Article

Headspace solid-phase microextraction for direct determination of volatile phenols in cider

Department of Chemistry, University of La Rioja, Logroño, La Rioja, Spain.
Journal of Separation Science (Impact Factor: 2.74). 11/2009; 32(21):3746-54. DOI: 10.1002/jssc.200900347
Source: PubMed

ABSTRACT

A headspace solid-phase microextraction coupled to gas chromatography-tandem mass spectrometry (GC-MS/MS) method was optimised and validated for the determination of 4-ethylguaiacol, 4-ethylphenol, 4-vinylguaiacol and 4-vinylphenol, involved in the presence of Brett character, in ciders. The influence of different parameters on extraction efficiency (fibre coating, salt addition, exposure time, extraction temperature and sample volume/total volume ratio) was evaluated. Divinylbenzene/carboxen/PDMS was selected as extraction fibre and the other optimised parameters were as follows: 10 mL of cider, temperature 70 degrees C, extraction time 60 min and addition of 0.4 g/mL of NaCl. The proposed method showed satisfactory linearity. The detection limits obtained were 0.01 microg/L for 4-ethylguaiacol, 0.02 microg/L for 4-ethylphenol, 0.08 microg/L for 4-vinylguaiacol and 0.03 microg/L for 4-vinylphenol. These detection limits were lower than those obtained in previous studies on the determination of volatile phenols in other alcoholic beverages. Good recoveries of over 95% were observed for all compounds, and the repeatability obtained was considered acceptable, ranging between 4 and 10%. To demonstrate the feasibility of the procedure, the method was applied to the analysis of commercial ciders. To our knowledge, this is the first time that the headspace solid-phase microextraction procedure has been optimised to determine specifically the Brett character responsible compounds in cider.

Full-text

Available from: Consuelo Pizarro, Oct 02, 2014
Research Article
Headspace solid-phase microextraction for
direct determination of volatile phenols in
cider
A headspace solid-phase microextraction coupled to gas chromatography-tandem mass
spectrometry (GC-MS/MS) method was optimised and validated for the determination of
4-ethylguaiacol, 4-ethylphenol, 4-vinylguaiacol and 4-vinylphenol, involved in the presence
of Brett character, in ciders. The influence of different parameters on extraction efficiency
(fibre coating, salt addition, exposure time, extraction temperature and sample volume/
total volume ratio) was evaluated. Divinylbenzene/carboxen/PDMS was selected as
extraction fibre and the other optimised parameters were as follows: 10 mL of cider,
temperature 701C, extraction time 60 min and addition of 0.4 g/mL of NaCl. The proposed
method showed satisfactory linearity. The detection limits obtained were 0.01 mg/L for
4-ethylguaiacol, 0.02 mg/L for 4-ethylphenol, 0.08 mg/L for 4-vinylguaiacol and 0.03 mg/L
for 4-vinylphenol. These detection limits were lower than those obtained in previous
studies on the determination of volatile phenols in other alcoholic beverages. Good
recoveries of over 95% were observed for all compounds, and the repeatability obtained
was considered acceptable, ranging between 4 and 10%. To demonstrate the feasibility of
the procedure, the method was applied to the analysis of commercial ciders. To our
knowledge, this is the first time that the headspace solid-phase microextraction procedure
has been optimised to determine specifically the Brett character responsible compounds
in cider.
Keywords: Brett character / Cider / Ethylphenols / Solid-phase microextraction
(SPME) / Vinylphenols / Volatile phenols
DOI 10.1002/jssc.200900347
1 Introduction
Cider is an alcoholic beverage made from the fermented juice
of apples. The production of cider is an important economic
resource in European regions where apple cultivars are viable
[1]. One of the main objectives of cider makers is to produce a
safe, high-quality product without reducing the rich flavours,
aroma and texture attributable to apple cider [2].
One of the off-flavours that can appear in apple ciders is
Brett character [3–6]. This defect is associated with the
presence of volatile phenols (4-ethylphenol, 4-ethylguaiacol,
4-vinylphenol and 4-vinylguaiacol). These compounds are
usually present in alcoholic beverages, but when present in
high concentrations they are considered as off-flavours that
negatively affect their quality [6]. The Brettanomyces/Dekkera
yeast seems to be mainly responsible for the production of
volatile phenols [6–8]. The Brettanomyces/Dekkera yeasts
were first described in 1903 by Claussen in beer production
[9]. This genus has been well known for a long time as a
contaminant in the beer, cider and carbonated drink
industries [10]. This yeast can develop in contaminated
ciders and ciders with residual sugar. Cider sulphiting, once
alcoholic and malolactic fermentations are complete, is a
solution often used to eliminate bacteria and yeasts that
cause undesirable alterations. However, this practice is
rarely used to produce natural cider. Indeed, it is preferred
to keep the addition of chemicals at a minimum level in
order to preserve the organoleptic qualities of the final
product [11, 12]. Nowadays, the determination of volatile
phenols in cider is of great interest to the cider industry so
as to guarantee cider quality and avoid financial losses.
