The contact of graphene with Ni(111) surface: description by modern dispersive forces approaches

Abstract

Here we present a Density Functional Theory (DFT) study on the suitability of modern corrections for the inclusion of dispersion related terms (DFT-D) in treating the interaction of graphene and metal surfaces, exemplified by the graphene/Ni(111) system. The Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional is used as basis, on top of which we tested the family of Grimme corrections (D2 and D3, including Becke-Jonson damping and Andersson approach) as well as different flavors of the approach by Tkatchenko and Scheffler (TS). Two experimentally observed chemisorbed states, top-fcc and bridge-top conformations, were examined, as well as one physisorbed situation, the hcp-fcc state. Geometric, energetic, and electronic properties were compared to sets of experimental data for our model system of graphene/Ni(111), but also for available data of bulk Ni, graphite, and free-standing graphene. Results show that two of the most recent approximations, the fully ab initio TS-MBD, and the semi-empirical Grimme D3 correction are best suited to describe graphene↔metal contacts, yet, comparing to earlier studies, the Rev-vdW-DF2 functional is also a good option, whereas optB86-vdW and optB88b-vdW functionals are fairly close to experimental values to be harmless used. The present results highlight how different approaches for the approximate treatment of dispersive forces yield different results, and so finetuning and testing of the envisioned approach for every specific system is advisable. The present survey clears the path for future accurate and affordable theoretical studies of nanotechnologic devices based on graphene-metal contacts.

Adding Pieces to the CO/Pt(111) Puzzle: The Role of Dispersion

“…This is true for D2, D3, and MBD corrections, yet it does not hold for D3BJ, where its application worsens the description for most cases, in line with a slightly worse description of graphene/Ni(111) interactions as earlier reported 35. Taking this into account, and focusing the attention to the PBE based cases, on observes how addition of D2 improves the description by 13 meV in average, whereas successive treatment by D3 and MBD further stepwise improves it by 20 and 8 meV, respectively.…”

“…Furthermore, the non-separable gradient approximation (NGA) N12 xc functional 43 and the TPSS 44 and M06-L 45 meta-GGA functionals have also been considered, again combined with different of the available methods aimed at introducing dispersion. [34][35] Among all these methods we chose Grimme D2 46 and D3 47 empirical corrections as well as the Becke-Jonson (BJ) damping on D3. 48 Indeed, PBE and RPBE in conjunction with D3 or D3BJ correction is a focus to our attention as has been successfully used computational levels for the treatment of complex molecular adlayers on inorganic surfaces.…”

“…53,54 These choices are supported from recent work showing that they provide the best treatment in systems where dispersive forces do play a key role such as graphene on Ni(111), H 2 storage on oxide based clusters, and many others. 34,35,55 Combining the different xc functionals and the above commented dispersion methods results in a total of 29 methods used to explore CO on Pt(111).…”

“…This is the case of graphene on Ni(111) highlighting both some significant discrepancies between results obtained from different dispersion corrected functionals and the importance of a correct treatment of this contribution when dealing with small energy differences right in the order of the hundredths of an eV. 34,35 From the previous discussion it is clear that even if the CO/Pt(111) puzzle was claimed to be solved, this was well before the blooming of methods aimed at introducing dispersion terms in the xc functionals. The results of Lazić et al 30 using the DFT-vdW functional clearly show the importance of dispersion in discriminating the preference for the two types of sites.…”

Step edge-mediated assembly of periodic arrays of long graphene nanoribbons on Au(111)

“…Perdew-Burke-Ernzerhof exchange-correlation functional, 32 which has a good performance in modeling metal substrate. 33 A non-local vdw-df functional 34 was used to achieve higher accuracy in the simulation of van der Waals interaction between polymer and Au, 35 which is shown to be a good choice to produce accurate adsorption energies for graphene on Au, 36 especially on substrate steps. 28 The energy cutoffs were 30 Ry and 300 Ry for the wavefunctions and charge density, respectively.…”