Nitrogen and oxygen tailoring of a solid carbon active site for two-electron selectivity electrocatalysis

Abstract

The synthesis of hydrogen peroxide (H₂O₂) via two-electron (2e⁻) oxygen reduction reaction (ORR) electrocatalysis has garnered extensive attention as an appealing green and safe production technology. However, the key progress is still strongly dependent on developing highly active, selective and stable electrocatalysts. Here, guided by first-principles calculations, we developed a rational design for nitrogen (N) and oxygen (O) tailoring of carbon materials by precisely controlling the pyridinic N (Pydn-N) structure and the oxygen doping. The second nearest C atoms of Pydn-N were identified to be the active site, and this is suppressed owing to its local chemical environment and increase in the oxygen dopant content. Experimentally, high-temperature hydrogenation (HTH) regulates the Pydn-N content and synchronous oxygen removal for N,O co-doped carbons (CNO-x). The increment of Pydn-N from 37.42% up to 39.73%, along with oxygen removal from 5.24 at% down to 3.80 at% after HTH, promotes the H₂O₂ selectivity and electron transfer number (n) to 82% and 2.36. The H₂O₂ production rate is stable at around 200 mmol g_cat⁻¹ h⁻¹ and a faradaic efficiency (F.E.) of up to 80% was recorded during the initial 4 h. The long-term H₂O₂ production further highlights the significance of sustaining the Pydn-N structure, modulating the solid carbon active site.

---

1. Introduction

Hydrogen peroxide (H₂O₂), as a green, versatile and multifunctional chemical, has been extensively applied in pulp and paper bleaching, medicine and environmental protection.^1–3 The global market demand for H₂O₂ is enormous, projected to reach 6000 kilotons in 2024.⁴ However, the traditional anthraquinone process for the synthesis of H₂O₂ is not a green process and generates substantial waste chemicals.^5,6 Alternatively, H₂O₂ production through the oxygen reduction reaction (ORR) is highly desirable with the advantages of green precursors and renewable energy sources.^7–9 However, the ORR involves two competitive processes in which the oxygen (O₂) can be converted to H₂O through a four-electron (4e⁻) pathway or a two-electron (2e⁻) one to form H₂O₂.¹⁰ For H₂O₂ production, owing to the difficult control in triggering the 2e⁻ pathway and the subsequent suppression of the following 2e⁻ step, electrocatalysts with a high activity and selectivity are a prerequisite for ORR for the 2e⁻ process.¹¹ Therefore, the rational design of a catalyst with a high selectivity, activity and stability is the key to the electrochemical synthesis of H₂O₂ through the ORR.

To overcome the dilemmas for electrocatalysts in terms of compatible activity and selectivity during the 2e⁻ ORR process, a substantial number of studies have focused on noble metals and their alloys (such as Pt,^12–15 Pd,^16,17 Pt–Hg¹⁸ and Pd–Hg^19,20), which exhibit a high efficiency for H₂O₂ production both in rate and selectivity. However, their nature, in terms of the scarcity and inferior stability of noble metal-based catalysts, hinders their large-scale application.^11,21 Cost-effective carbon materials with a nanoscale two-dimensional (2D) morphology and chemically functionalized surface have recently been regarded as promising catalysts for electrochemical synthesis of H₂O₂ through the ORR.^10,22–24 Currently, heteroatom doping is an effective route to regulate the electronic structure of carbon nanomaterials and simultaneously provide catalytic active sites for the ORR, thus enabling this approach to tailor the catalytic activity and selectivity towards H₂O₂ electrosynthesis.^25–28 In particular, nitrogen (N) doping plays an important role in enhancing the ORR activity and selectivity for H₂O₂ production.^29–31 For the N-doped carbon catalyst, the N atoms with a higher electronegativity create a net positive charge on the adjacent carbon atoms to influence the adsorption properties toward the intermediate (OOH* for 2e⁻ process).^10,25 Specifically, the contents and bonding states of nitrogen are crucial to affecting the catalytic activity and selectivity. Fellinger et al. reported an N-doped mesoporous carbon material with a favoured 2e⁻ ORR process, which was derived from the incremental nitrogen content and the pyrrolic nitrogen (Pyli-N) sites, which had a radical character. Moreover, pure pyridinic N (Pydn-N) doped graphene has been reported to be a good catalyst towards the synthesis of H₂O₂ with a high selectivity.³² It is preferential to accurately regulate the local chemical environment of the carbon atoms using N doping and to further elaborately study their correlations with the catalytic active site and selectivity. In addition to the N doping, the oxygen (O) is also considered to be an effective heteroatom for improving the selectivity of 2e⁻ ORR.^33–36 Cui and co-workers demonstrated the oxidized O-CNTs with the active site of the carbon atoms adjacent to several oxygen functional groups, showing an excellent activity and high selectivity toward H₂O₂ electrosynthesis.⁵

Notably, hydrogenation can modulate and redistribute the content and type of N atoms and partly remove the O atoms³⁷ for doped carbon materials. Inspired by this research progress, a large challenge remains in the design and synthesis of N and O co-doped carbon materials by chemically tailoring the catalytic active site for green and safe H₂O₂ production and to explore the mechanism of the local N and O microstructure determined active site for selective 2e⁻ ORR.

