Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites
Transcription
Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites
Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites A THESIS SUBMITTED TO THE UNIVERSITY OF LONDON FOR THE DEGREE OF DOCTOR OF PHILOSOPHY by Fawad Inam School of Engineering and Materials Science Queen Mary, University of London London E1 4NS, United Kingdom May 2009 Chapter 8. Mechanical properties of ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Declaration I certify that the present work is prepared solely by myself during the course of my studies at the Queen Mary, University of London. It has not been submitted for a degree at this or any other University. Any words and/ or figures from the work of other people are fully acknowledged according to standard referencing. This thesis fully complies with the regulations set by the University of London and the Queen Mary, University of London. Fawad Inam May 2009 1 Chapter 8. Mechanical properties of ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Abstract The increasing availability of nanopowders and nanotubes combined with new processing techniques is enabling the development of new multifunctional materials. Carbon Nanotubes (CNTs) are one of the recently discovered allotropic forms of carbon. They have exceptional mechanical, electrical and thermal properties. The application of CNTs in the reinforcement of ceramic nanocomposites has not yet been fully investigated and is the subject of this study. Alumina is the main matrix used in this study. CNTs need to be de-agglomerated and homogeneously distributed in ceramic nanocomposites. Dimethylformamide (DMF) produces fine and stable CNT and alumina dispersions. All nanocomposites were sintered by Spark Plasma Sintering (SPS). Nanocomposites prepared using DMF dispersions showed better dispersions, higher electrical conductivity and mechanical properties as compared to those prepared using ethanol dispersions. The addition of CNTs or Carbon Black (CB) to alumina significantly aids its densification. The CNTs produce significant grain growth retardation. CNTs were found to be well preserved in alumina after being SPSed up to 1900 o C. Structural preservation of CNTs in ceramic nanocomposites depends on the nature of ceramic and SPS processing conditions. The electrical conductivity of alumina – CNT nanocomposites is four times higher as compared to alumina – CB nanocomposites due to the fibrous nature and high aspect ratio of CNTs. Alumina coated CNTs were used for better interfacial adhesion with the matrix. Oxidative resistance of CNTs was increased by coating them with alumina and by decreasing the grain boundary area in alumina – CNT nanocomposites. Coated and uncoated CNTs showed higher mechanical reinforcement in alumina nanocomposite as compared to CB. The future for ceramic – CNT nanocomposites is very bright, especially for applications associated with the electrical and thermal properties. Apart from a good understanding of nanocomposites, the commercial development of CNT based technologies heavily relies on the availability and price of CNTs. 2 Chapter 8. Mechanical properties of ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table of Contents Acknowledgement 8 List of Figures and Tables 9 List of Technical Abbreviations 21 Preface 22 Chapter 1. Chapter 2. Introduction to Carbon Nanotubes (CNTs) 24 1.1. Introduction 24 1.2. Structure of CNTs 25 1.3. History of CNTs 30 1.4. Synthesis of CNTs 31 1.5. Biocompatibility and toxicology of CNTs 33 1.6. Market overview of CNTs 34 1.7. Summary 36 Introduction to Ceramic – CNT Nanocomposites 37 2.1. Introduction 37 2.2. Fabrication of ceramic – CNT nanocomposites 37 2.2.1. Pre-consolidation processing 37 2.2.2. Consolidation techniques 38 2.3. Mechanical properties of ceramic – CNT nanocomposites 40 2.3.1. The never ending controversy 41 2.3.2. Toughening mechanisms 42 2.3.3. Effect of CNT in alumina nanocomposites 43 2.3.4. Effect of CNT in other ceramic nanocomposites 48 2.4. Electrical properties of ceramic – CNT nanocomposites 2.4.1. Percolation threshold in ceramic – CNT nanocomposites 54 55 3 Chapter 8. Table of Contents Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 2 continued… Chapter 3. 2.4.2. Effect of CNT on the electrical conductivity 56 2.5. Thermal properties of ceramic – CNT nanocomposites 61 2.6. Miscellaneous effects of CNTs on ceramics 64 2.7. Summary 66 Materials and Experimental Techniques 67 3.1. Introduction 67 3.2. Materials 67 3.2.1. Carbon nanotubes 67 3.2.2. Carbon black 69 3.2.3. Alumina 70 3.2.4. Other ceramic matrices 72 3.3. Experimental techniques Chapter 4. 74 3.3.1. Nano-particle size analyzer 74 3.3.2. Spark Plasma Sintering (SPS) 75 3.3.3. Density measurement 79 3.3.4. Electrical conductivity measurements 80 3.3.5. Vickers indentation 82 3.3.6. Microscopic analysis 83 3.3.7. Thermo Gravimetric Analysis (TGA) 84 Homogenisation of CNTs in Ceramics 85 4.1. Introduction 85 4.2. Experimental techniques 87 4.2.1. Colloidal dispersions and characterisations 87 4.2.2. Nanocomposite powder preparation 88 4.2.3. Spark Plasma Sintering (SPS) 89 4.2.4. Nanocomposite characterisations 90 4.3. Results and discussion 90 4.3.1. Natural drying patterns 90 4.3.2. Agglomerate size analysis 92 4 Chapter 8. Table of Contents Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 4 continued… Chapter 5. 4.3.3. Re-agglomeration behaviour 93 4.3.4. Microstructure of nanocomposites 96 4.3.5. Mechanical and electrical properties 97 4.4. Conclusions 99 Sintering of Ceramic – CNT Nanocomposites 101 5.1. Introduction 101 5.2. Experimental techniques 103 5.2.1. CNTs purification 103 5.2.2. Nanocomposite powder preparation 104 5.2.3. Spark Plasma Sintering (SPS) 104 5.2.4. Nanocomposite characterisations 105 5.3. Results and discussion Chapter 6. 105 5.3.1. SPS of alumina and alumina – CNT nanocomposites 105 5.3.2. Sintering behaviour and possible mechanisms 107 5.3.3. Grain growth modification 112 5.3.4. Co-sintering of grain size laminate 115 5.3.5. Sintering behaviour of alumina-purified CNTs 118 5.4. Conclusions 120 Structural Stability of CNTs in Ceramics 121 (Preservation Studies) 6.1. Introduction 121 6.2. Experimental techniques 125 6.2.1. Starting materials 125 6.2.2. Spark Plasma Sintering (SPS) 125 6.2.3. Material characterisations 127 6.2.4. Raman Spectroscopy 127 6.2.5. X-Ray Diffraction (XRD) analysis 127 6.3. Results and discussion 6.3.1. SPS of alumina and alumina – CNT nanocomposites 128 128 5 Chapter 8. Table of Contents Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 6 continued… Chapter 7. Chapter 8. Chapter 9. 6.3.2. SPS of other ceramics and their CNT nanocomposites 133 6.3.3. SPS of bulk CNTs 135 6.4. Conclusions 139 Electrical Properties of Ceramic – CNT Nanocomposites 140 7.1. Introduction 140 7.2. Experimental techniques 140 7.3. Results and discussion 142 7.3.1. CNT vs. carbon black 142 7.3.2. Electrical conductivity as a function of grain size 144 7.3.3. Electrical conductivity as a function of temperature 146 7.4. Conclusions 147 Oxidative Stability of Ceramic – CNT nanocomposites 149 8.1. Introduction 150 8.2. Experimental techniques 150 8.3. Results and discussion 150 8.3.1. Oxidation of CNTs 150 8.3.2. Oxidation of alumina coated CNTs 152 8.3.3. Oxidation of alumina – CNT nanocomposites 154 8.4. Conclusions 155 Mechanical Properties of Ceramic – CNT Nanocomposites 156 9.1. Introduction 156 9.2. Experimental techniques 157 9.2.1. Flexural bending 158 9.2.2. Thermal shock resistance during SPS (observation) 159 9.3. Results and discussion 9.3.1. Surface finish 160 160 6 Chapter 8. Table of Contents Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 9 continued… 9.3.2. Vickers hardness 162 9.3.3. Indentation Fracture Toughness (IFT) 164 9.3.4. Flexural modulus 167 9.3.5. Flexural strength 168 9.3.6. Thermal shock resistance (qualitative assessment) 171 9.4. Summary of mechanical properties 173 9.5. Conclusions 174 Conclusions 175 Future Work 177 Appendix A. Appendix B. Properties of DMF 179 A.1. Introduction 179 A.2. Chemistry of DMF 179 Future of Ceramic – CNT Nanocomposites 181 B.1. Introduction 181 B.2. Applications of ceramic – CNT nanocomposites 182 B.2.1. Conductive ceramic – CNT nanocomposites 182 B.2.2. Electric Discharge Machining (EDM) 183 B.2.3. Special purpose electrodes 183 B.2.4. Heating elements 184 B.2.5. Porous structures 185 B.2.6. Ceramic armour 185 B.2.7. Functionally Gradient Materials (FGMs) 186 B.2.8. Nanoceramics 186 B.3. Challenges in the development of ceramic – CNT nanocomposites 186 B.3.1. CNT related issues 187 B.3.2. Understanding of nanocomposites 188 B.4. Conclusions 188 Appendix C. Weibull statistical analysis 189 Appendix D. Recent publications, based on carbon nanotubes 197 References 198 7 Chapter 10. Future of Ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Acknowledgement It is never superfluous to say thanks where acknowledgement is due. I would like to express my sincere gratitude and appreciation to my supervisors Prof. Dr. Ton Peijs and Dr. Michael J. Reece for their continual and valuable guidance and encouragement. I am also very thankful to my third supervisor cum friend, Dr. Haixue Yan, for supporting me in processing my materials. Their ideas and infectious enthusiasm always energized me to contribute greatly in the world of nanocomposite materials. Grateful acknowledgement is made to Queen Mary, University of London (QMUL) and Nanoforce Technology Limited for financially supporting my PhD studentship. I am highly obliged to Dr. Zofia Luklinska, Dr. Monisha Phillips, Dr. Rory Wilson, Nima Roohpur and Dr. Daniel Doni (Imperial College, UK) and for their technical support in the characterisation part. Special thanks to Prof. Dr. Alan Weimer (University of Colorado, USA) for supplying coated Carbon Nanotubes (CNTs). I am indebted to my research mates, Emilano, Chris, Deng, Jamie, Rui, Peng Peng, Manuela, Nancy, Paola, Peppe, Pavin, Edwin, Tilen, Lilly, Jean, Jianmin, Aqif Bhai, Boon and Eric for helping me in every possible manner. I would also like to record my gratitude to Bill Godwin, Mick Willis, Sandra Wells, Victoria Wells, Jonathan Hills, Daniella Samos and Raj Chadha and Tara Demetriou for their time and humble support at various stages of my work. Words cannot express my deepest gratitude for my parents for their enduring love and support, which means so much to me. On personal level, I wish to thank my wife, Nageen, for her love, moral support, encouragement and tireless proof reading of this work. Surely, this thesis is dedicated to my family. Last but certainly not the least; I would like to acknowledge all my teachers who showed me the wisdom. No doubt, I am at this stage because of all of them! 8 Chapter 10. Future of Ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites List of Figures and Tables Chapter 1. Figure 1.1. Publications based on CNTs over the years. Source: Web of Science. Figure 1.2. Different types of CNTs based on their number of graphene cylinders: (a) capped Single Wall CNT (SWNT); and (b) open Multi Wall CNT (MWNT). Figure 1.3. Different forms of carbon fibre. Figure 1.4. Different types of CNTs based on their chirality: a) armchair; b) zigzag; and c) chiral. Figure 1.5. Schematic diagram showing how a hexagonal sheet of graphite is “rolled” to form a carbon nanotube [23]. Figure 1.6. Historical forms of hollow carbon filaments. Reported in: (a) 1952 [47]; and (b) 1976. Modified from [49]. Figure 1.7. Schematic diagrams of arc-discharge method [58]. Figure 1.8. Vertical CVD furnace for mass production. Re-drawn from [65]. Figure 1.9. Schematic diagram of laser ablation method [58]. Table 1.1 Properties of different engineering fibres [11-22]. Table 1.2 Different ways of producing CNTs. Table 1.3 Some of the major CNT suppliers. Chapter 2. Figure 2.1. Colloidal processing route for making ceramic – CNT nanocomposites. 9 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 2.2. Common sintering techniques for making ceramic – CNT nanocomposites: (a) hot pressing; and (b) Spark Plasma Sintering (SPS). Figure 2.3. Fracture toughness of alumina – carbon nanocomposites. (a) By Vickers indentation method [140]; and (b) by SEVNB method [127]. Both papers [127, 140] were published in the same journal. Figure 2.4. Different toughening mechanisms in amorphous alumina – CNT (outer diameter: ~51-56 nm) nanocomposite: a) crack deflection; b) crack bridging; c) CNT pullout; and d) CNT collapse in shear band. Figure modified from [192]. Figure 2.5. Schematic of hybrid microstructure design of alumina reinforced by MWNTs and SiC nanoparticles. Modified from [120]. Figure 2.6. CNTs at the grain boundary: (a) schematics; and (b) high-resolution TEM micrographs of an alumina – 10 vol% SWNT nanocomposite showing SWNT bundles at alumina grain boundaries, with schematic diagrams indicating corresponding SWNTs orientations. Figure modified from [195]. Figure 2.7. Electrical conductivity (measured at room temperature) of ceramic – CNT nanocomposites as a function of CNT content. Percolation threshold is: (a) 0.64 vol% for MgAl2O4. Modified from [146]; and (b) 0.79 vol% for Al2O3 [96]. Figure 2.8. The electrical conductivity of various representative materials at room temperature. Note the more than 13 orders of magnitude increase in conductivity of the alumina – 15 vol% SWNT nanocomposite compared to monolithic alumina [138]. Figure 2.9. Thermal conductivity of different ceramic – CNT nanocomposites as a function of temperature. (a) Barium titanate. Modified from [222]; and (b) titanium nitride. Modified from [226]. Table 2.1. Summary of the fracture toughness of ceramic – CNT nanocomposites. 10 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 2.2. Summary of the electrical conductivity of ceramic – CNT nanocomposites. Table 2.3. Summary of the room temperature thermal conductivity of ceramic – CNT nanocomposites. Chapter 3. Figure 3.1. Different types of MWNTs used: (a,b) NC 7000; (c) GraphiStrength C100; and (d,e†) alumina coated (50 ALD cycles) GraphiStrength C100. Figure 3.2. Carbon black powders: (a) Vulcan XC72; and (b) Printex L6. Figure 3.3. Crystal structure of α – alumina. Figure modified from [256]. Figure 3.4. Phase transitions in alumina. Redrawn from [255]. Figure 3.5. Alumina powder used in this study. Figure 3.6. Reduced titanium dioxide: (a) as-received form; and (b) after ball milling. Figure 3.7. Boron carbide: (a) as-received form; and (b) after ball milling. Figure 3.8. Boron nitride: (a) as received form; and (b) nano-sized flakes. Figure 3.9. Working principle of Zeta particle size analyzer. Modified from [260]. Figure 3.10. Typical DC-pulsed current cycles (used in this study). Figure 3.11. Effect of DC pulse on the density of alumina [267]. Figure 3.12. SPS facility by FCT Systeme, Germany: (a) SPS facility at Queen Mary, University of London, UK; and (b) SPS at 1800 oC [277]. Figure 3.13. Cross-sectional view of carbon die set. Figure 3.14. Electrical conductivity measurement setup: (a) room temperature jig; and (b) high-temperature characterisation chamber in the tubular furnace. Figure 3.15. A typical Vickers indent. Modified from [281]. Table 3.1. CNTs (synthesised by CVD method) used in this study. 11 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 3.2. Carbon black powders used in this study. Table 3.3. Properties of reduced titanium dioxide as per supplier. Table 3.4. Properties of boron carbide as per supplier. Table 3.5. Properties of boron nitride as per supplier. Chapter 4. Figure 4.1. Ultrasonication bath: (a) Apparatus with dispersion bottle; (b) shock waves in distilled water, top view of the apparatus; and (c) schematic diagram. Figure 4.2. Pre-sintering processing of ceramic – CNT nanocomposite powder. Figure 4.3. Natural drying patters formed during processing of nanocomposite powder: (a) alumina – 2 wt% CNT; (b) alumina – alumina coated 2 wt% CNT. CNTs were supplied by NanoDynamics, USA; (c) alumina – 2 wt% carbon black (Vulcan XC72); and (d) alumina – 2 wt% carbon black (Printex L6). Figure 4.4. Alumina – 2 wt% CNT nanocomposite powder after sieving. Individual CNT can be seen. CNTs were supplied by Nanocyl, Belgium. Figure 4.5. Agglomerate size analysis with respect to ultrasonication time in different solvents: (a) CNTs, supplied by Nanocyl, Belgium; and (b) alumina. Figure 4.6. Colloidal dispersion stability comparison after 1 h bath sonication and 5 minutes hand shaking. The diameter of the bottles is 25 mm. CNTs were supplied by Nanocyl, Belgium. Figure 4.7. Re-agglomeration behaviour in different solvents after 30 minutes of ultrasonication. CNTs were supplied by NanoDynamics, USA: (a,b) pristine CNT; (c) alumina coated CNT (25 ALD cycles); and (d) alumina coated CNTs (50 ALD cycles). 12 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 4.8. Figure 4.8. Fractured surfaces of sintered alumina – 2 wt% CNT samples, SPSed at 1200 ◦C/100 MPa/ 3 minutes: (a) CNTs dispersed in ethanol and (b) CNTs dispersed in DMF. Figure 4.9. Vickers hardness and indentation fracture toughness of alumina – 2 wt% CNT nanocomposites prepared using different solvents. Figure 4.10. Density and electrical conductivity of alumina – 2 wt% CNT. Table 4.1. Samples SPSed for this chapter#. Chapter 5. Figure 5.1. Purification of CNTs by acid treatment. Figure 5.2. Moving punch speed as the function of temperature during SPS for alumina and alumina – 5 wt% (~11.2 vol%) CNT (uncoated) nanocomposite. CNTs were supplied by Nanocyl, Belgium. Inset shows SPS program details. Figure 5.3. Moving punch speed as the function of temperature during SPS for alumina – 5 wt% CNT (uncoated) and alumina – 5 wt% CNT (coated, 50 ALD cycles) nanocomposite. Heating rate: 300 oC/minute. CNTs were supplied by Arkema, France. Figure 5.4. Rel. theoretical density as a function of sintering temperature for: (a) alumina and nanocomposites; and (b) reduced titania and nanocomposites. Figure 5.5. Bulk density of the sintered product as the function of homogeneity of CNTs in alumina matrix. SPS conditions: 100 MPa/ 3 minutes. CNT content: 5 wt%. Figure 5.6. Schematics of graphite die sets: (a) alumina before SPS; (b) current passing through graphite only during SPS; and (c) current passing through graphite and alumina – CNT nanocomposite compact during SPS. 13 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 5.7. Compressibility and compactibility analysis performed at room temperature by uni-axial pressing. Figure 5.8. Rel. theoretical density as a function of grain size for alumina and nanocomposites. Figure 5.9. FE-SEM images of fractured surfaces of sintered nanocomposites processed at 1800 oC/ 100 MPa/ 3 minutes: (a) alumina; (b) alumina – 2 wt% carbon black (Printex L6); (c) alumina – 2 wt% CNT (Nanocyl, Belgium); and (d) alumina – 2 wt% alumina coated CNT (50 ALD cycles, NanoDynamics, USA). Figure 5.10. Grain size refinement effect of: (a) CNTs; and (b) CNTs and CB. Figure 5.11. Grain size refinement in isothermal conditions (1400 oC/ 100 MPa). Figure 5.12. FE-SEM images of grain size laminate showing interface between alumina and 2 wt% CNT nanocomposite: (a) high magnification fractured surface; and (b) low magnification polished surface. Figure 5.13. FE-SEM images of polished grain size laminates: (a) alumina and 2 wt% CNT nanocomposite; and (b) alumina and 5 wt% CNT nanocomposite showing cracking at interface. Figure 5.14. FE-SEM images of different areas of polished grain size laminates: (a) alumina region, grain size: 43 μm; (b) 2 wt% CNT nanocomposite region, grain size: 1.81 μm; and (c) 5 wt% CNT nanocomposite region, grain size: 0.99 μm. Figure 5.15. Thermo gravimetric analysis (TGA) of CNTs, before and after acid purification treatment. Heating rate: 5 oC/minute. Figure 5.16. Platinum pan (diameter: ~10 mm) used for TGA: (a) empty pan before analysis, (b) as received-CNTs and (c) impurities left after oxidation. Table 5.1. Electrical conductivity measurements of graphite die sets. Table 5.2. Density measurements for CNTs and alumina – ceramic nanocomposites. 14 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 6. Figure 6.1. Schematics of different Raman vibration in CNT. Figure modified from [342, 343]. Figure 6.2. Diamond particles formed during SPS of CNTs [344, 345]. Figure 6.3. Raman spectra of CNTs, alumina and alumina – 5 wt% CNT nanocomposites. Alumina and nanocomposites were SPSed at 100 MPa for 3 minutes. Figure 6.4. Intensity ratio (ID/IG) for CNT and alumina – 5 wt% CNT nanocomposites. All nanocomposites were SPSed at 100 MPa for 3 minutes. For 1600 oC, refer figure 6.5. Figure 6.5. Raman spectra of CNTs and alumina – 5 wt% CNT nanocomposites. All nanocomposites were sintered at 1600 oC/ 100 MPa. Figure 6.6. Intensity ratio (ID/IG) for CNT and alumina – 5 wt% CNT nanocomposites. All nanocomposites were SPSed at 1600 oC/ 100 MPa. Figure 6.7. HR-TEM of alumina – 5 wt% CNT nanocomposites showing electron diffraction patterns of different areas. SPSed at 1900 oC/ 100 MPa/ 3 minutes. Figure 6.8. HR-TEM of alumina – 5 wt% CNT nanocomposites showing agglomerates of CNTs at the grain boundary. Figure 6.9. XRD analysis of CNTs, alumina (SPSed) and alumina – 5 wt% CNT nanocomposite (SPSed). SPSed at 1900 oC/ 100 MPa/ 3 minutes. Figure 6.10. FE-SEM image of boron carbide – 5 wt% CNT nanocomposite. SPSed at 2000 oC/ 80 MPa/ 20 minutes: a) at lower magnification; and b) at higher magnification. Figure 6.11. Raman spectra of CNTs, boron carbide (SPSed) and boron carbide – 5 wt% CNT nanocomposites (SPSed). SPSed at 80 MPa for 20 minutes. Figure 6.12. Intensity ratio (ID/IG) for CNT and boron carbide – 5 wt% CNT nanocomposites. All nanocomposites were SPSed at 80 MPa for 20 minutes. 15 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 6.13. FE-SEM analysis of boron nitride – 5 wt% CNTs (rel. theoretical density: ~97.5%). SPSed at 2175 oC/ 80 MPa/ 20 minutes: (a) at lower magnification; and (b) at higher magnification. Figure 6.14. FE-SEM images of bulk CNTs SPS processed at different temperatures. a) Pressed at room temperature; b) SPS processed at 1000 oC; c) at 1500 oC; and c) at 2000 oC. Figure 6.15. HR-TEM images of bulk CNTs: a) as received; b) SPS processed at 1000 oC/ 100 MPa/ 3 minutes. Figure 6.16. HR-TEM of CNT SPS processed at 2000 oC/ 100 MPa/ 20 minutes: inset a) nano-onion; and inset b) CNT after SPS, maintaining its aspect ratio. Figure 6.17. Raman spectra of CNTs SPS processed at different conditions. All samples were processed at 80 MPa for 20 minutes. Table 6.1. Samples SPS processed for this chapter. Chapter 7. Figure 7.1. Electrical conductivities of alumina – carbon nanocomposites. Figure 7.2. HR-TEM image of alumina – 5 wt% CNT nanocomposite: (a) CNTs around grains; and (b) percolating network highlighted. Figure 7.3. Voltage vs. current measured at room temperature for alumina – CNT nanocomposites. Both nanocomposites were SPSed at 1800 oC/ 100 MPa/ 3 minutes. Figure 7.4. Electrical conductivities of alumina – carbon nanocomposites as the function of grain size. The grain sizes were coarsened by using higher sintering temperatures. Figure 7.5. Electrical conductivities of alumina – 5 wt% nanocomposites as the function of SPS processing durations. CNTs were supplied by Nanocyl, Belgium. 16 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 7.6. Electrical conductivities of alumina – 5 wt% nanocomposites as the function of temperature. The grain sizes were coarsened by using longer processing durations. Heating rate: 2 oC/ minute. Table 7.1. Samples SPSed for this chapter#. Chapter 8. Figure 8.1. TGA of raw CNTs from different suppliers. Heating rate: 5 oC/ minute. Figure 8.2. TGA of uncoated CNTs and alumina coated CNTs. CNTs were supplied by NanoDynamics, USA. Heating rate: 5 oC/ minute. Figure 8.3. Platinum pan used for TGA: a) alumina coated (50 ALD cycles) CNTs before oxidation; and b) alumina nanotubes left after oxidation of coated CNTs. CNTs were supplied by NanoDynamics, USA. Figure 8.4. Alumina nanotube left after the oxidation of coated CNTs. CNTs were supplied by NanoDynamics, USA: a) at lower magnification; and b) at higher magnification. Figure 8.5. TGA of raw CNTs and SPSed alumina – 5 wt% uncoated CNT nanocomposites. CNTs were supplied by Nanocyl, Belgium. Heating rate: 5 oC/ minute. Table 8.1. Alumina – 5 wt% CNT nanocomposites SPSed for this chapter. Chapter 9. Figure 9.1. Special jig for improved 3 – point flexural testing. Sample thickness: 1.5 mm. Figure 9.2. Chamfered edges to avoid stress concentration points on corners. Sample: Fractured alumina surface. 17 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 9.3. Polished surfaces of alumina – CNT (Nanocyl, Belgium) nanocomposites containing: (a) no CNTs, thermally etched at 1500 oC for 10 minutes; (b) 2 wt% (~4.7 vol%) CNTs; (c) 5 wt% (~11.2 vol%) CNTs; and (d) 10 wt% (~21 vol%) CNTs. Figure 9.4. Polished surface of alumina – 10 wt% CNT nanocomposite. Arrows showing alumina grain ex-sites. CNTs were supplied by Nanocyl, Belgium. Figure 9.5. Alumina coated CNT (50 ALD cycles) encapsulated in an alumina grain of nanocomposite. CNTs were supplied by NanoDynamics, USA. Figure 9.6. Different types of fracture modes in alumina – 2 wt% CNT nanocomposites; (a) intragranular fracture mode in coated CNT nanocomposite (50 ALD cycles, NanoDynamics, USA); and (b) intergranular fracture mode in CNT nanocomposite (Nanocyl, Belgium). Figure 9.7. Vickers indent after applying 2.5 kg load in: (a) alumina; and (b) alumina – 5 wt% CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Figure 9.8. Sub-surface damage analysis after Vickers indentation. Optical micrograph (dark field image) showing sub-surface cracking in: (a) alumina; and (b) alumina – 5 wt% uncoated CNT (Nanocyl, Belgium) nanocomposite. Figure 9.9. Sub-surface damage analysis after Vickers indentation. Cross-section of: (a) alumina, showing major crack; and (b) alumina – 5 wt% uncoated CNT (Nanocyl, Belgium) nanocomposite, showing no major sub-surface damage. Figure 9.10. Low energy failure in 3 – point flexural testing: (a) Optical micrograph (side view) of alumina – 5 wt% uncoated CNT nanocomposites; and (b) schematics as per ASTM C1161-02c [368]. Figure 9.11. Fractured surfaces of alumina – uncoated CNT nanocomposites. Two tensile surfaces are mounted together. Fracture origin can be identified. CNT concentration: (a) 2 wt%; and (b) 5 wt%. 18 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 9.12. Effect of rapid cooling (300 oC/ minute) from 1800 oC after sintering. Disk diameter: 20 mm; (a) alumina; (b) alumina – 2 wt% CB nanocomposite; and (c) alumina – 2 wt% uncoated CNTs (Nanocyl, Belgium) nanocomposite. Table 9.1. Vickers hardness of alumina and alumina – carbon nanocomposites. Table 9.2. Indentation fracture toughness of different materials. Table 9.3. Flexural modulus of different materials. Table 9.4. Flexural strength of different materials. Table 9.5. Comparison of thermal shock resistance for different materials. Table 9.6. Summary of mechanical properties of alumina and alumina – carbon nanocomposites#. Appendix A. Figure A. Chemical formula of DMF, C3H7ON. Table A. Properties of DMF [377, 381, 382]. Appendix B. Figure B.1. Different sports goods based on non-ceramic – CNT nanocomposites [386-389]. Figure B.2. Heating element based on alumina – CNT nanocomposite. Sample diameter is 20 mm: (a) alumina – 5 wt% CNT (Nanocyl, Belgium); (b) demonstration; and (c) after oxidation, white area shows oxidation of CNTs. Figure B.3. CNT network on the grain boundaries of alumina grains. (a) Before oxidation; and (b) After oxidation of CNTs, percolating porosity. 19 List of Figures and Tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Appendix C. Figure C1. Weibull plot for data shown in table D1. Weibull modulus: 9.05. Figure C2. Weibull plot for data shown in table D2. Weibull modulus: 6.52. Figure C3. Weibull plot for data shown in table D3. Weibull modulus: 9.26. Figure C4. Weibull plot for data shown in table D4. Weibull modulus: 13.77. Figure C5. Weibull plot for data shown in table D6. Weibull modulus: 13.35. Figure C6. Weibull plot for data shown in table D5. Weibull modulus: 7.38. Table C. Weibull modulus (fracture strength) of different materials. Table C1. Weibull statistical analysis (fracture strength) for alumina. Table C2. Weibull statistical analysis (fracture strength) for alumina – 5 wt% CB nanocomposite. CNTs were supplied by NanoDynamics, USA. Table C3. Weibull statistical analysis (fracture strength) for alumina – 2 wt% uncoated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Table C4. Weibull statistical analysis (fracture strength) for alumina – 2 wt% coated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Table C5. Weibull statistical analysis (fracture strength) for alumina – 5 wt% uncoated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Table C6. Weibull statistical analysis (fracture strength) for alumina – 5 wt% coated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. 20 List of figures and tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites List of Technical Abbreviations AFM Atomic Force Microscopy ALD Atomic Layer Deposition CB Carbon Black CCE Carbon Ceramic Electrode CNF Carbon Nanofibre CNT Carbon Nanotube CVD Chemical Vapour Deposition CoMoCAT Co-Mo CATalyst DMSO DiMethyl SulfOxide DMF N,N-DiMethylFormamide DWNT Double Wall Carbon Nanotube DDS Drug Delivery System DLS Dynamic Light Scattering EDM Electric Discharge Machining FE-SEM Field Emission Scanning Electron Microscopy FGM Functionally Gradient Material GPS Gas Pressure Sintering HR-TEM High Resolution Transmission Electron Microscopy HiPCO HIP High-Pressure CO Hot Isostatic Pressing IET Impulse Excitation Technique IFT Indentation Fracture Toughness JCPDS Joint Committee on Powder Diffraction Standards MWNT/ MWCN Multi Wall Carbon Nanotube NMP N-MethylPyrrolidone NAFB Nano Agglomerate Fluidised Bed PCS Photon Correlation Spectroscopy PVD Physical Vapour Deposition RBM Radial Breathing Mode SHS Self-propagating High-temperature Synthesis SWNT/ SWCN Single Wall Carbon Nanotube SEVNB Single-Edge V-Notched Beam SPS Spark Plasma Sintering TZP Tetragonal Zirconia Polycrystals TBC Thermal Barrier Coating TGA Thermo Gravimetric Analysis XRD X-Ray Diffraction 21 List of figures and tables Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Preface The discovery of new material structures is shifting research interest from developing traditional materials like metals, ceramics, polymers and composites, to the more novel nanocomposites. “Nanocomposites” are not new. Nature has evolved ways to make many nanocomposites, such as seashell [1], spider silk [2, 3], ivory [4] and bone [5]. Bone shows ceramic-like properties and it is a natural hybrid nanocomposite of plate-shaped hydroxyapatite mineral particles in an oriented collagen polymer matrix, with an attractive balance of stiffness, toughness, and vibrational damping properties [5]. In the context of scientific knowledge, the term “nanocomposite” can be found in literature as early as 1986, when Roy et al. [6] prepared hybrid ceramic – metal nanocomposite by the sol-gel method. Nanocomposite materials can be defined as composites of more than one Gibbsian solid phase where at least one of the phases shows dimensions in the nanometre range [6]. The solid phases can exist either in amorphous, semi-crystalline or crystalline states [7]. CNT is one of the recently discovered allotropes of carbon. A comprehensive introduction to CNTs is given in chapter 1. Since 1991 [8], CNT reinforced nanocomposites have been the focus of intense global research. The main emphasis of these global research efforts has been on the synthesis of CNTs, developments in electronics and development of CNT reinforced polymer nanocomposites. The application of CNTs in the reinforcement of ceramic nanocomposites has not yet been fully investigated and is the subject of this thesis. A literature review on the development of ceramic – CNT nanocomposites is presented in chapter 2. Alumina, which is the most commonly used structural [9] and bio- ceramic [10], is used as a model ceramic for this study. 22 Preface Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The main topics of the current study are: 1. Homogenisation and de-agglomeration of CNTs in ceramic matrices (chapter 4). 2. Effect of CNTs on the sintering and the grain growth behaviour of different ceramics (chapter 5). 3. Survivability of CNTs sintered at high temperatures and their suitability as a ceramic reinforcement (chapter 6). 4. Electrical (chapter 7), chemical (chapter 8) and mechanical (chapter 9) properties of ceramic – CNT nanocomposites. 23 Preface Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 1. Introduction to Carbon Nanotubes (CNTs) 1.1. Introduction Carbon Nanotubes (CNTs) are the subject of intense global research. In 2008, CNTs research produced more than 20 publications per day (figure 1.1). Among all other nanotubes (BN, BC3, BC2N, C3N4, CN), CNTs appear to have the highest commercial interest. Medical science, electronics and composite technology are the main sectors that might benefit most from the properties of CNTs (table 1.1). CNTs have tensile strengths significantly higher than steel and carbon fibre, electrical conductivity similar to silver and platinum, an ability to carry higher current densities than copper, thermal conductivity better than diamond and a density much lower than aluminium. Figure 1.1. Publications based on CNTs over the years. Source: Web of Science. 24 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 1.1. Properties of different engineering fibres [11-22]. Young's Tensile Thermal Electrical Modulus Strength Conductivity Conductivity (GPa) (GPa) (W/mK) (S/m) 0.01-0.04 1300 20-63 1800-6000 10 3 15-150 400 2.4 10 10-7-10 Boron 2.6 2.3-2.5 400 4 38 10 -10 HM Carbon 1.95 1.7-8 380 2-7 105 10 -10 HS Carbon 1.75 1.7-8 230 3.4 24 10 -10 Glass 2.56 11 76 2 0.05-13 10 -10 Aramid 1.4 12 70-180 3.6 0.3 Alumina 3.4 3-20 300 2 5 10 Cellulose 1 1-5 80 2 0.035-0.06 10 Fibres/ Density Diameter Properties (g/cm3) (microns) CNTs 1.3-2.1 SiC 7 -3 -4 -2 4 5 4 5 -6 -12 -2 10 -10 -13 -3 CNTs appear in many different forms, i.e. short, long, thick, thin, single wall, multi wall, functionalised, open, capped, stacked, containing different structural defects, spirality and rolling structures. Each type has its own advantages and disadvantages for different applications. This chapter comprehensively covers the structure, history, synthesis, biocompatibility, toxicology and market overview of CNTs. 1.2. Structure of CNTs CNTs can be conceptually visualized as rolled graphene. The main cylindrical part of CNTs is like chicken wire, based on hexagons, whereas the end caps correspond to 25 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites half a fullerene molecule, made of hexagons and pentagons. CNTs can be classified in many ways: 1. Based on the number of graphene cylinders (x), CNTs (figure 1.2) can be termed as Single Wall CNTs (SWNT, x = 1), Double Wall CNTs (DWNT, x = 2) or Multi Wall CNTs (MWNT, x > 2). For MWNTs, the intershell spacing between two successive tubes is in the range from 0.344 nm to 0.36 nm and the carbon-carbon bond length is 0.144 nm [23]. To date, Shanov and Schulz [24] have synthesised the longest CNT of 18 mm, which is 900,000 times larger than its diameter. The aspect ratio of CNTs makes them very appealing for many applications. On the other end, Peng et al. [25] produced the thinnest SWNT of 0.33 nm. Another variation of MWNT is Carbon Nanofibre (CNF). CNF is relatively thick (diameter exceeding 100 nm) and long (4-5 μm) nanotube [26]. To date, most of the work has been focused on SWNTs and MWNTs. For diameter comparison, different forms of carbon fibre are shown in figure 1.3. (b) (a) (b) Figure 1.2. Different types of CNTs based on their number of graphene cylinders: (a) capped Single Wall CNT (SWNT); and (b) open Multi Wall CNT (MWNT). 26 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Carbon Nanofibre (CNF ~ 100 nm) Single Wall CNT (SWNT ~1.5 nm) Multi Wall CNT (MWNT ~10 nm) Carbon fibre/ human hair (~ 5 μm) 100 nm Figure 1.3. Different forms of carbon fibre. 2. Based on the symmetry of the carbon bonds, they can be chiral or achiral [27]. In an achiral CNT, the cylindrical structure follows mirror symmetry in both axes, longitudinal and transverse. In chiral CNT (also called helical CNT), the mirror symmetry is not obeyed. There are two types of achiral CNTs, i.e. arm-chair and zigzag (figure 1.4). Another interesting structure is cup-stacked CNT that comprises several truncated conical graphene layers [28]. The various ways to roll graphene into tubes are mathematically defined by the vector of helicity C (figure 1.5), and the angle of helicity θ, as follows [23, 27, 29]: C = na 1 + ma 2 Equation 1.1 m 3 And tan 1 m 2n Equation 1.2 27 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) (c) Figure 1.4. Different types of CNTs based on their chirality: a) armchair; b) zigzag; and c) chiral. Where n and m are integers, and can be grouped together to make lattice translational indices (n,m). a1 and a 2 are the vectors of the hexagonal lattice that corresponds to a section of the nanotube perpendicular to the nanotube axis (figure 1.5). The angle of helicity θ is the tilt angle of the hexagons with respect to the rolling axis and determines the spiral symmetry. Due to the six fold symmetry of the honeycomb lattice, the value of angle of helicity θ falls in the range 0o – 30o. When n = m, the nanotube is armchair type (θ = 0o); when m = 0, then it is of the “zigzag” type (θ = 0o); and when n= m, it is a “chiral”. The value of (n,m) determines the chirality of CNT and affects the optical, mechanical and electronic properties. CNT with |n − m| = 3q are metallic and those with |n − m| = 3q ± 1 are semiconducting (where q is an integer) [23]. 