Bread-derived 3D macroporous carbon foams as high performance free-standing anode in microbial fuel cells

Author’s Accepted Manuscript Bread-Derived 3D Macroporous Carbon Foams as High Performance Free-Standing Anode in Microbial Fuel Cells Lijuan Zhang, Weihua He, Junchuan Yang, Jiqing Sun, Huidong Li, Bing Han, Shenlong Zhao, Yanan Shi, Yujie Feng, Zhiyong Tang, Shaoqin Liu www.elsevier.com/locate/bios

PII: DOI: Reference:

S0956-5663(18)30695-X https://doi.org/10.1016/j.bios.2018.09.005 BIOS10745

To appear in: Biosensors and Bioelectronic Received date: 10 July 2018 Revised date: 25 August 2018 Accepted date: 1 September 2018 Cite this article as: Lijuan Zhang, Weihua He, Junchuan Yang, Jiqing Sun, Huidong Li, Bing Han, Shenlong Zhao, Yanan Shi, Yujie Feng, Zhiyong Tang and Shaoqin Liu, Bread-Derived 3D Macroporous Carbon Foams as High Performance Free-Standing Anode in Microbial Fuel Cells, Biosensors and Bioelectronic, https://doi.org/10.1016/j.bios.2018.09.005 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Bread-Derived 3D Macroporous Carbon Foams as High Performance FreeStanding Anode in Microbial Fuel Cells Lijuan Zhanga,b, Weihua Hec, Junchuan Yangb, Jiqing Sunb, Huidong Lia, Bing Hanb, Shenlong Zhaoa, Yanan Shib, Yujie Fengc, Zhiyong Tangb* and Shaoqin Liua,c* a

School of Life Science and Technology, Key Laboratory of Microsystems and Microstructures Manufacturing, Ministry of Education, Harbin Institute of Technology, Harbin, 150080, P. R. China b

c

CAS Key Laboratory of Nanosystem and Hierarchical Fabrication, CAS Center for Excellence in Nanoscience, National Center for Nanoscience and Technology Beijing, 100190, P. R. China

State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin, 150080, P. R. China * Corresponding authors. E-mail addresses: [email protected] (S. Q. Liu) and [email protected] (Z. Y. Tang)

Highlights 

The N, P, S Co-doped 3D Carbon Foam (NPS-CFs) was fabricated by direct pyrolysis of the commercial bread.



The MFCs equipped with NPS-CF anodes generated a maximum areal power density of 3134 mW·m-2 and current density of 7.56 A·m-2, which is 2.57- and 2.63-fold that of the plain carbon cloth anodes.



The improved performance benefits from high conductivity, good biocompatibility and the improved bacteria-electrode interaction.



NPS-CFs electrode possess low-cost, high mechanical strength, and robust chemical and physical stability, which open up the possibility for ramping up the MFC commercialization.

1

Abstract Microbial fuel cells (MFCs) are a promising clean energy source to directly convert waste chemicals to available electric power. However, the practical application of MFCs needs the increased power density, enhanced energy conversion efficiency and reduced electrode material cost. In this study, three-dimensional (3D) macroporous N, P and S codoped carbon foams (NPS-CFs) were prepared by direct pyrolysis of the commercial bread and employed as free-standing anodes in MFCs. As-obtained NPS-CFs have a large specific surface area (295.07 m2·g-1), high N, P and S doping level, and excellent electrical conductivity. A maximum areal power density of 3134 mW·m-2 and current density of 7.56 A·m-2 are generated by the MFCs equipped with as-obtained NPS-CF anodes, which is 2.57and 2.63-fold that of the plain carbon cloth anodes (areal power density of 1218 mW·m-2 and current density of 2.87 A·m-2), respectively. Such improvement is explored to mainly originate from two respects: the good biocompatibility of NPS-CFs favors the bacterial adhesion and enrichment of electroactive Geobacter species on the electrode surface, while the high conductivity and improved bacteria-electrode interaction efficiently promote the extracellular electron transfer (EET) between the bacteria and the anode. This study provides a low-cost and sustainable way to fabricate high power MFCs for practical applications.

