Batch Operation of a Microbial Fuel Cell Equipped with Alternative Proton Exchange Membrane

international journal of hydrogen energy 40 (2015) 17323e17331

Available online at www.sciencedirect.com ScienceDirect

journal homepage: www.elsevier.com/locate/he

Batch operation of a microbial fuel cell equipped with alternative proton exchange membrane

* G. Hernandez-Flores a, H.M. Poggi-Varaldo a, , O. Solorza-Feria b, T. Romero-Castan~on c,E.Rıós-Leal d, J. Galıńdez-Mayer e, F. Esparza-Garcıá d a Environmental Biotechnology and Renewable Energies R&D Group, Dept. of Biotechnology and Bioengineering, Centro de Investigacion y de Estudios Avanzados del Instituto Politecnico Nacional, Av. Instituto Politecnico Nacional 2508, Col. San Pedro Zacatenco, Delegacion Gustavo A. Madero, Mexico D.F., C. P. 07360 Apartado Postal: 14-740, 07000 Mexico, D.F, Mexico b Dept. of Chemistry, Centro de Investigacion y de Estudios Avanzados del Instituto Politecnico Nacional, Av. Instituto Politecnico Nacional 2508, Col. San Pedro Zacatenco, Delegacion Gustavo A. Madero, Mexico D.F., C. P. 07360 Apartado Postal: 14-740, 07000 Mexico, D.F., Mexico c Electric Research Institute, Reforma 113, Col. Palmira, C. P. 62490, Cuernavaca, Morelos, Mexico d Dept. Biotechnology and Bioengineering, Centro de Investigacion y de Estudios Avanzados del Instituto Politecnico Nacional, Av. Instituto Politecnico Nacional 2508, Col. San Pedro Zacatenco, Delegacion Gustavo A. Madero, Mexico D.F., C. P. 07360 Apartado Postal: 14-740, 07000 Mexico, D.F., Mexico e ENCB del IPN, Division of Basic Sciences, Escuela Nacional de Ciencias Biologicas, ENCB, Unidad Profesional Lazaro Cardenas, Prolongacion de Carpio y Plan de Ayala s/n, Col. Santo Tomas, C. P. 11340, Delegacion Miguel Hidalgo, Mexico, D.F., Mexico article info abstract

Article history: We compared the effect of membrane type on the performance of microbial fuel cells (MFC) Received 30 December 2014 fed with an actual leachate and operated in batch for 15 days. The tested proton exchange Received in revised form membranes (PEMs) were Naﬁon 117 (NF) and a low cost membrane (LCM). The cell equipped 6 May 2015 with LCM outperformed the one equipped with NF. In the ﬁrst period of the batch, 0e8d, 3 Accepted 10 June 2015 average volumetric powers (PV) were 9000 and 4000 mW/m for the MFC equipped with LCM Available online 23 July 2015 and NF, respectively. In the second period (8e15 d) when the external resistances were 3 adjusted, the average PVs were 20 000 and 6800 mW/m for LCM and NF, respectively. At the Keywords: end of the batch, deposits of dry salts appeared on the external side of the cathode carbon Proton exchange membrane cloth of the cell equipped with NF. Likely, this could be related to the decrease of power Low cost membrane output in the last days of the batch (11e15 d) in the cell equipped with NF. Batch operation Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights Microbial fuel cell reserved. Leachate

* Corresponding author. CINVESTAV del IPN, Environmental Biotechnology and Renewable Energy Group, Dept. Biotechnology and Bioengineering, Apartado Postal 14-740, 07000, Mexico D.F., Mexico. Tel.: þ52 (55) 5747 3800x4321, þ52 (55) 5747 3800x4324. E-mail address: [email protected] (H.M. Poggi-Varaldo). http://dx.doi.org/10.1016/j.ijhydene.2015.06.057 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. 17324 international journal of hydrogen energy 40 (2015) 17323e17331

