Introduction

Hydrogen can be used in both power generation systems and direct combustion processes, providing the great advantage of clean combustion. Moreover, H2 is considered a very competitive energy carrier compared to other fuels, thanks to its high net heating value per unit volume. Nowadays, its use as a clean energy source is yet uncommon, while its main use is in ammonia production and hydrogenation of coal and petroleum during hydrocracking of traditional fuels (IEA 2019). However, the good environmental profile of H2 is commonly counteracted by the fact that it is still primarily derived from non-renewable sources, with a high associated energy consumption and relevant related CO2 emissions, posing an urgent need for sustainable production methods.

Several bioprocesses have been investigated over the last decades to produce H2 through sustainable methods (Hallenbeck et al. 2012). Among them, dark fermentation (DF) is considered one of the most promising options. The main reason is that DF averts the major drawbacks of other biological processes (including direct or indirect photolysis and photo-fermentation), related to the intermittent production of H2 and the need of a light source to support the process. Compared to the other biological processes, dark fermentative H2 production has the additional advantages of higher production rate, flexibility of operation under different temperature and pressure conditions, lower net energy input and, noteworthy, applicability to a range of renewable organic sources including organic residues and carbohydrate-based wastewaters (Ghimire et al. 2015; Silva et al. 2017; Park et al. 2021).

While DF of organic residues has been widely investigated over the past decades, the major current challenges include the poor stability of the biochemical process and the thermodynamic/biochemical limitations to the actual H2 production yield attainable. To this regard, when acetate is the final metabolic product of fermentation, a production of 4 mol of H2 per mole of hexose consumed is expected, which is regarded as an upper threshold for the H2 yield, known as the Thauer limit (Thauer et al.1977). Therefore, of the potential 12 mol of H2 that may be produced by one mole of glucose, only a third can be obtained biochemically. The actual H2 yield can even be lower than the mentioned limit if other more reduced metabolic products (e.g., butyrate, ethanol) are formed or additional competing metabolic pathways occur (e.g. propionic fermentation, homoacetogenesis).

Bioelectrochemical processes (BESs) have been proposed for a variety of applications aimed at improving the performance of biological systems. Their operating principle is based on the ability of specific microorganisms defined as electroactive bacteria (EAB) to interact with solid electrodes by forming a biofilm and catalyse the oxidation of organic matter by generating an electric potential. Microbial fuel cells (MFCs) are among the most widely investigated BESs, due to their capability of producing an electric power while simultaneously degrading an organic substrate. Generally, in the anodic chamber, where the EAB are attached to dedicated inert electrodes, the oxidation of organic substances takes place generating CO2 and protons, which migrate into the cathodic chamber through ion exchange membranes. The cathodic chamber is maintained under aerobic conditions, so that, in the presence of electrons, protons react with oxygen to produce water, resulting in the spontaneous production of electricity, the intensity and flow of which are functions of the construction features of the cell, the substrate characteristics, the inoculum and the operating conditions adopted.

A modified type of MFC, the microbial electrolysis cell (MEC), has been studied since 2005 (Liu et al. 2005b), and its scientific interest has strongly increased in recent years (Santoro et al. 2017). In that case, unlike the MFC, the cathodic chamber is maintained under anaerobic conditions; consequently, protons are reduced to H2, since there are no other electronegative species to intercept electrons. This process requires the supply of an electric current, since the electric potential naturally generated by microorganisms is not enough to reduce H+ to H2. Assuming acetate as a model organic source, the electrode reactions involve oxidation to CO2 at the anode and H+ reduction to H2 at the cathode (see Eqs. 1 and 2). Assuming that the open-circuit potential at the anode in an MFC is generally about E0 ~ – 300 mV (Liu et al. 2005b) and the minimum standard redox potential required for the cathodic reaction is E0 = – 410 mV (NHE) at pH 7.0, H2 can theoretically be obtained by applying a higher than 110 mV circuit voltage (typically 410–300 mV to overcome internal electric resistances). However, the voltage required is significantly lower than that used for conventional water electrolysis (1.21 V at neutral pH, which can increase up to 1.8–2.0 V under alkaline conditions due to electrode overpotentials), since the chemical energy extracted from organic substrates oxidized at the anode supplies most of the potential needed.

