AAAR 37th Annual Conference October 14 - October 18, 2019 Oregon Convention Center Portland, Oregon, USA
Abstract View
Evaluation of Low-Cost Particulate Matter Sensors in a Test House
JONATHAN GINGRICH, Sameer Patel, Elizabeth Graham, Erin K. Boedicker, Delphine K. Farmer, Richard Corsi, Marina Vance, University of Colorado Boulder
Abstract Number: 699 Working Group: Air Quality Sensors: Low-cost != Low Complexity
Abstract Exposure to particulate matter (PM) causes a wide spectrum of health effects1. There are many sources of indoor PM resulting in potentially high exposures to humans. Because of this and because humans spend most of their times indoors, indoor concentration measurements are important for assessing potential exposure and resulting health effects. Recently, low-cost air quality instruments marketed for consumers have become popular for measuring indoor air quality and contaminant levels by homeowners. We tested five popular commercially-available indoor air quality instruments (Air Quality Egg, Dylos, PurpleAir, Speck) as well as two prototypes developed by universities (Washington University in St. Louis and Harvard) and compared them to research- and industry-grade monitors (TSI Aerodynamic Particle Sizer (APS 3321), TSI DustTrak 8520, and Handix Scientific Portable Optical Particle Sizer (POPS)) in a three bedroom manufactured test house during the HOMEChem (House Observations of Microbial and Environmental Chemistry) campaign.
Five different categories were used to test: (1) accuracy during collocation, (2) time resolution to peaks, (3) accuracy at baseline, (4) accuracy during peaks, and (5) spatial variability of responses. In general, sensors performed well at baseline. PurpleAir and Air Quality Egg agreed best with the reference monitors during collocation. At high concentrations, almost all sensors overestimated PM concentrations compared to the reference monitors. Time response varied greatly based on the resolution of the monitors. Finally, for spatial variation, almost all sensors had lower concentrations in the living room than in the kitchen during cooking events, showing the spatial variability of PM during these events. While low-cost optical sensors show some ability to respond to peaks and are close to reporting an accurate PM concentration at the baseline, their considerable overestimation during peaks show that more development is needed before they can be considered accurate and robust sensors for consumers.
[1] Anderson, et al., J. Med. Toxicol. 8 (2012): 166-175.