Excitonic effects in coupled quantum wells

Electron-sublevel-anticrossing effects have been studied in coupled quantum wells where the exciton binding energy is comparable to the minimum sublevel splitting. The anticrossing was induced by applying an electric field to align the first and second sublevels of adjacent wells. In this situation the electron-hole Coulomb interaction has a strong effect on the splittings measured by optical techniques, because the optical spectra typically measure exciton energies rather than single-particle energies. The most striking effect is that the minimum splitting of the excitons associated with each of the split electron levels does not occur at the same field as for the minimum splitting of the bare-electron levels. One unexpected but readily observable consequence is that when the same electron-sublevel splitting is measured using two different pairs of intrawell and interwell exciton transitions, the field for minimum exciton splitting can differ by up to ∼10% from one pair of transitions to the other. We have constructed a variational model of the coupled excitons that explains these effects in terms of Coulomb mixing of the delocalized electron states. We have measured the exciton splittings directly by photocurrent spectroscopy in three GaAs/${\mathrm{Al}}_{0.3}$${\mathrm{Ga}}_{0.7}$As multiple-quantum-well structures. The samples were similar in design except that the ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As barrier thickness varied from 15 to 35 Å. By fitting our variational model to the experimental anticrossing data, we have been able to deduce the actual bare-electron level splittings rather than the exciton splittings. Within the experimental accuracy, we find that the minimum splitting decreased exponentially with increasing barrier thickness, as would be expected for simple quantum-mechanical tunneling.