This content has been downloaded from IOPscience. Please scroll down to see the full text.
Download details:
IP Address: 140.113.38.11
This content was downloaded on 25/04/2014 at 10:40
Please note that terms and conditions apply.
Performance Enhancement of a-Plane Light-Emitting Diodes Using InGaN/GaN Superlattices
View the table of contents for this issue, or go to the journal homepage for more 2009 Jpn. J. Appl. Phys. 48 04C136
(http://iopscience.iop.org/1347-4065/48/4S/04C136)
Performance Enhancement of a-Plane Light-Emitting Diodes
Using InGaN/GaN Superlattices
Shih-Chun Ling1, Te-Chung Wang1;2, Jun-Rong Chen1, Po-Chun Liu2, Tsung-Shine Ko1, Tien-Chang Lu1, Hao-Chung Kuo1, Shing-Chung Wang1, and Jenq-Dar Tsay2
1Department of Photonics and Institute of Electro-Optical Engineering, National Chiao Tung University,
1001 University Road, Hsinchu, Taiwan 300, R.O.C.
2Electronics and Optoelectronics Research Laboratories, Industrial Technology Research Institute,
195 Chung Hsing Rd., Sec. 4 Chu Tung, Hsinchu, Taiwan 310, R.O.C.
Received September 22, 2008; revised November 26, 2008; accepted December 11, 2008; published online April 20, 2009
We have fabricated a-plane light-emitting diodes (LEDs) with inserted InGaN/GaN superlattices. The structural characteristics and device performance of a-plane light-emitting diodes with/without superlattices were also identified. Transmission electron microscope (TEM) images revealed that the threading dislocation (TD) density for the sample using superlattices was reduced from 3 1010cm 2down to
9 109cm 2. The electroluminescence (EL) intensity of the sample with InGaN/GaN superlattices was enhanced by a factor of 3.42
times to that of the conventional sample without InGaN/GaN superlattices. Furthermore, we observed that the polarization degree of a-plane LEDs with superlattices (56.3%) was much higher than that without superlattices (27.4%). A series of experiments demonstrated that the feasibility of using InGaN/GaN superlattices for the TD reduction and the improvement of luminescence performance in a-plane III–nitride devices. #2009 The Japan Society of Applied Physics
DOI: 10.1143/JJAP.48.04C136
1. Introduction
Recently, nitride-based light-emitting diodes (LEDs) have been attracting great attention owing to the potential application in solid-state lighting. However, the conven-tional c-plane nitride-based quantum wells exhibit the quantum-confined Stark effect1,2)as a result of the existence of spontaneous and piezoelectric polarization fields that are parallel to [0001] c-direction. This effect results in spatial separation of the electron and hole wave functions in the quantum wells, which restricts the carrier recombination efficiency, reduces oscillator strength, and induces red-shifted emission.3) To avoid such polarization effects, growth along ½11220-oriented direction has been explored for planar a-plane GaN prepared on r-plane sapphire4)and a-plane SiC5) by metalorganic vapor phase deposition (MOCVD). Since these GaN surfaces contain an equal number of Ga and N atoms in each monolayer, a free electric field and nonpolar characteristic are obtained. The recent studies of a-plane AlGaN/GaN6,7) and InGaN/GaN8,9) multi-quantum wells (MQWs) demonstrate that it is pos-sible to eliminate such polarization fields along non-polar orientation. In addition, nonnon-polar LEDs have been shown to exhibit optically polarized spontaneous emis-sion, which is explained to be a result of the crystal field oriented along the c-axis of wurtzite GaN and its effect on the valence band splitting induced by large compressive strain within the wells.10,11) Therefore, the use of these polarized light emitters in liquid crystal display units is a great chance to save energy owing to the elimination of a sheet polarizer.12) However, the difficult issue is that there is no suitable substrate for heteroepitaxial planar a-plane GaN growth.
