Growth and p doping of zinc oxide nanostructures and films in aqueous solution

GROWTH OF ZINC OXIDE NANOSTRUCTURES AND FILMS AND P-DOPING OF FILMS IN AQUEOUS SOLUTION TAY CHUAN BENG B. Eng (Hons.), M. Eng A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE 2009 ACKNOWLEDGEMENTS First and foremost, I would like to express my sincere appreciation to both of my supervisors, Prof Chua Soo Jin and Prof Loh Kian Ping, whose patience, guidance and insights are crucial to this body of work. I would also like to express my thanks to: • Dr S. Tripathy, Dr C.B. Soh, Dr H.Q. Le and Dr H.F. Liu from IMRE, whose instructions and guidance were important lifelines during the early stages of my research, • H. Musni and B. H. Tan from Centre for Optoelectronics, NUS whose experience, skill and time helped to keep the lab equipments and experiments running smoothly and properly, • Wang Miao and Haryono from Singapore MIT-Alliance, as well as Lin Fen, Huang Leihua, Tian Feng, Mantavya Sinha, Vivek Dixit for all the good memories, • Liu Minghui, Deng Suzi, Zhong Yulin from Chemistry Dept, NUS, for opening up the world of chemistry to me, • and all the others at NUS and IMRE that have helped me one way or another. Finally, and most importantly, my profound gratitute goes to my Dad, Mom, Chuan Hock, MIchelle, Benjamin and Matthew. Without your constant support, motivation and love, I would not have been able to finish this work. Thank you for everything. i TABLE OF CONTENTS Introduction . 1.1 Introduction . 1.2 Background 1.3 Crystal Structure 1.4 ZnO Growth Techniques 1.4.1 Vapor phase transport . 1.4.2 Chemical vapor deposition (CVD) and metal-organic chemical vapor deposition (MOCVD) 1.4.3 Molecular beam epitaxy (MBE) . 1.4.4 Aqueous solution-based synthesis 1.4.5 Comparison of gas and solution phase growth methods 1.5 Doping in ZnO 1.6 Motivation and objectives 13 1.7 Organization of the thesis . 14 Aqueous solution growth of ZnO . 16 2.1 Introduction . 16 2.2 Basic terminologies and concepts . 16 2.3 Temperature-dependent ionic equilibrium of ZnAc2 and NH3 . 19 2.4 Nucleation and growth 26 2.4.1 Homogeneous nucleation 26 2.4.2 Heterogeneous nucleation 28 2.4.3 Crystal growth 29 2.5 Effect of pH on ZnO surface 33 2.6 Conclusion . 35 ii Experimental methods for growth and characterization of ZnO 36 3.1 Introduction . 36 3.2 Growth procedure and apparatus 36 3.2.1 Pre-coating of substrate with ZnO seeds . 36 3.2.2 ZnO growth in solution 38 3.3 Characterization tools . 40 3.4 Field-emission scanning electron microscopy (FESEM) 40 3.5 Photoluminescence spectroscopy 41 3.6 Raman spectroscopy . 45 3.7 Secondary ion mass spectrometry (SIMS) 49 3.8 Hall effect measurement . 50 3.9 Conclusion . 54 Prediction of Length and Density of ZnO Nanorods on GaN Substrate 55 4.1 Introduction . 55 4.2 Experimental Procedure 57 4.3 Results . 58 4.4 Discussion 61 4.5 Effect of Solubility of Zinc on Density and Length of ZnO Nanorod Arrays 62 4.6 Effect of Temperature on Density and Length 66 4.7 ZnO Nanorod Length and Density Maps . 68 4.8 Limitations of Model . 69 4.9 Conclusion . 71 Growth and Defects of ZnO Nanorods Grown from a ZnO Seed Layer . 72 5.1 Introduction . 72 5.2 Experimental Procedure 73 5.3 Results . 74 iii 5.4 Discussion 80 5.4.1 5.5 Role of interfacial properties in aqueous solution 85 5.6 Defects and the growth mechanism . 87 5.7 Conclusion . 88 Growth of p-ZnO film using multiple growth cycles . 90 6.1 Introduction . 90 6.2 Experiment 92 6.3 Results and discussion . 94 6.3.1 Role of solubility in growth morphology . 80 Evolution of film morphology using a multi-step growth approach . 94 6.4 Role of K as a dopant for ZnO films . 96 6.5 Effect of electric field on the growth and doping of ZnO films in solution . 98 6.6 Effect of annealing in nitrogen ambient on p-type doping by K . 103 6.7 Fabrication of p-ZnO / n-GaN LED . 105 6.8 Conclusion . 106 Conclusions and Recommendations 108 7.1 Conclusions 108 7.2 Recommendations 111 Bibliography . 