Growth kinetic aspects of dislocation evolution in SiC single crystals

• Applicants: Professor Dr. Stefan Sandfeld; Professor Dr.-Ing. Peter Wellmann
• Subject Area: Synthesis and Properties of Functional Materials
• Term: since 2018
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 394148498

The goal of this project is to significantly improve today's dislocation models for hexagonal semiconductors through conceptual derivation and computational implementation of models for dislocation flow and interaction based on experimental input and validation. Our predictive computational model will ultimately allow us to optimize process parameters (cooling and growth) and drastically reduce the number of dislocations in the crystal. Silicon carbide - which has a hexagonal crystal structure - is chosen as a model system for non-cubic materials; however, the knowledge gained will be representative and easily transferable to other hexagonal semiconductors such as GaN, AlN, ZnSe and ZnO. Our quantitative experimental characterization provides data on internal stresses and strains as well as details of the dislocation microstructure due to dislocation glide and dislocation pairing (in SiC at 2100 to 2300 C). The methods used are Raman spectroscopy, KOH defect etching, X-ray diffraction and, most importantly, synchrotron X-ray topography. During the first phase of the project, in conjunction with experimental work (crystal growth and characterization) and theoretical work (exploration of novel plasticity models for semiconductors), the local distribution of dislocations was studied for different crystal growth and thermal treatment boundary conditions. As a major result, it was shown that the newly developed continuum dislocation dynamics (CDD) simulation with source activation in 2D and 3D accurately describes the local dislocation density distribution very well. The CDD simulation outperforms the classical Alexander-Haasen model (without consideration of dislocation flow), which lacks a realistic prediction of the local dislocation arrangement. The follow-up proposal investigates two new phenomena for dislocation dynamics in semiconductors, i.e. the nucleation of dislocation pairs caused by growth perturbations and the influence of growth kinetic aspects in terms of correlation of surface step structure with dislocation propagation behavior on off-axis crystal surfaces. This experimental and simulation work is accompanied by machine learning and deep learning based analysis of the microscopy data

• DFG Programme: Research Grants