Microbial Fuel Cell Using a Novel Ionic Liquid-Type Membrane–Cathode Assembly for Animal Slurry Treatment and Fertilizer Production

Abstract

The implementation of a microbial fuel cell for wastewater treatment and bioenergy production requires a cost reduction, especially when it comes to the ion exchange membrane part and the catalysts needed for this purpose. Ionic liquids in their immobilized phase in proton exchange membranes and non-noble catalysts, as alternatives to conventional systems, have been intensively investigated in recent years. In the present study, a new microbial fuel cell technology, based on an ionic liquid membrane assembly for CoCu mixed oxide catalysts, is proposed to treat animal slurry. The new low-cost membrane–cathode system is prepared in one single step, thus simplifying the manufacturing process of a membrane–cathode system. The novel MFCs based on the new low-cost membrane–cathode system achieved up to 51% of the power reached when platinum was used as a catalyst. Furthermore, the removal of organic matter in suspension after 12 days was higher than that achieved with a conventional system based on the use of platinum catalysts. Moreover, struvite, a precipitate consisting of ammonium, magnesium, and phosphate, which could be used as a fertilizer, was recovered using this membrane–cathode system.

1. Introduction

Microbial fuel cells (MFCs) form part of an emerging technology that makes it possible to convert the chemicals and biowastes contained in wastewater into electrical energy and other high-value products. Most publications are related to laboratory-scale applications, although there are also pilot- or semi-industrial-scale publications. The change of scale requires the search for cheaper and more efficient materials than conventional materials [1,2,3,4,5,6]. A microbial fuel cell consists of a cathode, an anode, and an ion exchange membrane. Platinum (Pt) is widely used as a catalyst in cathode systems to combat slow reactions, but it is expensive and prevents commercialization progress; Pt is also used in the ion exchange membrane, which is usually a perfluorinated organic polymer [7,8]. The research in this area in recent years has focused on the development of new nanostructured chemical catalysts, which are comparable with platinum in terms of reaction velocity, by using low-cost catalysts based on non-noble metals, in the hope of developing technologies that can be industrially implemented. Specifically, the catalysts described below were developed and tested in MFCs: (i) iron streptomycin [9], (ii) MnO₂ [10,11], (iii) iron aminoantipyrine [12], (iv) ferroelectric materials such as LiTaO₃ [13], (v) activated carbon [14], and (vi) biocathodes [15,16,17].

Regarding the proton exchange membrane, ionic liquids are used in proton exchange membranes as the active phase being used in the innovative phase in proton exchange membranes. Ionic liquids (ILs) are salts that exist in a liquid state at or near room temperature. Ionic liquids are composed of organic cations, the most common of which are tetraalkylammonium, tetraalkylphosphonium, N-alkylpyridine, and N,N′-dialkylimidazolium, and anions, such as hexafluorophosphate, tetrafluoroborate, triflimide, and triflate anions. Due to their unique properties, such as high ionic conductivity, a wide electrochemical stability window, and thermal stability, they were extensively studied as alternative electrolytes for various electrochemical devices, including batteries, capacitors, sensors, and electroplating [18]. Furthermore, as was recently demonstrated, they allow for the selective transport of ions [19]. Three generations of proton exchange membranes composed of ionic liquids were used in MFCs: (i) supported ionic liquid membranes [20]; (ii) polymeric ionic liquid membranes, which are also called ionogel, based on ionic liquids [21] (the advantage of the latter over the former being that they allow for a much larger quantity of ionic liquids to be immobilized, while increasing stability); and (iii) polymeric ionic liquid inclusion membranes, based on ionic liquids assembled onto the catalyst [22]. For these last membranes, the catalyst was sprayed on a carbon cloth, and the ionogel was created on the carbon cloth, which allows for close contact between the membrane and the cathode, with a consequent reduction in electrical resistance. In this work, a new simple procedure was assayed, in which the catalyst is located within the ionogel, based on the ionic liquid phase, which also favors contact between the proton exchange membrane and catalyst. These new membranes were assayed for slurry treatment.

Livestock farms are one of the sectors with great economic activity in Europe. Increased demand for meat products led to the development of intensive farming, which results in large amounts of slurry, contaminating both our atmosphere and soils due to the nitrogenous content. This made conventional treatment for slurry infeasible and, therefore, motivated the scientific community and business sector to seek new and sustainable solutions for the treatment of slurry. Slurry is a mixture of animal excrement, unconsumed feed material, and water, and it is commonly used in agriculture as a fertilizer and soil conditioner. However, if not managed properly, slurry can cause environmental problems. The high concentration of nitrogen and phosphorus in slurry can contribute to eutrophication in watercourses and reservoirs, leading to ecological imbalances in aquatic ecosystems. Furthermore, slurry has high values of biochemical and chemical oxygen demand (BOD and COD), which make slurry treatment by conventional methods difficult. The high COD and BOD levels can lead to oxygen depletion in water bodies, causing harm to aquatic life [23].

