Mechanistic Study on Water Gas Shift Reaction on ... - ACS Publications

Article pubs.acs.org/JPCC

Mechanistic Study on Water Gas Shift Reaction on the Fe3O4 (111) Reconstructed Surface Liang Huang,†,‡ Bo Han,*,† Qingfan Zhang,† Maohong Fan,‡ and Hansong Cheng*,† †

Sustainable Energy Laboratory, Faculty of Material Science and Chemistry, China University of Geosciences (Wuhan), 388 Lumo Rd, Wuhan 430074, China ‡ Department of Chemical and Petroleum Engineering, University of Wyoming, Laramie, Wyoming 82071, United States S Supporting Information *

ABSTRACT: We present a first-principles study using periodic density functional theory on a water gas shift reaction on a Feoct2‑tet1terminated Fe3O4 (111) surface. We show that water can easily undergo dissociative adsorption to form OH and H adatom species on the surface. Three possible reaction mechanisms (i.e., redox mechanism, associative mechanism, and coupling mechanism) were systematically explored based on minimum energy path calculations. It was identified that the redox mechanism is the energetically most favorable pathway for the water gas shift reaction on the Feoct2‑tet1terminated Fe3O4 (111) surface. The COO* desorption was found to be the rate-limiting step with a barrier of 1.04 eV, and the OH dissociation has the second-highest activation barrier (0.81 eV). Our results are consistent with results of kinetic and isotope exchange experiments. Our studies suggest that it is necessary to develop a promoter to reduce the activation barriers of the COO* desorption and OH dissociation steps in order to improve the catalyst performance.

■

The associative mechanism was first proposed almost a century ago by Armstrong and Hiditch based on the study of WGSR on a copper chromite catalyst.23 It was suggested that CO and water first form various surface intermediates before the final products are produced:

INTRODUCTION Water gas shift reaction (WGSR, CO + H2O = CO2 + H2) is an important industrial process for hydrogen production in steam methane reform1−3 for ammonia and methanol synthesis4,5 as well as for direct PEM (proton exchange membrane) fuel cell applications.6,7 WGSR can be carried out in a wide temperature range, depending on catalysts and specific application conditions. In the temperature range of 350−500 °C, catalysts based on the oxides of iron and chromium (Fe2O3−Cr2O3) have been shown to be effective.8−12 In the low-temperature range of 190−250 °C, copper catalysts dispersed on ZnO and Al2O3 supports were found to be highly active.13−19 Catalysts for both high-temperature and low-temperature processes are commercially available and vary with targeted applications and product requirements. The shift converters can be either single stage or two-stage, depending on the requirement of hydrogen purity.20 For a two-stage process, an iron-based catalyst, such as Fe3O4/Cr2O3, is usually used at high temperature in the first stage.21,22 Upon cooling, the residual CO is shifted over a copper-based shift catalyst (e.g., Cu/ZnO/Al2O3) at a lowtemperature to react with the steam to produce high-purity hydrogen and CO2.11 The mechanisms of WGSR have long been a subject of extensive studies both experimentally and theoretically with the aim to develop more efficient catalysts for applications of interest. In general, two mechanisms have been proposed to dominate the shift reactions. One is an associative mechanism and another is a redox mechanism (regenerative mechanism). © 2015 American Chemical Society

CO + H 2O → (intermediates) → CO2 + H 2

The associative mechanism was found to be operative in most low-temperature water gas shift reactions by Burch.24 Subsequent studies25−27 identified that the reaction intermediates are formates, carbonyls, and carbonates. Indeed, a theoretical study by Gokhale et al.26 to investigate the lowtemperature WGSR on a Cu surface suggested that carboxyl and formate species are the main intermediates, which was verified by IR, HREELS, and EXAFS spectroscopic studies. Alternatively, the redox mechanism states that water first undergoes dissociation on catalyst surfaces to produce molecular hydrogen and an oxygen adsorbate followed by oxidation of CO by the oxygen adatom to form a CO2 molecule: H 2O + * → H 2 + O* CO + O* → CO2 + * Received: September 21, 2015 Revised: November 27, 2015 Published: December 2, 2015 28934

DOI: 10.1021/acs.jpcc.5b09192 J. Phys. Chem. C 2015, 119, 28934−28945

Article

The Journal of Physical Chemistry C where * represents the surface adsorption site, and O* is an adsorbed surface oxygen. Generally, conventional high-temperature shift catalysts contain roughly 80−90 wt % Fe2O3 and 8−10 wt % Cr2O3. To activate the catalyst, pretreatment is usually required by partially reducing the hematite (Fe2O3) to the magnetite (Fe3O4). In other words, the active phase of the catalyst is Fe3O4. The main function of Cr2O3 was deemed to prevent the catalyst from sintering and thus a loss of surface area of the iron oxide crystallites.21,22 Although tremendous effort has been made to understand the catalytic mechanisms, the detailed microscopic processes have still remained to be a subject of intense debates. Oki and co-workers28 conducted a series of isotope tracer experiments on WGSR over iron -based catalysts. The results were inconclusive as both the redox and the associative mechanisms were found to be consistent with experimental observations. Nevertheless, the transient experiments on WGSR over the industrial ferrochrome catalyst by Salmi et al.29,30 supported the redox mechanism. The subsequent experiments also found that the redox mechanism is consistent with the kinetic data.31 Recently, the redox mechanism was also confirmed to prevail over the associative mechanism for WGSR on Fe-based catalysts by in situ DRIFTS experiments.32 While extensive investigations on WGSR on Fe- and Cubased catalysts have been carried out, to the best of our knowledge, detailed mechanisms of WGSR on Fe3O4 surfaces have remained to be poorly understood. In light of the industrial importance of the WGSR process catalyzed by magnetite, it is essential to understand the elemental processes of the reactions on the catalyst surfaces at the atomistic scale in order to design more efficient catalysts or to optimize the current production process. In this paper, we present a firstprinciples study based on periodic density functional theory on WGSR on a Fe3O4 (111) surface. The elementary processes of the reaction along both the redox and associative pathways will be systematically examined. The objective of this study is to provide useful insight into the detailed reaction processes and to identify the key elementary steps that dictate the reaction rates.

Figure 1. Optimized structure of Feoct2‑tet1-terminated surface. (a) Side view and (b) top view.

All electronic structure calculations were carried out using density functional theory with the Perdew−Burke−Ernzerhof (PBE)41 exchange-correlation functional as implemented in the VASP code.42 The interactions between core electrons and ions were described by the projector augmented wave (PAW) method, and the valence electronic states were represented with a plane-wave basis set with the energy cutoff of 400 eV. A spinpolarization scheme was utilized to deal with the open shell electronic system intrinsic to magnetite surfaces. The Brillouin zone integration was performed using a Monkhorst−Pack grid of 3 × 3 × 1 special k-points. Geometry optimizations were carried out using the conjugate gradient algorithm. The computational mechanistic study requires extensive search for transition states (TS) for the prescribed reaction pathways. This was carried out by employing the “climbing images” nudged elastic band (CI-NEB) algorithm.43,44 Typically, six images were produced between the reactant and product states in each elementary step as the initial guess of reaction coordinates. Subsequently, each individual image was optimized based on the NEB algorithm. For calculations of the isolated molecules (CO and H2O), a cubic supercell with a lattice constant of 10 Å was used to avoid interactions between the periodic images, and a 2 × 2 × 2 k-points mesh was used for the Brillouin zone integration. Molecular adsorption energy, Eads, was calculated using the following equation:

■

SURFACE MODEL AND COMPUTATIONAL METHOD In a previous report, we studied water dissociative chemisorption on a Fe3O4 (111) surface,33 where a Feoct2-terminated slab model was employed to represent the surface. Fe3O4 (111) is the dominant natural growth facet,34 and the catalyst activity of Fe3O4 (111) is higher than that of other orientations.35,36 Among the six possible terminations of the Fe3O4 (111) surface, the Fetet1- and Feoct2‑tet1 (or Feoct2)-terminated surfaces were proposed to be the most stable.37,38 A previous STM experiment39 and several theoretical studies37,40 suggested that the Feoct2‑Fetet1-terminated surface is indeed energetically favored. Therefore, we chose the Feoct2‑tet1-terminated surface as the model for our mechanistic studies. The surface includes two exposed Fe atoms (Fetet1 and Feoct2) and one exposed O atom. In this study, the same surface model containing stoichiometrically eight layers of iron atoms and eight layers of oxygen atoms with a p (1 × 1) unit cell (see Figure 1) was utilized. The top eight layers of the surface were fully relaxed upon geometry optimization, and the bottom eight layers were kept fixed. The vacuum gap between the adjacent slabs was set to 13 Å to minimize the interaction between slabs.

