氮化鎵之發光二極體因為有效率下降的問題，所以限制了大尺寸和高功率的發展，我們利用氧化鋅摻鈷薄膜來降低效率下降的問題。本研究是將應用在發光二極體上的脈衝雷射鍍膜儀所成長的氧化鋅摻鈷薄膜，並與射頻濺鍍法鍍膜成長通入氫氣的薄膜來做比較，以利爾後在工業上的應用。因為以射頻濺鍍法所成長的氧化鋅摻鈷薄膜電阻值相當高，所以先將其結構特性做比較，之後再利用摻雜(鎵,鋁)等元素改善電性。本研究以x光吸收光譜和x光繞射法來確認其結構的變化，可以發現在調變不同的射頻濺鍍成長條件(壓力、基板溫度、瓦數等)會造成氧空缺和鈷間隙的改變，且並無二次相析出。利用x光吸收光譜和X光吸收光譜近緣結構以及延伸X光吸收光譜精細結構，可以觀察到氧和鈷未佔有軌域上的數量變化，和電子上的轉移，藉此來確認難判斷的氧空缺，並更精細確認鈷取代鋅的位置。我們也利用氧化鋅摻鈷和氧化鋅雙靶濺鍍，來調變鈷摻入氧化鋅的比例。最後，我們在(基板溫度400度、瓦數70瓦、壓力3.75E-3mbar、混入微量氫氣)成長出和脈衝雷射鍍膜儀結構相當的薄膜。

ABSTRACT

SUMMARY
Based on previous study, we have already improve the luminous efficiency of light-emitting diodes (LEDs) by depositing zinc oxide related materials thin film between electrode layer and n-GaN layer. Although we get similar results by depositing cobalt doping zinc oxide on pulse laser deposition(PLD), the precise mechanisms that improve luminous efficiency have not confirm. In this research, we assume zinc oxide thin film will reduce the Efficiency droop effect. In order to solve PLD method can't deposit large size chip which is hard to apply by LEDs industry, we use sputter to reconfirm our result. In this research we focus on cobalt doping zinc oxide structure deposited by sputter and compare with PLD thin film. We use X-ray absorption spectrum including (XAS,XANES,EXAFS) and X-ray diffraction method to analyze our thin film. By changing different depositing parameters(Hydrogen, Temperature, Pressure, Percentage), we can observe the difference between each structure. In this research, we find out that Hydrogen can induce Hydrogen interstitial or oxygen vacancy. When the temperature raising, there are few cobalt interstitial will be observed and the amount of oxygen vacancy will also change. Pressure can also affect thin film structure. We use two target to modulate the percentage of cobalt doping in zinc oxide. At the end, we find out that under depositing parameters(temperature:400℃,pressure: 3.75E-3 mbar, Hydrogen mixed, power: 70w)can get the similar structure to previous result.

Key word: Sputter, ZnO , XAS, Efficiency droop

INTRODUCTION
Light emitting diodes are one of the most important components in optoelectronic industry. Because of low power consumption, high lift time and fast reaction, LEDs research attracts increasing attention in the energy-price-up surging era. Some materials such as GaP, GaAs have been commonly used in yellow-blue LEDs. Facing strong global competition in full color LED industry, developing of high efficiency blue LEDs for illumination are ranged from 130-150 lm/W. Ideally, increasing the LED driving current enables the same lumen output to be achieved with less LED chips or using smaller LEDs. Unfortunately, this straightforward route to cost-per-lumen reduction is not readily available in actual GaN LEDs, One of the most significant challenges facing high-power and large size GaN-based LEDs is the efficiency droop effect which means the decrease in external quantum efficiency (EQE)of an LED with increasing injection current in quantum well. Peak internal quantum efficiency (IQE) occurs at relatively low-current densities , then rolls off as current density increases. Typical GaN-based LEDs have a peak in efficiency, typically at current densities less than 10 A/cm2, above which the efficiency gradually decreases. Despite having been the subject of extensive research efforts for a decade, the physical origin of droop has not been clarified. Several mechanism have been proposed to explain this LED droop effect, including electron leakage, poor hole injection, delocalization carriers, Auger recombination. In this research, we deposit cobalt doping zinc oxide thin film on n-GaN try to modulate mobility of n-GaN which can improve poor hole injection. Based on several reference, n-GaN mobility and carrier concentration will influence efficiency droop effect. Zinc oxide (ZnO), a wide bandgap (3.4 eV) II-VI compound semiconductor, has a stable wurtzite structure with lattice spacing a = 0.325 nm and c = 0.521 nm. It has attracted intensive research effort for its unique properties and versatile applications in transparent electronics. X-ray Absorption Spectroscopy (XAS) includes both Extended X-Ray Absorption Fine Structure (EXAFS) and X-ray Absorption Near Edge Structure (XANES). X-ray absorption spectra also contain information about the valence state of elements in materials. If the element is present as a cation, the absorption edge is shifted to lower (higher) photon energy because of the lower (higher) ionization potential The “X-ray absorption near edge structure” (XANES) is due to transitions into unoccupied bound states below the edge of the continuum. The “Extended X-ray absorption fine structure” (EXAFS) above the edge is due to backscattering of the photoelectron to the emitting atom. In our research, we use XRD method to check that lattice distance change and secondary phase form when growth parameters change. But XRD can't precise observe whether cobalt substitute zinc place or not, and this method can just determine cobalt interstitials and vacancy formed roughly by lattice distance changed. So we use XAS to analyze thin film structure more precisely.

