[molpro-user] Frequency calculation

Kirk Peterson kipeters at wsu.edu
Sun Oct 5 02:19:35 BST 2008


Ok, my mistake (sort of).  Computing numerical frequencies for  
multireference wavefunctions is tricky since some displacements will  
break symmetry and then your state designations in your casscf block  
make no sense.  That is what is happening in your current input.  The  
first displacement results in C1 symmetry, so asking for a state with  
symmetry 2 makes no sense.  I've attached an input which seems to work.

-Kirk


On Oct 4, 2008, at 5:04 PM, Jayashree wrote:

> I am now using my input with symmetry. and with the last line as you  
> suggested - {frequency,central} and I get the following message -
> PROGRAM * HESSIAN
>  Computing numerical hessian using default procedure for command  
> CASSCF
>  Symmetry elements for unique displacements: Y
>  Using no symmetry in wavefunction calculations
>  Running default procedure: CASSCF000
>   Numerically approximating hessian using central gradient differences
>  Task list generated. Total number of displacements:     21
>  ?STATE SYMMETRY NOT POSSIBLE: 2
> I have attached my input and output file.
>
> On Sat, Oct 4, 2008 at 6:41 PM, Kirk Peterson <kipeters at wsu.edu>  
> wrote:
>
> If you turn off symmetry, then both the A' and A" have symmetry A  
> and you do have two states of the same symmetry.  From the look of  
> this output though it appears that state-averaged gradients only  
> work with segmented basis sets which is not the newer program that  
> can do analytical hessians.  Your frequency calculation would then  
> calculate the hessian by numerical differentiation of the analytical  
> gradient (so disregard my cpmcscf,hess input).
>
> So I would recommend going back to your original input  (in  
> symmetry), removing the 2nd cpmcscf card, and in the frequency  
> command use:   freq,central
>
> -Kirk
>
>
> On Oct 4, 2008, at 3:49 PM, Jayashree wrote:
>
>> Two things -
>> 1. I need to have a UHF input, since I am dealing with a doublet.
>> 2. The present calculation is a conical intersection between A' and  
>> A" states. This is same as the conical intersection at the C3v  
>> geometry of methoxy molecule. In that case, I cannot have 2 states  
>> in the symmetry 1.
>>
>> But in any case, I tried the input given to me by you, and it  
>> terminates with the following message - (this is the section of  
>> output of MULTI program..the UHF and optg calculations run ok.
>> 1PROGRAM * MULTI (Direct Multiconfiguration SCF)       Authors:  
>> P.J. Knowles, H.-J. Werner (1984)     S.T. Elbert (1988)
>>  Number of closed-shell orbitals:  3 (  3 )
>>  Number of active  orbitals:       6 (  6 )
>>  Number of external orbitals:     37 ( 37 )
>>  State symmetry 1
>>  Number of electrons:    11    Spin symmetry=Doublet   Space  
>> symmetry=1
>>  Number of states:        2
>>  Number of CSFs:          6   (6 determinants, 6 intermediate states)
>>  NOEXTRA OPTION SET FOR FREQUENCY CALCULATION
>>  Molecular orbitals read from record     2200.2  Type=UHF/NATURAL  
>> (state 1.1)
>>  Wavefunction dump at record             2140.2
>>  Convergence thresholds  0.10E-05 (gradient)  0.10E-07 (energy)   
>> 0.10E-02 (step length)
>>  Weight factors for state symmetry  1:    0.50000   0.50000
>>  Number of orbital rotations:     351     (  18 Core/Active  111  
>> Core/Virtual   0 Active/Active  222 Active/Virtual)
>>  Total number of variables:       363
>>  ITER. MIC  NCI  NEG     ENERGY(VAR)     ENERGY(PROJ)   ENERGY  
>> CHANGE     GRAD(0)  GRAD(ORB)   GRAD(CI)     STEP       TIME
>>  ?WARNING, SMALL DIAGONAL HESSIAN ELEMENT FOR ROTATION 3.1 -  4.1   
>> D1E=-0.248D-06  D2E= 0.421D-07
>>  ?WARNING, SMALL DIAGONAL HESSIAN ELEMENT FOR ROTATION 3.1 -  5.1   
>> D1E=-0.632D-07  D2E= 0.135D-07
>>    1   60    7    0    -114.40430384    -114.40955138    
>> -0.00524754    0.05457236 0.00064851 0.00000000  0.44D-01      0.35
>>  ERROR EXIT
>>  CURRENT STACK:      MAIN
>>    2  201   15    0    -114.40957522    -114.40957745    
>> -0.00000223    0.00273117 0.00003244 0.00000000  0.10D-02      1.03
>>    3   60   16    0    -114.40957745    -114.40957745     
>> 0.00000000    0.00003241 0.00000000 0.00000000  0.28D-04      1.56
>>  ** WVFN ****  CONVERGENCE REACHED, FINAL GRADIENT:  0.42D-09
>>  CPMCSCF calculation required for state-averaged gradient  
