Exploitation of the Large-Area Basal Plane of MoS2 and Preparation of Bifunctional Catalysts through On-Surface Self-Assembly.

The development of nonprecious electrochemical catalysts for water splitting is a key step to achieve a sustainable energy supply for the future. Molybdenum disulfide (MoS2) has been extensively studied as a promising low-cost catalyst for hydrogen evolution reaction (HER), whereas HER is only catalyzed at the edge for pristine MoS2, leaving a large area of basal plane useless. Herein, on-surface self-assembly is demonstrated to be an effective, facile, and damage-free method to take full advantage of the large ratio surface of MoS2 for HER by using multiscale simulations. It is found that as supplement of edge sites of MoS2, on-MoS2 M(abt)2 (M = Ni, Co; abt = 2-aminobenzenethiolate) owns high HER activity, and the self-assembled M(abt)2 monolayers on MoS2 can be obtained through a simple liquid-deposition method. More importantly, on-surface self-assembly provides potential application for overall water splitting once the self-assembled systems prove to be of both HER and oxygen evolution reaction activities, for example, on-MoS2 Co(abt)2. This work opens up a new and promising avenue (on-surface self-assembly) toward the full exploitation of the basal plane of MoS2 for HER and the preparation of bifunctional catalysts for overall water splitting.

Introduction

Developing clean energy has been an irreversible momentum due to the gradual depletion of fossil fuel and the increasing release of carbon dioxide.1 Molecular n class="Chemical">hydrogen produced from electrochemical water splitting has been regarded as one of the cleanest energy carriers because of its nonpollution during the production and the combustion processes.2 Platinum (Pt) and its alloy are the most efficient electrochemical catalysts for hydrogen evolution reaction (HER) at present, whereas its high cost hinders the practical application severely.3, 4 Much effort has been devoted to developing nonprecious materials to replace Pt as highly active catalysts for HER.

With the emergence of 2D materials,5, 6, 7, 8, 9 several 2D materials have been demonstrated to be of high HER activity, such as n class="Chemical">MXenes,10, 11, 12, 13 C3N4,14, 15 boron monolayers,16 and molybdenum disulfide (MoS2).17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29 Among them, MoS2 is the most studied and is identified as the most promising alternative to Pt. Unfortunately, the active sites of pristine MoS2 are merely located at the edge, leaving a large area of in‐plane domains useless.17, 19, 29 Defect engineering of MoS2 is a common strategy toward the utilization of the basal plane,20, 23, 24, 25, 26, 27, 28 whereas it has a side effect on the stability of MoS2.20, 27 Another strategy is phase‐transition engineering of MoS2 from 2H to 1T (1T′) phases,18, 21, 22, 24 but the metastable nature21 of 1T or 1T'‐MoS2 hinders its practical applications.30, 31 Therefore, developing simple, effective, and damage‐free methods to exploit the large‐area basal plane of MoS2 is highly demanding.

Recently, molecular self‐assembly on gold and carbon‐based materials has been successfully used to prepare electrochemical catalysts toward n class="Chemical">O2 and CO2 reductions,32, 33, 34, 35 and 2D transition metal dichalcogenides were demonstrated to be suitable substrates for molecular self‐assembly.36 The organic molecules are physically self‐assembled on the underlying substrates via intermolecular weak interactions, thereby leading to no structural damage on the substrates. Inspired by such research, the motivation for the present work is to explore whether the basal plane of MoS2 can be fully exploited for HER through on‐surface self‐assembly. To this end, we need to verify that the following two conditions can be satisfied: (i) on‐MoS2 molecule possesses high HER activity; and (ii) the molecules self‐assembled on the basal plane of MoS2 via intermolecular weak interactions are stable and will not diffuse into water. By means of multiscale simulations combined density functional theory (DFT) and classical molecular dynamics (MD), we demonstrate that both the two preconditions can be fulfilled by M(abt)2 (M = Ni, Co; abt = 2‐aminobenzenethiolate) on MoS2, showing that on‐surface self‐assembly is indeed an effective approach to take advantage of the large‐area basal plane of MoS2 for HER without damage on MoS2. Also, a facile method toward the preparation of self‐assembled M(abt)2 monolayers is proposed, that is dropping M(abt)2 solution onto MoS2.

