Food Chemistry 113 (2009) 208–215

Contents lists available at ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Matrix effects on flavour volatiles release in dark chocolates varying in particle

size distribution and fat content using GC–mass spectrometry and GC–olfactometry
Emmanuel Ohene Afoakwa a,b,*, Alistair Paterson a, Mark Fowler b, Angela Ryan b
a
Centre for Food Quality, SIPBS Royal College Building, University of Strathclyde, 204 George Street, Glasgow G1 1XW, Scotland, UK
b
Nestlé Product Technology Centre York, P.O. Box 204, Haxby Road, York YO91 1XY, England, UK

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

Article history: Influences of matrix particle size distribution (PSD) (18, 25, 35 and 50 lm) and fat content (25%, 30% and
Received 26 April 2008 35%) on flavour release of dark chocolate volatiles were quantified by static headspace gas chromatogra-
Received in revised form 10 June 2008 phy using GC–MS. Sixty-eight (68) flavour compounds were identified, comprising alcohols, aldehydes,
Accepted 22 July 2008
esters, ketones, furans, pyrans, pyrazines, pyridines, pyroles, phenols, pyrones and thiozoles. From GC–
olfactometry, 2-methylpropanal, 2-methylbutanal and 3-methylbutanal had chocolate notes. With
cocoa/roasted/nutty notes were trimethyl-, tetramethyl-, 2,3-dimethyl-, 2,5-dimethyl-, 3(or 2),5-
Keywords:
dimethyl-2(or 3)-ethyl- and 3,5(or 6)-diethyl-2-methylpyrazine and furfuralpyrrole. Compounds with
Chocolate
Cocoa
fruity/floral notes included 3,7-dimethyl-1,6-octadien-3-ol and 5-ethenyltetrahydro-R,R,5-trimethyl-
Flavour release cis-2-furanmethanol. Caramel-like, sweet and honey notes were conferred by 2-phenylethanol, phenyl-
Pyrazines acetaldehyde, 2-phenylethylacetate, 2,3,5-trimethyl-6-ethylpyrazine, 2-carboxaldehyde-1H-pyrrole,
Acetic acid furancarboxaldehyde, furfuryl alcohol and 2,5-dimethyl-4-hydroxy-3(2H)furanone. There were direct
GC–MS relationships of fat content with 3-methylbutanal and branched pyrazines but inverse ones with 2-phen-
GC–olfactometry ylethanol, furfuryl alcohol, methylpyrazine, phenylacetaldehyde, 2, 3, 5-trimethyl-6-ethylpyrazine and 2-
carboxaldehyde-1-H-pyrrole. Particle size influenced higher alcohol, aldehyde, ester, ketone and pyrazine
concentrations at all fat contents. A multivariate product space suggested flavour effects of the interact-
ing factors.
Ó 2008 Published by Elsevier Ltd.

1. Introduction Fowler, 2007; Afoakwa, Paterson, & Fowler, 2008b; Do, Hargreaves,
Wolf, Hort, & Mitchell, 2007; Ziegler & Hogg, 1999; Ziegler, Mongia,
Flavour is central to acceptability in chocolate and is influenced, & Hollender, 2001). There are several studies of precursors for fla-
not only by volatile aroma compounds, but also by non-volatiles vour formation in cocoa and chocolate (Counet, Ouwerx, Rosoux, &
and behaviour of the continuous fat phase, influencing release of Collin, 2004; Kyi et al., 2005; Misnawi, Jinap, Jamilah, & Nazamid,
volatiles into the mouth headspace and taste perception. Precursor 2003).
composition depends on bean genotype and environmental effects, An appropriate cocoa bean composition can be converted,
particularly on contents of storage proteins and polyphenols (Afoa- through controlled post-harvest treatments and subsequent pro-
kwa, Paterson, Fowler, & Ryan, 2008a; Kim & Keeney, 1984; Sch- cessing technologies, to a high quality chocolate flavour character
wan & Wheals, 2004). Cocoa beans are rich in antioxidants, (Clapperton, 1994). Fermentation is crucial, not only to the forma-
including catechins, epicatechin and procyanidins, polyphenols tion of key volatile fractions (alcohols, esters and fatty acids) but
similar to those found in wine and tea (Carnesecchia et al., 2002; also provision of Maillard flavour precursors (amino acids and
Grassi, Lippi, Necozione, Desideri, & Ferri, 2005; Hatano et al., reducing sugars) (Buyukpamukcu et al., 2001; Kyi et al., 2005;
2002; Hermann et al., 2006; Lamuela-Raventos, Romero-Perez, An- Luna, Crouzillat, Cirou, & Bucheli, 2002). Drying reduces levels of
dres-Lacueva, & Tornero, 2005). Chocolate manufacture involves acidity and astringency in cocoa nibs, decreasing volatile acids
complex physical and chemical processes, determining rheological and total polyphenols. Maillard reactions during roasting convert
characteristics, flavour development, melting properties and ulti- these flavour precursors into two main classes of flavour-active
mately sensory perceptions of character (Afoakwa, Paterson, & component – pyrazines and aldehydes (Dimick & Hoskin, 1999;
Gill, MacLeod, & Moreau, 1984; Granvogl, Bugan, & Schieberle,
2006; Oberparlaiter & Ziegleder, 1997; Ramli, Hassan, Said, Samsu-
* Corresponding author. Address: Centre for Food Quality, SIPBS Royal College
Building, University of Strathclyde, 204 George Street, Glasgow G1 1XW, Scotland,
din, & Idris, 2006; Stark, Bareuther, & Hofmann, 2005). Flavour
UK. Tel.: +44 (0) 7984288727. development continues during conching, following the elimination
E-mail address: [email protected] (E.O. Afoakwa). of volatile acids and moisture, with associated viscosity changes,

0308-8146/$ - see front matter Ó 2008 Published by Elsevier Ltd.

doi:10.1016/j.foodchem.2008.07.088
 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215 209