In the present study, slurry at a high COD concentration was used as a fuel in a microbial fuel cell technology based on a new type of ionic-liquid-type membrane–cathode assembly. The methodology proposed in this study requires only one step, i.e., the impregnation of the diffusion layer with a mixture of the ionic liquid and catalyst. This novel approach simplifies the fabrication of the membrane–catalyst system, thus reducing costs, especially with a view of industrialization. The bioenergy produced by microbial fuel cells, as well as the reduction in COD, BOD, and other physical–chemical parameters, was analyzed. The chemical composition and the morphology of the membrane–catalyst system were extensively studied using SEM–EDX techniques.

2. Materials and Methods

2.1. Fuel and Chemicals

Livestock slurry from the livestock farm of the Veterinary Faculty of the Universidad de Murcia was used as fuel. The wastewater also acted as the inoculum for the formation of an anodic bacterial community. The wastewater’s soluble chemical oxygen demand (COD) was found to be 2540 mg/L.

Polyvinylidene chloride and ionic liquids, which were used for preparing the ionic liquid membranes, were purchased from Sigma-Aldrich-Fluka, Kawasaki, Japan.

2.2. Synthesis of Copper and Cobalt Mixed Valence Oxides

Thermal decomposition method was used to synthesize $\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4$ [9,23]. Previous studies prepared copper–cobalt oxides at varying atomic ratios of Cu/Co and found that $\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4$ had the highest power output [9]. This catalyst also showed up to three Co oxidation states in its structure [24]. To prepare the co-precipitates of copper and cobalt hydroxides (CoOH₂, CuOH₂), an excess quantity of 3 M NaOH solution was slowly added over a mixture of CuCl₂ and CoCl₂, stirred for 7 h at 25 °C, and filtered. The resulting precipitates $\left(\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4\right)$ were washed with deionized hot water and dried at 60 °C for 24 h. The dried powder was heated between 350 and 400 °C for 8 h.

The crystallographic structure and purity of the oxides were characterized using X-ray diffraction (XRD) with a Bruker D8 Advance diffractometer (Bruker; Karlsruhe, Germany) with a generator of X-ray, Kristalloflex K 7608-80 F, with a copper anode. Particle size was measured with a JEOL JEM2100 transmission electron microscopy (TEM).

The TEM images showed copper–cobalt oxides nanoparticle with dimensions ranging between 17 and 45 nm. The XRD patterns of the catalyst revealed that the oxides have spinel lattice packed in cubic structure and purity of 100% (semi-quantitative analysis), with no other phases appearing in the samples analyzed.

2.3. Preparation of New Proton Exchange Membranes with Catalytic Activity Based on Ionic Liquids

The embedded membrane–cathode assembly was a polymeric inclusion membrane type. Polymer inclusion membranes based on ionic liquids were obtained by casting methods using PVC polymer and methyltrioctyl ammonium chloride ([MTOA⁺][Cl⁻]). The amount of ionic liquid was 70% w/w of the PVC/IL mixture. Two new types of embedded membrane–cathode assemblies were prepared and tested in MFCs. In the first preparation (P1), the IL and PVC base polymer were dissolved in THF, and 10 mg of the catalyst was added to the solution, which was then stirred for 10 min. The resulting suspension was poured into a Fluka glass ring (28 mm inner diameter, 30 mm height) on a glass plate and left to settle overnight until complete evaporation of THF. After evaporation, a thin plastic membrane was obtained, which was carefully peeled off the glass plate. Finally, the membrane and a piece of 3 cm diameter carbon cloth (Fuel Cell Earth, EEUU) were fixed with a round joint clip. In the second procedure (P2), the mixture of the IL methyltrioctylammonium chloride, ([MTOA⁺][Cl⁻]), with PVC (polyvinyl chloride), THF (tetrahydrofuran), and 10 mg of catalyst were poured directly onto a piece of 3 cm diameter carbon cloth cathodes placed between a Fluka glass ring and a Fluka glass plate. This protocol produces a composite material comprising proton exchange membrane, catalyst, and diffusion layer.

