Zero-Valent Iron-Assisted Autotrophic Denitrification

Zero-Valent Iron-Assisted Autotrophic Denitrification

Downloaded from ascelibrary.org by Tampere Univ of Technology on 03/28/13. Copyright ASCE. For personal use only; all rights reserved.

Susham Biswas1 and Purnendu Bose2 Abstract: Porous reactive barriers containing metallic iron and hydrogenotrophic denitrifying microorganisms may potentially be suitable for in-situ remediation of nitrate-contaminated groundwater resources. The main objective of the research described here was to determine the type and concentration of metallic iron to be used in such reactive barriers so that ammonia formation through metallic iron-assisted abiotic nitrate reduction was minimized, while a reasonable rate of biological denitrification, sustained by hydrogen produced through metallic iron corrosion, was maintained. Initial experiments included the demonstration of autotrophic denitrification supported by externally supplied hydrogen, either from a gas cylinder or generated through anaerobic corrosion of metallic iron. Next, the effect of iron type on abiotic nitrate reduction was studied, and among those types of iron tested, steel wool, with its relatively low surface-area-toweight ratio, was identified as the material that exhibited the least propensity to abiotically reduce nitrate. Further, long-term experiments were carried out in batch reactors to determine the effect of steel wool surface area on the extent of denitrification and ammonia production. Finally, experiments carried out in up-flow column reactors containing sand and varying quantities of steel wool demonstrated biological denitrification occurring in such systems. Based on the results of the final set of experiments, it appeared that to minimize ammonia production, the steel-wool concentration up-flow columns must be even below the lowest value—0.5 g steel wool added to 125 cm3 of sand—used during this study. To counter any detrimental effect of lowered steel wool concentration on the extent of hydrogenotrophic denitrification, increase of the retention time in the columns to values higher than 13 days 共the maximum value investigated in this study兲 may be necessary. DOI: 10.1061/共ASCE兲0733-9372共2005兲131:8共1212兲 CE Database subject headings: Iron; Denitrification; Ground-water pollution; Abatement and removal.

Introduction Autotrophic denitrifying microorganisms use inorganic carbon as a food source, and rely on electron donors like hydrogen or reduced sulfur compounds for energy to reduce nitrate to nitrogen gas. Autotrophic denitrification using hydrogen 共Kurt et al. 1987兲, thiosulfate 共Claus and Kutzner 1985a,b, and sulfide 共Kleerebezem and Mendezà 2002兲 as electron donors has been reported in literature. Hydrogenotrophic denitrifiers—that is, denitrifying microorganisms utilizing hydrogen as energy source—are ubiquitous in nature 共Till et al. 1998; Gamble et al. 1976兲. Hydrogenotrophic denitrification of water was studied in a benchscale fluidized bed reactor 共Kurt et al. 1987兲, and Sulzer Water and Wastewater Treatment Company introduced a commercial design for hydrogenotrophic denitrification called “Denitropur” 共Hellekes 1986兲. In another study, Alcaligenes eutrophus, a hydrogenotrophic denitrifier, was immobilized in polyacrylamide and alginate copolymer to evaluate denitrification in fluidized-bed 1

Environmental Engineering and Management Program, Dept. of Civil Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, India. 2 Environmental Engineering and Management Program, Dept. of Civil Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, India. E-Mail: [email protected] Note. Discussion open until January 1, 2006. Separate discussions must be submitted for individual papers. To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor. The manuscript for this paper was submitted for review and possible publication on November 19, 2002; approved on January 26, 2005. This paper is part of the Journal of Environmental Engineering, Vol. 131, No. 8, August 1, 2005. ©ASCE, ISSN 0733-9372/2005/8-1212– 1220/$25.00.

and batch reactors 共Chang et al. 1999兲. Lee and Rittman 共2002兲 reported the application of a novel hollow-fiber membrane biofilm reactor for hydrogenotrophic denitrification of water. Several other researchers—Dries et al. 共1988兲, Gros and Tretler 共1986兲, Gros et al. 共1988兲, Ergas and Reuss 共2001兲, Liessens et al. 共1992兲—have also studied treatment systems for hydrogenotrophic denitrification of water. One of the main challenges of the hydrogenotrophic denitrification process is the high cost and technical difficulties in the production and supply of hydrogen gas 共Gayle et al. 1989兲. In this regard, hydrogen production through in-situ anaerobic corrosion of metallic iron seems to be a promising option. Metallic iron or Fe 共0兲 is oxidized in anaerobic environments, resulting in the reduction of water to hydrogen gas, as described by Eq. 共1兲. Anaerobic iron corrosion: Fe共0兲 + 2H2O → H2 + Fe2+ + 2OH− 共1兲 It is conceivable that this hydrogen may be used for autotrophic denitrification of nitrate as described by Eq. 共2兲. Biological nitrate reduction: 2NO−3 + 5H2 → N2 + 4H2O + 2OH− 共2兲 Till et al. 共1998兲 and Kielemoes et al. 共2000兲 showed that hydrogenotrophic denitrification could be sustained by hydrogen gas produced through anaerobic iron corrosion, using both pure culture 共Paracoccus denitrificans兲 and a mixed culture of hydrogenotrophic denitrifying microorganisms. Several researchers have studied metallic iron-aided abiotic nitrate reduction 共Murphy 1991; Ginner et al. 2004; Huang and Zhang 2004; Choe et al. 2004; Huang et al. 2003; Huang and

1212 / JOURNAL OF ENVIRONMENTAL ENGINEERING © ASCE / AUGUST 2005

J. Environ. Eng. 2005.131:1212-1220.

Zhang 2002; Devlin et al. 2000兲 and reported near stoichiometric quantities of ammonia being produced as the end product, as per Eq. 共3兲.

product distribution in batch reactors and in up-flow column reactors containing hydrogenotrophic denitrifying microorganisms and various amounts of the chosen type of metallic iron.

