Customizable design strategies for high-performance bioanodes in bioelectrochemical systems.

Bioelectrochemical systems (BESs) can fulfill the demand for renewable energy and wastewater treatment but still face significant challenges to improve their overall performance. Core efforts have been made to enhance the bioelectrode performance, yet, previous approaches are fragmented and have limited applicability, unable to flexibly adjust physicochemical and structural properties of electrodes for specific requirements in various applications. Here, we propose a facile electrode design strategy that integrates three-dimensional printing technology and functionalized modular electrode materials. A customized graphene-based electrode with hierarchical pores and functionalized components (i.e., ferric ions and magnetite nanoparticles) was fabricated. Owing to efficient mass and electron transfer, a high volumetric current density of 10,608 ± 1,036 A/m3 was achieved, the highest volumetric current density with pure Geobacter sulfurreducens to date. This strategy can be readily applied to existing BESs (e.g., microbial fuel cells and microbial electrosynthesis) and provide a feasibility for practical application.

Introduction

The accelerating modern consumption of traditional fossil energy and concomitant environmental deterioration have resulted in an urgent need for alternative energy sources and technologies (Rahimnejad et al., 2011). Bioelectrochemical systems (BESs) are emerging renewable technologies that can convert energy through the metabolism of electroactive bacteria (Gul and Ahmad, 2019; Logan et al., 2019), which have been exploited for diverse technologies, such as renewable energy generation (Sun et al., 2008), wastewater treatment (Du et al., 2007), biosensors (Kumlanghan et al., 2007), and bioremediation (Morris and Jin, 2007).

Despite several attempts enhance the performance of BESs, there remains a clear discrepancy between this prospective technology and real-world applications due to their relatively low power density output (Santoro et al., 2017). Apart from a vast array of factors (e.g., electroactive bacteria species, reactor architectures, ion exchange membranes, and substrates), bioelectrodes are considered the key factor in determining BES performance (Slate et al., 2019). Bioelectrodes serve as a platform where electroactive bacteria can exchange electrons, while providing a suitable environment for bacterial colonization and proliferation (Palanisamy et al., 2019). Therefore, the physicochemical properties and structure of electrodes significantly influence the electron transfer at the biological/inorganic interface and determine the maximum available surface area for electroactive bacteria attachment and growth. Previous studies have focused on high-performance electrodes with good biocompatibility and conductivity (Liu et al., 2020; Palanisamy et al., 2019; Rusli et al., 2019; Santoro et al., 2017; Slate et al., 2019), such as carbon cloth, graphite rod, carbon sponges, and metal oxide foams with conductive coatings (e.g., polypyrrole and poly (3, 4-ethylenedioxythiophene)) (Lv et al., 2013; Wang et al., 2013). Among them, porous electrodes are particularly promising, owing to their large surface area available for the adhesion and proliferation of bacteria. Nevertheless, pore clogging frequently occurs when using porous electrodes during the development of biofilms (Chong et al., 2019). Consequently, the transport of substrates and waste products inside 3D electrodes is largely restricted, resulting in ineffective pore utilization (i.e., an inert zone without bacteria) inside the electrodes, hindering performance improvement, contrary to the original intention of utilizing the large surface area of 3D porous electrodes. The random, long-distance, zigzag ion migration paths of the disordered electrode porous structure could hinder the rate of substrate transport and waste product removal, despite the biofilm not being completely clogged (Chen et al., 2012; Sun et al., 2017).

Three dimensional printing technology, which enables the fabrication of customizable 3D electrodes with expected geometric structures and channel sizes, can be used to explore a desired 3D electrode (Theodosiou et al., 2018, 2020). Recently, Bian et al. constructed a 3D-printed porous carbon anode using a carbonized UV-curable polymer resin with enhanced power generation in microbial fuel cells (MFCs) (Bian et al., 2018). Freyman et al. incorporated living electroactive bacteria (Shewanella Oneidensis) into 3D printing ink with alginate and carbon black to implement a 3D-printed bacterial structure as a living electrode for MFC (Freyman et al., 2020). These studies demonstrate the potential of fabricating electrodes via 3D printing technology; however, the low electron transfer rate at the biological/inorganic interface and the relatively low specific surface area available for bacterial adhesion limits power density. Therefore, the development of a high-performance bioelectrode with excellent biocompatibility, good conductivity, and optimized architecture for effective mass transfer is of great significance.

