Development of a ReaxFF Reactive Force Field for ... - ACS Publications

Subscriber access provided by HKU Libraries

Article

Development of a ReaxFF Reactive Force Field for the Pt-Ni Alloy Catalyst Yun Kyung Shin, Lili Gai, Sumathy Raman, and Adri C.T. van Duin J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.6b06770 • Publication Date (Web): 26 Sep 2016 Downloaded from http://pubs.acs.org on October 1, 2016

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 A 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 41

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

Development of a ReaxFF Reactive Force Field for the Pt−Ni Alloy Catalyst Yun Kyung Shin†, Lili Gai†, Sumathy Raman*, ‡ and Adri C.T. van Duin† †

Department of Mechanical and Nuclear Engineering, Pennsylvania State University,

University Park, Pennsylvania 16802, United States ‡ ExxonMobil

Research and Engineering Co., Annandale, New Jersey 08801, United States

ABSTRACT: We developed the ReaxFF force field for Pt/Ni/C/H/O interactions, specifically targeted for heterogeneous catalysis application of the Pt–Ni alloy. The force field is trained using the DFT data for equations of state of Pt3Ni, PtNi3 and PtNi alloys, the surface energy of the PtxNi1-x(111) (x = 0.67−0.83), and binding energies of various atomic and molecular species (O, H, C, CH, CH2, CH3, CO, OH and H2O) on these surfaces. The ReaxFF force field shows a Pt surface segregation at x ≥ 0.67 for the (111) surface and x ≥ 0.62 for the (100) surface in vacuum. In addition, from the investigation of the preferential alloy component of the adsorbates, it is expected that H and CH3 on the alloy surface to induce a segregation of Pt whereas the oxidation intermediates and products such as C, O, OH, H2O, CO, CH and CH2 are found to induce Ni segregation. The relative order of binding strengths among adsorbates is a function of alloy composition and the force field is trained to describe the trend observed in DFT calculations, namely, H2 < H2O < CH3 ≈ O2 ≈ CO < OH < CH2 < C ≈ CH on Pt8Ni4; H2 < H2O < CO < O2 ≈ CH3 < OH < CH2 < CH < C on Pt9Ni3; and H2 < H2O < O2 < CO < CH3 < OH < CH2 < C ≈ CH on Pt10Ni2. Using this force field, we performed the grandcanonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations for a Pt3Ni slab 1 ACS Paragon Plus Environment


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

and a truncated cuboctahedral nanoparticle terminated by (111) and (100) faces, to examine the surface segregation trend under different gas environments. It is found that Pt segregates to the alloy surface when the surface is exposed to vacuum and/or H2 environment while Ni segregates under the O2 environment. These results suggest that the Pt/Ni alloy force field can be successfully used for the preparation of Pt–Ni nanobimetallic catalysts structure using GCMC and run MD simulations to investigate its role and the catalytic chemistry in catalytic oxidation, dehydrogenation and coupling reactions. The current Pt/Ni force field still is found to have difficulties in describing the observed segregation trend in Ni–rich alloy compositions (x < 0.6), suggesting the need for additional force field training and evaluation for its application to describe the characteristics and chemistry of Ni–rich alloys.

INTRODUCTION Platinum is one of the most important and efficient metal catalysts. It has been applied to many industrial chemical processes, including the oxygen reduction reaction (ORR)1 occurring at the cathode of the fuel cells, the hydrogenation and dehydrogenation of hydrocarbons in petroleum refining industries,2,

3

and the CO oxidation/NOx reduction

reactions to reduce harmful emissions in many industrial processes.4-6 However, Pt as a noble metal is extremely expensive and limited in availability. Consequently, Pt−based alloys (Pt−M, M = Ni, Cu, Co, Cr, V, Fe) are being increasingly explored for their applications in catalysis.7-10 The surface properties and structures of the Pt−Ni alloy have been studied to design alternative catalysts with low cost and high catalytic activity and stability. The chemical 2 ACS Paragon Plus Environment

Page 2 of 41

Page 3 of 41

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


behavior of the alloy surfaces, in general is different from that of the pure metals. Hence, they can modify the catalytic performance of the alloy catalyst and may even have enhanced activity compared to that of pure Pt or Ni catalysts. The catalytic activity for the ORR on Pt3Ni(111) is found to be significantly enhanced, compared to pure metals, due to the surface-sandwich segregation resulting in a modified electronic structure of the platinum on the alloy surface.11 In addition, noble metals (Pd, Pt, Au, Rh and Ru) promoted Ni/α−Al2O3 catalysts are effective for oxidative steam reforming of methane (CH4 + H2O → CO + 3H2), compared to the monometallic Ni catalyst.12 The catalytic behavior of the alloy is known to be associated with its surface chemical composition and the relative adsorption strength of the reactants and products, including the intermediates for oxidation and reduction.13, 14 The surface composition of PtNi nanoparticle prepared under wet condition is found to change, depending on the adsorbing gaseous species. In the presence of O2 species, Ni segregates to the surface of PtNi nanoparticle, whereas Pt segregates under the H2 or the vacuum environment due to the difference of their affinity to Pt and Ni. Thus it is of great importance to study the correlation between the surface composition and the adsorption strength in the presence of various adsorption species on the Pt−Ni alloy catalysts. We recently published a ReaxFF reactive force field for Pt/C/H/O15 by coupling a Pt/O force field16 with Pt/H and Pt/C parameters, and evaluated the performance of this force field to predict the adsorption structure and isotherms of oxygen, hydrogen and OH over different surfaces of platinum and on nanoparticles of different sizes and shapes. We have earlier published a validated Ni/C/H/O force field17, 18 and our goal in this work is to develop a modular Pt/Ni bimetallic alloy parameters and couple them with Pt/C/H/O and 3 ACS Paragon Plus Environment


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

Ni/C/H/O force fields to obtain a bimetallic reactive force field for Pt/Ni/C/H/O, which is capable of describing the hydrocarbon conversion on a bimetallic catalyst. It is well accepted that ab initio or first-principles based DFT predictions are valuable to guide synthesis of new and/or improved catalytic materials by providing fundamental properties of reaction energetics and kinetics together with the stability of the catalyst.19-21 However, the DFT method is computationally expensive and is still limited to small system sizes of tens of Angstroms and a very short time scale of few picoseconds. ReaxFF which is trained against the first principles data set, is computationally less expensive than DFT and more efficient for a large-scale molecular dynamics simulation, and is uniquely capable of describing chemical reactions at a low computational cost but of reasonable accuracy. Importantly, bimetallic systems are associated with reactive surface segregation in the presence of adsorbates which could be addressed using reactive molecular dynamics or hybrid Monte Carlo−Molecular dynamics methods more readily in comparison to theoretically rigorous and computationally involved cluster expansion approaches22 at the DFT level to account for the effect of adsorbates23, 24 on surface segregation and ordering of alloy surfaces. Though cluster expansion approaches coupled with Monte Carlo simulations25-27 are increasingly being employed, reactive molecular dynamics or hybrid Monte Carlo−Molecular dynamics methods provide added advantage for modeling a large−scale complex alloy surface of which chemical composition and structure may not be well defined, and for examining reaction processes on it. In addition, this work also highlights the efforts and challenges involved in developing a reactive force field that is capable of describing all possible bonding scenarios among five atoms (Pt, Ni, C, H and O). As is shown below, the force field developed in this work performs reasonably well with a 4 ACS Paragon Plus Environment

Page 4 of 41

Page 5 of 41

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


Pt to Ni ratio of greater than 0.6. The following section presents the training set employed for a force field fitting. The ReaxFF training set includes equations of state of Pt–Ni alloys (Pt3Ni, PtNi and PtNi3), the surface energy of PtxNi1-x(111) (x = 0.67 – 0.83), and adsorption data for atomic (C, H and O) and molecular species (CH, CH2, CH3, OH, H2O and CO) binding on Pt−Ni(111) alloy surfaces calculated at the DFT level of theory. Subsequent section presents specific details of the ReaxFF force field parameterization by representing the adsorption behavior such as binding energies and site preferences between various Pt−Ni alloy surfaces. The Pt/Ni ReaxFF force field is verified in the subsequent section by performing the hybrid (GC)MC/MD simulations for investigation of the surface composition of the Pt3Ni slab and the truncated cuboctahedral nanoparticle in vacuum and in the presence of adsorbate (O and H).