In general, there is a threading dislocation (TD) density of 3 1010cm2 and a basal stacking fault density of
3:5 105cm1in a-plane GaN grown on r-plane sapphire
structure because of the large lattice mismatch.13)The TDs in GaN act as nonradiative recombination centers to restrict internal quantum efficiency. Therefore, the reduction of threading dislocation density is essential to improve device
performance. Lateral epitaxial overgrowth (LEO) techniques have been employed in the past to achieve defect reduction in nonpolar GaN.14–18)However, all of these LEO techniques involve ex-situ processing steps and it is difficult to control the uniformity of the regrowth thickness exceeding 20 mm required for coalescence. Another possible method for reducing threading dislocation density is to insert strain-layer superlattices into the epitaxial strain-layer. The strain which is due to different lattice constants in the superlattices, can deflect the threading dislocation into the interfacial plane and has been theoretically analyzed by Matthews and Blakeslee.19,20) The dislocation line that runs parallel to the growth direction is bowed as a result of the coherent strain present in the superlattices. Subsequently, the dislocations move laterally and are eliminated with other dislocations of the opposite Burgers vector, or combine with a dislocation of a different Burgers vector to form a third one, or even run to the edge of the wafer. Ultimately, dislocation density reduction is achieved by the insertion of superlattices. In this study, we utilize the strain mechanism to realize defect reduction in a-plane LEDs and demonstrate the performance improvement of a-plane LEDs using InGaN/GaN super-lattices. Unlike these traditional LEO techniques, this dislocation reduction method is highly advantageous due to the simplicity and low cost.
2. Experiments
First, a low temperature GaN nucleation layer was grown by low pressure metal-organic chemical vapor deposition (MOCVD) on r-plane sapphire substrates, followed by the growth of 0.5-mm-thick high temperature GaN. Then, a superlattices comprising 20 pairs of 5-nm-thick In0:2Ga0:8N
and 5-nm-thick GaN were inserted and subsequently a 1.5-mm-thick Si-doped n-GaN with an electron concentration of 3 1018cm3 was grown. Afterward, a MQWs consisting
10 pairs of 6-nm-thick wells and 15-nm-thick barriers was grown at a temperature of 827C and capped by a
0.15-mm-thick p-GaN layer with a hole concentration of 6 1017
cm3. The structure of the a-plane LED is shown in Fig. 1.
electron microscopy (TEM) to compare the microstructures of a-plane LEDs grown with and without InGaN/GaN superlattices. Temperature-dependent micro-photolumines-cence (m-PL) measurements were carried out using the 405 nm line of a CW InGaN laser diode with a spot size of 10 mm over the temperature range 90 – 300 K. Then, 300 300 mm2diode mesas were defined by chlorine-based
reactive ion etching. Ti/Au (100/200 nm) and Ti/Al/Pt/Au (30/180/40/150 nm) were used as p- and n-GaN contacts, respectively. The electroluminescence (EL) and polarization characteristics of the diodes were measured by on-wafer probing of the devices.
