Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI
Article 3
Mechanism of Methanol Decomposition on the PtNi(111) Surface: A DFT Study Pan Du, Ping Wu, and Chenxin Cai J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.7b01114 • Publication Date (Web): 17 Apr 2017 Downloaded from http://pubs.acs.org on April 18, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Mechanism of Methanol Decomposition on the Pt3Ni(111) Surface: A DFT Study
Pan Du,†‡ Ping Wu,† and Chenxin Cai*† † Jiangsu Key Laboratory of New Power Batteries, College of Chemistry and Materials Science, Jiangsu Key Laboratory for NSLSCS, Nanjing Normal University, Nanjing 210097, P.R. China ‡ College of Life Science and Chemistry, Jiangsu Second Normal University, Nanjing 210013, P.R. China
ACS Paragon Plus Environment
1
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 38
ABSTRACT: This work reports the detailed mechanism of methanol decomposition on Pt3Ni(111) based on self-consistent periodic density functional theory calculations. The geometries and energies of methanol and its intermediates are analyzed, and the decomposition network is mapped to illustrate the decomposition reaction mechanisms. On Pt3Ni(111), the less electronegative Ni atoms are more favorable for adsorbing radical intermediates and intermediates with lone-pair electrons (such as Ocontaining species). The possible pathways through initial scission of the O‒H, C‒H, and C‒O bonds in methanol are studied and discussed based on the steric effect and electronic structure of the related transition states and the Brønsted-Evans-Polanyi (BEP) relationships. The initial scission of the O‒H bond is the most favorable and bears the lowest energy barrier among the three decomposition modes (initial scission of O‒H, C‒H, and C‒O bonds). The decomposition of the energy barrier analysis indicates that the high energy barrier for initial C‒H and C‒O bond scission is caused by the large structural deformation, strong repulsive interaction, and the low adsorption ability of the decomposed species in their transition states. Potential energy surface (PES) analysis confirmed that the favorable decomposition pathway for methanol on Pt3Ni(111) proceeds via CH3OH → CH3O → CH2O → CHO → CO, in which scission of the O‒H bond is the rate-limiting step. The comparison between the current results and CH3OH decomposition on other systems shows that Pt3Ni(111) can efficiently promote methanol decomposition and alleviate the CO poisoning problem when it is used as an anode catalyst in direct methanol fuel cells (DMFCs).
ACS Paragon Plus Environment
2
Page 3 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
1. INTRODUCTION The adsorption and decomposition of methanol on transition-metal catalyst surfaces are important processes in direct methanol fuel cells (DMFCs) and in the chemistry of methanol steam reforming (MSR) to hydrogen gas.1‒6 Thus, the study of methanol decomposition on catalyst surfaces is of practical interest. Moreover, although methanol is the simplest alcohol, it contains three different types of bonds (C‒H, C‒O, and O‒H bonds). The methanol decomposition reaction has been selected as one of the prototypes in surface science to explore the selective activation of chemical bonds on metal surfaces because this reaction possesses many of the chemical characteristics of more complex carbohydrates used in fuel cell research1,7,9 and includes many of the same elementary reactions as surface-catalyzed mechanisms involving more complicated molecules. In addition, the study of methanol decomposition provides information about the surface structure of catalysts because the reaction is sensitive to catalyst structure.9,10 Therefore, it is also important to understand in detail the mechanism of methanol decomposition from the viewpoint of fundamental science. Pt and Pt-based alloys have long been regarded as the best catalysts for methanol decomposition and have been extensively used in both DMFCs and MSR. However, the detailed decomposition mechanisms of methanol on these catalysts (such as the selectivity and sequence of bond scission of the C‒O, C‒H, and O‒H bonds, elementary reaction steps, and competitive reaction routes) and the role of the second metal in the catalysts have not been convincingly determined. Although many works reported that methoxy (CH3O) is the dominant surface intermediate for CH3OH decomposition and could be easily formed through O‒H bond scission,11,12 Kaichev et al.13 found that the decomposition of methanol on atomically smooth and high-defect Pt(111) single-crystal surfaces occurred via cleavage of the C‒O bond, and others suggested that there would be an additional quasi-stable intermediate, COH,14 produced by sequential scission of C‒H. Moreover, the steps of the further decomposition of CH3O to CO are also different: some authors have suggested CHxO (x = 1, 2) as intermediates,15 and others have
ACS Paragon Plus Environment
3
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 38
observed C‒O bond cleavage,16 while a direct transition from CH3O to CO without any intermediates seems to be favored by still other researchers.12 In addition to these experimental studies, there are many theoretical studies related to CH3OH decomposition because theoretical investigations are of great help to clarifying the reaction mechanisms. Using a density functional theoretical (DFT) analysis, Zhao and co-workers suggested that alloying Pt with Ru on PtRu(111) helps to lower the activation energy barrier and thus promotes methanol decomposition.17 On the basis of a cluster model of PtnRu10‒n, Ishikawa et al. investigated the initial C‒ H scission pathway of methanol decomposition and found that the electronic effect is relatively important compared to the bifunctional mechanism.18 Studies of methanol decomposition on PtAu(111) surface have suggested that this reaction is favored by dehydrogenation to CH2O via initial O‒H activation19,20 rather than the C‒H on pure Pt(111) surfaces.8,21 A recent study of methanol decomposition on Pt3Sn(111) conducted by Lu and co-workers indicated that competitive methanol decomposition started with initial O‒H bond scission followed by successive C‒H bond scission.22 However, the Brønsted‒Evans‒Polanyi relations and energy barrier decomposition analysis identified that the initial C‒H and O‒H bond scissions on the Pt3Sn(111) surface were both favorable and more competitive than C‒O bond scission.22 Thus, although these studies have provided extensive information about CH3OH decomposition, the exact dehydrogenation mechanism of methanol decomposition on the surface of different metal catalysts remains to be disclosed. This work studies the detail mechanism of methanol decomposition on the Pt3Ni(111) surface. PtNi catalysts exhibit high catalytic activity for methanol decomposition and are considered to be more economical than other systems owing to the relatively low cost of Ni.23–28 However, the reason for the higher activity of the PtNi catalyst has not been clearly explained. In previous studies of the effects of the composition, morphology, and aspect ratio (for one-dimensional aligned PtNi catalysts) of PtNi catalysts on their catalytic activity for methanol decomposition, we concluded that the higher activity of PtNi catalysts could be explained by electron transfer from Ni to Pt, which was evidenced by shifts to lower values in Pt4f binding energy.3 This electron transfer modifies the electronic properties of the Pt ACS Paragon Plus Environment
4
Page 5 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
catalyst by changing the electron density of states of the d‒band and the Fermi level energy because of an electronic or ligand effect, hence enhancing the catalytic activity of the catalyst. We also studied the facet-dependent adsorption of methanol on the (111), (100), and (110) surfaces of a Pt3Ni catalyst using DFT calculations with van der Waals (vdW) corrections.29 The mechanism of the facet-dependent adsorption of methanol on the catalyst was rationalized in terms of the shifts of the d‒band center of the Ni component relative to the Fermi level, density of states, changes to the work function of each surface of the Pt3Ni catalyst, and polarization effects of the adsorbed methanol. The goal of this work is to understand the detailed mechanism of methanol decomposition on Pt3Ni(111) using periodic DFT calculations. The structure, adsorption stability, and site preference of CH3OH and the intermediates involved in its decomposition on Pt3Ni(111) are calculated. The possible pathways through initial scission of O‒H, C‒H, and C‒O bonds are discussed based on the steric effect and electronic structure of the related transition states and the Brønsted-Evans-Polanyi (BEP) relationships. Both the thermochemistry and the reaction barriers of each elementary reaction step along the decomposition pathway are addressed to obtain a reaction network. The vibrational frequencies for the adsorbed states of these intermediates are calculated so that the reaction rate constants for each possible elementary step can be calculated using harmonic transition state theory. The potential energy surface (PES) is constructed, and the rate-limiting step is determined. The current results and CH3OH decomposition on other systems are compared, demonstrating that Pt3Ni(111) can efficiently promote methanol decomposition in DMFCs. This work contributes to the fundamental understanding of the structural, energetic, and catalytic properties of PtNi catalysts, and the results gained through this work may help to guide the rational design and construction of nanoarchitectured PtNi surfaces for optimal heterogeneous catalysis. 2. THEORETICAL METHODS Self-consistent periodic DFT calculations were performed using the Vienna ab initio simulation package (VASP).30,31 The interaction between ionic cores and electrons was described by projector augmented ACS Paragon Plus Environment
5
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 38
