Recent Advances with Biomass‐Derived Carbon‐Based Catalysts for the High‐Efficiency Electrochemical Reduction of Oxygen to Hydrogen Peroxide

The oxygen reduction reaction (ORR) plays a pivotal role in electrochemical energy conversion and chemical production. Two‐electron (2e−) charge transfer for oxygen reduction is considered a promising method for the on‐site production of hydrogen peroxide (H2O2), which requires electrocatalysts with high H2O2 selectivity and ORR activity. Noble metal alloys (e.g., Pt‐Hg and Pd‐Hg) have been prevalent materials of choice due to their desirable intrinsic activity, but their scarcity and high cost seriously hinder their widespread application in practice. Self‐doped heteroatomic carbon‐based electrocatalysts, derived from abundant and inexpensive biomass, have emerged as attractive candidates for on‐site H2O2 production. This review summarizes the fundamentals and recent advances in H2O2 production via 2e− ORR, including basic catalytic mechanisms, the influence of electrolyte pH and porous structure of catalysts, selectivity assessment methods, determination of the cumulative H2O2 concentration, development of biomass‐derived carbon‐based catalyst, and electrochemical device designs. Current challenges and proposed opportunities are also presented with an emphasis on large‐scale electrochemical H2O2 synthesis.


Introduction
Considering increasing demand and environmental concerns, effective and environmentally friendly oxidants have attracted great attention from the chemical and medical industries. [1,2]Hydrogen peroxide (H 2 O 2 ) is an important basic chemical that is widely used as a disinfectant and oxidant in pulp bleaching, [3,4] chemical synthesis, [5,6] medical disinfection, [7] wastewater treatment, [8,9] and so on.In addition, H 2 O 2 is also considered as a potential energy carrier for generating electricity in a single-chamber H 2 O 2 fuel cell. [10,11]lobal H 2 O 2 consumption has been gradually increasing and its huge demand is 2.2 Mt per year, with a continuous annual increase of 4%. [12]In 2022, the global H 2 O 2 market estimated at US$10.4 billion, is projected to reach a revised size of US $15.4 billion in 2030, growing at a compound annual growth rate (CAGR) of 5% during the analysis period 2022-2030. [13]Currently, almost 98% of the world's H 2 O 2 production comes from conventional anthraquinone oxidation (AO). [12]However, this process requires centralized infrastructures and expensive palladium-based catalysts, leading to high energy consumption as well as organic waste. [14]In addition, during transportation and storage, the H 2 O 2 aqueous solution must be condensed to a concentration of up to 70 wt%, [15] which would lead to further costs and serious accidents.For this reason, there has recently been increasing interest in the exploration and development of simpler, energy-efficient, and environmentally friendly technologies for on-site H 2 O 2 production.
An alternative method, i.e., directly reacting H 2 and O 2 to produce metal-catalyzed H 2 O 2 was explored in the last decade.However, because mixing H 2 and O 2 with high concentrations in a chamber can result in a flammable and explosive regime, the research interest has gradually shifted toward safe protocols.In recent years, photo/electro-catalytic synthesis of H 2 O 2 has attracted a great deal of attention.Compared to the two mentioned routes, these new approaches are more economical and environmentally friendly.Belonging to direct synthesis methods, they can be performed using Earth-abundant H 2 O and O 2 on reagents, but are easier to handle and present lower risks.[18] Particularly, H 2 O 2 produced by oxygen electroreduction via a two-electron oxygen reduction reaction (2e À ORR) process, has attracted great interest due to its on-site, decentralized, and environmentally friendly character, Figure 1. [19,20]Once electricity, water, and air (O 2 ) are available, H 2 O 2 can be produced for immediate use, including in homes sanitation or cleaning products.This approach could avoid the expense and hazard associated with transporting and storing H 2 O 2 .However, the challenge is to develop selective 2e À ORR catalysts that can direct the reaction pathway to the H 2 O 2 product. [21]The first report, in 1887, was a Hg-Au catalyst developed by Traube, which showed high 2e À ORR selectivity for the production of H 2 O 2 . [22]Recently, a variety of materials, including noble metal alloys (e.g., Pt-Hg and Pd-Hg), carbon-based materials, single-atom catalysts, etc., have been discovered that show good H 2 O 2 selectivity. [23,24]oble metal alloys possess low overpotential and high H 2 O 2 selectivity (>97%) for 2e À ORR, but their scarcity and high costs seriously hinder their widespread application in practice.In contrast, due to their large specific surface area, high electrical conductivity, and abundant reserves in earth, carbon-based materials, have been recognized as nontoxic and cost-effective alternative catalysts for 2e À ORR. [25,26]However, the original carbon-based materials show an unsatisfactory tendency to desorb the reaction intermediate (*OOH) during the ORR process, as well as low selectivity and catalytic activity for H 2 O 2 due to their electronic structure. [1]In this context, considerable efforts have been devoted to improve the H 2 O 2 selectivity and ORR activity of carbon-based catalysts by various modification/decoration strategies including doping with N, P, O, and other heteroatoms. [27,28]n recent years, biomass-derived carbon materials have become of great interest due to their accessibility, low cost, and environmental friendliness. [29]It should be noted that biomass contains a certain amount of heteroatomic elements, which allow them to be introduced in situ into the material without the need for additional procedures. [30]As a result, the synthesis of in situ self-doped carbon-based porous material with heteroatoms, with a unique structure and chemical composition can be achieved by a simple biomass carbonization process.Moreover, these biomass-derived carbon materials possess abundant pore structures, which are favorable for offering more inner active sites and shortening the diffusion pathways of reactants/products during the ORR process. [31]More importantly, the conversion of biomass wastes into high-value-added electrocatalysts is beneficial for promoting resources-recycling and carbonfixation processes.
In this review, we provide an overview of recent advances in biomass-derived carbon materials that serve as efficient catalysts for 2e À ORR emphasizing the key role of heteroatom-doped strategy in enhancing H 2 O 2 selectivity and catalytic activity.First, the catalytic mechanism of 2e À ORR, the influence of electrolyte pH and porous structure of catalysts, the evaluation of selectivity, and the determination of the accumulated H 2 O 2 concentration are described.Then, the recent progress of biomass-derived carbon materials, i.e., heteroatom-doped carbon-based catalysts, for H 2 O 2 production is discussed in detail.Several selected examples of on-site H 2 O 2 production are given to investigate H 2 O 2 yield, expense, actual yield, and usefulness in practical small electrochemical devices.In the final section, we briefly point out current challenges and propose opportunities facing this rapidly growing field.

Catalytic Mechanism
The oxygen reduction reaction (ORR) is a reduction that takes place at the cathode, and is a complex process involving multiple electron transfer processes and reaction intermediates. [32]In ORR, O 2 molecules can be reduced directly to H 2 O via a fourelectron (4e À ) or follow a two-electron (2e À ) pathway to form H 2 O 2 .Since H 2 O is the thermodynamically favored product, producing H 2 O 2 is an obstacle through 2e À ORR pathway.
Generally, the 4e À pathway involves three reaction intermediates; *OOH, *O, and *OH, whereas only *OOH intermediate appears in the 2e À pathway. [33]The 2e À ORR pathway to H 2 O 2 is composed of two coupled electron-proton transfer steps, as indicated in Table 1.The first step consists of adsorption of the O 2 molecule on the active site, followed by its combination with H þ and acceptance of e À to generate *OOH species.In the second step, the *OOH intermediate reacts with another H þ to form H 2 O 2 , achieving the 2e À ORR pathway. [34]In contrast, for the 4e À pathway, the O-O bond in the *OOH also can cleave to form *O and *OH, successively receiving H þ and e À to generate H 2 O. [35] Subsequent reduction of H 2 O 2 to H 2 O can occur at the electrode/ solution interphase, Figure 2a, (orange part).It follows from the above that for H 2 O 2 production, the O-O bond must be Figure 1.Electrochemical cell stack powered by green electricity for H 2 O 2 production.Reproduced with permission. [129]Copyright , Wiley-VCH.
