What does your filter say about the air

What does your filter say about the air that you breathe? Federico Noris, Kerry Kinney, and Jeffrey Siegel Department of Civil, Architectural and Environmental Engineering The University of Texas at Austin IGERT: Indoor Environmental Science & Engineering The University of Texas 1

Particle Cycle HVAC Filter η Qr Ex / Infiltration Deposition Tracked-in dust 2

Sampling Approaches • Air – Short-term samples – Spatial variability • Settled dust – Sampling location (tracked-in dust) – Area or mass concentrations – Biased toward large particles • Alternative: HVAC filters? – Widely available – Easily collectable – Integrated (long-term) samples 3

Motivation • Microorganisms & heavy metals cause adverse health effects (Gyntelberg et al. , 1994; Moore, 1990) • Metal concentrations investigated in indoor dust (Adgate et al. , 1998; Oliver et al. , 1999; Turner and Simmonds, 2006) • Microbial concentrations studied in indoor air & settled dust (Andersson et al. , 1999; Bouillard et al. , 2002; Sessa et al. , 2002) • Inconsistent link between indoor microbial concentrations and respiratory symptoms (Peat et al. , 1998; Verhoeff and Burge, 1997) 4

Motivation • Airborne bacterial samples have high temporal variability (Brodie et al. , 2006; Fierer et al. , 2008) • HVAC filters show promise as long-term air samplers for microbial communities (Stanley et al. , 2008; Tringe et al. , 2008) • Majority of studies rely on culturable methods, <1% indoor microorganisms (Toivola et al. , 2002) • DNA-based approaches disclose a greater fraction of indoor communities (Kelley et al. , 2004; Pakarinen et al. , 2008, Rintala et al. , 2008) • Importance of microbial classification due to species association/origin 5

Overall Objectives Can we use filters as samplers for indoor contaminants? Explore the use of filters as samplers • Specific contaminants – Microorganisms (concentrations & communities) – Metals • Understand the strengths and weaknesses – Conditions in which could be used effectively – Important parameters 6

Study Phases Phase 1 Biological & heavy metal in residences Phase 2 Microbial composition in residences & full-scale test house Objectives 1. Compare concentrations in different sampling locations 2. Compare HVAC communities to other sampling locations 3. Influence of occupants vs outdoor contribution 4. Importance removal mechanisms Phase 3 Fate analysis of indoor particles 5. Role of critical parameters 7

Phase 1 Methodology • Collected – HVAC filter, floor & high surface dust from 8 residences – Bioaerosols from 5 of the above sites • Analyzed – Culturable microbial concentrations: bacteria & fungi (total & spores) – Heavy metals using atomic absorption spectroscopy: Pb, As, Cd • Compared – Sample locations – Filter efficiencies 8

Culturable concentration (CFU/g or m 3) Phase 1 Microbial Distributions Bacteria total Bacterial spores Fungi total Fungal spores 1 E+08 1 E+06 1 E+04 1 E+02 Filter High Surface Floor Air Sampling Location • Comparable trends across locations – Bacteria > Fungi – Spores ~100× < total 9

Phase 1 Metal Distributions Pb Cd As Concentration (μg/g) 300 100 30 10 3 1 0. 3 0. 1 Filter High Surface Floor Sampling Location • High surface statistically (p<0. 05) greater than filter for all metals • Floor statistically (p<0. 05) greater than filter, except for Cd 10 • Large particles may have greater metal concentrations (Al-Rajhi et al. , 1996)

Phase 1 Filter Efficiency Influence on Microbial Concentrations Filter MERV Bacteria total Low (n=7) 6 106 ± 6 105 Mid (n=7) High (n=4) Bacterial spores Fungi total Fungal spores 5 104 ± 9 103 4 105 ± 7 104 1 103 ± 1 103 9 105 ± 2 105 7 104 ± 7 103 6 105 ± 1 105 8 102 ± 1 103 3 105 ± 6 104 7 104 ± 3 103 1 105 ± 9 104 6 102 ± 7 102 CFU/g • Similar microbial concentrations across filters with different MERV ratings • Concentrations within range of literature values for indoor settled dust (Bouillard et al. , 2002; Ren et al. , 1999) • HVAC filters are hospitable environment for microorganisms 11

