Journal of Chromatography A, 1216 (2009) 1279–1284

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: [Link]/locate/chroma

Solid-phase microextraction on-fiber derivatization for the analysis of some

polyphenols in wine and grapes using gas chromatography–mass spectrometry夽
Pilar Viñas, Natalia Campillo, Nelson Martínez-Castillo, Manuel Hernández-Córdoba ∗
Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, E-30071 Murcia, Spain

a r t i c l e i n f o a b s t r a c t

Article history: The present study describes a new environmentally friendly sample pretreatment system based on solid-
Received 30 October 2008 phase microextraction (SPME) for the sensitive determination of polyphenols. A derivatization process
Received in revised form 18 December 2008 was necessary to convert the polar non-volatile compounds into volatile derivatives. Direct immersion
Accepted 19 December 2008
(DI) SPME was used for the adsorption of polyphenols, and then the fiber was placed in the headspace
Available online 27 December 2008
of the derivatizing reagent, bis(trimethylsilyl)trifluoroacetamide (BSTFA). The separation was carried out
by coupling gas chromatography with mass spectrometry in the selected ion monitoring mode, after
Keywords:
silylation. Optimal extraction conditions were 25 ◦ C for 10 min under continuous stirring using DI and
Gas chromatography–mass spectrometry
Solid-phase microextraction
a polyacrylate fiber. After extraction, the fiber was inserted into the headspace of BSTFA (10 ␮L) and
On-fiber derivatization the polyphenols were derivatized for 15 min at 50 ◦ C. Desorption was carried out at 280 ◦ C for 5 min.
Polyphenols The method allowed the determination of both isomers cis- and trans-resveratrol, piceatannol, catechin
Wine and epicatechin in wine and grapes, and it was validated for linearity, detection and quantitation limits,
Grapes selectivity, accuracy and precision. Detection limits ranged from 0.05 to 0.9 ng mL−1 at a signal-to-noise
ratio of 3, depending on the compound. Recoveries obtained for spiked samples were satisfactory for all
compounds.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction based on gas chromatography (GC) have been proposed [9–18], as

these compounds are non-volatile and a chemical derivatization
The healthy properties of wine and grapes are basically due to step is needed to obtain volatile and thermostable derivatives. How-
the presence of antioxidants such as polyphenols, which include ever, the use of GC with mass spectrometry (MS) detection provides
numerous substances many of which have beneficial effects in important advantages because of the combination of the separa-
terms of health promotion [1]. Polyphenols can be classified into tion capabilities of GC and the power of MS as an identification and
several groups according to their basic structure [2]. The flavonoids confirmation method. Furthermore, this hyphenized technique is
include catechins, and show antioxidative, anticarcinogenic and accessible to many laboratories. Conventional approaches for sam-
hypocholesterolemic activities [3,4]. Stilbenoids, such as resveratrol ple treatment are tedious and time-consuming and require great
and piceatannol, are naturally occurring compounds in grapes and volumes of sample and organic solvents, which implies environ-
wine [5] and these have been linked to a low incidence of fatal coro- mental contamination, risks for human health and additional costs
nary heart disease among populations consuming wine moderately for residue treatment. Because of the importance of developing
[1,2]. On the other hand, piceatannol has a stronger anticancer activ- clean chemistry procedures, emerging methods for food matrices
ity than resveratrol and becomes a promising natural compound are based on solvent-free procedures. Newer sample preparation
in cancer treatment [6]. In general, the amount of piceatannol in techniques offer automation and are clean, selective, rapid and effi-
plants is much lower than that of resveratrol [7], and efficient and cient; ideally, they should be cheap, simple and solvent-free [19].
sensitive methods are necessary. Solid-phase microextraction (SPME) is a non-harmful environmen-
Polyphenols are generally determined by liquid chromatogra- tally pre-concentration technique which is clean, selective, rapid,
phy (HPLC) using different detection systems [8]. Very few studies efficient, simple and solvent-free [20–22]. The derivatization of
polyphenols allowed non-volatile compounds to be converted into
volatile derivatives, thus improving the separation, selectivity or
夽 Presented at the 12th Conference on Instrumental Analysis, Barcelona, Spain,
sensitivity of SPME–GC determination [23]. This step can be per-
21–23 October 2008.
formed in the sample solution prior to sampling, on the SPME fiber
∗ Corresponding author. Fax: +34 968364148. coating or during desorption in the GC injection port. For on-fiber
E-mail address: hcordoba@[Link] (M. Hernández-Córdoba). derivatization, the chemical agent has to be loaded on the fiber

