Outdoor air and (non-combustion) appliances as sources of indoor particulate matter (PM)

Transcription

1 Outdoor air and (non-combustion) appliances as sources of indoor particulate matter (PM) Brent Stephens, PhD Assistant Professor Civil, Architectural and Environmental Engineering web

2 Indoor exposures to outdoor PM WHAT WE KNOW 2

3 Outdoor PM and adverse health effects Documented health effects include: Stroke Heart disease Lung cancer Chronic and acute respiratory diseases (including asthma) Lung function Mortality Measures of PM (some causal, some suggestive): PM 10 PM 2.5 Ultrafine particles (UFPs, less than 100 nm) Various chemical components of PM US EPA 2009 Integrated Science Assessment for Particulate Matter and many references therein 3

4 Outdoor PM 2.5 and mortality An estimated 130,000 deaths in 2005 in the US were due to elevated outdoor PM 2.5 Fann et al., 2012 Risk Analysis 4

5 Indoor exposures to outdoor PM We spend most of our time indoors Nearly 90% of the time, on average (~70% at home) Klepeis et al., 2001 J Exp Anal Environ Epidem Outdoor PM enters into buildings with varying efficiencies Outdoor PM becomes indoor PM Human exposure to outdoor PM often occurs indoors And often at home Meng et al., 2005 J Expo Anal Environ Epidem; Kearney et al., 2010 Atmos Environ; Wallace and Ott 2011 J Expo Sci Environ Epidem: MacNeill et al Atmos Environ; MacNeill et al Indoor Air 5

6 Indoor sources of outdoor PM and key definitions I/O ratio: Infiltration factor: Chen and Zhao, 2011 Atmos Environ 6

7 I/O PM ratios: Indoor + outdoor sources Means from 77 studies and over 4000 homes; includes indoor and outdoor sources Median I/O ratio for PM 2.5 : ~1 Median I/O ratio for PM 10 : ~0.8 Median I/O ratio for UFPs: ~0.8 Chen and Zhao, 2011 Atmos Environ [~UFPs] [0.5+ µm] 7

8 Infiltration factors: Outdoor PM sources only Means from 21 samples of over 20 homes (includes only outdoor PM infiltration) Total # of homes: ~1000 in the U.S. & ~150 in Europe Median F inf for PM 2.5 : ~0.55 Median F inf for PM 10 : ~0.3 Median F inf for UFPs: ~0.3 Kearney et al., 2010 Atmos Environ; Kearney et al., 2014 Atmos Environ; Stephens, 2015 Sci Technol Built Environ Chen and Zhao, 2011 Atmos Environ 8

9 Variability in infiltration factors F inf for UFPs in Windsor, ON F inf for PM 2.5 in Edmonton, AB Kearney et al., 2010 Atmos Environ Kearney et al., 2014 Atmos Environ F inf for PM 2.5 in 7 U.S. cities Betweenhome variations: 0.1 < F inf < 1 Allen et al., 2012 Environ Health Persp 9

10 Key drivers of variability in infiltration factors Source of ventilation air Infiltration (leaks) Mechanical ventilation Natural ventilation Human behaviors (e.g., window opening frequencies) Magnitude of the air exchange rate (AER) Meteorological conditions Sizes/classes/components of PM Building characteristics (e.g., airtightness) HVAC system design and operation Williams et al., 2003 Atmos Environ; Allen et al., 2012 Environ Health Persp; MacNeill et al, 2012 Atmos Environ; MacNeill et al., 2014 Indoor Air; El Orch et al., 2014 Build Environ; Chen et al., 2012 Epidemiology 10

11 Indoor exposures to outdoor PM WHAT WE DO NOT KNOW Or what do we know less about? 11

12 How does variability in F inf contribute to effect estimates? Accounting for variations in AERs and window opening PM 10 mortality (U.S.) PM 2.5 and acute myocardial infarction (NJ) Chen et al., 2012 Epidemiology Hodas et al., 2013 J Exp Sci Environ Epidem PM 2.5 and ER visits (ATL) Sarnat et al., 2013 J Exp Sci Environ Epidem 12

