Heteroepitaxial ZnO transparent current spreading layers with low sheet resistances were deposited on GaN-based light emitting diodes using aqueous solution phase epitaxy at temperatures below 90°C. The performance of the LEDs was analyzed and compared to identical devices using electron-beam evaporated indium tin oxide transparent current spreading layers. White LEDs with ZnO layers provided high luminous efficacy–157 lm/W at 0.5A/cm2, and 84.8 lm/W at 35A/cm2, 24% and 50% higher, respectively, than devices with ITO layers. The improvement appears to be due to the enhanced current spreading and low optical absorption provided by the ZnO.
© 2011 OSA
Highly efficient white light-emitting diodes (LEDs) can reduce energy use in interior, automobile, and display lighting. A common approach to efficiently generate white light is to couple a yellow-emitting phosphor with a high-brightness blue gallium nitride (GaN) based LED [1,2]. The ultimate performance depends on converting electrical energy to photons, and on extracting those photons to the phosphor, and out of the LED package. To efficiently accomplish this, current should be injected uniformly into the active layer of the GaN LED . Because of the low conductivity of p-type GaN, a transparent current-spreading layer is commonly used to minimize current-crowding and to avoid the optical absorption associated with extended top-side metal contacts [4,5]. Because of its combination of transparency and conductivity, indium tin oxide (ITO) is the transparent conductive material usually employed for these layers [6,7]. However, there is still significant room for improvement in the properties of transparent current spreading layers, especially considering the rising cost of ITO. ZnO is a lower cost material that can also display high transparency and electrical conductivity, and is of special interest in the context of GaN optoelectronics because of its closely-matched crystal lattice [8,9], which allows for low-strain epitaxial growth of ZnO . Single crystal epitaxial films should display higher conductivity and transparency compared to polycrystalline films due to a lack of grain boundary scattering. Thompson et al. recently demonstrated that epitaxial ZnO films deposited on GaN LEDs from low temperature aqueous solution served as effective current spreading layers . In this paper, we further investigate ZnO current-spreading layers as an alternative to ITO, and demonstrate that such films can provide significant enhancements in LED performance.
2. Experimental procedure
Blue-emitting (470nm) LED device wafers supplied by a commercial LED manufacturer were used for this research. The device wafers all used the same epitaxial (Al,Ga,In)N device structure grown by MOCVD on 2-inch diameter patterned c-plane sapphire substrates. All directly-compared LED measurements, including those from the control devices with ITO layers, refer to devices fabricated from a single LED wafer. Both the contact and sheet resistance vs. annealing condition optimization studies were also performed on single LED wafers.
ZnO layers were deposited on quarter sections of the wafers by a two-step aqueous solution deposition method, the details of which have been previously reported [11–14]. The first step is designed to create a high density of epitaxial ZnO seed crystals on the GaN surface of the substrate. This was achieved by rapidly initiating the formation of ZnO from a 90°C solution of zinc nitrate and ammonium nitrate with the addition of ammonium hydroxide. This step causes a thin, incompletely-coalesced seed-layer film of ZnO islands to be deposited on the wafer surface. These seed-layer films were then subjected to a 15 minute anneal at 400–600°C in an 80% N2/20% O2 atmosphere in a rapid thermal annealing (RTA) furnace. To produce thick, coalesced films, a second deposition step was performed, under conditions favoring growth on the existing seed-layer. The ZnO-seeded wafer was inserted into an aqueous solution of zinc nitrate, sodium citrate, and ammonium hydroxide at room temperature and placed in a sealed vessel in a 90°C oven for 18hrs. Post-annealing of the wafers was performed under the same conditions at temperatures from 150 to 250°C. The 240nm thick ITO film used for comparison was deposited using electron beam evaporation from a commercial ITO target. The as-deposited ITO layer was annealed at 600°C first for 10min in 80% N2/ 20% O2 to increase transparency, and then for 3min in pure N2 to increase its conductivity. This process has previously been used for high performance LEDs .
