非極性m平面氮化銦鎵/氮化鎵多重量子井之時間解析電激螢光研究與熱退火對以脈衝有機金屬化學氣相沉積成長的氮極性氮化銦鎵/氮化鎵多重量子井之影響

全文

(1)國立高雄大學應用物理學系研究所碩士論文. 非極性 m 平面氮化銦鎵/氮化鎵多重量子井之時間解析電激螢光研究與熱退火對以脈衝有機金屬化學氣相沉積成長的氮極性氮化銦鎵/氮化鎵多重量子井之影響 Time-resolved Electroluminescence Studies of Nonpolar m-plane InGaN/GaN Multiple Quantum Wells (MQWs) and Annealing effect on Nitrogen-Polar InGaN/GaN MQWs grown by Pulsed Metalorganic Chemical Vapor Deposition.. 研究生：尤昱翔撰指導教授：馮世維博士. 中華民國一百零四年七月.

(2) 非極性 m 平面氮化銦鎵/氮化鎵多重量子井之時間解析電激螢光研究與熱退火對以脈衝有機金屬化學氣相沉積成長的氮極性氮化銦鎵/氮化鎵多重量子井之影響指導教授：馮世維博士國立高雄大學應用物理學研究所. 學生：尤昱翔國立高雄大學應用物理學研究所. 摘要. 首先，我們呈現極性c平面與非極性m平面氮化銦家/氮化鎵多重量子井樣品與元件的掃描電子顯微鏡(SEM)、陰極發光(CL)、原子力顯微鏡(AFM)、光激螢光(PL)、電激螢光(EL)、電流-電壓(I-V)、以及時間解析電激螢光(TREL)的實驗結果。非極性m平面樣品相較於極性c平面樣品有較大的表面粗超度與較弱的CL強度，所以非極性m平面樣品有較高的缺陷密度與較低的晶體品質。陰極發光強度與原子力顯微鏡量測到的結果一致，極性c平面氮化鎵樣品有較高的陰極發光強度與較好的晶體品質。光激螢光的實驗結果顯示，非極性m平面樣品有較大的偏極化率，當溫度從20K升到300K，非極性m平面樣品的發光波長位置有藍移現象，而極性c平面樣品的發光波長位置有紅移現象。時間解析電激螢光的實驗結果顯示，非極性m平面元件比極性c平面元件有較短反應時間，這表示非極性m平面元件具有較好的載子注入效率。由於較大量子侷限史塔克效應 (QCSE)的影響，極性c平面元件有較長的載子複合時間。第二部份，我們呈現氮極性氮化銦家/氮化鎵量子井樣品的電子顯微鏡(SEM)、陰極發光(CL)、原子力顯微鏡(AFM)、X光繞射分析(XRD)、光激螢光(PL)的實驗結果與退. I.

(3) 火的影響。氮極性 t11@ t2 2 樣品比 t11@ t2 5 樣品有較大的表面粗糙度，所以通氮成長時間越短樣品有較大的表面粗糙度。氮極性 t11@ t2 2 樣品在單位面積裡有較多的氮化銦鎵晶粒。氮極性 t11@ t2 2 樣品氮化銦鎵的銦成分較 t11@ t2 5 樣品多。當溫度從20K升到 300K，氮極性 t11@ t2 5 樣品的PL積分強度衰減較快。退火後，兩個樣品表面都變得較為平坦，而CL頻譜的紫外光與可見光波長相對強度發生改變。根據X光繞射分析，氮化銦鎵訊號強度增強。退火後，樣品表面晶粒與發光密度都增加。. 關鍵字：非極性 m 平面氮化銦鎵/氮化鎵多重量子井發光二極體、載子傳輸、反應時間、氮極性氮化銦鎵/氮化鎵多重量子井、快速熱退火. II.

(4) Time-resolved Electroluminescence Studies of Nonpolar m-plane InGaN/GaN Multiple Quantum Wells (MQWs) and Annealing effect on Nitrogen-Polar InGaN/GaN MQWs grown by Pulsed Metalorganic Chemical Vapor Deposition. Advisor: Dr. Shih-Wei Feng Department of Applied Physics National University of Kaohsiung. Student: Yu-Siang You Department of Applied Physics National University of Kaohsiung. ABSTRACT. First, we have shown the experimental results of scanning electron microscope (SEM), cathodoluminescence (CL), atomic force microscopy (AFM), X-ray diffraction (XRD), photoluminescence (PL), electroluminescence (EL), current-voltage (I-V), and time-resolved electroluminescence (TREL) measurements of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples and LEDs. The larger surface roughness and weaker CL intensity of the nonpolar m-plane InGaN/GaN MQW sample than those of the polar c-plane one show a higher defect density and lower sample quality of the nonpolar m-plane one. The result of CL measurement is consistent of that of the AFM measurement. In addition, the DoP of PL of the nonpolar m-plane InGaN/GaN MQW sample is larger than the polar c-plane one. With III.

(5) increasing temperature from 20 to 300 K, PL position of the polar c-plane InGaN/GaN MQW sample is slightly blue-shifted, while that of the nonpolar m-plane one is red-shifted. The shorter response time of the nonpolar m-plane InGaN/GaN MQW LED than that of the polar c-plane one suggest a better injection efficiency. The longer recombination time of the polar c-plane InGaN/GaN MQW LED than that of the nonpolar m-plane one could be due to the larger QCSE and potential distribution in the MQWs. Second, we have shown the experimental results of SEM, CL, XRD, AFM, PL measurements and annealing affect of the N-Polar InGaN/GaN MQW samples. The surface roughness of the N-polar t11@ t2 2 sample is larger than that of the N-polar t11@ t2 5 one. The N-polar InGaN/GaN MQW and t11@ t2 2 InGaN/GaN MQW samples show a larger area ratio of InGaN mounds. The indium content of the N-polar t11@ t2 2 InGaN/GaN MQW sample is larger than those of the N-polar t11@ t2 5 one. With increasing temperature from 20 to 300 K, the integral PL intensity of the N-polar t11@ t2 5 InGaN/GaN MQW sample decays faster than those of the N-polar t11@ t2 2 one. The surface roughness of the two annealed samples becomes smoother. With thermal annealing, the relative intensities of UV and visible peaks in the CL spectra are changed and the InGaN intensity increases in XRD measurement. The mound and light densities of the annealed N-polar samples are larger than those of the as-grown ones.. Keywords: Nonpolar m-plane InGaN/GaN MQW LEDs；Carrier Transport；Response Time；N-Polar InGaN/GaN MQWs；Rapid Thermal Annealing. IV.

(6) Contents 中文摘要………………………………………………………….………………I Abstract………………………………………………………………………...III Contents………………………………………………………………………....V Figure Captions……………………………………………………………......VII Chapter 1 Time-resolved Electroluminescence Studies of Nonpolar m-plane InGaN/GaN Multiple Quantum Wells………………………………………..1 1.1 Introduction………………………………………………………………...…....…..….....1 1.2 Motivation and Investigation Flow Chart………………………………………….……...4 1.3 Sample Structures and Growth Conditions ...………………………………………...…...5 1.4 Anisotropic Characteristics of the Polar c-plane and Nonpolar m-plane InGaN/GaN MQW Samples………………………………………….…………………………….…..5 1.4.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Studies…..5 1.4.2 Atomic Force Microscopy (AFM) and X-ray Diffraction (XRD) Patterns Studies...6 1.4.3 Polarization- and Temperature-dependent Photoluminescence (PL) Studies............6 1.4.3.1 Low-temperature Polarized PL and Degree of Polarization (DoP)...............7 1.4.3.2 Temperature-dependent PL Study.................................................................7 1.5 Device Characteristics and Carrier Transport Properties of the Polar c-plane and Nonpolar m-plane InGaN/GaN MQW LEDs……………………………………………...8 1.5.1 Voltage- and Polarization-dependent Electroluminescence (EL) Studies………….8 1.5.1.1 Voltage-dependent EL…………………………………….………………...8 1.5.1.2 Polarized EL and Degree of Polarization (DoP) …………………………...8 1.5.2 Current-Voltage (I-V) and External Quantum Efficiency (EQE) Studies ……...…9 1.5.3 Time-resolved Electroluminescence (TREL) Study …….……………...…………9 1.6 Discussion and Summary...................................................................................................11 V.

