Microbial Fuel Cell as a Bioelectrochemical Sensor of Nitrite Ions

: The deteriorating environmental quality requires a rapid in situ real-time monitoring of toxic compounds in environment including water and wastewater. One of the most toxic nitrogen-containing ions is nitrite ion, therefore, it is particularly important to ensure that nitrite ions are completely absent in surface and ground waters as well as in wastewater or, at least, their concentration does not exceed permissible levels. However, no selective ion electrode, which would enable continuous measurement of nitrite ion concentration in wastewater by bioelectrochemical sensor, is available. Microbial fuel cell (MFC)-based biosensor offers a sustainable low-cost alternative to the monitoring by periodic sampling for laboratory testing. It has been determined, that at low (0.01–0.1 mg · L − 1 ) and moderate (1.0–10 mg · L − 1 ) concentration of nitrite ions in anolyte-model wastewater, the voltage drop in MFC linearly depends on the logarithm of nitrite ion concentration of proving the potential of the application of MFC-based biosensor for the quantitative monitoring of nitrite ion concentration in wastewater and other surface water. Higher concentrations (100–1000 mg · L − 1 ) of nitrite ions in anolyte-model wastewater could not be accurately quantiﬁed due to a signiﬁcant drop in MFC voltage. In this case MFC can potentially serve as a bioelectrochemical early warning device for extremely high nitrite pollution. real-time monitoring of nitrite ion concentration in and/or the investigation of a dual-chamber microbial fuel cell employing Shewanella putrefaciens as a bioelectrochemical biosensor for continuous monitoring of nitrite ion concentration in artiﬁcial wastewater. feed to the MFC was restored. While anolyte-model wastewater containing medium and high concentration of nitrite ions (1.0–1000 mg · L − 1 ) was fed to the MFC, nitrite ions were completely leached from the MFC only with the ﬁfth sample. In all cases, the sixth sample was free of nitrite ions. This retention of nitrite ions in the anode half-cell most likely determines the nature of the voltage-time dependence curves when the voltage drop is very sudden at the start of the supply of the nitrite-contaminated anolyte-model wastewater, whereas the recovery of the voltage level is much slower after the supply of the nitrite-free anolyte to the anode half-cell is resumed.


Introduction
The growing human population and rapid development of industrial activities result in the increasing contamination of water [1]. Rapid recognition of toxic compounds and in situ real-time monitoring of their concentration are essential for water quality detection. Recently, microbial fuel cell (MFC)-based biosensors have drawn increasing interest due to their sustainability and cost-effectiveness [2].
In MFC, the energy stored in chemical bonds in organic compounds is converted to electrical energy through the catalytic reactions by microorganisms [3]. Electrons produced by the exoelectrogens from organic substrates by microbial metabolic pathways are transferred to the anode and flow to the cathode via an external circuit, containing a resistor or operated under a load, and are consumed at the cathode during the reduction of the oxidizing agent [4,5].
MFC is a simple and robust technology, enabling direct conversion of organics energy into electrical energy, leading to simultaneous electricity generation and waste reclamation [6] and its application areas have been expanding exponentially [7]. For several decades, MFCs have been extensively studied and applied for the energy-harvesting [8,9], wastewater treatment [1,10,11], recovery of valuable compounds [12][13][14], and as biosensors. MFC-based biosensor is one of the most promising technologies for toxicant determination Two-chamber MFC with the network of interdigitate channels and~32 mL volume of anode and cathode chambers, separated by a 10 × 10 cm size proton-exchange membrane (PEM) Nafion ® NRE-212 (Fuel Cell Store, College Station, TX, USA) was constructed and used to investigate operation performance of the MFC as nitrite ion biosensor. MFC was operated in batch mode. Polyacrylonitrile-based graphite felt (GF) AvCarb G200 (Fuel Cell Store, TX, USA) square sheets with the dimensions of 7 × 7 × 0.65 cm (length x width x thickness) were used as MFC anode and cathode. Before filling the MFC with anolyte and catholyte both GF electrodes were hydrophilized (wetted) with aqueous 10% (v/v) propan-2-ol solution and washed well with distilled water.
