Download - caracterizarea-vinurilor
-
7/25/2019 caracterizarea-vinurilor
1/12
Characterization of wines using
compositional profiles andchemometricsJavier Saurina
This review discusses strategies for characterizing wines based on compo-
sitional profiles as sources of information. Contents of low molecular organic
acids, volatile species, polyphenols, amino acids, biogenic amines and
inorganic species seem to depend on climatic, agricultural and wine-making
factors. As a result, compositional profiles of these families of natural winecomponents can be exploited as potential descriptors of wine and its quality.
Most characterization studies rely on chemometrics to facilitate extraction
of information. Cluster analysis, principal component analysis and related
methods are currently used for discrimination, classification, modeling and
correlation.
2010 Elsevier Ltd. All rights reserved.
Keywords: Chemometrics; Classification; Cluster analysis; Compositional fingerprint;
Discrimination; Instrumental assay; Modeling; Principal component analysis; Sensory
property; Wine characterization
1. Introduction
In recent years, consumers have been
increasingly interested in information on
the characteristics and the quality of food
products [1]. A wide variety of analytical
methods have been published for charac-
terizing products (e.g., wine, honey, tea,
olive oil and juices) [2]. In the case of
wines, meticulous controls are required to
assess factors (e.g., geographical origins,
grape varieties, vintages and oenological
practices) as a way of evaluating quality
and detecting fraudulent adulterations[3].
Consumer preferences in the selection ofwines take into consideration several
factors (e.g., pleasant color, taste and aroma,
ecological production, guaranteed origin
and quality). These characteristics cannot
be described in a simple manner from
given individual compounds (or various)
of the sample but they result from complex
combinations of hundreds of components.
Typical tests to estimate wine quality rely
on sensorial assays carried out by a group
of expert panelists [4]. Although these
methods are probably the best option for
checking a small number of samples, they
have limitations when dealing with large
batches of wines (e.g., cost of assays, the
inter-individual variability in perception ofsensations and the unreliability of human
senses due to fatigue, health condition or
environmental interferences).
In order to overcome the limitations of
expert-panelist assays, some research
groups have pointed out the possibility of
correlating sensory characteristics with
physic-chemical variables [5,6]. Charac-
terization studies can therefore be carried
out in a simple, sensitive and reproducible
way using suitable analytical methods,
often in combination with chemometric
treatment of data[7]. Wine properties can
be assessed from the contents of organic
constituents, the elemental composition of
metals or isotopic analysis [8]. For ex-
ample, fraudulent practices (e.g., addition
of sugars or low-quality wines during fer-
mentation) can be detected from isotopic
analysis by nuclear magnetic resonance
(NMR) and mass spectrometry (MS) [9].
These methods rely on measurements of
isotopic ratios of 13C/12C, 15N/14N,
18O/16O and 2H/1H, which depend on
geographical origin and vintage. Well-defined authentic samples are needed as
references to establish the ranges of ratios
of genuine samples and to detect those
that are anomalous or different.
Isotopic analysis seems to be complex
and expensive for routine control of food
products, so simpler, cheaper and faster
analytical methods to satisfy demands on
wine analysis are welcome. In this way,
spectroscopic, spectrometric and separa-
tion methods have been proposed as being
Javier Saurina*
Department of AnalyticalChemistry, University of
Barcelona, Diagonal 647,
08028-Barcelona, Spain
*Tel.: +34 934 039 778;
Fax: +34 934 021 233;
E-mail:[email protected]
Trends Trends in Analytical Chemistry, Vol. 29, No. 3, 2010
234 0165-9936/$ - see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.trac.2009.11.008
mailto:[email protected]://dx.doi.org/10.1016/j.trac.2009.11.008http://dx.doi.org/10.1016/j.trac.2009.11.008mailto:[email protected] -
7/25/2019 caracterizarea-vinurilor
2/12
more straightforward for wine characterization [13].
Elemental analysis is mainly applied to the study of
origins and authenticities[9]. Gas chromatography (GC)
and MS are successful in varietal characterization. There
is special interest in MS combined with headspace
devices as electronic noses (e-noses), which have great
impact in evaluating aroma properties [6]. High-performance liquid chromatography (HPLC) performs
excellently in winemaking and vintage assays, and
characterizing sensorial properties of color and taste.
The following sections describe the most significant
achievements in wine characterization using chemo-
metric analysis of physic-chemical data. The methods
discussed come from representative papers published in
the last 10 years. Section 2 comments on the principal
types of data as sources of analytical information and
briefly refers to the chemometric methods utilized.
Section 3 contains analytical applications to wine anal-
ysis (andTable 1 provides a summary).
2. Data analysis
The type of data handled in wine characterization is,
in general, multivariate in nature. Data comprise a list
or array of values, so-called first-order data in che-
mometric terminology [10,11]. As shown in Fig. 1,
concentrations of organic and inorganic species,
instrumental responses, including spectra or chro-
matographic profiles, or more complex combinations of
data from various sources can be used as analytical
data to extract relevant information. Some commentsabout these data are as follows.
2.1. Concentration of wine components
Discrete concentration values of each desired species
result in a rich, simple source of data (Fig. 1a), so the
information is principally recovered from selected com-
ponents while irrelevant species do not influence the
study. The use of concentration values obviously re-
quires prior quantification, which is often complex and
time consuming. At this stage, uncontrolled calibration
errors affecting quantification may confound character-
ization studies and may lead to misinterpretation of re-sults.
2.2. Instrumental responses
Raw or preprocessed instrumental signals comprising
spectra and chromatographic profiles can be directly
used as chemical fingerprints of the corresponding
samples (Fig. 1b). Note that the use of such signals
makes the quantification step unnecessary. Besides, in
this strategy, identification and full separation of all
components occurring in the samples is not a funda-
mental issue. Although concentrations of components
are not explicitly known they are implicit in the inten-
sities of the signal features.
This approach is gaining popularity because it is so
simple, since, as mentioned, those time-consuming steps
devoted to resolving and to quantifying the desired
analytes can be avoided [12].
Other unspecific instrumental responses (e.g., fromelectronic noses and tongues) have demonstrated great
performance for modeling and estimating sensorial
properties of wines[13].
2.3. Combined data
Data from various analytical techniques can be com-
bined together as a way of enriching the quantity and
the quality of information and arriving at more feasible
conclusions. Hence, amounts of a wide range of con-
stituents and/or data from one or various instrumental
techniques can be brought together in a convenient
manner (see Fig. 1c). As shown in Section 3, various
applications have been reported joining physic-chemicaldata (e.g., pH, conductivity and acidity), concentrations
of relevant species, and chromatographic, MS and other
instrumental responses.
The simultaneous study of properties of a series of
wines is carried out using a table of data in which each
row corresponds to a given sample (Fig. 1f). In this
arrangement, dimensions arem n,m being the number
of samples and n the number of measurements of each
sample.
