Photoluminescence dynamics of cavity polaritons under resonant excitation in the picosecond range. J. Bloch and J. Y. Marzin in Phys. Rev. B 56:2103 (1997).  What the paper says!?

This theoretical paper «clarif[ies] the role of phonon scattering in [...] resonant photoluminescence experiments» performed experimentally with co-authors, using Boltzmann equations, exciting both the lower & upper branches (resonantly). Comparison to experiments are presented that reproduce «very well the short-time behavior of resonant luminescence measured in samples described in Ref. 6.»

There is a nice literature review of important/pioneering experiments of the time:

the Rabi splitting has been measured in reflectivity, absorption as well

as in photoluminescence.$^{1-5}$

time-resolved luminescence experiments were performed in thin semiconductor microcavities to study spontaneous emission

modification induced by cavity effect.$^{6-11}$

with some comments. For instance, the biexponential decay in PL following resonant excitation observed by Sermage et al. results from refilling of the radiant states from a reservoir of dark excitons.

They make a (rare) discussion of how they treat the ring degeneracy. It remains unclear to me how this properly accounts for Bose stimulation: a polariton is only stimulated by those on the very final state it goes to, no by others on the same ring. But they, curiously, don't appear to take into account Bose stimulation anyway in their rate equation!

There also seems to be a typo in their Eqs. (10) & (12), $dk$ should be $\delta k$. Here we see how they integrate over the ring:

Bose "Stimulation" is only mentioned here:

Notice that acoustic-phonon scattering has already been in- cluded in the calculation of stimulated emission in a micro- cavity by Pau et al.$^{13}$ Moreover we compare our results with a recent paper of Tassone et al.,$^{14}$ where the temporal re- sponse of an analogous system under nonresonant excitation

is considered.

The comparison is as follows:

As in Ref. 14, we also calculated the decay time of a sharply peaked excitonic population at the energy of the exciton con- tinuum and found, in agreement with Tassone et al., weak dependance of the overall dynamics on the detuning between cavity and exciton. In the present calculation however, since we create the initial carrier population resonantly on the up- per branch, we find a strong dependence of the rise time on the detuning in agreement with experiments. But as in Ref. 14, after a few picoseconds, the carriers thermally fill the reservoir and then the luminescence decay reflects the emp- tying of the reservoir and therefore weakly changes with de-

tuning.

But Tassone et al. do consider Bose stimulation in their rate equations, so this suggest it does not play a significant role in this dynamics. This is, however, not discussed!