On‐surface self‐assembly has been used to prepare single‐function catalysts.32, 33, 34, 35 However, to the best of our knowledge, the preparation of bifunctional catalysts has never been reported through on‐surface self‐assembly. The rational design and the facile preparation of bifunctional catalysts toward overall water splitting have attracted ever‐increasing attention in recent years because of the reduced production n class="Chemical">cost compared with two separate single‐function catalysts for HER and oxygen evolution reaction (OER).37, 38, 39, 40, 41, 42, 43, 44 Our studies show that as complementary to edge sites of MoS2 and N sites of on‐MoS2 Co(abt)2 for HER, OER can be catalyzed at Co sites of on‐MoS2 Co(abt)2, suggesting that on‐surface self‐assembly is also able to serve for overall water splitting. Our idea, via on‐surface self‐assembly to make the best use of the large‐area in‐plane domains of MoS2 for HER and prepare bifunctional catalysts for overall water splitting, is schematically displayed in Figure a.

Figure 1

a) Schematic used to reflect our idea. Structures of b) Ni(abt)2 crystal and c) Ni(abt)2 on MoS2. d) Calculated ΔG profile of HER at the surface N, S, and Ni sites as well as the inner N and S sites of Ni(abt)2 crystal. e) Calculated ΔG profile of HER at N, S, and Ni sites of on‐MoS2 Ni(abt)2. Black, red, gray, white, blue, yellow, and green balls stand for Ni, Co, C, H, N, S, and Mo atoms, respectively.

a) Schematic used to reflect our idea. Structures of b) Ni(abt)2 crystal and c) n class="Chemical">Ni(abt)2 on MoS2. d) Calculated ΔG profile of HER at the surface N, S, and Ni sites as well as the inner N and S sites of Ni(abt)2 crystal. e) Calculated ΔG profile of HER at N, S, and Ni sites of on‐MoS2 Ni(abt)2. Black, red, gray, white, blue, yellow, and green balls stand for Ni, Co, C, H, N, S, and Mo atoms, respectively.

Results and Discussion

HER Activity of On‐MoS2 Ni(abt)2

Ni(abt)2 crystal has been demonstrated to be of high n class="Chemical">HER activity experimentally,45 and the structure is shown in Figure 1b. We first explore the active sites of the bulk Ni(abt)2, including the surface and inner positions of the crystal. The HER and OER activities are both evaluated by the binding free energy (ΔG) that was developed by Nørskov and co‐workers,46, 47 which has proven a powerful way to predict new electrocatalysts and understand the reaction mechanisms.10, 11, 14, 27, 46, 47, 48, 49, 50, 51, 52 The HER is calculated in acidic circumstance and the process is described asin which the * stands for the catalytic site and the smaller the absolute value of ΔG is, the better HER performance is. On the surface, the ΔG of N, S, and Ni sites are −0.07, 0.67, and 1.02 eV, respectively (see Figure 1d). It is obvious that the HER activity of N site is far higher than that of S and Ni sites, in good agreement with the experiment.45 For the inner part, the N site (ΔG is −0.17 eV) also owns better HER activity in comparison with the S site (ΔG is 0.55 eV), while the Ni site is unfavorable for hydrogen adsorption due to easy migration to adjacent sites. By comparing ΔG of surface and inner N sites, it can be derived that the surface N site is superior to the inner of the bulk material in the HER activity.

