In solids the phonon-assisted, nonradiative decay from high-energy electronic excited states to low-energy electronic excited states is picosecond fast. It was hoped that electron and hole relaxation could be slowed down in quantum dots, due to the unavailability of phonons energy matched to the large energy-level spacings (“phonon-bottleneck”). However, excited-state relaxation was observed to be rather fast in InP, CdSe, and ZnO dots, and explained by an efficient Auger mechanism, whereby the excess energy of electrons is nonradiatively transferred to holes, which can then rapidly decay by phonon emission, by virtue of the densely spaced valence-band levels. The recent emergence of PbSe as a novel quantum-dot material has rekindled the hope for a slow down of excited-state relaxation because hole relaxation was deemed to be ineffective on account of the widely spaced hole levels. The assumption of sparse hole energy levels in PbSe was based on an effective-mass argument based on the light effective mass of the hole. Surprisingly, fast intraband relaxation times of were observed in PbSe quantum dots and have been considered contradictory with the Auger cooling mechanism because of the assumed sparsity of the hole energy levels. Our pseudopotential calculations, however, do not support the scenario of sparse hole levels in PbSe: Because of the existence of three valence-band maxima in the bulk PbSe band structure, hole energy levels are densely spaced, in contradiction with simple effective-mass models. The remaining question is whether the Auger decay channel is sufficiently fast to account for the fast intraband relaxation. Using the atomistic pseudopotential wave functions of and quantum dots, we explicitly calculated the electron-hole Coulomb integrals and the electron Auger relaxation rate. We find that the Auger mechanism can explain the experimentally observed intraband decay time scale without the need to invoke any exotic relaxation mechanisms.
REFERENCES
If strong electron-lattice coupling exists, the adiabatic approximation breaks down, leading to no phonon bottleneck, even if there is some energetic mismatch. See Refs. 14–16.
PbSe does have different properties relative to ordinary II-VI materials, e.g., its dielectric constant is 22.9, while the dielectric constant of CdSe is 6.3. However, this leads to only quantitative differences in carrier decay, not qualitative differences.
