MURDOCH RESEARCH REPOSITORY This is the author s final version of the work, as accepted for publication following peer review but without the publisher s layout or pagination. The definitive version is available at http://dx.doi.org/10.1016/j.biortech.2013.05.108 Cheng, K.Y., Kaksonen, A.H. and Cord-Ruwisch, R. (2013) Ammonia recycling enables sustainable operation of bioelectrochemical systems. Bioresource Technology, 143. pp. 25-31. http://researchrepository.murdoch.edu.au/16172/ Copyright: 2013 Elsevier Ltd. It is posted here for your personal use. No further distribution is permitted.
Accepted Manuscript Ammonia recycling enables sustainable operation of bioelectrochemical systems Ka Yu Cheng, Anna H. Kaksonen, Ralf Cord-Ruwisch PII: S0960-8524(13)00876-6 DOI: http://dx.doi.org/10.1016/j.biortech.2013.05.108 Reference: BITE 11895 To appear in: Bioresource Technology Received Date: 28 March 2013 Revised Date: 22 May 2013 Accepted Date: 25 May 2013 Please cite this article as: Cheng, K.Y., Kaksonen, A.H., Cord-Ruwisch, R., Ammonia recycling enables sustainable operation of bioelectrochemical systems, Bioresource Technology (2013), doi: http://dx.doi.org/10.1016/j.biortech. 2013.05.108 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Bioresource Technology Ammonia recycling enables sustainable operation of bioelectrochemical systems Ka Yu Cheng 1*, Anna H. Kaksonen 1, Ralf Cord-Ruwisch 2 1 CSIRO Land and Water, Floreat, WA 6014, Australia 2 School of Biological Science and Biotechnology, Murdoch University, WA 6150, Australia * Correspondence: +61(8) 9333 6158; fax: +61(8) 9333 6211; email: kayu.cheng@csiro.au 1
Abstract Ammonium (NH + 4 ) migration across a cation exchange membrane is commonly observed during the operation of bioelectrochemical systems (BES). This often leads to anolyte acidification (< ph 5.5) and complete inactivation of biofilm electroactivity. Without using conventional ph controls (dosage of alkali or ph buffers), the present study revealed that anodic biofilm activity (current) could be sustained if recycling of ammonia (NH 3 ) was implemented. A simple gas-exchange apparatus was designed to enable continuous recycling of NH 3 (released from the catholyte at ph > 10) from the cathodic headspace to the acidified anolyte. Results indicated that current (110 ma or 688 A m -3 net anodic chamber volume) was sustained as long as the NH 3 recycling path was enabled, facilitating continuous anolyte neutralization with the recycled NH 3. Since the microbial current enabled NH + 4 migration against a strong concentration gradient (~10 fold), a novel way of ammonia recovery from wastewaters could be envisaged. Keywords: microbial fuel cells, microbial electrolysis cells, proton gradient, ph split, cation exchange membrane 2
1 Introduction Bioelectrochemical systems (BES) have shown promise for the conversion of organic compounds in wastewaters into valuables (e.g. electricity, fuel gases, chemicals, etc) (Logan et al. 2008; Rabaey and Verstraete 2005; Rittmann 2008). Arguably, the microbial catalysed anodic reaction is the most critical reaction in a BES process as it transforms the chemical energy captured in the organics directly into electrical energy. Typically, this reaction not only produces electrons, but also liberates protons (H + ). For instance, anodic oxidation of one mole acetate liberates nine mole H + and eight moles electrons according to the following equation (Schroder 2007): CH 3 COO - +4H 2 O 2HCO - 3 + 9H + + 8e - Since the electrons are continuously scavenged by the anode, the liberated H + accumulate and lead to anolyte acidification (Marcus et al. 2011). This has been shown to severely inhibit the catalytic activity of anodic biofilms and consequently leading to a complete shutdown of the BES (Cheng et al. 2010; Clauwaert et al. 2007; 2008). In general, a ph neutral condition (ph 7, i.e. [H + ]= 10-7 mol L -1 ) is essential to sustain optimal anodic microbial activities (He et al. 2008; Patil et al. 2011). In theory, ph neutral anodic condition can be sustained if the requirement of cation migration (to establish charge balance) is satisfied exclusively by proton migration, resulting in the proton flux from anode to cathode being equal to the electron flow and the proton generation at the anode. However, in reality cations other than protons (Na +, K +, Ca 2+, Mg 2+, NH + 4 ) are available in much higher concentrations and will hence tend to migrate instead of protons. The higher the concentration of other cations is relative to protons, the lower is the likelihood of proton migration. Under typical experimental conditions the concentration of alkali cations such as Na + 3
