Open Access
Add to BibSonomy
Abstract
To elucidate the microscopic origin of the thermal droop, a blue-emitting indium gallium nitride (InGaN) quantum well grown on epitaxially laterally overgrown gallium nitride was investigated using temperature-dependent microphotoluminescence spectroscopy. Below 300 K, the sample exhibited a well-known dislocation-tolerant luminescence behavior. However, as temperature increases from 300 K to 500 K, the near band-edge emission at the wing region (with lower threading dislocation densities) was stronger than that at the seed region (with higher threading dislocation densities), indicating that threading dislocations are the microscopic origin of the thermal droop. Considering the carrier diffusion length, edge-type threading dislocations should play a major role in the thermal droop of heteroepitaxially grown InGaN-based LEDs.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Indium gallium nitride (InGaN) has attracted much attention as an emitting material throughout the visible spectral range. State-of-the-art InGaN-based blue LEDs have achieved a record external quantum efficiency of 84.3 % at 20 mA ($\sim 10\ \textrm {A}/\textrm {cm}^{2}$) at room temperature. [1] However, the external quantum efficiency of InGaN-based LEDs decreases as the current and temperature increase. The former is called the current or efficiency droop, while the latter is known as the thermal or temperature droop. Although the current droop mechanism is well studied, [2–6] the thermal droop mechanism has received less attention. [7] A promising application of InGaN-based LEDs is automotive headlights, where the surrounding temperatures can reach $80\ {}^\circ \mathrm {C}$. Current-induced heating further increases the headlamp and junction temperatures of LEDs to $110\ {}^\circ \mathrm {C}$ and $150\ {}^\circ \mathrm {C}$, respectively. [8,9] In such a situation, the thermal droop has an impact comparable to that of the current droop. [7] Consequently, the origin of the thermal droop must be thoroughly investigated.
Chhajed et al. studied the thermal droop using two InGaN-based LEDs with different threading dislocation densities of $5.3 \times 10^{8}\ \textrm {cm}^{-2}$ and $5.7\times 10^{9}\ \textrm {cm}^{-2}$. [10] Their temperature-dependent electroluminescence (EL) measurements from 293 K to 423 K showed that the former sample had a smaller thermal droop, which was attributed to the lower threading dislocation density. Santi et al. conducted temperature-dependent EL measurements from 83 K to 475 K using five InGaN-based single quantum wells with different point defect densities grown on Si substrates. [11] They estimated the point defect density using capacitance deep level transient spectroscopy, and proposed that the thermal droop is related to the point defect density. However, temperature-dependent EL measurements cannot distinguish between the effect of recombination and transport on the thermal droop. To remove the ambiguity, David et al. performed temperature-dependent photoluminescence (PL) and EL measurements between 298 K and 433 K. [12] They concluded that the thermal droop is dominated by the transport effects in InGaN-based LEDs grown on bulk GaN substrates (The threading dislocation density is $\sim 10^{6}\ \textrm {cm}^{-2}$). In short, the microscopic origin of the thermal droop remains unclear, especially in heteroepitaxially grown InGaN-based LEDs. We attribute this situation to the lack of microscopic optical investigations. Due to the low production cost, heteroepitaxially grown InGaN-based LEDs will continue to play an important role in industrial applications. Herein, temperature-dependent microphotoluminescence spectroscopy is performed for blue-emitting InGaN/GaN single quantum wells grown on epitaxially laterally overgrown GaN to extract the effect of recombination on the thermal droop. Our approach directly shows that threading dislocations are the microscopic origin of the thermal droop in blue-emitting InGaN/GaN quantum wells.