COMPUTATIONAL METHODS ReaxFF Method. The ReaxFF force field is a bond order dependent force field with instantaneous connectivity for the chemical bonds depending on the atomic local environment. The great advantage of ReaxFF is that it is reliable to describe the energetics for various reaction intermediates along a reaction path.28 The total system energy is described by physically meaningful many-body empirical potential terms: bond orderdependent energy terms such as bond, angle and torsion, and long range interaction terms such as van der Waals and Coulomb interactions. A distance-corrected Morse-potential is used to properly describe the short range interactions for the van der Waals energy. For the Coulomb interaction, ReaxFF employs the electronegativity equalization method (EEM) and 5 ACS Paragon Plus Environment


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

takes into account the shielding between two atoms at short distances. All force field parameters describing energy terms are optimized against the DFT data using a singleparameter based parabolic extrapolation method. More details about the ReaxFF force field are given in our earlier publications.29-31 DFT Calculations. The spin-polarized first-principles DFT calculations were performed using the Vienna ab initio simulation package (VASP)32. The DFT calculations use projector augmented wave (PAW)33 pseudo−potentials and Perdew−Burke−Ernzerhof (PBE)34 exchange−correlation functional with an energy cutoff of 600 eV. The Brillouin zone is sampled with an (8 × 8 × 1) Monkhorst−Pack k−point mesh. For the calculation of the surface energy of Pt–Ni alloys at various compositions (Pt8Ni4, Pt9Ni3 and Pt10Ni2), a (2 × 2) unit cell is used to construct a three layer Pt−Ni(111) slab. The models used for the calculations are shown in Figure 1. A set of nine slab models is used to calculate the alloy surface energy and adsorbate binding energy. These slabs are generated by cleaving a periodic supercell of fcc Pt3Ni along the (111) surface plane and adding a 20 Å thick of vacuum space. In these structures, ionic coordinates of all atoms of the slab were allowed to fully relax while the lattice parameters remained fixed to those of the bulk (a = 3.859 Å). The surface with Pt at.% below the Pt composition of the slab indicates Ni–segregation and is generated by exchanging the surface Pt and subsurface Ni atom while the composition of the bottom layer is fixed to its parent crystal phase, Pt3Ni. Similarly, the surface with Pt at.% above the Pt composition of the slab (Pt–segregation) is generated by exchanging the surface Ni and subsurface Pt atom. For example, in Pt3Ni slab models, three slabs were generated at different compositions of surface and subsurface: 50 at.% Pt on surface and 100 at.% Pt on subsurface, 75 at.% Pt on surface and 75 at.% Pt on subsurface, 100 at.% Pt 6 ACS Paragon Plus Environment

Page 6 of 41

Page 7 of 41

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


on surface and 50 at.% Pt on subsurface. However, the chemical composition and structure of the real alloy surfaces are much more complex, depending on different chemical environments, and inherently challenging to be fully characterized.26,

35

Thus readers

should be aware that these simplified models are chosen in order to generate a dataset of different compositions of alloys at a reduced computational cost and to be used for the force field development. For calculating the binding energy of a gaseous species, a slab for each adsorption site is generated based on a (2 × 2) unit cell. This slab model corresponds to a surface coverage of 0.25 ML when there is one adsorbate per unit cell. Adsorption is allowed on only one of the two surfaces and all three layers are relaxed. For the calculations of hydrogen binding energy on (111) surface of Pt, Ni and Pt3Ni, a seven layer slab is used. Seven–layer slabs are generated by employing a 20 Å thick of vacuum space to a periodic supercell of fcc Pt, Ni and Pt3Ni crystals cleaved along the (111) surface plane. Ionic coordinates of top four layers were allowed to fully relax while the bottom three layers were fixed. The lattice parameters were kept constant to those of the bulk. For the calculation of PtxOx and PtxO2x clusters (x = 1 – 7 ), a periodic supercell of 20 × 20 × 20 Å3 is created and k-points are generated with a single gamma point-centered mesh. Atomic coordinates of all atoms in the clusters were fully relaxed to obtain minimum energy structures.

7 ACS Paragon Plus Environment


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

Figure 1 Side views of (2 × 2) Pt–Ni(111) models used for the DFT calculations at different compositions of alloys, (a) Pt8Ni4, (b) Pt9Ni3 and (c) Pt10Ni2. A vacuum space of 20 Å thick is employed to create alloy surfaces. Various alloy surface compositions were constructed in the ranges from 25 at.% Pt to 100 at.% Pt. Adsorbates are adsorbed on the top, bridge, fcc and hcp sites at the coverage of 0.25 ML. RESULTS AND DISCUSSION ReaxFF Force Field Development. Based on our earlier published Pt/C/H/O15, 16, 36 and Ni/C/H/O17, 18 metal force fields, in the present work we develop the Pt/Ni alloy force 8 ACS Paragon Plus Environment

Page 8 of 41

Page 9 of 41

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


field combined with C, H and O. The Pt/Ni bonded interaction, van der Waals long range interaction terms, and the valence angle parameters for all combinations of Pt/Ni/C, Pt/Ni/H and Pt/Ni/O angles of the alloy are parameterized using the DFT data on equations of state of Pt–Ni alloys (fcc Pt3Ni, fcc PtNi3 and bcc PtNi), alloy surface energies and the adsorption energies of various species (C, H, O, CO, CH, CH2, CH3, OH and H2O). We first fit the Pt/Ni alloy force field with the equations of state for the alloy crystal structures in the volume range from 80% to 120% of the optimal volume as plotted in Figure 2. As shown in Figure 2, the heats of formation for Pt–Ni alloys are placed close to each other with small negative formation energies (≤−5 kcal/mol) and the Pt3Ni phase is slightly overestimated in ReaxFF. The energy−volume relationship of the alloys is relatively well reproduced in ReaxFF.

Figure 2 Heats of formation as a function of volume for Pt3Ni (fcc), PtNi (bcc) and PtNi3 (fcc) alloys at 0 K in ReaxFF and QM.



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 41

Figure 3 Surface energy (per atom) for the (111) surfaces of PtxNi1-x alloy at (a) x=0.67, (b) 0.75 and (c) 0.83. The fcc Pt and fcc Ni phases are used as reference states. The color blocks indicate the surface energy regimes in which the Pt segregation can occur in ReaxFF. When the surface Pt at.% is higher than the Pt composition of the slab (× in blue), it represents the Pt segregation. When the surface Pt at.% is lower than the Pt composition of the slab, it represents the Ni segregation. The snapshot of the slab (as side view) is taken from the most stable surface for each alloy composition. Yellow and cyan atoms represent Pt and Ni, respectively. In Pt–Ni alloy, it is experimentally and theoretically reported that Pt segregates at the atomically clean alloy surfaces (111, 100) over wide range of alloy compositions in vacuum.20, 37, 38 Surface segregation of alloy depends upon various factors such as the metal cohesive energies, the relative atomic sizes of the alloy components, the interatomic interactions between alloy components and alloy surface energies. First, the cohesive energy of Pt (133.1 kcal/mol in ReaxFF) is higher than that of Ni (102.7 kcal/mol in ReaxFF). Consequently, Pt has a higher surface energy, suppressing Pt segregation to the surface. In addition, it is expected that the strain relaxation effect promotes the segregation of Pt impurity due to the atomic size of Pt being larger than that of Ni. However, this factor would be less effective in Pt−rich alloys. Thus, these two factors do not favorably support the Pt segregation over the range of Pt–Ni alloy compositions investigated in this work and earlier studies. The interplay between surface energy and interatomic interactions26, 27 of 10 ACS Paragon Plus Environment

Page 11 of 41

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


alloy is also extensively used to predict the surface segregation behavior since surface composition at thermodynamic equilibrium is driven mainly by lowering the surface energy, which is closely related with the surface composition and relaxing the bulk strain energy. It has been reported that the surface segregation is strongly influenced by the interatomic interactions between alloy components (A, B), such as homonuclear (A–A, B–B) bonding and heteronuclear (A–B) bonding. The favorable bonding type may drive the atomic rearrangements of the surface structure to form a thermodynamically more stable configuration. To investigate the effect of these factors on the surface segregation, we calculated the surface energy at various combinations of surface and subsurface compositions in Pt−Ni alloys and compared to the DFT data. Figure 3 shows the surface energy per unit atom for the (111) surface of PtxNi1-x alloys at x = 0.67, 0.75 and 0.83 (Pt8Ni4, Pt9Ni3 and Pt10Ni2, respectively). It is observed in this alloy composition range that the surface energy decreases as the Pt concentration on the surface increases at the expense of the subsurface Pt. This indicates that Pt rather than Ni would be expected to segregate to the surface. It is noted that the segregation−induced enrichment of Pt is less substantial in the Pt−Ni alloy at x = 0.67 compared to the Pt–rich alloys and the amount of Pt segregation to the surface is underestimated in ReaxFF. Similarly, we also examined the surface energy of (100) surfaces. Four−layer slabs were generated by cleaving a (2 × 2 × 2) periodic supercell of fcc Pt3Ni along the (100) surface plane and adding a vacuum space of around 20 Å thick (Figure 4). As shown in Figure 5, the surface energy of the Pt−segregated (100) surface is lower than the non-segregated surface in ReaxFF, which is in qualitative agreement with experimental observations. The surface energy of Pt–Ni alloys correlated with surface segregation as shown in Figure 3 and 5 11 ACS Paragon Plus Environment


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 41

suggests that Pt surface segregation is mainly determined by the Pt−Ni alloy surface energy which is related with the surface and subsurface alloy compositions, and the atomic arrangement, rather than the atomic size or the metal cohesive energy. We note, however, that the Pt segregation is not observed in the PtxNi1-x alloys at x < 0.6.

