Microbial desalination cells for energy production ... - Semantic Scholar

Desalination 308 (2013) 122–130

Contents lists available at SciVerse ScienceDirect

Desalination journal homepage: www.elsevier.com/locate/desal

Microbial desalination cells for energy production and desalination Younggy Kim 1, Bruce E. Logan ⁎ Department of Civil and Environmental Engineering, 212 Sackett Building, Penn State University, University Park, PA, 16802, USA

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

► Bacteria generate electricity while oxidizing organic matter in wastewater. ► Electricity produced can be used to desalinate salt water without any other energy input. ► Salt removal can be very high (>95%) but this requires a large amount of wastewater. ► Stacked ion exchange, forward osmosis, or bipolar membranes can be used in MDCs. ► Low current densities, pH, membrane integrity, and safety issues need further investigation.

a r t i c l e

i n f o

Article history: Received 24 March 2012 Accepted 20 July 2012 Available online 19 August 2012 Keywords: Bioelectrochemical systems Sustainable desalination Renewable energy Exoelectrogenic microorganisms Microbial fuel cells Electrodialysis

a b s t r a c t Microbial desalination cells (MDCs) are a new, energy-sustainable method for using organic matter in wastewater as the energy source for desalination. The electric potential gradient created by exoelectrogenic bacteria desalinates water by driving ion transport through a series of ion-exchange membranes (IEMs). The specific MDC architecture and current conditions substantially affect the amount of wastewater needed to desalinate water. Other baseline conditions have varied among studies making comparisons of the effectiveness of different designs problematic. The extent of desalination is affected by water transport through IEMs by both osmosis and electroosmosis. Various methods have been used, such as electrolyte recirculation, to avoid low pH that can inhibit exoelectrogenic activity. The highest current density in an MDC to date is 8.4 A/m2, which is lower than that produced in other bioelectrochemical systems. This implies that there is a room for substantial improvement in desalination rates and overall performance. We review here the state of the art in MDC design and performance, safety issues related to the use of MDCs with wastewater, and areas that need to be examined to achieve practical application of this new technology. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Global water shortages have increased the need for desalination, as demonstrated by the large increase in installed desalination capacity since the 1980s [1,2]. As the use of desalination to produce potable water has increased, there have been shifts in technologies and major ⁎ Corresponding author. Tel.: +1 814 863 7908. E-mail address: [email protected] (B.E. Logan). 1 Department of Civil Engineering, McMaster University, Hamilton, ON, Canada L8S 4L7. 0011-9164/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.desal.2012.07.022

advances in minimizing the energy consumption for this process, although the energy demands remain a challenge. The theoretical minimum energy for desalination of typical seawater (35 g/L of total dissolved solids) is ~1.0 kWh/m3, assuming a thermodynamically reversible process at 50% water recovery [3–5]. The most energy efficient systems for seawater desalination have recently achieved an energy requirement of only 1.8–2.2 kWh/m3, suggesting that these systems are operating very close to the minimum electrical energy demands [3,5,6]. When other energy requirements are included in energy demands, such as pumping feed water to the system and water pre-treatment, the overall energy consumption is increased to 3 to 4 kWh/m3 in recently built desalination

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

plants using reverse osmosis (RO) [3,5]. It is thought that further advances in reverse osmosis systems will not be able to reduce the overall energy consumption below 3 kWh/m3 [3]. One new method that can reduce or eliminate the need for electrical power for desalination is the microbial desalination cell (MDC). The main feature of the MDC is that exoelectrogenic microorganisms produce electrical potential from the degradation of organic matter, which can then be used to desalinate water by driving ion transport through ion-exchange membranes (IEMs). If wastewater is used as the source of the organic matter, the MDC can achieve three goals: desalination, energy production, and wastewater treatment. The energy in domestic wastewater typically ranges from 1.8 to 2.1 kWh/m3 [7,8], which is comparable to the minimum energy needed for practical desalination of seawater (1.8– 2.2 kWh/m3). In addition, domestic wastewater usually has a relatively low ionic concentration (b0.8 g/L as total dissolved solids [9], compared to 35 g/L for seawater). This shows that there is additional energy available from the salinity gradient between the wastewater and seawater. The use of MDCs represents a new approach for desalination, but the operational conditions and reactor designs have varied widely. Wastewater can be a good source for energy to desalinate salt water, but acetate has been used as the fuel for most studies in order to create uniform operating conditions for testing desalination aspects of the system performance. We review here the principles of water desalination in MDCs, discuss the various reactor designs used in tests, assess the effectiveness of the different approaches on performance, and describe the challenges for practical applications of MDCs as a sustainable method for water desalination. 2. IEM-based MDCs 2.1. Reactor design and principles The first MDCs proposed for water desalination consisted of three chambers: the anode, a middle desalination chamber, and the cathode (Fig. 1A) [10]. At the anode, exoelectrogenic microbes oxidize organic matter and transfer electron to the anode. At the cathode, oxygen is usually used as the electron acceptor, with electrons and protons, to form

123

water. These electrode reactions create an electric potential gradient or electric field between the electrodes by up to about 1.1 V (open circuit condition with acetate; pH= 7; partial pressure of oxygen in air of pO =0.2 atm). An anion-exchange membrane (AEM) is placed next to the anode, and a cation-exchange membrane (CEM) next to the cathode. Under the electric field, anions are attracted to the anode and cations are driven toward to the cathode, desalinating the water in the middle chamber for the given IEM configuration (Fig. 1A). The ionic separation can be magnified by using more than one membrane pair between the electrodes [11], similar to the stack design used for electrodialysis (ED) desalination systems [12]. The IEM stack consists of alternating AEMs and CEMs, creating repeating pairs of desalting (diluate) and concentrating (concentrate) cells (Fig. 1B). In this stacked desalination system, a single electron transfer at the electrodes can separate as many ion pairs as the number of repeated cell pairs. For instance, in a 3-cell pair stack, a single electron transfer separates 3 pairs of mono-valent cations and anions with ideally permselective IEMs (Fig. 1B). The internal resistance of an MDC or ED system increases with the number of IEM pairs in the stack. In an ED system, the applied voltage is controllable depending on the stack size. In an MDC, however, the voltage used for desalination is limited to that produced by the electrode reactions, and therefore the voltage per cell pair decreases with an increase in the number of cell pairs. The stack in an MDC should be carefully designed to provide the best match between the potential energy generated by exoelectrogens with oxygen reduction (b1.1 V), and the resistance of individual cell pairs. Chen et al. [11] found that the rate of desalination with 2 cell pairs was faster than that with 3 cell pairs, using a cell with an inter-membrane distance of 1 cm. In conventional ED systems, however, the inter-membrane distance is only 0.2–3 mm [12]. In many MDC studies [11,13–17], the inter-membrane distance has been relatively large (1 to 2.4 cm), resulting in very high internal resistances due primarily to Ohmic resistances in the desalination chamber. To improve performance by reducing the internal resistance and allowing for a larger number of membrane pairs, the inter-membrane distance should be minimized. 2.2. Junction potential and water transport While the electric potential gradient, created by the electrode reactions, is mostly responsible for the ion separation in MDCs, the rate of desalination is also affected by other important factors, such as IEM junction potential and water transport across an IEM. These factors generally reduce desalination efficiencies in stacked systems having concentrate cells. However, in three-chamber MDCs, these factors work as additional driving forces for desalination. In three-chamber MDCs, wastewater in the anode chamber has a much smaller ion concentration than the desalination chamber. Typical domestic wastewater has an ion content of ~0.5 g/L and rarely exceeds 0.9 g/L of total dissolved solids [9]. Typical seawater in the middle chamber, with a total dissolved solid concentration of 30 to 40 g/L, creates a significant concentration gradient across an IEM with wastewater. Due to this concentration gradient, salt ions in the middle chamber of a MDC are driven to the adjacent anode and cathode chambers. This driving force between seawater (sea) and wastewater (ww) across an IEM can be quantified as junction potential (Δφjct) [18]:
!