Abiotic nitrate reduction:


4Fe共0兲 + NO−3 + 7H2O → 4Fe2+ + NH+4 + 10OH−

共3兲

Other researchers 共Choe et al. 2000兲 have reported complete metallic iron-assisted abiotic nitrate reduction to nitrogen gas under certain conditions, or formation of a mixture of ammonia and nitrogen gas as the end-product of nitrate reduction 共Chi et al. 2004; Huang et al. 1998兲. Hu et al. 共2001兲 reported reduction of nitrite, an intermediate product produced during nitrate reduction, to ammonia and nitrogen gas by metallic iron. In a field-scale study reported by Westerhoff and James 共2003兲, nitrate reduction in columns containing metallic iron resulted in apparent conversion of between 30–75% of the applied nitrate to nitrogen gas, with the conversion percentage increasing with increase in column run time. Kinetic studies on abiotic nitrate reduction by metallic iron in pure systems 共Alowitz and Scherer 2002兲 and in presence of competing anions 共Su and Puls 2004兲 have also been reported. In summary, formation of ammonia is an undesirable consequence of the metallic iron-assisted abiotic nitrate reduction process. In comparison, metallic iron-assisted hydrogenotrophic denitrification seems preferable, as the only end product of this process is nitrogen gas. However, the following concerns regarding the metallic iron assisted hydrogenotrophic denitrification process remain: • Abiotic reduction of nitrate to ammonia may occur in parallel with biological denitrification, resulting in substantial ammonia production. Proper reaction conditions must be maintained to prevent this occurrence. • Both abiotic and biological nitrate reductions result in the production of OH− ions 关see Eqs. 共2兲 and 共3兲兴, which may elevate the pH of the system. Hence, proper buffering must be provided such that pH increase is not excessive and detrimental to the biological nitrate reduction reactions. Considering the relative slowness of the anaerobic corrosion process 关Eq. 共1兲兴, and consequent hydrogen production, metallic iron-assisted hydrogenotrophic denitrification process might be best suited to in-situ nitrate remediation systems such as subsurface reactive barriers 共Blowes et al. 2000兲. The main objective of the research described here was to determine the effect metallic iron type and concentration on the end-product distribution during denitrification in flow-through biologically active reactive porous medium, such that conditions under which ammonia formation is minimized may be identified. Consequently, the two primary parameters, whose effect on hydrogenotrophic denitrification were studied in detail, were 共1兲 the type of metallic iron used, and 共2兲 metallic iron surface area. In addition, preliminary experiments concerning sustenance of hydrogenotrophic denitrification through externally supplied hydrogen, either from a gas cylinder or produced through anaerobic corrosion of metallic iron, were also conducted. Specifically, the research plan was as follows: • Demonstrate the ability of microorganisms cultured in the laboratory to sustain hydrogenotrophic denitrification reactions with external hydrogen supply; • Characterize different types of metallic iron samples in terms of abiotic nitrate-reduction capacity and ammonia production, and consequently select the most suitable iron type for further experimentation; and • Determine the extent of denitrification and consequent end-

Materials and Methods Analytical Procedures Nitrate and nitrite were measured using an ion chromatograph 共Metrohm 761兲 equipped with a Phenomenex STAR ION A 300 IC anion column and a conductivity detector with ion suppression. The minimum detection limit of the method used for these determinations was 0.1 mg/ L for both nitrate and nitrite. All samples for nitrate/nitrite determination were filtered through 0.2 ␮m filter paper before injection into the ion chromatograph. Ammonia was measured colorimetrically by Nesslerization 共method no. 417 B, APHA/WEF/AWWA 1985兲, using a spectophotometer 共Spectronic, 20 D+, India兲 with Borosil glass absorbance cells having 1 cm path length. Iron共II兲 was analyzed using the 1,10 phenanthroline method as described in APHA/ WEF/AWWA 共1995兲. The detection limit for this method was 0.1 mg/ L. Total iron estimation was carried out using an atomic absorption spectrophotometer 共spectrAA-20, Varian兲; samples were acidified with concentrated HNO3 before analysis. The pH was measured using a combination pH electrode 共Toshniwal CL-51, India兲 connected to a digital pH meter 共Toshniwal CL-54, India兲; pH also was checked using pH paper 共range pH: 2–10.5兲 when sample volume was low.

Development of Bacterial Culture A mixed culture of autotrophic denitrifying microorganisms was developed and maintained in a glass reactor. Initially, 1,000 mL of a mineral medium containing inorganic carbon 共HCO−3 兲, substrate 共NO−3 兲, buffer 共H2PO−4 兲, and other trace nutrients as per Till et al. 共1998兲 was prepared in distilled water and added to the reactor. The reactor was sterilized, and 10 mL of seed from an anaerobic reactor treating domestic sewage was added to it. The reactor contents were purged with nitrogen gas, followed by bubbling of hydrogen gas into the reactor for 5 min. The reactor was sealed and its contents were kept in continuously mixed condition. After every 6 days, reactor contents were purged with hydrogen. While purging, 100 mL of reactor contents was replaced with 100 mL of the sterilized mineral medium, containing nitrate. This corresponded to a mean cell residence time of 60 days. The effluent was analyzed for NO−3 -N , NO−2 -N , NH3-N, pH, and absorbance at 600 nm, using 4 cm path length spectroscopic cell 共Till et al. 1998兲. After bacterial growth was established, NO−3 -N, NO−2 -N, NH3-N concentrations in the effluent, as measured every 6 days, were below detection limits, while the pH stabilized around 8.5. The absorbance value of the reactor contents increased steadily and stabilized, indicating growth of microorganisms and subsequent maintenance of steady-state microbial concentration in the reactor. Several successive generations of reactors were prepared as described above, but using seed from the reactors of the immediately previous generation. From the second generation onward, the mineral medium was prepared using sterilized groundwater instead of distilled water. Adoption of these procedures ensured that a mixed culture of purely autotrophic denitrifying microorganisms was isolated 共Smith et al. 1994兲. Replacement of distilled water by groundwater ensured that the microorganisms were acclimatized to the groundwater

