MOVPE Growth of LWIR AlInAs/GaInAs/InP Quantum Cascade Lasers: Impact of Growth and Material Quality on Laser Performance

December 1, 2022

Massachusetts Institute of Technology, Lexington, Massachusetts

(a) Schematic of conduction band energy diagram and wavefunctions of a QCL structure. One QCL period consists of the injector and active region, and the period is typically repeated 30-40 times in a full QCL structure. (b) Cross-section of buried heterostructure QCL (right) and transmission electron microscopy cross-section of a portion of a QCL core.

Quantum cascade lasers (QCLs) are compact coherent optical sources that emit over a wide wavelength range in the mid- to long-infrared (3-25 μm) as well as into part of the terahertz spectrum. With recent developments of AlInAs/GaInAs/InP QCLs exhibiting watt-class output power levels at room temperature in the mid-wave infrared (MWIR, 3-7 μm) and long-wave infrared (LWIR, 8-12 μm) regions, QCLs have become increasingly attractive for numerous technological applications including infrared countermeasures, free-space communications, and chemical and biological sensing.

As interest in QCLs continues to grow, so does the desire to improve performance and understand factors that may ultimately limit these unique and complex devices. QCLs are unipolar devices based on tunneling and intersubband transitions between quantum-confined energy states in the conduction band of a coupled quantum-well structure. These structures are designed using band structure engineering to optimize optical transitions and electron transport for laser characteristics such as wavelength, threshold, power, efficiency, and high-temperature operation.

A typical QCL structure consists of a complex sequence of barrier and quantum well layers, totaling ~600-1000 layers, with thickness ranging between 0.6 to 6 nm. With the requirement of so many ultra-thin layers being reproducibly grown over microns of thickness, it is not surprising that while the first proposal to use intersubband transitions for radiation amplification was proposed in 1971, it was over 20 years before QCLs operating at cryogenic temperatures were first demonstrated in 1994, and another eight years for room temperature continuous-wave (cw) operation in 2002.

Exacting epitaxial growth of QCL structures goes hand-in-hand with optimization of its band-structure and wavefunction modeling, advanced processing involving fabrication and epitaxial regrowth of high-aspect ratio devices, and demanding heat-sinking packaging.

Impressive progress has been made in each of these areas and QCLs with improved operating temperature, wavelength range, output power, and efficiency are routinely possible. Record performance at room temperature is 5W cw output power and 21 percent wall plug efficiency (WPE) in the MWIR and 2W cw power and 10 percent WPE in the LWIR. Those QCLs were grown by molecular beam epitaxy (MBE) or gassource MBE, which are both high-vacuum growth processes.

Another viable growth technique for QCLs is metalorganic vapor phase epitaxy (MOVPE), which operates at or slightly below atmospheric pressure. It is the mainstream platform for more conventional p-n optoelectronic devices and was shown to be also suitable for QCL growth in the early 2000s. Comparable performance between MOVPE and MBE-grown QCLs was demonstrated in 2006, although the QCL structures were different. With these encouraging results, numerous groups have pursued MOVPE for development of QCLs, and have achieved a high level of success.

Despite these accomplishments, the epitaxial growth of QCL structures continues to be a challenge. In particular, the emission wavelength from QCLs of the same structure can not only be different when grown by MBE or MOVPE, but also different when grown by MOVPE at different organizations. For example, even though the same QCL structure was used, MBE- and MOVPE-grown QCLs emitted at different wavelengths. MBE-grown QCLs had emission at 9.3 μm, while MOVPE-grown QCLs from different organizations were reported with emission at 8.4 μm and at 9.2 μm. QCLs of that same structure grown elsewhere emitted around 10 μm. Thus, it is critical to gain additional knowledge of the relationships between epitaxial growth, their materials properties, QCL band structure modeling, and resulting QCL performance in order for MOVPE to be a more predictable growth process.

This work was performed by Christine A. Wang, Benedikt Schwarz, Dominic F. Siriani, Leo J. Missaggia, Michael. K. Connors, T. S. Mansuripur, Daniel R. Calawa, Daniel McNulty, M. Nickerson, J.P. Donnelly, and F. Capasso for the Massachusetts Institute of Technology. For more information, download the Technical Support Package (free white paper) below.