The ABINIT Users Mailing List ( CLOSED )

[Yong Liu <occupierliu@yahoo.com.cn>]

Dear Hamann:

   Thank you for your advice. I'm doing the test of the convergence on the ngkpt and ecut. So far I find that the 0.01% difference in total energy can cause 3% difference in relaxed-ion C11. And My C11 is 300GPa, larger than the results in the reference, 1.physics review B, 49:5828-5844(1994), C11=209GPa, 2.Journal of materials research, 12:2623-2627(1997), C11=237GPa. Is This attributed to the ABINIT code?

   Thanks again.

> Dear occupierliu (?):
>
> I'm not sure why your acoustic phonons are not closer to zero at gamma, 
> but the mathematics used to compute the atom-relaxation contribution to 
> the elastic constants projects the 3 acoustic modes out of the 
> interatomic-force-constant matrix prior to inverting it, so your 
> relaxed-atom elastic constants should not be sensitive to the explicit 
> errors this warning flags. Since you've done convergence testing 
> already, simply look at the elastic constants. If they appear to be 
> converged, they are. The warning about the acoustic mode eigenvalues 
> serves as a possible indication that the phonons that matter for the 
> relaxation correction, the various optic modes, might not be converged 
> either, so at least some testing is warranted when the warning appears.
>
> I hope this helps, and that your elastic constants are
>  adequately converged.
>
> Don Hamann


Sincerely.
Yong Liu

-------------------------------------------------------------------------------------
Yong Liu,
Institute of Inorganic Materials,
Department of Materials Science and Engineering,
Zhejiang University,PR China
Doctoral student

---

雅虎邮箱超强增值服务－2G超大空间、pop3收信、邮件提醒