Fundamentals of low-cost aerosol sensor design and operation
James Ouimette,W Patrick Arnott,Philip Laven,Richard Whitwell,Nagarajan Radhakrishnan,Suresh Dhaniyala,Michael Sandink,Jessica Tryner,John Volckens
DOI: https://doi.org/10.1080/02786826.2023.2285935
Abstract:Most evaluations of low-cost aerosol sensors have focused on their measurement bias compared to regulatory monitors. Few evaluations have applied fundamental principles of aerosol science to increase our understanding of how such sensors work and could be improved. We examined the Plantower PMS5003 sensor's internal geometry, laser properties, photodiode responses, microprocessor output, flow rates, and response to mono- and poly-disperse aerosols. We developed a physics-based model of particle light scattering within the sensor, which we used to predict counting and sizing efficiency for 0.30 to 10 μm particles. We found that the PMS5003 counts single particle scattering events, acting like an imperfect optical particle counter, rather than a nephelometer. As particle flow is not focused into the core of the laser beam, >99% of particles that flow through the PMS5003 miss the laser, and those that intercept the laser usually miss the focal point and are subsequently undersized, resulting in erroneous size distribution data. Our model predictions of PMS5003 response to varying particle diameters, aerosol compositions, and relative humidity were consistent with laboratory data. Computational fluid dynamics simulations of the PurpleAir monitor housing showed that for wind-speeds less than 3 m s-1, fine and coarse particles were representatively aspired to the PMS5003 inlet. Our measurements and models explain why the PurpleAir overstates regulatory PM2.5 in some locations but not others; why the PurpleAir PM10 is unresponsive to windblown dust; and why it reports a similar particle size distribution for coarse particles as it does for smoke and ambient background aerosol.