(7) References................................................................................................................................12. Chapter 2 Annealing effect on Nitrogen-Polar InGaN/GaN MQWs grown by Pulsed Metalorganic Chemical Vapor Deposition………………………29 2.1 Introduction ………………………………………………………………...…...…..…..29 2.2 Motivation and Investigation Flow Chart………………………………………….…….31 2.3 Sample Structures and Growth Conditions…………………………………………...….31 2.4 Material Characteristics of Nitrogen-Polar InGaN/GaN MQWs………………………...32 2.4.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Studies...32 2.4.2 X-ray Diffraction (XRD) Patterns...........................................................................33 2.4.3 Atomic Force Microscopy (AFM) Study……………………………………..... . 33 2.5 Temperature-dependent PL Studies Nitrogen-Polar InGaN/GaN MQWs……………….34 2.6 Material Characteristics of Annealed Nitrogen-Polar InGaN/GaN MQWs…………......35 2.6.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Studies...35 2.6.2 X-ray Diffraction (XRD) Patterns...........................................................................36 2.6.3 Atomic Force Microscopy (AFM) Study………………………………………. . 36 2.7 Temperature-dependent PL Studies Annealed Nitrogen-Polar InGaN/GaN MQWs….....37 2.8 Discussion and Summary...................................................................................................37 References................................................................................................................................39. VI.

(8) Figure Captions Figure 1.1 PL peak energies of the LEDs grown on m- and c-plane GaN substrates. InGaN well thickness was varied between 2.5 nm (circle) and 8.0 nm (triangle)………………..….14 Figure 1.2 Polarization ratio as a function of emission wavelength for the m-plane LEDs. InGaN well thickness was 8.0 nm……………………………………………………………14 Figure 1.3 (a) Room temperature EL spectra at different injection currents. (b) The FWHM and peak wavelength of EL spectra vs. current density……………………………….……..14 Figure 1.4 (a) Room temperature EL spectra with different polarizer angle. The angle of 0° corresponds to a polarization parallel to a-axis. (b) Variation of EL intensity at peak wavelength with angular orientation of the polarizer at 20 mA operation current…………...15 Figure 1.5 (a) Optical micrographs for Si-doped n-type GaN films with a varying misorientation angle. Angle indicates the misorientation towards the [000 1 ] c -direction. (b) Dependence of RMS roughness on the misorientation angle for n-GaN films………………15 Figure 1.6 Nomarski optical micrographs [(a) and (b)], fluorescence optical micrographs [(c) and (d)], L-I curves [(e) and (f)], and emission spectra with I  10mA , I  0.9 I th , and. I  I th [(g) and (h)] from several 2 μm × 500 μm LDs for samples A and B, respectively...16 Figure 1.7 Relative output power vs. EL peak wavelength for a large number of LDs grown on m- and (2021) -plane GaN substrates……………………………………………………16 Figure 1.8 Peak wavelength as a function of TMI flow for SQW LEDs grown on (a). (202 1) -, (2021) -, (303 1) -, (30 3 1) -, and m-planes at a growth temperature of 780°C and on (b) (202 1) - and (1122) - planes at a growth temperature of 830°C……………………17 Figure 1.9 Simulated band diagrams and emission wavelengths for SQW (with 25% indium composition) blue and green LEDs grown on the (a) (2021) -, (b) (202 1) -, (c) (1122) -, and (d) m-planes at a current density of 20 mA/cm2…………………………………………17 Figure 1.10 Experimental flow chart of this chapter………………………………………...18 VII.

(9) Figure 1.11 Sample structures of (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples………………………………………………………………………………..19 Figure 1.12 SEM images of the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples and monochromatic CL images (c) and (d) taken over the same regions with 11kV excitation electron voltage at room temperature, respectively………………………...20 Figure 1.13 CL spectra of the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples with the excitations of 5, 7, 9, and 11kV electron voltages at room temperature…..20 Figure 1.14 CL peak position as a function of electron voltage for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples………………………………………………21 Figure 1.15 AFM (5 × 5 μm2) [(a) and (b)] images and 3D AFM [(c) and (d)] images taken from the same regions of the polar c-plane (Rq：1.118 nm) and nonpolar m-plane (Rq： 17.957 nm) InGaN/GaN MQW samples, respectively. Surface roughness of each sample, Rq, is shown in the parentheses ……………………………………………………………….....21 Figure 1.16 XRD patterns for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples………………………………………………………………………………..22 Figure 1.17 Polarization-dependent PL at 20K for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples………………………………………………………...22 Figure 1.18 (Left coordinate) PL spectra with polarization degrees set at E // c-axis and E // m-axis for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples at 20K. (Right coordinate) The degree of polarization for the two samples is also shown…………..23 Figure 1.19 PL spectra as a function of temperature for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples………………………………………………………...23 Figure 1.20 PL peak position as a function of temperature for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples………………………………………………………...24 Figure 1.21 Experimental setup of EL measurement………………………………………..24 Figure 1.22 EL spectra as a function of CW applied voltage for the (a) polar c-plane and (b) VIII.

(10) nonpolar m-plane InGaN/GaN MQW LEDs at room temperature…………………………..25 Figure 1.23 EL peak position as a function of CW applied voltage for the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs at room temperature…………………………..25 Figure 1.24 Polarization-dependent EL at room temperature for the (a) polar c-plane InGaN/GaN MQW LED at 3.0V and (b) nonpolar m-plane one at 9.5V…………………….26 Figure 1.25 (Left coordinate) EL spectra with polarization degrees set at E // c-axis and E // m-axis for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW LEDs at room temperature. (Right coordinate) The degree of polarization for the two samples is also shown………………………………………………………………………………………...26 Figure 1.26 (a) Current density (I), (b) output power, and (c) external quantum efficiency (EQE) (right coordinate) as functions of applied voltage (V) for the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs…………………………………………………27 Figure 1.27 Experimental setup of TREL measurement…………………………………….27 Figure 1.28 TREL profiles of the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN LEDs at room temperature with 2.5-6V, 500ns pulse width, and 1kHz repetition rate applied pulse voltages………………………………………………………………………………...28 Figure 1.29 Response ( response) , rise ( rise) , delay ( delay ) and recombination. ( recombination ) times as functions of applied pulse voltage for the polar c-plane and nonpolar m-plane InGaN/GaN LEDs………………………………………………………...28 Figure 2.1 Schematic illustration of the temporal sequences of valve operation………........40 Figure 2.2 SEM images of N-polar MQW samples grown with (a) continuous-mode, (b). t2  5 s , and (c) t2  2 s , while keeping t1 at 3 s. (d) PL spectra of those N-polar samples at the room temperature………………………………………………………………………40 Figure 2.3 SEM images of the samples grown with QW numbers of (a) 1, (b) 3, (c) 5, and (d) 10. (e) PL spectra of those N-polar samples at the room temperature……………………….40 IX.

(11) Figure 2.4 In situ optical reflectance of N-face GaN growth for samples with nitridition temperatures (a) 1130°C, (b) 1030°C, (c) 980°C, and (d) 950°C………………….……......41 Figure 2.5 Nomarski optical microscope images of samples with nitridition temperatures (a) 1130°C, (b) 1030°C, (c) 980°C, and (d) 950°C………………………………………….….41 Figure 2.6 Experimental flow chart of this chapter………………………………………….42 Figure 2.7 Sample structures of the N-Polar (a) 131210A(SQW), (b) 131210B(MQW), (c) 140206A (t11@ t2 5) , and (d) 140206B (t11@ t2 2) samples………………………………...43 Figure 2.8 SEM [(a) and (b)] and panchromatic CL [(c) and (d)] images for the corresponding SEM regions using the 11kV excitation electron voltage for the N-polar InGaN/GaN SQW and MQW samples, respectively………………………………………...44 Figure 2.9 SEM [(a) and (b)], panchromatic CL [(c) and (d)], and monochromatic CL [(e) and (f)] images for the corresponding SEM regions using the 11kV excitation electron voltage. for. the. N-polar. t11@ t2 5. and. t11@ t2 2. InGaN/GaN. MQW samples,. respectively…………………………………………………………………………………...44 Figure 2.10 CL spectra of the N-polar InGaN/GaN (a) SQW and (b) MQW samples with the excitations of 5, 7, 9, and 11kV electron voltages at room temperature……………………..45 Figure 2.11 CL spectra of the N-polar (a) t11@ t2 5 and (b) t11@ t2 2 InGaN/GaN MQW samples with the excitations of 5, 7, 9, and 11kV electron voltages at room temperature…..45 Figure 2.12 XRD patterns for the N-polar (a) InGaN/GaN SQW and MQW and (b) t11@ t2 5 and t11@ t2 2 InGaN/GaN MQW samples…………………………………………………46 Figure 2.13 AFM images (5 × 5 μm2) of the N-polar InGaN/GaN (a) SQW (Rq：7.327 nm) and (b) MQW (Rq：24.985 nm) samples and the N-polar (c) t11@ t2 5 (Rq：27.422 nm) and (d) t11@ t 2 2 (Rq：28.958 nm) InGaN/GaN MQW samples. Surface roughness of each sample, Rq, is shown in the parentheses……………………………………………………..46 Figure 2.14 PL spectra as a function of temperature for the N-polar InGaN/GaN (a) SQW and (b) MQW samples……………………………………………………………………………47 X.

