A Density Functional Investigation - ACS Publications - American

Subscriber access provided by RMIT University Library

Article 16

Properties of HOPG Supported TaSi Clusters: A Density Functional Investigation Arpita Sen, and Prasenjit Sen J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.7b08858 • Publication Date (Web): 04 Dec 2017 Downloaded from http://pubs.acs.org on December 9, 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 32 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

Properties of HOPG Supported TaSi16 Clusters: A Density Functional Investigation Arpita Sen and Prasenjit Sen∗ Harish-Chandra Research Institute, HBNI, Chhatnag Road, Jhunsi, Allahabad, 211019, India. E-mail: [email protected] Phone: +91 532 2274307

Abstract Adsorption of the TaSi16 cluster on highly oriented pyrolytic graphite (HOPG) is studied using density functional methods. These calculations resolve some of the issues raised by the recent experiments of Shubita et al., 1 and provide additional insights into the system. For the first time, the ground state structure of TaSi16 is obtained through a global search employing evolutionary algorithm. A symmetric structure of the Si cage, together with a small charge transfer from Ta to the Si atoms explains the observed core-level photoemission spectra. The ground state isomer is deposited on HOPG at different points, and in different random orientations. The cluster is found to be physisorbed on HOPG, and retains the Ta-encapsulating cage structure, with little charge transfer between the cluster and the substrate. At finite temperatures the cluster does not diffuse over pico seconds time scales. The time-scale for diffusion is estimated to be a few tenths of a milli-second from the calculated barriers. No coalescence between the cluster and the substrate is found even at 700 K, ruling this ∗

To whom correspondence should be addressed

1

ACS Paragon Plus Environment

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

out as a possible cause for the observed core-level shift at elevated temperatures. Both at 400 K and 700 K, the cluster oscillates about its lowest energy structure after deposition, and is not found to transform to other isomers within the time-scale of the simulations. However, these oscillations cause large variations in the energies of the electronic levels which may explain the absence of definite peaks in the valence-level photoemission spectra.

Introduction Transition metal (TM) encapsulating silicon cage clusters have been studied extensively over the last couple of decades. 2–8 The initial motivation was to understand metal silicon interfaces. 2,3 After the discovery of C fullerenes, questions were asked if cages of Si would also be stable, and could be used as building blocks to synthesize novel materials. It was found that hollow Si cages are not stable because unlike C, sp2 hybridization in Si is unfavorable. Interestingly, encapsulation of TM atoms rendered Si cages stable. 4 In a pioneering experiment, Hiura et al. 5 showed that for a given TM atom, TMSin cluster of a particular size is more stable compared to its neighbors. For example, WSi12 was claimed to be the most stable cluster in the WSin series. The discussion then shifted towards discovering the most stable candidates among all TMSin clusters, and understanding the origin of their stability. Hiura et al. 5 suggested that the so-called 18-electron rule of the organometallic chemistry determines relative stability of these clusters. Assuming that each Si atom donates one electron to the central TM atom, clusters with 18 valence electrons on the TM atom were argued to be more stable. The WSi12 cluster satisfies this rule, and its presumed stability was cited as validity of this rule. Khanna and co-workers claimed CrSi12 9 and FeSi10 10 to be the most stable clusters in the respective TMSin series over limited ranges of size. Notably, both these clusters nominally satisfy the 18-electron rule. + In mass abundance experiments of Koyasu et al. ScSi− 16 , TiSi16 and VSi16 were found to be

most abundant among the anion, neutral and cation clusters, respectively. 11 This opened a 2


