Electronic Supplementary Material a Paper Being Published in Water Research

[Electronic Supplementary Material ]

Table S1Nucleotide sequence accession numbers of uncultivated bacteria from this study.

ID / Accession No. / ID / Accession No. / ID / Accession No.
MBfR_SS01 / EU923991 / MBfR_SS40 / EU924016 / MBfR_GSL102 / EU924041
MBfR_SS02 / EU923992 / MBfR_SS42 / EU924017 / MBfR_GSL103 / EU924042
MBfR_SS03 / EU923993 / MBfR_SS43 / EU924018 / MBfR_GSL110 / EU924043
MBfR_SS04 / EU923994 / MBfR_SS44 / EU924019 / MBfR_GSL112 / EU924044
MBfR_SS05 / EU923995 / MBfR_SS45 / EU924020 / MBfR_GSL113 / EU924045
MBfR_SS06 / EU923996 / MBfR_SS46 / EU924021 / MBfR_GSL115 / EU924046
MBfR_SS07 / EU923997 / MBfR_SS48 / EU924022 / MBfR_GSL117 / EU924047
MBfR_SS08 / EU923998 / MBfR_SS52 / EU924023 / MBfR_GSL118 / EU924048
MBfR_SS09 / EU923999 / MBfR_SS53 / EU924024 / MBfR_GSL119 / EU924049
MBfR_SS10 / EU924000 / MBfR_SS55 / EU924025 / MBfR_GSL120 / EU924050
MBfR_SS11 / EU924001 / MBfR_GSL02 / EU924026 / MBfR_FP01 / EU924051
MBfR_SS12 / EU924002 / MBfR_GSL04 / EU924027 / MBfR_FP07 / EU924052
MBfR_SS14 / EU924003 / MBfR_GSL06 / EU924028 / MBfR_FP08 / EU924053
MBfR_SS16 / EU924004 / MBfR_GSL07 / EU924029 / MBfR_FP15 / EU924054
MBfR_SS19 / EU924005 / MBfR_GSL10 / EU924030 / MBfR_FP17 / EU924055
MBfR_SS20 / EU924006 / MBfR_GSL14 / EU924031 / MBfR_FP23 / EU924056
MBfR_SS21 / EU924007 / MBfR_GSL16 / EU924032 / MBfR_FP25 / EU924057
MBfR_SS23 / EU924008 / MBfR_GSL34 / EU924033 / MBfR_FP38 / EU924058
MBfR_SS24 / EU924009 / MBfR_GSL43 / EU924034 / MBfR_FP40 / EU924059
MBfR_SS27 / EU924010 / MBfR_GSL44 / EU924035 / MBfR_FP43 / EU924060
MBfR_SS28 / EU924011 / MBfR_GSL63 / EU924036 / MBfR_FP47 / EU924061
MBfR_SS32 / EU924012 / MBfR_GSL64 / EU924037 / MBfR_FP48 / EU924062
MBfR_SS33 / EU924013 / MBfR_GSL66 / EU924038 / MBfR_FP52 / EU924063
MBfR_SS35 / EU924014 / MBfR_GSL68 / EU924039 / MBfR_FP53 / EU924064
MBfR_SS36 / EU924015 / MBfR_GSL82 / EU924040 / MBfR_FP105 / EU924065

[Electronic Supplementary Material– A paper being published in Water Research]

Kinetics of nitrate and perchlorate reduction in ion exchange brine using the membrane biofilm reactor (MBfR)

Steven W. Van Ginkel*a,b, Chang Hoon Ahnb, Mohammad Badruzzamanc, Deborah J. Robertsd, S. S. Geno Lehmanc, Samer S. Adhamc, Bruce E. Rittmannb

Corresponding author, Phone: 480-406-9877, Fax 480-727-0889, Email address:

aNational Risk Management Research Laboratory, Environmental Protection Agency, Cincinnati, OH45268

bCenter for Environmental Biotechnology, Biodesign Institute at Arizona State University, P.O. Box 875701, Tempe, AZ 85287-5701

cApplied Research Department, MWH Americas, Inc., 618 Michillinda Ave, Suite 200, Arcadia, CA91007

d School of Engineering, University of British Columbia, Okanagan, 3333 University Way, Kelowna, BC, Canada, V1V 1V7.