Table 1.Two-e À and four-e À reaction pathways in the ORR in acidic media."*" denotes an active site.
preserved to inhibit H 2 O generation during ORR.For an ideal catalyst for 2e À ORR, the binding free energy (ΔG *OOH ) of the *OOH species must be adequate to avoid the cleavage of O-O bond, Figure 2b.More specifically, the binding strength of O 2 molecules on the catalyst surface should be strong enough to facilitate the formation of *OOH, while the adsorption of *OOH should not be too strong or weak to rapidly desorb the generated H 2 O 2 .In a word, activity and selectivity are the key factors for an ideal 2e À ORR catalyst, the former is associated with the kinetic barriers of O 2 activation and *OOH binding strength on the catalyst surface, and the latter is related to the kinetic barrier of *OOH protonation versus dissociation. [36]ccording to the above-mentioned ORR mechanism, the adsorption of O 2 molecules on the catalyst surface is the first step and the one that determines the ORR rate.Therefore, the mode of O 2 adsorption plays a vital role in the H 2 O 2 selectivity of the catalyst.Yeager et al. [37] proposed three types of O 2 adsorption modes on metal sites, as described in Figure 2b, i Therefore, regulating the adsorption modes to optimize the binding strength of reaction intermediates is a rational strategy to design efficient 2e À ORR catalysts with high selectivity for H 2 O 2 .
Density functional theory (DFT) calculations work well to describe the adsorption energies of intermediates on the surface of catalysts, except the energy of solvated protons and electrons on the electrode at a given potential.Nørskov et al. [38,39] first developed a typical calculation method on the catalytic process based on the computational model of the hydrogen electrode (CHE) model.The free energy of a single proton-electron pair is defined as ÀeU relative to H 2 in the gas phase under standard conditions, where U is the electrode potential versus the reversible hydrogen electrode (RHE).The adsorption-free energy of an intermediate (with n proton-electron pairs) generally consists of several parts, including the calculated binding energy (ΔE ele ), adsorbate solvation (ΔE w ), electric field effects (ΔE field ), zeropoint energy (Δ ZPE ), and entropic corrections (ÀTΔS), as shown in Equation: Since the theoretical overpotential is a function of binding free energies of the three intermediates *OOH, *OH, *O, the limiting potential can be calculated from the *OH bond directly based Figure 2. a) Reactions that can occur during the ORR process, 2e À (red part) and 4e À (blue part) pathways in the ORR, and the H 2 O 2 reduction (orange part).Reproduced with permission. [130]Copyright 2021, Wiley-VCH.b) Schematic illustration of the reactions and the equilibrium potentials during the O 2 electroreduction.Reproduced with permission. [131]Copyright 2019, American Chemical Society.c) Schematic illustration of three adsorption modes of molecular oxygen on catalytic active sites.Reproduced with permission. [132]Copyright 2020, Royal Society of Chemistry.
on the scaling relationships (ΔG(*OOH) = ΔG(*OH) þ 3.2 and ΔG(*O) = 2ΔG(*OH)).Figure 3a shows the results of this calculation for the 4e À process in what is known as a "volcano" plot, due to the crossed relations for binding of *OH and *OOH.Nørskov et al. [40] concluded from these plots: 1) for catalysts that bind strongly with *OH, the reaction of *OH !H 2 O is potentially limiting (the solid blue line) for catalysts that show weak binding, O 2 !*OOH is potentially limiting (solid green line).This understanding creates an avenue to use various catalyst material engineering strategies, such as edge structure, strain, boundary modification, and doping, to engineer the binding strength of a catalyst and thereby achieve the minimum theoretical overpotential (%0.37 V) of the 4e À process.For the 2e À ORR process, since *OOH is the only intermediate involved in the reaction, its formation or removal becomes the only rate-limiting step (disregarding O 2 adsorption).Based on the above scaling relationships, a similar volcano plot, Figure 3b, can also be prepared to show the relationship between the limiting potential and the binding free energy of *OH.For catalysts with strong binding capacities, the desorption of *OOH to form H 2 O 2 will be the ratelimiting reaction step.Conversely, if the binding between O 2 and the catalyst is relatively weak, the activation of the O 2 and the reaction of O 2 !*OOH will be the rate-limiting step.Optimal catalyst binding capacity, in principle, can be achieved by appropriate material regulation, potentially giving a zero overpotential such as that of hydrogen evolution reactions (HER), Figure 3b. [41,42]n the volcano diagram, Figure 3c, the binding strength of *OOH determines the ORR pathway from the volcano peak to the left (4e À pathway forms water) and to the right (2e À pathway forms H 2 O 2 ). [43]A competitive reaction between the 4e À and 2e À pathways is observed.The left side of the peak exhibits a strong *OH binding site, and the free energy of *OH to form H 2 O decreases.This indicates that the selectivity of the 4e À ORR pathway is higher than that of the 2e À pathway.Moving from the double-electron volcano peak to the right, the two images overlap, showing that the O-O bond is more difficult to break, the weakening of the interaction with *O and *OH, and an increase in selectivity through chemical dissociation of *OOH or electrochemical reduction.Meanwhile, the activation of *OOH by O 2 decreases. [1]

pH Effect and Pore Structure
In previous studies, it has been observed that the selectivity and activity of 2e À ORR catalysts are strongly affected by the pH of the electrolyte, so it is necessary to understand the effect of pH for the design of efficient catalysts in different electrolytes. [44]As mentioned above, the overall reaction and thermodynamic potential versus RHE of the 2e À ORR are as follows: at pH However, when pH = 13, higher than the acid dissociation constant of H 2 O 2 (pK a = 11.7), the product of 2e À ORR changes from H 2 O 2 to HO 2 À , and the overall reaction can be expressed as: À ) = 0.74 V versus RHE.Although 2e À ORR can take place in the electrolytes with different pH value, such as potassium hydroxide (KOH), phosphate buffer (PBS), sulfuric acid (H 2 SO 4 ) and perchloric acid (HClO 4 ) solutions, [45] the production of H 2 O 2 in acidic medium is particularly attractive.On the one hand, the .Reproduced with permission. [40]Copyright 2018, American Chemical Society.c) The calculated ORR activity volcano plot for 2e À pathway to H 2 O 2 .Red and blue symbols indicate *OOH adsorption at C and Fe, respectively.The equilibrium potential of O 2 /H 2 O 2 is shown as a black dashed line.Reproduced with permission. [43]Copyright 2019, Springer Nature.
higher oxidation capacity of acidic H 2 O 2 allows its efficient use in various industrial processes.On the other hand, H 2 O 2 is more stable in acidic medium than in alkaline medium, since it would decompose more easily under higher pH condition.
For metal catalysts, it is more difficult to dissociate the O-O bond in acidic media than that in alkaline media due to a higher kinetic barrier for O-O bond dissociation, indicating that in an acidic medium the 2e À ORR pathway is more favored. [46]In contrast, for carbon-based catalysts, the ORR activity is lower due to their intrinsically weaker *OOH interaction, but the intermediates (H 2 O 2 and *OOH) tend to be desorbed from carbon-based catalysts to avoid O-O bond dissociation.Noffke et al. [47] revealed that based on a well-defined graphene nanostructure, the preference for the 4e À ORR pathway is observed in alkaline electrolytes by combining theoretical calculations and experimental results.This model also shows the preference of ORR for the 2e À pathway for H 2 O 2 production in the acidic electrolyte, Figure 4a.Conversely, under neutral or alkaline pH, a decrease in selectivity toward H 2 O 2 was observed as H 2 O formation through O-O cleavage became progressively more available.
Kim et al. [48] proposed that the electron density at the Fermi level of oxygen-functionalized sp 2 carbon, under operating ORR conditions, is the key determinant in the reaction mechanism, which is reflected in the measured electrical conductivity.Recently, research has been carried out to unravel the underlying mechanism of 2e À ORR catalyzed by slightly reduced graphene oxide grown on P50 carbon paper (mrGO/P50) by first-principles theory and experiments.Surprisingly, this catalyst showed remarkable selectivity (>99%) toward H 2 O 2 in both alkaline and acidic electrolytes.However, the activity is highly dependent on the pH of the electrolyte.They observed that a critical aspect was the availability of electron density at the Fermi energy of the catalyst, i.e., the electrical conductivity of the NrGO, which largely determined whether a coupled proton-electron transfer (CPET) process would dominate.If the catalyst was semiconducting (NrGO-2), then the only possible initial ORR process was an outer-sphere electron transfer to O 2 in solution to produce O 2 À (aq), with the source of the electron being the conducting support (P50 carbon paper).After this initial electron transfer, O 2 À or HO 2 À (depending on the pH of electrolyte) would adsorb on the semiconductor, followed by another uncoupled protonelectron transfer, and then produce H 2 O 2 (non-CPET 2e À ORR).In contrast, if there were sufficient electron density or electrical conductivity (NrGO-4), then the CPET 4e À ORR mechanism would probably dominate because the barriers for ORR were generally lower for coupled proton-electron steps than for uncoupled steps, Figure 4b.
