From owner-chemistry@ccl.net Mon Sep 11 09:47:01 2017 From: "Henrique C. S. Junior henriquecsj:gmail.com" To: CCL Subject: CCL:G: Correctly evaluating spin states of a cobalt trimer (using Single Points)? Message-Id: <-52994-170910171432-1974-fDIgmMx8v+7DrvWylEWtug*o*server.ccl.net> X-Original-From: "Henrique C. S. Junior" Content-Type: multipart/alternative; boundary="94eb2c0ba0fc7c107a0558dc4b23" Date: Sun, 10 Sep 2017 18:13:45 -0300 MIME-Version: 1.0 Sent to CCL by: "Henrique C. S. Junior" [henriquecsj|-|gmail.com] --94eb2c0ba0fc7c107a0558dc4b23 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Dear Andrew, thank you for putting all this information together. The problem is really much more complex than a careless first look may suggest. Luckily I'm already using ORCA so I know that I have the right tool for the task. Thank you again. On Sun, Sep 10, 2017 at 11:51 AM, Andrew Rosen rosen###u.northwestern.edu < owner-chemistry.:.ccl.net> wrote: > Henrique, > > I am going to preface this by saying that this is a deceivingly > challenging task. I am a young graduate student who is very much still > learning, so if others reading this disagree with something I say, please > feel free to chime in. In general, you have the right idea, although ther= e > are a few things to consider. My response is in the context of DFT, which= I > imagine is what you're using. > > 1. I assume you are using the structure from XRD in your electronic > structure calculations. It is likely that at the level of theory you > choose, the XRD structure is not exactly the minimum energy structure > (although it is hopefully quite close!). In that case, it may be advisabl= e > to do a geometry optimization from this experimental geometry instead of = a > single point energy calculation at each spin multiplicity. This may also = be > beneficial, albeit more computationally tasking, because different spin > states can lead to different molecular geometries. Oftentimes, this > difference may be small, but it has the possibility of impacting the > energetics. The chapter on "Spin interactions in cluster chemistry" in th= e > text "Advances in Inorganic Chemistry Volume 62: Theoretical and > Computational Inorganic Chemistry" may be quite useful. A publicly > accessible link to the relevant section on Google Books is found here > . > Pages 216-222 are extremely relevant to this discussion. > > 2. You should check to see the degree of spin contamination in your > calculations, as discussed here > . If > it is large, it could signify that the level of theory you chose is not > sufficient for the problem at hand and the energetics as well as other > molecular properties may not be accurate. In such cases, it may be > necessary to consider more accurate multireference methods. > > 3. It can often be difficult to accurately capture the relative energies > of various spin states for transition metal complexes using DFT, and this > can often be very sensitive to the choice of density functional. As > discussed here , > pure functionals tend to favor low-spin states whereas hybrid functions > tend to favor high-spin states, and the energy difference between low- an= d > high-spin states is often directly related to the amount of Hartree-Fock > exchange in a given functional. This is a limitation to keep in mind. > > 4. There is the possibility that a broken-symmetry state is most stable. > While this is likely a bit more involved than what you are looking to che= ck > for (especially given this is a trimer), it is worth realizing that such = a > possibility exists, as discussed here in the > context of Gaussian or here > = in > the context of ORCA. > > 5. At times, it may be necessary to check the stability of the > wavefunction when dealing with open-shell structures. A detailed discussi= on > on StackExchange can be found here > > . > > In the end, yes, you are correct that you must do this "manually" and > compare the energetics of different possible spin multiplicities. Your > approach may end up being sufficient, but as I mentioned, there are some > factors that you should at the very least keep in mind. > > Andrew > > On Sun, Sep 10, 2017 at 8:20 AM Henrique C. S. Junior henriquecsj+/- > gmail.com wrote: > >> Dear colleagues, I=E2=80=99m working with a Cobalt(II) trimer whose mole= cular >> structure was achieved by Single Crystal X-Ray Diffraction. My task now = is >> to check the spin states of the structure (High or Low spin). Since Co(I= I) >> can have 1 or 3 unpaired electrons, I=E2=80=99m approaching this problem= by >> calculating Single Points for every possible multiplicity (10, 8, 6, 4, = 2) >> and assuming that the most stable is the one that represents my structur= e >> (and my spin states). >> >> Is this approach correct? >> >> Thank you >> >> -- >> *Henrique C. S. Junior* >> >> --=20 *Henrique C. S. Junior* Industrial Chemist - UFRRJ M. Sc. Inorganic Chemistry - UFRRJ Data Processing Center - PMP Visite o Mundo Qu=C3=ADmico --94eb2c0ba0fc7c107a0558dc4b23 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Dear Andrew, thank you for putting all this information togeth= er. The problem is really much more complex than a careless first look may = suggest. Luckily I'm already using ORCA so I know that I have the right= tool for the task.

