Open Access
Add to BibSonomy
Abstract
We propose and experimentally demonstrate that the lasing power and characteristic temperature (T0) of 905 nm semiconductor lasers can be optimized by use of the high strain quantum well (HSQW). To fix the lasing wavelength around 905 nm, HSQW with a higher ndium (In) content of the InGaAs gain material than that of the commonly used low strain quantum well (LSQW) requires a thickness-reduced quantum well. Thus, the HSQW has the following two advantages: stronger quantum size effects caused by the deep and thin quantum well, and higher compressive strain caused by a high In content of the InGaAs gain material. With the similar epitaxial structure, laser diodes with HSQW have a characteristic temperature T0 of 207 K and can deliver a higher lasing power with less power saturations. The high strain quantum well optimization method can be extended to other laser diodes with a wavelength near 900 nm with low In content InGaAs quantum wells and other similar low-strain gain material systems.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Scanning LiDAR (Light Detection and Ranging) systems are key three-dimensional space sensors for autonomous driving. As one of the main light source, 905 nm semiconductor lasers with high peak power short optical pulses are usually operated within a wide ambient temperature range. In addition, 905 nm lasers can also be used as light sources of bioinspired sensor [1]. Thus, it is necessary for those laser sources to have high temperature stability. Quantum size effects (QSEs) allow the formation of discrete energy levels in quantum well [2], and electron-hole pairs confined to the energy levels will have enhanced band to band recombination processes, e.g. spontaneous or stimulated radiations. Semiconductor lasers with quantum wells as active layers have many advantages over their previous heterojunction counterparts, such as an increased differential gain and a high characteristic temperature (T0) [3–5]. Thus, the boosted QSE, determined by the depth and width of the finite potential well, will further enhance the beneficial behaviors. In addition to QSE, strains are also present in most current quantum wells [6]. The existence of the compressive strain or tensile strain can separate the light/heavy hole band energy levels, implying a “purer” heavy or light hole valence band state, at the Brillouin zone center in quantum wells [7,8]. The valence band energy levels splitting reduces the intra-valence band mixing, as well as the effective hole density of states at the Brillouin zone center [9–11]. The strain-induced “pure” valence band states can be regarded as a further quantization of the valence band energy levels [12].
For the InGaAs-GaAs material, commonly used quantum wells with a thickness of 7∼12 nm, lasing around 905 nm [13–16], have a relatively low indium (In) content (∼10%) and inherently sustain a weaker compressive strain. In order to further amplify the benefits caused by the compressive strain, as well as the splitting of the valence band energy levels, higher In content (>16%) quantum wells are employed. To compensate for the ground state shrinkage of the high In quantum well, the thickness of the quantum well is reduced from 8 nm to 4 nm to boost QSE to raise the sub-band energy level, so that the lasing wavelength is maintained around 905 nm. Thus, a deep and thin quantum well with a higher compressive strain is achieved. A higher upper limitation which is introduced by the degree of strain can be incorporated in a given InGaAs quantum well. As shown in Fig. 1, laser structures with a 4 nm In0.16Ga0.84As HSQW and an 8 nm In0.10Ga0.90As LSQW, respectively, are implemented as examples. The thickness difference of the N-waveguide is to keep the mode intensity confinement factors of the HSQW and LSQW consistent.
Fig. 1. Refractive index and near field distribution in the vertical direction of laser with HSQW (black) and LSQW (red). The insets show the details of the quantum well region.
A single 4 nm high strain quantum well cannot provide an enough optical gain, so a double 4 nm quantum well is adopted by the HSQW to keep the same effective gain thickness as the LSQW with a single 8 nm quantum well. An asymmetric AlGaAs cladding layer, with a low Al content for the thick N side and a high Al content for the thin P side, are used to ensure the fundamental mode operation. Doping profiles of the two laser structures are also remained identical. Within the above constraints, the laser performance changes can be attributed to the difference in quantum well structure.