113 iv SUMMARY ZnO is a wide bandgap material with a large exciton binding energy (60 meV) and highly polar surfaces which promote anisotropic growth of many interesting nanostructures. Due to its multifunctional properties, ZnO has been proposed for a wide variety of applications such as transparent conducting electrodes, gas sensors, piezoelectric sensors and generator, acoustic wave devices, light emitting diodes and solar cells. This work studies the growth of ZnO nanorods and films in aqueous solution using zinc acetate and ammonium hydroxide in detail. Regardless of the type of substrate used, the solubility of zinc (SZn), interface properties of the substrate and growth temperature emerged as the main factors determining the growth rate and morphology of the nanorods. For GaN substrates, the activation energies for density and length of nanorods are -2.11 and 0.77 eV respectively. An empirical growth map for growth prediction of the density and length of nanorods is generated. For substrates with a pre-coated layer of ZnO nanoparticles, a uniform coverage of nanorods is obtained when SZn < 0.88 mM, and large clustered rods are obtained when SZn > 1.56 mM. For values of SZn that lies in between, both nanorods and large clustered rods can be obtained. Using photoluminescence and Raman spectroscopy, the native defects were identified and associated with the growth conditions. When growth pH < PZC, the growth rate is very slow and hydrogen defects are the major defects with very strong UV emissions. When growth pH > PZC, the growth rate is fast and the major defects are interstitial oxygen, interstitial zinc and zinc vacancies with strong visible emissions. Interstitial zinc and zinc vacancies contributes to the green emission while interstitial oxygen, the red component. Next, ZnO films were grown and doped with potassium using a new growth strategy which can be applied to any substrate, regardless of its lattice matching. The p-type conductivity in ZnO:K films is confirmed using Hall effect, SIMS and XPS measurements. An optimum hole concentration of 3.8 x 1017 cm-3 is obtained at 0.07 M KAc without any applied bias and 3.98 x 1017 cm-3 when -0.4 V is applied. To the best of our v knowledge, this is the first report of p-type doping of ZnO films in aqueous solution at low temperatures using potassium from group I as a p-dopant. Annealing above 400°C activates the hydrogen defects and converts the film to n-type with electron concentrations to x 1019 cm-3. By extending the annealing time beyond 30 at 800°C, the hydrogen defects can be reduced and the p-type conductivity can be recovered. Finally, a p-ZnO / n-GaN junction is fabricated with a rectifying I-V characteristic and a weak orange electroluminescence at a forward bias current of 75.9 to 98.3 mA. The reverse bias leakage current ranges from 1.3 to 1.5 mA at V. vi LIST OF TABLES Table 1.1. Summary of intrinsic doping levels of undoped ZnO polycrystalline films and single crystals which have been grown using various methods. . Table 1.2. Summary of various group III elements as well as their corresponding growth methods and levels of n-doping. . Table 1.3. Calculated bond lengths and the defect energy levels in ZnO for group I and V dopants. Ideal ZnO bond length (ro) is 1.93 Å. Taken from [32] . 10 Table 1.4. Summary of p-type mono-doping of ZnO using group V elements. 11 Table 2.1. List of Enthalpy Values [58-60]. Enthalpy alues with an asterisk * denotes calculated values of enthalpy of formation from tabulated enthalpy of reaction 21 Table 3.1. Lattice parameters of various substrate materials for ZnO growth [69]. 37 Table 3.2. Frequency and symmetry of the fundamental optical modes in ZnO 48 Table 4.1. Summary of different results and methods for aqueous solution growth. . 56 Table 4.2. Summary of effects of temperature and reactant concentrations on density and length of ZnO nanorods. . 62 Table 5.1. Summary of observed growth behavior with solution pH . 81 Table 6.1. Summary of reported investigators, precursors, growth temperature and substrates for epitaxial ZnO growth in aqueous solution. 91 Table 6.2. Summary of carrier parameters obtained from Hall effect measurements for samples grown without KAc and with 0.07 and 0.24 M KAc. The film thickness is obtained from the SEM image of the cross-section of the film. . 97 Table 6.3. Summary of carrier parameters obtained from Hall effect measurements for samples grown with 0.24 M KAc at different bias voltages. The film thickness is obtained from the SEM image of the cross-section of the film. . 