Anolyte, supplied from the anolyte reservoir, was circulated with peristaltic pump BS100-1AQ (Fluid Technology Co., Ltd., Baoding, China) at 12 mL·min −1 flow rate under anaerobic conditions provided by constant flow of nitrogen gas (purity 99.996 % (v/v)) at 100 mL·min −1 flow rate in the anode part of the system. Catholyte, supplied from separate catholyte reservoir, was circulated with peristaltic pump BS100-1AQ (Fluid Technology Co., Ltd., Baoding, China) at 30 mL·min −1 flow rate. Continuous aeration of catholyte was performed by membrane air pump Schego M2K3 (SCHEGO Schemel&Goetz GmbH&Co KG, Germany) at constant~200 mL·min −1 flow rate of the air. Concentration of DO in catholyte both at the inlet to MFC and outlet from MFC was monitored with luminescent/optical DO probe connected to the multimeter Hach HQ40D (Hach Company, Loveland, CO, USA).
The influence of nitrite ions on the performance of the MFC, i.e., voltage generated by MFC, was evaluated by suddenly feeding nitrite ion-contaminated anolyte-model wastewater instead of the nitrite-free anolyte-model wastewater. The concentration of nitrite ions in the contaminated anolyte-model wastewater was changed from 0.01 mg·L −1 to 1000 mg·L −1 NO 2 ions. The contaminated anolyte-model wastewater of the certain nitrite ion concen- Concentration of nitrite ions in stock NaNO 2 solutions and prepared nitrite ioncontaminated anolyte-model wastewater solutions both at the inlet to MFC and outlet from MFC was monitored by a standardized molecular absorption spectrometric method [30]. Before the molecular absorption spectrometric analysis, all samples of nitrite ion-contaminated anolyte-model wastewater solutions were filtered through a sterile syringe filter (surfactant-free cellulose acetate membrane, 28 mm diameter and 0.45 mm pore size (Corning ® , Kaiserslautern, Germany). Measurements of the absorption of visible light at 540 nm wavelength of the analyzed solutions were performed in glass cuvettes with light path length of 10 mm by spectrophotometer PerkinElmer ® Lambda 25 (PerkinElmer Inc., Waltham, MA, USA).

Electrochemical Measurements
In order to determine the optimal electric circuit load (resistance), at which the current generated by MFC was the highest at the lowest voltage, the operation of MFC under different electric circuit loads was investigated by directly connecting the potentiostatgalvanostat SP-150 (BioLogic ® , Seyssinet-Pariset, France) to the MFC. First, MFC was operating in an open circuit mode until inoculation of S. putrefaciens was complete and stable open circuit voltage value was achieved. Afterward, the resistance of electric circuit was gradually decreased from 32 kΩ to 15 Ω programmatically by controlling the current such that the ratio of MFC voltage/current kept constant.
The operation of MFC was investigated by connecting an optimal passive electric load to it, i.e., a voltage divider consisting of 920 Ω and 80 Ω resistors connected in series. The total electric load was 1000 Ω. The voltage variation over time was measured across the 80 Ω resistor by using data logger USB TC-08 (Pico ® Technology Ltd., St Neots, UK) connected to the personal computer for data collection. The actual values of voltage and current generated by each MFC were calculated by applying Ohm's law for a part of the circuit. All MFC operation tests and electrochemical measurements were performed at room temperature (22 ± 1 • C).

