Understanding the Reactivity of Layered Transition-Metal Sulfides: A

Letter pubs.acs.org/JPCL

Understanding the Reactivity of Layered Transition-Metal Sulfides: A Single Electronic Descriptor for Structure and Adsorption Charlie Tsai,†,‡ Karen Chan,†,‡ Jens K. Nørskov,†,‡ and Frank Abild-Pedersen*,‡ †

Department of Chemical Engineering, Stanford University, Stanford, California 94305, United States SUNCAT Center for Interface Science and Catalysis, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States

‡

S Supporting Information *

ABSTRACT: Density functional theory is used to investigate the adsorption and structural properties of layered transition-metal sulfide (TMS) catalysts. We considered both the (1010̅ ) M-edge and (1̅010) S-edge terminations for a wide range of pure and doped TMSs, determined their sulfur coverage under realistic operating conditions (i.e, steady-state structures), and calculated an extensive set of chemisorption energies for several important reactions. On the basis of these results, we show that the d-band center, εd, of the edge-most metal site at 0 ML sulfur coverage is a general electronic descriptor for both structure and adsorption energies, which are known to describe catalytic activity. A negative linear correlation between adsorbate− S binding and S−metal binding allows εd to describe the adsorption of species on both metal and sulfur sites. Our results provide a significant simplification in the understanding of structure−activity relationships in TMSs and provides guidelines for the rational design and large-scale screening of these catalysts for various processes. SECTION: Surfaces, Interfaces, Porous Materials, and Catalysis

L

which correlates the shifts in the d-band center of the metal with the filling of antibonding states in the adsorbate. Further up-shifted (down-shifted) d-band centers lead to lesser (greater) filling of the antibonding states and hence stronger (weaker) binding. The catalytic activity of transition metals is thus described by a single electronic parameter. An analogous model based on surface resonance states has also been developed for the transition-metal carbides.31 However, the search for an analogous electronic descriptor for TMS activity faces several challenges: (1) the edge structure is highly dependent upon synthesis and reaction conditions, leading to different S terminations for different TMS catalysts under the same conditions; (2) adsorbates could bind to either a metal or a sulfur atom; (3) there exists two types of edges in TMS, either the (101̅0) M-edge or the (1̅010) S-edge,18 which are different in stability and activity. In this Letter, we use density functional theory (DFT) calculations to show that the d-band center, εd, of the edgemost metal atom at 0 ML sulfur coverage provides an electronic parameter that can simultaneously describe adsorption on the two types of edges across a wide range of TMS catalysts, as well as account for structural variations under operating conditions. By determining εd at 0 ML sulfur coverage, the descriptor is independent of the specific edge structure and adsorption site. We thus overcome all of the issues mentioned above. Our results demonstrate that, despite the complexity of the local

ayered transition-metal sulfide (TMS) catalysts have gained increased interest in recent years due to their unique chemical and physical properties resulting from their low dimensionality.1 Their general formula is MS2, where M is a transition metal. The most well studied member of the TMS is MoS2, which has been studied extensively for hydrodesulfurization,2−8 alcohol synthesis,9 electrochemical hydrogen evolution,10−13 and, recently, electrochemical CO2 reduction.14,15 The one-dimensional edge sites10,11,16−18 along the layers have been shown to be responsible for the catalytic activity. Recently, closely related structures such as WS212,19 and their selenide analogues13,20−22 have also been shown to exhibit high catalytic activity, which can be further modified by transition-metal doping into the edges.12,23−25 The successful application of TMS catalysts to both gas-phase thermocatalysis as well as electrocatalysis has led to intensified efforts in studying layered TMSs as a general class of active and earth-abundant catalysts. Despite the potential applications for TMS catalysts, there has not yet been a general theoretical framework for systematically understanding their structure and activity. If the description of catalytic activity could be reduced to a single electronic structural parameter, understanding of the activity trends would be greatly simplified, which allows for the rational design of next-generation catalysts.26,27 One major simplification arises from the linear scaling of chemisorption energies of key reaction intermediates with their activation barriers, which allows the trends in activity to be described from just the adsorption energies.28,29 For the transition metals and their alloys, the various adsorption energies are well-described by the d-band model,30 © 2014 American Chemical Society