Eads = Emol + Esur − Esur + mol

where Emol, Esur, and Esur+mol represent the energies of a gasphase molecule, the slab surface, and the molecule adsorbed on the surface, respectively. In order to analyze the influence of Hubbard U on the results, the adsorption energies of H2O on the Fe3O4 (111) surface with GGA and GGA+U (U − J = 4.0 eV)45 were calculated. With the use of the GGA+U method, the adsorption of H2O at the Feoct2 site was found to be slightly less exothermic by 0.04 eV with the adsorption energy of 1.52 eV; the calculated adsorption energy of H2O at the Fetet1 site is 0.72 eV, slightly increased from 0.70 eV obtained by the GGA method. Two typical reactions, OH dissociation and CO oxidation, along with the redox mechanism were also tested using the GGA+U method. It was found that the activation barrier for the OH dissociation is slightly decreased by 0.04 eV when the Hubbard U correction is included. For CO oxidation, 28935


Article

The Journal of Physical Chemistry C the energy difference is also marginal, only 0.01 eV. Our results suggest that the chosen theoretical method and surface model are sufficiently adequate without addition of the Hubbard U correction.

■

RESULTS AND DISCUSSION 1. Carbon Monoxide Adsorption. We first investigate the interaction of CO with the Fe3O4 (111) surface, which is an essential step of WGSR. As shown in Figure 1, there are only two possible adsorption sites for CO on the selected Fe3O4 (111) surface (i.e., on top of the Feoct2 atom and/or the Fetet1 atom). The relative adsorption strengths of CO on both sites were assessed by performing full geometry optimization with the C atom forming a covalent bond with the surface Fe atom. The optimized main structural parameters and adsorption energies of the adsorbed CO molecule on the two exposed Fe sites are labeled in Figure 2.

Figure 3. Calculated charge distribution and charge density difference for CO adsorption on the Fe3O4 (111) surface at the (a and c) Feoct2 site and the (b and d) Fetet1 site. The left panels (a and b) represent the bonding charge integrated in planes perpendicular to the surface and plotted as a function of the height from the surface. The right panels (c and d) display the charge density difference post and prior to CO adsorption at the value of 0.006 electrons/Å3. The yellow and the cyan regions represent the charge accumulation and depletion, respectively.

Figure 2. CO adsorption on the (a) Feoct2 site and the (b) Fetet1 site of the Fe3O4 (111) surface. The bond length unit is angstrom.

The calculated C−O bond distance of the gas phase molecule is 1.142 Å, in good agreement with the experimental value of 1.128 Å46 and the previously reported theoretical value.47 Upon adsorption on the Fe3O4(111) surface, the bond length is increased to 1.168 Å at the Feoct2 site and 1.173 Å at the Fetet1 site, respectively, modestly longer than the gas phase value. Since the C−O bond elongation upon adsorption results from the back-donation of d-electron to the antibonding π*orbital of CO, the molecule is somewhat activated on the oxide surface. The optimized C−Fe distances for the adsorption at the Feoct2 and Fetet1 sites are 1.811 and 1.825 Å, respectively. We note that CO adsorption on both sites results in surface relaxation to a certain extent by raising the Fe atoms up from the surface. The surface relaxation effect at the Fetet1 site is more pronounced than at the Feoct2 site as the Fetet1 atom is significantly raised from its original position by 0.745 Å upon CO adsorption. The calculated adsorption energies for CO at the Feoct2 site and the Fetet1 site are 1.90 and 1.30 eV, respectively. The calculated result is consistent with the previous DFT calculation,47 which reported an adsorption energy of 1.94 eV for CO on the Feoct2 site. Although the energetic preference is given to the Feoct2 site, it is likely both sites will be populated under a certain CO partial pressure as adsorption on these two sites is very strong. The charge density difference before and after CO adsorption on the Fe3O4 (111) surface was also calculated to analyze the charge transfer in the CO adsorption process. Figure 3 (panels a and b) display the charge density difference integrated in planes perpendicular to the surface, from which we can easily analyze the electron population distribution. The charge analysis clearly shows electron transfer from the Fe atom

to the C atoms, leading to the formation of C−Fe covalent bonds. While for the C−O bond, both charge accumulation and charge depletion take place, arising from the back-donation from the d-electrons of Fe atoms to the molecule and the forward donation of the 5σ-electrons of CO to the empty dorbitals of the Fe atoms, respectively. Our results indicate that the Fe atom transfers more charge to the C−Fe bond at the Feoct2 site than at the Fetet1 site (0.020 e− vs 0.010 e−), leading to stronger adsorption at the Feoct2 site. However, the charge transfer to the C−O bond at the Fetet1 site is slightly more than at the Feoct2 site (0.013 e− vs 0.009 e−), resulting in a modestly longer C−O bond than at the Feoct2 site. Figure 3 (panels c and d) display the calculated electron density difference defined by Δρ = ρCO + surf − ρsurf − ρCO

at the Feoct1 and Feoct2 sites, respectively, where ρCO+surf, ρsurf, and ρCO represent the electron densities of the surface with CO adsorption, the surface, and the gas-phase CO molecule. At the Feoct2 site, a considerable amount of electron density shifts away from the Feoct2 atom and a high electron density around the Fe−C bond is clearly visible, reflecting a strong adsorption strength. The adsorption gives rise to a relatively modest variation of the electron density around the binding site as the Fe atom is raised from the surface. At the Fetet1 site, a significant charge variation in the proximity of the adsorption center is observed. This is attributed to the considerable surface 28936


Article

The Journal of Physical Chemistry C relaxation upon CO adsorption, which gives rise to the superposition change of the calculated electron densities. The calculated density of states prior to and post CO adsorption is depicted in Figure 4. Figure 4a indicates that the

Figure 5. Optimized structures of water molecule adsorption at the Feoct2 site (a) and the Fetet1 site (b). (c) and (d) display the water dissociative chemisorption structure at the Fetet1 site and the associated potential energy curve of water dissociation, respectively.

