Results and Discussions

where a, is a mass-based stoichiometric coefficient for the reaction product / and Kom i (m3 jug1) is a normalized partitioning coefficient of product /. Eq 1 can be simplified to one-product model (i.e., /=1) or two-product model (i.e., i=2) by assuming that all semi-volatile products can be classified into one or two groups by their condensabilities. Parameters (a and Kom) are obtained by fitting experimentally determined SOA yields with least square method. Since hundreds of products are generated from ROGs oxidation, parameters obtained numerically have no actual physical meaning, but represent the overall properties of all the products (10).

Particle Wall Loss Correction

The SOA yields were calculated at the time when measured particle volume concentration reached maximum. A density of 1 g cm"3 was assumed to convert organic aerosol volume concentration to mass concentration. Here, the measured aerosol concentration should be corrected to take into account the particle wall loss due to deposition on the Teflon film. The correction method described in Takekawa et al. (12) is adopted since most experiments in this work were conducted with highly concentrated preexisting seed aerosols and coagulation may not be neglected. In this method, the aerosol deposition rate constant (kdep, h"1) is related with particle diameter (dp, nm) by a four-parameter equation:

kdep values of different dp are determined by monitoring the particle number decay under dark condition at a low number concentration (< 1000 particles cm" 3) to avoid coagulation. The optimized values of parameter a~d are 4.37x10"4, 0.925, 84.3 and 1.40, respectively. It should be noticed that the deposited aerosol concentration estimated by this method might introduce some error due to the low particle number concentration (12).

Survey Experiments for Dry (NH4)2S04 Seed Effect

Six survey experiments (S1-S6) were first conducted to test whether the highly concentrated dry (NH4)2S04 seed aerosols have effects on ozone and SOA formation in three aromatic hydrocarbon/NOx photooxidation systems, and their initial conditions and results are summarized in Table I. For each hydrocarbon, two similar experiments were conducted under identical initial conditions except the presence of (NH4)2S04 aerosols. Figure 2 shows the variations of NOx-NO, 03 and PMcorTCCted-PM0 (i.e. the generated SOA) concentration with time in these experiments. The results indicate that:

• The presence of highly concentrated dry (NH4)2S04 aerosols has no significant effect on gas-phase reactions. The trend of NOx-NO and 03 concentration variations with time is nearly the same for the particle-free and particle-introduced experiments, and the same statement can be made for NO and hydrocarbon consumptions. The small difference of the 03 and NOx-NO profiles may be due to the small discrepancy of initial experimental conditions.

• The presence of highly concentrated dry (NH4)2S04 aerosols obviously enhances organic aerosol generation and increases SOA yield. The surface of dry (NH4)2S04 particles may not be inert, and may influence the gas-particle partitioning process.

Table I. Initial Conditions and Results of Survey Experiments (303K)

Hydrocarbons_Toluene m-Xylene 1,2,4-TMBa

Table I. Initial Conditions and Results of Survey Experiments (303K)

Hydrocarbons_Toluene m-Xylene 1,2,4-TMBa

Experiment No.

SI

S2

S3

S4

S5

S6

[NO]0

(ppb)

151

151

161

166

136

137

[NOJ„

(ppb)

302

303

333

348

277

282

[HC]o

(ppm)

1.83

1.90

2.00

2.07

1.97

2.14

[PM]0

(um3cm'3)

0

95

0

74

0

51

[HCMNOJo

(ppm ppm1)

6.1

6.3

6.0

6.0

7.1

7.6

RH

(%)

61

61

53

51

57

59

[03]max

(ppb)

270

263

412

410

501

498

TimetO,]m«

00

3.5

3.5

1.5

1.5

2.8

2.8

[NOj-NO]max

(ppb)

230

232

286

297

230

232

Time[NO,-NO]m.x

00

1.4

1.4

0.5

0.6

1.0

1.0

A[HC]reacted

(ppm)

0.34

0.32

0.45

0.47

0.44

0.43

AM„

Cwgm3)

88

122

148

249

44

69

SOA Yield

(%)

7.1

10.6

7.6

12.3

2.1

3.3

a 1,2,4-Trimethylbenzene.

a 1,2,4-Trimethylbenzene.

Effect of Dry (NH4)2S04 Seeds on SOA Formation

Comparison of the SOA Yields between Aromatic Hydrocarbons

Using SOA yield curves calculated from eq 1, a series of experiments were further conducted to investigate how dry (NH4)2S04 particles affect SOA

formation. To avoid the potential impacts of different parameters on SOA formation, the hydrocarbon to NOx ratio, temperature and RH were fixed. The detailed experimental conditions are described later. Experiments in the absence of (NH4)2S04 seeds were first performed to obtain a base SOA yield curve for each hydrocarbon. Figure 3 shows the particle-free aerosol yield curves for three aromatic hydrocarbons. The yield curves are produced by fitting the experimental data to a one-product model based on eq 1. Values of a and Kom for each hydrocarbon are as follows: 0.148 and 0.022 for toluene, 0.109 and 0.017 for w-xylene, and 0.062 and 0.011 for 1,2,4-trimethylbenzene.