(12) Figure 2.15 (a) PL peak position and (b) normalized PL integral intensity in the 370-500 nm spectral range as a function of temperature for the N-polar InGaN/GaN SQW and MQW samples…………………………………………………………………….…………………47 Figure 2.16 PL spectra as a function of temperature for the N-polar (a) t11@ t2 5 and (b). t11@ t2 2 InGaN/GaN MQW samples………………………………………………………48 Figure 2.17 (a) PL peak position and (b) normalized PL integral intensity in the 370-430 nm and 430-500 nm spectral ranges as a function of temperature for the N-polar t11@ t2 5 and. t11@ t2 2 InGaN/GaN MQW samples………………………………………………...…….48 Figure 2.18 SEM [(a) and (b)], panchromatic CL [(c) and (d)], and monochromatic CL [(e) and (f)] images for the corresponding SEM regions using the 11kV excitation electron voltage for the annealed N-polar t11@ t2 5 and t11@ t2 2 InGaN/GaN MQW samples, respectively…………………………………………………………………………………...49 Figure 2.19 CL spectra of the annealed N-polar (a) t11@ t2 5 and (b) t11@ t2 2 InGaN/GaN MQW samples with the excitations of 5, 7, 9, and 11kV electron voltages at room temperature.…………………………………………………………………….……...49 Figure 2.20 XRD patterns for the as-grown (black line) and annealed (red line) N-polar (a). t11@ t2 5 and (b) t11@ t2 2 InGaN/GaN MQW samples ……………………......………..50 Figure 2.21 AFM images (5 × 5 μm2) of the as-grown [(a) (Rq：27.422 nm) and (b) (Rq： 28.958 nm)] and annealed [(c) (Rq：26.694 nm) and (d) (Rq：27.295 nm)] for the N-polar. t11@ t2 5 and t11@ t 2 2 InGaN/GaN MQW samples, respectively. Surface roughness of each sample, Rq, is shown in the parentheses………………………………………………..50 Figure 2.22 PL spectra as a function of temperature for the annealed N-polar (a) t11@ t2 5 and (b) t11@ t2 2 InGaN/GaN MQW samples……………………………………………..51 Figure 2.23 (a) PL peak position and (b) normalized PL integral intensity in the 370-430 nm and 430-500 nm spectral ranges as a function of temperature for the annealed N-polar XI.

(13) t11@ t2 5 and t11@ t2 2 InGaN/GaN MQW samples………………………………..…….51. XII.

(14) Chapter. 1. Time-resolved. Electroluminescence. Studies. of. Nonpolar m-plane InGaN/GaN Multiple Quantum Wells. 1.1 Introduction III-nitride semiconductors have become the key technological material for light-emitting diodes and laser diodes. These devices are generally fabricated along the c-axis, where piezoelectric and spontaneous polarizations lead to quantum-confined Stark Effect (QCSE) [1,2]. Nonpolar devices grown along the m- and a-axes have been demonstrated to be polarization-free, so can be used to reduce QCSE [3]. In our previous study, we report the roles of island coalescence rate and strain relaxation in the development of anisotropic in-plane strains and subsequent degree of polarization in m-plane GaN [4]. Figure 1.1 shows PL peak energies of the LEDs grown on m- and c-plane substrates, together with the data of InGaN single quantum well (SQW) with a well thickness of 100 nm. The following fitting data was obtained for the 100 nm SQW and 8.0 nm MQWs. With In composition in InGaN increasing, the peak position is red-shifted. It is clearly seen that the PL peak energy of InGaN well thickness of 2.5 nm is higher than that of 8.0 nm. The energy difference is estimated to be 0.2 - 0.3 eV, which is reasonable to explain the quantum confinement effect. This trend was also confirmed in m-plane [5] and a-plane InGaN/GaN MQWs [6,7]. Figure 1.2 shows the optical polarization ratios of the LED grown on m-plane substrate as a function of emission wavelength.. Degree of polarization is defined as.   ( I a  I c ) /( I a  I c ) , where I a and I c denote electroluminescence (EL) intensity at 1 mA with polarization parallel to the a- and c-axis, respectively. The polarization ratio increases from 0.27 to 0.89 with increasing emission wavelength from 383 to 476 nm, corresponding to the indium composition from 0.02 to 0.08. The increase in the polarization 1.

(15) ratio is most probably due to energy separation, which is caused by compressively strained InGaN QWs [7]. Figure 1.3 (a) shows the electroluminescence (EL) spectra of the m-plane LED for the DC current density ranging from 1.1 to 330 A/cm2, and the emission peak and full width at half-maxima (FWHM) of EL are shown in Figure 1.3 (b). The emission peak remains constant before suffering from excess heat at high current operation, indicating the absence of the polarization-induced electric fields in the m-plane MQWs. The initial blue-shift in the emission peak for driving current from 1.1 to 11 A/cm2 can be attributed to the band-filling of the localized states induced by alloy fluctuation in the InGaN QWs [8, 9]. Figure 1.4 (a) shows the EL polarization anisotropy [10]. The peak energy shift E and the degree of polarization of the EL intensity were analyzed by rotating a polarizer between the polarization angles 0 and 360°. The E is about 39.7 meV between the electric field perpendicular ( E ) to and parallel ( E// ) to the c-axis. The peak energy shift comes from the splitting of the valence bands caused by the in-plane compressive strain of the MQW, and the energy separation is proportional to the In composition of the active layer. Figure 1.4 (b) shows the angular dependence of the polarization ratio under the operation current of 20 mA. The degree of polarization is defined as   ( I   I // ) /( I   I // ) , where I  and I // are the intensities of the E and E // components, respectively. The degree of polarization in nonpolar LEDs gradually increases with a larger valance band splitting due to the higher In composition as well as higher compressive strain in the active region InGaN [11]. The large value of the degree of polarization in m-plane LEDs could be contributed from the higher quality of InGaN MQWs because the proper growth condition can prevent the InGaN from phase separation on the m-plane GaN surface and can result in higher compressive strain [9]. Figure 1.5 (a) shows optical micrographs of the Si-doped n-type GaN films. Angle 2.

(16) indicates the misorientation towards the [000 1 ] c -direction. For the on-axis free-standing (10 1 0) m-plane GaN substrates, MOCVD growth results in surfaces covered with shallow. four-sided pyramidal hillocks. The pyramidal hillocks could be eliminated by growing GaN films on misoriented m-plane GaN substrates instead of on-axis m-plane GaN substrates [12]. Figure 1.5 (b) shows the dependence of root-mean-square (RMS) roughness on misorientation angle for Si-doped n-type GaN films [13,14]. Figures 1.6 show Nomarski [(a) and (c)] and fluorescence [(b) and (d)] optical micrographs from samples A (nominally on-axis m-plane GaN substrate) and B (m-plane GaN substrate misoriented 1◦ towards the [000 1 ] c -direction), respectively. Figures 1.6 (e) and (f) show light versus current (L–I) curves from several 2 μm × 500 μm LDs for samples A and B, respectively. The threshold currents for the sample A (400-1000 mA) were higher and showed more variation than the threshold currents for the sample B. Figures 1.6 (g) and (h) show emission spectra for 2 μm × 500 μm LDs for samples A and B with drive currents of 10mA, 0.9 times the threshold current ( I  0.9 I th ) , and just above the threshold current. ( I  I th ) , respectively [14,15]. Figure 1.7 shows the dependence of relative output power on EL peak wavelength for a large number of LDs grown on m-plane and (2021) GaN substrates. The m-plane LDs show a drastic decrease in output power between 480 and 520 nm, while the (2021) LDs maintain relatively high output powers beyond 530 nm [14,16]. Figure 1.8 shows peak wavelength as a function of TMI flow for SQW LEDs grown on (a) (202 1) -, (2021) -, (303 1) -, (30 3 1) -, and m-planes at a growth temperature of 780°C and on (b) (202 1) - and (1122) -planes at a growth temperature of 830°C. The relevant nonpolar and semipolar planes in the wurtzite crystal structure are shown in the inset. The different TMI flow can affect EL wavelength position. The EL peak position is red-shifted under higher TMI flow [17]. 3.