28 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 1.5. Schematic diagram showing how a hexagonal sheet of graphite is “rolled” to form a carbon nanotube [23]. 3. Based on the presence of defects or an attached foreign molecule, CNTs can be functionalised for better performance. It has been widely accepted among polymer researchers that the functionalisation of CNTs significantly increases the interfacial interactions [30-32], dispersion [33-38] and the mechanical properties [33, 39-43] of polymer – CNT nanocomposites. Apart from the attachment of chemical groups, functionalisation can also be carried out by introducing different types of defects such as: point defects like vacancies, topological defects caused by forming pentagons and heptagons, hybridisation defects due to functionalisation [44]. However, the effect of functionalised CNTs on ceramic matrices has not yet been considered. 29 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 1.3. History of CNTs There is a lot of controversy about the discovery of CNTs [45]. The first patent on the forming of carbon filaments was issued in 1889 [46]. The patent proposed the use of such filaments in light bulbs [45]. Editors of the journal “Carbon” [45] believe that Radushkevich and Lukyanovich [47] should be credited for the discovery of carbon filaments or CNTs. These filaments were hollow and had a nanometre-size diameter (figure 1.6a). The paper [47] was published in the Russian language in the Journal of Physical Chemistry of Russia (1952). Due to the cold war, access to Russian scientific publications by Western scientists was not easy at that time, and the use of the Russian language made comprehension difficult [45]. In 1959, Walker et al. [48] synthesised carbon filaments for the first time by thermal decomposition (now called chemical vapour decomposition) of hydrocarbons. SWNTs were first reported in 1976 by Oberlin et al. [49] (figure 1.6b), but due to the low resolution of their TEM, they could not be discovered. The magnification used was too low and graphene fringes could not be resolved so the number of walls could not be determined [45]. However, considering the diameter of this tube (~5 nm), it may have been SWNT or DWNT. The discovery of C60, Buckminsterfullerene, in 1985 by a group led by R.E. Smalley [50] motivated a number of researchers to work on the nano structures of carbon. In 1990, Kratschmer et al. [51] found that the soot produced by graphite electrodes contained C60 and other fullerenes. The discovery of CNTs was an accidental event. An electron microscopist, Iijima from NEC, Japan reported the observation of “Helical microtubules of graphitic carbon” as a by-product of fullerene research [8]. Initially, CNTs were called “graphitic microtubules” [8, 52, 53]. In 1992, these microtubules were referred to as “carbon nanotubes” for the first time by Ajayan and Iijima [54]. Later, SWNTs were reported in 1993 in two papers that were published in the same issue of NATURE, one by Iijima and Ichihashi [55], the other by Bethune et al. [56]. In 1995, Bandow and Saito [57] made the first ceramic – CNT nanocomposite by surrounding ultrafine particles of ZrC and V4C3 in graphene layers by the arc burning of metal carbide graphite composites. 30 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 1.6. Historical forms of hollow carbon filaments. Reported in: (a) 1952 [47]; and (b) 1976. Modified from [49]. 1.4. Synthesis of CNTs The quality of CNTs is dependent on their method of production. As-synthesised, CNTs contain carbonaceous impurities, typically amorphous carbon and graphite nanoparticles, as well as particles of the transition-metal catalyst [58, 59]. To eliminate these unwanted impurities, several physical and chemical methods are used, e.g. filtration, centrifugation or microfiltration of ultrasonic-treated solutions, chromatography, oxidation, selective reduction, and sublimation in vacuum at high temperature [60]. CNTs are currently synthesised by different techniques (figures 1.7-1.8). The arcdischarge method (figure 1.7) is the one by which CNTs were first produced and recognised [58]. The final properties of CNTs and their market value are much dependant on their method of synthesis. The basic and most common methods are summarised in table 1.2. 31 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 1.2. Different ways of producing CNTs. Inventor Summary Yields (%) Spacetime yield [61] Amorphous carbon [61] Carbon purity [61] Arcdischarge Iijima, 1991 [8] Graphitic anode evaporates to form CNT by a plasma via high current (figure 1.7). > 70 [62] Low High High CVD Yacaman et al., 1993 [53] CNTs grow from a carbon source by the catalysis action of metal particles (figure 1.8). 100 [63] High Low Medium Laser ablation Guo et al., 1995 [64] Direct laser vaporisation of graphite based target (figure 1.9). 70 - 90 [44] Low High High Method Furnace Furnace Catalyst and carbon source CNTs collection Figure 1.7. Schematic diagrams of Figure 1.8. Vertical CVD furnace for arc-discharge method [58]. mass production. Re-drawn from [65]. 32 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 1.9. Schematic diagram of laser ablation method [58]. The most crucial factor for the success of CNT based technologies is the production cost, and unless cost is not reduced to a competitive level, large-scale use of nanotubes is unlikely [66]. Some of the other techniques for the synthesis of CNTs are HiPCoTM [67, 68], CoMoCATTM [69], electrolytic method [70], solar furnace method [71], arc-plasma jet method [72], floating catalyst method [73], templating technique [74], diffusion flame synthesis [75], Nano Agglomerate Fluidised Bed (NAFB) reactor [76] and pyrolysis for the production of aligned CNTs [77]. The synthesis of CNTs and their actual and potential industrial uses have attracted the attention of many scientists worldwide, but relatively little attention has been paid, so far, to their potential detrimental effects on human health and the environment [78]. 1.5. Biocompatibility and toxicology of CNTs CNTs could find application as local Drug Delivery Systems (DDSs) and scaffolds to promote and guide bone tissue regeneration. Usage of CNTs as biomaterials is still at 33 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites a preliminary stage [79]. Many reports [80, 81] have confirmed that functionalised CNTs have greater biocompatibility as compared to pristine CNTs. Kam et al. [82] reported that the selective cancer cell destruction can be achieved by functionalisation of SWNT without harming normal cells. In contrast, studies [83] examining the toxicity of engineered nanomaterials in cell cultures and animals have shown that size, surface area, surface chemistry, solubility and possibly shape all play a role in determining the potential for nanomaterials to cause harm [84]. CNTs fall in the category of nanoparticles and all nanoparticles are considered toxic [85]. SWNTs have been shown to be acutely toxic [86-88] in cells, and direct evidence was presented by Porter et al. [86]. CNTs have the ability to cause oxidative stress and inflammation just like asbestos and quartz [87]. Some researchers [89, 90] have argued over the toxicity of CNTs. Pulskamp et al. [90] reported that, instead of CNTs, the metallic impurities associated with the CNTs are solely toxic [90]. These metallic impurities can be reduced by: optimizing the synthesis of CNTs [91]; and, purifying CNTs after production [59]. Until a clear toxicity appraisal is available, CNTs should be treated as a toxic material [92] and strict preventive and protective measures should be taken to limit inhalation exposure of CNTs in occupational settings [78, 87]. It should be noted that the harmful effects of CNTs are significantly reduced, once they are encased into composite materials. 1.6. Market overview of CNTs Nanotubes and fullerenes already represent a significant niche market with global revenue exceeding US$300 million in 2008. By 2015 it is predicted that the market will exceed US$4.6 billion [93]. It is expected that CNT based composites will have largest share of the market by a wide margin. 34 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites At present, the major producer of CNTs is the USA [22]. Recently, there has been a shift of production from the USA to Asia Pacific. It is predicted that in the near future, the major producer and supplier of all types of CNTs will be China and Korea respectively [22]. Some of the leading suppliers of CNTs with their selling prices are shown in table 1.3. The prices are much dependant on the quality and production methods of CNTs. Table 1.3. Some of the major CNT suppliers.* Selling price (£/gram) Supplier Country SES Research SWNT DWNT MWNT USA 50 - 492 52 - 125 27 - 340 Nano Lab USA 112 - 1000 250 - 257 1.2 - 80 MER USA 17.5 - 30 175 - 200 2.5 - 30 Helix USA 3 - 105 62 - 105 2 - 41 NanoAmor USA 15 - 175 27 - 53 1.3 - 38 Cheap Tubes USA 5 - 150 NA 0.1 – 17.5 Bucky USA 62 - 75 NA 27 - 45 Nanocs USA 125 - 1250 NA 40 - 1000 SWeNT USA 25 - 250 NA NA Carbolex USA 30 - 50 NA NA Carbon Solutions USA 25 - 200 NA NA Apex USA 1.4 – 2.8 NA NA Nanocyl Belgium 154 - 347 116 - 231 2.3 - 77 Arry Germany 25 - 199 42 - 199 0.27 -26 Bayer Germany NA NA From 0.05 Nano Thinx Greece 47.8 – 96.3 NA 5.4 – 25.4 Rosseter Cyprus NA NA 10 - 12 Nano Carblab Russia 30 - 800 125 - 750 NA Chengdu Org. Chem. China 22 - 35 10 - 28 0.75 - 2 Alpha Nano powder China 20 - 75 25 - 50 0.075 - 10 NanoTechnologies China 40 - 50 40 - 125 NA * Surveyed on March 16, 2009, NA = Not available 35 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Other major suppliers are Thomas Swan (UK), NanoDynamics (USA), Arkema (France), Sun Nanotech (China), Hanwha (Korea) and Mitsubishi (Japan). 1.7. Summary CNTs were comprehensively introduced in this chapter. Their unique and outstanding properties and high aspect ratio make them an ideal reinforcement for composite materials as compared to other engineering fibres. SWNTs are seamless rolled graphene sheets. MWNTs are concentric graphene sheets or in other words, Russian dolls made out of SWNTs. MWNTs were first reported in 1952. However, the time of discovery of CNTs is still quite controversial. CNTs are currently synthesised by different techniques and their quality relies on the method of production. With the increase in global production, the price of CNTs is decreasing. The development of CNT based technologies heavily relies on the availability and price of CNTs. CNTs are potentially toxic because of their size and shape. Health hazards associated with CNTs are greatly reduced once they are encapsulated into ceramics to create nanocomposites. 36 Chapter 1. Introduction to Carbon Nanotubes (CNTs) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 2. Introduction to Ceramic – CNT Nanocomposites 2.1. Introduction Since 1991, CNT reinforced nanocomposites have been the focus of intense global research. The excitement for CNTs originates from their unique and unrivalled properties, which were discussed in the previous chapter. A majority of the work and reviews has been presented on the development of CNT reinforced polymer nanocomposites. However, the application of CNTs in the reinforcement of ceramic nanocomposites has not yet been fully exploited and is subject of major on-going research efforts. This chapter reviews the development of the fabrication methods and properties of ceramic – CNT nanocomposites. Important key results and mechanisms relating to the fracture toughness, electrical and thermal conductivity of the ceramic – CNT nanocomposites are discussed. 2.2. Fabrication of ceramic – CNT nanocomposites 2.2.1. Pre-consolidation processing The homogenisation and de-agglomeration of CNTs is a wide topic that falls outside of the context of this chapter. Briefly speaking, when dispersing nano particles in a 37 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites suspension, there are three stabilisation mechanisms: electrostatic stabilisation, steric stabilisation, and electrosteric stabilisation [94, 95]. For ceramic – CNT nanocomposites, the most common dispersion method before consolidation is the colloidal processing route [26, 96-144]. The sol-gel method is a variation of colloidal processing and has been used in some studies [145, 146]. The common steps in colloidal processing are summarized in figure 2.1. Purification of CNTs in acids and/or solvents De-agglomeration and homogenisation of CNTs in solvents with/without surfactants Mixing homogenised solution of CNTs with ceramic powder Thorough extraction of CNTs Sieving of the powder to break agglomerates Nanocomposite powder ready for consolidation Figure 2.1. Colloidal processing route for making ceramic – CNT nanocomposites. 2.2.2. Consolidation techniques Hot-pressing [104, 109-112, 114, 116, 117-119, 128, 130, 131, 133, 137, 143, 146150] and SPS [26, 96-98, 101, 103, 107, 113, 122, 136, 139-142, 145, 151-159] are the most common consolidation techniques (figure 2.2) used for making ceramic – CNT nanocomposites. 38 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites In hot pressing (figure 2.2a) the simultaneous application of pressure and heat to a ‘green’ component is responsible for the sintering of ceramic powders. Heat can be applied directly (induction or resistance heating) or indirectly (convection or radiation). Pressure is applied statically or dynamically to the heated component [160]. A vacuum or controlled atmosphere can also be applied to prevent degradation of the ceramic powder during consolidation. (a) (b) Optical pyrometer Optical pyrometer DC pulse generator/ Heat source Heat source Vacuum chamber Vacuum chamber Figure 2.2. Common sintering techniques for making ceramic – CNT nanocomposites: (a) hot pressing; and (b) Spark Plasma Sintering (SPS). SPS (figure 2.2b) is a variation of hot-pressing that involves a different mechanism of heat transfer. A detailed introduction of the technique is presented in section 3.2.2. Compared to SPS, hot-pressing methods, involving longer durations and higher temperatures, which can damage CNTs in the nanocomposite, leading to a decrease or total loss of reinforcing effects without producing fully dense nanocomposites [97, 138, 140, 141, 147]. Other techniques used for consolidation are pressureless sintering [123], hot isostatic pressing (HIP) [107, 124], hot extrusion [161], tape casting [102], plasma spraying [162-165], high pressure reactive sintering [144], microwave 39 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites sintering [166], melt-infiltration reaction bonding [167], thermolysis [168], solvothermal synthesis [169], laser surface alloying [100] and spray pyrolysis [170]. The effect of CNTs on the sintering behaviour of ceramics is given in chapter 5. 2.3. Mechanical properties of ceramic – CNT nanocomposites CNTs are one of the strongest and stiffest fibres known so far, due to the intrinsic strength of the carbon–carbon sp2 bond. The experimental evidence of such strength can be found in a short communication by Hung et al. [171], where a single CNT failed after stretching 280% before breaking at high temperature. Lourie et al. estimated that the stress required for producing buckling or collapse of a CNT is approximately 100–150 GPa [172]. Earlier work by Treacy et al. [173] showed individual CNTs with Young’s moduli of more than 3 TPa (highest: 3.8 TPa). It has been predicted that CNTs have the highest modulus of all the different types of nanotubes (BN, BC3, BC2N, C3N4, CN etc.) [174]. To translate the superior properties of CNTs to ceramics, the processing route is critical. CNTs can provide multi-axial damage tolerance to ceramic nanocomposites [175] if they are homogeneously dispersed without agglomeration in ceramic matrices. Mechanical properties are very dependent on the agglomeration of CNTs in ceramic matrices [109, 128]. Another important factor is the interfacial compatibility. Good interfacial bonding is required to achieve load transfer across the CNT – matrix interface, a condition necessary for improving the mechanical properties of ceramic nanocomposites [176]. CNTs and the matrix are bonded by a combination of residual thermal stress and a diffusion layer, which makes the interface shear strength as high as ~1 GPa [118]. CNTs can be added to ceramic matrices up to a limit. High volume contents of CNTs in ceramics produces inhomogeneous dispersion of CNTs and more porosity because of trapped gases in agglomerates, which reduce the mechanical and electrical properties of the nanocomposite. This has been widely reported in the literature [112, 176, 177]. 40 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 2.3.1 The never ending controversy Zhan et al. [138, 140] SPSed alumina – SWNT nanocomposites and reported a threefold improvement in the fracture toughness at the expense of a 20% decrease in hardness (figure 2.3a). The group used the Vickers indentation method [178] to evaluate KIC. However, in the last 5 years, the effect of CNTs on the toughness of ceramic has become a very controversial topic [127, 179-183]. It is due to the fact, that, to date, no research group has been able to reproduce Zhan et al.’s [138, 140] results. Several authors [111, 120, 128, 156, 162, 180, 182-189] argued over the validity of the techniques used for determining the fracture toughness of ceramics and their CNT nanocomposites (figure 2.3b). In 2007, Quinn and Bradt [185] suggested abandoning the indentation fracture method for the evaluation of the fracture toughness, but the research community [117, 130, 131, 136, 155, 156, 165, 190, 191] still use the method. This is primarily due to the unavailability of an easy, effective, quick and economical alternative method. The indentation fracture toughness method is only useful for comparison purposes. The technique should not be used to determine KIC as it tends to overestimate the value [185]. (a) (b) Figure 2.3. Fracture toughness of alumina – carbon nanocomposites. (a) By Vickers indentation method [140]; and (b) by SEVNB method [127]. Both papers [127, 140] were published in the same journal. 41 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 2.3.2 Toughening mechanisms Fracture toughness evaluates the ability of a material to resist the crack propagation until fracture. Different toughening mechanisms (figure 2.4) are reported for ceramic – CNT nanocomposites: crack deflection at the CNT – matrix interface [130, 140, 141, 152, 192]; crack bridging by CNTs [110, 112, 137, 141, 153, 163, 168, 192]; CNT pullout on the fracture surfaces [104, 110, 112, 141, 152, 153, 168, 192] and CNT shear band collapse [192]. (a) (b) (c) (d) Figure 2.4. Different toughening mechanisms in amorphous alumina – CNT (outer diameter: ~51-56 nm) nanocomposite: a) crack deflection; b) crack bridging; c) CNT pullout; and d) CNT collapse in shear band. Figure modified from [192]. 42 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Residual stresses are also responsible for enhancing the toughness of ceramic – CNT nanocomposites [192], as they draw the transverse cracks into the compression stresses produced at the CNT – matrix interface, at which point the crack must either deflect or penetrate the CNT [192]. 2.3.3 Effect of CNT in alumina nanocomposites After Zhan et al.’s [140] paper on the fracture toughness of alumina – CNT nanocomposites (section 2.3.1), Wang et al. [127] reported that alumina – SWNT and the alumina – graphite prepared by SPS are not tough, but their contact-damage resistance properties are attractive for applications where contact loading is prevalent. CNTs induced anelasticity/ viscoelasticity in the alumina matrix that enhanced the contact-damage resistance of the nanocomposite material [127]. In the same report, they showed that alumina – SWNT has inferior fracture toughness as compared to alumina – graphite. They used the single-edge V-notch beam method and showed only a 3% improvement in the fracture toughness of alumina – SWNT nanocomposite over monolithic alumina. However, the dependence of the notch radius on the toughness was not considered, especially when the notch radius is hardly well controlled using a razor blade. One can argue on the reliability of such results. Only a really sharp notch can be regarded as identical to a pre-existing crack for the purpose of meaningful long crack toughness testing [179]. Mo et al. [145] dispersed CNTs in alumina by sol-gel process followed by SPS. They reported an enhancement in Vicker’s hardness (~7%) as well as fracture toughness (~10%). Fan et al. [118] prepared SWNTs in alumina by heterocoagulation of SWNTs into matrix grains. The addition of 1 wt % SWNTs in the alumina nanocomposite increased the fracture toughness by 103% and flexural strength by 20% higher compared to unreinforced alumina ceramics. In another report, Fan et al. [109] hot-pressed alumina – MWNT nanocomposites and reported an 80% improvement in the fracture toughness at the cost of a 4% decrease in bending 43 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites strength. Sun and Gao [128] reported that 1 wt% addition of MWNTs to an alumina matrix produced a 32% and 10% improvements in fracture toughness and bending strength respectively compared to monolithic alumina. It was observed that hotpressing for longer duration at high temperature (1500 oC) sharply decreased the mechanical properties as compared to lower temperatures (< 1400 oC) [128]. In another report, Sun et al. [101] found that the addition of only 0.1 wt % CNTs in an alumina nanocomposite increased the fracture toughness from 3.7 to 4.9 MPa√m (~32%). Chang et al. [149] reported the fracture toughness for a nanocomposite containing 10 vol% of the MWNT was 4.2 MPa√m; an improvement of 24 % when compared with that of the monolithic alumina. Siegel et al. [143] used MWNTs to toughen an alumina matrix and achieved a 24% increase in fracture toughness. Cha et al. [142] fabricated alumina – MWNT nanocomposites by molecular level mixing and SPS and reported a ~15% and 30% improvements in the hardness and fracture toughness of the nanocomposite over pristine alumina. Recently, Wei et al. [110] reported a fracture toughness increase of 79% and bending strength increase of 13% for 3 vol% CNTs – alumina nanocomposite as compared to that of pure nanocrystalline alumina. Zhu et al. [111] aligned CNTs in alumina using an electric field and reported an improvement in the mechanical properties along the normal direction. The mechanical properties of the hot pressed nanocomposites were characterised in two directions, with fracture toughness of 4.66 MPa√m and 3.65 MPa√m, and flexural strengths of 390 MPa and 191 MPa, respectively in the normal and the perpendicular directions. These results showed significant improvements when compared with the respective fracture toughness and flexural strengths of 3.78 MPa√m and 302 MPa for pristine alumina, and 4.09 MPa√m and 334 MPa for alumina nanocomposites filled with 2 wt% MWNTs prepared without the effect of an electric field [111]. Balani et al. [162, 165] used plasma spraying to fabricate alumina – CNT nanocomposite coating. Better densification was achieved due to strong interfacial 44 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites bonding between MWNTs and the alumina matrix, which enhanced the hardness [165] and the fracture toughness [162, 165] as compared to pristine alumina. Maensiri et al. [137] fabricated alumina – CNF for the first time. An improvement in the fracture toughness of 13% over pure alumina was reported in the alumina reinforced with 2.5 vol% CNF. However, hardness and bending strength of the nanocomposites were reduced with increasing volume fraction of CNF. Hirota et al. [26] SPSed alumina – CNF nanocomposite and reported a ~25% and ~5% improvement in the bending strength and the fracture toughness, respectively, compared with those of monolithic alumina. Peigney et al. [148] used a catalytic method that produced alumina – Fe nanocomposite powders that contained in situ grown MWNTs and SWNTs. They found that the fracture strength and the fracture toughness of alumina – CNT nanocomposites were higher than that of alumina, but still lower than those of the carbon free alumina – Fe nanocomposites. They reported big agglomerates of CNTs, which resulted in poor mechanical properties. In other reports [147, 193], the same group reported the same observations when they processed these hybrid nanocomposites by hot pressing. Thomson et al. [136] added 5 vol% SWNTs to 10 vol% Nb–alumina and reported inferior indentation fracture toughness (~12% less) and hardness (~ 16% less) as compared to 10 vol% Nb–alumina due to poor dispersion of SWNTs in the hybrid nanocomposite. Yoo et al. [159] SPSed alumina – iron – MWNT nanocomposite and found that the fracture strength was less than that of the alumina – iron nanocomposite. However, Ahmad and Pan [120] SPSed alumina – SiC – MWNT hybrid nanocomposites (figure 2.5) and reported an improvement of ~117% in fracture toughness and ~44% in bending strength, while the hardness remained unaffected. 45 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Intra/ Inter type Mutually redundant mechanisms CNT Grain boundary strengthening SiC Alumina Nano-fibre type SiC CNT Alumina Fibre toughening Crack bridging by CNTs and SiC Figure 2.5. Schematic of hybrid microstructure design of alumina reinforced by MWNTs and SiC nanoparticles. Modified from [120]. Other mechanical properties were also characterized to study the effect of CNT addition in ceramics. Solvas et al. [194] SPSed alumina – SWNT nanocomposite and found an improvement of two orders of magnitude in the creep-resistance compared to a pure alumina, which was attributed to partial blocking of grain-boundary sliding by SWNTs in the nanocomposites. CNTs prevent grain-boundary sliding because of the entangled network of agglomerates (figure 2.6). This may not be true for MWNTs because of their telescoping effect in tensile stress. 46 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 2.6. CNTs at the grain boundary: (a) schematics; and (b) high-resolution TEM micrographs of an alumina – 10 vol% SWNT nanocomposite showing SWNT bundles at alumina grain boundaries, with schematic diagrams indicating corresponding SWNTs orientations. Figure modified from [195]. An et al. [132] studied the influence of MWNT content on the tribological properties of alumina – MWNT nanocomposites. A 56% reduction in the wear loss and 30% improvement in the microhardness of the nanocomposite were observed as compared to pure alumina. Xia et al. [196] reported an ~80% reduction in the coefficient of friction for well-aligned alumina – MWNT nanocomposites as compared to monolithic alumina. For high wear resistance and low friction coefficients of the nanocomposites, highly aligned CNTs and thick CNTs are useful [196]. The higher content of CNTs in wear debris resulted in a lower coefficient of friction for an alumina – CNTs nanocomposite [102, 197]. On the other hand, Wasche and Klaffke [198] reported that graphite powder had no effect on the tribological properties of the ceramics. Lim et al. [102] prepared alumina – SWNT nanocomposite by tape casting followed by hot pressing and reported 71% less wear loss in a 12 wt% alumina – SWNT nanocomposite as compared to monolithic alumina. For the same composition 47 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites of nanocomposite, they also reported a 68% improvement in the fracture strength as compared to monolithic alumina. Lu [199] observed that during processing, the green strength of the alumina – CNT nanocomposites increased with the CNT content when CNTs are well separated. 2.3.4. Effect of CNT in other ceramic nanocomposites Apart from alumina, CNTs were used to reinforce other ceramic matrices as well [103, 107, 112, 115-117, 119, 129, 134, 141, 200, 201]. Manocha [202] reported improved mechanical properties for C-C nanocomposite by the inclusion of CNTs. Huang et al. [151] commented on the phase transformation and stresses in BaTiO3 – CNT nanocomposites. Due to the presence of a high compressive stress at the grain boundary and the CNT – BaTiO3 interface, an increase of 143% in the fracture toughness over CNT-free BaTiO3 was reported [151]. Balazsi et al. [107, 200] compared the mechanical properties of silicon nitride – MWNT nanocomposite, prepared by SPS and hot-pressing. Only a 2% improvement in the fracture toughness was reported for the nanocomposite processed by SPS. On the other hand, 13% and 17% decrements were reported in modulus and hardness respectively for the nanocomposite as compared to pristine silicon nitride. Hot pressing damaged the MWNTs and produced coarser grains as compared to SPS. In other reports by Balazsi et al. [115, 133], the bending strength and elastic modulus of MWNT – silicon nitride nanocomposites showed a considerable improvement compared to matrices with carbon fibre, carbon black or graphite additions. But the mechanical properties of silicon nitride – MWNT were inferior as compared to pure silicon nitride. The decrease of modulus and strength can be related to the lower densification rate. [133]. Hwang and Hwang [99] compared the hardness of SiO2 – CNT glass micro-rods SiO2 nanocomposites and SiO2 glass micro-rods – SiO2 nanocomposites. They found 48 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites that the hardness of nanocomposite discs containing ~80 wt% SiO2 – CNT glass micro-rods increased by ~210% and ~65%, as compared to pure SiO2 and SiO2 glass micro-rods – SiO2 nanocomposite respectively. CNTs improved the wetting and interface between glass micro-rods and the matrix [99]. Ning et al. [119] hot-pressed SiO2 – MWNT nanocomposite and reported that the bending strength and fracture toughness of SiO2 – MWNTs nanocomposite, compared with the monolithic SiO2, were enhanced by 65 and 100%, respectively. Gou et al. [141] incorporated MWNTs in SiO2 and found a 158% and 38% improvement in the fracture toughness and the Young’s modulus of 10 vol% MWNT – SiO2 as compared to pure SiO2. Ye et al. [112] dispersed MWNTs in a glassy-ceramic (barium aluminosilicate) and reported an enhancement in flexural strength (~192%) and fracture toughness (~143%) with up to 10 vol% of MWNT content in the nanocomposite. Boccaccini et al. [116] dispersed 10 wt% MWNTs in a glassy-ceramic (borosilicate) and reported a 11% improvement in the fracture toughness of the nanocomposite. However, the presence of the CNT aggregates and porosity weakened the material rather than reinforcing it, which resulted in a 41%, 21% and 14% reduction in fracture strength, hardness and elastic modulus respectively. Wang et al. [105, 144] prepared SiC – CNT and SiC – diamond – CNT nanocomposites by high pressure reactive sintering and reported superior mechanical properties as compared to CNT free matrices. Thostenson et al. [167] made SiC – Si – MWNT nanocomposite and reported that a very small amount of CNT content (<2.1 vol%) has no influence on the mechanical properties of the nanocomposite. Morisada et al. [103] mixed nanometre-sized SiC powders with 1–5 vol% SiC – coated MWNTs and SPSed them. An improvement of ~20% in microhardness and ~13% in the fracture toughness was reported for the nanocomposites as compared to monolithic SiC, due to the improved adhesion between the MWNTs and the SiC matrix by the SiC coating. Ma et al. [134] hot pressed nano SiC – MWNT nanocomposite and reported a 10% enhancement in the three-point bending strength and fracture toughness over monolithic SiC ceramic. Hirota et al. [26] SPSed SiC – CNF nanocomposite and reported a ~23% improvement in the bending strength 49 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites compared to monolithic SiC. An et al. [201] employed MWNTs to reinforce a polymer-derived ceramic (SiCN) matrix and reported a 60% and 52% improvement in the modulus and hardness respectively. Katsuda et al. [168] showed marginal improvement in the modulus and 70% improvement in the fracture toughness of a ternary Si-C-N matrix reinforced with MWNTs. Burghard et al. [203] showed marginal improvement in the hardness and ~36% improvement in the modulus of SiC-N matrix reinforced with SWNTs. Wang et al. [104, 204] hot pressed a mulliteMWNT nanocomposite. The addition of 5 vol% MWNTs led to a 10% increase in bending strength and 78% increase in fracture toughness, respectively, compared with the monolithic Mullite [104]. Many researchers have reported the effect of CNTs on zirconia matrices [108, 124, 125, 130, 131, 205-207]. Ionascu and Schaller [125] and Daraktchiev et al. [108] showed that introducing MWNTs into zirconia reduces the grain boundary sliding and consequently the mechanical loss at high temperatures, leading to better creep resistance. Recently, this was also observed for alumina – SWNT nanocomposite [194]. For comparison, Daraktchiev et al. [108] compared yttria-stabilized zirconia – MWNT nanocomposite with silica doped yttria-stabilized zirconia, and found that MWNT based nanocomposite possessed better creep resistance at high temperatures. Ionascu and Schaller [125] compared yttria-stabilized zirconia – MWNT nanocomposite with yttria-stabilized zirconia – SiC whiskers, and found that MWNT based nanocomposite possessed better creep resistance at high temperatures. Duszova et al. prepared yttria-stabilized zirconia – CNF nanocomposite [131] and yttriastabilized zirconia – CNT nanocomposite [130] by hot pressing. In both nanocomposites, due to significant porosity, the hardness and fracture toughness of the nanocomposites were lower than that of pure zirconia. Sun et al. [139] reported that the existence of the agglomerated CNTs in the grain boundary of ceramic grains and the weak bonding between CNT and zirconia matrix led to the failure in reinforcement. Ukai et al. [124] prepared yttria-stabilized zirconia – 1 wt% MWNT nanocomposite and reported a ~60% loss in fracture strength as compared to pristine yttria-stabilized zirconia. 50 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Iron aluminide (Fe3Al) nanocomposites reinforced with CNTs were fabricated by hot press consolidation for the first time by Pang et al. [208]. As compared to iron aluminide, the hardness, compressive yield strength, and bending strength hardness of nanocomposites were higher by 63%, 1%, and 5%, respectively. Zhang et al. [150] hot pressed titanium diboride – 5 wt% nickel – 0.5 wt% MWNT nanocomposite and showed a clear increase in the hardness (17%), bending strength (15%) and fracture toughness (60%) of the nanocomposite as compared to that of titanium diboride – 5 wt% nickel nanocomposite. Zhang et al. [153] reported that the WC – Co – 0.5 wt% CNTs nanocomposites possess superior hardness and toughness. The hardness was about 15% and fracture toughness was about 40% higher than that of the pure nanoWC – Co cermets consolidated processed under the same conditions. Singh et al. [135] reported remarkable improvements of 1069% and 1101% in elastic modulus and hardness, when 0.1 wt% MWNTs were added to a PMMA modified HA matrix for biomedical bone cement and implant applications. Meng et al. [117] hotpressed MWNT – HA nanocomposite and reported (50% and 28%) improvements in the fracture toughness and flexural strength respectively when the volume percentage of MWNTs reached 7%. Balani et al. [163] used plasma spraying to fabricate HA – CNT nanocomposite coatings and reported a fracture toughness improvement of ~56%. Chen et al. [100] prepared MWNT reinforced hydroxyapatite nanocomposite coating by laser surface alloying and reported a 41% and 21% improvement in the hardness and the modulus respectively as compared to a MWNT-free nanocomposite coating. The higher the value of the modulus for the coating, the stronger the mismatch between the coating and the living bone tissues. This slight increase in the modulus is very beneficial for bio applications of such coatings [100]. Kealley et al. [106] reported that the inclusion of increasing amounts of CNT material has a slight effect on the strain in the HA, suggesting that there may be a small amount of bonding between the two materials. Wang et al. [174] showed that by addition of as-received MWNTs and bio-mineralized MWNTs, the compressive strength of the calcium phosphate cement increased by 24% and 120%, respectively. 