Content uploaded by Fernando Cebola Lidon
Author content
All content in this area was uploaded by Fernando Cebola Lidon on May 22, 2015
Content may be subject to copyright.
This article was downloaded by: [b-on: Biblioteca do conhecimento online UNL]
On: 20 January 2014, At: 12:19
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
International Journal of Food Properties
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/ljfp20
Identification of Chemical Clusters
Discriminators of Arabica and Robusta
Green Coffee
Natalina Cavaco Bicho a b , António Eduardo Leitão b , José Cochicho
Ramalho b , Nuno Bartolomeu de Alvarenga c & Fernando Cebola
Lidon a
a Departamento de Ciências e Tecnologia da Biomassa, Faculdade
de Ciências e Tecnologia , Universidade Nova de Lisboa , Caparica ,
Portugal
b Centro de Ecofisiologia, Bioquímica e Biotecnologia Vegetal,
Instituto de Investigação Científica Tropical, I.P., Quinta do
Marquês , Oeiras , Portugal
c Departamento de Tecnologias e Ciências Aplicadas , Escola
Superior Agrária, Instituto Politécnico de Beja , Beja , Portugal
Accepted author version posted online: 09 Aug 2012.Published
online: 27 Feb 2013.
To cite this article: Natalina Cavaco Bicho , António Eduardo Leitão , José Cochicho Ramalho ,
Nuno Bartolomeu de Alvarenga & Fernando Cebola Lidon (2013) Identification of Chemical Clusters
Discriminators of Arabica and Robusta Green Coffee, International Journal of Food Properties, 16:4,
895-904, DOI: 10.1080/10942912.2011.573114
To link to this article: http://dx.doi.org/10.1080/10942912.2011.573114
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
International Journal of Food Properties, 16:895–904, 2013
Copyright © Taylor & Francis Group, LLC
ISSN: 1094-2912 print / 1532-2386 online
DOI: 10.1080/10942912.2011.573114
IDENTIFICATION OF CHEMICAL CLUSTERS
DISCRIMINATORS OF ARABICA AND ROBUSTA
GREEN COFFEE
Natalina Cavaco Bicho1,2, António Eduardo Leitão2,José
Cochicho Ramalho2, Nuno Bartolomeu de Alvarenga3,
and Fernando Cebola Lidon1
1Departamento de Ciências e Tecnologia da Biomassa, Faculdade de Ciências e
Tecnologia, Universidade Nova de Lisboa, Caparica, Portugal
2Centro de Ecofisiologia, Bioquímica e Biotecnologia Vegetal, Instituto de
Investigação Científica Tropical, I.P., Quinta do Marquês, Oeiras, Portugal
3Departamento de Tecnologias e Ciências Aplicadas, Escola Superior Agrária,
Instituto Politécnico de Beja, Beja, Portugal
The chemical parameters pH, soluble solids, caffeine, trigonelline, total chlorogenic
acids, total caffeoylquinic acids, 3-O-caffeoylquinic acid, 4-O-caffeoylquinic acid, 5-
O-caffeoylquinic acid, total dicaffeoylquinic acids, 3,4-O-dicaffeoylquinic acid, 3,5-
O-dicaffeoylquinic acid, 4,5-O-dicaffeoylquinic acid, total feruloylquinic acids, 3-O-
feruloylquinic acid, and 5-O-feruloylquinic acid were measured in Arabica (C. arabica)
and Robusta (C. canephora) green coffees in order to determine discrimination parameters.
In general, Robusta green coffee showed higher values for pH, soluble solids, caffeine, total
caffeoylquinic acids, total dicaffeoylquinic acid, and total feruloylquinic acid, but the con-
tent of soluble solids was not significantly different in both species of green coffee. Through
application of a multivariate analysis, it was concluded that these chemicals form three clus-
ters, being the group of caffeine, trigonelline, 3-O-caffeoylquinic acid, 4-O-caffeoylquinic
acid, 3-O-feruloylquinic acid, 5-O-feruloylquinic acid, 3,4-O-dicaffeoylquinic acid, 3,5-O-
dicaffeoylquinic acid, and 4,5-O-dicaffeoylquinic acid highly discriminating for Arabica and
Robusta green coffees.
Keywords: Arabica green coffee, Caffeine, Chlorogenic acids, Robusta green coffee,
Trigonelline.
