The effect of heteroatom doping on the active metal site of CoS₂ for hydrogen evolution reaction

Abstract

The exploration of cost-effective hydrogen evolution reaction (HER) electrocatalysts through water splitting is important for developing clean energy technology and devices. The application of CoS₂ in HER has been drawing more and more attention due to its low cost and relatively satisfactory HER catalytic performance. And CoS₂ was found to exhibit excellent HER catalytic performance after appropriate doping according to other experimental investigations. However, the theoretical simulation and the intrinsic catalytic mechanism of CoS₂ remains insufficiently investigated. Therefore, in this study, density functional theory is used to investigate the HER catalytic activity of CoS₂ doped with a heteroatom. The results show that Pt-, N- and O-doped CoS₂ demonstrates smaller Gibbs free energies close to that of Pt, compared with the original CoS₂ and CoS₂ doped with other atoms. Furthermore, HER catalytic performance of CoS₂ can be improved by tuning d-band centers of H adsorption sites. This study provides an effective method to achieve modified CoS₂ for high-performance HER and to investigate other transition metal sulfides as HER electrode.

---

1 Introduction

The high combustion heat value, clean combustion products, and abundant resources make hydrogen (H₂) a distinguished energy storage material. Producing hydrogen from the electrochemical decomposition of water (H₂O), the commonly believed future energy carrier,^1,2 through hydrogen evolution reaction (HER) has been proved to be a clean and renewable method. However, the hydrogen production catalyst precious metal platinum (Pt), the most efficient electrocatalyst for HER owing to its high stability, low overpotential (about 0.08 V), as well low Tafel slope,³ has not been applied on large scales due to its high cost and scarcity. Hence, it is quite necessary to find a new high-performance HER catalyst.

In recent decades, transition metal sulfides (TMSs), such as MoS₂,^4–10 VS₂,^11–15 CoS₂,^2,16–20 NiS₂ (ref. 2, 7, 15, 21 and 22) have been attracting attention from researchers and industry practitioners. Though the effects of the aforesaid compounds as hydrogen production catalysts had been considered in some previous studies, the more recent ones show that some of them such as HER electrode material can exhibit outstanding electrocatalytic performance that is close to that of precious metals (such as Pt). To further optimize the performance of the TMSs, some other studies suggest the incorporating of a third heteroatom in a binary compound, which can improve the electrocatalytic activity of the catalyst.²³

As a representative of pyrite-type transition metal sulfides, CoS₂ is a low-cost catalytic material with excellent HER performance. Unfortunately, when doping CoS₂ with a third heteroatom, no good descriptor can be found to describe the relationship between the structure and activity of the doped CoS₂, and the inherent mechanism remains insufficiently explored.

It is noteworthy that Hoffmann et al. found that the center of the d-band formed by the TM turned out to be a good descriptor to predict the adsorption and reactivity of transition metal catalysts.²⁴ Furthermore, an analogous model of d-band based on surface resonance states has also been developed in the transition metal carbides (TMCs),²⁵ metal nitrides (MNs),²⁶ and layered TMSs.²⁷ This can significantly contribute to the analysis of structure–activity relationships in TMCs/TMSs, which can further help improve the design and high-throughput screening of catalysts. Many studies have also found that doping can change the energy band structure near the Fermi energy, so as to change the properties of materials.^10,28,29 Therefore, we take interest in investigating the possibility to doped CoS₂ for high-performance HER by tuning the d-band centers.

In this work, we study the effect of a heteroatom doping on HER electrochemical properties for CoS₂ electrode material based on density functional theory and explore the structure–activity relationships. Herein, the models of the pristine CoS₂ and doped CoS₂ were built, the stability, the catalytic activity, and the d-band centers of the active metal sites of CoS₂ doped with a heteroatom (Mn, Fe, Ni, Cu, Mo, Pt, Ru, C, N, O, P) are studied based on density functional theory (DFT). Our first-principles calculations exhibit that the structures of all doped CoS₂ except C-doped CoS₂ are stable, the catalytic behavior of CoS₂ can be effectively enhanced after Pt, N, and O doping, and the HER catalysis of CoS₂ can be changed by tailoring the d-band centers of active metal sites. Our studies are of great benefit to uncovering the intrinsic modulation mechanism of CoS₂ for HER catalysis.

