[molpro-user] Difference between gradient norm of OPTG and frequencies

[Jonathan Richard Church]

Hello,

I am searching for a transition state of an open shell system, so that I
may do an IRC run, but I am having a problem with a difference between the
gradient norm of OPTG and the frequencies program.

 For example, the OPTG portion of the program yields the following result:

 ITER.   ENERGY(OLD)    ENERGY(NEW)      DE          GRADMAX     GRADNORM
  GRADRMS     STEPMAX     STEPLEN     STEPRMS   CPU-time
   1  -611.71022228  -611.71022505    -0.00000277  0.00066570  0.00126076
0.00014008  0.01196757  0.01448116  0.00647617   3356.63
   2  -611.71022505  -611.71022505     0.00000000  0.00001985  0.00002256
0.00000251  0.00024485  0.00025603  0.00011450   3380.28
   3  -611.71022505  -611.71022505    -0.00000000  0.00000063  0.00000079
0.00000009  0.00000619  0.00000662  0.00000296   3401.71
   4  -611.71022505  -611.71022509    -0.00000004  0.00000010  0.00000011
0.00000001  0.00000602  0.00000604  0.00000270   3422.78

 END OF GEOMETRY OPTIMIZATION.

 Current geometry (xyz format, in Angstrom)

    5
 KS-SCF013/VQZ  ENERGY=-611.71022509
 O          1.7001672378        0.4242808265        0.2676071916
 H          2.7671125820        0.6580142934        0.3672988072
 O          1.7813402115       -0.7667779977       -0.1928082303
 H          2.8714069152       -0.8790215302       -0.2268308576
 Cl         4.3510332011       -0.0235287645        0.1171306274

However, when molpro calculates the vibrational frequencies of the
transition state that was found, the gradient norm produced is completely
different than the 0.00000011 value, i.e.

Atomic Coordinates

  Nr  Atom  Charge       X              Y              Z

   1   O     8.00    3.212850445    0.801774562    0.505704301
   2   H     1.00    5.229084936    1.243466801    0.694094151
   3   O     8.00    3.366245134   -1.449000414   -0.364354750
   4   H     1.00    5.426172662   -1.661109950   -0.428648198
   5   CL   17.00    8.222261109   -0.044462921    0.221344807

 Frequencies dumped to record   5400.2

 Gradient norm at reference geometry: 0.11657D-04

So the OPTG program is telling me that at a certain geometry the norm of
the gradient is about 1E-7 and the frequency program is telling me that the
norm is about 1E-5.

I am new to molpro, so I’m almost entirely sure this is user error. If
someone could help me and point out my mistake, I would greatly appreciate
it.


 Listed below is my dft calculation procedure and input from my input file.

proc freqdftTS
nogprint,variable
if(#functional.eq.0) then
  if(#df.ne.0) functional=df
  if(#dftname.ne.0) functional=dftname
  if(#functional.eq.0) functional=b3lyp
end if
ks,$functional, grid=1.d-10,
accu,16
optg, maxit=5000, hesscent, gradient=1.d-7, FREEZE=1.d-15;
NUMHESS, hstep=10, displace=CART;
root, 2;
coord,3n
{frequencies, central, sym=no
thermo,sym=c1
print,thermo}
set,program='FREQ[KS/$functional]'
energz=energy+zpe
saveenergy=energy
saveresults
set,program='ZPE'
saveenergy=zpe
saveresults
set,program='E+ZPE'
saveenergy=energz
saveresults
set,program='HTOTAL'
saveenergy=htotal
saveresults
set,program='GTOTAL'
saveenergy=gtotal
saveresults
printresults
endproc


Input file:

  include procedures
  gthresh,twoint=1.e-16
  gthresh,oneint=1.e-16
  gthresh,energy=1e-10
  symmetry,nosym,noorient
                ! don't use symmetry
  geometry={
   5
   TS2
 O          1.7007301530        0.4245034769        0.2657362287
 H          2.7703758943        0.6583135479        0.3693626602
 O          1.7802717109       -0.7677960654       -0.1921081306
 H          2.8679855246       -0.8821609295       -0.2285321442
 Cl         4.3516968648       -0.0198932024        0.1179389241}
                  !geometry
  basis=vqz
      !basisis
  $functional=M08-HX            !define fucntional (optional, b3lyp is
default)
  freqdftTS



Best,
Jon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20160804/7b7a1adb/attachment.html>