In addition, Figure S1d of the Supporting Information records the n class="Chemical">water distribution when Ni(abt)2 crystal is placed in water. The simulation shows that water molecules distribute on the surface of Ni(abt)2 crystal and do not penetrate into the inner part of the bulk. Considering that the inner HER activity is not high and water molecules are difficult to diffuse into the bulk, it can be concluded that HER mainly concentrates on the surface N sites of Ni(abt)2 crystal. In this respect, it will be highly material‐saving if Ni(abt)2 molecules are processed into ultrathin films or even monolayer, which can be potentially realized by bottom‐up on‐surface self‐assembly.53, 54, 55, 56, 57

Next, Ni(abt)2 molecule is placed on the surface of n class="Chemical">MoS2, and the HER activity of on‐MoS2 Ni(abt)2 is evaluated accordingly. As shown in Figure 1e, ΔG of N, S, and Ni sites of on‐MoS2 Ni(abt)2 are −0.02, 0.64, and 1.11 eV, respectively, and the N site is responsible for HER. The ΔG values are similar to the surface sites of Ni(abt)2 crystal, indicating that Ni(abt)2 is weakly coupling with MoS2 surface. Moreover, the distance between MoS2 and Ni(abt)2 is above 3.0 Å (a typical distance via weak interaction), further suggesting that Ni(abt)2 is physically self‐assembled on MoS2 surface via intermolecular weak interaction and thus will not bring damage on MoS2. Therefore, the self‐assembly of Ni(abt)2 molecules on MoS2 keeps high HER activities of both N sites from on‐MoS2 Ni(abt)2 and edge sites from MoS2. The coverage is further increased by two times, close to the coverage limit. Under such a high coverage, the N site still owns high HER activity and is considerably more superior than S and Ni sites (see Figure S2, Supporting Information), in good agreement with the results in Figure 1e, showing that the system is insensitive to the coverage.

Stability of On‐MoS2 Ni(abt)2

Another essential prerequisite is that n class="Chemical">Ni(abt)2 molecules physically self‐assembled on MoS2 via intermolecular weak interactions are stable on MoS2 and will not diffuse into water. Figure a depicts the initial structure of Ni(abt)2 molecules on MoS2 placed in water and Figure 2b presents the structure after 50 ns. From Figure 2b, it can be seen that all Ni(abt)2 molecules still adhere to the MoS2 surface and do not diffuse into water. The average height of Ni(abt)2 molecules relative to the MoS2 substrate is recorded in Figure 2c. The height only fluctuates in a tiny range (less than 0.1 Å), further demonstrating that on‐MoS2 Ni(abt)2 molecules are very stable in water. Like the case at 300 K, the system is still very stable even at a slightly higher temperature, e.g., 350 K (see Figure S3, Supporting Information). We also explore the opposite case, that is Ni(abt)2 molecules initially placed in water (see Figure 2d). Figure 2e–i records the evolution of the structure in Figure 2d with time. Obviously, the molecules fast escape from water and move onto the MoS2 surface, which unambiguously illustrates that Ni(abt)2 prefers staying on the surface of MoS2 rather than diffusing into water.

Figure 2

a) Initial structure of on‐MoS2 Ni(abt)2 molecules in water. b) Evolution of (a) after 50 ns. c) Average height of Ni(abt)2 molecules relative to MoS2 as a function of time. d–i) Dynamic process of Ni(abt)2 molecules from staying in water to lying on MoS2. The snapshots were taken every 2 ns.

a) Initial structure of on‐MoS2 n class="Chemical">Ni(abt)2 molecules in water. b) Evolution of (a) after 50 ns. c) Average height of Ni(abt)2 molecules relative to MoS2 as a function of time. d–i) Dynamic process of Ni(abt)2 molecules from staying in water to lying on MoS2. The snapshots were taken every 2 ns.