2. Experiments
The sample was an $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well grown on an epitaxially laterally overgrown GaN on a sapphire substrate. The seed (window) and wing regions had a width of 4 $\mu \textrm {m}$ and 16 $\mu \textrm {m}$, respectively. The quantum-well thickness was 3 nm. Figure 1(a) shows the stacking structure of the sample. According to atomic force microscopy, the sums of the screw- and mixed-type threading dislocation densities were $2\times 10^{8}\ \textrm {cm}^{-2}$ in the seed region and $<1\times 10^{7}\ \textrm {cm}^{-2}$ in the wing region. [13] In both the seed and wing regions, the estimated edge-type threading dislocation density was $7-10$ times higher than the sum of the screw- and mixed-type threading dislocation densities.
Fig. 1. (a) Stacking structure and (b) fluorescence microscopy image of a blue-emitting InGaN/GaN single quantum well. The fluorescence image is taken at 300 K. White scale bar denotes 20 $\mu \textrm {m}$. Red dotted square indicates the scanning area of the $\mu$-PL mapping ($50\times 50\ \mu \textrm {m}^{2}$). Dark line is utilized as an optical marker of the $\mu$-PL mapping.
To visualize the microscopic origin of the thermal droop, we developed a microscopic confocal PL apparatus. Our apparatus changed the experimental temperature from 4 K to 500 K. The excitation source, a continuous-wave laser diode with a 405-nm excitation wavelength, selectively excited the quantum well layer. [14] The laser beam was focused onto a sample surface by a long working distance objective with a numerical aperture of 0.42. The sample was placed in a liquid helium cooling cryostat. The luminescence signals were focused onto the cross-slit of a monochromator. The cross-slit acted as a pinhole of a confocal microscope. The lateral resolution of our confocal PL apparatus was approximately 2 $\mu \textrm {m}$. The scanning area of the microphotoluminescence ($\mu$-PL) mapping was $50\times 50\ \mu \textrm {m}^{2}$ with a scanning step of $1\ \mu \textrm {m}$. The excitation power density was set at 2.2 kW/$\textrm {cm}^{2}$. The estimated carrier density is on the order of $10^{17}\ \textrm {cm}^{-3}$, which is not in the current droop regime. [7]
Figure 1(b) shows a fluorescence microscopy image of the blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 300 K, where the scale bar denotes 20 $\mu \textrm {m}$. The excitation wavelength was 365 nm, which corresponded to non-selective excitation conditions. [14] The seed region exhibited darker emissions. The center of the wing region, which corresponded to the coalescence (growth front) region, showed the brightest emission. Because this fluorescence image was neither spectrally dispersed nor taken under selective excitation conditions, the near band-edge emission from an InGaN quantum well layer and the defect-related yellow luminescences (yellow luminescence (YL) band [15]) from GaN and InGaN layers were superimposed. To focus on the near band-edge emission from an InGaN quantum well layer, we performed temperature-dependent $\mu$-PL spectroscopy.
Figure 2 shows typical $\mu$-PL spectra of the blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 300 K in the wing and seed regions. Two luminescence peaks were observed. One was the near band-edge emission from an InGaN quantum well layer with a peak wavelength around 470 nm. The other was the defect-related yellow luminescence with a peak wavelength of $\sim 570$ nm. Because the vertical resolution of our confocal PL apparatus was 5 $\mu \textrm {m}$ at best, we could not conclude whether the defect-related yellow luminescence came from an InGaN quantum well layer or the surrounding layers. The near band-edge emission intensity at the seed region with higher dislocation densities was similar to that at the wing region with lower dislocation densities. Previous studies have also noted the dislocation-tolerant luminescence behavior of blue-emitting InGaN quantum wells at room temperature. [13,16–19] By contrast, the defect-related yellow luminescence intensity in the seed region was much stronger than that in the wing region at 300 K. This observation was consistent with a previous study. [20] Fig. 3 shows $\mu$-PL mapping images of the center wavelength for the blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 4 K, 300 K, 400 K, and 500 K. The center wavelength was defined by $\int \lambda \cdot I(\lambda )d\lambda /\int I(\lambda )d\lambda$, where $\lambda$ and $I(\lambda )$ were the wavelength and PL intensity, respectively. The integration range of the wavelength was $430-670$ nm. The location of an optical marker was identified thanks to Fig. 3. Figure 4 shows $\mu$-PL mapping images of the integrated PL intensity for the blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 4 K, 300 K, 400 K, and 500 K. The wavelength integration was performed on the defect-related yellow luminescence. Figure 4 shows a clear stripe pattern, indicating that the defect-related yellow luminescence intensity in the seed region was stronger than that in the wing region for all experimental temperatures. The less thermal quenching behavior of the defect-related yellow luminescence was reported. [15] The correlation between the defect-related yellow luminescence intensity and edge dislocation density was observed. [21] Our results are in agreement with those of the previous studies. Although we speculate that dislocations may be involved in Ga (In) vacancies and/or those complexes, [15,21] further discussion on the defect-related yellow luminescence is not the main scope of this paper.
