Coupling Ternary Selenide SnSb2Se4 with Graphene Nanosheets for High‐Performance Potassium‐Ion Batteries

Although chalcogenide anodes possess higher potassium storage capacity than intercalated‐based graphite, their drastic volume change and the irreversible electrochemical reactions still hinder the effective electron/ion transfer during the potassiation/depotassiation process. To solve the above problems, this article proposes the synthesis of a lamellar nanostructure where graphene nanosheets are embedded with SnSb2Se4 nanoparticles (SnSb2Se4/GNS). In the product, fine monodisperse SnSb2Se4 nanoparticles are coupled with graphene nanosheets to form a porous network framework, which can effectively mitigate the drastic volume changes during electrode reactions and guarantee efficient potassium‐ion storage through the synergistic interactions among multiple elements. Various electrochemical analyses prove that SnSb2Se4 inherits the advantages of the binary Sb2Se3 and SnSe while avoiding their disadvantages, confirming the synergistic effect of the ternary–chalcogenide system. When tested for potassium storage, the obtained composite delivers a high specific capacity of 368.5 mAh g−1 at 100 mA g−1 and a stable cycle performance of 265.8 mAh g−1 at 500 mA g−1 over 500 cycles. Additionally, the potassium iron hexacyanoferrate cathode and the SnSb2Se4/GNS anode are paired to fabricate the potassium‐ion full cell, which shows excellent cyclic stability. In conclusion, this strategy employs atomic doping and interface interaction, which provides new insights for the design of high‐rate electrode materials.


Introduction
Potassium-ion batteries (PIBs) are receiving increasing attention and are considered as the ideal substitute for lithium-ion batteries due to their low cost, abundant potassium source, and low redox potential of K/K + (À2.93 vs SHE). [1][2][3] Only with long cycle life and highpower density, PIBs can really be used for large-scale energy storage. However, these remain key challenges for electrode materials, especially anode materials. [4] The potassium ions (K + ), with a larger ionic radius (1.38 A), show poor kinetic characteristics in the charging and discharging processes, resulting in unsatisfactory electrochemical performance. [5][6][7][8][9] Therefore, proper control of the conductivity and structural stability of anode materials is the key to optimizing the PIBs.
A variety of metal oxide/chalcogenide compounds have excellent physical, chemical, and electronic properties due to their diversity of composition, adjustable crystal structure, and special morphology and architecture. Therefore, they have been considered as promising PIB anode materials and widely studied. Among them, sulfides (e.g., SnS, Bi 2 S 3 , and Sb 2 S 3 ) and selenides (e.g., SnSe, Bi 2 Se 3 , and Sb 2 Se 3 ) have been extensively explored owing to their synergistic conversion and alloying reactions that can ensure higher theoretical capacities than intercalated anodes (e.g., graphite). [10][11][12][13][14][15] Moreover, Se has higher conductivity (1 9 10 À3 S m À1 ) than S (5 9 10 À28 S m À1 ), indicating that its energy barrier of ion diffusion is lower than that of oxide and sulfide. [16] At the same time, metal selenides, with high theoretical capacities and superior redox reversibility, are attractive anode materials for PIBs and have been intensively investigated. [17] Nonetheless, substantial volume expansion occurs during potassiation/ depotassiation for metal selenides, leading to severe pulverization and rapid capacity decline. [18] In order to solve the above problems, carbon modification and the multi-element combination can be used to improve the electrochemical performance of selenides as PIB anode materials. [19][20][21] For example, Yi et al. [19] reported that layered SnSSe anodes ensure the high capacity and superior cycling stability of PIBs due to the increased interlayer spacing, the partial substitution of S by Se, and the improvement of the material's electrical conductivity with the help of the multilayer graphene sheets.
