Arsenic removal in synthetic ground water using iron electrolysis

Separation and Purification Technology 122 (2014) 225–230

Contents lists available at ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Arsenic removal in synthetic ground water using iron electrolysis Lei Li a,b, Jun Li a, Chen Shao a, Kejia Zhang a,c,⇑, Shuili Yu a,⇑, Naiyun Gao a, Yang Deng d, Daqiang Yin e a

State Key Laboratory of Pollution Control and Resources Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China Department of Civil and Environmental Engineering, University of California, Berkeley, Berkeley, CA 94720, USA c Department of Civil Engineering, Zhejiang University (Zijin gang Campus), Hangzhou 310058, China d Department of Earth and Environmental Studies, Montclair State University, Mallory Hall 252, 1 Normal Ave, Montclair, NJ 07043, USA e Key Laboratory of Yangtze River Water Environment, Ministry of Education, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China b

a r t i c l e

i n f o

Article history: Received 23 June 2013 Received in revised form 6 November 2013 Accepted 7 November 2013 Available online 15 November 2013 Keywords: Iron scrap Electrocoagulation Arsenic removal Sustainable water treatment Groundwater

a b s t r a c t Electrocoagulation (EC) using electrodes made from iron scrap is a novel and promising strategy for arsenic (As) contaminated ground water remediation. In synthetic groundwater, amorphous hydrous ferric oxide (HFO) precipitates formed by the rapid dissolution of a sacrificial iron scrap anode adsorbed As very effectively. A competitive adsorption model developed in this study with parameters fitted for As and other coexisting anions (phosphate and silicate) was in good agreement with the observed results. It is indicated that the maximum adsorption capacity of HFO generated in the EC system was 0.70 mol/mol. Reducing the As concentration in water from 500 lg/L As(V) and As(III) to below 50 lg/L (local drinking water standard in Bangladesh) required 8 mg/L and 32 mg/L iron respectively (pH = 7.1 ± 0.1, charge dosage rate = 3 coulomb/L/min). It was found that coexisting cations (Ca2+ and Mg2+) neutralized the HFO surface charge, promoted aggregation and resulted in greater As removal. The presence of humic acid exhibited a negligible effect on As removal and HFO precipitate settling. Jar tests showed that the turbidity of the solution could be reduced to <1 NTU with the addition of 2 mg/L Al3+. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Tens of millions of people worldwide are exposed to toxic concentrations of naturally occurred arsenic (As) in groundwater drinking supplies [1,2]. In Bangladesh, an estimated one third of tube wells deliver groundwater with As concentrations higher than Bangladesh’s standard of 50 lg/L [3], and As accounts for one fifth of adult deaths [4]. The United States Environmental Protection Agency (EPA), the World Health Organization (WHO) and some developing countries such as China have decreased the maximum contaminant level of arsenic in drinking water from 50 to 10 lg/L [5–7]. Among the several proven technical approaches [8–10] for removing arsenic from municipal water supplies, iron-based strategy (i.e. chemical coagulation (CC) by Fe(III)) is widely accepted because of its high capacity for arsenate (As(V)) adsorption and low cost [11,12]. However, arsenite (As(III)), which is much more mobile and toxic than As(V) and accounts for up to 67–99% of the total As in groundwater [13], has orders of magnitude less ⇑ Corresponding authors. Addresses: State Key Laboratory of Pollution Control and Resources Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China, Department of Civil Engineering, Zhejiang University (Zijin gang Campus), Hangzhou 310058, China (K. Zhang, S. Yu). E-mail addresses: [email protected] (K. Zhang), [email protected] (S. Yu). 1383-5866/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.seppur.2013.11.012

affinity with Fe(III) precipitate than As(V) and other coexisting competitors such as Phosphate (P) [12,14]. Therefore, As(III) is usually pre-oxidized to As(V) using an oxidant (e.g. ferrate , chlorine, UV radiation, permanganate) to improve the removal efficiency [9,15]. Much of recent research has focused on Fenton and Fenton-like reagents such as Fe(II)/H2O2, ZVI/O2, and Fe(II)/O2 [14,16–18], since the intermediate production could oxidize As(III) to As(V), which could be easily absorbed by amorphous hydrous ferric oxide (HFO) precipitates formed afterwards. As a result, less iron is required as compared to direct adsorption [14]. Electrocoagulation (EC), based on the generation of Fe(II) through the rapid dissolution of a sacrificial Fe(0) anode, is a promising As removal strategy for drinking water as, (1) it is efficient, low cost and easy to maintain and operate with locally available materials [19,20], (2) EC introduces Fe(II)/Fe(III) [21] without introducing undesirable anions into the solution, (3) the release of H2 (g) from the cathode [22] neutralizes the consumption of hydroxide by the Fe(III) hydrolysis and therefore likely to buffer the system better than CC, and (4) the gradual release of Fe(II)/Fe(III) in EC may produce intermediate oxidants that enhance the efficiency of As(III) oxidation as compared to CC [14,23]. In addition, using electrodes made of iron scrap, an abundant byproduct from iron planing machines, would further reduce the material cost of EC compared with using iron plates.

