GaN Semiconductors Heterojunction White

Subscriber access provided by READING UNIV

Article

Monolithic Inorganic ZnO/GaN Semiconductors Heterojunction White Light-Emitting Diodes Seonghoon Jeong, Seung Kyu Oh, Jae-Hyun Ryou, Kwang-Soon Ahn, Keun Man Song, and Hyunsoo Kim ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b15946 • Publication Date (Web): 10 Jan 2018 Downloaded from http://pubs.acs.org on January 11, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Monolithic Inorganic ZnO/GaN Semiconductors Heterojunction White Light-Emitting Diodes

Seonghoon Jeong,† Seung Kyu Oh,‡ Jae-Hyun Ryou,‡, § Kwang-Soon Ahn,,* Keun Man Song,⊥ and Hyunsoo Kim†,* †

School of Semiconductor and Chemical Engineering, Semiconductor Physics Research Center, Chonbuk National University, Jeonju 561-756, Korea

‡

Department of Mechanical Engineering, University of Houston, Houston, Texas 77204-4006, USA

§

Materials Science and Engineering Program and Texas Center for Superconductivity at UH (TcSUH), University of Houston, Houston, Texas 77204, USA School of Chemical Engineering, Yeungnam University, Gyeongsan 712-749, South Korea



⊥

Korea Advanced Nano Fab Center, Suwon 443-700, Republic of Korea

KEYWORDS: monolithic, white light emitting diodes, heterojunction, ZnO, GaN,

 ABSTRACT Monolithic light-emitting diodes (LEDs) that can generate white color at the one-chip level without wavelength conversion through packaged phosphors or chip integration for photon recycling are of particular importance to produce compact, cost-competitive, and smart lighting sources. In this study, monolithic white LEDs were developed based on ZnO/GaN semiconductor heterojunctions. The electroluminescence (EL) wavelength of the ZnO/GaN heterojunction could be tuned by a post-thermal annealing process, causing the generation of an interfacial Ga2O3 layer. Ultraviolet, violet-bluish, and greenish-yellow broad bands were observed from n-ZnO/p-GaN without an interfacial layer, while a strong greenish-yellow band emission was the only one observed from that with an interfacial layer. By controlled integration of ZnO/GaN heterojunctions with different post-annealing conditions, monolithic white LED

1

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

was demonstrated with color coordinates in the range (0.3534, 0.3710)–(0.4197, 0.4080) and color temperatures of 4778–3349 K in the CIE 1931 chromaticity diagram. Furthermore, the monolithic white LED produced approximately 2.1 times higher optical output power than a conventional ZnO/GaN heterojunction, due to the carrier confinement effect at the Ga2O3/nZnO interface.

 AUTHOR INFORMATION Corresponding Authors *

E-mail: [email protected], [email protected]

ORCID Hyunsoo Kim: 0000-0001-7142-7157 Notes The authors declare no competing financial interest.

2


Page 2 of 28

Page 3 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


 INTRODUCTION White light-emitting diodes (LEDs) nowadays become ubiquitous in everyday life as they are widely used in displays and general illumination.1,2 White LEDs are typically made with either inorganic III-nitride compound semiconductors1-4 or organic semiconductors for selected applications such as flexible, transparent, and planar lighting sources.5 The advantages of inorganic semiconductor white LEDs over organic LEDs are their higher external quantum efficiency, robustness, and long-term stability, which make them more suitable for general lighting applications. However, the inorganic white LEDs usually suffer from high production cost due to the requirements of delicate vacuum-based system for epitaxial growth of semiconductor layers, complicated chip fabrication, and packaging processes.6 The fabrication of more cost-competitive and brighter lighting sources using inorganic semiconductors would be the ultimate quest for the LED industry. Conventional inorganic white LED lamps are composed of blue LED chips and encapsulated phosphors where photons from the chip optically pumps phosphors to emit yellow light which is in turn mixed with blue light from the chip.1-4 In this case, however, the overall output efficiency of the white LEDs essentially is lowered, since the conversion efficiency of phosphors is low in the range of 50–80% caused by large Stokes loss. Accordingly, heat can be generated at the phosphors or encapsulants, resulting in package degradation.7 In addition, a delicate packaging process, which leads to higher packaging cost, is required to precisely control the amount of phosphor for an intended color conversion. To overcome these problems, several research groups have investigated phosphor-free white LEDs (so-called monolithic white LEDs), e.g., vertical stacks of multi-color-emitting GaN/InGaN quantum wells,8 photonrecycling concept using InGaN and AlGaInP active layers,9 or the integration of nanomaterials or nanostructures such as quantum dots,10 nanowires,11 and pyramids.12 However, these

3



methods are complicated and expensive for practical applications, while the efficiencies of the white LEDs are yet to be significantly improved to be comparable to those of phosphorconverted LEDs. Recently, ZnO/GaN heterojunction LEDs were extensively investigated owing to their potential merits including simple fabrication process, design capability to form efficient hybrid structures, tunability of emission wavelength, and possible use of ZnO semiconductors that are cheap, stable, and have large exciton binding energy (60 meV).13-19 Specifically, ZnO/GaN heterojunction LEDs have been found to produce various emission wavelengths such as a violet bluish band at around 430 nm, ultraviolet (UV) at around 370 nm, broad greenish yellow at around 560 nm, and red at around 700 nm depending on their structural design and/or process conditions. These results indicate that once the particular wavelength of interest is properly mixed, the monolithic white LED can be successfully demonstrated. In the present study, we investigate a method to fabricate monolithic white LEDs using ZnO/GaN heterojunctions. Investigations of the EL spectra and structural analysis reveals that the presence of interfacial Ga2O3 in the heterojunction layer plays a crucial role in determining the predominant emission wavelength. All possible emissions, e.g., 370, 430, 550–720 nm, is achieved from both p-GaN and n-ZnO sides when a negligible interfacial layer is formed. Conversely, a much stronger and singular broad emission at 550–720 nm is observed with a thick interfacial layer. By controlled combination of ZnO/GaN heterojunctions with and without interfacial layers within one chip, monolithic white LEDs can be demonstrated. Dual visible color emissions with peak wavelengths of ~430 and 550–720 nm are realized, while unnecessary UV emission can be completely suppressed. Remarkably, the color coordinates in the Commission Internationale de l’Eclairage (CIE) 1931 color space chromaticity diagram and the corresponding color temperatures can be controlled by adjusting the design factor. Moreover, the overall output efficiency can be increased by a factor of two as a result of the carrier

