Researchers Develop Novel Spectroscopy Technique to Identify Geographic Origin of Roasted Coffee – CoffeeTalk

An upcoming poster presentation at the American Chemical Society (ACS) Fall 2026 Meeting will focus on the innovative use of spectroscopy techniques to differentiate various roasted coffee types. Researchers from the University of Bergen developed a groundbreaking nondestructive method that allows for the rapid identification of the geographic origin of roasted coffee within minutes, eliminating the need for the destructive sampling methods of the past. This study contributes to the growing field of nondestructive technologies aimed at analyzing food products effectively.

The presentation will feature Sleshi Fentie Tadesse, who will discuss findings from the research during a poster session scheduled for Wednesday, August 26, from 7 to 9 p.m. EDT, at McCormick Place Convention Center in Chicago, Illinois. The research combines two key infrared (IR) spectroscopy techniques—attenuated total reflectance Fourier-transform infrared (ATR-FTIR) and near-infrared (NIR) spectroscopy—with statistical modeling to classify roasted, ground coffee by its origin, variety, and roast level.

Understanding the geographical origin of coffee is crucial for several reasons, including accurate market pricing and fostering consumer trust. Existing verification methods, such as chromatography and sensory panels, while accurate, are slow and require sample destruction, making them impractical for routine quality control in the coffee industry. The study analyzed coffee samples from renowned coffee-producing countries, including Ethiopia, Brazil, Colombia, Costa Rica, Kenya, Mexico, Peru, and Rwanda. Spectral data was collected across mid-infrared (MIR) and NIR ranges, and noise and scattering artifacts were systematically removed through processing.

The researchers evaluated two classification approaches: soft independent modeling of class analogy (SIMCA) and partial least squares discriminant analysis (PLS-DA). SIMCA achieved 100% specificity and over 92% accuracy in distinguishing coffee origins, although it was less reliable for sorting by roast level. Conversely, PLS-DA exhibited perfect classification performance across all metrics evaluated, supported by robust statistical confidence. The findings indicated that ATR-FTIR data provided superior predictive accuracy compared to NIR data, attributed to the former’s heightened sensitivity to specific molecular structures.

This study holds significant implications for the coffee industry, which currently lacks a scalable quality-control tool. Tadesse’s nondestructive method could potentially be integrated into existing inspection workflows without disrupting supply chains, positioning it as a strong candidate for routine authentication and quality assurance applications, subject to further validation regarding its effectiveness in distinguishing roast levels.

Read More @ Spectroscopy