JOURNAL OF ENVIRONMENTAL ENGINEERING © ASCE / AUGUST 2005 / 1213

J. Environ. Eng. 2005.131:1212-1220.

environment, and hence could be used in experiments involving the same groundwater. The fourth generation cultures prepared as above were used as seed for all subsequent experiments.


Experiment I Experiments were carried out in bottles that contained mineral medium similar to that used for development of bacterial culture, but with varying nitrate concentrations. Six 250 mL bottles were used as reactors for these experiments. At the start of the experiment, each bottle was filled with 200 mL mineral medium with varying nitrate concentrations; namely, 40, 80, 120, 160, 200, and 240 mg/ L 共as NO−3 -N兲. In addition, a seventh bottle with nitrate concentration of 40 mg/ L 共as NO−3 -N兲 was prepared as a nonbiological control. All bottles were sterilized. Then, 5 mL of seed from the stock culture was added to first six bottles. All seven bottles were purged with nitrogen gas for 10 min to remove any residual oxygen, followed by hydrogen bubbling for 5 min. Hydrogen was applied to all bottles every 3 days to ensure that the denitrification reaction in the biologically active bottles was never hydrogen limited. Every 7 to 10 days 10 mL samples were collected from each bottle, steam sterilized to ensure cessation of all bacterial activity, and stored for measurement of residual NO−3 -N, NO−2 -N, NH3-N, and pH.

Experiment II For the experiments of this type, hydrogen gas was generated by anaerobic corrosion of iron metal in a hydrogen-generation reactor of 1,000 mL volume. The gas thus generated was conveyed through a manifold to seven denitrification reactors of 250 mL volume each. The denitrification reactors, with 40, 80, 120, 160, 200, and 300 mg/ L 共as NO−3 -N兲 nitrate were seeded with autotrophic denitrifying microorganisms, while the seventh reactor contained 40 mg/ L nitrate 共as NO−3 -N兲 and was not seeded. Initially, 10 g of electrolytic iron power 共laboratory reagent grade, Reidel Chemicals, India兲, and 1 L of distilled water was added to the hydrogen generation reactor. Then, 200 mL of mineral medium, prepared with sterilized groundwater with different initial nitrate concentrations, as above, was added to each of the denitrification reactors. The hydrogen generation reactor and the denitrification reactors were then connected through the manifold, and the whole setup was steam sterilized. Contents of the denitrification reactors were purged with nitrogen gas for 15 min to remove all residual oxygen in the system. The valves of the reactor setup were maneuvered such that the hydrogen generated in the hydrogen-generation reactor could flow freely to the denitrification reactors. The system was kept in this anoxic and abiotic state for 7 days to ensure that a sufficient amount of hydrogen had migrated into the denitrification reactors to sustain a population of denitrifying microorganisms. After 7 days, 5 mL of seed from the stock culture was introduced into the denitrification reactors, while no seed was added to the blank reactor. Contents of the reactors were then mixed to ensure even distribution of microorganisms throughout each reactor. Every 7 to 10 days, 10 mL samples were collected from each denitrification reactor, steam sterilized to ensure cessation of all bacterial activity, and stored for measurement of residual NO−3 -N, NO−2 -N, NH3-N, and pH.

Experiment Type III These experiments were conducted in biochemical oxygen demand BOD bottles. A typical abiotic nitrate reduction experiment consisted of filling a BOD bottle with mineral medium made with sterilized groundwater. The nitrate concentration in these bottles was approximately 40 mg/ L 共as NO−3 -N兲. A predetermined quantity of iron metal was added to each bottle, after which the bottles were sealed and steam sterilized. After a specified reaction period, which varied from 2 to 7 days, the bottles were unsealed and sampled, samples were extracted and steam sterilized before being stored for analysis of NO−3 -N, NO−2 -N, NH3-N, and pH. Typical biological denitrification experiments were similar to the abiotic experiments just described, but with the following differences. After filling with sterilized groundwater and a predetermined mass of iron, the bottles were steam sterilized and left in sealed condition for 7 days to ensure that enough hydrogen had evolved through anoxic corrosion of iron metal to support microbial denitrification. Then a predetermined quantity of nitrate was added to each bottle such that nitrate concentration was approximately 40 mg/ L 共as NO−3 -N兲. Also, 5 mL of seed was added to each bottle. The bottles were then resealed and the contents well mixed. After a specified reaction period, which varied from 7 to 60 days, samples were extracted and sterilized before being stored for analyzing NO−3 -N, NO−2 -N, NH3-N, and pH. Twelve bottles were used for an initial set of abiotic nitrate reduction experiments, with three bottles seeded with each of the four different types of iron; namely, electrolytic iron powder 共laboratory grade, Reidel Chemicals, India兲, iron filings 共laboratory grade, Nice Chemicals Pvt. Ltd., India兲, iron shavings 共obtained from IIT Kanpur Central Workshop兲, and steel wool 共commercial grade, Rohit Industries, India兲. Initial nitrate concentration in each bottle was 40 mg/ L 共as NO−3 -N兲. Four bottles, each containing a different iron type, were sampled after the second, fourth, and the sixth days to determine the extent of abiotic nitrate reduction by different types of iron. In the second set of experiments, eight bottles were used, of which four were maintained in abiotic condition with steel wool concentration of 5, 4, 3, and 2 g. The other four bottles also had similar steel wool concentrations, but were seeded with denitrifying microorganisms. The initial nitrate concentration in each bottle was 42 mg/ L 共as NO−3 -N兲. All eight bottles were sampled after seven days for comparison of nitrate reduction through the abiotic and biological means. For the final set of experiments of this type, 18 seeded bottles were maintained, of which six bottles had steel wool concentration of 1.5 g / bottle, six other bottles had steel wool concentration of 1 g / bottle, while the steel wool concentration in the final six bottles was 0.5 g / bottle. Initial nitrate concentration in each bottle was 40 mg/ L 共as NO−3 -N兲. Three bottles, with steel wool concentrations of 0.5, 1.0, and 1.5 g / bottle were sampled after days the 7, 14, 20, 30, 40, and 60 to determine the extent of biological denitrification as a function of steel wool concentration.