Daniel et al. [49] reported graphitized nitrogen-doped singlewalled carbon nanohorns (CNHs) as electrocatalysts for selective ORR to H 2 O 2 .As shown in Figure 4c, in the first 6 h, the Faradaic efficiencies (FEs) at pH = 1 and pH = 13 remained almost constant.The drop in FE% was not related to the stability of the material but was most likely the result of further reduction of H 2 O 2 to H 2 O over long experimental times or simply decomposition of H 2 O 2 by chemical processes.Further analysis showed that under acidic condition, the resulting effective protonation of pyridinic N compromised the ability to weaken the O-O bond, leading to the preferential formation of H 2 O 2 .
In addition, porous structures, including pore size and porosity, are crucial factors for the catalytic performance of 2e À ORR, mainly influencing the mass transport process in catalytic layer. [50,51]In the case of mesoporous catalysts, the produced H 2 O 2 can be released in a relatively short contact time due to the ease of mass transport; therefore, H 2 O 2 can be produced with extremely high selectivity.In contrast, H 2 O 2 produced on catalysts with micropores is resident for a longer time and is therefore more likely to be further reduced to form H 2 O or decompose by disproportionation; therefore, ORR results in lower selectivity  [47] Copyright 2016, American Chemical Society.b) Scheme of coupled proton-electron transfers (CPET) versus non-CPET mechanisms on the nitrogendoped reduced graphene oxide (NrGO); Electron-transfer number (e-/O 2 ) for the three NrGO catalysts as a function of pH.Reproduced with permission. [48]Copyright 2019, American Chemical Society.c) Plots of FE% (H 2 O 2 ) versus time using a g-N-CNH modified electrode at pH = 1.0, 7.4, and 13.0; a stability test at pH 1.0, 7.4, and 13.0 over 24 h.Reproduced with permission. [49]Copyright 2018, Elsevier.d) H 2 O 2 selectivity of mesoporous carbon hollow spheres (MCHS-9:1) in different electrolyte and Tafel plots.Reproduced with permission. [133]Copyright 2020, American Chemical Society.toward H 2 O 2 .Park et al. [50] prepared a series of mesoporous nitrogen-doped carbon catalysts, which exhibited well-ordered mesopores with diameters of 3.4-4.0nm, and showed a high selectivity toward H 2 O 2 exceeding 90%.This high selectivity toward H 2 O 2 is probably due to good mass transport of the mesoporous structure and the abundant active sites exposed in the catalytic layer.Xiao et al. [52] described an rGO/PEI aerogel catalyst with higher selectivity in H 2 O 2 generation, related to the steric hindrance effect.The 3D porous structure of aerogels and the steric hindrance effect between PEI (polyethyleneimine) and rGO interface confer higher 2e À catalytic selectivity (90.7%), production rate (106.4mmol g catalyst À1 h À1 ) and durability for H 2 O 2 generation.A catalyst composed of mesoporous carbon hollow spheres (MCHS) showed high H 2 O 2 selectivity in neutral electrolyte (0.1 M PBS) attributed to suitable porosity and functional groups, cf. Figure 4d.DFT calculations manifested that MCHS samples related to the C-O group had the best H 2 O 2 formation performance in alkaline and neutral electrolytes.In an acidic electrolyte, the onset potentials of the metal-free carbon catalysts are relatively low because the neutralization of OOH À would result in energy loss for H 2 O 2 formation.

Evaluation of Selectivity
The 2e À ORR electrocatalytic activity is usually evaluated in terms of reduction peak position and current via cyclic voltammetry (CV).To investigate the ORR kinetics, linear scan voltammetry (LSV) is recorded using the rotating disk electrode (RDE).The LSV curves are then fitted to the Koutecký-Levich (K-L) model: [53] (1) where j, j k , and j d are the measured, kinetic, and limited-diffusion current densities, respectively; n is the number of transferred electrons; C O 2 is the concentration of O 2 in the electrolyte; D O 2 is the diffusion coefficient; F is the Faraday constant (96 486.4 C mol À1 ), v is the kinematic viscosity of the electrolyte, and k is the electron transfer rate constant; ω is the electrode rotating rate (rad s À1 ).The H 2 O 2 selectivity of the catalyst can be obtained by the rotating ring-disk electrode (RRDE) technique.RRDE works with a standard three-electrode cell system in a single chamber or Hcell, cf. Figure 5, which is a powerful electrochemical method to quantitatively measure H 2 O 2 production.Typically, a glassy carbon disk with Pt ring electrode serves as the working electrode.At the disk electrode, O 2 is reduced to H 2 O 2 or H 2 O, and then the generated H 2 O 2 is rapidly transferred to the concentric Pt ring electrode by the forced convection caused by the rotating motion of the electrode.The resulting H 2 O 2 is oxidized back to O 2 at the Pt ring electrode. [54]The measured potentials are quoted with respect to the reversible hydrogen electrode (RHE) and corrected for ohmic losses and background current.To quantify the H 2 O 2 production, the ring is set at %1.2 V versus RHE, a potential at which the ORR current is negligible and the oxidation of H 2 O 2 to O 2 at the ring electrode is limited by mass transport.Once the H 2 O 2 is generated at the disk electrode, there would be a positive current at the Pt ring electrode.
Based on the ratio between the O 2 consumption rate toward ) and the total consumption rate ( ṅO [55] the selectivity for H 2 O 2 production (H 2 O 2 % or O 2 efficiency%) can be calculated by the following Equation ( 3) where The number of transferred electrons during the ORR process, n, can be calculated from In addition, the selectivity of H 2 O 2 is often expressed as the Faradaic efficiency (FE) obtained by titration or spectroscopic Figure 5. a) Schematic of a RRDE set-up in a three-electrode electrochemical cell.Reproduced with permission. [1]Copyright 2021, Wiley-VCH.b) Schematic of the electrochemical H-cell.Reproduced with permission. [60]Copyright 2018, Springer Nature.quantification method.The selectivity of H 2 O 2 can be expressed by FE% [46] FE% ¼ 100 Â I R N À1 I D (5)   In addition, in-depth knowledge of reactive oxygen intermediate species (ROS), such as O 2 •À , is also important for fundamental exploration of the unambiguous mechanism of ORR and concern for the durability of electrocatalysts for practical applications.Here, ROS-triggered electrochemiluminescence (ECL) is proposed as a powerful tool with a high signal-to-noise ratio for screening the kinetics of ORR electrocatalysts, especially the temporal/spatial distribution of ROS intermediates in the vicinity of electrocatalysts. [56]The kinetic evaluation includes synchronous sequential ORR and ECL reactions driven by doublepotential step chronoamperometry and subsequent collection of ECL intensity.A commonly used potentiostat, glassy carbon electrodes, and photomultiplier (PMT) are required to drive the ORR and detect traces of intermediate ROS in the diffusion layer.ORR electrocatalysts with 2e À /4e À pathways could be reliably discriminated by ECL peak intensities using principal component analysis (PCA).Furthermore, due to multiple ultrasensitive stoichiometric reactions between ROS and the ECL luminophore, quantitative kinetic information, such as the apparent rate constant and potential-dependent temporal/spatial distribution of ROS, can be extracted by finite element analysis (FEA) of the ECL decay curves, Figure 6.There are four preconditions for ECL to be applicable in the evaluation of ORR kinetics using the instruments: 1) whether ROS are generated in ORR; 2) whether ROS trigger chemiluminescence of luminol; 3) whether short-lived ROS produced during ORR be stable until chemiluminescence is triggered; 4) whether the lateral influence of exotic luminol on ORR is negligible or not.

Determination of Cumulative H 2 O 2 Concentration
Currently, the concentration of H 2 O 2 in the electrochemical system is usually detected by three methods: UV-vis spectrophotometry, titration, and colorimetric test strips, Figure 7. Gil et al. [57] studied the accuracy of these three methods in the concentration range of 5-1000 ppm.The results showed that the UV-vis method with the cobalt carbonate assay is very robust and produces a relative measurement error of less than 5%.Titration with KMnO 4 provides comparable error metrics to UV-vis when the H 2 O 2 concentration is above 150 ppm.Colorimetric strips tend to be inaccurate under many conditions and should be used as "semiquantitative" measurement means.
The Beer-Lambert law is known to be the quantification principle of UV-vis spectrophotometry, [58] which states that the absorbance of a given sample at a specific wavelength is linearly proportional to the concentration of the absorbing assay.Although the H 2 O 2 by itself does not show adequate light absorption for UV-vis, it can react with another reagent (e.g., Co/CO 3 ) to form a compound that absorbs strongly in the UV/visible  [56] Copyright 2023, Wiley-VCH.
range.In the presence of bicarbonate, H 2 O 2 could oxidize Co 2þ and form a dark greenish "carbonato-cobaltate" complex (denoted as Co(CO 3 ) 3 3À ), whose maximum wavelength is at 257 nm [59] used for the quantification assay.It should be noted that the UV-vis method with the Co/CO 3 is compatible with several electrochemical systems, but the different electrolytes need to assemble their respective calibration curves.
