Cost-efficient microbial electrosynthesis of hydrogen peroxide on a facile-prepared floating electrode by entrapping oxygen

Microbial electrosynthesis of hydrogen peroxide is receiving growing interest for a green substitute for anthra-quinone process. However, poor oxygen transmission of electrode remains an obstacle to enhance H 2 O 2 production rate without aeration. Here, a superhydrophobic natural air diffusion floating electrode (NADFE), which naturally and efficiently entraps O 2 in the air, was proposed for the first time to improve microbial electro-synthesis of H 2 O 2 . Furthermore, a one-step calcined electrode preparation method was developed to reduce energy consumption further. In the microbial electrolysis cell with the NADFE, a high H 2 O 2 production rate of 39 mg/L/h and current efficiency of 86% were achieved without aeration. The production rate of H 2 O 2 was 2.2 times that of a gas diffusion electrode. Importantly, the energy consumption was 34.3 times lower than an electrochemical system. Therefore, the high H 2 O 2 production rate and current efficiency, and low energy consumption of the process provide a superior alternative for environmental remediation.


Introduction
As one of the most important fundamental chemicals, H 2 O 2 is widely used in chemical synthesis and environmental remediation (Wang et al., 2018;Yang et al., 2021;Zhu et al., 2012).The industrial production of H 2 O 2 is heavily dependent on the anthraquinone process, which is energy-intensive and involves expensive palladium catalysts for the hydrogenation of 2-alkylanthraquinone (Nishimi et al., 2011;Oor et al., 2000).
Recently, microbial electrolysis cells (MECs), a typical microbial electrochemical system, have attracted a significant amount of attention, because they can produce valuable products (e.g.H 2 and H 2 O 2 ) with a small voltage input by using electrical energy harvested from organic matter in wastewater (Huang et al., 2019;Logan et al., 2006;Rozendal et al., 2008).Compared to the industrial way to produce H 2 O 2 , O 2 reduction via the two-electron pathway in MECs is more environmentally friendly and cost-effective.
Traditionally, among others, electrode materials were critical factors affecting H 2 O 2 generation (Lee et al., 2021;Zhou et al., 2021).Archetypal electrodes, namely graphite plates, were widely used in MEC systems, as they consumed low amounts of energy (0.3 KWh/kg H 2 O 2 ) (Zou et al., 2021); however, energy consumption increased 331.3 times (99.6 KWh/kg) when taking aeration into account.Gas diffusion electrodes (GDEs), which were able to adsorb partial O 2 in the atmosphere, were developed and used in MEC cathodes to reduce part of energy costs and increase the efficiency of H 2 O 2 production (Li et al., 2016;Sim et al., 2018;Young et al., 2016).However, the H 2 O 2 generation rate was inadequate without aeration, due to its limited O 2 transmission efficiency of the gas diffusion layer.For example, the generation rate of H 2 O 2 was only 4 mg/L/h without aeration in a MEC using a GDE as the electrode, which was much lower than with aeration (8.8 mg/L/h) (Modin and Fukushi, 2012;Young et al., 2016).Therefore, it is crucial to improve cathodic O 2 transmission, to accelerate the H 2 O 2 generation rate without aeration.A superhydrophobic natural air diffusion electrode (NADE) was recently designed and applied in electrochemical systems for H 2 O 2 production (Li et al., 2021;Q. Zhang et al., 2020a;Zhang et al., 2021).Unlike traditional GDEs, which require an additional gas compression device, NADE allows rapid O 2 transportation from the atmosphere to the cathode by modifying the gas diffusion layer.A high H 2 O 2 production rate of 102 mg/h/cm 2 was obtained in the electrochemical system without active aeration (Q.Zhang et al., 2020a).Thus, NADE could be an ideal cathode to increase the H 2 O 2 generation rate in MECs, which has never been explored.Besides, the energy consumption of the MEC, which uses wastewater as a driving force, may reduce compared to that of an abiotic electrochemical system.
However, before integrating NADE into MECs for H 2 O 2 synthesis, several challenges need to be addressed.First, NADE was prepared through a traditional two-step high-temperature calcination method, thus it may have more energy consumption compared to the one-step option.(Li et al., 2021;Q. Zhang et al., 2020a;Zhang et al., 2021), thus increasing energy consumption compared to the one-step option.Second, electrodes placed on the side of the reactor may be subjected to excessive pressure and electrode solution leakage when the reactor is scaled up (Zhao et al., 2021;Zou et al., 2021).Third, NADE has been investigated in electrochemical systems, mostly under the condition of using 50 mM Na 2 SO 4 as the catholyte and Fenton application at pH 3, which requires the high consumption of Na 2 SO 4 and reagents for pH adjustment.
Thus, in this study, to cost-effectively produce H 2 O 2 without aeration, the performance of H 2 O 2 production in the MEC using a natural air diffusion floating electrode (NADFE) was investigated for the first time.
Furthermore, a one-step calcination process, as an alternative to the traditional two-step calcined electrode preparation method, was developed to reduce energy consumption involved in electrode preparation.Besides, the feasibility of floating the electrode on the surface of the catholyte was examined.Thereafter, the performance of H 2 O 2 generation at a neutral pH and low conductivity was investigated.Finally, H 2 O 2 produced by a MEC with one-step calcined NADFE was applied to a Fenton-like (Fe(II)/citric acid/H 2 O 2 ) system for the removal of 11 typical refractory micropollutants in the municipal effluent at a natural pH.

