|Arsenic Removal From Drinking Water by Point-of-Entry/Point-of-Use Adsorptive Media, U.S. EPA Demonstration Project at Oregon Institute of Technology at Klamath Falls, OR, Final Performance Evaluation Report (EPA/600/R-11/035) April 2011
This report documents the activities performed during and the results obtained from the arsenic removal treatment technology demonstration project at Oregon Institute of Technology (OIT) at Klamath Falls, Oregon. The objectives of the project were to evaluate the:
The project also characterized the water in the distribution system and residuals generated by the treatment process. The types of data collected included system operation, water quality, process residuals, and capital and O&M costs.
The three POE systems included one 30-gallon per minute (gpm) system installed in Purvine Hall, one 60-gpm system installed in the Residence Hall, and one 60-gpm system installed in the College Union. The 30-gpm system had two parallel treatment trains, each consisting of two 18-inch by 65-inch fiber reinforced plastic (FRP) vessels configured in series. Each vessel contained 5 cubic feet (ft3) of ArsenXnp (without underbedding), an iron-modified/resin-based media manufactured by Purolite. Each 60-gpm system consisted of two 36-inch by 2-inch FRP vessels configured in series. The Residence Hall system contained 20 ft3 per vessel of ARM 200 media (without underbedding), an iron oxide/hydroxide media developed by Engelhard Corporation. Because of an unexpected short run length, the lead vessel was first rebedded with ARM 300 (a newer version of ARM 200 media) and then rebedded, along with the lag vessel, with E33-S. E33-S is an iron-based media developed by Lanxess, formerly Bayer AG. The College Union system contained 16 and 19.5 ft3 of Adsorbsia GTO (a titanium dioxide-based media) in the lead and lag vessels, respectively. About 4 and 0.5 ft3 of Adsorbsia GTO media were washed from each vessel during initial system backwash, resulting in a large quantity of milky white effluent to be discharged to the sewer.
At the design flowrates of 30 and 60 gpm, the POE systems had a design hydraulic loading rate of 8.5 gpm per cubic foot and a design empty bed contact time (EBCT) of 2.5 minutes per vessel. Because the systems were operated on demand, actual flowrates were based on user consumption and could vary significantly from (but mostly lower than) the respective design flowrates. As such, actual hydraulic rates could be much lower than the design value of 8.5 gpm per cubic foot and actual EBCTs much longer than the design value of 2.5 minutes.
Source water at OIT was supplied by two wells (Well No. 1 and Well No. 4), with Well No. 1 as the primary well. After being chlorinated with a gas chlorine addition system, chlorinated water was stored in a 250,000-gallon aboveground storage tank before being distributed to the campus. At each building, water pressure was reduced to 50 to 76 pounds per square inch by a pressure-reducing valve before water was allowed to flow into the adsorption system. The inlet water contained 24.7 to 35.8 μg/L of total arsenic, existing predominately as soluble arsenic(V). Iron and manganese concentrations were low, typically less than method detection limits (MDLs) at 25 and 0.3 μg/L, respectively. pH values ranged from 7.3 to 8.5 and averaged 8.0. The only competing anion in the inlet water was silica, with concentrations ranging from 27.8 to 31.5 mg/L (as silicon dioxide).
The performance evaluation studies for the ArsenXnp and ARM 200 systems began on December 12, 2005. Because of the issues with media backwashing, the Adsorbsia GTO system required rebedding after initial media loading. Based on its pilot test results, Dow Chemical performed media backwashing in February 2006. The performance evaluation study for the Adsorbsia GTO system did not begin until February 17, 2006. The ArsenXnp system operated for a total of 1,353 days without media changeout. The ARM 200 system operated for 406 days. Because of early arsenic breakthrough, the lead vessel was rebedded with ARM 300 media on January 24, 2007, and then with E33-S media, together with the lag vessel, on October 3, 2007. The Adsorbsia GTO system operated for a total of 1,286 days without media changeout.
Under on-demand conditions with varying system flowrates and water usage, E33-S media performed the best, achieving approximately 50,000 bed volumes (BV) before reaching 10 μg/L arsenic in vessel effluent. Adsorbsia GTO media was the next best performer, achieving 33,500 BV. ARM 200, ARM 300, and ArsenXnp media had a rather common media life spanning from 13,940 to 16,200 BV.
Little or no pressure rise was observed across the adsorption vessels during the performance evaluation studies; therefore, the POE systems did not require backwash. However, one backwash was performed for all three systems on May 16 or 17, 2006. Backwash wastewater and solids samples were collected and analyzed for pH, total dissolved solids, total suspended solids, and/or metal contents.
Results of the distribution system water sampling indicated that arsenic concentrations decreased significantly from an average of 29.2 μg/L before system startup to an average of 0.8 and 0.6 μg/L in Purvine Hall and the College Union, respectively, after system startup. Distribution system water samples in the Residence Hall essentially mirrored ARM 200 system effluent. Post-baseline lead concentrations ranged from less than 0.1 to 1.6 μg/L, which were below the action level of 15 μg/L. Post-baseline copper concentrations ranged from 39.7 to 448 μg/L, which were below the action level of 1,300 μg/L.
Eight Kinetico POU units were installed either under a sink or inside a drinking water fountain in eight different buildings on campus, but only three were monitored for their performance. Each POU unit used a single cartridge to house 600 milliliter of ARM 200 media for arsenic removal. A shut-off assembly and an indicator on the outside of the filter head were used to measure and show the relative remaining cartridge capacity, based on a maximum capacity of 500 galloons. When 500 gallons of water was processed, the shut-off assembly was completely closed, preventing any more water from passing through the cartridge. About 11 months into the performance evaluation study, OIT began to install 40 AdEdge E33-S POU units and replace the eight Kinetico units with AdEdge units.
Total arsenic concentrations were consistently reduced to below 1.0 μg/L by two Kinetico ARM 200 POU units after treating up to 740 gallons of water. The third ARM 200 unit removed arsenic to less than 2.1 μg/L after treating 500 gallons of water; its effluent concentration rose steadily to 6.0 μg/L after treating 1,000 gallons of water. Arsenic was consistently removed to the MDL by three AdEdge E33-S POU units after treating 500 gallons of water. The E33-S unit could treat up to 3,000 gallons of water.
The normalized capital cost was $1,862 per gpm ($1.29 per gallon per day [gpd]) for the ArsenXnp system, $992 per gpm ($0.69 per gpd) for the ARM 200 system, and $1,221 per gpm ($0.85 per gpd) for the Adsorbsia GTO system. Based on the actual annual water use rate, the unit capital cost was $10.77 per 1,000 gallons for the ArsenXnp system, $0.93 per 1,000 gallons for the ARM 200 system, and $4.14 1,000 gallons for the Adsorbsia GTO system. The O&M costs included replacement media and labor to operate the system. Media replacement is the major cost for an adsorptive media system. Using vendor-provided quotes, a series of cost curves was constructed for all media tested (excluding ARM 300) in terms of media cost per 1,000 gallons of water treated as a function of the projected media run length to the 10 μg/L arsenic breakthrough.
You will need Adobe Reader to view some of the files on this page.