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 2015-12-02 17:51:05

Sreenithya
Member
Registered: 2015-12-02
Posts: 8

Rasscf energy diverges

Hi All,

I am new to Molcas.
When I attempt to do a RASSCF calculation on a triplet diradical, the following error comes:

Large orbital rotation.
Molecular orbital   4 of symmetry 1 MO space 1  weight is    0.427872
Molecular orbital  30 of symmetry 1 MO space 1  weight is    0.268001
Molecular orbital  31 of symmetry 1 MO space 1  weight is    0.257448
.
.
.
Molecular orbital 104 of symmetry 1 MO space 2  weight is    0.220323
Molecular orbital 106 of symmetry 1 MO space 2  weight is    0.382402

Rasscf energy diverges.
!!! The program was forced to stop !!!

Could you please help me understanding and solving this?

Rasscf input file is as follows:

>export MOLCAS_MOLDEN=ON
&RASSCF
Spin
3
Inactive
100
Nactel
10 1 1
Ras1
1
Ras2
8
Ras3
1
Alter
5
1 59 101
1 87 102
1 95 103
1 99 104
1 159 107

Thanks in advance,
Sreenithya

Offline

#2 2015-12-02 19:26:57

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,079

Re: Rasscf energy diverges

That seems to indicate your selected active space is not stable in your calculation. Try with CASSCF (Ras1, Ras3 = 0, Ras2 = 10). Try with a state average of several roots (CIROot keyword).

If you need more help, you'll have to show us how you create your starting orbitals.

Offline

#3 2015-12-02 19:43:33

Sreenithya
Member
Registered: 2015-12-02
Posts: 8

Re: Rasscf energy diverges

Thank you so much for your response. Let me try with Ras2=10.

I have generated MOs using guessorb from a previously optimized DFT geometry.
Could you please tell me what information should I share regarding starting orbitals?

For singlet diradicals, isn't necessary to mention Ras1=1 and Ras3=1?

Last edited by Sreenithya (2015-12-02 19:55:28)

Offline

#4 2015-12-02 21:01:02

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,079

Re: Rasscf energy diverges

Sreenithya wrote:

I have generated MOs using guessorb from a previously optimized DFT geometry.
Could you please tell me what information should I share regarding starting orbitals?

Well, enough information for anyone to reproduce the calculation: starting geometry, basis set, sequence of inputs...

For singlet diradicals, isn't necessary to mention Ras1=1 and Ras3=1?

That's unrelated to the diradical/closed shell nature of the result. Ras1 and Ras3 are only restrictions to the configurations used in MCSCF procedure, and if you only compute the ground state, you will still very likely get a closed shell system, I'd say.

Offline

#5 2015-12-03 04:45:24

Sreenithya
Member
Registered: 2015-12-02
Posts: 8

Re: Rasscf energy diverges

The input geometry is:

 58
Angstrom
 C                 -3.41090300   -0.47937300    1.54412500
 O                 -3.80729900   -0.63858300    2.69987600
 C                 -2.12999700   -0.92063700    1.10160700
 H                 -1.52555800   -1.55112700    1.73376500
 C                 -1.75616400   -0.68554200   -0.30924000
 O                 -2.72311600   -0.84745800   -1.21476800
 C                 -4.12786700   -0.70530100   -0.84376800
 C                 -4.24906700    0.11418200    0.44623300
 H                 -3.90069800    1.13025800    0.25580100
 H                 -5.28853400    0.16321600    0.75223200
 C                 -0.41416300   -0.58612100   -0.86514000
 O                 -0.20439900   -0.95892500   -2.01817600
 N                  0.60311200   -0.12997300   -0.06597900
 C                  1.94095900   -0.20851500   -0.57994500
 C                  2.85813500   -1.19350800   -0.18532800
 C                  2.31174500    0.78884200   -1.48529000
 C                  4.15079400   -1.08010300   -0.72310700
 C                  3.58703300    0.86063700   -2.00753100
 H                  1.56881800    1.51140600   -1.77728100
 C                  4.52052700   -0.08584100   -1.60854100
 H                  4.89816800   -1.80066600   -0.43900900
 H                  3.84804700    1.63992200   -2.70582400
 H                  5.53036800   -0.05542400   -1.98691600
 C                  0.41494600    0.74359700    1.09224500
 H                  1.17328700    0.48670400    1.82423400
 H                 -0.54368000    0.52345500    1.55052500
 C                  0.48604600    2.22617800    0.79905700
 C                  1.49357700    3.00238300    1.36100700
 C                 -0.46272200    2.84567400   -0.01307000
 C                  1.55972100    4.36969700    1.11553400
 H                  2.23432900    2.53569500    1.99261000
 C                 -0.39910200    4.20815800   -0.26342100
 H                 -1.25396000    2.25984600   -0.45536400
 C                  0.61468900    4.97564000    0.30054000
 H                  2.34984700    4.95585800    1.55814500
 H                 -1.14063200    4.67233600   -0.89478700
 H                  0.66378000    6.03554700    0.10626000
 C                  2.62700100   -2.38899900    0.76260900
 C                  3.52882600   -2.22593500    1.99706900
 H                  3.27520900   -1.32335900    2.55077700
 H                  3.39790500   -3.07577800    2.66537200
 H                  4.58174500   -2.17261600    1.73278400
 C                  1.19626200   -2.58793300    1.26755400
 H                  0.49271900   -2.75141500    0.45685300
 H                  1.18103600   -3.47824400    1.89465000
 H                  0.84635300   -1.76032100    1.87324900
 C                  3.01141100   -3.68633300    0.02931900
 H                  4.05445400   -3.70513000   -0.27254000
 H                  2.84504300   -4.53914700    0.68535300
 H                  2.40042000   -3.82445000   -0.86113700
 C                 -4.70385000   -2.10367600   -0.69389500
 H                 -5.76845500   -2.04418800   -0.48141200
 H                 -4.56733000   -2.66349200   -1.61539200
 H                 -4.22420100   -2.65399100    0.11202900
 C                 -4.76320500    0.02565000   -2.01023200
 H                 -4.64047000   -0.54434000   -2.92747300
 H                 -5.82673200    0.15659400   -1.82688600
 H                 -4.31234800    1.00544800   -2.14456300

