Energy Efficient Analog Mixed-Signal Front Ends for CNT-FET NO2 Air-Quality Nanosensors

Abstract

Alongside climate change, air pollution is one of the most concerning public health topics of the 21st century. Statistics estimate that more than seven million people die from air pollution yearly, especially in low- and middle-income countries, where people suffer from the highest exposure. Inhalable micrometer particulate matter (PM2.5 and P M10), nitrogen dioxide (NO2), ozone (O3), sulfur dioxide (SO2), and carbon monoxide (CO) are the most common pollutants, permanently monitored by World Health Organization (WHO). The six pollutants mentioned above are the main causes of a few million premature deaths annually and NO2 is one of the most important pollutants in the eye of public health. WHO’s new guidelines recommend an NO2 average level that should not exceed 107 ppb hourly, and 5 ppb annually. This thesis tackles the NO2 monitoring problem by employing carbon nanotube field-effect transistors (CNT-FETs) as sensing elements, hence extending the ubiquitous Internet-of-Things (IoT) applications, i.e., novel air quality monitoring systems. The first prototype design proposes an embedded system that can interface up to four CNT-FETs and may expand the IoT domain for environmental and lifestyle applications. The platform performance is demonstrated using a CNT-FET nanosensor, exposed to NO2 gas concentrations from 200 ppb down to 1 ppb. The sensor signals are measured for NO2 concentrations as low as 1 ppb, achieving a 3σ limit of detection (LOD) of 23 ppb with an R2 linear fit coefficient of 0.95. Although this prototype offers custom configuration, i.e., range, bandwidth, sampling rate, acquisition time intervals, SD card, and Bluetooth Low Energy (BLE) connection, its average power consumption of 64.5 mW is relatively high. Increasing the system’s power efficiency and downscaling its physical dimensions towards a fully integrated circuit (IC) is highly desired. Consequently, this thesis further presents the concept, design, and realization of an integrated signal acquisition system as a second prototype. The research advances a front-end mixed-signal solution composed of an adjustable CNT-FET voltage bias with a 28 mV step, which controls a regulated cascode for the current mode readout. This stage is followed by a transimpedance amplifier (TIA) with 109 dBOhm gain, 0.75 pA/√Hz noise, 4 μA input range with noise filtering included, and differential output. The latter is connected to a 9-bit SAR ADC of 91.7 fJ/conv. The design is realized in 180 nm CMOS technology and occupies a silicon area of 0.18 mm2. When supplied at 1.8 V, the system consumes an average power of 4.04 μW at an ADC sampling rate of 2.66 kSps and 200 ppb of NO2 gas concentration. The CNT-FET nanosensor connected to the proposed IC demonstrates NO2 gas concentration measurements from 0 ppb to 200 ppb in humid air, i.e., 50 % R.H. Lab measurement results shows that the full acquisition system achieves a 3σ LOD of 18.5 ppb with an R2 of 0.8. The integration of the IC and the CNT-FET as a part of a wireless sensor node, consuming an estimated power of 378 μW , is currently under development. This research presents the development of the prototypes and the demonstration of battery-powered air quality monitoring systems. Still, no comprehensive study characterizing the gas sensor was carried out in this context.