Abstract

Several sources of bacterial inocula were tested for their ability to reduce nitrate and perchlorate in synthetic ion-exchange spent brine (30-45 g/L) using a hydrogen-based membrane biofilm reactor (MBfR). Nitrate and perchlorate removal fluxes reached as high as 5.4 g N m-2d-1 and 5.0 g ClO4 m-2d-1, respectively, and these values are similar to values obtained with freshwater MBfRs. Nitrate and perchlorate removal fluxes decreased with increasing salinity. The nitrate fluxes were roughly first order in H2 pressure, but roughly zero-order with nitrate concentration. Perchlorate reduction rates were higher with lower nitrate loadings, compared to high nitrate loadings; this is a sign of competition for H2. Nitrate and perchlorate reduction rates depended strongly on the inoculum. An inoculum that was well acclimated (years) to nitrate and perchlorate gave markedly faster removal kinetics than cultures that were acclimated for only a few months. These results underscore that the most successful MBfR bio-reduction of nitrate and perchlorate in ion-exchange brine demands a well acclimated inoculum and sufficient hydrogen availability.

Key Words: nitrate, perchlorate, H2-based membrane biofilm reactor (MBfR), ion-exchange brine

Introduction

Perchlorate, primarily used in ordnance, explosives, and as a rocket propellant in the defense and aerospace industries, has contaminated the water supplies of more than 20 million people, primarily in the southwest United States (U.S. EPA, 2008). Perchlorate is manufactured in 44 states, and perchlorate contamination is known in 35 states (U.S. EPA, 2008). Because of its high solubility, non reactivity, and poor absorption to soils, perchlorate persists and spreads rapidly throughout the environment (Logan, 2001a). Because ammonium perchlorate absorbs water from the air, ammunitions and rockets have a limited shelf life and the ammonium perchlorate must occasionally be replaced. As of 2006, an expected 165 million pounds of perchlorate will be disposed and will require treatment (Wallace et al., 1998).

Perchlorate, a known competitive inhibitor of the sodium/iodide symporter (NIS) (De Groef et al. 2006), interferes with iodide uptake into the thyroid gland. In fetuses and children, perchlorate may interfere with proper development (U.S. EPA, 2008). Risks posed by perchlorate must be considered in conjunction with iodine intake, since a healthy diet consisting of iodine counteracts any negative effects induced by perchlorate (Kirk, 2006). The U.S. Environmental Protection Agency (EPA) set a preliminary remediation goal (PRG) for perchlorate of 24.5 µg/L to prevent adverse effects of exposure to unborn children. This goal assumes that exposure comes solely from drinking water and does not include exposures from food (i.e., mother’s milk, lettuce, etc.). However, Ginsberg et al. (2007) states that the PRG can lead to a 7-fold increase in breast milk concentration, causing 90% of nursing infants to exceed the reference dose (RfD) of 0.0007 milligrams of perchlorate per kilogram of body weight per day. Ting et al. (2006) state that 6 µg/L is adequately protective of sensitive subpopulations, including pregnant women, their fetuses, infants, and people with hypothyroidism, and the California Department of Public Health as of October 18, 2007 set 6 µg/L as a maximum contaminant level. Massachusetts promulgated a perchlorate MCL of 2 ug/L in 2006 (MADEP, 2006).