51 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The literature discussing the mechanical performance of CNT reinforced ceramic nanocomposites is at a relatively early stage. The reports of modest improvements in mechanical properties do not provide clear evidence linking the quantitative performance data to the actual mechanisms involved. In short, a brief summary of the CNT toughness effect on ceramics is shown in table 2.1. Table 2.1. Summary of the fracture toughness of ceramic – CNT nanocomposites. Research group † Year Ceramic matrix Processing method Reinforcement type Fracture toughness improvement (%) Zhan et al. [140] 2003 Alumina SPS SWNT 194† Guo et al. [141] 2007 Silica SPS MWNT 158† Ye et al. [112] 2006 Hot pressing MWNT 143 Huang et al. [151] 2005 Barium titanate SPS MWNT 143† Fan et al. [118] 2006 Alumina Hot pressing SWNT 103 Ning et al. [119] 2003 Silica Hot pressing MWNT 100† Fan et al. [109] 2006 Alumina Hot pressing MWNT 80 Wei et al. [110] 2008 Alumina Hot pressing MWNT 79 Wang et al. [104] 2007 Mullite Hot pressing MWNT 78 Wang et al. [144] 2006 Silicon carbide MWNT 75† Lei et al. [152] 2008 Alumina SPS MWNT 70† Katsuda et al. [168] 2006 Ternary Si-C-N Themolysis MWNT 70 Barium aluminosilicate glass High pressure reactive sintering Measured by Vickers indentation method 52 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 2.1. continued Research group Zhan et al. [150] 2006 Meng et al. [117] 2008 Balani et al. [162] 2007 Zhang et al. [153] 2005 Sun and Gao [128] 2005 Sun et al. [101] Ceramic matrix Titanium diboride + Processing method Reinforcement type Fracture toughness improvement (%) Hot pressing MWNT 60 Hydroxyapetite Hot pressing MWNT 50† Alumina Plasma spraying MWNT 43† SPS MWNT 40† Alumina Hot pressing MWNT 33† 2002 Alumina SPS MWNT 32† Cha et al. [142] 2005 Alumina SPS MWNT 30† Chang et al. [149] 2000 Alumina Hot pressing MWNT 24† Siegel et al. [143] 2001 Alumina Hot pressing MWNT 24† Zhu et al. [111] 2007 Alumina Hot pressing MWNT 23 Maensiri et al. [137] 2007 Alumina Hot pressing CNF 13† Morisada et al. [103] 2007 Silicon carbide SPS Ma et al. [134] 1998 Boccaccini et al. [116] Ukai et al. [124] Mo et al. [145] † Year 2007 2006 2005 nickel Tungsten carbide + cobalt Silicon carbide + boron carbide Borosilicate glass Yttria-stabilized zirconia Alumina SiC coated MWNT 13† Hot pressing MWNT 13 Hot pressing MWNT 11† MWNT 10† MWNT 10† Pressureless sintering + hot isostatic pressing SPS Measured by Vickers indentation method 53 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 2.1. continued Research group † Year Ceramic matrix Processing method Fracture Reinforcement type toughness improvement (%) Peigney et al. [148] 2000 Alumina + iron Hot pressing MWNT + SWNT 9 Thomson et al. [136] 2008 Alumina + niobium SPS SWNT 7 Hirota et al. [26] 2007 Alumina SPS CNF 5† Sun et al. [139] 2005 SPS MWNT 5† Wang et al. [127] 2004 Alumina SPS SWNT 3 Hirota et al. [157] 2007 Silicon carbide SPS CNF 2† Balazsi et al. [107] 2006 Silicon nitride SPS MWNT 2† Duszova et al. [131] 2008 Hot pressing CNF -10† Yttria-stabilized zirconia Yttria-stabilized zirconia Measured by Vickers indentation method 2.4. Electrical properties of ceramic – CNT nanocomposites The effect of the addition of CNT on the electrical conductivity of ceramics is much more significant than their effect on mechanical properties. Adding very little amount of CNT to a ceramic that is inherently an insulator can make it a good conductor. Carbon, a group IV element like Si, has a lot of potential in electronics. Prior to the discovery of nano fullerene structures of carbon, diamond (bandgap ~6 ev) and graphite (semi-metal) were the most commonly known forms [209]. The axial electrical conductivity of CNTs was found to be extremely high, reaching 2 x107 S/m [210], comparable to that of platinum i.e. 6.8 x107 S/m [211]. The electrical properties of CNTs vary with the type, diameters, chirality and defects in the structure. 54 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites SWNTs have much better electrical properties as compared to MWNTs due to their perfect structure as compared to MWNTs [97]. MWNTs can be metallic or semiconducting [158]. In semi-conducting category, MWNTs have a p-type conducting semiconducting behaviour, which can be converted to n-type semiconductor by doping alkali earth ions or annealing in a reducing atmosphere [158]. Based on the current-voltage characteristics, several authors [121, 122, 212] reported that the conductivity mechanisms involved in CNT are fluctuation-assisted tunnelling [213], whereas variable range hopping could be the conductivity mechanism as indicated in other reports [212]. 2.4.1. Percolation threshold in ceramic – CNT nanocomposites A very small volume content of CNTs, as low as 0.3 volume %, resulted in a 75% (measured by two-point method) reduction in the electrical resistivity of SiC – MWNTs nanocomposite [167]. Rul et al. [146] uniaxially hot pressed SWNT + MWNT – MgAl2O4 and reported a percolation threshold of 0.64 vol% (figure 2.7a). To date, this is the lowest percolation threshold reported. However, the waviness of CNTs is not considered in the scaling law of percolation theory. The effective (theoretical) percolation threshold may be lower than 0.64 vol% and the probable existence of paths with a low conductivity due to damage of some of the CNTs during hot pressing [146]. Ahmad et al. [96] showed that the electrical conductivity increased sharply as the content of MWNTs in alumina was close to percolation threshold of 0.79 vol% (figure 2.7b). This percolation threshold is 20 times smaller than that of micron scale two-phase random composites, and this low value is attributed to the enormous aspect ratio of MWNTs [96]. 55 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 0.64% 0.79% (a) (b) Figure 2.7. Electrical conductivity (measured at room temperature) of ceramic – CNT nanocomposites as a function of CNT content. Percolation threshold is: (a) 0.64 vol% for MgAl2O4. Modified from [146]; and (b) 0.79 vol% for Al2O3 [96]. Hirota et al. [26] SPSed alumina – CNF nanocomposite and reported a percolation threshold of 1.5 vol%. Shi and Liang [122, 126] SPSed 3 mol% yttria (Y) stabilized zirconia – MWNT and experimentally found a percolation threshold of 1.7 wt% (4.7 vol%), four times smaller than that of Ti3SiC2 – 3Y – TZP (Tetragonal Zirconia Polycrystals) nanocomposites [214]. 2.4.2. Effect of CNT on the electrical conductivity Zhan et al. [97, 138] SPSed alumina – SWNT nanocomposite and reported an increase of 13 orders of magnitude in the electrical conductivity over pure alumina. To date, this group has reported the highest electrical conductivity of 3345 S/m for 15 vol% SWNT – alumina nanocomposite (figure 2.8). 56 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 2.8. The electrical conductivity of various representative materials at room temperature. Note the more than 13 orders of magnitude increase in conductivity of the alumina – 15 vol% SWNT nanocomposite compared to monolithic alumina [138]. Modelling shows that the effective electrical conductivity of the MWNT – alumina nanocomposite in the aligned case is about three times of that in random distribution [215]. Peigney et al. [161] aligned CNTs in a Fe – Co – MgAl2O4 matrix by hightemperature extrusion and reported a 20 times greater conductivity in the direction of extrusion as compared to the transverse one. However, for MgO – Co – CNT nanocomposites, high-temperature extrusion damaged the structure of the CNT, which reduced the electrical conductivity from 20 S/m (un-extruded) to 10-6 S/m [161]. Recently, Zhu et al. [111] prepared alumina – 2 wt% MWNT nanocomposite by aligning MWNTs with AC electrical field. The electrical conductivities of these nanocomposites in both parallel and perpendicular directions to the MWNTs alignment were 6.2 x 10−2 Sm−1 and 6.8 x 10−9 Sm−1, respectively, compared with that of 4.5 x 10−15 Sm−1 for pristine alumina ceramics. Rul et al. [146] uniaxially hot pressed 24.5 vol% SWNT + MWNT – MgAl2O4 and reported an electrical 57 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites conductivity of 853 S/m. Flahaut et al. [147] uniaxially hot pressed 10 vol% CNT – 4.3 vol% Fe – Al2O3 and reported an electrical conductivity of 280-400 S/m. Thostenson et al. [167] fabricated SiC – Si – MWNT nanocomposites by melt infiltration of silicon and reported an electrical conductivity of 1538 S/m, a 96% decrease in electrical resistivity was observed for the ceramics with the highest CNT volume fraction of 2.1%. Balazsi et al. [133] prepared silicon nitride nanocomposites with MWNTs, carbon black and graphite by hot isostatic pressing. They reported a DC conductivity for a 5 wt% MWNT – silicon nitride nanocomposite of 85% less than that of 5 wt% carbon black – silicon nitride nanocomposite due to porosity and a poor interface at CNT/ matrix [133]. Tatami et al. [129] showed that hot pressing after Gas Pressure Sintering (GPS) of silicon nitride – MWNT nanocomposite improves the electrical conductivity from 30 to 79 S/m. Densification occurred at a much lower temperature using Y2O3 – Al2O3 – TiO2 – AlN as sintering aids during hot pressing [129]. Jian and Gao [123] reported a 44.7% and 11.5% enhancement in electrical conductivity when they added ~12 vol% CNT to TiN and Fe2N systems respectively. In another report, Jiang and Gao [169] reported that the addition of CNTs in a magnetite nanocomposite increased the electrical conductivity by about 32% from 1.9 to 2.5 S cm-1, compared with magnetite. Boccaccini et al. [116] dispersed 10 wt% MWNTs in glassy–ceramic (borosilicate) and reported an electrical conductivity of 7.7 S/m for the nanocomposite as compared to 10-3 S/m for monolithic Duran glass. Guo et al. [113] SPSed silica – MWNT nanocomposite and reported that the electrical conductivity of the system increased linearly with temperature from 5 to 300 K, showing a negative temperature coefficient of resistivity. Huang and Gao [114, 158] observed that the conductivity of BaTiO3 – MWNT decreased 47% when MWNTs are located within the grains. When CNTs are located 58 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites in the grain, they are not connected to the conductive network. The space between CNT and the grain is mainly responsible for the decreased electrical conductivity of the nanocomposite [114, 158]. Whereas in other case, MWNTs have been found to be effective in improving the electrical conductivity at the grain boundaries [96, 216] in ceramics. Zhan et al. [98] SPSed yttria stabilized zirconia – SWNT and alumina – yttria stabilized zirconia – SWNT nanocomposite and reported that the thermoelectric power of these nanocomposites increases with increasing temperature. But as compared to other thermoelectric materials, the electrical conductivity of the CNT/ ceramic nanocomposites is still low. Hirota et al. [26] SPSed alumina – 10 vol% CNF nanocomposite and reported an electrical conductivity of ~588 S/m. Duszova et al. [131] prepared yttria–stabilized zirconia – CNF nanocomposite and reported an electrical conductivity of 952 S/m, measured by the four point method. A brief summary of the room temperature electrical conductivity of ceramic – CNT nanocomposite is given in table 2.2. Table 2.2. Summary of the electrical conductivity of ceramic – CNT nanocomposites. Group Hirota et al. [26] Zhan et al. [97] Balazsi et al. [133] Shi and Liang [122] Guo et al. [113] Zhan et al. [98] Year Ceramic matrix Processing method Reinforcement type Electrical Percentage conductivity difference (S/m) (%) 2007 Alumina SPS CNF 588 1020 2003 Alumina SPS SWNT 3345 1017 2006 Silicon nitride Hot pressing MWNT 130 1016 SPS MWNT 65 1016 SPS MWNT 64.5 1016 SPS SWNT 55 1016 2006 2007 2006 Yttria stabilized zirconia Silica Yttria stabilized zirconia 59 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 2.2. continued. Group Tatami et al. [129] Zhu et al. [111] Duszova et al. [130] Rul et al. [146] Ukai et al. [124] Ahmad et al. [96] Boccaccini et al. [116] Thostenson et al. [167] Jian and Gao [123] Jian and Gao [169] Jian and Gao [123] Huang and Gao [114] Huang and Gao [158] Peigney et al. [161] Flahaut et al. [147] Flahaut et al. [147] Year Ceramic matrix 2005 Silicon nitride 2007 Alumina 2008 2004 2006 2006 2007 2005 Yttria stabilized zirconia Magnesium aluminate Electrical Percentage conductivity difference (S/m) (%) MWNT 79 1015 Hot pressing MWNT 0.062 1015 Hot pressing MWNT 952 1014 Hot pressing SWNT + MWNT 853 1014 MWNT 13.2 1013 Processing method Gas pressure sintering + hot pressing Reinforcement type Yttria stabilized Pressureless sintering zirconia + hot isostatic pressing Alumina SPS MWNT 7 105 Hot pressing MWNT 7.7 105 Melt infiltration MWNT 1538 1588 Pressureless sintering MWNT 73500 44.7 MWNT 2 32 Borosilicate glass Silicon carbide + silicon 2005 Titanium nitride 2003 Magnetite 2005 Iron nitride Pressureless sintering MWNT 88500 11.5 2004 Barium titanate Hot pressing MWNT 363 -47 2005 Barium titanate SPS MWNT 790 -64 Hot extrusion SWNT + MWNT 2000 - Hot pressing MWNT 400 - Hot pressing MWNT 180 - Solvothermal synthesis Magnesium 2002 aluminate + iron + cobalt 2000 Alumina + iron 2000 aluminate + iron Magnesium + cobalt 60 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 2.2. continued. Group Peigney et al. [161] Zhan et al. [98] Flahaut et al. [147] Yoo et al. [159] Peigney et al. [161] Year 2002 Ceramic matrix Alumina + iron Processing method Reinforcement type Electrical Percentage conductivity difference (S/m) (%) Hot extrusion SWNT + MWNT 158 - SPS SWNT 34 - Hot pressing MWNT 20 - SPS MWNT 8.47 x 10-4 - Hot extrusion SWNT + MWNT 1.8 x 10-6 - Alumina + yttria 2006 stabilized zirconia 2000 2006 2002 Magnesium oxide + cobalt Alumina + iron Magnesium oxide + cobalt 2.5. Thermal properties of ceramic – CNT nanocomposites Due to a large phonon mean free path in the strong carbon sp2 bond network of CNT walls, a very high theoretical value of thermal conductivity (~6600 W/mk) was predicted [217]. This high thermal conductivity is very much dependant on the structure, vacancies and defects associated with the individual CNTs [218]. Earlier results by Zhan et al. [138] showed that SWNTs decreased the thermal conductivity of the alumina nanocomposite as compared to pure alumina. Agglomeration and high thermal resistivity at the CNT – ceramic interface were mainly responsible for the poor behaviour. Calculations by Bakshi et al. [219] based on Xue’s model [220] for alumina – MWNT nanocomposite shows that thermal conductivity increases with the quality of the MWNTs dispersion. The strong tubetube coupling can decrease the thermal conductivity of MWNT bundles by an order of magnitude relative to individual tubes [138, 217]. Zhang et al. [221] reported that CNTs exhibited thermal conductivity as low as 4.2 W/m/K when SPSed into bulk 61 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites samples. The tube–tube interaction, entangled tubes, tube kinks and tube crosssectional heat diffusion were responsible for such a lower thermal conductivity. Huang et al. [222] reported a decrease in the thermal conductivity after adding MWNTs due to an interfacial thermal barrier between CNTs and the BaTiO3 matrix (figure 2.9a). The lattice mismatch between CNTs and BaTiO3 may account for this thermal barrier [222]. For other ceramics, the reduced thermal conductivity is useful for the applications of Thermal Barrier Coatings (TBCs) where they are used to protect critical components in the hot sections of gas turbine engines [223]. In ceramic – CNT nanocomposites, the thermal conductivity is affected by many factors like the crystallite boundaries, porosity, CNT content, interphase boundaries, and phase content [219]. Sivakumar et al. [224] reported that the thermal conductivity of the silica – MWNT nanocomposites was improved due to incorporation of MWCNT compared with a pure silica matrix. The improvement in thermal conductivity was enhanced with increasing amounts of MWNTs [224]. Chin et al. [225] fabricated a ceramic – CNT catalyst and reported that the novel structure possessed superior thermal conductivity. The CNTs allowed efficient heat removal from catalytic active sites during exothermic reaction. Jiang and Gao [226] found that the thermal conductivity increased with increasing MWNT content and temperature for a TiN – MWNT nanocomposite. In the presence of 5 wt% MWNTs, there was a 97% and 11% enhancement in the thermal conductivity at 430 oC and 27 oC, respectively, compared with that of TiN [226]. Apart from CNTs, Nakamatsu et al. [227] found that carbon coating on aluminium nitride powder increased the thermal conductivity of the final sintered product by 22%. 62 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 0 wt% CNT 0.1 wt% CNT 1 wt% CNT 5 wt% CNT 0 wt% CNT 0.5 wt% CNT 1 wt% CNT 3 wt% CNT (a) (b) Figure 2.9. Thermal conductivity of different ceramic – CNT nanocomposites as a function of temperature. (a) Barium titanate. Modified from [222]; and (b) titanium nitride. Modified from [226]. The thermal conductivity of the alumina – CNT nanocomposites [138, 219] decreased with increasing temperature due to the dominant effect of Umklapp scattering (phonon-phonon scattering) in reducing phonon mean-free path length [138]. This trend has also been reported for the thermal conductivity of CNTs [217, 221], diamond and graphite [217]. However, in titanium nitride – MWNT nanocomposite [226] phonons dominate thermal transport at all temperatures, which resulted in higher thermal conductivity at high temperatures (figure 2.9b). Ning et al. [228] showed that at 650 oC, the thermal conductivity of a 10 vol% CNT – silica nanocomposites was ~21% more as compared to that of silica. The exact mechanism for increased thermal conductivity at high temperatures was not reported. A brief summary of the room temperature thermal conductivity of ceramic – CNT nanocomposite is given in table 3.3. 63 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 2.3. Summary of the room temperature thermal conductivity of ceramic – CNT nanocomposites. Processing Reinforcement Percentage method type difference (%) Silica SPS MWNT 65 2008 Alumina Plasma spraying MWNT 15 Jiang and Gao [226] 2008 Titanium nitride SPS MWNT 11 Huang et al. [222] 2005 Barium titanate SPS MWNT -3 Zhan et al. [219] 2004 Alumina SPS SWNT -73 Thostenson et al. [167] 2005 Melt infiltration MWNT -9 Group Year Ceramic matrix Sivakumar et al. [224] 2007 Bakshi et al. [219] Silicon carbide + silicon 2.6. Miscellaneous effects of CNTs on ceramics Luo et al. [205] revealed that the addition of CNTs during the synthesis of ZrO2 assisted the transition from monoclinic ZrO2 to cubic ZrO2. CNTs stabilized cubic ZrO2 and prevented agglomeration of ZrO2 nanoparticles as well [205]. Balani et al. [163] found that CNTs improved the crystalline content of HA by 27%, which is beneficial for the bio applications of hydroxyapatite coatings [229]. Shi and Liang [122] reported that the dielectric constant was greatly increased when the MWCNT concentration in yttria stabilized zirconia was close to the percolation 64 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites threshold, which was attributed to dielectric relaxation, the space charge polarisation effect, and the percolation effect. Dou et al. [230, 231] and Kovac et al. [232] sintered magnesium boride – CNT nanocomposites and reported superior magnetic properties of the superconducting wires. Sun et al. [233] made iron oxide – CNT nanocomposites that showed ferromagnetic and super magnetic behaviours, when processed using different processing conditions. Huang and Gao [114] observed that BaTiO3 semiconductor transformed from n-type to p-type after doping with 0.1 wt% of CNTs due to the formation of a Schottky barrier constructed at the CNT–matrix contact. Huang and Gao [158] fabricated bi-layer ceramics, stacked by one layer of MWNT free BaTiO3 and another layer of 1 wt % MWNT – BaTiO3 nanocomposite that showed excellent rectification properties. For improved photovoltaic properties, Cao et al. [234] coated CdS and Lee et al. [235] coated TiO2 on MWNTs. Different authors found that ceramic coatings such as TiC [236], BN [237, 238], NiO [239] and BaO/ SrO [240] on the surfaces of CNTs improve the field emission properties of CNTs as compared to uncoated CNTs. However, a dielectric MgO coating on CNT decreased the strength of electric field on the CNT surface and increased the tunnel barrier for field emission [241]. Other ceramics like SiO2 [242], SnO2 [243], TiO2 [244] were coated on CNTs for the development of functional nanocomposites. Espinosa et al. [245] showed that the addition of a small quantity of MWNTs to different ceramics can significantly improve the detection capability of metal oxidebased sensors at low operating temperatures. Ma et al. [246] and Li et al. [247] showed that tungsten carbide – CNT nanocomposite has improved electro catalytic activity compared to tungsten carbide. Addition of MWNT improves the electrochemical activity of LiNi0.7Co0.3O2 [248] and LiCoO2 [249], which are widely used in batteries. 65 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 2.7. Summary Application of CNTs as a reinforcement in ceramic nanocomposites has not yet been fully exploited and is the subject of major on-going research efforts. This chapter reviewed recent studies conducted on the development of ceramic – CNT nanocomposites. CNTs have been demonstrated to increase the mechanical, electrical and thermal properties of the ceramic matrices. However, a significant degree of discrepancy still exists, especially with regard to the mechanical properties of these materials. The effect of CNTs on the electrical properties of the ceramics is the most promising area. An improved understanding of ceramic nanocomposites and breakthroughs in materials processing are need to be achieved for the successful placement of ceramic – CNT nanocomposites in high technology applications. 66 Chapter 2. Introduction to Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 3. Materials and Experimental Techniques 3.1. Introduction This chapter provides detailed information about the materials, equipments and experimental procedures used in this study. 3.2. Materials 3.2.1. Carbon Nanotubes A detailed introduction to CNTs was covered in Chapter 1. Pristine and alumina coated MWNTs were used in this study (table 3.1, figure 3.1). Table 3.1. CNTs (synthesized by CVD method) used in this study. Primary source / Nanocyl Arkema NanoDynamics Properties (Belgium) (France) (USA) Commercial grade NC 7000 GraphiStrength C100 NDCNT (E-1005-03) Density (g/cm3) 1.66 1.82 3.04 1.72 2.70 2.96 Average no. of graphitic shells 10 - 15 5 - 15 5 - 15 10 - 30 10 - 30 10 - 30 % wt carbon 90 > 90 ~46 > 92 ~ 56 ~ 45 Average diameter (nm) 9.5 10 - 15 20 - 45 10 - 30 15 - 35 20 - 40 Average length (µm) 1.5 0.1 - 10 0.1 - 10 1-2 1-2 1-2 ALD* coating cycles 0 0 50 0 25 50 * ALD was performed in Prof. A. Weimer’s lab, University of Colorado, USA. 67 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) (c) (d) (e) Figure 3.1. Different types of MWNTs used: (a,b) NC 7000; (c) GraphiStrength C100; and (d,e†) alumina coated (50 ALD cycles) GraphiStrength C100. Atomic Layer Deposition (ALD) is an ideal method for depositing thin films on high aspect ratio materials as it is independent of line of sight and self-limiting [250]. Sequential surface chemical reactions deposit highly conformal films with precise control at the atomic scale [250]. The method has been shown to be a viable technique to deposit a coating on a single CNT without adversely affecting its inherent † Figure 3.1e was provided by Prof. A. Weimer’s group 68 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites properties [251, 252]. Alumina coating on CNTs was done by elsewhere by ALD method [250]. 3.2.2. Carbon black Carbon-derived powders and particles comprise a family of synthetic materials, known under the generic term of carbon black, made by burning hydrocarbons in air. Carbon black (CB) are aggregates of graphite micro crystals, each only a few unit cells in size [253]. Some other names for CB are acetylene black, channel black, furnace black, lamp black, lampblack and thermal black [254]. For comparative studies, we used different types of carbon black powders (figure 3.2) summarised in table 3.2. (a) (b) Figure 3.2 Carbon black powders: (a) Vulcan XC72; and (b) Printex L6. Table 3.2. Carbon black powders used in this study. Primary source / Properties Cabot (USA) Degussa (Germany) Commercial grade Vulcan XC72 Printex L6 Density (g/cm3) 2.01 1.80 Average particle size (nm) 13 18 Particle shape Very spherical Roughly spherical 69 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.2.3. Alumina Alumina was used as the main ceramic matrix because of its industrial significance. Alumina is the most commonly used structural [9] and bio- ceramic [10]. It exists in a number of crystalline phases (polymorphs). The most important, and common, polymorphs are denoted α, γ, θ, and κ. These phases of alumina are unique for different applications. For example, the α and κ phases are widely used as wear resistant coatings due to their high hardness and thermal stability, while γ- and θalumina are more suited for catalytic applications due to their high surface energies, leading to larger active surface areas for catalytic reactions [255]. In addition to these, there are more than twenty other crystalline phases of alumina [255]. In this study α – alumina (figure 3.3) is the main phase that existed in the final sintered product. The α – alumina is also known as corundum (the name comes from the naturally occurring mineral corundum. It is used not only in materials science, but exists also as gemstones. Ruby is α – alumina doped with chromium, whereas sapphire is α – alumina doped with iron and titanium [255]. α – alumina has a variation of rhombohedral structure, whereas γ – alumina has a defected cubic spinel structure [255-256]. The thermodynamic stability of α-alumina makes it the most suited phase for use in many industrial applications. The corundum structure can be visualized as layers of hexagonal close-packed oxygen atoms with small Al atoms in two-thirds of the octahedrally coordinated holes between the oxygen atoms [255, 257]. The atomic positions consist of 12 aluminium atoms and 18 oxygen atoms. The unit cell dimensions are: a = b = 4.7588 A° and c = 12.992 A° [255, 257]. The α structure is thermodynamically stable at all temperatures up to its melting point at 2051 °C (figure 3.4). However, the metastable phases (e.g., γ and θ) still appear frequently in alumina growth studies [255]. The alumina powder used in this study was commercially available “544833 aluminium oxide” nanopowder from Sigma-Aldrich, UK (figure 3.5). As supplied by 70 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites the supplier, the main features of this product are: γ phase; particle size: < 50 nm; surface area 35-43 m2/g; melting point 2040 oC; and density 3.97 g/cm3. 0.4759nm c 1.2992nm 90o 120o Aluminium Oxygen Stacking order Unit cell (variation of rhombohedral crystal structure). Rhombohedras combine to form hexagon. Figure 3.3. Crystal structure of α – alumina. Figure modified from [256]. Figure 3.4. Phase transitions in alumina. Figure 3.5. Alumina powder used in Redrawn from [255]. this study. 71 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.2.4. Other ceramic matrices Some other ceramics studied for comparison in Chapter 4 and Chapter 6 are: 3.2.4a. Reduced titanium dioxide (Red. titania) Reduced titanium dioxide (Red. Titania/ TiO2) was supplied by Atraverda Inc., UK. This powder contained coalesced particles, which formed strong big agglomerates (figure 3.6a). After 24 hours of ball milling in ethanol, these micron-sized particles were broken up (figure 3.6b). As per supplier, the main features of this powder are given in table 3.3. This material is used as electrode because of its unique combination of metallic-like electrical conductivity along with the characteristic high corrosion resistance of ceramics [258]. (a) (b) Figure 3.6. Reduced titanium dioxide: (a) as-received form; and (b) after ball milling. Table 3.3. Properties of reduced titanium dioxide as per supplier. Surface area (m2/g) Grade Density (g/cm3) Melting point (oC) Ebonex 4.29 1800 † Measured by BET surface analyzer Mean particle size (µm) Before ball milling After ball milling† Before ball milling After ball milling* 0.32 – 0.34 1.6 – 2.1 52 24.5 * Measured by Nano-particle size analyzer (section 3.3.1) 72 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.2.4b. Boron carbide (B4C) Boron carbide was provided by H.C. Starck, Germany. The powder was ball milled for 24 hours (figure 3.7) in ethanol before mixing with CNTs. As per supplier, the main features of this powder are given in table 3.4. Typical applications of boron carbide are abrasive grit, polishing, lapping, light weight armour, wear resistant engineering components, sintering additives for different ceramics, neutron shielding [256]. Table 3.4. Properties of boron carbide as per supplier. Surface area (m2/g) Grade Density (g/cm3) B:C ratio HT 03 2.52 3.8 – 4.2 † Measured by BET surface analyzer Mean particle size (µm) Before ball milling After ball milling† Before ball milling After ball milling* 2.5 – 4 4 – 4.6 4.5 3.8 * Measured by Nano-particle size analyzer (section 3.3.1) (a) (b) Figure 3.7. Boron carbide: (a) as-received form; and (b) after ball milling. 3.2.4c. Boron nitride (BN) Boron nitride was supplied by H.C. Starck, Germany. As per supplier, the main crystalline phase of the raw powder is hexagonal. Details are given in table 3.5. The 73 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites raw powder consists of soft nano-sized flakes (figure 3.8). Typical applications of this material are solid lubricant for high temperature applications, mould release, raw material for cubic-BN (second hardest material known), evaporation boats, thermally conductive for polymers, refractories etc [259]. Table 3.5. Properties of boron nitride as per supplier. Grade C Density 3 (g/cm ) 2.1 % B2O3 Surface area (m2/g) 5-8 10 - 20 (a) (b) Figure 3.8. Boron nitride: (a) as received form; and (b) nano-sized flakes. 3.3. Experimental techniques 3.3.1. Nano-particle size analyzer Zetasizer nano-particle analyzer (series Malvern nano ZS) was used to study the colloidal stability and re-aggregation trend of the powder in Chapter 4. The results were analyzed using standard software “Dispersion Technology software, ver. 4.00”. The instrument (figure 3.9) performs size measurements using a process called “Dynamic Light Scattering (DLS)”, also known as “Photon Correlation Spectroscopy (PCS)”. The technique analyzes the Brownian motion of the suspended particles/ agglomerates and relates it to their size [260]. It does this by illuminating the 74 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites particles/ agglomerates with a laser and analyzing the intensity fluctuations in the scattered light. DLS is very sensitive to the intensity of light scattered by the particles/ agglomerates. Large particles/ agglomerates scatter more light than smaller ones. Hence, this technique is very good for studying particle/ agglomerate sizes and colloidal dispersions. Both, Zetasizer nano-particle analyzer and software were Attenuator Laser source supplied by Malven Instruments Ltd., UK. Digital signal processor Detector Cuvette Computer Detector Zetasizer Nano ZS Figure 3.9. Working principle of Zeta particle size analyzer. Modified from [260]. 3.3.2. Spark Plasma Sintering (SPS) Spark Plasma Sintering (SPS) is a variation of hot-pressing. It involves the rapid heating of graphitic dies by pulsed DC electric currents (figure 3.10). This rapid heating rate (up to 600 oC/ minute) combined with high pressure (up to 1 GPa) [261] is the main feature of SPS. During SPS, the detailed mechanism of enhanced densification is unclear. This is due to the particular electrical, thermal and 75 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites mechanical processes that are associated with the SPS process. Modelling suggests that high heating rates reduce the duration of densification-noncontributing surface diffusion, that favours powder systems’ sinterability and the densification is intensified by grain-boundary diffusion [262]. It has also been suggested that the direct current pulse could generate several effects such as spark plasma, spark impact, Joule heating, and electric field assisted diffusion [263-265]. Three factors that contribute to the rapid densification process can be discerned: (i) the application of a mechanical pressure; (ii) the use of rapid heating rates; and (iii) the use of pulsed direct current (figure 3.11), implying that the samples are also exposed to an electrical 15 ms 5 ms Transformer voltage (V) Electric current (kA) field [266]. Time (milliseconds) Figure 3.10. Typical DC-pulsed current Figure 3.11. Effect of DC pulse on the cycles (used in this study). density of alumina [267]. The commonly used name for this technique is very misleading and quite debatable. To date, there is no experimental observation of a “Spark Plasma” during SPS. Other names for SPS are pressure-assisted resistance sintering, electric-discharge sintering, discharge powder compaction, electro-consolidation, plasma activated sintering, fieldassisted sintering, electric pulse sintering, pulse electric current sintering and electromagnetic-field-assisted powder consolidation [262, 264]. SPS is also a cost-effective sintering technique. The entire processing time to sinter dense ceramic composites is about 1/10 of that required by conventional sintering processes [268]. It is also shown that SPS typically achieves maximum density at 76 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites temperatures of about 150-200 oC [269] and 250-300 oC [270] lower than hot pressing. SPS also limit the grain growth. Lee et al. [271] reported grain size of ~200 nm in 99% densified TiO2 by SPS. Compared to this, the microwave sintering resulted in grain size of ~300 nm, as against 1–2 μm grain size in conventionally sintered TiO2 [272]. Other techniques that also involve rapid heating and sintering are Self-propagating High-temperature Synthesis (SHS) [273, 274] and microwave sintering [275, 276]. However, temperature and heating rate cannot be practically controlled in these techniques. For ceramic – CNT nanocomposites, the use of the SPS technique allows sintering in very short times, limiting the matrix grain growth and damage to the CNTs [197]. Compared to SPS, hot-pressing methods, involving longer durations and high temperature, damage carbon nanotubes in the nanocomposite, leading to a decrease or total loss of reinforcing effects without producing fully dense nanocomposites [97, 138, 140, 141, 147]. In this study, all the samples were Spark Plasma Sintered (SPSed) in a HPD 25/1 furnace by FCT Systeme, Germany (figure 3.12). The current set up at Queen Mary, University of London allows samples of up to about 80 mm to be produced with sintering temperatures up to 2200°C. Details of the graphite parts and their assembly are shown in figure 3.13. The furnace has an optical pyrometer above the furnace and focussed inside a hole in the top graphite punch (figure 3.13). Typically, the voltage applied between the upper and lower punches is in the order for few voltages and the current can be as high as few thousand amperes. Graphite dies were lined with graphite paper (Le Carbone, UK, thickness: 0.38 mm) before pouring powder in them. It was used to prevent direct contact between graphite parts and the ceramic powder and to guarantee electrical contacts between all parts. Graphite dies were covered with carbon insulation (SGL, UK, thickness: 7mm) to avoid heat loss during SPS operation. 77 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 3.12. SPS facility by FCT Systeme, Germany: (a) SPS facility at Queen Mary, University of London, UK; and (b) SPS at 1800 oC [277]. Pyrometer measuring temperature through this channel Top alloy piston (moving) Carbon reducer Carbon die Carbon insulation jacket Powder (before sintering) Carbon punch Bottom alloy piston (fixed) Figure 3.13. Cross-sectional view of carbon die set. 78 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.3.3. Density measurements For the evaluation of bulk density, all of the sintered samples were ground using SiC paper to remove the carbon paper and diffused carbon layer. The bulk density was measured by water buoyancy method and the density of powders was measured by Helium pycnometery. 3.3.3a. Water buoyancy method The actual density ( A ) of the sintered samples was measured by weighing them in air ( mA ). The material was then submerged in distilled water and put on heating plate to boil the water. In this way, distilled water penetrated into the open porosities. After 10 minutes of boiling, the distilled water was left to cool down to room temperature, as the density of distilled water ( W ) changes with the temperature. Using a Archimedes density kit, the sintered material was submerged in the distilled water, and the submerged mass ( mw ) was recorded. The actual density ( A ) was then calculated using: mA m m A A W Equation 3.1 W XRD analysis confirmed that there was no reaction between the CNTs and the ceramics used in this project. Hence, the theoretical density ( T ) of the composites was calculated according to the rule of mixtures. Rel. theoretical density ( R ) was calculated by: R 100 T A Equation 3.2 79 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.3.3b. Helium pycnometery The density of the CNTs was measured by He pycnometer. The Micromeritics AccuPyc 1330 pycnometer is a fully automatic gas displacement pycnometer. The instrument determines the skeletal density and volume of a sample by measuring the pressure change of helium in a calibrated volume chamber. The instrument is very accurate as the Helium molecule has a diameter of less than 0.1 nm. The instrument has a density resolution of 0.001 g/cm3 [278]. The technique is good for powders only. It should not be used for sintered products, as it cannot evaluate the amount of closed porosity. 3.3.4. Electrical conductivity measurements The electrical conductivity of the sintered materials was measured using the twoprobe method [167] for the temperature range 30 – 500 oC. Silver electroded specimens (3 × 3 × 3 mm) were characterised (equation 3.3) with a high sensitivity digital micro-ohmmetre (Keithley 580). l R A Equation 3.3 Where, = electrical conductivity or specific conductance, l = sample thickness (3 mm), R = electrical resistance and A = cross-sectional area (9 mm2). The samples were held in a copper jig (figure 3.14a). A power supply (Keithley 2602) and digital multimeter (Keithley 6517A) were used to measure the current-voltage (I-V) characteristics of the samples. To measure electrical conductivity at high temperature, a specially designed alumina chamber in a tubular furnace was employed (figure 3.14b). Platinum wires and electrodes were selected as contacts. A heating rate of 2 o C/ minute was selected and the temperature was measured using a K-type thermocouple positioned next to the sample. 