INTRODUCTION
The cell wall of the coffee bean epidermis is surrounded by crystallized waxes,
whereas chlorogenic acid, terpenes, and derivatives of 5 hydroxytryptamine (serotonin)
prevail in the cuticular layer. The chlorogenic acids further accumulate in the cytoplasm of
epidermal and parenchyma cells, but larger quantities can be found in the periplasm. In the
cytoplasm of parenchyma cells, caffeine is also associated with chlorogenic acid (forming
Received 11 October 2010; accepted 15 March 2011.
Address correspondence to Fernando Cebola Lidon, Departamento Ciências e Tecnologia da Biomassa,
Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Campus da Caparica, Caparica 2829-516,
Portugal. E-mail: fjl@fct.unl.pt
895
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
896 BICHO ET AL.
a potassium chlorogenate complex) and metallic salts further occur in the form of calcium
ions of phosphate and potassium.[1,2]
In Arabica (Coffea arabica L.) and Robusta (Coffea canephora Pierre ex A.Froehn)
green coffees, pH usually is ca. 5.2–6.1,[3] while soluble solids correspond to 24–27 and
26–31 g/100 g coffee, respectively.[4] The caffeine levels of green coffee beans might vary
between 0.8–2.5 g/100 g coffee. Nevertheless, although some coffee species may present
low levels of caffeine, in Arabica and C. Canephora (for both cvs. Robusta and Conilon),
they are usually between 0.8–1.5 and 1.6–2.2 g/100 g coffee, respectively. Trigonelline
contents correspond to about 0.6 g/100 g coffee, but almost 50% of this compound is
degraded during roasting,[5] with the formation of other compounds, namely nicotinic
acid, pyridine, 3-methyl-pyridine, and methyl ester of nicotinic acid. Caffeoylquinic acids
(CQAs), dicaffeoylquinic acids (diCQA), and feruloylquinic acids (FQA) account for about
98% of total chlorogenic acids (CGA) in green coffee.[6] The hydroxycinnamic acids are
also particularly common in coffee beans, especially as chlorogenic acid (4–8 g/100 g
coffee) in the double form of caffeine and potassium chlorogenate,[7,8] but they are
significantly destroyed during roasting, with the release of the correspondent alkaloid.[9,10]
It has long been known that chemical composition of green coffee beans depends
on the genotypes and geographic area of origin, as well as of cultural practices, matura-
tion, and post-harvest conditions, particularly storage.[11] Considering these parameters,
this work aims to identify chemical discriminators that might be applied to differentiate the
majority of commercial Arabica and Robusta green coffee beans. Accordingly, the chem-
ical composition of Arabica and Robusta green coffees from Brazil and India was carried
out, being a multivariate analysis applied to identify chemical clusters that might be useful
as discriminators of these green coffee types.
MATERIALS AND METHODS
Sampling of Coffea arabica L. (from Brazil) and Coffea canephora Pierre ex
A.Froehn (from India) was carried out according to the Instrução Normativa No. 8,[12] NP
1666,[13] and ISO 4072,[14] as recommended by the ICO for sampling green coffee in bags.
The sampling process began with a randomized (a minimum of 10% of the lot) collection
of green coffee bags.[15] The selected bags were separated from the lot and each one was
collected in triplicate (with a probe of 30 ±6 g of green coffee beans) from three differ-
ent points in the bag (top, middle, and bottom). After extraction and homogenization, the
portions were joined, for an overall take of green coffee, with a minimum mass of 1.5 kg.
Soluble solids and pH were measured according to the AOAC,[16,17] at 25◦C, after cal-
ibration of the electrode with pH 4.0 and 7.0 buffer solutions. Ground green coffee (10 g ±
0.1 mg) mixed with water (200 mL) was boiled for 5 min, cooled at room temperature, and
the weight adjusted by adding water. After filtration with a Whatman No. 1 filter, the pH of
the filtrate was measured at room temperature. For quantification of soluble solids, 25 mL
of the filtrate were dried in a water bath until dryness, after which the residue was placed
in an oven at 105◦C, cooled in a desiccator, and weighed. Data is the average of triplicate
for each sample of green coffee.