$${\text{Anode}}:{\text{ CH}}_{{3}} {\text{COOH }} + {\text{ 2H}}_{{2}} {\text{O}} \to {\text{2CO}}_{{2}} + {\text{ 8e}}^{ - } + {\text{ 8H}}^{ + }$$
(1)
$${\text{Cathode}}:{\text{ 8H}}^{ + } + {\text{ 8e}}^{ - } \to {\text{4H}}_{{2}}$$
(2)

Some studies have successfully investigated BESs for the exploitation of volatile fatty acids (VFAs) or DF effluents into electricity or H2. Liu et al. (2005b) obtained 2.9 mol H2/mol acetate applying an additional voltage of 0.250 V in a MEC. Through optimization of materials and reactor configuration, Cheng and Logan (2007) achieved H2 yields between 2.0 and 3.9 mol H2/mol acetate at applied voltages of 0.2–0.8 V. Chae et al. (2008) showed that H2 production gradually increases as the applied voltage is increased from 0.1 to 1 V, reaching 2.1 mol H2/mol acetate. Liu et al. (2005a) tested power generation from acetate and butyrate in a MFC and observed that acetate is preferred over butyrate as the substrate, producing respectively 506 mW/m2 and 305 mW/m2.

The treatment of a real DF effluent was investigated by Chookaew et al. (2014) using both a MEC and a MFC. A power density of 92 mW/m2 in the MFC was achieved along with 50% COD removal. When treated in the MEC, the same substrate yielded 106 mL H2/g COD. Rivera et al. (2015) evaluated DF effluent exploitation as a substrate for a MEC. The highest production rate (81 mL H2/L/day) was obtained at a 550 mV voltage and was accompanied by 85% COD removal. Wang et al. (2011) performed a multi-stage process using a DF reactor for cellulose degradation, followed by two MFCs that were used as power sources for a subsequent MEC. The MFCs produced a maximum of 0.43 V using the fermentation effluent that induced H2 production in the MEC at a rate of 0.48 m3 H2/m3/d and with a yield of 33.2 mmol H2/g COD removed in the MEC. The authors observed a 41% overall improvement in H2 production for the integrated process compared with fermentation alone.

The integration of fermentation and electrochemical processes in the same unit has been the focus of specific studies on electro-fermentation (Moscoviz et al. 2016; Schievano et al. 2016; Yu et al. 2018). The fundamental concept is based on driving the fermentation process by modifying the redox potential through polarized electrodes placed in the reactor, which can either supply electrons or act as a sink under certain conditions. This could allow overcoming the metabolic limitations through direct electricity supply to the fermentation medium. Potential inocula include both electroactive and fermentative bacteria that can produce value-added organic acids and alcohols (Xue et al. 2018; Paiano et al. 2019), sometimes with concomitant production of H2 and/or CH4 (Nelabhotla and Dinamarca 2019; Toledo-Alarcón et al. 2019). Electro-fermentation has been rapidly gaining attention given the successful results. To date, the study of the process is still in a preliminary stage and future developments include the orientation of the metabolic pathways towards specific end products, the selection of efficient redox mediators, the application to complex substrates or suspended biomass configurations.

In the present work, an attempt was made at developing an innovative BES coupling DF with an electrochemical process, with the multiple aims of enhancing H2 generation, exploiting the fermentation products, produce electricity and provide an internal pH buffering effect. To the best of the authors’ knowledge, the concept behind the proposed process is novel and the BES developed has not been documented in other literature studies so far.

