Carbon nanotube fibres and yarns: production, properties and applications in smart textiles 9780081027226, 0081027222

Citation preview

CARBON NANOTUBE FIBERS AND YARNS

The Textile Institute Book Series Incorporated by Royal Charter in 1925, The Textile Institute was established as the professional body for the textile industry to provide support to businesses, practitioners and academics involved with textiles and to provide routes to professional qualifications through which Institute Members can demonstrate their professional competence. The Institute’s aim is to encourage learning, recognise achievement, reward excellence and disseminate information about the textiles, clothing and footwear industries and the associated science, design and technology; it has a global reach with individual and corporate members in over 80 countries. The Textile Institute Book Series supersedes the former ‘Woodhead Publishing Series in Textiles’ and represents a collaboration between The Textile Institute and Elsevier aimed at ensuring that Institute Members and the textile industry continue to have access to high calibre titles on textile science and technology. Books published in The Textile Institute Book Series are offered on the Elsevier web site at: store.elsevier.com and are available to Textile Institute Members at a substantial discount. Textile Institute books still in print are also available directly from the Institute’s web site at: www.textileinstitute.org To place an order, or if you are interested in writing a book for this series, please contact Matthew Deans, Senior Publisher: [email protected]

Recently Published and Upcoming Titles in The Textile Institute Book Series: New Trends in Natural Dyes for Textiles, Padma Vankar Dhara Shukla, 978-0-08-102686-1 Smart Textile Coatings and Laminates, William C. Smith, 2nd Edition, 978-0-08-102428-7 Advanced Textiles for Wound Care, 2nd Edition, S. Rajendran, 978-0-08-102192-7 Manikins for Textile Evaluation, Rajkishore Nayak Rajiv Padhye, 978-0-08-100909-3 Automation in Garment Manufacturing, Rajkishore Nayak and Rajiv Padhye, 978-0-08-101211-6 Sustainable Fibres and Textiles, Subramanian Senthilkannan Muthu, 978-0-08-102041-8 Sustainability in Denim, Subramanian Senthilkannan Muthu, 978-0-08-102043-2 Circular Economy in Textiles and Apparel, Subramanian Senthilkannan Muthu, 978-0-08-102630-4 Nanofinishing of Textile Materials, Majid Montazer Tina Harifi, 978-0-08-101214-7 Nanotechnology in Textiles, Rajesh Mishra Jiri Militky, 978-0-08-102609-0 Inorganic and Composite Fibers, Boris Mahltig Yordan Kyosev, 978-0-08-102228-3 SmartTextiles for In Situ Monitoring of Composites, Vladan Koncar,978-0-08-102308-2 Handbook of Properties of Textile and Technical Fibres, 2nd Edition, A. R. Bunsell, 978-0-08-101272-7 Silk, 2nd Edition, K. Murugesh Babu, 978-0-08-102540-6

The Textile Institute Book Series

CARBON NANOTUBE FIBERS AND YARNS Production, Properties, and Applications in Smart Textiles Edited by

MENGHE MIAO

An imprint of Elsevier

Woodhead Publishing is an imprint of Elsevier The Officers’ Mess Business Centre, Royston Road, Duxford, CB22 4QH, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, OX5 1GB, United Kingdom Copyright © 2020 Elsevier Ltd. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-08-102722-6 For information on all Woodhead publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Matthew Deans Acquisition Editor: Brian Guerin Editorial Project Manager: Aleksandra Packowska Production Project Manager: Joy Christel   Neumarin Honest Thangiah Cover Designer: Miles Hitchen Typeset by SPi Global, India

Contents Contributors ix

1. Introduction

1

Menghe Miao 1.1 A brief introduction to carbon nanotubes 1 1.2 CNT yarns versus conventional textile yarns 3 1.3 Prospects of high-strength CNT yarns 5 1.4 Potential applications 7 1.5 Organization of the book 7 References 10

Part I   Manufacturing methods 2. Yarn production from carbon nanotube forests

13

Menghe Miao 2.1 Synthesis of vertically aligned CNT arrays 13 2.2 Drawing a CNT web 15 2.3 Yarn formation 20 2.4 Production scale-up 30 2.5 Summary and conclusion 32 References 34

3. Carbon nanotube fibers spun directly from furnace

37

Guangfeng Hou, Mark J. Schulz 3.1 Introduction 37 3.2 Floating catalyst synthesis of CNTs 38 3.3 CNT assembly and fiber production 46 3.4 Structure and properties of CNT fibers 47 3.5 Conclusions and future directions 53 Acknowledgment 55 References 55 Further reading 59

v

vi

Contents

4. Solution-spun carbon nanotube fibers

61

Menghe Miao 4.1 Introduction 61 4.2 Spinning from surfactant-based solutions 62 4.3 Spinning from acid solutions 65 4.4 Alternative wet-spinning route 68 References 68

5. Interphase structures and properties of carbon nanotube-reinforced polymer nanocomposite fibers

71

Fengying Zhang, Yaodong Liu 5.1 Introduction 71 5.2 Unique properties for CNT-reinforced nanocomposites 73 5.3 Examples of high-performance CNT-reinforced polymeric fibers 85 5.4 Challenges 91 5.5 Perspective 95 References 96

6. Post-spinning treatments to carbon nanotube fibers

103

Hai Minh Duong, Sandar Myo Myint, Thang Quyet Tran, Duyen Khac Le 6.1 Twist insertion 104 6.2 Liquid densification 105 6.3 Coating and doping 106 6.4 Acid treatment 107 6.5 Mechanical densification and stretching 115 6.6 Infiltration 116 6.7 Irradiation 118 6.8 Hybrid treatments of CNT fibers 120 6.9 Conclusions and recommendations 129 References 131

Part II  Structures and properties 7. Carbon nanotube yarn structures and properties

137

Menghe Miao 7.1 CNT yarn geometry 7.2 Tensile strength of CNT fibers and yarns 7.3 Dynamic properties 7.4 Electrical conductivity

137 155 167 170



Contents

vii

7.5 Thermal conductivity 173 7.6 Outlook for CNT fiber strength 174 7.7 Summary and future prospect 177 References 177

8. Mechanics modeling of carbon nanotube yarns

183

Xiaohua Zhang 8.1 Introduction 183 8.2 Analytic models for CNT yarns 184 8.3 Load transfer between nanotubes 192 8.4 Microstructural evolution of CNT yarns 194 8.5 Summary and outlook 202 Acknowledgments 203 References 203

Part III  Applications 9. Sensors based on CNT yarns

213

Jude C. Anike, Jandro L. Abot 9.1 Introduction 213 9.2 Damage sensors 215 9.3 Torque sensors 219 9.4 Wearable sensors 221 9.5 Foil strain gauges 223 9.6 Thermal sensors 229 9.7 Biochemical sensors 234 9.8 Summary and prospects for future research 237 References 237

10. CNT yarn-based supercapacitors

243

Qiufan Wang, Sufang Chen, Daohong Zhang 10.1 Introduction 243 10.2 Electrochemical properties of CNTs 250 10.3 Architectures of threadlike supercapacitors 253 10.4 Symmetrical and asymmetrical (or hybrid) threadlike supercapacitors 254 10.5 Self-charging supercapacitors 264 10.6 Potential applications and future directions 265 Acknowledgments 266 References 267 Further reading 270

viii

Contents

11. Carbon nanotube yarn-based actuators

271

Xiaohui Yang, Menghe Miao 11.1 Introduction 271 11.2 Yarn transformation due to insertion of extremely high twist 272 11.3 Actuator architectures 275 11.4 Energy conversion mechanisms 279 11.5 Performance metrics for actuators 282 11.6 Potential applications 284 11.7 Future prospects 288 References 288 Index 293

Contributors Jandro L. Abot Department of Mechanical Engineering, The Catholic University of America, Washington, DC, United States Jude C. Anike Department of Mechanical Engineering, The Catholic University of America, Washington, DC, United States Sufang Chen Key Laboratory for Green Chemical Process of Ministry of Education, Wuhan Institute of Technology, Wuhan, China Hai Minh Duong National University of Singapore, Singapore, Singapore Guangfeng Hou Department of Mechanical and Materials Engineering, University of Cincinnati, Cincinnati, OH, United States Duyen Khac Le National University of Singapore, Singapore, Singapore Yaodong Liu Key Laboratory of Carbon Materials, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan, China Menghe Miao CSIRO Manufacturing, Geelong,VIC, Australia Sandar Myo Myint National University of Singapore, Singapore, Singapore Mark J. Schulz Department of Mechanical and Materials Engineering, University of Cincinnati, Cincinnati, OH, United States Thang Quyet Tran National University of Singapore, Singapore, Singapore Qiufan Wang Key Laboratory of Catalysis and Energy Materials Chemistry of Ministry of Education & Hubei Key Laboratory of Catalysis and Materials Science, South-Central University for Nationalities, Wuhan, China Xiaohui Yang College of Materials Science and Engineering, Guizhou Minzu University, Guiyang, China

ix

x

Contributors

Daohong Zhang Key Laboratory of Catalysis and Energy Materials Chemistry of Ministry of Education & Hubei Key Laboratory of Catalysis and Materials Science, South-Central University for Nationalities, Wuhan, China Fengying Zhang Key Laboratory of Carbon Materials, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan, China Xiaohua Zhang Division of Advanced Nano-Materials, Suzhou Institute of Nano-Tech and NanoBionics, Chinese Academy of Sciences, Suzhou; Innovation Center for Textile Science and Technology, Donghua University, Shanghai, China

CHAPTER 1

Introduction Menghe Miao

CSIRO Manufacturing, Geelong,VIC, Australia

1.1  A brief introduction to carbon nanotubes Nanotechnology and nanomaterials have now become common words in our daily life. Carbon nanotube (CNT) is at the center stage of this new nanomaterial world. CNTs exhibit extraordinary mechanical strength and unique electrical properties, and are efficient conductors of heat. These novel properties make CNTs potentially useful in a wide range of applications in nanotechnology, electronics, optics, energy storage, and other fields of materials science. CNTs are an allotrope of carbon and members of the fullerene structural family, which also includes buckyballs. The name nanotube is derived from its long and hollow shape, since the diameter of a nanotube is on the order of a few to tens of nanometers (the width of a human hair is typically 80 μm, or 80,000 nm) and can be up to several hundred millimeters in length.The wall of a CNT is formed by a one-atom-thick sheet of carbon, called graphene. The sheet is rolled at a specific and discrete angle (chirality). The combination of the rolling angle and the radius is critical to the nanotube properties. Fig. 1.1A shows an infinite graphene sheet. In order to form a seamless tube, certain geometrical conditions must be met. Nanotubes are named by their chirality (n, m) according to the chiral vector C h = na1 + ma 2 where a1 and a2 are unit vectors of the graphene. As the length of the CC bond in the graphene is 0.142 nm, the length of the unit vectors will be 0.246 nm.The structure of a single-walled carbon nanotube (Fig. 1.1B and C) is completely determined by its chirality. For armchair tubes, n=m; for ­zigzag tubes, m=0. For a given (n, m) nanotube, if n=m, the nanotube is metallic; if n−m is a multiple of 3 and n≠m and nm≠0, then the nanotube is ­quasi-metallic with a very small bandgap, otherwise the nanotube is a moderate semiconductor [1]. Carbon Nanotube Fibers and Yarns https://doi.org/10.1016/B978-0-08-102722-6.00001-8

Copyright © 2020 Elsevier Ltd. All rights reserved.

1

2

Carbon Nanotube Fibers and Yarns

Fig.  1.1  Carbon nanotube structures. (A) A graphene sheet is “rolled up” to make a nanotube. T denotes the tube axis, and a1 and a2 are the unit vectors of graphene in real space. (B) Armchair (n, n). (C) Zigzag (n, 0). (D) Triple-walled armchair carbon nanotube. (Courtesy of Wikipedia, https://en.wikipedia.org/wiki/Carbon_nanotube.)

CNTs are categorized as single-walled nanotubes (SWNTs), d­ ouble-walled nanotubes (DWNT), and multi-walled nanotubes (MWNTs). The MWNTs consist of multiple rolled layers of graphene (concentric tubes in Fig. 1.1D). The interlayer distance in MWNTs is close to the distance between graphene layers in graphite, approximately 3.4 Å (0.34 nm). CNTs are the strongest and stiffest materials yet discovered in terms of tensile strength and elastic modulus, respectively, owing to their covalent sp2 bonds between the individual carbon atoms. Mass-produced CNTs contain defects and possess considerably lower strength than that predicted from perfect graphene sheets, but still much higher than any existing commercial material. The challenge is to organize these nanotubes into macroscale structures without introducing further structural defects so that they can express similar properties as their constituent nanotubes.



Introduction

3

Without considering their detailed atomic structures, CNTs are nanoscale fibers that resemble nanofibrils in plant and animal fibers, such as cotton and wool. The CNTs are very long relative to their diameters, with aspect ratios one order of magnitude greater than common natural textile fibers. It is therefore a logical approach to align the CNTs in the form of a fiber or yarn that is expected to outperform conventional textile fibers. This book deals with the various aspects of such fibers or yarns produced from CNTs.

1.2  CNT yarns versus conventional textile yarns The use of the terms “CNT fiber” and “CNT yarn” has now become rather arbitrary. The term “CNT yarn” can be related to its manufacturing from vertically aligned CNT arrays, which bears close similarity to the production method used in traditional textile yarn spinning [2, 3] (Chapter 2). In comparison, the direct-spinning method [4] (Chapter 3) and the solution-spinning method [5, 6] (Chapter 4) more closely resemble reaction spinning and wet spinning of synthetic fibers, respectively. However, nowadays use of term “fiber” or “yarn” is more of a personal choice than a reference to its method of manufacture. The two terms are used interchangeably in this book. CNT yarns, especially twist-spun CNT yarns, are often compared with textile yarns in the analysis of their structure and tensile properties [3]. The insertion of twist to a textile yarn places individual fibers in approximately coaxial helix configuration and the fibers are pressed together because of the inward pressure generated by the tension in the helically disposed fibers. In a conventional textile yarn, interconnection between fibers relies on the fiber-fiber friction that arises from the pressure between fibers, which increases with the external tensile load applied to the yarn [7]. At low twist levels, due to low fiber-fiber friction, the yarn failure mechanism is dominated by fiber slippage. At high twist, fiber slippage is largely prevented by high fiber–fiber friction and thus the yarn fails due to fiber breakage. On the other hand, high twist reduces the contribution of fiber strength to the yarn strength due to fiber obliquity in the yarn. Therefore the maximum yarn specific strength is usually achieved at an intermediate level of twist, as illustrated in Fig. 1.2A. A twisted CNT yarn has a similar twist-strength relationship as the conventional textile yarns [12]. There is, however, a major difference in the mechanism of fiber-fiber (CNT-CNT) interaction. Because the nanotubes

4 Carbon Nanotube Fibers and Yarns

Fig. 1.2  (A) Twist-strength relationship for conventional textile yarns spun from staple fibers; (B) twist-spun CNT yarn from a CNT forest; (C) false twisted CNT yarn spun from a CNT forest [7]; (D) solvent-densified CNT fiber produced by the floating catalyst method [8]; (E) solution-spun CNT fiber [9]; (F) rub-densified CNT yarn produced from a CNT forest [10]; and (G) roll-pressed CNT fiber produced by the floating catalyst method [11]. (Source of (A–C): M. Miao, The role of twist in dry spun carbon nanotube yarns, Carbon 96 (2016) 819–826. Source of (D): J. Qiu, J. Terrones, J.J. Vilatela, M.E. Vickers, J.A. Elliott, A.H. Windle, Liquid infiltration into carbon nanotube fibers: effect on structure and electrical properties, ACS Nano 7 (10) (2013) 8412–8422. Source of (E): D.E. Tsentalovich, R.J. Headrick, F. Mirri, J. Hao, N. Behabtu, C.C. Young, et al., Influence of carbon nanotube characteristics on macroscopic fiber properties, ACS Appl. Mater. Interfaces 9 (41) (2017) 36189–36198. Source of (F): M. Miao, Production, structure and properties of twistless carbon nanotube yarns with a high density sheath, Carbon 50 (13) (2012) 4973–4983. Source of (G): J. Wang, X. Luo, T. Wu, Y. Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014) 3848.)



Introduction

5

have a nanoscale dimension, the van der Waals attraction (London dispersion force) between them plays an important role in the transfer of load between the nanotubes in CNT yarns. Unlike textile yarns, strong CNT yarns can be produced without the insertion of twist. For example, twistless CNT yarns can be produced by liquid shrinking without twist insertion [13, 14]. CNT yarns densified by polar solvent ethylene glycol [13] demonstrated a strength of 1.45 GPa, which is stronger than most twisted CNT dry spun yarns reported in the literature [15]. Mechanical rubbing [10] can also produce a strong CNT yarn without the insertion of twist. Van der Waals forces are inversely proportional to the second power of the distance between the surfaces of the particles so they become dominant for collections of very small nanotubes where there are no capillary forces present. Thus increasing the packing density of CNT is a very important strategy for improving the strength of CNT yarns and fibers. Nanotubes need to be very well aligned in order to pack them really close together. Twisting is a reversible process, that is, the twist in a yarn can be removed by inserting a twist of the same amount but in the opposite direction.When the twist in a staple fiber textile yarn is removed by untwisting, the yarn returns to a loose fiber strand with almost zero strength. On the other hand, when the twist of a twist-spun CNT yarn is removed by untwisting, the resulting yarn largely maintains its structural integrity and a major part of its original strength [7].

1.3  Prospects of high-strength CNT yarns The strength and modulus of monolayer graphenes (CNT shells) are generally considered to be about 130 GPa and 1.0 TPa, respectively [16], which can be taken as the theoretical values for CNTs. These extraordinary values drove the enthusiasm for extreme applications in the earlier days of CNT research, such as space elevator cables [17]. Indeed, bundles of carefully manufactured ultralong defect-free CNTs have been recently reported to possess a tensile strength over 80 GPa [18]. In reality, all mass-produced materials always have defects and thus do not achieve their theoretical strength. Although a CNT fiber with specific strength as high as 9 N/tex at 1-mm gauge length was reported [19], most commonly reported specific strength for CNT fibers is around 1 N/tex, which is already above the specific strength of common synthetic fibers, such as nylon and polyester.To put this in perspective, we plotted the specific

6

Carbon Nanotube Fibers and Yarns

strength of commercial textile fibers against the theoretical specific strength of their corresponding polymers in Fig. 1.3 (details are provided in Chapter 7). The reason for using specific strength in the plot is to take into account of any voids in fibers, which is a key feature for CNT fibers. Commodity textile fibers such as cotton, polyester, and nylon only achieve between 2.5% and 5% of their theoretical specific strength. High-performance fibers, such as Kevlar, spectra, and carbon fibers, which are produced with extremely high care and thus have fewer defects, achieve between 5% and 10% of their theoretical specific strength. It is rare for commercial fibers to achieve more than 10% of their theoretical specific strength. The theoretical specific strength (tenacity) of CNTs is 57.4 N/tex based on the strength (130 GPa) and density (2.266 g/cm3) of monolayer graphenes. The specific strength of future commercial CNT fibers will depend on the degree of structural perfection that can be achieved, as represented by the vertical thick bar on the right-hand side of the plot. For example, at 2.5% of the theoretical strength, similar to commodity synthetic fibers, the CNT fiber will achieve a specific strength of 1.44 N/tex, which translate into 1.87 GPa for a fiber density of 1.3 g/cm3; at 10% of

Fig. 1.3  Specific strength of commercial fibers plotted against the theoretical specific strength of their corresponding polymers.



Introduction

7

the theoretical strength, as achieved by high-performance synthetic fibers (Dyneema, Spectra, and Kevlar), the CNT fiber would achieve a specific strength as high as 5.7 N/tex or 10.3 GPa for a fiber density of 1.8 g/cm3.

1.4  Potential applications Although CNT fibers and yarns produced in laboratories are still not as strong as ultra-strong textile fibers, their multifunctional properties inspired a generation of researchers devoted to the development of a wide range of applications based on CNT fibers and yarns. The combination of strength, flexibility, electrical conductivity, electrochemical reactivity, and porosity makes CNT yarns excellent candidates for miniature sensors, actuators, and energy storage devices required in smart textiles for health care, sports, military, entertainment, and other aspects of modern digital life.

1.5  Organization of the book The science and manufacturing technology around CNT fibers and yarns are still evolving. This book is aimed at providing a snapshot of these developments to people in academia and research of CNT materials, as well as product designers and processing engineers interested in the science and technology for the production, further processing, and applications of emerging high-performance textile materials. Part 1 of the book deals with the production of CNT yarns and fibers, including “pure” CNT fibers and CNT-reinforced nanocomposite fibers. Chapter 2 discusses the probably most widely known two-step manufacturing method of CNT yarn. The first step is growing nanotubes, typically multi-walled carbon nanotubes (MWNTs) on a substrate, known variously as vertically aligned CNT arrays or CNT forests. In the second step, the CNTs in the forest are drawn out in the form of a continuous web, which is simultaneously densified into a yarn by twist insertion, liquid densification, mechanical rubbing, or other methods. CNT fibers can also be manufactured from gaseous feedstock directly in one step, a process bearing similarities to the production of silk fibers by spiders and silkworms, and to the reaction spinning of synthetic fibers. This process is often referred to as the “direct spinning” method because a fiber is pulled out from the high-temperature furnace directly, or referred to as the floating catalyst method in contrast with the deposition of catalyst on a substrate in the two-step method discussed in Chapter 2.

8

Carbon Nanotube Fibers and Yarns

The production of CNT fibers continuously from the furnace provides an effective method for production scaling up. The direct spinning method will be reviewed in Chapter 3, including the synthesis of the nanotubes, assembly of a continuous CNT network, and formation of a final fiber. Chapter 4 provides an overview of the wet spinning of neat or nearly neat CNT fibers from bulk-grown CNTs. Premade CNTs are dissolved into a solvent (usually a strong acid) or in a suspension with the aid of surfactant, which is then formed into a fiber using wet-spinning methods that are similar to the high-throughput extrusion of textile fibers from polymers. Because the synthesis of the CNTs is separated from the formation of fibers, the wet-spinning method provides the opportunity to optimize both processes independently. Instead of dissolving in a solvent, bulk-produced CNTs, usually a small percentage, can be dispersed in a polymer and then extruded using traditional textile fiber spinning methods. This approach is discussed in Chapter 5. Because of their superior properties and one-dimensional (1D) cylindrical geometry, CNTs are ideal fillers for reinforcing polymeric fibers. The reinforcement effect is beyond the rules-of-mixture effect because of the development of an interphase between the CNTs and the polymer. In this chapter, the structure development and property enhancement of such interphase are discussed in detail. Many treatments have been proposed to improve the mechanical, electrical, and thermal properties of neat and composite CNT fibers, including further densification treatments based on twist insertion, lateral compression, rubbing, liquid evaporation, purification, cross-linking treatments by irradiation and polymer infiltration, and combinations of two or more of these treatments. Chapter 6 reviews the principles and procedures of these post-spinning treatments and their effects on CNT fiber properties. Despite tremendous progresses in the last two decades, the properties of the CNT fibers and yarns produced around the world are far behind that of their constituent nanotubes. The challenge has been to organize CNTs into yarns with the best possible properties. Part 2 discusses the structures, properties, and methodology for improving the structure and properties of CNT fibers and yarns based on experiments and computational mechanics. Unlike conventional textile yarns, the strength of final CNT fibers and yarns can be rarely related back to the strength of their constituent nanotubes, mainly due to the complex nature of direct testing of individual nanotubes. Geometry of CNT yarn structure, such as nanotube alignment and packing density, is mainly investigated by adjusting the conditions of



Introduction

9

yarn manufacture and post-spinning treatments. Chapter  7 discusses the structures of CNT fibers and yarns manufactured by different methods and how their structures affect the mechanical, electrical, and thermal properties of the final fibers and yarns. Chapter 8 reviews mechanics models of CNT yarns. General analytic models only predict qualitatively the stress distribution inside the yarn and the trends of twist-dependent yarn performances, like in conventional textile yarn mechanics. Inter-tube sliding determines the precise mechanics of a dry CNT bundle in the yarn, which can be simulated using molecular dynamics.To deal with the large number of nanotubes in a yarn, coarse-grained molecular dynamics is employed to study the microstructural evolution of the CNT structure. Multi-scale modeling is becoming an increasingly important tool to deal with the hierarchical structure of CNT yarns. CNTs have superior mechanical, electrical, and thermal properties but their nanoscale dimensions restrict their applications. CNT yarns, being microscopic and continuous assemblies of CNTs, offer high potential for the development of applications. These multifunctional properties distinguish CNT yarns from textile fibers and metal wires, opening up the possibility of manufacturing a wide range of smart textile constructions. Part 3 reviews some of these applications, including sensing, energy storage, and artificial muscles. CNT yarns are piezo-resistive, which can be utilized for strain measurement, material damage detection, torque measurement and motion monitoring, as well as temperature measurement and detection of various chemicals. Chapter 9 presents the operating principles of CNT yarn sensors and experimental results. Flexible threadlike supercapacitors with high flexibility, tiny volume, and good specific performance have attracted extensive attention recently due to their potential in wearable electronics and smart textiles. CNT yarns have the advantages of high surface area, low mass density, outstanding chemical stability, and excellent electrical conductivity and thus are excellent electrode materials for threadlike supercapacitors. Chapter  10 discusses recent progresses in charge storage mechanisms, active materials, electrolytes, designs of threadlike architecture, and selfcharging supercapacitors. CNT yarns are also promising candidates for flexible actuators, also known as artificial muscles. Chapter 11 presents a brief review on the types of CNT yarn-based actuators developed in recent years and their energy conversion mechanisms, performance metrics, and potential applications.

10

Carbon Nanotube Fibers and Yarns

References [1] E.A. Laird, F. Kuemmeth, G.A. Steele, K. Grove-Rasmussen, J. Nygård, K. Flensberg, et al., Quantum transport in carbon nanotubes, Rev. Mod. Phys. 87 (3) (2015) 703. [2] K. Jiang, Q. Li, S. Fan, Spinning continuous carbon nanotube yarns, Nature 419 (2002) 801. [3] M. Zhang, K. Atkinson, R.H. Baughman, Multifunctional carbon nanotube yarns by downsizing an ancient technology, Science 306 (5700) (2004) 1358–1361. [4] Y. Li, I. Kinloch, A. Windle, Direct spinning of carbon nanotube fibers from chemical vapor deposition synthesis, Science 304 (5668) (2004) 276–278. [5] B. Vigolo, A. Penicaud, C. Coulon, C. Sauder, R. Pailler, C. Journet, et al., Macroscopic fibers and ribbons of oriented carbon nanotubes, Science 290 (2000) 1331–1334. [6] N. Behabtu, C.C. Young, D.E. Tsentalovich, O. Kleinerman, X. Wang, A.W. Ma, et al., Strong, light, multifunctional fibers of carbon nanotubes with ultrahigh conductivity, Science 339 (6116) (2013) 182–186. [7] M.  Miao, The role of twist in dry spun carbon nanotube yarns, Carbon 96 (2016) 819–826. [8] J. Qiu, J. Terrones, J.J. Vilatela, M.E. Vickers, J.A. Elliott, A.H. Windle, Liquid infiltration into carbon nanotube fibers: effect on structure and electrical properties, ACS Nano 7 (10) (2013) 8412–8422. [9] D.E. Tsentalovich, R.J. Headrick, F. Mirri, J. Hao, N. Behabtu, C.C. Young, et al., Influence of carbon nanotube characteristics on macroscopic fiber properties, ACS Appl. Mater. Interfaces 9 (41) (2017) 36189–36198. [10] M. Miao, Production, structure and properties of twistless carbon nanotube yarns with a high density sheath, Carbon 50 (13) (2012) 4973–4983. [11] J.  Wang, X.  Luo, T.  Wu, Y.  Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014) 3848. [12] M.  Miao, J.  McDonnell, L.  Vuckovic, S.C.  Hawkins, Poisson’s ratio and porosity of carbon nanotube dry-spun yarns, Carbon 48 (10) (2010) 2802–2811. [13] S. Li, X. Zhang, J. Zhao, F. Meng, G. Xu, Z. Yong, et al., Enhancement of carbon nanotube fibres using different solvents and polymers, Compos. Sci. Technol. 72 (12) (2012) 1402–1407. [14] X. Zhang, K. Jiang, C. Feng, P. Liu, L. Zhang, J. Kong, et al., Spinning and processing continuous yarns from 4-Inch wafer scale super-aligned carbon nanotube arrays, Adv. Mater. 18 (12) (2006) 1505–1510. [15] M. Miao,Yarn spun from carbon nanotube forests: production, structure, properties and applications, Particuology 11 (4) (2013) 378–393. [16] C. Lee, X. Wei, J.W. Kysar, J. Hone, Measurement of the elastic properties and intrinsic strength of monolayer graphene, Science 321 (5887) (2008) 385–388. [17] B.C.  Edwards, Design and development of a space elevator, Acta Astronaut. 47 (10) (2000) 735–744. [18] Y. Bai, R. Zhang, X. Ye, Z. Zhu, H. Xie, B. Shen, et al., Carbon nanotube bundles with tensile strength over 80 GPa, Nat. Nanotechnol. (2018) 1. [19] K.  Koziol, J.  Vilatela, A.  Moisala, M.  Motta, P.  Cunniff, M.  Sennett, et  al., High-­ performance carbon nanotube fiber, Science 318 (5858) (2007) 1892–1895.

PART I

Manufacturing methods

This page intentionally left blank

CHAPTER 2

Yarn production from carbon nanotube forests Menghe Miao

CSIRO Manufacturing, Geelong,VIC, Australia

2.1  Synthesis of vertically aligned CNT arrays The synthesis of aligned carbon nanotube (CNT) arrays on a substrate of mesoporous silica containing iron oxide nanoparticles using the chemical vapor deposition (CVD) method was first reported in the mid-1990s. Li et al. [1] reported that a mixture of 9% acetylene in nitrogen was introduced into the synthesis chamber at a flow rate of 110 cm3/min.The CNTs were formed on the substrate by the deposition of carbon atoms obtained by the decomposition of acetylene at 700°C. Ren et  al. [2] reported the growth of CNT arrays on nickel-coated glass at temperatures below 666°C by plasma-enhanced hot filament CVD. Acetylene gas was used as the carbon source and ammonia gas was used as a catalyst and dilution gas. Fan et al. [3] described the production of self-orientated CNT arrays on porous and plain silicon substrates. Silicon substrates were deposited with Fe films (5 nm thick) by electron beam evaporation and then annealed in air at 300°C overnight to oxidize the surface of the silicon and the iron. The substrate was placed in a cylindrical quartz boat sealed at one end and then inserted into the center of a quartz tube reactor housed in a tube furnace. The furnace was heated to 700°C in flowing Argon. Ethylene was then flown at 1000 sccm for 15–60 min, after which the furnace was cooled to room temperature. During the initial stage of CVD, ethylene molecules are catalytically decomposed on the iron oxide nanoparticles. As supersaturation occurs, a nanotube grows off each of the densely packed catalyst particles (average diameter 16 nm) and extends to open space along the direction normal to the substrate. As the nanotubes lengthen, their outermost walls interact with those of neighboring nanotubes via van der Waals forces to form a large bundle with sufficient rigidity. This rigidity enables nanotubes to keep growing along the original direction. Many research groups around the world have since investigated methods to fabricate CNT forests Carbon Nanotube Fibers and Yarns https://doi.org/10.1016/B978-0-08-102722-6.00002-X

Copyright © 2020 Elsevier Ltd. All rights reserved.