3. Results and Discussion
To demonstrate the dislocation reduction of the a-plane LED with the insertion of InGaN/GaN superlattices, TEM was performed to compare the microstructures of a-plane LEDs with and without superlattices. The typical bright-field cross-sectional TEM image of the sample with superlattices near ½11100GaN zone axis is shown in Fig. 2(a). The inset of
Fig. 2(a) presents the corresponding electron diffraction pattern. From the TEM image, significant blocking of TDs was observed at the n-GaN/superlattices interface. Thus it was apparent that the insertion of InGaN/GaN superlattices can reduce the dislocation density effectively. The estimated threading dislocation density was reduced from 3 1010
cm2 down to 9 109cm2. Figures 2(b) and 2(c) show the MQWs grown on the template without and with superlattices, respectively. As shown in Fig. 2(b), the boundaries between InGaN wells and GaN barriers were severely interfered by lots of TDs parallel to the ð11220Þ growth direction that would give rise to the poor carrier confinement and low internal quantum efficiency. In contrast, as shown in Fig. 2(c), the MQWs revealed a clearer boundary between InGaN wells and GaN barriers than that of Fig. 2(b). In order to confirm the quality improvement of a-plane MQWs using InGaN/GaN super-lattices, Arrhenius plots of the normalized integrated PL intensity for the a-plane LED without and with InGaN/GaN superlattices over the temperature range 90 – 300 K were measured. The 405 nm line of InGaN laser diode was used as a PL pumping source to avoid the absorption in GaN and
prevent the excess photo-generated carriers from being injected into MQWs. Therefore, the effective excitation conditions are the same for the MQWs grown on the templates with and without superlattices. Sample 1 and Sample 2 are the a-plane LED without and with InGaN/ GaN superlattices, respectively. Figures 3(a) and 3(b) show the normalized integrated PL as a function of 1000=T for Sample 1 and Sample 2, respectively, wherein the thermal activation energies of 86.15 and 118.7 meV were estimated from the Arrhenius plots. Furthermore, the integrated PL intensity ratio obtained at 90 and 300 K [IPL(300 K)/
IPL(90 K)] for Sample 2, which is approximately 16%, is
larger than that of Sample 1 (nearly 10.3%). In general, the temperature-induced quenching of luminescence involves Fig. 1. (Color online) Structure of a-plane LED with the insertion of
InGaN/GaN superlattices. [0001] 200 nm InGaN/GaN superlattices n-GaN u-doped GaN [11-20]
(a)
(c)
[0001] 100 nm [11-20](b)
100 nm [11-20] [0001]Fig. 2. (a) Cross-sectional TEM image of a-plane LED with super-lattices. Inset shows the corresponding electron diffraction pattern. Morphologies of MQWs grown on the templates (b) without and (c) with superlattices are also shown.
(a) 2 4 6 8 10 12 Sample 1 IPL(300 K) / IPL(90 K)=10.3% Ea=86.15meV Normailz ed intensity (arb. unit) 1000/T (b) 2 4 6 8 10 12 Sample 2 IPL(300 K) / IPL(90 K)=16% Ea=118.7meV Normaliz ed intensity (arb. unit) 1000/T
Fig. 3. Arrhenius plots of the normalized integrated PL intensity over the temperature range 90 – 300 K for (a) Sample 1 and (b) Sample 2.
the thermal emission of charge carriers out from confined quantum-well states into barrier states,21,22) the thermal dissociation of excitons into free electron–hole pairs,23)and the thermal activation of excitons to non-radiative defect states.24) Therefore, the thermal activation energy is a quantity to measure the exciton binding energy or the energy difference between the energy of the quantum-well confined state and the barrier continuum state or defect state. As a result, the carrier confinement of a-plane MQWs was indeed enhanced in terms of activity energy estimation by introducing InGaN/GaN superlattices.
Room-temperature EL spectra under 20 mA injection current is shown in Fig. 4(a). The same emission peaks were located at 449 nm for both samples. The EL full width at half maximum (FWHM) of Sample 2 was measured as 28.7 nm, which was narrower than that of Sample 1 (FWHM = 35.8 nm). The L–I–V curves were measured to further demonstrate the performance enhancement of a-plane LEDs with InGaN/GaN superlattices as shown in Fig. 4(b). These two I–V curves exhibited similar shapes, indicating that the absence of the degradation of electrical characteristics with the insertion of superlattices. The forward voltages of Sample 1 and Sample 2 were both 4.25 V at 20 mA operat-ing current. The series resistances of Sample 1 and Sample 2 were estimated to be 24 . The output power of Sample 2 had a 3.42- and 2.86-fold increase compared with those of Sample 1 at 20 and 100 mA injection current, respectively. Such an apparent enhancement was attributed to the dislocation reduction and the improvement of the MQW quality upon inserting InGaN/GaN superlattices.
