Recent experiments probing the conduction band electronic states in self-assembled dots, double dots and ring-like structures are reviewed. The results are discussed in view of contributions from quantization, magnetic forces and Coulomb interaction.
For intraband spectroscopic studies of fully quantized semiconductor structures, requirements have to be met which seem to be mutually exclusive: Whereas the investigated structures have to be in the nm-size range, the required sample sizes need to be large (of the order of mm2), to meet the required signal-to-noise ratio at the relevant wavelengths. The Stranski-Krastanow growth model-4 offers a convenient solution to the challenge of fabricating arrays of typically 108 structures with a sufficiently narrow size distribution. Even though this growth process can create the required nanostructure arrays with relative ease, its "selfordering" nature seems to limit the control over the desired shape of the individual structures. This review summarizes intraband spectroscopic studies on a variety of self-assembled InAs/GaAs nanostructures, ranging from individual quantum dots to coupled dots and to nanoscopic ring structures. Results from two spectroscopic techniques are discussed, namely capacitance spectroscopy, which probes the many-particle ground states of the electronic system, and far-infrared transmission spectroscopy, which probes its excitations. Making use of a special sample geometry, these complementary techniques can be performed simultaneously on a single sample, thus giving deep insight into the samples electronic structure.
The sample geometry is depicted in Fig. 1 for the case of quantum dots. The dots are embedded in a field-effect-transistor type structure between a semitransparent top gate and a metallic, highly doped GaAs back contact layer. The dot layer is separated from the back contact layer by an undoped GaAs layer of thickness tl, which is small enough to allow for tunneling of electrons from the back contact layer into the dots. On the contrary, the distance t2 between the dots and the gate is much larger and comprises a Al/GaAs superlattice as an additional barrier, so that no electron transfer is possible. When a voltage Vg is applied, the potential will drop to a good approximation linearly between the gate and the back contact, so that the induced energy shift at the location of the dots is given by AE = eAVg tl / (tl + t2). For the given low dot densities (typically 1010 cm-z) and small electron numbers (ne = l...S), this linearity is only little affected by charge in the dot layer5 and gives a convenient energy scale.