Open Access
Add to BibSonomyAbstract
In this work, we reported the fabrication of nitride-based hexagonal pyramids array (HPA) vertical-injection light emitting diodes (V-LEDs) by N-polar wet etching. The performance of HPA V-LEDs devices was significantly improved with 30% higher internal quantum efficiency compared with conventional roughened broad area V-LEDs. The simulated extraction efficiency by finite difference time domain method was 20% higher than typical roughened V-LEDs. The reversed leakage current of HPA V-LEDs was reduced due to better crystal quality, which was confirmed by conductive atomic force microscopy measurement. Furthermore, the efficiency droop for HPA V-LEDs were substantially alleviated.
©2013 Optical Society of America
1. Introduction
Nitride-based light emitting diodes (LEDs) have attracted significant attention for their applications in solid-state lighting. Despite the achievement of nitride based LEDs that led to the realization of commercial productions, there are still severe limitations such as poor efficiency and dramatic efficiency droop, for which the polarization field [1,2], dislocations [3,4], electron leakage [5,6] and Auger recombination [7,8] are considered as the major reasons. Progresses in low-dimensional micro- or nano-structure LEDs for overcoming the limitations have been steadily developed. Early studies showed that micro-size array LEDs fabricated by dry etching have achieved higher light output efficiency than conventional broad area (BA) LEDs [9,10]. However, a large number of threading dislocations (TDs) were still presented in micro-size LEDs, and material damages caused by dry etching processes are not negligible. Recently, remarkable advances have been made in respect of the low-dimensional nanostructure LEDs. In particular, the nanowire/nanorod LEDs of which the active region consisted of dot-in-a-wire nanoscale heterojunction structure, showed a brilliant quantum efficiency and improved efficiency droop [6,11]. It is believed that the improvement of quantum efficiency for such nanostructures devices originated from the diminished dislocations and the alleviation of strain induced piezoelectric polarization field.
On the other hand, nitride-based vertical-injection LEDs (V-LEDs), of which the insulated sapphire substrates are removed to transfer the epilayer to a new thermal and electrical conductive substrates, have been considered as a promising candidate for high power applications [12–14]. Compared with the conventional lateral injection LEDs fabricated on sapphire substrates, V-LEDs provide many advantages, such as better current injection, excellent heat dissipation, enhanced reliability with respect to electrostatic discharge, a good scalability of chip size, true Lambertian emission, and simple packaging process [15,16]. The wet etching process of N-polar () surface has been commonly adopted to acquire the textured surface in order to improve the extraction efficiency in the fabrication of conventional BA V-LEDs [17]. The hexagonal pyramid-shaped morphology produced on the N-polar face after the wet etching process has been reported by several groups [18–20]. It is believed that TDs may serve as nucleation points for the etch process, and other defects play an important role in the wet etching of N-polar () surface [17,21]. However, the development of V-LEDs in high power applications still suffers from several hindrances, including a large leakage current under reverse bias, a serious efficiency droop and limited light extraction, which could be induced by several factors. For example, it is suspected that surface damages caused by laser lift-off (LLO) and inductively coupled plasma (ICP) etching process, the TDs produced in the epitaxial growth process and the reduced sidewall surface of device due to the nontransparent of Cu submount are possible reasons.
In this paper, we presented the fabrication of hexagonal pyramids array (HPA) V-LEDs. The hexagonal pyramids with isolated active regions were produced by the N-polar () face wet etching. The eliminating of TDs in hexagonal pyramids was realized by the wet etching process. The improvement of internal quantum efficiency (IQE) and leakage current under reverse bias was achieved by the reduction of TDs in HPA V-LEDs. Compared with conventional BA V-LEDs, the efficiency droop and output power of the HPA V-LEDs exhibited a significant improvement, which could be due to the better crystal quality and the amelioration of extraction efficiency as the inherent advantage of hexagonal pyramid shape. We believe that N-polar wet etching provided an effective approach to fabricate the micro- to nano-size V-LEDs array without detrimental ICP process and artificial lithography mask. Concurrently, the improved crystal quality and enhanced extraction efficiency could be achieved as a result of the wet etching process.