2.4. Chemometric methods
Principal component analysis (PCA) has been used inmost recent exploratory studies of wines. Information
gained from PCA is often complemented with informa-
tion from other methods [e.g., groups of similar wines
can be assessed by cluster analysis (CA)]. Also, classifi-
cation of wines into pre-established categories or groups
is often performed by linear discriminant analysis (LDA)
and Soft Independent Modeling of Class Analogy
(SIMCA) methods. Artificial neural networks (ANNs)
and partial least squares (PLS) regression are sometimes
used for correlation purposes (e.g., in assessing rela-
tionships of physic-chemical variables with oenological
and sensorial attributes). PCA, CA, LDA, ANN and PLS
methods are detailed elsewhere [14].
3. Wine studies
3.1. Elemental analysis
The composition of trace and ultratrace mineral
elements of wines seems to be an excellent source of
information to be exploited in characterization studies
[15]. Grapes receive trace elements from the soil, so that
their elemental composition is reasonably correlated
with the geological characteristics of the producing
Trends in Analytical Chemistry, Vol. 29, No. 3, 2010 Trends
http://www.elsevier.com/locate/trac 235
-
7/25/2019 caracterizarea-vinurilor
3/12
areas. Mineral elements are not metabolized or trans-
formed during oenological processes and they remain
almost unaltered from musts to wines. The classification
of samples according to geographical regions or
denomination of origins and the detection of certain
adulterations may therefore rely on elemental composi-
tion. The mean contents of trace elements also dependson winemaking practices.
In the 1990s, the determination of trace elements in
wines was based mainly on flame atomic spectroscopies.
The absorption mode was, in general, sufficiently sensi-
tive to quantify transition metals (e.g., Fe, Cu, Ca, Co,
and Mn), often present at mg/L levels. Similarly, flame
emission spectroscopy was utilized for determining some
alkaline metals (e.g., Li, Na and K) and alkaline earth
metals (e.g., Ca and Sr).
More recently, the introduction of highly sensitive
atomic spectroscopies to wine analysis has expanded the
variety of elements quantified. Graphite furnace atomic
absorption spectrometry (GF-AAS), inductively coupledplasma optical emission spectroscopy (ICP-OES) and
inductively coupled plasma mass spectrometry (ICP-MS)
have been used to determine elements occurring at ng/L
levels in wines.
Apart from atomic spectroscopies, X-ray FL and elec-
trophoresis have been used to determine the elemental
composition of wines.
Table 1 summarizes recent applications.
3.2. Molecular spectroscopies
Ultraviolet visible (UV-Vis), near-infrared (NIR), mid-
infrared (MIR), fluorescence (FL) and NMR spectrosco-pies have been exploited in wine analysis (see Table 1).
UV-Vis spectra contain information regarding the
absorbing moieties, basically, polyphenolic compounds
[29]. The simplest compounds (e.g., phenolic, benzoic and
hydroxycinnamic acids) display absorption bands in the
range 300400 nm. More complex molecules (e.g., stil-
benes, flavanols, flavonols and anthocyanins) also absorb
moderate or strongly in the visible range. The composition
of these compounds is therefore the principal factor
modulating the color of wines and some taste character-
istics. Hence, significant relationships between spectral
data and oenological features can be assessed. Often, UV-
Vis spectral data arecomplemented with measurements in
the NIR range 8002500 nm. The richer source of
information resulting from these enlarged data sets has
been used in various classification studies[31].
FL of wines comes from naturally occurring fluoro-
phores, mainly polyphenols. In general, excitation
spectra have a maximum at 260270 nm and emission
spectra display characteristic features around 370 nm
and 315 nm. The shapes of excitation and emission
spectra vary slightly among wines because of differences
in the composition of native FL compounds. Table 1
includes applications based on FL.
MIR (4004000 cm1) comprises a complex spectral
range containing multiple vibration bands from func-
tional groups of major wine components, including
sugars and polysaccharides, polyphenols, tannins and
organic acids [38]. Currently, spectra obtained with a
Fourier-transform IR (FTIR) spectrophotometer operat-
ing under reflection or transmission modes are utilizedfor classification and quantification.
NMR is excellent for generating characteristic finger-
prints reflecting the complexity of compositional profiles
of wines[46,47]. The main 1H NMR signals are associ-
ated with major components including ethanol, organic
acids, sugars and amino acids. Data can be utilized to
estimate concentrations of alcohols and organic-acid
constituents [48].
From the point of view of information, UV-Vis-NIR and
FL spectra contain a limited number of spectral features,
so limiting their possibilities in discrimination. Despite
this apparent shortcoming, UV-Vis-NIR and FL tech-
niques offer advantages (e.g., simplicity, robustness,availability and minimal sample treatment).
By contrast, MIR and NMR spectroscopies provide
characteristic chemical fingerprints of great interest for
differentiation. However, drawbacks may arise (e.g.,
more tedious sample treatment procedures, more com-
plex measurements and less reproducible results).
3.3. Mass spectrometry
An emerging trend in wine analysis relies on MS for
describing complex aroma properties associated with
volatile components, which comprise hundreds of sub-
stances (e.g., alcohols, aldehydes, esters, acids, andterpenes). Some of them are already present in grapes,
while others are formed during fermentation and ageing.
Indeed, alcoholic fermentation is the vinification step
leading to formation of the greatest number of aroma
components. Further, ageing involves multiple aroma-
maturation processes (e.g., oxygenation and reactions
with molecules extracted from the oak barrels). In gen-
eral, the concentration of volatile compounds decreases
with ageing and strongly conditions the final flavor of
wines.
Contents of volatile compounds in wines vary by
several orders of magnitude, from major to trace con-
stituents. Compounds (e.g., ethanol and glycerol) occur
at concentrations of g/L. Ethyl acetate, isoamyl alcohol,
ethyl lactate, among others, are present at levels in the
range 10100 mg/L. There is a long list of components
present at concentrations of mg/L or below.
Methods reported in the literature (see Table 1) gen-
erally combine headspace systems for releasing the vol-
atile species with MS for analytical detection. This
approach is therefore recognized as a type of e-nose, in
which the mass spectra are utilized as fingerprints for
each sample with information related to aroma features.
The chemometric comparison of spectral data is then
Trends Trends in Analytical Chemistry, Vol. 29, No. 3, 2010
236 http://www.elsevier.com/locate/trac
-
7/25/2019 caracterizarea-vinurilor
4/12
Table 1. Recent methods for the characterization of wines
Samples Characterizationstudies
Instrumentaltechniques
Data analyzed Chemometricmethods
Remarks Ref.
