Auxiliary Materials for
Predicting hygroscopic growth using single particle chemical composition estimates
Robert M. Healy1,2, Greg J. Evans1, Michael Murphy1, Zsófia Jurányi3*, Torsten Tritscher3**,
Marie Laborde3***, Ernest Weingartner3*, Martin Gysel3, Laurent Poulain4, Katharina A.
Kamilli4, Alfred Wiedensohler4, Ian P. O’Connor2, Eoin McGillicuddy2, John R. Sodeau2, John
C. Wenger2
1
Southern Ontario Centre for Atmospheric Aerosol Research, University of Toronto, 200 College
Street, Toronto, Ontario, Canada
2
Department of Chemistry and Environmental Research Institute, University College Cork,
Ireland
3
Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, 5232 Villigen PSI, Switzerland
4
Leibniz Institute for Tropospheric Research, Leipzig, Germany
*Now at Institute of Aerosol and Sensor Technology, University of Applied Sciences,
Northwestern Switzerland 5210, Windisch, Switzerland
**Now at TSI GmbH, Aachen, Germany
***Now at Ecotech Pty Ltd, Knoxfield, Australia
Introduction:
This auxiliary material contains a comparison of aerosol composition at the LHVP and SIRTA
sites in Paris, located 20 km apart.
Bulk aerosol composition:
As shown in Fig. S1, on average, the bulk composition of aerosol detected (taken as the sum of
organic aerosol, ammonium, sulphate and nitrate measured by AMS and BC measured by
aethalometer) is quite similar across Paris, due to the strong influence of regional to continental
scale emissions upon local air quality, as discussed in detail by Crippa et al. (2013). Instrumental
details are also described by Crippa and coauthors. Temporally, good correlations are also
observed for hourly average mass concentrations of each species between the two sites (R2 = 0.560.89), as shown in Figs. S2-S6.
The distribution of hourly mean values for each chemical species can also be compared between
the two sites. Assuming that p < 0.05 indicates that there is a significant difference between
datasets, the hourly average mean mass concentration values of organic aerosol and nitrate
observed at the SIRTA and LHVP sites are highly similar, exhibiting no statistically significant
difference between sites (p = 0.56 and 0.08, respectively), as shown in Fig. S7. These two species
account for the majority of the aerosol mass at both sites. For ammonium, sulphate and black
carbon, however, a significant difference is observed (p < 0.05). This is likely due to differences
in local sources, such as vehicular and cooking activities discussed in the next section, at each site.
Despite these differences, median values agree within 10% and mean values agree within 30% for
all species when comparing the two sites.
Organic aerosol composition:
Good agreement is observed for the average bulk organic aerosol composition between sites as
shown in Fig. S8. Relatively poor temporal agreement is observed for HOA and COA (R2 = 0.35
and 0.14, respectively) between sites. However, BBOA and OOA are more associated with
regional scale emissions and thus the values observed at both sites are highly similar. These factors
also exhibit similar temporality at both sites (R2 = 0.62 and 0.96, respectively) and their combined
contribution to the total bulk organic aerosol mass is very high (>70%). Thus the bulk composition
of organic aerosol measured at both sites is relatively similar Figs. S9-12, as previously described
in detail by Crippa et al. (2013).
No significant difference is observed between hourly mean mass concentration values of BBOA
and OOA measured at the LHVP and SIRTA sites (p = 0.46 and 0.53, respectively), as shown in
Fig. S13. These factors combined account for >70% on average of the organic aerosol mass at each
site. For COA and HOA, however, there is a significant difference observed (p < 0.01), which is
expected considering the differences in traffic and cooking activities at each location.
Bulk aerosol composition-based predictions of hygroscopic growth:
A comparison of hygroscopic growth predictions based on bulk aerosol composition at the LHVP
and SIRTA sites, hygroscopic growth predictions based on size-resolved ATOFMS data at the
LHVP site, and HTDMA hygroscopic growth measurements are provided in Figs. S14-16.
Ammonium nitrate evaporation artefact:
The relationship between hygroscopic growth predictions based on bulk aerosol composition at
the SIRTA site and HTDMA hygroscopic growth measurements at the SIRTA site as a function
of nitrate content is provided in Fig S17.
References:
Crippa, M., et al. (2013), Wintertime aerosol chemical composition and source apportionment of
the organic fraction in the metropolitan area of Paris, Atmos. Chem. Phys., 13(2), 961-981,
doi:10.5194/acp-13-961-981.