2. Experiment
The InGaN/GaN multi-quantum wells LEDs were grown by AIXTRON metalorganic chemical vapor deposition (MOCVD) system. Triethylgallium (TEGa), trimethylindium (TMIn), trimethylaluminum (TMAl) were used as group III sources, whereas NH3 was used as group V source. Silicane and bis (cyclopentadienyl) magnesium were employed as the donor and accepter dopant, respectively. After a 25 nm thick GaN nucleation layer was deposited under 550°C on sapphire substrate, a 2 μm undoped GaN layer was grown under 1100°C. Then, the n-type GaN:Si layer with a thickness of 2 μm, the active region that consists of 8 periods of InGaN/GaN structure, the 20 nm p-type AlGaN:Mg electron blocking layer (EBL), and the 0.2 μm p-type GaN:Mg were grown subsequently.
Scanning electron microscopy (SEM) morphological analysis was performed using Hitachi 4800. The tilt-view SEM images were acquired at the angle of 25°. A FEI F20 microscope with acceleration voltage of 200 kV was used to obtain bright-field transmission electron microscopy (TEM) images. Temperature dependent of photoluminescence (PL) spectra and spatially integrated PL spectra were measured with a standard PL system. An InGaN laser diode with a power of 8 mW was employed as the excitation source with a wavelength of 405 nm, the energy of which is lower than the band gap of GaN. Time-resolved PL was performed using a mode-locked Ti:sapphire laser with double-frequency. The wavelength, width and repetition rate was 400 nm, 150 fs and 75 MHz, respectively. The average laser intensity at the sample surface was estimated to be 300 W/cm2. Meanwhile, a streak camera with the minimum temporal resolution lower than 16 ps was employed to detect the decay profile of PL. Monochromatic cathodoluminescence (CL) images were collected by Gatan MONO CL3 + system. Electrical properties were characterized using a LED model probe station. Duty-cycle operation with cycle period of 20 ms and pulse width of 1ms was adopted at room temperature for electrical measurement. The typical I-V property of single hexagonal pyramid was studied using conductive atomic force microscopy (C-AFM) in contact mode with an Au-coated tip. The bias voltage was applied to the single hexagonal pyramid between Au-coated tip and Cu submount.
3. Results and discussion
The schematic of key procedures for the fabrications of HPA V-LEDs are shown in Figs. 1(a) –1(c). After the epitaxial growth procedure, the high reflective Ni/Ag/Pt/Au metallization p-contact was deposited on p-GaN using electron-beam evaporator. Prior to the removal of sapphire substrate through LLO, the electroplating of Cu was performed on the metallization p-contact as the new substrate. The remaining epitaxial film was then subjected to a 2.5 h wet etching to prepare for the following fabrication of HPA V-LEDs. Dilute potassium hydroxide (KOH) solution with a concentration of 6 mol/L served as the etch electrolyte at a constant temperature of 70°C. BA V-LEDs sample with 10 min wet etching in the same KOH solution for a roughed N-polar face was fabricated as reference.
Fig. 1 The schematic of key procedures for the fabrications of HPA V-LEDs, including (a) hexagonal pyramids randomly formed on p-electrode after wet etching (the red parts of pyramids are separated active regions) (b) spin-coated silicon gel to protect the active regions and (c) deposition of ITO and n-electrode. Tilt-view SEM images of (e) hexagonal pyramids, (d) hexagonal pyramids with spin-coated silicon gel after plasma etching, (f) hexagonal pyramids with deposited ITO film.
