Open Access
Add to BibSonomy
Abstract
The influence of growth interruption on the surface and luminescence properties of AlGaN/GaN ultraviolet multi-quantum wells (UV MQWs) is investigated. It is found that when the well and barrier layers of MQW samples are continuously grown at the same temperature, they have lower edge dislocation density and flatter surface of MQWs compared to samples with interrupted well and barrier growth. Moreover, continuous growth of well and barrier layers is more conducive to improving the luminescence efficiency of MQWs. This phenomenon is attributed to more impurity incorporation induced by the growth interruption, while a continuous growth of well and barrier can reduce surface diffusion and migration processes of atoms, reducing the defects and surface roughness of MQWs. In addition, the continuous growth of well and barrier can better control the reaction between Al and N atoms, avoiding the formation of excessively high Al content AlGaN at the well/barrier interface, thus improving the luminescence of UV MQWs.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
AlGaN-based semiconductor materials have the advantages of high electron mobility, good thermal stability, and adjustable bandgap, making them ideal materials for manufacturing ultraviolet light-emitting diodes (UV-LEDs) and ultraviolet laser diodes (UV-LDs) [1–3]. AlGaN-based semiconductor UV-LEDs and LDs with multiple quantum wells (MQWs) structure can be widely used in fields such as ultraviolet curing, sterilization, microbial detection and photolithography, having great market potential [4–7]. As the core light-emitting region of UV LEDs and LDs, the quality of AlGaN/GaN UV MQWs directly affects the performance and application of the devices [8]. Luminescence efficiency and crystal quality are important parameters to characterize the quality of UV MQWs [9]. Actually, the large lattice mismatch between AlGaN and GaN produces a large number of mismatch dislocations during epitaxial growth, especially for AlGaN with high Al composition, exacerbating the defect and impurity densities. These defects and impurities can serve as non-radiative recombination centers, leading to a decrease in the luminescence efficiency of AlGaN/GaN MQWs [10]. Furthermore, the interface defects and interface states of AlGaN/GaN cause scattering and losing of electrons and photons, which affects carrier transport and leads to degradation of the crystal quality, thus degrading the performance of UV LEDs and LDs [11]. In addition, the epitaxial growth process of the metal-organic chemical vapor deposition (MOCVD) is very complex, and the changes in growth conditions such as growth temperature, growth rate, and carrier gas flow rate can cause changes in the migratory ability of Al and Ga atoms, which affects the surface morphology and the interface quality of AlGaN/GaN, and thus may lead to deterioration in the optoelectronic characteristics of the devices [12,13]. Several reports have proposed methods to improve the quality of AlGaN/GaN MQWs. For example, Craven et al. reported the growth of nonpolar (11$\bar{2}$0) a-plane GaN thin films with planar surfaces on (1$\bar{1}$02) r-plane sapphire substrates by MOCVD, then they grew 10-period a-plane and c-plane GaN/AlGaN MQW structures on a GaN/sapphire template layer and studied the PL emission characteristics versus the GaN quantum well width. They found that optimal photoluminescence was obtained for wide nonpolar GaN quantum wells in comparison to the thin optimal well width for c-plane structures [14,15]. Wang et al. grew ultraviolet AlGaN/GaN MQWs on low-dislocation-density AlN/sapphire templates, and optimized H2 purge time to reduce the defect density and improve the interface quality of the MQWs [16]. Li et al. prepared AlGaN/GaN MQWs with smooth surfaces and low dislocation densities using atomic layer deposition [2]. Son et al. utilized the localized surface plasmon effect of Al nanoring patterns to improve the luminescence efficiency of AlGaN/GaN MQWs [17]. However, these results mainly focused on the quantum well width, H2 purge time and template layer near the MQWs. The effect of well and barrier growth interruption on the performance of MQWs was rarely reported.