2. Gain analysis
The HSQW have different gain behaviors to those of the LSQW because of the modified band structure by the boosted QSE and stronger compressive strain [3,7]. A stronger compressive strain has a relatively small effect on the conduction band of direct-gap semiconductor but a greater effect on the degeneracy splitting of the light-hole (LH) and heavy-hole (HH) valence band maximum. The reduced 4 nm thickness quantum well can introduce considerable splitting between the first and second, HH1 and HH2, heavy-hole sub-band levels while avoiding the deleterious effect associated with the monolayer fluctuation in ultra-thin quantum wells [17]. Thus, the HSQW obtains a “purer” HH1 valence band maximum with a large effective mass perpendicular to the well plane but a LH behavior in the well plane, so affecting its gain behaviors significantly.
The reduced gain spectral width of HSQW in Fig. 2(a) is attributed to the further splitting of degenerate valence sub-bands. The “purer” HH1 valence sub-band structure at the Brillouin zone center promotes the recombination process between conduction band electrons and valence band heavy holes (C1-HH1), and thus, provides a stronger material gain [18]. Fig. 2(b) shows the peak material gain as a function of the carrier density. The curves of HSQW have a steeper gradient due to the promoted C1-HH1 sub-band recombination process for the thinner high strain quantum well [19]. However, the larger sub-band density of states also raises the needed carrier density for the population inversion (material gain > 0). Thus, the peak material gains curves of the HSQW and the LSQW have an intersection at a certain carrier density. The needed carrier density of the HSQW will be larger than that of the LSQW if the required peak material gain is located below the intersection point. Naturally, a smaller carrier density of the HSQW is needed if the required peak material gain is located above the intersection point. With the increased carrier density, LSQW shows a more severe gain compression than HSQW which means the differential gain decreases. Hence, the needed carrier density of the LSQW will sharply increase under higher gain requirements.
Fig. 2. (a) The calculated material gains spectra of HSQW and LSQW with the software [SimuPics3d], at 300 K. The carrier density range are [2.3E18∼5.0E18/cm3] and [1.8E18∼5.0E18/cm3], respectively. (b) Peak material gain vs. carrier density curves in quantum well at different temperatures.
Light powers and spectra of laser diodes with both anti-reflection-coated rear and front facets, prepared in the manner described in Section 3, are shown in Fig. 3. The threshold gain can be expressed as
where Γ is mode intensity confinement factors in the quantum well, Gth is the threshold gain, αi and αm are the internal optical loss and the mirror loss, respectively. The increased αm of laser diodes with both anti-reflection-coated rear and front facets needs to be compensated by a higher threshold gain Gth at a higher carrier density. Besides, the corresponding free-carrier absorption induced internal optical loss αi will also increase [20]. If there is always an acceptable carrier density nth, before the laser diodes saturate or fail, that satisfies the following formulaFig. 3. (a) The front facet light power vs current curves of laser structures with HSQW (HSQW_LD) and LSQW (LSQW_LD). The rear and front facets both were anti-reflection-coated (reflectivity 1.72% and 0.80%, respectively). (b) The corresponding spectra at 1A (green), 2A (blue) and 3A (red). 100 µm lateral stripe width, 2 mm cavity length, 25°C heatsink temperature.
3. Experimental results and discussion
Following the epitaxial growth by metal organic chemical vapor deposition (MOCVD), the wafer was wet etched into ridge laser diodes with a lateral stripe width of 100 µm and 200 µm, respectively. A 2 mm-long Fabry-Pérot cavity was cleaved. The rear and front facets were high- and low-reflection coated (reflectivity >99% and 2%, respectively). All laser diodes were bonded P-side down to COS sub-mount and tests were implemented with a thermo-electric-cooler (TEC) cooled heatsink.