Preparation of natural air diffusion floating electrode
Preparation of the NADFE was based on the traditional two-step calcined method of the NADE preparation (Li et al., 2021;Q. Zhang et al., 2020a;Zhang et al., 2021).In the two-step calcination method, briefly, the modified carbon felt was firstly obtained after the first calcination.Then, the modified carbon black was bonded to the modified carbon felt at the second calcination.
To reduce the calcination and annealing time, a facile one-step calcined preparation method was proposed and applied in this study.The modified carbon felt and carbon black were calcined simultaneously during the preparation of NADFE.Specifically, to obtain a superhydrophobic gas diffusion layer with a high O 2 transmission rate, after ultrasonically cleaned with deionized water and ethanol, carbon felt (length: 4.5 cm; width: 4 cm; thickness: 4 mm; weight: 1 g per piece, Beihai Carbon, China) was soaked in polytetrafluoroethylene (PTFE) suspension (2 wt%) for 2 min (Q.Zhang et al., 2020a).Meanwhile, to obtain a superhydrophobic catalytic layer, carbon black and PTFE suspension with a PTFE/carbon black mass ratio of 0.1 g/g were distributed (40 wt%) into absolute ethanol solution to form an initial mixture.The mixture was stirred and dried at 80 • C by a hot plate stirrer (Fisherbrand, USA), to produce a dough-like paste.Thereafter, 4 g of the paste was rolled onto one side of the pre-treated carbon felt.Finally, the pretreated carbon felt loaded with the catalytic layer was calcined at 350 • C for 25 min.Electrode obtained after calcination was represented as the one-step calcined NADFE in this work.To obtain the different ratios of modified carbon black and carbon felt, the different weights of paste were respectively rolled onto a fixed weight of pre-treated carbon felt, the other conditions were the same as the method mentioned above.

Establishment of microbial electrolysis cells
The schematic of the MEC with NADFE is shown in Fig. 1.The two chambers of the cell were separated by a cation exchange membrane (CMI 7001, Membrane international, NJ).A carbon fiber brush (diameter 5.9 cm, length 6.9 cm, Mill-Rose, USA) was used as the anode electrode.The one-step calcined NADFE was used as the cathode electrode, which floated on top of the catholyte to avoid an electrolyte leakage problem and to reduce the distance between the anode electrodes.O 2 in the atmosphere naturally diffused to the O 2 reduction interface to produce H 2 O 2 .The effective volumes of the anode compartment and the cathode compartment were 180 and 135 mL, respectively.The reactor was first inoculated with municipal wastewater (from the primary clarifier at the Lundtofte wastewater treatment plant, Lyngby City, Denmark), amended with sodium acetate (1 g/L), and was operated in microbial fuel cell (MFC) mode, in order to acquire a mature anodic biofilm (Zhang et al., 2015).The electrodes were connected with a titanium wire 1 mm in diameter through an external resistance of 1000 Ω, and a data logger (USB DAQ 280G, China) was used to record voltage across the external resistor.After one month of operation, the reactor was able to output a stable and repeatable voltage.The reactor was then switched to a continuous MEC flow mode with 0.3 V of applied voltage.Moreover, the series resistance between the two electrodes was reduced from 1000 Ω to 10 Ω.The anolyte (municipal wastewater) was replaced by a synthetic anolyte, with 1.0 g/L of sodium acetate as the carbon source (composition details can be found in a previous study (Wang et al., 2020)).Furthermore, the catholyte was replaced by tap water or different concentrations of Na 2 SO 4 solution.The anolyte and catholyte were both continuously fed through a multi-channel peristaltic pump (Longer, BT100-1F, China).

