This work presents the growth of high-quality Ge epilayers on Si (001) substrates using a reduced pressure chemical vapor deposition (RPCVD) chamber. Based on the initial nucleation, a low temperature high temperature (LT-HT) two-step approach, we systematically investigate the nucleation time and surface topography, influence of a LT-Ge buffer layer thickness, a HT-Ge growth temperature, layer thickness, and high temperature thermal treatment on the morphological and crystalline quality of the Ge epilayers. It is also a unique study in the initial growth of Ge epitaxy; the start point of the experiments includes Stranski–Krastanov mode in which the Ge wet layer is initially formed and later the growth is developed to form nuclides. Afterwards, a two-dimensional Ge layer is formed from the coalescing of the nuclides. The evolution of the strain from the beginning stage of the growth up to the full Ge layer has been investigated. Material characterization results show that Ge epilayer with 400 nm LT-Ge buffer layer features at least the root mean square (RMS) value and it’s threading dislocation density (TDD) decreases by a factor of 2. In view of the 400 nm LT-Ge buffer layer, the 1000 nm Ge epilayer with HT-Ge growth temperature of 650 °C showed the best material quality, which is conducive to the merging of the crystals into a connected structure eventually forming a continuous and two-dimensional film. After increasing the thickness of Ge layer from 900 nm to 2000 nm, Ge surface roughness decreased first and then increased slowly (the RMS value for 1400 nm Ge layer was 0.81 nm). Finally, a high-temperature annealing process was carried out and high-quality Ge layer was obtained (TDD=2.78 × 107 cm−2). In addition, room temperature strong photoluminescence (PL) peak intensity and narrow full width at half maximum (11 meV) spectra further confirm the high crystalline quality of the Ge layer manufactured by this optimized process. This work highlights the inducing, increasing, and relaxing of the strain in the Ge buffer and the signature of the defect formation.
Generating High-Quality Random Numbers by Next Nearest-Neighbor Cellular Automata