A previously enriched chemolithoautotrophic mixed microbial culture dominated by Acetobacterium spp. was used as the inoculum source (Roy et al., 2021). For the MES experiments, a minimal medium with pH 7 was used. It contained K₂HPO₄ (5.35 g/L), KH₂PO₄ (2.62 g/L), NH₄Cl (0.25 g/L), KCl (0.5 g/L), CaCl₂⋅2H₂O (0.15 g/L), MgCl₂⋅2H₂O (0.6 g/L), trace metal solution (1 ml/L), selenium-tungstate solution (1 ml/L), vitamin solution (2.5 ml/L), 2-bromoethanesulfonate (6.4 g/L), Na₂S⋅9H₂O (0.3 g/L), and resazurin (0.5 ml/L from a 0.1% stock solution) (Roy et al., 2021). It was cultivated and maintained with CO₂ and H₂ (H₂:CO₂ ratio was 4:1) as the sole carbon and energy sources, respectively, under anaerobic conditions at 28 ± 2°C. An active culture was inoculated in the cathode chamber of MES reactors to have 0.25 OD₆₀₀ in suspension during the start-up phase.

The biogas was collected from an operational and well-maintained anaerobic sewage treatment plant (Phase 3, Panchkula, Haryana). A vacuum pump (Model N86 KT.45.18; KNF pump) was used to collect biogas in 6 L passivated SilcoCan steel canisters (Restek, United States). A Teflon gas tubing was used to collect the biogas as it can withstand high temperature (>180°C) and pressure (>40 psi). A PTFE filter (0.2 µm) was used in the front of the tubing to restrict the particulate matter from the biogas from entering the canister. At the same time, the moisture content was minimized with the help of a moisture trap installed in gas tubing where magnesium perchlorate was used as the moisture-absorbing reagent. The biogas samples were analyzed by GC with TCD (Agilent 490 Micro GC, Analysis and Calculations). The major components of biogas were methane (60 ± 2%), CO₂ (28 ± 5%), and N₂ (10 ± 2%). A trace amount of H₂S was also present in the sample. 1 ± 0.1% O₂ was also observed, which was most likely present due to manual sampling error.

Custom-made double-chambered glass reactors with an empty bed volume of 700 ml (350 ml each chamber) were used for this study. The catholyte was the same as mentioned in Microbial Inoculum Source and Cultivation Conditions and the anolyte was 0.5 M Na₂SO₄ with pH 2.5 (adjusted with 1 M H₂SO₄). The experiments were conducted in three-electrode configuration mode under potentiostatically controlled conditions (VSP300, BioLogic Science Instruments, France). The working volume was 250 ml. A graphite plate with 10 cm² and dimensionally stable mixed metal oxide coated titanium plate with 7.5 cm² projected surface area were used as working (cathode) and counter (anode) electrodes, respectively. The anode and cathode chambers were separated by a 117 Nafion proton exchange membrane (Sigma-Aldrich). Water oxidation/oxygen evolution and CO₂ reduction reactions take place in the anode and cathode chambers, respectively. The cathode material and proton exchange membrane were pretreated as described elsewhere (Roy et al., 2021). All MES experiments were performed in duplicate at an incubation temperature of 28 ± 2°C. The performance of the duplicate reactors for reported parameters was reproducible and very close; therefore average data has been presented in all the figures. The electrode potential data are reported against Ag/AgCl (3.5 M KCl) reference electrode (0.205 V vs. SHE). The experimental setup and the reactions in MES reactors are illustrated in Supplementary Figure S1. Two types of control experiments were conducted. These include biotic open circuit potential (OCP) experiment (with all media components and carbon source but not electrically connected, i.e., no electron/energy source), and abiotic connected (with all the components and electrically connected but uninoculated) experiment.

Activation polarization was performed at −0.6 V before inoculation or starting the main experiment for electrochemical activation of the electrodes, membrane, and electrolyte. A constant potential of −1.0 V vs. Ag/AgCl was applied at the cathode to facilitate H₂-based bioproduction by using the chronoamperometry (CA) technique. The reduction current response was monitored at a constant time interval (5 min). Cyclic voltammetric analysis was conducted for two cycles within a potential window of −1.4 and −0.2 V at a constant scan rate of 1 mVs⁻¹ at different experimental conditions, i.e., before inoculation, after inoculation, and at the end of each biogas feed condition. The data of the second cycle was used for analysis.

