A customizable, low-power, wireless, embedded sensing platform for resistive nanoscale sensors

The paper addresses the need for portable, low-power, and customizable gas sensing platforms for improved air quality monitoring. Current solutions, such as gas chromatography, are expensive, bulky, and stationary. The demand for portable solutions is driven by the need for wider spatiotemporal coverage and reduced costs, facilitating community-based monitoring and increased data accessibility. The authors highlight the World Health Organization's (WHO) estimates of seven million premature deaths annually due to poor air quality and the limitations of conventional monitoring methods. The introduction references regulatory requirements and exposure limits for NO₂ from the EU and the US EPA, emphasizing the need for sensors operating within these limits. It also reviews existing commercial solutions, noting their limitations in terms of power consumption (2.5-6W), size, configurability (limited to hardware switches), resolution (8 bits), and lack of local storage or wireless data transfer. The authors emphasize the advantages of nanomaterials (nanowires, graphene, carbon nanotubes) for sensing applications due to their high surface-to-volume ratio, but also acknowledge challenges such as device variation and current decrease over time. This work proposes a versatile embedded system to address these challenges, providing a software-configurable front-end readout for nanosensors and demonstrating its performance with ultra-sensitive CNT nanosensors for NO₂ gas sensing.

The literature review section compares the proposed platform to existing commercial and research-based air quality monitoring systems. Commercial systems are criticized for high power consumption, limited configurability, low resolution, and lack of wireless capabilities and data storage. Research-based systems incorporating nanomaterials are discussed, noting their potential advantages but also highlighting the challenges associated with device variability and performance degradation over time. The review establishes the need for a versatile, low-power, customizable platform that can interface with various nanoscale sensors and transmit data wirelessly. Specific examples of existing systems with their limitations in terms of size, power consumption, sensing capabilities, resolution, and data transfer methods are provided. The focus is on highlighting the gaps that the proposed platform aims to fill.

The proposed platform centers around an ATmega2560 microcontroller, an SD card for data storage, and a Bluetooth Low Energy (BLE) module for wireless communication. The design utilizes a current-mode readout, suitable for resistive nanosensors. The sensor bias block (SBB) uses a 12-bit DAC (MCP4922) for voltage control, with a software-defined bias allowing for dynamic adjustments of measurement parameters. A cascaded charge pump (MAX6604) extends the voltage range to -10 V to +5 V. The sensor signal acquisition (SSA) employs a multichannel current-to-digital converter (CDC) based on time-interleaved integrators with programmable sampling rates and full-scale current ranges. The CDC's full-scale range is dynamically adjusted using two programmable parameters: Tconv (integration time) and Crange (integrator capacitance). An event-triggered Finite State Machine (FSM) on the microcontroller automates sensing routines, allowing for customizable bias schemes and data logging. The FSM's operation is defined by a CSV file stored on the SD card, controlling bias voltage amplitude, duration, CDC configuration, data storage, and BLE transmission. The platform is characterized using a sealed test chamber to control gas exposure, and the results are demonstrated using a smartphone connected via BLE. The characterization focuses on the full-scale range of the CDC, bandwidth, noise, and bias block performance.

The platform achieves a full-scale current range of 1.5 nA to 7.2 µA with 20-bit resolution and a variable sampling rate of up to 3.125 kSPS. The bipolar voltage is programmable from -10V to +5V with 3.65mV resolution. The average power consumption is 64.5 mW for a measurement protocol of three samples per second, including BLE advertisement. The noise floor is characterized, showing an input-referred current RMS noise around 1.15 nA. The platform's performance is demonstrated using a CNT nanosensor for NO₂ gas detection, achieving measurements down to 1 ppb, although the 3σ limit of detection is 23 ppb (1σ: 7 ppb). The system's versatility, low power consumption, and customizable features are highlighted as key advantages. The detailed characterization of the platform’s components, including the full-scale range, bandwidth, noise, and bias generation capabilities, is provided. The use of a custom-designed FSM allows for flexible and automated operation with various nanosensors. The results demonstrate the successful integration of the platform with a CNT-based NO₂ sensor, showing its capability to measure gas concentrations within the regulatory limits for ambient air quality monitoring.

The findings demonstrate the successful development and characterization of a highly versatile and low-power embedded sensing platform for resistive nanoscale sensors. The platform's ability to operate within the EU's MAK limits for NO₂ is a significant achievement, addressing a critical need in air quality monitoring. The use of a software-configurable FSM allows for a high degree of customization, making it suitable for various nanosensors and applications. The low power consumption is crucial for portable and battery-operated devices. The platform's ability to measure NO₂ concentrations down to 1 ppb, while having a 3σ LOD of 23 ppb, shows its potential for high-sensitivity gas sensing. Future work could focus on integrating more advanced sensor signal processing algorithms and exploring the application of the platform to other types of nanosensors and environmental monitoring.

The paper concludes by summarizing the key contributions: a customizable, low-power wireless sensing platform capable of interfacing with various resistive nanoscale sensors. The platform's successful operation within EU regulatory limits for NO₂ detection is highlighted. Future research directions could include the development of more advanced signal processing algorithms, investigation of different nanosensors, and expansion of the platform's capabilities for broader environmental monitoring applications.

The current study's limitations include the 3σ limit of detection (23 ppb) for NO₂ being higher than the ideal detection limit. While the platform can measure signals at 1 ppb, further improvements in signal processing or sensor design could lead to lower detection limits. The characterization primarily focuses on NO₂ sensing with CNT nanosensors; additional validation with other nanosensors and target analytes is necessary to fully assess the platform’s versatility. The platform's power consumption, while low compared to existing solutions, can still be further optimized. This may involve investigating other components with lower power consumption and exploring more energy-efficient data transmission methods.