Environ. Sci. Technol. 2002, 36, 3439-3445
Use of Polymer Mats in Series for Sequential Reactive Barrier Remediation of Ammonium-Contaminated Groundwater: Laboratory Column Evaluation B . M . P A T T E R S O N , * ,† M . E . G R A S S I , †,‡ G. B. DAVIS,† B. S. ROBERTSON,† AND A. J. MCKINLEY‡ CSIRO Land and Water, Private Bag No. 5, Wembley, Western Australia 6913, Australia, and Department of Chemistry, The University of Western Australia, Nedlands, Western Australia 6907, Australia
Large-scale column experiments were undertaken to evaluate the potential of in situ polymer mats (installed in series) to be used as permeable reactive barriers for delivery of oxidants and reductants to induce sequential bioremediation of ammonium-contaminated groundwater (∼120 mg L-1 NH4+-N), without bioaugmentation. The strategy was for the first group of polymer mats to deliver oxygen to induce bacterial nitrification of the ammonium to nitrite/ nitrate as the groundwater moved past and for the second group of polymer mats to deliver hydrogen or ethanol, to induce bacterial denitrification of the nitrite/nitrate to produce nitrogen gas. Once purging of the first polymer mat commenced, ammonium concentrations decreased downgradient from the polymer mats. Nitrification rates increased and stabilized over the 6-month experiment, with stable nitrification half-lives in the range 0.07-0.25 days. Nitrification most likely occurred in a biologically active zone at the polymer wall/aqueous interface. With hydrogen delivery via the polymer mats, a denitrification half-life (nitrate plus nitrite removal) of 3.5 days was induced. Denitrification rates were significantly enhanced when ethanol was delivered via a polymer mat, with denitrification half-lives in the range of 0.12-0.34 days. Nitrification/ denitrification rates were maintained for groundwater flow rates up to 300 m yr-1, suggesting oxygen and ethanol delivery rates via the polymer mats were sufficient not to limit nitrification or denitrification. In soil columns, the polymer mat delivery system provided an effective and reliable technique for delivery of oxygen and hydrogen or ethanol for sequential nitrification/denitrification of ammonium-contaminated groundwater. Scale-up of this concept to a field pilot-scale is currently underway.
Introduction Ammonium and nitrate contamination of groundwater comes from the compounding effects of multiple diffuse and * Corresponding author phone: 61-8-93336276; fax: 61-893336211; e-mail:
[email protected]. † CSIRO Land and Water. ‡ The University of Western Australia. 10.1021/es0157088 CCC: $22.00 Published on Web 07/04/2002
2002 American Chemical Society
point sources, from industry, sewage treatment lagoons, and fertilizer use (1). Conventional techniques for remediation often involve pumping and capture of polluted groundwater, with expensive long-term above ground treatment of large volumes of water. In situ remediation is possible through the use of permeable reactive barriers (PRBs) to treat contaminated water as it flows through the barrier. PRBs offer 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, i.e., no pumping or above ground treatment is involved. Recent reviews of the applicability of PRBs are found in Blowes et al. (2) and Davis and Patterson (3). The longevity of PRBs, such as zerovalent iron barriers (4) and sorption barriers made from ground rubber (5), is actively being researched. Regeneration of such PRBs may be prohibitively expensive. Repta and Gillham (6) observed progressive loss of catalytic trichloroethene degradation activity using a nickel-enhanced iron barrier. Innovation in the delivery mechanisms for amendments or in regeneration capabilities may increase the long-term cost-effectiveness of PRB systems. PRBs using polymer mats (constructed from coiled polymer tubing) have the potential to overcome regeneration problems as the delivery of amendments (via diffusion) can potentially be maintained indefinitely, potentially reducing costs associated with regenerating PRBs. Polymer-based diffusion systems have been used in laboratory scale membrane aeration bioreactors (MABR) for nitrification (7) and removal of dissolved organic carbon (DOC) in synthetic wastewaters and in a membrane reactor for the deoxygenation of a mineral medium (8). In situ diffusion systems using polymers have also been used to deliver dissolved benzene and trichloroethene (TCE) as groundwater tracers (9). Oxygen has also been delivered using polymers to stimulate in situ biodegradation of BTEX (benzene, toluene, ethylbenzene, and xylene) (10), methyl tertiary butyl ether (MTBE) (11), and atrazine (12) contaminated groundwaters. In this paper, laboratory column experiments are used to evaluate the concept of using polymer mats as permeable reactive barriers as a dual treatment/delivery system to remediate ammonium-contaminated groundwater via sequential nitrification and denitrification.
