Open Access
Add to BibSonomy
Abstract
The optical emission of four InGaN/InGaN light-emitting diode samples with double-quantum wells (DQW) with varying indium content in quantum barrier (QB) layers are investigated by high resolution X-ray diffraction, temperature-dependent photoluminescence excited by a 325-nm He-Cd laser and a 405-nm semiconductor laser, photoluminescence microscopy, and by atomic force microscope. We demonstrate that the introduction of a small amount of indium in quantum barriers can improve the uniformity and change the luminescence properties of quantum wells. It is found that the InxGa1-xN DQW has the best uniformity when the indium content x in QB layers is 0.15%.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
InGaN/InGaN double-quantum wells (DQW) are distinctly important for the GaN-based emitting devices, such as light-emitting diodes (LEDs) [1–3] and laser diodes (LDs) [4–6], which have extensive applications in solid state lighting, light communication and for mini projector currently [7,8]. Studies on the DQW diodes and related devices have obtained great achievements in recent decades. The main luminescence mechanism of InGaN/InGaN QWs is related to the fact that the injected carriers can be confined in quantum-dot-like regions, making contribution to the carrier localization effect. The localization is caused by the composition difference of the InGaN alloy [9] or the thickness fluctuation [10,11] in the well layers, which can reduce the negative influence of the threading dislocations on non-radiative recombination of carriers and improve the luminescence efficiency of GaN-based emitting devices [12]. It has been demonstrated that the radiative recombination of localized excitons dominates in the luminescence processes in InGaN alloys. Whereas, there are still some problems related to the luminescence mechanism not completely solved, and further studies on the photoluminescence characteristics of localized states are therefore necessary to understand the luminescence mechanism better. For example, Zheng et al reported that the increment of TMIn flow rate deteriorates the interfacial abruptness and decreases the nonradiative recombination efficiency based on results of temperature-dependent photoluminescence and high-resolution X-ray diffraction measurements [13]. However, Kuo et al [14], Xiong et al [15] and Chang et al [16] reported that device performance using InGaN quantum barrier is better over the conventional GaN quantum barrier due to the enhancement of electron confinement and the reduction of barrier height for the holes to transport in the active region according to the calculation results. In these investigations, the indium content in quantum barrier is equal or larger than 0.02, but a few works have investigated how is the influence of small indium content in quantum barriers on the uniformity and luminescence properties of quantum wells. In this work, four InGaN/InGaN LED samples with different indium content of quantum barrier layers are prepared by metal-organic chemical vapor deposition system (MOCVD). Their structure and emission properties are characterized in detail by high resolution X-ray diffraction (HRXRD), temperature-dependent photoluminescence (TDPL) excited by a 325-nm He-Cd laser and a 405-nm semiconductor laser, photoluminescence microscopy and atomic force microscope. It is found that introducing proper indium in QB layers is beneficial to improve the uniformity of DQW.
2. Methods
2.1 Materials
Four light emitting diodes (LEDs) with different InGaN/InGaN DQW are grown on c-plane sapphire substrate in an AIXTRON 3×2 in. Close-coupled showerhead reactor. Trimethylgallium (TMGa), trimethylindium (TMIn), and ammonia (NH3) are used for the epitaxial growth as Ga, In, and N source precursors, respectively. The structure of DQW samples is consisted of a GaN template layer, a 2-µm thick Si-doped n-type GaN layer(n-GaN), two periods of unintentionally doped InGaN/InGaN DQW and a120-nm thick Mg-doped p-type GaN layer(p-GaN). The n-GaN, DQW and p-GaN layers are grown at 1040°C, 750 °C and 1040°C, respectively. During growth of DQW, TMIn flow rate of quantum barrier (QB) layer of four different samples is varied, i.e. 0 mL/min 10 mL/min, 15 mL/min and 20 mL/min, which are named as samples A-D, respectively. The schematic structure of these four LED samples with DQW is shown in Fig. 1.