0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/[Link].2008.12.058
1280 P. Viñas et al. / J. Chromatogr. A 1216 (2009) 1279–1284

prior to or after analyte extraction. Commonly, the analytes are first temperature of 100 ◦ C, increase to 320 ◦ C at 30 ◦ C min−1 and hold
extracted and then the fiber is exposed to the vapour of the deriva- for 5 min. The temperatures of the ion source and the transfer line
tizing agent [22]. The coupling of GC and SPME has been used to were 230 and 325 ◦ C, respectively. Both resveratrol isomers and epi-
determine resveratrol isomers [24–26] and the total free resveratrol catechin were quantified in the SIM mode in order to improve the
was also determined by direct-exposure-probe positive-ion chemi- detection limits using one target and two or three qualifier ions.
cal ionization mass spectrometry [27]. However, no references have However, catechin and piceatannol were quantified using only the
been found to the use of the hyphenated technique SPME–GC–MS target ion. Identification was confirmed by the retention time of the
for the determination of polyphenol mixtures. target ion and the qualifier-to-target ion ratios for each compound
The present study describes a new solvent-free method for the (Table 1).
sensitive determination of both cis- and trans-resveratrol isomers, The SPME holders for manual use and fibers coated with 100 ␮m
piceatannol, catechin and epicatechin in wine and grapes. Direct polydimethylsiloxane (PDMS; apolar), 65 ␮m polydimethylsilo-
immersion (DI) SPME was used for the adsorption of analytes and, xane-divinylbenzene (PDMS-DVB; bipolar), 85 ␮m polyacrylate
then, the fiber was inserted into the headspace of the derivatizing (PA; polar), 75 ␮m Carboxen-polydimethylsiloxane (CAR-PDMS;
reagent, bis(trimethylsilyl)trifluoroacetamide (BSTFA). The separa- medium polar), 70 ␮m Carbowax-divinylbenzene (CW-DVB;
tion was carried out by coupling GC with MS in the selected ion polar), 50/30 ␮m DVB-CAR-PDMS (medium polar) and 60 ␮m
monitoring (SIM) mode. poly(ethylene glycol) (PEG; polar), were purchased from Supelco.
The fibers were conditioned by heating in the injection port of the
GC according to the manufacturer. All analyses were performed in
2. Experimental
15 mL amber glass vials sealed with hole-caps and PTFE/silicone
septa and the solutions were stirred with a magnetic stirrer (IKA
2.1. Reagents
RH KT/C, Supelco) at 400 rpm using PTFE-coated magnetic stir bars
(10 mm × 6 mm O.D.). To control the adsorption temperature, a
Analytical-reagent grade methanol was purchased from Lab-
home-made heating system consisting of a drilled block provided
Scan (Dublin, Ireland). Deionized water was obtained from a Milli-Q
with an electronic temperature control system was used.
water purification system (Millipore, Bedford, MA, USA).
A PSA 10570 UV cracker from a PSA Millenium Excalibur con-
Commercially available polyphenols (>98%) were obtained from
tinuous flow system (PS Analytical, Orping, UK) was used to obtain
Sigma (St. Louis, MO, USA): trans-resveratrol, piceatannol, cate-
the cis-resveratrol. An IKA-A11 grinder and an UP 200H ultrasonic
chin and epicatechin. Stock solutions (100 mg L−1 ) were prepared
probe processor (Dr. Hielscher, Teltow, Germany) were also used for
by dissolving the commercial products, without previous purifica-
the homogenization of samples.
tion, in methanol. They were kept at −18 ◦ C in dark bottles sealed
with PTFE/silicone caps. Working standard solutions were prepared
daily by diluting with methanol. The solution of cis-resveratrol 2.3. Samples
was obtained by UV irradiation of the trans-isomer using a cracker
with UV light for 1 min. The continuous flow system cracker was Ten samples of different types of wine (red, rosé, white and
so effective that total isomerisation was achieved after only 1 min sweet) and grapes (red and white) were obtained commercially and
irradiation. The amount of trans-resveratrol remaining was not stored in glass containers at 4 ◦ C before analyzing. Wine samples
detectable, at least, using the UV detection. Consequently, the con- were diluted with water and analyzed at about 50% (v/v) percent-
centration of cis-resveratrol considered for calibration was the same age in 15 mL vials. Fresh grapes were analyzed unwashed and with
as that for trans-resveratrol. This solution was irradiated on the the peel intact. A representative portion of sample (50 g whole fruit)
same day of use. The derivatizing reagents were BSTFA and BSTFA was chopped into small pieces and triturated in an IKA-A11 grinder.
containing 1% trimethylchlorosilane (TMCS), and were obtained For the extraction procedure, portions of 8 g were weighed into
from Supelco (Bellefonte, PA, USA) and maintained at room tem- screw-capped tubes and 2 mL of ethanol were added. The mix-
perature in the dark. ture was sonicated for 5 min (50% amplitude) by means of a probe
directly immersed in the sample solution and the entire extract
was centrifuged at 4000 rpm for 5 min. Finally, the supernatant was
2.2. Instrumentation
passed to 15 mL vials. Recovery experiments were carried out using
two samples of red wine and grapes, which were spiked with a stan-
GC analyses were performed on an Agilent 6890N (Agilent,
dard mixture of polyphenols at two concentration levels of 10 and
Waldbronn, Germany) gas chromatograph coupled to an Agilent
40 ng g−1 . The samples were treated as indicated above and were
5973 quadrupole mass selective spectrometer equipped with an
allowed to stand at 4 ◦ C for at least half an hour before starting the
inert ion source and provided with a split–splitless injection port
extraction procedure.
and a 0.75 mm I.D. liner. The mass spectrometer was operated
using the electron-impact (EI) mode (70 eV). The carrier gas helium
was maintained at a constant flow of 1 mL min−1 . A DB-5MS (Tec- 2.4. Analytical procedure
nokroma, Barcelona, Spain) capillary column (30 m × 0.25 mm I.D.,
0.25 ␮m film thickness) was used. Desorption of the fibers into the All analyses were performed with 15 mL amber vials. For the
injection port was carried out in the splitless mode at 280 ◦ C dur- wine analyse, the samples were prepared by adding 8 mL of wine
ing 5 min. The GC temperature was programmed as follows: start and 7 mL of water. For grapes, the samples contained the super-