Page 2 of 32

Page 3 of 32 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


new line of argument which claimed that TMSin clusters follow electron counting rules much like the shell models in simple metal clusters. 11–13,13 Thus clusters with 18 or 20 electrons on the TM atom should be stable. In fact, Reveles and Khanna claimed that a free-electron gas is formed inside the Si cage in these clusters. 14 However, all clusters that nominally satisfy 18- or 20-electron count were not found to have enhanced stability. Sen and Mitas were perhaps the first to point out that simple electron counting rules are not generally valid. 7 Guo et al.’s work supported these claims. 15 For example, Sen and Mitas found VSi12 rather than CrSi12 to be the most stable in the TMSi12 series. Initially, Khanna and co-workers argued that Sen et al. had not found CrSi12 (satisfying 18-electron rule) to be the most stable cluster because they failed to account for the so-called WignerWitmer spin conservation rule. 9 Once this rule was taken into account, CrSi12 turned out to be most stable. But more recent calculations from Khanna’s group 16–18 have established that TMSi14 are the most stable clusters for TM=Cr, Fe and W. This validates Sen et al.’s claim that nominal electron counting rules are not the only criteria determining stability of these clusters. Though the origin of enhanced stability of certain TMSin clusters remains a matter of debate, what is definitely established is that there are some clusters with enhanced stability. For example, experiments from Nakajima’s group established that TaSi+ 16 is more stable than neighboring sizes, and this was rationalized by electron counting rule (20-electron filled shell). 13 As is clear, numerous studies, both theoretical and experimental, have addressed fundamental properties of TMSin clusters in the gas phase. There were no studies trying to use them in applications, or studying their properties after deposition on a substrate, often a pre-condition for application. The first such attempt has come from Nakajima’s group. In a recent experiment, they have deposited stable TaSi+ 16 clusters on a highly oriented pyrolytic graphite (HOPG) substrate. 1 The main motivation in this set of experiments was to explore formation of assemblies of

3



TaSi16 on the HOPG substrate. Structure and electronic properties of the deposited clusters were studied using scanning tunneling microscopy (STM), X-ray photoemission spectroscopy (XPS) and ultra-violet photoemission spectroscopy (UPS) probes. While STM gives information about atomic structure, XPS and UPS probe energies of the core and valence electron levels respectively. STM and XPS measurements were performed both on the as-deposited samples, and after heating them at 400 K and 720 K. In order to check the so-called super atomic character of TaSi16 , they also studied reaction of the deposited clusters with oxygen. After deposition, the clusters retained their metal-enclosing cage structures and formed small islands. Height of the islands was found to be about 1 nm in STM experiments, indicating a one-layer thickness. In the as-deposited clusters, the Si cage was inferred to be highly symmetric, based on the XPS results. Individual clusters were also discernible at 400 K. There were some shifts in the XPS peaks after heating at 720 K which was presumed to be due to partial coalescence of the clusters or coalescence between the clusters and the HOPG substrate. UPS spectra did not show any definite peaks in the valence electron levels. This was supposed to be due to rapid transition of the cluster between close-lying isomers with very different electronic structures. The clusters showed considerable resistance to oxidation establishing their chemical stability. When they did react, the Si cage reacted more readily than the encapsulated Ta atoms, indicating that the cage structures were still maintained. While these experiments provided a lot of information on the behavior of TaSi16 clusters deposited on HOPG, it is important to recall that none of these provides microscopic insights. In order to have a microscopic picture of the system one needs atomistic calculations. Therefore, in order to better understand this important cluster-substrate system, and to further qualify the reported experimental results, we have performed first-principles calculations on free and HOPG-deposited TaSi16 clusters using density functional theory (DFT). One difference with the experiments however is that we study deposition of neutral clusters while in experiments cation clusters were deposited. We avoid study of cation clus-

4


Page 4 of 32

Page 5 of 32 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


ters due to technical reasons. Firstly, we cannot strictly ensure that the electron is removed from the cluster. Secondly, plane wave DFT codes cannot deal with charged systems. A compensating uniform background charge is introduced. Energy convergence with respect to cell size is slow in such cases. In any case, HOPG being a good electrical conductor, the charge on the cluster before deposition is perhaps not important, and hence studying an overall neutral system is appropriate. We confirm many of the conclusions derived from the experiments, but there are some differences too. Therefore these calculations complement the experimental studies.