is orders of magnitudes higher than that of protons resulting in the migration of Na + from anode to cathode (Harnisch et al. 2008; Rozendal et al. 2006a). This effect has been suggested to be industrially exploited for the production of caustic soda (Rabaey et al. 2010). In laboratory trials, the problem of anolyte acidification is often masked by dosage of ph buffer or alkali (Cheng et al. 2010; Rozendal et al. 2008a). However, these ph control strategies are not sustainable as ph controlling chemicals must be externally added to the process. To approach self-sustaining ph control for CEM-BES without using external chemicals, the migrating cation species across the CEM should posses four properties: (1) it must be an alkaline species neutralizing the excess protons in the anolyte; (2) upon reacting with proton it becomes a cation that migrates across the CEM to the catholyte to maintain charge balance; (3) in the catholyte, it readily dissociates and releases the proton and hence replenishes the proton consumed in the cathodic reaction; (4) upon releasing the proton in the catholyte, the species can be recycled for neutralizing the anolyte again (Cord-Ruwisch et al. 2011). Amongst all cation species typically found in wastewaters (e.g. Na +, K +, Ca 2+, Mg 2+, NH + 4 ), ammonium (NH + 4 ) is the only species that fulfils all the above criteria. It has a characteristic acid-dissociation constant (pka value) of 9.25 (25 o C). Hence, once it has migrated across a CEM to the catholyte, where the localized ph exceeds 9.25, it dissociates predominately as free volatile ammonia (NH 3 ) which can be + recovered as a gas (Figure 1). The concept of using NH 4 as a proton shuttle in a CEM-equipped BES (CEM-BES) has been evaluated in our recent work (Cord- Ruwisch et al. 2011). The ammonia recycling was achieved by continuously stripping the catholyte with nitrogen (N 2 ), which was directed through the acidified anolyte to close the loop. N 2 was used to maintain anaerobic condition required for the microbial 4
anodic reaction. However, N 2 stripping is impractical as it incurs substantial energy input and creates large volume of low-value off-gas. In order to use the very effective principle of ammonia recycling for the sustainable operation of BES, a simple low energy input approach is required. This work examined a new approach to sustain current generation in a CEM- BES by internal ammonia recycling without using N 2 stripping. The idea is based on the well-known phenomenon where under anaerobic and highly reducing conditions (e.g. -500 mv vs. Ag/AgCl), a BES cathode produces hydrogen gas. This gas stream is in principle a vector that could help driving the volatilized ammonia out of the cathodic half cell (Liu et al. 2005; Logan et al. 2008; Rozendal et al. 2006b; Rozendal et al. 2008b). The aim of this study was to develop an effective way of sustaining BES operation by maintaining suitable anodic ph levels using ammonia from the cathode as the alkalinity carrier. In contrast to our previous work (Cord- Ruwisch et al., 2011), a more efficient (10 fold higher current density) anodic biofilm was used to demonstrate the concept of ammonia recycling in a BES. 5
2 Materials and methods 2.1 Bioelectrochemical system configuration and process monitoring A two-chamber CEM-BES was used in this study. It was made of transparent Perspex and has a similar configuration as the one described in an earlier work (Cheng et al. 2008). The two half cells were of equal volume and dimension (316 ml, 14 cm 12 cm 1.88 cm). They were physically separated by a cation exchange membrane (CMI-7000, Membrane International Inc.) which has a surface area of 168 cm 2. Both chambers were filled with conductive granular graphite (3-6 mm diameter), which reduced the void volume of the working chamber from 316 to 160 ml. A graphite rod (diameter 5 mm) was inserted into each half cell to allow electric contact between the graphite granules and the external circuit. In this study, only one half cell was inoculated with bacteria and was operated as an anodic half cell, which is termed here as the working chamber. The other half cell (cathodic) is termed as the counter chamber. The graphite granules inside the working chamber (i.e. working electrode) was polarized against a silver-silver chloride (Ag/AgCl) reference electrode (saturated KCl) at a potential of -300 mv by using a potentiostat (Model no. 362, EG&G, Princeton Applied Research, Instruments Pty. Ltd.). This potential was selected as it facilitated effective anodic acetate oxidation that could significantly acidify the anolyte if no effective ph control was implemented. The Ag/AgCl electrode was mounted inside the working chamber at a distance of less than one cm away from the working electrode. All electrode potentials (mv) in this article refer to values against the Ag/AgCl reference (ca. +197 mv vs. standard hydrogen electrode, (Bard and Faulkner 2001)). The working electrolyte redox potential and the ph of both working and counter electrolyte were continuously monitored. A computer program 6
