Journal of the Science of Food and Agriculture J Sci Food Agric 87:586–592 (2007)

Influence of coffee/water ratioon the final quality of espresso coffee†

Susana Andueza, Marıa A Vila, M Paz de Pena and Concepcion Cid∗Food Science and Technology, and Toxicology Department, School of Pharmacy, University of Navarra, E-31080 Pamplona, Spain

Abstract: Espresso coffee is a polyphasic beverage in which the physico-chemical and sensory characteristicsobviously depend on both the selection of ground roasted coffee and the technical conditions of the percolationprocess. The aim of this work was to evaluate the influence of the coffee/water ratio on the physico-chemical andsensory quality of espresso coffee. Furthermore, the influence of botanical varieties (Arabica and Robusta) andthe type of roast (conventional and torrefacto) on the selection of coffee/water ratio was studied. The relationshipbetween pH and the perception of acidity intensity is discussed in relation to the influence of the coffee/water ratio,type of coffee and roast. The optimisation of other technical parameters in previous studies seemed to minimisethe influence of an increase in the coffee/water ratio on the extraction of soluble and solid compounds. In fact, onlysome sensory attributes, such as bitterness, astringency and burnt, acrid and earthy/musty flavours were proposedas relevant to the selection of 6.5 g 40 mL−1 or 7.5 g 40 mL−1 in conventional roasted coffees (Arabica 100% andRobusta blend), and 6.5 g 40 mL−1 in torrefacto roasted coffees. On the other hand, the addition of sugar during theroasting process in torrefacto roast coffees seemed to contribute to a higher generation of acids, melanoidins andother compounds by the Maillard reaction or caramelisation, which led us to select the lowest coffee/water ratio. 2007 Society of Chemical Industry

Keywords: coffee; espresso coffee; coffee/water ratio; sensory analysis; aroma; torrefacto roast

INTRODUCTIONEspresso Coffee is a polyphasic beverage prepared onlywith ground roasted coffee and water, and constitutedby a foam layer of small bubbles with a particular tiger-tail pattern, on the top of an emulsion of microscopicoil droplets in an aqueous solution of sugars, acids,protein-like material and caffeine, with dispersed gasbubbles and colloidal solids.1 These physico-chemicalcharacteristics of espresso coffee are responsible fortheir peculiar sensorial properties which include astrong body, a full fine aroma, a bitter/acid balancetaste and a pleasant lingering aftertaste, exempt fromunpleasant flavour defects.1

The physico-chemical and sensory characteristicsof an espresso coffee obviously depend on both theselection of ground roasted coffee and the technicalconditions of the percolation process, which shouldbe adjusted according to the coffee.2 In previousstudies, the optimal water temperature and pressure toobtain a good quality espresso coffee were establishedas 92 ◦C and 909 kPa (9 atm).3,4 However, othertechnical conditions related to coffee, such as gradeof grinding, should be different whether coffee wasroasted by the conventional or the torrefacto process.5

Torrefacto is a roasting process where sugar is addedto Robusta coffees. This type of roast contributesto the brownish colour of the coffee brew by

the caramelisation of sugar and the enhancing ofthe Maillard reaction products (MRPs). Also, thetorrefacto roast was initially used to mask the negativesensory characteristics of low quality Robusta coffees.This roasting technique is used in several countriesof southern Europe and South America, where somegroups of the population prefer espresso coffees witha high amount of foam, a dark brown colour, avery intense aroma, and a strong taste, with a bitterdominance.1

The coffee/water ratio is another factor that couldinfluence the extraction and quality of the coffeecompounds. An excessive amount of coffee would notallow sufficient expansion during wetting, thus causingover-compaction, which disturbs the percolation andresults in a deposit of solids in the cup. In contrast, toolittle coffee could result in a brew that is over-extractedand has a bitter flavour.1,6 Based on his commercialand technical experience Petracco1 proposed a rangebetween 5 g and 8 g of ground coffee for preparingone cup of espresso, depending on the coffee blend.However, only a few studies investigating the influenceof the coffee/water ratio on the coffee brew havebeen found but these focused on the kinetics andmechanisms of caffeine or solubles extraction inpressureless systems7,8 rather than on the chemicaland sensory characteristics of espresso coffees.

∗ Correspondence to: Concepcion Cid, Food Science and Technology, and Toxicology Department, School of Pharmacy, University of Navarra, E-31080Pamplona, SpainE-mail: ccid@unav.es†Part of this paper was presented at the 5th Pangborn Sensory Science Symposium, July 2003, Boston, MA, USA.Contract/grant sponsor: Comision Interministerial de Ciencia y Tecnologıa; contract/grant number: ALI-1999-0319Contract/grant sponsor: Ministerio de Ciencia y Tecnologıa EspanolContract/grant sponsor: Gobierno Vasco; contract/grant number: PN99-72681468(Received 13 December 2004; revised version received 6 March 2006; accepted 4 September 2006)Published online 15 January 2007; DOI: 10.1002/jsfa.2720

2007 Society of Chemical Industry. J Sci Food Agric 0022–5142/2007/$30.00

Coffee/water ratio on espresso coffee quality

The aim of this work was to evaluate the influence ofcoffee/water ratio on the physico-chemical and sensoryquality of espresso coffee. Furthermore, the influenceof botanical varieties (Arabica and Robusta) and thetype of roast (conventional and torrefacto) on theselection of the coffee/water ratio was studied.

