Environ. Sci. Technol. 2004, 38, 6846-6854
Use of Polymer Mats in Series for Sequential Reactive Barrier Remediation of Ammonium-Contaminated Groundwater: Field Evaluation B . M . P A T T E R S O N , * ,† M . E . G R A S S I , †,‡ B. S. ROBERTSON,† G. B. DAVIS,† A. J. SMITH,† AND A. J. MCKINLEY‡ CSIRO Land and Water, Private Bag No. 5, Wembley, Western Australia 6913, Australia, and Chemistry, School of Biomedical and Chemical Sciences, The University of Western Australia, Nedlands, Western Australia 6907, Australia
A pilot-scale field trial was undertaken to evaluate the potential of in situ polymer mats (installed in series) as permeable reactive barriers within a treatment wall remediation system to induce sequential bioremediation of ammonium-contaminated groundwater. The treatment wall consisted of 10 m wide impermeable wings on either side of a 0.75 m wide permeable reactive zone flowthrough box. Two polymer mats were positioned in the flowthrough box. The upgradient polymer mat within the flowthrough box was used to deliver oxygen to induce bacterial nitrification of the ammonium to nitrite/nitrate as the groundwater moved past. The downgradient polymer mat delivered ethanol to induce bacterial denitrification of the nitrite/nitrate to produce nitrogen gas. The field trial was carried out at a near-shore location. Initially the flowthrough box was left open; however, this resulted in substantial groundwater mixing, which inhibited sequential remediation. Once the flow-through box was in-filled with gravel, groundwater mixing was reduced, achieving a greater than 90% reduction in total N. Estimated firstorder half-lives for nitrification and denitrification rates were 1.2 and 0.4 d, respectively. Field nitrification halflives were approximately an order of magnitude greater than rates determined in large-scale columns using soil and groundwater from the site, while denitrification half-lives were similar. The results of this pilot-scale field trial indicate that sequential bioremediation of ammonium-contaminated groundwater at field scale is feasible using in situ polymer mats as permeable reactive barriers, although hydraulic conditions can be complex in such barrier systems.
Introduction Compounding effects of multiple diffuse and point sources, from industry, sewage treatment lagoons, and fertilizer use, can result in ammonium and nitrate contamination of groundwater (1). Conventional remediation techniques normally involve pumping and capture of polluted ground* Corresponding author phone: 61-8-93336276; fax: 61-893336211; e-mail:
[email protected]. † CSORO Land and Water. ‡ The University of Western Australia. 6846
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004
water, with expensive long-term above-ground treatment of large volumes of water. In situ permeable reactive barriers (PRBs) have been used for a range of groundwater contaminants such as dechlorination of trichloroethene-contaminated groundwater and denitrification of nitrate-contaminated groundwater. PRBs constructed with a slowly degrading carbon source, such as sawdust (2, 3) or cotton (4), have been investigated to promote an anaerobic environment and provide an energy source for denitrifying bacteria. The use of PRBs to treat contaminated groundwater offers the potential for low-cost remediation, as they rely on the natural hydraulic gradient of the aquifer to induce groundwater flow through the permeable barrier, thus avoiding pumping and above-ground treatment. To be cost-effective, long-term maintenance costs need to be low, including amendment or treatment delivery to the PRB. For a recent overview of bioreactive PRBs, see Davis and Patterson (5). For ammonium-contaminated groundwater, sequential nitrification followed by denitrification of nitrate to nitrogen gas is feasible (6). However, for in situ treatment, there are few technologies capable of sequential remediation treatment over a short time frame or groundwater flow distance, especially where there is a need to induce different geochemical redox conditions for contaminant degradation. For mixed groundwater contaminants (e.g., petroleum and chlorinated hydrocarbons), there have been limited in situ sequential remediation strategies investigated at field scale. Morkin et al. (7) conducted an in situ sequential field demonstration treatment of a mixed contaminant plume of petroleum and chlorinated hydrocarbons, using a combination of an iron reactive wall and air sparging. Fiorenza et al. (8) collated data from field evaluations involving a sequential iron reactive wall and slow release of oxygen or air sparging. Polymer-based diffusion systems designed as PRBs have the potential for targeted delivery of amendments to groundwater contaminants. These targeted PRBs may provide in situ sequential remediation strategies over a short time frame or groundwater flow distance. Also, they may overcome regeneration problems as the amendments can potentially be maintained indefinitely, leading to reduced costs associated with regenerating PRBs. Polymer-based diffusion systems have been used in laboratory-scale