Successful design of mid-IR lasers

The W-quantum well antimonide VCSEL

With the freedom to model any arbitrarily layered quantum semiconductor heterostructure, we can systematically investigate the effects of adding or removing materials of different composition. The direct band gap materials from the roster of III-V or II-VI compound semiconductors can be used to form ternary or quaternary compounds, as for example GaInAs, or AlGaInSb, with different stoichiometry. Each layer can have different composition. This wide array of parameters allows us to explore the full range of parameter space for the formation of heterostructures. The remarkable feature is that essentially all of them can be grown today by molecular beam epitaxy.

The 3-5 mm region of the electromagnetic spectrum did not have useful sources of coherent radiation. This region of the spectrum is significant since it provides an atmospheric window; water vapor does not absorb in this region of the spectrum. It is also useful for chemical detection/analysis and for military countermeasures. In 1995, the Hughes Research Laboratory developed a multiple quantum well laser using well layers of InAs and barrier layers made up of GaInSb and AlGaSb. While this was successful, the wavefunctions of carriers clearly showed poor overlap. This implies a weak optical transition matrix element. This is because of the nature of the interfaces in the Antimonides. The energy bandgap lineups are such that the narrower bandgap material, InAs, has its bandgap energies not contained within the energy bandgap of the barriers. This is called a Type-II alignment, and leads to electrons and holes being confined to different layers. The figure shown below clearly displayed this feature when the particular structure was modeled using an 8-band finite element method developed at QSA.

Click on the icon to see a larger picture of the bandstructure.

The challenge to improve this structure was taken up by J. Meyer at the Naval Research Laboratory in Washington, DC. By inserting a thin layer of GaInSb in the modeling, an asymmetric quantum well was established in the model. While this led to a better overlap, the carriers had only a quasi-1D confinement afforded by the wider well.

Click on the icon to see a larger picture of the bandstructure.

When a more symmetric W-quantum well structure is created with the insertion of one more layer, the overlap is significantly improved, while the confinement is tighter. This is shown in the following figure. The very first sample of this structure grown by MBE was shown to lase at the predicted ~4 mm wavelength. This is a triump of wavefunction engineering. It cut short a trial and test approach and essentially gave the specifications for the performance of the laser. The laser continues to operate at room temperature, albeit with less efficiency.

Click on the icon to see a larger picture of the bandstructure.

The W-structure multiple quantum wells based on the antimonides has been incorporated into a Vertical Cavity Surface Emitting Laser (VCSEL). This fairly complex structure has been modeled by Meyer and Vurgaftman using the software developed by QSA and the experimental work has been done at the NRL.

Click on the icon to see a larger picture of the Vertical Cavity Surface Emitting Laser.

The above evolution of the successful design of a quantum well laser in a region of the spectrum not accessible earlier, is a classic example of wavefunction engineering. The interfaces, the materials and the geometry of the structure, the strain built into the layers, ...; all these define its quantum mechanical properties and hence its dielectric, optical, and transport properties. It is clear that computational tools are going to be more and more important. It is only through reliable simulations that we can narrow down the material and geometry (layer thichnesses, in this case) parameters for the particular functionality that we are seeking.

We are still exploring the physics and device applications of layered structures. When we consider what the future holds for quantum wire and quantum dot structures, it is clear that exciting times are ahead for theorists, experimentalists, and "quantum" engineers. Our role at QSA is to provide simulation and modeling tools that will bring novel devices to full commerical development, and shorten the cycle of design and growth. characterization and modeling, and finally integrating such devices into more conventional VLSI.

The Interband Quantum Cascade Lasers

In 1994, Capasso, Faist, Sirtori, Cho and coworkers at Lucent Technologies developed the Inter-subband Quantum Cascade Laser. In this periodic structure, electrons are injected at one end of a multi-quantum well structure, and undergo transitions within the conduction band between specific energy levels. This leads to the emission of radiation. the same electrons are then driven into the next multi-quantum well structure via quantum mechanical tunneling, through the application of an electric field. In the new multi-quantum well period, once again the process of relaxation through the emission of a photon recurs. The injected electrons thus undergo a cascading process and more than one photon is emitted per injected carrier. This laser is again in the ~5 mm region of the spectrum. An energy level schematic of the structure is shown below.

Click on the icon to see a larger picture of the bandstructure for Lucent Technologies' intersubband cascade laser.

Inspired by the developments at Lucent Technologies, two groups, one at the University of Houston and the other at the NRL, independently proposed a new mechanism for the cascading of carriers through a periodic multi-quantum well structure. In this case, the relaxation is through interband recombination in the narrow gap antimonides. The modeling was done using software developed at QSA. This mechanism has the advantage of substantially reduced Auger excitation losses in which no photons are emitted in the relaxation process. It also does not involve the degradation of the energy of carriers through phonon emission. Both are important issues. The following two structures are among the most complex structures modeled using the 8-band k.P finite element method. The software was developed at QSA, the modeling was done at NRL and at WPI, the growth was done at the University of Houston. Recent experiments have successfully demonstrated the laser action in these structures. The following figure illustrates one of the laser structures with Type-I interfaces.

Click on the icon to see a larger picture of the bandstructure of the Type-I Intersubband Quantum Cascade Laser.

The figure below shows an Interband Cascade Laser structure with type-II quantum wells.

Click on the icon to see a larger picture of the bandstructure of the Type-II Intersubband Quantum Cascade Laser.

These examples illustrate that modeling tools allow the concept of wavefunction engineering to have concrete outcomes. It is clear that further developments along these lines will demonstrate the feasibility of novel electonic mechanisms for exciting new applications to optoelectonic devices.

Return to Outcomes

Click on the QSA logo below
to return to the home page