Phosphorus removal by sands for use as media in subsurface flow constructed reed beds

Sorption of P to the bed sand medium is a major removal mechanism for P in subsurface flow constructed reed beds. Selecting a sand medium with a high P-sorption capacity is therefore important to obtain a sustained P-removal. The objective of this study was to evaluate the P-removal capacities of 13 Danish sands and to relate the removal to their physico-chemical characteristics. The P-removal properties were evaluated in short-term isotherm batch-experiments as well as in 12-week percolation experiments mimicing the P-loading conditions in constructed reed bed systems. The P-removal properties of sands of different geographical origin varied considerably and the suitability of the sands for use as media in constructed reed beds thus differs. The P-removal capacity of some sands would be used up after only a few months in full-scale systems, whereas that of others would persist for a much longer time. The most important characteristic of the sands determining their P-removal capacity was their Ca-content. A high Ca content favours precipitation with P as sparingly soluble calcium phosphates particularly at the slightly alkaline conditions typical of domestic sewage. In situations where the wastewater to be treated is more acid, the contents of Fe and Al may be more important as the precipitation reactions with these ions are favoured at lower pH levels. The maximum P-sorption capacities estimated using the Langmuir-isotherm plots did not correspond to or correlate with the actual amount of P removed in the percolation columns. Hence, the Langmuir-isotherm does not estimate the P-removal capacities of sands. It is suggested that a suitable quick method of screening a selection of potential media for P-removal potential is to perform simple removal isotherm studies using water with a similar chemical composition as the wastewater to be treated. This method will not provide a direct estimate of the P-removal capacity that can be obtained in full-scale systems, but it is a means of comparing the relative performance of potential media. (C) 2001 Elsevier Science Ltd. All rights reserved.