Keywords: microbial fuel cells; three-dimensional electrode; N, P and S co-doped carbon electrode; extracellular electron transfer (EET)

2

Graphical abstract:

1. Introduction Microbial fuel cells (MFCs) are a promising clean energy source to directly convert waste chemicals to available electric power by using microorganisms as catalysts, which couple functions of waste removal and electricity generation (Logan and Rabaey, 2012). Despite great progress has been made in terms of the efficiency and applicability of MFCs in recent years, many issues are waiting to be tackled for practical applications of MFCs (Zhao et al., 2017). The most important challenges include the relatively low power density and poor energy conversion efficiency of MFCs, which mainly arise from low extracellular electron transfer (EET) efficiency between microorganisms and electrode, and the high cost of electrode materials. Thus, how to improve the electron transfer (ET) rate at the bacteria/electrode interface is a key issue in MFCs. In MFCs, electroactive microorganisms deliver the electrons produced from metabolism of organic matters to the anode via mediator-mediated ET through lowmolecular electron shuttling compounds (Chaudhuri and Lovley, 2003), direct ET 3

through tight contact (Malvankar et al., 2011) or microbial nanowires (De Volder et al., 2013) and electrode. Therefore, in regard of optimal bacterial activity and EET process in MFCs, desirable anodes are porous structure with large surface area, which could provide more sites for microorganism loading, facilitate the fast diffusion of mediators, help delivery of substrates and removal of products, and prevent clogging (Xie et al., 2015). In addition, to promote the efficient ET, the surface topography and interfacial charge resistance of anode should benefit establishing an active electric conduit. For increasing physical contact between electrode and outer-membrane proteins/microbial nanowires, a couple of the effective approaches have been exploited including the modification of the anode with conductive nanomaterials or direct utilization of nanostructured materials (Li et al., 2017). The efforts using nanostructured materials have led to a significant improvement in MFC performance. For example, various carbon-based nanomaterials (Yong et al., 2014; S. Zhao et al., 2015) and metal-based nanomaterials (Wu et al., 2011) have demonstrated as promising materials for anodes, giving rise to remarkable enhancement of EET at the bacteria/electrode interface and thus improving MFC performance. However, these carbon-based nanomaterials or metal-based nanomaterials often involve in complex and rigorous preparation processes or are expensive, so their use for scale-up MFCs would face a tremendous challenge. The exploration of inexpensive, readily available and high-efficient anode materials is highly desirable for the future development of MFCs. Compared to currently used carbon-based nanomaterials, carbon materials derived from natural resources that possess a naturally porous structure are likely obtained at a low cost (Chen et al., 2012; Liu et al., 2015). Most importantly, some of natural resources have high content of nitrogen, sulfur and phosphorus. The direct carbonization of such types of natural resources provides a low-cost and sustainable way to prepare heteroatom-doped carbon nanomaterials. Recent studies have proven that the doping of carbon materials with 4

heteroatom could break the electroneutrality of graphitic materials to create the positively charged sites, which favorable for adsorption of negative charged species (Gong et al., 2009). Moreover, the heteroatom introduction may change the spin density of graphitic materials. The change in the spin density or atomic charge density could strongly influence the biocompatibility, wettability, charge storage capacity and catalytic activity of carbon materials, thus affecting the affinity of the electrode surface for bacterial attachment and even the activity of microorganism (Yang et al., 2012; Yang et al., 2011; Kalathil and D. Pant, 2016; Ding et al., 2015; You et al., 2017). For example, the three-dimensional (3D) macroporous nitrogen-enriched carbon anode fabricated from commercially available melamine foam achieved a considerably increased power density (You et al., 2017). The enhanced performance was attributed to involvement of active nitrogen components (graphitic nitrogen, pyrrolic nitrogen and pyridinic nitrogen), which made the electrode positively charged and thus enhanced bacterial-electrode interaction through the electrostatic attraction. Furthermore, graphitic nitrogen and pyrrolic nitrogen took electrons from c-type cytochromes more easily and hence promoted the EET process. In this work, the daily available bread, which possesses a macroporous structure and high content of N, P and S, was used as raw material to prepare 3D macroporous N, P and S co-doped carbon foam (abbreviated as NPS-CF) for high performance and free-standing MFC anodes. The NPS-CF anodes were prepared by direct carbonization of the bread. The obtained 3D NPS-CFs has continuous macroporous structure with the size of several micrometers and high conductivity, and is enriched with the nitrogen, phosphorus and sulfur active components. All these properties benefit attachment of bacteria, establishing the bacteria-electrode electrical connection and promoting the EET process at the bacteria/electrode interface in bioelectrochemical systems, which would definitely improve the output power of MFCs.

