Breadcrumb
- Home
- Publications
- Proceedings
- 2025 AIChE Annual Meeting
- Annual Student Conference: Competitions & Events
- Undergraduate Student Poster Session: Computing and Process Control
- Deep Learning Approaches for Voltammetric Analysis of Coffee
While instruments such as high-performance liquid chromatography (HPLC) and mass spectrometry have commonly been used to accurately assess the chemical composition of coffee, they require expensive equipment, meticulous preparation, and specialized expertise. Emerging electronic tongue technologies have introduced methods for mimicking human tasting abilities. These tongues often utilize electrochemical techniques to detect redox-active analytes in complex solutions. This study demonstrates the use of a low-cost electrochemical system and screen-printed electrodes, combined with a neural network for data processing, to analyze key compounds in high-quality coffee.
We compare HPLC data on chlorogenic acids (CGAs) and caffeine in pure standards and coffee samples with cyclic voltammetry readings. Our voltammetry data, combined with deep learning techniques, suggest that we can estimate the concentration of caffeine and CGAs in coffee. Our findings highlight the potential for low-cost electrochemical methods for coffee analysis, while emphasizing the need for further refinement in sensor technology and data processing techniques to enhance sensitivity and reduce variability. Integrating machine learning, specifically neural network processing, shows promise in mitigating such limitations, such as the variability between screen-printed electrodes.
A low-cost approach could make quality and taste profile analysis accessible and easier to use for more researchers, coffee companies, or individual coffee enthusiasts. High-accuracy analysis could enable businesses and individuals to assess the taste profile and quality of their coffee in a non-subjective, accessible manner. These processing methods are being further explored to analyze more advanced coffee parameters such as total dissolved solids, roast level, and origin characteristics.