The effect of strain on the interdiffusion in InGaAs/GaAs quantum wells

Abstract

The effect of strain on the cation interdiffusion in InGaAs/GaAs quantum wells is described. It is found that the Fick’s diffusion equation does not properly describe the interdiffusion in the heterostructure with strained layers. It is believed that the strain changes crystal defect concentration and thus diffusivity is influenced by strain. Diffusion equation including the strain effect is formulated and solved numerically. The experimental photoluminescence peak shifts as a function of annealing time are well-fitted by this analysis and useful parameters such as diffusivity describing InGaAs/GaAs quantum well interdiffusion are extracted.

Effect of strain on the interdiffusion of InGaAs/GaAs heterostructures

“…It should be noted that in this case there is no error introduced into the determination of the diffusion coefficient. This is exactly the behavior observed in the data of Gillin et al 3 which Ryu et al 7 reanalyzed. If, on the other hand, a number of samples are used such that one sample receives a single anneal of 1 h, while another sample receives a single anneal of 10 h, then the samples which have the shorter anneal times see a bigger effect of the ramp phase.…”

“…The encapsulant used in these experiments has a higher refractive index and is known to contain less oxygen than those used previously, it is this change, we believe, that has reduced the surface effect to an immeasurable level. Given that these experiments show that there is no systematic variation in the diffusion coefficient with the strain in the quantum well, strains vary from 0.36% for the 5% indium well to 1.4% for the 20% indium well, it is necessary to address the analysis of Ryu et al 7 which appear to show a strain effect in this material. In their work, Ryu et al 7 grew a sample with two quantum wells, widths 60 and 120 Å, of In 0.2 Ga 0.8 As in GaAs barriers and annealed it in a furnace for times between 1.5 and 10 h, it appears from this article that each anneal was performed on a separate sample although this is not explicitly stated.…”

“…Given that these experiments show that there is no systematic variation in the diffusion coefficient with the strain in the quantum well, strains vary from 0.36% for the 5% indium well to 1.4% for the 20% indium well, it is necessary to address the analysis of Ryu et al 7 which appear to show a strain effect in this material. In their work, Ryu et al 7 grew a sample with two quantum wells, widths 60 and 120 Å, of In 0.2 Ga 0.8 As in GaAs barriers and annealed it in a furnace for times between 1.5 and 10 h, it appears from this article that each anneal was performed on a separate sample although this is not explicitly stated. Using the photoluminescence from the quantum wells and a standard Fick's law approach, identical to that used here, they produce data which show an apparent increase in diffusivity for short anneal times which reduces as annealing progresses.…”

“…For the III-V semiconductors, Gillin and co-workers 3-5 studied the InGaAs/GaAs system in detail and found no detectable effect of strain and in 1994 published an analysis showing that strain should not affect diffusion as the elastic strain energy is too small. 6 However, the following year Ryu et al 7 showed results for InGaAs/GaAs which suggested that strain does affect the diffusion in this system and reanalyzed some of the data of Gillin et al 3 to show that it was consistent with their model.…”

“…Finally, we need to address the mechanism through which Ryu et al 7 proposed that the strain enhanced the diffusion, changes in the vacancy creation rate due to strain. This theory is based on the work of Antonelli and Bernholc 9 who calculated that in Si:Ge there is a change in the vacancy energy in the presence of hydrostatic strain, however, Antonelli and Bernholc 9 also calculate the vacancy energy under biaxial strain, such as found in these samples, and they found that there was no change.…”