The definition of the density operator in Pathria

silverwhale
Messages
78
Reaction score
2
Hello Everybody,

I am working through Pathria's statistical mechanics book; on page 114 I found the following definition for the density operator:
\rho_{mn}= \frac{1}{N} \sum_{k=1}^{N}\left \{ a(t)^{k}_m a(t)^{k*}_n \right \},
where N is the number of systems in the ensemble and the a(t)'s are expansion coefficents.
Now my question is: what does this definition mean? Especially the term a(t)^{k}_m a(t)^{k*}_n. I do not get it.

Any help would be greatly appreciated!
 
Physics news on Phys.org
The formula you wrote above refers to the definition of density matrix for a finite ensemble of isolated systems (beware, not for a subsystem in interaction with environment). It works as follows.


Imagine that there is an ensemble of N copies of the system. Denote the wave function describing the k-th system by

<br /> \psi^{k}(\mathbf r).<br />

(Different copies have different wave functions).


It is assumed that this function can be expressed as a discrete linear combination of some basis functions \Phi_m, which are the same for all k:


<br /> \psi^{k}(\mathbf r) = \sum_k a_m^k \Phi_m(\mathbf r).<br />

The numbers a^k are complex expansion coefficients.

(Such expansion is possible if the set of functions \Phi_m is complete, like for Hamiltonian eigenfunctions of harmonic oscillator. In case of hydrogen eigenfunctions, things are more complicated, due to continuous spectrum of Hamiltonian).

The density matrix is introduced usually as a quantity \rho_{mn} that appears in the calculation of average value of some quantity f, say energy, over the ensemble.

The average over the ensemble is the weighted sum

<br /> <br /> \langle \langle f \rangle \rangle = \sum_k p_k \langle f \rangle^k,<br /> <br />

where p_k = 1/N is the probability that the system is in state described by k-th wave function.

The expression

<br /> <br /> \langle f \rangle^k<br />

used above is the average of f in a state described by \psi^k function and can be expressed as

<br /> \langle \psi^k | \hat f | \psi^k \rangle = \sum_{m,n} a_m^{k} a_n^{k*} f_{nm}. <br />

where f_{nm} = \langle \Phi_n|\hat f|\Phi_m\rangle.

Then, the average over the ensemble is

<br /> <br /> \langle \langle f \rangle\rangle = \sum_k \frac{1}{N} \sum_{m,n} a_m^{k} a_n^{k*} f_{nm}. <br />


This can be rewritten as

<br /> <br /> \sum_m \left( \rho_{mn}f_{nm} \right)<br /> <br />

where the quantity

<br /> \rho_{mn} = \sum_k \frac{1}{N} a_m^{k} a_n^{k*} <br />

was named the density matrix.
 
Perfect.

Many many thanks!
If you were living in Hamburg, Germany, I would give you a bag full of cookies!
I am really thankful! :)
 
Glad to be of help.
 
Hi. I have got question as in title. How can idea of instantaneous dipole moment for atoms like, for example hydrogen be consistent with idea of orbitals? At my level of knowledge London dispersion forces are derived taking into account Bohr model of atom. But we know today that this model is not correct. If it would be correct I understand that at each time electron is at some point at radius at some angle and there is dipole moment at this time from nucleus to electron at orbit. But how...

Similar threads

Back
Top