While constructed wetlands treat many pollutants well, they do not always attain high levels of phosphorus (P) removal, and are particularly poor for long-term P removal. Their performance for this pollutant, can, however, be improved by using suitable substrates to provide some physico-chemical removal capacity. Alum sludge (AS), a by-product of potable water treatment is produced in vast quantities world-wide, and increasingly stringent licensing regulations and costs associated with disposal, have led to investigations into beneficial reuse options.
This thesis describes a detailed study performed to determine whether waste alum sludge can be used as a substrate in constructed wetlands to optimise P removal. Residuals from two water treatment plants (WTPs) in Brisbane, Australia (North Pine (NP) and Mount Crosby (MC) WTPs) were selected on the basis that they were produced and dried by conventional processes, which are the most commonly used methods in Australia and around the world; and that they were comparable to sludges reported in the literature.
Laboratory-scale wetland mesocosm experiments operated for 7 months formed the foundation of the study. Specifically, these studies were used to determine the growth and elemental uptake characteristics of a common wetland macrophyte Phragmites australis when grown in substrates of 100% waste alum sludge, and to assess the treatment performance of these systems for all pollutant species, including the possibility of elemental leachate from the substrate under wetland conditions. Knowledge about the kinetics, capacity and mechanisms of P removal by AS under various conditions is crucial for wetland design. Batch phosphorus binding studies together with sequential extraction trials of phosphate bound to alum sludge were therefore used to explain the preferential binding of phosphorus by AS over time.
Batch studies revealed that AS has a very high capacity for P which varies depending upon initial P concentration (Co). Total capacities for P ranged from 11 mg P g-1 for oven dried AS incubated with an applied Co of 51 mgP L-1, to 113mgP g-1 (oven dry equivalent) for air dry AS from a 10 g P L-1 solution.
To date, studies designed to determine mechanisms of P retention by water treatment residuals have proven inconclusive, largely due to the complex mineral composition of AS. Batch adsorption studies performed here showed that P removal by AS was most likely controlled by amorphous Al-hydroxides and Fe-oxides. Sequential extraction experiments, which have not previously been used in association with AS, supported these findings, by confirming that the principal fractions responsible for removal of solution P were predominantly surface AI , and to a lesser extent, surface Fe elements. Further, sequential extraction experiments illustrated the likelihood that the mixture of elements in AS (including Al, Fe, and Ca elements) may result in sustainable P removal by preventing the re-release of bound P under wetland conditions.
Indeed, the superiority of AS substrates for P removal was particularly evident from the mesocosm experiments, which were loaded using a synthetic sewage solution with a mean COD and TKN concentrations of 164 mg L-1 COD, 40 mg L-1 TKN respectively and pH6.5. TP concentrations were increased throughout the experimental period, from 6.5 to 81 mg P L-1. Under these conditions, the mass based removal of TP was found to be >99%, 94% and 78% for MC AS, NP AS and sand, respectively. Mass balance calculations over the wetland mesocosms revealed that P uptake into plant tissues was minimal in the AS mesocosms, accounting for only 1.2% -2.9% of the influent P load, and thus confirming that AS substrates were responsible for binding 92% - 98% of influent P. This was in contrast to the sand controls, in which substrates were only responsible for storing 75% -87% of influent P, while plants removed around 10%.
Reduced levels of performance were observed for both COD and TN removal, compared to the sand controls. While COD removals of 93%, 80% and 74% (on a mass basis) for mesocosms of sand, NP and MC AS respectively were considered adequate, performance for nitrogen treatment was more compromised. Average removal rates of 85%, 74% and 59%, which resulted in mean effluent concentrations of 6, 10 and 16 mg N L-1 were found for sand, MC AS and NP AS mesocosms, respectively.
These studies showed that 100% AS was able to sustain vegetative growth of P. australis. demonstrating for the first time its possible use for such applications. The productivity of AS grown reeds was however inhibited by greater than 10% compared to those grown in sand and mass balance calculations over the systems revealed that the relative vigour of plant growth determined the overall success of N removal from the mesocosms. Effective ammonification and nitrification but incomplete denitrification was evident in all substrates.
Leachate studies showed that there was very little issue to be raised in relation to the environmental consequences of elemental leachate from AS under wetland conditions. Leachate concentrations of Al, Ca, Cu and Fe were not significantly higher than from the sand controls. Of the elements found to be released at significantly higher or lower concentrations than from the control systems (B, K, Mg, Mn, S and Zn), none were found at concentrations that exceeded local or international legislative guidelines for discharge to surface waters.
In summary, alum sludge shows great promise as a substrate for constructed wetlands, in that it will provide a high capacity, long-term sink for P in a wetland situation, with the potential to increase the life cycles of wetlands for P removal to around 20 years (based on conservative estimations). Analyses undertaken in this thesis show the decreased efficiency of N removal to be related to plant associated processes illustrating the importance of macrophytes in the functioning of wetlands. Because the growth of reeds in 100% AS is shown to be inhibited, concentration ranging tests of AS with alternative substrates are required to establish mixing rates that will achieve healthy macrophyte growth and adequate P removal. Alternatively, AS could be used as an isolated bed of filter material placed before the wetland outlet to reduce effluent P concentrations. This configuration would provide a simple retrofit to existing wetland systems that are achieving low P removal rates. A further advantage would be that this filter could easily be replaced after its capacity for P had been saturated.