Mechanism of Oxygen Reduction Reaction on Monolayer WTe2 with and without S Dopant at Low Coverage

a Ho Chi Minh City University of Technology (HCMUT), 268 Ly Thuong Kiet Street, District 10, Ho Chi Minh City, Vietnam b Vietnam National University Ho Chi Minh City, Linh Trung Ward, Thu Duc District, Ho Chi Minh City, Vietnam c Institute of Physical Chemistry “Ilie Murgulescu” of the Romanian Academy, Splaiul Independentei 202, Sector 6, 060021 Bucharest, Romania † Corresponding author: dnson@hcmut.edu.vn


I. INTRODUCTION
The development of catalysts for the cathode is a challenge to improve the performance of proton exchange membrane fuel cells and rechargeable metal-air batteries. The oxygen reduction reaction (ORR) on the cathode involves many intermediate steps and has a slow rate. Pure Pt, Pt-based alloys, and carbon-supported Pt nanoparticles were the most efficient catalytic materials [1]. However, they are too expensive to be widely commercialized. Pt-free alloys of less expensive metals such as Pd, Fe, and Co are the alternatives to significantly reduce cost and simultaneously improve the stability of catalysts [2−4]. Their ORR activity is comparable to that of the carbon-supported Pt materials [2]. Single-atom catalysts fabricated by dispersing metal atoms onto macromolecules and carbon-based materials like porphyrin and graphene were also investigated for the ORR [5]. The single-atom catalysts possess the maximum atom utilization efficiency. Besides, transition metal oxides such as cobalt oxides and copper oxides also exhibited a high per-formance [6]. Recently, transition metal dichalcogenides (TMDs) have attracted much attention for the ORR due to their low cost, high abundance, and high tolerance to organic molecules [7]. Perfect, defective, and doped structures of TMDs could significantly enhance the ORR activity [8−10]. Besides, the ORR mechanisms on WTe2, which is a new material of the TMD family, with Te vacancies have been recently investigated [11]. However, there is no research available to elucidate the ORR mechanism and the activity on perfect and doped monolayer WTe 2 . Especially, S is one of the most popular chalcogenide elements, which can form natural compounds like MoS 2 and WS 2 . Hence, S doping in WTe 2 is expected to be thermodynamically favorable. Indeed, we found in this work that the formation energy of S substitution for the Te atom in WTe 2 is significantly negative, which implies that S doping is an exothermic process, and substitution proceeds simultaneously. Other examples of S doping in TMDs such as WSe 2 and MoTe 2 were also found in literature [12−14]. Doping with electron-donating and electron-withdrawing elements is an effective method to modify the electronic structure of substrates and, therefore, it influences the adsorption strength of ORR intermediates and the efficiency of catalysts. However, due to the complicated nature of the ORR involving many intermediates, understanding the electron-withdrawing S dopant on the overall performance of WTe 2 remains unclarified. For these reasons, the present research will elucidate the ORR mechanisms on the perfect monolayer WTe 2 and S-doped monolayer WTe 2 (abbreviated as WTe 2 S) using density functional theory (DFT) calculations and a thermodynamics model.

