Molcas Forum

Support and discussions for Molcas and OpenMolcas users and developers

You are not logged in.

Announcement

Welcome to the Molcas forum.

Please note: The forum's URL has changed. The new URL is: https://molcasforum.univie.ac.at. Please update your bookmarks!

You can choose an avatar and change the default style by going to "Profile" → "Personality" or "Display".

Pages: 1

#1 2018-10-31 09:31:53

Fancy
Member
From: Beijng.China
Registered: 2016-07-09
Posts: 15

opt an excited state

Hi, everyone!
Recently in my work I optmized the 4th excited state, whoes conf is 15240, the input file of opt part is as below:


***
***
End of basis
>>> If ( Iter = 1 ) <<<
&SCF
Title
T5 minimum optimization
>>> EndIf <<<
&RASSCF &END
TITLE
 use scforb
 Spin
  3
 Nactel
 10 0 0
 Ras2
 10
 Inactive
 86
 Ciroot
 4 4 1
LumOrb
 Thrs
 1.0e-05,1.0e-1,1.0e-1
 Rlxroot
 4
 &SLAPAF
 CARTesian
 MaxStep
 0.1
 Thrs
 1.0D-04 1.0D-03
 >>>EndDo
 End of Input

however, it jumped to root 3 after smoothly optimized several steps, the out file is as follwoing:

      printout of CI-coefficients larger than  0.05 for root  3
      energy=   -2456.710017
      conf/sym  1111111111     Coeff  Weight
         11676  2200u02u20  -0.07378 0.00544
         15141  2u00202u20  -0.49976 0.24976
         15144  2u0020u220  -0.45864 0.21035
         15240  u200202u20  -0.64677 0.41831
         15243  u20020u220   0.08978 0.00806
         23516  2u00u02220   0.06472 0.00419
         23616  u200u02220   0.28413 0.08073

      printout of CI-coefficients larger than  0.05 for root  4
      energy=   -2456.708818
      conf/sym  1111111111     Coeff  Weight
         15141  2u00202u20  -0.65498 0.42900
         15144  2u0020u220  -0.09551 0.00912
         15240  u200202u20   0.50264 0.25264
         15243  u20020u220  -0.44995 0.20245
         23516  2u00u02220  -0.28442 0.08090
         23616  u200u02220   0.05367 0.00288

and then I continue following the tracks of conf 15240, change the  "Rlxroot 4" as "Rlxroot 3" in input file at the last optmized geometry, but after several optimization, the state I optmized come back to 4th excited sate again! (The out  file as following). Beside, no matter what the maxstep I set, this situation repeat again and again, I guess the the 3th and 4th state may occur state filp or something other mix.  

      printout of CI-coefficients larger than  0.05 for root  3
      energy=   -2456.710139
      conf/sym  1111111111     Coeff  Weight
         11676  2200u02u20   0.07298 0.00533
         15141  2u00202u20   0.64907 0.42129
         15144  2u0020u220   0.42722 0.18252
         15240  u200202u20   0.49915 0.24915
         15243  u20020u220  -0.18591 0.03456
         23516  2u00u02220  -0.12555 0.01576
         23616  u200u02220  -0.26188 0.06858

      printout of CI-coefficients larger than  0.05 for root  4
      energy=   -2456.708751
      conf/sym  1111111111     Coeff  Weight
         15141  2u00202u20   0.50487 0.25490
         15144  2u0020u220   0.19239 0.03701
         15240  u200202u20  -0.65221 0.42537
         15243  u20020u220   0.41963 0.17609
         23516  2u00u02220   0.26596 0.07074
         23616  u200u02220  -0.11283 0.01273

Could someone please tell me how to solve this problem?

Offline

#2 2018-10-31 16:17:00

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,085

Re: opt an excited state

First, just looking at "conf 15240" is not a good way to identify a state. Other than the configuration you need to look at the orbitals, which could have changed, and there is obviously mixing with other configurations which also play a role. I'd say in the two cases you show the roots don't really change if you look at the relative signs of the three main CSFs in each case (instead of just looking at which has the largest weight).