(17) Figure 1.9 shows simulated band diagrams and emission wavelengths for SQW (with 25% indium composition) blue and green LEDs grown on the (a) (2021) -, (b) (202 1) -, (c). (1122) -, and (d) m-planes at a current density of 20 mA/cm2. These differences are explained by the magnitudes and directions of polarization-related electric fields in the MQWs. For the (2021) - and (1122) -planes, the polarization-related electric field in the MQWs is in the. same direction as the p-n junction built-in electric field. For the (202 1) -plane, the polarization related electric field opposes the p-n junction built-in electric field. For the m-plane, the polarization-related electric field is zero, and only the p-n junction built-in electric field exists in the QWs [17].. 1.2 Motivation and Investigation Flow Chart Figure 1.10 shows the experimental flow chart of this chapter. By scanning electron microscope (SEM) and cathodoluminescence (CL) measurements, the microstructures and nano-photonics of samples are measured. By atomic force microscopy (AFM) measurement, the. surface. morphologies. of. samples. are. measured.. By. polarization-. and. temperature-dependent photoluminescence (PL) measurement, the optical properties of epilayers are investigated. By polarization-dependent and CW operation electroluminescence (EL) measurement, the electrical properties of LED devices are investigated. By current-voltage (I-V) and external quantum efficiency (EQE) measurements, the basic electrical properties of LEDs are studied. By time-resolved electroluminescence (TREL) measurement, response time and decay time for LED devices are determined. In this chapter, we investigate the anisotropic properties and carrier transport behaviors of the polar c-plane and nonpolar m-plane InGaN/GaN MQW epilayers and LED devices. According to the experimental results, we will discuss the anisotropic characteristics and carrier transport behaviors of the polar c-plane and nonpolar m-plane InGaN/GaN MQW 4.

(18) samples and LEDs.. 1.3 Sample Structures and Growth Conditions Figure 1.11 shows sample structures of (a) the polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples grown on the c- and m-plane sapphire substrates by metal organic chemical vapor deposition (MOCVD), respectively. The m-plane InGaN/GaN LED comprise a m-sapphire substrate, a 0.25 µm GaN buffer layer, a 3 µm undoped GaN layer, a 1.5 µm n-GaN contact layer, a 10-period InGaN/GaN multi quantum wells (MQWs), a 1.5 µm electron blocking layer, and a 0.25 µm p-GaN contact layer from bottom to top. The thicknesses of the InGaN well and GaN barrier were 3 and 10 nm in the MQWs, respectively. Except the substrate, the epilayer structures of the c-plane LED were the same as those of the m-plane one.. 1.4 Anisotropic Characteristics of the Polar c-plane and Nonpolar m-plane InGaN/GaN MQW Samples 1.4.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Studies In order to investigate the relation between optical property and microstructures, SEM and CL measurements were conducted at the same regions. SEM and CL images were acquired with a Gatan monoCL3 spectrometer in a JEOL SEM system (model JSM 7000F) under room temperature. The excitation electron voltages for CL measurement range from 5 to 11 kV. Figure 1.12 (a) and (b) show the SEM images for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples, respectively. The striation feature of the nonpolar m-plane InGaN/GaN MQW sample is easily observed. Figure 1.12 (c) and (d) show the panchromatic CL images for the corresponding SEM regions using the 11kV excitation 5.

(19) electron voltage, respectively. Both samples were emitted in all regions. Figure 1.13 shows the CL spectra of the two samples using excitation voltages of 5, 7, 9, and 11 kV at RT. It can be seen that the CL intensity of the polar c-plane InGaN/GaN MQW sample is stronger than that of the nonpolar m-plane one. In addition, Figure 1.14 shows that the CL peak position of the nonpolar m-plane InGaN/GaN MQW sample is shorter than the polar c-plane one. As the excitation voltage increases, the CL peak position is blue-shifted in the CL spectra.. 1.4.2 Atomic Force Microscopy (AFM) and X-ray Diffraction (XRD) Patterns Studies In order to study the surface morphology of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples, AFM measurement was conducted. AFM images were acquired with XE-70 and XE control electronics under non-contact mode. Figures 1.15 (a) and (b) show the AFM images of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples, while Figures 1.15 (c) and (d) show 3D AFM images taken from the same regions, respectively. The growth direction of GaN affects the surface roughness. The surface roughness in 5 μm × 5 μm area are 1.118 and 17.957 nm for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples, respectively. In general, the polar c-plane sample with high CL intensity usually has lower roughness. Stronger CL intensities of the polar c-plane InGaN/GaN MQW sample were observed in Figure 1.13. Figure 1.16 shows the XRD patterns of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples. Due to no signal, XRD pattern of the nonpolar m-plane InGaN/GaN sample can’t be measured.. 1.4.3 Polarization- and Temperature-dependent Photoluminescence (PL) 6.

(20) Studies PL is the most fundamental measurement for understanding the optical property of a material. PL measurement is carried out with the 325 nm line of a 55 mW He-Cd laser for excitation. The samples are placed in a cryostat for temperature-dependent PL measurement.. 1.4.3.1 Low-temperature Polarized PL and Degree of Polarization (DoP) Figure 1.17 (a) and (b) show the polarized PL for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples at 20K, respectively. The PL spectra show a high polarization anisotropy. The degree of polarization (DoP), ρ, can be expressed as. . I  c  I // c I  c  I // c. where I  c and I // c are the PL intensities for E  c and E // c , respectively. Figure 1.18 (a) and (b) show the DoP for the two samples. The DoP of the nonpolar m-plane InGaN/GaN MQW sample is larger than that of the polar c-plane one.. 1.4.3.2 Temperature-dependent PL Study Figure 1.19 (a) and (b) show the PL spectra of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples as a function of temperature, respectively. The emission peak of the polar c-plane InGaN/GaN MQW sample is around ~457 nm (2.713 eV), and that of the nonpolar m-plane one is around ~393 nm (3.155 eV). The emission peak position is related to indium composition in MQWs [7]. The PL intensities decay with increasing temperature. The PL peak positions as a function of temperature for the two samples are shown in Figure 1.20. The PL spectra of the polar c-plane InGaN/GaN MQW sample range from 454 nm to 459 nm, while those of the nonpolar m-plane one do from 383 nm to 402 nm. The peak position of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples are 458.6 and 383.8 nm at 20K, respectively. With increasing temperature form 20 to 300 K, PL position of 7.

(21) the polar c-plane InGaN/GaN MQW sample shows s-shape variation, and that of the nonpolar m-plane one is red-shifted.. 1.5 Device Characteristics and Carrier Transport Properties of the Polar c-plane and Nonpolar m-plane InGaN/GaN MQW LEDs 1.5.1 Voltage- and Polarization-dependent Electroluminescence (EL) Studies 1.5.1.1 Voltage-dependent EL Figure. 1.21. shows. the. experimental. setup. and. schematic. diagram. of. Electroluminescence (EL) under CW operation. A source meter (Keithley 2614B) was used to apply voltage to the LED devices. The luminescence from LED sample was focused into a spectrometer with a USB interface (Ocean Optics USB 2000+). The EL spectra were appeared in computer. Figures 1.22 (a) and (b) show the EL spectra of the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs at room temperature with 2.5-3.6 V and 2.5-9.5 V applied voltages, respectively. With higher applied voltages, the two LEDs show stronger EL intensities. The EL spectra of the polar c-plane InGaN/GaN MQW LED only have one peak, while those of the nonpolar m-plane InGaN/GaN MQW LED have one peak and small shoulders. Because the FWHM can be defined as the quality of the LED, the quality of the polar c-plane InGaN/GaN MQW LED is better than that of the nonpolar m-plane one. Figure 1.23 shows the peak position of EL spectra as a function of applied voltage for the two LEDs. With high applied voltages, the EL peak positions of the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs are slightly blue-shifted.. 1.5.1.2 Polarized EL and Degree of Polarization (DoP) Figure 1.24 (a) and (b) show the polarized EL for the polar c-plane and nonpolar 8.

(22) m-plane InGaN/GaN MQW LEDs at room temperature, respectively. The EL spectra show a low polarization anisotropy. The degree of polarization (DoP), ρ, can be expressed as. . I  c  I // c I  c  I // c. where I  c and I // c are the EL intensities for E  c and E // c , respectively. Figures 1.25 (a) and (b) show the DoP for the two LEDs. The DoP of the polar c-plane InGaN/GaN MQW LED is larger than that of the nonpolar m-plane one.. 1.5.2 Current-Voltage (I-V) and External Quantum Efficiency (EQE) Studies Figure 1.26 (a) shows current density (I) as functions of applied voltage (V) for the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs. The current density of the polar c-plane InGaN/GaN MQW LED is larger than that of the nonpolar m-plane one. Figure 1.26 (b) and (c) show the output power and external quantum efficiency (EQE ) of the two LEDs as functions of applied voltage. The output power (µW) of the two LEDs can be measured by a power meter (Newport Model 835). The external quantum efficiency. (EQE ) is defined by output power divided by the product of I*V. The  EQE of the polar c-plane InGaN/GaN MQW LED is much higher than that of the nonpolar m-plane one.. 1.5.3 Time-resolved Electroluminescence (TREL) Study In order to understand the carrier transport and recombination dynamics of the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs, TREL measurement was conducted. Figure 1.27 shows experimental setup and schematic diagram of TREL. The probe station can touch the electrodes by the probe. A pulse generator (Tektronix, AFC3011C) is used to apply 2.5-6V, 500ns pulse width, and 1kHz repetition rate applied pulse voltages to the devices. The 9.