The chemical titration method for the measurement of H 2 O 2 quantification mainly depends on the redox reaction between a standard sample and H 2 O 2 , in which the standard sample is gradually consumed and the amount of H 2 O 2 is determined by its consumption.The titration method is based on a resulting color change to know the stoichiometric point of the reaction.There are three common direct titration method: KMnO 4 titration, iodometric titration, and cerium titration.
In the case of titration with KMnO 4 , the dark purple MnO 4 À anion is highly oxidizing, and reduced by H 2 O 2 to yield a nearly colorless Mn 2þ product.
Once H 2 O 2 has fully reacted, further addition of the titrant leads to the end point of the titration, as indicated by a persistent faint pinkish color (lasting >30 s).The H 2 O 2 concentration, C H2O2 , can be calculated by the following Equation ( 5) where V EP is the volume of titrant added to reach the titration endpoint (mL); C KMnO 4 is the standard concentration of the titrant (mol L À1 ); V Al is the volume of the H 2 O 2 aliquot (L).
For the iodometric method, in the H 2 SO 4 medium, H 2 O 2 first oxidizes an excess of iodide to iodine, which is calibrated by a standard solution of sodium thiosulphate to determine the amount of iodine formed, thus extrapolating the H 2 O 2 concentration.The specific equations for the reactions are as follow At the beginning of the titration, the presence of iodine was easily observed by the characteristic yellow-brown color of the solution.In the final stages of the titration (after the solution had reached a pale, yellow color), starch was added as an indicator to determine the final endpoint. [60]The endpoint was reached when the characteristic blue color of the starch-iodine complex disappeared completely to give rise to a clear solution.The iodometric titration method has fewer side reactions compared to titration with KMnO 4 .However, the sodium thiosulphate solution is not stable enough, which requires an additional calibration process, resulting in a relatively low detection efficiency.
In the cerium titration method, an orange Ce 4þ solution is reduced by H 2 O 2 to colorless Ce 3þ , Equation (8). [61]An appropriate amount of o-diazafluoride is used as an indicator to determine the endpoint and extrapolate the H 2 O 2 content by the color change of the solution.
The concentration of Ce 4þ before and after the reaction can be measured by UV-vis spectroscopy.The wavelength used for the measurement is usually 316%318 nm. [62]Therefore, the H 2 O 2 concentration can be determined by M(H 2 O 2 ) = 2 Â M(Ce 4þ ), where M(Ce 4þ ) is the mole of Ce 4þ consumed.The cerium method produces no intermediates during the titration process and has fewer side reactions and simpler influencing factors, so in many cases, the test results are more accurate than those of the iodometric method.
Colorimetric test strips usually operate on a redox/color change principle like titration, but use solid-phase assays.The test strip is immersed in H 2 O 2 solution and then exposed to air for a period, hereafter referred to as the air exposure time.During this time, the strips undergo a color change, ranging from blue to yellow/brown.The color of the strip is correlated  [57] Copyright 2020, American Chemical Society.
to the H 2 O 2 concentration and can be approximated by eye or read by an electronic strip reader.Quantofix test strip is one of the commercial H 2 O 2 colorimetric test strips.It explained that in the 100 ppm strips, H 2 O 2 will react with peroxidase and an organic substrate to form a colored (blue) oxidation compound. [45] Biomass-Derived Heteroatom-Doped Carbon Catalysts for 2e À ORR Carbon materials possess high surface area and excellent electrical conductivity, which are beneficial for exposing abundant active sites and facilitating rapid electron transfer during the electrocatalysis process.However, the relatively high cost of raw materials and complex synthesis are not feasible for scalable production.To address these problems, biomass-derived porous carbon materials have recently emerged as low-cost, earth-abundant, and renewable materials for ORR catalysts.[63][64][65][66] In addition, biomass such as chitin, eggs, tobacco, Euonymus japonicas, lysine, and plant residues serve as precursors for both carbon and heteroatom, resulting in the self-doping of porous carbons with heteroatoms (N) without the use of any external heteroatom precursors, concentrated acids, or strong oxidants, thus making the synthesis green and environmentally friendly.However, many previous works have shown that electrocatalysts based on N-doped carbon accelerate the 4e À ORR pathway under alkaline conditions.[30,67] Many efforts have sought to establish a fundamental understanding of the nature of the active sites in nitrogen-doped carbon materials.Most propose that, in alkaline medium, both pyridinic-N and graphitic-N species are considered active sites that promote the 4e À ORR pathway, [68][69][70] while pyrrolic-N has a favorable effect on the 2e À ORR pathway.[71][72][73][74] It has been reported that nitrogen atoms with higher electronegativity could activate the π-conjugated system and impart positive charge on adjacent carbon atoms, thus facilitating the adsorption of *OOH intermediates.[69] However, the delocalized lone pair electrons of pyridinic-N could aggressively induce charge transfer from the π orbital to the antibonding orbitals on O 2 , resulting in a significantly weakened O-O bond and further dissociation of *OOH intermediate into *O and *OH.[49,75,76] Whereas in the acidic condition, the presence of protonation of the pyridinic atoms prevents this effect by occupying the lone pair electrons to form the N-H bond, causing the ORR reaction to convert to the 2e À pathway.[77] It has been reported that the graphitic-N is positively charged and that the carbon atoms surrounding graphitic-N can act as Lewis's acids, [78] which are not favorable for adsorption of intermediates.Qiao et al. [23] proposed that *OOH intermediates could be substantially preserved in the presence of a high amount of pyrrolic-N, leading to a two-electron ORR pathway on adjacent carbon atoms.The 4e À pathway was assumed to occur preferentially on carbon atoms adjacent to pyridinic-N rather than pyrrolic-N dopants.However, the influence of different N-doped active sites on 2e À ORR selectivity remains controversial, as it is difficult to fit a specific N-functionality to conventional measurements and to eliminate the effect of carbon material properties and reaction conditions.
Apart from N-doping, carbon centers functionalized with oxygen species have also been suggested to be the active sites in heteroatom-doped carbon catalysts that promote 2e À ORR, although the specific oxygen-containing functional groups vary among studies.The introduction of oxygen-containing functional groups (C=O, C-O, COOH, etc.) endows the carbon materials with striking conductivity, and electroactivity, as well as hydrophilicity that facilitates the transport of oxygen, in liquid phase, to the surface of the electrocatalyst. [79,80][83] Zhou et al. [84] modified graphite felt electrodes by simple electrochemical oxidation.The abundant surface oxygen groups (-COOH, -COH, -COO-, R-OH, >C=O) provided hydrophilicity to the modified carbon surface to produce readily accessible dissolved O 2 .Cui's group [85] employed DFT calculations to study the activities of a wide range of oxygen functional groups including carboxyl (-COOM, M=H and Na), carbonyl (C=O), etheric (O-C-O), and hydroxyl (-OH) introduced at different locations of the carbon matrix; e.g., the basal plane or edge.The ΔG *OOH was used as a descriptor and the activity volcano was plotted to highlight the activities of the different oxygen functional groups. [38,86]The calculated limit potential (U L ) was as a function of ΔG *OOH for the 2e À ORR to H 2 O 2 in these materials.The maximum limit potential was 0.70 V, representing zero overpotential at the top of the volcano.The calculated values suggested that the -OH functional group did not contribute significantly to the ORR.However, the C-O-C groups at the basal plane and at the edge of graphene (basal O and edge O) were very active for the two-electron reduction of oxygen to H 2 O 2 with overpotentials of 0.02 and 0.06 V, respectively, comparable to previously reported noble metal catalysts. [86]Among the different possible configurations for the -COOM functional group, they found that the arm-chair edge was the most active (COOM edge 2), resulting in an overpotential of 0.06 V.However, in another example, the results of DFT calculations showed that the C=O group at the defect edge exhibited the ΔG *OOH of 4.24 eV approaching the apex of the volcano (the ΔG *OOH of 4.22 eV corresponds to the thermodynamic equilibrium potential, U 0 H2O2 = 0.7 V).More importantly, the 2e À ORR activity of the C=O group at the defect edge was shown to exceed that of the CÀOH and COOH groups, which was in agreement with the chemical titration results. [87]ang et al. [88] used waste pitaya peels with high oxygen content, as carbon precursor, to prepare a hierarchically oxygen selfdoping porous electrocatalyst by a facile pyrolysis and activation method with KOH, as illustrated in Figure 8a.The pitaya peelderived carbon-based catalyst achieved H 2 O 2 yield of 41.6 mg h À1 cm À2 at a high current density of 100 mA cm À2 with a current efficiency of 65.5%.After a long-term electrolysis of 10 cycles (1 h each cycle) of successive tests at 70 mA cm À2 , the H 2 O 2 yields and the corresponding current efficiency are very stable, with a maximum decrease of 6.6%.This excellent performance could be ascribed to its abundant oxygen-containing self-doped groups, structural defects, and sp 3 -C bonds that serve as catalytic sites.O-induced active sites with high O content (>15%) can greatly enhance the 2e À ORR capability of O-doped carbon-based catalysts.Motivated by this, Guo et al. [89] prepared a carbon quantum dot (CQD) catalyst with ultra-high O content (30.4 at%) using glucose (C 6 H 12 O 6 ) as the carbon source due to its high atomic oxygen-to-carbon ratio, Figure 8b.The O-rich CQD catalyst showed excellent catalytic ability for H 2 O 2 production with selectivity close to 100%.In an H-type cell configuration of bulk H 2 O 2 electrosynthesis, the CQD catalyst showed a high yield of 10.06 mg cm À2 h À1 and Faraday efficiency of 97.7%, as well as good stability over 10 h, which showed a great potential in practical H 2 O 2 production.Experimental and theoretical studies confirmed that most of the C-O bonds were derived from ether groups in the CQD catalysts, and that the carbon atoms of the C-O bonds are the most active sites for 2e À ORR.
