# To reproduce results from Paper # Preparation using VASP ############################################################# The OMEN input are prepared in the PrepData folder. In PrepData run the script "create_symlinks.sh" to create all necessary symlinks. ## VASP You need to run the VASP simulations in the following folders 1. relax 2. scf 3. bandstructure 4. wannier 5. dielectric 1. Here a short relaxation is done. Simply run the VASP executable 2. Here a the plane wave functions are determined with high accuracy. The KPOINTS file corresponds to the one that reached convergence. Make sure to comment the NBANDS parameter first and then run a second simulations (with NELMIN = 5 for convergence) once the charge is determined to save simulation time. 3. Here the eigenvalues are calculated on a path. So that they can be compared to the results obtained by wannier90 on k-points that are not explicitely fitted. 4. Preparation of the wannier inputs. The following keywords need to be added to the INCAR of the scf calculation. LWAVE = .FALSE. LCHARG = .FALSE. ALGO = None LWANNIER90 = .TRUE. LWANNIER_MMN_AMN = .TRUE. The wannier90.win file provided corresponds to the one with optimal projections. In order to reproduced the results its best to delete the unit_cell, atoms_car, kpoints and mp_grid keywords. Sometimes the interface between VASP and wannier90 has problems with keywords from the wannier90.win file. In most cases simply commenting them suffices. 5. Calculation of the dielectric constant. The isotropicd dielectric constant eps_iso can be calculated as in the following paper "Sohier, T et. al. Phys. Rev. B, 2016, 94" ## wannier90 ############################################################# After the wannier90.mmn, wannier90.amn and wannier90.eig files are created by the VASP-wannier90 interface you run the wannier90 tool. Make sure the write_hr is set to obtained the wannier90_hr.dat file that contains the Wannier Hamiltonian. ## winterface ############################################################# The winterface connects the wannier90 and OMEN tools. Two command files - winput_old and winput_new are provided because the tool development continued during this study. The new tool should be used from here on. The main parameters to transfer, if not commented out in winput_old, into the winput_new are - xyz - C - l - tol To obtain the Hamiltonians run ltool -iwinput_new ltool -hwinput_new The files needed for the OMEN simulations are - H_3.bin, H_4.bin, H_5.bin containing the Hamiltonian - lattice_dat containing the channel structure - ph_mat_par containing material parameters The last two files may be compared to the ones found in the BinaryData folder. Differences may occur due to the latest Version of winterface. The best way to be sure of a correct reproduction is to compare the basis of as well as the number of all species of the lattice_dat obtained to the one in the BinaryData folder. Note that the basis vector in the confined direction might differ because it is changed during the OMEN simulation. # OMEN ############################################################# The OMEN simulations are carrier out as follows in the folder SimData/electron and SimData/hole run1: Is used to fit the grid to the poisson solver. Therefore grid.dat is then used to adapt tox run2: First simulations to fix the OFF-state current. Results in SCATTEL_Id_0.dat are used to determine new parameter phi_m. run3 and run4 As the first Simulation to determine the right phi_m iteratively Lg_15nm: Full IV curve starting from the results found in run4 Lg_Xnm with X={ 12.5, 10, 7.5, 5} Simulating two gate voltages with in the OFF-state domain to extract the subthreshold slope. To achieve this without again determining phi_m the gate voltages are shifted such that the currents are close enough to the OFF-state value. Transmission_new: Ballistic Simulations used to extract the pass factor from the transmission.