Figure 4 A four-layer Pt3Ni(100) slab used for generating various slabs with different compositions of surface and subsurface. A 20 Å thick of vacuum space is employed to create alloy surface.


Page 13 of 41

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


Figure 5 Surface energy (per atom) for the (100) surfaces of PtxNi1-x alloy at (a) x=0.625, (b) 0.69, (c) 0.75 and (d) 0.81 in ReaxFF. The fcc Pt and fcc Ni phases are used as reference states. The color blocks indicate the surface energy regimes in which the Pt segregation can occur in ReaxFF. When the surface Pt at.% is higher than the Pt composition of the slab (× in blue), it represents the Pt segregation. When the surface Pt at.% is lower than the Pt composition of the slab, it represents the Ni segregation. The snapshot of the slab (as top view) is taken from the most stable surface for each alloy composition. Yellow and cyan atoms represent Pt and Ni, respectively.

The segregation preference and extent are found to depend on the adsorption conditions and the environment. Different environments, e.g., O2, H2 or CO lead to different segregation characteristics. The segregation of one component of the bimetallic alloy can increase the adsorption of the gas species when the interaction between this alloy component and the gas species is more attractive than that between second component and the gas species. With the order reversed, the adsorbate can selectively enhance the particular segregation. The alloy component, say A, which segregates at the clean alloy surface can be replaced/replenished by the other component, say B, when the preferential attractions exist between the adsorbate and the component, B. Thus, the surface composition under reaction conditions may be different from that of the clean surface or under vacuum and could be made cyclic by changing the reaction condition. As is well known, with O adsorption, the Ni−rich surface is thermodynamically more 13 ACS Paragon Plus Environment


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 41

stable than the Pt−rich surface because the O adsorption energy gained from Ni is able to compensate for the loss of the segregation energy due to Ni−segregation. The binding behavior of various gaseous species has been studied on Pt or Ni surface.19, 39-41 Since we are interested in the catalytic oxidation and reduction on the Pt–Ni alloy surface, we investigated the adsorption behavior for the most commonly studied species on the Pt– Ni(111) facet: atomic hydrogen (H), oxygen (O) and carbon (C) are the most extensively studied adsorbates in catalytic reduction and oxidation processes. In addition, the heterogeneous catalytic conversion of hydrocarbon on the Pt/Ni alloy catalyst offers various intermediates and products, including hydroxyl (OH), water (H2O), carbon monoxide (CO) and CHx fragments (x=1−3). The binding energy (‫ܧ‬௕ (‫ ))ܣ‬is calculated with respect to a clean Pt−Ni(111) slab (‫ܧ‬௦௟௔௕ ) and the respective adsorbate (‫ܧ‬஺ ) in the gas phase. A E b ( A) = E slab − E slab − E A

The O2, H2, C, CHx, CO, OH and H2O species are used as reference states of the adsorbates. A negative binding energy indicates that a bound adsorbate is thermodynamically favorable and a more negative binding energy signifies a more strongly bound adsorbate to the surface site. Atomic hydrogen (H) is the least strongly bound atomic species in this study. Various binding sites available on Pt3Ni(111) surface are shown in Figure 6. As listed in Table 1, all the surface binding sites except for Ni top site appear to be stable for H adsorption in the binding energy range of –9.2 (bridge) to –13.4 (fcc) kcal/mol in ReaxFF and exhibit no significant binding energy difference between the surface sites which have Ni as a nearest


Page 15 of 41

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


neighbor (Ni) and/or only Pt as a nearest neighbor (Pt). This may indicate that the hydrogen has no strong preference for a particular adsorption site. Since the top(Ni) site is a thermodynamically unfavorable site (5.6 kcal/mol in ReaxFF) on the alloy surface, it leads to more thermodynamically favorable sites for H adsorption on Pt–rich surface than on Ni– rich surface. This may weakly drive the Pt–segregation at high H coverage, assuming all available binding sites are occupied, despite no strong preferential attractions between Pt and the adsorbate hydrogen atom. In addition, the top site on Ni would be rarely visited during the hydrogen surface diffusion because the binding on Ni top site is thermodynamically less stable and has a positive binding energy.

Figure 6 Surface sites for hydrogen binding on Pt3Ni(111). The (Ni) and (Pt) indicate that hydrogen atom has Ni as nearest neighbors and only Pt as nearest neighbors, respectively.



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 41

Table 1 Binding energy of H (0.25 ML) at various binding sites on Ni(111), Pt(111) and Pt3Ni(111) surface in ReaxFF and DFT (in kcal/mol). The Pt3Ni(Ni) and Pt3Ni(Pt) indicate that the hydrogen atom has Ni as a nearest neighbor and only Pt as nearest neighbors, respectively. The molecular hydrogen (H2) and a clean slab surface are used as reference states. ReaxFF

DFT Pt3Ni

site

Ni

Pt3Ni

Pt

Ni (Ni)

(Pt)

Pt (Ni)

(Pt)

fcc

−11.9

−10.2

−10.1

−13.4

−13.8

−10.3

−12.4

−11.7

hcp

−13.0

−8.0

−10.5

−11.3

−13.4

−9.9

−11.5

−11.4

bridge

−11.0

−9.1

−9.9

−9.2

−10.3

−9.9

−9.5

−11.3

3.8

−10.4

5.6

−12.9

0.22

−12.2

0.95

−10.3

top

The binding energy of oxygen (‫ܧ‬௕ (O)) is calculated for the fcc and hcp sites of Pt8Ni4, Pt9Ni3 and Pt10Ni2 alloys. Changes in oxygen binding energies as a function of the surface Pt concentration are shown in Figure 7a. In three alloys, the oxygen atom is most strongly bound to fcc site on Ni−rich surface (−59.3 (Pt8Ni4), −52.6 (Pt9Ni3), −33.0 (Pt10Ni2) kcal/mol in ReaxFF and −54.3, −47.9, −38.5 kcal/mol in DFT, respectively) while the oxygen is weakly bound to the hcp site of the Pt−rich surface. We note that the surface adsorption behavior at either extreme end (25 at.% Pt or 100 at.% Pt) is expected to be dominated by the majority element and the binding energy is close to the value of pure metal. Thus, under an oxidizing environment, it is expected that the adsorption of O2 on Pt−Ni alloy surface induces the Ni segregation due to the adsorption energy gain and lowers the surface concentration of Pt below its bulk concentration. It has also been reported experimentally that the presence of surface oxygen can induce Ni segregation in Pt−Ni alloys.38 Similarly, the OH species binds most strongly to the Ni−rich surface in all three alloys: – 16 ACS Paragon Plus Environment

Page 17 of 41

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


83.9 (Pt8Ni4), −82.2 (Pt9Ni3), −57.9 (Pt10Ni2) kcal/mol in ReaxFF and −82.3, −77.4, −64.6 kcal/mol in DFT, respectively (Figure 7b). It is found that in ReaxFF, the bridge site in Pt10Ni2 is also stable (−61.7 kcal/mol). On the other hand, the binding behavior of H2O species is different from that of H, O and OH species (Figure 7c). It is shown in three alloys that H2O binds strongly to the Ni rich surface at 50−75 at.% Pt with binding energies of −18.5 (Pt8Ni4), −16.3 (Pt9Ni3) and −16.4 (Pt10Ni2) kcal/mol in ReaxFF (−17.2, −18.2 and −17.8 kcal/mol, respectively in DFT), instead of showing strong preference for a particular alloy component as a function of Pt at.% on the surface. Since the H2O species has a relatively low binding energy, this oxidation product probably will be removed easily from the surface site and removed more easily from the Pt site, in particular if the species diffusion happens to occur from the Ni–rich surface sites to the Pt–rich surface sites.



