A novel All‐in‐One electrolysis electrode and bioreactor enable better study of electrochemical effects and electricity‐aided bioprocesses

Abstract

An autoclavable All‐in‐One electrolysis electrode in a rod shape assembly is developed as a new tool for bioelectrochemical systems and electricity‐aided bioprocesses. It can replace the classic two‐chamber bioelectrochemical system for electrolysis reactions, be inserted into conventional bioreactors and is easily adaptable as electrocatalytic surface or generator of super‐fine bubbles (H₂ and O₂) for bioconversion processes. Whereas the bioreactor itself functions as the working electrode chamber, a well‐integrated inner counter electrode chamber enables water electrolysis without the normally encountered undesired ion‐transfer effect. The efficiencies of the electrode are characterized and its advantages and usefulness compared to the classic H‐Cell bioelectrochemical system (BES) are demonstrated with glycerol fermentations by Clostridium pasteurianum DSM 525.

Abbreviations

BES

bioelectrochemical system

PDO

1,3‐propanediol

RCM

reinforced clostridia medium

1. Introduction

Electrobiotechnology (also called bioelectrochemistry or bioelectrotechnology) is gaining more and more importance for electricity‐driven bioproduction of chemicals and fuels 1, 2, 3. Even bioconversion of CO₂ to chemicals is now conceivable with extra accessible reductive power directly or indirectly from electricity 4. Especially electrochemically affected microbial processes have the potential to transfer classical chemical production routes using the support of excess regenerative energy 5.

In lab scale some successes in electrochemically affected fermentations have been reported 6, 7. To build on these progresses an effective implementation of electrochemical support in industrial fermentation scale is desirable. In general, all microbial electrochemical processes need at least two electrodes in the system that are inert against the culture medium, in many cases also autoclavable and structurally stable. The working electrode appears in the center of the whole process to provide or take electrons from the system 8.

Usually the counter electrode has no direct effect on the biosynthesis process but is needed to maintain the electric circuit. Often unused or toxic byproducts are formed at the counter electrode that makes a separation of the working and counter electrodes necessary. Hence an ion selective membrane is implemented to hold back undesired molecules or bacteria and maintain redox conditions. These described features led to a two chamber reactor that is often constructed as a H‐Cell system with an anode and a cathode chamber, respectively 9. For basic research in small scales such H‐Cell systems are feasible and often used 10.

However, because of the lateral tubular connection it is obvious that a stirred H‐shaped bioreactor possess poorly mixed zones next to the separator. Combined with unused volume for the counter electrode such a system is hardly up‐scalable for industrial purposes.

Furthermore, these two chamber reactor systems are primarily custom‐made (Fig. 2A). Thus experimental conditions and the final results from different groups are hardly comparable 11.

Figure 2.

H‐Cell setup (A) compared to the All‐in‐One electrolysis electrode implemented into a bioreactor filled with medium (B). C. pasteurianum DSM 525 fermentation with the All‐in‐One electrolysis electrode producing hydrogen (C). C. pasteurianum DSM 525 fermentation without the electrode (D).

In microbial fuel cell research it is already known that H‐shaped reactors are of disadvantage because of low power densities caused by long electrode differences and small membrane areas 12. An interesting step toward a standardized bioelectrochemical system was recently made by the group of Harnisch who constructed an expansion kit for bioreactors to perform microbial bio‐electrosynthesis 13. Because of the usually not entirely selective separation between two large chambers, undesired ion transfer and substrate/product diffusion could still interfere with the process 14, 15. As a consequence stable fermentation conditions such as constant salt and nutrient concentrations are not easy to achieve in these reactor systems. For a constant pH, an acidic supply is required in addition to alkalic addition. Additionally the selective membrane is often expensive, not autoclavable (e.g. not to be heated above 100°C according to recommendation of most manufacturers) and sensitive to fouling that consequently makes long‐term operation and scale‐up difficult.

By removing the strict membrane‐based electrode separation, reference electrode, and arranging the cathode and anode in a useful way, Giddings et al. simplified a bioelectrochemical system (BES) system and made an interesting step toward an up‐scalable BES 16. However, a separation‐free system could lead to problems for anaerobic processes due to the distribution of oxygen or oxidized products from the anode. Hu et al. 17 developed a tubular electrode assembly for testing new electrode materials supporting mixed culture biofilms in small reaction vessels. It was demonstrated that by reducing the space between the electrodes a decrease of the internal resistance can be achieved and thus a better efficiency was observed. This assembly itself was used as a reaction chamber and its use for large‐scale reactor is questionable.