(LabVIEW, National Instrument) was developed to continuously control and monitor the bioprocess. An analog input/ output data acquisition card (National Instrument TM ) was used to interface between the computer and the potentiostat. The BES current was monitored directly from the potentiostat. The electrode potentials, redox potentials and ph voltage signals were recorded at fixed time intervals and all data were regularly logged into an Excel spreadsheet. 2.2. Process start-up and general operation To start up the BES process, the working chamber was inoculated with returned activated sludge collected from a municipal sewage treatment plant (Subiaco, Perth, WA) (final mixed liquor suspended solid concentration was ca. 2 g L -1 ). A synthetic wastewater medium with a limited ph buffering capacity was used as both the working and counter electrolyte throughout the study. It consisted of (mg L -1 ): NH 4 Cl 125, NaHCO 3 125, MgSO 4 7H 2 O 51, CaCl 2 2H 2 O 300, FeSO 4 7H 2 O 6.25, and 1.25 ml L -1 of trace element solution, which contained (g L -1 ): ethylene-diamine tetraacetic acid (EDTA) 15, ZnSO 4 7H 2 O 0.43, CoCl 2 6H 2 O 0.24, MnCl 2 4H 2 O 0.99, CuSO 4 5H 2 O 0.25, NaMoO 4 2H 2 O 0.22, NiCl 2 6H 2 O 0.19, NaSeO 4 10H 2 O 0.21, H 3 BO 4 0.014, and NaWO 4 2H 2 O 0.050 (Cheng et al. 2010). The working electrolyte was supplemented with yeast extract (50 mg L -1 final concentration) as bacterial growth supplement during the initial start up period (ca. 2 weeks). During this period, the entire medium was refreshed once every 4 days. Acetate was used as the sole electron donor substrate for the anodic biofilm. A known amount of acetate standard solution (1 M) was added to the working electrolyte to obtain a desired acetate concentration (ranged from 1 to 50 mm). Unless stated otherwise, each half cell was hydraulically linked to a separate glass 7
recirculation bottle, which accommodated the extra volume of electrolyte; 0.5 and 0.7 L of anolyte and catholyte were continuously recirculating through the anodic and catholyte half cell, respectively at a rate of 6 L h -1. The CEM-BES was operated in batch mode at ambient temperature (25 ± 3 o C). 2.3. Ammonia recycling apparatus and its integration with the CEM-BES An ammonia-recycling apparatus was designed to enable recycling of ammonia from the cathodic headspace to the anolyte (Figure 2). It was constructed by gluing two 390-mL modified polyethylene terephthalate-made plastic containers together side by side with a common open window between them (5 cm 4 cm). The two containers were hydraulically separated from each other but were hydraulically connected with the anodic and cathodic half cell of the BES, respectively. As such, one container received the anolyte recirculation stream and the other the catholyte recirculation stream. The common window was sealed with a gas permeable fibre cloth (6 cm 5 cm), which was mounted at the anodic side and was continuously trickled with the anolyte. After trickling through the cloth, the anolyte was recirculated back to the BES anode. As both the influx and out flux rate of anolyte were identical (6 L h -1 ), a fixed volume of anolyte (ca. 50 ml) was always retained in the apparatus. At the cathodic side of the apparatus, the catholyte from the BES cathode was continuously sprinkled over the catholyte reservoir (ca. 50 ml) to facilitate ammonia volatilization within the cathodic container. Since the cathodic BES half cell was gas-tight, any build up of gas pressure would help drive the ammonia-containing gas in the cathodic container through the gas permeable cloth and the laminar flow of anolyte. This facilitated the dissolution of ammonia gas into the acidified anolyte. After retrofitted with the gas 8
