The ABINIT Users Mailing List ( CLOSED )

[Xifan Wu <wuxifan@physics.rutgers.edu>]

Hi Waigen,

The first-principles theories is a zero-temperature(ground state) theory,
by using quantum mechanics to solving the solid problems.
The interatomic force constant, the internal strain couplings and the
bare ealstic (clamped elastic constant) are accuratly and efficiently
predictly predicted by perturbation theories based on first-principles.

According to my experience, we compared the experimental results of
ZnO and rhombohedral BaTiO3 at low temperature, the comparison is
quite good within 10% error bar(PHYSICAL REVIEW B 72, 035105(2005)).

You might also interested in having  a look at the paper to 
Review of Modern Physics, 73, 515 (2001) 
and Phys. Rev. B 71, 035117 (2005) (the implementation paper of
bare elastic constant based on perturbation theories in ABINIT)


best,

Xifan Wu

On Thu, 3 May 2007, Zhang, Waigen wrote:

> Dear Wu,
> I really appreciate your reply. Right now I am struggling to make outputs seem
> more reasonable.However,no theoretical data are available now and all the 
> structural
> datas are from experiment for the material I am working on.
> The results from ABINIT seem not so agree with experiment. I checked the 
> convergence
> and still the answer seems not so friendly,not only for these elastic 
> constants.I am
> wondering how much should I trust my calculation,comparing to the experiment 
> datas.
>
> Best,
> Zhang,Waigen
>
>
> -----Original Message-----
> From: Xifan Wu [mailto:wuxifan@physics.rutgers.edu]
> Sent: Wed 5/2/2007 10:38 PM
> To: forum@abinit.org
> Subject: Re: [abinit-forum] negative elastic constant
>
> Hi Waigen,
>
> The physical elastic constant (relaxed-ion one) will be coupled with the
> pseudo-inverse of the force constant matrix through the internal-strain
> matrix. The output shows that there is probably a negative eigenvalue of
> force-constant matrix in your system, which means that the energy is not
> at the global minium point with respect to the ionic displacements and
> means that the claculated structure maybe not physical stable, or maybe
> wrong, or the calculations is not converged yet, etc.
>
> You could check the input structure, the convergences. And you could also
> take a reference at the paper, PHYSICAL REVIEW B 72, 035105(2005) and
> maybea good way to make sure the calculation is correct is try to
> reproduce some numbers already been calculated.
>
> good luck,
>
> Xifan Wu
>
> On Wed, 2 May 2007, Zhang, Waigen wrote:
>
> > Dear all,
> > I hate to ask question about Physics concept,but....
> > I calculated the elastic tensor with DFPT(monoclinic structure),and the 
> > outcomes are
> >
> > Elastic Tensor (relaxed ion) (unit:10^2GP):
> >  (at fixed electric field boundary condition)
> >
> >   1.1283604  -3.1643805  -2.0347373   0.0000001   1.3721603  -0.0000001
> >  -3.1643805 -13.2059771  -2.2664645   0.0000000   0.4939964   0.0000000
> >  -2.0347373  -2.2664645  -0.0102671   0.0000000   0.7978253  -0.0000001
> >   0.0000001   0.0000000   0.0000000  -1.1180576   0.0000000  -0.7148172
> >   1.3721603   0.4939964   0.7978253   0.0000000  -0.5816261   0.0000001
> >  -0.0000001   0.0000000  -0.0000001  -0.7148173   0.0000001  -2.3463341
> >
> > Unstable eigenvalue detected in force constant matrix at Gamma point
> >  The system under calculation is physically unstable.
> >
> > Dose that mean the structure is mechanically not stable? If so,can I say that 
> > this structure will transfer to some other structure without any external 
> > force?
> >
> > Thanks a million.
> >
> >
> > Zhang,Waigen