Modeled Global Impacts of Chlorine Oxidation and Temperature Dependence on the Atmospheric Lifetime and Concentrations of Volatile Methyl Siloxanes

TL;DR

This study models how temperature and chlorine reactions affect the lifespan and spread of siloxane chemicals in the atmosphere, finding shorter lifetimes and transport than previously thought.

Contribution

The study introduces chlorine oxidation and temperature-dependent reaction rates into global modeling of siloxanes for the first time.

Findings

VMS lifetimes were predicted to range between 2.7 and 6.7 days.

D5 concentrations in urban areas reached up to 160 ng m–3.

Chlorine oxidation significantly reduced LRET compared to previous models.

Abstract

Volatile methyl siloxanes
(VMS) are high production volume
chemicals
found in a wide range of consumer items such as personal care products.
VMS have attracted scrutiny due to long-range environmental transport
(LRET) concerns. However, their emissions, lifetimes, and concentrations
remain uncertain, in part because of limitations in previous atmospheric
modeling. Herein, we describe the global modeling of siloxanes: D4
(octamethylcyclotetrasiloxane), D5 (decamethylcyclopentasiloxane),
and D6 (dodecamethylcyclohexasiloxane) in the Community Earth System
Model (CESM2-SLH) using an updated chemical mechanism that includes
chlorine radical oxidation and temperature-dependent reaction rates.
With these previously unconsidered factors, we predicted VMS lifetimes
ranging between 2.7 and 6.7 days and annual average D5 near-surface
concentrations as high as 4.5 ng m–3 in the remote
Arctic and 160 ng m–3 in urban areas. These lifetimes
and the degree of LRET were significantly lower than previously reported.
OH remained the largest loss pathway, although it decreased at low
temperatures relative to previous modeling. The temperature-induced
lifetime increase was outweighed by the previously unconsidered chlorine
oxidation channel. Model results and previous measurements agreed
relatively well but exhibited negative normalized biases (−0.55
to −0.88), particularly in urban areas not well-resolved at
global model resolution, and for passive sampler measurements.

Figures