work is to use this technique to establish criteria for guaranteeing the absence of TCA in cork, consequently allowing the approval of cork planks intended for the manufacture of high-value wine stoppers. To establish these criteria, further spectra were analyzed for the two complete series of samples provided (A and R). The ATR spectra are shown in Figure 8. In the regions named above as 3 and 4, the TCA effects were tentatively quantified by the ratio between the intensity of a particular band to that of the 1161 cm−1 band: I1039/I1161 and I813/I1161, respectively. About 50 specimens (of the two series of samples A and R) were considered an acceptable number to carry out an analysis of variance (ANOVA) of those two intensity ratios. The extreme points were excluded from the statistical analysis, based on the modified Thompson Tau test for outlier determination.35 The Tukey test proved that the means difference between A and R series is significant at the 0.05 level, and the confirmation of equal variance for the two populations was obtained by using the Brown−Forsythe test. The samples used in the calculations, the results of the Brown− Forsythe test and ANOVA, are summarized in Table S1 in the Supporting Information. Because the two populations’ means are significantly different, the ratios I1039/I1161 and I813/I1161, the mean values, and corresponding confidence intervals are shown in Figure 9. The average intensity ratio I1039/I1161 is 0.96 ± 0.06 for the samples allegedly without TCA and 1.15 ± 0.06 for the samples with TCA, and the average ratio I813/I1161 is 0.42 ± 0.03 and 0.50 ± 0.04, respectively (95% confidence level). Accordingly, cork samples definitely not contaminated with TCA are expected to present simultaneously a I1039/I1161 ratio below 0.96 ± 0.06 and a I813/I1161 ratio below 0.42 ± 0.03. Cork being a natural product, the ideal sampling would consist of a larger number of specimens, identified by production region and pretreatments. The limitations of the previous statistical analysis on the data from spectral regions 3 and 4 may be partially overcome by complementing with the information from regions 1 and 2. A complete evaluation of the spectral effects of TCA is indispensable to establish the final criteria for deciding on the acceptance of a cork plank. The four effects are combined in Figure 10A.
Figure 8. ATR infrared spectra of cork stopper slices: series A, from batches in which TCA was not detected (blue spectra); series R, from batches in which TCA was detected (red spectra). The spectra were baseline corrected and normalized at 1161 cm−1.
In region 1, the TCA band near 1417 cm−1 grows, coincident with the observed for the sample with TCA embedded in the laboratory. In region 2, a very small band at ∼1314 cm−1 (almost a shoulder) is present; this new band was neither observed in the spectra of cork with incorporated TCA nor does it corresponds to a pure TCA band, but it may result from π−π interactions between the aromatic ring of TCA formed in situ and those of the suberin or lignin molecules. In region 3, the relative intensity of the band at ∼1039 cm−1 (δCO of polysaccharides) is much higher than in the spectrum of the sample without TCA. As referred to above, these polysaccharides may originate from natural cork or from fungi cells walls. Because TCA can be synthesized by cork fungi, this band may be an indirect indication of contamination. In region 4, the spectra between 950 and 800 cm−1 have a higher relative intensity that can be easily related with the presence of TCA, because its spectrum has very strong bands in this region, namely, the νCCl mode (that appears as a doublet at 815 and 801 cm−1 in pure TCA). In summary, the appearance of bands at ∼1417 and ∼1314 cm−1 and the intensity increase of the spectral region 950−800 cm−1 are direct indications of the presence of TCA in a cork stopper. The relative increase of the polysaccharide band at
Figure 9. Band intensity ratios for the samples used in the statistical analysis, ratio mean value (dashed line), and limits of the confidence interval (solid lines): (A) I1039/I1161; (B) I813/I1161; (blue squares) series A (allegedly without TCA); (red circles) series R (with TCA). 132
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
Figure 10. (A) Combined evaluation of the four effects of TCA on the ATR spectra of cork stoppers; (B) canonical scores for group A and R training data (TD) and corresponding canonical means: (blue squares) series A (allegedly without TCA); (red circles) series R (with TCA).