Following the ZnO or ITO depositions, the current-spreading and (Al,Ga,In)N layers of the wafers were processed, using several photolithography steps, into either structures for circular transmission line method (CTLM) electrical measurements or individual LED dies. Both processes consisted of a wet chemical etch in a dilute HCl solution to pattern the ZnO or ITO layer, a Cl2 inductively-coupled plasma etch to create mesas by etching trenches through the p-type and active epitaxial layers of the wafer, and electron-beam evaporation of low-resistance Ohmic metal contacts. Contacts to the n-GaN were according to . Contacts to the ZnO and layers were Ti/Au, with 15/400nm thickness. Wafers were ground and polished from the back of the sapphire substrate to 180–200μm thickness. The LEDs were mounted on silver headers using silicone and connected to the header leads by gold wire bonding. White LEDs were fabricated by encapsulating devices in a silicone dome and covering the dome with an yttrium aluminum garnet (YAG) phosphor-impregnated silicone cap.
In addition, the same deposition and processing methods were used to prepare van der Pauw clover-leaf type structures on unintentionally doped GaN/sapphire substrates for Hall measurements. Optical absorption measurements were made on ZnO and ITO films deposited on double-side polished (111) MgAl2O4 spinel substrates using a Shimadzu UV3600 UV-Vis-NIR spectrometer. X-ray diffraction (XRD) data was gathered from the LED wafers in a Panalytical MRD Pro diffractometer. The light-output performance of the LEDs was measured using an integrating sphere detector, under both DC and pulsed current injection with a 5% duty cycle. High aspect ratio LEDs with large contact spacing were tested using a probe station equipped with a microscope in order to observe the degree of current spreading. Monochrome images were captured at 50x magnification using a neutral density filter to prevent saturation of the CCD camera. These images were converted digitally to color-contrast images.
3. Results and discussion
The results of the CTLM electrical measurements examining annealing conditions are summarized in Tables 1 and 2 . The effect of seed layer annealing on contact resistance was examined through measurements on separate wafers. When the seed layer was not annealed, it was found that the ZnO film would occasionally delaminate from the LED wafer during or after the second-layer growth step. Adhesion to the GaN was improved greatly by annealing at temperatures above 400°C. It was found that annealing at temperatures higher than 400°C could decrease the contact resistance between ZnO and GaN—by 63% after a 450°C anneal, and by 70% after a 500°C anneal. This is in line with previously reported results indicating an improvement in crystal quality with annealing but a large increase in film resistivity above 500°C . The seed layer annealing did not appear to systematically affect the sheet resistance of the final ZnO film. All seed-layer annealing conditions resulted in Ohmic contact to the p-GaN, as indicated by their linear I-V behavior. A representative I-V curve of a sample with a 500°C seed layer anneal is shown in Fig. 1(c) . The TLM pattern used to measure ZnO/GaN contact resistance is shown in Fig. 1(a). Figure 1(b) shows the structure of the LED devices that were fabricated.
The effect of different post-processing anneals on sheet resistance are shown in Table 2. These measurements were all performed on the same set of CTLM structures found on a single wafer section. Annealing was performed subsequent to the metal contact deposition. The post-annealing conditions tested were all found to decrease the sheet resistance relative to the un-annealed ZnO. The lowest measured sheet resistance was obtained after the 225°C anneal, but the 250 °C anneal result was very similar. Higher annealing conditions resulted in degradation of the metal contacts, with haziness and cracking in the ZnO occurring at temperatures above 300°C. The film providing the results in Table 2 was produced with a 500°C seed-layer anneal. A Hall effect measurement was used to determine the carrier concentration and mobility of a ZnO film prepared using a 500 °C seed layer anneal and a 250°C post-anneal. In order to minimize the influence of carrier transport in the substrate on the measurement, a 3.1μm thick ZnO film was deposited on a low conductivity unintentionally doped GaN/Sapphire substrate. The Hall measurements determined the ZnO film to have an electron concentration of 2.3x1019 cm−3 with carrier mobility of 71 cm2V−1s−1. This indicates a significant improvement in mobility over sputtered polycrystalline ZnO films, with reported mobilities of ~10cm2V−1s−1 .