Nitrate is often a co-contaminant with perchlorate, since nitrogen is a major element in the manufacture of explosives, which primarily utilizes ammonium nitrate and diesel fuel (ITRC, 2000). In affected groundwater, the nitrate concentration is usually 2 to 5 orders of magnitude higher than the perchlorate concentration (Gu et al. 2003). Nitrate also has been shown to inhibit iodine uptake in the thyroid gland and accounts for a much larger proportion of iodine uptake inhibition than perchlorate, since at the legally accepted maximal nitrate levels in drinking water, nitrate far exceeds the potential effect of the proposed RfD for perchlorate (De Groef et al. 2006).

The most commonly used process today for removing perchlorate and nitrate from groundwater is ion-exchange (IX), in which NO3- and ClO4- are exchanged for harmless anions as the water passes through a bed of ion-exchange resin (Aldridge et al. 2004)). Although IX is efficient at removing nitrate and perchlorate from the water, it does not detoxify them. Instead, they accumulate on the IX resin or in the salt-brine used to regenerate the resin. Treatment of the IX brine so that it can be reused or disposed of safely is a major unresolved challenge. One approach for treating the brine is bioreduction.

Several studies have focused on nitrate and perchlorate bioreduction at high salinities typical of IX brine using an organic electron donor. Cang et al. (2004) and Okeke et al. (2002) observed rapid perchlorate reduction at 40 – 60 g/L NaCl when using acetate as the electron donor. Logan et al. (2001b) observed growth on acetate and perchlorate at salinities from 10 to 150 g/L NaCl. Chung et al. (2007) appears to be the only study that has investigatedbioreduction of nitrate andperchlorate at high salinities using hydrogen gas (H2) as the electron donor. This situation is called autohydrogenotrophic bioreduction, because the bacteria are autotrophs that utilize inorganic carbon as the carbon source. Chung et al. (2007) demonstrated autotrophic bioreductions for the entire range of salinities, but the kinetics slowed noticeably as the salinity was increased from 20 to 40 g/L NaCl.

The focus of the present study was to expand upon the work of Chung et al. (2007) by performing a much more comprehensive investigation of the factors that control nitrate and perchlorate bioreduction in ion-exchange brine. These factors include acclimation period, salinity, H2 availability, nitrate concentration, and inoculum source.

Working as an inter-institutional research team, the National Risk Management Research Laboratory’s Treatment Technology Evaluation Branch (TTEB) of the Environmental Protection Agency (EPA), the Center for Environmental Biotechnology at Arizona State University (ASU), and Montgomery Watson Harza (MWH) independently evaluated the H2-based MBfR (Membrane Biofilm Reactor) to remove nitrate and perchlorate from IX spent brine (30 to 45 g/L NaCl). MBfRs consist of a bundle of fibers that are encased in a continuous-flow reactor vessel. H2 gas is supplied to the inside of the fibers and diffuses to an autotrophic biofilm that grows on the outside of the fiber wall. The biofilm oxidizes H2 while reducing perchlorate and nitrate in the contaminated water.

The MBfR technology for brine treatment offers a potentially effective brine-treatment method, because it can remove the anions from the brine, thereby making it possible to reuse the brine or avoid problems of brine disposal. Since the MBfR technology treats the brine, a side stream, and not the water directly, the chance of bacteria being released into the distribution system is very low. Since this technology uses a biofilm, the bacteria largely remain attached to the membrane rather than living in a planktonic state and being carried with the treated brine back to the IX column. Besides, any bacteria that would be carried over in the treated brine would be subject to severe osmotic down shock in the IX column. Should the bacteria be released in the IX effluent, they are easilyremoved by usual post treatment systems, such as deep-bed or membrane filtration. The hydrogen gas, being insoluble especially in brine (1.3 mg H2/L at 1 atm of H2, 45 g/L NaCl), will not remain in the treated brine and will not cause bacterial growth on the IX resin or lead to the production of disinfection by-products (Weisenberger and Schumpe, 1996).

