Objective

Perchlorate is a strong oxidizer that is primarily used in solid rocket fuels, fireworks, explosives, and road flares. While perchlorate can be generated from natural processes, the majority of occurrence in the United States is from anthropogenic sources. Perchlorate is a human health concern because it can prevent assimilation of iodide in the thyroid by competitively inhibiting its uptake. Iodide regulates normal functions of the thyroid and is critical in the growth and development of fetuses, infants, and children. Nitrate (NO3-) is commonly found as a co-contaminant in water with perchlorate because ammonium nitrate is a main component in rocket fuel and explosives.

Anoxic biodegradation can be used to treat perchlorate and nitrate, and it can result in complete destruction of the contaminants. The membrane biofilm reactor (MBfR) process demonstrated in this project used the latest advances in membrane technologies and included anoxic biological reduction using a staged hydrogen-fed membrane biofilm reactor, aerobic biological stabilization, media filtration, and disinfection. The objective was to evaluate the feasibility of the MBfR to destroy perchlorate and nitrate in groundwater and produce potable water.

Technology Description

The MBfR process uses permeable, but non-porous hollow-fiber membranes pressurized with hydrogen gas to promote autotrophic bioreduction of perchlorate to chloride ion and nitrate to nitrogen gas. Hydrogen is fed to the lumen of hollow-fiber gas-transfer membranes, and bacteria grow naturally as a biofilm on the exterior of the membranes exposed to contaminated water. The treatment system included two anoxic MBfRs operated in series to reduce oxygen to water, nitrate to nitrogen gas, and perchlorate to the chloride ion. Post-MBfR treatment processes included aeration to re-oxygenate the water, media filtration supplemented with a coagulant/filter aid to remove suspended solids, and disinfection using sodium hypochlorite.

Demonstration Results

Perchlorate was reduced from an average of 154±5 μg/L to an average of 9.2±2.3 μg/L in the effluent of the lag reactor during Steady State (94.4% reduction). While the treatment objective of 6 μg/L was not met at all times, perchlorate removal was consistent (coefficient of variation was 0.75%). Special batch tests demonstrated that complete perchlorate removal was possible, but was commensurate with sulfate reduction and sulfide generation. Modeling and bench-scale studies conducted by Arizona State University (ASU) in parallel to the pilot studies demonstrated that complete perchlorate removal and minimal sulfide production could be achieved as long as the removal flux of nitrate and oxygen – expressed as stoichiometric hydrogen demand – was about 0.18 grams of hydrogen per meter squared per day (g-H2/m2-day) in the second stage of the MBfR. Operation under these conditions in the laboratory prevented overgrowth of sulfate reducing bacteria. Other differences between the laboratory and pilot-scale systems, such as trans-membrane liquid velocity and associated mass-transfer resistance, may have also led to different performance in the bench- and pilot-scale systems.

The MBfR was highly effective at removing nitrate. Total nitrogen (the sum of nitrate and nitrite) was reduced from an influent average of 9.0 milligrams nitrogen per liter (mg-N/L) to an average of 0.12±0.07 mg-N/L in the effluent of the lag reactor during Steady State (98.3% reduction). Nitrate reduction was consistently removed (coefficient of variation was 0.94%), with the highest effluent total nitrate only 0.24 mg-N/L.

Other drinking water treatment goals that were evaluated included disinfection, odor, turbidity, dissolved organic carbon (DOC), and pH. During Steady State, Escherichia coli, fecal coliforms, and total coliforms were below the detection limit (2/100 mL) in all post-disinfection samples. Heterotrophic plate counts (HPCs) were on average 43 most probable number per milliliter (MPN/mL), and no samples were greater than the maximum contaminant level (MCL) of 500 MPN/mL. Disinfection byproducts were below regulatory limits. Haloacetic acids (HAA5) were below detection (< 6 μg/L) and total trihalomethanes (TTHMs) averaged 4.8 μg/L compared to the MCL of 80 μg/L. Nitrosamines were not detected. The average threshold odor number (TON) during Steady State was 2.2 compared to the U.S. Environmental Protection Agency (USEPA) National Primary Drinking Water Regulation’s secondary standard for TON of three. An average turbidity of 0.27 nephelometric turbidity units (NTUs) was observed at the filter effluent during Steady State. Media-filter optimization would have resulted in lower turbidities, but was not part of the study.

Comparing the MBfR system with ion exchange (IX) showed that the MBfR was more economical for nitrate removal, particularly when wastewater disposal for IX regeneration is included, since IX resin regeneration disposal costs are site-specific. Wastewater from the MBfR system, which includes media backwash water and MBfR sparging water, can be discharged through the municipal sanitary sewer. However, wastewater generated during IX regeneration cannot be directly discharged to a municipal sewer mainly because of the high salt concentrations. MBfR costs also were compared with results from ESTCP project ER-200544, Direct Fixed-Bed (FXB) Biological Perchlorate Destruction Demonstration. The MBfR was shown to have similar or lower total treatment costs for nitrate removal.

Implementation Issues

The MBfR system for treatment of nitrate and production of potable water was shown to be effective. The MBfR system is ready for applications involving treatment of drinking water sources contaminated with nitrate. Treatment of perchlorate to less than 6 μg/L was not achieved consistently at the pilot scale. The parallel research conducted by ASU provides possible ways to address this current limitation. Conditional acceptance of the MBfR has been obtained from the California Department of Health (fgc). The first full-scale MBfR system for treatment of nitrate (ARoNite™) in drinking water is in the process of being permitted at Cucamonga Valley Water District.