Arsenic removal in a sulfidogenic fixed-bed column bioreactor

Journal of Hazardous Materials 269 (2014) 31–37

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Arsenic removal in a sulfidogenic fixed-bed column bioreactor Muslum Altun a,∗ , Erkan Sahinkaya b , Ilknur Durukan a , Sema Bektas a , Kostas Komnitsas c a

Hacettepe University, Department of Chemistry, Beytepe, Ankara, Turkey Istanbul Medeniyet University, Bioengineering Department, Goztepe, Istanbul, Turkey c Technical University of Crete, Department of Mineral Resources Engineering, Chania, Greece b

h i g h l i g h t s • • • •

Sulfidogenic treatment of As-containing AMD was investigated. High rate simultaneous removal of As and Fe was achieved. As was removed without adding alkalinity or adjusting pH. As and Fe removal mechanisms were elucidated.

a r t i c l e

i n f o

a b s t r a c t

Article history: Received 24 August 2013 Received in revised form 13 November 2013 Accepted 21 November 2013 Available online 27 November 2013

In the present study, the bioremoval of arsenic from synthetic acidic wastewater containing arsenate (As5+ ) (0.5–20 mg/L), ferrous iron (Fe2+ ) (100–200 mg/L) and sulfate (2000 mg/L) was investigated in an ethanol fed (780–1560 mg/L chemical oxygen demand (COD)) anaerobic up-flow fixed bed column bioreactor at constant hydraulic retention time (HRT) of 9.6 h. Arsenic removal efficiency was low and averaged 8% in case iron was not supplemented to the synthetic wastewater. Neutral to slightly alkaline pH and high sulfide concentration in the bioreactor retarded the precipitation of arsenic. Addition of 100 mg/L Fe2+ increased arsenic removal efficiency to 63%. Further increase of influent Fe2+ concentration to 200 mg/L improved arsenic removal to 85%. Decrease of influent COD concentration to its half, 780 mg/L, resulted in further increase of As removal to 96% when Fe2+ and As5+ concentrations remained at 200 mg/L and 20 mg/L, respectively. As a result of the sulfidogenic activity in the bioreactor the effluent pH and alkalinity concentration averaged 7.4 ± 0.2 and 1736 ± 239 mg CaCO3 /L respectively. Electron flow from ethanol to sulfate averaged 72 ± 10%. X-ray diffraction (XRD), X-ray fluorescence (XRF), scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) analyses were carried out to identify the nature of the precipitate generated by sulfate reducing bacteria (SRB) activity. Precipitation of arsenic in the form of As2 S3 (orpiment) and co-precipitation with ferrous sulfide (FeS), pyrite (FeS2 ) or arsenopyrite (FeAsS) were the main arsenic removal mechanisms. © 2013 Elsevier B.V. All rights reserved.

Keywords: Acid mine drainage Sulfate reduction Arsenic removal

1. Introduction During the processing of gold and other metal ores, arsenic disengagement occurs, due to the oxidation of arsenic bearing minerals [1–3]. Arsenopyrite-bearing sulfide ores and tailings may also be oxidized in a similar way releasing As and sulfate according to the following reaction (R1) [4,5]. FeAsS + 3.5O2 + H2 O → Fe

3+

+ SO4

2−

+ H2 AsO4

−

(R1)

release of As species. The biological iron oxidation at low pH and the chemical dissolution of arsenopyrite are shown in the following reactions (R2) and (R3) [6,7]. Some of the Fe(III) may react with the dissolved arsenate resulting in the precipitation as scorodite, FeAsO4 ·2H2 O (reaction (R4)) [5]. 2Fe2+ + 0.5O2 + 2H+

A. ferrooxidans

−→

2Fe3+ + H2 O

(R2)

FeAsS + 13Fe3+ + 8H2 O → 14Fe2+ + SO4 2− + 13H+ + H3 AsO4 (R3)

The presence of iron oxidizing bacteria, such as Acidithiobacillus ferrooxidans, accelerates the rate of arsenopyrite oxidation and the

H3 AsO4 + Fe3+ + 2H2 O → FeAsO4 ·2H2 O + 3H+

∗ Corresponding author at: Hacettepe University, Department of Chemistry, Beytepe, Ankara, Turkey. Tel.: +90 4143470820. E-mail address: [email protected] (M. Altun).

It has been reported that arsenic concentration ranges between 100 ␮g/L and 5000 ␮g/L in acidic leachates generated in areas where mining activities are carried out, while it normally resides between 1 ␮g/L and 10 ␮g/L in uncontaminated natural water.