Fig. 2. Typical confocal $\mu$-PL spectra of a blue-emitting InGaN/GaN single quantum well at 300 K in the wing and seed regions.
Fig. 3. $\mu$-PL mapping images of the center wavelength for a blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 4 K, 300 K, 400 K, and 500 K. Center wavelength unit is nm. White scale bar denotes 10 $\mu \textrm {m}$.
Fig. 4. $\mu$-PL mapping images of the integrated PL intensity for a blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 4 K, 300 K, 400 K, and 500 K. Integration is performed on the defect-related yellow luminescence. White scale bar denotes 10 $\mu \textrm {m}$. The exposure time are 0.1 s and 0.5 s for 4 K and $300-500$ K, respectively.
Figure 5 shows $\mu$-PL mapping images of the integrated PL intensity for the blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 4 K, 300 K, 400 K, and 500 K. The wavelength integration was performed on the near band-edge emission from the InGaN quantum well layer. It is emphasized that the PL intensity of the near band-edge emission was dislocation-tolerant at 4 K and 300 K. By contrast, the PL intensity in the seed region at 400 K and 500 K was smaller than that in the wing region. These observations provided direct evidence that threading dislocations were the microscopic origin of the thermal droop in blue-emitting InGaN/GaN quantum wells. Figure 5 also indicates that the near band-edge emission from the InGaN quantum well layer is not bright but rather dark at the center of the wing region compared to that at the edge of the wing region. Consequently, the brightest emission at the center of the wing region shown in Fig. 1 should originate from the defect-related yellow luminescence outside the InGaN quantum well layer.
Fig. 5. $\mu$-PL mapping images of the integrated PL intensity for a blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well at 4 K, 300 K, 400 K, and 500 K. Integration is performed on the near band-edge emission of an InGaN quantum well layer. White scale bar denotes 10 $\mu \textrm {m}$. The exposure time are 0.1 s and 0.5 s for 4 K and $300-500$ K, respectively.
3. Discussions
We statistically analyzed the observed thermal droop in the blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well. Figure 6(a) shows the thermal quenching of the near band-edge emission in seed and wing regions. $I^{seed}_{avg.}$ and $I^{wing}_{avg.}$ are the averaged PL intensity of the near band-edge emission at the seed and wing regions, respectively. The error bar of $I^{seed}_{avg.}$ in Fig. 6(a) represents the standard deviation. Figure 6(a) warrants that the differences between $I^{seed}_{avg.}$ and $I^{wing}_{avg.}$ above 300 K are statistically significant.
Fig. 6. Temperature dependence of (a) $I^{seed}_{avg.}$ and $I^{wing}_{avg.}$ and (b) $I^{seed}_{avg.}/I^{wing}_{avg.}$ for a blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well. $I^{seed}_{avg.}$ and $I^{wing}_{avg.}$ are the averaged PL intensity of the near band-edge emission at the seed and wing regions, respectively. Error bar of $I^{seed}_{avg.}$ in Fig. 6(a) represents the standard deviation.