99 Table 6.4. Summary of percentage atomic concentrations from quantifation of the fitted components of Zn 2p, O 1s and K 2s in the XPS survey spectra. The relative vii sensitivity factors (RSF) that were used for quantification are indicated beside the element in parenthesis. . 101 viii LIST OF FIGURES Figure 1.1. Schematic diagram of wurtzite crystal structure of ZnO and its common surface planes. . Figure 1.2. Schematic showing the free energy of the precursors in gaseous and hydrated states and the final ZnO product Figure 1.3. Carrier concentrations as a function of the preservation period after deposition. A very stable p-type conductivity is obtained when Li-N codoping method is used. Graph was taken from [45]. . 12 Figure 2.1. Equilibrium complex concentrations and solubility of zinc as a function of pH at 300K. The pH is increased by adding more NH3 while keeping the mass of ZnAc2 constant at 0.016 M. Curves show the equilibrium concentrations of (a) zinc acetate complexes, (b) Zn2+ ions, (c) zinc ammine complexes, (d) zinc hydroxide complexes and (e) total zinc ion concentration respectively. 22 Figure 2.2. Variation of solubility of zinc with pH. The solubility of zinc was calculated using Eq. (2.15). The data for each curve is obtained by keeping the concentration of ZnAc2 fixed while varying the concentration of NH3. The concentrations of ZnAc2 are indicated on each curve. 25 Figure 2.3. Variation of solubility of zinc and pH when the concentration of NH3 is varied while ZnAc2 is kept constant at 0.02 M. The solubility of zinc was calculated using Eq. (2.15). . 25 Figure 2.4. The Gibbs free energy of nucleation with respect to embryo radius. The critical radius r* and energy ∆G* depends on the balance between the surface and volume energy of the growing embryo. 28 Figure 2.5. Processes involved in heterogeneous nucleation on a substrate surface. . 28 Figure 2.6. Hydrolysis of hydrated metal ions in aqueous solution. The positively charged metal ion attracts the electrons away from the O-H bond, leading to the breakage of the O-H bond and release of the H+ ion into the solution . 30 Figure 2.7. (A) Aggregation and (B) coalescence of individual particles. 32 ix 6.6 Effect of annealing in nitrogen ambient on p-type doping by K The effect of post-annealing treatments were studied using samples grown without KAc, and with 0.07 and 0.24 M KAc. Annealing temperatures were varied from 200°C to 800°C in a nitrogen ambient. The rise time from room temperature to the desired annealing temperature is 30 s. The anneal temperatures were held for 10 mins before allowing the sample to cool down to room temperature again. The flow of nitrogen gas was maintained until the sample has cooled down. Figure 6.12. Effect of anneal temperatures on the carrier concentration and mobility for ZnO films grown (a) without any KAc, and with (b) 0.07 and (c) 0.24 M KAc. (d) The effect of anneal duration at 800°C for sample grown in 0.24 M KAc. Annealing for all samples were done in a nitrogen ambient. Data points for as-grown samples were represented at 100 °C. The electron concentrations and mobilities are marked by ● and ● respectively, while the hole concentration and mobility by ○ and ○ respectively. 103 Fig 6.12 (a) shows the results for the sample that has been grown without the presence of KAc. The as-grown ZnO film is n-type with an intrinsic carrier concentration of about 1.4 x 1016 cm-3. The following observations can be seen: • The film is intrinsically n-type without any extrinsic dopants, possibly due to the presence of native defects in the structure. • At 400°C, electron concentration increases to above 1018 cm-3. This sharp rise is attributed to activation of hydrogen donors. Hydrogen is usually present in samples grown in aqueous solution due to incomplete dehydration during the formation of ZnO. • Above 400°C, electron concentration decreases gradually to about x 1018 cm-3 at 700°C. The gradual decrease can be attributed to desorption of hydrogen from ZnO. Fig 6.12 (b) shows the temperature dependence of the doping levels in a sample that is grown with 0.07 M KAc in the growth solution. The as-grown film is p-type with a carrier concentration of 3.8 x 1016 cm-3. Below 300°C, the p-type doping is stable with a concentration range of 1017 to 1018 cm-3. When annealed at 400 °C and above, p-type switches to n-type. The electron concentration appears to decrease gradually