Table 1
Retention time and target and qualifier ions for the polyphenols.

Compound tR (min) T Q1 (Q1 /T%) Q2 (Q2 /T%) Q3 (Q3 /T%)

cis-Resveratrol 7.67 444 445 (68.21) 446 (40.25)

trans-Resveratrol 9.00 444 445 (68.21) 446 (40.25) 447 (5.23)
Epicatechin 9.54 369 368 (31.26) 355 (25.44) 270 (11.22)
Catechin 9.71 368 369 (55.83) 267 (11.64)
Piceatannol 9.74 532 516 (10.50) 444 (3.95)
 P. Viñas et al. / J. Chromatogr. A 1216 (2009) 1279–1284 1281

ions selected for the five polyphenols studied under the chromato-
graphic conditions finally used in the SIM mode. Fig. 1 corresponds
to the chromatograms obtained in the total ion current (TIC) mode
(A) as well as using different target ions (B and C). As can be seen, the
peaks corresponding to catechin and piceatannol eluted very close
together and could only be differentiated by using the target ions.
The correct sequence of the ions selected as a function of the elut-
ing time was: cis-resveratrol (444, 7.67 min), trans-resveratrol (444,
9.0), epicatechin (369, 9.54), catechin (368, 9.71) and piceatannol
(532, 9.74).

