Bioelectrochemical treatment of acid mine drainage dominated with iron

Journal of Hazardous Materials 241–242 (2012) 411–417

Contents lists available at SciVerse ScienceDirect

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

Bioelectrochemical treatment of acid mine drainage dominated with iron Olivier Lefebvre a,b , Carmen M. Neculita b , Xiaodi Yue a , How Yong Ng a,∗ Centre for Water Research, Department of Civil and Environmental Engineering, National University of Singapore, 1 Engineering Dr. 2, Singapore 117576, Singapore Research Institute - Mines and Environment, University of Quebec in Abitibi-Témiscamingue (UQAT), Rouyn-Noranda, Canada J9X 5E4

h i g h l i g h t s  Treatment of AMD dominated with Fe is problematic due to its low pH.  The goal of AMD treatment is pH increasing and Fe controlled removal.  An MFC could fulfill both these requirements by treating AMD at the aerated cathode.  The performance was proportional to the charge transfer up to 880 C.  The treated AMD met the discharge limits both for Fe (>99% recovery) and pH (7.9).

a r t i c l e

i n f o

Article history: Received 18 July 2012 Received in revised form 25 September 2012 Accepted 26 September 2012 Available online 5 October 2012 Keywords: Acid mine drainage Bioelectrochemical treatment Iron oxy(hydroxi)des Microbial fuel cell pH

g r a p h i c a l

a b s t r a c t

y = 8.6862ln(x) + 42.126 R² = 0.6709

12

100

10

75

8 y = 2.1253ln(x) - 5.4575 R² = 0.9453

6 4

pH Fe recovery

2 0 0

500 1,000 Charge (C)

50 25

Fe recovery (%)

b

pH

a

0 1,500

a b s t r a c t Treatment of acid mine drainage (AMD) dominated with iron (Fe), the most common metal, is a longterm expensive commitment, the goal of which is to increase the pH and remove Fe. In the present study, a proton exchange membrane microbial fuel cell (MFC) showed promise for the efficient treatment of an AMD dominated with ferric iron (pH 2.4 ± 0.1; 500 mg L−1 Fe3+ ). Briefly, Fe3+ was reduced to Fe2+ at the cathode of the MFC, followed by Fe2+ re-oxidation and precipitation as oxy(hydroxi)des. Oxygen reduction and cation transfer to the cathode of the MFC further caused a rise in pH. A linear relationship was observed between the charge transferred in the MFC and the performance of the system up to 880 C. Optimal conditions were found at a charge of 662 C, achieved within 7 d at an acetate concentration of 1.6 g L−1 in a membrane MFC. This caused the pH to rise to 7.9 and resulted in a Fe removal of 99%. Treated effluent met the pH discharge limits of 6.5–9. The maximum power generation achieved under these conditions averaged 8.6 ± 2.3 W m−3 , which could help reduce the costs of full-scale bioelectrochemical treatment of AMD dominated with Fe. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Acid mine drainage (AMD), which is produced through a combination of biochemical processes during weathering of metal sulfides in mine wastes, represents a challenging environmental problem worldwide [1]. AMD is characterized by low pH (as low as

−3.6) and high concentrations of dissolved metals (up to 200 g L−1 ) and sulfate (up to 760 g L−1 ) [2–5]. The most common metal sulfides in mine wastes are pyrite (FeS2 ) and pyrrhotite (Fe(1−x) S; x = 0–0.2). Thus, Fe concentrations in AMD are usually among the highest, often in excess of 1 g L−1 and can be as high as 111 g L−1 [2–5]. In the presence of oxygen, overall pyrite oxidation reaction is as follows: FeS2 (s) + 15/4O2 + 7/2H2 O → Fe(OH)3 (s) + 2·SO4 2− + 4H+

Abbreviations: Amd, acid mine drainage; Mfc, microbial fuel cell; Orr, oxygen reduction reaction. ∗ Corresponding author. Tel.: +65 6516 4777; fax: +65 6779 1635. E-mail address: [email protected] (H.Y. Ng). 0304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2012.09.062

(1)

Over time, the pH of the AMD progressively decreases, while dissolved metal concentrations (including Fe) in the mine waste increase [1]. The treatment of this highly contaminated water is very challenging as very few options are available for Fe removal

412

O. Lefebvre et al. / Journal of Hazardous Materials 241–242 (2012) 411–417

at very low pH, mainly because the Fe(OH)3 precipitation is the rate-limiting step of the process. Conventionally, the treatment of Fe-rich AMD uses active and passive technologies for precipitation of Fe-oxy(hydroxi)des and carbonates by raising the pH [1]. Active chemical treatment is performed through continuous addition of neutralizing materials (e.g., limestone), followed by agitation and forced aeration [2]. One challenge of this approach is the high energy consumption required to precipitate Fe to concentrations below the discharge limits and to maintain the pH of the treated effluent in the admissible range (6–9.5) [6]. An alternative approach is passive treatment, which has advantages over active treatment, including high metal removal at low pH, stable sludge production, and lower operational cost and energy consumption [7]. Treatment efficiency using biological processes, however, is limited for highly Fe-contaminated AMD because of the inhibitory effects of AMD’s low pH and Fe toxicity on the activity of anaerobic bacteria. In addition, the reactive mixture, used to fill the bioreactors, controls the long-term efficiency through substrate availability to anaerobic bacteria [7]. A microbial fuel cell (MFC) is a bioelectrochemical system, whereby bacteria catalyze the conversion of the chemical energy from organic substrates into electrical energy [8]. As the energetic performance of MFCs is still unsatisfactory to justify for scale-up, despite the intense research performed within the last decade, other applications could be explored [9]. The MFC could be an effective solution for the treatment of Fe-dominated AMD by the consumption of excess H+ to promote the oxygen reduction reaction (ORR) at the cathode of the system: O2 + 4e− + 4H+ → 2H2 O