Herein, we report the density functional theory (DFT) calculation guided rational design of N and O tailored carbon materials by precisely controlling the Pydn-N structure and oxygen doping for H₂O₂ synthesis. The second nearest C atoms of Pydn-N are predicted to act as the active site for 2e⁻ ORR. A series of N,O co-doped carbon materials (CNO-x) were synthesized, and following high-temperature hydrogenation (HTH) under Ar/H₂ atmosphere, were used to prepare the CNO-x-H. It was sufficiently demonstrated that HTH can generally regulate the Pydn-N structure and content, and the synchronous oxygen removal. The increment of the Pydn-N structure along with the oxygen removal for CNO-x-H, obviously promotes the H₂O₂ selectivity and electron transfer number (n) during the ORR. We also observed that the CNO-glu with a higher Pydn-N ratio exhibited a better H₂O₂ selectivity in comparison with the CNO-cyc and CNO-cel, while the CNO-glu-H, CNO-cyc-H and CNO-cel-H showed no difference in the Pydn-N ratio and 2e⁻ selectivity. The H₂O₂ production rate is stable during the initial 4 h. Long-term H₂O₂ production operation deteriorated the activity of CNO-x-H owing to breaking of the solid carbon active site caused by reduction of the Pydn-N structure.

2. Experimental

2.1 First-principles calculations

The DFT calculations were performed using the Vienna Ab initio Simulation Package (VASP).^38,39 Core electrons are described using the projector augmented wave (PAW) pseudopotentials.⁴⁰ We utilized the BEEF-vdW exchange correlation functional, which uses the vdW-DF2 nonlocal correlation energy and potential to estimate the effect of van der Waals interactions and this has been shown to accurately describe the chemisorption and physisorption properties on graphene.⁴¹ The plane-wave kinetic energy cutoff of 400 eV with a Gaussian smearing width of 0.1 eV, and Monkhorst–Pack k-point grids of 3 × 3 × 1 were adopted to ensure the convergence of the total-energy calculations. In addition, the convergence criterion of energy and forces were set to be 10⁻⁵ eV and 0.01 eV Å⁻¹, respectively. The constructed model structures were built by doping different types of N and O atoms into the 3 × 4 supercell of the orthogonal unit cell of single-layer graphite. The optimized lattice parameters for the model systems are 21.3 × 16.3 Å. All the single layer structures were separated by 20 Å of vacuum space perpendicular to the slab surface. Further details on the DFT calculations are provided in the ESI.†

2.2 Synthesis of CNO-x (x = glu, cyc and cel)

The CNO-x (x = glu, cyc and cel) were prepared according to the method described in our previous study.⁴² 1 g of glucose and 10 g of urea were mixed and ground using a mortar. Then, the mixture was placed in a tightly sealed crucible and calcined in a muffle furnace. The calcination temperature was first set at 550 °C for 1 h, then it was reduced down to 200 °C for 0.5 h. Subsequently, the temperature was raised to 900 °C and maintained for 1 h. The finally obtained black powder was named as CNO-glu. CNO-cyc and CNO-cel were prepared using the same method with cyclodextrin and cellulose being used separately to replace glucose.

2.3 Synthesis of CNO-x-H (x = glu, cyc and cel)

All these CNO-x-H materials were obtained by hydrogenation treatment, in which CNO-glu, CNO-cyc and CNO-cel were placed in a tube furnace at 500 °C for 1 h under an Ar/H₂ atmosphere, respectively.

2.4 Characterizations

The morphology was observed using scanning electron microscopy (SEM, Hitachi S-4800, 15 kV) and transmission electron microscopy (TEM, JEOL JEM F200). The X-ray diffractometry (XRD) patterns were obtained using a SmartLab (Rigaku 9 kW). The Raman spectra analysis were recorded on a Renishaw Invia Raman spectrometer. X-ray photoelectron spectroscopy (XPS) spectra were conducted with an ESCALAB 250 X-ray photoelectron spectrometer using Al Kα radiation as an excitation source with a pass energy of 30 eV.

2.5 Electrochemical characterization

The electrochemical performances were measured using a multichannel electrochemical workstation (Bio-Logic VMP-3) with a three-electrode system. The catalyst ink was prepared by dispersing 2 mg of the as-prepared catalyst in 1 ml ethanol with 30 μl of 5 wt% Nafion. A rotating ring-disk electrode (RRDE, 5 mm) was used as a working electrode (10 μL of catalyst ink was dropped onto the disk electrode). A platinum wire electrode and a KCl saturated Ag/AgCl electrode (or a Hg/HgO electrode) served as the counter electrode and the reference one, respectively. The activity and selectivity were investigated using linear sweep voltammetry (LSV) measurements in an O₂-saturated 0.1 M KOH electrolyte. During the testing process, a potential of 1.5 V versus reversible hydrogen electrode (RHE) was applied on the ring electrode. The working electrode worked under a speed of 1600 rpm. The electron transfer number (n) and the H₂O₂ selectivity were calculated using the following formulas:

In which, I_R is the ring current, I_D is the disk current and N is the collection efficiency of the ring electrode (N = 0.38).

The Nernst equation for converting the reference electrode (Ag/AgCl or Hg/HgO) potential into the reversible hydrogen electrode (RHE) potential was formulated according to the following formulas:

E(RHE) = E(Ag/AgCl) + 0.0591 pH + 0.1976 (3)

E(RHE) = E(Hg/HgO) + 0.0591 pH + 0.098 (4)

Electrocatalytical H₂O₂ production was performed in an H-cell with a Nafion 117 membrane as a separator. A carbon fibre paper (area 1 × 1 cm²) loaded with 1 mg of CNO-glu-H as the working electrode and a KCl saturated Ag/AgCl electrode (or Hg/HgO) as the reference electrode was placed in the cathode compartment. A platinum sheet electrode as the counter electrode was placed in the anode compartment. Both the cathode and anode compartment were filled with 35 mL 0.1 M KOH solution. The chronoamperometry curves applied at 0.5 V versus RHE were recorded in an O₂-saturated electrolyte.

