Room temperature surface emission is realized on a large area (1.5 mm × 1.5 mm) photonic crystal quantum cascade laser (PhC-QCL) driven under pulsed mode, at the wavelength around 8.75 μm. By introducing in-plane asymmetry to the pillar shape and optimizing the current injection with a grid-like window contact, the maximum peak power of the PhC-QCL is up to 5 W. The surface emitting beam has a crossing shape with 10° divergence.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Quantum cascade lasers (QCLs) are semiconductor laser sources based on intersubband transitions . In the mid-infrared region, QCLs are the sources of choice for many laser-based applications. Especially, high power QCLs are widely used for applications such as process control , remote sensing and counter-measures. Surface emitting lasers are favorable for two-dimensional (2D) integration and a stable power extraction, as they are insensitive to cleaving conditions . Because of the unique 2D in-plane coupling mechanism , photonic crystals (PhCs) [5–10] provide a powerful solution for surface-emitting QCLs. Furthermore, large-area surface-emitting lasers are advantageous for realizing high output power with a narrow and symmetrical beam shape [11–13].
Previously, our group has demonstrated room-temperature large-area PhC-QCLs by using the buried heterostructure and buried grating technique [13,14]. However, the peak output power was below 1W. In this work, by improving the design in several aspects, we present a large-area (> 500 periods in each in-plane dimension) PhC-QCL with 5W surface-emitting peak power driven in pulsed operation.
2. Design and fabrication
A schematic drawing is shown in Fig. 1(a). The 2.6 μm-thick active region is based on a strain-balanced InGaAs/AlInAs structure , grown by molecular beam epitaxy on a low-doped (1.3×1017 cm−3) InP substrate. A 600 nm-thick layer of Si-doped InP (InP:Si) is grown on top of the active region by metal-organic vapor phase epitaxy (MOVPE), to prevent the overgrowth of semi-insulating InP (InP:Fe) later in the process. A SiNx hard-mask is deposited for a smooth and deep etching. After deep-ultraviolet lithography (220 nm wavelength), the SiNx hard-mask is patterned with the reactive-ion etching. The active region is deeply etched into square-lattice PhC pillars by inductively coupled plasma (ICP) dry-etching with Cl2 and H2 at 200°C. As shown in the inset of Fig. 1(a), the top-view shape of the PhC pillar is a square with a missing corner. This asymmetry enhances the surface power extraction at the Γ2 point, by increasing the radiation constant in the surface-normal direction [3, 13]. A 600 nm-thick active region layer remains unetched to maintain a high optical overlap factor and to prevent over-etching into the InP substrate. Figure 1(b) shows the scanning electron microscopy (SEM) image of the PhC-QCL after ICP etching. The empty space of the PhC layer is filled with InP:Fe by hydride vapour phase epitaxy (HVPE), building the PhC layer with an index contrast between 3.08 (InP) and 3.35 (InGaAs/AlInAs). After removing the SiNx hard-mask by HF wet-etching, a 5.3 μm-thick InP:Si cladding layer is grown on top by MOVPE. The laser cavity is defined by wet-etching the cladding layer into 1.5mm × 1.5mm square mesas. Then, a SiNx layer is deposited on the boundaries of the mesa, serving to create absorbing boundaries. The substrate is lapped and polished to a thickness of 200 μm. The epi-side of the laser is fully covered by a 4 μm-thick gold contact, and the emission window is opened at the substrate side. The window contact is initially a square window with a width of 100 μm and a thickness of 4 μm, deposited by electroplating. Then an additional grid-like gold contact (268 nm thickness) is deposited on top of this square window to optimize the current injection. The overall shape of the window electrode is schematically shown in Fig. 1(a), where the thickness of the additional contact is made to the same thickness as the initial one for a clear view. After processing, the laser is cleaved and epi-down mounted on an AlN submount.
One limitation on the performance of large area surface-emitting PhC-QCLs is the uniformity of the current injection . In this work, the influence of the additional grid-like contact is investigated with a 2D COMSOL simulation, as shown in Fig. 2(a). The figure shows the current density distribution along the in-plane dimension at the center cross-layer of the active region. It is clear that, although the additional contact is only 50 μm wide and 268 nm thick, the current uniformity is significantly enhanced under the same voltage bias, and the current density at the center is more than doubled.