The a-plane (In,Ga)N films suffer from anisotropic in-plane compressive strain in the x–y in-plane (x is oriented along h11100i, y is oriented along h11220i, and z is oriented along h0001i) of the wurtzite crystal and the original jX iYi-like valence-band states of unstrained wurtzite crystal are split into jXi- and jYi-like states. The energy of the jXi-like state is raised by the strain, while that of the jYi-like state is pushed down to below that of the jZi-like state (see the inset of Fig. 5). Therefore, electronic transitions occur predom-inantly from the bottom of the conduction band to the topmost jXi-like state and the resulting light emission will have a strong x-polarized ðE ? cÞ characteristic.10,11) In order to analyze the linear polarization of the EL of our devices, we rotated a polarizer between the polarization angle of 0(parallel to the c-axis) and 360. The polarization
ratio is defined as ¼ ðImaxIminÞ=ðImaxþIminÞ, where Imax
is the intensity of light with polarization perpendicular to the c-axis and Imin is the intensity of light with polarization
parallel to the c-axis. Figure 5 shows the EL intensities of Sample 1 and Sample 2 at different polarization angles at room temperature. The degree of polarization of Sample 2, which was estimated to be approximately 56.3%, was larger than that of Sample 1 (nearly 27.4%). There are two possible reasons that account for this phenomenon. First, the TDs in GaN act as nonradiative recombination centers and could restrict the carrier transitions from the bottom of the conduction band to the topmost jXi-like state. Since the superlattices can reduce the dislocation density, the severe interruption of selective carrier transition was reduced. The other reason is that the superlattices may amplify the anisotropic in-plane compressive strain experienced by the MQWs and consequently give rise to a higher transition energy difference between the jXi-like state and jZi-like state. As a result, the polarized characteristic of a-plane LEDs with superlattices was enhanced.
4. Conclusions
We have demonstrated that the performance enhancement of a-plane LEDs was achieved by the insertion of InGaN/GaN superlattices. The threading dislocation density was reduced from 3 1010to 9 109cm2. From the Arrhenius plots,
the thermal activation energy of the sample with InGaN/ GaN superlattices which was estimated to be 105.3 meV, was larger than that of the sample without InGaN/GaN
0 20 40 60 80 100 0 2 4 6 8 Sample 1 Sample 2 Current (mA) V o lta g e (V) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Output po wer (arb. unit) 350 400 450 500 550 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 EL intensity (arb. uint) Wavelenghth (nm) Sample 1 Sample 2
(a)
(b)
Fig. 4. (Color online) (a) Room-temperature EL spectra and (b) L– I–V curves for Sample 1 and Sample 2.
-50 0 50 100 150 200 250 300 350 400 20 40 60 80 100 120 140 160 CB | X | Z | Y VB Normaliz ed EL intensity (arb. unit)
Polarizer angle (degree)
Sample 1 Sample 2
Fig. 5. (Color online) EL intensity of Sample 1 and Sample 2 at different polarization angles at 20 mA injection current.
superlattices (83.7 meV). Afterward, the EL intensity of the sample with InGaN/GaN superlattices exhibited improve-ment by a factor of 3 in comparison with that of the conventional sample without InGaN/GaN superlattices, which could be attributed to the TD reduction and the improvement of LED quality. Furthermore, we observed that the polarization degree of a-plane LEDs with superlattices (56.3%) was much higher than that of a-plane light emitting diodes without superlattices (27.4%).
Acknowledgements
This work was supported by the Ministry of Economic Affairs of the Republic of China (MOEA). The number of the project was 7301XS1G20 for the nonpolar GaN epitaxy and MOVPE part. Also, the authors would like to thank Professor Chia-Feng Lin at National Chung Hsing Univer-sity (NCHU) for assistance in the PL measurement.
1) T. Takeuchi, S. Sota, M. Katsuragawa, M. Komori, H. Takeuchi, H. Amano, and I. Akasaki:Jpn. J. Appl. Phys. 36 (1997) L382. 2) D. A. B. Miller, D. C. Chemla, T. C. Damen, A. C. Grossard, W.
Wiegmann, T. H. Wood, and C. A. Burrus:Phys. Rev. B 32 (1985) 1043.
3) J. S. Im, H. Kollmer, J. Off, A. Sohmer, F. Scholz, and A. Hangleiter: Phys. Rev. B 57 (1998) 9435.
