Nitrate reduction by fluoride green rust modified with copper

Available online at www.sciencedirect.com

Chemosphere 70 (2008) 1108–1116 www.elsevier.com/locate/chemosphere

Nitrate reduction by fluoride green rust modified with copper Jeongyun Choi a

a,*

, Bill Batchelor

b

Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, 373-1 Guseong-Dong, Yuseong-Gu, Daejeon 305-701, Republic of Korea b Department of Civil Engineering, Texas A&M University, College Station, TX 77843, USA Received 5 February 2007; received in revised form 23 July 2007; accepted 24 July 2007 Available online 14 September 2007

Abstract Nitrate reduction by the fluoride form of green rust modified with copper (GR-F(Cu)) was investigated using a batch reactor system. The extent of nitrate reduction was measured by measuring the increase in concentration of ammonia, which is the final product of nitrate reduction by GR. This approach was required, because nitrate could be removed from solution by ion exchange without reduction. The rate of ammonium production was investigated over the range of pH 7.8–11. The fastest reaction was achieved at pH 9 when GR was present at a concentration of 0.083 M as Fe(II) and 1 mM of Cu(II) was added. The rate at pH 9 was enhanced by a factor of 2.5 compared to that at pH 7.8 by comparing the time elapsed to transform all nitrate to ammonium. Kinetics of nitrate reduction by GR-F at pH 7.8 were affected by the concentration of Cu(II) added. The rate constants for ammonium production increased from 0.012 to 1.52 h1 as Cu(II) additions increased from 0 to 2.5 mM, but the reaction rate at 5 mM was slightly decreased to 1.25 h1. The mechanism of enhanced rates of nitrate reduction by addition of Cu(II) could not be fully determined in this study. However, XRD results showed that magnetite was produced in the reaction of Cu(II) and GR-F and SEM shows the production of nano-size particles which were not fully identified in this study. In addition, the concentration of Fe(II) in GR was observed to linearly decrease with concentration of Cu(II) added. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Nitrate; Nitrite; Green rust; Cu addition; Ammonium

1. Introduction Increasing nitrate concentration in groundwater is a considerable environmental problem, especially where groundwater is used as a source of drinking water. Excessive levels of nitrate can cause a serious illness called methemoglobinemia, which is also known as ‘‘blue baby syndrome’’, because it interferes with the oxygen-carrying capacity of the child’s blood (EPA, 1995). Additionally, several studies have reported that potentially carcinogenic N-nitroso compounds can be produced by reactions of nitrate with amines or amides (Canter, 1997). For these reasons, maximum contaminant levels have been set at

*

Corresponding author. Tel.: +82 42 869 3664; fax: +82 42 869 3610. E-mail address: [email protected] (J. Choi).

0045-6535/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.07.053

1 10 mg l1 for NO for NO 3 –N and 1.0 mg l 2 –N by the United States Environmental Protection Agency (EPA, 1995). The primary sources of nitrate in groundwater are the excessive usage of chemical fertilizers, soil nitrification, and animal manure. Waste disposal practices associated with land application of sludge, wastewater effluents, municipal or industrial landfills, and septic tank systems are considered as additional sources of nitrate contamination (Canter, 1997; Fanning, 2000; Haugen et al., 2002; Luk and Au-Yeung, 2002). To remove nitrate from groundwater, chemical reduction methods have been studied because they can degrade nitrate faster than biological methods and they are more cost-effective than many physico-chemical methods (Luk and Au-Yeung, 2002). Zero-valent iron (ZVI) has been studied as a chemical reductant to convert nitrate to ammonium and this reaction rate was strongly dependent

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

on the solution pH (Huang et al., 1998; Choe et al., 2000, 2004). In addition, green rusts (GRs) have been evaluated as chemical reductants for nitrate at neutral pH (Hansen et al., 2001). GR is a layered Fe(II)–Fe(III) hydroxide solid phase with an anion in the interlayer and the anion is often used to identify the type of GR. It was discovered as a transient corrosion product of iron pipe in the early 1900s and has been found to be an effective chemical reductant for organic and inorganic contaminants, including uranyl (O’Loughlin et al., 2003a), nitrate and nitrite (Hansen et al., 1994; Hansen and Koch, 1998; Hansen et al., 2001), selenate (Johnson and Bullen, 2003), chromate (Williams and Scherer, 2001; Legrand et al., 2004), and halogenated hydrocarbons (Lee and Batchelor, 2002; O’Loughlin et al., 2003c; O’Loughlin and Burris, 2004). Hansen and coworkers have studied nitrate reduction by GR-SO4 and GR-Cl. In their studies, nitrate was totally reduced to ammonium while GR was oxidized to a mixture of magnetite and Fe(II). The rate of reduction of nitrate by GR-Cl was faster than that by GR-SO4 (Hansen et al., 2001). One interesting aspect of GRs is that their reactivities can be enhanced when they are contacted with a transition metal such as Ag(I), Au(III), or Cu(II). The rate of dechlorination of several chlorinated hydrocarbons by GR with the addition of either Ag(I), Au(III), or Cu(II) was up to three orders of magnitude higher than for GR by itself (O’Loughlin et al., 2003c; O’Loughlin and Burris, 2004). Solids that contain Fe(II) have been observed to be good reductants for nitrate when they were combined with metallic catalysts. Ottley et al. (1997) combined Cu(II) with Fe(II)-containing solids and observed that the time required to achieve 15% reduction of nitrate was three orders of magnitude smaller than that achieved without Cu(II). The efficiency of nitrate reduction was improved by increasing the concentration of Cu(II). The results of these studies indicate that the addition of specific transition metals to GRs could enhance their reactivity for nitrate reduction as well as reductive dechlorination. A recent study conducted by the authors (Choi and Batchelor, 2004) found that the rates of nitrate reduction by GRs were affected by both type of GR (anion in the interlayer) and the metal used. They examined the reactivity for nitrate reduction of four different types of GR (GRCl, GR-SO4, GR-CO3, and GR-F) that were reacted with ten different trace metals (Ag, Ba, Co, Cu, Mn, Ni, Pb, Pt, Ti, and Zn). Fluoride green rust modified with Cu (GR-F(Cu)) had the greatest potential to treat groundwater polluted by nitrate, based on treatment efficiency and costs. The goals of this study are: (1) to characterize nitrate reduction by GR-F(Cu) and (2) to better understand the reaction mechanism of nitrate reduction. The effect of pH, Cu(II) concentration, and initial nitrate concentration on nitrate reduction by GR-F(Cu) was investigated.