Experiments
To examine the effect of the mass ratio of carbon black and carbon felt on H 2 O 2 production, one-step calcined NADFEs with different carbon black/carbon felt mass ratios (1, 2, 3 and 4 g/g) were studied in a continuous flow mode for 60 h.The cathode hydraulic retention time (HRT), anode HRT, pH, catholyte and aeration rate were set to 9 h, 12 h, 7, 50 mM Na 2 SO 4 and 5 mL/min, respectively.To investigate whether the oxygen transmission by natural diffusion can meet the demand for H 2 O 2 production, H 2 O 2 production performances at the MEC cathode, with and without active aeration, were compared.One-step calcined NADFE with carbon black/carbon felt mass ratios (3 and 4 g/g) was used as the cathode electrode.The concentration of H 2 O 2 was measured every 12 h.
The following experiments were all carried out without active aeration.First, a comparative study of the performance of the four electrodes (one-step calcined NADFE, two-step calcined NADFE, commercial GDE (length: 4.5 cm; width: 4 cm, Carbon cloth@vulcan carbon powder, CeTech, Taiwan China) and graphite plate (length: 4.5 cm; width: 4 cm, GM-10, Graphite store, USA)) to produce H 2 O 2 was carried out.Second, four cathode HRTs (3 h, 9 h, 18 h, 36 h) were employed to study the effect of HRT on H 2 O 2 production.Meanwhile, the water inlet rate of the anode was always the same as the cathode.Thereafter, four cathode pH conditions (5, 6, 7 and without pH control) were used to study the effect of pH.A pH controller (Prosystem Aqua, Spain) was used to control the pH automatically by adding 0.2 M H 2 SO 4 if the catholyte pH was higher than the set value.Finally, 50 mM Na 2 SO 4 , 25 mM Na 2 SO 4 , 10 mM Na 2 SO 4 and tap water were used as catholytes to investigate the effect of conductivity.Commercial GDE was used as a control when studying the effect of conductivity.
To verify the effectiveness of H 2 O 2 generated from the MEC cathode, H 2 O 2 generated from the MEC cathode using tap water as a catholyte was further applied in the Fe(II)/citric acid/H 2 O 2 system to remove micropollutants from the municipal effluent (23 mg/L of dissolved COD, Lundtofte wastewater treatment plant, Lyngby City, Denmark).First, a spiking solution with 11 micropollutants (Sigma, Denmark) was dissolved in the effluent, and the initial concentrations of the micropollutants ranged from 7.6 to 68 µg/L.Then, the H 2 O 2 was taken from the MEC with a fixed dosage of 10 mg/L, Fe(II) with a fixed dosage of 15 mg/L and different dosages of citric acid (0, 5 and 10 mg/L), which were added to the effluent separately.The batch experiments lasted 60 min.

Analytical methods and characterization
H 2 O 2 concentrations were detected by the spectrophotometric method (400 nm) as described previously (Sellers, 1980).Potassium titanium oxide oxalate forms a yellow pertitanic acid complex in the presence of H 2 O 2 .The COD was determined by the COD cuvette test (Hach, LCK114).The weight of the consumed 0.2 M H 2 SO 4 at the cathode for pH control was measured in terms of its weight change.Fe (II) was measured by a ferrozine method, as described previously (Stookey, 1970).Hydroxylamine hydrochloride (Sigma, Denmark) was used to reduce Fe(III) to Fe(II) when measuring Fe(III) (Viollier et al., 2000).Micropollutant concentration was analyzed using a highperformance liquid chromatography (Agilent 1290 Infinity, USA) system with a tandem mass spectrometer (Agilent 6470 series, USA), the details for which can be found in a previous paper (Wang et al., 2020).
Electrode morphologies were investigated with a scanning electron microscope (SEM, AFEG 250 Analytical ESEM), and the pore size distribution of the air diffusion layer was measured by a micromeritics surface area analyser (Micromeritics Gemini 2375, USA) at − 200 • C. The water contact angles of different materials were measured with a commercial drop shape analysis system (DSA100, KRüSS GmbH, Germany), using a sessile drop at three different points for each sample.Data treatment can be found in the supporting information.

