[3] Department of Environmental Science and Engineering, Fudan University, Shanghai, China.

Transcription

1 upplement for manuscript Influence of Intense secondary aerosol formation and long range transport on aerosol chemistry and properties in the eoul Metropolitan Area during spring time: Results from KORU-AQ Hwajin Kim 1,,Qi Zhang 3,* and Jongbae Heo [1] Center for nvironment, Health and elfare Research, Korea Institute of cience and Technology, eoul, Korea [] Department of nergy and nvironmental ngineering, University of cience and Technology, Daejeon, Korea [3] Department of nvironmental cience and ngineering, Fudan University, hanghai, China. [] Department of nvironmental Toxicology, University of California, Davis, CA 91, UA [] Center for Healthy nvironment ducation & Research, Graduate chool of Public Health, eoul ational University, eoul, Korea *Corresponding author: Qi Zhang Department of nvironmental Toxicology, University of California 1 hields Avenue, Davis, California 91 Phone: (3) mail: dkwzhang@ucdavis.edu 1 3 1

2 Table 1. Comparison of the average O/C, H/C, and OM/OC ratios of total OA and the four OA factors identified from PMF analysis calculated using the Aiken-Ambient method (Aiken et al., ) and the improved Canagaratna-Ambient method (Canagaratna et al., 1). pecies Ratio Aiken-Ambient Canagaratna- Ambient OA O/C.3.9 H/C OM/OC HOA O/C.11.1 H/C 1.. OM/OC COA O/C.1.19 H/C OM/OC VOOA O/C.33. H/C OM/OC LVOOA O/C.1.91 H/C OM/OC 1..3 Table. Comparison of aerosol properties and meteorological parameters at different stages during haze period. Average non-refractory submicrometer particulate matter (R- PM1) mass concentration (µg m -3 ) (Average ) Overall RH(%) / Temp( C) /19 9/1 7/1 7/1 /3 33 (m/s) Trace gas conc.(co./.1/3/././39/././3/./.1/37/.39/./.7 (ppm) and /O/ O /33 O3/(ppb)) PM, particulate matter; R-PM 1, non-refractory submicrometer particulate matter; RH, relative humidity 3

3 h h h h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) 1 13 ind peed (m/s) 1 13 ind peed (m/s) 1 13 ind peed (m/s) 31 h h h h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) h ind peed (m/s) 1 13 ind peed (m/s) 1 13 ind peed (m/s) 1 13 ind peed (m/s) 7 9 Figure 1. ind Rose plots for every hour colored by wind speed. Radial scales correspond to the frequency. 3

4 CDC Frequency (CDC) (a) /1 /1 (b). /1 / /7 /3. /3 / /9 Date and Time (KT).7 CDC /1 /1 /1 /1 / /7 /3. / / /.9 /11 / Figure. (a) Time series of composition dependent collection efficiency (CDC) and; (b) histogram of CDC, averaging. ±

5 Mass Conc. (µg/m 3 ) PM Density (g/cm 3 ) Mass Conc.(µg/m 3 ) (a) (b) /1 /1 /1 (c) r =.79 lope = 1.37 ±.1 / /7 /3 Date&Time(KT) /31/1 /11/1 /1/1 1 3 MP Volume (µm 3 /cm 3 ) Mass Conc. MP Vol. PM density /3 / /9 Frequency (Org. Density) Org. density /1 Date&Time (KT) (d) /1 /1 /1 / /7 /3 Org density PM density / / / Density bin mid-point (g cm -3 ) 1. /11 / Frequency (PM density) MP Vol. (µm 3 /cm 3 ) Org. Density (g/cm 3 ) Figure 3. (a) Time series of total particulate matter (PM1), scanning mobility particle sizer (MP) volume concentrations ; (b) Time series of the organic aerosol density estimated using the method reported in Kuwata et al. (1) ρorg = [1 1 (H/C) 1 (O/C)]/ [7 (H/C).1 (O/C)] and bulk aerosol density estimated from the measured chemical composition, known inorganic species density and the organic density estimated above (Zhang et al., ). (c) catter plot of the total PM1 mass (R-PM1 plus BC) versus MP volume, where the R-PM1 mass concentrations have been determined using the composition-dependent collection efficiencies; (d) histogram of organic aerosol density (average = 1.1 g cm -3 ) and bulk aerosol density (average = 1. g cm -3 ).