membrane-less MFC. In contrast, their MFC equipped with NF Introduction membrane displayed a lower power density of 28 mW/m2 but a higher CE of 28%. Nowadays the human-kind depend heavily on the use of pe- Regarding the use of new materials for PEMs that could troleum oil as well as other fossil energy sources, and conse- reduce costs there are few studies that have focused on glass quently faces at least two great risks: the inevitable depletion fibers or glass wool, salt bridge, as well as assemblies and the environmental pollution caused during exploration, membrane-cathode [14,15,22,23]. Yet, results so far indicate transport, combustion of fossil-based fuels. Thus, the inves- that either high Rint were found or low power was delivered by tigation to develop novel, renewable energy sources, particu- the cells or both. For instance, Li et al. indicated that some e larly bioenergies, has notably increased in the last years [1 3]. course-pore filters (i.e., glass fiber, nylon mesh, and porous Microbial fuel cells (MFCs) constitute a promising tech- fabrics) can be more economic than ionic exchange mem- nology that generates electrical power from a wide range of branes used in MFCs [15]. Yet, the first ones show the disad- soluble substrates (organic or inorganic), wastes included. The vantage of large pores that, in turn, would detract from the MFCs has become an interesting alternative to produce elec- effectiveness of the separator for decreasing the flux of oxy- trical energy and provide wastewater treatment simulta- gen to the anodic chamber [24]. Glass fiber mats show better e neously [4 8]. performance: they are resistant to biodegradation and In order to increase the MFC efficiency, several conditions biofouling, they exhibit a low O2 permeability, and relatively of MFC operation and components have been the subject of high CEs can be achieved (81%) [15,24]. Indeed, it has been intensive research such as the type of biocatalysts, membrane reported that glass wool can also serve as a low cost separator (electrolyte) or separators, temperature, pH, substrates, the in MFCs [25,26]. Yet, the power was low, with a maximum of type and materials of electrodes, electrode catalysts, cell 10.1 mW/m2. Besides the open circuit potential (OCP) was only e configuration and architecture, among others [2,8a 12]. 0.36 V with a resistance as high as 4000 U. The proton exchange membrane (PEM) is an important part In another work, Min et al. [22] determined the power of MFCs. The main features and purposes of the membranes in output in a MFC equipped with a salt bridge that replaced the e MFCs are listed below [13 15]: Nafion membrane; the inoculum was a pure culture of Geo- bacter metallireducens. The power density was very low, - to separate the anodic from the cathodic chamber in order 2 2.2 mW/m , whereas the Rint was nearly 20 kU. The authors to reduce the substrate flux from the anode to cathode, to ascribed the low power output to the higher Rint of the salt avoid the back-diffusion of the electron acceptor, and to bridge system. In contrast, the Rint of a similar device equipped isolate the catalyst from the cathode in single-chamber with Nafion membrane was 1286 U. Kargi & Eker worked with MFCs a two-chamber MFC separated by a salt-agar slab. A low power - to perform as a barrier to the transfer of other ions between density ca. 3mW/m2 and current intensity of 0.24 mA were the chambers obtained [23]. The authors claimed that these values were - to increase the Coulombic efficiency (CE) reducing the flux comparable with those reported in the literature for MFC uti- of the oxygen from the cathode chamber to the solution in lizing salt bridge, i.e., a work by Min et al. [22]. the anode chamber Therefore, the purpose of this work was to compare the - to ensure an efficient and sustainable operation along time effect of membrane type on the performance of MFC operated in long batch process, using actual leachates from Mexico City Yet, there are disadvantages related to the PEM use, such as sanitary landfill and inoculum previously enriched (E-In)in the high cost of standard membranes such as those made of electrochemically-active bacteria (EAB). The tested separators e the polymer Nafion [13 15]. For instance, Nafion cost has were a low cost membrane (LCM) and NF as reference. increased up to $1733/m2 [16]. Furthermore, its use might affect negatively the power generated by the MFC due to the increase of the internal resistance (Rint) [14,15,17]. Materials and methods Presently, one of the challenges of the MFCs is their scaling up; this mainly depends on the performance of MFC and cost Experimental design of materials [15,18]. So, in order to replace the Nafion 117 (NF), several polymeric membranes have been studied, such as The experiment consisted of the operation in a long batch ultrafiltration and microfiltration membranes, sulphonated process of the MFCs equipped either with a LCM or NF (as polyether ether ketone membrane, anion and cation exchange control). The MFCs were packed with graphite flakes (GF)as membranes, bipolar membrane, forward osmosis membrane anodic material and loaded with a mixture of inoculum pre- e [2,9,14,15,18 21]. However, these polymeric membranes can viously enriched in EAB and a very recalcitrant, actual also be expensive. leachate from Mexico City sanitary landfill. The mix was in a Membrane-less MFCs have also been tested, where the proportion 80% inoculum and 20% actual leachate. The MFCs water conducts the protons by itself. However, in most of the were operated for 15 d. works operated without a membrane, the CE is low or falls The long batch process was divided in two periods, a first e & down to unfeasible values [9,13 15,19]. Liu Logan explored one from 0 to 8 d, and a second one from 8 to 15 d. In the first the bioelectricity generation in a membrane-less MFC, in order period the cells were fitted with external resistances (Rext) to increase the energy output and reduce the cost [13]. They defined by the first electrochemical characterization of the Rint 2 reported a power density of 146 mW/m and 20% of CE for their at time 0 d. After a second electrochemical characterization international journal of hydrogen energy 40 (2015) 17323e17331 17325