Therefore, in this work, a lamellar nanostructure where graphene nanosheets are embedded with SnSb 2 Se 4 nanoparticles (denoted SnSb 2 Se 4 /GNS) is engineered. In the product, fine monodisperse SnSb 2 Se 4 nanoparticles are coupled with graphene nanosheets to form a porous network framework that can effectively mitigate the drastic volume changes during electrode reactions and guarantee efficient potassium-ion storage through the synergistic interactions among multiple elements. Various electrochemical analyses prove that SnSb 2 Se 4 inherits the advantages of the binary Sb 2 Se 3 and SnSe while avoiding their disadvantages, confirming the synergistic effect of the ternarychalcogenide system. The composite exhibits impressive cycling DOI: 10.1002/eem2.12617 Although chalcogenide anodes possess higher potassium storage capacity than intercalated-based graphite, their drastic volume change and the irreversible electrochemical reactions still hinder the effective electron/ion transfer during the potassiation/depotassiation process. To solve the above problems, this article proposes the synthesis of a lamellar nanostructure where graphene nanosheets are embedded with SnSb 2 Se 4 nanoparticles (SnSb 2 Se 4 /GNS). In the product, fine monodisperse SnSb 2 Se 4 nanoparticles are coupled with graphene nanosheets to form a porous network framework, which can effectively mitigate the drastic volume changes during electrode reactions and guarantee efficient potassium-ion storage through the synergistic interactions among multiple elements. Various electrochemical analyses prove that SnSb 2 Se 4 inherits the advantages of the binary Sb 2 Se 3 and SnSe while avoiding their disadvantages, confirming the synergistic effect of the ternary-chalcogenide system. When tested for potassium storage, the obtained composite delivers a high specific capacity of 368.5 mAh g À1 at 100 mA g À1 and a stable cycle performance of 265.8 mAh g À1 at 500 mA g À1 over 500 cycles. Additionally, the potassium iron hexacyanoferrate cathode and the SnSb 2 Se 4 /GNS anode are paired to fabricate the potassium-ion full cell, which shows excellent cyclic stability. In conclusion, this strategy employs atomic doping and interface interaction, which provides new insights for the design of high-rate electrode materials. performance and maintains 265.8 mAh g À1 over 500 cycles at a high current density of 500 mA g À1 . Hopefully, the results of this research can provide guidance for the design and manufacture of multi-element alloy-based anodes.

Results and Discussion
A straightforward high-energy ball milling method was initially employed to produce pure SnSb 2 Se 4 with comparatively large size. The  distribution of elements Sn, Sb, and Se in SnSb 2 Se 4 is homogeneous, as illustrated in Figures S1 and S2, Supporting Information, which is consistent with the initial molar ratio. Subsequently, SnSb 2 Se 4 was pulverized by high-energy ball milling and the expanded graphite was put under argon to obtain the terminal product, SnSb 2 Se 4 anchored on graphene nanosheets (SnSb 2 Se 4 /GNS) ( Figure 1). For comparison, Sb 2 Se 3 /GNS and SnSe/GNS were prepared via the same synthesis process ( Figures S3 and S4, Supporting Information). The field-emission scanning electron microscope (SEM) image of SnSb 2 Se 4 /GNS is shown in Figure 2a, which clearly illustrates the lamellar stacking microstructure of the product. According to the transmission electron microscope (TEM) image in Figure 2b and the high-resolution TEM (HRTEM) image in Figure 2c, the SnSb 2 Se 4 nanoparticles are evenly embedded between the graphene nanosheets, which can effectively improve the conductivity of the composite material and buffer the volume expansion during the discharge-charge process. Moreover, the HRTEM image in Figure 2d shows a set of parallel stripes with a d spacing of 0.287 nm, which corresponds to the (621) plane of SnSb 2 Se 4 . The element mapping images in Figure 2e confirm the even distribution of all elements. Figure 3a depicts the X-Ray diffraction (XRD) patterns of SnSb 2 Se 4 and SnSb 2 Se 4 /GNS. Compared with SnSb 2 Se 4 , the XRD pattern of SnSb 2 Se 4 /GNS exhibits a reduced peak intensity but the same diffraction peak position with an additional peak between 26°and 27°, which can be ascribed to the (002) plane of GNS. [22] In addition, Raman spectroscopy was employed to analyze the degree of graphitization of the composite. Two peaks at 1351 and 1580 cm À1 , which correspond to disordered carbon (D-band) and ordered carbon (G-band), respectively, can be seen in the Raman spectra of the SnSb 2 Se 4 /GNS ( Figure 3b). These two bands can reflect the number of structural defects and the degree of graphitization. [23][24][25] The outcomes demonstrate that following ball milling, the graphite structure becomes more disordered. Based on the thermogravimetric analysis (TGA) curves in Figure 3c, the carbon content of the GNS-containing SnSb 2 Se 4 can be determined to be 22.1 wt%. Figure 3d-f and Figure S5, Supporting Information show the chemical states of SnSb 2 Se 4 /GNS surface elements, which are illustrated by X-ray photoelectron spectroscopy (XPS) spectra. In comparison with pure SnSb 2 Se 4 , the peaks of Sn, Sb, and Se in SnSb 2 Se 4 /GNS are dramatically displaced, bringing about higher binding energy, which implies the reduced electron density of Sn, Sb, and Se. [26][27][28][29][30] Clearly, this also indicates the strong electron interaction between SnSb 2 Se 4 and GNS, which contributes to the charge redistribution at the coupled interface and further demonstrates the successful formation of SnSb 2 Se 4 /GNS. This agrees with the XRD and HRTEM results above.