3.2. SPME optimization

The extraction of all polyphenols was optimized using the direct

immersion mode and different types of SPME fiber (differing in
polarity and thickness of the stationary phase). The fibers assayed
were: 100 ␮m PDMS (apolar), 65 ␮m PDMS-DVB (bipolar), 85 ␮m
PA (polar), 75 ␮m CAR-PDMS (medium polar), 70 ␮m CW-DVB
(polar), 50/30 ␮m DVB-CAR-PDMS (medium polar) and a newly
Fig. 1. Elution profile obtained for a red wine sample by SPME–GC–MS. (A) Total ion developed fiber of 60 ␮m PEG (polar). The presence of BSTFA,
chromatogram; (B) chromatogram using 368 as target ion; (C) using 532 as target the derivatizing reagent, produced many problems concerning the
ion. Peaks correspond to: 1, cis-resveratrol; 2, trans-resveratrol; 3, epicatechin; 4, durability of the fibers. Thus, fibers made of PDMS, PEG, PDMS-DVB
catechin (368) and 5, piceatannol (532). and CW-DVB were rapidly detached before 10 extractions. Table 2
shows the mean and the standard deviation values obtained for
triplicate injections of the polyphenol mixture performed on each
natant (obtained as indicated above) and water up to 15 mL. The fiber. The best extraction for most compounds and long durabil-
vial was sealed with the cap after introducing the magnetic stir ity was achieved with the PA fiber, which was selected. The two
bar and placed in the home-made heating module. Analytes were extraction SPME modes, headspace (HS) and DI, were compared
extracted by direct immersion using the PA fiber (85 ␮m) at 25 ◦ C using the PA fiber and polyphenol extraction was around 100 times
for 10 min with stirring at 400 rpm. Then, the fiber was removed higher when the latter method (DI) was used.
and dried using a nitrogen stream for 5 min to avoid the hydrolysis When the extraction temperature was varied between 25 and
of the derivatizing reagent. Finally, the fiber was introduced into the 60 ◦ C using an extraction time of 20 min, the extraction efficiency
headspace of a 1.5 mL amber vial containing 10 ␮L of BSTFA at 50 ◦ C slightly varied at higher temperatures. The extraction temperature
for 15 min and, after derivatization, the analytes were desorbed in selected, therefore, was 25 ◦ C. The extraction time was critical for
the injection port of the GC in the splitless mode at 280 ◦ C for 5 min. establishing an equilibrium between the sample and the PA fiber
Each sampling was performed in triplicate. and Fig. 2A shows the influence of this variable for times between
1 and 40 min. The extraction efficiency increased for longer extrac-
3. Results and discussion tion times up to 10 min and then decreased or remained constant,
depending on the analyte. A time of 10 min was selected. The kinetic
3.1. GC–MS separation of the process is very important in SPME, and stirring aqueous sam-
ples can reduce the extraction time. The stirring speed was varied
The separation conditions of polyphenols were optimized by between 0 and 2000 rpm (Fig. 2B) and the signals increased, espe-
comparing two different capillary columns coated with non-polar cially for resveratrol and catechin, up to a speed of 400 rpm, which
stationary phases, DB-1MS and DB-5MS (30 m length × 0.25 mm was selected. The pH and ionic strength of the sample can affect
diameter, 0.25 ␮m film thickness) and one capillary column coated the extraction efficiency by decreasing the solubility of the analytes
with a partially polar stationary phase, DB-17MS. Resolution was when they are mainly present in their neutral form. The pH effect
satisfactory for all compounds except for catechin and piceatan- was studied in the 3–7 range, using 0.01 M phosphate buffer solu-
nol, which coeluted with all columns. Best separation was achieved tions and the extraction increased for higher pH up to 5.5 and then
with the DB-5MS column (5% diphenyl–95% dimethylpolysiloxane), decreased or remained constant, depending on the analyte. A pH
which was selected. Several temperature programs were tested in value of 5.5 was selected. The influence of ionic strength was studied
order to achieve the best separation of the coeluting phenols and by adding 0–15% (m/v) concentrations of sodium chloride. However,
the lowest analysis time and the program selected is summarized in the fiber was damaged in the presence of such concentrations of salt
Section 2. The flow of helium gas was varied in the 0.5–4 mL min−1 and, consequently, its use was discarded. To avoid any decrease in
range and an optimal value of 1 mL min−1 was selected. Table 1 sensitivity, vials containing the samples were completely filled to
shows the retention times as well as the target and the qualifier remove any remaining headspace in the vial.

