Heteroatom (Nitrogen/Sulfur)-Doped Graphene as an Efficient Electrocatalyst for Oxygen Reduction and Evolution Reactions

Abstract

Carbon nanomaterials are potential materials with their intrinsic structure and property in energy conversion and storage. As the electrocatalysts, graphene is more remarkable in electrochemical reactions. Additionally, heteroatoms doping with metal-free materials can obtain unique structure and demonstrate excellent electrocatalytic performance. In this work, we proposed a facile method to prepare bifunctional electrocatalyst which was constructed by nitrogen, sulfur doped graphene (NSG), which demonstrate superior properties with high activity and excellent durability compared with Pt/C and IrO₂ for oxygen reduction (OR) and oxygen evolution (OE) reactions. Accordingly, these phenomena are closely related to the synergistic effect of doping with nitrogen and sulfur by rationally regulating the polarity of carbon in graphene. The current work expands the method towards carbon materials with heteroatom dopants for commercialization in energy-related reactions.

1. Introduction

Graphene, a two-dimensional atom-thick conjugated structure, has drawn particular attention for its good conductivity, mechanical property, electrochemical stability and huge specific surface area as well as its wide potential applications in catalysts [1,2]. However, graphene oxide has serious structure disorder and contains oxygen-containing groups, which weaken its electrode transportation and conductivity [3,4,5,6]. Heteroatoms like N and S are adjacent elements to carbon. The doping of such elements to graphene may change its structure and improve the electrochemical property of graphene [4,7,8,9,10] due to its excellent stability, durability and controllability of nanoparticles [11,12].

On the other hand, fuel cells and metal-air batteries, as promising high-performance electrochemical energy-related devices, are suffering from bottlenecks because of its sluggish kinetics in oxygen reduction reaction (ORR, O₂ + H₂O + 4e⁻ → 4OH⁻) [13,14,15,16,17,18]. The platinum (Pt), with the advantage of low overpotential and high current density, has been served as the most promising ORR electrocatalyst regardless of the scarcity, high cost, and poor durability which obstruct its application in the energy field [7,19,20,21,22]. To address this problem, seeking alternative ORR and OER electrocatalysts with low overpotentials and cost is urgent for sustainable energy solutions. Heteroatom doped carbon-based materials as promising metal-free electrocatalysts have encouraged intensive research, while it is found that, the hybrids of doped graphene show multi-functions, and have a wide application prospect in energy storage and conversion and environmental detection. They can be used as catalysts for ORR and OER, but the relationship between precursor-doping pattern-ORR activity remains unclear [16,23,24,25,26]. Additionally, the performance of ORR and OER of some metal-free electrocatalysts can be optimized. Therefore, developing low cost and high electrocatalytic efficiency metal-free materials has become a focus of study.

Herein, we developed an efficient method to prepare heteroatom nitrogen-, sulfur-doped graphene as bifunctional materials for ORR/OER by hydrothermal synthesis of graphene oxide and potassium thiocyanate (KSCN), whether the guest gas (NH₃) exists or not.

2. Results and Discussion

2.1. Characterization of Electrocatalysts

The structure and morphology of materials were investigated through scanning electron microscopy (SEM) and transmission electron microscopy (TEM) [27]. As observed in Figure 1 and Figure S1, the elements of C, N, O and S were distributed uniformly in NSG, with the fabrication of stacking graphene sheets. This specific morphology could catalyze the reaction process derived from the exposing of more active sites. The SEM analyses (Figure 1b) confirm that the heteroatoms were doped successfully. The microstructure of nitrogen doped graphene (NG) and sulfur doped graphene (SG) were shown in Figures S2 and S3 where the twisted graphene layers were clearly observed.