In this paper, we propose an epitaxial growth method to improve the performance of AlGaN/GaN UV MQWs, where the barrier and well layers are continuously grown at the same temperature. The influences of growth interruption between AlGaN quantum barriers and GaN quantum wells as well as growth temperature modulation process on the morphology and luminescence properties of UV MQWs are investigated. The related physical mechanisms are analyzed. It is found that the defect density of MQWs decreases, the surface roughness value reduces, and the luminescence efficiency increases significantly when the well and barrier are grown continuously at the same temperature.
2. Experiment
Three AlGaN/GaN ultraviolet MQWs samples were grown on (0001) sapphire substrate by MOCVD with the same structure but different growth processes, denoted as sample A, sample B, and sample C. Trimethylaluminum (TMAl), trimethylgallium (TMGa), and ammonia (NH3) were used as reactant source materials to provide Al, Ga, and N, respectively. H2 is used as the carrier gas. The structure of all samples along the epitaxial growth sequence includes a 500 nm thick AlN stress regulating layer, a 2µm thick AlGaN template layer with Al composition of 11%, and three pairs of AlGaN/GaN QWs. Figure 1 shows the growth processes of AlGaN quantum barrier (QB) layers and GaN quantum well (QW) layers for the three samples. The growth temperatures of the AlGaN QBs and GaN QWs of sample A were 1070°C and 1060°C, respectively, and a 20s pause was made after growing the QB and before the GaN QW growth. The growth temperature of both well and barrier layers of sample B is 1070°C, and other growth conditions were the same as sample A, i.e. there is a pause of 20s between the well and barrier growth. In addition, sample C has the same MQW growth temperature as sample B, but instead of an interruption after growing the AlGaN QB, the metal-organic sources are continuously introduced to grow GaN QW.
Fig. 1. Schematic diagram of the growth process of AlGaN quantum barrier (QB) and GaN quantum well (QW) for three samples.
The high resolution X-ray diffraction (HRXRD) was used to evaluate the crystal quality of the three samples. The surface morphology of the samples was observed by atomic force microscope (AFM). The interfacial morphology and elemental distribution of the samples were analyzed by spherical aberration corrected transmission electron microscopy (TEM) and energy dispersive X-ray (EDS). Furthermore, the luminescence properties of three samples were investigated by room-temperature photoluminescence (PL) spectra using a He-Cd laser with wavelength of 325 nm as an excitation source.
3. Results and discussion
Defect densities are an important indicator for evaluating the quality of crystals. The higher defect density affects the conductivity and carrier mobility of crystals, increases the absorption loss of light, and leads to a degradation of the optical and electrical properties of AlGaN/GaN UV MQWs [18]. It is known that XRD is a commonly used non-destructive method for characterizing crystal defects, by which the rocking curve of a material is measured and thus the defect density of the material is analyzed [19]. However, due to the limitation of the equipment, the rocking curve is not suitable for measuring materials with very thin thickness. Considering that the thickness of AlGaN/GaN MQWs (only a few tens of nanometers) is too thin and the defect density of the 2µm-thick AlGaN template layer may well reflect the AlGaN/GaN MQWs, we measure the rocking curve of the AlGaN template layer of three samples. Figure 2(a) and (b) show the rocking curves of three samples at (002) and (102) planes. It can be seen that the difference in full width at half maximum (FWHM) of (002) is not significant for the three samples, while the difference in FWHM of (102) is significant. In GaN-based materials, the screw dislocations and edge dislocations are usually calculated by measuring the FWHM of the (002) and (102) planes [20,21]:
where Ds and DE are the screw dislocation density and the edge dislocation density, respectively, and b1 and b2 are the modes of the Bernoulli vectors for the two types of dislocations, which are 0.5185 nm and 0.3189 nm, respectively. β and θ are the FWHM of the (002) and (102) plane peaks, respectively, and φ is the inclination angle of (102) relative to (002) plane. According to the above equations, the density of screw dislocations of samples A, B, and C is calculated as 4.64 × 107 cm−2, 6.95 × 107 cm−2, 5.87 × 107 cm−2, and the density of edge dislocations as 9.23 × 108 cm−2, 7.07 × 108 cm−2, 5.89 × 108 cm−2, respectively. It is shown that the edge dislocation densities of all three samples are one order of magnitude higher than the screw dislocation density. In fact, the threading edge dislocations have a greater impact on the device performance of GaN-based materials. The edge dislocation density of the AlGaN template layer of sample C is lower compared to the other two samples A and B, which implies that the crystal quality of sample C is better.Fig. 2. Rocking curves of the AlGaN layer at (a) (002) and (b) (102) planes for the three samples A, B and C.