2.6 H₂O₂ concentration measurements

The cerium sulfate (Ce(SO₄)₂) titration method was employed to measure the H₂O₂ concentration based on the mechanism that the yellow transparent Ce⁴⁺ can be reduced to colourless Ce³⁺ by H₂O₂. Thus, the content of H₂O₂ can be calculated based on the consumption of Ce⁴⁺ determined using a ultraviolet-visible (UV-Vis) spectrometer (UV2310II, Techcomp). 404 mg of Ce(SO₄)₂ was dissolved in 500 mL of 0.5 M H₂SO₄ to prepare 10 mM of Ce(SO₄)₂ solution. For the calibration curve, a series of H₂O₂ solutions with known concentration were respectively added to Ce(SO₄)₂ solution and measured using the UV-Vis spectrometer (Fig. S8†). 1 mL of the electrolyte after the chronoamperometry measurement was added to the Ce(SO₄)₂ solution to calibrate the concentration of H₂O₂.

3. Results and discussion

DFT calculations were employed to explore the catalytic activity and active site for the two-electron (2e⁻) ORR to H₂O₂ on different types of N and O atoms doped carbon model structures. The model configurations are shown in Fig. 1A, in which all models contain one Pydn-N and Pyli-N on the zigzag and armchair edges, respectively. In addition, one O atom has been introduced at the edge or skeletal location of the graphene-like framework, except for the model labelled “Edge_C” without O doping. To understand the catalytic activities of all model structures, the binding energies of the reaction intermediates to the active sites (the carbon atoms denoted by dashed magenta circles in Fig. 1A) were calculated using DFT. The 2e⁻ ORR comprises two elementary steps:

O₂ + * + (H⁺ + e⁻) → OOH* (5)

OOH* + (H⁺ + e⁻) → H₂O₂ + * (6)

Fig. 1 Theoretical prediction for the 2e⁻ ORR activity for different model systems. (A) Schematic diagram of the topological models for the N and O tailoring of carbons examined in this study. Colour code: C, grey; N, blue; O, red; H, white. The carbon atoms denoted by dashed magenta circles are the active sites under investigation. (B) Calculated catalytic activity for 2e⁻ ORR to H₂O₂ displayed with the limiting potential plotted as a function of ΔG_OOH*. The equilibrium potential for the two-electron ORR is shown as the dashed black line.

Here, * denotes the active sites. Eqn (5) represents the direct proton/electron transfer process to the adsorbed molecular O₂ to produce OOH*. Furthermore, eqn (6) describes the production of H₂O₂via the second proton/electron transfer. Obviously, the pathway involves only OOH* as the reaction intermediate. Fig. 1B shows the volcano plot for the two-electron ORR activity with the descriptor of ΔG_OOH*. The calculated limiting potential (U_L) can be defined as the highest potential at which all the reaction steps are downhill in free energy. In addition, the theoretical overpotential is defined as the difference between the limiting potential and the equilibrium potential (U = 0.7 V vs. RHE).⁵ For the two-electron ORR, the overpotential is either due to the protonation of oxygen (eqn (5)) or the reduction of OOH* to form H₂O₂ (eqn (6)). The positioning of the structures at the left side of the volcano plot present strong binding toward OOH*, hence, eqn (6) is their rate-determining step. However, the ones located at the right side weakly bind OOH*, thus, eqn (5) is the rate-determining one. According to the calculations results, the second nearest carbon atom of Pyli-N (see model O1_PyliN_C2) and the carbon atom far from the N atoms (see model Edge_C) exhibit no significant contribution to the 2e⁻ ORR. In comparison, the second nearest carbon atoms of Pydn-N (see model O1_PydnN_C4 and O1_PydnN_C2) are highly active for the 2e⁻ ORR with overpotentials of 0.04 and 0.02 V, respectively. Moreover, when the location of the doped O atom is close to Pydn-N (see model O2_PydnN_C2), the overpotential increases to 0.12 V, and the overpotential further rises to 0.63 V when an O atom occurs in the skeletal location. Therefore, the second nearest C atoms of Pydn-N are predicted to be the active site for the 2e⁻ ORR, in which the catalytic activity would also be affected by the configuration and location of the doped O atom. Furthermore, the 2e⁻ ORR activity of the model structures with various O doping amounts have also been investigated. Models with no O, or only one O atom doped at an edge location, are predicted to give a superior 2e⁻ ORR performance (Fig. S1A–D†). As for the models with more O atoms (Fig. S1E–H†), the 2e⁻ ORR pathway is impractical owing to the uphill free energy in the first elementary step in eqn (5). Consequently, a rational design strategy for N and O tailored carbon materials for 2e⁻ ORR to H₂O₂ production will promote the Pydn-N structure and related content, and reduce oxygen doping to precisely control its chemical location.

Based on the DFT calculations, we focused on designing and synthesizing a series of N,O co-doped carbon materials and tailoring the catalytic active site of C atoms by regulating the content of Pydn-N and oxygen atoms. It has been reported that various types of N structures would be affected by high-temperature treatment and the oxygen atoms would be removed under the H₂ reducing atmosphere.^28,37 Hence, the hydrogenation temperature is critical in modulating the C atom active site for the 2e⁻ ORR performance improvement. The CNO-x (x = glu, cyc and cel) was synthesized using a one-step self-supporting solid-state pyrolysis (OSSP) technique, in which temperature terrace calcination was used to construct graphene-like oxygenated carbon nitride (OCN) materials.⁴² As shown in Scheme 1, at the first stage at 550 °C, the graphitic carbon nitride (g-C₃N₄) was produced by polycondensation of urea to serve as the 2D layered template. Simultaneously, amorphous carbon pitch with oxygen-containing functional groups was produced from glucose (cyclodextrin and cellulose). At the second stage at 900 °C, the g-C₃N₄ with amino groups reacted with the carbonyl groups of carbon pitch through the Maillard reaction⁴³ to chemically graft both components. Moreover, the evolved gases from the second stage expanded into the interlayers and finally generated the CNO-x nanosheets. Subsequently, the CNO-x was further calcinated by HTH³⁷ under an Ar/H₂ atmosphere to remove the O atoms to prepare the CNO-x-H.

Scheme 1 Schematic illustration of CNO-x-H prepared by OSSP followed by HTH.