RT
t ai;sea




i
∑ ln

Δφjct
¼ F
i zi ai;ww


Fig. 1. Schematic diagram for ionic separation in microbial desalination cells, (A) threechamber design and (B) stacked IEM design with three diluate and concentrate cell pairs.

ð1Þ

where R is the gas constant, T the absolute temperature, F Faraday's constant, z the ionic charge, and a the activity of ionic species i. The transport number (t) is defined as the fractional contribution of ionic flux to current density in the IEM. For an ideally permselective CEM in a NaCl solution, for instance, tNa+ will be unity and tCl− is zero.

124

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

Assuming that the activity ratio (ai,sea/ai,ww) equals the total dissolved solids ratio of typical seawater and domestic wastewater (35 g L −1/ 0.5 g L −1 = 70), the sum of junction potentials in a three-chamber MDC is approximately 0.2 V (25 °C; 90% permselective IEMs). This potential energy is ~30% of the potential energy generated by the electrode reactions (0.5–0.6 V [19]). Mehanna et al. [15] found that the concentration gradient contributed to the conductivity reduction in the desalination chamber by up to 68%. The concentration gradient across an IEM also induces osmotic water transport, and the osmotic water transport dilutes the middle chamber, improving the water recovery in the three-chamber MDCs. Osmotic pressure (π) is approximately 14.4 atm (212 psi) for an expected concentration difference (Δc = 35–0.5 = 34.5 g L −1) in an MDC (25 °C; total dissolved solids as NaCl) using π ¼ ΔcRT:

ð2Þ

Commercial IEMs are not designed to hold such a high pressure, and therefore osmotic water transport should be allowed to release the pressure during MDC operation. Osmotic water transport from the anode and cathode chambers to the desalination chamber dilutes the salinity of the middle chamber, improving desalination performance in a three-chamber MDC. Jacobson et al. [20] observed that 80.4 mL of water was transported into the middle chamber under open circuit conditions over 2 days (0.2 m 2 AMI-7001 and 0.27 m 2 CMI-7000), resulting in up to a 7% reduction of salinity (initially 35 g/L NaCl). While the osmotic water transport improves desalination performance of three-chamber MDCs, it should be noted that the most important driving force is the electric current generated by the exoelectrogenic microbes, as its contribution to the final extent of desalination ranged from 81 to 98% [21]. Another driver for water transport across an IEM is electroosmosis, where water transport occurs as a result of ions transported through the nano-scale pores in IEMs. In addition to solvated water molecules around individual ions, ionic movement also drag water molecules in IEM pores, transporting water in the direction of the ionic flux. As a result of this additional transport of water, electroosmosis reduces the water recovery of desalination and concentrates desalinated water.

Even though the effect of electroosmosis has not been clearly quantified in three-chamber MDCs, Jacobson et al. [20] reported increasing water volume in the middle chamber, implying that osmotic water transport dominated over that by electroosmosis in their experiments at current densities b 0.72 A/m 2. However, with the relatively low current densities, this result does not necessarily mean that osmosis is always dominant over electroosmosis in MDCs, because electroosmotic water transport increases with increasing current density through IEMs [22] and osmotic water transport can be reduced with decreasing hydraulic residence time. In three-chamber MDCs, junction potentials and osmosis improve desalination efficiencies because the ion concentration in wastewater is smaller than that in seawater. However, in a stacked MDC, these factors adversely affect desalination efficiencies since the ion concentrations in the concentrate cells are higher than those in the diluate cells as a result of desalination in the stack. With a 5-cell paired stack between a microbial anode and an air cathode, junction potential losses consumed up to 18% of the total potential created by the electrode reactions [19]. The junction potential losses resulted from the salinity difference between the concentrate (71 mS/cm) and the diluate (31 mS/cm). In a stacked system, osmosis and electroosmosis induce water transport through an IEM in the same direction, i.e., from the diluate to concentrate cells, reducing the water recovery and concentrating diluate stream in the stack. Osmosis and electroosmosis in this 5-cell stack resulted in an increase in the concentrate volume by 30%, with electroosmosis responsible for 65% of this increase (current density ~4 A/m2).

2.3. Performance 2.3.1. Salinity removal Salinity removals in three-chamber MDCs can be above 90% from 30–35 g/L NaCl solutions that have conductivities similar to seawater (Table 1, Fig. 2). However, very high salinity removals have required a large volume of non-salty water, primarily the anolyte but also the catholyte, with 55 to 133 times the volumes of desalinated water (Fig. 2). The use of stacked MDCs can reduce the need for large amounts of non-salty electrolyte. Up to 98% salinity removals from

Table 1 Comprehensive summary of MDC studies. (Reactor types, 3C: three-chamber MDC; Stack: stacked MDC; Bipolar: bipolar MDC; Osmotic: osmotic MDC.) Reactor types

Anolyte Sodium acetate (g/L)

PBS (mM)

Vol. (mL)

COD removal (%)

Reaction

Buffer

Vol. (mL)

NaCl (g/L)

Vol. (mL)

Salinity removal (%)

CE (%)

3C 3C 3C 3C 3C 3C 3C 3C Stack Stack Stack 3C 3C Bipolar 3C Osmotic Osmotic

1.6 4 4 3 3 1–2 2 1 1.64 1 1 –b 1c 1 4 4 4

50 10 10 10 10 50 50 50 47 100 100 None 25 50 10 10 10

300 1.0 L/da 1.0 L/da 5.7 L/da 5.7 L/da 14 14 100c 1000 120 3800 140 28 30 735 735 735

– – – 92 92 77–82 38–54 – – 91 91 53.8 74–79 – – – –

Fe(CN)63− ORR ORR ORR ORR ORR HERd HERe ORR ORR ORR Fe(CN)63− ORR ORRf Fe(CN)63− Fe(CN)63− Fe(CN)63−

PBS H2SO4 H2SO4 H2SO4 H2SO4 PBS PBS PBS PBS None None PBS PBS None PBS PBS PBS

100 4.3 L/da 4.3 L/da 5.8 L/da 5.8 L/da 14 27 144 300 72 2300 140 14 30 735 735 735

35 30 30 35 35g 20 20 10 20 35 35 14.2h 20 10 20 20 35g

3 0.1 L/da 0.4 L/da 0.2 L/da 0.2 L/da 14 14 17 21 60a 300a 60 14 10 75 75 75

93 100 11 94 74 63 37 98.8 72.1 44 98 66 34–55 86 41 60 63

– 11–17 11–17 – – 53–68 16–48 – – 80 80 25.4 22–52 97 – – –

a b c d e f g h

Fed by continuous flow. Wastewater. Xylose. Eap 0.55 V. Eap 0.8 V. Eap 1.0 V. Sea salt. Mixture of NaCl and NaHCO3.

