SURFACE COMPLEXATION MODELING OF COAL COMBUSTION RESIDUALS CONSTITUENTS ON NATURAL HYDROUS FERRIC OXIDE
Analytics
429 views ◎381 downloads ⇓
Abstract
Coal ash impoundments, along with other industrial and natural processes, can be sources of constituents of concern (COCs) to human health including arsenic (As), chromium (Cr), selenium (Se), and vanadium (V). Attempts to model the fate and transport of these COCs within subsurface aqueous environments often rely on extensive laboratory batch experiments to develop distribution coefficients (Kd values) for model input. Surface complexation theory can also explain the partitioning of constituents between the liquid and solid phases, with hydrous ferric oxide (HFO) often acting as a dominant solid phase adsorbent. While useful, such model descriptions generally apply to pure mineral phases and do not investigate natural adsorbents. This study applies surface complexation theory in the form of the generalized two-layer model (GTLM) to natural soils from coal ash sites by modeling the laboratory batch experiments used to generate soil isotherms and soil Kd values.Laboratory batch experiments equilibrated different ratios of soil solid phase with a synthetic groundwater solution liquid phase containing As, Cr, Se, and V, allowing for solution equilibrium and isotherm generation. Modeled batch solid phases were represented by surface concentrations of hydrous ferric oxide derived from two sequential extraction procedures: one utilizing a modified Citrate-Bicarbonate-Dithionite (CBD) method, and another using an acidified hydroxylamine (Chao) solution. Model liquid phases were input by creating representative synthetic groundwater solutions within the model. The defined phases were then equilibrated within the model so that isotherms and Kd values could be generated for all COCs.Experimental and modeled results were statistically analyzed and the relative error (RE) of the modeled Kd values to their experimental counterparts was calculated. Overall Chao extraction method showed little to no correlation with experimental results, shifting the focus of discussion to CBD method. Modeled CBD method As and Se results showed moderate correlations with experimental results; however, model Kd values tended to trend lower than experimental making them conservative. CBD method V results showed slight correlation with experimental data; however Kd values trended higher than experimental and were not conservative. CBD method Cr results indicated no correlation, with little apparent connection to experimental results and Kd values which were far larger and thus not conservative. Probable reasons for inconsistencies between model and experimental data include discrepancies between actual HFO content in soils and HFO calculated from sequential extraction, the exclusion of adsorptive mineral phases from the component additivity model, PHREEQC database equilibrium constants derived from pure mineral phase isotherms rather than natural soils, and the potential introduction of unknown constituents into solution from soil samples. While the Cr and V models used are likely too inaccurate for practical use, As and Se models show potential. With reasonable modification, As and Se models can return conservative Kd value estimates representative of natural soils.The main body of this work is supplemented with additional files including model and experimental isotherms, PHREEQC inputs and outputs, and the PHREEQC database utilized herein. Model and experimental isotherms are presented in the form of a spreadsheet with a .xlsx file type. PHREEQC files include both the .pqi and .pqo file types, and require the PHREEQC software provided by the United States Geological Survey (USGS). The database is included as a .txt file and must be designated as the default database within PHREEQC. Additional supplemental file information can be found in Appendix C.