key: cord-0101830-77lv31hs
authors: He, Jiayang; Huang, Ching-Hsuan; Yuan, Nanhsun; Austin, Elena; Seto, Edmund; Novosselov, Igor
title: Network of Low-cost Air Quality Sensor for Monitoring Indoor, Outdoor, and Personal PM2.5 Exposure in Seattle during the 2020 Wildfire Season
date: 2022-03-26
journal: nan
DOI: nan
sha: 6153ebaebf5c8e184e143d0f517a2771841d23e6
doc_id: 101830
cord_uid: 77lv31hs

The increased frequency of wildfires in the Western United States has raised public concerns. Exposure to wildfire smoke has been linked to an increased risk of cancer and cardiorespiratory morbidity. Evidence-driven interventions can alleviate the adverse health impact of wildfire smoke. Public health guidance during wildfires is based on regional air quality data with limited spatiotemporal resolution. We demonstrate the use of a network of low-cost particulate matter (PM) sensors to gather indoor, outdoor, and personal PM2.5 exposure data from seven locations in the urban Seattle area, along with a personal exposure monitor worn by a resident living in one of these locations during the 2020 Washington wildfire event. The data were used to determine PM concentration indoor/outdoor (I/O) ratios, PM reduction, and personal exposure levels. The result shows that locations equipped with high-efficiency particulate air (HEPA) filters and HVAC filtration systems had significantly lower I/O ratios (median I/O = 0.43) than those without air filtration (median I/O = 0.82). The median PM2.5 reduction for the locations with HEPA is 58 % compared to 20% for the locations without HEPA. The outdoor PM sensors showed a high correlation to the nearby regional air quality monitoring stations (R2 = 0.93). The personal monitor showed high variance in PM measurements as the user moved through different microenvironments and could not be fully characterized by the network of indoor or outdoor monitors. The findings imply evidence-based interventions can be developed for reducing pollution exposure based on the combination of indoor, outdoor sensors. Personal exposure monitoring in individuals' breathing zones provided the highest fidelity data capturing temporal spikes in PM exposure.

Climate-change-related wildfires have become more frequent and intense in the Western United States. Summer wildfire seasons are 40 to 80 days longer on average than they were 30 years ago [1] . Evidence suggests that California and other Western states will likely see everworsening fires for the coming century due to climate change and land management practices [2] [3] [4] [5] . The intensified wildfires will release more smoke into the atmosphere [6] , traveling significant distances [7] . Fine particulate matter (PM2.5), a major pollutant found in smoke from wildfires, can travel deep into the respiratory tract [8] . The combustion-generated aerosols consist of elemental carbon and organic carbon fraction, which may be more toxic than other PM2.5 sources and may have long-lasting impacts on health [9] [10] [11] [12] . Complex flow structures associated with large-scale flames and low flame temperature in biomass burning lead to low carbonization of organic carbon, thus --high levels of potentially carcinogenic polycyclic aromatic compounds [13] [14] [15] [16] . Exposure to PM2.5, particularly combustion-generated aerosols, has been linked to adverse respiratory and cardiovascular health effects, including ischemic heart disease, stroke, cardiovascular mortality, and exacerbations of asthma and chronic obstructive pulmonary disease [17] [18] [19] [20] [21] . More recently the wildfire PM exposures have been linked to higher severity and mortality of SARS-CoV-2 [22] [23] [24] [25] .

A series of large wildfires impacted air quality in western regions of the United States in 2020. The episode measured in this study (2020 Washington Labor Day fires) began on September 7, 2020, and were 90% contained by September 22. The fires burned over 41,000 acres of the forest [26] . Due to the SARS-CoV-2 pandemic shelter-in-place order by Washington state since early 2020, people probably spent a significant amount of time indoor during the 2020 wildfire season. The public health advice for protection from wildfire smoke exposure is to stay indoors, preferably in a "clean room" with filtered air, closed windows and doors, and minimize physical exertion. However, studies have shown that PM2.5 could penetrate indoors even with all the windows and doors closed [27, 28] . With limited access to portable air cleaners during wildfires and increasing awareness of the health impact from wildfire smoke exposure, monitoring indoor PM2.5 is critical to estimate household members' wildfire smoke exposure. Failure to assess the exposure to wildfire smoke could lead to misclassification of exposure in future epidemiology studies and have important public health implications for targeting smoke reduction interventions.