(2)

In an MFC, the protons required for the ORR are normally provided by the anode reaction, which can be described as follows when acetate is used as a substrate and a source of electrons: 3CH3 COO− + 6H2 O → 6CO2 + 24e− + 21H+

(3)

Proton availability and mobility generally constitute the main limitations in an MFC, where electroneutrality is achieved by transport of other more abundant cation species, such as Na+ , K+ , NH4 + , Ca2+ and Mg2+ [10]. However, it was recently found that in an acidic environment, the cathode can use the protons directly available in its vicinity, resulting in an overall transfer of acidity whereby the pH rises at the cathode and decreases at the anode [11]. By overcoming the shortage of protons at the cathode, a significant increase in power generation can be recorded. Consequently, if highly acidic Fe-dominated AMD was to be incorporated in the cathodic compartment of an MFC, it could provide protons for the ORR, while being treated in the same time. Furthermore, the Fe3+ present in the catholyte could provide an efficient cathodic electron mediator (catalyst) as documented in the literature [12–15]. In the present study, the bioelectrochemical treatment of AMD dominated by Fe was attempted using acetate solution (as substrate component of the reactive mixture) in the anode compartment and simulated AMD (FeCl3 solution) in the cathode compartment. The originality of the study lies in the proposed application more than in the MFC architecture. Yet, the study was performed with two different MFC designs: a salt bridge MFC and a membrane MFC. Overall, the salt bridge design – despite its rudimentary architecture – was intended to demonstrate the principle and feasibility of AMD bioelectrochemical treatment, while the membrane design was expected to improve Fe recovery, therefore demonstrating the potential of this technology. For this reason, both MFC designs are presented in this paper.

2. Experimental 2.1. MFC construction Two types of dual-chamber MFCs, differing in the nature of the separator, were designed and set up for this study (Fig. 1). The first type of MFC (Fig. 1a) was made of two 250 mL Duran bottles connected by a salt bridge (diameter of 15 mm) prepared according to Min et al. [16]. The composition of the salt bridge was as follows: 0.275 g of Na2 HPO4 , 0.422 g of NaH2 PO4 and 0.160 g of agar in 100 mL of deionized (DI) water. The second type of MFC (Fig. 1b) was made of two polyacrylic chambers (630 mL each) separated by a proton exchange membrane (Nafion 117; DuPont Co., USA). Nafion 117 was pretreated by boiling in H2 O2 (30% v/v) and DI water, followed by soaking in 0.5 M H2 SO4 and then DI water, each for 1 h. The apparent surface area of the membrane equaled 90 cm2 . The anode and the cathode chambers of both MFCs were filled with graphite granules (diameter between 1.5 and 5 mm; Carbone Lorraine, Belgium) as electrode material and graphite rods were used as current collector. The working volume of each chamber was 100 mL for the salt bridge MFC and 175 mL for the membrane MFC. A control set-up consisted of a 250 mL Duran bottle filled with graphite granules and artificial AMD (see Section 2.2), and aerated for the same duration during the experimental runs. 2.2. Operating conditions Obtaining inocula from other operating MFCs has become a common practice to reduce the startup time of an MFC. This ensures that electrochemically active bacteria are directly seeded into the new MFC [17]. Accordingly, both types of MFCs were seeded using the effluent of an active MFC available in our laboratories and fed with domestic wastewater [18]. Specifically, half of the anode chamber was filled with the effluent and the other half with the reactive mixture. The procedure was repeated until a significant increase of voltage was recorded, indicating successful inoculation. Subsequently, the reactive mixture alone was added into the anode compartment. The reactive mixture – used as the anolyte – consisted of a solution of nutrients, minerals and vitamins as per Oh et al. [19], to which sodium acetate was added as the carbon source and electron donor (substrate). The nutrient solution contained NaHCO3 (3.13 g L−1 ), NH4 Cl (0.31 g L−1 ), NaH2 PO4 ·H2 O (0.75 g L−1 ), KCl (0.13 g L−1 ), NaH2 PO4 (4.22 g L−1 ), Na2 HPO4 (2.75 g L−1 ), and metal (12.5 mL L−1 ) and vitamin (12.5 mL L−1 ) solutions. The metal and vitamin solutions were prepared following Balch et al. [20]. The acetate concentration used in the anolyte ranged from 100 to 12,800 mg L−1 . An artificial AMD, with 500 mg L−1 of Fe3+ , prepared using ferric chloride hexahydrate (FeCl3 ·6H2 O), was used as the catholyte. The selected Fe3+ concentrations correspond to values found in natural AMD (e.g., 551 mg L−1 Fe on Youngdong mine site, South Korea) [21]. The pH of the artificial AMD was left unadjusted at 2.4 ± 0.1. The experiment was carried out in batch mode. At the start of a batch test, the anode and the cathode chambers of the MFCs were filled with fresh solutions of anolyte and catholyte and the batch test was considered completed when the voltage recorded over an external resistance of 5  dropped below 0.2 mV. The anode chamber was kept anaerobic throughout the batch testing, while the cathode chamber was constantly aerated using an aquarium air pump connected to an air diffuser. The experiment was conducted at an ambient temperature (25 ◦ C). 2.3. Analytical methods and calculations The voltage of the MFCs was monitored continuously by a digital multimeter (M3500A; Array Electronic, Taiwan) connected to a

