Dear Guido
Thank you for your reply. I have tried the unconstrained configuration. I first
checked the convergence with NM, AFM and FM starting magnetic configurations
and compared the energy of the system. In that step I observed FM has lowest
energy. Then I wanted to study SOC. Thus I added the;
noncolin= .true.
lspinorb= .true.
to the FM structure file. But then it converged to a NM structure with:
total energy = -58.70841958 Ry
Harris-Foulkes estimate = -58.70841958 Ry
estimated scf accuracy < 1.9E-12 Ry
total magnetization = 0.00 0.00 0.00
Bohr mag/cell
absolute magnetization = 0.00 Bohr mag/cell
iteration # 29 ecut= 140.00 Ry beta=0.30
Davidson diagonalization with overlap
ethr = 1.00E-13, avg # of iterations = 1.0
That is why I used : constrained_magnetization= 'atomic',
My problem is whenever I give SOC to FM it converged to NM.
Thank you
Thank youDavid PhD student
Okayama UniversityDavid
On Tuesday, November 5, 2019, 5:03:01 a.m. CST, Guido Menichetti
<[email protected]> wrote:
Dear David
Did you try without constrained magnetization?
Are you sure that the ferromagnetic configuration with all spin parallel to the
x direction corresponds to the ground state?
The reason for high "constraint energy (Ryd)" could hide behind the choice of
the arrival state.
Re-Try the calculation without the constrained magnetization flag and Try not
to start with very highly symmetric configuration.
You can also study how the "constraint energy (Ryd)" varies as a function of
Lambda.
HTHBestGuido
Il giorno mar 5 nov 2019 alle ore 11:22 David Kostov <[email protected]> ha
scritto:
Dear Guido
Thank you for your valuable comment. I tried your advises and could get the
system get converged in the x direction magnetization with following options.
&SYSTEM
nat = 6,
ntyp = 2,
degauss = 0.12,
ibrav = 0,
ecutwfc = 140,
nbnd=50,
occupations = 'smearing',
smearing = 'mp',
starting_magnetization(1)=0.1,
starting_magnetization(2)=0.0,
constrained_magnetization= 'atomic',
angle1=90,
angle2=0,
noncolin= .true.
lspinorb= .true.
/
However I have got to know from previous pw_forum's posts that with such a
constrained calculation, the constrained energy in the convergence needs to be
very close to zero. But in my case it is not. (last iteration in the vc-relaxed
run copied below)
total magnetization = 5.77 0.00 0.00 Bohr mag/cell
absolute magnetization = 5.86 Bohr mag/cell
lambda = 1.00 Ry
iteration # 24 ecut= 140.00 Ry beta=0.30
Davidson diagonalization with overlap
ethr = 1.00E-13, avg # of iterations = 2.1
constraint energy (Ryd) = 11.31368053
Can you/someone else please let me know whether I can go forward with this
configuration or it needs to change.
I do not have any idea what to do to reduce constrained energy in this system,
please advice me.
Thank youDavid PhD student
Okayama University
On Saturday, November 2, 2019, 6:23:28 a.m. CST, Guido Menichetti
<[email protected]> wrote:
Dear Kostov
First of all, please add your signature (affiliation/university/position)
About the difference between SP and SOC calculation:The magnetism in SOC
calculation in non-collinear so if you give to the code just the magnitude of
the magnetic moments it could not be enough to reach the right/desired magnetic
configuration. If you look for a specific magnetic configuration you could also
try the "constrained magnetization".
Try to add a guess on the starting angles (of the magnetic moments)
About the vc-relax: It could happen if the relaxed structure is far from the
starting one.
Are you in the same magnetic configuration?Try the relaxation again.
HTHBestG.
Il giorno sab 2 nov 2019 alle ore 03:31 David Kostov <[email protected]> ha
scritto:
I am still waiting for a reply for this issue. Thank you in advance.
----- Forwarded Message ----- From: David Kostov <[email protected]>To: Quantum
ESPRESSO Users Forum <[email protected]>Sent: Wednesday, October
30, 2019, 2:43:25 p.m. CSTSubject: Problem with SOC
Dear QE community
I want to study my FM system with and without SOC included. Thus I tried
geometry optimization, vc-relax, for the two cases.
In the FM without SOC, I got the following magnetization. In my system, there
are two magnetic atoms. So I think one atom will have 6.48/2= 3.24 Bohr
mag/cell.
total magnetization = 6.37 Bohr mag/cell
absolute magnetization = 6.48 Bohr mag/cell
However, in the FM+SOC, the magnetization results for the final structure in
the vc-relax process is zero.
total magnetization = -0.00 -0.00 0.00 Bohr mag/cellabsolute
magnetization = 0.00 Bohr mag/cell
I am little bit not clear with this situation. Can someone please clarify me
what this results tells me in the FM+SOC case ? (is this converging to a NM
case? I mean since absolute mag=0?)
But in the input script, I defined in a way the system is FM.
input in the FM+SOC.................................
&SYSTEM
nat = 6,
ntyp = 2,
degauss = 0.12,
ibrav = 0,
ecutwfc = 140,
nbnd=50,
occupations = 'smearing',
smearing = 'mp',
starting_magnetization(1)=0.1,
starting_magnetization(2)=0.0,
noncolin= .true.
lspinorb=
.true.............................................................................................
The other problem I am facing is when I check the scf calculation with the
converged final lattice and atomic coordinates (of course with the same input
SYSTEM data above), the pressure is very high at the end, around 40GPa. I m not
sure what is wrong with the calculation.
Thank you in advance
D.
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
--
***************************************
Guido Menichetti
Post-Doc researcher in Condensed matter physics
Istituto Italiano di Tecnologia
Theory and technology of 2D materials
Address: Via Morego, 30, 16163 Genova
Email: [email protected] [email protected]
[email protected]
****************************************_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
--
***************************************
Guido Menichetti
Post-Doc researcher in Condensed matter physics
Istituto Italiano di Tecnologia
Theory and technology of 2D materials
Address: Via Morego, 30, 16163 Genova
Email: [email protected] [email protected]
[email protected]
****************************************_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users _______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
