From owner-chemistry ^at^ ccl.net Mon Aug 23 11:11:01 2021 From: "Muhammed mbtemiz3_-_gmail.com" To: CCL Subject: CCL: Can CASSCF energy be a descriptor for the stability of active spac Message-Id: <-54446-210823025905-9198-U42SOAEEE9EkqnK7zUBEyA##server.ccl.net> X-Original-From: Muhammed Content-Type: multipart/alternative; boundary="000000000000d2d1d705ca348c6e" Date: Mon, 23 Aug 2021 09:58:43 +0300 MIME-Version: 1.0 Sent to CCL by: Muhammed [mbtemiz3++gmail.com] --000000000000d2d1d705ca348c6e Content-Type: text/plain; charset="UTF-8" Dear Nuno, Thank you for the answer. You are right, having a balanced AS is not an easy thing. And a matching AS is needed for all spin states. However my problem is a bit more involved than that. I have a matching AS for all spin states, sure, but there is an additional solution - which does not include the orbitals I am after - with a lower energy for one of the spin states. What I am wondering is, does the lower energy solution tell me anything about the stability of the wavefunction. To simply put, does the lower energy - wrong AS for this particular problem - solution describe the system better and should I always aim for lower energy solutions? Consider this example: There are two solutions for a carbon-carbon (C1-C2) single bond for some same spin state; one with 2pz orbital of C1 AND 2pz orbital of C2, the another solution has 2pz and 2s orbitals of the C1 only. For the sake of the argument, let's assume the second solution is lower in energy than the former. Which one should I aim for? The second one clearly does not describe the system well if you are after the carbon-carbon interactions. Also, is it even justified to compare the energies of these two solutions? Thank you, Best regards. Muhammed. On Sat, Aug 21, 2021 at 7:00 PM Nuno A. G. Bandeira nuno.bandeira-x-tecnico.ulisboa.pt wrote: > > Sent to CCL by: "Nuno A. G. Bandeira" [nuno.bandeira++tecnico.ulisboa.pt] > [Resent to CCL by the Admin] > Having a balanced active space is not an easy thing to achieve. I find > that often its a tedious trial and error procedure to get to know how the > algorithm behaves. A robust second order algorithm will usually get to the > best solution but it always depends on the number of roots you ask for. > > > > First: is CASSCF the end game? Will you perturb the wavefunction > afterwards? If so then it wont matter so much to have a large AS. > > > > For different spin state it might be more important to correlate other > orbitals than the ones youre interested in. Theres nothing you can do about > that, but you can try to add more orbitals into the active space until you > have a matching AS across all spin states. > > > > I would start by looking at the MP2 natural orbitals at the frontier > region to see which orbitals of your ligand are more relevant. You might > not need that many but you need a matching set for every spin state. You > may not even need all the 3ds it just depends on the electronic structure > of the complex itself. > > > > Hope this helps. > > All the best, > > Nuno > > > > ----------- > > Nuno A. G. Bandeira, AMRSC > > Email: nuno.bandeira|-|ciencias.ulisboa.pt > > > > BioISI - Biosystems & Integrative Sciences Institute; > > 8.5.53 - C8, Faculty of Sciences, University of Lisbon > > Campo Grande- 1749-016 Lisboa Portugal > > > > Phone: +351 21 750 08 45 > > -------- > > > > https://www.researchgate.net/profile/Nuno_Bandeira > > http://www.researcherid.com/rid/B-6399-2012 > > http://orcid.org/0000-0002-5754-7328 > > http://www.cienciavitae.pt/pt/4A19-ACC3-E225 > > http://pt.linkedin.com/pub/nuno-a-g-bandeira/47/55a/2aa > > > > > > > From: Muhammed mbtemiz3:-:gmail.com > Sent: 20 August 2021 18:05 > To: Bandeira, Nuno A. G. > Subject: CCL: Can CASSCF energy be a descriptor for the stability of > active space? > > > > Dear CCLers, > > > > I am trying to construct an active space (AS) for a Co=O transition metal > complex. I am interested in the metal d orbitals, oxygen 2p orbitals and > one specific ligand orbital. I have managed to get the orbitals I desire > for three spin states (doublet, quartet, sextet) after a couple of orbital > rotations and got the energies with a relatively good basis set. (This is > the first AS) > > However while doing the orbital rotations, I found a lower energy solution > for the doublet state with the active space of 5 d orbitals, 3 Oxygen 2p > orbitals, and 1 Oxygen 2s orbital - ligand orbital is swapped with Oxygen > 2s orbital (This is the second AS). This AS is no good since I am also > interested in the ligand itself, however the energies made me question the > stability of the first active space. > > Now the problem is this: doublet has two solutions; first AS and second > AS. Second AS is around 30 kcal/mol lower in energy than the first AS, > however it does not include the orbital I am interested in. > > My question is this, do these results tell anything about the nature of > the active space? > Should I lean towards the second AS since it has a lower energy solution > than the first one? > > My reasoning is this; > 1- The lower energy active space tells me that the initial AS is not very > well balanced and additional orbitals might be included in the AS. However > this approach didn't solve my problem - got similar results even with the > highest electron/orbital pair I technically can run. (including double > shell orbitals are also not rewarding). And without the ligand orbitals, > this approach hit a wall. > > 2- These solutions live in different spaces and we can't compare them. > This is similar to constructing two different active spaces with different > electron/orbital pairs and trying to compare their energies - i.e comparing > CAS(10,10) energy with CAS(12,12). Although both solutions have equal > number of electrons/orbitals in this case, the nature of the wavefunction > is different for both. First solution includes configurations with > excitations from the ligand orbitals, however the second one considers the > excitations from oxygen 2s orbital. Therefore I can continue to work with > the first active space without worrying about the second AS solution. > > Although I am not sure about the last point, I would appreciate if someone > guides me to the right direction here. > > Thank you. > Muhammed. > > *I am aware of the double shell effect but ignored it here for the sake of > simplicity. > > > > -- > > Muhammed Buyuktemiz > > Chemistry Department, Gazi University> > > -- Muhammed Buyuktemiz Chemistry Department, Gazi University +90 554 844 11 25 --000000000000d2d1d705ca348c6e Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Dear Nuno,