As a control (C), a PILIM based on [MTOA⁺][Cl⁻] and PVC 70% w/w and a platinum cathode was used. The membrane was manufactured by dissolving the IL and PVC in THF, before pouring the solution into a Fluka glass ring (Sigma-Aldrich International GmbH; Buchs, Switzerland) and evaporating the THF [25]. The cathode, based on platinum on carbon cloth, was fixed together with the ionic liquid membrane with a round join clip. The cathode was a mixture of 10.5 mg of a Pt/C mixture (nominally 60% platinum on a high surface area advance carbon support, Thermo Scientific, Waltham, MA, USA), water, isopropanol, and PTFE sprayed onto a 3 cm diameter piece of 10% water-proof carbon cloth (Fuel Cell Earth, Sacramento, CA, USA). The final load of platinum on the carbon cloth was 0.5 mg Pt cm⁻².

2.4. MFC Studies

Newly embedded membrane–cathode assemblies were employed in one-chamber MFCs for slurry treatment and evaluated in terms of power output, COD, and BOD reduction (Figure 1).

The experimental setup consisted of reactors made from modified 250 mL glass bottles with cylindrical flanges, where the temperature was maintained at 25 °C. The proton exchange membrane–cation system was identified by the labels P1, P2, and C, as mentioned earlier. The cathode was connected to the anode using a 1 kΩ resistor. The anode was composed of 100 g of graphite granules with a diameter of 3–5 mm and a graphite rod with a diameter of 3.18 mm. Anode chambers contained 200 mL of feed and were sealed with a lid equipped with a sampling port, ensuring anaerobic conditions throughout the experiment. The membrane–cathode system was securely fastened to the reactor flange using a round joint clip (see Figure 1). All tests were conducted in batch mode using wastewater as the sole source of microorganisms and fuel.

To establish biofilm attachment and stabilization, an initial period was required, after which polarization experiment was carried out at 192 h. Anodes and membrane–cathode systems were replaced for each trial, and daily sampling was performed by extracting a 5 mL aliquot with a syringe, which was then filtered through a 0.45 µm pore diameter membrane filter. A 2 mL aliquot was retained for COD analysis, which was carried out throughout the experiment. To ensure accuracy, three MFC replicates were run, and the data presented are an average of three reactor runs for each system.

2.5. Analytical Methods

2.5.1. Chemical Analysis

During treatment, the wastewater was characterized in terms of COD (chemical oxygen demand), BOD₅ (biochemical oxygen demand at 5 days), and UV254 (absorbance at 254 nm). Wastewater samples were collected regularly, using 0.45 nm nylon syringe filters (Fisher brand of Fisher Scientific, Waltham, MA, USA). When necessary, the samples were digested in a TR 420 thermoreactor of Spectroquant, from Merck Millipore, Burlington, MA, USA. The chemical composition of the samples was determined with spectrophotometry using Spectroquant 300 prove equipment from Merck Millipore (Danvers, MA, USA) using kits. The methods used were as follows:

• COD (chemical oxygen demand): test conducted in COD 145541 Supelco cuvettes (Sigma-Aldrich). Procedure was according to DIN ISO 15705 and approved by the USEPA for wastewater. The relative standard deviation of the assay was found to be less than 3%, indicating a high degree of repeatability.
• Biochemical oxygen demand at 5 days (BOD₅): A system of six Velp Scientifica DBO sensors was used for the manometric determination of BOD, using dicyanamide as an inhibitor for nitrification and sodium hydroxide as alkali to capture CO₂.
• Absorbance at 254 nm (UV254): as described previously [26], there is a correlation between dissolved organic matter and absorbance at 254 nm. The reduction in UV254 has also been related to the removal of organic micropollutants [27].
• Other parameters, including pH, conductivity, and temperature, were determined with a digital multimeter (sensION + MM150 from Hach Company; Loveland, CO, USA).

2.5.2. Electrochemical Analysis

To analyze the performance of the new fuel material (wastewater) in terms of electricity generation, polarization, internal resistance, and Coulombic efficiency were calculated.

Polarization Test

The voltage was continuously monitored every minute by a PCI 6010 data acquisition system (National Instruments, Austin, TX, USA). The voltage was also read intermittently off-line using a DVM891 digital multimeter (HQ Power, Berlin, Germany). Polarization was measured using a variable resistor box (5.77 MΩ, 953 kΩ, 486 kΩ, 96.5 kΩ, 50 kΩ, 11 kΩ, 6 kΩ, 1.1 kΩ, 561 Ω, 94.5 Ω, and 1.5 Ω) after 192 h of operation.

