You will need Adobe Reader to view some of the files on this page. See EPA's PDF page to learn more.Arsenic Removal from Drinking Water by Process Modifications to Coagulation/Filtration
U.S. EPA Demonstration Project at Lidgerwood, ND Final Evaluation Report (78 pp, 2.9 MB) December 2006
This report documents the activities performed and the results obtained for the arsenic removal treatment technology demonstration project at the Lidgerwood, North Dakota, site. The objectives of the project were to evaluate: (1) the effectiveness of process modifications to an existing coagulation/gravity filtration plant in removing arsenic to meet the new arsenic maximum contaminant level (MCL) of 10 µg/L, (2) the reliability of the treatment system, (3) the required system operation and maintenance (O&M) and operator skills, and (4) the capital and O&M cost of the technology. The project also characterized water in the distribution system and process residuals produced by the treatment system.
The pre-existing 250 gal/min (gpm) treatment system consisted of pre-chlorination, forced draft aeration, KMnO4 oxidation, polymer addition, detention, gravity filtration, post-chlorination, and fluoridation. Chemicals were added into a rapid mix tank ahead of a 15,000-gal baffled detention tank, which provided about 60 min of detention time. Afterwards, water flowed into four 7.0 ft × 4.3 ft gravity filter cells, each containing a 24-in deep bed of manganese dioxide (MnO2)-coated anthrasand filter media manufactured by General Filter Products. The pre-existing treatment plant reduced total arsenic concentrations to an average level of 31 µg/L in the treated water, thus requiring process modifications to achieve arsenic levels below the new arsenic MCL.
The process modifications included the installation of an iron addition system and a supplemental polymer addition system. A series of jar and full-scale process tests were conducted to determine a set of optimum process conditions, which consisted of the addition of 1.2 mg/L (as Fe) of ferric chloride, 0.3 mg/L of Aqua Hawk 9207 PWG polymer (note that 0.1 mg/L of Aqua Hawk 9207 PWG polymer had already been added to the rapid mix tank prior to the demonstration study), and 0.5 mg/L of Aqua Hawk 127 polymer. These process conditions were implemented on January 1, 2005, and lasted until July 31, 2005, for the demonstration study.
During the seven-month demonstration study period, the system operated for a total of 1,300 hr with an average daily operating time of 6.1 hr/day. Based on wellhead totalizer readings, the system treated approximately 22,102,000 gal of water with an average daily water demand of 89,788 gal during this time period. The treatment system processed approximately 283 gpm of raw water from the wellhead and 26 gpm of reclaim water from the backwash recovery basin. This is equivalent to a hydraulic loading rate of about 2.6 gpm/ft2 to the filters.
The gravity filters were backwashed automatically every Monday, Wednesday, and Friday. The median filter run time was 13.3 hr with durations of run time ranging from 8.7 hr to 27.2 hr between two consecutive backwash cycles. This is equivalent to a median throughput of 225,834 gal of raw water without reclaim and a range of 147,726 to 461,856 gal of raw water throughput without reclaim. The longer filter run times up to 27.2 hr were associated with operations over the weekends (between Fridays and Mondays). Based on headloss measurements, it was determined that the rate of differential pressure (?p) buildup across the filters was 2.7 in of H2O/hr. Therefore, in order not to exceed 50 in of H2O headloss during the filter runs, the filter run times should be limited to no longer than 15 hr with a wellhead flowrate of 283 gpm and a reclaim flowrate of 26 gpm.
Total arsenic levels in raw water ranged from 113 to 158 µg/L with an average value of 129 µg/L. Arsenic was present primarily in the As(III) form at an average value of 125 µg/L. Total iron levels in source water averaged 1,344 µg/L and existed primarily in the soluble form. This amount of soluble iron corresponded to an iron:arsenic ratio of 9:1 given the average soluble iron and soluble arsenic levels in raw water. Because this was below the target ratio of 20:1 for effective arsenic removal, supplemental iron addition was required at an average dose of 1.2 mg/L (as Fe) using a ferric chloride solution.