Eventually, the epilayer was separated into hexagonal pyramids by wet etching, as shown in Fig. 1(d). The hexagonal pyramids are distributed randomly on the Ni/Ag/Pt/Au reflective metalized contact. Right following the wet etching process, the cover of silicone gel on isolated pyramids was implemented by spin-coating in order to protect the active regions of hexagonal pyramids. After that, the epilayer was subjected to O2 plasma etching to expose the apexes of the pyramids. As seen in Fig. 1(e), the dark areas in SEM image which submerge the bottom part of the pyramids indicate the cover of silicone gel because of its insulation to electron beam, while the top parts of pyramids that exhibit the bright hexagonal shape were exposed as a result of the O2 plasma etching. Then, a 200 nm layer of indium tin oxide (ITO) was deposited on the N-polar surface of epilayer, which serve as transparent electrode and current spreading layer. As displayed in Fig. 1(f), the deposition of ITO on hexagonal pyramids shows uniform covering on both pyramids and silicone gel, leading to the interconnection of hexagonal pyramids to realize current spreading. Cr/Pt/Au contacts were deposited on the ITO film as n-type electrode. Finally, the HPA V-LEDs and BA V-LEDs devices were fabricated into a 1200 μm × 1200 μm chip.
In the conventional diffraction contrast method, TEM images taken in two-beam diffraction conditions with diffraction vector g are used to investigate the Burgers vector b of dislocations, using the invisibility criterion g∙b = 0 when the dislocations are out of contrast [22]. Bright field TEM images of single pyramid in Fig. 2 were collected under two-beam condition of g = and g = to display the edge and screw dislocations, respectively. It is noteworthy that no obvious screw or edge dislocations are observed inside the pyramid in these diffraction conditions. Moreover, the active regions positioned in the bottom part of the hexagonal pyramids are separated by the wet etching process. The InGaN/GaN MQWs shows abrupt interfaces over the whole pyramid, which indicates MQWs structure closed to the edge of the pyramid are not affected by the wet etching. Furthermore, during the wet etching process, most TDs were connected with the grooves between the pyramids and terminated at the etched region during the wet etching process. As a result, it indicates that the dislocations which act as nonradiative centers may be important etching channels.
The top-view and birds-eye-view monochromatic CL images of single and arrayed pyramids samples were shown in Fig. 3 . The wet etching time for single pyramid and arrayed pyramid samples are specified to obtain an isolated pyramid and pyramids array, respectively. The top-view CL image of single pyramid displays the profile of active region, which is hexagonal shaped by the wet etching process. As the acceleration voltage increasing, the visible emitting area in top-view CL image increases due to the enhanced penetration depth of cathode ray through the n-type GaN in the top part of the pyramid, as shown in Figs. 3(a)–3(c). The whole area of active region within the pyramid is luminescent at 457 nm by the cathode ray under 20 kV with gradually decreased intensity distribution from edge to center, which is caused by the gradually increased thickness of n-type GaN from edge to center. The bright area in birds-eye-view CL image of single pyramid clearly indicates the position of active region, the dark area above and below it corresponds to the n-type and p-type GaN, respectively, as seen in Figs. 3(d)–3(f). No dark points or other wavelength are detected. Previously studies have revealed that the surface pits correlated with TDs, or regions containing high dislocation densities exhibit dark spots with reduced intensity in monochromatic CL images [23–25]. In this regard, these observations illustrate the formation of dislocation-eliminating hexagonal pyramids after wet etching process. The CL images of pyramids array which is prepared for the following fabrication processes are shown in Fig. 3(g). The size distribution of pyramids is relatively homogeneous with the average height of about 2 μm, while there are still small pyramids located on the spacing among or adjacent to the large pyramids, which are luminescent under cathode ray as well.
Fig. 3 (a-c) Top-view monochromatic CL images of single pyramid under 5 kV, 10 kV and 20 kV, respectively. (d-f) Birds-eye-view monochromatic CL images of single pyramid under 5 kV, 10 kV and 20 kV, respectively. (g) Birds-eye-view monochromatic CL image of large scale pyramids array under 10 kV.
