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: ceguohai@polyu.edu.hk

# Second corresponding author: chenghr@whu.edu.cn

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