2.2 Characterization
High resolution X-ray diffraction is used to characterize the DQW structural parameters and material quality by using Rigaku Ultima IV X-ray diffractometer with Cu-Ka radiation (λ = 1.54 Å). Moreover, temperature dependent photoluminescence measurements are employed to characterize the DQW optical properties, where the temperature is controlled by a closed-cycle refrigerator of CTI Cryogenics in the range from 30 to 300 K. Two kinds of excitation sources, i.e. a 325-nm He-Cd laser and a 405-nm semiconductor laser, are used to take TDPL measurements. In addition, photoluminescence-microscopy is performed to characterize the emission properties of the DQW samples in a micro-scale by using a confocal optical system with an excitation laser wavelength of 405 nm (Nikon A1). Moreover, atomic force microscope (AFM, Bruker Icon) measurements are taken to check the surface topography of the first quantum barrier (FQB) in samples.
To quantitatively check the luminescence characteristics of quantum wells of samples A-D, a model of localized-state ensemble (LSE) is used to simulate the dependence of luminescence peak energy on the temperature in TDPL measurements and analyze the distribution of localized states responsible for the QW luminescence. LSE model is based on an assumption of distribution function for localized carriers, developed by Q. Li et al [17]. Meanwhile, LSE model can also evolve into the well-known band-tail model [18–21] at high temperature and into the luminescence quenching model of a two-level system when the distribution of localized states approaches a δ-function. In LSE model, the relation of peak energy and temperature can be described as following equations [19]:
Based on the LSE model, $\frac{{{\tau _r}}}{{{\tau _{tr}}}}$ is defined as $\tau $, which shows the escape rate of carriers in localized states. A smaller value of $\tau $ means deeper localized level and harder escaping of carriers. Meanwhile, a smaller value of $\sigma $ means more concentrated Gaussian distribution of the localized states. The distribution of the localized states can reflect the dispersion of energy levels of these states. Thus, these two parameters are characteristic parameters of depth and uniformity of luminescence centers in QWs. The influence of indium content of quantum barriers on the uniformity of quantum wells is discussed in details based on the analysis of $\tau $ and $\sigma $ for samples A-D.
3. Results and discussion
3.1 HRXRD measurements
After epilayer growth, HRXRD measurements are used to check the structure of samples A-D. In Fig. 2, the (0002) Omega-2theta (ω-2θ) scan curves of samples A-D are depicted in black lines, and the fitting lines are depicted in red lines. The substrate peaks originate from GaN (0002) plane and three superlattice (SL) satellite peaks, i.e., -1st, -2nd and -3rd order SL peaks, can be observed clearly for samples A-D. The structural parameters of InGaN/GaN DQW of samples A-D can be obtained by fitting the measured curves. The material parameters, with certain limited error, may be obtained by the fitting of XRD curves. They show a clear difference between 4 samples. As shown in Table 1, the thickness of barrier layers does not have any remarkable change, but their indium content increases with increasing TMIn flow and is 0%, 0.15%, 0.24% and 0.41% for samples A-D, respectively.
Fig. 2. HRXRD Omega-2theta curves on GaN (0002) plane for samples A-D. The dark and red lines are the measurement data and fitting data, respectively.
Table 1. Structural parameter of InGaN/InGaN DQW of samples A-D determined by HRXRD measurements.
3.2 TDPL with a 325-nm He-Cd laser
It is known that luminescence property of InGaN/InGaN DQW LED is related to the localized states, and TDPL is an effective method to investigate localized states in quantum wells. Therefore, TDPL measurements of samples A-D excited by 325-nm He-Cd laser are taken to investigate how indium content of QB layer affects the localized states in quantum wells. Figure 3 shows the TDPL results for samples A-D. It can be seen that the peak energy position of these PL spectra is changed when the temperature increases from 30 K to 300 K. It is well known that the joint action of the thermal redistribution of localized carriers and the shrinkage of the band gap are the cause of the variation of luminescence peak energy with temperature. In detail, the thermal redistribution of carriers in localized states and the shrinkage of the band gap can result in blue or red shift of the PL peak energy with increasing temperature. Therefore, the PL spectra are fitted by Gaussian curves to find the peak energy, and the dependence of peak energy on the temperature is discussed in details below.
