Open Access
Add to BibSonomy
Abstract
Performance of InGaN-based blue-violet laser diodes (LD) with different waveguide structure were investigated by simulation and experimental methods. Theoretical calculation demonstrated that threshold current (Ith) can be reduced and slope efficiency (SE) can be improved by using an asymmetric waveguide structure. Based on the simulation results, a LD with 80-nm-thick In0.03Ga0.97N lower waveguide (LWG) and 80-nm-thick GaN upper waveguide (UWG) is fabricated with flip chip package. Under continuous wave (CW) current injection at room temperature, its optical output power (OOP) reaches 4.5 W at an operating current of 3 A and the lasing wavelength of 403 nm. The threshold current density (Jth) is 0.97 kA/cm2 and the SE is about 1.9 W/A.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
InGaN-based laser diodes (LD) have attracted great attentions due to their potentials in many applications, such as material processing, 3D printing, optical atomic clock, high-density and high-speed optical data storage, etc. [1–5] Although they have been commercially available, there is still large room for the improvement. The room for further improvement is mainly concentrated in two aspects: one is to reduce threshold current density (Jth) which makes device more suitable for low power consumption applications and the other is to increase slope efficiency (SE) which makes device more suitable for high optical output power (OOP) applications [4]. There are some common ways to improve the two aspects, such as improving optical field distribution to improve optical confinement factor (OCF) and reduce internal loss (IL) [6].
From the perspective of epitaxy structure design, a lot of research works have been reported on improving the optical field distribution by optimizing the structure of each layer, e.g., employing the plasmonic GaN substrate to reduce the optical mode leakage into the substrate [7], utilizing higher refractive index InGaN/InGaN multiple quantum well (MQW) to directly improve OCF [8], adjusting the position of electron blocking layer (EBL) to reduce optical loss [9], designing p-AlGaN cladding layer (CL) to reduce optical field proportion in the p-type side region [10]. The most significant work remains in the optimization of the waveguides, which is the key to achieving separate confinement structure for quantum well LDs [11–14]. The most commonly used waveguide structure is still symmetric structure, i.e., the upper waveguide (UWG) and lower waveguide (LWG) are basically the same. Among these efforts, the most remarkable work came from Panasonic corporation [13], who achieved a blue-violet laser preparation with 7.2 W light output power under CW operation through symmetric 80 nm In0.08Ga0.92N and 500 nm additional GaN thick waveguide structure. This result is the record OOP for InGaN-based blue-violet LD and has not been broken since 2016, but its Jth (1.6 kA/cm2) is believed to have room for further reduction. In fact, there are a large number of asymmetric properties in the InGaN material system. For instance, there is a difference between the mobility of electron and hole, and n-type doping is much easier compared to p-type doping. This intrinsic nature inspires us to adopt an asymmetric waveguide structure to cater for the asymmetry in the material. By using an asymmetric waveguide structure, the higher optical loss due to p-type layer can be reduced by shifting the optical field closer to the n-side, meanwhile, the hole injection can be improved by reducing the carrier loss in the p-side.
In this article, we started with a theoretical study of the waveguide layer structure design for the internal optical field distribution and output characteristics of the LD. A symmetric waveguide structure with both InGaN and GaN was modified into an asymmetric waveguide structure composed of InGaN LWG and GaN UWG. Finally, based on the asymmetric waveguide structure, a blue-violet LD with a light output power of 4.5 W is fabricated with a low Jth of 0.97 kA/cm2 under continuous wave (CW) operation at room temperature.