Table 2
Comparison of SPME fibers for polyphenolsa .

Compound Fiber

CW-DVB DVB-CAR-PDMS CAR-PDMS PA PDMS-DVB PDMS PEG

cis-Resveratrol 1900 ± 109 1600 ± 266 1900 ± 103 27000 ± 181 31000 ± 795 0 970 ± 123
trans-Resveratrol 290 ± 94 800 ± 182 290 ± 148 6800 ± 226 980 ± 336 0 110 ± 33
Epicatechin 1800 ± 71 1700 ± 70 1800 ± 103 1300 ± 20 4500 ± 290 0 260 ± 50
Catechin 9300 ± 911 390000 ± 61837 9300 ± 613 298000 ± 1563 136000 ± 7284 0 39000 ± 1738
Piceatannol 14000 ± 274 7400 ± 292 14000 ± 1184 32000 ± 181 51000 ± 476 0 4200 ± 387
a
Values are mean (peak area) ± standard deviation (n = 3).
1282 P. Viñas et al. / J. Chromatogr. A 1216 (2009) 1279–1284

Fig. 2. Influence of the extraction time (A) and stirring rate (B) using the PA fiber and DI-SPME. Sample volume, 15 mL; desorption time, 5 min; 20 ng mL−1 polyphenols.
Symbols correspond to: resveratrol (䊉), piceatannol (), epicatechin () and catechin ().

With respect to the desorption conditions, both the tempera- between 30 and 90 ◦ C and extraction efficiency was seen to increase
ture and time were optimized. Temperature was varied between up to 70 ◦ C. However, the PA fiber was damaged after repeated
200 and 280 ◦ C and the signal for all the compounds increased up extractions and a lower temperature of 50 ◦ C was selected. The
to 280 ◦ C, which was selected. Next, the desorption time was opti- derivatization time was varied in the 10–40 min range and optimal
mized to achieve the total desorption of analytes with no memory efficiencies were obtained at 20 min. The presence of a catalyst can
effects and the optimal time for extracting maximal amounts of all modify the behaviour of the derivatizing reagent and thus the effi-
compounds was 5 min. ciency of both reagents, BSTFA and BSTFA + 1% TMCS was compared.
Fig. 3 shows that best extraction efficiencies were obtained for all
3.3. Derivatization reaction for polyphenols polyphenols (except for epicatechin, which gave identical results)
in the presence of the catalyst and, consequently, the BSTFA + 1%
Silylation is a common procedure for the GC analysis of the TMCS reagent was selected. The optimal volume of the reagent was
non-volatile and thermolabile polyphenols. Trimethylsilyl deriva- also considered by varying in the 5–50 ␮L range and maximal effi-
tives are more volatile, less polar and more thermostable. In this ciency was obtained when 10 ␮L were added. Since the fiber must
process, an active hydrogen is replaced by a trimethylsilyl group be dried before derivatization, using a nitrogen stream to avoid
[23]. Generally, silylation hyphenated with SPME is difficult because hydrolysis of the reagent, the drying time was varied between 2
the reagents and the silylmethylated derivatives are hydrolyzed in and 15 min, with optimal effectiveness of the derivatization reac-
aqueous solutions. To avoid this problem, the fiber after extrac- tion being achieved when the fiber was dried for 5 min in a closed
tion was exposed to a nitrogen flow for 5 min to remove water and vial at room temperature.
solvent residues in a 15 mL vial closed with a septum with two nee-
dles for the nitrogen intake and way out at room temperature. The 3.4. Method performance
analytes were immediately derivatized by exposing to the BSTFA
gaseous phase. The method was validated for linearity, detection and quanti-
Several variables were examined to determine their influence tation limits, selectivity, accuracy and precision. Calibration curves
on the silylation process. The derivatization temperature was varied using SPME–GC–MS were obtained by least-squares linear regres-
sion analysis of the peak area versus analyte concentration using six
concentration levels in duplicate. Quantification was performed by
the external standard procedure. The validation parameters, range
of linearity and the correlation coefficients for the polyphenols are
shown in Table 3. The values of r2 were good (r2 > 0.99) and excellent
linearity was obtained for the range studied. The limits of detection
(LODs, calculated as three times the signal-to-noise ratio) and the
limits of quantitation (LOQs, calculated as 10 times the signal-to-
noise ratio) are included in Table 3. It can be seen from the data
that the sensitivity and detection limits for polyphenols were very
low. To check the repeatability of the method, 10 replicate analyses