exchange apparatus, the total volume of both the working and counter electrolyte was reduced to about 220 ml. The recirculation rates of both the anolyte and catholyte were kept at 6 L/h, corresponding to a hydraulic retention time of about 30 seconds for both anolyte and catholyte in the apparatus. 2.3 Experimental procedures Experiments were conducted to first investigate the effect of anolyte acidification on the anodic biofilm activity (current production). No ammonia recycling was implemented during this period. The initial biofilm establishment was facilitated by actively controlling the anolyte ph at around 7 using feedback dosing of NH 4 OH (1M). The dosage was computer recorded to obtain the NH 4 OH application rate. The catholyte was not ph controlled and hence it would become significantly alkaline, facilitating the NH 3 volatilization at the cathodic half cell. No aeration was provided to the catholyte. At the beginning of each batch run, the ammonia recycling apparatus was flushed with N 2 to obtain anaerobic condition. The effect of anolyte acidification was examined by comparing the current obtained with or without active dosing of NH 4 OH. Coulombic efficiencies of the anodic acetate oxidation (5mM) under ph neutral and acidified conditions were also compared (Logan et al. 2006). At selected time intervals, aliquots (1 ml) of anolyte and catholyte were sampled for chemical analysis. To demonstrate the effect of recycling the ammonia containing off-gas from the cathode to the anodic half cell, the catholyte in the recirculating bottle was continuously stripped with a pure nitrogen gas stream (~1 L min -1 ) which was then introduced back into the anolyte, completing an NH 3 recycling loop. Current, anolyte ph, acetate, ammonium, sodium and potassium concentrations were recorded over a 9
period of two days with or without NH 3 looping. To verify whether the ammonia transfer from the cathodic headspace to the anolyte could sustain the anodic biofilm activity, the BES was retrofitted with the ammonia recycling apparatus as described above. The effect of looping (i.e. gas exchanged was enabled via gas permeable cloth) on electrolyte ph and current was evaluated over an operational period of about 100 hours. 2.4 Chemical analysis Liquid samples taken from the BES were immediately filtered through a 0.2 µm filter (0.8/0.2 µm Supor Membrane, PALL Life Sciences) and were stored at 4 o C prior analysis. Acetate in the samples was analyzed using a Dionex ICS-3000 reagent free ion chromatography (RFIC) system equipped with an IonPac AS18 4 x 250 mm column (Cheng et al., 2012). 10 µl of the sample was injected into the column. Potassium hydroxide was used as an eluent at a flow rate of 1 ml min 1. The eluent concentration was 12-45 mm from 0-5 min, 45 mm from 5-8 min, 45-60 mm + from 8-10 min and 60-12 mm from 10-13 min. Ammonium (NH 4 -N), potassium and sodium in the filtered samples were measured with the same RFIC with a IonPac CG16, CS16, 5 mm column. Methansulfonic acid was used as an eluent with a flow rate of 1 ml min -1. The eluent concentration was 30 mm for 29 min. The temperature of the two columns was maintained at 30 C. Suppressed conductivity was used as the detection signal (ASRS ULTRA II 4 mm, 150 ma, AutoSuppressioin recycle mode). 10
3 Results and discussion 3.1 Anolyte acidification severely limited current generation The reactor described above was started up and operated to quantify the effect of ph drifts typically observed in CEM-equipped bioelectrochemical systems. After approximately two days of incubation, anodic current evolved gradually indicating that the microbial inoculum in the anodic chamber became electrochemically active. As expected, the current generation coincided with an acidification of the anolyte (ph 5.5) and an alkalization of the catholyte (ph 13), respectively (Figure 3A). As the working electrode (biofilm anode) was maintained at a constant potential (-300 mv) and the substrate was not limiting (acetate, > 10 mm), the observed current decline which followed the anodic acidification indicated that the low ph had suppressed the anodic activity of the biofilm. Similar observation has been reported by others (Cheng et al. 2010; Harnisch and Schroder 2009; He et al. 2008; Sleutels et al. 2010). 3.2 Effect of anolyte neutralisation on current production The ph neutralisation by the traditional NaOH addition was replaced by using an ammonium hydroxide solution. Neutralizing the anolyte acidity to ph 7 by dosing ammonium hydroxide immediately resumed current production (at 110 h, Figure 3). Prolonged current generation necessitated the demand of ammonium hydroxide, suggesting that this ph control approach could effectively sustain the anodic activity of the biofilm. As long as the anolyte ph was maintained neutral, the established 11