Galician(Spanish) wines
Recognition/discrimination ofprotecteddesignations oforigin. Detection ofadulterations
Flame AESFlame AAS
Li, Rb, Na, K,Mn, Fe, Ca
PCA, CA, LDA,KNN, SIMCA
Most discriminatingelements: Li, Rb and Fe
[16,17]
Oloroso sherry Classificationaccording toprovenance
Flame AESFlame AAS
Na, K, Al, SrCa, Mg, Fe, Cu, Mn, Zn
PCA, LDA Most discriminatingelements: Na, Mg, Fe, Aland Sr. 93% predictionability
[18]
Canary Islandswines
Classificationaccording todenomination oforigins
Flame AESFlame AAS
K, Na, Li, Rb, SrCa, Mg, Fe, Cu, Zn, Mn
PCA, CA, LDA,SIMCA
Significant differences incontents of Fe and Cuamong islands.Most discriminatingelements: Rb, Na, Mnand Sr. 95% predictionability
[19]
Serbian wines Classification of redand white wines
GF-AAS Cu, Mn, Fe, Cd, Pb PCA, FA, CA Discrimination betweenred and white wines wasimpossible
[20]
Sparkling wines Discriminationbetween cava(Catalonian) andchampagne wines
GF-AASHG-AASICP-OES
Cd, Ni, PbAsAl, Ba, Ca, Cu, Fe, K,Mg, Mn, Na, P, Sr, Zn
LDA, SIMCA Remarkableauthentication power.Zero false positives andnegatives
[21]
Valencia(Spanish) wines
Recognition/discrimination ofprotecteddesignations oforigin
ICP-OES AL, Ba, Be, Ca, Cd, Ce,Co, Cr, Cu, Dy, Er, Eu,Fe, Gd, Ho, K, La, Li, Lu,Mg, Mn, Mo, Na, Nd,Ni, Pb, Pr, Sc, Se, Sm, Sr,Tb, Ti, Tm, V, Y, Yb, Zn
PCA, CA Li and Mg contents werecharacteristic of Utiel-Requena and Jumillaregions
[22]
Bohemian(CzechRepublic) wines
Identification oforigins of Bohemianwines
ICP-OES ICP-MS Na, K, Ca, Fe, MgAl, As, Ba, Ce, Co, Cr,Cs, Cu, Li, Mn, Mo, Ni,Pb, Rb, Sb, Sr, Th, U, V,Y, Zn
PCA, DA Best identificationsbased on contents of Al,Ba, Ca, Co, K, Li, Mg,Mn, Mo, Rb, Sr, V andSr/Ba, Sr/Ca, Sr/Mg ratios
[23]
Canary Islandswines
Classification intored, rose and whitewines
ICP-MS Te, Re, Be, Cd, Sn, Sb,Co, As, V, Ni, Ti, Cu, Pt,U, Cs, La, Zr, As, Zn, Rb,Tl, Tb, Ce, Pr, Nd, W,Pb, Sr, Ba, Ho, Tm, Sm,Eu, Gd, Lu, Dy, Er, Yb,Th
LDA, ANN Most discriminatingelements: Sr, Rb, Pb, Be,Ba, Tl, Ti, and Au.Prediction ability > 95%
[24]
German wines Classificationaccording togrowing areas
ICP-MS Li, Be, Ti, Co, Ni, Ga,As, Rb, Y, Zr, Nb, Mo,Cd, b, Te, Cs, La, Ce, Pr,Nd, Sm, Eu, Gd, Tb, Dy,Ho, Er, Tm, Yb, Lu, W,
Tl, U
PCA, CA Prediction ability: 88% [25]
Nebbiolo winesfrom Piedmont(Italy)
Classificationaccording toprovenance into fivegrowing areas
ICP-MS Al, As, B, Ba, Be, Ca, Cd,Ce, Co, Cr, Cs, Cu, Dy,Er, Eu, Fe, Ga, Gd, Ge,Hf, Ho, I, K, La, Li, Mg,Mn, Mo, Na, Nb, Nd,Ni, P, Pb, Pd, Pr, Rb, Rh,Sb, Si, Sm, Sn, Sr, Tb, Te,Th, Ti, Tl, Tm, U, V, W,Y, Yb, Zn Zr
PCA, LDA Most discriminatingelements:Si, Mg, Ti, Mn and Mo
[26]
(continued on next page)
Trends in Analytical Chemistry, Vol. 29, No. 3, 2010 Trends
http://www.elsevier.com/locate/trac 237
-
7/25/2019 caracterizarea-vinurilor
5/12
Table 1 (continued)
Samples Characterizationstudies
Instrumentaltechniques
Data analyzed Chemometricmethods
Remarks Ref.
Ribera Sacra(Spanish) wines
Discrimination ofRibera Sacra winesfrom other Galician
wines
CE Na, K, Ca, Mn, Li PCA, LDA,KNN, SIMCA
Good categorizationaccording to origin
[27]
CabernetSauvignonwines
Differentiation ofwines produced invarious countries
X-rayfluorescence
K, Ca, Mn, Fe, Ni, Cu,Zn, Br, Rb, Sr
PCA, CA Most discriminatingelements:K, Fe and Rb
[28]
La Mancha(Spanish) wines
Classificationaccording toproducing area
UVVisspectroscopy
Absorption spectra: 300800 nm
PCA, SIMCA 90% prediction abilityaccording to origin. 75%prediction abilityaccording to grapevariety and ageing
[30]
Tempranillowines
Geographicclassification intoSpanish andAustralian groups
Vis and NIRspectroscopy
Absorption spectra: 4002500 nm
PCA, LDA, PLS Classification ability >70%
[32]
Riesling wines Preliminaryclassification
according to country
Vis and NIRspectroscopy
Absorption spectra: 4002500 nm
PCA, LDA, PLS Classification ability >67%
[33]
Australian whitewines
Classificationaccording to grapevarieties
Vis and NIRspectroscopy
Absorption spectra: 4002500 nm
PCA, PCR, DA-PLS
Classification ability >96%
[34]
Australian redwines
Assessment ofmodels betweenwine scores andspectra
Vis and NIRspectroscopy
Absorption spectra: 4002500 nm
PLS Ability for predictingsensory quality score:r = 0.61
[35]
Loire Valley(French) grapes
Study of ripening ofFrench Cabernetgrapes. Prediction ofanthocyanincontents
Front-facefluorescence
Excitation spectra ofgrape skins: 250350 nmat kem350 nm
PCA, DA, PLS Classification accordingto ripening.Prediction of malvidin-3-O-glucoside, totalanthocyanins and totalphenols: r > 0.8
[36]
French andGerman wines
Discriminationbetween French andGerman wines
Front-facefluorescence
Emission spectra: 275450 nm at kex261 nmExcitation spectra: 250350 nm at kem350 nm
PCA, DA Classification abilities >93%
[37]
Portuguesewhite wines
Classificationaccording to varietyand winemakingprocedure.Prediction ofmannose contents
FTIR (mid)spectroscopy
FTIR spectra of the winepolysaccharide dryExtracts: 1200800 cm1
PCA, CA, PLS Identification of thewinemaking process ofmust clarification and/ormaceration. Predictionability of mannose, r >0.97
[39]
Red, white andmodel wines
Classificationaccording to tanninlevels. Prediction oftannin contents
FTIR (mid)spectroscopy
FTIR transmissionspectra:400 (or 650)4000 cm1
DA, SIMCA, PLS Classification abilities >60%Prediction ability oftannins, r > 0.96
[40,41]
Austrian redwines
Discriminationamong varieties
FTIR (mid)spectroscopy
Attenuated totalreflectance spectra: 7004000 cm1
CA, SIMCA Classification ability >95%
[42]
Australian redand white wines
Discriminationamong varieties.Fingerprintauthentication
FTIR (mid)spectroscopy
FTIR transmissionspectra:9265012 cm1
PCA, LDA Classification ability >95%
[43,44]
Gamay wines Discriminationamong origin andvintage
FTIR (mid)spectroscopy
FTIR transmissionspectra:8001800 cm1
PCA, PLS Classification ability >70% (for origin) and >50% (for vintage)
[45]
Bordeaux(French) wines
Discriminationamong growingareas, cultivars andsoil types
1H NMRspectroscopy(500 MHz)
1D 1H NMR spectra(0.78.8 ppm) of driedtissue extracts D2O usingTSP as shift marker
PCA, PLS Significant clusteringaccording to growingareas. Effect of vintagemore noticeable thaneffect of soil type
[46,47]
(continued on next page)
Trends Trends in Analytical Chemistry, Vol. 29, No. 3, 2010
238 http://www.elsevier.com/locate/trac
-
7/25/2019 caracterizarea-vinurilor
6/12
Table 1(continued)
Samples Characterizationstudies
Instrumentaltechniques
Data analyzed Chemometricmethods
Remarks Ref.