O. Lefebvre et al. / Journal of Hazardous Materials 241–242 (2012) 411–417

413

Fig. 1. Schematic diagram of the (A) salt bridge and (B) membrane- microbial fuel cells used in this study.

performance of the system as a function of the acetate concentration is presented in Table 1 and the impact of the charge transfer on the pH of treated AMD and corresponding Fe recovery is further shown in Fig. 3. A linear relationship was observed for both parameters, demonstrating the possibility of elevating the pH of the AMD and recovering Fe in the MFC system proposed in the present study. A polarization and a power curve of the salt-bridge set-up are further shown in Fig. 4. As compared to the literature available on MFCs – a list of which is available elsewhere [11] – the Eemf of the salt bridge MFC was in the higher range at 0.9 ± 0.1 V but its Rint was also very high at 499 ± 75 m m3 , resulting in overall low power density of 0.5 W m−3 . Due to the high internal resistance, the Coulombic efficiency was low and further decreased with increasing charge transfer from 17% in the initial conditions to 1% at the highest charge tested (Table 1), as non-electrogenic reactions became predominant in the MFC, a phenomenon already reported 2 Salt bridge MFC 1600

3200

6400

12800

200

(B)

0 35 0

500 Time (h)

30

1600

800

25

1000 1500 Membrane MFC

3200

12800 6400

400

20 inoculation

Voltage (mV)

The salt bridge MFC was initially operated with 100 mg L−1 of acetate at the anode and 500 mg L−1 of Fe3+ at the cathode, at the beginning of each batch test. The voltage recorded was initially negligible (<0.1 mV, across an external resistance of 5 ); however, a steep increase of electricity production was observed after 170 h, which indicated successful colonization of the anode by electrochemically active bacteria (Fig. 2A). Thereafter, a good repeatability between batches was observed, characterized by an initially high voltage of around 1 mV (across a 5  resistance) for an initial period of around 5 h after the replenishment of the anodic and cathodic chambers with fresh solutions. The voltage then decreased slowly and regularly until the end of the batch test (Fig. 2A). Following successful inoculation, the salt bridge MFC was operated under increasing acetate concentrations of up to 12.8 g L−1 , resulting in increased charge transfer from the anode to the cathode (cf. typical profiles of electricity generation in Fig. 2A). The

800

1

3. Results 3.1. Salt bridge MFC

400

100

inoculation

(A) Voltage (mV)

desktop computer through a data acquisition system (PC1604; TTi, RS, Singapore). The current was then determined using the Ohm’s law and the Coulombic efficiency was calculated following Logan et al. [8]. Polarization curves were obtained at the start of each batch test by gradually decreasing the external resistance applied to the system and recording the pseudo steady-state voltage. The cell electromotive force (Eemf , V), internal resistance (Rint , m( m3 ) and maximum power (Pmax , W m−3 ) were then determined from these polarization curves following Lefebvre et al. [11]. In this paper, Rint and Pmax were normalized to the net cathodic compartment volume. The pH, acetate and total Fe concentrations were analyzed at the beginning and the end of each batch test following the Standard Methods [22]. Acetate concentration (detection limit of 10 ␮g L−1 ) was determined by gas chromatography–mass spectrometry after filtration using a 0.45 ␮m filter paper. Total Fe concentration was determined by inductively coupled plasma-optical emission spectroscopy (Optima 7200DV, PerkinElmer, USA). Mineralogical analysis of membrane was performed by energy-dispersive X-ray spectroscopy coupled with scanning electron microscopy (40-FE, Supra® , Carl Zeiss NTS GmbH, Germany). Selected membranes were sputtered with gold (SCD 005, Bal-Tec® , Micro Surface Engineering Inc., USA) prior to analysis.

15 10 5

200

100

0

0

250

500

750 1000 Time (h)

1250

1500

Fig. 2. Typical profiles of voltage generation across a 5  external resistance for (A) salt bridge microbial fuel cell and (B) membrane microbial fuel cell treating artificial AMD dominated with Fe. The numbers refer to the concentration of acetate provided in the reactive mixture (anolyte) in mg L−1 .