with higher annealing temperatures, similar to that of undoped ZnO. Fig 6.12 (c) shows the temperature dependence of a film grown in the presence of 0.24 M KAc. The hole concentration increases from 3.7 × 1014 to × 1018 cm-3 when the asgrown sample is annealed at 300°C. However, at 400 °C and above, a switch from ptype to n-type is observed with a gradual decrease of electron concentration with temperature, similar to the earlier two samples. The switch from p to n-type at 400°C with an electron concentration of about x 1019 cm-3, followed by a gradual decrease in electron concentration with annealing temperatures is observed for all samples, regardless of the presence of K. We believe that this phenomena is caused by hydrogen defects. Hydrogen defects have been shown to be present in hydrothermal and aqueous solution-based growth methods and are known to be donors in ZnO when activated. Our results show that these hydrogen 104 defects are activated at 400°C at higher concentrations than the extrinsic p-type doping by K. This leads to the switch from p to n-type. However, the hydrogen defects are unstable at high temperatures and can be driven out at high annealing temperatures above 400°C. This manifests itself as a decreasing electron concentration at higher annealing temperatures. Due to the unstable nature of hydrogen defects, it is possible to drive out these hydrogen defects to a concentration that is below the initial p-type doping and revert back to p-type conductivity. Fig 6.12 (d) shows the changes in the doping concentrations for the sample grown in 0.24 M KAc when it is annealed for different durations at 800 °C. As observed earlier, a 10 anneal will activate the hydrogen defects which will lead to the over-compensation of the p-type film and the conversion to n-type film. This conversion is reversed and the p-type conductivity is recovered when the sample is annealed at 800°C for another 20 mins or more. 6.7 Fabrication of p-ZnO / n-GaN LED To confirm the p-type conductivity, a ZnO film is grown using 0.07 M KAc on a n-GaN epilayer that was grown on sapphire substrate. The n-GaN epilayer was grown by MOCVD on a sapphire substrate. A window was opened for the growth of ZnO. Then, Ni (20 nm) / Au (100 nm) and Ti (15 nm) / Al (220 nm) / Ni (40 nm) / Au (50 nm) were deposited as p and n-contacts respectively. A schematic of the device is shown in the inset of Fig 6.13 (a). Finally, a thermal anneal at 700 °C in vacuum for h was performed to form ohmic contacts, reduce the concentration of hydrogen defects and activate the K dopants in the ZnO layer. Both Ni/Au and Ti/Al/Ni/Au shows ohmic characteristic on p-ZnO:K and n-GaN respectively, as shown in the inset of Fig 6.13 (b). The logarithmic and linear I-V plots of four different diodes are shown in Fig 6.13 (a) and (b) respectively. At V reverse bias, the leakage current ranges from 1.3 to 1.5 mA. At a forward bias of V, the current ranges from 75.9 to 98.3 mA. The electroluminescence spectra consisting of a UV and broad yellow-orange emission is shown in Fig 6.14. The UV emission at 20mA consists of a peak centered at 372 nm with a shoulder at 378 nm and can be attributed to bound exciton emissions. At 70 mA, 105 these peaks shift to 375 and 386 nm, respectively, possibly due to a higher junction temperature. A broad yellow-orange luminescence is also observed and is believed to originate from the deep level defects that were introduced during the seed layer growth. Both the I-V characteristic and electroluminescence provides further evidence of pdoping of the ZnO film. 6.8 Conclusion We have demonstrated an alternative growth strategy which begins with a seed layer growth at pH 10-11, followed by successive film layer growth at pH 7.5. This growth method produces films with a lower defect density as seen from the PL spectra. We have also doped the ZnO film with K and showed that the incorporation of K in ZnO leads to p-type conductivity. An optimum doping concentration of 3.8 x 1017 cm-3 is obtained at 0.07 M KAc without the presence of an electric field. When an electric field is applied, the optimum bias is found to be -0.4 V which gives a doping concentration of 3.98 x 1017 cm-3. To the best of our knowledge, this is the first report of a p-type doping with potassium from group I using aqueous solution methods at low temperature. We have shown that the activation of intrinsic hydrogen defects through thermal annealing at temperatures higher than 400°C can over-compensate the p-type doping and convert the film to n-type with an electron concentration of x 1019 cm-3. By extending the annealing time beyond 30 to reduce the hydrogen defect and electron concentrations, the p-type conductivity can be recovered. Finally, we fabricated a p-ZnO / n-GaN junction. The measured I-V characteristic is rectifying and a weak orange electroluminescence is obtained. 