The voltage, the lasing power and the power conversion efficiency (PCE) of the diodes are measured under continuous-wave (CW) bias condition at 25°C heatsink temperature as shown in Fig. 4. At high injection currents, HSQW_LD suffers less power saturation than LSQW_LD and can achieve a higher maximum lasing power. Detailed parameter results are shown in Table 1.
Fig. 4. Light-Voltage-Current (LIV) curves of HSQW_LD and LSQW_LD, as well as efficiency-current curves (blue). (a) lateral stripe width: 100 µm and (b) lateral stripe width: 200 µm.
Table 1. Detailed Parameter Results of Fig. 4
The slope efficiency (SE) is obtained by linear fitting of light-current curve at 1-4 A region. As the injection current increases, the free-carrier absorption and the corresponding internal optical loss will also suffer a significant rise [19], thus leading to a higher threshold gain. The threshold current density jth of a semiconductor laser can be expressed as
where ${d_{QW}}$ is the quantum well thickness, and A, B, C are the Shockley-Read-Hall coefficient, the spontaneous radiation coefficient and the Auger recombination coefficient, respectively. As shown in Fig. 2(b), when the needed threshold gains increases, the corresponding needed threshold carrier density in quantum well will also increase. However, the HSQW_LD has a less needed threshold carrier density rise than that of the LSQW_LD due to the steeper gradient of the peak material gain-carrier density curves. Therefore, an inferable assumption is that a smaller portion of the added operating current is used to compensate for the threshold current rise of HSQW_LD, and thus a higher lasing power with less power saturations is achieved.Characteristic temperatures T0 and T1 can be obtained by fitting the temperature-depended threshold current (Fig. 5(a)) and slope efficiency (Fig. 5(b)), according to the follow equation
where Ta, and Tr are the applied temperature and the reference temperature, respectively. As shown in Fig. 5, the evaluated characteristic temperature T0 of HSQW_LD is as high as 207 K and shows a higher temperature stability in the whole temperature range than that of LSQW_LD which is only 129 K. The realized high T0 of HSQW_LD is comparable to the high strain InGaAs-GaAs quantum well at longer wavelength [21,22], which is also better than the Al-free GaInP large waveguide 905 nm laser diodes with T0 = 175 K [23]. The SE of HSQW_LD also becomes more stable and a relatively consistent T1 can be evaluated in the whole temperature range to be as 556 K. However, the T1 of LSQW_LD undergoes a sudden drop from 550 K to 311 K when the heatsink temperature is over 25°C. The high and stable characteristic temperatures T0 and T1 are direct evidences of a stronger carrier confinement due to the deep band offset and an improved gain process of the HSQW [24,25]. Thus, the quantum well optimized HSQW_LD is more suitable for working under the CW condition than the previous LSQW_LD.Fig. 5. Threshold current Ith (a) and slope efficiency SE (b) of HSQW_LD and LSQW_LD at different heatsink temperature.
Benefit from the significantly reduced gain spectral width, lasing spectra FWHMs of HSQW_LD are always narrower than those of LSQW_LD at different temperatures and bias currents, as shown in Fig. 6. In addition, the FWHM of HSQW_LD shows a very high temperature stability in the whole temperature range, and broadens more slowly than that of LSQW_LD as the bias current increases as well.
Fig. 6. Full width at half maximum (FWHM) of lasing spectra at different temperatures (a) and bias currents (b).