Generally, the surface of the epitaxially grown MQW layers should be relatively flat on the AlGaN template layer of lower dislocation density [22]. To verify the surface characteristics of the three samples, the surface morphology of their AlGaN/GaN MQWs is measured by AFM, as shown in Fig. 3. Both sample A and sample B exhibit obvious step flow mode growth with step bunching, with root mean square roughness (Rq) of 0.49 nm and 0.33 nm, respectively. In contrast, sample C has a flatter surface with a lower Rq of only 0.26 nm. The reason for this phenomenon is attributed to changes in growth conditions. The AlGaN QB and GaN QW layers of sample A are grown at different temperatures, while sample B and sample C are grown at the same temperature. The growth of well and barrier layers at the same temperature helps to improve the interface matching between AlGaN and GaN, reducing the stress and dislocation density caused by lattice mismatch. At the same time, in this case the QB and QW growth rates can be kept consistent, which in turn reduces the formation of dislocations. So the edge dislocation density and surface roughness of sample A are the highest among the three samples. In addition, for sample B it stops for 20s after growing QB and before growing QW, while the QW of sample C is grown directly after growing QB. Interruption of growth can lead to an increased incorporation or desorption of impurities, resulting in interface defects. In contrast, the continuous growth of QW layer can reduce the surface diffusion and migration processes of atoms, avoiding impurity incorporation, and thus reducing defects and inhomogeneity on the surface of MQWs. Furthermore, the absence of growth interruption between QW and QB in sample C prevents the rearrangement of Al and Ga atoms during the long pause time, thus improving the flatness of the MQW surface. Therefore, the crystal quality of sample C is better than the other two samples, which is consistent with the above-mentioned observations.
Subsequently, the optical properties of three samples are studied by photoluminescence spectra. The room temperature PL spectra of samples A, B, and C are shown in Fig. 4(a), (b), and (c), respectively. It is found that all samples exhibit PL peaks at around 3.66 eV, 3.03 eV, and 2.44 eV, as marked by the black arrows. The peak at 3.66 eV originates from the AlGaN template layer, while the blue luminescence band at 3.03 eV and the yellow luminescence band at 2.44 eV may be related to defects such as dislocations and vacancies [23,24]. Notably, a luminescence peak appears near 3.39 eV in sample C, as marked by the red arrow in the figure, but not in the other two samples. This peak is attributed to the AlGaN/GaN MQW. The result indicates that the optical performance of the MQWs of samples A and B is poorer, while that of sample C is relatively better. Then, why do the MQWs of sample A and sample B emit weaker or even no corresponding PL peak is observed at room temperature? In order to find the reason, we investigate interface properties of their AlGaN/GaN MQWs as well as Al and Ga elemental distributions.
Fig. 4. Room temperature photoluminescence (PL) spectra of the three samples A, B and C. A distinguishing MQW PL peak is observed in sample C at nearly 3.39 eV.
Figures 5(a)-(c) show the cross-sectional STEM images of the AlGaN/GaN MQWs for samples A, B, and C, and Figs. 5(d)-(f) depict the corresponding EDS surface scans of the Al-element atoms. It can be seen from Fig. 5(a, b and c) that three straight dark line stripes appear in both sample A and sample B, while three bright stripes appear in sample C. The large differences in width, brightness and relative position of these stripes indicate that the thickness and layer composition of well layers in MQWs of the three samples are really quite different. Due to the different growth conditions of samples A, B, and C, the growth rate of their trap layers varies greatly, resulting in significant differences in composition and thickness. Moreover, it is found by EDS surface scanning of all the samples of Fig. 5(d, e, and f) that three dark stripes of samples A and B correspond to a denser distribution of Al element, while the three bright stripes of sample C correspond to almost zero Al element. Actually, the position corresponding to these dark or bright stripes is where the well layers are located. In other words, the GaN well layers of samples A and B are incorporated with more Al atoms during the growth process, while it does not occur in sample C, implying that there is much less Al atoms in the well layers. In addition, both sample A and sample C exhibit a clear distribution of Al element atoms above the AlGaN quantum barrier in the top layer, as marked by the white arrows in the Fig. 5(d) and (f). This phenomenon is attributed to the oxidation on the surface of AlGaN. Due to prolonged exposure of the samples to air, their surface is oxidized to form aluminum oxide, resulting in a higher Al intensity distribution on the sample surface [25].
