DESIGN AND FABRICATION OF FABRIC NEAR FIELD ANTENNA FOR WEARABLE APPLICATIONS

Transcription

1 DESIGN AND FABRICATION OF FABRIC NEAR FIELD ANTENNA FOR WEARABLE APPLICATIONS A thesis submitted to The University of Manchester for the degree of Master of Philosophy in the Faculty of Science and Engineering 2017 YUXI SHI SCHOOL OF MATERIALS

2 LIST OF CONTENTS LIST OF CONTENTS... 2 LIST OF TABLES... 6 LIST OF FIGURES... 7 ABSTRACT... 9 DECLARATION COPYRIGHT STATEMENT ACKNOWLEDGEMENT THE AUTHOR ABBREVIATIONS CHAPTER 1 INTRODUCTION Research Scope Research Gaps Research Aims and Objectives Research Methodology Research Value Outline of the Thesis CHAPTER 2 LITERATURE REVIEW Introduction Definition and Development of E-Textiles Definition of E-Textiles Current and Future E-Textiles Summary Conductive Materials Conductive Polymers Carbon-based Micro/Nano Materials Metallic Materials Summary

3 2.4 Fabrication and Integration Technologies Fabrication Technologies Integration Technologies Summary Devices and Applications Sensors Wearable Antennae Output Devices Summary Near Field Antennae Definition of Near Field Antennae Characterisation of Near Field Antennae Near Field Antennae Mode of Operation Summary Conclusions CHAPTER 3 MATERIAL PREPARATION AND ANTENNA DESIGN Introduction Material Preparation Experimental Design Instrument Material Experiment Test Result Antenna Design Resonant Frequency Inductance and Capacitance Magnetic Flux Density Quality Factor Reading Range

4 3.3.6 Geometric Dimensions Conclusions CHAPTER 4 FABRIC NFA FABRICATION PROCESS DEVELOPMENT Introduction Experiment Experimental Design Material Instrument Experimental Method Statistical Analysis and Discussion Factorial Design Experiment Process Development Discussion Conclusions CHAPTER 5 FABRIC NFA FABRICATION PROCESS OPTIMISATION Introduction Experiment Experimental Design Material Instrument Experimental Method Statistical Analysis and Discussion Steepest Ascent Central Composite Design Experiment Process Optimisation Discussion Conclusions CHAPTER 6 FABRIC NFA PERFORMANCE EVALUATION

5 6.1 Introduction Experiment Experimental Design Instrument Experimental Method Performance Evaluation Quality Factor and Reading Range Test Bending Test Conclusions CHAPTER 7 CONCLUSIONS AND FUTURE WORK Conclusions Future work Appendix І Appendix И REFERENCES The Final Word Count: 19,122 5

6 LIST OF TABLES Table 2-1 Patent search with the keywords of e-textiles (smart textiles, intelligent textiles, wearable textiles, smart clothing) and antennae. 29 Table 2-2 Electrical properties of normal metallic conductive wires Table 2-3 Comparison of different integration technologies Table 3-1 Conductivity and resistivity of conductive yarns Table 3-2 Conductivity of conductive yarns and metallic wires Table 4-1 Design factors and levels Table 4-2 The factorial design experiment Table 4-3 Descriptive analyses for process development Table 4-4 Factorial fit: resistance versus NLS, SD, ST Table 5-1 The central composite design for resistance Table 5-2 Descriptive analyses for process optimisation Table 5-3 Response of surface regression: resistance versus NLS, SD Table 6-1 Group set for results measurement Table 6-2 The test results for the quality factor and the reading range

7 LIST OF FIGURES Figure 1-1 Thesis structure and contents Figure 2-1 The products of e-textiles Figure 2-2 Citation report from Web of Science for the keyword of e-textiles (smart textiles, intelligent textiles, wearable textiles, smart clothing) Figure 2-3 Carbon-based micro/ nano materials Figure 2-4 The process to produce metallic fibres Figure 2-5 The diagram of stitch geometry Figure 2-6 Smart textile devices and applications Figure 2-7 NFA applications Figure 2-8 Citation report from Web of Science with the keyword of near field antennae Figure 2-9 NFA modes of operations Figure 3-1 The instruments and testing process Figure 3-2 The instruments for testing the width of yarns Figure 3-3 The instruments for testing the resistivity Figure 3-4 Digital image of conductive yarns Figure 3-5 Digital images of Statex conductive yarns Figure 3-6 Equivalent circuit of the NFA Figure 3-7 Rectangular antenna coil Figure 3-8 Transformer principle of the NFA Figure 3-9 Equivalent circuit of NFA reader antennae and tag antennae. 73 Figure 3-10 Simulation model of the NFA Figure 3-11 The NFA reader Figure 3-12 Sewn NFA Figure 4-1 The sewing machine Figure 4-2 The multimeter and the supply power

8 Figure 4-3 The fabrication process development experiment Figure 4-4 The resistance measurement experiment Figure 4-5 Main factors plot for resistance Figure 4-6 Interaction plot for resistance Figure 4-7 Contour plot of resistance vs SD, NLS Figure 4-8 Surface plot of resistance vs SD, NLS Figure 4-9 Contour plot of resistance vs ST, NLS Figure 4-10 Surface plot of resistance vs ST, NLS Figure 4-11 Residual plots for resistance for process development Figure 5-1 The fabrication process optimisation experiment Figure 5-2 Steepest ascent experiment Figure 5-3 Central composite design in the coded variables Figure 5-4 Contour plot of resistance vs SD, NLS Figure 5-5 Surface plot of resistance vs SD, NLS Figure 5-6 Contour plot of resistance Figure 5-7 Residual plots for resistance for process optimisation Figure 6-1 The FieldFox vector network analyser Figure 6-2 The NFA reader Figure 6-3 The experimental method for testing the reading range Figure 6-4 The experimental method for testing the quality factor Figure 6-5 The experimental method for evaluating the performance of the NFA under different bending conditions Figure 6-6 Contour plots for performance evaluation Figure 6-7 Line chart of resistance vs the reading range and the quality factor Figure 6-8 The centre frequency under different bending conditions

9 ABSTRACT E-textiles, also known as smart clothing or electronic textiles, are textiles or garments in which electronic devices and other wearable electronics, such as wiring and sensors, are embedded. E-textiles have received a lot of research attention in the past fifteen years for their use in wearable applications and usually refer to any electronic components that can be fixed on textile substrates. However, as e-textiles, very little research has been conducted to develop fabric near field antennae in literature. To achieve fabric near field antennae for wearable application requirements today in terms of comfort, function, sensing and ergonomics, this work introduces and realises a complex design and engineering methodology of fabric near field antennae using sewing technology. This research is completed by the appropriate conductive material preparation, the antenna design and fabrication, the process development and the optimisation, and the performance evaluation. In the stage of the design and fabrication processes, the near field antenna is designed and fabricated using sewing technology. The formulae for the parameters for a near field antenna are derived. Specifically, the relationships between the resistance and the quality factor (Q value) as well as the reading range are proved. In the stage of the development and optimisation processes, the main factors and the cross effects for the electrical performance of the conductive yarns are examined by the full factorial design experiment. Then the regression equation is established to simulate the performance of the near field antenna and the optimum resistance is calculated by the central composite design experiment. The results show that the quality factor (Q value) and the reading range increase accordingly with the decrease of resistance of the conductive yarns. The simulation results through process development and process optimisation are in correspondence with the experimental results qualitatively. Moreover, the near field antenna can work under different bending conditions. Thus, an optimum fabric near filed antenna for wearable applications has been designed and fabricated. 9

10 DECLARATION No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning. 10

11 COPYRIGHT STATEMENT The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the Copyright ) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the Intellectual Property ) and any reproductions of copyright works in the thesis, for example graphs and tables ( Reproductions ), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property University IP Policy (see in any 11

12 relevant Thesis restriction declarations deposited in the University Library, The University Library s regulations (see and in The University s policy on Presentation of Theses. 12

13 ACKNOWLEDGEMENT First and foremost, I would like to give my sincere gratitude to my chief supervisor, Prof. Henry Li, and my co-supervisors, Dr. Xuqing Liu, Dr. Jiashen Li, for their valuable guidance and constant support throughout my research studies. Prof. Henry Yi Li is an excellent supervisor who always encourages me to solve the problems and provides useful guidance for me. I would like to give my profound gratitude to Dr. Zhirun Hu, who provided me with excellent advice regarding electronics and antennae as well as allowed me to access his research facilities, and also to Mr. Kewen Pan, who helped me a lot in respect of antennae design and testing during my postgraduate research. Their professional advice and patient instruction helped me to improve myself. I would also like to give my particular thanks to my mother and my father for their constant love, support, and understanding. 13

14 THE AUTHOR During the academic period between September 2015 and April 2017, I majored in materials at the School of Materials, where I mastered the essential skills for research. While attending the University of Manchester, I have worked as a research student under the supervision of Prof. Henry Yi Li, Professor of Textiles Science and Engineering. In the process, I have assisted Prof. Henry Yi Li with the EU Horizon 2020 project with the project code ETEXWELD - H2020-MSCA-RISE I have also published two journal papers and one conference paper. Refereed Journal Article: 1. X. Liu, R. Guo, Y. Shi, Y. Li. Durable, Washable and Flexible Conductive PET Fabrics Designed by Fibre Interfacial Molecular Engineering. Macromolecular Materials and Engineering, 2016, 301(11), pp Y. Shi, Y. Li, Z. Hu, X. Liu. Advances in Smart E-Apparel Technology. Journal of Fibre Bioengineering and Informatics, Accepted. Conference Presentation and Publication: 1. Y. Shi, Y. Li, Z. Hu, X. Liu, Advanced in Smart E-textile Apparel Technology, Textile Bioengineering and Informatics Symposium Proceedings, Melbourne, Australia, July, 2016, pp

15 ABBREVIATIONS Ag Silver Au Gold CF Carbon fibre CI Confidence interval CNT Carbon nanotube CP Carbon particle Cu Copper ECG Electrocardiogram ELD Electroless deposition EMG Electromyogram HF High frequency HTG Hydrothermal growth H 2 S Hydrogen sulphide LED Light emitting diode NFA Near field antenna(e) Ni Nickel NLS Number of layers of stitching PAN Polyacrylonitrile 15

16 PANI Polyaniline PN Polynaphthalene PPy Polypyrrole PVA Polyvinyl alcohol PVDF-TrFE Polyvinylidenefluoride-tetrafluoroethylene RF Radio frequency RFID Radio frequency identification SD Stitch density spc Stitches per centimetre ST Stitch tension Stdv Standard deviation TPU Thermoplastic polyurethane WO 3 Tungsten oxide ZnO Zinc oxide 16

17 CHAPTER 1 INTRODUCTION 1.1 Research Scope E-textiles, also known as smart clothing or electronic textiles, are textiles or garments in which that electronic devices and other wearable electronics, such as wiring and sensors, are embedded [1]. E-textiles have received a lot of research attention in the last fifteen years for their use in wearable applications [2] and usually refer to any electronic components that can be fixed on textile substrates [3-4]. At the early stage of smart clothing development, however, electronic components were bulky, heavy, disassembled, obtrusively attached to clothing, and connected with wire cables [5]. It is not comfortable and convenient to users and cannot realise all the functions of e-textiles perfectly [6]. To achieve the required performance of wearable applications, new areas in wearable applications need to be considered. Metallic interconnects in a circuit board could be created by ink-jet printing and photolithography [7]. However, these advanced printing techniques are not good enough for textile substrates, because of abrasion, wicking, and roughness. To meet the needs in terms of comfort, function, sensing, and ergonomics, the main improvement of this work conducted is to unobtrusively integrate electronic sensors and natural fibre textile structures by embedding the sensors 17

18 into a fabric circuit with different integration techniques using sewing machines, embroidery machines, knitting machines, welding machines or glue [8-13]. E-textiles not only include integrated circuits on or into a textile substrate but, nowadays, electronic devices in the form of fibres or strips which can be sewn into textiles or garments directly [14]. The conductive yarns or threads of the textile can act as interconnects by weaving, knitting, embroidering, or welding techniques and can be applied to connect components in a garment as antennae [15]. Antennae are electrical devices that transform electric power to radio waves or convert radio waves into electric power [16]. The near field antenna (NFA) is a short-range, radio-frequency technology that enables smart electronic devices to establish wireless connective communication with each other by bringing them into close together [13, 17]. A fabric NFA is an NFA that is constructed on a fabric substrate using different integration technologies. After they were invented in the 100 years ago [18], e-textiles have been realised as the low-cost, flexible, and comfortable wearable applications to improve human activities and daily lives, and have acquired wide applications in health care and biomedical intervention, fashion design and home products, and military and industrial applications [17]. The most noticeable advantage of e-textiles lies in their seamless construction, which does not interrupt garments 18