4) M. D. Craven, S. H. Lim, F. Wu, J. S. Speck, and S. P. DenBaars: Appl. Phys. Lett. 81 (2002) 469.
5) M. D. Craven, A. Chakraborty, B. Imer, F. Wu, S. Keller, U. K. Mishra, J. S. Speck, and S. P. DenBaars:Phys. Status Solidi C 0 (2003) 2132.
6) G. A. Garrett, H. Shen, M. Wraback, B. Imer, B. Haskell, J. S. Speck, S. Keller, S. Nakamura, and S. P. DenBaars:Phys. Status Solidi A 202
(2005) 846.
7) H. M. Ng:Appl. Phys. Lett. 80 (2002) 4369.
8) A. Chakraborty, B. A. Haskell, S. Keller, J. S. Speck, S. P. DenBaars, S. Nakamura, and U. K. Mishra:Appl. Phys. Lett. 85 (2004) 5143. 9) A. Chakraborty, S. Keller, C. Meier, B. A. Haskell, S. Keller, P.
Waltereit, S. P. DenBaars, S. Nakamura, J. S. Speck, and U. K. Mishra:Appl. Phys. Lett. 86 (2005) 031901.
10) B. Rau, P. Waltereit, O. Brandt, M. Ramsteiner, K. H. Ploog, J. Puls, and F. Henneberger:Appl. Phys. Lett. 77 (2000) 3343.
11) Y. J. Sun, O. Brandt, M. Ramsteiner, H. T. Grahn, and K. H. Ploog: Appl. Phys. Lett. 82 (2003) 3850.
12) H. Masui, H. Yamada, K. Iso, S. Nakamura, and S. P. DenBaars:Appl. Phys. Lett. 92 (2008) 091105.
13) M. D. Craven, S. H. Lim, F. Wu, J. S. Speck, and S. P. DenBaars: Appl. Phys. Lett. 81 (2002) 469.
14) M. D. Craven, S. H. Lim, F. Wu, J. S. Speck, and S. P. DenBaars: Appl. Phys. Lett. 81 (2002) 1201.
15) C. Chen, J. Zhang, J. Yang, V. Adivarahan, S. Rai, S. Wu, H. Wang, W. Sun, M. Su, Z. Gong, E. Kuokstis, M. Gaevski, and M. A. Khan: Jpn. J. Appl. Phys. 42 (2003) L818.
16) B. Imer, F. Wu, S. P. DenBaars, and J. S. Speck:Appl. Phys. Lett. 88 (2006) 061908.
17) T. C. Wang, T. C. Lu, T. S. Ko, H. C. Kuo, M. Yu, and S. C. Wang: Appl. Phys. Lett. 89 (2006) 251109.
18) B. A. Haskell, F. Wu, M. D. Craven, S. Matsuda, P. T. Fini, T. Fujii, K. Fujito, S. P. DenBaars, J. S. Speck, and S. Nakamura:Appl. Phys. Lett. 83 (2003) 644.
19) J. W. Matthews and A. E. Blakeslee: J. Cryst. Growth 27 (1974) 118.
20) J. W. Matthews and A. E. Blakeslee: J. Cryst. Growth 32 (1976) 265.
21) S. Weber, W. Limmer, K. Thonke, R. Sauer, K. Panzlaff, G. Bacher, H. P. Meier, and P. Roentgen:Phys. Rev. B 52 (1995) 14739. 22) Y. T. Shih, Y. L. Tsai, C. T. Yuan, C. Y. Chen, C. S. Yang, and W. C.
Chou:J. Appl. Phys. 96 (2004) 7267.
23) I. Y. Gerlovin, Y. K. Dolgikh, V. V. Ovsyankin, Y. P. Efimov, I. V. Ignat’ev, and E. E. Novitskaya:Phys. Solid State 40 (1998) 1041. 24) Y. h. Wu, K. Arai, and T. Yao:Phys. Rev. B 53 (1996) 10485.