0304-3894/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2013.11.047

(R4)

32

M. Altun et al. / Journal of Hazardous Materials 269 (2014) 31–37

Although fairly high levels of arsenic concentrations are observed in acid mine drainage (AMD), the highest value reported is 72 mg/L at Zimbabwe Duke mining area [8]. Arsenic contamination of ground- and surface waters due to acid mine drainage (AMD) formation has been reported in many countries, such as Japan, Spain, India, Bangladesh, China, Chile, Argentina, Mexico, Taiwan, Vietnam, United States, and Turkey [5,9]. Due to its potential for combined removal of acidity, metals and sulfate, biological sulfate-reduction appears to be the most promising AMD treatment and metals recovery method. The process is based on biological hydrogen sulfide and alkalinity production by SRB (reaction (R5)): 2CH2 O + SO4 2− → H2 S + 2HCO3 −

(R5)

where organic matter (CH2 O) represents the electron donor. The biogenic hydrogen sulfide results in the precipitation of dissolved metals as low solubility sulfides, as indicated in reaction (R6): H2 S + M2+ → MS(s) + 2H+

(R6)

where M2+ denotes metal, such as Zn2+ , Cu2+ , Ni2+ , Co2+ or Fe2+ . Although there are several studies on sulfidogenic AMD treatment, very few studies are available in literature on sulfidogenic arsenic treatment. In the study of Battaglia-Brunet et al. [10] arsenic removal was investigated in a fixed bed sulfidogenic bioreactor in which glycerin or hydrogen gas, as electron sources, and 100 mg/L of As(V) were fed. Results showed that if the reactor is fed with glycerin, very low sulfate removal rates are obtained at pH 5 and the produced sulfide is just sufficient to remove arsenic as As2 S3 . However, when hydrogen gas was introduced in the reactor, sulfide concentration increased and resulted in dissolution of the precipitated As2 S3 . It is known that As2 S3 may dissolve at high pH and hydrogen sulfide concentration according to reaction (R7): 3/2As2 S3 (amorphous) + 3/2H2 S → H2 As3 S6 − + H+ log K = −5.0

(R7)

However, the co-presence of Fe and As in AMD may lead to the formation of arsenopyrite (FeAsS) and the removal efficiency of As in a sulfidogenic bioreactor may increase and become possible even at neutral pHs and high concentrations of sulfide [11]. It is also known that the generated FeS may precipitate during sulfidogenic AMD treatment and also adsorb arsenic [12]. Therefore, further studies are required for the sulfidogenic treatment of As containing AMD using highly efficient bioreactors in order to protect drinking water contamination from As. This study aims at investigating As removal from AMD in a sulfidogenic continuously fed fixed-bed bioreactor. The performance of the bioreactor was investigated in the presence or absence of Fe under varying operating conditions. 2. Materials and methods

Fig. 1. A laboratory scale fixed bed up-flow anaerobic glass bioreactor with dimensions of 5 cm (diameter) × 30 (length).

50 mL sludge containing active SRB obtained from a sulfate reducing anaerobic baffled reactor [13]. The reactor was covered with aluminum foil to prevent phototrophic bacterial activity. The active bed volume was considered for the calculation of HRT. Throughout the study, synthetic solution was fed to the bioreactor using a peristaltic pump at a flow rate of 1 L/day corresponding to 9.6 h HRT. The reactor was operated in a temperature controlled room at 30–32 ◦ C and the feed container was refrigerated (4 ◦ C) prior to use to prevent bacterial growth. 3. Experimental The bioreactor operated for 245 days under eight separate operating periods (Table 1) using a synthetic feed containing 1.480 g/L Na2 SO4 , 2.563 g/L MgSO4 ·7H2 O, 56 mg/L KH2 PO4 , 111 mg/L NH4 CI, 11 mg/L ascorbic acid and ethanol as carbon and electron source (1560 mg COD/L). In the first period (0–66 days), As(V) and Fe(II) free influent was fed to the bioreactor in order to enrich the ethanol oxidizing SRB. In the second period, As(V) was supplemented to the synthetic feed and ascorbic acid was excluded from the influent solution in order to prevent As(V) reduction. Stock solution of 1000 mg/L As(V) was prepared using Na2 HAsO4 ·7H2 O in deionized water. In the periods 2–6, As(V) concentration in the synthetic feed was gradually increased from 0.5 mg/L to 20 mg/L, while the influent pH was kept constant at 4. In the periods 7–9, FeSO4 ·7H2 O supplemented to the synthetic feed in order to evaluate the impact of Fe presence on As removal under sulfidogenic conditions. In the periods 7 and 8, Fe(II) concentration in the influent was 100 and 200 mg/L, respectively. In these periods the influent pH was decreased to 3.5 using HCl (Table 1). Throughout the study, influent sulfate concentration and HRT were kept constant at 2000 mg/L and 9.6 h, respectively. Influent and effluent of the bioreactor were sampled once and three times a week, respectively, for sulfate, dissolved sulfide (only in effluent), alkalinity, COD, pH and total As and Fe measurements. All chemicals were purchased from Merck, Germany.

