MOM without orbital number

Hii all,

I am trying to calculate the vertical ionization energy of a small water cluster. I am having issues with the perusal of MOM method here. So, basically I have an ion surrounded by water molecules that needs to be ionized. I calculate the GS energy normally, and then read the existing geometry for the second job, and ionize the ion using MOM method.

Now, one thing that I observe is that, the orbitals to be ionized(/excited) are mentioned using the “mom_occ/mom_vir” variables, for which I would need to know the exact orbital no. corresponding to the particular ion. Is there a way to make sure, that I can ionize the desired atom in a simpler manner? The reason I am asking is that, I have 100s of systems like these with different no.'s of water molecules, which would make it very cumbersome to know the exact orbital no. corresponding to the ions in those systems for every job.

Here is an example system :

molecule ref{
-1 1
Cl 0.000000 0.000000 0.000000
O 3.024572 0.909093 0.854854
H 3.608601 0.629260 0.142819
H 2.208308 1.189307 0.414287
O 0.538286 -0.156712 2.992781
H 0.545049 0.065416 2.084439
H 1.183132 0.415955 3.472148
O -0.979486 -2.574059 1.341671
H -0.762628 -1.667590 0.988259
H -1.826565 -2.714725 0.890969
O 2.560867 -2.284647 0.576523
H 2.406005 -2.326005 -0.380384
H 2.488851 -1.342974 0.854560
O -1.266869 2.412964 1.954434
H -1.059941 1.546410 1.623652
H -1.490492 2.242243 2.871224
}

Please let me know,

Thanks,
Best regards.

I want to make sure I understand you correctly. You want orbitals for the anion, with the negative charge on the metal, for several geometries. You want to do this by taking the converged orbitals for the neutral system, and applying MOM to those orbitals to generate a guess for the anion. Is that correct?

First of all, thanks for reaching out.

As you see my system, it’s a chloride ion with water around it. So, I want to ionize from the chloride ion. For the VIE, I plan to use the same geometry for the ionized state as the reference (ground) state by “set guess read”. And as chloride ion’s MOs are not the HOMO, I will be using MOM to maintain that hole in the chloride orbital during my ionized state calculation.

Now i have several clusters like this, with 4,5,6…20,30 water molecules around the chloride. I need to know to know an easy way how I can make the ionization from the chloride irrespective of the size of the size of the cluster.

Thanks.

set guess read has nothing to do with using the same geometry for two separate computations. MOM is implemented in Psi4 only for excited states. While there may be variants that can generate a guess for a different number of electrons, I’m not familiar with them.

Instead, why don’t you specify the charge of the metal separately from the charge of the water cluster? The procedure below is completely automatic.

molecule ref{
-1 1
Cl 0.000000 0.000000 0.000000
---
0 1
O 3.024572 0.909093 0.854854
H 3.608601 0.629260 0.142819
H 2.208308 1.189307 0.414287
O 0.538286 -0.156712 2.992781
H 0.545049 0.065416 2.084439
H 1.183132 0.415955 3.472148
O -0.979486 -2.574059 1.341671
H -0.762628 -1.667590 0.988259
H -1.826565 -2.714725 0.890969
O 2.560867 -2.284647 0.576523
H 2.406005 -2.326005 -0.380384
H 2.488851 -1.342974 0.854560
O -1.266869 2.412964 1.954434
H -1.059941 1.546410 1.623652
H -1.490492 2.242243 2.871224
}

Sorry for the late reply. I was trying the way you suggested.

However, when I used a cation such as Li+ in place of the anion, the difference of the reference and ionized came out to be an absurd value. (13.45 eV compared to the exp.value ~65 eV)

These are the kind of input files I am working with,

molecule ref{
1 1
Li 0.000000 0.000000 0.000000

0 1
water coordinates
}

set basis 6-31+G*
set reference rhf
memory 4000mb
ref_energy = energy(‘mp2’)

molecule ion{
2 2
Li 0.000000 0.000000 0.000000

0 1
water coordinates
}

set basis 6-31+G*
set guess read
set reference uhf
memory 4000mb
ion_energy = energy(‘mp2’)

whereby the ref_energy-ion_energy will give the ionization energy. As you see, during the SCF iterations the hole created from the Lithium orbitals in the ionized state, are filled up by aufbau principles, unless I am using MOM. However, in MOM, I need to specify the occupieds and virtuals which is not possible when I am working with varying number of water molecules. So I need to know, if there is a way around it.

Thanks once again,

Best regards.

The 13 eV looks more like the IE of water, which would be the expected outcome.
I have little hope for an automated procedure to get these very specific metal states. And not sure what physical meaning they really have.

One way to avoid ionizing the water is switching to point-charges or PCM, if electrostatic screening is what you need from the water cluster.

Exactly. The 13 eV is IE of water. The physical meaning is to study the electronic structure of the ions in the solvated phase. I am actually trying to calculate exact electrostatic screening for gas phase water clusters, thus needing explicit water molecules. In case it is not possible to automate the calculation, could anyone of you please tell me a couple of things :

  1. Is it possible to calculate Vertical Ionization Energy using Psi4?
  2. Is MOM only to be used for excited state calculations, and should not be used to ionizations?
  3. If I do not mention a “mom_occ/mom_vir” along with “mom_start” what is the default occupation of the electrons?

Lastly, thank you all once again for trying to help me. Best regards.