The impact of iron-dissolved organic matter colloids on the fate and bioavailability of trace metals and phosphorus in surface water.
Phosphorus (P) is a limiting nutrient in many aquatic ecosystems. In agriculture, P is present in fertilizer and in animal manure. Widespread and intensive use of fertilizer and animal manure cause diffuse P emissions to the environment, leading to eutrophication of water bodies and impaired water quality. The fate of P in natural waters is strongly linked to that of iron (Fe) bearing particles, to which P can bind. Streams contain heterogeneous associations between iron and natural organic matter (NOM) which range in size from a few nanometers to more than 100 micrometer. The iron in such particles can be present under a variety of forms, ranging from small mononuclear Fe(III)–NOM complexes to large mineral Fe(III) oxyhydroxide particles with surface adsorbed NOM. Phosphorus, in the form of the phosphate anion, can bind efficiently to the surface of Fe oxyhydroxide particles. It is poorly understood how P binding to Fe-rich particles in streams is related to their size and structure, and how, in turn, the size and structure of the particles is related to their formation and to the chemistry of the streams which carry them. In addition, it is unclear how binding to such particles affects fate and environmental effects of P. The objectives of this study were 1) to determine the structure, size, and composition of natural Fe-rich particles in streams; 2) to measure P binding by such particles, and to relate this to the properties of the particles and to the properties of the streams which carry them; and 3) to determine how binding to Fe-rich particles affects the fate and bioavailability of P in natural waters.
The structure, size, and composition of Fe-rich particles in streams was measured by field-flow fractionation and membrane filtration. Soft water streams draining acidic peatland contain small (<40 nm) associations of nanoparticulate Fe oxyhydroxides and humic substances. These particles bind P, but the larger particles contain more P than the smaller ones. This is likely due to competition by humic substances for binding on the surface of Fe oxyhydroxides: large particles are less covered by humic substances and may, therefore, bind P more strongly. In moderately hard to very hard water streams, the Fe oxyhydroxide particles are larger (>40 nm) than those in the soft waters. The P in such streams is bound to ferrihydrite colloids (40–1200 nm) and to associations between Fe and clay minerals (50–150 nm). It is concluded that P may be bound to a variety of Fe oxyhydroxide containing colloids. The results support the view that primary oxyhydroxide nanoparticles form increasingly larger aggregates with increasing water hardness, increasing pH, and increasing Fe:NOM ratio.
The effect of colloidal Fe oxyhydroxides on P bioavailability was determined in a model system with the freshwater green alga Raphidocelis subcapitata. Synthetic iron-organic matter colloids reduce the P uptake flux by algal cells compared to colloid-free test media. However, the P uptake flux from colloid containing solutions equals that from colloid-free ones if only the free orthophosphate concentrations are considered. This demonstrates that the colloidal P does not contribute to the P uptake flux, and hence that it is not readily bioavailable. The P added to post-synthesis ferrihydrite is bound less efficiently, and its bioavailability is higher, than if the P is present during ferrihydrite formation.
The formation and the fate of Fe-rich suspended particles were monitored in catchments which receive large Fe(II) inputs. Four Belgian lowland catchments fed by Fe-rich groundwater were sampled: the Kleine Nete catchment and three tributaries to the Demer. The groundwater contains, on average, 20 mg Fe/L and 0.4 mg P/L. As this groundwater surfaces, the soluble Fe(II) is oxidized to Fe(III) which readily forms particles (termed “authigenic”). The P is concomitantly removed from solution due to binding by the fresh Fe(III) minerals. The oxidation reaction proceeds as the groundwater surfaces and flows through the catchment into increasingly larger streams. The Fe(II) oxidation and the formation of authigenic particles is slower in winter than in summer, due to shorter travel times, lower pH, and lower temperature in winter. The authigenic particles are between 1 and 20 µm and consist of almost pure, poorly crystalline Fe oxyhydroxides similar to ferrihydrite. The mineralogy and composition of these particles change as they are transported into increasingly larger streams: the authigenic particles become larger in size due to aggregation, they become structurally more condense due to ageing reactions, and they are increasingly diluted by mixing with material from a different source. The removal of P from solution is much faster than that of Fe: it is already complete in the smallest headwater streams. The average P concentration in streams (42 µg/L) is one order of magnitude below that in groundwater (393 µg/L). The local environmental P limit for freshwater (140 µg/L) is between both values. Naturally occurring Fe in groundwater therefore alleviates the environmental risk of P in the receiving streams.
The formation of Fe-rich authigenic particles was monitored on the trajectory of draining groundwater: from the subsurface through Fe-rich sediments and into small drainage ditches of the Kleine Nete catchment. In the sediment, reductive dissolution of P-bearing Fe oxyhydroxides causes solubilization of Fe and P. Conversely, in the ditchwater, oxidative precipitation causes sequestration of Fe and P. Because the Fe is present in large excess, the removal of P is faster than that of Fe: ferric phosphate or P-saturated Fe oxyhydroxides are initially formed until P is nearly depleted. This yields a natural and highly efficient sink for P. In Fe-rich systems, the fate of P at the sediment-water interface is determined by reduction and oxidation of Fe.
Taken together, this study shows that P in streams may be bound to iron oxyhydroxide containing particles. These particles range in size between 1 nm and 100 µm. Small particles (1–40 nm) dominate in soft waters with low Fe:NOM ratio, whereas large particles (>40 nm) dominate in harder waters and at higher Fe:NOM ratio. In waters with low molar Fe:P ratio (“low Fe–high P waters”), the particles are saturated with P, and they contain Fe:P ratios around two. Conversely, in waters with high molar Fe:P ratio (“high Fe–low P waters”), most P is bound by Fe-rich particles. The strength of P binding also depends on the mechanism of particle formation: it is highest if the P is present during Fe(II) oxidation, due to coprecipitation with Fe oxyhydroxides. In contrast, it is lower if the P binds to existing Fe oxyhydroxides. The binding of P by Fe-rich particles reduces its bioavailability. In summary, the fate and effect of P in freshwater is intimately linked to the biogeochemical cycle of Fe and NOM.