2.1. Bioreactor 3.1. Analytical methods A laboratory scale glass column with dimensions of 5 cm (diameter) × 30 cm (length) was used as a fixed bed up-flow anaerobic bioreactor (Fig. 1). The column was packed with commercially available sand (particle diameter 1–1.5 mm) as biomass attachment medium (400 mL). The sand was washed with 10% nitric acid followed by rinsing with deionized water to eliminate possible contamination with organic matter which could be attached on the surface of the particles. The bioreactor was inoculated with

Prior to sulfate, dissolved sulfide, COD and total As and Fe measurements, samples were centrifuged at 3000 × g for 10 min (HettichRotofix 32) and then filtered using syringe filters (0.45 ␮m). Sulfate concentration was measured using a turbidimetric method [14]. Total dissolved sulfide concentration was measured using a spectrophotometric method [15]. COD was measured using a micro digestion and subsequent titration method according to APHA and

M. Altun et al. / Journal of Hazardous Materials 269 (2014) 31–37

33

Table 1 Operating conditions in the bioreactor. Periods

Days COD (mg/L) Influent As(V) (mg/L) Influent Fe(II) (mg/L) Influent pH

I

II

III

IV

V

VI

VII

VIII

IX

0–66 1560 0 0 4 ± 0.1

67–82 1560 0.5 0 4 ± 0.1

83–97 1560 1–25 0 4 ± 0.1

98–125 1560 5.0 0 4 ± 0.1

126–156 1560 10 0 4 ± 0.1

157–175 1560 20 0 4 ± 0.1

176–200 1560 20 100 3.5 ± 0.1

201–226 1560 20 200 3.5 ± 0.1

227–245 780 20 200 5 ± 0.1

WEF standard methods [14]. Prior to COD measurements, samples were acidified with concentrated H2 SO4 and then purged with nitrogen gas for 5 min in order to remove dissolved sulfide. Alkalinity was measured on unfiltered samples by decreasing pH to 4.5 with 0.1 N HCI [14]. For total Fe measurements, samples were first acidified approximately to pH 2.0 with concentrated HCI to solubilize Fe precipitates. Then, samples were filtered through 0.45 ␮m to remove biomass and other particles. Arsenic measurements were carried out after direct filtration through 0.45 ␮m. Total metal analysis in both influent and effluent were carried out by using atomic absorption spectrometry (AAS). A Perkin-Elmer Model Analyst 800 Atomic Absorption Spectrophotometer equipped with flame or electrothermal atomizer was used. Electron flow from ethanol oxidation to sulfate reduction was calculated according to Eq. (1), in which biomass growth was ignored. 0.67(SO4,o − SO4,e ) % Electron flow = 100 CODo − CODe

(1)

where SO4,o and SO4,e are influent and effluent sulfate concentrations (mg/L), CODo and CODe are influent and effluent COD concentrations (mg/L), respectively. A mass balance equation was used (Eq. (2)) to calculate biogenic sulfide recovery from sulfate reduction, assuming that 3 mmol of sulfide react with 2 mmol of As(III) and generate As2 S3 . % Sulfide recovery = 100

agarose gel electrophoresis and staining with ethidium bromide prior to DGGE analysis. DGGE was performed with the D-CODE System (BioRAD, The Netherlands). The electrophoresis was carried out at 60 ◦ C with 100 V for 16 h. After electrophoresis, the gel was stained in an SYBR Gold solution (100 ␮L/L inTAE) for 30 min and photographed on Vilber Lourmat Quantum St4 gel documentation system. Bands in DGGE gels were excised with a razor blade and placed in sterile 200 ␮L vials. DNA was eluted into 20 ␮L of water and stored overnight at 4 ◦ C. The eluted DNA was used as template in PCRs with the primers BacV3f (without GC clamp) and 907r using the same PCR program as described above. The sequencing of the purified products was performed at REFGEN (Ankara, Turkey). 4. Results and discussion 4.1. Sulfate reduction and COD oxidation performance The evolution of sulfate, dissolved sulfide and COD concentrations versus time in the effluent is presented in Fig. 2. In the first period of the reactor operation (0–66 days), sulfide concentration increased almost linearly while sulfate and COD concentrations decreased due to the enrichment and activity of sulfate reducing bacteria. Then, steady-state conditions were reached and the sulfate, COD, and sulfide concentrations between days 50 and 66

measured sulfide (mmol) + sulfide reacted for metal precipitation (mmol) reduced sulfate (mmol)

All analyses were carried out in duplicate.