The vertical profiles of the electric field |Ez| (blue curve) and the refractive index (black curve) are shown in Fig. 2(b). The electric field is simulated by COMSOL Multiphysics with the pillar filling factor of 0.45 and the band-edge mode E1. As will be explained later, the E1 mode is the lasing candidate in this structure, and the corresponding mode pattern can be found in Fig. 3. The field profiles in Fig. 2(b) are obtained by retrieving the data along two vertical cut-lines in the simulation, one penetrating the PhC pillar, and the other one not. The positions of the cut-lines are schematically shown in the insets of Fig. 2(b). The PhC layer and the unetched active region layer are marked light and dark in the figure. The uncolored regions represent the InP cladding layer and the substrate.
At the Γ2 point of the square-lattice PhC structure, there are four fundamental Bloch waves that dominate the energy, referred to as Rx, Sx, Ry, Sy [4, 16–18]. Figure 2(c) shows the coupling diagram of these Bloch waves in the momentum space of the photonic crystal. There are two major coupling mechanisms among them: the 180° 1D coupling κ1D and the 90° 2D coupling κ2D [4, 19]. The κ1D is intrinsically similar to the coupling strength in conventional distributed-feedback lasers [20, 21], whereas the κ2D is a unique property of the 2D photonic crystal structure [19,22].
As a result of the couplings among the fundamental Bloch waves, the eigensolutions of the PhC structure become four band-edge modes at the Γ2 point : mode A, B, E1 and E2. The electric field patterns of the four band-edge modes in a single lattice are shown in Fig. 3(a), where the color-map indicates the out-of-plane component Ez, and the vector map exhibits the in-plane components Ex and Ey. The normalized mode frequencies, the overlap factors, the vertical losses and the vertical radiation constants of the band-edge modes are obtained by sweeping the filling factor in the COMSOL simulation, as shown in Figs. 3(b)–3(e). Periodic boundary conditions are adopted in the simulation, which means the in-plane energy leakage is not considered. The material absorption of InP and the active region is neglected. The vertical losses are defined as the cavity losses of the eigenmodes in the simulation: α = 2π/(aQ), where a is the lattice constant of the PhC, and Q is the quality factor. The radiation constant is defined in the same manner, but the epi-side gold contact is replaced by a perfectly matched layer, thus the metal loss is eliminated in the simulation. At the filling factor of 0.45, the optical overlap factors with the active region () for modes A, B, E1, E2 are 34.1%, 20.5%, 42.8% and 33.2%, respectively. Due to the low overlap factor and the high vertical loss, mode B is unlikely to be the first lasing mode. Mode E1 instead, has the highest overlap factor and the lowest vertical loss, meaning the lasing candidate. It also has the second highest radiation constant, which is favorable for high output power.
The light-current-voltage (LIV) characteristics of a PhC-QCL with a period of 2.78 μm and the pillar filling factor of 0.45 is shown in Fig. 4(a). The laser is driven at pulsed operation (52 ns at 1.923 kHz) and the power is collected from the sample surface. The current density is obtained by having the current divided by the area of the active region area in the PhC layer. The measurement is conducted at both 289 K and 298 K. The thresholds and slope efficiencies are almost the same at both measurements. The threshold current density is around 5.1 kA/cm2. At 289 K, the maximum peak power is up to 5.02 W (at 88.8 A). At 298 K, the maximum peak power is 3.77 W (at 78.4 A). The limits on the maximum currents and the fluctuations of the measured curves, especially noticable at > 50 A currents, are caused by the imperfections of the electrical drive and measurement circuits. Since the roll-over is not reached, a higher emitting power of the same PhC-QCL is expected with a more powerful driver.