In the first phase, 30 ml of CO₂ containing biogas (with the composition stated in Biogas Sampling and Characterization) was added to the catholyte daily to establish a stable performing biocathode. The CO₂ present in the biogas got dissolved in the form of bicarbonate at pH 7. This was the sole carbon source for the microbes. A gas displacement assembly was attached to the reactor’s outlet to maintain the pressure condition inside the reactors. After the completion of two batch cycles, the biogas was fed in a continuous flow mode at different feed rates, viz., 0.5, 0.3 and 0.2 ml/min using the mass flow controllers (MFCs, Alicat/MCM-100SCCM-D) in the second phase. The MFC calibration was carried out with the air by the manufacturer (Alicat Scientific). The correction factors were applied for different gases depending on their viscosity, density, and compressibility. The standards used for calibrations were NIST (National Institute of Standards and Technology) traceable. The volumetric flow of the off-gas from MFCs was measured to confirm the used flow rates with the help of a water displacement assembly, and it was found to be the same. The volumetric flows of the off-gas from the microbial electrosynthesis reactors were also measured intermittently at each experimental condition and were found to be marginally lower than the applied feed rates. For instance, in the case of 0.5 ml/min biogas feed rate, the off-gas flow was about 0.47 ml/min. Similar trends were observed in other gas feed rates. The effect of different feed rates was evaluated on methane content increase as well as acetic acid production after acclimatization of the reactors for at least 2 days at each feed rate.

The bulk phase and gas samples from MES reactors were taken at a regular interval of 2 days. To assess the gas composition in the headspace and to know the methane concentration, the gas analysis was done by GC-TCD (Agilent 490 Micro GC) equipped with three channels for different gases [Channel 1: Column- Molecular sieve for H₂, carrier gas- Ar; Channel 2: Column- Molecular sieve for O₂, N₂, CO and CH₄, carrier gas–He; Channel 3: Column- Pora plot U for CO₂ and H₂S, carrier gas–He]. The bulk phase organics (C1-C4 organic acids and alcohols) analysis by HPLC (Agilent 1260 Infinity II, RID Detector, Hiplex H column, 5 µM H₂SO₄ as mobile phase, flow rate 0.5 ml/min, Temperature 50°C) was performed to evaluate the efficiency of microbes to produce organics from CO₂ in biogas. Additionally, monitoring of microbial growth via OD measurements at 600 nm (Photo-lab 7600 UV-VIS spectrophotometer) and pH (Oakton PC2700) in the bulk phase was done. Different parameters, such as product titer, production rates (normalized by cathode surface area and catholyte volume), coulombic efficiency (electron recovery into organic products), carbon recovery efficiency (Das et al., 2018) and energetic efficiency were calculated as described in supplementary section S1 (Patil et al., 2015b). Since the methane content increase was not directly related to energy consumption in this study, we did not consider it in CE and EE calculations.

To understand the microbial communities developed at the biocathode and in the bulk phase, V3-V4 16S amplicon sequencing-based analysis was conducted. For this, the genomic DNA of the inoculum source and from two duplicate MES reactors named R1 and R2 (biocathode and bulk phase) was isolated on the completion of the electrosynthesis experiments using QIAGEN DNeasy PowerSoil Pro Kit. Further DNA quantification was done by Nanodrop (Genova Nano—4359). Agarose gel electrophoresis (1%) was also performed to assess the integrity of the isolated genomic DNA. Since V3-V4 regions of 16S rRNA are the most conserved regions, they were amplified using specific V3 Forward (341F: 5′-GCCTACGGGNGGCWGCAG-3′) and V4 Reverse (805R: 5′-ACTACHVGGGTATCTAATCC-3′) primers. Again 2% agarose gel electrophoresis was performed to analyze the amplified products. Later, Illumina MiSeq 2,500 platform at Eurofins Genomics India Pvt. Ltd. (Bangalore, India) was used for amplicon sequencing. Trimmomatric v0.38 was used for trimming the adaptor and primers sequences. FLASH v1.2.11 was run for alignment as well as selecting the operational taxonomic unit (OTU). At last, the OTU products were run through the Silva database to find the most similar OTU hits using the Qiime1 pipeline and R packages for visualization. The raw amplicon sequences have been submitted to NCBI short read archive under bio-project PRJNA659908.

In order to form an efficient biocathode for CO₂ utilization, the biogas was first supplied in a fed-batch mode. Since the enriched mixed culture contained chemolithoautotrophic microbes (Roy et al., 2021), immediate organics production, in particular, acetic acid, was detected from the start of the experiment (Figure 1).