Fig. 5. The cross-sectional scanning transmission electron microscopy (STEM) images and corresponding Al-element energy dispersive X-ray (EDS) surface scan images for sample A (a) and (d), sample B (b) and (e), and sample C (c) and (f). The green arrows indicate the position of well layer regions in MQWs.
To visualize the distribution of N, Al and Ga elements in the MQWs more intuitively, the EDS element line scan mode is used for analysis, as shown in Fig. 6. Since the elemental distribution trends of sample A and sample B are similar, the distribution in sample B is not shown here in the figure. As can be seen from Fig. 6(a), the density distributions of Al (olive line) and Ga (orange line) in the MQW of sample A fluctuate greatly. The intensity of Al element at the interfaces between barrier and well layers is the highest, followed by a gradual decrease of Al content in the interior of well layers, as marked by the blue arrows. Moreover, in sample A, the intensity of the Al element in well layers is even significantly higher than that of the barrier layers, while the intensity of the Ga element is significantly lower than in barrier layers. However, the trend of distribution of Al and Ga elements in sample C is opposite to in sample A, as shown in Fig. 6(b). In addition, it is observed that sample A has a higher Al content at the well-barrier interface. It may be due to two reasons: (1) A 20s interruption in growth is introduced after the AlGaN barrier growth and before the growth of the GaN well layers. When an aluminum-rich environment exists in the reaction chamber, during the interruption of growth process, Al atoms have enough time to diffuse in the reaction chamber and undergo chemical reactions with GaN during the restart of growth, leading to nucleation or incorporation into the solid phase, forming a high Al content AlGaN interface film. (2) Since the Al-N bond energy is stronger than that of Ga-N, the desorption rate of Al atoms on the growing surface is much lower than that of Ga atoms [26,27]. Therefore, Al atoms enriched within the reaction chamber adsorb at the interface between the well and the barrier, resulting in a higher Al content in the grown AlGaN interface layer. In fact, for samples A and B, high-barrier “AlGaN quantum wells” are grown instead of GaN quantum wells, resulting in a much lower probability of radiative recombination of electrons and holes, and a poorer luminescence efficiency of the MQW samples. This is the reason why no MQW emission peak is observed in room temperature PL spectra of samples A and B, as shown in Fig. 4. In contrast, sample C is grown with no growth interruption between the growth of GaN well and AlGaN barrier layers. After the growth of the AlGaN barrier layers, the residual Al atoms in the reaction chamber are carried away by the carrier gas before they can diffuse, thereby the probability of Al atom incorporation is greatly reduced. Moreover, the atmosphere and temperature inside the reaction chamber do not change significantly during the continuous growth of the well and the barrier layers. When the GaN well layers are grown in a continuous way, the reaction of Al and N source gases is better controlled, which in turn avoids the formation of AlGaN with high Al content at the well-barrier interfaces. Therefore, the luminescence efficiency of sample C becomes the highest among three samples.
Fig. 6. EDS line scan images of elements Al (olive line), Ga (orange line), and N (pink line) for (a) sample A and (b) sample C.