Material characterization
It was observed that long, smooth fibers dispersed randomly with homogeneous large void spaces between them on the surface of the modified carbon felt (see Supplementary Material in online version).Wide pore distribution was also proven via surface area analysis (see Supplementary Material), thereby ensuring effective O 2 diffusion from the atmosphere into the reaction interface naturally (Q.Zhang et al., 2020b).Apart from the modified carbon felt, void spaces were found in the modified carbon black (see Supplementary Material), which helped transfer O 2 from the gas diffusion layer to the widely distributed catalytic sites.A large amount of white filamentous substance (indicated by the red arrow in Supplementary Material) was observed on the modified carbon black, which was proven to be PTFE according to the previous literature (Chen et al., 2015;Q. Zhang et al., 2020b).Only a small amount of filament structure was found in the pores of the modified carbon felt (see Supplementary Material), which facilitated the transmission of O 2 (Zhao et al., 2019).Moreover, the modified carbon black and carbon felt were closely intertwined (see Supplementary Material), which avoided excessive resistance due to the gaps between them.
Apart from morphological characteristics, electrode hydrophobicity was investigated by measuring the water contact angle (CA).A higher CA indicated better hydrophobicity, which would favour O 2 transfer to the reaction interface for fast O 2 reduction (Pérez et al., 2017).In the gas diffusion layer materials (see Supplementary Material), the highest hydrophobicity was observed in the one-step calcined carbon felt (CA = 141.6 • ).The hydrophobicity of two-step calcined carbon felt (CA = 134.4• ) was better than that of raw carbon felt (CA = 121.9• ).A previous report also showed that the hydrophobicity of the carbon felt was enhanced after modification with PTFE (Q.Zhang et al., 2020b).To address the reason why better hydrophobicity was obtained by one-step calcination, the concentration of PTFE on the electrode surface can affect the electrode hydrophobicity, and PTFE slowly decomposed at 350 • C (Schulze and Christenn, 2005;Zhang et al., 2019).The loss of PTFE in one-step calcined carbon felt may be smaller than that of twostep calcined carbon felt, due to the shorter calcination time.Furthermore, the superior hydrophobicity of one-step calcined carbon felt was also observed as an intuitive experimental phenomenon (see Supplementary Material).Water droplets were quickly adsorbed by raw carbon felt.However, the droplets can stand on one-step calcined carbon felt.Therefore, the obstruction of O 2 transmission by the immersion of catholyte was avoided.For the CA of the catalytic layer materials (see Supplementary Material), the hydrophobicity of modified carbon black (CA = 140.5 • ) was also the best, which can significantly affect cathode performance (Wang et al., 2021).The graphite plate had the lowest hydrophobicity, which may be the reason why it needed to rely on dissolved O 2 in water to generate H 2 O 2 (Li et al., 2017).Besides hydrophobicity, Zhang et al. (Q. Zhang et al., 2020b) proved that the O 2 transmission rate of the modified carbon felt was significantly higher than that of the traditional GDE, which makes it an important reason why it can eliminate active aeration.

