Phosphate removal from aqueous solution using ZVI/sand bed reactor: Behavior and mechanism

- Phosphate removal in an up-flow reactor filled with ZVI/sand support.
- Maximum phosphorus trapping capacity was 132 mgP/g Fe.
- Influence of preconditioning phase on ZVI/sand support efficiency.
- Homogeneous phosphate distribution along the column at saturation.
- Main ZVI byproduct was lepidocrocite.

This research reports on phosphate removal from aqueous solution using ZVI/sand packed columns. The influence of column preconditioning, consisting of ZVI pre-oxidation before feeding the columns with phosphate solution, revealed that a column aged for 1 day was more efficient than un-conditioned column, 5-days and 10-days preconditioned columns. The distribution of phosphate trapped inside the columns was evaluated by measuring phosphate concentration in the solids at different levels (P1, P2 and P3) along the depth of the columns. The distribution of phosphate inside the columns was determined for a time period up to 46 days, corresponding to column saturation. Results showed heterogeneous trapping along the column before saturation and homogeneous distribution upon saturation. The maximum cumulative trapped phosphate after column dismantling was determined before saturation (after 17 days running) at 130, 68 and 31 mgP/gFe at the inlet-P1, P1-P2 and P2-P3 layers, respectively, whereas the homogeneous distribution of phosphate upon saturation was determined at 132 mgP/gFe throughout the column. Solid supports were characterized using SEM, XRD and XPS. Lepidocrocite and maghemite/magnetite were the only iron oxidation products identified at the different layers inside the columns. XPS results confirmed the sorption of phosphate at the surface of ZVI and its oxidation products and highlighted the formation of an iron phosphate complex.

143