414

O. Lefebvre et al. / Journal of Hazardous Materials 241–242 (2012) 411–417

Table 1 Operating conditions and performance summary of the MFCs used in this study. System

Anolyte (reactive mixture)

Batch duration (d)

Charge transfer (C)

Coulombic efficiency (%)

pH

Acetate conc. (mg L−1 )

Catholyte (synthetic AMD) Fe conc. (mg L−1 )

pH

Fe recovery (%)

Initial

Final

Initial

Final

Initial

Final

Initial

Final

Salt bridge MFC

100 200 400 800 1600 3200 6400 12800

bdl bdl bdl bdl bdl bdl bdl bdl

7.7 7.5 8.0 7.5 7.8 7.7 7.9 7.8

7.8 7.5 7.6 7.6 8.0 8.7 8.8 9.1

1 3 4 6 7 8 9 11

21 29 45 70 94 122 143 201

17 12 9 7 5 3 2 1

2.4 2.4 2.4 2.4 2.4 2.5 2.5 2.4

2.4 2.5 2.4 2.6 2.6 3.0 3.1 4.2

500 500 500 500 500 500 500 500

450 431 394 398 381 332 323 282

10 14 21 20 24 34 36 44

Membrane MFC

100 200 400 800 1600 3200 6400 12800

bdl bdl bdl bdl bdl bdl bdl bdl

7.6 7.7 7.5 7.8 7.5 7.8 7.8 7.9

6.9 6.6 6.5 6.8 6.5 5.9 6.8 6.5

2 4 5 6 7 9 11 12

43 80 287 527 662 883 942 1204

38 36 32 30 19 10 7 4

2.4 2.4 2.4 2.4 2.5 2.5 2.4 2.4

2.8 4.1 5.8 7.1 7.9 9.8 9.9 9.4

500 500 500 500 500 500 500 500

191 37 12 13 5 1 bdl 2

62 93 98 97 99 >99 >99 >99

Control

NA NA NA NA NA NA NA NA

NA NA NA NA NA NA NA NA

NA NA NA NA NA NA NA NA

NA NA NA NA NA NA NA NA

1 3 4 6 7 8 9 11

NA NA NA NA NA NA NA NA

NA NA NA NA NA NA NA NA

2.4 2.4 2.4 2.4 2.4 2.5 2.5 2.4

2.3 2.2 2.4 2.5 2.4 2.4 2.5 2.6

500 500 500 500 500 500 500 500

470 468 465 470 447 388 372 301

6 6 7 6 11 22 26 40

Note. bdl: below detection limit; NA: not applicable.

3.2. Membrane MFC The membrane MFC was started similarly to the salt bridge MFC, at the same initial concentrations of acetate (100 mg L−1 ) and Fe3+ (500 mg L−1 Fe3+ ) at the anode and at the cathode, respectively. Likewise, the voltage output was initially low (<1 mV, across an external resistance of 5 ), but increased significantly after 80 h of

(A)

3

75 y = 0.0087x + 2.0458 R² = 0.8623

50

2 1

25

y = 0.1801x + 8.9314 R² = 0.9581

0

0

0

100

200

300

Charge (C) Fig. 3. Impact of the charge transfer on the pH of synthetic AMD dominated with Fe used as catholyte and corresponding Fe recovery in the salt bridge MFC.

0.6 0.4

0

Power (W m-3)

pH

4

Fe recovery (%)

100 pH Fe recovery

Salt bridge MFC Membrane MFC

0.2

(B) 5

1

0.8

Voltage (V)

by Lefebvre et al. [11]. Unexpectedly, the pH of the anolyte had a tendency to increase during the batch test, when the opposite trend was expected. This is, again, related to the limited performance of the system, where the charge produced was insufficient to lower the pH of the medium of Oh et al. [19]. This medium is indeed characterized by high concentration of sodium bicarbonate, which increased the pH of the anolyte in a way that could not be compensated by the generation of protons in this electrically highly resistant system. The limited electrical performance of a salt bridge MFC is a confirmation of previous findings [16]; however, the intent of this set-up was elsewhere as it served the purpose to successfully demonstrate the possibility to treat AMD in a MFC, with a linear relationship between the charge transfer and the pH change (and Fe recovery). At the highest charge tested (201 C), the pH increased to 4.2, inducing a Fe recovery of 44%. Higher concentrations of acetate were not tested with this set-up due to extended batch duration under these conditions (11 d, see Table 1). Instead, the effect of higher charge transfer was assessed in a more efficient MFC design, making use of a membrane separator.

15 12

9 6 3

0 0

2

4 Current (mA)

6

8

Fig. 4. Typical (A) polarization and (B) power curves obtained with MFCs treating artificial AMD dominated with Fe.

O. Lefebvre et al. / Journal of Hazardous Materials 241–242 (2012) 411–417

10

75

pH

8 y = 2.1253ln(x) - 5.4575 R² = 0.9453 pH Fe recovery

6 4 2 0 0

500

1,000

on its own further demonstrates the active role of the MFC in the treatment of AMD.