Table 3. Multivariate Statistics Linear Discriminant (LDA) Results Obtained for ATR Infrared Training Data of Samples A and R
In the absence of the bands at 1417 and 1341 cm−1 (quadrant I) and only within the shaded area (limited by I1039/I1161 ≤ 1.02 and I813/I1161 ≤ 0.45), all of the spectral observations point to safely concluding the nonexistence of TCA in the cork stoppers. In line with this conclusion, all of the samples within the shaded area come from batches in which TCA was not detected in the last cleaning solution. When the
bands at 1417 and 1341 cm−1 are both present (quadrant III), the cork stoppers are probably contaminated with TCA. Clearly, most of the samples included in this quadrant also present high values of the band ratios I1039/I1161 and I813/I1161 and therefore have a high possibility of TCA presence. Indeed, only one sample within this quadrant comes from A batches. If only the band at 1314 cm−1 is observed (quadrant IV), the 133
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
implementation, we are aware that a significantly larger number of samples should be tested.
conclusions can be more difficult. Only two samples are in these conditions, and both come from A batches. However, because the I1039/I1161 and I813/I1161 band ratios are high, the possibility of contamination in those cork stoppers may not be excluded. When only the band at 1417 cm−1 is observed (quadrant II), it is so unequivocal that it constitutes a strong enough indication of the presence of TCA in the cork. In this case, even if the band ratios do not support this hypothesis, the cork stoppers should be considered contaminated. To further discriminate samples with and without TCA, a multivariate statistics linear discriminant analysis (LDA) was applied to the ATR infrared data. The two data categories A and R were used, as previously, and the four spectral effects were considered as independent variables: X1 (the ratio I1039/ I1161), X2 (the ratio I813/I1161), X3 (the TCA band at 1417 cm−1), and X4 (the TCA band at 1314 cm−1). The dummy variables X3 and X4 were assigned either as 1 or 0, to account for the presence or absence of the bands, respectively. The output results (OriginPro 9.1) are summarized in Table 3, and the training data are reported in Table S2 in the Supporting Information. Equality of covariance matrices within groups is attested by similar log determinants and p value (0.319) of the likelihood-ratio test. Only one discriminant function and one eigenvalue accounting for 100% of the variance in the relationship are obtained, along with good canonical correlation between the discriminant function and the dependent variables. Statistical significance of the discriminant function is withdrawn from the low p value of the Wilks’ lambda test. According to the LDA analysis, the intensity ratio X1 (I1039/I1161) is the most relevant variable for the discrimination (canonical coefficient 0.886), followed by X3, X4, and X2 (I813/I1161). Strong validity of the variables chosen to distinguish between R and A samples is given by the absence of error in the classification (Table 3) and corresponding post probabilities obtained for the training data. Cross-validation (Table S2 in the Supporting Iinformation) evidences a higher error rate, most probably due to misclassification of some observations, namely those of group A that contain the relevant band at 1417 cm−1 (X3), pointing to a better allocation in group R. This corroborates the previous analysis, with the possibility of exclusion of some A samples. The canonical scores obtained for the ATR data along with the group centroids for A and R and the corresponding separation boarder line are represented in Figure 10B, confirming good quality of the discrimination attained. On the basis of this LDA, cork planks or stoppers can be safely rejected whenever they are allocated in the centroid region of group R. For acceptance of cork samples (guarantee of inexistence of TCA), however, the strong physical meaning of the bands at 1417 and 1341 cm−1 must be taken into