You are here

Heteroepitaxy of Indium Phosphide-Based Quantum Cascade Lasers on Silicon Substrates


OUSD (R&E) MODERNIZATION PRIORITY: Cybersecurity;General Warfighting Requirements (GWR);Microelectronics


TECHNOLOGY AREA(S): Materials / Processes


OBJECTIVE: Design and develop a heteroepitaxy growth process that enables epitaxial growth of high-performance and high-reliability Indium Phosphide-based Quantum Cascade Lasers on silicon substrates.


DESCRIPTION: Monolithic integration of Quantum Cascade Lasers (QCLs) on silicon (Si) would enable a mechanically stable substrate that could take advantage of the best of both worlds: existing high-performance Si-based electronic and optical circuits (e.g., multiple-function, high-speed electronic circuitry; low-loss passive Si optical waveguides; active Si optical modulators and phase-shifters; etc.); and III-V semiconductor-based photonics (e.g., high-performance QCLs, and photo-detectors, etc.). Such compact systems with monolithically integrated mid-infrared sources with Si electronics have applications in infrared countermeasures, integrated transceivers for free-space optical communications, phased-array beam-steerable sources for laser detection and ranging, various passive- and active-optical sensing systems, etc. Moreover, two- and three-photon absorption losses are minimal in the mid-infrared wavelength range, thereby enabling low-loss optical transmission over integrated Si waveguides.


Fabry-Perot (FP) [Ref 1] and distributed-feedback (DFB) [Ref 2] QCLs emitting at 4.6 µm have been demonstrated by wafer bonding on Silicon-on-Nitride-on-Insulator (SONOI) substrates. Transfer printing on silicon-on-sapphire has also enabled monolithic integration of mid-IR QCL on Si [Ref 3]. However, precise alignment limits further advance of such techniques making them less cost-effective. Direct heteroepitaxial growth of QCLs on Si would, potentially, offer a substantially lower cost, large-scale wafer-scale manufacturable approach for optoelectronic integration via growing III-V epitaxial layers on much cheaper and larger Si substrates, as the mature complementary metal oxide semiconductor (CMOS) processing on large Si wafers have proven excellent throughput and yields, thereby offering the most competitive performance and economic advantages.


Nevertheless, heteroepitaxy of III-V semiconductor alloys on Si is quite challenging due to: (a) 8% lattice mismatch between Indium Phosphide (InP) and Si; (b) 50% mismatch in thermal coefficient of expansion; and (c) the formation of antiphase boundaries and domains, which can occur during the growth of polar III-V compounds on nonpolar Si substrates. To overcome these issues, metamorphic-buffer-layers (MBLs) are generally required, which can provide a low-defect-density growth platform of same lattice constant as InP, for the subsequent growth of QCL device structures. Such approaches have been recently successful in realizing high-performance, quantum-dot, active-region diode lasers operating in the near-infrared wavelength regions (1.3-1.55 µm) on Si substrates [Ref 4]. III-V growth on patterned V-grooves alleviates the problems of antiphase domain formation and acts as a filter for dislocations and stacking faults [Ref 5]. Indium Arseide/Indium Aluminum Gallium Arsenide (InAs/InAlGaAs) quantum dots (QDs) have also shown to be effective threading-dislocation (TD) filters for InP MBLs [Ref 6]. However, there are very few studies reporting on direct growth of mid-IR QCLs on Si, in spite of the tremendous aforementioned size, performance, and cost advantages of the game-changing optoelectronic integration.