The X-ray diffraction results in Fig. 2 were obtained from a ZnO film produced using a 500°C seed-layer anneal and a 250°C post-deposition anneal. The high-resolution 2θ-scan, shown in Fig. 2(a), shows the respective (0002) reflections for ZnO and GaN. No reflections for other ZnO orientations or any other phases where observed. The full width at half-maximum of the ZnO (0002) ω-scan, shown in Fig. 2(b), was 0.365°, indicating epitaxy, but some low angle mosaicity in the out-of-plane orientation. The respective -scans of the ZnO and GaN reflections, shown in Figs. 2(c) and 2(d), display the same 6-fold symmetry, indicating that the ZnO and GaN LED also share the same in-plane crystal orientation. Together, the XRD results show that the ZnO film is nominally single crystalline and has formed epitaxially with the same in-plane and out-of-plane orientation as the GaN layers. This is in agreement with previous results from Kim et al. .
In order to measure the absorption coefficient of the ZnO films, the LED wafers were substituted with high transparency (111) MgAl2O4 single crystal substrates, which also allow for high quality epitaxial c-plane ZnO films using the same aqueous growth conditions . The film measured received a 500°C seed layer anneal and a 250°C post-anneal. The absorption was calculated to be 5.4% at 450 nm, and the film thickness was measured using profilometry to be 5.6 μm. This leads to a calculated absorption coefficient of 99.1 cm−1. In comparison, a 250 nm ITO was measured to have absorb 2.3% at 450nm, giving an absorption coefficient of 930 cm−1. Using the measured absorption coefficient for ZnO to estimate the absorption of the ~2.3 μm thick ZnO films on the LEDs gives 2.3% absorption, similar to the ITO despite being an order of magnitude thicker.
The LEDs with ZnO layers that were characterized were also prepared using a 500°C seed layer anneal and a 250°C post-anneal. The thickness of the ZnO was measured by profilometry after patterned etching to be between 2.1 and 2.3 μm. The e-beam deposited ITO thickness was confirmed to be ~240 nm. Using CTLM patterns included on the same LED wafer, the ZnO and ITO films were found to have respective sheet resistances of 26 and 91Ω/□. The specific contact resistances of the Ti/Au contacts to the ZnO and ITO layers were found to be 3 × 10−6 and 6 × 10−6 Ω-cm2, respectively. The slightly different sheet resistance, compared to the results in Table 2, could be the result of variation in the actual temperature achieved in the RTA furnace, or the thickness of the ZnO films.
The four different LED mesa and metal contact pattern types (insets) and their respective current-voltage (I-V) characteristics are shown in Fig. 3(a) . These four LED Types show the effect of increasing the contact spacing on devices using either a ZnO or ITO current spreading layer. For all four types, the devices with ZnO contacts display lower or similar series resistance compared to the devices with ITO contact. The increase in series resistance with increased spacing is much more significant for the ITO devices. This can be attributed to the higher sheet resistance of the ITO layers, which would be expected to add more to the total series resistance as contact spacing is increased.
The external quantum efficiencies (EQEs) and wall plug efficiencies (WPEs) of the same set of devices are shown respectively in Figs. 3(b) and 3(c). The EQEs and WPEs of all ZnO devices were higher than the otherwise identical ITO devices. The difference in performance was greater for devices with larger contact spacing. For the EQE, which does not take the operating voltage into account, the improvement with ZnO must be attributable to either a reduced degree of current-crowding, due to better current spreading, or otherwise enhanced light extraction. Current-crowding near the metal contacts can decrease EQE, because the higher local current density results in more efficiency droop, as well as more light simply being absorbed by the metal contact. Because the WPE also takes resistive losses into account, the plots in Fig. 3(c) display the combined effect of series resistance and EQE on LED performance. Again, the devices using solution-deposited ZnO show enhanced performance compared to those with ITO. For the Type 4 device, with the largest spacing, the WPE of the ITO device is less than 65% of that of the ZnO device at higher current density.