In this work, we describe the design and characterization of a simple and new All‐in‐One electrolysis electrode and corresponding BES that can well overcome the major problems mentioned above. Glycerol fermentations with Clostridium pasteurianum DSM 525 were used for the examination and comparison of the new All‐in‐One electrolysis BES with the classic H‐Cell BES. Glycerol fermentation in BES has been studied by several groups with different organisms 6, 7 in H‐Cell BES. Of particular notice is the study of Choi et al. (2014) for glycerol fermentation using C. pasteurianum DSM 525 in a conventional H‐Cell bioreactor system. These authors postulated a direct electron transfer from the cathode to the microorganism based on an observed slight metabolic shift in C. pasteurianum toward production of NADH‐consuming metabolites such as 1,3‐propanediol (PDO). This was one of the very few reports for a direct electron transfer to a microorganism in a fermentation process and raised some controversies. We set out to revisit the glycerol fermentation with C. pasteurianum DSM 525 in a conventional H‐Cell BES and compare it with the new BES with our All‐in‐One electrolysis electrode under high potentials producing hydrogen.

2. Materials and methods

2.1. Principle and construction of the All‐in‐One electrolysis electrode BES

The new electrode for electricity‐aided fermentations was designed by integrating all requirements mentioned above into one electrode arrangement that can be standardized and is suitable for typical bioreactors like stirred‐tank reactors as detailed in Fig. 1. An electrochemical experiment can be easily started by connecting the electrode contacts (3, 4) to a voltage source to apply a desired voltage or current between the outer working (9) and inner counter electrode (11). The outer surface working electrode faces toward the medium and produces the desired gas (H₂ or O₂ depending on the polarization). Superfine bubbles can form on the electrode surface during electrolysis that are then well distributed in a stirred bioreactor. The second gas produced at the counter electrode rod rising up through the inner gas channel (7) can be collected at the lateral outlet (5) and could be used for a parallel process.

Figure 1.

Construction and sketch of the All‐in‐One electrolysis electrode.

The ceramic separator (10) prevents mixing of the gases and diffusion of microbial cells but allows liquid to enter into the counter electrode channel (12) and to ensure a refill of the electrolytically consumed water. The crucial size difference of the inner counter electrode channel and the fermentation chamber (bioreactor) allows electrochemical processes without strongly affecting ion transfer or pH value of the bioreactor. After a short equilibrium phase mainly a H₃O⁺ /OH⁻ charge transfer proceeds.

An exact potential at the working electrode can be either calculated over the resistance between the working and counter electrodes or adjusted by implementing a reference electrode in the fermenter.

2.2. Characterization of the All‐in‐One electrolysis electrode in a bioreactor without cells

The medium used was the usual C. pasteurianum DSM 525 medium according to Biebl (2001) without the carbon source. The bioreactor used was a 2 L foil reactor (Bioengineering AG) (Fig. 2B–D). The bioreactor was filled up to 1.5 L, stirred at 350 rpm, controlled at 35°C by heating and had a preadjusted pH of 6. The All‐in‐One electrolysis electrode was implemented with a 74.8 cm² working grid electrode and a 6.7 cm² inner counter rod electrode made of platinized titan. A Gamry Interface 1000 potentiostat was connected to control the electrochemical parameters. The reactor was sparged with nitrogen to achieve anaerobic conditions and at a constant flow to carry the produced hydrogen. Two experiments with a stepwise current increase for each 60 min and one longer term experiment at 400 mA for 1270 min as described below were done. The energetic (η_energetic) and Faraday efficiency (η_faraday) of the All‐in‐One electrolysis electrode were calculated according to Eqs. (1) and (2):

$${\eta }_{\mathrm{energetic}}=\frac{{V_H}_2\times H_{s{H}_2}}{U\times I\times \mathrm{\Delta }t}$$ (1)

$${\eta }_{\mathrm{faraday}}=\frac{z_{H_2}\times F\times p\times V_{H_2}}{I\times \mathrm{\Delta }t\times R\times T}$$ (2)

Where ${V_H}_2$= hydrogen volume [m³], $H_{s{H}_2}$= heat of combustion for hydrogen [J/m³], $U$ = potential [V], I = current [A], t = time [s], F = Faraday constant [C/mol], p = atmospheric pressure [Pa], R = gas constant [J/(K mol)], T = temperature [K], and $z_{H_2}$= charge number of H₂.

The real H₂ production rate measured in the longer experiment at 400 mA was described with Eq. (3) while the theoretical rate was calculated with Eq. (4). The total amount of H₂ ($n_{H_2}$) was calculated by integrating ${Q_{H_2}}^{\ast }$using Eq. (5).

$${Q_{H_2}}^{\ast }=\frac{{Q_{H_2}}^{max}\times t}{b+t}$$ (3)

$$Q_{H_2}=\frac{I}{F\times z_{H_2}}$$ (4)

$$n_{H_2}=\underset{0}{\overset{t}{\int }}{Q_{H_2}}^{\ast }dt$$ (5)

Where, t = time [s], F = Faraday constant [C/mol], and $z_{H_2}$= charge number of H₂.