13

14

Carbon Nanotube Fibers and Yarns

with tuneable tube diameter, number of walls, length, alignment, covering areas, growth rate, etc. Growth enhancers (oxygen-containing molecules, e.g., water, alcohols, ethers, esters, ketones, aldehydes, and carbon dioxide) have been used to dramatically increase the growth efficiency, tube length, and alignment [4, 5]. Bedewy et al. [6] proposed a collective growth mechanism of vertically aligned CNT forests. The CNT forest growth starts with the formation of a thin “crust” of randomly oriented CNTs that resides on the top of the forest after alignment emerges. They postulated that the abrupt termination of CNT forest growth was caused by loss of the self-supporting structure. The conversion of vertically aligned CNT arrays or forests into a continuous length of interconnected CNT web was discovered accidentally by Jiang et al. [7]. While attempting to pull out a bundle of CNTs from an array several hundred micrometers high grown on a silicon substrate, they obtained instead a continuous ribbon-like web of pure CNTs, in a way similar to the drawing of a thread from a silk cocoon.The drawing of a web from a CNT forest is captured in Fig. 2.1. The drawing of a CNT forest into a continuous web is the key to the formation of a continuous CNT yarn. All known spinnable CNT forests are produced using the CVD method. Most spinnable CNT forests are grown on flat substrates using thermal CVD method. Typically, aligned CNTs are grown in a reaction furnace with flowing gaseous carbon feedstock in the presence of catalyst on a silicon wafer substrate. A layer of Fe catalyst nanoparticles is deposited on the silicon wafer by electron beam evaporation or magnetron sputtering. An Al2O3 buffer layer can also be introduced

Fig. 2.1  (A) SEM image showing that the MWNTs in a forest are rotated 90 degree to form a continuous ribbon web, and (B) TEM image of nanotube bundles in drawable forest [8]. (Panel (B) reprinted with permission from X. Zhang, K. Jiang, C. Feng, P. Liu, L. Zhang, J. Kong, et al., Spinning and processing continuous yarns from 4-inch wafer scale super-aligned carbon nanotube arrays, Adv. Mater. 18 (12) (2006) 1505–1510.)



Yarn production from carbon nanotube forests

15

before the catalyst deposition. A carrier gas, for example, helium, and a carbon source, such as acetylene [7, 9–12] or ethylene [11, 13, 14], are used. The diameter distribution and number of walls of the CNTs in a forest can be controlled to some extent by varying the thicknesses of the catalyst layer and the length of nanotubes by growth time [15]. Water vapor can be used to prolong CNT growth resulting in longer CNTs [11]. Cui et al. [16] studied the effects of reaction conditions on the growth rate of CNT forests (CNT length) in CVD using Fe catalyst on Si wafer, C2H2 as carbon source, and H2 as inhibitor for carbonaceous species. They found that the optimum reaction conditions were achieved at a temperature of 750°C, acetylene flow rate of 60 sccm, and H2 flow rate of 0.5 SLPM. The CNT growth rate became very low after 10 min. Huyhn et  al. [18] carried out a recycling analysis of Si substrate for spinnable CNT growth. A 100% regrowth in forest height and mass yield of CNTs were achieved in the first four cycles, but these parameters fell to about 20% in the fifth cycle. A decrease in nanotube diameter and increase in areal density were also observed. A floating catalyst CVD method was used for the production of spinnable CNT forests using a two-stage synthesis process. The CNT arrays were grown on silica (quartz flake) using ferrocene as catalyst precursor and cyclohexane as solvent and carbon source [19, 20]. The main advantages of the floating catalyst method are that it requires simple equipment and eliminates the procedure for the preparation of catalyst layers on substrates.

2.2  Drawing a CNT web 2.2.1  Formation of a continuous web from CNT forest Several groups studied forest drawability using experimental and modeling techniques and proposed a number of working mechanisms. Zhang et al. [21] ascribed drawability to the intermittent bundling of CNTs within the forest in which individual nanotubes migrate from one bundle of a few nanotubes to another. Bundled nanotubes are simultaneously pulled from different elevations in the forest sidewall, so that they join with bundled nanotubes that have reached the top and bottom of the forest. Disordered regions at the top and bottom of the forests, where a fraction of the nanotubes form loops, might help maintain continuity. They also reported that for forests having similar topology, the highest forests were easiest to draw into sheets, probably because increasing the nanotube length increases interfibril mechanical coupling within the web.

16

Carbon Nanotube Fibers and Yarns

Kuznetsov et al. [22] proposed a structural model for the drawing of a continuous web from a vertically oriented MWNT (multi-walled nanotube) forest (Fig.  2.2). The original forest consists of vertically oriented forest trees (big CNT bundles) interconnected by smaller bundles or individual nanotubes (called connects). The CNT web formation involves two principal processes: (1) unzipping by preferentially peeling off interconnections between the bundles in the forest and (2) self-strengthening of these interconnections by densification at the top and bottom of the forest during draw-induced reorientation of the bundles. Zhang et  al. [8] considered that CNTs from highly aligned arrays of clean surface CNTs form tight bundles due to strong van der Waals interactions between the tubes. When pulling the CNTs from a super-aligned array, van der Waals force makes the CNTs join end to end, thus, forming a continuous yarn. These end-to-end joints are enhanced by bridging CNT bundles to form a continuous ribbon web. Fallah Gilvaei et al. [12] measured the force required to separate CNTs from parent drawable and undrawable forests. The drawable forest showed high separation force at the top of the forest and low separation force in the middle of the forest height, while in the undrawable forests the separation force showed no marked difference at different locations. They 1

1

2

2

3

3

4

1 2

1 2

4

3

3 4

4 5

5

5

5

2 5

1 3 4

4 3

2

1

2

2 1

3 4

5

4 3

1

Fig. 2.2  Scheme of the process of pulling out a forest tree from the forest. Blue and red bars represent forest trees or bundles of CNTs, and the separation distance between them is increased for clarity. Red arrows show the external force applied to the first bundle. Small black arrows show the direction of the net movement of the interconnections [22]. (Reprinted with permission from A.A. Kuznetsov, A.F. Fonseca, R.H. Baughman, A.A. Zakhidov, Structural model for dry-drawing of sheets and yarns from carbon nanotube forests, ACS Nano 5 (2) (2011) 985–993.)



Yarn production from carbon nanotube forests

17

identified three factors involved in the bundle drawing process: end knots, entanglement in the middle of the CNT forest, and the interaction force between adjacent CNTs in the middle of the forest. High-density CNTs in the forest encourages the formation of entanglement at the top of the forest. Although the end knots are responsible for preserving the fiber continuity at the ends of the forest (top and bottom), the entanglement and the inter-tube interaction force in the middle of the forest cause the formation of new bundles. Entanglement in the middle of the forest plays the same role, but to a lesser extent. A high degree of drawability occurs when all these factors act simultaneously. While all the above studies placed emphases on the CNT entanglement at the top of the forest, Zhu et al. [17] demonstrated that drawable CNT forests remain to be drawable after removing the entangled layer at the top of the forest by reactive ion etching. They observed the formation of entangled structures when the pulling process approaches the bottom and top ends of the CNT arrays, as depicted in Fig. 2.3. These entangled structures were considered to be responsible for maintaining the continuity of the drawing process. CNT entanglements which are crucial to drawability require nanotube crossovers (loose entanglement) originated from the morphology of CNT forest. Manchard et  al. [23] observed that nanotubes rotated during synthesis. A group of rotating CNTs could therefore form a loosely entangled bundle during synthesis in a similar way as the formation of self-twist yarn from two or more twisted strands of conventional textile fibers [24]. When the loosely entangled bundle is pulled apart into two or more parts during web drawing, the twists or entanglements in the bundle are pushed to one end, forming a tight knot [17].

2.2.2  Improving drawability It is generally agreed that high level of CNT alignment in the forest is a prerequisite for drawability [8–10, 12, 20, 21, 25, 26]. Fig. 2.4A–C show SEM (scanning electron microscope) images of undrawable, marginally drawable, and fully drawable CNT forests. The SEM images of CNT forests and webs are often used to demonstrate CNT alignment. Only images taken at the same magnification should be compared for nanotube alignment. Fig. 2.4D–F [28] are SEM images of the same CNT forest taken at different magnifications, but the smaller magnification image (Fig.  2.4D) gives an impression of very high degree of CNT alignment while the larger magnification image (Fig.  2.5F) gives an impression of low CNT ­alignment.

18

Carbon Nanotube Fibers and Yarns

Fig.  2.3  Self-entanglement mechanism for continuous pulling of carbon nanotube bundles. The CNT array consists of crossing over and branched CNT bundles. (A) The bundle is first pulled out and detached from other bundles. (B) Cleaving of the front branched bundles and shifting of the crossing over structure. (C) Self-entanglement happens. (D) The twisting bundles entwine nearby branched bundles together to form an entangled structure. (E) The entangled structures pull out more bundles and the connection between the extracted bundles becomes firmer upon further pulling. (F) A similar entangled structure is formed on the top surface region [17]. (Reprinted with permission from C. Zhu, C. Cheng, Y.H. He, L. Wang, T.L. Wong, K.K. Fung, et  al., A self-entanglement mechanism for continuous pulling of carbon nanotube yarns, Carbon 49 (2011) 4996–5001.)



Yarn production from carbon nanotube forests

19

Fig. 2.4  SEM images of CNT forests with different levels of drawability: (A) forest of poor spinnability showing tortuous CNTs, (B) forest with intermediate drawability, and (C) highly spinnable forest showing high level CNT alignment [27]. (D)–(F) SEM images of a CNT forest at different magnifications [28]. (Panels (A–C) reprinted with permission from L. Zheng, G. Sun, Z. Zhan, Tuning array morphology for high-strength carbon-nanotube fibers, Small 6 (2010),132–137. Panels (D–F) reprinted with permission from W. Cho, M. Schulz, V. Shanov, Growth and characterization of vertically aligned centimeter long CNT arrays, Carbon 72 (2014) 264–273.)

CNT alignment can be characterized either in terms of directionality using a power spectrum [20] or small-angle X-ray scattering (SAXS) [30] or a tortuosity factor calculated from the geometry of individual tubes [19, 20]. Narrow size distribution of catalysts, high nucleation density, and very clean surfaces of CNTs are required for good drawability [31].The drawability of CNT forests is closely related to the morphology of the CNT arrays and can be altered, for example, by adjusting catalyst pretreatment time [32] or by introducing a small amount of hydrogen during CVD growth [27]. The shortest catalyst pretreatment time led to CNT arrays with the best drawability, while prolonged pretreatment resulted in coarsening of catalyst particles and non-drawable CNT forests [32]. Well-aligned CNT arrays were obtained from the hydrogen-assisted growth, and wave-like arrays could be obtained from the oxygen-assisted growth [27]. Huynh and Hawkins [10] and Kim et al. [25, 33] investigated the influences of catalyst deposition and synthesis conditions on CNT forest drawability. For a 44-mm internal diameter reactor, the optimum results were obtained by

20

Carbon Nanotube Fibers and Yarns

Fig. 2.5  Formation of twisted CNT yarn: (A) SEM image of CNT yarn formation zone, (B) evolution from a flat ribbon to a circular yarn by twist insertion (C and D) refer to positions in the left SEM image, and (C) SEM image of final twisted CNT yarn [29]. (Panel (C) reprinted with permission from M. Miao, Yarn spun from carbon nanotube forests: production, structure, properties and applications, Particuology 11 (4) (2013) 378–393.)

c­ ompromising a 2.3-nm thick Fe catalyst layer on a silicon substrate with 50 nm of thermal oxide, 670°C running temperature, 650 sccm helium, and 34 sccm acetylene for 20 min [10].

2.3  Yarn formation 2.3.1  Densification by twist The “yarn” drawn from a CNT forest by Jiang et al. [7] had a low mechanical strength. Subsequent research by Zhang et  al. [9] showed that twist insertion could convert the CNT web into a high-density yarn with dramatically increased mechanical strength. Fig. 2.5A is a SEM image of the



Yarn production from carbon nanotube forests

21

yarn forming zone. As a continuous flat ribbon of CNT web is pulled from the forest, it gradually converges into a triangle shape (starting from Position A in Fig. 2.5A). The ribbon starts to curve out of plane at approximately the middle line (position B) due to the torque transmitted by the already formed yarn (position D) from the twisting element. As the ribbon moves forward further, the middle part of the ribbon rolls further into a nuclei (Position C), on which neighboring nanotubes from both sides gather and wrap following the direction of the twisting action. Finally, the rest of the nanotubes are wrapped onto the growing nuclei to form a round yarn at the apex of the triangle zone (Position D). A SEM image of the final twisted yarn is shown in Fig. 2.5C. The spinning method initially used by Zhang et  al. [9] was a motor-assisted hand-spinning process. Twist was applied by the rotation of a motor while it is being moved away from the nanotube forest by sliding it along a table top. The achieved yarn length was limited to about 1 m, or the arm length of the person holding the motor. Other methods for spinning similar yarn lengths have been reported. Zhu’s group built a CNT yarn spinning machine from what appeared to be a modified lathe [34]. In this case, the yarn length is limited to the travel distance of the tool rest. University of Cincinnati [14] and Chinese Academy of Sciences [35] built continuous CNT yarn spinning machines with yarn collection bobbin perpendicular to the spindle (Fig. 2.6). The twisting and winding assembly consisted of a small motor, a number of pulleys, and a yarn winding bobbin. The bobbin shaft was carried on the spindle that inserts twist to the yarn. The yarn throughput and twist insertion speeds were regulated by two separate motors. The unbalanced design of the twisting-winding assembly results in a side to side movement of the forming CNT yarn, which destabilizes the spinning condition and is a source of yarn irregularity [14]. Because of the cumbersome design, the machines could run only at relatively low spindle speeds. 2.3.1.1  Flyer spinning Research efforts at the Commonwealth Scientific and Industrial Research Organization (CSIRO) resulted in two high-speed CNT yarn spinning machine designs.The first automated continuous CNT yarn spinning machine was built based on the conventional flyer spinning principle (Fig. 2.7). The twisting and winding operations are realized by two coaxial shafts that rotate at differential speeds, causing the yarn to be wound onto the yarn collection bobbin carried on the spindle. The spindle also performs a linear motion to

22

Carbon Nanotube Fibers and Yarns

Motor

CNT forest CNT fiber CNT shcct Substrate

(A)

Draw

(C)

Motor Twist

(B)

Fig.  2.6  A continuous CNT yarn spinning machine: (A) schematic of spinning method [35], (B) overview of spinning machine [14], and (C) construction of the twisting-winding assembly [35]. (Panels (A and C) reprinted with permission from J. Zhao, X. Zhang, Y. Huang, J. Zou, T. Liu, N. Liang, et al., A comparison of the twisted and untwisted structures for one-dimensional carbon nanotube assemblies, Mater. Des. 146 (2018) 20–27. Panels (B) reprinted with permission from C. Jayasinghe, S. Chakrabarti, M.J. Schulz, V. Shanov, Spinning yarn from long carbon nanotube arrays, J. Mater. Res. 26 (5) (2011) 645–651.)

a

Winding yarn guide Winding disk

Spindle pin

Initial yarn guide

Spindle base

Magnetic Disk

Yarn

Bobbin Spindle

Nanotube forest on substrate

Electromagnet

CNT forest

Adding chemicals

c

Twisting & winding Fig. 2.7  CSIRO flyer spinning machine: (A) schematic of flyer spinner, (B) photograph of a flyer spinner [29], and (C) application of the liquid chemical during spinning process [36]. (Panels (A and B) reprinted with permission from M. Miao, Yarn spun from carbon nanotube forests: production, structure, properties and applications, Particuology 11 (4) (2013) 378– 393. Panel (C) reprinted with permission from J.Y. Cai, J. Min, J. McDonnell, J.S. Church, C.D. Easton, W. Humphries, et al., An improved method for functionalisation of carbon nanotube spun yarns with aryldiazonium compounds, Carbon 50 (12) (2012) 4655–4662.)



Yarn production from carbon nanotube forests

23

spread the yarn on the bobbin in an orderly fashion. A computer is used to coordinate these motions. The machine can be run at twist insertion rate (spindle speed) up to 7000 rpm. A series of friction pins can be introduced between the CNT forest and the spindle on the flyer spinning machine [37]. These pins affect the spinning process in two ways: increasing the yarn tension and causing the twist to be inserted into the yarn in steps along the zones separated by the pins. The increased tension further increases yarn density, leading to a more compact yarn structure and higher stress-based strength and elastic modulus but lower breaking strain. The flyer spinner can also be used to add solvent, polymers, or other additives while a CNT yarn is being spun [36]. The additives can be conveniently applied on the CNT web and forming yarn between the feedstock forest and the yarn guide, as shown in Fig. 2.7C. 2.3.1.2 Up-spinning Fig. 2.8 shows what is dubbed as the CSIRO “up-spinner” [38]. The CNT forest is attached to a spindle that can be rotated at a high speed. The continuous CNT web drawn from the forest is pulled up (hence the name up-spinner) to the yarn bobbin while twist is being inserted by the spindle beneath it. On the up-spinner, the two essential functions of continuous yarn spinning are carried out independently: twist is inserted by the fast rotating vertical spindle that carries the feedstock (the CNT forest) while

Fig. 2.8  Main operating elements of the up-spinner.

24

Carbon Nanotube Fibers and Yarns

yarn collection is carried out by the slow rotating horizontal yarn bobbin. Additionally, a yarn guide or the bobbin itself traverses longitudinally to spread the yarn along the bobbin. The up-spinner is a much simpler machine design than the flyer spinner. The mass carried on the high-speed spindle is immensely smaller than that of the spindle assembly in the flyer spinning method. The yarn path on the up-spinner is essentially a straight line, resulting in much lower yarn tension during spinning so that delicate CNT webs can be spun into yarns at a high speed. For high-speed operation, the mass of the CNT forest and its holder must be counterbalanced by a weight. A liquid applicator is fitted next to the yarn guide for application of solvent, polymer, or other additives during spinning. The machine was run at spindle speeds up to 18,000 rpm during trials. 2.3.1.3 False-twisting Twisting is a reversible process, that is, the twist in a yarn can be removed by inserting a twist of the same amount but in the opposite direction (i.e., untwisting). When the twist in a conventionally spun textile yarn is removed by untwisting, the yarn returns to a loose fiber bundle with virtually zero strength. However, when the twist in a twisted CNT yarn is removed by untwisting, the resultant twistless yarn largely maintains its structural integrity due to the van der Waals force between the nanotubes [39]. Fig. 2.9 shows SEM images of (A) a twisted yarn and (B) an essentially twistless yarn obtained by untwisting the twisted yarn.

Fig.  2.9  Production of false-twisted CNT yarns: (A) Twisted yarn 5000 T/m. (B) Twistuntwisted yarn 5000 T/m [39]. (C) Schematic of false-twist CNT yarn spinning. (Panels (A and B) reprinted with permission from M. Miao, The role of twist in dry spun carbon nanotube yarns, Carbon 96 (2016) 819–826.)



Yarn production from carbon nanotube forests

25

The twisting-untwisting operation can be accomplished in one continuous operation, known as false-twisting. In a false-twist setup, as illustrated in Fig. 2.9C, turns of twist are present only in the region before the yarn reaches the false-twisting device [40]. This temporary twist disappears after the yarn passes the twisting device (thus called false twist). As false twist can be inserted into yarns without rotating a heavy yarn package or the feedstock, the method is used to process yarns at high speeds, as in the production of false-twist textured nylon yarns that are used in fine-gauge elastic fabrics, such as socks, tights, and leggings. 2.3.1.4  Core-spun yarns A core-spun CNT yarn consisting of a fine metal filament core and a CNT sheath was constructed for application in two-ply yarn supercapacitors (see Chapter 10) to take advantage of the high electrical conductivity of the metal filament in the core and the large specific surface area of the nanotubes in the sheath [41]. The core/sheath-structured CNT yarn can be manufactured on a flyer spinner, as shown schematically in Fig. 2.10A.

Fig. 2.10  Core-spun yarns: (A) schematic of core-spun yarn process and (B) cross-sectional view of core-spun CNT yarn [41]. (Reprinted with permission from D. Zhang, M. Miao, H. Niu, Z. Wei, Core-spun carbon nanotube yarn supercapacitors for wearable electronic textiles, ACS Nano, 8 (5) (2014) 4571–4579.)

26

Carbon Nanotube Fibers and Yarns

The CNTs forming the sheath are drawn from the CNT forest in the form of a continuous web. The core material is pulled out from a supply bobbin to merge with the CNT web at the center. The twisting action of the spindle at the right-hand side causes the metal filament and the CNT web to rotate together, resulting in wrapping of the CNT web around the metal filament to form a core/sheath structured yarn (Fig. 2.10B). Because of the very large width of the CNT web in relation to the diameter of the metal filament, the core is completely covered by the CNT sheath in the resulting core-spun yarn.

2.3.2  Rubbing densification Mechanical rubbing action can be used to produce a highly densified CNT yarn (Fig. 2.11A) [42]. The main working parts of the machine are a pair of padded rollers that participate in both rotation and axial oscillation, as shown in Fig. 2.11B.The rotational motion of the two rollers transports the CNT web drawn from the CNT forest to the yarn collection bobbin. The axial oscillations of the two rollers work in opposite directions to apply a rubbing action that densifies the CNT web into a yarn. During rubbing, the CNT yarn is deformed under the pressure between two elastomeric surfaces moving in opposite directions, as shown in Fig.  2.11C. The action of the moving surfaces is to rotate the yarn about its axis, causing the fibers to move around in a “race-track” fashion. CNTs in the race-track are unable to retain their positions relative to others in the yarn cross section, and the jockeying for position results in relative movement of CNTs in the yarn. On the other hand, the strong van der Waals attraction restricts the free relative movements between the individual nanotubes. The jockeying for positions caused by the rubbing action leads to the filling of voids between CNT bundles, forming a closely packed yarn structure. As illustrated in Fig. 2.11C, the outer layer on two sides of the yarn are directly driven by the two rubbing surfaces and thus are forced to move in opposite directions, so large-scale shears must take place in the intervening region, which causes the CNTs in the yarn core to be torn apart repeatedly as the process continues.When the yarn moves past the roller nip, the lateral compression applied to the yarn is released. The flattened yarn cross section opens up due to elastic recovery, leading to the formation of voids in the yarn core. The resulting yarn structure has thus a high-density sheath and a low-density core, as shown in the cross-sectional image in Fig. 2.11D. A high roller pressure causes severe yarn flattening at the roller nip, leading to a ribbon-like yarn structure without a distinctive porous core, as shown in Fig. 2.11E.



Yarn production from carbon nanotube forests

27

Fig. 2.11  CNT yarn production by rub densification: (A) working elements of the machine; (B) alternating twist introduced by reciprocating rubbing rollers; (C) schematic representation of the yarn structure formation mechanism; and (D–F) SEM images of rub-densified CNT yarns [42]. (Reprinted with permission from M. Miao, Production, structure and properties of twistless carbon nanotube yarns with a high density sheath, Carbon 50 (13) (2012) 4973–4983.)

Despite the high instantaneous twist insertion rate, the level of false twist generated in the rubbing roller system is relatively low because of the oscillating nature of the action. The false twist introduced to the yarn in the first half cycle of the reciprocation is in the direction opposite to the twist ­introduced in the next half cycle, and therefore the newly introduced

28

Carbon Nanotube Fibers and Yarns

twist in the same zone cancels the opposite directional twist introduced during the previous half cycle of reciprocation. So the CNTs in a rub densified CNT yarn are substantially straight and parallel to each other and are aligned in the direction of the yarn axis (Fig. 2.11F). Because of the small yarn diameter, twisted CNT yarns require many thousands of twist turns per meter.The large number of twists per meter limits the rate of yarn production and demands for highly engineered spinning machinery. In the rubbing machine, the yarn rotation is frictionally driven by the rubbing rollers so a small yarn diameter means that the yarn can be rolled at a high rotational speed by relatively low-speed rubbing rollers.Therefore a high production rate can be achieved without requiring high-speed machine parts.

2.3.3  Die drawing Die-drawing method was first used to produce a CNT fiber from single-walled carbon nanotubes (SWNTs) in the form of a film that was peeled off from the interior wall of a quartz tube used in floating catalytic CVD [43] (see Chapter 3).A series of diamond wire drawing dies were used.The drawing was carried out through 18 diamond dies with decreasing diameter from 1.2 to 0.2 mm. During drawing, the SWNT film formed a strand with the same diameter as that of the die used. Initially the as-grown nanotube fiber was black. The final SWNT fiber became gray with metallic luster.The obtained SWNT fibers were highly dense and aligned. The die-drawing method was adapted to densify webs drawn from CNT forests [44]. The nanotubes in the yarn are compressed together by the inner wall of the die, resulting in densification.The yarn diameter springs back by about 10% after coming out from the die. In the final yarn, the nanotubes are held together by van der Waals forces that arise from the intimacy between the tubes. Fig. 2.12 illustrates the CNT yarn production process and the design of the die. For the same CNT web (constant weight/unit length of web), a smaller diameter die results in a higher density yarn. The yarn produced using 35 μm die had the highest tensile strength over and above the yarn produced using 30 μm die. Although having a higher density (1.15 g/cm3) than typical twisted yarns (1300°C) the RBM signals drop significantly (Fig. 3.4), which indicates a reduction of the SWNT percentage in the sample. This transition could be related to the larger catalyst particles at a higher temperature.



Carbon nanotube fibers spun directly from furnace

(A)

(B)

(C)

(D)

43

Fig.  3.3  TEM images of CNTs synthesized at different temperatures: (A) 1200°C, (B) 1300°C, (C) 1400°C, and (D) 1500°C. (Reproduced with permission from Hou G, Chauhan D, Ng V, Xu C, Yin Z, Paine M, et al. Gas phase pyrolysis synthesis of carbon nanotubes at high temperature. Mater. Des. 132 (2017) 112–118.)

Fig. 3.4  Raman RBM peaks of CNTs at different synthesis temperature: (A) 514 nm laser, and (B) 785 nm laser. (Reproduced with permission from Hou G, Chauhan D, Ng V, Xu C, Yin Z, Paine M, et al. Gas phase pyrolysis synthesis of carbon nanotubes at high temperature. Mater. Des. 132 (2017) 112–118.)

44

Carbon Nanotube Fibers and Yarns

Fig.  3.5  TGA curves of CNTs from different synthesis temperatures: (A) 1200°C; (B) 1300°C; (C) 1400°C; and (D) 1500°C. (Reproduced with permission from Hou G, Chauhan D, Ng V, Xu C, Yin Z, Paine M, et al. Gas phase pyrolysis synthesis of carbon nanotubes at high temperature. Mater. Des. 132 (2017) 112–118.)

The purity of CNT samples varies at different temperature. The impurity level can be evaluated by thermal gravimetric analysis (TGA) method. Different burning events can be observed from the TGA curves (Fig. 3.5). The amorphous impurities burn out first at a temperature below 400°C, followed by two main oxidation events at ~520°C and ~660°C. These two events possibly correspond to SWNTs and MWNTs oxidation, respectively. The separated burning events for two types of CNTs were also observed by another group [22]. The amount of amorphous impurities increases at a higher temperature, which could be attributed to the non-catalytic decomposition of hydrocarbons at elevated temperature, introducing more amorphous impurities as by-products. From the Raman spectra in Fig. 3.6, there is a noticeable decrease in the D peak at a higher temperature, indicating a decrease in defects and graphitic impurities. In consideration of the more substantial amount of amorphous impurities at high temperatures, there should be a significant



Carbon nanotube fibers spun directly from furnace

45

Intensity (a.u.)

1500°C 1400°C 1300°C 1200°C 1150°C

0

1000

(A)

2000

3000

Raman shift (cm–1)

Intensity (a.u.)

1500°C 1400°C 1300°C 1200°C 1150°C

0

(B)

1000

2000

3000

Raman shift (cm–1)

Fig. 3.6  Raman spectra of samples at different temperatures: (A) 514 nm laser data normalized to the G peak; (B) 785 nm laser data normalized to the G peak. (Reproduced with permission from Hou G, Chauhan D, Ng V, Xu C, Yin Z, Paine M, et al. Gas phase pyrolysis synthesis of carbon nanotubes at high temperature. Mater. Des. 132 (2017) 112–118.)

46

Carbon Nanotube Fibers and Yarns

increase in CNT quality to mask the influence of the amorphous impurities. The IG/ID ratio increases by almost more than 200%, from 1150°C to 1500°C. This shows that CNTs of better quality with fewer defects can be obtained at higher temperatures.

3.3  CNT assembly and fiber production CNTs synthesized by floating catalyst method could be transformed into different macro assemblies. They can form unorganized entities, such as clusters or powders on the reactor wall or a collector [50], or alternatively CNTs from an aerosol that can be deposited on various substrates as thin films [51]. Meanwhile, these gas-carried CNTs can be assembled into organized entities, such as CNT array on a substrate placed inside the reactor [23, 38, 40, 41]. More importantly, an aerogel-like sock structure could be formed at relatively high temperature, which can easily be transformed into a CNT fiber [37, 52].

3.3.1  CNT sock The CNT sock assembly is an aerogel-like multiscale hierarchical structure (Fig. 3.7). Due to their high aspect ratio, the CNTs form bundles by contacting with each other.These bundles of CNTs serve as the building block for the sock. The critical question is how these μm-sized CNT bundles entangle and form a macroscale structure of centimeter dimension. There are a number of forces acting on the CNT bundles which cause them to stay together. It has been suggested that the sock forms due to thermophoresis or inertial migration, and the CNT may accumulate at a certain distance away from the wall [25]. The thermophoretic force is a result of

(B)

(A)

1 inch

(C)

1 µm

40 nm

Fig. 3.7  Hierarchical structure of CNT assembly: (A) cm-scale sock; (B) μm-scale network; and (C) nm-scale bundles. (Reproduced with permission from Hou G, Ng V, Xu C, Zhang L, Zhang G, Shanov V, et al. Multiscale modeling of carbon nanotube bundle agglomeration inside a gas phase pyrolysis reactor. MRS Adv. (2017) 1–6.)



Carbon nanotube fibers spun directly from furnace

47

temperature gradients and responsible for keeping CNTs away from the reactor wall. Small particles under aerodynamic inertial migration could segregate into annular regions, much similar to the shell of the CNT sock [53]. Other effects might also provide the bonding force, such as van der Waals attraction [24, 27, 37]. Gspann et  al. [11] suggested that due to the slow Poiseuille flow, the catalyst particles close to the reactor move slower and could grow larger through collision. It could be possible that the catalyst particle close to the wall has longer residence so CNTs around that regions will grow longer. The velocity gradient will help partial CNT connection and accumulation, preferentially close to the reactor wall. Some useful information could be learned from substrate-grown CNTs. Blakrishnan et al. found that the mechanical coupling between CNTs is critical for the self-organization of forest and it also introduces deformation and defects in CNT walls [54]. We have investigated the sock dynamics by controlling the feedstock type, injection rate, and carrier gas flow rate. A convection vortex has been identified, and a new convection vortex-driven model [55, 56] is proposed to explain the sock formation (Fig. 3.8A). We have also proposed a web-shell structure model (Fig. 3.8B) for the study of sock dynamics.The proposed model correlates well with experimental results.

3.3.2  CNT fibers The aerogel-like CNT sock in floating catalyst method can be transformed into a CNT fiber by direct-spinning [12, 50], bath-spinning [57], or ­rotating-anchor spinning method [26, 58, 59], as illustrated in Fig. 3.9. Since water does not wet or penetrate CNT [31], it is used in the bath for densifying the sock into a fiber. Alternatively, CNT sheets can be collected by directly winding the sock on a spool.

3.4  Structure and properties of CNT fibers CNT fibers collected from the direct-spinning technique have better alignment, smaller diameter, and linear density (0.02–0.5 tex) due to fast winding. CNT threads from the rotating-anchor method have the largest and a broad range of diameter and linear density (1–40 tex) [60]. The bath-­spinning method provides CNT fibers with intermediate diameter and linear density (0.1–1.0 tex). Representative images of CNT fibers from the three spinning methods are shown in Fig. 3.10. Measuring CNT fibers by their diameter can introduce errors up to a factor of five [22] due to their noncircular cross-sectional shape. Cross-sectional area

48

Carbon Nanotube Fibers and Yarns

Fig. 3.8  (A) Convection vortex-driven model for sock formation; (B) web-shell structure of the sock. (Reproduced with permission from Hou G, Su R, Wang A, Ng V, Li W, Song Y, et al. The effect of a convection vortex on sock formation in the floating catalyst method for carbon nanotube synthesis. Carbon 102 (2016) 513–519.)

measurement by FIB cutting and calculation directly from SEM (scanning electron microscopy) images can improve the accuracy but it is a time-consuming process. In light of this difficulty, Windle’s group suggested using specific stress expressed in N/tex for axial stress, which is precisely equivalent to the engineering stress/specific gravity in GPa/(g/cm3) [17].The specific stress at breakage is known as tenacity of the fiber in the textile industry, which requires only measuring the linear density (tex) and breaking load (N) of the fiber.