Herein, we demonstrate, to the best of our knowledge, for the first time a high-performance bioelectrode with designed 3D structures was fabricated by an adjustable printing ink using an extrusion-based 3D printing method. The customized ink for bioelectrode preparation consisted of well-dispersed graphene, ferric ions, and magnetite nanoparticles. Specifically, graphene oxide (GO) was selected as the primary component of the 3D printer ink, as its nanostructure can facilitate the electron transfer at the biological/inorganic interface (Li et al., 2012), and provide a large surface area for the adherence and proliferation of bacteria (Ren et al., 2016). Ferric ions and magnetite nanoparticles were incorporated as modular functional components into the 3D printer ink to promote bacterial growth and interfacial electron transfer (Cruz Viggi et al., 2014; Liu et al., 2018). Using extrusion-based 3D printing technology (Figure 1A) followed by freeze-drying for further pore formation, we obtained an integral composite GO aerogel electrode with hierarchical pores (macropores and micropores) (Figure 1B). The periodic macropores (>500 μm) of 3D printing frame structure, whose sizes were determined by the distance between the two nearby printed wires, were designed to provide open channels that can offer highly effective supply of nutrients and dissipation of waste inside and outside the electrode. The micropores (<100 μm) derived from graphene network on each printed wire significantly increased the bacterial colonization and proliferation surface area and can promote permeation of the substrate. Moreover, the non-conductive GO electrode could be converted in situ into a conductive graphene anode via an electroactive bacterial respiration process (Figure 1C) (De Silva et al., 2017). In this process, only Geobacter sulfurreducens were used for GO reduction, eliminating extra chemical conversion processes. Our work provides a facile strategy for an easily controllable, flexible, and highly producible bioelectrode fabrication, opening a new path to enhance the performance of BESs and related microbial electrochemical technologies.

Figure 1

Schematic illustration of 3D printing process and microbial GO reduction

(A) The customized GO-based ink was printed into a designed structure.

(B) Schematic illustration of the 3D-printed electrode hierarchical pores.

(C) Schematic illustration of the microbial GO reduction process.

Results

Morphology and characterization of 3D-printed GO aerogel electrodes

The electrode ink must possess a shear-thinning behavior to enable continuous extrusion and a suitable modulus to hold the expected structure. The ferric ions added to the customized ink can act as cross-linkers, as they can form electrostatic interactions and coordination bonds with negatively charged GO sheets (i.e., their carboxylic and hydroxyl functional groups (Jiang et al., 2018; Zhao et al., 2017). The resultant GO hydrogel inks showed a higher apparent viscosity with a clear shear-thinning behavior and higher storage modulus magnitudes, in comparison with pure GO solutions (Figure S1, Supplemental Information (SI)). Following the layer-by-layer 3D printing process and freeze-drying process, 3D-printed GO aerogel electrodes (1.6 cm in width, 2 cm in length, and 4 mm in thickness) with a lattice structure were obtained, as shown in Figure 2A. The open channel structure was easily observed from the top and the side (Figure S2, Supplemental Information). Figures 2B and 2C show representative scanning electron microscope (SEM) images of the 3D-printed GO aerogel electrode. It is obvious that the GO aerogel electrode possesses hierarchical pores, including macropores (∼hundreds of micrometers, up to the mm scale) between the two neighboring printed wires and micropores (∼tens of micrometers) on each wire of the aerogel. The macropores are designed to enable efficient mass transfer of substrates, while micropores enhancing the specific surface area, and promote the attachment and growth of electroactive bacteria. The strength test (Figures 2D and 2E) shows that the GO aerogel electrode can support 50 coins (∼160 g weight) without a clear shape change and damage. The X-ray diffraction (XRD) electrode pattern (Figure S3, Supplemental Information) shows the characteristic peaks at around 10.9° that were indexed to GO. In addition, the characteristic diffraction peaks at diffraction angles of 30.06°, 35.32°, 42.98°, 53.37°, 56.87°, and 62.51° (Figure 2F) correspond to the (220), (311), (400), (422), (511), and (440) crystal surfaces, respectively, which can be indexed to Fe3O4 nanoparticles (JCPDS 75–1609) (Zhao et al., 2010). X-ray photoelectron spectroscopy (XPS) analysis revealed the presence of Fe, C, and O elements inside the GO aerogel electrode (Figure S4A, Supplemental Information). In the high-resolution XPS spectra of Fe 2p (Figure S4B, Supplemental Information), the GO aerogel electrode clearly exhibited five peaks: 2p3/2 of Fe (III) ions, 2p1/2 of Fe (III) ions, 2p3/2 of Fe (II) ions, 2p1/2 of Fe (II) ions, and their satellite peaks, indicating the coexistence of Fe (II) and Fe (III) and hence the presence of Fe3O4. Energy-dispersive spectrum mapping (Figure 2G) showed a uniform distribution of iron corresponding to the SEM image (Fig C).