1109

2. Experimental section 2.1. Anoxic condition GR-F was modified and characterized under anoxic conditions in this study, because it is known to react with oxygen. Anoxic conditions were obtained by using an anoxic chamber (Coy Laboratory Products Inc.) in which the atmosphere was filled with a mixed gas containing 5% hydrogen and 95% nitrogen. The oxidation state of the chamber was maintained at below 218 mV at pH 9 using palladium catalysts which removed oxygen in the reaction with hydrogen. The colorimetric indicator solution (resazurin, Aldrich) was used to monitor the condition of chamber. 2.2. Materials Ferrous chloride (tetrahydrate, 99%, Aldrich), sodium hydroxide (97.0% min EM), and sodium fluoride (99.8%, J.T. Baker) were used in the procedure for synthesizing GR-F. Copper chloride (dihydrate, 99+%, Aldrich) was used as activating agents to modify GR-F. Sodium nitrate (Certified ACS, Fisher) was used to provide nitrate as the target compound. Two biological buffers were used in the experiment for determining the effect of pH on nitrate by GR-F(Cu). CAPS (3-[cyclohexylamion]-1-propanesulfonic acid, pKa = 10.4, Sigma) and CHES (2-[N-cyclohexylamino]ethanesulfonic acid, pKa = 9.3, Sigma) were used as a buffer for pH 11 and pH 9, respectively. Concentrations of ammonia formed as the final product of nitrate reduction were measured by the phenate method (APHA, 1995) using the following chemicals: phenol (liquefied, 88% min, EM), ethanol (99.5%, Aldrich), sodium nitroferricyanide (III) (dehydrate, 99%, Sigma–Aldrich), sodium citrate (anhydrous, 99.9%, Sigma), and sodium hydrochlorite (6%, VWR). 2.3. Experimental procedures Fluoride green rust was synthesized using the partial air oxidation method developed by Refait et al. (1997). Fe(OH)2 solid was freshly prepared in an anoxic chamber by mixing the equal volumes of 0.12 M of FeCl2 and 0.2 M of NaOH, and then adding 0.12 M of NaF as a source of interlayer anions. The mixture of Fe(OH)2 and NaF was taken out of the chamber and oxidized by oxygen by allowing it to contact the air. During the oxidation of Fe(OH)2, the pH in the solution was monitored. The oxidation process was stopped when the pH reached a maximum (around pH 7.8) and began to drop. The synthesized GR-F was stored in the chamber to prevent further oxidation. In addition, all of the following experimental preparation steps were conducted in an anoxic chamber. Anion content in the interlayer of the GR was measured to confirm production of GR-F, because both chloride and fluoride were present in solution during the synthesis. The

