Source Apportionment of PM 2.5 at Three Urban Sites Along Utah s Wasatch Front

Source Apportionment of PM 2.5 at Three Urban Sites Along Utah s Wasatch Front Robert Kotchenruther, Ph.D. EPA Region 10 October 6, 2011

Motivation for this analysis: Utah and Idaho share a PM2.5 nonattainment area in the Cache Valley, north of. There are very few speciated PM2.5 measurements in the Cache Valley, so Utah/Idaho will be using speciation data from sites along the Wasatch Front (closer to Salt lake City) as surrogates for PM2.5 speciation in the Cache Valley. Limited comparisons of speciation data between the Cache Valley and Wasatch Front sites indicate that chemical speciation is similar on wintertime high PM2.5 days (>50% ammonium nitrate). It is hoped that a better understanding of sources impacting sites along the Wasatch Front will inform the possible range of sources impacting PM2.5 in the Cache Valley To date no one has conducted a recent source apportionment analysis of PM2.5 for sites along the Wasatch Front.

Map and location of PM2.5 Speciation monitors Cache Valley NAA Idaho Utah PM2.5 Speciation Monitors used in source apportionment analysis:, Utah AQS site identifier 490110004 lat & lon: 40.902967, -111.884467, Utah AQS site identifier 490353006 lat & lon: 40.736389, -111.872222, Utah AQS site identifier 490494001 lat & lon: 40.341389, -111.713611

Comparison of Cache Valley PM2.5 Speciation and that in : A speciation monitor was recently installed in the Cache Valley (Logan) starting in the 2010/2011 winter. 5 - Day Average PM2.5 > 25 ug/m3 % NH4 %NO3 %SO4 %EC %OC Logan (Cache Valley) 22.0% 46.8% 3.3% 4.4% 17.8% Salt Lake City 16.8% 41.1% 4.8% 3.6% 21.9% (Source: Utah DEQ)

Details of this source apportionment work: Source apportionment model: Positive Matrix Factorization (PMF, EPA v4.0beta) Data: chemically speciated PM2.5 data, 24-HR average ~ 4 years 5/2007 5/2011, starting after switch in carbon sampling method (switched to URG3000N) Sites: (227 samples), (429 samples), and (228 samples) 28 chemical species: OC & EC thermal fractions, NO3, SO4, NH4, trace metals and soil elements. Method: Each site modeled separately and results compared

What is Positive Matrix Factorization (PMF)?? A form of principal component analysis (factor analysis). Model input -> Aerosol total mass, chemical composition, and measurement uncertainties Model process -> Looks for systematic patterns in the day-to-day chemical variations so that the data variability can be explained by a smaller subset of factors. These factors are often interpreted as sources of aerosol. Model output -> Each factor s chemical fingerprint Factor mass contributions to each sample

What is Positive Matrix Factorization (PMF)?? (continued) Model interpretation -> What is at PMF model Factor? Possibilities are: A single source (e.g. industrial facility) A source category (e.g. a bunch of sources that have very similar chemical fingerprints and emissions patterns e.g., cars, trucks, wood burning sources) Multiple sources or source categories grouped together (PMF may group together multiple sources because of insufficient variability in emissions, chemical fingerprint, and/or temporal/spatial resolution)

PMF Results for the Wasatch Front

10 12 Factors Found!?! An ammonium chloride aerosol factor is a surprising result, not found at other locations in the U.S., but found at all 3 sites along the Wasatch front. 10 factors 11 factors 12 factors Factors for this study Factors Factors Factors Ammonium Nitrate Ammonium Chloride Wood Smoke OP rich Gasoline Vehicles Not found Ammonium Sulfate Diesel I Fugitive Dust I Fugitive Dust II Fireworks Industrial/Urban Diesel II Not found Not found Other factors typical of urban U.S. locations. Factors are very similar site to site, suggesting relative homogeneity of sources in airshed.

Results: Mass impacts during winter high PM2.5 days Similar mass impacts site to site. Ammonium chloride factor is significant, 8-15% under winter high PM2.5 conditions. Is ammonium chloride aerosol a reasonable result?