5

2. Materials and methods 2.1 Preparation of NPS-CF anode materials The commercially available breads were purchased in Harbin Carrefour supermarket Co. Ltd. After drying naturally, the bread was first cut into smaller sizes (about 1.5 × 1.5 ×1.5 cm 3) and carbonized in a high-temperature furnace at different temperature for 2 h under high pure Ar atmosphere, with a heating rate of 2 oC·min-1. All the obtained samples were named as NPS-CF-X, where X represents the carbonized temperature. Subsequently, after washed by deionized water and ethanol followed by drying, the obtained NPS-CF was cut into the desired sizes (1 × 1 × 1 cm3) and connected with titanium wires to prepare 3D NPS-CF electrodes. 2.2 Characterization Powder X-ray diffraction (XRD) patterns were recorded on a Panaltical X’Pertpro MPD X-ray power diffract meter by using Cu Kα radiation (λ = 1.54056 Å). X-ray photon spectroscopy (XPS) spectra were performed to identify the elements on the surface of the samples with an ESCALAB 250 Xi XPS system of Thermo Scientific, where the analysis chamber was 1.5×10-9 mbar and the X-ray spot was 500 µm. Raman spectra were collected using RenishawinVia plus microscope. A He-Ne laser (514 nm) was used as the light source for excitation. The Brunauer-Emmett-Teller (BET) surface area was measured using the ASAP2420-4 instrument. Prior to each measurement, the sample was heated to 180°C and kept at this temperature for 12 hours. The morphology of the electrode and biofilm attached to anodes was observed under a Hitachi SU8220 scanning electron microscope (SEM). Transmission electron microscopy (TEM) was performed on a FEI Tecnai G2 F20 electron microscope operated at 200 kV. Thermo gravimetric (TG) and differential scanning calorimetry analysis (DSC) were carried out simultaneously using a Diamond TG/DTA instrument 6

-1

from room temperature to 1100 oC at a heating rate of 2 oC·min with an Ar flow rate of ca. 80 mL·min-1. Cyclic voltammograms(CVs) and differential pulse voltammetry (DPV) were performed with a CHI 660E electrochemical working station (CHI Instrument, Shanghai, China) in a standard three-electrode configuration consisting of an as-prepared electrode as working electrode, a saturated calomel reference electrode(SCE), and a platinum wire counter electrode. Electrochemical impedance spectroscopy (EIS) experiments were performed by using 1260 Impedance/gain-phase Analyzer (Solartron Metrology Inc., USA). 2.3 MFC Setup and Operation All MFC measurements were performed in batch mode using a dual-chamber Hshaped MFC device, which was constructed by connecting two equal rectangular glass bottles (with a volume of 100 mL) with N117 proton exchange membrane as a separator. The anode was as-prepared NPS-CF (1 × 1 × 1 cm3) or a plain carbon cloth (1 × 2 cm2), and the cathode was the commercial 3D carbon brush electrodes (with a diameter of 5 cm and height of 5 cm). The distance between the two electrodes was about 12 cm. For MFC start-up, the anode chamber was fed with 100 mL medium containing anhydrous sodium acetate (2.0 g·L-1), NaH2PO4·2H2O (2.77 g·L-1), Na2HPO4·12H2O (11.55 g·L-1), NH4Cl (0.31 g·L-1), KCl (0.13 g·L-1), trace minerals (82.1 mg·L-1), and vitamins (0.2 mg·L-1). And 5.0 mL of pre-acclimated bacteria from activated anaerobic sludge was added into the above anodic chamber. The cathode chamber contained 100 mL potassium ferricyanide (50 mM K3[Fe(CN)6] and 50 mM KCl). All MFCs were operated with an external loading resistance of 1 kΩ, and the curves of cell voltage versus time were collected using an online data acquisition system (2700, Keithly). The medium was refreshed when the voltage dropped below 50 mV. The polarization and power output curves were obtained by varying the 7

