New application of microbial fuel cell-based biosensor for monitoring the quality of actual potato chips’ processing wastewater

  • Ahmed Y. Radeef
  • Zainab Z. IsmailEmail author


In this study, a dual-chamber microbial fuel cell (MFC) fed with actual potato chips’ processing wastewater (PCPW) was tested as a biosensor. The performance of MFC-based biosensor was evaluated in terms of the current measurement range, toxicity detection and sensitivity, and the operational stability. The results revealed that the MFC can simply be converted to an online biosensor unit to detect the harmful effect of suspended solids and acidic content in the actual PCPW on the anodic attached biofilm and the values of the generated current as well. A notable decrease in the current values was observed indicating the adverse effects of the harmful matters in the PCPW fed to the biosensor unit. The results proposed a competition between the harmful components and the favorable substrate in binding to the redox complex. An excellent fitting was obtained between the experimental and predicted results by \( I_{{K_{m} }} \) model with determination coefficient (R2) and mean-square-error values of 0.927 and 0.363, respectively. Additionally, a new approach was developed based on direct measurement of actual field data to replace the conventional statistical methods.


Microbial fuel cell Biosensor Modeling Potato chips processing wastewater Biodegradation 


Microbial fuel cell (MFC) is a device in which the electrochemically active microorganisms can convert chemical energy into electrical energy by oxidizing the organic matter at anaerobic conditions and transfer the released electrons to the anodic electrode. In the MFC, the electrical current is a direct linear measure for metabolic activity of the electrochemically active bacteria [1]. MFC can be installed as part of a sensory network. It can be converted to a biosensor and used for the detection of soluble pollutants. Converting the MFC to biosensors can be accomplished by operating the MFC with a most sensitive external load that immediately makes changes in the electrical current value if the anodic attached biofilm has been poisoned. As the generated current can simply be monitored online, therefore, the MFC can play a role as an online biosensor for the microorganisms’ activity [2]. The MFC converted to online biosensor is typically used to protect the quality of the flow such as process water, drinking water, or wastewater. In an online biosensor, the wastewater should be continuously fed to the anodic chamber of the MFC, and then, the sensor gives an alarm and detects the toxic matters entering the system [3, 4]. One of the main advantages of using MFC as an online biosensor is the possibility to obtain a self-powered biosensor. It has the advantage of a long-term stability using electroactive biofilms as sensing element with continuous growth of new bacterial cells to replace the old/dead cells which allows the device to operate in a very long functional lifetimes (> 5 years of continuous operation), and minimize the replacement of sensing elements [5, 6, 7]. The biosensors were first reported by Karube et al. [8] when they positioned a pure culture Clostridium butyricum as a sensing element to transfer electrons using hydrogen as electron acceptor. Liu et al. [9] used MFC-based biosensor for monitoring of anaerobic digestion process over 6-month operation period and the results revealed a good correlation with gas flow rate, pH, and COD removal. Also, it was found out that the pH values were affected on the biofilm negatively when it shifted away from its normal value Liu et al. [9]. Stein et al. [10] modified a bioelectrochemical model combined with enzyme inhibition kinetics which described the polarization curve of an MFC-based biosensor monitoring four types of toxicity. Shen et al. [11] proved the availability of fast monitoring of acidic toxicity using MFC-based toxicity biosensor. Stein et al. [12] provided polarization curves under both non-toxic and toxic conditions in MFC-based biosensor. The results confirmed the adverse effect of toxic components on the electrochemically active bacteria in the MFC. Chen et al. [13] suggested a cost-effective and rapid analytical method for detecting toxic substances using two-chamber MFC as a biosensor for in situ monitoring of p-nitrophenol. Jia et al. [14] developed MFC-based biosensor to monitor the operation of an up-flow anaerobic sludge blanket. The biosensors’ properties were affected because of the excessive suffocation and acidification of the electron transport. Zhou et al. [15] used the MFC-based biosensor for monitoring toxic carbon monoxide. Schievano et al. [16] monitored the volatile fatty acids (VFAs) concentration in four MFC-based biosensors fueled in batch mode with four kinds of feedstock including kitchen waste, cheese whey, fishery waste, and citrus pulp. Jiang et al. [17] improved a novel gas diffusion (GD)–biocathode MFC sensor which can be used directly for formaldehyde detection from 0.0005 to 0.005% in both anaerobic and aerobic water bodies. Yue et al. [18] suggested that the optimal external load for MFC biosensor varied with each type of toxicants.