The morphology and microstructure were first investigated using SEM and TEM. According to Fig. 2A and B, both CNO-glu and CNO-glu-H present a two-dimensional (2D) nanosheet morphology with abundant wrinkles. In addition, the CNO-glu-H exhibits a much more wrinkled defect structure than that of CNO-glu, which was further confirmed by the TEM images (Fig. 2C and D). The elemental mapping analysis shows that the C, N and O elements were uniformly distributed among the 2D nanosheet, but the O content decreased for CNO-glu-H after the hydrogenation treatment. Then, XRD was carried out to identify the phase structure. As shown in Fig. 2E, the diffraction peaks centred at 26.2° and 42.2° are indexed to the (002) and (100) plane of graphite (JCPDS no. 75-1621), which demonstrates that no change in the graphitic phase and condensation state occurs after the HTH process. Following this, the Raman spectrum was measured to analyse the defect microstructure features. As shown in Fig. 2F, the Raman spectra between 1000 to 2000 cm⁻¹ are deconvoluted into four bands, including impurity (I), in-plane defect (D), interstitial defects (D′′) and graphitic structure (G).^44,45 The intensity ratio of D and G (I_D/I_G), an indicator for estimating the degree of disorder in carbon materials, in CNO-glu-H decreases slightly from 1.28 to 1.03, indicating the suppression of the in-plane defect by taking away the O atoms after HTH treatment.

Fig. 2 SEM images of (A) CNO-glu and (B) CNO-glu-H. TEM and elemental mapping images of (C) CNO-glu and (D) CNO-glu-H. (E) XRD patterns of CNO-glu and CNO-glu-H. (F) Raman spectra of CNO-glu and CNO-glu-H.

Moreover, XPS analysis was performed to detect the elemental constituents and bonding state of C, N and O. As shown in Table 1, around 13.23 at% of nitrogen and 5.24 at% of oxygen were measured in CNO-glu. After HTH treatment, the N content in CNO-glu-H increases slightly to 13.55 at%, while the O decreases to 3.80 at%. High resolution XPS spectroscopy further demonstrates the existence of multiple types of C and N bonding states. The deconvolution C 1s spectrum shown in Fig. 3A and B reveal four peaks: carbon in graphite (C–C) at 284.7 eV, carbon bound to nitrogen (C–N) at 285.7 eV, carbon singly bound to oxygen (C–O) at 286.6 eV, and carbon bound to two N/O atoms (O–C=O/N–C=N) at 288 eV.⁴⁶ The N 1s spectrum can be deconvoluted into four bands (Fig. 3C and D): Pydn-N at 398.4 eV; Pyli-N at 399.9 eV; graphitic N (Grap-N) at 401 eV; and oxidized N at 402.2 eV.^47,48 The ratios of the different C and N types are summarized in Tables 1 and S1.† For CNO-glu, the content of C–C, C–N, C–O and O–C=O/N–C=N is approximately 62.16%, 16.05%, 15.00% and 6.79%, respectively. For CNO-glu-H, the C–N concentration increases to 21.17% while the C–O concentration decreases to 10.41%, in accordance with the variation of the elemental constituent ratio. Notably, the content of Pydn-N in CNO-glu-H increases from 37.42% to 39.73% after HTH treatment. A similar phenomenon for elemental constituents and bonding states was also observed for CNO-glu-H. These results sufficiently demonstrate that the HTH can experimentally regulate the Pydn-N structure and content, and is synchronous to the removal of doping oxygen.

Fig. 3 Pydn-N structure and content determined by XPS. (A) and (B) High-resolution C 1s spectrum of CNO-glu and CNO-glu-H. (C) and (D) High-resolution N 1s spectrum of CNO-glu and CNO-glu-H.

Table 1 Atomic concentration (at%) of elements and the proportion (at%) of bonding types of nitrogen atoms determined by XPS

Sample	Composition (at%)			N 1s bonding types (at%)
	C	N	O	Pyridinic N	Pyrrolic N	Graphitic N	Oxided N
CNO-glu	81.53	13.23	5.24	37.42	24.54	31.45	6.59
CNO-glu-H	82.65	13.55	3.8	39.73	26.50	22.58	11.19
CNO-cyc	81.54	12.37	6.08	36.72	25.25	29.48	8.55
CNO-cyc-H	83.12	12.86	4.03	39.47	25.76	25.19	9.6
CNO-cel	86.05	8.25	5.7	35.96	26.67	28.47	8.88
CNO-cel-H	84.83	10.81	4.36	40.43	23.04	28.88	7.66

---

The 2e⁻ ORR activity and selectivity were investigated using a RRDE. The H₂O₂ oxidation currents (orange lines) were measured on a platinum ring electrode held at 1.5 V versus RHE, along with the oxygen reduction currents (cyan lines) measured on a disk electrode loaded the CNO catalysts. As shown in Fig. 4A, the CNO-glu exhibits catalytic activity for 2e⁻ ORR and an onset potential at 0.73 V versus RHE (defined as the potential³⁷ at a current density of 0.5 mA cm⁻²). The electron transfer number and H₂O₂ selectivity were calculated using eqn (1) and (2), respectively. The ORR takes place at the disk electrode, and the H₂O₂ produced at the disk electrode is radially transferred to the concentric platinum ring electrode by the forced convection caused by the rotating motion of the electrode. Subsequently, H₂O₂ is reoxidized back to O₂ at the platinum ring electrode. We can calculate the electron transfer number and H₂O₂ selectivity according to the disk current (I_D) and ring current (I_R), which indicates the fraction of O₂ used for H₂O₂. The CNO-glu displays a selectivity of 68% for 2e⁻ ORR with a transfer electron number of 2.64 at 0.5 V versus RHE (Fig. 4B). For the CNO-glu-H, as shown in Fig. 4C, a similar onset potential of 0.73 V versus RHE was performed, which indicates similar active sites present in both catalysts. Interestingly, the H₂O₂ selectivity and electron transfer number changed to 82% and 2.36 (Fig. 4D), which is better than some of the reported 2e⁻ ORR carbon catalysts (Table S2†). The catalytic performance of CNO-glu-H was also measured using a Hg/HgO reference electrode (Fig. S2†) to eliminate the effects of the reference in alkine solution. The CNO-glu-H shows a similar 2e⁻ ORR performances for the Hg/HgO reference electrode in the 0.1 M KOH solution. As predicted, the 2e⁻ ORR catalytic activity of CNO-glu is obviously improved after HTH treatment. This experimental demonstration directly denotes the key roles of tailoring the Pydn-N structure and oxygen dopants in promoting the solid carbon active sites for selective H₂O₂ production, which are in good accordance with the predictions from the DFT calculations.