external resistance over a range from 40 Ω to 2 kΩ. The current and power density was normalized by the projected area or volume of the anode for analysis. 3. Results and discussion 3.1. Preparation and characterization of NPS-CFs Preparation of N, P and S co-doped carbon foams (abbreviated as NPS-CF) was realized by drying and then pyrolyzing the commercial bread in a high-temperature furnace at different temperature for 2 h under high pure Ar atmosphere, with a heating rate of 2 oC·min-1. As shown in Fig. 1a, the dried bead possesses an interconnected, open and macroporous structure with the pore size ranging from 0.5 to 300 μm. After carbonization, the colour of the bread changes from brown to black (the inset in Fig. 1a and 1b), and the obtained NPS-CF has a good rectangular shape. The TG-DSC analysis was conducted in Ar atmosphere to monitor the carbonization process of the bread (Fig. S1). Three significant weight losses occur, corresponding to removal of H2O and air, removal of oil materials, H2O and air, and carbonization of the sample (see the detailed discussion in Supporting Information). SEM images in Fig. S2 and Fig. 1b reveal the structure evolution of the as-prepared NPS-CF-x at different temperatures. It was noticed that the macroporous structure of the bread is well preserved during carbonization, but the pore size slightly shrank with increasing temperature. Upon carbonization at 1000 oC, the pore size is still in the range of 0.5 ~ 300 μm, large enough for bacterial colonization and fast mass transfer. The cross-sectional image (Fig. 1c) shows an interconnected macroporous structure. HRTEM image discloses that the structures of the randomly orientated graphite and the amorphous carbon are discerned (Fig. 1d). The interlayer distance of the graphitic carbon for NPS-CF obtained at 1000 oC is 0.34 nm, corresponding to the graphite (002) plane. The XRD pattern of the resulting NPS-CFs exhibits two broad XRD diffraction peaks, attributing to the inter-plane (002) and the inner-plane (100)/(101) diffraction peaks of graphitic carbon at the 2θ angles of 23.7° and 43°(Fig. 1e). 8

Compared with pure graphitic carbon, the inter-plane (002) diffraction peak of NPS-CF shifts from 25o to 23.7o, confirming that introduction of nitrogen atoms in the carbon lattice leads to slight distortion in crystalline regularity along the a or b direction, because C-N bonding is shorter than C−C bonding (Qu et al., 2010). The formation of the graphitic carbon with the treatment temperature rises was also confirmed by the Raman spectra (Fig. S3).

Fig. 1. SEM images of bread (a) before and (b, c) after pyrolysis at 1000 oC. (a, b) Top view image and (c) cross-sectional view. The insets are photographs of bread (a) before and (b) after pyrolysis at 1000 oC. (d) HRTEM image, (e) XRD pattern , and (f) EDX spectrum of NPS-CF-1000. The inset in Fig. 1f summarizes the content of C, O, N, P, S and a small amount of K and Ca elements. (g) STEM−EDS mapping images of C, N, O, P and S of NPS-CF-1000

The elemental composition analysis from EDX pattern in Fig. 1f confirmed presence of elemental C, O, N, P and S in NPS-CF-1000. The N, P and S distribution in NPS-CF-1000 sample is relatively uniform, which is verified by STEM and the corresponding elemental mapping images (Fig. 1g). The elemental composition of the resulting NPS-CFs was further investigated by XPS. The XPS spectra of NPS-CFs (Fig. S4 and Table S1) show presence of C, O, N and P in all the samples, but the S content in NPS-CF-1100 is negligible. The N1s spectra in Fig. 2 manifest that three states of N exist in NPS-CFs, which are pyridinic (398.45 eV), pyrrolic (400.69 eV) and graphitic (401.5 eV) nitrogen (Gong et al., 2009). When the 9

pyrolyzing temperature is below 900 °Ϲ, the pyridinic and pyrrolic nitrogen are dominating; however, the ratio of pyridinic and pyrrolic nitrogen decreases with further elevation of the pyrolzing temperature due to their lower stability, while the proportion of graphitic nitrogen increases (Shi et al., 2013). Graphitic and pyrrolic nitrogen are known to be the active sites, because the graphitic and pyrrolic N atoms in the carbon lattice can take electrons from c-type cytochromes more easily than pyridinic nitrogen (You et al., 2017). The XPS spectra of C, P and S indicate that the NPS-CFs contain sp2 graphitic carbon, C-N and oxygen-containing groups (Fig. S5), P–O and P-C (Fig. S6) and oxidized sulphur groups (Fig. S7). The XPS results coupled with the elemental composition analysis from EDX strongly confirm that N, P and S atoms have been successfully introduced into the graphitic carbon structure. Moreover, the presence of P–O and oxidized sulphur groups would make the obtained carbon foams more hydrophilic and thus benefit the EET process at the bacteria/electrode interface (Kasmi et al., 1998).