that was carried out at 8 d, then, the Rext value was readjusted stainless steel 1 mm thickness. In the liquid side, the cathode and the second period started. was in contact with the PEM (NF or LCM) [16,28]. The main response variables determined in the electro- The NF was pretreated to activate and to remove impurities chemical characterization of our MFCs were the maximum before to use in the MFC. We describe a modified technique & volumetric power (PV,max) and the Rint of the MFCs. On the from Oh Logan [29]. The membrane was soaked first in H2O2 other hand, in the batch operation of the MFCs the response (3% v/v), followed by soaking in deionized water. Afterwards, variables were the average PV (with respect to time), the the membrane was immersed in 2 M H2SO4, followed by a last chemical oxygen demand (COD) removal (hCOD), the CE and the soaking in deionized water. Each stage lasted 1 h and was current intenstity (IMFC). performed at 80 C. The experiments were carried out at ambient temperature, The LCM was fabricated based on easily accessible, low cost with no mechanical mixing or heating, in a single compart- agar. A solution of agar (agar/agar, from red algae Gelidium ment, air-cathode MFCs. genus; purchased from Labcitec S.A. de C.V., Mexico City, Mexico) at 2% w/v was made by dissolving 0.64 g of agar in 32 mL of warm distilled water. Afterwards, while still warm, Microbial fuel cell the solution was poured in a Petri dish of 8.37 cm diameter. The Petri dish was placed in an oven at 70 C for 5 h. With this The MFCs consisted of horizontal cylinders built in Plexiglas treatment, the membrane achieved a dehydration extent of 80 mm long and 57 mm internal diameter. The anodic 97% [16]. Typical thickness of the dry membrane was chambers were packed with GF with a surface area of 98 ± 13 mm. Thickness was measured with an ultrasonic 0.28 m2. The surface area was obtained as follows: a large thickness gauge Minitest brand model 2100 from the company sample of material was screened and the fraction between ElektroPhysik. Finally, the dry membrane was painted with meshes 10 and 6 (diameters 2 mm and 3.55 mm, respec- 0.5 mg/cm2 platinum catalyst (Pt 10 wt%/C-ETEK) [28]. tively) was collected. From this fraction, five 20 g subsamples were taken and weighed. Afterwards, the number of parti- Enrichment of inocula cles in each subsample were determined and recorded. The average number of particles was estimated. Based on the An enrichment procedure based on selective pressure using latter, the average weight of particle of each material was Fe (III) as an electron acceptor and sodium acetate as carbon estimated. Afterwards, using the Eq. (1) shown below, it was source was implemented [16,30,31]. The departing inoculum possible to calculate the surface area of the mass of material for the enrichment procedure consisted of soil sampled from loaded into the MFC. The shape factor of the material (also an excavation made in the garden of Centro de Investigacion y called sphericity factor in other textbooks) was taken into de Estudios Avanzados del Instituto Politecnico Nacional account as described in Perry [27]. For instance, 0.43 was (193003300N, 990704600W) at a depth of 20 cm [32]. Afterwards, chosen for GF. gravel and vegetal matter were manually removed from the On the other hand, the net volume of the anodic chamber soil, and the soil was further screened in a screen mesh size in our MFCs was calculated as the geometric volume of the 14. chamber minus the physical volume of the anodic material. This conditioned soil was used as inoculum for the With the surface area of the anodic material and the net vol- enrichment procedure with FeO$OH, as follows [33]: a sample ume, the specific surface area of the anode ðA0 Þ was finally s of screened soil was transferred to an anaerobic bottle, after calculated with Eq. (1) below this, 5 g of soil sample was suspended in anaerobic saline ! 1=3 solution (50 mL); afterwards, 5 mL of sample was transferred 1=3 62 m2p3 M p M p 3 36 to 50 mL metal-reduction medium with acetate as electron Fs m p2r2 2 p Fs mpr 0 A s ¼ ¼ (1) donor and Fe(III) oxide-hydroxide as electron acceptor. The V M Vcell r enrichment of inocula was obtained with serial transfers [30,32,33]. where Duplicate enrichments were incubated at 30 C for 9 d in the dark condition. The enrichment procedure was repeated 3 F s shape factor of the particle defined as the quotient of the times. The culture medium consisted of (g/L): 2.5 NaHCO3, 0.25 $ area of a sphere equivalent to the volume of the particle NH4Cl, 0.6 NaH2PO4 H2O, 0.1 KCl, 10 mL vitamin solution and divided by the actual surface of the particle 10 mL mineral solution [34,35]. The Fe(III) oxide was synthe- $ mp average weight of a particle of the given size fraction sized as follows: a solution 0.4 M of FeCl3 6H2O (pH adjusted M total mass of anodic material loaded into the MFC to 7.0 with 10 M of NaOH) was added [35]. r actual density of the material At the end of the procedure, after several serial transfers,

Vcell geometric volume of the cell chamber the inert matter of soil was significantly diluted or lost; at this time just the biomass was retained. This biomass was likely The net volume of the MFC necessary for the denominator enriched in electrochemically active bacteria [36]. 0 in the calculation of A s was estimated as described above in the denominator of Eq. (1). Leachate The cathode of the MFCs was a flexible carbon-cloth con- taining 0.5 mg/cm2 platinum catalyst (Pt 10 wt%/C-ETEK). On The leachates used as substrate were sampled from the the air side, the cathode was limited by a perforated plate of Mexico City sanitary landfill “Bordo Poniente”. Two types of 17326 international journal of hydrogen energy 40 (2015) 17323e17331

leachates highly recalcitrant were provided: samples from The initial COD and biomass concentration in the MFC li-

Section (L-1) and samples from Section (L-4), where the quor were ca. 2000 mg O2/L and 1900 volatile suspended solids/ denomination ‘section’ is related to the chronological con- L (VSS/L), respectively. Biomass concentration was deter- struction of the landfill cells. Organic matter contents of mined in terms of VSS, according to the common practice in leachates were 4300 and 12 300 mg COD/L for L-1 and L-4, wastewater treatment and analysis, see Tchobanoglous et al. respectively. The MFC mas loaded with the sample from Sec- [40] and APHA [41]. The pH and the electrical conductivity tion (L-4) and the pH was slightly alkaline, ca. 8.0 (Table 1). were 9 and 39 mS/cm, respectively. The relatively high organic matter content and high value of BOD5/COD ratio indicated that the leachate is biodegradable Analyses and not quite aged [37,38]. Interestingly, we expected a lower pH consistent with fresh The COD, VSS, pH, BOD , total Kjeldahl nitrogen (TKN), sulfate leachate. That was not the case. It is known that the “Bordo 5 anion, electrolytic conductivity were determined according to Poniente” landfill is emplaced in a site characterized by sodic- the Standard Methods and others [41e43]. saline soil with pH of soil extract as high as 11. The local soil was likely used to cap the landfill cells during the daily operation of the landfill, possibly releasing sodium salts (carbon- ate, bicarbonate) as well as hydroxide that increased leachate Results and discussion pH. This explanation is supported by the high values of the electrolytic conductivity of the leachate (Table 1). Electrochemical characterization