Materials and methods

Integrated bio-electrochemical system; principle and setup

The integrated bio-electrochemical system (IBES) proposed here is based (see Fig. 1) on the electrochemical reduction of the protons released from the dissociation of the VFAs produced during the fermentation process, leading to additional H2 generation. The reduction reaction is mediated by the electrons released by the oxidation of a metallic element in the anodic chamber, which generates an electric current. An inert electrode, which does not take direct part in the reaction but rather plays the role of electron carrier, is placed in the fermentation medium, which is connected through an external electric circuit to the reducing electrode (anode) placed in an electrolytic solution in a dedicated chamber. The ion flow required to maintain the electroneutrality of the two electrolytes is attained through an appropriate connection between the two compartments. The reactions that occur in the cathodic (1) and anodic (2) compartments are shown below, along with the corresponding reduction potentials (E0) in accordance with the IUPAC standard potentials convention (298 K, 1 bar, 1 M), assuming metallic Zn as the anode:

$${\text{2H}}^{ + } + {\text{ 2e}}^{ - } \leftrightarrow {\text{ H}}_{{2}} \quad {\text{E}}^{0} \left( {{\text{2H}}^{ + } /{\text{H}}_{{2}} } \right) \, = \, 0.000{\text{V}}$$
(1)
$${\text{Zn }} \leftrightarrow {\text{ Zn}}^{{{2} + }} + {\text{ 2e}}^{ - } \quad {\text{E}}^{0} \left( {{\text{Zn}}_{{2}}^{ + } /{\text{Zn}}} \right) \, = \, - \, 0.{\text{762V}}$$
(2)
Fig. 1
figure 1

Layout of the IBES

Full size image

The overall cell electromotive force under standard conditions (∆E0), defined as the potential difference between the cathode and the anode, for this system is as follows:

$$\Delta {\text{E}}^{0} = {\text{ E}}^{0} \left( {{\text{2H}}^{ + } /{\text{H}}_{{2}} } \right) \, - {\text{ E}}^{0} \left( {{\text{Zn}}^{{{2} + }} /{\text{Zn}}} \right) \, = \, 0.{\text{762V}}$$
(3)

The fact that the Gibbs free energy ∆G0 = − nF∆E0 (with n = number of electrons exchanged in the reaction and F = Faraday’s constant = 9.64853 × 104 C mol−1) is negative (− 147 kJ) ensures that the redox reaction can take place spontaneously, as the reduction potential of the anode is adequately low.

Consequently, the IBES provides, compared to the biochemical process, an additional electrochemical generation of H2, exploiting the protons from the metabolic products. Moreover, since the system is designed as a galvanic cell, the process is energetically self-sufficient. Finally, the conversion of protons to H2 may offer the additional benefit of contrasting the acidification of the fermentation medium, which would otherwise require the external addition of buffering agents to maintain adequate pH levels for the microbial system.

In the present work, four different experimental setups (named A, B, C, D) were designed in order to identify the most suitable materials and configuration for the IBES (Fig. 2). The anode and cathode were mutually connected through an external electric circuit equipped with a system for continuous measurement and recording of electric current and cell potential produced under an electric load (resistor). As detailed in Table 1, the four systems differed for the compartment volume, the type of separation between the cathodic and anodic solutions (involving either a salt bridge or an anion exchange membrane, AEM) and (when applicable) the AEM surface-to-volume ratio (S/V). Both the salt bridge and the AEM served the purpose of allowing the ionic flow between the two compartments required to ensure the electroneutrality of the catholyte and anolyte. The AEM was specifically selected because of its recognized acid/proton blocking capability (Xu 2005; Guo et al. 2017).

Fig. 2
figure 2

Schematics of the IBES experimental setups

Full size image
Table 1 Summary of the key features of system configuration (synthetic model solution used)
Full size table

While in systems A and B the gas produced was allowed to evolve outside the system and therefore was not directly measured, systems C and D were gas-tight and also included collection, measurement and sampling of the gas generated. Dedicated H2 leakage tests were conducted for systems C and D in order to quantify potential gas losses during the tests due to the high fugacity of H2. The measured loss was found to lie in the range 0.12−0.16 mL/h for system C and 0.64− 0.95 mL/h for system D (likely due to some minor gas leakage through the AEM), which was accounted for to quantify the amount of gas produced.

In systems A and B, the following cathode materials were selected for the tests on the basis of electrical conductivity, recognized inert redox behaviour and absence of potential toxic effects on microorganisms: graphite sheet (15 cm2), Pt sheet (2 cm2), Ti grid (2 mm wire with mesh of 0.16 mm2) and Ni mesh (60 mesh with 0.18 mm wire).