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 41

Figure 7 Binding energy of (a) O, (b) OH, (c) H2O and (d) CO on (111) surface of Pt8Ni4 (left),Pt9Ni3 (middle) and Pt10Ni2 (right) alloys. The surface composition varies from 25 to 100 at.% Pt. The Pt at.% in gray represents a Pt−segregation surface. (F: fcc, H: hcp, B: bridge and T: top) 18 ACS Paragon Plus Environment

Page 19 of 41

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 CO species, an oxidation product of hydrocarbons binds to the alloy surface via a lone pair of the carbon atom. It is shown that the preferred surface for CO adsorption is the Ni−rich surface: the hcp site is preferred in Pt8Ni4, (–60.8 kcal/mol in ReaxFF and −48.7 in DFT) and the bridge site is preferred in Pt9Ni3 (−50.7 kcal/mol in ReaxFF and −49.6 kcal/mol in DFT). As shown in Figure 7d, a large discrepancy between DFT and ReaxFF occurs when CO binds to the Ni−rich surface (25 at.% Pt surface) in Pt8Ni4. As discussed earlier in the oxygen binding behavior, the CO binding at this surface is dominated by Ni element. The current ReaxFF force field overestimates the binding energy of CO on Ni(111) surface, e.g., –46.0 kcal/mol (ReaxFF) vs. –39.3 kcal/mol (DFT) at the hcp site. Because of the modular construction of the alloy force field based on earlier published force fields of the metals, namely, Ni/C/H/O and Pt/C/H/O, this overestimation of the CO binding energy on Ni–rich surface is inevitable. It is possible that this deviation causes the change in the relative strength between the adsorbates of similar binding strengths (e.g., O2, CO and CH3) on Ni−rich surfaces. In Pt10Ni2, CO prefers to bind at the bridge site in ReaxFF (−43.5 kcal/mol), however the hcp site is favored at the DFT level (−46.7 kcal/mol). As shown in Figure 7d, at both ReaxFF and DFT levels, the CO species tends to bind weakly to the Pt−rich surface, causing Ni to segregate in the place of Pt. The binding strength for atomic C and CHx (x = 1–3) species increases with decreasing hydrogen (Figure 8). Carbon atom tends to bind strongly to the fcc site on Ni−rich surface of Pt8Ni4, Pt9Ni3 and Pt10Ni2 alloys with the binding energy of −153.6, –149.2 and −150.2 kcal/mol in ReaxFF and −172.0, –166.8 and −166.4 kcal/mol in DFT, respectively, while binding weakly to the Pt−rich surface. We note, however that in ReaxFF, the carbon binding to the 75 at.% Pt surface in Pt9Ni3 alloy (fcc, −156.0 kcal/mol) is strong as well. According 19 ACS Paragon Plus Environment


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 41

to these carbon binding energy results, carbon adsorption on the alloy surface is expected to drive Ni segregation weakly. Similarly, the CH species which is the most strongly bound species among CHx species, binds strongly to the hcp site on Ni−rich surface: −155.3 (Pt8Ni4), −151.6 (Pt9Ni3), −152.3 (Pt10Ni2) kcal/mol in ReaxFF and −156.3, −158.3, −160.1 kcal/mol in DFT, respectively. As shown in Figure 8c, the CH2 species binds to both the Ni−rich surface (hcp site, –95.4 kcal/mol in ReaxFF, −99.6 kcal/mol in DFT) and the Pt−rich surface (bridge site, –97.7 kcal/mol in ReaxFF, −100.0 kcal/mol in DFT) in Pt8Ni4, and has no strong preferential attractions to a particular alloy component. On the other hand, in Pt9Ni3 and Pt10Ni2 alloys, the CH2 species shows preference for the Ni−rich surface to the Pt−rich surface. In contrast to C, CH and CH2 species, the CH3 species tends to be relatively unstable on Ni−rich alloy surface. The top site on the 100 at.% Pt surface is preferred by CH3 with binding energies of –58.9 (Pt8Ni4), –58.2 (Pt9Ni3) and −55.7 (Pt10Ni2) kcal/mol in ReaxFF. On the other hand, in DFT, CH3 binds most strongly at the 75 at.% Pt surface with binding energies of −48.4 and –49.0 kcal/mol in Pt8Ni4 and Pt9Ni3 alloy, and at 100 at.% Pt with the binding energy of –44.6 kcal/mol in Pt10Ni2. This indicates that CH3 adsorption drives Pt segregation more strongly. We also note that in ReaxFF, CH3 binding energy is overestimated on the 100 at.% Pt surface where the Pt metal character is dominant. Although there are some discrepancies in the most preferred binding sites between DFT and ReaxFF, we note that the relative binding strength and the binding preference between the C and CHx series are in agreement with DFT. The binding energies of the species at the most stable site in Pt8Ni4, Pt9Ni3, Pt10Ni2 alloys are summarized in Table 2.


Page 21 of 41

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


Figure 8 Binding energy of (a) C, (b) CH, (c) CH2 and (d) CH3 on (111) surface of Pt8Ni4 (left),Pt9Ni3 (middle) and Pt10Ni2 (right) alloys. The surface composition varies from 25 to 100 at.% Pt. The Pt at.% in gray represents a Pt−segregation surface. (F: fcc, H: hcp, B: bridge and T: top) 21 ACS Paragon Plus Environment


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 41

Table 2 Binding energy of gas phase atomic (C, O) and molecular species (CH, CH2, CH3, OH, H2O and CO) at the most stable binding site and surface Pt concentration in Pt8Ni4, Pt9Ni3 and Pt10Ni2 alloys. Ni segregation indicates when the surface Pt concentration is below the Pt concentration of the slab and Pt segregation indicates when the surface Pt concentration is above the Pt concentration of the slab. Alloy

Pt at.% on surface

Site

DFT

ReaxFF Segregation

O Pt8Ni4

25

fcc

-54.3

-59.3

Ni

Pt9Ni3

50

fcc

-47.9

-52.6

Ni

Pt10Ni2

75

fcc

-38.5

-33.0

Ni

OH Pt8Ni4

25

fcc

-82.3

-83.9

Ni

Pt9Ni3

50

fcc

-77.4

-82.2

Ni

Pt10Ni2

75

fcc

-64.6

-57.9

Ni

bridge

-61.7 H2O

Pt8Ni4

50

top

-17.2

-18.5

Ni

Pt9Ni3

75

top

-18.2

-16.3

Pt10Ni2

75

top

-17.8

-16.4

Ni

CO Pt8Ni4

25

hcp

-48.7

-60.8

Ni

Pt9Ni3

50

bridge

-49.6

-50.7

Ni

Pt10Ni2

75

hcp

-46.7

-41.6

Ni

bridge

-43.5 C

Pt8Ni4

25

fcc

-172.0

-153.6

Ni

Pt9Ni3

50

fcc

-166.8

-149.2

Ni

75

fcc

75

fcc

Pt10Ni2

-156.0 -166.4

-150.2

Ni

-156.3

-136.5

Ni

-155.3

Ni

CH Pt8Ni4

25

hcp

50

hcp

Pt9Ni3

50

hcp

-158.3

-151.6

Ni

Pt10Ni2

75

hcp

-160.1

-152.3

Ni

CH2


Page 23 of 41

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


Pt8N4

25

hcp

-99.6

75

bridge

-100.0

100

bridge

Pt9Ni3

50

hcp

Pt10Ni2

75

fcc

-95.4

Ni Pt

-97.7

Pt

-101.3

-107.7

Ni

-97.4

-97.8

Ni

-48.4

-55.4

Pt

-58.9

Pt

CH3 Pt8N4 Pt9Ni3 Pt10Ni2

75

top

100

top

75

top

100

top

100

top

-49.0 -44.6

-55.8 -58.2

Pt

-55.7

Pt

The overall adsorption energetics in ReaxFF agrees well with the DFT data. In ReaxFF, the binding strength for the adsorbates on the Pt−Ni (111) surface is ordered as follows. H2 < H2O < CH3 ≈ O2 ≈ CO < OH < CH2 < C ≈ CH

in Pt8Ni4

H2 < H2O < CO ≈ O2 < CH3 < OH < CH2 < CH < C

in Pt9Ni3

H2 < H2O < O2 < CO < CH3 < OH < CH2 < C ≈ CH

in Pt10Ni2

Since the binding energies of adsorbates depend on their coverages,42-44 it is possible that the relative binding strength between adsorbates of similar binding energies can be listed in a different order at different coverages. It is worthwhile to mention that the order listed above is examined at 0.25 ML to identify the relative stability of adsorbates. From a qualitative point of view, the adsorption of H and CH3 may induce the low surface concentration of Ni (Pt−segregation), or vice versa. On the other hand, oxidation intermediates and products of hydrocarbon (O, OH, H2O, CO, CH and CH2) increase the Ni concentration on the Pt−Ni alloy surface (Ni−segregation). The atomic C binds much more strongly than the other adsorbates with preferential attraction towards Ni. The interplay 23 ACS Paragon Plus Environment