bond in water that ultimately leads to dissociation of water molecules on the surface. The optimized local water dissociative adsorption structure at the Fetet1 site is displayed in Figure 5c. Upon dissociation at the Fetet1 site, one hydrogen atom is abstracted from the water molecule to form a surface hydroxyl species, leaving an OH species to be adsorbed at the bridge site between the Feoct2 and Fetet1 atoms. Our calculation shows that this process is moderately exothermic with a thermochemical energy of 0.95 eV. Our NEB calculation to locate a transition state of H2O dissociation at the Fetet1 site yields a reaction pathway with the potential energy declining smoothly and monotonically, indicating that the water dissociation at the Fetet1 site is facile. Compared to the adsorption strength of CO on the surface, water adsorption is significantly weaker on the Feoct2terminated Fe3O4 (111) surface. For both water and CO, the adsorption on the Feoct2 site is energetically more favorable than on the Fetet1 site, and both molecules are competing for these adsorption sites. From a thermodynamic point of view, there is virtually very little probability for these sites to be occupied by water. However, under the catalytic operating conditions, the partial pressure of water is much higher than that of CO.27 The CO gas must diffuse through the water vapor before it reaches the surface. Furthermore, water dissociative chemisorption is a prerequisite step for a WGS reaction. Therefore, it is reasonable to assume that before the CO molecules take over the adsorption sites, a sufficient amount of water molecules has already undergone adsorption and dissociation. 3. Activation Energy Barriers. To understand the reaction mechanisms on the Feoct2-terminated Fe3O4 (111) surface, coadsorption structures of CO and water need to be incorporated into the computational model. In the present study, we only consider the coadsorption structure with an OH group and a CO molecule. Table 1 outlines the reaction pathways considered for the WGS process on the Fe3O4 (111) surface. According to the redox mechanism, water completely dissociates into atomic H and O and the resulting O adatom reacts with the adsorbed CO to form a CO2 molecule. This mechanism was found to be valid

Figure 4. Calculated density of states of the (a) bare surface, (b) CO adsorption at the Feoct2 site, and (c) CO adsorption at the Fetet1 site.

magnetite is magnetic as expected. The d-electrons of iron dominate the low-lying states, which makes the surface highly sensitive to the adsorption of gas species. Indeed, CO adsorption gives rise to a significant electronic structural variation as is readily visible in Figure 4 (panels b and c). However, the change of the electronic structure is not only attributed to the adsorption itself but also to the considerable surface structural relaxation upon CO adsorption. Detailed analysis on the calculated partial DOS spectra suggests that CO withdraws charges from the d-orbitals of Fe to its antibonding π*-orbitals, consistent with the slightly elongated C−O bond length. Indeed, the Bader charge analysis reveals that there is 0.379e transferred from the surface to the adsorbed CO at the Fetet1 site and 0.322e transferred to the CO molecule at the Feoct2 site. 2. Water Dissociative Chemisorption. As shown in Figure 5 (panels a and b), both the Feoct2 atom and the Fetet1 atom can serve as the anchoring sites for water. For adsorption at the Feoct2 site, water molecules can spontaneously undergo dissociative chemisorption to form OH and H surface species upon geometry optimization as shown in Figure 5a, suggesting that the Feoct2 terminal surface is highly chemically active toward water adsorption. The calculated dissociative chemisorption energy is 1.56 eV. Water adsorption at the Fetet1 site was also considered as shown in Figure 5b. In contrast to the adsorption at the Feoct2 site, the water molecule is adsorbed nondissociatively at the Fetet1 site with an adsorption energy of 0.70 eV, indicating that the Feoct2 site is also more favorable than the Fetet1 site. At the Fetet1 site, the adsorption is dictated by the formation of a strong hydrogen bond between the surface O atom and an H atom in water and, to a lesser extent, a relatively weaker Fe−O bond with a bond distance of 2.230 Å. The O−H bond of water with the H atom attached to the closest surface O atom is elongated to 1.070 Å from the gasphase value of 0.973 Å. However, the formation of the hydrogen bond is responsible for the activation of the O−H 28937


Article

The Journal of Physical Chemistry C Table 1. Reaction Pathways for WGS on Feoct2-Terminated Fe3O4 (111) Surfacea

a

redox mechanism

associative mechanism

coupling mechanism

H2O(g) + * → H2O* H2O* + * → HO* + H* HO* + * → H* + O* CO(g) + * → CO* CO* + O* → CO2** CO2** → CO2 (g) + 2* H* + H* → H2 (g)

H2O(g) + * → H2O* H2O* + * → HO* + H* CO(g) + * → CO* HO* + CO* → COOH* + * COOH* + *→ CO2* + H* CO2* → CO2 (g) + * H* + H* → H2 (g)

H2O(g) + * → H2O* H2O* + * → HO* + H* CO(g) + * → CO* HO* + CO* → CO2* + H* CO2* → CO2 (g) + * H* + H* → H2 (g)

* indicates an adsorption site.

in both kinetic29,31 and isotope exchange experiments.48,49 In the case of the associative mechanism, water only partially dissociates and the adsorbed CO reacts with the surface OH species produced upon water dissociation to form a carboxyl intermediate, which subsequently decomposes into CO2 and an H atom. This mechanism was proposed to describe the lowtemperature shift on a copper surface by Gokhale et al.26 using density functional theory. The associative mechanism was also suggested to be the dominant pathway for the WGS reaction on surfaces of Cu and Pt catalysts in several experimental studies.50−52 In addition to the two pathways, we also propose a new route, named coupling pathway, on the Fe3O4 (111) surface. Here, the formation process of COOH* is skipped. Instead, a CO2* and a H* species are formed directly from the reaction between CO* and OH*. In this section, we describe the characteristics of the minimum energy path identified for each of the associated elementary steps (Table 1). 3.1. OH Dissociation, HO* + * → H* + O*. We first calculated the minimum energy pathway to break the O−H bond, which leads to formation of O and H adatoms on the surface. The calculated energy diagrams, along with the optimized structures, are shown in Figure 6. The configuration

Feoct2 site with a Fe−O bond length of 1.840 Å, and the distance between O and Fetet1 atom is reduced to 3.064 Å from 4.158 Å of the initial state structure. Upon moving to the bridge site, the OH bond tilts up with a bond angle of 64.6° and a bond length of 0.970 Å. The calculated distances of the Feoct2-O and Fetet1-O bonds are 1.992 and 2.048 Å, respectively. The calculated thermochemical energy is −0.22 eV. Upon the O−H bond dissociation, the H atom migrates further to the surface to combine with the nearby surface O atom to form a surface OH species with an optimized bond length of 0.978 Å, slightly longer than the value of a typical hydroxyl group. As a result of the O−H cleavage, the bond lengths of O−Feoct2 and O−Fetet1 become 1.865 and 1.873 Å, respectively. At the transition state, the O atom of the O−H bond is anchored at the bridge site with the O−Feoct2 distance of 1.906 Å and the O−Fetet1 distance of 1.985 Å, while the H atom points to the nearest surface O atom. The O−H bond is elongated to 1.985 Å, and the distance between H and the surface O atom is shortened to 1.153 Å, which falls into the typical hydrogen-bonding range. The O−H dissociation is moderately endothermic with the calculated reaction energy of 0.51 eV. The NEB calculation gives an activation energy of 0.81 eV. These results suggest that direct dissociation of the OH group on the Feoct2-terminated surface is energetically difficult. The dissociated OH species yields O and H adatoms located at the bridge and oxygen site, respectively, giving rise to a highly exposed O atom, which is chemically active (Figure 6). 3.2. CO Oxidation by Atomic O and CO2 Desorption. Upon OH dissociation, an oxygen atom is left on the bridge site to oxidize the CO molecule adsorbed on top of the adjacent Fe atoms. Here, we study the recombination reaction of the CO molecule and the O adatom. Two different initial structures were considered: one for CO adsorbed at the Feoct2 site and another for CO adsorbed at the Fetet1 site. 3.2.1. CO Oxidation at the Feoct2 Site. Since the preadsorbed O atoms may depress the surface activity, CO prefers to tilt to the surface with a lower adsorption energy of 1.05 eV than the value on the surface without O adsorption (1.90 eV). Upon CO adsorption, the calculated bond lengths of C−O and C−Fe are 1.160 and 1.786 Å, respectively. The separation between the C atom and the O adatom is 2.610 Å. The CO oxidation reaction results in formation of a carbonate intermediate structure as the CO gradually leans toward the O adatom (Figure 7). At the transition state, the C−O bond is essentially perpendicular to the surface with a distance of 1.161 Å. The calculated C−Fe bond length is 1.792 Å. The distance between the C atom and the O adatom is shortened to 2.181 Å and continues decreasing after going through the transition state to form a COO* species. The new C−O bond is almost parallel to the surface with a bond length of 1.258 Å. The formation of the surface COO* species requires a modest