The correct quantification of the volatile components in
alcoholic beverages is an important challenge for analysts,
and cider is no exception. Major volatiles in apple cider have
been analysed by direct injection [1, 13], since aroma
compounds are usually present at a very low concentration
level (minor volatiles), several procedures have been proposed
for isolating and concentrating them prior to gas chromato-
Consuelo Pizarro
Nuria Pe
´
rez-del-Notario
Jose
´
Marı
´
a Gonza
´
lez-Sa
´
iz
Department of Chemistry,
University of La Rioja, Logron
˜
o,
La Rioja, Spain
Received May 18, 2009
Revised July 31, 2009
Accepted August 1, 2009
Abbreviations: CAR, carboxen; DVB, divinylbenzene; HS,
headspace; SPME, solid-phase microextraction; Vs, sample
volume; Vt, total volume
Correspondence: Professor Consuelo Pizarro, Department of
Chemistry, University of La Rioja, C/ Madre de Dios 51, 26006
Logron
˜
o, La Rioja, Spain
E-mail: consuelo.pizarro@unirioja.es
Fax: 0034941299621
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
J. Sep. Sci. 2009, 32, 3746–3754
3746
Page 1
graphy. Sample preparation strategies that have been
proposed to analyse volatile compounds in cider comprise
classical liquid–liquid extraction with organic solvents [4, 14]
(which is time consuming, hazardous to health and expen-
sive due to the large amounts of organic solvent required),
purge and trap [1, 15], SPE [16] and solid-phase micro-
extraction (SPME) [2, 17]. Volatile phenols are usually
analysed by GC coupled to an adequate detection system
(flame ionisation detection or MS). HPLC (with fluorometric
or coulometric detection) has also been used, although less
extensively than GC [18]. These techniques are usually
preceded by an extraction step. Liquid–liquid extraction
methods with organic solvents were traditionally employed
[19–21], but nowadays, easier and more selective extraction
methods are used, such as SPE [22, 23], stir bar sorptive
extraction [24] and, more recently, dispersive liquid–liquid
microextraction [25]. SPME is a fast sensitive technique for
volatile extraction, which allows sample preparation time to
be reduced and does not require the use of solvents. In
addition, very good detection limits can often be achieved.
This technique has already been used satisfactorily to deter-
mine volatile phenols, in water, whisky, beer, wine and more
recently in oil [26–31], but to our knowledge, it has not been
suitably optimised yet for the simultaneous determination of
ethylphenols and vinylphenols in cider.
The aim of this study was to optimise and validate a
headspace (HS) SPME method for the determination of 4-
ethylphenol, 4-ethylguaiacol, 4-vinylphenol and 4-vinyl-
guaiacol in cider, which are the main compounds respon-
sible for Brett character. For this purpose, after selecting the
type of fibre, several SPME parameters that influence the
extraction process (the exposure time, extraction tempera-
ture and sample volume (Vs)/total volume (Vt) ratio) were
optimised by means of experimental design methodology.
Effects of experimental parameters were evaluated using a
Dohelert design. This experimental design allowed us to
study the effect of a variation in experimental factor levels
on response with a minimum number of experiments [32].
The volatile phenols were determined using GC/MS/MS
detection. The quality parameters of the HS-SPME method
were established and it was used for the determination of
the target compounds in real samples. To our knowledge,
this is the first time that the HS-SPME procedure has been
optimised to determine specifically Brett character respon-
sible compounds in cider. This method could be of great
interest for the cider industry, given the increasing impor-
tance of this organoleptical defect in ciders, above all when
materials that have not been adequately cleaned are used.
2 Materials and methods
2.1 Chemicals
4-Ethylguaiacol, 4-ethylphenol, 4-vinylguaiacol and 4-vinyl-
phenol were supplied by Aldrich Chemie (Steinheim,
Germany). The purity of all standards was above 98%.
Sodium chloride was obtained from Aldrich Chemie.
Methanol was purchased from Merck (Darmstadt,
Germany).
2.2 Standard solutions and samples
Individual standard stock solutions of each compound were
prepared in methanol. Work solutions used for further
studies were prepared by diluting different amounts of each
standard stock solution. Standard and work solutions were
stored in darkness at 41C.
Five ciders from different Spanish cider-producing
areas were selected to be analysed in the study. Samples
were spiked with different amounts of work solutions
containing the target analytes.
2.3 HS-SPME procedure
The evaluated fibres were purchased from Supelco (Belle-
fonte, PA, USA) and were coated with different stationary
phases: PDMS (100 mm), PDMS/divinylbenzene (DVB,
65 mm), polyacrylate (85 mm), carboxen (CAR)/PDMS
(75 mm), DVB/CAR/PDMS (50/30 mm) and carbowax/DVB
(70 mm). They were conditioned in accordance with the
producer’s specifications before use.
The main parameters that affect the SPME process (i.e.
ionic strength, fibre type, time, extraction temperature and
Vs) were studied. HS extraction was preferred to direct
extraction to prevent the fibre coming into direct contact
with the matrix.
For each SPME analysis, different aliquots of cider
(from 4 to 12 mL) were placed in a 20 mL HS vial and then
the vial was tightly sealed with a PTFE septum. Vials were
then incubated at the temperature selected for the extraction
(from 50 to 901C) for 5 min before SPME extraction. Then,
the fibre was exposed to the HS over the sample from 30 to
90 min, depending on the experiment. Once the extraction
step had been completed, the fibre was retracted and the
SPME device was removed from the vial. The SPME device
was then inserted into the injection port of a GC/MS/MS
system for thermal desorption at the maximum recom-
mended operating temperature for each fibre for 5 min
(2 min of splitless). Blank runs were completed at least once
daily before sampling to ensure no carryover of analytes
from previous extractions and to clean the fibre before the
analysis in order to remove potential interferences.