In order to clearly demonstrate the current spreading capability of the solution deposited ZnO relative to evaporation-deposited ITO, high aspect ratio LEDs with ~0.8 mm contact spacing were fabricated using both types of current spreading layers. Two illuminated LED strips, each using ZnO or ITO are shown in Fig. 4(a) . Neither device shows completely uniform emission resulting from perfect current spreading, as seen in Fig. 4(b), but the emission from the ITO LED clearly shows a higher degree of localized emission near the p-contact. The line scans of intensity shown in Figs. 4(c) and 4(d), which were taken along the center of each device, clearly show a more significant drop in light emission on the ITO sample when moving away from the metal p-contact.
The higher degree of localized light emission that is occurring in the ITO device helps explain the decreased EQEs of the ITO devices relative to the ZnO ones, as well as why the decrease becomes more significant for larger contact spacing. The enhanced current spreading may also explain why the WPEs of the ZnO devices were relatively unaffected by increasing the contact spacing, while those of the ITO devices were decreased.
The luminous efficacy of Type 3 devices encapsulated into a silicon dome and covered by a phosphor cap to produce white LEDs are shown in Table 3 . As expected, since the phosphor efficiency should be constant, these measurements show the same trends as the measured wall-plug efficiencies shown in Fig. 3. The ZnO device achieved 157lm/W at a low current density of 0.5A/cm2—24% higher than the ITO device, which only provided 126lm/W. An even larger improvement of about 50% was seen at the higher current density of 35A/cm2.
GaN LEDs with solution-deposited ZnO and electron-beam evaporated ITO current-spreading layers were fabricated. It was found that the series resistance, EQE, and luminous efficacy were improved in the devices with ZnO current spreading layers compared to those with ITO. This appears to be due to a combination of lower sheet resistance and lower optical absorption for the ZnO layers, which allowed current to be spread and injected into the active region over a larger area while simultaneously allowing more light to escape the LED die. This demonstrates that ZnO current spreading layers could be a viable indium-free alternative to ITO in the context of GaN LEDs.
This research was supported by the Solid State Lighting and Energy Center (SSLEC) at UCSB, as well as the UCSB nanofabrication facility, part of the NSF-funded NNIN network. This work was partially supported by the MRSEC Program of the NSF under Award No. DMR05-20415; a member of the NSF-funded Materials Research Facilities Network. This work was also supported by the NSF under grant No. 095254. We thank Stuart Brinkley and Nathan Pfaff for their help in packaging and characterizing the white LEDs. We also thank Walsin Lihwa Co. for contributing the LED wafers used in this study.
References and links
1. S. Nakamura, S. Pearton, and G. Fasol, The Blue Laser Diode: The Complete Story, 2nd ed. (Springer, 2000)
2. Y. Narukawa, I. Niki, K. Izuno, M. Yamada, Y. Murazaki, and T. Mukai, “Phosphor-conversion white light emitting diode using ingan near-ultraviolet chip,” Jpn. J. Appl. Phys. 41(Part 2, No. 4A), L371–L373 (2002). [CrossRef]
3. S. R. Jeon, Y. H. Song, H. J. Jang, G. M. Yang, S. W. Hwang, and S. J. Son, “Lateral current spreading in GaN-based light-emitting diodes utilizing tunnel contact junctions,” Appl. Phys. Lett. 78(21), 3265 (2001). [CrossRef]
