The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

--- Begin Message ---

• From: "Narayani Choudhury" <n.choudhury@gl.ciw.edu>
• To: <gonze@pcpm.ucl.ac.be>
• Subject: Re: [abinit-forum] ASR: large changes in phonon freq
• Date: Sat, 13 Dec 2003 22:40:54 -0500 (EST)
• Importance: Normal

Dear Xavier Gonze,

Thank you very much. yes, that resolves the convergence problem.

Thanks again,

Best wishes,
Narayani

> Dear Narayani Choudhury,
>
> You are studying Pb3MgNb2O9, that should behave, on average,
> like PbZrO3 or PbTiO3. Such materials have been already
> treated by ABINIT, and are good insulators.
> However, you seem to observe a convergence problem.
> So I suspect the following : there is a degeneracy in the
> conduction bands, and the automatic addition of conduction
> bands done by ABINIT, instead of stabilizing the convergence,
> screws it. In this case, the solution might be very easy :
> just set the "nband" input variable by hand, to 62 ,
> that is, the real number of occupied bands for your system.
> You will treat only valence bands, and the convergence problem should
> disappear.
>
> Xavier Gonze
>
> -------------------------------------------------------------------
>
> Narayani Choudhury wrote:
>> Hi,
>>
>> I undertook a linear response calc. of a supercell of a perovskite
>> structured (natom =15) material using ABINITv403.
>>
>> The calculations use OPIUM generated pseudopotentials: pspcod =6;
>> pspxc=2
>>
>> I undertook the long wavelength (q=0) rf calc using 2x2x2 k-pt
>> sampling gs: iscf=3; ixc =2; tolvrs=1.0d-18, kptopt =1
>> ddk: iscf=-3; ixc=2, tolwfr=1.0d-22; kptopt=2
>> dde: iscf=3; ixc=2, tolwfr=1.0d-22; kptopt=2
>> rf: iscf=3, ixc=2, tolwfr=1.0d-18, kptopt=2
>>
>> Apart from the ground state, the ddk, dde and rf runs had slow
>> convergence taking over 400 iterations to converge. The calculations
>> indicated several extremely imaginery frequencies (very
>> unrealistic). The acoustic modes were around 40 cm-1 without asr. I
>> then used anaddb to apply the acoustic sum rule corrections. It
>> changes the phonon frequencies just too much!!!!! can anyone help
>> with what's what's happening? The DDB fille is attached. will
>> improving the k-pts resolve this? Can anyone please help?


Narayani Choudhury
Geophysical laboratory,
Carnegie Institution of Washington,
5251, Broad Branch Road, N.W.,
Washington D.C. 20015
Phone: +1-202-478-8945
email: n.choudhury@gl.ciw.edu





--- End Message ---