19 usual function [9]. 1.2 Research Gaps Nowadays, the NFA for e-textiles is attractive. After reviewing previous literature in Chapter 2, however, it is found there are still some research gaps to be filled to fulfil the requirements of real applications. The remaining issues are summarised as follows. 1. Conductive materials have been broadly applied in e-textiles. However, there has been no systematic and comprehensive research for comparing and selecting suitable conductive materials for the fabric NFA in literature. 2. Although sewing technology has been used for decades and has been seen as a familiar daily technology, very little research has been conducted to develop the fabric NFA using sewing technology. 3. Many types of research on e-textiles have focused on antennae. However, currently developed antennae were tough and uncomfortable, such as those fabricated by etching or printing technology, while very little research has been conducted to develop the highly integrated fabric NFA for wearable applications. 4. Most researchers have focused on material selection and antenna design, but there is a lack of research on parameters of sewing methods, which have significant effects on the reading range and the quality factor (Q value). 19

20 1.3 Research Aims and Objectives This research aims to develop the fabric NFA for wearable applications. On the basis of the research scope, a series of research gaps has been identified as reported above. To fill these research gaps and address the key issues involved, the specific objectives of this research are as follows. 1. To compare and select the suitable conductive materials through a literature review; to calculate the conductivity and resistivity of conductive materials; to characterise the selected conductive yarns. 2. To establish a multi-disciplinary framework of the fabric NFA for wearable applications using sewing technology, which illustrates the knowledge, techniques, and logical relationships that are involved in achieving the fabric NFA. 3. To design and fabricate the fabric NFA using sewing technology; to realise the flexible and comfortable properties for wearable applications; to establish a novel integration method for the fabric NFA. 4. To investigate the influence of the various parameters on the electrical performance of the fabric NFA; to evaluate the main factors and the cross effects of resistance; to find the optimum resistance of the fabric NFA in order to decrease the energy loss, improve the working efficiency, and extend the reading range. 20

21 1.4 Research Methodology In order to achieve the research aims and objectives of this research, the following research methodology is performed in five stages. 1. In the initial stage, a literature review will be conducted for the general consideration of metallic yarns. The requirements of e-textiles will be specified in terms of the resistivity and conductivity of conductive materials. Material selection and preparation to achieve the above aspects will be further reviewed and conceived. After determining the suitable materials, the related electrical performances will be calculated. 2. In the second stage, based on the research objectives, the evaluation and identification of the sewing technology will be discussed according to the flexibility and comfort properties for wearable application requirements. The formulae of the quality factor and the reading range will be derived and the related data will be tested for the subsequent design of the fabric NFA. 3. In the third stage, the fabric NFA for wearable applications will be designed and fabricated using sewing technology. The methods for manufacturing the fabric NFA will be put forward using metallic yarns. 4. Fabric NFA developed in previous stages, with desired functional requirements, will be integrated by means of sewing techniques. The experimental data will be analysed using SPSS and Minitab statistical analysis 21

22 software. The main factors and the cross effects for electrical performances of metallic yarns will be examined by using the full factorial design experiment. The regression equations will be established to simulate the performances of the fabric NFA and the optimum resistance will be calculated by using the central composite design experiment. 5. Finally, as for the last stage, the research conclusions, limitations, and future work will be presented. 1.5 Research Value The significance of this project is the design and fabrication of the fabric NFA for wearable applications. Through previous studies, many types of NFA based on a variety of conductive materials have been investigated. The outstanding advantages of the fabric NFA are flexibility, high conductivity, low cost, durability, and stretchability. However, how to realise the fabric NFA through advanced integration technologies has not been systematically studied. This proposed research aims to fill this knowledge gap and develop scientific understanding and engineering principles for the design and fabrication of the fabric NFA. Furthermore, the development and optimisation processes of the electronic performances by changing the sewing methods will be adopted. The optimisation results of the fabric NFA will be found in order to decrease the energy loss, improve the working efficiency, and extend the reading range. 22

23 1.6 Outline of the Thesis Figure 1-1 Thesis structure and contents. Chapter 1 presents the research scope, research gap, research aims and objectives, research methodology, and research value adopted for this work. Chapter 2 comprises a critical literature review in multi-disciplinary areas, which illustrates the knowledge, techniques, and logical relationships that are involved in achieving the fabric NFA. Chapter 3 presents the conductive materials preparation and the fabric NFA design for wearable applications. 23

24 Chapter 4 and Chapter 5 discuss the process development and the process optimisation for the fabric NFA to decrease the energy loss, improve the working efficiency, and extend the reading range. Chapter 6 presents the performance evaluation for the fabric NFA under normal and bending conditions. Chapter 7 concludes this work and offers recommendations for future work. 24

25 CHAPTER 2 LITERATURE REVIEW 2.1 Introduction This chapter will present a literature review in five parts. Firstly, the definition of e-textiles, as well as their developmental processes will be introduced; the second part will introduce a comprehensive review of conductive materials that are used in e-textiles; the third part will introduce the fabrication and integration technologies; the fourth part will introduce the wearable applications and devices; the last part will introduce the definition, characterisation, and operation of the fabric near field antenna (NFA). 2.2 Definition and Development of E-Textiles Definition of E-Textiles Figure 2-1 The products of e-textiles. (a) Flexible embroidered electronics [19]; (b) Printing process for wearable applications [20]; (c) A schematic diagram of the woven textile with the conductive yarns [21]; (d) A schematic diagram of the knitting textiles with the conductive yarns [22]. 25

26 The term e-textiles, also known as smart clothing or intelligent textiles, refers to a broad field of products where electronic components are integrated into the fabrics and garments. Thus, yarns can be integrated into the textile structure as seen in Figure 2-1 by embroidering [23], sewing [24], knitting [25], nonwoven textiles [26], weaving [27], printing [28], braiding [29], and chemical treatments [30], which can build an intimate form that senses and responds to environmental stimuli [31]. The environment stimuli comprise physical, thermal, mechanical and chemical responses e.g., changes in temperature, pressure, or force [32]. According to a different degree of sensing, reacting and integrating, e-textiles can be divided into three subgroups: passive smart textiles, active smart textiles, and proactive textiles [32-33]. Passive smart textiles are only able to sense the environment; active smart textiles can sense and react to stimuli from the environment; and proactive textiles can sense, react and adapt their behaviour to the given circumstances based on sensors and actuators [32, 34-35]. Sensors are essential elements for passive smart textiles, as they provide a system to react to the signals from sensors. They can provide sensing properties of physical reactions and signals, including capacity, resistivity, optical and solar properties [36]. The actuators are able to detect the signals autonomously from their environment [37], together with the sensors, which 26

27 have been mentioned above, they are the basic parts in active smart textiles. For example, the biomedical sensors clothing, which is integrated with the electrocardiogram (ECG) [38], electromyography (EMG) [39], and electroencephalography (EEG) [40] or the shape-sensitive sportswear that can sense the environment and detect the specific muscle movements [41]. Active smart textiles include radio frequency (RF) technology [42], power generation and harvest devices [43], and human interface elements [44]. Active smart textiles usually require power, which can be acquired from piezoelectrical devices [45], while harvest devices can be achieved through photovoltaic elements [46]. However, NFA are merely conductive patterns that are sewn using conductive yarns of specific sizes and shapes and do not require any storage devices or power supply devices, so it s a simple application of smart textiles and can be realised [47]. Based on the research of previous e-textile products, proactive textiles are the expected products at the current time and in the future [48]. For example, smart shape memory sportswear will help athletes to monitor the physiological parameters to provide coherent physical health care; as well as adjust the clothing shape based on individual customers to optimise the training efficiency and security [49]. The most noticeable advantage in proactive textiles is the automatic adjustments according to different environmental conditions or different individual abilities [33]. 27

28 2.2.2 Current and Future E-Textiles Figure 2-2 Citation report from Web of Science for the keyword of e-textiles (smart textiles, intelligent textiles, wearable textiles, smart clothing) [50]. E-textiles have shown wide potential applications in recent years, especially since When typing e-textiles as the keyword in Web of Science, there are 1142 key papers and more than a half were written within the last ten years as seen in Figure 2-2 [50]. Furthermore, by searching related patents via Google scholar patent database, 2480 related patents from different areas can be found, such as communications, medical monitoring, military where e-textiles are designed to perform sensing and communication functions and support intelligent activities [51]. There are also wearable applications in everyday garments in the aesthetic and functional aspects. Developing e-textiles attracts considerable interest from fashion designers, costumers, venture capitalists and scientists and also requires a multi-disciplinary approach involving textile technology, electronic engineering, fibre science, 28

29 and human-computer interaction [13]. Table 2-1 Patent search with the keywords of e-textiles (smart textiles, intelligent textiles, wearable textiles, smart clothing) and antennae [51]. Patent Patent No. Date Applicants Electrically conductive metal composite embroidery yarn and embroidered circuit US /08/2013 Tae Jin Kang et al. Printable, flexible and stretchable diamond for thermal management US 12/418,071 25/06/2013 John A. Rogers et al. Conformable antennae using conducting polymers WO A1 16/02/2012 Ian W. Hunter et al. Method of printing electronic systems on textile substrates WO A1 12/09/2014 Davide Zanesi Textile material comprising a circuit module and an antennae US A1 21/08/2008 Anatoli Stobbe et al. Antenna stent device for wireless, intraluminal monitoring US A1 14/04/2005 Yogesh Gianchandani et al. Textile fabric with integrated electrically conductive fibres and clothing US A 25/05/1999 Michael S. Lebby et al Communication In recent years, personal information electronic devices including computers and mobile phones, are becoming more and more important in daily life and 29

30 no one can live without information technologies [52]. The tendency is towards mobility and comfort: people want personal electronic devices that are more light weight, thin profile, and unobtrusive [53]. Flexible and comfortable e-textiles integrated with a computer or mobile phone will have a huge market if the price is reasonable and acceptable [54] Medical Monitoring An evolution in healthcare in modern society has led to increased health consciousness as well as medical monitoring [55]. At the same time, the rising population of elderly people in developed countries is bringing a really heavy burden to bear on government and hospitals. In this case, wearable medical monitoring applications will provide an alternative solution for daily healthcare and personal conditions monitored [56]. For example, the wearable Cardioverter Defibrillator by Hulings RJ et al. [57] designed a chest band, which can provide emergency aid for taking medicine to people prone to heart attacks. Another example of an Electrocardiograph (ECG) is that invented by the Advanced Medical Corporation [58], which allows elderly people to monitor their personal data at home with data being sent to the doctors and nurses through the wireless connection. Moreover, for some people lost their senses like hearing or touch accidently, the e-textiles could help alleviate the suffering of such people in the form of artificial skins or artificial muscles [59-60]. 30

31 Military Application The military functional clothing is important for soldiers in combat, as they have to protect themselves during wars and conflicts through the functional clothing [54]. If e-textiles can be adopted in the military, soldiers would be connected to satellite systems via flexible miniature computers or other related electronic devices [61]. These devices would not only let them know their positions but also can provide some monitored data to help them avoid emergency accidents [62]. Most of the e-textiles are expensive, initially they may be only affordable by the government and the military where performances are in deep demand [63] Fashion Application In affluent societies, e-textiles have a competitive edge in fashion applications, especially in stage art. The first example of such an application was the Philips/Levis ICD+ that integrated mobile phones and music players into clothing [64]. After that, Nokia developed a snow jacket with a mobile phone and a wireless Bluetooth headset [65]. France Telecom tried to put an MP3 player on the collar [17]. Until now, various companies attempt to place mobile phones, Bluetooth headsets, light-emitting diode (LED) lights on the collars and other areas of clothing [66-68]. E-textiles can offer people an opportunity to create unique clothing, identifying and memorising different characterisation. Furthermore, the colour of clothing would detect the people s 31

32 mood, temperature, feelings, and even health conditions [69-70]. The customers, however, may focus more on the cost and quality, and require a high demand for easy care [71]. More importantly, e-textiles should be washable under standard washing conditions and ideally waterproof [72] Summary E-textiles that are integrated electronic devices into common textiles via knitting, weaving, sewing, or printing, have shown wide potential applications in recent years, especially since It is to be respected that e-textiles develop rapidly in various areas, including communications, medical monitoring, military applications, as well as fashion applications. The antenna can realise information transmission and data collection, and has attracted considerable interest in the above areas with regard to wearable applications. 2.3 Conductive Materials The conductive materials are essential parts for e-textiles based on the electrical, mechanical and chemical properties of conductive materials [73]. Conductive polymers, carbon-based micro/nano materials, and metallic materials are the most common conductive materials, in fibre form or as coating. They have been used widely in wearable applications such as sensors, actuators, transistors, flexible electronic devices, and field emission displays in textile systems [32, 74-76]. 32

