Technical Insight
Research Review: QCLs with internal frequency doubling target gas sensing
Intra-cavity frequency doubling creates 3 mm QCLs from the pairing of GaInAs and AlInAs
A US TEAM has substantially improved the output power of its short wavelength of GaInAs/AlInAs quantum cascade lasers (QCLs) that feature a frequency-doubling region.
The QCL made by engineers from The University of Texas at Austin and Adtech Optics delivers room temperature emission at 2.95 μm, which is within the ideal wavelength range for chemical sensing and spectroscopy applications. That’s because the spectral range 2.5-3.5 μm contains several important molecular absorption lines, including those for methane and ammonia.
Antimonide materials can also be used to produce diode lasers, QCLs and interband cascade lasers. However, these types of device are tricky to grow and process, according to corresponding author Mikhail Belkin from The University of Texas at Austin.
His view is that a far more cost-effective platform for making mid-infrared lasers for spectroscopy is the more mature InGaAs and AlInAs materials system that is also used to make telecom lasers. Conventional QCLs made from these materials can operate in continuouswave (CW) mode at room temperature and span the 3.5-12 μm range. To reach the shorter wavelengths, strain compensated superlattices are used to make deeper wells and higher barriers.
However, with this conventional design it is very challenging to produce QCLs emitting below 3.5 μm. “Although the gamma points in InGaAs and AlInAs have high energy offsets, electrons may scatter from a gamma valley in InGaAs to an Xvalley in AlInAs,” explains Belkin, who adds that at certain aluminium compositions, the bangap of AlInAs is indirect.
To extend GaInAs/AlInAs QCLs to 3 μm, the US team inserted a non-linear layer on top of the active region of the device that is tailored to have resonant optical nonlinearity for second harmonic generation.
First-generation devices that were reported by the Belking group in 2010 in collaboration with teams from Princeton University and the University of Maryland delivered peak output powers below 10 μW at 3.6 μm. This has now been increased to 35 μW at 2.95 μm, using a 10 μm-wide, 3 mm-long ridge laser that was driven with 50 ns pulses at a 100 kHz repetition rate. The latest device – which features a non-linear layer containing 28 repetitions of a multi-quantum well structure made from the pairing of In0.67Ga0.33As and Al0.57Ga0.43As – produced a room-temperature threshold current density of just 2.2 kA/cm2, a low enough value to suggest that this design is capable of yielding CW QCLs.
The increase in power of these novels lasers stemmed from improvements in material quality. Better materials were possible thanks to the collaboration of the university team with Adtech Optics, a commercial foundry that produces QCLs. “We also used strain compensated structures that allowed for more design flexibility in the non-linear layer,” adds Belkin.
Although this partnership between academia and industry led to success, there is still plenty of room for further improvement. “We had to overgrow our non-linear layer onto the Adtech Optics’ existing QCL structure, and as a result, the active region and waveguide configuration was not quite optimal for the second harmonic generation device,” explains Belkin.
The upshot of this lack of optimisation was a poor degree of overlap of the pump modes with the nonlinear section. Belkin admits that the team also made a slight mistake in its calculations of the bandstructure in the non-linear section of the device. This prevents a perfect double resonance from occurring in the QCL.
Once the team addresses these issues, it should be capable of fabricating lasers with output powers of at least 1 mW. “We then plan to process these lasers as buried heterostructure devices to demonstrate CW operation,” says Belkin. If successful, the researchers will have made devices meeting the typical output power requirements for chemical sensing and spectroscopy applications.
Another goal is to demonstrate externalcavity tuning of their devices, which will highlight their promise for spectroscopic applications. “Finally, we are also collaborating with Markus-Christian Amann’s group on designing secondharmonic generation lasers that can be tuneable over 0.5-1 μm in the 2.5-4 μm spectral range,” says Belkin. “This will require additional optimisation of the non linear section and waveguide structure that will support efficient quasi-phasematched second harmonic generation for different pump wavelengths.”
M. Jang et. al. Electron. Lett. 47 667 (2011)