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 41

between surface composition and adsorbates will be critical in determining the catalytic activity and selectivity of the alloy. The overall binding energy strength may demonstrate whether the Pt–Ni alloy catalyst could be successfully used for the catalytic oxidation and dehydrogenation reactions of hydrocarbon. Segregation of the Pt3Ni Alloy Slab and Nanoparticle. Pt−Ni alloys are considered as one of the promising catalysts for replacement of Pt since they show higher reactivity with less usage of Pt.7, 45 Surface composition and geometrical arrangements of the bimetallic catalysts play a key role in heterogeneous catalytic process, which always have a potential to change with reaction conditions.9, 46, 47 Thus, in this section, we investigate the atomic composition of the Pt–Ni alloy surface under different gaseous environments (vacuum, O2 and H2). Two systems of Pt3Ni alloys are chosen to demonstrate their segregation behavior, i.e., a low-index extended Pt−Ni alloy slab with (111) surface and a truncated cuboctahedron nanoparticle which has (100) face as well as (111) face. The slab system is composed of 15 Pt–Ni layers with 36 atoms per each layer. The orthogonal fcc lattice for Pt3Ni alloy is constructed with an optimized Pt−Ni lattice spacing determined from ReaxFF (a = 3.859 Å). The dimension of the simulation box is 16.4 × 14.2 × 50 Å3. Initially, the Pt and Ni atoms are distributed randomly over the slabs. The truncated cuboctahedron consists of 439 Pt and 147 Ni atoms with eight hexagonal (111) faces and six square (100) faces. In addition, there are edges of four atoms and corner sites of lower coordination. The diameter of the nanoparticle is around 24 Å and the cell dimension is set to 45.0 × 45.0 × 45.0 Å3. The GCMC/MD simulations are performed at T = 1200 K and PO2 = 10-4 atm in O2 environment and PH2 = 105 atm in H2 environment for the slab system. For the alloy nanoparticle, the simulation is performed at T = 600 K and PO2 = 10-23 atm and PH2 = 10-5 24 ACS Paragon Plus Environment

Page 25 of 41

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


atm. Although it is also of great interest to study the adsorption isotherm of adsorbates at low temperature, many previous experiments9, 46, 48-50 and simulations51, 52 on Pt/Ni alloy surfaces have been carried out at high temperature. Thus to compare with both experimental and theoretical studies, we also performed the GCMC/MD simulations at high temperature and investigated the effect of adsorbates on surface segregation. The chemical potentials of oxygen and hydrogen under these T and P conditions are calculated using thermodynamic tables.53 In GCMC/MD simulation, two alloy components (Pt, Ni) are allowed to exchange the position with an identity swap move under three different conditions, e.g., in vacuum and in the presence of O2 and H2. It should further be noted that the simulation performed in a clean alloy slab is a MC simulation with a constant number of atoms in the system. All accepted configurations are stored during the simulation and the last 100 configurations are averaged at the end of the simulation to give a good representation of the slab composition at equilibrium. Figure 9 shows the layer-by-layer metal atomic distributions of the alloy slabs at T = 1200 K by GCMC/MD simulations. The Pt concentration of the slab layer is defined in the present work as a ratio of the number of Pt to the total number of metals per layer, NPt /(NPt + NNi). Thus, when the Pt concentration in a layer is the same as the bulk Pt composition, the ratio is calculated to be 0.75. As discussed in Figure 3, the surface energy of Pt3Ni(111) decreases as the Pt concentration on surface increases and thus, Pt is expected to segregate to the atomically clean alloy surface. In the hybrid MC/MD simulation under vacuum, it is found that Ni atoms on the surface are replaced by Pt atoms in subsurface layers and a Pt−rich surface is achieved at thermodynamic equilibrium. From the equilibrium structures, the Pt concentration of the alloy surface is estimated to be 0.97. In the 25 ACS Paragon Plus Environment


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 41

experiment, Pt at.% on the surface in Pt78Ni22(111) has been measured to be about 97 at.% at 760 K using ISS54 and about 99±1 at.% at 1170 K using LEED55. In particular, the LEED measurement has also provided the Pt concentration on the first subsurface (30±5 at.%) and the second subsurface layer (87±10 at.%). Theoretical studies have predicted less amount of segregation in Pt78Ni22(111) – 93 at.% Pt on the surface, 74 at.% Pt on the first subsurface and 81 at.% Pt on the second subsurface at 1120 K56, and 87 at.% Pt on the surface, 72 at.% Pt on the first subsurface and 82 at.% Pt on the second subsurface at 1200 K57. In ReaxFF using the MC/MD simulation, it is observed that the Pt concentration in subsurface layers drops down below the bulk Pt composition – 0.68 on the first subsurface and 0.70 on the second subsurface, showing a little broader Ni−rich region and then weakly fluctuates around the bulk Pt composition in the more underlying layers, indicating the surface and near-surface regions to be similar to sandwich structure.


Page 27 of 41

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


Figure 9 GCMC/MD simulations of slab Pt3Ni under vacuum, O2 and H2 condition at T = 1200 K. (a) initial configuration, (b) segregation configuration at equilibrium and (c) atomic distribution of Pt, Ni and adsorbates (O, H) across the slab. Yellow, cyan, red and white atoms represent Pt, Ni, O and H, respectively. The dashed lines represent the bulk Pt concentration (gray), the per-layer number of Pt in bulk (black) and the per-layer number of Ni in bulk (blue). During GCMC simulations, besides the swap move of Pt and Ni atoms, atomic gas atoms are allowed to insert, delete and translate on the slab. When the oxygen is adsorbed to the alloy slab, the Pt atoms on surface are replaced by the Ni atoms from the subsurface layers, resulting in a Ni−rich surface. This agrees with the experimental observation and DFT studies, demonstrating a Ni–rich surface to be favored in the presence of adsorbed oxygen.50 The Pt concentration on the surface is estimated at 0.44 (0.42 on the top layer and 0.47 on the bottom layer). The oxygen atoms are found to occupy mainly the hollow 27 ACS Paragon Plus Environment


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 41

sites of the Ni−rich region and a Ni oxide island grows successively on the surface. The Pt concentration in the subsurface layers is higher than the bulk Pt concentration and then it fluctuates around the bulk Pt concentration in the underlying layers. In addition to the Ni(111) surface sites, Ni subsurface sites (octahedral and tetrahedral sites) are competitive in adsorbing oxygen than the Pt(111) surface sites with lower oxygen binding energies, e.g., –35.8 (octahedral site) and –27.4 kcal/mol (tetrahedral site) for Ni and –21.8 for Pt(fcc) in ReaxFF. Thus, it is possible that oxygen atoms occupy the subsurface sites with Ni neighbors. It is also interesting to note that the binding of OH at on-top sites of Pt (−45.5 kcal/mol at 0.25 ML in ReaxFF) is much stronger than the oxygen binding to Pt surface and Ni subsurface sites. Thus the formation of OH in the presence of oxygen and hydrogen adsorbates will make the surface structure more complicated, changing the preference of binding sites and in turn, surface composition as well. We also study the surface composition of the alloy under the H2 environment. In the presence of hydrogen, the Pt concentration on the surface layer is estimated to be 0.88, which is lower than the surface Pt concentration under vacuum condition, but higher than the bulk Pt concentration. Since Pt and Ni have similar affinity to hydrogen atom, the presence of hydrogen would not drive the strong segregation of a particular alloy component, inducing surface composition of the components to be placed between two values obtained in vacuum and in the bulk. It is observed that most of the hydrogen atoms sit on the Pt(top) site while a few occupies the bridge and/or the hollow sites as shown in Figure 9b. The surface Pt compositions at different conditions are listed in Table 3.