Caffeine and trigonelline contents were measured according to ISO 10095.[18]
Samples of ground green coffee (1 g ±0.1 mg) were homogenized with magnesium oxide
(4.5 ±0.5 g) and water (100 mL) and placed in a water bath at 90◦C with continuous
stirring for 20 min. After cooling, the weight was restored to the original level and the
mixture filtered through a Whatman No. 1 filter, without washing the residue. An aliquot
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
CHEMICAL DISCRIMINATORS OF GREEN COFFEE 897
of 2 mL of the filtered mixture was diluted with distilled water to a volume of 10 mL and
filtered through a 0.45 µm filter. Thereafter, caffeine quantification was carried out in an
integrated HPLC system (Waters, Milford, MA, USA; equipped with an UV-VIS detector,
model 440, column Lichrosorb 100 RP-18, Merck, Darmstadt, Germany; 5 µm particle
size, 4 ×250 mm), being used with 32 Karat Software (V. 7.0, Beckman Coulter, Inc.,
Brea, CA, USA). The elution of a 20 µL injection was performed at 25◦C, with a 1 mL
min−1flow rate, using phosphate buffer (20 mM; pH 4.3) and acetonitrile (9:1), with detec-
tion at 254 nm. Identification and quantification was performed using standard curves, with
concentrations between ca. 8–1000 µgmL
−1for caffeine, and between ca. 8–500 µgmL
−1
for trigonelline. Data were within the detection limits of the method. All extractions and
chromatographic analysis were performed in triplicate.
Chlorogenic acids extraction and analysis followed Correia[9] and Fortunato et al.[19]
After mixing 2 g ±0.1 mg of ground green coffee to 10 mL of methanol:water (40:60),
and mechanical agitation for 30 min, the mixture was centrifuged (9400 g, 5 min, 25◦C)
and the supernatant decanted. Thereafter, 1 mL of Carrez solutions I (aqueous solution of
Zn acetate dihydrate and glacial acetic acid, 10.95 g and 1.5 mL, respectively, to a final
volume of 50 mL) and II (aqueous solution of 5.3 g of potassium hexacyanoferrate II trihy-
drate in a final volume of 50 mL) were added for clarification of the sample, after which a
methanol:water (40:60) solution was added to a final volume of 100 mL. After resting for
15 min, the mixture was filtered through a Whatman No. 1 filter and an aliquot of 10 mL was
removed from the filtrate and filtered through 0.45 µm. For quantification, an HPLC system
(Beckman Coulter System Gold, Beckman Coulter, Inc.) equipped with a diode array detec-
tor (model 168), a reverse phase column (Spherisorb S5 ODS-2, Waters; 4.6 ×250 mm) and
32 Karat Software (V. 7.0, Beckman Coulter, Inc.) was used. The elution of a 20 µL injec-
tion was performed at ca. 25◦C,over45min,witha1mLmin
−1flow rate, using an opti-
mized linear gradient 20–70% of B (solvent A tripotassium citrate buffer solution 10 mM,
pH 2.5, and solvent B methanol 100%). Detection was performed at 325 and 330 nm.
For isomerisation of chlorogenic acid (caffeoylquinic acids), 200 mg of
5-caffeoylquinic acid was diluted in 20 mL of distilled water and the pH adjusted to 8 with
ammonium hydroxide (4 M). Then, the solution was boiled for 30 min in a water bath,
cooled, and the pH adjusted to 2.5 with HCl (4 M). After that filtration samples were used
for quantification. The identification of chromatographic peaks and quantification of results
was carried out using standard solutions of 5-CQA. To identify the isomers 3-CQA and 4-
CQA, the standard 5-CQA isomer was subjected to isomerization, as described. The peaks
appeared with the following sequence: 3-CQA, 3-FQA, 4-CQA, 5-CQA, 4-FQA, 5-FQA,
3,4-diCQA, 3,5-diCQA, and 4,5-diCQA. The calibration curve was obtained from 5-CQA
with readings at 325 and 330 nm. The quantification was done assuming the peak areas as
a reference and comparing them with the standard 5 CQA. To quantify each compound, the
following equation was used:[9,20]
c=Fr ×ε1×Mr2×A
ε2×Mr1
,
where c is the concentration of the isomer to quantify, in mg L−1; Fr is the response factor
of the standard 5-CQA in mg L−1per unit area; ε1is the molar absorption coefficient of the
standard 5-CQA in L mol−1cm−1;ε2is the molar absorption coefficient of the isomer to
quantify, in L mol−1cm−1;Mr
2is the molecular weight on the isomer under study—CQA
=354.31 g mol−1,FQA=368.28 g mol−1,diCQA=516.44 g mol−1;Mr
1is the molar
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
898 BICHO ET AL.
mass of acid 5-CQA; A is the peak area of the isomer to be quantified. The molar absorption
coefficients (3-CQA =18,400, 4-CQA =18,000, 5-CQA =19,500, 3,4-diCQA =31,800,
3,5-diCQA =31,600, 4,5-diCQA =33,200, with λ=330 nm; 3-FQA =19,000, 4-FQA
=19,500, 5-FQA =19,300, with λ=325 nm) indicated by Correia[9] and Farah et al.,[20]
in L mol−1cm−1, were used. Data were within the detection limits of the method. All
extractions and chromatographic analysis were performed in triplicate.