The voltage measurement was performed when the cell had reached a pseudo steady state under a specific resistor value. The time taken for the cell to reach the pseudo steady state was approximately 1 min. To ensure the reliability of the results, each measurement was conducted three times, and the mean value was reported. The relative standard deviation of the assay was found to be less than 10%, indicating a high degree of repeatability.

Internal Resistance

The internal resistance R_int (Ω) of each MFC was determined from Equation (1) [28]:

$${\mathrm{R}}_{\mathrm{int}}\left(\mathsf{\Omega }\right)=\frac{\mathrm{OCV}\left(\mathrm{V}\right)}{\mathrm{I}\left(\mathrm{A}\right)}-{\mathrm{R}}_{\mathrm{ext}}\left(\mathsf{\Omega }\right)$$

(1)

where OCV is the open circuit voltage, I is the current density at maximum power, and R_ext is the external resistance at maximum power.

Coulombic Efficiency

The concept of selectivity is used to evaluate Coulombic efficiency in microbial fuel cells, which measures the electric charge accumulated during substrate removal. Coulombic efficiency is determined by calculating the ratio of the number of coulombs transferred to the anode from the substrate to the maximum number of coulombs that would be transferred in the theorical case such that the entire substrate is transformed to current [28]:

$${\mathrm{Y}}_{\mathrm{Q}}=\frac{\mathrm{coulombs}\mathrm{transferred}}{\mathrm{total}\mathrm{theoretical}\mathrm{coulombs}\mathrm{produced}}·100$$

(2)

$${\mathrm{Y}}_{\mathrm{Q}}=\frac{{\mathrm{M}}_{\mathrm{m}}{{\displaystyle \int }}_0^{\mathrm{t}}\mathrm{i}\left(\mathrm{t}\right)\mathrm{dt}}{\mathrm{F}\Delta \mathrm{DQObV}}·100$$

(3)

where M_m is the molar mass of oxygen (32 gmol⁻¹), i(t) is the current intensity at a given time t, F is Faraday constant (96,485 Cmole⁻¹), ∆COD is the variation of COD in (mg/L), b is the number of moles of electrons produced per mole of oxygen (b = 4), and V is the volume of liquid in the anodic chamber (0.2 L).

2.6. SEM-EDX Characterization

To study the morphological characteristics, chemical composition, and distribution of elements in the membranes, a SEM equipped with an energy-dispersive X-ray (EDX) analyzer (APREO S and a JEOL-6100, from Jeol Ltd., Tokyo, Japan) was employed. The PILIMs were characterized using SEM-EDX immediately after preparation (referred to as fresh membranes) and after MFC operation (referred to as test membranes).

3. Results and Discussion

3.1. Electrochemical Analysis

Initially, 200 mL of slurry was added to form a biofilm around the graphite granules. The MFC voltage was continuously monitored for more than 280 h with a 1 kΩ external resistor, for different membrane–cathode systems. During the first 72 h (maturation period), the voltage increased until a plateau was reached. For the tests, fresh fuel was added, waiting 192 h for the voltage to reach a pseudo-steady state, after which the polarization tests were carried out. Polarization curves were obtained by varying the resistance between both electrodes and determining at which resistance the maximum current density is produced.

Figure 2 shows the polarization curves with different embedded membrane–cathode assemblies (P1, P2, and C). Figure 2 shows the maximum value of OCV for C at 500 mV. P1 and P2 OCV´s values were around 430 mV. At low intensities, data represent the activation overpotential of the electrodes, the linear part then represents the ohmic losses of the system and finally mass transfer losses can be determined at high current concentrations. P1 and P2 show two different slopes corresponding to activation and ohmic losses, with limiting current densities reached at 653 mA m⁻² and 1478 mA m⁻² for P1 and P2, respectively. In the case of membrane–cathode system C, polarization behavior was linear throughout, and with limiting current densities of around 1753 mA m⁻².

The power output curves (Figure 2B) showed that the highest power densities were obtained with the control membrane–cathode system based on platinum, with a maximum value of 230 mWm⁻². The power density for the P2 membrane–cathode system was 118 mWm⁻² (51% the value of C), and for P1, the maximum recorded power was 52 mWm⁻² (23% the value of C) (see Figure 2B). Kiely et al. [29] used an air cathode microbial fuel cell where the cathode is constructed by applying platinum (0.5 mg/cm²) and four diffusion layers (PTFE) to 30% wet-proofed carbon cloth. They used dairy manure wastewater with a soluble chemical oxygen demand of 450 mg/L. The maximum power density was 189 mW/m².

