Fermi-energy edge singularity and excitonic enhancement associated with the second subband in asymmetric modulation-doped quantum wells

A study of the Fermi-energy edge singularity (FEES) in the second (n=2) subband of asymmetric modulation-doped (AlGa)As-(InGa)As-GaAs quantum wells (AMDQW’s) is reported. In one of the AMDQW’s a Schottky gate is employed to vary the electron density in the n=2 subband (${\mathit{n}}_{\mathit{s},2\mathrm{}}$) from 0 to 1×${10}^{11}$ ${\mathrm{cm}}^{\mathrm{-}2}$. Temperature-dependent photoluminescence excitation (PLE) measurements clearly show that the n=2 PLE feature has FEES character for ${\mathit{n}}_{\mathit{s},2\mathrm{}}$≳0.4×${10}^{11}$ ${\mathrm{cm}}^{\mathrm{-}2}$. In contrast to PLE, photoluminescence (PL) intensity is not a true measure of oscillator strength, since PL intensity can be affected by competing recombination pathways. Temperature-dependent PL measurements have been performed on two types of AMDQW. One type has ${\mathit{n}}_{\mathit{s},2\mathrm{}}$∼0, with the Fermi energy close to the n=2 subband energy. The other type has ${\mathit{n}}_{\mathit{s},2\mathrm{}}$=1×${10}^{11}$ ${\mathrm{cm}}^{\mathrm{-}2}$ and a FEES associated with n=2 observed in PLE. We demonstrate that the very similar broadening and reduction in peak height of the n=2 PL peak with temperature for the two types of samples can be accounted for in terms of spreading of the electron or exciton populations near the n=2 subband edge. Therefore, we conclude that temperature-dependent PL does not provide unequivocal evidence for a many-body enhancement of the n=2 PL transition, in contrast to that reported by Chen et al. [Phys. Rev. Lett. 64, 2434 (1990)].