(23) emitting light of LEDs is focused by lens into the detector. Time-dependent EL signals are detected by the 50 Ω input resistance of a digital oscilloscope (Agilent, model DSO 6052A, 500 MHz/4Gs/s) together with a photomultiplier (Photo sensor Modules H10721 Series) located on top of the emitting areas. Figure 1.28 (a) and (b) show the TREL transit profiles of the two LEDs. With a larger applied pulse voltage, the rise profile exhibits a stronger intensity and the response time is shorter. The intensities of the polar c-plane InGaN/GaN MQW LED are stronger than that of the nonpolar m-plane one. The rising part of the nonpolar m-plane InGaN/GaN MQW LED rises more steeply and the decay part decays faster than that of the polar c-plane one. As shown in Figure 1.29, the response time ( response) can be determined by the time delay between addressing the device with a short voltage pulse and the first appearance of EL. Because a larger forward bias enhances the ability of the hole and electron leading fronts to meet faster more easily, a shorter response time is observed. The shorter response times of the nonpolar m-plane InGaN/GaN MQW LED imply a better carrier injection efficiency and a faster carrier recombination. The rising time ( rise ) defined by intercept of the tangents, as function of applied pulse voltage for two types of LED is shown in Figure 1.29. The  rise of the nonpolar m-plane InGaN/GaN MQW LED is shorter than that of the polar c-plane one for each applied voltage. This could also be determined by carrier localization, the strength of QCSE, and the height of potential barrier. When the voltage pulse is switched off (indicated with the vertical dotted line in Figure 1.28), TREL transit profiles of the polar c-plane InGaN/GaN LED show longer delay time. ( delay ) in reaching both the maximum intensity of transit EL and the subsequent recombination. Due to the combined effects of the weaker carrier localization, stronger QCSE, 10.

(24) and the potential barrier, the poorer relaxation efficiency of the polar c-plane InGaN/GaN LED needs more time to reach the maximum intensity of transit EL. The recombination time can be determined by fitting the TREL decay profile with a single exponential. As shown in Figure 1.29, the  recombination of the nonpolar m-plane InGaN/GaN MQW LED is shorter than that of the polar c-plane one for each applied pulse voltage. The shorter  recombination of m-plane InGaN/GaN MQW LED implies a fast carrier recombination. With a larger applied pulse voltage, the  recombination of the nonpolar m-plane InGaN/GaN MQW LED slightly increases, while that of the polar c-plane one increases more. This also be due to the larger QCSE and potential distribution upon a larger forward bias for the polar c-plane InGaN/GaN MQW LED.. 1.6 Discussion and Summary In summary, we have shown the experimental results of SEM, CL, AFM, PL, EL, I-V, and TREL measurements of the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples and LEDs. The larger surface roughness and weaker CL intensity of the nonpolar m-plane InGaN/GaN MQW sample than those of the polar c-plane one show a higher defect density and lower sample quality of the nonpolar m-plane one. The result of CL measurement is consistent of that of the AFM measurement. In addition, the DoP of PL of the nonpolar m-plane InGaN/GaN MQW sample is larger than the polar c-plane one. With increasing temperature from 20 to 300 K, PL position of the polar c-plane InGaN/GaN MQW sample is slightly blue-shifted, while that of the nonpolar m-plane one is red-shifted. The PL FWHM of the polar c-plane InGaN/GaN MQW LED is smaller than that of the nonpolar m-plane one. The shorter response time of the nonpolar m-plane InGaN/GaN MQW LED than that of the polar c-plane one suggest a better injection efficiency. The longer recombination time of the polar c-plane InGaN/GaN MQW LED than that of the nonpolar m-plane one could be due to the larger QCSE and potential distribution in the polar c-plane InGaN/GaN MQW LED. 11.

(25) References 1.. F. Bernardini, V. Fiorentini, and D. Vanderbilt, Phys. Rev. B 56, R10024 (1997).. 2.. P. Waltereit, O. Brandt, A. Trampert, H. T. Grahn, J. Menniger, M. Ramsteiner, M. Reiche, and K. H. Ploog, Nature 406, 865 (2000).. 3.. T. Paskova, Nitrides with Nonpolar Surfaces: Growth, Properties, and Devices (Weinheim: Wiley-Vch) (2008).. 4.. S. W. Feng, C. K. Yang, C. M. Lai, L. W. Tu, Q. Sun, and J. Han, J. Phys. D: Appl. Phys. 44, 375103 (2011).. 5.. K. Kim, M. Schmidt, H. Sato, F. Wu, N. Fellows, Z. Jia, M. Saito, S. Nakamura, S. DenBaars, and J. Speck, Appl. Phys. Lett. 91, 181120 (2007).. 6.. H. Yamada, K. Iso, M. Saito, H. Hirasawa, N. Fellows, H. Masui, K. Fujito, J. S. Speck, S. P. DenBaars, and S. Nakamura, Phys. Status Solidi: Rapid Res. Lett. 2, 89 (2008).. 7.. H. Yamada, K. Iso, M. Saito, H. Masui, K. Fujito, S. P. DenBaars, and S. Nakamura, Appl. Phys. Express 1, 041101 (2008).. 8.. T. Koyama, T. Onuma, H. Masui, A. Chakraborty, B. A. Haskell, S. Keller, U. K. Mishra, H. S. Speck, S. Nakamura, and S. P. Denbaars, Appl. Phys. Lett. 89, 091906 (2006).. 9.. S. P. Chang, T. C. Lu, L. F. Zhuo, C. Y. Jang, D. W. Lin, H. C. Yang, H. C. Kuo, and S. C. Wang, J. Electrochem. Soc. 157(5), H501-H503 (2010).. 10. N. F. Gardner, J. C. Kim, J. J. Wierer, Y. C. Shen, and M. R. Krames, Appl. Phys. Lett. 86, 111101 (2005). 11. J. Bhattacharyya, S. Ghosh, and H. T. Grahn, Appl. Phys. Lett. 93, 051913 (2008). 12. A. Hirai, Z. Jia, M. C. Schmidt, R. M. Farrell, S. P. DenBaars, S. Nakamura, J. S. Speck, and K. Fujito, Appl. Phys. Lett. 91, 191906 (2007). 13. R. M. Farrell, J. Cryst. Growth 313, 1 (2010). 14. R. M. Farrell, E. C. Young, F. Wu, S. P. DenBaars, and J. S. Speck, Semicond. Sci. Technol. 27, 024001 (14pp) (2012). 12.

(26) 15. Y. D. Lin, Appl. Phys. Express 2, 082102 (2009). 16. Y. D. Lin, PhD Thesis University of California, Santa Barbara (2010). 17. Y. Zhao, Q. Yan, C. Y. Huang, S. C. Huang, P. S. Hsu, S. Tanaka, C. C. Pan, Y. Kawaguchi, K. Fujito, C. G. Van de Walle, J. S. Speck, S. P. DenBaars, S. Nakamura, and D. Feezell, Appl. Phys. Lett. 100, 201108 (2012).. 13.

(27) Figure 1.1 PL peak energies of the LEDs grown on m- and c-plane GaN substrates. InGaN well thickness was varied between 2.5 nm (circle) and 8.0 nm (triangle) [7].. Figure 1.2 Polarization ratio as a function of emission wavelength for the m-plane LEDs. InGaN well thickness was 8.0 nm [7].. Figure 1.3 (a) Room temperature EL spectra at different injection currents. (b) The FWHM and peak wavelength of EL spectra vs. current density [9].. 14.

(28) Figure 1.4 (a) Room temperature EL spectra with different polarizer angle. The angle of 0° corresponds to a polarization parallel to a-axis. (b) Variation of EL intensity at peak wavelength with angular orientation of the polarizer at 20 mA operation current [9].. (a). (b). Figure 1.5 (a) Optical micrographs for Si-doped n-type GaN films with a varying misorientation angle. Angle indicates the misorientation towards the [000 1 ] c -direction. (b) Dependence of RMS roughness on the misorientation angle for n-GaN films [14].. 15.

(29) Figure 1.6 Nomarski optical micrographs [(a) and (b)], fluorescence optical micrographs [(c) and (d)], L-I curves [(e) and (f)], and emission spectra with I  10mA , I  0.9 I th , and I  I th [(g) and (h)] from several 2 μm × 500 μm LDs for samples A and B, respectively [14].. Figure 1.7 Relative output power vs. EL peak wavelength for a large number of LDs grown on m- and (2021) -plane GaN substrates [14].. 16.