33 2.3.1 Conductive Polymers If the electronic devices could be integrated into a fibre base, then the conductive fibre would provide an excellent fundamental unit for e-textiles, for example, as they can directly and naturally realise high electronic performance [77]. Conductive polymers are also called poly conjugated polymers, which are the organic materials that have single or double bonds in their chains, and are the most promising conductive materials that can integrate advanced electronic functionalities into fibres [32, 78]. The commonly used conductive polymers are polyaniline (PANI) and polypyrrole (PPy) [79]. There is a variety of ways to manufacture the conductive polymers. For example, PANI nanofibres are synthesised by polymerisation of aniline [32, 80]; PPy nanofibres are prepared in the presence of p-hydroxy-azobenzene sulfonic acid as a functional dopant [32, 81]. Conductive polymers have conductivity and water solubility at room temperature so that they were seen as ideal polymers that could be integrated into wearable applications using fabrication technologies. However, the high cost and the small conductivity (only S/cm) of conductive polymers limited their effective and wide usage in e-textiles. Thus, conductive polymers are only used in components of electronic devices or solar cells [32-33, 82]. 33

34 2.3.2 Carbon-based Micro/Nano Materials Figure 2-3 Carbon-based micro/ nano materials. (a) PVDF/ NaNbO 3 conductive fabric electrodes [83]; (b) Asymmetric (hybrid) ECs [84]. Carbon-based micro/nano materials, due to the unique properties including desirable intrinsic carrier mobility (10 6 cm 2 V -1 S -1 ) [85], offer superior conductivity (10 4 S/cm) [86], low cost, and good mechanical properties (elastic modulus in the order of 1 TPa) [87], have been frequently employed in e-textiles for many years. Nowadays, carbon-based micro/nano materials (e.g., carbon particles (CPs), carbon nanotubes (CNTs), carbon fibres (CFs), and graphene), have been wide used in fibre-based sensors, microelectrodes, and actuators [32, 83]. For instance, CP is a good commercial e-textiles material because of their high conductivity and low cost; CNTs can be used for flexible energy-storage because of the large specific surface area and the high electrical conductivity; CF is a good substrate for industrial production because of their high conductivity [32, 88-92]. However, the black colour of the traditional carbon materials may influence the aesthetics of e-textiles. The brittle nature of the carbon-based micro/nano materials is also a problem for 34

35 electronic devices [92] Metallic Materials Metallic materials are promising materials for flexible and wearable electronic devices. In e-textiles, the metal is directly used as a fibre or coated on traditional fibres as nano-materials [32, 93]. For first category, stainless steel yarns and copper fibres are the most common conductive materials; for the second category, the normal yarns, e.g., cotton yarns and nylon yarns would be plated with a metal film, such as gold, silver, copper, iron, aluminium, to acquire conductivity [94]. Thus, the metallic materials are introduced into e-textiles because of the high conductivity, low cost, light weight, thin profile, and high mechanical flexibility [95]. For example, polyacrylonitrile (PAN) textiles are coated with magnetic nanoparticles in coating baths with nickel (Ni), copper (Cu); metal conductors, such as copper, fabricated by catalytic printing; Silver (Ag)-plated Silicon fibres can show conductivity exceeding 470 S/cm; Kevlar plated with nickel (Ni) and gold (Au) reaches 6 S/cm; meanwhile, silver (Ag) coated nylon fibre can display electric conductivity values of 1800 S/cm [32, ]. So far, the metallic materials are one of the most adaptable technologies [32, 101]. The applications in the industry have been hindered, however, because of their problems in roughness, haze, and instability. Further development is being explored on improving the stability, durability and utilisation efficiency of metallic materials [83]. 35

36 Table 2-2 Electrical properties of normal metallic conductive wires [102]. Conductive materials Electrical conductivity (S/m) Copper wires ! Silver wires ! Gold wires ! Stainless steel wires ! Nickel wires ! Platinum wires ! Summary Conductive materials are fundamental for functional e-textiles. It is impossible to design e-textiles without conductive materials. Because of the limited conductivity in conductive polymers and the brittle nature of carbon based micro/nano materials, the metal coated yarns, which offer high conductivity, low cost, light weight, thin profile, and great mechanical flexibility are considered to be the ideal materials for this work. 2.4 Fabrication and Integration Technologies Fabrication and integration technologies are reported in two parts: one refers to the preparation of conductive substances, e.g., conductive yarns and conductive inks; the other is about various ways to integrate conductive substances with garments or clothing, e.g., weaving, knitting, sewing, and printing [103]. 36

37 2.4.1 Fabrication Technologies Conductive Yarn Fabricating conductive fibres is an effective approach to integrate electrical devices into fibres and an important way make the common textiles acquire a desirable interface layer for high conductivity [104]. Conductive fibres can be wide used in military garments [105], antistatic applications [106], medical apparel [107], and protective clothing [108]. Figure 2-4 The process to produce metallic fibres. (a) Metallic wire combined in iron tubes; (b) Reductions; (c) Bundling; (d) Leaching and realising [109]. The traditional processes to produce metallic fibres can be divided into four parts, namely metallic wire combined in iron tubes, reductions, bundling, and leaching and realising, as shown in Figure 2-4 [109]. Another fabrication method is coating conductive matter on traditional fibres no matter whether natural or synthetic fibres. After coating, the fibres would maintain the flexibility and stretchability. The coating technologies include 37

38 electrochemical plating, physical vapour deposition, chemical vapour deposition [ ], and electroless deposition (ELD) [112]. Surface modifications, like introducing ionic bonds power is essential for the coating process to improve the durability and washability of the fibres [113]. The key question of the surface modification is whether scientists can improve the production process under laboratory conditions without a vacuum environment to reduce the cost [101] Conductive Ink E-textiles can be produced and realised with conductive inks [114]. Conductive inks always contain a highly conductive metallic precursor, such as silver, copper, nickel, and gold [115]. Furthermore, the water-based carrier is needed so the materials can be printed onto the textile substrates [116]. Thus, the key issue during the ink-making process is to improve the electrical and mechanical properties of the conductive inks to enhance the performances of the respective products [117] Integration Technologies Integration technologies play a significant role in determining the physical properties (stability, durability, stretchability, flexibility, etc.), environmental stability, and the cost of the mass production of e-textiles as the conductive fabrics are the foundations of e-textiles [83]. Generally, various approaches to make conductive fabrics can be categorised into two groups. The first method 38

39 is to integrate conductive yarns made from conductive polymers, metallic materials, carbon-based nano materials, into textile structures by weaving and knitting [118]. The other method, that is complementary to the first category, is to attach conductive layers or patterns onto conventional fabrics or garments directly as the surface flexible electronic devices by sewing or printing [119] Weaving Weaving involves using a loom to interlace two perpendicular sets of yarns or threads at 90 degrees to form a fabric. Longitudinal yarns are named warp yarns and the lateral yarns are named weft yarns. During the weaving process, the warp yarns are moved up or down using a harness. This process creates many different weave structures and provides various functions and performances of fabrics [22, 32, 120]. In e-textiles, the conductive yarns can be used to replace some of the normal yarns and realise the conductive properties. The woven fabrics are characterised by the durability and dimensional stability [32, 121]. Recent publications [ ] have shown conductive threads and yarns that have integrated electronic devices into a loose textile structure via the weaving process. More specifically, the conductive threads or yarns can be used to replace the weft direction of the threads and use common threads as the warp direction. For example, Post et al. studied woven metallic organza and Edmison et al. developed piezoelectric materials for wearable electronic 39

40 textiles [124]. Researchers from Hong Kong Polytechnic University integrated the X-axis and Y-axis grids of Cu wires into woven textiles to realise conductive performance [125]. In addition to those two-dimensional textiles, three-dimensional textiles have become more popular in e-textiles. Three-dimensional textiles consider the thickness (Z-axis) relative to the planar (X-axis and Y-axis) dimensions, providing a variety of potential chances for yarn spacing in a three-dimensional space and fabricating more complicated composite three-dimensional fabrics. For instance, the Georgia Tech Wearable Motherboard used this method to reduce the need to connect wires during producing, cutting, and sewing to develop an uncut clothing [32, ]. This approach has not only resulted in the excellent conductive performance of e-textiles but also provided a complex network to have multiple layers and spaces to insert circuits and sensors [22] Knitting Similarity, knitting is a method by which yarns are manipulated to create interlocked loops into consecutive loops or stitches on needles. Knitted textiles consist of several consecutive rows of interlocking loops [32, 129]. By designing the sequence of knitting methods, complex patterns can be realised by using conductive yarns or threads in a commercial manufacturing knitting machine [32, 130]. Knitted fabrics have good elasticity, thermal retention, and humidity transport properties that are ideal for e-textiles [131]. Furthermore, 40

41 similarly to three-dimensional weaving fabrics, circular knitting technology allows the creation of three-dimensional seamless tubular smart clothing [32, 132]. Smart knitted textiles are made possible due to the above advantages. For example, Cottet et al. [132] acquired four types of geometrical fabric structures through knitting conductive yarns, which offer high frequency (HF) properties up to 6 GHz. Jost et al. [133] developed textile supercapacitors based on knitting technology and the capacitance can reach 0.50 F cm -2. Integrating wearable devices under the fibre base is recommended in order to maintain essential textile properties (e.g., durability, comfort). The flexible conductors would be integrated (woven or knitted) with normal yarns into an electric functioned textile substrate. However, integrating electronic devices at fibre level may meet with some difficulties during the process of production, as the conductive parts of clothing need to be designed before it is manufactured [33]. Besides, the fabric inserted flexibility and comfort performance may be influenced if electronic devices are rigid or thick. Despite researchers are studied over this problem persistently, most of the current electronic devices cannot fully conform to the normal fibre or fabrics [13] Sewing Traditionally, sewing technology was sewn decorations in or on textile fabrics and other surfaces (e.g., leather, plastic), through the application of yarns, 41

42 cords, and sequins [32, 134]. Gradually, sewing technology was used in mechanical engineering [135], wound dress [136] and e-textiles [137] because of the flexibility, washability, comfort, and aesthetic properties of the sewn textiles [32]. The normal process of sewing in e-textile comprises four steps: conductive yarns production, pattern design, digitisation, and pattern sewing [32, 138]. Figure 2-5 The diagram of stitch geometry. (a) The common types of stitch geometry; (b) The diagram of lockstitch geometry [139]. Sewn Radio Frequency Identification (RFID) has been investigated and produced. The RFID is the process by which items are uniquely identified using radio waves, while the NFA represents a specialized subset within the family of RFID technology. Specifically, the NFA forms a branch of High-Frequency (HF) RFID [140]. There is related research on sewn RFID. For instance, Loughborough University used sewing and related 42

43 manufacturing techniques to design wearable antennae [23]. Moradi et al. [141] designed sewn dipole-type radio frequency identification (RFID) tag antennae based on the effect of the stitch density of sewn RFID in Vena et al. [142] designed and produced stretchable sewn chipless RFID tags and sensors as well as investigated the main parameters that affect the working efficiency, including the stitch tension and the number of layers of sewing. In addition, Koski et al. [143] optimised the e-textiles ground planes for wearable RFID patch tag antennae based on the direction of the yarns, the number of yarns, the stitch geometry, and the stitch density. A diagram of stitch geometry is shown in Figure 2-5 [139]. By reviewing the related literature, the major factors that will influence the performance of e-textiles are the number of layers of sewing (NLS) [144], the stitch density (SD), and the stitch tension (ST) [145]. More specifically, the NLS equates to the number of repeated layers for sewing the same pattern. The SD is the number of stitches per unit length. The ST is the tension level that influences the degree of tightness and the situation of loops [ ]. Sewn electronic devices are one of the most suitable techniques for wearable applications due to their straightforward integration with clothing and low cost to realise electrical functionality [146]. Compared with knitting and weaving methods, sewing technology in e-textiles can be used to produce high-accuracy patterns and can be used on completed garments directly [147]. 43

44 Furthermore, sewing methods are recommended to sew conductive patterns on ready-to-wear garments directly. Based on the above advantages, sewing technology has become one of the popular techniques in e-textiles. The above-stated main parameters, along with their cross effects that will influence conductivity, need to be further explored Printing Printing technology, i.e., screen printing, inkjet printing, and dip-coating, has been used widely in the textile industry [148]. Printing is a process for copying text and images using a master form or template. In e-textiles, printing technology using conductive solutions is used for producing electronic devices. The conductive solution must contain an appropriate high conductive metal precursor, such as copper, silver, or nickel, to create conductive patterns [32, ]. Compared with other methods, the most noticeable advantage of printing is that it facilitates mass production under normal conditions [32-33]. The screen printing technique offers an easy way to adopt the fabrication of smart electronic devices, as the conductive layer would directly be printed onto the surface of the fabrics and does not require any photolithographic or chemical etching processes [151]. For example, the active electrodes for ECG monitoring have been realised by screen printing [152]. Compared with screen printing, inkjet printing technology has a much higher spatial precision of conductive ink droplets [153]. Metal, CNTs, and 44

