1
$\begingroup$

I came across a problem whose statement is as follows:

An electron moves in the potential well $P (x) = -\delta$ for $- a <x <0$ and $P (x) = \delta$ for $0 <x <a$ (Fig. 13.7). Use first-order perturbation theory to compute the first four energy levels. Set up the expression for the first-order expansion coefficients for the lowest energy state.enter image description here

I was trying to use the following equation that I saw throughout the chapter (number 13):

$$ E'_n - E_n = \int_{-\infty}^{\infty} \psi_n^{*} f(x) \psi_n dx $$

Treating $f(x) = P(x)$ and the $\psi_n$ function equal to

$$ \psi_n = \sqrt{\frac{2}{a}} \sin{\left( \frac{n \pi}{a} x\right)} $$

However, this gives me zero as a result of $E'_n - E_n$.

How can I proceed?

The textbook I'm using is Richtmyer F.K.,Kennard E.H.,Cooper J.N. - Introduction to modern physics-McGraw-Hill (1969).

$\endgroup$
2
  • $\begingroup$ Isn’t that $\psi_n$ for a square well from $x=0$ to $x=a$? $\endgroup$
    – G. Smith
    Commented Oct 26, 2020 at 21:03
  • $\begingroup$ That's right. First I was trying to use this $\psi_n$ everywhere, but then I realized that this is not good. And I'm having doubts about the usage of this in some specifc region in the figure. I mean, could I use it in the $x=-a$ to $x=0$ interval? $\endgroup$
    – Ron Stean
    Commented Oct 26, 2020 at 21:28

1 Answer 1

1
$\begingroup$

If you have a perturbation $P(x)$, then to 1st order, the energy shifts are:

$$ E^{(1)}_n - E^{(0)}_n = \langle n|P(x)|n\rangle $$

but you have to use the $|n\rangle$ that solve your unperturbed hamiltonian, $H^{(0)}(x)$, not the ones that solve $H^{(0)}(x'=\frac 1 2 (x+a))$.

You can either transform your $\psi_n(x')$ into $\psi(x)$ to match $H(x)$, or you can transform $H(x)$ to $H(x')$ to match you $\psi_n(x')$.

Note that as stated, you have an odd perturbation, so if an unperturbed solution spends 1/2 its time in the left (right) side of the box, the half of the time, the energy is lowered (raised), so you would expect the 1st order perturbations to be zero, esp. for even solutions, since that should be zero by symmetry.

You would expect the 1st order perturbation to the wave function to shift some weight to $x<0$, since the potential is deeper there.

$\endgroup$
1
  • $\begingroup$ You'd expect the 1st order perturbation to the ground state to shift towards $x\lt 0 $. For excited states the classical logic is that the particle moves faster in the lower potential regions and so spends less time there, thus they should shift to the right. $\endgroup$
    – jacob1729
    Commented Oct 26, 2020 at 23:04

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.