account: the allocation in the centroid region of group A should be combined with the absence of both bands directly related with TCA. This leads to rejecting all of the samples already rejected by other methods, usually performed on the cleaning solutions, plus some accepted by those methods. Thus, the result is a more restrictive method in terms of acceptance of cork planks for stopper manufacture, especially adequate for cases when the economical damage for TCA-contaminated stoppers is more significant, namely, when high-value wines are at stake. Besides, it is a direct method, nondestructive, simple, rapid, and applicable both on the cork planks and on the cork stoppers. Although the application of the proposed method renders valuable information, it is a proof of concept, and, prior to
■
ASSOCIATED CONTENT
S Supporting Information *
Tables S1 and S2. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Funding
This work was supported by Fundaçaõ para a Ciência e a Tecnologia (partly by Project RECI/CTM-POL/0342/2012). Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS We acknowledge the undisclosed stopper manufacturer for kindly providing the samples. REFERENCES
(1) Gibson, L. J.; Easterling, K. E.; Ashby, M. F. The structure and mechanics of cork. Proc. R. Soc. London A 1981, 377, 99−117. (2) Silva, S. P.; Sabino, M. A.; Fernandes, E. M.; Correlo, V. M.; Boesel, L. F.; Reis, R. L. Cork: properties, capabilities and applications. Int. Mater. Rev. 2005, 50, 345−365. (3) Pereira, H. Cork: Biology, Production and Uses, 1st ed.; Elsevier: Amsterdam, The Netherlands, 2007; pp 320−322. (4) Lopes, M. H.; Barros, A. S.; Pascoal Neto, C.; Rutledge, D.; Delgadillo, I.; Gil, A. M. Variability of cork from portuguese Quercus suber studied by solid-state 13C-NMR and FTIR spectroscopies. Biopolymers 2001, 62, 268−277. (5) Cordeiro, N.; Belgacem, M. N.; Silvestre, A. J. D.; Pascoal Neto, C.; Gandini, A. Cork suberin as a new source of chemicals. 1. Isolation and chemical characterization of its composition. Int. J. Biol. Macromol. 1998, 22, 71−80. (6) Shabaka, A. A.; Nada, A. M. A.; Fadly, M. Infrared spectroscopic study of thermally treated lignin. J. Mater. Sci. 1990, 25, 2925−2928. (7) Hatakeyama, H.; Hatakeyama, T. Lignin structure, properties, and applications. Adv. Polym. Sci. 2010, 232, 1−63. (8) Pereira, H. Cork chemical variability. BioResources 2013, 8, 2246− 2256. (9) Pereira, H. Chemical composition and variability of cork from Quercus suber L. Wood Sci. Technol. 1988, 22, 211−218. (10) Á lvarez-Rodríguez, M. L.; López-Ocaña, L.; López-Coronado, J. M.; Rodríguez, E.; Martínez, M. J.; Larriba, G.; Coque, J.-J. R. Cork taint of wines: role of the filamentous fungi isolated from cork in the formation of 2,4,6-trichloroanisole by O methylation of 2,4,6trichlorophenol. Appl. Environ. Microbiol. 2002, 68, 5860−5869. (11) Tindale, C. R.; Whitfield, F. B.; Levingston, S. D.; Ly Nguyen, T. H. Fungi Isolated from packaging materials: their role in the production of 2,4,6-trichloroanisole. J. Sci. Food Agric. 1989, 49, 437−447. (12) Miranda, I.; Gominho, J.; Pereira, H. Cellular structure and chemical composition of cork from the Chinese cork oak (Quercus variabilis). J. Wood Sci. 2013, 59, 1−9. (13) Rocha, S.; Delgadillo, I.; Ferrer Correia, A. J.; Barros, A.; Wells, P. Application of an electronic aroma sensing system to cork stopper quality control. J. Agric. Food Chem. 1998, 46, 145−151. (14) Barreto, M. C.; Vilas Boas, L.; Carneiro, L. C.; San Romão, M. V. Volatile compounds in samples of cork and also produced by selected fungi. J. Agric. Food Chem. 2011, 59, 6568−6574. (15) Simpson, R. F.; Sefton, M. A. Origin and fate of 2,4,6trichloroanisole in cork bark and wine corks. Aust. J. Grape Wine Res. 2007, 13, 106−116. (16) Daniels-Lake, B. J.; Prange, R. K.; Gaul, S. O.; McRae, K. B.; de Antueno, R.; McLachlan, D. Musty “off” flavor in Nova Scotia potatoes
134
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135
Journal of Agricultural and Food Chemistry
Article
