VB orbitals in $ORB and $GUS sections of the inputs in transition metals

[Oscar Ac]

Hello,

I want to run a L-VBSCF calculation of the Mn-Mn bond in Mn2(CO)10 but cannot find a Lewis structure or an alternative way to fill in the $ORB and $GUS sections.

My problem is that I have 95 MOs and therefore 96 VBOs. I have classified them as:

# 2 Cores of Mn

9*2= 18 orbitals       (Nine from 1s,2s,2px,2py,2pz,3s,3px,3py,3pz for each Mn atom).

# 10 Cores of C

1*10=10 orbitals     (1s orbitals of each C atom)

# 10 Cores of O

1*10=10 orbitals     (1s orbitals of each O atom)

# Mn-C bonds

1*10=10 orbitals     (5 bonds for each Mn)

# C-O bonds

1*10=10 orbitals     (A bond for each C)

# Active VBO

1*2=2 orbitals         (One for each Mn in the Mn-Mn bond)

Which sums to 60 VBOs. Now, I don't know an adequate way of deducing the other 36 VBOs using the lone pairs of the O and C atoms. If I use 3 orbitals (2s, 2px and 2py) for the lone pairs of the O atoms, I would have 30 orbitals and could not accommodate the other 6 VBOs using the 10 C atoms. I have also tried thinking about the 4s and 3d orbitals of the 2 Mn atoms but they do not sum up correctly in the end with the lone pairs. What am I doing wrong in my assignation of VBOs?

Afterwards, in the $GUS section, I must make the correspondence between the VBOs in the MOs and the MOs don't give a hint of an obvious way to relate them to some choice of the VBOs due to the delocalisation (which can be especially confusing for the ones involving the all the bonds except the Mn-Mn one).

I've dealt with this for a while and become a bit hopeless about it, I hope you can please help me.

Thank you very much for your help!