H2 is generated from large‐area in‐plane domains of n class="Chemical">MoS2, so it is desirable to explore the effect of H2 on the system stability. Figure S4 of the Supporting Information records the dynamic evolution of 80 Ni(abt)2 and 20 H2 molecules placed on MoS2. H2 molecules fast escape from MoS2 and diffuse into water, showing that on‐MoS2 H2 is very unstable. The intrinsic reason is owing to the extremely weak interaction between H2 and MoS2, only 0.06 eV. By contrast, Ni(abt)2 molecules have not been influenced by H2 and adhere to the MoS2 surface throughout. Therefore, we conclude that the generated H2 fast diffuses into water and does not give rise to the instability of Ni(abt)2 on MoS2.

Preparation of On‐MoS2 Ni(abt)2

We have demonstrated that on‐MoS2 n class="Chemical">Ni(abt)2 is of high HER activity above. Next, we will explore how to prepare self‐assembled Ni(abt)2 monolayers on MoS2. Many studies have shown that ultrathin self‐assembled organic monolayers can be prepared through a facile liquid‐deposition method,53, 54, 55, 56, 57 i.e., organic molecules are first dissolved in solvent and self‐assembled organic monolayers can be naturally generated on the substrate on which the mixed solution is dropped. From the preparation process, two key points toward the preparation of self‐assembled Ni(abt)2 monolayers on MoS2 can be derived: (i) an appropriate solvent that can dissolve Ni(abt)2 and (ii) dissolved Ni(abt)2 molecules that can deposit onto MoS2.

Figure a–c and Movie S1 (Supporting Information) describe the dynamic behavior of Ni(abt)2 molecules in n class="Chemical">water. Ni(abt)2 molecules are uniformly placed in water in the beginning (Figure 3a); thereafter, Ni(abt)2 molecules gradually aggregate (Movie S1, Supporting Information) till 12 ns at which all Ni(abt)2 molecules aggregate together as shown in Figure 3c. The results show that Ni(abt)2 is insoluble in water, in line with the experiment.45 Alternatively, Ni(abt)2 is known to be crystallized from diethyl ether,45 so MD simulation of Ni(abt)2 molecules in diethyl ether is performed. As expected, the Ni(abt)2 molecules are fast separated from each other in diethyl ether (see Figure 3d–f; Movie S2, Supporting Information) instead of aggregation as found in water, indicating that Ni(abt)2 molecules can be dissolved in diethyl ether well.

Figure 3

Dynamic behaviors of Ni(abt)2 in a–c) water and d–f) diethyl ether. a) Initial structure of Ni(abt)2 molecules in water and b,c) snapshots taken from the evolution of a) 6 and 12 ns later. d) Initial structure of the Ni(abt)2 cluster formed in c) placed in diethyl ether. e,f) Snapshots at 1.5 and 2.5 ns taken from the evolution of (d). Note that the hydrogen atoms of diethyl ether are not displayed for better visual effect.

Dynamic behaviors of Ni(abt)2 in a–c) n class="Chemical">water and d–f) diethyl ether. a) Initial structure of Ni(abt)2 molecules in water and b,c) snapshots taken from the evolution of a) 6 and 12 ns later. d) Initial structure of the Ni(abt)2 cluster formed in c) placed in diethyl ether. e,f) Snapshots at 1.5 and 2.5 ns taken from the evolution of (d). Note that the hydrogen atoms of diethyl ether are not displayed for better visual effect.

Next, we investigate whether n class="Chemical">Ni(abt)2 molecules in diethyl ether can deposit onto MoS2. Note that diethyl ether is highly volatile if exposed to air, so diethyl ether molecules must gradually diffuse into air when Ni(abt)2 solution is dropped on MoS2. Therefore, to reflect the actual situation more reasonably, the volatilization process is considered in our simulations. Figure a presents the initial structure of Ni(abt)2 solution dropped onto MoS2. Obviously, diethyl ether molecules gradually escape from the solution, but Ni(abt)2 molecules remain in solution and gradually deposit onto MoS2 (see Figure 4b–l). Finally, all Ni(abt)2 molecules deposit onto MoS2 as shown in Figure 4l, showing clearly that self‐assembled Ni(abt)2 monolayers on MoS2 can be prepared through a simple liquid‐deposition method.

