The Membrane-Less Microbial Fuel Cell (ML-MFC) with Ni-Co and Cu-B Cathode Powered by the Process Wastewater from Yeast Production

: Research related to measurements of electricity production was combined with parallel wastewater parameter reduction in a membrane-less microbial fuel cell (ML-MFC) fed with industry process wastewater (from a yeast factory). Electrodes with Ni–Co and Cu–B catalysts were used as cathodes. A carbon electrode (carbon cloth) was used as a reference due to its widespread use. It was demonstrated that all analyzed electrodes could be employed as cathodes in ML-MFC fed with process wastewater from yeast production. Electricity measurements during ML-MFC operations indicated that power (6.19 mW) and current density (0.38 mA · cm − 2 ) were the highest for Ni–Co electrodes. In addition, during the exploitation of ML-MFC, it was recorded that the chemical oxygen demand (COD) removal per time for all types of electrodes was similar to the duration of COD decrease in the conditions for wastewater aeration. However, the COD reduction curve for aeration took the most favorable course. The concentration of NH 4 + in ML-MFC remained virtually constant throughout the measurement period, whereas NO 3 − levels indicated almost complete removal (with a minimum increase in the last days of cell exploitation).


Introduction
For many years, studies have been focused on a technology that can offer both energy and raw material production in a process of bioconversion from liquid waste. Such a solution could apply microbial fuel cells (MFCs) that could provide manners of generating electricity directly from biodegradable raw materials using bacterial or fungal cells [1,2]. The volume of energy accumulated in sewage is nine times higher than that needed for its utilization by conventional methods, and microbiological cells are the adequate technology to enable the use of this potential [1,[3][4][5]. MFCs are bio-electrochemical systems (devices) in which bacteria are applied as catalysts for the oxidation of organic and inorganic matter, and for generating electricity [1,6]. In MFCs, oxidation of organic material occurs on the anode, and the results of this reaction include e − , H + ions, and CO 2 . An example reaction for pure glucose can be defined as follows [1,7,8].

Materials and Methods
Process wastewater derived from yeast production was applied for the present research. Yeast wastewater from vacuum filters was used in the measurements. Such filters are applied to extract water from yeast milk. This wastewater has the smallest concentration of pollutants of all process wastewater generated in the yeast production process [34]. Figure 1 presents the site where wastewater applied and the measurements were generated (Figure 1, W5).

Figure 1.
Source of wastewater (W5) applied in measurements: W1, wastewater after clarification of the molasses; W2, wastewater after the first centrifugation of the wort; W3, wastewater after the second centrifugation of the wort; W4, wastewater after the third centrifugation of the wort; W5, wastewater after obtaining yeast on vacuum filters. Yp is yeast for packaging and distribution.
In Table 1, a summary of the initial parameters of wastewater deployed in the measurements can be found (W5 in Figure 1). Table 1. Initial parameters of process wastewater used in research, yeast process wastewater derived from vacuum filters.
Measurements of sewage parameters were carried out for three different reactors: without aeration (R1), with aeration (R2), and an MFC (R3). The applied reactors R1 and R2 had a capacity of 15 L. In R1, the sewage was in contact with air only through the interface of the wastewater. In R3, the wastewater was constantly aerated (240 L·h −1 ). The third reactor (R3) was a ML-MFC. The Figure 1. Source of wastewater (W 5 ) applied in measurements: W 1 , wastewater after clarification of the molasses; W 2 , wastewater after the first centrifugation of the wort; W 3 , wastewater after the second centrifugation of the wort; W 4 , wastewater after the third centrifugation of the wort; W 5 , wastewater after obtaining yeast on vacuum filters. Yp is yeast for packaging and distribution.
In Table 1, a summary of the initial parameters of wastewater deployed in the measurements can be found (W 5 in Figure 1). Measurements of sewage parameters were carried out for three different reactors: without aeration (R1), with aeration (R2), and an MFC (R3). The applied reactors R1 and R2 had a capacity of 15 L. In R1, the sewage was in contact with air only through the interface of the wastewater. In R3, the wastewater was constantly aerated (240 L·h −1 ). The third reactor (R3) was a ML-MFC. The measurements were carried out until 90% of the COD reduction efficiency was obtained [64]. Figure 2 presents the three reactors, R1, R2, and R3, used in the measurements.