Fig. 3. TDPL spectra of samples A (a), B (b), C (c) and D (d) at a temperature range of 30-300 K excited by 325-nm He-Cd laser.
The circular dots of Fig. 4 show the results of TDPL measurement data for samples A-D excited by 325-nm He-Cd laser. For samples A-C, the peak energy increases first and then decreases with increasing temperature. It means the peak energy of samples A-C shows a blue shift first and then red shift. On the other side, peak energy of sample D always increases when the temperature increase from 30 K to 300 K. It suggests that for sample D the influence of the thermal redistribution of localized carriers is greater than the influence of the shrinkage of the band gap, so there is only a monotonous change of peak energy with increasing temperature. This may be attributed to the reason that the distribution of the localized states levels of sample D is wide and carriers redistribute into shallower localized states when the temperature increases. According to the HRXRD measurements, the indium content of QB layer of sample D is the largest among four samples. It suggests that the inhomogeneous distribution of indium composition, i.e. the uniformity of DQW, is enhanced when the indium content of QB layer increases to 0.41%.
Fig. 4. PL emission peak energy as a function of temperature for samples A (a), B (b), C (c) and D (d) excited by 325-nm He-Cd laser. The solid blue lines are theoretical fitting curves using LSE model and the circle dots are the experimental data.
To quantitatively check how indium content influences on the localized states in quantum wells, the temperature dependence of the luminescence peak of samples A-D excited with 325-nm He-Cd laser is analyzed. Fitting lines in LSE model of samples A-D are plotted in blue lines in Fig. 4, and the fitting parameters are listed in Table 2. First, for samples A-D, the value of $\tau $ is 0.0009, 0.1850, 0.2306 and 0.0074, respectively. It indicates that the carriers in sample A are hardest to escape out from the localized states. It means that the localized states of sample A are distributed in a deeper level (deep localization states). Second, for samples A-D, their standard deviation of Gaussian-type state density distribution of the localized states ($\sigma $) is 47.0 meV, 35.5 meV, 37.5 meV and 220.5 meV, respectively. It indicates that the distribution of the localized states of sample D is much wider than that of samples A, B and C. Therefore, combining the value of $\tau $ and $\sigma $, it is suggested that the localization state level in the quantum wells of sample A is the deepest and the distribution of localization states in the quantum wells of sample D are most inhomogeneous.
Table 2. Fitting parameters of PL peak energy in the LSE model for samples A-D excited by 325-nm He-Cd laser.
It is known that in general the localization states can be divided in two different types: deeper ones and shallower ones, which can be caused by the indium composition or thickness fluctuations in quantum wells. On one side, the strain and the piezoelectric field in the DQW decrease along with the increase of indium content of QB layers, because the difference of indium content between well and barrier layers is reduced, and the fluctuations are weakened. Thus, the uniformity of quantum wells of samples B and C is better than that of sample A. However, on the other side, it is also noted that the distribution of localization states in the quantum wells of sample D is widest and most inhomogeneous, which may be caused by weakening interface quality as more indium is introduced into the quantum barriers. It indicates that QB layers with relatively fewer indium may be beneficial to growing homogeneous quantum wells.
3.3 TDPL with a 404-nm semiconductor laser
The TDPL of samples A-D are also measured by using 40-mW 405 nm semiconductor laser excitation to double check the influence of InGaN quantum barriers on the uniformity of quantum wells. The experimental data of TDPL measurements and fitting lines by LSE model of samples A-D are plotted in Fig. 5 in square dots and blue lines, respectively. The fitting parameters in LSE model are listed in Table 3. It can be seen that $\tau $ of samples A-D is 0.0074, 0.0089, 0.0001 and 0.0013, respectively. In addition, $\sigma $ is 300.0 meV, 58.0 meV, 79.0 meV and 152.0 meV, respectively. These data show that the carriers of sample B are easier to escape out from the localized states, and the uniformity of quantum wells of sample A is the worst.
Fig. 5. PL emission peak energy as a function of temperature for samples A (a), B (b), C (c) and D (d) excited by 405-nm semiconductor laser with 40-mW incident optical power. The solid blue lines are theoretical fitting curves using LSE model, and the squares are the experimental data.