The opportunity exists to improve personal exposure assessment and design individualized intervention strategies that would significantly reduce the adverse impact of PM pollution on human health, including the severity and mortality of Covid-19 cases [29, 30] .

Recent advancements in low-cost particulate matter (PM) sensors led to their extensive use in various applications, such as air quality (AQ) monitoring in indoor [31] [32] [33] [34] and outdoor [35] [36] [37] [38] [39] environments, including large-scale deployments [40] [41] [42] [43] . Optical PM sensors rely on elastic light scattering providing size-resolved PM concentrations in the 0.3 -10.0 μm range. The low-cost sensor measurements may suffer from sensor-to-sensor variability due to a lack of quality control and differences between individual components. [44, 45] The scattering light intensity depends on particle size, morphology, complex index of refraction (CRI), and sensor geometry. [46] CRI sensitivity can be addressed by optimizing the design to measure scattered light at multiple angles simultaneously or by employing dual-wavelength techniques. [47, 48] However, these solutions are complex and involve expensive components that are not suitable for compact, low-cost devices. [31] Environmental conditions were reported to affect sensor output, e.g., a non-linear response has been reported with increasing RH. [49] [50] [51] [52] [53] High humidity (RH > 75%) creates challenges for particle instruments; e.g., significant variations were observed between different commercially available instruments, such as Nova PM sensor [49] and personal DataRAM. [51] In addition, the RH measurement approach could also affect the sensor output [49, 50] , e.g., the RH measurement based on reference monitoring site rather than inside the sensor enclosure may be different due to the microenvironment and transient effects. The selection of reference instruments with different measuring principles may also influence the calibration of low-cost sensors. For example, the calibration of the Plantower PM sensor in Jayaratne et al., 2018 was based on the tapered element oscillating microbalance (TEOM), while Zusman et al., 2020 calibrated the same sensor against the beta attenuation monitor (BAM) and federal reference method (FRM) measurements [50, 54] .

The integrated mass measurements cannot account for temporal particle size and concentration variation during the calibration experiment. The instruments that directly measure aerosol size and concentration, such as aerodynamic particle sizer (APS), can be a better fit for sensor calibration. [44, 55] As low-cost sensors find applications in pollution monitoring, various studies have evaluated the performance of low-cost PM sensors in laboratory and field settings. [44, 54, [56] [57] [58] [59] [60] [61] These reports show that low-cost sensors yield usable data when calibrated against research-grade reference instruments. [42, 62, 63] The low-cost sensor networks have the potential to provide high spatial and temporal resolution, identifying pollution sources and hotspots, which in turn can lead to the development of intervention strategies for exposure assessment and intervention strategies for susceptible individuals. Time-resolved exposure data from wearable monitors can be used to assess individual exposure in near real-time [64] .

This study utilized a network of indoor, outdoor, and wearable low-cost air quality sensors to evaluate 1) the effectiveness of intervention strategies used in different households in terms of PM2.5 I/O ratios and PM2.5 reduction during the 2020 Washington wildfire in seven locations, including residential and office buildings; 2) estimation of personal exposure by (i) wearable sensor and (ii) a combination of indoor and outdoor monitors, where the fraction of personal exposure from different microenvironments is determined based on the Global Positioning System (GPS) and time-resolved PM sensor data.