Fig. 2. High resolution N1s XPS spectra of NPS-CF-650 (a), NPS-CF-700 (b), NPS-CF-800 (c), NPS-CF900 (d), NPS-CF-1000 (e) and NPS-CF-1100 (f).

10

The surface area and pore size distribution of as-prepared NPS-CFs were determined by N2 adsorption/desorption isotherm. The BET surface area increases with elevation of the pyrolyzing temperature and reaches maximum at 1000oC (295.07 m2·g-1), and then slightly decreases when further increasing the temperature to 1100oC (Fig. S8 & Table S2). Moreover, it is discerned that the amount of mesopores also reached the maximum at 1000oC. Abundant mesopores benefit diffusion of substrate and removal of product. The conductivity of the obtained NPS-CFs was further measured by the four-probe method. It was found that the electrical conductivity increases drastically with elevation of the pyrolyzing temperature and is in the range of 4.191~19.014 S·cm-1 (Table S3), demonstrate the high conductivity of the obtained NPS-CFs. 3.2. MFC performance We further investigated the performance of dual-chamber MFCs equipped with asprepared NPS-CF electrodes as the free-standing anodes, with the commonly used plain carbon cloth (CC) electrode as control reference. The cell performance was evaluated in a classic H-shaped MFC (HMFC) in batch-fed mode. The cathode was the commercial 3D carbon brush electrodes and the cathode chamber contained 100 mL potassium ferricyanide (50 mM K3[Fe(CN)6] and 50 mM KCl). Fig.S9 shows the output voltage profiles of NPS-CFs and CC anodes. It is found that the MFCs equipped with NPS-CF anodes requires about 6 days activation and enrichment period to achieve a stable voltage, while the plain CC-based MFC takes at least 8 days to achieve its maximum voltage. This result indicates better affinity of NPS-CFs electrode for bacterial attachment. After the inoculation, all MFCs exhibits reproducible cycles of voltage output (Fig. 3a), and the cell voltages for NPS-CFs (0.68 - 0.61 V) are 1.17-fold higher than that of CC (0.58 V). Moreover, the voltage output of MFC equipped with NPS-CFs are stable during more than two months operation, while the voltage output of the plain CC-based MFC is stable for the first 2 months and then declines. The 11

polarization and power density curves in Fig. 3b & Fig. S10 illustrate that the performance of NPS-CF anodes are all superior to that of CC-based MFC. Moreover, it is clear that the pyrolyzing temperature strongly influences the areal or volumetric power density, current density, and cell voltage of NPS-CF anodes, following the order of NPS-CF-1000 > NPS-CF1100 > NPS-CF-900 > NPS-CF-800 > NPS-CF-700 > NPS-CF-650. The NPS-CF-1000 delivers the highest areal power density of 3134 mW·m-2 and the areal current density of 7.56 A·m-2, which are 2.57- and 2.63-fold higher than that of the plain CC anode (the areal power density of 1218 mW·m-2 and the areal current density of 2.87 A·m-2), respectively. The energy conversion efficiency of NPS-CF anodes is 10.83% - 7.41%, much higher than that of CC anode (6.91%). Fig. 3c confirmed that the potential of the cathodes in different MFCs are controlled to be almost same, whereas there is distinct variation in the anode potentials. Obviously, the higher power density produced by NPS-CF anodes should arise from their increased anode potential compared with CC anode, implying that there are lower overpotential (including activation polarization and ohmic loss) in NPS-CF anodes (Lefebvre et al., 2011).

12

Fig. 3. MFC performances. (a) Output voltage profiles amd (b) polarization and power density curves of MFCs equipped with NPS-CFs and CC anodes. (c) Electrode potentials (vs SCE) as a function of current density of NPS-CFs and CC anodes. (d) Nyquist plots of NPS-CFs and CC anodes after four replicated cycles of current generation. (e) The amplified Nyquist plots of Fig. 3d.