(30) Figure 1.8 Peak wavelength as a function of TMI flow for SQW LEDs grown on (a) (202 1) -, (2021) -, (303 1) -, (30 3 1) -, and m-planes at a growth temperature of 780°C and on (b). (202 1) - and (1122) - planes at a growth temperature of 830°C [17].. Figure 1.9 Simulated band diagrams and emission wavelengths for SQW (with 25% indium composition) blue and green LEDs grown on the (a) (2021) -, (b) (202 1) -, (c) (1122) -, and (d) m-planes at a current density of 20 mA/cm2 [17].. 17.

(31) Polar c-plane and nonpolar m-plane InGaN/GaN MQWs grown on sapphire and m-sapphire, respectively, by MOCVD. Sample preparation. Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Measurements. To investigate microstructures and nanophotonics of samples. X-ray Diffraction (XRD) and Atomic Force Microscopy (AFM) Measurements. To measure the structure characteristic and surface morphology of samples. Polarization- and Temperature-dependent Photoluminescence (PL) Measurements. To investigate optical properties of samples. Electroluminescence (EL) Measurement. To investigate spectra of LEDs. Current-Voltage (I-V Curve) and external quantum efficiency (EQE) Measurements. To study basic electrical properties of LEDs. Time-resolved Electroluminescence (TREL) Measurement. To investigate the carrier transport behaviors of LEDs. Figure 1.10 Experimental flow chart of this chapter.. 18.

(32) (a). (b). Figure 1.11 Sample structures of (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples.. 19.

(33) Figure 1.12 SEM images of the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples and panchromatic CL images (c) and (d) taken over the same regions with 11kV excitation electron voltage at room temperature, respectively. 240000. (a) polar c 5kV 7kV 9kV 11kV. 180000. Intensity (arb. unit). 120000 60000 0 40000. (b) nonpolar m. 30000 20000 10000 0 350. 400. 450. 500. 550. Wavelength (nm). Figure 1.13 CL spectra of the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples with the excitations of 5, 7, 9, and 11kV electron voltages at room temperature. 20.

(34) 460 polar c nonpolar m. CL Peak Position(nm). 450 440 430 420 410 400 390 380 370. 5. 6. 7. 8. 9. 10. 11. Electron voltage (kV). Figure 1.14 CL peak position as a function of electron voltage for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples.. Figure 1.15 AFM (5 × 5 μm2) [(a) and (b)] images and 3D AFM [(c) and (d)] images taken from the same regions of the polar c-plane (Rq：1.118 nm) and nonpolar m-plane (Rq：17.957 nm) InGaN/GaN MQW samples, respectively. Surface roughness of each sample, Rq, is shown in the parentheses. 21.

(35) 20000. Intensity (arb. unit). polar c nonpolar m 15000. 10000. 5000. 0 31. 32. 33. 34. 35. 36. 2degrees. Figure 1.16 XRD patterns for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples.. 0.30 (a) polar c. 0deg 15deg 30deg 45deg 60deg 75deg 90deg 105deg 120deg 135deg 150deg 165deg. 0.25 0.20. Intensity (arb. unit). 0.15 0.10 0.05 0.00. 180deg 195deg 210deg 225deg 240deg 255deg 270deg 285deg 300deg 315deg 330deg 345deg. 0.30 (b) nonpolar m 0.25 0.20 0.15 0.10 0.05 0.00 350. 400. 450. 500. 550. Wavelength (nm). Figure 1.17 Polarization-dependent PL at 20K for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples.. 22.

(36) 1.0 0.30 (a) polar c 0.25 0.20 0.5. 0.10 0.05 0.0 1.0. 0.00 0.30 (b) nonpolar m. E//c-axis E//m-axis DOP. 0.25. Degree of Polarization. Intensity (arb. unit). 0.15. 0.20 0.5. 0.15 0.10 0.05 0.00 350. 400. 450. 500. 0.0 550. Wavelength (nm). Figure 1.18 (Left coordinate) PL spectra with polarization degrees set at E // c-axis and E // m-axis for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples at 20K. (Right coordinate) The degree of polarization for the two samples is also shown. 0.30 (a) polar c 20K 40K 60K 80K 100K 120K 140K 160K. 0.25 0.20 0.15. Intensity (arb. unit). 0.10 0.05 0.00 0.30 (b) nonpolar m. 180K 200K 220K 240K 260K 280K 300K. 0.25 0.20 0.15 0.10 0.05 0.00 350. 400. 450. 500. 550. Wavelength (nm). Figure 1.19 PL spectra as a function of temperature for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW samples. 23.

(37) 470. PL Peak Position(nm). 460 450. polar c nonpolar m. 440 430 420 410 400 390 380. 0. 50. 100. 150. 200. 250. 300. Temperature (K). Figure 1.20 PL peak position as a function of temperature for the polar c-plane and nonpolar m-plane InGaN/GaN MQW samples.. Figure 1.21 Experimental setup of EL measurement.. 24.

(38) 70000. (a) polar c. 2.5V 3.0V 3.5V 4.0V 4.5V 5.0V 5.5V 6.0V. 60000 50000 40000 30000. Intensity (arb. unit). 20000 10000 0 20000 (b) nonpolar m. 2.5V 3.0V 3.5V 4.0V 4.5V 5.0V 5.5V 6.0V. 15000 10000 5000 0 350. 400. 450. 500. 6.5V 7.0V 7.5V 8.0V 8.5V 9.0V 9.5V. 550. Wavelength (nm). Figure 1.22 EL spectra as a function of CW applied voltage for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW LEDs at room temperature.. EL Peak Position(nm). 460 450 polar c nonpolar m Linear Fit of polar c Linear Fit of nonpolar m. 440 430 420. Equation y = a + b*x Adj. R-Square 0.86907 c-plane c-plane m-plane m-plane. Intercept Slope Intercept Slope. 0.83839 Value Standard Error 458.66381 0.87927 -1.37619 0.19976 419.65897 0.86371 -0.80703 0.11686. 410 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5. Applied Voltage (Volt). Figure 1.23 EL peak position as a function of CW applied voltage for the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs at room temperature.. 25.

(39) 60000 (a) polar c. 0deg 15deg 30deg 45deg 60deg 75deg 90deg 105deg 120deg 135deg 150deg 165deg. (b) nonpolar m. 180deg 195deg 210deg 225deg 240deg 255deg 270deg 285deg 300deg 315deg 330deg 345deg. 40000. Intensity (arb. unit). 20000. 0 15000. 10000. 5000. 0 350. 400. 450. 500. 550. Wavelength (nm). Figure 1.24 Polarization-dependent EL at room temperature for the (a) polar c-plane InGaN/GaN MQW LED at 3.0V and (b) nonpolar m-plane one at 9.5V. 1.0. 60000. (a) polar c. 40000. 0.5. Intensity (arb. unit). 0.0 1.0. 0 15000. (b) nonpolar m E//c-axis E//m-axis DOP. 10000. Degree of Polarization. 20000. 0.5 5000. 0. 350. 400. 450. 500. 0.0 550. Wavelength (nm). Figure 1.25 (Left coordinate) EL spectra with polarization degrees set at E // c-axis and E // m-axis for the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN MQW LEDs at room temperature. (Right coordinate) The degree of polarization for the two samples is also shown. 26.

(40) Output Power (mW) Current Density (A/cm2) EQE (%). 0.08. (a). polar c nonpolar m. (b). polar c nonpolar m. 0.06 0.04 0.02 0.00 2.5 2.0 1.5 1.0 0.5 0.0 1.0 0.8. (c). polar c nonpolar m. 0.6 0.4 0.2 0.0 2.5. 3.0. 3.5. 4.0. 4.5. 5.0. 5.5. 6.0. Applied Voltage (Volt). Figure 1.26 (a) Current density (I), (b) output power, and (c) external quantum efficiency (EQE) as functions of applied voltage (V) for the polar c-plane and nonpolar m-plane InGaN/GaN MQW LEDs.. Figure 1.27 Experimental setup of TREL measurement.. 27.

(41) 1.5. delay. (a) polar c. recombination. response. 1.0. rise. Intensity (arb. unit). 0.5. 0.0 0.6 (b) nonpolar m. 2.5V 3.0V 3.5V 4.0V 4.5V 5.0V 5.5V 6.0V. 0.4. 0.2. 0.0 0.0. 0.1. 0.2. 0.3. 0.4. 0.5. 0.6. 0.7. 0.8. Time (s). Figure 1.28 TREL profiles of the (a) polar c-plane and (b) nonpolar m-plane InGaN/GaN LEDs at room temperature with 2.5-6V, 500ns pulse width, and 1kHz repetition rate applied pulse voltages. 250 polar c. response. response, rise, delay, and recombination (ns). rise delay. 200. recombination. nonpolar m response. 150. rise delay recombination. 100. 50. 0. 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 Applied Pulse Voltage (volt). Figure 1.29 Response ( response) , rise ( rise) , delay ( delay ) and recombination. ( recombination ) times as functions of applied pulse voltage for the polar c-plane and nonpolar m-plane InGaN/GaN LEDs. 28.