The crystal quality of HPA V-LEDs and BA V-LEDs are investigated by temperature dependence of photoluminescence (PL) spectrum. Figure 4(a) shows the Arrhenius plot of the normalized integral PL intensity of HPA V-LEDs and conventional BA V-LEDs. The ratio of PL intensity at 300 K to that at 10 K is defined as the estimation of IQE. The measured IQE of HPA V-LEDs increases by 30% compared with that of BA V-LEDs, which suggests that the HPA V-LEDs has an improved crystallinity by the reduction of TDs density. Furthermore, the temperature at which the PL intensity starts to quench for HPA V-LEDs is much higher than that for BA V-LEDs. The quenching of PL intensity for HPA V-LEDs is well described using single activation energy. However, the PL intensity for BA V-LEDs first quenches with weak activation energy, then a second activation energy is dominated at high temperature. This observation indicates the larger activation energy for HPA V-LEDs, leading to the restricted activation of nonradiative recombination with increased temperature [26,27]. The inset of Fig. 4(a) shows the PL spectra of the two samples at 10 K and 300 K. The peak intensity of PL of HPA V-LEDs is 1.6 times stronger than that of BA V-LEDs, which may principally ascribe to the elimination of TDs and enhanced light extraction due to the large sidewall surface of hexagonal pyramid shape. The 3D finite difference time domain (FDTD) computational domain for HPA V-LEDs is a 10-period hexagonal pyramids array with a diameter and pitch of 1 and 2 μm, respectively. The theoretically calculated extraction efficiency for HPA V-LEDs using 3D FDTD method is about 71%, which is much higher than 52% for conventional roughened BA V-LEDs. As seen in Fig. 5 , the calculated far-field emission pattern of HPA V-LEDs is shown. The light intensity distribution of HPA V-LEDs spreads over the entire surface, suggesting that the light escape angle is significantly enlarged. Previous research by J. H. Son et al clearly revealed that the largest extraction efficiency enhancement was realized when the sidewall angle of nanostructure is approaching to 23.4°, which is the critical angle for total internal reflection (TIR) at the GaN/air interface, because of the effective elimination of TIR at the GaN/air interface [28]. The natural angle of sidewall () for hexagonal pyramid formed by wet etching in KOH solution is fixed at 31.6° [18], which is close to the critical angle, leading to the inherent advantage of high extraction efficiency for hexagonal pyramids shape. Elsewhere, although the removal of sapphire substrate has alleviated the strain between the epilayer and sapphire, the peak wavelength of PL at 10 K and 300 K for HPA V-LEDs still exhibits a blue shift of 5 nm compared with BA V-LEDs, which could be due to the further relaxed strain of epilayer by the isolated pyramids through wet etching.
Fig. 4 (a) Temperature dependence of integrated PL intensities (log scale) for HPA (red) and BA (green) V-LEDs. The inset shows the PL spectra for the two sample (b) Time-resolved PL decay curves for HPA (red) and BA (green) V-LEDs at 10 K and 300 K.
TRPL measurements were performed for the main emission peak of the two samples at 10 K and 300 K to study the recombination mechanism, as illustrated in Fig. 4(b). It is generally believed that the luminescence decay process is related to radiative and nonradiative recombination processes with the assumption that the nonradiative recombination is frozen at 10 K. The decay time of HPA V-LEDs at 10 K is slightly shorter than that of BA V-LEDs. The shorter life time at low temperature results from the enhanced oscillator strength due to the released quantum confine stark effect (QCSE) caused by the relaxed strain, which is in accordance with the blue shift observed in PL spectra for HPA V-LEDs. However, the decay time for HPA V-LEDs at room temperature is much longer than that for BA V-LEDs. Considering the predominant nonradiative recombination at room temperature, the longer decay time indicates the nonradiative lifetime is elongated in HPA V-LEDs due to the elimination of TDs and improved crystal quality. Furthermore, it is noteworthy that the discrepancy between the decay time of HPA V-LEDs at 10 K and 300 K is much smaller than that of BA V-LEDs, which suggests the stable radiative recombination rate and suppressed activation of non radiative recombination processes over the whole range of temperature for HPA V-LEDs. In contrast, the decay time for BA V-LEDs shows great degeneration as temperature increases from 10 K to 300 K. These results are consistent with the smaller intensity decrease of the integral PL intensity for HPA V-LEDs with increasing temperature compared with BA V-LEDs. Therefore, it is clear that the radiative recombination process is dominant in the HPA V-LEDs up to room temperature, resulting in the high brightness and high efficiency.