1110

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

type of GR produced was determined by measuring the change in concentration of each anion before and after GR synthesis based on assuming that all of the anion removed from solution became part of GR. These analyses showed that about 95% of the green rust produced was GR-F (data was not shown). This result is consistent with the anion selectivity for layered double hydroxides (LDH) such as GR. This selectivity is generally reported as OH > F > Cl > Br > I (Hansen, 2001). GR-F was washed twice with deaerated deionized water (DDW) to reduce the concentrations of undesirable ions before use. The concentration of Fe(II) in these suspensions of GR-F was in the range from 0.0831 to 0.0886 M. The pH in washed GR-F suspension was around 7.8 and fairly constant, because GR-F itself acted as a buffer. In experiments for determining the effect of pH, DDW in the second washing step was replaced with 0.1 M buffer solutions (CHES for pH 9 and CAPS for pH 11), which were able to maintain a constant pH throughout the experiment. The modification of GR-F was achieved by contacting it with Cu(II) at 1 mM for about 1 h. A magnetic stirrer was used to provide homogenous conditions during the modification period. Experiments to evaluate the effect of Cu(II) concentration (0.1–5 mM) were conducted by adding appropriate amounts of 0.4 M of Cu(II) stock solution into 250 ml of GR suspension. The batch kinetic experiments were conducted using a 250 ml polypropylene bottle (Nalgene) in an anoxic chamber. The batch reactor was open to atmosphere of the chamber during the reaction period. The reaction was started by spiking an appropriate volume of 400 mM nitrate stock solution into 200 mM GR-F(Cu) suspension resulting in initial concentrations of nitrate equal to 0.05, 0.1, 0.5, 1, and 1.2 mM. During the reaction, the suspension of modified GR was continuously mixed using a magnetic stirrer. Samples of about 5 ml were taken at each sampling time and were filtered using 0.2 lm cellulose nitrate membrane filters. The degradation kinetics were determined by monitoring nitrate, nitrite and ammonium concentrations over time. 2.4. Analytical procedures The phenate method was used to measure the concentration of ammonium after filtration with 0.2 lm cellulose nitrate membrane filters (Whatman) and dilution (APHA, 1995). The colored compound produced by reaction of ammonium with phenol, sodium nitroferricyanide(III) and alkaline citrate was measured by a UV–VIS spectrophotometer (Hewlett Packard G1103A) at 640 nm with a light path of 1 cm. Iron concentrations in the total suspension (solid and liquid phases) and in the liquid phase were analyzed by the ferrozine method (Hewlett Packard G1103A UV–VIS spectrophotometer) after dilution with 1.2 N HCl (Gibbs, 1976). Fe(II) was measured directly and total iron mea-

sured after reduction of Fe(III) to Fe(II) by hydroxylamine. The concentration of Fe(III) was calculated by difference. Concentrations in the liquid phase were measured after filtering through 0.2-lm cellulose nitrate membrane filters. Concentrations in the total suspension were measured after treatment with 1.2 N HCl. Concentrations in the solid were determined by difference.  Concentrations of anions (Cl, F, NO 3 , and NO2 ) were measured by an ion chromatograph (Dionex DX80) equipped with AS14A column, DS5 detection stabilizer (a conductivity detector), and MMS III suppressor. The eluent was a mixture of 8 mM Na2CO3 and 1 mM NaHCO3 and the flow rate was 1 ml min1. The samples were taken and filtered with 0.2 lm cellulose nitrate membrane filters in an anoxic chamber. Sample volumes of 1 ml were manually injected into the column, immediately after dilution with DDW. XRD analyses for GR-F and GR-F(Cu) were performed using a RigaKu automated diffractometer using Cu Ka radiation. Solids were separated by centrifugation at 2960g for 5 min (Beckman, model J-6M centrifuge, JS7.5 rotor) and carefully air dried in an anoxic chamber for more than 1 d. Dry solids were ground and transferred into 20 ml glass vials with three layered closures which were used to transport the GR-F and were found to successfully prevent the oxidation of solids. The samples were scanned between 2.1° to 70° 2h with a scan speed of 1° 2h min1. SEM analyses were performed to observe the morphology of GR-F and GR-F(Cu) using FEI XL-30 FEG. A droplet of suspensions was carefully dried on an aluminum foil under anoxic condition for several hours and coated with osmium. 3. Results and discussion 3.1. Nitrate reduction by GR-F(Cu) Fig. 1 shows the results of an experiment on nitrate reduction by GR-F at pH 7.8. The initial concentration of nitrate (1 mM) rapidly decreased to 0.4 mM in 30 min, remained constant for 4 h, and then slowly decreased. The concentration of nitrate remaining after 9.4 h was around 0.06 mM. The concentration of ammonium, the expected final product of nitrate reduction, continuously increased and reached a value of 0.12 mM in 9.4 h reaction time. The reason for the rapid drop of nitrate is believed to be due to anion exchange of nitrate with F in the interlayer of GR. Nitrate has been reported to be able to ion exchange with GR (Hansen and Koch, 1998). Therefore, it can be assumed that two reactions, anion exchange reaction and redox reaction, occur in parallel when nitrate reacts with GR-F. This means that measuring the loss of nitrate from solution is not an appropriate method to determine the kinetics of nitrate reduction by GR-F, especially at early reaction times. Nitrite, which is a possible intermediate of nitrate reduction, was not observed during the reaction period, which might be due to anion exchange

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

nitrate reduction. All nitrates were completely transformed to ammonium within 1.2 h by GR-F(Cu) as shown in Fig. 3a, while only about 20% of ammonium was produced in 10 h by unmodified GR-F at pH 7.8. No measurable nitrite was observed. This stoichiometric conversion of nitrate to ammonium was also reported in nitrate reduction by GR-Cl and GR-SO4 (Hansen et al., 1996, 2001). The first order kinetic model proposed by Hansen et al. (2001) shown in Eq. (1) was applied to simulate ammonium production.