100

50 25

Fe recovery (%)

y = 8.6862ln(x) + 42.126 R² = 0.6709

12

415

0 1,500

Charge (C) Fig. 5. Impact of the charge transfer on the pH of synthetic AMD dominated with Fe used as catholyte and corresponding Fe recovery in the membrane MFC.

operation (Fig. 2B). Hence, the time required for inoculation and acclimation was significantly reduced (by about 50%) in the membrane MFC as compared to the salt bridge MFC (80 h versus 170 h). Subsequently, the profile of every batch test was highly similar to that observed with the salt bridge MFC, but the voltage peaked at around 2 mV (across a 5  resistance), which was twice as high as that obtained with the salt bridge MFC (Fig. 2B). The voltage further increased considerably with increasing acetate concentration up to a plateau of about 25 mV (Fig. 2B). The impact of the charge transfer on the pH of the treated AMD under increasing acetate concentrations of up to 12.8 g L−1 and the corresponding Fe removal efficiency are shown in Fig. 5. The pH and Fe removal initially increased with the charge; however, a plateau was ultimately reached and charges higher than 883 C did not further increase the pH or the Fe recovery. The optimal conditions were found at a charge of 662 C, which value was achieved within 7 d at an acetate concentration of 1.6 g L−1 (Table 1). This caused the pH to rise from 2.5 to 7.9, resulting in a Fe recovery of 99%. The main issue of the membrane MFC was its long-term instability. With a fresh membrane, Eemf averaged 0.9 ± 0.1 V, while Rint averaged 26 ± 3 m m3 (Fig. 4). Thus the Eemf was found to be in the same range as what could be obtained from the salt bridge MFC, but the Rint induced by the proton exchange membrane was one order of magnitude lower. As a result, the power generation averaged 8.6 ± 2.3 W m−3 . However, the performance decreased with time and the membrane became ineffective after a period of 2 weeks. When replaced with a fresh membrane, the system recovered instantly. This could be explained by iron precipitation on the membrane that damaged it irreversibly with time, as previously reported by Ter Heijne et al. [14]. The Coulombic efficiency in this well performing system was higher than that of the salt bridge MFC (38% in the initial conditions) but decreased with increasing charge (Table 1). The pH of the anolyte decreased throughout the batch testing, but again only to a limited extent due to the buffering provided by the medium of Oh et al. [19].

3.3. Control A control, consisting of an aerated solution of synthetic AMD at the same concentration as in the fuel cells, was run in parallel to the MFCs used in this study (Table 1). In the control, the pH of the AMD toggled around 2.4–2.6, independently of the time. Under these conditions, substantial Fe removal occurred (up to 40% after 11 d). Moderate removal of Fe3+ at pH around 2.5 is not a surprise and can be explained by the strong dependency of Fe(OH)3 solubility on pH and total Fe concentration. Indeed, the literature confirms that partial hydrolysis and precipitation of Fe can happen at low concentration and pH (e.g., 1 g L−1 of Fe and pH around 1.6) [23]. Yet, the inability of the aeration alone to increase the pH of the AMD

3.4. Mineralogy of Fe precipitate At the end of the experimental runs, the precipitate on the graphite granules and on the membrane of the membrane-based MFC showed a reddish-brownish color typical of Fe-oxy(hydroxi)de minerals [2]. The membrane was further analyzed by SEM-EDX and the results are reported in Table 2. A fresh membrane was found to be mainly composed of fluorine (F), which constituted 76.7% of the membrane. Other elements included C (15.0%) and O (5.8%). Analysis of the membrane obtained after the end of the experimental runs, however, showed high Fe (63.8%) and O (34.3%) content, which can be interpreted as a confirmation of the presence of Feoxy(hydroxi)des, possibly as a mixture of iron (III) oxy(hydroxi)des (Fe(OH)3 , Fe2 O3 , FeOOH). 4. Discussion 4.1. MFC performance according to the design As mentioned above (Section 3.1), the limited electrical performance of the salt bridge MFC was expected and, consequently, the membrane MFC allowed faster start-up than the salt bridge MFC and displayed enhanced electrical and treatment performances. Yet, the salt bridge design was useful to demonstrate the principle of AMD treatment by MFC, with a linear relationship between the charge and the pH (and Fe recovery) of the treated AMD (Fig. 3). On the other hand, the membrane MFC showed the potential of the technology for AMD treatment at higher charge transfer. Indeed, the relationship between treatment performance and the charge transfer up to around 880 C could explain the enhanced performance in the membrane MFC. Below this value, the efficiency was limited by the availability of the electron donor (acetate) in the anode compartment; however, the trend was found to be reversed above 880 C in the membrane system, where the electron acceptor (Fe) became limiting and was almost fully precipitated (>99% recovery). However, Fe recovery in this design could partially be explained by precipitation on the membrane, resulting in performance deterioration with time. This is undeniably a drawback of the membrane MFC design at this time. Furthermore, both set-ups suffered from low Coulombic efficiencies (maximum of 17% with the salt bridge MFC and 38% with the membrane MFC), which suggests that alternative pathways independent from electrogenesis (e.g., methanogenesis) were mostly responsible for acetate degradation in the anode chamber, as reported by Lefebvre et al. [11]. This is another limitation of this study and of many MFCs in general. 4.2. Chemistry of Fe recovery in the MFC system Based on visual evaluation of its yellowish-brownish color and SEM-EDX analysis, the precipitate collected at the end of batch testing consisted of a mixture of iron (III) oxy-hydroxides (e.g., Fe(OH)3 , Fe2 O3 and FeOOH). These findings are consistent with previously reported results [23,24] that identified the Fe deposit recovered from an abiotic fuel cell oxidizing Fe2+ to Fe3+ to consist mainly of goethite (␣-FeOOH). In spite of the limited Coulombic recovery, the electromotive force averaged 0.9 V for both MFCs, one of the highest compared to the published literature [11]. As previously stated by Ter Heijne et al. [23], this very high value suggests the reduction of Fe3+ to Fe2+ as the dominant reaction, rather than the ORR at the cathode of the MFC. Indeed, in their study the performance of Fe3+ to Fe2+ reduction was found superior to oxygen reduction, even on a