H2

CNT fiber

Elastic CNT aerogel

Reaction solution carrier gas

Furnace

Ceramic tube

1) Carbon Gas Carrier Gas Catalist

Hollow cylindrical CNT assembly Container

(B)

3) Heat

4) CNTs

5) Aerogel

Winding system

Water or alcohol

6) Rotating anchor 8) CNT-Yarn 7) CNT Arrays

On-line condensation

(A)

2nd Rotational winder for sample collection 1st Rotational winder for condensation

2) Reactor

9) Bobbin

(C)

49

Fig. 3.9  Methods of producing CNT fibers from sock: (A) Direct-spinning method. (B) Bath-spinning method. (C) Rotating-anchor spinning method. (Panel (A) reproduced with permission from Gspann T, Smail F, Windle A. Spinning of carbon nanotube fibres using the floating catalyst high temperature route: purity issues and the critical role of sulphur. Faraday Discuss. 173 (2014) 2–7, (B) from Wang JN, Luo XG, Wu T, Chen Y. High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity. Nat. Commun. 5 (2014) 3848, and (C) from Misak HE, Mall S. Electrical conductivity, strength and microstructure of carbon nanotube multi-yarns. Mater. Des. 75 (2015) 76–84.)

Carbon nanotube fibers spun directly from furnace

Reactor temperature 1290°C

Catalyst particles

Hydrocarbon Ferrocene/He Thiophene/He

50

Carbon Nanotube Fibers and Yarns

(A)

(B)

(C)

Fig.  3.10  CNT fiber produced by different spinning methods. (A) Direct spinning. (B) Bath spinning. (C) Rotating anchor spinning. (Panel (A) reproduced with permission from Gspann TS, Juckes SM, Niven JF, Johnson MB, Elliott JA, White MA, et al. High thermal conductivities of carbon nanotube films and micro-fibres and their dependence on morphology. Carbon 114 (2016) 160–168, (B) from Zhong XH, Li YL, Liu YK, Qiao XH, Feng Y, Liang J, et  al. Continuous multilayered carbon nanotube yarns. Adv. Mater. 22 (2010) 692–696, and (C) from Schauer MW, White MA. Tailoring industrial scale CNT production to specialty markets. MRS Proceedings 2015;1752:mrsf14-1752-mm04-07.)

The nanoscale CNTs usually self-assemble into mesoscale bundles, which are the bridging units for forming the macroscopic fiber and sheet. Directly correlating the structure of the nanoscale CNTs with the properties of macroscopic fiber and sheet is challenging. In one study the CNT fiber tensile strength was found not dependent on the individual CNT diameter and wall number [48], probably due to the dominant effect of CNT bundle length on load bearing instead of the diameter of individual CNTs. Thus for assessing the properties of CNT macroscopic entities, the properties of the CNT bundles, instead of individual CNTs, are used. It has been suggested that few-layer larger diameter CNTs can autocollapse under certain conditions, which increases the contact area and load transfer between nanotubes, and thus improves bundle strength [52, 61]. It is widely accepted that longer CNTs or bundles could increase the strength and elongation at break [61, 62]. Chapters 7 and 8 provide detailed discussions on the structural mechanics of CNT fibers.



Carbon nanotube fibers spun directly from furnace

51

The axial stress in a CNT fiber is limited by the stress transfer between adjacent CNT bundles through interfacial shear [34, 61]. The strength of CNT fibers is usually limited by the coherence of the network. Thus the CNT fiber strength should be affected by the bundle length, the coefficient of static friction between bundles, and the surface contact area [61]. Longer gauge length used in tensile testing has an adverse effect on the measured yarn strength, with lower strength measured at longer gauge length. A detailed discussion on fiber strength variability is given in Chapter 7. The fiber tensile strength decreased with the linear density of the fiber samples [60] (Fig. 3.11). This is consistent with dry-spun fibers from CNT forests (see Chapter 7). If we continue reducing the linear density and diameter to the greatest extent, the ultimate scenario is an individual CNT which has an extremely high strength. So linear density should be included when reporting fiber strength for a fair comparison. The carbonaceous coatings and impurity clusters on the bundle surface could enhance the tensile strength and stiffness through inter-bundle adhesion. Large impurity clusters have a similar strength-enhancing effect especially at shorter testing gauge length, although it reduces the degree of alignment [21, 62]. With less impurity cluster the stiffness of the CNT fiber could be increased [11].

0.7

1400 1300

0.6

1200 1100

0.5

1000

0.4

800 700 600

0.3

UTS (N/TEX)

UTS (MPa)

900 Large Yarn PAR 915,919

500 0.2

400 300 200

0.1

100 0

0 0

10

20

30

40

TEX (g/km)

Fig. 3.11  The strength of CNT fiber as a function of linear density. (Reproduced with permission from Schauer MW, Lashmore D, White B. Synthesis and properties of carbon nanotube yarns and textiles. MRS Proceedings 2008; 1081:1081-P03-05.)

52

(A)

Carbon Nanotube Fibers and Yarns

(B)

Fig. 3.12  CNT alignment in fibers spun at different winding rates: (A) winding rate of 8 m/min and (B) winding rate of 55 m/min. (Reproduced with permission from Alemán B, Reguero V, Mas B, Vilatela JJ. Strong carbon nanotube fibers by drawing inspiration from polymer fiber spinning. ACS Nano (2015).)

The CNT bundle alignment significantly influences the measured mechanical properties. Increasing the winding rate can lead to a higher degree of alignment [22, 48, 52]. Compared to a winding rate of 8 m/min, the higher winding rate of 55 m/min increased the fiber tensile strength from 0.3 to 1 N/tex, and modulus from 5 to 40 N/tex (Fig. 3.12).This increase is more than what is expected of the effect caused just by reduced fiber linear density. The increase in winding rate is accompanied by a reduced concentration of CNTs in the gas phase which reduces CNT entanglements [48]. In another study, with a winding rate increase from 5 to 20 m/min, fiber tensile strength and modulus increased from 0.5 to 2 N/tex and from 10 to 80 N/tex, respectively [52]. Both improved CNT alignment and finer fiber contribute to the substantial increase in fiber strength. Gapann et al. [62] modeled the alignment effect on the thread strength, and suggested that perfect aligned bundles are inefficient for shear stress transfer compared to CNT fiber with less perfect orientation and with the presence of amorphous carbonaceous coating and impurity clusters. Densification could improve the properties of the CNT fiber and sheet. Acetone spraying merges CNT bundles, improves the bundle alignment, and enhances their packing efficacy [63]. Aleman et al. [48] reported that although spraying acetone could reduce the fiber diameter by a factor of about 11, the alignment, electrical resistance, and specific tensile strength of the resulting fiber stayed almost unchanged.They concluded that the moderate densification could not bring CNT bundles sufficiently close to improve charge and load transfer. Post-processing methods, such as stretching and rolling, could also be used to provide stronger densification effect. Stretching helps alignment and enhances the anisotropy of the material [64, 65].



Carbon nanotube fibers spun directly from furnace

53

By using a roller stretcher with N-methyl-2-pyrrolidone (NMP) as lubricant, Schauer et al. [26] showed that the CNT fiber strength could be improved from 1.2 to 2.2 GPa. Wang et al. reported improvement of mechanical strength and electrical conductivity by more than one order of magnitude with pressurized rolling [57]. Large strain stretching (~10%–20%) is often used for processing CNT fibers produced by floating catalyst method [29, 30, 48], although less improvements (2%–5%) were reported [9, 22, 52, 62]. Stronger fibers are often associated with lower breaking strains, thus a compromise can lead to optimum fiber toughness [48, 63]. This seemingly contradictory relationship between strength and strain could be explained by the density and compactness of the fiber. For electrical properties, the existence of impurities and defects in CNT macro entities causes electron scattering and contact resistance while insufficient CNT alignment leads to increased junction resistance. The CNT inter-bundle junctions are a key component of the conductive path in the CNT thread [31], whose separation and breakage will cause a decrease in conductivity. Impurities have an adverse effect on the electrical conductivity. Schauer et al. found that acid purification treatment could improve the CNT fiber conductivity from 0.36 × 104 to 0.8 × 104 S/cm [26]. Some CNT fibers from floating catalyst method have a metal-like behavior, where the electrical resistance increases with temperature [32] at a positive temperature coefficient of resistance of 4 × 10−4 /K. In comparison, conventional carbon fiber shows a decrease in resistance with temperature at a negative coefficient of 4 × 10−4/K. Terrones et  al. [31] passed a current through CNT fiber immersed in polar liquids.They found that the electrostatic force due to charge accumulation at open junctions brings the bundles closer, or even closes the junctions, which increases the conductivity. Typical properties of CNT fibers produced by the floating catalyst method are listed in Table 3.1.

3.5  Conclusions and future directions In this chapter, the method of spinning CNT fibers from floating catalyst furnace has been reviewed. To produce a high-performance fiber, it is critical to produce high-quality CNTs by optimizing the synthesis parameters. These individual CNTs serve as the building block of the final fiber. There is a balance between quality and yield in the current processes. Optimization in the parameter space and further understanding of the processing mechanism will benefit the upscaling and future commercialization

54

Linear density [tex]

Specific strength [N/tex]

Tensile strength [GPa]

Electrical [106 S/m]

Diameter [μm]

Gauge [mm]

Apparent density [g/cc]

Method

Reference

0.02–0.1

0.8–2

–

0.7–0.8

7

20

1.0

Direct-spinning

0.05–0.5 0.11–0.92 – 0.83–2.09 1.4–3.8 1–2.5 1.3–10

1.2–2.2 0.2–0.75 – 0.44–0.65a 0.9 0.84 0.45–0.5

– 0.4–1.25 3.3–4.2 0.25–0.28 – 0.8–1.3 –

– 0.5 2.24 0.3 0.36 – 0.19–0.22

2–20 10–150 – 50–190 57.3 50 –

– 10 20 50 3.2 4.8 –

1.0 – 1.3–1.8 0.38–0.64 – – –

Direct-spinning Bath-spinning Bath-spinning Rotating-anchor Rotating-anchor Rotating-anchor Rotating-anchor

[11, 30, 45, 52, 63, 66] [9] [37] [57] [58] [19, 67] [68–70] [71]

a

Estimated using original paper’s data.

Carbon Nanotube Fibers and Yarns

Table 3.1  Typical properties of CNT fibers produced by the floating catalyst method.



Carbon nanotube fibers spun directly from furnace

55

of the material. The interaction between CNTs inside the fiber plays an important role in the final fiber properties. A closely-packed fiber structure will benefit the fiber properties. Further densification can be achieved by post-processing methods. The fibers spun from floating catalyst furnace provide a promising solution for scaling up the nanomaterial production. A combination of improved individual CNT structure, assembly method, and post-processing process is required to further improve the fiber properties.

Acknowledgment This research was broadly supported by ONR Award N00014-15-1-2473; the NSF ERC EEC-0812348; the UCTAC Seed Grant under ESP TECH 15-0160; the Univ. of Cincinnati Education and Research Center Targeted Research Training program (UC ERC-TRT Program); the Water Environment & Reuse Foundation; and the NSF I/UCRC Center for Intelligent Maintenance Systems (IMS).

References [1] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58, https:// doi.org/10.1038/354056a0. [2] A.  Oberlin, M.  Endo, T.  Koyama, Filamentous growth of carbon through benzene decomposition, J. Cryst. Growth 32 (1976) 335–349, https://doi.org/10.1016/00220248(76)90115-9. [3] H.P. Boehm, Carbon from carbon monoxide disproportionation on nickel and iron catalysts: morphological studies and possible growth mechanisms, Carbon 11 (1973) 583–590, https://doi.org/10.1016/0008-6223(73)90323-0. [4] M. Monthioux, V.L. Kuznetsov, Who should be given the credit for the discovery of carbon nanotubes?, Carbon 44 (2006) 1621–1623, https://doi.org/10.1016/j.carbon.2006.03.019. [5] A.J.  Hart, Chemical, Mechanical, and Thermal Control of Substrate-Bound Carbon Nanotube Growth, 68, 2006. [6] C. Paukner, K.K.K. Koziol, Ultra-pure single wall carbon nanotube fibres continuously spun without promoter, Sci. Rep. 4 (2014) 3903https://doi.org/10.1038/srep03903. [7] W. Ren, F. Li, S. Bai, H.-M. Cheng, The effect of sulfur on the structure of carbon nanotubes produced by a floating catalyst method, J. Nanosci. Nanotechnol. 6 (2006) 1339–1345, https://doi.org/10.1166/jnn.2006.301. [8] V.  Reguero, B.  Alemán, B.  Mas, J.J.  Vilatela, Controlling carbon nanotube type in macroscopic fibers synthesized by the direct spinning process, Chem. Mater. 26 (2014) 3550–3557, https://doi.org/10.1021/cm501187x. [9] M. Motta, A. Moisala, I.A. Kinloch, A.H. Windle, High performance fibres from ‘dog bone’ carbon nanotubes, Adv. Mater. 19 (2007) 3721–3726, https://doi.org/10.1002/ adma.200700516. [10] K.-H. Lee, S.-H. Lee, J. Park, H.-R. Kim, J. Lee, Synthesis of high-quality carbon nanotube fibers by controlling the effects of sulfur on the catalyst agglomeration during the direct spinning process, RSC Adv. (2015) 41894–41900, https://doi.org/10.1039/ C5RA04691B.

56

Carbon Nanotube Fibers and Yarns

[11] T. Gspann, F. Smail, A. Windle, Spinning of carbon nanotube fibres using the floating catalyst high temperature route: purity issues and the critical role of sulphur, Faraday Discuss. 173 (2014) 2–7, https://doi.org/10.1039/C4FD00066H. [12] Y.-L. Li, K. I a, A.H. Windle, Direct spinning of carbon nanotube fibers from chemical vapor deposition synthesis, Science 304 (2004) 276–278, https://doi.org/10.1126/ science.1094982. [13] Y.-L.  Li, L.-H.  Zhang, X.-H.  Zhong, A.H.  Windle, Synthesis of high purity ­single-walled carbon nanotubes from ethanol by catalytic gas flow CVD reactions, Nanotechnology 18 (2007) 225604https://doi.org/10.1088/0957-4484/18/22/225604. [14] M.  Motta, I.  Kinloch, A.  Moisala, V.  Premnath, M.  Pick, A.  Windle, The parameter space for the direct spinning of fibres and films of carbon nanotubes, Phys. E 37 (2007) 40–43, https://doi.org/10.1016/j.physe.2006.07.005. [15] M.  Motta, Y.L.  Li, I.A.  Kinloch, A.H.  Windle, Mechanical properties of continuously spun fibers of carbon nanotubes, Nano Lett. 5 (2005) 1529–1533, https://doi. org/10.1021/nl050634. [16] Z.  Li, Z.  Liu, H.  Sun, C.  Gao, Superstructured assembly of nanocarbons: fullerenes, nanotubes, and graphene, Chem. Rev. (2015) https://doi.org/10.1021/acs.chemrev.5b00102. [17] J.J.  Vilatela, A.H.  Windle, Yarn-like carbon nanotube fibers, Adv. Mater. 22 (2010) 4959–4963, https://doi.org/10.1002/adma.201002131. [18] J. Benson, I. Kovalenko, S. Boukhalfa, D. Lashmore, M. Sanghadasa, G. Yushin, Multifunctional CNT-polymer composites for ultra-tough structural supercapacitors and desalination devices, Adv. Mater. 25 (2013) 6625–6632, https://doi.org/10.1002/ adma.201301317. [19] A.S.  Wu, X.  Nie, M.C.  Hudspeth, W.W.  Chen, T.W.  Chou, D.S.  Lashmore, et  al., Strain rate-dependent tensile properties and dynamic electromechanical response of carbon nanotube fibers, Carbon 50 (2012) 3876–3881, https://doi.org/10.1016/j. carbon.2012.04.031. [20] Z.P. Wu, X.L. Huang, B. Li,Y.H. Yin,Y.S. Li,Y.S. Chen, et al., Strong carbon nanotube macro-films with retained deformability at fairly low temperatures, Phys. E 47 (2013) 285–289, https://doi.org/10.1016/j.physe.2012.10.004. [21] P. Davies, P. Papakonstantinou, N. Martin, I. Kratochvílová, C. Ewels, M. Shaffer, et al., Synthesis in gas and liquid phase: general discussion, Faraday Discuss. 173 (2014) 115– 135, https://doi.org/10.1039/C4FD90042A. [22] T.S.  Gspann, S.M.  Juckes, J.F.  Niven, M.B.  Johnson, J.A.  Elliott, M.A.  White, et  al., High thermal conductivities of carbon nanotube films and micro-fibres and their dependence on morphology, Carbon 114 (2016) 160–168, https://doi.org/10.1016/j. carbon.2016.12.006. [23] Q. Zhang, J.-Q. Huang, M.-Q. Zhao, W.-Z. Qian, F. Wei, Modulating the diameter of carbon nanotubes in array form via floating catalyst chemical vapor deposition, Appl. Phys. A Mater. Sci. Process. 94 (2009) 853–860, https://doi.org/10.1007/s00339-0084904-5. [24] X.-H. Zhong,Y.-L. Li, J.-M. Feng,Y.-R. Kang, S.-S. Han, Fabrication of a multifunctional carbon nanotube “cotton” yarn by the direct chemical vapor deposition spinning process, Nanoscale 4 (2012) 5614, https://doi.org/10.1039/c2nr31309j. [25] D. Conroy, A. Moisala, S. Cardoso, A. Windle, J. Davidson, Carbon nanotube reactor: ferrocene decomposition, iron particle growth, nanotube aggregation and scale-up, Chem. Eng. Sci. 65 (2010) 2965–2977, https://doi.org/10.1016/j.ces.2010.01.019. [26] M.W. Schauer, D.S. Lashmore, D.J. Lewis, B.M. Lewis, E.C. Towle, Strength and electrical conductivity of carbon nanotube yarns, Mater. Res. 1258 (2010). [27] J. Chaffee, D. Lashmore, D. Lewis, J. Mann, M. Schauer, B. White, Direct synthesis of CNT yarns and sheets. Nsti Nanotech 2008, Tech. Proc. 3 (2008) 118–121.



Carbon nanotube fibers spun directly from furnace

57

[28] Z.P. Wu, J.N. Wang, J. Ma, Methanol-mediated growth of carbon nanotubes, Carbon 47 (2009) 324–327, https://doi.org/10.1016/j.carbon.2008.09.034. [29] B. Mas, B. Alemán, I. Dopico, I. Martin-Bragado, T. Naranjo, E.M. Pérez, et al., Group 16 elements control the synthesis of continuous fibers of carbon nanotubes, Carbon 101 (2016) 458–464, https://doi.org/10.1016/j.carbon.2016.02.005. [30] E. Senokos, V. Reguero, J. Palma, J. Vilatela, R. MARCILLA, Macroscopic fibres of CNTs as electrodes for multifunctional electric double layer capacitors: from quantum capacitance to device performance, Nanoscale (2016) https://doi.org/10.1039/ C5NR07697H. [31] J. Qiu, J. Terrones, J.J. Vilatela, M.E. Vickers, J. Elliott, A.H. Windle, Liquid Infiltration into carbon nanotube fibers: effect on structure and electrical properties, ACS Nano (2013) 8412–8422. [32] J.  Terrones, E.  J a, J.J.  Vilatela, A.H.  Windle, Electric field-modulated non-ohmic behavior of carbon nanotube fibers in polar liquids, ACS Nano (2014) 8497–8504, https://doi.org/10.1021/nn5030835. [33] R.M.  Sundaram, K.K.K.  Koziol, A.H.  Windle, Continuous direct spinning of fibers of single-walled carbon nanotubes with metallic chirality, Adv. Mater. 23 (2011) 5064–5068, https://doi.org/10.1002/adma.201102754. [34] J.J. Vilatela, A.H. Windle, A multifunctional yarn made of carbon nanotubes, J. Eng. Fibers Fabrics 7 (2012) 23–28. [35] C.  Hoecker, F.  Smail, M.  Bajada, M.  Pick, A.  Boies, Catalyst nanoparticle growth dynamics and their influence on product morphology in a CVD process for continuous carbon nanotube synthesis, Carbon (2015) https://doi.org/10.1016/j.carbon.2015.09.050. [36] A.G. Nasibulin, P.V. Pikhitsa, H. Jiang, E.I. Kauppinen, Correlation between catalyst particle and single-walled carbon nanotube diameters, Carbon 43 (2005) 2251–2257, https://doi.org/10.1016/j.carbon.2005.03.048. [37] X.H. Zhong,Y.L. Li,Y.K. Liu, X.H. Qiao,Y. Feng, J. Liang, et al., Continuous multilayered carbon nanotube yarns, Adv. Mater. 22 (2010) 692–696, https://doi.org/10.1002/ adma.200902943. [38] K.K.K. Koziol, C. Ducati,A.H. Windle, Carbon nanotubes with catalyst controlled chiral angle, Chem. Mater. 22 (2010) 4904–4911, https://doi.org/10.1021/cm100916m. [39] A.  Morelos-Gómez, M.  Fujishige, S.  Magdalena Vega-Díaz, I.  Ito, T.  Fukuyo, R. Cruz-Silva, et al., High electrical conductivity of double-walled carbon nanotube fibers by hydrogen peroxide treatments, J. Mater. Chem. A 4 (2016) 74–82, https://doi. org/10.1039/c5ta06662j. [40] S.W. Pattinson, K. Prehn, I.A. Kinloch, D. Eder, K.K.K. Koziol, K. Schulte, et al., The life and death of carbon nanotubes, RSC Adv. 2 (2012) 2909, https://doi.org/10.1039/ c2ra00660j. [41] C.  Singh, M.S.P.  Shaffer, A.H.  Windle, Production of controlled architectures of aligned carbon nanotubes by an injection chemical vapour deposition method, Carbon 41 (2003) 359–368, https://doi.org/10.1016/S0008-6223(02)00314-7. [42] M. Endo, Grow carbon fibers in the vapor phase, Chemtech 18 (1988) 568–576. [43] A.  Windle, Carbon nanotube fibres: science and technology transfer, in: A.  Misra, J.R. Bellare (Eds.), Nanoscience and Technology for Mankind, The National Academy of Sciences, India, 2014. [44] B. Alemán, M.M. Bernal, B. Mas, E.M. Perez,V. Reguero, G. Xu, et al., Inherent predominance of high chiral angle metallic carbon nanotubes in continuous fibers grown from molten catalyst, Nanoscale (2016) https://doi.org/10.1039/C5NR07455J. [45] K.L.  Stano, K.  Koziol, M.  Pick, M.S.  Motta, A.  Moisala, J.J.  Vilatela, et  al., Direct spinning of carbon nanotube fibres from liquid feedstock, Int. J. Mater. Form. 1 (2008) 59–62, https://doi.org/10.1007/s12289-008-0380-x.

58

Carbon Nanotube Fibers and Yarns

[46] M.S. Motta, A. Moisala, I.A. Kinloch, A.H. Windle,The role of sulphur in the synthesis of carbon nanotubes by chemical vapour deposition at high temperatures, J. Nanosci. Nanotechnol. 8 (2008) 2442–2449. [47] Y. Alinejad, A. Shahverdi, N. Faucheux, G. Soucy, Synthesis of single-walled carbon nanotubes using induction thermal plasma technology with different catalysts: thermodynamic and experimental studies, J. Phys. Conf. Ser. 406 (2012) 012019https://doi. org/10.1088/1742-6596/406/1/012019. [48] B. Alemán,V. Reguero, B. Mas, J.J. Vilatela, Strong carbon nanotube fibers by drawing inspiration from polymer fiber spinning, ACS Nano (2015) https://doi.org/10.1021/ acsnano.5b02408. [49] W. Ren, F. Li, H.M. Cheng, Evidence for, and an understanding of, the initial nucleation of carbon nanotubes produced by a floating catalyst method, J. Phys. Chem. B 110 (2006) 16941–16946, https://doi.org/10.1021/jp062526x. [50] H.W. Zhu, C.L. Xu, D.H. Wu, B.Q. Wei, R. Vajtai, P.M. Ajayan, Direct synthesis of long single-walled carbon nanotube strands, Science 296 (2002) 884–886, https://doi. org/10.1126/science.1066996. [51] A. Kaskela, P. Laiho, N. Fukaya, K. Mustonen,T. Susi, H. Jiang, et al., Highly individual SWCNTs for high performance thin film electronics, Carbon 103 (2016) 228–234, https://doi.org/10.1016/j.carbon.2016.02.099. [52] K.  Koziol, J.  Vilatela, A.  Moisala, M.  Motta, P.  Cunniff, M.  Sennett, et  al., High-­ performance carbon nanotube fiber, Science 318 (2007) 1892–1895, https://doi. org/10.1126/science.1147635. [53] J.-P.  Matas, J.F.  Morris, É.  Guazzelli, Inertial migration of rigid spherical particles in Poiseuille flow, J. Fluid Mech. 515 (2004) 171–195, https://doi.org/10.1017/ S0022112004000254. [54] V. Balakrishnan, M. Bedewy, E.R. Meshot, S.W. Pattinson, E.S. Polsen, F. Laye, et al., Real-time imaging of self-organization and mechanical competition in carbon nanotube forest growth, ACS Nano 10 (2016) 11496–11504, https://doi.org/10.1021/ acsnano.6b07251. [55] G. Hou,V. Ng,Y. Song, L. Zhang, C. Xu,V. Shanov, et al., Numerical and experimental investigation of carbon nanotube sock formation, MRS Adv. (2016) 1–6, https://doi. org/10.1557/adv.2016.632. [56] G. Hou, R. Su, A. Wang,V. Ng, W. Li,Y. Song, et al., The effect of a convection vortex on sock formation in the floating catalyst method for carbon nanotube synthesis, Carbon 102 (2016) 513–519, https://doi.org/10.1016/j.carbon.2016.02.087. [57] J.N. Wang, X.G. Luo, T. Wu, Y. Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014) 3848https://doi.org/10.1038/ncomms4848. [58] H.E.  Misak, S.  Mall, Electrical conductivity, strength and microstructure of carbon nanotube multi-yarns, Mater. Des. 75 (2015) 76–84, https://doi.org/10.1016/j.matdes.2015.03.020. [59] D.  Lashmore, B.  White, M.  Schauer, J.  Mann, Synthesis and electronic properties SWCNT sheets, in: Materials Research Society Symposium Proceedings, 2008, https://doi.org/10.1557/PROC-1081-P06-09. 1081E:No pp. given, Paper #: 1081P06-09. [60] M.W. Schauer, D. Lashmore, B. White, Synthesis and properties of carbon nanotube yarns and textiles, in: MRS Proceedings, 2008, https://doi.org/10.1557/PROC1081-P03-05. 1081:1081-P03-05. [61] J.J. Vilatela, J.A. Elliott, A.H. Windle, A model for the strength of yarn-like carbon nanotube fibers, ACS Nano 5 (2011) 1921–1927, https://doi.org/10.1021/nn102925a. [62] T.S. Gspann, N. Montinaro, A. Pantano, J.A. Elliott, A.H. Windle, Mechanical properties of carbon nanotube fibres: St Venant’s principle at the limit and the role of imperfections, Carbon 93 (2015) 1021–1033, https://doi.org/10.1016/j.carbon.2015.05.065.



Carbon nanotube fibers spun directly from furnace

59

[63] A. Mikhalchan,T. Gspann, A. Windle, Aligned carbon nanotube-epoxy composites: the effect of nanotube organization on strength, stiffness, and toughness, J. Mater. Sci. 51 (2016) 10005–10025, https://doi.org/10.1007/s10853-016-0228-6. [64] E.  Cimpoiasu, D.  Lashmore, B.  White, G.A.  Levin, Anisotropic magnetoresistance of stretched sheets of carbon nanotubes, in: MRS Proceedings, 2012, https://doi. org/10.1557/opl.2012.360. 1407:mrsf11-1407-aa05-41. [65] E. Cimpoiasu, V. Sandu, G.A. Levin, A. Simpson, D. Lashmore, Angular magnetoresistance of stretched carbon nanotube sheets, J. Appl. Phys. 111 (2012) https://doi. org/10.1063/1.4729538. [66]. Windle A. Carbon Materials Comprising Carbon Nanotubes and Methods of Making Carbon Nanotubes, 2013. [67] A.S.  Wu, T.-W.  Chou, J.W.  Gillespie, D.  Lashmore, J.  Rioux, Electromechanical response and failure behaviour of aerogel-spun carbon nanotube fibres under tensile loading, J. Mater. Chem. 22 (2012) 6792, https://doi.org/10.1039/c2jm15869h. [68] F.A. Hill,T.F. Havel, D. Lashmore, M. Schauer, C. Livermore, Powering electric systems using carbon nanotube springs, in: PowerMEMS 2011, 2011. [69]. Lashmore DS, Jared C, Mark S. Injector Apparatus and Methods for Production of Nanostructures. US 2009/0117025 A1, 2009. [70] F.A. Hill,T.F. Havel, D. Lashmore, M. Schauer, C. Livermore, Storing energy and powering small systems with mechanical springs made of carbon nanotube yarn, Energy 76 (2014) 318–325, https://doi.org/10.1016/j.energy.2014.08.021. [71] M.W.  Schauer, M.A.  White, Tailoring industrial scale CNT production to specialty markets, in: MRS Proceedings, 2015, https://doi.org/10.1557/opl.2015.90. 1752:mrsf14-1752-mm04-07.

Further reading [72] G. Hou, D. Chauhan,V. Ng, C. Xu, Z. Yin, M. Paine, et al., Gas phase pyrolysis synthesis of carbon nanotubes at high temperature. Mater. Des. 132 (2017) 112–118, https:// doi.org/10.1016/j.matdes.2017.06.070. [73] G. Hou, V. Ng, C. Xu, L. Zhang, G. Zhang, V. Shanov, et al., Multiscale modeling of carbon nanotube bundle agglomeration inside a gas phase pyrolysis reactor. MRS Adv. (2017) 1–6, https://doi.org/10.1557/adv.2017.371.

This page intentionally left blank

CHAPTER 4

Solution-spun carbon nanotube fibers Menghe Miao

CSIRO Manufacturing, Geelong,VIC, Australia

4.1 Introduction In the production of regenerated and synthetic textile fibers, a polymer material is converted into a fluid state and spun into a continuous bundle of fibers by extrusion through a spinneret. Polymer fibers can be produced using a number of extrusion techniques, known as wet spinning, dry spinning, melt spinning, gel spinning, and electrospinninga. In wet spinning, a polymer is dissolved in a solvent for spinning. Typically, the spinneret is submerged in a chemical bath that causes the fiber to precipitate, and then solidify, as it emerges. The solid-spun carbon nanotube fibers and yarns discussed in Chapters 2 and 3 require the carbon nanotubes to be organized in a specific manner, either in vertical arrays (forests) that can be drawn into a continuous CNT network, or in a continuous stream drawn out directly from a CNTsynthesis furnace. These approaches do not lend themselves to the typical easy scale-up of chemical processes and limit the options for process and material optimization. It would be ideal that carbon nanotube fibers can be produced from carbon nanotubes without being limited to a specific format of organization. This way, carbon nanotubes can be optimized according to their required properties rather than their organization for fiber manufacturing. A good example is the production of textile fibers from polymers in which polymer synthesis and fiber extrusion are independent of each other. CNTs tend to form bundles rather than dissolving because of the strong van der Waals forces between them. Various spinning methods have been investigated for spinning CNT composite fibers and pure CNT fibers [1, 2]. Obviously, melt spinning is an option only for CNT-containing thermoplastic composite fibers, not for pure CNT fibers. a

https://en.wikipedia.org/wiki/Spinning_(polymers).

Carbon Nanotube Fibers and Yarns https://doi.org/10.1016/B978-0-08-102722-6.00004-3

Copyright © 2020 Elsevier Ltd. All rights reserved.

61

62

Carbon Nanotube Fibers and Yarns

This chapter provides a brief overview of methods used to spin nearly pure CNT fibers from carbon nanotubes suspended or dissolved in a solvent, known as wet spinning or solution spinning. Carbon nanotubes cannot be controlled and aligned in solution unless they are dispersed at the molecular level, that is, single-nanotube level. A major challenge to the production of neat CNT fibers is dispersing the CNTs at high enough concentration suitable for efficient alignment and effective coagulation [3]. Solution spinning of CNT-based fibers is based on CNT dispersions stabilized in surfactant solution, superacids or other solvents.

4.2  Spinning from surfactant-based solutions In 2000,Vigolo et al. reported a coagulation spinning approach that could be used to assemble CNTs into long ribbons and fibers [4]. As shown in Fig.  4.1, SWNTs were sonicated in an aqueous solution of sodium ­dodecyl sulfate (SDS), a surfactant that adsorbs at the surface of the nanotube bundles and stabilizes the nanotubes against van der Waals attractions. The surfactant forms micellar structures around the individual CNTs, which are kinetically stable because the surrounding surfactant molecules prevent the CNTs from bundling together again.

