Vous êtes sur la page 1sur 3

Japanese Journal of Applied Physics Vol. 46, No. 7, 2007, pp.

L126L128 #2007 The Japan Society of Applied Physics

JJAP Express Letter

High Power and High External Eciency m-Plane InGaN Light Emitting Diodes
Mathew C. S CHMIDT, Kwang-Choong K IM, Hitoshi SATO, Natalie F ELLOWS, Hisashi M ASUI, Shuji N AKAMURA, Steven P. D ENB AARS, and James S. S PECK
Materials Department, and ERATO JST, UCSB Group, University of California, Santa Barbara, CA 93106, U.S.A. (Received December 17, 2006; accepted January 25, 2007; published online February 9, 2007)

 High power and high eciency nonpolar m-plane 1100 nitride light emitting diodes (LEDs) have been fabricated on low extended defect bulk m-plane GaN substrates. The LEDs were grown by metal organic chemical vapor deposition (MOCVD) using conditions similar to that of c-plane device growth. The output power and external quantum eciency (EQE) of the packaged 300 300 mm2 was 23.7 mW and 38.9%, respectively, at 20 mA. The peak wavelength was 407 nm and <1 nm redshift was observed with change in drive current from 1 20 mA. The EQE shows a minimal drop o at higher currents. [DOI: 10.1143/JJAP.46.L126]
KEYWORDS: GaN, nonpolar, m-plane, LED, high power, high efciency

Nonpolar nitride light emitting diodes (LEDs) have the potential to out perform polar c-plane devices due to their lack of a built-in polarization-related electric elds. Devices grown on c-plane GaN suer from large internal electric elds due to discontinuities in spontaneous and piezoelectric polarization eects which cause charge separation between holes and electrons in quantum wells and limits the radiative recombination eciency.1) This is one of the primary reasons why thin quantum wells (typically 2.5 nm) are employed in c-plane LEDs.2) Additionally, the peak emission wavelength of LEDs grown on c-plane GaN blue-shift with increasing drive current due to screening of the internal elds.3) Nonpolar GaN devices, such as in the m-plane   1100 and a-plane 1120 are free from polarization-related electric elds since the polar c-axis is parallel to any heterointerfaces.4) However, the very limited availability of low defect density nonpolar GaN substrates has hindered the development of high power optical devices. Growth of nonpolar GaN on dissimilar substrates such as m-plane SiC5) or LiAlO2 6,7) yields lms with threading dislocation densities on the order of 1010 cm2 and basal plane stacking fault densities on the order of 105 cm1 . Extended defects, namely threading dislocations, act as non-radiative recombination centers and subsequently lead to low eciencies and low output powers when their density is high.8) The rst m-plane LED, reported by Chakraborty et al. in 2005, was on a free-standing m-plane GaN substrate grown by hydride vapor phase epitaxy (HVPE) with a threading dislocation density of 4 109 cm2 and a stacking fault density of 1 105 cm1 .9) This device emitted at a peak wavelength of 450 nm and produced an output power of 240 mW and an external quantum eciency (EQE) of 0.43% at a 20 mA drive current. More recently, Okamoto et al. reported on a dislocation-free m-plane LED also using an m-plane GaN free-standing substrate.10) The peak wavelength emission of this device was at 435 nm and had an output power of 1.79 mW and an EQE of 3.1% at 20 mA. This is the highest reported output power from an m-plane GaN LED until now. However, the power of this device is roughly an order of magnitude lower than conventional cplane LEDs.11) The present letter reports an output power of 23.7 mW for an m-plane InGaN/GaN multi-quantum well (MQW) LED emitting at a peak wavelength of 407 nm at 20 mA. The EQE was 38.9% at 20 mA. This is the rst report

of a nonpolar nitride LED that has comparable performance to current state-of-the-art c-plane LEDs. The LEDs were grown by metal organic chemical vapor deposition (MOCVD) on free-standing bulk m-plane GaN substrates produced by Mitsubishi Chemical Co., Ltd, which were grown by HVPE. The exposed m-plane surface of these slices was prepared by chemical and mechanical surface treatment techniques. These substrates have threading dislocation densities of less than 5 106 cm2 , carrier concentrations of approximately 1 1017 cm3 , and surface roughness (RMS) of less than 1 nm, as measured by the manufacturer. The growth conditions of the LEDs were very similar to that of optimized c-plane LEDs. All MOCVD growth was performed at atmospheric pressure (AP), at typical V/III ratios (>3000), and at typical temperature ranges (875 to 1185  C). A GaN:Si layer was grown rst on top of the substrate without a low temperature (LT) nucleation layer, followed by a 6-period InGaN/GaN (8/ 18 nm) undoped MQW active region. In contrast to c-plane LEDs, relatively thick 8 nm quantum wells were employed to achieve high eciency devices. After the active region, a 10-nm-thick undoped AlGaN electron blocking layer was grown on top of the last barrier, followed by a 160 nm p-GaN layer. The as grown surface has an RMS roughness of approximately 10 nm on a 100 100 mm atomic force microscope (AFM) scan. An annealed SnO2 -doped In2 O3 (ITO) p-type GaN contact was utilized as a transparent p-electrode. LED mesas were etched by reactive ion etching (RIE). Ti/Al/Au n-contacts were deposited by e-beam evaporation. The LED active area was a conventional 300 300 mm2 design with a U-shaped n-contact, as shown in Fig. 1. After device fabrication, the LEDs were diced and packaged using standard die and wire bonding techniques and then molded with epoxy. The power and emission spectra measurements of the packaged LEDs were performed in an integrating sphere at room temperature under DC operation. Figure 2 shows the dependence of both power and EQE on drive current. At 20 mA, the output power was 23.7 mW and the EQE was 38.9%. Figure 3 shows the IV curve for an m-plane LED. The Vf was 5.2 V at 20 mA. We attribute the high operating voltage to the unoptimized ITO p-contacts. In the case of our m-plane devices, the EQE appears to be constant after approximately 5 20 mA, and does not display as signicant