due to emulsification and tannin oxidation (Mermet, Cros, & analysis, during refining, to ensure D90 values. Refined chocolate
Georges, 1992; Plumas, Hashim, & Chaveron, 1996; Reineccius, flakes were placed in plastic containers and conditioned at 50–
2006). Afoakwa et al. (2008a) reviewed relationships between ini- 55 °C for 24 h to ensure melting of fat prior to conching in a Lipp
tial composition and post-harvest treatments of cocoa beans and Conche (Model IMC-E10, Boveristr 40–42, D-68309, Mannhein,
subsequent processing (roasting and conching) and technological Germany) at low speed for 3.5 h at 60 °C. Lecithin and cocoa butter
effects on final flavour character in chocolate. were added and the mixtures conched at high speed for 30 min to
Particle size distribution influences dark chocolate structure – effect adequate mixing and liquefaction. Samples were stored in
specifically inter-particle interactions and network microstructure, sealed plastic containers at ambient temperature (20–22 °C) and
rheology and texture. Specific surface area and mean particle size moisture and fat contents determined using Karl Fischer and Soxh-
influence yield stress, plastic viscosity, product spread and hard- let methods (ICA, 1988; ICA, 1990), respectively.
ness (Afoakwa et al., 2008b; Afoakwa, Paterson, Fowler, & Vieira,
2008c; Beckett, 2008; Chevalley, 1999). Genovese, Lozano, and 2.2. Tempering procedure
Rao (2007) suggested that non-hydrodynamic parameters, such
as particle shape, particle size and size distribution, particle defor- Samples were melted at 50 °C for 4 h and tempered using a con-
mability and liquid polarity, influence food structure and flow tinuous three-stage tempering unit (Model AMK 10, Aasted Mikr-
behaviours. Such factors dictate the space dimension of a suspen- overk A/S, Farum, Denmark), pumping chocolate through multi-
sion, whether strongly or weakly flocculated, with influence on stage units with a worm screw, driving product through heat
yield stress and plastic viscosity. Although key flavour compounds exchangers. Sensors in equipment measured temperature of both
of milk and dark chocolates have been reported (Cerny & Fay, chocolate and coolant fluid at each stage. Based on our previous
1995; Counet, Callemien, Ouwerx, & Collin, 2002; Reineccius, work on modelling temperature controls to study tempering
2006; Schieberle & Pfnuer, 1999; Schnermann & Schieberle, behaviour (Afoakwa, Paterson, Fowler, & Vieira, 2008d), the tem-
1997; Taylor, 2002; Taylor & Roberts, 2004), their abundancy, re- peratures of each of the coolant fluids were thus set and controlled
lease and contribution to product character and matrix effects re- independently to obtain a final chocolate at 27 °C to promote
main unclear. crystal growth of the desired triacylglyceride fractions. Pre-crystal-
Modern healthier foods (less fat and low sugar products) re- lization was measured with a computerised tempermeter (Exo-
quire modifications in ingredients and recipe formulation with im- therm 7400, Systech Analytics, SA, Switzerland), using a built-in
pacts on flavour release and product rheology, structure and algorithm to ensure an optimal temper regime of slope 0 ± 0.3
texture. Knowledge of how variations in PSD and continuous phase (Afoakwa et al., 2008d; Nelson, 1999). Tempered chocolate was
fat content would influence flavour is useful for product develop- formed using plastic moulds, 80 by 20 by 8 mm, allowed to cool
ment and manufacture. The objectives of this study were to char- at 10 ± 2 °C for 2 h before de-moulding onto plastic trays and con-
acterise and quantify volatile flavour constituents in dark ditioned at 20 ± 2 °C for 14 days before analysis.
chocolates, and to evaluate matrix effects from varying PSD and
fat content on release of flavour volatiles using headspace HRGC, 2.3. Determination of particle size distribution
identifying components with GC–MS and flavour notes by GC–
olfactometry. A MasterSizerÒ Laser Diffraction Particle Size Analyzer equipped
with MS 15 Sample Presentation Unit (refractive index 1.590)
2. Materials and methods (Malvern Instrument Ltd., Malvern, England) was used. About
0.2 g of refined dark chocolate was dispersed in vegetable oil
2.1. Materials (refractive index 1.450) at ambient (20 ± 2 °C) until an obscuration
of 0.2 was obtained. Samples were placed under ultrasonic disper-
Cocoa liquor of central west African origin was obtained from sion for 2 min to ensure that particles were independently dis-
Cargill Cocoa Processing Company (York, UK), sucrose (pure cane persed and suspensions thereafter maintained by stirring. Size
extra fine) from British Sugar Company (Peterborough, UK) and distribution was quantified as the relative volume of particles in
pure prime pressed cocoa butter and soy lecithin from ADM Cocoa size bands presented as size distribution curves (Malvern Master-
Limited (Koog aan de Zaan, Netherlands) and Unitechem Company SizerÒ Micro Software v 2.19). PSD parameters obtained included
Ltd. (Tianjin, China), respectively. specific surface area, largest particle size (D90), mean particle vol-
Recipe (Table 1) and sample formulations have been described ume (D50), smallest particle size (D10) and Sauter mean diameter
previously (Afoakwa et al., 2008b). Chocolates were formulated (D[3,2]).
with total fat of 25–35% (w/w) from cocoa liquor and cocoa butter
with >34% total cocoa: composition as specified for dark chocolate 2.4. Quantitation of flavour volatiles by gas chromatography
(Codex Revised Standard, 2003; European Commission Directive,
2000). Sucrose and cocoa liquor (5 kg per formulation) were mixed Static headspace isolation of volatile compounds was per-
in a Crypto Peerless Mixer (Model K175, Crypto Peerless Ltd., Bir- formed using solid phase micro extraction (SPME) for 30 min at
mingham, UK) at low speed for 2 min and then at high for 3 min 55 °C onto a polydimethylsiloxane-divinylbenzene, 65 lm fibre
then, using a 3-roll refiner (Model SDX 600, Buhler Ltd., CH-9240 (Supelco, Bellafonte, PA, USA). Chocolate (4 g) was previously
Uzwil, Switzerland) to a specified particle size (D90:18 ± 1 lm, heated to 55 °C and intermittently stirred for 60 min for headspace
25 ± 1 lm, 35 ± 1 lm and 50 ± 1 lm), conducting particle size equilibration. Each experiment had a system control sample, made
by stirring an empty vial under the same conditions. Volatile com-
pounds were desorbed (5 min) into the splitless injector (220 °C) of
Table 1
an Agilent Technologies 6890N-5793 Network GC–MS system
Recipes used for the formulation of the dark chocolate
(Agilent Technologies, Santa Clara, CA, USA) and separated on a
Ingredient 25% Fat (% (w/w)) 30% Fat (% (w/w)) 35% Fat (% (w/w)) J&W 60 m DB-Wax capillary column (0.22 mm i.d., 0.25 lm film
Sucrose (%) 59.0 49.9 40.8 thickness). The temperature programme was: 5 min at 40 °C; 3 °C
Cocoa liquor (%) 35.5 44.6 53.7 min 1 to 230 °C; finally 15 min at 230 °C. Compounds were frag-
Cocoa butter (%) 5.0 5.0 5.0
mented using electron-impact ionisation (70 eV), with a source
Lecithin (%) 0.5 0.5 0.5
temperature of 200 °C, a scan range of 30–300 amu and a scan rate
210 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215