is associated with 2,4,6-trichloroanisole released from pesticide-treated soils and high soil temperature. J. Am. Soc. Hortic. Sci. 2007, 132, 112− 119. (17) Olivella, M. A.; Caixach, J.; Planas, C.; Oliveras, A.; Jové, P. Concentrations of organochlorine pesticides and 2,4,6-trichloroanisole in cork bark. Chemosphere 2012, 86, 754−758. (18) Juanola, R.; Subirá, D.; Salvadó, V.; Garcia Regueiro, J. A.; Anticó, E. Migration of 2,4,6-trichloroanisole from cork stoppers to wine. Eur. Food Res. Technol. 2005, 220, 347−352. (19) Rocha, S.; Delgadillo, I.; Ferrer Correia, A. J. Improvement of the volatile components of cork from Quercus suber L. by an autoclaving procedure. J. Agric. Food Chem. 1996, 44, 872−876. (20) Taylor, M. K.; Young, T. M.; Butzke, C. E.; Ebeler, S. E. Supercritical fluid extraction of 2,4,6-trichloroanisole from cork stoppers. J. Agric. Food Chem. 2000, 48, 2208−2211. (21) Recio, E.; Á lvarez-Rodríguez, M. L.; Rumbero, A.; Garzón, E.; Coque, J. J. R. Destruction of chloroanisoles by using a hydrogen peroxide activated method and its application to remove chloroanisoles from cork stoppers. J. Agric. Food Chem. 2011, 59, 12589−12597. (22) Pereira, C.; Gil, L.; Carriço, L. Reduction of the 2,4,6trichloroanisole content in cork stoppers using gamma radiation. Radiat. Phys. Chem. 2007, 76, 729−732. (23) Garde-Cerdán, T.; Lorenzo, C.; Zalacain, A.; Alonso, G. L.; Salinas, M. R. Using near infrared spectroscopy to determine haloanisoles and halophenols in barrel aged red wines. LWT−Food Sci. Technol. 2012, 46, 401−405. (24) Boudaoud, N.; Eveleigh, L. A new approach to the characterization of volatile signatures of cork wine stoppers. J. Agric. Food Chem. 2003, 51, 1530−1533. (25) Sefton, M. A.; Simpson, R. F. Compounds causing cork taint and the factors affecting their transfer from natural cork closures to wine − a review. Aust. J. Grape Wine Res. 2005, 11, 226−240. (26) Buffeteau, T.; Desbat, B.; Eyquem, D. Attenuated total reflection Fourier transform infrared microspectroscopy: theory and application to polymer samples. Vib. Spectrosc. 1996, 11, 29−36. (27) Faniran, J. A.; Shurvell, H. F. Raman and infrared spectra of 2,4,6-trichlorophenol and 2,4,6-tribromophenol. J. Raman Spectrosc. 1980, 9, 73−75. (28) Hoffmann, L. J. H.; Marquardt, S.; Gemechu, A. S.; Baumgärtel, H. The absorption spectra of anisole-h8, anisole-d3 and anisole-d8. The assignment of fundamental vibrations in the S0 and the S1 states. Phys. Chem. Chem. Phys. 2006, 8, 2360−2377. (29) Ortega-Fernández, C.; González-Adrados, J. R.; García-Vallejo, M. C.; Calvo-Haro, R.; Cáceres-Esteban, M. J. Characterization of surface treatments of cork stoppers by FTIR-ATR. J. Agric. Food Chem. 2006, 54, 4932−4936. (30) González-Adrados, J. R.; Garcia-Vallejo, M. C.; Caceres-Esteban, M. J.; Garcia de Ceca, J. L.; Gonzalez-Hernandez, F.; Calvo-Haro, R. Control by ATR-FTIR of surface treatment of cork stoppers and its effect on their mechanical performance. Wood Sci. Technol. 2012, 46, 349−360. (31) González-Gaitano, G.; Ferrer, M. A. C. Definition of QC parameters for the practical use of FTIR-ATR spectroscopy in the analysis of surface treatment of cork stoppers. J. Wood Chem. Technol. 2013, 33, 217−233. (32) Rocha, S. M.; Goodfellow, B. J.; Delgadillo, I.; Neto, C. P.; Gil, A. M. Enzymatic isolation and structural characterisation of polymeric suberin of cork from Quercus suber L. Int. J. Biol. Macromol. 2001, 28, 107−119. (33) Olivella, M. À .; Fernández, I.; Cano, L.; Jové, P.; Oliveras, A. Role of chemical components of cork on sorption of aqueous polycyclic aromatic hydrocarbons. Int. J. Environ. Res. 2013, 7, 225− 234. (34) Oliveira, L.; Evtuguin, D. V.; Cordeiro, N.; Silvestre, A. J. D.; Silva, A. M. S.; Torres, I. C. Structural characterization of lignin from leaf sheaths of “Dwarf Cavendish” banana plant. J. Agric. Food Chem. 2006, 54, 2598−2605.
(35) Anbarasi, M. S.; Ghaayathri, S.; Kamaleswari, R.; Abirami, I. Outlier detection for multidimensional medical data. Int. J. Comput. Ssci. Inf. Technol. 2011, 2, 512−516.
135
dx.doi.org/10.1021/jf503309a | J. Agric. Food Chem. 2015, 63, 128−135