Received: September 27, 2014 Accepted: October 22, 2014 Published: October 22, 2014 3884

dx.doi.org/10.1021/jz5020532 | J. Phys. Chem. Lett. 2014, 5, 3884−3889

The Journal of Physical Chemistry Letters

Letter

stable edge structure at 0 VRHE and a H2S pressure between 10−5 and 10−8 bar (standard corrosion resistance), we determined the stable coverage to be where hydrogen evolution is more downhill than either H2S desorption or further H adsorption14,22 (details are in the Supporting Information). Because chemisorption energies are known to describe catalytic activity,28,29 we considered the adsorption of H, SH, O, OH, CHO, COOH, NNH, and NH2, which are key activity descriptors for the hydrogen evolution reaction,34,35 the oxygen evolution reaction,36 the electrochemical CO2 reduction reaction,37 and the electrochemical NH3 synthesis reaction,38 respectively. We considered adsorption at both the sulfur and the sulfur defect sites. The sulfur defect site is defined as the vacancy formed when a sulfur atom at the edge is desorbed as H2S(g). Adsorption energies were calculated using DFT on an infinite stripe model for modeling the TMS edge sites. This has been described in detail in a number of recent studies.10,18,22 The adsorption energies for H,10 S and SH,22 O and OH,36 CHO and COOH,14,39 and NNH and NH2,38 are also defined in the same way as in past studies for pure transition metal surfaces. The adsorption energies for H, O, OH, CHO, and COOH were calculated for a S site, whereas the adsorptions of SH, NNH, and NH2 were calculated for a defect site. In almost all cases, NNH and NH2 bind too weakly onto the S, and defects would be necessary to stabilize them on the edge. SH is used as a descriptor for stability in a reducing environment because the degradation of the TMS catalyst is related to the removal of SH as H2S (computational details are in the Supporting Information). The d-projected densities of states were calculated on the edge-most metal with a S coverage of 0 ML (i.e., the bare metal edge with no S present in Figure 1a), and the d-band center εd is defined as

catalyst structure, it is still possible to significantly reduce the relevant parameters and describe key quantities without making unrealistic assumptions and simplifications. Because the most studied layered TMS catalysts, MoS2 and WS2, exist primarily in the 2H trigonal prismatic structure, we focus our study on 2H structures in this study (2H-MoS2, 2HWS2, 2H-NbS2, and 2H-TaS2). To show greater generality, we consider both types of edges (the (101̅0) M-edge and the (1̅010) S-edge) and also consider transition-metal-doped Sedges of MoS2 with a wide range of metal dopants: Ag, Au, Co, Ni, Os, Pd, Pt, Rh, and Ru. Here, we have assumed 100% substitution of Mo atoms at the edge as a first approximation in order to obtain broader trends. We focus on the edge sites in this study because they are usually the active sites for catalysis. In general, synthesis and reaction conditions will determine the stable coverage of terminating sulfur atoms (i.e., the edge structure), which can lead to vastly different catalytic activities.18,32,33 In this study, one set of experimental reaction conditions was used to determine all edge structures, several of which are shown in Figure 1a. The S coverage θS is defined as a fraction of a monolayer with respect to S atoms at the edge (a full row of S dimers is 1.0 ML), and the H coverage θH is defined as a fraction of a monolayer with respect to available adsorption sites. We consider S and H coverages from 0 to 1.0 ML for all systems. Starting with the thermodynamically most