226

L. Li et al. / Separation and Purification Technology 122 (2014) 225–230

Previous EC researchers have mostly focused on the effect of design and operation parameters (current density (current per electrode area, amps/cm2), electrode distance and conductivity, pH) or water matrix (different initial As, P and Si concentrations) on As removal, and proposed some qualitative conclusions [11,12,19, 22,24], which were mainly applicable only under the particular conditions in which the data were taken. In addition, the effect of typical co-existing substances such as Ca2+, Mg2+ and natural organic matter (NOM) on HFO precipitate formation and As removal in EC systems has not been previously reported, despite the fact that these substances exhibited great effects on HFO particle size and As removal efficiency in other similar As removal methods [15,25–27]. Moreover, the separation of HFO precipitates from EC system, one of the most significant steps for drinking water treatment, has rarely been explored in the literature. This paper reports the successful application of iron scrap electrolysis on arsenic remediation, discusses the redox reactions in the EC system, quantifies the effects of P and Si on As(V) and As(III) adsorption based on a competitive adsorption model and X-ray diffraction patterns, evaluates the effect of other co-existing substances (Ca2+, Mg2+ and humic acid) in groundwater on As oxidation and adsorption, and finally optimizes Al3+ concentration and other parameters for better HFO sludge separation. 2. Experimental procedures 2.1. Chemicals and experimental set up All chemicals were reagent grade or higher. Experiments were conducted in synthetic Bangladesh ground water (SBGW) containing 8.2 mM NaHCO3, 2.5 mM CaCl2 , 1.6 mM MgCl2, 500 lg/L As(III)/(V), 3 mg/L P and 30 mg/L Si according to British Geological Survey (BGS) [28] and previous study [14] for better comparison unless otherwise noted. Batches of SBGW were prepared by adding NaHCO3, MgCl2, Na2HPO4, and Na2SiO3 as solids to ultrapure 18 MX water in sequence under vigorous stirring. The pH was then reduced to 8 by bubbling CO2(g). CaCl2 stock solution (5% w/v) was added subsequently [14]. As(III) or As(V) was added before adjusting pH to 7.0 by bubbling CO2(g). All batches were aged at least 1 h after all components had been added and then sampled to verify the initial concentrations of P, Si, As(III), and As(tot) [23]. All experiments were conducted in a 1-L glass beaker. The spring-shaped iron scrap (origin steel type: 41CrAlMo74) were twist together to form solid electrodes before submerged in the SBGW and connected with the power supply by copper wires. Prior to experiments, electrodes were chemically cleaned with 1% HCl and rinsed by ultrapure water 3 times to remove the iron oxides and any passive film that may have formed. Samples taken from the reactor under various electrolysis duration were mixed for 2 h allowing complete Fe(II) oxidation and maximum As adsorption. Then, unfiltered samples were taken to determine total (dissolved and adsorbed) As (As(tot)). A second set of samples was filtered through 0.45-lm nylon filters to determine dissolved/ aqueous As, P and Si. 2.2. Chemical analysis As(III) and As(tot) concentrations were determined using a hydride generation atomic fluorescence spectrometer (AFS-230E). To selectively detect As(III), procedures were adopted from Roberts et al. [14]. P, Si and Fe(III) were determined with ICP-OES or ICPMS for low concentrations. All the detection procedures were done right after the experiment and the relative standard deviations (RSD) for all the determinations were <10% (normally <4%). A ferrozine method was used to determine the concentrations of