The MFC resistance was further measured using EIS. Before inoculation, as shown in Fig. S11 and Table S4, the NPS-CF electrodes possess smaller ohmic resistance (Rs), lower charge-transfer resistance (Rct) and higher ion diffusion coefficient (D). Both the Rs and Rct values of NPS-CF electrodes decrease with increasing the pyrolyzing temperature, reaching the minimum (4.36 Ω) at 1000°C, demonstrating an extremely fast electron transfer process at the electrode/electrolyte interface. Fig. 3d and 3e show Nyquist plots of NPS-CFs and CC anodes after four replicated cycles of current generation. The growth of the biofilm on the NPS-CF anodes results in decrease in the Rct values. The Rct value of NPS-CF-1000 anode decreases from 4.36 Ω before inoculation to 2.81 Ω after biofilm formation, much lower than that of the CC anode (36.20 Ω), indicating a faster electron transfer process between the bacteria and the electrode surface of NPS-CF-1000. SEM images shown in Fig. 4a, 4b and Fig. S12 disclose that both the entire surface and interior of NPS-CF electrodes are covered by rod-shaped bacterial cells to form a thick biofilm with the thickness of 2-4 μm. Noteworthily, the NPS-CF anodes still maintain their interconnected and open macroporous structure with pore size ranging from 0.5 to 300 μm even after a thick biofilm grows on it, which allows sufficient substrate diffusion to support internal bacterial biofilm growth. The alive-dead stained assay confirms negligible cell death on the NPS-CF-1000 anode after more than two months operation (Fig. 4e), indicating the good biocompatibility. In sharp contrast, for the commercial CC, the bacterial biofilms are 13

formed only on the surface (Fig. 4c and 4d). Moreover, it is easily discerned that after two months discharge, only a few bacteria attached to the surface of CC are alive (Fig. 4f), which might arise from relatively low porosity of CC and limited substrate diffusion.

Fig. 4. SEM images of biofilm growing on the surface of NPS-CF-1000 (a-b) and CC (c-d). CLSM images of biofilm growing on the surface of NPS-CF-1000 (e) and CC (f).

Furthermore, to elucidate the ability of NPS-CF to promote EET of bacteria, CV and DPV were conducted in a three-electrode cell. Fig. 5a outlines CVs of NPS-CFs and CC anodes after stable operation for 30 days under turnover. In the presence of sodium acetate, typical sigmoidal CVs are observed for all the anodes, demonstrating the catalytic oxidation 14

of acetate by these anodes. No obvious redox peak is distinguished in 50 mM PBS in the absence and presence of 2 g·L-1 sodium acetate without the biofilm (Fig. S13), confirming that the NPS-CF and CC anodes have no electrocatalytic activity toward oxidation of acetate and the catalysts are the biofilm attached on these anodes. The catalytic current density of NPSCF anodes (10.47-17.34 mA·cm-2) were 3.08~5.09 times of that of CC anode (3.52 mA·cm-2), indicating the improved catalytic activity. After acetate is depleted, CVs of CC anode displays a redox couple with cathodic peak at -0.42 V and anodic peak at -0.38 V (black curve in Fig. 5b), consistent with the redox potential of outer membrane c-type cytochromes OmcA (Eggleston et al., 2008). As for NPS-CF-1000 anode, aside from the characteristic redox peak

of OmcA (-0.42 V (cathodic) and -0.38 (anodic)), there is a weak redox couple with cathodic peak at -0.048 V and anodic peak at -0.019 V (red curve in Fig. 5b), close to the redox potential of outer membrane c-type cytochromes MtrC (Meitl et al., 2009).

Fig. 5. (a) CVs of NPS-CF and CC anodes under turnover conditions (b) and after acetate is depleted (b). Electrolyte: 50 mM PBS containing 2 g·L-1 sodium acetate. Scan rate: 5 mV·s-1. (c) DPVs of NPS-CF-1000 and CC anodes. Electrolyte: 50 mM PBS. (d) Structure of microbial community at different anodes. (e) Maximum power density of MFCs, which is closely related to the amount of biomass and the amount of electroactive Geobacter.

