[Pw_forum] Help: the error of computaion single atom

[Xunlei Ding]

Hi,

If you are calculating for free N atom, I think you may have made some
mistakes.

The first question:
N atom has 5 valence electrons. So it can't be calculated with spin=1,
because it has at least one unpaired electrons.

The second question:
Total_magnatization= (number of unpaired electrons)/ (total number of
electrons), and starting_magnetization is the initial value of it.
For spin=2 calculation, the starting_magnetization must be set with a
nonzero value.
As for N, it has 5 valence electrons in which 3 electrons are unpaired.
So it is reliable to set starting_magnetization(1)=3/5=0.6 . Actually
other nonzero value will also work.
And, for N atom, multiplicity=4, not 3.

btw: maybe nosym is needed in calculation of only one atom.

Yours,
Ding

shangyi wrote:

> Dear all:
> I'm a newer of pwscf(3.1.1) and troubled by the error of computation
> single N atom energy! I computed the N atom without spin using the
> following input:
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> &control
> calculation = 'scf'
> restart_mode='from_scratch',
> prefix='N.atom',
> pseudo_dir = '/home/pub/pw/pseudo/',
> outdir='/home/shyma/tmp/'
> /
> &system
> ibrav= 1, celldm(1) =20, nat= 1, ntyp= 1,
> ecutwfc =40,
> /
> &electrons
> diagonalization='cg'
> mixing_mode = 'plain'
> mixing_beta = 0.7
> conv_thr = 1.0d-8
> /
> ATOMIC_SPECIES
> N 14.0067 N.pbe-rrkjus.UPF
> ATOMIC_POSITIONS
> N 0 0 0
> K_POINTS AUTOMATIC
> 1 1 1 0 0 0
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> and it stop with the error:
> WARNING: integrated charge= 4.30000000, expected= 5.00000000
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> from electrons : error # 1
> charge is wrong
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> I have searched the maillist, but the information I got didn't solve
> my problem, so what's wrong with my input file and how to correct it?
> Another question is that I added the keyword "nspin=2, multiplicity=3"
> when I want to computate the N atom with spin, it stoped and displayed
> "some starting_magnetization MUST be set". The INPUT_PW says that
> "multiplicity fixes the final value of the magnetization".
> So is it need the starting_magnetization value when I have already set
> the multiplicity? If it's needed, what's the meaning of
> starting_magnetization and how ot set its value?
> Any advice is appreciated!
> Best wishes!
> shuix
>
> ------------------------------------------------------------------------
> Mp3疯狂搜-新歌热歌高速下
> <http://music.yahoo.com.cn/?source=mail_mailbox_footer>