Thank you for the answer.=C2= =A0

You are right, having a balanced AS is not an = easy thing. And a matching AS is needed for all spin states. However my pro= blem is a bit more involved than=C2=A0that. I have a matching AS for all sp= in states, sure, but there is an additional solution - which does not inclu= de the orbitals I am after - with a lower energy for one of the spin states= . What I am wondering is, does the lower energy solution tell me anything a= bout the stability of the wavefunction.=C2=A0

To s= imply put, does the lower energy - wrong AS for this particular problem - s= olution describe the system better and should I always aim for lower energy= solutions?=C2=A0

Consider this example:=C2=A0
There are two solutions for a carbon-carbon (C1-C2) single bond for = some same spin state; one with 2pz orbital of C1 AND 2pz orbital of C2, the= another solution has 2pz and 2s orbitals of the C1 only. For the sake of t= he argument, let's assume the second solution is lower in energy than t= he former. Which one should I aim for? The=C2=A0second one clearly does not= =C2=A0describe the system well if you are after the carbon-carbon interacti= ons. Also, is it even justified to compare the energies of these two soluti= ons?

Thank you,
Best=C2=A0regards.=C2=A0=

Muhammed.

On Sat, Aug 21, 2021 at 7:00 PM Nu= no A. G. Bandeira nun= o.bandeira-x-tecnico.ulisboa.pt <owner-chemistry[A]ccl.net> wrote:

Sent to CCL by: "Nuno A. G. Bandeira" [nuno.bandeira++tecnico.uli= sboa.pt]
[Resent to CCL by the Admin]
Having a balanced active space is not an easy thing to achieve. I find that= often its a tedious trial and error procedure to get to know how the algor= ithm behaves. A robust second order algorithm will usually get to the best = solution but it always depends on the number of roots you ask for.



First: is CASSCF the end game? Will you perturb the wavefunction afterwards= ? If so then it wont matter so much to have a large AS.



For different spin state it might be more important to correlate other orbi= tals than the ones youre interested in. Theres nothing you can do about tha= t, but you can try to add more orbitals into the active space until you hav= e a matching AS across all spin states.