2.3. Glycerol fermentation with C. pasteurianum DSM 525 in H‐Cell and e‐bioioreactor with the All‐in‐One electrolysis electrode

For the glycerol fermentations with C. pasteurianum DSMZ 525 two precultures were made.

The first one (50 mL working volume in 100 mL anaerobic bottles) with reinforced clostridia medium (RCM) was inoculated with 2 mL directly thawed cryo stock culture. After 24 h incubation at 35°C 2 mL of the RCM culture were transferred to 100 mL anaerobic bottles with 50 mL modified Biebl Medium and incubated for 24 h at 35°C as well. This medium was previously enriched with a FeSO₄ · 7H₂O solution to a concentration of 5 mg/L. The cryo stocks were made by freezing 2 mL aliquots of this 24 h old C. pasteurianum culture in 20% glycerol at –80°C. The Biebl preculture medium components in 1 L distilled water were: glycerol, 20 g; K₂HPO₄, 0.5 g; KH₂PO₄, 0.5 g; MgSO₄ ·7H₂O, 0.2 g; (NH₄)₂SO₄, 5 g; CaCl₂ · 2H₂O, 0.02 g; cysteine HCl, 0.3 g; Resarzurin, 0.005 g; CaCO₃, 2 g; 2 mL of a trace element solution and 1 g yeast extract. The medium was prepared in anaerobic serum bottles, degassed with N₂ and autoclaved for 20 min at 121°C. The trace elements solution consisted of ZnCl_2, 70 mg; MnCl₂ · 4H₂O, 100 mg; H₃BO₃, 60 mg; CoCl₂ · 6H₂O, 200 mg; CuCl₂ · 2H₂O, 20 mg; NiCl₂ · 6H₂O, 20 mg; Na₂MoO2 · 2H₂O, 40 mg and 1 mL HCL (25% v/v) per 1 L distilled water.

In the H‐Cell BES four fermentations were carried out under two different iron conditions with and without electrochemistry; in the bioreactor with the All‐in‐One electrolysis electrode two fermentations were run under iron limited conditions with and without electrochemistry. The fermentation medium for bioreactor or H‐Cell cultivations was based on the Biebl medium with 80 g/L glycreol but without CaCO₃. Additionally 0.5 g/L cysteine‐HCl and for one fermentation 5 mg/L FeSO₄ · 7H₂O in the H‐Cell was added. The current for hydrogen production was applied when the redox potential of the medium reached a relatively constant and minimum value (around –580 mV in all the fermentations) that corresponded to the end of the lag phase with OD around 2–3. The H‐Cell chambers of 500 mL total volume each were filled each up to 300 mL (600 mL in the whole H‐Cell BES) with a predetermined medium described above. The chambers were separated by a Nafion® 117 proton exchange membrane (Fig. 2A). The cathode side was used for the fermentation and equipped with a pH probe, a redox probe, cooled gas outlet, and a gas sparging ring. Both sides were stirred with a magnet bar. Platinized titanium grids were used as electrodes. The total electrode area in contact with the medium was 50.4 cm², which conforms to 0.034 cm² per mL medium. The H‐Cell was doubled‐jacketed for temperature control at 35°C with water circulation. After filling the vessels with medium the chambers were closed and the whole reactor was autoclaved at 121°C for 20 min. Sparging the reactor for 1 h with N₂ before inoculating provided anaerobic conditions in the fermentation chamber.

During fermentation in the H‐Cell BES a potentiostat (Gamry Interface 1000) applied a potential to maintain a current of 150 mA between the electrodes (0.25 mA/mL overall volume). The potential was only applied when the pH of the fermentation broth was decreasing below pH 6 to regulate the pH initially without any chemicals and later only with base. Consequently in the lag and exponential phases the potential was applied in pulses. The cathode produced H₂ and OH⁻. By having a cationic charge transfer through the Nafion membrane toward the cathode, OH⁻ ions were hold on the fermentation side and neutralized microbially produced acids. With the increasing electrolysis the anode chamber became more and more acidic and the charge transfer smoothly switched to proton transmission. In this phase the electrolysis worked constantly and the pH was regulated additionally by adding a 5 M KOH solution with a pump.