2 Computational methods

All density functional theory (DFT) simulations were employed using the Generalized Gradient Approximation-Perdew Burke Ernzerhof (GGA-PBE) was used for the exchange-correlation term in the Spanish Initiative for Electronic Simulations with Thousands of Atoms (SIESTA).^30,31 The geometric optimization was employed using a conjugate gradient method until the maximum force was less than 0.02 eV Å⁻¹. An energy cut-off for all simulations was set to be 150 Ry.

2.1 Bulk CoS₂

The k-point mesh is set as Γ-centered 7 × 7 × 7 to relax the CoS₂ unit cell with a space group of Pa3̄. The calculated CoS₂ unit cell parameter is 5.544 Å, which is highly consistent with the experimental value (5.538 Å).^23,32 The optimized bulk structure is shown in Fig. 1, where the big and small balls represent Co and S atoms, respectively.

Fig. 1 The optimized structure of the bulk CoS₂.

2.2 CoS₂(001) surface

The (001) surface of CoS₂ is selected as the optimization model owing to its stability,³² which is built as a 2 × 2 repeating surface unit cells consisting of four S–Co–S atomic layers and 8 CoS₂ units. Cation-doped CoS₂ is acquired by replacing one Co atom with one metal atom, and anion-doped CoS₂ is acquired by replacing one S atom with one nonmetal atom. A vacuum layer with a thickness of 15 Å is introduced to reduce interaction between periodic images.³³ The CoS₂(001) crystalline structure is represented in Fig. 2.

Fig. 2 The crystallographic structure of CoS₂(001) surface.

2.3 Binding energy

The binding energies (E_b) were first calculated to measure the stability of CoS₂ after doping, E_b is the difference between the cohesive energy (E_coh) of cubic CoS₂ and the embedding energy of the dopants (E_emb) on the CoS₂ surface.³⁴

E_coh = E_bulk − E_iso (1)

E_emb = E_dopant+subs − E_subs − E_iso (2)

E_b = E_coh − E_emb (3)

where E_bulk is the total energy of an atom in the bulk, E_iso is the total energy of an isolated atom, E_dopants+subs is the total energy of doped CoS₂, and E_subs is the total energy of CoS₂ with one vacancy.

2.4 Gibbs free energy of H adsorption

Nørskov et al.³ showed that with the absolute Gibbs free energy of hydrogen adsorption (ΔG_H*) approximating zero, the binding between HER intermediates and the electrode surface was neither too strong nor too weak, demonstrating a high catalytic performance of HER. Hence, the (ΔG_H*) as a significant factor of measuring the HER catalytic activity is quite essential to is calculated using reliable DFT simulations.³⁵ In our work, the HER catalytic properties in acidic media were investigated. While considering the HER reaction process in acid, the Volmer reaction³ was selected under our calculations, because most reactions occur in protic solution or involve proton as the reactant:³⁶ H⁺(sol) + e⁻ → H*, where H* refers to H adsorbed on the active site, and many studies analyze the electrochemical catalytic properties of HER based on the Volmer reaction.³ The Gibbs free energy of hydrogen adsorption ((ΔG_H*)) as a significant factor of measuring the HER catalytic activity is calculated based on the Volmer reaction using the DFT method. Its definition is as follows³⁵

ΔG_H* = ΔE_H + ΔE_ZPE − TΔS (4)

where ΔE_ZPE is the difference in zero-point energy between the adsorption state of H and gas phase, while ΔS is the difference in entropy between the adsorption state of H and gas phase.³ ΔE_H is the hydrogen adsorption energy. Given that the vibrational entropy of H* in the adsorbed state is small, the entropy of adsorption of −1/2H₂ is simplified as ΔS_H ≈ −1/2S_H2⁰, where S_H2⁰ is the entropy of H₂ in the gas phase at standard conditions. Therefore, the overall corrections of ΔG_H* are³