of 5 s 1. Components were identified based on comparison of mass and solids specific surface area were two key parameters in manu-
spectra with those of spectral libraries NIST 05 and Wiley 7N Reg- facture. The former determines chocolate coarseness and textural
istry of GC Mass Spectral Data, (John Wiley, New York, USA). character, the latter desirable flow properties. Specific surface area
was inversely correlated with the different component of PSD
2.5. Gas chromatography-olfactometry analytical conditions (Afoakwa et al., 2008b; Beckett, 1999; Sokmen & Gunes, 2006; Zie-
gler & Hogg, 1999). Fat contents were 25%, 30% and 35 ± 1% (each)
The GC–O analyses were conducted using an Agilent Technolo- and moisture was in the range 0.90–0.98%.
gies instrument (6890N Network Systems, CA, USA) with analyses
as before, diverting the effluent to a humidified sniffing port. Two
chromatographic runs were assessed by two trained assessors 3.2. Characterisation of flavour compounds in dark chocolates
(alternating for 20 min periods). Only matching descriptors for an
aroma attribute were retained. Criteria for selection of the key volatiles were presence in head-
spaces at >106 abundant units) quantified by GC–FID and also
2.6. Experimental design and statistical analysis detection and intensities by the GC–olfactometric techniques. In
all, 68 flavour compounds (Table 2), comprising nitrogen and oxy-
Two experimental variables, comprising PSD and fat contents, gen heterocycles, aldehydes and ketones, esters, alcohols, hydro-
were used with other variables, including refiner temperature carbons, nitriles and sulphides, were identified by GC–MS in dark
and pressure, conching time and temperature held constant. A chocolates. A typical chromatogram is shown in Fig. 2.
4  3 factorial experimental design was used with PSD (D90): 18, Compounds quantified included: 1-pentanol (1), 3-(methylthi-
25, 35 and 50 lm; fat was 25%, 30% and 35% (w/w). A Statgraphics ol)-propionaldehyde (12), methylbenzene (38), methylpyrazine
Plus 4.1 (Graphics Software System, STCC, Inc., Rockville, USA) (41), ethenylpyrazine (47), pyridine (55), 2-methylpyridine (56),
examined quantitative data, using two-way analysis of variance 1-(2-furanylmethyl)-1H-pyrrole (62), 1H-indole (63) and dimethyl
(ANOVA) and multiple range tests to determine effects of factors disulphide (67) (Table 1). Two others, benzyl alcohol (5) and dihy-
and interactions. Multivariate techniques, comprising principal dro-2-methyl-3(2H)-furanone (30) were only recently reported in
component analysis and multiple regression analysis, were used dark chocolates (Counet et al., 2002). Specific nitrogen heterocycles
to evaluate relationships between selected flavour volatiles ob- from Maillard reactions included: 3(or 2),5-dimethyl-2(or 3)-eth-
tained by quantification of GC–FID data and influential factors. Tu- ylpyrazine (50), 3,5-(or 6)-diethyl-2-methylpyrazine (53), 2,3-di-
key multiple comparisons at 95% significance level were conducted methyl-1H-pyrrole (59), 3-ethyl-2,5-dimethyl-1H-pyrrole (61)
to determine differences between factor levels. All process treat- and 10(2-furanylmethyl)-1H-pyrrole (furfurylpyrrole) (62) (Table
ments and analyses were conducted in duplicate and the mean val- 1). All had cocoa, praline, chocolate and roasted notes identified
ues reported. as important. The ethyl group in two pyrazine compounds suggests
key roles for alanine and/or its Strecker aldehyde, acetaldehyde, in
3. Results and discussion dark chocolate flavour (Cerny & Fay, 1995).
Flavour-active compounds identified as having strong chocolate
3.1. Particle size distribution of dark chocolates characters included 2-methylpropanal (8), 2-methylbutanal (9)
and 3-methylbutanal (10). Compounds derived from Maillard reac-
Particle size distributions (Fig. 1), previously reported (Afoakwa tions were 2,3-dimethylpyrazine (45), 2,5-dimethylpyrazine (42),
et al., 2008b), show volume histograms consisting of narrow 2,6-dimethylpyrazine (43), trimethylpyrazine (47), tetramethyl-
(18 lm PS) and wide (25 lm PS) bimodal and narrow (35 lm PS) pyrazine (51), 3(or 2),5-dimethyl-2(or 3)-ethylpyrazine (50),
and wide (50 lm PS) multimodal size distributions. This PSD range 3,5(or 6)-diethyl-2-methylpyrazine (53) and furfurylpyrrole (60),
18–50 lm, using D90 values (>90% finer), covers optimum mini- exhibiting cocoa/roasted/nutty/cooked notes. Counet et al. (2002)
mum and maximum sizes with direct effects on texture and sen- identified such flavour volatiles in dark chocolates after conching,
sory character in manufacture (Afoakwa et al., 2007; Beckett, suggesting that these are formed during cocoa processing.
2008; Ziegler & Hogg, 1999). Data from PSD showed variations in Volatiles, such as 2-phenylethanol (7), phenylacetaldehyde
specific surface area, mean particle volume D(v,50), Sauter mean (15), 2-phenylethylacetate (22), 2,3,5-trimethyl-6-ethylpyrazine
(D[3,2]) and mean particle diameter (D[4,3]) with increasing D90 (54) and 2-carboxaldehyde-1H-pyrrole (60), were characterised
particle sizes. Beckett (1999) concluded that largest particle size by sweet, candy and honey flavours. Furancarboxaldehyde (furfu-
ral) (31), furfuryl alcohol (furfurol) (32) and 2,5-dimethyl-4-hydro-
xy-3(2H)furanone (furaneol) (35) were also characterised by
caramel-like, sweet and honey notes, likely derivatives of Strecker
10 100 degradation and caramelization reactions developed during cocoa
90 processing and transformed during chocolate flavour synthesis in
80 conching (Afoakwa et al., 2008a; Cerny & Fay, 1995).
70 Eight heterocyclic compounds, including 2,3-dimethylpyrazine
Volume (%)

60 (45), 2,5-dimethylpyrazine (42), 2,6-dimethylpyrazine (43), trim-

ethylpyrazine (47), tetramethylpyrazine (51), 3(or 2),5-dimethyl-
50
2-(3)-ethylpyrazine (50), 3,5(or 6)-diethyl-2-methylpyrazine (53)
40
d and 2,3,5-trimethyl-6-ethylpyrazine (54) were identified (Table
c 30 1). Characteristic key chocolate flavours, such as fruity and floral,
b
20 likely derived from cocoa, were found in 3,7-dimethyl-1,6-octadi-
a
10 en-3-ol (linalool) (6) and 5-ethenyltetrahydro-R,R,5-trimethyl-cis-
0 0 2-furanmethanol (linalool oxide) (36). Ethyl cinnamate (23) and
0.1 1.0 10.0 100.0 1000.0
acetic acid (24), not previously reported important in dark
Particle Diameter (µm)
chocolates, were characterised by fruity-spicy and astringent-
Fig. 1. Particle size distribution in dark chocolate with D90 of (a) 18 lm (b) 25 lm vinegar notes, respectively. Tetramethylpyrazine (51), the most
(c) 35 lm and (d) 50 lm. abundant flavour compound in dark chocolate, exhibited milk
 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215 211

Table 2 Table 2 (continued)