MATERIALS AND METHODSMaterialsThree ground roasted coffee samples, pure Coffeaarabica from Colombia (conventional espresso roast,2% water content) (A100); Arabica/Robusta 20:80blend, (conventional espresso roast, 2% water content)(A20:R80); and a blend of Arabica/Robusta 20:80 with50% of torrefacto roast Robusta coffee (A20:R80 50%torrefacto, 1.8% water content) were provided by alocal company. Two batches of each coffee samplewere used. Samples were stored in similar conditions(4 ◦C, vacuum package, less than 2 days) before andduring analysis.

Pure reference standards of acetaldehyde, 2-methylpropanal, 3-methylbutanal, 2,3-butandione,2,3-pentandione and 2-ethyl-3,5dimethylpyrazinewere purchased from Acros (New Jersey, USA); hex-anal, 2-methoxyphenol (guaiacol), propanal, caffeineand 5-caffeoylquinic acid (5-CQA) were obtainedfrom Sigma (Steinheim, Germany).

Selection of the coffee/water ratioTo select coffee/water ratios, espresso coffees werebrewed for each sample with the experimentalprototype espresso coffee-maker under the conditionsgiven below. A volume of 40 ± 2 mL, time ofpercolation between 18 and 24 s, and the absenceof particles at the bottom of the cup were the maincriteria for selecting the coffee/water ratio. Weights of6.5, 7.5 and 8.5 g of coffee to prepare an espresso cupof 40 ± 2 mL were selected as low, medium and highcoffee/water ratios, respectively.

Espresso coffee samples and preparation foranalysisEspresso coffees were prepared from each selectedcoffee/water ratio with the use of an experimentalprototype espresso coffee-maker. The conditions forpreparing espresso coffee were fixed at 92 ◦C watertemperature (corresponding to erogation temperatureof 86 ± 2 ◦C), relative water pressure of 909 kPa (9atm), extraction time of 21 ± 3 s and holder filterdiameter of 38 mm. Twenty espresso coffees of eachcoffee/water ratio were prepared and mixed togetherin order to have sufficient coffee volume for analysis.Every parameter was determined in triplicate.

pH, density, viscosity and surface tensionEspresso coffee samples were immediately cooledto 20 ◦C, and pH (Orion 420A benchtop pHmeter, Beverly, MA), density (densimeter), viscosity(Ostwald viscosimeter, Proton, Barcelona, Spain),

and surface tension (Traube estalagmometer, Oppac,Pamplona, Spain) were measured.

Foam index and persistence of foamFoam index was defined as the volume of espresso cof-fee in millilitres, referred to 100 mL of espresso coffeetotal volume.1 Volumes were measured immediatelyafter the extraction of espresso coffee using a 100-mLgraduated cylinder. Persistence of foam was definedas the time (in minutes) that the liquid phase belowthe cream layer took to appear during cooling at roomtemperature.1

Total solids, extraction and total solids on filtrateTotal solids were determined by oven drying 40 mL ofespresso coffee to a constant weight (14 h, 102 ± 3 ◦C).Extraction was defined as the percentage of total solidswith respect to the dose of ground roasted coffee. Totalsolids in the filtrate (or soluble solids) were definedas the dry residue, expressed in mg mL−1, obtainedby oven drying the eluate obtained by filtration withWhatman no. 1 filter paper of 40 mL espresso coffeeto constant weight (14 h, 102 ± 3 ◦C).

Total lipidsTwenty millilitres of espresso coffee was extracted byadding 20 mL of trichloromethane three times in aseparating funnel. The organic fraction was washedwith distilled water three times. Total lipids werequantified by weight after evaporation of the solvent.

Caffeine and trigonellineExtract preparation, clean-up and high-performanceliquid chromatography (HPLC) analysis have alreadybeen described by Maeztu et al.9 HPLC analysis wasachieved with an analytical HPLC unit (Hewlett-Packard 1100, Madrid, Spain). A reversed-phaseHypersil-ODS (5 µm particle size, 250 × 4.6 mm)column was used. The mobile phase was acetoni-trile–water (15:85) under isocratic conditions at aconstant flow rate of 2.0 mL min−1 at 25 ◦C. Com-pounds were detected by using a diode-array detector,and chromatograms were recorded at 280 nm.

Chlorogenic acid (5-CQA)Extraction of 5-CQA and clean-up were carriedout according to the method of Bicchi et al.10 withHPLC equipment described above. Conditions of thegradient solvent system used were 100% citrate–aceticacid buffer solution (pH 3.0) for 2 min, 85:15buffer–methanol for 8 min, both at a flow rate of0.8 mL min−1, and 85:15 buffer–methanol for 5 minat a flow rate of 1.2 mL min−1, at 25 ◦C. The detectionwavelength was 325nm.

