QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: wot19920302 on April 25, 2016, 14:06

Title: repetitions and k-points sampling for dynamical matrix calculation with DFT
Post by: wot19920302 on April 25, 2016, 14:06

Dear Quantum Wise staffs:
                       After reading tutorials and posts,I have conlusions about calculating dynamical matrix with DFT, I will thank you very much if you can point out my mistakes.
                       For a nanoribbon, I can calculate  electrical energy bands of the unit cell(as shown below) using k-points 1*1*27. When I calculate dynamical matrix, I can set repetitions  1*1*3, then k-points used to DFT calculation is 1*1*9(assuming the values of repetitions along the z direction, 3,are enough).
                       For a nanosheet,  I can calculate electrical energy bands of  the unit cell using k-points 1*27*27. When I calculte dynamical matrix, I can set repetitions  1*3*3, then k-points used to DFT calculation is 1*9*9(assuming the values of repetitions along the y and z direction, 3,are enough).
                       For  1D devices,  I can calculate the electrical transport properties of the device using k-points 1*1*100(for electrodes). When I calculate dynamical matrix, I can set repetitions  1*1( the z directions is enable，the log file may yield 1*1*1, I guess) , so k-points when calculating dynamical matrix with DFT is still 1*1*100
                       By the way, when I want to calculate transmission spetrum and phonon tramsmission spetrum of 1D devices both using DFT, the order is New calculation(DFT)-Transmission spetrum-New calculation(DFT) -Dynamical Matrix-Phonon Transmission Spetrum, is that correct?
                        Best regards
                         
                       

Title: Re: repetitions and k-points sampling for dynamical matrix calculation with DFT
Post by: Jess Wellendorff on April 26, 2016, 08:36

That all sounds about right, including the order in which you compute electron and phonon transmission.

Title: Re: repetitions and k-points sampling for dynamical matrix calculation with DFT
Post by: wot19920302 on April 26, 2016, 09:38

              Thanks for your guide :D!   I konw the calculation of dynamical matrix with DFT is much time-consuming, but the clusters have been calculating electron transimission spectrum and phonon transmission spectrum of a MoS2 nanoribbon device( shown below) with DFT for  24 hours, is that normal?

Title: Re: repetitions and k-points sampling for dynamical matrix calculation with DFT
Post by: Jess Wellendorff on April 26, 2016, 10:03

You are right, calculating the dynamical matrix can be time consuming. You should be able to see from the log file if the electron transmission spectrum has finished or not. If the CPU time is currently spent on the dynamical matrix, I would let the job keep running.

Title: Re: repetitions and k-points sampling for dynamical matrix calculation with DFT
Post by: wot19920302 on April 27, 2016, 05:08

thanks for your help, sir. ;D