Fig. 4.1  Schematic representation of a rotating bath used for coagulating surfactant-­ dispersed SWNTs into a fiber [2]. (Reprinted with permission from N. Behabtua, M.J. Greena, M. Pasquali. Review: carbon nanotube-based neat fibers. Nano Today 3(5–6) (2008) 24–34.)



Solution-spun carbon nanotube fibers

63

At the right concentrations of SDS, the SWNTs were homogeneously dispersed to form a single-phase solution with a viscosity similar to that of pure water. An optimum was found at about 0.35 wt% of nanotubes and 1.0 wt% of SDS. This corresponds to the maximum amount of SWNTs achievable in homogeneous dispersions. The surfactant-stabilized SWNT suspension was then injected through a cylindrical spinneret in the stream of a polyvinyl alcohol (PVA) solution (5 wt%), which induced the coagulation of the nanotubes through bridging flocculation. As a result, a mesh could be obtained, which was then washed several times with pure water to remove most of the surfactant and polymer. The mesh was finally pulled out of the water and collapsed into a fiber consisting of a densely interconnected nanotube network making use of the capillary effect at drying. X-ray scattering analysis showed that the fibers were composed of SWCNT bundles, PVA chains, graphitic objects, and catalyst particles. The nanotubes, graphitic objects, and catalyst came from the synthesis of the raw nanotubes, whereas the PVA chains were introduced during the elaboration process, where they were adsorbed onto the nanotube bundles. The diameter of the resulting CNT fibers varied from several to  100 μm depending on the processing conditions, such as the diameter of the syringe needle, the flow rate of the injected solution, and the co-flowing polymer solution. The tensile strength and Young’s modulus of the resulting fibers were around 300 MPa and 40 GPa, respectively. The electrical conductivity at room temperature was about 10 S/cm, and a nonmetallic behavior was observed when the temperature was decreased. Posttreatments, such as hot-drawing, could enhance nanotube alignment and hence improved the mechanical performance of the CNT fibers. Vigolo et  al. [5] later applied a stretching treatment to improve the alignment of the single-wall carbon nanotubes in the fibers. The fibers were rewetted, swollen, and redried vertically under a tensile load with a weight attached to the fiber end. Once rewetted and swollen, they could be stretched up to 160%.The alignment of the nanotubes was studied by X-ray scattering and characterized by the full-width at half-maximum (FWHM) of the azimuthal intensity distribution. A smaller FWHM value indicates a higher level of CNT alignment. The FWHM varied from 75 to 80 degree for raw fibers to values smaller than 50 degree for stretched fibers, showing a substantial improvement of the SWNTs orientations after stretching treatment. As a result of the improvement in SWNT orientation, there was

64

Carbon Nanotube Fibers and Yarns

a substantial increase of the Young’s modulus, with values up to 40 GPa, in addition to a significant improvement of the tensile strength. Badaire et al. [6] studied the influence of stretching and heat treatment on the properties of SWNT fibers prepared according to Vigolo’s spinning method. The fibers were dried, rewetted under tensile load, and redried to improve the alignment. The PVA polymer was removed by annealing the fiber in hydrogen at 1000°C. The FWHM measured from X-ray scattering decreased linearly from 27.5 degree in the initial extruded fiber to 14.5 degree after stretching by 80%. Both electrical and thermal conductivity of neat or composite fibers were improved upon the alignment of the carbon nanotubes. However, the relative improvement due to further alignment was modest, by a factor of 3–4 for electrical conductivity for fibers having the same chemical nature. In sharp contrast, annealing the fibers to remove insulating polymers resulted in neat nanotube fibers with a significantly increased conductivity, by several orders of magnitude. Dalton et  al. [7–9] modified the method initially used by Vigolo et  al. and produced a nanotube gel fiber that was then converted into a 100-m-long solid nanotube composite fiber. Lithium dodecyl sulfate (LDS) was used as a surfactant in preparing the CNT solution. The solution was injected into the center of a cylindrical pipe in which the PVA coagulation solution flows. The resulting fibers were about 50 μm in diameter and contained around 60 wt% SWNTs and 40 wt% PVA. These fibers showed a tensile strength up to 1.8 GPa and a Young’s modulus up to 80 GPa. Although PVA chains in the CNT fiber enhanced the load transfer efficiency between CNTs and consequently improved the fiber mechanical performance, they resulted in lower electrical and thermal conductivity due to the high loading of PVA. Kozlov et  al. [10] described another flocculation-based wet-spinning process that resulted in polymer-free carbon nanotube fibers. Like the polymer-based coagulant spinning method used by Vigolo et  al. [4], the polymer-free spinning process utilized dilute, low-viscosity dispersions of carbon nanotubes (about 0.6 wt% or lower SWNT content). The spinning solution used was essentially the same lithium-dodecyl-sulfate-stabilized (LDS-stabilized) aqueous dispersion employed for spinning SWNT/PVA composite fiber used by Dalton et al. [7]. HiPco SWNTs of 0.6 wt% were dispersed using a horn sonicator in an aqueous solution of 1.2 wt% LDS surfactant. A narrow jet of this spinning solution was injected into the flocculation bath containing 37% hydrochloric acid, which rotated at 33 rpm. Flocculation of nanotubes in the spinning solution to form a gel fiber



Solution-spun carbon nanotube fibers

65

­ ccurred very close to the point of contact of the spinning solution and the o acid in the bath. This gel fiber (containing 90 wt% volatilizable liquid) was washed in methanol to remove the hydrochloric acid. The fiber was then pulled from the wash bath, stretched over a frame, and dried under tension. After annealing at 1000°C in argon to remove possible residual impurities, a pure SWNT fiber was obtained.The mechanical properties of the fiber were relatively low, with a specific breaking stress of 65 MPa/(g/cm3), Young's modulus of 12 GPa/(g/cm3), and a strain-to-failure of about 1%. However, it had an electrical conductivity of 140 S/cm, which was much higher than the SWNT-PVA composite fibers obtained by Dalton et al. [8,9].

4.3  Spinning from acid solutions Strong acids such as fuming sulfuric acid are used in the commercial production of high-performance synthetic fibers composed of rod-like polymers. SWNTs behave as rigid rods when dissolved in superacids [3]. Acid solvents have the unique ability to form liquid-crystalline dopes with a high concentration of SWNTs (Fig. 4.2). The protonation of single-walled carbon nanotubes in superacids allows them to be dispersed at high concentration, more than an order of magnitude higher than typical concentrations achieved in surfactants or organic solvents [11].The protonation is also fully reversible. At a high enough concentration (>4 wt%), the SWNTs coalesce and form ordered domains, behaving similarly to nematic liquid crystalline rod-like polymers. The ensuing electrostatic repulsion counteracts the attractive van der Waals interaction between CNTs.

Fig.  4.2  Microscopy images under cross polars (rotated by 0 and 45 degree, respectively) of SWNT (8 wt%) dissolved in sulfuric acid, showing the typical birefringent texture of liquid-crystalline solutions [2]. (Reprinted with permission from N. Behabtua, M.J. Greena, M. Pasquali. Review: carbon nanotube-based neat fibers. Nano Today 3(5–6) (2008) 24–34.)

66

Carbon Nanotube Fibers and Yarns

Ericson et al. [3] prepared an 8 wt% dispersion of purified SWNTs in 102% sulfuric acid (2 wt% excess SO3) in a nitrogen-purged dry box. The mixture was manually mixed and then transferred to a mixing apparatus via a stainless steel syringe. Extensive mixing was accomplished by two alternating pneumatic pistons, which pushed the SWNT dope back and forth through an actively rotating shear cell within an evacuated housing. When the viscosity reached a steady state, the SWNT material was extruded through a small capillary tube (Tg

∆

20

(A)

∆

Heat

0

Heat

2

Polymer

>Tg

CNT/polymer ∆L2 < ∆L1

Temperature (°C)

∆L1

16

A2

>Tm

∆

12

Heat

∆

Heat

Shrinkage (%)

14

Control PAN PAN/VGCNFs PAN/MWNTs PAN/SWNTs >Tm

CNT/polymer

Polymer

∆L′2 < ∆L′1

10

∆L′1

8

6

(B)

0

2

4

6

8

10

12

14

16

18

CNT surface area (m2/g)

Fig. 5.7  (A1) Thermal shrinkages of various types of CNT/PAN nanocomposites fibers as a function of temperature; (A2) corresponding thermal shrinkage of various fibers at 160°C as a function of CNT surface area [25a]; and (B) a scheme of the changes of molecular structure during thermal shrinkage of polymer and CNT/polymer fibers. (Source of (A1) and (A2): H.G. Chae, T.V. Sreekumar, T. Uchida, S. Kumar, A comparison of reinforcement efficiency of various types of carbon nanotubes in polyacrylonitrile fiber, Polymer 46 (24) (2005) 10925–10935.)

c­ onductivity five orders of magnitude lower than randomly aligned PMMA/ SWNT nanocomposites [51]. Wang et al. also reported a 25-fold reduction of electrical conductivity attributable to cold drawing of gel-spun CNTreinforced UHMW-PE composite fiber [52]. By comparison, Choi et  al. found that the electrical conductivity of Epoxy/SWNT nanocomposite increased after being moderately aligned under magnetic field [53]. Winey et al. systematically studied the electrical conductivity of SWNT/PMMA nanocomposite and drawn the conclusion that the CNT percolation structure depended on both CNT loading and its alignment. At a given CNT

84

Carbon Nanotube Fibers and Yarns

content, the conductivity of the nanocomposite with slightly anisotropic CNT orientations was higher than that with isotropic CNT orientations, however highly oriented CNTs led to a significantly reduced conductivity [54]. Potschke et al. reported that PC/MWNT isotropic film was conductive but highly drawn melt-spun PC/MWNT fiber lost its conductivity [55]. During CNT/polymer fiber drawing process, electrical conductivity of CNT-containing fiber decreases up to several orders of magnitude with increase of draw ratio [7b, 56]. For an isotropic nanocomposite, the CNT percolation threshold concentration is normally lower than 0.1 wt% [5a], and even as low as 0.045 vol% in PVC [57]. By comparison, the percolation threshold in a polymer/CNT fiber is much higher, for example, 0.5–2 wt% for PMMA/SWNT fiber [51, 54] and 1 wt% for PVA/MWNT [56a]. This is because the CNT conductive networks in an isotropic composite are disrupted by the alignment of CNTs during drawing. Whereas, if a highly drawn CNT/polymer nanocomposite fiber is further annealed at a temperature above its glass transition or melting point, the relaxation of oriented polymer chains will distort the orientation of CNTs and promotes the formation of CNT conductive paths, therefore improves the electrical conductivity. Peijs et al. annealed highly aligned PP/MWNT tape with a CNT concentration of 5.4 wt%, and observed that the conductivity was in the order: drawn and annealed films > isotropic film > drawn films [7f]. The increase in conductivity was ascribed to the thermal relaxation of anisotropic CNT bundles which reconstructed the CNT conductive networks. Similar phenomenon was also observed for CNT-PE/PP fiber [7b]. Since the percolation status of the CNT conductive network in CNT-polymer nanocomposites could change during deformation, many researchers proposed to use CNT-polymer nanocomposites as strain sensors [9b, 58]. CNT-containing polymer nanocomposite fibers have been found to have better solvent resistance than pure polymer fibers. The CNT-containing fibers can remain intact in a chemical solvent at a higher temperature and for a longer time than their pristine polymeric fibers, such as PVA [30] and PAN [27]. The addition of CNTs in polymer fibers is expected to improve fiber thermal conductivity, since individual CNTs have unusually high thermal conductivity, for example, 6600 W/mK for SWNT [59] and 3000 W/ mK for MWNT [60] although the high tube-tube and tube-polymer resistance can greatly hinder the thermal transfer [61]. Moderate improvements of thermal conductivity have been observed in many polymer/CNT nanocomposite films and bulk materials, but the thermal conductivity of



Carbon nanotube-reinforced polymer nanocomposite fibers

85

CNT-reinforced polymeric fiber has been rarely reported. Highly oriented CNTs and polymer chains as well as the existence of interphase improves the thermal transportation from CNT to polymer in nanocomposite fibers, and thus the enhancement of thermal conductivity by CNTs in an anisotropic fiber is also greater than that in an isotropic nanocomposite. This is an underexplored area and more works are needed.

5.3  Examples of high-performance CNT-reinforced polymeric fibers Over 3000 journal papers on the topic of CNT-reinforced polymers are published each year. Even though many of these papers show improved performances with the addition of CNTs in polymer matrices, practical applications of CNTs in high-performance polymer fibers are still very rare. The current fiber market is very cost sensitive. Unless CNT-reinforced fibers offer superior performances than current products, the processing cost of CNTs and nanocomposites could make the applications of CNTs unrealistic. Additionally, there are some huge technical barriers between laboratory scale samples and industrial products. The technology readiness levels (TRLs) of these works are mostly between 1 and 3 (the proof of concept). A TRL level of 5 or higher is required to reach the stage of technology demonstration.The investment barrier to bridge this gap is high while there is no guarantee of a successful outcome. In this section, we discuss the advancements of CNT-reinforced high-performance fibers with potential for commercial utilization, including CNT-reinforced PAN-based carbon fiber, PVA fiber, and aromatic fiber.

5.3.1  CNT-reinforced carbon fibers Carbon fiber, with its high tensile properties and relatively low density, is widely used as reinforcement in composite materials. Carbon fiber is made from polymer fibers, predominantly PAN fiber. The structures and properties of the PAN precursor fiber strongly affect the quality of the resulting carbon fiber. The traditional manufacturing methods for PAN precursor fibers are wet spinning and dry-jet wet spinning, which give PAN precursor fibers a strength of 0.4–0.6 GPa and a modulus of 8–12 GPa. Gel spinning is a relatively new spinning technology used to produce high-performance fibers [27, 62]. During gel-spinning, PAN polymer solution is gelled in a cold methanol bath and the gelled fiber is further hot-drawn to a very high draw ratio. Gel spinning was recently adopted for extrusion of PAN fibers.

86

Carbon Nanotube Fibers and Yarns

It is found that the gel-spun PAN-based carbon fibers exhibit both high strength (5.5–5.8 GPa) and high modulus (353–375 GPa) over conventional wet-spun PAN-based carbon fibers [63]. Incorporating a small amount of CNTs (0.25–1 wt%) in gel-spun PAN fiber increases the tensile strength of the precursor from 0.90 to 1.07 GPa and the Young’s modulus from 22.1 to 28.7 GPa [27].The addition of CNTs also extends the PAN molecular chains along their axis [25b, 27] mainly due to CNT-nitrile interactions [64], facilitates the formation of conjugated nitrile and reduces the formation of β-amino nitrile [25b, 33c, 65] during stabilization, and leads to highly ordered graphitic structures after carbonization [35b, 66]. 1 wt% CNT increases the tensile strength of the final carbon fiber from 1.9 to 3.4 GPa and Young’s modulus from 286 to 425 GPa [33c]. TEM images in Fig. 5.8B shows that the PAN molecules in the interphase regions are better aligned than in the polymer matrix [27]. The highly ordered PAN structure develops into an annular graphitic structure after stabilization and carbonization (Fig. 5.8D) that did not exist in pure PAN-based carbon fiber control sample (Fig. 5.8C). The formation of annular graphitic phase in carbonized CNT/PAN fiber has also been verified by TEM [67], Raman [33c], and WAXD [28]. Processing tension [33c] and temperature [66b, 67a] are known to affect the formation of graphitic phase. Higher tension during stabilization and carbonization improves the quality of carbon fiber [66b], and the addition of CNT increases the fiber strength required to bear higher processing tension [28]. With the addition of CNTs and the formation of annular graphitic structures, the electrical conductivity and thermal conductivity of the resulting carbon fibers are significantly improved [35a]. The 1 wt% CNTcontaining PAN-based carbon fiber showed 146% and 103% higher electrical conductivity and 37% and 25% higher thermal conductivity than the pure PAN-based carbon fibers T300 and IM7, respectively (Fig.  5.9A and B). These improvements pointed to high-performance multifunctional carbon fibers. The strength of a fiber is limited by the existence of defects in the fiber [1]. With the decrease of fiber diameter, the defect number inside a unit length of a fiber decreases, resulting in an increase of the fiber tensile strength while other physical characteristics of the fiber are kept unchanged. Electrospinning produces fibers with diameters in the range from 40 to 800 nm. However, polymer chains relax very fast in solution state, which can result in poor chain orientation and low tensile properties of the ­electro-spun fiber. Island-in-a-sea bicomponent spinning method has also been used to

Carbon nanotube-reinforced polymer nanocomposite fibers

87

Fig. 5.8  HR-TEM images of carbonized (A, C1, C2) gel-spun PAN fiber, (B2, B3, D1, D2) CNT/PAN fiber, and schematic representation of the composite fiber [33c, 35b]. (Sources: H.G. Chae, M.L. Minus, A. Rasheed, S. Kumar, Stabilization and carbonization of gel spun polyacrylonitrile/single wall carbon nanotube composite fibers, Polymer 48 (13) (2007) 3781–3789; B.A. Newcomb, L.A. Giannuzzi, K.M. Lyons, P.V. Gulgunje, K. Gupta, Y. Liu, M. Kamath, K. McDonald, J. Moon, B. Feng, High resolution transmission electron microscopy study on polyacrylonitrile/carbon nanotube based carbon fibers and the effect of structure develop­ment on the thermal and electrical conductivities, Carbon 93 (2015) 502–514.)

88

Carbon Nanotube Fibers and Yarns

80

50

T300 IM7 GT PAN GT PAN/CNT(0.42%) GT PAN/CNT(0.5%) GT PAN/CNT(1.0%)

Thermal conductivity (W/m-k)

Electrical conductivity (kS/m)

90

70

60

50 280

(A)

290

300

310

320

330

340

T300 IM7 GT PAN GT PAN/CNT(0.42%) GT PAN/CNT(0.5%) GT PAN/CNT(1.0%)

40

30

20

10 280

350

290

(B)

Temperature (K)

300

310

320

330

340

350

Temperature (K)

Carbon fiber diameter (µm) 1

8 7

Tensile strength (GPa)

10 Carbonized PAN Carbonized PAN/CNT (99/1)

6 5 4 3 2 1 0 0.01

(C)

0.1

1

10

100

1000

Cross-sectional area (µm2)

Fig.  5.9  (A) Electrical conductivities and (B) thermal conductivities of gel-spun PANbased carbon fiber (GT PAN) and gel-spun PAN/CNT-based carbon fibers (GT PAN/CNT) [35b]; and (C) diameter-dependent tensile strength of carbonized gel-spun PAN and PAN/CNT fibers [28]. (Source of (A) and (B): B.A. Newcomb, L.A. Giannuzzi, K.M. Lyons, P.V. Gulgunje, K. Gupta, Y. Liu, M. Kamath, K. McDonald, J. Moon, B. Feng, High resolution transmission electron microscopy study on polyacrylonitrile/carbon nanotube based carbon ­fibers and the effect of structure development on the thermal and electrical conductivities, Carbon 93 (2015) 502–514. Source of (C): H.G. Chae, Y.H. Choi, M.L. Minus, S. Kumar, Carbon nanotube reinforced small di­ameter polyacrylonitrile based carbon fiber, Compos. Sci. Technol. 69 (3–4) (2009) 406–413.)

p­ roduce fine PAN-based carbon fibers with diameters down to 500 nm. The strength of the resulting carbon fiber increases with decreasing fiber diameter (Fig. 5.9C).The strength of CNT-containing carbon fiber is 30%–60% higher than control pure PAN-based carbon fiber. The reinforcement efficiency of CNT becomes higher at a smaller diameter with the effective CNT reinforcement stresses of 67, 61, and 28 GPa for carbon fiber diameters of 1, 6, and



Carbon nanotube-reinforced polymer nanocomposite fibers

89

12 μm, respectively. CNTs have also been used as surface modifier for carbon fibers to reinforce the interfacial stress transfer between the carbon fibers and the matrixes, leading to improvements to interface adhesion, interfacial strength, and in-plane shear strength of the resulting composites [68].

5.3.2  PVA fiber PVA crystal has a planar zig-zag structure similar to that of PE.The theoretical modulus of PVA fiber is as high as 250–300 GPa. Kuraray uses a gel spinning technique to produce high-performance PVA fibers with excellent tensile properties (tensile modulus: 11–43 GPa, tensile strength: 0.9–1.9 GPa) and high alkali resistance. CNTs show strong nucleate and templating effects on PVA crystals [23, 30, 69] and can be used to reinforce PVA fibers. Zhang et al. found that by the addition of 3 wt% CNTs, the tensile strength and modulus of gel-spun PVA fibers increased from 0.9 to 1.1 GPa and from 25.6 to 35.8 GPa, respectively [44b]. Miaudet et al. added 0.35 wt% of SWNT in wet-spun PVA fibers and found that the tensile modulus increased from 2.5 to 14.5 GPa and the strength increased from 0.6 to 1.6 GPa [70]. Xu et al. dispersed 0.3 wt% of SWNT into gel-spun PVA fibers and found that the tensile modulus was improved from 28 to 36 GPa and the strength from 1.7 to 2.2 GPa [71]. Dai et al. used 0.6 wt% tea polyphenol-functionalized MWNT to reinforce PVA fibers and observed a strength improvement of more than 160% [72]. Zhou et al. prepared PVA composite fibers with 20 wt% MWNT and obtained a high modulus of 41.5 GPa and an electrical conductivity of 6.85 × 10−3 S/m [73]. Minus et al. attempted to optimize gel spinning and drawing parameters [30]. When the fiber was finally drawn at 290°C, the modulus and strength of the PVA fiber reached as high as 48 and 1.6 GPa, respectively. Under the same spinning and drawing conditions, 1 wt% SWNT/PVA fiber showed a modulus of 71 GPa and a strength of 2.6 GPa. In a later study, Minus et al. used steady shear flow gel-spinning method to produce a PVA/SWNT nanocomposite fiber with tensile strength, modulus, and toughness of 4.9 GPa, 128 GPa, and 202 J/g, respectively, which are the highest reported values up till now [74]. The key factor for the excellent reinforcement of CNTs on PVA fibers is the development of highly crystallized and oriented interphase structures.

5.3.3  Aromatic fiber Poly(p-phenylene benzobisoxazole) (PBO) is a rigid-rod polymer [75]. PBO fiber has high tensile strength, stiffness, and thermal stability. A PBO fiber was commercialized by Toyobo Co. (Japan) in 1998 with the trade name Zylon. PBO exhibits lyotropic liquid crystalline behavior similar to

90

Carbon Nanotube Fibers and Yarns

poly(p-phenylene terephthalamide) (PPTA, Kevlar). It has been observed that CNTs orient in liquid crystals [76]. Kumar et  al. synthesized PBO with well-dispersed SWNTs in poly(phosphoric acid) and used dry-jet wet-spinning method to produce PBO/SWNT nanocomposite fibers [77]. The cross-polarized optical images of PBO and PBO/SWNT solutions (Fig.  5.10A) suggest a good dispersion of SWNT in the dope without noticeable aggregation. With the addition of 10 wt% SWNT to the PBO fibers, the tensile modulus increased from 138 to 167 GPa and the strength from 2.6 to 4.2 GPa, which are approximately 20% and 60% increases, respectively (Fig.  5.10B). The commercial PBO fiber (Zylon HM) has a strength of 5.8 GPa. If a similar reinforcement can be achieved, the strength of PBO/SWNT fiber could be expected to be higher than 8 GPa.

Fig.  5.10  Cross-polarized optical micrographs of (A1) 14 wt% PBO in PPA and (A2) 14 wt% SWNT/PBO (10/90) in PPA; (B) typical stress-strain curves for PBO and SWNT/ PBO (10/90) fibers [77]. (Source: S. Kumar, T.D. Dang, F.E. Arnold, A.R. Bhattacharyya, B.G. Min, X. Zhang, R.A. Vaia, C. Park, W.W. Adams, R.H. Hauge, R.E. Smalley, S. Ramesh, P.A. Willis, Synthesis, structure, and properties of PBO/SWNT composites, Macromolecules 35 (24) (2002) 9039–9043.)



Carbon nanotube-reinforced polymer nanocomposite fibers

91

Kevlar may be reinforced in the same way. Kim et al. used in-situ polymerization to synthesize PPTA/CNT nanocomposite solution and found a huge improvement of electrical conductivity [78]. Sainsbury et al. grafted PPTA onto MWNTs to achieve a better dispersion than directly mixing unmodified MWNTs with PPTA [79]. PPTA was dissolved in fuming sulfuric acid (H2SO4) to protonate SWNTs and obtain exfoliated SWNT [32]. Cao et  al. blended PPTA/H2SO4 solution with SWNTs and studied the rheology of the blend. When the SWNT concentration was higher than 0.2 wt%, single-phase nematic liquid crystals were formed at 85°C, which could be used to produce superior-performance fibers [80]. Deng et  al. compared the structures and tensile properties of PPTA and PPTA/SWNT fibers, and observed that the CNT orientation was lower than PPTA chains in the nanocomposite fibers. The measured tensile curves suggested that reinforcement only occurred at low draw ratios (~2) and could become detrimental to the fiber at a draw ratio higher than 4 [81]. Instead of mixing CNT with polymers, O’Connor et al. developed a new approach to produce PPTA/CNT fibers by swelling Kevlar fibers in CNT suspensions.The fiber strength increased from 3.9 to 4.8 GPa and the modulus from 120 to 130 GPa when 1 wt% CNT was absorbed by Kevlar [82]. CNTs could also be used as a bridging agent to enhance the interfacial interaction between Kevlar [83] or PBO [84] fibers and epoxy resins with significantly improved performances for the resulting composites.

5.4 Challenges CNTs have exceptional mechanical properties. However, in most of the reported cases, the properties of CNTs are not fully utilized due to a number of reasons.

5.4.1  CNT dispersion Fig. 5.11A shows the dependence of specific surface area of CNTs on the wall number of the CNTs. With the increase of CNT wall number, the specific surface area of the CNTs dramatically decreases. For the reinforcement of polymer fibers, SWNT is theoretically superior to MWNT. Since CNT would always be damaged during processing, there are experimental evidences that double-walled and few-walled CNTs (DWNT and FWNT) are preferred over SWNT for retaining mechanical properties of CNTs in the final nanocomposites. Fig. 5.11B shows the specific surface area of a SWNT bundle as a function of the number

Carbon Nanotube Fibers and Yarns

103

Surface area (m2/g)

Surface area (m2/g)

92

102

101

100

0

10

20

30

102

40

(B)

CNT wall number

CNT bundle diameter (nm)

(A)

103

0

200

400

600

800

1000

CNT number in a SWNT bundle

25

20

15

10

5

(C)

DMAc DMF

0

1000

2000

3000

CNT aspect ratio

Fig. 5.11  (A) CNT surface area as a function of wall number; (B) surface area as a function of CNT number in a SWNT bundle; and (C) dispersed CNT bundles diameters in a solvent as a function of CNT aspect ratio [85]. (Source: X. Yan, H. Dong, Y. Liu, B.A. Newcomb, H.G. Chae, S. Kumar, Z. Xiao, T. Liu, Effect of processing conditions on the dispersion of carbon nanotubes in polyacrylonitrile solutions, J. Appl. Polym. Sci. 132 (26) (2015) 42177.)

of CNTs in the bundle. Although CNTs have high surface areas up to ~1330 m2/g (SWNT), the interfacial area between the CNTs and the matrixes is dramatically reduced if the CNTs aggregate in bundles instead of being exfoliated individual tubes. The interfacial area plays an important role in stress transfer between the CNTs and the polymer [10], which is critical for reinforcement of composites. Our studies found that the dispersed CNT bundle diameter (bundle size) is dependent on the aspect ratio of the CNT (Fig. 5.11C) [85]. The interactions among CNTs, solvent molecules, and polymer chains can affect the dispersion of CNTs [86]. Chemical modifications and surfactants have been used to improve the dispersion of CNTs in solvents and polymer matrices, but it is not clear how the grafting groups and surfactants affect interphase development. Poor CNT dispersion is a detriment to the reinforcement of polymer nanocomposites fibers.



Carbon nanotube-reinforced polymer nanocomposite fibers

93

5.4.2  CNT alignment The alignment of CNTs is critical to the modulus of nanocomposite fibers. According to continuum mechanics, the effect of SWNT rope orientation on its reinforcement modulus inside a nanocomposite could be estimated by the following equation:  1 2γ  1 1 1 = cos4 θ + sin 4 θ +  − 12  sin 2 θ cos2 θ Ex E1 E2 E1   G12

(5.1)

where E1 is the longitude modulus and E2 is the transverse modulus of CNT ropes, G12 is the in-plane shear modulus, Ex is the reinforcement modulus of the CNT rope along the axis direction, and θ is the angle between CNT rope and fiber axis. For aggregated CNTs, G12 depends on the lower values of in-plane shear modulus of CNT ropes or CNT/polymers; for exfoliated CNTs, G12 is the in-plane shear modulus of CNT/ polymers. Normally, the full-width of half-maximum (FWHM) of the CNT orientation angle distribution could be used to represent the orientation factor of CNTs in a polymer fiber. In Eq. (5.1), CNT orientation is the primary factor affecting the modulus while in-plane shear modulus G12 plays a secondary role (Fig. 5.12). G12 depends on the lower value of

Fig. 5.12  Relationships between CNT rope orientation and its reinforcement modulus inside a nanocomposites fibers [87]. (Source: T. Liu, S. Kumar, Effect of orientation on the modulus of SWNT films and fibers, Nano Lett. 3 (5) (2003) 647–650.)

94

Carbon Nanotube Fibers and Yarns

the modulus between CNT tubes inside a bundle and modulus between the CNT and the polymer matrix. Once CNTs aggregate into bundles, CNTs can easily slip between each other and G12 between CNTs inside a CNT bundle becomes as low as 1 GPa. Thus, a highly drawn nanocomposite fiber with well-aligned individual CNTs will maximize the reinforcement effect.

5.4.3  CNT/matrix interfacial shear strength Stress transfer between CNT and polymer matrix is essential for strength of CNT-reinforced polymer composite fibers. A typical scheme of a CNT embedded in polymer matrix is shown in Fig. 5.13. During stretching or compressing, stress is transferred from the matrix to the CNT. A critical length (Lc) of CNT is defined as the minimum length required to avoid CNT being pulled out from the polymer matrix: σ D (5.2) L c = CNT CNT 2τ c where τc is the interfacial shear strength between the CNT and the polymer matrix, σCNT is the CNT strength, and DCNT is the diameter of the CNT. The critical aspect ratio (Rc) of CNT is therefore: Rc =

Lc σ = CNT DCNT 2τ c

(5.3)

Fig. 5.13 shows the relationship between Rc and τc of a CNT in polymer matrix. It is clear that a high strength CNT gives a high Rc value. For a

s = 20GPa s = 30GPa s = 40GPa

750

s (x)

Rc

500

250 30 MPa

0

25

50

75 100 tc (MPa)

125

150

Fig. 5.13  Relationships between Rc and τc. The strength of CNT is assumed to be 20, 30, and 40 GPa, respectively.



Carbon nanotube-reinforced polymer nanocomposite fibers

95

perfect SWNT with a strength greater than 140 GPa, the Rc is extremely high. As discussed earlier, as CNT aspect ratio increases, it becomes more and more difficult to disperse individual CNTs uniformly in the nanocomposites. On the other hand, a higher τc leads to a lower Rc.The formation of interphase between the CNT and the polymer matrix improves the stress transfer and the interfacial strength (τc). Different CNT surface chemical modification treatments have been designed to improve CNT dispersion and the interfacial stress transfer in nanocomposites [88]. However, the existence of CNT surface functional groups, surfactant and interface compatibilizer can also adversely affect the formation of highly ordered interphase structures. In summary, the CNT length, dispersion, orientation, and interfacial shear strength are key factors for the mechanical performances of nanocomposite fibers.These factors can work against each other. For example, a high CNT aspect ratio is preferred for interfacial stress transfer, but it makes the dispersion of CNTs in the nanocomposites extremely difficult.

