[Pw_forum] negative frequency in high q-mesh

Wei-Bing Zhang weibingzhangavh at gmail.com
Thu Jan 31 11:56:54 CET 2013


Dear Dr.  Paulatto

Thanks very much for your suggestions.


> * that after the relax you are putting back the final parameters in the pw
 > input


 Actually, I performed the relaxation firstly, then use some script
 written by myself based on ASE to read the final geometry and prepare
 the input file for phonon calculations.

> * that in the SCF calculation that precedes the phonon one, forces and

> stress are actually zero.

The final geometry relaxation give me the below information about
 forces and stress. is it small enough ??

 Total force =     0.000085     Total SCF correction =     0.000000

          total   stress  (Ry/bohr**3)                   (kbar)     P=   -0.04
   -0.00000041   0.00000000   0.00000000         -0.06      0.00      0.00
    0.00000000  -0.00000041   0.00000000          0.00     -0.06      0.00
    0.00000000   0.00000000  -0.00000003          0.00      0.00      0.00



> I would also suggest to NOT play to input parameters: you are using
 > input_dft and exx_div_threatement both of which do not apply in your case.


Right. But it should be harmless in this case( except for calculation
 using other functional pz-vbc.upf). The additional parameters are
 induced since i used a  template for scf which is prepared by PWGUI.


> Finally, ecutwfc = 65 is incredibly high for a Si ultrasoft pseudopotential.
 > Try to read something about this subject i.e. here
 > <http://www.stanford.edu/~hkulik/www/Tutorials/Entries/2012/4/17_Convergence_101.html>
 > (the "dual" is ecutrho/ecutwfc)hog


Yes, it is really good Tutorial. I have already learn a lot from here.
 ecutwfc is indeed quite large already( Although the potential
(Si.pbe-rrkj.UPF)  used here is not ultrasoft pseudopotential,
right?).  Actually, I also  perform a very strick convergence test for
cutoff and kpoints, I increase this value just want to ensure a more
safe calculation.

Best


-- 
----------------------------------------------------------------------
Dr. Wei-Bing Zhang

School of Physics and Electronic Sciences
Changsha University of Science and Technology,
Changsha,  People’s Republic of China
Phone : +86 186 73108538
E-mail: zwb256 at hotmail.com, weibingzhangavh at gmail.com
----------------------------------------------------------------------




More information about the users mailing list