The electrical property of HPA V-LEDs and BA V-LEDs were investigated on the Model LED tester. The HPA V-LEDs show the higher forward voltage, as seen in current versus voltage (I-V) characteristic of Fig. 6(a) . This is probably owing to the insulated parts of contact between silicone gel and ITO, as well as the limited effective contact area of hexagonal pyramids. Furthermore, the leakage current under reverse bias of 10 V for HPA V-LEDs is reduced to 0.87 μA compared with 3.76 μA for BA V-LEDs as shown in Fig, which is attributed to the improved crystal quality. In order to explore the origin of the improved leakage current for HPA V-LEDs devices, localized I-V properties of single pyramid and BA V-LEDs were investigated using C-AFM at room temperature. I-V properties of various contact points on the single pyramid and N-polar surface of BA V-LEDs were measured. As seen in Fig. 6(b), the average leakage current for single pyramid has the constant value of 0.3 nA under reversed bias up to 10 V. It indicates that such a saturated current under reversed bias originate from the absence of defect induced trap levels within the band gap [29]. However, the reversed leakage current for BA V-LEDs significantly increases with reversed bias, resulting in the leakage current of 457 nA at 10 V. Such a large leakage current may attribute to the ICP etching damage and defects formed during the epitaxy. Moreover, the leakage current for BA V-LEDs exhibits a strong field dependent, indicating that defect-assisted tunneling is the dominant transport mechanism. In this regard, it is evidenced that the better leakage current for HPA V-LEDs devices was originated from the improved crystal quality of hexagonal pyramids. Elsewhere, the threshold voltages of BA V-LEDs and single pyramid measured by C-AFM are much larger than that of corresponding devices in typical I-V measurement, which may attribute to the poor ohmic contact between Au-coated tip and N-polar surface.
Fig. 6 (a) I-V properties for BA and HPA V-LEDs, the inset shows the leakage current under reversed bias. (b) Localized I-V curves (log scale) of HPA (red) and BA (green) V-LEDs measured by C-AFM.
The light output power versus forward current (L-I) for HPA V-LEDs and conventional BA V-LEDs are shown in Fig. 7(a) . Under the injection of 350 mA, the light output of HPA V-LEDs increases about 35% compared with BA V-LEDs. The enhanced light extraction allied with the improved crystal quality contributes to the increased light output of HPA V-LEDs. However, the increase in light output power for HPA V-LEDs is less than the multiply raise in IQE and extraction efficiency. This may ascribe to several reasons: primarily, the decreased area of active region for HPA V-LEDs due to the wet etching; secondly, the increased power dissipation caused by the raised forward voltage for HPA V-LEDs. Additionally, the limited area of detector may not be capable of completely collecting the more diverging light emitted from HPA V-LEDs. The microscope images of 20 mA driven HPA V-LEDs is shown in the inset of Fig. 7(a), which is taken from n-type metal contact surface. Relative external quantum efficiency (EQE) versus current density for HPA V-LEDs and conventional BA V-LEDs are measured at 300 K, depicted in Fig. 7(b). Relative EQE droop is defined as the ratio of the decreased quantity of EQE at certain temperature to the maximum value. The HPA V-LEDs exhibits significant improved efficiency droop with a peak value of 1.575, which is 12.5% more than that of BA V-LEDs. The efficiency droop at 350 mA for HPA V-LEDs is 3.3%, which is significantly alleviated compared with 19.7% for BA V-LEDs. It is noteworthy that although the effective active region of individual chip was reduced caused by wet etching process, the HPA V-LEDs devices still demonstrate significantly improved performance. Since N-polar wet etching process of GaN exhibited the feature of dislocation-selected, the eliminated part of epilayer by wet etching was almost dislocated material, yet the remaining epilayer was nearly dislocation-free. As a consequence, the wet etching process could decrease the density of TDs, which is randomly distributed in typical GaN V-LEDs. And this is chief reason that the performance of HPA V-LEDs was enhanced and able to overcome the loss of active regions. In addition, despite of the inherent advantage of the enhanced extraction efficiency for hexagonal pyramid shape, the random distribution of pyramids reduced the recapture possibility of extracted light at large escape angle from individual pyramid by adjacent pyramids.