4


Page 4 of 28

Page 5 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


confinement effect at the Ga2O3/ZnO interface, suggesting that our novel method will open a new avenue for phosphor-free monolithic white LEDs.

 EXPERIMENTAL SECTION A p-type GaN wafer grown on c-plane sapphire (Al2O3) substrate by metal organic chemical vapor deposition (MOCVD) was used in this study. The structure consisted of 0.8-µm-thick Mg-doped p-GaN and 1-µm-thick unintentionally doped GaN on a c-Al2O3 substrate. Halleffect measurement of p-GaN wafer yielded a carrier concentration (p) of 2.7 × 1017 cm–3 and Hall mobility (µ) of 12 cm2·V–1·s–1. Prior to the fabrication of p-GaN/n-ZnO heterojunction LEDs, a 100-nm-thick SiO2 layer was e-beam evaporated onto the entire p-GaN wafer to prevent sputtering damage of the surface. Then, the circular active layer (diameter = 200 µm) was patterned and selectively etched until p-GaN was exposed, using conventional photolithographic technique and wet etching using HCl:deionized (DI) water (1:20) solution (see Figure 1a). Next, a 150-nm-thick ZnO film was deposited over the entire p-GaN wafer at room temperature with a radio frequency (RF) magnetron sputtering system. Sputtering was performed using an Al2O3 (1 wt%)-doped ZnO target at a working pressure of 10 mTorr, Ar/O2 gas ratio of 2/1, and RF power of 120 W. After ZnO deposition, the sample was selectively wetetched using HCl:DI water solution and buffered oxide etchant to expose the p-GaN surface for Ohmic contact, followed by rapid thermal annealing (RTA) at 850 or 950 °C for 3 min in N2. Note that post-RTA would be the key process for tuning the EL of the heterojunction, although it was originally performed to optimize the epitaxial quality and electrical conductivity of the n-ZnO layer. Indeed, to optimize the n-ZnO layer, the samples were annealed in either N2 or O2 ambient. However, the n-ZnO layer annealed in O2 produced poorer electrical resistivity than that did in N2, resulting in even non-Ohmic contact to n-ZnO (annealed in O2). This could be attributed to the significantly reduced electron concentration, which is possibly by the elimination of oxygen vacancies (VO) present in ZnO.20 Therefore, all RTA process was 5



Page 6 of 28

performed in N2 in this experiment. Hereinafter, the ZnO/GaN heterojunction fabricated with post-RTA temperatures of 950 and 850 °C will be referred to as HJ-a and HJ-b, respectively. For the electrical p-contact, 50/30-nm-thick Ni/Au layers were deposited on p-GaN using an ebeam evaporator followed by RTA at 550 °C for 1 min in O2. Finally, 30/50-nm-thick Ti/Au layers were deposited on n-ZnO as an n-type contact. The electrical, optical, and structural properties of ZnO films were investigated using Hall-effect measurement (CMT-SR1000N), photoluminescence (PL, SpectroPro 500i), and Xray diffraction (XRD, X’pert Pro Powder). The interface of the ZnO/GaN heterojunction was investigated using secondary-ion mass spectrometry (SIMS, IMS 7f) and transmission electron microscopy (TEM, JEM-2100F). The contact properties of Ni/Au formed on p-GaN were evaluated by the transmission line model method.21 The monolithic white LED was made by combining HJ-a and HJ-b in one chip, as discussed later in detail. The electrical and optical characteristics of LEDs were measured using a probe-station system equipped with a parameter analyzer (HP4156A), photodiode (883-UV), and optical spectrometer (Ocean Optics USB2000).

 RESULTS AND DISCUSSION Figure 1b shows the current-voltage (I−V) curves of LEDs fabricated with HJ-a and HJ-b. First, note that the I−V curve of HJ-a LED shows an excellent characteristic in terms of current injection. The threshold voltage is ~3.0 V and current density (J) at 5 V is ~8.1 A/cm2, which is believed to the best result among ZnO/GaN heterojunctions.13-19,

22-24

This is due to the

combined effect of efficient current injection by the surrounding circular p-contact structure and the formation of an Ohmic contact to p-GaN avoiding sputtering damage.22 The specific contact resistance of Ni/Au contact to p-GaN is as low as 9.8 × 10−3 Ω·cm2. Second, HJ-a LED shows significantly reduced resistance in forward I−V curve as compared to HJ-b LED. This is