Key flavour volatiles identified in dark chocolate
No. Flavour compound Odour descriptiona
No. Flavour compound Odour descriptiona
54 2,3,5-Trimethyl-6-ethylpyrazine Candy, sweet,
Alcohols
Pyridines
1 1-Pentanol
55 Pyridine
2 2,4-Hexadien-1-ol
56 2-Methylpyridine Caramel-like, sweet
3 3-Methyl pentanol
57 2-Pyridinamine
4 2-Heptanol
58 1-(2-Pyridinyl)-1-propanone
5 Benzyl alcohol
6 3,7-Dimethyl-1,6-octadien-3-ol (linalool) Flowery, floral, Pyrroles
fruity (low) 59 2,3-Dimethyl-1H-pyrrole Cocoa, praline, chocolate
7 2-Phenylethanol Caramel-like, 60 2-Carboxaldehyde-1H-pyrrole Honey, candy (low)
sweet, honey 61 3-Ethyl-2,5-dimethyl-1H-pyrrole Cocoa, coffee
62 1-(2-Furanylmethyl)-1H-pyrrole Cocoa, roasted (low)
Aldehydes
(furfurylpyrrole)
8 2-Methylpropanal (isobutanal) Chocolate
63 1H-Indole Chocolate, green (low)
9 2-Methylbutanal Chocolate
10 3-Methylbutanal Chocolate Phenols
11 2-Methyl-2-butenal 64 Phenol
12 3-(Methylthio)propionaldehyde (methional) Potato 65 4-Methylphenol
13 Heptanal Pyrones
14 Benzaldehyde Nutty 66 3-Hydroxy-2-methyl-4-pyrone (maltol)
15 Phenylacetaldehyde Flowery, sweet,
honey Sulphur compounds
16 Nonanal 67 Dimethyl disulphide Meaty (low)
17 2-Phenyl-2-butenal Cocoa, roasted Thiozoles
18 2-Phenyl-5-methyl-2-hexenal Roasted 68 4,5-Dihydro-2-methylthiazole
Esters a
Odour quality and intensity at GC–O outlet.
19 Ethylbenzoylformate
20 Ethylbenzoate
21 Ethyloctanoate
22 2-Phenylethylacetate Honey, sweet
23 Ethyl cinnamate Fruity, floral (low) coffee-roasted-cooked notes, and trimethylpyrazine (47) had co-
24 Acetate (acetic acid) Astringent, vinegar
coa-roasted-cooked characters (Table 2).
Ketones
25 2,3-Butanedione (diacetyl) Buttery (low)
3.3. Effects of particle size distribution (PSD) on flavour volatile release
26 2-Heptanone
27 4-Methylcyclohexanone
28 3-Hydroxy-2-butanone Effects of PSD and fat content on the release of selected abun-
29 3,4,4-Trimethyl-2-cyclopenten-1-one dant (>106 U) flavour volatiles, characterised by distinct aroma,
Furans were evaluated using SPME-HRGC with FID detection. Data from
30 Dihydro-2-methyl-3(2H)-furanone ANOVA indicated that, with the exception of 3,7-dimethyl-1,6-
31 Furancarboxaldehyde (furfural) Caramel-like, sweet
octadien-3-ol (linalool) and 2-carboxaldehyde-1-H-pyrrole
32 Furfuryl alcohol (furfurol) Caramel-like, sweet
33 1-(2-Furanyl)ethanone (acetylfuran)
(P = 0.965 and 0.854, respectively), increasing particle size (PS)
34 5-Methyl-2-furancarboxaldehyde caused significant reduction in the release of all selected com-
35 2,5-Dimethyl-4-hydroxy-3(2H)furanone (Furaneol) Caramel-like, sweet pounds measured in the sample headspace, with P < 0.001 for 3-
36 5-Ethenyltetrahydro-R,R,5-trimethyl-cis-2- Fruity, floral/ methylbutanal, 2-phenylethanol, furfuryl alcohol (furfurol), acetic
furanmethanol (linalool oxide) flowery (low)
acid, methylpyrazine, 2,3-dimethylpyrazine, 2,5-dimethylpyrazine,
37 3-Phenylfuran Cocoa, green, nutty
trimethylpyrazine, tetramethylpyrazine and 2,3,5-trimethyl-6-
Hydrocarbons
ethylpyrazine, and P < 0.05 for 2-phenylethylacetate, 2-methylbut-
38 Methylbenzene (toluene)
anal and 5-ethenyltetrahydro-R,R,5-trimethyl-cis-2-furanmethanol
Nitrogen compounds
(linalool oxide), with significant interactions noted with fat con-
39 Benzonitrile
tent (Table 5).
Pyrans
The decreasing flavour volatiles release with increasing PS could
40 3,4-Dihydro-8-hydroxy-3-methyl-1H-2-benzopyran-
1-one be related to increased matrix retention through structural, rheo-
logical and textural differences (Afoakwa et al., 2008b; Afoakwa
Pyrazines
41 Methylpyrazine Nutty, green et al., 2008c; Afoakwa, Paterson, Fowler, & Vieira, 2008e). Beckett
42 2,5-Dimethylpyrazine Roasted, cooked (2008) noted that movement of volatiles was related to an initial
43 2,6-Dimethylpyrazine Roasted, cooked concentration gradient between phases, and refining (degree of
44 Ethylpyrazine Nutty, roasted
particle sizes) in production may influence release during manu-
45 2,3-Dimethylpyrazine Cooked, nutty
46 2-Ethyl-5(or 6)-methylpyrazine Cocoa, roasted,
facture. Beckett (2008) also correlated compositional and sensory
green analyses for differences in flavour profile, with preference for lower
47 Trimethylpyrazine Cocoa, roasted, PS (thick and pasty) chocolates, rather than the higher PS (thin and
cooked runny). These results suggest that dark chocolates with finer PS (18
48 2-Ethyl-3-methylpyrazine Hazelnut, roasted
and 25 lm) would release more cocoa-chocolate-praline and cara-
49 2-Ethenyl-6-methylpyrazine Roasted, smoky
50 3(or 2),5-Dimethyl-2(or 3)-ethylpyrazine mel-like-sweet-honey notes than would those with larger PS (35
51 Tetramethylpyrazine Milk-coffee, and 50 lm), predicting perceived differences in flavour with vary-
roasted, ing PS. The increase in surface area with deceasing PS (D90) would
52 2,3-Dimethyl-5-ethylpyrazine Cocoa, chocolate
be predicted to facilitate volatiles release. The lack of significant ef-
53 3,5(or 6)-Dimethyl-2-ethylpyrazine Cocoa, praline,
chocolate
fects of 3,7-dimethyl-1,6-octadien-3-ol (linalool) and 2-carboxal-
dehyde-1-H-pyrrole (P = 0.965 and 0.854, respectively) would
not be expected to influence flavour character from headspace
212 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215

Fig. 2. Typical GC–MS chromatogram used to identify flavour volatiles.

contents and odour intensities (Tables 2–4). Voltz and Beckett ship with fat content at all PS (Tables 3 and 4). Volatiles release
(1997) and Ziegler et al. (2001) reported that finer (smaller PS) data suggested that chocolates of higher fat content would exhibit
chocolates tended to be sweeter in taste than coarser (larger PS) greater release of components with cocoa-chocolate-praline notes
ones, attributed to relative crystal sizes and melting behaviour. Par- than would those with lower fat. This decreased matrix retention,
ticle size influences perceptions of creaminess and flavour release in could be related to differences in (micro)structure as inter-particle
soft model systems (Engelen & Van der Bilt, 2008; Engelen et al., flocculation and aggregates are reduced with higher fat contents
2005; Kilcast & Clegg, 2002). Concentration of flavour volatiles in (Afoakwa et al., 2008c), releasing more Strecker degradation com-
headspaces has been reported as a function of diffusion in the solid pounds with cocoa-chocolate notes. Concentrations of less volatile
phase (Carr et al., 1996; Engelen et al., 2003; Guinard & Marty, heterocyclic compounds were increased, notably polysubstituted
1995; Kersiene, Adams, Dubra, De Kimpe, & Leskauskaite, 2008). ethyl- and isobutylpyrazines, tri- and tetramethylpyrazine and
furans (linalool oxide), suggesting structural and rheological
3.4. Effects of fat content on flavour volatile release effects as major determinants of chocolate character (Afoakwa
et al., 2008b; Afoakwa et al., 2008c; Afoakwa et al., 2008e;
Fat content influenced the headspace concentration of volatiles, Do et al., 2007).
independently of PSD (Table 3). ANOVA showed that 3,7-dimethyl- By contrast, volatiles with caramel-like, sweet, honey and candy
1,6-octadien-3-ol (linalool), 2-methylbutanal, 2-phenylethylace- notes included: 2-phenylethanol, furfuryl alcohol (furfurol), meth-
tate and 2-carboxaldehyde-1-H-pyrrole lacked significant effects ylpyrazine, phenylacetaldehyde, 2, 3, 5-trimethyl-6-ethylpyrazine
(P > 0.05). Fat content significantly influenced headspace concen- and 2-carboxaldehyde-1-H-pyrrole. All showed an inverse rela-
trations of all other quantified volatiles (P < 0.001) at all PSD with tionship with fat content at all PS (Tables 3 and 4), primarily due
significant interactions amongst factors studied (Table 5). Volatiles to lipophilic matrix–flavour interactions. The major influence of
characterised by cocoa, chocolate, praline, fruity and roasted notes fat content was observed with the most lipophilic compounds (Ta-
included: trimethypyrazine, 3-methylbutanal, 2,3-dimethylpyr- bles 3 and 4), particularly with fat contents above 25%. These re-
azine, 2,5-dimethylpyrazine, tetramethylpyrazine, linalool oxide sults are consistent with earlier reports (Doyen, Carey, Linforth,
and 2,3,5-triethyl-5-methylpyrazine. All showed a direct relation- Marin, & Taylor, 2001; Jo & Ahn, 1999) and are also consistent with