In all systems, the anolyte was 0.5 M Zn sulphate, while a metallic Zn plate was used as the anode.

Materials and electrochemical/bio-electrochemical tests

CW was collected at an Italian dairy industry producing mozzarella cheese from a mixture of cow and buffalo milk. The characteristics of CW are reported in Table 2. The samples were stored at − 18 °C and thawed at room temperature for approximately 24 h before use. The pH of CW was adjusted to 7.5 at the beginning of the DF experiments using 2 M NaOH, while no further pH control was performed during the tests. The indigenous microorganisms in CW were the only active biomass source in the system, as it was previously demonstrated (Asunis et al. 2019) that it can successfully produce H2 through dark fermentation with no need of an external inoculum. The DF tests were operated under mesophilic conditions (38 ± 1 °C).

Table 2 Characterization parameters of CW used for IBES in system D and stand-alone DF
Full size table

In the preliminary electrochemical tests, conducted in systems A, B and C, either diluted acetic (HAc) or butyric acid (HBu) was used as the model substrate to simulate the metabolic products of DF. The IBES was then tested in a batch mode using the optimized configuration D with cheese whey (CW) as the substrate; in this case, a stand-alone batch DF reactor fed with CW was employed as a reference. All tests were performed in duplicate. A summary of the experiments is reported in Table 3.

Table 3 Experimental design
Full size table

Analytical methods

In systems C and D, a volumetric gas counter with a 2 mL capacity was used for gas volume measurement, while a gas bag was employed for gas storage. In all cases the measured volume was converted under standard pressure and temperature conditions (T = 273.15 K, P = 105 Pa).

The biogas was periodically sampled from the gas bag with a 25-mL gastight syringe and analysed through a gas chromatograph (Model 3600 CX, VARIAN) equipped with a thermal conductivity detector and 2-m stainless-steel packed column (ShinCarbon ST) with an inner diameter of 1 mm. The operating temperatures of the injector and detector were 100 and 130 °C, respectively, with He as the carrier gas. The oven temperature was initially set at 80 °C and subsequently increased to 100 °C at 2 °C/min.

The VFAs (acetate, butyrate, propionate, valerate, caproate, heptanoate) and ethanol concentrations were determined in 0.2-µm filtered and HCl acidified (pH = 2) liquid effluent (1 µl) with a gas chromatograph equipped with a flame ionization detector (FID) and a 30-m capillary column (TRB-WAX) with an inner diameter of 0.53 mm. The temperatures of the detector and the injector were 270 and 250 °C, respectively. The oven temperature was initially set at 60 °C, held for 3 min at this value, subsequently increased to 180 °C at a rate of 10 °C/min and finally increased to 220 °C at a rate of 30 °C/min and held for 2 min.

Sulphates, total and volatile solids were measured according to the Standard Methods for the Examination of Water and Wastewater (APHA AWWA and WEF 2005).

Total organic carbon (TOC) was measured using a Shimadzu TOC analyser (TOC-VCHS and SSM-5000 module, Shimadzu, Japan).

Carbohydrates were analysed through the colorimetric phenol–sulfuric acid method using glucose as the standard (Dubois et al. 1956).

The electrochemical process was monitored through cell voltage (∆V) and electric current intensity (I) measurements. The acquisition system NI cDAQ-9174 was used for this purpose, and a potentiometer with a resistive load ranging from 500 Ω to 1.3 Ω was used to obtain the power curves for each system configuration. The cell voltage was measured continuously for a few minutes following each resistance variation, in order to avoid significant changes in the electrolyte solutions, at the same time ensuring the achievement of equilibrium conditions. The measurement system was combined with LabVIEW as the data acquisition software.

The total amount of electric charge, Q, generated during the electrochemical process was calculated as the integral of the measured electric current, and the related theoretical amount of H2 produced was also derived.

