[QE-users] Electron Phonon Calculation: tolerance limit for P

Thu Oct 27 18:41:22 CEST 2022

I'm not telling that one should run phonon calculations at 200 Ry. 200 Ry is to get the asymptotic behavior upon vc-relax with PseudoDojo NC PP. Then one should decrease the cutoff until the asymptotic error is around 1 meV/atom. This cutoff when you get 1 meV/atom energy error is the minimum to be used if you want accurate forces to calculate phonon vibrations.

Maybe you can expect less stringent cutoffs to work with non-organic crystals, but the figure below shows the behavior I get from various NC PP for the vc-relax of a molecular crystal (beta-HMX). ONCV and HGH require very large cutoffs for top accuracy, whereas PseudoDojo needs only 90-130 Ry (converged and accurate forces -> converged mass density). Again, the verification of the pressure tensor at the verification/validation step at the end of the output of QE is a very good proxy to get an insight on force accuracy.

[cid:f5bad075-70fd-4b5f-b74a-eaf3a20d1c79]
Xavier
________________________________
From: users <users-bounces at lists.quantum-espresso.org> on behalf of mkondrin at hppi.troitsk.ru <mkondrin at hppi.troitsk.ru>
Sent: Thursday, October 27, 2022 11:03 AM
To: Quantum ESPRESSO users Forum <users at lists.quantum-espresso.org>
Subject: Re: [QE-users] Electron Phonon Calculation: tolerance limit for P

Hello, Xavier! Is all these efforts (increasing ecut to 200 Ry) are just
make one to believe that frequency of acoustic modes at Gamma point is
zero?

Sincerely yours,
M. V. Kondrin

On Oct 27 2022, Bidault, Xavier wrote:

>Hello Ramesh,
>
> Norm-conserving (NC) pseudopotentials (PP) require higher energy cutoffs.
> You need NC PP for electron-phonon calculations. NC PP from PseudoDojo
> website are very good, at least for what I've tested on organic
> materials. Then you don't need ecutrho for NC PP, which are made to work
> with default ecutrho = 4 x ecutwfc.
>
> So, you should try first to increase your cutoff, up to 200 Ry should do
> the job to have the asymptotic behavior with PseudoDojo NC PP
> (unfortunately, other NC PP may require much higher cutoffs to get the
> asymptotic behavior). Then you can choose the cutoff where the asymptotic
> error is just below 1 meV/atom.
>
> After you do a vc-relax simulation, always check the verification step at
> the end of the output. If the pressure tensor is different from the
> pressure tensor of the last optimization step, then it means your cutoff
> is not large enough, or your system has converging issues (initial
> geometry, k-points...).
>
> I think your forc_conv_thr of 1e-4 is good. You can tighten a little bit
> etot_conv_thr to 1e-5 and press_conv_thr to 0.1 kbar.
>
> Regards, Xavier ________________________________ From: users
> <users-bounces at lists.quantum-espresso.org> on behalf of Ramesh Kumar
> Kamadurai via users <users at lists.quantum-espresso.org> Sent: Thursday,
> October 27, 2022 3:23 AM To: users at lists.quantum-espresso.org
> <users at lists.quantum-espresso.org> Subject: [QE-users] Electron Phonon
> Calculation: tolerance limit for P
>
> Dear users I am trying to calculate el-Ph coupling constant and other
> properties for a superconductor. The input file is attached. I am a bit
> stuck with VC-Relax calculation as the system is not able to stabilize at
> 0.0 kbar. I followed all the suggestions given in forum 1. Increased
> Cut-off to maximum 2. Tried USPP,NC, Mixed and currently running with
> PBE-Sol 3. Sequentially done 2 VC Relax calculation but i am not able to
> sort out the issue. &CONTROL
>    calculation   = "vc-relax"
>    etot_conv_thr =  4.00000e-05
>    forc_conv_thr =  1.00000e-04
>    max_seconds   =  1.72800e+05
>    nstep         = 100
>    outdir        = "/home/sjacobs/tmp/"
>    prefix        = "PDZ-VCRELAX"
>    pseudo_dir    = "C:\Users\RAMESH\.burai\.pseudopot"
>    tprnfor       = .TRUE.
>    tstress       = .TRUE.
>    verbosity     = "high"
>/
>
>&SYSTEM
>    a           =  6.36512e+00
>    degauss     =  7.34986e-03
>    ecutrho     =  6.40000e+02
>    ecutwfc     =  8.00000e+01
>    ibrav       = 2
>    nat         = 4
>    nosym       = .FALSE.
>    nspin       = 1
>    ntyp        = 3
>    occupations = "smearing"
>    smearing    = "marzari-vanderbilt"
>/
>
>&ELECTRONS
>    conv_thr         =  8.00000e-10
>    electron_maxstep = 80
>    mixing_beta      =  4.00000e-01
>    startingpot      = "atomic"
>    startingwfc      = "atomic+random"
>/
>
>&IONS
>    ion_dynamics = "bfgs"
>/
>
>&CELL
>    cell_dofree    = "all"
>    cell_dynamics  = "bfgs"
>    press          =  0.00000e+00
>    press_conv_thr =  5.00000e-01
>/
>
>K_POINTS {automatic}
>12 12 12  0 0 0
>
>ATOMIC_SPECIES
>Ga     69.72300  Ga.pbesol-dnl-rrkjus_psl.1.0.0.UPF
>Pd    106.42000  Pd_ONCV_PBEsol-1.0.upf
>Zr     91.22400  zr_pbesol_v1.uspp.F.UPF
>
>ATOMIC_POSITIONS {crystal}
>Zr      0.000000   0.000000   0.000000
>Ga      0.493162   0.493162   0.493162
>Pd      0.739743   0.739743   0.739743
>Pd      0.246581   0.246581   0.246581
>
>The output stress is given below
>entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= 2.67
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.41
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.82
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= 0.56
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= 0.13
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.08
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.23
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.35
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.50
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.63
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.77
>     entering subroutine stress ...
>          total stress (Ry/bohr**3) (kbar) P= -0.83 My question is What
> is the optimal pressure value to avoid negative frequencies for electron
> phonon calculation ? Is there any mistake in my input file ? Thanks in
> advance Regards Ramesh Asst. Professor GITAM University Vizag, India
>
>
_______________________________________________
The Quantum ESPRESSO community stands by the Ukrainian
people and expresses its concerns about the devastating
effects that the Russian military offensive has on their
country and on the free and peaceful scientific, cultural,
and economic cooperation amongst peoples
_______________________________________________
Quantum ESPRESSO is supported by MaX (https://nam04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.max-centre.eu%2F&data=05%7C01%7Cxavbdlt%40uic.edu%7C9c72e72d28f64de72bfa08dab834e707%7Ce202cd477a564baa99e3e3b71a7c77dd%7C0%7C0%7C638024834676401041%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=znbFHXFsDoeMMjZ5VWVWQ7b8hHxS6KjzB2wAxiNenVU%3D&reserved=0)
users mailing list users at lists.quantum-espresso.org
https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.quantum-espresso.org%2Fmailman%2Flistinfo%2Fusers&data=05%7C01%7Cxavbdlt%40uic.edu%7C9c72e72d28f64de72bfa08dab834e707%7Ce202cd477a564baa99e3e3b71a7c77dd%7C0%7C0%7C638024834676557254%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Gpwaxd%2BgDVBcWmccCNxOg2hYGeHVFUmKX%2FmIjZmgakE%3D&reserved=0
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20221027/e1ce9614/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 107355 bytes
Desc: image.png
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20221027/e1ce9614/attachment.png>