416

O. Lefebvre et al. / Journal of Hazardous Materials 241–242 (2012) 411–417

Table 2 Mineralogical analysis of the Nafion 117 membrane used in this study by SEM-EDX. All values are given in weight percentage.

Blank Experimental

Fe

O

F

C

S

P

Ca

Na

Total

bdl 63.8

5.8 34.3

76.7 bdl

15.0 bdl

2.5 0.2

bdl 1.3

bdl 0.2

bdl 0.2

100 100

Note. bdl: below detection limit.

Pt-coated electrode at pH 4 using pure oxygen. A better affinity of graphite for Fe3+ relative to O2 could explain these results [13,14]. In the study of Ter Heijne et al. [23], Fe3+ reduction at the cathode was furthermore combined with biological Fe2+ re-oxidation to regenerate the Fe3+ . Most likely, in our system, re-oxidation phenomena also occurred, due to the presence of O2 in the catholyte. In fact, after cathodic reduction of Fe3+ to Fe2+ , several simultaneous re-oxidation reactions could occur in the catholyte in the presence of O2 [25], following Eqs. (4)–(6): Fe2+ → Fe3+ + e− +

Fe(OH) → Fe(OH) 0

(4) 2+

+e

+

−

Fe(OH)2 → Fe(OH)2 + e

(5)

−

(6)

Eqs. (4)–(6) are abiotic and spontaneous in the presence of O2 , especially at pH values above 5, depending on the total Fe concentration [25,26]. Fe2+ re-oxidation was then followed by precipitation of Fe(III) in the MFC, according to Eq. (7): Fe

3+

+ 3H2 O → Fe(OH)3 (s) + 3H

+

(7)

Finally, the electrons produced by Eqs. (4)–(6) and the protons generated by Eq. (7) served to promote ORR in the MFC according to Eq. (2) resulting in an overall pH rise. When the pH rises above ∼3.5, the Fe2+ oxidation proceeds more rapidly, improving the Fe3+ hydrolysis and precipitation [2]. Therefore, the MFC increased the efficiency of AMD treatment (i.e., ORR caused a rise in pH that enhanced Fe3+ removal). Eq. (2) is most likely not the only one occurring in the cathode compartment that is responsible for the pH increase of the catholyte, as the transport of cation species other than protons (e.g., Na+ , K+ , NH4 + , Ca2+ and Mg2+ ) also probably contributed to the pH rise, as explained by Rozendal et al. [10]. 4.3. Comparison with other fuel cell applications for metal recovery Applications of fuel cells for metal removal are limited so far. Cheng et al. [27] successfully treated artificial AMD by oxidizing Fe2+ to Fe3+ at the anode of an abiotic fuel cell. However, the reaction generated protons that resulted in Fe(III) dissolution if the pH dropped below 2.5 and consequently Fe removal in this study relied on active pH control using HCl. Furthermore, cathodic reduction of Cr and Cu in an actual MFC has also been studied. However, metals were recovered from solution in their elemental forms. Thus, reactions did not involve protons and, consequently, were inefficient at altering the pH of the catholyte, and because of that, metal recovery was metal-dependent. For example, Cu precipitation independently of the pH allowed Ter Heijne et al. [28] and Wang et al. [29] to successfully recover pure Cu crystals for synthetic streams enriched with Cu. However, when Wang et al. [30] carried out the reduction of Cr(VI) in a two-chamber MFC, complete reduction of Cr(VI) to Cr(III) was achieved but Cr(III) remained soluble in water and could not be recovered. Similarly, in the studies of Ter Heijne et al. [13,14], the iron (either Fe2+ and/or Fe3+ ) remained soluble in solution due to the low pH maintained throughout the tests. Noteworthy, the main goal of these last studies was to achieve the maximum energy efficiency of the MFC and not to recover Fe. Yet, while energy production in an MFC should be maximized as much