II. COMPUTATIONAL METHOD
We performed the DFT-based calculations using the Vienna ab initio simulation package [15] with the Perdew-Burke-Ernzerhof generalized gradient approximation (PBE-GGA) functional for exchange-correlation [16,17] and the projector augmented wave (PAW) method for the core-valence interaction [18−20]. The cut-off energy for the plane-wave basis set expansion was 400 eV. The k-point mesh grid of 3 × 3 × 1 was sampling by the Monkhorst-Pack technique. We took into account the dipole corrections for excluding the interaction of images in periodic supercells.
We modeled the supercell of monolayer WTe2 by a slab of 5 × 5 unit cell and a vacuum of about 13 Å, as shown in Figure 1. S doping at a low coverage was created by replacing one Te atom in perfect monolayer WTe 2 to create the WTe 2 S substrate. The substrates have the dimensions of a = b = 17.61 Å [11]. All atomic positions in the substrates and the ORR intermediates were fully optimized. The formation energy of −6.64 eV for S doping implies that the process is exothermic and thermodynamically favorable. The formation energy f was defined by Here, Sub S , Sub v , and S i are the total energies of the WTe 2 S substrate, the WTe 2 substrate with a Te vacancy, and the isolated S atom, respectively. The formation energy was calculated by Eq. (1), because, in reality, monolayer WTe 2 is not formed from powder but bulk crystal by the mechanical exfoliation technique [21]. The chalcogenide vacancies often occur as an accompanying effect during the fabrication process [22]. Besides, substitutional doping with dopant atoms covalently bonded at vacancy sites is a promising technique [23]. Therefore, the doping process can be interpreted as generating vacancies and updating dopant atoms into the vacancies [24].
The adsorption energy a for each ORR intermediate was estimated through the total energy of each component as The Gibbs free energy was studied to understand the progress of the reaction steps and the thermodynamic stability of reaction intermediates [25,26]. The Gibbs free energy was calculated by the formula; Here, we ignored the contribution of the acid medium by taking pH = 0. The reaction energy, the zero-point energy correction, and the entropy change are denoted as Δ , Δ , Δ , respectively. is the electrode potential relative to the standard hydrogen electrode.

