Luminescence and Crystalline Properties of InGaN-based LED on Si Substrate with AlN/GaN Superlattice Structure

A crack-free indium gallium nitride (InGaN) based light emitting diode (LED) grown on silicon (Si) substrate was successfully demonstrated by introducing aluminium nitride/gallium nitride (AlN/GaN) superlattice structure (SLS) in the growth of the LED. The luminescence and the crystalline properties of the LED were discussed. From photoluminescence (PL) surface mapping measurement, the emission wavelength of the LED (453 nm) was almost uniform across the LED epi-wafer area. Temperaturedependent PL revealed that the dominant emission peak of the LED was 2.77 eV at all temperatures. The emission peak was related to the quantum wells of the LED. Some additional peaks were also observed, in particular at lower temperatures. These peaks were associated to alloy fluctuations in the In0.11Ga0.89N/ In0.02Ga0.98N multiquantum wells (MQWs) of the LED. Furthermore, the dependence of PL intensity and PL decay time on temperature revealed the evidence related to indium and/or interface fluctuations of the quantum wells. From X-ray diffraction (XRD) ω-scan measurements, fringes of the AlN/GaN SLS were clear, indicating the SLS were grown with good interface abruptness. However, the fringes for the MQWs were less uniform, indicating another evidence of the alloy fluctuations in the MQWs. XRD-reciprocal surface mapping (RSM) measurement showed that all epitaxial layers of the LED were grown coherently, and the LED was fully under strain.

Entropy-stable space–time DG schemes for non-conservative hyperbolic systems

We propose a space–time discontinuous Galerkin (DG) method to approximate multi-dimensional non-conservative hyperbolic systems. The scheme is based on a particular choice of interface fluctuations. The key difference with existing space–time DG methods lies in the fact that our scheme is formulated in entropy variables, allowing us to prove entropy stability for the method. Additional numerical stabilization in the form of streamline diffusion and shock-capturing terms are added. The resulting method is entropy stable, arbitrary high-order accurate, fully discrete, and able to handle complex domain geometries discretized with unstructured grids. We illustrate the method with representative numerical examples.