The MBfR provides an efficient and simple means to deliver H2 to a contaminant-reducing biofilm. Directly supplying H2 via membranes is becoming a proven technology and is more efficient than using an organic donor, since the biofilm uses only the stoichiometric amount of hydrogen it needs to reduce the contaminants in the feed stream; this leads to reduced costs and precludes leakage of oxygen demand to the effluent. For these reasons, the MBfR is being researched and field tested for a range of bioreduction applications when the water can be routed through an ex situ bioreactor, as is the case with IX brines.

Materials & Methods

Media

At the EPA, ammonium perchlorate and ammonium nitrate served as the sole electron acceptors, and ammonium bicarbonate (1.32 g/L) served as the sole carbon source for bacterial growth in the MBfRs. At ASU, sodium perchlorate, sodium nitrate, and sodium bicarbonate (0.5 g/L) were used. At MWH, ammonium perchlorate, ammonium nitrate, and sodium bicarbonate (0.2 g/L) were used. Hydrogen gas was the sole electron donor in all cases, but it was not present in the medium, since it was delivered by diffusion through the MBfR membrane. At the EPA, the salinity in the MBfR media was 1, 30, and 45 g/L NaCl. At ASU and MWH, the salinity was 30 g/L NaCl. At all three locations, MgCl2 • 6H2O was added at a Mg:Na ratio of 0.11:1, since magnesium has been shown to increase osmotolerance (Lin et al. 2006).

At the EPA, macronutrients (g/L) used include KH2PO4 (0.56), CaCl2 • 2H2O (0.1), and MnCl2 • 2H2O (0.6). Trace nutrients (mg/L) include FeCl3 • 6H2O (97), CoCl2 • 6H2O (82), ZnCl2 (47), Al2 (SO4)3 • 18 H2O (26), NiCl2 • 6H2O (20), CuSO4 • 5H2O (10), H3BO3 (10), Na2MoO4 • 2H2O (10), Na2WO4 • 2H2O (10), and Na2SeO3 • 5 H2O (2). EDTA was also added to the trace nutrient solution at 0.5 g/L. The effect of adding yeast extract (0.5 g/L) was tested.

At ASU,macronutrients (g/L) used include KH2PO4 (0.13) and NaHPO4 (0.43). Trace nutrients (mg/L) include CaCl2 • 2H2O (20), MnCl2 • 2H2O (400), FeSO4 • 7H2O (20), CoCl2 • 6H2O (200), ZnSO4 • 7H2O (100), NiCl2 • 6H2O (10), CuCl2 • 2H2O (10), H3BO3 (300), Na2MoO4 • 2H2O (30), and Na2SeO3 • 5 H2O (10).

At MWH, 1 g/L sodium acetate was added during startup, but removed for normal operation. Macronutrients (g/L) used include KH2PO4 (0.05), CaCl2 • 2H2O (1.4), and KCl (0.72). The following mineral solution (g/L) was added at 1 mL/L media: MnCl2 • 2H2O (0.03), FeCl2 • 4H2O (1.5), CoCl2 • 6H2O (10), ZnCl2 (0.05), NiCl2 • 6H2O (0.03), H3BO3 (0.3), and (NH4)6Mo7O24 • 4H2O (10).

Analytical Methods

At the EPA, automated colorimetry (EPA Method 353.2) was used for the nitrate and nitrite measurements, and ion chromatography (EPA Method 314.0) was used for the perchlorate measurements. Nitrate was determined by automated cadmium reduction colorimetry using a SmartChem Autoanalyzer (Westco Scientific, Danbury CT). Perchlorate was measured using an ion chromatograph (IC) (DX2500; Dionex) equipped with an AS16 column and guard column, a CD20 Conductivity Detector, a self-regeneratingsuppressor, and an autosampler (Dionex Corporation, Sunnyvale, California). At ASU and MWH, all anions were measured using ion chromatography as described above, while EPA Method 300.0 was used for the nitrate and nitrite measurements.