Distillers' grains (DG) are solid residues generated during the brewing process, and rich in crude protein, crude fiber, and crude fat of which the element O is the elemental composition.DGs are an ideal precursor to obtain O-doped porous biochar (OPB) due to their good accessibility and cost-effectiveness.Hu et al. [90] used distillers' grains as biomass precursor to prepare O-doped porous biochar (OPB) catalyst by pyrolysis with KHCO 3 .The elemental mapping result showed that the OPB possessed a high O content of 19.41 wt%.The O 1s spectrum was fitted to three peaks at 536.1, 534.3, and 533.5 eV, corresponding to -OH, -COOH, and C-O groups, respectively, verifying the effective introduction of oxygen functional groups.The H 2 O 2 yield of the OPB catalyst was 15.46 mmol L À1 at 0.7 V at pH = 2.In another case, natural bamboo biomass was also used as a carbon precursor to prepare the oxygen-doped rock electrocatalyst for electrochemical generation of H 2 O 2 . [91]The three-dimensional interconnected hierarchical pores in the rock electrocatalysts exposed abundant electroactive sites and facilitated the mass transfer of reactive species.As shown in Figure 8c, this rock electrocatalyst exhibited H 2 O 2 selectivity up to 80%.At an optimum catalyst loading of 2 mg cm À2 , H 2 O 2 productivity of 1037 mmol g À1 catalyst h À1 with a high current efficiency of up to 74.1% at 30 mA cm À2 was achieved using an airbreathing cathode.Higher H 2 O 2 productivity (1525 mmol g À1 catalyst h À1 ) was obtained by further increasing the current density at 50 mA cm À2 .
Miao Gao et al. [92] fabricated O-doped carbon-based catalyst (MBCs) by a simple mechanical peanut shell grinding treatment.This ball milling treatment not only effectively increases the stacking density of the catalyst, but also introduces abundant defects and oxygen-containing functional groups.As a result, the treated MBCs biochar has higher electrical conductivity, and its H 2 O 2 production activity was 4.2 times higher than that of pristine biochar.Under neutral pH conditions, treated biochar showed 87% selectivity for H 2 O 2 production.Compared with commercial acetylene black and state-of-the-art electrocatalysts, the ball-milled biochar showed superior performance in H 2 O 2 .Reproduced with permission. [88]Copyright 2022, Elsevier.b) D-glucose formula structure and preparation process for the O-rich CQDs; High-resolution XPS spectra of O 1s of the CQDs; FE of the CQDs.Reproduced with permission. [89]opyright 2022, Royal Society of Chemistry.c) Schematic of the preparation process for bamboo-derived O-doped rocky electrocatalyst; RRDE curves; Stability test of the O-BC-3h catalyst-coated air-breathing cathode in 0.05 M Na 2 SO 4 .Reproduced with permission. [91]Copyright 2022, Elsevier.
production and maintained a stable yield during the 6 h reaction period.
Nitrogen atom (N) is the most widely used dopant atom, and its lone pair of electrons has a conjugation effect with π-electrons in carbon materials, which can induce efficient catalytic sites with good electrochemical properties, helping to facilitate the adsorption and transfer of proton from O 2 molecules by the catalyst.At present, ORR mechanisms catalyzed by N-doped materials are proposed in the following three aspects: 1) Dipole-dipole interaction.Carbon atoms adjacent to nitrogen were found to exhibit higher positive charge density, higher electron affinity, and better catalytic activity; [93] 2) Nitrogen hydrogenation.The results of DFT calculations indicated that at a potential relevant to the reaction, half of the nitrogen atoms in the material would hydrogenate.This hydrogenation process would destabilize some carbon atoms in the lattice and provide a segregated charge. [94]ue to the low destabilization of carbon sites, the resulting molecular oxygen with a chemisorbed state, possessing the characteristics of a superoxide species, would be only slightly stable, promoting the formation of H 2 O 2 ; 3) Optimization of surface properties.It was reported that N doping could change the alkalinity and hydrophilicity of carbon nanomaterials, and increase the charge mobility on the surface of carbon materials. [95]he group of Randriamahazaka [96] successfully prepared N-doped nanoscale carbon dots (CDs) using the microwave-assisted method, as shown in Figure 9a.Glutamine and glucose in ionic liquid (1-ethyl-3-methylimidazole ethyl sulfate) were used as nitrogen and carbon sources, respectively, and the obtained sample was denoted as CDs-1.CDs were generated from glucose in ionic liquid, and were denoted as CDs-2.The as-prepared CDs displayed nanometer-sized particles and N-doping with the presence of a thin layer of the ionic liquid.Both CD materials showed significant electrocatalytic activity and a prevalent 2-electron pathway leading to H 2 O 2 production.More interestingly, CDs-2 synthesized in the presence of glucose and ionic liquid showed selective 2-electron reduction of O 2 , with H 2 O 2 production exceeding 90% from 0.6 V to À0.2 V versus RHE.
Different types of N-containing functional groups in N-doped carbon catalysts have different influence on 2e À ORR selectivity. [61]For example, Panomsuwan et al. [97] prepared a series of nitrogen-doped carbon catalysts (CL-NCs) derived from Cattail leaf for ORR, Figure 9b.The ammonia concentration (1.0, 1.5, and 2.0 M) was varied to alter the nitrogen doping content.Characterization results revealed that the CL-NCs exhibited an amorphous structure, while the density of structural defects increased as the ammonia concentration was increased.The nitrogen-doping content in the CL-NCs ranged from 0.65 to 1.55 at%, with the predominant proportions of pyridinic-N and graphitic-N.Within the potential range of À0.3 to À1.0 V versus RHE, the n value for CL-NC-0 was about 2.72 to 2.85, and H 2 O 2 Figure 9. a) Synthetic pathway of carbon dots in ionic liquids media; RRDE curves; Chronoamperometric response of CDs-2 relative current variation at 0.2 V as a function of time.Reproduced with permission. [96]Copyright 2018, Elsevier.b) Schematic illustration of the preparation of CL-NCs by hydrothermal ammonia treatment and pyrolysis processes; peroxide yields of all catalysts calculated from RRDE; long-term stability tests.Reproduced with permission. [97]Copyright 2022, Elsevier.c) Schematic illustration of the preparation of biochar from water hyacinth plants and the production of H 2 O 2 from Z0 to Z4 and GP samples.Reproduced with permission. [98]Copyright 2018, Elsevier.
production was about 60%, indicating that ORR proceeded mainly through a 2e À pathway; CL-NC-1.0 showed the highest value of n (3.23 to 3.55) and H 2 O 2 production was the lowest (30% to 40%).In contrast, Sun et al. [98] found that the dominant role in H 2 O 2 production was the total nitrogen species content rather than the ratio between different nitrogen species.A series of nitrogen-doped biochar ORR catalysts were obtained from water hyacinth by carbonization of molten ZnCl 2 salts, and the Z0-Z4 catalysts represent different ratios of water hyacinth powder and zinc chloride salt, Figure 9c.As the carbonization progressed from Z0 to Z4, the pyridinic-N content increased from 41.21 to 56.32 at%, while graphitic-N decreased from 58.79 to 43.68 at%.The ratio of pyridinic-N to graphitic-N was better in Z2 than in Z3, but the H 2 O 2 yield of the former was not as high as that of the latter.Compared with other samples, the current efficiency of Z3 for H 2 O 2 production after 5 min of reaction at À0.7 V reached 81.2 AE 2.5%, which has high energy utilization efficiency.The H 2 O 2 production rate reached an outstanding yield of 21.7 Â 10 À3 (mmol L À1 min À1 cm À2 ).