ΔG_H* = ΔE_H + 0.24 eV (5)

The hydrogen adsorption energy (ΔE_H) on the electrode surfaces is defined as

where E_subs+H* is the total energy of doped CoS₂(001) with H* adsorption, E_subs is the total energy of doped CoS₂(001) without H* adsorption, and E_H2 is the total energy for one hydrogen molecule in the gas phase.

2.5 d-band centers

To further explore the possibility of enhancing electrochemical performance of CoS₂ via doping, we calculated the d-band center of the metal at active site.

The d-band center ε_d is defined as

where ρ are the densities of states and ε are the energies.

3 Results and discussion

Some materials demonstrate good computational performance but are difficult to synthesize due to their aggregation- or clustering-related instability, which limits their application in experiments or in practice.³⁷ So, it is important to first investigate the stability of CoS₂ in order to explore the possibility of its synthesis. We use the binding energy (E_b) descriptor to measure the stability of doped CoS₂, if, E_b < 0, the atoms shows more tendency of being incorporated into the CoS₂(001) surface rather than clustering or aggregation, and then we regard that the material satisfies the stability criterion.³⁴ The calculated binding energies are listed in Table 1. From Table 1 we can see, E_b of all cation and anion doped CoS₂ except for C-doped CoS₂ are less than zero, which demonstrates that it is energetically possible for one cation dopant to replace one Co atom in the CoS₂(001) surface and for one anion dopant to replace one S atom in the CoS₂(001) surface. However, the binding energy is 1.96 eV when one S atom in the CoS₂(001) surface is substituted by one C atom, which is much higher than zero, indicating that when one C is incorporated into the CoS₂(001), it tends to take the form of cluster or aggregation, which causes an unstable C-doped CoS₂(001) structure. Therefore, C will not be considered as a heteroatom to form a HER catalyst.

Table 1 Binding energies of doped CoS₂ (E_b, eV)

Cation	Model	Mn	Fe	Ni	Cu	Mo	Pt	Ru
	Eb/eV	−2.89	−2.02	−1.67	−0.96	−1.78	−1.41	−1.97

