Thesis Examination Committee
Prof Cameron Dougall CAMPBELL, SOSC/HKUST (Chairperson)
Prof Amine BERMAK, ECE/HKUST (Thesis Supervisor)
Prof Chi Ying TSUI, ECE/HKUST (Thesis Co-supervisor)
Prof Jing KONG, Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology (External Examiner)
Prof Wing Hung KI, ECE/HKUST
Prof Zhiyong FAN, ECE/HKUST
Prof Yi-Kuen LEE, MAE/HKUST
Portable and wearable gas sensors have witnessed an expanding market application in areas related to air quality monitoring and toxic gas detection. Large deployment of such devices in IoT applications is challenging because of the large power consumption, large system size, poor selectivity, high maintenance cost, to name few. This thesis is devoted to address these issues and in particular contributes to the development of a low power autonomous integrated gas sensor array with gas classification ability.
First, a multi-mode gas monitoring circuit is presented for the widely developed resistive gas sensor array. Read-out of large resistance range of the gas sensors from 400kΩto 1GΩ is achieved. In the monitoring phase, the gas detection mode will detect the signal change and trigger the ratio-metric sensitivity quantizer only when the gas exposure activity is detected. The sensitivity in the transient region is automatically quantized and multiple features in the transient region are extracted for improved classification performance. In addition, the transmission bandwidth is largely reduced according to gas exposure activity.
Second, Gas-sensitive FETs (Gas-FETs) in standard CMOS is introduced. 8x8 Gas-FETs are integrated with the readout and quantization circuits. The top metal is exploited as the extended gate with a lateral control gate or reset switch. Various sensing materials can be deposited on top to form a diverse sensor array. The drain current in the weak inversion region is readout then quantized by column SAR ADCs integrated on-chip side-by-side with the sensor array.
Third, temperature modulation with an embedded heater is applied for better selectivity. A 16x8 array has been fabricated in standard CMOS. Each local site can sense the temperature and the gas response. Meanwhile, the temperature change represents the different thermal conductivity at different gas exposure, which offers another sensing mechanism for better gas classification. The sensing principle is extended to measure flow rate from the temperature change within the sensor array. This provides the very first of its kind sensing platform that can not only quantify gases in an efficient way but also sense the flow of gases in the environment. To the best of our knowledge, this is the very first sensing platform that provides gas sensing and classification as well as flow measurement leading to huge potential future applications.