Catholyte

Salt water

Ref.

[10] [21] [21] [20] [20] [15] [14] [13] [11] [19] [19] [17] [16] [34] [42] [42] [42]

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

125

of wastewater is also dependent of the initial salinity of salt water in MDCs.

Fig. 2. Salinity removal in MDCs from initial NaCl concentration of 20 g/L or higher. The number on the top of each bar indicates the ratio of used non-salty electrolyte solution to desalinated water volume.

35 g/L NaCl were achieved using stacked MDCs (5 cell pairs) with a reduced anolyte volume (13 times the desalinated water) [19]. MDCs can be more effective for partial removals of salinity. For instance, the required wastewater was only two to three times the desalinated water while MDCs achieved 40 to 60% salinity removal (Fig. 2). These results imply that, for practical applications, MDCs are more likely to be used for partial salt removal from seawater. MDCs can also be used for brackish water desalination. Luo et al. [13] achieved up to 99% removal from 10 g/L NaCl while the volume ratio between non-salt electrolyte and desalinated water was ~ 14 (Table 1). Similarly, 86% removal from 10 g/L NaCl was reported at the volume ratio of 6. These results indicate that the required amount

Fig. 3. (A) Maximum current normalized by cross-sectional area of current (normalized by AEM area for cylindrical MDCs) and (B) maximum power normalized by anode chamber volume.

2.3.2. Maximum current vs. maximum power The operational conditions chosen for an MFC (microbial fuel cell) or an MDC always represent a tradeoff between maximizing current density, and therefore the rate of treatment, and the power density. More rapid desalination is achieved under higher current conditions; from the perspective of minimizing reactor hydraulic retention times and maximizing rates of treatment, this is a preferable operational condition. Maximum current densities reported for MDCs have ranged from 0.7 to above 8.4 A/m 2 (Fig. 3A). In many of these studies the use of higher current densities (with less power recoveries) could have likely produced better desalination results. High current conditions can be achieved by using a low external resistance (b 10 Ω), or by short-circuiting the electrodes. However, Chen et al. [11] reported that smaller external resistances (5 Ω) did not necessarily result in higher current densities. The lack of improved performance at lower resistances could be due to unfavorable anode potentials, which can adversely affect current generation by exoelectrogenic bacteria in MFCs [23–25]. At lower external resistances, power densities can rapidly drop off and current densities can double back to lower values, a phenomenon called Type D power overshoot [26]. However, power overshoot can be overcome through proper acclimation of the reactors to lower resistances over time [27]. The other choice for MDC operation is to maximize electrical power production. When operated under maximum power conditions, MDCs have generated up to 65 W/m 3 (normalized by anolyte volume), although this required the use of a ferricyanide catholyte [10] (Fig. 3B). With oxygen reduction reaction, the highest power reported for MDCs was 31 W/m 3 [21]. Future experiments with MDCs should be designed to address the tradeoff in these operational conditions. If electrical power costs are low, it is likely that the economics of the process will favor achieving higher current densities than maximum power densities. The power and current relationship of MDCs often forms a semicircle (Fig. 4), similar to other fuel cell systems [19–21]. In this situation, the maximum power operation will induce currents about a half of the maximum possible value, decreasing the rate of desalination. However, as mentioned by Jacobson et al. [21], the harvested electric energy from MDCs can be used in downstream desalination processes, such as reverse osmosis processes.

Fig. 4. Correlation between current and power generation from stacked MDCs. The power is normalized by the cross sectional area of the cathode (7 cm2), and 1000 mW/m2 is equivalent to 23.3 W/m3 normalized by anolyte volume (30 mL). (Reprinted with Permission from Kim and Logan [19]. Copyright (2011) American Chemical Society.)

126

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

2.3.3. Current efficiency Current efficiency (ηi) is the amount of separated ions divided by the amount of electrons transferred at the electrodes in MDCs [28], and it is calculated as: ηi ¼

FzVΔc NCP ∫idt

ð3Þ

where Δc is the reduction in the concentration of the salt water, V the volume desalinated, NCP the number of cell pairs, and i the electric current. Current efficiencies have varied from 50 to 120% in MDCs (Table 2). While ED systems typically have current efficiencies ~ 90% [12], the wide variation in the current efficiencies in MDCs, even above 100%, implies that ionic separation is greatly influenced by osmotic water transport to the middle chambers, diluting salty water and increasing current efficiencies [11]. The current efficiency is a very important parameter in IEM separation processes. The integrity of used IEMs can be often estimated in ED by measuring the current efficiency. In well-controlled ED systems, the current efficiency should be between 90 and 100% [29]. For instance, current efficiencies below 90% indicate failures in IEM integrity or significant current losses along feed water channels [30]. Current efficiencies >100% reported in MDCs [11] (Table 2) must have occurred due to significant water transport in through the IEMs, resulting in reduced salt concentrations. While the current efficiency was not always reported in previous MDC studies, we encourage that this parameter be reported in future studies to confirm the IEM integrity, and to estimate the influence of water transport on desalination performance. 2.3.4. Coulombic efficiency The Coulombic efficiency (CE) is the ratio of total electrons transferred from the anode to the maximum possible electrons generated from the oxidation of substrate [31], and it is calculated as: CE ¼