Figure 6. OH dissociation on the Feoct2-terminated surface. The left part of the curve describes the migration process of the OH species and the right part describes the OH dissociation process. For clarity, the energy of the structure with OH adsorption at the Feoct2 site is set to zero.

of the reactant was obtained from the water dissociative chemisorption structure at the Feoct2 site. Initially, a hydroxyl species is adsorbed at the Feoct2 site with the Fe−O and O−H bond distances of 1.808 and 0.972 Å, respectively. The O−H bond was found to be nearly parallel to the surface, with a long distance between the H atom and the nearby surface O atom. Subsequently, the hydroxyl species overcomes a small barrier of 0.07 eV to move to the bridge site between Fetet1 and Feoct2. At the transition state, the OH species remains essentially at the 28938


Article

The Journal of Physical Chemistry C

C−O bond is elongated to 1.226 Å. The configuration of the COO* species is shown in Figure 8 with a slightly longer C−O bond of 1.281 Å than the value at the Feoct2 site (1.258 Å). The activation barrier of 0.32 eV for the combination between CO and the O adatom to form the COO* species is quite modest and is 0.21 eV lower than the value for CO oxidation at the Feoct2 site. Desorption of the COO* to form a CO2 molecule requires an activation barrier of 1.05 eV, and the process is moderately endothermic with a thermalchemical energy of 0.81 eV, similar to the process at the Feoct2 site. Summarizing the calculated reaction energies and barriers of all the elementary reactions of the redox mechanism, we found that two processes exhibit relatively high energy barriers: OH dissociation (0.81 eV) and COO* desorption (1.04/1.05 eV). We thus conclude that desorption of CO2 from the surface is the rate-limiting step along the reaction route. 3.3. COOH Formation and Decomposition. Once the OH species is formed on the surface through the dissociation of water, it may directly react with the adsorbed CO to form COOH*. The process was named associative mechanism.53,54 In this section, we discuss the possible reaction pathways of COOH formation and its subsequent decomposition. This process was found to be critical for the WGS reaction on the Cu(111)55 and the Pt(111)52 surfaces in the associative reaction pathway. The OH adsorption configuration was obtained from the water dissociation step, in which the OH species occupies the Feoct2 site. For the coadsorption of CO and OH, there are two possible adsorption sites. One is for CO and OH to be coadsorbed at the Feoct2 site, and the other is for the two species to be separated with CO landed on the Fetet1 site. 3.3.1. CO and OH Coadsorption at Feoct2 site. We consider the case of CO and OH coadsorption at the Feoct2 site. The calculated energy profiles and the optimized local structures are shown in Figure 9. Initially, the CO and OH surface species are

Figure 7. Combination of CO and O to form COO* at the Feoct2 site and COO desorption. The left part of the curve describes the combination of CO and O to form COO, and the right part describes the desorption of CO2 from the surface. For clarity, the energy of the CO and O coadsorption structure is set to zero.

barrier of 0.53 eV and is essentially thermal neutral with the calculated reaction energy of −0.09 eV. The calculated energy variation along the reaction pathway is shown in Figure 7. The results indicate that the combination between CO and the O adatom on the surface is both thermochemically and kinetically facile. Figure 7 suggests that desorption of CO2 from the surface is considerably endothermic with a thermochemical energy of 0.83 eV. Search of a transition state yields a structure with a virtually broken C−Fe bond of 3.249 Å. The two C−O bonds are of similar lengths of 1.177 and 1.175 Å, close to the gas phase value. The calculated activation energy is 1.04 eV, higher than the barriers for other elementary steps. Therefore, CO2 desorption is identified to be the rate-limiting step for the redox mechanism. 3.2.2. CO Oxidation at Fetet1 Site. The process starting from CO adsorption at the Fetet1 site was also investigated, and the calculated energy profile is shown in Figure 8. Similar to the

Figure 9. Energy profile of the recombination of CO and OH at the Feoct2 site and the decomposition of COOH. The left part of the curve describes the recombination of CO and OH, and the right part describes the decomposition of COOH. For clarity, the energy of the structure of the CO and OH coadsorption at the Feoct2 site is set to zero.

Figure 8. Combination of CO and O to form COO* at the Fetet1 site and COO desorption. The left part of the curve describes the combination of CO and O to form COO, and the right part describes the desorption of CO2 from the surface. For clarity, the energy of the CO and O coadsorption structure is set to zero.

coadsorbed at the Feoct2 site with a distance of 2.505 Å between the O atom of OH and the C atom of CO. A COOH intermediate is then formed at the Feoct2 site with the C−Fe bond distance of 1.873 Å and the OH bond pointing to the surface. The reaction goes through a transition state, where the O atom of the OH species is 2.465 Å from the Feoct2 atom and 1.435 Å from the C atom, respectively. The reaction is considerably endothermic with the calculated thermochemical

case at the Feoct2 site, initially, the CO molecule also tilts from the surface normal. This step is an exothermic process with a thermochemical energy of 0.64 eV. Upon reaching the transition state, the CO molecule becomes perpendicular to the surface with the calculated distances of the C−Fe and C−O bonds of 1.845 and 1.172 Å, respectively. The length of the C− O bond is shortened to 1.281 Å, while the length of the original 28939


Article

The Journal of Physical Chemistry C energy of 0.57 eV. The calculated activation barrier is 1.24 eV, suggesting that COOH formation is a difficult process. We next consider the decomposition of the COOH* species, which includes two possible routes. The first one is direct decomposition, in which H is transferred to the surface oxygen site to form an O−H species and the COO* species is anchored at the Feoct2 site (Figure 9). Alternatively, in the second route, the COOH* first rotates around the C−Fe bond to reorient the O−H bond toward the Fetet1 site; subsequently, the H atom is transferred to the Fetet1 site, forming a H−Fe hydride species on the surface (Figure 10).

Figure 11. Energy profile of COO* desorption from the Feoct2 site. The left part of the curve describes the conformational change of COO* species, and the right part describes the desorption of COO*. For clarity, the energy of the COO* and H coadsorption structure is set to zero.

is an endothermic process with a reaction energy of 0.57 eV and an activation barrier of 0.79 eV. 3.3.2. CO on the Fetet1 Site and OH on the Feoct2 Site. COOH formation may also take place on the Fetet1 site starting from the CO and OH coadsorption at the Fetet1 and Feoct2 sites, respectively, as depicted in Figure 12. The two species are Figure 10. Energy profile of COOH decomposition and CO2 desorption from the Feoct2 site. The left part of the curve describes the reorientation of the COOH* species, and the right part describes the decomposition of COOH. For clarity, the energy of the adsorbed trans-COOH structure is set to zero.

