Jump to main content.


  Arsenic Removal From Drinking Water by Adsorptive Media, EPA Demonstration Project at Rollinsford, NH, Six-Month Evaluation Report (EPA/600/R-05/116) October 2005

This report describes the activities and results of the first six months of the arsenic removal treatment technology demonstration project at the Rollinsford Water and Sewer District facility in Rollinsford, New Hampshire. The objectives of the project are to evaluate the:

  • Effectiveness of AdEdge Technologies' AD-33 media in removing arsenic to meet the new arsenic maximum contaminant level of 10 micrograms per liter (µg/L)
  • Reliability of the treatment system (Arsenic Package Unit [APU]-100)
  • Simplicity of the required system operation and maintenance (O&M) and operator skill level
  • Capital and O&M costs of the technology

The project is also characterizing the water in the distribution system and residuals produced by the treatment system process. The types of data collected include system operation, water quality (both across the treatment train and in the distribution system), process residuals, and capital and O&M costs.

The APU-100 treatment system consisted of two 36-inch-diameter by 72-inch-tall fiberglass-reinforced plastic vessels in parallel configuration, each containing approximately 27 cubic feet of AD-33 media. AD-33 is an iron-based adsorptive media developed by Bayer AG and marketed under the name of AD-33 by AdEdge. (It is identical to Severn Trent Services' SORB 33 media used at larger arsenic removal systems.) The system was designed for a peak flow rate of 100 gallons per minute (gpm) (50 gpm to each vessel) corresponding to a design empty-bed contact time of about 4 minutes per vessel and a hydraulic loading to each vessel of about 7 gallons per minute per square foot (gpm/ft2).

The AdEdge treatment system began regular operation on February 9, 2004. From February 9 to August 13, 2004, the system treated approximately 7,158,000 gallons of water, or about 19,500 bed volumes.

Breakthrough of total arsenic concentrations above the 10 µg/L target level was first observed during the May 25, 2004 sampling event at 12,500 bed volumes. Concentrations in the treated water were below 10 µg/L during the next sampling event on June 8 but again exceeded the target level of 10 µg/L on June 22. Based on these data, it appears that breakthrough of arsenic at concentrations above the target level occurred when 4.5 to 5.5 million gallons (12,500 and 15,000 bed volumes) of water had been treated. This volume represents about 15 to 20 percent of the vendor-estimated working capacity of the AD-33 media.

Prior to breakthrough, the system reduced total arsenic levels from 28.7–46.3 µg/L in raw water to less than 10 µg/L in the treated water. The soluble arsenic concentration in the raw water included an average of 18.3 µg/L of arsenic (III) and 14.8 µg/L of arsenic (V).

In March, 2004 total arsenic levels in the treated water were observed at concentrations of 5.5 to 7.7 µg/L, and the majority of arsenic passing through the AD-33 media was arsenic (III). Prechlorination was added to the treatment train on March 24, 2004, and was effective at oxidizing arsenic (III) to arsenic (V). Following the switch to prechlorination, the average arsenic (III) concentration in the treated water dropped to 0.6 µg/L, which was very similar to the arsenic (III) concentration seen in untreated water sampled upstream of the adsorption system.

Total and free chlorine residuals measured before the adsorption vessels ranged from 0.05 to 0.40 mg/L (as chlorine) for free chlorine and 0.20 to 0.71 mg/L (as chlorine) for total chlorine.

Total and free chlorine residuals measured after the adsorption vessels ranged from 0.04 to 0.05 mg/L (as chlorine) for free chlorine and 0.23 to 0.26 mg/L (chlorine) for total chlorine.

This indicates little or no chlorine consumption by the AD-33 media.

Influent total iron concentrations ranged from 37 to 489 and averaged 156.4 µg/L, with the majority of iron present in the soluble iron (II) form. Upon prechlorination, iron precipitated immediately and was filtered by the media.

Influent total manganese levels ranged from 52 to 245 µg/L and averaged 114.0 µg/L, with the majority of manganese present in the soluble manganese (II) form. Prior to prechlorination, manganese quickly broke through the AD-33 media, reaching about 100 percent breakthrough after about 3,700 bed volumes of water treated. Unlike iron, manganese remained mostly in the soluble form upon prechlorination, indicating slow oxidation kinetics. However, following adsorption, manganese was decreased to less than 10 µg/L, suggesting that the presence of chlorine promoted the removal of manganese on the surface of the AD-33 media.

Results of the distribution samples collected before and after the installation and operation of the APU-100 system showed no discernable trend in any of the distribution sampling results collected, indicating that the treatment system had little to no effect on the water quality in the distribution system. This was probably caused by the blending of the treated water with untreated water from another well location used to supply water to the town's looped distribution system. The blending of the treated water with the untreated water might have masked any detectable effects of the APU-100 system on the water quality in the distribution system.

Three backwash water samples were collected during the first six months of system operation. Arsenic concentrations in the backwash water ranged from 11.1 to 33.4 µg/L. In most cases, arsenic, iron, and manganese concentrations were lower than those in the raw water (backwash was performed using raw water from the supply wells), indicating some removal of these metals by the media during backwash.

Since startup, the APU-100 system experienced higher than expected pressure drops across the treatment system and elevated inlet pressure. In multiple attempts to address these elevated pressure conditions, backwashing was conducted repeatedly with flow rates up to 11 gpm/ft2, as recommended by the vendor. However, the aggressive backwashing did not appear to be effective in solving the elevated pressure problems. In addition, there were periods when the system was bypassed due to the elevated pressure conditions.

Extensive troubleshooting and replacement of certain system components were performed to address the problems encountered. However, at the end of the first six months of the evaluation period, the system continued to operate under elevated pressure higher than that expected based on original design information.

The capital investment of $106,568 included $82,081 for equipment, $4,907 for site engineering, and $19,580 for installation. Using the system's rated capacity of 100 gpm (144,000 gallon per day [gpd]), the capital cost was $1,066 per gpm of design capacity ($0.74 per gpd) and equipment-only cost was $821 per gpm of design capacity ($0.57 per gpd). These calculations did not include the cost of the building construction.

O&M costs included only incremental costs associated with the adsorption system, such as media replacement and disposal, chemical supply, electricity, and labor. Although not incurred during the first six months of system operation, the media replacement cost represented the majority of the O&M cost and was estimated to be $16,810 to change out both vessels. This cost was used to estimate the media replacement cost per 1,000 gallons of water treated as a function of the projected media run length to the 10-µg/L arsenic breakthrough.


Thomas Sorg

See Also

Arsenic Research

You will need Adobe Reader to view some of the files on this page.
See EPA's PDF page to learn more.

Office of Research & Development | National Risk Management Research Laboratory

Local Navigation

Jump to main content.