Interband Cascade Lasers Shed Their Ultra-cool Credentials
The US-led coalition s fight against terrorism and insurgents in Iraq and Afghanistan is aided by control of the skies. But this is under threat from shoulder-launched heat-seeking missiles that are made around the globe and are widely available.
These missiles can be prevented from hitting fighter jets by flares that divert them away from the aircraft s engine. However, this is not foolproof and the flares pose a risk of starting fires on the ground.
A more promising form of countermeasure is based on high-power mid-infrared lasers that can jam the guidance systems of heat-seeking missiles. Compact, low-cost sources needed for this form of defense are not available today, but excellent progress has been made through the development of novel interband cascade lasers (ICLs) by our team at the Naval Research Laboratory in Washington DC.
Improvements in infrared laser performance would also bring other benefits, like the detection of methane, ethane, hydrogen chloride, formaldehyde, hydrogen sulfide, nitrous oxide, and carbon monoxide and dioxide, which all have strong absorption lines between 3.3 and 4.6 µm. Methane is the primary component of natural gas, and a room-temperature laser with a narrow emission profile could provide the key component in a methane detector. Similar instruments could map out the distribution of this potent greenhouse gas in the Earth s atmosphere.
Making a suitable laser for gas sensing and infrared counter measures is not easy, however. Although solid-state parametric sources exist, their usefulness is limited by high cost. Lead salt IV–VI lasers also have their downsides, including a need for cryogenic cooling and output powers of less than 1 mW.
One alternative is III-V laser diodes, which are a multibillion-dollar business in the 0.8–1.6 µm spectral range. But it s very tricky to extend these lasers into the mid-infrared region of 3–5 µm. Carrier absorption losses scale as λ2 to λ3, and Auger non-radiative decay rates shoot up exponentially with increases in wavelength.
On top of this the GaSb-based material system that s needed for producing emission in this spectral region is relatively immature. Consequently, this type of laser has fallen short of what s needed for potential applications and the development of infrared sources has only come to the fore in the last decade.
Our recent success unites two different types of emitter: quantum cascade lasers (QCLs) and GaSb-based interband lasers. Neither of these can perform well in continuous-wave (CW) mode throughout the mid-infrared, but their marriage forms a hybrid design that can span the 3–4 µm range.
QCLs work best at 4.5–10 µm, although Manijeh Razeghi s group at Northwestern University, IL, has produced a 3.8 µm room-temperature laser. However, further improvements will be tough because there is a physical limit to the maximum barrier height that dictates the emission wavelength in a strain-balanced InGaAs/InAlAs/InP quantum well (QW).
GaSb-based interband lasers are now able to produce CW room-temperature lasing throughout the 2–3 µm band, thanks to highly strained InGaAsSb QWs. But extending this design much beyond 3 µm is problematic because the lasing wavelength is determined by the valence band offset that confines the holes in the active QWs.
Our ICLs, which are evolutions of Rui Yang s invention at the University of Toronto in 1995, produce lasing through interband optical transitions. This is the process found in other types of laser diode, but the crucial difference with an ICL is that it also employs multiple active stages cascaded in series (see box "How an ICL works").
Cascading means that several photons are emitted for each injected electron. This increases output power at the expense of an increased bias voltage, which is needed to activate all of the cascaded stages simultaneously. However, the positives outweigh the negatives because lower current densities for a given output power reduce the effects of parasitic ohmic and non-ohmic voltage drops.
We fabricate our ICLs by MBE growth on tellurium-doped GaSb substrates. Our design features an active region with 5–10 stages, which are clad with two n-type InAs/AlSb superlattices (figure 1).
Our first lasers, which we made in August 2005, shared two of the weaknesses of many early ICLs – high threshold current densities and a low temperature for CW output. However, the ICL community received a boost in 2006, thanks to the efforts of another team headed by Yang – by then working at the Jet Propulsion Laboratory, CA. The breakthrough was the fabrication of 12-stage, 3.3 µm ICLs with an operating temperature of 264 K. Close proximity to room-temperature operation is significant because it allows device cooling via a compact, energy-efficient thermoelectric cooler.