At the transition state of the direct decomposition route (Figure 9), the O atom combines with the Feoct2 atom and the O−H bond becomes slightly closer to the surface with a bond length of 0.978 Å. The decomposition of COOH results in an essentially broken O−H bond with a distance of 1.818 Å and a COO* species anchored at the Feoct2 site. This process is quite exothermic (−0.75 eV) with a very small activation barrier of 0.17 eV, indicating that the COOH dissociation is facile. In the second route (Figure 10), the COOH* species reorients itself around the C−Fe bond to form an intermediate with one C−O bond forming a triangle with the underneath Fe atom and another C−O bond forming a 6-membered ring structure with the surface atoms to minimize the activation barrier. The structural variation incurs a small activation energy of 0.2 eV and is nearly thermal neutral. This orientation allows the COOH* species to deliver the H atom to the Fetet1 atom to form a surface hydride species upon CO2 desorption, which incurs an activation energy of 0.89 eV required to break up the C−Fe and O−Fe bonds. As a result, the desorption of CO2 becomes modestly endothermic with a thermochemical energy of 0.17 eV. Figure 11 displays the calculated energy profile and structural variation of COO* desorption following the direct decomposition of COOH* species. Note that the distance between the surface hydrogen atom and an O atom of COO* in the initial state is about 1.818 Å, which falls into the conventional hydrogen bond range. We thus divide the CO2 desorption process into two steps. The COO* species first goes through a conformational change to break away from the hydrogen bond and subsequently desorbs from the Feoct2 site. The first step is slightly endothermic with a reaction energy of 0.17 eV and a moderate activation barrier of 0.54 eV. Again, CO2 desorption

Figure 12. Energy profile of COOH formation and decomposition from the Fetet1 site. The left part of the curve describes the recombination of CO and OH, and the right part describes the decomposition of COOH. For clarity, the energy of the adsorbed trans-COOH structure is set to zero.

adsorbed vertically on the surface separated by 3.279 Å. A COOH* intermediate is then formed at the Fetet1 site with the Fetet1−C distance of 2.066 Å and the O−Feoct2 distance of 1.990 Å. The reaction is a moderately endothermic process (0.45 eV) with a substantial activation barrier of 1.44 eV, suggesting that the occurrence of the process is energetically difficult compared to other reaction routes that lead to formation of the COOH surface species with smaller activation energies. At the transition state, the OH species breaks the bond with the Feoct2 atom and shifts nearly completely to the CO species on the Fetet1 site, which requires a high activation energy. The decomposition of COOH* is also energetically challenging as shown in Figure 12. The process is moderately exothermic, but the activation barrier of 1.08 eV is again substantial. The transition state structure requires breakup of the 5-membered ring formed by one of the C−O bond with the Fetet1−O−Feoct2 surface atoms, leading to a slightly elongated O−H bond of 1.007 Å with the H atom interacting with a 28940


Article


shown in Figure 13b. The CO and OH species are coadsorbed with a distance of 2.505 Å between the C atom of CO and the O atom of OH. The OH bond points to the surface oxygen atom with a distance of 2.117 Å. The NEB calculation gives an activation barrier of 1.02 eV, nearly the same as the value in the first case. At the transition state, the O−H bond is essentially broken, leaving an O atom adsorbed on the Feoct2 atom with a distance of 2.390 Å from the C atom. This reaction is again thermal neutral with a reaction energy of −0.01 eV. The decomposition of the COO* species on surface follows the same steps as in the associative mechanism (Figure 10) with an activation barrier of 0.79 eV. 3.5. H2 Production. As one of the main products of WGSR, hydrogen is produced via recombination of H atoms on the surface. A previous study suggests that the coadsorption structure with one H atom anchored at the Feoct2 site and another H atom bonded with a surface O atom in the proximity is energetically most stable.56 We thus only consider the coadsorption structure for H2 production. A key step is to investigate whether it is energetically feasible for a H atom to be adsorbed on the Feoct2 atom. Two possible processes leading to the hydride formation were investigated. One involves H migration from an adjacent surface O atom to the Feoct2 atom and another is derived from water dissociative chemisorption. 3.5.1. H Migration. There are three possible adsorption sites for hydrogen atom: a surface oxygen site, the Feoct2 site, and the bridge site between Feoct2 and Fetet1. Figure 14 depicts the

surface O atom via H bonding. As a consequence, a COO* species and H adatom are formed on the surface. Desorption of the COO* species is similar to what is shown in Figure 8 (i.e., CO2 desorption is thermochemically very endothermic with a high activation barrier). Summarizing the calculated results on the elementary steps following the associative mechanism, we identified that the lowest energy reaction pathway is (1) coadsorption of CO and OH species at the same Feoct2 site, (2) combination of the CO and OH species to form a COOH intermediate, (3) decomposition of the COOH species to form COO* and H*, and (4) desorption of CO2 (with a moderate activation barrier of 0.79 eV). In the associative mechanism, the formation of COOH* was found to be the rate-limiting step with an activation barrier of 1.24 eV. 3.4. Simultaneous CO Oxidation and OH Dissociation. In addition to the redox mechanism and the associative mechanism for WGSR on the Fe3O4 (111) surface, here we propose a coupling mechanism to describe the reaction phenomenon. We first investigate the case of OH and CO coadsorption at Feoct2 and Fetet1, separately, to form COO* and H* species. The calculated energy profile and the structures are shown in Figure 13a. In the initial configuration, the hydroxyl species is

Figure 13. Energy profile of the coupling reaction between OH* and CO*. (a) CO and OH coadsorption at different Fe sites and (b) CO and OH coadsorption at the Feoct2 site. The energy of the coadsorbed CO* and OH* structure is set to zero. Figure 14. Energy profile of hydrogen migration on the Fe3O4 (111) surface. The energy of the structure with H adsorbed at surface oxygen site is set to zero.

adsorbed at the Feoct2 site with a bond length of 0.973 Å. A CO molecule is adsorbed at the Fetet1 site nearby with a distance of 3.366 Å between the C atom in CO and the O atom of the OH species. Subsequently, the adsorbed CO and OH species lean to each other to form a transition state as the distance between the C atom and the O atom is reduced to 1.809 Å. The O−H bond points to a surface O atom with a slightly elongated bond length of 0.998 Å; the C−O bond is also stretched to 1.180 Å. The calculated activation energy is 1.01 eV, slightly smaller than the value of COOH formation of 1.24 eV (Associative mechanism) but higher than the value of OH dissociation of 0.81 eV (redox mechanism). In the final state, the O−H bond becomes fully dissociated and the H atom lands on a nearby surface O atom. This reaction is essentially thermoneutral with a reaction energy of −0.01 eV. Desorption of the COO* species is similar to what was shown in Figure 8 (i.e., CO2 desorption is thermochemically very endothermic with a high activation barrier). Alternatively, a coadsorption structure of CO and OH at the same Feoct2 site was examined to form COO* and H* species, and the calculated energy profile and structural variation are

minimum energy pathway for an H atom on the surface oxygen site to migrate to the neighboring bridge site, and then to move to the atop site of the adjacent Feoct2 atom, and finally to reach another surface oxygen site. The hydrogen atom clearly prefers the Fe site with ∼0.15 eV lower in energy than at the surface oxygen site. H migration from the O site to the bridge site requires an activation energy of 0.78 eV. However, from the bridge site to the atop site of Feoct2, the barrier is only 0.25 eV. Finally, migration from the Feoct2 site to the surface oxygen atom results in a considerable energy barrier of 1.03 eV, which arises from the long migration distance between the O site and the Fe site. Table 2 displays the charge variation of the H atom along the migration paths. Clearly, the charges on the hydrogen atom are negative at both the bridge and the Feoct2 sites. In contrast, the charge becomes positive once the atom migrates to the oxygen site. At the transition states of hydrogen migration from the O site to the bridge site and from the Feoct2 site to the surface O 28941


Article

The Journal of Physical Chemistry C Table 2. Calculated Bader Charge of the H Atom along the Migration Pathways Bader charge of H atom (e−) reactant transition state product

O → bridge

bridge → Feoct2

Feoct2 → O

0.709 −0.055 −0.464

−0.464 −0.389 −0.347

−0.347 0.001 0.709

atom, hydrogen remains essentially in its atomic state, which is unstable and thus requires to overcome high-energy barriers. 3.5.2. Hydride Formation via Water Dissociation. We have shown that a water molecule can readily undergo dissociative chemisorption on the Fe3O4 (111) surface, with one hydrogen atom firmly anchored to the surface oxygen. If most of the surface oxygen atoms are bonded with hydrogen, the water dissociation process may be partially obstructed. Hence we consider two different surface iron sites, Feoct2 and Fetet1, for water adsorption with surface oxygen atoms occupied by hydrogen. The calculated energy profile and the structures of the reactive species are shown in Figure 15.