MBfR Inoculation

Inocula were sediments from three saline sources - Freeport, TX, the Great Salt Lake, and the Salton Sea; and one freshwater source – backwash sludge from a perchlorate degrading packed bed reactor at the EPA adapted to acetate (‘EPA’). The MBfRs at each location were inoculated differently as described in the following sections. After startup, the MBfRs were operated in batch mode for several days, during which time the medium was circulated around the membrane fibers to allow biofilm colonization. Subsequently, feed pumps were turned on to initiate continuous feeding. The systems behaved as a completely mixed biofilm reactors because of the high recirculation ratio (150:1), which also provided a high-flow velocity that helped maintain a consistent biomass thickness on the hollow fibers. Schematics and physical characteristics of the MBfRs can be found in Chung et al. (2006 a,b,c), Chung et al. (2007), and Nerenberg and Rittmann (2004). Briefly, the main membrane module contained a bundle of 32 hydrophobic hollow-fiber membranes (Model MHF 200TL, Mitsubishi Rayon) inside a glass shell. At EPA, the volume encasing the MBfR fibers was approximately 13 mL while at ASU the volume encasing the MBfR fibers was approximately 11.7 mL. The total volume including a recirculation loop was approximately 30 mL. At MWH, a large recirculation loop was used, and the total volume was 160 mL. The fiber surface area at ASU and EPA was the same at 72.6 cm2. The fiber surface area of MWH’s MBfRs was 78.6 cm2.

EPA MBfRs

At the EPA, all four inocula were acclimated in batch over several weeks to 1, 30, or 45 g/L NaCl and a 100% H2 headspace. These cultures were then injected into the MBfRs using a sterile plastic syringe. Four MBfRs were operated at 45 g/L NaCl using each inoculum (Table 1). The first four conditions in Table 1 had lower nitrate concentrations, since high nitrate concentrations may inhibit perchlorate reduction. The last three conditions in Table 1 had higher nitrate concentrations and longer hydraulic retention times, which are relevant to improving the cost effectiveness of the MBfR design.

In addition to the four MBfRs operated at 45 g/L NaCl salinity, two MBfRs were operated at 30 g/L NaCl using Freeport batch cultures adapted to 30 g/L NaCl (Table 2). Initially, ‘F-30,1’ was fed only nitrate, while ‘F-30, 2’ was fed nitrate and perchlorate. At the loading rates of 0.37 and 0.43 g N m-2d-1, nitrate, perchlorate, and yeast extract (0.5 g/L) were fed to F-30,1.

Two additional MBfRs were operated to serve as ‘controls’ for the other MBfRs. The first control MBfR was inoculated with a known salt-tolerant denitrifier, Bacillus halodenitrificans (ATCC 49067), operated at 45 g/L NaCland designated Bh 45 g/L. Bh 45 g/L was given the same media as the other MBfRs at 45 g/L NaCl, expect acetate (1 g/L) was added to the media for one month before H2 became the sole electron donor in the MBfR.

The second control MBfR was inoculated with the EPA culture, operated at 1 g/L NaCl and designated EPA 1 g/L. For startup, backwash sludge from the packed bed reactor at the EPA was directly injected into the MBfR. The membrane fibers of EPA 1 g/L were citric acid washed before the nitrate loading rate of 1.75 g N m-2d-1 to observe whether or not the fiber pores were fouled by precipitates which would reduce hydrogen transfer to the biofilm.

ASU and MWH MBfRs

At ASU, small amounts of the three different saline sediments were directly injected into five separate MBfRs, and triplicate MBfRs were operated using the Freeport sediment (Table 3). The inoculum for the MWH MBfR was obtained by taking a 100 mL subsample of a culture that had been adapted to growth on acetate, nitrate, perchlorate, and 30 g/L NaCl for over 4 years as described by Cang et al. (2004), Hiremath et al. (2006), and Lin et al. (2007) and feeding it hydrogen under batch conditions at the University of Houston. Once the new culture had been shown to degrade perchlorate and hydrogen simultaneously over three consecutive feedings it was sent to MWH for use in MBfRs. A total of three different nitrate and perchlorate loading rates were tested at MWH (Table 4).