Yet, to the authors’ knowledge, none of the previously published studies dealt with evaluating the performance of MFC-based biosensor fueled with real raw potato chips processing wastewater. The potato processing industry including potato chips manufacturing factories is one of the most worldwide spread food industries, which utilizes large volumes of fresh water in different subsequent operations. It contains high concentrations of total suspended solids (TSS) that must be removed and acidic or alkaline materials that should be subjected to neutralization prior to biological treatment to enhance the removal of high organic matters (carbohydrates) load resulting in high biochemical oxygen demand (BOD) and chemical oxygen demand (COD) [19, 20].

This study aimed to investigate for the first time the effects of actual raw potato chips’ processing wastewater (PCPW) on the anodic biofilm stabilization in a dual-chamber MFC-based biosensor. Unlike previous studies which concerned of single toxicant, the present study takes into consideration the effect of the dual inhibition of two toxicants; pH and TSS on the bioactivity of microorganisms. Also, the Butler–Volmer–Monod model suggested by Stein et al. [10] to convert MFC to a biosensor unit was developed and applied in this study.

This study was carried out after continuous operation for a relatively long-time duration (120 days) at a steady-state conditions regarding the biofilm activity and power generation. Then after, the pH adjustment unit and the suspended solids removal unit which proceeded the MFC system were eliminated to investigate the performance of MFC and the disturbance of the stabilized biofilm in the anode chamber of the MFC system in the absence of these two preliminary treatment units. Hence, the MFC-based biosensor models were developed and applied for the first time to estimate the changes occurred during the treatment of real raw PCPW in the MFC during the absence of pH adjustment and the suspended solid removal units.

Materials and methods

System configuration and operation

A horizontal dual-chamber MFC having an effective volume of 1900 mL for each chamber was designed and set up in this study. The anodic and cathodic chambers were separated by a cation exchange membrane (CEM) type CMI-7000. Graphite plain electrodes were used as the anodic and cathodic electrodes; each had a projected surface area of 128 cm2. Before introducing the biomass in the anodic section of MFC, the mixed bacterial cells were anaerobically enriched and acclimated using peptone-loaded MSM, following the procedure suggested by Huang et al. [21]. The Ag–AgCl reference electrode was introduced in the cathodic electrode upon operating the MFC as biosensor and the working electrode potential was recorded against the Ag–AgCl reference electrode. The anode chamber was continuously fed with actual potato chips’ processing wastewater (PCPW) via a peristaltic pump. The average concentrations of the major constituents in the raw PCPW were 7830, 2580, and 2930 mg/L for COD, TSS, and TDS at pH 5.2–5.8. Phosphate buffer saline consisted of 8.5 g/L NaCl, 1.91 g/L Na2HPO4, and 0.38 g/L KH2PO4 at pH 7.2 ± 0.2 was used as a catholyte and continuously sparged with an air pump at a rate of 10 mL/min. The MFC system preceded by holding-neutralization tank and a sedimentation tank was continuously operated for 4 months as a normal major MFC unit proceeded by the holding-neutralization tank and sedimentation tank as a primary treatment unit (Fig. 1a). Table 1 presents the average concentrations of the constituents in the PCPW after the primary treatment which included neutralization and sedimentation units for pH adjustment and TSS removal, respectively.
Fig. 1

MFC system: a at normal operation with primary treatment (pH adjustment and TSS removal); b MFC as biosensor (without primary treatment)

Table 1

Characterization of PCPW before and after primary treatment



Average concentrations

In the raw PCPW

After primary treatment before entering the MFC

After pH adjustment in the neutralization tank

After TSS removal in the sedimentation tank








7 ± 0.2

7 ± 0.2
















After 120 days of continues operation at a steady-state conditions, the MFC was converted to act as a biosensor. The transformation of MFC to biosensor requested: (1) elimination of the primary treatment units to keep the initial values of pH and TSS concentration without change, and (2) introduce the reference electrode Ag/AgCl in the cathodic section as given in Fig. 1b.

