Atmospheric Degradation of Cyclic Volatile Methyl Siloxanes: Radical Chemistry and Oxidation Products

Abstract

Cyclic volatile methyl siloxanes (cVMS) are anthropogenic chemicals that have come under scrutiny due to their widespread use and environmental persistence. Significant data on environmental concentrations and persistence of these chemicals exists, but their oxidation mechanism is poorly understood, preventing a comprehensive understanding of the environmental fate and impact of cVMS. We performed experiments in an environmental chamber to characterize the first-generation oxidation products of hexamethylcyclotrisiloxane (D3), octamethylcyclotetrasiloxane (D4), and decamethylcyclopentasiloxane (D5) under different peroxy radical fates (unimolecular reaction or bimolecular reaction with either NO or HO 2 ) that approximate a range of atmospheric compositions. While the identity of the oxidation products from D3 changed as a function of the peroxy radical fate, the identity and yield of D4 and D5 oxidation products remained largely constant. We compare our results against the output from a kinetic model of cVMS oxidation chemistry. The reaction mechanism used in the model is developed using a combination of previously proposed cVMS oxidation reactions and standard atmospheric oxidation radical chemistry. We find that the model is unable to reproduce our measurements, particularly in the case of D4 and D5. The products that are poorly represented in the model help to identify possible branching points in the mechanism, which require further investigation. Additionally, we estimated the physical properties of the cVMS oxidation products using structure−activity relationships and found that they should not be significantly partitioned to organic or aqueous aerosol. The results suggest that cVMS first-generation oxidation products are also long-lived in the atmosphere and that environmental monitoring of these compounds is necessary to understand the environmental chemistry and loading of cVMS.

Comparison of the Yield and Chemical Composition of Secondary Organic Aerosol Generated from the OH and Cl Oxidation of Decamethylcyclopentasiloxane

“…Corresponding average Van Krevelen slopes were −0.27 (OH-SOA) and −0.11 (humid Cl-SOA). This overall trend was consistent with expected CH 3 → OH and/or CH 3 → CH 2 OH reactions at lower OH exp or Cl exp , 10,50 followed by the formation of carbonyl, carboxylic acid, and/or formate ester functional groups at higher oxidant exposures. On the other hand, H/C of dry Cl-SOA decreased continuously from 2.5 to 1.4 with increasing O/C, with an average Van Krevelen slope ranging from −1.03 (OFR254/313-iC 2 Cl 2 O 2 ) to −1.07 (OFR369-iCl 2 ).…”

“…These formulas can also explain our signals at those nominal m / z values but require significantly more complex fragmentation in the AMS than our proposed assignments. Other signals that were present at large abundances in both OH-SOA and Cl-SOA spectra include m / z 15 (CH 3 + ), which corresponds to methyl groups that were bonded to Si atoms, m / z 29 (CHO + ), a marker for alcohols, and m / z 44 (CO 2 + ), a marker for organic acids and potentially also formate esters. , …”

“…The ions at m / z 363 (C 5 H 18 O 9 Si 5 H + , detected previously in D 5 SOA) and m / z 365 (C 4 H 16 O 10 Si 5 H + ) were the largest signals in the OH-SOA spectrum. We hypothesize that C 5 H 18 O 9 Si 5 and C 4 H 16 O 10 Si 5 are carbonyl siloxanols generated following three and four CH 3 → OH fragmentation reactions, respectively, , plus the replacement of two adjacent methyl groups with a carbonyl [SiO 2 (CH 3 ) 2 → Si­(O)­O 2 ]. Similarly, carbonyl siloxanols detected at m / z 361 (C 6 H 20 O 8 Si 5 H + ) and m / z 367 (C 3 H 14 O 11 Si 5 H + ) were likely generated by two and five CH 3 → OH fragmentation reactions plus one SiO 2 (CH 3 ) 2 → Si­(O)­O 2 fragmentation reaction.…”

“…Both D 5 OH-SOA and Cl-SOA spectra were dominated by C x H y O z Si n + ion groups, which contributed 61 and 69% of the total OH-SOA and Cl-SOA signal, respectively (panels e and f of Figure 3 47 Han et al 44 ), which corresponds to methyl groups that were bonded to Si atoms, m/z 29 (CHO + ), a marker for alcohols, 38 and m/z 44 (CO 2 + ), a marker for organic acids 48 and potentially also formate esters. 49,50 Figure S2 of the Supporting Information shows AMS C x H y O z Si n + signals that were detected in D 5 OH-SOA and Cl-SOA spectra above m/z 350. In general, AMS signals in this range of higher m/z are less common as a result of high-energy electron impact ionization of 70 eV and flash vaporization at T = 600 °C that result in thermal decomposition and fragmentation of the analyte molecules; their presence here underscores the stability of siloxanes that undergo both processes.…”

“…Here, we assumed RO 2 reacted with OH, HO 2 , Cl, ClO, RO 2 , and/or underwent autoxidation via sequential isomerization and O 2 addition. Alton and Browne50 proposed a humidity-dependent mechanism for the gas-phase OH/Cl oxidation of D 3 in which RO that was generated following RO 2 isomerization reacted with H 2 O, generating a silanol and OH as co-products. To evaluate the potential effect of this reaction on the RO 2 fate, we set the RO + H 2 O rate coefficient to 4 × 10 −12 cm 3 s −1 to make this reaction proceed at the same rate as RO 2 isomerization/autoxidation reactions and then calculated F ROd 2 in D 5 /Cl experiments with and without this reaction enabled in the model.…”