2. Device simulation and fabrication
The LD structures taken for simulation study consist of a 1-µm-thick n-type GaN layer, 800-nm-thick n-type Al0.07Ga0.93N CL (n-CL), LWG layer, an unintentional doped InGaN/GaN MQW active region with 6% In content of InGaN, UWG layer, a 20-nm-thick p-type Al0.2Ga0.8N EBL, a 600-nm-thick p-type Al0.7Ga0.93N CL (p-CL), a 50-nm-thick p-type GaN layer and a 20-nm-thick heavily p-type doped GaN contact layer. An initial structure for LWG and UWG is symmetric and consist of both In0.01Ga0.99N layer and GaN layer, where In0.01Ga0.99N layer is closer to the active region. The final optimized waveguide structure is asymmetric and consist of 80-nm-thick In0.03Ga0.97N LWG and 80-nm-thick GaN UWG layers. The structures of these two LDs can be seen in Fig. 1. All theoretical calculation on LD performances were carried out with the help of LASTIP program (Crosslight Software), which is a powerful two-dimensional semiconductor LD simulator [15–17]. The ridge size of all LDs was set as 3 µm × 600 µm and the screening factor of polarization was taken as 0.25. Some other parameter settings and the detailed models, including band-gap energy, ionization energy for doping element, absorption coefficient, spontaneous and piezoelectric polarization and the refractive index of ternary alloys, can be referred to our previous work [6].
Fig. 1. Schematic diagram of the epitaxial structure of LDs with symmetric waveguide structure and with asymmetric waveguide structure.
Two LDs (LD1 and LD2) were grown on c-plane free-standing GaN substrate by metal-organic chemical vapor deposition (MOCVD). The structure parameters of LDs are the same as the simulation of the initial one (with symmetric waveguide) and the final optimized one (with asymmetric waveguide). During growth, trimethylgallium (TMG), triethylgallium (TEG), trimethylindium (TMI), trimethylaluminum (TMA) and ammonia (NH3) were used as the precursors of Ga, In, Al, and N, respectively. Silane (SiH4) and bis-cyclopentadienyl magnesium (Cp2Mg) were used as n-type and p-type dopants. Hydrogen (H2) and nitrogen (N2) were used as carrier gases. After growth, a 30-µm-wide ridge stripe was formed by laser scribing along the <$1\bar{1}00$> direction with silicon dioxide (SiO2) as electron blocking layer, and a 1200-µm-long cavity was fabricated by cleaving along the <$11\bar{2}0$> direction. The Ti/Pt/Au and Pd/Pt/Au were used as n-type and p-type metal electrodes, respectively. The front and rear cleaved cavity facets were coated with high reflectivity (HR) of 90% and anti-reflectivity (AR) of 10% films in order to improve the output power. It is noted that the actual cavity width and cavity length of the fabricated laser is longer than the simulation setting, which results in a lower Jth obtained from the experiments. However, this does not change the inherent physical pattern on output characteristics between the simulated samples with different waveguide structures, so the actual fabrication process of waveguide structure design can refer to the simulation results.
3. Results and discussions
3.1 Comparison between symmetric and asymmetric waveguide structure
Our first attempt was to adopt symmetrical waveguide structures of InGaN/GaN composite waveguide layers. The thickness of InGaN layer in UWG was adjusted from 0 to 160 nm and the thickness of other layers in waveguide were fixed at 80 nm. The simulation results are plotted in Fig. 2. It is found that using InGaN layer in UWG has only a small performance improvement (higher OOP and lower Ith values) at the thinner thickness of 20 nm. In Fig. 2 (b), OCF shows a maximum value at the thicker thickness of 80 nm, while IL decreases with the increase of thickness. These changes can be analyzed according to the optical field distribution curves which are shown in Fig. 3 (a). When no InGaN UWG is introduced, the center of the optical field falls within the InGaN LWG, but not in the MQW due to the higher refractive index of the InGaN/GaN LWG compared to the GaN UWG. As the thickness of the InGaN UWG layer increases to 20 nm, the optical field center shifts into the MQW region. When the thickness of the InGaN UWG continues to increase and reaches 80 nm, the waveguide structure is completely symmetrical, so that the center of the optical field is almost at the center of the MQW. Meanwhile, OCF reaches a maximum value. Then the center of the optical field shifts toward the p-type region with a further increasing of InGaN UWG thickness, but it remains at MQW region. From the view of full width at half maxima (FWHM) of the optical field shown in Fig. 3 (b), under the influence of InGaN UWG, FWHM decreases first compared to the case of UWG without InGaN. This is attributed to the larger difference in refractive index between UWG and p-CL, leading to less light to enter p-CL and a better optical confinement in MQW and WG, the proportions of optical field in p-CL and UWG are also shown in Fig. 3 (b). As the incomplete ionization of the p-type region results in higher absorption loss, and the optical field distribution in p-CL decreases with increasing InGaN UWG thickness, the IL also decreases as shown in Fig. 2 (b). It can be seen that although the light is reduced in the p-CL, the light generated from MQW will more enter the high refractive index InGaN UWG with increasing thickness, resulting in an increase in FWHM and a corresponding decrease in OCF. Since both Ith and OOP are closely related to OCF and IL [18], we thought that it should be reasonable to find the best performance of LD happening at the thickness of 80 nm, i.e., when the LD has a completely symmetrical waveguide structure. However, in fact, there is really a small improvement of the performance at 20 nm, which shows that there are other factors affecting the output characteristics of LD.