---

Anion	Model	C	N	O	P
	Eb/eV	1.96	−1.67	−2.27	−1.52

---

After identifying the stability for a cation/anion heteroatom incorporated into the CoS₂(001) surface, the state of the adsorbed H was discussed according to H adsorption energy. The negative value means that H is easy to bind to the CoS₂(001).^38–40 The value of adsorption energy for 11 doped-CoS₂ are from −1.43 eV to −2.46 eV, which indicates that it is feasible to H bind to the surface of CoS₂(001). Moreover, the catalytic performance for HER of the CoS₂(001) with and without cation/anion heteroatom doping was investigated. To determine whether CoS₂ is an outstanding HER catalyst, the Gibbs free energies (ΔG_H*) for hydrogen adsorption at active metal sites were calculated based on the DFT method. Fig. 3(a) is the relaxed crystallographic structure of H adsorption at the Co site for the pure CoS₂(001), in which the green ball refers to H atom. The Co–H bond length is 1.50 Å. The calculated |ΔG_H*| of the pure CoS₂(001) is 0.25 eV, as shown in Fig. 3(b), which is larger than that of the precious metal Pt (0.08 eV).³ In order to further probe the effect of heteroatom atoms on electrochemical HER catalytic performance for CoS₂(001) as electrode material, the ΔG_H* of cation- (Mn, Fe, Ni, Cu, Mo, Pt and Ru) and anion- (N, O and P) doped CoS₂ are calculated using the DFT method. Fig. 4(a) and (c) show the calculated ΔG_H* of H adsorption at the active metal sites for cation- and anion-doped CoS₂(001), respectively. In Fig. 4(a) and (c), ΔG_H* of Ni-doped CoS₂(001) is −0.12 eV, that of Pt-doped CoS₂(001) is 0.03 eV, that of N-doped CoS₂(001) is 0.07 eV, and that of O-doped CoS₂(001) is −0.01 eV. Compared with that (0.25 eV) of the pure CoS₂ electrode, the |ΔG_H*| of Ni-, Pt-, N-, and O-doped CoS₂(001) electrode are much smaller when H is adsorbed at metal sites. It is noteworthy that |ΔG_H*| of Pt-, N-, and O-doped CoS₂(001) is smaller than that of Pt (0.08 eV), revealing their potential excellent electrochemical HER catalytic performance. And |ΔG_H*| of O-doped CoS₂(001) is closest to zero, which suggests that O-doped CoS₂ show the best HER catalytic activity among these doped CoS₂ material. Fig. 4(b) and (d) show the optimized structures of the Pt-doped CoS₂(001) with H adsorption at Pt site, Pt–H bond length is 1.66 Å, and that of the O-doped CoS₂(001) with H adsorption at Co site, Co–H bond length is 1.51 Å, respectively. To further explore why the electrochemical HER performance of CoS₂ at active metal sites via doping can be enhanced, the d-band centers of the active metal sites were calculated using the DFT method. We plotted in Fig. 5 the density of states for active metal sites used to H adsorption and marked the d-band center at active metal site with a dashed line. The d-band center of Co active site for the pristine CoS₂ is −2.06 eV and it is closest to the Fermi level. The d-band center of Cu active site for Cu-doped CoS₂ is −6.01 eV and it is farthest to the Fermi level. Meanwhile, ΔG_H* for Cu-doped CoS₂ is 0.71 eV. To further probe the relationship between the d-band center of active metal site used to H adsorption and the Gibbs free energy of H adsorption on the active metal sites, a more detailed analysis diagram is plotted in Fig. 6. We can see from Fig. 6 that ΔG_H* increases as the d-band centers of metal active sites decrease, which clearly illustrates that the binding between H adsorption atom and metal atom at the active site is weakened as the d band center is far away from the Fermi level. Our results just described are in line with investigations in the ΔG_H* and d-band centers based on other materials.²⁶ This study reveals an intrinsic mechanism that d-band center of active metal site moves downwards after doping, leading to the weakening of hydrogen adsorption, which facilitates the desorption of H from the surface of CoS₂ as HER electrode material.

Fig. 3 (a) Represents the H adsorption structure of the pristine CoS₂, where H adsorbs at Co site. The green ball refers to H atom. (b) is ΔG_H* diagram of the pristine CoS₂.

Fig. 4 (a) and (b) are ΔG_H* diagram and the structure of Pt-doped CoS₂(001) with H adsorption at Pt site, respectively. (c) and (d) are ΔG_H* diagram and the structure of O-doped CoS₂(001) with H adsorption at Co site, respectively. The blue, dark red, dark green, red balls and light green ball are Co, S, Pt, O and H atoms, respectively. Pt–H bond length is 1.66 Å and Co–O bond length is 2.21 Å.

Fig. 5 The density of states at active metal sites. The dashed line marked the d-band center. The Fermi energy is set to zero.

Fig. 6 The linear relationship between ΔG_H* and d-band centers of H adsorption sites.

Based on our DFT calculations, we can find that Ni-, Pt-, O-, and N-doping can intrinsically promote the electrochemical HER catalytic performance of CoS₂ and a line in ΔG_H* and d-band centers of H adsorption sites is achieved. The Pt-, O-, and N-doped CoS₂ show overpotentials of 0.03 V, 0.01 V, and 0.07 V, respectively, which are very close to values when using Pt, which is believed to be the best catalyst for the HER because of its small overpotential (about 0.08 V).³ Therefore, the CoS₂, after doping by Pt, O, and N becomes a promising catalyst.

