ars.els-cdn.com · Web viewText S1 Measurements of VOCs at HT and TC. Text S2 Calculations of...
Transcript of ars.els-cdn.com · Web viewText S1 Measurements of VOCs at HT and TC. Text S2 Calculations of...
Supplement of the manuscript “New particle formation and growth at a suburban site and
a background site in Hong Kong”
X.P. Lyu 1,2, H. Guo 1,2*, H.R. Cheng 3# and D.W. Wang 4
1 Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University,
Hong Kong
2 Research Institute for Sustainable Urban Development, The Hong Kong Polytechnic University,
Hong Kong
3 Department of Environmental Engineering, School of Resource and Environmental Sciences,
Wuhan University, Wuhan, China
4 State Key Laboratory of atmospheric boundary layer physics and atmospheric chemistry,
Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
* First corresponding author: [email protected]
# Second corresponding author: [email protected]
List of Texts:
Text S1 Measurements of VOCs at HT and TC.
Text S2 Calculations of formation rate of 5.5 nm particles, condensation sink, sulfuric acid
concentration and production of organic condensable vapor peroxies.
List of Figures:
Figure S1 Comparison of hourly VOC mixing ratios between on-line and off-line analyses
during the sampling campaign at TC in autumn of 2013.
Figure S2 Box plots of PNNuc, PNAit, PNAcc and total PN at (a) HT and (b) TC. The statistics
presented include minimum, 10% percentile, 25% quartile, median, mean, 75% quartile, 90%
percentile and maximum.
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure S3 Temporal variations of trace gases monitored at TC through January 2011 to
December 2013.
Figure S4 Surface plot of particle number distribution and the time-dependent variation of CS on
(a) December 10, 2011 at HT and (b) October 3, 2013 at TC. The white makers represent CS.
Figure S5 Diurnal variations of PNNuc, PNAit and PNAcc on (a) November 08, (b) November 09, (c)
November 19 and (d) November 26 observed at a mountainous site in Hong Kong in 2010. The
increases of PNNuc in the afternoon when PNAit and PNAcc were decreasing are highlighted in the
red box.
Figure S6 Scatter plots of (a) GRNuc and GR0 at HT, (b) GRAit and GR0 at HT, (c) GRNuc and GR0
at TC and (d) GRAit and GR0 at TC. Points above and on the one-to-one line represent that 100%
of the observed GR are explained by sulfuric acid.
List of Tables:
Table S1 Instruments, measurement techniques and detection limits for trace gases monitoring at
HT and TC.
Table S2 Average mixing ratios of trace gases concurrently monitored at HT and TC from
August to December, 2011 (unit: ppbv).
Table S3 Average proxies productions of the organic condensable vapors for the excluded and
included samples in Figure 6 (Unit: molecules/cm3/s×104). Bold font denotes significant
differences between the two groups of samples (p<0.05).
Text S1 Measurements of VOCs at HT and TC.
Thirty VOC species (11 alkanes, 10 alkenes and 9 aromatics) were measured at half an hour
interval by an online analytical system (Syntech Spectra GC 955, Series 600/800, the
Netherlands) at HT. The quality of VOC measurement was regularly checked by Dr. Donald
Blake’s group in University of California, Irvine, an internationally recognized group in VOCs
analysis (Colman et al., 2001; Simpson et al., 2011). The accuracy, precision and detection limit
of the real-time VOC measurements were 1-10%, 2.5-20% and 2-787 pptv, respectively.
2
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Procedures about quality control and assurance can be found in Lyu et al. (2016). At TC, the
whole air samples (WAS) were collected using the 2L stainless steel canisters. Prior to sampling,
the canisters were repeatedly filled with humidified pure nitrogen and evacuated for at least three
times. During sampling, a valve was used to control the flow rate to make the sampling last for 1
hour. Afterwards, the samples were delivered to the State Key Laboratory of organic
geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences for chemical
analysis. In total, 84 VOC species were quantified with the combined application of gas
chromatography (GC) and mass selective detector (MSD)/ flame ionization detector (FID)/
electron capture detector (ECD). Details about the instruments, configurations and quality
control protocols are given in Zhang et al. (2012). It is noteworthy that monoterpenes (α-/β-
pinenes) at TC were analyzed by the offline detection system, which were not available at HT
due to the limitation of the online measurement technique. Moreover, the gas standard used for
VOC analysis was provided by Dr. Donald Blake’s group at UC Irvine (Colman et al., 2001;
Simpson et al., 2011), which included 26 halocarbons, 8 alkyl nitrates and 61 hydrocarbons.