3.3. DGGE analysis DNA was extracted with PowerSoil DNA kit (MoBio). Extracted DNA samples were stored at −20 ◦ C and were used as template for polymerase chain reaction (PCR) without further treatment. Fragments corresponding to nucleotide positions 341–926 of the Escherichia coli 16S rRNA gene sequence were amplified with the forward primer GC-BacV3f (5 -CCT ACG GGA GGC AGC AG-3 ) [16]. PCR amplification was performed using a Thermocycler T3000 (TECHNE). The presence of PCR products was confirmed by 1% (w/v)

Sulfate (mg/L)

2000

Influent Effluent

1500 1000 500 0 1600

COD (mg/L)

Dried sludge containing support material and metal precipitates collected from the fixed bed up-flow anaerobic bioreactor was subjected to mineralogical analyses. XRF analysis was performed by a Bruker-AXS type S2Range energy dispersive spectrometer. XRD analysis was performed by a Bruker D8 Advance diffractometer using a Cu tube and a scanning range from 3◦ to 70◦ 2 with a step of 0.03◦ and 4 s/step measuring time. Qualitative analysis was carried out using the Diffracplus Software (Bruker AXS) and the PDF database. SEM and EDS studies were also carried out for precipitates. Secondary and back-scattered electron images were obtained using a JEOL-6380LV SEM (Tokyo, Japan).

2500

Influent Effluent

1200 800 400 0

Sulfide (mg/L)

3.2. Mineralogical analysis

(2)

600 400 200 0 0

50

100

150

200

250

Day Fig. 2. Evolution of sulfate, COD and dissolved sulfide concentration vs. time in the bioreactor effluent.

M. Altun et al. / Journal of Hazardous Materials 269 (2014) 31–37

120 100

pH

80 60 40 Electron flow to sulfate Sulfide recovery

20 0

0

50

100

150

200

250

Day Fig. 3. Profiles of electron flow to sulfate and % sulfide recovery.

averaged 379 ± 72 mg/L, 72 ± 13 mg/L, and 566 ± 71 mg/L, respectively. The corresponding sulfate and COD removal rates were 4 g/(L d) and 3.72 g/(L d), respectively, and are comparable with literature data. Sahinkaya et al. [17] observed maximum sulfate reduction rate of 4.6 g/(L d) in a fluidized bed reactor (FBR) fed with sulfate (2.5 g/L) and ethanol (COD/sulfate = 0.85) at an HRT of 12 h. Kaksonen et al. [18] reported maximum sulfate reduction rate in an ethanol-fed FBR treating synthetic AMD of about 4 g/(L d). In another study, Sahinkaya and Yucesoy [19] reported maximum sulfate reduction rate of about 3 g/(L d) in an ethanol-fed anaerobic baffled reactor treating synthetic AMD containing Cu and Zn. In periods 2–6, the influent As concentration increased steadily from 0.5 mg/L to 20 mg/L. The presence of As in the influent did not adversely affect reactor performance while the effluent sulfate, COD and sulfide concentrations averaged 402 ± 114 mg/L, 85 ± 30 mg/L, and 496 ± 88 mg/L, respectively. The corresponding removal percentages of sulfate and COD were around 80% and 95%, respectively. In periods 7 and 8, the influent As concentration was kept constant at 20 mg/L and Fe2+ was supplemented to the synthetic feed at 100 mg/L and 200 mg/L, respectively. Addition of Fe to the synthetic feed did not adversely affect the system performance and the effluent sulfate, COD and sulfide concentrations averaged 370 ± 179 mg/L, 140 ± 69 mg/L and 475 ± 74 mg/L (Fig. 2), respectively. In the last period, influent COD concentration was decreased to its half (780 mg/L) to decrease the sulfide concentration in the reactor while the influent Fe2+ concentration remained at 200 mg/L. According to reaction (R7), high sulfide concentration may solubilize As2 S3 while low sulfide concentration may increase As removal efficiency. After decreasing the feed COD concentration to 780 mg/L, the effluent sulfate, COD and sulfide concentrations averaged 1527 ± 47 mg/L, 64 ± 18 mg/L and 46 ± 35 mg/L, respectively. The decrease of influent COD concentration resulted in appreciable decrease in sulfate reduction efficiency and sulfide production as it was expected since the system contained low carbon and electrons. Electrons generated by ethanol oxidation are used in several reactions, such as sulfate reduction, biomass growth, and methane generation. The ratio of the electrons used for sulfate reduction is depicted in Fig. 3. In the first 50 days, the percentage of electrons utilized for sulfate reduction was around 40% and then increased sharply to around 70%. The reason for this observation is the probable inhibition of non-SRB due to their exposure to high sulfide concentrations for a long period (50 days), which increased carbon availability for sulfate reduction. Until the last period in which ethanol concentration was decreased to its half, electron flow to sulfate reduction was stable and averaged 72 ± 10%. Previous studies [18,20] have shown that, depending on reactor configuration and operating conditions, 0.05–0.15 mg of volatile suspended solids (VSS) are produced per mg of reduced sulfate. Similarly, Bayrakdar et al. [21] and Sahinkaya and Yucesoy [19] reported that in a sulfidogenic ABR the percent of electron flow to sulfate reduction was