4. J.-C. Chen, G.-J. Sheu, F.-S. Hwu, H.-I. Chen, J.-K. Sheu, T.-X. Lee, and C.-C. Sun, “Electrical-optical analysis of a GaN/sapphire LED chip by considering the resistivity of the current-spreading layer,” Opt. Rev. 16(2), 213–215 (2009). [CrossRef]
5. H. Kim, J. M. Lee, C. Huh, S. W. Kim, D. J. Kim, S. J. Park, and H. Hwang, “Modeling of a GaN-based light-emitting diode for uniform current spreading,” Appl. Phys. Lett. 77(12), 1903–1904 (2000). [CrossRef]
6. T. Margalith, O. Buchinsky, D. A. Cohen, A. C. Abare, M. Hansen, S. P. DenBaars, and L. A. Coldren, “Indium tin oxide contacts to gallium nitride optoelectronic devices,” Appl. Phys. Lett. 74(26), 3930 (1999). [CrossRef]
7. D. W. Kim, Y. J. Sung, J. W. Park, and G. Y. Yeom, “A study of transparent indium tin oxide (ITO) contact to p-GaN,” Thin Solid Films 398–399, 87–92 (2001). [CrossRef]
8. S. J. Pearton, D. P. Norton, K. Ip, Y. W. Heo, and T. Steiner, “Recent progress in processing and properties of ZnO,” Prog. Mater. Sci. 50(3), 293–340 (2005). [CrossRef]
9. P. L. Washington, H. C. Ong, J. Y. Dai, and R. P. H. Chang, “Determination of the optical constants of zinc oxide thin films by spectroscopic ellipsometry,” Appl. Phys. Lett. 72(25), 3261 (1998). [CrossRef]
10. R. D. Vispute, V. Talyansky, S. Choopun, R. P. Sharma, T. Venkatesan, M. He, X. Tang, J. B. Halpern, M. G. Spencer, Y. X. Li, L. G. Salamanca-Riba, A. A. Iliadis, and K. A. Jones, “Heteroepitaxy of ZnO on GaN and its implications for fabrication of hybrid optoelectronic devices,” Appl. Phys. Lett. 73(3), 348 (1998). [CrossRef]
11. D. B. Thompson, J. J. Richardson, S. P. DenBaars, and F. F. Lange, “Light emitting diodes with ZnO current spreading layers deposited from a low temperature aqueous solution,” Appl. Phys. Express 2, 042101 (2009). [CrossRef]
12. J. J. Richardson and F. F. Lange, “Controlling low temperature aqueous synthesis of ZnO: 1. thermodynamic analysis,” Cryst. Growth Des. 9(6), 2570–2575 (2009). [CrossRef]
13. J. J. Richardson and F. F. Lange, “Controlling low temperature aqueous synthesis of ZnO: 2. a novel continuous circulation reactor,” Cryst. Growth Des. 9(6), 2576–2581 (2009). [CrossRef]
14. J. H. Kim, E. M. Kim, D. Andeen, D. Thomson, S. P. DenBaars, and F. F. Lange, “Growth of Heteroepitaxial ZnO Thin Films on GaN-Buffered Al2O3 (0001) Substrates by Low-Temperature Hydrothermal Synthesis at 90°C,” Adv. Funct. Mater. 17(3), 463–471 (2007). [CrossRef]
15. K. J. Vampola, N. N. Fellows, H. Masui, S. E. Brinkley, M. Furukawa, R. B. Chung, H. Sato, J. Sonoda, H. Hirasawa, M. Iza, S. P. DenBaars, and S. Nakamura, “Highly efficient broad-area blue and white light-emitting diodes on bulk GaN substrates,” Phys. Status Solidi A 206(2), 200–202 (2009). [CrossRef]
16. Z. Fan, S. N. Mohammad, W. Kim, Ö. Aktas, A. Botchkarev, and H. Morkoç, “Very low resistance multilayer Ohmic contact to n-GaN,” Appl. Phys. Lett. 68(12), 1672 (1996). [CrossRef]
17. D. H. Zhang and D. E. Brodie, “Effects of annealing ZnO films prepared by ion-beam-assisted reactive deposition,” Thin Solid Films 238(1), 95–100 (1994). [CrossRef]
18. O. Caporaletti, “Electrical and optical properties of bias sputtered ZnO thin films,” Sol. Energy Mater. 7(1), 65–73 (1982). [CrossRef]
19. D. J. Andeen, J. H. Kim, F. F. Lange, G. K. L. Goh, and S. Tripathy, “Lateral Epitaxial Overgrowth of ZnO in Water at 90°C,” Adv. Funct. Mater. 16(6), 799–804 (2006). [CrossRef]