Ambient Ammonium Chloride Aerosol: Ambient ammonium chloride aerosol has been previously documented and studied in lab analyses. Sample of literature references where ambient ammonium chloride was observed: Yoshizumi and Okita (1983) found NH4Cl concentrations between 7.76-15.5 ug/m3 in Riverside, CA. Pio and Harrison (1987) found NH4Cl concentrations up to 10 ug/m3 in Northwest England. Possanzin et al. (1992) found NH4CI represents approximately one-fifth of NH4NO3 and one-tenth of the total ammonium species by mass in Rome, Italy. (Most references appear to be from the 1980 s and 1990 s).

Ammonium Chloride Aerosol Chemistry: Pio and Harrison (1987) determined that the thermodynamics of NH4Cl formation is very similar to that of NH4NO3. NH3(g) + HCl(g) NH4Cl(s) NH3(g) + HNO3(g) NH4NO3(s) Both have reversible reactions and similar dependency on temperature and RH. From Pio and Harrison (1987): NH4Cl(s) is slightly more volatile than NH4NO3(s) [for the same NH3(g) concentration. equilibrium concentrations of HCI(g) are 1.5-2 times higher than HNO3(g) levels] Taking typical atmospheric concentrations of gaseous NH3 and HCI, it is unlikely that NH4CI aerosol will exist at temperatures much above 10 deg C.

Ammonium Chloride Aerosol Chemistry: NH3(g) + HCl(g) NH4Cl(s) What might be sources of HCl(g) in the Wasatch Front? Chlorine replacement in reactions with NaCl aerosol releases HCl(g) H2SO4(g) + 2NaCl(s) => 2HCl(g) + Na2SO4(s) HNO3(g) + NaCl(s) => HCl(g) + NaNO3(s) The Salt Lake has high NaCl concentrations. Heavy road salting is done in the winter in the area. MagCorp (a magnesium chloride plant) is a known large point source of chlorine the Salt Lake airshed (~40 miles west of ) Poorly controlled coal combustion is a known historical source of HCl(g). Waste incineration can release HCl(g).

PMF factor Ammonium Chloride at each location Chemical fingerprint Br Cl NH 4

PMF factor Ammonium Chloride at each location 2.5 Mass impacts 2.0 Factor Monthly Average Mass Impacts (ug/m3) 16 14 12 10 8 6 4 2 0 Factor Mass Impacts (ug/m3) Factor time series 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 1.5 1.0 0.5 0.0 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011 Monthly mean conc. 1 2 3 4 5 6 7 8 9 10 11 12 Month Note: Mainly winter impacts agree with understood thermodynamics of secondary NH4Cl Concentrations in line with those observed elsewhere

Other analysis to test the PMF finding of NH 4 Cl aerosol: Further analysis of PMF results Do NH 4 Cl and NH 4 NO 3 PMF factors form on the same days? -> look at NH 4 Cl and NH 4 NO 3 factor mass scatter plot What is the temperature dependence of the NH 4 Cl factor? -> look at NH 4 Cl factor mass vs. max daily temperature Outside of the PMF analysis, other data analyses Are we missing any major ions that could account for the Cl? -> look at ion balance between available cations (+) and anions (-), is there a charge imbalance indicating possible missing ions? If we have ion balance, what cations are available in the sample that could be associated with Cl? -> look at samples with high Cl mass and see what cations could be associated with the Cl.

Are conditions conducive to a high NH 4 NO 3 factor mass also conducive to a high NH 4 Cl factor mass? Comparing PMF ammonium nitrate and ammonium chloride factors Ammonium Chloride Factor (ug/m 3 ) 20 18 16 14 12 10 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 Ammonium Nitrate Factor (ug/m 3 ) Conditions leading to significant NH 4 NO 3 are also more conducive to NH 4 Cl, consistent with similarity between NH 4 NO 3 and NH 4 Cl thermodynamics.