15

To exclude the interference from the capacitive and background signals, more sensitive DPV was employed. The CC anode shows one major redox peak with the peak potential of 0.42 V (cathodic) versus -0.38 V (anodic) (black curve in Fig. 5c), which are attributed to outer membrane c-type cytochromes. For the NPS-CF-1000 anode, the biofilm gives rise to two redox peaks at -0.42 V (cathodic) and -0.38 V (anodic), -0.048 V (cathodic) and -0.019 V (anodic), respectively. Evidently, the redox pair appearing at -0.048 V and -0.019 V is assigned to MtrC. This result clearly verifies that graphitic nitrogen and pyrrolic nitrogen are able to take electrons from c-type cytochromes more easily (You et al., 2017). Very interestingly, the anodic and cathodic potentials have almost the same value, implying that the redox reaction of MtrC on NPS-CF-1000 anode is subject to the surface-absorbed reactantcontrolled process. In fact, the relationship of peak current with respect to scan rate discloses that as for NPS-CF-1000, the cathodic peak current is proportional to the scan rate (as shown in Fig. S14). It also provides another strong evidence for the surface reactant-controlled process. However, with respect to CC anode, the cathodic peak current versus square root of scan rate shows a linear relationship, suggesting that the redox reaction of OmcA is subject to a diffusion-controlled process. Altogether, above results clearly manifest that the NPS-CF1000 can facilitate extracellular electron transfer between electroactive bacteria and electrode via direct electron transfer through the redox active cytochrome proteins on the bacterial outer membrane. Comparing the cathodic peak current of different anodes at -0.42 V, it can found that the NPS-CF-1000 anode delivers a 5-fold higher cathodic peak current than that of CC anode, suggesting that the NPS-CF anode has more redox active species in the biofilm (Liu et al., 2013). To confirm this deduction, we measured the amount of biomass and the microbial community structure of the biofilm on different anodes via a bacterial 16S rRNA gene clone library. As shown in Fig. 5d &S15, a diverse community including Geobacter, Advenella, Acidovorax and Sphaeroeldelais present on NPS-CF and CC anodes. Geobacter 16

sulfurreducens is dominant microorganism in all the samples, but its amount depends on the anode composite. A positive correlation is revealed between the maximum power density, the amount of biomass and the amount of electroactive Geobacter species (Fig. 5e). The NPS-CF1000 anode, which has the optimal geometrical structure and the dopant element condition, captures more bacteria and benefits enrichment of electroactive Geobacter species, thus produces highest power density. Finally, the NPS-CF anodes possess outstanding properties, including low-cost, high mechanical strength, robust chemical and physical stability, and good reproducibility (Fig.S16 & 17). It deserves to be noted that the capital cost of bread-derived NPS-CF electrodes is evaluated to be about $ 5-8 per m3, or $5-8 per m2 for 1 cm thick pieces, based on the cost raw of materials and energy consumption, at least three order of magnitude less than most commercial CC electrodes (approximately $1,000/m2) and carbon brush (approximately $270/m3) (Wang et al., 2009). Low-cost, high mechanical strength, and robust chemical and physical stability provided by the resulting NPS-CF materials will open up the possibility for mass production of the MFCs. 4. Conclusion In conclusion, N, P and S co-doped carbon foams are successfully prepared by direct pyrolysis of the commercial bread. As-obtained NPS-CFs have a large specific surface area (295.07 m2·g−1), high N, P and S doping level, and excellent electrical conductivity. Impressively, a maximum areal power density of 3134 mW·m-2 and a current density of 7.56 A·m-2 are generated by the MFCs equipped with as-obtained NPS-CF anodes, surpassing all the reported 3D anodes by far (Table S5). Detailed characterizations reveal that such a remarkable performance is attributed to two respects: (1) N, P and S doping makes the obtained carbon foams more hydrophilic, thus favoring the bacterial adhesion and enrichment of electroactive Geobacter species on the electrode surface; (2) the high conductivity and improved bacteria-electrode interaction efficiently promote the EET between the bacteria and 17

the anode. In addition, low-cost, high mechanical strength, and robust chemical and physical stability provided by the resulting NPS-CF materials will open up the possibility for ramping up the MFC commercialization. However, the mechanism of how the surface physiochemical properties affect the affinity of the electrode surface for bacterial attachment and even the activity of microorganism is unclear. Further study will focus on establishing an in-situ platform to understand the mechanism. Acknowledgements This research was financially supported by the National Key R&D Program of China (2017YFA0207201), National Natural Science Foundation of China (No. 21775032, 51372054) and HIT Environment and Ecology Innovation Special Funds (No. HSCJ201618).

