Mechanisms controlling the recycling of nutrient silicon in freshwater sediments: An experimental study of the interactions between silicon, iron and phosphorus

Abstract

Dissolved silicon (DSi) is an essential nutrient for numerous terrestrial and aquatic organisms. In freshwater systems, including streams, rivers and lakes, an important class of siliceous algae are diatoms. Human activities, including land use changes, nitrogen (N) and phosphorus (P) enrichment, and hydrological alterations, have caused a decrease of DSi availability relative to N and P. In turn, these changes in macronutrient availability may contribute to shifts in phytoplankton communities that increase the likelihood of nuisance and harmful algal blooms. Internal loading of nutrient silicon (Si) from bottom sediments is one key process regulating the availability of DSi in the overlying water column. The magnitude and timing of internal DSi loading in freshwater bodies are controlled by biogeochemical reactions in sediments whose mechanisms and kinetics remain to be fully understood. In this thesis, I use controlled laboratory experiments to unravel the roles of different reaction pathways in the immobilization and release of DSi in freshwater sediments. Starting with initially very simple synthetic reaction systems, I progressively include additional components, specifically iron (Fe) and P, in order to mimic more realistic biogeochemical reaction networks, and ultimately, perform experiments with real freshwater sediments. After an introduction of the Si biogeochemical cycle, a review of the literature, and an outline of my thesis (Chapter 1), I present a study on the dissolution kinetics of amorphous silica (ASi) as a function of pH and salinity across the ranges typically found in natural freshwater (Chapter 2). The surface properties of ASi of various synthetic and natural sources are characterized with potentiometric titrations whose results are interpreted with the help of a constant capacitance model. Next the dissolution kinetics of ASi are measured in batch experiments, and the observed dissolution kinetics of ASi are fitted to a surface reaction model. The results confirm the previously reported non-linear relationship between the dissolution rate of ASi and the degree of undersaturation, implying that at least two dissolution rate constants are needed to describe the dissolution kinetics at high (typically, >0.4) and low (typically, <0.4) degree of undersaturation. In addition, the lack of correlation between the total measured electrical charge and dissolution rate constants provide a way to estimate the fraction of internal silanol groups in porous ASi materials. The quantitative relationships between the dissolution rate constant of ASi and environmental variables, including pH, degree of undersaturation, and salinity obtained in Chapter 2 contribute to the general framework for predicting dissolution rates of ASi in freshwater environments. The dissolution of ASi is the first step to recycling ASi back to bioavailable DSi in surface water. Based on the findings in Chapter 2, the effects of Fe(II) adsorption on the dissolution kinetics of ASi are assessed in Chapter 3. A series of batch reactor experiments with variable amounts of aqueous Fe(II) added to ASi suspensions are conducted under anoxic conditions. Experimental results show that the presence of Fe(II) under anoxic conditions retards the release of DSi, with the magnitude of the retardation dependent on the initial Fe(II) concentration. Trace amounts of Fe(II) slow down the release of DSi probably by forming bidentate surface complexes which block reactive sites on ASi, rather than through the formation of new ferrous iron silicate phases. A Langmuir adsorption model that incorporates two types of surface groups (silicate groups bonded to the silica lattice via two bridging oxygens, Q2 groups, and silicate groups bonded to the silica lattice via three bridging oxygens, Q3 groups) is used to describe the effect of Fe(II) on the dissolution kinetics of ASi. The modeling results suggest that Fe(II) preferentially adsorbs to the Q2 groups. In addition, Fe(II) ions adsorbed to the two types of surface sites have contrasting effects on the dissolution kinetics of ASi, inhibiting dissolution by stabilizing Q2 sites, and enhancing dissolution by catalyzing the detachment of Q3 groups. Thus, the redox cycling of Fe can induce an apparent redox dependence of ASi dissolution, which consequently affects the recycling rate of ASi and the timing of DSi release in freshwater systems. In Chapter 4, I investigate the effects of the oxidative precipitation of Fe(II) on the immobilization of DSi in the absence of ASi. I present kinetic data on the concurrent consumption of aqueous Fe(II) and DSi during their co-precipitation in batch experiments, at different pH values and in the presence of variable initial dissolved phosphate (DP) concentrations. The data, combined with kinetic modeling, indicate that the consumption of DSi during Fe(II) oxidation proceeds along two pathways. At the beginning of the experiments, the oxidation of Fe(II)-DSi complexes induces the fast removal of DSi. Upon complete oxidation of Fe(II), further DSi removal is due to adsorption to surface sites of the Fe(III) oxyhydroxides. The presence of DP effectively competes with DSi via both of these pathways during the initial and later stages of the experiments, with as a result more limited removal of DSi during Fe(II) oxidation. Additionally, results from heterogeneous column experiments show that in porous media the transport of dissolved reactants imposes further controls on the oxidation kinetics of Fe(II) and, therefore, on the removal kinetic of DSi. Overall, I conclude that the oxidation of Fe(II) can immobilize DSi, but that DSi immobilization is strongly diminished in the presence of DP. In Chapter 5, I present the results of experiments using a natural sediment collected from a pond in Cootes Paradise marsh, Ontario. Flow-through experiments with sediment columns are carried out by flowing anoxic solutions containing variable concentrations of Fe(II), DSi and DP via the lower inlet. The outflow side of each columns is exposed to aerated water, hence creating an oxic upper layer of sediment in the columns. The inflow of Fe(II) causes the retention of DP in the sediment as a result of the precipitation of Fe(III) in the uppermost sediment. However, the Fe(III)-enriched surface layer is unable to completely retain all the DP supplied to, as well as produced within, the sediment columns, resulting in considerable DP flux across the sediment water interface. In contrast to Fe(II) and DP, there is little evidence of DSi retention within surficial Fe(III) rich precipitates when the column surface is exposed to oxic conditions. When anoxia is induced in the overlying water, the release of Fe(II) and DP from the columns is enhanced significantly. However, no corresponding changes to DSi efflux are observed upon switching to anoxic overlying water. The constant net production of DSi in the sediment is likely due to the dissolution of naturally occurring biogenic silica. Overall, the results with the natural sediment confirm that, at near-neutral pH, the presence of high DP concentrations inhibit the co-precipitation and adsorption of DSi with Fe(III) oxyhydroxides, hence preventing DSi retention in sediments at the interface with oxic overlying water. Therefore, I conclude that the speciation and stability of legacy P pools in sediment, as well as recent P inputs exert a control on the capacity of sediment to release or retain Si by decoupling Fe and Si cycling. In the final chapter of my thesis (Chapter 6), I summarize the main findings of my research, as well as their implications for the recycling of nutrient Si in freshwater environments. I further present possible future research directions into the biogeochemical cycle of this often overlooked nutrient element.