45 graphene-based conductive inks are developed for inkjet printing [154]. Dip-coating cannot be ignored in printed electronic devices. For example, polyamide yarns can be coated with polyvinyl alcohol/ carbon nanotubes (PVA/CNTs) via dip-coating [155] Summary The most noticeable advantage of knitting, weaving, sewing, and printing technologies is portability, flexibility, and comfort compared to traditional electronic devices. However, integrating sensors and circuits at fibre level may have difficulties during the process of production, as the conductive parts of clothing need to be designed before it is manufactured. Besides, the fabric inserted flexibility and comfort performance may be influenced if electronic devices are rigid or thick. Compared with knitting and weaving methods, sewing technology in e-textiles can help produce high-accuracy patterns and can be used on completed garments directly. Furthermore, the sewing method is inexpensive in terms of mass production compared with printing. The comparison of different integration technologies is summarised in Table

46 Table 2-3 Comparison of different integration technologies. Advantages Weaving Knitting Sewing Printing Relatively straightforward production No No Yes Yes Adverse influence on conductivity Yes Yes Yes No Adverse influence on flexibility Yes Yes No Yes High accuracy No No Yes Yes Mass production No No Yes Yes Low cost Yes Yes Yes No 2.5 Devices and Applications Figure 2-6 Smart textile devices and applications. (a) Mamagoose pyjama developed by Verhaert [13]; (b) Smart shirt with sensors developed by WEALTHY [13]; (c) Smart jackets with LED lights [156]. Typical examples are electronic textiles or garments, such as sensors, wearable antennae, and output devices, as shown in Figure 2-6, which can change their conductivity and resistivity when the environmental impact changes. 46

47 2.5.1 Sensors Sensors are the most mature and successful electronic devices that can measure physical quantities (e.g., temperature, pressure, and force) and convert them into specific signals that can be stored and analysed. Fibres may acquire sensing signals via electrically conductive materials. In e-textiles, sensors are often inserted into bio-monitoring clothing to monitor the physiological parameters in humans to detect changes in the environment [32, ]. Because of its application areas, the most important features of sensors are a compromise between sensitivity, durability, and environmental stability [32, 160]. Polypyrrole (PPy) was the first conductive material that was printed onto polyamide 6 and Lycra fabrics to fabricate resistive fabric strain sensors, which can detect the posture and gesture of people [161]. Xue et al. [162] developed strain sensors made from PPy coated fibres. Compared with traditional strain sensors, the sensitivity of the strain sensor made from PPy coated fibres is much higher. However, the stability and durability of PPy strain sensors would influence the working efficiency and future development. Carbon nanotubes (CNTs) and graphite flakes on natural rubber substrates can produce flexible and stretchable high-strain sensors, which are ~5 and ~12 times greater than conventional PPy strain sensors [163]. However, the potential toxicity of the CNT sensors or multiwall CNT sensors could cause 47

48 health problems when used in wearable applications. In this context, Zhang et al. [164] studied thermoplastic polyurethane (TPU) copolymers and carbon particle (CP) elastomeric polymer composite fabric strain sensors, which can show good recoverability and durability after being washed with cyclic loading. It is worth noting that advanced high-strain sensors need to have high sensitivity as well as reproducibility after several washing cycles. Similarly, pressure sensors are used to identify muscle movement [165] and body posture [113] by measuring modulation parameters including resistance, voltage, light intensity, and wavelength. The piezoresistivity of graphene/polyurethane has been developed in pressure sensors having linear response in a range of pressures from 0.26kPa to 200kPa [166], with stable and recoverable piezoresistive properties. It is possible to develop wearable applications such as pressure sensors, strain sensors, and artificial skin, in respect of stretchability, flexibility, resilience, and magnetic properties. Piezoelectricity from polyvinylidenefluoride-tetrafluoroethylene (PVDF-TrFE) co-polymer film can be used to detect pressure [167]. These pressure sensors have light weight, mechanical flexibility, robust and chemical resistance, and can be commercially mass produced [16]. Wearable applications can also function as chemical sensors, such as humidity sensors and temperature sensors [29]. The ideal materials for chemical sensors are the materials sensitive to different environmental stimuli (e.g., humidity, 48

49 temperature, and ph), such as PPy, zinc oxide (ZnO), or graphene oxide [168]. Su and Peng [169] adopted polypyrrole and tungsten oxide (PPy/WO 3 ) nanocomposite films to fabricate H 2 S gas sensors at room temperature by in-situ photopolymerisation method, the improved PPy/WO 3 has a stronger response at a low concentration of H 2 S gas than pure PPy or WO 3 film. Tsai et al. [170] studied ZnO nanosheet humidity sensors at room temperature by the hydrothermal growth (HTG) method. The sensing response has a good sensing range from 12% to 96% RH at a fast feedback speed at 600ms. Borini et al. [171] explored graphene oxide based humidity and temperature sensors with a variety of advantages such as flexibility, stretchability, transparency and durability for mass production. The main problem of existing chemical sensors is that they are sensitive to several environmental stimuli rather than one ideal response, which may influence their applications in industry [172] Wearable Antennae The antenna is an electrical device that transform electric power to radio waves or convert radio waves into electric power [16]. Antennae usually consist of arrangements of metallic conductors and electrical connections to receivers or transmitters [173]. Radio waves are electromagnetic waves that can carry information through space at the speed of light and without losing transmission. The first antenna was built by the famous German physicist called Heinrich Hertz in 1888, and thus proved the existence of 49

50 electromagnetic waves [174]. Nowadays, there is a rapid development of wearable textile applications in wireless communications with health monitoring functionality integrated into garments. A wireless communication device between the receiver and the transmitter requires the antenna to transmit information and collect data. Traditional wearable antenna is made of metal patterns or copper wires on circuits and is wide used in the area of communication, military, and personal healthcare [156]. Wearable antennae are in the form of patch antennae because of their advantages of high integration, miniaturisation, and excellent working efficiency [175]. The upper conductive layer, the lower conductive layer, and the dielectric substrate in the middle ground plan are the fundamental compositions of patch antennae [176]. Furthermore, the microstrip feedline is an essential part for patch antennae. Related research for wearable antennae has been found on the main types of conductive yarns, the stitch density, the types of woven patterns or knitted patterns, and the dielectric properties of textile fabrics [177]. Apart from this, the flexible and stretchable wearable antenna that was fabricated from conductive materials and elastomeric yarns has caused concern. Further studies will focus on durability, washability, flexibility, and bending of wearable antennae [173]. This kind of wearable antennae can be 50

51 attached to different parts of garments to measure the specific data or can be used for wireless transmission [178]. Wearable antennae require conductive fabrics as well as dielectric textile substrates. The resistivity of conductive fabrics should normally be lower than 1Ω/m 2 [16]. Coated conductive fabrics may not be as good as the woven/knitted conductive fabrics that are fabricated by conductive yarns. Furthermore, woven conductive fabrics perform better than knitted conductive fabrics due to the fact that they have higher geometrical accuracy [178]. As for dielectric textile substrates, the dielectric constant and its thickness will influence the bandwidth and working efficiency of wearable antennae. Based on previous studies, this may be affected by the moisture absorption and the thickness change when bending or stretching [179]. Wearable antennae, however, still face some challenges in practical applications. The biggest problem is that the antennae should increase the comfort of wearability which can be done by reducing mining sizes and adopting flexible materials [180]. It is easy to understand people would move when wearing e-textiles, so the wearable antennae under different bending, stretching and abrasion conditions need to be reported [181]. While the traditional wearable antenna is rigid, which would become permanently deformed, or even break when moved or stretched [182]. Therefore, wearable antennae require good conformality, flexibility, light weight and are subject to 51

52 continuous stress caused by body movements [32, 183] Output Devices Output devices are used to communicate changes in the textile system with the ambient environment. In most cases, e-textiles are interconnected with traditional display devices (e.g., coloured LEDs, optical fibres, and watch faces) [32, 44]. In e-textiles, transistors are another typical output devices [ ]. Nowadays, various products to integrate transistors into textiles have been developed [13]. Furthermore, artificial muscles and artificial skins are other examples of textile output devices that have attracted considerable attention [32, 59-60] Summary The antenna is a passive device that require no external batteries, just uses magnetic induction coupling between the tag and the reader within the certain distance. Wearable antennae are in the form of patch antennae because of their advantages of high integration, miniaturisation, and excellent working efficiency. 52

53 2.6 Near Field Antennae Definition of Near Field Antennae Figure 2-7 NFA applications [186]. The near field antenna (NFA) is a short-range, radio-frequency technology that enables smart electronic devices to establish wireless connective communication with each other by bringing them into close proximity (within 10cm distance) [13, 17]. The maximum transmitting speed of the NFA can be optimised up to about 424 kbps with ASK modulation [92]. It can be used for both-way communication over a small distance (less than 10cm) and works by magnetic field induction. The NFA operates within the RF band of MHz [187]. The most important advantage of this contactless reader is that the NFA is a passive communication tag that allows battery saving implementations [188]. This technology was first proposed by the Philips company and gradually was popularised by Nokia and Sony companies [189]. Figure 2-7 shows the applications for NFA technology [186]. 53

54 Figure 2-8 Citation report from Web of Science with the keyword of near field antennae [190]. When typing near field antennae as the topic in Web of Science, there are approximately 2622 results and most of them have been written within the latest five years as shown in Figure 2-8. Moreover, there is no related patent in the fabric NFA when searching patents via Google scholar [191]. That is to say, there s a research gap in the fabric NFA for wearable applications Characterisation of Near Field Antennae The NFA go by the acronym NFA, which is a set of standards for mobile phones or other electronic devices by placing them into close proximity (within 10cm distance) [192]. The NFA is just like b or n for WIFI, which set the protocols to send as well as receive signals [193]. The NFA is used widely, for example, swiped wireless payment (such as ApplePay for Mcdonald s), information exchange at a certain distance (such as touching smartphones to send and receive information or data), as well as simplified setup of electronic devices (such as Wifi and Bluetooth) [194]. Furthermore, 54

55 communication is also a possibility between the NFA and the unpowered NFA chips, named NFA tags [195]. Clearly, the NFA simply represents a way of communicating with each other between two electrical devices at a short distance [196]. The NFA works at approximately MHz frequencies. The corresponding wavelength is approximately 22 metres [197]. However, the radiating structures limit the NFA design, and the maximum linear distance is about 0.5% of a wavelength, that means 10 cm [198]. So the NFA operates at low frequency (large wavelengths) and the radiation efficiency will approach zero. Hence, the NFA is more like the big inductor, not the real antenna. The NFA requires the mutual coupling to complete the contactless energy transfer when it works [199] Near Field Antennae Mode of Operation Figure 2-9 NFA modes of operations [200]. Based on its characterisation, the NFA has become one of the most common 55

56 electronic devices nowadays. Specifically, the NFA modes of operations can be divided into three parts, namely card emulation, card writer/reader, and peer-to-peer, as shown in Figure 2-9 [198] Card Emulation Mobile phones with the NFA can act as a contactless card or a tag in card emulation mode. This mode can be used in existing traditional card readers based on payment systems [201]. For example, this mode is used as an identity card to enter the office or home or use as contactless payment at the point of sale machine in stores [202]. The NFA works as a gateway to transfer data and information from the card applications on the mobile phones to the receiving devices but doesn t carry out any computation [203] Reader/Writer The NFA can read data and related information via mobile phones or smart cards. At the mean time, the mobile phone can be used in a writer to input data [204]. For example, the body health data can be acquired and read via mobile phones containing NFA if the related sensors have been connected to the NFA [205] Peer-to-Peer The peer-to-peer mode can exchange information and data between two mobile phones. One can act as the sender and the other can act as the receiver 56

57 [ ]. For example, two business partners can transfer their card information with each other via their NFA inserted into mobile phones by placing them close to each other [208]. Furthermore, the technologies of Wi-Fi or Bluetooth can be used to set up connections at the same time [209] Summary Through the above introduction, it is found that the NFA are one of the most typical and wide spread antennae. They can achieve contactless information transfer without any batteries, which helps to create light weight and comfortable e-textiles. However, very little research has been conducted to develop the fabric NFA. Based on the advantages as well as the research gaps, this work plans to fabricate the fabric NFA for wearable applications. 2.7 Conclusions Through the literature review, a clear image about the fabric NFA in wearable applications has been acquired. The conductive materials and fabrication technology are fundamental for functional e-textiles. Metallic materials are suitable for fabricating wearable applications and are usually selected by researchers because of their perfect performance properties, such as high conductivity, low cost, light weight, thin profile, and great mechanical flexibility. Compared to other methods, sewing technology is the ideal technique to 57

58 realise wearable applications, especially in ready-to-wear clothing. The most considerable advantage of sewing technology that has been mentioned in literature reviews is the convenience for producing high accuracy conductive patterns directly on textile substrates. Moreover, this technology helps to acquire flexible and comfortable properties, which are important for e-textiles. Furthermore, It is good for mass production as it saves a lot of money and time. Hence, this work will use sewing technology to realise wearable applications. As for electronic devices, the fabric NFA is chosen as an ideal production. They are not big and can achieve contactless information transfer without any batteries. Passive antennae help acquire light weight e-textiles. However, there are no related fabric NFA using metallic coated yarns and sewing technology. To achieve the perfect performance and functionality of the fabric NFA, Chapter 3 will focus on the design and fabrication of the fabric NFA using metallic coated yarns and sewing technology. 58