Thank you again.

On Sun, Sep 10, 2017 at 11:51 AM, And= rew Rosen rosen###u.northwestern.edu<= /a> <owner-chemistry.:.ccl.net> wrote:
Henrique,=

I am going to preface this by saying that this is a deceivingly challen= ging task. I am a young graduate student who is very much still learning, s= o if others reading this disagree with something I say, please feel free to= chime in. In general, you have the right idea, although there are a few th= ings to consider. My response is in the context of DFT, which I imagine is = what you're using.

1. I assume you are using the structure from XRD i= n your electronic structure calculations. It is likely that at the level of= theory you choose, the XRD structure is not exactly the minimum energy str= ucture (although it is hopefully quite close!). In that case, it may be adv= isable to do a geometry optimization from this experimental geometry instea= d of a single point energy calculation at each spin multiplicity. This may = also be beneficial, albeit more computationally tasking, because different = spin states can lead to different molecular geometries. Oftentimes, this di= fference may be small, but it has the possibility of impacting the energeti= cs. The chapter on "Spin interactions in cluster chemistry" in th= e text "Advances in Inorganic Chemistry Volume 62: Theoretical and Com= putational Inorganic Chemistry" may be quite useful. A publicly access= ible link to the relevant section on Google Books is found=C2=A0here. Pages 216-222 are extremely relevant to this discussion.
<= div style=3D"font-size:small">
2. Y= ou should check to see the degree of spin contamination in your calculation= s, as discussed=C2=A0here. If it is large, it cou= ld signify that the level of theory you chose is not sufficient for the pro= blem at hand and the energetics as well as other molecular properties may n= ot be accurate. In such cases, it may be necessary to consider more accurat= e multireference methods.

3. It can often be difficult to accurately capt= ure the relative energies of various spin states for transition metal compl= exes using DFT, and this can often be very sensitive to the choice of densi= ty functional. As discussed=C2=A0here, pure functionals tend to= favor low-spin states whereas hybrid functions tend to favor high-spin sta= tes, and the energy difference between low- and high-spin states is often d= irectly related to the amount of Hartree-Fock exchange in a given functiona= l. This is a limitation to keep in mind.

4. There is the possibility that= a broken-symmetry state is most stable. While this is likely a bit more in= volved than what you are looking to check for (especially given this is a t= rimer), it is worth realizing that such a possibility exists, as discussed= =C2=A0here=C2=A0= in the context of Gaussian or=C2=A0here=C2=A0= in the context of ORCA.

5. At times, it may be necessary to check the sta= bility of the wavefunction when dealing with open-shell structures. A detai= led discussion on StackExchange can be found=C2=A0here.

= In the end, yes, you are correct that you must do this "manually"= and compare the energetics of different possible spin multiplicities. Your= approach may end up being sufficient, but as I mentioned, there are some f= actors that you should at the very least keep in mind.

Andrew
=
On Sun, Sep 10, 2017 at 8:2= 0 AM Henrique C. S. Junior henriquecsj+/-gmail.com <owner-chemistry^^ccl.net> wrote:

Dear coll= eagues, I=E2=80=99m working with a Cobalt(II) trimer whose molecular struct= ure was achieved by Single Crystal X-Ray Diffraction. My task now is to che= ck the spin states of the structure (High or Low spin). Since Co(II) can ha= ve 1 or 3 unpaired electrons, I=E2=80=99m approaching this problem by calcu= lating Single Points for every possible multiplicity (10, 8, 6, 4, 2) and a= ssuming that the most stable is the one that represents my structure (and m= y spin states).

Is this approach correct?

Thank you


--
Henri= que C. S. Junior




--
Henrique C. S. Junior
Industrial= Chemist - UFRRJ
M. Sc. Inorganic Chemi= stry - UFRRJ
Data Processing Center - PMP
<= div>Visite o Mundo= Qu=C3=ADmico
--94eb2c0ba0fc7c107a0558dc4b23--