[QE-users] pw2casino mismatch energy error

[Wai-Ga Ho]

Hello QE community,


I am currently attempting to extract the kinetic energy contributions to the total energies of different polymorphs of MgO.


I understand that we need to use the -pw2casino flag if we want our PWSCF output to print a more detailed breakdown of the total energy in terms of its contributions. This means that using the command:


mpirun -np 4 pw.x  <project.in> project.out


will produce a project.out file which contains some text like this:


!    total energy              =    -183.20894402 Ry
     Harris-Foulkes estimate   =    -183.20894409 Ry
     estimated scf accuracy    <       0.00000009 Ry

     total all-electron energy =      -551.326109 Ry

     The total energy is the sum of the following terms:

     one-electron contribution =     -85.45759185 Ry
     hartree contribution      =      47.93001020 Ry
     xc contribution           =     -25.22704816 Ry
     ewald contribution        =     -95.91150866 Ry
     one-center paw contrib.   =     -24.54280555 Ry

while adding the -pw2casino flag and running:

mpirun -np 4 pw.x -pw2casino <project.in> project.out


will print the original lines above, as well as additional lines like:

 Energies determined by pw2casino tool
 -------------------------------------
 Kinetic energy      21.442363717309750       au  =     42.884727434619499       Ry
 Local energy       -84.609446621660297       au  =    -169.21889324332059       Ry
 Non-Local energy    20.438286977076853       au  =     40.876573954153706       Ry
 Ewald energy       -47.955754330605572       au  =    -95.911508661211144       Ry
 xc contribution    -12.613524077939315       au  =    -25.227048155878631       Ry
 hartree energy      23.965005100370561       au  =     47.930010200741123       Ry
 Total energy       -79.333069235448022       au  =    -158.66613847089604       Ry

>From the above output lines, we can see that the total energies (-183.20894402 Ry and  -158.66613847089604 Ry) are pretty terribly mismatched. I find that this mismatch occurs when using either PWSCF v5.0.2 and PWSCF v6.0. When I was using PWSCF v6.0, quantum-espresso actually prints an "Energy mismatch error" (I'd paste the error below but my laptop, which runs v6.0, is currently in the shop), so at least it recognizes that something is wrong.

I've tried looking through some other threads and attempted to apply fixes that worked for other people, such as adding the following line:

assume_isolated = 'mp'

to my &CONTROL in my input file (source: http://lists.quantum-espresso.org/pipermail/users/2015-April/032059.html), but nothing so far has worked.

Does anyone have any idea what the issue is and how to resolve it?  Because with the demonstrated mismatch in total energies, I do not have faith in the kinetic energy value that is output. If this helps, I am using the following pseudopotentials:

file Mg.pbe-spnl-kjpaw_psl.1.0.0.UPF: wavefunction(s)  2P renormalized
file O.pbe-nl-kjpaw_psl.1.0.0.UPF: wavefunction(s)  2S 2P renormalized

and below is my input file for MgO in a CsCl-type crystalline arrangement (which generated the energy values I gave above, using PWSCF v5.0.2):

&CONTROL
    calculation = "scf"
    max_seconds =  8.64000e+04
    outdir      = "./"
    prefix      = "MgO_CsCl"
    pseudo_dir  = "/home/davidho/.burai/.pseudopot"
    title       = "Working MgO CsCl Structure(SCF)"
    wf_collect  = .TRUE.
/

&SYSTEM
    a           =  2.65994e+00
    degauss     =  1.00000e-02
    ecutrho     =  2.00000e+03
    ecutwfc     =  2.00000e+02
    ibrav       = 1
    nat         = 2
    ntyp        = 2
    occupations = "fixed"
    smearing    = "gaussian"
    assume_isolated = 'mp'
/

&ELECTRONS
    conv_thr         =  1.00000e-06
    electron_maxstep = 200
    mixing_beta      =  7.00000e-01
    startingpot      = "atomic"
    startingwfc      = "atomic+random"
/

K_POINTS {automatic}
10 10 10  0 0 0

ATOMIC_SPECIES
Mg     24.30500  Mg.pbe-spnl-kjpaw_psl.1.0.0.UPF
O      15.99940  O.pbe-nl-kjpaw_psl.1.0.0.UPF

ATOMIC_POSITIONS {angstrom}
Mg      0.000000   0.000000   0.000000
O       1.329970   1.329970   1.329970


Any help and further insight into both the problem with and/or the method of energy decomposition would be greatly appreciated.

Thank you,
Wai-Ga David Ho, Department of Physics at Florida State University



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190221/5030611c/attachment.html>