Red, roseandwhite wines
Exploratory analysis.Quantification ofwine components
from 1H NMR data
1H NMRspectroscopy(400 MHz)
1H NMR spectra (0.56 ppm) of wines in D2Ousing TSP as shift marker
PLS Prediction abilities of ethanol, glycerol, lacticacid, methanol and
malic acid: r = 0.120.96
[48]
Spanish wines Differentiation andclassificationaccording to origin,variety and ageing
Headspace-MS(e-nose)
MS by EI ionization at70 eV inm/zrange 50200
PCA, SIMCA More difficult to separateaccording to ageing
[49]
Australian whitewines
Classificationaccording to variety(Chardonnay andRiesling)
Headspace-MS(e-nose)
MS by EI ionization at70 eV in m/zrange 50180
PCA, LDA, PLS Classification abilityaccording to variety >73%
[50]
Australian redwines
Monitoring winespoilage
Headspace-MS(e-nose)
MS by EI ionization at70 eV in m/zrange 50180
PCA, LDA, PLS Successful classificationusing 4-ethylphenol asindex of spoilage
[51]
AustralianRiesling wines
Prediction of aromaproperties
Headspace-MS(e-nose)
MS by EI ionization at70 eV in m/zrange 50
180
PCA, PLS Prediction abilities ofscores of aroma notes:
estery (r = 0.75),perfume floral (r = 0.89),lemon (r = 0.82), stewedapple (r = 0.82), passionfruit (r = 0.67) and honey(r = 0.90)
[52]
Italian wines Classification ofwine varietiesaccording to aroma(volatile) fingerprints
Headspace-MS(e-nose)
MS by EI ionization at70 eV in m/zrange 50250
PCA Prediction ability >83.3%
[53]
White wines Classificationaccording to originand variety
Differentialpulsevoltammetry:PEDT-modifiedelectrodes
Voltammogram range:0.10.7 V
PCA, PLS Sample clustering withrespect to varieties andorigin
[54]
Italian wines Classificationaccording to originand variety
Headspace thin-filmmultisensoryarrayOthers
Gas sensor e-nose:current resistance of fourdifferent sensorspH, Conductivity andethanol content
PCA, ANN Prediction ability > 78% [55]
Croatian redwines
Influence ofpolyphenoliccomposition onsensory perceptions
RP HPLC-UV Major polyphenols PCA Significant correlationbetween polyphenolsand astringency,bitterness and overallsensory impression
[59]
Lacrima diMorro dAlbared wines
Classificationaccording topolyphenols.Influence on sensoryperceptions
RP HPLC-UV(DAS)RP HPLC-MS(ion trap)
Major polyphenols PCA, FA, PLS Significant correlationbetween anthocyaninconcentrations andpigmentation, sour tasteand astringency
[60]
Greek red wines Differentiationbased on variety andorigin
RP HPLC-UV(DAS)
19 polyphenols DA Sample clustering withrespect to varieties andgeographical origin
[61,62]
South Africanred and whitewines
Classificationaccording to grapevariety
RP HPLC-UV(DAS)
22 major polyphenols ANOVA, PCA,LDA
Classification ability intored and white groups >97%
[63]
Slovenian wines Classificationaccording to variety
RP HPLC-UVRP HPLC-MS
13 anthocyanins ANOVA, CA,PCA, LDA
Classification ability100%
[64]
Fino sherry Study of browning RP HPLC-UV Polyphenols PCA, CA Discrimination betweennatural and acceleratedbrowning
[65]
German redwines
Classificationaccording to variety
RP HPLC-UV 9 anthocyanins PCA, CA Sample clustering withrespect to varieties
[66]
(continued on next page)
Trends in Analytical Chemistry, Vol. 29, No. 3, 2010 Trends
http://www.elsevier.com/locate/trac 239
-
7/25/2019 caracterizarea-vinurilor
7/12
Table 1 (continued)
Samples Characterizationstudies
Instrumentaltechniques
Data analyzed Chemometricmethods
Remarks Ref.
