Reclaiming the health of British rivers Optimising conditions in low-cost systems for treating diffuse water pollution Carr 1, S., Heal 1, K.V., Lumsdon 2, D., and Vinten 2, A. 1 The University of Edinburgh, Edinburgh. 2 The Macaulay Institute, Aberdeen. 1 Research rationale This project focuses upon the design and implementation of a suite of filters comprising ‘iron’ ochre, suitable for removing phosphates from anthropogenically enriched surface waters. Ochre (amorphous Fe precipitates) is formed when treating mine-water which is contaminated with potentially toxic elements such as Fe, Al and Ca. At least 2.7x10 4 tonnes of ochre is generated p.a. in the UK with no current end-use. Ochre is largely comprised of Fe(OH) 3 and FeO.OH with smaller proportions of aluminium oxides and calcium carbonates. Due to this ochre has a high adsorption capacity for phosphorus (up to 26mg P g -1 ochre) and is therefore proposed as a suitable filter substrate for phosphorus adsorption. Selection of Ochre Across the UK there are over 177 abandoned metal and coal mines producing discharges affecting over 600km of river reach. To alleviate this, mine-water treatment plants have been created across the UK. These act to settle potentially toxic elements out of the water column prior to discharge to natural water course, leaving behind an iron rich sludge; ochre. The elemental and structural composition of ochre is highly site specific, depending upon factors such as mine chemistry and the conditions of settlement. Ochres were chosen for this project based upon; Phosphorus adsorption capacity and elemental composition Body of previous research Availability of the ochre For more details contact Stephen Carr at Photograph: Polkemmet, west Lothian mine-water treatment plant settlement pond Figure: Location of mine-water treatment plant ochres used for this project Ochre characterisation Previous research on ochre has shown its high capacity and ability to remove phosphorus from solution under laboratory conditions. Expansion to field trials has been less successful, with far lower rates of phosphorus adsorption. This may be for a range of hydrological and chemical reasons. A detailed chemical and physical examination is ongoing in order to characterise and develop a more through understanding of ochre. This can then be used to optimise filters of ochre to remove phosphorus from aquatic environments. Project Methodology The project has been divided into a series of experiments; laboratory, computer and field based which when completed will achieve the objectives of the research. pH equilibrium experiments pH equilibrium experiments have been conducted for each ochre over an 8 day period. Ochre was suspended (15mg l -1 ) in 30ml 0.1M NaNO 3 solution with the addition of 1ml of either HCl or NaOH to adjust the pH of the solution over the desired range (pH 4-10). The solution was then allowed to equilibrate. The sweep of pHs was achieved by conducting 9 trials with varying acid/base concentrations for each ochre. Conclusion A range of ochres from across the UK are being characterised chemically and physically. This will form the basis of a chemical model using the software ORCHESTRA. From this filter design will be optimised and implemented. Shaking experiment (with complex solutions) Column experiments Complex column experiments ORCHESTRA filter modelling Arsenic trial: Validate the model Filter design Field implementation Diffusion into a sphere BET surface area Acid digest Filter evaluation Pellet stability Density/ porosity experiments Oxalate- extractable Fe X-ray diffraction pH equilibrium Isotherm exp. Shaking experiment Flow chart of intended experimentation for the project Figure: pH equilibrium for Acomb unpelletised ochre. Legend denotes strength of acid or base (-) added. Data points are the mean of triplicate experiments with the error bars indicating 1 σ Figure: Final pH value after equilibration for Acomb pellets and Acomb unpelletised. Y- axis shows the strength of either acid (HCl) or base (NaOH) addition. Data points are the mean of triplicate experiments with the error bars indicating 1σ. 6