Electrochemical characterization of the microbial fuel cells The MFCs equipped with their respective PEMs were electrochemically characterized at 0 d. The procedure lead to deﬁne

The Rint and power density curve of the MFC were determined the Rext to start the power generation; a value of Rext close to ' by duplicate, using the polarization curve method by varying that of Rint was chosen based on Jacobi s Theorem for elec- the external resistance and recording both the voltage and the tromotive forces [8,44]. The MFC equipped with LCM showed current intensity [14,17,28,39]. The MFCs were operated at discouraging values (Fig. 1, Table 2). 3 open circuit for 1 h; afterwards different resistors were varied, The Rint and Pv,max were 650 U and 9 mW/m , whereas the U 10 to 1 M and viceversa, to determine the power generation an MFC equipped with NF displayed a Rint and Pv,max of 350 U and another response variables as a function of load. After this, 1100 mW/m3, respectively (Fig. 2, Table 2). In this first test, the the cell was set to open circuit conditions for 1 h in order to cell fitted NF performed better than that with LCM. check the adequacy of the procedure (values of initial and final open circuit voltages should be close). The voltage was measured and recorded with a Multimeter ESCORT 3146A. The current was calculated by the Ohm's law (Eq. (2)) and the Rint was calculated as the slope of the linear section of the curve voltage versus the current intensity [19,32]. The PV was calculated according the Eq. (3) below.

EMFC IMFC ¼ (2) Rext

2 ðEMFCÞ PV ¼ (3) Vcell Rext where IMFC is the current intensity of the MFC in A, EMFC is the voltage delivered by the cell in V, Rext is the external resistance U connected to the cell in and Vcell is the net volume of the anodic chamber.

Table 1 e Characteristics of municipal leachate. Parameters Value pH 8.26 ± 0.02 Electrical conductivity (mS/cm) 36.7 ± 0.1 Total Kjeldahl nitrogen (g/L) 2.9 ± 0.03 2 SO4 (g/L) 0.281 ± 0.01 COD (g/L)a 12.3 ± 0.5 b ± BOD5 (g/L) 10.6 0.2

BOD5/COD 0.86 a Chemical oxygen demand. Fig. 1 e Electrochemical characterization of the MFC b Biochemical oxygen demand. equipped with LCM at (a) 0 d and (b) 8 d. international journal of hydrogen energy 40 (2015) 17323e17331 17327

Table 2 e Electrochemical characterization of microbial fuel cells equipped with different membranes. Parameter Time (d) 0d 8d LCMa NFb LCM NF U ± ± ± ± Rint ( ) 649.3 21.8 350.0 218.1 40.8 6.7 79.7 1.3 3 c Pv,max (mW/m ) 9.31 ± 3.2 1142.7 ± 379.2 22560.0 ± 2727.0 8594.9 ± 1069.8 d IMFC (mA) 0.026 ± 0.004 0.49 ± 0.08 8.59 ± 0.05 2.75 ± 0.17 2 e ± ± ± ± Pcath (mW/m ) 0.26 0.09 32.34 10.73 638.40 77.20 243.22 30.30 f ± ± ± ± EMFC,max (mV) 25.6 4.3 163.0 27.7 189.0 11.5 225.2 14.02 g EMFC,OC (mV) 29.4 ± 19.4 380.0 ± 29.5 619.4 ± 47.7 549.7 ± 28.3 Notes. a Low cost membrane. b Naﬁon 117 membrane. c Maximum volumetric power. d Current intensity value at the maximum power. e Maximum power density based on surface area of electrode (cathode). f Potential value at the maximum power. g Open circuit potential.

However, after 8 d, in the second electrochemical charac- So, characteristics of the MFC equipped with LCM were supe- terization the MFC equipped with LCM improved its perfor- rior to that of the MFC equipped with NF, Table 2. 3 mance, reaching a Pv,max of 22 500 mW/m (Fig. 1, Table 2). The

Rint decreased 94% (40 U) and the Pv,max increased by 4 orders of magnitude. Performance of the MFC equipped with low cost membrane In the second characterization of the MFC equipped with and Nafion membrane in long batch operation U 3 NF, the Rint and Pv,max observed were 80 and 8600 mW/m e (Table 2, Fig. 2). The Rint only decreased 77%, whereas the Pv,max In the first period, 0 8 d, the load resistances used were 470 increased 87% with respect to the corresponding initial values. and 680 U for the MFCs equipped with NF and LCM, respectively (defined by previous electrochemical characterizations).

The average volumetric powers (PVs) recorded were 9000 and 4000 mW/m3 for MFC equipped with LCM and NF, respectively (Fig. 3, Table 3). The MFC equipped with LCM reached an

average PV higher than MFC equipped with NF, although the values in the ﬁrst characterization. The second electrochemical characterization was carried out at day 8 in order to check for possible variations of

the Rints. In the second period, 8e15 d, new load resistances were applied, 47 and 82 U in the systems using LCM and NF, respectively. In this period the MFC equipped 3 with LCM reached and average PV of 20 000 mW/m ,a

value very close to the maximum PV,max recorded in the second electrochemical characterization (Tables 2 and 3).