Xue et al. [65] fabricated nitrogen-doped activated carbon derived from ramie biomass as an efficient electrocatalyst for H 2 O 2 production.The high-temperature carbonization heat treatment significantly improved the specific surface area, mesoporous structure, and degree of graphitization of the activated carbon material.When the initial concentration of H 2 SO 4 electrolyte was 0.05, 0.01, and 0.005 M, the amount of H 2 O 2 after 60 min was 18.1, 14.1, and 11.9 μM, respectively.Since H 2 O 2 generation increased with decreasing pH, the results indicated that the higher H þ concentration and lower initial pH (pH = 1) are beneficial for higher amount of H 2 O 2 .In the 0.05 M H 2 SO 4 þ 0.05 M Na 2 SO 4 solution, a slight difference (approximately 4%) was observed after 14 consecutive cyclic utilization compared to the first cycle, indicating the excellent stability and reusability of the NAC electrocatalyst.
Doping with phosphorus (P), boron (B), and sulfur (S), has been identified as an effective strategy to improve the ORR activity and H 2 O 2 selectivity of carbon-based materials due to their electron-withdrawing effect on adjacent carbon atoms. [99]As a result, a positive electron density is created in the carbon matrix, which can act as an active site to attract O 2 molecules, which are then chemisorbed and reduced. [95,100,101]Compared to the doping of O and N elements, the study of carbon materials monodoped with P, B, and S heteroatoms derived from biomass resources for H 2 O 2 production is less.For example, litchi shell was investigated as the precursor of an electrocatalyst based on Pdoped carbon toward ORR to generate H 2 O 2 . [66]The obtained carbon with a BET-specific surface area of 693 m 2 g À1 exhibited hierarchical porous structures with micropores and macropores derived from the special structure of the litchi shell.The hierarchical porous structure was beneficial for high catalytic activity.Onset and half-wave potentials of %0.98 and %0.80 V RHE were obtained in the electrolyte of 0.1 M KOH solution, respectively.A high H 2 O 2 -efficiency (%60%) was achieved with porous carbon.[104][105] The interaction, between doped heteroatoms contributes to improve the efficiency H 2 O 2 generation. [106]As a remarkable example of multidoping strategy, Wang et al. [71] prepared an S, N self-doped biomass carbon catalyst, termed SN-BC derived from waste ginkgo leaves without additional support templates and activation processes, Figure 10a.X-ray photoelectron spectroscopy (XPS) analysis revealed that N and S heteroatoms were successfully codoped into the porous carbon material.Interestingly, N or S doping can make adjacent carbon atoms positively charged, which facilitates oxygen adsorption. [107,108]The SN-BC catalyst showed a H 2 O 2 production of 98.9 mg L À1 in 120 min.The SNJC Janus cathode was composed of a hydrophobic gas diffusion layer at the center and hydrophilic SN-BC catalytic membrane at both ends.The SNJC showed a high CE of 62.4%, and excellent H 2 O 2 generation performance of 98.9 mg L À1 at 120 min with a lower energy consumption of 12.5 Wh g À1 .It is more likely that this phenomenon was associated with the three-phase interface in SNJC as highly active sites for ORR and sufficient O 2 supply from hydrophobic gas diffusion layer.In another case, black liquor (BL), as a waste liquor produced by pulping with sulfate or caustic soda in paper mills, composed of 60% organic matter and 40% inorganic chemicals, was used as feedstock to prepare a porous carbon cofunctionalized with O and B as a highly active and selective 2e À ORR catalyst. [109]lectrochemical results showed that C 3 -O 3 -B 10 exhibited excellent 2e À ORR performance than C 3 -O y and C 3 -B z under alkaline conditions, including high activity for H 2 O 2 production with early onset potential (0.804 V vs RHE), remarkable H 2 O 2 selectivity of 77%, and long-term electrochemical stability in the 5 h chronoamperometry test.The intrinsic causes of the superior 2e À ORR activity and selectivity of C 3 -O 3 -B 10 can be attributed to the following aspects: 1) C 3 -O 3 -B 10 possessed a lower activation energy (E a ) of 6.7 kJ mol À1 @ 0.4 V compared to C 3 -O y and C 3 -B z ; 2) the doped B atom activated the conjugated π electrons system, causing partial localization; 3) the B atom acted as a conduit for electron transfer.O 2 was more prone to form the Pauling adsorption mode at C 3 -O 3 -B 10 , which was attributed to the fact that the lowest unoccupied molecular orbital (LUMO) of an O 2 triplet would have maximum overlap with the highest occupied molecular orbital (HOMO) of the boron-doped carbon.A similar phenomenon also occurred in cotton-stalk-derived activated carbon fibers (CSCFs) with monodoped N and dual-doped N/P catalysts. [110]The maximum H 2 O 2 generation of the CSCF-N/P and CSCF-N electrodes were 41% and 25% greater than that of pure CSCF.The N/P dual-doped CSCF-N/P exhibited the highest H 2 O 2 generation capacity due to the enhanced ion transport and electron conductivity resulting from the combined action of the N and P functional groups.
Copyright 2023, Elsevier.b) The stability of the NO/PC-500-GDE (a) and the restored cathode (b) at 100 mA cm À2 .(NO/PC-500-GDE without a hydrophobic layer electrolyzed at 30 mA cm À2 after 240 min electrolysis is defined as the deactivated cathode.The restored cathode was obtained by introducing a hydrophobic layer on the deactivated cathode).The adsorption energies of OOH and H 2 O 2 molecules on pure carbon and various N/O containing functional group doped carbon substrates; the mechanism of the two-electron ORR for H 2 2 electrogeneration on NO/PC catalysts.Reproduced with permission. [134]Copyright 2021, Royal Society of Chemistry.
a one-electron ORR to form *OOH; finally, the *OOH on the NO/PC surface was bonded to H þ to generate H 2 O 2 through proton-coupled electron transfer.This work provides insights into the potential application of biomass-derived heteroatomdoped porous carbon as an effective electrocatalyst for H 2 O 2 generation.Chen et al. [112] regulated the doping amounts of N, P heteroatoms as well as the content of active species (pyrrolic-N and P-C bonding motifs) to improve the 2e À ORR selectivity of the carbonaceous catalyst derived from Millettia speciosa champ.Thanks to the high N, P-doped amounts (both were about 2.9 at.%, respectively), the large specific surface area (1388 m 2 g À1 ) as well as appropriate pore sizes (1.5/3.8 nm).The obtained carbonaceous catalyst displayed a high accumulative H 2 O 2 concentration of 168.7 mg L À1 in 140 min.The current density performs about 10% fall off after 30 000 s [142]  summarizes the 2e À ORR performance, including H 2 O 2 selectivity, production rate and stability, of biomass-derived and nonbiomass-derived heteroatom-doped carbon-based catalysts demonstrating that biomass-derived heteroatom-doped carbonbased catalysts show comparable electrocatalytic performance for H 2 O 2 production compared to nonbiomass-derived catalysts.

Electrochemical Device Design for H 2 O 2 Production
Along with promising catalysts for the preparation of 2e À ORR, the design and development of well-configured electrochemical devices to carry out the bulk production of H 2 O 2 is another crucial part to achieve industrial-scale application.In the design of electrochemical devices, many critical factors, such as mass transport, gas diffusion, and H 2 O 2 decomposition, during long-term electrolysis, must be considered.As described above, rotating ring disk electrode (RRDE) is often used in the laboratory to examine H 2 O 2 production rapidly and quantitatively in catalysts.However, it is difficult to reflect the influence of mass transport limitation and long-term stability by RRDE measurement because it is performed under forced convection condition in a short period. [113]The H-type cell configuration, in which the porous electrode is directly immersed in liquid electrolyte, can simulate the mass production of H 2 O 2 . [60,85,114]Yamanaka et al. designed an H-cell with a Nafion membrane.With 2 mol L À1 NaOH, as electrolyte, 7 wt% H 2 O 2 was produced at the mixed-carbon cathode with a current efficiency of 93% (CE). [115]With 1.2 mol L À1 H 2 SO 4 , as electrolyte, 3.5 wt% H 2 O 2 was produced at a Mn porphyrin-derivatized cathode with a current efficiency of 45%. [116]However, the performance deteriorated rapidly, resulting in low reaction rate and product concentration.Although H-cells are widely used to select electrocatalysts for electrochemical H 2 O 2 production, they fail to achieve a high conversion rate or industrial-grade current density (e.g., >100 mA cm 2 ) or overcome the mass transport limitation problem.