Alkalinity (mg CaCO3/L)

Electron Flow to Sulfate or Sulfide Recovery, %

34

9 8 7 6 5 4 3

Influent Effluent

2000 1500 1000 500 0

0

50

100

150

200

250

Day Fig. 4. Evolution of influent and effluent pH and effluent alkalinity vs. time during operation of the bioreactor.

slightly higher than 85%. Sahinkaya [20] reported an electron flow of 83% in a sulfidogenic continuously stirred tank reactor (CSTR) treating acidic wastewater containing zinc. Finally, Kaksonen et al. [21] reported that in a mesophilic ethanol-fed fluidized bed bioreactor (FBR) the average percentage of electrons utilized for sulfate reduction was 76 ± 10%. 4.2. Alkalinity and pH evolution in the bioreactor The evolution of influent and effluent pHs as well as the effluent alkalinity concentrations are illustrated in Fig. 4. Although the influent pH was 3.5–5 throughout the study, the effluent pH remained always at neutral or slightly alkaline values due to alkalinity production as a result of sulfate reduction according to reaction (R5). During the first 50 days, alkalinity concentration increased slowly from 1000 to 1500 mg CaCO3 /L due to the relatively low sulfate reduction rate observed in this first period. After day 50, the effluent alkalinity concentration increased to around 1900 mgCaCO3 /L and did not change significantly until the last period. Between days 50 and 226, the effluent pH and alkalinity concentration averaged 7.4 ± 0.2 and 1736 ± 239 mg CaCO3 /L respectively. Even addition of high As and Fe concentrations to the feed solution did not significantly affect effluent pH and alkalinity. Although influent COD concentration was lowered to half of its initial value (780 mg/L) in the last period the effluent pH did not change (7.3 ± 0.18), whereas the alkalinity concentration decreased appreciably to 456 ± 228 mg CaCO3 /L as expected. These results indicate that AMD can be neutralized in sulfidogenic bioreactors due to adequate alkalinity generation by sulfate reducing bacteria. 4.3. Metal removal The evolution of arsenic and iron concentrations in the effluent throughout the experiment are shown in Fig. 5. Until period 7, arsenic was the sole component present in the synthetic feed and its concentration increased gradually from 0.5 mg/L to 20 mg/L within 175 days (Table 1 and Fig. 4). In period 7, 100 mg/L of Fe2+ was supplemented to the feed in order to explore its effect on As removal. In the periods 8 and 9, Fe2+ concentration increased to 200 mg/L and the influent COD concentration decreased to half of its initial value only in period 9 to further improve As removal efficiency. Between periods 2 and 6, in which As(V) was supplemented as sole component, the influent and the effluent total As concentrations were quite similar and As removal percentage was no higher than 8–9%. Neutral to slightly alkaline pH and high concentration of

M. Altun et al. / Journal of Hazardous Materials 269 (2014) 31–37

35

25

15

5 0 200 Fe (mg/L)

Q

8000

10

Lin (Counts)

As (mg/L)

10000

Influent Effluent

20

Influent Effluent

150

6000

4000

A

Q

Q

2000 S

S

100 0

50

P A

A

4

C

20

AA

A Q A A P A Q Q A

Q

A

A

A S A Q

40

Q P

Q

Q

60

2-Theta - Scale

0 60

90

120

150 Day

180

210

240

Fig. 5. Evolution of As and Fe concentration in the effluent throughout the study.