Furthermore, the PV was stable throughout this period (Fig. 3). Indeed, the CE of the MFC equipped with LCM was more than double of that of the Naﬁon-equipped MFC (Table 3). On the other hand, the MFC equipped with NF, reached 3 only an average PV of 6800 mW/m during the days 8e11. Af-

terwards, the PV began to fall down (Fig. 3). This pattern could be explained because of deposits of dry salts appeared on the external side of the cathode carbon cloth of the MFC equipped with NF, starting at day 11 until the end of the batch (Fig. 4). The inﬂuent fed to the MFCs had a high salinity and high pH (Table 1) and it seems that the NF was affected by these parameters. However, this effect was not observed for the MFC equipped with LCM. It seems that the salt deposits could be responsible for the Fig. 2 e Electrochemical characterization of the MFC decrease of power output during 11e15 d in the cell equipped equipped with NF at (a) 0 d and (b) 8 d. with NF. 17328 international journal of hydrogen energy 40 (2015) 17323e17331

Fig. 4 e Deposits of dry salts on the external side of the cathode carbon cloth of the MFC equipped with NF from the day 11th.

Regarding the organic matter of the effluent, both MFCs had enough “fuel” supply to convert into electrical energy, so, no refeed of leachate was performed during the batch. The Fig. 3 e Time course of voltage and volumetric power h at the end of the batch operation were 39.32 and 28.29% outputs of MFCs using (a) NF and (b) LCM. COD for the MFCs equipped with LCM and NF, respectively (Table 3). Once more, the MFC equipped with the LCM, exhibited a h higher COD than the device with NF. Interestingly, the influent color at the end of the batch operation was significantly

Table 3 e Performance of the microbial fuel cells equipped with low cost and Naﬁon membranes. Parameter First period (0e8 d) Second period (8e15 d) LCMa NFb LCM NF

c Rext (U) 680 470 47 82 P (mW)d 0.65 ± 0.01 0.30 ± 0.01 1.40 ± 0.04 0.50 ± 0.01 3 e ± ± ± ± Pv (mW/m ) 8954 139 4104 80 19343 570 6876 162 f ± ± ± ± IMFC (mA) 0.98 0.01 0.79 0.01 5.45 0.08 2.46 0.03 2 g Pcath (mW/m ) 253.4 ± 3.9 116.1 ± 2.3 547.4 ± 16.1 194.6 ± 4.6 h EMFC (mV) 663.0 ± 5.2 373.2 ± 3.6 256.2 ± 3.8 201.8 ± 2.4 ɳ i k COD(%) NA NA 39.32 28.29 ɳ j Coul (%) NA NA 71.72 32.67 Cost (US$/m2) 14 1733 14 1733 Ratio Power to-Cost (mW/US$) 18.10 0.07 39.10 0.11

Notes. a Low cost membrane. b Nafion 117 membrane. c External resistance. d Power. e Volumetric power. f Current intensity. g Power density based on surface area of electrode (cathode). h Potential value. i Chemical oxygen demand removal efficiency. j Coulombic efficiency. k Not applicable. international journal of hydrogen energy 40 (2015) 17323e17331 17329

reduced and cell liquors became much clearer in both systems The effect of NF versus agar membrane in our case was (Fig. 5) at the end of the batch. Overall, the performance of the associated to a significantly low performance of the first MFC equipped with LCM was much better than MFC equipped compared to the second one. This was in contrast with the with NF. experiences in the literature: Min et al. [22] observed a From Table A1 in the Appendix A, Supplementary Material, dramatically higher power density in their cell equipped with it is apparent that our MFC equipped with an agar membrane NF membrane than that of the one fitted with salt bridge (4000 outcompeted results of previous works. Indeed, our average vs 2 mW/m2, Table A1), whereas Patil et al. [45] reported a 65% power density achieved in the batch operation of actual increase in power density using NF membrane versus salt leachate was ca. 550 mW/m2, significantly higher than the bridge (Table A1). As we have mentioned above, the reason for range of reported values (2.2e145 mW/m2) that were observed lower power densities of NF-equipped MFC in our case could with a cells that used agar-like separators and a variety of be associated to the presence of precipitates on the external influents (from synthetic ones with sodium acetate and face of the cathode (salt crystals) that started to accumulate at glucose to dilute industrial effluents such as chocolate in- the 11th day of operation (Fig. 4). Furthermore, the fall of dustry and molasses) [22,23,45]. Our high value of power power in our cell equipped with the expensive NF proton ex- density was accompanied by a low Rint of 51 U, 400-fold lower change membrane could be related to the pH of our influent than the Rint reported by Min et al. for a two-chamber cell (ca. 9). It has been observed and discussed in the open litera- equipped with an agar salt-bridge [22]. Furthermore, the ture that NF could not be ideal for MFC because at neutral or average potential of our cells during the batch operation was higher pHs, cations are moving from the anode to cathode also in the high side of the range of potentials observed in rather than protons [46e51]. other works with agar-like separators (ca. 260 mV versus range Sivasankaran & Sangeetha [20] developed a sulphonated of 20e190 mV). So, the set of electrochemical variables in our polyether ether ketone (SPEEK) to use in a MFC instead of NF.

