Non-invasive monitoring of exhaled acetone-a recognized biomarker of diabetes-has offered a promising and convenient diagnostic approach for diabetes. However, conventional optical and metal oxide semiconductor sensors suffer from bulky instrumentation, high power consumption, and poor portability. Metal-organic framework (MOF)-based sensors can overcome these drawbacks but still require improvements in response time and stability. Here, we develop a gate-sensitive field-effect transistor (GS-FET) gas sensor functionalized with a sensitive MOF for ultrafast and noninvasive acetone detection. The MOF serves as a chemical-sensitive gate, modulating the polysilicon channel current, while a solvent-modification strategy promotes the density of edge-unsaturated sites with enhanced adsorption activity, as confirmed by density functional theory. Benefiting from these optimizations, the GS-FET sensor achieves a sub-500 ppb detection limit toward acetone and enables real-time breath analysis when integrated into a portable mobile-linked device. To further improve the practicality and convenience of the gas sensor, we have proposed a data analysis algorithm to predict the concentration of acetone based on the initial response of the sensors within 5 s with high data reliability. This work demonstrates a practical pathway for leveraging MOF-based architectures in ultrafast, noninvasive diabetes diagnosis and provides new insights into the development of high-performance gas sensors.
Liu et al. (Thu,) studied this question.