Table 3. Fitting parameters in the LSE model for samples A-D excited by 405-nm semiconductor laser with 40-mW incident optical power.
Furthermore, it is also noted that $\sigma $ of samples B is 35.5 meV and 58.0 meV, respectively, when the excited source is a 325 nm laser or a 405-nm laser. It indicates that the $\sigma $ values and Gaussian-type state density distribution of the localized states of samples B are smallest among these four sample, and the $\tau $ value of sample B is the largest when the excited source is 405-nm laser It indicates that the uniformity of quantum wells in sample B is the best.
According to the analysis of $\tau$ and σ in Table 3, the uniformity of quantum wells in sample B is the best and it is consistent with result from Table 2. There are two reasons for the discrepancy in the values of E0, Ea, α, $\tau$ and σ in the two tables. First, incident optical power between a 325-nm He-Cd laser and a 405-nm semiconductor laser is different, and it results in different concentration of photo-generated carriers. Second, besides the quantum wells, 325-nm laser can excite carriers from other layers in the DQW structure, including quantum barriers. However, 405-nm laser can only excite carriers from the QW layer. Thus, the number of photon-generated carriers will be different, thus the localized states participating in the carrier recombination will be varied. Therefore, peak energy of the TDPL spectra will be different, and the fitting parameters in Tables 2 and 3 are also different.
To further check the uniformity of quantum wells in sample B and sample C, TDPL spectra excited by 405-nm laser with a lower incident optical power of 1 mW are measured. In Fig. 6, the experiment data and fitting lines in LSE mode are depicted in square dots and blue lines, respectively, and the fitting parameters in LSE mode are listed in Table 4. It can be seen that the shapes of peak energy with increasing temperature of samples B and C are completely different, which is “S” shape and “reversed V-shape”, respectively. Meanwhile, $\tau $ of sample B is about 15000 times larger than that of sample C, and $\sigma $ of samples B is only almost one third of sample C. It indicates that compared to samples C, the carriers in sample B are easier to escape out from the localized states in quantum wells, and the distribution of the related localized states is more concentrated. It also demonstrates that the uniformity of quantum wells in sample B is better than that of sample C, and their luminescence efficiency is also better.
Fig. 6. PL emission peak energy as a function of temperature for samples B (a) and C (b) excited by 405-nm semiconductor laser with 1-mW incident optical power. The solid blue lines are theoretical fitting curves using LSE model, and the squares are the experimental data.
Table 4. Fitting parameters of PL peak energy in the LSE model for samples B and C excited by 405-nm semiconductor laser with 1-mW incident optical power.
3.4 Photoluminescence microscopy and atomic force microscope results
Moreover, it is known that photoluminescence microscopy can be used to check the uniformity through the difference of luminescence intensity in the luminescence micro-image. Thus, photoluminescence microscopy measurements of four samples A-D are employed to check the uniformity of quantum wells, and white and red circle is used to mark the bigger and smaller dark region, respectively. In Fig. 7, it is clear that the size and quantity of dark regions in photoluminescence microscopy image of sample B is the smallest. On the other side, there are many larger dark regions in sample A. Compare sample B, the size and quantity of dark regions in samples C and D increase. It indicates that the difference of luminescence intensity in the photoluminescence microscopy image of the sample B is smallest which means the uniformity of quantum wells in sample B is best.
Fig. 7. Photoluminescence microscopy images of samples A-D, the bigger and smaller dark region are marked in white and red circle, respectively.
Furthermore, to confirm the influence of small indium content in the quantum barrier, atomic force microscopy measurements and related analysis are added in the revised manuscript. Four other samples are prepared to take AFM measurements. All of these four samples consist of a GaN template layer, a n-GaN layer and an undoped GaN layer (u-GaN). The growth conditions of the u-GaN layer of these four samples are the same as the first quantum barrier (FQB) in samples A-D, respectively. In Fig. 8, the atomic force microscopy images indicate that, in a 5 µm × 5 µm region of FQB in samples A-D, the root-mean-square (RMS) roughness is 0.86 nm, 0.43 nm,0.85 nm and 0.89 nm, respectively.
