Dielectric cap quantum well disordering for band gap tuning of InGaAs/InGaAsP quantum well structure using various combinations of semiconductor-dielectric capping layers

Abstract

The spatially selective band gap tuning of quantum well structure has been an essential tool to realize the integration of optoelectronic or photonic devices on a single wafer [1, 2]. Quantum well disordering (QWD) techniques are widely employed for the band gap tuning, which induces a change of shape of the quantum well profile by intermixing the well and the barrier materials during annealing. Among the QWD techniques Dielectric cap QWD (DCQWD) method uses the vacancies created both in the dielectric capping layer and at the dielectric-semiconductor interface. There have been many studies of dielectric capping layers, such as SiO2[3], SiNx [4, 5], SrF2[6], and WNx [7], to control the QWD. In this study, DCQWD method was employed for band gap tuning. We were concerned about the effect of the various semiconductor-dielectric capping layer combinations on the band gap tuning as a function of the annealing temperature. The vertical structure of the samples used in this study is schematically shown in Fig. 1. We employed chemical beam epitaxy (CBE) to grow the undoped In0.53Ga0.47As/InGaAsP (Q1.25) single quantum well and the subsequent semiconductor capping layers. The CBE growth temperature was 500 ◦C. The single quantum wells (In0.53Ga0.47As/InGaAsP (Q1.25)) capped with different semiconductor capping layers (InP, In0.53Ga0.47As and InGaAsP (Q1.25)) with 50 nm-thick were prepared, and each was then subjected to topmost capping with dielectric material. The dielectric layers of SiO2 and SiNx (300 nm thick and 100 nm thick, respectively) were deposited by plasma enhanced chemical vapor deposition technique. For SiNx capping layers, dilute Silane (5% SiH4 in N2, flow rate of 40 sccm) and NH3 (99.999%, flow rate of 25 sccm) were used as reactant gases, while Silane (5% SiH4 in N2, flow rate of 40 sccm) and N2O (99.999%, flow rate of 40 sccm) were used for SiO2 layer. The process pressure and temperature were maintained at 0.9 torr and 200 ◦C, respectively. The RF power was 30 W. The growth time was 6 min 40 s for SiNx , and 3 min for SiO2. The refractive index of SiNx film was 1.91. The band gap of virgin quantum well prior to capping was 0.8 eV (λPL = 1550 nm) at room temperature. Thermal treatment was accomplished in N2 atmosphere at various annealing temperatures (600 ◦C– 800 ◦C) for 8 min. PL measurement was carried out at room temperature to identify the degree of band gap tuning. Fig. 2 shows the energy shifts achieved by the various annealing temperatures. The maximum energy shift occurred at 800 ◦C for all samples studied in this work. The energy shift generally increased with the increase of the annealing temperature, which has been reported previously [8]. Fig. 3 shows the blue shift difference between SiNx capped samples and SiO2 capped samples. In cases of InP and InGaAsP semiconductor capping layers, the difference amounted to about 135 meV after annealing at 800 ◦C, while the difference for InGaAs reached about 80 meV. As a result of interdiffusion of atoms between quantum well and barrier, a change in band gap width of the quantum well may happen when the structure is heated above a specific temperature. In our work, the threshold temperature was around 750 ◦C. The overall experimental results of this study, which can be seen in Fig. 2, is that SiNx capping layer caused larger blue shifts than SiO2 capping layer, which means that SiNx layer has many more vacancies than SiO2 layer. However, there was another report for band gap tuning of InGaAsP/InP