Fig. 4 N and O tailoring of the solid carbon active site for 2e⁻ ORR selectivity. (A) Polarization curves (orange line) and simultaneous H₂O₂ detection currents at the ring electrode (cyan line) for CNO-glu in O₂-saturated 0.1 M KOH solution at 1600 rpm. (B) H₂O₂ selectivity and transfer electron number for CNO-glu. (C) Polarization curves (orange line) and simultaneous H₂O₂ detection currents at the ring electrode (cyan line) for CNO-glu-H in O₂-saturated 0.1 M KOH solution at 1600 rpm. (D) H₂O₂ selectivity and transfer electron number for CNO-glu-H.

To demonstrate the general principle of precisely tailoring the Pydn-N structure and content, and strictly controlling the oxygen dopants using the HTH approach for N,O co-doped carbon to improve the 2e⁻ ORR, we further examined two other CNO-cyc and CNO-cel samples. Similar to CNO-glu-H, the CNO-cyc-H and CNO-cel-H show the same trends in the structural, morphological and constituent evolution. A more wrinkled surface (Fig. S3†) without phase change (Fig. S4†) was observed after HTH treatment. For Raman spectra (Fig. S5†), the CNO-cyc-H and CNO-cel-H also display a reduced I_D/I_G of 1.19 and 1.18 in comparison to that of CNO-cyc (1.21) and CNO-cel (1.20), indicating a similar suppression of the in-plane defects. In detail, the XPS survey results show that the O content from 6.08 at% for CNO-cyc and 5.7 at% for the CNO-cel decreases down to 4.03 at% for CNO-cyc-H and 4.36 at% for the CNO-cel-H. The N content, from 12.37 at% for CNO-cyc and 8.25 at% for CNO-cel, increases to 12.86 at% for CNO-cyc-H and 10.81 at% for CNO-cel-H (Fig. S6†). The Pydn-N ratio rises up to 39.47% for CNO-cyc-H and 40.43% for CNO-cel-H, relative to that of 36.72% for CNO-cyc and 35.96% for CNO-cel (Fig. 5C–F and Table 1). The C 1s spectrum exhibits four similar deconvoluted peaks attributed to the C–C, C–N, C–O, and O–C=O/N–C=N bonds for the CNO-cyc-H and CNO-cel-H, but presents significantly more C–N bonds than CNO-cyc and CNO-cel owing to the removal of O atoms after HTH treatment (Fig. S7†).

Fig. 5 General principle of Pydn-N structure modulating carbon active site. (A) Polarization curves (solid lines) and simultaneous H₂O₂ detection currents at the ring electrode (dashed lines) in O₂-saturated 0.1 M KOH solution at 1600 rpm for CNO-cyc, CNO-cel, CNO-cyc-H and CNO-cel-H. (B) H₂O₂ selectivity and transfer electron number for CNO-cyc, CNO-cel, CNO-cyc-H and CNO-cel-H. High-resolution N 1s peaks of (C) CNO-cyc, (D) CNO-cel, (E) CNO-cyc-H and (F) CNO-cel-H.

In addition, according to the RRDE results shown in Fig. 5A, these four catalysts (CNO-cyc and CNO-cyc-H, CNO-cel and CNO-cel-H) display a similar onset potential, polarization current and 2e⁻ ORR activity. Similar to the CNO-glu-H series catalyst, the CNO-cyc-H and CNO-cel-H exhibits a higher H₂O₂ selectivity (column charts) and 2e⁻ ORR pathway (point segment charts) (see Fig. 5B). The H₂O₂ selectivity increases from 55.4% for CNO-cyc and 54.3% for CNO-cel to 78.5% for CNO-cyc-H and 82.3% for CNO-cel-H, respectively. Simultaneously, the electron transfer number for CNO-cyc and CNO-cel decreases down to 2.43 for CNO-cyc-H and 2.36 for CNO-cel-H from 2.89 and 2.91, respectively. This is consistent with the observations for CNO-glu and CNO-glu-H. Moreover, the effects of the structure and content of Pydn-N and the oxygen dopants on the solid carbon active sites for selective 2e⁻ ORR are consequently confirmed. Interestingly, the CNO-glu has a higher Pydn-N ratio with a better H₂O₂ selectivity among the initial three catalysts (CNO-glu, CNO-cyc and CNO-cel), while the CNO-glu-H, CNO-cyc-H and CNO-cel-H show no difference in the Pydn-N ratio and 2e⁻ selectivity. This analysis experimentally confirms our DFT prediction that the second nearest C atom of Pydn-N is a highly active site for 2e⁻ ORR and the reducing O content is further beneficial to the 2e⁻ ORR process.