The electrochemical performance of SnSb 2 Se 4 /GNS and its counterparts, SnSe/GNS and Sb 2 Se 3 /GNS, was evaluated by the selected 100 mA g À1 charge-discharge test, as shown in Figure 4a-c and Figure S6, Supporting Information. According to the figures, the Sb 2 Se 3 /GNS delivers a discharge capacity of 787.9 mAh g À1 at the beginning of the cycle and an initial Coulombic efficiency (ICE) of 63.1%. It maintains 323.7 mAh g À1 after the 50th cycle but diminishes to 218 mAh g À1 after the 100th cycle, creating a rapid capacity decay of 32.7% in 50 cycles (Figure 4a). The discharge capacity of SnSe/GNS is likewise unstable after 50 cycles (Figure 4b). Nevertheless, compared with the erratic performance of SnSe/GNS and Sb 2 Se 3 /GNS, SnSb 2 Se 4 /GNS exhibits a notable performance improvement as displayed in Figure 4c. The SnSb 2 Se 4 /GNS electrode has an ICE of 69.5% and becomes rather stable after 50 cycles, with discharge-charge curves at the 50th and the 100th cycles nearly overlapping. Apart from its superior stability, the SnSb 2 Se 4 /GNS electrode also provides a lower working voltage for the full battery. In conclusion, compared to SnSe/ GNS and Sb 2 Se 3 /GNS, SnSb 2 Se 4 /GNS has a higher ICE, greater stability, and more appropriate discharge platforms.
To investigate the potassiation/depotassiation process of the products, cyclic voltammetry (CV) scans were performed on SnSb 2 Se 4 /GNS, SnSe/GNS, and Sb 2 Se 3 /GNS. Figure 4d displays the first three CV scans of SnSb 2 Se 4 /GNS. In the first cathode scan, the prominent peak at 0.89 V refers to the formation of the solid electrolyte interphase (SEI) layer, and the peak at 0.38 V denotes the alloy reaction between K and Sn-Sb alloy, which is slightly skewed due to the formation of SEI. In the subsequent scanning process, the cathode peaks shift to 1.39 and 0.53 V and then maintain their position and intensity, indicating the stability of the SEI film after the first scanning. The peak at 1.39 V can be attributed to the conversion reaction, which is obscured by a strong peak from SEI during the first scan, and the peak at 0.53 V can be ascribed to the alloying reaction between K and Sn-Sb. In the anodic scan, the peaks at 1.47 and 2.3 V correspond to the depotassiation reaction. Furthermore, CV curves of Sb 2 Se 3 /GNS and SnSe/GNS are shown in Figure S7, Supporting Information. Their peak positions are similar to those of SnSb 2 Se 4 /GNS. After the initial scan, however, the cathodic and anodic peaks of SnSb 2 Se 4 /GNS are more intense than those of Sb 2 Se 3 /GNS and SnSe/GNS. This means that the discharge-charge platform of SnSb 2 Se 4 /GNS is necessarily wider than those of Sb 2 Se 3 /GNS and SnSe/GNS, which will be further verified below.
