# How to do first principle calculation in electron gas condition

1. Oct 19, 2005

### pigbear

I try to to some first principle calculation in the environment of free electron gas. For intance, the total energy of one atom or a cluster in the free electron gas.
What kind of calculation program could I choose ? I only know that VASP provide an opion to create charged background. However, it 's designed for periodic structures intead of atom or molecular calculation? For calculate few atom, a trick is required and it's so time and memory consuming!!.

2. Oct 19, 2005

### ZapperZ

Staff Emeritus
Let me understand you clearly here. You want to put an atom in a free electron gas, and you want <giggle> to calculate, from First Principle, all the interactions of that atom with, let's say, an Avogadro's number of electron?!

I had to clarify that this is what I understood from what you posted, before I fall off my chair. :)

Zz.

Last edited: Oct 19, 2005
3. Oct 19, 2005

### pigbear

Sorry for unclarity of the question!

we know that the model of free electron gas is just one model for explaination the property of metal. Though it is quite old and out of fashion, it might be still qualified to simplify the system containing dilute alloy atoms in the metal matrix, for instance Al, in which alloy atom will only feel sitting in the background of electron gas, which screen the nucleus of Al atom. The concertration of electron gas is definated.

4. Oct 19, 2005

### ZapperZ

Staff Emeritus
Er.. no. The free-electron gas is simply one, elementary model, not THE model, for metals. If you open Ashcroft and Mermin's Solid State text, you'll see Chapter 1 dealing with such a model (i.e. the Drude model), but then look at Chapter 3 - Failure of the free electron model!

There are already several different levels of theoretical treatment of metals - the most relevant of which is Landau's Fermi Liquid theory. I suggest you look into this before proceedings. Beyond that, you will start to get into more exotic conduction behavior exhibited by TiO or Cr, for instance. You then get into Mott-Hubbard model, Brinkmann-Rice model, etc... etc. The point here is that if you think the free-electron gas is the ONLY model describing conduction in metals, you would be wrong.

Zz.

5. Oct 19, 2005

### pigbear

But the point now is not on Metal, I don't care the conductivity or magnetism of metal matrix. My interest is now on the solute atoms in the metal matrix. Metal matrix is just an environment of solute atoms, like the function of water for the ions reaction in solution, which provide an environment of dipole. And here metals provide an environment of charge.

6. Oct 19, 2005

### ZapperZ

Staff Emeritus
Then I would again ask you the same question I asked in the very beginning. Are you trying to describe an atom in a free electron gas?

BTW, what is a "metal matrix"? Are you refering to the crystal lattice as a metal matrix?

Zz.

7. Oct 19, 2005

### pigbear

Yes, my purpose is try to describe atoms and moleculars in the charge background . The number of background electrons can be derived by the density of free electron gas multify the volume occupied by solute atom.

in solid solution , one type of element will keep its crystal lattice almost unchanged, its lattice is called matrix, other atoms (normally the concentration of them in matrix C(solue atom)= No.(solute atom)/ {No.(matrix atom)+No.(solute atom)} is lower than 50% ) are called solute atoms.

8. Oct 19, 2005

### ZapperZ

Staff Emeritus
Then why didn't you just say "yes" when I originally asked

Unless you're willing to deal with the mean-field approximation, I have nothing useful to tell you.

Zz.

9. Oct 19, 2005

### pigbear

In fact, my idea is like that the reaction of two atoms in the water solution is unlike the reaction in air condition, since water provide an environment of dipole. Analogically, it should also be different if we put them in metal matrix.

10. Oct 19, 2005

### pigbear

I was afraid of saying yes since you said "an Avogadro's number of electron?!"

11. Oct 19, 2005

### ZapperZ

Staff Emeritus
I said "let's say, Avogadro's number..." to emphasize the HUGE quantity of interactons. It is impossible to write down all and every interactions in such a case. In condensed matter physics, we never calculate such thing without invoking many-body physics - which is why I mentioned the mean-field approximation. From what I have read, I don't think you are either aware of it, or maybe that's the avenue you choose not to pursue.

Zz.

12. Oct 19, 2005

### Gokul43201

Staff Emeritus
pigbear : I think (if I understand correctly, what you are trying to say) that what you want is to do a DFT calculation based on the Extended Huckel Model (or similar method). This is the popular approach for molecular orbital calculations of the binding of particles/clusters to a metallic substrate.

Last edited: Oct 19, 2005
13. Oct 20, 2005

### pigbear

Yes. Gokul43201
Those particles / clusters embeded in metallic materials. And if we ignore the influence of necleus. We can say those clusters are in the electron gas, can't we.

14. Oct 20, 2005

### Gokul43201

Staff Emeritus
I don't know how good/bad an approximation that would provide, or what level of accuracy you desire, but this is definitely NOT what current computational methods use. Did you try doing a search with the keywords I provided above ?

15. Oct 21, 2005

### pigbear

I'm reading one book " Methods of Electronic structure calculations", in which Extended Huckel Model" is described and also some papers about it. Thanks for your suggestion. However, I don't think it works for my purpose. as you said, it' Not what current computational methos use.

In fact, I did find one software "VASP", which provide charged background. This program is based on " pseudopotential" method, designed for crystal structure. If I want to calculate the property of cluster, a big super cell have to be created, which induce time and memerory consuming problem.

G03 seems good for clusters and moleculars, however, it provides only dipole backing ground.

16. Oct 29, 2005

### msdeng

molecular electronics

I think so, G03 is wonderful!
But I have some questions on molecular electronics, some as:
the dimensionality of metal contacts have a strong effect on the I-V characteristics of the molecular junction. How to understand?