Figure 16. Energy profile of H2 formation from two H adatoms. The energy of the structure of the coadsorbed H atoms is set to zero.

Initially, the distance between the coadsorbed H atoms is 1.628 Å with the calculated Fe−H and O−H bond lengths of 1.667 and 0.977 Å, respectively. Subsequently, the two H atoms approach to each other to form a new H−H bond. At the transition state, the distance between the two atoms is shortened to 0.914 Å. A hydrogen molecule is then formed and adsorbed at the Feoct2 site (Figure 16). This process was found to be moderately endothermic by 0.15 eV with a calculated activation barrier of 0.49 eV. The last step is the hydrogen desorption from the surface. The desorption process is endothermic with a thermodynamic energy of 0.47 eV and an activation barrier of 0.65 eV. Both the recombination and desorption processes were found to be endothermic with moderate activation barriers, suggesting that hydrogen formation is relatively facile in WGSR. 4. Discussion. Figure 17 summarizes the calculated thermochemical energies and the associated activation barriers

Figure 15. Energy profile of the dissociation process of H2O. (a) H2O adsorption at the Feoct2 site and (b) H2O adsorption at the Fetet1 site. The energy of the coadsorbed H2O* and H* structure is set to zero.

We first explore the dissociation process starting from water adsorption at the Feoct2 site. The results are shown in Figure 15a. Initially, the molecule is adsorbed at the Feoct2 site with a Fe−O bond of 2.131 Å. One of the OH bonds points to the nearby Fetet1 atom with a distance of 2.469 Å, and the OH bond is slightly elongated to 1.004 Å. The other bond points away from the surface with a bond length of 0.977 Å. Upon dissociation, the water molecule is decomposed into a surface OH species and H atom. The OH species occupies the Feoct2 site, and the H atom is located at the Fetet1 site with a Fe−H bond distance of 1.528 Å. This reaction is a slightly endothermic process (0.24 eV) with a moderate activation barrier of 0.44 eV. We next investigate water dissociation at the Fetet1 site. The calculated energy profile and the local structures of the reactive species are shown in Figure 15b. This reaction starts from water adsorption at the Fetet1 site. Upon dissociation, the hydrogen atom moves to the Feoct2 site, and the OH species is located at the bridge site between the two surface Fe atoms. The reaction is highly exothermic with the calculated thermodynamic energy of −0.85 eV. The calculated activation barrier of 0.43 eV is also moderate. Our results suggest that the H atom may also be adsorbed on the Fe atoms upon water dissociation. 3.5.3. Hydrogen Recombination and Desorption. We now discuss hydrogen recombination to form H2 on the surface and its subsequent desorption. The results are shown in Figure 16.

Figure 17. Overall energy profiles for WGSR on the Fe3O4 (111) surface along the redox, the associative, and coupling reaction pathways.

of the elementary steps of the WGS reaction for the three proposed reaction pathways. The WGS reaction may undergo the redox, the associative, and the coupling routes individually or simultaneously. The three pathways, represented by black, blue, and red lines, respectively, undergo the same adsorption processes of CO and H2O followed by water dissociation. The main difference lies in the O−H dissociation step. After the COO* and H* species are produced, the three pathways are merged into the same surface state. The last step for all the three pathways is the desorption of CO2 and H2 from the surface. 28942


Article


abstracted from the water molecule, (2) an associative mechanism, in which a COOH intermediate species is initially formed and subsequently decomposed into COO* and H*, and (3) a coupling mechanism, in which a COO* species is directly formed through the surface rearrangement of the OH and CO species without going through OH dissociation and COOH formation. The last step for all the three mechanistic processes is the desorption of CO2 and H2. The results indicate that the redox mechanism is energetically the most favorable pathway for the water gas shift reaction on the Fe3O4 (111) surface with the rate-limiting step being the desorption of CO2. The associated activation barrier of the ratelimiting step was calculated to be 1.04 eV, which is good agreement with the reported experimental result. The OH dissociation was found to exhibit the second highest activation barrier of 0.81 eV along the redox pathway. In order to further improve the catalyst performance, an appropriate promoter is needed to reduce the activation barrier for the desorption of CO2 and the dissociation of the OH species.

Clearly, for the redox mechanism (black), the highest barrier is 1.04 eV. We thus conclude that the rate-limiting step is the CO2 desorption step. For the associative mechanism (red), the rate-limiting step is the formation of the COOH intermediate via CO and OH recombination with an activation barrier of 1.24 eV. In the coupling mechanism (blue), there is a smaller barrier for the COO* formation than the value found in the associative mechanism, but this barrier is higher than the value in the redox mechanism. Similar to the redox process, CO2 desorption was also found to be the rate-limiting step for the coupling pathway. Overall, these results indicate that the redox mechanism is the dominate pathway for the WGS reaction on Fe3O4 surface with CO2 desorption as the rate-limiting step. The calculated rate constants ln k for the desorption of CO2 and the OH dissociation steps at different temperatures (from 275 to 1000 K) are shown in Table S1 and Figure S1. The analysis57 of the rate constants reveals that the CO2 desorption step is the rate-limiting step for the redox mechanism. Many previous studies focus mostly on the associative mechanism for the low-temperature stage of WGSR,26,52,58 and to the best of our knowledge, there has been essentially no report about the associative mechanism for an iron oxide based catalyst. It is important to note that the isotope exchange experiments28,59 indicate that CO → CO (ads) and 2H* → H2 are the slow steps for the WGSR over the iron oxide catalysts. The redox mechanism proposed by Salmi and co-workers29,30 suggests that CO + O* → CO2*, CO2* → CO2 + *, H2O + 2* → 2H* + O*, and 2H* → H2 + 2* are all the slow steps in the reaction over the iron oxide catalysts. Our computational study identified that the desorption of carbon dioxide and hydrogen and the dissociation of OH species are the slow steps. In particular, the desorption of CO2 was found to be the ratelimiting step for the redox mechanism. In contrast to the conclusions drawn by Salmi and co-workers, the carbon monoxide reaction with oxygen to form carbon dioxide was found to be a fast step with a moderate activation barrier of 0.53 eV and a thermochemical energy of −0.09 eV. The effective activation barrier for the WGS reaction over the iron oxide catalysts has also been measured in numerous experimental studies with a value ranging from 1.04 to 1.55 eV.11,20,31 The highest barrier of 1.04 eV found for the redox mechanism certainly falls into the range of the reported experimental values.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b09192. Kinetic analysis for the desorption of CO2 and OH dissociation steps and rate constants for the desorption of CO2 and OH dissociation steps at different temperatures on the Fe3O4 (111) surface graph and table (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge support of the research by the National Natural Science Foundation of China (Grants 21473164, 21203169, and 21233006), the Fundamental Research Funds for the Central Universities, China University of Geosciences, and Air Products and Chemicals, Inc.