MO2 ∫idt ne FV an ΔCOD

discharge. Identified exoelectrongenic microbes in MDCs (Geobacter sulfurreducens and Pelobacter propionicus) do not differ from those active exoelectrogens in MFCs [14], suggesting from a microbiological perspective that MDCs have a very similar capability to MFCs for wastewater treatment. Thus, MDCs can be used for treating various types of wastewater that have been proven to work for MFCs [32]. 2.3.6. Effects of electrolyte pH in MDCs The separation of the electrodes into two different chambers using IEMs in MDCs can create pH imbalances. Protons are produced at the anode as a result of oxidation of organic matter, while hydroxyl ions are generated by cathodic reactions, such as oxygen reduction or hydrogen evolution. In a single chamber MFC, overall pH is neutralized by mixing within the reactor, maintaining a bulk neutral pH (although there can be local gradients at the electrodes). In MDCs, however, the majority of the ionic flux through the IEMs, which is needed to balance charge, is due to salt ions such as sodium and chloride. Thus, protons remain in the anode chamber and hydroxyl ions in the cathode chamber, which can create substantial pH imbalances. The most deleterious pH change is at the anode, where a low pH can seriously reduce or eliminate current generation. Despite the use of a 100 mM phosphate buffer, the anolyte pH in a stacked MDC decreased to 4.7 over a fed-batch cycle, eventually eliminating current generation [19]. This limited performance at a moderately low pH is consistent with the reported sensitivity of exoelectrogens in MFCs. He et al. [33] reported that the exoelectrogenic activity was substantially inhibited at pH 5 compared to that at pH 6. This microbial sensitivity to pH indicates that the anolyte acidification is one of the most critical limiting factors of MDC applications. In the absence of a buffer, the catholyte can increase to a pH of 12 [19,34], but a pHs above 10 have so far not been shown to adversely affect power production [35]. The effect of catholyte pH on maximum possible potential of the electrode reactions can be calculated using the Nernst equation as: 0

ð4Þ

where Van is the volume of the anode chamber, ΔCOD the change in chemical oxygen demand during MDC operation, MO2 the molecular mass of O2 (32), and ne the required number of electrons to reduce O2 to water (4). The Coulombic efficiency indicates how much of the substrate was used to produce current by the exoelectrogenic bacteria. Coulombic efficiencies in MDCs have varied over a wide range, and have been as high as 97% (Table 1). This very high Coulombic efficiency was likely a result of the anode chamber being well isolated from oxygen leaking through the air cathode due to the presence of multiple IEMs (Fig. 1). 2.3.5. COD removal The majority of MDC studies have been performed with an easily biodegradable substrate, such as acetate or xylose, and the resulting COD removals have been relatively high (mostly over 70%) (Table 1). In one recent study that used domestic wastewater as an anolyte solution, there was only 54% COD removal in a three-chamber MDC [17]. This result confirms the capability of MDCs in treating wastewater, but it is not clear if this would be a sufficient level of COD removal for wastewater

Table 2 Current efficiencies reported in MDC studies. Current efficiency (ηi)

Ref.

50% 93% 81–99% 51–120% 72%

[10] [19] [21] [11] [20]

E¼E þ

RT fOxg ln zF fRedg

ð5Þ

where {Ox} is the oxidizing species (e.g., O2 or H+) and {Red} is the reducing species (e.g., H2O or H2) in the half-cell electrode reactions. For oxygen reduction (O2 + 4 H+ + 4e− → 2H2O) and hydrogen evolution reactions (2H+ + 2e− → H2), a one unit pH increase (i.e., 10-fold decrease in H+ activity) could produce a 59-mV reduction in cell voltage (25 °C). For example, a catholyte pH 12 will decrease cell voltages by (12–7)× 59= 295 mV compared to neutral pH under open circuit conditions. One way to solve the pH imbalance is to recirculate the anolyte and catholyte solutions [16]. Recirculating these solutions in a three chamber MDC stabilized the anolyte pH above 6 even with a relatively low buffer concentration (25 mM, phosphate). However, the authors found that the Coulombic efficiency decreased from 60 to 22% with the recirculation (0.1 mL/min, Van = 28 mL, Vcat = 14 mL) because this operation cycled substrate into the cathode chamber where oxygen was available for aerobic substrate degradation by microorganisms. Other flow rates were not examined, and thus it may be possible to better optimize recirculation rates for minimizing substrate losses while one can maintain pH control. 2.3.7. Salinity effects on exoelectrogenic activity During MDC operation, chloride ions are transported from salt water through the AEM into the anode chamber (Fig. 1). As a result, the anode chamber can become enriched in chloride ions, which can adversely affect exoelectrogenic microbes. For instance, oxidation of wastewater containing 1000 mg/L COD can result in the transfer of 125 mM of chloride ions into the anode chamber (Eq. (4)). In MFCs, power and current generation have not been adversely affected by chloride concentrations of up to 300 mM [36,37], but maximum power was decreased by 12% at

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

500 mM [37]. Oh and Logan [37] also found that increasing KCl concentrations improved power generation, but only up to 300 mM. These results suggest that exoelectrogenic activity would only be adversely affected by very high concentrations of anions produced when using high concentrations of organic matter in the anode chamber. The addition of the salt to the anode chamber can affect the bacterial communities that develop in MDCs, relative to those in MFCs. Mehanna et al. [14] found that microbes with sequences most similar to P. propionicus were reduced, and those of G. sulfurreducens increased, with increasing NaCl concentration in MDCs. This suggests that a given salinity level can provide more favorable conditions for certain types of microbes than others. 2.3.8. Cathode reactions: O2 reduction vs. H2 evolution Three different cathodic reactions have been used for MDCs: oxygen reduction reaction (ORR), hydrogen evolution reaction (HER), and ferricyanide reduction (Table 1). While the use of ferricyanide increases power production relative to oxygen (Fig. 3B), ferricyanide is not a sustainable option for MDC operation. With oxidation of acetate at the anode, ORR could provide up to 1.1 V (open circuit condition; neutral pH; 0.2 atm O2) [38], but available potential energy for desalination is 0.5–0.6 V due to losses by electrode overpotentials [19]. HER is not a thermodynamically favorable reaction with acetate oxidation, and therefore MDCs with HER require an external potential energy of more than Eap =0.11 V (at neutral pH). However, the invested external energy can be offset by recovery of H2 produced at the cathode. One benefit of HER in MDC operation is that current densities can be substantially increased by using higher applied potentials. In a single chamber MEC (microbial electrolysis cell), current densities can be very high, for example, 36 A/m2 [39]. The highest current density with HER in an MDC study was only 5 A/m2 (Eap = 0.8 V), which is even lower than that obtained in MDC studies with ORR (Fig. 3A). This comparison implies that there is room for improvement in the current generation in HER-based MDCs.

127

or not, it is clear that using FO membranes will reduce energy needs for downstream RO by accomplishing dilution of the feedwater. In MODCs, desalination is accelerated by water introduction to the salt water chamber from mainly the anode chamber. Conductivity reduction (which was contributed by both the water transport and salt removal) was 60% using 20 g/L NaCl, and 63% with 35 g/L of a sea salt solution (Fig. 2) [42]. In the study [42], three-chamber MDCs (with an AEM replacing the FO membrane) showed greater reduction in total NaCl in the middle chamber than the MODCs. However, the conductivity reduction in the three-chamber MDCs was only 41% with 20 g/L NaCl (Table 1), indicating the improved conductivity removal (60%) was due to water transport through the FO membrane. The water volume in the middle chamber of MODCs was twice the initial volume with 20 g/L NaCl and 35 g/L sea salt solutions [42]. This result gives a remarkable desalinated water recovery of 200%. However, it should be noted that the consumed volume of anolyte (10 mM phosphate buffer) and catholyte (100 mM ferricyanide+ 100 mM phosphate buffer) was about 10 times the desalinated water (Fig. 2). Future studies should be directed toward minimizing the use of anolyte and catholyte solutions. The main technical challenge with using FO membranes in an MODC with real wastewaters will be fouling. The dilution of the seawater requires a high flux of water from the wastewater through the membrane, while avoiding carryover of ammonia, organic molecules, viruses, and bacteria. In one MODC test, the electrical resistance of the FO membrane increased by more than 200% (from 1.9 to 6.0 Ω) after 10 days of operation [42]. Although this is a large increase, the final

