# Statistical mechanics: Heat capacity

1. Aug 11, 2010

### espen180

I am trying to work out the heat capacity of a body-centered cubic iron lattice using stat.mech., but am having some trouble.

Firstly, I assumed that the iron atoms behaved as harmonic occilators, not taking electronic or nuclear spin into account. Is this a good or bad approximation?

Then, when I compute the partition function and calculate the heat capacity $$C_V=\frac{dU}{dT}|_V$$, I get

$$C_V=\frac{N\hbar^2\omega^2}{kT^2}\left(\frac{e^{\frac{\hbar\omega}{kT}}}{\left(1-e^{\frac{\hbar\omega}{kT}}\right)}+\frac{e^{\frac{2\hbar\omega}{kT}}}{\left(1-e^{\frac{\hbar\omega}{kT}}\right)^2}\right)$$.

I can provide intermediate steps if neccesary.

For large T, the value of this expression is $$C_V=Nk$$. I find this an indication that either my model isn't working or I have messed up, since I think the heat capacity should be dependent on the bond strength and packing density of the lattice.

If I want to make a better approximation, how do take the influence the different iron atoms have on each other into account?

Thanks in advance.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Can you offer guidance or do you also need help?
Draft saved Draft deleted