Determination of bacteria-specific odor signatures using low-cost metal oxide gas sensors

In this study, a feasibility assessment was conducted to evaluate the potential of commercially available low-cost gas sensors for differentiating bacterial species inoculated into fish muscles. The signals were obtained by determining how low-cost gas sensors detected the volatile organic compound (VOC) patterns produced by each bacterial species (Enterococcus faecalis, Pseudomonas luteola, Proteus mirabilis, and Photobacterium damselae) in trout tissue. It was observed that MQ3 and MQ4 sensors yielded high rates for all bacteria, indicating a strong increase in broad VOC load consistent with the known cross-sensitivity range of MQ sensors as a result of bacterial metabolism. In addition to cumulative VOC patterns, time-dependent derivative analysis (dVOC/dt) was applied to reveal bacterium-specific metabolic differences in more detail. Thus, it was determined that Photobacterium damselae, unlike other bacteria, exhibited suppressed VOC production kinetics in the early stages, demonstrating that dynamic VOC fingerprints are a powerful tool for bacterial differentiation. The results demonstrate the feasibility of dynamic VOC fingerprinting using minimal sensor configurations for exploratory bacterial discrimination.