Technical Insight
Powerful QCLs stretch to shorter wavelengths
Strain-balanced QCLs deliver continuous-wave output powers in excess of 500 mW at wavelengths as short as 3.4 µm
Engineers at Northwestern University, Illinois, have extended the spectral coverage of powerful quantum cascade lasers (QCLs) to shorter wavelengths.
Prior to their efforts, the shortest wavelength continuous-wave QCL, which was reported by Corning in late 2011, delivered a room-temperature output of 40 mW at 3.55 µm. This bar has now been raised with a pair of highly strain-balanced, InP-based AlInAs/GaInAs QCLs that emit 504 mW at 3.39 µm and 576 mW at 3.56 µm when driven in continuous-wave (CW) mode at room-temperature.
These powerful emitters can be used for remote sensing and jamming. In the former application, QCLs can unveil the presence of many hydrocarbons, which have strong absorption in the 3.3 µm to 3.6 µm range due to stretching modes of the carbon-hydrogen bonds. These organic materials include nerve agents such as cyclosarin, soman and tabun gas, and some explosives.
“For jamming applications, the 3.0 – 3.5 µm range is used by missile seekers to look for hot metal objects,” explains Manijeh Razeghi, head of the group at Northwestern University. “A laser in this wavelength range can be used to ‘dazzle’ the missile seeker so that it cannot acquire a target.”
Manijeh Razeghi’s group produces powerful QCLs on InP.
Another class of laser that emits around 3 µm is the interband cascade laser. However, the performance of this class of device tends to vary far more with changes in temperature than a QCL. Historically, shorter wavelength QCLs have lagged behind their longer wavelength cousins in temperature insensitivity, but lasers from Northwestern have significantly closed this gap.
“We believe our success is partly due to the use of better electron confinement and a three-well active region,” reveals Razeghi.
The QCL community has made pulsed-mode QCLs operating at room-temperature from three different types of material system: InAs/AlSb/InAs, GaInAs/AlAs(Sb)/InP and GaInAs/AlInAs/InP. Razeghi and her co-workers use latter of these because its yields the best material quality, and it is best suited to commercial device manufacture. “A large infrastructure exists that is dedicated to the growth of antimony-free InP-based compounds,” says Razeghi, who adds that antimony-containing systems would require significant ‘retooling’ and process development. The QCLs that she and her co-workers fabricate span an emission range from 3 µm to the terahertz range.
When GaInAs/AlInAs structures are lattice-matched to InP, the quantum-well depth is just 505-520 meV, which in insufficient for electron confinement for making QCLs operate below 6 µm. It is possible to increase the depth of the well a great deal by switching from GaInAs to InAs and replacing AlInAs with AlAs. However, this alternative material system is not lattice-matched to InP, preventing the growth of high-quality films that are a pre-requisite for making devices.
“To grow a device with a deep quantum well, good interfaces and without dislocations, the compressive strain of the InAs-rich material must be balanced – on the scale of one cascade stage – with the tensile strain of the AlAs-rich layer,” explains Razeghi. Those strain-balanced layers have a depth of about 1 eV. “This allows us to reach much shorter emission wavelengths with good electron confinement.”
Using this approach led to the construction of a 3.39 µm, double-channel ridge-waveguide QCL with an 8.6 µm width and 5 mm laser cavity. Driven with 500 ns pulses at a 100 kHz repetition rate, this laser produces 1.1 W at room temperature. Switch to CW operation and output falls to 504 mW, 403 mW and 88 mW at 15 °C, 25 °C and 55 °C, respectively. The QCL’s 3.56 µm cousin has a 10.5 µm-wide ridge, and produces 1.66 W when driven in pulsed mode. In CW operation, output falls to 576 mW, 437 mW and 45 mW at 15 °C, 25 °C and 55 °C, respectively.
QCL performance can be improved with a buried heterostructure geometry that trims ridge widths and optical losses, while increasing thermal conductance. Performance then improves on several fronts: Threshold current falls; slope efficiency rises, resulting in higher output power; and internal temperature increases are smaller at all drive currents. The two major benefits that result from all these gains are an increase in the power saturation current density and a threshold current that is low, even at high temperatures. “This leads naturally to a higher operation temperature for CW operation,” remarks Razeghi.
Along with her co-workers, she is trying to improve material homogeneity, which will lead to better QCL performance. “We are systematically trying to identify the source of inhomogeneity. This includes investigations of reactor hardware limitations, growth conditions and interface effects.”
