Source Attribution: Understanding a Complex Source

Transcription

1 Source Attribution: Understanding a Complex Source Presentation at Institute of Medicine s workshop: From Exposure to Human Disease: Research Strategies to Address Current Challenges September 14-15, 2006 Lynn M. Hildemann Civil & Environmental Engineering Dept. Stanford University hildemann@stanford.edu

2 Goal of Source Attribution n To quantify contributions of various sources of toxic chemicals to a site of interest (in this case, a person s bodily exposure) This talk will focus on identifying and quantifying the sources of inhaled airborne particulate matter (PM)

3 Current State of the Science (1) Chemical Mass Balance Model Known Source #1 Known Source #2 Known Source #3??? Sample from Site

4 Information Needed for Chemical Mass Balance Model n Chemical composition of PM emissions from each source n Chemical composition of PM for site/exposure of interest Important Assumption: Tracer species must be nonreactive.

5 Example: Major Sources of Ambient Fine PM (PM2) Sources of Ambient PM2, Pasadena (from Schauer et al., Atm.Env.30:3837, 1996) Paved road dust 12% Veget. Debris 1% Wood smoke 10% Cigarettes 1% Meat cooking 9% Diesel vehicle exhaust 18% Gas. vehicle exhaust 6% Tire wear 1% Background+ 2ndary 42%

6 Example: Chemical Composition of Diesel Exhaust PM2 (% by mass) Unresolved/ UnIDed Organics 38% n-alkanoic Acids 20% Benzoic Acids 3% IDed Organics 2% Trace metals, ions 2% PAHs 3% oxy-pahs 3% Steranes 3% Hopanes 4% Elem. Carbon 40% Other (not identified) 18% n-alkanes 62% Other IDed 2% [Data from Hildemann et al., ES&T 25:744, 1991; Rogge et al., ES&T 27:636, 1993]

7 Outcome: Can ID Source(s) of (Stable) Toxic Chemicals at an Ambient Site n This is possible because we know: (1) The concentration of PM from each source present at the ambient site (in µg PM/m 3 ) (2) The mass fraction of toxic chemical emitted from each source (in ng toxin/µg PM) Why can t we use this to assess exposure?

8 Personal Exposure is Due to Multiple Sources and Locations Sources are indoors as well as outdoors Exposure varies with proximity to source SOURCE Outdoor RECEPTOR Indoor (Passive Exposure) SOURCE Personal Activities

9 The average U.S. resident spends ~90% of the time indoors Personal Indoor Outdoor Activity Experimental set-up 2 Persons walk Dry dust Vacuum first floor Vacuum bedrooms Collocation period :00 13:00 14:00 15:00 16:00 17:00 18:00 Time [from Ferro et al., J. Expos. Anal. Env. Epid. 14:S34, 2004] PM5 Concentration (ug/m^3)

10 Should Use Personal Monitoring Instead of Ambient Measurements: n Will capture proximity effect and incomplete mixing effects, but n Is more difficult to ID all the potentially major sources

11 Current State of the Science (2) Positive Matrix Factorization To Be IDed Source #1 To Be IDed Source #2 To Be IDed Source #3 X 100 or more Large Pool of Samples from Site

12 Information Needed for Positive Matrix Factorization n Chemical composition of PM for >100 similar sites/exposures n Ability to identify sources based on deduced chemical composition Important Assumption: Tracer species must be nonreactive.

13 Example: Source Attribution With Exposure Data [Yakovleva et al., ES&T 33:3645 (1999)] Samples: 178 personal PM10 exposure measurements Tracers: 18 trace elements No. of sources retained: 5 Source assignments: 1 = soil 2 = secondary sulfate 3 = sea salt 4 = smelters and motor vehicles 5 = personal activities 1: soil 2: secondary sulfate 3: sea salt 4: smelters, vehicles 5: personal activities

14 Example (cont.): Source Attribution With Exposure Data [Yakovleva et al., ES&T 33:3645 (1999)] n Source contributions can also be found: ~60% of total exposure due to indoor sources

15 More Research is Needed to Better Understand: n Nature/Sources of personal activities n Emission rates from specific activities (=sources) n Effects of proximity to source on exposure

16 Example: Emission Rates for a Few Personal Activities (mg PM/min) Persons walk (FF) Dance on rug (BM) PM-2.5 PM Person walk (FF) Vacuum (BR) Dance on wood (BM) 0 Smoke cigarette Dry dust (FF) Vacuum (FF) Make bed/fold clothes (BR) 1 Person walk (BM) Fold blankets (BR) [from Ferro, Kopperud and Hildemann, 2004, Envir. Sci. & Technol. 38: ]

17 Barriers Remaining 1) Particle sizes change with time n ETS particles increase in size by 20-50% within mins after release inside typical homes [Morawska et al., Sci. Total Env. 196:43, 1997] This will affect the deposition efficiency in the respiratory tract -- expect fresh (small) PM to deposit most efficiently

18 Barriers Remaining 2) Chemical composition changes with time n Some compounds readily degrade within a timescale of minutes to hours n Reactions can form new (and sometimes more toxic?) chemicals Difficult to predict levels of reactive chemicals in aged source emissions

19 Barriers Remaining 3) Particle composition varies with size 0.4 Cigarette has a peak in the mass distribution at 0.4 µm µm PM Mass But V peaks at 0.6 µm, while Sb peaks at 0.06 µm 0.6 µm 0.06 µm [modified from Kleeman et al., ES&T 33:3516, 1999]

20 Summary: To ID Sources of Exposure to Toxins, We Need to Better Understand: n Pollutant exposures from specific human activities n Effects of proximity to sources n Effects of emissions aging on toxin levels, respiratory uptake Thank you

21

22 Importance of Stable Tracers [from JJ Schauer et al., Atm.Env. 30:3837 (1996)] n Some organic compounds degrade significantly, while others have significant secondary formation pathways Significant secondary formation in the atmosphere Significant chemical degradation in the atmosphere