4. Conclusion
High In content of InGaAs-GaAs material system will introduces high compressive strain in the quantum well, but the In content related ground energy is restricted by the lasing wavelength. However, the energy of the quantized sub-band levels is lifted above the ground state by the QSEs. Thus, the wavelength shift, caused by the increased In content, can be compensated by enhancing the QSE and be fixed at a desired range. Guided by the above ideas, the laser structure with a high strain double quantum well (HSQW_LD) is designed in this work. The laser wavelength is fixed around 905 nm by the compensation between high In content and enhanced QSE, which results in higher material gain, in the quantum well. Laser performance has been optimized by the HSQW, especially the maximum lasing power and the characteristic temperature T0. As shown in Table 1, compared with the LSQW_LD, a higher maximum lasing power can be achieved by the HSQW_LD at a lower bias current, which means that the HSQW_LD suffers less power saturations and power losses at high bias currents. Furthermore, the temperature-depended threshold current and slop efficiency curves show that the HSQW_LD has a higher temperature stability. It is meaningful that the T0 of HSQW_LD is as high as 207 K with an increase of >60% compared to that of the LSQW_LD. The above performance improvement is significant with only a slight change on the quantum well. We believe that the optimization method in this work is also suitable for other laser diodes lasing around 900 nm with low In content InGaAs quantum wells and other similar low-strain gain material systems.
Funding
Key Technology Research and Development Program of Shandong (2023ZLYS03, 2022CXGC020104); Key-Area Research and Development Program of Guangdong Province (2020B090922003).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. A. Leal-Junior, L. Avellar, V. Biazi, M. S. Soares, A. Frizera, and C. Marques, “Multifunctional flexible optical waveguide sensor: on the bioinspiration for ultrasensitive sensors development,” Opto-Electron. Adv. 5(10), 210098 (2022). [CrossRef]
2. N. Holonyak, R. Kolbas, R. Dupuis, and P. Dapkus, “Quantum-well heterostructure lasers,” IEEE J. Quantum Electron. 16(2), 170–186 (1980). [CrossRef]
3. M. Asada, A. Kameyama, and Y. Suematsu, “Gain and intervalence band absorption in quantum-well lasers,” IEEE J. Quantum Electron. 20(7), 745–753 (1984). [CrossRef]
4. Y. Arakawa and Y. Amnon, “Theory of gain, modulation response, and spectral linewidth in AlGaAs quantum well lasers,” IEEE J. Quantum Electron. 21(10), 1666–1674 (1985). [CrossRef]
5. N. K. Dutta, “Calculated threshold current of GaAs quantum well lasers,” J. Appl. Phys. 53(11), 7211–7214 (1982). [CrossRef]
6. L. J. Mawst, H. Kim, G. Smith, W. Sun, and N. Tansu, “Strained-layer quantum well materials grown by MOCVD for diode laser application,” Prog. Quantum. Electron. 75, 100303 (2021). [CrossRef]
7. S. L. Chuang, “Efficient band-structure calculations of strained quantum wells,” Phys. Rev. B 43(12), 9649–9661 (1991). [CrossRef]
8. J. J. Coleman, “Strained-layer InGaAs quantum-well heterostructure lasers,” IEEE J. Sel. Top. Quantum Electron. 6(6), 1008–1013 (2000). [CrossRef]
9. D. Ahn, S. J. Yoon, S. L. Chuang, and C.-S. Chang, “Theory of optical gain in strained-layer quantum wells within the 6×6 Luttinger-Kohn model,” J. Appl. Phys. 78(4), 2489–2497 (1995). [CrossRef]
10. S. Seki, T. Yamanaka, W. Lui, Y. Yoshikuni, and K. Yokoyama, “Theoretical analysis of pure effects of strain and quantum confinement on differential gain in InGaAsP/lnP strained-layer quantum-well lasers,” IEEE J. Quantum Electron. 30(2), 500–510 (1994). [CrossRef]
11. A. R. Adams, “Strained-layer quantum-well lasers,” IEEE J. Sel. Top. Quantum Electron. 17(5), 1364–1373 (2011). [CrossRef]
12. D. Ahn and S. L. Chuang, “Optical gain in a strained-layer quantum-well laser,” IEEE J. Quantum Electron. 24(12), 2400–2406 (1988). [CrossRef]
13. S. O. Slipchenko, A. A. Podoskin, A. V. Rozhkov, N. A. Pikhtin, I. S. Tarasov, T. A. Bagaev, V. P. Konyaev, M. A. Ladugin, A. A. Marmalyuk, A. A. Padalitsa, and V. A. Simakov, “A study of nonlinear lasing dynamics of an InGaAs/AlGaAs/GaAs heterostructure power laser-thyristor emitting at 905 nm,” J. Appl. Phys. 116(8), 084503 (2014). [CrossRef]
14. O. S. Soboleva, V. S. Golovin, V. S. Yuferev, P. S. Gavrina, N. A. Pikhtin, S. O. Slipchenko, and A. A. Podoskin, “Modeling the Spatial Switch-On Dynamics of a Laser Thyristor (λ= 905 nm) Based on an AlGaAs/InGaAs/GaAs Multi-Junction Heterostructure,” Semiconductors 54(5), 575–580 (2020). [CrossRef]
15. A. Knigge, H. Christopher, A. Liero, J. Fricke, A. Klehr, T. Honig, E. Langenbach, H. Wenzel, and G. Tränkle, “Wavelength stabilized high pulse power laser bars for line-flash automotive LIDAR,” Proc. SPIE 11262, 112620F (2020).