biofilm could effectively catalyze acetate oxidation with a good coulombic recovery (>80%) (Figure 4). The intermittent dosing of ammonium hydroxide into the anolyte resulted in a gradual increase of ammonium concentration in the catholyte, reaching a level of about 90 mm NH + 4 -N (Figure 3B). Although ammonium hydroxide was continuously added to the anolyte, its concentration stayed approximately 10-fold lower than in the catholyte (Figure 3B). Clearly, the ammonium kept migrating against its concentration gradient from the anolyte to the catholyte across the cation exchange membrane. Of the amount of ammonium migrated from the anolyte to the catholyte, 88.7% of the ammonium was lost from the catholyte (data not shown). Such a loss was most likely due to the high catholyte ph (>11) that favours the dissociation of ammonium as free ammonia and its volatilization from the catholyte. 3.3 Effect of NH 3 recycle from cathode to anolyte via N 2 gas stream To quantify to what extent recycling the migrated ammonium from catholyte back to the anolyte could overcome the detrimental anolyte acidification, N 2 gas was purged in series through the catholyte and then through the anolyte (Figure 5). Before allowing the N 2 flow, the CEM-BES responded to the addition of an acetate spike (10 mm) not only by producing an anodic current (Figure 5A) and degrading acetate, but also, as in Figure 4, by an associated ph drop in the anolyte (Figure 5B). As expected, the anolyte acidification had slowed down the anodic activity of the biofilm and the current declined significantly even though acetate was still present in excess (>3 mm). When switching on the N 2 flow (~21-27 h, Figure 5B), the current and acetate degradation resumed as explained the rise in ph caused by the alkalinity transfer from 12
the catholyte (as ammonia) to the anolyte. The BES now operated sustainably without a marked drift in anolyte ph. Repeating the stopping and starting of ammonia exchange by N 2 transfer showed the same principal effect. Further, the result also suggested that cations other than ammonium (sodium and potassium) also migrated across the CEM from the anolyte to the catholyte, but unlike ammonium which could be removed from the catholyte as a gas, both sodium and potassium accumulated in the catholyte during current production (Figure 5 C and D). This observation clearly highlights the uniqueness of ammonium as both the charge-balancing species and recyclable alkalinity carrier in a BES system. 3.4 Diffusional ammonia transfer from cathode to anode Using nitrogen to strip off the ammonia gas from the catholyte and recycle the ammonia containing nitrogen gas stream into the anolyte incurs extra energy to the extent where the energetic sustainability of a BES would become questionable. Further, in BES where H 2 is produced at the cathode, this valuable fuel gas would be lost via dilution with N 2. Instead of continuously purging N 2 through the cell, the recycling of ammonia could in theory be done by allowing the ammonia gas to vent from the catholyte and diffuse back to the anolyte. For this purpose a sufficiently large gas exchange window between the gas space of anodic and cathodic chamber would need to be used. This could be accomplished in a number of ways, including the use of a gas permeable membrane. To test whether diffusional gas exchange between the gas spaces of anodic and cathodic chamber enable adequate ammonia transfer, a separate apparatus was 13
designed to act as the extended gas space of the two chambers (Figure 2). The anolyte recycle was used to wet a vertically mounted gas permeable cloth that acted as the gas diffusion window, while the catholyte recycle was via a spray to facilitate NH 3 transfer from the alkaline catholyte to the gas phase (Figure 2). The vented NH 3 in the gas phase of the apparatus is expected to readily dissolve in the water saturated cloth which served as an ammonia scrubber. The anolyte acidity was thus neutralized with the aim of sustaining the anodic microbial activity. Below the effectiveness of this ammonia looping technique for sustaining the microbe-driven current of a CEM-BES is described (Figure 6). With the ammonia recycling apparatus (Figure 2) disconnected, the CEM-BES responded to the addition of an acetate spike (10 mmols) by instantly producing an anodic current and a decrease in anolyte ph to about 5.8. At this stage, another acetate addition (Ac 2 in Figure 6) did not resume the current demonstrating again that the CEM-BES had become inactive due to acidification of the anolyte. As soon as the gas exchange by the described apparatus was enabled, the anolyte ph was neutralised almost immediately (Figure 6B) allowing sustained current close to the maximum rate of this particular anodic biofilm as long as acetate was available. Renewed acetate addition at 55 h resumed the current. When the gas exchange between anolyte and catholyte was interrupted (~78 h) the anodic current could not be sustained again due to the anolyte acidification. By merely allowing the gas diffusion window between anodic and cathodic chamber (at ~96 h), continued current production could be sustained for several days as long as the anodic substrate (acetate) was not limiting (data not shown). 14