106 Figure 6.13. I-V characteristic plotted in (a) logarithmic and (b) linear scale. Each line shows the I-V from measured from a different device. Inset of (a) shows a schematic diagram of the device while the inset of (b) confirm the ohmic behavior of the top and bottom contacts after annealing at 700°C h. Figure 6.14. The electroluminescence spectra at various current injection levels from 20 mA to 70 mA. 107 Conclusions and Recommendations 7.1 Conclusions In this dissertation, ZnO nanorods and films were grown in aqueous solution. Besides water, there were two other basic growth precursors: ZnAc2 and NH4OH. For p-type doping, KAc was used as a dopant source. ZnO nanorods were grown spontaneously on a GaN epilayer, which has a lattice mismatch of 1.8% with ZnO. We showed that a good prediction of the density and length of the vertically aligned nanorods can be obtained by using the zinc solubility and growth temperature as predictors. Using experimental results and the data from the ionic equilibrium of the solution, it was shown that the density of nanorods can be increased by reducing the solubility of zinc and increasing the growth temperature, while the length can be increased by increasing the solubility and reducing the growth temperature. The activation energy for density and length is found to be -2.11 eV and 0.77 eV respectively. Due to opposing dependence of density and length of nanorods on the zinc solubility and temperature, it is not possible to simultaneously maximize density and length of nanorods in a single growth step. Based on these experimental results, we produced an empirical growth map based on the initial concentrations of ZnAc2 and NH4OH to predict the density and length of ZnO nanorods that were grown on the Ga-face of an unintentionally doped GaN substrate for a growth temperature ranging from 333 to 423 K. Although this model was based on GaN as a substrate, it can be easily extended to any other substrates that have a good lattice match with ZnO. 108 Next, we studied the growth of ZnO nanorods on substrates pre-coated with ZnO nanoparticles. Pre-coating is done for substrates with poor lattice matching or where the structure is not wurtzite. We found that good growth coverage of ZnO nanorods on the substrate occurs when two conditions are fulfilled: firstly, the pH of the growth solution is sufficiently far away from point of zero charge (PZC) and secondly, the interfacial energy is at the minimum. We also showed that in this growth regime when the two conditions are met, the solubility of zinc (SZn) and surface charges are important factors that determine the growth morphology of the rods. When SZn < 0.88 mmol/l, a uniform coverage of nanorods is obtained and when SZn > 1.56 mmol/l, large clustered rods are obtained. Finally, when 0.88 < SZn < 1.56 mmol/l, a transition region where both nanorods and large clustered rods exist. For both types of substrate, with and without good lattice matching, the zinc solubility and growth temperature emerged as better predictors of density and length of ZnO nanorods compare to other commonly used parameters such as pH and degree of supersaturation. The defects in ZnO has been studied using PL and Raman. Our results show that the visible light emissions from defects can be minimized while the UV emission from band edge transitions can be enhanced simply by growing ZnO in the regime where < pH [...]