Analytical methods

The MFC system was operated at a fixed external load of 100Ω, and the voltage was continuously measured using a voltage data logger (model: Lascar EL-USB-3, USA). The voltage values were converted to current according to Ohm’s law: I = V/Rext, where I is the current (mA), V is the voltage (mV), and Rext is the external resistance (Ω). Polarization curves were obtained by changing the external resistances within a range from 60,000 to 5 Ω.

Description of biosensor model

Stein et al. [10] used the main equation of Butler–Volmer–Monod model as a starting point for the development of a biosensor model. This equation is a simple and easy mathematical formula used to describe the conditions of the substrate degradation in the anodic chamber.

This equation was presented by Hamelers et al. [22] as follows:
$$ I = I_{{\rm max} } \frac{{1 - e^{ - f \cdot \eta } }}{{k_{1} \cdot e^{ - (1 - \alpha )f \cdot \eta } + k_{2} \cdot e^{ - f \cdot \eta } + \left( {\frac{{k_{M} }}{S} + 1} \right)}}, $$
where η = overpotential (voltage loss) (mV), Imax = maximum current determined by maximum enzymatic rates of microorganisms (mA), Km = substrate affinity constant (mol/L), K1 = lumped parameter describes the ratio between biochemical and electrochemical rate constants, K2 = lumped parameter describes the forward over backward biochemical rate constants, and S = substrate concentration (mol/L):
$$ f = F /(R*T), $$
where F = Faraday’s constant (96,485 A/mol), T = temperature (K), and R = gas constant (8.314 J/mol K).

The presence of inhibiting components was not included or taken into account in this equation. Therefore, in this section, the model was extended by introducing an inhibition term to detect the toxicity effect. Then after, the sensitivity of the sensor for toxic compounds will be determined as a function of overpotential.

Description of MFC-based biosensor

The proposed model was developed to design MFC-based biosensor considering the microbial populations and their bioelectrochemical reactions in the anodic chamber. The model was applied for the MFC fed with actual PCPW and operated in a continuous mode. The following assumptions were considered:
  1. 1.

    The microorganisms’ activity depends on pH and temperature; therefore, pH and temperature should to be under control to maintain an accurate signal from the MFC-based biosensor.

  2. 2.

    Under non-toxic conditions, the overpotential must be controlled to get a stable baseline, because the microorganisms’ activity also depends on the energy obtained from the consumption of substrate.

The Butler–Volmer–Monod model (Eq. 1) was used as a starting point. It was assumed that the toxicity effect can be modeled and described by its effect on the kinetic reaction rates which presented by the electron transfer. The toxic components can influence several reaction rates depending on the mechanism of influence. An inhibition term [IT = (\( \frac{K_i}{K_i + X_i} \))] was involved in the four proposed models, in which:
$$ \begin{aligned} & K_i = {\text{Affinity}}\,{\text{constant}}\,({\text{mmol/L}}) \\ & X_i = {\text{Concentration}}\,{\text{of}}\,{\text{toxic}}\,{\text{component}}\,({\text{mmol/L}}). \\ \end{aligned} $$
A smaller Ki value means that the component is more toxic to the microorganisms.
  1. 1.
    The first model was symbolized as (Itox). This model assumed that the toxic components play their role as an irreversible inhibitor or as a noncompetitive inhibitor [23]. Thus, the model was arranged as follows:
    $$ I_{\text{tox}} = I_{{\rm max} } \frac{{1 - e^{ - f \cdot \eta } }}{{k_{1} \cdot e^{{ - \left( {1 - \alpha } \right)f \cdot \eta }} + k_{2} \cdot e^{ - f \cdot \eta } + \left( {\frac{{k_{M} }}{S} + 1} \right)}} \cdot \frac{K_i}{K_i + X_i}. $$
  2. 2.
    The relation between electrochemical and biochemical reaction rate which was described by k1 constant might change when the toxic components may play a role as an electron acceptor. In this case, the current decreases due to the fact that redox complex is not fully oxidized at the anode due to the existence of toxic components. This model is called (\( I_{{k_{1} }} \)) and calculated by:
    $$ I_{k_1} = I_{{\rm max} } \frac{{1 - e^{ - f \cdot \eta } }}{{k_{1} \cdot \left( {\frac{K_i + X_i}{K_i}} \right) \cdot e^{ - (1 - \alpha )f \cdot \eta } + k_{2} \cdot e^{ - f \cdot \eta } + \left( {\frac{{k_{M} }}{S} + 1} \right)}}. $$
  3. 3.
    The relationship between the backward and forward reaction rate constants of substrate oxidation was described by k2 constant. This relation can be changed when no products are formed anymore or much more products (CO2 and protons) are formed. When the toxic component is acidic, this relation also changes. Thus, the reduction of the forward reaction rate is caused by the increased concentration of protons (H+). This model is called (\( I_{{K_{2} }} \)) and is given by:
    $$ I_{{K_{2} }} = I_{{\rm max} } \frac{{1 - e^{ - f \cdot \eta } }}{{k_{1} \cdot e^{{ - \left( {1 - \alpha } \right)f \cdot \eta }} + k_{2} \cdot \left( {\frac{K_i + X_i}{K_i}} \right) \cdot e^{ - f \cdot \eta } + \left( {\frac{{k_{M} }}{S} + 1} \right)}}. $$
  4. 4.
    When there is a competition between toxic component and substrate to bind to the redox complex, the substrate affinity constant Km might change [22]. This model is called (\( I_{{K_{m}}}\)) and calculated as follows:
    $$ I_{{K_{m}}} = I_{{\rm max} } \frac{{1 - e^{ - f \cdot \eta } }}{{k_{1} \cdot e^{{ - \left( {1 - \alpha } \right)f \cdot \eta }} + k_{2} \cdot e^{ - f \cdot \eta } + \left( {\frac{{k_{M} }}{S} \cdot \left( {\frac{K_i + X_i}{K_i}} \right) + 1} \right)}}. $$