While α-/β-pinenes were available, d-limonene was not included in the mixed standard.
Therefore, d-limonene which might also play important role in NPF was not measured in this
study.
Text S2 Calculations of formation rate of 5.5 nm particles, condensation sink, sulfuric acid
concentration and production of organic condensable vapor peroxies.
According to Sihto et al. (2006), J5.5 was calculated according to Equation (S1) (Sihto et al.,
2006; Guo et al., 2012). J5.5 was calculated due to the fact that 5.5 nm was the smallest size that
can be detected by our SMPS. Since 5.5 nm was not the initial size of clusters formed by
condensable vapors, J5.5 actually reflected the overall rate of cluster formation and growth of
particles to 5.5 nm, with the consideration of sinks of particles due to coagulation and further
growth into larger size particles.
J5.5=dN 5.5−10
dt+CoagSdp=7nmN 5.5−10+
14.5nm
GR5.5−10N 5.5−10 (Equation S1)
3
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
where N5.5−10 refers to the total particle number in size range of 5.5-10 nm. Coagulation sink of
particle in this size range is approximated by the coagulation loss of 7 nm particles, under the
assumption that 7 nm is approximately the geometric mean diameter (GMD) of particles between
5.5 and 10 nm. GR5.5−10 denotes the growth rate of 5.5-10 nm size range particles, which was
calculated through the linear regression between GMD5.5-10 and time. Apart from GR5.5−10, growth
rates for particles in modes of nucleation (GR5.5-24.7) and Aitken (GR24.7-101.4) were also calculated
in the same way.
Condensation sink (CS) is an important pathway for particle loss, which is an integration of
particle loss due to condensation onto pre-existing particles over the whole size spectrum.
Therefore, it was calculated as follows (Kulmala et al., 2005).
CS=2πD∑iβM DpN (Equation S2)
where D represents the diffusion coefficient of the condensable vapor. βM is the transitional
regime correction factor, which increases with the growth of particles and approaches 1 for large
size particles. D p and N are the particle diameter and number in a specific size bin (i),
respectively. i is an arbitrary size bin of the 44 size bins divided by SMPS. In this study, CS was
calculated over the size range of 5.5-350.4 nm. It was expected that the particles larger than
350.4 nm made very minor contribution (<4%) to total CS, according to the observations of 10-
1000 nm particles at a roadside site in Hong Kong (unpublished data).
It has been a consensus that sulfuric acid vapor plays key roles in NPF. The concentration of
sulfuric acid (Qsa) was estimated following Equation (S3) (Mikkonen et al., 2011).
Qsa=8.21×10−3 · k · Radiation·[ SO2]0.62 ·(CS ·RH )−0.13 (Equation S3)
where k is a constant (1.035), and RH represents relative humidity. Qsa required for the particle
growth at the rate 1 nm/h in different size ranges was calculated according to Nieminen et al.
(2010), with the consideration of hydration effect.
In addition, to understand the roles of condensable organic vapors (particularly HOMs) in NPF,
the production rates of the first-generation oxidation products of VOCs were used as the proxies
of organic condensable vapors in this study. For example, the potential of condensable vapor
4
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
formation from ozonolysis of α-pinene was represented by kO 3+α− pinene · [O¿¿3] ·[α−pinene ]¿, so
were for O3 and OH oxidation of other VOCs. The reaction rates of O3+VOCs and OH+VOCs
were taken from the website of Master Chemical Mechanism (MCM, accessible at
http://mcm.leeds.ac.uk/MCM/). OH concentration was simulated with a photochemical box
model incorporating MCM (Lyu et al., 2016). The VOC species considered include isoprene, α-
pinene, β-pinene, benzene, toluene, ethylbenzene, xylenes and trimethylbenzenes. It should be
kept in mind that the proxies of the production of organic condensable vapors estimated in this
method should have uncertainties, due to different volatilities of the oxidation products.