Polarization curves provide information on the electrochemical behavior of system and help determine the internal resistance (see Figure 2). The different embedded membrane–cathode assembly systems showed different internal resistances due to their composition and the way in which they are manufactured. The control (C), which is composed of platinum, showed the lowest internal resistance (833 Ω), and consequently, the greatest maximum power. The resistance of P2 (1093 Ω) was similar to that of the MFC with platinum and lower than a half of the resistance of P1 (2465 Ω). The greater power produced by P2 compared with P1 could be due to the membrane–cathode system of P2 being manufactured in one step. For P1, the catalytic membrane is placed on a carbon cloth after the ionogel is prepared, which would increase the internal resistance compared with P2.

In all cases, the Coulombic efficiency increased with time, which could be due to a more efficient oxygen reduction reaction and, consequently, a better utilization of the chemical energy contained in organic matter. In general, the Coulombic efficiency of P2 was always higher than that of P1 (Figure 3). This could be explained by the fact that the new composite material P2 offers less ionic resistance since it is prepared in one single step.

The most common method for preparing the cathode involves spraying Pt/C, water, isopropanol, and PTFE on a diffusion layer and used a proton exchange membrane on the diffusion layer [9]. The proposed method is simpler, as it is based on mixing the ionic liquid phase, which also acts as a membrane, and the catalyst. The new preparation technique is also malleable if intended for industrial purposes.

3.2. Slurry Wastewater Treatment Using Microbial Fuel Cell

The initial COD of the slurry wastewater sample was found to be 2540 mg/L, and its behavior was assessed over the test time (288 h) for the three types of microbial fuel cells (P1, P2, and C). The data obtained provided the COD reduction percentage for each MFC (Figure 4).

The rate of COD reduction was nearly constant during the first 96 h with a 41% reduction obtained for P2, and around 25% for P1 and C. After this time, the rate of COD reduction decreased, reaching a COD reduction during the last phase (192–288 h) of 54% for the P2 system, 41% for the P1 system, and 31% for C. Considering the removal of COD at initial time (until 96 h), the COD removal rate values were 10 mg L⁻¹ h⁻¹ for P1 and 6.3 mg L⁻¹ h⁻¹ for P2 and C. Furthermore, taking into account the higher Coulombic efficiency of platinum compared with P1 and P2, the higher values of COD reduction for P1 and P2 could be explained due to the organic matter oxidation in the anodic compartment. The formulation of the platinum cathodes could affect the activity of microorganisms in the anodic compartment decreasing their natural activity towards the degradation of organic matter. The results of the analysis of dissolved organic compound (DOC) based on the absorbance at 254 nm are shown in

Table 1. % DOC reduction in MFC based on membrane–cathode systems of P1, P2, and C.

Absorbance at 254 nm (Dilution 1:10)
MFC	${\left[UV254\right]}_0$	${\left[UV254\right]}_{t=288 h}$	${\left[\%DOC\right]}_{t=288 h}$
P1	0.847	0.3465	59.09
P2	0.847	0.269	68.24
C	0.847	0.393	53.60

.

A similar trend was observed for %COD and %DOC removal, since higher values were observed for P2 and lower values for P1. Indeed, these two parameters correlated well with the literature. The initial value of biodegradable organic matter was 1167 mg L⁻¹. The % of BOD reduction reached at the end of the experiment was 89%, 77%, and 63% for P1, P2, and C, respectively.

In the literature, there are very few research works that analyze the use of slurry as fuel in microbial fuel cells. Furthermore, other works usually consider slurry with considerably less COD than the one used in this work. Yokohama et al. [30] studied the purification of slurry containing the feces and urine of Holstein cows using cathode-in-air microbial fuel cells. The cathode contained 0.5 mg/cm² of platinum catalyst, attached to a Nafion-115 proton exchange membrane. The initial COD values were 1010 mg/L. In this work, the BOD reduction was around 80% at 70 h, and the COD reduction was around 70%. The maximum power density was around 0.34 mW/m². Similar BOD reduction values were found in both research works; however, higher COD reduction was found in the Yokoyama’s work [30]. The lower COD reduction could be due to the higher ammonium concentration in our sample. The initial ammonium concentration in our wastewater was around 620 mg/L and in the previous work was around 38.9 mg/L. The high values of COD could involve a high value of ammonium, which is toxic and hinders the degradability of the COD. However, the BOD values are relatively high, and they are similar to those obtained in the literature.