Since the inoculum source contains mainly Acetobacterium spp. that follow the Wood-Ljungdahl pathway for CO₂ fixation, microbial growth and acetic acid concentration are directly correlated. The maximum titer achieved was 0.6 ± 0.1 g/L with a production rate of 0.06 ± 0.001 g/L/d (Figure 1). Along with the acetic acid production increase in the methane content was also evident. As a result of microbial CO₂ utilization from biogas, the methane concentration increased from 61 ± 3% to 71 ± 2% (Figure 1). These observations are close to the previously reported data on MES from biogas (Das and Ghangrekar, 2018). The volumetric measurement of the off-gas was not performed regularly in this study. It is important to take into account the changes in the off-gas volume during continuous experiments for the optimization and development of such processes. The batch mode operation and consistent acetic acid production for two consequent cycles suggested biocathode development with biogas feed. No microbial growth and acetic acid production were observed in control MES reactors (Supplementary Figure S2).

After developing stable performing biocathodes in a fed-batch mode operation (Biocathode Development Through Batch Mode Operation), biogas was fed at different feed rates in a continuous mode. In all cases, instant microbial growth and acetic acid production were observed (Figure 2), implying successful acclimatization of the biocathode (Marshall et al., 2012). A trace amount of formic acid (up to 0.120 g/L) was also detected. As microbial growth is directly linked with organics production, the OD and acetic acid titer data correlated well (Ragsdale, 2008). Compared to the batch mode operation, the acetic acid production was considerably higher in the case of continuous feed operation. It is mainly due to the unlimited and more CO₂ supply to the microbes in the continuous mode than a limited and less CO₂ availability in the batch mode operation. With the highest biogas feed rate of 0.5 ml/min, the maximum acetic acid titer reached to 1.5 ± 0.5 g/L with a volumetric production rate of 0.107 ± 0.02 g/L/d and a cathode surface area-based rate of 27 ± 9 g/m²/d after 14 days (Figure 2A). The acetic acid production was leveled off after 10 days. For 0.3 ml/min feed rate, the maximum acetic acid titer was 2.5 g/L with a production rate of 0.17 ± 0.3 g/L/d and a cathode surface area-based rate of 46 ± 7 g/m²/d (Figure 2B). The maximum acetic acid titer of 3.4 g/L was achieved with the lowest feed rate of 0.2 ml/min (Figure 2C). It was produced at 0.23 ± 0.04 g/L/d volumetric and 60 ± 11 g/m²/d cathode surface-based rates. The acetic acid titer tends to increase with the decrease in biogas feed rates based on a 2-fold higher titer at 0.2 ml/min feed rate than 0.5 ml/min. This can be attributed to the retention and thereby microbial utilization of both CO₂ and H₂ (produced at the cathode) in reactors for a longer duration at the lower feed rates.

The methane concentration in the off-gas was analyzed at a regular time interval in a continuous mode operation. No methane upgradation was observed in the case of control MES reactors (Supplementary Figure S2). Irrespective of the feed rates, an increase in methane content was prominent in all cases (Figure 3). Since microbes utilized CO₂ present in biogas as the sole carbon source, the methane content started to get enriched from the very beginning of the reactor operation at different feed rates.

At the highest feed rate, the methane concentration increased from 62 ± 3 to 82 ± 2% on the 4^th day (Figure 3A) and remained within that range thereafter. A maximum methane concentration of 84 ± 2% was observed on the 14^th day. The hydrogen concentration was mostly about 12% in the off-gas. Most of the hydrogen was getting flushed out in the off-gas due to the higher feed rate. A significant increase in the methane content from 55 ± 6 to 90 ± 4% was seen at a 0.3 ml/min feed rate after the 10^th day (Figure 3B). Hydrogen was mostly utilized in this particular case and was detected when methane concentration was lower. The highest methane concentration was observed with the lowest feed rate after the 12^th day (from 62 ± 4 to 93 ± 3%) (Figure 3C). After that, the methane concentration remained almost at the same level. The hydrogen concentration remained at a low level (<7%) in this case. The methane concentration enhancement was pronounced in the case of both 0.3 and 0.2 ml/min feed rates. This can be justified by the fact that at lower feed rates, microbes are able to utilize most of the carbon source for growth and acetic acid production (MES Experiments in a Continuous Mode Operation). With the goal of futuristic green fuel, a combination of hydrogen and methane can be more promising (Nanthagopal et al., 2011).

