Development of Nanocomposites Based Sensors Using

Home , Liu Chang (Southern Han), Nanofoam, Silver nanoparticle

DEVELOPMENT OF NANOCOMPOSITES BASED SENSORS USING

MOLECULAR/POLYMER/NANO-ADDITIVE ROUTES

Dissertation

Submitted to

The School of Engineering of the

UNIVERSITY OF DAYTON

In Partial Fulfillment of the Requirements for

The Degree of

Doctor of Philosophy in Engineering

Chang Liu

Dayton, Ohio

May 2019

DEVELOPMENT OF NANOCOMPOSITES BASED SENSORS USING

MOLECULAR/POLYMER/NANO-ADDITIVE ROUTES

Name: Liu, Chang

APPROVED BY:

______Khalid Lafdi, Ph.D. Donald A. Klosterman, Ph.D. Advisory Committee Chairman Associate Professor Professor; Wright Brothers Endowed Department of Chemical & Materials Chair in Nanomaterials Engineering Department of Chemical & Materials Engineering

______Erick S. Vasquez, Ph. D. Vikram K. Kuppa, Ph.D. Committee Member Committee Member Assistant Professor Research Scientist Department of Chemical & Materials University of Dayton Research Institute Engineering

______Robert J. Wilkens, Ph.D., P.E. Eddy M. Rojas, Ph.D., M.A., P.E. Associate Dean for Research and Innovation Dean Professor School of Engineering School of Engineering

Chang Liu

2019

ABSTRACT

DEVELOPMENT OF NANOCOMPOSITES BASED SENSORS USING

MOLECULAR/POLYMER/NANO-ADDITIVE ROUTES

Name: Liu, Chang University of Dayton

Advisor: Dr. Khalid Lafdi

In this study, multiple approaches were explored for building advanced nanocomposite sensors intended for use in fiber reinforced organic matrix composite structures. One expected application of such technology is sensing of chemical penetration in the walls of large chemical tanks. The work described herein involved development and characterization of various novel conductive nanocomposites from polymeric feedstocks as well as carbon nanoparticles.

The first approach consisted of using pitch based, liquid crystal molecular additives to polyacrylonitrile (PAN) to create novel electrospun carbon nanofibers. Raman spectroscopy confirmed the increase of an ordered structure in PAN/pitch based carbon nanofibers by analyzing the sharpness of the G band. As a result, the addition of pitch increased the degree of graphene alignment because of the high amount of liquid crystal present in the pitch. This structure led to enhanced physical properties of the carbon nanofibers.

The second approach used a conductive network of conjugated polymer (polyaniline,

PAni) nanoparticles dispersed in a blend of polyvinylpyrrolidone (PVP) and polyurethane (PU).

PAni was synthesized using an in situ polymerization method which resulted in colloidal PAni or

PAni nanowires. PAni nanowires self-assembled into scattered fractal networks. After adding PU, a concentrated PAni/PVP phase occurred. Such a phenomenon was attributed to the balance between blocking force and van der Waals force. When the surface tension is the determining factor in the ‘island’, the round shaped phase separation occurs. The surface tension and Van der

Waals force were two determining factors in the formation of bi-continuous phase separation.

iii

When the forces were in equilibrium, a fractal network structure was formed and the polymer blends were very stable. A flexible conductive fabric was successfully prepared by coating the conductive ternary mixture onto a non-woven fabric.

The last approach uses carbon nanoparticles (carbon nanotube and carbon black) as PAni as additives to an epoxy matrix to alter conductivity in order to predict the chemical penetration in a composite structure. In this study, two nanocomposite formulations were produced: one is based on polyaniline and the second uses CNT as additives. These materials were dispersed in an epoxy resin system and cured into a solid plate which also contained embedded metal electrodes.

The sensor assembly was then immersed in an acid solution to evaluate its ability to detect the ingress of ions. It appears as the amount of nano-additives increased, the conductivity increased and the response time towards acid penetration was shorter. The sensing mechanism was depicted using a Fickian model and the experimental and theoretical data were in agreement. Indeed, the penetration and diffusion of hydrogen ions were responsible in connecting the CNT aggregates by forming a continuous conductive network. Finally, the sensor was connected to a radio frequency based wireless system to demonstrate its ultimate use in the field.

DEDICATION

Dedicated to my parents

Yumen Liu & Aizhu Zhang

ACKNOWLEDGEMENTS

My special thanks are giving to Dr. Khalid Lafdi, my advisor and friend, for providing the time, material and equipment necessaries, for guiding me to be a competent researcher, for directing this dissertation and bringing it to an accomplishment patiently and professionally, for being a kind and inspiring mentor. Great thanks to my committee members for their guidance.

I would also express my appreciation to everyone who helped me. This includes but not limited to Dr. Donald A. Klosterman, Dr. Erick S. Vasquez, Dr. Cao Li, Dr. Francisco Chinesta,

Dr. Abdulaziz Baçaoui, Robyn Braford, Qichen Fang, Yuhan Liao, Jean-Baptiste Dumuids, Nuha

Al Habis, Saja M. Nabat Al-ajrash, Ali A. Muhsan, Tseng-Hsiang Ho, Robert Busch, and

Shuangshan Li.

The support and accompany from Yufei Liu is invaluable.

