Reductive transformation of pentachloronitrobenzene by zero-valent iron and mixed anaerobic culture

Chemical Engineering Journal 210 (2012) 309–315

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Reductive transformation of pentachloronitrobenzene by zero-valent iron and mixed anaerobic culture Weizhao Yin a,b,c, Jinhua Wu a,b,c,⇑, Ping Li a,b,c, Guanghui Lin a,b,c, Xiangde Wang a,b,c, Bin Zhu a,b,c, Bo Yang d a

The Key Laboratory of Environmental Protection and Eco-Remediation of Guangdong Regular Higher Education Institutions, Ministry of Education, China The Key Lab of Pollution Control and Ecosystem Restoration in Industry Clusters, Ministry of Education, China c School of Environmental Science and Engineering, South China University of Technology, Guangzhou 510006, China d School of Chemistry and Chemical Engineering, Shenzhen University, Shenzhen 518060, China b

h i g h l i g h t s " The bio-iron system showed sufficient p-CNB removal capacity. " Optimum pH, high iron dosage and ethanol enhanced the p-CNB removal. " Nitrate and sulfate suppressed the bio-iron system performance. " Dissolved oxygen did not influence the bio-iron system significantly. " Reduction and adsorption/co-precipitation was the major p-CNB removal mechanisms.

a r t i c l e

i n f o

Article history: Received 24 May 2012 Received in revised form 14 August 2012 Accepted 3 September 2012 Available online 11 September 2012 Keywords: Pentachloronitrobenzene (p-CNB) Zero-valent iron Microorganism Reduction

a b s t r a c t The simultaneous pentachloronitrobenzene (p-CNB) reduction by both zero-valent iron (ZVI) and microorganisms was investigated through batch experiments. Compared to the mono-iron system, the bio-iron system showed more sufficient p-CNB removal capacity. A p-CNB removal rate of 94.2% and 59.8% was obtained in the bio-iron system and the mono-iron system with a corresponding pentachloroaniline (p-CAN) recovery rate of 59.4% and 29.8%, respectively. The p-CNB removal rate increased significantly from 26.9% to 92.2% with an increase of iron dosage from 0 to 3.0 g L1. The results also showed that the maximum p-CNB removal rate and corresponding p-CAN recovery rate were 95.6% and 56.4% at an initial pH of 6.0, while lower pH values would inhibit the removal of p-CNB in this combined system. It is found that organic substrates can improve the p-CNB removal rate, among which ethanol was found to be the most effective electron donor for this bio-iron system and the corresponding p-CNB removal rate in presence of ethanol was 98.2%. Common electron acceptors in groundwater were found to inhibit 2 the reduction of p-CNB in the following order: NO 3 > SO4 > O2 and it suggested that sulfate and oxygen exert limited inhibition effects on the reduction. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction As the most common chloronitrobenzenes (CNBs), pentachloronitrobenzene (p-CNB) is widely used in various industrial processes such as for the manufacture of insecticides, herbicides, dyes and explosives as well as serving as an agricultural fungicide in crops such as potatoes, wheat, onions and others, resulting in groundwater pollution and chronic threat to water safety [1]. China has produced more than 60% of p-CNB annually in the world. The potential genotoxicity and carcinogenicity of p-CNB was ⇑ Corresponding author at: School of Environmental Science and Engineering, South China University of Technology, Guangzhou 510006, China. Tel./fax: +86 20 39380568. E-mail address: [email protected] (J. Wu). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.09.003

reported [2] and p-CNB is listed as a pollutant that needed to be immediately controlled in China and other countries [3]. The maximum concentration of CNBs is regulated to be 0.05 mg L1 in drinking water according to the Quality Standard for Surface Water in China (GB 3838-2002). Up to date, very few literatures were found on methods and processes for p-CNB removal. Studies on the metabolism of CNBs were mainly focused on oxidative pathways and mechanisms using advanced oxidation processes [4–6]. However, the expensive operation cost limited their applications. Because of the electronwithdrawing property of the nitro and chlorine groups in the aromatic ring, a more cost-effective microbial reductive transformation of chlorinated compounds was extensively used as a predominant biodegradation pathway for removing chloro-aromatics from water [7–9]. Specific bacterial strains were found to be able