6


Page 7 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


attributed to reduced electrical resistivity of the n-ZnO layer (ρn-ZnO) after higher temperature annealing. The resistivity values of n-ZnO, ρn-ZnO annealed at 950 °C and 850 °C are 1.25 × 10−2 Ω·cm and 1.96 × 10−2 Ω·cm, respectively. In addition, Hall-effect measurement yields n = 1.25 × 1019 cm–3, µ = 36.1 cm2·V–1·s–1 when annealed at 950 °C and n = 1.97 × 1019 cm–3, µ = 16.3 cm2·V–1·s–1 when annealed at 850 °C. Given the relation 1/ρn-ZnO = qnµ, where q is the electronic charge, the 36% lower ρn-ZnO of the 950 °C-annealed sample than that of the 850 °Cannealed sample is attributed to the 122% higher µ value in spite of the 37% lower n value. This indicates that higher temperature annealing is more effective for the improvement of the ZnO layer quality by reducing crystalline defects present in ZnO, typically, donor-like point defects such as VO. To further support our hypothesis, room-temperature PL measurements were performed for the ZnO layers sputtered on sapphire substrate, as shown in Figure 1c. The ZnO layer annealed at 950 °C shows significantly higher peak intensity than that annealed at 850 °C. The XRD spectra for the (002) ZnO plane also shows higher peak intensity in the 950 °Cannealed sample than in the 850 °C-annealed one (see Figure 1d). These results indicate that the crystalline defects present in ZnO must be reduced by higher temperature annealing. In Figure 1c, the PL intensity of n-ZnO is approximately 7 times higher than that of p-GaN, which is due to the higher carrier density and larger exciton binding energy of n-ZnO than those of pGaN.18 Lastly, despite the improved crystallinity of the n-ZnO layer after RTA at 950 °C, the reverse leakage current becomes more significant in HJ-a LED than in HJ-b LED (see the inset of Figure 1b). This contradictory result suggests that the interface of the ZnO/GaN heterojunction may play a crucial role in carrier transport compared to the bulk or surface of the ZnO layer. Indeed, this hypothesis is quite important to explain the EL mechanism of our samples. Figure 2a and 2b show room-temperature EL spectra of HJ-a LED and HJ-b LED as 7



a function of injection current. The optical oscillation is due to interference fringes. The insets show the EL images of fabricated LEDs taken at I = 20 mA. Note that both LEDs have quite different EL spectra. Specifically, HJ-a LED shows a greenish yellow band in the very broad wavelength range 500–750 nm with a peak emission wavelength of ~570 nm. This broad band may originate from all possible radiative recombinations through deep-level defects present in n-ZnO,13,14, 22-24 including optical transitions between the conduction band (EC) and VO (750 nm), VO and valence band (EV) (~700 nm), Zn interstitials (Zni) and O interstitials (Oi) (601 nm), Zni and VO (551 nm), EC and Oi (543 nm), and EC and VO (502 nm). On the other hand, HJ-b LED shows additional emission peaks including the UV (~370 nm), violet-bluish band (~400–460 nm) besides broad greenish-yellow band (500–750 nm). The UV and violet-bluish emissions may originate from near band-edge transition in n-ZnO and from the optical transition involving Mg-related deep-level defects in p-GaN, respectively.13,14, 22-24 It is also meaningful to discuss the EL spectra of the LEDs as a function of the injection current. For HJ-b LED, at a low injection current of 5 mA, the EL peak starts to appear at around 500–750 nm (greenish-yellow band). At 10 mA, the violet-blue emission at around 420 nm appears. At a higher injection current of 20 mA, the violet-blue emission begins to overcome the greenish-yellow band. In addition, the UV emission starts to appear at around 375 nm and becomes stronger with increasing current. Consequently, at 60 mA, the predominant EL emission is UV, while the greenish-yellow band is the weakest. This indicates that the predominant radiative recombination occurs sequentially from the deep-level defects in ZnO, through the Mg-related deep-level defects in p-GaN, to the near band-edge transition in n-ZnO. On the other hand, the HJ-a LED shows insignificant EL spectra dependence on the injection current. No discernible EL is observed from p-GaN and the near band-edge transition from nZnO. The inspection of the EL spectra raises several questions that should be addressed:

8


Page 8 of 28

Page 9 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


-

Why are the UV and violet-bluish bands not observed in HJ-a LED?

-

Despite the improved crystallinity of the ZnO layer in HJ-a, why does the greenishyellow band become much stronger than that of HJ-b?

-

Why is the integrated EL intensity of HJ-a approximately two times higher than that of HJ-b?

-

According to the PL spectrum of ZnO, the greenish-yellow band is negligibly small when compared to the band-edge emission (Figure 1c). However, why does the greenish-yellow band associated with the deep-level related emission in ZnO dominate the EL spectrum? To elucidate the EL mechanism, we performed structural analysis of ZnO/GaN

heterojunction. Figure 3a and 3b show cross-sectional TEM images of samples HJ-a and HJ-b. Both samples show nearly the same microstructure in the large scale view. The thickness of ZnO is 150 nm. However, energy-dispersive X-ray spectroscopy (EDX) depth profiling near the ZnO/GaN interface shows different characteristics in terms of the O content, as shown in Figure 3c. For example, at position α or β—on the ZnO or GaN side adjacent to the ZnO/GaN interface—both samples show nearly the same O amounts. However, at position γ (on the GaN side approximately 20 nm underneath ZnO/GaN interface), while sample HJ-b shows no measurable O content, sample HJ-a shows the same amount of O as those at positions α and β. This indicates that O significantly out-diffused from ZnO to GaN in sample HJ-a. Accordingly, sample HJ-a shows the clear presence of an interfacial layer, shown in Figure 3d. The interfacial layer is clearly observed at the ZnO/GaN interface of sample HJ-a (see arrow 1) and is found to extend toward the GaN side to a thickness of ~20 nm (see the arrow 2). On the other hand, sample HJ-b shows insignificant generation of an interfacial layer except at the exact position of the ZnO/GaN interface. Figure 3e shows the SIMS depth profiles of sample HJ-a before and after RTA at 950

