Mulliken population

[Shan]

Hello all...

I calculated the mulliken population of isolated boron where i got a value of 3. I used it in pyridine molecule in place of the carbon just beside the nitrogen, where i got a mulliken population of 3.537. What exactly can i infer from these values.... does  it mean the boron gained a charge of 0.537 after doping  it to pyridine?

If so, when i performed the same calculation using Gaussian03 it shown a charge loss of 0.176.

I really stuck in big confusion, please help me in this matter....

Thank you in advance....

[zh]

This is a quite common problem in the Mulliken population analysis. The charge may strongly depend on the used basis set.

[Anders Blom]

As ZH mentions, Mulliken charges are not particularly "physical". You need to define more clearly what you mean by "charge transfer" and then study the literature for how people traditionally compute this quantity. And more importantly, I think: why do you want to calculate it, what will the precise number tell you?

[Shan]

Dear professor Zh and Professor Anders,

Thank You for the kind reply.
But I am sorry to say, my doubt did not get cleared yet.
 
As I mentioned in the question earlier, did the boron gained charge in my case in atk?
I used DZP here in atk, when plotted electron density using the same basis set, it shows charge loss at the position of boron.

Even if I used different basis set in both tools, how can the basic property can change from charge losing to gaining?

expecting your kind reply soon.

[Anders Blom]

Because the Mulliken charges - as mentioned - are not always a good measure of the physical charge transfer. It's really just a number corresponding to the limited trace (over a set of orbitals, centered on the same atom) of the density matrix. Of course, in certain cases (mostly for physisorption) it can give a reasonable estimate of the charge transfer (at least its sign) but if we are talking about isolating individual atoms in a strongly convelently bonded system (like inside a molecule) I don't think it's a proper measure.

[Shan]

Thank you very much professor Anders...

[Jenny]

Because the Mulliken charges - as mentioned - are not always a good measure of the physical charge transfer. It's really just a number corresponding to the limited trace (over a set of orbitals, centered on the same atom) of the density matrix. Of course, in certain cases (mostly for physisorption) it can give a reasonable estimate of the charge transfer (at least its sign) but if we are talking about isolating individual atoms in a strongly convelently bonded system (like inside a molecule) I don't think it's a proper measure.

Dr. Blom,

I've just found your old post here may solve my current problem. Now I have a new question that how can I find whether there is a covalent bonding between in the system via mulliken population or other methods in QuantumWise.

Best,

Jenny

[Anders Blom]

You can calculate the Mulliken bond population for pairs of atoms. A high number indicates a stronger ("more covalent") bond.

However, as we know, bonds are not real, they are just used to make pretty pictures :-) What matters is the sharing of electrons, and for that there is always a floating scale from non-bonded to strongly bonded, but there is no universal cut-off defining what is a covalent bond or not.

[Jenny]

You can calculate the Mulliken bond population for pairs of atoms. A high number indicates a stronger ("more covalent") bond.

However, as we know, bonds are not real, they are just used to make pretty pictures :-) What matters is the sharing of electrons, and for that there is always a floating scale from non-bonded to strongly bonded, but there is no universal cut-off defining what is a covalent bond or not.

Dr. Blom,

Thank you for your answer. So how can I investigate the Mulliken bond population for paris of atoms as you mentioned? just by adding the mulliken population of single one atom and its nearby atom? If not, how can i do the calculation?

Best,
Jenny

[Anders Blom]

See the manual - always a good resource
http://quantumwise.com/documents/manuals/latest/ReferenceManual/index.html/ref.mullikenpopulation.html