MFC equipped with agar membrane on the one hand were The PV,max produced by their system, using dairy wastewater internally consistent, on the other hand they indicated a and domestic wastewater as inﬂuent were 5700 and 3200 mW/ 3 better performance compared to previous similar works. m , respectively. Our average PV obtained during the batch Possibly, the fact that the agar membrane in our case had a operation was ca. 20 000 mW/m3. It seems that our membrane lower thickness than the agar slab of Kargi & Eker [23] and the LCM could be considered a promising alternative for using as a salt bridges used by Min et al. [22] and Patil et al. [45], could separator in a MFC system. have contributed to our low Rint value. Also, the use of an E-In It is worth highlighting that the cost of our LCM was $14/ that was acclimated to the use of Fe(III) as electron acceptor m2, compared to the unit cost of $1733/m2 of the Naﬁon (enriched in dissimilatory metal-reducing microorganisms) in membrane (Table 3). Thus the cost ratio NOM/NF was quite our cell, could also explain the better performance of our MFC low, ($14/m2)/($1733/m2) ~ 1/120e0.8%. Moreover, the ratio equipped with an agar membrane. It has been reported in the Power-to-Cost (Table 3) suggests that the values of this indi- open literature that there is a close relationship between the cator obtained for the LCM is outstandingly higher than the metal-reduction capability of microorganisms and their corresponding values for the cells equipped with NF, i.e., 39.1 exoelectrogenic capability [44]. However, Min et al. worked and 0.11 mW/US$, respectively. with a strain Geobacter metallirreducens which is a known exoelectrogenic microorganism [22]. Thus, if the inoculum were the determining factor, a higher power could have been also expected in their salt-bridge equipped cell [22]. Conclusion

Our research was focused on the treatment of leachate from an actual sanitary landﬁll using MFCs equipped with a novel LCM. In spite of the aggressivity of the leachate, the cells removed up to 30e40% of the organic matter of the inﬂuent. The LCM tested in this work is a new alternative to use as separator in MFCs. Along the batch operation, the MFC performance using LCM was more stable and delivered a volumetric power more than 2-fold the power of the MFC equipped with NF. An outstanding ratio of 39 mW/US$ was obtained for the MFC equipped with LCM, compared to a very low value 0.11 for the NF-equipped MFC. Moreover, our LCM does not need any pretreatment; in contrast, the NF membrane requires a pretreatment with hydrogen peroxide and sulfuric acid that in turn generates hazardous wastes, besides the increased costs of membrane fabrication and conditioning. Fig. 5 e Visual evidence of leachate depuration after 15 d of In this work, the combination of the MFC technology and batch operation in the MFC equipped with (a) LCM and (b) the development and application of a novel, low cost sepa- NF. rator, showed promising results for leachate treatment. It 17330 international journal of hydrogen energy 40 (2015) 17323e17331

