Use of non-destructive methods to differentiate cocoa beans based on variety and fermentation level

Theobroma cocoa is a cash crop found in all cocoa-producing countries. In the Republic of Congo, there are three main varieties: Criollo, Forastero, and Trinitario. Determining cocoa bean quality (i.e. fermentation level) is an important production and trade issue. This study aimed to (i) determine whether the variety and geographical origin of whole fermented dried cocoa beans could be distinguished using Raman spectrometry, hyperspectral imaging (HSI), and near-infrared spectrometry (NIRS) and (ii) assess whether these non-destructive methods could characterize bean fermentation level. The latter was determined using the cut test and the fermentation index.

The main peaks of the Raman, HSI, and NIR spectra were associated with chemical compounds and groups when possible. Bean variety could be distinguished (accuracy = 98.2%, 91.4%, and 80.2% for Raman, HSI, and NIRS, respectively) as could bean geographical origin (accuracy = 99.4%, 97.3%, and 97.1% for Raman, HSI, and NIRS, respectively). All three methods yielded very good predictions of actual fermentation levels, determined using the cut test (accuracy: 97%); the most effective methods were HSI followed by Raman spectroscopy. All three methods could also yield very good predictions of fermentation index values using models containing a selection of 9-12 spectral bands (Raman: R² = 0.92, HSI: R² = 0.99, and NIRS: R² = 0.997; model errors < 0.04).

These non-destructive methods are thus demonstrably effective and versatile and could be used by industry to assess cocoa bean quality, and even authenticate beans, if a wider database is built. © 2026 Society of Chemical Industry.