∞

εd =

∫−∞ ρ(ε)ε dε ∞

∫−∞ ρ(ε) dε

(1)

where ρ are the densities of states and ε are the energies. We begin with the following observation: in the case of H adsorption on a S site (the simplest case), when the d-projected densityof states (a selection is shown in Figure 1b) is arranged in order of decreasing hydrogen adsorption energy ΔEH at the stable S and H coverages. A more down-shifted εd corresponds to a more negative ΔEH (i.e., how strongly H is adsorbed). The edges states are metallic;17,18 therefore, this trend is analogous to the d-band model for pure metals,30 except the shift in adsorption is in the opposite direction. εd of the metal can therefore describe chemisorption at a sulfur site for various S and H coverages (i.e., steady-state edge structures). To quantify this trend, we begin by defining the characteristic S−metal binding at the edge in terms of the adsorption of a single S atom at the bare metal-terminated edge, ΔE S (equivalent to the lowest possible S coverage in our unit cell, θS = 0.125 ML). This choice is made to compare the S−metal binding strengths across TMS structures for a fixed coverage. By minimizing the coverage of S at the edge, the energetic contribution due to the rearrangement of S atoms is also reduced.22,40,41 We define the characteristic H−S binding at the edge as the binding of a single H atom on that S atom, ΔEH−S, (equivalent to θH = 0.25 ML at θS = 0.125 ML). As shown in Figure 2a, ΔES versus ΔEH−S has a linear correlation with a negative slope. This shows that binding of the H adsorbate to the S atom is directly determined by how strongly the S atom is

Figure 1. (a) Various edge structures with different S coverages, θS, and H coverages, θH. The coverages are defined as a fraction of a monolayer with respect to the number of sites available for adsorption. The location of the edge-most metal is highlighted in green. This is also the location where the dopant metals will substitute. (b) Selection of the d-projected density of states calculated on the edge-most metal site at θS = 0 ML (no sulfur), arranged in order of increasing hydrogen binding strength (decreasing ΔEH), at their stable S and H coverages (ML, shown as subscripts of each species). We show different pure TMSs with different edges, as well as doped MoS2. The shaded regions indicate the occupied states up to the Fermi level εF. The binding strength of hydrogen increases (ΔEH decreases) with a downward shift of εd, which is the opposite trend seen in pure transition metals. 3885

dx.doi.org/10.1021/jz5020532 | J. Phys. Chem. Lett. 2014, 5, 3884−3889

The Journal of Physical Chemistry Letters

Letter

entire density of d-states, it is intimately related to the bond order of S−metal. ΔEH−S is simply the difference between ΔESH and ΔES (indicated by the shaded region in Figure 2c); therefore, the slope of ΔEH−S versus εd should have a magnitude that is −1/2 of that of ΔES versus εd, which is what we find approximately (0.55 and −0.87, respectively). Figure 3 shows that the trends in the stable S coverage θS for both the M-edge and S-edge are also described by εd. This is

Figure 3. Plot of the stable S coverage at each type of edge, θS, as a function of the d-band center εd. The S coverages for the M-edge and S-edge increase with εd because a more up-shifted εd corresponds to stronger S binding. Figure 2. (a) H adsorption energy ΔEH−S on a S site as a function of the S adsorption energy ΔES on the edge-most metal. Adsorption energies were calculated at θS = 0.125 ML and θH = 0.25 ML. In the unit cell, this represents one S atom on the edge and one H atom on that S atom. (b) ΔES and ΔEH−S as a function of εd. (c) Linear fits corresponding to ΔEH−S, ΔES, and ΔESH as a function of εd. The shaded region indicates the difference between ΔESH and ΔES, which corresponds to ΔEH−S = ΔESH − ΔES. The differences in SH and S binding allows for a direct description of H−S binding using εd.