1.2

Nitrate Ammonium

0.6

0.4

dC NH4 ¼ k ðC NH4 ;max  C NH4 Þ; dt

0.2

0.0 0

2

6

4

10

8

Time (hr)

Fig. 1. Nitrate reduction by GR-F only at pH 7.8. Initial nitrate concentration was set to 1 mM and initial solid-phase Fe(II) in GR was 85 mM.

Fig. 2. Hypothetical reaction scheme of the nitrate reduction by GR-F.

reaction with GR-F. Therefore, ammonium concentrations were measured over time in this research to determine the kinetics of nitrate reduction by GR-F. Fig. 2 expresses the simplified reaction scheme that is believed to occur in the reaction of nitrate with GR-F. Fig. 3 shows results of nitrate reduction using GR-F modified with 1 mM Cu(II) at three different pH (pH 7.8, 9, and 11). These results demonstrate that the addition of Cu(II) to GR-F effectively increased the reaction rate of

1.2

1.2

0.6 NO3 + NH4 + NH4 model

Concentration (mM)

Concentration (mM)

0.8

0.2

(c) pH 11

1.0

1.0

0.4

1.2

(b) pH 9

(a) pH 7.8

1.0

0.8 0.6 NO3 NO2 + NH4 + NH4 model

0.4 0.2 0.0

0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

Time (h)

ð1Þ

where k is the rate constant for ammonium production, C NH4 ;max is the maximum concentration of ammonium produced by the reaction, which is equivalent to initial nitrate concentration, and C NH4 is the concentration of ammonium in solution. The value of the rate constants was obtained by conducting a nonlinear regression on aqueous phase ammonium concentrations using the Gauss-Newton method coded in the nlinfit function of MATLABÒ (MathWorks Inc.). The MATLAB function ‘nlparci’ was used to calculate 95% confidential levels of the rate constants. Predictions made using this model are shown as a solid line in Fig. 3. This model reasonably describes ammonium production by GR-F(Cu) at pH 7.8. The rate constant (k) that was obtained at pH 7.8 was 1.3(±0.2) and the R2 value for the fit of the kinetic model to ammonium concentrations was calculated to be 0.97. Fig. 3b and c show the results of nitrate reduction at pH 9 and 11, respectively. The rates of nitrate removal were identical at pH 9 and pH 11, with 95% removal within about 0.3 h, which was twice as fast as that observed at pH 7.8. Nitrite was detected as a by-product of nitrate reduction by GR-F at these pH values, but this was not observed at pH 7.8. The concentration of nitrite was 0.25 mM after 0.23 h at pH 9 and 0.4 mM after 0.3 h at pH 11. The ammonium production rate at pH 9 was faster

Concentration (mM)

Concentration (mM)

1.0

0.8

1111

0.8 0.6 NO3 NO2 + NH4 NH4+ model

0.4 0.2 0.0

0.0

0.1

0.2

0.3

Time (h)

0.4

0.5

0.6

0.0

0.2

0.4

0.6

0.8

Time (h)

Fig. 3. Results of kinetic experiments on nitrate reduction by GR-F(Cu) at (a) pH 7.8, (b) pH 9 and (c) pH 11. Initial nitrate concentration was 1 mM, Cu(II) addition was 1 mM, and initial solid-phase Fe(II) in GR was 87 mM. The solid lines represent predictions of ammonium production using the kinetic model (Eq. (1)). The dot lines are plotted (not fit) to guide the eye.

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

 þ NO 3 ! NO2 ! NO ! N2 O ! NH4

ð2Þ

If one of these steps happens slowly, the production of ammonium at early reaction times would be observed to be much lower even though the rate of removal of nitrate could be fast. One thing that would make nitrate/nitrite reduction slow at high pH is the surface charge of GR. GRs have a pH dependent charge on their surfaces as do other iron (oxy)hydroxide minerals. The surface dependent charge of GRs is positive at pH below the zero point charge (ZPC) and negative at the pH above ZPC (Lee and Batchelor, 2002). Based on the reaction mechanism for nitrate reduction proposed by Hansen and Koch (1998), nitrate was reduced on the surface of GR and the approach of nitrate/nitrite to GR occurred through electrostatic attraction. Therefore, nitrate or nitrite might be retarded in approaching GR surfaces at higher pH, since the surface of GR would be more negatively charged. This would result in slower nitrate and nitrite reduction as well as slower initial ammonium production at high pH. The total times elapsed in reducing 1 mM of nitrate to ammonium at pH 7.8, 9 and 11 were observed to be 1.5, 0.6 and 0.8 h, respectively. It means that the nitrate was reduced relatively faster at high pH. This might be related to the reactivity of GR itself. GR showed high reactivity at high pH during reductive dechlorination of tetrachloroethylene, resulting in faster reaction kinetics with higher pH (pH 6.8–10.1) (Lee and Batchelor, 2002). In addition, anion exchange of nitrate and nitrite with F in the inter-

layer of GR at higher pH might be less favorable compared to that at lower pH since nitrate and nitrite have to compete with hydroxide. Hydroxide was been reported to have the highest preference of any anion for staying in the interlayer of a LDH (Hansen, 2001). Therefore, it can be reasonably presumed that the tendency of nitrate to exist in solution rather than in the interlayer might be amplified with increasing pH. It means that the possibility of contacting nitrate/nitrite in solution with GR was increased at high pH, resulting in the faster transformation of nitrate to ammonium. This can also explain the absence of nitrite in solution at pH 7.8, but the presence of nitrite in solution at pH 9 and 11. In general, it has been reported that the reduction of nitrate by ZVI is less favorable at high pH, because of hydrogen ion consumption and precipitation of iron oxide on the surface of ZVI (Alowitz and Scherer, 2002). However, this explanation may not be applicable to this study, because it was observed that ammonium production rate increased at high pH and hydrogen ion was not consumed but produced.