Volatile compoundsThe extraction of volatile compounds and GC analysiswere carried out by using the method describedby Sanz et al.11 adapted to espresso coffee by

J Sci Food Agric 87:586–592 (2007) 587DOI: 10.1002/jsfa

S Andueza et al.

Maeztu et al.12 Volatiles were extracted using a staticheadspace sampler (Hewlett-Packard model 7694).GC analysis was achieved with a capillary HP-Wax(60 m × 0.25 mm × 0.5 mm film thickness) columnin a HP 6890 gas chromatograph with a HP 5973mass selective detector (Hewlett-Packard). Volatilecompounds were identified by referring to the massspectra in the Wiley database, their retention timesand Kovats indices.13,14 Thirteen key odorants werequantified, and results were expressed as relativepercentages from the total amounts of volatiles.

Sensory descriptive analysisTwenty judges were recruited from members of theFood Science and Technology Department at theUniversity of Navarra. Selection and training werecarried out as described by Maeztu et al.12 to obtain a10-member panel. Odour, body, acidity, bitterness,astringency, flavour and aftertaste intensities wererated on 11-point scales from ‘none’ (0) to ‘veryhigh’ (10). The mean and standard deviation for eachattribute in each espresso coffee sample were obtained.

Sensory flavour profileThe most frequently described odour/flavour attributesby judges during the training process were written inthe same score card in two columns: one for positiveand another for negative flavour attributes. Posi-tive flavour attributes were fruity/winey, malty/cereal,freshness, straw, caramel-like, equilibrate, chocolate-like, spicy, nutty, tobacco, and buttery. Negativeflavour attributes were woody/papery, burnt/roasty,acrid, fermented, earthy/musty, rancid, burnt rub-bery, sulfurous, flat, grassy/green/herbal, animal-like,motor-oil and ashy. In both columns, one line for‘other flavours’ was added. The flavour profile of eachespresso coffee sample was defined by the percentageof judges that perceived each positive and negativeflavour attribute.

Sensory descriptive evaluation of espresso coffeesamples was carried out in triplicate over 18 sessions.Three espresso coffees were analysed per session.Each espresso coffee was prepared immediately beforetasting, and served monadically in white porcelaincoffee cups labelled with a three-digit code. The orderof presentation was randomised among judges andsessions. All evaluations were conducted in isolatedsensory booths illuminated with white light in thesensory laboratory under standardised conditions asgiven by UNE 87-004-79.15 Rinse water was providedbetween individual samples.

Statistical analysisAnalysis of variance (ANOVA) was applied for eachtype of coffee. The source of variation was thecoffee/water ratio. The Tukey T-test was applied aposteriori with a level of significance of 95%. Allstatistical analyses were performed using the SPSSv.10.0 software package. T

able

1.In

fluen

ceof

coffe

e/w

ater

ratio

(g40

mL−1

)on

phy

sico

-che

mic

alp

aram

eter

sof

esp

ress

oco

ffee

sam

ple

s

A10

0A

20:R

80A

20:R

80,5

0%to

rref

acto

6.5

7.5

8.5

6.5

7.5

8.5

6.5

7.5

8.5

pH5.

5;0.

0b5.

4;0.

0a5.

4;0.

0a5.

9;0.

1b5.

9;0.

0b5.

8;0.

0a5.

8;0.

0c5.

7;0.

0b5.

6;0.

0a

Den

sity

(gm

L−1)

1.00

5;0.

001a

1.01

0;0.

000b

1.01

0;0.

000b

1.01

0;0.

000a

1.01

0;0.

000a

1.01

2;0.

000b

1.00

5;0.

001a

1.01

0;0.

000b

1.01

0;0.

000b

Vis

cosi

ty(m

Nm

−2s−

1)

1.23

;0.0

1a1.

29;0

.02c

1.26

;0.0

1b1.

14;0

.01a

1.17

;0.0

2b1.

17;0

.01b

1.15

;0.0

0a1.

19;0

.04b

1.21

;0.0

0b

Sur

face

tens

ion

(mN

m−1

)48

.37;

0.90

a48

.81;

0.99

ab50

.61;

0.96

b39

.39;

0.00

a51

.58;

1.05

b51

.00;

0.00

b47

.81;

0.00

a51

.58;

1.05

b50

.90;

0.00

b

Foam

inde

x(%

)12

.3;0

.3a

14.4

;1.0

b17

.7;0

.6c

15.8

;0.0

a21

.9;0

.0b

30.8

;0.7

c20

.4;2

.7a

20.8

;6.6

a22

.9;0

.8b

Per

sist

ence

offo

am(m

in)

25.3

3;0.

51a

30.0

0;0.

00b

30.0

0;0.

00b

30.0

0;0.

