The dual isotopic signatures of particulate nitrate (hereafter as delta N-15-NO3- and delta O-18-NO3-) have been extensively used to imprint the source and chemical transformation of atmospheric NOx (NOx = NO + NO2). For instance, the delta O-18-NO3- elevated proportionally when NOx converted by O-3. In the present study, daily PM2.5 samples (n = 91) were collected in winter Beijing (December to February in 2017-2018) and a two-endmember linear isotopic mixing model was used to model the endmember delta O-18 values of NO3- oxidized by O-3 (termed as delta O-18(noct)) and the contribution of the different pathways (i.e., daytime and nocturnal oxidation pathways). During the campaign, the NO3- concentrations in PM2.5 varied from 0.3 to 46.3 mu g m(-3) (8.1 +/- 9.8 mu g m(-3)), delta N-15-NO3- from +1.0 parts per thousand to +19.6 parts per thousand (12.5 +/- 3.6 parts per thousand) and delta O-18-NO3- from +50.7 parts per thousand to +103.5 parts per thousand (74.9 +/- 13.3 parts per thousand). The Keeling plots indicated that the delta O-18(noct) endmember value was within the ranges based on theoretical approaches. The contribution of the nocturnal pathway to NO3- in PM2.5 ranged from 8.5 +/- 3.2% in background days to approximately 100% in extremely polluted days, with a mean of 52.0 +/- 25.5%. The determined delta N-15-NO3- values (12.5 +/- 3.6 parts per thousand) were comparable with previous studies conducted in recent winter Beijing (2013-2017, i.e., average values of 11.9 parts per thousand-13.8 parts per thousand). Due to the optimization of energy structure in Beijing, we inferred that the dominant NOx source in recent wintertime was the vehicular exhaust. The Bayesian mixing model also confirmed that the contribution of vehicle exhaust/biomass burning sources to the NO3- was up to 70%. This study may further improve the understanding of NOx emission source and atmospheric processes in urban environments.