Experiment IV These experiments were conducted in anoxic flow-through columns operated in the up-flow mode. Nitrate-containing mineral medium flowed through columns containing sand and various amounts of metallic iron. Abiotic columns were maintained in sterile condition throughout the experiment, while the biological


J. Environ. Eng. 2005.131:1212-1220.


Fig. 1. Schematic of the experimental apparatus used for abiotic/ biological nitrate reduction in flow-through columns

denitrifying columns were seeded with microorganisms from the stock culture. All columns were operated in the intermittent-flow mode so that sufficient retention time could be provided in the columns for nitrate-reduction reactions to take place. Samples were collected daily for analysis. Each column consisted of two parts—a small inlet chamber and a main column—separated by a porous sintered disk. Water influent to the column first entered the inlet chamber and then passed through the porous sintered disk to reach the main column, which contained the media. The reactive zone of glass columns used for these experiments were approximately 12 cm in length and 4 cm in diameter. The height of media in each column was 10 cm, which corresponded to a media volume of 125 cm3. Porosity of the media was approximately 50%. The sintered disk also acted as a support for the media in the reactive zone. The media in the column consisted of sand graded to a 1–2 mm diameter, used after thorough washing with acid and water followed by drying in an oven. This sand was mixed with varying quantities of steel wool, and loaded into the columns, after which the columns were steam sterilized. Each column was sealed at the top by a rubber stopper, with an inserted glass tube for sample extraction. The schematic of the experimental apparatus used is shown in Fig. 1. Two columns were used for the abiotic experiments using this apparatus, one containing only sand, and the other sand mixed with 1.0 g steel wool. At the start of the experiment, both chambers of the two columns were fully flooded with freshly prepared warm 共40°C兲 distilled water, which contained little or no dissolved oxygen. The columns were then maintained in sealed condition for 7 days to ensure onset of anaerobic conditions. The columns were then connected to a reservoir containing mineral medium prepared with sterilized groundwater. In addition to containing approximately 35 mg/ L of nitrate 共NO−3 -N兲, 10 mg/ L HgCl2 was added to the mineral medium in the reservoir to inhibit any microbial activity; the whole setup, including the reservoir, columns, and connectors, was steam sterilized in assembled condition to ensure complete cessation of all bacterial activity. Each column was then flushed with approximately 250 mL of mineral medium from the reservoir so that the nitrate concentration in the effluent from the columns was approximately the same as the initial nitrate concentration in the reservoir. Next, the valve connecting the reservoir to the column and the valve controlling water flow out of the column were closed, thus maintaining the column in sealed condition. Once every 24 hours a fixed amount of effluent, either 50, 25, 10, or 5 mL corresponding to retention times of 1.3, 2.6, 6.5, or 13 days respectively, was passed through the columns and collected. The samples were steam sterilized, and stored for measurement of residual NO−3 -N, NO−2 -N, NH3-N, and pH.

Fig. 2. Schematic of the experimental apparatus used for promoting attached growth of hydrogenotrophic denitrifying microorganisms in flow-through columns

Three columns were used for the biological denitrification experiments, containing 0.5, 1.0, and 1.5 g steel wool, mixed with 125 cm3 of sand. These columns were first steam sterilized in sealed condition to ensure cessation of all bacterial activity. For the purpose of seeding with denitrifying microorganisms, the columns were connected to a stock culture reservoir. The schematic of the setup used for seeding the columns is shown in Fig. 2. A multichannel peristaltic pump 共IKEA, Germany兲 was employed for recirculating the stock culture through the columns. Initially, the stock culture reactor was periodically provided with hydrogen from an external source. Colonization of the sand media in the columns by microorganisms was considered to have occurred once, and despite stoppage of the external hydrogen supply, continued nitrate reduction was observed in the stock culture reservoir. To further demonstrate this point, a reservoir of sterilized mineral medium with 40 mg/ L of nitrate 共as NO−3 -N兲 was purged with nitrogen gas and attached to the columns in place of the stock culture reservoir. Continued recirculation of this initially sterilized mineral medium through the column also resulted in reduction in nitrate concentration in the reservoir. At this point, a fresh reservoir of sterilized mineral medium containing 40 mg/ L of nitrate 共as NO−3 -N兲 was attached to the columns, and biological denitrification experiments were carried out in exactly the same way as the abiotic experiments described earlier.

