Photocurrent (PC) spectroscopy is employed to study several important aspects of the interband optoelectronic properties of InAs/GaAs self-assembled quantum dots (SAQDs). PC spectroscopy is first shown to be a highly sensitive, quantitative technique to measure the interband absorption spectra of SAQDs. Up to four well-defined features are observed in the spectra arising from transitions between confined hole and electron levels. The transition energies are shown to agree well with those observed in electroluminescence, with negligible Stokes shifts found, contrary to previous reports. Large quantum-confined Stark shifts of the transitions are observed in PC spectroscopy as discussed in detail elsewhere [Fry et al. Phys. Rev. Lett. 84, 733, (2000)]. Discrete interband transitions are observed superimposed on a broad background signal, shown to arise in part from field-dependent transitions into tail states of the two-dimensional wetting layer and the GaAs cladding region. A field-independent contribution to the background is also found, possibly from dots with larger size and shape fluctuations than those which give rise to the resolved interband transitions. By comparison of photocurrent signals from quantum dots and the wetting layer within the same sample, it is demonstrated that the quantum-dot oscillator strength is not significantly modified relative to that of a quantum well of the same surface area, consistent with performance found from quantum-dot laser devices. Polarization studies for in-plane light propagation are reported. The measurements show that the observed interband transitions involve predominantly heavy-hole-like levels, consistent with an assumption of theoretical modelling of Stark-effect results. Finally carrier escape mechanisms from the dots are deduced, with tunneling found to dominate at low temperature, and thermally activated escape becoming increasingly important at temperatures above ∼100 K. Carrier escape is shown to occur from a common level, the ground state, demonstrating that excited-state to ground-state relaxation is faster than direct excited-state escape.

• Received 21 April 2000

DOI:https://doi.org/10.1103/PhysRevB.62.16784