59 CHAPTER 3 MATERIAL PREPARATION AND ANTENNA DESIGN 3.1 Introduction Material preparation and antenna design is one of the most important parts to achieve the high electrical conduction property of the fabric NFA. The conductivity and resistivity of metallic yarns, and the geometric dimension of the fabric NFA influence the fabrication process of the fabric NFA. Specifically, the quality factor and reading range are the most important factors that affect the working efficiency of the fabric NFA. Through the literature review, the definition, characterisation, and operation for the fabric NFA have been discussed in Chapter 2. This chapter will focus on material preparation and antenna design based on accommodated knowledge. The goals of this chapter are to make a clear plan for the fabrication the fabric NFA and find the key factor that will influence the quality factor and the reading range of the fabric NFA. 3.2 Material Preparation Experimental Design To investigate the most suitable conductive yarns that can be used in the later sewing process, three different conductive yarns are tested in this part. The conductivity and resistivity of conductive yarns are tested and calculated. 59

60 3.2.2 Instrument Figure 3-1 The instruments and testing process. (a) Materials and instruments of testing resistance; (b) Digital image of measuring resistance. Figure 3-2 The instruments for testing the width of yarns. 60

61 Figure 3-3 The instruments for testing the resistivity. (a) Analysis part of the machine; (b) Testing part of the machine. The resistance of conductive yarns was measured using a multimeter (type number: U1251A) and a power supply (type number: E3641A), which were produced by the Voltcraft Co. Ltd, Germany. The width of conductive yarns was determined using an optical microscope (type number: ), which was produced by the Projectina Co. Ltd, USA. The resistivity of conductive yarns was measured using a resistivity tester (type number: 4200-SCS), which was produced by the Keithley Co. Ltd, USA Material Conductive Yarn 61

62 Figure 3-4 Digital image of conductive yarns. (a) Statex; (b) Bekaert; (c) HK. The scale bar is 200µm. Silver high conductive yarns (2-ply) were purchased from the Statex Co. Ltd, Germany. Stainless steel conductive yarns (1-ply) were obtained from the Bekaert Co. Ltd, Belgium. Silver conductive yarns (2-ply) were obtained from Hongkong and they are pictured in Figure Textile Substrate The textile substrate used in this work featured the composition of 1:1 plain woven fabric with 50% Meta-aramid and 50% FR Lenzing fibre Experiment Resistance Four-point probe resistance measurement as shown in Figure 3-1 was used to measure the resistance of the conductive yarns [210]. The electric current I (A) was set to 0.100A. After acquiring a constant voltage, the measured voltage U (V) could be read through the multimeter (at 45%RH, 21.5 C) and the resistance R (Ω) of conductive yarns could be calculated as seen in equation 3.1 [210]. 62

63 R =!! (3.1) The test samples were measured five times. Then, the mean and the standard deviation of resistance could be calculated Average Width The magnification of the optical microscope was chosen to measure the width of the conductive yarns as shown in Figure 3-2. After obtaining a clear image on the computer screen, the average width d (mm) of the conductive yarns could be multiplied by the measured width L (mm), measured times n, and magnification m. Each sample was tested at least five times. The width of the conductive yarns could be expressed in the form as seen in equation 3.2 [211]. d =! m (3.2)! Surface Resistivity One of the basic characteristics of conductive yarns is surface resistivity ρ (Ω m). According to the digital image of Figure 3-4, the conductive yarns are not circular, which means that the calculated resistivity is not accurate. The surface resistivity could be measured using a resistivity tester machine (type number: 4200-SCS), which was produced by the Keithley Co. Ltd, USA, as shown in Figure 3-3. The test samples were measured five times. Then, the mean and the standard deviation of surface resistivity could be calculated [212]. 63

64 Conductivity Another basic characteristic is conductivity σ (S/m) of conductive yarns characterising their ability to conduct electrical current. S is siemens,1s=1/1ω. The conductivity σ (S/m) could be calculated by surface resistivity ρ (Ω m) as seen in equation 3.3 [213]. σ =!! (3.3) Test Result Conductivity According to the results, it can be seen that the resistance of Statex conductive yarns is much little than HK conductive yarns. Besides, Statex 2-ply conductive yarns are suitable for sewing than Bekaert 1-ply conductive yarns. The conductive yarns that can be used in this work are Statex 235/34 dtex 2-ply (Statex, USA) [214]. The yarns are made from polyamide 6.6 filaments yarn plated with 99% nanoparticle-sized pure silver. The conductivity of the yarns is ! S/m. A sewing technique was used to apply the conductive yarns to the textile substrate with a specifically designed pattern. 64

65 Table 3-1 Conductivity and resistivity of conductive yarns. Yarn Yarn Width Stdv Resistance Stdv Resistivity Stdv Conductivity source properties (mm) (mm) (Ω) (Ω) (Ωm) (Ωm) (S/m) Statex Silver coated !! !! ! !! !! !! !! ! ! Bekaert HK Stainless steel Silver coated !! !! Table 3-2 Conductivity of conductive yarns and metallic wires. Source Experiment Handbook [102] Materials Conductivity (S/m) Statex yarns ! Bekaert yarns HK yarns ! ! Copper wires ! Silver wires ! Gold wires ! Nickel wires ! Microscopy Figure 3-5 Digital images of Statex conductive yarns. (a) With 400 magnification; (b) With 5000 magnification. To visualise the structure of Statex conductive yarns, the yarns were examined 65

66 using a scanning electron microscopy (SEM) (type number: SU8220), which was produced by the Hitachi Co. Ltd, Japan. Figure 3-5 shows the SEM microscopy image of the surface (top view) of the conductive yarns with 400 and 5000 magnifications. The visible appearance of the conductive yarns shows that these yarns have an average width of 370 µμm (standard deviation 10 µμm). It can be seen that the conductive yarns have a large scale of silver coating on the surface smoothly and densely. 3.3 Antenna Design Resonant Frequency Figure 3-6 Equivalent circuit of the NFA [213]. Figure 3-6 shows the equivalent circuit of the NFA. L! (H) is the coil inductance, R! (Ω) represents the losses in the coil, C! (F) is the sum of the 66

67 coil capacitance and resonance capacitor which force the loop coil antenna to work around MHz. Normally, the resistance 𝑅! (Ω) for a metal coil is extremely small [215]. However, the conductivity of conductive yarns is approximately hundreds of times smaller than copper wires, as shown in Table 3-2 [102]. The impact of the yarn on the NFA s large parasitic resistance not only affects the magnetic field strength but also affects the antenna resonant frequency and its quality factor. The quality factor (Q value) is a dimensionless parameter in physics and engineering and is the physical quantity of the vibrator damping property and the resonant frequency of the oscillator with respect to the size of the bandwidth [216]. The input impedance 𝑍!" (Ω) can be represented as equation 3.4 [215].!! 𝑍!" = (!!!!"!! )(!!!! )!!!!!"!!!!!!!!!!!!!!!! =!!!!!(!!!!!!)!!!!!!! (3.4) Assuming the antennae works in a resonance state, the imaginary part should be cancelled, thus:!! 𝑗! =𝑗!!!!!!!!!!!!!!!! (3.5) Thus:!!!!!! =!!!!!!!!!!!!!!!!!!! =! 𝜔!! 𝐿!! 67 (3.6) (3.7)

68 ω!! =!!!!!!!!!!! (3.8) Because the angular velocity ω (rad/s) equals [217]:!!! =!! (3.9) Thus, the resonant frequency f (MHz) can be found as in equation f! =!!!!!!!!!!!!! =!!!!! 1!! (3.10)!!!!!!!! The formula of resonant frequency of a NFA is different from the metal component due to the larger parasitic resistance. The larger resistance decreases the resonant frequency [215]. Overall, the parasitic resistance must be reduced by using different sewing methods [ ] Inductance and Capacitance Figure 3-7 Rectangular antenna coil. Parasitic inductance and coil inductance are the most important factors that affect the impedance matching and the radiation efficiency. Calculation of the 68

69 antenna inductance is complex. However, here are some simplified formulae that use certain assumptions to get an estimation of the inductance [218]. Also, these assumptions must be considered when choosing different inductor shapes. In this work, an approximation method for a rectangle coil is selected as shown in equation 3.11 [219]. L! = N!!! 2 W + h + 2! h! + W! h ln!!!!!!!! W ln!!!!!!!! + h ln!"! + W ln!!! (3.11) Where L! (H) is the parasitic inductance, N is the number of turns, μ! is the magnetic permeability, which is a constant value for the same conductive yarns, W (mm) is the average width of the rectangle, and h (mm) is the average height of the rectangle, and a (mm) represents the yarns width. Thus, L! (H) is decided by the geometric size of the antennae. C! (F) is the sum of the coil capacitance and the resonant capacitor which force the loop coil antenna to work around MHz, which consisted of two parts, the first part is the coil capacitance, this is quite small and can be ignored [220]. The second part is the resonant capacitor, which depends on the antenna frequency, which is a constant value [221] Magnetic Flux Density For direct current analysis of the fabric NFA, a model with pure resistances gives a close resistance value of the sewing lines. For radio frequency (RF) 69

70 analysis of the fabric NFA, a simple model cannot be established because the current cannot flow on the surface. Hence, the skin effect and the coupling effect between the conductive yarns are obvious although in the high frequency (HF) band. Moreover, the conductivity of conductive yarns varies a lot depending on the sewing methods [ ], and bending also changes the antenna s conductivity and geometry [222]. Many studies show that a simple model of a sewn strip can be adopted in simulation software [223]. A sewn strip can be approximated by an effective bulk conducting strip with the same size as the pattern. For the NFA, energy and signal transmission depend on the magnetic fields that are produced by the NFA reader. The parasitic resistance has significant influences on the magnetic field strength. Assuming that the conductive yarn is a long, straight wire, due to Ampere's law [224]: B =!!!!!!" (3.12) Where B (T) is magnetic flux density, μ! is the magnetic permeability, which is a constant value, I! (A) is the current in the wire, and D (m) is the distance between the wire and the point. Due to ohm s law as seen in equation 3.1, the magnetic flux density is inversely proportional to the resistance of the conductive yarns [224]. The fabric features show quite significant electric current losses when the conductive yarns go through the fabric, however, it is little, so the overall conductive losses will be mostly affected by the resistance 70

71 of the conductive yarns [225] Quality Factor The quality factor (Q value) is another important factor that must be discussed for the NFA. The quality factor is a dimensionless parameter in physics and engineering and is the physical quantity of the vibrator damping property and the resonant frequency of the oscillator with respect to the size of the bandwidth [216]. A large quality factor Q always leads to a narrow bandwidth and a small quality factor Q reduces the mutual inductance and the reading range [226].!"#$%!"!#$%&'()*!$+#!"!#$%!"!"#!$"%!"#$%!"#$%&%'" Q = 2π!"#$%&!"##!"!"#!"#$%&%'"!"!#$ (3.13) The total electromagnetic energy W ω! (W) in the circuit under resonance is [216]: W ω! = T! I!! R! (3.14) Where the W (ω! ) (W) is the total electromagnetic energy, T! (s) is the period time, R! is the resistance, I! (A) is the current that would flow through an equivalent resistance. The energy loss E ω! (W) in this circuit can be represented as [216]: E ω! =! L!!! 2I! =! L! I! (3.15) Where L! (H) is the inductance. Hence, the quality factor for the fabric NFA can be represented as in equation 3.16 [216]. Q = 2π!!!!!! = 2π!!!!!! =!!!!!! =!!!!!!! 1 (3.16) For a normal metal NFA, small parasitic resistance leads to a high quality 71

72 factor. A resistor that is series connected to the NFA is needed to decrease the quality factor and increase its bandwidth. Higher quality factor and inductance values will still function but with a reduced range and performance. But for the fabric NFA, parasitic resistance is always dozens of ohms. So the quality factor is always under 20 (the best quality factor tradeoff [227]). Thus, it is necessary to increase the quality factor. Based on equation 3.11, it is figured out that the L! (H) and C! (F) are decided by the geometric size of the NFA rather than the coil electronic properties. In other words, if this work wants to acquire higher quality factors, the resistance of the fabric NFA should be decreased Reading Range Figure 3-8 Transformer principle of the NFA [226]. The reading range is the distance between the tag antenna and the reader antenna under working conditions [228]. Assuming both the reader antenna and the tag antenna are circular, the reader transmitting coil generates an 72

73 electromagnetic field with a frequency of MHz as in Figure 3-8. A small part of the emitted field penetrates the tag antenna, which is some distance away from the reader antenna. Simultaneously, voltage is generated in the tag antenna by its inductance. Figure 3-9 Equivalent circuit of NFA reader antennae and tag antennae [226]. The amount of magnetic flux that penetrates through the tag antenna is affected by the coil itself, the substrates, and the geometrical quantity. Figure 3-9 gives the equivalent circuit of the above structure [229]. L! (H) and L! (H) in the left side of Figure 3-9 are the self-inductance of the reader antenna and the tag antenna. C! (F) is the resonant capacitor which forces the loop coil antenna to work around MHz. The parasitic resistance of conductive yarns is equivalent to R! (Ω), which is connected to the antenna in parallel. A typical parallel resonance circuit consists of these components. Going a step further, the left side circuit can be transformed into an air-core transformer model in the right hand of Figure 3-9. The mutual 73

