I am familirazing myself with the lattice dynamics reading Ashkroft-Mermin (p2 ch 22). My question is what it the mathematics behind the deriviation

$U^{harm}=1/2 K \sum_{n} [u(na)-u([n+1]a)]^2$ (22.22)

$\frac{\delta U^{harm}}{\delta u(na)}=-K[2u(na)-u([n-1]a)-u([n+1]a)]$ (22.23)

I feel like it is very basic and I fail to see it or I totally misunderstand the notation. ( I get something like $-K(u(na)-u[(n+1)a]$).

How it comes, that one of the displacements is for (n-1)a atom? Is the 22.23 has this form to fit the concept of this lattice to be a system of balls connected by spring?

  • $\begingroup$ $u(na)$ appears in two terms in $U^\text{harm}$. $\endgroup$
    – Meng Cheng
    Commented May 16, 2015 at 15:41

1 Answer 1


Make sure when you vary your first equation that you use a different index instead of $n$. Do the following:

$$\frac{\delta U^{harm}}{\delta u(ma)}$$

this gives

$$ \frac{1}{2}K \sum_{n} 2 [u(na) - u([n+1]a)]\left[\frac{\delta u(na)}{\delta u(ma)} - \frac{\delta u([n+1]a)}{\delta u(ma)} \right] $$ by the chain rule. But $$ \frac{\delta u(na)}{\delta u(ma)} = \delta_{n,m} \text{ and } \frac{\delta u([n+1]a)}{\delta u(ma)} = \delta_{n+1,m} $$ Now we have four terms: \begin{equation} K\sum_{n}[u(na)\delta_{n,m} - u(na)\delta_{n+1,m} - u([n+1]a)\delta_{n,m} + u([n+1]a)\delta_{n+1,m}] \end{equation} and performing the sum for each term forces $n$ to take on the necessary values. This gives $$ K[u(ma) - u([m-1]a) - u([m+1]a) + u(ma)] = K[2u(ma) - u([m-1]a) - u([m+1]a)] $$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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