Multi-elemental Analysis of Philippine Coffee Using X-ray Fluorescence Spectrometry for Varietal and Geographical Discrimination

Specialty coffee from specific regions in the Philippines holds high value, making it susceptible to adulteration. To ensure authenticity in terms of variety and origin, chemical analysis of coffee beans is crucial. This study, part of a larger multi-year project, used handheld X-ray Fluorescence (hXRF) to analyze 11 green coffee bean samples (Arabica, Robusta, Excelsa, and Liberica) from five regions (CAR, CALABARZON, Western Visayas, Central Visayas, and Caraga). Among the samples, high levels of potassium (K), magnesium (Mg), and sulfur (S) were observed, while arsenic (As), bismuth (Bi), and yttrium (Y) were detected in low amounts. Samples from each coffee variety and region exhibited distinct elemental compositions, allowing for differentiation. The study utilized a Random Forest classification model to predict coffee variety and geographical origin based on XRF-derived elemental data. As an estimation of performance, the OOB error and a test dataset was utilized, confirming the accuracy of the classification model. Furthermore, the MDS plots obtained for the varietal and geographical classification demonstrated a distinct clustering of samples. This suggests that multi-elemental profiles serve as effective discriminants and can be used as elemental fingerprints for identifying coffee variety and establishing provenance. Overall, the study highlights the potential of XRF-based multi-elemental profiling combined with machine learning algorithms as a promising method for coffee authentication and fraud detection.