74 inductance M (H) is represented as follows [230]. M = K L! L! (3.17) Where M (H) is the mutual inductance, K is the coupling coefficient which is a constant value, L! (H) and L! (H) are the coil inductance of the reader antenna and the tag antenna, which are decided by the geometric sizes of the NFA which has been certificated in equation From the transformer model, this work is able to find the amplitude of the output voltage that is produced on the tag antenna [230]. V!"# = ω M I! (3.18) Where V!"# (V) is the generated output voltage, ω (rad/s) is the angular velocity (constant value), M (H) is the mutual inductance, I! (A) is the current flowing through the tag side. It is necessary to make the output voltage as large as possible. Every NFA chip has a threshold voltage. If the NFA output voltage is not higher than the chip s threshold voltage, the chip will not activate. In other words, the higher output voltage increases the reading range. Furthermore, M (H) is a constant value for a NFA, as M (H) is decided by the geometric sizes of the NFA, as certificated in equation I! (A) should be increased if it is wished to increase the output voltage of the tag antenna. Due to ohm s law as seen in equation 3.1 [224], the only way to increase I! (A) is to reduce the resistance 74

75 in the circuit and the NFA. So, optimising the sewing methods is significant Geometric Dimensions Figure 3-10 Simulation model of the NFA. Figure 3-11 The NFA reader (type number: TRF7970AVM). The geometric dimension and the images of the fabric NFA are shown in Figure The sizes of the fabric NFA are decided by a NFA reader (type 75

76 number: TRF7970AVM), which was produced by the Texas Instruments Co. Ltd, USA, as seen in Figure As can be seen, the outer length of the NFA is 70 mm and the outer width of the NFA is 50 mm. The conductive yarn width (the dark part with four yarns in parallel) is 1.48 mm. The gap between the two lines is 2 mm. Figure 3-12 Sewn NFA. (a) Sewing process; (b) Prepared fabric NFA. After completing the antenna design, the samples were sewn using a sewing machine (type number: LS2-190), which was produced by the Mitsubishi Co. Ltd, UK, under satandard laboratory conditions of 20% relative humidity and 20 C room temperature. The fabrication process and prepared fabric NFA can be seen in Figure Conclusions In this chapter, the material preparation and the antenna design have been discussed. The initial NFA has been prepared in terms of conductive yarns and derived formulae. During the process, a hypothesis that the resistance is the 76

77 main factor influencing the performance of the NFA is shown. Then this work will pay more attention to process development and process optimisation for the NFA to acquire the simulated NFA to improve the quality factor and extend the reading range. 77

78 CHAPTER 4 FABRIC NFA FABRICATION PROCESS DEVELOPMENT 4.1 Introduction Material preparation and antenna design have been reported in Chapter 3. It has been found that the resistance is the main factor that influences the quality factor and the reading range, which are the main parameters for influencing the performance of the fabric NFA. Thus, these two chapters will focus on reducing the resistance to improve and optimise the fabric NFA. The development and optimisation part for the fabric NFA will give a lot of help for acquiring more efficient and effective wearable applications. Based on the formulae derived in equation 3.16 and equation 3.18, it can be hypothesised that the resistance is the main factor that influences the working efficiency of the fabric NFA. More specifically, the lower resistance increases the quality factor and extends the reading range. There are numbers of different factors in the fabricating process that influence the resistance of the fabric NFA. For example, the conductive path will be widened through increasing the number of layers of sewing and thereby reducing the resistance by introducing the parallel connections. The stitch density, in other words, the number of stitches per unit length, will influence the resistance because of the differences between the conductive yarns parallel to the textile substrates and going through the fabrics. Moreover, the stitch tension will influence the 78

79 degree of tightness and the situation of loops, which may affect the resistance of the fabric NFA. However, the specific sewing methods of fabricating the fabric NFA have not been adequately studied or researched in previous literature. Therefore, it is necessary to investigate the influence of various stitch structural parameters on the performance of the fabric NFA. The next two chapters will pay attention to reduce the resistance of conductive patterns based on changing the sewing methods, in order to make the fabric NFA work over a longer distance with a higher accuracy. This chapter will focus on the influences of the main factors and their cross effects on the resistance of conductive patterns; defining an initial regression equation of the resistance and analysing influencing factors for the identification of major parameters. 4.2 Experiment Experimental Design To investigate the influence of different sewing methods (the number of layers of sewing, the stitch density, and the stitch tension) on the resistance of conductive patterns, and the realisation of the development progress through the 2! experimental design, a full factorial experiment, is designed to test the predictions. 79

80 4.2.2 Material The conductive yarns used in this work are Statex 235/24 dtex 2-ply conductive yarns (Statex, USA). The textile substrate used in this work is Composition: 50% Meta-aramid and 50% FR Lenzing fibre 1:1 plain woven fabric. The materials have been discussed in Chapter Instrument Figure 4-1 The sewing machine (type number: LS2-190). All fabric samples were sewn using a sewing machine (type number: LS2-190), which was produced by the Mitsubishi Co. Ltd, UK, under standard laboratory conditions of 20% relative humidity and 20 C room temperature. 80

81 Figure 4-2 The multimeter (type number: U1251A) and the power supply (type number: E3641A). The resistance of the conductive patterns was measured using a multimeter (type number: U1251A) and a power supply (type number: E3641A), both of them were produced by the Agilent Co. Ltd, USA. Four-point probe resistance measurement as shown in Figure 3-1, was used to measure the resistance of the conductive patterns [210] Experimental Method In this chapter, the process development is used to investigate the main factors for resistance and their cross effects as well. In each complete trial or replicate of the fabrication process development experiment, all possible combinations of the levels of the factors are investigated. The whole factorial design experiment can be divided into five steps as illustrated in Figure

82 Figure 4-3 The fabrication process development experiment. Step 1: observation the samples of the experiment. Step 2: developing a reasonable assumption to explain the observations. Step 3: using the assumption to make predictions. Step 4: doing the experiment to test the predictions. Step 5: coming to the conclusions or modifying the assumption. 82

83 4.3 Statistical Analysis and Discussion Factorial Design Experiment In this work, the sewn patterns are inhomogeneous since the resistance is decided by various parameters, including the number of layers of sewing (NLS), the stitching density (SD), and the stitch tension (ST) [145]. More specifically, the NLS equates to the number of repeated layers for sewing the same pattern. The SD is the number of stitches per unit length. The ST is the tension level that influences the degree of tightness and the situation of loops [ ]. To investigate the influence of sewing methods on resistance, a full factorial design, three factors in two levels, is designed with a total of twenty-four runs (by three replicates). The response is the resistance of the conductive patterns. The NLS, the SD, and the ST were applied for three factors of sewing methods, each of them is assigned to two levels: low level (1) and high level (2), see Table 4-1. The level ranges are chosen according to previous literature [ ]. Table 4-1 Design factors and levels. Level NLS SD ST (layers) (spc) (circles) Low (1) High (2) Note: spc=stitches per centimetre. 83

84 Figure 4-4 The resistance measurement experiment. Then the resistances are measured according to the AATCC standard (23 C, 20% RH) as shown in Figure 4-4. For reference purposes, a current of 100mA was used for a minimum period of 1 min until a constant voltage was acquired at an electrode separation of 10 mm. As required, three test specimens were prepared for each sample and computed the resistance in ohms per 10 mm per strand of yarn as seen in equation 4.1 [231].! 𝑅 =!!!!!!!!!!! (4.1) Where R is the resistance in ohms per 10 mm per strand, S is the number of strands per specimen, D is the distance between the electrodes in 10 mm, r is the resistance of the individual specimens tested, n is the total number of specimens tested. 84

85 Experimental results of twenty-four tests are summarised in Table 4-2. Minitab 17.0 software was used for experimental design and the statistical analysis of the experimental results. The influence of investigated factors can be estimated using the p-values of ANOVA results. The results were tested at random to improve the experimental accuracy. Table 4-2 The factorial design experiment. StdOrder RunOrder NLS SD ST Observation (layers) (spc) (circles) (Ω)

86 4.3.2 Process Development After completing the design of the experiment, SPSS 15.0 software was used for basic statistical analysis. The total valid cases were 24, which mean all the results can be used in this experiment. Then descriptive analyses are shown in Table 4-3. The mean resistance is Ω. All the results are in a range of Ω. The maximum result is Ω and the minimum result is Ω. Most of the results (95% CI) are from Ω to Ω. In order to check the data, statistical summary, stem-and-leaf plot, histogram, interval plot and the box-plot chart were used as shown in AppendixⅠ. Table 4-3 Descriptive analyses for process development. Mean 95% CI (Ω) Max Min Median Range Std.V (Ω) Lower Upper (Ω) (Ω) (Ω) (Ω) (Ω) Table 4-4 shows the results of the factorial analysis that was performed. The p-value is the probability (within a 95% confidence level) that the selected factors have a significant impact on the response (resistance of conductive patterns). It was noted that the impact is significant when its p-value 0.05 [232]. In this experiment, the main factors of resistance, NLS, SD, and ST are significant on resistance. It was also noted that the cross effects of NLS SD, NLS ST are significant. Furthermore, the coefficient of determination shows that the NLS has much larger effects than the SD and the ST. R-sq (adj) value 86

87 of 85.80% indicates that the proposed model can explain 85.80% of the variation of load hysteresis. The VIF is the variance inflation factor. It was noted that the proposed model is reliable when the VIF 10 [232]. Table 4-4 Factorial fit: resistance versus NLS, SD, ST. Term Coef SE Coef t-value p-value VIF Constant *** NLS *** 1.00 SD ** 1.00 ST ** 1.00 NLS SD *** 1.00 NLS ST *** 1.00 SD ST NLS SD ST Model Summary S R-sq R-sq(adj) R-sq(pred) % 85.80% 77.77% Note: *** stands 1% significance level, ** stands 5% significance level. The prediction model of resistance (%) can be explained by the following regression equation (4.2): Resistance= NLS SD ST NLS SD NLS ST (4.2) 87

88 Figure 4-5 Main factors plot for resistance. After determining the model to be valid, the main factors and the cross effects for resistance were generated as in Figure 4-5 and Figure 4-6. Regarding the main factor plot for resistance, NLS plays the most important role in affecting the resistance. For the NLS, the more layers of sewing acquire a much lower resistance of conductive patterns. This phenomenon may be explained by the fact that a higher NLS will increase the width of the conductive patterns and thereby reducing the resistance by introducing the parallel connections. Similarly, a higher ST acquires a lower resistance of conductive patterns. With increasing the ST, on the one hand, the conductive yarns become tighter due to a higher ST; on the other hand, it decreases the length of conductive yarns because it reduces the length of yarns used. Conversely, a higher SD leads to a higher resistance of conductive patterns. In lower SDs, yarns are partly paralleled to the textile substrate or floated on the surface of the textile 88

89 substrate, which reduces the yarns contacting points in the textile substrate and substantially decreases the resistance in this case. Figure 4-6 Interaction plot for resistance. According to the interaction plot for resistance, the intersection of the main factors shows they have cross effects, which verify the proposed regression equation 4.2. The changing trend between the factors shows that the change in resistance is not a linear change. The interaction plot is a schematic diagram of the relationship between the main factors in Table

90 Figure 4-7 Contour plot of resistance vs SD, NLS. Figure 4-8 Surface plot of resistance vs SD, NLS. The contour plot of resistance versus SD, NLS and the surface plot of resistance versus SD, NLS are plotted in Figure 4-7 and Figure 4-8. In this case, the figures show the lower NLS and higher SD results in maximised resistance, which is an unexpected result. At the same time, it is concluded 90

91 that the higher NLS and higher SD results in minimised resistance clearly, which is the expected result. In other conditions, the resistance is around Ω. Moreover, the lower SD can acquire median resistance no matter how the NLS is. Figure 4-9 Contour plot of resistance vs ST, NLS. Figure 4-10 Surface plot of resistance vs ST, NLS. 91

92 The contour plot of resistance versus ST, NLS and the surface plot of resistance versus ST, NLS are shown in Figure 4-9 and Figure In this case, the figures show the lower NLS and lower ST results in maximised resistance, which is an unexpected result. At the same time, it is concluded that the higher NLS and lower ST results in minimised resistance clearly, which is the expected result. In other conditions, the resistance is around Ω. Furthermore, the higher ST can acquire median resistance no matter how the NLS is. Figure 4-11 Residual plots for resistance for fabrication process development. Then the residuals are checked for normal distribution and constant variance of the data to confirm the adequacy of the previously proposed regression equation through normal probability, versus fit, histogram, and versus order. 92