Methods Column Experimental Setup and Operation. Two aluminum columns (2.0 m long, 14 cm i.d. diameter) were constructed with fine stainless steel mesh, 10 cm from the top and bottom of the column to prevent soil depositing in the feed and effluent lines. Six polymer mats were installed in each column, emplaced horizontally within the circumference of the column and orthogonal to the water flow direction. These coil mats were placed at 73, 83, 93, 103, 113, and 123 cm from the base of the column. The mats were arranged in this way so amendments could be delivered through a number of mats until complete nitrification and then denitrification of the groundwater occurred. Twelve water sampling ports, each consisting of an inverted mesh covered needle protruding 4 cm into the column, were also positioned along the length of each column, at 4 (port A), 68 (port B), 78 (port C), 88 (port D), 98 (port E), 108 (port F), 118 (port G), 128 (port H), 138 (port I), 153 (port J), 168 (port K), and 182 cm (port L) from the base of the column. The first port was set below the stainless steel mesh (below the soil) and was used to monitor influent aqueous concentrations to the column. In situ oxygen probes were also installed at 68 (probe 1), 78 (probe 2), 88 VOL. 36, NO. 15, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3439
FIGURE 1. Schematic of the flow through column. Insert: photo of polymer mat. (probe 3), 98 (probe 4), and 108 cm (probe 5) from the base of the column. A schematic of a column is shown in Figure 1. Each polymer mat consisted of a 5.0 m length of silicone tubing (2.0 mm i.d., 3.0 mm o.d.,) with a fine stainless steel spring inserted into the center of the polymer tubing to provide support and to prevent twisting or collapsing of the tubing. The polymer tubing was then woven through a 13.5 cm diameter flexible plastic support frame. The theoretical cross-sectional area of a mat was 150 cm2, giving a crosssectional coverage of the column of 97%. Due to the weaving and overlapping of the tubing, the thickness of the mats was up to 1.5 cm and the density of the coverage was reduced. Hydraulic conductivity estimates, based on hydraulic head differences and water fluxes, were relatively high for regions of the columns with (2.9 m day-1) and without (4.9 m day-1) the polymer mats, indicating the mats did not provide significant resistance to water flow in the soil columns. The mats were installed in the column with the inlet and outlet ports of the polymer tubes attached to the side of the column. For oxygen and hydrogen delivery, gases were flushed through the inner gas space of the polymer tubing via the inlet port on the side of the column, with effluent gas discharging via the outlet port which was open to the atmosphere. For ethanol delivery, 2-3 L of an aqueous ethanol solution (50-100 g L-1) was continuously recycled through a polymer mat. For both columns (column 1, oxygen/hydrogen amended; column 2, oxygen/ethanol amended), the first two upgradient polymer mats (at 73 and 83 cm from the base of the column) were purged with oxygen. For column 1, the remaining four polymer mats (at 93, 103, 113, and 123 cm from the base of 3440
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 15, 2002
the column) were purged with hydrogen gas. For column 2, of the remaining four polymer mats, only the polymer mat at 93 cm from the base of the column was purged with a ethanol solution. The remaining three mats had their ends sealed and were not used. The columns were filled with low organic-content (0.04% w/w) leached Spearwood sand. Preliminary experiments showed no significant retardation of ammonium, nitrate, or nitrite using this sand. The columns were operated in saturated up-flow mode, under a hydraulic head which varied from 1.0 to 1.5 m. Gravity volumetric flow through each column was restricted using a peristaltic pump on the effluent line to a constant flow of 720 mL day-1 giving a flow velocity through the sediments of 42 m yr-1 (assuming a soil porosity of 0.41). This flow is similar to groundwater velocities found on the Swan Coastal Plain in Perth, Western Australia (13). This construction and operation were used in order to avoid entrapment of gas and oxygen ingress into the column influent groundwater through contact between plastic tubing used in the peristaltic pump. Experiments were run at room temperature, which was controlled coarsely at 22 °C. Groundwater for the columns was collected every 2 weeks from a bore within a known ammonium groundwater plume and stored under an atmosphere of nitrogen to maintain anaerobic conditions. Groundwater chemical data are shown in Table 1. The soil and groundwater in the columns were used without bacterial amendment (no bioaugmentation). Purging Rates through Polymer Tubing. Gases or liquids were continuously purged through the inner volume of the tubing to ensure a uniform gas/liquid concentration (and therefore chemical gradient) along the internal volume of the tubing. The amendment (oxygen, hydrogen, or ethanol)
TABLE 1. Groundwater Chemistry Data parameter NH4+-N NO3--N NO2--N SO42--S PO4--P ClHCO3Na K Mg Ca Fe S2dissolved organic carbon dissolved oxygen
concentration (mg
L-1)
120
