Theoretial simulations for - Biphenylene network: A nonbenzenoid carbon allotrope

This database contains all the necessary data (and metadata) for DFT calculations, PP-AFM and PP-STM simulations presented in:

Qitang Fan, Linghao Yan, Matthias W. Tripp, Ondřej Krejčí, Stavrina Dimosthenous, Stefan R. Kachel, Mengyi Chen, Adam S. Foster, Ulrich Koert, Peter Liljeroth and J. Michael Gottfried :

Biphenylene network: A nonbenzenoid carbon allotrope ,

Science 372, pp. 852-856 (2021); DOI: 10.1126/science.abg4509

Altogether 12 different geometries of Biphenylene ribbons - 6 different widths and types of 2 edges (H and HF) - were calculated.

All the DFT calculations, were obtained with FHI-AIMS[1], HSE06[2] & TS-vdW[3] and light basis set (see `control.in` for more details)

The optimisations were done via FHI-AIMS (v. aims.191119.mpi.scalapack.x ). The inputs (and optimised geometries) can be found in zipped folder `opt` with separated folder for each type of ribbon. The parameters for the calculations can be found in `control.in` file. Input geometry in `geometry.in` and the optimized geometry in `geometry-opt.in`. The `run_aims_scalapack.slrm` was used for starting the calculations.

The calculations of bands (and plotted pDOS). The inputs, scripts and outputs can be found in zipped foled `bands` The FHI-aims version for these calculations was: aims.200229.mpi.scalapack.x, except for Biphenyl ribbon 6 ( aims.191119.mpi.scalapack.x ) and Biphenyl ribbon 21 ( aims.200422.mpi.scalapack.x ). The FHI-aims calculations were performed using standard inputs — `control.in` with physical settings and geometry.in with the geometries. For plotting the band structure the same way as it was done in the publication, the edges of the valence bands and conduction bands were extracted by means of the `grep_from_output.sh` scripts (except for 21H and 21HF, which are metallic) that produced `for_shift.txt` file. The important data in each directory —  `atom_proj_dos_*****_raw.dat` and `band1001.out` together with the energy shifts — were then processed and plotted with Wolfram Mathematica (11.3.0.0) notebook `GNR_DOS_plot_and_view.nb` (optionally see the pdf). The plots for each width of the ribbons are in`bandsNdos_thor***.png`. 

The PP-AFM[4] and PP-STM[5] calculations simulating CO-tip AFM and STM image are described in zipped folder `PP_simulations`; The calculations started with creation of 9x1x1 cell, because the PP-STM simulations can be done only at the Gamma k-point at the time. At the same time. the z-vector of the cell is lowered to 30 Å in order to save the memory space on the PP-AFM calculations. All of this was done via ASE[6] (by hand, not shown here). The (processed) geometries are written in `geometry.in`. The FHI-aims computations performed with aims.200422.scalapack.mpi.x. (with options saved in `control.in`) produced `cube_001_hartree_potential.cube`, containing the Hartree potential, and `KS_eigenvectors.band_1.kpt_1.out`, with the Eigen-energies and Eigen-vectors for the STM simulations. Both of these files are not saved (due to space reasons).

The PP-AFM simulations continued using the standard CO settings stored in `params.ini` using the `run_PPAFM.sh` script. The PP-AFM version was the Master branch:

>>> Krejci Ondrej committed on 25 Mar 2020 

>>> 1 parent 805f0bd

>>> commit 76f6104d52b9ce44b6bbf76ed96846bcc5433c96

The resulting images are `df_atoms_002.png`. For plotting the overlaid atoms, it is important to have `input_plot.xyz` file with geometry (in the same directory), that will be plotted. This file was done from the 1x1x1 cell geometry through ASE.

The PP-STM simulations (dI/dV simulations) were done with the (last) python2 master version:

>>> ondrejkrejci committed on 25 May 2020

>>> 1 parent b66fa0d

>>> commit 2b088e274240fd756bc8c2f53f870fe7a2e833ac

The original PPSTM_simple.py script was changed to the 2 new scripts used for the STM simulations: `PPSTM_cb.py` (conduction band) and `PPSTM_vb.py` (valence band). The Fermi Level was changed to the centre of the gap, with energies obtained of from the `for_shift.txt` created by the `grep_from_output.sh` script. The exceptions are 21H and 21HF, where the valence band was computed (just because of a history reason) through two separate runs with `PPSTM_e0.2_s.py` and `PPSTM_e0.2_pxy.py` and the final image was then created with `SUM_e0.2.py`. In these two cases, the Fermi Level was not changed, since these ribbons are metallic. 

The 2 outputs (shown in fig. S18 in the Supplementary Information) are always in `didv_******.png` files

References:

[1] V. Blum et al., Ab initio molecular simulations with numeric atom-centered orbitals. Comput. Phys. Commun. 180, 2175-2196 (2009).

[2] J. Heyd, G. E. Scuseria, M. Ernzerhof, Hybrid functionals based on a screened Coulomb potential. J. Chem. Phys. 118, 8207-8215 (2003).

[3] A. Tkatchenko, R. A. DiStasio Jr., R. Car, M. Scheffler, Accurate and Efficient Method for Many-Body van der Waals Interactions. Phys. Rev. Lett. 8 108, 236402 (2012).

[4] P. Hapala et al., Mechanism of high-resolution STM/AFM imaging with functionalized tips. Phys. Rev. B 90, 085421 (2014); P. Hapala, R. Temirov, F. S. Tautz, P. Jelínek, Origin of high-resolution IETS-STM images of organic molecules with functionalized tips. Phys. Rev. Lett. 113, 226101 (2014); https://github.com/ProkopHapala/ProbeParticleModel .

[5] O. Krejčí, P. Hapala, M. Ondráček, P. Jelínek, Principles and simulations of high-resolution STM imaging with a flexible tip apex. Phys. Rev. B 95, 045407 (2017); https://github.com/ondrejkrejci/PPSTM .