[Pw_forum] PHonon underestimates the dielectric tensor

[stefano de gironcoli]

Dear Mohamed,
   are you interested on the static dielectric constant (\epsilon_0) or 
on the so called "high frequency" static dielectric constant 
(\epsilon_\infty)?
   the second one is the one computed by the PHonon code and corresponds 
to the response to time dependent macroscopic electric fields varying 
with a frequency large compared to phonon frequencies but small compared 
to electronic excitations, the second one is the real static dielectric 
constant and includes the coupling to the lattice.
   In simple polar semiconductors they differ by 10-20%, in strongly 
polar materials I can imagine a larger effect. In the extreme case of 
liquid water the electronic epslilon_infty is something like 3-5, while 
epsilon_0 is about 80 !

   In solids, it can be easily computed from the knowledge of 
\epsilon_\infty, the effective charges and the transverse dynamical 
matrix at gamma, all quantities computed by PHonon.

   kpoint sampling is usually more delicate for dielectric properties 
than other microscopic responses, but neglecting the coupling with the 
lattice is likely a major problem.

stefano


On 06/12/2015 18:07, Mohamed Elshazly wrote:
> Hello Apsi,
>
> Thank you for your feedback.
>
> Yes, these are relaxed atomic positions. The values I got for energy and the equilibrium lattice constant were very good, and that's why I went ahead with the phonon calculation. I'm using a Martins-Troullier norm-conserving pseudopotential with a PBE XC functional.
>
> The problem is that I also tried LDA on this material system and PBE on a different material system, same problem. I'm running version 5.2.1 and the examples worked just fine, so it doesn't seem like a bug either.
>
> I will, however, try better convergence thresholds and update the thread if it fixes the problem. Thank you!
>
> ________________________________________
> From:pw_forum-bounces at pwscf.org  <pw_forum-bounces at pwscf.org>  on behalf of Ari P Seitsonen<Ari.P.Seitsonen at iki.fi>
> Sent: December-06-15 11:50 AM
> To: PWSCF Forum
> Subject: Re: [Pw_forum] PHonon underestimates the dielectric tensor
>
> Dear Mohamed,
>
>     This probably does not resolve your problem, but still: Did you check
> the convergence of the properties with respect to the cut-off energy,
> 'ecutwfc'? We traditionally used values of 80-100 Ry, but I must confess
> that I do not know exactly this potential. Also I would converge the
> final electronic structure a bit better than the default value of 10^-6
> ('conv_thr'). Are these relaxed atomic positions?
>
>       Greetings,
>
>          apsi
>
> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
>     Ari Paavo Seitsonen /Ari.P.Seitsonen at iki.fi  /http://www.iki.fi/~apsi/
>     Ecole Normale Supérieure (ENS), Département de Chimie, Paris
>     Mobile (F) : +33 789 37 24 25    (CH) : +41 79 71 90 935
>
>
> On Sun, 6 Dec 2015, Mohamed Elshazly wrote:
>
>> Hello everyone,
>>
>>
>> I'm trying to reproduce the dielectric constant of alpha-alumina using
>> PHonon. However, I'm getting values close to third of the actual dielectric
>> constant and cannot figure out why. I also had the same issue with
>> Columbite-(Mg), so it's not a particular problem with that material system.
>>
>>
>> Here are my SCF and PH input files:
>>
>>
>> &control
>> calculation='scf',
>> outdir='./out',
>> prefix='alo',
>> pseudo_dir='.',
>> tstress=.true.,
>> tprnfor=.true.,
>> forc_conv_thr=1.0D-8,
>> etot_conv_thr=1.0D-6,
>> /
>>
>> &system
>> ibrav=0,
>> celldm(1)=9.69051561,
>> nat=10,
>> ntyp=2,
>> ecutwfc=50,
>> /
>>
>> &electrons
>> diago_full_acc=.true.,
>> /
>>
>> &ions
>> /
>>
>> ATOMIC_SPECIES
>> O 15.9994 O.pbe-mt.UPF
>> Al 26.981539 Al.pbe-mt_fhi
>>
>> CELL_PARAMETERS alat
>>     0.444822702  -0.256818507   0.805845773
>>    -0.000000000   0.513637014   0.805845773
>>    -0.444822702  -0.256818507   0.805845773
>>
>> ATOMIC_POSITIONS crystal
>> O        0.058814216   0.749985721   0.441239527
>> O        0.250014279   0.558760473   0.941185784
>> O        0.441239527   0.058814216   0.749985721
>> O        0.558760473   0.941185784   0.250014279
>> O        0.749985721   0.441239527   0.058814216
>> O        0.941185784   0.250014279   0.558760473
>> Al       0.148217545   0.148217545   0.148217545
>> Al       0.351772827   0.351772827   0.351772827
>> Al       0.648227173   0.648227173   0.648227173
>> Al       0.851782455   0.851782455   0.851782455
>>
>> K_POINTS automatic
>> 16 16 16 0 0 0
>>
>> PH:
>>
>> alo_dielectric
>>
>> &INPUTPH
>> tr2_ph=1e-15,
>> outdir='./out',
>> prefix='alo',
>> epsil=.true.,
>> /
>>
>> 0 0 0
>>
>> Any help would be greatly appreciated.
>>
>> Mohamed Elshazly
>> Edward S. Rogers Department of Electrical and Computer Engineering
>> University of Toronto
>>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum