Molcas Forum
Support and discussions for Molcas and OpenMolcas users and developers
You are not logged in.
- Topics: Active | Unanswered
Announcement
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".#1 2018-11-20 00:45:01
- Leo
- Member
- Registered: 2018-11-20
- Posts: 15
Configurations mixing in CASSCF S1 optimization
Dear all,
I am trying to optimize the S1 geometry of a molecule at CASSCF (12,12) level.
Here, I reported the wavefunction printout section on the first and last iteration (converged structure). It is evident how for S1 on the last iteration the closed shell character increased at the expense of the singly excited configuration. Also, by the occupation numbers, looks like on the last iteration the wave function is no more correctly describing any excitation.
First iteration:
************************************************************************************************************************
Wave function printout:
occupation of active orbitals, and spin coupling of open shells (u,d: Spin up or down)
************************************************************************************************************************
Note: transformation to natural orbitals
has been made, which may change the order of the CSFs.
printout of CI-coefficients larger than 0.05 for root 1
energy= -868.913199
conf/sym 111111111111 Coeff Weight
1 222222000000 0.86993 0.75677
8 2222u2d00000 -0.14849 0.02205
9 2222u20d0000 0.06586 0.00434
69 2222udu000d0 -0.05033 0.00253
164 222202200000 -0.09741 0.00949
179 222202u0000d -0.05846 0.00342
184 222202000002 -0.07326 0.00537
186 222ud2ud0000 -0.06086 0.00370
201 222ud20u000d -0.06133 0.00376
242 2u22d2u0000d -0.07249 0.00525
247 2u22d2000002 -0.06191 0.00383
327 2u2d220u000d -0.05090 0.00259
416 202222200000 -0.05988 0.00359
436 202222000002 -0.05473 0.00300
594 2u22u2d0000d 0.07263 0.00528
printout of CI-coefficients larger than 0.05 for root 2
energy= -868.802725
conf/sym 111111111111 Coeff Weight
1 222222000000 0.17062 0.02911
8 2222u2d00000 0.77229 0.59643
9 2222u20d0000 0.05447 0.00297
13 2222u200000d -0.25227 0.06364
26 2u2222d00000 0.11086 0.01229
27 2u22220d0000 -0.06025 0.00363
31 2u222200000d 0.10137 0.01028
69 2222udu000d0 0.05896 0.00348
164 222202200000 -0.19761 0.03905
165 222202ud0000 -0.06168 0.00380
179 222202u0000d 0.12130 0.01471
188 222ud2u0d000 0.05898 0.00348
200 222ud2u0000d -0.08408 0.00707
227 2u22d2200000 -0.11544 0.01333
228 2u22d2ud0000 0.05296 0.00280
233 2u22d2u00d00 -0.07308 0.00534
249 u222d2ud0000 0.05458 0.00298
258 u222d2u000d0 -0.06249 0.00391
594 2u22u2d0000d 0.05918 0.00350
Natural orbitals and occupation numbers for root 1
sym 1: 1.939987 1.899234 1.974538 1.926189 1.853056 1.949355 0.158258 0.069586 0.044979 0.054515
0.032486 0.097818
Natural orbitals and occupation numbers for root 2
sym 1: 1.930258 1.871137 1.980297 1.912239 1.099799 1.944932 0.897805 0.081420 0.068203 0.052344
0.036775 0.124790Last iteration:
************************************************************************************************************************
Wave function printout:
occupation of active orbitals, and spin coupling of open shells (u,d: Spin up or down)
************************************************************************************************************************
Note: transformation to natural orbitals
has been made, which may change the order of the CSFs.
