Dominant role of NO2 oxidation in SO2 conversion to sulfate revealed by synchronous measurements of gas and particle sulfur isotopes in haze episodes

The oxidation pathways of SO ₂ conversion to sulfate remain controversial. Sulfur isotope (δ ³⁴ S) has been used to trace SO ₄ ²⁻ formation pathways based on sulfur isotope fractionation. Accurately assessing the isotope fractionation is crucial for quantifying the oxidation pathways of SO ₄ ²⁻ formation. However, previous studies have used particle δ ³⁴ S (δ ³⁴ SO ₄ ²⁻ ) to estimate the isotope fractionation (α ³⁴ S _g→p -estimated), leading to significant uncertainties in SO ₄ ²⁻ formation. This study synchronously measured δ ³⁴ S values of gas-phase SO ₂ and particle-phase SO ₄ ²⁻ to uncover isotope fractionation (α ³⁴ S _g→p ) in SO ₄ ²⁻ formation during haze episodes. Results found that the α ³⁴ S _g→p values (-3.7 ~ + 9.9‰) obtained by gas-to-particle δ ³⁴ S showed a significant difference with α ³⁴ S _g→p -estimated values(-6.4 ~ + 1.4‰) obtained by δ ³⁴ SO ₄ ²⁻ , implying different results for SO ₄ ²⁻ formation using the two methods. Among them, α ³⁴ S _g→p results indicated the prominent contribution of NO ₂ oxidation(48–56%), followed by TMI-catalyzed O ₂ (transition-metal ions, 26–40%). While α ³⁴ S _g→p -estimated (-6.4 ~ + 1.4‰) suggested the dominant role of TMI-catalyzed O ₂ (54–80%). Compared to α ³⁴ S _g→p -estimated, α ³⁴ S _g→p results show more reasonable response to SO ₄ ²⁻ formation and consistent trends with oxidant concentrations. α ³⁴ S _g→p -estimated analysis overestimated the TMI-catalyzed O ₂ pathway contribution (38–47%) to SO ₄ ²⁻ formation. This is the first study to employ gas-to-particle δ ³⁴ S to demonstrate the dominant role of NO ₂ oxidation in SO ₄ ²⁻ formation, implying the importance of synchronous measurements gas-to-particle δ ³⁴ S. Furthermore, a functional relationship between D-values (difference values of α ³⁴ S _g→p and α ³⁴ S _g→p -estimated) and impact factors was established, simulating actual α ³⁴ S _g→p in the absence of gas-phase δ ³⁴ S. This approach offers a possible correction for α ³⁴ S _g→p -estimated values, providing new insight into using single δ ³⁴ SO ₄ ²⁻ for the analysis of SO ₄ ²⁻ formation.