9



Page 10 of 28

°C. Consistent with the TEM observation, significant inter-diffusion is observed at the ZnO/GaN interface with annealing, particularly for the Ga, O, and Zn atoms. Therefore, the GaO phase is regarded to form at the interface while a highly defective layer localized at the ZnO interface is expected to form as a result of significant O and Zn out-diffusion from ZnO. The presence of an interfacial Ga-O phase was further investigated using XRD θ/2θ scans of HJ-a and HJ-b (see Figure 3f). The predominant peaks for HJ-a are located at 2θ = 34.60°, 34.69°, and 38.35°, verified to be (002) ZnO, (002) GaN, and (31-1) Ga2O3 phases, respectively.25 In contrast, HJ-b shows negligible Ga2O3 peak intensity. These results clearly indicate that the main structural difference between HJ-a and HJ-b is the presence/absence of an interfacial 20 nm-thick Ga2O3 layer. Here, the Ga2O3 layer could be formed by the thermodynamic interaction between out-diffused O (from ZnO) and GaN, since the thermodynamic tendency to form Ga2O3 is more favorable than GaN. As a result, the O vacancy is assumed to generate on the adjacent ZnO surface. Actually, the RTA process performed at above 950 °C may generate thicker Ga2O3 layer. However, this could not be done because of the limitation of our RTA system. On the one hand, according to the literatures, the β-Ga2O3 phase was shown to form on thick GaN films (when exposed to dry air) at temperatures of 900-1000 °C.26,27 In addition, thermodynamically, GaN phase was shown to be unstable at 890 °C (at 1 atm), i.e., the decomposition of GaN into liquid Ga and N2 gas occurs. These reports suggest that the RTA at 950 °C might be enough to generate interfacial Ga2O3.28 Based on the structural features of the heterojunction, the energy band diagram can be constructed under forward bias condition as shown in Figure 4, according to the Anderson model and the following relation29:

∆EC = χZnO − χGa2O3

(1)

∆EV = Eg,Ga2O3 + χGa2O3 − Eg, p-GaN − χp-GaN,

(2)

where ∆EC represents the conduction band offset at the n-ZnO/Ga2O3 interface and ∆EV

10


Page 11 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


represents the valence band offset at the Ga2O3/p-GaN interface. Here, the bandgap energies (Eg) of n-ZnO, Ga2O3, and p-GaN were set at 3.37, 5.0, and 3.4 eV, and the electron affinity (χ) of each layer was 4.35, 2.7, and 4.2 eV, respectively. The calculated ∆EC (1.65 eV) is much larger than ∆EV (0.1 eV), indicating that the interfacial Ga2O3 layer can act as an excellent electron blocking layer (EBL),30,31 while it can allow hole transport from p-GaN to n-ZnO. In addition, substantial deep-level states are expected to localize at ZnO of the Ga2O3/ZnO interface because significant out-diffusion of O and Zn atoms from ZnO to the GaN side generates crystalline imperfections in the ZnO. Consequently, the presence of Ga2O3 EBL and the localized deep-level states at the ZnO interface can explain all the possible EL mechanisms. At low forward bias voltages, holes are likely to move from p-GaN to n-ZnO overcoming the low ∆EV, while electrons accumulate near the Ga2O3/ZnO interface owing to the high ∆EC. Thereby, the carriers confined near the Ga2O3/ZnO interface recombine through the deep-level states residing at the ZnO interface resulting in a broad greenish-yellow band emission (denoted recombination I). With increasing bias voltage, electrons from ZnO begin to reach GaN via tunneling of the Ga2O3 EBL, resulting in a violet-blue emission via recombination through Mg-related deep-level defects in p-GaN (recombination II). At very high bias voltages, hole injection from p-GaN to n-ZnO significantly increases, causing near-band transitions or UV emission on the ZnO side (recombination III). Here, recombination II is expected to occur preferentially over III because of its much higher electron density, higher mobility (associated with lighter effective mass) of electrons over holes, and larger exciton binding energy of ZnO over GaN. In sample HJ-b, all recombination processes (I, II, and III) are expected to occur sequentially depending on the forward bias since the interfacial barrier is thin enough to tunnel through. However, in sample HJ-a, only recombination processes I and III will be available to occur owing to its much thicker interfacial barrier, i.e., nearly complete suppression of electron

11



tunneling. In this case, the radiative recombination efficiency is expected to increase as a result of the carrier confinement effect,3 namely, effective suppression of carrier diffusion over long distances across the neutral region of n-ZnO sides that can increase the probability of nonradiative recombination. Notably, for the sample HJ-a, the recombination process III was also negligibly observed even at the highest injection current of 60 mA. Actually, at this current levels (or high bias voltage), the recombination process III should be observable since a number of injected holes are likely to overcome the ∆EV and to reach the neutral region ZnO layer. These findings suggest that, for the heterojunction having interfacial Ga2O3 layer, the recombination process I is quite efficient as compared to II or III, namely, most injected holes are consumed via recombination process I before they start process III. Our hypothesis that the deep-level states are localized at the ZnO interface can explain the different PL and EL spectra. For example, the PL spectrum reflects the optical and structural properties of ZnO surface or bulk regions (because of the limited penetration depth of the excitation laser source, i.e., 67 nm),23 while the EL spectrum is influenced by the carrier recombination at the Ga2O3/ZnO interface with a highly defective layer. Further, the higher leakage currents of HJ-a LED over HJ-b LED are also understood in terms of the presence of a thick interfacial layer in which a number of deep-level states can induce generation/recombination currents via trap-and-release events of charge carriers. It is evident that the post-RTA of the ZnO/GaN heterojunction can easily tune the predominant emission wavelength and output efficiency. For the purpose of fabricating white LEDs, however, HJ-a is not suitable due to the lack of bluish color. On the one hand, HJ-b is likely to be more suitable to make the white color. However, HJ-b has the critical demerit of unnecessary UV emission and weak overall output efficiency. To combine the advantages of each LED, we designed and fabricated a monolithic white LED by integrating HJ-a and HJ-b within one chip, as schematically described in Figure 5a. After forming a gear-shaped HJ-a on