The discharge capacities of SnSb 2 Se 4 /GNS, Sb 2 Se 3 /GNS, and SnSe/ GNS at different current densities are shown in Figure 4e. At 50 mA g À1 , the discharge capacity of SnSb 2 Se 4 /GNS can reach 460.2 mAh g À1 , which is 11.9% higher than Sb 2 Se 3 /GNS and 35.2% higher than SnSe/GNS, indicating the superior electrical conductivity of the material. When the current density of the SnSb 2 Se 4 /GNS electrode increases to 50, 100, 200, 500, 1000, 2000, and 4000 mA g À1 , the discharge capacity decreases to 408.5, 380.9, 328.7, 279.6, 221.7, and 124.1 mAh g À1 , respectively. Even at a high current density of 6000 mA g À1 , the SnSb 2 Se 4 /GNS electrode maintains a high capacity of 74.3 mAh g À1 (Figure 4e and Figure S8, Supporting Information). Moreover, after undergoing the high current density, SnSb 2 Se 4 /GNS maintains a stable discharge capacity of about 380.2 mAh g À1 when the current density drops to 50 mA g À1 , indicating its excellent rate capability. Generally, the SnSb 2 Se 4 /GNS electrode shows significantly optimized rate performance. The long-term cycling performance of the electrode at 500 mA g À1 is revealed in Figure 4f and Figure S9, Supporting Information. After 500 cycles, SnSb 2 Se 4 /GNS reaches a high charging capacity of 265.8 mAh g À1 , achieving a capacity retention of 81.3% compared with the second cycle. It also shows superior rate capability in comparison with the anodes of many reported PIBs (Figure 4g and Table S1, Supporting Information). [31][32][33][34][35][36][37][38][39][40][41][42][43] In conclusion, the ternary synergistic effect can effectively improve the rate performance and long-term cycle stability of SnSb 2 Se 4 /GNS.
To gain insight into the electrode process dynamics of SnSb 2 Se 4 / GNS, galvanostatic intermittent titration technique (GITT) was performed. As shown in Figure 5a-d and Figure S10, Supporting Information, the diffusion coefficient of K + of the prepared ternary compound during the discharging and charging processes is above 10 À8 cm 2 s À1 , which is larger than that of other compounds. Electrochemical impedance spectroscopy (EIS) results (Figure 5e,f and Table S2, Supporting Information) further show that SnSb 2 Se 4 /GNS has the smallest semicircle, the smallest charge transfer resistance, and the highest slope among the three compounds, which confirm the excellent conductivity consistent with GITT. These results prove that the synergistic effect in the ternary compound significantly enhances the K + diffusion kinetics. Based on these advantages, the SnSb 2 Se 4 /GNS composite shows excellent rate property at different current densities, as shown in Figure 4.
In addition, we explored the interaction among SnSe, Sb 2 Se 3 , SnSb 2 Se 4 , and GNS, respectively, using density functional theory (DFT) calculations ( Figure 6) to understand why ternary compounds can enhance battery performance. The model systems in our calculations are based on SnSe, Sb 2 Se 3 , SnSb 2 Se 4 unit cells, and single-layer graphene. The optimized results of the interaction among SnSe, Sb 2 Se 3 , SnSb 2 Se 4 , and GNS are shown in Figure 6. As we can see, the results show that the adsorption energy between SnSb 2 Se 4 and GNS is À1.571 eV, which is higher than the other two selenides (Figure 6a-c). Furthermore, the distance between SnSb 2 Se 4 and GNS is 2.364 A, which is smaller than that of the other two selenides (Figure 6d-f). Firstprinciples calculations allow us to confirm that GNS has a stronger affinity for SnSb 2 Se 4 , which can significantly improve cyclic stability.
As shown in Figure 7b, ex situ XRD patterns of SnSb 2 Se 4 /GNS under different discharge-charge states (from states A to M in Figure 7a) were recorded to investigate the structural evolution mechanism of potassiation/depotassiation during the first cycle. According to the XRD value of the fresh electrode material, the peak at 31.06°can be attributed to SnSb 2 Se 4 and the one at 26.41°can be attributed to GNS. The characteristic peak of SnSb 2 Se 4 /GNS disappears when the battery is discharged from 3.0 to 0.01 V, indicating that the SnSb 2 Se 4 /GNS structure is damaged during the conversion and alloying process. At the same time, the signal peaks of K 2 Se and K 3 Sb gradually become more intense. [44][45][46] When the electrode is discharged to 1.5 V, the diffraction peaks of the XRD pattern correspond to K 4 Sn 23 , K 2 Se, and K 3 Sb, respectively, indicating that the conversion and alloy reactions between SnSb 2 Se 4 /GNS and K form K 2 Se, K 4 Sn 23 , and K 3 Sb. During the charging process, the signals of Sn, K 2 Se, and K 3 Sb gradually fade, and the peaks of Sn and K 3 Sb disappear at 3.0 V, indicating the occurrence of the depotassiation reaction. [47] Finally, the weakened characteristic peak of SnSb 2 Se 4 /GNS reappears after the composite is charged to 3.0 V, implying that the depotassiation of SnSb 2 Se 4 /GNS is partially reversible.