(42) Chapter 2 Annealing effect on Nitrogen-Polar InGaN/GaN MQWs. grown. by. Pulsed. Metalorganic. Chemical. Vapor. Deposition. 2.1 Introduction It is reposted that luminescence properties of N-polar InGaN/GaN MQWs grown with short pulse durations are better than those of N-polar one grown with conventional continuous growth method [1]. Using pulsed-mode creates a high density of hexagonal mounds with an increased InGaN growth rate and enhanced In composition around screw-type dislocations, resulting in improved luminescence properties [1]. Figure 2.1 shows pulsed-mode for N-Polar InGaN/GaN MQWs growth. During the growth of InGaN/GaN MQW in N-Polar GaN, Ga and N sources are alternately injected into the reactor to alter the surface stoichiometry. A (black line) and NH3 (blue line) sources were switched ON/OFF alternately, while TMIn (red line) was continuously injected at a constant flow rate [1]. The durations of TEGa ON (NH3 OFF) and TEGa OFF (NH3 ON) are defined as t1 and t2 , respectively. In general, the durations of precursors switching “ON” and “OFF” are very important in pulsed growth mode. Many groups have conducted the pulsed growth mode for different purposes and have conducted investigations on the effect of pulse durations, such as the ability to influence the impurity incorporation rate in growing high quality GaAs [2], lateral/vertical growth rate ratio in epitaxial lateral overgrowth [3, 4], and the aspect ratio of GaN nanorods [5]. To get a sense about the effect of pulse durations on the properties of N-polar MQWs, the influence of duration t1 and t2 on the properties of N-polar MQWs have been studied. Figure 2.2 shows SEM images and PL spectra at room temperature of the N-polar samples 29.

(43) grown with different durations of t2 . The density of hexagonal mounds on the surface increases with decreasing t2 , showing a similar trend as the PL intensity. This indicates that the enhancement of the PL intensity may be caused by these mounds [1]. Figure 2.3 shows the SEM images and PL spectra at room temperature of the N-polar samples grown with different QW numbers. t1 and t2 for this set of experiments are kept at 3 and 5 s, respectively. The intensity of PL peak at 450 nm increases gradually by increasing the number of QWs. The size of mounds increases monotonically with increasing the number of QWs. The enhanced PL intensity cannot be simply due to the increase of density of mounds, but relates to the increase of total area of mounds [1]. A natural concept regarding c-plane sapphire nitridation is that the higher the temperature, the better the sapphire surface preparation. When the nitridation temperature is too high, sapphire surface maybe damaged. The sapphire was nitridized at 1130°C for 10min (sample A), 1030°C (sample B), 980°C (sample C), and 950°C (sample D) for 30s, respectively. With an aggressive nitridation for the c-plane sapphire, the in situ optical reflectance of the N-face GaN growth quickly damps in amplitude in Figure 2.4 (a), and the as-grown surface shows a high density of hexagonal hillocks in Figure 2.5 (a). As the nitridation temperature decreases, the damping of the reflectance oscillation is relieved in Figure 2.4 (b)-(d), and the density of hexagonal hillocks is greatly reduced in Figure 2.5 (b)-(d). The slight damping in the reflectance oscillation of sample D can be completely eliminated by using 2°off-cut sapphire substrates. With the usage of 2°offcut sapphire, N-face GaN with structural quality comparable to Ga-face GaN has been achieved [6]. The reflectance damping in Figure 2.4 (a) starts from the first oscillation, indicating that the onset of hexagonal hillock formation occurs at the very beginning of high temperature growth [7]. Therefore, the hexagonal hillock formation originates from the severe nitridation of sapphire. At a very high temperature, the strong chemical reaction between sapphire and NH3 can cause local damages for the sapphire surface [6,8]. 30.

(44) 2.2 Motivation and Investigation Flow Chart Figure 2.6 shows the experimental flow chart of this chapter. By scanning electron microscope (SEM) and cathodoluminescence (CL) measurement, the microstructures and nano-photonics of samples are measured. By X-ray diffraction (XRD) measurements, the indium composition of sample is determined. By atomic force microscopy (AFM) measurement, the surface morphologies of samples are measured. By temperature-dependent photoluminescence (PL) measurement, the optical properties of samples are investigated. According to the results of the experiments, we will discuss the material and optical characteristics of N-Polar InGaN/GaN MQW samples. Thermal annealing with 900°C for 60 seconds in Argon environment for the N-polar InGaN/GaN MQW samples was conducted.. 2.3 Sample Structures and Growth Conditions Prior to the growth of InGaN MQWs, N-polar GaN epilayers used as templates were grown on c-plane (000 1 ) sapphire with 2°off-cut toward the a-axis according to the procedure reported earlier [6,9]. Sapphire was heated up for nitridation in a mixture of ammonia (NH3) (3 slm) and N2 (4 slm) in a horizontal metalorganic chemical vapor deposition (MOCVD) reactor to 950°C for 30 s. A 20 nm GaN nucleation layer was then grown on nitridized sapphire at 600°C in H2 carrier gas, followed by growing 1 μm N-polar GaN at 1055°C and 100 mbar, with a NH3 flow rate of 0.5 slm and a trimethylgallium (TMGa) flow of 66 μmol/min. During the growth, TMGa and NH3 were used as Ga and N sources, respectively. The surface of the N-polar GaN template is very smooth and featureless under optical microscopy. The N-polarity of GaN templates was confirmed by convergent beam electron diffraction and wet etching in KOH solution [6,9]. After growth of N-polar GaN templates, the temperature was ramped down to 830°C to grow five pairs of InGaN MQWs under N2 carrier gas. Triethylgallium (TEGa), trimethylindium (TMIn) and NH3 were 31.

(45) used as Ga, In, and N sources, respectively. During the growth of MQWs, a pulsed-mode was employed to grow InGaN quantum wells (QWs), whereas continuous-mode was used to grow GaN quantum barriers (QBs) [1]. Figures 2.7 (a)-(d) show sample structures of the 131210A(SQW), 131210B(MQW), 140206A (t11@ t2 5) , and 140206B (t11@ t2 2) samples, respectively. t1 is the flow time of TEGa and t2 is the flow time of NH3. TMIn was continuously injected at a constant flow rate. The N-polar InGaN/GaN samples of 131210A and 131210B were grown with 1 and 3 QWs, respectively. t1 and t2 are kept at 3 and 5 s, respectively. In addition, the N-polar InGaN/GaN MQW samples of 140206A and 140206B were grown with t2 at 5 and 2 s while keeping t1 at 1 s, respectively. The number of QW for the two samples is 5.. 2.4 Material Characteristics of Nitrogen-Polar InGaN/GaN MQWs 2.4.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Studies In order to investigate the relation between optical property and microstructures, scanning electron microscope (SEM) and cathodoluminescence (CL) measurements were conducted at the same regions. Figure 2.8 shows the SEM [(a) and (b)] and panchromatic CL [(c) and (d)] images for the corresponding SEM regions using the 11kV excitation electron voltage for the N-polar InGaN/GaN SQW and MQW samples, respectively. The total area ratio of mounds of the MQW sample is larger than that of the SQW. CL in both samples was observed in InGaN mounds. Figure 2.9 shows the SEM [(a) and (b)], panchromatic CL [(c) and (d)], and monochromatic CL [(e) and (f)] images for the corresponding SEM regions using the 11kV excitation electron voltage for the N-polar t11@ t2 5 and t11@ t2 2 samples, respectively. 32.