3. Other types of bioelectrochemical systems for desalination 3.1. Osmotic MDCs The use of forward osmosis (FO) in wastewater treatment was originally proposed as a method to dilute a brine solution using wastewater in order to enhance water recovery from brine solutions in downstream RO processes [40,41]. In FO processes, brine solutions are called draw solutions, as they pull water from relatively dilute feed solutions through the FO membrane. In osmotic MDCs (MODCs), seawater is the draw solution and wastewater is used as the feed solution. The low ion concentration of a typical domestic wastewater (0.25 to 0.85 g/L of total dissolved solids) [9] creates a significant osmotic pressure difference with seawater. An MODC has the same reactor design as a threechamber MDC, except that the FO membrane replaces the AEM separating the anode and middle chambers [42] (Fig. 5A). This membrane maximizes osmotic water transport from the anode chamber to the middle chamber, diluting the middle chamber salinity and increasing water recovery. However, by replacing the AEM with a FO membrane, the system loses the capability of selectively separating anions, decreasing the rate of overall ionic separation from the middle chamber with reduced current efficiencies. The benefits of using FO membranes in bioelectrochemical systems (BESs) with respect to power densities have not been well investigated. Ionic flux across a FO membrane is thought to be due to non-ideal salt rejection of the membrane. Thus, FO membranes with higher salt rejection are expected to increase the internal resistance of BESs. Researchers have claimed that the FO membrane increased power generation in two-chamber MFCs compared to those with a CEM [43,44]. However, a more useful comparison would have been to an MFC with an AEM, as the AEM provides more favorable conditions for current generation with a salinity gradient. Whether the FO membrane increases power

Fig. 5. Schematic diagram for (A) ionic separation and water transport in osmotic MDCs and (B) ionic separation and acid/base production in bipolar membrane MDCs.

128

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

resistance was still very small relative to other internal resistance components of the MDC. In another MODC study, there was a surprising result that better current densities were obtained with fouled FO membranes than with clean FO membranes [44]. These contradictory results indicate that the use of FO membranes in BES systems is not clearly understood, especially relative to the role of the membrane in ionic transport. Fouling can lead to a reduction in the water flux over time, but this aspect has not been sufficiently addressed in MODC studies. Fouling issues for MODCs should be very similar to other bioreactor systems that incorporate FO membranes. Fouling has been effectively addressed in lab-scale osmotic membrane bioreactors by reversing the water flux (osmotic backwashing) [45]. In this particular study, irreversible fouling was found to occur only in the early stages of operation (2 weeks), as the water flux was fully restored with osmotic backwashing thereafter. These issues of fouling and cleaning need to be more carefully addressed in future MODC studies.

was good production of hydrochloric acid in the bipolar MDC, with a final of 0.2 N HCl (pH = 0.7) [34]. When the applied external voltage decreased from 1.0 to 0.3 V, the produced acid decreased to 0.03 N (pH = 1.5), demonstrating that sufficient external power is needed to ensure quality of the produced chemicals. Without an external voltage supply, the acid chamber pH was neutral, confirming that the given electrode reactions by themselves cannot drive the water splitting reaction with a bipolar membrane. In the cathode chamber, pH increased up to 12.9 (0.08 N NaOH) with an external voltage of 1.0 V [34]. A basic solution of similar quality (pH =12.2) was also produced in the cathode chamber of a MDC or a microbial reverseelectrodialysis cell lacking the bipolar membrane in the absence of a pH buffer [19,50], and the pH was very high, reaching 13.93 in another study [51]. These results show that acid and base production using bipolar membrane MDCs can be more efficient at chemical production when the reactor design and operation are optimized. 4. Challenges and perspective

3.2. Bipolar membrane MDCs 4.1. Control of pH in MDCs Additional modifications of MDCs have been proposed to achieve not only desalination, but also simultaneous production of acid (HCl) and base (NaOH) solutions. A bipolar membrane was placed into the MDC next to the anode chamber, creating a four-chamber MDC (Fig. 5B) [34]. A bipolar membrane consists of a CEM and an AEM laminated together to form a single membrane. Instead of ions crossing the membrane, water is split at the interface of the laminated IEMs into protons and hydroxyl ions with the application of a sufficiently large electrical potential gradient. When organic matter is degraded in the anode chamber, OH − is released from the bipolar membrane into the anode chamber and H + into the adjacent chamber, forming hydrochloric acid (HCl). Salt water is desalinated in the chamber next to the cathode, and sodium hydroxide is produced in the cathode chamber. The bipolar membrane therefore plays a key role in avoiding a decrease in the pH of the anode chamber, which is a major challenge for MDC operation. One limitation of the process is that voltage must be added when using a bipolar membrane MDC. The theoretical minimum potential energy to split water is 0.83 V ([H+]=[OH−]=1 M; 25 °C) [46,47], with >1.2 V needed in practice [48,49]. Cell voltages produced by air-cathode MFCs alone are therefore insufficient to the needed potentials for using a bipolar membrane. Consequently, an applied voltage of up to 1 V was used in this bipolar membrane MDC for production of these chemicals and desalination [34]. If the value of these chemicals is sufficiently high, then the costs for the bipolar membrane and applied potentials may be recovered. Salinity removals in MDCs with a bipolar membrane MDC depend on the applied potential, with up to 86% removal using a 10 g/L NaCl solution at an applied voltage of 1 V (Table 1), but only ~ 50% removal at 0.3 V [34]. When the external voltage was removed, the salinity removal dropped to ~ 5%, indicating that water splitting at the bipolar membrane played an important role in desalinating the water. The extent of desalination in a four-chamber bipolar membrane MDC can be greater than that of a conventional three-chamber MDC due to maintenance of the anode pH [34]. Salinity removal in a three-chamber MDC was 60%, which was smaller than the 86% achieved by using a bipolar membrane. This improved performance was due to the stable anolyte pH of ~7 in the bipolar membrane MDC throughout the fed-batch cycle. In contrast, the three-chamber MDC had an acidified anolyte (pH of 4.5) at the end of fed batch cycle, resulting in a drop in current generation at the low pH, and incomplete substrate removal due to the shortened cycle time. This comparison implies that bipolar membranes can be used in other BESs to provide neutral pH conditions for exoelectrogenic microbes. Even with the need for an external power source, MDCs with bipolar membranes can be very efficient at chemical production. There