Table 3
Flavour volatiles in dark chocolates varying in PSD and fat contenta

Volatile compound (abundance  106) 25% 30% 35%

18 lm 25 lm 35 lm 50 lm 18 lm 25 lm 35 lm 50 lm 18 lm 25 lm 35 lm 50 lm
3,7-Dimethyl-1,6-octadien-3-ol (linalool) 1.93 1.74 1.24 1.03 2.07 1.72 1.37 9.99 2.66 1.91 1.85 1.26
2-Phenylethanol 46.8 32.4 26.2 25.2 27.8 24.5 22.3 22.0 26.4 24.0 22.0 17.1
2-Methylbutanal 2.64 2.24 2.02 2.01 2.92 2.24 2.14 1.99 3.04 2.23 2.25 1.99
3-Methylbutanal 8.12 8.17 7.24 7.11 8.82 8.26 7.46 7.16 9.42 9.19 8.86 7.14
Phenylacetaldehyde 20.9 12.8 10.2 10.0 9.78 9.44 7.99 7.86 6.70 6.35 6.37 3.99
2-Phenylethylacetate 11.9 8.75 7.09 6.82 6.74 6.08 5.86 5.56 6.14 6.01 5.80 5.32
Furfuryl alcohol (furfurol) 4.29 3.21 2.44 2.24 4.08 2.84 2.17 2.05 2.89 2.67 1.99 1.82
5-Ethenyltetrahydro-R,R,5-trimethyl-cis-2-furanmethanol 5.58 5.22 4.63 4.60 5.45 5.05 4.73 4.52 5.87 5.32 5.16 5.01
(linalool oxide)
2-Carboxaldehyde-1-H-pyrrole 2.74 1.53 1.02 1.02 1.21 0.89 0.75 0.68 1.01 0.68 0.54 0.40
Acetic acid 130 78.6 58.8 56.1 112 65.3 48.8 42.0 30.8 28.7 28.1 27.2
a
Quantification was by GC–MS, expressed as mean peak area.
 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215 213

Table 4
Abundant pyrazines in dark chocolates varying in PSD and fat contenta

Volatile compound (abundance  106) 25% 30% 35%

18 lm 25 lm 35 lm 50 lm 18 lm 25 lm 35 lm 50 lm 18 lm 25 lm 35 lm 50 lm
Methylpyrazine 5.48 5.08 3.78 2.92 4.25 3.30 2.94 2.60 3.51 3.09 2.78 2.46
2,3-Dimethylpyrazine 5.77 6.20 6.16 5.57 6.56 5.88 5.49 5.39 6.89 7.10 7.02 5.46
2,5-Dimethylpyrazine 9.27 9.04 8.43 8.43 10.09 8.95 7.43 6.60 10.24 9.65 9.43 8.88
Trimethylpyrazine 28.6 28.9 28.6 28.5 29.5 29.1 28.8 28.8 33.1 32.8 32.1 30.4
Tetramethylpyrazine 110 105 96.7 96.4 98.9 96.9 96.8 96.1 113 107 107 96.9
2,3-Diethyl-5-methylpyrazine 4.66 4.59 4.29 3.79 4.80 4.54 4.51 3.89 4.89 4.64 4.59 3.89
2,3,5-Trimethyl-6-ethylpyrazine 3.86 3.05 2.68 1.93 2.25 2.11 1.82 1.78 1.89 1.90 1.78 1.56
a
Quantification was by GC–MS, expressed as mean peak area.

Table 5 3.9
ANOVA summary, showing F-values and regression coefficients of flavour compounds B: caramel-like, sweet, honey
identified in dark chocolates with varying PSD and fat content
C: influential factors Phenylacetaldehyde

Component 2 (26.5%)
Volatile compound PSD Fat: B Interactions: R2a 1.9 2,3,5-trimethyl-6-ethylpyrazine
(D90): A AB Acetic acid
PSD
Methylpyrazine
3,7-Dimethyl-1,6-octadien-3- 1.81 2.64 1.03 7.11 2-phenylethanol
ol (linalool) Furfuryl alcohol
2-Phenylethanol 1305.56*** 1906.11*** 265.68*** 75.21*** -0.1 2-carboxaldehyde-1-H-pyrrole
2-Methylbutanal 5.83* 0.48 4.76* 21.2
3-Methylbutanal 32.36*** 20.79*** 27.92*** 84.3*** Linalooloxide
2, 3-diethyl-5-methylpyrazine
Phenylacetaldehyde 8.62*** 29.46*** 10.28** 81.8*** Tetramethylpyrazine
2-Phenylethylacetate 3.23* 1.67 3.28* 9.47 -2.1 2, 5-dimethylpyrazine
Furfuryl alcohol (furfurol) 70.57*** 19.16*** 27.82*** 87.7*** Fat 2, 3-dimethylpyrazine
3-methylbutanal
5-Ethenyltetrahydro-R,R,5- 4.89* 5.34* 5.71* 7.86 Trimethylpyrazine
trimethyl-cis-2- A: chocolate, cocoa, cooked, roasted
furanmethanol (linalool -4.1
oxide) -4.1 -2.1 -0.1 1.9 3.9 5.9
2-Carboxaldehyde-1-H-pyrrole 1.15 1.77 0.95 17.9
Acetic acid 13.67*** 26.31*** 12.62*** 75.0*** Component 1 (65.2%)
Methylpyrazine 30.15*** 34.81*** 28.63*** 86.4***
2,3-Dimethylpyrazine 8.93*** 11.26*** 10.56** 51.6* Fig. 3. PCA biplots of dark chocolate flavour volatiles as influenced by PSD and fat
2,5-Dimethylpyrazine 15.62*** 12.32*** 18.63*** 61.4* content.
Trimethylpyrazine 13.01*** 795.09*** 18.52*** 81.9***
Tetramethylpyrazine 13.68*** 17.20*** 12.29*** 51.0*
may be related to higher plastic viscosity and yield values (Afoa-
2,3-Diethyl-5-methylpyrazine 312.88*** 19.24*** 48.78** 86.9***
2,3,5-Trimethyl-6- 9.67*** 29.59*** 13.36** 76.6*** kwa et al., 2008b), and greater flocculation and aggregation of in-
ethylpyrazine ter-particle network structure (Afoakwa et al., 2008c), influencing
release and volatilisation in conching. High acetic acid levels in
Significant F-ratios at *P 6 0.05, **P 6 0.01, ***
P 6 0.001.
a
R-squares from multiple regression.
low (25%) fat chocolates may reduce acceptability scores: effective
elimination of volatile free fatty acids (e.g. acetic acid) and mois-
ture during conching is crucial for development of final flavour
the suggestion that the more lipophilic the volatile, the less lipid is character and texture in chocolates (Beckett, 2008; Kealey et al.,
needed to reduce its headspace concentration (Roberts, Pollien, & 2001; Mermet et al., 1992; Plumas et al., 1996; Pontillon, 1995).
Watzke, 2003). More lipid generally reduces volatility of lipophilic As demand for healthier (low fat) chocolate has increased in recent
components such as long-chain aldehydes and esters (Kersiene years, good process optimisation to effect adequate release of ace-
et al., 2008). Lack of significant effect on overall flavour character tic acid during manufacture of low (25%) fat dark chocolates
from 3,7-dimethyl-1,6-octadien-3-ol (linalool), 2-methylbutanal, would be necessary to obtain well-balanced flavour characters.
2-phenylethylacetate and 2-carboxaldehyde-1-H-pyrrole would
be predicted (Tables 2 and 3). Studies from emulsions showed that 3.5. Relating flavour volatiles release to PSD and fat content: product
release of lipophilic compounds is decreased with limited amounts spaces
of lipid (Carey et al., 2002; Roberts et al., 2003). Factors, such as
lipophilicity or hydrophobicity of compounds could modulate the Multivariate principal component analysis (PCA) generated a
effect of fat content on release, specifically in confectionery (Bar- product space exploring influence of PSD and fat content on head-
ylko-Pikielna & Szczesniak, 1994; Hyvönen et al., 2003), as well space volatiles data of dark chocolates. The PCA space (Fig. 3) ex-
as mouth-feel (De Wijk, Terpstra, Janssen, & Prinz, 2006) and ther- plained >91% variance in two factors, and showed two flavour
mal perceptions (Engelen et al., 2002). volatiles clusters with loadings for PSD and fat content as influen-
A further key finding was related to headspace acetic acid con- tial factors. Fat content had polar influences on PC1 (65.2% vari-
tents with high value for products with 25% and 30% fat at lower ance) score while particle size had marginal influence on PC2
(18 and 25 lm) PS, inversely related to fat content. Greater reduc- (25.6% variance) score. The PCA loading showed distinct relation-
tion in acetic acid (4-fold) was noted with 35% fat at all PS, than ships. Two components were extracted with eigenvalues P1, and
with 25% and 30% fat (Table 3). Similarly, increasing PS from 25– volatiles segregated into two groups, labelled A and B. Group A vol-
50 lm reduced contents by 2 to 3-fold with 25% and 30% fat, atiles were: trimethypyrazine, 3-methylbutanal, 2,3-dimethylpyr-
whereas only minimal (5%) reductions were noted with 35% fat. azine, 2,5-dimethylpyrazine, tetramethylpyrazine, linalool oxide
From ANOVA there were highly significant effects of PSD and fat and 2,3,5-triethyl-5-methylpyrazine, all characterised by cocoa,
content (P = 0.001) on acetic acid release, with significant interac- chocolate, praline and roasted notes, possibly originating in cocoa.
tions (Table 5). Acetic acid in 25% and 30% fat chocolate headspaces Group B consisted of: 2-phenylethanol, furfuryl alcohol (furfurol),
214 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215