13 Window opening frequencies and impact on AER Johnson and Long, 2005 J Expo Anal Environ Epidem Price and Sherman 2006 LBNL Report El Orch et al., 2014 Build Environ Limited data on air exchange rate multipliers with open windows: Typically ~2-4 times higher, depending on area of openings, I/O ΔT Marr et al., 2012 HVAC&R Research; Wallace et al., 2002 J Expo Anal Environ Epidem; Johnson et al., 2004 J Expo Anal Environ Epidem; Chen et al., 2012 Epidemiology 13

14 Underlying mechanisms that govern F inf Outdoor particles P AER Loss Loss AER Penetration from outdoors Penetration Factor If P = 1: The envelope offers no protection If P = 0: The envelope offers complete protection Removal by air exchange Removal by deposition to surfaces, phase changes, or control by filters or air cleaners 14

15 Envelope penetration factors Chen and Zhao, 2011 Atmos Environ Liu and Nazaroff 2001 Atmos Environ 19 homes in TX Rim et al., 2010 Environ Sci Technol P for UFPs # of homes with penetration factors measured: Size-resolved PM: < 20 homes UFPs: ~30-50 homes PM 2.5 : estimated in 100s of homes But seldom (never?) measured Stephens and Siegel, 2012 Indoor Air 15

16 Other unknowns (or less knowns) Associations between F inf (or P) and building characteristics Some evidence of associations w/ AC usage, year of construction, and envelope airtightness How do they change after building retrofits? Allen et al., 2012 EHP; MacNeill et al., 2012 Atmos Environ; Stephens and Siegel 2012 Indoor Air Chemical transformations e.g. evaporative losses Hodas et al., 2014 Aerosol Sci Technol High spatial- and temporal-resolution data for: Outdoor particle size distributions Outdoor size-resolved aerosol composition 16

17 Summary of outdoor indoor PM transport research needs We need more integration between epidemiologists and exposure scientists Address exposure misclassification Improve health effect estimates We need more data on window opening frequencies and their impact on air exchange rates We need more field measurements of size-resolved, UFP, and PM 2.5 penetration factors And explorations of associations with building characteristics 17

18 INDOOR PM SOURCES Specifically: Non-combustion appliances 18

19 Non-combustion appliances as indoor PM sources Several (non-combustion) appliances emit PM indoors Mostly in the UFP size range Laser printers Desktop 3D printers He et al., 2010 J Aerosol Sci Afshari et al., 2005 Indoor Air 19 Stephens et al., 2013 Atmos Environ

20 Non-combustion appliances as indoor PM sources More desktop 3D printers UFP emission rate (#/min) Heated surfaces Wallace 2015 ISES Conference Azimi et al., 2016 Environ Sci Technol Procedure: Wash, expose to indoor air, burn until UFP reaches zero Proposed mechanisms: 1. Deposition of material on surface 2. Desorption of SVOCs on the heated surface 3. Followed by nucleation and particle growth in cooler air Wallace et al., 2015 Indoor Air 20

21 Typical indoor UFP emission rates UFP emitting device Size range Emission rate (#/min) Reference Flat iron with steam nm Afshari et al. (2005) Electric frying pan nm Buonnano et al. (2009) 3D printer w/ PLA nm ~ Stephens et al. (2013) Vacuum cleaner nm Afshari et al. (2005) Scented candles nm Afshari et al. (2005) Gas stove nm Afshari et al. (2005) 3D printer w/ ABS nm ~ Stephens et al. (2013) Cigarette nm Afshari et al. (2005) Electric stove nm Afshari et al. (2005) Frying meat nm Afshari et al. (2005) Radiator nm Afshari et al. (2005) Desktop 3D printers nm ~10 8 to ~10 12 Azimi et al. (2016) Laser printers nm to He et al. (2010) Cooking on a gas stove nm Buonnano et al. (2009) Items in red are non-combustion sources Items in black are combustion-related 21

22 Summary of (non-combustion) appliance emissions We continue to find new indoor sources of PM Mostly UFPs We need to continue to gather emission rate data for these and other sources Including size-resolved emission rate data We need to continue to explore source control and exposure mitigation strategies 22