3.2. Effect of Cu(II) concentration The effect on nitrate reduction of the concentration of Cu(II) added to GR suspensions was investigated over the range of 0–5 mM. Fig. 4 shows the effects of the amount of Cu(II) added on the ammonium production rate constant (k) and the solid-phase Fe(II) content in GRF(Cu). The Fe(II) content was measured right before nitrate reduction by GR-F(Cu) was initiated. The initial pH decreased slightly after Cu(II) was added to GR. The measured pH of GR-F(Cu) was 7.8, 7.7, 7.5, and 7.4 at 0, 1, 2.5, and 5 mM Cu(II), respectively. Rate constants for nitrate reduction increased with increasing Cu(II) additions in the range of 0 to 1 mM and the rate constant at 2.5 mM was almost the same as that at 1 mM. The rate

2.0

90 Rate constant Fe(II) concentration

1.8

85

1.6 1.4

80

1.2 1.0

75

0.8 0.6

70

0.4 0.2

65

0.0

Fe(II) concentration in GR solid (mM)

than that at pH 11. All nitrates were transformed to ammonium within 0.6 h at pH 9, while it took more than 0.8 h to produce 90% of the maximum ammonium at pH 11. Lower recovery of ammonium at the final sampling point in the case of pH 11 might be due to the loss of ammonia by the volatilization because most of ammonia (around 98%) would exist in the form of NH3(g) (the pKa of NHþ 4 was 9.3). The kinetic model was not able to describe the experimental data at pH 9 and 11 as well as it did for the data at pH 7.8. The predicted ammonium concentrations were somewhat higher than the measured concentrations at early reaction time and this difference was amplified with increasing pH. Ammonium production at pH 11 was less than 2% in 15 min, but 90% of ammonium was produced in the next 30 min. This early lag of ammonium production at high pH might be explained with the combination of the following two concepts; the nitrate reduction occurred through sequential step reaction and the early step of these sequential steps might be inhibited at higher pH. Several other studies also observed that nitrate was initially reduced to nitrite and finally transformed to ammonium (Siantar et al., 1996; Gao et al., 2003). In addition, NO and N2O were proposed as possible intermediates of nitrate reduction even though these compounds were not measured in this study (Fanning, 2000). Therefore, it can be presumed that nitrate reduction can take place by the following steps.

Rate constant (h-1)

1112

60 0

1

2

3

4

5

6

Cu(II) addition

Fig. 4. Dependence of ammonium production rate and Fe(II) content in GR solid with respect to Cu(II) addition. Initial nitrate concentration was 1 mM and no pH buffer was used. Error bars for rate constant represent 95% confidential intervals.

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

constant at 2.5 mM Cu(II) was 1.52 h1 which was almost 100 times higher than that observed without Cu(II) addition (0.012 h1). Although no previous studies have reported on the effect of Cu(II) on kinetics of nitrate reduction by GR, several studies have presented results showing that Cu(II) addition affects rates of reductive dechlorination. The addition of Cu(II) into GR-SO4 increased the rate of reduction of carbon tetrachloride by up to 2 orders of magnitude and increased the rate of reduction of 1,1,1trichloroethane by a factor of 500 (O’Loughlin et al., 2003c; O’Loughlin and Burris, 2004). In addition, Cu(II) was also effective in enhancing the rate of trichloroethylene reduction by GR-Cl (Maithreepala and Doong, 2005). XRD and SEM analyses were conducted in order to better understand the reaction of GR-F with Cu(II). Results of XRD analysis of GR-F(Cu) produced with 1 mM Cu(II) are shown in Fig. 5 and are compared with those of GR-F. The results show that some of the GR-F was transformed to magnetite by reaction with Cu(II), but no products containing Cu(II) are observed. Formation of magnetite means that there was a redox reaction between GR-F and Cu(II), because magnetite is often observed as a product of the oxidation of GRs (Lee and Batchelor, 2002; Maithreepala and Doong, 2005). Fig. 6 shows the results of surface analyses of GR-F and GR-F(Cu) by SEM. GR-F particles are hexagonal plates with lengths of about 400 nm (Fig. 6a). The addition of Cu(II) produced particles of about 10 nm and the number of these particles increased with increasing Cu(II) addition as shown in Figs. 6b–d. In this study, we could not identify these particles, but they could be metallic copper (Cu(0)) or

