Field Demonstration of a Novel Biotreatment Process for Perchlorate Reduction

Dr. Marc Deshusses | Duke University

ER-200636

Objectives of the Demonstration

The main objective of this project was to test and demonstrate the efficacy of zero valent iron (ZVI)-supported biological reduction of perchlorate. Additional objectives were to obtain pertinent data that will guide full-scale design and operation, provide relevant data for treatment cost estimation and comparison, provide the necessary data leading to possible permitting of the process by the California Department of Public Health (DPH), and disseminate the results in various forms to promote technology transfer. Significant treatment performance issues occurred that motivated further laboratory studies to help identify the causes of the problems observed in the field.

Back to Top

Technology Description

In this process, treatment relies on autotrophic perchlorate-reducing bacteria immobilized on ZVI. As ZVI corrodes in water, hydrogen is released from the reduction of water, which is then used by perchlorate-reducing bacteria as a source of electrons. Hydrogen minimizes biomass growth and has a low potential for disinfection by-product precursors compared to organic substrates such as acetate. Other potential advantages include the possible reduction (biotic or abiotic) of nitrate, trichloroethene (TCE), hexavalent chromium, and the possible control of metals by adsorption.

Back to Top

Demonstration Results

A trailer-mounted pilot demonstration system was designed, built, and mobilized at Well #2 in Rialto, California. It consisted of a water holding tank, the ZVI packed bed (300 gal) with approximately 4400 lb cast iron aggregate (mesh size 3/5), two parallel sand filters for post-treatment removal of bacteria and iron leaving the ZVI bioreactor, and ancillary monitoring and control equipment. Shortly after the start of the system, a pretreatment unit was installed to remove some of the influent water dissolved oxygen (DO), as the influent water was found to be saturated with oxygen. The treatment system was designed to treat a nominal water flow of 20 gpm, corresponding to an empty bed water residence time of 15 minutes, i.e., similar to many of the units that were operated in the laboratory. The experimental plan called for a number of tests at various flow rates, perchlorate concentrations, and overall operating conditions to fully evaluate the process.

The pilot reactor was operated at the Well #2 site from August 2007 to May 2008, during which time the uptime exceeded 98%. However, treatment performance varied considerably, with essentially 3 months of flawless operation, followed by numerous treatment performance problems that eventually forced the shutdown of the system.

During the initial 3 months, the average effluent concentration of perchlorate was 1.8 ± 0.9 parts per billion (ppb); nitrate was <0.01 mg L-1 NO3-N; iron ranged from 0 to 0.05 mg L-1; and coliforms, fecal coliforms, and E. coli in the reactor effluent were all below the detection limits. Extensive characterization of the ZVI bioreactor was achieved. One caveat was that these results were obtained at a flow of 4 gpm, i.e., well below the nominal treatment capacity of 20 gpm. Problems occurred before the treatment performance at high flow could be determined.

After approximately 3 months of operation, perchlorate and nitrate removal began to deteriorate, with less than 40% perchlorate and nitrate removal for the remaining 5 months of testing. This was unexpected as laboratory bioreactors ran over 2 years without a problem. Several hypotheses for the loss in treatment efficacy were tested. These included a severe reduction of the ZVI bed hydraulic conductivity, loss of perchlorate-reducing bacteria or of biological activity, or loss in iron reactivity. Various attempts were made to recover full treatment capacity, but all failed and full treatment capacity was never recovered. The most likely explanation was a loss of hydraulic conductivity, although it is likely that biological factors and iron reactivity also played a role.

Significant efforts were then placed on a series of laboratory experiments with scaled-down ZVI columns aimed at determining the causes of the failure of the field unit. These led to the conclusion that a single factor was not responsible for the failure of the demonstration unit. Instead, a combination of adverse conditions, with perhaps design and operating choices (nature of pretreatment, low water flow), were likely responsible for the failure of the field bioreactor. Thus, the water chemistry was instrumental, and it remains unclear whether similar failure would have been observed had the demonstration been carried out at a different facility with groundwater with a different composition.

Back to Top

Implementation Issues

Perchlorate biotreatment using bacteria immobilized on ZVI is novel and can potentially offer cost and other benefits for perchlorate and nitrate removal. However, it is susceptible to environmental factors. The technology is not yet demonstrated, and it requires further study prior to full-scale implementation. In particular, the following unresolved issues should be considered prior to implementation of the technology:

  • Water chemistry, in particular alkalinity, nitrate concentration, and DO are important for ZVI biosystems; greater attention should be given to these parameters.
  • While there is a large body of literature on abiotic ZVI reactive systems, the effects of bacteria on the ZVI corrosion and on the longevity and stability of ZVI bioreactors remain poorly understood, and further studies are warranted.
  • Mineral supplementation used for biological stimulation can have undesirable consequences and result in mineral deposits that can either reduce the bed porosity or passivate iron surfaces and thus should be exerted with care.
  • Greater attention should be placed on the iron balance in ZVI packed beds and the effects of iron mesh size, bed geometry, and bed height/water velocity relationships. Experiments in the laboratory showed that beds with fine ZVI rapidly plugged when beds of coarse ZVI did not.
  • Bed porosity, hydraulic issues, and ZVI passivation are the greatest challenges to long-term sustained treatment performance in ZVI biotreatment systems. Reactor designs other than packed beds should be considered, with designs that can effectively deal with the adverse effects identified above. Possible designs include fluidized beds, circulating or moving beds, and other yet to be developed designs.

Back to Top

Points of Contact

Principal Investigator

Dr. Marc Deshusses

Duke University

Phone: 919-660-5480

Program Manager

Environmental Restoration

SERDP and ESTCP

Share