00a

30.0

0;0.

00a

30.0

0;0.

00a

20.0

0;0.

00a

30.0

0;0.

00b

30.0

0;0.

00b

Tota

lsol

ids

(mg

mL−1

)34

.03;

0.21

a37

.24;

0.73

b42

.07;

0.39

c35

.58;

0.41

a36

.01;

0.41

b40

.76;

0.60

c36

.17;

0.17

a38

.20;

0.97

b43

.11;

0.33

c

Ext

ract

ion

(%)

20.9

;0.2

b19

.9;0

.4a

19.8

;0.2

a20

.7;0

.2b

19.2

;0.4

a19

.2;0

.2a

22.3

;0.1

b20

.4;0

.5a

20.3

;0.2

a

Tota

lsol

ids

onfil

trat

e(m

gm

L−1)

32.0

;0.3

a35

.2;0

.5b

39.7

;0.9

c30

.7;0

.9a

35.6

;0.7

b38

.8;0

.8c

33.0

;0.1

a34

.4;0

.7b

40.9

;0.1

c

Tota

llip

ids

(mg

mL−1

)4.

80;0

.17a

5.06

;0.0

4b5.

13;0

.09b

3.34

;0.1

0a3.

77;0

.03b

3.92

;0.0

8c3.

46;0

.06a

3.76

;0.0

2b4.

06;0

.08c

Caf

fein

e(m

gm

L−1)

1.80

;0.0

4a1.

88;0

.12a

2.21

;0.0

5b3.

01;0

.08a

3.17

;0.0

7a3.

31;0

.18b

2.49

;0.2

3a2.

76;0

.07b

3.15

;0.0

5c

Trig

onel

line

(mg

mL−1

)0.

73;0

.06a

0.72

;0.0

5a0.

94;0

.09b

1.49

;0.1

7b1.

55;0

.15b

1.29

;0.0

8a1.

66;0

.17b

1.74

;0.0

7b0.

93;0

.06a

Chl

orog

enic

Aci

d(5

-CQ

A)(

mg

mL−1

)1.

16;0

.17a

1.38

;0.0

4b1.

80;0

.06c

1.18

;0.0

7a1.

45;0

.02b

1.52

;0.0

3c1.

28;0

.07a

1.39

;0.0

2b1.

50;0

.08c

All

valu

esar

esh

own

asm

ean;

stan

dar

dd

evia

tion

(n=

6).

Inea

chro

w,d

iffer

ent

lett

ers

ind

icat

esi

gnifi

cant

diff

eren

ce(P

<0.

05)a

mon

gd

iffer

ent

coffe

e/w

ater

ratio

inea

chty

pe

ofco

ffee.

588 J Sci Food Agric 87:586–592 (2007)DOI: 10.1002/jsfa

Coffee/water ratio on espresso coffee quality

RESULTS AND DISCUSSIONThe results of the physico-chemical parameters andsensory attributes of espresso coffee samples are shownin Tables 1 and 2, respectively. pH values were in therange proposed by Petracco1 as normal for espressocoffee (5.2–5.8), except for 6.5 and 7.5 g 40 mL−1

A20:R80 blend espresso coffees with 5.9. Arabica(A100) espresso coffees showed lower pH valuesthan Robusta blends, as previously reported by otherauthors.16 These results could partially explain thehigh perception of acidity by the judges of A100espresso coffees (Table 2). In A100 espresso coffee,a significant increase of the acidity with coffee/waterratio was observed. However, in Robusta blends,a clear tendency in the perception of acidity withcoffee/water ratio was not found. Espresso aciditycannot be described only by pH1 and, in fact,many studies have shown that there is only moderatecorrelation between pH and the acidity perception.16

Consequently, although a higher extraction of acidscould be proposed when the coffee/water ratio wasincreased, the diversity of the acids, volatiles andnon-volatiles, organic acids such as citric and malic,as well as chlorogenic acids and its hydrolysisderivatives (quinic, ferulic, cafeic acids), and inorganicacids, such as phosphoric, contribute in differentproportions to acidity.17 Furthermore, some of theseacids, such as chlorogenic, together with caffeine andother compounds, can also contribute to bitterness,modifying the typical bitterness–acidity balance ofespresso coffees.1 For all these reasons, pH does notseem to be a good reference in espresso coffee forproposing the best coffee/water ratio when the othertechnical parameters were previously optimised.

On the other hand, torrefacto espresso coffees hadlower pH values than did A20:R80 espresso coffee.This may be due to a greater amount of acids generatedby the Maillard reaction, or by caramelisation18

because of the addition of sugars during the roastingprocess.