Molecular beam epitaxy-grown mid-IR QCLs, operating at low temperatures (170 K), have been demonstrated on Si substrates with 6°-miscut towards crystal orientation [111], by employing both a Germanium (Ge) buffer and a compositionally graded Aluminum Indium Arsenide (AlInAs) MBL to target the InP lattice constant [Ref 7]. MBLs, based on QD-dislocation filtering on exact (001) Si, have also been employed for the growth of QCL active regions by MOCVD [Ref 8]. Residual threading dislocation densities have been estimated to be rather high (1E8 cm² range) in both cases. The use of (001)-oriented Si substrates is key to achieving compatibility with Si-CMOS processing. Since QCLs are unipolar devices, they are expected to be insensitive to nonradiative recombination centers. However, dislocations can perturb the QCL superlattice active region and thus interfere with the coherent tunneling process. Thus, it is the objective of this project to reduce the residual-dislocation densities substantially and provide a low-surface roughness platform for the growth of high-performance, high-reliability QCLs on Si, equal with the performance specifications of 5 Watts continuous wave (CW) output at room temperature, wall-plug efficiency no less than 25%, and almost diffraction-limited beam quality with M2 < 1.5.


PHASE I: Develop a path for achieving low-defect density (< 1 x 1E7 /cm²) buffer layers on Si suitable for the growth of mid-IR QCLs. Complete the design of experiments for Phase II to establish room-temperature CW QCL operation on Si substrates. The Phase I effort will include prototype plans to be developed under Phase II.


PHASE II: Demonstrate room-temperature CW QCL operation on Si substrates employing direct-growth methods based on the epitaxial growth methods, and conditions, discovered in Phase I. The performance requirements of the QCL on Si substrates include 5 Watts CW output at room temperature, wall-plug efficiency no less than 25%, and almost diffraction-limited beam quality with M2 < 1.5.


PHASE III DUAL USE APPLICATIONS: Fabricate, test, and finalize the technology based on the design and demonstration results developed during Phase II. Develop a prototype using the finalized design and transition the technology with the final specifications for DoD applications in the areas of Directed Infrared Countermeasures (DIRCM), advanced chemicals sensors, and Laser Detection and Ranging (LIDAR).


The commercial sector can also benefit from this crucial, game-changing technology development of monolithic integration of QCLs with electronics on silicon substrate in the areas of detection of toxic gas environmental monitoring, non-invasive health monitoring and sensing, and industrial manufacturing processing.



  1. Spott, A. et al.“Quantum cascade laser on silicon.” Optica, 3(5), 545-551.
  2. Spott, A. et al. “Heterogeneously integrated distributed feedback quantum cascade lasers on silicon.” Photonics, 3(2) 35.
  3. Jung, S., Kirch, J., Kim, J. H., Mawst, L. J., Botez, D., & Belkin, M. A. (2017, November 20). “Quantum cascade lasers transfer-printed on silicon-on-sapphire.” Applied Physics Letters, 11(211102).
  4. Jung, D., Herrick, R., Norman, J., Turnlund, K., Jan, C., Feng, K., Gossard, A. C., & Bowers, J. E. (2018, April). “Impact of threading dislocation density on the lifetime of InAs quantum dot lasers on Si.” Applied Physics Letters, 112(15) 153507.
  5. Li, Q. & Lau, K. M. (2017, December). “Epitaxial growth of highly mismatched III-V materials on (001) silicon for electronics and optoelectronics.” Progress in Crystal Growth and Characterization of Materials, 63(4), 105-120.
  6. Shi, B., Li, Q., & Lau, K. M. (2017, April 15). “Self-organized InAs/InAlGaAs quantum dots as dislocation filters for InP films on (001) Si.” Journal of Crystal Growth, 464, 28–32.
  7. Go, R. et al. (2018). “InP-based quantum cascade lasers monolithically integrated onto silicon.” Optics Express, 26(17), 22389-22393.
  8. Rajeev, A. et al.(2018, October 10). “III-V superlattices on InP/Si metamorphic buffer layers for ?˜4.8 µm quantum cascade lasers.” Physica Status Solidi, 216(1).


KEYWORDS: Silicon; quantum cascade laser; QCL; monolithic integration; complementary metal oxide semiconductor; CMOS; heteroepitaxial; distributed feedback

US Flag An Official Website of the United States Government