Next, we considered which type of threading dislocation is the dominant origin of the observed thermal droop in blue-emitting InGaN/GaN quantum wells. Table 1 shows the number of threading dislocations in the photoexcitation spot ($\sim 3.1\ \mu \textrm {m}^{2}$) of the sample. As already noted, the edge-type threading dislocation density was $7-10$ times higher than the other threading dislocation densities in the seed and wing regions. Previous studies have reported that the carrier diffusion length in InGaN quantum wells was 60 nm [22] and 200 nm [23] at cryogenic temperatures. The room temperature value was reported to be less than 100 nm [13] or around 100 nm. [24] Generally, the carrier diffusion length in semiconductors decreases as temperature increases. [25] Thus, we concluded that the diffusion length in our sample was less than 200 nm at all the experimental temperatures. Herein we assumed that the carrier diffusion length of our sample was 200 nm (the upper limit) at 400 K and 500 K. The maximal effect of threading dislocations on the PL intensity could be estimated by the product of the number of threading dislocations (Table 1) and the diffusion area ($\pi \times 0.1^{2}\ \mu \textrm {m}^2$). Only for the edge-type threading dislocations at the seed region, the product ($1.9\ \mu \textrm {m}^{2}$) was comparable to the photoexcitation spot area ($3.1\ \mu \textrm {m}^{2}$). Therefore, edge-type threading dislocations should be the dominant origin of the observed thermal droop in blue-emitting InGaN/GaN quantum wells. Screw- and mixed-type threading dislocations should play a minor role considering their lower dislocation densities and shorter carrier diffusion length. Because our experiments were not conducted in the current droop regime, the present results indicate that edge-type threading dislocations induce Shockley-Read-Hall nonradiative recombination processes. However, it should be noted that we cannot distinguish whether the nonradiative recombination center is edge-type dislocation or edge-type dislocation induced point defects. [21] Either way, our results indicate that edge type dislocations cause thermal droop in blue-emitting InGaN/GaN quantum wells.
Table 1. Number of threading dislocations in the photoexcitation spot ($\sim 3.1\ \mu \textrm {m}^{2}$) of a blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well.
Finally, we evaluated the impact of edge-type threading dislocations on the thermal droop in heteroepitaxially grown blue-emitting InGaN/GaN quantum wells. Figure 6(b) shows the temperature dependence of $I^{seed}_{avg.}/I^{wing}_{avg.}$ for a blue-emitting $\textrm {In}_{0.20}\textrm {Ga}_{0.80}\textrm {N}/\textrm {GaN}$ single quantum well. As the edge-type threading dislocation density increased from $\sim 10^{8}\ \textrm {cm}^{-2}$ (wing region) to $\sim 10^{9}\ \textrm {cm}^{-2}$ (seed region), a PL thermal quenching around 15 % was observed at 425 K compared to the $I^{seed}_{avg.}/I^{wing}_{avg.}$ at 300 K. In a previous study, a blue-emitting InGaN-based LED with a threading dislocation density of $5.7\times 10^{9}\ \textrm {cm}^{-2}$ exhibited around 60 % EL thermal quenching from 300 K to 423 K under a low injection regime. [10] Considering our experimental results, about quarter of the EL thermal quenching should be attributed to nonradiative recombinations at edge-type threading dislocations or edge-type threading dislocation induced point defects. Therefore, our results suggest that both transport and recombination effects should be leading factors for the thermal droop in heteroepitaxially grown InGaN-based LEDs (with an edge-type threading dislocation density of $\sim 10^{9}\ \textrm {cm}^{-2}$).