310

W. Yin et al. / Chemical Engineering Journal 210 (2012) 309–315

to transform CNBs to corresponding chloroanilines (CANs) by cometabolism in presence of additional carbon and nitrogen sources. It is also reported that Pseudomonas putida [10] and Pseudomonas stutzeri [11] could mineralize o-CNB as sole carbon, nitrogen and energy source. The CANs can be mineralized if aniline dioxygenase and catechol 1,2-dioxygenase were available in the bacteria under anaerobic condition [10]. However, the secondary pollution resulted from additional carbon source and the slow mineralization rate of CNBs would significantly disturb the promotion of this biological treatment in groundwater remediation. Fe0 has been recently added into anaerobic wastewater treatment reactors to enhance the conversion efficiency of various contaminants [12–14]. H2 evolved during iron corrosion was found to promote the dechlorination of CNBs and, additionally, the coprecipitation from released ferrous iron and free sulfide could alleviate inhibition effect on the dechlorination efficiency by sequestering sulfide [12]. Research on bio-iron systems was mostly focused on positive effects of additional iron on traditional bioreactors in which metallic iron did not serve as a major reductant to contaminants. Gillham [15] and Agrawal et al. [16] reported that metallic iron could induce the reduction of chlorinated aromatics and nitro aromatics in permeable reactive barrier (PRB). CNBs are readily reduced to corresponding CANs by undergoing a six electrons reduction process in the Fe0–H2O system [17]. Besides the metallic iron and corresponding iron corrosion products, relevant microbial communities were also observed in long-term PRB systems [18,19]. Actually, in a Fe0–H2O system, hydrogen released from iron corrosion can function as a potential electron donor for autotrophic hydrogen-consuming bacteria such as acetogenic, methanogenic, denitrifying, sulfate-reducing (SRB) and iron-reducing bacteria (IRB). In addition, the existence of these microbial communities could promote specific reactive precipitations [20,21]. However, few reports were available on the performance of bio-iron systems where contaminants could be reduced by both of metallic iron and bacteria, as well as the influence of operational factors and field conditions such as iron dosage, pH value and corresponding effects resulted from the presence of common anions, oxygen, specific organic matters, etc. Based on the above consideration, a batch reactor was established to estimate the combined effect of iron and microorganism on p-CNB removal in groundwater. The p-CNB reduction in this integrated system was of special interest since the transformation of p-CNB involved two coupled reactions: one abiotic, where metallic iron served as a reductive agent and released hydrogen by corrosion; and one biotic, which was mediated by iron and corresponding hydrogenotrophic microorganisms or/and even heterotrophic bacteria. In this study, the performance of the bio-iron system on p-CNB reduction was evaluated under different iron dosages, pH values and organic substrates. The effects of common electron acceptors such as sulfate, nitrate and oxygen in groundwater on p-CNB reduction were also studied.

2. Materials and methods 2.1. Microorganism and synthetic groundwater A mixed culture of p-CNB-reducing bacteria was prepared in an anaerobic bioreactor (5L) from sludge obtained freshly from an anaerobic digester of a wastewater treatment plant. The acclimation process started at an initial suspended solid (SS) and a granular iron dosage of 5 g L1 and 10 g L1, respectively. The feeding p-CNB concentration was gradually increased from 5 to 50 mg L1. The nutrient medium contained glucose 150 mg L1, NH4Cl 500 mg L1, K2HPO4 250 mg L1, MgCl2 200 mg L1, CaCl2

100 mg L1, NaHCO3 1500 mg L1 and trace metals (CoCl25H2O, CuCl22H2O, MnCl24H2O, NiCl26H2O, NH4MO3, and ZnSO4, each at 0.5 mg L1). The hydraulic retention time was set at 1 d. After 6 months of acclimation, p-CNB removal in the anaerobic bioreactor was over 85% and corresponding p-CAN recovery was about 80%. This mixed culture was inoculated in the biotic ZVI batch experiment without further acclimation and enrichment. The initial biomass concentration was adjusted to approximately 880 mg L1 of mixed liquor suspended solids (MLSS) with 400 mg L1 of volatile suspended solid (VSS). The background composition of simulated groundwater was modified and designed according to the National Standard of Groundwater (GB/T14848-93) and nutritional requirement for microorganisms. The simulated groundwater was made by deionized water and contained 150 of NH4Cl, 50 of KH2PO4, 240 of MgSO4 7H2O, 16 of CaCl22H2O, 60 of FeSO47H2O, 0.5 of CoCl25H2O, 0.5 of CuCl22H2O, 0.5 of MnCl24H2O, 0.5 of NiCl26H2O, 0.5 of NH4MO3, 0.5 of ZnSO4 in mg L1. The medium was buffered with 900 mg L1 of NaHCO3. In order to simulate anaerobic conditions in groundwater, the deionized water was purged with N2 (1.0 L min1) for 30 min prior to use. Sulfate, nitrate and acetate were presented in sodium salt forms. All chemicals were analytical grade. 2.2. Batch experiment The iron powder (Guangdong Metals Co.) was approximately 0.05 mm in diameter and with a Brunauer-Emmett-Teller (BET) surface area of 0.024 m2 g1. It has a metallic glaze surface and was used without further pre-treatment. The p-CNB containing groundwater was prepared by dissolving a known amount of p-CNB (Aladdin-reagent Co., Shanghai, China) into the simulated groundwater. The initial p-CNB concentration was set to be 30 mg L1. For this p-CNB containing groundwater, the initial pH value was adjusted to 7.0 by HCl or NaOH. All the reactors (300 mL conical flasks) contained 250 mL of solution and 50 mL of headspace. After the addition of metallic iron powder, p-CNB containing groundwater, and mixed culture, the reactors were capped with rubber plugs. The headspaces were flushed with nitrogen gas for 20 s (0.5 L min1) by inserting two needles through the rubber plugs. Then the reactors were put in an anaerobic glove box for static reaction at 25 °C. The experiment process for investigating the effect of oxygen and the sampling procedure followed our previous work [22]. The pH value was monitored by a pH electrode set at the top of the reactor. 2.3. Analytical methods p-CNB and p-CAN were analyzed by a high-performance liquid chromatography (HPLC) (L-2000, Hitachi). A Luna 5u C18 column was used for reversed-phase separation, using a mobile phase of MeOH/H2O (50/50 v/v) with a flow rate of 1.0 mL min1. The injection volume for all samples was 10 lL and the detection wavelength was 254 nm. p-CNB and p-CAN concentrations were quantified with the external reference method. The concentrations of sulfate and nitrate after reaction were determined by an ion chromatography (ICS1000, Dionex). The pH and DO values were monitored using a pH meter (PHS-3C, Sanxin, China) with a pH electrode (E-201-C, Sanxin, China) and a DO meter (550A, YSI, America), respectively. 2.4. Data analysis The p-CAN recovery (Rp-CAN) in this study was calculated according to Eq. (1):

