Knowledge of the molecular-level chemistry of brown carbon (BrC) is important in reducing the uncertainties in aerosol radiative forcing. Time-resolved ambient PM 2.5 samples were collected during a severe pollution episode in January 2017 over Xi'an, China for a comprehensive nontarget and full scanning of BrC molecules and their absorption properties using electrospray ionization Fourier transform-ion cyclotron resonance mass spectrometry combined with partial least squares regression analysis, which apportioned the overall ultraviolet absorption to individual molecules. The estimated absorption of CHNO and CHNOS molecules exhibited nighttime prevalence, whereas CHOS, CHNS, CHN, CHO, CHS, and CH molecules presented a dynamic trend. Carbon conjugation was positively correlated with estimated absorption by CHO and CHNO molecules, while exhibiting a mixed relationship with CHNOS. Higher nitrogen content was associated with enhanced light-absorption properties of BrC molecules, while higher oxygen and sulfur content appeared to be associated with photobleaching during secondary transformation.Plain Language Summary Individual molecular absorption has significant implications for ambient brown carbon analysis, but existing studies mostly focused on bulk absorption properties for the sample as a whole. A novel method is presented here to evaluate individual brown carbon absorption at the molecular level through high resolution mass spectrometry, and partial least squares regression analysis was applied to PM 2.5 samples collected during a heavy pollution episode in Xi'an, China. Nitrogen-containing compounds were found to be the strongest contributor to brown carbon absorption, especially during the nighttime, while processes involving oxidation and sulfur addition appeared to weaken the molecular absorption. This study bridges the gap between real atmospheric PM 2.5 absorption and knowledge from controlled laboratory studies, which support these findings.