Figure 2

Morphology and characterization of 3D-Printed GO aerogel electrodes

(A) Photographical image of a typical GO aerogel electrode.

(B and C) SEM images of the hierarchical porous structure.

(D and E) Strength test of the GO aerogel electrodes pressed by 50 coins (∼160 g weight).

(F) XRD analyses of GO aerogel electrodes.

(G) Energy dispersive spectrum elemental map of iron for GO aerogel electrode corresponding to the SEM image.

In situ microbial reduction of 3D-printed GO aerogel electrodes

The low conductivity of GO, caused by the destruction of its π-conjugated structure, does not meet the requirements of an electrode for efficient electron transfer and electron conduction. Therefore, the 3D-printed GO electrodes must be reduced to rGO electrodes with high conductivity. Recent studies have shown that several electroactive bacteria (e.g., Shewanella (Salas et al., 2010)) are capable of directly utilizing GO as the extracellular electron acceptor for the reduction of GO to rGO. The microbially reduced graphene possessed better biological properties (e.g., biocompatibility) than chemically reduced graphene, exhibiting significantly enhanced microbial electricity generation owing to the formation of bacteria/graphene scaffolds (Fan et al., 2014; Virdis and Dennis, 2017). Herein, the GO electrode was microbially reduced to graphene by keeping bioelectrochemical reactors under open circuit conditions following G. sulfurreducens inoculation. As shown in Figure 3A, the potential of the GO electrode decreased with operating time, suggesting the colonization and growth of G. sulfurreducens on the surface of the GO electrode (Geelhoed and Stams, 2011), and the concomitant reduction of GO to rGO by G. sulfurreducens. After 48 hr of inoculation, rGO formation and its interaction with bacteria were investigated by chemical tests.

Figure 3

Characterization of microbial reduced graphene

(A) Variation of the potential (vs. Ag/AgCl) of a 3D-printed rGO aerogel electrode with time at an open circuit condition after the inoculation with G. sulfurreducens.

(B) XPS spectra of the 3D-printed rGO electrode before (0 hr) and after (48 hr) the microbial reduction by G. sulfurreducens.

(C) FTIR spectra of G. sulfurreducens, and the 3D-printed rGO electrode before (0 hr) and after (48 hr) the microbial reduction.

Figure 3B shows the high-resolution C1s XPS spectrum of the rGO electrodes. Before the inoculation of G. sulfurreducens, four peaks at 284.8, 285.9, 286.9, and 288.2 eV were clearly resolved by fitting experimental data. The peak at 284.8 eV is a typical C-C bond value in C1s, which exists in graphene and graphite. The peaks at 285.9, 286.9, and 288.2 eV are consistent with the C1s binding energy in C-OH, epoxy (C-O), and C=O, respectively (Lei et al., 2012; Mai et al., 2011). In contrast, the intensity of the C-C peak increased following cultivation for 48 hr, indicating that the sp2 carbon character of graphite was restored, and the GO was successfully reduced during the microbial reduction process. The intensity of the peaks corresponding to oxygen-containing functional groups of the electrodes became much weaker, while the peak at 285.8 eV (corresponding to C-N, C-OH) increased following the microbial reduction process. This can possibly be ascribed to the formation of C-N, whose binding energy was close to that of C-OH. As shown in the high-resolution N1s (Figure S5B, Supplemental Information), a peak at 399.9 eV (assigned to C–N) was observed (Zhao et al., 2019), further revealing the formation of C–N bonds during the microbial reduction of GO. The results of the Raman test (Figure S6, Supplemental Information) also confirmed the successful reduction of GO, as the intensity ratio value (1.08) of the disorder D peak to the graphitic G peak (ID/IG) increased in contrast to that of GO (0.88), which can be ascribed to the presence of unrepaired defects following the removal of partial oxygen groups. In addition, the XPS spectrum of the rGO electrode interior also showed an increased C-C peak intensity at 284.8 eV post-cultivation for 48 hr (Figure S7, Supplemental Information), indicating an efficient GO reduction of the entire electrode.