III. RESULTS AND DISCUSSION
Generally, the ORR in an acidic medium, O 2 + 4(H + + e − ) → 2H 2 O, can proceed through dissociative and associative mechanisms [1,27] as described in Scheme 1. Both pathways begin with the adsorption of oxygen molecules onto the catalytic surface. The associative mechanism (path 2 → 4) remains the molecular state of the oxygen molecule during the electron and proton transfer, while the dissociative mechanism (path 1 → 6 → 8) involves the dissociation of the oxygen molecule to form two oxygen atoms before the hydrogenation steps. Partially, the associative mechanism can turn to the dissociative one at steps 3 and 5. The second hydrogenation step on HOO * and O * + HO * will generate O * + H 2 O, while the third hydrogenation step on HOOH * , 2HO * , and O * + H 2 O will create HO * + H 2 O. The last hydrogenation step 12 will form the by-product of the ORR.
We have searched the adsorption site and configuration for the possible intermediates of the ORR on WTe 2 and WTe 2 S by designing their geometric structures and performing the DFT optimization. Scheme 1 listed the adsorption energy for the most stable configuration of intermediates. The positive adsorption energy of the ORR intermediates in the pathways 2 → 4 and 1 → 6 → 8 implies that adsorption of the ORR intermediates is thermodynamically unfavorable, which is in good agreement with the work of Liu et al. [28]. Even with HOO * on WTe 2 S, the adsorption energy is negative, although the vibrational (zero-point) energy is much higher than its adsorption energy. This result indicates that HOO * is unstable on the WTe 2 S surface due to a large fluctuation. Notably, only two intermediates, i.e., O * + H 2 O and HO * + H 2 O adsorb stably with significant negative adsorption energies on WTe 2 and WTe 2 S. The analysis of the adsorption energy suggests that the ORR should proceed through the chain: i.e., pathway 7 → 10 → 12. Furthermore, the adsorption energy of O * and HO * in the absence of water (the gas phase intermediates) was also presented in Scheme 1 for comparison with O * + H 2 O and HO * + H 2 O. It exhibits that these ORR intermediates on the same substrate in the presence of water adsorb stabler than those without water. Also, the adsorption energy of O * and HO * with and without H 2 O is more negative on WTe 2 S than on WTe 2 . The result implies that the S dopant increases the adsorption strength of O * and HO * . Figure 1 shows  Figure 2), the O * p x orbital is more dominant than the O * p z orbital around −2.0 eV, which mainly interacts with the Te p x and p z states of the WTe 2 and WTe 2 S substrates due to the overlapping with the DOS of the Te atoms. However, the O * p z orbital is more significant than the O * p x orbital at lower energies. Contrastingly, for WTe 2 S, there are the peak resonances of the O * p z orbital with the S p z orbital and the O * p x orbital with the S p x orbital around −6.5 eV. For HO * + H 2 O (the right column of Figure 2), the DOS of O * disperses in a broad range of energy, and they are similar for both the WTe 2 and WTe 2 S substrates. The S p orbitals do not show a resonance with the O * p orbitals because the most favorable configuration of the HO * + H 2 O intermediate locates far away from the S atom. Even though, we have to note that the S atom site is the second most favorable position of HO * + H 2 O on the WTe 2 S substrate.  The interaction of the DOS between the substrates and the intermediates leads to charge rearrangement. We calculated the point charge of the atoms in the substrate-intermediate systems by the Bader partition technique to understand the charge exchange tendency.
Equation (4) involves transferring of two protons and two electrons at once. We can imagine that O 2 (gas phase) combines with two protons and two electrons during approaching the surface of WTe 2 and WTe 2 S to form the stable intermediates, while the common dissociative and associative ORR mechanisms always start with adsorption of the oxygen molecule. We obtained the Gibbs free energy for the intermediate steps, i.e., Eqs. (4) to (6), from the calculated zero-point energy, total energy, and the thermodynamics model [25,26] at the electrode potential of 1.23 V, pH = 0, the pressure of 1 bar, and the temperature of 300 K. The Gibbs free energy for the proposed intermediate steps were calculated by the following formulae: The corresponding values of the Gibbs free energy were listed in Table 3.
The Gibbs free energy diagram in Figure 3(a) shows that, in the presence of H 2 O, both steps of the O * and HO * formation on WTe 2 are going uphill. In Figure 3(b), the step of the O * formation on WTe 2 S is downhill before uphill again at the step of the HO * formation. The reason for going downhill is due to the stronger adsorption of the O * intermediate upon the S doping. In the presence of H 2 O, we find on both substrates that the first step transfers two protons and two electrons at the same time. This process is different compared to the first hydrogenation step with only one hydrogen atom at a time for various catalytic materials in the literature [1,11,27]. The second step is the rate-limiting step of the reaction, which is the transfer of the third proton and electron. This result is also different com-pared to the literature in that the first proton and electron transfer is the rate-limiting step of the reaction [1,11,25,27]. The activation barrier for the whole process in the forward direction from left to right of the Gibbs free energy diagram was calculated by the energy difference between two steps of the HO * formation and O 2 + 4H for WTe 2 and between two steps of the HO * and O * formation for WTe 2 S. The barriers of 1.22 and 1.63 eV were obtained for the ORR in the presence of H 2 O on WTe 2 and WTe 2 S, respectively. The result implies that S doping decreases the performance of WTe 2 toward the ORR, which also agrees with literature for other electron-withdrawing elements or p-type doping [29−31]. Furthermore, the presence of H 2 O decreases the activation barrier by 0.1 eV compared to the absence of H 2 O on both WTe 2 and WTe 2 S substrates. The ORR mechanism involves four electrons in the current work.

IV. CONCLUSION
In summary, the density functional theory calculations and the thermodynamics model revealed that the ORR mechanism on perfect monolayer WTe 2 and S-doped monolayer WTe 2 begins with a simultaneous transfer of two protons and two electrons to dissociate and stabilize O 2 in the form of O * + H 2 O. The second step, i.e., the transfer of the third proton and electron, is the rate-limiting step with the thermodynamics barrier of about 1.22 and 1.63 eV for WTe 2 and WTe 2 S, respectively. The final step forms water, the by-product of the ORR. We have to emphasize that the mechanism does not start with the adsorption of the oxygen molecules. Also, the rate-determining step is not the first hydrogenation step as the common dissociative and associative mechanisms. Interestingly, the ORR proceeds via a four-electron process on the transition-metal dichalcogenides. The S dopant harms the ORR performance of WTe2 because it enhances the adsorption strength of O * and HO * intermediates and leads to the increase of the ORR thermodynamics barrier.