dissolved and total Fe(II) [29]. The suspension under various electrolysis duration in EC system were subjected to particle size detection at a 90 degree scattering angle using Dynamic Light Scattering, as well as zeta potential detection using Laser Doppler Microelectrophoresis (Zetasizer Nano ZS90, Malvern) within 10 min after sampling. The deviations between repeated experiments were <10% if error bars are not shown. 2.3. X-ray diffraction Powder diffraction data were collected at beamline 11-ID-B of the Advanced Photon Source located at Argonne National Laboratory (Argonne, IL). Air-dried samples were packed into 3 mm diameter polyimide tubes and diffraction data were obtained with 58 keV (0.2128 Å) X-rays. Radiation scattered from the sample was collected on an amorphous silicon MAR-345 image plate detector and processed using the Fit2-D program [30]. Diffraction data from a CeO2 standard were used to calibrate the sample-to-detector distance and tilt angle of the detector. Diffraction data were also collected for 2-line ferrihydrite prepared following the Schwertmann and Cornell recipe [31]. 2.4. Jar test procedure Al2(SO4)3 was newly made and added to the post-electrolysis solution prior to coagulation/flocculation in a 1 L jar tester run at 200RPM for 2 min, 90RPM for 6 min and 30RPM for 9 min. The supernatant was then sampled after different settling times for As and turbidity tests. 3. Results and discussion 3.1. Role of the electrodes Which redox reactions (e.g. the mechanism of Fe(0) dissolution and As(III) oxidation) occur on the electrodes in the EC system has been debated [21,32]. Reactions that could potentially be occurring on the electrodes are discussed in this section. As electrodes were exposed to the water and oxygen, Fe(0) would be sacrificed through electrolysis (Eqs. (1), (2)) and natural corrosion (Eq. (3)). Fig. SI1 shows that Fe(II) (total bivalent Fe) generated in the system accounted for over 90% of the total Fe (Fe(II) + Fe(III)) while the solution was sealed and purged with N2 during the electrolysis (DO was kept below 1 mg/L), implying that Fe(II) instead of Fe(III) was generated through the sacrifice of the anode. This conclusion is consistent with previous research with a similar ground water recipe in an EC system [21]. In addition, Fig. SI2 shows that the total iron concentration matched well (error <15%) with the value calculated using Faraday’s law (Eq. (4), where m is mass (g) of iron oxidized at a specific current, I is current (A), t is the time (s), M is the molecular weight of iron, Z is the number of electrons involved, and F is the Faraday constant (96,485.3 coulomb/mole)) under both high and low charge dosage rates (24 and 3 coulomb/L/min, the corresponding currents and voltages were 0.4 A, 24 V and 0.05 A, 3 V). This finding is consistent with the report from Lakshmanan et al. [21]. When the electrode was submerged in the SBGW without current, 1.1 mg/L iron was detected after 30 min exposure, indicating that natural corrosion of the electrode contributed negligibly to HFO generation.

Fe  2e ! FeðIIÞ

ð1Þ

Fe  3e ! FeðIIIÞ

ð2Þ

Fe þ O2 ! FeðIIÞ=FeðIIIÞ

ð3Þ

m¼

ItM ZF

ð4Þ

227

L. Li et al. / Separation and Purification Technology 122 (2014) 225–230

It should be kept in mind that the redox reaction on the electrodes strongly depends on the water matrix. For instance, increasing P concentration to 0.01 M in SBGW decreased the iron concentration to less than half of that in the control test. Similarly, the detected iron concentration after the electrolysis was <10% of the calculated value through Faraday’s law when the test water merely contained NaNO3 (0.1 M) rather than all the chemicals used in other tests. This probably indicated some other electron donors (such as OH) outcompeted Fe(0) to be oxidized on the anode during electrolysis. 3.2. Adsorption properties of HFO precipitates The X-ray diffraction pattern of the iron precipitate obtained in SBGW under the current dosage rate of 3 coulomb/L/min (i.e. SBGW-3C/L/min) is presented in Fig. 1. It exhibits two peaks similar to the characteristic peaks of 2-line ferrihydrite (2L-Fh) at d space approximately 0.148 and 0.255 nm. The peaks in this study were broader and the primary peak was shifted to 0.270 nm compared to 0.255 nm for 2L-Fh, which was in good agreement with Voegelin et al. [33] and Carlson and Schwertmann [34], who claimed the high concentration of Si reduced the crystallinity and resulted in the peak drifting. For a much higher charge dosage rate (24 coulomb/L/min) (i.e. SBGW-24C/L/min in Fig. 1) in the EC process and higher Fe(II) concentration in the CC processes (i.e. SGBG-CC(Fe(II) in Fig. 1), X-ray diffraction exhibited a similar pattern, aligned with our wet chemistry analysis that the adsorption capacity of As(V) and P under various charge dosage rates were quite similar (data not shown). This conclusion is consistent with an EXAFS study by van Genuchten et al. [35], which showed that both Fe and As K-edge spectra were similar between samples across a wide range of charge dosage rates (0.06–18 C/L/min). Besides the dosage rate, HFO adsorption properties might depend on water matrix. In Fig. 1, different peak patterns were observed for SBGW without Si. The peak at 0.270 nm was replaced by two peaks at 0.252 and 0.323 nm, and a new distinct peak at 0.197 nm appeared. The altered and sharper peaks indicated that the HFO was more lepidocrocite (Lp)-like, was more crystalline [26] and might has less adsorption capacity. Recent reports indicated that dissolved Si at Si/Fe molar ratios >0.5 would result in the formation of amorphous short-range-ordered hydrous Fe(III)-precipitates [33]. Given that Si concentrations of 19.2 ± 4.7 mg/L have been detected in water from Bangladesh ground water wells [28], both SBGW and real ground water in Bangladesh probably belong to the same category of silicate-rich hydrous ferric oxide (Si-HFO). 0.255nm 0.27nm

0.148nm

Intensity (Arbitrary Units)

0.197nm

2-Line Ferrihydrite

Based on the results presented in this section, HFO in the EC system might be considered to be a uniform adsorbent regardless of the charge dosage rate and water matrix as long as the Si concentration is in similar level.