The slope efficiencies in both measurements are around 130 mW/A, which are significantly enhanced compared to our previous published result (17.5 mW/A) . This improvement can be attributed to the strain-balanced active region used for this laser , the stronger effective index contrast (2 μm-thick PhC layer) and the additional grid-like contact which enhances the current injection. In the ideal case (without process imperfections), the total area of the window-side gold contact is 0.85 mm2, which equals to 37.8% of the surface emitting area. The complex refractive index of the substrate can be calculated with the Drude model, which gives the value of 3.08 + 8.32 × 10−4 j at the doping of 1.3×1017 cm−3. The corresponding energy absorption from the substrate in a single pass αs = 1 − e−4πκd/λ0 is 21.3%. The reflectivity R from InP (refractive index 3.08) to the air is 0.260. Considering the multi-pass reflection and the area blocked by the electrode, the ratio between the externally collected surface-emitting power and the overall extracted power in the out-of-plane direction is:Appendix, Sec. A. The edge-emitting slope efficiency of this PhC-QCL is 73 mW/A. Considering the in-plane symmetry, the total power leakage through the four side facets is approximately 292 mW/A. By summing up with the out-of-plane power, the overall internal slope efficiency is around 593 mW/A.
The surface-emitting spectrum is shown in Fig. 4(b), measured with a Fourier-transform infrared spectrometer at the spectral resolution of 0.125 cm−1. The PhC-QCL is multi-mode lasing at the wavelength around 1140 cm−1, which can be attributed to two possible reasons. Firstly, the regrowth is not perfect. There are some random air holes between the InP:Fe grown by HVPE and the InP:Si grown by MOVPE. Such voids are commonly seen among all the inspected devices (for SEM analysis) from the same wafer as the presented QCL. An example is shown in the Appendix, Sec. B. In our opinion, the voids are formed by the combination of unplanarized HVPE regrowth and the subsequent anisotropic growth of MOVPE. The growth rate of MOVPE is different on different crystal planes. The vertical growth rate on (001) surface is higher than that on the facet planes of unplanarized HVPE growth between mesas, which causes the holes to be closed before the voids can be filled . The unintended voids lead to a random and strong index contrast, which weakens the periodicity of the PhC. Secondly, our coupled wave theory model  shows that the |κ1DL| of a square pillar PhC-QCL with the same structural parameters is larger than 10. The square-shape pillar (without a missing corner) is used in the model for the simplicity of the Fourier transformation. The model gives a κ1DL of −10.2 and a κ2DL of 8.1. Although both the major couplings have similar amplitudes and the opposite signs, the |κ1DL| beyond the moderate-coupling region (between 1 and 4) might be harmful for single-mode operation, due to the spatial hole burning effect [16,20,23]. Besides, the large |κ1DL| might compact the in-plane intensity envelope of the fundamental mode, leading to a divergent surface-emitting beam. In the future, the design can be improved by maintaining |κ2DL| and meanwhile suppressing |κ1DL|, towards the case where |κ1DL| is considerably smaller than |κ2DL| .
The surface-emitting far-field pattern is shown in Fig. 5(a), measured by a mid-infrared beam profiler 3 cm away from the laser surface. The polarization characteristics are shown in Figs. 5(b)–5(e), where the white arrows represent the direction of the polarizer in each measurement. The results are calibrated against the image when the laser is switched off. The beam pattern is a symmetrical crossing with a divergence of around 10 °. One possible reason for the crossing-shape is the grid-like window-side electrode. A theoretical calculation on the impact of the window shape is performed by taking the Fourier transform of the near-field intensity [24, 25]. See Appendix, Sec. C for more details. Another reason could be the multi-mode operation. The overall emitting beam consists of several patterns with different polarizations. The main-lobe in the center with 1° divergence has a major polarization component along the vertical direction in Fig. 5, whereas the line-shaped lobe below the center has a polarization along the horizontal axis. This indicates that the lasing modes can be finite-size PhC modes from different bands. The bright line-shaped lobe in Fig. 5(d) might also suggest a 1D mode lasing.
To conclude, surface-emission lasing is realized with a large-area (1.5 mm × 1.5 mm) PhC-QCL at room temperature driven under pulsed operation, with the wavelength around 8.75 μm. By optimizing the PhC design and the window-side electrode and selecting the proper active region, the maximum peak power is up to 5 W. With a more powerful electrical driver, a higher power is expected with the presented laser.