12


Page 12 of 28

Page 13 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


a substrate by sequential processes of ZnO sputtering, patterning, wet etching, and RTA at 950 °C, the circle-shaped HJ-b was then stacked onto the HJ-a and uncovered region by the same procedures. The second RTA process performed at 850 °C to form HJ-b should not alter the characteristics of the HJ-a due to its lower thermal budget. The monolithic white LED was completed by forming n- and p-electrodes. The diameter of the active region was 200 µm and overall chip dimension was 400 µm × 400 µm. To further tune the white color, the relative area ratio of HJ-a over HJ-b (AHJ-a/AHJ-b) was changed with 1/9, 3/7, 5/5, 7/3, and 9/1 (see Figure 5b). The monolithic white LED was not packaged, i.e., the electrical and optical characteristics were measured on-wafer condition. Figure 5c shows the EL images of monolithic white LEDs taken at 20 mA. It is evident that the two active regions emit distinctive colors as shown in Figure 2, indicating that the monolithic white LEDs are constructed. The typical reverse leakage current of monolithic white LEDs measured at −4 V was around −3 µA. The LEDs show consistent I−V curves and EL spectra, as shown in Figure 5d and 5e, i.e., larger AHJ-a/AHJ-b yields steeper I−V curves and stronger greenish-yellow emission (or weaker violet-blue emission). The EL was taken at an injection current of 20 mA at room temperature. In addition, the optical output power of the monolithic white LED (AHJ-a/AHJ-b=5/5) was 2.1 times higher than that of the pure HJ-b LED, which is a typical heterojunction structure (see Figure 5f). This indicates that the use of HJ-a or an interfacial Ga2O3 layer can also be beneficial in terms of output enhancement for monolithic white LEDs. In addition, the HJ-b LED showed insignificant output increase at higher injection current (e.g., > 30 mA), while the HJ-a LED showed noticeable and gradual output increase with increasing current. This finding shows that the Joule heating would be much less significant in the HJ-a LED than HJ-b LED.32 Nevertheless, further inspection of EL images—particularly for monolithic white LEDs with AHJ-a/AHJ-b=1/9 (see Figure 5a)—show that current crowding occurred at the circular mesa edge, giving a ring-like emission pattern

13



with a violet-bluish color. This is due to the mismatch of electrical resistivity between p-GaN and n-ZnO, causing most current to flow through the top n-ZnO layer to the surrounding circular p-contact.33 This indicates that proper device design should be taken into consideration for better device performance, which is presently under investigation. Based on the EL spectra, the colors of white LEDs can be expressed in the CIE 1931 color space chromaticity diagram as shown in Figure 6a. The chromaticity coordinates are close to the Planckian locus and change depending on AHJ-a/AHJ-b, i.e., (0.3534, 0.3710), (0.3857, 0.3841), (0.3912, 0.3906), (0.4197, 0.4080), and (0.4426, 0.4376) with correlated color temperatures (TC) of 4778, 3922, 3831, 3349, and 3165 K when AHJ-a/AHJ-b = 9/1, 7/3, 5/5, 3/7, and 1/9, respectively. These results indicate that relatively warm white colors can be obtained from our monolithic white LEDs, which is consistent with actual photographic images of luminescent LEDs (taken at 20 mA) as shown in Figure 6b. Although the monolithic white LEDs fabricated with ZnO/GaN heterojunctions are very valuable in terms of compact integration and cost-competition against conventional while LED lamps (comprising GaN-based blue LED chip and phosphors), it’s still necessary to improve the output efficiency. The exact luminous efficiency of monolithic white LED could not be obtained at the moment because of unavailable package process. Nevertheless, it was certain that, based on relative comparison of EL under nominal operation condition (by naked eyes) between monolithic white LED (wafer probing) and commercially available white LED lamp, the overall output power of monolithic LED was quite low. Essentially, the relatively poor output efficiency of monolithic white LED is due to the poor carrier injection through pGaN (causing current crowding and degraded I-V curves) and possible presence of nonradiative recombination centers at the Ga2O3/ZnO active region because the recombination process I requires a lot of deep-level defects. Indeed, this may also limit the reproducibility of HJ LED, indicating that careful device design and process optimization should be further carried out.

14


Page 14 of 28

Page 15 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In addition, it would be also meaningful to address the reliability of monolithic white LED because it is the critical factor for a commercialization. According to our previous studies concerning about the reliability of commercial-grade white LED lamp,34-38 the primary failure origins were the generation of deep-level states (or defects) at the InGaN/GaN active regions and/or the degradation of packaging materials. These two degradation mechanisms were also found to strongly depend on the junction temperature of LED chip, namely, optical degradation was a thermally activated process. In this connection, we can approximately estimate the reliability of monolithic white LEDs based on several facts observed as following. First, since the monolithic white LEDs produced poorer device resistance (or worse forward voltage) compared to the conventional white LED lamp, the Joule heating would be more significant in the monolithic white LEDs, i.e., Q=I2R, where Q is the generated heat and R is the resistance. Essentially, thereby, less stable operation is expected in the monolithic white LED than the conventional lamp. In addition, the HJ-a is regarded to have structural instability against HJ-b, since a lot of deep-level defects are present at the ZnO surface (near Ga2O3/ZnO interface). This was also responsible for the large reverse leakage current of HJ-a against HJ-b. Actually, three years of room-temperature aging of monolithic white LED, which has been kept in a vacuum package, resulted in a systematic I-V degradation, i.e., the forward voltage at 10 mA before and after aging was 10.93 and 10.87 V for AHJ-a/AHJ-b=1/9, 10.58 and 11.03 V for AHJ-a/AHJ-b=5/5, 9.69 and 10.97 V for AHJ-a/AHJ-b=9/1, respectively. Note that the higher AHJ-a/AHJ-b is, the larger degradation of forward voltage is observed, which is in good agreement with our hypothesis that the HJ-a may have structural instability against HJ-b. These facts indicate that, for practical application, the reliability of heterojunction LEDs should be greatly improved, more specifically, the device resistance should be decreased and the stability of ZnO or Ga2O3/ZnO interface should be improved.