wave (PAW) pseudo-potentials.32,33 The 5d and 6s electrons of Pt, the 3d and 4s electrons of Ni, and the 2s and 2p electrons of O were used as the valence electrons. The exchange-correlation functional was described within the generalized gradient approximation (GGA)34 proposed by Perdew, Burke, and Ernzerhof (PBE).35 The vdW correction proposed by Grimme (DFT+D3) was employed to improve the description of the PBE functional, in which a correction was added to the selected exchange-correlation functional.36,37 The Kohn-Sham equations were solved using a plane-wave basis set with a kinetic energy cutoff of 400 eV. Brillouin zone integrations were performed using Monkhorst-Pack grids33 of 11 × 11 × 11 and 5 × 5 × 1 for the bulk and slab calculations, respectively. For the density of states (DOS) calculations, the k–point mesh was set to 7 × 7 × 1 because increasing the k–point mesh to 11 × 11 × 1 did not produce observable energy differences, suggesting that the 7 × 7 × 1 mesh was sufficient to achieve high calculation accuracy. Spin polarization was considered in the calculations, and the Gaussian smearing method was employed to determine electron occupancies, with a smearing parameter σ of 0.2 eV. The equilibrium geometries were obtained when the atomic forces were less than 0.01 eV/Å on each atom, using a total energy convergence of 10−6 eV. Because both Pt and Ni form closely packed face-centered cubic (fcc) structures, and because the experimental results demonstrated that the Pt3Ni alloy also has an fcc structure,3,38 we used an fcc unit cell for our calculations, in which all corner sites were occupied by Ni atoms and the remaining sites were occupied by Pt atoms. This configuration gives the Pm3m space group. The unreconstructed Pt3Ni(111) surface was modeled using a three-layer slab with (2 × 2) surface unit cells (Figure S1) because the surface energy has converged at the three-layer slab based on the energy calculation of the two-, three-, four-, and five-layer slab with (2 × 2) surface unit cells. The atoms in the topmost layer were allowed to relax according to their atomic forces, whereas the atoms on the remaining two layers were fixed at their ideal bulk positions to mimic bulk behavior. A vacuum gap of ~14 Å and a dipole correction along the z–direction were introduced. Under the current computational conditions, the lattice constant of bulk Pt3Ni was calculated to be 3.88 Å, which is identical to the previously calculated value from Yang et al. (3.88 Å)39 and agrees well with the experimental value of 3.89 Å.40 It is also consistent ACS Paragon Plus Environment
6
Page 7 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
with Vegard’s law,41 which predicts that the lattice constant of PtNi alloys decreases linearly with increasing amount of Ni substituted into the Pt matrix because both Pt and Ni are fcc structures. These results suggest that the present model is appropriate for its current application. We also optimized the configurations of isolated CH3OH molecules in the gas phase in a 20 × 20 × 20 Å3 cubic box and calculated their bond lengths. The bond lengths were 1.44 Å for the C–O bond, 1.10 Å for the C–H bond, and 0.96 Å for the O–H bond, agreeing well with both the available experimental values of 1.43, 1.10, and 0.96 Å42 and the theoretical values of 1.41, 1.09, and 0.95 Å,43 further demonstrating that our calculations are reliable. In the CH3OH/Pt3Ni configurations, only one CH3OH molecule per super cell was adsorbed on one side of the slab (corresponding to the surface coverage of 1/4 ML) to reduce the lateral interactions between adsorbates. CH3OH was almost vertical to the surface and interacted with the surfaces via its hydroxyl group. The stable adsorption state of methanol on the Pt3Ni(111) surface was used to initiate our exploration of the reaction and the pathways of methanol decomposition. In addition, we also calculated the adsorption of CH3OH molecule (on Ni-top site) on a larger (2 × 4) Pt3Ni(111) cell. The relevant adsorption energy was found to be ‒0.81 eV, which is slightly lower than that calculated on the (2 × 2) cell (‒0.83 eV), implying that the effect of surface coverage is not significant for adsorption. We optimized the geometries of CH3OH and the intermediates. We calculated the adsorption energies (Eads) of these species at different adsorption sites using the following definition: Eads = Eadsorbate+Pt3Ni ‒ Eadsorbate – EPt3Ni
(1)
where Eadsorbate+Pt3Ni is the total energy of the adsorbed species and the catalyst (Pt3Ni), EPt3Ni is the energy of the clean catalyst, and Eadsorbate is the energy of the isolated species in the gas phase calculated using a 20 Å cubic box. Using this definition, a negative value of Eads denotes an exothermic adsorption process, indicating that the adsorption of those species on the catalyst surface is energetically favorable. In contrast, a positive value of Eads denotes an endothermic adsorption process, indicating that the adsorption of those species on the catalyst surface is energetically unfavorable. The minimum energy paths (MEPs) of various elementary steps involved in methanol decomposition ACS Paragon Plus Environment
7
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 38
on Pt3Ni(111) were identified by the climbing image nudged elastic band (CI-NEB) method.44‒47 The transition-state search was initiated by interpolating intermediate images of the system between the initial and final states on the PES. A spring interaction between the images was added to maintain the spacing between adjacent images. The vibrational frequencies were calculated by numerical differentiation of the forces using a second-order finite difference approach with a step size of 0.015 Å. The highest energy along the reaction coordinate relative to that of the initial state gave the activation energy barrier (Ea). Ea = ETS ‒ Ereactant
(2)
where ETS and Ereactant are the energies of the transition state (TS) and reactant, respectively. The reaction energy, ∆H, was calculated according to the formula: ∆H = Eproduct ‒ Ereactant
(3)
where Eproduct is the energy of the product. Zero-point energy (ZPE) corrections, which were calculated based on the vibrational frequencies (eq 4), were included in the barrier and reaction energy calculations.
ZPE = ∑ ℎ
(4)
where h is Planck’s constant and νi is the vibration frequency. The reaction rate constant, k, of each elementary step depends on the magnitude of Ea and the preexponential factors, A. The values of k and A of each step were calculated using conventional transition state theory (eqs 5, 6)48 = exp(− / )
(5)
where A is the pre-exponential factor, Ea is the activation energy barrier, kB is Boltzmann’s constant, and T is temperature (300 K in this work). The A values are given by eq 6:
=
(6)
where QTS and QIS are the partition functions per unit volume for TS and the initial state (IS), respectively. ACS Paragon Plus Environment
8
Page 9 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
3. RESULTS For clarity, this section is organized as follows. First, we present the structures and Eads of the most stable adsorbed intermediates. Then, we calculate the structure of TS and the values of Ea and ∆H of each possible elementary reaction to obtain the thermodynamic and kinetic data of CH3OH decomposition on the Pt3Ni(111) surface. 3.1. Structures and Adsorption Energies. The adsorption of methanol and various intermediates, including CH3OH, CH3O, CH2O, CHO, CH2OH, CHOH, COH, CO, CH3, OH, and atomic H, were examined. The most stable configurations are shown in Figure 1, and the corresponding Eads and structural parameters are listed in Table 1 (for comparison, the Eads of these intermediates on Pt(111) and Ni(111), which were obtained from previous reports, are also listed in Table 1).
Figure 1. The most stable adsorption configurations of CH3OH and its intermediates involved in CH3OH decomposition on the Pt3Ni(111) surface. Atomic color code: yellow, platinum; green, nickel; gray, carbon; red, oxygen; and white, hydrogen.
Adsorption of CH3OH. CH3OH prefers a Ni top site of Pt3Ni(111) surface via donation of the lonepair electrons from oxygen to the metallic surface (Figure 1), with an Eads of ‒0.83 eV (Table 1). The C‒ O bond is inclined at an angle of 28 ° from the surface normal, and the O‒H bond is oriented toward an
ACS Paragon Plus Environment
9
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 38
hcp site, facilitating the binding of CH3OH to the surface via the oxygen lone-pair orbital. The large O‒ surface distance (2.05 Å), as well as small changes in the structure of CH3OH upon adsorption, suggest the relatively weak adsorption of CH3OH on Pt3Ni(111). However, the value of Eads (‒0.83 eV) is much greater than that on pure Pt(111) (‒0.33 eV)8 and Ni(111) (‒0.17 eV)49 surfaces (Table 1), indicating that CH3OH decomposition on the Pt3Ni(111) surface is more favorable than on other catalyst surfaces. Table 1. Adsorption Configurations, Adsorption Energies (Eads), and Structural Parameters for CH3OH and Its Intermediates Involved in CH3OH Decomposition on the Pt3Ni(111) Surface. a αb (°)
Eads (Pt) (eV) c
Eads (Ni) (eV) d
2.10
28
‒0.33
‒0.17
1.43
1.85
28
‒1.54
‒2.58
47
species
site
configuration
Eads (eV)
dC‒O (Å)
dO‒Ni (Å)
CH3OH
Ni-top
η1(O)
‒0.83
1.44
‒2.61
CH3O CH2O CH2O-t CHO CH2OH CHOH
Ni-top Ni-top PtNi-bridge PtNi-bridge PtNi-bridge Pt2-bridge
1
η (O) 1
η (O)
dC‒Pt (Å)
‒0.64
1.23
1.88
1
1
‒1.08
1.34
1.89
2.13
87
‒0.50
‒1.03
1
1
‒3.20
1.25
2.07
1.96
92
‒2.36
‒2.41
1
1
‒2.24
1.46
2.13
2.07
90
‒1.98
‒1.68
‒3.42
1.32
2.03
58
‒3.24
0
‒4.45
0
‒1.82
‒2.31
‒2.71
‒2.94
‒1.93
‒2.04
‒2.12
‒3.19
η (C)-η (O) η (C)-η (O) η (C)-η (O) 2
η (C)
2.26 COH
Pt3-hcp
3
η (C)
‒4.07
1.33
2.00 2.01 2.03
CO
PtNi-bridge
η2(C)
‒2.05
1.18
2.01 1.92
H
fcc
η2(H) 1
‒2.87
CH3
Pt-top
η (C)
‒2.98
OH
PtNi-bridge
η2(O)
‒3.43
2.08 1.96 2.19
e
f
a
The values of Eads on Pt(111) and Ni(111), which were obtained from previous reports, are also listed for comparison. b The angle of the O‒C axis inclines from the surface normal. c The values are Eads on Pt(111), taken from Greeley et al. [Greeley, J.; Mavrikakis, M. J. Am. Chem. Soc. 2004, 126, 3910‒3919]. d The values are Ni(111), taken from Wang et al. [Wang, G.-C.; Zhou, Y.-H.; Morikawa, Y.; Nakamura, J.; Cai, Z.-S.; Zhao, X.-Z. J. Phys. Chem. B 2005, 109, 12431‒12442]. e Lengths of the C‒Ni bonds. f Length of the O‒Pt bond.
