Thesis presented October 10, 2024

Abstract:
III-N semiconductors (AlN, GaN, InN) and their alloys are widely used in the production of light-emitting devices such as LEDs and lasers. One of the major challenges for the development of high efficiency light sources is the large polarisation fields (spontaneous and piezoelectric). Spontaneous polarisation results from the difference in electronegativity between nitrogen and the group III metal, and the absence of symmetry in the (0001) plane for these hexagonal materials. The piezoelectric fields are due to strain induced by lattice mismatch. The internal polarisation fields along the growth direction are generally perpendicular to the quantum well layer in heterostructures produced by growth [0001]. The polarisation mismatch results in an electric field that separates the electron and hole wave functions and redshifts the emission energy, known as the quantum confined stark effect (QCSE), leading to a decrease in the quantum efficiency of light emission. Various methods have been proposed and studied to reduce internal polarisation fields and improve LED efficiency, such as the use of alternative growth planes (semi-polar planes or non-polar planes) to reduce spontaneous polarisation, or doping such as the incorporation of free carriers into the QCSE by screening.
The characterisation of these polarisation fields is of main importance in order to improve the understanding and performance of the applications of luminous devices in order to control or reduce their effect. Electron holography appears to be a very promising tool because it allows direct measurement of the total electrostatic potential (Vtot), with high spatial resolution (a few nanometres) and great sensitivity. The total potential measured in holography comprises two main contributions: the mean internal potential (MIP) and the polarisation potentials (Vpol), which result from the internal electric field generated by spontaneous and piezoelectric polarisation, and which can be partially masked by the presence of free carriers (doping). We propose a method for separating these components in order to study the effects of polarisation in more detail. This method requires precise knowledge of the value of the MIP, which is typically calculated for bulk samples in the literature, far from the reality of heterostructures where the materials under a lot of strain. We propose here new DFT calculations that include the effects of strain, as well as experimental methods for measuring the MIP in bulk samples, as well as the difference in MIP between 2 materials in heterostructures, which highlight the effects of strain.
Finally, the effects of doping to screen polarisation fields are studied. Firstly, AlN/GaN samples in which the GaN layers are Ge-doped demonstrate potential screening due to doping. We were also able to demonstrate a potential inversion in AlN layers due to doping migration and segregation at the interfaces. This delta-doping was then explored as a means of controlling electrostatic fields locally. Using Si doping, we were able to locally reverse the electrostatic field in AlGaN/GaN structures. However, at equivalent or even higher concentrations, Ge doping proved difficult to control, and dopant migration was observed, preventing its use for fine control of electrostatic fields.

Keywords:
Electron holography, III/V materials, LEDs