Data were statistically analyzed using a one-way ANOVA (P≤0.05). Based on
the ANOVA results, a Tukey’s test was performed for mean comparison, for a 95% con-
fidence level. Different letters indicate significant differences for 95% confidence level.
Multivariate analysis was carried out with STATISTICA 6.0 software (Copyright StatSoft,
Inc., Tulsa, OK, USA), following several authors.[21–25]
RESULTS AND DISCUSSION
Coffee acidity depends on the geographic origin of green coffee beans, fruits matu-
rity, harvest process, weather conditions during harvesting and drying, and post-harvest
processing.[26,27] Additionally, coffee acidity might also be determined by growth condi-
tions, such as altitude[26,28] and shading,[28] but Robusta coffees are considered to have
low acidity, unlike the Arabica coffees. Although the average values were not dramatically
apart, the pH of Arabica green coffee was significantly lower than the Robusta green coffee,
with values ranging between 5.26–6.11 and 5.27–6.13, being in the range pointed out by
Leroy et al.[3] for these coffee types.
Mendonça et al.[4] supported that soluble solids for Arabica and Robusta green cof-
fees might vary between 23.85–27.31 and 26.07–30.6 g/100 g coffee, respectively, whereas
Esteves and Oliveira[29] pointed that in Robusta green coffees those values might vary
between 32.46–34.91 g/100 g coffee. The obtained content of soluble solids (Table 1)
was similar to those obtained by Mendonça et al.[4] for Arabica coffees, and Esteves and
Oliveira[29] for Robusta coffees. The soluble solids assayed in Arabica and Robusta green
coffees were not significantly different, but were slightly higher in Robusta green coffee
(Table 1).
Moreover, caffeine contents varied significantly being higher in the Robusta green
coffee (Table 1), showing much higher values (10- to 20-fold higher) than reported in
leaves, which did not present differences among the Coffea spp. genotypes,[19] contrary
to what is reported here for the bean. In fact, our results followed previous studies in
which Robusta coffees usually have a higher caffeine content than Arabica ones, although
in green coffee the levels of caffeine might vary widely, mostly due to inter-specific
differences.[7,30–33] Indeed, according to Viani,[7] the levels of caffeine might vary between
0.9–1.4 g/100 g in Arabica coffees, and between 1.5–2.6 g/100 g in Robusta ones, whereas
Macrae[30] presented a review of results obtained by various authors, pointing out average
values from 1.16–1.35 and 1.72–2.76 g/100 g coffee for Arabica and Robusta types, respec-
tively. Additionally, Silvarolla et al.[31] in 99 progenies from Ethiopia found caffeine values
varying between 0.42 and 2.9 g/100 g coffee, while in another study of 16 progenies of
C. arabica, tolerant and susceptible to Hemileia vastatrix, the levels ranged from 0.95 to
1.24 g/100 g coffee.[34] Salva and Lima[32] further reported that the caffeine content of
Arabica coffees varies between 1.1 and 1.8 g/100 g coffee.
The content of trigonelline in green coffees might vary with the kind of coffee and
geographical origin, being suggested that this compound could be used to discriminate the
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
CHEMICAL DISCRIMINATORS OF GREEN COFFEE 899
Tab le 1 Levels of pH, soluble solids, caffeine, trigonelline, 3-CQA, 4-CQA, 5-CQA, CQAtotal,
3,4-diCQA, 3,5-diCQA, 4,5-diCQA, diCQAtotal, 3-FQA, 5-FQA, FQAtotal ,andCGAof
Arabica and Robusta green coffee.∗
Green coffee samples
Parameters Arabica Robusta
pH 5.62 ±0.00r5.71 ±0.00s
Soluble solids (g/100 g) 33.32 ±0.07r33.96 ±0.52r
Caffeine (g/100 g) 1.454 ±0.007r2.382 ±0.019s
Trigonelline (g/100 g) 1.385 ±0.014r1.009 ±0.032s
CQA (g/100 g)
3-CQA 0.563 ±0.002r0.631 ±0.004s
4-CQA 0.711 ±0.008r0.801 ±0.002s
5-CQA 4.430 ±0.036r4.703 ±0.007s
CQAtotal 5.704 ±0.044r6.135 ±0.005s
diCQA (g/100 g)
3,4-diCQA 0.201 ±0.002r0.572 ±0.005s
3,5-diCQA 0.395 ±0.003r0.592 ±0.005s
4,5-diCQA 0.226 ±0.003r0.515 ±0.005s
diCQAtotal 0.822 ±0.007r1.680 ±0.014s
FQA (g/100 g)
3-FQA 0.029 ±0.002r0.059 ±0.000s
5-FQA 0.242 ±0.004r0.594 ±0.002s
FQAtotal 0.271 ±0.005r0.653 ±0.002s
CGA (g/100 g) 6.797 ±0.054r8.467 ±0.013s
∗Each value is the mean ±SE (n=3). Different letters (r, s) indicate significant differences,
in a multiple range analysis, for 95% confidence level.
