Assessment of oxidative potential by hydrophilic and hydrophobic fractions of water-soluble PM2.5 and their mixture effects

Highlights: • Oxidative potential of water-soluble PM_2.5 is mainly from the hydrophilic fraction. • Cu is the dominant contributor of oxidative potential of water-soluble PM_2.5 • Mixture effects of the hydrophilic-hydrophobic fractions lead to suppression on OP._{• OH} Transition metals (TMs) (e.g. copper (Cu) and iron (Fe)) and certain organic compounds are known active constituents causing oxidative potential (OP) by inhaled ambient fine particulate matter (PM_2.5) in lung fluid. Humic-like substances (HULIS), isolated from atmospheric PM_2.5, are largely metal-free and contain mixtures of organics that are capable of complexing TMs. TMs and HULIS co-exist in the water-extractable part of PM_2.5. In this work, we used a solid phase extraction procedure to isolate the water-soluble TMs in the hydrophilic fraction (HPI) and HULIS in the hydrophobic fraction (HPO) and carried out this isolation procedure to a set of 32 real-world PM_2.5 samples collected in Beijing and Hong Kong, China. We quantified two OP endpoints, namely hydroxyl radical formation (denoted as OP_{• OH}) and ascorbic acid depletion (denoted as OP_AA), by the two fractions separately and combined, as well as by the bulk water-soluble aerosols. OP_{• OH} and OP_AA were well-correlated in both separate fractions and their combined mixtures or bulk water-soluble aerosols. OP by HPI far exceeded that by HPO. On a per unit PM_2.5 mass basis, the Hong Kong samples on average had a higher OP_AA and OP_{• OH} than the Beijing samples due to more water-soluble Cu. For HPI, Cu was a dominant OP_{• OH} and OP_AA contributor (>80%), although water-soluble Fe was present at a concentration approximately one order of magnitude higher. Suppression effects on OP_{• OH} were observed through comparing the OP of the bulk water-soluble aerosol with that of HPI. Our work reveals the importance of monitoring PM_2.5 chemical compositions (especially water-soluble redox active metals). Furthermore, we demonstrate the need to consider metal-organic interactions when evaluating the aggregate OP by PM_2.5 from individual components or apportioning OP by PM_2.5 to specific chemical components.