The decrease in the pH of the anode solution is a major challenge for practical applications of MDCs. A pH below 5 can prevent exoelectrogenic microbes from generating current. To avoid low pH, most MDC studies have been conducted with high concentrations of a pH buffer (e.g., 50 to 100 mM phosphate solutions) (Table 1), equivalent to 25 to 50 mN of alkalinity at an initial pH=7. This is much higher than that of a domestic wastewater, which typically has an alkalinity of 1.2–2.4 mN (60 to 120 mg/L as CaCO3) [9]. The organic matter concentration of domestic wastewaters (120 to 380 mg/L, biochemical oxygen demand [9]) is much lower than that of the acetate solutions typically used in MDC tests, resulting in an alkalinity (mg/L as CaCO3) to BOD ratio of 0.16 to 1. In typical MDC tests with 1 g/L sodium acetate (780 mg/L BOD), the ratio is much greater at 1.6 to 3.2. This comparison shows that the use of actual wastewaters in MDCs will likely not improve the situation relative to pH decreases in the anode chamber. Recirculation of anolyte and catholyte can stabilize electrolyte pH, but it reduces the Coulombic efficiency with oxygen reduction reaction at the cathode [16]. One possible solution for this low Coulombic efficiency is to have hydrogen evolution at the cathode rather than oxygen reduction. Under anaerobic conditions in the cathode chamber, recirculation would not affect the Coulombic efficiency. In addition, when designing recirculation systems, one should make sure to minimize ionic flux through recirculation tube lines. This ionic flux will reduce the current flowing through IEMs, decreasing the current efficiency. One way to avoid this short circuiting is to use periodic recirculation, resulting in no flow of ions through the tubing in between pumping periods. An alternative method for pH control in MDCs with oxygen reduction at the cathode is to blend the catholyte effluent, which will have a high pH, with wastewater flowing into the anode chamber. Oxygen reduction at the cathode will produce an equivalent number of moles of hydroxyl ions as the protons released at the anode. Thus, blending this stream into the anolyte could neutralize pH. A high anolyte pH is not an issue for the MDC operation, as exoelectrogenic activity is not adversely affected by pHs up to 10 [33]. This suggests adding catholyte into the anolyte could effectively control pH in MDCs, although the catholyte volume should be kept small in order to minimize dilution of organic matter concentration in the incoming wastewater. Another approach for avoiding a loss of current in the anode with proton production is to seed the anode with acidophilic exoelectrogenic microorganisms, such as Acidiphilium cryptum. MFCs inoculated with A. cryptum produced open circuit potentials ~0.3 V at pH= 4 and generated power up to 12.7 mW/m 2 [52]. This result suggests that MDCs could also be operated under low pH conditions, although this approach has so far not been tested in MDCs.

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

4.2. Improving the performance of stacked MDCs In stacked MDCs, osmotic water transport should be minimized in order to avoid water losses from the diluate into the concentrate cells, as this increases the salinity of the diluate and reduces water recoveries. Preventing this water loss is important, as it was shown in tests with a 5 cell-paired stack that water transport limited the possible extent of desalination in the MDCs [19]. To reduce osmotic water transport, hydraulic residence times (HRTs) in the stack should be minimized by increasing salt water flow rates. To maintain sufficient desalination in the diluate effluent, current densities would also need to be increased. The highest current density observed in MDCs so far is only 8.4 A/m 2 (Fig. 3A) [34], even though they can be much higher in MECs (e.g., 27 A/m 2 [53]; 36 A/m 2 [39]). At high current conditions, ionic separation will be fast enough to effectively reduce the salt water residence time in the IEM stack, minimizing the osmotic water transport. The amount of water transported by osmosis is proportional to the IEM surface area in the stack. As a result, water transport will be reduced in proportion to reductions in the total cross-sectional area of IEMs in the stack. However, reducing the cross-sectional area may increase the electrical resistance of the stack, which would increase electrical potential losses. Current densities in MDCs are much lower than those used in conventional ED processes, which usually exceed 100 A/m 2 [12]. This suggests that surface areas in MDCs could be reduced by an order of magnitude, without any significant limitation in current flowing through the IEM stack, although this would likely require much higher applied voltages than those that have traditionally been used in MDC stacks. Reducing the area of the IEM would not only decrease water losses by osmotic water transport, but also reduce the capital costs of MDCs. IEMs cost ~$150 US per m 2, and therefore contribute to ~ 80% of the initial cost for ED desalination processes [12]. However, these costs are likely high due to relatively small markets for IEMs compared to other types of pressurized membranes such as RO membranes. Thus, the cost of the IEMs could likely be reduced with higher production rates and improvements in materials and production methods. An alternative solution for minimizing water transport in stacked MDCs would be the development of new types of IEMs that have reduced water contents. Water transport by electroosmosis decreases with the water contents of the IEMs [22]. Commercial IEMs are usually manufactured for minimizing electric resistance so that losses in potential energy are small under high current conditions in conventional ED systems. MDCs do not generate very high electrical current densities compared to conventional ED processes, and therefore electrical potential energy losses due to the IEMs are less sensitive to membrane resistance. This insensitivity suggests that IEMs could be more specifically designed for MDCs to minimize water transport even though the low water content sacrifices high ionic conductivities. 4.3. IEM integrity under high microbial activities The AEM adjacent to the anode chamber is exposed to various microorganisms and organic compounds. Thus, the growth of a biofilm on this surface is inevitable, and this could alter or change the chemistry of the functional groups of the membrane, or compromise its polymeric structure. There is little experience with the use of these membranes in conventional ED for conditions typical of wastewater and biological treatment applications, and therefore it is difficult to predict long-term performance of IEMs under these conditions. In one study, a liter-scale MDC was operated over 8 months with synthetic wastewater (acetate as substrate), but the authors did not examine the integrity of used IEMs, for example through changes in permselectivity, following this period of operation [20]. While a biofilm will form on an IEM, it is not clear to what extent the presence of this biofilm will affect long-term operation. In one