In the case of control MES reactors, the reduction current was almost negligible (Supplementary Figure S3). A significant increase in reduction current was observed with all feed rates, and it was maximum for the lowest feed rate. These observations align well with the acetic acid production data (Supplementary Section S4, Supplementary Figure S4). Cyclic voltammograms (CVs) were recorded at different experimental conditions. The blank CV was performed before inoculation, and the inoculated and biocathode CVs were performed just after microbial source inoculation and at the end of the cycle, respectively. No redox peaks in the blank CV indicate the absence of any redox-active moieties at the cathode surface or in the catholyte in the uninoculated condition (Figure 4). By comparing blank CV with biocathode CV, it can be seen that the H₂ evolution potential of biocathode is higher than the bare cathode (−0.95 V). For different feed rates of 0.5, 0.3 and 0.2 ml/min, the H₂ evolution potentials are −0.92, −0.77 and −0.76 V. The shift is most prominent (∼0.2 V) in case of the lower feed rates (Figure 4C). This observation aligns well with the acetic acid production and methane upgradation data (Figures 2, 3).

The shift in H₂ evolution potential suggests lowering of the H₂ evolution over-potential at the biocathode due to the presence of microbes. Moreover, the reduction current drawn at any cathode potential is higher in the case of biocathode than the abiotic cathode due to the presence of microbial catalysts. For instance, at −1 V, the current density (in mA/cm²) of biocathodes are −0.3 (for 0.5 ml/min), −0.4 (for 0.3 ml/min), and −1.2 (for 0.2 ml/min) in comparison to −0.08 mA/cm² of the bare cathode. Similar findings have been reported for CO₂ reducing biocathodes earlier (Jiang et al., 2013; Jourdin et al., 2015a; Patil et al., 2015a; Bajracharya et al., 2015). The improved electrocatalytic activity and decreased H₂ evolution overpotential can be attributed to microbial activity at the cathode of the MES systems (Rozendal et al., 2008; Jourdin et al., 2015a; Patil et al., 2015a).

Acetobacterium, Desulfovibrio, Sulfurospirillum, and Lentimicrobium genera dominated the mixed culture (Figure 5). The relative sequence abundance of different microbial genera changed according to conditions. For example, the relative abundance of Acetobacterium increased in the MES reactors (38 ± 3%) compared to the inoculum source (9%). The relative sequence abundance of Sulfurospirillum was prominent in the inoculum source (16%) and bulk phase of reactors (12 ± 1%) but less at the cathodes (3 ± 2%). In the MES reactors, Acetobacterium spp. was the most dominant genus with a relative abundance of 18 ± 2% in the bulk phase and 19 ± 2% on the biocathode. This genus is known as acetogens for its H₂ and CO₂ fixing abilities into acetic acid (Ragsdale, 2008). Its primary function is to fix CO₂ through the Wood-Ljungdahl pathway for energy conservation and synthesize acetyl-CoA and biomass (Labelle et al., 2020). This genus played a likely role in acetic acid production in this study (MES Experiments in a Continuous Mode Operation). Another dominant genus Sulfurospirillum is microaerophilic, which means that they can scavenge small amounts of O₂ that may get transferred from the anode to the cathode chamber. Thus anaerobic conditions in the MES reactors can be maintained (Goris et al., 2014). As they can oxidize acetic acid using O₂ as the electron acceptor, the acetic acid titer may decrease slightly (Labelle et al., 2020). As there are more chances of O₂ presence in the bulk phase, its abundance is higher in the bulk phase compared to the cathode.

Desulfovibrio was one of the dominant genera in this study, with a relative abundance of 14%, 12 ± 1%, and 6 ± 2% at the inoculum source, bulk phase, and biocathode, respectively. These sulfate reducers are capable of direct electron transfer (Aulenta et al., 2012). They can also facilitate H₂ and formate production with the help of cytochromes, hydrogenases, and formate dehydrogenase (Labelle et al., 2020). These can then be utilized as an energy source by the acetogens present in the MES system. Lentimicrobium genus was found to be dominant in the MES reactors. In the inoculum source, its relative abundance was only 0.5%. Its relative abundance increased to 13 ± 1% and 10 ± 1% in the MES reactor bulk phase and biocathode, respectively. It is a genus of strict anaerobic bacterium previously reported in biogas upgradation studies (González-Cortés et al., 2021). It has also been reported for acetic acid production under certain conditions (Sun et al., 2016). Apart from these genera, Pseudomonas was also present at a low relative abundance of up to ∼3.5 ± 0.5% in bulk phase and ∼1.5 ± 1% in biocathode. It is known for hydrogen production via electron transfer through redox mediators as well as hydrogenase enzymes and has been reported in bioelectrochemical reactors (Mateos et al., 2019). Some other bacterial genera, namely, Pirellula, Sphaerochaeta and Enterobacter, were also observed in very low abundances, whose role is not very clear in MES systems.