MATERIALS AND METHODS
We deposit cobalt doping zinc oxide by sputter. First, we compare two conditions, one is pure Ar gas, the other is 5％ H2 gas mixed. Second, we modulate different parameters, for example, temperature:(400-750℃),pressure(1E-2,3.75E-3,1E-3mbar),
power:70w.Third,we use two targets(ZnO,CZO) with different power to modulate cobalt percentage in zinc oxide. Last, we compare with PLD thin film, which substrate temperature is 300℃,frequency is 2HZ, energy is 40mJ. X-ray Absorption(XAS,XANES, EXAFS) is measured in NSRRC 20A,07 beamline in Taiwan. Structural characterization was carried out by X-ray diffraction (XRD). Surface characterization was carried out by SEM, thickness of thin film was carried out by α-step, all were measuring in NCKU.

RESULTS AND DISCUSSION
First, we compare with the pure argon gas condition and five percentage hydrogen gas mixed condition. In XRD result, we can rule out secondary phase formed and we can observe a slightly move at 34 degree toward small angle when hydrogen gas mixed. It implies lattice distance increase that may cause by hydrogen interstitial. In XAS spectrum we can observe oxygen vacancy formed. In cobalt L-edge spectrum can find out that unoccupied states of cobalt are increasing when hydrogen mixed with argon and oxygen unoccupied states are decreasing. Because of oxygen vacancy formed, the amount of oxygen is decrease which hybridization with cobalt. This cause the peak and area in oxygen K-edge spectrum decrease. In XANES and EXAFS spectrum, we can rule out cobalt is metallic or cobalt oxide form in crystal, because of the shape of our spectrum is totally different to reference. So that we can confirm that cobalt substitute zinc. Second, we change different substrate temperature from 400-750℃. In XRD spectrum, we observe the lattice distance decrease when temperature raised. So we assume that hydrogen will escape from substrate when temperature raised. So the ability of forming oxygen vacancy decrease. In XAS spectrum, we can find out that the amount of oxygen vacancy decrease and unoccupied states increase in oxygen K-edge spectrum when temperature raised. At 750℃,the unoccupied states of oxygen are decrease. Based on reference, it may cause by cobalt interstitial. In XANES and EXAFS spectrum, we can confirm that cobalt substitute zinc whether temperature change or not. Because the cobalt interstitials amount are small, so it can't influence XRD and EXAFS spectrum. Third, we modulate pressure parameter (1E-2,3.75E-3,1E-3mbar),and we find out 3.75E-3 mbar is the better parameter of our sputter system. Obviously, the 3.75E-3 mbar condition FWHM of ZnO(002) peak is more narrow then the other. And photon energy of1E-2,1E-3mbar condition are shifting toward small angle that means the cobalt valence are more approaching to zero. Fourth, we can successfully use ZnO and CZO targets to deposit different cobalt percentage thin film by modulate different power of target, but sacrifice thin film quality. In XAS spectrum, different power of ZnO target can lead different cobalt behavior hybrid with oxygen. Last, we compare PLD with sputter result, we can observe at (temperature:400℃,pressure: 3.75E-3 mbar, Hydrogen mixed, power: 70w) this condition the structure that sputter is similar to PLD result whether in XRD or XAS spectrum. Although, the peak shape are slightly different, we approach our initial goal.

CONCLUSION
In Hydrogen condition, argon mix Hydrogen can produce oxygen vacancy but also formed some interstitial site. In Temperature condition, when the substrate temperature raise up, we can observe that may decrease hydrogen interstitial and little cobalt interstitials formed at 750.In Pressure condition,3.75E-3mbar is a better growth pressure in our system. In Co-sputter condition, we can modulate Co concentration by using CZO and ZNO target Co-sputter. But quality of thin film we be destroy slightly. we find out the growth condition at(temperature:400℃,pressure: 3.75E-3 mbar, Hydrogen mixed, power: 70w)can approach our previous PLD result. Because of the poor electrical properties of CZO film by sputter, we have to co-doping other materials( Al, Ga…) to improve its electrical properties. Then we can deposit on GaN substrate to check our result and apply on LED.