Fig. 2. (a) Simulated OOP at the injecting current of 120 mA and threshold current (Ith) versus thickness of InGaN UWG; (b) OCF and IL versus the thickness of InGaN UWG.
Fig. 3. (a) Optical filed distribution variation with different thickness values of InGaN UWG layer (0, 20, 40, 60, 80, 120 and 160 nm). The inset shows position of the center of the optical field relative to the active region (cyan); (b) the FWHM of the optical field (black), proportion of optical field in p-CL (blue) and UWG (red) versus the thickness of InGaN UWG layer.
In general, the carrier injection efficiency will also affect the output performance [6,13]. The carriers distributions and hole current density distribution were depicted in Fig. 4. From the perspective of carriers concentration distributions, as the thickness of InGaN UWG increases from 0 to 160 nm, the electron concentration in UWG increases as shown in Fig. 4 (a), indicating that the electron leakage will become more serious, and the increase of electron overflow leads to an increase of residual holes in UWG as shown in Fig. 4 (b). The increase of residual holes in UWG induces two results: on the one hand, it leads to the decrease of holes injected into quantum wells, i.e., the decrease of hole injection efficiency, which can be seen in Fig. 4 (c). On the other hand, the Shockley-Read-Hall (SRH) recombination, spontaneous radiative recombination (SRR) and Auger recombination in UWG will all increase, as shown in Fig. 5. The former two recombination rates are in the same order of magnitude and the lower rate of Auger recombination is only discussed when the carrier concentration is much higher (the peak of the Auger recombination rate is 10−8 cm-3/s). These undesired recombination processes also cause the loss of holes, which induces a lower hole injection efficiency. The reduced injection efficiency causes the output characteristics of the symmetric structure to be not as good as expected, although the symmetric structure is better in terms of optical field distribution (higher OCF and lower IL). Under the joint effect of OCF, IL and carrier injection efficiency, the LD has the best output characteristics when the thickness of InGaN UWG is 20 nm.
Fig. 4. Electron (a) and hole (b) concentration distribution around the MQW region with different UWG layer thickness values of 0, 20, 40, 60, 80, 120 and 160 nm, respectively. (c) the hole current density distribution in the active region.
Fig. 5. (a) the SRH recombination rate distribution, (b) SRR rate distribution, (c) Auger recombination rate distribution around the MQW region for different thickness values of InGaN UWG layer of 0, 20, 40, 60, 80, 120 and 160 nm, respectively.
Considering that the use of InGaN in UWG has a limited performance improvement, the more complex structure will cause problems in actual growth, such as the formation of non-radiative recombination centers (for instance, our previous work showed the formation of V-pits [19], which will cause degradation of laser output performance), therefore, it is believed that the InGaN alloy is not necessarily to be used in the UWG. At the same time, it is also found that the use of GaN LWG has no positive influence effect on the improvement of output characteristics (not shown here). Hence, the waveguide structure is finally chosen as an asymmetric one composed of InGaN LWG and GaN UWG.