To further investigate the reduction of GO electrodes and the colonization of G. sulfurreducens, we performed an Fourier transform infrared (FTIR) analysis of GO, rGO, and G. sulfurreducens. The spectrum (Figure 3C) of the original GO sample showed a number of peaks corresponding to oxygenated functional groups, including the alkoxy C-O (1,078 cm−1), epoxy C-O (1,229 cm−1), C=O (1,729 cm−1), and O-H peak (3,399 cm−1). In contrast, C-O, C=O, and O-H oxygenated functional groups significantly decreased after the microbial reduction process. The G. sulfurreducens sample exhibited special peaks, such as the 1,541 cm−1 region (closely corresponding to the N-H and C-N from amide) and the region at around 1,656 cm−1 (associated with the C=O from amide). Notably, the FTIR spectrum of the rGO electrode showed a comprehensive characteristic line, resulting from GO and G. sulfurreducens samples, reflecting the colonization of G. sulfurreducens on the surfaces of the rGO electrode. The potential source for these amide groups in the rGO sample was probably G. sulfurreducens, as well as their extracellular polymeric substances on the surface of the electrodes, which were in accordance with those of previous studies (Chen and Chen, 2015; Lee and Hur, 2016). Moreover, the electroactive accessible surface area (ECSA) of rGO aerogel and carbon felt anode were estimated by the cyclic voltammetry (CV) with K3[Fe(CN)6] solution (Figure S8, Supplemental Information). The rGO aerogel anode exhibited the highest peak current density of 90.79 A/m2, with a corresponding ECSA of 18.31 cm2/cm2 (subject to the surface area of the electrodes), which was about four times larger than that of carbon felt and much higher than that of the modified carbon cloth in previous studies (Table S1, Supplemental Information). The results indicated that the 3D-printed rGO aerogel electrode can provide more electrochemically active sites to accept electrons from G. sulfurreducens.

BES performance with customized composite electrodes

To evaluate the bioelectrochemical performance of the customized composite electrodes, four kinds of rGO aerogel electrodes were fabricated by altering the composition of the printing ink with/without the addition of ferric ion or magnetite nanoparticles, represented by rGO, rGO/Fe3+, rGO/Fe3O4, and rGO/Fe3+/Fe3O4 aerogel electrodes, respectively. Following the open circuit-based microbial reduction of GO, all the electrodes were poised at 50 mV vs. Ag/AgCl for current generation.

Figure 4A illustrates the current generation of bioelectrochemical reactors using rGO, rGO/Fe3+, rGO/Fe3O4, and rGO/Fe3+/Fe3O4 aerogel, as well as carbon felt as anodes. A repeatable current generation was obtained 13 days post cultivation, indicating successful bioanode start-up. The maximum current density of each rGO aerogel bioanode was at least two times higher than that of the conventional 3D carbon felt anode, suggesting the excellent properties of rGO as an electrode material. Notably, the rGO/Fe3+/Fe3O4 aerogel bioanode achieved the highest maximum current generation of 18.99 ± 2.28 A/m2, which was 44.3, 29.4, 13.4% higher than that of the rGO, rGO/Fe3+, rGO/Fe3O4 aerogel anodes, respectively. These results indicate that the combination of graphene, ferric ions, and magnetite nanoparticles all had positive effects on the formation of biofilm or the proliferation of G. sulfurreducens, and there perhaps existed a synergistic effect between ferric ions and magnetite nanoparticles on promoting biocatalysis.