Finally, as critical criteria for practical applications, the H₂O₂ production rate and long-term stability were also examined. The H₂O₂ production rate on CNO-glu-H was measured using an H-cell with chronoamperometry operation for hours. As shown in Fig. 6A, the H₂O₂ production rate is stable around 200 mmol g_cat⁻¹ h⁻¹ (the calibration curve of H₂O₂ is shown in Fig. S8†) and the faradaic efficiency (F.E.) is maintained at up to 80% during the initial 4 h. However, an obvious decrease occurs down to 126 mmol g_cat⁻¹ h⁻¹ at the 5^th hour, indicating that the catalyst has an inferior catalytic stability above a 4 h long-term extension. The i–t curves show the disk current loss of CNO-glu-H is 26% after 5 h and 31% after 6 h (Fig. 6B). In addition, SEM and XPS analysis were employed to investigate the morphology and composition information during the stability test. The SEM image in Fig. 6C reveals that the CNO-glu-H retains a two-dimensional (2D) morphology without obvious change. However, the N 1s spectrum of CNO-glu-H displays an evident decrease of Pydn-N from 39.73% to 30.88% after the 5 h stability test (Fig. 6D), which further demonstrates the remarkable tailoring function of Pydn-N in determining the solid carbon active site for 2e⁻ ORR. After the long-term chronoamperometry operation, the Pydn-N structure was partly destroyed leading to the reduction of the Pydn-N tailored solid C active sites for the 2e⁻ ORR process. Hence, it is still a big challenge for these carbon electrocatalysts to sustain the Pydn-N structure modulating active site, namely the second nearest C atom of Pydn-N, under long-term H₂O₂ production operation conditions. A novel strategy for dynamic active site regeneration still needs to be explored in the future.

Fig. 6 H₂O₂ production and long-term stability. (A) H₂O₂ production rate on CNO-glu-H at various electrolysis extensions. (B) Current–time (i–t) chronoamperometric responses of CNO-glu-H in O₂-saturated 0.1 M KOH solution at 1600 rpm. (C) SEM image of CNO-glu-H after 5 h chronoamperometry test. (D) High-resolution N 1s spectrum of CNO-glu-H after 5 h H₂O₂ production test.

4. Conclusions

In this work, we have theoretically and experimentally demonstrated the modulation of Pydn-N and O dopants on a solid carbon active site for 2e⁻ ORR selectivity. The second nearest C atoms of Pydn-N are the predicted active sites, and would be suppressed upon an increase in the doped oxygen or reducing the Pydn-N structure. A novel strategy for dynamic active site regeneration still needs to be explored to sustain the Pydn-N structure modulated active site under long-term H₂O₂ production operation conditions. This work provides a prospective and practical avenue to combine theoretical predications with experimental demonstrations to develop advanced carbons with active sites for 2e⁻ ORR electrocatalysis for green and safe H₂O₂ synthesis.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 21961024, 21961025 and 11904187), the Inner Mongolia Natural Science Foundation (No. 2018JQ05, 2019BS02007), the Inner Mongolia Autonomous Region Science & Technology Planning Project for Applied Technology Research and Development (No. 2019GG261), Incentive Funding from Nano Innovation Institute (NII) of Inner Mongolia University for Nationalities (IMUN), the Inner Mongolia Autonomous Region Incentive Funding Guided Project for Science & Technology Innovation (2016), the Inner Mongolia Autonomous Region Funding Project for Science & Technology Achievement Transformation (No. CGZH2018156), the Doctoral Scientific Research Foundation of Inner Mongolia University for Nationalities (No. BS480, BS437) and the Scientific Research Program of Inner Mongolia University for Nationalities (No. NMDYB19040).