80 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Connecting to micro ohm-metre Copper connects (a) Alligator gold connects Spring for uniform contact pressure Silver electroded sample Polymer (insulating) base (b) Vertical tube furnace Silver electroded ceramic plate Support for hollow alumina tube chamber Thermocouple measurement Silver electroded sample (placed next to thermocouple) Platinum wires (protected in alumina tube) Micro ohm-metre Figure 3.14. Electrical conductivity measurement setup: (a) room temperature jig; and (b) high-temperature characterisation chamber in the tubular furnace. 81 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.3.5. Vickers indentation Due to its simplicity, its non-destructive nature, and the fact that minimal machining is required to prepare the sample, the use of the Vickers indentation method to quantify toughness has become quite popular [279]. A diamond indenter was applied to the surface of the specimens. Upon removal, the impression of indent was used for the quantification of hardness. The length of the radial cracks (figure 3.15) reflects of the crack toughness of the material which can be used to calculate the toughness of the material by semi-empirical formulation. Load of 2.5 kg was used for a duration time of 5 seconds. Vickers hardness was evaluated in accordance with ASTM C132703 [280]. H 0.0018544 P d2 Equation 3.4 Figure 3.15. A typical Vickers indent. Modified from [281]. 82 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Where, H = hardness, P = indentation load and d = average length of indentation diagonals (figure 3.15). For indentation fracture toughness Anstis’s equation [178] was used: P E IFT 0.016 3/ 2 c H Equation 3.5 Where, IFT = indentation fracture toughness and c = half of the mean radial crack length (figure 3.15). 3.3.6. Microscopic analysis 3.3.6a. Optical Microscopy For microstructural characterisations in Chapter 4, 5 and 9, optical microscopy was performed using an Olympus BX60F fitted with a live camera assembly. 3.3.6b. Field-Emission Scanning Electron Microscopy (FE-SEM) FE-SEM was used extensively (chapter 3-10) in this study. Two different FE-SEMs were used, a JEOL (JSM-6300, 20 kV) and FEI (Inspect F, 20 kV). All of the powders, fractured and polished surfaces were gold coated prior to SEM examination. 3.3.6c. High Resolution transmission Electron Microscopy HR-TEM In Chapter 6 and chapter 7, HR-TEM (JEOL 2010, 200 kV) is used to study the survivability of CNTs after SPS. The electron transparent nanocomposite films (thickness <100 nm) were prepared by mechanical grinding, polishing, dimpling and focused ion milling. 83 Chapter 3. Materials and Experimental Techniques Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 3.3.7. Thermo gravimetric analysis (TGA) TGA was performed in chapter 4 and 8 using TA Instruments SDT Q600 TGA thermo gravimetric analyzer. All specimens were examined on platinum pans in the range 30 – 1000 oC. A heating rate of 5 oC/ minute in flowing air (at 180 ml/ minute) was used. Powder sample masses ranged from 30 – 40 mg, whereas sintered sample masses ranged from 30 – 50 mg. 84 Chapter 3. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 4. Homogenisation of CNTs in Ceramics 4.1. Introduction A critical step in the processing of ceramic – CNT nanocomposites is the preparation of a suspension of homogeneously isolated CNTs that can be added to different ceramic powders to make nanocomposites. Ultrasonication (figure 4.1) in solvents is a common primary step, and high power bath ultrasonication has been shown to be one of the best methods for producing homogeneous and relatively aggregate-free dispersions [95]. The dispersant used needs to overcome the strong van der Waals force between CNTs and then resist their re-agglomeration [95]. (a) (b) (c) Metallic chamber Wave generators Dispersion bottle De-ionised water bath Figure 4.1. Ultrasonication bath: (a) Apparatus with dispersion bottle; (b) shock waves in distilled water, top view of the apparatus; and (c) schematic diagram. The best solvents reported for generating CNT dispersions are amides, particularly N,N-Dimethylformamide (DMF) and N-methylpyrrolidone (NMP) [282-284]. All of 85 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites these solvents are characterized by high values for β (electron pair donicity), negligible values for α (the hydrogen bond donation parameter of Taft and Kamlet), and high values for π (solvochromic parameter) [284, 285]. Thus, Lewis basicity (i.e., the availability of a free electron pair) without hydrogen donors is key to the good dispersion of CNTs [284, 286]. However, this seems to be a necessary but not complete set of conditions, as Dimethyl Sulfoxide (DMSO), a mediocre solvent, meets these criteria [284, 287]. Ham et al. [288] illustrated that solvents with high values of dispersion component (δd) of the Hildebrand solubility parameter (δt) are the best for making homogeneous and agglomerate-free dispersions of CNTs. In this regard, DMF (δd = 17.4 MPa1/2) proves to be better than ethanol (δd = 15.8 MPa1/2), water (δd = 15.6 MPa1/2), acetone (δd = 15.5 MPa1/2) and methanol (δd = 15.1 MPa1/2) for making CNT dispersions [288]. Other approaches to make stable dispersions are the use of surfactants [289-291], acid treatments [123, 286, 290], and chemical functionalisation [32, 43, 292, 293], which change the surface energy of CNTs, improving their adhesion/wetting characteristics and reducing their tendency to agglomerate in the solvents [95]. However, in all these approaches, the selection of the solvent still remains a very important factor. Despite the significant differences in the chemical properties of various solvents [284, 285, 288], many authors [111, 112, 114, 119, 129, 137, 140, 153, 181] have repeatedly employed ethanol for dispersing CNTs in different ceramics. This seems primarily due to the fact that alcohols are a common media for ball milling of ceramics. Wang et al. [127] compared the use of methanol and DMF for the dispersion of CNTs in an alumina matrix, and reported that the choice of dispersant made no difference to the nanocomposites in terms of their densities and microstructures. They did not describe the processing details for their ultrasonic agitation, so there is insufficient evidence to assess their observations. The effectiveness of a dispersion route depends upon various factors like solvent properties, bath properties, energy applied, solution concentration, geometry of the vessel, and the vessel position [95]. 86 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites A significant amount of work has been done on the dispersion of CNTs in solvents. DMF (appendix A) is well recognized among polymer researchers [31, 92, 294-297] as a good dispersant for processing CNT – based nanocomposites. Lau et al. [294] reported solvent effects in the order of DMF > ethanol > acetone for making polymer – CNT nanocomposites. After synthesizing, Moniruzzaman et al. [295] stored CNTs in DMF to avoid agglomeration and later also employed DMF to disperse their CNTs in epoxy nanocomposites. Ciselli [92] showed smaller agglomerate sizes for CNTs in DMF solution as compared to other solvents. The use of DMF as a dispersant by the ceramic community making CNT nanocomposites is very rare [127] and not fully realized. In this chapter, we study the agglomeration and re-aggregation behaviour of coated and uncoated CNTs, and carbon black in ethanol and DMF. This chapter also compares the use of ethanol and DMF for making alumina – CNT nanocomposites by analyzing the pre-sintering (colloidal stability and agglomerate size analysis) and post-sintering (dispersion profile and electrical conductivity measurements) stages. 4.2. Experimental procedure 4.2.1. Colloidal dispersions and characterisations To monitor the colloidal stability, a 77 mg/l concentration of CNTs in DMF was hand mixed for 15 seconds and high power bath ultrasonicated (Engisonic plus, Engis Ltd., UK) for 1 hour. It was then hand shaken for another 5 minutes to remove any gradients produced by non-uniform ultrasonication. The dispersion was then placed in front of a luminescent light box to observe its re-aggregation behaviour. For agglomerate size analysis, CNTs were hand mixed for 15 seconds in DMF solution (100 mg/l) and high power bath ultrasonicated for different durations. The solutions were then transferred to standard glass cuvettes (10 mm × 10 mm × 45 mm) and placed in a Malvern Zetasizer nanoparticle size analyser (Nano ZS). The software was programmed to record the average of at least 30 readings for the quantification of the 87 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites agglomerates’ size and distribution. Using the same procedure as above, alumina dispersion in DMF was also characterized (150 mg/l). All of the above procedures were then repeated with ethanol to allow a comparison between the behaviour of DMF and ethanol. Re-agglomeration behaviour of CNTs was studied systematically using CNTs supplied by NanoDynamics, USA. To study the re-agglomeration behaviour, a 1.75 g/l concentration of CNTs in DMF was hand mixed for 15 seconds and high power bath ultrasonicated in a glass cuvette for 30 minutes. The solution was then placed in a nanoparticle size analyzer. The software was programmed to record the average of at least 30 readings. The re-agglomeration behaviour was studied for 20 minutes. The same experiment was conducted for alumina coated CNTs (25 cycles and 50 cycles). All of the above procedures were then repeated with ethanol to allow a comparison between the behaviour of DMF and ethanol. 4.2.2. Nanocomposite powder preparation Alumina – 2 wt% (~4.7 vol%) CNT nanocomposites were prepared. The CNTs were dispersed in DMF via high power bath sonication for 2 h and then hand mixed with the alumina nanopowder for another 5 minutes. The liquid mixture was transferred to another jar filled with zirconia balls (milling media) of two different sizes (10 and 5 mm, mass ratio: 3:2). The jar was sealed and rotation ball milled for 8 h at ~200 rpm. The milled powder was then shifted through a steel pan. The milled slurry mixture was dried at 75 oC for 12 hours on a heating plate and then transferred to a vacuum oven (100 oC) for 3 days for complete removal of the dispersant. A solvent trap (filled with ice) was connected between the vacuum pump and the oven. The same procedure was followed for making alumina – alumina coated CNTs (50 cycles) nanocomposite powder and alumina – carbon black nanocomposite powder. 88 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The dried mixture (alumina – CNT nanocomposite powder) was ground and sieved using a 250 mesh and then returned to the vacuum oven for another 4 days at the same temperature for thorough extraction of the solvent. This lengthy drying procedure was followed because any residual solvent has a detrimental effect on the properties of CNT-reinforced nanocomposites [287, 288]. The same method was employed to make nanocomposite powder using ethanol. A brief summary of the dispersion process is given in figure 4.2. Mixing solvent with CNTs Ultrasonicating mixture for 2 hours Mixing CNT dispersed mixture with ceramic powder Ball milling mixture for 8 hours Drying milled mixture on hot-plate for 12 hours Vacuum drying powder mixture for 7 days Nanocomposite powder ready for Spark Plasma Sintering Figure 4.2. Pre-sintering processing of ceramic – CNT nanocomposite powder. 4.2.3. Spark Plasma Sintering (SPS) Dried nanocomposite powder (~2 g) was poured into a carbon die and cold pressed at 0.62 MPa for 5 s before sintering. Nanocomposite discs (thickness 2 mm and diameter 89 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 20 mm) were prepared by SPS. A pressure of 100 MPa was applied concurrently with the heating (rate 300 oC/ minute) and released at the end of the sintering time, which was 3 minutes for all of the samples. The sintering temperatures were in the range 1200 – 1950 oC. All of the samples were slowly cooled (~50 oC/ minute) to avoid fracture due to thermal shocks and differential contractions. The same sintering procedure was repeated for the dried nanocomposite powder dispersed using ethanol. 4.2.4. Nanocomposite characterisations The SPSed samples were ground using SiC paper and diamond polished down to 1 μm. Density measurements were conducted using helium pycnometer (AccuPyc 1330, Micrometics) and water buoyancy methods. There were no significant differences in the results from both techniques, so the mean was used to characterize the density of the SPSed samples. SPSed samples were fractured in order to observe the agglomeration and dispersion of CNTs. Nanocomposite powder (alumina – CNT) and the fractured surfaces were gold coated and observed in a field emission scanning electron microscope (FE-SEM). The electrical conductivities of the samples were measured (section 3.3.4) with a high sensitivity digital micro-ohmmetre (Keithley 580) using the two-point method on silver electroded specimens (3 mm × 3 mm× 3 mm) prepared using a diamond cutting machine. 4.3. Results and discussion 4.3.1. Natural drying patterns Before sieving, the natural drying patterns (figure 4.3) of the nanocomposite powder hinted at the strength of the secondary bonding of nano carbon fillers. Because of high entanglement and aspect ratio of CNTs, they formed centimetre-sized agglomerates with alumina (figure 4.3a). It appeared that coating CNTs with alumina decreased the van der forces between CNTs, as they did not form entangled networks (figure 3b). 90 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The same was also observed for alumina – carbon black nanocomposite powders because of the non-fibrous nature of carbon black (figure 3c and 3d). This is another qualitative assessment that highlights the strong entanglement in CNTs because of their fibrous nature. A good dispersion was observed after sieving alumina – 2 wt% CNT (figure 4.4). (a) (b) (c) (d) Figure 4.3. Natural drying patters formed during processing of nanocomposite powder: (a) alumina – 2 wt% CNT; (b) alumina – alumina coated 2 wt% CNT. CNTs were supplied by NanoDynamics, USA; (c) alumina – 2 wt% carbon black (Vulcan XC72); and (d) alumina – 2 wt% carbon black (Printex L6). 91 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 4.4. Alumina – 2 wt% CNT nanocomposite powder after sieving. Individual CNT can be seen. CNTs were supplied by Nanocyl, Belgium. 4.3.2. Agglomerate size analysis The presence of agglomerates in powders with very fine grain size contributes to grain coarsening during sintering and produces non-uniformity in the resulting microstructure [261]. Agglomeration is particularly significant in CVD-grown nanotubes because substantial entanglement of the tubes occurs during nanotube synthesis [167]. CNTs and alumina powder were separately ultrasonicated for various durations and their agglomerate sizes were then measured immediately (figure 4.5). For all of the ultrasonication durations, DMF disperses CNTs more efficiently as compared to ethanol, by reducing the agglomerate size (figure 4.5a). It should be noted that these results are presented for comparison and do not represent the optimum conditions for dispersing the CNTs. DMF also showed better dispersion of alumina compared to ethanol (figure 4.5b). All these observations related to colloidal dispersions can be explained by the higher values of Hildebrand solubility parameter (δt) and Lewis basicity of DMF as compared to ethanol [284, 285, 288]. Looking at the lengthy processing procedures followed by many researchers [38, 110, 111, 114, 92 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 153, 294] before sintering (for ceramics) or curing (for polymers), the fact cannot be ignored that the efficiency of DMF for de-bundling and making stable CNT dispersion is much better as compared to that of ethanol. (b) (a) Figure 4.5. Agglomerate size analysis with respect to ultrasonication time in different solvents: (a) CNTs, supplied by Nanocyl, Belgium; and (b) alumina. 4.3.3. Re-agglomeration behaviour Successful fabrication of nanocomposites depends crucially on maintaining stable colloidal mixtures of the nanotubes and matrix phase [298] before ceramic sintering or polymer curing. Figure 4.6 shows the colloidal dispersion for CNTs in DMF and ethanol after ultrasonication and at different time intervals. The DMF dispersion is very stable, showing no signs of agglomeration even after several months (figure 4.6). These observations are consistent with previous work that showed that CNT – DMF dispersions aggregate on a timescale of days [283] and weeks [299]. The CNT – ethanol dispersion re-agglomerated significantly within half an hour of ultrasonication. This qualitative analysis shows that CNTs are much more stable in DMF as compared to ethanol. 93 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites DMF Ethanol Soon after ultrasonication After 24 hours After 15 months Figure 4.6. Colloidal dispersion stability comparison after 1 h bath sonication and 5 minutes hand shaking. The diameter of the bottles is 25 mm. CNTs were supplied by Nanocyl, Belgium. To quantify such observation, re-agglomeration with the passage of time was observed for all types of CNTs in a nano particle size analyzer. The slopes (figure 4.7) indicated the rate of re-agglomeration of the CNTs in ethanol and DMF. The higher the magnitude of slope, the faster the re-agglomeration and vice versa [300]. At time = 0 minutes, it appears that the dispersion process employed was not appropriate for all samples because of the presence of large agglomerates. As mentioned before, these results are for comparison and do not represent the optimum conditions for dispersing the CNTs. 94 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) (c) (d) Figure 4.7. Re-agglomeration behaviour in different solvents after 30 minutes of ultrasonication. CNTs were supplied by NanoDynamics, USA: (a,b) pristine CNT; (c) alumina coated CNT (25 ALD cycles); and (d) alumina coated CNTs (50 ALD cycles). Comparing the slopes, faster re-agglomeration was observed in ethanol as compared to DMF for all types of CNTs. Higher colloidal stability is only possible if the CNTs have a charge on their surface preventing aggregation after dispersion by repulsive electrostatic forces [301]. The lower charge density of the nanotubes dispersed by ethanol is responsible for their lower stability in ethanol compared to DMF. DMF (pH = 9) is also more basic as compared to ethanol (pH = 6), which is another important factor for its better dispersion properties [286]. Other important factors are the high values of Hildebrand solubility parameter (δt) and Lewis basicity of DMF as compared to ethanol [284, 285, 288]. It was also observed that coating CNTs with alumina reduced the re-agglomeration rate. The larger the thickness of alumina coating, the slower the re-agglomeration. This is attributed to weaker van der Waals attraction between coated CNTs. 95 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 4.3.4. Microstructure of nanocomposites Samples SPSed for this chapter are summarised in table 4.1. Table 4.1. Samples SPSed for this chapter#. # Rel. theoretical Matrix Weight % Dispersant SPS conditions Alumina 2 Ethanol 1200 oC/ 100 MPa/ 3 minutes ~77 Alumina 2 Ethanol 1240 oC/ 100 MPa/ 3 minutes ~80 Alumina 2 Ethanol 1400 oC/ 100 MPa/ 3 minutes ~87 Alumina 2 Ethanol 1600 oC/ 100 MPa/ 3 minutes ~96 Alumina 2 Ethanol 1900 oC/ 100 MPa/ 3 minutes ~100 Alumina 2 Ethanol 1950 oC/ 100 MPa/ 3 minutes - Alumina 2 DMF 1200 oC/ 100 MPa/ 3 minutes ~78 Alumina 2 DMF 1400 oC/ 100 MPa/ 3 minutes ~87 Alumina 2 DMF 1600 oC/ 100 MPa/ 3 minutes ~97 Alumina 2 DMF 1900 oC/ 100 MPa/ 3 minutes ~100 Alumina 2 DMF 1950 oC/ 100 MPa/ 3 minutes - density (%) All CNTs were supplied by Nanocyl, Belgium. For CNTs, 2 wt% = ~4.7 vol% 96 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Representative images of the fractured surfaces of the SPSed nanocomposites were selected for studying the distribution of the CNTs (figure 4.8). Bright micron-sized agglomerates of CNTs on alumina grains (100-300 nm) are visible in the sample prepared using ethanol as the dispersant (figure 4.8a). These agglomerates were produced because of the inability of ethanol to make an agglomerate-free dispersion of CNTs before sintering. The sample prepared using DMF as the dispersant has a homogeneous distribution of individual CNTs (figure 4.8b). It is interesting to note the different grain sizes of the nanocomposites produced with ethanol and DMF. Those prepared with ethanol have a noticeably larger grain size (figure 4.8a) compared to the equivalent nanocomposite prepared with DMF (figure 4.8b). The effect of CNTs on the grain size refinement of nanocomposites is the subject of Chapter 5. (a) (b) Figure 4.8. Fractured surfaces of sintered alumina – 2 wt% CNT samples, SPSed at 1200 ◦C/100 MPa/ 3 minutes: (a) CNTs dispersed in ethanol and (b) CNTs dispersed in DMF. 4.3.5. Mechanical and electrical properties The presence of agglomerates in nanocomposites is property limiting. They reduce the mechanical properties [302, 303] and electrical properties [303, 304] of the nanocomposites. Figure 4.9 shows the Vickers hardness and indentation fracture 97 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites toughness for alumina – 2 wt% nanocomposites prepared using different solvents. The densities and electrical conductivities of the nanocomposites are shown in figure 4.10. The SPS processing of the colloidally dispersed starting powder mixtures produced nanocomposites of high density with well-distributed CNTs. There was no major difference between the densities of the SPSed samples prepared from the different dried nanocomposite powders. However, nanocomposite prepared using ethanol showed inferior Vickers hardness and indentation fracture toughness as compared to nanocomposite prepared using DMF. Agglomeration favours an inhomogeneous densification and thus inhomogeneous grain size distribution [177]. Agglomerates contain very fine porosity that is responsible for easy crushing under load resulting poor hardness. The toughness improving mechanisms like fibre pull-out, crack deflection and crack bridging are not possible in the presence of agglomerates of CNTs. This resulted in poor indentation fracture toughness of nanocomposite prepared using ethanol as compared to the nanocomposite prepared using DMF (figure 4.9). Figure 4.9. Vickers hardness and indentation fracture toughness of alumina – 2 wt% CNT nanocomposites prepared using different solvents. 98 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Samples melted Figure 4.10. Density and electrical conductivity of alumina – 2 wt% CNT. Error bars are not marked in the electrical conductivity measurements as they are very small (figure 4.10). There is a significant difference in the electrical conductivities of the SPSed samples prepared from ethanol and DMF dispersions, particularly as the density of the nanocomposites increases. This is due to a better homogeneous dispersion of the conductive CNTs in DMF solution as compared to ethanol. Alumina is inherently an insulator (electrical conductivity: 10−13 S/m [211]), so a uniform distribution of highly conductive CNTs is critical for making the nanocomposites good electrical conductors and reducing the percolation threshold. 4.4. Conclusions A prerequisite for the ceramic nanocomposites with good electro-mechanical properties is the homogeneous dispersion and distribution of the CNTs in the ceramic 99 Chapter 4. Homogenisation of CNTs in Ceramics Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites matrices. The extraordinary high specific surface area of CNTs results in very high van der Waals forces between them, inducing a strong tendency to agglomerate. The selection of the ultrasonication medium is very important for the final properties of the nanocomposite. Non-hydrogen bonding Lewis bases are the best solvents for CNT dispersions. From dispersion stability observations and agglomerate size measurements, it is clear that DMF produces fine and stable CNT and alumina dispersions. Faster re-agglomeration was observed in ethanol as compared to DMF for pristine and coated CNTs. Coating CNTs with alumina reduced the re-agglomeration rate. The larger the thickness of alumina coating, the slower the re-agglomeration. No evidence of agglomeration and a good distribution of the CNTs was observed in FESEM micrographs of the SPSed samples when they were mixed with alumina in DMF. Nanocomposites prepared using DMF dispersions showed better dispersions and higher electrical conductivity as compared to those prepared using ethanol dispersions. Therefore, it is concluded that DMF is a good dispersant for making homogeneous and agglomerate-free slurries by any type of colloidal processing. 100 Chapter 5. Sintering of Ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 5. Sintering of Ceramic – CNT Nanocomposites 5.1. Introduction Control of microstructure and improvement in densification is one of the objectives of using dopants/ reinforcements in ceramics. Improved mechanical [127, 140, 200], electrical [129] and thermal properties [226] have also been reported for ceramic – CNT nanocomposites produced by rapid processing using SPS. However, the role of CNTs in the sintering of ceramics is not clear in the literature, possibly because of differences in the materials and, particularly, dispersion and mixing of the CNTs. An and Lim [305] suggested that CNTs in alumina decreased mass transportation during sintering, which inhibited the densification process. Tatami et al. [129] and Jiang and Gao [226] reported inhibitation of densification of silicon nitride – CNT and titanium nitride – CNT nanocomposites. For glass-ceramics, Boccaccini et al. [116, 306] and Ning et al. [119] observed that the presence of MWNTs in a glass matrix hindered the densification of the material; the CNTs served as nucleation points, and the crystallized phases acted as a rigid body, which hindered densification. However, Morisada et al. [103] and Wei et al. [110] reported that the addition of CNTs to silicon carbide and alumina had no effect on the densification behaviour of the nanocomposites. These negative and nil effects are possibly because of differences in the materials studied, dispersion and mixing of the CNTs, and the presence of agglomerates. Guo et al. [141] reported that CNTs improved the densification and 101 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites mechanical properties of silica nanocomposite. Huang et al. [222] attributed this effect to the good combination of SPS sintering and the high electrical and thermal conductivity of CNTs. In some studies [110, 114, 140, 147, 158, 161, 222] it was reported that the addition of CNTs modified the grain size. Zhan et al. [140] prepared, using SPS, 100% dense alumina and alumina – 10 vol% CNTs at the same processing temperature (1150 oC); and the grain size of the alumina in the nanocomposite was ~39% smaller as compared to in the alumina. Even carbon nanofibres (diameter 100200 nm) were found to retard alumina grain growth [137]. However, there has been no systematic investigation into the effect of CNTs on the sintering behaviour and grain growth of ceramics. It is well known that carbon black is one of the best sintering aids for various ceramics [307-310], and very little addition is required as compared to other additives [311, 312]. Erkalfa et al. [312] reported that due to the self-lubricating nature of carbon, it enhances compactibility and compressibility, which favours densification. The increased densification of non-oxide ceramics is possibly due to the removal of surface oxide layer on the powder by the addition of carbon, which increases surface diffusion during sintering [227, 313]. In this chapter, the sintering and grain growth behaviour of alumina – CNTs and alumina – carbon black nanocomposites, alumina prepared by Spark Plasma Sintering (SPS) were studied. The influence of CNT addition on the sintering and grain growth of reduced titanium dioxide was also studied. In the last section of this chapter, the effect of residual impurities (left during the synthesis of CNTs) on the sintering behaviour of alumina – CNT nanocomposites is analysed. 102 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 5.2. Experimental procedure 5.2.1. CNTs purification As per supplier (Nanocyl, Belgium), CNTs were >90% pure. An acid treatment was performed using a mixture of concentrated nitric (HNO3, 90%) and sulfuric (H2SO4, 90%) acids. Distilled water (~20 vol%) was used to dilute the acids. In order to produce pure CNTs, the as-received CNTs (400 mg) were mixed with 200 ml dilute acidic solution. Both acids were equally mixed in the solution. The acid-CNT mixture was homogenized by stirring with a glass rod on heating plate (~85 oC) for 30 minutes and then bath ultrasonicated for 2 hours. The resulting CNT dispersion was thoroughly washed with distilled water until the filtrate was colourless and neutral (pH ~7) after filtration. A Whatman filter paper of 1 μm was used. The purified CNTs were then dried for 48 hours at 100 oC in an oven. The quality of CNTs was quantified by thermo gravimetric analysis (TGA). A short summary of the purification process is shown in figure 5.1. Mixing CNTs with dilute acid mixture Mixing on hot-plate (~110 oC) for 30 minutes Bath Ultrasonicating for 2 hours Filtering with ultra fine filter paper Cleaning with distilled water until pH ~ neutral Drying CNTs at 100 oC for 48 hours Purified CNTs Figure 5.1. Purification of CNTs by acid treatment. 103 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 5.2.2. Nanocomposite powder preparation In this section, composite powders were prepared using different dispersion routes. Those prepared by a colloidal dispersion method (using DMF, section 4.2.1) gave the best results. This method will be referred to as “optimum method of dispersion” in this study. To compare the role of dispersion on the sintering behaviour, different routes were adopted to prepare in-homogenous dispersions of CNTs. The properties of the ceramic nanocomposites prepared by the different routes were then compared. 1. Viscous solution mixing: The same process was followed as in section 4.2.1, with the exception of using 1 gram of CNTs in 150 ml of DMF. In “optimum method of dispersion”, 400 mg of CNTs in 150 ml of DMF was used. 2. Hand mixing: The same process was followed as in section 4.2.1, with the exception of doing hand mixing of CNTs in DMF in place of ultrasonication and ball milling. The hand mixing was done for 30 minutes. 5.2.3. Spark Plasma Sintering (SPS) Ceramic and nanocomposite pellets (diameter 20 mm and thickness 2 mm) were prepared by SPS. A pressure of 100 MPa was applied concurrently with the heating (rate 300 oC/ minute) and released at the end of the sintering period (3 minutes) for all samples. An alumina / alumina – CNTs / alumina laminate sample was prepared by subsequent compacting of alumina nanopowder, nanocomposite powder and alumina nanopowder in a graphite die. The powder compact was cold pressed at 0.64 MPa for a few seconds and co-sintered at 1800 oC for 3 minutes under a pressure of 85 MPa. 104 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 5.2.4. Nanocomposite characterisations All of the sintered samples were ground using SiC paper down to 4000 grit. The density of the ground samples was measured using the water buoyancy method. Selected samples were thermally etched at 250 oC less than the SPS temperature for 10 minutes. Field emission scanning electron microscopy (FE-SEM) was used to observe fractured surfaces in order to determine the grain sizes. The polished laminate and fractured surfaces were coated with a very thin layer of gold and transferred to an FE-SEM for examination. Grain sizes were measured with the aid of software (Image tool for Windows, version 3.00, developed by UTSHCSA, USA). A minimum of 200 readings was taken to measure the grain sizes of each material. 5.3. Results and discussion 5.3.1. SPS of alumina and alumina – CNT nanocomposites With the aid of user-friendly, in-built touch screen, SPS can be monitored very intelligently during sintering. Whenever a material shrinks, sinters or melts the speed of the moving punch increases noticeably at that instant. This observation is reproducible and a very clever way of knowing the important events during sintering of unknown and new materials. The bulk initial phase of alumina was γ. During SPS, the phase transformation (γ to α) in alumina and alumina – CNT nanocomposite was observed distinctly (figure 5.2) at ~1200 oC. Such phase transition is reported to result in the formation of interlocking vermicular structure, which is detrimental to densification [272]. 105 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 5.2. Moving punch speed as the function of temperature during SPS for alumina and alumina – 5 wt% (~11.2 vol%) CNT (uncoated) nanocomposite. CNTs were supplied by Nanocyl, Belgium. Inset shows SPS program details. At 1600-1700 oC, there is another peak in the speed of the moving punch. This relates to the bulk sintering or liquid phase sintering (figure 5.2). In both peaks, CNTs enhanced the compressibility and compactibility of the nanocomposite, which is evident from the higher speed of the moving punch as compared to the lower speed in alumina (figure 5.2). However, when CNTs are encased within the alumina coating, the speed at these temperatures (~1200 oC and 1650 oC) was reduced and the events were delayed as well (figure 5.3). This illustrates that uncoated CNTs have a very clear effect on the sintering of alumina – CNT nanocomposites. 106 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 5.3. Moving punch speed as the function of temperature during SPS for alumina – 5 wt% CNT (uncoated) and alumina – 5 wt% CNT (coated, 50 ALD cycles) nanocomposite. Heating rate: 300 oC/minute. CNTs were supplied by Arkema, France. 5.3.2. Sintering behaviour and possible mechanisms Figure 5.4 illustrates the rel. theoretical density versus sintering temperature of ceramic – CNT nanocomposites SPS processed for 3 minutes at their sintering temperature at a pressure of 100 MPa. The addition of CNTs significantly reduces the sintering temperature required to achieve full densification of the nanocomposites as compared to pristine ceramics (figure 5.4). For example, at 1200 oC, the alumina – 5 wt% CNT has 100% rel. theoretical density, while the alumina has 59% rel. theoretical density. 107 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 5.4. Rel. theoretical density as a function of sintering temperature for: (a) alumina and nanocomposites; and (b) reduced titania and nanocomposites. 108 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites For comparison, the sintering behaviour of alumina – 2 wt% carbon black nanocomposites prepared by the same processing route were also investigated. The effect of CNTs and carbon black, for the same carbon content, on densification behaviour is very similar (figure 5.4a). Homogenous dispersion of CNTs is crucial for good densification of CNT nanocomposites. The increased density of the nanocomposite with increasing CNT content (figure 5.4) suggests that the CNTs were well dispersed, which is not true for nanocomposites prepared using in-homogenous powder mixtures (figure 5.5). Figure 5.5. Bulk density of the sintered product as the function of homogeneity of CNTs in alumina matrix. SPS conditions: 100 MPa/ 3 minutes. CNT content: 5 wt%. In electrically insulating powders, like alumina, the current only flows through the graphite dies and punches during SPSing (figure 5.6a and figure 5.6b). Huang et al. [222] speculated that CNTs promote densification of ceramic powders due to the flow of electrical current through the conductive powder (figure 5.6c) as well as graphite dies and punches during SPS. The high pulse current in SPS is preferentially transported through CNTs due to their low resistance, which locally increases the 109 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites temperature near CNTs. In this manner, CNTs become the dominant heat generator in the composites [222]. However, no evidence of such phenomenon was presented. Table 5.1 shows the electrical resistance of graphite die set and its contents. It is confirmed that as opposed to alumina compact, current flows through alumina – CNT nanocomposite compact during SPS. Electrical current flows through CNTs that promotes local Joules heating in the powder compact. (a) (b) (c) Figure 5.6. Schematics of graphite die sets: (a) alumina before SPS; (b) current passing through graphite only during SPS; and (c) current passing through graphite and alumina – CNT nanocomposite compact during SPS. Table 5.1. Electrical conductivity measurements of graphite die sets. Graphite die set Electrical resistance† (mΩ) Comments Empty die 2.85 ± 0.3 - Die + CNTs 2.97 ± 0.4 - Die + air cavity 3.81 ± 0.1 - Die + alumina powder 3.72 ± 0.8 Before/ after sintering Die + alumina – 5 wt% CNT 3.31 ± 0.3 Before sintering Die + alumina – 5 wt% CNT 3.02 ± 0.4 After sintering † All die sets contain same volume of powder/ cavity (0.25 cm3) and pressed to 1 ton load. 110 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 5.7 shows the displacement of the rams during cold pressing (loading and unloading cycle) for alumina and alumina – 5 wt% CNT powders. Subtracting the effect of the compliance of the SPS loading train and the die set, the alumina – 5 wt% CNT powder has better compactibility as compared to alumina powder. The green relative density for the alumina – 5 wt% CNT compact is 47.6%. It was not possible to measure the same for alumina, as the compact did not stay intact after removal from the die. The improved sinterability of alumina with carbon additions is partly due to improved self-lubricating properties, which promotes compactibility and compressibility of the nanocomposite powder. Figure 5.7. Compressibility and compactibility analysis performed at room temperature by uni-axial pressing. 