Triplet State Formation of Chromophoric Dissolved Organic Matter in Atmospheric Aerosols: Characteristics and Implications

a School of Environmental Science and Engineering, Shaanxi University of Science and Technology, Xi’an 710021, China b Environment Research Institute, Shandong University, Qingdao, 266237, China c School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, China d Department of Building Science, School of Architecture, Tsinghua University, Beijing, 100084, China e College of Resource and Environment, Anhui Science and Technology University, 233100 Anhui, China f State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China g College of Resources and Environment, University of Chinese Academy of Sciences, 100190, Beijing, China h Department of Environmental Science and Engineering, Xi’an Jiaotong University, Xi’an 710049, China

Based on the off-line analysis mode of the TOC analyzer, each sample was 165 continuously analyzed 3 times, and the average value after subtracting the background 166 value was the final detection value. The relative standard deviation of the WSOC 167 content was 1.5%. 168

Optical Absorption and EEM Fluorescence Spectra. 169
Absorption and EEMs of the extracts were obtained using a fluorescence 170 spectrophotometer (Aqualog, Horiba Science, America). Detection conditions: the 171 excitation wavelength range is 200-600 nm and the emission wavelength range is 172 250-800 nm. The wavelength interval is 5 nm, and the integration time is 0.5 s. The 173 background samples are also analyzed under the same detection conditions and 174 deducted from the sample signal. The WSOC concentration in the sample was diluted 175 to within 10 ppm so that the absorbance at 250 nm was less than 0.5. The inner filter 176 effect has little influence on the results because the sample was fully diluted. 177 Correction of the inner filter effect for the EEMs is also performed.  Compared with the previous study, the concentration of TMP used in the paper is 203 higher; therefore, we compared high-concentration TMP with low-concentration TMP. 204 The results show that under our reaction conditions, the high-concentration TMP may 205 have a relatively low background and a higher reaction rate constant (the results are 206 shown in Figure S3 of SI). 207 The sulfate in aerosols may produce sulfate free radicals under illumination, which 208 can possibly consume TMP. Simulated TMP consumption by a sulfate ion solution 209 was also examined in this study. Three parallel groups of background and control 210 experiments are compared. We studied the effect of salts on the formation of triplet 211 states. As shown in Figure S4, in the reaction system with or without (NH4)2SO4, no 212 significant difference exists in the decay rate of TMP. To further study the effect of 213 salts on the formation system of triplet states, we used solid-phase extraction to 214 separate high-polar substance salts and low-polar HULISs (Chen, et al., 2016a). We 215 determined the effects of salts, HULISs, and a salt and HULIS mixture on TMP 216 attenuation. As shown in Figure S5, we found that when the salts were mixed with 217 https://doi.org/10.5194/acp-2019-1032 Preprint. Discussion started: 24 January 2020 c Author(s) 2020. CC BY 4.0 License.

9/25
low-polar substances, no significant effect on TMP attenuation was identified in the 218 low-polar reaction system. (1) The driving effects of the triplet state on 1 O2 were studied. 229 2,2,6,6-tetramethyl-piperidine (TEMP, cTEMP = 0.25 M) was used as a scavenger for 230 1 O2, and SA (cSA = 4×10 2 µM) was added into the reaction system as a triplet state 231 quencher. After 60 min of illumination (the illumination device is shown in Figure S1  formed in the reaction system is constant and the formation rate of 3 CDOM* is the 275 same as its quenching rate. In this case, the quenching mechanism of 3 CDOM* 276 conforms to the paths (2)-(3) and (5)-(7) described in Scheme 1. In the reaction 277 process, 3 CDOM* may not be consumed, but 3 CDOM* mainly promotes energy 278 transfer, such as converting O2 to 1 O2. 279 As shown in Figure 1C Therefore, the ability of the ambient PM to form 3 CDOM* is greater in winter. In 341 particular, the highest kTMP value was 0.046 min -1 in the winter samples, which was 342 similar to coal combustion and wood burning samples. This result is consistent with 343 the fact that coal combustion is an important source of ambient PM in winter in Xi'an. 344

376
The 3 CDOM* formation ability depends on the CDOM type. In this study, five 377 types of CDOM were identified through the PARAFAC model, as shown in Figure  378 4C and E, and Figure 4A

394
The ability of different CDOM to form 3 CDOM* is different. The structure-activity 395 relationship between the CDOM type and 3 CDOM* formation rate was established by 396 the improved PARAFAC model in equation (2). Figure 4B illustrates the relative 397 contributions of the different types of CDOM to the total formation rate of 3 CDOM*. According to the contribution of 3 CDOM* to TMP consumption, the results of this 415 study indicated that the 3 CDOM* formed by HULISs and phenol-like substances are 416 not able to transfer electrons. The quenching mechanism is mainly energy transfer, 417 which means that this 3 CDOM* has more significant effect of driving ROS. However, 418 typical N-containing chromophores such as amino acids may include both of the 419 above quenching modes. Figure 3D illustrates that the contribution of C1 to 420 fluorescence in the POA and ambient PM samples is positively correlated with kTMP, 421 indicating that C1 is the most important high-energy 3 CDOM* precursor in aerosols.

19/25
In this study, the structure-activity relationship between the CDOM type and 465 3 CDOM* formation rate was established by the EEM-PARAFAC approach. We 466 identify that the C1 and C3 chromophores, which may be attributed to N-containing 467 substances, significantly contribute to 3 CDOM* formation, although C1 and C3 468 contribute little to the total fluorescence intensity. The results showed that C1 and C3 469 chromophores are the main precursors for the formation of 3 CDOM* in aerosols. In 470 contrast, HULIS and phenol-like chromophores do not contribute significantly to 471 TMP attenuation. However, the above do not mean that these substances do not have 472 the ability to form 3 CDOM*. In this case, as shown in Scheme 1, 3 CDOM* through 473 self-quenching and energy transfer does not consume TMP, and low-energy 3 CDOM* 474 cannot react with TMP. 475

Data availability 476
The PM 2.5 data used in this paper are from http://www.cnemc.cn (China National 477 Environmental Monitoring Center). 478

Supporting information 479
Additional details, including Tables S1−S3, Figures S1−S10, calculation of the 480 formation rate of 3 CDOM* and the consumption rate of TMP due to 3 CDOM* 481 formation in aerosols under solar illumination, are contained in the SI. 482