It was noticed that a net gas pressure build-up in the system occurred, presumably due to cathodic hydrogen production as the system was anaerobic and the cathode was maintained at highly reducing (negative) potentials (< -2V vs. Ag/AgCl, Figure 6C). This gas was released from the system via the vent (Figure 2), and may have participated in stripping of the NH 3 from the catholyte to the gaseous phase. Hydrogen transfer from the cathodic chamber to the ammonia recycling apparatus can be explained by minute bubbles visible in the catholyte. However, due to its poor solubility compared to ammonia the transfer of hydrogen to anodic chamber is expected to be minimal. 3.5 Practical considerations The experiments suggest that sustained current production in a CEM-BES is possible only if the anodic biofilm was operated under a neutral ph condition. Maintaining such a condition is difficult particularly for systems designed for high current output (Marcus et al. 2011; Picioreanu et al. 2010). This study demonstrated that without using conventional active ph control methods (e.g. regular dosage of alkali hydroxide or ph buffering chemicals), a relatively high anodic current (110 ma; 688 A m -3 net anodic chamber volume; 500 A m -3 total anolyte volume) could be sustained if a simple ammonia recycle via gas diffusion was implemented. Although the diffusivity of protons is about five times higher than that of ammonium (9.31 10-5 vs. 1.96 10-5 cm 2 s -1 (Vanysek 2000)), in the absence of ammonium the current is limited by the slow proton flux because of the low proton concentrations (10-6 M at ph 6). The presence of about 10 mm (10-2 M) of ammonium to the anode can enable an up to 10 4 times higher cation flux from anode 15
to cathode and hence a higher current. While high currents can also be guaranteed with the addition of other cations such as sodium, as is the case when using traditional sodium hydroxide based ph control of the anolyte (Cheng et al. 2008; Cheng et al. 2010), dosage of sodium hydroxide requires substantial energy costs (Cord-Ruwisch et al 2011) and intermittent renewal of the catholyte due to sodium accumulation (Rabaey et al. 2010). Nevertheless, the proposed concept has several constraints that may limit its practical use. For instance, this approach may only be applied in a microbial electrolysis mode as the production of hydrogen is essential to create a carrier stream to bring the ammonia from the catholyte to the anolyte. The high ph environment (>ph 9.25) required for the shift of NH + 4 /NH 3 equilibrium in favour of ammonia stripping may also limit the use of this approach for (cathodic) bioelectrosynthesis, which typically occurs under neutral ph condition. Further, the use of ammonia recycling in a BES can be unacceptable if the presence of ammonia in the anolyte is undesirable (e.g. wastewater treatment). If relying on ammonia as the proton carrier in a BES, the use of N 2 as the recycling agent of ammonia has been described as a proof of concept (Cord-Ruwisch et al 2011). However, due to costs and dilution of cathodic hydrogen the use of external N 2 is likely to be impractical for most applications. The gas exchange apparatus described in the present study showed immediate effects on anolyte neutralization and the production of anodic current, suggesting that the current was not limited by the rate of ammonia recycle but by the rate of microbial electron delivery to the anode. Accordingly, simpler and smaller gas exchange devices could be designed. Further research is warranted to explore how efficient a BES process 16
with the proposed NH 3 recycling mechanism could sustain cathodic hydrogen production. A further step towards implementing low energy ammonia recycle as a buildin proton carrier in BES could be the use of a gas permeable membrane between anolyte and catholyte or to develop and use a gas permeable cation exchange membrane. Such a system would not depend on the physical movement of anolyte and catholyte also should enable a more effective low-energy NH 3 recycle. To what extent membrane-free BES could profit from ammonium as proton carriers is yet to be investigated. It is conceivable that by using gas permeable membranes as part of the ion exchange separator (e.g. a CEM that is also gas permeable) the ammonia recycle could become a seamless feature of a BES. The presented work shows that ammonia readily migrates from the anode to the cathode against a rather strong diffusion gradient (~10 fold, Figure 3B). In principle that would mean that ammonium containing wastewaters treated in the BES anode would be likely to lose ammonia by migration to the catholyte. If this effect can be used to concentrate up ammonia by 10 or even 100 fold a novel way of ammonia recovery from wastewater could be envisaged. 4 Conclusions Overall, this study demonstrates that by continuously recycling ammonia from the catholyte to the anolyte in a CEM-BES, neutral ph condition desirable for the anodic biofilm could be maintained without using conventional ph control methods (e.g. regular dosage of alkaline hydroxide or ph buffering chemicals). The use of the 17