129

study, performance of the cathode was examined over a 14 month period, and over time power generation was decreased by 22% and a thick biofilm was observed on the cathode [54]. It was concluded, however, that the cathode biofilm did not appreciably affect overall performance as removal of this biofilm did not improve performance. The main reduction in performance was due to internal fouling of the catalyst. Biofilm formation also did not affect current generation in an MODC, even though fouling of the FO membrane reduced water flux [42]. Thus, we infer from these studies that biofilm development on the IEM will not be critical to the long term performance of the IEMs. Coatings on the AEM, which help to reduce organic matter transport into the IEM, however, might be beneficial to the stability and extended performance on the AEM. 4.4. Water safety There is no information available yet on water safety issues for MDCs. However, we can assess the potential for crossover of contaminants in wastewater into the desalinated water. When IEMs are used, the primary barrier will be the AEM, and therefore the properties of this membrane will affect the safety of the desalinated water. To be exclusive to co-ions (e.g., cations for AEMs), IEMs must have very small pore sizes, in the range of 1 nm [12,29]. Due to this small pore size, even counter-ions (e.g., anions for AEMs) cannot move through an IEM if their molecular weight is greater than 350 Da [12]. Considering typical virus sizes range from 20 to 300 nm [55], AEMs should provide an effective barrier against viruses or larger pathogenic bacteria during MDC operation, as long as the physical integrity of the membrane is maintained. The AEM will be less effective in inhibiting the crossover of organic substrates or inorganic ions through the AEM. Dissolved organic compounds with negative or neutral charges, such as glucose (180 Da) and acetate (60 Da), can be transported through the AEM. The rate of transport can be accelerated by osmotic water transport to the salt water chamber. Organic matter that is moved through the AEM could support microbial growth within the desalinated water stream, potentially enhancing the growth of pathogenic microorganisms that might be present in that water stream. There have been no studies of organic matter crossover through AEMs in MDC studies, but crossover of acetate is known to occur in MFCs [56]. Transport of ions and organic matter will be much more substantial when using FO membranes than that for IEMs. Further study is needed to better understand and estimate the possible degree of contamination of the desalinated water, and to develop effective downstream treatment processes to enhance the removal of such contaminants. 5. Outlook MDCs are new technologies that are just beginning to be developed. Many of the challenges for designing cost-efficient MDC systems can be anticipated from experience gained with MFCs and MECs, and also from ED, RO, and FO systems. However, MDCs will have unique challenges and opportunities for development compared to these other systems. The IEMs will have lower current densities than EDs, and thus can likely be designed differently from those used in conventional ED systems. The integration of MDCs into a wastewater treatment plant will present greater challenges for biofouling, but these may be offset by helping to improve wastewater treatment efficiency or reduce the overall cost of the combined wastewater-water treatment plant. When potable water production is needed, MDCs will best be used as a pre-treatment step for RO. Reducing the salinity will proportionally decrease the energy requirement for downstream RO. If desalinated water produced by MDCs is used for other purposes, such as for irrigation or more conventional grey water applications, then full desalination may not be needed and partial desalination of a brackish

130

Y. Kim, B.E. Logan / Desalination 308 (2013) 122–130

water might be acceptable. This suggests that MDCs may have multiple applications and uses, and that different types of systems may be developed depending on the intended use of both the desalinated product water, and the wastewater. Acknowledgments This research was supported by funding through the King Abdullah University of Science and Technology (KAUST) (Award KUS-I1-003-13). References [1] S. Lattemann, M.D. Kennedy, J.C. Schippers, G. Amy, Global desalination situation, in: I.C. Escobar, A.I. Schafer (Eds.), Sustainable Water for the Future: Water Recycling Versus Desalination, Elsevier, Amsterdam, 2010. [2] L.F. Greenlee, D.F. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis desalination: water sources, technology, and today's challenges, Water Res. 43 (2009) 2317–2348. [3] M. Elimelech, W.A. Phillip, The future of seawater desalination: energy, technology, and the environment, Science 333 (2011) 712–717. [4] C. Liu, K. Rainwater, L. Song, Energy analysis and efficiency assessment of reverse osmosis desalination process, Desalination 276 (2011) 352–358. [5] R. Semiat, Energy issues in desalination processes, Environ. Sci. Technol. 42 (2008) 8193–8202. [6] A. Subramani, M. Badruzzaman, J. Oppenheimer, J.G. Jacangelo, Energy minimization strategies and renewable energy utilization for desalination: a review, Water Res. 45 (2011) 1907–1920. [7] I. Shizas, D.M. Bagley, Experimental determination of energy content of unknown organics in municipal wastewater streams, J. Energy Eng. 130 (2004) 45–53. [8] E.S. Heidrich, T.P. Curtis, J. Dolfing, Determination of the internal chemical energy of wastewater, Environ. Sci. Technol. 45 (2011) 827–832. [9] G. Tchobanoglous, F.L. Burton, H.D. Stensel, Wastewater Engineering, Treatment and Reuse, Metcalf & Eddy, Inc. 4th edition McGraw-Hill, Boston, 2003. [10] X. Cao, X. Huang, P. Liang, K. Xiao, Y. Zhou, X. Zhang, B.E. Logan, A new method for water desalination using microbial desalination cells, Environ. Sci. Technol. 43 (2009) 7148–7152. [11] X. Chen, X. Xia, P. Liang, X. Cao, H. Sun, X. Huang, Stacked microbial desalination cells to enhance water desalination efficiency, Environ. Sci. Technol. 45 (2011) 2465–2470. [12] H. Strathmann, Ion-exchange Membrane Separation Processes, Elsevier B. V., Amsterdam, 2004. [13] H. Luo, P.E. Jenkins, Z. Ren, Concurrent desalination and hydrogen generation using microbial electrolysis and desalination cells, Environ. Sci. Technol. 45 (2011) 340–344. [14] M. Mehanna, P.D. Kiely, D.F. Call, B.E. Logan, Microbial electrodialysis cell for simultaneous water desalination and hydrogen gas production, Environ. Sci. Technol. 44 (2010) 9578–9583. [15] M. Mehanna, T. Saito, J. Yan, M. Hickner, X. Cao, X. Huang, B.E. Logan, Using microbial desalination cells to reduce water salinity prior to reverse osmosis, Energy Environ. Sci. 3 (2010) 1114–1120. [16] Y. Qu, Y. Feng, X. Wang, J. Liu, J. Lv, W. He, B.E. Logan, Simultaneous water desalination and electricity generation in a microbial desalination cell with electrolyte recirculation for pH control, Bioresour. Technol. 106 (2012) 89–94. [17] H. Luo, P. Xu, T.M. Roane, P.E. Jenkins, Z. Ren, Microbial desalination cells for improved performance in wastewater treatment, electricity production, and desalination, Bioresour. Technol. 105 (2012) 60–66. [18] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, 2nd ed. John Wiley & Sons, New York, 2001. [19] Y. Kim, B. Logan, Series assembly of microbial desalination cells containing stacked electrodialysis cells for partial or complete seawater desalination, Environ. Sci. Technol. 45 (2011) 5840–5845. [20] K.S. Jacobson, D.M. Drew, Z. He, Use of a liter-scale microbial desalination cell as a platform to study bioelectrochemical desalination with salt solution or artificial seawater, Environ. Sci. Technol. 45 (2011) 4652–4657. [21] K.S. Jacobson, D.M. Drew, Z. He, Efficient salt removal in a continuously operated upflow microbial desalination cell with an air cathode, Bioresour. Technol. 102 (2011) 376–380. [22] N. Lakshminarayanaiah, Electroosmosis in ion-exchange membranes, J. Electrochem. Soc. 116 (1969) 338–343. [23] C.I. Torres, R. Krajmalnik-Brown, P. Parameswaran, A.K. Marcus, G. Wanger, Y.A. Gorby, B.E. Rittmann, Selecting anode-respiring bacteria based on anode potential: phylogenetic, electrochemical, and microscopic characterization, Environ. Sci. Technol. 43 (2009) 9519–9524. [24] P. Aelterman, S. Freguia, J. Keller, W. Verstraete, K. Rabaey, The anode potential regulates bacterial activity in microbial fuel cells, Appl. Microbiol. Biotechnol. 78 (2008) 409–418. [25] R.C. Wagner, D.F. Call, B.E. Logan, Optimal set anode potentials vary in bioelectrochemical systems, Environ. Sci. Technol. 44 (2010) 6036–6041. [26] V.J. Watson, B.E. Logan, Analysis of polarization methods for elimination of power overshoot in microbial fuel cells, Electrochem. Commun. 13 (2011) 54–56.