I would start by looking at the MP2 natural orbitals at the frontier region= to see which orbitals of your ligand are more relevant. You might not need= that many but you need a matching set for every spin state. You may not ev= en need all the 3ds it just depends on the electronic structure of the comp= lex itself.



Hope this helps.

All the best,

Nuno



-----------

Nuno A. G. Bandeira, AMRSC

Email: nuno.bandeira|-|ciencias.ulisboa.pt



BioISI - Biosystems & Integrative Sciences Institute;

8.5.53 - C8, Faculty of Sciences, University of Lisbon

Campo Grande- 1749-016 Lisboa Portugal



Phone: +351 21 750 08 45

--------



https://www.researchgate.net/profile/Nuno_Bandeira=

http://www.researcherid.com/rid/B-6399-2012

http://orcid.org/0000-0002-5754-7328

http://www.cienciavitae.pt/pt/4A19-ACC3-E225

http://pt.linkedin.com/pub/nuno-a-g-bandeira/= 47/55a/2aa





> From: Muhammed mbtemiz3:-:gmail.com
Sent: 20 August 2021 18:05
To: Bandeira, Nuno A. G.
Subject: CCL: Can CASSCF energy be a descriptor for the stability of active= space?



Dear CCLers,



I am trying to construct an active space (AS) for a Co=3DO transition metal= complex. I am interested in the metal d orbitals, oxygen 2p orbitals and o= ne specific ligand orbital. I have managed to get the orbitals I desire for= three spin states (doublet, quartet, sextet) after a couple of orbital rot= ations and got the energies with a relatively good basis set. (This is the = first AS)

However while doing the orbital rotations, I found a lower energy solution = for the doublet state with the active space of 5 d orbitals, 3 Oxygen 2p or= bitals, and 1 Oxygen 2s orbital - ligand orbital is swapped with Oxygen 2s = orbital (This is the second AS). This AS is no good since I am also interes= ted in the ligand itself, however the energies made me question the stabili= ty of the first active space.

Now the problem is this: doublet has two solutions; first AS and second AS.= Second AS is around 30 kcal/mol lower in energy than the first AS, however= it does not include the orbital I am interested in.

My question is this, do these results tell anything about the nature of the= active space?
Should I lean towards the second AS since it has a lower energy solution th= an the first one?

My reasoning is this;
1- The lower energy active space tells me that the initial AS is not very w= ell balanced and additional orbitals might be included in the AS. However t= his approach didn't solve my problem - got similar results even with th= e highest electron/orbital pair I technically can run. (including double sh= ell orbitals are also not rewarding). And without the ligand orbitals, this= approach hit a wall.

2- These solutions live in different spaces and we can't compare them. = This is similar to constructing two different active spaces with different = electron/orbital pairs and trying to compare their energies - i.e comparing= CAS(10,10) energy with CAS(12,12).=C2=A0 Although both solutions have equa= l number of electrons/orbitals in this case, the nature of the wavefunction= is different for both. First solution includes configurations with excitat= ions from the ligand orbitals, however the second one considers the excitat= ions from oxygen 2s orbital. Therefore I can continue to work with the firs= t active space without worrying about the second AS solution.

Although I am not sure about the last point, I would appreciate if someone = guides me to the right direction here.

Thank you.
Muhammed.

*I am aware of the double shell effect but ignored it here for the sake of = simplicity.



--

Muhammed Buyuktemiz

Chemistry Department, Gazi University



-=3D This is automatically added to each message by the mailing script =3D-=
E-mail to subscribers: CHEMISTRY[A]ccl.net or use:
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/cgi-bin/ccl/s= end_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST[A]ccl.net or use
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/cgi-bin/ccl/s= end_ccl_message
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/chemistry/sub_un= sub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.net/chemist= ry/announcements/conferences/

Search Messages: http://www.ccl.net/chemistry/sear= chccl/index.shtml
=C2=A0 =C2=A0 =C2=A0 http://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/ins= tructions/




--
Muhammed Buyukt= emiz
Chemistry Department,= =C2=A0Gazi University=C2=A0
+90 554 844 11 25=C2=A0
--000000000000d2d1d705ca348c6e--