Inorganic chlorine budget and gas-particle partitioning in the winter lower troposphere over the northeast United States

Abstract

Reactions of halogens occurring in fine mode aqueous aerosols are regulated by aerosol pH and liquid water content which affect ozone concentrations, tropospheric oxidant concentrations, sulfate and nitrate aerosol mass loadings, and hydrocarbon concentrations. However, the rates of heterogeneous and aqueous halogen reactions under atmospheric conditions, and, consequently the inorganic tropospheric halogen budget and its overall impact on air quality, remain highly uncertain. In this work, measurements from the 2015 Wintertime Investigation of Transportation, Emissions, & Reactivity (WINTER) aircraft campaign are used to show that the inorganic chlorine budget is dominated by HCl (g) and total particulate chloride, accounting for greater than 85% of the total chlorine budget within the boundary layer. The total amount and the range of all chlorine compounds sampled are elevated in marine environments with the total mass of inorganic chlorine compounds found over marine regions being 1014 pptv and 609 pptv over continental regions. The median observed mass fraction of chlorine in the fine mode particle phase (pCl-) is 0.054 (Q25% = 0.038, Q75% = 0.084) for the whole campaign, with large variability about this average. ClNO2 was observed to peak at night containing ~15% of the overall chlorine budget (>1000 pptv). HOCl was observed to peak during the day and with a median of 100.8 pptv over the ocean and 21.3 pptv over land. WINTER observations of gas and particle composition were used as inputs to the offline thermodynamic equilibrium model, ISORROPIA II, to compare observed and modeled chlorine gas-to-particle partitioning. Observations show 0-20% of available, submicron non-refractory chlorine partitions into the particle (0 - 0.2 μg m-3amb) at sufficiently high relative humidities, low temperatures, and pH above 0.8. The thermodynamic model significantly over-predicted particulate chloride by approximately a factor of two and under-predicted gas phase HCl by the same. Errors in the modeled HCl gas-particle partitioning may be caused by the presence of unmeasured refractory sodium chloride. The chlorine partitioning can be brought into agreement with the observations without significantly perturbing the nitrate partitioning by artificially increasing fine mode sodium chloride particle mass in the model to simulate an unmeasured sea salt fraction by <0.1μg m-3 in the median. The disagreement could also be caused by an over prediction in the effective equilibrium constant for HCl. An implied equilibrium constant function is derived from the WINTER data set and ISORROPIA II estimates of pH, LWC, and activity (Keq= 4.5533 x 106 exp[3012 (1/T- 1/T0)] mol2 kg-2soln atm-1 ) which has a distinctly lower temperature dependency than the ISORROPIA II default value and is expected to be a lower limit of the equilibrium constant. Ultimately this work highlights the sensitivity of chlorine partitioning predictions by ISORROPIA II to minor changes chemical and environmental inputs. Such sensitivity will ultimately propagate to predictions of heterogeneous halogen chemistry and its effects on air quality and climate.