23 23

24 Bonus slides 24

25 outline Importance of outdoor PM Outdoor PM comes in Define Finf, then show data Define underlying mechanisms for Finf, show data Then get to what we don t know We don t have great exposure assessment for health data (misclassification charlie s and sarnat s work w AER) Size-resolved P knowledge weak UFP and PM2.5 P knowledge weak Window opening behavior/frequency and impact on AER and P Associations w/ building characteristics / predictive ability / how P changes in future Cautionary tale w/ Austin mech vent data Then transition to indoor sources 25

26 Cleaning activities Ironing Vacuuming cleaning Air freshener sprays Printers Laser 3d Cooking Electric stove SVOC desorption (hot surfaces) Others Humidifiers 26

27 RQs Envelope penetration factors Size resolved PM2.5 Chemical constituents Changes after retrofits (relationships w/ air tightness) Window opening Behaviors/frequencies Impacts on AER Non-combustion appliances Source control Be ever vigilant (e.g., 3D printers) 27

28 Don t forget cautionary tale of mechanical ventilation in Austin (worst and best protector) Azimi et al size distributions (comp to Jaenicke) Our new P and K work 28

29 Several mechanisms govern indoor pollutant concentrations Outdoor PM Ventilation/ Air Exchange Intentional: Mechanical Natural (windows) Unintentional: Air infiltration (leakage) Homogeneous Chemistry Adsorption/ Desorption Phase change, partitioning & byproduct formation Indoor Emissions Filtration Deposition/S urface Reactions Ventilation/ Air Exchange Temp RH Primary pollutant classes: (1) Particulate matter (both biological and non-biological origin) (2) Non-reactive gases (e.g., CO, CO 2, VOCs, formaldehyde) (3) Reactive gases (e.g., O 3, NO 2 ) 29

30 Indoor proportion of outdoor particles Chen and Zhao, 2011 Atmos Environ Kearney et al., 2010 Atmos Environ Outdoor particles infiltrate into and persist within buildings with varying efficiencies Williams et al., 2003 Atmos Environ Exposure to outdoor PM often occurs indoors Often at home Meng et al., 2005 J Expo Anal Environ Epidem Kearney et al., 2010 Atmos Environ Wallace and Ott 2011 J Expo Sci Environ Epidem MacNeill et al Atmos Environ 30

31 Outdoor UFP source terms and airtightness (blower door) C C in out = P AER AER + Loss Leakier homes had much higher outdoor particle source rates Leaky homes are also older predictive ability? Potential socioeconomic implications: low-income homes are also leakier Chan et al., 2005 Atmos Environ Stephens and Siegel, Indoor Air (6):

32 PM infiltration and age of homes C C in out = P AER AER + Loss C C in out = P AER AER + Loss Penetration Factor, P P = [m] yearbuilt + [b] [m] = ± [b] = ± R 2 = Year Built Outdoor Source Term, P AER ( (hr ) -1 ) Source = [m] yearbuilt + [b] [m] = ± [b] = ± R 2 = Year Built Older homes also had much higher outdoor particle source rates Stephens and Siegel 2012 Indoor Air 32

33 Implications for submicron PM exposure: 19 homes Combined effects: Lower bound Upper bound Penetration factor, P Air exchange rate, AER (1/hr) Outdoor source term, P AER (1/hr) Indoor loss rate, β + ηq/v (1/hr) Fractional HAC operation, f 55.3% 10.7% I/O submicron ratio (F inf ) Factor of ~60 to ~70 difference in indoor proportion of outdoor particles between: A new airtight home with a very good filter and high HAC operation, and A leaky old home with a poor filter and low HAC operation Some potential for predictive ability using: Age of home Knowledge of HAC filter type Building airtightness test results I/O climate conditions Stephens 2015 Sci Technol Built Environ 33

34 Envelope penetration factors 34

35 Envelope penetration factors 35

36 We gathered 194 long-term average (1-year or more) outdoor particle size distributions from the literature from all over the world Azimi et al., 2014 Atmos Environ 36

37 We gathered 194 long-term average (1-year or more) outdoor particle size distributions from the literature from all over the world Azimi et al., 2014 Atmos Environ 37

38 Penetration results from Thatcher et al. (2003) Richmond: Leakier house ELA = 148 cm 2 Clovis: Tighter house ELA = 87 cm 2 * Estimated Leakage Area (ELA) = f (blower door air leakage coefficients & ΔP) 38