2.4 Equipment and chromatographic conditions
The HS-SPME-GC/MS/MS analyses were performed with a
Varian 3800 gas chromatograph (Walnut Creek, CA, USA)
equipped with a Combipal Autosampler (CTC Analytics,
Zwingen, Switzerland) and connected to an ion-trap mass
spectrometer (Varian Saturn 2200). Compounds were
J. Sep. Sci. 2009, 32, 3746–3754 Gas Chromatography 3747
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 2
separated using a CP-WAX 52-CB column (30 m 0.25 mm
id, 0.25 mm film thickness) from Varian. Helium, at a flow
of 1 mL/min, was used as a carrier gas. Oven temperature
was programmed as follows: 351C for 2 min, heated at 201C/
min to 1701C and kept for 1 min and finally raised to 2101C
at 31C/min and held for 12 min. Injection was performed in
splitless mode for 2 min and then split was set at 30 mL/
min. An inlet of 0.75 mm id was used and the injector
temperature, after the optimisation stage, was fixed at
2701C. The manifold, GC/MS interface and ion trap
temperatures were set at 60, 280 and 2001C, respectively.
Mass spectra were obtained using electron impact ionisation
(70 eV). Precursor ions were isolated using a 3 amu isolation
window and subjected to CID. For operating in MS-MS
mode, the emission current was fixed at 80 mA and scan
time at 0.46 s/scan. The other MS/MS parameters are
summarised in Table 1.
2.5 Software
The construction and analyses of the experimental design
for reaching the optimum HS-SPME conditions were
carried out using the Nemrod-W statistical package [33].
3 Results and discussion
3.1 Optimisation of microextraction conditions
The influence of different parameters that can affect
extraction efficiency was evaluated using cider samples
spiked with ethylphenols and vinylphenols (40 mg/L for each
compound). Of the different variables that could improve
SPME efficiency, we selected the following parameters for
optimisation: fibre coating, extraction time, extraction
temperature and Vs/Vt ratio. Salt addition was also studied
to prove its effect on the microextraction procedure [34].
3.1.1 Fibre selection
In the first stage of SPME optimisation, the effect of fibre
coating on the extraction yield of the selected volatiles was
evaluated using the spiked cider samples. The influence of
physical and chemical properties of fibre coating on SPME
efficiency is clear. Fibre composition determines the nature
of analytes extracted from the sample, because it determines
the distribution coefficients of the compounds between the
fibre and the sample. As previous studies to determine
volatile phenols [27, 30, 35–38] showed differing results as
regards the efficiency of fibre coatings, it was decided to
perform a thorough study in order to test the optimal
conditions for extracting the target compounds in ciders. In
this research, six different fibre coatings (PDMS 100 mm,
PDMS/DVB 65 mm, polyacrylate 85 mm, CAR/PDMS 75 mm,
DVB/CAR/PDMS 50/30 mm and carbowax/DVB 70 mm)
and five temperatures were evaluated in order to compare
their efficiency. Figure 1 shows the chromatograms
obtained after the extraction with the six different fibre
coatings studied at 801C. Figure 2 summarizes the results of
the fibre screening process. The extraction behaviour
pattern for 4-ethylguaiacol and 4-vinylguaiacol was very
similar. The highest responses were achieved using DVB/
CAR/PDMS and CAR/PDMS fibres, whereas the lowest
extraction efficiency for these compounds was obtained by
PDMS. In the case of 4-ethylphenol, except for PDMS, all
fibres provided satisfactory results, with DVB/CAR/PDMS
fibre efficiency standing out from the rest. Finally, 4-
vinylphenol extraction profiles showed a relevant depen-
dence on temperature. Low extraction temperatures
provided poor responses. When temperature was increased
extraction efficiency improved for most of the fibre coatings
studied. This could be due to an increase in extraction
temperature that translates to increased diffusion coeffi-
cients and Henry’s constants, and decreased distribution
constant. As Henry’s constants increase, HS concentrations
increase too [39]. Taking into account the relative olfactory
level thresholds of the studied compounds [28, 30, 40], we
mainly looked for the highest sensitivity for 4-ethylphenol
(which has the lowest olfactory threshold), followed by
4-ethylguaiacol and 4-vinylguaiacol; whereas 4-vinlyphenol
is the compound with highest olfactory threshold. This
higher sensitivity could be achieved by using the CAR/
PDMS and DVB/CAR/PDMS fibres, but with the former
the reproducibility of the method decreased, above all when
temperature was increased. Because of this increase in the
variability of the procedure, DVB/CAR/PDMS fibre was
selected to improve the precision of the proposed method.
This fibre coating had already been proposed in previously
published methods to determine volatile compounds in
ciders [17] and volatile phenols in other matrices, such as
wine, beer and oil [27, 30, 31].
Table 1. Retention time and MS/MS detection parameters for volatile phenols using the proposed method
Compound Retention time (min) Precursor ion (m/z) Quantification ions (m/z) CID parameters
Storage level (m/z) Amplitude (V)
4-Ethylguaiacol 13.005 137 91 75 80
4-Ethylphenol 15.205 107 77 60 69
4-Vinylguaiacol 15.557 150 107 80 72
4-Vinylphenol 19.342 120 91 65 64
J. Sep. Sci. 2009, 32, 3746–37543748 C. Pizarro et al.
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 3
Based on these preliminary studies, the experimental
domain was successfully reduced for a subsequent optimi-
sation of the Vs, extraction time and extraction temperature
in order to obtain an optimal compromise situation for all
compounds.