93 The normal probability plot shows that the results are close to a straight line, which indicates a normal distribution and the proposed regression model is reasonable. The histogram shows that the distribution frequency of residuals is basically normal. The fits versus residual and the order versus residual can help to check whether there are any patterns in the results that may lead to non-constant variance or not. The normality and constant variance of the data are reasonable because the residuals versus the fitted values are distributed along the x-axis horizontal band evenly and correspondingly. Furthermore, the residuals versus the observation order show irregularity and randomness, hence the independence assumption is accepted Discussion Through the previous experiment, the NLS, the SD, and the ST significantly influence all investigated responses. The main factor of NLS is the most important factor in this experiment. Applying increased layers of sewing is an effective way to reduce the resistance. This phenomenon may be explained by the fact that a higher NLS will increase the width of the conductive patterns and thereby reducing the resistance by introducing the parallel connections. Similarly, the main factor of ST significantly influences all responses and interacts with other factors. With increasing ST, on the one hand, the conductive yarns become tighter by higher ST; on the other hand, it decreases 93

94 the length of conductive yarns because it reduces the length of yarns used. Furthermore, the main factor of SD also significantly affects the resistance. A lower SD leads to lower resistance than higher SD. This effect may be due to the change of stitch configuration. In lower SDs, yarns are partly paralleled to the textile substrates or floated on the surface of the fabric, which reduces the yarns contacting points in textile substrates and substantially decreases the resistance in this case. 4.4 Conclusions In this chapter, the resistance of conductive yarns using different sewing methods is observed in the attempt to understand and interpret the main factors and their cross effects in influencing resistance. A fabrication process development experiment is designed to improve the performance of the fabric NFA. The resistance is further defined to indicate the working efficiency. Full factorial experiments were carried out to observe the influence of the NLS, the SD, and the ST. Through the experimental design and data analysis, following results have been obtained. 1). Different sewing methods significantly influence all investigated responses, which can be observed through the changes of resistance. Using suitable sewing methods to fabricate the conductive pattern is an effective way to 94

95 decrease the resistance. 2). The NLS is the most important factor in this experiment. Applying more layers of sewing is an effective way to reduce the resistance. Similarly, a higher ST shows a lower resistance in the conductive patterns. In contrast, lower resistance can be obtained by applying a lower SD. 3). According to the interaction plot for resistance, the intersection of the main factors shows they have cross effects, which verify the proposed regression equation 4.2. The changing trend between the factors shows that the change in resistance is not a linear change. 4). During the previous experiment, the main parameters in the sewing methods that influence the resistance of conductive patterns have been discovered. The initial regression equation for the resistance has been obtained through the fabrication process development experiment. Later, the NLS and the SD will be chosen as the main factors and will be used to determine the fitted regression equation in a central composite design. During this process, the minimum resistance will be obtained in detail. 95

96 CHAPTER 5 FABRIC NFA FABRICATION PROCESS OPTIMISATION 5.1 Introduction The previous chapter focused on factorial and fractional factorial designs, which are really important for fabrication process development. The most important factors that affect the resistance of the fabric NFA have been identified in Chapter 4. The fabrication process development experiment helps to find the appropriate subset of process variables. Over the original region of experimentation in Chapter 4, the minimum resistance that can be obtained is approximately 0.50 Ω. It is suggested to run this process at a resistance of Ω. However, they may still be far away from the optimal results. Therefore, it is necessary to leave the previous experimental region of the process development experiment to decrease the resistance. In order to find the minimum resistance, in other words, finding the optimum areas of the resistance that lead to the best performance of the fabric NFA, the fabrication process optimisation experiment is introduced. In this chapter, firstly, the steepest ascent will be used to find the near optimum area. Then the minimum response location will be approximated and the central composite design in the central point will be used. Lastly, the 96

97 second-order equation for resistance will be found and the ideal areas for the resistance of the fabric NFA will be identified. Through the above methodology, an optimum result for optimising functional performance for the fabric NFA will be found. 5.2 Experiment Experimental Design The previous chapter focused on the fabrication process development experiment. During that process, the most important factor that affects the resistance of the fabric NFA was identified. This chapter will pay attention to the fabrication process optimisation, in other words, finding the optimum areas of the resistance that lead to the best performance of the fabric NFA. Thus, the central composite design is used Material The conductive yarns used in this work are Statex 235/24 dtex 2-ply conductive yarns (Statex, USA). The textile substrate used in this work is Composition: 50% Meta-aramid and 50% FR Lenzing fibre 1:1 plain woven fabric. The materials have been discussed in Chapter Instrument All fabric samples were sewn using a sewing machine (type number: 97

98 LS2-190), which was produced by the Mitsubishi Co. Ltd, UK, under standard laboratory conditions of 20% relative humidity and 20 C room temperature. The resistance of the conductive patterns was measured using a multimeter (type number: U1251A) and a power supply (type number: E3641A), both of them were produced by the Agilent Co. Ltd, USA. Four-point probe resistance measurement as shown in Figure 3-1 was used to measure the resistance of the conductive patterns [210] Experimental Method Figure 5-1 The fabrication process optimisation experiment. Step 1: Based on the previous chapter, a first-order model and the relatively 98

99 minimum result in the fabrication process development part can be figured out. Step 2: finding the main factors of the regression model (always two or three factors) so that subsequent experiments are more effective. Step 3: checking whether the proposed levels are near the optimum or not. This fabrication process optimisation experiment was adopted to find a region of improved response, called the method of steepest ascent. Step 4: After completing the steepest ascent, the levels approximate the optimum. The minimum response location can be approximated. This point is called the central point. Step 5: At the central point, the central composite design experiment was conducted, and the true first-order model within a relatively small region around the minimum resistance was fitted to the data. 5.3 Statistical Analysis and Discussion Steepest Ascent By using ANOVA, it was found that two of the three factors, the NLS and the SD, significantly affect resistance. If a model is fitted using only these main factors, the following equation can be obtained: y = NLS SD (5.1) 99

100 as a prediction equation for the resistance. Over the original region of experimentation as shown in Figure 4-7 that is, for NLS between 1 and 3 layers, and SD between 3 and 5 stitches per centimetre (spc). Based on the previous fabrication process development experiment, the minimum resistance that can be obtained is about 0.50 Ω. It is suggested to run this process at a resistance of Ω. Thus, it is necessary to leave the previously experimental region of the process development experiment to decrease the resistance. From the fitted model (equation 5.1), it is seen that to move away from the design centre - the point ( NLS = 0, SD = 0) along with the path of steepest ascent, units in the SD direction for every units in the NLS directions are moved. Thus the path of steepest ascent passes through the point ( NLS = 0, SD = 0) and has slope

101 Figure 5-2 Steepest ascent experiment. It was decided to use one layer of NLS as the basic step size. Thus, one layer of NLS is equivalent to a step in the NLS coded variable of NLS = 1. Thus, the steps along the path of steepest ascent are NLS = 1 and SD = NLS 3 = A change of SD = 0.33 in the SD coded variable is equivalent to approximately 0.33 spc. An actual observation on resistance will be obtained by running the process at each point Central Composite Design Experiment To investigate the small religion of optimisation of the fabric NFA, a central composite design was adopted. The design of the experiment and statistical analysis was designed using the Minitab 17.0 software. Figure 5-2 shows four points along the path of steepest ascent and the 101

102 observation results (resistance). At the second point, the observed resistance is the lowest point of Ω. Therefore, NLS=4 layers and SD=5.33 stitches per centimetre (spc) with an observed resistance of Ω could be used as the central point for the central composite design experiment. The steepest ascent procedure terminates in the vicinity of NLS=4 layers and SD=5.33 spc with an observed resistance of Ω. This region is close to the desired operating region for the process. The NLS=4 layers and SD=5.33 spc are chosen as the central point for the central composite design. Figure 5-3 Central composite design in the coded variables. The central composite design in the coded values is shown in Figure 5-3. Based on the steepest ascent, the central point is set at NLS = 0 and SD = 0, which represents NLS=4 layers and SD=5.33 spc respectively. By 102

103 using this information, a linear equation can be derived: NTS uncoded value = 1 (coded values) + 4 (layers) (5.2) SD uncoded value = 1 coded values (spc) (5.3) If the coded value is 1.414, the uncoded value of NLS and SD will be: 1(1.414) + 4=5.414 (layers) 1(1.414) =6.733 (spc) If the coded value is , the uncoded value of NLS and SD will be: 1(-1.414) + 4=2.586 (layers) 1(-1.414) =3.916 (spc) The central composite design is investigated in the two factors that influence the resistance, called the NLS (number of layers of sewing) and the SD (stitch density) for the replication of three times. In the central composite design, the cube points are 12, the centre points in the cube are 15, and the axial points are 12, so the total runs are 39. In order to improve the experimental accuracy, the results were tested at random. 103

104 Table 5-1 The central composite design for resistance. StdOrder RunOder Coded Coded Uncoded Uncoded Observation NLS SD NLS SD (layers) (spc) (Ω)

105 5.3.3 Process Optimisation In the fabrication process optimisation experiment, SPSS 15.0 software and Minitab 17.0 software were used for basic statistical analysis and surface response regression. The total valid cases are 39, the valid percentage is 100%. The descriptive analyses are shown in Table 5-2. The mean resistance is Ω and the median resistance is Ω. All the results are in a range of Ω. The maximum result is Ω and the minimum result is Ω. 95% results are in the range of Ω to Ω. In order to check the data, statistical summary, stem-and-leaf plot, histogram, interval plot and the box-plot chart were used, as shown in AppendixⅡ. Table 5-2 Descriptive analyses for process optimisation. Mean 95% CI (Ω) Max Min Median Range Std.V (Ω) Lower Upper (Ω) (Ω) (Ω) (Ω) (Ω) Table 5-3 shows the results of the central composite design that is performed. As has been said in Chapter 4, the p-value is the probability (within a 95% confidence level) that the selected factors have a significant impact on the response (resistance of conductive patterns). It was noted that the impact is significant when the p-value 0.05 [232]. In this experiment, the main factor of the NLS is significant in respect of the resistance. It is also noted that the cross effects of NLS NLS and NLS SD, are significant. Furthermore, the 105

106 factors of the NLS and NLS SD are the most significant factors in this model. R-sq (adj) value of 66.02% indicates that the proposed model can explain 66.02% of the variation of load hysteresis. The VIF is the variance inflation factor. It was noted that the proposed model is reliable when the VIF 10 [232]. Table 5-3 Response of surface regression: resistance versus NLS, SD. Term Coef SE Coef t-value p-value VIF Constant *** NLS *** 1.00 SD NLS NLS ** 1.02 SD SD NLS SD *** 1.00 Model Summary S R-sq R-sq(adj) R-sq(pred) % 66.02% 55.65% Note: *** stands 1% significance level, ** stands 5% significance level. For the resistance, the prediction model of Resistance can be explained by the following regression equation (5.4): Resistance = NLS NLS NLS NLS SD (5.4) 106

107 Figure 5-4 Contour plot of resistance vs SD, NLS. After determining the model to be valid, the contour plot of resistance versus SD and NLS are generated in Figure 5-4. In this case, the figure shows the resistance results with the NLS and SD changing. It was found that the NLS=4 layers of sewing mainly results in minimised resistance. Furthermore, both the less or more layers of sewing lead to higher resistance. Figure 5-5 Surface plot of resistance vs SD, NLS. 107

108 Figure 5-5 gives the response surface plot for resistance. It is seen that the optimisation results are around the central point ( NLS = 0, SD = 0), and the minimum resistance is approximately Ω. According to the above two figures, the second-order model for resistance versus SD and NLS can be observed clearly. Figure 5-6 Contour plot of resistance. To determine an optimum resistance that can improve the working efficiency and extend the reading range, an overlay plot for resistance is used as shown in Figure 5-6. Variables of the number of layers of sewing and the stitch density are nominal and set as deciding factors. The target resistance is set from y = 0.40 Ω to y = 0.50 Ω. The white areas of this plot identify the ideal resistance results and its NLS and SD measurements. 108

109 Figure 5-7 Residual plots for resistance for fabrication process optimisation. Then the residuals are checked for normal distribution and constant variance of the data to confirm the adequacy of the previously proposed regression equation through normal probability, versus fit, histogram, and versus order. The normal probability plot shows that the results are close to a straight line, which indicates a normal distribution and the proposed regression model is reasonable. The histogram shows that the distribution frequency of residuals is basically normal. The fits versus residual and the order versus residual can help to check whether there are any patterns in the results that may lead to non-constant variance or not. The normality and constant variance of the data are reasonable because the residuals versus the fitted values are distributed along the x-axis horizontal band evenly and correspondingly. Furthermore, the residuals versus the observation order show irregularity and randomness, hence the independence assumption is accepted. 109