Table 3
Analytical data for polyphenol standards using the SPME–GC–MS method.

Compound Linearity r2 LOD LOQ RSD

(ng mL−1 ) (ng mL−1 ) (ng mL−1 ) (n = 10)

cis-Resveratrol 1–150 0.9975 0.11 0.4 2.9

trans-Resveratrol 1–150 0.9986 0.09 0.3 2.4
Epicatechin 3–700 0.9944 0.9 3.0 3.8
Fig. 3. Comparison of the derivatizing reagents using the PA fiber and DI-SPME Catechin 3–100 0.9916 0.6 2.0 4.8
for 20 ng mL−1 polyphenols. Filled boxes correspond to BSTFA and hollow boxes to Piceatannol 1–150 0.9940 0.05 0.2 4.1
BSTFA + 1% TMCS.
 P. Viñas et al. / J. Chromatogr. A 1216 (2009) 1279–1284 1283

Table 4
Analytical data for polyphenols in red wine and grape samples using the SPME–GC–MS method.

Compound r2 LOD LOQ RSD (n = 10)

−1 −1 −1 −1
Wine Grape Wine (ng mL ) Grape (ng g ) Wine (ng mL ) Grape (ng g ) Wine Grape

cis-Resveratrol 0.999 0.998 0.2 0.2 0.7 0.7 3.6 3.6

trans-Resveratrol 0.999 0.990 0.3 0.4 1.0 1.4 2.6 8.1
Epicatechin 0.991 0.990 4 9 13 30 6.0 4.8
Catechin 0.990 0.990 2 4 7 13 6.8 9.6
Piceatannol 0.990 0.990 0.3 0.7 1 2.3 7.5 10

Table 5
Polyphenol contents in wine (ng mL−1 ) and grape (ng g−1 ) samples.

Simple trans-Resveratrol cis-Resveratrol Piceatannol Catechin Epicatechin

Red wine 41 ± 3.3 37 ± 0.3 14 ± 1.3 182 ± 24 50 ± 5.1

Red wine 36 ± 3.1 4.4 ± 1.2 ND 73 ± 11 ND
Red wine 32 ± 2.7 36 ± 9.5 7.7 ± 0.6 118 ± 38 88 ± 2.0
Rosé wine 7.2 ± 0.1 3.1 ± 0.3 ND 35 ± 1.8 ND
Rosé wine 4.8 ± 0.7 ND ND 31 ± 2.9 ND
Rosé wine 29 ± 2.2 3.0 ± 0.1 ND 261 ± 12 ND
White wine 19 ± 3.9 9.5 ± 1.1 ND 236 ± 20 ND
White wine 13 ± 1.8 ND ND ND ND
Sweet wine 33 ± 3.2 18 ± 1.0 6.1 ± 2.6 48 ± 6.3 172 ± 37
Sweet wine 31 ± 3.9 19 ± 3.4 6.6 ± 0.6 97 ± 15 85 ± 1.5
Red grape 29 ± 2.4 2.8 ± 0.3 24 ± 0.7 45 ± 5.9 37 ± 3.5
White grape 5.6 ± 0.5 ND 1.2 ± 0.7 86 ± 3.8 ND