dissolved sulfide concentration in the reactor were the reasons for this behavior. As5+ in the influent was reduced to As3+ within the reactor by dissolved sulfide, resulting in the formation of the wellknown mineral and pigment orpiment (As2 S3 ) as it was indicated from the yellow-orange crystal precipitate formed by reducing the pH of an aliquot of solution in a separate flask. Similar results were obtained by Newman et al. [11], but it has been stated that the generated As2 S3 can redissolve according to reaction (R7) when the reactor operates at high pH and sulfide concentrations. In period 7, the addition of 100 mg/L Fe2+ to the synthetic feed resulted in a sharp decrease of the effluent total As concentration to 7.5 ± 3.5 mg/L, corresponding to around 63% As removal. In period 8, the increase of Fe2+ concentration to 200 mg/L decreased the effluent As concentration to 3.72 ± 1.72 mg/L, corresponding to over 81% As removal. No external Fe addition to increase As removal is required during the sulfidogenic AMD treatment, as the real AMD already contains high Fe concentration, and this makes the process more efficient and cost effective [15]. In the last period, the influent COD concentration was decreased to its half in order to further increase As removal efficiency. This option decreased appreciably the effluent sulfide concentration to 46 mg/L (Fig. 2) and subsequently the dissolution of As2 S3 according to reaction (R7). The effluent As concentration in the last period decreased further to 0.92 ± 0.2 mg/L corresponding to 95.4% As removal, and proving that decreased sulfide concentration enhances As removal efficiency even at neutral to alkaline pHs. Therefore, the presence of Fe in the influent and the decrease of the sulfide concentration in the reactor resulted in an appreciable decrease of the total As concentration in the effluent. The concomitant removal of arsenic via adsorption and/or co-precipitation using Fe in a sulfidogenic environment has several advantages including low cost, simultaneous removal of other metals and sulfate as well as increase of pH as a result of alkalinity generation during sulfate reduction. Since Fe and As already exist in numerous mining, metallurgical and industrial effluents, the potential of the proposed process is very high. If adequate amount of Fe is present, high arsenic removal is possible even at high pHs and sulfide concentrations. Eary [4] stated that orpiment (As2 S3 ) precipitates mostly in acidic pH and low sulfide concentrations. If the dissolved sulfide concentration exceeds 1 mM

Fig. 6. XRD pattern of precipitate (Q, quartz SiO2 ; A, aragonite CaCO3 ; C, calcite CaCO3 ; P, pyrite FeS2 ; S, sulfur).

for equal mM of As, then soluble thioarsenic species may dominate the solution at high pHs according to reaction (R7) [21]. Therefore, the As removal mechanism must be carefully controlled to eliminate the negative effect of high pH and sulfide concentration. The increase of As removal in the presence of Fe may be also due to the formation of FeAsS instead of As2 S3 [11,21] and the adsorption of As by FeS formed in the reactor [22]. 4.4. Mineralogical analysis The composition of the precipitate (sludge) produced in the reactor using XRF analysis is given in Table 2. The chemical analysis, apart from Cr, Cl and Cu, is shown in the form of oxides. The XRD pattern of the precipitate is presented in Fig. 6. The presence of quartz, aragonite and calcite is attributed to the supporting material mixed with sludge. Sulfur and FeS2 are also detected. No arsenic compounds are detected due to (i) their presence in low quantities, (ii) their partial amorphous nature, and (iii) limitations of the instrument used (detection limit around 3–5%). Fig. 7 shows cross sectional SEM–backscattered electron (BSE) images as well as the representative EDS analyses of the precipitate. All BSE images (Fig. 7a, c and e) show the presence of substantial amounts of Fe, As and S in bigger or lesser percentages. It is most probable that arsenic precipitates mainly in the form of As2 S3 . This is clearly shown in Fig. 7a and b which indicates that the precipitate contains 30.76% As, 14.82% S and 10.6% Fe. Considering that iron present in the stock solution will precipitate as FeS2 as a result of the sulfidogenic activity, as indicated by XRD analysis, or as FeS and FeAsS, it is assumed that part of arsenic may be also adsorbed on these precipitates. This is more clearly indicated in Fig. 7d and f, in which point analysis shows high content of Fe (27.9–31.4%), and lower contents of As (3.7–16.8%) and S (6.4–13.7%). It has to be mentioned also that part of the precipitated compounds may present a degree of amorphicity. The presence of Ca, O and Si shown in all BSE images is attributed to the presence of quartz and calcite in the support material. Some S present in sludge may also belong to hydrated sulfates. The traces of P shown in Fig. 7f are due to the presence of KH2 PO4 in the synthetic feed. Thus, SEM studies indicate that the precipitation of As in the form of As2 S3 and its co-precipitation with FeS, FeS2 or FeAsS are the main As removal mechanisms in the studied system.

Table 2 Composition of precipitate. Compound

SiO2

CaO

Fe2 O3

SO3

Al2 O3

Na2 O

MgO

K2 O

P2 O5

ZnO

SrO

Cr

Cl

Cu

SUM

%

16.24

34.21

16.71

26.35

0.94

2.63

1.83

0.38

0.41

0.09

0.13

0.01

0.05

0.01

99.99

36

M. Altun et al. / Journal of Hazardous Materials 269 (2014) 31–37

Fig. 7. Cross sectional SEM – backscattered electron (BSE) images (a, c, and e) as well as the representative EDS analyses of the precipitate (b, d, and f). (a) and (b) Indicate the presence of As2 S3 while (c)–(f) indicate co-precipitation of As with FeS, FeS2 or FeAsS.