110 5.3.4 Discussion Through the previous experiment, the second-order optimum model has been developed based on the main factors of the NLS and the SD. During the steepest ascent experiment, the uncoded value of central point of NLS = 4 layers, SD = 5.33 spc. That means before NLS = 4 layers, SD = 5.33 spc, the resistance decreases, and after NLS = 4 layers, SD = 5.33 spc, the resistance increases. The minimum results can be acquired around the central point (NLS = 4 layers, SD = 5.33 spc). As for the NLS, on the one hand, higher NLS increases the width of the conductive patterns and thereby reduces the resistance by introducing the parallel connections; on the other hand, too many NLS extends the length of conductive yarns, which increases the resistance at the same time. Similarly, as for the SD, on the one hand, the lower SD reduces the yarns contacting points in textile substrates and substantially decreases the resistance; on the other hand, too lower SD also affects the stitch tension because of the lack of a control point, which increases the resistance. After confirming the central point, the second-order optimum model can be investigated by central composite design experiments. The main factor of the NLS is significant on all responses and interacted with itself. This factor may be due to the parallel connections of conductive patterns and the length of 110

111 conductive yarns, which has been explained in detail (Chapter 4). It is interesting that the main factor of SD has no factor on resistance, while the cross factors of the SD and the NLS have a significant factor on resistance. This phenomenon may be explained by the fact that both of the factors influence the length of conductive yarns accordingly. 5.4 Conclusions In this chapter, the resistance of the conductive yarns based on two main factors, NLS and SD, and their cross effects have been observed in the attempt of interpreting the fabrication process optimisation experiment. Through the experimental design and data analysis, the following results were obtained. 1). In this experiment, the main factor of resistance - NLS is significant in aspect of the resistance. It is also noted that the cross effects of NLS NLS and NLS SD, are significant. The NLS and NLS SD are the most significant factors, and the influence of NLS NLS is a relatively weak. Furthermore, the SD, SD SD have little significant influence on the resistance of conductive patterns. 2). According to the contour plot and response surface plot, it is possible to calculate the optimisation areas and related NLS, SD measurements in the process optimisation. During this process, the minimum resistance of the 111

112 fabric NFA can be obtained and will be checked in Chapter 6. The interval plot chart and box-plot chart show all the results in different groups and indicate the mean or median of each group clearly. All the above charts and tables give a clear image of the experimental results, providing the design of the experiment is valid and reasonable, as well as supporting fundamental evidence to support later analytical processes. 112

113 CHAPTER 6 FABRIC NFA PERFORMANCE EVALUATION 6.1 Introduction The fabric near field antenna (NFA) fabrication process development and optimisation has been discussed in Chapter 4 and Chapter 5. The regression equation for resistance has been obtained and established during the factorial design experiments and the central composite design experiments. The optimisation resistance of the fabric NFA has been calculated to improve the quality factor and extend the reading range. This chapter will test the quality factor and the reading range to verify the hypothesis in Chapter 3. Furthermore, as the fabric NFA is always distorted when applied to clothes and on human-bodies, the performance of the fabric NFA in terms of the S11 value under different bending conditions will be evaluated. 6.2 Experiment Experimental Design To verify the hypothesis for the relationship between the resistance and the quality factor, the reading range in Chapter 3, and to evaluate the performance of the fabric NFA under different bending conditions, the fabric NFA performance evaluation is designed. 113

114 6.2.2 Instrument Figure 6-1 The FieldFox vector network analyser (type number: N9918A). Figure 6-2 The NFA reader (type number: TRF7970AVM). 114

115 The quality factor and the reading range were measured using a FieldFox vector network analyser (type number: N9918A), which was produced by the Agilent Technologies Co. Ltd, USA, and a NFA reader (type number: TRF7970AVM), which was produced by the Texas Instruments Co. Ltd, USA Experimental Method Figure 6-3 The experimental method for testing the reading range. For testing the reading range of the NFA, the experimental method is shown in Figure 6-3. The reading range can be measured using the ruler. 115

116 Figure 6-4 The experimental method for testing the quality factor. For testing the quality factor, the fabric NFA should be kept as flat and connected to the vector network analyser. Then the quality factor can be read directly through the vector network analyser. Figure 6-5 The experimental method for evaluating the performance of the NFA under different bending conditions. For evaluating the performance of the fabric NFA under different bending conditions, the NFA should be bent inwards. The results were read directly via 116

117 the analyser and the line chart was completed accordingly. 6.3 Performance Evaluation Quality Factor and Reading Range Test After the fabrication process optimisation experiment, several groups were adopted to test the quality factor and the reading range to verify the hypothesis that lower resistance will lead to higher quality factors and longer reading ranges. The values of the NLS and the SD in Table 6-1 were got from the fabrication process optimisation experiment described in Chapter 5. The group set is shown as blow, in Table 6-1. Table 6-1 Group set for results measurement. NLS (layers) SD (spc) A B C D E F G H I According to the test results, it is verified that the resistance is inversely proportional to the quality factor. That is to say, the lower resistance would lead to higher quality factors, and vice versa. Of course, the quality factor is decided by several factors, such as the matching problem of the fabric NFA and the test chip, the sewing variability and so on. The test data, however, are roughly equivalent to the theoretically derived formulae that were used in Chapter 3. When the resistance equals Ω, the highest quality factor is

118 Furthermore, the ideal reading range in the group A is about 70 mm (increases by 35%), which is notably longer than that in other groups. Further information is shown in Table 6-2. Table 6-2 The test results for the quality factor and the reading range. A B C D E F G H I Resistance (Ω) Quality factor Stdv Reading range (mm) Stdv (mm) Figure 6-6 Contour plots for performance evaluation. 118

119 Unit: mm The reading range(mm) The quality factor Resistance (Unit:Ω) Figure 6-7 Line chart of resistance vs the reading range and the quality factor. Figure 6-6 and Figure 6-7 show the relationship between the resistance and the reading range, the quality factor. The contour plots as seen in Figure 6-6 were created from the results in Table 6-1 and Table 6-2 using the Minitab 17.0 software. These schematic diagrams show that the higher resistance areas correspond to the higher quality factor areas and the longer reading range areas. The line chart, as seen in Figure 6-7, was created from the results in Table 6-2 using the Excel 2013 software. The results show that the changing trend in the quality factor and the changing trend in the reading range are similar. Figure 6-6 and Figure 6-7 verify the hypothesis in Chapter 3 that the resistance determines the quality factor and the reading range. Next, the measurements were conducted with a vector network analyser and 119

120 thus evaluated the performance of the fabric NFA in terms of the S11 value and the reading range. The fabric NFA was connected to the vector network analyser (type number: N9918A) via conductive epoxy (Circuit works CW2400). Firstly, the measured output impedance was 145-j85 Ω using the vector network analyser. Thus, the input impedance was adjusted to 145+ j85 Ω, which requires it to be matched with the output impedance at MHz [215]. The S11 value represents the antenna resonant frequency and the quality factor. Then the fabric NFA was matched to the impedance of the vector network analyser, which is a constant value at 50 Ω, by a matching circuit in order to measure to the S11 value [215]. Figure 6-8 The centre frequency under different bending conditions. As seen in Figure 6-8, the centre frequency of the fabric NFA is close to MHz (the original line) and well matched which indicates the measured result is close to the theoretical simulation result. The measured quality factor is 120

121 approximately 2.1 which brings a wide bandwidth but reduces its reading range Bending Test The fabric NFA was always bent on clothes and when worn on human bodies. Hence, measuring the NFA performance under different bending conditions is important. The fabric NFA was axisymmetric and was bent in both the X and the Y inward directions (the X direction refers to the width direction, the Y direction refers to the length direction), for example, the normal condition (X-0 and Y-0 ), the X-90 bending condition, the X-180 bending condition, the Y-90 bending condition, the Y-180 bending condition, and the on-body condition (the arm line). Several important conclusions can be achieved from the measured results as seen in Figure 6-6. It was observed that the centre frequency rises a little bit to 14 MHz but still within the NFA s bandwidth when bent at 90 degrees in the X and the Y directions. The S11 value of the X direction at 90 degrees the bent NFA in the centre frequency is -27 db, which is 10 db higher than the un-bent NFA. The current flow in the two sides of the X direction is in opposite direction. When the fabric NFA is bent, the conductive yarns of two sides are closer together, thus the induced magnetic fields cancelled each other out [229]. In the Y direction, the minimum S11 value increases to -23 db. The reason is that the two side yarns of the Y direction are longer than the X direction, which has more cancelling effects. The S11 value is measured to find the NFA performance under extreme 121

122 condition. The centre frequency moves to 12.5 MHz and the S11 value increases to -18 db. The S11 value of the X direction and the S11 value of the Y direction coincide. It is reasonable because the current flow directions in the yarns are opposite to each other. No magnetic field can be generated. The reading distance should be 0. The fabric NFA is also put on the arm and the results were acquired. 6.4 Conclusions The measurement results show that the simulation results were in correspondence with the experimental results qualitatively. The results verify the hypothesis that the resistance determining the quality factor and the reading range, and thus the resistance is the main factor that influencing the working efficiency of the fabric NFA. During the stimulation process, the fabric NFA was optimised to achieve up to a 70 mm reading range when using 4 layers of sewing and a stitch density of 5.33 spc. After the bending test, this work demonstrates that the fabric NFA can be a viable alternative for wearable applications under different bending conditions. 122

123 CHAPTER 7 CONCLUSIONS AND FUTURE WORK 7.1 Conclusions To achieve the fabric near field antenna (NFA) in wearable applications requirement today in terms of comfort, function, sensing and ergonomics, this work aims to introduce and realise a complex design and engineering methodology of the fabric NFA using sewing technology. It has been completed by integrating appropriate conductive material preparation, antenna design, process development and optimisation, and performance evaluation. The four research aims and objectives have been achieved as follows. 1. Comparison and selection of suitable conductive materials through the literature and the characterisation of the selected conductive yarns. 2. Establishment of a multi-disciplinary framework for the fabric NFA in wearable applications using sewing technology. 3. Design and fabrication of the fabric NFA to realise the flexibility and comfort properties in wearable applications. 4. Realisation of a highly conductive fabric NFA simulation model through process development and process optimisation according to the statistical analysis software. 123

124 7.2 Future work The research aims and objectives of this work have been achieved. Future work for this research, however, should be considered for the development of a more systematic and complex the fabric NFA for wearable applications as follows. 1. Optimise the fabric NFA to improve the comfort and electrical performances. 2. Develop RFID in wearable applications according to sewing and embroidering technologies based on this research. 3. Develop new conductive materials to design and realise the fabric NFA, for example, conductive polymers or carbon-based nanomaterials, especially graphene and hybrid conductive materials, which can be attractive prospects in conductive materials because they have excellent transparency, electrical conductivity and good flexibility. 4. Expand new realisation methods for the fabric NFA, for example, printing, weaving or knitting. Based on established multi-disciplinary framework, novel integration technologies for the fabric NFA in wearable applications may be achieved in the future. 124

125 Appendix І The tables and figures in Appendix І are the data check for Chapter 4. The statistical summary is shown as follows. The total valid cases are 24, which means all the results can be used in this experiment. Table І-1 Statistics summary for process development. Cases Valid Missing Total N Percent N Percent N Percent Resistance % 0 0.0% % All the results are shown as follows. The stem-and leaf plot shows more than a half of the results are around 0.6 Ω. Table І-2 Stem-and-leaf plot for process development. Frequency Stem&Leaf Stem width Each leaf 1 case(s) 125

126 Figure І-1 Histogram of resistance for process development. A normal distribution of the data is shown using the histogram and distribution curve, which indicates the experimental processes are reasonable and the results are acceptable. Figure І-2 Interval plot of resistance for process development. 126

127 Figure І-3 Interval plot of resistance for process development. The interval plot chart and box-plot chart show all the results in different groups and indicate the mean or median of each group clearly. All the above charts and tables give a clear image about the experimental results, proving the design of the experiment is valid and reasonable, as well as supporting fundamental evidence to support later analytical processes. 127

128 Appendix И The tables and figures in Appendix И are the data check for Chapter 5. The statistical summary is shown as follows. The total valid cases are 39, which means all the results can be used in this experiment. TableИ-1 Statistics summary for process optimisation. Cases Valid Missing Total N Percent N Percent N Percent Resistance % 0 0.0% % The stem-and-leaf plot is shown as follows. Most of the results are within Ω. For the details of results distribution, please see the steam-and-leaf plot. TableИ-2 Stem-and-leaf Plot. Frequency Stem&Leaf Extremes (. 77) Stem width Each leaf 1 case(s) 128

129 Figure И-1 Histogram of resistance for process optimisation. A normal distribution of the data is shown using the histogram and distribution curve, which indicates the experimental processes are reasonable and the results are acceptable. Figure И-2 Box-plot chart of resistance for process optimisation. 129

130 Figure И-3 Interval plot of resistance for process optimisation. The interval plot chart and box-plot chart show all the results in different groups and indicate the mean or median of each group clearly. All the above charts and tables give a clear image about the experimental results, proving the design of the experiment is valid and reasonable, as well as supporting fundamental evidence to support later analytical processes. 130