Page 29 of 41

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


Table 3 Summary of surface segregation in the Pt3Ni alloy slab and truncated cuboctahedron nanoparticle under vacuum, H2 and O2 environments in GCMC/MD simulation. The value in the parentheses indicates the Pt concentration on the surface. Segregation component condition Slab

Cuboctahedron

vacuum

Pt (0.97)

Pt

O2

Ni (0.44)

Ni

H2

Pt (0.88)

Pt

Besides the extended slab surface, we also investigated the surface composition of the alloy nanoparticle under vacuum, O2 and H2 by GCMC/MD at T = 600 K. We analyze the atomic distribution of the alloy components (Pt, Ni) and the effect of adsorbate (O and H) on the structures using the GCMC/MD simulations as shown in Figure 10. Calculating the composition of the nanoparticle surface is not accurate as the slab system since the spherical surface with adsorbates brings an additional complexity to the analysis. Thus we compare the qualitative trend of the surface atomic distributions between the different conditions. The Pt and Ni references in Figure 10 are the number distributions of Pt and Ni, respectively, at an atomic ratio of NPt : NNi = 3 : 1 along the distance from the nanoparticle center. For the clean alloy nanoparticle, Ni distribution shifts to the left slightly, namely towards the center of the nanoparticle. In addition, the number of Ni atoms on the surface is much lower than the number of Pt atoms and in fact, only a few Ni atoms remain on the surface, leading to a Pt skin shell (Figure 10b). It is interesting to note that Pt exchanges with Ni at edge and corner sites. This may indicate that the lower coordination sites are favored for segregation.



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 41

On the other hand, with oxygen adsorbate at the surface, the number of Pt atoms on the surface becomes lower than the Pt reference while the number of Ni atoms is much higher than the Ni reference, indicating an oxygen-induced surface segregation of Ni. The oxygen atoms mainly occupy the hollow sites with Ni neighbors, similar to the slab system. It is found that the Ni oxide is formed on the (111) face rather than at the edges or corners. In the presence of hydrogen on the surface, the surface composition shifts slightly to the Ptrich surface. However, the surface segregation is not as apparent as shown in the presence of oxygen adsorbate or under vacuum. The surface segregations of nanoparticle at different conditions are summarized in Table 3.

Figure 10 GCMC/MD simulations of truncated cuboctahedron Pt3Ni under vacuum, O2 and H2 condition at T = 600 K. (a) initial configuration, (b) segregation configuration at equilibrium and (c) atomic distribution of Pt, Ni and adsorbates (O, H) from the center of the nanoparticle to the surface. Yellow, cyan, red and white atoms represent Pt, Ni, O and H, respectively. 30 ACS Paragon Plus Environment

Page 31 of 41

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


We also investigated the surface segregation in Pt2Ni alloy slab and nanoparticle. Since the segregation trend shows similar results to the Pt3Ni systems, we do not present them in this work. As noted earlier in this work, Pt segregation is not present on a clean PtxNi1-x (111) surface at x < 0.67. We also performed a GCMC/MD simulation in a 1:1 PtNi truncated cuboctahedral nanoparticle at 600 K under different gas environments. It is observed that a weak segregation of Ni occurs in vacuum while a weak segregation of Pt is observed at high surface coverage of hydrogen. A strong Ni segregation occurs in O2 environment. The surface morphology and the atomic distributions of the alloy nanoparticle are presented in Supporting Information (Figure S1). In summary, the segregation phenomena observed from the slab and the nanoparticle in a given alloy composition range (x > 0.6) are in agreement with the segregation trend reported in earlier studies.46, 51 Decomposition of the PtxOy Nanoparticle. As mentioned earlier, a reactive force field development is modular in nature − implying a Pt/O force field is combined with a Pt/C and a Pt/H to form a Pt/C/H/O which in turn, is combined with a similarly developed Ni/C/H/O force field and a Pt/Ni alloy force field to arrive at a Pt/Ni/C/H/O reactive force field. Though the alloy force field development in this work focused on enriching the training set with data to account for Pt/Ni interaction and alloy interactions with C, H and O, it is worthwhile to evaluate the performance of the alloy force field for a subsystem to understand the fitting. In addition, as mentioned earlier, the training set for a Pt/O force field contains data corresponding to bulk systems, making it worthwhile to explore its limits and constantly identify domains to improve. Traditionally, for the very same reason, while employing reactive force fields, it is customary to validate and/or retrain the force 31 ACS Paragon Plus Environment


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 41

field to describe the salient chemistry in the application of interest. It has been reported that when a platinum wire is heated above the temperature at which Pt oxide cannot exist as a solid, it loses weight rapidly.58-60 Herein, to understand the Pt oxide thermal decomposition, we created a 1:1 PtO cluster with a diameter of around 12 Å by performing the GCMC/MD simulation at T = 600 K and PO2 = 10-6 atm. As the temperature of the PtO nanoparticle is raised above 825 K, it is observed that PtO begins thermal decomposition and breaks up into simpler Pt oxide clusters (Pt6O8). To examine the thermodynamic stability of the Pt oxide clusters, we carry out a series of GCMC /MD simulations of various sizes of Pt-clusters (Pt1 to Pt7) under O2 environment at 300 K and PO2 = 10–40 – 105 atm. The clusters formed from the GCMC simulations are shown in Figure 11. The formation energies of these Pt oxide clusters are compared to the DFT structures. As shown in Figure 11, PtxO2x clusters are more stable than PtxOx clusters. Also, the formation of small size Pt oxide clusters (x < 5 for PtxOx and x < 4 for PtxO2x) is not thermodynamically favorable. From the comparison of the formation energy between the clusters, a highly symmetric Pt6O8 cluster is found to be most stable in ReaxFF, which supports the observation in the PtO cluster thermal decomposition at high temperature. In DFT, as the cluster grows, the chain structure which is shown at small clusters becomes flexible enough to form a packed cluster shape. The structural and energetic discrepancy in small size clusters between the ReaxFF and DFT data can be explained by the fact that the Pt/O force field has been trained strongly against the training data set for the bulk oxides (PtO, α−PtO2, β−PtO2 and Pt3O4) with hardly any energetics and geometry of small clusters. Thus, it is expected that ReaxFF may provide a reasonable description of the thermodynamic stability for the relatively large and stable Pt–oxide clusters. We believe 32 ACS Paragon Plus Environment

Page 33 of 41

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


that improving the force field using this observation would lead to a further progress in the study of the growth and the decomposition process of Pt oxide nanoclusters.

Figure 11 Formation energy of PtxOx and PtxO2x clusters (x = 1–7) with respect to fcc Pt and O2 gas. *Cyan and pink dashed lines indicate the formation energy of bulk PtO and α−PtO2, respectively (reference 16). The data in black color represents the Pt6O8 cluster. Yellow and red atoms in the structures represent Pt and O, respectively.

CONCLUSIONS We developed the ReaxFF force field for Pt/Ni/C/H/O interactions for the 33 ACS Paragon Plus Environment


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 41

heterogeneous catalytic application of the Pt–Ni alloy. Since the interplay between surface composition and adsorbates is critical in determining the catalytic activity and selectivity of the alloys, the force field parameterization are performed using a series of DFT data sets for the adsorption strength of various atomic and molecular adsorbates (O, H, C, CH, CH2, CH3, CO, OH and H2O) on (111) surfaces of the Pt8Ni4, Pt9Ni3, and Pt10Ni2 alloys, and the surface energy of these three alloys at different surface compositions. Pt surface segregation which is known to be observed on (111) and (100) surfaces, is successfully demonstrated in PtxNi1-x alloy phases at x ≥ 0.67 for (111) surface and x ≥ 0.62 for (100) surface. However, it is worth mentioning that the Pt/Ni alloy force field in this study is not successful in reproducing the Pt segregation of PtxNi1-x alloys at x < 0.67 on (111) surface and at x < 0.62 on (100) surface. The orders of binding strength between the adsorbates at the surface coverage of 0.25 ML in three alloys are as follows: H2 < H2O < CH3 ≈ O2 ≈ CO < OH < CH2 < C ≈ CH

(Pt8Ni4)

H2 < H2O < CO ≈ O2 < CH3 < OH < CH2 < CH < C

(Pt9Ni3)

H2 < H2O < O2 < CO < CH3 < OH < CH2 < C ≈ CH

(Pt10Ni2)