3.1.2 Sodium chloride addition
The effect of altering the ionic strength of the matrix was
studied by adding different amounts of sodium chloride.
Analyte solubility usually decreases when ionic strength
increases. A decrease in analyte solubility improves
sensitivity by promoting analyte partitioning into the
stationary phase, but the ‘‘salting-out’’ effect is compound
dependent [34, 39]. Three different levels were evaluated in
cider analyses: no sodium chloride addition, undersatura-
tion (0.2 g/mL) and supersaturation (0.4 g/mL). Each
experiment was performed in triplicate. Figure 3 shows
the influence of sodium chloride concentration on the
efficiency of the SPME. As it can be seen, both families of
compounds, i.e. ethylphenols and vinylphenols, showed
similar behaviour patterns. When the amount of sodium
chloride was increased, a significant improvement in
volatile phenol sensitivity was observed. Therefore, in order
4-EG
0
200000
400000
600000
800000
1000000
30ºC 40ºC 60ºC 80ºC 95ºC
30ºC 40ºC 60ºC 80ºC 95ºC 30ºC 40ºC 60ºC 80ºC 95ºC
30ºC 40ºC 60ºC 80ºC 95ºC
TEMPERATURE
AREA (counts)
4-EP
0
500000
1000000
1500000
2000000
TEMPERATURE
AREA (counts)
4-VG
0
200000
400000
600000
800000
TEMPERATURE
AREA (counts)
4-VP
0
200000
400000
600000
800000
1000000
TEMPERATURE
AREA (counts)
AB
CD
Figure 2. Influence of fibre type
and extraction temperature on
the HS-SPME process: (A) 4-
ethylguaiacol, (B) 4-ethylphenol,
(C) 4-vinylguaiacol and (d) 4-
vinylphenol.
Figure 1. GC-MS/MS chromatograms of
volatile phenols for the studied fibres.
J. Sep. Sci. 2009, 32, 3746–3754
Gas Chromatography 3749
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 4
to improve the sensitivity of the method for all compounds,
it was decided to carry out extraction with the addition of
0.4 g/mL of sodium chloride in further studies.
3.1.3 Doehlert experimental design: Evaluation of the
influence of the exposure time, extraction tempera-
ture and Vs/Vt
The influence of each experimental factor on the extraction
of volatile phenols was evaluated by means of Doehlert
experimental design. Doehlert design requires fewer experi-
ments than other experimental designs, as it is more
efficient and permits the choice of the number of levels for
each optimised variable according to their impact on the
response. The suitability of Doehlert experimental designs
for the optimisation of SPME procedures has been
previously demonstrated [41, 42]. Three factors were
included in the Doehlert design, namely: extraction time,
extraction temperature and Vs/Vt ratio. Assigned values to
each factor were selected taking into account the results of
the preliminary experiments. DVB/CAR/PDMS fibre was
selected and a narrower range of temperatures (50–90 1C)
that provided the best extraction yield was selected. Since
extraction temperature is one of the most important factors
affecting microextraction procedures, it was studied at seven
levels. Extraction time was studied at five levels, between 30
and 90 min, and Vs/Vt ratio was evaluated at three levels in
the range 0.20–0.60. Doehlert’s experimental matrix [32, 43]
comprising 19 experiments was used (experiments were
performed in triplicate). In order to estimate experimental
repeatability, three of the experiments were performed at
the centre of the domain. Four test points fixed at the
vertices of a regular tetrahedron within a sphere with a
radius of 0.5 (coded value) were also included. This model
was used to obtain the surface response fitting the data to a
polynomial model, the evaluation of the effects of each
factor and the interaction effects between factors. All
experiments were randomly performed. The Doehlert
factors considered, the experimental conditions studied
and the average value of the experimental responses
obtained are shown in Table 2.
The estimates of the coefficients for the models of each
response were calculated by least squares linear regression
and these models were analysed and validated by the
analysis of variance and the test points using Nemrod-W
software [33]. It was demonstrated that the proposed
mathematical models were significant for all compounds,
and correctly explain the behaviour of the compounds
throughout the experimental domain. Therefore, the models
were accepted and the results analysed in detail. Model
coefficients for each response are shown in Table 3. As
might be expected, extraction time and temperature were
significant for all compounds. For 4-ethylguaiacol the
influence of all factors in the extraction efficiency was
0
20
40
60
80
100
120
4-EG 4-EP 4-VG 4-VP
Compound
Normalised response (%)
no addition 0.2g/ml NaCl 0.4g/ml NaCl
Figure 3. Influence of NaCl concentration on the efficiency of
SPME for 4-ethylguaiacol, 4-ethylphenol, 4-vinylguaiacol and
4-vinylphenol using a DVB/CAR/PDMS fibre (n 5 3). Responses
were normalized to the maximum signal achieved for each
response.