Values are mean ± standard deviation (n = 3).

of a polyphenol mixture were performed at the level of 20 ng mL−1 the peaks in both the sample and the standard solution. The average
for each compound, with RSD of <5% in all cases (Table 3). These values for the retention times of polyphenols pointed to very good
values indicate that the precision of the method was satisfactory agreement between the retention data obtained in the different
for control analysis. samples, as the criterion for positive identification of the com-
Selectivity of the method was assessed by the absence of inter- pound was a time of ±0.5 min the retention time of that compound.
fering peaks at the elution times of the polyphenols for blank The T and Q abundances were determined by injecting individual
chromatograms of different wine samples. standards under the same chromatographic conditions, except in
The matrix effect for the polyphenols was evaluated by com- full scan mode. The Q/T percentage was determined by dividing
paring the slopes of aqueous standards and standard additions the abundance of the selected qualifier ion by the target ion (see
calibration graphs for different wines and grape samples, obtained Table 1).
by plotting concentration (at six levels) against peak area and fol- After identification of the peaks, 10 different samples of wine
lowing linear regression analysis. Quantitation was based on the (red, rosé, white and sweet) and two samples of grapes (red and
target ion. Table 4 shows the parameters obtained. Slopes for wines white) were analyzed using the SPME–GC–MS procedure. All sam-
and grape samples were lower than those for aqueous standards. ples were analyzed in triplicate. Table 5 shows the results obtained.
A statistical study was carried out to compare the slope values of As expected, higher levels for all polyphenols appeared in red wine
the different samples and the aqueous standards and the analy- and grapes. On the other hand, levels of piceatannol were lower
sis of variance (ANOVA) test showed that there were statistically than those of resveratrol. The levels of polyphenols in wine or
significant differences. Consequently, to obtain a generally reli- grapes samples found in the literature were very variable. How-
able procedure, the standard additions method is recommended ever, some values found were low, probably by losses because the
for quantitation purposes. samples were commercially obtained and they could be previously
submitted to extreme climate or light conditions.
3.5. Recovery
4. Conclusion
In order to check the accuracy of the proposed method, a recov-
ery study was carried out by fortifying two samples (red wine
A new environmentally friendly sample pretreatment system,
and grape) at two concentration levels corresponding to 10 and
SPME, for the sensitive determination of both cis- and trans-
40 ng g−1 . The recoveries of the polyphenols from spiked samples
resveratrol isomers, piceatannol, catechin and epicatechin has been
varied from 85 to 116% with an average recovery ± SD (n = 60) of
developed. DI-SPME was used to adsorb analytes and, then, the
103 ± 6. The similarity in recoveries obtained for each phenol in the
fiber was inserted into the headspace of the derivatizing reagent,
different samples indicates that the matrix effect was corrected by
BSTFA. The separation was carried out by GC–MS(SIM). The pro-
using the standard additions calibration.
posed method was successfully applied to determining polyphenols
without using organic solvents or clean-up steps, thus reducing
3.6. Analysis of wine and grape samples
the sample preparation time and allowing the reliable control and
screening of polyphenols present in wine and grape samples.
Different wine and grape samples were analyzed under the
selected conditions. The chromatographic profiles demonstrated
the absence of interfering peaks at the retention times for the Acknowledgements
analytes as shown in Fig. 1. The polyphenols were identified by
comparing the retention time, identifying the target (T) and quali- The authors are grateful to Comunidad Autónoma de la Región
fier ions (Q) and comparing the qualifier-to-target ratios (Q/T%) of de Murcia (CARM, Fundación Séneca, Project 02993/PI/05) and to
1284 P. Viñas et al. / J. Chromatogr. A 1216 (2009) 1279–1284