In the control fermentations in the H‐Cell a pH of 6 was controlled automatically since the beginning with a 5 M KOH solution. Electrodes, anode chamber, and membrane were also implemented but without electricity. The microbially produced hydrogen in electrochemical affected H‐Cell cultivations (bio‐hydrogen) was calculated by subtracting the theoretically electrochemically produced hydrogen at 150 mA (Eq. (6)) from the overall measured hydrogen production.

$$n_{H_2}=\frac{I\times t}{z_{H_2}\times F}$$ (6)

Where I = current [A], t = time [s], F = Faraday constant [C/mol] and $z_{H_2}$= charge number of H₂. For the All‐in‐One electrolysis electrode experiments the bioreactor was filled with medium up to 1.5 L and stirred at 350 rpm using a bottom driven six blade stirrer (Fig. 2B). A pH probe was implemented to maintain the pH at 6 by pumping a 5 M KOH solution. Degassing with N₂ and checking with an installed redox probe achieved an anaerobic environment before inoculation. After filling the reactor with medium the reactor was in situ autoclaved at 121°C for 20 min. Sparging the reactor for 1 h with N₂ before inoculating provided anaerobic conditions. The All‐in‐One electrolysis electrode was driven with a constant 400 mA current (0.266 mA/mL) and equipped as in the characterization experiment described above. The microbiologically produced hydrogen (bio‐hydrogen) was calculated by subtracting the calculated electrochemically produced hydrogen at 400 mA as described in Eq. (5) from the overall produced hydrogen. After the fermentation process the All‐in‐One electrolysis electrode was in situ cleaned using a usual bioreactor cleaning protocol. To this end, the reactor was cleaned with 0.05 M NaOH followed by 0.05 M citric acid for 20 min each at 80°C. After the last cleaning step with water under same conditions, the electrode was ready to use for the next fermentation. In case of significant fouling the ceramic can be replaced and the electrodes were cleaned with peroxymonosulfuric acid.

2.4. Analytics

The gas analysis was done with a combination of a volumetric gas measurement using a Miligascounter (Dr.‐Ing. RITTER Apparatebau GmbH & Co. KG, Bochum) and gas composition measurement using a Balzers Omnistar 3000 mass spectrometer (Pfeiffer Vacuum GmbH, Asslar).

For the iron measurement in the H‐Cell experiments, liquid samples were either promptly frozen to break up the cell envelopes to extract the iron from the cells for the measurement of intracellular and extracellular irons or cell‐free filtered before freezing to get only the extracellular iron concentration. For analysis of unfiltered samples (with extracellular and intracellular iron) they were first filtered and photometrically measured. Cell‐free filtered samples were measured directly. An average error of 4% was calculated by triplet determination experiments.

For each of fermentation normally 15–16 samples were taken, which were cell‐free filtered and prepared for HPLC (Kontron Instruments) analysis to measure glycerol, n‐butanol, 1,3‐propanediol (1,3‐PDO), ethanol, butyrate, acetate, formate, succinate, lactate, and propionate. Separated by an Animex HPX‐87H column the sample components were detected using a refracting index and ultraviolet detector at 210 nm. The HPLC operating temperature was 60°C at a flow rate of 0.6 mL/min of the 5 mM H₂SO₄ mobile phase. For the HPLC analysis an average error of 5% was calculated by triplet determination experiments.

The OD of the samples was measured by using a UV/vis V 1200 Spectrometer (VWR) at 600 nm. An OD/DM correlation factor of 0.337 was established to obtain the biomass concentration in g/L. Furthermore, the specific production rate q _product for the exponential growth phase was calculated using Eq. (7).

$$q_{\mathrm{product}}=\frac{\mathrm{\Delta }{c}_{\mathrm{product}}}{\mathrm{\Delta }{c}_{BM}}\times \mu$$ (7)

Where μ = growth rate [h⁻¹], $c_{BM}$= biomass concentration [g/L] and c _product= product concentration [g/L]

According to the stoichiometric calculations of the carbon balance in C. pasteurianum DSM 525 glycerol fermentations published in 18, Eq. (8) 19 was applied to calculate the carbon recovery [%] (C_recovery) in batch culture. The redox balance was calculated using Eq. (9) to determine the NADH_recovery [%] averaged over the last 10 samples 19.

$${\mathrm{C}}_{\mathrm{recovery}}=\frac{{{\mathrm{C}}^{\ast }}_{\mathrm{PDO}}+{{\mathrm{C}}^{\ast }}_{\mathrm{Butanol}}+{{\mathrm{C}}^{\ast }}_{Butyrate}+{{\mathrm{C}}^{\ast }}_{\mathrm{Ethanol}}+{{\mathrm{C}}^{\ast }}_{\mathrm{Lactate}}+{{\mathrm{C}}^{\ast }}_{\mathrm{Acetate}}+{{\mathrm{C}}^{\ast }}_{\mathrm{Formate}}+{{\mathrm{C}}^{\ast }}_{\mathrm{Succinate}}+{{\mathrm{C}}^{\ast }}_{\mathrm{CO}2}+{{\mathrm{C}}^{\ast }}_{BM}}{\mathrm{\Delta }{{\mathrm{C}}^{\ast }}_{\mathrm{Glycerol}}}$$ (8)

$$\mathrm{NAD}{\mathrm{H}}_{\mathrm{recovery}}=\frac{{\mathrm{c}}_{\mathrm{PDO}}+{\mathrm{c}}_{\mathrm{\Delta }{\mathrm{H}}_2}}{2{\mathrm{c}}_{\mathrm{acetat}}+2{\mathrm{c}}_{\mathrm{butyrat}}+{\mathrm{c}}_{\mathrm{lactate}}+{\mathrm{c}}_{\mathrm{BM}-\mathrm{H}}}$$ (9)

Where C*_i = Concentration c _i [mol/L] of compound i multiplied with the number of its carbon atoms and c_{BM − H} = NADH₂ associated with biomass [mol/L].