proposed gas exchange device was effective to recycle the ammonia without using energy-intensive N 2 stripping of the catholyte. This approach of ph control in BES systems has not been reported elsewhere, and it has the potential to be further developed to achieve recovery of ammonia from wastewaters. 5. Acknowledgement This work was funded by the CSIRO Water for a Healthy Country Flagship. 6. References 1. Bard AJ, Faulkner LR. 2001. Electrochemical methods: fundamentals and applications. New York, USA: John Wiley & Sons, Inc. 2. Cheng KY, Ho G, Cord-Ruwisch R. 2008. Affinity of microbial fuel cell biofilm for the anodic potential. Environ. Sci. Technol. 42, 3828-3834. 3. Cheng KY, Ho G, Cord-Ruwisch R. 2010. Anodophilic biofilm catalyzes cathodic oxygen reduction. Environ. Sci. Technol. 44, 518-525. 4. Cheng KY, Ginige MP, Kaksonen, AH. 2012. Ano-cathodophilic biofilm catalyzes both anodic carbon oxidation and cathodic denitrification. Environ. Sci. Technol. 46, 10372-10378. 5. Clauwaert P, Van Der Ha D, Boon N, Verbeken K, Verhaege M, Rabaey K, Verstraete W. 2007. Open air biocathode enables effective electricity generation with microbial fuel cells. Environ. Sci. Technol. 41, 7564-7569. 6. Cord-Ruwisch R, Law Y, Cheng KY. 2011. Ammonium as a sustainable proton shuttle in bioelectrochemical systems. Bioresource Technology 102, 9691-9696. 18
7. Harnisch F, Schroder U. 2009. Selectivity versus mobility: How to separate anode and cathode in microbial bioelectrochemical systems? Chem. Sus. Chem. 2, 921-926. 8. Harnisch F, Schroder U, Scholz F. 2008. The suitability of monopolar and bipolar ion exchange membranes as separators for biological fuel cells. Environ. Sci. Technol. 42, 1740-1746. 9. He Z, Huang Y, Manohar AK, Mansfeld F. 2008. Effect of electrolyte ph on the rate of the anodic and cathodic reactions in an air-cathode microbial fuel cell. Bioelectrochemistry 74, 78-82. 10. Liu H, Grot S, Logan BE. 2005. Electrochemically assisted microbial production of hydrogen from acetate. Environ. Sci. Technol. 39, 4317-4320. 11. Logan BE, Call D, Cheng S, Hamelers HVM, Sleutels THJA, Jeremiasse AW, Rozendal RA. 2008. Microbial electrolysis cells for high yield hydrogen gas production from organic matter. Environ. Sci. Technol. 42, 8630-8640. 12. Logan BE, Hamelers B, Rozendal RA, Schroder U, Keller J, Freguia S, Aelterman P, Verstraete W, Rabaey K. 2006. Microbial fuel cells: methodology and technology. Environ. Sci. Technol. 40, 5181-5192. 13. Marcus AK, Torres CI, Rittmann BE. 2011. Analysis of a microbial electrochemical cell using the proton condition in biofilm (PCBIOFILM) model. Bioresource Technol. 102, 253-262. 14. Patil SA, Harnisch F, Koch C, Hubschmann T, Fetzer I, Carmona-Martinez AA, Muller S, Schröder U. 2011. Electroactive mixed culture derived biofilms in microbial bioelectrochemical systems: the role of ph on biofilm formation, performance and composition. Bioresource Technol. 102, 9683-9690. 19
15. Picioreanu C, Loosdrecht MCM, Cirtis TP, Scott K. 2010. Model based evaluation of the effect of ph and electrode geometry on microbial fuel cell performance. Bioelectrochemistry 78, 8-24. 16. Rabaey K, Butzer S, Brown S, Keller J, Rozendal RA. 2010. High current generation coupled to caustic production using a lamellar bioelectrochemical system. Environ. Sci. Technol. 44, 4315-4321. 17. Rabaey K, Verstraete W. 2005. Microbial fuel cells: novel biotechnology for energy generation. Trends Biotechnol. 23, 291-298. 18. Rittmann BE. 2008. Opportunities for renewable bioenergy using microorganisms. Biotechnol. Bioeng. 100, 203-212. 19. Rozendal RA, Hamelers HVM, Buisman CJN. 2006a. Effects of membrane cation transport on ph and microbial fuel cell performance. Environ. Sci. Technol. 40, 5206-5211. 20. Rozendal RA, Hamelers HVM, Euverink GJW, Metz SJ, Buisman CJN. 2006b. Principle and perspectives of hydrogen production through biocatalyzed electrolysis. Int. J. Hydrogen Energy 31, 1632-1640. 21. Rozendal RA, Hamelers HVM, Rabaey K, Keller J, Buisman CJN. 2008a. Towards practical implementation of bioelectrochemical wastewater treatment. Trends Biotechnol. 26, 450-459. 22. Rozendal RA, Jeremiasse AW, Hamelers HVM, Buisman CJN. 2008b. Hydrogen production with a microbial biocathode. Environ. Sci. Technol. 42, 629-634. 23. Schroder U. 2007. Anodic electron transfer mechanisms in microbial fuel cells and their energy efficiency. Phys. Chem. Chem. Phys. 9, 2619-2629. 20
