[Wannier] MLWFs convergence

Arash Mostofi a.mostofi at imperial.ac.uk
Mon Feb 11 10:52:52 CET 2013


Dear Barak,

Increasing the kinetic energy cutoff of your plane wave basis will give 
you more accurate description of the ground state of your system, but 
there is nothing particular about the MLWF localization that requires 
use of a cutoff higher what you would usually consider to be sufficient 
for the ground state calculation.

The sampling density of the Brillouin zone can have an impact because it 
determines the accuracy of the finite-difference representation of the 
gradient operator in reciprocal space.

If your system has structural symmetry, then you would expect the 
resulting MLWFs to also have some related symmetries, which would be 
reflected in the distribution of MLWF centres and spreads.

Best wishes,

Arash

--
Dr Arash A Mostofi
Departments of Materials and Physics
Deputy Director, CDT on Theory and Simulation of Materials
Imperial College London, London SW7 2AZ, UK
+44 (0)207 594 8154 | www.cmth.ph.ic.ac.uk/people/a.mostofi

On 09/02/2013 08:24, Barak Hirshberg wrote:
> Dear Arash,
>
> Thank you for your response.
> I have tried several initial guesses and now I get good convergence on
> the spread (1 A^2 and less) for all 40 MLWFs.
> However, the maximum Im/Re ratio is still similar to what I got
> previously. (see results below)
> I was wondering if you have some advice on how to improve this ratio?
> Would using a denser k-grid be helpful?
> Should I increase my ecutwfc?
>
> In addition, I was wondering how reliable are just the wannier centers
> in this case? are they completely non- reliable or should I still be
> able to use just them. (instead of trying to plot the whole MLWF..)
>
> Thank you again for answering my very basic questions.
>
> Barak
>
> ____________________________
>
>   Final State
>    WF centre and spread    1  ( -2.569023,  0.644867, -0.100107 )
> 0.45251536
>    WF centre and spread    2  ( -1.486477,  0.440403,  0.168346 )
> 0.64402381
>    WF centre and spread    3  ( -3.048456,  0.378654, -0.122575 )
> 1.14255333
>    WF centre and spread    4  ( -0.534422,  0.581838,  0.028593 )
> 0.57677559
>    WF centre and spread    5  ( -4.492827,  1.158239, -1.043528 )
> 0.70919022
>    WF centre and spread    6  ( -3.532550,  0.742776, -0.548374 )
> 0.58766708
>    WF centre and spread    7  ( -1.418104,  0.873635, -0.024045 )
> 0.71402860
>    WF centre and spread    8  ( -1.441444,  0.490485, -0.310155 )
> 0.68002955
>    WF centre and spread    9  ( -2.285484, -0.399393, -3.731327 )
> 0.54667930
>    WF centre and spread   10  (  0.247320, -0.736405, -2.253549 )
> 0.81216245
>    WF centre and spread   11  ( -0.197519, -0.515099, -3.033429 )
> 0.55434842
>    WF centre and spread   12  ( -0.524216, -0.168141, -3.868716 )
> 0.63981439
>    WF centre and spread   13  ( -4.100941, -0.618594,  2.899069 )
> 0.71635089
>    WF centre and spread   14  ( -2.326516, -0.641118,  2.277893 )
> 0.81307041
>    WF centre and spread   15  ( -0.883865, -0.685486,  3.533116 )
> 0.73113692
>    WF centre and spread   16  ( -0.851913, -1.147348,  3.331060 )
> 0.73440009
>    WF centre and spread   17  (  3.232525,  0.916311,  0.721760 )
> 0.69679315
>    WF centre and spread   18  (  6.254851,  0.797776, -0.637205 )
> 0.70513909
>    WF centre and spread   19  (  4.394716,  0.785610,  0.562636 )
> 0.47891015
>    WF centre and spread   20  (  4.790188,  0.496984,  0.483876 )
> 1.11417634
>    WF centre and spread   21  (  0.824277, -0.511198, -2.766208 )
> 0.46445255
>    WF centre and spread   22  (  1.886636, -0.544648, -3.248580 )
> 0.66878224
>    WF centre and spread   23  (  1.968522, -0.213019, -2.913186 )
> 0.72339456
>    WF centre and spread   24  (  1.965068, -0.677025, -2.788499 )
> 0.63771742
>    WF centre and spread   25  ( -2.511676, -0.535107,  2.770006 )
> 1.16342688
>    WF centre and spread   26  ( -4.006714, -1.072571,  3.029986 )
> 0.66309170
>    WF centre and spread   27  ( -4.967071, -0.934258,  2.828672 )
> 0.57695735
>    WF centre and spread   28  ( -4.049865, -0.947572,  2.562237 )
> 0.65605546
>    WF centre and spread   29  (  5.433117,  1.141925, -0.860434 )
> 0.69987115
