Molcas Forum

sourav ghoshal

Member

Registered: 2017-09-21

Posts: 25

Is there any problem in SLAPAF_TS module?

I am using MOLCAS_UU binary. I am trying to locate the transition state of H abstraction reaction between acid and radical. I have chosen an active space 9 electron and 9 orbitals to locate TS.

Initially, I have got success in locating TS using cc-pVDZ basis set. Now, I am facing the problem in aug-cc-pVDZ basis by using the same orbital of cc-pVDZ basis set (EXPBASIS).

Although, I have able to locate the TS using contained bond length (constant 1 or more bonds). But, when I am trying to relocate the partially optimize TS in $SLAPAF TS module without not bond constant the TS immediate fall into the stable minimal. I thing my input has some error. Please rectify me if I have made any mistake.

Any assistance from any other will be highly appreciate !!! 

************** Input (For Constrained Optimization)**************

&GATEWAY
coord
9
Geom section
basis=cc-pVDZ
group=NoSymm
Constraint
R1 = Bond O2 O1
R2 = Bond O1 O5
Value
R1 = 2.16 Angstrom
R2 = 2.41 Angstrom
End of Constraint

>>> EXPORT MOLCAS_MAXITER=500
>>> DO WHILE
&SEWARD
&RASSCF
fileorb=$CurrDir/3so.RasOrb
Spin=2
Nactel=9 0 0
Inactive=25
Ras2=9

&Slapaf
TS

>>> END DO

&McKinley
Perturbation
Hessian

************* My Input (Without any constant) *************

&GATEWAY
partially optimized geom
basis=cc-pVDZ
group=NoSymm

&SEWARD
&RASSCF
fileorb=$CurrDir/fullopt3so.RasOrb
Spin=2
Nactel=9 0 0
Inactive=25
Ras2=9

&McKinley
Perturbation
Hessian
NoMCLR

&MCLR
Iterations = 500

Last edited by sourav ghoshal (2018-08-30 07:36:55)

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Is there any problem in SLAPAF_TS module?

You may have more success using FindTS instead of TS

sourav ghoshal

Member

Registered: 2017-09-21

Posts: 25

Re: Is there any problem in SLAPAF_TS module?

Yes, I am trying to do that.

But, I have noticed in several times that after few Iterations, especially, when Hessian Index appear as 1, than, the constant parameters disappear and normal optimization start with no constrained by showing a warning that "FindTS specified with no soft constratint".

How to specify the soft constraints or others, so that the partial optimization run smoothly to maintain constant bond parameters what I have specified in my input.

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Is there any problem in SLAPAF_TS module?

But, I have noticed in several times that after few Iterations, especially, when Hessian Index appear as 1, than, the constant parameters disappear and normal optimization start with no constrained by showing a warning that "FindTS specified with no soft constratint".

That's exactly how FindTS is supposed to work: start a constrained optimization and release (i.e. remove) the constraints when a TS appears to be nearby. If it succeeds, at the end you have a true, unconstrained TS.

For FindTS to work, it needs to know which constraints are real (hard) and must be satisfied no matter what, and which are just a guide (soft) intended for approaching the TS and will be released at the end. This can be done by specifying "Constraints" (hard) in GATEWAY and "TSConstraints" (soft) in SLAPAF, or by specifying only "Constraints" in GATEWAY but adding the word "soft" in the definition of the constraints intended to be soft. You probably did neither of those, by default all constraints are hard, which makes no sense for FindTS, so it assumed they were all soft instead and gave you a warning.

How to specify the soft constraints or others, so that the partial optimization run smoothly to maintain constant bond parameters what I have specified in my input.

Check the documentation for FindTS and TSConstraints http://molcas.gitlab.io/doc/sphinx/user … lapaf.html