6 Ratios Canagaratna method H/C r =.9 lope = 1. ± OM/OC r =.99 lope = 1. ±.1 O/C r =.997 lope = 1. ±.1 1. Ratios Aiken method Figure. catter plot of OM/OC, O/C, and H/C calculated with the Canagaratna method vs. those with the Aiken-Ambient method

7 Residual (µg/m 3 ) (Resid / )/Q exp (Resid / )/Q exp caled Residual Mass Conc. (µg/m Q/Qexpected (c). Mass fraction (e) (f) (g) (h) (i) (a) Residual O3 O LV-OOA V-OOA COA HOA fpeak Measured Total Mass Reconstructed Total Mass /1/1 /1/1 /11/1 /1/1 /31/1 /1/1 Date and Time (KT) C x H y C x H y O 1 C x H y O H y O 1 C x H y p C x H y O z p 1 3 m/z Q for fpeak Best olution 3 7 umber of factors (p) 9 Q/Qexpected R, Time eries m/z (d) 1_ 1 _ (b) fpeak or seed. 1 3_ 1_3_3 3 9 Residual = (Measured - Reconstructed) 9 Q for p Current olution min Q for fpeakoreed.. R, Mass pectra

8 Figure. ummary of the key diagnostic plots of the chosen -factor solution ( OA factor solution) from PMF analysis of the organic aerosol fraction: (a) Q/Qexp as a function of the number of factors (p) explored in PMF analysis, with the best solution denoted by the open orange circle. Plots b-i are for the chosen solution set, containing factors: (b) Q/Qexp as a function of fpeak; (c) mass fractional contribution to the total mass of each of the PMF factors, including the residual (in purple), as a function of fpeak; (d) Pearson s r correlation coefficient values for correlations among the time series and mass spectra of the PMF factors. Here, 1 = O3, = O, 3 = LV-OOA, = V-OOA, = COA, = HOA; (e) box and whiskers plot showing the distributions of scaled residuals for each m/z; (f) time series of the measured mass and the reconstructed mass from the sum of the factors; (g) time series of the variations in the residual (= measured reconstructed) of the fit; (h) the Q/Qexp for each point in time; (i) the Q/Qexp values for each fragment ion.

9

10 (e) % of total 1 C x H y C x H y O 1 C x H y O H y O 1 C x H y p C x H y O z p O x O x H x O/C =.1, H/C =., /C =., OM/OC = 1.3 Factor 7 O/C =.1, H/C = 1.9, /C =.3, OM/OC = 1.33 Factor O/C =.37, H/C = 1.7, /C =., OM/OC = 1. Factor O/C =., H/C = 1., /C =.17, OM/OC =. Factor O/C =.7, H/C = 1.37, /C =.3, OM/OC =.19 Factor 3 O/C =.3, H/C = 1.9, /C =.19, OM/OC =.3 Factor (f) Mass Concentration (µg/m 3 ) O/C =., H/C = 1.33, /C =.3, OM/OC = m/z (amu) Factor /1/1 /11/1 /31/1 Date and Time (KT) Figure. Overview of two other solution sets and -factor solution including inorganic factors and signals from inorganic included PMF analysis: (a)(b) High resolution mass spectra and time series of the different OA factors from the -factor solution (3-OA); (c)(d) High resolution mass spectra and time series of the different OA factors from the -factor solution (-OA); (e)(f) High resolution mass spectra and time series of the different OA factors from the 7-factor solution (- OA). The mass spectra are colored by different ion families. 7 1

11 % of total (a) C x H y C x H y O 1 C x H y O H y O 1 C x H y p C x H y O z p O/C =.17, H/C = 1.9, /C =., OM/OC = 1. O/C =., H/C = 1.79, /C =.3, OM/OC = 1.9 O/C =.3, H/C = 1., /C =.1, OM/OC =. O/C =.7, H/C = 1., /C =.7, OM/OC =.1 3 m/z (amu) 7 Factor Factor 3 Factor Factor (b) Mass Concentration (µg/m 3 ) 1 /1/1 /11/1 /31/1 Date and Time (KT) % of total (c) C x H y C x H y O 1 C x H y O H y O 1 C x H y p C x H y O z p O/C =.1, H/C = 1.99, /C =., OM/OC = 1. O/C =.1, H/C = 1., /C =., OM/OC = 1.3 O/C =., H/C = 1.73, /C =., OM/OC = 1.79 O/C =.9, H/C = 1.3, /C =.19, OM/OC =. O/C =.71, H/C = 1.7, /C =., OM/OC =.1 3 m/z (amu) 7 Factor Factor Factor 3 Factor Factor (d) 1 Mass Concentration (µg/m 3 ) 3 1 /1/1 /11/1 /31/1 Date and Time (KT) 7 9 Figure 7. Overview of the -factor solution (a and b) and the -factor solution (c and d) from PMF analysis of the organic mass spectra only. The mass spectra are colored by different ion families. 3 11

