First-line electroanalytical screening platform for rapid detection of Alternaria toxins in fresh tomatoes based on chemically-modified screen-printed gold electrodes and Machine Learning predictive models

• Modified-electrode electrochemical platform for rapid screening of Alternaria toxins • DoE-optimized dodecanethiol SAM deposition for improved AOH and AME detection • Machine learning-driven Alternaria toxin detection • Excellent compliance to green analytical principles confirmed by three metric tools • Implementation of the ML algorithm required for data processing on the edge The detection of Alternaria mycotoxins in food is a growing concern due to their potential health effects. Although HPLC-MS-based techniques are the most widely used for the accurate determination of mycotoxins, they are not suitable for rapid on-site screening, which emphasises the need for portable, cost-effective and environmentally friendly analytical instruments. In this study, we developed a first-line electroanalytical platform for the rapid detection of alternariol (AOH) and alternariol-9-methyl ether (AME) using screen-printed gold electrodes modified with dodecanethiol self-assembled monolayers (SAMs). Dodecanethiol SAM deposition was optimized using a 2 2 full factorial design to enhance and differentiate the electrochemical responses of each of the two toxins with even minimal structural differences, as confirmed by cyclic voltammetry and impedance spectroscopy. Good results with LOD and LOQ at the low µg/kg level were achieved, approaching the indicative levels recommended for the target compounds in processed tomato products. Very good inter-electrode repeatability with RSD