Figure 4

Electrochemical characteristics of carbon felt, rGO, rGO/Fe3+, rGO/Fe3O4, and rGO/Fe3+/Fe3O4 aerogel bioanodes

(A) Current generation under a constant potential of 50 mV vs. Ag/AgCl.

(B) Cyclic voltammetry (CV) under turnover conditions in a PBS solution (50 mM, pH = 7.2) containing 3.38 g/L acetate at 0.1 mV/s (the inset plots show the first derivative of cyclic voltammograms).

(C) Differential pulse voltammetry (DPV) characterized in acetate-free PBS solution with 300 ms pulse width, 50 mV amplitude, 2,000 ms step time, and 0.5 mV step height.

(D) Electrochemical impedance spectroscopy (EIS) analysis characterized at a working potential of 50 mV vs. Ag/AgCl from 100 kHz to 10 mHz at AC amplitude of 10 mV.

Cyclic voltammetry (Figure 4B) of bioanodes under turnover conditions illustrated a similar sigmoidal shape for all the rGO aerogel anodes in the scan range of −0.6 to 0.4 V, demonstrating the catalytic oxidation of acetate by the biofilms on these bioelectrodes in contrast to the abiotic control electrode (Figure S9A, Supplemental Information). Interestingly, the first derivative of the cyclic voltammograms clearly showed that the redox active potential range changed. The rGO, rGO/Fe3+, rGO/Fe3O4, and rGO/Fe3+/Fe3O4 biofilm aerogel anodes exhibited increasing electroactivity over a wider potential range from −0.47 to −0.25 V, while the redox active potential range for carbon felt was narrower from −0.4 to −0.25 V, indicating that the G. sulfurreducens electron transfer could probably be changed using 3D-printed rGO aerogel anodes.

To further investigate the electron transfer, differential pulse voltammetry (DPV)diff was performed in depleted sodium acetate culture medium. Figure 4C shows the different oxidation peaks for the different samples of the bioanodes. Two dominant oxidation peaks at −0.25 and −0.31 V were observed in the carbon-felt bioanode, while two new oxidation peaks (i.e., p3 at −0.43 V and p4 at −0.38 V) appeared in the DPV of the 3D printed rGO aerogel bioanodes. In contrast, no clear redox peak was observed in the DPV of the bare rGO/Fe3O4 aerogel electrode prior to G. sulfurreducens inoculation, suggesting the contribution of G. sulfurreducens to the oxidation peaks (Figure S9B, Supplemental Information). Interestingly, the potential of the p3 and p4 peaks were close to the midpoint potentials of typical cytochrome OmcZ −0.42 V), OmcB (−0.39 V), OmcS (−0.43 V), and periplasmic cytochrome PpcA (−0.37 V) of G. sulfurreducens (Katuri et al., 2010; Zhu et al., 2012). The results indicated that the ferric ions probably have positive effects on the synthesis of c-type cytochromes, which could promote electron transfer and thus enhance the performance of bioanodes (Zhao et al., 2016).

Figure 4D shows the Electrochemical impedance spectroscopy (EIS) analysis of the bioanodes. The semicircle portion at high frequency corresponded to the interfacial charge transfer resistance (Rct). For the rGO aerogel anodes (10.75 Ω for rGO, 10.67 Ω for rGO/Fe3+, 8.45 Ω for rGO/Fe3O4, 8.26 Ω for rGO/Fe3+/Fe3O4), the Rct values were smaller than the carbon felt bioanode (15.45 Ω), and the rGO/Fe3+/Fe3O4 aerogel anode showed the smallest Rct, which was beneficial for the electron transfer from bacteria to the solid anode (Table S2, Supplemental Information). The curve at low frequency represents the mass transfer resistance (Rmt), and the Rmt of the rGO aerogel anodes (987 Ω for rGO, 896 Ω for rGO/Fe3+, 945 Ω for rGO/Fe3O4, 953 Ω for rGO/Fe3+/Fe3O4) was much lower than that of the carbon felt group (2382 Ω), indicating that the interconnected porous structure of the rGO aerogel electrode was beneficial for mass transfer to some degree. This customized composite rGO/Fe3+/Fe3O4 electrode acts as an excellent carrier of biocatalysts for high performance.

