The speciation, oxidation states, and relative abundance of iron (Fe) phases in PM2.5 samples from two locations in urban Los Angeles were investigated using a combination of bulk and spatially resolved, element-specific spectroscopy and microscopy methods. Synchrotron X-ray absorption spectroscopy (XAS) of bulk samples in situ (i.e., without extraction or digestion) was used to quantify the relative fractions of major Fe phases, which were corroborated by spatially resolved spectro-microscopy measurements. Ferrihydrite (amorphous Fe(III)-hydroxide) comprised the largest Fe fraction (34-52%), with hematite (α-Fe2O3; 13-23%) and magnetite (Fe3O4; 10-24%) identified as major crystalline oxide components. An Fe-bearing phyllosilicate fraction (16-23%) was fit best with a reference spectrum of a natural illite/smectite mineral, and metallic Fe(0) was a relatively small (2-6%) but easily identified component. Sizes, morphologies, oxidation state, and trace element compositions of Fe-bearing PM from electron microscopy, electron energy loss spectroscopy (EELS), and scanning transmission X-ray microscopy (STXM) revealed variable and heterogeneous mixtures of Fe species and phases, often associated with carbonaceous material with evidence of surface oxidation. Ferrihydrite (or related Fe(III) hydroxide phases) was ubiquitous in PM samples. It forms as an oxidation or surface alteration product of crystalline Fe phases, and also occurs as coatings or nanoparticles dispersed with other phases as a result of environmental dissolution and re-precipitation reactions. The prevalence of ferrihydrite (and adsorbed Fe(III) has likely been underestimated in studies of ambient PM because it is non-crystalline, non-magnetic, more soluble than crystalline phases, and found in complex mixtures. Review of potential sources of different particle types suggests that the majority of Fe-bearing PM from these urban sites originates from anthropogenic activities, primarily abrasion products from vehicle braking systems and engine emissions from combustion and/or wear. These variable mixtures have a high probability for electron transfer reactions between Fe, redox-active metals such as copper, and reactive carbon species such as quinones. Our findings suggest the need to assess biological responses of specific Fe-bearing phases both individually and in combination to unravel mechanisms of adverse health effects of particulate Fe.
Keywords: PM2.5; X-ray absorption spectroscopy; electron energy loss spectroscopy; exhaust emissions; iron speciation; non-exhaust emissions.