[Pw_forum] Problem with constrained_magnetization='atomic'

[Fabio Bernardini]

Dear all,

I want to try a calculation of constrained_magnetization='atomic' for 
Fe, but it was impossible.
I started with an example I found in the pw's forum for Cr AFM and it 
works perfectly (ref: Cyrille Barreteau cyrille.barreteau at cea.fr
Wed Dec 2 14:19:44 CET 2009)
Then I substituted Cr with Fe and I tried to fix the atomic magnetic 
moments to a given value.

Here my imput:
&control
     calculation = 'scf'
     restart_mode='from_scratch',
     prefix='chromium',
     tstress = .true.
     tprnfor = .true.
     pseudo_dir = './',
     outdir='./scratch/'
/
&system
     ibrav=  1, celldm(1) =5.4235, nat=  2, ntyp= 2, nspin = 2,
     constrained_magnetization='atomic', lambda=1.0,
     starting_magnetization(1)=1.0,starting_magnetization(2)=1.0,
     ecutwfc =90,ecutrho = 900,
     occupations='smearing',smearing='mp',degauss=0.005

  /
&electrons
     diagonalization='cg'
     mixing_mode = 'plain'
     mixing_beta = 0.7
     conv_thr =  1.0d-8
  /
ATOMIC_SPECIES
Fe1 55.847  Fe.pbesol-spn-rrkjus_psl.0.2.1.UPF
Fe2 55.847  Fe.pbesol-spn-rrkjus_psl.0.2.1.UPF
ATOMIC_POSITIONS
Fe1 0.00 0.00 0.00
Fe2 0.50 0.50 0.50
K_POINTS automatic
10 10 10 0 0 0

It never converges to the constrained value I want.
Here a digest of the output


      Program PWSCF v.5.0.2 (svn rev. 9392) starts on  8Apr2015 at 17:29:26

omissis ....

     Parallel version (MPI), running on     8 processors
      R & G space division:  proc/nbgrp/npool/nimage =       8

omissis ....

  ==============================================================================
      atom number    1 relative position :    0.0000   0.0000   0.0000
      charge :    14.420685
      magnetization :          2.265917
      magnetization/charge:    0.157130
      constrained moment :     1.000000

  ==============================================================================

  ==============================================================================
      atom number    2 relative position :    0.5000   0.5000   0.5000
      charge :    14.420691
      magnetization :          2.265879
      magnetization/charge:    0.157127
      constrained moment :     1.000000

  ==============================================================================

the final value of the magnetization is always very different from the 
target value.

I made a lot of tests. I changed the pseudopotential, I changed the 
value of lambda to 5, I tried AFM ordering, never getting the correct 
constrained value.

I tried instead the tot_magnetization=2.0 and it worked.

I use INTEL compiler :
Intel(R) Fortran Intel(R) 64 Compiler Professional for applications 
running on Intel(R) 64, Version 11.1    Build 20100414 Package ID: 
l_cprof_p_11.1.072
Copyright (C) 1985-2010 Intel Corporation.  All rights reserved.
FOR NON-COMMERCIAL USE ONLY

with this UNIX version:
Linux tako 2.6.32-5-amd64 #1 SMP Tue May 13 16:34:35 UTC 2014 x86_64 
GNU/Linux

I ask for help and I am ready to do more tests if it helps to find out 
where is the problem.

Thank you in advance

Fabio Bernardini

-- 
                   Fabio Bernardini
   ADDRESS :
             Dipartimento di Fisica
             Universita' di Cagliari,
             Cittadella Universitaria
             S.P. Monserrato-Sestu Km 0.700
             I-09042 MONSERRATO (CA), Italy
             Cod. Fis.: 80019600925                            |
             P. IVA : 00443370929                                |

      fax +39 070 6753191
      e-mail: fabio.bernardini at dsf.unica.it