The air quality in Chinese megacities has been improved as indicated by large decreases in fine particulate matter (PM2.5) due to remarkable decreases in key precursors (e.g., SO2, NOx) after the implementation of strict mitigation strategies. However, nitrate concentrations in PM2.5 (p-NO3-) have not decreased and mass fractions of p-NO3- in PM2.5 have increased, especially during wintertime haze events. Discerning chemical mechanisms leading to nitrate growth during haze events is critical to implement effective mitigation policies. Chemical transport models incorporating oxygen isotope anomaly of nitrate (Δ17O(NO3-)) have been widely used to investigate nitrate formation mechanisms, showing general consensus on the modelled and observed Δ17O(NO3-). However, under Beijing haze days, the same model tends to underestimate observed Δ17O(NO3-). Here we compiled reported Δ17O(NO3-) data in Beijing haze along with relevant observational parameters (e.g., OH total reactivity, peroxyl radical concentrations), tested assumptions on Δ17O of key precursors (e.g., OH and NO2), re-calculated Δ17O(NO3-) and compared with observations. Our results indicate that considering heterogeneous N2O5 reactions on Cl–containing aerosols with a ClNO2 yield of ~ 0.75 can explain the observed high Δ17O(NO3-). According to the Δ17O(NO3-) data, this heterogeneous N2O5 + Cl- chemistry can explain ~ 60% of nighttime nitrate production and makes daytime and nocturnal pathways equally important in winter Beijing haze. Meanwhile, the high yield of ClNO2 means that on the following day the subsequent photolysis of ClNO2 would enhance atmospheric oxidation capacity and promote haze pollution, highlighting the critical role of reactive chlorine chemistry in air pollution/chemistry in inland cities