4. Conclusion
Three different samples are epitaxially grown using MOCVD, which are designed to have the same MQW structure, but with and without interruption between the growth of barrier layer and well layer. Their surface morphology and luminescence characteristics are studied. The results of HRXRD and AFM indicate that the dislocation density is much less and surface quality is much better for the sample in which well and barrier layers are grown continuously at the same temperature in comparison with those samples which barrier and well layers are grown with an interruption or at different temperatures. The reason for this phenomenon is that the growth of well and barrier layers at the same temperature helps to improve the interface matching between AlGaN barrier and GaN well layers, reducing stress and dislocation density caused by lattice mismatch. At the same time, continuous growth of well and barrier layers can reduce the surface diffusion and migration processes of atoms, improving the surface morphology of UV MQWs. The room temperature PL spectra show that no MQW emission peak is observed in the samples with interrupted well and barrier growth, while the MQW emission is significantly strong in the sample grown with continuous well and barrier growth. This is due to the interrupted growth leading to the formation of a layer of high Al composition AlGaN at the well-barrier interface, which results in the deterioration of luminescence efficiency of MQWs. In contrast, the continuous growth of well and barrier layers can effectively reduce the probability of Al atom incorporation at the MQW interfaces, avoiding the formation of high Al composition AlGaN film at the interface, which improves the luminescence efficiency of UV MQWs.
Funding
National Key Research and Development Program of China (2022YFB3605104); Key Research and Development Program of Jiangsu Province (BE2021008-1); Shanxi-Zheda Institute of Advanced Materials and Chemical Engineering (2022SX-TD016); Youth Innovation Promotion Association of the Chinese Academy of Sciences (2019115, 2023124); National Natural Science Foundation of China (61904172, 61974162, 62034008, 62074140, 62074142, 62250038, 62274157).
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. Khan, K. Balakrishnan, and T. Katona, “Ultraviolet light-emitting diodes based on group three nitrides,” Nat. Photonics 2(2), 77–84 (2008). [CrossRef]
2. Z. Y. Li, M. H. Lo, C. T. Hung, et al., “High quality ultraviolet AlGaN/GaN multiple quantum wells with atomic layer deposition grown AlGaN barriers,” Appl. Phys. Lett. 93(13), 648 (2008). [CrossRef]
3. L. Li, Y. Xu, F. Yang, et al., “Optical properties of AlGaN-based deep-ultraviolet LED grown on graphene/SiC,” Mater. Lett. 317, 132104 (2022). [CrossRef]
4. M. Kneissl, T. Y. Seong, J. Han, et al., “The emergence and prospects of deep-ultraviolet light-emitting diode technologies,” Nat. Photonics 13(4), 233–244 (2019). [CrossRef]
5. J. Yang, D. G. Zhao, Z. S. Liu, et al., “A 357.9 nm GaN/AlGaN multiple quantum well ultraviolet laser diode,” J. Semicond. 43(1), 010501 (2022). [CrossRef]
6. S. Tanaka, Y. Ogino, K. Yamada, et al., “Low-threshold-current (∼85 mA) of AlGaN-based UV-B laser diode with refractive-index waveguide structure,” Appl. Phys. Express 14(9), 094009 (2021). [CrossRef]
7. D. Li, K. Jiang, X. Sun, et al., “AlGaN photonics: recent advances in materials and ultraviolet devices,” Adv. Opt. Photonics 10(1), 43–110 (2018). [CrossRef]
8. J. Han, M.H. Crawford, R. J. Shul, et al., “AlGaN/GaN quantum well ultraviolet light emitting diodes,” Appl. Phys. Lett. 73(12), 1688–1690 (1998). [CrossRef]