16. G. Pan, M. Xun, Z. Zhao, Y. Sun, J. Zhou, and D. Wu, “High Slope Efficiency Bipolar Cascade 905 nm Vertical Cavity Surface Emitting Laser,” IEEE Electron Device Lett. 42(9), 1342–1345 (2021). [CrossRef]
17. P. Ei. O’Reilly, “Valence band engineering in strained-layer structures,” Semicond. Sci. Technol. 4(3), 121–137 (1989). [CrossRef]
18. S. W. Corzine, R. H. Yan, and L. A. Coldren, “Theoretical gain in strained InGaAs/AlGaAs quantum wells including valence-band mixing effects,” Appl. Phys. Lett. 57(26), 2835–2837 (1990). [CrossRef]
19. M. W. Prairie and R. M. Kolbas, “A general derivation of the density of states function for quantum wells and superlattices,” Superlattices Microstruct. 7(4), 269–277 (1990). [CrossRef]
20. D. A. Veselov, Y. K. Bobretsova, A. Y. Leshko, V. V. Shamakhov, S. O. Slipchenko, and N. A. Pikhtin, “Measurements of internal optical loss inside an operating laser diode,” J. Appl. Phys. 126(21), 213107 (2019). [CrossRef]
21. J. J. Lee, L. J. Mawst, and D. Botez, “Asymmetric broad waveguide diode lasers (λ = 980 nm) of large equivalent transverse spot size and low temperature sensitivity,” IEEE Photonics Technol. Lett. 14(8), 1046–1048 (2002). [CrossRef]
22. Z. Qiao, X. Li, H. Wang, T. Li, X. Gao, Y. Qu, B. Bo, and C. Liu, “High-performance 1.06-µm InGaAs/GaAs double-quantum-well semiconductor lasers with asymmetric heterostructure layers,” Semicond. Sci. Technol. 34(5), 055013 (2019). [CrossRef]
23. M. Krakowski, M. Meghnagi, P. Afuso-Roxo, P. Resneau, E. Izquierdo, C. Theveneau, E. Vinet, Y. Robert, J. P. Legoec, M. Garcia, O. Parillaud, and B. Gerard, “High optical power pulsed laser structures at 905 nm for automotive LIDAR,” Proc. SPIE 12440, 124400J (2023). [CrossRef]
24. C. Y. Liu, S. F. Yoon, Q. Cao, C. Z. Tong, and Z. Z. Sun, “Comparative analysis of cavity length-dependent temperature sensitivity of GaInNAs quantum dot lasers and quantum well lasers,” Nanotechnology 17(22), 5627–5631 (2006). [CrossRef]
25. L. V. Asryan and Z. N. Sokolova, “Optical power of semiconductor lasers with a low-dimensional active region,” J. Appl. Phys. 115(2), 023107 (2014). [CrossRef]