311

W. Yin et al. / Chemical Engineering Journal 210 (2012) 309–315


 ¼ C p-CAN  M p-CNB = C 0p-CNB  Mp-CAN

ð1Þ

(a) 35

where and Cp-CAN were the concentration of removal p-CNB and the final concentration of p-CAN, respectively. Mp-CNB and Mp-CAN were the corresponding molar mass of p-CNB and p-CAN. Pseudo zero-order kinetic model (Eq. (2)) and pseudo first-order kinetic model (Eq. (3)) were employed as the preliminary assessments of the p-CNB removal.

30

Rp-CAN

C 0p-CNB

ð2Þ

ln ðC t =C 0 Þ ¼ k1 t

ð3Þ

-1

p-CNB/mg.L

C 0  C t ¼ k0 t

25

ð4Þ

where V (L) is the solution volume, a (m2 g1) is the specific surface area of iron particles and m (g) is the iron dosage.

15 10

Medium control cell 0 Fe 0 Fe +cell

5

where C0 is the initial concentration of p-CNB, and Ct is the concentration of p-CNB at reaction time t (h). In order to present comparable data for different iron specimens, the rate coefficients ksa based on specific surface area was calculated by the following model (Eq. (4)).

K sa ¼ ðk1  VÞ=ða  mÞ

20

0 0

(b)

-1

p-CNB þ 6Hþ þ 3Fe0 ! p-CAN þ 3Fe2þ þ 2H2 O

ð5Þ

2Fe0 þ p-CNB þ 2H2 O ! 2c-FeOOH þ p-CAN

ð6Þ

8c-FeOOH þ Fe0 ! 3Fe3 O4 þ 4H2 O

ð7Þ

2c-FeOOH þ Fe2þ ! Fe3 O4 þ 2Hþ

ð8Þ

However, only 29.8% of total p-CNB removal was recovered as p-CAN in the Fe0 control experiment. Since bimetallic material was not used in our experiment, no dechlorination and no aniline (AN) was observed in this abiotic experiment. Therefore, most of the p-CNB was removed because of the adsorption and/or coprecipitation by relative iron oxyhydroxides (e.g. FeOOH and Fe3O4) generated in situ during iron corrosion (Eq. (6)–(8)) [23–26]. In this physical removal process, the FeOOH was primarily generated because of iron corrosion (Eq. (6)). Since FeOOH is a sufficient absorbent and is ready to turn into Fe3O4 (a dense material) (Eq. (7) and (8)), p-CNB could be trapped in the iron corrosion layer during the Fe3O4 formation. Although Fe2+ was regarded as an electron conductor, most of these ferrous ions also can be consumed as Fe3O4 being generated. Therefore, the reduction of p-CNB was limited in the pure iron treatment. In the cell control reactor, although

30

40

50

30

40

50

Medium control cell 0 Fe 0 cell+Fe

15

p-CAN/mg.L

Compared experiments were conducted to investigate the performance of the bio-iron and the control systems. The p-CNB removal efficiency in corresponding batch reactors is shown in Fig. 1. Almost 95% of the initial p-CNB was removed with 14.2 mg L1 p-CAN production within 48 h in the reactor containing both Fe0 and microorganism. In contrast, only 59.8% and 27.1% of p-CNB removal rate was obtained with 4.59 and 0.9 mg L1 p-CAN production in the Fe0 and cell control reactors, respectively. Correspondingly, the final p-CAN concentration accounts for about 58.5% of the total p-CNB removal (on molar basis) in the bio-iron reactor compared to 29.8% in the Fe0 control reactor. The p-CNB reduction might occur in the cell control reactor since the final p-CAN concentration was observed to be 0.9 mg L1 in this reactor. According to the pathway for abiotic p-CNB reduction by Fe0 suggested by Xu et al. [17], p-CNB could be reduced to p-CAN by accepting six electrons (Eq. (5)):

20

t/h

3. Results and discussion 3.1. p-CNB removal in batch reactors

10

10

5

0 0

10

20

t/h Fig. 1. p-CNB removal rate (a) and corresponding p-CAN production (b) in different batch reactors (C0: 30 mg L1; iron dosage: 3.0 g L1; initial pH: 7.0; T: 25 °C).

