Leveraging DOM UV absorbance and fluorescence to accurately predict and monitor short-chain PFAS removal by fixed-bed carbon adsorbers

Abstract

Carbon adsorbent fouling by dissolved organic matter (DOM) inhibits the ability of the widely-used rapid small-scale column test (RSSCT) to accurately predict the removal of organic micropollutants (OMP) from water by full-scale carbon adsorbers. Here, the adsorption of 11 short-chain per-/poly-fluoroalkyl substances (PFAS) from groundwater, surface water, and wastewater was examined in pilot columns as well as RSSCTs using constant diffusivity (CD) and proportional diffusivity (PD) designs. Neither the CD- or PD-RSSCT accurately predicted pilot adsorber breakthrough of PFAS using standard diffusional mass transfer models. However, PFAS breakthrough relative to optical property (e.g., peak C, UV absorbance at 254 nm) breakthrough remained constant between pilot column, CD-RSSCT, and PD-RSSCT designs. This finding permitted accurate breakthrough predictions for the sum of PFAS and for 9 of the 11 PFAS on an individual basis in pilot columns using RSSCTs. Multiple linear regressions incorporating influent and treated water optical parameters enabled the modeling approach to be applied to water sources with heterogeneous DOM characteristics. It is hypothesized that this methodology was successful because (i) optical parameters adequately quantified the competitive nature of DOM and their adsorption behaved similar to OMP and (ii) competitive adsorption by low-molecular weight DOM was the predominant fouling mechanism. An OMP monitoring approach was developed for waters containing DOM with heterogenous characteristics that also relied on raw and treated water optical properties. UVA254 and fluorescence monitoring could therefore enable water treatment to remove PFAS in a variety of scenarios that face inhibitory cost and analytical limitations, such as decentralized and low-resource settings.