as possible, it is a lower priority than human and environmental health. In contrast, the MFC disclosed in the present paper showed a promising approach to address the main drawback of available systems for AMD treatment dominated by Fe, which is the decrease of pH to values that cannot sustain the precipitation of Fe(III) oxy(hydroxi)de minerals (pH below 2–3.5, depending on the total Fe concentration) [2]. Instead of making use of the natural acidity of AMD to improve electricity generation, our MFC successfully increased the pH of the artificial AMD, leading to the precipitation of Fe(III) below the detection limit. Unlike currently available active and passive technologies, this was achieved without use of chemicals, making MFC a promising approach for the treatment of acid mine drainage (AMD) dominated by Fe. 4.4. Practical applications Based on the findings from this study, AMD shows potential to generate substantial amount of power (up to 8.6 ± 2.3 W m−3 ) in an MFC. In the laboratory testing, the power generated was too low to be of use; however, the recorded voltage is a useful indicator of the treatment performance as it reflects the operational conditions (Fig. 2). This can be used typically to control the availability of the reactive mixture in the anode compartment, indicating when it needs to be replenished. In the field, power generation could alleviate the cost of aeration or the cost associated to bringing and disposing of the reactive mixture. The MFC typically will result in a net transfer of acidity from the cathode to the anode. However, this study has shown that it is possible to limit the pH drop at the anode by making use of the natural buffering capacity of the reactive mixture. Associating a naturally alkaline reactive mixture (e.g., ammonia-rich streams, wastewater, and leachate from municipal and agricultural wastes), could help reduce the costs. Finally, the capital costs associated with this system can be seen as a major obstacle of the treatment system proposed by the present work; however, these are expected to drop drastically in the future as the MFC technology matures [31]. In a broader perspective, the treatment process developed in this study could be attractive as a sustainable alternative for the treatment of AMD dominated with Fe in remote areas. This could involve the precipitation of Fe prior to other challenging metals (e.g., Mn) or the co-precipitation of Fe-oxy(hydroxi)des with other dissolved metals in AMD. 4.5. Ongoing and future research Firstly, the long-term stability of membrane MFCs must be improved in the membrane design. The use of an anion exchange membrane or a bipolar membrane, as suggested by Ter Heijne et al. [14]. could alleviate the issue of membrane fouling. Secondly, scaled-up MFC testing systems operating in a continuous manner would allow a better evaluation of treatment efficiency and maximum power generation prior to field application. Thirdly, the tests should be performed using artificial AMD comprised with ferrous iron (Fe2+ ), which is the most likely form to be present at high concentrations in anaerobic AMD dominated with Fe, especially at low pH and limited dissolved oxygen concentrations. Fourthly, the substrate for anaerobic bacteria at the anode of the MFC should

O. Lefebvre et al. / Journal of Hazardous Materials 241–242 (2012) 411–417

be changed for a more realistic source in the form of wastewater or sludge from municipal wastewater treatment plants or leachate from agricultural activities. This could further impact the Coulombic efficiency of the system. Higher concentrations of calcium carbonate could also be tested as neutralizing and buffering agent in the anolyte composition to withstand the pH dropping over time. Finally, the recovery and the prospective use of the precipitate recovered at the end of the treatment process for practical applications, such as pigment for paints, should be considered to further reduce the overall treatment cost associated with the MFC technology. Acknowledgements We acknowledge the financial support from the Singapore National Research Foundation and the Economic Development Board (SPORE, COY-15-EWI-RCFSA/N197-1). References [1] D.W. Blowes, C.J. Ptacek, J.L. Jambor, C.G. Weisener, The geochemistry of acid mine drainage, in: B. Sherwood Lollar (Ed.), Treatise on Geochemistry. Environmental Geochemistry, Elsevier, Toronto, 2003, pp. 149–204. [2] C.C. Hustwit, T.E. Ackman, P.M. Erickson, The role of oxygen transfer in acid mine drainage treatment, Water Environ. Res. 64 (1992) 817–823. [3] D.K. Nordstrom, C.N. Alpers, C.J. Ptacek, D.W. Blowes, Negative pH and extremely acidic mine waters from Iron Mountain, California, Environ. Sci. Technol. 34 (2000) 254–258. [4] K.D. Nordstrom, C.N. Alpers, Negative pH, efflorescent mineralogy, and consequences for environmental restoration at the Iron Mountain Superfund site, California, in: National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare”, Irvine, CA, USA, 1999, pp. 3455–3462. [5] M.C. Moncur, C.J. Ptacek, D.W. Blowes, J.L. Jambor, Release, transport and attenuation of metals from an old tailings impoundment, Appl. Geochem. 20 (2005) 639–659. [6] C.L. Kairies, C.L. Capo, G.R. Watzlaf, Chemical and physical properties of iron hydroxide precipitates associated with passively treated coal mine drainage in the Bituminous Regions of PA and MD, Appl. Geochem. 20 (2005) 1445–1460. [7] C.M. Neculita, G.J. Zagury, B. Bussière, Passive treatment of acid mine drainage in bioreactors using sulphate-reducing bacteria: critical review and research needs, J. Environ. Qual. 36 (2007) 1–16. [8] B.E. Logan, B. Hamelers, R. Rozendal, U. Schrorder, J. Keller, S. Freguia, P. Aelterman, W. Verstraete, K. Rabaey, Microbial fuel cells: methodology and technology, Environ. Sci. Technol. 40 (2006) 5181–5192. [9] D. Pant, A. Singh, G. Van Bogaert, S. Irving Olsen, P. Singh Nigam, L. Diels, K. Vanbroekhoven, Bioelectrochemical systems (BES) for sustainable energy production and product recovery from organic wastes and industrial wastewaters, RSC Advances 2 (2012) 1248–1263. [10] R.A. Rozendal, H.V.M. Hamelers, C.J.N. Buisman, Effects of membrane cation transport on pH and microbial fuel cell performance, Environ. Sci. Technol. 40 (2006) 5206–5211.

