Insights into low-cost pm sensors using size-resolved scattering intensity of cooking aerosols in a test house

Abstract

Spatio-temporal particulate matter (PM) measurements are critical to understanding emission profiles to develop effective mitigation strategies. However, high-resolution spatio-temporal measurements using research-grade instruments are cost-prohibitive, making low-cost sensors (LCS) a promising alternative. LCS measures particle light scattering intensity (I_sc) and converts it to mass and number concentrations. Such measurements are influenced by particles’ physico-optical properties, which make LCS appropriate for measuring PM with similar properties to those used for calibration. This work evaluates the performance of an LCS (PurpleAir) against reference instruments (SMPS + APS) during cooking experiments performed in a test house where PM_0.3/PM₁₀ as high as 90% was observed. LCS demonstrated better correlations (R²) for PM₁ (0.98–1), PM_2.5 (0.96–1), and PM₁₀ (0.88–1) compared to coarser size bins, i.e., PM_1–2.5 (0.04–0.99) and PM_2.5–10 (0.02–0.93). Measurements from different cooking activities show different linear fits. I_sc calculated using size-resolved measurements of reference instruments also had different linear fits with corresponding mass concentrations, even after assuming the same refractive index for all cases, highlighting the dependence of size distribution on LCS measurements. On average, sub-300 nm particles accounted for up to 75 ± 18% of the total I_sc. Such dominance of sub-300 nm particles leads to size misclassification by LCS, whose algorithm assumes the inability to measure sub-300 nm particles. Moreover, the I_sc calculated using the size distribution reported by LCS is considerably lower than that of reference instruments. This mismatch can be attributed to truncated angle error and positioning of particles in the sensing zone relative to the photodiode.

Copyright © 2024 American Association for Aerosol Research

Editor:

• Andrea Ferro

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data is available upon request.

Additional information

Funding

This experimental work was supported by Alfred P. Sloan Foundation, which funded the HOMEChem campaign (G-2017-9944, G-2016-7050, G-2019-11412, and G-2019-12301). The analysis and modeling work was supported by the Science and Engineering Research Board (SERB), Department of Science and Technology, Government of India (Grant# SRG/2021/1001050). Partial funding from the Indian Institute of Technology Gandhinagar is also acknowledged. The first author (AKT) acknowledges the Prime Minister Research Fellowship received from the Ministry of Education, Government of India.