(1) David S. Meyaard , Guan-Bo Lin et al. "Asymmetry of carrier transport leading to efficiency droop in GaInN based light-emitting diodes" Appl. Phys. Lett. 99, 251115 (2011)
(2) Guan-Bo Lin, David Meyaard et al." Analytic model for the efficiency droop in semiconductors with asymmetric carrier-transport properties based on drift-induced reduction of injection efficiency" Appl. Phys. Lett. 100, 161106 (2012)
(3)Giovanni Verzellesi, Davide Saguatti et al."Efficiency droop in InGaN/GaN blue light-emitting diodes: Physical mechanisms and remedies"Journal of Applied Physics 114, 071101 (2013)
(4) Jong-In Shim, Dong-Pyo Han et al." Efficiency droop in AlGaInP and GaInN light-emitting diodes" Applied Physics Letters 100, 111106 (2012)
(5) Jaehee Cho, E. Fred Schubert et al." Efficiency droop in light-emitting diodes: Challenges and countermeasures" Laser Photonics Rev. 7, No. 3, 408–421 (2013)
(6)W. G. Scheibenzuber et al."Recombination coefficients of GaN-based laser diodes" J. Appl. Phys. 109, 093106(2011)
(7) 許庭維,氧化鋅摻雜鈷在低載子濃度下磁性、電性及光學性質研究, 國立成功大學物理學系碩博士班, 2011
(8) 陳金宏,鈷鎵共摻雜氧化鋅薄膜覆蓋於氮化鎵發光二極體之光、電特性研究, 國立成功大學物理學系碩博士班, 2012
(9)發光二極體之原理與製程,陳隆建 編著,全華科技圖書有限公司(2010)
(10)LED 原理與應用,郭浩中 賴芳儀 郭守義著,五南圖書出版公司(2013)
(11)K.Sato and H.Katayama-Yoshida,Jpn.J.Appl.Phys,40,L334(2001)
(12)J.M.D.Coey, A.P.Douvalis, C.B.Fitzgerald, and M. Venkatesan,Appl.Phys.Lett,84,1332(2004)
(13)J.J. Rehr and R.C.Aleber,Rev.Mod.Phys.,72,621(2000)
(14)Koningsberger and R.Prins,"X-ray Absorption principles, applications, techniques of EXAFS,XANES"(1988)
(15)J.E.Muller and J.W.Wilkins,Phys.Rev.B,29,4331(1984)
(16) 同步辐射应用概论（第二版）,马礼敦 杨福家 , 复旦大学出版社(2005)
(17) M. G. Wardle, J. P. Goss, and P. R. Briddon" Theory of Fe, Co, Ni, Cu, and their complexes with hydrogen in ZnO" PHYSICAL REVIEW B 72, 155108(2005)
(18) M. D. McCluskey et al. " Defects in ZnO", Journal of Applied Physics 106, 071101 (2009)
(19) F M F de Groott, M Abbatet et al." Oxygen 1s and cobalt 2p x-ray absorption of cobalt oxides" J. Phys: Condens. Matter 5 (1993) 277-2288
(20) J. W. Chiou, H. M. Tsai, C. W. Pao et al." Role of valence-band Co 3 d states on ferromagnetism in Zn 1 x Co x O nanorods" Applied Physics Letters 90, 062103 (2007)
(21) C. McGuinness, L. S. Dorneles et al. "Soft-x-ray spectroscopic investigation of ferromagnetic Co-doped ZnO", Journal of Applied Physics 99, 08M111 (2006)
(22) J-H Guo, L Vayssieres et al." Polarization-dependent soft-x-ray absorption of highly oriented ZnO microrod arrays" J. Phys.: Condens. Matter 14 (2002)
(23) Tao Liu, Hairuo Xu et al." Local Structural Evolution of Co-Doped ZnO Nanoparticles upon Calcination Studied by in Situ Quick-Scan XAFS ", J. Phys. Chem. C 2008, 112, 3489-3495
(24) Tongfei Shi,1 Zhenguo Xiao et al." The role of Zn interstitials in cobalt-doped ZnO diluted magnetic semiconductors" APPLIED PHYSICS LETTERS 96, 211905 (2010)
(25) Tongfei Shi, Zhenguo Xiao et al."The role of Zn interstitials in cobalt-doped ZnO diluted magnetic semiconductors", APPLIED PHYSICS LETTERS 96, 211905( 2010)
(26) K. Potzger, K. Kuepper et al. " High cluster formation tendency in Co implanted ZnO" Journal of Applied Physics 104, 023510 (2008);
(27) G. S. Chang et al." Effect of Co and O defects on the magnetism in Co-doped ZnO: Experiment and theory ", PHYSICAL REVIEW B 75, 195215,(2007)