Figures:
1. fs01.pdf (Figure S1): Average bulk aerosol composition at the SIRTA and LHVP sites
2. fs02.pdf (Figure S2): Temporal variation and linear regression of organic aerosol mass
concentrations detected by AMS instruments at the LHVP and SIRTA sites (N = 496).
3. fs03.pdf (Figure S3): Temporal variation and linear regression of ammonium mass
concentrations detected by AMS instruments at the LHVP and SIRTA sites (N = 496).
4. fs04.pdf (Figure S4): Temporal variation and linear regression of nitrate mass concentrations
detected by AMS instruments at the LHVP and SIRTA sites (N = 496).
5. fs05.pdf (Figure S5): Temporal variation and linear regression of sulphate mass concentrations
detected by AMS instruments at the LHVP and SIRTA sites (N = 496).
6. fs06.pdf (Figure S6): Temporal variation and linear regression of black carbon mass
concentrations detected by aethalometer instruments at the LHVP and SIRTA sites (N = 496).
7. fs07.pdf (Figure S7): Box plot of mass concentrations of each species (hourly mean values, N
= 496) measured at the LHVP and SIRTA sites. Horizontal lines, boxes and whiskers represent
the median, 75th and 90th percentile values, respectively.
8. fs08.pdf (Figure S8): Average bulk organic aerosol composition derived from AMS PMF
analysis at the SIRTA and LHVP sites
9. fs09.pdf (Figure S9): Temporal variation and linear regression of biomass burning organic
aerosol (BBOA) mass concentrations detected by AMS instruments at the LHVP and SIRTA
sites (N = 496).
10. fs10.pdf (Figure S10): Temporal variation and linear regression of oxidized organic aerosol
(OOA) mass concentrations detected by AMS instruments at the LHVP and SIRTA sites (N
= 496).
11. fs11.pdf (Figure S11): Temporal variation and linear regression of hydrocarbon-like organic
aerosol (HOA) mass concentrations detected by AMS instruments at the LHVP and SIRTA
sites (N = 496).
12. fs12.pdf (Figure S12): Temporal variation and linear regression of cooking organic aerosol
(COA) mass concentrations detected by AMS instruments at the LHVP and SIRTA sites (N
= 496).
13. fs13.pdf (Figure S13): Box plot of mass concentrations of each organic aerosol factor (hourly
mean values, N = 496) measured at the LHVP and SIRTA sites. Horizontal lines, boxes and
whiskers represent the median, 75th and 90th percentile values, respectively.
14. fs14.pdf (Figure S14): Comparison of HTDMA measured growth factors at the SIRTA site
and ATOFMS predicted growth factors at the LHVP site for particles of different size for
each hour of the measurement period, together with bulk composition predicted growth
factors (N=496 hours). Horizontal lines, boxes and whiskers represent the median, 75th and
90th percentile values, respectively. The “x” markers represent mean values.
15. fs15.jpg (Figure S15) Top panel: Temporal trends for ATOFMS mean predicted GF values
(265±20 nm particles) and measured HTDMA mean GF values for 265 nm particles, together
with bulk aerosol mean predicted GF values at both sites. Middle panel: ATOFMS predicted
growth factor probability density function for 265±20 nm particles. Bottom panel: HTDMA
measured growth factor probability density function for 265 nm particles.
16. Fs16.pdf (Figure S16) Top left: Comparison of bulk aerosol predicted mean GF values at the
SIRTA site and measured HTDMA mean GF values for 265 nm particles for each hour of the
entire measurement period (N=496 hours), coloured as a function of inorganic aerosol content.
Bottom left: The same comparison but for periods influenced by continental transport only
(N=175 hours). Top middle: Comparison of bulk aerosol predicted mean GF values at the
LHVP site and measured HTDMA mean GF values for 265 nm particles for each hour of the
entire measurement period (N=496 hours), coloured as a function of inorganic aerosol content.
Bottom middle: The same comparison but for periods influenced by continental transport only
(N=175 hours). Top right: Comparison of ATOFMS predicted GF values (265±20 nm) and
measured HTDMA mean GF values for 265 nm particles for each hour of the entire
measurement period (N=496 hours), coloured as a function of inorganic aerosol content.
Bottom right: The same comparison but for periods influenced by continental transport only
(N=175 hours).
17. fs17.pdf (Figure S17) Comparison of bulk aerosol predicted mean GF values at the SIRTA site
and measured HTDMA mean GF values for 265 nm particles for each hour of the entire
measurement period (N=496 hours), coloured as a function of nitrate content.