Re: CCL:Help!: BaO optimization with GAMESS
On Fri, Jul 16, 1999 at 01:15:38PM -0400, taeyun { *at * } mail.com wrote:
> Hi all:
>
> I'm tring to perform a geometry optimization and hessian analysis for BaO
with GAMESS. Unfortunately, SCF never converge, which I can not figure out.
>
> It would be greatly appreciated if you could check my input file given as
follow, and drop me a line of advice.
>
> Please send message to me directly by the email address " taeyun { *at
* } mail.com " as this would only be interested in to a person like me who
is a beginer in computational chmistry.
>
> Many thanks in advance.
>
> Yeon Ami
>
> ------input file--------
>
> $CONTRL SCFTYP=RHF MULT=1 COORD=ZMTMPC NZVAR=1
> ECP=SBK RUNTYP=OPTIMIZE MAXIT=70 $END
> $SYSTEM TIMLIM=1440 $END
> $BASIS GBASIS=SBK $END
> $FORCE METHOD=NUMERIC VIBANL=.TRUE.
> TEMP(1)=298.15,398.15,498.15,598.15,698.15
> PRTSCN=.TRUE. $END
> $GUESS GUESS=HUCKEL $END
> $SCF DIRSCF=.TRUE. $END
> $STATPT HSSEND=.TRUE. NSTEP=100 $END
> $DATA
> BaO_3.mpc
> C1
> Ba 0.000000 0 0.000000 0 0.000000 0 0 0 0
> O 2.620010 1 0.000000 0 0.000000 0 1 0 0
> $END
>
>
OK. I had a bit of time to potter about on a Sunday morning from home,
so here is what I did and some advice about how to handle this type of
problem. I am copying to CCL as others might want to add to what I
say. I'll put the final result in a FAQ on my gamess page at:-
http://lacebark.ntu.edu.au/gamess/
1. Never run a runtyp=optimize case from guess=huckel unless it is
the most trivial of organic molecule (all 1st row atoms). Always
get a single point to converge first and use the $VEC group from
the PUNCH file (*.dat) with guess=moread.
2. Always run a exetyp=check. This gave a warning about the geometry
which in hindsight I should perhaps have followed up first.
3. If straight guess=huckel does not converge try different convergence
methods, such as level shifting (shift=.T.) and diis=.T.
4. For heavy atoms sometimes guess=hcore works better.
5. Use symmetry. For BaO we expect the px and py MOs to be degenerate
so I used C4v.
I simplified the input to use straight cartesians as these are
usually the best and easiest for linear molecules.
In this case I remembered a problem with guess=huckel for ECP so
I went straight to guess=hcore with diis=.T. and shift=.T. You have
to turn off soscf - soscf=.F. I raised maxit although this rarely
works. It did not in this case. I then lowered R to 1.62. It then
converged in 23 iterations.
At this point and only at this point, you can start an optimize
run as below. It gave a final R of 2.356.
Note I commneted out $FORCE and $STATPT as neither is really needed.
$CONTRL SCFTYP=RHF MULT=1 RUNTYP=OPTIMIZE
ECP=SBK MAXIT=200 $END
$SYSTEM TIMLIM=1440 $END
$BASIS GBASIS=SBK $END
! $FORCE METHOD=NUMERIC VIBANL=.TRUE.
! TEMP(1)=298.15,398.15,498.15,598.15,698.15
! PRTSCN=.TRUE. $END
$GUESS GUESS=MOREAD NORB=16 $END
$SCF DIRSCF=.TRUE. $END
! $STATPT HSSEND=.TRUE. NSTEP=100 $END
$DATA
BaO_3.mpc
CNV 4
BA 56.0 .0000000000 .0000000000 .0000000000
SBKJ 0
O 8.0 .0000000000 .0000000000 1.6200100000
SBKJ 0
$END
--- RHF ORBITALS --- GENERATED AT 18:44:16 CST 17-JUL-1999
BaO_3.mpc
E(RHF)= -15.9445190712, E(NUC)= 3.9198072795, 23 ITERS
$VEC
1 1 3.45898912E-02 0.00000000E+00 0.00000000E+00-4.70877960E-02-2.48386369E-02
1 2 0.00000000E+00 0.00000000E+00 1.88681050E-02 7.89229375E-01 0.00000000E+00
1 3 0.00000000E+00-1.15745561E-01 2.54683877E-01 0.00000000E+00 0.00000000E+00
1 4 9.81385505E-02
2 1 0.00000000E+00-3.45717848E-02 0.00000000E+00 0.00000000E+00 0.00000000E+00
2 2 1.68715512E-02 0.00000000E+00 0.00000000E+00 0.00000000E+00 7.20510533E-01
2 3 0.00000000E+00 0.00000000E+00 0.00000000E+00 4.23850210E-01 0.00000000E+00
2 4 0.00000000E+00
3 1 0.00000000E+00 0.00000000E+00-3.45717848E-02 0.00000000E+00 0.00000000E+00
3 2 0.00000000E+00 1.68715512E-02 0.00000000E+00 0.00000000E+00 0.00000000E+00
3 3 7.20510533E-01 0.00000000E+00 0.00000000E+00 0.00000000E+00 4.23850210E-01