12 C3: µg/m 3 C:1 µg/m 3 3 1% % 7% 19% % % 3 1% 1% % C:1 µg/m 3 1% 1% % 3 % C: µg/m % % % 11% % % 3 C1: µg/m 3 3% 1% 1% % 9% % % % 7% 13% % Cluster1 : 3% Cluster : % Cluster3 : 3% Cluster : 1% Cluster : % Figure. Averaged compositional pie chart of PM1 species (non-refractory-pm1 plus black carbon (BC)) in different clusters from the five cluster solution. The trajectories were released at half of the mixing height at the KIT (latitude: 37.; longitude: 17.) and the average arriving height for the back trajectories for this study was approximately 19 m. 1

13 H predicted (µg/m 3 ) =O x x 1/9 O 3 x 1/ Cl x 1/ r =.997 lope = 1. ± Date /31/1 /11/1 /1/1 1 1 H measured (µg/m 3 ) 3 Figure 9. catterplot that compares predicted H versus measured H concentrations. The predicted values were calculated assuming full neutralization of the anions (e.g., sulfate, nitrate, and chloride). The data points are colored by date

14 7 9 7 dorg/dlog 1 D va (µg m -3 ) 1h h 3h h h h 7h h 9h 1h 11h 1h 13h 1h 1h 1h 17h 1h 19h h 1h h 3h h (a) do /dlog 1 D va (µg m -3 ) 1h h 3h h h h 7h h 9h 1h 11h 1h 13h 1h 1h 1h 17h 1h 19h h 1h h 3h h (b) do 3 /dlog 1 D va (µg m -3 ) D va (nm) 1h h 3h h h h 7h h 9h 1h 11h 1h 13h 1h 1h 1h 17h 1h 19h h 1h h 3h h D va (nm) (c) dh /dlog 1 D va (µg m -3 ) D va (nm) 1h h 3h h h h 7h h 9h 1h 11h 1h 13h 1h 1h 1h 17h 1h 19h h 1h h 3h h D va (nm) (d) 77 7 Figure 1. ize distributions for (a) organic; (b) ulfate: (c) itrate; (d) ammonium for every hour. Y axis are shared for all distributions which are maximized to the axis Acculation mode Org. (µg/m 3 ) Fine mode Org. (µg/m 3 ) (a) nm Hour of Day (b) -11 nm Hour of Day 91 9 Figure 11. One-hour averaged diurnal profiles of organics in (a) accumulation mode and (b) fine mode. 1

15 ws 1 ws 1 ws 1 ws 1 ws Organic mass conc. ( g m 3 ) 1 3 itrate mass conc. ( g m 3 ) 1 ws ulfate mass conc. ( g m 3 ) Ammonium mass conc. ( g m 3 ) 1 ws 1 ws Chloride mass conc. ( g m 3 ) 1 ws VOOA mass conc. ( g m 3 ) LVOOA mass conc. ( g m 3 ) COA mass conc. ( g m 3 ) HOA mass conc. ( g m 3 ) ws 1 ws 1 ws 1 ws 1 ws BC mass conc. ( g m 3 ) CO (ppm) O (ppb).. O (ppb) 3 3 O 3 (ppb) ws umber Conc.(#/cm3) Figure 1. Polar plots of hourly averaged PM1 species concentrations (top row), mass concentrations of the five OA factors identified from PMF analysis (middle row), the mixing ratios of various gas phase species (second row from the bottom), and the particle number concentrations (bottom row) as a function of and direction. 1

16 fo (a) r =.9 lope =.3 ± Date 1//1 1/1/1 1/31/1 1/1/1 1/11/1 fo...3. r =.7 lope =.3 ± Date /31/1 /11/1 /1/ RH (%) 1. RH (%) Figure 13. catterplot of the variations of fo ratios as a function of RH (a) during winter (Kim et al., 17); (b) during spring in this study

17 wfact_ts fp_ 37 R m/z 9wfact_ts fp_ R m/z 9wfact_ts fp_ R m/z 9wfact_ts_3_fp_ R m/z Figure 1. Correlations between four OA factors and HRM ions that are segregated into five categories (CxHy, CxHyO, CxHyO, CxHyp and CxHyOzq ).From the top, HOA, COA, V- OOA and LV-OOA. 17