Table 2. Experimental design matrix and response (mean
values) obtained for volatile phenols
No.
exp
Time
(min)
Temperature
(1C)
Vs/Vt Results (area counts)
EG EP VG VP
1 90 70 0.40 127 350 181 050 631 357 224 792
2 30 70 0.40 122 250 172 650 333 262 64 376
3 75 90 0.40 56 075 127 950 620 412 252 288
4 45 50 0.40 75 711 121 200 200 841 114 364
5 75 50 0.40 104 900 160 150 283 841 133 323
6 45 90 0.40 65 869 133 350 538 393 231 251
7 75 77 0.60 104 158 159 600 685 448 279 206
8 45 63 0.20 107 350 166 150 328 884 177 828
9 75 63 0.20 122 750 186 200 477 786 218 327
10 60 83 0.20 93 927 155 500 497 610 163 207
11 45 77 0.60 125 300 163 350 656 367 360 615
12 60 57 0.60 147 900 174 750 424 919 360 615
13 60 70 0.40 151 100 180 900 619 945 295 701
14 60 70 0.40 154 800 184 900 701 419 270 916
15 60 70 0.40 154 100 174 600 621 069 305 548
16 48 65 0.36 151 400 178 750 454 750 235 194
17 72 65 0.36 155 200 190 400 579 515 209 646
18 60 81 0.36 115 123 157 100 650 761 217 465
19 60 70 0.52 140 450 180 850 643 743 352 065
Table 3. Estimates of model coefficients for the responses
Coefficient EG EP VG VP
b
0
153 008 181 299 635 133 282 051
b
1
2743 8291 118 310 36 693
b
2
21 458 9993 196 194 58 126
b
3
9958 2028 94 007 65 282
b
11
27 479 4061 153 741 138 050
b
22
93 893 59 685 247 923 89 341
b
33
24 307 4529 85 496 10 069
b
12
22 282 25 567 1425 4308
b
13
14 261 5495 73 782 72 879
b
23
31 080 13 682 72849 115 247
Bold numbers denote significant effects (5%).
J. Sep. Sci. 2009, 32, 3746–3754
3750 C. Pizarro et al.
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 5
statistically significant. Since interaction coefficients, time-
extraction temperature and extraction temperature-Vs/Vt
ratio were also significant, it was necessary to study the
existing interactions with the help of response surfaces.
Several conclusions could be drawn from the response
surface for interaction time-extraction temperature (Fig.
4A). When Vs was fixed at an intermediate level (8 mL), a
maximum response zone was clearly defined (region in red),
this zone comprises extraction times between 45 and 85 min
and temperatures in the range of 60–751C. In the case of the
response surface for extraction temperature-Vs/Vt interac-
tion, fixing extraction time at 60 min (Fig. 4B), optimum
values for response were achieved varying temperature
between 50 and 851C, regardless of the Vs employed. For 4-
ethylphenol only extraction time–temperature interaction
coefficient was statistically significant. When this interac-
tion was graphically represented (Fig. 4C), a maximum
response region was noticeable for times of more than
75 min, although 50 min were enough to obtain acceptable
responses, varying temperature in the range of 60–701C.
When extraction time and temperature increased, analyte
response decreased, probably due to the fact that increased
temperature decreases the distribution constant of the
analyte between the sample matrix and the fibre coating
and/or due to displacement effects that occur at long
extraction times. When analytes with more affinity with the
coating displace the analytes with less affinity for 4-vinyl-
guaiacol significant coefficients were as follows: extraction
time, temperature and Vs/Vt ratio. The first two factors were
significant in their quadratic form, implying non-linear
behaviour for the response. The sign of Vs/Vt coefficient
was positive, providing bigger signals when extraction
volume was increased. If Vs was fixed at 8 mL (Fig. 4D),
optimum conditions were achieved with extraction times
between 60 and 80 min, with temperatures settled in a range
of 70–801C. When 4-vinylphenol coefficients were analysed,
it was observed that the extraction time effect was statisti-
cally significant in its quadratic form. As extraction time was
increased, we observed an increase in the response, until the
equilibrium situation was reached at 60 min. The interac-
Figure 4. Response surface extraction time versus extraction temperature for 4-ethylguaiacol (A), 4-ethylphenol (C), 4-vinylguaiacol
(D) and temperature versus Vs/Vt ratio for 4-ethylguaiacol (B) and 4-vinylphenol (E).
J. Sep. Sci. 2009, 32, 3746–3754
Gas Chromatography 3751
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 6
tion coefficient temperature-Vs/Vt ratio was also significant,
which meant that both factors had to be studied together for
this compound. Figure 4E shows the response surface for
extraction temperature versus Vs/Vt ratio for the 4-vinyl-
phenol (time fixed at 60 min), the higher responses for this
compound were obtained with temperatures higher than
751C and Vs higher than 11 mL.
To select the ideal conditions for the HS-SPME process,
we took into account all the results obtained, bearing in
mind the olfactory thresholds of the compounds, and trying
to save time in the extraction process. Thus, the acceptable
optimum compromise situation for the extraction of volatile
phenols in cider was reached when the extraction time was
set at 60 min, the extraction temperature at 701C and the
Vs/Vt ratio at 0.50 (10 mL of cider).
3.2 Method performance
After the optimisation of the SPME influencing parameters
analytical quality parameters were established using cider
samples spiked with the four volatile phenols. Linearity of
the response, limits of detection and quantification of the
target compounds, repeatability and recovery results were
evaluated in order to assess the performance of the HS-
SPME-GC/MS/MS proposed method. Linearity was studied
at five different concentration levels, between 2 and
2000 m g/L for 4-ethylphenol, 4-ethylguaiacol and 4-vinyl-
guaiacol, and from 2 to 1500 mg/L for 4-vinylphenol. Linear
range, correlation coefficients and quantification and detec-
tion limits are shown in Table 4. An adequate linearity was
obtained for all analytes, with correlation coefficients
varying between 0.993 and 0.998 over the linear range.
