0
$\begingroup$

I don't understand the integral expansion done for equation 2.73 in Solid State Physics by Ashcroft and Mermin.

Specifically the equation deals with the Sommerfeld Expansion and looks like this: $$\int_0^\mu H(\epsilon) d\epsilon = \int_0^{\epsilon_f} H(\epsilon) d\epsilon + (\mu-\epsilon_f)H(\epsilon_f) + ...$$

I am trying to extend this to:

$$\int_0^\mu H(\epsilon) d\epsilon = \int_0^{\epsilon_f} H(\epsilon) d\epsilon + (\mu-\epsilon_f)H(\epsilon_f) + \frac12(\mu-\epsilon_f)^2H'(\epsilon_f) + ...$$

But need to understand the source of the first equation before I can do so. It feels like I'm overthinking this or missing something obvious! Am I just being dumb?!

$\endgroup$
  • $\begingroup$ I don't have the book at hand now. Do you mean $\int_0^{\epsilon_f}{H(\epsilon)d\epsilon}$ on the right hand side (first term)? $\endgroup$ – DelCrosB Oct 3 '16 at 21:26
  • $\begingroup$ Ah yes, my bad. Good catch. $\endgroup$ – onb Oct 3 '16 at 21:37
1
$\begingroup$

Define the function $$\boxed{f(x) := \int_0^x H(\epsilon) d\epsilon}$$

Suppose we want to expand this function around the value $\epsilon_f$, i.e. $x = \epsilon_f + \delta x$ with $\delta x$ small. Then of course $$ f(x) = f(\epsilon_f) + \delta x \; f'(\epsilon_f) + \frac{\left(\delta x\right)^2}{2} \; f''(\epsilon_f) + \mathcal O \left( (\delta x)^3 \right)$$

Note that $f'(x) = H(x)$ (which also directly gives $f''(x) = H'(x)$). In case this is not clear, it is just a one-liner using the definition $f'(x) = \lim_{\alpha \to 0} \frac{f(x+\alpha) - f(x)}{\alpha} $.

Plugging this in, we get $$\boxed{f(x) = \int_0^{\epsilon_f} H(\epsilon) d\epsilon + \delta x \; H(\epsilon_f) + \frac{\left(\delta x\right)^2}{2} \; H'(\epsilon_f) + \mathcal O \left( (\delta x)^3 \right) }$$

Simply setting $x = \mu$ (and hence $\delta x = \mu -\epsilon_f$), we're done.

| cite | improve this answer | |
$\endgroup$
1
$\begingroup$

I think that the idea is: $$\int_0^{\mu}{H(\epsilon)d\epsilon}=\int_0^{\epsilon_f}{H(\epsilon)d\epsilon}+\int_{\epsilon_f}^{\mu}{H(\epsilon)d\epsilon}\approx\int_0^{\epsilon_f}{H(\epsilon)d\epsilon}+(\mu-\epsilon_f)H(\epsilon_f)$$ where the last step is motivated by the fact that the interval $\mu-\epsilon_f$ is very narrow and so the function $H$ can be assumed to be constant and equal to its value at $\epsilon_f$ in that region. I imagine that the book is showing how to do a low $T$ approximation, and so the chemical potential is only slightly different from the Fermi energy.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.