Figure 3 shows that among the three aromatic hydrocarbons, the SOA yield of toluene is the highest, and that of 1,2,4-trimethylbenzene is the lowest, with m-xylene in the middle. This result is consistent with previous work of Odum et al. (//) and Takekawa et al. (12). Odum et al. (11) indicated that the aromatic species containing one or fewer methyl substituent have higher SOA yield than those containing two or more methyl substituents. Thus, toluene is classified as a "high-yield" aromatic species, and m-xylene and 1,2,4-trimethylbenzene as "low-yield" species. Takekawa et al. (12) pointed out that an aromatic hydrocarbon with a lower reaction rate constant with OH radicals may have a higher secondary to primary reaction rate ratio, which is proportional to the SOA yields. Since the order of reaction rate constants of the three aromatic hydrocarbons with OH radicals are toluene, m-xylene and 1,2,4-trimethylbenzene from low to high, our experimental results are in good agreement with Takekawa's theory.

Toluene/NOx Photooxidation Systems

The initial experimental conditions and results for toluene experiments are shown in Table II, and the SOA yields versus M0 are shown in Figure 4. The aerosol yields obtained with dry (NH4)2S04 aerosols are higher than particle-free ones, and it confirms the conclusion derived from survey experiments that the presence of highly concentrated dry (NH4)2S04 aerosols increases SOA yields.

As shown in Figure 4, the presence of dry (NH4)2S04 seed aerosols increases SOA yields, and the data appears to fall into several different yield curves. To identify which factor is the intrinsic cause for the effect of dry (NH4)2S04, detailed seed information of experiments AS1-AS9 are listed in Table III. PMv>g and PMs>g in Table III stand for the volume and surface PM (i.e. (NH4)2S04) concentrations when SOA starts to generate. Here, the time when SOA begins to form is selected at the time when measured PM decreases to a minimum. Before SOA generation, the measured aerosol volume concentration will continue decreasing due to the deposition. Since the SOA generation process is very fast and die measurement interval of SMPS is 6 minutes, the PM minimum point is approximately equal to the real SOA formation point. After checking the data, it is reasonable to categorize these data into four groups based

Figure 2. Variations of NOx-NO, 03 and PMcorrectecrPMo concentration in survey experiments of toluene (A), m-xylene (B) and 1,2,4-trimethylbenzene (C).

Figure 2. Variations of NOx-NO, 03 and PMcorrectecrPMo concentration in survey experiments of toluene (A), m-xylene (B) and 1,2,4-trimethylbenzene (C).

Figure 3. SOA yield (Y) variation with organic aerosol mass concentration (MJ for seeds-free aromatic hydrocarbons/NOx photooxidation experiments.

Figure 3. SOA yield (Y) variation with organic aerosol mass concentration (MJ for seeds-free aromatic hydrocarbons/NOx photooxidation experiments.

Group 1, without (NH4)2S04

Group 2, PMs g~13 cm2 m4

Figure 4. SOA yields variation with organic aerosol mass concentration for toluene/NOx photooxidation experiments.

Group 1, without (NH4)2S04

Group 2, PMs g~13 cm2 m4

Figure 4. SOA yields variation with organic aerosol mass concentration for toluene/NOx photooxidation experiments.

Table II. Summary of Initial Experiment Conditions and Results for Toluene/NOx Photooxidation Experiments. (303K, 58~62%RH)

Experiment

HCo

NO0

NOx,o

PM0

HC0

A HC

M0

Y

No.

(ppm) (ppb) (ppb) (um cm )

NOIfi

(fgm) (f*gm~)

(%)

Tol-1

1.09

49

96

0

11.4

585

43

1A

Tol-2

2.02

107

216

0

9.4

941

91

9.6

Tol-3

2.50

125

251

0

9.9

1090

112

10.2

Tol-4

2.85

143

289

0

9.9

1194

134

11.2

Tol-5

4.30

212

443

0

9.7

1655

200

12.1

Tol-6

5.10

250

511

0

10.0

1911

240

12.5

Tol-7

1.03

46

94

50

11.0

583

56

9.7

Tol-8

3.26

165

325

73

10.0

1297

164

12.7

Tol-9

4.14

209

442

72

9.4

1579

214

13.6

Toi-10

4.92

260

517

78

9.5

1816

260

14.3

Tol-11

1.24

56

118

67

10.5

634

76

12.0

Toi-12

1.93

105

208

74

9.3

840

112

13.3

Tol-13

2.52

122

244

65

10.3

1107

177

16.0

Toi-14

3.10

155

305

60

10.2

1382

217

15.7

Toi-15

1.04

49

102

75

10.2

534

75

13.9

Toi-16

1.41

70

142

85

9.9

604

92

15.3

Toi-17

2.01

104

210

70

9.6

951

172

18.1

Tol-18

2.51

123

245

66

10.2

1121

209

18.7

Table III. Information and Results of Categorization for Experiments with/without (NH4)2S04 Seed Aerosols

Group No.

Experiment No.