Effect of carbon black proportion and oxygen transfer performance
The mass ratio of PTFE and carbon black was investigated in a previous study of the NADE (Q.Zhang et al., 2020b).The mass ratio of carbon black and carbon felt may also affect the performance of H 2 O 2 production as a result of the number of active sites and O 2 diffusion distance (Chung et al., 2020).To verify this notion, one-step calcined NADFE with different carbon black/carbon felt mass ratios was used as the cathode electrode (Fig. 2).When the mass ratios of carbon black/ carbon felt were 1, 2, 3 and 4 g/g, the stable concentrations of generated H 2 O 2 were 196, 205, 369 and 339 mg/L, respectively (Fig. 2a), thus demonstrating that the concentration of H 2 O 2 increased in line with the mass ratio of carbon black/carbon felt when the mass ratio was less than or equal to 3 g/g.Interestingly, the best H 2 O 2 production performance was obtained when the mass ratio was 3 g/g, while the highest current was obtained at a mass ratio of 4 g/g.Notably, a higher amount of carbon black increased the thickness of the catalytic layer, which increased the transmission distance of oxygen (Wang et al., 2021).Furthermore, apparent gaps between carbon black and carbon felt were formed during the calcination process when the mass ratio of carbon black/carbon felt was 4 g/g (see Supplementary Material), which may affect the O 2 transmission and resistance of the electrode.As a result, the current efficiency for a mass ratio of 3 g/g was 23% higher than that of a 4 g/g mass ratio (Fig. 2b).Besides, the highest H 2 O 2 generation rate (41 mg/L/h) was also obtained when the mass ratio was 3 g/g.Therefore, the 3 g/g mass ratio was used in the following experiments.
To verify the efficient oxygen transfer performance of one-step calcined NADFE, H 2 O 2 production performances at MEC cathodes, with and without aeration, were compared.When the mass ratio of carbon black and carbon felt was 3, the concentrations of H 2 O 2 , with and without aeration, were 369 and 355 mg/L, respectively (Fig. 2c).H 2 O 2 concentration only dropped by less than 5% without aeration.A similar phenomenon was found when the mass ratio of carbon black and carbon felt was 4 g/g.Additionally, no apparent difference in current efficiencies was found in either the presence or absence of aeration (Fig. 2d).Specifically, when the mass ratio of carbon black and carbon felt was 3 g/g, current efficiencies were both above 86%, with and without aeration.The results show that additional aeration at the catholyte had a slight effect on H 2 O 2 yield, which indicates the one-step calcined NADFE's good O 2 diffusion performance using natural air.In a previous study of H 2 O 2 production using the NADE in an electrochemical system, minimal H 2 O 2 accumulation was found when a diphase system only trapped dissolved O 2 in solution (Q.Zhang et al., 2H Furthermore, the anode COD degradation rate and usage efficiency were measured (detailed in Supplementary Material).No apparent differences in the anode COD degradation rate and usage efficiency were observed between the aeration and without-aeration conditions.Moreover, the theoretical H + concentration required to generate H 2 O 2 was calculated (see Supplementary Material).Theoretically, 2 M of H + is required to generate 1 M of H 2 O 2 according to Equation 1. Besides, actual H + consumption per mole H 2 O 2 production was also measured.When one-step calcined NADFE was used as the cathode electrode, actual H + consumption was only 0.4 M H + / M H 2 O 2 , which reduced 80% compared to that of the theoretical consumption.This means that only 0.7 mL of 98% H 2 SO 4 was consumed to generate 1 L of 355 mg/L H 2 O 2 .The possible reason was that 2.4 M H + / M H 2 O 2 was produced when COD in the anode was consumed at a rate of 24.4 mg/L/h, according to Equations 2 and 3, which in turn reduced actual acid consumption as H + transferred from the anode to the cathode through the cation exchange membrane.This advantage can reduce the need for chemical reagents and the costs involved in practical application.

