EUVIMEDObjevte EUVIMED
Evaluationsstudie

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

Bell-Blonde Danielcia Biyendolo Loumpangou iD, Daoud Ounaissi iD, Vanessa Lançon-Verdier iD, Jean-Mathurin Nzikou iD, Chantal Maury iD

Journal of the Science of Food and Agriculture · 2026

DOI: 10.1002/jsfa.70463 ↗

At a glance

Evaluationsstudie in Journal of the Science of Food and Agriculture, 2026. Written by Bell-Blonde Danielcia Biyendolo Loumpangou and 4 co-authors. Institution: GRAPPE, Ecole Supérieure des Agricultures, USC INRAE 1422, SFR 4207 QUASAV, Angers, France. Major MEDLINE topics: Cacao, Seeds, Spectrum Analysis, Raman, Hyperspectral Imaging. 91 references. Funded by: Institut International 2i.

AbstraktSubjectsBibliografické údajeCiteRelated studies

Abstract (original language)

Background

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.

Results

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: R 2 = 0.92, HSI: R 2 = 0.99, and NIRS: R 2 = 0.997; model errors < 0.04).

Conclusion

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.

Subjects

MeSH terms (MEDLINE)

CacaoSeedsSpectrum Analysis, RamanHyperspectral ImagingFermentationSpectroscopy, Near-Infrared

Author keywords

NIRSRamancocoa beansfermentationhyperspectral imagingprediction

Authors and institutions

  • GRAPPE, Ecole Supérieure des Agricultures, USC INRAE 1422, SFR 4207 QUASAV, Angers, France.
  • UNESCO-ENSP CHAIR, Marien NGOUABI University, Brazzaville, Congo.

ORCID profiles: Bell-Blonde Danielcia Biyendolo Loumpangou, Daoud Ounaissi, Vanessa Lançon-Verdier, Jean-Mathurin Nzikou, Chantal Maury

Funding

  • Institut International 2i

Cite

APA

Biyendolo Loumpangou, B. D., Ounaissi, D., Lançon-Verdier, V., Nzikou, J. M., & Maury, C. (2026). Use of non‐destructive methods to differentiate cocoa beans based on variety and fermentation level. Journal of the Science of Food and Agriculture. https://doi.org/10.1002/jsfa.70463

Vancouver

Biyendolo Loumpangou BD, Ounaissi D, Lançon-Verdier V, Nzikou JM, Maury C. Use of non‐destructive methods to differentiate cocoa beans based on variety and fermentation level. Journal of the Science of Food and Agriculture. 2026. doi:10.1002/jsfa.70463

BibTeX
@article{BiyendoloLoumpangou2026,
  author = {Biyendolo Loumpangou, Bell-Blonde Danielcia and Ounaissi, Daoud and Lançon-Verdier, Vanessa and Nzikou, Jean-Mathurin and Maury, Chantal},
  title = {Use of non‐destructive methods to differentiate cocoa beans based on variety and fermentation level},
  journal = {Journal of the Science of Food and Agriculture},
  year = {2026},
  doi = {10.1002/jsfa.70463}
}
Start for free →