3.3. Characterization of the Membrane–Catalyst Assembly

The membrane–catalyst assembly systems (P1, P2, and C) were characterized using SEM-EDX before and after use in the MFCs. Figure 5 shows the SEM micrographs corresponding to the P1 membrane on the side in contact with the anodic solution before testing and after >280 h of testing. The air side after operation is also shown in Figure 5.

Figure 5A shows a relatively homogeneous surface, with some roughness and grooves. After using the catalytic membranes (Figure 5B), the surface becomes more homogeneous, and deposits appear on it, which are assumed to be salts from the anodic solution as are discussed below in

Table 2. Peak element vs. weight (%) and atomic (%) for EDX spectra of PEM–catalyst assembly (P1) based on [MTOA⁺][Cl⁻] ionic liquids, 30% PVC, and $\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4$. (A) Anodic solution side before operation; (B) anodic solution side after 280 h operation; (B*) analysis of the surface deposits observed on the anodic solution side after 280 h operation; and (C) air side after 280 h operation.

PEM–Catalyst Assembly (P1) ([MTOA+][Cl−]-PVC-$\left(C{u}_{0.3}C{o}_{0.7}\right)C{o}_2{O}_4$)
Peak Element	Weight % A	Atomic % A	Weight % B	Atomic % B	Weight % B*	Atomic % B*	Weight % C	Atomic % C
C K	56.25	74.46	53.87	75.30	16.69	33.55	40.16	65.15
N K	1.57	1.78	−0.43	−0.51	3.11	5.36	2.45	3.40
O K	8.23	8.18	3.62	3.81	-	-	4.38	5.34
Na K	1.82	1.26	4.54	3.32	22.75	23.89	-	-
S K	-	-	0.79	0.41	0.49	0.37	1.64	1.00
Cl K	31.65	14.20	34.22	16.22	51.03	34.76	37.88	20.82
K K	-	-	0.80	0.34	0.44	0.27	-	-
Ca K	-	-	2.13	0.89	0.84	0.51	2.16	2.05
Co K	0.36	0.10	0.38	0.11	0.37	0.15	6.49	2.15
Cu K	0.13	0.03	0.07	0.02	0.07	0.03	0.62	0.9
Zr K	-	-	-	-	4.22	1.12	4.22	0.9

. Figure 5C shows the SEM micrograph of the side in contact with the carbon cloth after 280 h of operation. This side presents a more homogeneous appearance than the opposite face in contact with the anodic solution (Figure 5B), with no deposits on its surface.

Table 2 shows the results of the EDX technique applied to the surface of the catalytic membrane of P1 before and after use in a single-chamber microbial fuel cell for 280 h. Note that the EDX spectra are from a sample of up to a few micrometers thick/deep so surface catalytic membrane is analyzed.

The EDX spectrum of PVC presented the characteristic peaks assigned to the C and N K lines corresponding to the chemical formulation of PVC. Hydrogen and other light elements were not detected using EDX. The EDX study of the membrane–cathode assembly was based on the selection of characteristic elements: N and Cl for [MTOA⁺][Cl⁻], Cl for PVC, and Co and Cu for the metallic catalyst $\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4$. The relative peak heights of identical elements in the different compounds were roughly related with their respective concentrations.

The nitrogen peaks are mainly due to the presence of the ionic liquid [MTOA⁺][Cl⁻]. Nitrogen is present in fresh (unused) P1 (Table 2, column A), which disappears when EDX is applied to the entire surface in contact with the anodic solution for 280 h (Table 2, column B). However, when the EDX was carried out for the deposits observed on the surface (see Figure 2B), traces of nitrogen were evident (Table 2, column B*). On the face in contact with the carbon cloth (external face), the presence of nitrogen was still evident after 280 h of use (Table 2, column C). As previously mentioned, EDX provides information on the composition of the first few micrometers of the membrane. Taking this into account and from the results discussed above, we can assume that the ionic liquid has been washed, to a certain extent, on the outermost surface of the face of P1 in contact with the anodic solution meanwhile the ionic liquid remained on the outer face (in contact with the carbon cloth). The selection of the ionic liquid ([MTOA⁺][Cl⁻]) was made taking into account the low solubility of this ionic liquid in water (<0.02% (v/v)), in order to increase the stability of the membrane in aqueous solutions. Furthermore, the inclusion method with PVC has been demonstrated to increase the stability of the liquid membrane in the face of aqueous media [31].

The deposits found in the membranes after use (Figure 2B), besides the ionic liquid, included other elements (Table 2, column B*) such as sodium, sulfur, potassium, and calcium, probably from salts contained in the anodic solution.