3.3. Adsorption modeling According to the previous section, it is valid to treat the HFO as a uniform adsorbent under the various conditions in this study, and necessary to develop an adsorption model to quantitatively evaluate the effect of the anions that compete with As for adsorption. As assumed in Eq. (5), the number of adsorption sites („FeAOH) of HFO („FeAOH) should be proportional with Fe(III) concentration and the adsorption site concentration q (site/Fe(III)). The competitive adsorption of the anions is described as a set of fast and reversible equilibrium reactions with the sorption sites as shown in Eq. (6).

BFe  OH ¼ FeðIIIÞ  q

ð5Þ

X þ BFe  OH BFe  X KX

ð6Þ

where X represents As(V), As(III), P or Si and Kx represents the corresponding adsorption constant. The modeled adsorption constants of As(V), As(III), P and Si to Fe(III) precipitate and the adsorption site concentration (q) in SBGW in the EC system are summarized in Table 1 according to the data from Figs. 2 and 3 and Table SI1 (refer SI for detail modeling procedure). As shown, the adsorption site concentration for HFO produced by EC was 0.70 mol/mol iron, quite similar to the adsorption concentration for HFO produced by CC (0.71 mol/mol iron) [14], considering the experimental and fitting errors. In addition, adsorption capacity of precipitate derived from Fe(II) is considerably higher than that of precipitate coming directly from Fe(III) (0.44 mol/mol iron) [14]. However, diffractgram patterns did not show a crystallinity difference (Fig. 1). The difference in adsorption capacity likely ascribed to quick cross-linkage of the small Fe(III) polymeric units to form larger aggregates in which a considerable portion of the adsorption sites were no longer accessible [14], while in the case of Fe(II) or EC, Fe(III) precipitate was gradually formed from the Fe(II) oxidation, giving time for As, P and Si to migrate to the adsorption sites before they aggregated. As(V) removal in SBGW as a function of iron concentration in the EC system is shown in Fig. 2. Approximately 8 mg/L of iron was required to reduce As(V) concentration from 500 to 50 lg/L (local standard of Bangladesh). P removal followed the same trend as As, while Si was hardly removed since its adsorption constant is orders of magnitude less than those of As(V) and P (Table 1). The experimental and modeled data matched very well.

SBGW-3C/L/min SBGW-24C/L/min 0.33nm

0.252nm

SBGW-CC(Fe(II)) SBGW-NO Si SBGW-CC(Fe(III)) SBGW-Coagulation with Alum

0.40

Table 1 Fitted adsorption constants and sorption site concentrations in EC and CC systems.

0.20

0.10

0.05

D-spacings (nm) Fig. 1. Background-subtracted diffraction data for precipitates generated in various chemical conditions and operating parameters. For all patterns, the intensity of the highest peak is normalized to 1.

Log(KAs(V)) Log(KAs(III)) Log(KSi) Log(KP) [„Fe]/[Fe0]

ECa

Fe(II)b

Fe(III)b

5.42 ± 0.07 4.10 ± 0.04 3.04 ± 0.01 6.48 ± 0.06 0.70 ± 0.03

5.72 ± 0.23 3.67 ± 0.11 2.77 ± 0.08 5.87 ± 0.15 0.71 ± 0.05

5.97 ± 0.12 4.47 ± 0.07 3.36 ± 0.11 6.07 ± 0.04 0.44 ± 0.02

a Data were subdivided into 2 independent sets each by taking every second data point through the successive data points shown in Figs. 2 and 3, Fig. 4(a) and Table SI1. Similar optimized values were obtained, indicating the model were well constrained. b Data adopted from Roberts et al. [14].