Fig. 7 (a) Light output power versus current for BA (solid circle) and HPA (solid triangle) V-LEDs. The inset shows the optical photograph of HPA V-LEDs under 20 mA. (b) Relative efficiency droop versus current for HPA (red) and BA (green) V-LEDs.
4. Conclusion
In conclusion, we fabricated the high performance HPA V-LEDs through N-polar face wet etching. The TEM and CL analysis indicate that TDs in HPA V-LEDs are eliminated significantly by wet etching process. The estimated IQE for HPA V-LEDs by temperature dependence of PL showed an increase of 30% than that for BA V-LEDs due to the better crystal quality. Furthermore, the TRPL measurements showed the more stable radiative recombination rate and suppressed activation of nonradiative recombination process with increased temperature for HPA V-LEDs. As a result, the performance of fabricated HPA V-LEDs exhibits substantial improvement compared with that of conventional BA V-LEDs. The reverse leakage current for HPA V-LEDs is reduced from 3.76 μA to 0.87 μA. It is revealed that the reduced leakage current is caused by the reduction of TDs in hexagonal pyramids, which is evidenced by the constant leakage current of 0.3 nA under reversed bias up to 10 V by C-AFM measurement. The output power of HPA V-LEDs is found to be about 1.35 times higher than that of BA V-LEDs, which reflect the improvement of crystal quality and extraction efficiency. Moreover, the HPA V-LEDs suffer from alleviated efficiency droop in comparison with BA LEDs, which is only about one sixth of the efficiency droop for BA V-LEDs. It is demonstrated that the N-polar wet etching process to separate the active regions is an effective approach to fabricate the micro- to nano-size V-LEDs array without detrimental ICP process and artificial lithography mask. Meanwhile, it provides a simple but effective way to further improve the crystal quality and light extraction of GaN based LEDs based on the existing N-polar roughening method.
Acknowledgments
This work was financially supported by the National High Technology Research and Development Program of China (No. 2011AA03A105). Jun Ma and Liancheng Wang contributed equally to this work.