acetic acid, methylpyrazine, phenylacetaldehyde, 2, 3, 5-trimethyl- anal, 2,3-dimethylpyrazine, 2,5-dimethylpyrazine, tetramethylpyr-

6-ethylpyrazine and 2-carboxaldehyde-1-H-pyrrole, characterised azine, linalool oxide and 2,3,5-triethyl-5-methylpyrazine, at all
by caramel-like, sweet, honey and candy notes, developed during particle size distributions. By contrast, there was an inverse rela-
chocolate manufacture. PC1, and to some extent PC2, differentiated tionship between matrix fat content and headspace concentrations
chocolate samples, with groupings clearly characterised by specific of 2-phenylethanol, furfuryl alcohol (furfurol), methylpyrazine,
flavour notes from GC olfactometry (Table 2). phenylacetaldehyde, 2, 3, 5-trimethyl-6-ethylpyrazine and 2-car-
The PCA loadings on PC1 showed that flavour volatiles within boxaldehyde-1-H-pyrrole, likely due to lipophilic matrix–flavour
Group A, characterised by cocoa, chocolate, praline and roasted interactions.
notes, were highly related in abundancy to fat content. Our work- The PCA and regression models predicted contribution of vola-
ing hypothesis is that Group A flavour volatiles are primary origins tiles to overall flavour character. One Strecker aldehyde, 3-methyl-
of cocoa and chocolate notes in dark chocolates, with trimethylpyr- butanal, and two nitrogen heterocycles derived from Maillard
azine and 3-methylbutanal central to these characters (Fig. 3). By reactions, trimethylpyrazine and 2,3-diethyl-5-methylpyrazine,
contrast, the PCA loading showed a polar relationship of fat content provided cocoa, praline-chocolate and nutty notes, with three oth-
with Group B, observed to have caramel-like, sweet, honey and ers, 2,3-dimethylpyrazine, 2,5-dimethylpyrazine and tetramethyl-
candy notes (Table 2), suggesting that increasing fat content re- pyrazine, likely making little contribution to cocoa-chocolate
duces the influence of such notes on flavour character in dark flavour. Ethyl groups in pyrazine compounds suggest the key role
chocolates. of alanine and its Strecker aldehyde and acetaldehyde in dark choc-
Regression models were developed to predict contribution of olate flavour formation. Others, 2-methylbutanal, 5-ethenyltetra-
specific flavour volatiles to overall flavour character. One Strecker hydro-R,R,5-trimethyl-cis-2-furanmethanol (linalool oxide) and
aldehyde and two nitrogen heterocycles derived from Maillard 3,7-dimethyl-1,6-octadien-3-ol (linalool), probably had no effect
reactions had high regression coefficients, respectively, 3-methyl- on dark chocolate flavour character. Likewise, 2-phenylethanol,
butanal, R2 = 0.843, P = 0.001, trimethylpyrazine, R2 = 0.819, furfuryl alcohol (furfurol), methylpyrazine, phenylacetaldehyde
P = 0.001 and 2,3-diethyl-5-methylpyrazines, R2 = 0.869, P = 0.001 and 2, 3, 5-trimethyl-6-ethylpyrazine emerged as compounds con-
(Table 4). These emerged as probably the most interesting tributing caramel-like, sweet, honey and candy notes, with acetic
compounds in dark chocolates, providing cocoa, praline-chocolate acid contributing to acid-sour sensations. Matrix effects on flavour
and nutty flavours. Three other heterocycles, 2,3-dimethylpyr- release in dark chocolate merit attention as new product develop-
azine, 2,5-dimethylpyrazine and tetramethylpyrazine, showed ment and consumers demand a wider range of origins and defined
less but still significant effects with R2 = 0.516, 0.614 and products, and their influence on sensory effects with PSD and fat
0.510 (P < 0.05), respectively, contributing cocoa-chocolate notes. content remains unclear.
Others, 2-methylbutanal, 5-ethenyltetrahydro-R,R,5-trimethyl-
cis-2-furan-methanol (linalool oxide) and 3,7-dimethyl-1,6-octadi- Acknowledgements
en-3-ol (linalool), had no significant influence (P > 0.05) (Table 4),
possibly due to low contents in the central west African cocoa This study was co-funded by the Government of Ghana and
(Tables 3 and 4). Nestlé Product Technology Centre (York, UK): sponsorship is
On the other hand, the regression models developed showed gratefully acknowledged. The authors thank Drs. Steve Beckett,
high and significant regression coefficients (R2 = 0.75 0.88, Joselio Vieira and Angel Manez (Nestlé PTC, York) for useful
P = 0.001) for Group B compounds. These predict likely contribu- discussions.
tions of 2-phenylethanol, furfuryl alcohol (furfurol), methylpyra-
zine, phenylacetaldehyde and 2, 3, 5-trimethyl-6-ethylpyrazine of References
caramel-like, sweet, honey and candy notes. Acetic acid had high
regression coefficient (R2 = 0.75, P = 0.001), and likely contributed Afoakwa, E. O., Paterson, A., & Fowler, M. (2007). Factor influencing rheological and
textural qualities in chocolate – a review. Trends in Food Science and Technology,
astringent-sour characters to dark chocolates. Others, 2-phenylac- 18, 290–298.
etate and 2-carboxaldehyde-1-H-pyrrole, showed no significant ef- Afoakwa, E. O., Paterson, A., Fowler, M., & Ryan, A. (2008a). Flavor formation and
fect (R2 = 0.095 and 0.179, P > 0.05, respectively), indicating character in cocoa and chocolate: A critical review. Critical Reviews in Food
Science and Nutrition, 48, 1–18.
minimal impacts on flavour character in dark chocolate. This prod-
Afoakwa, E. O., Paterson, A., & Fowler, M. (2008b). Effects of particle size distribution
uct space from PCA (Fig. 3), demonstrated the importance and rela- and composition on rheological properties of dark chocolate. European Food
tionships of the different flavour volatiles and likely effect on Research and Technology, 226, 1259–1268.
Afoakwa, E. O., Paterson, A., Fowler, M., & Vieira, J. (2008c). Microstructure and
flavour character and, furthermore, the influence of solids PSD
mechanical properties relating to particle size distribution and composition in
and continuous phase matrix fat content on the overall volatiles re- dark chocolate. International Journal of Food Science and Technology,
lease into the headspace in dark chocolates. doi:10.1111/j.1365-2621.2007.01677.x.
Afoakwa, E. O., Paterson, A., Fowler, M., & Vieira, J. (2008d). Modelling tempering
behaviour of dark chocolates from varying particle size distribution and fat
content using response surface methodology. Innovative Food Science and
4. Conclusion Emerging Technologies, published online, doi:10.1016/j.ifset.2008.02.002.
Afoakwa, E. O., Paterson, A., Fowler, M., & Vieira, J. (2008e). Particle size distribution
and compositional effects on textural properties and appearance of dark
Variations in flavour volatile release in dark chocolate matrices chocolates. Journal of Food Engineering, 87, 181–190.
varying in particle size distribution and fat content were noted, Barylko-Pikielna, N., & Szczesniak, A. S. (1994). Taste and flavour perception in
suggesting potential effects of matrix structure and lipophilic-fla- mayonnaise-type emulsion of various fat level. Poland Journal of Food &
Nutritional Sciences, 3, 57–70.
vour interactions. Increasing PS significantly reduced the release
Beckett, S. T. (1999). Industrial chocolate manufacture and use (3rd ed.). Oxford:
of 3-methylbutanal, 2-phenylethanol, furfuryl alcohol (furfurol), Blackwell Science. pp. 153–181, 201–230, 405–428, 460–465.
acetic acid, methylpyrazine, 2,3-dimethylpyrazine, 2,5-dimethyl- Beckett, S. T. (2008). The science of chocolate (2nd ed.). London, Cambridge, UK:
Royal Society of Chemistry.
pyrazine, trimethylpyrazine, tetramethylpyrazine and 2,3,5-tri-
Buyukpamukcu, E., Goodall, D. M., Hansen, C. E., Keely, B. J., Kochhar, S., & Wille, H.
methyl-6-ethylpyrazine, 2-phenylethylacetate, 2-methylbutanal (2001). Characterization of peptides formed during fermentation of cocoa bean.
and 5-ethenyltetrahydro-R,R,5-trimethyl-cis-2-furanmethanol Journal of Agricultural & Food Chemistry, 49, 5822–5827.
(linalool oxide). Fat content was directly related to headspace Carnesecchia, S., Schneidera, Y., Lazarusb, S. A., Coehloc, D., Gossea, F., & Raul, F.
(2002). Flavanols and procyanidins of cocoa and chocolate inhibit growth and
concentrations of compounds characterised by cocoa, chocolate, polyamine biosynthesis of human colonic cancer cells. Cancer Letters, 175,
praline, fruity and roasted notes: trimethypyrazine, 3-methylbut- 147–155.
 E.O. Afoakwa et al. / Food Chemistry 113 (2009) 208–215 215

