Exploring portable NIR spectroscopy in transmittance and reflectance modes for the authentication of Brazilian coffees with geographical indications

BACKGROUND Geographical indications (GIs) certify the link between coffee origin and quality, enabling premium pricing and protecting producers. However, the high market value of GI coffees increases their vulnerability to fraud, underscoring the need for reliable and practical authentication methods. Portable near‐infrared (NIR) spectroscopy represents a rapid and non‐destructive alternative, but its capability to discriminate Brazilian GI coffees requires systematic assessment. RESULTS This study compared NIR transmittance spectra of aqueous coffee extracts with NIR reflectance spectra of the corresponding ground samples for the authentication of four Brazilian GIs from southeastern Brazil: Cerrado Mineiro, Mogiana Paulista, Mantiqueira de Minas, and Matas de Minas. Distinct spectral signatures were observed in the 900–1650 nm range. Data‐driven soft independent modeling of class analogy (DD‐SIMCA) was employed, resulting in excellent classification performance. Although both acquisition modes showed satisfactory performance during calibration and validation, reflectance consistently outperformed transmittance in the prediction of a test set, achieving accuracies ranging from 97% to 100%. The superior performance of reflectance‐based models was attributed to the preservation of chemically informative features in the solid coffee matrix, including lipids and other compounds poorly extracted into water, whereas aqueous extracts were dominated by water absorption and exhibited greater intraclass variability. CONCLUSION Portable NIR spectroscopy, particularly in reflectance mode combined with DD‐SIMCA, provides a fast, non‐destructive, and highly reliable approach for authenticating Brazilian GI coffees. These findings highlight its potential as a practical tool to protect producers and consumers against fraud and to ensure the integrity of products bearing protected geographical indications.