1113

Fig. 5. Results of XRD analysis for GR-F and GR-F modified with 1 mM of Cu(II). G represents GR-F and M represents magnetite. D-space values for GR-F were 7.43, 3.83, 2.67, 2.36, 1.58, and 1.55. D-space values for ˚. magnetite were 2.99, 2.55, 2.10, 1.72, 1.61, and 1.48 A

nano-size magnetite. The formation of similar particles (10–20 nm size) was observed in the reaction of GR-Cl and GR-SO4 with Cu(II) and analysis by x-ray photoelectron spectroscopy and X-ray absorption fine-structure spectroscopy showed that they were zero-valent copper (O’loughlin et al., 2003b; Maithreepala and Doong, 2005). Therefore, it can be assumed that the added Cu(II) is reduced to elemental Cu, even though direct evidence of this is not available. The following equation was proposed to express the reaction of GR-F with Cu(II)

Fig. 6. SEM image of GR-F and GR-F(Cu). (a) GR-F alone, (b) GR-F with 1 mM Cu(II), (c) and (d) GR-F with 5 mM Cu(II). SEM analysis conducted before GR-F(Cu) reacted with nitrate.

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

2þ III 6FeII 3 Fe1 ðOHÞ8  F þ 5Cu

!

III 8FeII 1 Fe2 O4

0



þ

þ 5Cu þ 6F þ 16H þ 16H2 O

ð3Þ

Results presented in Fig. 4 show that the concentration of Fe(II) in the solid phase decreased in proportion to the amount of Cu(II) added with a molar Fe/Cu ratio of 5.3, which is higher than that calculated using Eq. (3). This means that more Fe(II) was removed from GR by reaction with Cu(II), which might be due to dissolution of GR caused by hydrogen ion being produced (Eq. (3)). The dissolution of GR can be expressed by the following reaction, which was modified from the reaction proposed by Hansen et al. (2001). III þ II III 2þ 2FeII 3 Fe1 ðOHÞ8  F þ 8H ! Fe1 Fe2 O4 þ 5Fe

þ 2F þ 12H2 O

ð4Þ

Fe(II) in solution increased from 0.5 mM to 7.3 mM when Cu(II) concentration increased from 0 to 5 mM (data not shown), which is additional evidence of the dissolution of GR. Based on the above observations and previous results, it can be expected that the reduced form of Cu might play an important role in enhancing nitrate reduction rate by GRF. The main role of Cu(0) in improving rates of redox reactions is by facilitating electron transfer from reductant to target compounds (Ottley et al., 1997; O’Loughlin et al., 2003c; Maithreepala and Doong, 2005). The addition of Cu(II) accelerated the rate of nitrate reduction by Fe(II) sorbed onto ferric iron oxy-hydroxides and the mechanism was assumed to be production of a solid form of Cu, probably metallic Cu, that acted catalytically (Ottley et al., 1997). Similar behavior of solid phase Cu was also mentioned in the studies associated with the reductive dechlorination reaction by GRs. The transition metals added into GR were reduced to their zero-valent forms, which were assumed to play an important role in facilitating electron transfer from GR to chlorinated compounds (O’Loughlin et al., 2003c; Maithreepala and Doong, 2005). The rate constant observed with 5 mM of Cu(II) was found to be decreased compared to that at 2.5 mM Cu(II). This is somewhat unexpected and may be the result of Cu(II) consuming too much Fe(II), so that there would not be enough to rapidly reduce nitrate. Similar behavior was observed in experiments on reductive dechlorination by GR-Cl with Cu(II), where the fastest reaction was also achieved at an intermediate dose of Cu(II) (Maithreepala and Doong, 2005); the fastest reaction rate was observed when 1 mM Cu(II) was added to a 3 g l1 suspension of GR-Cl. The loss of Fe(II) and the decrease of pH in the reaction with Cu(II) were proposed as possible reasons for observing lower reaction rates at higher Cu(II) doses (Maithreepala and Doong, 2005). This may be an appropriate explanation for our experimental results as well, because their experiment differed from ours primarily in the target compound and the anion contained in GR. In addition, we also observed that pH decreased from 7.8 to 7.3 with the addi-

tion of 5 mM Cu(II) and that the Fe(II) concentration in GR decreased proportionally with increasing Cu(II) addition as shown in Fig. 4. 3.3. Effect of initial nitrate concentration The effect of initial nitrate concentration on nitrate reduction by GR-F(Cu) at pH 7.8 was studied over the range of 0.05 to 1.2 mM and the results are shown in Fig. 7. The rate constant of ammonium production (k) decreased from 4.14(±1.22) to 1.24(±0.19) h1 as initial nitrate concentration increased from 0.05 to 1.2 mM. Initial reduction rates were calculated by multiplying the reduction rate constant by the initial nitrate concentration and were found to be nonlinearly related to initial nitrate concentrations, as shown in Fig. 7. This relationship was described with the following saturation model, r0 ¼

rmax C init ; ðK m þ C init Þ

ð5Þ

where r0 is the initial degradation rate, rmax is the maximum initial degradation rate, Km is the nitrate concentration at its half maximal degradation rate, and Cinit is the initial nitrate concentration. The values of rmax and Km obtained through nonlinear regression using MATLAB are 2.38(±0.16) mM h1 and 0.55(±0.1) mM, respectively. This result supports the hypothesis that nitrate reduction was a surface saturation reaction in which reduction rate approaches a maximum value at high nitrate concentration and the reaction kinetics shift from first-order to zeroth-order. Furthermore, the results of these experiments might be also useful in estimating reaction rates at higher concentrations.