Figure 4

Liquid deposition of Ni(abt)2 molecules in diethyl ether solution onto MoS2 with the volatilization of diethyl ether molecules. a) Initial structure and b–l) snapshots taken from the evolution of (a) after 6, 12, 18, 24, 30, 36, 42, 48, 54, 60, and 64 ns, respectively. For better visual effect, the hydrogen atoms of diethyl ether are not displayed.

Liquid deposition of Ni(abt)2 molecules in n class="Chemical">diethyl ether solution onto MoS2 with the volatilization of diethyl ether molecules. a) Initial structure and b–l) snapshots taken from the evolution of (a) after 6, 12, 18, 24, 30, 36, 42, 48, 54, 60, and 64 ns, respectively. For better visual effect, the hydrogen atoms of diethyl ether are not displayed.

Bifunctional Catalysts for Overall Water Splitting

We have shown that the on‐surface self‐assembly of Ni(abt)2 rn class="Gene">etains the merits of high HER activities from both MoS2 and Ni(abt)2. Also, the integration of various active sites from components of the complex is promising for bifunctional or multifunctional catalysts, as learned from nanoscale alloys37, 38, 39, 40, 41 and doping carbon‐based materials,42, 43, 44 which are common approaches toward the preparation of bifunctional catalysts for overall water splitting. Hence, we want to explore whether on‐surface self‐assembly has potential to be a new strategy to prepare bifunctional catalysts. The self‐assembled system consisting of MoS2 and Ni(abt)2 has been demonstrated to be of high HER activity, so if on‐MoS2 Ni(abt)2 is also of high OER activity, it can be served as the bifunctional catalyst for overall water splitting. For OER, the reaction consists of four elementary processes where the * represents the catalytic site. The smaller the overpotential η is, the higher the OER activity is. η is defined asin which ΔG 1, ΔG 2, ΔG 3, and ΔG 4 are the free energy differences of the four processes mentioned above, respectively. In general, oxygen evolution or reduction occurs at transition metal atoms. As shown in Figure b, for the Ni site of on‐MoS2 Ni(abt)2 the calculated ΔG 1, ΔG 2, ΔG 3, and ΔG 4 are 2.00, 0.04, 2.91, and −0.03 eV, respectively. The rather positive ΔG 1 and ΔG 3 suggest the weak bonding strength between the Ni site and OH* and OOH*. The process from O* to OOH* is the potential‐determining step and the overpotential η reaches up to 1.68 V, showing that the OER activity is very poor.

Figure 5

a) Calculated ΔG profile of HER at N, S, and Co sites of on‐MoS2 Co(abt)2. ΔG diagrams for OER at b) the Ni site of on‐MoS2 Ni(abt)2 and c) the Co site of on‐MoS2 Co(abt)2. d) Calculated projected density of states (PDOS) of d band for the Ni atom of Ni(abt)2 on MoS2 and the Co atom of Co(abt)2 on MoS2. The d band center is marked by the red dashed line and the Fermi level is set as zero.

a) Calculated ΔG profile of HER at N, S, and n class="Chemical">Co sites of on‐MoS2 Co(abt)2. ΔG diagrams for OER at b) the Ni site of on‐MoS2 Ni(abt)2 and c) the Co site of on‐MoS2 Co(abt)2. d) Calculated projected density of states (PDOS) of d band for the Ni atom of Ni(abt)2 on MoS2 and the Co atom of Co(abt)2 on MoS2. The d band center is marked by the red dashed line and the Fermi level is set as zero.