no external energy source (e.g., organic substrate and iron) was added, the residual organic substrates that carried over from the seed culture and/or the endogenous respiration can be considered as an energy source for p-CNB reduction during the experimental period. Major p-CNB removal in the cell control experiment was attributed to cell adsorption. More sufficient p-CNB removal efficiency was obtained in the integrated bio-iron system. The data show that the p-CAN production in the bio-iron reactor was 2 times higher than that in the Fe0 control reactor, indicating that the p-CNB reduction was significantly promoted by the interaction between metallic iron and microorganisms. This can be explained that cathodic hydrogen generated in the anaerobic iron corrosion process was utilized by the autotrophic bacteria (e.g., IRB and SRB) to reduce p-CNB [19]. Instead of passive secondary iron corrosion, positive precipitations that are always assigned as iron sulfide and carbonate green rusts which were promoted by specific bacteria (e.g. IRB and SRB) in the bio-iron system [26–28]. They maintained the transportation of electron from Fe0 to p-CNB and the formation of sufficient hydrogen. In addition, since p-CAN could hardly be absorbed by the iron oxyhydroxides and the cell (data not shown), more reactive sites were regenerated after p-CNB reduction, which favored the total p-CNB removal in the bio-iron system. Kinetic models were employed to elucidate the p-CNB removal process in different batch reactors. Related rate coefficients (k) and corresponding standard errors are given in Table 1. The pseudo

312

W. Yin et al. / Chemical Engineering Journal 210 (2012) 309–315

Table 1 Kinetic parameters of p-CNB removal rate in different batch reactors. Reactor

Initial p-CNB (mg L1)

Iron dosage (g L1)

Pseudo zero-order kinetics k0 (mg L

Cell Fe0 Fe0 + cell

30 30 30

3.0 3.0 3.0

1

h

1

)

0.148 ± 0.005 0.363 ± 0.001 0.596 ± 0.002

Pseudo first-order kinetics

R20

k1 (h1)

R21

0.7678 ± 0.038 0.9467 ± 0.023 0.8904 ± 0.039

0.006 ± 0.001 0.018 ± 0.001 0.061 ± 0.004

0.8077 ± 0.026 0.9894 ± 0.011 0.9962 ± 0.007

k: the observed rate constant; R: the related coefficients. ksa, the rate coefficient bases on surface area, was determined as 0.0625 L/(h m2) for the Fe0 treatment.

(a) 36

0

-1

0

Fe =0 gL 0 -1 Fe =3.0 gL

-1

-1

Fe =1.0 gL 0 -1 Fe =5.0 gL 0 -1 Fe =10.0 gL

30

p-CNB/mg.L

first-order rate coefficient (k1) of the bio-iron system was estimated to be 0.061 h1, obviously higher than that of the Fe0 control. Since the reactive area was important and limited (resulted from iron corrosion) in the chemical reduction, the regeneration of reactive sites was the rate-limited step in the p-CNB reduction process. Alternatively, the pseudo zero-order kinetics model (R02 = 0.9467) can be applied to the p-CNB removal process in the iron control reactor. However, since p-CNB in the Fe0 control system was eliminated mainly by physical removal mechanism, the mass transfer process had more significant effect on the pCNB removal. Thus the pseudo first-order model (R21 ¼ 0:9894) was better than the pseudo zero-order one for kinetic modeling of the mono-iron experiments. For the bio-iron system, the pCNB removal process would not merely depend on the reactive site/area of Fe0, rather on the combined effects from biotransformation, reduction by Fe0 and sorption/co-precipitation. It can hence be modeled by the pseudo first-order kinetics.

24 18 12 6 0 0

10

20

30

40

50

40

50

t/h 3.2. Effect of iron dosage

(b) 20 0

-1

15

p-CAN/mg.L

The effect of iron dosage on p-CNB removal in the bio-iron system was explored by varying the iron dosage from 0 to 10.0 g L1. As metallic iron and other energy sources were absent, only about 26.9% of the total initial p-CNB could be removed in the mono-cell system. Fig. 2a shows an obvious improvement of the p-CNB removal rate from 72.5% to 92.2% with an increase of iron dosage form 1.0 to 3.0 g L1, consistent with our previous report [22]. The increase of iron dosage would result in more reactive sites for p-CNB reduction by Fe0 and more hydrogen for the biotransformation of p-CNB in this bio-iron system. The pseudo first-order degradation rate coefficients (k1) were estimated by non-linear regression between concentrations versus time. The rate coefficients for these treatments containing different iron dosages were 0.025 ± 0.001 h1 (R2 = 0.96 ± 0.01), 0.054 ± 0.003 h1 (R2 = 0.99 ± 0.01), 0.070 ± 0.003 h1 (R2 = 0.99 ± 0.01) and 0.074 ± 0.004 h1 (R2 = 0.98 ± 0.01) at an iron dosage of 1.0, 3.0, 5.0 and 10.0 g L1, respectively. No further enhancement was obtained on the p-CNB removal when the iron dosage was over 3.0 g L1, indicating that except the amount of reactive sites, p-CNB mass transfer from solution to iron surface might be the rate-limited step during the iron reduction process. Moreover, the p-CNB biological reduction rate was limited by the biomass concentration since H2 was overloaded for the hydrogen-using bacteria at a relative high iron to microorganism ratio. The p-CAN production was presented in Fig. 2b, in which a p-CAN recovery rate of 40.1%, 57.5%, 65.2% and 70.0% was observed at an iron dosage of 1.0, 3.0, 5.0 and 10.0 g L1, respectively. It implied that p-CNB reduction was favorable in higher iron dosage in the bio-iron system. Although adsorptive iron corrosion products were found to be important during contaminant removal in a Fe0–H2O system, the larger iron corrosion scale did not bring any enhancement on the p-CNB removal efficiency in this experiment. It possibly resulted from the complicated interaction between mass transport and chemical reaction [29].