Red, rose andwhite Spanishwines
Classificationaccording to type ofwine
RP HPLC-UV,precolumnderivatization
with Dabsyl-Cl
9 biogenic amines PCA, CA Most discriminatingvariables: putrescine,histamine and tyramine
[67]
Tokaji Aszu(Hungarian)wines
Authentication ofwine specialtiesmade frombotrytized grapes
RP-HPLC-UV(DAS)RP-HPLC-fluorimetry (kex345 nm, kem455 nm), post-columnderivatizationwith OPA
Tartaric, malic, shikimic,lactic, acetic, citric andfumaric acids13 biogenic amines
PCA, LDA Identification ability ofauthentic samples >96%
[68,69]
Spanish redwines
Classificationaccording to ageing
RP HPLC-UV(DAS),precolumnderivatizationwith NQS
8 biogenic aminesRaw chromatographicprofiles
PCA, PLS Sample clustering withrespect to ageing. Gooddifferentiation amongyoung and aged wines
[12]
Hungarian redand white wines
Classificationaccording towinemakingtechnology
Ion-exchangeHPLC-UV, post-columnderivatizationwith ninhydrin
8 biogenic amines and20 amino acids
PCA, LDA Classification ability>65% (origin), >62%(variety) and >73%(vintage)
[70]
Pinotage wines Classificationaccording to vintageand origin
SBSE-(headspace) andGC-MS
39 volatile compounds ANOVA, PCA,FA
Reasonable clusteringwith respect to originand vintage
[72]
South Africanred and whitewines
Classificationaccording to cultivar
SBSE-(headspace) andGC-MS
Major volatile and semi-volatile components
PCA, CA, DA Successful classificationof wines according tovariety
[73]
Spanish wines Differentiation ofcertified brands oforigins
Headspace-SPME and GC-FID
18 major volatilecomponents
PCA, LDA, ANN Classification ability100% with respect toorigins using ANN
[74]
Rioja (Spanish)wines
Classificationaccording towinemaking
Headspace-SPME GC-MS
11 major volatilecomponents
PCA, CA, LDA Most discriminatingvariables: 3-methyl-butylacetate, ethyloctanoate,diethylsuccinate,hexanoic acid, 2-phenylethanol anddecanoic acid.Prediction ability 100%
[75]
Andalusian(Spanish) sweetwines
Classificationaccording to grapevariety
Headspace- andGC-FID
>30 volatile components PCA, CA, LDA Prediction ability 100% [76]
Slovakian wines Classificationaccording to variety
Headspace-SPME GC-FID
65 volatile components PCA, CA, ANN,LDA, KNN
Classification ability >95%
[77]
Greekdry redwines
Prediction of sensoryscores from volatilecomponents
Headspace-SPME and GC-MS
GC fingerprint profile PCA, PLS Reasonable relationshipbetween sensoryproperties and GCprofiles
[78]
Madeira wines Characterization ofaroma profiles offive varieties
Headspace-SPME and SBSEand GC-MS
Major aromacomponents
PCA Most discriminatingvariables among dry andsweet wines: cis-oaklactone, diethylsuccinateand ethyl octanoate
[79]
CabernetSauvignonwines
Classificationaccording to origin
Microchip CE-electrochemicaldetection
Electrophoretic profiles PCA, LDA Preliminary insight toclassification
[81]
Hungarianwines
Classification intored and wine classes
Overpressured-thin layerchromatography
Polyphenols andbiogenic amines
PCA, CA Sample clustering withrespect to type of wine
[82]
(continued on next page)
Trends Trends in Analytical Chemistry, Vol. 29, No. 3, 2010
240 http://www.elsevier.com/locate/trac
-
7/25/2019 caracterizarea-vinurilor
8/12
Table 1(continued)
Samples Characterizationstudies
Instrumentaltechniques
Data analyzed Chemometricmethods
Remarks Ref.
South Africanwines
Classificationaccording to variety
LLE and GC-FIDFTIR (mid)spectroscopy
Volatile componentsFTIR transmissionspectra:
9295011 cm1
ANOVA, PCA,LDA
White wines: predictionability >98% using FTIRdata
Red wines: predictionability >86% usingFTIR + GC data
[83]
Italian wines Classificationaccording to origin
Variousanalyticaltechniques
pH, conductivity,acidity, ethanol, totalextract, redox potential,absorbance at fourwavelengths, e-tongueand e-nose profiles
CAIMAN Preliminary assessmentof a new classificationmethod
[84]
Italian wines Authentication oforigin
AAS, HPLC 35 chemical descriptors(polyphenols, metals,organic acids,alcohols, . . .)
SIMCA, UNEQ Most discriminatingvariables:Cu, Zn, anthocyans andSO2
[85]
Spanish (rose)wines
Classificationaccording to origin
Variousanalytical
techniques
19 chemical descriptors(polyphenols, metals,
classical oenologicalparameters, colorparameters)
LDA, ANN Prediction ability 100% [86]
Greek wines Classificationaccording to origin
HPLCCG-MSAAS
Polyphenols, metals,other oenologicalparameters, sensory data
PCA Satisfactory classificationaccording to origin
[87]
Gamay wines Classificationaccording to origin
FTIR (mid)spectroscopyVariousanalyticaltechniques
FTIR transmissionspectra:8001800 cm1
pH, acidity, organicacids, phenols, colorparameters
DA Higher performance of MIR spectra with respectto classical oenologicalparameters
[88]
Slovenian wines Study ofchaptalizationpractices.
Classificationaccording to origin
SNIF-NMRIRMS
D/H isotopic ratios andd13C
PCA, CA, KNN Possibility of detectingadulterations
[89]
Canary Islands(Spanish) wines
Comparison ofwines and musts
Variousanalyticaltechniques
Acidity, pH, reducingsugars, B, Ca, ashes, Cu,density, Fe, alcohol, K,Mg, Na, tartaric acid, Pb,SO2, Zn
PCA, LDA, FA Differentiation betweenwines and musts
[90]
Galician(Spanish) wines
Authentication of aGalician certifiedbrand of origin
Variousanalyticaltechniques
34 chemical variablesincluding metals,organic acids, volatilecompounds,polyphenols
PCA, CA, LDA,KNN, SIMCA
Most discriminatingvariables:Fe, Li, Rb, delphinidinand epicatechin
[91]
Slovenian andItalian wines
Classificationaccording to origin
Ion exclusion-HPLCICP-OES1H-NMR
Metals, organic acids,amino acids
PCA, CA Higher performance of NMR spectra comparedwith other data
[92]
Rioja (Spanish)wines
Classificationaccording to winetype (claret, roseandblend)
UV-VisspectroscopyColorimetry
Color parameters ANN, SIMCA Classification ability >50%
[93]
AAS, Atomic absorption spectroscopy; AES, Atomic emission spectroscopy; GF, Graphite furnace; HG, Hydride generation; ICP, Inductively coupled plasma;OES, Optical emission spectroscopy; MS, Mass spectrometry; UV, Ultraviolet; Vis, Visible; NIR, Near infrared; FTIR, Fourier transform infrared; HPLC, Highperformance liquid chromatography; GC, Gas chromatography; CE, Capillary electrophoresis; EI, Electronic impact; DAS, Diode array spectrophotometer; kex,Excitation wavelength; kem, Emission wavelength; FID, Flame-ionization detector; SBSE, Stir-bar sorptive extraction; SPME, Solid-phase microextraction; LLE,Liquid-liquid extraction; SNIF, Site-specific natural-isotope fractionation; NMR, Nuclear magnetic resonance; IRMS, Isotope-ratio mass spectrometry; TSP,(Trimethyl)-propionic-2,2,3,3-d4 acid; PEDT, Poly(3,4-ethylenedioxythiophene); Dabsyl-Cl, (4-Dimethylamino-azobenzene-4-sulphonyl chloride); OPA, o-Phthaldialdehyde; NQS, 1,2-Naphthoquinone-4-sulfonate; PCA, Principal component analysis; CA, Cluster analysis; DA, Discriminant analysis; LDA, Lineardiscriminant analysis; KNN,K-Nearest neighbors; SIMCA, Soft independent modeling of class analogy; ANN, Artificial neural networks; FA, Factor analysis;PLS, Partial least square (regression); CAIMAN, Classification and influence matrix analysis; UNEQ, Unequal class modeling; r, Coefficient of correlation.