A schematic illustration of each component of the liquid flow electrochemical device for H 2 O 2 production is shown in Figure 11a.The main components are the gas flow channels, catalyst layers, gas diffusion layers, exchange membrane, electrolyte compartment, etc.To improve the H 2 O 2 performance, two important factors should be considered in the design of liquid flow electrochemical devices: 1) H 2 O 2 produced in situ should be discharged and collected in time to avoid decomposition; 2) a stable three-phase gas-liquid-solid interface is needed to effectively catalyze the 2e À ORR due to the low solubility of O 2 (0.9 mM at 298 K and 1 atm pressure).In this case, the mass-transport rate of O 2 in the gas flow channel is obviously faster than in the solution.In addition, to further alleviate the gas transport limitation, a gas diffusion electrode (GDE) is developed, in which O 2 with a high concentration is imported from the gas phase through a hydrophobic gas diffusion layer (GDL) to directly reach the catalyst layer in front of the electrolyte, [117] forming a three-phase gas-liquid-solid interface, as described in Figure 11b,c.The three-phase gas, ion, and electron interface provide a more accessible catalytic region, thus contributing to a remarkable improvement in current density and O 2 to H 2 O 2 conversion rate. [118,119]However, when operated at a high current, the hydrophobicity of the catalytic layer decreases, leading to an increased diffusion resistance of gaseous O 2 toward the catalytic sites, thus causing a low efficiency for H 2 O 2 production.Worse, electrode flooding by the electrolyte, inevitably causes blockage of the gas diffusion channels and destruction of the three-phase interfaces, which aggravates the performance deterioration.Thus, accelerating the transport and trapping of O 2 in the catalytic region, while avoiding the flooding problem, should be feasible to achieve long-term high-current H 2 O 2 electrosynthesis.To improve gas transport to the catalytic region, several strategies have been described to improve the hydrophobicity of the catalytic layer by adding hydrophobic materials or depositing a hydrophobic oil layer. [118,120,121]Zhou's group [118] used hydrophobic poly(tetrafluoroethylene) (PTFE) material to successfully improve the hydrophobicity of the GDE, achieving remarkably accelerated O 2 diffusion and reaching a high current density of 200 mA cm À2 with an efficiency of 65.7% for H 2 O 2 production.Cordeiro-Junior et al. [122] developed a continuous flow reactor based on the carbon-based Printex L6 GDE as cathode, which allowed the generation of cumulative H 2 O 2 of more than 3 g L À1 at 200 mA cm À2 .Specific H 2 O 2 production of up to 131 g kWh À1 at 25 mA cm À2 was obtained.The durability/lifetime test performed showed that the 40% PTFE-GDE recorded a lifetime of 48 Ah (corresponding to 10 days of uninterrupted use) at 200 mA cm À2 .The lifetime of the 40% PTFE-GDE was found to be 1.3 times longer than that of the 20% PTFE-GDE; in essence, this result demonstrates that an increase in PTFE loading on the GDE resulted in an increase in electrode lifetime.
In another interesting case, Cao et al. [123] presented a highly hydrophobic gas-liquid-solid three-phase architecture formed by densely distributed N-doped carbon (NPC) nanopolyhedra, which exhibited the characteristic of super-aerophilicity to achieve fast transport and trapping of gaseous O 2 even under high-current operation by virtue of electrolytic flooding resistibility.The aerophilicity of the hydrophobic NPC architecture was visibly proven by the fast underwater gaseous O 2 trapping, in stark contrast to the difficult O 2 trapping by the hydrophilic NPC surface.Using the aerophilic three-phase NPC architecture, it can deliver a current of 50-250 mA cm À2 with a current efficiency of 83%-99%, achieving a H 2 O 2 production rate of 8.53 mol g cat À1 h À1 (at 100 mA cm À2 ), which enables the fabrication of high-concentration of H 2 O 2 (0.66-5.38 wt%).The high hydrophobicity feature of the three-phase NPC architecture confers the flood-proof capability that ensures O 2 transport and trapping without blockage, thus enabling the durability for electrocatalytic H 2 O 2 synthesis for 200 h at 100 mA cm À2 that far outperforms its hydrophilic NPC counterparts.
Considering that the H 2 O 2 product is produced in a mixture, additional separation and purification processes are required to recover the pure H 2 O 2 solution.Therefore, it is highly desirable to carry out H 2 O 2 production without electrolyte.For instance, Xia et al. [124] reported a direct electrosynthesis strategy to achieve direct production of pure H 2 O 2 solution in a H 2 /O 2 solid electrolyte reactor (H 2 //SE//O 2 ), in which the H þ and HO 2 À generated through the anode and cathode recombined to form H 2 O 2 , as illustrated in Figure 11d.High selectivity (> 90%) was achieved at current densities up to 200 mA cm À2 for pure H 2 O 2 using functionalized carbon black as the 2e À ORR catalyst, representing a H 2 O 2 productivity of 3.4 mmol cm À2 h À1 .A wide range of pure H 2 O 2 solution concentrations up to 20 wt% could be obtained by tuning the flow rate of water through the solid electrolyte, and the catalyst retained activity and selectivity for 100 h.
Other types of reactors have also been proposed for the electrochemical production of H 2 O 2 , such as phase-transfer device, [125] electrochemical jet-cell, [126] and pressurized reactor. [127]urray et al. [125] reported a redox-mediated phase-transfer (RMPT) electrocatalysis, in which 2,7-disulfonyl anthraquinone (AQDS 2À ) could efficiently catalyze the selective conversion of O 2 to H 2 O 2 with simultaneous separation into an electrolyte-free aqueous product stream, Figure 10e.AQDS 2À was first reduced to AQDSH 2 2À by accepting two electrons, subsequently performed a phase-transfer between the aqueous electrolyte and the organic phase, and finally extracted from the organic phase to form a pure aqueous H 2 O 2 solution.This phase-transfer device can continuously produce H 2 O 2 at 2-3 μmol min À1 cm À2 with a concentration of 33 mM for many hours.Excluding resistivity losses, the process shows an energy efficiency of 40% over a wide range of current densities.In both flow cells and solid-polymer electrolyte (SPE) devices, product crossover is avoided by using polymer electrolyte membranes (PEM), which are known to be degraded by radicals generated from H 2 O 2 self-decomposition, limiting the long-term stability of the device.To this end, Chen et al. [128] developed a membrane-less reactor with carbon-based catalyst for H 2 O 2 Figure 11.a) The components and key operating principles of devices for H 2 O 2 electrosynthesis.Reproduced with permission. [129]Copyright 2022, Wiley-VCH.b) GDE, comprises current collector, gas diffusion layer and catalyst layer, and current collector.Reproduced with permission. [117]Copyright 2021, Royal Society of Chemistry.c) Continuous flow cell by adopting a GDE.Reproduced with permission. [135]Copyright 2020, American Chemical Society.d) Solid-electrolyte fuel cell.Reproduced with permission. [124]Copyright 2019, AAAS.e) Phase-transfer device.Reproduced with permission. [125]Copyright 2019, Elsevier.f ) Membrane-free electrochemical H 2 O 2 generator.Reproduced with permission. [128]Copyright 2017, Royal Society of Chemistry.
production with high Faradaic efficiencies of >90%, requiring cell voltages of only %1.6 V, Figure 11f.In the cathodic compartment, oxygen was reduced to H 2 O 2 and the hydrophobic polymer-coated GDE-blocked H 2 O 2 diffusion, allowing the concentrated H 2 O 2 solution to be collected.