As Co‐based materials are anotn class="Chemical">her commonly applied and nonprecious electrocatalysts, we further investigate the catalytic behavior of on‐MoS2 Co(abt)2. Like Ni(abt)2 for HER, the N site owns high HER activity (ΔG is only 0.04 eV) and its HER activity is comparable to that of on‐MoS2 Ni(abt)2 (see Figure 5a,e). Figure S5 of the Supporting Information further demonstrates that Co(abt)2 molecules on MoS2 are very stable in water and they prefer staying on MoS2 instead of in water as well. The high HER activity and water stability guarantee that self‐assembled Co(abt)2 monolayers can be used to effectively utilize the large‐area basal plane of MoS2 for HER. In addition, self‐assembled Co(abt)2 monolayers on MoS2 can also be prepared by a simple operation similar to on‐MoS2 Ni(abt)2, as shown in Figure S6 of the Supporting Information. Unlike Ni(abt)2 for OER, each step is moderate for the Co site of on‐MoS2 Co(abt)2 (see Figure 5c). The potential‐determining step is the process from OH* to O* and the overpotential η is only 0.43 V, comparable to traditional precious metal‐based OER catalysts in a range of about 0.3–0.7 V,51, 58, 59 showing that OER can be efficiently catalyzed at the Co site of on‐MoS2 Co(abt)2. Our calculations show that OER can be catalyzed by Co(abt)2 along the four‐electron pathway. However, this does not mean that OER is impossible to occur at Co(abt)2 along other pathways. There may even exist better pathways with lower OER overpotentials compared to the four‐electron pathway, but this will not change our conclusion that Co(abt)2 can effectively catalyze OER. Therefore, we can conclude that on‐MoS2 Co(abt)2 is of both high HER and OER activities, suggesting that on‐surface self‐assembly can serve for the preparation of efficient bifunctional catalysts toward overall water splitting.

The binding energies of MoS2 and the intermediates of Mn class="Chemical">(abt)2 during HER and OER (see Figure S7a–e, Supporting Information) are −1.76, −1.59, −1.56, −1.43, and −1.75 eV, respectively, which is comparable to that between MoS2 and M(abt)2 (−1.63 eV). This shows that M(abt)2 on MoS2 is still stable during HER and OER. The distinct catalytic performance for the Ni site of Ni(abt)2 and the Co site of Co(abt)2 can be understood well by the theory of d band center (εd). The closer to the Fermi level the εd is, the stronger the binding of the adsorbate and the catalytic site is. As shown in Figure 5d, εd of the Co atom is closer to the Fermi level; therefore the binding of the adsorbates and the Co atom is stronger, leading to lower ΔG of the adsorbate H, OH, and OOH (see Figures 1e and 5a–c). However, the adsorbate O is abnormal whose binding with Co(abt)2 is much weaker. Such an abnormal phenomenon can be explained by the structure difference. As opposite to the other cases, the O atom is unable to steadily adsorb on the Ni atom and it insets into the Ni—S bond (see Figure S7, Supporting Information), which strengthens the bonding of the O atom and Ni(abt)2, thus resulting in a lower ΔG.

Conclusion

In summary, we have reported a new and promising avenue (on‐surface self‐assembly) to make full use of the large‐area in‐plane domains of MoS2 for n class="Chemical">HER and prepare bifunctional catalysts for overall water splitting. We demonstrated that the weak interaction between M(abt)2 molecules and MoS2 plays an important role in the effective bottom‐up on‐surface self‐assembly from three aspects: (a) rendering a facile preparation by liquid‐deposition method possible, (b) keeping M(abt)2 molecules distributing on the MoS2 surface instead of diffusing into water, and (c) preserving high HER activities of both edge sites of MoS2 and N sites of M(abt)2 molecules. Therefore, efficient utilization of the large‐area basal plane of MoS2 for HER is achieved based on significantly increased active sites by on‐surface self‐assembly, which is also anticipated to be generally applicable to exploit the basal plane of other transition metal dichalcogenides. In addition, the bottom‐up self‐assembly of molecules onto MoS2 surface can bring other desired properties for multifunctional applications. For example, our results show that on‐MoS2 Co(abt)2 is able to catalyze OER, suggesting that on‐surface self‐assembly can also be used to prepare bifunctional catalysts for overall water splitting. This work provides an effective and facile strategy, on‐surface self‐assembly, to prepare efficient catalysts for water splitting, thereby offering high possibility toward achieving a sustainable energy supply.