simply a consequence of ΔES being determined by εd. Further down-shifted εd corresponds to weaker S adsorption and a lower steady-state S coverage, while an up-shifted ε d corresponds to stronger S adsorption and hence higher stable S coverages. In reality, θS should vary continuously, but the finite size of our unit cells prevents the sampling of certain stable coverages. Having understood how the εd successfully describes adsorption onto the metal site (ΔES), adsorption onto the S site (ΔEH−S), and the S coverage or steady-state structure (θS), we extend the analysis to the adsorption of other species on the stable structures under operating conditions. The H, SH, O, OH, CHO, COOH, NNH, and NH2 adsorption energies are plotted as a function of the d-band center in Figure 4. All of the adsorption energies scale linearly with the d-band center. As with ΔEH−S, all adsorbates bound at a S site (O, OH, CHO, COOH) scale positively with the d-band center. Again, this is due to the negative linear scaling between adsorbate−S binding and S−metal binding and the positive scaling between S−metal binding and εd. All adsorbates bound directly to a metal site at the S defect (SH, NNH, NH2) scale negatively with the d-band center, similarly to that of the d-band model for transition metals. Comparing ΔEH−S versus εd (Figure 2c), which has a constant S coverage, with ΔEH versus εd at the various stable coverages (Figure 4), the effect of varying the S coverage is to decrease the slope of the scaling line (from 0.55 to 0.17). Structures with down-shifted εd will have low steady-state S coverages, and ΔEH will be similar to ΔEH−S, which was calculated at a constant low S coverage; structures with more up-shifted εd will have higher steady-state S coverages, which weakens the binding of each S on the edge and strengthens the H binding on the S sites; thus, ΔEH < ΔEH−S. The d-band center εd is thus a valid general descriptor independent of the specific operating conditions, which determine edge structure (S coverage). The preservation of the trends at varying stable coverages at given reaction conditions for both types of edges

held at the edge; stronger (weaker) S−metal binding at the edge-most sites will have correspondingly weaker (stronger) H−S binding. Both M-edges and S-edges across several transition-metal-containing structures have been considered, which suggests that the negative scaling between ΔES and ΔEH−S is general for a sulfur-terminated site. The linear scaling between ΔES and ΔEH−S indicates that an electronic descriptor that describes ΔES should also describe ΔEH−S. Figure 2b shows that indeed the d-band center εd correlates linearly with ΔES. Because the edges are metallic, this is analogous to the d-band model for pure transition metals,30 where down-shifted d-states lead to more filling of the adsorbate antibonding states, which weakens the bonding strength of S. Due to the negative linear relation between ΔES and ΔEH−S (Figure 2a), the d-band center also determines how strongly H is bound to S through the negative correlation between εd and ΔEH−S. Because the d-band center can describe both ΔES and ΔEH−S, the d-band center thus has the potential to simultaneously describe the coverage of S on the edge (through the ΔES) as well as the adsorption of H and other reaction intermediates adsorbed on S at the stable coverages. An explanation for the negative linear scaling between ΔES and ΔEH−S (and, hence, the scaling between εd and ΔEH−S) arises from the differences in S and SH binding on the edge. We first define a characteristic SH−metal binding energy on the edge as ΔESH = ΔES + ΔEH−S. Both ΔES and ΔESH scale linearly with εd in Figure 2c, but the ΔESH line has a smaller slope. This is due to bond order conservation, which means that the S−metal bond order is reduced by half when the S−H bond is formed. Because εd involves an integration over the 3886

dx.doi.org/10.1021/jz5020532 | J. Phys. Chem. Lett. 2014, 5, 3884−3889

The Journal of Physical Chemistry Letters

Letter

under other conditions as well. The change in environment would affect the steady-state sulfur coverage but consistently across the TMS catalysts based on how strongly sulfur is adsorbed on the edge, which is described by εd. In this Letter, we have shown the d-band center of the edgemost metal at a S coverage of 0 ML to be a single electronic descriptor for the structure and adsorption on layered TMS. This descriptor is completely general in layered TMSs because the structure and adsorption energies at various structures, edge types, and both pure and doped TMS are all well-described. Our results provide a systematic framework for understanding and predicting the structure and reactivity of layered TMS compounds. We have recently shown the similarities between TMSs and TM selenides, specifically with regards to the negative linear scaling between ΔES and ΔEH−S, which suggests that the results presented here could be easily extended to layered transition-metal dichalcogenides in general. The compensation effect between S−metal binding and S− adsorbate binding at the edge enables the description of adsorption onto sulfur and metal sites using a single parameter from the local electronic structure of the metal. Although we have focused on the edge sites of layered TMS in this study, the compensation effect could be a much more general phenomenon, and there are potential implications for other structures of sulfides as well as transition-metal oxides, nitrides, and phosphides. Having established the importance of the edge-most metal in determining adsorption at both S and defect sites, our results provide a general electronic structurebased guideline for understanding and designing the next generation of TMS catalysts by fine-tuning the adsorption of key reaction intermediates via modifications to the edge-most metal.

