Extraction of dust collected in HVAC filters for quantitative filter forensics

Abstract

The analysis of dust collected on the filters installed in the heating, ventilation, and airconditioning (HVAC) systems, filter forensics, is a useful approach to explore concentration, size distribution, and composition of indoor particles. The extraction of dust from filters represents one of the biggest challenges to obtain accurate results from filter forensics. Although vacuuming is one of the most common dust extraction techniques, it is unclear how efficient it is and whether it provides a representative sample in terms of particle size. In this article, we used a high-capacity vacuum sampler to extract dust from 20 filters artificially loaded with well-characterized test dust as well as from 41 filters naturally loaded in residential HVAC systems. After all extractions, we recovered 0.1-5.5 g and 0.02-11.4 g of dust from the artificially and naturally loaded filters, respectively. These ranges were equivalent to 11.3-52.2% and 1.8-72.9% recovery efficiency, the ratio of dust recovered to the dust loaded in the filters. Multiple extractions were found to be an effective strategy to add to the recovery to enable filter forensics for the detection of multiple analytes. The recovered samples were slightly over-representative of particles greater than 10 mm. Therefore, caution should be taken when applying filter forensics for studying contaminants associated with smaller particles.

Quantitative filter forensics: Size distribution and particulate matter concentrations in residential buildings

“…Particularly, we observed that the vacuuming extraction technique that we developed underrepresented particles <10 µm in the recovered dust from HVAC filters due to the loss of particles during the extraction and/or the strong binding forces of these particles to the filter fibers. 25 Also, we barely detected particles smaller than approximately 1 µm in the LDPS although the in-situ filter efficiencies for this range were greater than zero. 23 The non-zero in-situ filter efficiencies for particles <1 µm show the existence of these particles in the filters captured from the upstream flow.…”

“…We detected one more peak with 41 samples and two more peaks with two of the samples all at sizes larger than that of the domi- The fractional volume at these sizes was very small, however, compared with the entire volume distribution (eg, <5%). 25 Therefore, the presence of larger particles is likely due to the indoor sources of these large particles rather than filter fibers.…”

“…All the particle size distributions were measured assuming the indices of refraction and absorption of 1.6 and 0.1 (-), respectively, and a density of 2.8 (g/cm 3 ). 25 For comparison purposes, we also measured the size distribution of six additional filter dust and three settled dust samples from homes in semi-rural areas of Central Texas, United States which were collected as part of a different project. Additional explanations regarding dust collection and extraction from the filters in this project are explained elsewhere.…”

“…Other LDPS measurement details are described elsewhere. 25 Briefly, each LDPS measurement was performed for a particle size range of 0.1-3500 µm for a total of 92 channels. The results were normalized (ie, the sum of the volume fractions of all the channels are 100%).…”

Distribution of SARS-CoV-2 RNA Signal in a Home with COVID-19 Positive Occupants

“…The following parameters were used for the calculation: m was the mass of dust (g) collected in the HVAC filter, f was the concentration of SARS-CoV-2 (N2 gene copies/g) in the dust collected on the filter, η was the integrated particulate matter filtration efficiency of the MERV-4 filter, Q was the volumetric air flowrate (m 3 /h) through the filter (median for the summer season for the same geographical location, (Givehchi et al, 2019) and t was the runtime of the HVAC system (h) over the duration of the SARS-CoV-2 collection time -approximately one month-, median for the summer season for the same geographical location, (Givehchi et al, 2019). A few considerations that ought to be taken into account are: (1) our estimate does not account for the attenuation of the signal over time or the losses due to deposition, (2) it considers the estimate that approximately 5.5% of the viral signal is recovered through RNA extraction from the dust matrix (as estimated in this work via the spike and recovery tests for the surrogate BRSV virus), (3) the mass recovered by the participants was estimated to be 68% of that recoverable by trained researcher (based on previous experiments comparing researcher-collected samples to participant-collected samples), and (4) that approximately 27% of the accumulated dust can be recovered from the filter (Mahdavi and Siegel, 2020). Owing to the use of averaged parameters from a similar population of homes and analysis approaches, our estimate should be considered a scaling approach rather than a precise calculation of SARS-CoV-2 airborne concentration.…”