Zhu Jianxiong, Sun Zhongda, Xu Jikai, Walczak Rafal D, Dziuban Jan A, Lee Chengkuo
School of Mechanical Engineering, Southeast University, Nanjing 211189, China; Department of Electrical and Computer Engineering, National University of Singapore, Singapore 117576, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore 117576, Singapore; NUS Suzhou Research Institute (NUSRI), Suzhou 215123, China.
Department of Electrical and Computer Engineering, National University of Singapore, Singapore 117576, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore 117576, Singapore; NUS Suzhou Research Institute (NUSRI), Suzhou 215123, China.
Sci Bull (Beijing). 2021 Jun 30;66(12):1176-1185. doi: 10.1016/j.scib.2021.03.021. Epub 2021 Mar 23.
Ion mobility analysis is a well-known analytical technique for identifying gas-phase compounds in fast-response gas-monitoring systems. However, the conventional plasma discharge system is bulky, operates at a high temperature, and inappropriate for volatile organic compounds (VOCs) concentration detection. Therefore, we report a machine learning (ML)-enhanced ion mobility analyzer with a triboelectric-based ionizer, which offers good ion mobility selectivity and VOC recognition ability with a small-sized device and non-strict operating environment. Based on the charge accumulation mechanism, a multi-switched manipulation triboelectric nanogenerator (SM-TENG) can provide a direct current (DC) bias at the order of a few hundred, which can be further leveraged as the power source to obtain a unique and repeatable discharge characteristic of different VOCs, and their mixtures, with a special tip-plate electrode configuration. Aiming to tackle the grand challenge in the detection of multiple VOCs, the ML-enhanced ion mobility analysis method was successfully demonstrated by extracting specific features automatically from ion mobility spectrometry data with ML algorithms, which significantly enhance the detection ability of the SM-TENG based VOC analyzer, showing a portable real-time VOC monitoring solution with rapid response and low power consumption for future internet of things based environmental monitoring applications.
离子迁移率分析是一种在快速响应气体监测系统中用于识别气相化合物的知名分析技术。然而，传统的等离子体放电系统体积庞大，在高温下运行，不适用于挥发性有机化合物（VOC）浓度检测。因此，我们报道了一种基于摩擦起电电离器的机器学习（ML）增强型离子迁移率分析仪，它以小型设备和宽松的操作环境提供了良好的离子迁移率选择性和VOC识别能力。基于电荷积累机制，多开关操作摩擦纳米发电机（SM-TENG）可以提供几百量级的直流（DC）偏置，利用特殊的尖端-平板电极配置，该偏置可进一步用作电源，以获得不同VOC及其混合物独特且可重复的放电特性。为应对多种VOC检测中的重大挑战，通过使用ML算法从离子迁移谱数据中自动提取特定特征，成功展示了ML增强型离子迁移率分析方法，这显著提高了基于SM-TENG的VOC分析仪的检测能力，为未来基于物联网的环境监测应用展示了一种具有快速响应和低功耗的便携式实时VOC监测解决方案。