Figure 4. Adsorption energy of various key reaction intermediates ΔE as a function of the d-band center εd. H, O, OH, CHO, and COOH are adsorbed onto S sites at the stable S and H coverages, whereas SH, NH2, and NNH are adsorbed into a S defect at the stable coverage. Typical adsorption configurations for a stable structure of θS = 0.5 ML and θH = 0.75 ML are shown as examples (blue = Mo, yellow = S, white = H, red = O, gray = C, dark blue = N, green = edge-most metal).

■

COMPUTATIONAL SECTION Plane-wave density functional theory (DFT) with ultrasoft pseudopotentials as implemented in the Quantum ESPRESSO code,45 were used to calculate all structures. All calculations were done using the BEEF-vdW exchange−correlation functional.46−49 A plane-wave cutoff and density cutoff of 500 and 5000 eV, respectively, were used, and the structures were relaxed until all force components were less than 0.05 eV/Å. Further details including the calculated lattice parameters are summarized in the Supporting Information.

indicates that the edge-most metal plays a central role in determining the trends in the overall activity. There is noticeably more scatter in the linear relations compared to the d-band model30 and adsorption scaling relations42 for transition metals. One reason for this scatter is that the adsorbates often induce significant reorganization of the S atoms at the edge. This generally involves repulsive interactions between neighboring S atoms as well as the breaking of a S−S bond if the adsorbate is bonded to a S dimer.40,41 Adsorption at defect sites leads to the greatest amount of edge rearrangement, which explains the larger scatter for the SH, NNH, and NH2 data points in Figure 4. The scatter also increases with the size of the molecules as various adsorption configurations exist at each site. In light of these considerations, a mean absolute error of 0.23 eV appears reasonable. The model could be further refined by incorporating other electronic effects, as we have demonstrated recently.43,44 It has also been suggested previously that the dyz states and the dx2−y2 states are especially important for adsorption on MoS2 edges,32 but we find that the weighted centers for each of these quantities scale linearly with εd (Supporting Information), which suggests that shifts in both the dyz states or dx2−y2 states are adequately captured by εd. Although we have only demonstrated the descriptive capability of εd at one set of operating conditions, it is expected to work

■

ASSOCIATED CONTENT

S Supporting Information *

Computational details, lattice parameters, definition of adsorption energies, description of the method for determining stable structures, and a summary of stable structures. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge support from the Center on Nanostructuring for Efficient Energy Conversion (CNEEC) at Stanford University, an Energy Frontier Research Center 3887