Although foam index was significantly increasedwith coffee/water ratio, all espresso coffees hada sufficient amount of consistent, persistent andhazelnut foam. So, the coffee/water ratio in theproposed range seems to have less influence on the

amount and quality of the foam than do other technicalparameters such as water extraction pressure3 andtemperature,4 and grinding.5 On the other hand,Robusta blends espresso coffees had higher foamindices than Arabica, maybe because the presenceof unknown tensioactive substances which increasefoam.1

Density, viscosity, surface tension, total solids, totalsolids on filtrate and total lipids were significantlyincreased with coffee/water ratio (Table 1). However,only the significant stronger body was appreciatedby the judges of the Robusta blends, mainly withtorrefacto roast (Table 2). Furthermore, a higherviscosity was observed in Arabica espresso coffee,which also had higher amounts of lipids, becauseviscosity is influenced by the amount of lipid dropletsin the emulsion.1 Additionally, although the totalsolids increased with coffee/water ratio, extractionyields, which are dose dependent, were slightlydecreased or maintained because the extractionconditions for espresso coffee had been optimisedin previous studies.3–5 A similar, but stronger patternwas observed by Cammenga et al.8

The extraction of caffeine and chlorogenic acid,compounds related to bitterness and astringency,increased when the coffee/water ratio was higher.However, significant increases in bitterness andastringency were perceived only in Robusta blendsespresso coffees, and mainly in torrefacto coffees. Thiscould be partially due to the higher amount of caffeineand trigonelline in the Robusta variety,1,19,20 and,consequently, higher extraction in espresso coffees.Furthermore, the formation of other unidentifiedbitter compounds derived from Maillard reactionsand caramelisation during roasting of Robusta coffees,and mainly in torrefacto roast, could also contributeto the greater bitterness and astringency of Robustablends espresso coffee. A bitterness intensity higherthan 7.5 on a 10-scale should be proposed to reject8.5 g 40 mL−1 in both Robusta blends espresso coffeesand also 7.5 g 40 mL−1 in torrefacto coffees.

Results for aroma/flavour are shown in Table 3 andFig. 1. In Arabica (A100) espresso coffee, profiles ofthe majority of key odorants were significantly similarthroughout the three coffee/water ratios. However,

Table 2. Influence of coffee/water ratio (g 40 mL−1) on sensory attributes of espresso coffee samples

A100 A20:R80 A20:R80, 50% torrefacto

6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5

Odour intensity 6.3; 0.8a 6.5; 0.7a 6.6; 0.7a 6.3; 0.8a 6.2; 0.5a 6.2; 1.0a 6.1; 0.8a 6.0; 0.7a 6.4; 0.8b

Body 6.1; 0.7a 6.0; 0.6a 6.1; 0.7a 6.1; 0.7a 6.2; 0.6a 7.0; 0.9b 6.4; 0.9a 6.9; 1.0b 7.3; 0.7c

Acidity 5.4; 0.8a 6.0; 1.0b 6.4; 1.0c 1.6; 0.6a 2.0; 0.7b 1.4; 0.7a 3.5; 0.7b 3.6; 1.0b 1.8; 0.9a

Bitterness 6.6; 1.3a 6.7; 1.0a 7.0; 0.8a 6.9; 1.1a 7.3; 1.2a 8.2; 1.2b 6.8; 0.9a 7.5; 0.9b 8.2; 1.0c

Astringency 6.1; 1.0a 5.8; 0.9a 6.0; 0.9a 6.0; 0.9a 6.4; 0.9a 7.6; 1.4b 5.7; 0.6a 6.4; 0.8b 7.2; 1.0c

Flavour intensity 6.7; 0.9a 6.7; 0.9a 6.8; 0.6a 6.5; 0.9a 6.6; 0.8a 7.7; 0.9b 6.4; 0.9a 6.7; 0.7b 7.1; 0.8c

Aftertaste intensity 6.1; 0.8a 6.3; 0.8a 6.6; 0.7b 6.8; 1.0a 6.9; 0.9a 7.7; 1.1b 6.6; 1.0a 6.9; 0.9ab 7.2; 0.9b

All values are shown as mean; standard deviation (n = 6).In each row, different letters indicate significant difference (P < 0.05) among different coffee/water ratio in each type of coffee.

J Sci Food Agric 87:586–592 (2007) 589DOI: 10.1002/jsfa

S Andueza et al.

Tab

le3.

Influ

ence

ofco

ffee/

wat

erra

tio(g

40m

L−1)o

nth

ere

lativ

ep

erce

ntag

eof

key

odor

ants

ines

pre

sso

coffe

esa

mp

les

A10

0A

20:R

80A

20:R

80,5

0%to

rref

acto

KI1

ID2

Key

odor

ant

6.5

7.5

8.5

6.5

7.5

8.5

6.5

7.5

8.5

SU

LFU

RC

OM

PO

UN

DS

635

CM

etha

noth

iole

0.16

;0.0

1a0.

24;0

.01b

0.22

;0.0

2b0.

10;0

.00a

0.12

;0.0

1b0.

12;0

.02b

0.15

;0.0

1a0.

16;0

.00b

0.16

;0.0

0b

ALD

EH

YD

ES

645

AA

ceta

ldeh

yde

0.49

;0.0

4a0.

54;0

.04a

0.51

;0.0

5a0.

