Halperin 1982 investigated the Landau level edge state spectrum by solving the harmonic oscillator on a half-infinite line $x \in (-\infty,s]$, where varying $s$ is equivalent to varying the momentum parallel to the edge:

$$[-\frac{d^2}{dx^2} + x^2] g_{\nu,s}(x) = E_\nu g_{\nu,s}(x)$$

where I've set all physical constants to 1, and the boundary conditions are

$$\lim_{x\to s} g_{\nu,s}= \lim_{x\to-\infty} g_{\nu,s} = 0.$$

He claims:

  • When $s \to \infty$, $E_{\nu} \to \nu+\frac12$, which makes sense since this is the usual harmonic oscillator
  • As $s\to -\infty$, $E_\nu$ passes through $2\nu+\frac32$ at $s=0$ and then grows as $s^2$.

This leads to the well-known energy levels for the quantum Hall edge states and shows they are chiral. But how does one derive the second claim? Also, is there a simple solution for $g_{\nu,s}$?


1 Answer 1


For the $s = 0$ case we just have to look at the eigenmodes of the usual quantum harmonic oscillator (the $s \to \infty$ case) which have a node at $x = 0$. They are precisely the ones with odd $\nu$ energies $E_\nu = \nu + \frac{1}{2}$, so writing $\nu = 2 n + 1$ we get $E_n = 2n + \frac{3}{2}$.

For the case $s < 0$ I only have a crappy guess to why it scales as $\sim s^2$. The harmonic solutions $|\Psi_\nu \rangle$ for the $s \to \infty$ still span the Hilbert space, so we can expand the eigenstates $ | \Phi_n \rangle$ of any $s < 0$ problem into the $| \Psi_\nu \rangle$ basis. Then, my guess is that the $| \Psi_\nu \rangle$ state which will most contribute to the expansion of the $s < 0$ ground state $| \Phi_0 \rangle$ will be the lowest energy one that has a node close to $s$

Image taken from the wikipedia article https://en.wikipedia.org/wiki/Quantum_harmonic_oscillator Image by AllenMcC, distributed under a CC BY-SA 3.0 license.

Looking at this image, I suppose that the location of the outermost node of the $ |\Psi_\nu \rangle$ state goes as $ r_\nu = - \sqrt{E_\nu} $, the classical turning point for the harmonic oscillator. Thus, the ground state energy should go as $E \sim E_\nu = r_\nu^2 \sim s^2$.


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.