-1

Fe =0 gL 0 -1 Fe =1.0 gL 0 -1 Fe =3.0 gL 0 -1 Fe =5.0 gL 0 -1 Fe =10.0 gL

10

5

0 0

10

20

30

t/h Fig. 2. p-CNB removal (a) and p-CAN production (b) in the bio-iron system at different iron dosages (C0: 30 mg L1; initial pH: 7.0; T: 25 °C).

3.3. Effects of initial pH The effect of initial pH on p-CNB removal in the bio-iron treatment is presented in Fig. 3a and b. About 86.8%, 95.8%, 74.3%, 34.9% and 24.5% of the initial p-CNB was removed, accommodated with 12.8, 14.2, 10.6, 2.3 and 0.9 mg L1 of p-CAN production at an initial pH value of 5.0, 6.0, 7.0, 8.0 and 9.0, respectively. The bio-iron system was found to be highly pH-dependent. A serious inhibition on p-CNB reduction was observed when the initial pH value was higher than 8.0. The p-CNB removal rate increased as the initial pH values decreased from 8.0 to 6.0. However, the maximum p-CNB removal efficiency occurred at initial pH 6.0 instead of the

313

W. Yin et al. / Chemical Engineering Journal 210 (2012) 309–315

(a)

35

pH=5.0 pH=6.0 pH=7.0

30

pH=8.0 pH=9.0

p-CNB/mg.L

-1

25 20 15 10

of aqueous iron corrosion. In this bio-iron system, the complicated interaction among microbial activity, ferrous (hydroxide) and specific anions resulted in positive precipitation species such as carbonate green rust which has considerable reactive area for contaminant reduction, adsorption and even precipitation [27,28]. During the growth of these iron corrosion products, the aqueous contaminants were adsorbed and enmeshed in the matrix of the iron oxide. Accordingly, the contaminants were progressively reduced and even incorporated in the growing oxide film on the iron surface [29] under relative low pH condition (5.0–7.0) due to the combination of the iron corrosion process and microbial activity.

5

3.4. Effects of additional organic substrates 0 20

30

40

50

30

40

50

t/h

(b)

15

pH=5.0 pH=6.0 pH=7.0 pH=8.0 pH=9.0

p-CAN/mg.L

-1

12

9

6

3

0 0

10

20

t/h Fig. 3. p-CNB removal (a) and p-CAN production (b) in the bio-iron system at different initial pH (C0: 30 mg L1; iron dosage: 1.0 g L1; T: 25 °C).

lowest initial pH value of 5.0. This suggested a characteristic difference from the contaminant reduction in the mono-iron system. Huang and Zhang [30] reported that in the Fe0–H2O system, lower pH could eliminate the passive oxide layer on the iron surface and result in releasing of Fe2+ (reported as a ‘‘chmic-conductor’’), more reactive sites and lower ORP value, which favored the contaminant reduction. Similarly, the p-CNB removal rate would increase when the initial pH value decreased from 9.0 to 6.0. Because the absorbed p-CNB was partly reduced by hydrogen-using bacteria in the bio-iron system, when the initial pH value changed from 6.0 to 5.0, the microbial activity decreased and the bio-reduction efficiency dropped. To get better understanding of the removal mechanism in the combined bio-iron treatment, a primary p-CAN recovery analysis was conducted. About 56.7%, 56.4%, 54.2%, 25.7% and 13.5% pCAN recovery was found at an initial pH value of 5.0, 6.0, 7.0, 8.0 and 9.0, respectively. The p-CAN recovery increased about 40% when the initial pH value dropped from 9.0 to 7.0, but no significant fluctuation of p-CAN recovery occurred between pH 5.0 and 7.0. As aforementioned, the p-CNB removal contributed from the iron reduction is highly dependent on the available reactive sites. The lower initial pH in this integrated process would accelerate the regeneration of reactive sites on Fe0 available for the p-CNB reduction by metallic iron, although the bio-reduction was inhibited at the same time. Noubactep [29] reported that the oxide scale formation on Fe0 surface at pH > 4.5 is a fundamental characteristic

Since groundwater could be contaminated by common organic compounds, experiment was conducted to investigate the influence of acetate, glucose and ethanol on this bio-iron treatment at a relative low concentration (50 mg L1 as COD). As Fig. 4 shown, after 48 h reaction time, the total p-CNB removal rate increased from 71.0% (control) to 75.0%, 82.1% and 98.3% with a corresponding increase of p-CAN production (not shown in the figure) from 10.19 (control) to 11.73, 14.21 and 17.56 mg L1 at a dosage of 50 mg L1 (as COD) acetate, glucose and ethanol in 48 h. The pCAN recovery (not shown in the figure) increased from 51.4% (control) to 62.4%, 69.1% and 71.4% respectively, for acetate, glucose and ethanol. The presence of organic substrates in the bio-iron system leads to an increase of the p-CNB removal rate which was attributed to microorganisms that utilized the organic materials for p-CNB bio-reduction through co-metabolism [8,9]. Because ethanol was an effective electron donor during the contaminant bio-reduction and an easily biodegradable substrate, the ethanol system showed higher synergistic effect than acetate and glucose. In contrast to the control sample, about 27.3% increase in p-CNB removal and 7.37 mg L1 increase in p-CAN production was found in the presence of ethanol after the reaction. However, the maximum p-CAN production after 48 h reaction time was only 17.56 mg L1 and about 30% of the total p-CNB removal was still achieved in a physical way. Limited p-CAN production also occurred in the aforementioned iron dosage test. These results could be attributed to the following mechanism: (I) some p-CNB molecules were adsorbed on the surface of iron oxide (e.g. FeOOH), waiting for reduction by metallic iron [30]; (II) some other p-CNB molecules enmeshed in the matrix of the iron oxide and then were incorporated in the growing oxide film [29]; (III) p-CNB biotransformation