[27] Y. Hong, D.F. Call, C.M. Werner, B.E. Logan, Adaptation of high current using low external resistances eliminates power overshoot in microbial fuel cells, Biosens. Bioelectron. 28 (2011) 71–76. [28] L. Shaffer, M. Mintz, Electrodialysis, in: K.S. Spiegler, A.D.K. Laird (Eds.), Principles of Desalination, Academic Press, New York, 1980. [29] F.H. Meller, Electrodialysis and electrodialysis reversal, in: P. Murray (Ed.), Manual of Water Supply Practices — M38, American Water Works Association, Denver, 1995. [30] J. Veerman, J.W. Post, M. Saakes, S.J. Metz, G.J. Harmsen, Reducing power losses caused by ionic shortcut currents in reverse electrodialysis stacks by a validated model, J. Membr. Sci. 310 (2008) 418–430. [31] B.E. Logan, P. Aelterman, B. Hamelers, R. Rozendal, U. Schröder, J. Keller, S. Freguiac, W. Verstraete, K. Rabaey, Microbial fuel cells: methodology and technology, Environ. Sci. Technol. 40 (2006) 5181–5192. [32] D. Pant, G.V. Bogaert, L. Diels, K. Vanbroekhoven, A review of the substrates used in microbial fuel cells (MFCs) for sustainable energy production, Bioresour. Technol. 101 (2010) 1533–1543. [33] Z. He, Y. Huang, A.K. Manohar, F. Mansfeld, Effect of electrolyte pH on the rate of the anodic and cathodic reactions in an air-cathode microbial fuel cell, Bioelectrochemistry 74 (2008) 78–82. [34] S. Chen, G. Liu, R. Zhang, B. Qin, Y. Luo, Development of the microbial electrolysis desalination and chemical-production cell for desalination as well as acid and alkali productions, Environ. Sci. Technol. 46 (2012) 2467–2472. [35] J.-Y. Nam, B.E. Logan, Enhanced hydrogen generation using a highly saline catholyte in a two chamber microbial electrolysis cell, Int. J. Hydrogen Energy 36 (2011) 15105–15110. [36] H. Liu, S. Cheng, B.E. Logan, Power generation in fed-batch microbial fuel cells as a function of ionic strength, temperature, and reactor configuration, Environ. Sci. Technol. 39 (2005) 5488–5493. [37] S.-E. Oh, B.E. Logan, Proton exchange membrane and electrode surface areas as factors that affect power generation in microbial fuel cells, Appl. Microbiol. Biotechnol. 70 (2006) 162–169. [38] B.E. Logan, Microbial Fuel Cells, John Wiley & Sons, Inc., Hoboken, NJ, 2008. [39] J.-Y. Nam, J.C. Tokash, B.E. Logan, Comparison of microbial electrolysis cells operated with added voltage or by setting the anode potential, Int. J. Hydrogen Energy 36 (2011) 10550–10556. [40] T.Y. Cath, S. Gormly, E.G. Beaudry, B.T. Flynn, V.D. Adams, A.E. Childress, Membrane contactor processes for wastewater reclamation in space Part I. Direct osmotic concentration as pretreatment for reverse osmosis, J. Membr. Sci. 257 (2005) 85–98. [41] T.Y. Cath, D. Adams, A.E. Childress, Membrane contactor processes for wastewater reclamation in space II. Combined direct osmosis, osmotic distillation, and membrane distillation for treatment of metabolic wastewater, J. Membr. Sci. 257 (2005) 111–119. [42] B. Zhang, Z. He, Integrated salinity reduction and water recovery in an osmotic microbial desalination cell, RSC Adv. 2 (2012) 3265–3269. [43] F. Zhang, K.S. Brastad, Z. He, Integrating forward osmosis into microbial fuel cells for wastewater treatment, water extraction and bioelectricity generation, Environ. Sci. Technol. 45 (2011) 6690–6696. [44] Z. Ge, Z. He, Effects of draw solutions and membrane conditions on electricity generation and water flux in osmotic microbial fuel cells, Bioresour. Technol. 109 (2012) 70–76. [45] A. Achilli, T.Y. Cath, E.A. Marchand, A.E. Childress, The forward osmosis membrane bioreactor: a low fouling alternative to MBR processes, Desalination 239 (2009) 10–21. [46] H. Strathmann, Ion-exchange membrane processes in water treatment, in: I.C. Escobar, A.I. Schafer (Eds.), Sustainable Water for the Future: Water Recycling Versus Desalination, Elsevier, Amsterdam, 2010. [47] H. Strathmann, Electrodialysis, a mature technology with a multitude of new applications, Desalination 264 (2010) 268–288. [48] Astom, Neosepta bipolar specification, http://www.astom-corp.jp/en/en-main2bipolar.html, February, 2012. [49] Fumatech, Fumasep–bipolar membrane specification, http://www.fumatech. com/EN/Products/fumasep/Bipolar-membranes/February, 2012. [50] Y. Kim, B.E. Logan, Hydrogen production from inexhaustible supplies of fresh and salt water using microbial reverse-electrodialysis electrolysis cells, Proc. Natl. Acad. Sci. U. S. A. 108 (2011) 16176–16181. [51] K. Rabaey, S. Butzer, S. Brown, J. Keller, R.A. Rozendal, High current generation coupled to caustic production using a lamellar bioelectrochemical system, Environ. Sci. Technol. 44 (2010) 4315–4321. [52] A.P. Borole, H. O'Neill, C. Tsouris, S. Cesar, A microbial fuel cell operating at low pH using the acidophile Acidiphilium cryptum, Biotechnol. Lett. 30 (2008) 1367–1372. [53] S. Cheng, B.E. Logan, High hydrogen production rate of microbial electrolysis cell (MEC) with reduced electrode spacing, Bioresour. Technol. 102 (2011) 3571–3574. [54] F. Zhang, D. Pant, B.E. Logan, Long-term performance of activated carbon air cathodes with different diffusion layer porosities in microbial fuel cells, Biosens. Bioelectron. 30 (2011) 49–55. [55] M.T. Madigan, J.M. Martinko, Brook Biology of Microorganisms, 11th ed. Pearson Education, Upper Saddle River, 2006. [56] J.R. Kim, S. Cheng, S.-E. Oh, B.E. Logan, Power generation using different cation, anion and ultrafiltration membranes in microbial fuel cells, Environ. Sci. Technol. 41 (2007) 1004–1009.