 CONCLUSION

15



A novel method to fabricate monolithic white LEDs using n-ZnO/p-GaN semiconductor heterojunctions was investigated. The EL wavelength of the heterojunction could be determined by the presence of an interfacial Ga2O3 layer, i.e., all possible emissions including UV, violetbluish band, and greenish-yellow broad band were possible with heterojunctions having interfacial layers, while strong greenish-yellow band emission occurred in heterojunctions having thick interfacial layers. By mixing two heterojunctions (with and without interfacial layer) within one chip scale, monolithic white LEDs were demonstrated with predominant dual emission of violet-bluish band and greenish-yellow colors. Accordingly, the monolithic white LED displayed warm white colors that were very close to the Planckian locus, i.e., color coordinates were in the range (0.3534, 0.3710)–(0.4197, 0.4080) and the color temperatures were in the range 4778–3349 K depending on the design factor. Further, the monolithic white LED showed approximately two times higher optical output power than the conventional ZnO/GaN heterojunction, indicating that our novel method holds substantial potential for practical applications.

 ACKNOWLEDGMENTS This research was supported in part by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (2017R1A2B4007182) and in part by the Development of R&D Professionals on LED Convergence Lighting for Shipbuilding/Marine Plant and Marine Environments (Project No: N0001363) funded by the Ministry of TRADE, INDUSTRY & ENERGY(MOTIE, Korea)

 REFERENCES (1) Schubert, E. F.; Kim, J. K. Solid-state light sources getting smart. Science. 2005, 308, 12741278. (2) Pimputkar, S.; Speck, J. S.; DenBaars, S. P.; Nakamura, S. Prospects for LED lighting. Nat.

16


Page 16 of 28

Page 17 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Photonics.2009, 3, 180-182. (3) Schubert, E. F. Light Emitting Diodes; Cambridge University Press, Cambridge, England, 2006. (4) Nakamura, S.; Pearton, S.; Fasol, G. The Blue Laser Diode: The Complete Story; SpringerVerlag: New York, 2000. (5) D’Andrade, B. W.; Forrest, S. R. White organic light-emitting devices for solid-state lighting. Adv. Mater. 2004, 16, 1585-1595. (6) Jung, S.; Song, K.-R.; Lee, S.-N.; Kim, H. Wet chemical etching of semipolar GaN planes to obtain brighter and cost-competitive light emitters. Adv. Mater. 2013, 25, 4470-4476. (7) Meneghini, M.; Trevisanello, L. R.; Meneghesso, G.; Zanoni, E. A review on the reliability of GaN-based LEDs. IEEE Trans. Device Mater. Rel.2008, 8, 323-331. (8) Damilano, B.; Grandjean, N.; Pernot, C.; Massies, J. Monolitic white light emitting diodes based on InGaN/GaN multiple-quantum wells. Jpn. J. Appl. Phys.2001, 40, L918-L920. (9) Guo, X.; Graff, J. W.; Schubert, E. F. Photon recycling semiconductor light emitting diode. IEDM Tech. Dig.1999, IEDM’99, 600-605. (10)Damilano, B.; Grandjean, N.; Semond, F.; Massies, J.; Leroux, M. From visible to white light emission by GaN quantum dots on Si (111) substrate. Appl. Phys. Lett. 1999, 75, 962964. (11)Nguyen, H. P. T.; Cui, K.; Zhang, S.; Djavid, M.; Korinek, A.; Botton, G. A.; Mi, Z. Controlling electron overflow in phosphor-free InGaN/GaN nanowire white light-emitting diodes. Nano Lett.2012, 12, 1317-1323. (12)Lim, S. H.; Ko, Y. H.; Rodriguez, C.; Gong, S. H.; Cho, Y. H. Electrically driven, phosphorfree, white light-emitting diodes using gallium nitride-based double concentric truncated pyramid structures. Light Sci. Appl. 2016, 5, e16030. (13)Vispute, R. D.; Talyansky, V.; Choopun, S.; Sharma, R. P.; Venkatesan, T.; He, M.; Tang,

