Item

Abstract

Air pollution remains a critical global health challenge, with nitrogen dioxide (NO₂) playing a significant role in adverse health outcomes. Low-cost sensors (LCS) offer promising opportunities for accessible and high-resolution air quality monitoring but face scrutiny over their accuracy and reliability. This study evaluates the performance of electrochemical LCS for NO₂ measurement in comparison to high-precision reference instruments—cavity attenuated phase shift (CAPS) and chemiluminescence NO₂ monitors—at eleven temporal resolutions (between 10-s and 6-h). Using three EarthSense Zephyrs containing electrochemical sensors, data were collected over six months at an urban-traffic air quality monitoring site in Berlin. Sensor performance was assessed based on statistical metrics, including R2, relative error, and mean bias error (MBE). Results revealed that LCS exhibit good agreement with reference instruments at coarse time resolutions (≥1-h averages, R2 > 0.8), but accuracy diminishes significantly at higher resolutions (<1-min, R2 < 0.5). Overall, LCS perform better when trained against CAPS monitors than against chemiluminescence monitors. This performance is largely influenced by chemistry and emissions, with poorer performance during the daytime than at night, a pattern which is exacerbated at high time resolutions. CAPS-calibrated predictive models outperform those calibrated against chemiluminescence monitors in capturing short-term concentration peaks. These findings suggest that while LCS are suitable for coarse-resolution measurements of NO2, their limitations in high temporal resolution dynamic environments pose significant challenges for their use in exposure studies and mobile measurements. Recommendations for improved LCS deployment include careful calibration, strategic experimental design, and focused application on lower time-resolution monitoring.