Explaining the Unexpected Spikes in a Quantum Well Absorption Spectrum

[jeebs]

I'm looking in a textbook at a discussion of a semiconductor quantum well which is being irradiated and an absorption spectrum produced. Electrons are being promoted across the band gap and holes are being created. There are heavy hole and light hole transitions, and the spectrum shown rises in steps (corresponding to the energy levels formed from the 1-dimensional confinement characteristic of the quantum well). The threshold energy for each step is given by \hbar\omega = E_g + \frac{\hbar^2n^2\pi^2}{2m_e^*d^2} + \frac{\hbar^2n'^2\pi^2}{2m_h^*d^2} where d is the well width, m_h* can be the heavy hole or light hole mass, and n/n' = 0,1,2,3...

Having followed the derivation through I am happy about why these steps at certain threshold energies are predicted and observed. However, I notice that, on the absorption vs. photon energy plot, rather than being a perfectly square step, there will be a large spike at the heavy hole n = n' = 1, followed immediately by a smaller spike for the light hole, then the flat top of the step appears until the next threshold. The next threshold (n=n'=2) has the spikes separated a bit more, and the heavy hole peak is only slightly higher than the light hole peak. I cannot think of a reason for this.

Can anyone explain this behaviour?

 

[Cthugha]

Are you looking at room temperature or low temperature data?

The step-like absorption spectrum is the prediction for free electrons and holes in a quantum well. At low temperatures the Coulomb interaction between electrons and holes can become strong enough to form excitons, bound hydrogen-like complexes of electrons and holes. Energetically those are located at or to be more precise slightly below the band gap. That causes a peak in the absorption seen near the step edges.

At higher temperatures k_B T becomes large compared to the eciton binding energy and the peaks should vanish.