[CPMD-list] population analysis mulliken
Axel Kohlmeyer
axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Mon May 30 18:31:27 CEST 2005
On Mon, 30 May 2005, zhen gong wrote:
ZG> hi, everyone!
ZG> I have a problem about the population analysis mulliken.
dear zhen gong,
there are a couple of issues with your input, that are
unrelated to the population analysis. i suspect, that
you do not have a proper wavefunction to boot.
before you can do a population analysis, you need
a good wavefunction. assuming, that you have done the
previous wavefunction optimization with the same
parameters in your input, you won't get it.
ZG> This is a test example, but the results is incredible, the "unassigned charge" is 14.
ZG> Do I have a mistake? What's the population analysis results mean?
ZG>
ZG>
ZG> ** &CPMD **
ZG> ** MIRROR **
ZG> ** RESTART WAVEFUCNTION LATEST **
ZG> ** PROPERTIES **
ZG> ** CONVERGENCE ORBITALS **
ZG> ** 1.0D-3
this is _far_ too large. to analyse a wavefunction,
you need a well converged wavefunction (1.d6-1.d8,
the lower). the default value (1.d5) is good enough
to start a CP-MD, where your wavefunction will oscillate
around the 'real' optimal structure, anyways.
**
ZG> ** ODIIS NO_RESET=200 **
ZG> ** 10 **
since you include a semicore pseudopotentials,
convergence will be somewhat tricky. you should
try with a smaller DIIS vector (e.g. 5) or use
PCG MINIMIZE.
ZG> ** MAXSTEP **
ZG> ** 3000 **
ZG> ** &END **
ZG> ** **
ZG> ** &SYSTEM **
ZG> ** ANGSTROM **
ZG> ** CHARGE **
ZG> ** 0 **
hmmm. you have Na atoms in your system,
but use charge 0 and don't do an open shell
calculation, i would (naiively?) expect, that you
either have to set charge to 4.0 or do an LSD
calcuation with MULTIPLICITY 5.
ZG> ** SYMMETRY **
ZG> ** 1 **
ZG> ** CELL **
ZG> ** 7.0 1.0 1.0 0.0 0.0 0.0 **
with a cell of this size, you probably need to add
TESR
1
to have a better convergence of the ewald sum.
ZG> ** CUTOFF **
ZG> ** 70.0 **
i don't think that 70ry is a sufficient cutoff
for a sodium semicore pseudopotential. you may
get 'away' with ~90.0 for an md (hoping for cancellation
of errors), but for analysing the wavefunction
you probably need at least 100ry, maybe 120ry.
[...]
ZG> ** 1.735 -3.545 0.479 **
ZG> ** *O_MT_HCTH.psp **
you should add the KLEINMAN-BYLANDER
flag here...
ZG> ** *Na_MT_HCTH_SEMI.psp **
... and here.
ZG> ** LMAX=P **
also the &INFO section of this file tells
you to use LOC=S here
ZG> ** 4 **
ZG> ** -1.105 1.821 -0.987 **
ZG> ** 1.275 1.207 3.051 **
ZG> ** -0.865 -0.893 0.042 **
ZG> ** -3.057 0.840 1.147 **
ZG> ** &END **
ZG> ** **
ZG> ** &DFT **
ZG> ** FUNCTIONAL HCTH **
ZG> ** &END **
ZG> ** **
ZG> ** **
ZG> ** **
finally, you should, for the sake of consistency
poject onto the pseudowavefunctions, as they are
contained in the pseudopotential files. for the
semi-core sodium potential, this makes less sense,
since you don't have a wavefunction for the valence s
state. so your &BASIS section should look like this:
&BASIS
PSEUDO AO 1
0
PSEUDO AO 2
0 1
SKIP
&END
ZG> 31 Na 1.048 1.141 0.822
ZG> UNASSIGNED CHARGE 14.061 14.061
with all those changes, even without a well converged
wavefunction (i just made a quick check). the amount
of unassigned charge will easily go below 2. and your
results should make much more sense, too.
regards,
axel.
ZG> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡zhen gong
ZG> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡legend_gongzhen at hotmail.com
ZG> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡2005-05-30
ZG>
ZG>
ZG> _______________________________________________
ZG> CPMD-list mailing list
ZG> CPMD-list at cpmd.org
ZG> http://cpmd.org/mailman/listinfo/cpmd-list
ZG>
ZG>
--
=======================================================================
Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045
D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.
More information about the CPMD-list
mailing list