5. DGGE analysis The presence of sulfate reducing bacteria groups was confirmed via DGGE analysis using the samples obtained from the bioreactor. Five different bands were identified in DGGE gels (Fig. 8). The excision of the generated bands was carried out prior to PCR amplification and then subjected to BLAST algorithm in order to reveal the similarity of the clones to reference strains in the GenBank databases; the results are shown in Table 3. Analysis of the active sludge of the bioreactor, resulted in identification of different microorganism species corresponding to 5

Table 3 Results of the BLAST analysis using DNA sequences in isolated DGGE bands. Band no.

Description

Accession

Query cover (%)

Max ident (%)

1 2 3 4 5

Desulfomicrobium baculatum Desulfovibrio desulfiricans Desulfovibrio africanus Desulfovibrio magneticus Desulfovibrio sp.

NC013173.1 NC011883.1 NC016629.1 NC011567.1 N.1

100 99 100 100 99

97 92 91 97 93

M. Altun et al. / Journal of Hazardous Materials 269 (2014) 31–37

37

treated in a sulfidogenic bioreactor without the addition of any external source of alkalinity. Acknowledgements This study was funded by Istanbul Medeniyet University Research Fund (Project No: FBA-2012-174) and the Scientific & Technological Research Council of Turkey (TUBITAK Project No: 109Y374). References

Fig. 8. The inverted image of the gene sequences isolated from bands using DGGE.

different bands within the gel medium. Results of the BLAST algorithm also identified a few different bacteria species which do not use sulfate as a terminal electron acceptor (not shown in Table 3). The most predominant species identified are Desulfomicrobium baculatum, Desulfovibrio desulfiricans, Desulfovibrio africanus and Desulfovibrio sp. Although some species are aerotolerant, the majority of the community is composed of anaerobic bacteria groups and is mostly active in mesophilic environment as sulfate reducers using sulfate and thiosulfates as electron acceptors. One of the interesting species detected during DGGE is Desulfurovibrio magneticus which is a magnetotactic bacterium and generates extracellular magnetic iron sulfide [23]. 6. Conclusion In this study, As(V) removal from AMD was studied in a sulfidogenic up flow fixed bed bioreactor under varying operating conditions. The steady increase of the influent As(V) concentration within 175 days to 20 mg/L did not adversely affect reactor performance while the sulfate removal rate averaged 4 g/(L d). Although the influent pH was 3.5–5.0, the effluent pH averaged 7.5 throughout the study due to the alkalinity produced as a result of sulfate reduction. In the absence of iron no significant arsenic removal was observed at high sulfide concentrations. When the synthetic feed was supplemented with 100 mg/L and 200 mg/L Fe2+ , arsenic removal increased to 63% and over 80%, respectively, even at high sulfide concentration, due to the formation of FeS or FeAsS instead of As2 S3 . Decrease of the dissolved sulfide production by limiting the carbon source, but maintaining Fe2+ concentration in the feed at 200 mg/L, further increased arsenic removal to around 96%. The use of analytical techniques indicates that precipitation of arsenic in the form of As2 S3 , and co-precipitation with FeS, FeS2 or FeAsS are the main As removal mechanisms in the studied system. Hence, AMD containing high concentration of As and Fe may be effectively