5.5 Perspective Due to the large surface area and 1D rod structure, CNTs are a preferred reinforcement material for polymer fibers. Incorporating CNTs in polymers has been found to improve the tensile properties of almost all fibers, including but not limited to PE, PP, PVA, PC, polymethyl methacrylate, PBO, cellulose, PAN, and its derived carbon fibers. CNTs act as nucleating agent for polymer crystallization and template for polymer orientation. In a CNT-reinforced nanocomposite fiber, the reinforcement of tensile property comes not only from the superior properties of CNTs, but also from the formation of highly ordered interphase polymer structures in many cases. The first-generation nanocomposite fibers are based on the full utilization of the excellent properties of the nano-fillers as well as the dispersion, orientation, and interfacial shear strength of the nano-fillers in the polymer matrixes. Beside tensile properties, the addition of CNTs in polymeric fibers can also improve other properties of the composite fibers, including electrical and thermal conductivity. By adjusting the type of CNTs as well as their concentrations and orientations, electrical and thermal conductivity can be improved and tailored for different applications. The electrical and thermal conductivity of CNT/polymer nanocomposite fibers can be further tuned during the fiber extrusion and drawing stages.

96

Carbon Nanotube Fibers and Yarns

Besides the above key factors in first-generation nanocomposite fibers, the development of the next-generation nanocomposite fibers will focus on the developments of highly crystallized and oriented interphase polymer structures. There are still some uncertainties in the developments of nanocomposite interphase, including (1) how nanocomposite solution preparation and fiber processing methods affect the formation of interphase structures; (2) how the curvature and chirality of CNTs affect the interphase structures; (3) how the CNT surface chemical structures affect the interphase structures and properties; and (4) how the interphase structure develops under external stimulations. A good understanding of the interphase formation, microstructures, and properties are required for the development of the next-generation nanocomposite fibers.

References [1] H.G. Chae, S. Kumar, Making strong fibers, Science 319 (5865) (2008) 908–909. [2] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (6348) (1991) 56–58. [3] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang,Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Electric field effect in atomically thin carbon films, Science 306 (5696) (2004) 666. [4] O. Breuer, U. Sundararaj, Big returns from small fibers: a review of polymer/carbon nanotube composites, Polym. Compos. 25 (6) (2004) 630–645. [5] (a) Z. Spitalsky, D. Tasis, K. Papagelis, C. Galiotis, Carbon nanotube-polymer composites: chemistry, processing, mechanical and electrical properties, Prog. Polym. Sci. 35 (3) (2010) 357–401. (b) E.T. Thostenson, Z. Ren, T.-W. Chou, Advances in the ­science and technology of carbon nanotubes and their composites: a review, Compos. Sci. ­Technol. 61 (13) (2001) 1899–1912. (c) M. Moniruzzaman, K.I. Winey, Polymer ­nanocomposites containing carbon nanotubes, Macromolecules 39 (16) (2006) 5194–5205. (d) J.N. Coleman, U. Khan, W.J. Blau, Y.K. Gun’ko, Small but strong: a review of the mechanical properties of carbon nanotube-polymer composites, Carbon 44 (9) (2006) 1624–1652.(e)M.R. Loos, K. Schulte, Is it worth the effort to reinforce polymers with carbon nanotubes? Macromol. Theory Simul. 20 (5) (2011) 350–362. [6] (a) P.M. Ajayan, Nanotubes from carbon, Chem. Rev. 99 (7) (1999) 1787–1800. (b) R.H.  Baughman, A.A.  Zakhidov, W.A.  de Heer, Carbon nanotubes—the route toward applications, Science 297 (5582) (2002) 787–792. [7] (a) R. Jain, Carbon nanotube reinforced polyacrylonitrile and poly(etherketone) fibers, PhD Thesis, Georgia Institute of Technology, United States—Georgia, 2009. (b) X. Gao, S. Zhang, F. Mai, L. Lin, Y. Deng, H. Deng, Q. Fu, Preparation of high ­performance conductive polymer fibres from double percolated structure, J. Mater. Chem., 21 (17) (2011) 6401–6408. (c)A.  Soroudi, M.  Skrifvars, Melt blending of carbon nanotubes/polyaniline/polypropylene compounds and their melt spinning to conductive fibres, Synth. Met. 160 (11−12) (2010) 1143–1147. (d)K. Liu, Y. Sun, X. Lin, R. Zhou, J. Wang, S. Fan, K. Jiang, Scratch-resistant, highly conductive, and high-strength carbon nanotube-based composite yarns, ACS Nano 4 (10) (2010) 5827–5834.



Carbon nanotube-reinforced polymer nanocomposite fibers

97

(e)E.  Bilotti, R.  Zhang, H.  Deng, M.  Baxendale, T.  Peijs, Fabrication and property prediction of conductive and strain sensing TPU/CNT nanocomposite fibres, J. Mater. Chem. 20 (42) (2010) 9449–9455. (f) H. Deng, R. Zhang, C.T. Reynolds, E. Bilotti, T. Peijs, A novel concept for highly oriented carbon nanotube composite tapes or fibres with high strength and electrical conductivity, Macromol. Mater. Eng. 294 (11) (2009) 749–755. (g) P. Xue, K.H. Park, X.M. Tao, W. Chen, X.Y. Cheng, Electrically conductive yarns based on PVA/carbon nanotubes, Compos. Struct. 78 (2) (2007) 271–277. [8] S. Uchida, A. Martinez,Y.-W. Song, T. Ishigure, S. Yamashita, Carbon nanotube-doped polymer optical fiber, Opt. Lett. 34 (20) (2009) 3077–3079. [9] (a) N.  Lachman, C.l.  Bartholome, P.  Miaudet, M.  Maugey, P.  Poulin, H.D.  Wagner, Raman response of carbon nanotube/PVA fibers under strain, J. Phys. Chem. C 113 (12) (2009) 4751–4754. (b) S. Gong, D. Wu, Y.X. Li, M.Y. Jin, T. Xiao, Y. Wang, Z. Xiao, Z.H. Zhu, Z. Li, Temperature-independent piezoresistive sensors based on carbon nanotube/polymer nanocomposite, Carbon 137 (2018) 188–195. (c) M. Amjadi, A. Pichitpajongkit, S. Lee, S. Ryu, I. Park, Highly stretchable and sensitive strain sensor based on silver nanowire-elastomer nanocomposite, ACS Nano 8 (5) (2014) 5154–5163. [10] M.  Cadek, J.N.  Coleman, K.P.  Ryan, V.  Nicolosi, G.  Bister, A.  Fonseca, J.B.  Nagy, K. Szostak, F. Béguin, W.J. Blau, Reinforcement of polymers with carbon nanotubes: the role of nanotube surface area, Nano Lett. 4 (2) (2004) 353–356. [11] B.P. Grady, Recent developments concerning the dispersion of carbon nanotubes in polymers, Macromol. Rapid Commun. 31 (3) (2010) 247–257. [12] (a) J. Sandler, G. Broza, M. Nolte, K. Schulte,Y.M. Lam, M.S.P. Shaffer, Crystallization of carbon nanotube and nanofiber polypropylene composites, J. Macromol. Sci. B 42 (3–4) (2003) 479–488. (b) A.R. Bhattacharyya, T.V. Sreekumar, T. Liu, S. Kumar, L.M. Ericson, R.H. Hauge, R.E. Smalley, Crystallization and orientation studies in polypropylene/single wall carbon nanotube composite, Polymer 44 (8) (2003) 2373–2377. (c) B.P. Grady, F. Pompeo, R.L. Shambaugh, D.E. Resasco, Nucleation of polypropylene crystallization by single-walled carbon nanotubes, J. Phys. Chem. B 106 (23) (2002) 5852–5858. [13] O. Probst, E.M. Moore, D.E. Resasco, B.P. Grady, Nucleation of polyvinyl alcohol crystallization by single-walled carbon nanotubes, Polymer 45 (13) (2004) 4437–4443. [14] J.K.W.  Sandler, S.  Pegel, M.  Cadek, F.  Gojny, M.  van Es, J.  Lohmar, W.J.  Blau, K.  Schulte, A.H.  Windle, M.S.P.  Shaffer, A comparative study of melt spun ­polyamide-12 fibres reinforced with carbon nanotubes and nanofibres, Polymer 45 (6) (2004) 2001–2015. [15] R. Haggenmueller, J.E. Fischer, K.I. Winey, Single wall carbon nanotube/polyethylene nanocomposites: nucleating and templating polyethylene crystallites, Macromolecules 39 (8) (2006) 2964–2971. [16] K.A.  Anand, T.S.  Jose, U.S.  Agarwal, T.V.  Sreekumar, B.  Banwari, R.  Joseph, PETSWNT nanocomposite fibers through melt spinning, Int. J. Polym. Mater. 59 (6) (2010) 438–449. [17] G.-W. Lee, S. Jagannathan, H.G. Chae, M.L. Minus, S. Kumar, Carbon nanotube dispersion and exfoliation in polypropylene and structure and properties of the resulting composites, Polymer 49 (7) (2008) 1831–1840. [18] L.  Li, C.Y.  Li, C.  Ni, Polymer crystallization-driven, periodic patterning on carbon nanotubes, J. Am. Chem. Soc. 128 (5) (2006) 1692–1699.

98

Carbon Nanotube Fibers and Yarns

[19] S. Zhang, W. Lin, C.-P. Wong, D.G. Bucknall, S. Kumar, Nanocomposites of carbon nanotube fibers prepared by polymer crystallization, ACS Appl. Mater. Interfaces 2 (6) (2010) 1642–1647. [20] (a) S. Zhang, M.L. Minus, L. Zhu, C.-P. Wong, S. Kumar, Polymer transcrystallinity induced by carbon nanotubes, Polymer 49 (5) (2008) 1356–1364. (b) J.P. Abdou, K.J. Reynolds, M.R. Pfau, J. van Staden, G.A. Braggin, N. Tajaddod, M.  Minus, V.  Reguero, J.J.  Vilatela, S.J.  Zhang, Interfacial crystallization of isotactic polypropylene surrounding macroscopic carbon nanotube and graphene fibers, Polymer 91 (2016) 136–145. [21] W.  Ding, A.  Eitan, F.T.  Fisher, X.  Chen, D.A.  Dikin, R.  Andrews, L.C.  Brinson, L.S.  Schadler, R.S.  Ruoff, Direct observation of polymer sheathing in carbon ­nanotube−polycarbonate composites, Nano Lett. 3 (11) (2003) 1593–1597. [22] S. Zhang, S. Kumar, Shaping polymer particles by carbon nanotubes, Macromol. Rapid Commun. 29 (7) (2008) 557–561. [23] M.L. Minus, H.G. Chae, S. Kumar, Single wall carbon nanotube templated oriented crystallization of poly(vinyl alcohol), Polymer 47 (11) (2006) 3705–3710. [24] K.A.  Anand, U.S.  Agarwal, R.  Joseph, Carbon nanotubes induced crystallization of poly(ethylene terephthalate), Polymer 47 (11) (2006) 3976–3980. [25] (a) H.G. Chae, T.V. Sreekumar, T. Uchida, S. Kumar, A comparison of reinforcement efficiency of various types of carbon nanotubes in polyacrylonitrile fiber, Polymer 46 (24) (2005) 10925–10935. (b) Y. Liu, H.G. Chae, S. Kumar, Gel-spun carbon nanotubes/polyacrylonitrile composite fibers. Part I: effect of carbon nanotubes on stabilization, Carbon 49 (13) (2011) 4466–4476. [26] J.N.  Coleman, M.  Cadek, R.  Blake, V.  Nicolosi, K.P.  Ryan, C.  Belton, A.  Fonseca, J.B. Nagy,Y.K. Gun’ko,W.J. Blau, High performance nanotube-reinforced plastics: understanding the mechanism of strength increase, Adv. Funct. Mater. 14 (8) (2004) 791–798. [27] H.G. Chae, M.L. Minus, S. Kumar, Oriented and exfoliated single wall carbon nanotubes in polyacrylonitrile, Polymer 47 (10) (2006) 3494–3504. [28] H.G. Chae, Y.H. Choi, M.L. Minus, S. Kumar, Carbon nanotube reinforced small diameter polyacrylonitrile based carbon fiber, Compos. Sci. Technol. 69 (3–4) (2009) 406–413. [29] Y. Li,Y. Yu,Y. Liu, C. Lu, Interphase development in polyacrylonitrile/SWNT nanocomposite and its effect on cyclization and carbonization for tuning carbon structures, ACS Appl. Nano Mater. 1 (7), (2018), 3105–3113. [30] M.L.  Minus, H.G.  Chae, S.  Kumar, Interfacial crystallization in gel-spun poly(vinyl alcohol)/single-wall carbon nanotube composite fibers, Macromol. Chem. Phys. 210 (21) (2009) 1799–1808. [31] A.H. Barber, S.R. Cohen, H.D. Wagner, Measurement of carbon nanotube—polymer interfacial strength, Appl. Phys. Lett. 82 (23) (2003) 4140–4142. [32] L.M. Ericson, H. Fan, H. Peng, V.A. Davis, W. Zhou, J. Sulpizio, Y. Wang, R. Booker, J.  Vavro, C.  Guthy, A.N.G.  Parra-Vasquez, M.J.  Kim, S.  Ramesh, R.K.  Saini, C. Kittrell, G. Lavin, H. Schmidt,W.W. Adams,W.E. Billups, M. Pasquali,W.-F. Hwang, R.H. Hauge, J.E. Fischer, R.E. Smalley, N. Macroscopic, Single-walled carbon nanotube fibers, Science 305 (5689) (2004) 1447–1450. [33] (a) K. Livanov, L. Yang, A. Nissenbaum, H.D. Wagner, Interphase tuning for stronger and tougher composites, Sci. Rep. 6 (2016) 26305. (b) Y.  Zhang, K.  Song, J.  Meng, M.L.  Minus, Tailoring polyacrylonitrile interfacial morphological structure by crystallization in the presence of single-wall carbon nanotubes, ACS Appl. Mater. Interfaces 5 (3) (2013) 807–814. (c) H.G. Chae, M.L. Minus, A. Rasheed, S. Kumar, Stabilization and carbonization of gel spun polyacrylonitrile/single wall carbon nanotube composite fibers, Polymer 48 (13) (2007) 3781–3789.



Carbon nanotube-reinforced polymer nanocomposite fibers

99

[34] (a) Y. Zare, Development of Halpin-Tsai model for polymer nanocomposites assuming interphase properties and nanofiller size, Polym. Test. 51 (2016) 69–73. (b) S. Chopra, K.A. Deshmukh, A.D. Deshmukh, C.L. Gogte, D. Peshwe, Prediction, evaluation and mechanism governing interphase strength in tensile fractured PA-6/ MWCNT nanocomposites, Compos. Pt. A-Appl. Sci. Manuf. 112 (2018) 255–262. [35] (a) B.A.  Newcomb, P.V.  Gulgunje, K.  Gupta, M.G.  Kamath, Y.  Liu, L.A.  Giannuzzi, H.G. Chae, S. Kumar, Processing, structure, and properties of gel spun PAN and PAN/CNT fibers and gel spun PAN based carbon fibers, Polym. Eng. Sci. 55 (11) (2015) 2603–2614. (b) B.A. Newcomb, L.A. Giannuzzi, K.M. Lyons, P.V. Gulgunje, K. Gupta,Y. Liu, M. Kamath, K. McDonald, J. Moon, B. Feng, High resolution transmission electron microscopy study on polyacrylonitrile/carbon nanotube based carbon fibers and the effect of structure development on the thermal and electrical conductivities, Carbon 93 (2015) 502–514. [36] Y.  Liu, S.  Kumar, Polymer/carbon nanotube nano composite fibers—a review, ACS Appl. Mater. Interfaces 6 (9) (2014) 6069–6087. [37] Y.Y. Zhang, N. Tajaddod, K.A. Song, M.L. Minus, Low temperature graphitization of interphase polyacrylonitrile (PAN), Carbon 91 (2015) 479–493. [38] J. Brent, P.K.P. Carey, L. Ci, G.G. Silva, P.M. Ajayan, Observation of dynamic strain hardening in polymer nanocomposites, ACS Nano 5 (4) (2011) 2715–2722. [39] P.S.  Owuor, C.S.  Tiwary, R.  Koizumi, M.  Soto, A.C.  Hart, E.V.  Barrera, R.  Vajtai, J. Lou, P.M. Ajayan, Self-stiffening behavior of reinforced carbon nanotubes spheres, Adv. Eng. Mater. 19 (5) (2017) 1600756. [40] E. Senses, P. Akcora, An interface-driven stiffening mechanism in polymer nanocomposites, Macromolecules 46 (5) (2013) 1868–1874. [41] L. Cao,Y. Wang, P. Dong, S. Vinod, J.T. Tijerina, P.M. Ajayan, Z. Xu, J. Lou, Interphase induced dynamic self-stiffening in graphene-based polydimethylsiloxane nanocomposites, Small 12 (27) (2016) 3723–3731. [42] Y. Li, P. Zhou, F. An,Y. Liu, C. Lu, Dynamic self-stiffening and structural evolutions of polyacrylonitrile/carbon nanotube nanocomposites, ACS Appl. Mater. Interfaces 9 (6) (2017) 5653–5659. [43] T.V. Sreekumar,T. Liu, B.G. Min, H. Guo, S. Kumar, R.H. Hauge, R.E. Smalley, Polyacrylonitrile single-walled carbon nanotube composite fibers, Adv. Mater. 16 (1) (2004) 58–61. [44] (a) N.  Minoo, et  al., Electrospun single-walled carbon nanotube/polyvinyl alcohol composite nanofibers: structure–property relationships, Nanotechnology 19 (30) (2008) 305702. (b) X. Zhang,T. Liu,T.V. Sreekumar, S. Kumar, X. Hu, K. Smith, Gel spinning of PVA/ SWNT composite fiber, Polymer 45 (26) (2004) 8801–8807. [45] K.W.  Putz, C.A.  Mitchell, R.  Krishnamoorti, P.F.  Green, Elastic modulus of ­single-walled carbon nanotube/poly(methyl methacrylate) nanocomposites, J. Polym. Sci. B Polym. Phys. 42 (12) (2004) 2286–2293. [46] J.-T. Yeh,Y.-C. Lai, H. Liu,Y.-C. Shu, C.-Y. Huang, K.-S. Huang, K.-N. Chen, Ultradrawing properties of ultrahigh-molecular-weight polyethylene/carbon nanotube fibers prepared at various formation temperatures, Polym. Int. 60 (1) (2011) 59–68. [47] R. Jain, Y.H. Choi, Y. Liu, M.L. Minus, H.G. Chae, S. Kumar, J.-B. Baek, Processing, structure and properties of poly(ether ketone) grafted few wall carbon nanotube composite fibers, Polymer 51 (17) (2010) 3940–3947. [48] H. Ma, J. Zeng, M.L. Realff, S. Kumar, D.A. Schiraldi, Processing, structure, and properties of fibers from polyester/carbon nanofiber composites, Compos. Sci. Technol. 63 (11) (2003) 1617–1628. [49] R.  Jain, M.L.  Minus, H.G.  Chae, S.  Kumar, Processing, structure, and properties of PAN/MWNT composite fibers, Macromol. Mater. Eng. 295 (8) (2010) 742–749. [50] V. Mottaghitalab, G.M. Spinks, G.G. Wallace, The influence of carbon nanotubes on mechanical and electrical properties of polyaniline fibers, Synth. Met. 152 (1–3) (2005) 77–80.

100

Carbon Nanotube Fibers and Yarns

[51] F. Du, J.E. Fischer, K.I. Winey, Coagulation method for preparing single-walled carbon nanotube/poly(methyl methacrylate) composites and their modulus, electrical conductivity, and thermal stability, J. Polym. Sci. B Polym. Phys. 41 (24) (2003) 3333–3338. [52] J. Wang, M. Miao, Z. Wang,W. Humphries, Q. Gu, A method of mobilizing and aligning carbon nanotubes and its use in gel spinning of composite fibres, Carbon 57 (2013) 217–226. [53] E. Choi, Enhancement of thermal and electrical properties of carbon nanotube polymer composites by magnetic field processing, J. Appl. Phys. 94 (9) (2003) 6034. [54] F. Du, J.E. Fischer, K.I. Winey, Effect of nanotube alignment on percolation conductivity in carbon nanotube/polymer composites, Phys. Rev. B 72 (12) (2005) 121404. [55] P. Pötschke, H. Brünig, A. Janke, D. Fischer, D. Jehnichen, Orientation of multiwalled carbon nanotubes in composites with polycarbonate by melt spinning, Polymer 46 (23) (2005) 10355–10363. [56] (a) Y. Bin, M. Mine, A. Koganemaru, X. Jiang, M. Matsuo, Morphology and mechanical and electrical properties of oriented PVA–VGCF and PVA–MWNT composites, Polymer 47 (4) (2006) 1308–1317. (b) S. Hooshmand, A. Soroudi, M. Skrifvars, E ­ lectro-conductive composite fibers by melt spinning of polypropylene/polyamide/carbon nanotubes, Synth. Met. 161 (15– 16) (2011) 1731–1737. [57] Y.  Mamunya, A.  Boudenne, N.  Lebovka, L.  Ibos, Y.  Candau, M.  Lisunova, Electrical and thermophysical behaviour of PVC-MWCNT nanocomposites, Compos. Sci. Technol. 68 (9) (2008) 1981–1988. [58] (a) A.N. Mallya, P.C. Ramamurthy, Design and fabrication of a highly stable polymer carbon nanotube nanocomposite chemiresistive sensor for nitrate ion detection in water, ECS J. Solid State Sci. Technol. 7, (7) (2018) Q3054–Q3064. (b) N. Hu,Y. Karube, H. Fukunaga, A Strain Sensor from a Polymer/Carbon Nanotube Nanocomposite, vol. 19, 2010, pp. 77–86. (c)B. Pradhan, K. Setyowati, H. Liu, D.H. Waldeck, J. Chen, Carbon nanotube—polymer nanocomposite infrared sensor, Nano Lett., 8 (4) (2008) 1142–1146. (d)N. Hu, Y. Karube, C. Yan, Z. Masuda, H. Fukunaga, Tunneling effect in a polymer/carbon nanotube nanocomposite strain sensor, Acta Mater. 56 (13) (2008) 2929–2936. [59] S. Berber, Y.-K. Kwon, D. Tománek, Unusually high thermal conductivity of carbon nanotubes, Phys. Rev. Lett. 84 (20) (2000) 4613. [60] P. Kim, L. Shi, A. Majumdar, P.L. McEuen, Thermal transport measurements of individual multiwalled nanotubes, Phys. Rev. Lett. 87 (21) (2001) 215502. [61] Z. Han, A. Fina, Thermal conductivity of carbon nanotubes and their polymer nanocomposites: a review, Prog. Polym. Sci. 36 (7) (2011) 914–944. [62] (a) W.  Li, J.  Hao, P.  Zhou, Y.  Liu, C.  Lu, Z.  Zhang, Solvent-solubility-parameter-­ dependent homogeneity and sol–gel transitions of concentrated polyacrylonitrile solutions, J. Appl. Polym. Sci. 134 (41) (2017). (b) J. Hao, F. An, Y. Yu, P. Zhou, Y. Liu, C. Lu, Effect of coagulation conditions on solvent diffusions and the structures and tensile properties of solution spun polyacrylonitrile fibers, J. Appl. Polym. Sci. 134 (5) (2017) 44390. [63] H.G. Chae, B.A. Newcomb, P.V. Gulgunje,Y. Liu, K.K. Gupta, M.G. Kamath, K.M. Lyons, S. Ghoshal, C. Pramanik, L. Giannuzzi, High strength and high modulus carbon fibers, Carbon 93 (2015) 81–87. [64] J. Lee, J.I. Choi, A.E. Cho, S. Kumar, S.S. Jang,Y.H. Kim, Origin and control of polyacrylonitrile alignments on carbon nanotubes and graphene nanoribbons, Adv. Funct. Mater. (15) (2018) 28. [65] Y. Liu, H.G. Chae, S. Kumar, Gel-spun carbon nanotubes/polyacrylonitrile composite fibers. Part II: stabilization reaction kinetics and effect of gas environment, Carbon 49 (13) (2011) 4477–4486. [66] (a) N. Tajaddod, H. Li, M.L. Minus, Low-temperature graphitic formation promoted by confined interphase structures in polyacrylonitrile/carbon nanotube materials, Polymer 137 (2018) 346–357.



Carbon nanotube-reinforced polymer nanocomposite fibers

101

(b) Y. Liu, H.G. Chae, S. Kumar, Gel-spun carbon nanotubes/polyacrylonitrile composite fibers. Part III: effect of stabilization conditions on carbon fiber properties, Carbon 49 (13) (2011) 4487–4496. [67] (a) P. Sabina, et al., Carbonization of electrospun poly(acrylonitrile) nanofibers containing multiwalled carbon nanotubes observed by transmission electron microscope with in situ heating, J. Polym. Sci. B Polym. Phys. 48 (20) (2010) 2121–2128. (b) P. Sabina, Z. Eyal, C. Yachin, The effect of embedded carbon nanotubes on the morphological evolution during the carbonization of poly(acrylonitrile) nanofibers, Nanotechnology 19 (16) (2008) 165603. [68] (a) Y. Lv, C. Peng, Chemically grafting carbon nanotubes onto carbon fibers for enhancing interfacial strength in carbon fiber/HDPE composites, Surf. Interface Anal. 50 (5) (2018) 552–557. (b) H. Cui, Z. Jin, D. Zheng, W. Tang,Y. Li,Y. Yun, T.Y. Lo, F. Xing, Effect of carbon fibers grafted with carbon nanotubes on mechanical properties of cement-based composites, Constr. Build. Mater. 181 (2018) 713–720. (c) H.S.  Bedi, S.S.  Padhee, P.K.  Agnihotri, Effect of carbon nanotube grafting on the wettability and average mechanical properties of carbon fiber/polymer multiscale composites, Polym. Compos. 39 (2018) E1184–E1195. (d) L.  Zhang, N.  De Greef, G.  Kalinka, B.  Van Bilzen, J.-P.  Locquet, I.  Verpoest, J.W. Seo, Carbon nanotube-grafted carbon fiber polymer composites: damage characterization on the micro-scale, Compos. B Eng. 126 (2017) 202–210. (e) L. Feng, K.-Z. Li, J.-H. Lu, L.-H. Qi, Effect of growth temperature on carbon nanotube grafting morphology and mechanical behavior of carbon fibers and carbon/ carbon composites, J. Mater. Sci. Technol. 33 (1) (2017) 65–70. (f) M.S. Islam,Y. Deng, L. Tong, S.N. Faisal, A.K. Roy, A.I. Minett,V.G. Gomes, Grafting carbon nanotubes directly onto carbon fibers for superior mechanical stability: towards next generation aerospace composites and energy storage applications, Carbon 96 (2016) 701–710. (g) W. Fan,Y. Wang, C. Wang, J. Chen, Q. Wang,Y. Yuan, F. Niu, High efficient preparation of carbon nanotube-grafted carbon fibers with the improved tensile strength, Appl. Surf. Sci. 364 (2016) 539–551. . (h) L. Feng, K.-z. Li, Z.-s. Si, Q. Song, H.-j. Li, J.-h. Lu, L.-j. Guo, Compressive and interlaminar shear properties of carbon/carbon composite laminates reinforced with carbon ­nanotube-grafted carbon fibers produced by injection chemical vapor deposition, Mater. Sci. Eng. A 626 (2015) 449–457. (i) N. Subramanian, B. Koo, K.R. Venkatesan, A. Chattopadhyay, Interface mechanics of carbon fibers with radially-grown carbon nanotubes, Carbon 134 (2018) 123–133. [69] M.L. Minus, H.G. Chae, S. Kumar, Observations on solution crystallization of poly(vinyl alcohol) in the presence of single-wall carbon nanotubes, Macromol. Rapid Commun. 31 (3) (2010) 310–316. [70] P. Miaudet, S. Badaire, M. Maugey, A. Derre,V. Pichot, P. Launois, P. Poulin, C. Zakri, Hot-drawing of single and multiwall carbon nanotube fibers for high toughness and alignment, Nano Lett. 5 (11) (2005) 2212–2215. [71] X.  Xu, A.J.  Uddin, K.  Aoki, Y.  Gotoh, T.  Saito, M.  Yumura, Fabrication of high strength PVA/SWCNT composite fibers by gel spinning, Carbon 48 (7) (2010) 1977–1984. [72] L.L. Lu, W.J. Hou, J. Sun, J.J. Wang, C.X. Qin, L.X. Dai, Preparation of poly (vinyl alcohol) fibers strengthened using multiwalled carbon nanotubes functionalized with tea polyphenols, J. Mater. Sci. 49 (9) (2014) 3322–3330. [73] Y.Z. Wei, D.P. Lai, L.M. Zou, X.L. Ling, H.W. Lu, Y.J. Xu, Facile fabrication of PVA composite fibers with high fraction of multiwalled carbon nanotubes by gel spinning, Polym. Eng. Sci. 58 (1) (2018) 37–45. [74] J.S.  Meng, Y.Y.  Zhang, K.N.  Song, M.L.  Minus, Forming crystalline polymer-nano interphase structures for high-modulus and high-tensile/strength composite fibers, Macromol. Mater. Eng. 299 (2) (2014) 144–153.

102

Carbon Nanotube Fibers and Yarns

[75] (a) M. Afshari, D.J. Sikkema, K. Lee, M. Bogle, High performance fibers based on rigid and flexible polymers, Polym. Rev. 48 (2) (2008) 230–274. (b) H.G.  Chae, S.  Kumar, Rigid-rod polymeric fibers, J. Appl. Polym. Sci. 100 (1) (2006) 791–802. [76] (a) J.P.F. Lagerwall, G. Scalia, Carbon nanotubes in liquid crystals, J. Mater. Chem. 18 (25) (2008) 2890–2898. (b) M.D. Lynch, D.L. Patrick, Organizing carbon nanotubes with liquid crystals, Nano Lett. 2 (11) (2002) 1197–1201. [77] S. Kumar, T.D. Dang, F.E. Arnold, A.R. Bhattacharyya, B.G. Min, X. Zhang, R.A. Vaia, C. Park,W.W. Adams, R.H. Hauge, R.E. Smalley, S. Ramesh, P.A. Willis, Synthesis, structure, and properties of PBO/SWNT composites, Macromolecules 35 (24) (2002) 9039–9043. [78] K. Hun-Sik, M. Seung Jun, J. Rira, J. Hyoung-Joon, Preparation and characterization of poly(p-phenylene terephthalamide)/multiwalled carbon nanotube composites via in-situ polymerization, Mol. Cryst. Liquid Cryst. 492 (1) (2008) 20–27. [79] T. Sainsbury, K. Erickson, D. Okawa, C.S. Zonte, J.M.J. Fréchet, A. Zettl, Kevlar functionalized carbon nanotubes for next-generation composites, Chem. Mater. 22 (6) (2010) 2164–2171. [80] C. Yutong, L. Zhaofeng, G. Xianghua,Y. Junrong, H. Zuming, L. Ziqi, Dynamic rheological studies of poly(p-phenyleneterephthalamide) and carbon nanotube blends in sulfuric acid, Int. J. Mol. Sci. 11 (4) (2010) 1352–1364. [81] L. Deng, R.J. Young, S. van der Zwaag, S. Picken, Characterization of the adhesion of single-walled carbon nanotubes in poly(p-phenylene terephthalamide) composite fibres, Polymer 51 (9) (2010) 2033–2039. [82] I.  O’Connor, H.  Hayden, J.N.  Coleman, Y.K.  Gun’ko, High-strength, high-­ toughness composite fibers by swelling Kevlar in nanotube suspensions, Small 5 (4) (2009) 466–469. [83] (a) S. Sharma, A.K. Pathak, V.N. Singh, S. Teotia, S.R. Dhakate, B.P. Singh, Excellent mechanical properties of long multiwalled carbon nanotube bridged Kevlar fabric, Carbon 137 (2018) 104–117. (b) E.D.  LaBarre, X.  Calderon-Colon, M.  Morris, J.  Tiffany, E.  Wetzel, A.  Merkle, M. Trexler, Effect of a carbon nanotube coating on friction and impact performance of Kevlar, J. Mater. Sci. 50 (16) (2015) 5431–5442. [84] C.H. Zhang, H.F. Xu, Z.X. Jiang, F.L. Zhu, Y.D. Huang, Carbon nanotubes grafting PBO fiber: a study on the interfacial properties of epoxy composites, Polym. Compos. 33 (6) (2012) 927–932. [85] X. Yan, H. Dong,Y. Liu, B.A. Newcomb, H.G. Chae, S. Kumar, Z. Xiao, T. Liu, Effect of processing conditions on the dispersion of carbon nanotubes in polyacrylonitrile solutions, J. Appl. Polym. Sci. 132 (26) (2015) 42177. [86] C. Pramanik, J.R. Gissinger, S. Kumar, H. Heinz, Carbon nanotube dispersion in solvents and polymer solutions: mechanisms, assembly, and preferences, ACS Nano 11 (12) (2017) 12805–12816. [87] T.  Liu, S.  Kumar, Effect of orientation on the modulus of SWNT films and fibers, Nano Lett. 3 (5) (2003) 647–650. [88] P.H.  Wang, S.  Sarkar, P.  Gulgunje, N.  Verghese, S.  Kumar, Fracture mechanism of high impact strength polypropylene containing carbon nanotubes, Polymer 151 (2018) 287–298.

