Substrate engineering of 1.55 μm lasers

Lasers grown on InP suffer from a lot of disadvantages compared to those grown on GaAs: the price of the substrate, the high temperature sensitivity of the 1.55 μm lasers and the limitations imposed by the miscibility gap of the InGaAsP system on the design of the active regions. We propose the use of substrate engineering (in this case the growth on In_1-xGa_xP) to solve these problems. It was recently demonstrated that a thin crystal layer bonded on a bulk crystal of the same nature with a high enough angular misalignment can be used as a compliant universal substrate. When the screw dislocation array at the interface is dense enough (misalignment angle higher than about 15°) the stress introduced by an epitaxial layer on the thin film can be released at that highly dislocated interface. This concept makes it possible to create new substrates when no ideal crystal is available. In the case of the 1.55 μm lasers it is now possible to imagine a new substrate to get rid of many of the problems of the InP substrate. The structure we propose consists of growing the 1.55 μm active regions on an In_1-xGa_xP substrate created on a compliant universal substrate with x equal to about 0.1. For x=0.1, the band gap energy of In_1-xGa_xP is 1.43 eV, i.e. 80 meV higher than that of InP. Most of this energy difference increases the energy of the conduction band edge. Moreover we demonstrate that it is possible to grow strained compensated quantum wells (QW) active regions emitting at 1.55 μm on that type of substrate using only InGaAs alloys