Results and discussion

Electrochemical tests

A summary of the preliminary tests using systems A and B is provided in Table 4. The cathode characteristics were found to affect the electrochemical profile of the process, with the highest (2.2 mA) and lowest (0.77 mA) current intensities in system A being obtained with the Ti grid (run A-Ti-HAc) and graphite sheet (run A-G-HAc) electrodes, respectively. However, system A also showed the presence of a high overpotential, likely due to the fact that salt bridges are known to generate high internal resistances (Logan et al. 2006). When the salt bridge was replaced by the AEM (system B) using a Ti cathode, the cell voltage and the current intensity increased from 0.20 (run A-Ti-HAc) to 0.43 V (run B-Ti-HAc) and from 2.2 to 4.7 mA (average of the values measured during the first hour of the tests at 90 Ω as the external load), respectively.

Table 4 Main results for systems A and B (average values during the first hour with a fixed external resistive load of 90 Ω)
Full size table

Figure 3 depicts the power curves derived to describe the electrical characteristics of the optimized systems (runs B-Ti-HAc, C-Ti-HAc and D-Ti-HAc). The maximum power, P, observed was 5.1 mW for D-Ti-HAc (at I = 9 mA and R = 62 Ω), 2.5 mW for B-Ti-HAc (at I = 8.5 mA and R = 33.5 Ω) and 0.6 mW for C-Ti-HAc (at I = 2.1 mA and R = 110 Ω). Since the experimental conditions of the three systems were the same apart from the S/V ratio of the AEM, the results show that this parameter played a key role in determining the power efficiency of the electrochemical cell. Run D-Ti-HAc, where the geometrical configuration was arranged to maximize the S/V ratio, indeed showed to have been improved with regard to the electrical performance.

Fig. 3
figure 3

Power curves for B-Ti-HAc, C-Ti-HAc and D-Ti-HAc showing the influence of the S/V ratio

Full size image

The process evolution over time displayed a similar profile in all the investigated systems, although with different absolute values and rates of variation of the investigated parameters. In particular, as shown in Fig. 4 for C-Ti-HAc and D-Ti-HAc, the catholyte pH displayed an increasing trend as a result of proton conversion into H2, with the typical shape of an acid–base titration curve that reached a final plateau as soon as the acid dissociation was complete. The current intensity mirrored the pH evolution, decreasing to almost zero as pH levelled off at the plateau. When comparing C-Ti-HAc and D-Ti-HAc, it is clear that the lower S/V ratio of the former resulted in a considerably slower current evolution and remarkably lower current intensities.

Fig. 4
figure 4

Time evolution of catholyte and anolyte pH and current intensity for a C-Ti-HAc and b D-Ti-HAc; c catholyte pH and acetate partitioning among the two chambers for D-Ti-HAc

Full size image

It is also worth mentioning that in all systems the electric current flow was observed to recover upon renewed addition of the proton source (data not shown), as would happen during continuous fermentation.

The anolyte pH was found to slightly increase within the first 5−6 h of the test from an initial value of ~ 4.3 to a value of ~ 5.7, likely due to the migration of hydroxide ions from the catholyte through the AEM. The final constant pH value achieved is in good agreement with that expected in a solution in equilibrium with a Zn(OH)2 precipitate (Ks = 2 × 10−17 − 4 × 10−17). Acetate and butyrate were also clearly observed to migrate (most likely in the dissociated form, given the nature of the AEM used) to the anodic chamber over time, in compliance with electroneutrality constraints, and virtually fully transferred to this compartment at the end of the test (see the values of the percent partitioning of acetate at the anode shown in Fig. 4c) for D-Ti-HAc and Fig. 5b) for D-Ti-HBu).

Fig. 5
figure 5

Time evolution of catholyte and anolyte pH and current intensity for D-Ti-HBu (a); (b) catholyte pH and butyrate partitioning among the two chambers for D-Ti-HBu

Full size image

For D-Ti-HBu, the observed trends (see Fig. 5) of the electric current, pH of the cathodic and anodic solutions as well as the acid dissociation behaviour and migration of anionic species through the AEM were identical to run D-Ti-HAc. This prospectively indicates that the electrochemical process investigated can be applied to a fermentation system where an array of organic acids is generated.

A summary of the assessment of the process performance of C-Ti-HAc, D-Ti-HAc and C-Ti-HBu is provided in Table 5. The reported data provide a comparison between the observed cumulative H2 production and the total theoretical H2 yield expected based on either the overall amount of electric charge generated or the overall amount of protons derived from acid dissociation.