${\mathrm{c}}_{\mathrm{\Delta }{\mathrm{H}}_2}$represents the reducing power potentially released as H₂ through the two ferredoxin‐dependent oxidoreductases in the metabolism of pyruvate and crotonyl‐CoA, respectively. It can be calculated from the overall measured H₂ evolution rate $({\mathrm{c}}_{{\mathrm{H}}_2\mathrm{measured}}$) in the effluent gas subtracted from the H₂ generated by the electrode $({\mathrm{c}}_{{\mathrm{H}}_2\mathrm{electrode}})$ and biologically generated H₂ by assuming no involvement of the ferredoxin‐dependent oxidoreductases (${\mathrm{c}}_{{\mathrm{H}}_2\mathrm{assumed}}$):

$${\mathrm{c}}_{\mathrm{\Delta }{\mathrm{H}}_2}={\mathrm{c}}_{{\mathrm{H}}_2\mathrm{measured}}-{\mathrm{c}}_{{\mathrm{H}}_2\mathrm{electrode}}-{\mathrm{c}}_{{\mathrm{H}}_2\mathrm{assumed}}$$ (10)

$${\mathrm{c}}_{{\mathrm{H}}_2\mathrm{assumed}}=({\mathrm{c}}_{\mathrm{ethanol}}+{\mathrm{c}}_{\mathrm{acetate}}+2{\mathrm{c}}_{\mathrm{butyrate}}+2{\mathrm{c}}_{\mathrm{butanol}})$$ (11)

${\mathrm{c}}_{{\mathrm{H}}_2\mathrm{electrode}}$ was calculated from the operation parameters and time of the electrolysis electrode according to experimental characterization (see Results below).

3. Results

3.1. Characterization of the All‐in‐One electrolysis electrode in bioreactor without microorganism

The electrode was first tested in the Biebl medium without microorganisms by applying potentials to generate currents between 100 mA and 600 mA in 50 mA steps for one hour respectively (Fig. 3). The working electrode functioned as a cathode to produce hydrogen. For the first step to 100 mA a constant voltage of 3.15 V was reached after 200 s. Afterwards for the stepwise increases of 50 mA stable potentials were reached below 1 s. The output (P) resulting from these currents was between 1.3 and 2.4 W with an average energetic efficiency of 31.9 ± 2.2% after an initial lower efficiency of 25.2% at a current lower than 150 mA calculated with Eq. (1). The efficiency slightly decreased with higher current outputs. An average Faraday efficiency of 82.4 ± 1.2% was reached (Eq. (2)). The measured time resolved H₂ volume was converted into time dependent H₂ production rates in mmol/min ($Q_{H_2}$). The electrolysis experiment retained a relatively stable pH until 400 mA. After applying a current over 400 mA the pH decreased slowly.

Figure 3.

Experimental results of electrolysis with different outputs (P) raising from currents between 100 and 600 mA. Red arrow = H₂ production rate $Q_{H_2}$ at 400 mA, measured and calculated.

The estimation of generation rate of H₂ of the electrode at a given current and for a given time period is of practical importance, especially regarding the deviation of the maximal production rate from the theoretical maximum at the given current. The latter can be calculated with Eq. (4). For example, the theoretical maximum H₂ production rate $Q_{H_2}$with I = 400 mA is calculated to be 0.109 mmol/min. The experimental value is normally lower (Fig. 3, indicated by the red arrow) because of ohmic and efficiency losses (e.g. cathodic O₂ reduction, gradient forming). In order to estimate such losses and to assess the kinetics of H₂ production a longer experiment was carried out at I = 400 mA for 1270 min (Fig. 4). The other conditions were the same as for the experiments in Fig. 3 but with a more sensitive gas volume measurement (updated each minute). It was observed that the electrode has an acceleration phase until it reaches approximately its maximal rate. The real H₂ production rate (${Q_{H_2}}^{\ast }$ in mmol/min) can be described by Eq. (3) with a maximum production rate (${Q_{H_2}}^{\max }$) and a time constant b. ${Q_{H_2}}^{\max }$ was estimated as 0.089 mmol/min that corresponds to 81.65% of the theoretical maximum; the time constant b was estimated to be 27.7 min. With these parameters the amount of H₂ ($n_{H_2}$) in mmol produced by the electrode working at a given current (e.g. I = 400 mA) and for a given time period can be calculated from Eq. (5) as the integral of Eq. (3).