>> calculation. Please add
>>  CPMCSCF,GRAD,STATE=istate.isym
>>  to the mcscf/casscf input, where istate.isym specifies the state  
>> for which the gradients are required.
>>  Note that gradients for state-averaged MCSCF can only be computed  
>> with segmented basis sets.
>>  For more details, please refer to the MOLPRO manual.
>>  ? Error
>>  ? Missing cpmcscf
>>  ? The problem occurs in musave
>> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>> If I use the following input indicating that the intersection is of  
>> states of different symmetry, it terminates with the following  
>> message -
>>  NOEXTRA OPTION SET FOR FREQUENCY CALCULATION
>>
>>  ?STATE SYMMETRY NOT POSSIBLE: 2
>>
>> My input looks like this -
>> basis=6-311++G
>> geomtype={zmat}
>> geometry={Angstrom;nosym
>>           c;
>>           o,c,oc;
>>           h1,c,h1c,o,h1co;
>>           h2,c,h1c,o,h1co,h1,120.0;
>>           h3,c,h1c,o,h1co,h1,240.0}
>> oc=1.410000, h1c=1.100000, h1co=109.100000
>>
>>  {uhf} !uhf input without symmetry
>>  put,molden,met_uhf_6311++g.molden
>>
>>  {optg,space=zmat}
>>  put,molden,met_opt_6311++g.molden
>>
>>  {casscf;
>>   occ,9;
>>   closed,3;
>>   wf,17,1,1;
>>   state,1;
>>   weight,0.5
>>   wf,17,2,1;
>>   state,1;
>>   weight,0.5
>>   start,orbital=2200.2
>>
>>   CPMCSCF,HESS}
>>  {optg,startcmd=casscf}
>>  put,molden,met_cas_6311++g_con.molden
>>  {frequencies}
>>
>> Any hope of calculating frequencies keeping symmetry in the input?
>> Thanks
>> Jayashree
>>
>> On Sat, Oct 4, 2008 at 11:21 AM, Kirk Peterson <kipeters at wsu.edu>  
>> wrote:
>> The analytical MCSCF hessian calculation can only be performed  
>> without symmetry.  I'm not sure about state-averaged wavefunctions,  
>> but you can try put a "nosym" in your geometry block after  
>> "Angstrom", comment out your rhf input, and then in your casscf  
>> input change this to something like:
>>
>>  {casscf;
>>   occ,9;
>>   closed,3;
>>   wf,17,1,1;
>>   state,2;
>>   CPMCSCF,HESS}
>>
>>
>> -Kirk
>>
>> On Oct 3, 2008, at 1:18 PM, Jayashree wrote:
>>
>>> I have located a conical intersection corresponding to the C3v  
>>> geometry in methoxy molecule. I would like to calculate the normal  
>>> mode frequencies of the molecule at this geometry. If I try to  
>>> give the command 'frequencies' below 'optg', I get the following  
>>> error:
>>>  ?STATE SYMMETRY NOT POSSIBLE: 2
>>> How do I calculate the frequencies?
>>>
>>> This is how my input file looks like -
>>> ***,methoxy
>>> memory,128,m
>>> file,2,met_cas_6311++g.wf,new;
>>> punch,met_cas_6311++g.pun
>>>
>>> gprint,basis,orbitals=6,civector
>>>
>>> basis=6-311++G
>>> geomtype={zmat}
>>> geometry={Angstrom;
>>>           c;
>>>           o,c,oc;
>>>           h1,c,h1c,o,h1co;
>>>           h2,c,h1c,o,h1co,h1,120.0;
>>>           h3,c,h1c,o,h1co,h1,240.0}
>>> oc=1.410000, h1c=1.100000, h1co=109.100000
>>>
>>>  {uhf !uhf input with symmetry
>>>   occ,7,2
>>>   open,7.1
>>>   wf,17,1,1}
>>>  put,molden,met_uhf_6311++g.molden
>>>
>>>  {optg,space=zmat}
>>>  put,molden,met_opt_6311++g.molden
>>>
>>>  {casscf;
>>>   occ,7,2;
>>>   closed,3,0;
>>>   wf,17,1,1;
>>>   state,1;
>>>   weight,0.5
>>>   wf,17,2,1;
>>>   state,1;
>>>   weight,0.5
>>>   start,orbital=2200.2
>>> !put,molden,met_cas_6311++g_gs.molden
>>>
>>>   CPMCSCF,GRAD,1.1,spin=0.5,accu=1.0d-7,record=5101.1
>>>   CPMCSCF,GRAD,1.2,spin=0.5,accu=1.0d-7,record=5102.1}
>>>   put,molden,met_cas_6311++g.molden
>>>
>>>   {Force
>>>   SAMC,5101.1
>>>   CONICAL,6100.1}
>>>
>>>   {Force
>>>   SAMC,5102.1
>>>   CONICAL,6100.1}
>>>
>>>  {optg,startcmd=casscf}
>>>  put,molden,met_cas_6311++g_con.molden
>>>  {frequencies}
>>>
>>>
>>> -- 
>>> Jayashree
>>> _______________________________________________
>>> Molpro-user mailing list
>>> Molpro-user at molpro.net
>>> http://www.molpro.net/mailman/listinfo/molpro-user
>>
>>
>>
>>
>> -- 
>> Jayashree
>
>
>
>
> -- 
> Jayashree
> <met_cas_6311++g.inp><met_cas_6311++g.out>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20081004/e7ecf3df/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: freq.inp
Type: application/octet-stream
Size: 1803 bytes
Desc: not available
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20081004/e7ecf3df/attachment.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20081004/e7ecf3df/attachment-0001.html>


More information about the Molpro-user mailing list