The input for guessorb is :

&GATEWAY
COORD
1-2-DR.xyz           
BASIS
C.ANO-RCC...3s2p.,N.ANO-RCC...3s2p.,H.ANO-RCC...2s.,O.ANO-RCC...3s2p.
GROUP=NOSYM

&SEWARD
amfi
doug
SDIPolar
Cholesky
&GUESSORB
PRMO
3
PRPOpulation

I am trying to do an excited state calculation.
In this input geometry, radical centers are C3 and C5 which are generated from a previous C3-C5 double bond.
In the next step, C3 radical abstracts H26.

Hence in my active space (10,7), I tried to include: C24-H26 sigma, C24-H26 sigma*, C3-C5 pi, C3-C5 pi*, lone pairs of O12, O2, and N13.

Offline

#6 2015-12-03 15:59:46

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,079

Re: Rasscf energy diverges

Sreenithya wrote:

Hence in my active space (10,7), I tried to include: C24-H26 sigma, C24-H26 sigma*, C3-C5 pi, C3-C5 pi*, lone pairs of O12, O2, and N13.

But your calculation is not (10,7), it's (10,10): Ras1+Ras2+Ras3=10.

You probably want to include the pi orbitals in C1, O2 and C11 as well, which would make it a (12,10) active space. Also guess orbitals are often not the best start. Try this:

&GATEWAY
  COORD = 1-2-DR.xyz           
  BASIS = C.ANO-RCC...3s2p.,N.ANO-RCC...3s2p.,H.ANO-RCC...2s.,O.ANO-RCC...3s2p.
  GROUP = NOSYM

&SEWARD
  Cholesky
  amfi
  doug
  SDIPolar

&SCF

&LOCALISATION
  File = $Project.ScfOrb
  Occupied

&LOCALISATION
  File = $Project.LocOrb
  Virtual

>copy $Project.LocOrb INPORB
>export MOLCAS_MOLDEN=ON
&RASSCF
  Spin = 3
  Inactive = 99
  NActEl = 12 0 0
  Ras2 = 10
  Alter = 8
    1 45 101
    1 61 102
    1 80 103
    1 86 105
    1 123 106
    1 134 107
    1 145 108
    1 152 109

At least it converges, but you'll see that the active space at the end is not correct. You may need to include the in-plane lone pairs in O6, O12 and/or perform a state averaging (CIROot keyword).

Offline

#7 2015-12-03 19:13:23

Sreenithya
Member
Registered: 2015-12-02
Posts: 8

Re: Rasscf energy diverges

Ignacio, this is awesome...!!!
I will try this.

Could you please tell me how do we decide the CiRoot and it's weightage?
Is it possible to control the radical nature using CiRoot?

Offline

#8 2015-12-03 19:26:50

Ignacio
Administrator
From: Uppsala
Registered: 2015-11-03
Posts: 1,079

Re: Rasscf energy diverges

Sreenithya wrote:

Could you please tell me how do we decide the CiRoot and it's weightage?

There's no sure recipe. It depends on the system and process you are studying. You should include a number of roots that's enough to (1) include the state you are interested in, (2) produce a stable active space, and (3) give a balanced description of all the states of interest. In general, to get a stable active space with lone-pair orbitals, you'd have to include at least as many states as lone pairs... but's that's just a guideline.

The weight is easier: equal weights for all roots, unless you know what you are doing (and even then). That is, if you want N states, write "CIROot = N N 1" (this is a shortcut)

Is it possible to control the radical nature using CiRoot?

No, this simply selects a number of states according to their energy.

Offline

#9 2015-12-16 18:52:07

Sreenithya
Member
Registered: 2015-12-02
Posts: 8

Re: Rasscf energy diverges

I have tried the calculation after including the C-O pi orbitals.
But when I check the orbitals after rasscf calulation, C-H and C-H* orbitals are not there in the active space.

Offline

Pages: 1

Links: OpenMolcas | Molcas | Manual | Facebook

Board footer

Powered by FluxBB 1.5.11

Last refresh: Today 22:12:04