Abstract:
Cocoa bean quality is influenced by preharvest and postharvest factors. Rapid
evaluation is therefore required to aid in decision-making. In this study, four
experiments were performed separately using a completely randomized design
with age class, pollination type, production method, cocoa-producing region
and geographical location as the main factors. The novel handheld NIR
spectrometer combined with multivariate qualitative algorithms gave a 100 %
classification rate for cocoa beans; from the seven cocoa-growing regions in
Ghana, four geographical locations in Africa, fennented against unfermented,
and organic against conventional. Quantitatively, the perfonnance of the
regression models for simultaneous prediction of fennentation index, pH, fat,
polyphenols, flavonoids , and antioxidant capacity was in the range of:
0.87 < R2
cal < 0.99 and 0.88 < R2pr~ < 0.99 in calibration and prediction sets
respectively. Cocoa beans' physical, chemical and mineral propelties were
significantly impacted by age of the cocoa tree and pollination type. Calcium,
magnesium, phosphorus and potassium were found in the range of 111.44 -
125.23, 238.79 - 249.05, 528.24 - 541.40 and 473 .05 - 631.34 mgllOOg,
respectively whereas sodium, iron, copper and zinc were found in the range of
7.08 - 11.54,5.80 - 8.83,1.34 -3 .33 and 2.36 - 5.14 mg/l00 g, respectively for
cocoa bean categories examined. Generally, the NIR spectroscopic technique
developed cOlTelated well with the wet chemistry method (R2 = 0.93). The
outcome of the study reveals that the handheld NIR spectroscopic technique can
be used for rapid, non-destructive and on-site measurement of cocoa beans
quality parameters qualitatively and quantitatively.
