Wireless food-freshness monitoring and storage-time prediction based on ammonia-sensitive MOF@SnS2 PN heterostructure and machine learning

Conductive metal-organic frameworks (C-MOF) materials possess a high absorbability and structural tunability. However, their low sensitivity and poor specificity limit their applications. Constructing PN heterostructures with other semiconductor materials can regulate the C-MOF energy band structure, reduce particle stacking, and boost gas-sensing performance. Here, we synthesize Cu3(HHTP)2 (C-MOF) material in-situ onto SnS2 nanolayers to form PN heterojunctions that facilitate high-performance NH3 sensing, a four times higher response compared to pristine MOF-based sensors, ultralow detection limits (experimental: 125 ppb; theoretical: 9.84 ppb), and improved selectivity against various rotten gases. Machine learning is applied to analyse the effects of temper-ature, time, humidity, and category on the sensing response, achieving 78.5 % accuracy in predicting meat storage time. By adding a Bluetooth module and cloud-based signal analysis, we provide a proof-of-concept for a portable, fast response, non-destructive remote gas sensing device for real-time food-freshness monitoring deployable at any stage along the food supply chain.

Automated summary

By synthesizing Cu3(HHTP)2 (C-MOF) material in-situ onto SnS2 nanolayers to form PN heterojunctions, the NH3 sensing performance is improved with a four times higher response compared to pristine MOF-based sensors, ultralow detection limits, and improved selectivity against various rotten gases. Machine learning is applied to analyze the effects of temperature, time, humidity, and category, achieving 78.5% accuracy in predicting meat storage time. With the addition of a Bluetooth module and cloud-based signal analysis, a proof-of-concept for a portable, fast response, non-destructive remote gas sensing device for real-time food freshness monitoring is provided.