CHAPTER 6

Post-spinning treatments to carbon nanotube fibers Hai Minh Duong, Sandar Myo Myint, Thang Quyet Tran, Duyen Khac Le National University of Singapore, Singapore, Singapore

Aligned CNT fibers can be fabricated by three methods: spinning from CNT solution [1,2], spinning from CNT arrays [3–5], and direct spinning via floating catalyst method [6–12].The CNT fibers spun from the first two methods are relatively clean, whereas the direct-spun CNT fibers contain many impurities of catalyst and amorphous carbon due to their single-step fabrication process [6, 13]. These residual impurities lower the fiber performance and limit their applications. Although the as-spun CNT fibers possess excellent mechanical and electrical properties, many studies have been conducted to further improve their properties by different posttreatments [3, 14–21], such as densification treatments. Densification treatments can be classified into indirect methods (such as twisting [22,23], liquid densification [23], and drawing through dies [24]) and direct methods (such as rubbing [25] and pressurized rolling [26]). The indirect approaches are limited by their low densifying forces [22–24], whereas the direct approaches are more effective as higher densifying forces can be applied directly to the CNT fibers, resulting in much denser CNT fiber structures [26]. CNT fiber properties can also be enhanced by polymer infiltration, in which the formation of cross-links between the CNT bundles through the infiltration process can effectively improve the inter-tube load transfer efficiency of the CNT fibers, resulting in their better mechanical performance [22, 27, 28]. Although many studies have investigated the effects of each posttreatment on the properties of CNT fibers, few researchers have reported their combined effects [28–31]. Since each treatment has specific positive effects on the properties of CNT fibers, the combination of two posttreatment methods may be expected to improve the fiber performance more

Carbon Nanotube Fibers and Yarns https://doi.org/10.1016/B978-0-08-102722-6.00006-7

Copyright © 2020 Elsevier Ltd. All rights reserved.

103

104

Carbon Nanotube Fibers and Yarns

significantly. This chapter reviews the effects of several posttreatments on the mechanical performance of the CNT fibers.

6.1  Twist insertion The first CNT fiber was successfully prepared through spinning a CNT homogeneous dispersion into a polyvinyl alcohol (PVA) coagulation bath [32]. This approach was modified by Baughman’s group to make single-walled CNT composite fibers with very high strength [33, 34]. The major issues with this approach include a relatively high fraction of remaining polymer volume and short individual CNTs, which limits the fiber strength, and electrical and thermal conductivity [35]. The load transfer between CNTs in a fiber depends on both the contact areas and the inter-tube spaces between the CNTs [17]. Twisting is an effective posttreatment to densify the CNT fibers by reducing the inter-tube spaces. At high twist angle, the CNTs in the fibers are in closer contact with each other, enhancing the van der Waals force and friction between the bundles and, hence, increasing the fiber strength. Zhang et al. [20] investigated the effect of fiber twist on the mechanical properties of the CNT fibers spun from 650-μm-long arrays. CNT fibers were spun from a CNT array using a spindle made of a microprobe. In the post-spin twisting process, a proper weight was hung on one end of a fiber to provide tension in the axial direction, while the other end of the fiber was attached to a rotator. The extent of post-spin twisting depends on the twisting speed and duration. In their study, a CNT fiber 5 cm long was typically twisted at a rotation rate of 500 rpm for 2 min, which gave 20,000 turns/m of twist. The tensile strength of the CNT fibers significantly increased from 0.85 to 1.9 GPa after post-spin twisting. Moreover, the reduction in the diameter of the fibers from 4 to 3 μm after twisting evidenced the densification effect of the treatment. However, research conducted by Zhao et  al. showed that twist also had a negative contribution to the fiber strength if fiber was over-twisted [3, 21]. This could be explained by the fact that at a higher twisting angle, the fibers are more misaligned with the fiber axis, therefore lowering the fiber strength, although twisting reduces the inter-tube spaces and minimizes the contact resistance between the CNTs [15, 18]. Miao reported that the CNT fibers with higher density (i.e., lower fiber porosity) could be achieved by increasing their twisting angle [18]. Interestingly, he found that the electrical conductivity of the CNT fibers decreased with



4.0E+04

8.0E–05

3.5E+04

7.0E–05

3.0E+04

6.0E–05

2.5E+04

5.0E–05

2.0E+04

4.0E–05

1.5E+04

3.0E–05

1.0E+04

2.0E–05 Conductivity

5.0E+03 0.0E+00 0.5

0.55

0.6

(A)

Resistivity

0.65 0.7 Yarn porosity

105

Resistivity (Ωm)

Conductivity (S/m)

Post-spinning treatments to carbon nanotube fibers

1.0E–05

0.75

0.0E+00 0.8

Specific conductivity (S·cm2·g–1)

600 500 400 300 200 100 0 0.5

0.6

(B)

0.7

0.8

0.9

1

CNT yarn porosity

Fig. 6.1  (A) Effect of fiber porosity on the electrical conductivity and resistivity of pure CNT fibers; (B) effect of yarn porosity on the specific electrical conductivity of CNT fibers. (Reproduced with permission from M. Miao, Electrical conductivity of pure carbon nanotube yarns, Carbon 49 (2011) 3755–3761.)

increasing fiber porosity (Fig. 6.1A) but their specific conductivity was almost independent of the fiber porosity (Fig. 6.1B).

6.2  Liquid densification As-spun CNT fibers can also be densified by liquid densification [23]. In this method, a liquid is absorbed by the CNT fibers and then it is evaporated. Due to the surface tension of the solvents, the fibers are densified as their diameter reduces. Liu et al. [23] reported a simple and continuous

106

Carbon Nanotube Fibers and Yarns

spinning method that combines twisting and shrinking processes to produce CNT yarns. In their method, a yarn freshly spun from a super-aligned CNT array is first twisted and then passes through a volatile solvent for shrinking. The as-produced yarn consists of densely packed CNTs and thus has a tensile strength up to about 1 GPa. The tensile strength depends on the diameter and the twisting angle of the yarn. Different kinds of solvents, such as water, ethanol, and acetone, are used to shrink the twisted yarns, and acetone shows the best shrinking effect. They compared the mechanical properties of the CNT fibers spun from vertically aligned CNT arrays before and after liquid densification treatment. Their stressstrain curves show that the maximum strain of the fibers was unchanged while their strength and Young’s modulus enhanced significantly after the application of liquid densification. The fiber diameters decreased from 11.5 to 9.7 μm after the treatments, indicating a denser structure of the fibers obtained [23]. The percentage of load increase and diameter reduction of the CNT fibers after acetone shrinking ranged from 15% to 40% and from 15% to 24%, respectively.The method is suitable for continuous mass production of high strength CNT yarns with a wide range of diameters, especially ultrathin yarns.

6.3  Coating and doping The goal of CNT fiber production and treatment is to translate the excellent properties of individual CNTs to large CNT assemblies. Among these, a macroscopic cable that would replace metals as a universal conductive wire would have large volume applications in electricity transmission, aerospace, and automobile industry. Several methods have been developed for creating multi-walled, double-walled, and single-walled CNT to generate nanotube fibers with good mechanical properties. The electrical resistivities reported are over a large range between 7.1 × 10−3 and 2 × 10−6 Ω m. The CNT resistivity values reported up to now are 2–3 orders higher than that for oxygen-free Cu (1.68 × 10−8 Ω m), one of the most conducting metals widely used in current carrying applications. The electrical conductivity of the CNT fibers can be significantly increased by nanoparticle coating or doping [36]. Randeniya et al. [36] compared the electrical performance of the pure CNT fibers and the fibers doped with metal nanoparticles, including Cu, Au, Pd, and Pt. They reported that the CNT fibers doped with Cu and Au possessed an electrical conductivity of up to 3 × 102 S/cm at room temperature, which was



Post-spinning treatments to carbon nanotube fibers

107

600 times higher than that of the pure CNT fibers. Similarly, the Pd- and Pt-coated CNT fibers showed high electrical conductivity of 2 × 104 and 5 × 103 S/cm, respectively. Instead of coating nanoparticles on the surface of the CNT fibers, Zhao et al. [37] introduced iodine atoms into the structure of double-walled CNT fibers (DWNT); the diameters were in the range of 2–3 nm. They reported the fabrication and doping of carbon nanotube cables with resistivity one order of magnitude closer to the resistivity of Cu than the predecessors. The superior conductivity was achieved by a synergistic effect between the unique structure of the CNTs and the rational design of processing and doping. The nanotubes aligned in one direction and the nanotube bundles were interconnected and formed into a continuous network.The small bundles and the nanotubes by themselves were several micrometers long. After the nanotubes were fabricated into the cable, the natural alignment of the nanotube “stocking” was retained, which turned out to be beneficial for the conductivity of the cable. Interestingly, the electrical conductivity of the CNT fibers could reach up to 6.7 × 108 S/cm. Owing to the low density of the fibers (0.33 g/cm3), their specific conductivity was higher than that of copper and aluminum.

6.4  Acid treatment Impurities in the CNT fibers can be reduced by optimizing their synthesis process [6, 38, 39] and/or adopting purification treatments [29–31, 40, 41]. Due to their simplicity, cost efficiency, and scalability, oxidative purification in liquid phase or gas phase has been widely employed as a powerful posttreatment for purifying CNTs [42]. While the liquid-phase oxidation uses a base or acid solution, oxidation in gas phase uses air, oxygen, or other gases, and might be followed by acid leaching [43]. Although successful purification of CNT fibers by liquid-phase oxidation methods has been widely reported [29–31], studies on purifying CNT fibers by gas-phase oxidation is very limited, especially for CNT fibers spun by the floating catalyst method.

6.4.1  Purification of CNT fibers The dissolution of CNTs in chlorosulfonic acid (HSO3Cl), the true solvent for CNTs [44], has drawn great attention owing to their ability to obtain well-controlled CNT morphologies [44] for production of different macroscopic CNT assemblies [1, 45]. While many studies have been conducted

108

Carbon Nanotube Fibers and Yarns

on the dissolution of CNT powders in superacid [1, 44, 45], few works on the dissolution of the direct-spun CNT fibers have been reported. An oxidative purification approach using air, followed by HCl washing, is applied to purifying CNT fibers spun by the floating catalyst method. Fig. 6.2 shows TEM, TGA, and Raman results of the as-spun and purified CNT fibers produced by the floating catalyst method. Only CNT fibers spun from toluene source are used for the purification due to their high purity. As can be seen in Fig. 6.2A, the as-spun CNT fibers mainly consist of double-walled nanotubes (DWNTs) with an average diameter of 5.5 ± 0.4 nm. Iron impurities encapsulated by graphitic layers can be observed in the fiber structure (Fig. 6.2B). After applying the purification treatment, almost no iron impurities are observed in the CNT structure shown in Fig. 6.2C, suggesting that the CNT fibers are much cleaner. In TGA analysis of CNTs, the weight loss below 400°C usually corresponds to the removal of amorphous and disordered carbon from the sample [31, 46]. Generally speaking, CNTs start to be oxidized at above 400°C. The TGA result shown in Fig. 6.2E indicates that carbonaceous impurities in the purified CNT fibers are below 1% and the purification treatment reduces their iron impurities from 11.9% to 4.2%.The results are supported by a remarkable temperature upshift (50°C) of the CNT oxidative decomposition after the treatment (Fig.  6.2E). Due to the presence of the iron impurities, the activation energy of the CNTs is lowered and their parasitic oxidation is catalyzed [46]. Therefore, the reduction of the iron impurities improves the thermal stability of the CNT fibers after the purification treatment. These findings are in good agreement with the enhanced thermal stability of the purified CNT films reported by Lin et al. [43]. The findings suggest that the purification treatment using air and HCl leaching can effectively reduce the iron and carbonaceous impurities from the direct-spun CNT fibers. The iron impurities in the as-spun CNT fibers could not be removed by simple acidic washing since carbon shells encapsulate them. However, these impurities catalyze the oxidation of the protective carbon shells, resulting in their lower oxidation resistance [42, 43]. As the oxidation resistances between the CNTs, amorphous carbon, and multishell carbon nanocapsules are different, the catalyst impurities are exposed and can be dissolved by HCl. Therefore, the purification treatment reduces the impurity content in the CNT fibers. However, due to the highly packed structures of the CNT fibers, not all CNTs are exposed to air for oxidation during the purification process. Therefore, a small amount of catalyst impurities are still left in the fiber structure [46].



Post-spinning treatments to carbon nanotube fibers

(A)

109

(B) Iron impurities

10 nm

40 nm

(C)

(D)

Damaged CNT 5 nm

40 nm

(E)

(F)

100

As-spun fiber Purified fiber

90

G

80 60

50C

50 40 30

2D

D

Intensity (a.u.)

TGA (%)

70

As-spun fiber Purified fiber

ID/IG = 0.56

20 ID/IG–0.17

10 0

0

100 200 300 400 500 600 700 800 900 1000 Temperature (C)

0

1000 2000 Wave number (cm–1)

3000

Fig. 6.2  (A, B) TEM images of as-spun CNT fibers; (C) TEM images of purified CNT fibers; (D) damaged CNT due to purification, (E) TGA results, and (F) Raman results of the asspun and purified CNT fibers.

The intensity ratio (ID/IG) is used to evaluate the defect present on the CNTs [46]. The increase in the ID/IG ratio of the CNT fibers shown in Fig. 6.2F from 0.17 to 0.56 after the purification process suggests that the treatment introduced defects into the CNT structures. The defect increase in the purified CNT structures might stem from the fact that the CNTs themselves are attacked at their defective sites during the oxidation

110

Carbon Nanotube Fibers and Yarns

although they have better oxidation resistance than the amorphous carbon [46]. The purification treatment might also decrease the number of CNTs and left surviving CNTs with damage (Fig.  6.2D), resulting in a higher defect density [42]. These findings are consistent with the results of the increasing defects on the structure of CNT films [43] and CNT powders [46] purified by air oxidation treatment. Therefore, the oxidative treatment should be carefully designed and optimized to balance the impurity removal requirements, CNT structure preservation, and enhanced performance of CNT fibers for desired applications. The CNT length distribution was estimated by using measured isotropic cloud point, CNT diameter measured by TEM and viscosityaverage aspect ratio under the assumption of a log-normal distribution [47]. The average CNT length was estimated to be 2 μm and about 90% of the CNTs were shorter than 6 μm. The results were in good agreement with the aforementioned formation of the CNT tactoids and the required length of the short F-Actin tactoids reported by Oakes et  al. [48]. The length of the CNTs constituting the fibers was still longer than that of many commercial SWNTs and DWNTs such as HiPco 183.6 (1.51 μm), HiPco 188.3 (0.29 μm), UniDym OE (1.92 μm), and SWeNT CG300 (0.71 μm) [47]. The CNTs were likely shortened during purification treatment by oxidation from their ends. Therefore, the average CNT length in the as-spun fibers might be longer than 2 μm. Furthermore, the CNT length and CNT aspect ratio (length/diameter ratio) of the fibers can be improved by optimizing the oxidative purification procedure [42, 43] or improving the synthesis process by controlling the iron catalyst size to reduce the CNT diameters or employing different carbon sources and synthesis temperatures to enhance CNT growth [6, 7, 49].

6.4.2  Effects of acidization on mechanical properties of CNT fibers The tensile performance of CNT fibers can be improved by CNT surface modification through acid treatment [50]. Meng et al. immersed CNT fibers spun from CNT arrays in HNO3 (16 M) for several hours to modify the CNT surface by introducing various functional groups, including hydroxyl (–OH), methyl (–CH3)/methylene (–CH2–), and carbonyl (–C=O) [50]. The surface modification improved the load transfer between the CNTs by enhancing the inter-tube interaction and interfacial shear property. The changes were reflected in the fiber’s mechanical property with 50% increase



Post-spinning treatments to carbon nanotube fibers

111

in strength after a 2-h treatment. However, too long acidization time could destroy the crystalline structure of the CNTs, leading to a decrease in the fiber strength. Fig. 6.3 compares the surface morphologies, and mechanical and electrical properties of the as-spun and purified CNT fibers. The as-spun CNT fibers in Fig.  6.3A and B consisted of aligned CNT bundles along the fiber axis and their diameters were about 8 ± 0.2 μm. Due to the impurity removal after the purification treatment, the fiber diameter was reduced by nearly 19%, reaching 6.5 ± 0.18 μm (Fig.  6.3C). The good alignment of CNT bundles observed in the structure of the purified CNT fibers (Fig.  6.3D) suggests that the purification treatment preserved the fiber structure. Additionally, the larger bundle size observed in Fig. 6.3D suggests the positive effects of the purification treatment on the impurity removal, leading to stronger van der Waals interactions between the CNTs and CNT

(A)

10 µm

(B)

2 µm

(C)

10 µm

(D)

2 µm

0.4

4679 Purified CNT fiber As-spun CNT fiber

0.3 Strength (GPa)

As-spun fibers Purified fibers

0.2

20.66

8.41

0.1 0.32 0.0 0.00

(E)

2369

0.04

0.08 Strain

0.12

0.16

(F)

0.38

Strength (GPa)

Young’s modulus (GPa)

Electrical conductivity (S/cm)

Fig. 6.3  SEM images of (A, B) as-spun CNT fiber and (C, D) purified CNT fiber, (E) stressstrain curves, and (F) mechanical and electrical properties of the as-spun and purified CNT fibers.

112

Carbon Nanotube Fibers and Yarns

bundles. Due to the volume shrinkage after the impurity removal, the CNT fibers were strongly condensed inward, leading to a slightly wrinkled surface morphology as shown in Fig. 6.3D. The as-spun CNT fibers had an average tensile strength and Young’s modulus of 0.32 and 8.41 GPa, respectively (Fig. 6.3E and F). Although the CNTs may be damaged due to defect introduction during the purification process, the significant reduction of the CNT fiber diameter after purification contributed positively to the properties of the fibers, resulting in their improved mechanical and electrical performance. Since the cross-sectional area was reducted by a factor of 1.5 after purification, both mechanical and electrical properties of the CNT fiber are expected to change by the same ratio. In fact, the strength and Young’s modulus of the purified CNT fibers were 0.38 and 20.66 GPa, corresponding to 119% and 246% of those of the as-spun CNT fibers, respectively. However, the maximum tensile load before fiber breakage decreased slightly from 1.6 to 1.2 cN due to the reduction of fiber diameter. More importantly, the CNT fibers exhibited a nearly two-fold increase in electrical conductivity after the purification treatment, from about 2400 to 4600 S/cm. The impressive enhancements in the properties of CNT fibers might be explained by the improved inter-tube interactions and increased CNT contact areas from the denser and cleaner structure of the CNT fibers after the purification [30]. Therefore, effective purification of the CNT fibers can result in significant improvements in the electrical and mechanical properties. These findings are in good agreement with the improvements of morphology and properties of CNT films and twisted CNT fibers purified by air oxidation methods [43, 51]. The as-spun CNT fibers produced by the floating catalyst method were immersed in concentrated HNO3 for purification. Fig.  6.4 shows SEM images, diameters, and mechanical properties of the CNT fibers before and after the acidization. As can be seen, the diameter of the acidized CNT fibers became slightly smaller than that of the as-spun fibers (Fig. 6.4B). This diameter reduction was due to the densification effects of the acid treatment on the CNT fibers [50, 52]. Additionally, the CNT bundle size increased and the inter-tube space reduced after the treatment, suggesting that the acidized CNT fibers possessed slightly denser structures and better CNT alignment than those of the as-spun CNT fibers. Fig. 6.4D presents the tensile strength and Young’s modulus of the CNT fibers at different acidization times. Acidization could considerably improve the mechanical performance of the CNT fibers. The as-spun CNT



Post-spinning treatments to carbon nanotube fibers

113

Fig. 6.4  (A) SEM images of the 15-min acidized CNT fibers, (B) diameter of the CNT fibers before and after acid treatment, (C) surface morphologies of the CNT fibers before and after acidization, and (D) effect of acid treatment time on the mechanical properties of the CNT fibers.

fibers had average tensile strength and Young’s modulus of 0.41 ± 0.05 and 14.6 ± 1.5 GPa, respectively. Their mechanical properties increased significantly to a strength of 0.73 ± 0.07 GPa and Young’s modulus of 26.12 ± 4.32 GPa after a 15-min treatment, corresponding to 178% and 179% those of their as-spun counterparts. However, prolonged treatment beyond 15 min degraded the mechanical properties of the CNT fibers.The CNT fibers acidized at 120 min showed strength and Young’s modulus as low as 0.48 ± 0.06 GPa and 24.12 ± 3.24 GPa, respectively. TEM images and Raman results of the as-spun and the 15-min acidized CNT fibers are compared in Fig. 6.5. In Fig. 6.5C, iron catalysts can be observed on the surface of the acidized CNTs, suggesting that the treatment did not remove the iron impurities of the CNT fibers due to the protective carbon layers outside the catalysts and the short treatment time [30, 31]. While the CNT surface of the as-spun fibers was attached with many carbonaceous impurities, the acidized CNTs were thinner with a cleaner surface due to the purification effect of the acid treatment on the CNT structures (Fig. 6.5A and B). These findings are supported

114

(A)

Carbon Nanotube Fibers and Yarns

(B)

Amorphous carbon

Less amorphous carbon

10 nm

(C)

(D) Iron impurities

Fig. 6.5  TEM images of (A) the as-spun CNT fibers, (B, C) 15-min acidized CNT fibers, and (D) Raman result of the as-spun and 15-min acidized CNT fibers.

by the Raman results shown in Fig. 6.5D. The CNT fibers experienced a slight increase in IG/ID ratio from 2.34 to 3.25 due to the acid treatment, indicating that the acidized CNTs became less defective. These findings are in good agreement with the purification effects of concentrated HNO3 on rolled CNT fibers and as-spun CNT films reported by Liu et al. [29]. The improved mechanical properties of the acidized CNT fibers can be explained by the purification effects of the acidization. As amorphous carbon impurities are removed by the acid treatment, stronger van der Waals interactions between clean CNT bundles are obtained, resulting in the enhanced stacking of the CNT bundles in the acidized CNT fibers [30, 31]. Since the acid treatment increases the CNT bundle sizes and inter-CNT contacts and interactions, better load transfer efficiency and enhanced mechanical performance of the CNT fibers are achieved. The results are in good agreement with the improved mechanical properties of the CNT fibers spun from CNT arrays by acidization process using a mixture of nitric and sulfuric acids [52].



Post-spinning treatments to carbon nanotube fibers

115

The purification effect of the acid treatment is consistent with the reduction of the mass density of the CNT fibers after the treatment. Specifically, the mass density of the 15-min acidized fibers was about 1.53 g/cm3, 15% lower than that of the as-spun fibers (1.80 g/cm3). Since the amorphous carbon impurities are reactive and invariably exist in the as-spun CNT fibers, the acid treatment may oxidize them to form oxidized debris before they are removed by subsequent aqueous washing [53, 54]. However, prolonged treatment beyond 15 min cannot further improve the mechanical performance of the CNT fibers, as presented in Fig. 6.4D. This result can be explained by the fact that the CNT structures are possibly damaged at the prolonged treatment [50, 55] although denser fiber structure can be obtained. The competition between improved interfacial shearing strength and damaged CNT structures would finally determine the mechanical performance of the CNT fibers. Therefore, the acidization parameters, such as acid concentration and treatment time need to be carefully controlled to optimize the acidization process.

6.5  Mechanical densification and stretching Among all the densification methods, mechanical densification is one of the best approaches to produce highly dense CNT structures since the densifying forces are applied directly to the fibers [26, 28]. Several traditional textile twistless methods based on mechanical densification have been used to densify the CNT fibers [25, 26]. Miao [25] used a rubbing roller system (Fig. 6.6A) to densify the CNT web drawn from a vertically aligned CNT array into compact twistless fibers. The resulting densified fibers consisted of a high packing density sheath with CNTs lying straight and parallel to the fiber axis and a low density core with microscopic voids. Due to the improved contact length between CNTs in the dense sheath, the coresheath structured, twistless CNT fibers exhibited high specific modulus up to 59 N/tex and specific strength up to 75 cN/tex. Wang et al. [26] used a pressurized rolling system to densify the CNT fibers spun by the floating catalyst method to a high density structure (Fig. 6.6B).With a densification factor up to 10, their densified fibers showed significant improvement in strength from 0.3 to 4.3 GPa. This result was the highest strength reported for the CNT fibers at the gage length of 10 mm in literature while their Young’s modulus remained at 90 GPa, as shown in Fig. 6.6C. This impressive enhancement was due to improved load transfer between the nanotubes and bundles after the treatments. Badaire et al. [14]

116

Carbon Nanotube Fibers and Yarns

Press system

vs v r

v

Fiber-supplying system

(A)

(B)

5

Rolled CNT fiber

As prepared CNT fiber

0.03

10 mm gauge length 3.76–5.53 GPa; Average 4.34 GPa

4

Resistivity (Ω*cm)

Tensile strength (GPa)

6

Fiber-collecting system

3 2 1

0.02

0.01

0

(C)

0

2

4

6

8

Strain (%)

10

12

14

0.00

(D)

0

20

40

60

80

100

Stretch %

Fig.  6.6  (A) Schematic diagram of experimental setup for the fabrication of twistless CNT fibers by a rubbing roller system, (B) schematic diagram of the experimental setup for the densification of high-density CNT fibers by a pressurized rolling system, (C) stress-strain curve of densified CNT fibers after the mechanical treatment, and (D) electrical resistivity at room temperature vs the stretch ratio of two sets of annealed SWNT fibers. ((A) Reproduced with permission from M. Miao, Production, structure and properties of twistless carbon nanotube yarns with a high density sheath, Carbon 50 (2012) 4973–4983; (C) Reproduced with permission from J.N. Wang, X.G. Luo, T. Wu, Y. Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014); (D) Reproduced with permission from S. Badaire, V. Pichot, C. Zakri, P. Poulin, P. Launois, J. Vavro, et al., Correlation of properties with preferred orientation in coagulated and stretch-aligned single-wall carbon nanotubes, J. Appl. Phys. 96 (2004) 7509–7513.)

reported that the electrical resistance of single-walled CNT fibers obtained by high-temperature annealing of a CNT/PVA composite fiber could be reduced by improving the CNT alignment by stretching, as shown in Fig. 6.6D. The better CNT alignment of the stretched fibers enhanced the contact areas between the CNTs, resulting in their improved electrical performance.

6.6 Infiltration Together with densification treatments, polymer infiltration is an effective method to improve CNT fiber strength. The reinforcement effect is due to the enhanced inter-tube load transfer as well as the crystallinity



117

Post-spinning treatments to carbon nanotube fibers

of the polymers [17, 56, 57]. There are various types of polymers that can be used for the treatment, such as PVA, polyethyleneimine (PEI), and epoxy [56, 58]. Liu et al. [22] compared the tensile strength of the CNT fibers experiencing three types of treatments: twisting, twisting and shrinking, and PVA infiltration, as shown in Fig. 6.7A. According to their results, the CNT/PVA composite fibers could reach a high strength of 1.95 GPa, which was 255% higher than that of the simple twisting fibers and 103% higher than the value of the CNT fibers subjected to twisting and shrinking treatment. As shown in Fig.  6.7B, the high strength of the infiltrated fibers could be attributed to two main factors: (1) the decrease of fiber diameter due to high wettability between CNTs and dimethyl sulfoxide (DMSO) and (2) the increase of tensile load due to improved load transfer efficiency between the CNTs after the infiltration treatments. Wang et al. [55] introduced graphene oxide (GO) into the CNT fiber structures to enhance their interfacial shear strength (Fig. 6.8). Because the sizes of GO and the void within the fibers matched closely, the CNT bundles were interlocked, therefore enhancing their shear interactions. Thus, the GO-infiltrated CNT fibers showed significant improvements of 100% in Young’s modulus, 110% in yield strength, 56% in tensile strength, and 30% in energy to failure.

2.0

PVA(2500)/DMSO Twisting & shrinking Simply twisting

Diameter (µm) Tensile load (mN) Tensile strength (GPa)

1.95

Stress (GPa)

1.5 1.16

1.0

230 0.5 19.0

188

152 0.54 14.4

12.3

0.0 –0.5

(A)

0.0

0.5

1.0 1.5 Strain (%)

2.0

Twisting

2.5

(B)

Twisting & shrinking

SACNT/PVA

Types of yarns

Fig. 6.7  (A) Stress-strain curves of a typical CNT/PVA fiber and two types of pure CNT fibers and (B) comparison of the diameter, tensile load, and tensile strength of a simply twisting, a twisting, and shrinking (by DMSO), and a CNT/PVA fiber. (Reproduced with permission from K. Liu, Y. Sun, X. Lin, R. Zhou, J. Wang, S. Fan, et al., Scratch-resistant, highly conductive, and high-strength carbon nanotube-based composite yarns, ACS Nano 4 (2010) 5827–5834.)

118

Carbon Nanotube Fibers and Yarns

Pristine MWCNT bundles

One MWCNT bundle GO infiltrated MWCNT bundles GO

SiO2

250 nm GO size < 50 nm

Fig.  6.8  (A) SEM image of an as-spun CNT fiber surface (top view), (B) schematic 3D model of CNT bundles intertwined with each other, (C) AFM image of a single GO particle on a SiO2 surface, and (D) schematic 3D model of GO infiltrated CNT bundles. (Reproduced with permission from Y. Wang, G. Colas, T. Filleter, Improvements in the mechanical properties of carbon nanotube fibers through graphene oxide interlocking, Carbon 98 (2016) 291–299.)

6.7 Irradiation Electron- and ion-beam irradiations have been employed to engineer CNTs and strengthen CNT assemblies. Miao et al. [59] used gamma irradiation posttreatment in the air to increase the lateral interactions between CNTs within fibers spun from CNT arrays, as shown in Fig. 6.9A. The irradiated CNT fibers exhibited a significant enhancement in their mechanical performance with an increase in the average breaking stress from 0.66 to 0.84 GPa whereas their average Young’s modulus increased from 13.9 to 23.3 GPa (Fig. 6.9B). Since the improvement in the mechanical performance of the CNT fibers was accompanied by increasing concentration of oxygen related to the CNTs, it was hypothesized that the improved mechanical properties of the fibers may stem from the chemical reaction and the resulted CNT cross-links.



Post-spinning treatments to carbon nanotube fibers

900

(B)

800

No irradiation Irradiated

700 Stress (MPa)

(A)

119

600 500 400 300 200 100 0 0

10 um

(C)

0.04

0.06

0.08

Strain

1×1013 atoms/cm2 Unirradiated

250

0.02

(D)

Stress (MPa)

200 150 100 50 0

0

1

2

3

4

5

Strain (%)

6

7

8 300 nm

Fig. 6.9  (A) CNT fibers spun from CNT array for gamma irradiation treatment, (B) stressstrain curve of the gamma irradiated and unirradiated CNT fibers, (C) stress-strain curves of the unirradiated and irradiated fibers by ion beam, and (D) SEM of the fiber location with good conditions for ion beam welding (encircled region).

Gigax et  al. [60] investigated the effects of proton irradiation on the mechanical properties of the CNT fibers spun from CNT arrays. The posttreated CNT fibers were found to have higher tensile strength and lower fracture strain after the irradiation treatment, as presented in Fig.  6.9C. This might be explained by the fact that the CNTs within bundles of the irradiated fibers were tightly packed and inter-tube linkages were likely formed in these regions. The encircled region in Fig.  6.9D might be a point of bonding since these CNTs were compressed by shear forces from the fiber production process. The ID/IG ratio decreased at low ion fluence, which indicates ion beam-induced defect repair. The repair might be explained by a thermal effect from the local beam heating and an athermal effect through the formation of mobile point defects and defect recombination with those introduced in the production process [60].