To further determine the potassiation/depotassiation process of SnSb 2 Se 4 /GNS, the XPS spectra of Sn 3d, Sb 3d, and Se 3d were obtained as shown in Figure 7c. After the electrode is discharged to 0.01 V, the peak of Se shifts to lower binding energy, probably due to the electron-donating effect of K + in the newly formed K 2 Se. [48] The reduction of the binding energy of Sn and Sb peaks indicates the alloying reaction among K, Sn, and Sb simultaneously. [49,50] When the charge voltage of the SnSb 2 Se 4 /GNS electrode reaches 3.0 V, the peaks of Sn 3d, Sb 3d, and Se 3d all shift to the left, indicating the reversibility of the electrochemical reaction. In the meantime, the structural evolution of SnSb 2 Se 4 /GNS was confirmed by ex situ TEM characterization (Figure 7d-f). It demonstrates the conversion-alloying mechanism of the SnSb 2 Se 4 /GNS electrode, which is consistent with the results of ex situ XRD. The element mapping image of the electrode in full discharge state after 50 cycles under 100 mA g À1 is shown in Figure S11, Supporting Information, which displays the uniform distribution of K, Sn, Sb, and Se.
Finally, to further study the application potential of the prepared SnSb 2 Se 4 /GNS, a potassium-ion full battery with SnSb 2 Se 4 /GNS as the anode and potassium iron hexacyanoferrate (KFeHCF) as the cathode was constructed and tested in a voltage window of 1.0-3.8 V. KFeHCF was synthesized via the coprecipitation method, [51] and its structural characterization is shown in Figure S12, Supporting Information. According to the charge-discharge curves, the KFeHCF cathode shows a specific capacity of 86.3 mAh g À1 at 50 mA g À1 when the operating voltage ranges from 2.0 to 4.0 V ( Figure S13, Supporting Information). Figure 8a illustrates the working mechanism of the full battery. K + is extracted from the KFeHCF cathode and inserted into the SnSb 2 Se 4 /GNS anode during the charging process, and the reverse process takes place during the discharging process. The representative charge-discharge curve of the whole battery is shown in Figure 8b and the SnSb 2 Se 4 / GNS//KFeHCF full battery is operated at 50 mA g À1 . Its rate performance is shown in Figure 8c,d. Figure 8c shows that the average discharge capacities of the SnSb 2 Se 4 /GNS//KFeHCF full battery are 163.1, 121.6, 107.2, 87.5, and 69.1 mAh g À1 when the current densities are 50, 100, 200, 500, and 1000 mA g À1 , respectively. Figure 8d displays the potassiation-depotassiation profiles of the SnSb 2 Se 4 /GNS//KFeHCF full battery at diverse current densities. The curves at different current densities are similar in shape, reflecting the outstanding reversibility of the full battery. After the rate performance test, the SnSb 2 Se 4 /GNS// KFeHCF full battery is cycled at 50 mA g À1 . It shows a discharge capacity of 160.1 mAh g À1 after 200 cycles, which confirms its excellent cycling and rate performance. These results indicate that SnSb 2 Se 4 /GNS is a promising anode material for PIBs.

Conclusion
According to the synthesis strategy proposed by this article, the ternary SnSb 2 Se 4 /GNS hybrid can be prepared via high-energy ball milling, thus obtaining the PIB anode with high efficiency. The phase evolution of SnSb 2 Se 4 /GNS during the discharge/charge process has been determined by ex situ XRD, XPS, and TEM analyses. The SnSb 2 Se 4 /GNS electrode delivers considerable specific capacity (368.5 mAh g À1 at 100 mA g À1 ), outstanding rate performance (156.2 mAh g À1 at 2000 mA g À1 ), and superior cycle performance (265.8 mAh g À1 after 500 cycles at 0.5 A g À1 ) when applied in potassium-ion half-cells. In addition, when paired with the KFeHCF cathode, the anode manifests excellent full-cell performance. With the help of the synergistic effect of the ternary chemicals, the performance of the compound which functions poorly in binary systems is also greatly improved. Hopefully, this research can help extend the concept of ternary or multi-element and facilitate the application of these elements in large-scale energy storage for future grids.