■

■

CONCLUSION Water gas shift reaction (WGSR, CO + H2O = CO2 + H2) is an important industrial process for hydrogen production. Catalysts based on the oxides of iron and chromium (Fe2O3−Cr2O3) have been shown to be effective in the high-temperature shift process. In the catalyst working condition, the active phase was found to be Fe3O4. Although tremendous effort has been made to understand the catalytic mechanisms, the detailed microscopic processes have still been a subject of intense debates. In this paper, the mechanisms of water gas shift reaction on the Fe3O4 (111) surface with the Feoct2‑tet1 terminal have been studied using the first-principles calculations. It was found that Fe3O4 (111) surface terminated by Feoct2 displays high activity for water adsorption. Subsequently, the water molecule undergoes nearly spontaneous dissociation into the OH and H atom. Upon the water dissociation, a CO molecule reacts with the surface OH species. Three different reaction pathways were then systematically explored: (1) a redox mechanism, in which CO is oxidized by the oxygen atom

REFERENCES

(1) Abbas, H. F.; Wan Daud, W. M. A. Hydrogen Production by Methane Decomposition: A Review. Int. J. Hydrogen Energy 2010, 35, 1160−1190. (2) Armor, J. N. The Multiple Roles for Catalysis in the Production of H2. Appl. Catal., A 1999, 176, 159−176. (3) Vanhook, J. P. Methane-Steam Reforming. Catal. Rev.: Sci. Eng. 1980, 21, 1−51. (4) Shelef, M.; Gandhi, H. S. Ammonia Formation in the Catalytic Reduction of Nitric Oxide. III. The Role of Water Gas Shift, Reduction by Hydrocarbons, and Steam Reforming. Ind. Eng. Chem. Prod. Res. Dev. 1974, 13, 80−85. (5) Klier, K. Methanol Synthesis. Adv. Catal. 1982, 31, 243−313. (6) Galvita, V.; Schroder, T.; Munder, B.; Sundmacher, K. Production of Hydrogen with Low COx-Content for PEM Fuel Cells by Cyclic Water Gas Shift Reactor. Int. J. Hydrogen Energy 2008, 33, 1354−1360. (7) Ruettinger, W.; Ilinich, O.; Farrauto, R. J. A New Generation of Water Gas Shift Catalysts for Fuel Cell Applications. J. Power Sources 2003, 118, 61−65. 28943


Article

The Journal of Physical Chemistry C (8) Lee, D. W.; Lee, M. S.; Lee, J. Y.; Kim, S.; Eom, H. J.; Moon, D. J.; Lee, K. Y. The Review of Cr-Free Fe-Based Catalysts for HighTemperature Water-Gas Shift Reactions. Catal. Today 2013, 210, 2−9. (9) Jeong, D. W.; Subramanian, V.; Shim, J. O.; Jang, W. J.; Seo, Y. C.; Roh, H. S.; Gu, J. H.; Lim, Y. T. High-Temperature Water Gas Shift Reaction over Fe/Al/Cu Oxide Based Catalysts Using Simulated Waste-Derived Synthesis Gas. Catal. Lett. 2013, 143, 438−444. (10) Martos, C.; Dufour, J.; Ruiz, A. Synthesis of Fe3O4-Based Catalysts for the High-Temperature Water Gas Shift Reaction. Int. J. Hydrogen Energy 2009, 34, 4475−4481. (11) Rhodes, C.; Peter Williams, B.; King, F.; Hutchings, G. J. Promotion of Fe3O4/Cr2O3 High Temperature Water Gas Shift Catalyst. Catal. Commun. 2002, 3, 381−384. (12) Atwood, K.; Arnold, M. R. Activity of an Iron Oxide-Chromium Oxide Water-Gas Shift Catalyst. Ind. Eng. Chem. 1953, 45, 424−426. (13) Rasmussen, D. B.; Janssens, T. V. W.; Temel, B.; Bligaard, T.; Hinnemann, B.; Helveg, S.; Sehested, J. The Energies of Formation and Mobilities of Cu Surface Species on Cu and ZnO in Methanol and Water Gas Shift Atmospheres Studied by DFT. J. Catal. 2012, 293, 205−214. (14) Kowalik, P.; Prochniak, W.; Borowiecki, T. The Effect of Alkali Metals Doping on Properties of Cu/ZnO/Al2O3 Catalyst for Water Gas Shift. Catal. Today 2011, 176, 144−148. (15) Atake, I.; Nishida, K.; Li, D.; Shishido, T.; Oumi, Y.; Sano, T.; Takehira, K. Catalytic Behavior of Ternary Cu/ZnO/Al2O3 Systems Prepared by Homogeneous Precipitation in Water-Gas Shift Reaction. J. Mol. Catal. A: Chem. 2007, 275, 130−138. (16) Shishido, T.; Yamamoto, M.; Li, D.; Tian, Y.; Morioka, H.; Honda, M.; Sano, T.; Takehira, K. Water-Gas Shift Reaction over Cu/ ZnO and Cu/ZnO/Al2O3 Catalysts Prepared by Homogeneous Precipitation. Appl. Catal., A 2006, 303, 62−71. (17) Ayastuy, J. L.; Gutierrez-Ortiz, M. A.; Gonzalez-Marcos, J. A.; Aranzabal, A.; Gonzalez-Velasco, J. R. Kinetics of the Low-Temperature WGS Reaction over a CuO/ZnO/Al2O3 Catalyst. Ind. Eng. Chem. Res. 2005, 44, 41−50. (18) van Herwijnen, T.; Guczalski, R. T.; de Jong, W. A. Kinetics and Mechanism of the CO Shift on CuZnO: II. Kinetics of the Decomposition of Formic Acid. J. Catal. 1980, 63, 94−101. (19) van Herwijnen, T.; Jong, W. A. d. Kinetics and Mechanism of the CO Shift on Cu/ZnO: I. Kinetics of the Forward and Reverse CO Shift Reactions. J. Catal. 1980, 63, 83−93. (20) Newsome, D. S. The Water-Gas Shift Reaction. Catal. Rev.: Sci. Eng. 1980, 21, 275−318. (21) Kung, M. C.; Kung, H. H. The Surface Cation Densities of Iron Oxide-Chromium Oxide Solid Solutions. Surf. Sci. 1981, 104, 253− 269. (22) Keiski, R. L.; Salmi, T. Deactivation of the High-Temperature Water-Gas Shift Catalyst in Nonisothermal Conditions. Appl. Catal., A 1992, 87, 185−203. (23) Armstrong, E. F.; Hilditch, T. P. A Study of Catalytic Actions at Solid Surfaces. IV. The Interaction of Carbon Monoxide and Steam as Conditioned by Iron Oxide and by Copper. Proc. R. Soc. London, Ser. A 1920, 97, 265−273. (24) Burch, R. Gold Catalysts for Pure Hydrogen Production in the Water-Gas Shift Reaction: Activity, Structure and Reaction Mechanism. Phys. Chem. Chem. Phys. 2006, 8, 5483−5500. (25) Goguet, A.; Meunier, F. C.; Tibiletti, D.; Breen, J. P.; Burch, R. Spectrokinetic Investigation of Reverse Water-Gas-Shift Reaction Intermediates over a Pt/CeO2 Catalyst. J. Phys. Chem. B 2004, 108, 20240−20246. (26) Gokhale, A. A.; Dumesic, J. A.; Mavrikakis, M. On the Mechanism of Low-Temperature Water Gas Shift Reaction on Copper. J. Am. Chem. Soc. 2008, 130, 1402−1414. (27) Ovesen, C. V.; Clausen, B. S.; Hammershoi, B. S.; Steffensen, G.; Askgaard, T.; Chorkendorff, I.; Norskov, J. K.; Rasmussen, P. B.; Stoltze, P.; Taylor, P. Microkinetic Analysis of the Water-Gas Shift Reaction under Industrial Conditions. J. Catal. 1996, 158, 170−180.