3.2 Optimization process of asymmetric waveguide structure
The effect of the thickness of InGaN LWG and GaN UWG is investigated first. The thickness of InGaN LWG was firstly fixed at 80 nm and the In concentration was kept at 1%. The impacts of the thickness of GaN UWG are plotted in Fig. 6. As shown in Fig. 6 (a), the best output characteristics appear when the GaN UWG thickness is 80 nm. Due to the higher refractive index of InGaN LWG, the distribution of overall optical field is biased toward the n-type region. As the thickness of GaN UWG increases, the optical field gradually shifts toward the center of the MQW region, so the OCF continues to rise. However, when the optical field shifts toward the p-type region, its overall movement is less than the increase of GaN UWG layer thickness. Therefore, more light enters the UWG waveguide and the light in the p-CL is reduced, inducing a less IL. Since the increase of GaN UWG thickness may also cause an increase of overflow carriers, the injection efficiency decreases, resulting in a decrease of device output characteristics, i.e., OOP decrease and Ith increases when the thickness further increases above 80 nm. The effect of the thickness of InGaN LWG on the LD performance and optical field is also investigated and analyzed. The variation of InGaN LWG thickness has an effect similar to what the thickness of GaN UWG induces. The simulation results are plotted in Fig. 7. The OOP and Ith reach the maximum when the thickness of InGaN LWG is 80 nm. Unlike the case of varying the thickness of GaN UWG, an excessively thick InGaN LWG will pull the center of optical field away from the active region, causing a decrease in OCF. It is also found that the best output performance appears when the thickness of InGaN LWG is taken as 80 nm.
Fig. 6. (a) and (b) Simulated OOP, Ith, OCF and IL in dependence on the thickness of GaN UWG layer. The injecting current is 120 mA; (c) The optical field distribution with the thickness of GaN UWG changing from 40 to 160 nm. The inset shows position of the center of optical field relative to active region; (d) The proportion of optical field in p-CL and UWG at different thickness of GaN UWG layer.
Fig. 7. (a) Simulated OOP at the injecting current of 120 mA and Ith versus the thickness of InGaN LWG; (b) the optical field distribution with different thickness of InGaN LWG. Note that the left side of the horizontal coordinate is the p-type side. The inset shows the position of the center of optical field relative to active region; (c) OCF and IL at different thickness of InGaN LWG layer.
Subsequently, the influence of In component of the InGaN LWG is studied and the simulation results are shown in Fig. 8. From the view of optical field distribution, the center of the optical field shifts toward the n-type region with higher In component, and by the time the component reaches 3%, the center deviates out of the MQW, but the OCF continues to increase, somewhat different from the phenomenon observed in Fig. 7. This is mainly because as the In component increases, more light enters the InGaN LWG with a fixed thickness (instead of an increasing thickness), leading to a constant decrease in the FWHM of the optical field. By the way, InGaN has a smaller material band gap, and due to the increased In component, the InGaN LWG forms a potential well with GaN QB and n-CL, which accumulates carriers (shown in Fig. 9 (a) and (b)) and causes a slight decrease in carrier injection efficiency, even causing a leakage hole current at the In component of 5% as shown in Fig. 9 (c). The decrease of the carrier injection and the hole leakage will lead to a best output performance when the In component is 3%.
Fig. 8. (a) Simulated OOP at the injecting current of 120 mA and Ith versus the In content in InGaN LWG layer; (b) The optical field distribution when In content in InGaN LWG layer changes from 1% to 5%. The inset shows the position of the center of optical field relative to active region; (c) OCF and IL when In content in InGaN LWG layer changes from 1% to 5%.
Fig. 9. Electron (a) and hole (b) concentration distribution in the InGaN LWG and MQW; (c) Hole current density distribution in the MQW region. The leakage hole current occurs when the In content in InGaN LWG layer reaches 5% as marked by black arrow.