111 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 5.3.3. Grain growth modification The density versus grain size dependence of alumina, alumina – 2 wt% carbon black and alumina – 2 wt% CNT nanocomposites are shown in Figure 5.8. This shows that it is possible to produce completely dense ceramics with sub-micrometer grain structures by the addition of CNTs. The grain growth during densification is significantly less in the alumina – 2 wt% CNT and alumina – 2 wt% carbon black nanocomposites, compared to the alumina even though they are dense and the alumina is not. Figure 5.8. Rel. theoretical density as a function of grain size for alumina and nanocomposites. Figure 5.9 shows the microstructure of alumina (figure 5.9a) and the nanocomposites (figure 5.9b and c) sintered under the same conditions of 1800 C for 3 minutes. The CNTs are located at the grain boundary (figure 5.9c), which has been previously 112 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites reported [195]. Comparing the microstructures of the monolithic alumina and the nanocomposites, the addition of carbon black and CNTs retards grain growth. The CNTs have a greater grain size retardation effect than the carbon black, which may be attributed to the different geometric contributions of CNTs and carbon black powder. (a) (b) (c) (d) Figure 5.9. FE-SEM images of fractured surfaces of sintered nanocomposites processed at 1800 oC/ 100 MPa/ 3 minutes: (a) alumina; (b) alumina – 2 wt% carbon black (Printex L6); (c) alumina – 2 wt% CNT (Nanocyl, Belgium); and (d) alumina – 2 wt% alumina coated CNT (50 ALD cycles, NanoDynamics, USA). The grain sizes versus sintering temperature are shown in figure 5.10. There is a very large difference in the grain size of the alumina and the alumina – 5 wt% CNT nanocomposite (figure 5.10a). Figure 5.10b shows the comparison of the effect of CNTs and carbon black, and CNT content on grain size. The carbon black, like the alumina, shows an exponential dependence of grain size with temperature (same sintering time of 3 minutes). An parabolic dependence of the grain size with 113 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites temperature is reported for monolithic ceramics [279], particulate composites [314317] and doped ceramics [266, 318-320] The alumina – CNT nanocomposites show a decreasing rate of grain growth with increasing CNT content (figure 5.10a). The 5 wt% CNT nanocomposites show a nearly linear dependence (figure 5.10b), which suggests a different type of mechanism for grain growth. The CNTs form a strong entangled network around the grains, which appears to constrain the grain growth. This effect was not observed when CNTs were coated with alumina. There was no web of CNTs, therefore large grains are very clear in figure 5.9d. Figure 5.10. Grain size refinement effect of: (a) CNTs; and (b) CNTs and CB. In order to observe the grain growth retardation after full densification, lengthy dwell times were used (figure 5.11). The grain size strongly depends on the density of sintered samples [321], therefore, the comparison of grain growth between the two materials should be determined at the same density. In general, grain growth law for polycrystalline materials during isothermal holding can be described using equation 5.1 and 5.2 [321]. D G n Gon t T Equation 5.1 E D Do exp RT Equation 5.2 114 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites where G and Go are the grain sizes at holding time t=t and t=0, n is a constant related to grain growth mechanism, D is diffusion coefficient related to the grains, and Do is diffusion activation energy. The equations are applicable for the solid state sintering of alumina [321]. Because of the decrease in atomic diffusion coefficient caused by the presence of CNTs at the grain boundary, the presence of CNTs reduced the grain growth by a factor of ~5.3 (figure 5.11). Grain growth: 80 nm/minute 15 nm/minute Figure 5.11. Grain size refinement in isothermal conditions (1400 oC/ 100 MPa). 5.3.4. Co-sintering of grain size laminate In monolithic ceramics, a coarse-grained microstructure is desired for applications requiring higher modulus [322], higher creep resistance [279], higher thermoelectric properties (Hall mobility and figure of merit) [323], higher thermal conductivity [323, 324] higher electrical conductivity [323, 325] and higher optical transparency [279]. 115 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites On the other hand, a very fine-grained microstructure is desired for applications requiring higher strength [279], higher wear resistance [326], higher thermal shock resistance [327], higher cyclic fatigue resistance [279], low dielectric loss [279] and higher optical transparency [279]. Grain-size FGMs are advantageous for bio-medical applications [328]. Morsi et al. [328] hot-pressed alumina with different particulate sizes and produced grain-size FGMs with a difference of only ~2 times. Moreover, there was cracking at the interface due to poor bonding. An alumina / 2 wt% CNT nanocomposite laminate structure was fabricated in the current work to demonstrate the grain refinement size effect of CNTs and their ability to produce grain size laminated materials (figure 5.12). (a) (b) Figure 5.12. FE-SEM images of grain size laminate showing interface between alumina and 2 wt% CNT nanocomposite: (a) high magnification fractured surface; and (b) low magnification polished surface. It should be noted that the monolithic alumina layers and the nanocomposite layers were co-sintered at the same temperature (1800 oC). Figure 5.12a shows a fractured surface and figure 5.11b a polished and thermally etched surface. Both layers were successfully co-sintered without any cracking at the interface (figure 5.12a). The upper half part of figure 5.11b shows large grains of alumina (~ 20 μm), whereas the other lower part of the image shows very fine grains (~ 2 μm) in the nanocomposite layer. This image illustrates the potential of CNTs to control the microstructure of ceramics. 116 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites An alumina/ alumina – 5 wt% CNT nanocomposites/ alumina laminate was also SPSed (figure 5.13). Due to large difference in the grain size, there were cracks at the interface (figure 5.13b). These cracks were developed during the slow cooling of the laminate. Further optimisation of sintering process may avoid this cracking. However, clear grain growth retardation effect was observed in these materials (figure 5.14). (a) (b) Figure 5.13. FE-SEM images of polished grain size laminates: (a) alumina and 2 wt% CNT nanocomposite; and (b) alumina and 5 wt% CNT nanocomposite showing cracking at interface. (a) (b) (c) Figure 5.14. FE-SEM images of different areas of polished grain size laminates: (a) alumina region, grain size: 43 μm; (b) 2 wt% CNT nanocomposite region, grain size: 1.81 μm; and (c) 5 wt% CNT nanocomposite region, grain size: 0.99 μm. 117 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 5.3.5. Sintering behaviour of alumina – purified CNTs CNTs supplied by Nanocyl, Belgium were studied in this section. To the best of authors’ knowledge, there is no comparative study for analyzing the effects of residual impurities (left during the synthesis of CNTs) on the sintering behaviour of ceramic – CNT nanocomposites. Figure 5.15 shows the oxidation behaviour of as received-CNT and pure-CNTs. 94.7% loss 98.6% loss Figure 5.15. Thermo gravimetric analysis (TGA) of CNTs, before and after acid purification treatment. Heating rate: 5 oC/minute. Thermo gravimetric analysis (TGA) is a good tool for quantifying the noncarbonaceous species e.g. catalytic metals and oxides (figure 5.16). The as receivedCNT were ~95 % pure, whereas purified CNTs were ~98.5% pure, at the cost of decreased oxidation resistance. 118 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) (c) Figure 5.16. Platinum pan (diameter: ~10 mm) used for TGA: (a) empty pan before analysis, (b) as received-CNTs and (c) impurities left after oxidation. No significant effect of residual impurities was observed on the densification of the alumina – CNT nanocomposite (table 5.2) as appeared in the literature as well [155]. The CNT content used in this chapter was up to 5 wt%. Therefore, the catalytic impurities and oxides have no influence on the densification because of the lower content (<0.25 wt%) in the final product. Table 5.2. Density measurements for CNTs and alumina – ceramic nanocomposites. Rel. theoretical density (%) of alumina – 5 wt% Material Powder density CNT nanocomposite (g/cm3)† (g/cm3)* SPSed at 1400 oC SPSed at 1600 oC As received CNTs 1.66 ± 0.4 99.7 ± 0.8 100 ± 0.6 Purified CNTs 1.59 ± 1.1 99.5 ± 0.5 100 ± 0.4 * Measured by Helium pycnometery. † Measured by water buoyancy method. 119 Chapter 5. Sintering of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 5.4. Conclusions The addition of CNTs or carbon black to alumina significantly increases its sintering rate. The sintering temperature required to achieve full densification of alumina – CNT nanocomposites was reduced by 500 o C as compared to alumina. An improvement in the densification was also observed in reduced titanium dioxide – CNT nanocomposite by the addition of CNTs. Clear evidence is presented of the effect of CNTs on grain growth. The CNTs, which form entangled networks at the grain boundaries, produce significant grain growth retardation. Using this effect, an alumina / nanocomposite laminate structure with a grain size difference of about ten times was successfully co-sintered. However, the grain growth retardation was not evident when alumina coated CNTs were employed in the alumina matrix. The effect of residual impurities (left during synthesis) was not observed on the densification of alumina – CNT nanocomposites. CNTs should not only be considered as an additive for improving the properties of ceramics because of their excellent intrinsic physical properties, but also as a means of controlling their sintering behaviours and microstructures. This will allow materials with improved and novel microstructures to be fabricated. This includes nanoceramics, co-fired ceramic multilayered structures, and functionally gradient materials. 120 Chapter 8. Mechanical properties of ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) 6.1. Introduction Significant improvements in the electrical and mechanical properties of polymer [31, 329] and ceramic [98, 111, 127, 140, 146, 179, 330] based CNT nanocomposites were reported. CNTs are well preserved in polymers due to low processing temperatures and pressures [31, 329]. However, to date, there is a lot of controversy about the stability of CNTs during high temperature processing in ceramic composites [111, 112, 134, 146, 164, 179, 181, 330-333]. It is difficult to examine the exact amount of structural defects in CNTs. Raman spectroscopy is a well-known tool to characterise graphitic carbons [334, 335]. Different types of Raman vibrations of CNT are illustrated in figure 6.1. The peak at ~100-400 cm–1 corresponds to a Radial Breathing Modes (RBMs). RBMs can provide information about CNT chirality (i.e., (n,m) indices) and diameter distributions since the RBM frequency is inversely proportional to nanotube diameter [181, 336]. However, RBMs signals (~100-400 cm–1) are very weak for thicker MWNTs. RBMs cannot be used for quality assessment of MWNTs [335, 337]. The peak at ~13201350 cm–1 corresponds to a disorder–induced phonon mode (D band) of MWNTs, and 121 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites the peak at ~1500-1600 cm–1 can be assigned to the G band of MWNTs or tangential stretching of sp2 bonded carbon atoms in a two-dimensional hexagonal lattice [179, 226, 335]. The D band is the signature of defects. The D band is a double–resonance Raman mode, which provides a measure of the structural disorder produced by amorphous carbon and any defects [335]. Local defects in the walls of the CNTs lead to a reduction in the activation energy and lower oxidation temperatures [335]. According to Tuinstra and Koenig [338], an increase in ID/IG corresponds to an increase in the amount of “unorganized” carbon and/or decrease in the mean crystal size [335]. ID/IG > 2 indicates a highly disordered form of carbon [338, 339]. There is a lot of experimental evidence that any interfacial interaction with a matrix [226] and sidewall derivatisation [340] of CNTs significantly increase ID/IG. Moreover, during graphitisation, the ID becomes smaller than IG which indicates a more improved graphene structure [341]. Therefore, the authors used ID/IG ratios supported with highresolution transmission electron microscopic analysis to study the degradation of CNTs. Radial Breathing Modes (RBM) -1 Range: 100-400 cm Tangential stretching (G band) -1 Range: 1500-1600 cm Defects/ disorders (D band) -1 Range: 1320-1350 cm SWNT/ DWNT SWNT/ DWNT/ MWNT SWNT/ DWNT/ MWNT Figure 6.1. Schematics of different Raman vibration in CNT. Figure modified from [342, 343]. 122 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites During high temperature processing of CNTs and ceramic – CNT nanocomposites, many different outcomes have been reported. This is mainly attributed to the difference in processing and characterisation methods. Menchavez et al. [330, 331] reported the formation of some refractory phase, aluminium carbide, during the sintering of alumina – carbon nanocomposites at 1700 oC. Rul et al. [146] reported damage of the SWNT structure during hot–pressing in vacuum, which caused a decrement in the electrical conductivity of magnesium aluminate – CNT nanocomposite. Jiang et al. [179] reported conversion of SWNT to graphite at higher sintering temperatures (>1150 oC) due to the disappearance of a shoulder on the G band peak in the Raman spectra. This resulted in a decrease in the fracture toughness of the alumina – CNT nanocomposite [179]. Poyato et al. [286] reported that the SPS process (1550 oC/ 40 MPa/ 3 minutes) is responsible for selective destruction of SWNTs, and the conversion of some SWNTs into disordered graphite, diamond, and carbon ‘nano-onions’. In another study, diamond particles were found with diameters close to 10 μm after SPS of CNTs at 1500 oC for 20 minutes [344]. Diamond particles (figure 6.2) were formed from CNTs by nucleation from the cores of carbon nanoonions, which are formed from CNTs [344]; subsequently these crystals grow as the sintering time increases [345]. Many other techniques have also been successful in synthesising diamond from CNTs. These include laser irradiation [346], shock waves [347] and radio-frequency hydrogen plasma [348]. (a) (b) Figure 6.2. Diamond particles formed during SPS of CNTs [344, 345]. 123 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites In contrast, Thomson et al. [181] found that the SWNT structure was preserved in alumina nanocomposites at SPS temperatures up to 1250 oC, even when a very high pressure (105 MPa) was used. Wang et al. [127] reported SWNT in an alumina composite were undamaged when processed by SPS at 1550 oC. Recently, Wei et al. [332] reported a 200% improvement in the fracture toughness (indentation fracture toughness) of alumina – 1 wt% SWNT as compared to monolithic alumina. The nanocomposite was hot–pressed at 1600 oC for one hour. MWNTs are less sensitive to high temperature degradation as compared to SWNTs because of their concentric shells [349]. MWNT bulk samples were prepared by SPS at 1600 oC/ 60 MPa/ 1 minute and were found to be well-preserved at this temperature [350]. Ye at al. [112] reported an increase in fracture toughness and fracture strength of bariumaluminosilicate glass – 10 vol% MWNT after hot–pressing the composite at 1600 oC/ 20 MPa/ 1 hour. In another report, MWNTs were found well preserved at 1500 oC / 40 MPa/ 1 hour based on Raman spectroscopy data [111]. Raman spectrum of sintered MWNTs shows the intensity of the G peak was about 1.3 times stronger than that of the D peak, which demonstrated that the graphite sheet structure of MWNTs was markedly improved after hot press sintering [111]. Balaszi et al. [333] sintered silicon nitride – MWNT nanocomposite by hot isostatic pressing (HIP) at 1700 oC/ 3 hours and reported preserved MWNTs as observed by showing HR-TEM images. Ma et al. [134] hot-pressed silicon carbide – 10 wt% MWNTs at 2000 oC for 1 hour (25 MPa in Ar environment). The MWNTs were not damaged in the composite and 10% improvements in bending strength and fracture toughness over monolithic SiC were reported. In some instance CNTs were subjected to very high temperatures. MWNTs were clearly observed by Laha et al. [164] after plasma spraying blended powder (Al – S – MWNTs) to a rotating metallic mandrel. In the same report, MWNTs were sprayed at very high temperatures (9,700 – 14,700 oC), but for very short durations. Also, the graphitisation of CNTs starts from 2000 oC in an inert atmosphere or vacuum. In the graphitisation process, defects are removed, leading towards a more perfect graphene cylinder, which can have a large radius curvature [341]. The graphitisation 124 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites temperature of carbon can be reduced to 1900 oC by the addition of alumina [351] and to 1000 oC by the addition of catalytic metals (Fe, Co and Ni) [352]. Alumina is the most common structural ceramic [9] and very popular ceramic matrix for CNTs [98, 111, 127, 140, 179, 181, 286, 330-332]. So far there has been no systematic study of the stability of CNTs during high temperature processing in alumina. To study the stability of the CNTs in ceramics that require very high temperature (>2000 oC) and long sintering times, the authors SPSed boron carbide – CNT nanocomposites and boron nitride – CNT nanocomposites. To date, these nanocomposites have not been sintered to full density. These ceramics would be difficult to sinter by conventional techniques because of their high sintering temperatures. This chapter studies the structure of CNTs after SPS of alumina – CNT, boron carbide – CNT and boron nitride – CNT nanocomposites. 6.2. Experimental procedure 6.2.1. Starting materials CNTs were supplied by Nanocyl, Belgium. Details of alumina, boron carbide and boron nitride are provided elsewhere (section 3.2.4). 6.2.2. Spark Plasma Sintering (SPS) Composite powders were prepared by colloidal dispersion method (using DMF, section 4.2.1). Bulk CNTs, ceramic and nanocomposite pellets (diameter 20 mm and thickness 2 mm) were prepared by SPS. A pressure of 80-100 MPa was applied concurrently with the heating (rate 300 oC/ minute) and released at the end of the sintering period for all samples. All powder compacts were cold pressed at 0.64 MPa 125 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites for a few seconds before SPS. Samples SPSed for this chapter are summarised in table 6.1. Table 6.1. Samples SPS processed for this chapter. Matrix CNT weight % SPS conditions Rel. theoretical density (%) Alumina 0 1800 oC/ 100 MPa/ 3 minutes ~100 Alumina 5 1110 oC/ 100 MPa/ 3 minutes ~98.5 Alumina 5 1400 oC/ 100 MPa/ 3 minutes ~100 Alumina 5 1600 oC/ 100 MPa/ 3 minutes ~100 Alumina 5 1600 oC/ 100 MPa/ 13 minutes ~100 Alumina 5 1600 oC/ 100 MPa/ 20 minutes ~100 Alumina 5 1600 oC/ 100 MPa/ 30 minutes ~100 Alumina 5 1700 oC/ 100 MPa/ 3 minutes ~100 Alumina 5 1900 oC/ 100 MPa/ 3 minutes ~100 Boron carbide 0 2000 oC/ 80 MPa/ 20 minutes ~99 Boron carbide 5 1400 oC/ 80 MPa/ 20 minutes ~85 Boron carbide 5 1600 oC/ 80 MPa/ 20 minutes ~93 Boron carbide 5 1800 oC/ 80 MPa/ 20 minutes ~100 Boron carbide 5 2000 oC/ 80 MPa/ 20 minutes ~100 Boron nitride 5 2175 oC/ 80 MPa/ 20 minutes ~97.5 Bulk CNT - 1000 oC/ 100 MPa/ 3 minutes - Bulk CNT - 1000 oC/ 80 MPa/ 20 minutes - Bulk CNT - 1500 oC/ 80 MPa/ 20 minutes - Bulk CNT - 2000 oC/ 80 MPa/ 20 minutes - 126 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 6.2.3. Material characterisations All of the sintered samples were ground using SiC paper down to 4000 grit. The density of the ground samples was measured using the Archimedes’ method. Field emission scanning electron microscopy (FE-SEM) was used to observe fractured surfaces. The fractured surfaces were coated with a very thin layer of gold and transferred to an FE-SEM for examination. High-resolution transmission electron microscopy (HR-TEM) was carried to study the structure of CNTs after SPS. The electron transparent nanocomposite films (thickness less than 100 nm) were prepared by mechanical grinding, polishing, dimpling and focus ion milling. Bulk CNT SPS processed disks were cut and specimens were scratched with a razor blade from the centre. The scratched specimens were ultrasonicated in ethanol. A drop of suspension was left to evaporate on a carbon coated copper grid for HR-TEM analysis. 6.2.4. Raman spectroscopy Structural characterisation of CNTs in ceramic nanocomposites was performed by Raman spectroscopy. A Nicolet Almega Dispersive Raman Spectrometer was used. Raman spectra were excited with a 488 nm Ar+ laser line at a power of 35 mW. Spectra were detected with an imaging photomultiplier (1024 x 1024) with 5 cm-1 resolution. All samples were cut and Raman laser was focused on the different areas of the cross section. Typical collection times were 4 minutes. At least 10 locations were examined to determine the ratio of the intensities of the D and G band. 6.2.5. X-Ray Diffraction (XRD) analysis The X-Ray Diffraction (XRD) patterns of different phases were obtained by using Cu Kα radiation, generated with X’PERT PRO (Phillips) at 45 kV, 40 mA. The data was compared with JCPDS standards using the in-built software. 127 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 6.3. Results and discussion 6.3.1. SPS of alumina and alumina – CNT nanocomposites Raman spectra of alumina – 5 wt% CNTs processed at different conditions is shown in figure 6.3. Alumina has no Raman signals for the given range (1100-1800 cm–1) that may interfere with the Raman signals of CNTs. Slight peak shifts indicate the presence of residual stresses [18, 20, 24, 39]. Figure 6.4 shows the intensity ratio of the D and G bands. The CNTs were well preserved in the composite as ID/IG is < 2 for all samples. It should be noted that sample prepared at 1400 oC and above were 100% dense. o Composite (1900 C) o Relative Raman Intensity Composite (1700 C) o Composite (1400 C) o Composite (1100 C) o Alumina (1800 C) D G CNT -1 Wavelength (cm ) Figure 6.3. Raman spectra of CNTs, alumina and alumina – 5 wt% CNT nanocomposites. Alumina and nanocomposites were SPSed at 100 MPa for 3 minutes. 128 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 6.4. Intensity ratio (ID/IG) for CNT and alumina – 5 wt% CNT nanocomposites. All nanocomposites were SPSed at 100 MPa for 3 minutes. For 1600 oC, refer figure 6.5. Figure 6.5 shows the stability of CNTs with dwell time when sintered at 1600 oC. The CNTs were well preserved in the nanocomposite sintered for 3 minutes (figure 6.6). The ratio ID/IG was observed to be greater than 2 for dwell times of more than 13 minutes. The structural stability of CNTs is dependent on the sinter dwell time. 129 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Relative Raman Intensity 30 minutes 20 minutes 13 minutes 3 minutes D G CNT Wavelength (cm-1) Figure 6.5. Raman spectra of CNTs and alumina – 5 wt% CNT nanocomposites. All nanocomposites were sintered at 1600 oC/ 100 MPa. Figure 6.6. Intensity ratio (ID/IG) for CNT and alumina – 5 wt% CNT nanocomposites. All nanocomposites were SPSed at 1600 oC/ 100 MPa. 130 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 6.7 shows the HR-TEM image of the composite SPSed at 1900 oC/ 100 MPa/ 3 minutes. Electron diffraction patterns from the different areas confirmed the structural preservation of CNTs after SPS. It was difficult to visualize individual CNT as they were overlapping with each other (figure 6.8). Alumina CNT Figure 6.7. HR-TEM of alumina – 5 wt% CNT nanocomposites showing electron diffraction patterns of different areas. SPSed at 1900 oC/ 100 MPa/ 3 minutes. The grain boundaries and ceramic – CNT interfaces play an important role in the resulting properties of the ceramic composites. From HR-TEM analysis (figure 6.7 and figure 6.8) and XRD studies (figure 6.9), no grain boundary or other phases were observed. All XRD peaks observed (figure 6.9) were representing α alumina (corundum). 131 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites CNTs CNTs Alumina Alumina Figure 6.8. HR-TEM of alumina – 5 wt% CNT nanocomposites showing agglomerates of CNTs at the grain boundary. Alumina – 5 wt% CNT CNTs x 10 times Alumina Figure 6.9. XRD analysis of CNTs, alumina (SPSed) and alumina – 5 wt% CNT nanocomposite (SPSed). SPSed at 1900 oC/ 100 MPa/ 3 minutes. 132 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 6.3.2. SPS of other ceramics and their CNT nanocomposites Using SPS, we prepared 100% dense boron carbide – CNT nanocomposite without adding any sintering aid. FE-SEM inspection of the fracture surface showed fibrous CNTs (figure 6.10). Figure 6.10 shows a FE-SEM image of a sample that was SPSed at 2000 oC/ 80 MPa/ 20 minutes. (a) (b) Figure 6.10. FE-SEM image of boron carbide – 5 wt% CNT nanocomposite. SPSed at 2000 oC/ 80 MPa/ 20 minutes: a) at lower magnification; and b) at higher magnification. Raman spectroscopy (figure 6.11) revealed that when a holding time of 20 minutes was used, the CNTs were structurally degraded (ID/IG increase) with the increase in processing temperature. ID/IG > 2 indicates a highly disordered form of carbon [35], which was observed at SPS temperatures >1400 oC (figure 6.12). In figure 6.11, at temperature > 1800 oC, no shoulder (~1060 cm-1) was observed in boron carbide peak. This may be an indication of CNTs reaction with boron carbide that is not the subject of this chapter. At 2000 oC, no Raman peak for CNT was detected, which indicates that there was severe structural degradation of the CNTs during SPS. 133 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites o Composite (2000 C) o Relative Raman Intensity Composite (1800 C) o Composite (1600 C) o Composite (1400 C) o Boron carbide (2000 C) D G CNT Wavelength (cm-1) Figure 6.11. Raman spectra of CNTs, boron carbide (SPSed) and boron carbide – 5 wt% CNT nanocomposites (SPSed). SPSed at 80 MPa for 20 minutes. Figure 6.12. Intensity ratio (ID/IG) for CNT and boron carbide – 5 wt% CNT nanocomposites. All nanocomposites were SPSed at 80 MPa for 20 minutes. 134 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites To observe the severity of degradation at high temperatures, boron nitride – 5 wt% CNTs was SPSed using the highest furnace operating temperature of 2175 oC. To produce dense boron nitride, a high sintering temperature because of its inherent strong covalent bonding [51]. No Raman peaks of CNTs were observed in boron nitride – CNT nanocomposite. However, like boron nitride – CNT nanocomposite, the CNTs were found to be fibrous and with a high aspect ratio (figure 6.13). (a) (b) Figure 6.13. FE-SEM analysis of boron nitride – 5 wt% CNTs (rel. theoretical density: ~97.5%). SPSed at 2175 oC/ 80 MPa/ 20 minutes: (a) at lower magnification; and (b) at higher magnification. 6.3.3. SPS of bulk CNTs To analyze the degree of structural deterioration of the CNTs during high temperature processing, monolithic bulk CNTs disks were SPS processed. Fabricating bulk CNT for various applications is not new [350]. However, none of the previous works discussed the structure of the CNTs after the hot pressing. FE-SEM analysis of bulk CNTs prepared at different temperatures are shown in figure 6.14. 135 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) (c) (d) Figure 6.14. FE-SEM images of bulk CNTs SPS processed at different temperatures. a) Pressed at room temperature; b) SPS processed at 1000 oC; c) at 1500 oC; and c) at 2000 oC. CNTs appear thick as the function of SPS temperature (figure 6.14). The CNTs used in this study were >90% pure. It can be concluded that the impurities (catalytic metals and oxides) might have reduced the graphitisation temperature of the CNTs and graphitisation caused the formation of additional graphene layers on CNTs. This is previously reported in the literature [351, 352]. It was proposed that during hot press sintering, the graphite sheets will grow along their original orientations (the axial and circumferential directions) with diffusions of the carbon atoms, and the growing graphite sheets may join together to form bigger sheets [111]. However, HR-TEM analysis (figure 6.15) shows that CNTs are well deformed, de-shaped and wavy at higher SPS temperatures and pressures (1000 oC/ 100 MPa). Because of the limited FE-SEM resolution, the deformed CNTs appeared thick in figure 6.14. This is more 136 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites clear in figure 6.16, where CNTs were SPS processed at 2000 oC/ 80 MPa/ 20 minutes. (b) (a) Figure 6.15. HR-TEM images of bulk CNTs: a) as received; b) SPS processed at 1000 oC/ 100 MPa/ 3 minutes. Amorphous carbon (a) CNT shells Maintaining high aspect ratio (b) Figure 6.16. HR-TEM of CNT SPS processed at 2000 oC/ 100 MPa/ 20 minutes: inset a) nano-onion; and inset b) CNT after SPS, maintaining its aspect ratio. 137 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites During high-pressure SPS, the ends of the CNTs are more likely to deform. The stress concentrates in defective areas damaging the end caps and the defective regions along the nanotubes axis. Electron diffraction pattern of different areas of bulk CNTs SPS processed at 2000 oC revealed partly damaged CNT (figure 6.16). These observations are well supported with Raman spectroscopy (figure 6.17). A very slight peak shift was observed in the Raman spectroscopy, which is an indication of structural transformations. Intensity of D band line is lower than that of G band line, which indicates high temperature graphitisation, leading towards large CNTs [341]. This is well supported with the FE-SEM result (figure 6.14). At these conditions, CNTs are going through a series of transformation as reported in the literature. Here, such transformations are not fully complete to reveal a clear microscopic evidence. It can be graphitisation and/or transformation into nano-onions and diamond. Some HRTEM observations revealed that the layers of the outer shells of the nanotubes break and transform into curled graphitic structures, termed as nano-onions (figure 6.16, inset a). Relative Raman Intensity Amorphous carbon 1332 Cubic diamond o 2000 C o 1500 C o 1000 C Raw form Wavelength (cm-1) Figure 6.17. Raman spectra of CNTs SPS processed at different conditions. All samples were processed at 80 MPa for 20 minutes. 138 Chapter 6. Structural Stability of CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites In figure 6.17, the Raman signal for bulk CNTs SPS processed at 2000 oC shows a Raman peak at 1332 cm-1, which is a Raman signal for the cubic diamond phase (C–C sp3 bond) [353]. It should be noted that the motive of the work in this chapter was not to synthesize diamond from CNTs. Also, no diamond particle or electron diffraction pattern of diamond was observed during HR-TEM analysis. This is because of low processing durations and pressures. The Raman signal also shows amorphous carbon, which was observed in the HR-TEM analysis (figure 6.16). 6.4. Conclusions The structure of CNTs should be preserved in the ceramic matrix if they are to provide an effective reinforcement. CNTs were found to be preserved in alumina after being SPSed up to 1900 oC/ 100 MPa/ 3 minutes. In alumina and boron carbide matrices, structural degradation of CNTs started from 1600 oC. This can be avoided by the use of additional sintering aids that lowers the sintering temperature. CNTs maintained their high aspect ratio and fibrous nature after being SPSed in boron nitride at 2175 oC for 20 minutes. However, no Raman vibrations of CNTs were observed for nanocomposites processed at temperatures > 2000 oC. Structural preservation of CNTs in ceramic nanocomposites depends on the nature of ceramic, SPS temperature and dwell times. CNTs are not suitable for matrices that require higher temperatures (> 1600 oC) and longer processing times (> 13 minutes). CNTs went through a series of incomplete transformations during high temperature SPS processing of bulk CNTs. Partly amorphous CNT and nano-onions were very distinctly observed. Even in bulk form, CNTs maintained their high aspect ratio and fibrous nature after being SPSed at 2000 o C/ 80 MPa/ 20 minutes. Raman spectroscopy provided peaks for cubic diamond phase. However, in contrast to the literature, no diamond phase was seen in the electron microscopic analysis. This is because of low processing durations and pressures. 139 Chapter 6. Structural Stability CNTs in Ceramics (Preservation Studies) Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites 7.1. Introduction The addition of CNTs for improving electrical conductivity of ceramics is widely appreciated (table 2.2). Alumina is inherently an insulator (electrical conductivity: 10-13 S/m). Adding a small amount of CNT (~0.79 vol%) [96] to alumina can make it electrically conductive (10-4 S/m). A detailed literature review is presented elsewhere (section 2.4). In this chapter, the electrical conductivity of alumina – CNT and alumina – carbon black nanocomposites were compared. The effect of grain size and sintering conditions on the electrical conductivity of the alumina – carbon nanocomposites was also studied. For alumina – CNT nanocomposites, electrical conductivities were also measured as a function of temperature. 7.2. Experimental techniques MWNTs (supplied by Nanocyl, Belgium) and CB (supplied by Degussa, Germany) were used to prepare alumina nanocomposites. The preparation method is described in chapter 4 and 5. All nanocomposites were fully dense (rel. theoretical density: ~100%). Samples prepared for this chapter are described in table 7.1. 140 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table 7.1. Samples SPSed for this chapter#. # Matrix Filler (wt %) SPS conditions Rel. theoretical density (%) Alumina - 1800 oC/ 100 MPa/ 3 minutes ~100 Alumina CB (2) 1400 oC/ 100 MPa/ 3 minutes ~84 Alumina CB (2) 1600 oC/ 100 MPa/ 3 minutes ~98 Alumina CB (2) 1800 oC/ 100 MPa/ 3 minutes ~99 Alumina CNT (2) 1200 oC/ 100 MPa/ 3 minutes ~78 Alumina CNT (2) 1400 oC/ 100 MPa/ 3 minutes ~88 Alumina CNT (2) 1600 oC/ 100 MPa/ 3 minutes ~97 Alumina CNT (2) 1800 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (3.5) 1000 oC/ 100 MPa/ 3 minutes ~67 Alumina CNT (3.5) 1100 oC/ 100 MPa/ 3 minutes ~84 Alumina CNT (3.5) 1200 oC/ 100 MPa/ 3 minutes ~98 Alumina CNT (3.5) 1400 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (3.5) 1600 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (3.5) 1800 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (5) 1200 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (5) 1400 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (5) 1400 oC/ 100 MPa/ 7 minutes ~100 Alumina CNT (5) 1400 oC/ 100 MPa/ 13 minutes ~100 Alumina CNT (5) 1400 oC/ 100 MPa/ 20 minutes ~100 Alumina CNT (5) 1600 oC/ 100 MPa/ 3 minutes ~100 Alumina CNT (5) 1800 oC/ 100 MPa/ 3 minutes ~100 For CNTs, 2 wt% = ~4.7 vol%, 3.5 wt% = ~8 vol% and 5 wt% = ~11.2 vol% 141 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The electrical conductivities of the samples were measured (section 3.3.4) with a high sensitivity digital micro-ohmmetre (Keithley 580) using the two-point method on silver electroded specimens (3 mm × 3 mm× 3 mm) prepared using a diamond cutting machine. Details are given elsewhere (section 3.3.4). A power supply (Keithley 2602) and digital multimeter (Keithley 6517A) were used to measure the current-voltage (IV) characteristics of the samples. 7.3. Results and Discussion 7.3.1. CNT vs. carbon black The DC electrical conductivities of alumina – CNT and alumina – carbon black nanocomposites are shown in figure 7.1. The conductivity mechanisms involved in CNT nanocomposites could be variable range hopping [212] or fluctuation-assisted tunnelling [213]. Figure 7.1. Electrical conductivities of alumina – carbon nanocomposites. 142 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The higher electrical conductivity of the ceramic – CNT nanocomposite is the result of the large aspect ratio of CNTs (~150) as compared to that of carbon black (~1). This is well supported in figure 7.1. The large aspect ratio of CNTs resulted in entangled network of conductive pathways (figure 7.2) on the grain boundaries of alumina, which did not occur in alumina – carbon black nanocomposites (figure 5.9b). As compared to alumina – 2 wt% carbon black (30 S/m), the electrical conductivity of alumina – 2 wt% CNT nanocomposites is 125 S/m, i.e. four times higher than alumina – carbon black nanocomposite. A larger CNT content resulted in higher electrical conductivity. The increased electrical conductivity is attributed to the presence of undamaged CNTs, due to the utilisation of SPS technique that allowed lower sintering temperatures and shorter sintering times. Figure 7.3 shows the voltage – current relations measured at room temperature for alumina – CNT nanocomposite. The current increased linearly with voltage for any instance, indicating a good ohmic behaviour. The slope of the current – voltage curve corresponds to the CNT content (figure 7.3). Higher CNT content in the nanocomposite resulted in a lower slope/ electrical resistance and vice versa. (a) (b) Figure 7.2. HR-TEM image of alumina – 5 wt% CNT nanocomposite: (a) CNTs around grains; and (b) percolating network highlighted. 