35

control -1 acetate=50 mgL as COD -1 glucose=50 mgL as COD -1 ethanol=50 mgL as COD

30 25 -1

10

p-CNB/mg.L

0

20 15 10 5 0 0

10

20

30

40

50

t/h Fig. 4. Effects of additional organic substrates on p-CNB removal in the bio-iron treatment (C0: 30 mg L1; iron dosage: 1.0 g L1; initial pH: 7.0; T: 25 °C).

W. Yin et al. / Chemical Engineering Journal 210 (2012) 309–315

(a)

3.5. Effects of oxygen, sulfate and nitrate

nitrogen air oxygen

25 20 15 10 5 0 0

10

20

30

40

50

t/h

(b)

35

2-

SO4 =0 mgL 2-

-1

SO4 =50 mgL

30

-1

2-

-1

2-

-1

40

50

SO4 =100 mgL SO4 =500 mgL

25

ð10Þ

Fig. 5b shows the p-CNB removal rate at different sulfate concentrations in the bio-iron system. The p-CNB removal rate decreased from 91.4% to 84% then to 78.4% with an increase of sulfate concentration from 0 to 50 then to 100 mg L1after 48 h reaction time, respectively. It slowly decreased as the sulfate dosage increased from 100 to 500 mg L1. The final sulfate concentration was 25.3, 66.9 and 455.1 mg L1 as the initial sulfate concentration was 50, 100, 500 mg L1 respectively. In our previous investigation, no sulfate was reduced by Fe0, but SO2 4 was found to be able to enhance nitrobenzene and p-CNB removal [22,32] since SO2 4 could promote the contaminant reduction by Fe0 through removing the iron oxides and hydroxides from the iron surface [30]. However, in the system containing metallic iron and hydrogen-using bacteria, obvious suppression on p-CNB removal occurred in the presence of sulfate. Since hydrogen-using bacteria (sulfate reducing bacteria, SRB) could promote sulfate reduction by utilizing H2 as energy source, thus, the amount of p-CNB biotransformation would be diminished due to the electron competition with sulfate reduction. Lin et al. [12] found that the free sulfide generated from the SO2 4 reduction process was also an inhibitor to certain autotrophic bacteria. But in this bio-iron system, the suppression effect from sulfate was found to be finite. Since the sulfate bio-reduction occurred and sulfide was generated accordingly, this free sulfide would co-precipitate with ferrous and then form iron sulfide [27,26,31]. Iron sulfide was also an efficient reactive mineral for contaminant reduction in the Fe0–H2O system. In addition, due to FeS production, it partially weakened the poisonous effect of free sulfide on the bacteria. Nitrate was also one of the common contaminants in groundwater. Fig. 5c shows the nitrate ions effect on p-CNB removal in the bio-iron system. About 72.3%, 60.0% and 45.7% p-CNB removal rate was obtained at 50, 100 and 300 mg L1 initial concentration of nitrate, while 93.9% p-CNB was removed by the free nitrate treatment after 48 h reaction time. The final nitrate concentration (not shown in the figure) was found to be 23.6, 38.4, 167.0 mg L1 at the initial nitrate concentration of 50, 100, 300 mg L1, respectively. Compared to sulfate, nitrate would lead to more serious depression for p-CNB removal. Many studies have examined the

20 15 10 5 0 0

10

20

30

t/h

(c)

35

-

NO3 =0mgL -

30

-1

NO3 =50mgL

-1

-

-1

-

-1

NO3 =100mgL

25 -1

6Fe3 O4 þ O2 þ 4Hþ ! 8c-Fe2 O3 þ 2Fe2þ þ 2H2 O

ð9Þ

p-CNB/mg.L

4Fe2þ þ O2 þ 4Hþ ! 4Fe3þ þ 2H2 O

p-CNB/mg.L

-1

Some common electron acceptors such as oxygen, sulfate and nitrate are considered to be vital components of groundwater that might influence the performance of PRB systems. The bio-iron treatment performance with different electron acceptors is shown in Fig. 5. No significant inhibitory effect was observed as increasing the oxygen pressure in the headspace of the reactor (Fig. 5a). This observation was different from our previous result [22] that dissolved oxygen could act as an effective inhibitor during nitrobenzene reduction in the Fe0–H2O system. It revealed that microorganisms presented in the Fe0–H2O system could weaken the negative effect of dissolved oxygen since oxygen is a favorable electron acceptor during the bio-chemical reaction in a cell. Due to the microbial activity, the majority of the dissolved oxygen was consumed immediately by the bacteria; only small amount of oxygen could react with iron corrosion products and produce non-reactive ferric oxide mineral (Eq. (9) and (10)) [26,27].