From the study of the preferential alloy component of the adsorbates, it is expected that the adsorption of H and CH3 on the alloy surface induces the Pt−segregation, or vice versa. On the other hand, oxidation intermediates and products such as C, O, OH, H2O, CO, CH and CH2 increase the Ni concentration on the Pt−Ni alloy surface, resulting in the Ni−segregation. Using the ReaxFF force field, we performed the GCMC/MD simulations in Pt3Ni slab and truncated cuboctahedron nanoparticle to demonstrate the segregation trend in vacuum and under different gas environments (O2 and H2). Adsorbates can selectively enhance the 34 ACS Paragon Plus Environment

Page 35 of 41

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


particular segregation of one component of the bimetallic Pt–Ni alloy when the preferential attractions exist between the adsorbate and the alloy component. From the GCMC structures at equilibrium, Pt segregation is observed under vacuum and in the presence of hydrogen adsorbate in both slab and nanoparticle surfaces. On the other hand, oxygen adsorption on the alloy surface induces Ni segregation. The segregation phenomena are in good agreement with the findings reported in the earlier works, verifying the applicability of the alloy force field and the feasibility of the GCMC/MD method in modeling and studying the Pt–Ni catalysts. As a part of the alloy force field validation, we also investigated the formation of small PtO and PtO2 clusters over a range of cluster size (from Pt1 to Pt7) using GCMC/MD simulation. While the structural and energetic discrepancy exists between the ReaxFF and DFT in the thermodynamically unfavorable small size clusters, the force field can still describe reasonably geometrical structures and energetics for the relatively large Pt–oxide clusters. As a result, we suggest that the Pt–Ni alloy force field targeted for heterogeneous

catalysis

application

can

be

used

for

the

catalytic

oxidation,

dehydrogenation and coupling reactions.

ASSOCIATED CONTENT Supporting Information GCMC/MD simulation results for a PtNi cuboctahedral nanoparticle give in Figure S1. ReaxFF Reactive Force Field Parameters for Pt/Ni/C/H/O. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION 35 ACS Paragon Plus Environment


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 41

Corresponding Author *E-mail: [email protected] Notes The authors declare no competing financial interest.

ACKNOWLEDGEMENTS This work was funded by a grant from the Corporate Strategic Research, ExxonMobil Research and Engineering, Clinton, NJ. We wish to thank the PSU Institute for CyberScience Acvanced Cyberinfrastructure group for providing the computational resources.

REFERENCES 1. Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Markovic, N. M.; Ross, P. N. Oxygen reduction on high surface area Pt-based alloy catalysts in comparison to well defined smooth bulk alloy electrodes Electrochim. Acta 2002, 47, 3787-3798. 2. Belgued, M.; Pareja, P.; Amariglio, A.; Amariglio, H. Conversion of Methane into Higher Hydrocarbons on Platinum Nature 1991, 352, 789-790. 3. Velichkina, L. M.; Pestryakov, A. N.; Vosmerikov, A. V.; Tuzovskaya, I. V.; Bogdanchikova, N. E.; Avalos, M.; Farias, M.; Tiznado, H. Catalytic activity in the hydrocarbon conversion of systems containing platinum, nickel, iron, and zinc nanoparticles (communication 2) Petrol Chem+ 2008, 48, 355-359. 4. Satterfield, C. N. In Heterogeneour Catalysis in Industrial Practice, McGraw-Hill: New York, 1991. 5. Arenz, M.; Mayrhofer, K. J. J.; Stamenkovic, V.; Blizanac, B. B.; Tomoyuki, T.; Ross, P. N.; Markovic, N. M. The effect of the particle size on the kinetics of CO electrooxidation on high surface area Pt catalysts J. Am. Chem. Soc. 2005, 127, 6819-6829. 6. Captain, D. K.; Mihut, C.; Dumesic, J. A.; Amiridis, M. D. On the mechanism of the NO reduction by propylene over supported Pt catalysts Catal. Lett. 2002, 83, 109-114. 7. Yu, W.; Porosoff, M. D.; Chen, J. G. Review of Pt-Based Bimetallic Catalysis: From Model Surfaces to Supported Catalysts Chem. Rev. 2012, 112, 5780-5817. 8. Siani, A.; Alexeev, O. S.; Lafaye, G.; Amiridis, M. D. The effect of Fe on SiO2-supported Pt catalysts: Structure, chemisorptive, and catalytic properties J. Catal. 2009, 266, 26-38. 9. Mu, R. T.; Guo, X. G.; Fu, Q.; Bao, X. H. Oscillation of Surface Structure and Reactivity of 36 ACS Paragon Plus Environment

Page 37 of 41

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


PtNi Bimetallic Catalysts with Redox Treatments at Variable Temperatures J. Phys. Chem. C 2011, 115, 20590-20595. 10. Chen, S.; Ferreira, P. J.; Sheng, W. C.; Yabuuchi, N.; Allard, L. F.; Shao-Horn, Y. Enhanced activity for oxygen reduction reaction on "Pt(3)CO" nanoparticles: Direct evidence of percolated and sandwich-segregation structures J. Am. Chem. Soc. 2008, 130, 13818-13819. 11. Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Improved oxygen reduction activity on Pt3Ni(111) via increased surface site availability Science 2007, 315, 493-497. 12. Garcia-Dieguez, M.; Pieta, I. S.; Herrera, M. C.; Larrubia, M. A.; Alemany, L. J. Nanostructured Pt- and Ni-based catalysts for CO2-reforming of methane J. Catal. 2010, 270, 136-145. 13. Su, H. Y.; Bao, X. H.; Li, W. X. Modulating the reactivity of Ni-containing Pt(111)-skin catalysts by density functional theory calculations J. Chem. Phys. 2008, 128. 14. Ahmadi, M.; Behafarid, F.; Cui, C. H.; Strasser, P.; Cuenya, B. R. Long-Range Segregation Phenomena in Shape-Selected Bimetallic Nanoparticles: Chemical State Effects Acs Nano 2013, 7, 9195-9204. 15. Gai, L. L.; Shin, Y. K.; Raju, M.; van Duin, A. C. T.; Raman, S. Atomistic Adsorption of Oxygen and Hydrogen on Platinum Catalysts by Hybrid Grand Canonical Monte Carlo/Reactive Molecular Dynamics J. Phys. Chem. C 2016, 120, 9780-9793. 16. Fantauzzi, D.; Bandlow, J.; Sabo, L.; Mueller, J. E.; van Duin, A. C. T.; Jacob, T. Development of a ReaxFF potential for Pt-O systems describing the energetics and dynamics of Pt-oxide formation Phys. Chem. Chem. Phys. 2014, 16, 23118-23133. 17. Mueller, J. E.; van Duin, A. C. T.; Goddard, W. A. Development and Validation of ReaxFF Reactive Force Field for Hydrocarbon Chemistry Catalyzed by Nickel J. Phys. Chem. C 2010, 114, 4939-4949. 18. Zou, C. Y.; Shin, Y. K.; van Duin, A. C. T.; Fang, H. Z.; Liu, Z. K. Molecular dynamics simulations of the effects of vacancies on nickel self-diffusion, oxygen diffusion and oxidation initiation in nickel, using the ReaxFF reactive force field Acta Mater. 2015, 83, 102-112. 19. Papoian, G.; Norskov, J. K.; Hoffmann, R. A comparative theoretical study of the hydrogen, methyl, and ethyl chemisorption on the Pt(111) surface J. Am. Chem. Soc. 2000, 122, 41294144. 20. Weigand, P.; Novacek, P.; Vanhusen, G.; Neidhart, T.; Mezey, L. Z.; Hofer, W.; Varga, P. Surface-Composition of Ptxni1-X Single-Crystal Alloys Nucl. Instrum. Meth. B 1992, 64, 9397. 21. Peng, G. W.; Sibener, S. J.; Schatz, G. C.; Ceyer, S. T.; Mavrikakis, M. CO2 Hydrogenation to Formic Acid on Ni(111) J. Phys. Chem. C 2012, 116, 3001-3006. 22. Sanchez, J. M.; Ducastelle, F.; Gratias, D. Generalized Cluster Description of Multicomponent Systems Physica A 1984, 128, 334-350. 23. Christoffersen, E.; Stoltze, P.; Norskov, J. K. Monte Carlo simulations of adsorption37 ACS Paragon Plus Environment