The monitor used in this study consists of a PM sensor, a temperature/humidity/pressure sensor, a GPS module (ublox SAM-M8Q), and a display. The PM sensor (Plantower PMS A003, Beijing Plantower Co., Ltd, China; referred to as PMS hereafter) is an optical scattering-based sensor with a photodiode positioned normal to the excitation beam. The scattering light intensity is converted to a voltage signal to estimate PM number concentration and mass concentration using a proprietary calibration algorithm. The PMS provides estimated particle counts in six size bins with the optical diameter in 0.3-10 μm (#/0.1L) range and mass concentration (μg/m 3 ) for PM1, PM2.5, and PM10. The mass concentrations can be set to "standard" and "atmospheric", altering the assumed particle density. The "standard" condition is designed to be used in industrial settings, whereas the "atmospheric" condition best measures particles in the ambient environment. When collecting the data, the "ATM" setting for PM2.5 concentration was used; the sampling interval was set to be 10 seconds. In the analysis, our own calibration was developed. This was motivated by a recent report showing that including PM-specific terms (such as PM CRI and density) in the calibration algorithm improves the correlation between the data from Plantower PMS A003 sensor with reference particle counter for both number density and mass concentration. In contrast, the RH term did not improve calibration in the range of RH=17-80%. [62] Thus, the specific calibration for this wildfire event was developed and used in the analysis. The same device was also used as the wearable persona monitor. GPS data were used to coordinate the personal data to a specific location and attribute the PM exposures to the user's microenvironment. 

The monitors were deployed in seven urban Seattle locations (see Figure A1 ). Each location had one outdoor sensor and at least one indoor sensor. The L2 location had two indoor sensors. One user from the sampling site L2 wore an additional personal monitor for the duration of the study. The study covered the wildfire episode between September 10 and September 21, 2020. The sampling sites included two University of Washington (UW) buildings and five residences in Seattle. Data from the nearby Puget Sound Clean Air Agency (PSCAA) regional stations were used for the sensor calibration. Before the study, information about the sites such as housing type, size, HVAC, primary indoor PM sources, and locations of the sensors were collected (see Table 1 ). Three sites (L1-L3) had portable air purifiers or built-in high-efficiency air filtration in HVAC systems. Location 2 (L2) had two indoor monitors in separate rooms. Both UW buildings (L6 and L7) were largely unoccupied due to the shelter-in-place order and were not equipped with HEPA filtration units at the time. The residents at the sampling sites were not given specific instructions on whether to keep the windows and doors open or closed but were asked about this after the sampling was completed. The research staff performed the sampling at the UW buildings, and the windows were kept closed during sampling. 

The low-cost sensor data were corrected against the average of PM2.5 from the two nearby PSCAA regional monitors. The correction model was generated using a data subset from the outdoor sensor outside a UW building (L7) during the wildfire event. L7 was largely unoccupied during the wildfire with minimum local activities that may influence PM2.5 measurement compared to the other sampling sites. L7 is located at least 200 meters from any major traffic arterials. Sensor to sensor difference was within 10%, as shown in our previous aerosol chamber experiments [65] .

Informed by our previous PMS sensor calibration study, a linear model or a quadratic were tested to fit using the outdoor sensor data from L7:

where 0 , 1 , and 2 are the regression coefficients, is the hourly reference PM2.5

concentrations from the nearby PSCAA monitoring stations, and and 2 are the linear and quadratic coefficients of the raw PM2.5 data from the sensor, respectively. The fits with zero intercept ( 0 = 0 ) and non-zero intercept ( 0 ≠ 0 ) were tested. The Bayesian Information Criterion (BIC) was used to select the optimal calibration model.

The time-resolved PM concentration I/O ratio was calculated to assess the smoke infiltration. We conducted the Wilcoxon signed-rank tests (for paired comparison) to compare the I/O ratio during the wildfire to the I/O ratio post the wildfire. To assess the reduction in PM levels,

we compared indoor and outdoor time-resolved PM concentrations. We calculated PM2.5 reduction for each site and the personal exposure as:

where is the average outdoor PM2.5 concentration (µg/m 3 ) during the wildfire and is the average indoor or personal PM2.5 concentration (µg/m 3 ).

To understand the contribution of each microenvironment to personal exposure, we attributed the user's daily PM2.5 exposures in each location using GPS data with a 2.5m horizontal accuracy from the wearable monitor. The personal exposure attribution was done using Python 3.7.1. The raw PM2.5 data were first aggregated into 10-min averages to reduce the data size without losing significant spatial resolution. The geocoordinates, recorded in conjunction with the PM2.5 concentration, were grouped into three categories where the user spent most of the time:

home, office, and other. We defined the buffer zones encircling the residence and the office locations with a 10-meter clearance to minimize misclassification error caused by the GPS drift.