18 Mass Fraction CO CH3O O O O O CH9 CH11 CHO CH1O C7H1O 3 HOA COA V-OOA LV-OOA O O Figure 1. Mass fractional contribution of the factors (four OA factors two inorganic factors) from PMF analysis to various ions that are relevant to each significant tracer (a) 1 3% LVOOA 7% VOOA HOA COA % 17% Mass Percentages (%) (b) C x H y O z p C x H y O z p C x H y p H y O 1 C x H y O C x H y O 1 C x H y.%.%.3% 3.%.7%.7% LVOOA VOOA.1%.%.3% 1.% 7.7% COA.% 1.%.7% 1.% 71.9% HOA.13% 1.3%.7%.13%.%.9%.%.9%.%.9% 11.% 1.% Mass Percentages (%) 1 (c) HOA COA VOOA LVOOA.9%.%.7% 1.%.1% 3.7% 1.% 1.3% 17.%.%.3% 3.% 19.% 1.1%.% 1.1% 3.%.%.% 7.% C x H y C x H y O 1 C x H y O C x H y p C x H y O z p Figure 1. (a) Compositional pie chart of the average fractional contribution of each of the OA factors to the total OA over the campaign; (b) Average mass fractional contributions of seven ion families to each of the OA factors and; (c) Average mass fractional contributions of four OA factors to each ion families

19 3 33 /31/1... C3H7 (µg/m3) /1/1 /1/1 /1/1. r =. lope = 1.31 ±.1 r =.7 lope = 3.17 ± CH9 (µg/m3) r =.1 lope =.7 ± CH11 (µg/m3).3 r =.9 lope = 3.3 ± /11/1 BC(µg/m3) r =.9 lope = 39. ± 1. /11/ C3H3O (µg/m3) C3HO (µg/m3) r =.7 lope = 13. ± 1. r =.7 lope = 1 ±.7 1 /31/1 /11/1 /1/1-3 x1 CH1O (µg/m3) 1 1 /31/1 /31/1 /11/1 /11/1 x1-3 C7H1O (µg/m3) 1 1 /31/1 /11/1 /1/1 r =.7 lope = 1. ±.91 /1/1 /1/1 LV-OOA LV-OOA COA x1-3 CHO (µg/m3) 1 1 /11/1 /1/1 /1/1. 1 /31/1 /31/1 1 /31/1 39 /1/1 COA 391. /1/1 39 /1/1. COA COA HOA 1 39 /11/1.... CH7 (µg/m3) /31/1 /31/1 /11/1 3 HOA /11/1 1 /31/1 /11/1. r =.97 lope =.31 ± /31/ r =.9 lope = 1.79 ±.7 1 /11/1 37 r =.1 lope = 1. ±.3 HOA COA HOA 3 r =.7 lope = 1. ±.7 HOA CO (µg/m3) CH3O (µg/m3) Figure 17. catter plots of factors and relevant tracer ions colored by measurement date

20 f f (%) OOA,sub (%) 3 (a) (b) 1 VOOA (c) COA line Date /31/1 /11/1 LVOOA /1/1 VOOA COA HOA 1 1 f3 (%) LVOOA COA HOA HOA line 1 f7 OOA,sub (%) 1 17 f (%) 1 1 LVOOA VOOA 1 19 HOA 1 COA 3 f (%) Figure 1. Triangular plots of (a) f versus f3 and (b) f versus f (c) f,ooa sub versus f7, OOA sub for the five OA factors and all of the measured OA data (dots), colored by the time of the day. f3, f, and f are the ratios of the organic signal at m/z = 3,, and to the total organic signal in the component mass spectrum, respectively. f,ooa sub and f,ooa sub are the ratios of the organic signal at m/z, 7 after subtracting the contributions from V-OOA and LV-OOA (e.g., f,ooa sub = m/z - m/zv-ooa - m/zlv-ooa ; f7,ooa sub = m/z 7- m/z 7V-OOA - m/z 7LV- OOA)

21 LV-OOA (µg/m 3 ) (a) Date r =.3 lope =.7 ±.3 /31/1 /11/1 /1/1 V-OOA (µg/m 3 ) 1 1 Date (b) r =.1 lope =.1 ±.3 /31/1 /11/1 /1/ O x (ppb) 1 O x (ppb) Figure 19. catter plots of (a) LV-OOA vs. Ox; (b) V-OOA vs Ox over the entire period in the spring. ote that organic dominant period (/, 17: - /, :) are excluded. All scatter plots are colored by measurement date