143 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 7.3. Voltage vs. current measured at room temperature for alumina – CNT nanocomposites. Both nanocomposites were SPSed at 1800 oC/ 100 MPa/ 3 minutes. 7.3.2. Electrical conductivity as a function of grain size Figure 7.4 shows the electrical conductivity of alumina – carbon nanocomposites as a function of grain size. The grain sizes were coarsened by using higher sintering temperatures. For alumina – CNT nanocomposites, the electrical conductivity increased significantly with increasing grain size. In large grained nanocomposites, higher electrical conductivity is due to the availability of fewer paths for the current to flow. Large grains of alumina concentrate CNTs in fewer conductive paths. However, due to the particulate nature of carbon black, this was not observed for the alumina – carbon black nanocomposites (figure 7.4). With the growth of alumina grains, the particulates or agglomerates of carbon black may isolate, resulting in higher percolation threshold and lower electrical conductivity. Due to the fibrous nature and 144 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites aspect ratio of CNT, such isolation is not possible that resulted higher electrical conductivity for alumina – CNT nanocomposite. o SPSed at 1800 C o SPSed at 1600 C Figure 7.4. Electrical conductivities of alumina – carbon nanocomposites as the function of grain size. The grain sizes were coarsened by using higher sintering temperatures. Apart from addition of CNTs and using higher SPS temperatures, another way to tailor electrical conductivity of alumina – CNT nanocomposites is to use longer processing durations. In figure 7.5, alumina grains were slightly coarsened using longer SPS processing time at 1400 oC. The grain growth in alumina – 5 wt% is not significant (figure 5.11). In this context, this route can be adopted to slightly increase the electrical conductivity of these nanocomposites. For example, increasing processing time by 13 minutes resulted in a 12% increase in the electrical conductivity. However, using higher sintering temperatures (figure 7.4) proved to be 145 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites more effective in increasing the electrical conductivity as compared to using longer dwell times (figure 7.5). Figure 7.5. Electrical conductivities of alumina – 5 wt% nanocomposites as the function of SPS processing durations. CNTs were supplied by Nanocyl, Belgium. 7.3.3. Electrical conductivity as a function of temperature Figure 7.6 shows the electrical conductivity of alumina – CNTs nanocomposites as a function of temperature. It should be noted that CNTs are the only conductor in the nanocomposite. MWNTs show metallic behaviour [158]. This is evident from figure 7.6, where the electrical conductivity is decreasing with the rise in temperature. At a particular temperature, CNTs started oxidizing and a sharp decrease in the nanocomposite’s electrical conductivity was observed. For instance, alumina – 5 wt% CNT nanocomposite (SPSed at 1400 oC/ 100 MPa/ 3 minutes) showed an abrupt 146 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites decrease in electrical conductivity at 449 oC, where as alumina – 5 wt% CNT nanocomposite (SPSed at 1400 oC/ 100 MPa/ 20 minutes) showed such behaviour at 496 oC. This is due to higher oxidation resistance in large grained nanocomposites. In large grained nanocomposites, there is a small volume of grain boundary and it is difficult for more oxygen to diffuse in and react with the CNTs. In this way, the electrical conductivity of alumina – CNT nanocomposites can be improved. Alumina grain size o Electrical Conductivity (S/m) 496 C 701 nm 653 nm 364 nm o 471 C o 449 C Temperature (oC) Figure 7.6. Electrical conductivities of alumina – 5 wt% nanocomposites as the function of temperature. The grain sizes were coarsened by using longer processing durations. Heating rate: 2 oC/ minute. 7.4. Conclusions The electrical conductivity of alumina – CNT nanocomposites is four times higher as compared to alumina – carbon black nanocomposites for the same mass content of 147 Chapter 7. Electrical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites carbon due to the fibrous nature and high aspect ratio of CNTs. The conductive network structure allowed the percolation of CNTs at low volume fractions [96] and thus increased the electrical conductivity compared to alumina – carbon black nanocomposites. The electrical conductivity of alumina – CNT nanocomposite increased with increasing grain size due to the concentration of CNTs in fewer conductive paths. Because of the electronic properties of CNTs, the electrical conductivity of alumina – CNT nanocomposites varied with temperature for different grain sizes. 148 Chapter 8. Mechanical properties of ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites 8.1. Introduction Ceramics are used in high temperature applications [279]. In order to get most out of CNTs in ceramic – CNT nanocomposites, it is necessary to retain their structure and properties, and avoid degradation. The potentially improved properties of nanocomposites can then be explored. However, because of their low oxidation temperature resistance (~500 o C) of carbon [354], using ceramic – CNT nanocomposites at high temperatures is an obstacle for their commercial success. Most of previous reports about oxidation resistance of CNTs are based on polymer based CNT composites [355-357]. For improving electronic and electrical properties, ceramic layers were coated on CNTs [234, 358-361]. However, no comments were presented on the stability against oxidation of the coated CNTs. Wang et al. [362] coated 10 nm of silicon layer and reported an improvement of 105 oC in oxidation resistance. Li et al. [363] reported that the increase in Ni catalyst during CVD growth improves the oxidation resistance of multiwall CNTs. The catalyst composition had a significant effect on the CNT structure and stability, and is important for large-scale CNT synthesis [363]. To date, there are no reports on the effect of the ceramic matrices and ceramic coatings on the oxidation resistance of CNTs. Coating CNTs by atomic layer 149 Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites deposition (ALD) provides shielding that improves oxidative stability of the encapsulated CNTs. This chapter reports the high temperature protection effect of alumina coating on CNTs. It also reports the effect of grain boundary area on the oxidative stability of CNT in alumina – CNT nanocomposites. 8.2. Experimental techniques MWNTs supplied by Nanocyl, Belgium and NanoDynamics, USA were used in this chapter. The preparation method is described in chapter 4 and 5. All nanocomposites were fully dense (rel. theoretical density: ~100, table 8.1). The oxidation resistance of CNTs in different systems was characterised by Thermo Gravimetric Analysis (TGA) as described in section 3.3.7. Table 8.1. Alumina – 5 wt% CNT nanocomposites SPSed for this chapter. Alumina coating Weight % CNT supplier SPS conditions Rel. theoretical density (%) No 5 Nanocyl, Belgium 1200 oC/ 100 MPa/ 3 minutes ~100 No 5 Nanocyl, Belgium 1800 oC/ 100 MPa/ 3 minutes ~100 8.3. Results and Discussion 8.3.1. Oxidation of CNTs Figure 8.1 shows the oxidation behaviour of CNTs obtained from different suppliers. 477 oC is the onset of oxidation (intersection point of first two slopes in figure 8.1) for CNTs provided by NanoDynamics, USA and 507 oC is the onset of oxidation for CNTs provided by the other source. CNTs provided by NanoDynamics, USA have larger average diameter (~20 nm) as compared to the CNTs provided Nanocyl, 150 Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Belgium (~12 nm). CNTs provided by NanoDynamics, USA showed slightly less oxidation resistance as compared to CNTs provided by Nanocyl, Belgium. The oxidative stability of CNTs is influenced by defects [363] and nanotube diameter [363, 364]. Oxygen molecules react easily with larger surface areas, resulting in decreased oxidative stability of CNTs (supplied by NanoDynamics, USA). The oxidation of CNTs is not rapid and acute like combustion, which is also evident in other report [365]. Because the kinetic energy of oxygen varies with temperature, there is not a critical temperature when the oxidation of CNTs starts [365] as shown in figure 8.1. During the initial stage of TGA, all samples showed a slight mass loss due to the presence of amorphous carbon, as reported in the literature [363, 366]. In the second stage of TGA, the curve slope remains almost the same in the definite temperature range for both types of CNTs. In the third stage of TGA, there was no weight gain observed during thermal treatment, since no oxidation of the impurities occurred. The weight loss for both types of CNTs was not 100 % due to the presence of impurities (section 5.3.5). o 477 C 507 oC Figure 8.1. TGA of raw CNTs from different suppliers. Heating rate: 5 oC/ minute. 151 Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 8.3.2. Oxidation of alumina coated CNTs To improve the oxidation resistance of CNTs, the CNTs were coated with alumina [250]. The oxidative stability of CNTs was distinctly improved due to the protective alumina coating (figure 8.2). o 553 C o 477 C 530 o C Figure 8.2. TGA of uncoated CNTs and alumina coated CNTs. CNTs were supplied by NanoDynamics, USA. Heating rate: 5 oC/ minute. The onset oxidisation temperature for sample coated for 50 ALD cycles is now as high as 553 oC in air atmosphere, which is 76 oC higher than that of uncoated CNTs. Once the oxidation started, the degradation was also reduced to 0.41%/oC, compared to 0.92 %/oC for uncoated CNTs. The degradation process was delayed because it became more difficult for oxygen molecule to approach CNTs after coating. By analyzing the third stage (> 700 oC, figure 8.2), it is possible to quantify the mass content of the alumina on the coated CNTs (figure 8.3a). The coated CNTs with 25 152 Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites ALD cycles had a residual mass of 56.4 % of CNTs of the original mass and the 50 ALD cycles had a 44.9 % residual mass. A thicker alumina coating could further inhibit the oxidisation of CNTs but it may decrease the mechanical properties of CNTs by making the coated CNTs brittle. After cooling down the TGA furnace, white coloured alumina nanotubes (figure 8.3b) were left in the platinum pan, which were previously surrounding CNTs. This could be a route to mass-produce alumina nanotubes (figure 8.4). (a) (b) Figure 8.3. Platinum pan used for TGA: a) alumina coated (50 ALD cycles) CNTs before oxidation; and b) alumina nanotubes left after oxidation of coated CNTs. CNTs were supplied by NanoDynamics, USA. 153 Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 8.4. Alumina nanotube left after the oxidation of coated CNTs. CNTs were supplied by NanoDynamics, USA: a) at lower magnification; and b) at higher magnification. 8.3.3. Oxidation of alumina – CNT nanocomposites Dense 5 wt% CNT (uncoated) dispersed alumina nanocomposites with different grain sizes were fabricated using SPS. In figure 8.5, the onset oxidisation temperature for sample sintered at 1800 oC is 588 oC in air, which is 81 oC higher than that of raw CNTs. Once the oxidation started, the degradation was 0.026%/oC, which is 97% less than that of raw CNTs. Oxidative reactivity in these nanocomposites is influenced by the grain boundary area. The sample sintered at 1200 oC and 1800 oC have grain sizes of 0.15 μm and 0.66 μm (figure 5.10b), respectively. The larger grain size material therefore had a smaller total area of grain boundaries, making the ingression of oxygen slower and increasing the oxidation resistance. Thus the oxidative stability of alumina – CNT nanocomposites can be tailored by changing the grain boundary area. 154 Chapter 8. Oxidative Stability of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites o 588 C o 547 oC 507 C Figure 8.5. TGA of raw CNTs and SPSed alumina – 5 wt% uncoated CNT nanocomposites. CNTs were supplied by Nanocyl, Belgium. Heating rate: 5 oC/ minute. 8.4. Conclusions It is necessary to preserve the chicken wire hexagonal distribution of CNTs in ceramics for high-temperature applications. It was observed that TGA is a good tool to evaluate the mass content of CNTs in the coated CNT and SPSed nanocomposites. Coating CNTs by ALD provides shielding that improves oxidative stability of the encapsulated CNTs. A thicker alumina coating could further inhibit the oxidisation of CNTs and enhance the thermal stability of CNTs. Coarser grained materials have higher oxidation resistance of CNT due to the presence of fewer grain boundaries. The oxidative stability of alumina – CNT nanocomposites can be tailored by changing the grain boundary area. SPS is a good processing method to modify the grain size (and grain boundary area) of alumina – CNT nanocomposites. 155 Chapter 8. Oxidative stability of Ceramic – CNT nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites 9.1. Introduction Ceramics are inherently brittle. Zhan et al. [138, 140] SPSed alumina – SWNT nanocomposites and reported a threefold improvement in the fracture toughness (section 2.3.1). Since then, a number of research groups have investigated various routes to fabricate strong and tough ceramic – CNT nanocomposites [110-112, 134, 137, 141-145, 153, 161-163, 168, 192, 332]. A detailed literature review is presented in chapter 2. However, to date, the role of CNTs on the mechanical properties of ceramic nanocomposites remains inconclusive. This is because of the usage of different types and compositions of CNTs, inhomogeneous distribution of CNTs and differences in experimental conditions. In many reports [106, 109, 125, 131, 134, 332], the presence of good quality CNTs was not confirmed after conventional processing at high temperatures. This is also one of the reasons for variation in the mechanical properties of CNT reinforced ceramic nanocomposites. The structural stability of CNTs after sintering was confirmed in chapter 6. In this chapter, alumina, alumina – carbon black and alumina – CNTs nanocomposites were prepared using SPS. Hardness, indentation fracture toughness, flexural modulus, flexural strength and thermal shock resistance (qualitative assessment) were analysed. Good interfacial bonding is required to achieve load transfer across the CNT-matrix interface, a condition necessary for improving the mechanical properties of ceramic nanocomposites [176]. The toughening enhancement produced by ceramic coated 156 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites CNTs was much better than that resulted from application of either CNTs or nanometre ceramic powders individually [187]. To improve interfacial bonding, alumina coated CNTs were used. The mechanical properties of nanocomposite reinforced with alumina coated CNTs are compared with nanocomposite reinforced with CNTs. The outcomes of these comparisons are discussed in relation with their microstructures. 9.2. Experimental techniques MWNTs (supplied by NanoDynamics, USA and Nanocyl, Belgium) and CB (supplied by Cabot, USA) were used to prepare alumina nanocomposites. All coated CNTs were coated for 50 cycles of ALD, which resulted in ~5-10 nm coating of alumina (figure 3.1e). All materials were prepared as explained in chapter 4 and 5. Alumina and alumina – carbon nanocomposites were sintered at 1800 oC/ 100 MPa/ 3 minutes. To exclude the effect of porosity on the mechanical properties, all specimens were fully densified (rel. theoretical density: ~100%) using SPS. The sintering conditions used for all samples were the same for the sake of comparison. The high residual stresses at the surface of alumina grains may cause the nanotube/ matrix interface debonding [196]. To avoid the influence of residual stresses on the mechanical properties, a relatively slow cooling rate (~50 oC/ minute) was adopted. Vickers hardness and indentation fracture toughness (IFT) were measured as described in section 3.3.5 using 2.5 kg load. For surface roughness analysis, all specimens were ground using SiC paper and diamond polished down to 1 μm. The data presented for Vickers hardness and IFT is the average values obtained from at least ten indentations on the same specimen. 157 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.2.1. Flexural bending Rectangular bars of dimensions ~ 2 x 1.5 x 25 mm were cut using a high speed diamond cutter. The flexural modulus was measured by non-destructive 3-point bending method using Dynamic Mechanical Analyzer (DMA Q800, TA Instruments). A small load (up to 18 N) was applied and deflection of the beam was measured. The flexural modulus was calculated by measuring the slope of the load-deflection curve. The technique is quite accurate and reliable. Gou [367] reported similar values of flexural modulus for different ceramic materials using an Impulse Excitation Technique (IET) and flexural bending method. Flexural strength was evaluated by 3-point bending in accordance with ASTM C1161-02c [368]. For achieving good results, a special jig was used (figure 9.1). The bending tests were performed at room temperature on an Instron 6025 using a load cell of 1 kN. A cross-head speed of 0.2 mm/ minute was used for all specimens. Sample dimensions were ~ 2 x 1.5 x 25 mm. All edges and corners on these samples were chamfered on 4000 grit size SiC grinding paper in order to avoid stress concentration points (figure 9.2). Configuration A [368] was used for all specimens. The flexural strength of the specimens was determined from the failure load and the geometry of the test piece using equation 9.1. S= 3PL 2bd 2 Equation 9.1 Where, S = flexural strength, P = maximum loading level, L = 20 mm, outer (support) span, b = specimen width and d = specimen thickness. The data presented for fracture strength and flexural modulus is the average values obtained from at least ten test specimens of the same composition. Fractographic analysis was performed on selected fractured samples after flexural testing to identify the cause of failure. The origin of the fracture was identified by carefully monitoring the surfaces under an optical microscope and a field emission SEM. 158 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Figure 9.1. Special jig for improved 3 – Figure 9.2. Chamfered edges to avoid point flexural testing. stress concentration points on corners. Sample thickness: 1.5 mm. Sample: Fractured alumina surface. 9.2.2. Thermal shock resistance during SPS (observation) A SPS furnace was used to compare the thermal shock resistance of alumina and alumina – carbon nanocomposites. This analysis is possible by rapidly cooling a thin sample after sintering. Dried alumina powder (0.07 cm3) was poured into a carbon die and cold pressed at 0.62 MPa for 5 s before sintering. Nanocomposite discs (thickness 0.2 mm and diameter 20 mm) were prepared by SPS. A pressure of 100 MPa was applied concurrently with the heating (rate 300 oC/ minute). The sample was sintered at 1800 oC. After sintering, temperature was dropped to 450 oC using a cooling rate of 100 – 300 oC/ minute. A constant load of 100 MPa was used while cooling. Application of pressure while cooling is necessary in order to remove heat from the system. The carbon die was given sufficient time for cooling. The sintered material was then carefully taken out from the die and analysed for cracks. The same procedure was followed for alumina – carbon nanocomposites. 159 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.3. Results and Discussion 9.3.1. Surface finish Figure 9.3 shows the surface finish of different compositions of alumina – CNT nanocomposites prepared using CNTs supplied by Nanocyl, Belgium. All samples were polished in the same manner. (a) (b) (c) (d) Figure 9.3. Polished surfaces of alumina – CNT (Nanocyl, Belgium) nanocomposites containing: (a) no CNTs, thermally etched at 1500 oC for 10 minutes; (b) 2 wt% (~4.7 vol%) CNTs; (c) 5 wt% (~11.2 vol%) CNTs; and (d) 10 wt% (~21 vol%) CNTs. It is difficult to achieve good surface finish for alumina – CNT nanocomposites. Because of the strong covalent bonding in alumina and due to lubricating nature of carbon, there was poor interfacial adhesion at the ceramic/ CNT interface. On 160 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites polishing, alumina grains were selectively plucked out (figure 9.3b, 9.3c and 9.3d) because of poor interfacial adhesion. This is very obvious in nanocomposite containing 10 wt% CNT (figure 9.4). The arrows show the location of alumina grain before plucking (figure 9.4). Such surface will result in poor mechanical properties because of significant number of surface flaws. Figure 9.4. Polished surface of alumina Figure 9.5. Alumina coated CNT (50 – 10 wt% CNT nanocomposite. Arrows ALD cycles) encapsulated in an alumina showing alumina grain ex-sites. CNTs grain of nanocomposite. CNTs were were supplied by Nanocyl, Belgium. supplied by NanoDynamics, USA. Apart from surface flaws, high weight content of CNTs (10 wt%) yields an inhomogeneous dispersion of CNTs and a porous microstructure. Trapped gases in agglomerates reduce the mechanical and electrical properties of the nanocomposite [112, 177]. In this context, the maximum amount of CNT in alumina – CNT nanocomposites should be restricted to 5 wt%. However, when alumina coated CNTs were used, CNTs were located inside the grains (figure 9.5) and much better surface finishes could be achieved. 161 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.3.2. Vickers hardness Table 9.1 shows the grain size and Vickers hardness of alumina and alumina – carbon nanocomposites. Alumina is one of the hardest structural ceramics. The Vickers hardness of the nanocomposites decreased monotonously with increasing carbon content. It is quite well known that the hardness of alumina increases with decreasing grain size [369]. However, this is not true for alumina – carbon nanocomposites. Table 9.1. Vickers hardness of alumina and alumina – carbon nanocomposites. § Material Alumina coating Grain size (µm) Vickers hardness (GPa) Alumina - 43.20 18.9 ± 0.3 Alumina – 2 wt% CNT§ - 1.81 10.1 ± 0.3 Alumina – 2 wt% CNT# - 1.79 10.2 ± 0.3 Alumina – 2 wt% CNT# ~5-10 nm 3.23 14.9 ± 0.2 Alumina – 5 wt% CNT# - 0.66 12.0 ± 0.3 Alumina – 5 wt% CNT# ~5-10 nm 2.34 18.6 ± 0.2 Alumina – 5 wt% CB - 4.67 11.0 ± 0.3 CNTs were supplied by Nanocyl, Belgium # CNTs were supplied by NanoDynamics, USA The presence of a soft phase (hardness of MWNT in radial axis ~6 – 10 GPa [370]) eases the penetration of diamond indenter during testing. Such effect can be estimated by hardness rule-of-mixtures for composites [371] as indicated in equation 9.2: H vc = H vm (1 - v f )+ H vf v f Equation 9.2 Where, H vc = hardness of composite, H vm = intrinsic hardness of the matrix phase, v f = volume fraction of the filler phase and H vf = intrinsic hardness of the filler phase. It must be noted here, that equation 9.2 is a useful tool for approximation and 162 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites does not consider the dispersion and quality of CNTs. The reduced hardness values of the nanocomposites are due to the presence of very soft phase at the grain boundaries, which is nullifying the effect of fine grains. However, there may be some contribution coming from grain size when hardness values of alumina – 5 wt% CNT and alumina – 5 wt% CB nanocomposites are compared. The presence of alumina coating increased the hardness of alumina – CNT nanocomposites. The alumina coated CNTs are located inside the grains (figure 9.5 and figure 9.6a) unlike uncoated CNTs, which form a network of entangled CNTs on the grain boundaries (figure 9.6b). This gives the flexibility of making CNT based intragranular nanocomposites (figure 9.6a) and intergranular nanocomposite (figure 9.6b) or both. In alumina coated CNT nanocomposites, no soft lubricating phase is present at the grain boundaries. Higher frictional forces between alumina grains are resisting the penetration of indenter and grain sliding that is improving the hardness of the nanocomposite. (a) (b) Figure 9.6. Different types of fracture modes in alumina – 2 wt% CNT nanocomposites; (a) intragranular fracture mode in coated CNT nanocomposite (50 ALD cycles, NanoDynamics, USA); and (b) intergranular fracture mode in CNT nanocomposite (Nanocyl, Belgium). 163 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.3.3. Indentation Fracture Toughness (IFT) Table 9.2 shows the indentation fracture toughness (IFT) for different materials. Mechanical properties are influenced by the type of carbon nanotubes used in ceramic nanocomposites [107]. Zhan et al. [140] reported a value of 9.7 MPa√m (measured by IFT method) for alumina – 10 vol% SWNT. In the current study MWNTs were used. The internal shells of MWNT are unable to bond to the alumina matrix and therefore tensile loads are carried entirely by the external shell [140]. An increase in IFT could be obtained even by improving the quantity of carbon nanotubes. A 78% improvement in the IFT was found for alumina – 5 wt% CNT nanocomposite as compared to alumina. Figure 9.7a shows alumina having longer and wider radial cracks as compared to alumina – 5 wt% CNT nanocomposite (figure 9.7b). This is due to the presence of strong entangled network of CNTs at the grain boundaries (figure 9.6b). Such a phenomenon was not observed in alumina – CB nanocomposite (figure 5.9b). Table 9.2. Indentation fracture toughness of different materials. § Material Alumina coating Indentation Fracture Toughness (MPa√m) Alumina - 3.3 ± 0.2 Alumina – 2 wt% CNT§ - 5.7 ± 0.1 Alumina – 2 wt% CNT# - 5.6 ± 0.3 Alumina – 2 wt% CNT# ~5-10 nm 5.5 ± 0.5 Alumina – 5 wt% CNT# - 5.9 ± 0.3 Alumina – 5 wt% CNT# ~5-10 nm 5.8 ± 0.2 Alumina – 5 wt% CB - 3.4 ± 0.3 CNTs were supplied by Nanocyl, Belgium # CNTs were supplied by NanoDynamics, USA 164 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 9.7. Vickers indent after applying 2.5 kg load in: (a) alumina; and (b) alumina – 5 wt% CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Figure 9.8a shows the sub-surface damage caused by 2.5 kg loading on alumina and alumina – 5 wt% uncoated CNT. It was not possible to observe the same sub-surface optical effect for alumina – CNT nanocomposite because of the opaque nature of sample (figure 9.8b). (a) (b) Figure 9.8. Sub-surface damage analysis after Vickers indentation. Optical micrograph (dark field image) showing sub-surface cracking in: (a) alumina; and (b) alumina – 5 wt% uncoated CNT (Nanocyl, Belgium) nanocomposite. 165 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Sub-surface damage was analyzed by observing the cross-section of indented specimen under SEM (figure 9.9). The sample was carefully fractured to observe the cross-section. As compare to alumina (figure 9.9a), no major damage was observed for alumina – 5 wt% CNT nanocomposite (figure 9.9b). Significant redistribution of stresses by the strong entangled network of CNTs was the main reason for improved sub-surface damage resistance [127]. It prevented the formation of cracks that were observed for brittle alumina. These properties are important for ceramics used in contact-mechanical applications such as bearings, valves, nozzles, seals, wear parts, armour and prostheses [127]. (a) (b) Figure 9.9. Sub-surface damage analysis after Vickers indentation. Cross-section of: (a) alumina, showing major crack; and (b) alumina – 5 wt% uncoated CNT (Nanocyl, Belgium) nanocomposite, showing no major sub-surface damage. In uncoated CNT nanocomposites, fracture occurred along the grain boundaries (intergranular) and not within the alumina grains (transgranular). In coated CNT nanocomposite (figure 9.6a), the fracture path was intergranular and transgranular. Increasing the quantity of CNTs in alumina – coated CNT nanocomposite decreased the IFT further. Here transgranular fracture indicates weak grains and this is due to poor contact between coated CNTs and alumina grains. The objective of using alumina coated CNTs in alumina for improved IFT was not successful. 166 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.3.4. Flexural modulus Table 9.3 shows the flexural modulus for different materials. Previous publications [107, 116, 133, 200] reported a decrease in flexural modulus of ceramic – CNT nanocomposites due to poor density and damage caused to CNTs while processing. Chapter 5 confirms the improved density of raw CNT reinforced nanocomposite and chapter 6 confirms the structural stability of CNTs after SPSing. The decrease in flexural modulus is due to grain size refinement, lubricating nature of CNTs and poor adhesion between CNTs and alumina grains. The argument is also valid for alumina – CB nanocomposite. These properties may be useful in applications where mechanical flexibility and good electrical conductivity is desired. A detailed discussion on the applications of ceramic based CNT nanocomposites is presented in appendix B. In coated CNT nanocomposites, CNTs were individually encapsulated in coarse alumina grains (figure 9.5). Unlike uncoated CNT nanocomposite, no soft lubricating phase was present between alumina grains that resulted in slightly higher flexural modulus of coated CNT nanocomposites. Table 9.3. Flexural modulus of different materials. # Material Alumina coating Grain size (µm) Flexural modulus (GPa) Alumina - 43.20 308 ± 2.9 Alumina – 2 wt% CNT# - 1.81 234 ± 1.9 Alumina – 2 wt% CNT# ~5-10 nm 3.23 263 ± 3.9 Alumina – 5 wt% CNT# - 0.66 165 ± 2.1 Alumina – 5 wt% CNT# ~5-10 nm 2.34 185 ± 3.1 Alumina – 5 wt% CB - 4.67 201 ± 1.3 CNTs were supplied by NanoDynamics, USA 167 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.3.5. Flexural strength The variability of the flexural strength of a ceramic can be analysed in a number of ways. The most commonly used is Weibull statistical analysis (appendix C). Table 9.4 shows the fracture strength for different materials. Table 9.4. Flexural strength of different materials. # Material Alumina coating Grain size (µm) Flexural strength (MPa) Alumina - 43.20 241 ± 8.9 Alumina – 2 wt% CNT# - 1.81 253 ± 9.3 Alumina – 2 wt% CNT# ~5-10 nm 3.23 278 ± 6.9 Alumina – 5 wt% CNT# - 0.66 261 ± 11.2 Alumina – 5 wt% CNT# ~5-10 nm 2.34 289 ± 7.3 Alumina – 5 wt% CB - 4.67 233 ± 12.3 CNTs were supplied by NanoDynamics, USA According to the Griffith criterion, the fracture stress is expected to decrease with increasing grain size [279]. The inherent flaw size is increases with the increase in grain size. Minor improvement in flexural strength was observed for alumina – uncoated CNT nanocomposites as compared to alumina. The grain size of alumina – 5 wt% CNT is >43 times smaller than alumina, but only 15% improvement in flexural strength was observed (table 9.4). There is no major effect of the fine grains on the 168 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites fracture strength of alumina – uncoated CNT nanocomposites. The effect of fine grain size is very significant in monolithic alumina. A 21% improvement in fracture strength was reported in monolithic alumina when grain size was reduced from 8 to 2 microns [372]. Fracture strength depends on the flaw geometry and dimensions (equation 9.3), i.e. larger the flaw size, lower the fracture strength [373]. K = c Equation 9.3 Where, K = stress-intensity factor, = fracture strength, = dimensionless geometric constant and c = flaw (including surface flaws) size. There was no significant reinforcing effect of uncoated CNTs on the fracture strength of alumina. Poor grain boundary adhesion and surface flaws (figure 9.3b and 9.3c) are responsible for poor fracture strength of alumina – uncoated carbon nanocomposites. Using alumina coated CNTs as reinforcement for alumina proved slightly effective (table 9.4). As explained in section 9.3.2 and 9.3.3, in alumina – coated CNT (figure 9.6a), fracture occurred in an intergranular and transgranular fashion. A mixed inter/ transgranular fracture mode is a representation of improved fracture strength [147]. All samples failed forming a compression curl on the loading surface, which is a representation of low energy failure (figure 9.10). It was observed that all samples failed by a surface flaw located on the tensile surface. Figure 9.11 shows alumina – CNT nanocomposites mounted in a way that the two surfaces are in direct contact. A typical half-penny shaped region at the centre can be observed in figure 9.11a. Figure 9.11b shows the failure originated from sub-surface agglomerate/ porosity. It is difficult to identify the fracture path because of the higher sub-surface damage resistance of alumina – CNT nanocomposites (figure 9.9b). 169 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (a) (b) Figure 9.10. Low energy failure in 3 – point flexural testing: (a) Optical micrograph (side view) of alumina – 5 wt% uncoated CNT nanocomposites; and (b) schematics as per ASTM C1161-02c [368]. (a) (b) Figure 9.11. Fractured surfaces of alumina – uncoated CNT nanocomposites. Two tensile surfaces are mounted together. Fracture origin can be identified. CNT concentration: (a) 2 wt%; and (b) 5 wt%. 170 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.3.6. Thermal shock resistance (qualitative assessment) Table 9.5 and figure 9.12 show different materials after rapid cooling from 1800 oC. Thermal shock resistance of ceramics improve with decreasing grain size and porosity. All materials were fully densified before rapid cooling. Apart from alumina – CNT nanocomposites, alumina and alumina – CB nanocomposites were found fractured after rapid thermal shocking (table 9.5). Table 9.5. Comparison of thermal shock resistance for different materials. § Alumina – 2 Alumina – 2 wt% CB wt% CNT§ 43.20 5.48 1.81 Fracture status (Cooling rate: 100 oC/ minute) NC NC NC Fracture status (Cooling rate: 200 oC/ minute) PC NC NC Fracture status (Cooling rate: 300 oC/ minute) CR CR NC Materials/ Properties Alumina Grain size (μm) CNTs were supplied by Nanocyl, Belgium NC: Not cracked, PC: Partially cracked and CR: Cracked into two or more fragments. (a) (b) (c) Figure 9.12. Effect of rapid cooling (300 oC/ minute) from 1800 oC after sintering. Disk diameter: 20 mm; (a) alumina; (b) alumina – 2 wt% CB nanocomposite; and (c) alumina – 2 wt% uncoated CNTs (Nanocyl, Belgium) nanocomposite. 171 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites The results can be explained by Hasselman’s theory on the thermal shock resistance of brittle ceramics [374]. According to the theory, the residual strength of material after water quenching is an important index of the thermal shock resistance. The thermal stress fracture resistance parameter R was used to estimate the thermal stress fracture resistance of materials [374]: Tmax = f (1- ) E R K IC 2 Equation 9.4 Equation 9.5 Where, Tmax = maximum quenching temperature before fracture, f = flexural strength, = Poisson ratio, E = elastic modulus, = coefficient of thermal expansion and K IC = fracture toughness. The coefficient of thermal expansion of CNT nanocomposites is quite high when compared to their matrices [196, 375]. It is clear in table 9.6 that alumina – CNT nanocomposites have higher indentation fracture toughness, higher flexural strength, finer grains, lower elastic modulus and higher coefficient of thermal expansion. As per equation 9.4 and 9.5, these mechanical properties are dominating in increasing the thermal shock resistance of alumina – CNT nanocomposites. This is a comparative assessment that infers the alumina – CNT nanocomposites have better thermal shock resistance as compared to alumina and alumina – CB nanocomposites. 172 # - - - - 18.9 3.3 308 241 Cracked Vickers hardness (GPa) Indentation fracture toughness (MPa√m) Flexural modulus (GPa) Fracture strength (MPa) Thermal shock resistance analysis† § 5.48 43.20 Grain size (μm) CNTs were supplied by NanoDynamics, USA - - Alumina coating (nm) - 233 201 3.4 11.0 4.67 - Alumina – 5 wt% CB CNTs were supplied by Nanocyl, Belgium Cracked Alumina – 2 wt% CB Alumina Materials/ Properties - 261 165 5.9 12.0 0.66 - Alumina – 5 wt% CNT Cooling rate: 300 oC/ minute - Not cracked§ † 278 263 5.5 14.9 3.23 ~5-10 Alumina – 2 wt% CNT 253 234 5.6 10.2 1.81 - Alumina – 2 wt% CNT Table 9.6. Summary of mechanical properties of alumina and alumina – carbon nanocomposites#. - 289 185 5.8 18.6 2.34 ~5-10 Alumina – 5 wt% CNT Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.4. Summary of mechanical properties 173 Chapter 9. Mechanical Properties of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 9.5. Conclusions Raw CNTs form strong entangled networks around alumina grains that is useful in making flexible (low flexural modulus) and tough ceramic nanocomposites. CNTs reduced sub-surface damage in alumina – CNT nanocomposites as compared to alumina. Poor hardness was observed for alumina – uncoated CNTs nanocomposites, because of the soft lubricating nature of CNTs. A marginal improvement in flexural strength was observed for alumina – uncoated CNT nanocomposites as compared to alumina. Uncoated CNTs reduced the grain size of alumina nanocomposite (as compared to alumina) which proved beneficial for fracture strength and thermal shock. However, for flexural strength, refining grain size for monolithic alumina is more effective [372] as compared to carbon addition. Wide scattering in fracture strength data (low Weibull moduli, appendix C) was observed for alumina – uncoated CNTs nanocomposites due to poor surface finish as opposed to alumina – coated CNT nanocomposites. Owing to the poor surface finish, it was difficult to analyse the intrinsic contribution of uncoated CNTs on the fracture strength of alumina – CNT nanocomposites. Effective bonding of CNTs with the matrix plays a vital role in the mechanical properties of CNT reinforced ceramics. The effectiveness of using alumina coated CNTs for improved toughness of the nanocomposite was not very significant. Poor grain boundary adhesion reduced the strength of nanocomposite’s grains that resulted failure in transgranular mode. Alumina coated CNTs may be useful for applications where improved IFT as well as good hardness is desired for ceramic nanocomposite. A good interfacial adhesion is required to increase the stress transfer ability in alumina – CNT nanocomposites. This will be the subject of future research. Alumina coated and uncoated CNTs showed higher mechanical reinforcement effectiveness in alumina nanocomposite as compared to carbon black. This is because of their geometry and outstanding mechanical properties (table 1.1). 