Detection and quantification of fish spoilage marker using polymer-coated micro-electro-mechanical systems sensor array-based electronic nose—a computational study

Abstract

This study explores the feasibility of polymer functionalized microelectromechanical systems (MEMS) sensor array-based Electronic nose for fish spoilage detection. Using vapor-polymer partition coefficients as key features, we employ data-driven techniques, including principal component analysis (PCA), fuzzy c-means (FCM), fuzzy subtractive clustering (FSC), and particle swarm optimization (PSO), to optimally select five polymers from a pool of 39 commercial candidates. The selected polymers’ effectiveness in generating diversified sensor responses is evaluated through a virtual sensor array model. The MEMS sensor array, functionalized with these polymers, operates in both static and dynamic sensing modes. By leveraging experimentally reported volatile organic compounds (VOCs) in fish headspace, we demonstrate effective separation of spoilage markers, particularly trimethylamine (TMA), dimethylamine (DMA), 1-Octanol, and Indole, in principal component space. The study further validates the system’s calibration for TMA quantification using generalized regression neural networks (GRNN) and radial basis function (RBF) neural networks, achieving an RMSE of 0.0002 and R2 = 0.999. These findings highlight the feasibility of MEMS-based E-Nose technology for real-time fish spoilage monitoring, with potential for industrial applications. © Association of Food Scientists & Technologists (India) 2025.