geographical origin of coffee.[35] However, Aguiar et al.,[36] after examining different vari-
eties of C. canephora, concluded that the differences in content of trigonelline are too small
to constitute a useful tool to discriminate this kind of coffee. In this study, it was found that
the content of the alkaloid trigonelline was significantly higher in Arabica green coffee
(Table 1), therefore showing a pattern similar to previous reports.[30,36,37] Indeed, accord-
ing to Macrae,[30] trigonelline levels in Arabica and Robusta green coffee are around 1 and
0.7 g/100 g coffee, respectively, while Aguiar et al.[36] pointed values between 0.97 to
1.01 g/100 g coffee for green coffee Robusta. Ky et al.,[37] using seeds of C. arabica
and C. canephora from various geographical origins, reported trigonelline levels between
0.88–1.77 and 0.75–1.24 g/100 g coffee, respectively, showing a large variation that turns
this compound not suitable to discrimination.
CQA, diCQA, and FQA represent about 98% of total chlorogenic acids in coffee.[6]
Concerning CQAs, it was found (Table 1) that 5-CQA was the most representative com-
pound both in Arabica and Robusta green coffees, representing ca. 77% of total CQA
content, similarly to what was reported in the leaves of several Coffea spp. genotypes where
it accounts for ca. 85%.[19] Also, the levels of all caffeoylquinic acid isomers (3-CQA, 4-
CQA, and 5-CQA), as well as CQAtotal, were significantly higher in the Robusta green
coffee, within the range pointed out previously.[7,9,37,38]
The content of each isomer of diCQA was clearly and significantly higher in the
Robusta green coffee (Table 1). Quite similar amounts of 3,4-diCQA (0.572 g/100 g cof-
fee), 3,5-diCQA (0.592 g/100 g coffee), and 4,5-diCQA (0.515 g/100 g coffee) were found
in Robusta green coffee, while the isomer 3,5-diCQA (0.395 g/100 g coffee) predominated
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
900 BICHO ET AL.
in green Arabica coffee (Table 1). These patterns further followed the values indicated by
several authors.[37,38,39]
In agreement with previous results,[7,9,37,38] the 5-FQA isomer predominates (ca. 90%
of total FQA), being the contents of the others (3-FQA and 4-FQA) much less relevant.
Following previous works,[7,9,37,38] it was also found that Robusta green coffee had signifi-
cantly higher levels of 3-FQA and 5-FQA, as well as a significantly higher total amount of
FQA (Table 1).
The total content of each group of chlorogenic acid (CQAtotal,diCQA
total, and
FQAtotal), as well as CGAtotal , tended to be higher in Robusta green coffees (Table 1),
prevailing 5-CQA in green coffee. In fact, 5-CQA represented more than half (65 and 55%)
of total CGAs in Arabica and Robusta coffees (Table 1). That led to the higher weight
of CQAs, representing 84 and 72% of total CGAs in Arabica and Robusta green coffees,
respectively (Table 1). diCQA acids were the second most abundant group in the diCQAtotal,
corresponding to ca. 12 and 20% of total CGAs, while the FQAs represented ca. 4 and 8%
of the CGAtotal value, assayed in samples of Arabica and Robusta green coffees, respec-
tively. The content of total CGA in Robusta green coffee was ca. 25% higher than in the
Arabica green coffee (Table 1).
A cluster analysis might generate homogeneous groupings and segmentation based
on variables. Considering the pH, soluble solids, caffeine, trigonelline, CQAtotal,di-
CQAtotal, and FQAtotal , a cluster analysis, using a linkage distance around 0.6 (Fig. 1)
allowed a discrimination of Arabica and Robusta samples. Forming of clusters by the cho-
sen data set can result in a new variable that identifies cluster members among the cases.