Trends in Analytical Chemistry, Vol. 29, No. 3, 2010 Trends
http://www.elsevier.com/locate/trac 241
-
7/25/2019 caracterizarea-vinurilor
9/12
exploited to differentiate and to classify wines according
to variety and other oenological features.
3.4. Electrochemical techniques
To date, electroanalytical data have had little signifi-
cance in characterizing food products. Table 1contains
some preliminary studies for evaluating the possibilities
of these techniques in wine analysis.
3.5. Separation techniques
Separation techniques, especially HPLC and GC, have
been widely used for wine characterization and classifi-
cation. Data of different degrees of complexity (e.g.,
analyte concentrations, peak areas and chromatograms)
can be generated and processed.
In the case of concentrations, the range of analytes
considered is limited to identifiable and quantifiable
components. For peak-area data, identification and
quantification can be omitted. Hence, areas of charac-
teristic peaks at well-defined retention times are taken for
analysis.
In the case of chromatographic profiles, the whole run
or selected time windows can be used as analytical data
[12].
Fig. 2shows an example to illustrate these possibilities
with a chromatogram of the separation of polyphenols
in a red wine by reversed-phase HPLC with UV detection.
As in this case, the great complexity of the wine sample
hinders full separation of all components of interest and
multiple unknown peaks may eventually appear in the
chromatograms. In these circumstances, the separation
of all components and the identification of all chro-
matographic features may not be fully achieved. How-
ever, the occurrence of unresolved peaks should not be
250 300
0
10
20
30
40
50
250 300
mAU
0
10
20
30
40
50
Spectrum
Wine 1
Wine 2
Wine 3
Wine m
Responses 1 to n
Wines
1tom
D(m x n)
Responses 1 to n
Analyte
1
Analyte
2
Analyte
3
Concentrations
(a)
(b)
(d)
(c) Concentrations Spectrum Chromatogram Others
...
Figure 1. Types of data of interest in wine characterization. (a) Array from discrete concentration (or physico-chemical) values from the analysisof a wine. (b) Instrumental (spectral) data profile. (c) Combined array from concentrations and instrumental data. (d) Scheme for constructing adata matrix to be used in the study of a series of wines.
Trends Trends in Analytical Chemistry, Vol. 29, No. 3, 2010
242 http://www.elsevier.com/locate/trac
-
7/25/2019 caracterizarea-vinurilor
10/12
taken as a negative issue, since chromatograms may be
exploited as chemical fingerprints that provide specific
information on the samples analyzed.
The two principal families of wine components inves-
tigated by HPLC comprise polyphenols and amino com-
pounds (biogenic amines and amino acids).
Compositional profiles regarding polyphenolic and/or
amino species have been correlated with significant
factors (e.g., organoleptic properties, winemaking prac-
tices and grape varieties) [56,57]. However, extracting
information on the influence of oenological factors may
be difficult due to the multifactorial nature of the pro-
cesses involved [58]. Here, again, application of
chemometrics may assist in reaching reliable conclu-
sions.Table 1 provides some representative examples of
relevant HPLC applications.
GC is typically utilized for the analysis of the volatile
fraction of wines. Headspace and solid-phase microex-
traction (SPME) have been extensively used for wine
pretreatment in order to recover volatile components. In
the headspace technique, species are volatilized from the
sample under controlled experimental conditions (either
in static or dynamic mode) and the gas phase is directly
transferred to the chromatograph using an appropriate
interface[71]. Alternatively, gas-phase components can
be adsorbed on a fiber to be further introduced into the
injection port of the GC instrument to run the separa-
tion. In SPME mode, the fiber is immersed in the wine
phase to retain the desired components. In all cases,
analytes are further released from the fibers by thermal
desorption. The chromatographic detection is often
accomplished by MS as a way of enhancing the sensi-
tivity and the selectivity of data. As mentioned in
Table 1, compositional profiles of volatile compounds
have been exploited to extract information on chemical
and sensory features (e.g., taste and aroma).
To date, CE has been of limited significance in wine
analysis, and only a few methods have been proposed for
quantifying wine components[80]. In one example, the
performance of a CE-microchip device with electrochemical
0 5 10 15 20 25
0
100
200
300
400
gallicacid
3,4-dihydroxybenzoicacid
Absorbance(mAU)
Time (min)
catechine
4-hydroxybenzoicacid
vanillicacid
caffeicacid
syringicacid
coumaricacid
ferulicacid
t-resveratrol
Peak
areas
Concentrations
Chromatographic
profile
(Absorbancevstime)
a1
a2
a4 a
5...a
3
c1
c2 c
3 c
5
calibrations
...Figure 2. Possibilities of generating data of different complexity: chromatographic profile (absorbance vs. time), peak areas and concentrations.
Trends in Analytical Chemistry, Vol. 29, No. 3, 2010 Trends
http://www.elsevier.com/locate/trac 243
-
7/25/2019 caracterizarea-vinurilor
11/12
detection was assessed for preliminary classification of
wines[81].
3.6. Combined data sets
The combination of data from various analytical sources
is a powerful strategy to enrich the information content,
thus improving characterization results. For example,physico-chemical parameters, analyte concentrations,
spectra, and chromatograms can be analyzed together
from an extended data set. Note that, in these cases,
differences in the nature and the magnitude of the data
handled have to be compensated for. Typically, nor-
malization or standardization pretreatments are required
to provide similar weight to all variables. Table 1com-
ments on some relevant applications based on combined
data sets.
4. Conclusions
Characterization of wines based on analytical methods
combined with chemometric treatment of data
provides excellent robustness and efficiency, and re-
duces costs compared with sensory tests by expert
panelists. Physico-chemical parameters, concentrations
of wine components and instrumental signals can be
used as multivariate data. Wine features, including
origin, variety and winemaking practices, can be
evaluated from this source of information.
Because of the multiparametric nature of wines,
chemometics makes interpretation of data more feasible.
Contents of mineral elements result in appropriate
descriptors of origin and authenticity. The volatile frac-
tion of wines is useful for categorizing wines based on
aroma characteristics. Polyphenolic fingerprints can be
exploited in analyzing varieties and evaluating color and
taste. Finally, amino compounds are suitable markers for
studying winemaking practices and ageing of wines.
Acknowledgement
This work was supported by the Spanish Ministerio de
Ciencia y Tecnolog a, project CTQ2008-04776/BQU.
References[1] D.M.A.M. Luykx, S.M. Van Ruth, Food Chem. 107 (2008) 897.[2] L.M. Reid, C.P. ODonnell, G. Downey, Trends Food Sci. Technol.
17 (2006) 344.
[3] I.S. Arvanitoyannis, M.N. Katsota, E.P. Psarra, E.H. Soufleros, S.
Kallithraka, Trends Food Sci. Technol. 10 (1999) 321.
[4] N. Cayot, Food Chem. 101 (2007) 154.
[5] S. Lopez-Feria, S. Cardenas, M. Valcarcel, Trends Anal. Chem. 27
(2008) 794.