2.5

2.0

r0 (mM/h)

1114

1.5

1.0

0.5

0.0 0.0

0.2

0.4 0.6 0.8 1.0 Initial nitrate concentration (mM)

1.2

1.4

Fig. 7. Dependence of initial nitrate reduction rate on initial nitrate concentration. Cu(II) addition was 1 mM and initial solid-phase Fe(II) in GR was 87 mM. Error bars for r0 represent 95% confidential intervals. The solid line represents predictions of a saturation model.

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

Surface saturation behavior has also been observed in several studies on the reaction between chlorinated organics and solid reductants such as ZVI (Arnold and Roberts, 2000; Schafer et al., 2003; Janda et al., 2004), cement slurries containing Fe(II) (Hwang and Batchelor, 2002), and GR (Lee and Batchelor, 2002).

4. Conclusions The results of experiments on nitrate reduction by GR-F modified with Cu(II) demonstrated that the addition of Cu(II) into GR-F was effective in improving nitrate reduction rates. Ammonium concentration rather than nitrate concentration was measured in this study to determine the amount of nitrate reduced, because removal of nitrate could occur by both anion exchange and reduction reaction. The reaction kinetics were influenced by solution pH, with the fastest reaction being observed at pH 9. Kinetics of ammonium production were described by a first-order rate model. The model was able to adequately describe the behaviors of ammonium during reaction at pH 7.8, but was less successful at higher pH (pH 9 and 11). The errors of the model at high pH were mainly due to its inability to describe a delay in ammonium production relative to nitrate removal, which was observed to be more important at higher pH. This delay could be due to slow reduction of intermediates, such as nitrite, that interacted with the GR surface in different ways at higher pH. The rates of nitrate reduction rate increased when concentrations of Cu(II) added increased up to 2.5 mM, but a lower rate was observed at 5 mM of Cu(II). This might be due to loss of Fe(II) from GR due to the reaction with Cu(II). XRD results confirmed the formation of magnetite as an oxidation product of GR. SEM analysis revealed the formation of nano-size particles (about 10 nm), which might be metallic Cu or magnetite, but were not identified by this study. The initial ammonium production rate showed a saturation relationship with respect to nitrate concentration, which indicates that nitrate reduction was occurring on the surface of the GR solid. The results obtained from this study can be applied to develop a technology to treat water and wastewater contaminated by nitrate. However, the extent of fluoride release should be evaluated for any specific application. In addition, the results of this study might be helpful to improve our knowledge of the mechanism of nitrate reduction by GR as well as factors affecting the reactivity of green rusts.

Acknowledgements This material is based in part upon work supported by the Texas Advanced Technology Program under Grant No. 000512-0066-2001.