In this context, Robusta green coffee showed a tendency to, in general, present higher values
for all parameters except trigonelline, but the content of soluble solids was not significantly
different in both species of green coffee (Table 1). These data suggest that the parameters
Figure 1 Samples (n=3) dendogram of Arabica (A) and Robusta (R) coffees considering the parameters pH,
soluble solids, caffeine, trigonelline, CQAtotal,di-CQA
total, and FQAtotal based on Euclidean distances between
them. (Color figure available online.)
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
CHEMICAL DISCRIMINATORS OF GREEN COFFEE 901
Figure 2 Green coffees dendogram of: (a) pH, soluble solids, caffeine, trigonelline, CQAtotal,diCQA
total,and
FQAtotal based on Euclidean distances between them; (b) pH, soluble solids, caffeine, trigonelline, the 3-CQA, 4-
CQA, 5-CQA, 3,4-diCQA, 3.5-diCQA, 4.5-DICQ, 3-FQA, and 5-FQA, based on the Euclidean distances between
them. (Color figure available online.)
can be subdivided into three groups, according to their importance for samples discrimi-
nation achieved with the cluster analysis (Figs. 1 and 2). Although the content of soluble
solids were not relevant, the pH and CQAtotal, particularly the fraction 5-CQA, constitute a
second group in terms of its importance in the discrimination of samples (Fig. 2), whereas
caffeine, trigonelline, 3-CQA, 4-CQA, 3-FQA, 5-FQA, 3,4-diCQA, 3,5-diCQA, and 4,5-
diCQA formed a third group of substances, highly discriminating for Arabica and Robusta
green coffees, because they have higher and significantly different values. Accordingly,
and considering that the chemical characterization of Arabica and Robusta green coffee
provided in this study, in general, parallels with the worldwide parameters for these coffee
species, it might be concluded that this third cluster might be used as a universal discrim-
inator. Moreover, considering the roast coffee chemical characteristics, this model cannot
be applied to roast coffee.[40,41]
CONCLUSION
The cluster analysis used for classification of chemical data of Arabica and Robusta
green coffee allowed a partitionation into groups based on a distance/dissimilarity func-
tion. In this context, caffeine, trigonelline, 3-O-caffeoylquinic acid, 4-O-caffeoylquinic
acid, 3-O-feruloylquinic acid, 5-O-feruloylquinic acid, 3,4-O-dicaffeoylquinic acid, 3,5-O-
dicaffeoylquinic acid, and 4,5-O-dicaffeoylquinic acid can be considered a chemical cluster
discriminator for Arábica and Robusta green coffees. Moreover, pH, soluble solids, 5-O-
caffeoylquinic acid, total chlorogenic acids, total caffeoylquinic acids, total feruloylquinic
acids, and total dicaffeoylquinic acids should not be used as discriminators, because their
contents are not significantly different between Arabica and Robusta green coffees.
NOMENCLATURE
CGA Total chlorogenic acids
CQAtotal Total caffeoylquinic acids
3-CQA 3-O-Caffeoylquinic acid
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
902 BICHO ET AL.
4-CQA 4-O-Caffeoylquinic acid
5-CQA 5-O-Caffeoylquinic acid
diCQAtotal Total dicaffeoylquinic acids
3,4-diCQA 3,4-O-Dicaffeoylquinic acid
3,5-diCQA 3,5-O-Dicaffeoylquinic acid
4,5-diCQA 4,5-O-Dicaffeoylquinic acid
FQAtotal Total feruloylquinic acids
3-FQA 3-O-Feruloylquinic acid
5-FQA 5-O-Feruloylquinic acid
ICO International Coffee Organization.
ACKNOWLEDGMENTS
The authors wish to thank Drs. Joel I. Fahl and Maria Luíza Carelli (IAC) for the supply of
seed material and Professor Dr. Santos Oliveira (FCT/UNL) for his scientific suggestions.
REFERENCES
1. Dentan, E. The microscopic structure of the coffee bean. In: Coffee, Botany, Biochemistry and
Production of Beans and Beverage; Clifford, M.N.; Wilson, K.C.; Eds.; Croom Helm & Methuen
Inc: New York, 1987; 284–304.
2. Nestlé. Coffee. Documentation Centre, Corporate Affairs Department, Nestec Ltd: Vevey,
Switzerland, 1991; 1–60.
