[Wien] lapwdm (bis)

[Stefaan Cottenier]

Dear Wien-users,

I did some more tests on the lapwdm 'problem' (?) I posted a few days ago
(see copy below). I could reproduce the problem in bulk bcc Fe, which makes
testing much easier. First I converge the case with spin-orbit and with the
moment in case.inso either along 001 or 100 (010 doesn't seem to work, but
all these 3 cases should be equivalent anyway). Then I run 'x
lapwdm -c -so -up' twice: once with '0 0' in case.indmc, once with '1 3'. In
both cases I see the following output in case.scfdmup:

For '0 0':
:ORB001:  ORBITAL MOMENT:  0.04682  0.00000  0.00000 PROJECTION ON M
0.04682

For '1 3':
:XOP001  2    -0.02690     0.00000     0.07372     0.04682

This is entirely as I expect: both procedures yield the same orbital moment.

But then, instead of 001 or 100, I specify 110 in case.inso (which gives 8
instead of 16 symmetry operations). Now both procedures yield *different*
orbital moments, just as I reported a few days ago for a more complicated
cell:

For '0 0':
:ORB001:  ORBITAL MOMENT:  0.03325  0.03325  0.00000 PROJECTION ON M
0.04702

For '1 3':
:XOP001  2     0.00000     0.00000     0.00000     0.00000

I can now repeat my original question on firmer grounds: "Should both
procedures (00 and 13) be equivalent as I assumed up to now, or do they
calculate different quantities ?"

Stefaan



----- Original Message ----- 
From: "Stefaan Cottenier" <Stefaan.Cottenier at fys.kuleuven.ac.be>
To: <wien at zeus.theochem.tuwien.ac.at>
Sent: Monday, September 15, 2003 12:33 PM
Subject: [Wien] lapwdm


> Dear Wien-users,
>
> I have a problem interpreting the output of lapwdm. After a k-point
parallel
> calculation with spin-orbit coupling included, I run 'x
> lapwdm -c -so -up -p' with '0 0' in the last line of case.indmc. In
> case.scfdmup, I see -- among much other information -- the orbital moment
of
> the first atom in my case.struct:
>
> :ORB001:  ORBITAL MOMENT:  0.03087  0.02869  0.00000 PROJECTION ON M
> 0.04210
>
> Now I change the last line of case.indmc to '1 3', in order to get only
the
> orbital moment written to case.scfdmup, and run lapwdm again.
Surprisingly,
> now it reads
>
> :XOP001  2     0.00047     0.00000     0.00105     0.00152
>
> In my understanding, these are two equivalent ways to obtain the orbital
> moment in wien2k, so I didn't expect numbers that are so different as
0.0421
> and 0.0015. Also, in all other cases I saw before, these two procedures do
> yield identical numbers. Is this a bug, or do these two numbers maybe
refer
> to 2 different quantities, and not both to 'the' orbital moment ? Can
> somebody comment?
>
> While trying to find an answer, I found the following information that
maybe
> can be helpful. In case.outputdm_1up, I read for the first set of k-points
> and for the first atom this sentence for '1 3' :
>
> DENSITY MATRIX IN LOCAL COORD. SYSTEM of repre. atom
>  spin coordinate system according to case.inso
>
> Apparantly, with '1 3' lapwdm takes into account information from
case.inso,
> while with '0 0' it maybe does not? This is my case.inso:
>
> WFFIL
>  4  1  0                      llmax,ipr,kpot
>  -10.0000   1.50000           emin,emax (output energy window)
>    1.  1.  0.                 direction of magnetization (lattice vectors)
>  0                            number of atoms for which RLO is added
>  0 0 0 0 0                    number of atoms for which SO is switch off;
> atoms
>
> All other cases that worked well, had two zero's in the direction of
> magnetization, this case is the first I try with the moment not along one
of
> the main axes of case.struct. Maybe that is related to my problem?
>
> Thanks,
> Stefaan