The home and office geolocation data were visually confirmed on a map for each occurrence when the user with the wearable monitor was at the location. Records collected outside these two buffer zones were classified as "other locations". Then the PM2.5 exposures attributed to each microenvironment for each day was calculated as:

where represents the attributable exposures of microenvironment to the total personal exposure on day ; is the hourly average PM2.5 concentrations (µg/m 3 ) of microenvironment on day ; and is the fraction of time spent in microenvironment on day .

The data from L7 outdoor sensor shows a good agreement with the regional monitors with the pre-calibration R 2 = 0.92. The linear model shows the lower root-mean-square-error (RMSE) and BIC, with the overall RMSE improved from 18.47 μg/m 3 to 14.35 μg/m 3 against the regional monitors with the post-calibration R 2 = 0.94 (see Figure A2 ). The quadratic model did not result in a significant improvement; thus, the data from all the other sensors were corrected with the linear model. Note that while the PM2.5 correction agreed with reference monitors using the linear model, the particle number density data from PMS were not evaluated for wildfire smoke. The particle size-dependent data from low-cost optical sensors are significantly influenced by the particle index of refraction, particle morphology, particle loading, sensor geometry, and environmental conditions [44, 46, 57, [66] [67] [68] . The locations with active air filtration had a significantly lower PM concentration, while locations without HEPA filters had only slightly lower PM levels. Occasional spikes in PM concentration, above the already high baseline, were observed due to cooking activities. Note that even with the active PM control strategies implemented in several households, the average indoor PM2.5 is still much higher than the typical (non-wildfire season) outdoor PM2.5 levels (< 10 μg/m 3 )

in this region [69, 70] . 

We present a case-by-case analysis to provide insight into the effectiveness of aerosol mitigation strategies and the significance of the quantity and placement of sensors within the residences. Figure 3 shows the 1-hour average indoor, and outdoor data for all sites with the averaged PM2.5 from the two reference monitors in green color for comparison.

Overall, the outdoor sensors are in close agreement with PSCAA data. In some cases, the outdoor sensors reported spikes in PM concentration likely due to the local activities near monitoring sites, which the PSCAA monitors could not detect (see Figure 3 ). These differences can be explained by (i) spikes in the local PM concentration or (ii) mismatch in sensor sampling rate as the network sensors sampling interval was set to ~ 10 seconds, while the reference monitors'

reporting interval was 1 hour. Though both scenarios are possible, the PM levels difference between the locations suggests that some differences in outdoor PM2.5 concentration were driven by local events that the regional monitors could not capture.

Only a moderate reduction in PM levels (16-29%) was observed when low-grade PM filers were used at L5, L6, and L7. Table 2 lists the mean and maximum of PM2.5 for each sampling site.

PM2.5 reduction levels were calculated for each sampling location to indicate the effectiveness of mitigation strategies for households. Location L1 had the indoor monitor placed in a relatively small room (home-office ~150 ft 2 ) with a high-volume HEPA filter for the entire wildfire episode. This relatively small "clean room" environment strategy resulted in the study's lowest median I/O ~ 0.2 (see Table A1 ).

However, data for other locations (e.g., bedrooms, living room) within the residence are not available, which is problematic for assessing personal exposure. The I/O comparison is shown in Figure 4 .

The residents of the L2 residence had the two monitors in separate rooms: L2a was placed in a larger living room (~350 ft 2 ) and L2b -in the adjacent home office (120 ft 2 ). Two portable HEPA filtration units were used: one in the living room, and the other was moved from the office to the bedroom at nighttime. The data from the living room has relatively low variance; however, the larger room was not cleaned as effectively as the smaller home office or bedroom, I/O ratio stayed relatively constant ~0.5. When the filtration unit was positioned in the office, the I/O ratios dropped to ~ 0.4. When the filter was moved from the home office to the bedroom, the PM level in the office increased to the level of the adjacent living room. The bedroom was not monitored by a fixed sensor; however, the resident's wearable sensor recorded a significant PM reduction in the bedroom during the night with the I/O ratio close to 0.1 (see Figure 5 ).