Adsorption of CHxO (x = 0‒3). The methoxy (CH3O) formed from O‒H bond scission of CH3OH also prefers the Ni top site, with an Eads of ‒2.61 eV, which is much greater than that of CH3OH due to the abstraction of one H. One effect of the adsorption is a substantial stretching of the C‒O bond (1.43 Å ACS Paragon Plus Environment
10
Page 11 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
versus 1.34 Å) of CH3O due to electron donation from metallic 3d orbitals to the π* orbital of C‒O in CH3O.43 Our calculation indicates that there is 0.188 |e| moving from Ni 3d orbitals to the π* orbital of C‒O in CH3O. The removal of one H atom from CH3O generates formaldehyde (CH2O), which is a critical intermediate in CH3OH decomposition and synthesis. Because the O atom of CH3O binds strongly to the metal surface, when the C‒H bond breaks, CH2O adsorbs on the metal surface through the oxygen lone-pair electrons in an upright structure, with an Eads of ‒0.64 eV. However, the preferred mode of CH2O adsorption on Pt(111) is CH2O binding to the surface through the carbon π orbital and through simultaneous overlap between the metal d state and the carbonyl π* orbital.50 We calculated the adsorption of CH2O on the bridge site and found that CH2O is apt to adsorb on the Pt‒Ni bridge site via the η1(C)-η1(O) mode, with the O atom binding to an Ni atom and the CH2 group pointing to a Pt atom (denoted as CH2O-t), as shown in Figure 1. The Eads is ‒1.08 eV, which is similar to that obtained on Ni(111) (‒1.03 eV,49 adsorbed above a bridge site through the C and O atoms in a di-σ mode). This relatively low Eads is predicted by its closed-shell configuration. Similar to CH2O, formyl (CHO) also prefers to adsorb on the Pt‒Ni bridge site via η1(C)-η1(O) mode, simultaneously forming Pt‒C (1.96 Å) and Ni‒O (2.07 Å) bonds. The alloying strengthens the adsorption of CHO, with an Eads of ‒3.20 eV, compared to that on Pt(111) (‒2.36 eV)8 and Ni(111) (‒2.41 eV)49 (Table 1). Adsorption of CHxOH (x = 0‒2). Hydroxymethyl (CH2OH), formed by C‒H scission of CH3OH, prefers to bond at the Pt‒Ni bridge site on Pt3Ni(111) via η1(C)-η1(O) mode with the O atom binding to a Ni atom and the C atom binding with a Pt atom (Figure 1). The C‒O axis is parallel to the Pt3Ni(111) surface (α = 90 °), and an Eads of ‒2.24 eV is obtained, which is higher than those reported on Pt(111) (‒ 1.98 eV)8 and Ni(111) (‒1.68 eV).49 Although hydroxymethylene (CHOH) has a closed-shell electronic configuration, its CH end is still active and can bind stably at the Pt2-bridge site through its C atom (in η2(C) mode), with a high Eads of ‒3.42 eV. Hydroxymethylidyne (COH) prefers to bind to the Pt3-hcp site (η3(C) mode) through its carbon atom, with an Eads of ‒4.07 eV, indicating a strong interaction between the C atom and Pt3Ni because the C atoms have no H neighbors. Removal of one H atom from the CHx end of CHxOH results in enhancement of the Eads by approximately 0.6‒1.2 eV. This may be ACS Paragon Plus Environment
11
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 38
due to the alteration of the binding nature to the Pt3Ni(111) surface because CH2OH, CHOH and COH bind with Pt3Ni(111) via η1(C)-η1(O), η2(C), and η3(C) modes, respectively. Generally, bonding through a carbon atom is stronger than through an oxygen atom; therefore, the increase in the number of C‒Pt bonds formed in adsorbed CH2OH, CHOH and COH results in the enhancement of Eads. This phenomenon is also observed on Pt(111),8 PtRu(111),17 and Ru(0001) surfaces.51 Adsorption of CO and Atomic H. CO interacts with the Pt3Ni(111) surface through the carbon atom via η2(C) mode, forming Pt‒C (2.01 Å) and Ni‒C (1.92 Å) bonds. The Eads is ‒2.05 eV, which is higher than that reported on Pt(111) (‒1.82 eV)8 and Pt2Ni(111) (‒1.37 eV)52. The H atom prefers the fcc site with a distance of 1.15 Å to the surface, slightly longer than that on Ni(111) (0.905 Å)49 and much shorter than that on Pt(111) (1.57 Å).8 The Eads is ‒2.87 eV, which is higher than that reported on Pt(111) (‒2.71 eV).8 The high Eads may be due to its radius being so small that it sinks deep into the surface, as indicated by its short distance from the surface (1.15 Å). Adsorption of CH3 and OH. Methyl (CH3), formed by the scission of the C‒O bond in CH3OH, sits at the Pt top site, with an Eads of ‒2.98 eV, which is higher than that reported on Ni(111) (‒1.75 eV,49 on the top site), Pt(111) (‒1.93 eV),8 and Pt3Sn(111) (‒1.98 eV,22 on the Sn top site). The hydroxyl radical adsorbs preferentially on the Pt‒Ni bridge site (in η2(O) mode), and the distances of the oxygen to Pt and Ni are 2.19 and 1.96 Å, respectively. The Eads is ‒3.43 eV (Table 1), which is stronger than the value of ‒2.12 eV obtained on Pt(111),8 indicating that OH radical species form strong bonds with Pt3Ni(111). In view of the electronegativity of the OH radical and its partially vacant highest-lying antibonding 1π orbital, the binding of the OH radical to nickel is expected to involve electron donation from the Pt3Ni(111) surface. The partially vacant antibonding π orbitals of the OH radical play a major role in chemisorptive binding on Pt3Ni(111) and result in the higher Eads value. 3.2. Decomposition Reaction. After determining the perfected adsorption site and Eads for each intermediate, we explore the detailed reaction mechanism using Ea and ∆H calculations. The structures of the IS, TS, and final state (FS) for each elementary step are depicted in Figure 2. The relevant Ea and ∆H values are listed in Table 2. Generally, we choose the most stable sites of intermediates as the IS, ACS Paragon Plus Environment
12
Page 13 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
and the corresponding product species in the most stable positions as the FS. However, to facilitate the reaction on the Pt3Ni(111) surface, the dissociated H atom is located at the substable Pt-top site as the final state of the dissociation step.
Figure 2. The structures of the IS, TS, and FS for each elementary step of CH3OH decomposition on Pt3Ni(111) via initial O‒H bond scission.
ACS Paragon Plus Environment
13
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 38
Hydrogen Abstraction from Methanol. The first decomposition step involves scission of the C‒O, C‒ H, or O‒H bond of methanol to initiate the catalytic cycle. However, our calculations indicate the initial C‒O and C‒H bond scissions have rather high Ea (2.08 and 1.63 eV for scission of the C‒O and C‒H bonds, respectively) and thus are too difficult to occur in general. Therefore, we analyze the abstraction of hydroxyl hydrogen from adsorbed CH3OH on Pt3Ni(111). We further analyze the initial C‒O and C‒ H bond scission in the next section. Table 2. Reaction Energy (∆H, in eV), Activation Energy Barrier (Ea, in eV), Pre-Exponential Factor (A, in s‒ 1 ), and Rate Constant (k, in s‒1) at 300 K for the Elementary Steps Involved in CH3OH Decomposition on the Pt3Ni(111) Surface. a reactions
∆H (eV)
Ea (eV)
Ea (Pt) (eV)
A (s‒1)
k (s‒1)
R1
CH3OH → CH3O + H
0.41
0.69
0.81 b
1.32 × 1013
3.38 × 101
R2
CH3OH → CH3 + OH
‒0.22
2.08
2.19 c
4.79 × 1014
5.45 × 10‒21
R3
CH3OH → CH2OH + H
0.64
1.63
0.67 c
8.45 × 1013
3.49 × 10‒14
R4
CH3O → CH2O + H
‒0.09
0.11
1.03 × 1014
1.46 × 1010
R5
CH3O → CH3 + O
‒0.79
1.06
5.62 × 1012
8.74 × 10‒6
R6
CH2O → CH2O-t
‒0.35
0.04
2.16 × 1012
4.59 × 1011
R7
CH2O-t → CHO + H
‒0.68
0.29
6.84 × 1013
9.18 × 108
R8
CH2O-t → CH2 + O
‒0.03
2.21
1.52 × 105
1.13 × 10‒22
R9
CHO → CO + H
‒0.74
0.43
2.33 × 1014
6.33 × 1014
R10 CHO → CH + O
‒0.01
2.23
3.23 × 1013
1.11 × 10‒24
a
The values of Ea for the initial O‒H, C‒O, and C‒H scissions in CH3OH decomposition on Pt(111) are also listed for comparison. These values were obtained from previous reports. b The values are the Ea on Pt(111), taken from Greeley et al. [Greeley, J.; Mavrikakis, M. J. Am. Chem. Soc. 2002, 124, 7193‒7201]. c The values are the Ea on Pt(111), taken from Greeley et al. [Greeley, J.; Mavrikakis, M. J. Am. Chem. Soc. 2004, 126, 3910‒3919].