Inoculation and Electric Load Selection of MFC
The scheme of a setup of equipment used to investigate a MFC as a biosensor of nitrite ions is presented in Figure 1. At the initial stage, potentiostat-galvanostat SP-150 was connected to the MFC for the measurement of open circuit voltage and determination of the optimal electrical circuit load (resistance). When the further measurements were conducted, the voltage divider with optimal 1000 Ω resistor and data logger USB TC-08 were connected to the MFC instead of the potentiostat-galvanostat SP-150. connected to the MFC for the measurement of open circuit voltage and determination of the optimal electrical circuit load (resistance). When the further measurements were conducted, the voltage divider with optimal 1000 Ω resistor and data logger USB TC-08 were connected to the MFC instead of the potentiostat-galvanostat SP-150. After filling the MFC system with appropriate electrolytes, particular care was taken to ensure that the catholyte circulating in the cathode half-cell of the MFC was sufficiently aerated and that there was no DO in the anolyte-model wastewater circulating in the anode half-cell. The change in DO concentration over time was monitored in both electrolytes with luminescent/optical DO probe connected to the multimeter Hach HQ40D (Hach Company, USA). During these measurements, the flow rates of catholyte and anolytemodel wastewater through the respective MFC half-cells was 30 mL·min −1 and 12 mL·min −1 , respectively. Air was passing through the catholyte at a flow rate of ~200 mL·min −1 , and nitrogen gas was passing through the anolyte-model wastewater at a flow rate of ~100 mL·min −1 to ensure anaerobic conditions. As it has been determined previously [44], higher catholyte circulation rate and more intensive aeration ensure that the MFC cathodic process is not limiting. The maximum concentration of DO (7.2 mg·L −1 ) in the catholyte was achieved after approx. 15 min from the beginning of aeration. Meanwhile, complete deaeration of the anolyte-model wastewater to reach a concentration of DO <0.02 mg·L −1 took approx. 35 min ( Figure 2). After filling the MFC system with appropriate electrolytes, particular care was taken to ensure that the catholyte circulating in the cathode half-cell of the MFC was sufficiently aerated and that there was no DO in the anolyte-model wastewater circulating in the anode half-cell. The change in DO concentration over time was monitored in both electrolytes with luminescent/optical DO probe connected to the multimeter Hach HQ40D (Hach Company, USA). During these measurements, the flow rates of catholyte and anolyte-model wastewater through the respective MFC half-cells was 30 mL·min −1 and 12 mL·min −1 , respectively. Air was passing through the catholyte at a flow rate of~200 mL·min −1 , and nitrogen gas was passing through the anolyte-model wastewater at a flow rate of 100 mL·min −1 to ensure anaerobic conditions. As it has been determined previously [44], higher catholyte circulation rate and more intensive aeration ensure that the MFC cathodic process is not limiting. The maximum concentration of DO (7.2 mg·L −1 ) in the catholyte was achieved after approx. 15 min from the beginning of aeration. Meanwhile, complete deaeration of the anolyte-model wastewater to reach a concentration of DO <0.02 mg·L −1 took approx. 35 min ( Figure 2).
After assuring that the anolyte-model wastewater was completely deaerated, a suspension of S. putrefaciens bacteria (10 mL of suspension to 390 mL of anolyte) was added.  After stable MFC open circuit voltage was reached, MFC was periodically loaded with different electrical circuit resistance. The MFC was first loaded with a maximum re sistance of 32 kΩ, which was gradually reduced to 15 Ω. The duration of loading with different resistances was 15 min, followed by an open circuit period of 45 min. It had been observed that as the resistance of the external circuit connected to the MFC decreased, it voltage dropped steadily to almost 0 V (Figure 4a). This is typical for any electrochemica current source when the circuit resistance approaches 0 Ω (approaching a short circuit) Meanwhile, the current generated by the MFC increased steadily as the electrical circui load decreased (Figure 4b) until it reached a maximum average current of ~33 μA at an   After stable MFC open circuit voltage was reached, MFC was periodically loaded with different electrical circuit resistance. The MFC was first loaded with a maximum re sistance of 32 kΩ, which was gradually reduced to 15 Ω. The duration of loading with different resistances was 15 min, followed by an open circuit period of 45 min. It had been observed that as the resistance of the external circuit connected to the MFC decreased, its voltage dropped steadily to almost 0 V (Figure 4a). This is typical for any electrochemica current source when the circuit resistance approaches 0 Ω (approaching a short circuit) Meanwhile, the current generated by the MFC increased steadily as the electrical circui load decreased (Figure 4b) until it reached a maximum average current of ~33 μA at an external circuit resistance of 1 kΩ. By further reducing the resistance of the external circuit After stable MFC open circuit voltage was reached, MFC was periodically loaded with different electrical circuit resistance. The MFC was first loaded with a maximum resistance of 32 kΩ, which was gradually reduced to 15 Ω. The duration of loading with different resistances was 15 min, followed by an open circuit period of 45 min. It had been observed that as the resistance of the external circuit connected to the MFC decreased, its voltage dropped steadily to almost 0 V (Figure 4a). This is typical for any electrochemical current source when the circuit resistance approaches 0 Ω (approaching a short circuit). Meanwhile, the current generated by the MFC increased steadily as the electrical circuit load decreased (Figure 4b) until it reached a maximum average current of~33 µA at an external circuit resistance of 1 kΩ. By further reducing the resistance of the external circuit, the average current generated by the MFC decreased steadily, although slightly. In addition, current fluctuations were observed at resistance below 125 Ω, which were particularly significant at the lowest circuit resistance (15 Ω). Thus, for the further investigation of the MFC as a nitrite ion sensor, the passive resistance of 1 kΩ, at which MFC generated the highest average current at a voltage of approx. 25-30 mV, was chosen. the average current generated by the MFC decreased steadily, although slightly. In addition, current fluctuations were observed at resistance below 125 Ω, which were particularly significant at the lowest circuit resistance (15 Ω). Thus, for the further investigation of the MFC as a nitrite ion sensor, the passive resistance of 1 kΩ, at which MFC generated the highest average current at a voltage of approx. 25-30 mV, was chosen.