>    WF centre and spread   30  (  5.318741,  0.855016,  0.058766 )
> 0.59653164
>    WF centre and spread   31  (  2.343587,  0.657031,  0.670116 )
> 0.57788495
>    WF centre and spread   32  (  3.262562,  0.665619,  0.316861 )
> 0.66928716
>    WF centre and spread   33  (  6.420104, 10.384046,  0.116003 )
> 0.86756069
>    WF centre and spread   34  ( 13.512421,  7.658182, -1.030438 )
> 0.71526427
>    WF centre and spread   35  (  3.222608, -0.878029,  3.697845 )
> 0.57667797
>    WF centre and spread   36  (  5.090252,  0.655475,  0.931847 )
> 0.84379196
>    WF centre and spread   37  (  7.689943,  9.028615, -3.084478 )
> 0.64125235
>    WF centre and spread   38  (  2.890957,  0.899496, -6.064272 )
> 0.66930839
>    WF centre and spread   39  (  3.520372,  5.671793,  2.631666 )
> 0.45895973
>    WF centre and spread   40  ( -2.794288, -4.098386, -2.708644 )
> 1.15117600
>    Sum of centres and spreads ( 36.245395, 30.601379, -7.517394 )
>   28.07120954
>
>        Wannier Function Num:    1       Maximum Im/Re Ratio =    0.028881
>        Wannier Function Num:    2       Maximum Im/Re Ratio =    0.001965
>        Wannier Function Num:    3       Maximum Im/Re Ratio =    0.001104
>        Wannier Function Num:    4       Maximum Im/Re Ratio =    0.000506
>        Wannier Function Num:    5       Maximum Im/Re Ratio =    0.006402
>        Wannier Function Num:    6       Maximum Im/Re Ratio =    0.002556
>        Wannier Function Num:    7       Maximum Im/Re Ratio =    0.001814
>        Wannier Function Num:    8       Maximum Im/Re Ratio =    0.003038
>        Wannier Function Num:    9       Maximum Im/Re Ratio =    0.018246
>        Wannier Function Num:   10       Maximum Im/Re Ratio =    0.044236
>        Wannier Function Num:   11       Maximum Im/Re Ratio =    0.010289
>        Wannier Function Num:   12       Maximum Im/Re Ratio =    0.035488
>        Wannier Function Num:   13       Maximum Im/Re Ratio =    0.007110
>        Wannier Function Num:   14       Maximum Im/Re Ratio =    0.008226
>        Wannier Function Num:   15       Maximum Im/Re Ratio =    0.023429
>        Wannier Function Num:   16       Maximum Im/Re Ratio =    0.021357
>        Wannier Function Num:   17       Maximum Im/Re Ratio =    0.000995
>        Wannier Function Num:   18       Maximum Im/Re Ratio =    0.005712
>        Wannier Function Num:   19       Maximum Im/Re Ratio =    0.003256
>        Wannier Function Num:   20       Maximum Im/Re Ratio =    0.002103
>        Wannier Function Num:   21       Maximum Im/Re Ratio =    0.009841
>        Wannier Function Num:   22       Maximum Im/Re Ratio =    0.052406
>        Wannier Function Num:   23       Maximum Im/Re Ratio =    0.020228
>        Wannier Function Num:   24       Maximum Im/Re Ratio =    0.044602
>        Wannier Function Num:   25       Maximum Im/Re Ratio =    0.011048
>        Wannier Function Num:   26       Maximum Im/Re Ratio =    0.030242
>        Wannier Function Num:   27       Maximum Im/Re Ratio =    0.001397
>        Wannier Function Num:   28       Maximum Im/Re Ratio =    0.019344
>        Wannier Function Num:   29       Maximum Im/Re Ratio =    0.001303
>        Wannier Function Num:   30       Maximum Im/Re Ratio =    0.002092
>        Wannier Function Num:   31       Maximum Im/Re Ratio =    0.001234
>        Wannier Function Num:   32       Maximum Im/Re Ratio =    0.001839
>        Wannier Function Num:   33       Maximum Im/Re Ratio =    0.022842
>        Wannier Function Num:   34       Maximum Im/Re Ratio =    0.032049
>        Wannier Function Num:   35       Maximum Im/Re Ratio =    0.036329
>        Wannier Function Num:   36       Maximum Im/Re Ratio =    0.001437
>        Wannier Function Num:   37       Maximum Im/Re Ratio =    0.003444
>        Wannier Function Num:   38       Maximum Im/Re Ratio =    0.000813
>        Wannier Function Num:   39       Maximum Im/Re Ratio =    0.028603
>        Wannier Function Num:   40       Maximum Im/Re Ratio =    0.028500
>
>
>
> _______________________________________________
> Wannier mailing list
> Wannier at quantum-espresso.org
> http://www.democritos.it/mailman/listinfo/wannier
>



More information about the Wannier mailing list