Results and Discussion Experiment I sought to determine the ability of hydrogenotrophic anaerobic denitrifying microorganisms cultured for this study to denitrify aqueous solutions containing nitrate using externally supplied hydrogen as energy source. Results of these experiments are presented in Fig. 3. Nitrate was completely removed within 15–30 days in all reactors 关Fig. 3共a兲兴. The time required for complete nitrate removal increased with increasing initial nitrate concentrations in the bottles. Formation of nitrite as an intermediate product of nitrate reduction was observed in all reactors 关Fig. 3共b兲兴. Nitrite concentration in all reactors initially increased with time to a maximum value of up to 40% of initial nitrate concentration before decreasing with time. This suggests slower nitrite removal kinetics, as compared to nitrate removal rates. Samples were also tested for presence of ammonia; no ammonia could be detected in any of the samples tested. Thus, the extent of denitrification—that is, the removal of nitrogen from the aqueous phase—can be determined from the evolution of total aquatic nitrogen, which is the sum of nitrate and nitrite nitrogen 关Fig. 3共c兲兴. Complete denitrification was observed in reactors with initial nitrate concentrations of 40, 80, and 120 mg/ L nitrate 共as N兲 within 30 days. In reactors with higher nitrate concentra-


J. Environ. Eng. 2005.131:1212-1220.


Fig. 4. Autotrophic denitrification in batch reactors with hydrogen supplied by anaerobic corrosion of metallic iron

Fig. 3. Autotrophic denitrification in batch reactors with external hydrogen supply

supply. This may be attributed to the fact that the microbial growth, and hence denitrification, was limited by the lower supply of hydrogen. Another difference between results presented in Figs. 3 and 4 is the fact that no nitrite production was noticed in the latter case. This difference may be explained as follows. In Experiment II 共see Fig. 4兲, the rate of nitrate reduction was hydrogen limited, and hence considerably slower than in experiment I 共see Fig. 3兲, while the rate of nitrite reduction was not affected as severely. Under the circumstances, any nitrite produced due to nitrate reduction was readily reduced further to nitrogen gas, thus resulting in very little nitrite accumulation in the reactor. The next experimental objective was to demonstrate the denitrification process in a mixed system containing nitrate-containing aqueous solution, metallic iron, and hydrogenotrophic anaerobic denitrifying microorganisms. The choice of metallic iron type to be used for this purpose is of importance, and the iron type with the least ability to abiotically reduce nitrate to ammonia should be favored. Based on the results presented in Fig. 5 it was concluded that over the experimental duration of 6 days, steel wool was least able to abiotically reduce nitrate to ammonia, as compared to the other iron types. Hence, steel wool was chosen as the iron type to be used for all subsequent experiments. The reason for steel wool

tions, approximately 80% denitrification was observed up to the observation period of 35 days. Residual nitrogen species in these reactors after 30 days was entirely nitrite. It is conceivable that this residual nitrite would have been removed if longer reaction time had been provided. Experiment II sought to determine the ability of hydrogenotrophic anaerobic denitrifying microorganisms to denitrify nitrate-contaminated aqueous solutions using hydrogen generated through anaerobic corrosion of metallic iron. The main difference between this and the previous set of experiments was that the amount of hydrogen available in this case was expected to be much less than in the previous case, and hydrogen-limiting conditions were expected to persist. Nitrate concentrations over time are shown in Fig. 4. Nitrate was completely removed within 50 days in reactors having initial nitrate concentration of 40, 80, and 120 mg/ L. Approximately 90% of the nitrate was removed within 60 days in reactors having initial nitrate concentrations of 160, 200, and 300 mg/ L. No ammonia formation and very little nitrite production 共less that 1% of total nitrogen added was converted to nitrite兲 were observed in these reactors. The nitrate removal or denitrification rates were slower during these experiments compared to the previous experiments using external hydrogen

Fig. 5. Abiotic nitrate reduction by various types of metallic iron in batch reactors 共iron quantity: 1 g; reactor volume: 300 mL兲


J. Environ. Eng. 2005.131:1212-1220.


Fig. 6. By-product formation by abiotic and biological nitrate reduction processes in batch reactors over 7 days 共reactor volume: 300 mL兲

being the least effective in reducing nitrate to ammonia appears to be the low surface-area-to-mass ratio of this substance as compared to other iron types like iron powder or iron filings. As per values reported by Till et al. 共1998兲, typical acid-washed metallic iron powder was reported to have a surface-areato-weight ratio of 2 m2 / g, as compared to 1 and 0.568 m2 / g, respectively for the iron filing and iron shaving used in this study, and 0.0075 m2 / g for steel wool. Since the rate of reaction between metallic iron and nitrate is a function of metallic iron surface area, an equal mass of steel wool is expected to interact far less with nitrate and hence produce less ammonia than iron powder. An undesirable corollary of the above observation is the fact that steel wool is also likely to corrode more slowly through interaction with water due to the lower surface area it offers compared to other iron types. This would imply a decline in hydrogen production by anaerobic corrosion, resulting in the retardation of biological denitrification reactions due to the potential of hydrogen-limiting conditions in the system. Questions still remained regarding the amount of steel wool to be used to minimize ammonia formation, while maintaining reasonable biological denitrification rate. It was to answer these questions that the next set of experiments was conducted. These experiments involved adding 2, 3, 4, and 5 g of steel wool in reactors containing 40mg/ L nitrate 共as NO−3 -N兲. Two kinds of reactors were prepared for each steel wool amount, one seeded and the other sterile. Results obtained after 7 days are shown in Fig. 6. Similar results were also obtained after 14 days. Based on the results presented in Fig. 6, the following may be concluded. First, the total aqueous nitrogen concentration—that is, the sum of nitrate, nitrite, and ammonia, expressed as nitrogen—remained nearly unchanged in the sterile 共abiotic兲 reactors. Second, the total nitrogen concentration declined in the seeded 共biological兲 reactors; this is attributed to biological denitrification reactions in these reactors. Third, ammonia formation was noticed in both sterile and seeded reactors, with the amount of ammonia formed increasing with increasing steel wool concentration in the reactors. Fourth, in the seeded reactors, the extent of biological denitrification increased very slightly with increase in steel-wool concentration in each reactor. Fifth, even at the lowest steel wool concentration tested 共2 g兲, considerable amount of ammonia formation was noticed in both seeded and sterile reactors. Based on these results, it seems probable that employing even lower steel wool concentrations in subsequent biological denitrification experiments of similar type may be necessary to minimize ammonia formation. However, this also may reduce hydrogen formation, thus affecting the rate of biological denitrification adversely. Efforts were made to determine long-term biological denitrification characteristics in batch reactors containing 0.5,