34;0

.02a

0.35

;0.0

3a0.

38;0

.02a

0.37

;0.0

1b0.

32;0

.02a

0.39

;0.0

3b

712

AP

ropa

nal

0.71

;0.1

0a0.

74;0

.08a

0.74

;0.0

9a0.

47;0

.04a

0.49

;0.0

4a0.

54;0

.01b

0.45

;0.0

3ab0.

43;0

.02a

0.47

;0.0

3b

747

A2-

Met

hylp

ropa

nal

2.36

;0.3

8a2.

75;0

.26a

2.74

;0.2

3a2.

00;0

.18a

2.27

;0.1

9b2.

41;0

.19b

2.05

;0.1

5ab1.

93;0

.21a

2.17

;0.0

8b

880

C2-

Met

hylb

utan

al1.

36;0

.20a

1.60

;0.2

0a1.

53;0

.12a

1.26

;0.1

4a1.

43;0

.15ab

1.58

;0.2

0b1.

24;0

.09a

1.20

;0.1

7a1.

25;0

.05a

884

A3-

Met

hylb

utan

al3.

60;0

.51a

4.28

;0.4

8a4.

09;0

.37a

2.35

;0.2

9a2.

65;0

.23ab

2.79

;0.1

5b2.

56;0

.18a

2.44

;0.3

0a2.

46;0

.33a

1084

AH

exan

al0.

08;0

.02a

0.07

;0.0

3a0.

06;0

.01a

0.10

;0.0

3a0.

09;0

.02a

0.07

;0.0

1a0.

07;0

.01a

0.04

;0.0

1a0.

04;0

.01a

KE

TON

ES

962

A2,

3-B

utan

dion

e0.

53;0

.09a

0.57

;0.0

8a0.

54;0

.04a

0.34

;0.0

4a0.

32;0

.02a

0.36

;0.0

1a0.

32;0

.02ab

0.30

;0.0

1a0.

33;0

.02b

1058

A2,

3-P

enta

ndio

ne0.

83;0

.12a

0.87

;0.0

7a0.

85;0

.09a

0.46

;0.0

4a0.

45;0

.04a

0.48

;0.0

1a0.

44;0

.01b

0.40

;0.0

1a0.

45;0

.04b

PY

RA

ZIN

ES

1359

CE

thyl

pyra

zine

0.14

;0.0

1b0.

11;0

.02a

0.11

;0.0

1a0.

17;0

.01ab

0.16

;0.0

1a0.

19;0

.01b

0.12

;0.0

1ab0.

12;0

.00a

0.13

;0.0

1b

1411

C2-

Eth

yl-6

-met

hylp

yraz

ine

0.04

;0.0

0a0.

04;0

.00a

0.04

;0.0

0a0.

07;0

.00b

0.06

;0.0

0b0.

07;0

.00a

0.06

;0.0

1b0.

05;0

.01b

0.05

;0.0

0a

1475

A2-

Eth

yl-3

,5-d

imet

hylp

yraz

ine

0.05

;0.0

0a0.

05;0

.00a

0.04

;0.0

0a0.

08;0

.00a

0.08

;0.0

1a0.

08;0

.01a

0.13

;0.0

4a0.

06;0

.00a

0.06

;0.0

0a

PH

EN

OLI

CC

OM

PO

UN

DS

A2-

Met

hoxy

phen

ol(g

uaia

col)

0.03

;0.0

0a0.

04;0

.00ab

0.08

;0.0

4b0.

04;0

.00a

0.04

;0.0

0a0.

04;0

.00a

0.04

;0.0

0a0.

04;0

.00a

0.04

;0.0

0a

All

valu

esar

esh

own

asm

ean;

stan

dar

dd

evia

tion

(n=

6).

Inea

chro

w,

diff

eren

tle

tter

sin

dic

ate

sign

ifica

ntd

iffer

ence

(P<

0.05

)am

ong

diff

eren

tco

ffee/

wat

erra

tioin

each

coffe

esa

mp

le.1

KI,

Kov

ats

ind

exca

lcul

ated

for

the

HP

-Wax

cap

illar

yco

lum

n.2

The

relia

bili

tyof

the

iden

tifica

tion

pro

pos

alis

ind

icat

edb

yth

efo

llow

ing:

A,m

ass

spec

trum

,ret

entio

ntim

e,an

dK

ovat

sin

dex

acco

rdin

gto

stan

dar

ds;

C,m

ass

spec

trum

,com

par

edw

ithth

eW

iley

mas

ssp

ectr

ald

atab

ases

.