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

induced segregation Surf. Sci. 2002, 505, 200-214. 24. Han, B. C.; Van der Ven, A.; Ceder, G.; Hwang, B. J. Surface segregation and ordering of alloy surfaces in the presence of adsorbates Physical Review B 2005, 72, 205409. 25. Muller, S. Bulk and surface ordering phenomena in binary metal alloys J Phys-Condens Mat 2003, 15, R1429-R1500. 26. Chen, W.; Dalach, P.; Schneider, W. F.; Wolverton, C. Interplay between Subsurface Ordering, Surface Segregation, and Adsorption on Pt-Ti(111) Near-Surface Alloys Langmuir 2012, 28, 4683-4693. 27. Stephens, J. A.; Ham, H. C.; Hwang, G. S. Atomic Arrangements of AuPt/Pt(111) and AuPd/Pd(111) Surface Alloys: A Combined Density Functional Theory and Monte Carlo Study J. Phys. Chem. C 2010, 114, 21516-21523. 28. Chenoweth, K.; van Duin, A. C. T.; Persson, P.; Cheng, M. J.; Oxgaard, J.; Goddard, W. A. Development and application of a ReaxFF reactive force field for oxidative dehydrogenation on vanadium oxide catalysts J. Phys. Chem. C 2008, 112, 14645-14654. 29. Chenoweth, K.; van Duin, A. C. T.; Goddard, W. A. ReaxFF reactive force field for molecular dynamics simulations of hydrocarbon oxidation J. Phys. Chem. A 2008, 112, 1040-1053. 30. Shin, Y. K.; Shan, T. R.; Liang, T.; Noordhoek, M. J.; Sinnott, S. B.; van Duin, A. C. T.; Phillpot, S. R. Variable charge many-body interatomic potentials MRS Bull. 2012, 37, 504-512. 31. Senftle, T. P.; Hong, S.; Islam, M. M.; Kylasa, S. B.; Zheng, Y.; Shin, Y. K.; Junkermeier, C.; Engel-Herbert, R.; Janik, M. J.; Aktulga, H. M.; Verstraelen, T.; Grama, A.; van Duin, A. C. T. The ReaxFF reactive force-field: development, applications and future directions npj Comput Mater 2016, 2, 15011. 32. Kresse, G.; Furthmuller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set Phys. Rev. B: Condens. Matter 1996, 54, 1116911186. 33. Blochl, P. E. Projector Augmented-Wave Method Phys. Rev. B: Condens. Matter 1994, 50, 17953-17979. 34. Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple Phys. Rev. Lett. 1996, 77, 3865-3868. 35. Cao, L.; Mueller, T. Rational Design of Pt3Ni Surface Structures for the Oxygen Reduction Reaction J. Phys. Chem. C 2015, 119, 17735-17747. 36. Ludwig, J.; Vlachos, D. G.; van Duin, A. C. T.; Goddard, W. A. Dynamics of the dissociation of hydrogen on stepped platinum surfaces using the ReaxFF reactive force field J. Phys. Chem. B 2006, 110, 4274-4282. 37. Hofer, W. Surface Segregation of Ptni Alloys - Comparing Theoretical and Experimental Results Fresenius J. Anal. Chem. 1993, 346, 246-251. 38. Weigand, P.; Novacek, P.; Vanhusen, G.; Neidhart, T.; Varga, P. Surface-Analysis of Ptxni1-X Single-Crystals Surf. Sci. 1992, 269, 1129-1134. 38 ACS Paragon Plus Environment

Page 38 of 41

Page 39 of 41

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


39. Podkolzin, S. G.; Shen, J. Y.; de Pablo, J. J.; Dumesic, J. A. Equilibrated adsorption of CO on silica-supported Pt catalysts J. Phys. Chem. B 2000, 104, 4169-4180. 40. Michaelides, A.; Hu, P. A density functional theory study of hydroxyl and the intermediate in the water formation reaction on Pt J. Chem. Phys. 2001, 114, 513-519. 41. Ma, Y. G.; Balbuena, P. B. Pt surface segregation in bimetallic Pt3M alloys: A density functional theory study Surf. Sci. 2008, 602, 107-113. 42. Schmidt, D. J.; Chen, W.; Wolverton, C.; Schneider, W. F. Performance of Cluster Expansions of Coverage-Dependent Adsorption of Atomic Oxygen on Pt(111) J Chem Theory Comput 2012, 8, 264-273. 43. Michaelides, A.; Hu, P. A density functional theory study of the reaction of C+O, C+N, and C+H on close packed metal surfaces J. Chem. Phys. 2001, 114, 5792-5795. 44. Hamada, I.; Morikawa, Y. Density-Functional Analysis of Hydrogen on Pt(111): Electric Field, Solvent, and Coverage Effects (vol 112C, pg 10897, 2008) J. Phys. Chem. C 2008, 112, 17486-17486. 45. Berkeley, Argonne team report promising Pt/Ni nanocatalysts Fuel Cells Bulletin 2014, 2014, 11. 46. Li, L.; Zhou, L.; Ould-Chikh, S.; Anjum, D. H.; Kanoun, M. B.; Scaranto, J.; Hedhili, M. N.; Khalid, S.; Laveille, P. V.; D'Souza, L.; Clo, A.; Basset, J.-M. Controlled Surface Segregation Leads to Efficient Coke-Resistant Nickel/Platinum Bimetallic Catalysts for the Dry Reforming of Methane ChemCatChem 2015, 7, 819-829. 47. Bao, H.; Li, J.; Jiang, L.; Shang, M.; Zhang, S.; Jiang, Z.; Wei, X.; Huang, Y.; Sun, G.; Wang, J.-Q. Structure of PtnNi Nanoparticles Electrocatalysts Investigated by X-ray Absorption Spectroscopy J. Phys. Chem. C 2013, 117, 20584-20591. 48. Deckers, S.; Habraken, F. H. P. M.; van der Weg, W. F.; Denier van der Gon, A. W.; Pluis, B.; van der Veen, J. F.; Baudoing, R. Segregation at the Pt0.5Ni0.5(111) surface studied by medium-energy ion scattering Phys. Rev. B: Condens. Matter 1990, 42, 3253-3259. 49. Menning, C. A.; Chen, J. G. Thermodynamics and kinetics of oxygen-induced segregation of 3d metals in Pt–3d–Pt(111) and Pt–3d–Pt(100) bimetallic structures J. Chem. Phys. 2008, 128, 164703. 50. Menning, C. A.; Hwu, H. H.; Chen, J. G. Experimental and Theoretical Investigation of the Stability of Pt−3d−Pt(111) Bimetallic Surfaces under Oxygen Environment J. Phys. Chem. B 2006, 110, 15471-15477. 51. Wang, G.; Van Hove, M. A.; Ross, P. N.; Baskes, M. I. Monte Carlo simulations of segregation in Pt-Ni catalyst nanoparticles J. Chem. Phys. 2005, 122, 024706. 52. Stadler, H.; Hofer, W.; Schmid, M.; Varga, P. Embedded-atom method calculations applied to surface segregation of Pt-Ni single crystals Surf. Sci. 1993, 287–288, Part 1, 366-370. 53. In Chase, M. W. and National Institute of Standards and Technology, NIST-JANAF Thermochemical Tables (American Chemical Society, Washington, D.C., 1998). 54. Detemmerman, L.; Creemers, C.; Vanhove, H.; Neyens, A.; Bertolini, J. C.; Massardier, J. 39 ACS Paragon Plus Environment


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

Experimental-Determination of Equilibrium Surface Segregation in Pt-Ni Single-Crystal Alloys Surf. Sci. 1986, 178, 888-896. 55. Gauthier, Y.; Joly, Y.; Baudoing, R.; Rundgren, J. Surface-Sandwich Segregation on Nondilute Bimetallic Alloys - Pt50ni50 and Pt78ni22 Probed by Low-Energy ElectronDiffraction Physical Review B 1985, 31, 6216-6218. 56. Treglia, G.; Legrand, B. Surface-Sandwich Segregation in Pt-Ni and Ag-Ni Alloys - 2 Different Physical Origins for the Same Phenomenon Physical Review B 1987, 35, 43384344. 57. Lundberg, M. Surface Segregation and Relaxation Calculated by the Embedded-Atom Method - Application to Face-Related Segregation on Platinum-Nickel Alloys Physical Review B 1987, 36, 4692-4699. 58. Krier, C. A.; Jaffee, R. I. Oxidation of the Platinum-Group Metals J Less-Common Met 1963, 5, 411-431. 59. Chaston, J. C. Reaction of Oxygen with the Platinum Metals Platinum Met. Rev. 1964, 8, 50-54. 60. Jehn, H. Platinum Losses during High-Temperature Oxidation J Less-Common Met 1981, 78, P33-P41.


Page 40 of 41

Page 41 of 41

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


TABLE OF CONTENTS (TOC) IMAGE


Development of a ReaxFF Reactive Force Field for ... - ACS Publications

Recommend Documents