The ML-MFC took the shape of a pipe (120 × 250 mm). The level of wastewater in the MFC was about 200 mm. A plexiglass pipe was used in the design of the cell plexi (7 in Figure 3). From the bottom, the pipe was closed with a plug made with 3D-printing technology and sealed with silicone. The anode section contained the wastewater intake (11 in Figure 3), and the cathode section contained the wastewater outflow (1 in Figure 3). ML-MFC was combined with the external wastewater tank by the intake and outtake of the wastewater. Wastewater was slowly circulated (0.05 L·h −1 ) in a closed loop. The volume of sewage in the entire system was 15 L (ML-MFC + an external tank), which   measurements were carried out until 90% of the COD reduction efficiency was obtained [64]. Figure  2 presents the three reactors, R1, R2, and R3, used in the measurements.   . Scheme of the ML-MFC applied in measurements: 1-wastewater outflow (to wastewater tank); 2-glass balls; 3-glass wool; 4-plastic mesh; 5-anode; 6-plug; 7-plexiglass pipe; 8cathode; 9-pipe with holes; 10-air intake; 11-wastewater intake (from wastewater tank); 12cathode section; 13-anode section; and R-load.
The ML-MFC took the shape of a pipe (120 × 250 mm). The level of wastewater in the MFC was about 200 mm. A plexiglass pipe was used in the design of the cell plexi (7 in Figure 3). From the bottom, the pipe was closed with a plug made with 3D-printing technology and sealed with silicone. The anode section contained the wastewater intake (11 in Figure 3), and the cathode section contained the wastewater outflow (1 in Figure 3). ML-MFC was combined with the external wastewater tank by the intake and outtake of the wastewater. Wastewater was slowly circulated (0.05 L·h −1 ) in a closed loop. The volume of sewage in the entire system was 15 L (ML-MFC + an external tank), which Figure 3. Scheme of the ML-MFC applied in measurements: 1-wastewater outflow (to wastewater tank); 2-glass balls; 3-glass wool; 4-plastic mesh; 5-anode; 6-plug; 7-plexiglass pipe; 8-cathode; 9-pipe with holes; 10-air intake; 11-wastewater intake (from wastewater tank); 12-cathode section; 13-anode section; and R-load.
The ML-MFC took the shape of a pipe (120 × 250 mm). The level of wastewater in the MFC was about 200 mm. A plexiglass pipe was used in the design of the cell plexi (7 in Figure 3). From the bottom, the pipe was closed with a plug made with 3D-printing technology and sealed with silicone. The anode section contained the wastewater intake (11 in Figure 3), and the cathode section contained the wastewater outflow (1 in Figure 3). ML-MFC was combined with the external wastewater tank by the intake and outtake of the wastewater. Wastewater was slowly circulated (0.05 L·h −1 ) in a closed loop. The volume of sewage in the entire system was 15 L (ML-MFC + an external tank), which corresponded to the capacity of the other reactors ( Figure 2). The cell was divided into an anode (in the lower part) and a cathode (in the upper part). As a consequence of the location of the cathode chamber in the upper part of the cell, the possibility of anode aeration was largely eliminated. After the aeration of the cathode chamber, air was drained directly into the environment. In addition, the chambers (sections) were separated by a layer of glass wool and glass balls (2 and 3 in Figure 3) [62,63]. This procedure was designed to separate the aerated part (cathode space) from the unaerated section (anode space). To ensure that the glass wool and beads were kept at a given location, a spatial ABS (acrylonitrile butadiene styrene) plastic mesh was installed under a layer of mineral wool. This mesh was made by 3D printing and ensured the free flow of wastewater. ABS material in filament form (diameter 1.75 mm) was used as material for 3D printer. The 3D print was made with high accuracy (single print layer, 0.09 mm). The air was fed directly under the cathode. The cathode aeration performance was 5 L·h −1 .
In this research, the role of the anode was played by a carbon cloth, whereas the cathode was an electrode (copper mesh) with a Ni-Co catalyst (15% Co), an electrode (copper mesh) with a Cu-B catalyst (9% B), or a carbon cloth applied as the reference for the Ni-Co and Cu-B cathodes. The surface area of the anode was 15 cm 2 , whereas for the anode it was 10 cm 2 . The ML-MFC had an external charge in the form of an external 100 Ω resistor.
Ni-Co and Cu-B electrodes were designed and manufactured by electrolytic melt deposition [65,66]. A mixture comprised mainly of NiSO 4 and CoSO 4 was utilized to obtain the alloy of Ni-Co [65,67]. For the purpose of conducting electrochemical deposition, the temperatures of 293-323 K, at a current density of 1-3 A·dm −2 and a pH of 2.0-5.5, were used [25,65]. The mixture consisting mainly of NaBH 4 and CuSO 4 was utilized to obtain the Cu-B alloys [26]. To conduct this process, the temperatures of 355-365 K, at a current density 1-3 A·dm −2 and a pH of 2.0-5.0, were applied [26,66]. For the purpose of ensuring appropriate wettability of the copper mesh, it underwent washing with 25% aqueous KOH before the application process of electrolytic deposition of the alloy. Subsequently the copper mesh was washed in acetic acid and later in alcohol.