References and links
1. M. F. Schubert, J. Xu, J. K. Kim, E. F. Schubert, M. H. Kim, S. Yoon, S. M. Lee, C. Sone, T. Sakong, and Y. Park, “Polarization-matched GaInN/AlGaInN multi-quantum-well light-emitting diodes with reduced efficiency droop,” Appl. Phys. Lett. 93(4), 041102 (2008). [CrossRef]
2. F. Bernardini, V. Fiorentini, and D. Vanderbilt, “Spontaneous polarization and piezoelectric constants of III-V nitrides,” Phys. Rev. B 56(16), R10024–R10027 (1997). [CrossRef]
3. B. Monemar and B. E. Sernelius, “Defect related issues in the “current roll-off” in InGaN based light emitting diodes,” Appl. Phys. Lett. 91(18), 181103 (2007). [CrossRef]
4. Y. Yang, X. A. Cao, and C. H. Yan, “Investigation of the nonthermal mechanism of efficiency rolloff in InGaN light-emitting diodes,” IEEE Trans. Electron. Dev. 55(7), 1771–1775 (2008). [CrossRef]
5. M. H. Kim, M. F. Schubert, Q. Dai, J. K. Kim, E. F. Schubert, J. Piprek, and Y. Park, “Origin of efficiency droop in GaN-based light-emitting diodes,” Appl. Phys. Lett. 91(18), 183507 (2007). [CrossRef]
6. H. P. T. Nguyen, K. Cui, S. Zhang, M. Djavid, A. Korinek, G. A. Botton, and Z. Mi, “Controlling electron overflow in phosphor-free InGaN/GaN nanowire white light-emitting diodes,” Nano Lett. 12(3), 1317–1323 (2012). [CrossRef] [PubMed]
7. Y. C. Shen, G. O. Mueller, S. Watanabe, N. F. Gardner, A. Munkholm, and M. R. Krames, “Auger recombination in InGaN measured by photoluminescence,” Appl. Phys. Lett. 91(14), 141101 (2007). [CrossRef]
8. M. Zhang, P. Bhattacharya, J. Singh, and J. Hinckley, “Direct measurement of auger recombination in In0.1Ga0.9N/GaN quantum wells and its impact on the efficiency of In0.1Ga0.9N/GaN multiple quantum well light emitting diodes,” Appl. Phys. Lett. 95(20), 201108 (2009). [CrossRef]
9. H. W. Choi, M. D. Dawson, P. R. Edwards, and R. W. Martin, “High extraction efficiency InGaN micro-ring light-emitting diodes,” Appl. Phys. Lett. 83(22), 4483 (2003). [CrossRef]
10. S. X. Jin, J. Li, J. Z. Li, J. Y. Lin, and H. X. Jiang, “GaN microdisk light emitting diodes,” Appl. Phys. Lett. 76(5), 631 (2000). [CrossRef]
11. H.-M. Kim, Y.-H. Cho, H. Lee, S. I. Kim, S. R. Ryu, D. Y. Kim, T. W. Kang, and K. S. Chung, “High-brightness light emitting diodes using dislocation-free indium gallium nitride/gallium nitride multiquantum-well nanorod arrays,” Nano Lett. 4(6), 1059–1062 (2004). [CrossRef]
12. O. B. Shchekin, J. E. Epler, T. A. Trottier, T. Margalith, D. A. Steigerwald, M. O. Holcomb, P. S. Martin, and M. R. Krames, “High performance thin-film flip-chip InGaN–GaN light-emitting diodes,” Appl. Phys. Lett. 89(7), 071109 (2006). [CrossRef]
13. Ch.-F. Chu, F.-I. Lai, J.-T. Chu, C.-C. Yu, C.-F. Lin, H.-C. Kuo, and S. C. Wang, “Study of GaN light-emitting diodes fabricated by laser lift-off technique,” J. Appl. Phys. 95, 3916 (2004).