Influence of NO2 − Concentration on MFC Voltage
Total of 200 mL of contaminated anolyte-model wastewater solution of the certain nitrite ion concentration was prepared by taking a portion of the nitrite-free anolytemodel wastewater from an anolyte reservoir of the stabilized MFC and adding an appropriate volume of concentrated stock NaNO2 solution into it (Table 1). This solution of nitrite ion-contaminated anolyte-model wastewater was poured into a separate reservoir and deaerated with nitrogen gas at a flow rate of ~100 mL·min −1 analogously as nitritefree anolyte-model wastewater. A sudden drop in MFC voltage was observed after the

Influence of NO 2 − Concentration on MFC Voltage
Total of 200 mL of contaminated anolyte-model wastewater solution of the certain nitrite ion concentration was prepared by taking a portion of the nitrite-free anolyte-model wastewater from an anolyte reservoir of the stabilized MFC and adding an appropriate volume of concentrated stock NaNO 2 solution into it (Table 1). This solution of nitrite ion-contaminated anolyte-model wastewater was poured into a separate reservoir and deaerated with nitrogen gas at a flow rate of~100 mL·min −1 analogously as nitrite-free anolyte-model wastewater. A sudden drop in MFC voltage was observed after the abrupt replacement of the supply of nitrite-free anolyte-model wastewater to the anode half-cell of the MFC by the feed of contaminated anolyte-model wastewater at the same flow rate (12 mL·min −1 ) ( Figure 5). After passing 80 mL of nitrite ion-contaminated anolyte-model wastewater through the anode half-cell of the MFC, the supply of nitrite-free anolyte-model wastewater to the MFC was restored. The rise of the MFC voltage was observed at once with the subsequent recovery to the baseline. Duration of such a recovery of the generated voltage level was 30-60 min, when concentration of nitrite ions in the contaminated anolyte-model wastewater was low (0.01-0.1 mg·L −1 ) and moderate (1.0-10 mg·L −1 ). It is important to note, that the moderate concentration range covers the permissible levels of nitrite set by governing bodies worldwide [25,27,28], thus proving the applicability of such a MFC-based biosensor for accurate quantification (Figure 6) of nitrite ions at the edge of the allowed contamination level. However, when the concentration of nitrite ions in the contaminated anolyte was from 100 mg·L −1 to 1000 mg·L −1 , the MFC voltage did not reach the initial level even after more than 2 h of supplying nitrite-free anolyte to the MFC. It has been assumed that such a failure in recovery is due to the strong toxic effect of nitrite ions resulting in partial loss of the inoculated bacteria at the bioanode. abrupt replacement of the supply of nitrite-free anolyte-model wastewater to the anode half-cell of the MFC by the feed of contaminated anolyte-model wastewater at the same flow rate (12 mL·min −1 ) ( Figure 5). After passing 80 mL of nitrite ion-contaminated anolyte-model wastewater through the anode half-cell of the MFC, the supply of nitritefree anolyte-model wastewater to the MFC was restored. The rise of the MFC voltage was observed at once with the subsequent recovery to the baseline. Duration of such a recovery of the generated voltage level was 30-60 min, when concentration of nitrite ions in the contaminated anolyte-model wastewater was low (0.01-0.1 mg·L −1 ) and moderate (1.0-10 mg·L −1 ). It is important to note, that the moderate concentration range covers the permissible levels of nitrite set by governing bodies worldwide [25,27,28], thus proving the applicability of such a MFC-based biosensor for accurate quantification (Figure 6) of nitrite ions at the edge of the allowed contamination level. However, when the concentration of nitrite ions in the contaminated anolyte was from 100 mg·L −1 to 1000 mg·L −1 , the MFC voltage did not reach the initial level even after more than 2 h of supplying nitrite-free anolyte to the MFC. It has been assumed that such a failure in recovery is due to the strong toxic effect of nitrite ions resulting in partial loss of the inoculated bacteria at the bioanode. anolyte through the anodic half-cell, which is essentially in line with the extension of the analyzed anolyte-wastewater hydraulic retention time or by using multi-stage MFCs connected in series [20]. Therefore, the obtained results prove the possibility to use MFC as a nitrite ion biosensor for the accurate quantitative determination of nitrite ions in wastewater at lower concentrations or as an early warning system in the event of particularly high nitrite ion wastewater pollution.  