Measurements in the first two reactors (R1 and R2 in Figure 2) included measurements of wastewater parameters (COD, NH 4 + , and NO 3 − ), whereas the measurements related to the ML-MFC included both analysis of wastewater parameters and electrical parameters (COD, NH 4 + , and NO 3 − , and power, current density, and cell voltage). Measurements of the wastewater parameters and the cell voltage were carried out at one day intervals. Each such measurement was carried out in ten trials. The wastewater parameters (COD, NH 4 + , and NO 3 − ) can be determined by the analytical methods [68][69][70], however in this work, the colorimetric methods were used for the measurement of COD, NH 4 + , and NO 3 − . All measurements were carried out at 20 • C.
To prepare the 3D objects (plastic mesh and MFC plug) for printing, and for printing the spatial plastic mesh and the MFC plug, (4 and 6 in Figure 3) a M200 printer with Z-Suite software (Zortrax S.A, Olsztyn, Poland) was used. During measurements of the wastewater parameter a HI 83224 colorimeter (HANNA Instruments, Woonsocket, RI, USA) was used. During measurements of the current and cell voltage, an 8840A multimeter (Fluke Corporation, Everett, WA, USA) was used. For measurements of the current density a System 500 potentiostat (Amel S.l.r., Milano, Italy) with CorrWare software (Scribner Associates Inc., Southern Pines, NC, USA) was used.  As we can see from Figure 4, a COD removal rate of 90% (1050 mgL −1 ) from the initial value for the reactor without aeration (R1 in Figure 2) took 37 days. In the case of the reactor with aeration, the COD reduction to the assumed level (i.e., removal by 90%) took 17 days (R2 in Figure 2), which is about half of the time recorded for the case of reactor without aeration. The removal rate of this parameter by 90% (19,20, and 21 days depending on the type of electrode) for the ML-MFC was similar to the removal time in the process with aeration. It should be noted that, for all cathodes applied in the testing (Ni-Co, Cu-B, carbon cloth), the time it took for COD concentration to decrease in ML-MFC was similar. Despite the same COD reduction time for an aerated reactor (R2 in Figure 2) and for the ML-MFC (R3 in Figure 2), the curve characteristics for aeration were much more favorable. In the initial phase, the increase in the COD reduction rate took a shorter period of time. A removal efficiency of 80% (up to 2100 mgL −1 ) was obtained after just four days. After this time, the COD reduction rate decreased considerably. When the ML-MFC was used, the COD reduction remained virtually constant throughout the measurement period. The differences in the rate of reduction of this parameter for different catalysts were small. We should note, however, that for Ni-Co catalyst, 50% efficiency of COD reduction (up to 5250 mgL −1 ) was obtained over the first eight days, whereas for Cu-B catalysts and carbon cloth took over 9 and 10 days, respectively. After 11 days, however, COD concentration dropped the fastest when carbon cloth was applied as the cathode. Figure 5 presents details of the variations in NH4 + concentration over time. The variations in concentration were recorded for the three types of reactors and the three types of ML-MFC cathode (Ni-Co, Cu-B, carbon cloth). As we can see from Figure 4, a COD removal rate of 90% (1050 mg·L −1 ) from the initial value for the reactor without aeration (R1 in Figure 2) took 37 days. In the case of the reactor with aeration, the COD reduction to the assumed level (i.e., removal by 90%) took 17 days (R2 in Figure 2), which is about half of the time recorded for the case of reactor without aeration. The removal rate of this parameter by 90% (19,20, and 21 days depending on the type of electrode) for the ML-MFC was similar to the removal time in the process with aeration. It should be noted that, for all cathodes applied in the testing (Ni-Co, Cu-B, carbon cloth), the time it took for COD concentration to decrease in ML-MFC was similar. Despite the same COD reduction time for an aerated reactor (R2 in Figure 2) and for the ML-MFC (R3 in Figure 2), the curve characteristics for aeration were much more favorable. In the initial phase, the increase in the COD reduction rate took a shorter period of time. A removal efficiency of 80% (up to 2100 mg·L −1 ) was obtained after just four days. After this time, the COD reduction rate decreased considerably. When the ML-MFC was used, the COD reduction remained virtually constant throughout the measurement period. The differences in the rate of reduction of this parameter for different catalysts were small. We should note, however, that for Ni-Co catalyst, 50% efficiency of COD reduction (up to 5250 mg·L −1 ) was obtained over the first eight days, whereas for Cu-B catalysts and carbon cloth took over 9 and 10 days, respectively. After 11 days, however, COD concentration dropped the fastest when carbon cloth was applied as the cathode.  