What temperatures are conducive to high NH 4 Cl factor mass? Max daily temperature and ammonium chloride factor Max Daily Temperature (Deg. C) 40 40 30 20 10 0 Significant NH 4 Cl factor mass occurs mostly when max daily temperatures are less than 10C, consistent with Pio and Harrison (1987) prediction about the temperature dependence of NH 4 Cl aerosol. -10 0 2 4 6 8 10 12 14 16 Ammonium Chloride Factor (ug/m 3 )

Are we missing any major ions that could be associated with Cl? To check we can see if there is charge balance between the expected major ions: Cl - NO 3 - SO 4 2- Na + K + NH 4 + Plotting anion charge equivalence vs. cation checks this. anion equivalence = [Cl - ]/35.453 + [NO 3- ]/62.005 + [SO 4 2- ]/48.03 cation equivalence = [Na + ]/23.0 + [K + ]/39.098 + [NH 4+ ]/18.04 Cation Equivalence 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 1:1 line There is very close charge balance between the expected anions and cations, so it does not appear we are missing any major species. 0.2 0.1 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Anion Equivalence

When we have high Cl equivalence, what are the available cations that could be associated? Highest Cl samples at, ranked by Cl charge equiv. [rescaled from right] Anion Equivalence Cation Equivalence 0.10 0.08 0.06 0.04 0.02 0.00 0.00 0.02 0.04 0.06 0.08 0.10 High Cl Sample Days Anion Equivalence Cation Equivalence 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 High Cl Sample Days Anion Equivalence NO3 SO4 Cl Cation Equivalence Na K NH4 This shows that the highest Cl concentration samples at are mostly associated with NH 4

When we have high Cl equivalence, what are the available cations that could be associated? Highest Cl samples at, ranked by Cl charge equiv. [rescaled from right] Anion Equivalence Cation Equivalence 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 High Cl Sample Days Cation Equivalence Anion Equivalence 1.0 0.8 0.6 0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0 High Cl Sample Days Anion Equivalence NO3 SO4 Cl Cation Equivalence Na K NH4 This shows that the highest Cl concentration samples at Salt Lake City are mostly associated with NH 4

When we have high Cl equivalence, what are the available cations that could be associated? Highest Cl samples at, ranked by Cl charge equiv. Anion Equivalence 0.8 0.6 0.4 0.2 0.0 0.0 High Cl Sample Days Anion Equivalence NO3 SO4 Cl [rescaled from right] Cation Equivalence 0.2 0.4 0.6 0.8 Cation Equivalence Na K NH4 High Cl Sample Days This shows that the 2 highest Cl concentration samples at are mostly associated with NH 4, the next two highest could be associated with any of the cations.

For the Cache Valley, is there significant NH 4 Cl? Trying to get speciated PM2.5 data from Cache Valley to see.

References: Kim, Eugene, Philip K. Hopke, and Eric S. Edgerton, Improving source identification of Atlanta aerosol using temperature resolved carbon fractions in positive matrix factorization, Atmospheric Environment 38, 3349 3362, 2004. Lewis, C.W., Norris, G.A., Henry, R.C., and Conner, T.L., Source Apportionment of Phoenix PM2.5 Aerosol with the Unmix Receptor Model. Journal of the Air & Waste Management Association 53, 325-338, 2003. Maykut, N., J. Lewtas, E. Kim, and T. Larson, Source Apportionment of PM2.5 at an Urban IMPROVE Site in Seattle, Washington, Environ. Sci. Technol., 37, 5135-5142, 2003. Ogulei, D., Sources of Fine Particles in the Wapato Hills-Puyallup River Valley PM2.5 Nonattainment Area, Washington State Department of Ecology, Publication No. 10-02-009, April 2010. Pio, C. and R. Harrison, The Equilibrium of Ammonium Chloride Aerosol with Gaseous Hydrochloric Acid and Ammonia under Tropospheric Conditions, Atmospheric Environment, 21, 1243 1246, 1987. Possanzini, M., P. Masia, and V. Di Palo, Speciation of Ammonium-Containing Species in Atmospheric Aerosols, Atmospheric Environment, 26A, 1995-2000, 1992. Ramadan, Z., Song, X.-H., and Hopke, P.K., Identification of Sources of Phoenix Aerosol By Positive Matrix Factorization. Journal of the Air & Waste Management Association 50, 1308-1320, 2000. Yoshizumi, K. and T. Okita, Quantitative Estimation of Sodium- and Ammonium-Nitrate, Ammonium Chloride, and Ammonium Sulfate in Ambient Particulate Matter, Journal of the Air Pollution Control Association, Volume 33, No. 3, 224-226, 1983. Zhao, W. and P. Hopke, Source apportionment for ambient particles in the San Gorgonio wilderness, Atmospheric Environment 38, 5901 5910, 2004.