22 T (ºC) ind (º) CO (ppm) O 3 (ppb) PM 1 (µg/m 3 ) Organic & O 3 (µg/m 3 ) % of PM 1 LV-OOA & V-OOA O/C... (µg/m 3 ) %OA OM/OC Organic dominant period Org O 3 O H BC dat HOA COA VOOA LVOOA CO O O 3 O /19 / /1 / /3 / / Date&Time(KT) Figure. Overview of the temporal variations of submicron aerosols at the Korea Institute of cience and Technology (KIT) in MA from May 19 to May including organic dominant period colored by yellow (May, 17:-May ::): (a) Time series of ambient air temperature (T), relative humidity (RH), and precipitation (Precip.); (b) Time series of wind direction (D), with colors showing different wind speeds (); (c) Time series of CO and O; (d) Time series of O3, and O; (e) Time series of total particulate matter (PM1), scanning mobility particle sizer (MP) volume concentrations and also shown are the h averaged PM1BC with bars. (f) Time series of the organic (Org.), nitrate (O3 - ), sulfate (O - ), ammonium (H ) and BC aerosols; (g) Time series of the mass fractional contribution of organic aerosols (Org.), nitrate (O3 - ), sulfate (O - ), ammonium (H ), chloride (Cl - ), and BC to total PM1 together with isoprene and toluene time series; (h) Time series of each factor RH (%) O (ppb) O (ppb) (µm 3 /cm 3 HOA & COA ) (µg/m 3 ) Tolune(ppb) (µg/m 3 ) H/C ms Prec. (m) MP Volume O,H x1-3 x /C Isoprene (ppb)

23 3 derived from the positive matrix factorization (PMF) analysis; (i) Time series of the mass fractional contribution to total organic aerosol (OA); (j) organic matter to organic carbon (OM/OC), oxygen to carbon (O/C), hydrogen to carbon (H/C) and nitrogen to carbon (/C) (Canagaratna et al., 1)

24 Figure 1. Long range transportation of plums from China to Korea during Haze period. Plots are from MODI, terra Figure. Two cluster solution of backtrajectroy analysis during haze period from (/ 7:- / :). The trajectories were released at half of the mixing height at the KIT (latitude: 37.; longitude: 17.) and the average arriving height for the back trajectories for this study was approximately 11 m.

25 r =.77 lope =.9 ± O3 1 /31/1 1 /9/1 /7/1 19 RH //1 1 Figure 3. catterplots between nitrate vs. RH during haze period

26 References Aiken, A. C., Decarlo, P. F., Kroll, J. H., orsnop, D. R., Huffman, J. A., Docherty, K.., Ulbrich, I. M., Mohr, C., Kimmel, J. R., ueper, D., un, Y., Zhang, Q., Trimborn, A., orthway, M., Ziemann, P. J., Canagaratna, M. R., Onasch, T. B., Alfarra, M. R., Prevot, A.. H., Dommen, J., Duplissy, J., Metzger, A., Baltensperger, U., and Jimenez, J. L.: O/C and OM/OC ratios of primary, secondary, and ambient organic aerosols with high-resolution time-of-flight aerosol mass spectrometry, nvironmental cience & Technology,, 7-, 1.11/es739q,. Canagaratna, M. R., Jimenez, J. L., Kroll, J. H., Chen, Q., Kessler,. H., Massoli, P., Hildebrandt Ruiz, L., Fortner,., illiams, L. R., ilson, K. R., urratt, J. D., Donahue,. M., Jayne, J. T., and orsnop, D. R.: lemental ratio measurements of organic compounds using aerosol mass spectrometry: characterization, improved calibration, and implications, Atmospheric Chemistry and Physics, 1, 3-7, 1.19/acp-1-3-1, 1. Kim, H., Zhang, Q., Bae, G.., Kim, J. Y., and Lee,. B.: ources and atmospheric processing of winter aerosols in eoul, Korea: insights from real-time measurements using a high-resolution aerosol mass spectrometer, Atmos. Chem. Phys., 17, 9-33, 1.19/acp , 17. Kuwata, M., Zorn,. R., and Martin,. T.: Using lemental Ratios to Predict the Density of Organic Material Composed of Carbon, Hydrogen, and Oxygen, nvironmental cience & Technology,, 77-79, 1.11/esq, 1. Middlebrook, A. M., Bahreini, R., Jimenez, J. L., and Canagaratna, M. R.: valuation of Composition- Dependent Collection fficiencies for the Aerodyne Aerosol Mass pectrometer using Field Data, Aerosol cience and Technology,, -71, 1.1/7.11.1, 1. Zhang, Q., Canagaratna, M. R., Jayne, J. T., orsnop, D. R., and Jimenez, J. L.: Time- and size-resolved chemical composition of submicron particles in Pittsburgh: Implications for aerosol sources and processes, Journal of Geophysical Research-Atmospheres, 11, 1.19/jd9,.