Quantum Monte Carlo

Strong correlation matters in graphene and carbon compounds

Prof Sandro Sorella - SISSA (Scuola Internazionale Superiore di Studi Avanzati)

Benchmarking the interaction strength between water and graphene

Dr Andrea Zen - University College London

Thursday 27th September 2018
Time: 4pm
Venue: Ramsay LT, Christopher Ingold Building, followed by drinks and nibbles in Physics E3/7
Contact: Karen Stoneham
Tel: 0207 679 7306

Strong correlation matters in graphene and carbon compounds - Sandro Sorella

We review our understanding of the strong electron correlation by means of
a correlated ansatz based on a symmetric Jastrow factor applied to a  mean-field antisymmetric state. This wavefunction  is able to describe conventional and non conventional phases of matter, from simple metal to non Fermi liquids, from band  to Mott insulators and  from magnetic to  spin liquid phases.  

Here we apply this ansatz to study the phase diagram of isotropically strained graphene by an efficient Quantum Monte Carlo (QMC) method, allowing us to compute not only energy but also enthalpy and tensile strain, clearly fundamental thermodynamic quantities.  Following undistorted semimetallic graphene (SEM) at low strain, multi-determinant Heitler-London correlations stabilize between ~8.5\% and ~15\% strain an insulating Kekule'-like dimerized (DIM) state. The DIM state prevails over the competing antiferromagnetic insulating (AFI) state favored by  density-functional calculations which we conduct in parallel.  The DIM stressed graphene insulator, whose gap is predicted to  grow in excess of 1 eV before failure near 15\% strain,  is similar to the excitonic one reported in [2] for Carbon nanotubes, but with a different order parameter that does nor break A-B sublattice symmetry.

If time is left we will show preliminary results on small molecules with a new ansatz, based on the pfaffian mean-field[3], which looks very promising  to understand the chemical bond between atoms with non zero spins, and in particular carbon based materials.

[1] S.Sorella, K. Seki, O. O. Brovko, T. Shirakawa, S. Miyakoski, S. Yunoki  and E. Tosatti, arXiV: 1804.004479  (2018)
[2] Daniele Varsano, Sandro Sorella, Davide Sangalli, Matteo Barborini, Stefano Corni, Elisa Molinari and Massimo Rontani, Nature Communications, volume 8, 1461 (2017).
[3] M. Bajdich, L. Mitas, and L. K. Wagner, Phys. Rev. B 77,115112 (2008).

Sandro Sorella has been a professor in physics at the Scuola Internazionale Superiore di Studi Avanzati in Trieste, Italy, since 2005 and a senior visiting scientist at the Riken Advanced Institute for Computational Science in Kobe, Japan, since 2013. His scientific activity has been devoted to the study of strongly correlated electron systems by advanced numerical simulation techniques based on quantum Monte Carlo, and has covered several topics, such as quantum spin liquids, high-temperature superconductivity, hydrogen at high pressures and several carbon-based compounds, such as graphene and nanotubes. He has developed several efficient algorithms for describing the ground state properties of correlated systems, from the auxiliary field quantum Monte Carlo during his Ph.D. (1989), to stochastic optimization methods (stochastic reconfiguration in 1998) for variational Monte Carlo and the alleviation of the so called ‘fermion sign problem.’ Moreover, starting in 2008, he has been the pioneer developer of efficient molecular dynamics and structure optimization methods for the electronic simulation of material properties by quantum Monte Carlo.


Benchmarking the interaction strength between water and graphene - Andrea Zen

Molecular adsorption on surfaces plays an important part in catalysis, corrosion, desalination, and various other processes that are relevant to industry and in nature. The adsorption energy of water on graphene is particularly interesting as it calls for an accurate treatment of electron correlation effects, as well as posing a practical challenge to experiments and electronic structure methods.

Here, we employ diffusion Monte-Carlo (DMC) for calculating adsorption energies of water on graphene.  Additionally, water adsorption on benzene and coronene is investigated using DMC and linear scaling coupled cluster (CCSD(T)).  These systems are frequently considered as finite-sized graphene-like clusters and we establish excellent agreement of DMC and CCSD(T).

Different adsorption motifs on graphene have a very similar binding energy though the contributions to the interaction are quite distinct as shown by charge density analysis and symmetry adapted perturbation theory.  The binding energies from the many-body electronic structure methods agree within sub-chemical accuracy, our best estimate from DMC is -99 +/- 6 meV.  The vanishing band gap makes the convergence of the substrate towards the periodic graphene layer particularly slow and thus binding energy estimations from small clusters unreliable.

Most van der Waals inclusive DFT methods yield reasonably accurate result for the finite clusters.  In contrast, beyond atom-pairwise van der Waals interactions seem to be necessary to capture all details of the different binding motifs on graphene.


Andrea holds a M.Sc. cum laude in Theoretical Physics from the University of Padua, and a Ph.D. in Statistical and Biological Physics. from the International School for Advanced Studies of Trieste. Afterwards, he was awarded a research fellowship from the University of Rome La Sapienza, and since 2014 he has been working as a research associate at the University College London, in the group of Prof. Angelos Michaelides. 

His research interests are mainly focused on the theoretical physics of matter, and he has an extensive experience in application and development of sophisticated electronic structure computational approaches, such as quantum Monte Carlo (QMC). He has done important implementations in several ab initio packages and two years ago he introduced an improved diffusion QMC algorithm, yielding a speed-up of up to 100x (Phys. Rev. B, 2016, 93, 241118(R)). Andrea also performed the first molecular dynamics simulation of liquid water with QMC (J. Chem. Phys., 2015, 142, 144111), and he recently showed that diffusion QMC is one of the most accurate approaches computationally affordable to assess the stability of molecular crystals (Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 1724-1729).

  • Sorella_Zen.png

Follow @tyc_london for updates from the Thomas Young Centre.