bjb2chen / VEmodel

2 stars 0 forks source link

RhF3 model #24

Open bjb2chen opened 10 months ago

bjb2chen commented 10 months ago

Today I started doing RhF3.

Started with the no valence electron energy calculation because RhF3 has the same valence electrons as CoF3 (however it's 4d8 5s1, not 3d7 4s2). This meant that I encountered an error where gbasis=cct does not suffice. I elected to use SPKrTZP, adding relwfn=LUT-IOTC in $contrl.

Finally, kept getting the THERE ARE ATOMS LESS THAN 0.100 APART, QUITTING.. error. So I took my D3H structure, opened the Coordinates tab, and selected the Fluorine atom with zeroes in y, z columns:

Rh 45.0 -0.00000 0.00000 0.00000 F 9.0 2.07

If there are two non-zeroes in the x,y,z the D3H calculation will fail. So you need to choose the atom with only one non-zero value and initiate from there.

bjb2chen commented 10 months ago

Could not converge in this manner with filling by bunches of electrons at a time.

What worked was doing a one-shot calculation and filling all electrons in their expected occupation and orbitals from the very start. So this meant mult=5, icharg=0, ROHF, SPKrTZP. This actually worked! Was able to optimize_gh calculation and then 1st, 3st, and 5st diabatization.

bjb2chen commented 10 months ago

Toby was happy and then recommended a new calculation at 3 state reference geometry: RhF3_SPK_gmcpt_C1_e_mult5_diis_3st_diab_refG_SOC.inp

bjb2chen commented 9 months ago

We looked at the 15 state orbitals and found that they were mixed in $d{xy}$ and $d{x^2-y^2}$, with some p orbitals. This meant that we had to extract the pure orbitals from the original ROHF calculation and insert them into the DMO group for the 15 state calculation to replace the mixed degenerate ones.

bjb2chen commented 9 months ago

Now we will take dmo.dat and renamed it $VEC, so when we loaded this file into MacMolPlt, it shows as the "Actually Used MOs" and then we will plug them in as DMO group for diab calculation to get refdet. Which we will discuss more on tomorrow.

bjb2chen commented 9 months ago
Screenshot 2024-01-16 at 2 00 43 PM

This has nothing to do with RhF3. It is actually part of CoF3 calculation that I found pretty cool and just wanted to keep around for posterity.