Epitaxial growth optimization, measurement and theoretical analysis of strain-compensated QCL grown on (511)A InP

TL;DR

This work addresses how crystal orientation and interface roughness influence mid‑IR QCL performance by growing strain‑compensated InGaAs/AlInAs devices on the non‑standard $(511)A$ InP surface and systematically optimizing growth conditions using AFM/XRD. It demonstrates the first lasing on $(511)A$, but the devices show lower performance than the $(100)$ reference, partly due to impurity incorporation along the exposed direction, which also causes a ~7% redshift not explained by changes in conduction‑band offset or electron mass. The authors develop a general framework to evaluate CBO changes and effective mass under arbitrary growth directions via a Van de Walle‑based strain model and a Kane‑Hamiltonian treatment, confirming that these strain effects cannot account for the observed spectral shift. Overall, while non‑trivial growth directions may offer some control over interface roughness, the current results emphasize the challenges and the need for rigorous impurity management to realize practical gains. The work provides a systematic methodology for combining growth optimization, device characterization, and first‑principles‑informed band structure analyses in unconventional orientations.

Abstract

Interface roughness scattering is an important limiting factor for achieving high performance Quantum Cascade Lasers. Following recent results, we study the growth conditions for a strain-compensated QCL emitting around 4.6 μm grown on a (511)A InP substrate using AFM and XRD measurements. We find that modulating the arsenic flux and correctly tailoring the III/V ratio is fundamental to achieve a good quality material. We report the first lasing device on such a platform with a current density threshold of 1.34 kA/cm2 and a slope efficiency of 1.1 W/A, which result suboptimal compared to the (100) reference. Finally, we find a 7% redshift of the (511)A spectrum which we attribute to an impurity scattering due to the increased incorporation along the exposed (111) direction. We validate this statement by verifying that the change in CBO and effective electron mass due to strain along a non-trivial direction cannot cause such a shift by using the k-p method generalized to arbitrary growth directions

Figures (7)

• Figure 1: a) Schematic of the InP lattice oriented along the (100) and (511) direction. In both cases the monolayer planes are highlighted. b) Computed scattering rate for a transition energy of 270meV for the (100) and (511) case, as a function of the correlation length and assuming as $\Delta$ the monolayer thickness.
• Figure 1: AFM measurements of the surface roughness of 0.5$\mu$m bulk InGaAs grown on (100), (511) and (411) crystal direction. V/III flux ratios are reported as bare fluxes and normalized to the (100). RMS values of the surface roughness are also shown for 2 different integration areas.
• Figure 2: Correlation plots of surface roughness (nm) as a function of the V-III ratio of AlAs and InGaAs grown on (100) and (511) oriented InP substrates. The red line indicated the optimal growth windows.
• Figure 2: AFM measurements of the surface roughness of 0.5$\mu$m bulk AlInAs grown on (100), (511) and (411) crystal direction. V/III flux ratios are reported as bare fluxes and normalized to the (100). RMS values of the surface roughness are also shown for 2 different integration areas.
• Figure 3: Example of an XRD measurement performed on a AlAs/In$_{0.7}$Ga$_{0.3}$As superlattice grown along a (100) and (511) direction.