printout of CI-coefficients larger than 0.05 for root 1
energy= -868.890074
conf/sym 111111111111 Coeff Weight
1 222222000000 -0.82328 0.67779
8 2222u2d00000 0.31601 0.09986
9 2222u20d0000 -0.07076 0.00501
62 2222udu0d000 0.05411 0.00293
164 222202200000 0.13167 0.01734
184 222202000002 0.06456 0.00417
186 222ud2ud0000 0.07229 0.00523
201 222ud20u000d 0.05239 0.00274
242 2u22d2u0000d 0.08955 0.00802
247 2u22d2000002 0.05101 0.00260
254 u222d2u00d00 -0.05165 0.00267
416 202222200000 0.06695 0.00448
436 202222000002 0.05049 0.00255
554 222uu2dd0000 -0.05991 0.00359
594 2u22u2d0000d -0.11275 0.01271
639 2u2u22d0000d 0.05050 0.00255
printout of CI-coefficients larger than 0.05 for root 2
energy= -868.826103
conf/sym 111111111111 Coeff Weight
1 222222000000 -0.35604 0.12676
8 2222u2d00000 -0.73226 0.53621
13 2222u200000d -0.05061 0.00256
26 2u2222d00000 0.17810 0.03172
31 2u222200000d -0.05507 0.00303
62 2222udu0d000 -0.05569 0.00310
69 2222udu000d0 -0.05775 0.00334
164 222202200000 -0.16872 0.02846
179 222202u0000d -0.13075 0.01710
186 222ud2ud0000 -0.09248 0.00855
191 222ud2u00d00 -0.05085 0.00259
200 222ud2u0000d 0.05107 0.00261
227 2u22d2200000 0.07414 0.00550
228 2u22d2ud0000 -0.06438 0.00414
242 2u22d2u0000d -0.09194 0.00845
254 u222d2u00d00 0.06016 0.00362
269 222022200000 0.05537 0.00307
416 202222200000 0.07224 0.00522
594 2u22u2d0000d 0.09466 0.00896
3639 2022u2d00002 0.06578 0.00433
Natural orbitals and occupation numbers for root 1
sym 1: 1.943251 1.890464 1.983245 1.927777 1.799614 1.953001 0.197686 0.073344 0.042670 0.054676
0.034755 0.099516
Natural orbitals and occupation numbers for root 2
sym 1: 1.928667 1.859521 1.975170 1.902761 1.646764 1.938589 0.362856 0.091695 0.045348 0.066811
0.053221 0.128596
************************************************************************************************************************I would really be grateful if you could help me understand what could possibly determine this "configurations mixing" which I never found in any of my models. Could that possibly be determined by the degeneracy of the excited states? Or might that be determined by intruder states? If those are the case, what should I try to solve this problem? I have already tried Track in SLAPAF but that doesn't do the trick.
Here is my input, the coordinates and basis set (6-31G*) are given explicitly:
> EXPORT MOLCAS_MOLDEN=ON
> COPY $InpDir/$Project.xyz $WorkDir
> COPY $InpDir/$Project.key $WorkDir
> COPY $InpDir/file.prm $WorkDir
> COPY $InpDir/$Project.Espf.Data $WorkDir
> COPY $InpDir/$Project.JobIph $WorkDir/$Project.JobOld
> EXPORT CONT=1
> export MOLCAS_MAXITER = 100
> Do while
&SEWARD
&ESPF
External
Tinker
lamorok
&RASSCF
spin
1
nActEl
12 0 0
Inactive
76
Ras2
12
JobIph
cirestart
ciroot
3 3 1
rlxroot
2
> COPY $WorkDir/$Project.JobIph $InpDir
> COPY $Project.JobIph $Project.JobOld
&ALASKA
&SLAPAF
cartesian
rHidden
10.0
> COPY $WorkDir/$Project.xyz $InpDir/$Project.Final.xyz
> COPY $WorkDir/$Project.Espf.Data $InpDir
> EXPORT EST=00$CONT
> If ( -FILE $Project.$EST )
> COPY $WorkDir/$Project.$EST $InpDir
> EndIf
> EXPORT EST=0$CONT
> If ( -FILE $Project.$EST )
> COPY $WorkDir/$Project.$EST $InpDir
> EndIf
> EVAL CONT=$CONT+1
> EndDoThank you in advance,
Leo
Last edited by Leo (2018-11-20 08:28:16)
Offline
#2 2018-11-20 09:19:10
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,085
Re: Configurations mixing in CASSCF S1 optimization
Going from 60% to 54% is not a huge change. Anyway, just a couple of things to keep in mind:
Electronic configurations, on their own, don't mean anything. They must always be associated with the particular set of orbitals. Change the orbitals and the configurations and coefficients change (without changing the wave function).
Some states can be reasonably well represented with a single configuration, some states require a few or many configurations. This happens when the configurations (not the states) are almost degenerate. There's nothing wrong or surprising in finding this kind of mixing with CASSCF.
Offline