35 30

-1

was achieved by the bacteria after adsorption. It is concluded that the physical removal mechanism would significantly contribute to the p-CNB elimination in the bio-iron treatment during the observation time although the additional organic substrates could improve the removal rate of p-CNB.

p-CNB/mg.L

314

NO3 =300mgL

20 15 10 5 0 0

10

20

30

40

50

t/h Fig. 5. p-CNB removal performance of the bio-iron treatment in present of oxygen (a), sulfate (b) and nitrate (c) (C0: 30 mg L1; iron dosage: 3.0 g L1; initial pH: 7.0; T: 25 °C).

feasibility of nitrate reduction by metallic iron. Positive effect of nitrate on contaminant reduction by metallic iron was observed in our previous reports [22,32]. In the Fe0–H2O system, the nitrate reduction process could promote the generation of magnetite (electronic conductive iron corrosion) which would benefit contaminant reduction by iron [30]. In addition, the presence of nitrate was able to accelerate the iron corrosion rate and improve the

W. Yin et al. / Chemical Engineering Journal 210 (2012) 309–315

release of electron and hydrogen. However, nitrate enhancement on p-CNB removal did not occur in this bio-iron system. Since nitrate has also been found to be easier for biotransformation, large amount of H2 was used for the biological denitrification and p-CNB biotransformation process was restrained. Different from the inhibitory effect of sulfate, the p-CNB removal rate decreased significantly with an increase of nitrate concentration, indicating that the p-CNB removal was directly suppressed by the electron competition from the nitrate reduction. 4. Conclusions This study has evaluated the bio-iron system performance under various conditions. The integrated system possessed more sufficient capacity for p-CNB removal compared to the system only containing metallic iron. The removal mechanism included chemical reduction by Fe0, bio-reduction and physical removal (e.g. adsorption and co-precipitation by secondary minerals). Although low pH could promote the reduction of contaminants in the Fe0 system, the bio-iron reactor exhibited relatively low removal efficiency for p-CNB at an initial pH of below 6.0. This result implied that microorganisms could improve the performance of the Fe0– H2O system in neutral and even relatively high pH condition. Additionally, microorganisms could also prevent negative effect on contaminant removal from dissolved oxygen. But sulfate and nitrate could lead to significant suppression on the bio-iron system performance. Sulfate exhibited complicated inhibition on the bio-iron system through the interaction among electron competition from sulfate bio-reduction, biological poisonous effect of the free sulfide generated during the sulfate bio-reduction and iron sulfide formation during the iron corrosion process. In contrast, nitrate was found to directly diminish the p-CNB removal since nitrate could be reduced by Fe0 and denitrifying bacteria, resulting in the competition for limited electrons between nitrate and p-CNB. In addition, the results indicate the performance of PRB system for groundwater remediation only using metallic iron could be improved by inoculating anaerobic culture. However, in this combined bio-iron batch reactor, the interaction among chemical, biological and physical removing mechanism has not been fully understood. Kinetics experiments and column studies under specific field conditions including flow rate, temperature and relative geological conditions are needed for further investigation. Acknowledgments The authors thank the National Natural Science Foundation of China (50708039, 51178191), the National High-Tech Research and Development Program of China (863 Pro-gram; Grant 2009AA063902), the Fundamental Research Funds for the Central Universities (2012ZZ0047) and Guangdong Provincial Natural Foundation for Science (05300188). References [1] R.S. Nair, F.R. Johannsen, G.J. Levinskas, J.B. Terrill, Subchronic inhalation toxicity of p-nitroaniline and p-nitrochlorobenzene in rats, Fundam. App. Toxicol. 6 (1986) 618–627. [2] J. Feng, Nitrochlorobene market shrinks gradually, Chin. Chem. Rep. 2 (6) (2005) 21 (in Chinese). [3] USEPA, National Pollutant Discharge Elimination System, Code Of Federal Regulations, Agency EP, US Government Printing Office, Washington, DC, 1988. [4] Z.L. Chen, J.M. Shen, X.Y. Li, Ozonation degradation of p-nitrochlorobenzene in aqueous solution: kinetics and mechanism, J. Chem. Ind. Eng. 7 (2006) 2439– 2444.

