excitation of electron in an impurity level.

[souffler]

Hi all.

I study the recombination in a direct band-gap semiconductor.

My questioning is that which energy-level the exciting electron settle down when a direct band-gap material exposed by photon (Photoluminescence).

The electrons near the valence band maximum(VBM) jump to the energy states near conduction band minimum(CBM) by the exposure of photon. Then they recombine with the electron-holes in VBM emitting photons. As far as I know, I can understand the excitation and recombination using the concept of the exciton.

By the way, what happens in the case of (electron occupied) impurity level (between band-gap) forming by doping impurity element to the direct band-gap semiconductor?    

So, which energy level the electrons in the impurity level jump to? Do they also jump to the energy level near CBM, or the other state of the impurity itself, or the other way of excitation?  In my opinion, there would be no corresponding energy state of the impurity level such as the bonding - antibonding state. 

In addition, in the condition of the excitation or recombination, should there be spacial overlap between the wavefunctions of the starting and arriving states? 

[Enthalpy]

Hi Souffler,

a photon can and does take electrons from the depth of the valence band too, and send them deep in the conduction band. That's why materials absorb many different wavelength, not just light with energy equalling the bandgap. In silicon (indirect gap) this process is much more efficient: you can check that photons with just the bandgap energy are little absorbed.

Other bands can contribute the photon absorption too, or secondary extrema in the bands.

Recombination occurs most often among the band edges because this is where there are candidates, at least near thermal equilibrium. But under strong injection, hotter electrons and holes (=empty energy levels) can be available for recombination. This happens at laser diodes, whose wavelength is imperfectly defined.

The spin and momentum conservation impose other restriction that may need the help of a phonon, making the process less probable.

Impurities are isolated in the crystal if the doping isn't too heavy, and then the doping level is nearly a single energy. Excited states can exist, but they differ by few meV, often neglected.

These impurity levels do lose and get electrons to and from both bands, and not only at the edges. Choosing the dopant hence its energy level defines the colour of GaP leds.

In both bands, the electrons are largely delocalized. If they are states of perfectly defined energy or momentum, they are delocalized to the whole crystal or doping zone. So overlapping is trivial. In dopant levels, the electrons are localized to the atom, so it can jump to a band, but jumping to an other dopant is difficult.