590 J Sci Food Agric 87:586–592 (2007)DOI: 10.1002/jsfa

Coffee/water ratio on espresso coffee quality

b

b a aa

c

c

c

ac

b

ba

c

b a b aa

a

a a ba b

a a0

20

40

60

80

Mal

ty/c

erea

l

Fre

sh

Str

aw

Equ

ilibr

ate

Woo

dy/p

aper

y

Bur

nt/r

oast

y

Acr

id

Fer

men

ted

Ear

thy/

mus

ty

6.57.5

8.5

(a) A100

Per

cent

of j

udge

s

a b aa b b

ab

a a a

a aa

b b

b b b ab a

aa

020406080

Mal

ty/c

erea

l

Fre

sh

Str

aw

Woo

dy/p

aper

y

Bur

nt/r

oast

y

Acr

id

Fer

men

ted

Ear

thy/

mus

ty

6.57.5

8.5

(b) A20:R80

Per

cent

of j

udge

s

b b ab

c c

a a

cb

b ab c b

a ab

020406080

100

Mal

ty/c

erea

l

Woo

dy/p

aper

y

Bur

nt/r

oast

y

Acr

id

Fer

men

ted

Ear

thy/

mus

ty

6.57.5

8.5

(c) A20:R80, 50% Torrefacto

Per

cent

of j

udge

s

Figure 1. Influence of coffee/water ratio (g 40 mL−1) on espresso coffee flavour profile. For each parameter, different letters indicate significantdifferences (P < 0.05) among different coffee/water ratios.

in 8.5 g 40 mL−1 A100 espresso coffee, burnt/roasty,acrid and fermented flavours were perceived bysignificantly higher percentage of judges, and freshnesswas less detected with coffee/water ratio increase. Thissensory flavour profile led us to reject 8.5 g 40 mL−1

coffee/water ratio in Arabica coffee.2-Methylpropanal, 2-methylbutanal, 3-methylbuta-

nal, Strecker degradation products of valine, isoleucineand leucine related to the malty flavour in coffeebrews,21,22 were extracted in lower amounts in 6.5 g40 mL−1 A20:R80 espresso coffees. On the otherhand, pyrazines, which have been related to negativeflavours such as burnt/roasty, woody/papery andearthy/musty in ground and brewed coffees, includingespresso coffee,12,23,24 were in similar percentages inall A20:R80 espresso coffees. However, surprisingly,more judges perceived malty/cereal flavours in 6.5 g40 mL−1 A20:R80 espresso coffees whereas in 7.5 and

particularly 8.5 g 40 mL−1 A20:R80 espresso coffees,negative flavours (burnt/roasty, acrid, fermented andearthy/musty) were better perceived. These differencesbetween instrumental and sensory analyses, andamong coffee/water ratios could be explained by thewell-known greater sensitivity of judges, and by themasking effect over positive flavours of some potentodorants, such as pyrazines23 and others that stillremain unidentified, which could be extracted inhigher amounts as the coffee/water ratio increased.

A similar pattern in the analysis of volatilecompounds and sensory flavour profile was observedin both Robusta blends espresso coffees. Nevertheless,in torrefacto roast coffees, less positive flavour notesrelated to freshness were perceived, maybe due tothe ability of melanoidins to bind specific volatilecompounds.25 Furthermore, an acrid flavour wassignificantly perceived by higher percentage of judges

J Sci Food Agric 87:586–592 (2007) 591DOI: 10.1002/jsfa

S Andueza et al.

in torrefacto espresso coffees prepared with morethan 6.5 g 40 mL−1 coffee/water ratio. This latterobservation, combined with an intensity of bitternesshigher than 7.5, led us to the selection of 6.5 g 40 mL−1

coffee/water ratio as the most suitable in A20:R80,50% torrefacto coffee.

In conclusion, the previous optimisation of othertechnical parameters such as water extraction, pressureand temperature, and grinding3–5 seemed to minimisethe influence of the increase of coffee/water ratio onthe extraction of soluble and solid compounds. Infact, only some sensory attributes, such as bitterness,astringency and burnt, acrid and earthy/musty flavourswere proposed as relevant to the selection of 6.5 g40 mL−1 or 7.5 g 40 mL−1 in conventional roastedcoffees (A100 and A20:R80), and 6.5 g 40 mL−1 intorrefacto roasted coffees (A20:R80, 50% torrefacto).On the other hand, the addition of sugar during theroasting process in torrefacto roast coffees seemedto contribute to a higher generation of acids,melanoidins and other compounds by the Maillardreaction or caramelisation which led us to select thelowest coffee/water ratio. Further investigations insoluble compounds related to taste and in volatilesrelated to aroma profiles, and their relationships, arerequired because, so far, sensory analysis rather thaninstrumental analysis seems to be more sensitive anddefinitive for evaluating coffee quality.

ACKNOWLEDGEMENTSWe thank the panel of judges, without whom thisstudy could not have been carried out. We thank theComision Interministerial de Ciencia y Tecnologıaproject (ALI-1999-0319) for their contribution tothe financial support of this work. We also thankthe Ministerio de Ciencia y Tecnologıa Espanol andGobierno Vasco for the grants given to S. Andueza(PN99-72681468) and M. Vila, respectively.

REFERENCES1 Petracco M. The cup, in Espresso Coffee: The Science of

Quality, ed. by Illy A and Viani R. Academic Press, London,pp. 290–315 (2005).

