Development and Optimization of a Cocoa Incubation Process to Mimic Spontaneous Fermentation

Fermentation is an essential step for the development of cocoa's sensory characteristics, and under field conditions, the lack of control over variables such as temperature and pH results in quality variations between batches. Therefore, incubation under controlled conditions emerges as an alternative to reproduce the fermentation process. This work aimed to develop a microorganism-independent cocoa incubation method, capable of mimicking the biochemical transformations of spontaneous fermentation in a controlled and reproducible manner based on different biochemical markers and modeling strategies. Based on literature data and preliminary tests, non fermented and dried cocoa seeds (70 g) were rehydrated in 210 mL of water (3:1 v/w) for 24 hours. A Central Composite Design (CCD), with 17 runs, was used to evaluate the effects of time (24, 72, and 120 h), temperature (40, 50, and 60 °C), and pH (3, 5, and 7, all containing 5% ethanol) on incubation. The analyzed responses, also determined for non-fermented and reference fermented cocoa samples, were: total phenolics compounds (Folin-Ciocalteu), total free amino acids (ninhydrin method), reducing sugars (DNS), and fermentation index (based on anthocyanin degradation), all quantified by spectrophotometry. Using R, the data were subjected to normality (Shapiro-Wilk) and homoscedasticity (Levene) tests, ANOVA with Tukey's HSD post-hoc test, response surface generation, principal component analysis (PCA), and multivariate optimization using desirability methodology, targeting the reference fermented cocoa values. Under some tested conditions, the fermentation index, sugar, and phenolic parameters showed behavior similar to that of fermented cocoa. Model residuals showed normal distribution, except for the fermentation index, as well as homogeneity of variances. The model for total phenolics was not statistically significant but was retained in the optimization due to its relevance to the fermentation process. PCA explained 74.31% of the data variance, showing that the incubation trials were distributed across the first two principal components, yet clearly separated from the non-fermented sample, indicating that the incubation process promoted biochemical changes consistent with fermentation. Desirability optimization indicated the predicted condition to be 51.11 °C, 120 hours, and pH 3, with individual desirabilities ranging from 0.80 to 0.99 and an overall desirability of 0.90, obtained from the geometric mean of the individual responses. The results suggest that, through incubation, it is possible to reproduce the biochemical changes characteristic of spontaneous cocoa fermentation, approaching the reference profile across multiple parameters simultaneously, which can be modulated to obtain different chemical profiles in a reproducible manner.

Acknowledgements:The authors acknowledge the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) (Process 001) and CNPq (Process:304472/2026-3) for financial support.