However, to some extent it can reflect the potentials of VOCs in forming condensable vapors,
because all the following reactions are initiated by the first-generation oxidations of VOCs.
Figure S1 Comparison of hourly VOC mixing ratios between on-line and off-line analyses
during the sampling campaign at TC in autumn of 2013.
5
106
107
108
109
110
111
112
113
114
115
116
117
118
119
Figure S2 Box plots of PNNuc, PNAit, PNAcc and total PN at (a) HT and (b) TC. The statistics
presented include minimum, 10% percentile, 25% quartile, median, mean, 75% quartile,
90% percentile and maximum.
6
120
121
122
123
124
125
Figure S3 Temporal variations of trace gases monitored at TC through January 2011 to
December 2013.
Figure S4 Surface plot of particle number distribution and the time-dependent variation of
CS on (a) December 10, 2011 at HT and (b) October 3, 2013 at TC. The white makers
represent CS.
7
126
127
128
129
130
131
132
133
Figure S5 Diurnal variations of PNNuc, PNAit and PNAcc on (a) November 08, (b) November
09, (c) November 19 and (d) November 26 observed at a mountainous site in Hong Kong in
2010. The increases of PNNuc in the afternoon when PNAit and PNAcc were decreasing are
highlighted in the red box.
8
134
135
136
137
138
139
140
Figure S6 Scatter plots of (a) GRNuc and GR0 at HT, (b) GRAit and GR0 at HT, (c) GRNuc and
GR0 at TC and (d) GRAit and GR0 at TC. Points above and on the one-to-one line represent
that 100% of the observed GR are explained by sulfuric acid.
Table S1 Instruments, measurement techniques and detection limits for trace gases
monitoring at HT and TC.
Site Species Instrument Measurement technique Detection limit
HT SO2 API 100 E UV fluorescence 0.4 ppbv
CO API 300 Red absorption with gas filter
correlation
<0.050 ppm
NO-NO2-
NOx
API 200A Chemiluminescence 0.4 ppbb
O3 TEI 49, 49C and
49i
UV absorption 1.0 ppbv
TC SO2 API 100E UV fluorescence 0.4 ppbv
CO API 300 Red absorption with gas filter
correlation
<0.050 ppm
NO-NO2-
NOx
API 200A Chemiluminescence 0.4 ppbb
O3 API 400 UV absorption 0.6 ppbv
9
141
142
143
144
145
146
147
148
Table S2 Average mixing ratios of trace gases concurrently monitored at HT and TC from
August to December, 2011 (unit: ppbv).
HT TC p value of t-test
SO2 4.0 ± 0.3 4.5 ± 0.3 <0.05
CO 235.2 ± 19.0 541.7 ± 12.8 <0.001
NO 1.5 ± 0.2 18.0 ± 1.1 <0.001
NO2 10.5 ± 0.7 18.8 ± 0.9 <0.001
O3 27.1 ± 1.8 20.1 ± 1.7 <0.001
Table S3 Average proxies productions of the organic condensable vapors for the excluded
and included samples in Figure 6 (Unit: molecules/cm3/s×104). Bold font denotes significant
differences between the two groups of samples (p<0.05).
Pathway Excluded samples Included samples
OH + isoprene 87.5 ± 47.1 77.8 ± 31.9
O3 + isoprene 8.5 ± 2.0 11.4 ± 2.4
OH + α-pinene 1.9 ± 1.4 0.7 ± 0.3
O3 + α-pinene 2.2 ± 0.3 1.5 ± 0.3
OH + β-pinene 6.0 ± 1.8 4.5 ± 1.4
O3 + β-pinene 1.1 ± 0.8 1.1 ± 0.3
OH + aromatics 130.5 ± 69.1 149.5 ± 83.0
10
149
150
151
152
153
154
155