315

[5] M.H. Priya, G. Madras, Photocatalytic degradation of nitrobenzenes with combustion synthesized nano-TiO2, J. Photochem. Photobiol. A: Chem. 178 (2006) 1–7. [6] T.Y. Zhang, L.Y. You, Y.L. Zhang, Photocatalytic reduction of pchloronitrobenzene on illuminated nano-titanium dioxide particles, Dyes Pigments 68 (2006) 95–100. [7] E. Katsivela, V. Wray, D.H. Pieper, R.F. Wittich, Initial reactions in the biodegradation of 1-chloro-4-nitrobenzene by a newly isolated bacterium strain LW1, Appl. Environ. Microbiol. 65 (1999) 1405–1412. [8] H.S. Park, S.J. Lim, Y.K. Chang, Degradation of chloronitrobenzenes by a coculture of Pseudomonas putida and Rhodococcus sp, Appl. Environ. Microbiol. 65 (1999) 1083–1091. [9] A. Kuhlmann, W. Hegemann, Degradation of mono-chloronitrobenzenes by Pseudomonas acidovorans CA50, Acta Hydrochim. Hydrobiol. 25 (1997) 298– 305. [10] H.Z. Wu, C.H. Wei, Y.Q. Wang, Q.C. He, S.Z. Liang, Degradation of ochloronitrobenzene as the sole carbon and nitrogen sources by Pseudomonas putida OCNB-1, J. Environ. Sci. 21 (2009) 89–95. [11] H. Liu, S.J. Wang, N.Y. Zhou, A new isolate of Pseudomonas stutzeri that degrades 2-chloronitrobenzene, Biotechnol. Lett. 27 (2005) 275–278. [12] H. Lin, L. Zhu, X. Xu, L. Zhang, Y. Kong, Reductive transformation and dechlorination of chloronitrobenzenes in UASB reactor enhanced with zerovalent iron addition, J. Chem. Technol. Biotechnol. 86 (2011) 290–298. [13] X. Xu, H. Lin, L. Zhu, Y. Yang, J. Feng, Enhanced biodegradation of 2chloronitrobenzene using a coupled zero-valent ironcolumn and sequencing batch reactor system, J. Chem. Technol. Biotechnol. 86 (2011) 993–1000. [14] D.O. Tas, S.G. Pavlostathis, The influence of iron reduction on the reductive biotransformation of pentachloronitrobenzene, Eur. J. Soil Biol. 43 (2007) 264– 275. [15] R.W. Gillham, Cleaning halogenated contaminants from groundwater, US Patent 5, 1993, pp. 206–213. [16] A. Agrawal, P.G. Tratnyek, Reduction of nitro aromatic compounds by zerovalent iron metal, Environ. Sci. Technol. 30 (1996) 153–160. [17] X. Xu, J. Wo, J. Zhang, Y. Liu, Catalytic dechlorination of p-NCB in water by nanoscale Ni/Fe, Desalination. 242 (2009) 346–354. [18] A.D. Henderson, A.H. Demond, Long-term performance of zero-valent iron permeable reactive barriers: a critical review, Environ. Eng. Sci. 24 (2007) 401– 423. [19] T.V. Noothen, D. Springael, L. Bastiaens, Positive impact of microorganisms on the performance of laboratory-scale permeable reactive iron barriers, Environ. Sci. Technol. 42 (2008) 1680–1686. [20] R. Gerlach, A.B. Cunningham, F. Caccavo, Dissimilatory iron-reducing bacteria can influence the reduction of carbon tetrachloride by iron metal, Environ. Sci. Technol. 34 (2004) 2461–2464. [21] T.V. Nooten, F. Lieben, J. Dries, E. Pirard, D. Springael, L. Bastiaens, Impact of microbial activities on the mineralogy and performance of column-scale permeable reactive iron barriers operated under two different redox conditions, Environ. Sci. Technol. 41 (2007) 5724–5730. [22] W.Z. Yin, J.H. Wu, P. Li, P.X. Wu, N.W. Zhu, X.D. Huang, B. Yang, Experimental study of zero-valent iron induced nitrobenzene reduction in groundwater: The effects of pH, iron dosage, oxygen and common dissolved anions, Chem. Eng. J. 182 (2012) 198–204. [23] C. Noubactep, An analysis of the evolution of reactive species in Fe0/H2O systems, J. Hazard. Mater. 168 (2009) 1626–1631. [24] C. Noubactep, The fundamental mechanism of aqueous contaminant removal by metallic iron, Water SA. 36 (2010) 663–670. [25] C. Noubactep, Metallic iron for water treatment: A knowledge system challenges mainstream science, Fresenius Environ. Bull. 20 (2011) 2632–2637. [26] T. Kohn, J.T. Livi, A.L. Roberts, P.J. Vilesland, Longevity of granular iron in groundwater treatment processes: corrosion product development, Environ. Sci. Technol. 39 (2005) 2867–2879. [27] I.H. Yoon, K.W. Kim, S. Bang, M.G. Kim, Reduction and adsorption mechanisms of selenate by zero-valent iron and related iron corrosion, App. Catal. B: Environ. 104 (2011) 185–192. [28] M.C. Lo, S.C. Lam, C.K. Lai, Hardness and carbonate effects on the reactivity of zero-valent iron for Cr(VI) removal, Water Res. 40 (2006) 595–605. [29] C. Noubactep, Metallic iron for safe drinking water production, Freiberg Online Geol. 27 (2011) 1–43. [30] Y.H. Huang, T.C. Zhang, Reduction of nitrobenzene and formation of corrosion coatings in zero-valent iron systems, Water Res. 40 (2006) 3075–3082. [31] D.H. Phillips, T.V. Nooten, L. Bastiaens, M.I. Russell, J.M. Ahad, T. Newton, T. Elliot, R.M. Kalin, Ten year performance evaluation of a field-scale zero-valent iron permeable reactive barrier installed to remediate trichloroethene contaminated groundwater, Environ. Sci. Technol. 44 (2010) 3861–3869. [32] C. Le, J.H. Wu, S.B. Deng, L. Ping, X.D. Wang, N.W. Zhu, P.X. Wu, Effects of common dissolved anions on the reduction of para-chloronitrobenzene by zero-valent iron in groundwater, Water Sci. Technol. 63 (2011) 1485–1490.