There is still a substantial room to enhance the laser performance. For instance, the regrowth planarization can be improved to avoid the air holes. Other PhC pillar shapes that are less sensitive to fabrication imperfections may benefit the single-mode selectivity. The device area can be modified for a better balance between the lasing threshold and the radiation efficiency. The PhC structure can be optimized to reduce the |κ1DL| and effectively flatten the in-plane mode profile . Furthermore, anti-reflection coating on the substrate surface will also increase the emission efficiency. Although our PhC-QCL process is far from maturity, it already competes with state of the art performances. In the end, it is our belief that high power surface-emitting PhC-QCLs will eventually play a significant role in industrial and academic applications in the mid-infrared region.
A. Characterization of edge emission
The edge emitting LIV of the presented PhC-QCL is shown in Fig. 6. The slope efficiency is 73 mW/A.
B. Regrowth imperfections
A cross-sectional SEM image of a PhC-QCL is shown in Fig. 7. This PhC-QCL is from the same wafer as the presented PhC-QCL in the main text. Random-size regrowth voids can be seen.
C. Far-field pattern
The impact of the electrode shape on the far-field pattern is theoretically investigated. Figures 8(a) and 8(b) show the near-field intensity with and without the additional grid-like gold contact. As the simplest case, the intensity of the exposed area is assumed unity everywhere. Figures 8(c) and 8(d) show the corresponding far-field patterns of Figs. 8(a) and 8(b), respectively. In each figure, the profile is normalized to 1. The maximum (saturation) colorbar value in Figs. 8(c) and 8(d) is set to 0.1, in order to illustrate the side-lobes clearly. Since this model is focusing on the influence of the electrode shape, the PhC pattern is neglected for simplicity. The in-plane phase in both cases is assumed to be equal in the whole plane. In both cases, the full width half maximums of the central lobes are around 0.4°. The additional grid-like electrode enhances the brightness of the side-lobes.
H2020 European Research Council (ERC) (724344); FP7 People: Marie-Curie Actions (PEOPLE) (FEL-27 14-2); Vetenskapsrådet (VR) (2015-05470).
2. N. Lang, J. Röpcke, S. Wege, and A. Steinbach, “In situ diagnostic of etch plasmas for process control using quantum cascade laser absorption spectroscopy,” Eur. Phys. J. Appl. Phys. 49, 13110 (2010). [CrossRef]
3. K. Hirose, Y. Liang, Y. Kurosaka, A. Watanabe, T. Sugiyama, and S. Noda, “Watt-class high-power, high-beam-quality photonic-crystal lasers,” Nat. Photonics 8, 406–411 (2014). [CrossRef]
4. Z. Wang, Y. Liang, X. Yin, C. Peng, W. Hu, and J. Faist, “Analytical coupled-wave model for photonic crystal surface-emitting quantum cascade lasers,” Opt. Express 25, 11997–12007 (2017). [CrossRef] [PubMed]
5. R. Colombelli, K. Srinivasan, M. Troccoli, O. Painter, C. F. Gmachl, D. M. Tennant, A. M. Sergent, D. L. Sivco, A. Y. Cho, and F. Capasso, “Quantum cascade surface-emitting photonic crystal laser,” Science 302, 1374–1377 (2003). [CrossRef] [PubMed]
6. Y. Bai, B. Gokden, S. Darvish, S. Slivken, and M. Razeghi, “Photonic crystal distributed feedback quantum cascade lasers with 12 W output power,” Appl. Phys. Lett. 95, 1105 (2009). [CrossRef]
7. G. Xu, R. Colombelli, R. Braive, G. Beaudoin, L. Le Gratiet, A. Talneau, L. Ferlazzo, and I. Sagnes, “Surface-emitting mid-infrared quantum cascade lasers with high-contrast photonic crystal resonators,” Opt. Express 18, 11979–11989 (2010). [CrossRef] [PubMed]
8. L. Mahler and A. Tredicucci, “Photonic engineering of surface-emitting terahertz quantum cascade lasers,” Laser Photonics Rev. 5, 647–658 (2011).