L126

Jpn. J. Appl. Phys., Vol. 46, No. 7 (2007)

Express Letter

M. C. SCHMIDT et al.

Fig. 4. Dependence of the electroluminescence spectra on LED drive current.

Fig. 1. Top-view image of a processed 300 300 mm2 m-plane LED.

Fig. 2. The dependence of output power (measured in an integrating sphere) and external quantum eciency of 407 nm m-plane LEDs on drive current.

Fig. 5. Peak emission wavelength dependence on LED drive current.

Fig. 3. IV curve of m-plane LED grown on bulk GaN substrate and processed with ITO p-contact.

of a decrease in eciency as seen in typical c-plane LEDs. The slight decrease in eciency at higher current densities is most likely due to the temperature rise caused by excess

heat from power dissipation at currents above 20 mA. This is consistent with the red shift observed at higher drive currents. Figure 4 illustrates the electro-luminescence spectra at increasing drive currents. The peak emission wavelength at various drive currents is shown in Fig. 5. The peak wavelengths at 1 and 20 mA are 407.7 and 407.4 nm, respectively, showing negligible blue-shift. Collectively, these results demonstrate the viability of high performance nonpolar GaN LEDs. The small electroluminescence peak shift with increasing drive current is expected due to the absence of polarization-related internal electric elds in the active region of the device. It is unclear if the fairly at EQE at drive currents above 5 mA is related to the absence of polarization-related electric elds in these structures or is related to the relatively short wavelength of these emitters. In conclusion, we have demonstrated the rst nonpolar nitride LED that has comparable performance to that of current c-plane devices. The LED was grown by MOCVD on a free-standing m-plane bulk GaN substrate. At 20 mA, the output power and EQE were 23.7 mW and 38.9%, respectively. The peak wavelength emission was approximately 407 nm.

L127

Jpn. J. Appl. Phys., Vol. 46, No. 7 (2007)

Express Letter

M. C. SCHMIDT et al.

The authors acknowledge the support of the Solid State Lighting and Display Center at UCSB and the NICP/JST Exploratory Research for Advanced Technology (ERATO) Program.
1) F. Bernardini, V. Fiorentini, and D. Vanderbilt: Phys. Rev. B 56 (1997) R10024. 2) S. Nakamura, M. Senoh, N. Iwasa, and S. Nagahama: Jpn. J. Appl. Phys. 34 (1995) L797. 3) T. Takeuchi, S. Sota, M. Katsuragawa, M. Komori, H. Takeuchi, H. Amano, and I. Akasaki: Jpn. J. Appl. Phys. 36 (1997) L382. 4) P. Walterweit, O. Brandt, A. Trampert, H. T. Grahn, J. Menniger, M. Ramsteiner, M. Reiche, and K. H. Ploog: Nature 406 (2000) 865.

5) B. Imer, F. Wu, M. D. Craven, J. S. Speck, and S. P. DenBaars: Jpn. J. Appl. Phys. 45 (2006) 8644. 6) B. A. Haskell, T. J. Baker, M. B. McLaurin, F. Wu, P. T. Fini, S. P. DenBaars, J. S. Speck, and S. Nakamura: Appl. Phys. Lett. 86 (2005) 111917. 7) B. A. Haskell, A. Chakraborty, F. Wu, H. Sasano, P. T. Fini, S. P. DenBaars, J. S. Speck, and S. Nakamura: J. Electron. Mater. 34 (2005) 357. 8) J. S. Speck and S. J. Rosner: Physica B 273 (1999) 24. 9) A. Chakraborty, B. A. Haskell, S. Keller, J. S. Speck, S. P. DenBaars, S. Nakamura, and U. K. Mishra: Jpn. J. Appl. Phys. 44 (2005) L173. 10) K. Okamoto, H. Ohta, D. Nakagawa, M. Sonobe, J. Ichihara, and H. Takasu: Jpn. J. Appl. Phys. 45 (2006) L1197. 11) Y. Narukawa, J. Narita, T. Sakamoto, K. Deguchi, T. Yamada, and T. Mukai: Jpn. J. Appl. Phys. 45 (2006) L1084.

L128

Vous aimerez peut-être aussi