LOQ and LOD were calculated for a S/N of 10 and 3,
respectively. These detection limits were lower than those
obtained in other studies relating to the determination of
volatile phenols in other alcoholic beverages [29, 35, 36, 44].
The repeatability of the proposed method was evaluated by
performing five analyses of spiked cider samples prepared at
three different concentration levels on the same day. The
results, expressed in terms of RSDs, are shown in Table 5,
and ranged from 4.5 to 10%. For recovery experiments each
compound was added at two different concentration levels,
at each level three replicates were performed. Relative
recoveries can also be seen in Table 5. Recoveries higher
than 95% were obtained for all compounds.
3.3 Application of the method to real samples
Once the method had been optimised and evaluated, we
applied it to determine the quantity of volatile phenols
present in cider commercial samples. In order to avoid
possible matrix effects, the standard addition technique was
proposed for analyte quantification. Five different ciders
extracted under the optimised HS-SPME conditions were
analysed by GC-MS/MS. Each determination was made in
triplicate. A summary of the results obtained in the analyses
is shown in Table 6.
As shown in the table, all analysed ciders contained all
the ethylphenols and vinylphenols. Ciders B and D were
selected because they came from a press-house where the
Brettanomyces yeast was suspected to be present. Cider
quantification results were compared with their corre-
sponding sensorial analyses. Ciders B and D showed a
significant ‘‘bretty’’ note; in these cases, the levels of volatile
phenols were very high compared to commercial ciders A, C
and E.
This study confirms the potential impact of volatile
phenols on organoleptical defects present in cider. More-
over, it proves the suitability of this technique to assess the
presence of target compounds in this alcoholic beverage.
Table 4. Linear range, correlation coefficients, LOD and LOQ for
the proposed method (n 5 3)
Compound Linear
range
(mg/L)
Correlation
coefficient
(r
2
)
LOQ
S/N 5 10
(mg/L)
LOD
S/N 5 3
(mg/L)
4-Ethylguaiacol 2.71–2263 0.994 0.03 0.01
4-Ethylphenol 2.72–1829 0.994 0.08 0.02
4-Vinylguaiacol 2.90–1474 0.993 0.26 0.08
4-Vinylphenol 2.67–1801 0.998 0.09 0.03
Table 5. Summary of repeatability and recovery studies for the
proposed method (n 5 3)
Compound Repeatability (RSD%)
n 5 5
Average recoveries
7RSD (%)
10 mg/L 100 mg/L 1000 mg/L Low level
a)
High level
b)
4-Ethylguaiacol 4.56 8.37 5.03 98.1776.58 99.2275.48
4-Ethylphenol 8.10 7.54 6.45 95.2477.25 97.4176.63
4-Vinylguaiacol 6.72 5.25 9.58 96.6276.13 95.7378.42
4-Vinylphenol 5.30 5.45 6.39 97.2175.27 98.3474.91
a) Spiked concentration 10 mg/L.
b) Spiked concentration 1000 mg/L.
Table 6. Results of an analysis of commercial cider samples by
the HS-SPME-GC/MS/MS proposed method (n 5 3)
Compound Concentration 7 SD (mg/L)
Cider A Cider B Cider C Cider D Cider E
4-Ethylguaiacol 2574 109753072 130792172
4-Ethylphenol 9578 749711 12579 983715 7576
4-Vinylguaiacol 3975 22278 18378 324711 1073
4-Vinylphenol 7176 1125710 11477 1241716 9077
J. Sep. Sci. 2009, 32, 3746–37543752 C. Pizarro et al.
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 7
4 Concluding remarks
In this paper, a method based on HS-SPME coupled to GC-
MS/MS was optimised for the determination of volatile
phenols in ciders. A study of the different types of coatings
at diverse temperatures was performed, which demon-
strated that the DVB/CAR/PDMS fibre was the most
appropriate for the determination of the target compounds
in cider. Salt addition was also studied to prove its effect on
the microextraction procedure. Then, to optimise the
procedure, the influence of each experimental factor
(exposure time, extraction temperature and Vs/Vt) on the
extraction of volatile phenols was evaluated using the
Doehlert experimental design. Optimised HS-SPME condi-
tions were as follows: 10 mL of cider placed in a 20 mL HS
vial at 701C for 60 min, with the addition of 0.4 g/mL of
NaCl. The proposed method showed satisfactory linearity,
precision and detection limits. These detection limits were
lower than those obtained in previous studies to determine
volatile phenols in other alcoholic beverages. The applic-
ability of the proposed method was demonstrated by
analysing several ciders. To our knowledge, this is the first
time that the HS-SPME procedure has been optimised to
determine simultaneously the Brett character responsible
compounds in cider. This method could be of great interest
for the cider industry, taking into account the increasing
importance of this organoleptical defect in ciders.
The authors thank the Spanish Government Ministerio de
Ciencia e Innovacio
´
n for its financial support (project
CTQ2008-03493/BQU) and Professor R. Phan-Tan-Luu of the
University of Marseille (France) for providing the NEMROD-W
software.
The authors have declared no conflict of interest.
5 References
[1] Rodrı
´guez-M
adrera, R., Garcı
´a-Hevia,
A., Palacios-
Garcı
´
a, N., Su
´
arez-Valles, B., LWT Food Sci. Technol.