Figure 4.

H₂ production rate trend of the electrolysis electrode working at 400 mA. Equation (3) describes the red fitted curve.

The fast adjustment of a constant output (P) until the next current increase step observed in the electrode characterization experiment (Fig. 3) indicated no limitations in the counter electrode chamber. This means that neither gas bubbles block the electrode surface nor limitation of liquid transfer over the ceramic occurs. Otherwise an increased potential would be needed to maintain the current. During both experiments the O₂ headspace concentration in the reactor was continuously measured that was never above 0.03%. Consequently, it can be assumed that O₂ hardly passes through the working electrode into the reactor medium. This means an efficient gas transport through the inner channel to the outlet. However, a small amount of O₂ could pass through the ceramic and is then immediately reduced at the cathode, leading to a certain loss of the energetic efficiency mentioned above (Fig. 3).

3.2. Comparison of the All‐in‐One electrolysis electrode e‐bioreactor with the H‐Cell BES for real fermentation

Electrochemically affected glycerol fermentations were first investigated in the H‐Cell BES with C. pasteurianum DSM 525 under varied initial iron concentrations. After experiencing several obstacles (described below) in the H‐Cell cultivations fermentations with identical initial conditions were carried out in the All‐in‐One electrolysis electrode e‐bioreactor for comparison of the results.

3.2.1. Iron migration in the H‐Cell BES interfered with C. pasteurianum metabolism

Choi et al. (2014) performed their experiments in an H‐Cell BES system with a cathode compartment poised at +0.045 V vs. SHE and claimed a direct electron transfer from the cathode to the microbes. In this work, we tested the growth and metabolism of C. pasteurianum DSM 525 in a similar H‐shaped bioreactor, but applied a relatively high potential (11 ± 2 V between the anode and cathode) to generate a current of 150 mA and hydrogen electrochemically from water. As shown in Table 1 and Fig. 5 with electrochemistry under medium iron conditions a lower 1,3‐PDO yield was observed compared to the control fermentation (0.20 mol/mol vs. 0.32 mol/mol). Final butanol yields of 0.26 mol/mol with electricity and 0.14 mol/mol without electricity were reached. The same trend was obtained in fermentations with iron limitation, with 0.19 mol/mol vs. 0.10 mol/mol with and without electricity for butanol and 0.22 mol/mol vs. 0.38 mol/mol for 1,3‐PDO, respectively (Fig. 5). The NADH recovery observed in the electrofermentations was slightly but not significantly increased compared to the control fermentations (Table 1).

Table 1.

Final glycerol consumed, product concentrations (g/L or mmol/L for CO₂ and H₂), carbon, and NADH recoveries (%) in H‐Cell fermentations

	Iron limited, E	Iron limited, Control	Iron modicum, E	Iron modicum, Control
Glycerol	29.5	33.1	48.9	72.2
1,3‐PDO	5.5	10.4	8.0	19.4
Ethanol	0.0	0.0	0.7	1.2
Butanol	4.5	2.6	10.2	7.9
Butyrate	1.5	2.3	2.3	4.9
Acetate	0.5	0.8	0.6	1.5
Formate	0.5	0.6	0.6	0.9
Succinat	0.1	0.0	0.4	0.0
Lactate	3.4	5.0	0.0	2.7
BM	2.2	2.2	5.2	4.6
H2	231.8	187.0	398.7	470.5
CO2	184.6	135.1	345.3	376.3
C‐Recovery	96	100	100	94
NADH recovery	109.4 ± 18.8	104.6 ± 6.3	107.8 ± 12.1	105.4 ± 12.3

Figure 5.

(A) Substrate consumption, cell growth, product formation, and headspace hydrogen concentration in iron limited H‐Cell fermentation, (B) Results of iron‐modicum H‐Cell fermentation, (C) final yields of butanol, 1,3‐PDO and biomass based on glycerol (mol/mol).

The data obtained under the electrochemical influence were in fact not fully in agreement with those of Choi et al. (2014). We did not observe a general yield increase of 1,3‐PDO. Instead of that there was a shift of the product distribution from 1,3‐PDO to butanol. It is noticed that we found a slightly higher amount of reducing equivalents in the reducing power consuming products (pathways) than in the reducing power generating products (pathways) according to Eq. (9) (Table 1). Because of the relatively large deviations in the calculation of the redox recovery, especially in association with the calculation and measurement of H₂ we cannot conclude with certainty if electrons have been transferred from the cathode to the cells electrode. It should be mentioned that the potential applied in this work was much higher than that of Choi et al. (2014) and we did not observe biofilm growth. Khosravanipour et al. 20 also reported a butanol yield increase in an electrified H‐Cell with glucose consuming C. pasteurianum but only a minor biofilm due to H₂ production at the cathode. The authors observed a pH shift in their study but not further investigated its effect. In fact, the observed metabolic shifts in these studies might be also due other effect(s) as described below.