Experimental Section
Material synthesis: First, Sn, Sb, and Se, with a molar ratio of 1:2:4, were ground with a high-energy ball mill for 100 min in an Ar atmosphere to prepare the SnSb2Se4 powder. Second, the as-prepared SnSb2Se4 powder and the expanded graphite, with a weight ratio of 3:1, were further ground with a high-energy ball mill for 100 min under an Ar protection to obtain the SnSb2Se4/GNS composite. During the ball milling process, stainless steel balls with a diameter of 10 mm were used, and the ball-to-powder weight ratio was 20:1. For comparison, SnSe/ GNS and Sb 2 Se 3 /GNS were also synthesized in stoichiometric ratios under the same conditions, respectively.
Materials characterization: The morphology and structure of the samples were characterized by SEM (JSM-7600F) and TEM (FEI Talos F200X). XRD (Rigaku SmartLab) was also used to examine the structure of the samples. The cycled electrode was placed on a glass slide and protected by a Kapton film for ex situ XRD measurements. Raman spectrum was obtained on a Labram HR800 with a laser wavelength of 514 nm. Furthermore, the composition of the samples was analyzed quantitatively via TGA on a NETZSCH STA 449F3 thermal analyzer in airflow at a heating rate of 10°C min À1 to 800°C. XPS analysis was performed on an ESCALAB Xi+ electron spectrometer.
Electrochemical measurements: First, SnSb 2 Se 4 /GNS was mixed with Ketjen black and sodium carboxymethyl cellulose binder in deionized water with a mass ratio of 8:1:1 to obtain the electrode slurry. Then, the slurry was coated on a Cu foil and dried in a vacuum at 80°C for 12 h. The average mass loading of each electrode was about 1.0-1.5 mg cm À2 . Finally, the coin cells were assembled in an argon-filled glovebox (MBRAUN, Unilab1200), with potassium metal as the counter electrode and glass fiber (Whatman) as the separator. Meanwhile, a solution of 3 M potassium bis(fluorosulfonyl)imide (KFSI) in 1,2-dimethoxyethane was utilized as the electrolyte. To examine the electrochemical properties of the product, the galvanostatic discharging/charging surveys were conducted between 0.01 and 3.0 V versus K/K + on a Land CT2001A battery tester. The K + storage mechanism of the product was investigated by CV and EIS on a PARSTAT 4000 electrochemical workstation. Additionally, before GITT experiments, cells were initially activated for five cycles at 0.1 A g À1 and subsequently charged or discharged for 10 min at the pulse current of 0.1 A g À1 . They were then kept still for 20 min to reach the potassium equilibrium potential. All capacities of the product were calculated based on the mass of the active material. For the fabrication of the full cells, the SnSb 2 Se 4 /GNS anode was pre-activated in half batteries at 0.1 A g À1 before assembly. KFeHCF was prepared via the previously reported method, [51] and 1 M KFSI in dimethyl carbonate was used as the electrolyte. The full battery voltage range was set to 1.0-3.8 V.
Theoretical calculations: All DFT calculations within the generalized gradient approximation using the Perdew-Burke-Ernzerhof formula were performed Figure 7. a) Discharge/charge curves at 100 mA g À1 and b) ex situ XRD patterns of SnSb 2 Se 4 /GNS at varied statuses during the first cycle. XPS spectra of the SnSb 2 Se 4 /GNS electrode at different states: c) Sn 3d, Sb 3d, and Se 3d. HRTEM images of SnSb 2 Se 4 /GNS at different states: d) initial discharge to 0.6 V, e) initial discharge to 0.01 V, and f) initial charge to 3.0 V. Energy Environ. Mater. 2023, 6, e12617 7 of 9 using the Vienna Ab Initio Package. [52][53][54] The projected-enhanced wave potential was applied to describe ionic nuclei. Valence electrons were studied by using a plane-wave basis set with a kinetic energy cutoff of 520 eV. [55,56] Partial occupation of Kohn-Sham orbitals was allowed using a Gaussian smearing method and a width of 0.05 eV. Electronic energies were considered selfconsistent when the energy change was <10 À5 eV. The geometry optimization was considered converged when the energy change was <0.03 eV A À1 . The vacuum spacing was 20 A perpendicular to the plane of the structure. Brillouin zone integration utilized a 3 9 3 9 1 Monkhorst-Pack k-point sampled surface structure. Adsorption energies (E ads ) were calculated as E ads = E ad/ sub À E ad À E sub , where E ad/sub , E ad , and E sub were the total energies of the optimized adsorbate/substrate system, adsorbate in the structure, and clean substrate, respectively.