CHARACTERIZATION OF MECHANICAL PROPERTIES OF TUNGSTEN NANOWIRES LI QI A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF ENGINEERING DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE 2005 Acknowledgement Acknowledgement I would like to express my deepest appreciation and thank to my supervisor, Associate Professor John Thong Thiam Leong, for my research opportunities at the Center for Integrated Circuit Reliability and Failure Analysis (CICFAR), National University of Singapore. With his critical reviews of my work, his guidance and his utmost friendship, I have been and continue to be part of a revolutionary area of nanoscience. I wish also to thank the staff of CICFAR, especially Mr Koo Chee Keong for his help during my project, and Mrs Ho Chiow Mooi and Mr Goh Thiam Pheng for valuable assistance given. A very special thanks to my colleagues at CICFAR, Mr Yeong Kuan Song, Mr You Guo Feng, Ms Law Bee Khuan Jaslyn, Mr Lin Soon Huat and others, for their timely assistance and inspiring discussions. Finally, I would like to especially thank my wife Yanyu for her support and encouragement during this work. i Contents Contents ACKNOWLEDGMENTS i CONTENTS ii SUMMARY iv LIST OF TABLES vi LIST OF FIGURES vii LIST OF ABBREVIATIONS xi CHAPTER 1 – INTRODUCTION 1 1.1. Properties of materials at nanoscales 2 1.2. Motivation 3 1.3. Objectives of the experiment 5 1.4. Outline of the thesis 6 CHAPTER 2 – LITERATURE REVIEW 7 2.1. Characterization methods with electron microscopy 7 2.1.1. Characterization methods with scanning electron microscopy 7 2.1.2. Characterization methods with transmission electron microscopy 10 Mechanical testing using SPM Techniques 14 2.2.1. bending nanowires with an AFM tip 14 2.2.2. Probing on protruding nanotubes and nanorods using the AFM 18 2.2.3. Tensile loading of carbon nanotubes 21 Characterization methods with the TEM/AFM 24 2.2. 2.3. ii Contents CHAPTER 3 – LATERAL FORCE MEASUREMENT 26 3.1. Principle of lateral force measurement 27 3.2. Derivation of Young’s modulus from measurement and cantilever calibration 30 3.3. Growth of nanowire AFM tip 38 3.4. Measurement procedures 43 3.5. Results and discussions 48 CHAPTER 4 – NANOWIRE LOADING WITH AFM TIP 53 4.1. Principle of bending test 55 4.2. Sample preparation 58 4.3. Measurement procedure 66 4.4. Results and discussions 69 4.4.1. Young’s modulus of as-gown nanowires 69 4.4.2. Effects of annealing on Young’s modulus of nanowires 74 4.4.3. Effects of contamination coating on the derivation of Young’s modulus 77 CHAPTER 5 – CONCLUSIONS 79 5.1. Summary 79 5.2. Future work 81 REFERENCES 83 APPENDIX I 88 iii Summary Summary It has long been established that the mechanical properties of solids are affected by the microstructure of the material. For nanoscale structures, their properties are additionally affected by the dimensions of the structure especially as we approach length scales comparable to the grain structure. This project aims to develop techniques to measure the Young’s modulus of nanocrystalline tungsten nanowires grown by a field-emission induced growth technique, with the capability of examining the microstructure of the nanowire in a transmission electron microscope (TEM). Two techniques based on the scanning probe microscope were considered. The first uses lateral force measurements in which a tungsten nanowire grown on a standard atomic force microscope (AFM) tip is bent against the corner of a step. The dimensions and microstructure of the nanowire could be partially observed in the TEM by mounting the entire cantilever in the microscope. The Young’s modulus obtained from the lateral force curve, which is less than 40 GPa, suggests a gross underestimation of the Young’s modulus of tungsten nanowire, when it is compared to the bulk tungsten value of 411 GPa. The underlying reason is postulated to be a weak point near the base of the nanowire as evidenced by frequent occurrence of breakages of nanowire tips at the base. Such non- uniformities in the growth of the nanowire could not be confirmed in the TEM due to obfuscation of the lower length of the nanowire by the AFM cantilever. iv Summary An alternative measurement method was developed in which the AFM tip is used to load the midpoint of a suspended nanowire that is grown and aligned across a slot. The nanowire was clamped at both ends by electron-beam induced contamination in the scanning electron microscope. The slotted substrate was fabricated by focused-ion beam milling and thinning of an anisotropically etched silicon wafer, in which micron-sized