(28) Oki, S.; Mezaki, R. Identification of Rate-Controlling Steps for the Water-Gas Shift Reaction over an Iron Oxide Catalyst. J. Phys. Chem. 1973, 77, 447−452. (29) Salmi, T.; Boström, S.; Lindfors, L.-E. A Dynamic Study of the Water-Gas Shift Reaction over an Industrial Ferrochrome Catalyst. J. Catal. 1988, 112, 345−356. (30) Salmi, T.; Lindfors, L. E.; Boström, S. Modelling of the High Temperature Water Gas Shift Reaction with Stationary and Transient Experiments. Chem. Eng. Sci. 1986, 41, 929−936. (31) Keiski, R. L.; Salmi, T.; Niemisto, P.; Ainassaari, J.; Pohjola, V. J. Stationary and Transient Kinetics of the High Temperature Water-Gas Shift Reaction. Appl. Catal., A 1996, 137, 349−370. (32) Natesakhawat, S.; Wang, X. Q.; Zhang, L. Z.; Ozkan, U. S. Development of Chromium-Free Iron-Based Catalysts for HighTemperature Water-Gas Shift Reaction. J. Mol. Catal. A: Chem. 2006, 260, 82−94. (33) Zhou, C.; Zhang, Q.; Chen, L.; Han, B.; Ni, G.; Wu, J.; Garg, D.; Cheng, H. Density Functional Theory Study of Water Dissociative Chemisorption on the Fe3O4(111) Surface. J. Phys. Chem. C 2010, 114, 21405−21410. (34) Yu, X.; Huo, C.-F.; Li, Y.-W.; Wang, J.; Jiao, H. Fe3O4 Surface Electronic Structures and Stability from GGA+U. Surf. Sci. 2012, 606, 872−879. (35) Yu, L.; Liu, Y.; Yang, F.; Evans, J.; Rodriguez, J. A.; Liu, P. CO Oxidation on Gold-Supported Iron Oxides: New Insights into Strong Oxide-Metal Interactions. J. Phys. Chem. C 2015, 119, 16614−16622. (36) Van Natter, R. M.; Coleman, J. S.; Lund, C. R. F. DFT Models for Active Sites on High Temperature Water-Gas Shift Catalysts. J. Mol. Catal. A: Chem. 2008, 292, 76−82. (37) Zhu, L.; Yao, K. L.; Liu, Z. L. First-Principles Study of the Polar (111) Surface of Fe3O4. Phys. Rev. B: Condens. Matter Mater. Phys. 2006, 74, 035409. (38) Shimizu, T. K.; Jung, J.; Kato, H. S.; Kim, Y.; Kawai, M. Termination and Verwey Transition of the (111) Surface of Magnetite Studied by Scanning Tunneling Microscopy and First-Principles Calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 235429. (39) Lennie, A. R.; Condon, N. G.; Leibsle, F. M.; Murray, P. W.; Thornton, G.; Vaughan, D. J. Structures of Fe3O4 (111) Surfaces Observed by Scanning Tunneling Microscopy. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 53, 10244−10253. (40) Ahdjoudj, J.; Martinsky, C.; Minot, C.; Van Hove, M. A.; Somorjai, G. A. Theoretical Study of the Termination of the Fe3O4 (111) Surface. Surf. Sci. 1999, 443, 133−153. (41) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (42) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for ab initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (43) Sheppard, D.; Terrell, R.; Henkelman, G. Optimization Methods for Finding Minimum Energy Paths. J. Chem. Phys. 2008, 128, 134106. (44) Henkelman, G.; Uberuaga, B. P.; Jonsson, H. A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901−9904. (45) Yu, X.; Li, Y.; Li, Y.-W.; Wang, J.; Jiao, H. DFT+U Study of Molecular and Dissociative Water Adsorptions on the Fe3O4(110) Surface. J. Phys. Chem. C 2013, 117, 7648−7655. (46) Huber, K. P.; Herzberg, G. Molecular Spectra and Molecular Structure, IV. Constants of Diatomic Molecules; D.Van Nostrand Company Inc, 1979; Vol. 4. (47) Huang, D. M.; Cao, D. B.; Li, Y. W.; Jiao, H. Density Function Theory Study of CO Adsorption on Fe3O4(111) Surface. J. Phys. Chem. B 2006, 110, 13920−13925. (48) Tinkle, M.; Dumesic, J. A. Isotopic Exchange Measurements of the Rates of Adsorption/Desorption and Interconversion of CO and CO2 over Chromia-Promoted Magnetite: Implications for Water-Gas Shift. J. Catal. 1987, 103, 65−78. 28944


Article

The Journal of Physical Chemistry C (49) Tinkle, M.; Dumesic, J. A. Use of Hydrogen/Water Gas Mixtures to Study Adsorption on Chromia-Promoted Magnetite at Water-Gas Shift Temperatures. J. Phys. Chem. 1984, 88, 4127−4130. (50) Rodriguez, J. A.; Liu, R.; Hrbek, J.; Perez, M.; Evans, J. WaterGas Shift Activity of Au and Cu Nanoparticles Supported on Molybdenum Oxides. J. Mol. Catal. A: Chem. 2008, 281, 59−65. (51) Kalamaras, C. M.; Olympiou, G. G.; Efstathiou, A. M. The Water-Gas Shift Reaction on Pt/Gamma-Al2O3 Catalyst: Operando Ssitka-Drifts-Mass Spectroscopy Studies. Catal. Today 2008, 138, 228−234. (52) Grabow, L. C.; Gokhale, A. A.; Evans, S. T.; Dumesic, J. A.; Mavrikakis, M. Mechanism of the Water Gas Shift Reaction on Pt: First Principles, Experiments, and Microkinetic Modeling. J. Phys. Chem. C 2008, 112, 4608−4617. (53) Ammal, S. C.; Heyden, A. Water−Gas Shift Catalysis at Corner Atoms of Pt Clusters in Contact with a TiO2(110) Support Surface. ACS Catal. 2014, 4, 3654−3662. (54) Vecchietti, J.; Bonivardi, A.; Xu, W.; Stacchiola, D.; Delgado, J. J.; Calatayud, M.; Collins, S. E. Understanding the Role of Oxygen Vacancies in the Water Gas Shift Reaction on Ceria-Supported Platinum Catalysts. ACS Catal. 2014, 4, 2088−2096. (55) Tang, Q. L.; Chen, Z. X.; He, X. A Theoretical Study of the Water Gas Shift Reaction Mechanism on Cu(111) Model System. Surf. Sci. 2009, 603, 2138−2144. (56) Yang, T.; Wen, X.-D.; Huo, C.-F.; Li, Y.-W.; Wang, J.; Jiao, H. Structure and Energetics of Hydrogen Adsorption on Fe3O4(111). J. Mol. Catal. A: Chem. 2009, 302, 129−136. (57) Stegelmann, C.; Andreasen, A.; Campbell, C. T. Degree of Rate Control: How Much the Energies of Intermediates and Transition States Control Rates. J. Am. Chem. Soc. 2009, 131, 8077−8082. (58) Rodriguez, J. A.; Ma, S.; Liu, P.; Hrbek, J.; Evans, J.; Perez, M. Activity of CeOx and TiOx Nanoparticles Grown on Au(111) in the Water-Gas Shift Reaction. Science 2007, 318, 1757−1760. (59) Oki, S.; Mezaki, R. Mechanistic Structure of the Water-Gas Shift Reaction in the Vicinity of Chemical Equilibrium. J. Phys. Chem. 1973, 77, 1601−1605.

28945