Summary and Perspectives
The electrocatalytic 2e À oxygen reduction reaction (ORR) constitutes an innovative and sustainable way to produce H 2 O 2 .We summarized recent advances on the 2e À ORR research, including basic catalytic mechanisms, influence of pH effect and porous structure on selectivity and activity, selectivity assessment methods, determination of cumulative H 2 O 2 concentration, catalyst development, and electrochemical device design.Several promising biomass-derived carbon-based materials, especially self-doping heteroatom from biomass have been analyzed.The introduction of oxygen groups on biochar can modify the surface physical properties, such as conductivity, reactivity, and hydrophilicity, thus promoting H 2 O 2 production through 2e À ORR pathway.Biochar functionalized with various N-species or multiple heteroatoms is suggested as the active sites facilitating 2e À ORR, although the specific nitrogen-containing functional groups varied among studies.Despite the remarkable progress made in the last decade in porous carbon-based materials doped with biomass-derived heteroatoms, the electrochemical production of H 2 O 2 is still in its infancy and has a long way to go toward its applications at the practical level.The following aspects and research directions deserve attention: 1) Developing of optimal 2e À ORR catalysts: Due to base-catalyzed decomposition, H 2 O 2 is less stable under alkaline conditions, which may limit its applications.Thus, more attention should be paid in the future to explore catalysts that can effectively drive 2e À ORR under acidic or neutral conditions.The combination of experimental measurements and first-principal calculations can help to identify the most desirable structural features of catalyst materials for 2e À ORR.In addition, carbon-based catalysts prepared from Earth-abundant feedstocks, such as biomass, are highly desirable for practical-scale applications.2) Improve catalyst stability: Although it is urgent to solve the problem of activity and selectivity, the long-term stability of the catalyst is essential for reliable and sustainable electrochemical production.In most current studies, the evaluation of catalytic stability usually takes several hours or several tens of hours, which is too short to convincingly demonstrate long-term durability.In addition, because the produced H 2 O 2 must accumulate to a certain high concentration during continuous operation, the catalyst materials must be stable under strong oxidation conditions.Therefore, an extended durability test of the catalyst, e.g., more than 100 h is required, in an environment of high H 2 O 2 concentration, e.g., >3 wt%, and at high current density, e.g., >100 mA cm À2 .3) Understanding the catalytic mechanism: So far, the fundamental understanding of the 2e À ORR reaction mechanisms and stability degradation mechanisms is still insufficient.A combination of operando spectroscopic studies and computational simulations will help unravel the 2e À ORR mechanism for H 2 O 2 selectivity, which in turn can be used to design advanced carbon-based catalysts with higher 2e À ORR activity, H 2 O 2 selectivity, and long-term stability.4) Electrochemical devices design: For industrial-scale application of H 2 O 2 production, in addition to efficient catalysts, other factors such as electrode materials (substrates, binders, and current collectors), electrolytes, ion exchange membrane, reactor configurations, and operating conditions (e.g., temperature, flow rate, applied potential/ current and O 2 partial pressure) must be taken into account.In addition, energy costs must also be taken into account.
.e., the Pauling model, the Griffiths model, and the Bridge model: 1) Pauling model also referred to as "end-on" adsorption through a single bond, in a σ-type bond in which the σ-orbital of O 2 donates electrons to an acceptor d z 2 orbital on the metal; this model prefers to preserve the O-O bond and thus produce H 2 O 2 through the 2e À ORR pathway; 2) Griffiths model, a molecular adsorption of O 2 to the edge of an active site, the bond is formed mainly between the π-orbital of O 2 and an empty d z 2 orbital on the metal surface with a π-back-bonding to form partially filled d xy and d yz orbitals of the metal and an antibonding π* orbital of O 2 ; which would weaken the O-O bond and undergoing 4e À ORR pathway; 3) For the Bridge model, both side-on molecular O 2 atoms are adsorbed on the two active sites, which can elongate and cleave the O-O bond to form the *O intermediate and then reduced to H 2 O via a 4e À pathway.

Figure 3 .
Figure 3. a) Volcano plot of limiting potentials as a function of *OH free energy for the 4e À process, highlighting the regions of strongly bound *OH (solid blue line) and weakly bound *OOH (solid green line); b) volcano plot of limiting potentials as a function of *OH free energy for the 2e À process, highlighting the regions of strongly bound *OH (solid purple line) and weakly bound *OOH (solid green line).Reproduced with permission.[40]Copyright 2018, American Chemical Society.c) The calculated ORR activity volcano plot for 2e À pathway to H 2 O 2 .Red and blue symbols indicate *OOH adsorption at C and Fe, respectively.The equilibrium potential of O 2 /H 2 O 2 is shown as a black dashed line.Reproduced with permission.[43]Copyright 2019, Springer Nature.
D is the total disk current and is the sum of the O 2 reduction currents to H 2 O (I H 2 O ), and to H 2 O 2 (I H 2 O 2 ): I D = I H 2 O þ I H 2 O 2 ; I R is the ring current; N is the collection efficiency of the ring current (I H 2 O 2 ¼ I R N À1 ); F is Faraday constant.

Figure 6 .
Figure 6.Principle and setup of ECL for ORR kinetics evaluation.a) Responses to H 2 O 2 at different concentrations using ECL and RRDE methods.Colored areas represent standard errors.b) Reaction pathways of ORR in the modified Damjanovic model.c) Brief setup of the time-dependent (td)-ECL method for ORR kinetics analysis and electrochemical/ECL signals at different times during td-ECL measurement.The enlarged area and line charts show the distributions of dissolved O 2 , ROS, and Ap* in the diffusion layer.Reproduced with permission.[56]Copyright 2023, Wiley-VCH.

Figure 7 .
Figure7.Schematic of the H 2 O 2 concentration in electrochemical systems determined by three methods: UV-vis spectrophotometry, titration, and colorimetric test strips.Reproduced with permission.[57]Copyright 2020, American Chemical Society.

Figure 8 .
Figure 8. a) Schematic of the manufacturing process of the hierarchically porous catalyst derived from waste pitaya peels; SEM image of PC-1; H 2 O 2 yields and corresponding CE of different cathodes at 70 mA cm À2.Reproduced with permission.[88]Copyright 2022, Elsevier.b) D-glucose formula structure and preparation process for the O-rich CQDs; High-resolution XPS spectra of O 1s of the CQDs; FE of the CQDs.Reproduced with permission.[89]Copyright 2022, Royal Society of Chemistry.c) Schematic of the preparation process for bamboo-derived O-doped rocky electrocatalyst; RRDE curves; Stability test of the O-BC-3h catalyst-coated air-breathing cathode in 0.05 M Na 2 SO 4 .Reproduced with permission.[91]Copyright 2022, Elsevier.
At a potential of 0.423 V versus Ag/AgCl, the H 2 O 2 selectivity of NO/ PC reached 85.3%.The H 2 O 2 concentration only decreased by 51.3 mg L À1 during 10 consecutive experiments (2400 min) at 100 mA cm À2 and pH = 3.0, suggesting that the NO/PC-500 gas diffusion electrode showed good H 2 O 2 electrogeneration performance and appreciable stability at high current density.As shown in Figure 10b, the adsorption energy of *OOH molecule on a pure carbon substrate was À0.35 eV, higher than that of most carbon substrates doped with N/O functional groups, except for carbon substrates doped with pyridinic N and C-O-C (C-Pyri-N-C-O-C, À0.26 eV), indicating that N/O functional groups could enhance the *OOH adsorption ability of carbon materials.The 2e À ORR to H 2 O 2 on NO/PC catalyst was proposed: first, heteroatom doping could lead to charge delocalization of nearby C atoms due to the higher electronegativities of O (3.44) and N (3.04) compared to C (2.55) atom, which resulted in a higher electron density on the C atoms close to the N and O atoms; next, O 2 molecules were more readily absorbed on the NO/PC surface than pure surface due to pyridine N and C-O-C coping; H þ outside bonded to the adsorbed O 2 through

Figure 10 .
Figure 10.a) Schematic diagram of S, N catalytic hydrophobic cathode (SNHC) and S, N Janus Cathode (SNJC) electrode preparation; H 2 O 2 accumulation comparison among various electrodes at 30 mA with air flow rate of 30 mL min À1; current efficiencies and O 2 utilization efficiencies.Reproduced with permission.[71]Copyright 2023, Elsevier.b) The stability of the NO/PC-500-GDE (a) and the restored cathode (b) at 100 mA cm À2 .(NO/PC-500-GDE without a hydrophobic layer electrolyzed at 30 mA cm À2 after 240 min electrolysis is defined as the deactivated cathode.The restored cathode was obtained by introducing a hydrophobic layer on the deactivated cathode).The adsorption energies of OOH and H 2 O 2 molecules on pure carbon and various N/O containing functional group doped carbon substrates; the mechanism of the two-electron ORR for H 2 2 electrogeneration on NO/PC catalysts.Reproduced with permission.[134]Copyright 2021, Royal Society of Chemistry.