4. Conclusion
In conclusion, a blue-emitting InGaN/GaN quantum well grown on epitaxially laterally overgrown GaN on a sapphire substrate was investigated using temperature-dependent microphotoluminescence spectroscopy. The results confirm that threading dislocations are the microscopic origin of the thermal droop in heteroepitaxially grown InGaN-based LEDs. Considering the dislocation densities and carrier diffusion length, edge-type threading dislocations should play a dominant role in the thermal droop. To suppress the thermal droop of blue-emitting InGaN-based LEDs, the threading dislocation densities should be less than $10^{9}\ \textrm {cm}^{-2}$.
Funding
Japan Society for the Promotion of Science (JP16H06426, JP17H04810, JP19H02615, JP20H05622).
Acknowledgments
The authors thank Z. Zhu for the assistance with the microphotoluminescence experiments.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data is available from the authors upon request.
References
1. Y. Narukawa, M. Ichikawa, D. Sanga, M. Sano, and T. Mukai, “White light emitting diodes with super-high luminous efficacy,” J. Phys. D: Appl. Phys. 43(35), 354002 (2010). [CrossRef]
2. J. Piprek, “Efficiency droop in nitride-based light-emitting diodes,” Phys. Stat. Sol. (a) 207(10), 2217–2225 (2010). [CrossRef]
3. G. Verzellesi, D. Saguatti, M. Meneghini, F. Bertazzi, M. Goano, G. Meneghesso, and E. Zanoni, “Efficiency droop in InGaN/GaN blue light-emitting diodes: Physical mechanisms and remedies,” J. Appl. Phys. 114(7), 071101 (2013). [CrossRef]
4. Y. Kawakami, A. Kaneta, A. Hashiya, and M. Funato, “Impact of Radiative and Nonradiative Recombination Processes on the Efficiency-Droop Phenomenon in InxGa1−xN/GaN Single Quantum Wells Studied by Scanning Near-Field Optical Microscopy,” Phys. Rev. Appl. 6(4), 044018 (2016). [CrossRef]
5. A. David, N. G. Young, C. Lund, and M. D. Craven, “Review—The Physics of Recombinations in III-Nitride Emitters,” ECS J. Solid State Sci. Technol. 9(1), 016021 (2020). [CrossRef]
6. C. Weisbuch, “Review—On The Search for Efficient Solid State Light Emitters: Past, Present, Future,” ECS J. Solid State Sci. Technol. 9(1), 016022 (2020). [CrossRef]
7. M. Meneghini, C. D. Santi, A. Tibaldi, M. Vallone, F. Bertazzi, G. Meneghesso, E. Zanoni, and M. Goano, “Thermal droop in III-nitride based light-emitting diodes: Physical origin and perspectives,” J. Appl. Phys. 127(21), 211102 (2020). [CrossRef]
8. S. Chhajed, Y. Xi, Y. L. Li, T. Gessmann, and E. F. Schubert, “Influence of junction temperature on chromaticity and color-rendering properties of trichromatic white-light sources based on light-emitting diodes,” J. Appl. Phys. 97(5), 054506 (2005). [CrossRef]
9. See https://compoundsemiconductor.net/article/97529-lasers-light-the-road-ahead.html
10. S. Chhajed, J. Cho, E. F. Schubert, J. K. Kim, D. D. Koleske, and M. H. Crawford, “Temperature-dependent light-output characteristics of GaInN light-emitting diodes with different dislocation densities,” Phys. Stat. Sol. (a) 208(4), 947–950 (2011). [CrossRef]
11. C. D. Santi, M. Meneghini, M. L. Grassa, B. Galler, R. Zeisel, M. Goano, S. Dominici, M. Mandurrino, F. Bertazzi, D. Robidas, G. Meneghesso, and E. Zanoni, “Role of defects in the thermal droop of InGaN-based light emitting diodes,” J. Appl. Phys. 119(9), 094501 (2016). [CrossRef]