3. Leroy, T.; Ribeyre, F.; Bertrand, B.; Charmetant, P.; Dufour, M.; Montagnon, C.; Marraccini, P.;
Pot, D. Genetics of coffee quality. Brazilian Journal of Plant Physiology 2006,18, 229–242.
4. Mendonça, L.M.V.L.; Pereira R.G.F.A.; Mendes, A.N.G. Parâmetros bromatológicos de grãos
crus e torrados de cultivares de café (Coffea arabica L.). Ciências Tecnologia Alimentar 2005,
25, 239–243.
5. Blitz, H.B.; Grocsh, W. Química de los alimentos. Editorial Acribia, S.A., Zaragoza, Espanha
1988, 1–6, 747–760.
6. Clifford, M.N.; Staniforth. P. S. A critical comparison of six spectrophotometric methods
for measuring chlorogenic acids in green coffee beans. In: Huitième Colloque Scientifique
International sur le Café; Abidjan, ASIC: Paris, 1977; 109–114.
7. Viani, R. The composition of coffee. In: Caffeine, Coffee, and Health; Garattini, S.; Ed.; Raven
Press, Ltd: New York, 1993; 17–41.
8. Clifford, M.N. Chlorogenic acids and other cinnamates-nature, occurrence and dietary burden.
Journal of the Science of Food and Agriculture 1999,79, 362–372.
9. Correia, M.N.G. Influência da torra na evolução dos ácidos clorogénicos do café; Instituto
Superior de Agronomia, Universidade Técnica de Lisboa, Lisboa, 1990.
10. Correia, M.N.G. Manual da Tecnologia do Café. Cultivar; Associação dos Técnicos de Culturas
Tropicais: Porto, 1995.
11. Smith, A.W. Introduction. In: Coffee, I: Chemistry; Clark, R.J.; Macrae, R.; Eds.; Elsevier
Applied Science Publishers Co Ltd: London and New York, 1989; 1–41.
12. Instrução Normativa no. 8. Regulamento técnico de identidade e de qualidade para a classificação
do café beneficiado grão cru. Ministério de Estado da Agricultura, Pecuária e Abastecimento:
Brasil, 2003.
13. NP 1666. Café–Colheita de amostras de café verde em sacos. Direcção Geral de Qualidade:
Lisboa, 1980.
14. ISO 4072. Green coffee in bags-sampling. International Organization for Standardization,
Geneva, Switzerland, 1982.
15. PSCB no. 36/02. Code of practrice: Enhancement of coffee quality through prevention of mould
formation. International Coffee Organization, London, England, 2002.
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
CHEMICAL DISCRIMINATORS OF GREEN COFFEE 903
16. AOAC. Coffee and Tea, Vol. 2; AOAC International–Association of Official Analytical Chemists,
Gaithersburg, MD, USA, 1996, Chapter 30.
17. AOAC. Definition of terms and explanatory notes. In: Official Methods of Analysis, Vol. 1;
AOAC International, Gaithersburg, MD, USA, 1996; note 11.
18. ISO 10095. Coffee–Determination of caffeine content–Method using high-performance liquid
chromatography. International Organization for Standardization, Geneva, Switzerland, 1992.
19. Fortunato, A.S.; Lidon, F.C.; Batista-Santos, P.; Leitão, A.E.; Pais, I.P.; Ribeiro, A.I.; Ramalho,
J.C. Biochemical and molecular characterization of the antioxidative system of Coffea sp. under
cold conditions in genotypes with contrasting tolerance. Journal of Plant Physiology 2010,167,
333–342.
20. Farah, A.; De Paulis, T.; Trugo, L.C.; Martin, P.R. Effect of roasting on the formation of
chlorogenic acid lactones in coffee. Journal of Agricultural and Food Chemistry 2005,53,
1505–1513.
21. Neto, J.M.M.; Moita, G.C. Uma introdução à análise exploratória de dados multivariados.
Química Nova 1998,21, 467–469.
22. Alvarenga, N.; Silva, P.; Rodriguez-Garcia, J.; Sousa. I. Changes in chemical, physical and rhe-
ological properties of “Serpa” cheese during the ripening. Journal of Dairy Research 2008,75,
233–239.
23. Chapman, K.W.; Lawless, H.T.; Boor, K.J. Quantitative descriptive analysis and principal com-
ponent analysis for sensory characterization of ultrapasteurized milk. Journal of Dairy Science
2001,84, 12–20.
24. Maeztu, L.; Andueza, S.; Ibañez, C.; de Peña, M.P.; Bello, J.; Cid, C. A multivariate method
for differentiation of espresso from botanical varieties and types of roast by foam, taste and
mouthfeel characteristics. Journal of Agricultural and Food Chemistry 2001,49, 4743–4747.