membrane aeration bioreactors (MABRs) for nitrification (9) and removal of dissolved organic carbon (DOC) in synthetic wastewaters, and in a membrane reactor for the deoxygenation of a mineral medium (10). In situ diffusion systems using polymers have also been used to deliver dissolved benzene and trichloroethene (TCE) as groundwater tracers (11). Oxygen has also been delivered using polymers to stimulate in situ biodegradation of BTEX (benzene, toluene, ethylbenzene, and xylene) (12), methyl tert-butyl ether (MTBE) (13), and atrazine (14) contaminated groundwaters. Recently this polymer membrane approach has been applied to ammonium-contaminated groundwater as a sequential nitrification and denitrification treatment/ delivery system in laboratory column experiments (6). Scale-up of this polymer-based delivery system for the remediation of ammonium-contaminated groundwater has now been undertaken. In this paper we present results from a pilot-scale field demonstration trial using polymer mats as permeable reactive barriers for a dual treatment/delivery system to remediate ammonium-contaminated groundwater via sequential nitrification and denitrification. 10.1021/es0497781 CCC: $27.50
2004 American Chemical Society Published on Web 11/06/2004
FIGURE 1. Field site showing the layout of the treatment wall (excavated soil to 1 m below ground) prior to installation. The flow-through treatment box is on the left of the photo, prior to lifting and installation. Sheet piles used for the wings are shown in the background.
Site Description The site for the field demonstration is near Perth, Western Australia, on the Swan Coastal Plain. The site was located within 30 m of the shoreline of Cockburn Sound (15). The stratigraphy of the site consisted of approximately 11 m of medium to fine dune sand overlying a clay aquitard. The water table was approximately 1.0 m below ground. Leaks and spills of ammonium products from a fertilizer factory approximately 300 m upgradient of the site had resulted in an ammonium-contaminated groundwater plume. Multilevel piezometer data at the location of the treatment wall showed the plume was approximately 6 m thick, extending from approximately 5 m below ground to the clay aquitard at 11 m below ground. At this location, the maximum total N concentration of 110 mg L-1 was located 6 m below ground. The ammonium-contaminated plume was anaerobic, but only mildly reductive (Eh between +80 and -50 mV) with a low DOC content of 7 mg L-1. Further details of the site groundwater chemistry are given by Patterson et al. (6).
Materials and Methods Design of the Treatment Wall. The pilot-scale treatment wall was emplaced within an ammonium plume to test, at pilot scale, the feasibility of the in situ sequential polymer mat remediation system. The test was not designed to capture and remediate the entire ammonium plume. The treatment wall was designed with impermeable wings to direct a proportion of the ammonium-contaminated groundwater plume through the permeable reactive zone. Initial column experiments using a polymer mat delivery system showed that the oxygen and ethanol delivery rates did not limit nitrification and denitrification rates at groundwater flow rates up to 300 m yr-1 (6). As these flow rates were approximately an order of magnitude greater than estimated groundwater velocities at the field site, the treatment wall was designed with 10 m wide impermeable wings on either side of a 0.75 m wide permeable reactive zone flow-through box. Due to cost limitations, the treatment wall was designed using 7 m deep impermeable wings and a 7 m deep permeable reactive zone flow-through box. As a result, the treatment wall was not keyed into the clay aquitard at the base of the sand aquifer (11 m below ground), and only targeted the top of the ammonium plume, where the total N (sum of NH4+
N, NO3--N, and NO2--N concentrations) concentration was ∼10 mg L-1. Preliminary modeling used an aquifer permeability of 25 m d-1, a wing permeability 3 orders of magnitude lower, and a flow-through box permeability 3 orders of magnitude greater than the aquifer permeability. Modeling predicted that using 10 m wide (and 7 m deep) impermeable wings on either side of a 0.75 m wide permeable reactive zone, not keyed into the clay aquitard at 11 m below ground, would increase the groundwater flux rates through the flow-through box and reactive zone by a factor of 2.7, but flow rates would be less if the flow-through box was in-filled with gravel. Since the treatment wall was not keyed into the clay aqutard, only a proportion of the groundwater flowing to the treatment wall would flow through the flow-through box. Groundwater close to the base of the impermeable wings would tend to travel under the wings rather than through the permeable reactive zone, reducing the efficiency of the wings with depth. Also, groundwater close to the sides of the impermeable wings would tend to travel around the wings. Construction of the Treatment Wall. The 0.75 m wide, 1 m long, 7 m deep permeable reactive zone flow-through box was constructed of galvanized mild steel (