120

Carbon Nanotube Fibers and Yarns

6.8  Hybrid treatments of CNT fibers A hybrid posttreatment method consisting of acidization and epoxy infiltration was developed to enhance the mechanical performance of the CNT fibers spun by the floating catalyst method. The treatment time for acidization was optimized to balance between purification effects and structural damage on the CNTs; whereas for the epoxy infiltration step, the concentration of the epoxy solution was optimized to achieve high load transfer efficiency between CNTs. This method was proposed to provide a strategy to improve the multifunctional properties of as-spun CNT fibers, in particular, those produced by the floating catalyst method. Many studies have reported the successful improvement of the inter-tube interactions of CNT fibers by enhancing their van der Waals interactions through densification [23, 26, 28, 61] and purification [31, 40, 62] or creating new cross-links between the CNTs through polymerization [22, 27, 28, 63]. Among these posttreatments, densification and polymer infiltration are the most effective methods because they can result in the highest enhancement factor for the mechanical and electrical properties of the CNT fibers [22, 26–28].

6.8.1  Combined purification and epoxy infiltration The 15-min acidized CNT fibers mentioned before were used for the epoxy infiltration treatment to further enhance their mechanical properties. Since the epoxy viscosity was high (up to 17 mPa s), a set of diluted epoxy solutions in acetone (10, 20, 30, 40, and 50 wt.%) were prepared to study their infiltration effectiveness. Fig. 6.10A shows an SEM image of the epoxy-infiltrated CNT fibers using 30 wt.% epoxy solution and the

10.83 µm

20 µm

(A)

(B)

Fig.  6.10  (A) SEM image of the CNT fibers after epoxy infiltration at 30 wt.% and (B) effect of epoxy infiltration on diameter of the CNT fibers.



Post-spinning treatments to carbon nanotube fibers

121

diameters of the CNT fibers treated at different epoxy concentrations. The diameter of the infiltrated CNT fibers increased with increasing epoxy concentration, from 9.5 ± 0.05 μm for 0 wt.% epoxy solution to 15.9 ± 0.14 μm for 50 wt.% the epoxy solution, as shown in Fig. 6.10B. The surface morphologies of the CNT fibers infiltrated with 10, 30, and 50 wt.% epoxy solutions are presented in Fig. 6.11. The epoxy was well infiltrated into the fiber structure, and the amount of epoxy in the infiltrated fibers increased with increasing epoxy concentration. Many pores and CNT bundles were still visible in the 10 wt.% epoxy-treated fibers (Fig. 6.11A). At higher epoxy concentration (30 wt.%), the amount of epoxy on the fiber surface increased, resulting in reduced number of pores, but the CNT bundles were still visible on the treated fiber surface, as shown in Fig. 6.11B. The infiltration of the highest epoxy concentration employed in this study (50 wt.%) results in no pores and only few CNT bundles visible on the fiber surface (Fig. 6.11C), suggesting that a large amount of epoxy covers the fiber surface. This increasing epoxy amount in the fiber structure was the main reason for the increased diameters of the epoxy-infiltrated fibers. Fig.  6.12 compares the tensile strength and Young’s modulus of the CNT fibers infiltrated at different epoxy concentrations. When the epoxy

2 µm

(A)

2 µm

(B)

2 µm

(C) Fig. 6.11  SEM images of surface morphology of the CNT fibers after being infiltrated with (A) 10 wt.%, (B) 30 wt.%, and (C) 50 wt.% epoxy solutions.

122

Carbon Nanotube Fibers and Yarns

Fig.  6.12  Effect of epoxy concentration on the mechanical properties of the CNT fibers.

concentration increased from 0 to 30 wt.%, the mechanical properties of the CNT fibers significantly improved, with an increase in strength from 0.73 ± 0.07 GPa to 1.1 ± 0.02 GPa and an increase in Young’s modulus from 26.12 ± 4.32 GPa to 68.78 ± 1.33 GPa, which are 150% and 263%, respectively, compared to the 15-min acidized counterparts. These impressive improvements could be explained by the fact that the epoxy was well-infiltrated between the CNT bundles and cross-linked them after curing [27, 28]. Consequently, their inter-tube interaction was stronger, minimizing the inter-tube slippage and substantially improving the stress transfer efficiency between CNTs. On the other hand, the CNT fibers infiltrated by epoxy with concentrations higher than 30 wt.% exhibited a reduction in mechanical strength down to as low as 0.44 ± 0.06 GPa, which is even lower than that of the 15-min acidized fibers. The low fraction of CNTs in those CNT fibers with excessive epoxy resulted in ineffective reinforcements of the CNTs in the fiber structure, therefore lowering their mechanical performance [31]. The 30 wt.% epoxy solution provided an optimum amount of epoxy for the improvement of the fibers’ mechanical performance.

6.8.2  Combined densification and epoxy infiltration A simple but effective direct densification method was reported to mechanically densify CNT fibers spun by the floating catalyst technique into high density structures.The CNT fibers were spun directly from a horizontal CVD system using methane as a carbon source at a winding rate of 15 m/min. This low winding rate was used to obtain CNT fibers with good flexibility for mechanical densification treatment [31].The fibers were sandwiched between



Post-spinning treatments to carbon nanotube fibers

123

two sheets of A4 paper before being mechanically densified into CNT ribbons by a spatula. Then, 30 wt.% epoxy solution was used to infiltrate the ribbon structure to improve its mechanical performance. As shown in Fig. 6.13A, the as-spun CNT fibers had a uniform diameter of 13.5 ± 0.21 μm. Due to their low spinning rate, the alignment degree of the as-spun CNT fibers used in this study was expected to be lower than that of those spun at 20 m/min [31].The many pores and spaces on the fiber surface (Fig. 6.13B) indicate a porous structure. After the application of mechanical densification, the CNT fibers were turned into ribbons with a width of ~22 ± 1.1 μm and a thickness of ~0.65 ± 0.12 μm, as shown in Fig. 6.14A and B. Comparing to the as-spun CNT fibers, the CNT ribbons had a more packed structure with better CNT alignment along the fiber axis (Fig. 6.14C). After epoxy-infiltration, the CNT ribbons showed smoother surfaces as the epoxy filled up most pores and spaces (Fig. 6.14D–F). The epoxy coated ribbons had width and thickness of 23.5 ± 1.2 μm and 1 ± 0.2 μm, respectively. Fig. 6.15A shows stress-strain curves of the as-spun CNT fibers, CNT ribbons, and epoxy-infiltrated CNT ribbons. As can be seen, a transition from elastic to plastic deformation appeared before failure in all curves. The curves are linear and typically present a sharp increase in slope at low strains (1%–2%) and a gradual decrease in slope at high strains, indicating that the samples had ductile behavior. Overall, the strength and modulus of the CNT samples increased significantly after each treatment was applied. Fig.  6.15B compares the mechanical strength and modulus of the CNT fibers, CNT ribbons, and cross-linked CNT ribbons. The asspun CNT fibers had tensile strength, Young’s modulus and elongation of 0.27 ± 0.01 GPa, 4.28 ± 0.38 GPa, 12%, respectively [26], which were within the range of the as-spun MWNT fibers reported in the literature

Pores

50 µm

(A)

CNT fiber direction

2 µm

(B)

Fig. 6.13  SEM images of (A) the as-spun CNT fiber and (B) its surface morphology.

124

Carbon Nanotube Fibers and Yarns

(A)

50 µm

(D)

50 µm

(B)

5 µm

(E)

5 µm

Pores

CNT fiber direction

CNT fiber direction

(C)

2 µm

(F)

2 µm

Fig. 6.14  SEM images showing the (A) width, (B) thickness, and (C) surface morphology of the CNT ribbons, and the (D) width, (E) thickness, and (F) surface morphologies of the epoxy-infiltrated CNT ribbons.

(0.15–0.46 GPa) [17]. The performance may be explained by the loose structures of the CNT fibers (Fig. 6.13B) and weak interactions between the CNTs and CNT bundles [20]. The results indicated that liquid densification, specifically the ethanol spraying, was not an effective method to produce high-performance CNT fibers. After the mechanical densification, the strength and Young’s modulus increased to 2.81 ± 0.07 GPa and 78.72 ± 6.51 GPa, or nearly 10 and 18 times, respectively. This remarkable improvement was mainly due to the reduction in the cross-sectional area of the CNT fibers from 143.1 ± 5.2 μm2 to 14.3 ± 1.4 μm2 after the treatment because the breaking tensile load was increased only marginally, as presented in Fig. 6.15C.



125

Post-spinning treatments to carbon nanotube fibers

Cross-linked CNT ribbon 4.0

CNT ribbon

2

1

200

2.5 2.0

150

1.5

100

1.0 CNT fiber

0 0.00

250

3.0 Strength (GPa)

Strength (GPa)

3

300

Strength Young’s modulus

3.5

0.05

0.10

50

0.5

0.15

Strain

(A)

Young’s modulus (GPa)

4

0.0

CNT fibers

(B)

CNT ribbons

Cross-linked CNT ribbons

0

Cross-sectional area (µm2) Tensile load (cN) Failure elongation (%)

143.07 12.08 11.28

8.45

4.79 4.01

3.81 14.30

(C)

CNT fibers

CNT ribbons

23.50

Cross-linked CNT ribbons

Fig. 6.15  (A) Stress-strain curves of the CNT samples, (B) tensile strength and Young’s modulus of the CNT samples, and (C) cross-sectional area, tensile load, and elongation to failure of the CNT samples.

The highly dense structures of the CNT ribbon observed in Fig. 6.14C suggest that the mechanical densification treatment increased the CNT bundle sizes and inter-CNT contacts, and induced better alignments, leading to a slight increase in their breaking load from 3.81 ± 0.13 to 4.01 ± 0.1 cN (Fig. 6.15C).The decrease in the failure strain of the CNT ribbons after the densification was a result of their stronger inter-CNT interactions. These findings were in agreement with that of the CNT ribbons densified by pressurized rolling methods [26], suggesting that mechanical densification could be an effective posttreatment to produce highly densified CNT structures with improved mechanical performance. In the epoxy infiltration treatment, the cross-sectional area of the coated CNT ribbons increased by more than 60%, as presented in Fig. 6.15C.This result was opposite to the reduction in the cross-sectional area of the CNT fibers after the combined treatment of liquid densification and polymer infiltration [22, 27].The large amount of epoxy coating on the CNTs and CNT bundles of the CNT ribbons was the

126

Carbon Nanotube Fibers and Yarns

main reason for their increased cross-sectional area [31].Although the breaking load of the CNT ribbons increased by more than 100% after the infiltration treatment, their strength only showed a slight net increase of more than 20% to 3.6 ± 0.16 GPa. The enhancement in Young’s modulus by epoxy-coated CNT ribbons was remarkable although the improvement in the strength of the epoxy-coated CNT ribbons was much lower than that of the CNT fibers treated by combined treatment of liquid densification and polymer infiltration [22, 27]. Specifically, their Young’s modulus increased by more than three times with an average value of 266 ± 14.47 GPa after the infiltration treatment. These results may be attributed to significant enhancement in the interfacial load transfer between the CNTs and between the CNT bundles. The epoxy infiltrated into the CNT structures, cross-linked the CNTs and CNT bundles with strong bonds, and minimized the inter-tube slippage [64]. Consequently, the inter-tube interactions of the cross-linked CNT ribbons were significantly enhanced, resulting in the large increase in their Young’s modulus. Moreover, the reductions in both failure elongation and plastic region of the coated CNT ribbons can be attributed to the hindering effects of the infiltrated epoxy on the inter-tube slippage [64].

6.8.3  Advantages of hybrid posttreatments Mechanical densification has been used in many studies to densify different CNT assemblies such as CNT arrays with biaxial densification [65] and rolling [66], and CNT fibers with pressurized system [26]. In the densification technique described above, compressive stress is applied perpendicular the fiber axis to densify the CNT fibers into highly densified structures. The resulting shear stress during the treatment causes CNTs to slide slightly along the fiber axis, improving the CNT alignments. Moreover, the use of two protective layers can prevent damage caused by the shear and compressive effects during the treatment while maximizing the packing effects [28]. The thickness of the densified samples is determined by the gap between the two protective sheets under the compressive forces during the densification treatment. There was no damage observed in the CNT fiber structures although excessive densification forces of 100 N were applied to the samples many times. This technique can be used to densify CNT fibers of continuous length using a mechanism similar to the pressurized rolling system [26]. The hybrid posttreatment described in this chapter achieved excellent improvement in the mechanical performance of the as-spun MWNT fibers. Fig. 6.16A compares the improvement factors in the mechanical performance of CNT fibers using different posttreatments, including liquid densification

80

Strength Young’s modulus

Improvement factor

70 60 50 40 30 20 10 0

liquid densification

acid treatment

(A)

laser treatment

liquid Mechanical Mechanical densification densification densification + infiltration + infiltration

Post-treatment methods CNT ribbon from aerogel spinning

Tensile strength (GPa)

6 5 4 3

PAN carbon fiber

2

CNT fiber from wet spinning

1 0

(B)

0

100

CNT fiber from array spinning

This work

CNT fiber CNT ribbon Cross-linked CNT ribbon

200 300 400 500 Young’s modulus (GPa)

600

Fig. 6.16  (A) Comparisons of improvement factors of different posttreatments in literature, including liquid densification [22], acid treatment [52], laser treatment [67], hybrid treatment of liquid densification and polymer infiltration [22], mechanical densification [26], and a combination of mechanical densification and polymer infiltration, (B) comparisons of mechanical properties of the best CNT fibers from array-spinning [20] and wet-spinning [1], ribbons from aerogel spinning, and PAN carbon fibers [26]. (Sources of (A): K. Liu, Y. Sun, X. Lin, R. Zhou, J. Wang, S. Fan, et al., Scratch-resistant, highly conductive, and high-strength carbon nanotube-based composite yarns, ACS Nano 4 (2010) 5827–5834; K. Wang, M. Li, Y.N. Liu, Y. Gu, Q. Li, Z. Zhang, Effect of acidification conditions on the properties of carbon nanotube fibers, Appl. Surf. Sci. 292 (2014) 469–474; K. Liu, F. Zhu, L. Liu, Y. Sun, S. Fan, K. Jiang, Fabrication and processing of high-strength densely packed carbon nanotube yarns without solution processes, Nanoscale 4 (2012) 3389–3393; J.N. Wang, X.G. Luo, T. Wu, Y. Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014) 3848. Sources of (B): X. Zhang, Q. Li, Y. Tu, Y. Li, J.Y. Coulter, L. Zheng, et al., Strong ­carbon-nanotube fibers spun from long carbon-nanotube arrays, Small 3 (2007) 244–248; N. Behabtu, C.C. Young, D.E. Tsentalovich, O. Kleinerman, X. Wang, A.W.K. Ma, et al., Strong, light, multifunctional fibers of carbon nanotubes with ultrahigh conductivity, Science 339 (2013) 182–186; J.N. Wang, X.G. Luo, T. Wu, Y. Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014) 3848.)

128

Carbon Nanotube Fibers and Yarns

[22], acid treatment [52], laser treatment [67], hybrid treatment of liquid densification and polymer infiltration [22], and mechanical densification [26].The hybrid posttreatment is the most effective method, with improvement factors of more than 13.5 for tensile strength and 63 for Young’s modulus. The strength and Young’s modulus of the CNT ribbons and epoxyinfiltrated CNT ribbons are much higher than those of the CNT fibers posttreated by the other methods. As shown in Fig. 6.16B, the tensile strength of the densified CNT ribbons is higher than that of the best CNT fibers spun from wet-spinning and array-spinning methods. After further infiltrated with the epoxy, the CNT composite ribbons reach a high strength of up to 5.2 GPa and Young’s modulus of up to 444 GPa, which are comparable to those of commercial PAN carbon fibers. Moreover, the strength of the CNT-epoxy composite ribbons is comparable to that of the best DWNT ribbons fabricated by the floating catalyst method (5.53 GPa) [26] but with much higher Young’s modulus. Knot efficiency of fibers is the ratio between the strength of the knotted fibers and their unknotted counterparts. Fig. 6.17 shows that the CNT fibers and CNT ribbons exhibited excellent knot performance, with 100% knot efficiency. Due to the alteration of their yarn structures, the ­CNT-epoxy

Carbon T300 Dyneema Kevlar 49 Twaron Nylon Fiber glass Silk Wool Cotton CNT fiber CNT ribbon Cross-linked CNT ribbon

Knot strength efficiency (%)

100

80

60

40

20

0

0

1

2 3 Tensile strength (GPa)

4

5

Fig. 6.17  The plot of knot-strength efficiency against fiber strength for a variety of fibers. The values of Kevlar 49, Dyneema, Carbon T300, and Twaron are derived from Ref. [9] while the others are derived from Ref. [68]. (Sources: J.J. Vilatela, A.H. Windle, Yarn-like carbon nanotube fibers, Adv. Mater. 22 (2010) 4959–4963; W.W. Morton, J.W. Hearle, Physical Properties of Textile Fibers, fourth ed., The textile Institute, Manchester, 2008.)



Post-spinning treatments to carbon nanotube fibers

129

composite ribbons show a slightly lower knot-strength efficiency of 78%. However, their knot performance is still comparable to that of silk, wool, nylon, and cotton, and much better than that of other high-strength commercial fibers [9].

6.9  Conclusions and recommendations We first focused on the dissolution of CNT fibers spun by the floating catalyst method in chlorosulfonic acid (CSA). The as-spun CNT fibers were not fully dissolved in the CSA because the impurities within the fibers hindered the protonation effects of the superacid. After the removal of these impurities by purification, the CNT fibers were dissolved successfully in CSA, leading to the observation of isotropic, biphasic, and liquid crystal phases of the CNT solutions [40, 44]. The length of the CNTs constituting the fibers was characterized by extensional viscosity method [40, 47]. The CNT length was determined to be 2 μm in average, which was still longer than many commercial SWNTs and DWNTs, such as SweNT CG300, UniDym OE, HiPco 188.3, and HiPco 183.6 (0.71–1.92 μm). A second focus of the chapter is posttreatment methods based on mechanical densification, which could significantly improve the mechanical and electrical properties of the CNT fibers.The outstanding feature of this method is that very high densifying force of 100 N could be applied directly to the CNT fibers while minimizing the damage of the fiber structure.The method improved the CNT alignment of the fibers and reduced spaces and pores within the fiber structure. After the densification, the CNT fibers exhibited increases in strength and Young’s modulus by 10 and 18 times, respectively.These are comparable to the improved performance of the CNT fibers densified by pressurized rolling process [26] and much better than twisting [22, 23], liquid densification [23], drawing through dies [24], and rubbing [25].The technique can also be applied to aligned CNT fibers and CNT thin films fabricated from the wet-spinning and array-spinning methods with many potential applications such as composite reinforcements and electrically conducting wires [31, 40, 69, 70]. Effects of different posttreatment methods, especially the hybrid posttreatments, such as purification combined with epoxy infiltration and densification combined with epoxy infiltration, on the electrical and mechanical performances of the CNT fibers spun from the floating catalyst method were reviewed comprehensively. In particular, oxidative purification removes catalyst impurity of the CNT fibers spun from toluene source by up to 65%. The purified CNT fibers exhibited strength,Young’s modulus, and

130

Carbon Nanotube Fibers and Yarns

electrical conductivity of 0.38 GPa, 20.66 GPa, and 4679 S/cm, respectively, corresponding to 120%, 250%, and 200% of their as-spun fibers. The hybrid treatment consisting of acidization and epoxy infiltration treatments, on the other hand, showed a much higher improvement factor in mechanical properties of the CNT fibers. After the treatment, the CNT fibers exhibited a 2.8 times increase in strength and 4.2 times increase in Young’s modulus, reaching 1.1 GPa and 68.78 GPa, respectively. The CNT fibers treated by the hybrid treatment of mechanical densification and epoxy infiltration possessed excellent strength of 3.6 GPa and Young’s modulus of 266 GPa, representing enhancement factors for strength and Young’s modulus of 13.5 times and 62 times, which are the highest reported in the literature [28]. Their performance was comparable to many commercial high strength fibers, such as PAN-based carbon fibers, suggesting great potentials for the mass production of high-performance CNT fibers. Although the experimental setups for the hybrid-posttreatments can be only used to produce short CNT fibers, the posttreatments can be scaled up to synchronize with the synthesis process for online fabrication of high-performance posttreated CNT fibers with unlimited length. Only CNT fibers spun from methane source were treated with the hybrid posttreatments so far. Fibers spun from toluene source and their purified counterparts have lower impurities and less defective structures. The potential effects of the posttreatments such as acidization, mechanical densification, and epoxy infiltration on their performance need to be investigated in the future. Although the oxidative purification could significantly remove impurities in the CNT fibers spun from the floating catalyst method, the process might damage the CNTs by introducing defects. These defects were undesirable since they lowered the fiber performance and, therefore, limited the efficiency of the purification method. Physical methods, such as centrifugation [70] high-temperature annealing [71], filtration [72], and multistep purification methods, such as microfiltration in combination with oxidation [62, 73], high-temperature annealing in combination with extraction [74], and sonication in combination with oxidation [75] are known not to severely damage the CNT structures [42] These methods should be investigated for purification of the as-spun CNT fibers to improve fiber performance. Finally, as the performance of the posttreated CNT fibers was comparable to many commercial high-strength fibers, such as carbon fibers T300, Dyneema, and Twaron, they can be used as reinforcement for advanced composites. Nanotube-based composites fabricated from unstructured CNT powders have been widely used for a broad range of



Post-spinning treatments to carbon nanotube fibers

131

applications, such as structural materials for automotive and aerospace applications [76], electrical and thermal conductors for energy applications [33, 77], ­nano-biotechnology [78], and in other disciplines [44, 79]. The CNT fibers with their aligned CNT structures and excellent mechanical and electrical properties are promising materials to fabricate high performance, lightweight, and multifunctional composites.

References [1] N. Behabtu, C.C. Young, D.E. Tsentalovich, O. Kleinerman, X. Wang, A.W.K. Ma, et al., Strong, light, multifunctional fibers of carbon nanotubes with ultrahigh conductivity, Science 339 (2013) 182–186. [2] L.M.  Ericson, H.  Fan, H.  Peng, V.A.  Davis, W.  Zhou, J.  Sulpizio, et  al., Macroscopic, neat, ­single-walled carbon nanotube fibers, Science 305 (2004) 1447–1450. [3] S. Fang, M. Zhang, A.A. Zakhidov, R.H. Baughman, Structure and process-dependent properties of solid-state spun carbon nanotube yarns, J. Phys. Condens. Matter 22 (2010) 334221. [4] T. Mirfakhrai, M. Kozlov, S. Fang, M. Zhang, R.H. Baughman, J.D. Madden, Carbon nanotube yarns: sensors, actuators and current carriers, Proc. SPIE Int. Soc. Opt. Eng. 6927 (2008) 692708. [5] M. Zhang, K.R. Atkinson, R.H. Baughman, Multifunctional carbon nanotube yarns by downsizing an ancient technology, Science 306 (2004) 1358–1361. [6] T.S. Gspann, F.R. Smail, A.H. Windle, Spinning of carbon nanotube fibres using the floating catalyst high temperature route: purity issues and the critical role of sulphur, Faraday Discuss. 173 (2014) 47–65. [7] Y.L. Li, I.A. Kinloch, A.H. Windle, Direct spinning of carbon nanotube fibers from chemical vapor deposition synthesis, Science 304 (2004) 276–278. [8] R.M. Sundaram, K.K.K. Koziol, A.H. Windle, Continuous direct spinning of fibers of single-walled carbon nanotubes with metallic chirality, Adv. Mater. 23 (2011) 5064–5068. [9] J.J.  Vilatela, A.H.  Windle, Yarn-like carbon nanotube fibers, Adv. Mater. 22 (2010) 4959–4963. [10] H.M. Duong, F. Gong, P. Liu,T.Q.Tran, Advanced fabrication and properties of aligned carbon nanotube composites: experiments and modeling, in: R.M. Berber (Ed.), Carbon Nanotubes, InTech, 2016, pp. 47–72. [11] P. Liu,T.Q. Tran, Z. Fan, H.M. Duong, Formation mechanisms and morphological effects on multi-properties of carbon nanotube fibers and their polyimide aerogel-coated composites, Compos. Sci. Technol. 117 (2015) 114–120. [12] T.Q. Tran, Z. Fan, P. Liu, H.M. Duong, Advanced morphology-controlled manufacturing of carbon nanotube fibers, thin films and aerogels from aerogel technique, in: Asia Pacific Confederation of Chemical Engineering Congress 2015: APCChE 2015, Incorporating CHEMECA 2015, Engineers Australia, 2015, p. 2444. [13] H. Khoshnevis, T.Q. Tran, S.M. Mint, A. Zadhoush, H.M. Duong, M. Youssefi, Effect of alignment and packing density on the stress relaxation process of carbon nanotube fibers spun from floating catalyst chemical vapor deposition method, Colloids Surf. A Physicochem. Eng. Asp. 558 (2018) 570–578. [14] S. Badaire,V. Pichot, C. Zakri, P. Poulin, P. Launois, J. Vavro, et al., Correlation of properties with preferred orientation in coagulated and stretch-aligned single-wall carbon nanotubes, J. Appl. Phys. 96 (2004) 7509–7513. [15] M.B. Jakubinek, M.B. Johnson, M.A. White, C. Jayasinghe, G. Li, W. Cho, et al., Thermal and electrical conductivity of array-spun multi-walled carbon nanotube yarns, Carbon 50 (2012) 244–248.

132

Carbon Nanotube Fibers and Yarns

[16] Q.  Li, Y.  Li, X.  Zhang, S.B.  Chikkannanavar, Y.  Zhao, A.M.  Dangelewicz, et  al., ­Structure-dependent electrical properties of carbon nanotube fibers, Adv. Mater. 19 (2007) 3358–3363. [17] W. Lu, M. Zu, J.H. Byun, B.S. Kim, T.W. Chou, State of the art of carbon nanotube fibers: opportunities and challenges, Adv. Mater. 24 (2012) 1805–1833. [18] M. Miao, Electrical conductivity of pure carbon nanotube yarns, Carbon 49 (2011) 3755–3761. [19] P.  Miaudet, C.  Bartholome, A.  Derré, M.  Maugey, G.  Sigaud, C.  Zakri, et  al., ­Thermo-electrical properties of PVA-nanotube composite fibers, Polymer 48 (2007) 4068–4074. [20] X. Zhang, Q. Li, Y. Tu, Y. Li, J.Y. Coulter, L. Zheng, et al., Strong carbon-nanotube fibers spun from long carbon-nanotube arrays, Small 3 (2007) 244–248. [21] J. Zhao, X. Zhang, J. Di, G. Xu, X. Yang, X. Liu, et al., Double-peak mechanical properties of carbon-nanotube fibers, Small 6 (2010) 2612–2617. [22] K. Liu,Y. Sun, X. Lin, R. Zhou, J. Wang, S. Fan, et al., Scratch-resistant, highly conductive, and high-strength carbon nanotube-based composite yarns, ACS Nano 4 (2010) 5827–5834. [23] K. Liu,Y. Sun, R. Zhou, H. Zhu, J. Wang, L. Liu, et al., Carbon nanotube yarns with high tensile strength made by a twisting and shrinking method, Nanotechnology 21 (2010) 045708. [24] K. Sugano, M. Kurata, H. Kawada, Evaluation of mechanical properties of untwisted carbon nanotube yarn for application to composite materials, Carbon 78 (2014) 356–365. [25] M. Miao, Production, structure and properties of twistless carbon nanotube yarns with a high density sheath, Carbon 50 (2012) 4973–4983. [26] J.N. Wang, X.G. Luo, T. Wu, Y. Chen, High-strength carbon nanotube fibre-like ribbon with high ductility and high electrical conductivity, Nat. Commun. 5 (2014) 3848. [27] S. Ryu, Y. Lee, J.W. Hwang, S. Hong, C. Kim, T.G. Park, et al., High-strength carbon nanotube fibers fabricated by infiltration and curing of mussel-inspired catecholamine polymer, Adv. Mater. 23 (2011) 1971–1975. [28] T.Q. Tran, Z. Fan, P. Liu, S.M. Myint, H.M. Duong, Super-strong and highly conductive carbon nanotube ribbons from post-treatment methods, Carbon 99 (2016) 407–415. [29] P. Liu, Z. Fan, A. Mikhalchan, T.Q. Tran, D. Jewell, H.M. Duong, et al., Continuous carbon nanotube-based fibers and films for applications requiring enhanced heat dissipation, ACS Appl. Mater. Interfaces 8 (2016) 17461–17471. [30] P. Liu, D.C.M. Hu, T.Q. Tran, D. Jewell, H.M. Duong, Electrical property enhancement of carbon nanotube fibers from post treatments, Colloids Surf. A Physicochem. Eng. Asp. 509 (2016) 384–389. [31] T.Q.  Tran, Z.  Fan, A.  Mikhalchan, P.  Liu, H.M.  Duong, Post-treatments for multifunctional property enhancement of carbon nanotube fibers from the floating catalyst method, ACS Appl. Mater. Interfaces 8 (2016) 7948–7956. [32] B. Vigolo, A. Penicaud, C. Coulon, C. Sauder, R. Pailler, C. Journet, et al., Macroscopic fibers and ribbons of oriented carbon nanotubes, Science 290 (2000) 1331–1334. [33] A.B.  Dalton, S.  Collins, E.  Muñoz, J.M.  Razal, V.H.  Ebron, J.P.  Ferraris, et  al., ­Super-tough carbon-nanotube fibres, Nature 423 (2003) 703. [34] G. Gu, M. Schmid, P.-W. Chiu, A. Minett, J. Fraysse, G.-T. Kim, et al.,V2O5 nanofibre sheet actuators, Nat. Mater. 2 (2003) 316. [35] M.E. Kozlov, R.C. Capps,W.M. Sampson,V.H. Ebron, J.P. Ferraris, R.H. Baughman, Spinning solid and hollow polymer-free carbon nanotube fibers,Adv. Mater. 17 (2005) 614–617. [36] L.K. Randeniya, A. Bendavid, P.J. Martin, C.D. Tran, Composite yarns of multiwalled carbon nanotubes with metallic electrical conductivity, Small 6 (2010) 1806–1811. [37] Y. Zhao, J. Wei, R. Vajtai, P.M. Ajayan, E.V. Barrera, Iodine doped carbon nanotube cables exceeding specific electrical conductivity of metals, Sci. Rep. 1 (2011) 83. [38] K.  Koziol, J.  Vilatela, A.  Moisala, M.  Motta, P.  Cunniff, M.  Sennett, et  al., High-­ performance carbon nanotube fiber, Science 318 (2007) 1892–1895.