For O‒H scission, the IS is the Ni top adsorbed CH3OH, while coadsorbed CH3O and H are the FS. In this process, the reaction coordinate begins with an O‒H stretch vibration at 3539 cm‒1 (the frequency of this mode becomes imaginary at the TS; the calculated vibrational frequencies for CH3OH and its intermediates are listed in Table S1) and ends with CH3O bound through the O atom at Ni top site and H adsorbed to an adjacent Pt top site (step (a) in Figure 2). The activated O‒H distance is elongated from 0.982 Å in the IS to 1.574 Å in the TS (TS1 in Figure 2), indicating that the O‒H bond becomes weaker in the TS. The distance from the C atom to the catalyst surface is 3.09 Å, much shorter than its initial ACS Paragon Plus Environment
14
Page 15 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
value in CH3OH (3.43 Å), which indicates that a portion of the CH3O in the TS moves closer to the catalyst surface. Beyond the TS, the CH3O formed is left at the Ni top site, while the H atom moves to an adjacent Pt top site. This step has an Ea of 0.69 eV and is endothermic by 0.41 eV (Table 2). Hydrogen Abstraction from Methoxy. Subsequently, methyl H abstraction from CH3O affords coadsorbed formaldehyde and hydrogen. The IS is selected as the most stable CH3O configuration (Ni top), and the FS consists of a hydrogen atom at a Pt top site and formaldehyde on an adjacent Ni top site, as shown in Figure 2 (step (b)), which guarantees that the repulsion between them is smaller. This process begins with a C‒H stretch vibration at 2985 cm‒1. To favor C‒H bond activation and scission, the C‒O bond inclines and brings the hydrogen atom close to the surface (the angle α changes from 28° in the IS to 69° in the TS). The TS (TS2 in Figure 2) is formaldehyde-like, with the other hydrogen atom essentially moving straight away from the initial location in the CH3O radical to the final Pt top site, and the formaldehyde part rises from the surface, as indicated by the angle α of 47° in the FS. This step has a small Ea of 0.11 eV, and the reaction is exothermic by ‒0.09 eV (Table 2). The C‒O bond scission in CH3O has an Ea of 1.06 eV (Table 2). The TS for the C‒O bond scission in CH3O is depicted in Figure S2, TS6, and this value is much higher than that of the C‒H bond scission (0.11 eV), suggesting that C‒O bond scission in CH3O is not a possible pathway for CH3OH decomposition on the Pt3Ni(111) surface. The Ni top site adsorbed CH2O then turns its C‒O axis downward (α changes from 47° to 87°), with the C‒O axis almost parallel to the Pt3Ni(111) surface and the C atom in the CH2 group binding to the Pt atom, forming CH2O-t (step (c) in Figure 2). This structure conversion occurs via TS3 (Figure 2). In TS3, the angle α increases to 61° in the TS from 47° in the IS, and the distance between the C atom and Pt atom decreases to 3.05 from 3.47 Å. After TS3, the CH2O group binds to the surface through C and O atoms. This transition needs to overcome a very small Ea (0.04 eV), with an exothermic energy of ‒0.35 eV. Hydrogen Abstraction from Formaldehyde. After formaldehyde is formed, its decomposition to yield the desired final products CO and H may be determined by the competition between the reaction of ACS Paragon Plus Environment
15
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 38
formaldehyde dehydrogenation to yield adsorbed formyl and a hydrogen atom and formaldehyde desorption. Our calculation indicated that the Ea for formaldehyde (CH2O-t) desorption is greater than 0.91 eV, which is much higher than that for formaldehyde decomposition (0.29 eV). Therefore, formaldehyde decomposes to formyl. In the formation of CHO, we use CH2O as the IS, CHO as its favorable site (i.e., Pt‒Ni bridge site) and hydrogen at the Pt top site as the FS (step (d) in Figure 2). This process begins with a C‒H stretch vibration at 2987 cm‒1. Since the TS (TS4, Figure 2) has little change in geometry relative to the IS, this step has a low Ea (0.29 eV). This step is exothermic, with ∆H of ‒0.68 eV. We also perform a similar calculation for the C‒O bond scission in formaldehyde (CH2O-t). The Ea is 2.21 eV (TS7 in Figure S2), which is much higher than that of the C‒H bond-breaking reaction, suggesting this step is not a possible pathway in CH3OH decomposition. To the best of our knowledge, there is no experimental result confirming the existence of the CH2 species in methanol decomposition. Hydrogen Abstraction from Formyl. After CHO has been produced on the Pt3Ni(111) surface, it can either produce CO and H via C‒H scission or CH and O via C‒O scission. The C‒H bond scission in formyl begins with a C‒H stretch vibration at 2914 cm‒1 and ends with CO and H (step (e) in Figure 2), with an Ea is 0.43 eV (Table 2). The Ea for C‒O scission in CHO is 2.23 eV (Table 2; the TS for this scission reaction is depicted in TS8, Figure S2), limiting the occurrence of the reaction. Therefore, CHO is likely to decompose readily into CO and hydrogen and will not poison the catalyst surface toward CH3OH decomposition. In the TS (TS5 in Figure 2), the C‒H bond is elongated, and the departing H atom is located at the adjacent Pt top site (Figure 2). The Ni‒O bond is broken at a distance of 2.97 Å, and CHO forms a monodentate structure through the C atom. After the TS, the O atom moves to the top of the C atom, which forms bonds with the Pt and Ni atoms. The H atom departs from C and adsorbs stably at a Pt top site in the FS. This process is exothermic, with ∆H of ‒0.74 eV. To identify the possible products under suitable conditions, we study the decomposition of the adsorbed CO. The Ea of C‒O bond-breaking on Pt3Ni(111) is calculated to be 2.86 eV, indicating that
ACS Paragon Plus Environment
16
Page 17 of 38
this reaction is not important under any conditions. This value (2.86 eV) is higher than the desorption energy of CO (2.05 eV), implying that CO will prefer desorption from Pt3Ni(111) to decomposition. 4. DISCUSSION In this section, we first analyze the adsorption sites and the adsorption nature of CH3OH and its intermediates on Pt3Ni(111). Then, we analyze the factors influencing the decomposition reaction path (especially the initial scission reaction in CH3OH decomposition). PESs are built, and the rate constants of each elementary step are calculated and discussed. Finally, the CH3OH decomposition mechanism on the Pt3Ni(111) surface is depicted and compared with that on other catalyst surfaces. 8 6 4 2 0 8 6 4 2 0
d−PDOS
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
8 6 4 2 0 8 6 4 2 0
Pt(111)
Pt3Ni(111)
Pt3Ni(111)−Pt
Pt3Ni(111)−Ni
8 6 Ni(111) 4 2 0 -10
-5
0
5
Energy (eV) Figure 3. d‒Partial density of states of Pt3Ni(111), Pt in Pt(111) and Pt3Ni(111), and Ni in Pt3Ni(111) and Ni(111). The vertical green line represents the Fermi level. ACS Paragon Plus Environment
17
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 18 of 38
4.1. Analysis of Adsorption Sites. From the Eads calculations for CH3OH and its intermediates, we can conclude that the intermediates are apt to adsorb around the Ni site rather than the Pt site. This situation may be induced by the electronegativity of the metal atoms. The electronegativities of Ni and Pt atoms are 1.91 and 2.28, respectively. Therefore, the Ni atoms in Pt3Ni(111) have slight positive charges, while the Pt atoms are negatively charged. Bader analysis indicates that the Ni and Pt atoms on the Pt3Ni(111) surface are populated by 0.3715 and ‒0.2025 |e|, respectively. Thus, compared to the Pt atom, the Ni atom favors adsorption of the intermediates of radical or having lone-pair electrons (such as Ocontaining species). In comparison with those on the Pt(111) surface, the adsorption of CH3OH and its intermediates on their stable adsorption sites is strengthened by the effect of Ni, especially for those containing the η1(O) mode (such as CH3OH, CH3O, CH2O, CHO, CH2OH, and OH; the Eads of these intermediates on the Pt(111) and Ni(111) are listed in Table 1). This can be understood by analyzing the d‒partial density of states (d‒PDOS) of Pt(111), Ni(111), and Pt3Ni(111), which is depicted in Figure 3. Compared to Pt(111), the d‒band of Pt3Ni(111) is broad and the number of the d‒band peaks decreases, suggesting stronger delocalizability of d electrons. In particular, the significant increase in the d‒PDOSs for the Pt3Ni(111) surface near the Fermi level favors the reaction activity. To evaluate the adsorption activity of Pt3Ni(111), the d–band center was calculated based on the Hammer-Nørskov d–band model.53 It is ‒ 2.68 eV for Pt in Pt(111), ‒2.78 eV for Pt in Pt3Ni(111), ‒1.28 eV for Ni in Pt3Ni(111), and ‒1.51 eV for Ni in Ni(111). After alloying Pt with Ni, the downward shift (from ‒2.68 to ‒2.78 eV) of the d‒band center of the Pt atoms accounts for the weaker adsorption at Pt sites than on pure Pt (111), while the higher d‒band center of Ni than Pt in Pt3Ni(111) (‒1.28 vs ‒2.78 eV) results in stronger adsorption at the Ni sites than at the Pt sites. We can also conclude that radical intermediates (such as CH3O, CHOH, COH, OH, and CH3) have much stronger interactions with Pt3Ni(111) than do the molecules (such as CH3OH, CH2O, and CO), which is due to the unpaired electrons of the radicals. Therefore, the radicals are highly reactive and ACS Paragon Plus Environment
18
Page 19 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
have stronger charge acceptance tendency from the metal surface atoms during chemisorption. Those intermediates without a steric hindrance (such as CHOH, COH, and CO) interact with Pt3Ni(111) through the C atom, not through the O atom, because the C atom usually has a stronger tendency than the O atom to form a bond with a metal surface. However, the steric hindrance plays such an important role that those radicals bonding through the O atom (such as CH3O) also have a higher binding energy. 4.2. Analysis of the Initial Scission in CH3OH Decomposition. The first step of CH3OH decomposition involves scission of a C‒O, C‒H, or O‒H bond. Our calculations show that scission of the O‒H bond is more favorable than scission of the C‒O and C‒H bonds because scission of O‒H has a low Ea value (Table 2). There are several reasons why the O‒H bond is activated in preference to the C‒ O and C‒H bonds. First, the favorable O‒H scission can be understood based on steric effects. The structure of adsorbed CH3OH in Figure 1 shows that to reduce the steric hindrance of methyl, the O‒H bond has to be closer to the Pt3Ni(111) surface than the C‒O and C‒H bonds, which causes the O‒H bond to be more activated by the catalyst and the bond length to change to a greater extent (the bond lengths of C‒H, C‒O, and O‒H in CH3OH are 1.10, 1.44, and 0.98 Å, respectively, and the bond lengths of the CH3OH molecule in the gas phase are 1.10 Å for C‒H, 1.43 Å for C‒O, and 0.96 Å for O‒H bond). Therefore, O‒H bond scission may be the favorable step. Moreover, the structure of the TS in the O‒H bond scission (TS1 in Figure 2) is similar to the reactant CH3OH, and H and CH3O still bind together, with low repulsive interaction. In contrast, as shown in Figure S2, the TS in C‒O and C‒H scission carries two dissociated fragments, which are so close that they have a strong repulsive interaction and bring the TS to a higher-energy state (the distance between the C atom in CH2OH and the dissociated H is 2.49 Å in TS9 for initial scission of the C‒H bond, and the distance between the C atom in CH3 and the O atom in dissociated OH is 2.45 Å in TS10 for initial scission of the C‒O bond). Second, the nature of initial O‒H scission can be better understood based on the electronic structures of the O‒H, C‒H, and C‒O bonds. We calculate the d‒PDOS of the surface Pt and Ni atoms on the bare surface and in the TS for initial C‒H, O‒H, and C‒O bond scission. Figure 4 shows the ∆d‒PDOS of ACS Paragon Plus Environment
19
The Journal of Physical Chemistry
the surface atoms, which is obtained by subtracting the d‒PDOS of the bare surface from that of the TS for initial C‒H, O‒H, and C‒O bond scission. The stabilization energy (Ed) of the d‒states is calculated using the following equation (eq 7):54,55 +
, ! = "-. # ⋅ (%!&' − %!()* ) /#
(7)
where nd is the normalized DOS of the surface atoms (in electrons/eV) with and without bonds to TS adsorbates, EF is the Fermi level, and ε is the energy level. Generally, a larger absolute value of Ed indicates stronger interaction between the adsorbates and the surface, resulting in lower activation energy. The Ed is ‒0.22 eV, ‒1.78 eV, and ‒2.98 eV for the C‒O, C‒H, and O‒H bonds, respectively, suggesting that among the three bond scissions, O‒H bond scission results in the strongest interaction between the fragment and Pt3Ni(111) surface, producing the lowest Ea, which is in good agreement with the Ea of 2.08, 1.63, and 0.69 eV for initial C‒O, C‒H, and O‒H bond scission, respectively, in Table 2. 10 O−H C−H C−O
5
∆d−PDOS
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 38
0
-5
-10
-5
0
5
Energy (eV)
Figure 4. Variation of the d‒projected density of states (d‒PDOS) of the surface metal atoms (Pt and Ni atoms) due to their interaction with the TS complex in the initial C‒H, C‒O, and O‒H bond scission in CH3OH decomposition on the Pt3Ni(111) surface. The vertical green line represents the Fermi level.