In addition, it is noted that v-pits are formed in the first quantum barrier (FQB) of samples A-D during growth. Thus, the distributions of v-pit depth in the “FQB in samples A-D” are plotted in Fig. 9. It is found that the average depth of v-pits in the first quantum barrier for the sample B is lowest, and the density of v-pits in the “FQB of the sample B” is small as shown by the images of Figs. 8 and 9. These results indicate that the quality of InGaN FQB of sample B with small indium content is best. It means that that surface topography of the FQB could be improved by introducing small amount of indium. However, it will be deteriorated when the indium content is too large. Thus, results of the atomic force microscopy measurements demonstrate that surface topography of the FQB could be improved by introducing small indium, but it will be deteriorated when the indium content is larger.
Therefore, TDPL, photoluminescence microscopy and AFM measurements demonstrate that introducing proper indium content in quantum barriers can result in a good uniformity of quantum wells. It is suggested that the strain and the piezoelectric field in the quantum wells have been reduced after introducing indium in QB layers and the material quality of quantum barriers has been improved due to catalysis of indium atoms during the film growth. Thus, the uniformity of quantum wells in sample B is better than that of sample A. On the other side, the interface quality becomes worse as too more indium is introduced into the quantum barriers, which can weaken the uniformity of quantum wells. Thus, the uniformity of quantum wells in samples C and D becomes worse than that of sample B.
4. Conclusions
We take a detailed analysis of illumination properties of InGaN/InGaN double-quantum wells with varying indium content in quantum barrier layers by TDPL measurements, HRXRD, photoluminescence microscopy and AFM. It is found that when introducing a small amount of indium in quantum barriers, and keeping the indium content a proper value, a good uniformity of quantum wells can be obtained.
Funding
National Key R&D Program of China (2018YFB0406903, 2016YFB0400801, 2016YFB0400803); National Natural Science Foundation of China (61574134, 61574135, 61604145, 61674138, 61674139).
References
1. M. Kneissl, T. Y. Seong, J. Han, and H. Amano, “The emergence and prospects of deep-ultraviolet light-emitting diode technologies,” Nat. Photonics 13(4), 233–244 (2019). [CrossRef]
2. A. David, N. G. Young, C. A. Hurni, and M. D. Craven, “Quantum Efficiency of III-Nitride Emitters: Evidence for Defect-Assisted Nonradiative Recombination and its Effect on the Green Gap,” Phys. Rev. Appl. 11(3), 031001 (2019). [CrossRef]
3. T. D. Moustakas and R. Paella, “Optoelectronic device physics and technology of nitride semiconductors from the UV to the terahertz,” Rep. Prog. Phys. 80(10), 106501 (2017). [CrossRef]
4. L. R. Jiang, J. P. Liu, A. Q. Tian, Y. Cheng, Z. C. Li, L. Q. Zhang, S. M. Zhang, D. Y. Li, M. Ikeda, and H. Yang, “GaN-based green laser diodes,” J. Semicond. 37(11), 111001 (2016). [CrossRef]
5. S. Masui, Y. Nakatsu, D. Kasahara, and S. Nagahama, “Recent improvement in nitride lasers,” Proc. SPIE 10104, 101041H (2017). [CrossRef]
6. D. G. Zhao, J. Yang, Z. S. Liu, P. Chen, J. J. Zhu, D. S. Jiang, Y. S. Shi, H. Wang, L. H. Duan, Liqun Zhang, and Hui Yang, “Fabrication of room temperature continuous-wave operation GaN-based ultraviolet laser diodes,” J. Semicond. 38(5), 051001 (2017). [CrossRef]