(46) The total area ratio of mounds of the N-polar t11@ t2 2 sample is larger than that of the N-polar t11@ t2 5 one. Monochromatic CL mapping was collected at wavelengths of 450 nm and 460 nm for the N-polar t11@ t2 5 and t11@ t2 2 samples, respectively. The emitting InGaN mounds of the t11@ t2 2 sample are larger than those of the t11@ t2 5 one. Figure 2.10 shows the CL spectra of the two samples using excitation voltages of 5, 7, 9, and 11 kV at RT. Both samples have one CL peak around ~360 nm. Figure 2.11 shows the CL spectra of the N-polar t11@ t2 5 and t11@ t2 2 samples, in which there are two CL peaks: one in ~360 nm and the other one in ~450 nm and ~460 nm, respectively.. 2.4.2 X-ray Diffraction (XRD) Patterns Figure 2.12 shows the XRD patterns for the four samples. The diffraction peaks corresponding to InGaN and GaN can be identified. The GaN diffraction peak is mainly from the contact and barrier layers. The side shoulder with a broad distribution below the GaN main peak is attributed to InGaN with various indium contents, sizes, and shapes in the quantum wells [10]. The indium contents of the N-polar InGaN/GaN MQW sample are larger than those of the N-polar InGaN/GaN SQW one, and the indium contents of the N-polar. t11@ t2 2 sample are larger than that of the N-polar t11@ t2 5 one.. 2.4.3 Atomic Force Microscopy (AFM) Study In order to study the surface morphology of the N-polar samples, AFM measurement was conducted. Figures 2.13 (a) and (b) show the AFM images of the N-polar InGaN/GaN SQW and MQW samples, respectively. The InGaN mounds affect the surface roughness. The surface roughness in 5 μm × 5 μm area are 7.327 and 24.985 nm for the N-polar InGaN/GaN SQW and MQW samples, respectively. As the growth number of QW increases, a larger surface roughness is observed. 33.

(47) Figures 2.13 (c) and (d) show the AFM images of the N-polar t11@ t2 5 and t11@ t2 2 samples, respectively. The InGaN mounds affect the surface roughness. The surface roughness in 5 μm × 5 μm area are 27.422 and 28.958 nm for the N-polar t11@ t2 5 and. t11@ t2 2 samples, respectively. Although the growth time (t2 ) for the two samples is different, the surface roughness is almost similar.. 2.5 Temperature-dependent PL Studies of Nitrogen-Polar InGaN/GaN MQWs Figures 2.14 (a) and (b) show the PL spectra of the N-polar InGaN/GaN SQW and MQW samples as a function of temperature, respectively. The PL spectra of the two samples are categorized into two parts. One spectral range is from 350 nm to 370 nm, and the other one is from 370 nm to 500 nm. The PL intensity decays with increasing temperature. The PL peak positions as a function of temperature for the two samples are shown in Figure 2.15 (a). The N-polar InGaN/GaN SQW and MQW samples have three peaks: ~360 nm (3.444 eV), ~379 nm (3.272 eV), and ~390nm (3.179 eV). The dominate peak is around ~379 nm (3.272 eV). The emission peak positions of ~360 nm (3.444 eV) are related to GaN, while those of ~379 nm (3.272 eV) and ~390nm (3.179 eV) are related to InGaN. The peak positions of the SQW and MQW samples are nearly the same. With increasing temperature form 20 to 300 K, PL positions of the two samples are nearly temperature-independent. In addition, the normalized integral PL intensities as a function temperature of the two samples are shown in Figure 2.15 (b). The integral PL intensities of the N-polar SQW sample decay faster than those of the N-polar MQW one after 80K. Figure 2.16 (a) and (b) show the PL spectra of the N-polar t11@ t2 5 and t11@ t2 2 samples as a function of temperature, respectively. The PL spectra of the two samples are categorized into two parts. One spectral range is from 350 nm to 430 nm, and the other one is from 430 nm to 500 nm. The PL intensity decays with increasing temperature. 34.

(48) The PL peak positions as a function of temperature for the two samples are shown in Figure 2.17 (a). The N-polar t11@ t2 5 sample has three peaks: ~379 nm (3.272 eV), ~390nm (3.179 eV), and ~450 nm (2.756 eV). The N-polar t11@ t2 2 sample also has three peaks: ~379 nm (3.272 eV), ~390nm (3.179 eV), and ~470 nm (2.638 eV). The dominate peaks of the two samples are around ~379 nm (3.272 eV). The three emission peak positions of the two samples are related to InGaN with different Indium contents. The peak positions around ~379 nm (3.272 eV) and ~390 nm (3.179 eV) for the two samples are nearly the same. With increasing temperature form 20 to 300 K, PL peak positions of the two samples are nearly temperature-independent. In addition, the normalized integral PL intensities as a function temperature of the two samples are shown in Fig 2.17 (b). The integral PL intensity of the N-polar t11@ t2 5 sample decays faster than those of the N-polar t11@ t2 2 one.. 2.6 Material Characteristics of annealed Nitrogen-Polar InGaN/GaN MQWs Thermal annealing with 900°C for 60 seconds in Argon environment was conducted for the N-polar t11@ t2 5 and t11@ t2 2 InGaN/GaN MQW samples.. 2.6.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL) Studies Figure 2.18 shows the SEM [(a) and (b)], panchromatic CL [(c) and (d)], and monochromatic CL [(e) and (f)] images for the corresponding SEM regions using the 11kV excitation electron voltage for the annealed N-polar t11@ t2 5 and t11@ t2 2 samples, respectively. The total area ratio of mounds of the annealed N-polar t11@ t2 2 sample is larger than that of the annealed N-polar t11@ t2 5 one. Monochromatic CL mapping was collected at a wavelength of 450 nm and 460 nm for the annealed N-polar t11@ t2 5 and 35.

(49) t11@ t2 2 samples, respectively. The emitting InGaN mounds of the annealed N-polar t11@ t2 2 sample are larger than those of the annealed N-polar t11@ t2 5 one. The mound and light densities of the annealed N-polar t11@ t2 5 and t11@ t2 2 samples are larger than those of the as-grown ones Figure 2.19 shows the CL spectra of the two annealed samples using excitation voltages of 5, 7, 9, and 11 kV at RT. The CL spectra of the annealed N-polar t11@ t2 5 and t11@ t2 2 InGaN/GaN MQW samples show two CL peaks: one in ~360 nm and the other one in ~450 nm and ~460 nm, respectively. The CL peak positions of the annealed samples are the same as those of as-grown samples in Figure 2.11, but the CL intensities of the annealed N-polar. t11@ t2 5 and t11@ t2 2 samples are weaker than those of the as-grown ones. After annealing, the relative intensities of two CL peaks are different.. 2.6.2 X-ray Diffraction (XRD) Patterns Figure 2.20 shows the XRD patterns for the two anneals samples. The diffraction peaks corresponding to InGaN and GaN can be identified. The GaN diffraction peak is mainly from the contact and barrier layers. The side shoulder with a broad distribution below the GaN main peak is attributed to InGaN with various indium contents, sizes, and shapes in the quantum wells [10]. The indium contents of the annealed N-polar t11@ t2 2 sample are larger than those of the annealed N-polar t11@ t2 5 one. After annealing, the density of InGaN mounds increases.. 2.6.3 Atomic Force Microscopy (AFM) Study Figures 2.21 (c) and (d) show the AFM images of the annealed N-polar t11@ t2 5 and. t11@ t2 2 samples, respectively. The InGaN mounds affect the surface roughness. The surface roughness in 5 μm × 5 μm area are 26.694 and 27.295 nm for the annealed N-polar 36.

(50) t11@ t2 5 and t11@ t2 2 samples, respectively. The surface roughness of the annealed samples becomes smaller than that of the as-grown ones.. 2.7 Temperature-dependent PL Studies of annealed Nitrogen-Polar InGaN/GaN MQWs Figures 2.22 (a) and (b) show the PL spectra of the annealed N-polar t11@ t2 5 and. t11@ t2 2 samples as a function of temperature, respectively. The peak positions of the two samples are categorized into two parts. One spectral range is from 350 nm to 430 nm, and the other one is from 430 nm to 500 nm. The PL intensity decays with increasing temperature. After annealing, the relative intensities of PL peaks are different. The PL peak positions as a function of temperature for the two samples are shown in Figure 2.23 (a). The annealed N-polar t11@ t2 5 sample has three peaks: ~379 nm (3.272 eV), ~390nm (3.179 eV), and ~450 nm (2.756 eV). The annealed N-polar t11@ t2 2 sample also has three peaks: ~379 nm (3.272 eV), ~390nm (3.179 eV), and ~470 nm (2.638 eV). The dominate peaks of the two samples around ~379 nm (3.272 eV). The three emission peak positions of the two samples are related to InGaN. The peak positions around ~379 nm (3.272 eV) and ~390nm (3.179 eV) for the annealed N-polar t11@ t2 5 and t11@ t2 2 samples are nearly the same. With increasing temperature form 20 to 300 K, PL positions of the two samples are nearly temperature-independent. In addition, the normalized integral PL intensities as a function temperature of the two annealed samples are shown in Figure 2.23 (b). The integral PL intensity of the annealed N-polar t11@ t2 5 sample decays slower than those of the annealed N-polar t11@ t2 2 one.. 2.8 Discussion and Summary In summary, we have shown the experimental results of SEM, CL, XRD, AFM, PL measurements of N-Polar InGaN/GaN SQW and MQW samples. The surface roughness of 37.