EFFECT OF THE MATURATION PROCESS ON FERMENTED COCOA BEAN I : FREE AMINO ACIDS AND VOLATILE

Absfractr The maturation step ,in cocoa processing was originally introduced to reduce acidity of the cocoa bean. This study shpws that maturation causes a decrease in free amino acids except glutamic and aspartic acid. Maturation also showed an increase in volatile carbonyls. The main increases were in diacetyl, formaldehyde, acetone, acetaldehyde. butyraldehyde (normal plus iso-) and valeraldehyde ( normal plus iso-) levels. Although there appears to be a relationship between these two trends, the appearance of carbonyls appears to be caused not only by degradation of amino acihs but also other processes for example the oxidative degradation of fatty acids. Roasting caused a further decline of amino acids and increase of carbonyls presumably due t o the Saecker reaction.


Introduction
One of the problems of fermentation of cocoa beans is the prevalence of high acidity.This is the characteristic of cocoa beans produced in Malaysia, New Guinea and Uganda where the pH of the fermented bean has been reported t o be between 4.6 -4.8.2*3 9 4 In order to circumvent this problem ~i a u ~ proposed the process of "maturation" which simply meant holding the fermented cocoa beans in thin layers at ambient temperatures during which time little or no drying took place.During this time acetic acid is oxidised to C 0 2 by the Krebs cycle.This rocess was primarily proposed to overcome high acidity.andSchwartz e t nLg worked out several simple technologies to reduce acidity .andthereby enhance the flavour of cocoa t o be used in chocolate manufacture.In Sri Lanka acidity of cocoa bean had not been a major problem, mainly due t o the fact that cocoa beans are largely under fermented.However, if cocoa is properly fermented (for four days or more) the acidity problem does arise in this country too.
In this study fermented cocoa beans that were treated by the maturation process were subjected t o analysis for free amino acid and volatile carbonyl contents as it was felt that the process not only affected the

7,
acidity but might also affect the flavour and flavour precursors.This was indicated by a study of packiyasothy6 which 'showed that volatile carbonyl content increased with the maturation process.

Trapping o f volatile carbonyls as 2,4-DNP derivatives
is tilled water (400 ml) was added t o finely ground shell-free cocoa bean (100 g) and the mixture was distilled over 4h.The volatile compounds &ere trapped in isopentane : diethyl ether (1:l) using a modified Clavenger apparatus fitted with an iced water condenser: ( ~O C , Fryka-Therm water cooler).The extract was collected at half hourly intervals while adding fresh solvent (isopentane : diethyl ether, 1:1), The carbonyls present in the total extract (% 35 ml) were then converted t o their 2,4-DNP derivatives.'' Although it is probable that recovery of carbonyls and derivatisation is not 100% as standard conditions were followed any errors introduced by these standard procedures would be common to different treatments.Fermented, matured, fermented and roasted, matured and roasted cocoa was subjected t o such analysis.

Standard curves
Standard curves were plotted using' standard solutions (50 mg1100 ml) after suitable dilutions using chloroform as solvent..An aliquot of 10111 (1-5pg) was spotted (keeping the spotting volume constant at 1m1) on silica gel G-60 TLC plates ( 3 0 0 ~) developed as described in 2.6.3.The dried plates were scanned transversely for the standard curve.The technique of dual 1 2 mm non-overlapping bands was employed to ensure that the entire spot is covered when spot width reached 1 2 mm or a near value.
The standard curves were plotted accordingly (weight pg versus peak area).The standard curves gave straight lines in the ranges of (0.5pg).As there was some doubt regarding reproducibility from plate t o plate, the standard curves were used only t o ensure that the peak area was proportional t o the 2,4-DNP derivative levels (pg).Standard spots were run on each plate for quantification purposes.

2.6.5.Separation o f 2,4-DNP derivatives of carbonyls of cocoa
The 2,4-DNP derivatives of carbonyls from cocoa distillates were prepared.These were dissolved in chlorofrom (50 mg/25 ml) and aliquots (5yl) were spotted in TLC plates (silica gel G-60-500~) and developed as described above.The plates were subjected t o TLC-UV densitometric analysis as in 2.6.3.
Each peak area was also calculated using a computer programme on a Commodore 64 Personal Computer.The peaks which were not picked up by the integrator were quantified by this method.The calibration of each run was carried out by spotting a known quantity of a sample on the same plate.

Incubation Experiments with Amino Acids
Samples of-fermented cocoa beans (500 g) were macerated in a Waring blender.The macerated samples were treated as follows -One sample was held at 5 0 ' ~ using 1N Na2Ca0 These treatments (50 g) Was retained as a control.Another sample (50 g) for 2h.A thkd sample (50 g) was made alkaline (pH 11.0) l 3 after which the pH was reduced to (4.2) using 2N HCl. were introduced in an effort to deactivate enzymes and which was partially successful (Table 4).
The remaining samples (50g each) were incubated with the following amino acids (50 mg/100ml distilled water) alanine, valine, leucine, threonine and glutamic acid.To all samples a solution (5 ml) of sodium benzoate fixed concentration (200 ppm) was added to prevent growth of microorganisms.
The samples were allowed t o stand for 48h.(covered with polythene to prevent drying).
The volatiles were trapped and the carbonyls converted to their 2,4-DNP derivatives and quantified.

Results
The process of maturation resulted in the pH of the cocoa beans rising from 4.8 to 5.2.

Effect on Amino Acids
Maturation caused a decline in amino acids.The only amino acids that showed increases were glutamic acid and aspartic acid.The roasting process also caused a decline in amino acid content.Details of results are shown in Table 1

Effect on Carbonyls
Maturation had a dramatic effect on volatile carbonyls.The results are shown in Table 2.A TLC densitometric scan of the matured sample is shown in Figure 1 and the quantification of carbonyls is shown in Table 3.
In the case of butyraldehyde and valeraldehyde, it was not possible to separate the normal and iso-forms.However, the bulk is likely t o be the iso-form?
Roasting at 165' C resulted in an increased level o f carbonyls as compared to 1 2 5 ' ~ The peaks between 1 and 2 are unidentified.
A peak No. 4 (not in this particular scan) normally seen between 3 and 5 is normally propanaldehyde.
The large peak between 1 and 2 was unidentified and was possibly the mono-DNP derivative of diacetyl (Peak No. 1 was confirmed t o be the di DNP derivative of diacetyl)

Table 1 .
. D. D. Abeygunasekera and E. R. Jansz Effect of maturation and roasting of cocoa on individual amino acids