Method All calculations were performed within the framework of plane wave DFT as implemented in the VASP code. 19–23 An energy cutoff of 600 eV was used for the plane wave basis set. Interaction between the valence electrons and the ion cores was represented by the projector augmented wave (PAW) potentials. 24 Brillouin zone integrations were performed with the Γ point only both for isolated clusters and HOPG supported clusters. The HOPG surface was represented by a repeated slab geometry. The graphite slab in the simulation cell contained three carbon layers with a vacuum layer of 25 ˚ A separating two successive slabs. Each layer was in the shape of a rhombus with sides of 14.88 ˚ A. Such a large surface supercell ensures that the distance between the cluster in the simulation cell and its periodic images is at least ∼ 12 ˚ A so that there are no interactions between them. Carbon atoms in the two top layers were fully relaxed while the bottom layer of atoms was kept fixed to simulate a bulk-like termination. Additional carbon layers, or relaxation of all three layers did not change the adsorption energies of the deposited clusters. Before the TaSi16 cluster was deposited on HOPG, a global search was performed to locate its minimum energy structure using an evolutionary algorithm. The USPEX code 25,26 was used for this. The initial population was generated without any constraint on point group

5



symmetry, i.e., the atomic co-ordinates of the cluster were generated in a completely random manner. Population in each subsequent generation was produced through the application of variational operations of heredity, mutation and soft-mutation. 26 A population size of 34 was maintained in each generation. The fitness criterion for choosing the best structures was the energy of the clusters. Calculation was continued up to a maximum of 25 generations subject to the condition that if the energy of the best structure did not change over 17 successive generations, it would be stopped. Energies were calculated using the plane wave PAW method within the DFT using the VASP code. The clusters were placed inside a box whose size was generated automatically by USPEX. Brillouin zone integrations were performed using the Γ-point only. For each structure, energy calculations in VASP were performed at five increasing levels of accuracy in terms of both the force convergence criterion, and energy cut-off. For the first two levels, the energy cut-off was determined by the ENMAX of the POTCAR files (245 eV). For the next three levels it was 320 eV. For the first level the force convergence criterion was 0.2 eV/˚ A, for the next two levels it was 0.01 eV/˚ A, and in the final level it was 0.001 eV/˚ A. The global search was performed using the spin un-polarized PBE-GGA exchange-correlation functional. At the end of the global search we collected the best structure from each generation. These structures were further optimized using the vdW-DF2 functional in a spin-polarized calculation. vdW-DF2 functional was used because this functional is used for study of TaSi16 on graphite. Use of vdW-DF2 is necessitated by the fact that graphite is not described well by the local density (LDA) or semi-local (GGA) approximations in DFT. 27,28 We discuss this further in the following section. For the re-optimization calculation the energy cut-off was set at 600 eV and the force convergence criterion was 0.001 eV/˚ A. We re-optimized each of the structures selected from USPEX in the doublet, quartet, and sextet spin states. Sometimes, different initial structures converged to the same final structure after relaxation. For every structure the doublet state was found to have the lowest energy. For the study of reaction with oxygen molecules, we used both the PBE and vdW-DF2

6


Page 6 of 32

Page 7 of 32 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


functionals. We got qualitatively same results in the two methods. All other parameters were the same as stated above. For the finite temperature MD calculations, the temperature was fixed by a Nos´e-Hoover thermostat. 29–31 Two carbon layers were used in order to reduce computational costs. Verlet algorithm was used to integrate the equations of motion of the ions. Other parameters were the same as in zero temperature calculations. As mentioned earlier, MD calculations were performed at 400 K and 700 K as the deposited clusters were heated to these two temperatures. At 400 K, a time step of 2 fs was used, and at 700 K the time step was set at 1.5 fs due to faster motion of the ions.

Results and Discussion Bulk graphite In this subsection, we briefly discuss our results for bulk graphite to emphasize the point that inclusion of dispersion interactions is absolutely important for its proper description. As LDA and GGA functionals do not incorporate dispersion interactions, we used methods that incorporate these. In particular, DFT-D2, vdW-DF and vdW-DF2 were used. In the DFT-D2 a pair-wise semi-empirical attractive term is added to the DFT total energy to include medium and long distance dispersion forces. The total energy in this method is written as

Etot = EDFT + Edisp . EDFT is the DFT total energy for a given exchange-correlation functional, and Edisp is the semi-empirical pair-wise dispersion correction,

Edisp = −s6

N X C ij

6 f (Rij ). 6 damp Rij i

A Density Functional Investigation - ACS Publications - American

Recommend Documents