3.3 Fabrication of the LDs and their performances
Based on the simulation results, two blue-violet LDs (LD1 and LD2) were fabricated on free-standing GaN substrate. The optimized asymmetric waveguide structure (InGaN LWG and GaN UWG) was used in LD1, while in LD2 a symmetric waveguide (InGaN/GaN LWG and GaN/InGaN UWG) was used. The selected parameters of waveguide structures and corresponding LD fabrication processes for LD1 and LD2 are taken as what have been described in the above-mentioned simulation and fabrication chapters. Figure 10 (a) shows the measured output light power before package as a function of the forward direct current (P-I curves) under pulsed current injection with a pulse duration of 500 ns and a repetition frequency of 10 kHz. As expected, it is demonstrated that in comparison with LD2, the threshold current density (Jth) of LD1 decreases from 1.05 kA/cm2 to 0.83 kA/cm2 and the slope efficiency (SE) of LD1 increases from 0.9 W/A to 1.6 W/A. The Jth is reduced by 21% and SE is increased by 77%. The result shows that when the waveguide structure changes from the symmetric one to the optimized asymmetric one, the optical performances of LD have been remarkably improved.
Fig. 10. (a) P-I curves of two LDs before package under the pulsed operation. LD1 uses optimized asymmertric waveguide structure while LD2 uses symmertric waveguide strucuture. (b) P-I-V curves of LD1 after package under the CW operation. The insets show the lasing wavelength and the WPE versus the applied current.
The CW performance of LD1 is investigated further. The chip was packaged with p-side down on the sub-mount and P-I-V curves of LD1 is shown in Fig. 10 (b). At room temperature, the measured Jth is 0.97 kA/cm2 and SE is 1.9 W/A. The maximum OOP reaches 4.5 W at an operating current of 3 A and the lasing wavelength is 403 nm. The highest wall plug efficiency (WPE) is about 31.0% at an injecting current of 1.55 A. It is worth mentioning that the SE after package (1.9 W/A) is higher compared to that before package (1.6 W/A). This is mainly because the detector is located closer to the LD front edge during CW testing condition and more light can be received by the detector. Some other reports about the performance of InGaN-based blue-violet LDs were listed in Table 1 (the method to improve LD output performance is not limited to waveguide designing). The most remarkable work came from Panasonic corporation [13], who achieved a blue-violet laser preparation with 7.2 W light output power under CW operation. Its SE is 2.5 W/A and Jth is about 1.6 kA/cm2. Compared to that work, it seems that our LD with asymmetric waveguide structure would have an advantage with respect to reduce the Jth. As can be seen from our simulation results, although the asymmetric waveguide structure makes the optical field biased towards the n-side, which results in a reduction of OCF, it also reduces the hole consumption in the UWG and improves the hole injection. Therefore, the LD with asymmetric structure is able to exhibit lower Jth. The reduction in Jth can also be seen in the results of our tests on unpacked LDs.
Table 1. A list of previous reports on the performance of InGaN-based blue-violet LDs at the lasing wavelength around 405 nm
4. Conclusion
In summary, output performance of blue-violet LD with different waveguide structures are investigated by theoretical simulation method. It is found that LD with an asymmetric waveguide structure (with optimized layer thickness and In component) may possess higher OCF, carrier injection efficiency and lower IL than the LD with a complex symmetrical waveguide structure. Therefore, lower threshold current and higher OOP are exhibited. Finally, an LD with high output power of 4.5 W at an operating current of 3 A and the wavelength of 403 nm in CW mode at room temperature is fabricated by using the optimized asymmetric waveguide structure and a flip chip package. The Jth is 0.97 kA/cm2 and SE is 1.9 W/A.
Funding
Youth Innovation Promotion Association (2019115); National Natural Science Foundation of China (61904172, 61974162, 62034008, 62074142, 62074140, 62250038); Beijing Municipal Science and Technology Project (Z161100002116037, Z211100007921022); Beijing Nova Program (202093); Strategic Priority Research Program of Chinese Academy of Sciences (XDB43030101); Shanxi-Zheda Institute of Advanced Materials and Chemical Engineering (2022SX-TD016).
