Dear amber developers and users,
I am trying to reproduce PW12O40(3-)'s resp charge in the supplementary
infomation of the 2008 jpcb paper, "Molecular Dynamics Simulations of
Dendrimer-Encapsulated r-Keggin Ions in Trichloromethane Solution". I
follow their quantum chem procedure (the description is in the end of this
email): B3LYP+SBKJC basis set, and my gaussian input file reads:
%Mem=4GB
%nprocshared=24
%chk=test.chk
#P B3LYP/Gen Integral=Ultrafine Pseudo=Read
SCF(Conver=8,XQC,MaxConventionalCycles=64,MaxCycle=128) Symm=loose
5D 7F Pop=(chelpg,ReadRadii,Regular) IOp(6/33=2) GFinput GFprint
MEP step
-3 1
P -0.003 -0.000 29.997
O 0.631 1.267 30.631
...
O -0.140 -5.285 29.860
P O W 0
CEP-121G
****
P O W 0
CEP-121G
W 1.8
After gaussian calculation is done, I use antechamber:
~/software/amber12/AmberTools/bin/antechamber -i ../CEP-121G-P.log -fi
gout -o KEG.mol2 -fo mol2 -c resp -nc -3 -s 0
My electronic energy from gaussian is -1461.3 a.u., agreeing with theirs
-1462.4 a.u. but my resp charge looks strange:
1 P1 -0.0090 0.0090 0.0000 p5 1 MOL 4.362211
2 O1 0.6240 -0.6260 1.2670 o 1 MOL -1.703139
3 O2 0.2560 1.5370 0.0000 o 1 MOL -1.703139
4 O3 -1.5380 -0.2520 0.0000 o 1 MOL -1.703139
5 O4 0.6240 -0.6260 -1.2670 o 1 MOL -1.703139
6 W1 0.0590 -0.0590 3.5980 W 1 MOL 1.785263
......
17 W12 0.0590 -0.0590 -3.5980 W 1 MOL 1.785263
18 O5 0.0890 1.7440 2.9340 os 1 MOL -0.610356
......
basically, my resp charges for the center PO4 are 4.36 (P) and -1.70 (O),
for W are 1.79, and for the other 36 oxygens are -0.61. However, their resp
charges for the 12 outermost O are ~-0.4, and for other else atoms are
close to 0. I am wondering why my resp charges are so different? Am I doing
anything wrong?
I also try 6-311G(d) for P or O atoms: it does not affect resp charge much,
but lead to electronic energy of about -4100 a.u.
In addition, I assign 1.8 for W's radius. If I calculate some molecule
conataining
smaller atom, e.g. Cr or Mo, can I use the same radius of 1.8?
If I use the chelpg charge from gaussian to run MD simulation, can I get
similar
results to resp charge?
Any suggestion or comment are very welcome. I really appreciate it.
yours
bob
================= supplementary info of
JOURNAL OF PHYSICAL CHEMISTRY B 112(16): 5153–5162. *DOI*. 10.1021/jp710215u
=====================================
All Kohn–Sham density functional calculations for polyoxometalates (POM)
and dendrimer
ions were performed with the Gaussian03 program package,1 using the
gradient-corrected
B3LYP hybrid density functional2,3 (with the VWN-III, not the VWN-V
functional). Effective
core potentials (ECP) and associated spherical Gauss-type basis sets of
Stevens and
coworkers4,5 were used for all centres (no ECP, only basis set, for
hydrogen). For the polyoxometalates,
polarization and diffuse functions (see Table 1 for their exponents) were
added,
this basis set is termed SBKJC+(d, f). The basis sets for the dendrimers
were augmented by
polarization functions (taken from the Gausssian basis set database) on all
non-hydrogen
centres, this basis is termed SBKJC(d). Incremental buildup of the Fock
matrix was observed
to cause convergence problems for the POMs, and therefore was disabled. For
the structure
optimizations we used an integration grid of 99 radial shells with 590
angular points per
shell for all atoms (pruned for nonmetals, unpruned for metals). The
threshold for maximum
force was 4.5 · 10−4 a.u., the threshold for maximum displacement in
internal coordinates was
1.8 · 10−3 a.u. This lead to equilibrium structures with energies
consistent to 10−7 a.u. (the
same accuracy as for the electronic energy).
State-averaged MCSCF calculations (using the SBKJC+(d, f) basis set) for
the states
5d46s2 5D, 5d56s1 7S, and 5d6 5D of the tungsten atom were performed with
the Gamess
program.6 From the electron densities of these states, radii of density
isovalue surfaces with
½ = 0.002 a.u. were obtained. The arithmetic mean of these radii was taken
as the Lennard-
Jones parameter ¾ii for the tungsten atoms.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Oct 14 2015 - 11:30:03 PDT