Phase 1 Filter Efficiency Influence on Metals Filter MERV Pb Low (n=7) Mid (n=7) High (n=4) 18. 5 ± 0. 79 12. 9 ± 0. 54 7. 49 ± 0. 44 Cd μg/g 2. 00 ± 0. 062 1. 64 ± 0. 035 1. 54 ± 0. 027 As 4. 61 ± 0. 24 1. 89 ± 0. 32 0. 912 ± 0. 22 • Cd concentrations similar across filter efficiencies • Low MERV filters statistically greater Pb & As concentrations than high MERV filters – Low MERV filters collect a greater fraction of large particles – Large particles may be associated with greater metal concentrations (Al. Rajhi et al. , 1996) – High MERV filters collect more mass of metals • Possible site specific indoor sources – lead-based paint (Kim et al. , 2002; Tong et al. , 1998) 12

Phase 1 Summary • HVAC filter dust colonized by microorganisms • Similar microbial concentrations in different sampling locations • Greater metal concentrations on high surfaces than filters • Greater metal concentrations on low MERV filters • Several phenomena need further investigation Phase 2 – Microbial communities could differ – Influence of occupants Phase 2 – Filter efficiency, HVAC cycling – Influence of particle size Phase 3 13

Study Phases Phase 1 Biological & heavy metal in residences Phase 2 Microbial composition in residences & full-scale test house Objectives 1. Compare concentrations in different sampling locations 2. Compare HVAC communities to other sampling locations 3. Influence of occupants vs outdoor contribution 4. Importance removal mechanisms Phase 3 Fate analysis of indoor particles 5. Role of critical parameters 14

Phase 2 Objectives • Develop culture-independent technique to assess microbial communities • Investigate 4 residences & test house • Characterize & compare microbial communities – Different sampling locations – Occupied & unoccupied house • Effects of humans • Outdoor contribution 15

Phase 2 UTest House • Controlled unoccupied environment (1, 300 ft 2) • Fan operated continuously • Detailed measurements: daily indoor & outdoor bioaerosols Composited into month-long samples 16

Phase 2 Sample Collection • High surface dust with vacuum mechanism (two-month study) – Avoiding track-in dust – Previously cleaned surfaces • High MERV filter dust – 9 pieces 1 x 1 inch filter • Daily indoor & outdoor bioaerosols (test house) Bioaerosols HVAC filter dust High surface dust 17

Phase 2 Molecular Tools • Extract DNA • Amplify (PCR)! • Clone • Sequence 18

$Phase 2 Phylogenetic Information Unifrac Significance • Developed by Pace group at CU (Lozupone$

Phase 2 Phylogenetic Information Unifrac Significance • Developed by Pace group at CU (Lozupone and Knight, 2005) • Applied to several settings, including indoor environment studies (Lauber et al. , 2008; Fierer et al. , 2008; Rintala et al. , 2008) Completely different 0 Identical 1 P-value Different 0. 1 Similar

Phase 2 Bacterial classification • Common phyla: Proteobacteria, Actinobacteria & Firmicutes • Firmicutes & Actinobacteria > in residences than Test House occupants • Proteobacteria > in Test House than residences outdoor origin

Phase 2 Fungal classification • Common classes: Dothideomycetes, Sordariomycetes & Agaricomycetes • Scarcity of Penicillium & Aspergillus spp. - bias of culturable methods (Pitkäranta et al. , 2008) • Sordariomycetes > in Test House than residences 21

Phase 2 Potential Pathogens Several potential opportunistic pathogens (Taylor et al. , 2001): Bacteria: • Pantoea agglomerans • Ralstonia pickettii • Enterobacter hormaechei • Staphylococcus aureus and epidermidis • Bacillus cereus, pumilus and subtilus • … Fungi: • Alternaria alternata and tenuissima • Fusarium proliferatum and oxysporum • Nigrospora sphaerica • Cladosporium cladosporioides • …

Phase 2 Residences Filter vs. High surface (P-value) Site Number 2 5 6 7 Bacteria 0. 41 0. 35 0. 48 0. 10 Fungi 0. 30 0. 65 0. 55 0. 16 • Microbial communities on filters & high surfaces are similar • The information provided by filter not statistically different than high surface dust 23

$Phase 2 Test House Unifrac Significance (P-value) Bacteria Filter-indoor air 0. 14 Fungi •$

Phase 2 Test House Unifrac Significance (P-value) Bacteria Filter-indoor air 0. 14 Fungi • • Filter-high surface 0. 13 0. 01 0. 24 High surface- Indoor-outdoor indoor air 0. 01 0. 12 0. 03 0. 19 Occupants impact microbial communities Filters can be used as long-term air samplers (Tringe et al. , 2008) High surface communities differ from air (larger particles? ) Indoor and outdoor air are similar in an unoccupied building 24