Post-spinning treatments to carbon nanotube fibers

133

[39] H. Khoshnevis, S.M. Mint, E. Yedinak, T.Q. Tran, A. Zadhoush, M. Youssefi, et al., Super high-rate fabrication of high-purity carbon nanotube aerogels from floating catalyst method for oil spill cleaning, Chem. Phys. Lett. 693 (2018) 146–151. [40] T.Q. Tran, R.J. Headrick, E.A. Bengio, S. Myo Myint, H. Khoshnevis,V. Jamali, et al., Purification and dissolution of carbon nanotube fibers spun from the floating catalyst method, ACS Appl. Mater. Interfaces 9 (2017) 37112–37119. [41] H. Cheng, K.L.P. Koh, P. Liu, T.Q. Thang, H.M. Duong, Continuous self-assembly of carbon nanotube thin films and their composites for supercapacitors, Colloids Surf. A Physicochem. Eng. Asp. 481 (2015) 626–632. [42] P.X. Hou, C. Liu, H.M. Cheng, Purification of carbon nanotubes, Carbon 46 (2008) 2003–2025. [43] Y. Lin, J.W. Kim, J.W. Connell, M. Lebrõn-Colõn, E.J. Siochi, Purification of carbon nanotube sheets, Adv. Eng. Mater. 17 (2015) 674–688. [44] V.A. Davis, A.N.G. Parra-Vasquez, M.J. Green, P.K. Rai, N. Behabtu, V. Prieto, et al., True solutions of single-walled carbon nanotubes for assembly into macroscopic materials, Nat. Nanotechnol. 4 (2009) 830–834. [45] F.  Mirri, A.W.K.  Ma, T.T.  Hsu, N.  Behabtu, S.L.  Eichmann, C.C.  Young, et  al., High-performance carbon nanotube transparent conductive films by scalable dip coating, ACS Nano 6 (2012) 9737–9744. [46] S. Osswald,Y. Gogotsi, In situ Raman spectroscopy of oxidation of carbon nanomaterials, in: Challa S.S.R. Kumar (Ed.), Raman Spectroscopy for Nanomaterials Characterization, Springer Nature, Berlin, 2012, pp. 291–351. [47] D.E. Tsentalovich, A.W.K. Ma, J.A. Lee, N. Behabtu, E.A. Bengio, A. Choi, et al., Relationship of extensional viscosity and liquid crystalline transition to length distribution in carbon nanotube solutions, Macromolecules 49 (2016) 681–689. [48] P.W. Oakes, J. Viamontes, J.X. Tang, Growth of tactoidal droplets during the first-­order isotropic to nematic phase transition of F-actin, Phys. Rev. E Stat. Nonlinear Soft Matter Phys. 75 (2007) 061902. [49] M.S. Motta, A. Moisala, I.A. Kinloch, A.H. Windle,The role of sulphur in the synthesis of carbon nanotubes by chemical vapour deposition at high temperatures, J. Nanosci. Nanotechnol. 8 (2008) 2442–2449. [50] F. Meng, J. Zhao,Y. Ye, X. Zhang, Q. Li, Carbon nanotube fibers for electrochemical applications: effect of enhanced interfaces by an acid treatment, Nanoscale 4 (2012) 7464–7468. [51] S. Li,Y. Shang, W. Zhao,Y. Wang, X. Li, A. Cao, Efficient purification of single-walled carbon nanotube fibers by instantaneous current injection and acid washing, RSC Adv. 6 (2016) 97865–97872. [52] K. Wang, M. Li,Y.N. Liu,Y. Gu, Q. Li, Z. Zhang, Effect of acidification conditions on the properties of carbon nanotube fibers, Appl. Surf. Sci. 292 (2014) 469–474. [53] L. Shao, G. Tobias, C.G. Salzmann, B. Ballesteros, S.Y. Hong, A. Crossley, et al., Removal of amorphous carbon for the efficient sidewall functionalisation of single-walled carbon nanotubes, Chem. Commun. (2007) 5090–5092. [54] L.  Stobinski, B.  Lesiak, L.  Kövér, J.  Tóth, S.  Biniak, G.  Trykowski, et  al., Multiwall carbon nanotubes purification and oxidation by nitric acid studied by the FTIR and electron spectroscopy methods, J. Alloys Compd. 501 (2010) 77–84. [55] Y. Wang, G. Colas, T. Filleter, Improvements in the mechanical properties of carbon nanotube fibers through graphene oxide interlocking, Carbon 98 (2016) 291–299. [56] S. Li, X. Zhang, J. Zhao, F. Meng, G. Xu, Z. Yong, et al., Enhancement of carbon nanotube fibres using different solvents and polymers, Compos. Sci.Technol. 72 (2012) 1402–1407. [57] P. Liu, A.L. Cottrill, D. Kozawa,V.B. Koman, D. Parviz, A.T. Liu, et al., Emerging trends in 2D nanotechnology that are redefining our understanding of “nanocomposites”, Nano Today 21 (2018) 18–40. [58] J.Y. Cai, J. Min, M. Miao, J.S. Church, J. McDonnell, R. Knott, et al., Enhanced mechanical performance of CNT/polymer composite yarns by γ-irradiation, Fibers Polym. 15 (2014) 322–325.

134

Carbon Nanotube Fibers and Yarns

[59] M. Miao, S.C. Hawkins, J.Y. Cai, T.R. Gengenbach, R. Knott, C.P. Huynh, Effect of gamma-irradiation on the mechanical properties of carbon nanotube yarns, Carbon 49 (2011) 4940–4947. [60] J.G. Gigax, P.D. Bradford, L. Shao, Radiation-induced mechanical property changes of CNT yarn, Nucl. Instrum. Methods Phys. Res., Sect. B 409 (2017) 268–271. [61] A. Mikhalchan, Z. Fan, T.Q. Tran, P. Liu,V.B.C. Tan, T.-E. Tay, et al., Continuous and scalable fabrication and multifunctional properties of carbon nanotube aerogels from the floating catalyst method, Carbon 102 (2016) 409–418. [62] R.M. Sundaram, A.H. Windle, One-step purification of direct-spun CNT fibers by post-production sonication, Mater. Des. 126 (2017) 85–90. [63] P. Liu, A. Lam, Z. Fan, T.Q. Tran, H.M. Duong, Advanced multifunctional properties of aligned carbon nanotube-epoxy thin film composites, Mater. Des. 87 (2015) 600–605. [64] W. Ma, L. Liu, Z. Zhang, R. Yang, G. Liu,T. Zhang, et al., High-strength composite fibers: realizing true potential of carbon nanotubes in polymer matrix through continuous reticulate architecture and molecular level couplings, Nano Lett. 9 (2009) 2855–2861. [65] A.V. Krasheninnikov, F. Banhart, Engineering of nanostructured carbon materials with electron or ion beams, Nat. Mater. 6 (2007) 723–733. [66] C.L. Pint,Y.Q. Xu, M. Pasquali, R.H. Hauge, Formation of highly dense aligned ribbons and transparent films of single-walled carbon nanotubes directly from carpets, ACS Nano 2 (2008) 1871–1878. [67] K. Liu, F. Zhu, L. Liu, Y. Sun, S. Fan, K. Jiang, Fabrication and processing of highstrength densely packed carbon nanotube yarns without solution processes, Nanoscale 4 (2012) 3389–3393. [68] W.W. Morton, J.W. Hearle, Physical Properties of Textile Fibers, fourth ed., The textile Institute, Manchester, 2008. [69] P.  Liu, Y.F.  Tan, D.C.M.  Hu, D.  Jewell, H.M.  Duong, Multi-property enhancement of aligned carbon nanotube thin films from floating catalyst method, Mater. Des. 108 (2016) 754–760. [70] A.  Yu, E.  Bekyarova, M.E.  Itkis, D.  Fakhrutdinov, R.  Webster, R.C.  Haddon, Application of centrifugation to the large-scale purification of electric arc-produced ­single-walled carbon nanotubes, J. Am. Chem. Soc. 128 (2006) 9902–9908. [71] J.M. Lambert, P.M. Ajayan, P. Bernier, J.M. Planeix,V. Brotons, B. Coq, et al., Improving conditions towards isolating single-shell carbon nanotubes, Chem. Phys. Lett. 226 (1994) 364–371. [72] J.M. Bonard,T. Stora, J.P. Salvetat, F. Maier,T. Stöckli, C. Duschl, et al., Purification and size-selection of carbon nanotubes, Adv. Mater. 9 (1997) 827–831. [73] Y. Kim, D.E. Luzzi, Purification of pulsed laser synthesized single wall carbon nanotubes by magnetic filtration, J. Phys. Chem. B 109 (2005) 16636–16643. [74] H. Zhang, C.H. Sun, F. Li, H.X. Li, H.M. Cheng, Purification of multiwalled carbon nanotubes by annealing and extraction based on the difference in van der waals potential, J. Phys. Chem. B 110 (2006) 9477–9481. [75] P. Hou, C. Liu,Y. Tong, S. Xu, M. Liu, H. Cheng, Purification of single-walled carbon nanotubes synthesized by the hydrogen arc-discharge method, J. Mater. Res. 16 (2001) 2526–2529. [76] B.L. Wardle, D.S. Saito, E.J. García, A.J. Hart, R. Guzmán De Villoria, E.A. Verploegen, Fabrication and characterization of ultrahigh-volume-fraction aligned carbon ­nanotube-polymer composites, Adv. Mater. 20 (2008) 2707–2714. [77] K. Jiang, J. Wang, Q. Li, L. Liu, C. Liu, S. Fan, Superaligned carbon nanotube arrays, films, and yarns: a road to applications, Adv. Mater. 23 (2011) 1154–1161. [78] J.M.  Razal, K.J.  Gilmore, G.G.  Wallace, Carbon nanotube biofiber formation in a ­polymer-free coagulation bath, Adv. Funct. Mater. 18 (2008) 61–66. [79] Z. Huang, M. Gao, T. Pan,Y. Zhang, B. Zeng, W. Liang, et al., Microstructure dependence of heat sink constructed by carbon nanotubes for chip cooling, J. Appl. Phys. 117 (2015) 024901.

PART II

Structures and properties

This page intentionally left blank

CHAPTER 7

Carbon nanotube yarn structures and properties Menghe Miao

CSIRO Manufacturing, Geelong,VIC, Australia

Carbon nanotubes (CNTs) have been shown to possess extraordinarily high mechanical properties [1, 2] combined with good electrical and thermal conductivity [3]. The challenge is to organize these nano-sized building blocks into macroscale structures that express similar properties. Without considering their detailed atomic structures, CNTs can be considered as nanoscale fibers that resemble the diameter of fibrils in plant and animal fibers such as cotton and wool. It is, therefore, a logical approach to align the CNTs in the form of a fiber or yarn that is expected to outperform their conventional textile counterparts [4].

7.1  CNT yarn geometry 7.1.1 Twist Twist insertion has been an important method to densify CNT webs drawn from vertically aligned CNT arrays (CNT forests) since it was first reported in 2004 [5].The method imitates the spinning of short textile fibers (staple fibers) into a continuous yarn, which has been traditionally used and is still the predominant method of staple fiber yarn production in the textile industry today. Twist is applied to a yarn by rotating one end of the yarn while it is wound on a bobbin. In the textile industry, twist is measured by counting the number of turns required per unit length (T, turns per unit length) to cancel out the twist applied to the yarn during spinning. In spinning mills, the twist applied to a yarn is set by tuning the ratio between the rotational speed of the spindle (revolutions per minute) and the throughput speed of the yarn (length per minute) set on the spinning machine. The geometry of twisted yarns is often represented by a series of coaxial helices (Fig. 7.1), a model first proposed by Gégauff in 1907 [7].The model has been widely adopted, sometimes with minor modifications, in yarn structural Carbon Nanotube Fibers and Yarns https://doi.org/10.1016/B978-0-08-102722-6.00007-9

Copyright © 2020 Elsevier Ltd. All rights reserved.

137

138

Carbon Nanotube Fibers and Yarns

Fig.  7.1  Coaxial helix model of twisted yarns [6]. (Reprinted with permission from M. Miao, The role of twist in dry spun carbon nanotube yarns, Carbon 96 (2016) 819–826.)

mechanics analysis [8].When describing the helical path of a fiber in a yarn, we use the angle between the fiber and the yarn axis rather than the rising angle of the helix.The length of one turn of the helices is a constant (h) independent of the radial position of the fiber in the yarn, which is equal to the reciprocal of the twist of the yarn (T), h = 1/T. Consequently, the helix angle of a fiber varies according to its radial position in the yarn. At the center of the yarn (r = 0), the fiber helix angle is zero; and at the yarn surface (r = d/2, where d is the yarn diameter), the fiber helix angle is the maximum (θ in Fig. 7.1). The helix angle of fibers on the yarn surface is commonly referred to as the twist angle of the yarn. The twist angle of a yarn is related to its twist T and diameter d: tan θ = πdT. Although it is convenient to measure the twist level of a yarn by the number of turns per meter (TPM), the helical angle of fibers on the yarn surface should be used when comparing the degree of twist in yarns with different diameters. Clearly, for two yarns with the same TPM, a thicker yarn has a larger twist angle than a finer yarn.

7.1.2  Yarn diameter and linear density Measuring the diameter of a CNT yarn from SEM images has been a common practice in earlier research works [5, 9–11]. This method has some



Carbon nanotube yarn structures and properties

139

inherent shortcomings. Diameter measurement taken from random points on a short yarn specimen is often not representative of the yarn sample. It is especially problematic if the diameter measurement was taken on the fracture end of a specimen, outside of the tensile test specimen, or from another part of the yarn sample. Many CNT yarns and fibers have irregular cross-sectional shapes (e.g., yarns produced by liquid densification) and thus it is difficult to determine their diameter. Laser diffraction method is widely used for monitoring wire diameter [12]. This method has been used for measuring CNT yarn diameter [13– 15]. A laser diffraction system with multiple laser beams may be mounted on a tensile tester to measure the diameter of the tensile specimen at several points during tensile testing [14, 15] to allow the calculation of instantaneous stress and Poisson’s ratio. Fig. 7.2 shows how the diameter of twisted CNT yarns is affected by the strain applied to the yarns. The low twist yarn had an initial diameter of 33.1 μm. At 0.0536 axial strain, the yarn diameter contracted to 20.9 μm. This gives a huge Poisson’s ratio of 6.8, which is more than 20 times higher than common solid materials (Poisson's ratio is about 0.3). A very large disparity can appear when yarn diameter is used for calculating yarn tensile stress [14]. On the other hand, the diameter change of highly twisted yarns is much less dramatic, giving a Poisson’s ratio less than 1. Textile yarns, especially yarns spun from staple fibers, do not have well-defined cross-section boundaries although they are often approximated to a circular shape to simplify analysis. The variability of yarn cross section is associated with the random number of fibers along the yarn length, the irregularity of constituent fibers, and the imperfect condition in forming the yarn. Because porosity is inevitable in yarns produced from staple fibers, the application of a relatively small tensile load to a twisted yarn, which produces only a small change in the yarn length, can cause a significant decrease in yarn diameter. Generally speaking, it is difficult to make a precise determination of yarn diameter that is applicable to all downstream processes and specifications. For these reasons, textile technologists prefer to specify yarn size in terms of count number (traditionally used, expressed in length/ mass, e.g., in metric count, 1 Nm = 1 m/g) or linear density (preferred unit, expressed in tex, 1 tex = 1 g/km = 1 mg/m). Unlike diameter, linear density is an average value from a long length of yarn. CNT fibers and yarns are also porous and often have irregular and inconsistent cross-sectional shape within a sample. Measuring CNT yarn thickness by linear density has now become increasingly common. The ­linear

140

Carbon Nanotube Fibers and Yarns

(A)

(B)

(C)

(D) 9 8

1

7

0.8 0.6 Low twist yarn

0.4

High twist yarn

0.2 0

(E)

Poisson’s ratio

Relative yarn diameter (d/D)

1.2

6 5 4

Low twist yarn

3

High twist yarn

2 1 0

0

0.02

0.04

0.06

Axial strain

0.08

0.1

0

(F)

0.02

0.04

0.06

0.08

0.1

Axial strain

Fig. 7.2  SEM images of yarns under different levels of axial strain. (A) Low twist yarn at zero axial strain; (B) low twist yarn at 0.0536 axial strain; (C) high twist yarn at zero axial strain; (D) high twist yarn at 0.061 axial strain; (E, F) yarn diameter and Poisson’s ratio at different axial strain levels [14]. (Reprinted with permission from M. Miao, J. McDonnell, L. Vuckovic, S.C. Hawkins, Poisson’s ratio and porosity of carbon nanotube dry-spun yarns, Carbon 48 (10) (2010) 2802–2811.)

density of a CNT yarn in tex can be determined by weighing a length of the CNT yarn sample using a microgram balance [14]. Alternatively, a Vibrascope [16] may be used to determine the linear density of a relatively short yarn sample directly.



141

Carbon nanotube yarn structures and properties

7.1.3  Nanotube packing density, bulk density, and porosity If all nanotubes are perfectly straight and aligned in one direction (a bundle of parallel cylinders), they can lie next to each other and the nanotube assembly achieves the maximum packing density. Fig.  7.3 shows a cross-sectional view of parallel cylinders closely packed in a hexagonal array. Using the parallelogram “unit cell” [14], we can find that the proportion of unfilled area (including spaces inside the nanotubes), or the minimum porosity of the closely packed yarn, is

ϕmin

π 1 − (d /D )2  =1− 2 3

(7.1)

where D (nm) is the outer diameter of the CNT and d (nm) is the inner diameter of the CNT. If the nanotubes are treated as solid cylinders, i.e., d = 0, Eq. (7.1) gives the minimum porosity of 9.3%, or the maximum packing density of 90.7%. If the space inside the nanotube is counted as voids in the yarn, the minimum porosity will be higher. For example, when d/D = 0.4, the minimum porosity is 23.8% [14]. The maximum number of nanotubes that can be packed in a 1 μm2 cross section (nmax, tubes/μm2) can be calculated from nmax =

2 3 × 106 3D 2

(7.2)

Unit cell

Fig. 7.3  Close hexagonal packing of parallel cylinders [14]. (Reprinted with permission from M. Miao, J. McDonnell, L. Vuckovic, S.C. Hawkins, Poisson’s ratio and porosity of carbon nanotube dry-spun yarns, Carbon 48 (10) (2010) 2802–2811.)

142

Carbon Nanotube Fibers and Yarns

In practice, the nanotubes in CNT yarns are not perfectly aligned and closely packed, and thus the yarn porosity is greater and the nanotube packing is lower than the above values, respectively. The real yarn porosity can be derived from the bulk density of the CNT yarn (ρyarn) and the density of the constituent nanotubes (ρcnt) ρ yarn (7.3) ϕ =1− ρ =1− ρcnt where ρ is the CNT packing fraction of the yarn. The average yarn density ρyarn can be calculated from the linear density and the average cross-sectional area of the yarn. The density of CNTs is strongly related to the nanotube diameter and number of walls [17]. The number of walls has a strong relationship with the diameter for chemical vapor deposition (CVD) grown nanotubes [18]. Fiber packing fraction in an assembly is affected by both the compression and the alignment between the fibers.The pressure P required to compress randomly orientated elastic fibers into a structure with a fiber packing fraction ρ follows the well-known van Wyk power relationship [19]. P = kE ρ 3

(7.4a)

where E is the Young’s modulus of the fiber and k is a proportionality factor to be determined by experiment. In a structure comprised of perfectly aligned fibers, almost no pressure is required to lay the fibers next to each other to obtain the maximum packing fraction. In practice, fibers in aligned structures such as yarns are not perfectly straight (i.e., with some level of waviness or crimp) and there are local misalignment between fibers. In such a case, the van Wyk relationship can be modified to [20]. P = kE ρ n

(7.4b)

The exponent n is greater than 3 and can be as high as 15 for highly aligned glass fiber rovings (no crimp). As the value of fiber packing fraction ρ is always smaller than unity, it is easier to densify aligned fibers than misaligned fibers. For yarns formed from textile fibers, some level of fiber misalignment is essential to achieve a self-locking structure. Fiber misalignment can be introduced during yarn formation due to twist insertion, fiber migration [21], fasciation [22], felting [23], and interlacing [24]. The pressure between staple fibers and the resulting interfiber friction are the primary forces that give a textile yarn its tensile strength.



Carbon nanotube yarn structures and properties

143

Unlike conventional textile fibers, van der Waals forces between nanotubes are of high importance to the strength of unbonded CNT yarn. The magnitude of the van der Waals force depends strongly on the distance between the nanotubes, which is a function of nanotube packing fraction in the yarn. 7.1.3.1  Twisted yarns Generally speaking, CNTs are not uniformly distributed in CNT yarns. The local nanotube packing fraction in a twisted CNT yarn can vary considerably. Sears et  al. [25] showed that nanotube packing fraction decreases from the center to the peripheral in twisted CNT yarns, as shown in the SEM images taken from focused ion beam (FIB) sectioned yarns in Fig. 7.4A. When studying the porous structure of CNT yarns, two types of voids may be distinguished, i.e., voids between nanotubes in the same bundle and voids between the bundles, as shown in Fig. 7.4B and C. Due to

Fig. 7.4  Nanotube packing density distribution in twisted CNT yarns. (A) SEM image of CNT yarn cross section, showing radial density distribution [25]. (B, C) SEM images showing CNT bundles and pores between CNT bundles [26]. (Panel (A) reprinted with permission from K. Sears, C. Skourtis, K. Atkinson, N. Finn, W. Humphries, Focused ion beam milling of carbon nanotube yarns to study the relationship between structure and strength, Carbon 48 (15) (2010) 4450–4456.; Panels (B and C) reprinted with permission from D. Zhang, M. Miao, H. Niu, Z. Wei, Core-spun carbon nanotube yarnsupercapacitors for wearable electronic textiles, ACS Nano 8 (5) (2014) 4571–4579.)

144

Carbon Nanotube Fibers and Yarns

the strong van der Waals force, CNTs tend to form bundles during growth and processing. Undensified CNT webs drawn from a CNT forest (Chapter 2) have an extremely high porosity, estimated to be in the vicinity of 99.97% [27]. The porosity of a twisted CNT yarn decreases with the degree of twist inserted into the yarn, as shown in Fig. 7.5. A highly twisted CNT yarn can have a packing fraction as high as 0.6 (a porosity of 0.4) [14]. This is similar to the fiber packing fraction of a highly densified textile yarns [28]. If the twist in a twisted yarn is removed, the yarn diameter can increase substantially, as shown in Fig. 7.6.The change of yarn porosity is believed to be mainly due to the widening of the voids between CNT bundles while voids within a CNT bundle do not change so significantly [6]. 1

1.2

Forest 1 Forest 2

1

Forest 3

0.8

Forest 3

0.7

0.8 0.6

0.6 0.5 0.4

0.4

0.3

0.2

0.2

φ = 0.238

0.1

0

(A)

Forest 1 Forest 2

0.9

Porosity

Bulk density (g/cm3)

1.4

0

10 20 30 40 50 Surface twist angle (degrees)

5000 t/m

(C)

60

70

0

(B)

0

10

20 30 40 50 60 Surface twist angle (degrees)

70

15000 t/m

(D)

Fig.  7.5  CNT yarn bulk density and porosity. (A) Relationship between twist angle and yarn bulk density, (B) relationship between twist angle and yarn porosity [14], and (C, D) SEM images of FIB sections milled through CNT yarns with two levels of twist [25]. (Panels (A and B) reprinted with permission from M. Miao, J. McDonnell, L. Vuckovic, S.C. Hawkins, Poisson’s ratio and porosity of carbon nanotube dry-spun yarns, Carbon 48 (10) (2010) 2802–2811.; Panels (C and D) reprinted with permission from K. Sears, C. Skourtis, K. Atkinson, N. Finn, W. Humphries, Focused ion beam milling of carbon nanotube yarns to study the relationship between structure and strength, Carbon 48 (15) (2010) 4450–4456.)



Carbon nanotube yarn structures and properties

(B)

0.8

1

0.7

0.9 0.8

0.6 Yarn porosity

Yarn density (g/cm3)

(A)

0.5 0.4 0.3

0.7 0.6 0.5 0.4 0.2

0.1

0.1 0

0 0

5000 10,000 15,000 20,000 25,000 Twist or false twist (T/m)

Twisted Twist-untwisted Trend-twisted Trend-twist-untwisted

0.3

0.2

(C)

145

0

(D)

5000 10,000 15,000 20,000 25,000 Twist or false twist (T/m)

Fig. 7.6  (A, B) SEM images of twisted and twist-untwisted CNT yarns. (C, D) Relationship between twist/false twist and yarn density and porosity [6]. (Reprinted with permission from M. Miao, The role of twist in dry spun carbon nanotube yarns, Carbon 96 (2016) 819–826.)

7.1.3.2  Rub-densified yarns Rub-densified twistless yarn produced at different rubbing roller pressures showed different CNT packing density distributions across the yarn [29]. At low roller pressure, the yarn showed a high-density shell and a low-density core, as shown in Fig. 7.7A. Due to its very low density, we do not expect the core to make a significant contribution to the yarn strength. It was estimated that the core occupies about 40% of the total yarn cross-sectional area. The average density of the sheath could be estimated from the average yarn density (0.64 g/cm3) divided by the area proportion of the sheath (i.e., 0.6), giving a sheath density of 1.07 g/cm3. The low-density core can be eliminated by increasing the pressure between the rubbing rollers and by lowering the yarn tension, resulting in a ribbon-like yarn cross section with seemingly uniform CNT packing density (Fig. 7.7B). 7.1.3.3  Die-drawn yarns Cross sections of die-drawn yarns are shown in Fig.  7.8 [30]. The yarn diameter was controlled by adjusting the die diameter. As Fig. 7.9 shows,

146

Carbon Nanotube Fibers and Yarns

Fig. 7.7  SEM images of FIB milled rub-densified yarn cross sections. (A) Low-pressure rub-densified CNT yarn with a high-density sheath and a low-density core and (B) high-pressure rub-densified CNT yarn showing a uniform CNT packing density and a flat cross section [29]. (Reprinted with permission from M. Miao, Production, structure and properties of twistless carbon nanotube yarnswith a high density sheath, Carbon 50 (13) (2012) 4973–4983.)

30 µm

(A)

30 µm

(B)

300 nm

(E)

30 µm

(C)

300 nm

(F)

30 µm

(D)

300 nm

(G)

300 nm

(H)

Fig. 7.8  Die-drawn CNT yarns. SEM images showing cross sections of CNT yarns with diameters of (A, E) 30 μm, (B, F) 35 μm, (C, G) 55 μm, and (D, H) 75 μm, respectively [30]. (Reprinted with permission from K. Sugano, M. Kurata, H. Kawada, Evaluation of mechanical properties of untwisted carbon nanotube yarn for application to composite materials, Carbon 78 (2014) 356–365.)

with the increase of die diameter, which means less compression during die drawing, the density of the resulting yarn decreased. Note that the yarns produced using the two small diameter dies (30 and 35 μm) showed similar yarn density as the high twist CNT yarns in Fig. 7.5A.



Carbon nanotube yarn structures and properties

147

Fig. 7.9  Yarn diameter and density as a function of die diameter, based on the data from Ref. [30]. (Source: K. Sugano, M. Kurata, H. Kawada, Evaluation of mechanical properties of untwisted carbon nanotube yarn for application to composite materials, Carbon 78 (2014) 356–365.)

7.1.3.4  Liquid-densified fibers Twistless liquid-densified fibers often demonstrate irregular cross-section shape and the shape usually changes along the fiber length. This makes it difficult to estimate the fiber porosity and the nanotube packing density. Liquid-densified yarns show more uniform nanotube packing density in the yarn cross section than twisted yarns. Qiu et al. [31] characterized diameters of the nanotube bundles and the ­inter-bundle pores using the longitudinal section SEM images of as-spun fibers (Fig. 7.10A, B, and E). The distribution of the bundle diameters was found to be similar to the diameter observed for inter-bundle pore diameters (Fig. 7.10D and E). Fig. 7.10F shows that the distribution of bundle size peaked at about 20–30 nm, while that of the pores reached a maximum around 30–40 nm. Cho et  al. [32] measured nanotube distribution in the cross section of acetone-densified CNT yarns drawn directly from a floating catalyst CVD furnace (Fig.  7.11 and Table  7.1). The as-spun acetone-densified yarn had a density of 0.24 g/cm3 and a porosity of 0.84, which indicates a rather low densification in comparison with the twisted yarns in Fig.  7.5. When the acetone-densified yarn was further treated with solvent 1-methyl-2-pyrrolidinone (NMP) or chlorosulfonic acid (CSA), the yarn density was more than doubled (Table 7.1) and the yarn porosity decreased to the level comparable to the twisted yarns with a 40-degree twist angle [14]. Nanotube flattening was found with the NMP and CSA

148 Carbon Nanotube Fibers and Yarns

Fig. 7.10  SEM images of a CNT fiber cut using a focused ion beam (FIB) showing the longitudinal (A, B) and radial (C, D) cross sections of a solvent-densified CNT yarn produced by the floating catalyst chemical vapor deposition method. (E) A SEM image at a higher magnification of a FIB-cut longitudinal section of a CNT fiber showing the inter-bundle pores and CNT bundles, and (F) histogram of the lateral bundle dimensions and pore dimensions [31]. (Reprinted with permission from J. Qiu, J. Terrones, J.J. Vilatela, M.E. Vickers, J.A. Elliott, A.H. Windle, Liquid infiltration into carbon nanotube fibers: effect on structure and electrical properties, ACS Nano 7 (10) (2013) 8412–8422.)



Carbon nanotube yarn structures and properties

149

Fig. 7.11  TEM images of cross sections of CNT bundles showing tube-to-tube spaces in (A) acetone-densified CNT yarn, (B) further treated in NMP, and (C) further treatment of (A) in CSA, showing nanotube flattening. All scale bars: 10 nm [32]. (Reprinted with permission from H. Cho, H. Lee, E. Oh, S.-H. Lee, J. Park, H.J. Park, et al., Hierarchical structure of carbon nanotube fibers, and the change of structure during densification by wet stretching, Carbon 136 (2018) 409–416.)

treatments. The increase of yarn density was accompanied by an increase of bundle size from 12 to 40–60 tubes/bundle.The bundle size was further increased to 86 tubes with the application of a 13% wet stretch. However, the yarn density showed a decrease from 0.74 to 0.69 g/cm3 as the stretch ratio was increased from 7% to 13%. Wang et al. [33] used a pair of calendar rollers to compress an initially solvent-densified yarn. The initial yarn was directly produced from a floating catalyst CVD furnace by passing the CNT yarn through a water or alcohol bath. Up to five passes of calendaring action consolidated the wet yarn into a ribbon-shape with substantially increased yarn density, estimated to be 1.3–1.8 g/cm−3, which may be the highest value reported in literature.

7.1.4  CNT alignment CNTs packed in fibers and yarns are far from ideally parallel-laid cylinders. The wavy (crimpy) configuration, misalignment, and bundling of CNTs in the forest and the drawn web can be observed from the SEM images in Fig. 7.12 [14]. The CNTs in the original forest in Fig. 7.12A have crimps but most of them are not entangled with each other. During web drawing, the CNTs are turned 90 degrees from the vertical direction in the forest to the horizontal direction in the newly formed web (Fig. 2.1A, Chapter 2). During this process, the CNTs inevitably interfere with each other and form loops, hooks, reversals, and crossings that constitute an entangled CNT network, as shown in Fig. 7.12B. By comparing the configurations of the CNTs in the forest and in the drawn web, it is clear that the tension

150

As-spun NMP NMP (7% stretch) CSA CSA (7% stretch) CSA (13% stretch) a

Yarn linear density (tex)

Yarn crosssection area (μm2)

Between-bundle distance (nm)

Tubes per bundle

Yarn density (g/cm3)a

Yarn porositya

0.1 0.12 0.14 0.13 0.11 0.11

417 208 243 218 148 160

61.8 64.9 39.8 46.1 40 33

11.7 44.6

0.240 0.577 0.576 0.596 0.743 0.688

0.85 0.64 0.64 0.63 0.54 0.57

64.2 85.8 3

Calculated values from reported yarn linear density/cross-sectional area and a CNT density of 1.6 g/cm (i.e., nanotubes treated as solid cylinders).

Carbon Nanotube Fibers and Yarns

Table 7.1  Solvent-densified yarn morphology [32].



Carbon nanotube yarn structures and properties

151

Fig. 7.12  Tortuosity and misalignment of CNTs in forest and drawn web. (A) SEM image of CNTs in vertically aligned CNT array (forest) and (B) SEM image of CNTs in a drawn web [14]. (C) Schematic of misaligned straight fibers (longitudinal view) and (D) schematic of a void formed between misaligned straight fibers (end view of schematic C). (Panels (A and B) reprinted with permission from M. Miao, J. McDonnell, L. Vuckovic, S.C. Hawkins, Poisson’s ratio and porosity of carbon nanotube dry-spun yarns, Carbon 48 (10) (2010) 2802–2811.)

applied to the CNTs during web drawing removed some of the crimps present in the CNT.The van der Waals force and the entanglement between the CNTs in the drawn web give the web its strength for the dry-spinning process. Unlike conventional staple or short fiber textile webs such as those of wool or cotton, the CNT web does not draft or stretch easily once formed. Under sufficient tensile load, the CNT web will break sharply instead of drafting apart through slippage between individual CNTs. This is because the constituent CNTs are strongly connected to each other by van der Waals and other interactions (e.g., amorphous carbon bridging) and thus do not slip readily. The tortuosities (the hooks, reversals, and crimps) and misalignment of the CNTs observed in the web will thus persist into the yarn while the straight CNTs will take up the tension applied to the yarn.The presence of crimps, hooks, and misalignment keep the CNTs apart from each other, resulting in voids between them, as illustrated in Fig. 7.12C and D. Enabling drafting of CNT strands (webs, yarns) has been considered an effective strategy for improving CNT alignment and thus the strength of the strands. This usually requires the use of a lubricating liquid or an u ­ ncured

152

Carbon Nanotube Fibers and Yarns

resin, resulting in a CNT nanocomposite material [34–38]. In some cases, the polymer matrix may be removed after the alignment treatment to obtain a pure CNT structure. As shown in Fig.  7.13, a relatively low draft (