228

L. Li et al. / Separation and Purification Technology 122 (2014) 225–230

required to reduce As(III) from 500 to 50 lg/L in the EC system (pH 7.1 ± 0.1, charge dosage rate 3 coulomb/L/min) was 32 mg/ L, considerably below that in a Fe(II) CC system (50 mg/L) under the same ions background [14]. This improvement greatly reduces the iron scrap dosage and the cost of iron sludge disposal, as well as DO consumption caused by Fe(II) oxidation would therefore guarantee the oxidizing condition in the system. Simultaneous with As(III) oxidation, total As(V) concentration greatly increased in the initial stage and leveled off to 335 lg/L as illustrated in Fig. 3. The corresponding aqueous As(V) increased with the same trend when iron dosage was below 3 mg/L, indicating minimum As(V) was adsorbed at this stage. It might be due to the presence of a much greater amount of P. Afterwards, aqueous As(V) concentration decreased and leveled off to 0 due to the complete removal of P.

1.2

Remaining fraction

1

0.8 As(V) P

0.6

Si 0.4

0.2

0 0

2

4

6

8

10

12

Iron (mg/L) Fig. 2. Removal of dissolved As(V), silicate, and phosphate as a function of iron concentration. The lines represent the output from the model calculation. Initial concentrations: 500 lg/L As(V), 3.0 mg/L P, and 30 mg/L Si; pH 7.1 ± 0.1.

ð7Þ

FeðIVÞ þ FeðIIÞ ! 2FeðIIIÞ K1

ð8Þ

FeðIVÞ þ AsðIIIÞ ! AsðVÞ þ FeðIIIÞ K2

ð9Þ

FeðIVÞAsðIIIÞ ¼ K2 ½AsðIIIÞ=ðK2 ½AsðIIIÞ þ K1 ½FeðIIÞÞ

ð10Þ

FeðIVÞAsðIIIÞ ¼ 1=ðK1 ½FeðIIÞ=K2 ½As½III þ 1Þ

ð11Þ

Total As(III)

1

Total As(V) Aqueous As(V)

0.8

Remaining fraction

3FeðIIÞ þ O2 ! . . . ! FeðIVÞ

Aqueous P

3.5. Effects of Ca2+ and Mg2+

0.6

In SBGW, Ca2+/Mg2+ were observed to have great impacts on precipitate formation and As removal (Fig. 4). Without Ca2+/Mg2+

0.4

0.2

0 0

5

10

15

20

25

30

35

Iron (mg/L) Fig. 3. Removal of As, P, and Si as a function of iron concentration. Initial concentrations: 500 lg/L As(III), 3.0 mg/L P, and 30 mg/L Si; pH 7.1 ± 0.1.

3.4. As(III) oxidation One of the recognized advantages of the EC process for As(III) removal is the As(III) oxidation during the simultaneous oxidation of Fe(II) by DO [24]. As(III) oxidation and adsorption as well as the removal of other co-existing anions in the EC system were investigated in detail in this study, allowing for a deeper understanding of how these ions behave during EC. To examine the intermediate oxidant generated in EC system, 2-propanol in excess (14 mM) on As(III) oxidation was found negligible in this study, which is consistent with the As(III) oxidant being an Fe(IV) species generated as an intermediate during the oxidation of Fe(II) in SBGW [36]. As a result, Fe(IV) was supposed as the plausible intermediate product of oxidizing Fe(II) with DO as briefly expressed in Eq. (7). Then, Fe(IV) would be consumed by Fe(II) and As(III) according to Eqs. (8) and (9). K1 and K2 are the reaction rates of Fe (IV) with Fe (II) and As (III), respectively. Therefore, the amount of Fe(IV) consumed by As(III) (Fe(IV)As(III)) can be expressed by Eqs. (10) and (11). Accordingly, the mole ratio of As(III) to Fe(II) (RAs(III)/Fe(II)) was the controlling factor for As(III) oxidation efficiency [23]. Fortunately, Fe(II) is generated gradually during EC as opposed to the one step Fe(II) dosage in CC. Gradual release of Fe(II) would result in a high RAs(III)/Fe(II) value and therefore enhance As(III) oxidation. As a result, the amount of iron

Fig. 4. Effect of Ca2+/Mg2+ on precipitate formation, As removal ((a) with and (b) without Ca2+/Mg2+); Initial concentrations: 500 lg/L As(III), 3.0 mg/L P, and 30 mg/L Si; pH 7.1 ± 0.1.

229

L. Li et al. / Separation and Purification Technology 122 (2014) 225–230 Table 2 Effect of Ca2+/Mg2+ on the HFO particle size and zeta potential (initial concentrations: 500 lg/L As(III), 3.0 mg/L P, and 30 mg/L Si with and without Ca2+/Mg2+). Iron concentration (mg/L)

5.2 10.1 18.2

Without Ca2+/Mg2+

Iron concentration (mg/L)

Zeta potential (mV)

Average particle size (nm)

28.8 26.0 28.4

15.3 48.2 123.5

4.8 9.6 19.3

With Ca2+/Mg2+ Zeta potential (mV)

Average particle size (nm)

8.9 9.3 11.6

884 906 964

the remaining HA was under 10% of the initial concentration, implying different removal pathways for HA and As in this system. Moreover, the turbidity of HFO and its settling properties were not significantly affected by the addition of HA. The lack of an effect of HA finding in this research is inconsistent with previous reports [15,37]. This is probably due to the co-existing Ca2+/Mg2+ in ground water neutralizing the surface charge of the Fe(III)-HA complex [25] and therefore minimizing the effect of HA on HFO properties and As removal. Further study is warranted to make discussion in detail.