Disclosures
The authors declare that there are no conflicts of interest related to this article
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. S. P. Najda, S. Stanczyk, A. Kafar, P. Perlin, T. Suski, L. Marona, S. Grzanka, K. Gibasiewicz, P. Wisniewski, R. Czernecki, D. Schiavon, and M. Leszczynski, “Tapered waveguide high power AlGaInN laser diodes and amplifiers for optical integration and quantum technologies,” in Conference on Quantum Information Science and Technology III(Warsaw, POLAND, 2017).
2. J. Yang, D. G. Zhao, Z. S. Liu, F. Liang, P. Chen, B. B. Wang, Y. H. Zhang, and Z. Z. Zhang, “Regulating absorption loss and carrier injection efficiency in ultraviolet laser diodes by changing waveguide layer structure,” Opt. Laser Technol. 156, 108574 (2022). [CrossRef]
3. T. Czyszanowski, S. Stanczyk, A. Kafar, and P. Perlin, “Optical optimization of InGaN/GaN edge-emitting lasers with reduced AlGaN cladding thickness,” Jpn. J. Appl. Phys. 53(3), 032701 (2014). [CrossRef]
4. M. T. Hardy, D. F. Feezell, S. P. DenBaars, and S. Nakamura, “Group III-nitride lasers: a materials perspective,” Mater. Today 14(9), 408–415 (2011). [CrossRef]
5. J. Yang, D. G. Zhao, Z. S. Liu, F. Liang, P. Chen, L. H. Duan, H. Wang, and Y. S. Shi, “A 357.9 nm GaN/AlGaN multiple quantum well ultraviolet laser diode,” J. Semiconductors 43, 010501 (2022). [CrossRef]
6. Z. Z. Zhang, J. Yang, D. G. Zhao, F. Liang, P. Chen, and Z. S. Liu, “Theoretical Optical Output Power Improvement of InGaN-Based Violet Laser Diode Using AlGaN/GaN Composite Last Quantum Barrier,” Nanomaterials 12(22), 3990 (2022). [CrossRef]
7. P. Perlin, K. Holc, M. Sarzynski, W. Scheibenzuber, L. Marona, R. Czernecki, M. Leszczynski, M. Bockowski, I. Grzegory, S. Porowski, G. Cywinski, P. Firek, J. Szmidt, U. Schwarz, and T. Suski, “Application of a composite plasmonic substrate for the suppression of an electromagnetic mode leakage in InGaN laser diodes,” Appl. Phys. Lett. 95(26), 261108 (2009). [CrossRef]
8. Y. H. Ben, F. Liang, D. G. Zhao, J. Yang, Z. S. Liu, and P. Chen, “The role of InGaN quantum barriers in improving the performance of GaN-based laser diodes,” Optics and Laser Technology 145, 107523 (2022). [CrossRef]
9. P. Chen, M. X. Feng, D. S. Jiang, D. G. Zhao, Z. S. Liu, L. Li, L. L. Wu, L. C. Le, J. J. Zhu, H. Wang, S. M. Zhang, and H. Yang, “Improvement of characteristics of InGaN-based laser diodes with undoped InGaN upper waveguide layer,” J. Appl. Phys. (Melville, NY, U. S.) 112(11), 113105 (2012). [CrossRef]
10. Y. Hou, D. Zhao, F. Liang, Z. Liu, J. Yang, and P. Chen, “Enhancing the efficiency of GaN-based laser diodes by the designing of a p-AlGaN cladding layer and an upper waveguide layer,” Opt. Mater. Express 11(6), 1780–1790 (2021). [CrossRef]
11. G. MuzioŻ, H. Turski, M. Siekacz, M. Sawicka, P. Wolny, C. Cheze, G. Cywiński, P. Perlin, and C. Skierbiszewski, “Waveguide Design for Long Wavelength InGaN Based Laser Diodes,” Acta Phys. Pol. A 122(6), 1031–1033 (2012). [CrossRef]
12. G. Alahyarizadeh, Z. Hassan, S. M. Thahab, F. K. Yam, and A. J. Ghazai, “Performance characteristics of deep violet InGaN DQW laser diodes with InGaN/GaN superlattice waveguide layers,” Optik 125(1), 341–344 (2014). [CrossRef]
13. M. Kawaguchi, O. Imafuji, S. Nozaki, H. Hagino, S. Takigawa, T. Katayama, and T. Tanaka, “Optical-loss suppressed InGaN laser diodes using undoped thick waveguide structure,” in Gallium Nitride Materials and Devices XI (International Society for Optics and Photonics2016), pp. 974818.