Optimization of the channel size of 3D-printed rGO aerogel electrodes

As mentioned above, microscale pores are prone to be clogged by the growing biofilm, which is the primary reason for the insufficient utilization of the inner zone in typical 3D electrodes. Therefore, to exploit the internal surface of rGO/Fe3+/Fe3O4 aerogel electrode and to enable efficient mass transfer into the 3D structure, we designed five kinds of rGO/Fe3+/Fe3O4 aerogel electrodes with different channel sizes, according to the distance between the two nearby 3D-printed wires (i.e., 1, 1.5, 2, 2.5, and 3 mm). All these 3D rGO/Fe3+/Fe3O4 electrodes possessed hierarchical porous structures ranging from nanometers to millimeters, and the thickness of all the electrodes was 8 mm. Figure 5A shows the current generation of rGO/Fe3+/Fe3O4 aerogel anodes with different channel sizes after nine days of operation. With the increase in the channel size of the rGO aerogel electrodes, the maximum current generation exhibited a tendency of increasing first and then decreasing. The maximum projected current density (84.9 ± 8.3 A/m2) (equals to 10,608 ± 1,036 A/m3 of volumetric current density) was achieved by rGO/Fe3+/Fe3O4-2mm electrode, which is 7.9-fold higher than that (10.8 ± 2.2 A/m2) of carbon felt with the same thickness. To our knowledge, the volumetric current density (10,608 ± 1,036 A/m3) in this study had the highest values for pure culture of G. sulfurreducens to date and was much higher than that of most previously reported bioanodes (Table 1). Then, the bioelectrocatalytic activity of the anodes was further identified using CV. From Figure 5B, it is apparent that the biocatalytic current densities were in the same order as well as stable current densities. Notably, when the pore size was reduced by 0.5 mm from 2 mm to 1.5 mm, the biocatalytic current density of rGO/Fe3+/Fe3O4 decreased by 33%, which was thought to be a limitation of substance transfer in the pore size of 1.5 mm.

Figure 5

Electrochemical characteristics of 3D-printed rGO/Fe3+/Fe3O4 anodes with different channel sizes and carbon felt anode

(A) Current generation under a constant potential of 50 mV vs. Ag/AgCl.

(B) EIS characterized at a working potential of 50 mV, recording from 100 kHz to 10 mHz at AC amplitude of 10 mV.

(C) CV in a PBS solution (50 mM, pH = 7.2) containing 3.38 g/L acetate at 0.1 mV/s.

(D) Long-term current generation of rGO/Fe3+/Fe3O4-2mm anode in a batch mode (400 hr).

Table 1

Comparisons of the performance of 3D porous anodes

Anode	Operating conditions	Inoculation/substrate	Anode performance		Reference
			Projected current density(A/m2)	Volumetric current density (A/m3)
Graphene/polyacrylamide	Single chamber, batch fed, 30°C	Mixed culture,1 g/L acetate	3.6	251.7	(Chen et al., 2019)
FeS2-decorated graphene	H-cell, batch fed, 37°C	Mixed culture, 2 g/L acetate	3.06	/	(Wang et al., 2018)
Onion	Single chamber, discontinuous mode, 30°C	Mixed culture, 1 g/L acetate	9.77	/	(Li et al., 2018)
3D porous CNT sponge	H-cell, batch fed, room temperature	Mixed culture, 1 g/L acetate	6.51	2170	(Xu et al., 2019)
CNT/PANI modified graphite felt	H-cell, batch fed, 30°C	Mixed culture, 1.64 g/L acetate	1.9	95	(Cui et al., 2015)
3D-weaved carbon mesh anode	H-cell, continuous mode (2.17 mL/min)	G. sulfurreducens, 0.64 g/L acetate	/	22.6	(Luo et al., 2020)
Polydopamine-decorated graphitic carbon foam	H-cell, batch fed	S. putrefaciens cells, 1.6 g/L lactate	10.6	3533	(Jiang et al., 2017)
3D printed carbonaceous porous anode	Single chamber, batch fed, room temperature	S. Oneidensis MR-1, 1.6 g/L lactate	1.72	344	(Bian et al., 2018)
3D-printed carbon black-decorated alginate living electrode	H-cell, batch fed	S. Oneidensis MR-1, Tryptic soy broth	0.092	18.5	(Freyman et al., 2020)
3D-printed Fe3O4-decorated graphene	H-cell, batch fed, 30°C	G. sulfurreducens, 2 g/L acetate	84.9	10,608 ± 1,036	current study

CNT, carbon nanotubes; PANI, polyaniline.