9. Z. Diao, C. Bonzon, G. Scalari, M. Beck, J. Faist, and R. Houdré, “Continuous-wave vertically emitting photonic crystal terahertz laser,” Laser Photonics Rev. 7, L45–L50 (2013). [CrossRef]
10. D.-Y. Yao, J.-C. Zhang, O. Cathabard, S.-Q. Zhai, Y.-H. Liu, Z.-W. Jia, F.-Q. Liu, and Z.-G. Wang, “10-W pulsed operation of substrate emitting photonic-crystal quantum cascade laser with very small divergence,” Nanoscale Res. Lett. 10, 177 (2015). [CrossRef] [PubMed]
11. C. Sigler, J. Kirch, T. Earles, L. Mawst, Z. Yu, and D. Botez, “Design for high-power, single-lobe, grating-surface-emitting quantum cascade lasers enabled by plasmon-enhanced absorption of antisymmetric modes,” Appl. Phys. Lett. 104, 131108 (2014). [CrossRef]
12. C. Boyle, C. Sigler, J. Kirch, D. Lindberg III, T. Earles, D. Botez, and L. Mawst, “High-power, surface-emitting quantum cascade laser operating in a symmetric grating mode,” Appl. Phys. Lett. 108, 121107 (2016). [CrossRef]
13. Y. Liang, Z. Wang, J. Wolf, E. Gini, M. Beck, B. Meng, J. Faist, and G. Scalari, “Room temperature surface emission on large-area photonic crystal quantum cascade lasers,” Appl. Phys. Lett. 114, 031102 (2019). [CrossRef]
14. R. Peretti, V. Liverini, M. J. Süess, Y. Liang, P.-B. Vigneron, J. M. Wolf, C. Bonzon, A. Bismuto, W. Metaferia, M. Balaji, S. Lourdudoss, E. Gini, M. Beck, and J. Faist, “Room temperature operation of a deep etched buried heterostructure photonic crystal quantum cascade laser,” Laser Photonics Rev. 10, 843–848 (2016). [CrossRef]
15. 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, 24272–24279 (2012). [CrossRef] [PubMed]
17. Y. Liang, C. Peng, K. Sakai, S. Iwahashi, and S. Noda, “Three-dimensional coupled-wave model for square-lattice photonic crystal lasers with transverse electric polarization: A general approach,” Phys. Rev. B 84, 195119 (2011). [CrossRef]
18. Y. Yang, C. Peng, Y. Liang, Z. Li, and S. Noda, “Three-dimensional coupled-wave theory for the guided mode resonance in photonic crystal slabs: TM-like polarization,” Opt. Lett. 39, 4498–4501 (2014). [CrossRef] [PubMed]
19. M. Yoshida, M. De Zoysa, K. Ishizaki, Y. Tanaka, M. Kawasaki, R. Hatsuda, B. Song, J. Gelleta, and S. Noda, “Double-lattice photonic-crystal resonators enabling high-brightness semiconductor lasers with symmetric narrow-divergence beams,” Nat. Mater. 18, 121 (2019). [CrossRef]
20. H. Kogelnik and C. Shank, “Coupled-wave theory of distributed feedback lasers,” J. Appl. Phys. 43, 2327–2335 (1972). [CrossRef]
21. Z. Wang, H. Zhang, L. Ni, W. Hu, and C. Peng, “Analytical perspective of interfering resonances in high-index-contrast periodic photonic structures,” IEEE J. Quantum Electron. 52, 1–9 (2016). [CrossRef]
22. L. Ni, Z. Wang, C. Peng, and Z. Li, “Tunable optical bound states in the continuum beyond in-plane symmetry protection,” Phys. Rev. B 94, 245148 (2016). [CrossRef]
23. W. Rabinovich and B. Feldman, “Spatial hole burning effects in distributed feedback lasers,” IEEE J. Quantum Electron. 25, 20–30 (1989). [CrossRef]
24. M. N. Sadiku, Elements of electromagnetics (Oxford University, 2007).
25. Y. Liang, C. Peng, K. Sakai, S. Iwahashi, and S. Noda, “Three-dimensional coupled-wave analysis for square-lattice photonic crystal surface emitting lasers with transverse-electric polarization: finite-size effects,” Opt. Express 20, 15945 (2012). [CrossRef] [PubMed]