Carey, M. E., Asquith, T., Linforth, R. S. T., & Taylor, A. J. (2002). Modeling the ICA (1990). Determination of fat content of cocoa and chocolate products. Analytical
partition of volatile aroma compounds from a cloud emulsion. Journal of Method 37. CAOBISCO, Brussels, Belgium.
Agricultural and Food Chemistry, 50, 1985–1990. Jo, C., & Ahn, D. U. (1999). Fat reduces volatiles production in oil emulsion system
Carr, J., Baloga, C., Guinard, X., Lawter, L., Marty, C., & Squire, C. (1996). The effect of analyzed by purge-and-trap dynamic headspace/gas chromatography. Journal of
gelling agent type and concentration on flavour release in model systems. In R. Food Science, 64, 641–643.
J. McGorrin & J. V. Leland (Eds.), Flavor–food interactions (pp. 98–108). Kealey, K. S., Snyder, R. M., Romanczyk, L. J., Geyer, H. M., Myers, M. E., Withcare, E.
Washington, DC: American Chemical Society. J., Hammerstone, J. F., and Schmitz, H. H. (2001). Cocoa components, edible
Cerny, C., & Fay, L. B. (1995). Mechanism of formation of alkylpyrazine in the products having enhanced polyphenol content, methods of making same and
Maillard reaction. Journal of Agricultural & Food Chemistry, 43, 2818–2822. medical uses. Patent Cooperation Treaty (PCT) WO 98/09533, Mars
Chevalley, J. (1999). Chocolate flow properties. In S. T. Beckett (Ed.), Industrial Incorporated, USA.
chocolate manufacture and use (pp. 182–199). New York: Chapman & Hall. Kersiene, M., Adams, A., Dubra, A., De Kimpe, N., & Leskauskaite, D. (2008).
Clapperton, J. F. (1994). A review of research to identify the origins of cocoa flavour Interactions between flavour release and rheological properties in model
characteristics. Cocoa Growers’ Bulletin, 48, 7–16. custard desserts: Effect of starch concentration and milk fat. Food Chemistry,
Codex Revised Standard (2003). Codex Alimentarius commission revised standard 108(2008), 1183–1191.
on cocoa products and chocolate. In Report of the 19th session of the codex Kilcast, D., & Clegg, S. (2002). Sensory perception of creaminess and its relationship
committee on cocoa products and chocolate (pp. 1–37). Alinorm 03/14. with food structure. Food Quality & Preference, 13(2002), 609–623.
Counet, C., Callemien, D., Ouwerx, C., & Collin, S. (2002). Use of gas Kim, H., & Keeney, P. G. (1984). Epicatechin content in fermented and unfermented
chromatography–olfactometry to identify key odorant compounds in dark cocoa beans. Journal of Food Science, 49, 1090–1092.
chocolate. Comparison of samples before and after conching. Journal of Kyi, T. M., Daud, W. R. W., Mohammad, A. B., Samsudin, M. W., Kadhum, A. A. H., &
Agricultural & Food Chemistry, 50, 2385–2391. Talib, M. Z. M. (2005). The kinetics of polyphenol degradation during the drying
Counet, C., Ouwerx, C., Rosoux, D., & Collin, S. (2004). Relationship between of Malaysian cocoa beans. International Journal of Food Science & Technology, 40,
procyanidin and flavor contents of cocoa liquors from different origins. Journal 323–331.
of Agricultural & Food Chemistry, 52, 6243–6249. Lamuela-Raventos, R. M., Romero-Perez, A. I., Andres-Lacueva, C., & Tornero, A.
De Wijk, R. A., Terpstra, M. E. J., Janssen, A. M., & Prinz, J. F. (2006). Perceived (2005). Review: Health effects of cocoa flavonoids. Food Science & Technology
creaminess of semi-solid foods. Trends in Food Science & Technology, 17, International, 11, 159–176.
412–422. Luna, F., Crouzillat, D., Cirou, L., & Bucheli, P. (2002). Chemical composition and
Dimick, P. S., & Hoskin, J. C. (1999). The chemistry of flavour development in flavor of Ecuadorian cocoa liquor. Journal of Agricultural & Food Chemistry, 50,
chocolate. In S. T. Beckett (Ed.), Industrial chocolate manufacture and use (3rd 3527–3532.
edition, pp. 137–152). Oxford: Blackwell Science. Mermet, G., Cros, E., and Georges, G. (1992). Etude preliminaire de l’optimisation
Do, T-A. L., Hargreaves, J. M., Wolf, B., Hort, J., & Mitchell, J. R. (2007). Impact of des parametres de torrefaction du cacao. Consommation des precurseurs
particle size distribution on rheological and textural properties of chocolate d’arome, development des pyrazines, qualite organoleptique. Café, Cacao, The,
models with reduced fat content. Journal of Food Science, 72(9), E541–E552. 36, 285–290.
Doyen, K., Carey, M., Linforth, R. S., Marin, M., & Taylor, A. J. (2001). Volatile release Misnawi, S., Jinap, B., Jamilah, N., & Nazamid, S. (2003). Effects of incubation and
from an emulsion: Headspace and in-mouth studies. Journal of Agricultural and polyphenol oxidase enrichment on colour, fermentation index, procyanidins
Food Chemistry, 49, 804–810. and astringency of unfermented and partly fermented cocoa beans.
Engelen, L., De Wijk, R. A., Prinz, J. F., Van Der Bilt, A., Janssen, A. M., & Bosman, F. International Journal of Food Science & Technology, 38, 285–295.
(2002). The effect of oral temperature on the temperature perception of liquids Nelson, R. B. (1999). Tempering. In S. T. Beckett (Ed.), Industrial chocolate
and semisolids in the mouth. European Journal of Oral Science, 110, 412–416. manufacture and use (pp. 347–376). New York: Chapman & Hall.