9. J. Bai, T. Wang, P. Comming, et al., “Optical properties of AlGaN/GaN multiple quantum well structure by using a high-temperature AlN buffer on sapphire substrate,” J. Appl. Phys. 99(2), 023513 (2006). [CrossRef]
10. C. Singha, S. Sen, A. Das, et al., “GaN/AlN multiple quantum wells grown by molecular beam epitaxy: effect of growth kinetics on radiative recombination efficiency,” Thin Solid Films 709, 138216 (2020). [CrossRef]
11. T. J. Anderson, A. D. Koehler, J. D. Greenlee, et al., “Substrate-dependent effects on the response of AlGaN/GaN HEMTs to 2-MeV proton irradiation,” IEEE Electron Device Lett. 35(8), 826–828 (2014). [CrossRef]
12. L. Tang, B. Tang, H. Zhang, et al., “Review of research on AlGaN MOCVD growth,” ECS J. Solid State Sci. Technol. 9(2), 024009 (2020). [CrossRef]
13. Q. Cai, H. You, H. Guo, et al., “Progress on AlGaN-based solar-blind ultraviolet photodetectors and focal plane arrays,” Light: Sci. Appl. 10(1), 94 (2021). [CrossRef]
14. M.D. Craven, S.H. Lim, F. Wu, et al., “Structural characterization of nonpolar (11$\bar{2}$0) a-plane GaN thin films grown on (1$\bar{1}$02) r-plane sapphire,” Appl. Phys. Lett. 81(3), 469–471 (2002). [CrossRef]
15. M.D. Craven, P. Waltereit, J.S. Speck, et al., “Well-width dependence of photoluminescence emission from a-plane GaNÕAlGaN multiple quantum wells,” Appl. Phys. Lett. 84(4), 496–498 (2004). [CrossRef]
16. S. Wang, X. Zhang, Z. Wu, et al., “The impact of purging on the quality of AlGaN/GaN multiple quantum wells grown on AlN/sapphire template,” J. Phys.: Conf. Ser. 844(1), 012015 (2017). [CrossRef]
17. K .R. Son, B. R. Lee, M. H. Jang, et al., “Enhanced light emission from AlGaN/GaN multiple quantum wells using the localized surface plasmon effect by aluminum nanoring patterns,” Photonics Res. 6(1), 30–36 (2018). [CrossRef]
18. M. Hanzaz, A. Bouhdada, P. Gibart, et al., “Impact of the defects on the electrical and optical properties of AlGaN ultraviolet photodetectors,” J. Appl. Phys. 92(1), 13–18 (2002). [CrossRef]
19. H. Heinke, V. Kirchner, S. Einfeldt, et al., “X-ray diffraction analysis of the defect structure in epitaxial GaN,” Appl. Phys. Lett. 77(14), 2145–2147 (2000). [CrossRef]
20. T. Metzger, R. Höpler, E. Born, et al., “Defect structure of epitaxial GaN films determined by transmission electron microscopy and triple-axis X-ray diffractometry,” Philos. Mag. A 77(4), 1013–1025 (1998). [CrossRef]
21. M. A. Moram and M. E. Vickers, “X-ray diffraction of III-nitrides,” Rep. Prog. Phys. 72(3), 036502 (2009). [CrossRef]
22. U. Zeimer, J. Jeschke, A. Mogilatenko, et al., “Spatial inhomogeneities in AlxGa1-xN quantum wells induced by the surface morphology of AlN/sapphire templates,” Semicond. Sci. Tech. 30(11), 114008 (2015). [CrossRef]
23. B. B. Wang, F. Liang, D. G. Zhao, et al., “Transient behaviours of yellow and blue luminescence bands in unintentionally doped GaN,” Opt. Express 29(3), 3685–3693 (2021). [CrossRef]
24. Z. Xie, Y. Sui, J. Buckeridge, et al., “Prediction of multiband luminescence due to the gallium vacancy–oxygen defect complex in GaN,” Appl. Phys. Lett. 112(26), 262104 (2018). [CrossRef]
25. X .L. Wang, D. G. Zhao, J. Chen, et al., “Effect of oxidation on the optical and surface properties of AlGaN,” Appl. Surf. Sci. 252(24), 8706–8709 (2006). [CrossRef]
26. G. Chen, Z.L. Li, X. Q. Wang, et al., “Effect of polarization on intersubband transition in AlGaN/GaN multiple quantum wells,” Appl. Phys. Lett. 102(19), 192109 (2013). [CrossRef]
27. A. Fan, X. Zhang, Z. Wu, et al., “Enhanced hole concentration and improved surface morphology for nonpolar a-plane p-type AlGaN/GaN superlattices grown with indium-surfactant,” Superlattices Microstruct. 130(1), 396–400 (2019). [CrossRef]