There's no sure way of dealing with root flipping, among other things because the states can mix in any way and there's no one-to-one correspondence of the states at one geometry and the next (especially near a conical intersection).

Anyway, I would suggest to use the "Track" keyword in SLAPAF, it will try to follow your initial state through (approximate) wavefunction overlaps, which take into account changes in both MO and CI coefficients. You could also locate a conical intersection between states 3 and 4, and explore the surfaces according to its topography.

Offline

#3 2018-11-01 04:56:31

Fancy
Member
From: Beijng.China
Registered: 2016-07-09
Posts: 15

Re: opt an excited state

Thank you for your kind ang immediately reply, I'll take your advice and resubmit the work again. Actually, I always checked the orbital after every step, and the "conf 15240" is really the state I need. What I wonder is that after I added “Track” in &SLAPFA part, the molcas will always trace and optimized the state that I set at the begaining, right?

Offline

#4 2018-11-01 09:15:02

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,085

Re: opt an excited state

I will try, by picking at each iteration the state that's "most similar" to the state of the previous iteration. But if the changes are gradual enough, it may end up in a completely different state.

Offline

#5 2018-11-01 11:28:14

Fancy
Member
From: Beijng.China
Registered: 2016-07-09
Posts: 15

Re: opt an excited state

Unfortunately, after 1 iteracion, the work I have added "Track" in &SLAPFA part reported error, and it said that the problem is "TRACK"

input file:

>>> If ( Iter = 1 ) <<<
&SCF
Title
T5 minimum optimization
>>> EndIf <<<
&RASSCF &END
TITLE
 use scforb
 Spin
  3
 Nactel
 10 0 0
 Ras2
 10
 Inactive
 86
 Ciroot
 4 4 1
LumOrb
 Thrs
 1.0e-05,1.0e-1,1.0e-1
 Rlxroot
 4
 &SLAPAF
 Track
 MaxStep
 1.0
 Thrs
 1.0D-04 1.0D-03
 >>>EndDo
 End of Input

out file:

--- Stop Module:  rasscf at Thu Nov  1 18:17:10 2018 /rc=0 ---
--- Module rasscf spent 1 hour and 27 minutes
--- Start Module: slapaf at Thu Nov  1 18:17:11 2018
 ###############################################################################
 ###############################################################################
 ###                                                                         ###
 ###                                                                         ###
 ###    Error in RdCtl_Slapaf                                                ###
 ###                                                                         ###
 ###                                                                         ###
 ###############################################################################
 ###############################################################################

  *********** ERROR ***********
  The program has been supplied
  with an unknown command.
  *****************************
TRACK                                                                                                                                                        
--- Stop Module:  slapaf at Thu Nov  1 18:17:11 2018 /rc= _INPUT_ERROR_ ---
Non-zero return code - check program input/output
--- Stop Module:  auto at Thu Nov  1 18:17:12 2018 /rc= _INPUT_ERROR_ ---
--- Module auto spent 7 hours and 53 minutes
--- Stop Module:  auto at Thu Nov  1 18:17:12 2018 /rc= _INPUT_ERROR_ ---
--- Module auto spent 7 hours and 53 minutes

So, could you please tell me did I do something wrong or this is another problem in this work?

Offline

#6 2018-11-01 13:12:08

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,085

Re: opt an excited state

See here: https://molcasforum.univie.ac.at/viewto … 1060#p1060

Offline

#7 2018-11-03 04:11:49

Fancy
Member
From: Beijng.China
Registered: 2016-07-09
Posts: 15

Re: opt an excited state

Thank you for your patient reply!
I'm sorry, it may have some diffculty to get OpenMolcas in my lab recently, I wonder is there any other method to solve this state filpping?

Offline

Pages: 1

Links: OpenMolcas | Molcas | Manual | Facebook

Board footer

Powered by FluxBB 1.5.11

Last refresh: Today 19:06:33