dx.doi.org/10.1021/jz5020532 | J. Phys. Chem. Lett. 2014, 5, 3884−3889

The Journal of Physical Chemistry Letters

Letter

(19) Füchtbauer, H. G.; Tuxen, A. K.; Moses, P. G. Morphology and atomic-scale structure of single-layer WS2 nanoclusters. Phys. Chem. Chem. Phys. 2013, 15, 15971−15980. (20) Wang, H.; Kong, D.; Johanes, P.; Cha, J. J.; Zheng, G.; Yan, K.; Liu, N.; Cui, Y. MoSe2 and WSe2 nanofilms with vertically aligned molecular layers on curved and rough surfaces. Nano Lett. 2013, 13, 3426−3433. (21) Tang, H.; Dou, K.; Kaun, C. C.; Kuang, Q. MoSe2 nanosheets and their graphene hybrids: Synthesis, characterization and hydrogen evolution reaction studies. J. Mater. Chem. A 2014, 2, 360−364. (22) Tsai, C.; Chan, K.; Abild-Pedersen, F.; Nørskov, J. K. Active edge sites in MoSe2 and WSe2 catalysts for the hydrogen evolution reaction: A density functional study. Phys. Chem. Chem. Phys. 2014, 16, 13156−13164. (23) Moses, P. G.; Hinnemann, B.; Topsøe, H.; Nørskov, J. K. The effect of Co-promotion on MoS2 catalysts for hydrodesulfurization of thiophene: A density functional study. J. Catal. 2009, 268, 201−208. (24) Merki, D.; Vrubel, H.; Rovelli, L.; Fierro, S.; Hu, X. Fe, Co, and Ni ions promote the catalytic activity of amorphous molybdenum sulfide films for hydrogen evolution. Chem. Sci. 2012, 3, 2515−2525. (25) Tsai, C.; Chan, K.; Nørskov, J. K.; Abild-Pedersen, F. Rational design of MoS2 catalysts: Tuning the structure and activity via transition metal doping. Catal. Sci. Technol. 2014, DOI: 10.1039/ C4CY01162G. (26) Nørskov, J. K.; Bligaard, T.; Rossmeisl, J.; Christensen, C. H. Towards the computational design of solid catalysts. Nat. Chem. 2009, 1, 37−46. (27) Nørskov, J. K.; Abild-Pedersen, F.; Studt, F.; Bligaard, T. Surface chemistry special feature: Density functional theory in surface chemistry and catalysis. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 937− 943. (28) Nørskov, J. K.; Bligaard, T.; Logadóttir, Á .; Bahn, S.; Hansen, L. B.; Bollinger, M. V.; Bengaard, H.; Hammer, B.; Sljivancanin, Z.; Mavrikakis, M. Universality in heterogeneous catalysis. J. Catal. 2002, 209, 275−278. (29) Bligaard, T.; Nørskov, J. K.; Dahl, S.; Matthiesen, J.; Christensen, C. H.; Sehested, J. The Brønsted−Evans−Polanyi relation and the volcano curve in heterogeneous catalysis. J. Catal. 2004, 224, 206−217. (30) Hammer, B.; Nørskov, J. K. Electronic factors determining the reactivity of metal surfaces. Surf. Sci. 1995, 343, 211−220. (31) Vojvodic, A.; Hellman, A.; Ruberto, C.; Lundqvist, B. I. From electronic structure to catalytic activity: A single descriptor for adsorption and reactivity on transition-metal carbides. Phys. Rev. Lett. 2009, 103, 146103. (32) Raybaud, P.; Hafner, J.; Kresse, G.; Kasztelan, S.; Toulhoat, H. Structure, energetics, and electronic properties of the surface of a promoted MoS2 catalyst: An ab initio local density functional study. J. Catal. 2000, 190, 128−143. (33) Schweiger, H.; Raybaud, P.; Kresse, G.; Toulhoat, H. Shape and edge sites modifications of MoS2 catalytic nanoparticles induced by working conditions: A theoretical study. J. Catal. 2002, 207, 76−87. (34) Trasatti, S. Work function, electronegativity, and electrochemical behaviour of metals. J. Electroanal. Chem. 1971, 33, 351−378. (35) Nørskov, J. K.; Bligaard, T.; Logadóttir, Á .; Kitchin, J. R.; Chen, J. G.; Pandelov, S.; Stimming, U. Trends in the exchange current for hydrogen evolution. J. Electrochem. Soc. 2005, 152, J23−J26. (36) Man, I. C.; Su, H. Y.; Calle Vallejo, F.; Hansen, H. A.; Martínez, J. I.; Inoglu, N. G.; Kitchin, J. R.; Jaramillo, T. F.; Nørskov, J. K.; Rossmeisl, J. Universality in oxygen evolution electrocatalysis on oxide surfaces. ChemCatChem 2011, 3, 1159−1165. (37) Peterson, A. A.; Nørskov, J. K. Activity descriptors for CO2 electroreduction to methane on transition-metal catalysts. J. Phys. Chem. Lett. 2012, 3, 251−258. (38) Skúlason, E.; Bligaard, T.; Gudmundsdóttir, S.; Studt, F.; Rossmeisl, J.; Abild-Pedersen, F.; Vegge, T.; Jónsson, H.; Nørskov, J. K. A theoretical evaluation of possible transition metal electro-catalysts for N2reduction. Phys. Chem. Chem. Phys. 2012, 14, 1235−1245.

funded by the U.S. Department of Energy, Office of Basic Energy Sciences under Award Number DE-SC0001060. J.K.N. and F.A-P. acknowledge financial support from the U.S. Department of Energy, Office of Basic Energy Sciences to the SUNCAT Center for Interface Science and Catalysis. C.T. acknowledges support from the National Science Foundation GRFP Grant DGE-114747, and K.C. acknowledges support from the Air Force Office of Scientific Research through the MURI program under AFOSR Award No. FA9550-10-1-0572.