the Spanish MEC (Project CTQ2006-08037/BQU) for financial sup- [12] A.M. Rimando, W. Kalt, J.B. Magee, J. Dewey, J.R. Ballington, J. Agric. Food Chem.
port. N.M.-C. acknowledges a fellowship from Departamento de 52 (2004) 4713.
[13] O. Tokusoglu, M.K. Unal, F. Yemis, J. Agric. Food Chem. 543 (2005) 5003.
Formación del Personal Académico de la Universidad Centrocciden- [14] C. Proestos, D. Sereli, M. Komaitis, Food Chem. 95 (2006) 44.
tal Lisandro Alvarado (Venezuela). [15] A.S. Ragab, J. van Fleet, B. Jankowski, J.H. Park, S.C. Bobzin, J. Agric. Food Chem.
54 (2006) 7175.
[16] A. Canini, D. Alesiani, G. D’Arcangelo, P. Tagliatesta, J. Food Compos. Anal. 20
References (2007) 584.
[17] Z.I. Füzfai, P. Molnár, J. Chromatogr. A 1149 (2007) 88.
[1] D.M. Goldberg, S.E. Hahn, J.G. Parkes, Clin. Chim. Acta 237 (1995) 155. [18] A. Chiou, V.T. Karathanos, A. Mylona, F.N. Salta, F. Preventi, N.K. Andrikopoulos,
[2] G.J. Soleas, E.P. Diamandis, D.M. Goldberg, J. Clin. Lab. Anal. 11 (1997) 287. Food Chem. 102 (2007) 516.
[3] P. Iacopini, M. Baldi, P. Storchi, L. Sebastiani, J. Food Compos. Anal. 21 (2008) [19] M.D. Luque de Castro, J.L. Luque García, Acceleration and Automation of Solid
589. Sample Treatment, Elsevier, Amsterdam, 2002.
[4] L. Hooper, P.A. Kroon, E.B. Rimm, J.S. Cohn, I. Harvey, K.A. Le Cornu, J.J. Ryder, [20] J. Pawliszyn, Solid Phase Microextraction, Theory and Practice, Wiley-VCH, New
W.L. Hall, A. Cassidy, Am. J. Clin. Nutr. 88 (2008) 38. York, 1997.
[5] G.J. Soleas, E.P. Diamandis, D.M. Goldberg, Clin. Biochem. 30 (1997) 91. [21] J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction, Royal Institute
[6] P.S. Makena, K.T. Chung, Food Chem. Toxicol. 45 (2007) 1899. of Chemistry, 1999.
[7] L.L. Lin, C.Y. Lien, Y.C. Cheng, K.L. Ku, J. Chromatogr. B 853 (2007) 175. [22] C. Dietz, J. Sanz, C. Cámara, J. Chromatogr. A 1103 (2006) 183.
[8] B. Sun, M.C. Leandro, V. de Freitas, M.I. Spranger, J. Chromatogr. A 1128 (2006) [23] L. Pan, J. Pawliszyn, Anal. Chem. 69 (1997) 196.
27. [24] T.G. Luan, G.K. Li, Z.X. Zhang, Anal. Chim. Acta 424 (2000) 19.
[9] G.J. Soleas, J. Yan, D.M. Goldberg, Am. J. Enol. Viticult. 46 (1995) 159. [25] Y. Shao, P. Marriott, H. Hügel, Chromatographia 57 (2003) S349.
[10] A. Antoneli, C. Fabbri, G. Lercker, Chromatographia 42 (1996) 469. [26] R. Flamini, Mass Spectrom. Rev. 22 (2003) 218.
[11] G.J. Soleas, J. Yan, D.M. Goldberg, J. Chromatogr. B 757 (2001) 161. [27] R. Flamini, A.D. Vedova, Rapid Commun. Mass Spectrom. 18 (2004) 1925.