25. Da Silva, A.F.S.; Minim, V.P.R.; Chaves, J.B.P.; Stringheta, P.C.; Ribeiro, M.M. Avaliação do
gosto amargo da bebida de café (Coffea arábica L.) orgânico por meio da análise intensidade–
tempo. Ciências Tecnologia Alimentar 2004,24, 468–472.
26. Clifford, M. N. Chemical and physical aspects of green coffee products.InClifford,M.N.;
Wilson, K.C.; Eds. Coffee, Botany, Biochemistry and Production of Beans and Beverage; Croom
Helm & Methuen Inc.: New York, 1987; 305–374.
27. Salva, T.J.G. A composição química dos grãos e a qualidade da bebida de café como consequên-
cia do método de preparo e cultivar. IAC -Instituto Agronómico, Centro de Análise e Pesquisa
Tecnológica do Agronegócio do Café “Alcides de Carvalho,” Brasil, 2007.
28. Witegens, J.N. Factores que influencian la calidad del café. XV Simpósio Latinoamericano
de Cafeicultura, Xalapa, Veracruz, Mexico, 21–24 July 1992, Instituto Interamericano de
Cooperaction para la Agricultura, México, 1992.
29. Esteves, A.B.; Oliveira, J.S. Contribuição para o estudo das características dos cafés de Angola.
Estudos, Ensaios e Documentos no. 126, Junta de Investigações do Ultramar Lisboa 1970,
126–177.
30. Macrae, R. Nitrogenous components. In: Coffee, I: Chemistry; Clarke, R.J.; Macrae, R.; Eds.;
Elsevier Applied Science Publishers Co Ltd: London and New York, 1989; 115–152.
31. Silvarolla, M.B.; Mazzafera, P.; Lima, M.M.A. Caffeine content of Ethiopian Coffea arabica
beans. Genetics and Molecular Biology 2000,23, 213–215.
32. Salva, T.J.G.; Lima, V.B. A composição química do café e as características da bebida e do grão.
O Agronómico 2007,59, 57–59.
33. Coste, R. Coffee: The Plant and the Product; MacMillan Press Ltd: London, 1992.
34. Mendonça, L.M.V.L.; Pereira, R.G.F.A.; Mendes, A.N.G.; Borém, F.M.; Marques, E.R. Chemical
composition of raw grains of cultivars the Coffea arabica L. susceptible and resistant to Hemileia
vastatrix Berg et Br. Ciência & Agrotecnologia 2007,31, 413–419.
35. Bicho, N.C.C. Qualidade e segurança alimentar no café verde e aclimatação do género
coffea em condições de stress ambiental. Master’s Thesis, Faculdade de Ciências e Tecnologia,
Universidade Nova de Lisboa, Portugal, 2005, 276 pp.
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
904 BICHO ET AL.
36. Aguiar, A.T.E.; Fazuoli, L.C.; Garcia, T.J.G.; Favarin J.F. Diversidade química de cafeeiros na
Espécie Coffea canephora. Bragantia 2005,64, 577–582.
37. Ky, L.C.; Louarn, J.; Dussert, S.; Guyot, B.; Hamon, S.; Noirot, M. Caffeine, trigonelline,
chlorogenic acids and sucrose diversity in wild Coffea arabica L. and C. canephora P. accessions.
Food Chemistry 2001,75, 223–230.
38. Clifford, M.N. Chlorogenic acids. In: Coffee, I: Chemistry; Clarke, R.J.; Macrae, R.; Eds.;
Elsevier Applied Science Publishers Co Ltd: London and New York, 1989; 153–202.
39. Clifford, M.N. The nature of chlorogenic acids. Are they advantageous compounds in coffee?.
In: 17eme Colloque Scientifique International sur le Café–Nairobi; ASIC: Paris, 1997; 79–91.
40. Casal, S.; Oliveira, M.B.P.P.; Alves, M.R.; Ferreira, M.A. Discriminate analysis of roasted coffee
varieties for Trigonelline, nicotinic acid, and caffeine content. Journal of Agricultural and Food
Chemistry 2000,48 (8), 3420–3424.
41. Severa, L.; Buchar, J.; Nedomová, S. Shape and size variability of roasted Arabica coffee beans.
International Journal of Food Properties 2012,15 (2), 426–437.
Downloaded by [b-on: Biblioteca do conhecimento online UNL] at 12:19 20 January 2014
