System for performance evaluation and calibration of low-cost gas sensors applied to air quality monitoring

The increasing emissions of air pollutants have been a current issue and with great interest by regulatory agencies, industries, and society in general as they represent a threat to health and to the environment. The monitoring using low-cost atmospheric sensors has been an emerging alternative with great prospect of use to complement the data obtained by traditional air quality monitoring stations. However, the data obtained by these low-cost sensors are still questionable and can often be unreliable. The objective of this work was to build a system for calibration and evaluation of performance parameters of low-cost monitor comparing their results to reference analyzers. In this sense, three low-cost sensors for CO, O-3 and SO2 of a monitor were calibrated and evaluated against the performance parameters defined by baseline drift, linearity, precision, bias, accuracy, response time, interference of pollutant gases and long-term drift. There was no significant deviation from baseline drift for any sensor. All sensors showed a linear profile (R-ADJ(2) similar to 1), good precision (>95%), and a response time of less than 2 min. After calibration, satisfactory accuracy (>89%) and bias (close to zero) data were obtained. The SO2 sensor showed signal disturbances in the presence of different concentrations of CO. Long-term drift suggests a calibration periodicity of less than 60 days. To use these sensors to monitor air quality, prior calibration is necessary, and the determination of performance parameters is suggested.

Automated summary

The increasing emissions of air pollutants have raised concerns over health and the environment. Low-cost atmospheric sensors are emerging as an alternative to traditional air quality monitoring stations. However, the reliability of data obtained by these sensors is still questionable. This study aimed to calibrate and evaluate the performance of low-cost sensors, comparing their results to reference analyzers. Satisfactory accuracy and minimal bias were achieved after calibration.