4 Conclusions

In summary, we studied HER catalytic activities of cation/anion-doped CoS₂ as HER catalyst based on the DFT method. Our results reveal that the Pt-, N- and O-doped CoS₂ possesses an outstanding HER catalytic activity comparable to Pt. And a negative linear relationship was identified between ΔG_H* and d-band centers of active metal sites. This exhibits that d-band center of active metal site moves downwards after cation/anion doping, leading to the weakening of hydrogen adsorption and facilitating the H desorption from the surface of CoS₂. Therefore, the intrinsic HER catalytic performance of CoS₂ can be improved by tailoring the d-band center of the active metal site. Based on this study, we can further explore the relationship between the structure and activity of CoS₂, which may shed light on the exploration of the potential use of other TMSs as HER electrode materials.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work is supported by the Fund Project of Chengdu Technological University under Grant No. 2020RC004.

References

1. Y. Tachibana, L. Vayssieres and J. R. Durrant, Nat. Photonics, 2012, 6, 511–518.
2. B. Qiu, Y. Zhang, X. Guo, Y. Ma, M. Du, J. Fan, Y. Zhu, Z. Zeng and Y. Chai, J. Mater. Chem. A, 2022, 10, 719–725.
3. J. K. Nørskov, T. Bligaard, A. Logadottir, J. R. Kitchin, J. G. Chen, S. Pandelov and U. Stimming, ChemInform, 2005, 152, L5.
4. D. Saha, V. Patel, P. R. Selvaganapathy and P. Kruse, Nanoscale Adv., 2022, 4, 125–137.
5. B. Hinnemann, P. G. Moses, J. Bonde, K. P. Jørgensen, J. H. Nielsen, S. Horch, I. Chorkendorff and J. K. Nørskov, J. Am. Chem. Soc., 2005, 127, 5308–5309.
6. R. Li, J. Liang, T. Li, L. Yue, Q. Liu, Y. Luo, M. S. Hamdy, Y. Sun and X. Sun, Chem. Commun., 2022, 15, 633–644.
7. J. Tian, X. Xing, Y. Sun, X. Zhang, Z. Li, M. Yang and G. Zhang, Chem. Commun., 2022, 58, 557–560.
8. S. Mansingh, K. K. Das and K. Parida, Sustainable Energy Fuels, 2021, 5, 1952–1987.
9. H. Wang, X. Xiao, S. Liu, C.-L. Chiang, X. Kuai, C.-K. Peng, Y.-C. Lin, X. Meng, J. Zhao, J. Choi, Y.-G. Lin, J.-M. Lee and L. Gao, J. Am. Chem. Soc., 2019, 141, 18578–18584.
10. J. Lee, J. Heo, H. Y. Lim, J. Seo, Y. Kim, J. Kim, U. Kim, Y. Choi, S. H. Kim, Y. J. Yoon, T. J. Shin, J. Kang, S. K. Kwak, J. Y. Kim and H. Park, ACS Nano, 2020, 14, 17114–17124.
11. Y. Qu, H. Pan, C. Tat Kwok and Z. Wang, Phys. Chem. Chem. Phys., 2015, 17, 24820–24825.
12. Y. Lin, M. Yu, X. Li, W. Gao, L. Wang, X. Zhao, M. Zhou, X. Yao, M. He and X. Zhang, J. Mater. Chem. C, 2021, 9, 15877–15885.
13. J. Xu, Y. Zhu, B. Yu, C. Fang and J. Zhang, Inorg. Chem. Front., 2019, 6, 3510–3517.