Notes and references

1. C. A. Martinez-Huitle and S. Ferro, Electrochemical oxidation of organic pollutants for the wastewater treatment: direct and indirect processes, Chem. Soc. Rev., 2006, 35, 1324–1340.
2. J. M. Campos-Martin, G. Blanco-Brieva and J. L. Fierro, Hydrogen peroxide synthesis: an outlook beyond the anthraquinone process, Angew. Chem., Int. Ed., 2006, 45, 6962–6984.
3. I. Yamanaka, T. Onizawa, S. Takenaka and K. Otsuka, Direct and continuous production of hydrogen peroxide with 93% selectivity using a fuel-cell system, Angew. Chem., Int. Ed., 2003, 42, 3653–3655.
4. H. W. Kim, M. B. Ross, N. Kornienko, L. Zhang, J. Guo, P. Yang and B. D. McCloskey, Efficient hydrogen peroxide generation using reduced graphene oxide-based oxygen reduction electrocatalysts, Nat. Catal., 2018, 1, 282–290.
5. Z. Lu, G. Chen, S. Siahrostami, Z. Chen, K. Liu, J. Xie, L. Liao, T. Wu, D. Lin, Y. Liu, T. F. Jaramillo, J. K. Nørskov and Y. Cui, High-efficiency oxygen reduction to hydrogen peroxide catalysed by oxidized carbon materials, Nat. Catal., 2018, 1, 156–162.
6. W. Zhou, X. Meng, J. Gao and A. N. Alshawabkeh, Hydrogen peroxide generation from O₂ electroreduction for environmental remediation: A state-of-the-art review, Chemosphere, 2019, 225, 588–607.
7. S. Yang, A. Verdaguer-Casadevall, L. Arnarson, L. Silvioli, V. Čolić, R. Frydendal, J. Rossmeisl, I. Chorkendorff and I. E. L. Stephens, Toward the Decentralized Electrochemical Production of H₂O₂: A Focus on the Catalysis, ACS Catal., 2018, 8, 4064–4081.
8. E. Pizzutilo, S. J. Freakley, S. Cherevko, S. Venkatesan, G. J. Hutchings, C. H. Liebscher, G. Dehm and K. J. J. Mayrhofer, Gold–Palladium Bimetallic Catalyst Stability: Consequences for Hydrogen Peroxide Selectivity, ACS Catal., 2017, 7, 5699–5705.
9. Y. Lei, S. Lei and L. Piao, Study on the preparation of H₂O₂ by photocatalytic reduction of oxygen, Prog. Chem., 2020 DOI:10.7536/PC200463.
10. Y. Jiang, P. Ni, C. Chen, Y. Lu, P. Yang, B. Kong, A. Fisher and X. Wang, Selective Electrochemical H₂O₂ Production through Two-Electron Oxygen Electrochemistry, Adv. Energy Mater., 2018, 8, 1801909.
11. S. Chen, Z. Chen, S. Siahrostami, T. R. Kim, D. Nordlund, D. Sokaras, S. Nowak, J. W. F. To, D. Higgins, R. Sinclair, J. K. Nørskov, T. F. Jaramillo and Z. Bao, Defective Carbon-Based Materials for the Electrochemical Synthesis of Hydrogen Peroxide, ACS Sustainable Chem. Eng., 2017, 6, 311–317.
12. C. M. Sanchez-Sanchez and A. J. Bard, Hydrogen peroxide production in the oxygen reduction reaction at different electrocatalysts as quantified by scanning electrochemical microscopy, Anal. Chem., 2009, 81, 8094–8100.
13. R. Shen, W. Chen, Q. Peng, S. Lu, L. Zheng, X. Cao, Y. Wang, W. Zhu, J. Zhang, Z. Zhuang, C. Chen, D. Wang and Y. Li, High-Concentration Single Atomic Pt Sites on Hollow CuSx for Selective O₂ Reduction to H₂O₂ in Acid Solution, Chem, 2019, 5, 2099–2110.
14. S. Yang, Y. J. Tak, J. Kim, A. Soon and H. Lee, Support Effects in Single-Atom Platinum Catalysts for Electrochemical Oxygen Reduction, ACS Catal., 2017, 7, 1301–1307.
15. S. Yang, J. Kim, Y. J. Tak, A. Soon and H. Lee, Single-Atom Catalyst of Platinum Supported on Titanium Nitride for Selective Electrochemical Reactions, Angew. Chem., Int. Ed., 2016, 55, 2058–2062.
16. D. C. Ford, A. U. Nilekar, Y. Xu and M. Mavrikakis, Partial and complete reduction of O₂ by hydrogen on transition metal surfaces, Surf. Sci., 2010, 604, 1565–1575.
17. J. S. Jirkovsky, I. Panas, E. Ahlberg, M. Halasa, S. Romani and D. J. Schiffrin, Single atom hot-spots at Au-Pd nanoalloys for electrocatalytic H₂O₂ production, J. Am. Chem. Soc., 2011, 133, 19432–19441.
18. S. Siahrostami, A. Verdaguer-Casadevall, M. Karamad, D. Deiana, P. Malacrida, B. Wickman, M. Escudero-Escribano, E. A. Paoli, R. Frydendal, T. W. Hansen, I. Chorkendorff, I. E. Stephens and J. Rossmeisl, Enabling direct H₂O₂ production through rational electrocatalyst design, Nat. Mater., 2013, 12, 1137–1143.
19. A. Verdaguer-Casadevall, D. Deiana, M. Karamad, S. Siahrostami, P. Malacrida, T. W. Hansen, J. Rossmeisl, I. Chorkendorff and I. E. Stephens, Trends in the electrochemical synthesis of H₂O₂: enhancing activity and selectivity by electrocatalytic site engineering, Nano Lett., 2014, 14, 1603–1608.
20. D. Deiana, A. Verdaguer-Casadevall, P. Malacrida, I. E. L. Stephens, I. Chorkendorff, J. B. Wagner and T. W. Hansen, Determination of Core-Shell Structures in Pd-Hg Nanoparticles by STEM-EDX, ChemCatChem, 2015, 7, 3748–3752.
21. X. Xiao, T. Wang, J. Bai, F. Li, T. Ma and Y. Chen, Enhancing the Selectivity of H₂O₂ Electrogeneration by Steric Hindrance Effect, ACS Appl. Mater. Interfaces, 2018, 10, 42534–42541.
22. T. P. Fellinger, F. Hasche, P. Strasser and M. Antonietti, Mesoporous nitrogen-doped carbon for the electrocatalytic synthesis of hydrogen peroxide, J. Am. Chem. Soc., 2012, 134, 4072–4075.
23. Y. Sun, S. Li, Z. P. Jovanov, D. Bernsmeier, H. Wang, B. Paul, X. Wang, S. Kuhl and P. Strasser, Structure, Activity, and Faradaic Efficiency of Nitrogen-Doped Porous Carbon Catalysts for Direct Electrochemical Hydrogen Peroxide Production, ChemSusChem, 2018, 11, 3388–3395.
24. J. Park, Y. Nabae, T. Hayakawa and M. a. Kakimoto, Highly Selective Two-Electron Oxygen Reduction Catalyzed by Mesoporous Nitrogen-Doped Carbon, ACS Catal., 2014, 4, 3749–3754.