Sensitivity for toxic components

At the maximum change of the generated electrical current with equivalent change in the concentration of toxic components, the MFC-based biosensor will be most sensitive for toxic compounds. Therefore, the derivative of the current (as a function of overpotential η) with respect to the concentration of toxic compound should be at the maximum. The overpotential can be calculated using the following equation:
$$ \eta^{*} = \mathop {{ \arg }{\rm max} }\limits_{{\eta \in \left[ {0,700} \right]}} \frac{{{\text{d}}I\left( \eta \right)}}{{{\text{d}}X_i}}, $$
where \( \frac{{{\text{d}}I\left( \eta \right)}}{{{\text{d}}X_i}} \) = the sensitivity of the current for toxic compounds. \( \eta^{*} \) = overpotential value within the interval [0, 700] mV, which gives maximum sensitivity.

This objective can be achieved by plotting the derivative of current to toxicity \( (\frac{{{\text{d}}I}}{{{\text{d}}X_i}}) \) as a function of overpotential, η.

Development of MFC-based biosensor

As given in Eqs. 25, the inhibition term (\( \frac{K_i}{K_i + X_i} \)) was involved in the four proposed models, in which Ki usually estimated by statistical methods for the best fit and Xi can be estimated by direct measurement of a specific toxicant in the bulk solution before and after the exposure to toxicity. This method is only suitable for laboratory synthetic wastewater and it is difficult to be applied for real wastewaters which may contain more than one type of toxic matters. Therefore, in the present study, the inhibition term was calculated and estimated by new approaches based on the values which can be measured directly in real wastewater as follows:
  1. 1.
    The inhibition term was extended to contain at least one parameter that can be measured directly in wastewater. Therefore, in case of adding the difference in the measured concentration of TDS between influent and effluent (∆TDS) to the inhibition term, a new term named (IT1) can be obtained as follows:
    $$ {\text{IT}}1 = \frac{K_i}{{K_i + X_i + \Delta {\text{TDS}}}}. $$
  2. 2.
    The inhibition term was calculated as a fraction of the ratio between (∆TDS) and the measured TDS concentration in the influent stream (Influent TDS); this term named (IT2) and calculated as shown below:
    $$ {\text{IT}}1 = \frac{K_i}{{K_i + X_i + \Delta {\text{TDS}}}}. $$
  3. 3.
    The toxicity can be determined as inhibition ratio (IR) by measuring the changes of the maximum output voltage in MFC, when inhibition ratio represents the ratio between maximum voltage output at toxicity-free condition and maximum voltage output a toxic condition as given below [24, 25]:
    $$ {\text{IR}}\;(\% ) = 100 \cdot \frac{{(V_{\text{clean}} - V_{\text{toxic}} )}}{{V_{\text{clean}} }}, $$
    where Vclean is the maximum voltage output resulted from clean anolyte without toxic matters and Vtoxic is the maximum voltage output resulted from toxic matters-loaded anolyte.
MATLAB software version (R2017b) was used to solve a set of equations from Eqs. (1) to (6). The useful parameters as well as the constants related to the performance of biosensor are summarized in Table 2.
Table 2