Fig. 7. Abiotic and biological nitrate reduction in batch reactors containing various amounts of steel wool 共reactor volume: 300 mL兲

1.0, and 1.5 g steel wool. Results are presented in Fig. 7 . These results indicate that for all three cases nitrate concentrations in the reactor were nearly zero after 60 days. Formation and subsequent disappearance of nitrite also was noticed in all cases. Ammonia formation was low 共⬍5 mg/ L as N兲 up to 40 days in all reactors, after which sharp increases in ammonia concentrations were observed. Extent of denitrification in all reactors was approximately the same, with the total aqueous nitrogen concentration declining from approximately 40 mg/ L to approximately 25 mg/ L in 60 days. The following conclusions may be drawn from the results presented in Fig. 7. First, despite relatively low steel wool concentrations in these reactors, nitrate reduction was not hampered, as seen from the reduction of nitrate concentrations to below detection levels. Second, nitrate reduction was only partially biological, as seen from the buildup of ammonia in all reactors. It appears that for the first 40 days, the nitrate reduction was mostly biological, as seen by the low ammonia concentrations in reactors up to that time. Beyond 40 days, depleted nitrate concentrations in the reactors probably reduced the rate of biological denitrification and hence the abiotic nitrate reduction process gained precedence. This was demonstrated by the steep increase in ammonia production in all reactors after 40 days. Experiment IV was conducted to determine the extent of nitrate reduction during the flow of nitrate-containing water through porous media containing steel wool. Two of these experiments were conducted with unseeded columsn maintained in sterile conditions by addition of 10 mg/ L HgCl2. One of these sterile columsn contained only sand and the other sand mixed with 1.0 g of steel wool. The influent nitrate concentrations in both columns were approximately 33 mg/ L 共as NO−3 -N兲. Initial rate of effluent extraction from the columns corresponded to a retention time of 1.3 days. The rate of effluent extraction was progressively decreased such that the retention time of water in the columns increased in steps to a final value of 13 days.


J. Environ. Eng. 2005.131:1212-1220.


Fig. 8. Abiotic nitrate reduction in flow-through reactors containing sand and various amounts of steel wool: by-product evolution 共volume of sand: 125 cm3兲

The reactors were maintained at this retention time for a period of 12 days starting from the 24th day after the start of effluent extraction. Effluent nitrate and ammonia concentrations from both columns were measured every few days. However, effluent nitrite concentrations could not be measured due the increased chloride concentration in the effluent, resulting from addition of HgCl2. This resulted in the nitrite peak in the ion chromatograph being swamped by the chloride peak. The results of the two experiments described above are shown in Fig. 8. In the case of the column containing only sand 共Fig. 8兲, ammonia concentration in the effluent was found to be very low at all retention times, including the final retention time of 13 days. This was expected because no nitrate reduction or nitrite/ammonia formation mechanisms existed in this column. In case of the column containing both sand and steel wool, ammonia concentration in the effluent was much higher than in the previous column at all retention times, including the final retention time of 13 days. This was consistent with the abiotic nitrate reduction and consequent ammonia formation in presence of metallic iron, as demonstrated earlier 共see Figs. 5 and 6兲. Three additional experiments were conducted using columns seeded with autotrophic denitrifying microorganisms. The steel wool concentration in these columns was 0.5, 1.0, and 1.5 g. In addition to effluent nitrate and ammonia concentrations, nitrite concentrations also were measured in the effluents from each column; the nitrite concentration was found to be negligibly small in most cases. Total nitrogen balance for the effluents from the columns was performed to determine the extent of biological denitrification. The results of the three experiments described above are shown in Fig. 9. Based on these results, it appears that the retention time of 13 days in the reactive media was sufficient for reduction of nitrate to achieve low effluent values in all three cases studied. However, in all three cases, the effluent contained substantial amounts of ammonia, an undesirable by-product of abiotic nitrate reduction. The fraction of initial nitrate converted

Fig. 9. Abiotic and biological nitrate reduction in flow-through reactors containing sand and various amounts of steel wool: by-product evolution 共volume of sand: 125 cm3兲

to ammonia by abiotic nitrate reduction increased with increasing iron concentration in the reactive media. Consequently it appears that to minimize ammonia production, the steel wool concentration of the reactive media has to be lowered even further. Reduction of the steel-wool concentration as suggested above would, however, also reduce the rate of hydrogen production, which is undesirable from the biological denitrification perspective. It is entirely possible that lower metallic iron concentrations in the reactive media may result in increased nitrate concentration in the effluent in lieu of the decline in ammonia concentration. Preliminary indications of this eventuality can be obtained through careful observation of the effluent nitrate data for the three cases presented in Fig. 9. In the case where metallic iron concentration was lowest 共0.5 g兲, the steady-state effluent nitrate concentration corresponding to the retention time of 13 days was seen to be higher than that in cases where metallic iron concentration was 1.0 and 1.5 g. Increasing the retention time can of course reduce higher nitrate concentration in the effluent. However, increasing the retention time also increases the possibility of producing more ammonia in lieu of reducing the nitrate concentration.

