Efficacious paper-based colorimetric detection of bacterial contamination in vegetables utilizing indicator dyes and machine learning.

Abstract

Food contamination from bacteria and resulting spoilage has been a persistent problem in the supply chain, leading to substantial waste and financial loss. Likewise, vegetables are prone to microbial contamination due to poor/unhygienic agricultural practices/distribution. This work reports simple, cost-effective, non-destructive, non-contact, and continuous detection of bacteria. Employing a colorimetric approach, paper-based sensors were fabricated while utilizing pH-sensitive indicator dyes. The sensors exhibited noticeable color change on exposure to VOCs emitted by bacterial strains. The pH/color change relies on deprotonation and restructuring of indicator molecules during VOC interactions. The sensors successfully detected the presence of bacterial contamination in vegetables and enabled associated spoilage estimation. Recorded colors were digitized to RGB values and combined with random forest classifier for an erudite distinction of contaminated samples. The results accentuate the feasibility of the approach for real-time and automated detection of bacterial contamination in food supply chain with better accessibility and high throughput.