18

References Chaudhuri S. K., Lovley D. R., 2003. Nat. Biotechnol. 21, 1229-1232. Chen S., Liu Q., He G., Zhou Y., Hanif M., Peng X., Wang S., Hou H., 2012. J. Mater. Chem. 22, 18609-18613. Ding C., Lv M., Zhu Y., Jiang L., Liu H., 2015. Angew. Chem. 54, 1446-1451. Eggleston C. M., Vörös J., Shi L., Lower B. H., Droubay T. C., Colberg P. J. S., 2008. Inorg. Chim. Acta. 361, 769-777. Gong K., Du F., Xia Z., Durstock M., Dai L., 2009. Science 323, 760-764. Kalathil S., Pant D., 2016. RSC Adv. 6, 30582-30597. Kasmi A. El, Wallace J. M., Bowden E. F., Binet S. M., Linderman R. J., 1998. J. Am. Chem. Soc. 120, 225-.226 Lefebvre O., Uzabiaga A., Chang I. S., Kim B.-H., Ng H. Y., 2011. Appl. Microbio. Biotechnol. 89, 259-270. Logan B. E., Rabaey K., 2012. Science 337, 686-690. Liu C., Li J., Zhu X., Zhang L., Ye D., Brown R. K., Liao Q., 2013. Int. J. Hydrogen Energ. 38, 15646-15652. Liu Q., Zhou Y., Chen S., Wang Z., Hou H., Zhao F., 2015. J. Power Sources 273, 1189-1193. Li S., Cheng C., Thomas A., 2017. Adv. Mater. 29, 1602547. Malvankar N. S., Vargas M., Nevin K. P., Franks A. E., Leang C., Kim B.-C., Inoue K., Meitl L. A., Eggleston C. M., Colberg P. J. S., Khare N., Reardon C. L., Shi L., 2009. Geochim. Cosmochim. Acta 73, 5292-5307. Mester T., Covalla S. F., Johnson J. P., Rotello V. M., Tuominen M. T., Lovley D. R., 2011. Nat. Nanotechol. 6, 573-579. Qu L., Liu Y., Baek J.-B., Dai L., 2010. ACS Nano 4, 1321-1326. Shi Q., Peng F., Liao S., Wang H., Yu H., Liu Z., Zhang B., Su D., 2013. J. Mater. Chem. A 1, 14853-14857. 19

Volder M. F. L. De, Tawfick S. H., Baughman R. H., Hart A. J., 2013. Science 339, 535-539. Wang X, Cheng S, Feng Y, Merrill M D, Saito T, Logan B E., 2009. Environ. Sci. Technol. 43, 6870-6874. Wu X., Zhao F., Rahunen N., Varcoe J. R., Avignone-Rossa C., Thumser A. E., Slade R. C. T., 2011. Angew. Chem. 123, 447-450. Xie X., Criddle C., Cui Y., 2015. Energy Environ. Sci. 8, 3418- 3441 Yang L., Jiang S., Zhao Y., Zhu L., Chen S., Wang X., Wu Q., Ma J., Ma Y., Hu Z., 2011. Angew. Chem. 50, 7132-7135. Yang Z., Yao Z., Li G., Fang G., Nie H., Liu Z., Zhou X., Chen X., Huang S., 2012. ACS Nano 6, 205-211. Yong Y. C., Yu Y. Y., Zhang X., Song H., 2014. Angew. Chem. Int. Ed. 53, 4480-4483. You S., Ma M., Wang W., Qi D., Chen X., Qu J., Ren N., 2017. Adv. Energy Mater. 7, 1601364. Zhao C.-e., Gai P., Song R., Chen Y., Zhang J., Zhu J.-J., 2017. Chem. Soc. Rev. 46, 15451564. Zhao S., Li Y., Yin H., Liu Z., Luan E., Zhao F., Tang Z., Liu S., 2015. Sci. Adv. 1, e1500372.

20