We report a room-temperature eight-element phase-locked quantum cascade laser array emitting at 8 µm with a high continuous-wave power of 8.2 W and wall plug efficiency of 9.5%. The laser array operates primarily via the in-phase supermode and has single-mode emission with a side-mode suppression ratio of ~20 dB. The quantum cascade laser active region is based on a high differential gain (8.7 cm/kA) and low voltage defect (90 meV) design. A record high wall plug efficiency of 20.4% is achieved from a low loss buried ridge type single-element Fabry-Perot laser operating in pulsed mode at 20 °C.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
The infrared spectral region between 3 and 12 µm is technologically important because it contains strong fingerprint-like absorption features of many molecules and two atmospheric transmission windows at 3-5 µm (MWIR) and 8-12 µm (LWIR) . Compact, room-temperature, high-power, and high-efficiency long-wavelength infrared lasers are of high interest for spectroscopy sensing . Also, high average power lasers emitting in the two transmission windows are highly desirable for infrared countermeasures, which primarily aims at blinding heat-seeking missiles . Quantum cascade lasers (QCLs) have become the leading laser source emitting between 3 and 12 µm due to their compact form, ability to deliver high power continuous wave (CW) radiation at room temperature, and emission wavelength flexibility through bandgap engineering [4,5]. So far, due to the tremendous effort on high-quality epitaxial material growth, bandgap engineering, development of low-loss waveguides, and advanced thermal packaging, CW output power of 5.1 W has been achieved from one-side of a single-element QCL emitting in the MWIR band . However, CW output power is limited to about 1.0 W from one-side of a QCL emitting in the LWIR band [7–11]. This is due to the decrease of wall plug efficiency (WPE) for longer wavelength QCLs. The decrease of WPE is because of lower voltage efficiency which results from small photon energy over voltage defect ratio, and higher free carrier absorption which is proportional to the square of the wavelength .
2. High-differential-gain QCL active region design for λ ~8 µm
Here, we report a high-differential-gain QCL active region design targeting λ ~8 µm, which gives a high WPE of 20.4% in pulsed mode operation and a CW output power of 2.0 W is achieved from one-side of a single-element QCL using epi-layer down bonding. Figure 1 shows the conduction band diagram and wavefunctions of relevant energy levels of the active region design, which is based on a strain balanced Al0.64In0.36As/In0.59Ga0.41As structure. This structure also features a diagonal optical transition compared to the design used in , which means the overlap between upper laser level 2 and lower laser level 1 is reduced. This helps to increase the lifetime of electrons on level 2 from 0.5 ps to 0.75 ps and to thus also increase the differential gain and quantum efficiency. Most previous QCL designs in the LWIR range target a voltage defect of Δinj = E1-Eg = 120 meV, at the resonant field, to largely suppress population backfilling to the lower lasing levels. However, for QCL designs with increased differential gain and quantum efficiency, a relatively small value of Δinj can be used without introducing significant population backfilling . This will greatly enhance the voltage efficiency and therefore WPE. Our designed QCL active region targets a smaller voltage defect Δinj of 90 meV. Experimentally, a laser structure consisting of 45-stages of the strain balanced design is grown on a n- (Si, 2 × 1017 cm−3) InP substrate by gas-source molecular beam epitaxy (GSMBE). The detailed layer structure from the InP substrate consists of a 4 µm InP buffer layer (Si, ~2 × 1016 cm−3), the 45-stage laser core, a 50 nm InP spacer (Si, ~2 × 1016 cm−3), and a 300 nm InGaAs grating layer (Si, ~2 × 1016 cm−3) for single mode selection. After GSMBE growth, metal organic chemical vapor deposition (MOCVD) is used for the growth of a 4 µm InP cladding layer (Si, ~2-5 × 1016 cm−3) and 1 µm InP cap layer (Si, ~5 × 1018 cm−3).
The complete wafer is processed into a standard, low loss, buried ridge waveguide with a ridge width of 8 µm. Laser bars of 2 mm, 3 mm and 4 mm length are cleaved and epi-layer up bonded without any facet coating for testing in pulsed mode at 20 °C, which is regulated by a thermoelectric cooler (TEC). The power-current-voltage (P-I-V) and WPE result is shown in Fig. 2(a). A maximum WPE of 20.4% is achieved from a 3-mm long device, which is the highest among all reported LWIR QCLs. This is partially due to a low threshold operating voltage, which is 10% lower than the laser structure with a similar core thickness reported in . As shown in Fig. 2(b), threshold current density and slope efficiency are fitted as function of mirror loss and inverse mirror loss respectively, which depend solely on laser length for uncoated devices. The wafer has a differential gain gd as high as 8.6 cm/kA, a transparency current density Jtr of 1.13 kA/cm2, internal quantum efficiency ηi of 66.1%, and the waveguide loss is only 1.34 cm−1, all of which have contributed to the high WPE.