2 Petracco M. Percolation, in Espresso Coffee: The Science ofQuality, ed. by Illy A and Viani R. Academic Press, London,pp. 259–289 (2005).

3 Andueza S, Maeztu L, Dean B, de Pena MP, Bello J and Cid C,Influence of water pressure on the final quality of arabicaespresso coffee. Application of multivariate analysis. J AgricFood Chem 50:7426–7431 (2002).

4 Andueza S, Maeztu L, Pascual L, Ibanez C, de Pena MP andCid C, Influence of extraction temperature on the final qualityof espresso coffee. J Sci Food Agric 83:240–248 (2003).

5 Andueza S, de Pena MP and Cid C, Chemical and sensorialcharacteristics of espresso coffee as affected by grinding andtorrefacto roast. J Agric Food Chem 51:7034–7039 (2003).

6 Lingle TR, The Coffee Brewing Handbook. A Systematic Guide toCoffee Preparation. Speciality Coffee Association of America,Long Beach, California (1996).

7 Spiro M and Selwood RM, The kinetics and mechanism ofcaffeine infusion from coffee: the effect of particle size. J SciFood Agric 35:915–924 (1984).

8 Cammenga HK, Eggers R, Hinz T, Steer A and Waldmann C,Extraction in coffee-processing and brewing, in 17th IntColoq Chem Coffee, ASIC (International Association on CoffeeScience), Nairobi, pp. 219–226 (1997).

9 Maeztu L, Andueza S, Ibanez C, de Pena MP and Cid C, Amultivariate method for differentiation of espresso coffeesfrom different botanical varieties and types of roast byfoam, taste and mouthfeel characteristics. J Agric Food Chem49:4743–4747 (2001).

10 Bichi CP, Binello AE, Pellegrino GM and Vanni AC, Charac-terisation of green and roasted coffees through the chlorogenicacid fraction by HPLC-UV and principal component analysis.J Agric Food Chem 45:3238–3243 (1997).

11 Sanz C, Ansorena D, Bello J and Cid C, Optimizing headspacetemperature and time sampling for identification of volatilecompounds in ground roasted Arabica coffee. J Agric FoodChem 49:1364–1369 (2001).

12 Maeztu L, Sanz C, Andueza S, de Pena MP, Bello J and Cid C,Characterisation of espresso coffee aroma by HS-GC-MSand sensory flavor profile. J Agric Food Chem 49:5437–5444(2001).

13 Tranchant J, Manuel pratique de chromatographie en phase gazeuse.Masson, Paris (1982).

14 Kondjoyan N and Berdague JL, A Compilation of RelativeRetention Indices for the Analysis of Aromatic Compounds. Theix,France (1996).

15 AENOR, Analisis sensorial. Tomo 1. Alimentacion. Recopilacionde normas UNE. AENOR (Asociacion Espanola de Normal-izacion y Certificacion), Madrid (1997).

16 Baltzer HH, Acids in coffee, in Coffee: Recent Developments, ed.by Clarke RJ and Vitzthum OG. Blackwell Science, Oxford,pp. 18–32 (2001).

17 Engelhardt UH and Maier HG, The acids of coffee. 12. Thecontribution of individual acids to the sour taste. Z LebensmUnters-Forsch 181:20–23 (1985).

18 Bonnlander B, Eggers R, Engelhardt UH and Maier HG,Roasting, in Espresso Coffee: The Science of Quality, ed. byIlly A and Viani R. Academic Press, London, pp. 179–214(2005).

19 Macrae R, Nitrogenous compounds, in Coffee Chemistry,Volume 1., ed. by Macrae R. Elsevier Applied Science,London, pp. 115–151 (1985).

20 Andueza S, Influence of technological variables on espressocoffee quality. Antioxidant and pro-oxidant capacity of coffee.Doctoral thesis, University of Navarra. (2003).

21 Semmelroch P and Grosch W, Analysis of roasted coffeepowders and brews by gas chromatography – olfactometryof headspace samples. Lebensm-Wiss-Technol 28:310–313(1995).

22 Semmelroch P and Grosch W, Studies on character impactodorants of coffee brews. J Agric Food Chem 44:537–543(1996).

23 Blank I, Sen A and Grosch W, Aroma impact compoundsof arabica and robusta coffee. Qualitative and quantitativeinvestigations. 14th Int Coloq Chem Coffee, ASIC (InternationalAssociation on Coffee Science), San Francisco, California,pp. 117–129 (1991).

24 Holscher W, Vitzthum OG and Steinhart H, Identification andsensorial evaluation of aroma impact compounds in roastedColombian coffee. Cafe, Cacao, The 34:205–212 (1990).

25 Hofmann T, Czerny M, Calligaris S and Schieberle P, Modelstudies on the influence of coffee melanoidins on flavourvolatiles of coffee beverages. J Agric Food Chem 49:2382–2386(2001).

592 J Sci Food Agric 87:586–592 (2007)DOI: 10.1002/jsfa