N. Bandyopadhyay et al. Appl. Phys. Lett. 100 212104 (2012)
Prior to their efforts, the shortest wavelength continuous-wave QCL, which was reported by Corning in late 2011, delivered a room-temperature output of 40 mW at 3.55 µm. This bar has now been raised with a pair of highly strain-balanced, InP-based AlInAs/GaInAs QCLs that emit 504 mW at 3.39 µm and 576 mW at 3.56 µm when driven in continuous-wave (CW) mode at room-temperature.
These powerful emitters can be used for remote sensing and jamming. In the former application, QCLs can unveil the presence of many hydrocarbons, which have strong absorption in the 3.3 µm to 3.6 µm range due to stretching modes of the carbon-hydrogen bonds. These organic materials include nerve agents such as cyclosarin, soman and tabun gas, and some explosives.
“For jamming applications, the 3.0 – 3.5 µm range is used by missile seekers to look for hot metal objects,” explains Manijeh Razeghi, head of the group at Northwestern University. “A laser in this wavelength range can be used to ‘dazzle’ the missile seeker so that it cannot acquire a target.”
Manijeh Razeghi’s group produces powerful QCLs on InP.
Another class of laser that emits around 3 µm is the interband cascade laser. However, the performance of this class of device tends to vary far more with changes in temperature than a QCL. Historically, shorter wavelength QCLs have lagged behind their longer wavelength cousins in temperature insensitivity, but lasers from Northwestern have significantly closed this gap.
“We believe our success is partly due to the use of better electron confinement and a three-well active region,” reveals Razeghi.
The QCL community has made pulsed-mode QCLs operating at room-temperature from three different types of material system: InAs/AlSb/InAs, GaInAs/AlAs(Sb)/InP and GaInAs/AlInAs/InP. Razeghi and her co-workers use latter of these because its yields the best material quality, and it is best suited to commercial device manufacture. “A large infrastructure exists that is dedicated to the growth of antimony-free InP-based compounds,” says Razeghi, who adds that antimony-containing systems would require significant ‘retooling’ and process development. The QCLs that she and her co-workers fabricate span an emission range from 3 µm to the terahertz range.
When GaInAs/AlInAs structures are lattice-matched to InP, the quantum-well depth is just 505-520 meV, which in insufficient for electron confinement for making QCLs operate below 6 µm. It is possible to increase the depth of the well a great deal by switching from GaInAs to InAs and replacing AlInAs with AlAs. However, this alternative material system is not lattice-matched to InP, preventing the growth of high-quality films that are a pre-requisite for making devices.
“To grow a device with a deep quantum well, good interfaces and without dislocations, the compressive strain of the InAs-rich material must be balanced – on the scale of one cascade stage – with the tensile strain of the AlAs-rich layer,” explains Razeghi. Those strain-balanced layers have a depth of about 1 eV. “This allows us to reach much shorter emission wavelengths with good electron confinement.”
Using this approach led to the construction of a 3.39 µm, double-channel ridge-waveguide QCL with an 8.6 µm width and 5 mm laser cavity. Driven with 500 ns pulses at a 100 kHz repetition rate, this laser produces 1.1 W at room temperature. Switch to CW operation and output falls to 504 mW, 403 mW and 88 mW at 15 °C, 25 °C and 55 °C, respectively. The QCL’s 3.56 µm cousin has a 10.5 µm-wide ridge, and produces 1.66 W when driven in pulsed mode. In CW operation, output falls to 576 mW, 437 mW and 45 mW at 15 °C, 25 °C and 55 °C, respectively.
QCL performance can be improved with a buried heterostructure geometry that trims ridge widths and optical losses, while increasing thermal conductance. Performance then improves on several fronts: Threshold current falls; slope efficiency rises, resulting in higher output power; and internal temperature increases are smaller at all drive currents. The two major benefits that result from all these gains are an increase in the power saturation current density and a threshold current that is low, even at high temperatures. “This leads naturally to a higher operation temperature for CW operation,” remarks Razeghi.
Along with her co-workers, she is trying to improve material homogeneity, which will lead to better QCL performance. “We are systematically trying to identify the source of inhomogeneity. This includes investigations of reactor hardware limitations, growth conditions and interface effects.”
N. Bandyopadhyay et al. Appl. Phys. Lett. 100 212104 (2012)