A 5-mm long device is high reflectance (HR)-coated with Y2O3/Au/Ti (400 nm/100 nm/30 nm) on the back facet, anti-reflectance (AR)-coated with Y2O3 (880 nm) on the front facet, and epi-layer down bonded to a diamond sub-mount before bonding to a copper mount for testing. The device is first tested in pulsed mode at 20 °C and a temperature range from 15 °C to 85 °C with a step of 10 °C. At 20 °C, the maximum WPE is 17.0%, with the result shown in Fig. 2(c). By analyzing temperature dependent threshold current density and slope efficiency, as shown in Fig. 2(d), characteristic temperatures T0 and T1 are fitted to be 170 K and 450 K respectively, which are adequately high and crucial for high temperature operation of the device. In CW mode operation, a high output power of 2.0 W is achieved from the AR-coated side. Maximum WPE in CW mode operation is 12.8%, which is the highest among all previously reported LWIR QCLs operating in CW mode. Emission spectrum of the CW device at I = 1.5 A ranges between 8.05 and 8.3 µm, as shown in the inset of Fig. 2(c).
3. High-power, continuous-wave, phase-locked QCL arrays
The increase of WPE is important for reducing excessive heat, and therefore increases CW output power. But achieving significantly higher output power can only be achieved by increasing the volume of laser active region while keeping the same WPE. Power scaling of QCLs has been successfully demonstrated using broad area QCLs  and photonic crystal QCLs . But these power scaling schemes are limited to pulsed mode operation due to insufficient heat extraction from a wide waveguide (>50 µm), considering the relatively high threshold current density and operating voltage of QCLs. A promising approach for power scaling of QCLs in CW mode is to use an array of narrow-width QCLs that can be separated far enough for necessary heat removal. It has to be noted that for a non-coherent laser array, of which the phase of each element is not locked, the far field brightness decreases rapidly with the pitch of the array, and therefore phase-locked laser arrays are highly desirable .
Phase-locked QCL arrays have been realized with several architectures. These include leaky wave coupled , evanescent coupled , Talbot-cavity coupled [18,19], tree-array Y-junction combined  and tree-array multimode interferometer (MMI) combined [21,22] QCL arrays. Among them, the MMI combined QCL array has the highest potential for realizing power scaling in CW operation thanks to the high efficiency of the MMI for beam combining/splitting, the narrow ridge width across the entire device, and a flexible choice of array element separation (pitch). Previously, we have demonstrated high-power phase-locked QCL arrays in pulsed mode operation using the MMI combined laser array based on a dry-etched waveguide . Here, we take advantage of the high-efficiency LWIR QCL wafer and optimize the MMI combined laser array structure for CW operation by using a buried ridge waveguide, a larger array pitch, and applying micro-impingement cooling to remove waste heat.
Figure 3(a) shows the schematic structure of a 6-mm long, eight-element QCL array combined by a tree array MMI. The 6-mm length is chosen based on a tradeoff between output power and device fabrication yield. Except the MMI section, the ridge width of all other sections is as narrow as 5 µm and is based on a buried ridge type waveguide for efficient heat removal and a low waveguide loss. The single section side, which is 750 µm long, contains a uniform distributed feedback (DFB) grating for single wavelength selection. The product of the DFB section length and coupling strength is ~5. In addition, the single-section side will be HR-coated for further reduction of the lasing threshold. The 1.8-mm long array section has a pitch of 200 µm, which is selected based on thermal simulation as shown in the next paragraph. The multi-section side will be AR-coated to extract more output power, and targeted facet reflectivity is 8%. S-bend waveguides and MMIs are used to connect the DFB section and the array. The radii of curvature for the s-bend waveguides is 1010, 1340 and 1260 µm for the three stages of beam splitting, respectively, which is chosen based on beam propagation simulation. Every MMI is 21 µm wide and its length is determined to be 130 µm, also by beam propagation simulation. Simulated beam splitting efficiency of a MMI for each channel is >48% (ideal is 50%). In total, the power splitting efficiency from the DFB section to the array is >10% for each channel (ideal is 12.5%), when not considering amplification of the tree-array beam splitter.