EIS tests were performed to investigate the performance of these graphene electrodes with different channel sizes. As shown in Figures 5C and Table S3 (Supplemental Information), the mass transfer resistance of carbon felt was 2,854.2 Ω, which accounted for almost 99% of the total resistance (2,872.1 Ω). In contrast, the mass transfer resistance of rGO aerogel anodes was less than 5% of that of carbon felt, suggesting that the lattice structure of the rGO aerogel fabricated by 3D printing significantly reduced the mass transfer resistance. The mass transfer resistance of the rGO aerogel anodes varied with the channel size. The mass transfer resistance of rGO aerogel anodes decreased by an order of magnitude from 115.6 Ω for rGO/Fe3+/Fe3O4-1mm to 11.0 Ω for rGO/Fe3+/Fe3O4-1.5mm. When the channel size of the rGO aerogel electrode continued to increase, the mass transfer resistance decreased to 6.9, 6.5, and 6.4 Ω, corresponding to rGO/Fe3+/Fe3O4-2mm, rGO/Fe3+/Fe3O4-2.5mm, and rGO/Fe3+/Fe3O4-3mm (Table S3, Supplemental Information), respectively. The results of EIS analysis strongly suggested that the electrode structure was the primary determinant of mass transfer. The value of mass transfer resistance was significantly reduced when the pore size was raised close to the millimeter scale and almost stabilized when the pore sizes continued to increase, indicating that the pore size at the millimeter scale could assure an adequate mass transfer for bioanodes, resulting in the highest current density of rGO/Fe3+/Fe3O4-2mm.

To further elaborate the influence of pore size, the structure and morphology of the biofilms on graphene anodes were subsequently characterized by SEM after 35 days of steady current generation. As shown in Figure 6, the entire surface of each 3D-printed graphene wire of all graphene anodes was covered with a thick and dense biofilm, in contrast to the relatively sparse biofilm on the carbon felt anode (Figure S11, Supplemental Information), and the 3D-printed macropores were not clogged by the biofilm after 1 month of operation. The SEM images of the biofilm in the center of the graphene anodes show quite different morphologies. The internal biofilms of rGO/Fe3+/Fe3O4-1mm to rGO/Fe3+/Fe3O4-2mm appeared to increase density, and the internal biofilms appeared similar anew from the rGO/Fe3+/Fe3O4-2mm to rGO/Fe3+/Fe3O4-3mm, which was in accordance with the value of EIS analysis. The results indicated that the ordered macropores at the millimeter scale provided an unobstructed channel for a large population of bacteria to colonize while ensuring sufficient inter mass transfer of substrates and products. Generally, the channel size of the electrodes primarily affects the performance of bioelectrodes by influencing the biofilm formation and mass transfer in the electrodes, which also mutually influence each other. For a typical porous electrode, pores with a few tens of micrometers often become clogged during the growth and maturation of the biofilm, which can achieve a thickness of tens or hundreds of micrometers. According to a previous study on pores ranging from 100 to 500 μm (Chen et al., 2014; Yuan et al., 2013), the clogging can be avoided following long-term operation, while the mass transfer (primarily acetate and H+), as the considerable rate-limiting step, will hinder the catalytic activity of bacteria in the internal of the electrode and restrict the overall performance, resulting in scattered bacteria or sparse biofilm inside of porous electrodes. Significantly, a thick and dense biofilm formed on the outside and inside of the rGO/Fe3+/Fe3O4-2mm (Figures 6 and 7), indicating that this 3D-printed electrode provided a genuine 3D structure for biofilm development and effective mass transfer, thus guaranteeing the overall catalytic ability of G. sulfurreducens and achieving a remarkable current density.