constitutes an attractive contribution to cost-effective and a [9] Yang Y, Sun G, Xu M. Microbial fuel cells come of age. J Chem more sustainable remediation of aggressive, actual leachates. Technol Biotechnol 2010;86:625e32. [10] Kim Y, Shin SH, Chang IS, Moon SH. Characterization of uncharged and sulfonated porous poly(vinylidene fluoride) membranes and their performance in microbial fuel cells. J Acknowledgments Membr Sci 2014;463:205e14. [11] Belleville P, Strong PJ, Dare PH, Gapes DJ. Influence of The authors gratefully acknowledge the input and suggestions nitrogen limitation on performance of a microbial fuel cell. Water Sci Technol 2011;63:1752e7. of the Editors and the anonymous Reviewers of the Journal, [12] Zhou M, Chi M, Luo J, He H, Jin T. An overview of electrode that allowed to significantly improving the manuscript. The materials in microbial fuel cells. J Power Sources authors are also grateful to CINVESTAV-IPN and ICYTDF (now 2011;196:4427e35. SECITI-GDF), Mexico, for financial support to this research [13] Liu H, Logan BE. Electricity generation using an air-cathode (PICCO-10-27 and PICCO-10-28), and CONACYT for the Infra- single chamber microbial fuel cell in the presence and structure Project 188281. Giovanni Hernandez-Flores received absence of a proton exchange membrane. Environ Sci Technol 2004;38:4040e6. a graduate scholarship from CONACYT, Mexico. Also the au- [14] Logan BE. Microbial fuel cells. New Jersey, USA: John Wiley- thors thank Mr. Rafael Hernandez-Vera, and technicians of Interscience; 2007. the Environmental of Biotechnology and Renewable Energy [15] Li W-W, Guo-Ping S, Xian-Wei L, Han-qing Y. Recent R&D Group, CINVESTAV-IPN for their excellent technical help. advances in the separators for microbial fuel cells. Bioresour Technol 2011;102:244e52. [16] Hernandez-Flores G. Interim report. Sc D Thesis. Mexico, D.F: Appendix A. Supplementary data CINVESTAV-IPN; 2013. [17] Xia X, Tokash JC, Zhang F, Liang P, Huang X, Logan BE. Oxygen-reducing biocathodes operating with passive oxygen Supplementary data related to this article can be found at transfer in microbial fuel cells. Environ Sci Technol http://dx.doi.org/10.1016/j.ijhydene.2015.06.057. 2013;47(4):2085e91. [18] Wei J, L P, H Xia. Recent progress in electrodes for microbial fuel cells. Bioresour Technol 2011;102:9335e44. references [19] Logan BE, Hamelers B, Rozendal R, Schroder€ U, Keller J, Freguia S, et al. Microbial fuel cells: methodology and technology. Environ Sci Technol 2006;40:5181e92. [1] Cheng-Dar Y, Chung-Ming L, Liou EML. A transition toward a [20] Sivasankaran A, Sangeetha D. Development of MFC using sustainable energy future: feasibility assessment and sulphonated polyether ether ketone (SPEEK) membrane for development strategies of wind power in Taiwan. Energy electricity generation from waste water. Bioresour Technol e Policy 2001;29:951e63. 2011;102(24):11167 71. [2] Logan BE, Regan JM. Microbial challenges and harnessing the [21] Zhang F, Brastad KS, He Z. Integrating forward osmosis into metabolic activity of bacteria can provide energy for a variety microbial fuel cells for wastewater treatment, water of applications, once technical and cost obstacles are extraction and bioelectricity generation. Environ Sci Technol e overcome. Environ Sci Technol 2006:5172e80. 2011;45:6690 6. [3] Das D, Veziroglu TN. Hydrogen production by biological [22] Min B, Cheng S, Logan BE. Electricity generation using processes: a survey of literature. Int J Hydrogen Energy membranes and salt bridge microbial fuel cells. Water Res e 2001;26:13e28. 2005;39:1675 86. [4] Logan BE, Rabaey K. Conversion of wastes into bioelectricity [23] Kargi F, Eker S. Electricity generation with simultaneous and chemicals by using microbial electrochemical wastewater treatment by a microbial fuel cell (MFC) with Cu technologies. Science 2012;337:686e90. and Cu-Au electrodes. J Chem Technol Biotechnol e [5] Hou J, Liu Z, Yang S, Zhou Y. Three-dimensional 2007;82:658 62. macroporous anodes based on stainless steel fiber felt for [24] Zhang XY, Cheng S, Wang X, Huang X, Logan BE. Separator high-performance microbial fuel cells. J Power Sources characteristics for increasing performance of microbial fuel e 2014;258:204e9. cells. Environ Sci Technol 2009;43:8456 61. [6] Pant D, Bogaert GV, Diels L, Vanbroekhoven K. A review of [25] Ghangrekar MM, Shinde VB. Performance of membrane-less the substrates used in microbial fuel cells (MFCs) for microbial fuel cell treating wastewater and effect of sustainable energy production. Bioresour Technol electrode distance and area on electricity production. e 2010;101:1533e43. Bioresour Technol 2007;98:2879 85. [7] Modin O, Gustavsson DJ. Opportunities for microbial [26] Mohan SV, Raghavulu SV, Sarma PN. Biochemical evaluation electrochemistry in municipal wastewater treatment e an of bioelectricity production process from anaerobic overview. Water Sci Technol 2014;69(7):1359e72. wastewater treatment in a single chambered. Biosens e [8] Poggi-Varaldo HM, Munoz-Paez KM, Escamilla-Alvarado C, Bioelectron 2008;23:1326 32. Robledo-Narvaez PN, Ponce-Noyola MT, Calva-Calva G, et al. [27] Perry R. Chemical engineers handbook. 4th ed. New York: & Biohydrogen, biomethane and bioelectricity as crucial McGraw-Hill Co; 1963. pp 5.50 ff. components of biorefinery of organic wastes: a review. [28] Vazquez-Larios AL, Solorza-Feria O, Vazquez-Huerta G, Esparza- Waste Manage Res 2014;32:353e65. http://dx.doi.org/10.1177/ Garcia FJ, Rios-Leal E, Rinderknecht-Seijas N, et al. A new design 0734242X14529178. improves performance of a single chamber microbial fuel cell. J e a) Poggi-Varaldo HM, Carmona-Martıńez A, Vazquez- New Mater Electrochem Syst 2010;13:219 26. Larios AL, Solorza-Feria O. Effect of inoculum type on the [29] Oh SE, Logan BE. Proton exchange membrane and electrode performance of a microbial fuel cell fed with spent organic surface areas as factors that affect power generation in extracts from hydrogenogenic fermentation of organic solid microbial fuel cells. Appl Microbiol Biotechnol e wastes. J New Mater Electrochem Syst 2009;12:49e54. 2006;70:162 9. international journal of hydrogen energy 40 (2015) 17323e17331 17331