Iron is known to significantly affect the metabolism of C. pasteurianum, and several iron containing enzymes are involved in clostridia metabolism, e.g. nitrogenases, ferredoxin coupled enzymes, and alcohol dehydrogenases 16. Early studies showed a metabolic shift toward butanol production in iron excess medium 21. The effects of iron availability and the possible mechanisms of metabolism shift were investigated with more detailed metabolic and proteomic analyses in our recent systems biology study of the glycerol fermentation with C. pasteurianum DSM 525 19. We observed a significantly higher butanol/1,3‐PDO ratio under iron excess conditions. Hence we examined the glycerol fermentations in the H‐Cells with particular attention to the remaining iron concentrations in the different chambers.

At an initial iron concentration of 5 mg/L and despite the use of a proton exchange Nafion membrane, we found that iron migrated through the membrane from the anode to the cathode chamber (Fig. 6A). The migration rate of iron was estimated to be 4.8 times higher in the electrochemically affected fermentation than in the control fermentation.

Figure 6.

Iron concentrations in the cathodic (I) and anodic (without cells) (II) chambers of the electrochemically affected fermentation (A), compared to the control cultivation without current (B). The blue curve (triangles) describes the extracellular iron concentration while the red lines display the total iron concentration measured after cell disruption as a linear fit of the black curve (dots).

As shown in Fig. 5 and Table 1 the higher butanol yields achieved in the iron modicum H‐Cell fermentation with such a high potential could be caused by the excess iron migration through the Nafion membrane. Beside the iron migration it can be assumed that further cations have transferred through the membrane and interfered with the cultivation. Arising electrochemical effects could be overlaid by ion migration through the membrane and iron limitations due to deposition on the cathode during the process caused by insufficient iron concentrations. These effects combined with other drawbacks such as the more complex pH regulation, shape induced mixing limitations, and ohmic losses render H‐Cells not a reliable BES for performing and examining high potential electro‐fermentations.

3.2.2. The All‐in‐One electrolysis electrode and e‐bioreactor enable a better study of electrochemical effects on fermentation

Glycerol fermentations with C. pasteurianum grown in an e‐bioreactor with a working All‐in‐One electrolysis electrode and without the working electrode (control fermentation) were started with the same preculture and slightly different ODs of 0.23 and 0.27, respectively (Fig. 7A, t ₁). After the lag phase the All‐in‐One electrolysis electrode was put into operation in one of the fermentations (Fig. 7A, t ₂, red) to maintain a current of 400 mA that was achieved with a stable potential of 3.27 V for 12.25 h followed by a rise of the needed potential to 3.81 V. This potential increase could be explained by the simultaneously increasing butanol and 1,3‐PDO concentrations that led to a lower permitivity constant than water and glycerol (butanol = 17.51 F/m; 1,3‐PDO = 35 F/m; water = 78.39 F/m; glycerol = 42.5 F/m) or a blocked ceramic separator by microorganisms. The redox potential in both fermentations stabilized around –587 mV after the lag phase. In the electricity‐affected fermentation the redox potential could be maintained, while the control fermentation showed a slightly increasing redox potential. At t ₃ the culture reached an OD value around 10.2 with the working electrode and 7.9 without electrochemical effects (control). Here the cultures reached a stationary phase and were aborted. During the growing phase the electricity‐affected fermentation achieved a higher specific growth rate of 0.19 h⁻¹ compared to 0.11 h⁻¹ in the control. Also here a slightly but statistically not significantly NADH recovery over 100% in the electrofermentation was observed (Table 2).

Figure 7.

(A) Redox potential, hydrogen evolution, BM progress of fermentations and applied potential in electro‐fermentation. T₁: Inoculation, t ₂: Start of the electrode, t ₃: End of the cultivation; (B) Product formation and headspace hydrogen concentrations of All‐in‐One electrolysis affected and control glycerol fermentation with C. pasteurianum DSM 525 under iron limited conditions.

Table 2.