Figure 2) using yeast effluent, was it reduced by 70%. For the reactor without aeration (R1 in Figure 2), NH 4 + levels remained constant throughout the measurement period. However, during measurements using the ML-MFC (R3 in Figure 2) for all cathodes used (Ni-Co, Cu-B, carbon cloth), a slight increase in NH 4 + concentration was recorded.  Analysis of measurements of the shift in NH4 + concentration demonstrated that, only in the case of aeration (R2 in Figure 2) using yeast effluent, was it reduced by 70%. For the reactor without aeration (R1 in Figure 2), NH4 + levels remained constant throughout the measurement period. However, during measurements using the ML-MFC (R3 in Figure 2) for all cathodes used (Ni-Co, Cu-B, carbon cloth), a slight increase in NH4 + concentration was recorded.     Analysis of measurements of the shift in NH4 + concentration demonstrated that, only in the case of aeration (R2 in Figure 2) using yeast effluent, was it reduced by 70%. For the reactor without aeration (R1 in Figure 2), NH4 + levels remained constant throughout the measurement period. However, during measurements using the ML-MFC (R3 in Figure 2) for all cathodes used (Ni-Co, Cu-B, carbon cloth), a slight increase in NH4 + concentration was recorded. Figure 6 contains details of the shift in concentration of NO3 − over time. The variations in concentration were recorded for the three types of reactors and the three types of ML-MFC cathode (Ni-Co, Cu-B, carbon cloth).  During the measurement of shifts in NO 3 − concentration in the ML-MFC (for all cathodes used:

Results and Discussion
Ni-Co, Cu-B, and carbon cloth), it was observed that this level was reduced by 70% within 12 days, but that after that time, the concentration started to increase again. The increase in concentration after 12 days was, however, insignificant, and dependent on the type of the cathode applied in the test. The increase in NH 4 + concentration was due to the capture of hydrogen ions by the ammonia particle (occurring, e.g., during putrefaction) [71][72][73], whereas the increase in NO 3 − (Figures 5 and 6) concentration resulted from the nitrification process during bacterial growth [1,73,74]. The cell voltage (with a Cu-B, Ni-Co, or carbon cloth cathode) obtained during the operation of the ML-MFC (Figure 3) is shown in Figure 7. Measurements were made at one-day intervals.
During the measurement of shifts in NO3 − concentration in the ML-MFC (for all cathodes used: Ni-Co, Cu-B, and carbon cloth), it was observed that this level was reduced by 70% within 12 days, but that after that time, the concentration started to increase again. The increase in concentration after 12 days was, however, insignificant, and dependent on the type of the cathode applied in the test.
The cell voltage (with a Cu-B, Ni-Co, or carbon cloth cathode) obtained during the operation of the ML-MFC (Figure 3) is shown in Figure 7. Measurements were made at one-day intervals. As demonstrated in Figure 7, the cell voltage for each type of cathode was at a similar level. However, slightly higher voltage (0.1 V on average) was noted for the cathode with a Ni-Co catalyst, as well as the one comprising carbon cloth. Higher voltage for the Ni-Co cathode was noted especially in the early days of the operation of the ML-MFC cell. When Cu-B catalyst was used, the cell voltage was noticeably lower (compared to the use of Ni-Co and C catalysts) after 8 days of the operation of the ML-MFC.
The power curves of the ML-MFC (with a Cu-B, Ni-Co, or carbon cloth cathode) obtained during the operation of the ML-MFC (Figure 3) is shown in Figure 8. As demonstrated in Figure 7, the cell voltage for each type of cathode was at a similar level. However, slightly higher voltage (0.1 V on average) was noted for the cathode with a Ni-Co catalyst, as well as the one comprising carbon cloth. Higher voltage for the Ni-Co cathode was noted especially in the early days of the operation of the ML-MFC cell. When Cu-B catalyst was used, the cell voltage was noticeably lower (compared to the use of Ni-Co and C catalysts) after 8 days of the operation of the ML-MFC.
The power curves of the ML-MFC (with a Cu-B, Ni-Co, or carbon cloth cathode) obtained during the operation of the ML-MFC (Figure 3) is shown in Figure 8.