However, the proportion of catalytic metals, Co and Cu, remained practically constant before and after the test in the fuel cell (Table 2, columns A and B), which demonstrated the stability of the catalyst in the membrane during the test time (280 h).

The morphology of the P2 (Figure A1) type membrane–cathode assembly differs from that of P1. In P2, the ionic liquid and PVC solution, together with the catalyst, are poured onto a carbon cloth that acts as a diffusion layer and electrical conductor. The fiber-like appearance is due to the carbon cloth. P2 has a larger active surface than P1, which may be important for the rate of the chemical reactions taking place in it and, therefore, in the power of the fuel cell. SEM micrographs corresponding to the side in contact with the anodic solution before the test (A), the same side after 280 h of testing (B), and the side that is in contact with air after 280 h (C) (Figure 1, Appendix A). There is no great morphological difference between them, although, in micrograph C, a uniform layer is evident on the carbon cloth. This morphology may be explained by the fact that this side was the one that remained in contact with a glass sheet (upside down) when P2 was formed and, consequently, the ionic liquid and PVC mixture accumulated on this face.

Table 3. Peak element vs. weight (%) and atomic (%) for EDX spectra of PEM–catalyst assembly (P2) based on [MTOA⁺][Cl⁻] ionic liquids, 30% PVC, and $\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4$. (A) Anodic solution side before operation; (B) anodic solution side after 280 h operation; and (C) air side after 280 h operation.

PEM–Catalyst Assembly (P2) ([MTOA+][Cl−]-PVC-$\left(C{u}_{0.3}C{o}_{0.7}\right)C{o}_2{O}_4$)-Carbon Cloth
Peak Element	Weight % A	Atomic % A	Weight % B	Atomic % B	Weight % C	Atomic % C
C K	68.46	86.56	74.92	90.76	52.08	75.34
N K	1.88	2.04	−3.17	−3.29	0.46	0.57
F K	-	-	5.85	4.48	-	-
O K	-	-	-	-	2.84	3.09
Si K	-	-	-	-	0.62	0.38
S K	-	-	0.29	0.13	1.10	0.59
Cl K	23.51	10.07	15.41	6.32	39.43	19.39
Ca K	-	-	2.22	0.80	-	-
Co K	3.02	0.78	0.69	0.17	0.47	0.14
Cu K	0.41	0.10	0.27	0.06	−0.10	−0.03
Zr K	2.72	0.45	3.53	0.56	3.09	0.59

shows the results of the EDX technique applied to the surface of the catalytic membrane of P2 before use and after use in a single-chamber microbial fuel cell for 280 h.

As a first observation, it should be noted that the % weight of the carbon element is greater in P2 than in P1, which would be related with the composition of P2 based on carbon cloth. The nitrogen element, related with the composition of the ionic liquid, [MTOA⁺][Cl⁻], disappears from the side in contact with the anode compartment due to the partial washing of the first micrometers of depth of the catalytic membrane (column B). However, it does not disappear from the side of the catalytic membrane in contact with air (column C). Furthermore, both Co and Cu catalysts remain on the membrane after use (columns C and B). Other elements present in the sample such as calcium, silicon, or fluorine may come from the residual water or impurities of the raw materials.

In the case of PEM–catalyst assembly control system (C) (Figure A2), the fresh membrane (A) had a smooth surface that began to show deposits after 280 h of operation (B). The side of the membrane in contact with the carbon cloth (C) had less deposits than B after 280 h of operation. The morphology of the carbon cloth in contact with the membrane–catalyst assembly after 280 h of operation (D) showed a conventional crossed-fiber appearance. The side of the carbon cloth in contact with the air after operation (data not shown) was similar in appearance to the side in contact with the membrane–catalyst assembly (D).

The behavior of the control system (composed of ionic liquid-based proton exchange membrane and diffusion layer impregnated with Pt with Teflon) (

Table 4. Peak element vs. weight (%) and atomic (%) for EDX spectra of PEM–catalyst assembly (C-control) based on a PEM ([MTOA⁺][Cl⁻]/PVC, 70/30) and $\left({\mathrm{Cu}}_{0.3}{\mathrm{Co}}_{0.7}\right){\mathrm{Co}}_2{\mathrm{O}}_4 \mathrm{supported}\mathrm{on}\mathrm{carbon}\mathrm{cloth}$. (A) Anodic solution side before operation; (B) anodic solution side after 280 h operation; and (C) air side after 280 h operation.