1115

References Alowitz, M.J., Scherer, M.M., 2002. Kinetics of nitrate, nitrite, and Cr(VI) reduction by iron metal. Environ. Sci. Technol. 36, 299–306. APHA, 1995. Standard Methods for the Examination of Water and Wastewater, 19th ed. American Public Health Association, Washington, DC. Arnold, W.A., Roberts, A.L., 2000. Pathways and kinetics of chlorinated ethylene and chlorinated acetylene reaction with Fe(0) particles. Environ. Sci. Technol. 34, 1794–1805. Canter, L.W., 1997. Nitrates in Groundwater. CRC Press, Boca Raton, Florida. Choe, S., Chang, Y., Hwang, K., Khim, J., 2000. Kinetics of reductive denitrification by nanoscale zero-valent iron. Chemosphere 41, 1307– 1311. Choe, S., Liljestrand, H.M., Khim, J., 2004. Nitrate reduction by zerovalent iron under different pH regimes. Appl. Geochem. 19, 335–342. Choi, J., Batchelor, B., 2004. Methods for producing high-activity modified green rusts. Remediation of chlorinated and recalcitrant compounds – 2004. In: Proceedings of the International Conference on Remediation of Chlorinated and Recalcitrant Compounds, 4th. Monterey, CA, United States. EPA, 1995. National Primary Drinking Water Regulations. Office of Ground Water & Drinking water, Washington, DC, US. Fanning, J.C., 2000. The chemical reduction of nitrate in aqueous solution. Coordin. Chem. Rev. 199, 159–179. Gao, W., Guan, N., Chen, J., Guan, X., Jin, R., Zeng, H., Liu, Z., Zhang, F., 2003. Titania supported Pd–Cu bimetallic catalyst for the reduction of nitrate in drinking water. Appl. Catal. B-Environ. 46, 341–351. Gibbs, C.R., 1976. Characterization and application of ferrozine iron reagent as a ferrous iron indicator. Anal. Chem. 48, 1197–1201. Hansen, H.C.B., 2001. Environmental chemistry of Iron(II)–Iron(III) LDHs (green rusts). In: Rives, V. (Ed.), Layered Double Hydroxides: Present and Future. Nova Science Publishers, New York. Hansen, H.C.B., Borggaard, O.K., Sorensen, J., 1994. Evaluation of free energy of formation of Fe(II)–Fe(III) hydroxide-sulphate (green rust) and its reduction of nitrate. Geochim. Cosmochim. Acta 58, 2599– 2608. Hansen, H.C.B., Guldberg, S., Erbs, M., Koch, C.B., 2001. Kinetics of nitrate reduction by green rusts – effects of interlayer anion and Fe(II):Fe(III) ratio. Appl. Clay Sci. 18, 81–91. Hansen, H.C.B., Koch, C.B., 1998. Reduction of nitrate to ammonium by sulphate green rust: activation energy and reaction mechanism. Clay Miner. 33, 87–101. Hansen, H.C.B., Koch, C.B., Nancke-Krogh, H., Borggaard, O.K., Sorensen, J., 1996. Abiotic nitrate reduction to ammonium: key role of green rust. Environ. Sci. Technol. 30, 2053–2056. Haugen, K.S., Semmens, M.J., Noval, P.J., 2002. A novel in situ technology for the treatment of nitrate contaminated groundwater. Water Res. 36, 3497–3506. Huang, C.-P., Wang, H.-W., Chiu, P.-C., 1998. Nitrate reduction by metallic iron. Water Res. 32, 2257–2264. Hwang, I., Batchelor, B., 2002. Reductive dechlorination of chlorinated methanes in cement slurries containing Fe(II). Chemosphere 48, 1019– 1027. Janda, V., Vaseka, P., Bizova, J., Belohlav, Z., 2004. Kinetic models for volatile chlorinated hydrocarbons removal by zero-valent iron. Chemosphere 54, 917–925. Johnson, T.M., Bullen, T.D., 2003. Selenium isotope fractionation during reduction of Fe(II)–Fe(III) hydroxide-sulfate (green rust). Geochim. Cosmochim. Acta 67, 413–419. Lee, W., Batchelor, B., 2002. Abiotic reductive dechlorination of chlorinated ethylenes by iron-bearing soil minerals. 2. Green rust. Environ. Sci. Technol. 36, 5348–5354. Legrand, L., Figuigui, A.E., Mercier, F., Chausse, A., 2004. Reduction of aqueous chromate by Fe(II)/Fe(III) carbonate green rust: kinetic and mechanistic studies. Environ. Sci. Technol. 38, 4587–4595.

1116

J. Choi, B. Batchelor / Chemosphere 70 (2008) 1108–1116

Luk, G.K., Au-Yeung, W.C., 2002. Experimental investigation of the chemical reduction of nitrate from groundwater. Adv. Environ. Res. 6, 441–453. Maithreepala, R.A., Doong, R.-A., 2005. Enhanced dechlorination of chlorinated methanes and ethenes by chloride green rust in the presence of copper(II). Environ. Sci. Technol. 39, 4082–4090. O’Loughlin, E.J., Burris, D.R., 2004. Reduction of halogenated ethanes by green rust. Environ. Toxicol. Chem. 23, 41–48. O’Loughlin, E.J., Kelly, S.D., Cook, R.E., Csencsits, R., Kemner, K.M., 2003a. Reduction of uranium(VI) by mixed iron(II)/iron(III) hydroxide (green rust): formation of UO2 nanoparticles. Environ. Sci. Technol. 37, 721–727. O’loughlin, E.J., Kelly, S.D., Kemner, K.M., Csencsits, R., Cook, R.E., 2003b. Reduction of AgI, AuIII, CuII, and HgII by FeII/FeIII hydroxysulfate green rust. Chemosphere 53, 437–446. O’Loughlin, E.J., Kemner, K.M., Burris, D.R., 2003c. Effects of Ag(I), Au(III), and Cu(II) on the reductive dechlorination of

carbon tetrachloride by green rust. Environ. Sci. Technol. 37, 2905– 2912. Ottley, C.J., Davison, W., Edmunds, W.M., 1997. Chemical catalysis of nitrate feduction by iron(II). Geochim. Cosmochim. Acta 61, 1819– 1828. Refait, P.H., Drissi, S.H., Pytkiewicz, J., Genin, J.M.R., 1997. The anionic species competition in iron aqueous corrosion: Role of various green rust compounds. Corros. Sci. 39, 1699–1710. Schafer, D., Kober, R., Dahmke, A., 2003. Competing TCE and cis-DCE degradation kinetics by zero-valent iron-experimental results and numerical simulation. J. Contam. Hydrol. 65, 183–202. Siantar, D.P., Schreier, C.G., Chou, C.-S., Reinhard, M., 1996. Treatment of 1,2-dibromo-3-chloropaopane and nitrate contaminated water with zero-valent iron or hydrogen/palladium catalysts. Water Res. 30, 2315–2322. Williams, A.G.B., Scherer, M.M., 2001. Kinetics of Cr(VI) reduction by carbonate green rust. Environ. Sci. Technol. 35, 3488–3494.