174 Future Work Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Conclusions 1. In the literature, the role of CNTs on the properties of ceramic matrices remains inconclusive. This is because of the usage of different types and compositions of CNTs, inhomogeneous distribution of CNTs and difference in experimental conditions. 2. The selection of the ultrasonication medium for homogenous dispersion of CNTs is very important for the final properties of the nanocomposite. From dispersion stability observations and agglomerate size measurements, it is clear that DMF produces fine and stable CNT and alumina dispersions. Faster re-agglomeration was observed in ethanol as compared to DMF for pristine and coated CNTs. Coating CNTs with alumina reduced the re-agglomeration rate. No evidence of agglomeration and a good distribution of the CNTs was observed in FE-SEM micrographs of the SPSed samples when they were mixed with alumina in DMF. 3. Nanocomposites prepared using DMF dispersions showed better dispersions, better mechanical properties and higher electrical conductivity as compared to those prepared using ethanol dispersions. Therefore, it is concluded that DMF is an ideal dispersant for making homogeneous and agglomerate-free slurries by any type of colloidal processing. 4. The addition of CNTs or carbon black to alumina significantly increases its sintering rate. The sintering temperature required to achieve full densification of alumina – CNT nanocomposites was reduced by 500 oC as compared to alumina. 5. CNTs form entangled networks at the grain boundaries, which produce significant grain growth retardation. Using this effect, an alumina / nanocomposite laminate structure with a grain size difference of about ten times was successfully co-sintered. 175 Conclusion and Future Work Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 6. CNTs were found to be well preserved in alumina after being SPSed up to 1900 oC/ 100 MPa/ 3 minutes. CNT may not be suitable for ceramics that require high processing temperatures and longer processing durations. Structural preservation of CNTs in ceramic nanocomposites depends on the nature of ceramic, SPS temperature and dwell times. 7. The electrical conductivity of alumina – CNT nanocomposites is four times higher as compared to alumina – carbon black nanocomposites for the same mass content of carbon due to the fibrous nature and high aspect ratio of CNTs. The electrical conductivity of alumina – CNT nanocomposite increased with increasing grain size due to the concentration of CNTs in fewer conductive paths. 8. Fabricating ceramic – CNT nanocomposite by ALD coating of the CNTs and SPS densification provide shielding that improves the oxidative stability of the encapsulated CNTs. In ceramic – CNT nanocomposites, coarser grained materials have higher oxidative resistance as compared to the finer ones, due to the presence of fewer grain boundaries. The oxidative stability of alumina – CNT nanocomposites can be tailored by changing the grain boundary area. 9. CNTs in alumina reduce indentation sub-surface damage as compared to alumina. The effectiveness of using alumina coated CNTs for improved toughness of the nanocomposite was not very significant. However, from a mechanical properties point of view, alumina coated CNT nanocomposite is a good compromise between hard (alumina) and tough (uncoated CNT nanocomposite) materials. 10. Owing to the poor surface finish, it was difficult to analyse the intrinsic contribution of uncoated CNTs on the fracture strength of alumina – CNT nanocomposites. This was not true for coated CNT nanocomposites where marginal and significant improvements in fracture strength and Weibull modulus were observed respectively. 176 Conclusion and Future Work Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 11. Alumina coated and uncoated CNTs showed higher mechanical reinforcement effectiveness in alumina nanocomposite as compared to carbon black. This is because of the geometry and outstanding mechanical properties of CNTs. 12. A qualitative assessment indicated that alumina – CNT nanocomposite has better thermal shock resistance as compared to alumina and alumina – CB nanocomposites. Future Work 1. CNTs form a strong entangled network around alumina grains. Using longer CNTs will result in forming much stronger network that will be beneficial for grain growth retardation and creep resistance at high temperatures. The influence of quality of CNTs (particularly aspect ratio) on the sintering, grain growth and properties should be studied. 2. Breakdown of the conductive networks through deformation or micro-cracking leads to a large change in the electrical conductivity of alumina – CNT nanocomposites. The electrical properties could be used to monitor the damage in sintered nanocomposites. The electrical conductivity could be analysed during mechanical testing to monitor damage evolution. 3. The increased electrical conductivity (section 7.3.1) is attributed to the presence of undamaged CNTs, due to the utilisation of SPS technique that allowed lower sintering temperatures and shorter sintering times. Other conventional techniques (hot-pressing and pressureless sintering) should be compared with SPS to gain deeper understanding. 177 Conclusion and Future Work Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites 4. Alumina – CNT nanocomposites were prepared using 2 – 10 wt% (~ 4.7 – 21 vol%) of CNTs. As compared to alumina and coated CNT nanocomposites, a poor surface finish was observed for uncoated CNT nanocomposites. Better properties are likely to be achieved at lower volume contents of about 0.5 – 1.5 wt%. The influence of using lower concentrations of CNTs should be studied. 5. The potential of using different types of CNTs (i.e. DWNT and SWNT) for improved mechanical properties of ceramic nanocomposites should be investigated. 6. The improvement in mechanical properties of alumina coated CNT nanocomposites was not significant. A better coating like monazite [376] should be used in future studies. Monazite is already well used to improve the interfacial properties of ceramic – ceramic fibre nanocomposites [376]. Coated CNTs are much easier to disperse than the uncoated CNTs and the coating may significantly improve the strength and toughness of the nanocomposite. 7. CNTs have outstanding thermal properties. Thermal conductivity and thermal shock resistance of ceramic – CNT nanocomposites should be investigated. 8. The potential of various prospective applications (appendix B) will be analysed in detail, particularly heating element (section B.2.4) and percolating nano-porosity (section B.2.5). 178 Appendix A. Properties of DMF Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Appendix A. Properties of DMF A.1. Introduction N,N-Dimethylformamide (DMF) is a colourless organic liquid, miscible with water and other organic solvents. It is uniquely versatile and powerful solvent that has a wide liquid range, good chemical and thermal stability, a high polarity, and a wide solubility range for both organic and inorganic compounds [377]. The principal stabilising mechanism for DMF is electrostatic, i.e. the overlap of similarly charged electric double layers [378]. DMF increases the stability of CNTs by dispersal of charge [378, 379]. Martelli [380] reported that DMF affects living organs (kidneys and livers) and digestive tract of pigs when given orally or subcutaneously. Liver disturbances, stomach complaints, headache, loss of appetite and nausea were found in workers subjected to less than 20 ppm vapours [377]. A.2. Chemistry of DMF DMF (figure A) is a derivative of formamide, the amide of formic acid. It is a polar aprotic solvent, which has dipole moments that help to solvate cations by electron donation from an oxygen atom [379]. The dipole moments of DMF are several times as large as that of water. DMF lacks the ability to form hydrogen bonds because all hydrogen atoms are bonded with carbon [379]. Some of the properties of DMF are given in table A. 179 Appendix A. Properties of DMF Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Represents resonance of double bond Figure A. Chemical formula of DMF, C3H7ON. Table A. Properties of DMF [377, 381, 382]. # Properties Values/ description Molecular weight 73.1 Boiling point 153 oC Melting point -60.4 oC Flash point 67 oC Viscosity (at 20 oC) 0.92 cP Viscosity (at 40 oC) 0.74 cP Surface tension (at 20 oC) 0.0368 N/m Surface tension (at 40 oC) 0.0344 N/m Solubility in water Infinite Critical volume 0.25 litres/mol Dipole moment 3.86 D Dielectric constant (at 25 oC) 36.71 Electrical conductivity (at 25 oC) 6 x 10-6 S/m Toxicity threshold limit# 10 ppm or 30 mg/m3 In vapour form 180 Appendix A. Properties of DMF Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Appendix B. Future of Ceramic – CNT Nanocomposites B.1. Introduction The possibilities for the application of CNTs are intriguing and challenging. CNT based nanocomposites may significantly increase in the near future [383]. Apart from making nanocomposites, the main avenues of potential applications of CNTs are: ultimate reinforcement fibres, conducting nanowires, field emitters, nano-tools, energy storage and energy conversion devices, sensors, drug delivery, medical diagnostics and cancer therapy [303, 384]. At the moment, commercial usage of CNTs is quite limited i.e. AFM tips [385] and specialized sports equipments (figure B.1), like bicycle frames [386], baseball bats [387], tennis rackets [388], ice hockey sticks [389]. Figure B.1. Different sports goods based on non-ceramic – CNT nanocomposites [386-389]. 181 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites For the development of commercial products, research into ceramic – CNT nanocomposites is at a very preliminary stage. A commercial product based on ceramic – CNT nanocomposite is not available yet in the market. This chapter discusses the future prospects of ceramic – CNT nanocomposites and the issues relevant to their potential commercial application. B.2. Applications of ceramic – CNT nanocomposite CNTs can be added in ceramics for the following reasons: 1. As a sintering aid 2. To control microstructure 3. To improve electrical and thermal conductivity 4. To improve mechanical properties The first three reasons are the most promising ones. In particular, CNTs are one of the potential reinforcements for ceramic matrices in microelectronics devices, microwave devices, consumer products, medical devices, batteries, solid oxide fuel cells, chemical sensors, gas turbine engines, high-temperature reactors and structural components that are exposed to high-temperature and aggressive environments [390]. Some of the potential applications of ceramic – CNT nanocomposites are discussed below: B.2.1. Conductive ceramic – CNT nanocomposites Ceramic composites with tailorable electrical conductivity have many industrial applications. Dense electrically conductive ceramics are used for static charge dissipation, lightning protection, ceramic heaters, electric discharge machining 182 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites (EDM), electromagnetic interference shielding in electronic, mechanical, structural, chemical, and vacuum applications and anti-static floor tiles [324]. In particular, ceramic alumina with added electrically conductive fillers has been used to fabricate substrates for handling semiconductor wafers that require static protection. In contrast, porous electrically conductive ceramics have applications in highperformance radiative heaters, filters for the aeration of liquids, ceramic foam heaters, and exhaust traps for automotive applications, as well as for the combustion of diesel soot and the non-catalytic oxidation of noxious gases [324]. B.2.2. Electric Discharge Machining (EDM) Ceramic materials with precise and complicated shapes can be manufactured by electric discharge machining. EDM requires materials with a low electrical resistivity, below ~1 Ωm [26]. Hence insulating ceramics, like alumina cannot be electric discharge machined. Adding CNTs to alumina significantly reduces the electrical resistivity (section 7.3.1) and this enhances the electric discharge machinability of alumina. This can be further employed to other insulating matrices, e.g., boron nitride, zirconia, silicon nitride. CNTs have a low oxidation temperature (~ 477 oC, section 8.3.1). If CNTs are not required in the microstructure, ceramic – CNT nanocomposite can then be very easily converted into monolithic ceramic completely by a heat treatment (up to 700oC, section 8.3.3). B.2.3. Special purpose electrodes CNTs can replace graphite in Carbon Ceramic Electrodes (CCEs) for molecule sensing [391] and water treatment in aggressive environments [392]. This is due to their large surface area and superior electrochemical properties [393]. Other 183 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites applications of ceramic – CNT nanocomposites include ceramic microelectrodes for biomedical applications [394], where electrical conductivity is the main concern. B.2.4. Heating elements As a result of outstanding electrical and thermal properties of CNTs, ceramic – CNT nanocomposites can be used as heating elements (figure B.2). To date, Zhan et al. [97] reported the highest electrical conductivity for ceramic – CNT nanocomposite, (3345 S/m), which is quite high as compared to some of the commercially available SiC heating elements [395]. To prevent oxidation of CNTs, these heating elements would need to be covered with a protective ceramic layer. Other areas of interest that might require high electrical and thermal conductivity of ceramic – CNT nanocomposites include electrical contacts, electrical switches, electromagnetic interference shielding, electronic devices, thermal plates and electrical/ thermal pastes. (a) (b) (c) Figure B.2. Heating element based on alumina – CNT nanocomposite. Sample diameter is 20 mm: (a) alumina – 5 wt% CNT (Nanocyl, Belgium); (b) demonstration; and (c) after oxidation, white area shows oxidation of CNTs. 184 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites B.2.5. Porous structures CNTs exist at the grain boundaries (section 5.3.3). Oxidising CNTs leaves a nanoscale, porous network in the ceramic matrix (figure B.3). Because of the low percolation threshold for CNTs, a nanoporous material could be produced with a small content (<1 vol%) of porosity. These materials could be used as nano-filters, bio-scaffolds, thermal insulations and water purification membranes [392]. (a) (b) Figure B.3. CNT network on the grain boundaries of alumina grains. (a) Before oxidation; and (b) After oxidation of CNTs, percolating porosity. B.2.6. Ceramic armour Adding CNTs to ceramic armour (e.g., silicon carbide, alumina, boron carbide, titanium diboride etc.) may improve their ballistic properties. One of the key properties required for improved ballistic resistance of ceramic armour is fracture toughness [396]. Ceramic – CNT nanocomposites have demonstrated an increased sub-surface damage resistance and fracture toughness (section 9.3.3). Apart from ductility, adding CNTs to ceramics can also improve their sinterability and refine their microstructure (section 5.3.1 and 5.3.3), which is beneficial for strength of ceramic nanocomposites as observed in many reports [26, 102, 110, 111, 118-120, 128]. 185 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites B.2.7. Functional Gradient Materials (FGMs) Functional Gradient Materials (FGMs) are required for the application demanding different and unique set of properties in different areas of a designed component. Grain-size FGMs are very advantageous for bio-medical applications [328]. An alumina / nanocomposite laminate structure was fabricated in this work (section 5.3.4) to demonstrate the grain refinement effect of the CNTs and their ability to produce grain size laminated materials. B.2.8. Nanoceramics By reducing the grain size to nanoscale, the nanocrystalline monoliths exhibit strength, hardness and wear resistance, in contrast to those achieved by conventional composite formation [272]. In section 5.3.4 (figure 5.12-5.14), sub-micron alumina after oxidizing CNTs from the alumina – CNT nanocomposite was prepared. It is a lengthy process to make nanoceramics. However, further optimisation of the process may lead to an advance, commercial way of preparing monolithic nanoceramics using CNTs. B.3. Challenges in the development of ceramic – CNT nanocomposite Section B.2 describes some of the applications that may emerge for ceramic – CNT nanocomposites. However, for the commercial success of these applications, there are some immediate obstacles that have to be overcome. 186 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites B.3.1. CNT related issues The commercial success of any technology relies on the availability of materials at a reasonable price. Very little SWNT material is available and most investigators have been limited to working with milligram quantities. This may be problematic in the development of ceramic – SWNT nanocomposites, since bench top processing may not always be representative of larger scale processing [66]. However, the situation for MWNTs is different as industrial quality MWNTs is available in bulk. The 2006 global production capacity of MWNTs was more than 300 tons/year with the potential to grow at a significant rate [92]. International manufacturer Bayer Material Science AG has an annual capacity of 60 tons, and is planning to scale up to 3000 tons by 2010 [61]. Availability in bulk and low cost are the prime factors for the commercial use of MWNTs rather than SWNTs. But unbundled SWNTs have several advantages over MWNTs, i.e. enabling lower percolation threshold for electrical conductivity, reducing the required loading levels in nanocomposites [303]. SWNTs may be more important is microelectronics, where the quantities is much less of an issue. Within the time frame of years, prices have been predicted to decrease between 10 to 100 times, depending on the type of nanotubes [22]. Already, Bayer Material Science AG has started providing industrial grade MWNTs as cheap as £ 0.05/ gram [397]. Future market developments of ceramic – CNT nanocomposites will be fuelled by the declining prices of high quality CNTs. Standards for terminology and testing are required to improve understanding and confidence. They are very important for new materials and designs, as they aid in determining the reliability and effectiveness of the new products. A nanotechnology standards debate has been going since 2004. Some organisations, like ASTM, ANSI, IEEE, BSI and ISO have set some terminologies and standards for nanotechnology [398]. However, there are no standardized protocols for the evaluation of the quality of CNTs and their nanocomposites, which is one of the pre-requisites for the commercial success of ceramic – CNT nanocomposites. 187 Appendix B. Future of Ceramic – CNT Nanocomposites Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Other related issues are the de-bundling of SWNTs, toxicology and the health hazards associated with CNT exposure. B.3.2. Understanding of nanocomposites CNTs exist in different forms, such as CNF, SWNT, DWNT and MWNT. A comparison needs to be done to see the effectiveness of the different types of CNTs for ceramic – CNT nanocomposite. A deeper understanding is required of the fundamental mechanisms associated with the properties of ceramic – CNT nanocomposites. In this work, the interfacial bonding was improved by using alumina coated CNTs. A constitutive modelling approach has to be followed to study the interface, load transfer mechanisms and fracture mechanics of CNT reinforced ceramics. The interfacial adhesion to the matrix [399], dispersion and stress transfer [31] of polymer – CNT nanocomposites can be enhanced by chemical functionalisation of CNTs. However, the role of functionalised CNTs on the properties and sintering of ceramic – CNTs nanocomposite has not yet explored. B.4. Conclusions The future for ceramic – CNT nanocomposites is very bright, especially for applications concerned with the electrical and thermal properties. However, research into ceramic – CNT nanocomposites is at very early stage. CNT reinforced ceramic nanocomposites are not available as a product in the market. Much of the global research interest is in the development for polymer – based nanocomposites, which has resulted in some limited commercial success of polymer – CNT nanocomposites. Apart from a deeper understanding of CNT nanocomposites, CNT related issues must be resolved for the substitution of CNT based technology in the real world 188 Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Appendix C. Weibull Statistical Analysis Weibull analysis is the empirical statistical distribution function describes the scatter in strength values of ceramic materials. It is used to assign mechanical properties to brittle materials in probabilistic terms, and to define design requirements in terms of strength and reliability [400]. To increase the confidence level, the data sample has to be sufficiently large (generally ≥10), as a small uncertainty in Weibull modulus can result in large uncertainties in the survival probabilities. Metals typically have values of m ~100, while traditional ceramics have values ~5 and engineering ceramics are in the range 10-25 [400]. The survival probability, i.e., the fraction of samples that would survive a given stress level can be given by equation C1 [279]. m S p = exp o Equation C1 Where S p = survival probability, = maximum design stress, o = normalising parameter (when lnln(1/S) = 0) and m = Weibull modulus. The Weibull modulus is relatively low for alumina, alumina – uncoated CNT and alumina – CB nanocomposites (table C). For alumina – carbon nanocomposites, this is attributed to the poor surface finish as reported in section 9.3.1. Higher Weibull 189 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites moduli were observed for coated CNT nanocomposites due to good surface finish (comparatively) that resulted in less variation in the fracture strength data (table C). Table C. Weibull modulus (fracture strength) of different materials. Material Alumina coating Weibull modulus Alumina - 9.05 Alumina – 2 wt% CNT - 9.26 Alumina – 2 wt% CNT ~5-10 nm 13.77 Alumina – 5 wt% CNT - 7.38 Alumina – 5 wt% CNT ~5-10 nm 13.35 Alumina – 5 wt% CB - 6.52 CNTs and CB were supplied by NanoDynamics, USA and Cabot, USA respectively. Tables C1 – C6 and figures C1 – C6 show Weibull statistical analysis for different materials used in this study. The reason that -lnln(1/S) is plotted rather than lnln(1/S) is aesthetic, such that the high survival probabilities appear on the upper left-hand sides of the plots [279]. Note for Weibull modulus analysis, S j was calculated using equation C2. j - 0.3 S j = 1- N +0.4 Equation C2 Where S j = survival probability of jth sample, j = sample rank and N = total no. of samples tested, which was 10 for this study. 190 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table C1. Weibull statistical analysis (fracture strength) for alumina. Flexural Strength, (MPa) ln ( ) Rank, j S -lnln (1/S) 204.97 5.32 1 0.93 2.66 205.37 5.32 2 0.84 1.72 227.80 5.43 3 0.74 1.20 228.76 5.43 4 0.64 0.82 229.59 5.44 5 0.55 0.51 242.87 5.49 6 0.45 0.23 253.66 5.54 7 0.36 -0.03 254.91 5.54 8 0.26 -0.30 259.50 5.56 9 0.16 -0.59 300.95 5.71 10 0.07 -0.99 -lnln(1/S) m = 9.05 ln (σ) Figure C1. Weibull plot for data shown in table C1. Weibull modulus: 9.05. 191 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table C2. Weibull statistical analysis (fracture strength) for alumina – 5 wt% CB nanocomposite. CNTs were supplied by NanoDynamics, USA. Flexural Strength, (MPa) ln ( ) Rank S -lnln (1/S) 178.30 5.18 1 0.93 2.66 195.24 5.27 2 0.84 1.72 201.08 5.30 3 0.74 1.20 212.29 5.36 4 0.64 0.82 217.99 5.38 5 0.55 0.51 247.98 5.51 6 0.45 0.23 250.89 5.53 7 0.36 -0.03 254.82 5.54 8 0.26 -0.30 264.32 5.58 9 0.16 -0.59 307.43 5.73 10 0.07 -0.99 -lnln(1/S) m = 6.52 ln (σ) Figure C2. Weibull plot for data shown in table C2. Weibull modulus: 6.52. 192 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table C3. Weibull statistical analysis (fracture strength) for alumina – 2 wt% uncoated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Flexural Strength, (MPa) ln ( ) Rank S -lnln (1/S) 199.68 5.30 1 0.93 2.66 229.12 5.43 2 0.84 1.72 230.64 5.44 3 0.74 1.20 244.74 5.50 4 0.64 0.82 247.80 5.51 5 0.55 0.51 255.90 5.54 6 0.45 0.23 257.63 5.55 7 0.36 -0.03 282.87 5.64 8 0.26 -0.30 287.27 5.66 9 0.16 -0.59 293.05 5.68 10 0.07 -0.99 -lnln(1/S) m = 9.26 ln (σ) Figure C3. Weibull plot for data shown in table C3. Weibull modulus: 9.26. 193 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table C4. Weibull statistical analysis (fracture strength) for alumina – 2 wt% coated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Flexural Strength, (MPa) ln ( ) Rank S -lnln (1/S) 244.47 5.50 1 0.93 2.66 257.36 5.55 2 0.84 1.72 262.30 5.57 3 0.74 1.20 269.21 5.60 4 0.64 0.82 270.61 5.60 5 0.55 0.51 273.56 5.61 6 0.45 0.23 284.13 5.65 7 0.36 -0.03 295.84 5.69 8 0.26 -0.30 303.32 5.71 9 0.16 -0.59 315.21 5.75 10 0.07 -0.99 -lnln(1/S) m = 13.77 ln (σ) Figure C4. Weibull plot for data shown in table C4. Weibull modulus: 13.77. 194 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table C5. Weibull statistical analysis (fracture strength) for alumina – 5 wt% uncoated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Flexural Strength, (MPa) ln ( ) Rank S -lnln (1/S) 184.20 5.22 1 0.93 2.66 239.14 5.48 2 0.84 1.72 242.34 5.49 3 0.74 1.20 249.53 5.52 4 0.64 0.82 257.28 5.55 5 0.55 0.51 268.82 5.59 6 0.45 0.23 277.22 5.62 7 0.36 -0.03 281.40 5.64 8 0.26 -0.30 297.28 5.69 9 0.16 -0.59 310.85 5.74 10 0.07 -0.99 -lnln(1/S) m = 7.38 ln (σ) Figure C5. Weibull plot for data shown in table C5. Weibull modulus: 7.38. 195 Appendix C. Weibull Statistical Analysis Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Table C6. Weibull statistical analysis (fracture strength) for alumina – 5 wt% coated CNT nanocomposite. CNTs were supplied by NanoDynamics, USA. Flexural Strength, (MPa) ln ( ) Rank S -lnln (1/S) 248.95 5.52 1 0.93 2.66 261.19 5.57 2 0.84 1.72 267.76 5.59 3 0.74 1.20 290.76 5.67 4 0.64 0.82 291.28 5.67 5 0.55 0.51 292.07 5.68 6 0.45 0.23 303.40 5.72 7 0.36 -0.03 304.21 5.72 8 0.26 -0.30 308.41 5.73 9 0.16 -0.59 323.59 5.78 10 0.07 -0.99 -lnln(1/S) m = 13.35 ln (σ) Figure C6. Weibull plot for data shown in table C6. Weibull modulus: 13.35. 196 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites Appendix D. Recent Publications (on CNT nanocomposites) 1. F. Inam, H.X. Yan, T. Peijs, M.J. Reece, "The sintering and grain growth behaviour of ceramic – carbon nanotube nanocomposites", Journal of the American Ceramic Society, Ready for submission. 2. F. Inam, H.X. Yan, M.J. Reece, T. Peijs, “Stability of multiwall carbon nanotubes in sintered ceramic nanocomposite”, Advances in Applied Ceramics, Accepted (2009). 3. J. Dusza, G. Blugan, J. Morgiel, J. Kuebler, F. Inam et al., “Hot pressed and Spark Plasma Sintered zirconia / carbon nanofibre composites”, Journal of the European Ceramic Society, Article in Press (2009). DOI: 10.1016/j.jeurceramsoc.2009.05.030 4. F. Inam, H.X. Yan, T. Peijs, M.J. Reece, "Electrically conductive alumina – carbon nanocomposites prepared by Spark Plasma Sintering", Journal of the European Ceramic Society, Article in Press (2009). DOI: 10.1016/j.jeurceramsoc.2009.05.045 5. F. Inam, H.X. Yan, M.J. Reece, T. Peijs, "Dimethylformamide: an effective dispersant for making ceramic – carbon nanotube composites", Nanotechnology, Vol. 19, No. 19 (2008) 195710 (5 pages) 6. F. Inam, H.X. Yan, R. Zhang, D. Hua, M. Reece, T. Peijs, "Firing up on all cylinders: Carbon nanotube based nanocomposites", Materials World, Vol. 15, No. 10 (2007) pp. 24-25 7. F. Inam, T. Peijs, "Re-aggregation of Carbon Nanotubes in two-component epoxy system", Journal of Nanostructured Polymers and Nanocomposites, Vol. 2, No. 3 (2006) pp. 87-95 8. F. Inam, T. Peijs, "Transmission light microscopy of Carbon Nanotubes – epoxy nanocomposites involving different dispersion methods", Advanced Composite Letters, Vol. 15, No. 1 (2006) pp. 7-15 197 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites References [1] Li, X. and Nardi, P., Nanotechnology, Vol. 15 (2004) pp. 211-217 [2] Stubbs, D.G., Tillinghast, E.K. and Townley, M.A., Vol. 79 (1992) pp. 231-234 [3] Gosline, J.M., Guerette, P.A., Ortlepp, C.S. and Savage, K.N., Journal of Experimental Biology, Vol. 202 (1999) pp. 3295-3303 [4] Jakubinek, M.B., Samarasekera, C.J. and White, M.A., Journal of Materials Research, Vol. 21 (2006) pp. 287-292 [5] Porter, D., Materials Science and Engineering A, Vol. 365 (2004) pp. 38-45 [6] Roy, R., Roy, R.A., Roy, D.M., Materials Letters, Vol. 4 (1986) pp. 323-328 [7] Sternitzke, M., Journal of the European Ceramics Society, Vol. 17 (1997) pp. 1061-1082 [8] Iijima, S., Nature, Vol. 354 (1991) pp. 56-58 [9] Okada, A., Journal of the European Ceramic Society, Vol. 28 (2008) pp. 1097-1104 [10] Rahaman, M.N., Yao, A., Bal, B.S., Garino, J.P. and Ries, M.D., Journal of the American Ceramic Society, Vol. 90 (2007) pp. 1965–1988 [11] Hull, D. and Clyne, T.W., An Introduction to Composite Materials, Second Edition, Cambridge Solid State Science Series, UK (1996) [12] Kutz, M., Handbook of Materials Selection, John Wiley & Sons, Inc., USA (2002) [13] Miracle, D.B. and Donaldson, S.L., ASM Handbook: Composites v. 21, ASM, USA (2001) [14] Somiya, S. and Inomata, Y., Silicon Carbide Ceramics-2, Elsevier Applied Science, USA (1988) [15] Kostikov, V.I., Fibre Science and Technology, Springer, USA (1995) [16] Hosseini, S.H. and Pairovi, A., Iranian Polymer Journal, Vol. 14 (2005) pp. 934-941 [17] Cho, J.W. and Jung, H., Journal of Materials Science, Vol. 32 (1997) pp. 5371-5376 [18] Harris, P.J.F., International Materials Reviews, Vol. 49 (2004) pp. 31-43 [19] Saccani, A., Sandrolini, F., Barbieri, L., Corradi, A. and Lancellotti, I., Journal of Materials Science, Vol. 36 (2001) pp. 2173-2177 [20] EURIMA Fact Sheet, No. 88, Cellulose Fibre Insulation, Belgium (2004) [21] Cellulose, Materials Information, www.goodfellow.com [22] Carbon Nanotubes, www.cientifica.com [23] Belin, T. and Epron, F., Materials Science and Engineering B, Vol. 119 (2005) pp. 105118 [24] UC Researchers Shatter World Records with Length of Carbon Nanotube Arrays, www.uc.edu/news 198 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [25] Peng, L.M., Zhang, Z.L., Xue, Z.Q., Wu, Q.D., Gu, Z.N. and Pettifor, D.G., Physics Review Letters, Vol. 85 (2000) pp. 3249-3252 [26] Hirota, K., Takaura, Y., Kato, M. and Miyamoto, Y., Journal of Materials Science, Vol. 42 (2007) pp. 4792-4800 [27] Thostenson, E.T., Ren, Z. and Chou, T.W., Composites Science and Technology, Vol. 61 (2001) pp. 1899-1912 [28] Choi, Y.K., Gotoh, Y., Sugimoto, K., Song, S.M, Yanagisawa, T. and Endo, M., Polymer, Vol. 46 (2005) pp. 11489-11498 [29] Qin, L.C., Physical Chemistry Chemical Physics, Vol. 9 (2007) pp. 31-48 [30] Ramanathan, T., Liu, H. and Brinson, L.C., Journal of Polymer Science: Part B: Polymer Physics, Vol. 43 (2005) pp. 2269-2279 [31] Coleman, J.N., Khan, U. and Gun’ko, Y.K., Advanced Materials, Vol. 18 (2006) pp. 689706 [32] Dyke, C.A. and Tour, J.M., Journal of Physical Chemistry: A, Vol. 108 (2004) pp. 1115111159 [33] Song, Y.S. and Youn, J.R., Korea-Australia Rheology Journal, Vol. 16 (2004) pp. 201-212 [34] Li, S., Qin, Y., Shi, J., Guo, Z.X., Li, Y. and Zhu, D., Chemistry of Materials, Vol. 17 (2005) pp. 130-135 [35] Wichmann, M.H.G., Sumfleth, J., Gojny, F.H., Quaresimin, M., Fiedler, B. and Schulte, K., Engineering Fracture Mechanics, Vol. 73 (2006) pp. 2346-2359 [36] Gojny, F.H., Nastalczyk, J., Roslaniec, Z. and Schulte, K., Chemical Physics Letters, Vol. 370 (2003) pp. 820-824 [37] Zhu, J., Khabashesku, V., Imam, A., Crane, R., Lozano, K. and Barrera, E., Materials Science Forum, Vols. 475-479 (2005) pp. 1059-1062 [38] Gojny, F.H., Wichmann, M.H.G., Fiedler, B. and Schulte, K., Composites Science and Technology, Vol. 65 (2005) pp. 2300-2313 [39] Namilae, S., Chandra, N. and Shet, C., Chemical Physics Lettersa, Vol. 387 (2004) pp. 247-252 [40] Askeland, D.S., The Science and Engineering of Materials, Third Edition, PWS Publishing, USA (1994) [41] Chandra, N. and Ghonem, H., Composites: Part A, Vol. 32 (2001) pp. 575-584 [42] Buryachenko, V.A., Roy, A., Lafdi, K., Anderson, K.L. and Chellapilla, S., Composites Science and Technology, Vol. 65 (2005) pp. 2435-2465 [43] Wang, J., Fang, Z., Gu, A., Xu, L. and Liu, F., Journal of Applied Polymer Science, Vol. 100 (2006) pp. 97-104 [44] Bellucci, S., Physica Status Solidi C, Vol. 2 (2005) pp. 34-47 [45] Monthioux, M. and Kuznetsov, V.L., Carbon, Vol. 44 (2006) pp. 1621-1623 199 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [46] Hughes T.V. and Chambers C.R., US Patent 405480, 1889 [47] Radushkevich, L.V. and Lukyanovich, V.M., Zurn Fisic Chim, Vol. 26 (1952) pp. 88-95 [48] Walker, P.L., Rakszawski, J.F. and Imperial, G.R., Journal of Physical Chemistry, Vol. 63 (1959) pp. 133-140 [49] Oberlin, A., Endo, M. and Koyoma, T., Journal of Crystal Growth, Vol. 32 (1976) pp. 335349 [50] Kroto, H.W., Heath, J.R., O'Brien, S.C., Curl, R.F. and Smalley, R.E., Nature, Vol. 318 (1985) pp. 162-163 [51] Kratschmer, W., Lamb, L.D., Fostiropoulos, K. and Huffman, D.R., Nature, Vol. 347 (1990) pp. 354-358 [52] Hamada, N., Sawada, S. and Oshiyama, A., Physics Review Letters, Vol. 68 (1992) pp. 1579-1581 [53] Yacaman, M.J., Yoshida, M.M., Rendon, L., Santiesteban, J.G., Applied Physics Letters, Vol. 62 (1993) pp. 202-204 [54] Ajayan, P.M. and Iijima, S., Nature, Vol. 358 (1992) pp. 23 [55] Iijima, S. and Ichihashi, T., Nature, Vol. 363 (1993) p. 603 [56] Bethune, D.S., Kiang, C.H., Vries, M.S.D., Gorman, G., Savoy, R., Vazquez, J. and Beyers, R., Nature, Vol. 363 (1993) p. 605 [57] Bandow S. and Saito Y., Japanese Journal of Applied Physics: Part 2, Vol. 32 (1995) pp. 1677-1680 [58] Ando, Y., Zhao, X., Sugai, T., Kumar, M., Materials today, Vol. 7 (2004) pp. 22-29 [59] Vivekchand, S.R.C., Jayakanth, R., Govindaraj, A. and Rao, C.N.R., Small, Vol. 1 (2005) pp. 920-923 [60] Bellucci, S., Physica Status Solidi C, Vol. 2 (2005) pp. 34-47 [61] Bierdel, M., Buchholz, S., Michele, V., Mleczko, L., Rudolf, R., Voetz, M. and Wolf, A., Physica Status Solidi B, Vol. 244 (2007) pp. 3939-3943 [62] Popov, V.N., Materials Science and Engineering R, Vol. 43 (2004) pp. 61-102 [63] Li, H. C. Shi, X. Du, C. He, J. Li, N. Zhao, Materials Letters, Vol. 62 (2008) pp. 14721475 [64] Guo, T., Nikolaev, P., Thess, A., Colbert, D.T. and Smalley, R.E., Chemical Physics Letters, Vol. 243 (1995) pp. 49-54 [65] Nalwa, H.S., Encyclopedia of Nanoscience and Nanotechnology, American Scientific Publishers, USA, 2003 [66] Breuer, O., and Sundararaj, U., Polymer Composites, Vol. 25 (2004) pp. 630-645 [67] Smalley, R.E., Colbert, D.T., Rinzler, A.G., Nikolaev, P., Guo, T. and Thess, A., US Patent 6183714, 1996 200 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [68] Nikolaev, P., Bronikowski, M.J., Bradley, R.K., Rohmund, F., Colbert, D.T., Smith, K.A. and Smalley, R.E., Chemical Physics Letters, Vol. 313 (1999) pp. 91-97 [69] Resasco, D.E., Kitiyanan, B., Alvarez, W.E. and Balzano, L., US Patent 6919064, 2005 [70] Hsu, W.K., Hare, J.P., Terrones, M., Kroto, H.W., Walton, D.R.M. and Harris, P.J.F., Nature, Vol. 377 (1995) pp. 687 [71] Laplaze, D., Bernier, P., Maser, W. F., Flamant, G., Guillard, T. and Loiseau, A., Carbon, Vol. 36 (1998) pp. 685-688 [72] Ando, Y., Zhao, X., Hirahara, K., Suenaga, K., Bandow, S. and Iijima, S., Chemical Physics Letters, Vol. 323 (2000) pp. 580-585 [73] Rohmund, F., Falk, L.K.L. and Campbell, F.E.B., Chemical Physics Letters, Vol. 328 (2000) pp. 369-373 [74] Kyotani, T., Tsai, L.F. and Tomita, A., Chemistry of Materials, Vol. 8 (1959) pp. 21092113 [75] Vanderwal, R.L., Ticich, T. and Curtis, V.E., Chemical Physics Letters, Vol. 323 (2000) pp. 217-223 [76] Wei, F., Zhang, Q., Qian, W.Z., Yu, H., Wang, Y., Luo, G.H., Xu, G.H. and Wang, D.Z., Powder Technology, Vol. 103 (2008) pp. 10-20 [77] Li, K.Z., Wei, J., Li, H.J., Zhang, Y.L., Wang, C. and Hou, D.S., Applied Surface Science, Vol. 253 (2007) pp. 7365-7368 [78] Muller, J., Huaux, F., Lison, D., Carbon, Vol. 44 (2006) pp. 1048-1056 [79] Saito, N., Usui, Y., Aoki, K., Narita, N., Shimizu, M., Ogiwara, N., Nakamura, K., Ishigaki, N., Kato, H., Taruta, S. and Endo, M., Current Medical Chemistry, Vol. 15 (2008) pp. 523-527 [80] Sanchez, J.C.C., Elias, A.L., Mancilla, R., Arrellin, H.G., Terrones, H., Laclette, J.P. and Terrones, M., Nano Letters, Vol. 8 (2006) pp. 1609-1616 [81] Schipper, M.L., Ratchford, N.N., Davis, C.R., Kam, N.W.S., Chu, P., Liu, Z., Sun, X.M., Dai, H.J. and Gambhir, S.S., Nature Nanotechnology, Vol. 3 (2008) pp. 216-221 [82] Kam, N.W.S., O'Connell, M., Wisdom, J.A. and Dai, H.J., Proceedings of the National Academy of Sciences of the USA, Vol. 102 (2005) pp. 11600-11605 [83] Oberdorster, G., Oberdorster, E. and Oberdorster, J., Environmental health perspectives, Vol. 113 (2005) pp. 823-839 [84] Maynard, A.D. et al., Nature, Vol. 444 (2006) pp. 267-269 [85] Warheit, D.B., Materials Today, Vol. 7 (2004) pp. 32-35 [86] Porter, A.E., Gass, M., Muller, K., Skepper, J.N., Midgley, P.A. and Welland, M., Nature Nanotechnology, Vol. 2 (2007) pp. 713-717 [87] Donaldson, K., Aitken, R., Tran, L., Stone, V., Duffin, R., Forrest, G. and Alexander, A., Toxicological Science, Vol. 92 (2006) pp. 5-22 201 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [88] Cherukuri, P., Bachilo, S.M., Litovsky, S.H. and Weisman, R.B., Journal of the American Chemical Society, Vol. 126 (2004) pp. 15638-15639 [89] Pulskamp, K., Knirsch, J.M.W. and Krug, H.F., Toxicology Letters, Vol. 172 (2006) pp. S35-S36 [90] Pulskamp, K., Diabte, S. and Krug, H.F., Toxicology Letters, Vol. 168 (2007) pp. 58-74 [91] Zheng, L., Liao, X. and Zhu, Y.T., Materials Letters, Vol. 60 (2006) pp. 1968-1972 [92] Ciselli, P., The Potential of Carbon Nanotubes in Polymer Composites, PhD Thesis, Eindhoven University of Technology, Holland (2007) [93] Global Market and Applications for Carbon Nanotubes and Fullerenes, Market Assessment Report, www.nanoposts.com (2009) [94] Lu, K., Powder Technology, Vol. 177 (2007) pp. 154-161 [95] Hilding, J., Grulke, E.A., Zhang, E.A. and Lockwood, F., Journal of Dispersion Science and Technology, Vol. 24 (2003) pp. 1-41 [96] Ahmad, K., Pan, W. and Shi, S.L., Applied Physics Letters, Vol. 89 (2006) p. 133122 [97] Zhan, G.D., Kuntz, J.D., Garay, J.E. and Mukherjee, A.K., Applied Physics Letters, Vol. 83 (2003) pp. 1228-1230 [98] Zhan, G.D., Kuntz, J.D., Mukherjee, A.K., Zhu, P. and Koumoto, K., Scripta Materialia, Vol. 54 (2006) pp. 77-82 [99] Hwang, G.L. and Hwang, K.C., Journal of Materials Chemistry, Vol. 11 (2001) pp. 17221725 [100] Chen, Y., Gan, C., Zhang, T., Yu, G., Bai, P. and Kaplan, A., Applied Physics Letters, Vol. 86 (2005) pp. 251905 [101] Sun, J., Gao, L. and Li, W., Chemistry of Materials, Vol. 14 (2002) pp. 5169-5172 [102] Lim, D.S., You, D.H., Choi, H.J., Lim, S.H. and Jang, H., Wear, Vol. 259 (2005) pp. 539544 [103] Morisada, Y., Miyamoto, Y., Takaura, Y., Hirota, K. and Tamari, N., International Journal of Refractory Metals and Hard Materials, Vol. 25 (2007) pp. 322-327 [104] Wang, J., Kou, H., Liu, X., Pan, Y. and Guo, J., Ceramics International, Vol. 33 (2007) pp. 719-722 [105] Wang, Y. and G.A., Zerda, Journal of Physics: Condensed Matter, Vol. 19 (2007) pp. 356205 [106] Kealley, C., Elcombe, M. and Riessen, A.V., Journal of Synchrotron Radiation, Vol. 15 (2008) pp. 86-90 [107] Balazsi, C., Weber, F., Kover, Z., Shen, Z., Konya, Z., Kasztovszky, Z., Vertesy, Z., Biro, L.P., Kiricsi, I. and Arato, P., Current Applied Physics, Vol. 6 (2006) pp. 124-130 [108] Daraktchiev, M., Moortele, B.V.D., Schaller, R., Couteau, E. and Forro, U., Advanced Materials, Vol. 17 (2005) pp. 88-91 202 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [109] Fan, J., Zhao, D., Wu, M., Xu, Z. and Song, J., Journal of the American Ceramic Society, Vol. 89 (2006) pp. 750-753 [110] Wei, T., Fan, Z., Luo, G. and Wei, F., Materials Letters, Vol. 62 (2007) pp. 641-644 [111] Zhu, Y.F., Shi, L., Zhang, C., Yang, X.Z. and Liang, J., Applied Physics A, Vol. 89 (2007) pp. 761-767 [112] Ye, F., Liu, L., Wang, Y., Zhou, Y., Peng, B. and Meng, Q., Scripta Materialia, Vol. 55 (2006) pp. 911-914 [113] Guo, S., Sivakumar, R., Kitazawa, H. and Kagawa, Y., Journal of the American Ceramic Society, Vol. 90 (2007) pp. 1667-1670 [114] Huang, Q. and Gao, L., Journal of Materials Chemistry, Vol. 14 (2004) pp. 2536-2541 [115] Balazsi, C., Konya, Z., Weber, F., Biro, L.P. and Arato, P., Materials Science and Engineering C, Vol. 23 (2003) pp. 1133-1137 [116] Boccaccini, A.R., Thomas, B.J.C., Brusatin, G. and Colombo, P. , Journal of Materials Science, Vol. 42 (2007) pp. 2030-2036 [117] Meng, Y.H., Tang, C.Y., Tsui, C.P. and Chen, D.Z., Journal of Materials Science: Materials in Medicine, Vol. 19 (2008) pp. 75-81 [118] Fan, J.P., Zhuang, D.M., Zhao, D.Q., Zhang, G., Wu, M.S., Wei, F. and Fan, Z.J., Applied Physics Letters, Vol. 89 (2006) p. 121910 [119] Ning, J.W., Zhang, J.J., Pan, Y.B. and Guo, J.K., Materials Science & Engineering A, Vol. 357 (2003) pp. 392-396 [120] Ahmad, K. and Pan, W., Composite Science and Technology, Vol. 68 (2008) pp. 13211327 [121] Fonseca, F.C., Muccillo, R., de Florio, D.Z., Ladeira, L.O. and Ferlauto, A.S., Applied Physics Letters, Vol. 91 (2007) pp. 243107 [122] Shi, S.L. and Liang, J., Journal of the American Ceramic Society, Vol. 89 (2006) pp. 35333535 [123] Jiang, L.Q. and Gao, L., Journal of Materials Chemistry, Vol. 15 (2005) pp. 260-266 [124] Ukai, T., Sekino, T., Hirvonen, A., Tanaka, N., Kusunose, T., Nakayama, T. and Niihara, K., Key Engineering Materials, Vol. 317-318 (2006) pp. 661-664 [125] Ionascu, C. and Schaller, R., Materials Science and Engineering A, Vol. 442 (2006) pp. 175-178 [126] Shi, S.L. and Liang, J., Journal of Applied Physics, Vol. 101 (2007) pp. 023708 [127] Wang, X.T., Padture, N.P. and Tanaka, H., Nature Materials, Vol. 3 (2004) pp. 539-544 [128] Sun, J., Gao, L., and Jin, X., Ceramics International, Vol. 31 (2005) pp. 898-896 [129] Tatami, J., Katashima, T., Komeya, K., Meguro, T. and Wakihara, T., Journal of the American Ceramic Society, Vol. 88 (2005) pp. 2889-2893 203 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [130] Duszova, A., Dusza, J., Tomasek, K., Blugan, G. and Kuebler, J., Journal of the European Ceramic Society, Vol. 28 (2008) pp. 1023-1027 [131] Duszova, A., Dusza, J., Tomasek, K., Morgiel, J., Blugand, G. and Kuebler, J., Scripta Materialia, Vol. 58 (2008) pp. 520-523 [132] An, J.W., You, D.H. and Lim, D.S., Wear, Vol. 255 (2003) pp. 677-681 [133] Balazsi, C., Fenyi, B., Hegman, N., Kover, Z., Weber, F., Vertesy, Z., Konya, Z., Kiricsi, I., Biro, L.P. and Arato, P., Composite: Part B, Vol. 37 (2006) pp. 418-424 [134] Ma, R.Z., Wu, J., Wei, B.Q., Liang, J. and Wu, D.H., Journal of Materials Science, Vol. 33 (1998) pp. 5243-5246 [135] Singh, M.K., Shokuhfar, T., Gracio, J.J.D.A., Sousa, A.C.M.D., Fereira, J.M.D.F., Garmestani, H. and Ahzi, S., Advanced Functional Materials, Vol. 18 (2008) pp. 694-700 [136] Thomson, K.E., Jiang, D., Lemberg, J.A., Koester, K.J., Ritchie, R.O. and Mukherjee, A.K., Materials Science and Engineering A, Vol. 493 (2008) pp. 256-260 [137] Maensiri, S., Laokul, P., Klinkaewnarong, J. and Amornkitbamrung, V., Materials Science and Engineering A, Vol. 447 (2007) pp. 44-50 [138] Zhan, G.D. and Mukherjee, A.K., International Journal of Applied Ceramic Technology, Vol. 1 (2004) pp. 161-171 [139] Sun, J., Gao, L., Iwasa, M., Nakayama, T. and Niihara, K., Ceramics International, Vol. 31 (2005) pp. 1131-1134 [140] Zhan, G.D., Kuntz, J.D., Wan, J. and Mukherjee, A.K., Nature Materials, Vol. 2 (2003) pp. 38-42 [141] Guo, S., Sivakumar, R. and Kagawa, Y., Advanced Engineering Materials, Vol. 9 (2007) pp. 84-87 [142] Cha, S.I., Kim, K.T., Lee, K.H., Mo, C.B. and Hong, S.H., Scripta Materialia, Vol. 53 (2005) pp. 793-797 [143] Siegel, R.W., Chang, S.K., Ash, B.J., Stone, J., Ajayan, P.M., Doremus, R.W. and Schadler, L.S., Scripta Materialia, Vol. 44 (2001) pp. 2061-2064 [144] Wang, Y., Voronin, G.A., Zerda, T.W. and Winiarski, A., Journal of Physics: Condensed Matter, Vol. 18 (2006) pp. 275-282 [145] Mo, C.B., Cha, S.I., Kim, K.T., Lee, K.H. and Hong, S.H., Materials Science and Engineering A, Vol. 395 (2005) pp. 124-128 [146] Rul, S., Lefèvre-schlick, F., Capria, E., Laurent, C. and Peigney, A., Acta Materialia, Vol. 52 (2004) pp. 1061-1067 [147] Flahaut, E., Peigney, A., Laurent, C., Marliere, C., Chastel, F. and Rousset, A., Acta Materialia, Vol. 48 (2000) pp. 3803-3812 [148] Peigney, A., Laurent, C., Flahaut, E. and Rousset, A., Ceramics International, Vol. 26 (2000) pp. 677-683 204 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [149] Chang, S., Doremus, R.H., Ajayan, P.M. and Siegel, R.W., 24th Annual Conference on Composites, Advanced Ceramics, Materials, and Structures: A: Ceramic Engineering and Science Proceedings, Vol. 21 (2000) pp. 653 [150] Zhang, X.H., Hong C.Q., Han, J.C., Zhang, H.X., Meng, S.H. and Qu, W., Journal of Inorganic Materials, Vol. 21 (2006) pp. 899-905 [151] Huang, Q., Gao, L. and Sun, J., Journal of the American Ceramic Society, Vol. 88 (2005) pp. 3515-3518 [152] Lei, S., Zhu, Y.F., Chan, Z. and Ji, L., Journal of Reinforced Plastics and Composites, Vol. 27 (2008) pp. 245-253 [153] Zhang, F., Sun, J. and Shen, J., Materials Science Forum, Vol. 475-479 (2005) pp. 989-992 [154] Estili, M., Takagi, K. and Kawasaki, A., Scripta Materialia, Vol. 59 (2008) pp. 703-705 [155] Yamamoto, G., Omori, M., Hashida, T. and Kimura, H., Nanotechnology, Vol. 19 (2008) pp. 315708 [156] Estili, M., Kawasaki, A., Sakamoto, H., Mekuchi, Y., Kuno, M. and Tsukada, T., Acta Materialia, Vol. 56 (2008) pp. 4070-4079 [157] Hirota, K., Hara, H. and Kato, M., Materials Science and Engineering A, Vol. 458 (2007) pp. 216-225 [158] Huang, Q. and Gao, L., Applied Physics Letters, Vol. 86, (2005) pp. 123104 [159] Yoo, S.H., Wang, H.J., Oh, S.T., Kang, S.G. and Choa, Y.H., Key Engineering Materials, Vol. 317-318 (2006) pp. 665 - 668 [160] Viswanathan, V., Laha, T., Balani, K., Agarwal, A. and Seal, S., Materials Science and Engineering R, Vol. 54 (2006) pp. 121-285 [161] Peigney, A., Flahaut, E., Laurent, C., Chastel, F. and Rousset, A., Chemical Physics Letters, Vol. 352 (2002) pp. 20-25 [162] Balani, K., Bakshi, S.R., Chen, Y., Laha, T. and Agarwal, A., Journal of Nanoscience and Nanotechnology, Vol. 7 (2007) pp. 3553-3562 [163] Balani, K., Anderson, R., Laha, T., Andara, M., Tercero, J., Crumpler, E. and Agarwal, A., Biomaterials, Vol. 28 (2007) pp. 618-624 [164] Laha, T., Agarwal, A., McKechnie, T. and Seal, S., Materials Science and Engineering A, Vol. 381 (2004) pp. 249-258 [165] Balani, K., Zhang, T., Karakoti, A., Li, W.Z., Seal, S. and Agarwal, A., Acta Materialia, Vol. 56 (2008) pp. 571-579 [166] Wang, Y., Iqbal, Z. and Mitra, S., Carbon, Vol. 44 (2006) pp. 2804-2808 [167] Thostenson, E.T., Karandikar, P.G. and Chou, T.W., Journal of Physics D: Applied Physics, Vol. 38 (2005) pp. 3962-3965 [168] Katsuda, Y., Gerstel, P., Narayanan, J., Bill, J. and Aldinger, F., Journal of the European Ceramic Society, Vol. 26 (2006) pp. 3399-3305 205 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [169] Jiang, L. and Gao, L., Chemistry of Materials, Vol. 15 (2003) pp. 2843-2853 [170] Kamalakaran, R., Lupo, F., Grobert, N., Scheu, T., Phillipp, N.Y.J. and Ruhle, M., Carbon, Vol. 42 (2004) pp. 1-4 [171] Huang, J.Y., Chen, S., Wang, Z.Q., Kempa, K., Wang, Y.M., Jo, S.H., Chen, G., Dresselhaus, M.S. and Ren, Z. F., Nature, Vol. 439 (2006) p. 281 [172] Lourie, O., Cox, D.M. and Wagner, H.D., Physics Review Letters, Vol. 81 (1998) pp. 1638-1641 [173] Treacy, M.M.J., Ebbesen, T.W. and Gibson, J.M. , Nature, Vol. 381 (1996) pp. 678-680 [174] Wang, X., Ye, J., Wang, Y. and Chen, L., Journal of the American Ceramic Society, Vol. 90 (2007) pp. 962-964 [175] Sealy, C., Materialstoday, Vol. 7 (2004) p. 15 [176] Gao, L., Jiang, L. and Sun, J., Journal of Electroceramics, Vol. 17 (2006) pp. 51-55 [177] Zhao, Z., Buscaglia, V., Bowen, P. and Nygren, M., Key Engineering Materials, Vol. 264268 (2004) pp. 2297-2300 [178] Anstis, G.R., Chantikul, P., Lawn, B.R. and Marshall, D.B., Journal of the American Ceramic Society, Vol. 64 (1981) pp. 533-538 [179] Jiang, D., Thomson, K., Kuntz, J.D., Ager, J.W. and Mukherjee, A.K., Scripta Materialia, Vol. 56 (2007) pp. 959-962 [180] Curtin, W.A. and Sheldon, B.W., Materialstoday, Vol. 7 (2004) pp. 44-49 [181] Thomson, K.E., Jiang, D., Ritchie, R.O. and Mukherjee, A.K., Applied Physics A, Vol. 89 (2007) pp. 651-654 [182] Jiang, D. and Mukherjee, A.K., Scripta Materialia, Vol. 58 (2008) pp. 991-993 [183] Padture, N.P. and Curtin, W.A., Scripta Materialia, Vol. 58 (2008) pp. 989-990 [184] Sheldon, B.W. and Curtin, W.A., Nature Materials, Vol. 3 (2004) pp. 505-506 [185] Quinn, G.D. and Bradt, R.C., Journal of the American Ceramic Society, Vol. 90 (2007) pp. 673-680 [186] Gatto, A., Journal of Materials Processing Technology, Vol. 174 (2006) pp. 67-73 [187] Zhu, Y.F., Shi, L., Liang, J., Hui, D. and Lau, K.T., Composites Part B: Engineering, Vol. 39 (2008) pp. 1136-1141 [188] Peigney, A., Nature Materials, Vol. 2 (2003) pp. 15-16 [189] Belmonte, M., Advanced Engineering Materials, Vol. 8 (2006) pp. 693-703 [190] Madtha, S., Lee, C. and Chandran, K.S.R., Journal of the American Ceramic Society, Vol. 91 (2008) pp. 1319-1321 [191] Li, J.Y., Dai, H., Zhong, X.H., Zhang, Y.F., Ma, X.F., Meng, J. and Cao, X.Q., Journal of Alloys and Compounds, Vol. 452 (2008) pp. 406-409 [192] Xia, Z., Riester, L., Curtin, W.A., Li, H., Sheldon, B.W., Liang, J., Chang, B. and Xu, J.M., Acta Materialia, Vol. 52 (2004) pp. 931-944 206 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [193] Laurent, C., Peigney, A., Dumortier, O. and Rousset, A., Journal of European Ceramic Society, Vol. 18 (1998) pp. 2005-2013 [194] Solvas, E.Z., Poyato, R., Garcia, D.G., Rodriguez, A.D., Radmilovic, V. and Padture, N.P., Applied Physics Letters, Vol. 92 (2008) pp. 111912 [195] Vasiliev, A.L., Poyato, R. and Padture, N.P., Scripta Materialia, Vol. 56 (2007) pp. 461463 [196] Xia, Z.H., Lou, J. and Curtin, W.A., Scripta Materialia, Vol. 58 (2008) pp. 223-226 [197] Peigney, A., Rul, S., Schlick, F.L. and Laurent, C., Journal of European Ceramic Society, Vol. 27 (2007) pp. 2183-2193 [198] Wasche, R. and Klaffke, D., Wear, Vol. 249 (2001) pp. 220-228 [199] Lu, K., Journal of Materials Science, Vol. 43 (2008) pp. 652-659 [200] Balazsi, C., Shen, Z., Konya, Z., Kasztovszky, Z., Weber, F., Vertesy, Z., Biro, L.P., Kiricsi, I. and Arato, P., Composite Science and Technology, Vol. 65 (2005) pp. 727-733 [201] An, L., Xu, W., Rajagopalan, S., Wang, C., Wang, H., Fan, Y., Zhang, L., Jiang, D., Kapat, J., Chow, L., Guo, B., Liang, J. and Vaidyanathan, R., Advanced Materials, Vol. 16 (2004) pp. 2036-2040 [202] Manocha, L.M., Materials Science and Engineering A, Vol. 412 (2005) pp. 27-30 [203] Burghard, Z., Schon, D., Garstel, P., Bill, J. and Aldinger, F., International Journal of Materials Research, Vol. 97 (2006) pp. 1667-1672 [204] Wang, J., Kou, H.M., Pan, Y.B. and Guo, J.K., Key Engineering Materials, Vol. 313 (2006) pp. 145-150 [205] Luo, T.Y., Liang, T.X. and Li, C.S., Materials Science and Engineering A, Vol. 336 (2004) pp. 206-209 [206] Sun, Z., Zhang, X., Na, N., Liu, Z., Han, B. and An, G., Journal of Physical Chemistry, Vol. 110 (2006) pp. 13410-13414 [207] Shan, Y. and Gao, L., Nanotechnology, Vol. 16 (2005) pp. 625-630 [208] Pang, L.X., Sun, K.N., Ren, S., Sun, C. and Bi, J.Q., Journal of Composite Materials, Vol. 41 (2007) pp. 2025-2031 [209] Moriarty, P., Reports on Progress in Physics, Vol. 64 (2001) pp. 297-381 [210] Ebbesen, T.W., Lezec, H.J., Hiura, H., Bennett, J.W., Ghaemi, H.F. and Thio, T., Nature, Vol. 382 (1996) pp. 54-56 [211] Callister, W.D.J., Materials Science and Engineering: An Introduction, Sixth Edition, Wiley, USA (2003) [212] Goh, S.K., Kaiser, A.B., Lee, S.W., Lee, D.S., Yu, H.Y. and Park, Y.W., Current Applied Physics, Vol. 6 (2006) pp. 919-924 [213] Sheng, P., Physics Review B, Vol. 21 (1980) pp. 2180-2195 207 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [214] Shi, S.L., Pan, W., Fang, M.H. and Fang, Z.Y., Applied Physics Letters, Vol. 87 (2005) pp. 172902 [215] Deng, F. and Zheng, Q.S., Applied Physics Letters, Vol. 92 (2008) pp. 071902 [216] Kaiser, A.B., Duesberg, G. and Roth, S., Physical Review B: Condensed Matter, Vol. 57 (1998) p. 1418 [217] Berber, S., Kwon, Y.K. and Tomanek, D., Physical Review Letters, Vol. 84 (2000) pp. 4613-4616 [218] Che, J., Cagin, T. and Goddard, W.A., Nanotechnology, Vol. 11 (2000) pp. 65-69 [219] Bakshi, S.R., Balani, K. and Agarwal, A., Journal of the American Ceramic Society, Vol. 91 (2008) pp. 942-947 [220] Xue, Q.Z., Physica B: Condensed Matter, Vol. 368 (2005) pp. 302-307 [221] Zhang, H.L., Lia, J.F., Yao, K.F. and Chen, L.D., Journal of Applied Physics, Vol. 97 (2005) pp. 114310 [222] Huang, Q., Gao, L., Liu, Y. and Sun, J., Journal of Materials Chemistry, Vol. 15 (2005) pp. 1995-2001 [223] Meyyappan, M., Carbon nanotubes: Science and Applications, CRC Press Inc., USA (2004) [224] Sivakumar, R., Guo, S., Nishimura, T. and Kagawa, Y., Scripta Materialia, Vol. 56 (2007) pp. 265-268 [225] Chin, Y.H., Hu, J., Cao, C., Gao, Y. and Wang, Y., Catalysis Today, Vol. 110 (2005) pp. 47-52 [226] Jiang, L. and Gao, L., Ceramics International, Vol. 34 (2008) pp. 231-235 [227] Nakamatsu, T., Pomar, F. and Ishizaki, K., Journal of the Materials Science, Vol. 34 (1999) pp. 1553-1556 [228] Ning, J., Zhang, J., Pan, Y. and Guo, J., Journal of Materials Science Letters, Vol. 22 (2003) pp. 1019-1022 [229] Lima, R.S., Khor, K.A., Li, H., Cheang, P. and Marple, B.R., Materials Science and Engineering A, Vol. 396 (2005) pp. 181-187 [230] Dou, S.X., Yeoh, W.K., Shcherbakova, O., Weyler, D., Li, Y., Ren, Z.M., Munroe, P., Chen, S.K., Tan, K.S., Glowacki, B.A. and Driscoll, J.L.M., Advanced Materials, Vol. 18 (2006) pp. 785-788 [231] Yeoh, W.K., Horvat, J., Kim, J.H., Xu, X. and Dou, S.X., IEEE Transactions on Applied Superconductivity, Vol. 17 (2007) pp. 2929-2932 [232] Kovac, P., Husek, I., Skakalova, V., Meyer, J., Dobrocka, E., Hirscher, M. and Roth, S., Superconductor Science and Technology, Vol. 20 (2007) pp. 105-111 [233] Sun, Z., Liu, Z., Wang, Y., Han, B., Du, J. and Zhang, J., Journal of Materials Chemistry, Vol. 15 (2005) pp. 4497-4501 208 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [234] Cao, J., Sun, J.Z., Hong, J., Li, H.Y., Chen, H.Z. and Wang, M., Advanced Materials, Vol. 16 (2004) pp. 84-87 [235] Lee, T.Y., Alegaonkar, P.S. and Yoo, J.B., Thin Solid Films, Vol. 515 (2007) pp. 51315135 [236] Pan, L., Shoji, T., Nagataki, A. and Nakayama, Y., Advanced Engineering Materials, Vol. 9 (2007) pp. 584-587 [237] Morihisa, Y., Kimura, C., Yukawa, M., Aoki, H., Kobayashi, T., Hayashi, S., Akita, S., Nakayama, Y. and Sugino, T., Journal of Vacuum Science and Technology B, Vol. 26 (2008) pp. 872-875 [238] Su, C.Y., Juang, Z.Y., Chen, Y.L., Leou, K.C. and Tsai, C.H., Diamond and Related Materials, Vol. 16 (2007) pp. 1393-1397 [239] Yang, C.J., Park, J. and Cho, Y.R., Advanced Engineering Materials, Vol. 9 (2007) pp. 8891 [240] Jin, F., Liu, Y. and Day, C.M., Applied Physics Letters, Vol. 88 (2006) pp. 163116 [241] Pan, L., Konishi, Y., Tanaka, H., Chakrabarti, S., Hokushin, S., Akita, S. and Nakayama, Y., Journal of Vacuum Science and Technology B, Vol. 25 (2007) pp. 1581-1583 [242] Seeger, T., Köhler, T., Frauenheim, T., Grobert, N., Rühle, M., Terrones, M. and Seifert, G., Chemical Communication, Vol. 1 (2002) pp. 34-35 [243] Han, W.Q. and Zettl, A., Nano letters, Vol. 3 (2003) pp. 681-683 [244] Fan, W., Gao, L. and Sun, J., Journal of the American Ceramic Society, Vol. 89 (2006) pp. 731-733 [245] Espinosa, E.H., Ionescu, R., Chambon, B., Bedis, G., Sotter, E., Bittencourt, C., Felten, A., Pireaux, J.J., Correig, X. and Llobet, E., Sensors and Actuators B: Chemical, Vol. 127 (2007) pp. 137-142 [246] Ma, C.A., Tang, J.Y., Li, G.H. and Sheng, J.F., Acta Chimica Sinica, Vol. 64 (2006) pp. 2123-2126 [247] Li, G., Maa, C., Tang, J. and Sheng, J., Electrochemica Acta, Vol. 52 (2005) pp. 20182023 [248] Li, X., Kang, F. and Shen, W., Carbon, Vol. 44 (2006) pp. 1334-1336 [249] Zhang, Q.T., Qu, M.Z., Niu, H. and Yu, Z.L., New Carbon Materials, Vol. 22 (2007) pp. 361-364 [250] Hakim, L.F., King, D.M., Zhou, Y., Gump, C.J., George, S.M. and Weimer, A.W., Advanced Functional Materials, Vol. 17 (2007) pp. 3175-3181 [251] Zhan, G.D., Du, X., King, D.M., Hakim, L.F., Liang, X.H., McCormick, J.A. and Weimer, A.W., Journal of the American Ceramic Society, Vol. 91 (2008) pp. 831–835 [252] Farmer, D.B. and Gordon, R.G., Nano Letters, Vol. 6 (2006) pp. 699-703 209 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [253] Pierson, H.O., Handbook of Carbon, Graphite, Diamond and Fullerenes, Noyes Publications, USA (1993) [254] Carbon Black, www.ccohs.ca/oshanswers/chemicals/chem_profiles/carbonbl/ [255] Andersson, J.M., Controlling the Formation and Stability of Alumina Phases, PhD Thesis, Linköping University, Sweden (2005) [256] Feret, F.R., Roy, D. and Boulanger, C., Spectrochimica Acta Part B, Vol. 55 (2000) pp. 1051-1061 [257] Wyckoff, R.W.G., Crystal Structures, Second Edition, Vol. 2, John Wiley & Sons, USA (1964) [258] Ebonex® Technology, www.atraverda.co.uk [259] Advanced Ceramics, www.hcstarck.com [260] Zetasizer Nano Series User Manual, Man 0317, Malvern Instruments Ltd., UK (2004) [261] Tamburini, U.A., Garay, J.E. and Munir, Z.A., Scripta Materialia, Vol. 54 (2006) pp. 823828 [262] Olevsky, E.A., Kandukuri, S. and Froyen, L., Journal of Applied Physics, Vol. 102 (2007) pp. 114913 [263] Omori, M., Materials Science and Engineering A: Structural Materials: Properties, Microstructure and Processing, Vol. 287 (2000) pp. 183-188 [264] Munir, Z.A., Tamburini, U.A. and Ohyanagi, M., Journal of Materials Science, Vol. 41 (2006) pp. 763-777 [265] Yamamoto, G., Sato, Y., Takahashi, T., Omori, M., Hashida, T., Okubo, A. and Tohji, K., Advanced Functional Materials, Vol. 21 (2006) pp. 1537-1542 [266] Shen, Z., Johnsson, M., Zhao, Z. and Nygren, M., Journal of the American Ceramic Society, Vol. 85 (2002) pp. 1921-1927 [267] Mishra, R.S., Mukherjee, A.K., Yamazaki, K. and Shoda, K., Journal of Materials Research, Vol. 11 (1996) pp. 1144-1148 [268] Nygren, M. and Shen, Z., Solid State Science, Vol. 5 (2003) pp. 125-131 [269] Hong, J., Gao, L., Torre, S.D.D.L., Miyamoto, H. and Miyamoto, K., Materials Letters, Vol. 43 (2000) pp. 27-31 [270] Harada, Y., Uekawa, N., Kojima, T. and Kakegawa, K., Journal of the European Ceramic Society, Vol. 28 (2008) pp. 235-240 [271] Lee, Y.I., Lee, J.H., Hong, S.H. and Kim, D.Y., Materials Research Bulletin, Vol. 38 (2003) pp. 925-930 [272] Mukhopadhyay, A. and Basu, B., International Metals Reviews, Vol. 52 (2007) pp. 257288 [273] Lee, J.H., Ko, S.K. and Won, C.W., Materials Research Bulletin, Vol. 36 (2000) pp. 989996 210 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [274] Meng, F., Fu, Z., Zhang, J., Wang, H., Wang, W., Wang, Y. and Zhang, Q., Journal of the American Ceramic Society, Vol. 90 (2007) pp. 1262-1264 [275] Reisel, A.D., Schops, S., Lenk, A. and Schmutzler, G., Advanced Engineering Materials, Vol. 9 (2007) pp. 400-405 [276] Freim, J., McKittrick, J., Katz, J. and Sickafus, K., Nanostructured Materials, Vol. 4 (1994) pp. 371-385 [277] FAST, Field Assisted Sintering Technique, www.fct-ceramic.com [278] AccuPycTM Gas Displacement Pycnometre System, Product Manual, Micromeritics, USA (2003) [279] Barsoum, M.W., Fundamentals of Ceramics, Second Edition, Institute of Physics, UK (2003) [280] ASTM C 1327-03, Standard test method for Vickers indentation hardness of advanced ceramics [281] Strecker, K., Ribeiro, S., and Hoffmann, M.J., Materials Research, Vol. 8 (2005) pp. 121124 [282] Valentin, E., Auvray, S., Goethals, J., Lewenstein, J., Capes, L., Filoramo, A., Ribayrol, A., Tsui, R., Bourgoin, J.P. and Patillon, J.N., Microelectronic Engineering, Vol. 61-62 (2002) pp. 491-496 [283] Liu, J., Casavant, M.J., Cox, M., Walters, D.A., Boul, P., Lu, W., Rimberg, A.J., Smith, K.A., Colbert, D.T. and Smalley, R.E., Chemical Physics Letters, Vol. 303 (1999) pp. 125129 [284] Ausman, K.D., Piner, R., Lourie, O., Ruoff, R.S. and Korobov, M., The Journal of Physical Chemistry B, Vol. 104 (2000) pp. 8911-8915 [285] Marcus, Y., Journal of Solution Chemistry, Vol. 20 (1991) pp. 929-944 [286] Poyato, R., Vasiliev, A.L., Padture, N.P., Tanaka, H. and Nishimura, T., Nanotechnology, Vol. 17 (2006) pp. 1770-1777 [287] Wang, Z., Ciselli, P. and Peijs, T., Nanotechnology, Vol. 18 (2007) pp. 455709 [288] Ham, H.T., Choi, Y.S. and Chung, I.J., Journal of Colloid and Interface Science, Vol. 286 (2005) pp. 216-223 [289] Lisunova, M.O., Lebovka, N.I., Melezhyk, O. V. and Boiko, Y.P., Journal of Colloid and Interface Science, Vol. 299 (2006) pp. 740-746 [290] Chen, Q., Saltiel, C., Manickavasagam, S., Schadler, L.S., Siegel, R.W. and Yang, H., Journal of Colloid and Interface Science, Vol. 280 (2005) pp. 91-97 [291] Jiang, L., Gao, L. and Sun, J., Journal of Colloid and Interface Science, Vol. 260 (2003) pp. 89-94 [292] Banerjee, S. and Wong, S.S., Nano Letters, Vol. 2 (2002) pp. 49-53 211 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [293] Bahr, J.L., Mickelson, E.T., Bronikowski, M.J., Smalley, R.E. and Tour, J.M., Chemical Communication, Vol. 2 (2001) pp. 193-194 [294] Lau, K., Lu, M., Lam, C., Cheung, H., Sheng, F.L. and Li, H.L., Composites Science and Technology, Vol. 65 (2005) pp. 719-725 [295] Moniruzzaman, M., Du, F., Romero, N. and Winey, K.I., Polymer, Vol. 47 (2006) pp. 293298 [296] Chae, H.G., Sreekumar, T.V., Uchida, T. and Kumar, S., Polymer, Vol. 46 (2005) pp. 10925-10935 [297] Bryning, M.B., Islam, M.F., Kikkawa, J.M. and Yodh, A.G., Advanced Materials, Vol. 17 (2005) pp. 1186-1191 [298] Bae, J., Jang, J. and Yoon, S.H., Macromolecular Chemistry and Physics, Vol. 203 (2002) pp. 2196-2204 [299] Boul, P.J., Liu, J., Mickelson, E.T., Huffman, C.B., Ericson, L.M., Chiang, I.W., Smith, K.A., Colbert, D.T., Hauge, R.H., Margrave, J.L. and Smalley, R.E., Chemical Physics Letters, Vol. 310 (1999) pp. 367-372 [300] Inam, F. and Peijs, T., Journal of Nanostructured Polymers and Nanocomposites, Vol. 2 (2006) pp. 87-95 [301] Tkac, J. and Ruzgas, T., Electrochemistry Communications, Vol. 8 (2006) pp. 899-903 [302] Lee, K.C., Yu, H.H., Hwang, S.J., Li, Y.S., Cheng, M.H. and Lin, C.C., Materials Science Forum, Vols. 505-507 (2006) pp. 1075-1080 [303] Baughman, R.H., Zakhidov, A.A. and Heer, W.A.D., Science, Vol. 297 (2002) pp. 787792 [304] Grossiord, N., Regev, O., Loos, J., Meuldijk, J. and Koning, C.E., Analytical Chemistry, Vol. 77 (2005) pp. 5135-5139 [305] An, J.W. and Lim, D.S., Journal of Ceramic Processing Research, Vol. 3 (2002) pp. 201204 [306] Boccaccini, A.R., Acevedo, D.R., Brusatin, G. and Colombo, P., Journal of European Ceramic Society, Vol. 25 (2005) pp. 1515-1523 [307] Zackrisson, J., Andren, H.O., International Journal of Refractory Metals and Hard Materials, Vol. 17 (1999) pp. 265-273 [308] Lee, H. and Speyer, R.F., Journal of American Ceramic Society, Vol. 86 (2003) pp. 14681478 [309] Schwetz, K.A. and Grellner, W., Journal of Less-Common Metals, Vol. 82 (1981) pp. 3741 [310] Zorzi, J.E., Perottoni, C.A. and Jornada. J.A.H.D., Materials Letters, Vol. 59 (2005) pp. 2932-2935 212 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [311] Baharvandi, H.R., Hadian, A.M., Abdizadeh, A. and Ehsani, N., Journal of Materials Science, Vol. 41 (2006) pp. 5269-5272 [312] Erkalfa, H., Misirli, Z. and Baykara, T., Journal of Materials Processing Technology, Vol. 62 (1996) pp. 108-115 [313] Mishra, S.K. and Das, S.K., Materials Letters, Vol. 59 (2005) pp. 3467-3470 [314] Wang, H.Z., Gao, L. and Guo, J.K., Ceramics International, Vol. 26 (2000) pp. 391-396 [315] Park, S.Y., Jung, S.W. and Chung, Y.B., Ceramics International, Vol. 29 (2003) pp. 707712 [316] Yang, H., Shang, F. and Gao, L., Ceramics International, Vol. 33 (2007) pp. 1521-1524 [317] Tuan, W.H., Lin, M.C. and Tzing, W.H., Materials Chemistry and Physics, Vol. 48 (1997) pp. 156-159 [318] Fang, J.X., Thompson, A.M., Harmer, M.P. and Chan, H.M., Journal of the American Ceramics Society, Vol. 80 (1997) pp. 2005-2012 [319] Toyoda, T. and Shimamoto, S., Materials Science and Engineering B, Vol. 54 (1998) pp. 29-32 [320] Berry, K.A. and Harmer, M.P., Journal of American Ceramics Society, Vol. 69 (1965) pp. 143-149 [321] Xie, Z., Yang, J. and Huang, Y., Materials Letters, Vol. 37 (1998) pp. 215-220 [322] Yeheskel, O., Chaim, R., Shen, Z. And Nygren, M., Journal of Materials Research, Vol. 20 (2005) pp. 719-725 [323] Nakagawa, H., Tanaka, H., Kasama, A., Anno, H. And Matsubara, K., Proceedings of 16th International Conference on Thermoelectrics, (1997) pp. 351-355 [324] Soyez, G., Eastman, J.A., Thompson, L.J., Bai, G.R., Baldo, P.M., McCormick, A.W., DiMelfi, R.J., Elmustafa, A.A., Tambwe, M.F. and Stone, D.S., Applied Physics Letters, Vol. 77 (2000) pp. 1155-1157 [325] Tschope, A., and Birringer, R., Journal of Electroceramics, Vol. 7 (2001) pp. 169-177 [326] Goh, G.K.L., Lim, L.C., Rahman, M. and Lim, S.C., Wear, Vol. 206 (1997) pp. 24-32 [327] You, X.Q., Si, T.Z., Liu, N., Ren, P.P., Xu, Y.D. and Feng, J.P., Ceramics International, Vol. 31 (2005) pp. 33-38 [328] Morsi, K., Keshavan, H. and Bal, S., Journal of Materials Science: Materials in medicine, Vol. 15 (2004) pp. 191-197 [329] Sandler, J.K.W., Kirk, J.E., Kinloch, I.A., Shaffer, M.S.P. and Windle, A.H., Polymer, Vol. 44 (2003) pp. 5893-5899 [330] Menchavez, R.L., Fuji, M. and Takahashi, M., Advanced Materials, Vol. 20 (2008) pp. 2345-2351 [331] Menchavez, R.L., Fuji, M., Yamakawa, T., Endo, T. and Takahashi, M., Materials Science Forum, Vol. 2123 (2007) pp. 561-565 213 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [332] Wei, T., Fan, Z., Luo, G., Wei, F., Zhao, D. and Fan, J., Materials Research Bulletin, Vol. 43 (2008) pp. 2806-2809 [333] Balazsi, C., Sedlackova, K. and Czigany, Z., Composites Science and Technology, Vol. 68 (2008) pp. 1596-1599 [334] Dresselhaus, M.S. and Eklund, P.C., Advacnes in Physics, Vol. 49 (2000) pp. 705-814 [335] Osswald, S., Havel, M. and Gogotsi, Y., Journal of Raman Spectroscopy, Vol. 38 (2007) pp. 728-736 [336] Dresselhaus, M.S., Dresselhaus, G., Jorio, A., Filho, A.G.S. and Saito, R., Carbon, Vol. 40 (2002) pp. 2043-2061 [337] Dresselhaus, M.S., Dresselhaus, G., Saito, R. and Jorio, A., Physics Reports, Vol. 409 (2005) pp. 47-99 [338] Tuinstra, F. and Koenig, J.L., Journal of Chemical Physics, Vol. 53 (1970) pp. 1126-1130 [339] Cuesta, A., Dhamelincourt, P., Laureyns, J., Alonso, A.M. and Tascon, J.M.D., Journal of Materials Chemistry, Vol. 8 (1998) pp. 2875-2879 [340] Bahr, J.L. and Tour, J.M., Journal of Materials Chemistry, Vol. 12 (2002) pp. 1952-1958 [341] Delhaes, P., Couzi, M., Trinquecoste, M., Dentzer, J., Hamidou, H. and Guterl, C.V., Carbon, Vol. 44 (2006) pp. 3005-3013 [342] Jorio, A., Pimenta, M.A., Filho, A.G.S., Saito, R., Dresselhaus, G. and Dresselhaus, M.S., New Journal of Physics, Vol. 5 (2003) pp. 139.1-139.17 [343] Krasheninnikov, A.V. and Banhart, F., Nature Materials, Vol. 6 (2007) pp. 723-733 [344] Zhang, F., Shen, J., Sun, J., Zhu, Y.Q., Wang, G. and McCartney, G., Carbon, Vol. 43 (2005) pp. 38-42 [345] Shen, J., Zhang, F.M., Sun, J.F., Zhu, Y.Q. and McCartney, D.G., Nanotechnology, Vol. 17 (2006) pp. 2187-2191 [346] Wei, B., Zhang, J., Liang, J. and Wu, D. , Carbon, Vol. 36 (1998) pp. 997-1001 [347] Zhu, Y.Q., Sekine, T., Kobayashi, T., Takazawa, E., Terrones, M. and Terrones, H., Chemical Physics Letters, Vol. 287 (1998) pp. 689-693 [348] Sun, L.T., Gong, J.L., Zhu, D.Z., Zhu, Z.Y. and He, S.X., Advanced Materials, Vol. 16 (2004) pp. 1849-1853 [349] Monthioux, M., Burteaux, B., Claye, A., Fischer, J.E. and Luzzi, D.E., Carbon, Vol. 39 (2001) pp. 1251-1272 [350] Li, J.L., Wang, L.J., He, T. and Jiang, W., Carbon, Vol. 45 (2007) pp. 2636-2642 [351] Yu, J.K., Ueno, S., Li, H.X. and Hiragushi, K., Journal of the European Ceramic Society, Vol. 19 (1999) pp. 2843-2848 [352] Sinclair, R., Itoh, T. and Chin, R., Microscopy and Microanalysis, Vol. 8 (2002) pp. 288304 214 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [353] Wei, B., Zhang, J., Liang, J., Liu, W., Gao, Z. and Wu, D., Journal of Materials Science Letters, Vol. 16 (1997) pp. 402-403 [354] Yamamoto, O., Sasamoto, T. and Inagaki, M., Journal of Materials Science Letters, Vol. 19 (2000) pp. 1053-1055 [355] Costache, M.C., Wang, D.Y., Heidecker, M.J., Manias, E. and Wilkie, C.A., Polymer for Advanced Technologies, Vol. 17 (2006) pp. 272-280 [356] Yang, J., Lin, Y.H., Wang, J.F., Lai, M., Li, J., Liu, J., Tong, X. and Cheng, H., Journal of Applied Polymer Science, Vol. 98 (2005) pp. 1087-1091 [357] Bocchini, S., Frache, A., Camino, G. and Michael, C., European Polymer Journal, Vol. 43 (2007) pp. 3222-3235 [358] Pan, L., Konishi, Y., Tanaka, H, Chakrabarti, S., Hokushin, S., Akita, S. and Nakayama, Y. (2007) Journal of Vacuum Science and Technology B, Vol. 25 (2007) pp. 1581-1583 [359] Son, Y.W., Han, S and Ihm, J., New Journal of Physics, Vol. 5 (2003) pp.152.1-152.9 [360] Kawasaki, S., Catalan, G., Fan, H., Saad, M.M., Gregg, J.M., Duarte, M.A.C., Rybczynski, J., Morrison, F.D., Tatsuta, T., Tsuji, O. and Scott, J.F., Applied Physics Letters, Vol. 92 (2008) pp. 053109 [361] Fu, L., Liu, Y., Liu, Z., Han, B.X., Cao, L.C., Wei, D.C, Yu, G. and Zhu, D.B., Advanced Materials, Vol. 18 (2006) pp. 181-185 [362] Wang, Y.H., Li, Y.N., Lu, J., Zang, J.B. and Huang, H., Nanotechnology, Vol. 17 (2006) pp. 3817-3821 [363] Li, H., Zhao, N., He, C., Shi, C., Du, X. and Li, J., Materials Science and Engineering A, Vol. 473 (2008) pp. 355-359 [364] Yao, N., Lordi, V., Ma, S.X.C, Dujardin, E., Krishnan, A., Treacy, M.M.J and Ebbesen, T.W., Journal of Materials Research, Vol. 13 (1998) pp. 2432-2437 [365] Tian, C.H., Ren, H.F., Geng, D.T. and Tian, X.X., 7th International Symposium on Test and Measurement, Vol. 1-7 (2007) pp. 4880-4882 [366] Chen, C.M., Chen, M., Leu, F.C., Hsu, S.Y., Wang, S.C., Shi, S.C. and Chen, C.F., Diamond and Related Materials, Vol. 13 (1998) pp. 1182-1186 [367] Guo, Z., Development of Si3N4 based composites for high-temperature sensor application, PhD Thesis, University of London, UK (2006) [368] ASTM C 1161-02c, Standard test method for flexural strength of advanced ceramics at ambient temperature, re-approved in 2008 [369] Krell, A. and Blank, P., Journal of the American Ceramic Society, Vol. 78 (1995) pp. 1118-1120 [370] Zhu, C., Guo, W., Yu , T.X. and Woo, C.H., Nanotechnology, Vol. 16 (2005) pp. 10351039 215 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [371] Rice, R.W., Mechanical properties of ceramics and composites – grain and particle effects, Marcel Dekker Inc., USA (2003) [372] Teng, X., Liu, H. and Huang, C., Materials Science and Engineering A, Vol. 452-453 (2007) pp. 545-551 [373] Lawn, B., Fracture of Brittle Solids, Second Edition, Cambridge University Press, UK (1993) [374] Hasselman, D.P.H., Journal of the American Ceramic Society, Vol. 52 (1969) pp. 600-604 [375] Deng, C.F., Ma, Y.X., Zhang, P., Zhang, X.X. and Wang, D.Z., Materials Letters, Vol. 62 (2008) pp. 2301-2303 [376] Davis, J.B., Marshall, D.B. and Morgan, P.E.D., Journal of the European Ceramic Society, Vol. 20 (2000) pp. 583-587 [377] Riddick, J.A. and Bunger, W.B., Organic Solvents; Physical Properties and Methods of Purification, Third Edition, John Wiley & Sons Inc., USA (1970) [378] Shaw, D.J., Introduction to Colloid and Surface Chemistry, Fourth Edition, ButterworthHeinemann Ltd., UK (1991) [379] Clayden, J., Greeves, N., Warren, S., Wothers, P., Organic Chemistry, Oxford University Press, UK (2007) [380] Martelli, D., La Medicina del Lavoro, Vol. 51 (1960) pp. 123-127 [381] Dreisbach, R.R., Physical Properties of Chemical Compounds III, American Chemical Society, USA (1961) [382] Stokinger, H.E., Documentation of Threshold Limit Values, American Conference of Governmental Industrial Hygienists, USA (1966) [383] Port, O., Business Week, http://www.businessweek.com, November 27 (2000) Issue [384] Chapana, J.A.R. and Giersig, M., Journal of Nanoscience and Nanotechnology, Vol. 6 (2006) pp. 316-321 [385] Li, J., Cassell, A.M. and Dai, H., Surface and Interface Analysis, Vol. 28 (1999) pp. 8-11 [386] BMC Swiss Cycling Technology, www.bmc-racing.com [387] SC 900 Easton Stealth Baseball Bat, www.anacondasports.com [388] Babolat: Tennis, Badminton, Paddle, www.babolat.com [389] Montreal Hockey Sticks, www.montrealhockey.com [390] Zhan, G., Kuntz, J.D. and Mukherjee, A.K., US Patent 6976532, 2005 [391] Wang, P. and Zhu, G., Electrochemistry Communications, Vol. 4 (2002) pp. 36-40 [392] Baroch, P., Saito, N. and Takai, O., Journal of Physics D: Applied Physics, Vol. 41 (2008) pp. 085207 [393] Hu, G., Wang, W.L., Wang, S.X., Zhu, W. and Li, Y., Diamond and Related Materials, Vol. 12 (2003) pp. 1295-1299 216 References Development of Ceramic – Carbon Nanotube (CNT) Nanocomposites [394] Perale, G., Giordano, C., Daniele, F., Tunesi, M., Colombo, P., Gottardo, L., Maccagnan, S. and Masi, M., International Journal of Artificial Organs, Vol. 31 (2008) pp. 272-278 [395] Pelissier, K., Chartier, T. and Laurent, J.M., Ceramics International, Vol. 24 (1998) pp. 371-377 [396] Vyshnyakov, L.R., Mazna, A.V., Neshpor, A.V., Kokhanyi, V.A. and Oleksyuk, O.N., Strength of Materials, Vol. 36 (2004) pp. 643-648 [397] Bayer Baytubes, www.baytubes.com [398] Heintz, M.E., Nanotechnology Law Report, www.nanolawreport.com (May 19, 2008) [399] Fiedler, B., Gojny, F.H., Wichmann, M.H.G., Nolte, M.C.M. and Schulte, K., Composites Science and Technology, Vol. 66, No. 16 (2006) pp. 3115-3125 [400] Todd, A., Dictionary of Ceramics, Third Edition, The Institute of Materials, UK (1994) 217