slots were subsequently milled. This sample configuration allows the entire length of the nanowire to be mounted and viewed in the TEM to determine the nanowire diameter and microstructure. The Young’s modulus of nanowires deduced from this method is comparable to that of bulk polycrystalline tungsten. The tungsten nanowire was annealed at 850 degree Celsius and the Young’s modulus is found to be slightly higher than before annealing. The corresponding change observed in the microstructure is an increase in the average grain size due to Oswald ripening. v List of Tables List of Tables Table 3.2.1 Specifications of CSC 38 AFM tip 34 Table 3.2.2 Results of AFM cantilever calibration 37 Table 3.5.1 Slope of force over distance for different nanowire AFM tip 50 Table 4.4.1 Young’s modulus of two different nanowires 72 vi List of Figures List of Figures Fig. 1.2.1 High-resolution TEM image of tungsten nanowire [Tay et al., 2004] 4 Fig. 1.2.2 TEM image of tungsten nanowire after annealing 5 Fig. 2.1.1 Setup for monitoring mechanical vibrations by Fujita et al. [2001] 9 Fig. 2.1.2 Setup for Vibrating Nanowires by Dikin et al. [2003] 9 Fig. 2.1.3 A silicon nanowire (a) stationary, (b) at the first harmonic resonance with the vibration plane parallel to the viewing direction, and (c) the resonance with the vibration plane perpendicular to the viewing direction [Wang, 2000] 12 Fig. 2.1.4 Electromechanical resonance of a ZnO nanobelt: (a) perpendicular to the viewing direction and (b) nearly parallel to the viewing direction [Wang , 2003] 13 Fig. 2.2.1 Loading nanowire beam using an AFM tip [Salvetat et al., 1999] 15 Fig. 2.2.2 A microtubule lying on a porous substrate under two nominal loading forces: (a) 100 pN, and (b) 150pN [Kis et al., 2002] 16 Fig. 2.2.3 SEM image of a Au nanowire mechanically fixed by ebeam induced deposition of Pt lines. The scale bar is 500 nm [Wu et al., 2005] 17 Fig. 2.2.4 Method to Probe Nanotubes and Nanorods: (a) randomly dispersed nanorods and nanotubes on a MoS2 substrate; (b) SiO2 pads to fix one end of beam; (c) location of nanobeam with AFM; (d) lateral probing with AFM tip; (e) forcedisplacement results by calculation [Wong et al., 1997] 18 Fig. 2.2.5 (a) SEM image of 2µm long cantilevers with a width of 150 nm and thickness of 50 nm (b) Principle of mechanical properties testing [Nilsson et al., 2003] 20 Fig. 2.2.6 Rope of single-wall nanotubes freely suspended over trench in silicon (a) before and (b) after being stressed past its elastic limit [Walters et al., 1997] 21 vii List of Figures Fig. 2.2.7 Extension of a nanotube stretched by two AFM tips [Yu et al., 2000] 23 Fig. 2.3.1 (a) CNT testing with buckling beam to measure both deformation and force of CNT. The left portion is a comb drive actuator, the central portion is a frame enclosing a long slender beam, and the right portion is the sample. (b) CNT testing with load sensor to measure the force of CNT, while the deformation is observed by atomic probe microscopy. The left portion is a comb drive actuator, central portion is the sample, and the right portion is load sensor. [Espinosa et al. 2002] 28 Fig. 3.1.1 Principle of lateral force microscopy measurement 28 Fig. 3.1.2 Change of θ with movement of nanowire across the step. (a) 29 Approach to the step, (b) bending over the step corner, (c) nanowire tracing the top of the step, (d) retrace on the top of the step, (e) departure from the step edge Fig. 3.2.1 Analysis of a suspended cantilever beam 31 Fig. 3.2.2 Deflection of Cantilever When Pushed Against (a) a Rigid sample and (b) a Flexible Spring Sheet by Ruan & Bhushan [1994] 35 Fig. 3.2.3 Force vs. distance curve when pushing tip to be calibrated toward (a) rigid substrate, and (b) cantilever with known spring constant. Both the curves show trace and retrace lines 36 Fig. 3.3.1 Schematic setup of field emission induced growth 40 Fig. 3.3.2 Experimental setup for field emission induced growth 40 Fig. 3.3.3 Forking of tungsten nanowire grown on AFM tip 41 Fig. 3.3.4 Field emission induced growth of tungsten nanowire on AFM tip 41 Fig. 3.3.5 TEM image of tungsten nanowire, a is the tungsten core of the nanowire, b is the original carbon coating formed from the precursor, and c is the contamination coating after the nanowire has been viewed in the SEM chamber 42 Fig. 3.4.1 Pushing the suspended end of nanowire against a sidewall of a silicon grating 43 viii List of Figures Fig. 3.4.2 Tip is raised to a position (a) too low, (b) just right and (c) too high 44 Fig. 3.4.3 Nanowire cannot be observed in a