Other PMF Factors

Ammonium Nitrate Factor NH 4 NO 3

Ammonium Nitrate Factor Factor time series 50 45 40 35 Factor Mass Impacts (ug/m3) 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 Month Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. 30 25 20 15 10 5 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Wood Smoke Factor ( factor omitted here) Lack of K unusual, but OC1 and EC1 similar to Tacoma wood smoke factor. K OC1 EC1 Tacoma WA Wood Smoke Factor

Wood Smoke Factor Factor time series 9 8 7 6 3.0 2.5 2.0 1.5 1.0 Factor Mass Impacts (ug/m3) 0.5 0.0 Monthly mean conc. 1 2 3 4 5 6 7 8 9 10 11 12 Month Factor Monthly Average Mass Impacts (ug/m3) 5 4 3 2 1 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011 Seasonal pattern of mass impacts matches that expected of wood smoke from home heating

OP Rich Factor ( factor omitted here) OP EC1 Tacoma WA OP Rich Factor

OP Rich Factor 2.5 2.0 Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. Factor time series 10 9 8 7 1.5 1.0 Factor Mass Impacts (ug/m3) 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 6 5 4 3 2 1 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011 The source(s) of this factor are uncertain.? Aged POA/SOA More heavily oxidized POA/SOA would be less volatile and perhaps more likely to char rather than volatilize during OC thermal analysis

Gasoline Vehicles Factor (not found at ) Similar to factors found in Zhao and Hopke, 2004 Kim et al. 2004 Maykut et al. 2003 OC2 OC3 OC4 EC1 Tacoma WA Gasoline Vehicles Factor

Gasoline Vehicles Factor 2.5 2.0 Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. Factor time series 8 7 1.5 1.0 Factor Mass Impacts (ug/m3) Salt Lake 0.5City 6 5 4 3 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 2 1 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Ammonium Sulfate Factor NH 4 SO 4

Ammonium Sulfate Factor 2.0 1.5 Monthly mean conc. Factor Monthly Average Mass Impacts (ug/m3) Factor time series 10 9 8 7 Factor Mass Impacts (ug/m3) 1.0 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 6 5 4 3 2 1 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Diesel I Factor Cu Fe Mn Zn EC1

Diesel I Factor 1.0 0.8 Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. Factor time series 4 3 Factor Mass Impacts (ug/m3) 0.6 0.4 0.2 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 2 1 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Fugitive Dust I Factor Al Ca Fe Ti Si

Fugitive Dust I Factor 1.5 Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. 1.0 Factor time series 6 5 4 Factor Mass 0.5Impacts (ug/m3) 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 3 2 1 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Fugitive Dust II Factor Al Ca Fe Ti Si

Fugitive Dust II Factor 3.0 2.5 Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. 2.0 Factor time series 50 45 40 35 Factor Mass Impacts (ug/m3) 1.5 1.0 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 30 25 20 15 10 5 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Fireworks Factor Cu Mg K

Fireworks Factor 2.5 2.0 Monthly mean conc. Factor Monthly Average Mass Impacts (ug/m3) Factor time series 16 14 12 10 8 6 1.5 1.0 Factor Mass Impacts (ug/m3) 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 4 2 0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Industrial/Urban Factor Cr Pb Ni NO3 EC2 EC3 SO4

Industrial/Urban Factor 1.0 0.8 Factor Monthly Average Mass Impacts (ug/m3) Monthly mean conc. Factor time series 1.5 1.0 Factor Mass Impacts (ug/m3) 0.6 0.4 0.2 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 0.5 0.0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011

Diesel II Factor Cu Mn Zn NO3 EC1

Diesel II Factor 0.2 Monthly mean conc. Factor Monthly Average Mass Impacts (ug/m3) Factor time series 1.5 Factor Mass Impacts (ug/m3) 0.1 1.0 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month 0.5 0.0 4/1/2007 7/1/2007 10/1/2007 1/1/2008 4/1/2008 7/1/2008 10/1/2008 1/1/2009 4/1/2009 7/1/2009 Date 10/1/2009 1/1/2010 4/1/2010 7/1/2010 10/1/2010 1/1/2011 4/1/2011