25. L. Qu, Y. Liu, J. B. Baek and L. Dai, Nitrogen-doped graphene as efficient metal-free electrocatalyst for oxygen reduction in fuel cells, ACS Nano, 2010, 4, 1321.
26. D. Guo, R. Shibuya, C. Akiba, S. Saji, T. Kondo and J. Nakamura, Active sites of nitrogen-doped carbon materials for oxygen reduction reaction clarified using model catalysts, Science, 2016, 351, 361.
27. K. Zhao, Y. Su, X. Quan, Y. Liu, S. Chen and H. Yu, Enhanced H₂O₂ production by selective electrochemical reduction of O₂ on fluorine-doped hierarchically porous carbon, J. Catal., 2018, 357, 118–126.
28. Y. H. Lee, F. Li, K. H. Chang, C. C. Hu and T. Ohsaka, Novel synthesis of N-doped porous carbons from collagen for electrocatalytic production of H₂O₂, Appl. Catal., B, 2012, 126, 208–214.
29. Y. Sun, I. Sinev, W. Ju, A. Bergmann, S. Dresp, S. Kühl, C. Spöri, H. Schmies, H. Wang, D. Bernsmeier, B. Paul, R. Schmack, R. Kraehnert, B. R. Cuenya and P. Strasser, Efficient Electrochemical Hydrogen Peroxide Production from Molecular Oxygen on Nitrogen-Doped Mesoporous Carbon Catalysts, ACS Catal., 2018, 8, 2844–2856.
30. D. Iglesias, A. Giuliani, M. Melchionna, S. Marchesan, A. Criado, L. Nasi, M. Bevilacqua, C. Tavagnacco, F. Vizza, M. Prato and P. Fornasiero, N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O2 Reduction to H₂O₂, Chem, 2018, 4, 106–123.
31. L. Han, Y. Sun, S. Li, C. Cheng, C. E. Halbig, P. Feicht, J. L. Hübner, P. Strasser and S. Eigler, In-Plane Carbon Lattice-Defect Regulating Electrochemical Oxygen Reduction to Hydrogen Peroxide Production over Nitrogen-Doped Graphene, ACS Catal., 2019, 9, 1283–1288.
32. B. Zheng, X. L. Cai, Y. Zhou and X. H. Xia, Pure Pyridinic Nitrogen-Doped Single-Layer Graphene Catalyzes Two-Electron Transfer Process of Oxygen Reduction Reaction, ChemElectroChem, 2016, 3, 2036–2042.
33. Y. J. Sa, J. H. Kim and S. H. Joo, Active Edge-Site-Rich Carbon Nanocatalysts with Enhanced Electron Transfer for Efficient Electrochemical Hydrogen Peroxide Production, Angew. Chem., Int. Ed., 2019, 58, 1100–1105.
34. G. F. Han, F. Li, W. Zou, M. Karamad, J. P. Jeon, S. W. Kim, S. J. Kim, Y. Bu, Z. Fu, Y. Lu, S. Siahrostami and J. B. Baek, Building and identifying highly active oxygenated groups in carbon materials for oxygen reduction to H₂O₂, Nat. Commun., 2020, 11, 2209.
35. K. H. Wu, D. Wang, X. Lu, X. Zhang, Z. Xie, Y. Liu, B. J. Su, J. M. Chen, D. S. Su, W. Qi and S. Guo, Highly Selective Hydrogen Peroxide Electrosynthesis on Carbon: In Situ Interface Engineering with Surfactants, Chem, 2020, 6, 1443–1458.
36. X. Lu, D. Wang, K. H. Wu, X. Guo and W. Qi, Oxygen reduction to hydrogen peroxide on oxidized nanocarbon: Identification and quantification of active sites, J. Colloid Interface Sci., 2020, 573, 376–383.
37. S. Chen, Z. Chen, S. Siahrostami, D. Higgins, D. Nordlund, D. Sokaras, T. R. Kim, Y. Liu, X. Yan, E. Nilsson, R. Sinclair, J. K. Norskov, T. F. Jaramillo and Z. Bao, Designing Boron Nitride Islands in Carbon Materials for Efficient Electrochemical Synthesis of Hydrogen Peroxide, J. Am. Chem. Soc., 2018, 140, 7851–7859.
38. G. Kresse and J. Furthmüller, Efficiency of Ab initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set, Comput. Mater. Sci., 1996, 6, 15–50.
39. G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
40. G. Kresse and D. Joubert, From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758.
41. J. Wellendorff, K. Lundgaard, A. Møgelhøj, V. Petzold, D. Landis, J. Nørskov, T. Bligaard and K. Jacobsen, Density functionals for surface science: Exchange-correlation model development with Bayesian error estimation, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 85, 235149.
42. J. Liu, W. Li, L. Duan, X. Li, L. Ji, Z. Geng, K. Huang, L. Lu, L. Zhou, Z. Liu, W. Chen, L. Liu, S. Feng and Y. Zhang, A Graphene-like Oxygenated Carbon Nitride Material for Improved Cycle-Life Lithium/Sulfur Batteries, Nano Lett., 2015, 15, 5137–5142.
43. J. E. Hodge, Dehydrated Foods, Chemistry of Browning Reactions in Model Systems, J. Agric. Food Chem., 1953, 1, 625–651.
44. H. Shao, X. Zhang, H. Huang, K. Zhang, M. Wang, C. Zhang, Y. Yang, M. Wen and W. Zheng, Magnetron Sputtering Deposition Cu@Onion-like N-C as High-Performance Electrocatalysts for Oxygen Reduction Reaction, ACS Appl. Mater. Interfaces, 2017, 9, 41945–41954.
45. R. Zhou and S. Z. Qiao, An Fe/N co-doped graphitic carbon bulb for high-performance oxygen reduction reaction, Chem. Commun., 2015, 51, 7516–7519.
46. Z. H. Sheng, L. Shao, J. J. Chen, W. J. Bao, F. B. Wang and X. H. Xia, Catalyst-free synthesis of nitrogen-doped graphene via thermal annealing graphite oxide with melamine and its excellent electrocatalysis, ACS Nano, 2011, 5, 4350.
47. C. Zhang, W. Zhang, S. Yu, D. Wang, W. Zhang, W. Zheng, M. Wen, H. Tian, K. Huang, S. Feng and J. J. Bentzen, Cover Picture: Unlocking the Electrocatalytic Activity of Chemically Inert Amorphous Carbon-Nitrogen for Oxygen Reduction: Discerning and Refactoring Chaotic Bonds (ChemElectroChem 6/2017), ChemElectroChem, 2017, 4, 1266–1266.
48. C. Hu, Y. Zhou, R. Ma, Q. Liu and J. Wang, Reactive template synthesis of nitrogen-doped graphene-like carbon nanosheets derived from hydroxypropyl methylcellulose and dicyandiamide as efficient oxygen reduction electrocatalysts, J. Power Sources, 2017, 345, 120–130.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0qi01089h

‡ These two authors contributed equally.

---