7. Y. Nakatsu, Y. Nagao, K. Kozuru, T. Hirao, E. Okahisa, S. Masui, T. Yanamoto, and S. Nagahama, “High-efficiency blue and green laser diodes for laser displays,” Proc. SPIE 10918, 109181D (2019). [CrossRef]
8. F. Liang, J. Yang, D. G. Zhao, Z. S. Liu, P. Chen, J. J. Zhu, D. S. Jiang, Y. S. Shi, H. Wang, L. H. Duan, Liqun Zhang, and Hui Yang, “Room-temperature continuous-wave operation of GaN-based blue-violet laser diodes with a lifetime longer than 1000 h,” J. Semicond. 40(2), 022801 (2019). [CrossRef]
9. Y. Narukawa, Y. Kawakami, M. Funato, S. Fujita, S. Fujita, and S. Nakamura, “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]
10. D. M. Graham, A. Soltani-Vala, P. Dawson, M. J. Godfrey, T. M. Sexton, J. S. Barnard, M. J. Kappers, C. J. Humphreys, and E. J. Thrush, “Optical and misconstruction studies of InGaN∕GaN single-quantum-well structures,” J. Appl. Phys. 97(10), 103508 (2005). [CrossRef]
11. N. K. van der Laak, R. A. Oliver, M. J. Kappers, and C. J. Humphreys, “Role of gross well-width fluctuations in bright, green-emitting single InGaN∕GaN quantum well structures,” Appl. Phys. Lett. 90(12), 121911 (2007). [CrossRef]
12. S. Nakamura, “The Roles of Structural Imperfections in InGaN-Based Blue Light-Emitting Diodes and Laser Diodes,” Science 281(5379), 956–961 (1998). [CrossRef]
13. X. H. Zheng, H. Chen, Z. B. Yan, H. B. Yu, D. S. Li, Y. J. Han, Q. Huang, and J. M. Zhou, “Effects of TMIn flow rate of barrier layer on the optical and structural properties of InxGa1-xN/InyGa1-yN multiple quantum wells,” J. Cryst. Growth 257(3-4), 326–332 (2003). [CrossRef]
14. Y. K. Kuo, J. Y. Chang, M. C. Tsai, and S. H. Yen, “Advantages of blue InGaN multiple-quantum well light-emitting diodes with InGaN barriers,” Appl. Phys. Lett. 95(1), 011116 (2009). [CrossRef]
15. J. Y. Xiong, S. W. Zheng, and G. H. Fan, “Performance Enhancement of Blue InGaN Light-Emitting Diodes With InGaN Barriers and Dip-Shaped Last Barrier,” IEEE Trans. Electron Devices 60(11), 3925–3929 (2013). [CrossRef]
16. J. Y. Chang, Y. A. Chang, F. M. Chen, Y. T. Kuo, and Y. K. Kuo, “Improved Quantum Efficiency in Green InGaN Light-Emitting Diodes With InGaN Barriers,” IEEE Photonics Technol. Lett. 25(1), 55–58 (2013). [CrossRef]
17. Q. Li, S. J. Xu, M. H. Xie, and S. Y. Tong, “A model for steady-state luminescence of localized-state ensemble,” Europhys. Lett. 71(6), 994–1000 (2005). [CrossRef]
18. Y. H. Cho, G. H. Gainer, A. J. Fischer, J. J. Song, S. Keller, U. K. Mishra, and S. P. DenBaars, “S-shaped temperature-dependent emission shift and carrier dynamics in InGaN/GaN multiple quantum wells,” Appl. Phys. Lett. 73(10), 1370–1372 (1998). [CrossRef]
19. P. G. Eliseev, P. Perlin, J. Lee, and M. Osinski, “Blue temperature-induced shift and band-tail emission in InGaN-based light sources,” Appl. Phys. Lett. 71(5), 569–571 (1997). [CrossRef]
20. K. G. Zolina, V. E. Kudryashov, A. N. Turkin, A. E. Yunovich, and S. Nakamura, “Luminescence Spectra Of Superbright Blue and Green InGaN/AlGaN/GaN Light-Emitting Diodes,” MRS Internet J. Nitride Semicond. Res. 1, e11 (1996). [CrossRef]
21. Q. Li, S. J. Xu, W. C. Cheng, M. H. Xie, S. Y. Tong, C. M. Che, and H. Yang, “Thermal redistribution of localized excitons and its effect on the luminescence band in InGaN ternary alloys,” Appl. Phys. Lett. 79(12), 1810–1812 (2001). [CrossRef]