Third, the initial scission reaction in CH3OH decomposition can be further analyzed by the relationships between the Ea and ∆H in each elementary step, which are known as Brønsted-EvansPolanyi (BEP) relationships,54,56 because BEP relationships can be analyzed to distinguish the selectivity of the C‒H, O‒H, and C‒O bond scissions.22 Figure 5 depicts three distinct BEP ACS Paragon Plus Environment
20
Page 21 of 38
relationships for the C‒O (a), C‒H (b), and O‒H bond scission reactions (c). Although no single BEP relationship can be established for all the elementary reactions on the Pt3Ni(111) surface, a rough linear BEP relationship is obtained for each type of bond scission reaction, and the BEP lines for C‒H, O‒H, and C‒O bond scissions are nearly parallel. The intercepts of C‒O (2.26) and C‒H bond scission (0.95) are much larger than that of O‒H scission (0.26), indicating that C‒O and C‒H bond scissions do not readily occur compared to O‒H bond scission. 2
1
ETS (eV)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
O−H C −H C −O
0
-1
-2 -2.5
-2.0
-1.5
-1.0
-0.5
0.0
EFS (eV)
Figure 5. Plot of the TS energy against the FS energy for all bond scission reactions involved in methanol decomposition on Pt3Ni(111). The energy reference for each step is the clean slab plus the gas-phase reactants. The square, circle, and triangle symbols denote the O‒H, C‒H, and C‒O bond scission reactions, respectively. The linear regression equation for the O‒H bond scission is ETS (TS energy, in eV) = 0.89 EFS (FS energy) + 0.26 (eV), with a correlation coefficient (R) of 0.89. The linear regression equation for C‒H bond scission is ETS = 0.92 EFS + 0.95 (eV), with an R of 0.99, and the linear regression equation for the C‒O bond scission is ETS = 1.09 EFS + 2.26 (eV), with an R of 0.83.
4.3. Analysis of the Energy Barrier of the Initial Scission of the O‒H, C‒H, and C‒O Bonds. To provide novel insight into the factors governing initial bond activation and scission, the Ea of the initial scission of the O‒H, C‒H, and C‒O bonds is decomposed using the following formula (eq 8):57‒59 Ea = ∆Esub + ∆EABdef + EABIS + EintTS ‒ EATS ‒ EBTS
(8)
where ∆Esub is the effect of the structural change of the substrate from the IS to the TS on the activation barrier (∆Esub = EsubTS ‒ EsubIS, EsubTS and EsubIS refer to the energy of substrate at TS and IS, ACS Paragon Plus Environment
21
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 22 of 38
respectively); ∆EABdef is the deformation energy, which refers to the effect of the structural deformation of AB on the barrier and is defined as the difference of the energy of AB in the TS and IS, where AB represents CH3OH in the present work; EABIS is the adsorption energy of AB in the IS; EintTS is the interaction between A and B in the TS and is defined as the energy difference of the (A + B) and AB in the TS, where A represents CH3O, CH2OH, or CH3 and B represents H, or OH; and EATS (EBTS) is the adsorption energy of A(B) in the TS geometry without B(A). The first four terms increase the Ea, while the last two terms decrease the Ea. All the terms for the initial scission of the O‒H, C‒H, and C‒O bonds in CH3OH are presented in Table 3. Table 3. Energy Barrier and Contribution Factors (in eV) for the Initial Scission of the O‒H, C‒H, and C‒O Bonds in CH3OH Decomposition on Pt3Ni(111). initial reactions CH3OH → CH3O + H
∆Esub (eV) 0.001
∆EABdef (eV)
EABIS (eV)
EintTS (eV)
EATS (eV)
EBTS (eV)
Ea (eV)
2.98
0.83
2.11
2.54
2.69
0.69
CH3OH → CH2OH + H
‒0.04
3.09
0.83
2.40
1.82
2.83
1.63
CH3OH → CH3 + OH
‒0.02
3.43
0.83
2.46
1.25
3.37
2.08
Table 3 shows that the substrate undergoes a slight change in all reaction modes (∆Esub ranges from ‒ 0.04 to 0.001). For the three decomposition modes of CH3OH (initial scission of the O‒H, C‒H, and C‒ O bonds), EABIS has the same value of 0.83 eV for O‒H, C‒H, and C‒O scission because they involve the same IS. The EintTS for the C‒O and C‒H scission (2.46, and 2.40 eV, respectively) is greater than that for O‒H scission (2.11 eV), indicating that the repulsive interaction of the TS for the C‒O and C‒H scission (TS10 and TS9 in Figure S2) is indeed greater than that for O‒H scission (TS1 in Figure 2), which is in agreement with our structural analysis of the TS. For each C‒O bond scission, the deformation energy ∆EABdef (3.43 eV) is the largest value among the three initial scission modes, which means that structural deformation via CH3OH → CH3 + OH is strong and plays the dominant role in the high Ea for the C‒O bond scission reaction. In addition, the low value of EATS (A is CH3 here) also contributes to the high Ea of C‒O scission. These results explain why the C‒O bond scissions are too difficult to occur. ACS Paragon Plus Environment
22
Page 23 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
For C‒H and O‒H scission, the values of ∆EABdef are similar (3.09 eV vs 2.98 eV). Moreover, the value of EBTS (B represents an H atom for C‒H and O‒H scission) also shows a minor difference (2.83 eV vs 2.69 eV). This small difference probably originates from tiny changes in the adsorption energy of H at different sites. Therefore, the ∆EABdef and EBTS exert less influence on the different Ea for C‒H and O‒H scission. However, the EintTS for C‒H scission (2.40 eV) is greater than that for O‒H scission (2.11 eV), and the EATS for C‒H scission (1.82 eV, A represents CH2OH) is lower than that for O‒H scission (2.54 eV, A represents CH3O). The high EintTS and low EATS account for the high Ea (1.63 eV) for the initial C‒H bond scission compared with O‒H scission (0.69 eV).
Figure 6. Plots of the electron density difference (∆ρ) (upper panels) and the atomic positions in the TS (lower panels) for initial O‒H (TS1), C‒H (TS9), and C‒O bond scission (TS10) in CH3OH decomposition on the Pt3Ni(111) surface. The scale bar unit is 1 × 10‒2 e/Bohr3. Orange and red indicate an increase in electron density; green and blue indicate a decrease in electron density.
The interaction of the dissociated fragments (AB) with the Pt3Ni(111) surface in the TS for initial C‒ H (TS9) and C‒O (TS10) bond scission is lower than that for initial O‒H bond scission (TS1), which can also be demonstrated by the polarization effects in their TSs, as illustrated by the electron density ACS Paragon Plus Environment
23
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 24 of 38
difference (∆ρ). We calculated ∆ρ using eq 10:29 ∆ρ = ρTS ‒ ρDF ‒ ρPt3Ni
(10)
where ρTS, ρDF, and ρPt3Ni are the electron densities of the TS system (TS1, TS9, or TS10), dissociated fragments (CH3OH in TS1, CH2OH and H in TS9, and CH3O and OH on IS10), and Pt3Ni(111), respectively. The results depicted in Figure 6 show that the induced electron density redistribution in TS1 is much larger than those in TS9 and TS10, indicating interaction of the reactant with the Pt3Ni(111) surface in TS1 is much stronger than those in TS9 and TS10, in accordance with the calculated values listed in Table 3 (the values of (EATS + EBTS) are 5.23, 4.65, and 4.62 eV for initial scission of the O‒H, C‒H, and C‒O bonds, respectively). The low interaction of the reactant with the catalyst surface in TS9 and TS10 also contributes to the high Ea for scission of the C‒H and C‒O bonds. 4.4. Potential Energy Surface (PES) and Decomposition Mechanism. The PES for CH3OH decomposition on the Pt3Ni(111) surface is presented in Figure 7. To elucidate the reaction mechanisms, the rate constants (including the pre-exponential factor) of each elementary step are calculated at 300 K, as shown in Table 2.
Figure 7. PES of CH3OH dehydrogenation on the Pt3Ni(111) surface. Energies (in eV) are related to the total energy of one gaseous CH3OH molecule and the clean slab.
