Push-Pull Tests to Support In Situ Chemical Oxidation System Design

Abstract

The problems associated with the contamination of groundwater environments by non-aqueous phase liquids (NAPLs) such as chlorinated solvents, gasoline and manufacturing gas plant (MGP) residuals, including their distribution and persistence, are well accepted. The treatment of groundwater by in situ chemical oxidation (ISCO) relies on the oxidation potential of chemical reagents to destroy harmful organic compounds. The interaction of these oxidants with target and non-target compounds in the subsurface will help determine effectiveness and efficiency of an ISCO treatment system. Push-pull tests (PPTs) have the utility to estimate key properties in situ and allow for sampling a larger volume of aquifer to yield more representative estimates as compared to conventional bench-scale tests. The scale and cost-effectiveness of a PPT make it an ideal tool to collect valuable information on subsurface system behaviour so that uncertainties can be minimized. The use of PPTs to provide insight into treatment expectations or to support the design of an ISCO system requires a suitable interpretation tool. A multi-species numerical model (‘PPT-ISCO’) in a radial coordinate system was developed to simulate a PPT with the injection of a conservative tracer and oxidant (persulfate or permanganate) into the saturated zone of a porous medium environment. The pore space may contain variable amounts of immobile, multicomponent, residual NAPL. The aquifer material contains a natural organic matter (NOM) fraction and/or other oxidizable aquifer material (OAM) species. The model is capable of simulating mass transport for an arbitrary number of conservative and reactive tracers and NAPL constituents subjected to chemical reactions. The ability of PPTs to capture the in situ natural oxidant interaction (NOI) was tested with PPTISCO. Breakthrough curve (BTC) data collected from permanganate and persulfate PPTs conducted in the field were compared to simulated BTCs by assigning the same field operational parameters to the model and applying NOI kinetic information obtained from batch tests. These tests confirmed the usability of the model and PPTs to obtain the NOI kinetics from PPT BTCs. The sensitivity of PPT BTCs to variations in the field operating and NOI parameters were investigated. The results of varying the field operating parameters indicated that the oxidant BTCs could be scaled to match varying injection and extraction flow rates. Variations in NOI parameters revealed that the permanganate BTC is primarily controlled by the permanganate fast reaction rate coefficient and the quantity of OAM present in the aquifer. The spatial profiles of OAM across the test zone revealed that the majority of the OAM consumption is from the fast fraction and occurs in the vicinity of the well where the permanganate concentration is greatest. An estimate of the permanganate fast reaction rate coefficient can be obtained from a permanganate PPT BTC by employing the model to simulate the PPT with the operational parameters (used in the field) and literature estimates of the remaining NOI parameters. Calibration between the simulated and observed BTCs can be undertaken to adjust the permanganate fast reaction rate coefficient to fit the permanganate PPT BTC. Persulfate NOI sensitivity investigations revealed that persulfate PPT BTCs can be characterized by a concentration plateau at early times as a result of the increased ionic strength in the area around the injection well. The ionic strength is primarily controlled by the injected persulfate concentration, and as persulfate degrades into sulphate and acid, the ionic strength is enhanced. Graphical analysis of the BTC revealed that an underestimated value of the persulfate degradation rate coefficient can be obtained from the PPT BTC. A more representative estimate of the persulfate degradation rate coefficient can be achieved after fitting the field BTC to the simulated results, applying the underestimated value as a starting point. PPTs investigating ISCO treatability have the ability to provide insight into the effect of the NOI on the oxidation of target compounds, site-specific oxidant dosage requirements and NAPL treatment expectations. NAPL component BTCs from treatability PPTs are primarily controlled by the mass in the fast region, and the fast region mass transfer rate coefficient. Oxidation estimates extracted from NAPL component BTCs were shown to accurately approximate the mass of each NAPL component oxidized when compared to model calculations. The mass of NAPL oxidized for each of the components yields a site-specific oxidant dosage. This estimate exceeds what is prescribed by the stoichiometry between permanganate and the contaminant of concern due to the effect of the NOI. The utility of PPTs to study and quantify the interaction between injected oxidants and the aquifer material has been demonstrated with PPT-ISCO. In addition, PPT-ISCO has revealed that treatability PPTs can be tailored to investigate the dosage requirements and treatment expectations of residual NAPLs. Results from this effort will be used to support ongoing field research exploring the use of PPTs to assist in understanding the competing subsurface processes affecting ISCO applications.