ShakeNBreak: Identifying Ground-State Defect Structures (GRC 2022 Poster)

Poster presented at the 2022 GRC Defects in Semiconductors conference, in New Hampshire, USA.

Code docs here: https://shakenbreak.readthedocs.io/en/latest/
Paper here: https://arxiv.org/abs/2207.09862

Other references:
Matter Preview of Defect Structure Searching: https://www.sciencedirect.com/science/article/pii/S2590238521002733
Metastable defects : https://doi.org/10.1039/D2FD00043A
Recombination at V_Cd in CdTe (case study): https://pubs.acs.org/doi/abs/10.1021/acsenergylett.1c00380

For other research articles and updates, check out my website at:
https://seankavanagh.com/

Transcript

1. Irea Mosquera-Lois,‡ Seán R. Kavanagh, ‡ Aron Walsh, David O. Scanlon
   Identifying the Ground State Structures of Point Defects in Solids
   Scan for the
   paper link and
   a slideshow
   summary!
   I. Mosquera-Lois‡ & S.R. Kavanagh‡, A. Walsh, D.O. Scanlon, arXiv: 2207.09862 (2022)
   I. Mosquera-Lois & S.R. Kavanagh, Matter 4, 2602 10.1016/j.matt.2021.06.003 (2021)
   Arrigoni, M. & Madsen, G. K. H. npj Comp Mater 7, (1) 10.1038/s41524-021-00537-1 (2021)
   S.R. Kavanagh, D.O. Scanlon, A. Walsh, C. Freysoldt, Faraday Discuss. 10.1039/D2FD00043A (2022)
   Lany, S. & Zunger, A. Phys. Rev. Lett. 93, 156404 10.1103/PhysRevLett.93.156404 (2004)
   S.R. Kavanagh, A. Walsh, D.O. Scanlon, ACS Energy Lett. 6, 1392 10.1021/acsenergylett.1c00380 (2021)
   I. Mosquera-Lois‡ & S.R. Kavanagh‡, A. Walsh, D.O. Scanlon, JOSS (In Submission) (2022)
   The Problem: How prevalent are defect reconstructions?
   We see it in every material we investigated, including:
   Sb2
   Se3
   , Si, ZnO, a-TiO2
   , r-TiO2
   , CdTe, GaAs, CeO2
   , Sb2
   S3
   , In2
   O3
   – Pretty prevalent!
   We find a range of physical origins for these reconstructions:
   Typical defect modelling approach:
   1. Create ‘ideal’ defect in host supercell on known crystal site.
   2. Gradient-based geometry optimisation.
   This initial configuration can reside in a local minimum on the potential
   energy surface as above for VCd
   in CdTe, yielding a metastable structure
   (Tetrahedral) instead of the true ground state arrangement (Te dimer).
   Standard defect supercell relaxation
   How important can it be?
   How can we obtain the correct defect structure?
   Reconstructions can drastically
   affect:
   o Formation energies (ΔE>1 eV)
   o Concentrations
   o Transition levels
   o Recombination activity
   o Many other properties…
   a -> VSb
   formation energy
   diagram with the (metastable)
   results obtained using the
   standard supercell approach
   (dashed faded lines), and the
   correct ground states
   identified with our method
   (solid lines)
   b -> VSb
   energy level diagram
   c -> Corresponding structures
   We propose a method exploiting the localised nature of defect distortions
   and chemical intuition to efficiently navigate the potential energy surface.
   a b
   c
   Standard defect calculations can
   yield incorrect structures
   Testing our approach on a range of diverse semiconductors and insulators (listed
   above), we find:
   o High prevalence of energy-lowering reconstructions for point defects, which are
   missed by standard defect calculations.
   o Robust performance (>95% success rate for all known cases)
   o Minimal additional computational cost, demonstrated by the range of materials
   and defects included in this study.
   o Using hybrid DFT with coarse numerical accuracy for expedient qualitative
   identification of defect reconstructions, followed by fully-converged calculations.
   Code openly-available at GitHub.com/SMTG-UCL/ShakeNBreak (Python & CLI)
   User-friendly, compatible with VASP, CASTEP, FHI-AIMs, Quantum Espresso, CP2K.
   @Kavanagh_Sean_ sean.kavanagh.19@ucl.ac.uk
   

   View Slide