To optimize the array pitch for efficient heat removal, the laser core temperature increase as a function of pitch width is simulated by a finite element method (FEM). To achieve a CW output power of ~10 W, the input power may reach 100 W, considering a CW WPE of ~10%. Therefore, the 1.6 mm* 6 mm QCL array chip will first be epi-layer down bonded to a diamond sub-mount. Then the diamond sub-mount will be bonded to a state-of-art micro-impingement cooler, which provides direct water cooling under the cooler surface. Under a water flow of 2 L/min, the cooler has a thermal resistance of 0.025 K⋅cm2/W. This sets the interface between diamond and micro-impingement cooler to a heat flux boundary condition with a heat transfer coefficient of 40 W/K⋅cm2. Thermal conductivity of the QCL core is 4 W/m⋅K and 2 W/m⋅K in the lateral and vertical direction respectively . Thermal conductivity of the rest of material used for modeling could be found in . Figure 3(b) shows the laser core temperature increase as a function of pitch width. The amount of laser core temperature increase decreases rapidly when the pitch increases from 50 µm to 200 µm and then decreases only slowly after that. Considering the total size of the array, we have chosen 200 µm as the pitch. The estimated laser core temperature increase is 72 K at the estimated rollover current density of 4.0 kA/cm2, assuming a WPE of 10%.
The insulating material (Fe-doped InP) of conventional buried ridge type QCLs are grown after the etching of laser core and InP cladding and cap layers, which requires an etching depth of usually >10 µm. However, a deep etch is not desirable for the QCL arrays which contain curved waveguides. Because the crystal direction of curved waveguides varies at different position, a deep etching may result in profile distortion. To address this issue, we have developed a shallow etch process that grows the Fe-doped InP insulating material before cladding and cap layer regrowth. This way, the required etching depth is around 5 µm, which ensures a high-quality waveguide profile. After Fe-doped InP regrowth following the shallow channel etching and SiO2 hard mask removal, InP cladding and cap layers were grown by MOCVD. Then, wide isolation channels were etched outside the Fe-doped InP area for current confinement. The rest of the fabrication is the same as for standard buried ridge type QCLs. Figure 4(a) shows scanning microscope image (SEM) of the cross section of the shallow etched buried ridge structure and the laser core area, which has a width of 5 µm. There is no void on the interface between the laser core and Fe-doped InP. Figure 4(b) shows the region close to the two-output side of a MMI. High quality regrowth without any void has been achieved in this region as well.
Figure 5 shows the P-I-V testing result of a 6-mm long HR-AR coated eight-element array epi-layer down bonded to a diamond submount, which is bonded to a micro-impingement cooler chilled at 15°C. The array device is uniformly biased with one current source. In pulsed mode operation, the threshold current density is 1.5 kA/cm2, and a high output power of 13 W is achieved at I = 8 A. The maximum WPE is 13.5%. It is worth noting that the array is epi-layer down bonded to a diamond submount coated with only 100 nm thick gold. There is a series resistance of 0.1 Ω from the sub-mount surface, which leads to an extra voltage drop of up to 0.8 V. After voltage correction, device efficiency is about 14.2%. This number is close to the WPE of the 8-µm wide buried ridge HR-AR coated Fabry-Perot (FP) laser. This indicates that the splitting loss of the MMIs and the scattering loss of the curved waveguides are very low. In CW operation, the threshold current density is 1.7 kA/cm2. The threshold current density increase translates to a laser core temperature increase of only 25 K. A maximum CW power of 8.2 W is achieved at I = 7.2 A, which is limited by thermal rollover. At this current value, the thermal efficiency (ratio of CW power to pulsed mode power) is as high as 75%, and estimated laser core temperature increase is 65 K, which matches well with the simulation. The maximum WPE in CW mode is as high as 9.5% nominal and 10.1% after serial voltage correction, despite a total input power of more than 80 W. It is attributed to the high gain of the wafer, low-loss resonator design and excellent thermal management.
Figure 6 shows the measured emission spectrum of the array device operating in CW mode. Single mode emission is obtained at all current levels from 3A to 7A with an increase step of 1 A, and side mode suppression ratio (SMSR) is >19 dB. Wavelength shifts with current are observed, which is due to internal heating effects.