12. A. David, N. G. Young, C. Lund, and M. D. Craven, “Thermal droop in high-quality InGaN LEDs,” Appl. Phys. Lett. 115(22), 223502 (2019). [CrossRef]
13. A. Kaneta, M. Funato, and Y. Kawakami, “Nanoscopic recombination processes in InGaN/GaN quantum wells emitting violet, blue, and green spectra,” Phys. Rev. B 78(12), 125317 (2008). [CrossRef]
14. R. Ishii, A. Yoshikawa, K. Nagase, M. Funato, and Y. Kawakami, “265nm AlGaN-based deep-ultraviolet light-emitting diodes grown on AlN substrates studied by photoluminescence spectroscopy under ideal pulsed selective and non-selective excitation conditions,” Appl. Phys. Express 13(10), 102005 (2020). [CrossRef]
15. M. Reshchikov and H. Morkoç, “Luminescence properties of defects in GaN,” J. Appl. Phys. 97(6), 061301 (2005). [CrossRef]
16. S. Chichibu, T. Azuhata, T. Sota, and S. Nakamura, “Spontaneous emission of localized excitons in InGaN single and multiquantum well structures,” Appl. Phys. Lett. 69(27), 4188–4190 (1996). [CrossRef]
17. Y. Narukawa, Y. Kawakami, M. Funato, Sz. Fujita, and Sg. Fujita, “Role of self-formed InGaN quantum dots for exciton localization in the purple laser diode emitting at 420 nm,” Appl. Phys. Lett. 70(8), 981–983 (1997). [CrossRef]
18. T. Mukai, K. Takekawa, and S. Nakamura, “InGaN-Based Blue Light-Emitting Diodes Grown on Epitaxially Laterally Overgrown GaN Substrates,” Jpn. J. Appl. Phys. 37(Part 2, No. 7B), L839–L841 (1998). [CrossRef]
19. A. Cieślińska, L. Marona, J. Koziorowska, S. Grzanka, J. Weyher, D. Schiavon, and P. Perlin, “Role of dislocations in nitride laser diodes with different indium content,” Sci. Rep. 11(1), 21 (2021). [CrossRef]
20. M. Sugiyama, T. Shioda, Y. Tomita, T. Yamamoto, Y. Ikuhara, and Y. Nakano, “Optical and Structural Characterization of InGaN/GaN Multiple Quantum Wells by Epitaxial Lateral Overgrowth,” Mater. Trans. 50(5), 1085–1090 (2009). [CrossRef]
21. D. G. Zhao, D. S. Jiang, H. Yang, J. J. Zhu, Z. S. Liu, S. M. Zhang, J. W. Liang, X. Li, X. Y. Li, and H. M. Gong, “Role of edge dislocations in enhancing the yellow luminescence of n-type GaN,” Appl. Phys. Lett. 88(24), 241917 (2006). [CrossRef]
22. S. Chichibu, K. Wada, and S. Nakamura, “Spatially resolved cathodoluminescence spectra of InGaN quantum wells,” Appl. Phys. Lett. 71(16), 2346–2348 (1997). [CrossRef]
23. D. Cherns, S. J. Henley, and F. A. Ponce, “Edge and screw dislocations as nonradiative centers in InGaN/GaN quantum well luminescence,” Appl. Phys. Lett. 78(18), 2691–2693 (2001). [CrossRef]
24. M. Mensi, R. Ivanov, T. K. Uždavinys, K. M. Kelchner, S. Nakamura, S. P. DenBaars, J. S. Speck, and S. Marcinkevičius, “Direct Measurement of Nanoscale Lateral Carrier Diffusion: Toward Scanning Diffusion Microscopy,” ACS Photonics 5(2), 528–534 (2018). [CrossRef]
25. C. Netzel, V. Hoffmann, J. W. Tomm, F. Mahler, S. Einfeldt, and M. Weyers, “Temperature-Dependent Charge Carrier Diffusion in [000$\bar{1}$] Direction of GaN Determined by Luminescence Evaluation of Buried InGaN Quantum Wells,” Phys. Stat. Sol. (B) 257(6), 2000016 (2020). [CrossRef]