As shown in Figure 7, the initial C‒O and C‒H bond scission pathways in CH3OH decomposition ACS Paragon Plus Environment
24
Page 25 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
reaction do not occur because of the high Ea (2.08 and 1.63 eV for C‒O and C‒H bond scission, respectively). Compared to the initial scission of the C‒O and C‒H bonds, the initial scission of the O‒ H bond to form CH3O has an Ea of 0.69 eV and is thus more favorable. Further calculations show that the rate constant of O‒H bond scission (33.8 s‒1 at 300 K, in Table 2) is much larger than those of C‒O (5.45 × 10‒21 s‒1) and C‒H scission (3.49 × 10‒14 s‒1). For the subsequent decomposition reactions, the CH3O at the most stable adsorption site is not kinetically facile via C‒O bond scission (Ea = 1.06 eV), but the reaction via C‒H bond scission is more favorable due to the very low Ea of 0.11 eV. The relevant rate constant is 1.46 × 1010 and 8.74 × 10‒6 s‒1 for CH3O decomposition via C‒H and C‒O bond scissions, respectively, at 300 K. Before decomposition of the formed CH2O (adsorbed on Ni-top site), the adsorption structure must change to the CH2O-t configuration (adsorbed on the Pt‒Ni bridge site via η1(C)-η1(O) mode). This structure transformation occurs rapidly with an Ea as low as 0.04 eV, and the rate constant for this transformation is as high as 4.59 × 1011 s‒1. For CH2O-t decomposition, C‒O bond scission is not competitive due to the very high Ea (2.21 eV), but C‒H bond scission has a low Ea of 0.29 eV. The relevant rate constants are 1.13 × 10‒22 s‒1 for C‒O bond scission and 9.18 × 108 s‒1 for C‒ H bond scission. For CHO, C‒H bond scission to CO occurs with an Ea of 0.43 eV, which is considerably less than that for C‒O bond scission to CH and an O atom (2.23 eV). The rate constant of the C‒H bond scission is 38 orders of magnitude greater than that of C‒O scission (6.33 × 1014 s‒1 vs 1.13 × 10‒24 s‒1), suggesting that C‒H bond scission in CHO is much easier than C‒O bond scission. C‒ O bond scission cannot occur due to the very high energy barrier. Thus, the PES and rate constant analyses indicate that initial C‒H and C‒O bond scission are not competitive for CH3OH decomposition, and C‒O bond scission is not competitive for the decomposition of the intermediates involved in CH3OH decomposition. Therefore, the most favorable reaction pathway for CH3OH decomposition on Pt3Ni(111) proceeds via CH3OH → CH3O → CH2O → CHO → CO (Figure 8).
ACS Paragon Plus Environment
25
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 26 of 38
Figure 8. Illustration of the most favorable pathway of CH3OH decomposition on the Pt3Ni(111) surface proceeding via CH3OH → CH3O → CH2O → CHO → CO.
4.5. Comparisons to Other Systems. To better understand CH3OH decomposition on Pt3Ni(111), the current results are compared to those of previous studies of CH3OH decomposition on Pt(111), Ni(111), Ir(111), PtRu(111), Pt3Sn(111), and Pt5Ni2 cluster. The adsorption of CH3OH on the Pt3Ni(111) surface is strong, as revealed by its Eads of ‒0.83 eV (Table 1), which is higher than that reported on Pt(111) (‒0.33 eV),8 Ni(111) (‒0.17 eV),49 PtRu(111) (‒ 0.70 eV, on Ru top site)17, Pt3Sn(111) (‒0.57 eV, on Sn top site),22 and Ir(111) (‒0.38 eV).60 The high Eads on the Pt3Ni(111) surface indicates that desorption of the adsorbed CH3OH is not easy and thus favors decomposition. This can be verified by the fact that the Ea of the initial O‒H bond scission of CH3OH on Pt3Ni(111) (0.69 eV) is lower than its desorption energy (0.83 eV). In contrast, the initial reaction energy barrier of CH3OH decomposition on Pt(111) (0.67 eV for initial scission of the C‒H bond8 and 0.81 eV for initial scission of the O‒H bond,61 Table 2), Ni(111) (0.4 eV),49 Ir(111) (0.51 eV),60 and Pt3Sn(111) (1.01 eV)22 is much greater than the desorption energy of CH3OH on those catalyst surfaces (the desorption energy of CH3OH on Pt(111), Ni(111), Ir(111), and Pt3Sn(111) is 0.33,8 0.17,48 0.38,59 and 0.57 eV (on Sn top site),22 respectively), implying that desorption of CH3OH is ACS Paragon Plus Environment
26
Page 27 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
easier than decomposition on these catalyst surfaces. Experimental studies of CH3OH adsorption on Pt(111) have shown that more than 90% of the adsorbed CH3OH desorbs without decomposition, and only a small percentage decomposes to CO and H2.62,63 These results suggest that Pt3Ni(111) is more feasible for CH3OH decomposition than other catalysts. The initial scission reaction for CH3OH decomposition on Pt3Ni(111) is scission of the O‒H bond, which is fairly different from that on Pt(111) and Ir(111) surfaces, i.e., initial scission of the C‒H bond is favorable on Pt(111)8 and initial scission of the C‒H and O‒H bonds are both favorable on Ir(111).60 This difference can be explained by the relative stabilities of the resulting CH3O and CH2OH intermediates on the different catalyst surfaces with respect to that of gas-phase CH3OH. The adsorption of CH2OH is more stable than that of CH3O on Pt(111) (‒1.98 eV vs ‒1.54 eV).8 In the absence of energy barriers, or at equilibrium, the increased stabilities ensure that the reaction almost certainly proceeds primarily through the more stable intermediates, accounting for the preference of the initial C‒ H bond scission on Pt(111). However, the adsorption of CH3O and CH2OH on the Ir(111) surface (‒ 2.18 eV vs ‒2.01 eV)60 is nearly equally stable, which explains why the initial C‒H and O‒H bonds are both favorable for CH3OH decomposition on Ir(111). The adsorption of CH3O on the Pt3Ni(111) surface (‒2.61 eV) is more stable than that of CH2OH (‒2.24 eV); thus, initial O‒H bond scission is more competitive than initial C‒H bond scission. The same result is found on the Ni(111)49 and Pt3Sn(111) surfaces,22 which provide greater adsorption stability for CH3O than CH2OH (‒2.58 eV vs ‒1.68 eV on Ni(111);49 ‒1.71 eV vs ‒0.82 eV on the Sn top site of Pt3Sn(111)22). PES analysis (Figure 7) demonstrates that the rate-limiting step (RLS) of CH3OH decomposition on Pt3Ni(111) is abstraction of the hydroxyl hydrogen, i.e., CH3OH → CH3O + H, with Ea of 0.69 eV. The RLS is different for CH3OH decomposition on Ni(111), PtRu(111), and Pt5Ni2 cluster.17,49,64 The RLS of CH3OH decomposition on Ni(111) and PtRu(111) is abstraction of hydrogen from CH3O, i.e., CH3O → CH2O + H. The Ea of the RLS on Ni(111) and PtRu(111) are 0.89 and 1.01 eV, respectively. The RLS of CH3OH decomposition on Pt5Ni2 cluster is abstraction of hydrogen from CH2O, i.e., CH2O → CHO + H with an Ea of 1.36 eV.64 Although the RLS on Pt3Ni(111) is the same as that on the Pt3Sn(111) ACS Paragon Plus Environment
27
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 28 of 38
surface, the Ea is much lower than that on the Pt3Sn(111) surface (1.01 eV).22 These results further indicate that CH3OH is more easily decomposed on Pt3Ni(111) than on other catalyst surfaces. Many works reported that the RLS of CH3OH decomposition on metal catalyst surface is abstraction of hydrogen from CH3O,17,49,64 to confirm the reliability of our determination of the RLS of CH3OH decomposition on Pt3Ni(111), we refined the TS structures of the scission of the O‒H bond in CH3OH and the C‒H bond in CH3O and calculated the related Ea again using the DIMER approach, which is also implemented in VASP. The calculation indicates that the values of the Ea for the scission of the O‒ H bond in CH3OH and the C‒H bond in CH3O are 0.66 and 0.17 eV, respectively, which are similar to those values obtained from CI-NEB (0.69 and 0.11 eV for scission of the O‒H bond in CH3OH and the C‒H bond in CH3O, respectively). The Ea of the scission of the O‒H bond in CH3OH (0.66 eV) is much higher than that of scission of C‒H bond in CH3O (0.17 eV), suggesting that the scission of the O‒H bond in CH3OH is an RLS for CH3OH decomposition on Pt3Ni(111). These results indicate that our determination of the RLS of CH3OH decomposition on Pt3Ni(111) is reliability. CH2O, which is a saturated molecule, is an important intermediate in CH3OH decomposition. On the Pt3Ni(111) surface, it readily decomposes into CHO via C‒H bond scission, with a low Ea of 0.29 eV, which is lower than that on Pt3Sn(111) (0.75 eV),22 PtRu(111) (0.38 eV),17 Ni(111) (0.48 eV),49 and Pt5Ni2 cluster (1.36 eV).64 In addition, as desired in DMFCs, CH3OH and its intermediates act as donors to consecutively provide electrons to the catalyst surface, primarily via dehydrogenation reactions. On some Pt-based catalyst surfaces, for example PtRu(111), the CH2O and CHO intermediates can react with a surface OH group to form CH2OOH and CHOOH, respectively, suggesting that CH2O and CHO oxidation spontaneously occur more easily than further dehydrogenation to CO. Our calculations indicate that oxidation of CH2O and CHO on the Pt3Ni(111) surface is more difficult than dehydrogenation of these intermediates. For example, the Ea for the oxidation of CHO, i.e., CHO + OH → CHOOH, is 0.67 eV, which is much higher than that for the dehydrogenation of CHO, i.e., CHO → CO + H (0.43 eV). These results suggest that Pt3Ni(111) may be a good catalyst in DMFCs because CH3OH and its intermediates are apt to decompose on its surface to provide electrons to the catalyst. ACS Paragon Plus Environment