Assumed, measured, and estimated parameters in addition to the physical constants used in models







Faraday’s constant





Gas constant


J/mol K








The maximum current





The overpotential





A lumped parameter describing the ratio between biochemical and electrochemical rate constants





A lumped parameter describing the forward over backward biochemical rate constants





Substrate affinity constant





Substrate concentration





the affinity constant





the concentration of toxic component





Transfer coefficient










The change in total dissolved solids concentrations





Inhibition term





Inhibition term include ∆TDS





Inhibition term based on TDS concentration





Inhibition ratio




These models are based on enzyme inhibition kinetics, while combined effects may also play a role. In practice and reality, the biomass inhibition may not be due to single type of toxicants. To the authors’ knowledge, combined inhibition was not taken into full consideration in the previously reported studies [27]. Most recently, Zhao et al. [28] investigated the use of sediment microbial fuel cell (SMFC) as a biosensor for in situ real-time monitoring of Cr(VI), in which the organic substrate was oxidized in the anode and Cr(VI) is reduced at the cathode simultaneously. The system showed high specificity for Cr(VI) which was considered as the solely toxicant, as other co-existing ions such as Cu2+, Zn2+, and Pb2+ did not interfere with Cr(VI) detection. Hence, the current study considered the effect of dual toxicants; pH and TSS which were commonly found in real samples of PCPW.

Results and discussion

Performance of MFC-based biosensor

To convert the MFC to a biosensor unit, the MFC should be operated after achieving the steady-state conditions at the optimum value of overpotential. In this study, the proper value for the inhibition term was equal to 0.065 which gave the smallest mean-square error (MSE) within the range of the inhibition term of (0–1) to the original curve (normal condition with primary treatment of PCPW). The affinity constant (Ki) and the concentration of toxic component (Xi) were found to be 0.7 and 10 mmol/L, respectively. These values were estimated by trial-and-error method due to complexity of the actual PCPW. Figure 2 illustrates the plot of the current values when MFC run at steady-state conditions and the expected changes for each model when considering the existence of toxic components in the PCPW.
Fig. 2

Original current curve and expected curves for models; Itox, \( I_{{k_{1} }} \), \( I_{{K_{2} }} \), and \( I_{{K_{m} }} \) for MFC biosensor fed with PCPW without primary treatments

MFC sensitivity

As mentioned above, the MFC should be operated at the optimum value of overpotential after achieving the steady-state conditions which makes the MFC more sensitive to the toxic components. For example, when the toxic components affect the cell, the most sensitive setting must be at the overpotential value which gives the maximum change of the electrical current for any toxic concentration. The derivative of current to toxicity, dI/dXi, is plotted versus overpotential. At maximum value of dI/dXi, the sensor is the most sensitive at the corresponding overpotential. The optimal setup values of overpotential with corresponding maximum current changes are given in Fig. 3 and Table 3.
Fig. 3

The change in current value (Sensitivity) for models Itox, \( I_{{k_{1} }} \), \( I_{{K_{2} }} \), and \( I_{{K_{m} }} \) when toxic compounds affect the biomass

Table 3

Maximum changes of current due to the toxicity and their corresponding overpotentials




\( I_{{k_{1} }} \)

\( I_{{K_{2} }} \)

\( I_{{K_{m} }} \)













In practical systems, this method is considered as an efficient approach to keep higher sensitivity for each model using four electrodes in one anodic chamber in a small lab-scale MFC and set each one to detect any type of toxic compounds which may minimize the COD removal efficiency. For example, the effect of toxic compounds can be estimated by model \( I_{{k_{1} }}\) at any time based on the ratio between biochemical and electrochemical rate constants and so on for other models.