14. W. Y. Lin, D. S. Wuu, K. F. Pan, S. H. Huang, C. E. Lee, W. K. Wang, S. C. Hsu, Y. Y. Su, S. Y. Huang, and R. H. Horng, “High-power GaN-mirror-Cu light-emitting diodes for vertical current injection using laser liftoff and electroplating techniques,” IEEE Photon. Technol. Lett. 17(9), 1809–1811 (2005). [CrossRef]
15. C.-F. Chu, C.-C. Yu, H.-C. Cheng, C.-F. Lin, and S.-C. Wang, “Comparison of p-Side Down and p-Side Up GaN Light-Emitting Diodes Fabricated by Laser Lift-Off,” Jpn. J. Appl. Phys. 42(Part 2, No. 2B), L147–L150 (2003). [CrossRef]
16. J.-T. Chu, H.-W. Huang, C.-C. Kao, W.-D. Liang, F.-I. Lai, C.-F. Chu, H.-C. Kuo, and S.-C. Wang, “Fabrication of large-area GaN-based light-emitting diodes on Cu substrate,” Jpn. J. Appl. Phys. 44(4B), 2509–2511 (2005). [CrossRef]
17. T. Fujii, Y. Gao, R. Sharma, E. L. Hu, S. P. DenBaars, and S. Nakamura, “Increase in the extraction efficiency of GaN-based light-emitting diodes via surface roughening,” Appl. Phys. Lett. 84(6), 855 (2004). [CrossRef]
18. H. M. Ng, N. G. Weimann, and A. Chowdhury, “GaN nanotip pyramids formed by anisotropic etching,” J. Appl. Phys. 94(1), 650 (2003). [CrossRef]
19. S. L. Qi, Z. Z. Chen, H. Fang, Y. J. Sun, L. W. Sang, X. L. Yang, L. B. Zhao, P. F. Tian, J. J. Deng, Y. B. Tao, T. J. Yu, Z. X. Qin, and G. Y. Zhang, “Study on the formation of dodecagonal pyramid on nitrogen polar GaN surface etched by hot H3PO4,” Appl. Phys. Lett. 95(7), 071114 (2009). [CrossRef]
20. Y. Gao, T. Fujii, R. Sharma, K. Fujito, S. P. DenBaars, S. Nakamura, and E. L. Hu, “Roughening hexagonal surface morphology on laser lift-off (LLO) N-face GaN with simple photo-enhanced chemical wet etching,” Jpn. J. Appl. Phys. 43(No. 5A), L637–L639 (2004). [CrossRef]
21. J.-H. Kim, C.-S. Oh, Y.-H. Ko, S.-M. Ko, K.-Y. Park, M. Jeong, J. Y. Lee, and Y.-H. Cho, “Dislocation-eliminating chemical control method for high-efficiency GaN-based light emitting nanostructures,” Cryst. Growth Des. 12(3), 1292–1298 (2012). [CrossRef]
22. P. B. Hirsch, A. Howie, R. B. Nicholson, D. W. Pashley, and M. J. Whelan, “Electron Microscopy of Thin Crystals” (Krieger, New York, 1977).
23. J. S. Speck and S. J. Rosner, “The role of threading dislocations in the physical properties of GaN and its alloys,” Physica B 273–274, 24–32 (1999). [CrossRef]
24. T. J. Badcock, R. Hao, M. A. Moram, M. J. Kappers, P. Dawson, and C. J. Humphreys, “The effect of dislocation density and surface morphology on the optical properties of InGaN/GaN quantum wells grown on r-plane sapphire substrates,” Jpn. J. Appl. Phys. 50(8), 080201 (2011). [CrossRef]
25. Y. B. Tao, T. J. Yu, Z. Y. Yang, D. Ling, Y. Wang, Z. Z. Chen, Z. J. Yang, and G. Y. Zhang, “Evolution and control of dislocations in GaN grown on cone-patterned sapphire substrate by metal organic vapor phase epitaxy,” J. Cryst. Growth 315(1), 183–187 (2011). [CrossRef]
26. F. E. Williams and H. Eyring, “The mechanism of the luminescence of solids,” J. Chem. Phys. 15(5), 289–304 (1947). [CrossRef]
27. P. J. Dean, “Absorption and luminescence of excitons at neutral donors in gallium phosphide,” Phys. Rev. 157(3), 655–667 (1967). [CrossRef]
28. J. H. Son, J. U. Kim, Y. H. Song, B. J. Kim, C. J. Ryu, and J.-L. Lee, “Design rule of nanostructures in light-emitting diodes for complete elimination of total internal reflection,” Adv. Mater. (Deerfield Beach Fla.) 24(17), 2259–2262 (2012). [CrossRef] [PubMed]
29. P. Kozodoy, J. P. Ibbetson, H. Marchand, P. T. Fini, S. Keller, J. S. Speck, S. P. DenBaars, and U. K. Mishra, “Electrical characterization of GaN p-n junctions with and without threading dislocations,” Appl. Phys. Lett. 73(7), 975 (1998). [CrossRef]
