Modeling and experimental approaches for determining fluoride diffusion kinetics in bone char sorbent and prediction of packed-bed groundwater defluoridator performance.

Fluoride (F) in groundwater (GW) in excess of 1.5 mg/L is a globally distributed problem impacting the health of hundreds of millions of people, many of whom cannot access centralized treatment infrastructure. Animal (e.g., cow) bone char has received emerging interest as a low-cost F sorbent for use in decentralized household and community water treatment. Pilot column tests using full-sized granular bone char particles can be used to assess treatment performance of fixed-bed contactors, but are costly, time consuming, and require large amounts of test water. Rapid small-scale column tests (RSSCTs) can be used to simulate F uptake in bone char contactors if the relationship between F intraparticle diffusion kinetics and bone char particle size is known. Two common approaches to the RSSCT assume either constant (CD) or linear proportional (PD) sorbate diffusivity as a function of sorbent particle size. This study used experimentally determined pseudo-equilibrium and kinetic F sorption data in model groundwater as inputs to the homogeneous surface diffusion model (HSDM) to determine F intraparticle diffusion coefficients for different-sized bone char particles, and to fit RSSCT and pilot column breakthrough data to evaluate CD and PD approaches. Results of this study, corroborated by incorporation of additional literature data, indicate approximately linearly proportional diffusivity of F as a function of bone char particle size. Congruently, the PD-RSSCT approach provided a superior simulation of pilot column F breakthrough compared to the CD-RSSCT. PD-RSSCT breakthrough data closely matched pilot breakthrough on a scaled service time basis up to around 500 bed volumes, corresponding to a relative F breakthrough of about 40%, and provided a slightly conservative indicator of F removal thereafter. The PD-RSSCT was compared with a hybrid modeling and empirical workflow using the HSDM with experimentally determined pseudo-equilibrium and kinetic parameter inputs as time-and-cost-saving approaches to evaluating full-sized groundwater treatment system performance. This comparison and a sensitivity analysis of HSDM input parameters used in the hybrid workflow indicated that greater precision can be obtained using the PD-RSSCT.