We followed this up with an ICL featuring five active stages that lased at up to 257 K. Using fewer stages failed to provide sufficient gain at high temperatures to overcome the structure s internal losses, and pulsed operation peaked at 295 K. However, we overcame this weakness with a new design that produced a 4.1 µm ICL operating at 288 K. Increases in high-temperature capability resulted from a doubling of the number of active stages to increase gain, improvements to the design of the band structure, lower Auger losses and a reduction in free carrier absorption.
Our latest ICL designs feature an additional GaSb QW for hole confinement, which reduces electron tunneling leakage that would bypass the holes. The dimensions of this well must be carefully chosen to ensure a low hole density in the GaSb layer because this minimizes intervalence absorption – one of the primary sources of optical loss.
We have optimized the electron free-carrier absorption in the optical cladding layers, which is governed by the doping density. If doping is too low, this can lead to excessive series resistance; if it s too high laser thresholds and external efficiencies suffer.
Auger recombination – a non-radiative process that involves the transfer of energy from electron-and-hole recombination to a third carrier – plagues the performances of all mid-infrared laser diodes and gets worse as you move farther into the infrared. We minimized this loss mechanism by optimizing the composition and thickness of the GaInSb QW that provides hole confinement, and we were able to produce 3–4 µm ICLs with room-temperature Auger coefficients of 3–4 × 10–28 cm6/s. Our low laser thresholds are a direct consequence of this low Auger coefficient – it is by far the lowest value for any III-V laser covering this wavelength range.
Our lasers have low threshold currents, such as 2.5 A/cm2 at 78 K. When pulsed at 300 K this rose to 360–400 A/cm2 (figure 2). Pulsed input power densities are lower than QCLs and W diodes (figure 3).
Recent efforts have focused on improvements in ICL efficiency, driven by reductions in internal optical losses. According to pulsed room-temperature slope efficiencies, our most recent device is twice as efficient as our best effort from 2006 (figure 4).
Reducing internal losses has a secondary benefit – less gain is needed, so we have been able to revert to cascading only five stages. We produced this type of ICL in March, which lased in pulsed mode at temperatures well beyond 300 K and had an efficiency of 160 mW/A. A big advantage of this design is that optical losses are actually lower than those for its 10-stage cousin, thanks to fewer net holes that contribute to unwanted intervalence processes.
To close in on the holy grail of ICLs that perform in CW at practical system temperatures, more attention was given to thermal management. Heat that must be dissipated to run these diodes is equivalent to the pulsed input power density (the product of threshold current density and applied bias) required for lasing (figure 3). Biasing the ICL is proportional to its number of stages, which makes our five-stage device the best for room-temperature CW lasing.
We have produced the first CW ICLs operating at room temperature by fabricating narrow-ridge waveguides into the five-stage design and adding a gold electroplated film for improved thermal dissipation when the laser is mounted epitaxial side up. Devices with 5, 9 and 11 µm ridges have delivered CW lasing up to 313, 319 and 317 K, and output powers of more than 10 mW at 300 K (figure 5).
These recent successes should pave the way for CW, room-temperature ICLs throughout the 3–4 µm band. Nothing should obstruct this advance because the threshold current densities are broadly similar across this spectral range. There is also good reason to believe that further improvements in device performance are possible because ICLs are still a relatively immature laser technology.
Our efforts are shifting to the development of packaged devices for real-world applications. We are aiming to equip products with robust, spectrally pure beams by incorporating features such as gratings based on distributed feedback or photonic crystals. Spectroscopy applications can be targeted with this type of design, but we will need to deliver far higher output powers for missile jamming. This could be realized through further reductions in the internal loss at high temperatures, optimization of narrow-ridge fabrication and better heat sinking.
T Hosoda et al. 2008 Appl. Phys. Lett. 92 091106.
K Mansour et al. 2006 Electron. Lett. 42 1034.
J R Meyer et al. 1998 Appl. Phys. Lett. 73 2857.
I Vurgaftman et al. 2001 Phil. Trans. R. Soc. London A 359 489.
R Q Yang 1995 Superlatt. Microstruct. 17 77.
J S Yu et al. 2006 Appl. Phys. Lett. 88 251118.
View pdf of article