Detection of PCPW toxicity

To detect the type of toxic compounds in the actual PCPW and their effects on the four biosensor models, the MFC-based biosensor was fed with raw PCPW without pH adjustment and suspended solid removal. The current values were plotted versus the overpotential values and fitted with the four models to estimate the factors which may adversely affect the biomass activity. The experimental and predicted results for MFC as a biosensor are given in Fig. 4.
Fig. 4

Experimental data fitted with predicted results for models Itox, \( I_{{K_{1} }} \), \( I_{{K_{2} }}\), and \( I_{{K_{m} }} \) for MFC biosensor fed with real PCPW without primary treatment based on (IT)

The data of the polarization curves presented in Fig. 4 were collected and recorded, while the biofilm was still exposing to the toxicants; low pH and high concentration of TSS.

However, in spite of the toxicity shock, it exhibited a stable electrical current signal. Than after, when the primary treatment was re-introduced to the MFC system and re-functioned again, it took a period of 3–5 days for the biofilm to restore its normal function before exposing to the toxic shock in the system after eliminating the primary treatment units represented by a holding-neutralization and the sedimentation tanks.

Table 4 summarizes the mean-square error (MSE) and the coefficient of determination (R2), which were considered to assess the goodness of the models fitting.
Table 4

Mean-square error (MSE) and coefficient of determination (R2) values for the models Itox, \( I_{{k_{1} }} \), \( I_{{K_{2} }} \), and \( I_{{K_{m} }} \) based on (IT)



\( I_{{k_{1} }} \)

\( I_{{K_{2} }} \)

\( I_{{K_{m} }} \)











In the present study, the major pollutants in the real PCPW were COD, TSS, and acidic pH. When the raw PCPW directly fed to the MFC without primary treatment, the activity of the biomass inhibited and the polarization curve fit with the (\( I_{{K_{m}}}\)) curve, indicating that the existence of TSS and acidic pH affected the substrate affinity constant (km) which refers to the substrate influence on the biochemical conversion. This observation is in a good agreement with the fact that a competition happened between toxic component and substrate to restrict the redox complex and affect the \( I_{{K_{m} }} \), which means a disturbance occurred and prevented the redox component to transfer electrons correctly to the anode surface [23]. However, when the pH and TSS concentrations simultaneously changed in the anodic section, the variation in the value of the potential was relatively higher compared to the potential change when the pH is the sole toxicant as reported by Stein et al. [3] and Chen et al. [13]. A previous study reported by Stein et al. [12] suggested that other toxicants including nickel, bentazon, and sodium dodecyl sulfate (SDS) well fitted with model (Itox), whereby ferricyanide was best fitted to model (\( I_{{K_{2} }} \)) or (Itox), depending on its initial concentration.

Due to complexity of the real wastewater composition, a new approach for estimating the inhibition term was developed in this study based on real-field measured data instead of the conventional statistical method. For \( I_{{k_{1}}}\) estimation, the difference in total dissolved solids concentrations (ΔTDS) was introduced into the inhibition term, whereby the calculation of \( I_{{k_{2}}}\) was totally based on real-field measured data represented by the ΔTDS divided by TDS values in the influent. On the other hand, IR values were assessed using the following formula:
$$ {\text{IR}} \, \left( \% \right) = 100* \frac{{\left( {V_{\text{clean}} - V_{\text{toxic}} } \right)}}{{V_{\text{clean}} }}, $$
where Vclean is the cell-voltage value at the toxicity absence; Vtoxic is the cell-voltage value at the toxicity existence.
The results revealed that the new calculated values of IT1, IT2, and IR were significantly similar to \( I_{{k_{m} }} \) and the statistically estimated IT value. Figures 5, 6, and 7 and Table 5 indicated an excellent fitting of real data with \( I_{{K_{m} }} \).
Fig. 5

Experimental and predicted data for models Itox, \( I_{{K_{1}}}\), \( I_{{K_{2}}}\), and \( I_{{K_{m} }} \) for MFC biosensor fed with real PCPW without primary treatment based on (IT1)

Fig. 6

Experimental and predicted data for models Itox, \( I_{{K_{1} }} \), \( I_{{K_{2} }} \), and \( I_{{K_{m} }} \) for MFC biosensor fed with real PCPW without primary treatment based on (IT2)

Fig. 7

Experimental and predicted data for models Itox, \( I_{{K_{1} }} \), \( I_{{K_{2} }} \), and \( I_{{K_{m} }} \) for MFC biosensor fed with real PCPW without primary treatment based on (IR)

Table 5

Mean-square error (MSE) and coefficient of determination (R2) values for the models Itox, \( I_{{k_{1} }} \), \( I_{{K_{2} }} \), and \( I_{{K_{m} }} \) based on IT1, IT2, and IR

Inhibition term



\( I_{{k_{1} }} \)

\( I_{{K_{2} }} \)

\( I_{{K_{m} }} \)


































These results demonstrated that for IT1 calculation, the value of R2 increased and MSE decreased indicating the validity of introducing ΔTDS values in the estimation of inhibition term. However, for IT2, the R2 increased with very slight increase in MSE value, whereby similar R2 with decreased MSE were observed for IR, indicating its potential.