The far field distribution of the array optical output is measured to evaluate its phase–locking characteristics. As shown in Fig. 7, the distribution is stable when injection current increases from 4.0 A to 6.5 A. The measured far fields are compared to the intensity distribution of an eight-slit Fraunhofer diffraction, and a close match is achieved, which indicates the laser array is operating via the in-phase mode. Beside the main peaks, there is a slight secondary set of peaks, which is most likely caused by fabrication related phase-error, such as a non-uniform path length introduced by material defects. Also, the secondary set of peaks could be related to a non-idea MMI geometry, where the out-of-phase mode is not completely suppressed. In future work, phase refinement could be achieved using either on-chip thermal phase tuner [25,26] or external phase controller .
In conclusion, we report the design and growth of a LWIR QCL active region with a high differential gain and low voltage defect. By utilizing a low-loss buried ridge waveguide, a record high WPE of 20.4% is achieved in pulsed mode operation at 20 °C, and a CW power of 2.0 W is obtained from the AR-coated side of a HR-AR coated epi-layer down bonded device. Furthermore, we demonstrate power scaling in CW operation using a MMI combined, phase-locked, eight-element QCL array. A high CW power of 8.2 W is achieved, with a WPE of 9.5%. The array operates primarily via the in-phase supermode and exhibits single mode spectra with a SMSR of ~20 dB. Imminent future work is the development of either on-chip or external phase tuners for beam steering demonstrations. Since the array structure is adaptive in wavelength, it could be applied to the MWIR wavelength band, where significantly higher CW powers (>20 W) are expected since up to 5.1 W has been demonstrated from a single device. Lastly, the MMI combined array structure could also be applied to other types of semiconductor lasers for power scaling, especially those have a stringent requirement on heat dissipation.
Office of Naval Research (N00014-17-1-2836); Naval Air Systems Command (N68936-17-C-0063).
The published material represents the position of the author(s) and not necessarily that of the Navy. The authors would also like to acknowledge the encouragement and support of all of the involved program managers, especially Dr. K. K. Law, Dr. Daniel Green and Dr. Kevin Leonard.
1. A. Schliesser, N. Picque, and T. W. Hansch, “Mid-infrared frequency combs,” Nat. Photonics 6(7), 440–449 (2012). [CrossRef]
2. R. F. Curl, F. Capasso, C. Gmachl, A. A. Kosterev, B. McManus, R. Lewicki, M. Pusharsky, G. Wysocki, and F. K. Tittel, “Quantum cascade lasers in chemical physics,” Chem. Phys. Lett. 487(1–3), 1–18 (2010). [CrossRef]
3. C. K. N. Patel, “Quantum cascade lasers: a game changer for defense and homeland security IR photonics,” Proc. SPIE 8031, 803126 (2011).
5. M. Razeghi, W. Zhou, S. Slivken, Q. Y. Lu, D. Wu, and R. McClintock, “Recent progress of quantum cascade laser research from 3 to 12 μm at the Center for Quantum Devices [Invited],” Appl. Opt. 56(31), H30–H44 (2017). [CrossRef] [PubMed]
6. Y. Bai, N. Bandyopadhyay, S. Tsao, S. Slivken, and M. Razeghi, “Room temperature quantum cascade lasers with 27% wall plug efficiency,” Appl. Phys. Lett. 98(18), 181102 (2011). [CrossRef]
7. Q. Lu, D. Wu, S. Sengupta, S. Slivken, and M. Razeghi, “Room temperature continuous wave, monolithic tunable THz sources based on highly efficient mid-infrared quantum cascade lasers,” Sci. Rep. 6(1), 23595 (2016). [CrossRef] [PubMed]
8. M. Troccoli, A. Lyakh, J. Y. Fan, X. J. Wang, R. Maulini, A. G. Tsekoun, R. Go, and C. K. N. Patel, “Long-Wave IR Quantum Cascade Lasers for emission in the λ=8-12 µm spectral region,” Opt. Mater. Express 3(9), 1546–1560 (2013). [CrossRef]