4. Jar tests Fig. 5. Remaining turbidity as a function of settling time and Al3+ concentration (subject to 60 min of mixing before the jar test).

(using equivalent NaCl to keep the same iron strength of the solution), no higher than 13%, 15% and 19% of iron, As and P respectively were filterable by a 0.45 lm membrane when the iron concentration were below 17 mg/L (Fig. 4(b)). With the present of Ca2+/Mg2+ (Fig. 4(a)), on other hand, nearly 100% of iron was filterable, leading to much greater removal rate of P and As. This may be mainly ascribed to the neutralization of surface charge of HFO by Ca2+/Mg2+, which facilitates its aggregation. As shown in Table 2, the adding of Ca2+/Mg2+ greatly increased the zeta potential of HFO from 30 mV to 10 mV, and enhanced HFO aggregation with the particle size increased from 15.3, 48.2 and 123.5 to 884.4, 905.6 and 963.6 nm accordingly, which is consistent with previous report [25]. Of note, the portion of filterable iron dramatically increased when the iron concentration was over 18 mg/L, which probably due to the increasing of HFO particles size caused by a higher particles density. In addition, the Ca2+/Mg2+ in HFO precipitate (filterable Ca2+/Mg2+) accounted for 10% of the total amount of Ca2+/Mg2+, indicating that 0.40 mM Ca2+/Mg2+ might be sufficient for HFO surface charge neutralization. Fortunately, most of the ground water found in Bangladesh contains sufficient levels of these cations [3].

Although the Fe(II) has been shown to have higher adsorption capacity than Fe(III) (Table 2), the precipitate attained from Fe(II) was harder to settle than that attained from Fe(III) according to our tests and the literature [14]. Jar tests were therefore conducted to optimize operation parameters to enhance HFO settling and sludge separation.Subsequent to electrolysis, the solution was mixed for 60 min to ensure the complete oxidation of Fe(II) (final Fe(II)/Fe(III) ratio was <3%). Of note, to simulate field conditions, pH was not controlled, and increased to 7.7 during the mixing. Results in Fig. 5 showed that the turbidity would decrease from 140 to 54 NTU after 30 min of settling without coagulant, while the turbidity removal was greatly enhanced with the addition of Al3+ at concentrations of 1, 2, 3, 5, and 8 mg/L. The corresponding terminal turbidies were 1.2, 0.67, 0.91, 1.2 and 1.5 NTU after 120 min of settling. In other words, 2 mg/L Al3+ was sufficient to reduce turbidity to below 1 NTU after the clarification process, implying no need for further filtration. In addition, the corresponding remaining arsenic concentrations of the supernatants for the sets were 54.2, 49.9, 52.1, 61.3 and 54.7 lg/L, which were in agreement with the directly filtrated sample (49.2 lg/L), indicating that enhanced aggregation or coagulation by Al3+ has no obvious effect on the adsorption capacity of HFO. In agreement with the wet chemical analysis, there were no obvious differences between the X-ray diffraction patterns (Fig. 1) of the HFO before and after coagulation.

3.6. Effect of natural organic matter (NOM) 5. Conclusions and implications NOM, another common compound in shallow ground water or contaminated deep ground water, has been reported to interfere with As removal by competing for adsorption sites [27,37], increasing time required to reach equilibrium [27], reducing the settling ability of the precipitate [15], and potentially interfering with As(III) oxidation by interacting with reactive oxygen species. As previously reported research of NOM effects on As removal were conducted under various conditions and with different water matrices, which may be different from As removal under SBGW background in the EC system, this research also investigated the effects of NOM. Humic acid (HA) was added as the surrogate to the water matrix to evaluate the overall effect of NOM on As(III) oxidation and adsorption (Fig. SI3). For HA concentrations of 0, 5 and 15 mg/L the remaining respective As(III) concentrations were 150, 137 and 125 lg/L total As(III), 47, 45, and 45 lg/L. In addition,

In this study, many of the uninvestigated engineering and scientific aspects of As(III) removal by EC were comprehensively studied. Our results indicate that As(III) removal efficiency was considerably higher than with the widely used CC process and other iron based processes. Iron scrap, an abundant byproduct from iron planing machines, exhibited the same capacity as iron plates for As removal during electrolysis, further reducing the material cost of EC. Settle-ability of HFO precipitates can be greatly enhanced by the addition of 2 mg/L Al3+, which reduced the turbidity of the supernatant to <1 NTU (drinking water standard for many countries or regions) after 120 min settling. These conclusions offer new scientific and engineering insights for practitioners and other researchers seeking better and more efficient treatment of As-contaminated ground water.