[30] Lovley DR, Phillips EJP. Organic matter mineralization with community in the anode chamber. Bioresour Technol reduction of ferric iron in anaerobic sediments. Appl Environ 2009;100:5132e9. Microbiol 1986;51:683e9. [46] Fornero JJ, Rosenbaum M, Angenent LT. Electric power [31] Galıńdez-Mayer J, Ramon-Gallegos J, Ruiz-Ordaz N, Juarez- generation from municipal, food, and animal wastewaters Ramı´rez C, Salmeron-Alcocer A, Poggi-Varaldo HM. Phenol using microbial fuel cells. Electroanalysis and 4-chlorophenol biodegradation by yeast Candida tropicalis 2010;22(7e8):832e43. in a fluidized bed reactor. Biochem Eng J 2008;38:147e57. [47] Fornero JJ, Rosenbaum M, Cotta MA, Angenent LT. Microbial [32] Vazquez-Larios AL, Solorza-Feria O, Barrera Cortes J, Rıós- fuel cell performance with a pressurized cathode chamber. Leal E, Ponce-Noyola MT, Esparza-Garcıá F, et al. In: Battelle Environ Sci Technol 2008;42(22):8578e84. 9th International Conference on Remediation of Chlorinated [48] Rozendal RA, Hamelers HVM, Buisman CJN. Effects of and Recalcitrant Compounds, May 19e22, 2014. Monterey, membrane cation transport on pH and microbial fuel cell CA, USA. Platform Session H8 “Managing recalcitrant performance. Environ Sci Technol 2006;40:5206e11. compounds in landfill leachate”. [49] Zhao F, Harnisch F, Schroder€ U, Scholz F, Bogdanoff P, [33] Vazquez-Larios AL, Solorza-Feria O, Poggi-Varaldo HM, Herrmann I. Challenges and constraints of using oxygen Gonzalez-Huerta RG, Ponce-Noyola MT, Rıós-Leal E, et al. cathodes in microbial fuel cells. Environ Sci Technol Bioelectricity production from municipal leachate in a 2006;40:5193e9. microbial fuel cell: effect of two cathodic catalysts. Int J [50] Hernandez-Flores G, Poggi-Varaldo HM, Romero-Castan~on T, Hydrogen Energy 2014;39:16667e75. Solorza-Feria O. Harvesting energy from leachates in [34] Lovley DR, Phillips EJP. Rapid assay for microbially reducible microbial fuel cells using an anion exchange membrane. ferric iron in aquatic sedimens. Appl Environ Microbiol Article submitted to the 15th Int Congr Hydrogen Energy. 1987;53:1536e40. [51] Hernandez-Flores G, Poggi-Varaldo HM, Solorza-Feria O, [35] Lovley DR, Phillips EJP. Novel mode of microbial energy Ponce Noyola MT, Romero-Castan~on T, Rinderknecht- metabolism: organic carbon oxidation coupled to Seijas N. Tafel equation based model for the performance of dissimilatory reduction of iron or manganese. Appl Environ a microbial fuel cell. Int. J. Hydrogen Energy 2015. Microbiol 1988;54:1472e80. [36] Bond DR, Lovley DR. Electricity production by Geobacter Notation sulfurreducens attached to electrodes. Appl Environ Microbiol 2003;69(3):1548e55. A0 : relationship between the anode surface area to cell volume, [37] Singh SK, Tang WZ. Statistical analysis of optimum Fenton s also known as specific surface area of the anode oxidation conditions for landfill leachate treatment. Waste CE: coulombic efficiency Manage 2013;33(1):81e8. COD: chemical oxygen demand [38] Poggi-Varaldo HM, Rinderknecht-Seijas N. A differential EAB: electrochemically-active bacteria availability enhancement factor for the evaluation of E : voltage delivered by the cell pollutant availability in soil treatments. Acta Biotechnol (Eng MFC E : cell potential at which the maximum volumetric power Life Sci Ed John Wiley) 2003;23:271e80. MFC,max is registered [39] Damiano L, Jambeck JR, Ringelberg DB. Municipal solid waste E : cell potential value at open circuit potential landfill leachate treatment and electricity production using MFC,OC GF: graphite flakes microbial fuel cells. Appl Biochem Biotechnol I : current intensity of the MFC 2014;173:472e85. MFC LCM: low cost membrane [40] Tchobanoglous G, Burton FL, Stensel HD. Wastewater M: total mass of anodic material loaded into the MFC engineering-treatment and reuse. 4th ed. McGraw-Hill; 2003. MFC: microbial fuel cell [41] APHA. Standard methods for examination of water and m : average weight of a particle of the given size fraction wastewater. 17th ed. Washington DC: APHA-AWWA-WEF; p NF: Nafion 117 1989. P : power density based on surface area of electrode (cathode) [42] Poggi-Varaldo HM, Alzate-Gaviria LM, Perez-Hernandez A, cath PEM: proton exchange membrane Nevarez-Morillon VG, Rinderknecht-Seijas N. A side-by-side P : volumetric power comparison of two systems of sequencing coupled reactors V P : maximum volumetric power for anaerobic digestion of the organic fraction of municipal v,max R : external resistance solid waste. Waste Manage Res 2005;23(3):270e80. ext R : internal resistance [43] Valdez-Vazquez I, Poggi-Varaldo HM. Alkalinity and high int SPEEK: sulphonated polyether ether ketone total solids affecting H2 production from organic solid waste V : geometric volume of the cell chamber by anaerobic consortia. Int J Hydrogen Energy cell 2009;34:3639e46. [44] Poggi-Varaldo HM, Vazquez-Larios A, Solorza-Feria O. Greek characters Microbial fuel cells. In: Rodrı´guez-Varela FJ, Solorza-Feria O, h Hernandez-Pacheco E, editors. Fuel cells. Montreal, Canada: COD: chemical oxygen demand removal Book Livres; 2010. p. 124e61. r: actual density of the material F [45] Patil SA, Surakasi VP, Koul S, Ijmulwar S, Vivek A, s: shape factor of the particle defined as the quotient of the area Shouche YS, et al. Electricity generation using chocolate of a sphere equivalent to the volume of the particle divided by industry wastewater and its treatment in activated sludge the actual surface of the particle based microbial fuel cell and analysis of developed microbial