Phase 2 Summary • Similar filter & high surface microbial communities in residences • Filters can be used as long-term air samplers • Actinobacteria and Firmicutes associated with occupants • Proteobacteria dominate air and may be of outdoor air origin • Presence of opportunistic pathogens on filter dust • Occupants influence communities – Carry microorganisms – Generate particles 25

Study Phases Phase 1 Biological & heavy metal in residences Phase 2 Microbial composition in residences & full-scale test house Objectives 1. Compare concentrations in different sampling locations 2. Compare HVAC communities to other sampling locations 3. Influence of occupants vs outdoor contribution 4. Importance removal mechanisms Phase 3 Fate analysis of indoor particles 5. Role of critical parameters 26

Phase 3 Objectives • Scaling analysis to evaluate key removal mechanisms – 0. 001 -100 μm indoor airborne particles • Investigate filters as samplers – Likelihood of filters to capture a particle • Identify the most critical parameters involved 27

Phase 3 Mechanisms • Typical residence (V= 392 m 3) • Mechanisms: – Deposition β (Riley et al. , 2002 , ES&T) – HVAC Filtration λr η (Waring & Siegel, 2008, Indoor Air) – Exfiltration λ (Murray & Burmaster, 1995, Risk Analysis) • Model parameters: – Filtration efficiency (MERV): <5, 6, 11 – ASHRAE St. 52. 2 – Air recirculation rate (λr): 1. 1 & 5. 2 h-1 – Air exchange rate (λ): 0. 2, 0. 5 & 1. 3 h-1 BASELINE 28

Phase 3 Baseline Scenario Settled dust samples Air samples Settled dust samples

Phase 3 Filter Efficiency Scenarios

Phase 3 Air Recirculation Rate Scenarios λr=5. 2 h-1 MERV 11

Phase 3 Air Exchange Rate Scenarios

Phase 3 Summary • Deposition dominant mechanism for small (< 0. 02 μm) & large particles (> 2 μm) • Exfiltration important for mid-size particles (0. 02 - 2 μm) • High efficiency filters can develop into effective overall samplers – Especially if HVAC system runs continuously – In particular for 0. 3 – 3 μm particles – Building tightness has minimal importance 33

Conclusions • Filter viable option to assess culturable concentrations indoors • Some evidence of greater metal concentrations in larger particles • Similar filter & high surface microbial communities in residences • Filters could be used to detect pathogens • Filter efficiency & air recirculation rate important • Filters alternative to periodic air measurements • Filters can be used as integrated overall samplers 34

Limitations • Geographical, seasonal & socioeconomic • HVAC system operation (Stephens et al. , 2010; Thornburg et al. , 2004) • Return duct leakage • Occupants have significant control on parameters – System operation – System maintenance – Filter efficiency 35

Future Research • Sources & sinks of contaminants are not evenly distributed in building • Characterization of microbial growth conditions on filters • Other particle-bound contaminants (PBDEs, Phthalates) • Link to occupant exposure – Occupants personal monitors & filter analysis – Can filter concentrations predict exposure? 36

Acknowledgments Financial support • ASHRAE Graduate Student Grant-in-aid • NIOSH Pilot Project Research Training Program • NSF IGERT: Indoor Environmental Science and Engineering Technical and moral support • Jim Rosenthal • Dr. Kinney and Dr. Siegel and their research groups

Thank you! Questions? 38

Backup Slides 39

Phase 1 Microbial Distribution by Site Fungi total Fungal spores 1 E+08 1 E+06 1 E+04 Filter HS Floor 3 Filter HS Floor 2 Filter HS Floor 1 E+02 Filter HS Floor Mean Concentration (CFU/g) Bacteria total Bacterial spores 4 5 6 7 8 Site • Comparable concentrations across locations • Often same location has highest concentrations of contaminants 40

Phase 1 Metal Distribution by Site As Cd 213. 3 100 80 60 40 Filter HS Floor Filter HS Floor 1 Filter HS Floor 20 Filter HS Floor Mean Concentration (µg/g) Pb 2 3 4 5 6 7 8 Site • Less similarity (sources, particle size? ) • In 6 out of 8 sites, HVAC has the lowest Pb concentration 41

Standard 52. 2 • Measure upstream & downstream concentrations in 12 bins six times (clean filter to filter loaded with standard synthetic dust). • Challenge with poly disperse solid-phase KCl particle aerosol. For each bin, consider the lowest efficiency (of 6 tests), average into 3 groups (E 1, E 2, E 3) • Classify into MERV using E 1, E 2 and E 3 42

Rarefaction Curves Bacteria Fungi

Air Exchange Rates for High Efficiency Filters