Figure 6

SEM images of biofilm formation for rGO/Fe3+/Fe3O4 anodes with different channel sizes

Figure 7

Cross section SEM images of one graphene wire located in the center of rGO/Fe3+/Fe3O4-2mm electrode, indicating the presence of large quantities of attached bacteria on the graphene sheets

To evaluate the long-term stability of the 3D-printed rGO bioanodes, a dual-chamber bioelectrochemical reactor using an rGO/Fe3+/Fe3O4-2mm anode was operated for 400 hr with periodic medium exchange. Figure 5D illustrates the electricity generation cycle using rGO/Fe3+/Fe3O4-2mm as the anode. The maximum current densities were relatively stable during 16 batch-feeding cycles, which could be attributed to the consistently sufficient mass transfer owing to the open channels. Moreover, the designed 3D-printed electrode was not destroyed and clogged during the operation. This demonstrates that the rGO/Fe3+/Fe3O4-2mm anode can support a suitable environment for bacterial growth and achieve electricity generation in a long-term operation.

Discussion

In this paper, we proposed and demonstrated a scalable strategy to fabricate a designed 3D structure electrode by functional tailored printing ink and the bioelectrode achieved an unprecedented high volumetric current density of 10,608 ± 1,036 A/m3, the highest volumetric current density with pure G. sulfurreducens to date. This strategy brings new insights into electrode fabrication and has the merit of wide applicability in BESs. We can flexibly adjust the physicochemical and structural properties of electrode by simply altering the base component and functional modular components, such as polymeric matrices, redox species, or conductive nanomaterials. Besides, taking the advantage of 3D printing technology, we can freely fabricate electrodes with arbitrary structure for various demands in BESs, and it is expected to be a feasibility for scale-up application due to the high productivity of 3D printing technology.

Future development of electrode is necessary for scale-up and practical application of bioelectrochemical technique. During the past decades, there are considerable researches regarding the cellular microbiology (bioelectroecology of biofilm, electrons exchange and etc.), electrode (materials, structure, and etc.), reactor configuration, and operation parameters. Our strategy opens up opportunities to integrate the tremendous efforts so far, and achieve functional integration of BESs electrodes, which allows us to modularize major or critical function components of the electrode and match the electrode with the most appropriate properties for specific electrochemical application scenarios. Still, technologies based on electroactive microorganisms are in nascent stage, there is much left to be further explored.

Limitations of the study

The base component and functional modular components of the customized printing ink can be broadened to various materials (e.g., conductive polymer, nanoscale carbon) to endow the electrode with different characteristics. The interactions between different functional materials and electroactive microorganisms deserve to be further studied. This article shows the channel sizes close to millimeter scale can assure an adequate mass transfer and avoid clogging of G. sulfurreducens biofilm. For the optimization of channel sizes in other situations, the influence of electroactive microorganism species and system operation parameters needs to be further considered. In addition to experimental measurements, modeling and simulation work can be further integrated to provide theoretical guidance for the design of electrode structure.

Conclusion

In summary, we have demonstrated a strategy to fabricate a designed 3D structure electrode using a functionally tailored printer ink. The 3D printed rGO/Fe3+/Fe3O4-2mm aerogel bioelectrode achieved an unprecedented high volumetric current density of 10,608 ± 1,036 A/m3, which was 7.9-fold higher than that of the control carbon felt bioanode and was the highest volumetric current density with pure G. sulfurreducens. The mass transfer resistance of the optimized electrode was almost three orders of magnitude lower than that of the carbon felt. The SEM images and electrochemical analysis suggested that the high performance of rGO bioelectrodes was attributed to the hierarchical pores of the rGO aerogel, which ensured sufficient mass transfer of the substrate and high specific surface area for the attachment and growth of bacteria, and the addition of magnetite nanoparticles, which probably enhanced the interfacial electron transfer. This strategy provides new approaches for electrode fabrication and has a great potential for wide applicability in BESs.

Resources availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Qian Fu (fuqian@cqu.edu.cn)

Materials availability

This study did not generate new unique reagents.

Data and code availability

This study did not generate/analyze data sets/code.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.