Comparison with different cathode materials
To examine further the H 2 O 2 production performance of the one-step calcined NADFE (mass ratio of carbon black/carbon felt: 3 g/g) in a MEC without active aeration, its performance was compared with those of two-step calcined NADFE and other conventional electrode materials (i.e. commercial GDE and graphite plates).As illustrated in Fig. 3a, H 2 O 2 concentration and current followed the order of one-step calcined NADFE > two-step calcined NADFE > commercial GDE > graphite plate.Furthermore, the H 2 O 2 generation rate, using one-step calcined NADFE, was 39.4 mg/L/h at HRT 9 h without aeration, which was 1.8, 2.2, and 9.2 times that of the two-step calcined NADFE, commercial GDE and graphite plate, respectively (Fig. 3b).Moreover, the current efficiency, COD degradation rate and usage efficiency of one-step calcined NADFE were all the highest compared with the other three cathodes (Fig. 3b-c).
Besides, actual anode H + production and cathode H + consumption were also calculated when per mole H 2 O 2 was generated.Compared with the other three cathodes, even though the anode H + production was the lowest using the one-step calcined NADFE, actual H + consumption at of one-step calcined NADFE was still the least (Fig. 3d).The lowest actual acid consumption of one-step calcined NADFE may be due to its highest current efficiency, which may reduce unnecessary H + consumption in side reactions such as four-electron reduction.Additionally, energy consumption for H 2 O 2 production also followed the order of one-step calcined NADFE < two-step calcined NADFE < commercial GDE < graphite plate.In conclusion, the results indicate that one-step calcined NADFE has great potential to replace normal electrodes for microbial electrosynthesis of H 2 O 2 , which can not only produce H 2 O 2 efficiently but also use the current wisely.
To address the reasons for the best performance, hydrophobicity was the best option for the one-step calcined NADFE compared with the other three cathodes, as it reduced the impact of catholyte immersion into the gas diffusion layer.Furthermore, the carbon felt modified by PTFE demonstrated better O 2 transmission efficiency than traditional GDEs in the electrochemical system (Q.Zhang et al., 2020a).In addition to the diffusion layer, the catalytic layer of the one-step calcined NADFE (modified carbon black) was proven to be a more efficient catalyst for H 2 O 2 production than graphite and carbon power in previous MECs studies (Barazesh et al., 2015;Q. Zhang et al., 2020a).A comparison of H 2 O 2 production by the different electrodes in the literature will be discussed further in Section 3.5.

