GaAs quantum cascade lasers offer a potential route to mid-infrared emission near 16 µm, a wavelength relevant to sensing gases such as uranium hexafluoride (UF6) and BTEX compounds. At this wavelength, the choice of semiconductor substrate matters: InP, commonly used for quantum cascade lasers (QCLs), has a two-phonon resonance close to the energy of 16 µm light, creating a significant absorption concern. A preliminary GaAs/AlGaAs design instead explores how quantum-well engineering can support the target transition while addressing the trade-offs that affect gain and optical loss.
Why consider GaAs for 16 µm quantum cascade lasers?
QCLs produce light through transitions between quantized energy levels within a semiconductor’s conduction band. This differs from conventional diode lasers, which emit through transitions between the conduction and valence bands. Because the QCL transition energy is set largely by quantum-well structure rather than the bulk material bandgap, the emission wavelength can be tailored through layer design.
InP is widely used in QCL structures because InGaAs and AlInAs can be combined with it to form active regions and waveguides. But its two-phonon resonance is reported at about 626 cm-1, while 16 µm light corresponds to about 625 cm-1. This close match raises the likelihood of photon absorption in InP and makes it an unattractive substrate for this particular wavelength.
GaAs paired with AlGaAs is a possible alternative. The alloy system allows the band offset to be adjusted through composition, and GaAs is produced at industrial scale. However, the material pairing also presents design challenges: its smaller band offsets can make it harder to prevent electrons escaping into the continuum, and higher electron effective mass can shorten scattering lifetimes relative to the InGaAs/InP system. Active-region and waveguide decisions therefore need to account for both the wavelength target and these material limitations.
How the active region creates population inversion
A QCL active region must direct electrons through a sequence of energy states so that the upper laser level holds more carriers than the lower level. This population inversion enables optical gain. After an electron reaches the upper state, it can emit a photon as it transitions to a lower state. The lower state must then empty efficiently, allowing the cycle to continue.
In the design approach described here, longitudinal optical (LO) phonon scattering is used to help depopulate the lower laser state. The desired energy separation between the lower states is therefore set near the LO-phonon energy, while the upper-to-lower laser transition is set for 16 µm emission. This balance is central to achieving inversion: if the lower state empties too slowly, it can become populated and absorb light rather than support net gain.
Active-region geometries also affect the optical transition strength and carrier lifetimes. Vertical transitions have substantial spatial overlap between the upper- and lower-state wavefunctions, supporting a strong optical dipole transition. That overlap can also shorten the upper-state lifetime, however, so transition strength and carrier residence time must be considered together. The preliminary design selected a vertical-transition structure in part for its potential temperature resilience and relative ease of adjustment.
From individual wells to a biased three-well design
The design process built up the active region in stages. A single GaAs/AlGaAs quantum well was first adjusted to produce an energy separation of approximately 77.5 meV, corresponding to about 16 µm. Coupling a second well split the energy levels into pairs, but the resulting separation between the lower states was too small relative to the LO-phonon energy. Making the coupling barrier thinner could improve the alignment, but the required layer dimensions were judged impractical to fabricate in that design.
A three-well structure provided additional control over the relevant energy levels. The wells and barriers were adjusted to target both the optical transition and the lower-state phonon-assisted transition. Applying an electric bias then tilts the energy profile, enabling electrons to tunnel between states and across the injector region into the next active region.
The proposed active and injector period is 520.4 Å. Its reported differential-gain figure of merit is 1.32. The design exhibits carrier inversion and was assessed as a preliminary structure expected to lase, but the figure of merit is low. It should therefore be understood as a starting point for further optimization, not as evidence of a high-performance or experimentally validated device. The identified direction is to improve gain while reducing losses.
Waveguide choice is part of the design
Even a well-designed active region cannot produce net amplification if optical losses exceed gain. Important loss mechanisms include imperfect resonator reflections, absorption in doped semiconductor layers and metal contacts, and unwanted resonant intersubband absorption. Light must be confined in the device long enough to be amplified, while minimizing absorption along its path.
Dielectric waveguides work well in some InP-based QCLs because the active-region materials can have a higher refractive index than the InP cladding. GaAs-based structures present a different problem: GaAs has a higher refractive index than AlGaAs, so the same arrangement cannot simply use AlGaAs as a conventional cladding layer for total internal reflection. The design discussion therefore identifies a plasmon-enhanced dielectric waveguide as a more suitable approach to investigate.
In a surface-plasmon waveguide, light is confined near a metal–semiconductor boundary. The proposed approach is relevant to long wavelengths, although metal-related optical loss remains a consideration. Gold is identified as the preferred metal among those considered for wavelengths at or above 15 µm. The waveguide must also support the transverse-magnetic polarization associated with the QCL intersubband transition.
Practical implications for QCL development
The 16 µm GaAs/AlGaAs concept illustrates why long-wavelength QCL design is a system-level problem. Selecting a substrate to avoid a nearby phonon resonance is only one step. The active region must establish suitable energy spacing, carrier inversion and useful transition strength, while the injector, waveguide and resonator must transport carriers and confine light without overwhelming absorption.
For engineers evaluating long-wavelength QCLs, the preliminary design is most useful as a set of design constraints and trade-offs. Its 520.4 Å period and low 1.32 differential-gain figure of merit provide a starting point for refinement. Further work would need to improve gain and reduce optical losses while preserving the required transition and carrier dynamics. GaAs may address a key substrate limitation at 16 µm, but device performance depends on how effectively the complete layered structure solves the remaining challenges.