X-ray photoelectron spectroscopy (XPS) tests were the valid method to analyze the ingredients and chemical valence of materials [28]. As shown in Figure 2a, Compositions of C, N, O and S were evaluated through XPS survey spectrum whose atomic percentages respectively are 91.90, 4.18, 3.90 and 0.02 at %, demonstrating the successful doping with nitrogen and sulfur. As observed from Figure 2b, sp² hybridized graphitic carbon, C–S and C=O and C=N were corresponding to 284.5 eV, 285.5 eV and 287.0 eV, respectively. Pyridinic N (397.8 eV), pyrrolic N (398.9 eV) and graphitic N (400.3 eV) all existed (Figure 2c). As we know, graphitic N is viewed as active sites in the oxygen reduction reaction [23,28]. Moving to Figure 2d, high resolution S 2p, the peaks of 163.9 and 165.0 eV can ascribe to the S 2p spin-orbit doublet (S 2p1/2 and S 2p3/2) [29], affected by C–S–C, deeply demonstrating the successful addition with sulfur [30]. Compared with the high resolution N 1s of NG (Figure S4), the existence of C–S–C and synergistic effects of N and S doping may be the essential factors of performance improvement for NSG. All of this demonstrated the synergy of N and S catalyzed the activity of NSG.

Figure 3a shows distinct Raman spectra at 1345 and 1600 cm⁻¹ which is ascribed to the D and G band. A recent report indicated that the G band represents the graphitization degree of materials, structural defects are reflected by the D band [31]. To our best knowledge, the I_D/I_G ratio represents structural disorder in graphitic materials [32]. As observed from Figure 3a, the value of NSG (1.09) was lower than that of SG (1.15) and NG (1.29), indicating the increasing graphitic degree of NSG which may be derived from the “self-repairing” of intermediate products under experimental conditions with the repairing of partial sp²-bonded C atoms. The components of the electrocatalysts were also analyzed. The graphitic C structure existed in NSG, SG and NG at 24° with the observation of X-ray diffraction (XRD) spectrum in Figure 3b [33]. In addition, for SG and NSG, the inconspicuous peak at 667 cm⁻¹ belonged to the vibration of C–S bond which was making a clear indication that S element was doped with graphene in Figure S5 [5,8].

2.2. Electrocatalytic Properties of Catalysts

In O₂-saturated 0.1 M KOH electrolyte, a usual three-electrode electrochemical device consisted of Ag/AgCl, Pt wire and rotating ring-disk electrode (4 mm diameter) which was orderly used as a reference electrode, counter electrode and working electrode with the continuous flowing of O₂ operated to test the ORR performances of different catalysts which was revealed by linear sweep voltammetry (LSV) [34]. According to the conversion equation (V_RHE = V_Ag/AgCl + 0.059pH + 0.197), the potentials tested through Ag/AgCl were converted into reversible hydrogen electrode (RHE).

Compared with commercial Pt/C (20 wt %), the ORR performance of NSG, SG and NG was investigated via LSV in 0.1 M KOH electrolyte (Figure 4a). Obviously, NSG shows higher onset potential (0.95 V vs. RHE) compared with SG and NG (0.90 and 0.89 V). For half-wave potential, as shown in Figure 4b, the value of NSG (0.84 V) is prior to that of Pt/C, SG and NG (0.82, 0.8, 0.77 V), indicating its preponderant electrocatalytic performance for ORR. In addition, NSG displays large limited current density, demonstrating a more efficient mass transfer among such electrocatalysts [5].

In order to quantitatively analyze the mechanism of synthesized catalysts for ORR, further LSVs were recorded at 400, 800, 1200 and 1600 rpm (Figure 4c, Figure S6a,c). Koutecky-Levich (K-L) plots were illustrated in Figure 4c and Figure S5b,d. To the best of our knowledge, a function relationship exists between inverse current density and square root of the rotation rate at a different range of potentials. According to the calculating formula of K-L equation, 3.86, 3.72 and 3.73, the electron transfer numbers (n) of NSG, SG and NG, indicate the ORRs mechanism is a four-electron process.

To further explore the performance of synthesized electrocatalysts, the electron transfer number (n) and percentage (%) of peroxide species were tested through a rotating ring disk electrode (RRDE). As shown in Figure 4d, the four-electron process of the as-prepared NSG for ORR is consistent with the RDE results. Additionally the average percentage (%) of peroxide yield of NSG is the lowest compared with SG and NG, demonstrating the multiple doping.

Adding 8.5 mL methanol into 70 mL 0.1 M KOH electrolyte with continuous O₂, the tolerance of NSG to methanol was investigated. Compared with the sharp decline of current density of Pt/C (current loss of ~60%), NSG exhibits a little fluctuation with a retention ratio of ~93%, demonstrating the decreased activity of Pt/C affected by methanol, and NSG possesses a well tolerance to methanol (Figure 4e).