2008, 41, 2064–2069.
[2] Rye, G. G., Mercer, D. G., Food Res. Int. 2003, 36,
167–174.
[3] Bueno, G., Aramburu, T., Gonza
´
lez, E., Cabranes, C.,
The cider book (El libro de la sidra). Ed. Pentalfa, Oviedo
1991, p. 529.
[4] Le Que
´
re
´
, J. M., Husson, F., Renard, C. M. G. C.,
Primault, J., LWT Food Sci. Technol. 2006, 39,
1033–1044.
[5] Cocolin, L., Rantsiou, K., Iacumin, L., Zironi, R., Comi,
G., Appl. Environ. Microbiol. 2004, 70, 1347–1355.
[6] Etievant, P. X., in: Maarse, H. (Ed.), Volatile Compounds
in Foods and Beverages, Marcel Dekker, New York 1991,
pp. 483–546.
[7] Heresztyn, T., Arch. Microbiol. 1986, 146, 96–98.
[8] Chatonnet, P., Dubourdieu, D., Boidron, J. N., Pons, M.,
J. Sci. Food Agric. 1992, 60, 165–178.
[9] Gilliland, R. B., J. Inst. Brew. 1961, 67, 257–261.
[10] Deak, T., Beuchat, L. R., Handbook of Spoilage Yeast,
CRC Press, New York 1996.
[11] del Campo, G., Berregi, I., Santos, J. I., Duen
˜
as, M.,
Irastorza, A., Bioresour. Technol. 2008, 99, 2857–2863.
[12] Duen
˜
as, M., Irastorza, A., Munduate, A., Santos, J. I.,
Berregi, I., Del Campo, G., J. Inst. Brew. 2002, 108,
243–247.
[13] Rodrı
´
guez-Madrera, R., Sua
´
rez-Valles, B., J. Chroma-
togr. Sci. 2007, 45, 428–434.
[14] Williams, A. A., May, H. V., J. Inst. Brew. 1981, 87,
372–375.
[15] Rodrı
´guez-Mad
rera, R., Garcı
´a-Hevia,
A., Palacios-
Garcı
´a,
N., Su
´
arez-Valles, B., J. Chromatogr. A 2005,
1069, 245–251.
[16] Mangas, J. J., Gonza
´
lez, M. P., Rodrı
´guez,
R., Blanco, D.,
Chromatographia 1996, 42, 101–105.
[17] Xu, Y., Fan, W., Qian, M. C., J. Agric. Food Chem. 2007,
55, 3051–3057.
[18] Larcher, R., Nicolini, G., Puecher, C., Bertoldi, D., Moser,
S., Favaro, G., Anal. Chim. Acta 2007, 582, 55–60.
[19] Pollnitz, A. P., Pardon, K. H., Sefton, M. A., J. Chroma-
togr. A 2000, 874, 101–109.
[20] Pe
´
rez-Prieto, L. J., Lo
´
pez-Roca, J. M., Martı
´
nez-Cutillas,
A., Pardo-Mı
´
nguez, F., Go
´
mez-Plaza, E., J. Agric. Food
Chem. 2003, 51, 5444–5449.
[21] Falque
´
, E., Darriet, P., Ferna
´
ndez, E., Dubourdieu, D., Int.
J. Food Sci. Technol. 2008, 43, 464–475.
[22] Lo
´
pez, R., Aznar, M., Cacho, J., Ferreira, V., J. Chroma-
togr. A 2002, 966, 167–177.
[23] Domı
´nguez,
C., Guille
´
n, D. A., Barroso, C. G., Anal.
Chim. Acta 2002, 458, 95–102.
[24] Monje, M. C., Privat, C., Gastine, V., Nepveu, F., Anal.
Chim. Acta 2002, 458, 111–117.
[25] Farin
˜
a, L., Boido, E., Carrau, F., Dellacassa, E., J. Chro-
matogr. A 2007, 1157, 46–50.
[26] Mousavi, M., Noroozian, E., Jalali-Heravi, M., Molla-
hosseini, A., Anal. Chim. Acta 2007, 581
, 71–77.
[27] Rodrigues, F., Caldeira, M., Ca
´
mara, J. S., Anal. Chim.
Acta, 2008, 609, 82–104.
[28] Boutou, S., Chatonnet, P., J. Chromatogr. A 2007, 1141,
1–9.
[29] Pizarro, C., Pe
´
rez-del-Notario, N., Gonza
´
lez-Sa
´
iz, J. M.,
J. Chromatogr. A 2007, 1143, 176–181.
[30] Vichi, S., Romero, A., Tous, J., Tamames, E. L., Buxa-
deras, S., J. Chromatogr. A 2008, 1211, 1–7.
[31] Pizarro, C., Pe
´
rez-del-Notario, N., Gonza
´
lez-Sa
´
iz, J. M.,
J. Chromatogr. A 2007, 1166, 1–8.
[32] Lewis, G. A., Mathieu, D., Phan-Tan-Luu, R., Pharma-
ceutical Experimental Design, Marcel Dekkler, New York
1999.
[33] Mathieu, D., Nony, J., Phan-Tan-Luu, R., NEMROD-W,
version 2001, LPRAI, Marseille 2001.
[34] Pawliszyn, J., Solid-Phase Microextraction-Theory and
Practise, Wiley-VCH, Weinheim 1997.