9. F. Xie, C. Caneau, H. P. Leblanc, D. P. Caffey, L. C. Hughes, T. Day, and C. E. Zah, “Watt-Level Room Temperature Continuous-Wave Operation of Quantum Cascade Lasers With λ > 10 µm,” IEEE J. Sel. Top. Quantum Electron. 19(4), 1200407 (2013). [CrossRef]
10. B. Schwarz, C. A. Wang, L. Missaggia, T. S. Mansuripur, P. Chevalier, M. K. Connors, D. McNulty, J. Cederberg, G. Strasser, and F. Capasso, “Watt-Level Continuous-Wave Emission from a Bifunctional Quantum Cascade Laser/Detector,” ACS Photonics 4(5), 1225–1231 (2017). [CrossRef] [PubMed]
11. A. Lyakh, R. Maulini, A. Tsekoun, R. Go, and C. K. N. Patel, “Multiwatt long wavelength quantum cascade lasers based on high strain composition with 70% injection efficiency,” Opt. Express 20(22), 24272–24279 (2012). [CrossRef] [PubMed]
12. A. Bismuto, R. Terazzi, B. Hinkov, M. Beck, and J. Faist, “Fully automatized quantum cascade laser design by genetic optimization,” Appl. Phys. Lett. 101(2), 021103 (2012). [CrossRef]
13. D. Heydari, Y. Bai, N. Bandyopadhyay, S. Slivken, and M. Razeghi, “High brightness angled cavity quantum cascade lasers,” Appl. Phys. Lett. 106(9), 091105 (2015). [CrossRef]
14. B. Gökden, Y. Bai, N. Bandyopadhyay, S. Slivken, and M. Razeghi, “Broad area photonic crystal distributed feedback quantum cascade lasers emitting 34 W at λ ~4.36 µm,” Appl. Phys. Lett. 97(13), 131112 (2010). [CrossRef]
15. M. J. R. Heck, “Highly integrated optical phased arrays: photonic integrated circuits for optical beam shaping and beam steering,” Nanophotonics 6(1), 93–107 (2017). [CrossRef]
16. J. D. Kirch, C. C. Chang, C. Boyle, L. J. Mawst, D. Lindberg III, T. Earles, and D. Botez, “5.5W near-diffraction-limited power from resonant leaky-wave coupled phase-locked arrays of quantum cascade lasers,” Appl. Phys. Lett. 106(6), 061113 (2015). [CrossRef]
17. Y. H. Liu, J. C. Zhang, F. L. Yan, F. Q. Liu, N. Zhuo, L. J. Wang, J. Q. Liu, and Z. G. Wang, “Coupled ridge waveguide distributed feedback quantum cascade laser arrays,” Appl. Phys. Lett. 106(14), 142104 (2015). [CrossRef]
18. L. Wang, J. Zhang, Z. Jia, Y. Zhao, C. Liu, Y. Liu, S. Zhai, Z. Ning, X. Xu, and F. Liu, “Phase-locked array of quantum cascade lasers with an integrated Talbot cavity,” Opt. Express 24(26), 30275–30281 (2016). [CrossRef] [PubMed]
19. B. Meng, B. Qiang, E. Rodriguez, X. N. Hu, G. Liang, and Q. J. Wang, “Coherent emission from integrated Talbot-cavity quantum cascade lasers,” Opt. Express 25(4), 3077–3082 (2017). [CrossRef] [PubMed]
20. A. Lyakh, R. Maulini, A. Tsekoun, R. Go, and C. K. N. Patel, “Continuous wave operation of buried heterostructure 4.6 µm quantum cascade laser Y-junctions and tree arrays,” Opt. Express 22(1), 1203–1208 (2014). [CrossRef] [PubMed]
21. W. Zhou, S. Slivken, and M. Razeghi, “Phase-locked, high power, mid-infrared quantum cascade laser arrays,” Appl. Phys. Lett. 112(18), 181106 (2018). [CrossRef]
23. L. Missaggia, C. Wang, M. Connors, B. Saar, A. Sanchez-Rubio, K. Creedon, G. Turner, and W. Herzog, “Thermal management of quantum cascade lasers in an individually addressable monolithic array architecture,” Proc. SPIE 9730, 973008 (2016).
24. H. Lee, K. Chung, and J. Yu, “Thermal analysis of InP-based quantum cascade lasers for efficient heat dissipation,” Appl. Phys. B 93(4), 779–786 (2008). [CrossRef]
25. W. Zhou, N. Bandyopadhyay, D. Wu, R. McClintock, and M. Razeghi, “Monolithically, widely tunable quantum cascade lasers based on a heterogeneous active region design,” Sci. Rep. 6(1), 25213 (2016). [CrossRef] [PubMed]
26. W. Zhou, D. H. Wu, R. McClintock, S. Slivken, and M. Razeghi, “High performance monolithic, broadly tunable mid-infrared quantum cascade lasers,” Optica 4(10), 1228–1231 (2017). [CrossRef]