417

[11] O. Lefebvre, Y. Shen, Z. Tan, A. Uzabiaga, I.S. Chang, H.Y. Ng, Full-loop operation and cathodic acidification of a microbial fuel cell operated on domestic wastewater, Bioresour. Technol. 102 (2011) 5841–5848. [12] H. Eom, K. Chung, I. Kim, J.-I. Han, Development of a hybrid microbial fuel cell (MFC) and fuel cell (FC) system for improved cathodic efficiency and sustainability: the M2FC reactor, Chemosphere 85 (2011) 672–676. [13] A. Ter Heijne, H.V.M. Hamelers, C.J.N. Buisman, Microbial fuel cell operation with continuous biological ferrous iron oxidation of the catholyte, Environ. Sci. Technol. 41 (2007) 4130–4134. [14] A. Ter Heijne, H.V.M. Hamelers, V. De Wilde, R.A. Rozendal, C.J.N. Buisman, A bipolar membrane combined with ferric iron reduction as an efficient cathode system in microbial fuel cells, Environ. Sci. Technol. 40 (2006) 5200–5205. [15] K. Chung, I. Lee, J.-I. Han, Effect of Fe-chelating complexes on a novel M2FC performance with ferric chloride and ferricyanide catholytes, Chemosphere 86 (2012) 415–419. [16] B.K. Min, S.A. Cheng, B.E. Logan, Electricity generation using membrane and salt bridge microbial fuel cells, Water Res. 39 (2005) 1675–1686. [17] B.E. Logan, Essential Data, Techniques for conducting microbial fuel cell and other types of bioelectrochemical system experiments, ChemSusChem 5 (2012) 988–994. [18] Y.J. Shen, O. Lefebvre, Z. Tan, H.Y. Ng, Microbial fuel-cell-based toxicity sensor for fast monitoring of acidic toxicity, Water Sci. Technol. 65 (2012) 1223–1228. [19] S. Oh, B. Min, B.E. Logan, Cathode performance as a factor in electricity generation in microbial fuel cells, Environ. Sci. Technol. 38 (2004) 4900–4904. [20] W.E. Balch, G.E. Fox, L.J. Magrum, C.R. Woese, R.S. Wolfe, Methanogens: reevaluation of a unique biological group, Microbiology Review 43 (1979) 260–296. [21] S. Ji, S. Kim, J. Ko, The status of the passive treatment systems for acid mine drainage in South Korea, Environ. Geol. 55 (2008) 1181–1194. [22] APHA, Standard Methods for Water and Wastewater Examination, 21st ed., American Public Health Association, Washington, 2005. [23] A. Ter Heijne, L.S. van Rijnsoever, M. Saakes, H.V.M. Hamelers, C.J.N Buisman, F. Liu, Performance of a scaled-up microbial fuel cell with iron reduction as the cathode reaction, J. Power Sources 196 (2011) 7572–7577. [24] S. Cheng, J.-H. Jang, B.A. Dempsey, B.E. Logan, Efficient recovery of nano-sized iron oxide particles from synthetic acid-mine drainage (AMD) water using fuel cell technologies, Water Res. 45 (2011) 303–307. [25] W. Stumm, J.J. Morgan, Aquatic Chemistry, Chemical Equilibria and Rates in Natural Waters, 3rd ed., John Wiley & Sons, Inc., New York, 1996. [26] B. Morgan, O. Lahav, The effect of pH on the kinetics of spontaneous Fe(II) oxidation by O2 in aqueous solution – basic principles and a simple heuristic description, Chemosphere 68 (2007) 2080–2084. [27] S. Cheng, B.A. Dempsey, B.E. Logan, Electricity generation from synthetic acidmine drainage (AMD) water using fuel cell technologies, Environ. Sci. Technol. 41 (2007) 8149–8153. [28] A. Ter Heijne, F. Liu, R. van der Weijden, J. Weijma, C.J.N. Buisman, H.V.M. Hamelers, Copper recovery combined with electricity production in a microbial fuel cell, Environ. Sci. Technol. 44 (2010) 4376–4381. [29] Z. Wang, B. Lim, H. Lu, J. Fan, C. Choi, Cathodic reduction of Cu2+ and electric power generation using a microbial fuel cell, Bull. Korean Chem. Soc. 31 (2010) 2025–2030. [30] G. Wang, L.P. Huang, Y.F. Zhang, Cathodic reduction of hexavalent chromium [Cr(VI)] coupled with electricity generation in microbial fuel cells, Biotechnol. Lett. 30 (2008) 1959–1966. [31] R.A. Rozendal, H.V.M. Hamelers, K. Rabaey, J. Keller, C.J.N. Buisman, Towards practical implementation of bioelectrochemical wastewater treatment, Trends Biotechnol. 26 (2008) 450–459.