Effect of HRT, pH and conductivity on H 2 O 2 production
Apart from the cathodes, HRT was another crucial limiting factor on H 2 O 2 generation.Its impact on system performance was investigated and is shown in Fig. 4. The concentration of H 2 O 2 decreased in line with a decrease in HRT (Fig. 4a).When the cathode HRT was 36, 18, 9 and 3 h, stable H 2 O 2 production was 424, 352, 355, and 138 mg/L, respectively.When the HRT was shortened from 36 to 3 h, the current increased from 5.2 to 10.2 mA.The highest H 2 O 2 generation rate (46 mg/L/h) and current efficiency were both obtained when the HRT was 3 h (Fig. 4b).Specifically, when HRT was 36, 18, 9, and 3 h, current efficiency was 48%, 59%, 86%, and 96%, respectively.The same current efficiency trend was also obtained in a previous study of the MEC for H 2 O 2 production (Sim et al., 2015).Furthermore, according to the performance of NADE in the electrochemical system (Q.Zhang et al., 2020a), O 2 transported from air was sufficient for H 2 O 2 production at a higher current density.The H 2 O 2 generation rate may still have the potential to increase when current density is further enhanced in the MEC.
Moreover, the anode COD degradation rate and usage efficiency were measured (see Supplementary Material), demonstrating an increasing trend with the decrease of HRT.When HRT was 36, 18, 9, and 3 h, COD degradation rates were 12, 17, 25, and 23 mg/L/h, respectively.COD usage efficiency was as high as 2 g H 2 O 2 / g COD at HRT 3 h.Furthermore, actual H + consumption decreased in line with a decrease in HRT.The lowest actual H + consumption obtained at HRT 3 h was only 0.36 M H + / M H 2 O 2 , which is much lower than the theoretical H + consumption for H 2 O 2 (see Supplementary Material).However, when the HRT was 3 h, the COD removal efficiency of the anode was only 9%, and the accumulation concentration of H 2 O 2 was also significantly lower than when the HRT was 9 h.Therefore, a HRT of 9 h was adopted in the following experiments.
According to Equation 1, H + is a reactant for H 2 O 2 production, and thus the pH of the catholyte may affect the rate of H 2 O 2 production.Four pH conditions, including three near-neutral pHs (5, 6, 7) and one without pH control, were used to investigate their effect on H 2 O 2 production.As shown in Fig. 4c, the highest generated concentrations of H 2 O 2 (392 mg/L) and current (11.3 mA) were both obtained at pH 5. The performance of H 2 O 2 generation at pH 6 and pH 7 was close, in that the generated concentrations of H 2 O 2 were 360 and 355 mg/L, respectively.However, the stable concentration of H 2 O 2 without pH control showed a noticeable decline compared with that of pH 7 (from 355 mg/L to 310 mg/L), while the current was similar in both conditions.The pH of the catholyte increased to 11.2 without pH control, because the proton consumption rate at the cathode when synthesizing H 2 O 2 was faster than its transmission rate through the cation exchange membrane (Zou et al., 2021).This high pH meant that the generated H 2 O 2 was easily decomposed by equations 4 and 5 (Chung et al., 2020;Young et al., 2016).Similarly, when the pH was 5, 6, 7 and without control, the H 2 O 2 generation rates were 44, 40, 39 and 34 mg/L/h, respectively (Fig. 4d).Furthermore, current efficiencies were all above 80% at nearneutral pHs, whilst COD removal rates were all greater than 20 mg/L/h (detailed in Supplementary Material).High COD usage efficiency (greater than 1.6 g H 2 O 2 / g COD) was obtained for all pH conditions.In previous MEC studies for H 2 O 2 production using GDEs as cathodes, higher H 2 O 2 generation rates were also found in acid conditions (Chung et al., 2020;Young et al., 2017).However, even though no buffer solution was used in the catholyte, actual H + consumption at pH 5 was still higher than that of pH 7 (detailed in Supplementary Material).So pH 7 was used in subsequent studies.Overall, the results showed that the onestep calcined NADFE could meet the demand for the efficient generation of H 2 O 2 at multiple near-neutral pHs.
In addition to catholyte pH, catholyte conductivity was another key factor for H 2 O 2 production.A 50 mM Na 2 SO 4 solution is widely used as a catholyte to increase conductivity, but it increases the costs involved in using chemical reagents.To reduce the consumption of chemical reagents, the effect of Na 2 SO 4 concentration on H 2 O 2 production and the feasibility of using tap water as a catholyte were investigated (Fig. 4e-f).H 2 O 2 and current concentrations both gradually decreased in line with a decrease in conductivity.This phenomenon was consistent with previous MEC studies for H 2 O 2 production (Zhou et al., 2019;Zou et al., 2021).It is worth mentioning that when tap water was used as the influent, the average concentration of H 2 O 2 for the one-step calcined NADFE was 227 mg/L, which was much higher than the required concentration of Fenton reaction (Hu et al., 2017;Wang et al., 2020).Interestingly, as shown in Fig. 4f, when the cathode influent was adjusted from 50 mM Na 2 SO 4 to tap water, the H 2 O 2 generation rate using one-step calcined NADFE dropped by 34% (from 39 to 25 mg/L/ h).However, when commercial GDE was used, the H 2 O 2 generation rate dropped by 60% (from 18 to 7 mg/L/h).This result indicates that the advantage of using one-step calcined NADFE as the cathode electrode over commercial GDE was more obvious when using tap water as the cathode influent.
Furthermore, the COD removal rate and usage efficiency of one-step calcined NADFE were both higher than that of commercial GDE (see Supplementary Material).However, actual acid consumption was lower than that of commercial GDE.Moreover, conductivities in catholyte influent and effluent were measured (see Supplementary Material).Effluent conductivity significantly increased compared to that of the influent.Among them, the effluent conductivity increased by 186%, using tap water as the cathode influent.In contrast, the control experiment was done by adding the same concentration of commercial H 2 O 2 into the influent, to avoid its effect on conductivity.When 227 mg/L of H 2 O 2 was added to the tap water, conductivity only increased by less than 5%.The result indicates that other chemicals (such as Na 2 SO 4 and K 2 SO 4 ) might form in the cathode, thereby increasing cathode conductivity.Cations (such as H + , Na + and K + ) can pass through the cation exchange membrane from the anode to the cathode (Zou et al., 2021).As the pH was fixed, anion concentration increased during the pH control, due to H 2 SO 4 being added into the catholyte.In addition to obtaining H + from wastewater, another advantage of using wastewater as a driving force was the transportation of cations from the wastewater to the cathode, which increased catholyte conductivity and further increased the H 2 O 2 generation rate.

Comparison with existing cathodes and electrochemical systems
To evaluate comprehensively H 2 O 2 generation, using one-step calcined NADFE at the MEC, its performance was compared with different MEC cathodes (see Supplementary Material).First, the H 2 O 2 generation rate of one-step calcined NADFE in the MEC was 39.4 mg/L/ h, i.e. much higher than those of GDEs (From 0.019 to 8.8 mg/L/h) in MECs (Li et al., 2016;Sim et al., 2018;Young et al., 2016).Besides, when using GDEs in MECs, mechanical aeration was mostly required, to increase dissolved O 2 concentration in the catholyte, which in turn made their energy consumption (from 1.1 to 56 KWh/kg) much higher compared to this study (0.55 KWh/kg).Modin et al. investigated H 2 O 2 generation without aeration, using the GDE as the cathode in a MEC (Modin and Fukushi, 2012).The generation rate for H 2 O 2 was 4 mg/L/h, which was only 11% of this study, but its energy consumption (1.8 KWh/ kg) was 3.4 times that identified herein.Furthermore, Zou et al. reported a scaled-up MEC reactor, where graphite plates were used as the cathode electrode (Zou et al., 2021).Energy consumption from an external power supply was relatively low in the scaled-up MEC reactor (0.3 KWh/ kg); however, when aeration was included in the calculation, it increased by 300 times (99.6 KWh/kg).Low O 2 utilization rate was pointed out as the main reason for high energy consumption.Overall, compared to the listed cathodes studied in the MEC, the one-step calcined NADFE had the lowest energy consumption while still achieving the highest generation rate and current efficiency.
Apart from the MECs, the two-step calcined NADE has been used in an abiotic electrochemistry system (see Supplementary Material).The NADE eliminated the need to pump O 2 /air to overcome gas diffusion layer resistance, resulting in fast H 2 O 2 production (102 mg/h/cm 2 ) (Q. Zhang et al., 2020a).Compared with the electrochemistry system using a NADE as a cathode, the MEC with one-step calcined NADFE in this study had several advantages' First, energy consumption (0.55 KWh/kg) was far less than that of the NADE in the electrochemistry system (19.4KWh/kg), as the bioanode harvested electrons from wastewater and contributed part of the energy for H 2 O 2 production in the MEC.Second, one-step calcination method was used in electrode preparation, which reduced energy consumption compared to that of the two-step method.Third, electrode floated on the catholyte surface can avoid possible electrolyte leakage encountered by placing a cathode on the side of a reactor.Fourth, the proton transport from the anodic wastewater to catholyte reduced acid consumption for H 2 O 2 production and enhanced catholyte conductivity.Fifth, and finally, the MEC reactor was additionally operated with tap water as the catholyte and under neutral pH conditions, which reduced the consumption of chemical reagents.However, compared to the electrochemical system, the main limitation of the MEC system developed in this study was lower current density (0.6 mA/cm 2 ).It has been proven in electrochemical systems that O 2 in the NADE air diffusion layer can still meet demand at higher current densities (Q.Zhang et al., 2020b).In the future, the current density of the MEC with NADFE may increase by expanding the volume of the anode chamber and the number of electrodes, thereby enhancing the H 2 O 2 generation rate.G. Wang et al.

Fig 1 .
Fig 1.Schematic diagram of the MEC using NADFE as a cathode for H 2 O 2 production.

Fig. 2 .
Fig. 2. Effect of the mass ratio of carbon black and carbon felt during cathode preparation: (a) Changes in H 2 O 2 concentration and current and (b) H 2 O 2 generation rate and current efficiency; Oxygen transfer performance: (c) Changes in H 2 O 2 concentration and current and (d) H 2 O 2 generation rate and current efficiency.Entrapping oxygen in the atmosphere was sufficient for producing hydrogen peroxide.

Fig. 3 .
Fig. 3. Comparison of different cathodes on H 2 O 2 production: (a) Changes in H 2 O 2 concentration and current; (b) H 2 O 2 generation rate and current efficiency; (c) Anode COD degradation rate and usage efficiency; (d) Actual H + consumption during H 2 O 2 generation at cathode, and H + production during COD degradation at anode.Note that active aeration was not involved.

Fig. 4 .
Fig. 4. Effect of HRT, pH and conductivity on H 2 O 2 production: (a, c and e) Changes in H 2 O 2 concentration and current, and (b, d and f) H 2 O 2 generation rate and current efficiency.Note that active aeration was not involved.The flow rate of the anode was always the same as that of the cathode.

Fig. 5 .
Fig. 5.The removal efficiency of 11 micropollutants by generated H 2 O 2 in a Fenton-like system.