Molcas Forum

lwang

Member

Registered: 2024-10-14

Posts: 7

Swap track error using the keyword ALTEr

I calculated the guanine anionic caspt2 energy of the PCM water solvent with MOLCAS, the electronic states of the first root and the second root were flipped, and then I did the 40 and 41 orbital swap in the second step RASSCF, I wanted to correct it, but it didn't work successfully, what is the reason for this?
The input files are as follows：

&GATEWAY
coord =g.XYZ
 basis =Aug-cc-pvdz
 group = C1
 RF-input
 PCM-model; solvent= water
 End of RF-input
 RICD
>>> EXPORT MOLCAS_MAXITER = 500
&SEWARD
&RASSCF
   Charge =-1
   spin=2
   nActEl=13 0 0
   Ras2=12
   ciroot =1 1 1
   ALTEr= 2; 1 40 42; 1 41 42
&CASPT2
  Maxit =500
  Lroot = 1
  RFPert
&RASSCF
   Charge =-1
   spin=2
   nActEl=13 0 0
   Ras2=12
   ciroot =8 8 1
   ALTEr= 1; 1 40 41
   NONEquilibrium
&CASPT2
   Maxit =500
   Multistate=4 1 2 3 4
   RFPert

The configuration of the rasscf output is as follows：

 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=    -539.612740
      conf/sym  111111111111     Coeff  Weight
             1  222222u00000  -0.08358 0.00699
             2  2222220u0000   0.91643 0.83984
             3  22222200u000   0.11245 0.01265
             9  22222uud0000   0.12186 0.01485
            10  22222u020000  -0.06011 0.00361
           161  2u22220d0u00   0.06012 0.00361
           223  2222202u0000  -0.09332 0.00871
           230  2222200u2000   0.05804 0.00337
           298  2222uduud000   0.08097 0.00656
           650  2222020u2000   0.05261 0.00277

      printout of CI-coefficients larger than  0.05 for root  2
      energy=    -539.599858
      conf/sym  111111111111     Coeff  Weight
             1  222222u00000   0.89516 0.80131
             2  2222220u0000   0.10844 0.01176
             3  22222200u000  -0.18123 0.03285
             4  222222000u00  -0.05481 0.00300
            10  22222u020000   0.11293 0.01275
            11  22222ud0u000  -0.12899 0.01664
            17  22222u0d0u00   0.05046 0.00255
            43  2222u2200000   0.05556 0.00309
           156  2u22220du000  -0.07285 0.00531
           224  222220u20000   0.08122 0.00660
           229  222220u02000   0.07336 0.00538
           295  2222ud20u000  -0.06748 0.00455

      printout of CI-coefficients larger than  0.05 for root  3
      energy=    -539.536336
      conf/sym  111111111111     Coeff  Weight
             1  222222u00000   0.14663 0.02150
             2  2222220u0000  -0.07780 0.00605
             3  22222200u000   0.85703 0.73451
             7  22222u200000  -0.25442 0.06473
             8  22222udu0000   0.05572 0.00311
             9  22222uud0000  -0.05629 0.00317
            10  22222u020000   0.06108 0.00373
            11  22222ud0u000  -0.10111 0.01022
            15  22222u002000   0.07962 0.00634
            17  22222u0d0u00  -0.06429 0.00413
            19  22222uu00d00   0.06185 0.00383
            20  22222u0u0d00  -0.05563 0.00310
            43  2222u2200000   0.06158 0.00379
            46  2222u2020000  -0.06705 0.00450
           153  2u2222ud0000  -0.10474 0.01097
           154  2u2222020000  -0.05015 0.00251
           225  22222020u000  -0.08112 0.00658
           227  22222002u000  -0.07976 0.00636
           299  2222udu02000   0.06241 0.00390

Use the MULTIWFN program to view the molecular orbitals and their occupancy numbers below：
Orb:    39 Ene(au/eV):    -0.000400      -0.0109 Occ: 1.495010 Type:A+B (A   )
Orb:    40 Ene(au/eV):    -0.000900      -0.0245 Occ: 0.604830 Type:A+B (A   )
Orb:    41 Ene(au/eV):    -0.345800      -9.4097 Occ: 0.670920 Type:A+B (A   )
Orb:    42 Ene(au/eV):    -0.002800      -0.0762 Occ: 0.313780 Type:A+B (A   )

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,097

Re: Swap track error using the keyword ALTEr

Are 40 and 41 active orbitals? Then it makes no sense swapping them, the order of active orbitals is completely irrelevant (at least within the same RAS space).

lwang

Member

Registered: 2024-10-14

Posts: 7

Re: Swap track error using the keyword ALTEr

The two orbitals are in the activation space, because the electronic states are flipped, I want to make them exchange the two roots to form the correct configuration, is there any way to solve it?

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,097

Re: Swap track error using the keyword ALTEr

Why do you say they're flipped? Because "2222220u0000" is lower than "222222u00000". As I said, the order is irrelevant, the program will order the orbitals by average occupation or something, but to analyze the states you should look at the shape of the orbitals, not their order.

lwang

Member

Registered: 2024-10-14

Posts: 7

Re: Swap track error using the keyword ALTEr

After guanine is attached, the first root (i.e., the anionic ground state) should be "222222u00000", but my first root is "2222220u0000". My molecular orbital is the correct Π* orbital, but the order of the electron configuration is reversed because I want to calculate the value of VEA, is the calculated caspt2 energy correct? Is there a reasonable explanation for this phenomenon?

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,097

Re: Swap track error using the keyword ALTEr

Which one is lowest depends probably on the basis set, computational level (PT2 or not), environment, etc. Your results say that, at CASSCF level (and state-averaged over 8 states, etc., I presume those values are from RASSCF), "2222220u0000" is lower, not by much. Changing the order of the orbitals will change nothing, what matters is the actual shape of those orbitals. The CASPT2 calculation may change the order of the states.