An ab-initio study of bilayer graphene using higher order quantum chemical methods

Graphite and carbon-based structures represent important test case systems for the validation of novel computational methodologies aimed at accurately describing van der Waals interactions at the nanoscale. In this context, Density Functional Theory (DFT)71,90 has played a key role when describing the ground-state properties of a wide variety of molecular systems, under local and semi-local approximations150,152 to the exchange-correlation (XC) potential. Nevertheless, non-covalent systems, such as bilayer graphene, cannot yet be accurately described under these DFT functionals due to the lack of long range correlation. In addition, they suffer from spurious self- interaction as well as the absence of discontinuity in the chemical potential58,93,140. In this sense, exchange-correlation functionals have been specifically developed to account for van der Waals interactions34,35,57,145, commonly as a post-processing correction to an initial calculation based on local or semi-local XC functionals. One successful example is the PBE+vdW functional, by A. Tkatchenko and M. Scheffler184. On the other hand highly accurate wavefunction-based methods represent an alternative, even though they come at a high-computational expense. Such methods mainly involve Møller- Plesset perturbation theory and the Random Phase Approximation (RPA) and Coupled Cluster (CC). In this work we focus mainly on the former two methods, which represent a good compromise between accuracy and computing time and are considered as the cheapest alternative to DFT. Indeed they scale as the N5 and N4 power of the number of wave functions N respectively, against the N3 scaling of local and semi-local DFT functionals. Furthermore, a representative π−π system, like the benzene dimer, shows excessive binding in the MP2 method while underbinding in the RPA method compared to accurate coupled cluster calculations84. It is expected that in an infinite system like bilayer graphene, the binding would be between the two. Moreover, RPA theory describes the dispersion interactions correctly in the infinite limit of electron densities and interlayer distances42,47,122 (i.e. the distances between two subsystems taken apart), and for such a reason the binding in a bulk material like graphite, where dispersion effects are larger due to a larger number of neighbouring atoms, is expected to be described more accurately. In our study we implemented, for such a purpose, the Møller-Plesset perturbation theory at the second order (MP2) in our in-house code FHI-AIMS19, based on numerical atomic orbitals (NAO)19, and we have applied it along with PBE+vdWand RPA/RPA+ techniques to the study of mono and bilayer graphene with the aim of understanding in particular the role of dispersion in such a bilayer graphene system.

Author dc:creator

• Sanfilippo, Andrea G.

• vdw
• van der Waals
• bilayer, graphene
• graphite
• mp2
• rpa
• rpa+
• dft

Licence dc:rights.uri

• http://www.fu-berlin.de/sites/refubium/rechtliches/Nutzungsbedingungen

Language dc:language

eng

Identifier URI

http://dx.doi.org/10.17169/refubium-6818

OAI record oai:identifier

https://refubium.fu-berlin.de/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Arefubium.fu-berlin.de%3Afub188%2F2617