As seen from the results presented in Figure 6, in the concentration range from 0.01 mg·L −1 to 10 mg·L −1 of nitrite ions in nitrite ion-contaminated anolyte-model wastewater, there is a clearly expressed linear dependence of MFC voltage drop on logarithm of nitrite ion concentration in a contaminated anolyte-model wastewater. The lowest detectable nitrite ion concentration can be defined by a limit of detection (LOD). LOD is calculated as the ratio of the standard error of the regression (σ) over the slope of the linear regression curve (m) [45]: The calculated LOD for nitrite ion concentration is~0.0028 mg·L −1 .
The linear dependence of MFC voltage drop on logarithm of nitrite ion concentration is also observed in the range of concentrations from 10 mg·L −1 to 1000 mg·L −1 . In this case, the MFC voltage drop values are very high (reaching 89% and 95%, respectively) and differ among themselves insignificantly, therefore, the slope of the curve is lower what determines the lower measurement accuracy. In practice, the concentration of nitrite ions in wastewater very rarely exceeds 1000 mg·L −1 [46]. Therefore, 1000 mg·L −1 was used as the highest concentration in the present study. It should be noted that the upper limit of the linear range is restricted by the capacity of the electrochemically active biofilm [47], thus, the anode area should be increased to provide a larger area for biofilm formation. It could be expected that the application of a MFC for the more accurate measurement of the higher concentration of nitrite ions could be achieved by reducing the flow rate of the anolyte through the anodic half-cell, which is essentially in line with the extension of the analyzed anolyte-wastewater hydraulic retention time or by using multi-stage MFCs connected in series [20]. Therefore, the obtained results prove the possibility to use MFC as a nitrite ion biosensor for the accurate quantitative determination of nitrite ions in wastewater at lower concentrations or as an early warning system in the event of particularly high nitrite ion wastewater pollution.

Change of Nitrite Ion Concentration in the Anolyte at the MFC Outlet
While investigating the effect of different concentrations of nitrite ions in contaminated anolyte-model wastewater that was fed to the MFC on its voltage response, samples of the anolyte at the outlet from the MFC anode half-cell (40 mL each) were collected simultaneously to determine the nitrite ion concentration in them. The diagrams of the distribution of nitrite ion concentration in individual samples of anolyte entering and leaving the MFC (Figure 7) show that when contaminated anolyte-wastewater with low concentration of nitrite ions (0.01-0.1 mg·L −1 ) was supplied to the MFC, nitrite ions were completely leached from the MFC with the fourth sample after the nitrite-free anolyte-model wastewater feed to the MFC was restored. While anolyte-model wastewater containing medium and high concentration of nitrite ions (1.0-1000 mg·L −1 ) was fed to the MFC, nitrite ions were completely leached from the MFC only with the fifth sample. In all cases, the sixth sample was free of nitrite ions. This retention of nitrite ions in the anode half-cell most likely determines the nature of the voltage-time dependence curves when the voltage drop is very sudden at the start of the supply of the nitrite-contaminated anolyte-model wastewater, whereas the recovery of the voltage level is much slower after the supply of the nitrite-free anolyte to the anode half-cell is resumed.