■

REFERENCES

(1) Chhowalla, M.; Shin, H. S.; Eda, G.; Li, L.-J.; Loh, K. P.; Zhang, H. The chemistry of two-dimensional layered transition metal dichalcogenide nanosheets. Nat. Chem. 2013, 5, 263−275. (2) Topsøe, N.-Y.; Topsøe, H. Characterization of the structures and active sites in sulfided Co−Mo/Al2O3 and Ni−Mo/Al2O3 catalysts by NO chemisorption. J. Catal. 1983, 84, 386−401. (3) Harris, S.; Chianelli, R. R. Catalysis by transition metal sulfides: The relation between calculated electronic trends and HDS activity. J. Catal. 1984, 86, 400−412. (4) Harris, S.; Chianelli, R. R. Catalysis by transition metal sulfides: A theoretical and experimental study of the relation between the synergic systems and the binary transition metal sulfides. J. Catal. 1986, 98, 17− 31. (5) Topsøe, H.; Clausen, B. S. Active sites and support effects in hydrodesulfurization catalysts. Appl. Catal. 1986, 25, 273−293. (6) Ma, X.; Sakanishi, K.; Isoda, T.; Mochida, I. Hydrodesulfurization reactivities of narrow-cut fractions in a gas oil. Ind. Eng. Chem. Res. 1995, 34, 748−754. (7) Vasudevan, P. T.; Fierro, J. A review of deep hydrodesulfurization catalysis. Catal. Rev. 1996, 38, 161−188. (8) Hensen, E. J. M.; Brans, H. J. A.; Lardinois, G. M. H. J.; de Beer, V. H. J.; van Veen, J. A. R.; van Santen, R. A. Periodic trends in hydrotreating catalysis: Thiophene hydrodesulfurization over carbonsupported 4d transition metal sulfides. J. Catal. 2000, 192, 98−107. (9) Shi, X.-R.; Jiao, H.; Hermann, K.; Wang, J. CO hydrogenation reaction on sulfided molybdenum catalysts. J. Mol. Catal. A: Chem. 2009, 312, 7−17. (10) Hinnemann, B.; Moses, P. G.; Bonde, J. L.; Jørgensen, K. P.; Nielsen, J. H.; Horch, S.; Chorkendorff, I.; Nørskov, J. K. Biomimetic hydrogen evolution: MoS2 nanoparticles as catalyst for hydrogen evolution. J. Am. Chem. Soc. 2005, 127, 5308−5309. (11) Jaramillo, T. F.; Jørgensen, K. P.; Bonde, J. L.; Nielsen, J. H.; Horch, S.; Chorkendorff, I. Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts. Science 2007, 317, 100−102. (12) Bonde, J. L.; Moses, P. G.; Jaramillo, T. F.; Nørskov, J. K.; Chorkendorff, I. Hydrogen evolution on nano-particulate transition metal sulfides. Faraday Discuss. 2008, 140, 219. (13) Kong, D.; Wang, H.; Cha, J. J.; Pasta, M.; Koski, K. J.; Yao, J.; Cui, Y. Synthesis of MoS2 and MoSe2 films with vertically aligned layers. Nano Lett. 2013, 13, 1341−1347. (14) Chan, K.; Tsai, C.; Hansen, H. A.; Nørskov, J. K. Molybdenum sulfide and selenides as possible electrocatalysts for CO2reduction. ChemCatChem 2014, 6, 1899−1905. (15) Asadi, M.; Kumar, B.; Behranginia, A.; Rosen, B. A.; Baskin, A.; Repnin, N.; Pisasale, D.; Phillips, P.; Zhu, W.; Haasch, R. Robust carbon dioxide reduction on molybdenum disulphide edges. Nat. Commun. 2014, 5, 4470. (16) Byskov, L. S.; Nørskov, J. K.; Clausen, B. S.; Topsøe, H. Edge termination of MoS2 and CoMoS catalyst particles. Catal. Lett. 2000, 64, 95−99. (17) Bollinger, M. V.; Lauritsen, J. V.; Jacobsen, K. W.; Nørskov, J. K.; Helveg, S.; Besenbacher, F. One-dimensional metallic edge states in MoS2. Phys. Rev. Lett. 2001, 87, 196803. (18) Bollinger, M. V.; Jacobsen, K. W.; Nørskov, J. K. Atomic and electronic structure of MoS2 nanoparticles. Phys. Rev. B 2003, 67, 085410. 3888