[1] J.M. Azcue, J.O. Nriagu, Impact of abandoned mine tailings on the arsenic concentrations in Moira Lake, Ontario, J. Geochem. Explor. 52 (1995) 81–89. [2] M. Williams, F. Fordyce, A. Paijitprapapon, P. Charoenchaisri, Arsenic contamination in surface drainage and groundwater in part of the southeast Asian tin belt, Nakhon si Thammarat Province, southern Thailand, Environ. Geol. 27 (1996) 16–33. [3] P.L. Smedley, W.M. Edmunds, K.B. Pelig-Ba, Mobility of arsenic in groundwater in the Obuasi gold-mining area of Ghana: some implications for human health, in: J.D. Appleton, R. Fuge, G.J.H. McCall (Eds.), Environmental Geochemistry Health, Geol. Soc. Spec. Publ. 113, Geological Society, London, 1996, pp. 163–181. [4] L.E. Eary, The solubility of amorphous As2 S3 from 25 to 90 ◦ C, Geochim. Cosmochim. Acta 56 (1992) 2267–2280. [5] U. Gemici, G. Tarcan, C. Helvaci, A.M. Somay, High arsenic and boron concentrations in groundwaters related to mining activity in the Bigadic¸ borate deposits (Western Turkey), Appl. Geochem. 23 (2008) 2462–2476. [6] K.A. Natarajan, Microbial aspects of acid mine drainage and its bioremediation, Trans. Nonferrous Met. Soc. China (English Ed.) 18 (2008) 1352–1360. [7] C. Komnitsas, F.D. Pooley, Optimization of the bacterial oxidation of an arsenical gold sulphide concentrate from Olympias, Greece, Miner. Eng. 4 (1991) 1297–1303. [8] M. Williams, Arsenic in mine waters: international study, Environ. Geol. 40 (2001) 267–278. [9] P.S. Herrera, H. Uchiyama, T. Igarashi, K. Asakura, Y. Ochi, F. Ishizuka, S. Kawada, Acid mine drainage treatment through a two-step neutralization ferrite-formation process in northern Japan: Physical and chemical characterization of the sludge, Miner. Eng. 20 (2007) 1309–1314. [10] F. Battaglia-Brunet, C. Crouzet, A. Burnol, S. Coulon, D. Morin, C. Joulian, Precipitation of arsenic sulphide from acidic water in a fixed-film bioreactor, Water Res. 46 (2012) 3923–3933. [11] D.K. Newman, T.J. Beveridge, F.M.M. Morel, Precipitation of arsenic trisulfide by Desulfotomaculum auripigmentum, Appl. Environ. Microbiol. 63 (1997) 2022–2028. [12] A.A. Carbonell-Barrachina, J.D. Jorda, F.M. Burlo, F. Martinez-Sanchez, Evaluation of arsenite sorption in Spanish soils, Commun. Soil Sci. Plant Anal. 31 (2000) 2865–2879. [13] E. Sahinkaya, M. Altun, S. Bektas, K. Komnitsas, Bioreduction of Cr(VI) from acidic wastewaters in a sulfidogenic ABR, Miner. Eng. 32 (2012) 38–44. [14] APHA, WEF, Standard Methods for the Examination of Water and Wastewater, 21st ed., American Public Health Association, American Water Works Association, Water Environmental Federation, Washington, DC, USA, 2005. [15] R. Cord-Ruwisch, A quick method for the determination of dissolved and precipitated sulfides in cultures of sulfate-reducing bacteria, J. Microbiol. Methods 4 (1985) 33–36. [16] Denaturing gradient gel electrophoresis of PCR-amplified 16S rDNA – a new molecular approach to analyse the genetic diversity of mixed microbial communities, in: G. Muyzer, S. Hottentrager, A. Teske, C. Wawer, A.D.L. Akkermans, J.D. Van Elsas, F. De Bruijn (Eds.), Molecular Microbial. Ecology Manual, Kluwer Academic Publishers, Dordrecht, 1996. [17] E. Sahinkaya, F.M. Gunes, D. Ucar, A.H. Kaksonen, Sulfidogenic fluidized bed treatment of real acid mine drainage water, Bioresource Technol. 102 (2011) 683–689. [18] A.H. Kaksonen, P.D. Franzmann, J.A. Puhakka, Effects of hydraulic retention time and sulfide toxicity on ethanol and acetate oxidation in sulfate-reducing metalprecipitating fluidized-bed reactor, Biotechnol. Bioeng. 86 (2004) 332–343. [19] E. Sahinkaya, Z. Yucesoy, Biotreatment of acidic zinc- and copper-containing wastewater using ethanol-fed sulfidogenic anaerobic baffled reactor, Bioprocess Biosyst. Eng. 33 (2010) 989–997. [20] E. Sahinkaya, Biotreatment of zinc-containing wastewater in a sulfidogenic CSTR: performance and artificial neural network (ANN) modelling studies, J. Hazard. Mater. 164 (2009) 105–113. [21] A. Bayrakdar, E. Sahinkaya, M. Gungor, S. Uyanik, A.D. Atasoy, Performance of sulfidogenic anaerobic baffled reactor (ABR) treating acidic and zinc-containing wastewater, Bioresource Technol. 100 (2009) 4354–4360. [22] D. Teclu, G. Tivchev, M. Laing, M. Wallis, Bioremoval of arsenic species from contaminated waters by sulphate-reducing bacteria, Water Res. 42 (2008) 4885–4893. [23] T. Sakaguchi, A. Arakaki, T. Matsunaga, Desulfovibrio magneticus sp. nov., a novel sulfate-reducing bacterium that produces intracellular single-domain-sized magnetite particles, Int. J. Syst. Evol. Microbiol. 52 (2002) 215–221.