Dear Prem,
Please do not forget to write your affiliation. > Dear All, > I am a new user of QE and interested to calculate the optical absorption > spectra in atomic level using > TDDFT code. I have checked the ecut & > ecutfock convergence and also relaxed the structure. I can understand so far > what is going on. But when I am running turbo_lanczos.x and turbo_davidson.x, > I am getting output but how can I check that my calculation is correct ?? There is a reference data provided with the examples for the TDDFPT package. You may compare your converged spectra with the experimental ones and with other theoretical spectra in the literature. > After running turbo_davidson.x, I am getting a file named as "CH4.plot". If I > am plotting it using xmgrace, it is showing the x-axis in Ry scale but that I > want in eV ?? If you use turbo_davidson.x, plot the second column of CH4.plot (labeled as "total", oscillator strength S(E)) as a function of the first column (energy E in Ry). In xmgrace you can easily do a rescaling of the axes: in your case just make a conversion of the energy from Ry to eV by multiplying by 13.6. If you use turbo_lanczos.x, you will obtain a file "CH4.plot_chi.dat". This file contains an information about the dipole polarizability tensor (chi_i_j) and about the oscillator strength S(E) which is: S(E) ~ E * sum_i Im[ chi_i_i(E) ]. Extract the oscillator strength "S(E)" from this file using the grep command, and then plot S(E) as a function of E which is the absorption spectrum. The two codes (Lanczos and Davidson) must give you the same absorption spectrum (when both are converged). More details can be found in this paper: X. Ge et al., Computer Physics Communications, 185, 2080 (2014). > Thanks, > Prem HTH Regards, Iurii -- Dr. Iurii Timrov Postdoctoral Researcher Swiss Federal Institute of Technology Lausanne (EPFL) Laboratory of Theory and Simulation of Materials (THEOS) Lausanne, Switzerland
_______________________________________________ Pw_forum mailing list [email protected] http://pwscf.org/mailman/listinfo/pw_forum