28
Page 29 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
One of the most promising applications of CH3OH decomposition is as the anodic reaction in DMFCs. However, the CO product would poison Pt-based anodic catalysts, causing significant performance loss. Figure 7 shows that the total energy of the final decomposition products CO and H (both CO and H are in an adsorbed state) is ‒2.28 eV lower than that of the IS. However, the state of CO(g) and 2H2(g) is 1.61 eV higher than the IS, indicating that the adsorbed CO may be too stable on the catalyst surface to be released into the gas phase. The calculated desorption energy of the CO adsorbed on Pt3Ni(111) is 2.05 eV. Although this value is lower than that on Ni(111) (2.31 eV),49 it is almost identical to those on Pt (111) (1.82 eV),8 Ir(111) (2.10 eV),60 and PtRu(111) (2.10 eV).17 These results imply that desorption of the CO adsorbed on the Pt3Ni(111) catalyst is relatively easier than that on Ni(111), however, it is similar to that on Pt(111), Ir(111), and PtRu(111) surface. Note that, although the desorption energy of CO adsorbed on Pt3Ni(111) is obtained under vdW correction and those values reported on Pt(111), Ni(111), Ir(111), and PtRu(111) surfaces are obtained without the vdW correction, the vdW correction contributes little to the strong adsorption,65 such as CO adsorption. For example, we calculated the adsorption energy of CO on Pt3Ni(111) to be 2.03 eV without the vdW correction. The adsorption energy of CO on Pt(111) is also calculated with and without the vdW correction, and it is 2.20 and 2.18 eV, respectively. The absorbed CO can also react with the surface OH group to form CO2, which can be desorbed from the catalyst surface easily.4,66 Our calculation showed that the adsorbed CO on the Pt3Ni(111) can be converted to CO2 by reacting with surface OH group via an intermediate of COOH. The Ea for the decomposition of COOH to CO2 is 0.92 eV, which is similar to that obtained on PtRu(111) (0.97 eV)17 and lower than that obtained on Co(0001) surface (1.01‒1.21 eV).4 These results imply that Pt3Ni(111) possesses good anti-poisoning of CO ability when it is used as an anode catalyst in DMFCs. 5. CONCLUSIONS In summary, methanol decomposition on Pt3Ni(111) has been explored using the self-consistent periodic DFT method. Several useful conclusions have been obtained. The Pt and Ni atoms on the ACS Paragon Plus Environment
29
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 30 of 38
Pt3Ni(111) surface are slightly negatively and positively charged, respectively; thus, most of the intermediates (such as radical intermediates and species with lone-pair electrons) are apt to adsorb around the Ni sites. The initial dehydrogenation of CH3OH favors O‒H bond scission rather than C‒H and C‒O bond scissions. The most probable pathway occurs via CH3OH → CH3O → CH2O → CHO → CO, in which scission of the O‒H bond is the RLS. Thus, the subsequent steps occur relatively rapidly without accumulation of intermediates. The comparison between the current results and CH3OH decomposition in other systems has demonstrated that Pt3Ni(111) can efficiently promote methanol decomposition and alleviate the CO poisoning problem in DMFCs. The results presented provide a better understanding of the PtNi anode characteristics in DMFCs and a way to improve the performance of DMFCs. ASSOCIATED CONTENT Supporting Information Supporting Information Available: Calculated vibrational frequencies of the stable intermediates of
methanol decomposition on the Pt3Ni(111) surface; slab models of Pt3Ni(111) with (2 × 2) unit cells; and structure of the TS for methanol decomposition via initial C‒O and C‒H bond scissions. This material is available free of charge on the ACS Publications website at http://pubs.acs.org.
AUTHOR INFORMATION Corresponding Authors *E-mail:
[email protected]. Tel.: 86 25 85891780. Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS
ACS Paragon Plus Environment
30
Page 31 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
This work is supported by NSFC (21335004, 21375063, 21405083, and 21675088), the Program of Outstanding Innovation Research Team of Universities in Jiangsu Province, and the Priority Academic Program Development of Jiangsu Higher Education Institutions.
REFERENCES (1) Kramer, Z. C.; Gu, X.-K.; Zhou, D. D. Y.; Li, W.-X.; Skodje, R. T. Following Molecules through Reactive Networks: Surface Catalyzed Decomposition of Methanol on Pd(111), Pt(111), and Ni(111). J. Phys. Chem. C 2014, 118, 12364‒12383. (2) Palo, D. R.; Dagle, R. A.; Holladay, J. D. Methanol Steam Reforming for Hydrogen Production. Chem. Rev. 2007, 107, 3992‒4021. (3) Wu, P.; Zhang, H.; Qian, Y.; Hu, Y.; Zhang, H.; Cai, C. Composition- and Aspect-Ratio-Dependent Electrocatalytic Performances of One-Dimensional Aligned Pt-Ni Nanostructures. J. Phys. Chem. C 2013, 117, 19091‒19100. (4) Luo, W.; Asthagiri, A. Density Functional Theory Study of Methanol Steam Reforming on Co(0001) and Co(111) Surfaces. J. Phys. Chem. C 2014, 118, 15274‒15285. (5) Gu, X. K.; Li, W. X. First-Principles Study on Origin of the Different Selectivities for Methanol Steam Reforming on Cu(111) and Pd(111). J. Phys. Chem. C 2010, 114, 21539‒21547. (6) Hung, T.-C.; Liao, T.-W.; Liao, Z.-H.; Hsu, P.-W.; Cai, P.-Y.; Lee, H.; Lai, Y.-L.; Hsu,Y.-J.; Chen, H.-Y.; Wang, J.-H.; Luo, M.-F. Dependence on Size of Supported Rh Nanoclusters in the Decomposition of Methanol. ACS Catal. 2015, 5, 4276‒4287. (7) Cortright, R. D.; Davda, R. R.; Dumesic, J. A. Hydrogen from Catalytic Reforming of BiomassDerived Hydrocarbons in Liquid Water. Nature 2002, 418, 964‒967. (8) Greeley, J.; Mavrikakis, M. Competitive Paths for Methanol Decomposition on Pt(111). J. Am. Chem. Soc. 2004, 126, 3910‒3919. (9) Lin, S.; Xie, D.; Guo, H. Pathways of Methanol Steam Reforming on PdZn and Comparison with ACS Paragon Plus Environment
31
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 32 of 38
Cu. J. Phys. Chem. C 2011, 115, 20583‒20589. (10) Choksi, T.; Greeley, J. Partial Oxidation of Methanol on MoO3(010): A DFT and Microkinetic Study. ACS Catal. 2016, 6, 7260‒7277. (11) Liu, Z. X.; Sawada, T.; Takagi, T.; Watanabe, K. Reaction Intermediates in the Oxidation of Methanol on a Pt(111)-(2×2)O Surface. J. Chem. Phys. 2003, 119, 4879‒4886. (12) Endo, M.; Matsumoto, T.; Kubota, J.; Domen, K.; Hirose, C. In Situ IRAS Observation of Catalytic Deep Oxidation of Methanol on Pt(111) Under Ambient Pressure Conditions. J. Phys. Chem. B 2001, 105, 1573‒1577. (13) Kaichev, V. V.; Prosvirin, I. P.; Bukhtiyarov, V. I. Decomposition and Oxidation of Methanol on Platinum: A Study by In Situ X-ray Photoelectron Spectroscopy and Mass Spectrometry. Kinetics and Catalysis 2014, 55, 509‒519. (14) Johnson, S.; Madix, R. J. Sulfur Induced Selectivity Changes for Methanol Decomposition on Ni(100). Surf. Sci. 1981, 103, 361‒396. (15) Knop-Gericke, A.; Hävecker, M.; Schedel-Niedrig, T.; Schlögl, R. Characterization of Active Phases of a Copper Catalyst for Methanol Oxidation Under Reaction Conditions: An In Situ X-ray Absorption Spectroscopy Study in the Soft Energy Range. Top. Catal. 2001, 15, 27‒34. (16) Chen, J. J.; Jiang, Z. C.; Zhou, Y.; Chakraborty, B. R.; Winograd, N. Spectroscopic Studies of Methanol Decomposition on Pd(111). Surf. Sci. 1995, 328, 248‒262. (17) Zhao, L.; Wang, S.; Ding, Q.; Xu, W.; Sang, P.; Chi, Y.; Lu, X.; Guo, W. The Oxidation of Methanol on PtRu(111): A Periodic Density Functional Theory Investigation. J. Phys. Chem. C 2015, 119, 20389‒20400. (18) Ishikawa, Y.; Liao, M.-S.; Cabrera, C. R. Oxidation of Methanol on Platinum, Ruthenium and Mixed Pt-M Metals (M = Ru, Sn): A Theoretical Study. Surf. Sci. 2000, 463, 66‒80. (19) Zhong, W.; Liu, Y.; Zhang, D. Theoretical Study of Methanol Oxidation on the PtAu(111) Bimetallic Surface: CO Pathway vs Non-CO Pathway. J. Phys. Chem. C 2012, 116, 2994‒3000. (20) Yuan, D.; Gong, X.; Wu, R. Decomposition Pathways of Methanol on the PtAu(111) Bimetallic ACS Paragon Plus Environment
32