230

L. Li et al. / Separation and Purification Technology 122 (2014) 225–230

Acknowledgments We gratefully acknowledge support for this work by The Richard C. Blum Center for Developing Economies, USEPA P3 Phase II award, support from China Postdoctoral Science Foundation (2013M531215), State Key Laboratory of Pollution Control and Resource Reuse Foundation (PCRRY12002), and China National Major Science and Technology Project of China (No. 2012ZX07403-001). We also thank Mr. John Erickson from UC Berkeley for editing the English and anonymous reviewers and editors for their efforts on this paper. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.seppur.2013.11.012.

[17]

[18]

[19] [20] [21] [22]

[23]

[24]

[25]

References [1] M. Karim, Arsenic in groundwater and health problems in Bangladesh, Water Res. 34 (2000) 304–310. [2] N. Singh, D. Kumar, A.P. Sahu, Arsenic in the environment: effects on human health and possible prevention, J. Environ. Biol. 28 (2007) 359–365. [3] D.G. Kinniburgh, P.L. Smedley, Arsenic contamination of ground water in Bangadesh. Final Report Summary, British Geological Survey, Bangladesh Department for Public Health, Engineering, 2001. [4] M. Argos, T. Kalra, P.J. Rathouz, Y. Chen, B. Pierce, F. Parvez, T. Islam, A. Ahmed, M. Rakibuz-Zaman, R. Hasan, G. Sarwar, V. Slavkovich, A. van Geen, J. Graziano, H. Ahsan, Arsenic exposure from drinking water, and all-cause and chronicdisease mortalities in Bangladesh (HEALS): a prospective cohort study, Lancet 376 (2010) 252–258. [5] Who, Guidelines for drinking-water quality, second ed., vol. 1. Recommendations, WHO {a}, Geneva, Switzerland, 1993. [6] USEPA, Arsenic treatment technologies for soil, waste and water, Tech. Report No. EPA-542-R-02-004, 2002. [7] CMOH, China Ministry of Health (CMOH), Standards for Drinking Water Quality-GB5749-2006, 2006 (in Chinese). [8] K. Ng, Z. Ujang, P. Le-Clech, Arsenic removal technologies for drinking water treatment, Rev. Environ. Sci. Biotechnol. 3 (2004) 43–53. [9] M. Bissen, F.H. Frimmel, Arsenic – a review. Part II: Oxidation of arsenic and its removal in water treatment, Acta Hydrochim. Hydrobiol. 31 (2003) 97–107. [10] M. Bissen, F.H. Frimmel, Arsenic – a review. – Part 1: Occurrence, toxicity, speciation, mobility, Acta Hydrochim. Hydrobiol. 31 (2003) 9–18. [11] X.G. Meng, G.P. Korfiatis, S.B. Bang, K.W. Bang, Combined effects of anions on arsenic removal by iron hydroxides, Toxicol. Lett. 133 (2002) 103–111. [12] X.G. Meng, S. Bang, G.P. Korfiatis, Effects of silicate, sulfate, and carbonate on arsenic removal by ferric chloride, Water Res. 34 (2000) 1255–1261. [13] A.B. Mukherjee, P. Bhattacharya, Arsenic in groundwater in the Bengal Delta Plain: slow poisoning in Bangladesh, Environ. Rev. 9 (2001) 189–220. [14] L.C. Roberts, S.J. Hug, T. Ruettimann, M. Billah, A.W. Khan, M.T. Rahman, Arsenic removal with iron(II) and iron(III) waters with high silicate and phosphate concentrations, Environ. Sci. Technol. 38 (2004) 307–315. [15] X.H. Guan, H.R. Dong, J. Ma, L. Jiang, Removal of arsenic from water: effects of competing anions on As(III) removal in KMnO4–Fe(II) process, Water Res. 43 (2009) 3891–3899. [16] S.-Y. Pang, J. Jiang, J. Ma, Oxidation of sulfoxides and Arsenic(III) in corrosion of nanoscale zero valent iron by oxygen: evidence against ferryl ions (Fe(IV)) as

[26]

[27]

[28]

[29] [30]

[31] [32]

[33]

[34]

[35]

[36]