Final glycerol consumed, product concentrations (g/L and mmol/L for CO₂ and H₂), carbon, and NADH recoveries (%) in the All‐in‐One electrolysis electrode e‐bioreactor and control fermentations

	Iron limited, Electrode	Iron limited, Control
Glycerol	56.7	41.7
1,3‐PDO	11.6	9.9
Ethanol	0.0	0.0
Butanol	11.3	6.3
Butyrate	0.7	1.5
Acetate	0.7	0.9
Formate	0.2	0.1
Succinat	0.1	0.0
Lactate	6.5	7.9
BM	3.4	2.7
H2	470.03	292.6
CO2	330.8	214.9
C‐Recovery	98	103
NADH recovery	107.7 ± 10.8	105.6 ± 10.2

Further significant differences in product concentrations and yields between the electrochemically affected fermentation and the control one were observed (Table 2 and Figs. 7 and 8). The butanol production was enhanced from 0.19 to 0.25 mol/mol while the 1,3‐PDO production was decreased (0.29–0.25 mol/mol) with the ectrochemical effect. The biomass related yields for butanol was increased from 3.24 g/g in the control fermentation to 4.52 g/g in the fermentation with electrochemical effect. The 1,3‐PDO yield decreased from 4.96 g/g to 4.50 g/g. The specific yields (Eq. (7)) for the exponential growth phase revealed similar trends (electrochmically affected q _butanol= 0.54 g/(g h) and $q_{PDO}$= 0.24 g/(g h) vs. control q _butanol= 0.26 g/(g h) and $q_{PDO}$= 0.24 g/(g h)).

Figure 8.

(A) Final yields per glycerol of the electrode and control glycerol fermentation with C. pasteurianum DSM 525. (B) Final yields per BM of the respective fermentations.

The growth and yield differences between the fermentations in the e‐bioreactor with the All‐in‐One electrolysis electrode and the H‐Cell compared to the respective control gave a similar tendency.

Electrochemically produced H₂ obviously led to a metabolic shift toward butanol in the product distribution, which was also noticed in ABE fermentation with C. acetobutylicum by in increasing the H₂ partial pressure 22. With the All‐in‐One electrolysis electrode this was shown unambitiously and the effects of iron shift can be excluded.

4. Conclusion and perspectives

The All‐in‐One electrolysis electrode developed in this work successfully integrates the anode and cathode of a typical electrochemical system into a simple, rod‐shaped structure and can be easily integrated into conventional bioreactors such as stirred tank reactors to convert them into a useful bio‐electrochemical reactor system for electricity‐aided bioprocesses. It has a high Faraday efficiency over 80% as demonstrated for in situ water electrolysis for hydrogen generation in a stirred e‐bioreactor. The new e‐bioreactor was applied for glycerol fermentation by C. pasteurianum to demonstrate its functionality and usefulness for real fermentation process, especially in comparison to presently mostly used H‐Cell two‐chamber bioreactor. In addition to the inherent limitations of H‐Cell reactor such as the use of ion selective membrane for separating the two reactor chambers and the related problems of small charge transfer area and scale‐up issue, we showed here that iron transfer between the two chambers significantly interfere with the cell growth and metabolism. The new e‐bioreactor with the All‐in‐One electrolysis electrode can well solve this problem and is thus more suitable for assessing the electrochemical effects on fermentation. In terms of scale‐up that is a major disadvantage of the two‐chamber bioreactor system (“unused” counter electrode chamber, limited area of separation membrane, heterogeneity and transport limitation) the new e‐bioreactor is also of advantage: multiple All‐in‐One electrolysis electrodes of enlarged size can be relatively easily integrated into the bioreactor in different ports and places. This distributed solution can also warranty a better distribution of the electrochemically active surface and related function over the whole reactor volume. For the glycerol fermentation with in situ hydrogen generation we observed a general trend of shifting of the metabolism toward butanol formation. More detailed work is needed to clarify this metabolic shifting. The energetic efficiency of the electrode can be also further improved. It is known that energetic efficiency for electrical hydrogen production in specialized water electrolyzers can reach as high as 70% 23. To this end, the structure and the materials used can be further improved.

The All‐in‐One electrolysis electrode could be also explored for a number of other applications in biotechnology, for example for the control of redox during fermentation, for bioelectrochemical transformation with immobilized enzymes on the electrode surface, for combined electrochemical and biological conversion of CO₂. For potentiostatic biofilm experiments a direct or salt‐bridge utilizing reference electrode can be integrated into the scaffold above the electrode assembly. It is also thinkable to make use of both hydrogen and oxygen from water cleavage for bioproduction in one and the same system. Related studies are on‐going in our lab.

Practical applications

Electrobiotechnology is gaining more and more attention for electricity‐driven or aided bioproduction of chemicals and fuels. It calls for new bioelectrochemical systems (BES) or electro‐bioreactors (e‐bioreactors) that are feasible for efficient and large‐scale operation. We have designed a new electrode assembly to integrate the working and counter electrodes of a conventional two‐chamber bioelectrochemical system into a rod shaped All‐in‐One electrolysis electrode. It can be easily put into common bioreactors to perform electrochemically supported bioprocesses. This All‐in‐One electrolysis electrode can be used as hydrogen or oxygen donor, electrodialysis unit, and/or electrocatalytic surface for biological or chemical conversion processes.

The authors have declared no conflict of interest.