17



X.; Halpern, J. B.; Spencer, M. G.; Li, Y. X.; Salamanca-Riba, L. G.; Iliadis, A. A.; Jones, K. Heteroepitaxy of ZnO on GaN and its implications for fabrication of hybrid optoelectronic devices. Appl. Phys. Lett. 1998, 73, 348-350. (14)Alivov, Y. I.; Van Nostrand, J. E.; Look, D. C.; Chukichev, M. V.; Ataev, D. M. Observation of 430 nm electroluminescence from ZnO/GaN heterojunction light-emitting diodes. Appl. Phys. Lett. 2003, 83, 2943-2945. (15)Park, W. I.; Yi, G. C. Electroluminescence in n-ZnO nanorod arrays vertically grown on pGaN. Adv. Mater. 2004, 16, 87-90. (16)Hwang, D. K.; Kang, S. H.; Lim, J. H.; Yang, E. J.; Oh, J. Y.; Yang J. H.; Park, S. J. pZnO/n-GaN heterostructure ZnO light-emitting diodes. Appl. Phys. Lett. 2005, 86, 222101. (17)Zhang, X. M.; Lu, M. Y.; Zhang, Y.; Chen, L. J.; Wang, Z. L. Fabrication of a highbrightness blue-light-emitting diode using a ZnO-nanowire array grown on p-GaN thin film. Adv. Mater. 2009, 21, 2767-2770. (18)Choi, Y. S.; Kang, J. W.; Hwang D. K.; Park, S. J. Recent advances in ZnO-based lightemitting diodes. IEEE Trans. Electron Devices. 2010, 57, 26-41. (19)Forsberg, M.; Hemmingsson, C.; Amano, H.; Pozina, G. Dynamic properties of excitons in ZnO/AlGaN/GaN hybrid nanostructures. Sci. Rep. 2015, 5, 7889. (20)Janotti, A.; Van de Walle, CG. Oxygen vacancies in ZnO. Appl. Phys. Lett. 2005, 87, 122102. (21)Schroder, D. K. Semiconductor Material and Device Characterization, second ed.; John Wiley and Sons Inc: New York, 1998. (22)Jeong, S.; Kim, H. Enhanced performance characteristics of n-ZnO/p-GaN heterojunction light0emitting diodes by forming excellent Ohmic contact to p-GaN. Mater. Sci. Semicond. Process. 2015, 39, 771-774. (23)Chen, H. C.; Chen, M. J.; Huang, Y. H.; Sun, W. C.; Li, W. C.; Yang, J. R.; Kuan, H.; Shiojiri,

18


Page 18 of 28

Page 19 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


M. White-light electroluminescence from n-ZnO/p-GaN heterojunction light-emitting diodes at reverse breakdown bias. IEEE Trans. Electron Devices. 2011, 58, 3970-3975. (24)Alvi, N. H.; ul Hasan, K.; Nur, O.; Willander, M. The origin of the red emission in n-ZnO nanotubes/p-GaN white light emitting diodes. Nanoscale Res. Lett. 2011, 6, 130. (25)Geller, S. Golden Book of Phase Transitions; Wroclaw, 2002. (26)Wolter, S. D.; Mohney, S. E.; Venugopalan, H.; Wickenden, A. E.; Koleske, D. D. Kinetic study of the oxidation of gallium nitride in dry air. J. Electrochem. Soc. 1998, 145, 629631. (27)Wolter, S. D.; Luther, B. P.; Waltemyer, D. L.; Onneby, C.; Mohney, S. E.; Molnar, R. J. Xray photoelectron spectroscopy and x-ray diffraction study of the thermal oxide on gallium nitride. Appl. Phys. Lett. 1997, 70, 2156. (28)Karpinski, J.; Porowski, S. High pressure thermodynamics of GaN. J. Cryst. Growth. 1984, 66, 11. (29)Anderson, R. L. Experiments on ge-gaas heterojunctions. Solid State Electron. 1962, 5, 341-351. (30)You, J. B.; Zhang, X. W.; Zhang, S. G.; Wang, J. X.; Yin, Z. G.; Tan, H. R.; Zhang, W. J.; Chu, P. K.; Cui, B.; Wowchak, A. M.; Dabiran, A. M.; Chow, P. P. Improved electroluminescence from n-ZnO/AlN/p-GaN heterojunction light-emitting diodes. Appl. Phys Lett. 2010, 63, 201102. (31)Huang, H.; Fang, G.; Li, S.; Long, H.; Mo, X.; Wang, H.; Li, Y.; Jiang, Q.; Carroll, D.; Wang, J.; Wang, M.; Zhao, X. Ultraviolet/orange bicolour electroluminescence from an nZnO/n-GaN isotype heterojunction light emitting diode. Appl. Phys. Lett. 2011, 99, 263502. (32)Kim, H.; Yang, H.; Huh, C.; Kim, S.; Park, S.; Hwang, H. Electromigration-induced failure of GaN multi-quantum well light emitting diode. Electron. Lett. 2000, 36, 908. (33)Kim, H.; Lee, J. M.; Huh, C.; Kim, S. W.; Kim, D. J.; Park, S. J.; Hwang, H. Modeling of

19



a GaN-based light-emitting diode for uniform current spreading. Appl. Phys. Lett. 2000, 77, 1903. (34)Jung, E.; Ryu, J.; Hong, C.; Kim, H. Optical degradation of phosphor-converted white GaNBased light-emitting diodes under electro-thermal stress. J. Electrochem. Soc. 2011, 158, H132. (35)Jung, E.; Kim, M.; Kim, H. Analysis of contributing factors for determining the reliability characteristics of GaN-based white light-emitting diodes with dual degradation kinetics. IEEE Trans. Electron. Devices. 2013, 60, 186. (36)Jung, E.; Kim, H. Rapid optical degradation of GaN-based light-emitting diodes by a current-crowding-induced self-accelerating thermal process. IEEE Trans. Electron. Devices. 2014, 61, 825. (37)Jung, E.; Kim, H. Analysis of GaN-based light-emitting diodes degraded by generation of deep-level states. Phys. Status Solidi A-Appl. Mat. 2014, 211, 1764. (38)Choi, Y.; Jung, E.; Kim, Y.; Oh, S.; Yoo, M.; Kim, H. Inhomogeneous carrier transport at contact/GaN interfaces and thermal instability of vertical light-emitting diodes. ACS Appl. Mater. Interfaces. 2014, 6, 18673.

20


Page 20 of 28

Page 21 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


 FIGURE CAPTIONS Fig. 1

(a) Schematic of the fabrication procedure for heterojunction LEDs. (b) I−V curves of LEDs fabricated with HJ-a and HJ-b. The inset shows the semi-logarithmic I−V plots. (c) Room-temperature PL spectra and (d) XRD spectra of n-ZnO/sapphire samples annealed at 850 and 950 °C.

