• s.SNOM
• near-field
• ISB transitions
• FTIR

The manipulation of electronic states in artificial crystal structures is of primary importance for the development of many photonic technologies, such as lasers, LEDs and photodetectors. For the Mid-IR and THz spectral regions the transitions between the quantized levels of the conduction band confined in nano-scale quantum wells (QW) is of particular interest.

The QW structure consists of a heterogeneous semiconductor in which charge carriers behave like a 2D electron (or hole) gas with quantized states in the direction orthogonal to the QW plane. Transitions between these confined states are called intersubband transitions (ISB) and can be tuned by changing the QW thickness. The advancement in fabrication techniques, such as molecular beam epitaxy, permits the fabrication of structures with atomic precision, thus allowing the realization of QW structures with suitably tailored properties. Thanks to such precise control of the QW thickness, ISB transitions are now at the heart of mid-IR and THz technologies like quantum well infrared photodetectors (QWIPs) and quantum cascade lasers (QCLs).

Typically, ISB transitions are characterized with the so-called Fourier Transform Infrared (FTIR) Spectroscopy, which is a far-field interferometric method. FTIR has indeed emerged as a powerful spectroscopic tool for investigating intersubband transitions of many semiconductors. This technique provides a complete characterization of the optical properties of the sample, but since it involves the far-field, it also shows some drawbacks. First, ISB transitions can only couple with light polarized in the out-of-plane direction. It is therefore impossible to probe such transitions by shining light perpendicular to the quantum well; specific geometries must then be designed to drive the beam with a non-zero incident angle, usually through a prism shaped sample. In addition, the spatial resolution of the FTIR method is limited by the large typical radius of hundreds of microns of the collimated light beam. This makes FTIR unsuitable for applications where the examination of nano-objects is required. As an example, the peak absorption of a QW measured with FTIR is usually broadened by the average fluctuations in the thickness of the well, which varies on a nanometric scale.

Recently, another measurement technique has been developed, exploiting the properties of near-field interaction: scattering-Scanning Near Field Optical Spectroscopy (s-SNOM). This technique consists of an illuminated AFM tip which provides nanoscale resolution, well below the diffraction limit. Thanks to the near field of the tip, which provides an out-of-plane component of the electric field independently of the direction of the impinging light, s-SNOM can excite ISB transitions and in effect it was recently employed to measure ISB transitions in a few layer exfoliated 2D material.

Due to the evanescent nature of the near field, the s-SNOM technique can probe only the superficial region of the sample, down to few tens of nanometres in depth. In order to measure the quantum well ISB absorption with the s-SNOM, the quantum well must therefore be placed as close to the surface as possible.

Bloch's theorem describes what happens in the bulk of semiconductor crystals, where the translational symmetry is exploited to simplify the calculation of energy levels. However, when dealing with real crystals, surface effects must be taken into account. Common III-V semiconductors like GaAs/AlGaAs show charge depletion at the surface: it is a region of few tens of nanometres where the charge carriers are pushed away from the surface. For this reason, with this kind of semiconductors it is impossible to build surface nanostructures like nanodots, nanowires or shallow quantum wells . On the other hand InAs, shows opposite behaviour: charges accumulate at the surface, thus creating a thin layer of charges.
The accumulation of charges can be helpful for the operation of structures like nanowires. However, it is detrimental for the confinement of charges in shallow QW. Moreover, a high charge density at the surface might shield the near field of the s-SNOM tip, thus preventing its interaction with deeper objects like the QW 2DEG.

The goal of this thesis is to characterize ISB transitions in two different shallow single QWs of InAs/AlSb, in which a cap of InAlSb is able to suppress surface effects. This is designed to have the bottom of the conduction band aligned with the surface levels so that Fermi level pinning does not cause major band-bending.
Firstly we perform FTIR measurements, thus observing the presence of ISB transitions in QWs at only 6 nm in depth from the surface. We then designed and measured a particular geometry that allows for carrier density modulation inside the QW. The modulation of the carrier density will be employed for more complex s-SNOM measurements. Preliminary s-SNOM measurements are finally discussed.

This thesis is organized as follow: in the first chapter, we give a background to our work, introducing motivations and scopes of this thesis in relation with the latest literature. In the second chapter, we overview ISB transitions theory, highlighting the main differences between far-field and near-field interactions with the heterostructure. In the third chapter, we introduce a model we customized to allow for numerical simulations of the tip-sample interaction. The effects of the s-SNOM tip momentum distribution on the absorption of the QW are explored as function of the many parameters involved in the experiment. The fourth chapter presents the experimental methods and techniques, with an introduction to the FTIR method and the preparation of the samples for FTIR measurements. We also introduce the working principle of the s-SNOM, focusing on the tip momentum distribution. At the end of the chapter, we present the doping modulation technique we employ to isolate the ISB contribution to the absorption from different mechanisms, such as free-carrier absorption or background absorption. We then show our experimental findings in the fifth chapter and the issues remaining to be address with the described sample designs. Finally, we summarize the results and discuss the future outlook of this work.