In semiconductors, can phonons create excitons?

[jeebs]

As I understand it, an exciton is created when an electron is liberated from the valence band of a semiconductor, but becomes bound to the hole that it leaves vacant via Coulomb attraction.
We can calculate the quantized energy levels of the exciton just like hydrogen. I have done this for the ground state of an exciton in GaN, which turns out to be about 0.023eV.
At room temperature, the thermal energy is about kT = 0.026eV. These are two close figures. Am I correct in thinking kT can be considered the phonon energy?

I was wondering, do excitons get created by phonons knocking the electrons out of the valence band, or is it only able to happen when there is a photon of wavelength equal to the ground state (or an excited state) of the exciton?

 

[Cthugha]

No, excitons will usually not be created thermally. The minimum energy needed to create an exciton is the band gap energy minus the exciton binding energy. GaN is a wide band gap material and the band gap should be somewhere near 3.2 eV at room temperature. Although the exciton binding energy is quite large in GaN, it is still much smaller than the band gap, so you will usually not get the (almost) 3.2 eV needed to form an exciton out of phonons.

However, excitons are usually destroyed thermally at high temperatures. In this case the thermal excitation creates free electrons and holes instead of the bound electron-hole state.