Engelen, L., de Wijk, R. A., Prinz, J. F., Janssen, A. M., van der Bilt, A., Weenen, H., et al. Oberparlaiter, S., & Ziegleder, G. (1997). Amyl alcohols as compounds indicative of
(2003). A comparison of the effect of added saliva, alpha-amylase and water on raw cocoa bean quality. Zeitschrjft Lebensmittel Untersuchung und Forschung A,
texture and flavor perception in semisolid foods. Physiology & Behaviour, 78, 204(2), 156–160.
805–811. Plumas, B., Hashim, L., & Chaveron, H. (1996). Measurement of the olfactive
Engelen, L., De Wijk, R. A., Bilt, VanDer, Prinz, J. F., Janssen, A. M., & Bosman, F. intensity of chocolate by differential olfactometry. Food Control, 7, 117–120.
(2005). Relating particles and texture perception. Physiology & Behaviour, 86, Pontillon, J. (1995). La fabrication du chocolat. In Pour la Science (pp. 118–126).
111–117. Berlin: Paris.
Engelen, L., & Van Der Bilt, Prinz (2008). Oral physiology and texture perception of Ramli, N., Hassan, O., Said, M., Samsudin, W., & Idris, N. A. (2006). Influence of
semisolids. Journal of Texture Studies, 39, 83–113. roasting condition on volatile flavour of roasted Malaysian cocoa beans. Journal
European Commission Directive (2000) European Commission Directive 2000/36/ of Food Processing & Preservation, 30(3), 280–298.
EC on cocoa and cocoa products. Official Journal of the European Communities, L Reineccius, G. (2006). Flavour chemistry and technology (2nd ed.). Boca Raton, USA:
197/19–25. CRC Press.
Genovese, D. B., Lozano, J. E., & Rao, M. A. (2007). The rheology of colloidal and Roberts, D. D., Pollien, P., & Watzke, B. (2003). Experimental and modeling studies
noncolloidal food dispersions. Journal of Food Science, 72(2), R11–R20. showing the effect of lipid type and level on flavour release from milk-based
Gill, M. S., MacLeod, A. J., & Moreau, M. (1984). Volatile components of cocoa with liquid emulsions. Journal of Agricultural and Food Chemistry, 51, 189–195.
particular reference to glucosinolate products. Phytochemistry, 23(9), Schieberle, P., & Pfnuer, P. (1999). Characterization of key odorants in chocolate. In
1937–1942. Teranishi et al. (Eds.), Flavor chemistry: 30 years of progress (pp. 147–153). New
Granvogl, M., Bugan, S., & Schieberle, P. (2006). Formation of amines and aldehydes York: Kluwer Academic/Plenum.
from parent amino acids during thermal processing of cocoa and model Schnermann, P., & Schieberle, P. (1997). Evaluation of key odorants in milk
systems: New insights into pathways of the Strecker reaction. Journal of chocolate and cocoa mass by aroma extract dilution analyses. Journal of
Agricultural & Food Chemistry, 54, 1730–1739. Agricultural & Food Chemistry, 45, 867–872.
Grassi, D., Lippi, C., Necozione, S., Desideri, G., & Ferri, C. (2005). Short term Schwan, R. F., & Wheals, A. E. (2004). The microbiology of cocoa fermentation
administration of dark chocolate is followed by a significant increase in insulin and its role in chocolate quality. Critical Review in Food Science & Nutrition,
sensitivity and a decrease in blood pressure in healthy persons. American 44, 205–221.
Journal of Clinical Nutrition, 81, 611–614. Sokmen, A., & Gunes, G. (2006). Influence of some bulk sweeteners on rheological
Guinard, J.-X., & Marty, C. (1995). Time-intensity measurement of flavour release properties of chocolate. LWT- Food Science & Technology, 39, 1053–1058.
from a model gel system: Effect of gelling agent type and concentration. Journal Stark, T., Bareuther, S., & Hofmann, T. (2005). Sensory-guided decomposition of
of Food Science, 60(4), 727–730. roasted cocoa nibs (Theobroma cacao) and structure determination of taste-
Hatano, T., Miyatake, H., Natsume, M., Osakabe, N., Takizawa, T., Ito, H., et al. (2002). active polyphenols. Journal of Agricultural & Food Chemistry, 53, 5407–5418.
Proanthocyanidin glycosides and related polyphenols from cacao liquor and Taylor, A. J. (2002). Food flavour technology. Sheffield, UK: Sheffield Academic Press.
their antioxidant effects. Phytochemistry, 59, 749–758. Taylor, A. J., & Roberts, D. D. (2004). Flavour perception. Oxford, UK: Blackwell
Hermann, F., Spieker, L. E., Ruschitzka, R., Sudano, I., Hermann, M., Binggeli, C., et al. Publishing Limited.
(2006). Dark chocolate improves endothelial and platelet function. Heart, 92, Voltz, M., & Beckett, S. T. (1997). Sensory of chocolate. The manufacturing
19–120. confectioner, 77, 49–53.
Hyvonen, L., Linna, M., Tuorila, H., & Dijksterhius, G. B. (2003). Perception of melting Ziegler, G., & Hogg, R. (1999). Particle size reduction. In S. T. Beckett (Ed.), Industrial
and flavour release of ice cream containing different types and contents of fat. chocolate manufacture and use (pp. 182–199). New York: Chapman & Hall.
Journal of Dairy Science, 86, 1130–1138. Ziegler, G. R., Mongia, G., & Hollender, R. (2001). Role of particle size distribution of
ICA (1988). Determination of moisture content of cocoa and chocolate products, suspended solids in defining the sensory properties of milk chocolate.
Analytical Method 26. CAOBISCO, rue Defacqz 1, B-1000 Bruxelles, Belgium. International Journal of Food Properties, 4, 353–370.