Results Discussion

Tables 1 – 4 present the nitrate removal fluxes for all experiments. They were calculated using the tabulated data according to the following equation:

J = (So-S)Q/A

where So and S are the influent and effluent nitrate concentrations (g/L), Q is the volumetric flow rate through the main membrane module (L/hr), and A is the membrane surface area (m2). The flux of the rate-limiting substrate is the most fundamental kinetic parameter for a biofilm process, as it integrates substrate utilization and mass transport (Rittmann and McCarty, 2001). In addition, normalizing the removal rate [(So-S)Q] to the membrane surface area (A) makes it possible to compare directly the performance of MBfRs according to the amount of membrane surface area, which is of practical and economic importance.

The nitrate removal fluxes in this study compare fairly well with other autohydrogenotrophic studies using freshwater. H2-based MBfR nitrate removal fluxes have ranged from 1.0 to 5.4 g N m-2d-1 with fresh water (Ergas and Reuss, 2001; Lee and Rittmann, 2000, 2002; Nerenberg et al. 2002; Ho et al. 2001). The nitrate flux observed by MWH at 30 g/L NaCl is at the high end of this range (4.0 g N m-2d-1). At ASU, the maximum removal flux observed at 30 g/L NaCl was 1.9 g N m-2d-1, while at the EPA, the maximum removal flux observed at 45 g/L NaCl was 0.38 g N m-2d-1. The highest nitrate removal flux observed by Chung et al. (2007) was 0.034 g N m-2d-1 at 45 g/L NaCl. In Chung et al. (2007), they started their MBfR at 10 g/L NaCl and then increased the salinity to 20 g/L NaCl after 80 days and then to 40 g/L NaCl after 100 days of incubation. This procedure contrasts with the present study, where the cultures were initially incubated in batch at 30 g/L NaCl, and the salinity was increased to 45 g/L NaCl over several weeks. The salinity was 45 g/L NaCl at the time of MBfR inoculation. Also, in the present study, the MBfRs were operated at 45 g/L NaCl for 329 days, compared to 50 days in Chung et al. (2007). Furthermore, magnesium was added in the present study, which may have increased the osmotolerance of the MBfR cultures (Lin et al. 2006).

The removal fluxes observed using the MWH MBfRs were two orders of magnitude higher than the removal fluxes observed for the ASU and the EPA MBfRs. At the time of inoculation the culture used to inoculate the MWH MBfRs was well acclimated to reducing perchlorate using acetate as the electron acceptor and this culture was able to switch to autohydrogenotrophic growth and continue to reduce perchlorate. Since the cultures used to inoculate the MWH MBfRs was initially obtained from the same sediment source as used for the ASU and the EPA inocula, the MWH data show that it is possible to have high perchlorate and nitrate fluxes with adaptation.

Effect of nitrate concentrations on nitrate and perchlorate removal

If the nitrate removal flux (J) is controlled by the effluent nitrate concentration (S) in a first-order manner, then the effluent normalized flux, i.e., J/S (m/d), must be constant throughout the range of effluent nitrate concentrations. On the other hand, if the effluent normalized flux declines with increasing concentration, the flux is closer to zero order in nitrate concentration, and the nitrate removal flux does not increase in proportion to an increase in the effluent (in-reactor) nitrate concentration (Chung et al. 2006 a,b). Figure 1 shows that, at 30 g/L NaCl, the nitrate removal flux never has a first-order relationship, but is much closer to zero order in nitrate concentration. The nitrate removal flux is limited either by active biomass or H2 availability. At 45 g/L NaCl (Figure 1 inset), the nitrate removal flux also approaches zero-order kinetics.