Summary and Conclusions Our study shows that passing nitrate-contaminated water through reactive porous media comprising of 125 cm3 of sand containing 0.5, 1.0, or 1.5 g of steel wool and seeded with hydrogenotrophic denitrifying microorganism was sufficient to reduce nitrate


J. Environ. Eng. 2005.131:1212-1220.


concentrations from 40 mg/ L 共as NO−3 -N兲 to less than 2 mg/ L 共as NO−3 -N兲 with a retention time of 13 days. However, in all three cases, the effluent contained substantial amounts of ammonia, which was an undesirable by-product of abiotic nitrate reduction. It was postulated that further decrease in the steel wool concentration in the reactor media and/or increase in retention time may further reduce ammonia concentrations, while promoting biological denitrification in favor of abiotic nitrate reduction. Determination of the optimal composition of the porous media for denitrification thus remains an unfinished task. The hydrogenotrophic denitrification process described above may find application in the design and construction of in-situ reactive barrier systems for removal of nitrate from groundwater. For example, considering the groundwater flow rate to be in the order of 10–40 cm/ day, an approximate thickness of a barrier of this type is likely to be 260–1,040 cm 共assuming porosity of barrier material to be 0.5兲 for a retention time of 13 days. Future research may include extension of the work presented in this study to further optimize the composition of the porous media to obtain a more favorable end-product distribution. Another important area of research is the problem of long-term pH increase in porous media. This occurs due to the release of OH− ions during the biological denitrification process, which may raise the pH of the porous media to values beyond the limit for sustenance of denitrifying microorganisms. In this study, the systems chosen were buffered to prevent precipitous pH increase. However, in practical systems, sufficient buffering may not be available. In such cases, the inherent buffering capacity of the porous media must be increased through the amendment of the media with suitable additives. Another possible cause of concern is the release of ferrous ions as a consequence of the biological denitrification process. However, measurement of iron concentration in the effluent from the reactive media during this research 共data not shown兲 indicated that dissolved iron concentration is rarely more that 2 mg/ L. Dissolved iron released due to corrosion of metallic iron may have precipitated as amorphous Fe共OH兲2 inside the porous media under the prevalent reducing conditions 共Kamolpornwijit et al. 2003兲. Such precipitation and progressive growth of microbial biomass inside the reactive media could result in reduction of porosity of the reactive media, thus affecting the long-term efficiency of such media in regard to denitrification capacity.

Acknowledgment The writers gratefully acknowledge the financial help provided by the Department of Science and Technology of the Government of India through Project No. III.5共119兲2001-SERC Engg. in carrying out this study.

References Alowitz, M. J., and Scherer, M. M. 共2002兲. “Kinetics of nitrate, nitrite, and Cr共VI兲 reduction by iron metal.” Environ. Sci. Technol., 36共3兲, 299–306. American Public Health Association/Water Environment Federation/ American Water Works Association 共APHA/WEF/AWWA兲. 共1985兲. Standard methods for examination of water and wastewater, 16th ed., APHA, Washington D.C. American Public Health Association/Water Environment Federation/ American Water Works Association 共APHA/WEF/AWWA兲. 共1995兲.