TEM sample holder 46 Fig. 3.4.4 Tip of nanowire can be observed with an additional support 46 Fig. 3.4.5 A tungsten nanowire grown on the top of an AFM tip 47 Fig. 3.4.6 The diameter of the nanowire on AFM tip is around 6 nm 47 Fig. 3.5.1 Lateral force curve when a nanowire knocks a step of silicon grating 49 Fig. 3.5.2 Lateral force curve obtained by knocking Nanowire against silicon grating step 49 Fig. 3.5.3 AFM topography scanned by (a) nanowire AFM tip and (b) AFM tip with broken nanowire. Z-scale – black level corresponds to 0nm, while white level corresponds to 538 nm for (a) and 527 nm for (b) 51 Fig. 4.1.1 Sketch of clamped beam model 55 Fig. 4.2.1 A tungsten nanowire grown on the substrate with circular trenches. Z-scale – black level corresponds to 0nm, while white level corresponds to 879nm 59 Fig. 4.2.2 A nanowire beam across one of the trenches. Z-scale – black level corresponds to 0nm, while white level corresponds to 247 nm 59 Fig. 4.2.3 A silicon substrate with etched windows 60 Fig. 4.2.4 Schematic of the cross section near the window edge 60 Fig. 4.2.5 Two-step FIB milling process 61 Fig. 4.2.6 Long slots penetrating the silicon substrate 62 Fig. 4.2.7 Schematic setup of field emission induced nanowire growth 63 Fig. 4.2.8 Directing the growth of a nanowire by moving the anode 63 Fig. 4.2.9 A tungsten nanowire grown across the slot on silicon substrate 64 ix List of Figures Fig. 4.3.1 Nanowire scanned with different loading forces 67 Fig. 4.3.2 AFM images of nanowire under relatively (a) small (b)large 68 loads Z-scale – black level corresponds to 0nm, while white level to 879 nm for (a) and 920 nm for (b) Fig. 4.4.1 AFM topography of tungsten nanowire. Z-scale – black level corresponds to 0nm, while white level corresponds to 214nm 69 Fig. 4.4.2 TEM image of tungsten nanowire with thick contamination coating 70 Fig. 4.4.3 The AFM tip penetrates into the gap and the AFM image is enlarged 70 Fig. 4.4.4 The convolution of the AFM tip with the nanowire diameter 71 Fig. 4.4.5 TEM image of tungsten nanowire without contamination coating 72 Fig. 4.4.6 Young’s modulus of nanowire before and after annealing 74 Fig. 4.4.7 Tungsten nanowire structure before annealing 76 Fig. 4.4.8 Tungsten nanowire structure after annealing 76 Fig. 4.4.9 Contamination cantilevers grown in the SEM 78 x List of Abbreviations List of Abbreviations NEMS Nano-electro-mechanical Systems SEM Scanning Electron Microscope / Microscopy SPM Scanning Probe Microscope STM Scanning Tunneling Microscope AFM Atomic Force Microscope / Microscopy FEIG Field Emission Induced Growth TEM Transmission Electron Microscope / Microscopy CNT Carbon Nanotube LFM Lateral Force Microscope / Microscopy xi [...]... convolution of the AFM tip with the nanowire diameter 71 Fig 4.4.5 TEM image of tungsten nanowire without contamination coating 72 Fig 4.4.6 Young’s modulus of nanowire before and after annealing 74 Fig 4.4.7 Tungsten nanowire structure before annealing 76 Fig 4.4.8 Tungsten nanowire structure after annealing 76 Fig 4.4.9 Contamination cantilevers grown in the SEM 78 x List of Abbreviations List of Abbreviations...List of Figures Fig 4.3.1 Nanowire scanned with different loading forces 67 Fig 4.3.2 AFM images of nanowire under relatively (a) small (b)large 68 loads Z-scale – black level corresponds to 0nm, while white level to 879 nm for (a) and 920 nm for (b) Fig 4.4.1 AFM topography of tungsten nanowire Z-scale – black level corresponds to 0nm, while white level corresponds to 214nm 69 Fig 4.4.2 TEM image of tungsten. .. structure before annealing 76 Fig 4.4.8 Tungsten nanowire structure after annealing 76 Fig 4.4.9 Contamination cantilevers grown in the SEM 78 x List of Abbreviations List of Abbreviations NEMS Nano-electro -mechanical Systems SEM Scanning Electron Microscope / Microscopy SPM Scanning Probe Microscope STM Scanning Tunneling Microscope AFM Atomic Force Microscope / Microscopy FEIG Field Emission Induced Growth ... Young’s modulus of two different nanowires 72 vi List of Figures List of Figures Fig 1.2.1 High-resolution TEM image of tungsten nanowire [Tay et al., 2004] Fig 1.2.2 TEM image of tungsten nanowire... LIST OF TABLES vi LIST OF FIGURES vii LIST OF ABBREVIATIONS xi CHAPTER – INTRODUCTION 1.1 Properties of materials at nanoscales 1.2 Motivation 1.3 Objectives of the experiment 1.4 Outline of the... underestimation of the Young’s modulus of tungsten nanowire, when it is compared to the bulk tungsten value of 411 GPa The underlying reason is postulated to be a weak point near the base of the nanowire