The L3 site had an electrostatic precipitator installed in the HVAC system. The HVAC system was controlled by a thermostat, which explains the periodic pattern of the PM concentration. The PM concentration (measured in the bedroom) dipped during the daytime when the forced air HVAC system was ON and went up during the nighttime when the HVAC was OFF.

L4 indoor monitor was placed in the kitchen where it detected the spikes from cooking activities in addition to the high background level from wildfire smoke intrusion. L4 residents kept

their kitchen windows open during the wildfires, which explains the highest background indoor PM level among all the sampling sites. L4 outdoor sensor stopped working three days after the deployment because it was accidentally unplugged from the AC outlet. I/O ratio for L4 was calculated using the data collected before September 13. The sampling site L4 without air filtration and closed windows had only a 13% indoor-outdoor difference.

The L5 residence did not have a central HVAC system. A portable AC unit (AeonAir Model #RPAC08EE) with a low-grade PM filter was used by the residents during the wildfire.

The window in residence was closed for the entire duration of sampling (self-reported). L5 monitor was placed in the apartment's kitchen/living room area (200 ft2). PM concentration was lower than the outdoor level during the wildfire but higher than other residences with air filtration units. L6

and L7 are two UW buildings with HVAC systems but low-grade filtration units. The data is similar to the L5 residential site. The buildings were largely unoccupied during the wildfire due to COVID lockdown, and the windows were closed. The indoor PM2.5 at L6 and L7 are lower than outdoor. The I/O ratios were similar to L5 (~0.7-0.8). Only a moderate difference in PM levels (16-29%) was observed when low-grade PM filers were used at L5, L6, and L7.

Interestingly, sites L3 and L7 with central filtration units show similar trends. PM concentration dipped rapidly when the HVAC was turned ON (controlled by a thermostat). As the central HVAC unit was OFF, the PM concentration climbed up due to infiltration of smoke.

Though the analysis of HVAC performance is beyond the scope of this manuscript, these data can be used to design and optimize HVAC performance, such as the use of economizers, sensor-based controls, filter upgrades, etc. 

To assess personal exposure as a function of the microenvironment, we compared the personal data measured by the wearable monitor with the wearer's home-based monitor data. Figure 5 shows the 10-minute average of PM2.5 concentration measured by the user's personal and home-based monitors for reference. The wearable sensor PM2.5 data showed a 68% PM2.5 reduction, which is lower than the PM2.5 reduction estimated using the wearer's indoor monitors.

The shaded areas mark the nighttime (10:00 pm -6:00 am). The personal monitor recorded significantly lower PM2.5 concentration at night compared to the other two home-based monitors located in the living room and the home office, which can be explained by the colocation of the HEPA air cleaner and wearable monitor. The user is an aerosol researcher who monitored his exposure during the wildfire. The difference in the personal exposures, measured by the wearable and the home-based monitor, indicates that access to special hyperlocal resolution (Indoor -room level, Outdoor -2.5 meters, Wearable sensor level -2.5m with 10 m buffer) for PM concentration data can enhance the efficacy of PM exposure interventions. We also apportioned the user's exposure based on the GPS data. Figure 6a shows the 10minute average of PM2.5 concentration measured by the personal sensor color-coded based on the microenvironment. The user spent most of his time at home and with 15% time in the office and 9% in other locations. Exposure outside the home and the office was categorized into other locations. The wearable monitor recorded higher PM2.5 levels in the user's workplace and other locations than at the residence. The work location did not have an efficient filtration system based on the geolocation and PM concentration data. Figure 6b shows the weighted daily average personal exposure in different microenvironments. The personal exposure contribution from the office and other locations was 36% of the total smoke exposure during the wildfire, while the time spent in these environments was 24% (see Figure A3 ). The attribution of personal exposure to microenvironments can help personal activities during wildfire seasons. a b 

This study demonstrates the application of a low-cost sensor network for air quality monitoring during the 2020 Washington wildfire event. The outdoor PM2.5 data from the sensor network had an excellent agreement with the nearby PSCAA regional monitors. The spatial variance for PM2.5 in the urban area was low during the wildfire event. Our results showed that during the 2020 Washington wildfire, the outdoor PM2.5 level was as high as >200 μg/m 3 in the Seattle area. Using portable HEPA air cleaner was associated with lower indoor PM2.5 levels during the wildfire episode, with the PM2.5 reduction of 50 -77 % among the sampling sites.