Accordingly, these findings proved that replacing the conventional statistical approach for IT calculation by real-field measured data to; (1) develop the IT to IT1, and (2) introduce the IT2 and IR for inhibition term estimation, was favorable and significantly increased the results’ accuracy.


This study investigated, for the first time, the feasibility of converting a long-term operated MFC to a biosensor unit to detect the adverse effects of eliminating the primary treatment units for the raw PCPW before entering the MFC. Results demonstrated that MFC can be converted to biosensor to detect the harmful effect of suspended solids and acidic content in PCPW on the biomass activity. An obvious decrease in current values was observed when harmful substances enter the biosensor. The results revealed that a competition between the toxic components and the substrate to bind to the redox complex was occurred by excellent fitting of experimental data with curves of \( I_{{K_{m} }} \) model with values of R2 and MSE equal 0.927 and 0.363, respectively. The development of biosensor method depending on directly measured values in the real field showed a significant similarity with the conventional statistical methods. This unique biosensor could be used in situ to protect the whole system without using a costly sensor.



The authors sincerely acknowledge the staff of the Salah Al-din Bakery & Pastry factory in Tikrit city, Iraq for being helpful and supportive in providing the actual samples of wastewater. Also, the authors are thankful to the Department of Environmental Engineering at University of Baghdad for the technical support.