dx.doi.org/10.1021/jz5020532 | J. Phys. Chem. Lett. 2014, 5, 3884−3889

The Journal of Physical Chemistry Letters

Letter

(39) Peterson, A. A.; Abild-Pedersen, F.; Studt, F.; Rossmeisl, J.; Nørskov, J. K. How copper catalyzes the electroreduction of carbon dioxide into hydrocarbon fuels. Energy Environ. Sci. 2010, 3, 1311. (40) Byskov, L. S.; Bollinger, M. V.; Nørskov, J. K.; Clausen, B. S.; Topsøe, H. Molecular aspects of the H2 activation on MoS2 based catalysts  The role of dynamic surface arrangements. J. Mol. Catal. A: Chem. 2000, 163, 117−122. (41) Tsai, C.; Abild-Pedersen, F.; Nørskov, J. K. Tuning the MoS2 edge-site activity for hydrogen evolution via support interactions. Nano Lett. 2014, 14, 1381−1387. (42) Abild-Pedersen, F.; Greeley, J.; Studt, F.; Rossmeisl, J.; Munter, T. R.; Moses, P. G.; Skúlason, E.; Bligaard, T.; Nørskov, J. K. Scaling properties of adsorption energies for hydrogen-containing molecules on transition-metal surfaces. Phys. Rev. Lett. 2007, 99, 016105. (43) Vojvodic, A.; Nørskov, J. K.; Abild-Pedersen, F. Electronic structure effects in transition metal surface chemistry. Top. Catal. 2013, 57, 25−32. (44) Xin, H.; Vojvodic, A.; Voss, J.; Nørskov, J. K.; Abild-Pedersen, F. Effects of d-band shape on the surface reactivity of transition-metal alloys. Phys. Rev. B 2014, 89, 115114. (45) Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; et al. Quantum ESPRESSO: A modular and open-source software project for quantum simulations of materials. J. Phys.: Condens. Matter 2009, 21, 395502. (46) Wellendorff, J.; Lundgaard, K. T.; Møgelhøj, A.; Petzold, V.; Landis, D. D.; Nørskov, J. K.; Bligaard, T.; Jacobsen, K. W. Density functionals for surface science: Exchange−correlation model development with Bayesian error estimation. Phys. Rev. B 2012, 85, 235149. (47) Dion, M.; Rydberg, H.; Schröder, E.; Langreth, D. C.; Lundqvist, B. I. van der Waals density functional for general geometries. Phys. Rev. Lett. 2004, 92, 246401. (48) Thonhauser, T.; Cooper, V. R.; Li, S.; Puzder, A.; Hyldgaard, P.; Langreth, D. C. Van der Waals density functional: Self-consistent potential and the nature of the van der Waals bond. Phys. Rev. B 2007, 76, 125112. (49) Román-Péréz, G.; Soler, J. M. Efficient implementation of a van der Waals density functional: Application to double-wall carbon nanotubes. Phys. Rev. Lett. 2009, 103, 096102.

3889

dx.doi.org/10.1021/jz5020532 | J. Phys. Chem. Lett. 2014, 5, 3884−3889