[Pw_forum] strange population analysis result, in PW v 5.4

[Yi Wang]

Dear Dr. Paolo,

Many thanks for the explanation!

I read make_pointlists.f90 to understand this. About the radius, I think,  
would not it be reasonable to calculate distmin in a.u. cartesian  
coordinates, and then directly get r_m(nt) in a.u. unit, finally give its  
equivalent of alat unit as r_m(nt)/alat? That would ensure a same  
integration for all cases (and won't be harmful if an option for r_m is  
directly allowed in input)

Yi Wang

-- 
Yi Wang
Ph.D candidate at Nanjing University of Science and Technology

On Mon, 05 Sep 2016 18:00:25 +0800, Paolo Giannozzi  
<p.giannozzi at gmail.com> wrote:

> In the first case, lattice parameter alat = the value of celldm(1);
> in the second case, alat = length of the first cell vector ("a" in  
> crystallography).
> There is a factor sqrt(3) difference.
> The "atomic magnetization" (whatever it means) is computed by  
> integrating over a sphere whose radius is a fixed >fraction of the  
> lattice parameter (look for r_m in the output). So the two  
> magnetizations differ even if >everything else is the same
>
> Paolo
>
>
>
> On Sat, Sep 3, 2016 at 5:19 AM, Yi Wang <wangwy13 at gmail.com> wrote:
>> Hi, dear developers,
>>
>> I'm using PW v5.4 to do some calculation, I found the population  
>> analysis
>> printed during scf calculation has a strange behavior depending on the
>> settings of cell_parameter.
>>
>> bellowing is the input where I found the strange thing:
>> as you may see, here, I use celldm(1) and alat to control the cell
>> parameters. I got a m=1.36 in this case.
>> If I change the cell parameters in the form of "CELL_PARAMETERS bohr", I
>> got m=1.96.
>> In both cases, the "total magnetization" is 1.96 (the cutoff is not
>> converged for the potential, but anyway, potentials and cutoffs are
>> irrelevant, I've checked both of them)
>> The geometry is essentially the same, but the analysis is very  
>> different,
>> energy, "total" and "absolute" magnetizations are not affected, of  
>> course.
>>
>> May I also ask is the analysis printed during scf is the Mulliken  
>> analysis?
>>
>>
>> &control
>>    calculation = 'scf' ,
>>    prefix = 'pwscf' ,
>>    outdir = './tempfft/' ,
>>    pseudo_dir = './' ,
>>    restart_mode = from_scratch ,
>>    disk_io= 'none'
>> /
>> &system
>>    ibrav = 0 ,
>>    ecutwfc = 80 ,
>>    occupations = 'tetrahedra' ,
>>    nspin=2,
>>    nbnd= 18 ,
>>    nat = 1 ,
>>    ntyp = 1 ,
>>    use_all_frac=.TRUE.
>>    celldm(1)=2.7120201830,
>>    starting_magnetization( 1 )= 0.26,
>> /
>> &electrons
>>    conv_thr = 1d-12 ,
>>    diagonalization = 'david' ,
>>    mixing_mode = 'plain' ,
>>    startingpot = 'atomic' ,
>>    startingwfc = 'atomic+random' ,
>>    mixing_beta = 0.12 ,
>>    mixing_ndim =  10,
>> /
>> CELL_PARAMETERS alat
>> -1 1 1
>> 1 -1 1
>> 1 1 -1
>> ATOMIC_SPECIES
>>   Fe 55.845 Fe.gth.upf
>> ATOMIC_POSITIONS bohr
>> Fe       0 0 0
>> K_POINTS automatic
>>    20 20 20 0 0 0
>>
>>
>> Thanks for your attention.
>>
>>>> Yi Wang
>>
>> --
>> Yi Wang
>> Ph.D candidate at Nanjing University of Science and Technology
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://pwscf.org/mailman/listinfo/pw_forum
>
>
>
> --Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160905/8e746827/attachment.html>