14. F. Liang, D. G. Zhao, D. S. Jiang, Z. S. Liu, J. J. Zhu, P. Chen, J. Yang, W. Liu, S. T. Liu, Y. Xing, L. Q. Zhang, W. J. Wang, M. Li, Y. T. Zhang, and G. T. Du, “Performance enhancement of the GaN-based laser diode by using an unintentionally doped GaN upper waveguide,” Jpn. J. Appl. Phys. 57(7), 070307 (2018). [CrossRef]
15. J. Yang, D. G. Zhao, J. J. Zhu, Z. S. Liu, D. S. Jiang, P. Chen, F. Liang, S. T. Liu, and Y. Xing, “Effect of Mg doping concentration of electron blocking layer on the performance of GaN-based laser diodes,” Appl. Phys. B 125(12), 235 (2019). [CrossRef]
16. P. Chen, D. G. Zhao, D. S. Jiang, J. J. Zhu, Z. S. Liu, L. C. Le, J. Yang, X. Li, L. Q. Zhang, J. P. Liu, S. M. Zhang, and H. Yang, “The effect of composite GaN/InGaN last barrier layer on electron leakage current and modal gain of InGaN-based multiple quantum well laser diodes,” Phys. Status Solidi A 212(12), 2936–2943 (2015). [CrossRef]
17. F. Liang, D. G. Zhao, D. S. Jiang, Z. S. Liu, J. J. Zhu, P. Chen, J. Yang, W. Liu, S. T. Liu, Y. Xing, L. Q. Zhang, W. J. Wang, M. Li, Y. T. Zhang, and G. T. Du, “Output light power of InGaN-based violet laser diodes improved by using a u-InGaN/GaN/AlGaN multiple upper waveguide,” Chinese Phys. B 26(12), 124210 (2017). [CrossRef]
18. J. Yang, B. B. Wang, D. G. Zhao, Z. S. Liu, F. Liang, P. Chen, Y. H. Zhang, and Z. Z. Zhang, “Realization of 366 nm GaN/AlGaN single quantum well ultraviolet laser diodes with a reduction of carrier loss in the waveguide layers,” J. Appl. Phys. (Melville, NY, U. S.) 130(17), 173105 (2021). [CrossRef]
19. F. Liang, D. G. Zhao, D. S. Jiang, W. J. Wang, Z. S. Liu, J. J. Zhu, P. Chen, J. Yang, and L. Q. Zhang, “Performance deterioration of GaN-based laser diode by V-pits in the upper waveguide layer,” Nanophotonics 9(3), 667–674 (2020). [CrossRef]
20. C. Zeng, S. M. Zhang, L. Ji, H. B. Wang, D. G. Zhao, J. J. Zhu, S. Z. Liu, D. S. Jiang, Q. Cao, M. Chong, L. H. Duan, H. Wang, Y. S. Shi, S. Y. Liu, H. Yang, and L. H. Chen, “Room-Temperature Continuous-Wave Operation of InGaN-Based Blue-Violet Laser Diodes with a Lifetime of 15.6 Hours,” Chinese Phys. Lett. 27(11), 114215 (2010). [CrossRef]