[6] B. Plutowska, W. Wardencki, Food Chem. 101 (2007) 845.
[7] M. Kozak, C.H. Scaman, J. Sci. Food Agric. 88 (2008) 1115.
[8] M. Gishen, R.G. Dambergs, D. Cozzolino, Aust. J. Grape Wine Res.
11 (2005) 296.
[9] N. Ogrinc, I.J. Kosir, J.E. Spangenberg, J. Kidric, Anal. Bioanal.
Chem. 376 (2003) 424.
[10] S. Sentellas, J. Saurina, S. Hernandez-Cassou, M.T. Galceran, L.
Puignou, J. Chromatogr., A 909 (2001) 259.
[11] S. Sentellas, J. Saurina, J. Sep. Sci. 26 (2003) 1395.
[12] N. Garca-Villar, S. Hernandez-Cassou, J. Saurina, J. Agric. Food
Chem. 55 (2007) 7453.
[13] M. Scampicchio, D. Ballabio, A. Arecchi, S.M. Cosio, S. Mannino,
Microchim. Acta 163 (2008) 11.
[14] S.D. Brown, R. Tauler, B. Walczak, Comprehensive Chemometrics,
Chemical and Biochemical Data Analysis, Vol. 3, Elsevier,
Amsterdam, The Netherlands, 2009.
[15] M. Suhaj, M. Korenovska, Acta Aliment. 34 (2005) 393.
[16] M.J. Latorre, C. Garca Jares, B. Medina, C. Herrero, J. Agric. Food
Chem. 42 (1994) 1451.
[17] R.M. Pena, M.J. Latorre, S. Garca, A.M. Botana, C. Herrero, J. Sci.
Food Anal. 79 (1999) 2052.
[18] P. Paneque, M.T. Alvarez-Sotomayor, I.A. Gomeza, Food Chem.
117 (2009) 302.
[19] S. Fras, J.E. Conde, J.J. Rodrguez-Bencomo, F. Garca-Montelon-
go, J.P. Perez-Trujillo, Talanta 59 (2003) 33 5.
[20] S. Razic, D. Cokesa, S. Sremac, J. Serb. Chem. Soc. 72 (2007)
1487.
[21] A. Jos, I. Moreno, A.G. Gonzalez, G. Repetto, A.M. Camean,
Talanta 63 (2004) 377.
[22] A. Gonzalvez, A. Llorens, M.L. Cervera, S. Armenta, M. de la
Guardia, Food Chem. 112 (2009) 26.
[23] J. Sperkova, M. Suchanek, Food Chem. 93 (2005) 659.
[24] J.P. Perez-Trujillo, M . Barbaste, B. Medina, Anal. Lett. 36 (2003)
679.
[25] G. Thiel, G. Geisler, I. Blechschmidt, K. Danzer, Anal. Bioanal.
Chem. 378 (2004) 1630.
[26] E. Marengo, M. Aceto, Food Chem. 81 (2003) 621.
[27] M. Nunez, R.M. Pena, C. Herrero, S. Garca-Martn, Analusis 28
(2000) 432.
[28] S.J. Haswell, A.D. Walmsley, J. Anal. Atom. Spectrom. 13 (1998) 131.
[29] V. Cheynier, Am. J. Clin. Nutr. 81 (2005) 223S.
[30] M. Urbano, M.D. Luque de Castro, P.M. Perez, J. Garca-Olmo,
M.A. Gomez-Nieto, Food Chem. 97 (2006) 166.
[31] D. Cozzolino, R.G. Dambergs, L. Janik, W.U. Cynkar, M. Gishen, J.
Near Infrared Spectrosc. 14 (2006) 279.
[32] L. Liu, D. Cozzolino, W.U. Cynkar, M. Gishen, C.B. Colby, J. Agric.
Food Chem. 54 (2006) 6754.
[33] L. Liu, D. Cozzolino, W.U. Cynkar, R.G. Dambergs, L. Janik,
B.K. ONeill, C.B. Colb, M. Gishen, Food Chem. 106 (2008) 781.
[34] D. Cozzolino, H.E. Smyth, M. Gishen, J. Agric. Food Chem. 51
(2003) 7703.
[35] D. Cozzolino, G. Cowey, K.A. Lattey, P. Godden, W.U. Cynkar, R.G.
Dambergs, L. Janik, M. Gishen, Anal. Bioanal. Chem. 391 (2008)
975.
[36] M. Le Moigne, E. Dufour, D. Bertrand, C. Maury, D. Seraphin, F.
Jourjon, Anal. Chim. Acta 621 (2008) 8.
[37] E. Dufour, A. Letort, A. Laguet, A. Lebecque, J.N. Serra, Anal.
Chim. Acta 563 (2006) 292.
[38] X.Y. Niu, Y.B. Ying, H.Y. Yu, L.J. Xie, X.P. Fu, Spectrosc. Spec.Anal. 28 (2008) 804.
[39] M.A. Coimbra, F. Goncalves, A.S. Barros, I. Delgadillo, J. Agric.
FoodChem.50 (2002) 3405.
[40] K. Fernandez, E. Agosin, J. Agric. Food Chem. 55 (2007)
7294.
[41] K. Fernandez, X. Labarca, E. Bordeu, A. Guesalaga, E. Agosin,
Appl. Spectrosc. 61 (2007) 1163.
[42] A. Edelmann, J. Diewok, K.C. Schuster, B. Lendl, J. Agric. Food
Chem. 49 (2001) 1139.
[43] C.J. Bevin, A.J. Fergusson, W.B. Perry, L.J. Janik, D. Cozzolino, J.
Agric. Food Chem. 54 (2006) 9713.
[44] C.J. Bevin, R.G. Dambergs, A.J. Fergusson, D. Cozzolino, Anal.
Chim. Acta 621 (2008) 19.
Trends Trends in Analytical Chemistry, Vol. 29, No. 3, 2010
244 http://www.elsevier.com/locate/trac
-
7/25/2019 caracterizarea-vinurilor
12/12
[45] D. Picque, T. Cattenoz, G. Corrieu, J.L. Berger, Sci. Aliments 25
(2005) 207.
[46] G.E. Pereira, J.P. Gaudillere, C. Van Leeuwen, G. Hilbert, O.
Lavialle, M. Maucourt, C. Deborde, A. Moing, D. Rolin, J. Agric.
Food Chem. 53 (2005) 6382.
[47] G.E. Pereira, J.P. Gaudillere, C. Van Leeuwen, G. Hilbert, M.
Maucourt, C. Deborde, A. Moing, D. Rolin, J. Int. Sci. Vigne Vin 41
(2007) 103.
[48] F.H. Larsen, F. van den Berg, S.B. Engelsen, J. Chemom. 20
(2006) 198.
[49] M.P. Mart, O. Busto, J. Guasch, J. Chromatogr., A 1057 (2004)
211.