Fig. 2 Room-temperature EL spectra of (a) HJ-a LED and (b) HJ-b LED as a function of injection current. The images of EL for fabricated LEDs taken at 20 mA are shown in the inset. Fig. 3

Cross-sectional TEM images of (a) HJ-a and (b) HJ-b; and the (c) corresponding EDX depth profiling in terms of the O element. (d) High-resolution cross-sectional TEM images of HJ-a and HJ-b samples. (e) SIMS depth profiles of sample HJ-a before and after post-RTA at 950 °C. (f) XRD θ/2θ scans of HJ-a and HJ-b.

Fig. 4

Schematic energy band diagram of the n-ZnO/Ga2O3/p-GaN heterojunction under forward bias.

Fig. 5

(a) Schematic of the fabrication procedure for monolithic white LEDs. (b) Design splits and (c) the images of EL for monolithic white LEDs operating at I = 20 mA. (d) I−V curves and (e) EL spectra (I = 20 mA) of monolithic white LEDs with different AHJ-a/AHJ-b ratios. (f) The optical output power versus current plots of HJ-a LED, HJb LED, and monolithic white LED (AHJ-a/AHJ-b=5/5).

Fig. 6

(a) Color coordinates of monolithic white LEDs as a function of AHJ-a/AHJ-b (I = 20 mA) on the CIE 1931 color space chromaticity diagram and (b) corresponding actual photographic images of LED luminescence (taken at I = 20 mA).

21


a p-GaN

p-GaN

n-ZnO Ni/Au

SiO2 deposition & patterning

c

10-3

I (A)

ZnO sputtering

10-6 10-9

HJ-a LED HJ-b LED

-12

10 -20

-10

0 V (V)

10

20

HJ-a LED HJ-b LED -5

0

5

V (V)

10

15

Wet etching & RTA

n-ZnO: 950 oC n-ZnO: 850 oC p-GaN

x7

300

ACS Paragon Plus Environment 400 500 600

Wavelength (nm)

700

Electrodes deposition

d

n-ZnO: 950 oC n-ZnO: 850 oC

XRD intensity (a. u.)

Wafer

b

I (mA)

Ti/Au22 of 28 Page

SiO2

PL intensity (a. u.)

1 2 3 4 5 6 7 20 8 9 10 11 15 12 13 14 10 15 16 17 5 18 19 20 0 21 22 -10 23

ACS Applied Materials & Interfaces n-ZnO

30

32

34

36

2θ (degree)

38

40

aPage8023 of 28ACS Applied Materials & Interfaces

HJ-a LED

HJ-a LED

70

Intensity (a. u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Intensity (a. u.)

b

60 50 40 30 20

Current (mA) 5 10 20 30 40 50 60

20mA

10 0 300

400

500

600

700

800

Wavelength (nm) 25

HJ-b LED Current (mA) 5 10 20 20 30 40 15 50 60 10

HJ-b LED

20mA

5

300

ACS Paragon Plus Environment 400 500 600 700

Wavelength (nm)

800

ZnO α β γ GaN

ZnO

GaN δ

50 nm

c

Page 24 of 28

8

HJ-a HJ-b 6

4

2

50 nm 0

α

β

γ

δ

Position

HJ-a

d 1 2

HJ-b

e

before RTA

Intensity (c/s)

δ

HJ-b


Zn

Ga

O

N

Mg 0 50 after RTA: HJ-a

100

150

200

Zn

250 Ga

O

N Mg

0


Depth (nm)

200

250

f

ZnO (002)

XRD intensity (a. u.)

α β γ

b

Intensity (c/s)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

HJ-a

O atomic percent (%)

a

HJ-a HJ-b

GaN (002)

Ga2O3 (31-1)

32

34

36

2Θ (degree)

38

40

deep-level states Page 25ACS of 28 Applied Materials & Interfaces electron hole

1 ∆EC = 1.65 eV EC 2 I 3 III 4 5 6 7 EV 8 II 9 ∆EV = 0.10 eV 10 11 ACS Paragon Plus Environment 12 13 n-ZnO Ga2O3 14 p-GaN

a

b

HJ-a

AHJ-a/AHJ-b= 1/9

3/7

5/5

7/3

9/1

ZnO & RTA at 950 °C Ti/Au

100 µm

c

Ni/Au

AHJ-a/AHJ-b= 1/9

3/7

5/5

7/3

9/1

Electrodes deposition

e

AHJ-a/AHJ-b 1/9 3/7 5/5 7/3 9/1

I (mA)

EL intensity (a. u.)

d

V (V)

Page 26 of 28

5

10

15

300

ACS Paragon Plus Environment 400 500 600 700

Wavelength (nm)

800

f

HJ-a LED HJ-b LED monolithic LED (5/5)

Output power (a. u.)

p-GaN 1 2 3 4 Wafer 5 6 HJ-b 7 8 9 10 11 12 13 ZnO & RTA at 850 °C 14 15 16 20 AHJ-a/AHJ-b 17 18 1/9 19 15 3/7 20 5/5 21 7/3 9/1 22 10 23 24 25 5 26 27 28 0 29 30 -10 -5 0 31


0

10

20

30

40

Current (mA)

50

60

aPage 27 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


AHJ-a/AHJ-b 9/1 7/3 5/5 3/7 1/9

b


9/1

7/3

5/5

3/7

1/9


ZnO:a

ZnO:b

GaN/ZnO:a GaN/ZnO:b

EL intensity (a. u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Page 28 of 28

p-GaN


600

700

Wavelength (nm)

800

GaN Semiconductors Heterojunction White

Recommend Documents