However, the observed levels were still higher than the typical Seattle outdoor PM levels (< 10 μg/m 3 ). The I/O ratio was driven by the smoke infiltration and the quality of air filtration. The personal monitoring results highlighted the influence of microenvironments on an individual's exposure to PM2.5. Although this study had a relatively small sample size, it demonstrated that personal action, such as staying indoors and using HEPA air cleaner, can reduce personal exposure to wildfire smoke. The personal exposure analysis suggests knowledge about PM levels can lead to a reduction in exposure. More extensive studies and a collection of time-activity information are warranted to investigate the source of PM2.5 exposure and the health impact of PM exposure.

Climate-induced variations in global wildfire danger from 1979 to 2013

Does hot and dry equal more wildfire? Contrasting short-and long-term climate effects on fire in the Sierra Nevada

Climate change presents increased potential for very large fires in the contiguous United States

Impacts of climate change from 2000 to 2050 on wildfire activity and carbonaceous aerosol concentrations in the western United States

The changing risk and burden of wildfire in the United States

Ensemble projections of wildfire activity and carbonaceous aerosol concentrations over the western United States in the mid-21st century

Assessment of PM2. 5 and PM10 over Guwahati in Brahmaputra River Valley: Temporal evolution, source apportionment and meteorological dependence

The changing risk and burden of wildfire in the United States

Do 16 polycyclic aromatic hydrocarbons represent PAH air toxicity? Toxics

Wildfire smoke impacts respiratory health more than fine particles from other sources: observational evidence from Southern California

Differential Cardiopulmonary Health Impacts of Local and Long-Range Transport of Wildfire Smoke

The delayed effect of wildfire season particulate matter on subsequent influenza season in a mountain west region of the USA

Molecular Composition and the Optical Properties of Brown Carbon Generated by the Ethane Flame

Excitation-Emission Matrix Spectroscopy for Analysis of Chemical Composition of Combustion Generated Particulate Matter. Environmental Science & Technology

Nanostructure transition of young soot aggregates to mature soot aggregates in diluted diffusion flames. Carbon

Soot morphology and nanostructure in complex flame flow patterns via secondary particle surface growth

Cardiopulmonary Effects of Fine Particulate Matter Exposure among Older Adults, during Wildfire and Non-Wildfire Periods, in the United States

Health Impact Assessment of the 2020 Washington State Wildfire Smoke Episode: Excess Health Burden Attributable to Increased PM2.5 Exposures and Potential Exposure Reductions

The effects of bushfire smoke on respiratory health

Critical review of health impacts of wildfire smoke exposure

Health impact analysis of PM2. 5 from wildfire smoke in Canada

Effect of environmental pollutants PM-2.5, carbon monoxide, and ozone on the incidence and mortality of SARS-COV-2 infection in ten wildfire affected counties in California

The association of wildfire air pollution with COVID-19 incidence in New South Wales

SARS-CoV-2 test positivity rate in Reno, Nevada: association with PM2. 5 during the 2020 wildfire smoke events in the western United States

Wildland firefighter exposure to smoke and COVID-19: A new risk on the fire line. Science of The Total Environment

National Large Incident Year-to-Date Report

Investigation and modeling of the residential infiltration of fine particulate matter in Beijing

Estimation of Outdoor PM2.5 Infiltration into Multifamily Homes Depending on Building Characteristics Using Regression Models. Sustainability

Personal Interventions to Reduce Exposure to Outdoor Air Pollution. Annual Review of Public Health

Forecast-based interventions can reduce the health and economic burden of wildfires. Environmental science & technology

Methodology for Addressing Infectious Aerosol Persistence in Real-Time Using Sensor Network. Sensors

Indoor household particulate matter measurements using a network of low-cost sensors. Aerosol and Air Quality Research

Spatiotemporal distribution of indoor particulate matter concentration with a low-cost sensor network

