Quantum dot spectroscopy

In semiconductor quantum dots, the three-dimensional (3D) confinement leads to very specific features like a delta-like density of states. This delta-like density of states which is similar to that observed in atoms is expected to lead to strong modifications of the optoelectronic properties Optical transitions can be observed between the three-dimensionally confined states of the quantum dots either between the conduction states or the valence states. These transitions are usually referred to as intersublevel or intraband transitions. The measurement of the infrared absorption is the most direct path to observe intersublevel transitions in quantum dots. Absorption usually results from the transition of carriers from the ground state of the dot, either in the conduction band or in the valence band, towards an excited state. For the InAs/GaAs self-assembled quantum dots, the absorption associated with the intersublevel transitions occurs in the midinfrared and in the far infrared spectral range. Intersublevel absorption spectroscopy is a powerful technique to get access to a very detailed experimental description of the electronic structure of the dots. We emphasize that intersublevel absorption spectroscopy is very complementary to interband spectroscopy such as photoluminescence. Intersublevel absorption corresponds to a unipolar measurement since it only involves one type of carrier (electron or separately hole) while interband transitions involve confined excitons. Infrared absorption spectroscopy directly provides the difference between confinement energies (i.e. without any difference associated with the exciton binding energy). In addition, intersublevel absorptions satisfy distinct transition rules as compared to interband transitions. An interband transition essentially follows the overlapping integral between envelope functions of the initial and final state while an intersublevel transition is governed by the dipole matrix element between the initial and the final envelope wave functions. Intersublevel absorption may therefore give access to states that are not observed in interband measurements. At last the information on polarization allows a fine tracking of the main origin of the excited levels (in-plane confinement, growth axis confinement).

Infrared Absorption of InAs/GaAs self-assembled quantum dots

The figure below shows the mid and far infrared absorption of n-doped InAs/GaAs self-assembled quantum dots measured by Fourier transform spectroscopy. The dots are populated by inserting a Si doping plane 2 nm beneath the wetting layer. The sheet doping density is nominally chosen to correspond to 2 and 1.5 electrons per dot in the samples of the left and right hand side of the figure respectively. Two different configurations are used.
On the left hand side, the absorption is measured in a normal incidence configuration in a sample containing 30 quantum dot planes separated by 50 nm GaAs barriers. The temperature is 5 K and the polarization is set along either cleaved edge directions. Two clear infrared resonances are observed at 56 and 63 meV. The full width at half maximum is 6 meV. Following the calculated electronic structure, these infrared resonances are attributed the intersublevel transitions of the electrons from the s ground state to the p excited state. The exact agreement (within less than 1 meV) between the predicted transition energies and the experiment is fortuitous but demonstrates an overall validity of the model and the input parameters. In particular the observed 7 meV s-p splitting energy agrees with a 10% elongation of the dot along the [100] direction. The broadening of the absorption is attributed to the size fluctuation of the dots. According to the multiband 3D simulation, the observed broadening correspond to a ±9 % dot size fluctuation in good agreement with structural characterizations.
On the right hand side of the figure, the absorption is measured in the 90-400 meV energy range in a wave guide configuration. The sample contains 40 quantum dot planes inserted into the core of a GaAs/Al_0.9Ga_0.1As midinfrared waveguide. The light is injected with an infrared microscope coupled to the Fourier transform spectrometer. Polarization can be set either TE, in the layer plane, or TM along the growth axis. In this configuration, spectrally large infrared resonances are observed around 8-10 µm wavelength. In TM polarization, the absorption is maximum at 155 meV with a ~ 100 meV full width at half maximum. In TE polarization, the absorption is maximum at 200 meV with the same width. Considering the theoretical electronic structure, the TM resonance is attributed to the bound-to-continuum absorption from the s ground state to the delocalized state of the wetting layer continuum. The polarization is reasonable since the expected transfer of charge (i.e. the dipole) is mostly along the growth axis from the dot to the wetting layer. The TE resonance on the contrary is attributed to the transition towards the bulk GaAs barriers, at a slightly higher energy. The 45 meV difference between the TE and TM resonances corresponds roughly to the wetting layer subband depth and, as compared to the calculation, integrates the fact that the oscillator strength of a bound-to-continuum transition is maximum at higher energy than the continuum onset. The 100 meV broadening of the absorption is attributed to three factors : the dot size fluctuation, the bound-to-continuum nature of the transition, and a possible contribution of the d-states around 120 meV (in TM polarization).