ultaneously to determine the nitrite ion concentration in them. The diagrams of the distribution of nitrite ion concentration in individual samples of anolyte entering and leaving the MFC (Figure 7) show that when contaminated anolyte-wastewater with low concentration of nitrite ions (0.01-0.1 mg·L −1 ) was supplied to the MFC, nitrite ions were completely leached from the MFC with the fourth sample after the nitrite-free anolyte-model wastewater feed to the MFC was restored. While anolyte-model wastewater containing medium and high concentration of nitrite ions (1.0-1000 mg·L −1 ) was fed to the MFC, nitrite ions were completely leached from the MFC only with the fifth sample. In all cases, the sixth sample was free of nitrite ions. This retention of nitrite ions in the anode half-cell most likely determines the nature of the voltage-time dependence curves when the voltage drop is very sudden at the start of the supply of the nitrite-contaminated anolytemodel wastewater, whereas the recovery of the voltage level is much slower after the supply of the nitrite-free anolyte to the anode half-cell is resumed.
(c) (f) After the nitrite ion concentration in all collected samples was determined, the total nitrite content in the MFC anolyte effluent (totally 6 samples; 40 mL volume each) was calculated and compared with the total nitrite ion content in nitrite ion-contaminated anolyte-model wastewater that was fed into the MFC (2 samples; 40 mL volume each) ( Table 2). In all cases, the total nitrite concentration in the anolyte-model wastewater samples collected at the outlet of the MFC anode half-cell was lower than that in the nitrite ion-contaminated anolyte-model wastewater samples collected at the inlet. A particularly significant decrease in nitrite ion concentration was observed in the case of anolyte-model wastewater containing low nitrite ion concentration (0.01-0.1 mg·L −1 ). It was approx. 25-30%, whereas the decrease in nitrite ion concentration in the samples of moderate concentration of nitrite ions (1.0-10 mg·L −1 ) was approx. 5-10%. This decrease in nitrite ion concentration can be attributed to the partial reduction of these ions to N 2 gas (denitrification) at the MFC bioanode [46,48]. The overall electrochemical processes, which take place on the surface of the anode and cathode during MFC operation, can be described by the following reaction equations [3]: Anodic process: CH 3 COO -(aq) + 2H 2 O(l) → 2CO 2 (g) + 7H + (aq) + 8eacetate (2) Cathodic process: O 2 (g) + 4H + (aq) + 4e -→ 2H 2 O(l) Since bacteria-exoelectrogens oxidize acetate ions at the bioanode and transfer electrons to the anode, nitrite ions entering the MFC anode half-cell compete for the electrons transferred by bacteria-exoelectrogens with the MFC anode. The reduction of nitrite ions at the bioanode when they gain electrons donated by bacteria-exoelectrogens can be described by the following electrochemical denitrification reaction equation [48]: It is very likely that "deficit" of electrons due to their transfer to nitrite ions by bacteriaexoelectrogens determines the increase of anode electrochemical potential and, therefore, the MFC voltage drop.

Statistical Analysis
All experiments were repeated at least three times, calculating the mean and standard deviation. The data were processed using Microsoft Office Excel 2016 software (Microsoft Corporation, Redmond, WA, USA). Statistical analysis was performed by using Student's t-test. The level of significance was set as p < 0.05.

Conclusions
In summary, it has been demonstrated, that at low (0.01-0.1 mg·L −1 ) and moderate concentration (1.0-10 mg·L −1 ) of nitrite ions in anolyte-model wastewater, the voltage drop in MFC linearly depends on the logarithm of concentration of nitrite ions proving the potential of the use of MFC-based biosensor for the quantitative monitoring of nitrite ion concentrations in wastewater and other surface water. However, higher concentrations (100-1000 mg·L −1 ) of nitrite ions in anolyte-model wastewater could not be accurately quantified due to a significant drop in MFC voltage. In this case MFC can potentially serve as a bioelectrochemical early warning device for extremely high nitrite pollution. MFCbased biosensor offers a sustainable low-cost real-time in situ alternative to the monitoring by periodic sampling for laboratory testing.