14. Z. Wang, K. Yu, R. Huang and Z. Zhu, CrystEngComm, 2021, 23, 5097–5105.
15. N. Ran, B. Sun, W. Qiu, E. Song, T. Chen and J. Liu, J. Phys. Chem. Lett., 2021, 12, 2102–2111.
16. G. Chen, H. Li, Y. Zhou, C. Cai, K. Liu, J. Hu, H. Li, J. Fu and M. Liu, Nanoscale, 2021, 13, 13604–13609.
17. Y. Dong, H. Sun and G. Liu, Sustainable Energy Fuels, 2021, 5, 4115–4125.
18. Z. Shi, X. Qi, Z. Zhang, Y. Song, J. Zhang, C. Guo, W. Xu, K. Liu and Z. Zhu, Nanoscale, 2021, 13, 6890–6901.
19. Y. Jiang, S. Gao, J. Liu, G. Xu, Q. Jia, F. Chen and X. Song, Nanoscale, 2020, 12, 11573–11581.
20. Y. Zang, B. Yang, A. Li, C. Liao, G. Chen, M. Liu, X. Liu, R. Ma and N. Zhang, ACS Appl. Mater. Interfaces, 2021, 13, 41573–41583.
21. C. n. Karakaya, N. Solati, U. Savacı, E. Keleş, S. Turan, S. Çelebi and S. Kaya, ACS Catal., 2020, 10, 15114–15122.
22. C. Karakaya, N. Solati, U. Savacı, E. Keleş, S. Turan, S. Çelebi and S. Kaya, ACS Catal., 2020, 10, 15114–15122.
23. J. Zhang, Y. Liu, B. Xia, C. Sun, Y. Liu, P. Liu and D. Gao, Electrochim. Acta, 2018, 259, 955–961.
24. G. Papoian, J. K. Nørskov and R. Hoffmann, J. Am. Chem. Soc., 2000, 122, 4129–4144.
25. A. Vojvodic, A. Hellman, C. Ruberto and B. I. Lundqvist, Phys. Rev. Lett., 2009, 103, 146103.
26. Z. Chen, Y. Song, J. Cai, X. Zheng, D. Han, Y. Wu, Y. Zang, S. Niu, Y. Liu, J. Zhu, X. Liu and G. Wang, Angew. Chem., Int. Ed., 2018, 57, 5076–5080.
27. C. Tsai, K. Chan, J. K. Nørskov and F. Abild-Pedersen, J. Phys. Chem. Lett., 2014, 5, 3884–3889.
28. A. Maibam, S. Krishnamurty and M. Ahmad Dar, Mater. Adv., 2022, 3, 592–598.
29. A. Maibam, S. K. Das, P. P. Samal and S. Krishnamurty, RSC Adv., 2021, 11, 13348–13358.
30. J. M. Soler, E. Artacho, J. D. Gale, A. García, J. Junquera, P. Ordejón and D. Sánchez-Portal, J. Phys.: Condens. Matter, 2002, 14, 2745–2779.
31. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
32. J. Zhang, Y. Liu, C. Sun, P. Xi, S. Peng, D. Gao and D. Xue, ACS Energy Lett., 2018, 3, 779–786.
33. J. Shi, Z. Wang and Y. Q. Fu, J. Phys. D: Appl. Phys., 2016, 49, 505601.
34. W. I. Choi, B. C. Wood, E. Schwegler and T. Ogitsu, Adv. Energy Mater., 2015, 5, 1501423.
35. J. K. Nørskov, T. Bligaard, J. Rossmeisl and C. H. Christensen, Nat. Chem., 2009, 1, 37–46.
36. D. Kim, J. Shi and Y. Liu, J. Am. Chem. Soc., 2018, 140, 9127–9131.
37. Q. Sun, Q. Wang, P. Jena and Y. Kawazoe, J. Am. Chem. Soc., 2005, 127, 14582–14583.
38. Y. Sun, A. Huang and Z. Wang, RSC Adv., 2019, 9, 26321–26326.
39. G. Feng, M. V. Ganduglia-Pirovano, C.-F. Huo and J. Sauer, J. Phys. Chem. C, 2018, 122, 18445–18455.
40. A. Rasool, I. Anis, M. Dixit, A. Maibam, A. Hassan, S. Krishnamurty and M. A. Dar, Catal. Sci. Technol., 2022, 12, 310–319.

---