Control System (C) [MTOA+][Cl−]-PVC-Pt-Carbon Cloth
Peak Element	Weight % A	Atomic % A	Weight % B	Atomic % B	Weight % C	Atomic % C	Weight % D	Atomic % D
C K	64.53	83.05	46.31	74.12	55.82	77.85	71.91	80.45
N K	2.22	2.44	−3.98	−5.47	1.17	1.40	0.73	0.70
F K	-	-	-	-	-	-	26.54	18.77
OK	-	-	-	-	4.26	4.46	-	-
S K	-	-	2.87	1.72	2.10	1.10	-	-
Cl K	33.25	14.50	53.09	28.79	31.15	14.42	0.06	0.02
K K	-	-	1.72	0.85	-	-	-
Pt M	-	-	-	-	5.51	4.47	0.75	0.05

) was similar to that found in previous studies (P1 and P2) (Table 2 and Table 3). The side in contact with the anodic dilution lost ionic liquid from the surface after 280 h of operation (column B), while the liquid remained on the membrane side in contact with the diffusion layer (carbon cloth) because the N element appeared in the EDX spectrum (column C). This surface also contained Pt since it was in contact with the diffusion layer, which was impregnated with Pt. The diffusion layer in contact with the ionic liquid-based PEM also contained Pt and F (column D). F element was used in the protocol to support Pt on the carbon cloth. In regards the SEM-EDX analysis, it is important to point out that crystals of struvite were found on the anodic side of the membrane–catalyst assembly systems (P1, P2, and C) (see Figure A3). Struvite (magnesium ammonium phosphate) is a phosphate mineral with the formula NH₄MgPO₄·6H₂O, which crystallizes with an orthorhombic geometry. Struvite is used as an agricultural fertilizer since it contains P and N, two of the three major plant macronutrients, along with Mg as a minor macronutrient. SEM micrographs showed orthorhombic crystals on the surface of the membrane–catalyst systems. The EDX spectrum of struvite crystal showed two big peaks, which corresponded to Mg K and P K.

Recently, there has been a great deal of interest in the production of struvite in microbial fuel cells. Furthermore, previous work reported on the accelerated recovery of struvite with the addition of magnesium [32]. The effluent from this process was used to obtain bioenergy and to reduce its COD more efficiently.

When the concentrations of Mg²⁺, NH₄⁺, and PO₄³⁻ exceed the limit for struvite formation, it leads to the precipitation of struvite crystals. The solubility of struvite decreases as the pH increases, and it is also influenced by the ionic strength of the solution [33]. The decrease in pH and the reduction in ionic strength favor the formation of struvite. In this work, it is observed that struvite precipitation occurred on the membrane–catalyst assembly, indicating that the pH near the cathode should increase.

During oxygen reduction reactions, protons are consumed at the electrode surface, and an increase in pH in the vicinity of the electrode surface is expected [34]. Furthermore, the effect of the ionic liquid of the membrane in the cathode microenvironment also needs to be considered. Similar results were found by Ichihashi and Hirooka [35] when they treated swine wastewater by using microbial fuel cell technology. According to their proposal, struvite crystal formation was triggered by the cathodic reaction and resulted from the rise in pH near the cathode.

Although we have worked in an experimental one-chamber lab system, for an industrial application, we propose to work with a stack made up of different membrane–electrode assemblies (MEAs). During the MEA cleaning process, the struvite could be detached and recovered for fertilizer application since struvite is precipitated as microcrystals, which are easy to remove from the MEA. Also, it is important to point out that in this study, the lack of any biofilms observed in the membranes based on ionic liquids was due to the antibiofilm activity of ionic liquids at high concentration as it seen in ionic liquid membranes.

4. Conclusions

In this work, microbial fuel cell technology based on a new type of membrane–cathode systems was studied for slurry purification. In these new membrane–cathode systems, the non-noble catalyst copper and cobalt mixed valence oxides are suspended in the ionic liquid phase and not sprayed on the diffusion layer as it is conventionally conducted. In the P2 membrane–cathode assembly, the internal resistance was significantly reduced, and the maximum power reached was 120 mWm⁻², which corresponds to 51.3% of the value obtained with the control using a platinum-based cathode. This membrane–cathode system P2 attained >50% organic matter removal at 288 h, which was higher than that was achieved with the platinum based-cathode. This novel membrane–cathode system allows easier manufacturing and production of cheaper membranes and catalysts than that conventionally used in microbial fuel cells. This would reduce the capital and the processing costs involved in the case of the industrial manufacture of membrane–cathode systems. Furthermore, struvite fertilizer was obtained as a precipitate in all the membrane–cathode assemblies. The proposed MFC permits the optimal use of slurry wastewater for bioenergy and fertilizer production, accompanied by water decontamination.