' ) (um cm

PMsg 3) (cm2 m3)

a

K-om (M* m3)

R2

1

Tol-l~Tol-6

0

0.148

0.022

0.99

2

Tol-7

50

46

12.2

0.161

0.025

0.98

Tol-8

73

57

12.9

Tol-9

72

62

13.2

Toi-10

78

57

13.2

3

Tol-11

67

57

14.7

0.198

0.020

0.95

Toi-12

74

68

15.3

Toi-13

65

59

15.7

Toi-14

60

56

14.4

4

Toi-15

75

71

16.6

0.230

0.021

1.00

Toi-16

85

82

18.5

Toi-17

70

66

17.4

Toi-18

66

62

17.1

on the different PMs>g concentration. Group 1 includes results from particle-free experiments (Tol-l~Tol-6), while group 2 to 4 consist of those from experiments with PMSjg of about 13, 15 and 17 cm2 m"3, respectively. Using best-fit one-product model (/=1 in eq 1), four aerosol yield curves are obtained in Figure 4, and it indicates that aerosol yield increases with increasing particle surface concentration under the same organic aerosol mass concentration. PMsg characterizes surface concentration of dry (NH4)2S04 aerosol when organic components start to partition onto the seed particles. Therefore, the dry (NH4)2S04 aerosol effect is believed to correlate with the variation of aerosol seed surface area.

The fitted values of a and Kom for each group are listed in Table III. The fitting results show that partitioning coefficient Kom has little change with the variation of PMsg value, which implies that the overall condensability of condensable organic compounds (CCs) does not change. However, stoichiometric coefficient a increases with PMsg, which means that more aerosol-forming products (i.e. CCs) are generated with a higher surface concentration of (NH4)2S04 aerosol.

m-Xylene/NOx and 1,2,4- Trimethylbenzene/NOx Photooxidation Systems

The effect of (NH4)2S04 PMs g values on SOA curves are also investigated for m-xylene and 1,2,4-trimethylbenzene (25). All experiments are conducted at 303K and about 60% RH. The hydrocarbon to NOx ratios for m-xylene and 1,2,4-trimethylbenzene were fixed at about 6 and 8, respectively. The results are shown in Figure 5. Similar to toluene, the SOA yield data points can be categorized into several different groups based on the PMs g values. In addition, the stoichiometric coefficient a increases with PMsg increase, while the partitioning coefficient Kom is nearly unchanged. In the next section, we will provide one possible interpretation for the effect of (NH4)2S04 aerosol surface concentration on SOA formation.

Hypothesis of Dry (NH4)2S04 Seeds Effect

Unlike (NH4)2S04, it was shown that the presence of highly concentrated dry CaS04 and aqueous Ca(N03)2 seed aerosols do not have significant effect on SOA formation (25, 26). After comparing the molecular composition of these three different inorganic compounds, we can find that both CaS04and Ca(N03)2 are neutral inorganic seeds, while (NH4)2S04 has some weak acidity, which might account for its effect on SOA formation. It was reported that organic compounds have little effect on the DRH of pure inorganic particles (27), which means the dry (NH4)2S04 particle will continue to be solid when it is covered by an organic layer. The organic layer contains both hydrophilic and hydrophobic

o Group 2, PMsg~13 cm2 m"3

100 150

M0 Ufl m3)

Figure 5. SOA yields variation with organic aerosol mass concentration for m-xylene/NOx (A) and l,2,4-trimethylbenzene/NOx (B) photooxidation experiments.

species, and the hydrophilic species could absorb water molecules from the gasphase, especially under high RH condition. Then, under the interaction of (NH4)2S04 and water-contained organic material, the aerosol surface may show some acidity due to the hydrolysis of NH4+.

It has been reported that strong acidic aqueous seeds can cause the heterogeneous reactions of carbonyl compounds, and enhance SOA yield by accelerating the formation of larger oligomers (13, 14). It is possible that the acidic (NH4)2S04 surface, similar to this strong acid effect, could also facilitate the acid-catalyzed heterogeneous reactions of carbonyl compounds, and produce oligomer from carbonyl monomer (15). The oligomerization of condensable organic compounds (CCs) will not change the amount of aerosol-forming products themselves, which means the mass-based stoichiometric coefficient a should not be changed. However, the optimized a increases with PMsg (Table III), which indicates that some incondensable compounds (ICs, do not contribute SOA formation) could in some way produce aerosol-forming products in the presence of (NH4)2S04. We propose that some ICs can participate in the heterogeneous reactions occurring at the surface of (NH4)2S04, resulting in three types of oligomerization processes as shown in Table IV. The comprehensive results of these three types of reactions increase the amount of CCs produced, and therefore makes a higher.

The partition coefficient Kom was relatively unchanged through these four series of experiments, which can also be explained by the proposed ICs oligomerization. Both type 2 and type 3 oligomerization processes could produce a larger CC molecule, which lowers the volatility and increases the partition coefficient. However, type 1 process may generate CCs with smaller molecular weight and low condensability. The integrated results of these three different oligomerization processes can give a similar overall condensability (Kom) of CCs as when no acid-catalyzed oligomerization can take place.

Was this article helpful?

0 0
Renewable Energy Eco Friendly

Renewable Energy Eco Friendly

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable.

Get My Free Ebook


Post a comment