[50] D. Cozzolino, H.E. Smyth, W. Cynkar, R.G. Dambergs, M. Gishen,
Talanta 68 (2005) 382.
[51] W. Cynkar, D. Cozzolino, B. Dambergs, L. Janik, M. Gishen, Sens.
Actuators B 124 (2007) 167.
[52] D. Cozzolino, H.E. Smyth, W. Cynkar, L. Janik, R.G. Dambergs, M.
Gishen, Anal. Chim. Acta 621 (2008) 2.
[53] C. Armanino, M.C. Casolino, M. Casale, M. Forina, Anal. Chim.
Acta 614 (2008) 134.
[54] L. Pigani, G. Foca, K. Ionescu, V. Martina, A. Ulrici, F. Terzi, M.
Vignali, C. Zanardi, R. Seeber, Anal. Chim. Acta 614 (2008) 213.
[55] M. Penza, G. Cassano, Food Chem. 86 (2004) 283.
[56] D.P. Makris, S. Kallithraka, P. Kefalas, J. Food Comp. Anal. 19
(2006) 396.
[57] C. Ancin-Azpilicueta, A. Gonzalez-Marco, N. Jimenez-Moreno,
Crit. Rev. Food Sci. Nutr. 48 (2008) 257.
[58] J. Saurina, Characterization of wines through the compositional
profiles of biogenic amines and related compounds, focusing on
the description of toxicological and organoleptic features, in: P.
OByrne (Editor), Red Wine and Health, Nova Science Publishers,
Inc., NY, USA, 2009, pp. 630.
[59] I. Budic-Leto, G. Zdunic, J.G. Kljusuric, I. Pezo, I. Alpeza, T. Lovric,
J. Food Agric. Environ. 6 (2008) 138.
[60] E. Boselli, A. Giomo, M. Minardi, N.G. Frega, Eur. Food Res.
Technol. 227 (2008) 709.
[61] D.P. Makris, S. Kallithraka, A. Mamalos, Talanta 70 (2006)
1143.
[62] S. Kallithraka, A. Mamalos, D.P. Makris, J. Agric. Food Chem. 55
(2007) 3233.
[63] A. de Villiers, P. Majek, F. Lynen, A. Crouch, H. Lauer, P. Sandra,
Eur. Food Res. Technol. 221 (2005) 520.
[64] L.J. Kosir, B. Lapornik, S. Andrensek, A.G. Wondra, U. Vrhovsek,
J. Kidric, Anal. Chim. Acta 513 (2004) 277.
[65] M. Palma, C.G. Barroso, J.A. Perez-Bustamante, Analyst (Cam-
bridge, UK) 125 (2000) 1151.
[66] B. Berente, D.D.L. Garca, M. Reichenbacher, K. Danzer, J.
Chromatogr., A 871 (2000) 95.
[67] R. Romero, M. Sanchez-Vinas, D. Gazquez, M.G. Bagur, J. Agric.
Food Chem. 50 (2002) 4713.
[68] J. Kiss, A. Sass-Kiss, J. Agric. Food Chem. 53 (2005) 10042.
[69] A. Sass-Kiss, J. Kiss, B. Havadi, N. Adanyi, Food Chem. 110
(2008) 742.
[70] K. Heberger, E. Csomos, L. Simon-Sarkadi, J. Agric. Food Chem.
51 (2003) 8055.
[71] R. Castro, R. Natera, E. Duran, Garca-Barroso, Eur. Food Res.
Technol. 228 (2008) 1.
[72] B.T. Weldegergis, A.M. Crouch, J. Agric. Food Chem. 56 (2008)
10225.
[73] A. Tredoux, A. de Villiers, P. Majek, F. Lynen, A. Crouch, P.
Sandra, J. Agric. Food Chem. 56 (2008) 4286.
[74] J.M. Jurado, O. Ballesteros, A. Alcazar, F. Pablos, M.J. Martn, J.L.
Vlchez, A. Navalon, Anal. Bioanal. C hem. 390 (2008) 961.
[75] S. Cabredo-Pinillos, T. Cedron-Fernandez, C. Saenz-Barrio, Eur.
Food Res. Technol. 226 (2008) 1317.
[76] R. Marquez, R. Castro, R. Natera, C. Garca-Barroso, Eur. Food
Res. Technol. 226 (2008) 1479.
[77] D. Kruzlicova, J. Mocak, J. Hrivnak, J. Food Nutr. Res. 47 (2008)
37.
[78] E. Koussissi, A. Paterson, J.R. Piggott, Eur. Food Res. Technol. 225
(2007) 749.
[79] R.F. Alves, A.M.D. Nascimento, J.M.F. Nogueira, Anal. Chim. Acta
546 (2005) 11.
[80] N. Garca-Villar, J. Saurina, S. Hernandez-Cassou, Electrophoresis
27 (2006) 474.
[81] M. Scampicchio, S. Mannino, J. Zima, J. Wang, Electroanalysis
(NY) 17 (2005) 1215.
[82] E. Csomos, K. Heberger, L. Simon-Sarkadi, J. Agric. Food Chem.
50 (2002) 3768.
[83] L. Louw, K. Roux, A. Tredoux, O. Tomic, T. Naes, H.H.
Nieuwoudt, P. van Rensburg, J. Agric. Food Chem. 57 (2009)
2623.
[84] D. Ballabio, A. Mauri, R. Todeschini, S. Buratti, Anal. Chim. Acta
570 (2006) 249.
[85] F. Marini, R. Bucci, A.L. Magri, A.D. Magri, Chemom. Intell. Lab.
Syst. 84 (2006) 164.
[86] S. Perez-Magarino, M. Ortega-Heras, M.L. Gonzalez-San Jose, Z.
Boger, Talanta 62 (2004) 983.
[87] S. Kallithraka, I.S. Arvanitoyannis, P. Kefalas, A. El-Zajouli, E.
Soufleros, E. Psarra, Food Chem. 73 (2001) 501.
[88] D. Picque, T. Cattenoz, G. Corrieu, J. Int. Sci. Vigne Vin 35 (2001)
165.
[89] I.J. Kosir, M. Kocjancic, N. Ogrinc, J. Kidric, Anal. Chim. Acta 429
(2001) 195.
[90] G. Gonzalez, E.M. Pena-Mendez, Eur. Food Res. Technol. 212
(2000) 100.
[91] S. Rebolo, R.M. Pena, M.J. Latorre, S. Garca, A.M. Botana, C.
Herrero, Anal. Chim. Acta 417 (2000) 211.
[92] M.A. Brescia, I.J. Kosir, V. Caldarola, J. Kidric, A. Sacco, J. Agric.
Food Chem. 51 (2003) 21.
[93] M.E. Melendez, M.S. Sanchez, M. Inguez, L.A. Sarabia, M.C. Ortiz,
Anal. Chim. Acta 446 (2001) 159.
Trends in Analytical Chemistry, Vol. 29, No. 3, 2010 Trends
http://www.elsevier.com/locate/trac 245