  1. 1.
    Logan BE, Hamelers B, Rozendal R, et al. Microbial fuel cells: methodology and technology. Environ Sci Technol. 2006;40:5181–92.CrossRefGoogle Scholar
  2. 2.
    Feng Y, Kayode O Jr, Harper WF. Using microbial fuel cell output metrics and nonlinear modeling techniques for smart biosensing. Sci Total Environ. 2013;449:223–8.CrossRefGoogle Scholar
  3. 3.
    Stein NE, Hamelers HV, Buisman CN. Stabilizing the baseline current of a microbial fuel cell-based biosensor through overpotential control under non-toxic conditions. Bioelectrochemistry. 2010;78:87–91.CrossRefGoogle Scholar
  4. 4.
    Lvars-Barceló F, Zuliani A, Fallah M, et al. Novel applications of microbial fuel cells in sensors and biosensors. Appl Sci. 2018;8:1184–99.CrossRefGoogle Scholar
  5. 5.
    Gil GC, Chang IS, Kim BH, et al. Operational parameters affecting the performance of a mediator-less microbial fuel cell. Biosens Bioelectron. 2003;18:327–34.CrossRefGoogle Scholar
  6. 6.
    Chang IS, Moon H, Jang JK, et al. Improvement of a microbial fuel cell performance as a BOD sensor using respiratory inhibitors. Biosens Bioelectron. 2005;20:1856–9.CrossRefGoogle Scholar
  7. 7.
    Grattier M, Minteer SD. Self-powered biosensors. ACS Sens. 2017;5:44–53.Google Scholar
  8. 8.
    Karube I, Matsunaga T, Mitsuda S, et al. Microbial electrode BOD sensors. Biotechnol Bioeng. 1977;19:1535–47.CrossRefGoogle Scholar
  9. 9.
    Liu Z, Liu J, Zhang S, et al. Microbial fuel cell based biosensor for in situ monitoring of anaerobic digestion process. Bioresour Technol. 2011;102:10221–9.CrossRefGoogle Scholar
  10. 10.
    Stein NE, Keesman KJ, Hamelers HV, et al. Kinetic models for detection of toxicity in a microbial fuel cell based biosensor. Biosens Bioelectron. 2011;26:3115–20.CrossRefGoogle Scholar
  11. 11.
    Shen YJ, Lefebvre O, Tan Z, et al. Microbial fuel-cell-based toxicity sensor for fast monitoring of acidic toxicity. Water Sci Technol. 2012;65:1223–8.CrossRefGoogle Scholar
  12. 12.
    Stein NK, Hamelers HVM, Straten GV, et al. Effect of toxic components on microbial fuel cell-polarization curves and estimation of the type of toxic inhibition. Biosensors. 2012;2:255–68.CrossRefGoogle Scholar
  13. 13.
    Chen Z, Niu Y, Zhao S, et al. A novel biosensor for p-nitrophenol based on an aerobic anode microbial fuel cell. Biosens Bioelectron. 2016;85:860–8.CrossRefGoogle Scholar
  14. 14.
    Jia H, Yang G, Ngo HH, et al. Enhancing simultaneous response and amplification of biosensor in microbial fuel cell-based upflow anaerobic sludge bed reactor supplemented with zero-valent iron. Chem Eng J. 2017;327:1117–27.CrossRefGoogle Scholar
  15. 15.
    Zhou S, Huang S, Li Y, et al. Microbial fuel cell-based biosensor for toxic carbon monoxide monitoring. Talanta. 2018;186:368–71.CrossRefGoogle Scholar
  16. 16.
    Schievano A, Colombo A, Cossettini A, et al. Single-chamber microbial fuel cells as on-line shock-sensors for volatile fatty acids in anaerobic digesters. Waste Manage. 2018;71:785–91.CrossRefGoogle Scholar
  17. 17.
    Jiang Y, Liang P, Huang X, et al. A novel microbial fuel cell sensor with a gas diffusion biocathode sensing element for water and air quality monitoring. Chemosphere. 2018;203:21–5.CrossRefGoogle Scholar
  18. 18.
    Yi Y, Xie B, Zhao T, et al. Effect of external resistance on the sensitivity of microbial fuel cell biosensor for detection of different types of pollutants. Bioelectrochemistry. 2019;125:71–8.CrossRefGoogle Scholar
  19. 19.
    Malladi B, Ingham SC. Thermophilic aerobic treatment of potato processing wastewater. World J Microbiol Biotechnol. 1993;31:426–8.Google Scholar
  20. 20.
    Davis M, Cornwell D. Introduction to environmental engineering. New York: McGraw-Hill; 1999.Google Scholar
  21. 21.
    Huang LP, Jin B, Lant P. Direct fermentation of potato starch wastewater to lactic acid by Rhizopus oryzae and Rhizopus arrhizus. Bioprocess Biosyst Eng. 2005;27:229–38.CrossRefGoogle Scholar
  22. 22.
    Hamelers HV, Ter Heijne A, Stein N, et al. Butler–Volmer–Monod model for describing bio-anode polarization curves. Bioresour Technol. 2011;102:381–7.CrossRefGoogle Scholar
  23. 23.
    Bisswanger H. Enzyme kinetics principles and methods. Hoboken: Wiley-VCH Verlang GmbH; 2008.CrossRefGoogle Scholar
  24. 24.
    Kim M, Hyun MS, Gadd GM, et al. A novel biomonitoring system using microbial fuel cells. J Environ Monit. 2007;9:1323–8.CrossRefGoogle Scholar
  25. 25.
    Yu D, Bai L, Zhai J, et al. Toxicity detection in water containing heavy metal ions with a self-powered microbial fuel cell-based biosensor. Talanta. 2017;168:210–6.CrossRefGoogle Scholar
  26. 26.
    Wang SF, Chen T, Zhang ZL, et al. Direct electrochemistry and electrocatalysis of heme proteins entrapped in agarose hydrogel films in room-temperature ionic liquids. Langmuir. 2005;21:9260–6.CrossRefGoogle Scholar
  27. 27.
    Zhou T, Han H, Liu P, et al. Microbial fuels cell-based biosensor for toxicity detection: A Review. Sensors. 2017;17:2230.CrossRefGoogle Scholar
  28. 28.
    Zhao S, Liu P, Niu Y, et al. A novel early warning system based on a sediment microbial fuel cell for in situ and real time hexavalent chromium detection in industrial wastewater. Sensors. 2018;18:642.CrossRefGoogle Scholar

Copyright information

© Zhejiang University Press 2019

Authors and Affiliations

  1. 1.Department of Environmental EngineeringUniversity of BaghdadBaghdadIraq

Personalised recommendations