Table 5 Summary of preliminary results for systems C and D
Full size table

Taking into account the potential gas losses (see Sect. 2.1), the data in Table 5 show that the total volume of H2 produced is consistent with the total theoretical volume of H2 calculated from the mobilized electrons. The released protons, assuming complete dissociation of the acid, appear to have been totally reduced to H2 for C-Ti-HAc (VH2,th/VH2,max=99%), albeit at a slower production rate, and for D-Ti-HAc (VH2,th/VH2,max=102%) , while for D-Ti-HBu the conversion was slightly lower (90%).

Bio-electrochemical tests

Figure 6a shows the evolution of the cumulative H2 production for the stand-alone DF reactor and the IBES, while the profiles of pH of the biological compartment and current intensity are reported in Fig. 6b. The experimental data indicate a final yield of, respectively 22.4 NL H2/kg TOC and 68.7 NL H2/kg TOC (the latter corresponding to 75.5–78.8 NL H2/kg TOC considering the H2 leakage through the AEM), and a total duration of H2 production of 32 and 44 h. As observed in our previous experiments on dark fermentation of various organic substrates, the biological process stopped as soon as the degradation of carbohydrates, which are the preferred substrate for H2 generation, was complete. The generation of organic acids as the metabolic products of fermentation was therefore virtually complete after 32 h (DF alone) and 44 h (IBES) from the start of the experiments, concomitantly with the pH plateau at 5–5.5.

Fig. 6
figure 6

a Comparison between the stand-alone DF and IBES in terms of H2 yield and theoretical electrochemical H2 yield calculated from charge (green dashed line); b pH evolution over time in both systems and electricity production in the IBES.

Full size image

From the data in Fig. 6a, it is therefore evident that the IBES attained a significant improvement in H2 production (by 3 times) over stand-alone DF. In order to assess the advantages of the IBES over the conventional DF process, an attempt was made at separating the contributions of the biological and the electrochemical processes to the total H2 yield. The green dashed curve in Fig. 6a represents the theoretical volume of H2 that would be expected from the electrochemical reactions on the basis of the electric charge mobilized. If the biological and the electrochemical processes were additive, such a volume would add up to the volume generated by the biochemical process alone. According to such hypotheses, a total yield of 63.4 NL H2/kg TOC would be expected, that is 19–24% lower than the actual value obtained for the IBES. It is thus tempting to hypothesize that, further to the additional volume of H2 produced by chemical reduction of protons, the electrochemical process exerted a synergistic effect on the fermentation reactions, enhancing H2 generation associated to the biochemical metabolic pathways. While such a result could in principle be related to a pH buffering effect caused by the conversion of protons to H2, this was in fact found to be relatively minor (see the pH profiles in Fig. 6b). Other mechanisms are thus claimed to have played a role during the process, including changes in the redox potential of the fermentation medium caused by the electricity flow. Concerning this, several studies on BESs, and more specifically on electro-fermentation, have demonstrated that fermentation pathways can be electrically enhanced if improved electron transport routes and energy conservation mechanisms in biomass take place. It is recognized that changes in the extracellular redox potential (ORP) may affect the intracellular redox homeostasis and metabolism of microbial cells. As a consequence, changes in fermentation metabolites (Liu et al. 2013) can occur. This concept can be advantageously exploited in anaerobic processes, because the standard ORP for the redox pair O2/H2O (E0 =  + 820 mV) is the highest among the other pairs associated with intracellular metabolism (i.e., H+/H E0 = – 420 mV; NAD+/NADH E0 = – 320 mV; NADP+/NADPH = – 315 mV). Therefore, with no O2 in the system, electrons can be more easily accepted by intermediate metabolites. Possible ways through which electron supply may affect the fermentation process include NADH reduction and may also promote the production of additional ATP (Schievano et al. 2016). Toledo-Alarcón et al. (2019) showed that even a small amount of electrons can have a significant effect on the metabolic pathways, affecting H2 production (with an improvement in H2 yield by 1.8–2.5 times compared to conventional DF), almost irrespective of the applied potential. The authors suggested that small changes in ORP can affect the regulation of hydrogenase involved in H2 production, due to its high sensitivity to ORP variations.