ML-MFC power measurements demonstrated that the highest values were more clearly obtained for electrodes (cathodes) made of carbon cloth and Ni-Co catalysts (Figure 8, black and red markers). When the cathode with a Cu-B catalyst was applied, the power output was clearly lower (Figure 8, yellow markers). In addition, based on the cell voltage, current level, and electrode surface, the maximum current densities obtained in the ML-MFC (Figure 3) for the various cathodes used were calculated. The following results were recorded: 0.38 mA·cm −2 and power output of 6.19 mW for the Ni-Co catalyst (Figure 8, red markers), 0.32 mA·cm −2 and a power output of 5.43 mW for the Cu-B catalyst (Figure 8, yellow markers), and 0.19 mA·cm −2 and a power output of 5.99 mW for the carbon cloth catalyst (Figure 8, black markers), respectively. Table 2 summarizes the electrical parameters of the ML-MFC for different types of cathodes. ML-MFC power measurements demonstrated that the highest values were more clearly obtained for electrodes (cathodes) made of carbon cloth and Ni-Co catalysts (Figure 8, black and red markers). When the cathode with a Cu-B catalyst was applied, the power output was clearly lower (Figure 8, yellow markers). In addition, based on the cell voltage, current level, and electrode surface, the maximum current densities obtained in the ML-MFC (Figure 3) for the various cathodes used were calculated. The following results were recorded: 0.38 mA·cm −2 and power output of 6.19 mW for the Ni-Co catalyst (Figure 8, red markers), 0.32 mA·cm −2 and a power output of 5.43 mW for the Cu-B catalyst (Figure 8, yellow markers), and 0.19 mA·cm −2 and a power output of 5.99 mW for the carbon cloth catalyst (Figure 8, black markers), respectively. Table 2 summarizes the electrical parameters of the ML-MFC for different types of cathodes.

Conclusions
The measurements demonstrated that the assumed level of 90% COD concentration removal was obtained by all investigated reactors ( Figure 4). However, the time needed to remove the COD was different for each type of reactor. The most adequate characteristics of the COD removal curve were recorded for the reactor with aeration ( Figure 4, blue markers). At this point, we should note that the final COD removal rate equal to 90% for the ML-MFC was similar to that of the reactor with aeration for all types of cathodes used (Figure 4, red, black, and yellow markers). However, the shortest removal time was obtained by a Ni-Co catalyst (Figure 4, red markers). In the case of the measurements of NH4 + concentration for ML-MFC cells, this parameter did not decrease and the concentration level even slightly increased ( Figure 5). The measurements of the variations in NO3 − concentration in the ML-MFC (for all cathodes used: Ni-Co, Cu-B, and carbon cloth) demonstrated that after 12 days this parameter was reduced by 70%; however, after that time the concentration began to increase again ( Figure 6). Cell voltage measurements indicated that for all cathodes the ML-MFC voltage remained at a similar level (0.9-1.0 V) for about 2 to 12 days. After 14 day, the voltage

Conclusions
The measurements demonstrated that the assumed level of 90% COD concentration removal was obtained by all investigated reactors ( Figure 4). However, the time needed to remove the COD was different for each type of reactor. The most adequate characteristics of the COD removal curve were recorded for the reactor with aeration ( Figure 4, blue markers). At this point, we should note that the final COD removal rate equal to 90% for the ML-MFC was similar to that of the reactor with aeration for all types of cathodes used (Figure 4, red, black, and yellow markers). However, the shortest removal time was obtained by a Ni-Co catalyst (Figure 4, red markers). In the case of the measurements of NH 4 + concentration for ML-MFC cells, this parameter did not decrease and the concentration level even slightly increased ( Figure 5). The measurements of the variations in NO 3 − concentration in the ML-MFC (for all cathodes used: Ni-Co, Cu-B, and carbon cloth) demonstrated that after 12 days this parameter was reduced by 70%; however, after that time the concentration began to increase again ( Figure 6). Cell voltage measurements indicated that for all cathodes the ML-MFC voltage remained at a similar level (0.9-1.0 V) for about 2 to 12 days. After 14 day, the voltage began to gradually decrease to about 0.8 V (the lowest for the Cu-B cathode). We can also note that the highest values were obtained for the Ni-Co cathode (similar values were obtained for the carbon cloth), whereas the lowest values were obtained for the Cu-B cathode. The highest ML-MFC power values were also obtained using a Ni-Co cathode, although the parameters were similar to those obtained using a carbon fabric cathode. A similar result was recorded when the levels of the current density were recorded. On the basis of the results of the testing, we can conclude that the cathode with the Ni-Co catalyst was just as suitable for use in the L-MFC as the carbon cloth cathode. The Cu-B alloy had a slightly poorer performance parameters, and noticeably deteriorated after several days of the operation of the ML-MFC (cell voltage, current density).