3 4 0.00000000E+00
4 1-5.70200110E-01 0.00000000E+00 0.00000000E+00 4.60069283E-01 2.21679249E-01
4 2 0.00000000E+00 0.00000000E+00 9.71434611E-02-5.75772983E-02 0.00000000E+00
4 3 0.00000000E+00 6.19368532E-01 7.67026987E-02 0.00000000E+00 0.00000000E+00
4 4 1.53949681E-01
5 1 3.05436239E-01 0.00000000E+00 0.00000000E+00-1.38642019E-01 6.13393382E-01
5 2 0.00000000E+00 0.00000000E+00 2.93957167E-02-8.88766045E-02 0.00000000E+00
5 3 0.00000000E+00 4.07511454E-01-6.51816753E-02 0.00000000E+00 0.00000000E+00
5 4 1.31924368E-01
6 1 0.00000000E+00-7.41809083E-02 0.00000000E+00 0.00000000E+00 0.00000000E+00
6 2 1.06874072E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00-2.00615874E-02
6 3 0.00000000E+00 0.00000000E+00 0.00000000E+00-7.16661799E-02 0.00000000E+00
6 4 0.00000000E+00
7 1 0.00000000E+00 0.00000000E+00-7.41809083E-02 0.00000000E+00 0.00000000E+00
7 2 0.00000000E+00 1.06874072E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
7 3-2.00615874E-02 0.00000000E+00 0.00000000E+00 0.00000000E+00-7.16661799E-02
7 4 0.00000000E+00
8 1 2.68770175E-01 0.00000000E+00 0.00000000E+00-7.66507354E-01-5.00892068E-02
8 2 0.00000000E+00 0.00000000E+00 1.47300984E+00 1.18090345E-03 0.00000000E+00
8 3 0.00000000E+00-2.38828490E-01-8.61285734E-02 0.00000000E+00 0.00000000E+00
8 4-6.31283095E-02
9 1 0.00000000E+00 1.84224646E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
9 2-1.45092573E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00-4.75495917E-02
9 3 0.00000000E+00 0.00000000E+00 0.00000000E+00-1.48782475E-01 0.00000000E+00
9 4 0.00000000E+00
10 1 0.00000000E+00 0.00000000E+00 1.84224646E+00 0.00000000E+00 0.00000000E+00
10 2 0.00000000E+00-1.45092573E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00
10 3-4.75495917E-02 0.00000000E+00 0.00000000E+00 0.00000000E+00-1.48782475E-01
10 4 0.00000000E+00
11 1 5.91968317E-01 0.00000000E+00 0.00000000E+00 1.78058479E+00-1.53756833E-01
11 2 0.00000000E+00 0.00000000E+00-1.10157534E+00-3.75786092E-02 0.00000000E+00
11 3 0.00000000E+00-2.44075957E-01-4.50927262E-01 0.00000000E+00 0.00000000E+00
11 4-2.89157377E-01
12 1-1.90928085E+00 0.00000000E+00 0.00000000E+00 4.85872686E-02 1.99448038E+00
12 2 0.00000000E+00 0.00000000E+00-6.79296832E-02 1.63063587E-01 0.00000000E+00
12 3 0.00000000E+00-5.03792440E-01-1.47454809E-01 0.00000000E+00 0.00000000E+00
12 4 5.42310037E-01
13 1 0.00000000E+00-3.84858838E-01 0.00000000E+00 0.00000000E+00 0.00000000E+00
13 2 1.94375882E-01 0.00000000E+00 0.00000000E+00 0.00000000E+00-9.19484227E-01
13 3 0.00000000E+00 0.00000000E+00 0.00000000E+00 1.15309208E+00 0.00000000E+00
13 4 0.00000000E+00
14 1 0.00000000E+00 0.00000000E+00-3.84858838E-01 0.00000000E+00 0.00000000E+00
14 2 0.00000000E+00 1.94375882E-01 0.00000000E+00 0.00000000E+00 0.00000000E+00
14 3-9.19484227E-01 0.00000000E+00 0.00000000E+00 0.00000000E+00 1.15309208E+00
14 4 0.00000000E+00
15 1 3.26416733E-01 0.00000000E+00 0.00000000E+00-4.39984400E-01-5.43097900E-01
15 2 0.00000000E+00 0.00000000E+00 1.51911747E-01-1.02476175E+00 0.00000000E+00
15 3 0.00000000E+00-5.66848942E-01 1.22626822E+00 0.00000000E+00 0.00000000E+00
15 4 8.46511108E-01
16 1-1.17349267E+00 0.00000000E+00 0.00000000E+00-5.35157245E-01 8.36680217E-01
16 2 0.00000000E+00 0.00000000E+00 2.86528886E-01-1.08975712E+00 0.00000000E+00
16 3 0.00000000E+00 3.07828227E-01 1.55537045E+00 0.00000000E+00 0.00000000E+00
16 4-5.36504385E-01
$END
8POPULATION ANALYSIS
BA .90933 1.09067 .98984 1.01016
O 7.09067 -1.09067 7.01016 -1.01016
0MOMENTS AT POINT 1 X,Y,Z= .000000 .000000 .318170
0DIPOLE .000000 .000000 -4.144363
Well, there you have it. Anyone care to improve my tips. This
really is a FAQ.
Cheers, Brian.
--
Associate Professor Brian Salter-Duke (Brian Duke)
Chemistry, Faculty of Science, IT and Education, Northern Territory University,
Darwin, NT 0909, Australia. Phone (61) (0)8-89466702
e-mail: b_duke { *at * } lacebark.ntu.edu.au or b_duke { *at * }
quoll.ntu.edu.au