24. Sleutels THJA, Hamelers HV, Buisman CJ. 2010. Reduction of ph buffer requirement in bioelectrochemical systems. Environ. Sci. Technol. 44, 8259-8263. 25. Vanysek P. 2000. Ionic conductivity and diffusion at infinite dilution. In: Lide DR, editor. CRC Handbook of chemistry and physics. 81 ed. Boca Raton: CRC Press. p 2556. 21
Figure Captions Figure 1. The principle of using ammonium/ ammonia as a proton shuttle in a CEM- BES. Here, the cathode is anaerobically operated enabling abiotic hydrogen gas formation. At neutral anolyte ph (ph 6.5-7.5), ammonia predominately exists as NH + 4, whereas free volatile NH 3 dominates in the catholyte (ph>10). Figure 2. Schematic of the proposed ammonia recycling apparatus connected in-series with a CEM-BES. The process is anaerobic and hence favoring cathodic hydrogen production. Diagram not drawn to scale. Figure 3. Effect of continuous anodic ph control by ammonium hydroxide addition on the current of the BES and ammonium migration to the cathode. At time zero, the anodic chamber was inoculated with biofilm fragments from a microbial fuel cell operated with acetate. Arrow 1 and 2 indicate additions of sodium acetate: 0.5 mmole (1 mm) and 2 mmole (4 mm), respectively. Figure 4. Current production from a 5 mm acetate spike (dotted vertical lines) in the presence (0-22h) and absence (22-40h) of anodic ph control by automated ammonium hydroxide addition. The anode was potentiostatically controlled at -300 mv vs. Ag/AgCl. Figure 5. Effect on current generation of the CEM-BES of recycling ammonia from the cathode to the anolyte via a gas stream. Legend: (1) acetate was added to the anolyte; (2) purged N 2 through the catholyte and the off-gas was introduced directly into the anolyte (N 2 flow rate was about 1 L/min); (3) 22
sodium acetate was added to the anolyte; (4) nitrogen purging was terminated; (5) acetate was added to the anolyte. Figure 6. Effect of looping the gas-permeable path in the ammonia recycling apparatus on current production by the CEM-BES. Anodic potential was controlled at -300 mv vs. Ag/AgCl throughout the experiment. The BES was operated under anaerobic condition. Ac 1-4 represent additions of 10 (45 mm), 5 (23 mm), 9 (41 mm) and 9 (41 mm) mmole of sodium acetate to the anolyte, respectively. 23
Figure 1 Electron flow 0.25 CH 2 O 0.25 H 2 O 0.25CO 2 Bacterium 1e - H + NH 4 + NH 3 CEM H + Gas Liquid 1e - 0.5H 2 % formation 100 75 50 25 0 Bio-anode >98% as NH 4 + NH 4 + NH 3 Cathode >84% as freenh 3 5 7 9 11 13 ph 24
Figure 2 H 2 (g) + CO 2 (g) H + (aq) Acidified Stream NH 4 + (aq) NH 3(g) +H 2 (g) NH 3 Scrubber (Fibre Cloth) Ammonia Recycling Apparatus ph7 ph>10 Neutralized Stream Electrical current Anodic Chamber Cathodic Chamber CO 2 + H + COD Anode Cathode H 2 (g) H + (aq) Bioelectrochemical System NH 4 + (aq) NH 3(aq) Cation Exchange Membrane 25
Figure 3 Current (ma) Cummulative NH 4 OH Added (mmole) 150 100 50 0 200 150 100 50 (A) (B) No ph Control Current Anolyte ph Catholyte ph Acetate NH 4 OH Added NH 4+ -N Anolyte NH 4+ -N Catholyte ph Controlled with NH 4 OH (2) (1) 12 8 4 0 80 60 40 20 ph/ Acetate Conc. (mm) NH 4 + -N (mm) 0 24 48 72 96 120 144 168 192 96 Time (h) 0 26
Figure 4 Current (ma) 200 150 100 50 0 13 11 Neutral Anolyte ph (A) Acetate (B) Coulombic Recovery = 81 % Catholyte ph Acetate Anolyte Acidification Coulombic Recovery = 36 % ph 9 7 Anolyte ph Cumulated NH 4 OH (mmole) 40 20 0 (C) 0 10 Time (h) 20 30 40 27
Figure 5 No NH 3 Recycle NH 3 Recycled No NH 3 Recycle Current (ma) [1] [2] [3] [4] [5] 200 150 100 50 A 0 13 Acetate (mm) 10 5 Acetate Catholyte ph Anolyte ph B 11 ph 9 7 NH4 + -N (mm) Na + (mm) 0 80 Catholyte 60 Anolyte 40 20 0 80 60 40 20 0 Na + (Catholyte) K + (Catholyte) Na + (Anolyte) K + (Anolyte) 0 10 30 Time (h) C D 5 4 3 2 1 0 K + (mm) 28
Figure 6 Un-Loop NH 3 -Loop Un-Loop NH 3 -Loop Ac 1 Ac 2 Ac 3 Ac 4 A B C Time (h) 29
Graphical abstract Electron flow 0.25 CH 2 O 0.25 H 2 O 0.25CO 2 Bacterium 1e - H + NH 4 + NH 3 CEM H + Gas Liquid 1e - 0.5H 2 % formation 100 75 50 25 0 Bio-anode >98% as NH 4 + NH 4 + NH 3 Cathode >84% as freenh 3 5 7 9 11 13 ph 30
Highlights Anolyte acidification (ph<5.5) severely inhibited current generation Ammonium is selectively migrated across a cation exchange membrane to the catholyte Recycling the ammonia to anolyte neutralized the acidity and sustained current Ammonia recycling was achieved without using external carrier gases 31