After detention and prior to the filter, approximately 38% of arsenic was removed through settling within the baffled detention tank. Based on the average iron dose of 1.2 mg/L and the total iron levels in the raw water, approximately 37% of the iron particulates also were removed within the baffled detention tank.
After the filters, total arsenic levels were reduced to 6.3 to 14.3 µg/L and averaged 8.5 µg/L. Arsenic in the treated water was present primarily as As(V) at an average of 5.7 µg/L. Particulate arsenic levels ranged from <0.1 to 4.9 µg/L and averaged 1.1 µg/L. Total iron levels in the treated water (existing solely as particulates) ranged from <25 to 64 µg/L.
Due to particulate arsenic breakthrough (up to 14.3 µg/L) from the filters, an increase in backwash frequency would be required to maintain the filter performance to achieve levels consistently below the 10 µg/L MCL. Additional process modifications were implemented based on recommendations developed from this demonstration study. The modifications included: (1) installing a 40-gpm backwash reclaim pump to provide additional capacity for daily backwash, (2) implementing a more frequent backwash schedule, and (3) reducing the wellhead pump flowrate to lower the hydraulic loading rate to the filters. The 40-gpm reclaim pump was installed at the plant on October 18, 2005. The wellhead flowrate was reduced to an average value of 239 gpm, which after including the 40 gpm reclaim flowrate, would yield a hydraulic loading rate of 2.3 gpm/ft2 to the filters. The operator also performed filter backwash over the weekends in October 2005 and anticipated performing daily backwash as the water demand increased in the spring and summer.
The existing plant was backwashed automatically on Mondays, Wednesdays, and Fridays. This backwash schedule was maintained during the demonstration study period due to the limited capacity for backwash reclaim given the original plant infrastructure. The rate of backwash water production was approximately 5.5% of the amount of treated water produced. The backwash water contained relatively low levels of soluble arsenic (i.e., 9.8 µg/L on average) and soluble iron (i.e., <25 µg/L on average). The solids in the backwash water contained 7.63E+03 to 1.15E+04 µg/g of arsenic and 1.99E+05 to 3.07E+05 µg/g of iron. The backwash solids passed the Toxicity Characteristic Leaching Procedure (TCLP) test with arsenic in the leachate at <0.5 mg/L. Only barium at 0.069 mg/L and chromium at 0.054 mg/L were detected in the leachate. The TCLP regulatory limit set by EPA is 5 mg/L for arsenic, 100 mg/L for barium, and 5 mg/L for chromium. As such, the backwash solids were non-hazardous and could be accumulated and disposed of at a landfill.
Arsenic levels in water samples collected from the distribution system averaged 12.1 µg/L after process modifications, which was higher than the average arsenic level of 8.5 µg/L in the treated water. The higher levels in the distribution system might be due to longer filter runs over the weekends or solubilization, destablization, and/or desorption of arsenic-laden particles/scales within the distribution system. More frequent backwash as implemented in October 2005 would help to eliminate the longer filter run times over the weekends. Since the process modifications, iron levels in the distribution system remained at non-detectable levels at <25 µg/L. Manganese levels were generally lower in the distribution system samples at 6.7 µg/L compared to 17.9 µg/L in the treated water. Lead and copper levels in the distribution system were not affected by the process modifications.
The capital investment cost was $57,038 which included $32,452 for equipment, $5,786 for engineering, and $18,800 for installation. The capital cost was solely for the new equipment required for the iron addition system, second polymer mixer, and reclaim pump. This does not include the cost for the second polymer feed system because an existing spare chemical feed pump and tank were used. The incremental O&M cost was estimated at $0.04/1,000 gal based on the supplemental iron and polymer dosages required to achieve the target process conditions. Including the O&M cost for all chemical supplies (i.e., chlorine, potassium permanganate, Aqua Hawk 9207 PWG polymer, Aqua Hawk 127 polymer, and fluoride), electrical usage, and labor, the total O&M cost was estimated at $0.52/1000 gal of treated water.