Real-time sensors for indoor air monitoring and challenges ahead in deploying them to urban buildings

Use of low-cost particle monitors to calibrate traffic-related air pollutant models in urban areas

Low-cost sensors as an alternative for long-term air quality monitoring

Air quality during and after the Commonwealth Games 2018 in Australia: Multiple benefits of monitoring

Community Air Sensor Network (CAIRSENSE) project: evaluation of lowcost sensor performance in a suburban environment in the southeastern United States

A distributed network of low-cost continuous reading sensors to measure spatiotemporal variations of PM2.5 in Xi'an, China. Environmental Pollution

The rise of low-cost sensing for managing air pollution in cities. Environment international

Integrating low-cost air quality sensor networks with fixed and satellite monitoring systems to study ground-level PM2. 5. Atmospheric Environment

Integrating Fixed Monitoring Systems with Low-Cost Sensors to Create High-Resolution Air Quality Maps for the Northern China Plain Region. ACS Earth and Space Chemistry

Improving data reliability: A quality control practice for low-cost PM2. 5 sensor network. Science of The Total Environment

Laboratory Evaluation of the Shinyei PPD42NS Low-Cost Particulate Matter Sensor

Inter-comparison of low-cost sensors for measuring the mass concentration of occupational aerosols

Design and optimization of a compact low-cost optical particle sizer

LOAC: a small aerosol optical counter/sizer for ground-based and balloon measurements of the size distribution and nature of atmospheric particles-Part 2: First results from balloon and unmanned aerial vehicle flights. Atmospheric Measurement Techniques

Numerical and experimental study of the performance of the dual wavelength optical particle spectrometer (DWOPS)

Performance Assessment of a Low-Cost PM2.5 Sensor for a near Four-Month Period in Oslo

The influence of humidity on the performance of a low-cost air particle mass sensor and the effect of atmospheric fog

Performance evaluation of the active-flow personal DataRAM PM2.5 mass monitor (Thermo Anderson pDR-1200) designed for continuous personal exposure measurements

Field evaluation of a modified DataRAM MIE scattering monitor for real-time PM2.5 mass concentration measurements

Light scattering characteristics of aerosols as a function of relative humidity: Part I -A comparison of measured scattering and aerosol concentrations using the theoretical models

Calibration of low-cost particulate matter sensors: Model development for a multi-city epidemiological study

Laboratory assessment of low-cost PM monitors

Using statistical methods to carry out in field calibrations of low cost air quality sensors

Performance evaluation of twelve low-cost PM2. 5 sensors at an ambient air monitoring site

Ambient and laboratory evaluation of a low-cost particulate matter sensor. Environmental pollution

Long-term field evaluation of the Plantower PMS low-cost particulate matter sensors. Environmental pollution

Laboratory evaluation of low-cost PurpleAir PM monitors and in-field correction using co-located portable filter samplers

Performance of low-cost indoor air quality monitors for PM2. 5 and PM10 from residential sources. Building and Environment

Assessing the value of complex refractive index and particle density for calibration of low-cost particle matter sensor for size-resolved particle count and PM2. 5 measurements. Plos one

Evaluation of nine low-cost-sensor-based particulate matter monitors. Aerosol and Air Quality Research

Usability of a personal air pollution monitor: Design-feedback iterative cycle study

Assessing the value of complex refractive index and particle density for calibration of low-cost particle matter sensor for size-resolved particle count and PM2.5 measurements

Assessing the Value of Complex Refractive Index and Particle Density for Calibration of Low-Cost Particle Matter Sensor for Size-Resolved Particle Count and PM2

Assessing the accuracy of low-cost optical particle sensors using a physics-based approach. Atmospheric measurement techniques

Low-cost PM2. 5 sensors: An assessment of their suitability for various applications. Aerosol and Air Quality Research

Impacts of the COVID-19 responses on traffic-related air pollution in a Northwestern US city. Science of The Total Environment

Impacts of using auto-mode portable air cleaner on indoor PM2.5 levels: An intervention study. Building and Environment

Field measurements of PM2.5 infiltration factor and portable air cleaner effectiveness during wildfire episodes in US residences