Frank van Raffe

Waste generation is increasing globally. Governments are striving to manage the gener ated waste via sustainable strategies such as recycling, yet a share of (untreatable) waste needs to be landfilled. Landfilled waste may lead to adverse impacts on human health and the environment, as landfills are a significant source of the greenhouse gas methane and of leachate rich in contaminants. Such adverse impacts are traditionally reduced via sealing of the waste with impermeable barriers together with capture and treatment of the gas and leachate; an effective method, yet in which degradation processes are halted and the potential to pollute remains. Consequently, costs of replacing the barrier and treating the leachate are shifted to future generations. A potentially more sustain able method of landfill management is to instead actively stabilize the waste via active introduction of air (aeration) or recirculation of the leachate (leachate recirculation). This so-called ‘active treatment’ intends to either degrade, flush, or retain contaminants, to a point where the waste ultimately no longer poses a threat to the environment. To assess the potential of active treatment, three Dutch landfill-scale pilots were instigat ed with the infrastructure to extensively monitor changes in gas quality, leachate quality, and changes in the solid waste. This thesis specifically focused on changes in leachate quality during active treatment and aimed to elucidate the underlying geochemical mechanisms. This was done by conducting experiments from the lab- to the field scale, for each scale identifying the main controlling parameters. The lab experiments were supported by multi-surface geochemical model calculations – calculations which com bine models predicting ion-binding to reactive surfaces (clay, metal (hydr)oxides, organic matter) with a thermodynamic database to calculate inorganic speciation and mineral precipitation. Thus, the distribution of contaminants across different chemical species could be identified, which can be linked to processes controlling contaminant solubility. Mechanisms were first identified in chapter 2 in the most well-defined lab set-up: the pH-dependent test. The pH-dependent solubility of contaminants was determined ex perimentally, which was supported by geochemical model calculations to identify three distinct contaminant groups: non-reactive contaminants, reactive anions, and reactive cations. Non-reactive contaminants had only limited interaction with the solid waste and did not show any pH-dependent solubility. Reactive ions did have a pH-dependent solu bility, with a maximum solubility around the upper and lower ends of the pH range. Reac tive anions had a minimum solubility around pH 5, steered by binding to metal (hydr) oxides. Reactive cations had a minimum solubility at varying pH, based on the cations’ tendency to hydrolyse (resulting in a solubility minimum at higher pH) and their affin ity to bind to organic matter (OM) (resulting in a solubility minimum at lower pH). The mechanisms steering solubility varied between reactive cations and included binding to OM, binding to metal (hydr)oxides, and the formation of mineral precipitates. Ammonium (NH4 + ) surprisingly showed behaviour akin to a reactive cation with a solubility minimum around pH 8. The distinction between non-reactive contaminants, reactive anions, and reactive cations formed the basis of contaminant behaviour interpretation in the next chapters. Additionally, a discrepancy was observed between geochemical model calcula tions and experimental results for NH4 + , due to the absence of binding parameters, and chromium (Cr), likely due to poor OM-binding parameters. A first step in improving model calculations was taken in chapter 3, where new OM binding parameters were derived for Cr. Adsorption experiments were conducted where at different pH levels and in the presence of humic- or fulvic acid, the solution was titrated with Cr and free Cr was measured to determine total adsorption. This data was used to fit NICA-Donnan parameters, which could only be done by considering the binding of the free Cr3+ species and the binding of Cr2(OH)2 4+. The derived parameters were com pared against the originally existing parameters using three independent datasets. The newly derived parameters showed a significant improvement in predicting Cr binding to specific OM fractions, and in predicting free Cr concentrations. However, in soil samples marked with low amounts of (dissolved) OM, binding of Cr was underestimated and the prediction of dissolved Cr worsened when compared to predictions using the original parameter set. It was thus concluded that the newly derived Cr parameters were a sig nificant improvement in predicting environmental Cr speciation, yet that care should be taken when applying the parameters to low OM environments. On a larger scale than the pH-dependent test, 60L landfill simulation reactors containing waste from all three landfills were applied to identify effects of leachate recirculation, aeration, and their combination on contaminant solubility in chapter 4. Leachate recircu lation was marked by strongly reduced conditions, where dissolved OM concentrations were stable and pH increased. Based on their identified pH-dependent solubility, the change in pH resulted in an increase for most contaminants due to hydrolyzation and mineral dissolution. Contrary to the pH-dependent test, the landfill simulation reactors supported by geochemical modelling identified that for several contaminants, i.e. Cd, Cu, Fe, Mn, and Zn, their solubility was steered by the formation of sulphide minerals. Aeration resulted in oxidized conditions, decreased dissolved OM, and decreased pH. The oxidized conditions led to a dissolution of sulphide minerals with a corresponding increased solubility of associated elements, which subsequently decreased as other min eral precipitates were formed or the elements bound to different reactive surfaces. Am monium concentrations strongly decreased during aeration as the presence of oxygen allowed for the process of nitrification and NH4 + was transformed to nitrate. The landfill simulation reactors ultimately showed the importance of redox conditions, and that in active treatment there is a precarious balance, and switching treatments and changing redox conditions may lead to a decrease in some contaminants and an increase in others. Solubility of contaminants was finally explored on the landfill scale in chapter 5 by sampling pore water from one of the landfills using vadose monitoring systems. A high covariance could be observed between the different parameters. Concentrations of dissolved OM and contaminants increased with depth yet sharply decreased below 12 metres, likely due to the presence of an impermeable layer. By combining the insights of the previous chapters and estimating solid waste properties based on solid waste extrac tions, the geochemical model could capture these trends well for most contaminants. Thus, the importance of binding to specific reactive surfaces, formation of minerals, and importance of redox could be coupled to specific contaminants, and their concentrations could be well-predicted based mostly on measurements of pH and dissolved OM. This shows the applicability of geochemical models to identify contaminant speciation even in such complex systems. Finally, all content chapters were synthesized in chapter 6, and their findings were out lined against the overall research goal of elucidating mechanisms underlying contami nant solubility in landfills. A reflection was made of the used methodologies in this thesis, particularly regarding the investigation of the different scales (what scales tell what) and the different assumptions made and problems encountered when applying the multi surface model. Considering the poor model description of NH4 + in the different content chapters, an attempt was made at improving this description by performing NH4 + -OM adsorption experiments and by modelling pH-dependent NH4 + solubility considering selective binding to clay; particularly the latter proved promising for future studies. The implications of this thesis’ findings for active treatment were discussed by interpreting changes in leachate quality of the full treated landfills and by discussing the criteria for when a landfill can be considered safe for the environment. Ultimately, this thesis’ find ings show the potential and the pitfalls when actively treating landfill waste to reduce contaminant emissions.