Experimental Section

DFT Calculations: All DFT calculations were carried out through the projector augmented wave method60 with spin polarization and van der Waals (vdW) modification (D3)61 as implemented in the Vienna ab initio simulation package.62 The exchange‐n class="Chemical">correlation functional was built on the functional of Perdew, Burke, and Ernzerhof of generalized gradient approximation.63 A 7 × 3√3 × 1 supercell was adopted for the MoS2 substrate. The Brillouin zone was sampled using the Monkhorst–Pack scheme with k‐point mesh of 3 × 3 × 1 for 2D systems and 3 × 3 × 3 for 3D systems. The kinetic energy cutoff for the plane‐wave basis set was set as 500 eV. All structures were fully relaxed until reaching the convergence threshold of 0.02 eV Å−1 for force and 10−4 eV for energy.

MD Simulations: All simulations were performed by using the software Gromacs version 464 under the condition of 300 K. Intermolecular interactions n class="Chemical">consist of the vdW and electrostatic interactions, which were calculated according to Coulomb's law and 12–6 Lennard‐Jones (LJ) potential, respectively. The cutoff distance of intermolecular interactions was set to 1.5 nm outside which the smoothed particle mesh Ewald sum65 was used to deal with the long‐range electrostatic interaction and the vdW interaction was not considered. The time step was set to 1 fs and the berendsen thermostat and barostat were utilized to control temperature and pressure, respectively. Periodic boundary conditions were employed to avoid the edge effect. Force field (FF) parameters of Ni(abt)2 or Co(abt)2 were constructed by combining all‐atom Amber99sb FF66, 67 with universal FF.68 The reliability of the constructed FF parameters is guaranteed by the good agreement between simulation and experiment for the lattice parameters of Ni(abt)2 crystal (see Table S1). Moreover, the simulated dynamic behavior of Ni(abt)2 in water and diethyl ether, i.e., insoluble in water and soluble in diethyl ether, agrees with the experimental results, further guaranteeing the reliability of the constructed FF parameters. The FF parameters for diethyl ether were built on all‐atom Amber99sb FF and water was described by extended simple point charge model. Molecular partial charges were obtained based on the Chelpg methodology.69 The atomic charges of MoS2 were obtained from ref. 70 and its LJ parameters were derived according to the calculated molecule‐MoS2 interaction from DFT calculations. The size of MoS2 used in simulations is 10.18592 nm × 9.92124 nm consisting of 1152 Mo and 2304 S atoms. The evaporation process of diethyl ether in Figure 4 was mimicked by exhausting the escaped molecules every 2 ns.

ΔG Calculations: ΔG is the difference of the free energy (G) between products and reactants and G is calculated aswhere E, E ZPE, T, and S represent the energy, the zero‐point energy, the reaction temperature, and the entropy, respectively. The difference of the zero‐point energies is obtained via vibrational frequency calculation for the catalyst with and without adsorbed species. The difference between the entropies of the catalyst with and without adsorbed species is very small and neglected. All calculations were done under the standard n class="Chemical">hydrogen electrode in which the free energy of protons and electrons (H+ + e −) is taken as a half of the free energy of gas H2. The solvation corrections to account for the effect of water as well as the zero‐point energies and the entropies of gas H2 and liquid H2O were obtained from ref. 49. The free energy of gas O2 () is calculated byin which represents the free energy of gas H2, stands for the free energy of liquid H2O, and 4.92 eV is the reaction free energy for splitting liquid H2O into gas H2 and O2.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary

Click here for additional data file.

Supplementary

Click here for additional data file.

Supplementary

Click here for additional data file.