Page 33 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Surface: A First-Principles Study. J. Chem. Phys. 2008, 128, 064706. (21) Desai, S. K.; Neurock, M.; Kourtakis, K. A Periodic Density Functional Theory Study of the Dehydrogenation of Methanol over Pt(111). J. Phys. Chem. B 2002, 106, 2559‒2568. (22) Lu, X.; Deng, Z.; Guo, C.; Wang, W.; Wei, S.; Ng, S.-P.; Chen, X.; Ding, N.; Guo, W.; Wu, C.M. L. Methanol Oxidation on Pt3Sn(111) for Direct Methanol Fuel Cells: Methanol Decomposition. ACS Appl. Mater. Interfaces 2016, 8, 12194‒12204. (23) Hu, Y.; Wu, P.; Yin, Y.; Zhang, H.; Cai, C. Effects of Structure, Composition, and Carbon Support Properties on the Electrocatalytic Activity of Pt-Ni-Graphene Nanocatalysts for the Methanol Oxidation. Appl. Catal., B 2012, 111-112, 208‒217. (24) Hu, Y.; Wu, P.; Zhang, H.; Cai, C. Synthesis of Graphene-Supported Hollow Pt-Ni Nanocatalysts for Highly Active Electrocatalysis toward the Methanol Oxidation Reaction. Electrochim. Acta 2012, 85, 314‒321. (25) Hu, Y.; Shao, Q.; Wu, P.; Zhang, H.; Cai, C. Synthesis of Hollow Mesoporous Pt-Ni Nanosphere for Highly Active Electrocatalysis toward the Methanol Oxidation Reaction. Electrochem. Commun. 2012, 18, 96‒99. (26) Cao, L.; Mueller, T. Theoretical Insights into the Effects of Oxidation and Mo-Doping on the Structure and Stability of Pt-Ni Nanoparticles. Nano Lett. 2016, 16, 7748‒7754. (27) Tu, K.; Lardner, M. J.; Morhart, T. A.; Rosendahl, S. M.; Creighton, S.; Burgess, I. J. Spatial Mapping of Methanol Oxidation Activity on a Monolithic Variable-Composition PtNi Alloy Using Synchrotron Infrared Microspectroscopy. J. Phys. Chem. C 2016, 120, 23640‒23647. (28) Mueller, J. E.; Krtil, P.; Kibler, L. A.; Jacob, T. Bimetallic Alloys in Action: Dynamic Atomistic Motifs for Electrochemistry and Catalysis. Phys. Chem. Chem. Phys. 2014, 16, 15029‒15042. (29) Du, P.; Wu, P.; Cai, C. Mechanistic Insight into the Facet-Dependent Adsorption of Methanol on a Pt3Ni Nanocatalyst. J. Phys. Chem. C 2015, 119, 18352‒18363. (30) Kresse, G.; Hafner, J. Ab Initio Molecular Dynamics for Open-Shell Transition Metals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 48, 13115‒13126. ACS Paragon Plus Environment
33
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 34 of 38
(31) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169‒11186. (32) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953‒17979. (33) Kresse, G.; Joubert, D. From Ultrasoft Pseudopotentials to the Projector Agumented-Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 1758‒1775. (34) Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C. Atoms, Molecules, Solids, and Surfaces: Applications of the Generalized Gradient Approximation for Exchange and Correlation PW91 Reference. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 46, 6671‒6687. (35) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865‒3868. (36) Grimme, S. Semiempirical GGA-Type Density Functional Constructed with a Long-Range Dispersion Correction. J. Comput. Chem. 2006, 27, 1787‒1799. (37) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010, 132, 154104. (38) Vasiliev, M. A. Surface Effects of Ordering in Binary Alloys. J. Phys. D: Appl. Phys. 1997, 30, 3037‒3070. (39) Yang, Z.; Wang, J.; Yu, X. The Adsorption, Diffusion and Dissociation of O2 on Pt-Skin Pt3Ni(111): A Density Functional Theory Study. Chem. Phys. Lett. 2010, 499, 83‒88. (40) Leroux, C.; Cadeville, M.; Pierron-Bohnes, V.; Inden, G.; Hinz, F. Comparative Investigation of Structural and Transport Properties of L10 NiPt and CoPt Phases. The Role of Magnetism. J. Phys. F: Met. Phys. 1988, 18, 2033‒2051. (41) Bond, G. C. The Electronic Structure of Platinum-Gold Alloy Particles. Platinum Met. Rev. 2007, 51, 63‒68. ACS Paragon Plus Environment
34
Page 35 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
(42) Lide, D. R. CRC Handbook of Chemistry and Physics, 81st Ed.; CRC Press: Boca Raton, FL, 2000; Section 9, Structure of Free Molecules in the Gas Phase. (43) Jiang, R.; Guo, W.; Li, M.; Fu, D.; Shan, H. Density Functional Investigation of Methanol Dehydrogenation on Pd(111). J. Phys. Chem. C 2009, 113, 4188‒4197. (44) Ulitsky, A.; Elber, R. A New Technique to Calculate Steepest Descent Paths in Flexible Polyatomic Systems. J. Chem. Phys. 1990, 92, 1510‒1511. (45) Mills, G.; Jonsson, H.; Schenter, G. K. Reversible Work Transition State Theory: Application to Dissociative Adsorption of Hydrogen. Surf. Sci. 1995, 324, 305‒337. (46) Henkelman, G.; Jónsson, H. Improved Tangent Estimate in the Nudged Elastic Band Method for Finding Minimum Energy Paths and Saddle Points. J. Chem. Phys. 2000, 113, 9978‒9985. (47) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901‒9904. (48) Wynne-Jones, W. F. K.; Eyring, H. The Absolute Rate of Reactions in Condensed Phases. J. Chem. Phys. 1935, 3, 492‒502. (49) Wang, G.-C.; Zhou, Y.-H.; Morikawa, Y.; Nakamura, J.; Cai, Z.-S.; Zhao, X.-Z. Kinetic Mechanism of Methanol Decomposition on Ni(111) Surface: A Theoretical Study. J. Phys. Chem. B 2005, 109, 12431‒12442. (50) Delbecq, F.; Sautet, P. Low-Temperature Adsorption of Formaldehyde on a Platinum (111) Surface. A Theoretical Study. Langmuir 1993, 9, 197‒207. (51) Zhang, S. T.; Yan, H.; Wei, M.; Evans, D. G.; Duan, X. Hydrogenation Mechanism of Carbon Dioxide and Carbon Monoxide on Ru(0001) Surface: A Density Functional Theory Study. RSC Adv. 2014, 4, 30241‒30249. (52) Xu, C.; Hou, J.; Pang, X.; Li, X.; Zhu, M.; Tang, B. Nanoporous PtCo and PtNi Alloy Ribbons for Methanol Electrooxidation. Int. J. Hydrogen Energy 2012, 37, 10489‒10498. (53) Kitchin, J.; Nørskov, J. K.; Barteau, M.; Chen, J. Modification of the Surface Electronic and Chemical Properties of Pt(111) by Subsurface Transition Metals. J. Chem. Phys. 2004, 120, 10240‒ ACS Paragon Plus Environment
35
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 36 of 38
10246. (54) Deng, Z.; Lu, X.; Wen, Z.; Wei, S.; Zhu, Q.; Jin, D.; Shi, X.; Guo, W. Decomposition Mechanism of Methylamine to Hydrogen Cyanide on Pt(111): Selectivity of the C‒H, N‒H and C‒N Bond Scissions. RSC Adv. 2014, 4, 12266‒12274. (55) Wang, H. F.; Liu, Z. P. Comprehensive Mechanism and Structure-Sensitivity of Ethanol Oxidation on Platinum: New Transition-State Searching Method for Resolving the Complex Reaction Network. J. Am. Chem. Soc. 2008, 130, 10996‒11004. (56) Nørskov, J. K.; Bligaard, T.; Logadottir, A.; Bahn, S.; Hansen, L. B.; Bollinger, M.; Bengaard, H.; Hammer, B.; Sljivancanin, Z.; Mavrikakis, M.; Xu, Y.; Dahl, S.; Jacobsen, C. J. H. Universality in Heterogeneous Catalysis. J. Catal. 2002, 209, 275‒278. (57) Cao, Y.; Chen, Z. X. Slab Model Studies of Water Adsorption and Decomposition on Clean and X- (X = C, N and O) Contaminated Pd(111) Surfaces. Phys. Chem. Chem. Phys. 2007, 9, 739‒746. (58) Deng, Z. G.; Lu, X. Q.; Wen, Z. Q.; Wei, S. X.; Liu, Y. J.; Fu, D. L.; Zhao, L. M.; Guo, W. Y. Mechanistic Insight into the Hydrazine Decomposition on Rh(111): Effect of Reaction Intermediate on Catalytic Activity. Phys. Chem. Chem. Phys. 2013, 15, 16172‒16182. (59) Lu, X. Q.; Deng, Z. G.; Chau, K. S.; Li, L. F.; Wen, Z. Q.; Guo, W. Y.; Wu, C.-M. L. Mechanistic Insight into Catalytic Oxidation of Ammonia on Clean, O- and OH-Assisted Ir(111) Surfaces. ChemCatChem 2013, 5, 1832‒1841. (60) Wang, H.; He, C.-Z.; Huai, L.-Y.; Liu, F.-Y. Decomposition and Oxidation of Methanol on Ir(111): A First-Principles Study. J. Phys. Chem. C 2013, 117, 4574‒4584. (61) Greeley, J.; Mavrikakis, M. A First-Principles Study of Methanol Decomposition on Pt(111). J. Am. Chem. Soc. 2002, 124, 7193‒7201. (62) Gibson, K. D.; Dubois, L. H. Step Effects in the Thermal Decomposition of Methanol on Pt(111). Surf. Sci. 1990, 233, 59‒64. (63) Diekhöhner, L.; Butler, D. A.; Baurichter, A.; Luntz, A. C. Parallel Pathways in Methanol Decomposition on Pt(111). Surf. Sci. 1998, 409, 384‒391. ACS Paragon Plus Environment
36
Page 37 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
(64) Guo, W.; Tian, W. Q.; Lian, X.; Liu, F.; Zhou, M.; Xiao, P.; Zhang, Y. A Comparison of the Dominant Pathways for the Methanol Dehydrogenation to CO on Pt7 and Pt7‒xNix (x = 1, 2, 3) Bimetallic Clusters: A DFT Study. Comput. Theor. Chem. 2014, 1032, 73‒83. (65) García-Muelas, R.; Li, Q.; López, N. Density Functional Theory Comparison of Methanol Decomposition and Reverse Reactions on Metal Surfaces. ACS Catal. 2015, 5, 1027‒1036. (66) Wu, P.; Du, P.; Zhang, H.; Cai, C. Graphyne-Supported Single Fe Atom Catalysts for CO Oxidation. Phys. Chem. Chem. Phys. 2015, 17, 1441‒1449.
ACS Paragon Plus Environment
37
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 38 of 38
TOC Graphic
ACS Paragon Plus Environment
38