Standard methods for examination of water and wastewater, 19th ed., APHA, Washington D.C. Blowes, D. W., Ptacek, C. J., Benner, S. G., McRae, C. W. T., Bennett, T. A., and Puls, R. W. 共2000兲. “Treatment of inorganic contaminants using permeable reactive barriers.” J. Contam. Hydrol., 45共1/2兲, 123– 137. Chang, C. C., Tseng, S. K., and Huang, H. K. 共1999兲. “Hydrogenotrophic denitrification with immobilized Alcaligenes eutrophus for drinking water treatment.” Bioresour. Technol., 69共1兲, 53–58. Chi, I., Zhang, S. T., Lu, X., Dong, L. H., and Yao, S. L. 共2004兲. “Chemical reduction of nitrate by metallic iron.” J. Water Supply: Res. Technol.-AQUA, 53共1兲, 37–41. Choe, S., Chang, Y. Y., Hwang, K. Y., and Khim, J. 共2000兲. “Kinetics of reductive denitrification by nanoscale zero-valent iron.” Chemosphere, 41共8兲, 1307–1311. Choe, S. H., Ljestrand, H. M., and Khim, J. 共2004兲. “Nitrate reduction by zero-valent iron under different pH regimes.” Appl. Geochem., 19共3兲, 335–342. Claus, G., and Kutzner, H. J. 共1985a兲. “Autotrophic denitrification by thiobacillus denitrificans.” Appl. Microbiol. Biotechnol., 22共2兲, 289– 296. Claus, G., and Kutzner, H. J. 共1985b兲. “Physiology and kinetics of autotrophic denitrification by Thiobacillus denitrificans.” Appl. Microbiol. Biotechnol., 22共2兲, 283–288. Devlin, J. F., Eedy, R., and Butler B. J. 共2000兲. “The effects of electron donor and granular iron on nitrate transformation rates in sediments from a municipal water supply aquifer.” J. Contam. Hydrol., 46共1/2兲, 81–97. Dries, D., Liessens, J., Verstraere, W., Stevens, P., de Vost, P. and de Lay J. 共1988兲. “Nitrate removal from drinking water by means of hydrogenotrophic denitrifiers in a polyeurethane carrier reactor.” Water Supply, 6共3兲, 181–192. Ergas, S. J., and Reuss, A. 共2001兲. “Hydrogenotrophic denitrification of drinking water using a hollow fiber membrane bioreactor.” J. Water Supply: Res. Technol.-AQUA, 50共3兲, 161–171. Gamble, T. N., Betlach, M. R., and Tiedje, J. M. 共1976兲. “Numerically dominant denitrifying bacteria from world soils.” Appl. Environ. Microbiol., 33共4兲, 926–939. Gayle, B. P., Boardman, G. D., Sherrard, J. H., and Benoit, R. E. 共1989兲. “Biological denitrification of water.” J. Environ. Eng., 115共5兲, 930– 943. Ginner, J. L., Alvarez, P. J. J., Smith, S. L., and Scherer, M. M. 共2004兲. “Nitrate and nitrite reduction by Fe-0: Influence of mass transport, temperature, and denitrifying microbes.” Environ. Eng. Sci., 21共2兲, 219–229. Gros, H., and Treutler, K. 共1986兲. “Biological denitrification process with hydrogen-oxidizing bacteria for drinking water treatment.” J. Water Supply: Res. Technol.-AQUA, 5, 288–290. Gros, H., Schnoor, G., and Ruthen, P. 共1988兲. “Biological denitrification process with hydrogenoxidizing bacteria for drinking water treatment.” Water Supply, 6, 193–198. Gros, H., Schnoor, G., and Treutler, K. 共1986兲. “Nitrate removal from groundwater by autotrophic microorganisms.” Water Supply, 4共1兲, 11–21. Hu, H. Y., Goto, N., and Fujie, K. 共2001兲. “Effect of pH on the reduction of nitrite in water by metallic iron.” Water Res., 35共11兲, 2789–2793. Huang, C. P., Wang, H. W., and Chiu, P. C. 共1998兲 “Nitrate reduction by metallic iron.” Water Res., 32共8兲, 2257–2264. Huang, Y. H., and Zhang, T. C. 共2002兲. “Kinetics of nitrate reduction by iron at near neutral pH. J. Environ. Eng., 128共7兲, 604–611. Huang, Y. H., and Zhang, T. C. 共2004兲. “Effects of low pH on nitrate reduction by iron powder.” Water Res., 38共11兲, 2631–2642. Huang, Y. H., Zhang, T. C., Shea, P. J., and Comfort, S. D. 共2003兲. “Effects of oxide coating and selected cations on nitrate reduction by iron metal.” J. Environ. Qual., 32共4兲, 1306–1315. Kamolpornwijit, W., Liang, L., West, O. R., Moline, G. R., and Sullivan, A. B. 共2003兲. “Preferential flow path development and its influence on long-term PRB performance: Column study.” J. Contam. Hydrol., 66共3/4兲, 161–178.


J. Environ. Eng. 2005.131:1212-1220.


Kielemoes, J., De Boewer, P., and Verstraete, W. 共2000兲. “Influence of denitrification on the corrosion of iron and stainless steel powder.” Environ. Sci. Technol., 34共4兲, 663–671. Kleerebezem, R., and Mendezà, R. 共2002兲. “Autotrophic dentrification for combined hydrogen sulfide removal from biogas and postdentrification.” Water Sci. Technol. 45共10兲, 349–356. Kurt, M., Dunn, I. J., and Bourne, J. R. 共1987兲. “Biological denitrification of drinking water using autotrophic organisms with H2 in a fluidizedbed biofilm reactor.” Biotechnol. Bioeng., 29共4兲, 493–501. Lee, K. C., and Rittmann, B. E. 共2002兲. “Applying a novel autohudrogenotrophic hollow-fiber membrane biofilm reactor for denitrification of drinking water.” Water Res., 36共8兲, 2040–2052. Liessens, J., Vanbrabant, J., de Vos, P., Kersters, K., and Verstraete, W. 共1992兲. “Mixed culture hydrogenotrophic denitrification of drinking

water.” Microb. Ecol., 24共3兲, 271–290. Murphy, A. 共1991兲. “Chemical removal of nitrate from water.” Nature, 350共6315兲, 223–225. Smith, R. L., Ceazan, M. L., and Brooks, M. H. 共1994兲. “Autotrophic, hydrogen-oxidizing, denitrifying bacteria in groundwater: Potential agents for bioremediation of nitrate contamination.” Appl. Environ. Microbiol., 60共6兲, 1949–1955. Su, C. M., and Puls, R. W. 共2004兲. “Nitrate reduction by zerovalent iron: Effects of formate, oxalate, citrate, chloride, sulfate, borate, and phosphate.” Environ. Sci. Technol., 38共9兲, 2715–2720. Till, B. A., Weathers, L. J., and Alvarez, P. J. 共1998兲. “Fe共0兲-supported autotrophic denitrification.” Environ. Sci. Technol., 32共5兲, 634–639. Westerhoff, P., and James, J. 共2003兲. “Nitrate removal in zero-valent iron packed columns.” Water Res., 37共8兲, 1818–1830.


J. Environ. Eng. 2005.131:1212-1220.