A portable acetone detector based on SmFeO3 can pre-diagnose diabetes through breath analysis

Abstract

Breath analysis using gas sensors is becoming one of the important supplementing ways for an early diagnosis of diseases. The feasibility of detecting acetone concentration in exhaled breath using a SmFeO3-based sensor has been studied herein. The detection limits of acetone using a SmFeO3-based sensor are extremely low (1.56–0.01 ppm; 1.87–0.02 ppm; 2.39–0.05 ppm; 3.22–0.1 ppm; 4.1–0.2 ppm; 4.79–0.5 ppm; 5.92–1 ppm). To remove the effects of carbon dioxide and relative humidity (RH), a mathematical model between the response and acetone concentration, carbon dioxide concentration and RH in a person’s exhaled breath is established via machine learning. Volunteer studies have proved the accuracy of this mathematical model. Compare with the acetone concentration obtained by GC-MS method, the accuracy of the proposed method exceeded 85 %. Finally, a portable acetone detector based on the SmFeO3 sensor is designed and realized to monitor the dynamic change in acetone concentration in the exhaled breath of volunteers before and after meals, and the experimental results have proved that the method based on the gas sensor and Machine learning (ML) is feasible and effective in accurately detecting acetone concentration in persons’ exhaled breath.