Virtual Winter School on Computational Chemistry

Cecam Logo

Density-Functional Tight-Binding for the Predictive Simulation of Complex Systems

Speaker: Dr Stephan Irle
Speaker Link: https://www.ornl.gov/staff-profile/stephan-irle
Institute: Oak Ridge National Laboratory
Country: USA

Dr Stephan Irle

Computational Chemistry and Nanomaterials Sciences Group, Computational Sciences and Engineering Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6129, U.S.A.

The density-functional tight-binding (DFTB) method [1] is an approximation to density functional theory (DFT) allowing a speedup of first principles electronic structure calculations by two to three orders of magnitude. This is achieved by solving the Kohn-Sham equations for valence electrons using a parameterized two-center Hamiltonian in a minimum pseudoatomic orbital basis set. Since electronic structure is explicitly computed for each atomic configuration, DFTB is capable of simulating chemical processes including the breaking of covalent bonds, changes in aromatic electronic structure, charge transfer, charge polarization, etc. [2]. DFTB methods can therefore be employed in atomistic molecular dynamics (MD) simulations of processes that involve complex chemical processes, electron transfer, and/or mass and ion transport. Its applicability is limited in part due to the unfavorable cubic scaling of computer time with system size, and in part due to the necessity of parameterization for element pairs. Linear-scaling algorithms for massively parallel computation [3,4] and semiautomatic parameterization codes [5] have been developed to address these shortcomings. Recently, systematic bias corrections were proposed based on a D-machine learning approach employing neural network potentials [6].

In this talk, I will first briefly review the DFTB method and its various “flavors” for including Coulombic interactions, before highlighting challenges associated with the parameterization of the Hamiltonian. DFTB-based simulations of nanoscale materials self-assembly will illustrate the predictive power of the method to unravel complex chemical processes occurring in nonequilibrium on large length scales [6].

Recording:

Video is available only for registered users.

pdfPresentation slides

References

[1] a) Christensen, A. S.; Kubar, T.; Cui, Q.; Elstner, M. Semiempirical Quantum Mechanical Methods for Noncovalent Interactions for Chemical and Biochemical Applications, Chem. Rev. 2016, 116, 5301-5337; b) http://www.dftbplus.org

[2] Cui, Q.; Elstner, M. Density functional tight binding: values of semi-empirical methods in an ab initio era, Phys. Chem. Chem. Phys. 2014, 16, 14368-14377.

[3] Nishizawa, H.; Nishimura, Y.; Kobayashi, M.; Irle, S.; Nakai, H. Three pillars for achieving quantum mechanical molecular dynamics simulations of huge systems: Divide-and-conquer, density-functional tight-binding, and massively parallel computation, J. Comp. Chem. 2016, 37, 1983-1992.

[4] a) Nishimoto, Y.; Fedorov, D. G.; Irle, S. Density-Functional Tight-Binding Combined with the Fragment Molecular Orbital Method, J. Chem. Theory Comput. 2014, 10, 4801-4812; b) Vuong, V. Q.; Nishimoto, Y.; Fedorov, D. G.; Sumpter, B. G.; Niehaus, T. A.; Irle, S. The Fragment Molecular Orbital Method Based on Long-Range Corrected Density-Functional Tight-Binding, J. Chem. Theory Comput. 2019, 15, 3008-3020.

[5] Chou, C.-P.; Nishimura, Y.; Fan, C.-C.; Mazur, G.; Irle, S.; Witek, H. A. Automatized Parameterization of DFTB using Particle Swarm Optimization, J. Chem. Theory Comput. 2016, 12, 53-64.

[6] Zhu, J.; Vuong, V. Q.; Sumpter, B. G.; Irle, S. Artificial Neural Network Correction for Density-Functional Tight-Binding Molecular Dynamics Simulations, MRS Commun. 2019, 9, 867-873 (2019).

[7] Irle, S; Page, A. J.; Saha, B.; Wang, Y.; Chandrakumar, K. R. S.; Nishimoto, Y.; Qian, H.-J.; Morokuma, K. Atomistic mechanism of carbon nanostructure self-assembly as predicted by nonequilibrium QM/MD simulations, in: J. Leszczynski, M. K. Shukla, Eds. “Practical Aspects of Computational Chemistry II: An Overview of the Last Two Decades and Current Trends”, Springer-European Academy of Sciences, Chapter 5, pp. 105-172 (April 2, 2012). ISBN 978-94-007-0922-5. DOI: 10.1007/978-94-007-0923-2_5 Preprint: https://www.dropbox.com/s/n2o3sjnb0t1z6mr/5_Online%20PDF.pdf?dl=0

4 thoughts on “Density-Functional Tight-Binding for the Predictive Simulation of Complex Systems”

  1. Tuesday, 16 February 2021 22:26

    Hi Dr. Irle,

    Thanks for the wonderful talk! Quick question, is DFTB applicable for open-shell systems? Ran some issues before in GAMESS DFTB where frequency calculations would have really large imaginary frequencies (~5000i). The structures themselves were okay when confirmed with other methods but frequencies were not. (P.S.: I know you would not remember but I was a JENESYS exchange student before at FIFC back in 2010 and have seen you around. 😁)

    1. Wednesday, 17 February 2021 02:39

      Hi Benjoe,

      Thank you for your message and also thank you for joining the JENESYS program, those were exciting times.  There should be no problem for performing open-shell vibrational frequency calculations with the GAMESS implementation of DFTB, both in spin-restricted and unrestricted (spin-polarized) versions: Nishimoto et al, Quantum chemical prediction of vibrational spectra of large molecular systems with radical or metallic electronic structure, Chem. Phys. Lett. 2017, 667, 317-321, https://doi.org/10.1016/j.cplett.2016.11.014.  However, we did not extensively test the unrestricted version, and I would be happy to take a look at the problems in your calculations if you send them to me at  

      With best regards

      Stephan

  2. Monday, 15 February 2021 21:59

    Will the recording of this interesting lecture be uploaded? Hope so!

    1. Tuesday, 16 February 2021 18:36

      Hi Moyassar, indeed, I believe it will.  In the meantime, everyone reading this, please feel free to post questions in this space, as I am checking it during the next few days.