As a state-of-the-art catalyst for ORR, chronoamperometric measurement was used to assess the durability at a constant cathodic voltage of 0.57 V. As observed from Figure 4f, NSG exhibits an outstanding ORR stability with a weak attenuation over 30,000 s, maintaining 96% of the initial current. Compared with approximately 15% loss of initial current, NSG was better than Pt/C for ORR in alkaline electrolyte. OER activity of various as-prepared samples was also tested in the same condition (Figure 5a). To our best knowledge, the quantitative value of E_j = 10 (potentials to deliver 10 mA cm⁻² current density) represents 10% efficient solar water-splitting cell which is used to make comparison with various catalysts [35]. For NSG, E_j = 10 (1.62 V) is lower than that of 1.65 V (SG), 1.66 V (NG) and some of state-of-the-art carbon-based catalysts, just like N, S-CN [5] and CN nanocables [36] and so on. The Tafel slope of NSG (105 mV dec⁻¹) is lower than that of SG (121 mV dec⁻¹) and NG (124 mV dec⁻¹), indicating its favorable reaction kinetics (Figure 5b). In addition, the representation of charge transfer kinetics, electrochemical impedance spectroscopy (EIS) of NSG, NG and SG, were also investigated. The Nyquist plots in Figure S7 demonstrates a lower charge transfer resistance towards the OER process of NSG than that of NG and SG, which is consistent with the OER performance, revealing faster Faradaic process in OER kinetics of NSG.

Considering the structure-property relationship, the better catalytic performance towards ORR and OER may be a result of the stable covalent C–N which could form high positive charge density on neighboring carbon atoms, and the mismatch of outermost orbitals between C and S [37] and large surface area [38] which could facilitate the charge transfer, further endowing more accessible catalytic surfaces.

3. Materials and Methods

3.1. Preparation of Electrocatalysts

At first, GO was prepared according to the procedure used by Hummer [39]. To begin this process, GO and KSCN (20.0 mg/20.0 mg) were intensively mixed in 30.0 mL deionized water. Then, a homogeneous solution was fabricated through ultrasonication and kept in Teflon-lined stainless-steal autoclave (150 °C) for 15 h. The as-prepared solution was treated, at freezing, with a vacuum dryer overnight to form powder. After dried, the powder was placed in a tube furnace which was programmed at 800 °C for 2 h with 5 °C/min of rising rate, keeping Ar flowing of 150 mL/min. In the end, with Ar and NH₃ whose flow rate is 500 mL/min, the intermediate product was annealed at 800 °C for 1 h (Scheme 1). Up to now, the target materials were obtained and defined as nitrogen-, sulfur- co-doped graphene (NSG). For the sake of contrast, only S-doped or N-doped graphene catalysts were synthesized under the same operating conditions which was defined as SG or NG, respectively. In other words, the only source of N (NH₃) or S (KSCN) was used as introduction.

3.2. Characterization of Electrocatalysts

A S-4800 SEM instrument (Hitachi High-Technology Co., Ltd., Tokyo, Japan) was used to test the surface characterization of different electrocatalysts. TEM (JEOL JEM-2100, Tokyo, Japan) was operated at 200 kV. With radiation of Cu-Kα, X-ray diffraction (XRD, D/Max2000, Rigaku, Japan) was investigated. Fourier transformed infrared (FTIR) spectra were observed through a TENSOR 27 FT-IR spectrometer (Scotia, NY, USA) in the range of 4000–500 cm⁻¹, after the samples were dried. Escalab 250 xi (Thermo Scientific, Loughborough, UK) was used to record X-ray photoelectron spectroscopy (XPS), providing a base pressure of 5 × 10⁻⁹ Torr radiated from monochromatic Al Kα. Raman spectra were investigated by using a Renishaw Raman spectroscope (Renishaw plc., Gloucestershire, UK).

4. Conclusions

To conclude this work, the four-electron pathway for ORR on N, S-co-doped graphene is revealed and synergy effects between dopants are discussed. The synergy effect is ascribed to the increasing spin density with the dopant of S and graphitic N. The as-prepared catalyst exhibits excellent performance for ORR and OER which is originated from the unique structure of NSG, fortunately, the unique structure is to the benefit of mass transfer. Overall, this work provides a carbon-based bifunctional electrocatalyst of dual doping in ORR and OER on the promising widespread application in energy-related devices.