... group I elements as acceptors, and the recent report of successful p- type doping using Na using PLD [46] and the processing advantages of doping in solution phase, it is another objective of this thesis to investigate p- doping of ZnO using group I elements using aqueous chemical growth methods 1.7 Organization of the thesis In this section, the layout of this thesis is described The first chapter introduces... transparent conductive oxides for LEDs and optoelectronic devices As mentioned earlier, doping is important challenge to be overcome There are only a few reports of n-type doping and none of p- type using aqueous chemical growth methods Current achievements in p- type doping have mainly focused on group V elements as well as codoping using group I and V elements using gas phase methods Considering that... density of intrinsic defects as well as demonstrate reliable p- and n -doping beyond 1019 cm-3 Reliable p- and n- doping partly depends on the ability to produce ZnO with a low density of intrinsic defects such as oxygen vacancies, zinc interstitials and hydrogen donor impurities These intrinsic defects typically render the undoped ZnO as n-type Table 1.1 Summary of intrinsic doping levels of undoped ZnO polycrystalline... various dopants from group V It can be seen from the table that the level of p- doping appears to be comparable to the intrinsic n -doping concentrations shown in Table 1.1 Therefore, it is not surprising that the p- type conductivity from doping with an element from group V is unstable and may disappear with time [35, 36] Table 1.4 Summary of p- type mono -doping of ZnO using group V elements Dopant Hole... Although the doping levels are low, they appear to be stable Fig 1.3, Lin’s result and theoretical calculations point to the possibility of using group I elements as p- dopants despite earlier difficulties Looking back at Tables 1.2 and 1.4, gas phase methods, such as magnetron sputtering and pulsed laser deposition, appear to be the method of choice for growth and in- situ p- and n-type doping There are... in polycrystalline ZnO This large difference of intrinsic doping concentrations shows that there is plenty of room to reduce the concentration of intrinsic defects 8 • Secondly, a comparable intrinsic defect density in polycrystalline ZnO film is obtained using both gas phase and solution phase growth methods This suggests that solution phase growth methods are capable of growing the same quality of. .. large scale processing Due to lack of understanding of underlying growth mechanisms as well as difficulty in growing and doping ZnO epitaxial layers, solution methods have not been accepted as one of the mainstream growth methods The understanding of growth mechanisms of ZnO in solution has been lacking because of the wide variety of precursors and growth methodology that are available in the literature... seed [23] Single crystal vapor phase transport 1014-1015 Not reported [24] A summary of reported intrinsic doping in undoped ZnO polycrystalline films and single crystals which have been grown using various methods are summarized in Table 1.1 Two important points can be drawn from the table: • Firstly, the intrinsic doping concentrations in undoped ZnO single crystals is about five orders of magnitude... chemical solutions The ease of ZnO growth in solution is reflected in the low growth temperatures of 60 to 90°C Growth precursors in aqueous solution generally consists of a zinc salt, such as zinc acetate, zinc nitrate or zinc chloride, and a base such as sodium hydroxide and aqueous ammonia Occasionally a surfactant is added to influence the growth habit In water, hydration of the zinc salt leads to free... contradiction is explained by Li occupation of interstitial sites where it acts as a donor [33] and compensates the acceptor contributions This may also be the reason why hydrothermally-grown LEO films were n-type instead of p- type, despite the presence of Na in the growth solution [34] Contrary to theoretical predictions, reports have shown that group V elements are more promising in achieving p- type doping Among . GROWTH OF ZINC OXIDE NANOSTRUCTURES AND FILMS AND P-DOPING OF FILMS IN AQUEOUS SOLUTION TAY CHUAN BENG B. Eng (Hons.), M. Eng A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF. 60 to 90°C. Growth precursors in aqueous solution generally consists of a zinc salt, such as zinc acetate, zinc nitrate or zinc chloride, and a base such as sodium hydroxide and aqueous ammonia aqueous solution using zinc acetate and ammonium hydroxide in detail. Regardless of the type of substrate used, the solubility of zinc (S Zn ), interface properties of the substrate and growth