The enhancement of the fermentation reactions when coupling the biological process with an electrochemical system was also reported by Paiano et al. (2019), who observed an increased amount of metabolic products when exogenous redox mediators were used.

Table 6 shows the concentration of the metabolic products in the cathodic and anodic compartments of the IBES and the stand-alone DF reactor. These mainly included HAc, HBu and EtOH, although EtOH concentration did not differ considerably from the initial value. It can be observed that the concentrations of HBu and the total metabolic products were higher in the IBES than in the stand-alone DF. The higher amounts of metabolic products measured in the IBES are consistent with its larger observed H2 generation. While clarifying the electro-stimulation of the fermentation pathways was beyond the scope of the present study, the results yet show that the integration of the biochemical and electrochemical processes produced larger than additive effects on H2 generation.

Table 6 Metabolic products at the end of IBES (run D-Ti-CW) and stand-alone DF
Full size table

The data in Table 6 also show that unlike the tests with the model acid solutions, the extent of migration of the organic metabolites towards the anode was only minor. This was due to the larger complexity of the chemical composition of CW compared to the model solutions used in the previous tests, with other anionic species (presumably with smaller ionic radiuses) being likely responsible for ensuring the electroneutrality balance.

As mentioned above, the buffering effect associated with proton consumption in the IBES was mainly visible during the first stages of the process (see pH profiles in Fig. 6b), but was not capable of contrasting the progressive acidification of the cathode solution caused by the accumulation of volatile fatty acids produced during fermentation. Towards the end of the test, pH approached a value of ~ 3.5 for both tests, which was mirrored by a drop in the current flow. The unexpected decline in current intensity after ~ 45 h was probably related to a passivation effect of the anode likely caused by the precipitation of Zn(OH)2 onto its surface (also evident upon visual inspection).

Conclusions

The purpose of this preliminary work was exploring the feasibility and performance of a novel IBES which is aimed at improving the H2 yield during DF through protons reduction into H2 mediated by a self-generated electron flow, with no need for external energy supply (as otherwise required by other conventional systems such as MECs) and taking advantage from the coupling of the biochemical and electrochemical components of the system.

Preliminary tests were carried out through different configurations using a model solution to assess the electrochemical process. They highlighted the importance of optimizing the reactor design in order to reduce internal resistances and provide an appropriate H2 generation rate along with electricity production. According to the electric current production that was measured and the flow of dissociated organic acid through the AEM, it was concluded that the model solution was almost completely (90–100%) dissociated and the protons were virtually fully converted into H2.

The experimental tests on the bio-electrochemical integration were performed through DF of CW, without inoculum, in the optimized reactor configuration. Under this condition, IBES achieved a H2 yield of 75.5–78.8 NL H2/kg TOC that, compared to the stand-alone DF (22.4 NL H2/kg TOC) shows a 3-times improvement. When converted into an energy yield assuming a heating value of ~ 140 kJ/g H2, these figures correspond to an energetic equivalent of ~ 940−1000 kJ/kg TOC; conversely, the energy associated with the amount of electrons mobilized during the process represents a minor additional contribution (~ 6.6 kJ/kg TOC) to the total energy output of the process. The scale at which the process was tested in the present study does not allow deriving energetic/economic considerations about the industrial-scale implementation of the IBES. However, it can be inferred that, for the process to be energetically sustainable, the additional energy requirements associated with the electrochemical compartment (i.e. mixing of the anodic chamber, indirect energy consumption for the electrodes manufacturing, energy demand for treatment of the final ZnSO4 solution) must be well below the above-mentioned gross energy gain of 940−1000 kJ/kg TOC. While a comprehensive assessment of the energetic profile of the IBES was beyond the scope of the work, this issue is certainly of extreme relevance in view of a thorough assessment of the feasibility of the process.

Based on the discussion provided about the IBES, the results appear to be promising. However, to derive a better understanding of the process, future aspects need to be investigated in further detail; these include the potential electrochemical stimulation effects on biomass and the strategies to prevent the passivation of the anode, as such phenomena likely played a role in the experiments performed.