[QE-users] Magnetization Keeps Converging to Zero in vcrelax

[Ing. Martin Matas]

Dear Stephen,

Make sure that your pseudopotentials contain all the electrons responsible for the magnetization. For example, based on the file name, I guess your cerium PP lacks the f-electrons.
Hope that helps

Martin Matas
Ph.D. student
University of West Bohemia


 Pátek, Srpen 07, 2020 07:40 CEST, Stephen Zhang <lolzen at berkeley.edu> napsal:

> Hi everyone,
> 
> Recently for many of my vc-relaxation calculations, the total magnetization
> of the crystal keeps falling to zero. I know this is not correct since the
> non-zero magnetization has been verified by others and this fall off to
> zero has been happening for several compounds I've tried. Could someone
> possibly enlighten me on what might be going on? I've attached my input
> file below.
> #######################################################
> &control
>   calculation = 'vc-relax'
>   prefix = 'CeAlCu'
>   outdir = './outdir'
>   pseudo_dir = '/global/scratch/lolzen/qe-6.5/SSSP_precision_pseudos'
>   etot_conv_thr = 1e-6
>   forc_conv_thr = 1e-5
> /
> &system
>     ibrav=4, celldm(1)=10.04189124, celldm(3)=0.7585576804,
>     nat=6, ntyp=6,
>     ecutwfc=60,
>     ecutrho=600,
>     occupations='smearing', smearing='gaussian', degauss=0.001,
>     nspin=2,
>     starting_magnetization(1)=0.1,
>     starting_magnetization(2)=0.11,
>     starting_magnetization(3)=0.12,
>     starting_magnetization(4)=0.099,
>     starting_magnetization(5)=0.098,
>     starting_magnetization(6)=0.097,
> /
> &electrons
>   conv_thr=1e-7,
> /
> &ions
> /
> &cell
>   cell_dofree='all'
> /
> ATOMIC_SPECIES
>  Ce1  140.116 Ce.pbe-spdn-kjpaw_psl.1.0.0.UPF
>  Al2  26.981539 Al.pbe-n-kjpaw_psl.1.0.0.UPF
>  Al3  26.981539 Al.pbe-n-kjpaw_psl.1.0.0.UPF
>  Cu4  63.546 Cu.pbe-dn-kjpaw_psl.1.0.0.UPF
>  Cu5  63.546 Cu.pbe-dn-kjpaw_psl.1.0.0.UPF
>  Cu6  63.546 Cu.pbe-dn-kjpaw_psl.1.0.0.UPF
> 
> ATOMIC_POSITIONS crystal
>  Ce1 0 0 0
>  Al2 0.333333 0.666667 0
>  Al3 0.666667 0.333333 0
>  Cu4 0.5 0 0.5
>  Cu5 0 0.5 0.5
>  Cu6 0.5 0.5 0.5
> 
> K_POINTS (automatic)
>  4 4 3 0 0 0
> #########################################################
> 
> Thanks,
> Stephen
> 
> 
> -- 
> *University of California, Berkeley*
> *Department of Letter and Sciences*