J. Sep. Sci. 2009, 32, 3746–3754 Gas Chromatography 3753
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 8
[35] Castro, R., Natera, R., Garcı
´a,
M. V., Garcı
´a
Barroso, C.,
J. Chromatogr. A 2003, 995, 11–20.
[36] Martorell, N., Martı
´,
M. P., Mestres, M., Busto, O.,
Guasch, J., J. Chromatogr. A 2002, 975, 349–354.
[37] Monje, M. C., Privat, C., Gastine, V., Nepveu, F., Anal.
Chim. Acta 2002, 458, 111–117.
[38] Romano, A., Perello
´
, M. C., Lonvaud-Funel, A., Sicard,
G., de Revel, G., Food Chem. 2009, 114, 15–19.
[39] Lord, H., Pawliszyn, J., J. Chromatogr. A 2000, 902, 17–63.
[40] Cullere, L., Escudero, A., Cacho, J., Ferreira, V., J. Agric.
Food Chem. 2004, 52, 1653–1660.
[41] Martı
´nez-Urun˜
uela, A., Gonza
´
lez-Sa
´
iz, J. M., Pizarro, C.,
J. Chromatogr. A 2004, 1056, 49–56.
[42] Pizarro, C., Pe
´
rez-del-Notario, N., Gonza
´
lez-Sa
´
iz, J. M.,
J. Chromatogr. A 2007, 1143, 26–35.
[43] Doehlert, D. H., Appl. Statist. 1970, 19, 231–239.
[44] Domı
´nguez,
C., Guille
´
n, D. A., Barroso, C. G., Anal.
Chim. Acta 2002, 458, 95–102.
J. Sep. Sci. 2009, 32, 3746–37543754 C. Pizarro et al.
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
Page 9
  • Source
    • "Unlike Brettanomyces/Dekkera sp., other wine associated yeasts such as S. cerevisiae, Pichia sp., Torulaspora sp., and Zygosaccharomyces sp. can produce vinylphenols but not ethylphenols under normal oenological conditions [15]. Analysis of volatile phenols in alcoholic beverages is an active research area [16] [17] [18] [19] [20] [21] [22] [23]. High performance liquid chromatography (HPLC) is a frequently used analytical technique [18] [19] [22] [24]. "
    Full-text · Dataset · Oct 2015
  • Source
    • "Unlike Brettanomyces/Dekkera sp., other wine associated yeasts such as S. cerevisiae, Pichia sp., Torulaspora sp., and Zygosaccharomyces sp. can produce vinylphenols but not ethylphenols under normal oenological conditions [15]. Analysis of volatile phenols in alcoholic beverages is an active research area1617181920212223. High performance liquid chromatography (HPLC) is a frequently used analytical technique [18,19,22,24]. "
    [Show abstract] [Hide abstract] ABSTRACT: An ethylene glycol (EG)/polydimethylsiloxane (PDMS) copolymer based stir bar sorptive extraction (SBSE)-GC-MS method was developed for the analysis of volatile phenols (4-ethylphenol, 4-vinylphenol, 4-ethylguaiacol, and 4-vinylguaiacol) in alcoholic beverages. The beverage samples were diluted with phosphate buffer (1M, pH 7) and extracted with an EG/PDMS stir bar. Volatile phenols were thermally desorbed and analyzed by GC-MS. Parameters affecting extraction efficiency were studied including ionic strength, pH, extraction time, ethanol content and nonvolatile matrix. Good correlation coefficients with R(2) in the range of 0.994-0.999 were obtained for volatile phenol concentration of 5-500μg/L. Recovery for all phenols were from 95.7% to 104.4% in a beer matrix and 81.4% to 97.6% in a wine matrix. The method had a standard deviation less than 5.8% for all volatile phenols. The limit of quantification (LOQs) in beer samples was lower than 3μg/L. The method was further applied to analyze the concentrations of volatile phenols in beer, wine and other alcoholic beverage samples. Copyright © 2015. Published by Elsevier B.V.
    Full-text · Article · Feb 2015 · Journal of Chromatography A
  • [Show abstract] [Hide abstract] ABSTRACT: An ultrasound-assisted emulsification microextraction (USAEME) based on low-density solvents was successfully applied for the extraction and pre-concentration of four toxic nitrophenols in water samples. The extracted analytes were analyzed by high-performance liquid chromatography-UV detection. The important parameters influencing the extraction efficiency were studied and optimized utilizing two different optimization methods: one variable at a time (OVAT) and central composite design (CCD). The results showed that the emulsification process can be completed in a few seconds using low-density solvents, but almost 10–20 min is necessary for high-density solvents. Under the optimum conditions (extraction solvent, 1-octanol; extraction solvent volume, 40 µL; sample pH, 3.0; salt concentration, 20% (w/v) NaCl; extraction temperature, 40 (±3)°C), limits of detection of the method were in the range of 0.25 to 1 µg L and the repeatability and reproducibility of the proposed method, expressed as relative deviation, varied in the range of 2.2–4.2% and 4.7–6.9%, respectively. Linearity was found to be in the range of 1 to 200 µg L and the preconcentration factors (PFs) were between 77 and 175. The relative recoveries of the four nitrophenols from water samples at spiking level of 10.0 µg L were in the range of 92.0 to 115.0%.
    No preview · Article · Jan 2011 · International Journal of Environmental Analytical Chemistry
Show more