[37]

active intermediates in Fenton reaction, Environ. Sci. Technol. 45 (2010) 307– 312. A.K. Leuz, S.J. Hug, B. Wehrli, C.A. Johnson, Iron-mediated oxidation of antimony(III) by oxygen and hydrogen peroxide compared to arsenic(III) oxidation, Environ. Sci. Technol. 40 (2006) 2565–2571. I.A. Katsoyiannis, T. Ruettimann, S.J. Hug, PH dependence of Fenton reagent generation and As(III) oxidation and removal by corrosion of zero valent iron in aerated water, Environ. Sci. Technol. 42 (2008) 7424–7430. P.K. Holt, G.W. Barton, C.A. Mitchell, The future for electrocoagulation as a localised water treatment technology, Chemosphere 59 (2005) 355–367. P.R. Kumar, S. Chaudhari, K.C. Khilar, S.P. Mahajan, Removal of arsenic from water by electrocoagulation, Chemosphere 55 (2004) 1245–1252. D. Lakshmanan, D.A. Clifford, G. Samanta, Ferrous and ferric ion generation during iron electrocoagulation, Environ. Sci. Technol. 43 (2009) 3853–3859. X. Zhao, B.F. Zhang, H.J. Liu, J.H. Qu, Removal of arsenite by simultaneous electro-oxidation and electro-coagulation process, J. Hazard. Mater. 184 (2010) 472–476. L. Li, C.M. van Genuchten, S.E.A. Addy, J. Yao, N. Gao, A.J. Gadgil, Modeling As(III) oxidation and removal with iron electrocoagulation in groundwater, Environ. Sci. Technol. 46 (2012) 12038–12045. W. Wan, T.J. Pepping, T. Banerji, S. Chaudhari, D.E. Giammar, Effects of water chemistry on arsenic removal from drinking water by electrocoagulation, Water Res. 45 (2011) 384–392. X.H. Guan, J. Ma, H.R. Dong, L. Jiang, Removal of arsenic from water: effect of calcium ions on As(III) removal in the KMnO4–Fe(II) process, Water Res. 43 (2009) 5119–5128. R. Kaegi, A. Voegelin, D. Folini, S.J. Hug, Effect of phosphate, silicate, and Ca on the morphology, structure and elemental composition of Fe(III)-precipitates formed in aerated Fe(II) and As(III) containing water, Geochim. Cosmochim. Acta 74 (2010) 5798–5816. A.D. Redman, D.L. Macalady, D. Ahmann, Natural organic matter affects arsenic speciation and sorption onto hematite, Environ. Sci. Technol. 36 (2002) 2889– 2896. D.G. Kinniburgh, P.L. Smedley, Arsenic contamination of ground water in Bangadesh.Final Report Summary, British Geological Survey, Bangladesh Department for Public Health Engineering, 2001. B.M. Voelker, B. Sulzberger, Effects of fulvic acid on Fe(II) oxidation by hydrogen peroxide, Environ. Sci. Technol. 30 (1996) 1106–1114. A.P. Hammersley, S.O. Svensson, M. Hanfland, A.N. Fitch, D. Hausermann, Twodimensional detector software: from real detector to idealised image or twotheta scan, High Pressure Res. 14 (1996) 235–248. U. Schwertmann, R.M. Cornell, Iron oxides in the laboratory preparation and characterization, 1991. D. Lakshmanan, D.A. Clifford, G. Samanta, Comparative study of arsenic removal by iron using electrocoagulation and chemical coagulation, Water Res. (Oxford) 44 (2010) 5641–5652. A. Voegelin, R. Kaegi, J. Frommer, D. Vantelon, S.J. Hug, Effect of phosphate, silicate, and Ca on Fe(III)-precipitates formed in aerated Fe(II)- and As(III)containing water studied by X-ray absorption spectroscopy, Geochim. Cosmochim. Acta 74 (2010) 164–186. L. Carlson, U. Schwertmann, Natural ferrihydrites in surface deposits from Finland and their association with silica, Geochim. Cosmochim. Acta 45 (1981) 421–429. C.M. van Genuchten, S.E.A. Addy, J. Pena, A.J. Gadgil, Removing arsenic from synthetic groundwater with iron electrocoagulation: an Fe and As K-Edge EXAFS study, Environ. Sci. Technol. 46 (2012) 986–994. S.J. Hug, O. Leupin, Iron-catalyzed oxidation of arsenic(III) by oxygen and by hydrogen peroxide: pH-dependent formation of oxidants in the Fenton reaction, Environ. Sci. Technol. 37 (2003) 2734–2742. V. Pallier, G. Feuillade-Cathalifaud, B. Serpaud, Influence of organic matter on arsenic removal by continuous flow electrocoagulation treatment of weakly mineralized waters, Chemosphere 83 (2011) 21–28.