TABTLE OF CONTENTS

ABSTRACT ...... iii

DEDICATION ...... v

ACKNOWLEDGEMENTS ...... vi

LIST OF FIGURES ...... ix

LIST OF TABLES ...... xiii

LIST OF EQUATIONS ...... xiv

LIST OF ABBREVIATIONS AND NOTATIONS ...... xv

1. CHAPTER I INTRODUCTION ...... 1

1.1 Nanocomposites ...... 1 1.2 Nano-additives ...... 2 1.3 Nano Fabrication ...... 5 1.4 Conductive Network ...... 8 1.5 Structural Health Monitoring ...... 8 1.6 Thesis Statement ...... 9 2. CHAPTER II THE EFFECT OF CARBONACEOUS MOLECULAR ADDITIVE ON

ELECTROSPUN CARBON NANOFIBER ...... 11

2.1 Background ...... 11 2.2 Materials and Characterization ...... 14 2.3 Results and Discussion ...... 15 2.4 Conclusion ...... 25 3. CHAPTER III SELF-ASSEMBLY AND SURFACE TENSION INDUCED FRACTAL

CONDUCTIVE NETWORK IN TERNARY POLYMER SYSTEM ...... 27

3.1 Background ...... 27 3.2 Materials and Characterization ...... 29 3.3 Results and Discussion ...... 30 3.4 Conclusion ...... 39 4. CHAPTER IV CNT AND POLYANILINE BASED SENSORS FOR THE DETECTION OF

ACID PENETRATION IN POLYMER COMPOSITE ...... 41

vii

4.1 Background ...... 41 4.2 Materials and Methods ...... 44 4.3 Result and Discussion ...... 47 4.4 Conclusion ...... 56 5. CHAPTER V FABRICATION OF HIGH PERFORMANCE NANOCOMPOSITE BASED

CHEMICAL SENSOR USING LOW CONCENTRATION ADDITIVES ...... 58

5.1 Background ...... 58 5.2 Materials and Methods ...... 61 5.3 Result and Discussion ...... 65 5.4 Conclusion ...... 73 6. CHAPTER VI CARBON NANO MATERIAL BASED CHEMICAL SENSING: WIRELESS

DESIGN ...... 74

6.1 Background ...... 74 6.2 Experiment Setup ...... 75 6.3 Sensing Result ...... 76 6.4 Conclusion ...... 77 7. CHAPTER VII CONCLUSION AND PERSPECTIVE ...... 79

REFERENCES ...... 82

APPENDIX A Arduino-RFID-RC522-LCD Connections ...... 109

APPENDIX B Arduino Code for Wireless Sensing ...... 110

viii

LIST OF FIGURES

Figure 1.1 Dispersion (aggregation) of nano additive and formation of conductive path in

nanocomposite...... 2

Figure 1.2 Carbon nanomaterials: (a) Carbon nanotube; (b) Graphene; (c) Carbon black...... 3

Figure 1.3 Illustration of (a) spin coating and (b) electrospinning experiment setup...... 7

Figure 2.1 Average chemical structure of A240 molecule...... 13

Figure 2.2 DSC results of PAN, pitch, and their blends. (a) Glass transition of the blends;

(b) Derivate curves of (a)...... 17

Figure 2.3 FTIR results of PAN, pitch, and their blends. (a) 3000 cm-1 region for C-H. (b)

900–1800 cm-1 region for C-O bonds and C-C bonds; (c) 2250 cm-1 region for cyano

bond of PAN...... 18

Figure 2.4 (a) Rheology result of PAN and the blends. (b) TGA result of PAN and the

blends...... 20

Figure 2.5 Optical (a–c) and SEM (d–f) images of PAN (a,d), Pitch0.5PAN1 (b,e),

Pitch1PAN1 (c,f) as-spun fibers...... 23

Figure 2.6 SEM images of stabilized (a–c) and carbonized (d–f) PAN (a,d), Pitch0.5PAN1

(b,e), Pitch1PAN1 (c,f) fibers...... 23

Figure 2.7 Raman spectroscopy results. (a) Difference on D band and G Band; (b)

sharpness of G band and D/G ratio...... 24

Figure 2.8 Conductivity and modulus of the carbon nanofibers...... 25

Figure 3.1 Optical (a, b, d, e) and SEM (c, f) images for thin film (a, d) and bulk (b, c,

e, f) of the 1-0-1 sample (a, b, c) and the 1-1-0 sample (d, e, f). The scale bars in

optical images are 100 μm and for SEM images are 10 μm...... 31

Figure 3.2 Optical (a, b, d, e, g, h) and SEM (c, f, i) images for thin film (a, d, g) and bulk

(b, c, e, f, h, i) of the 1-1-1 sample (a, b, c), 1-5-10 sample (d, e, f), and 1-10-5 sample

(e, h, i). The scale bars in optical images are 100 μm and for SEM images are 10 μm...... 32

Figure 3.3 Phase separation situations in the observed ternary polymer blends. (a) Van der

Waals force and surface tension are balanced with friction. (b) Van der Waals force

balanced with surface tension. (c) Surface tension dominates. (d) Illustration of the

balance between forces. (e) SEM image of part a...... 34

Figure 3.4 (a) Sample points prepared and measured in experiments. (b) General physical

property of the polymer blends. (c) Initial and (d) stable conductivity of polymer blends.

The unit of scale bar on the right side is log ohm and (c, d) are using the same scale. (e)

Stability of conductivity for various polymer blends...... 37

Figure 3.5 (a) Flexible 1-5-10 sample treated nonwoven fabric mat. (b) Origami crane made

from the fabric in part a. (c) Test of the origami crane as flexible and conductive fabric.

(d) SEM image of pristine nonwoven fabric mat. (e) SEM image of coated nonwoven

fabric mat. The scale bars in parts d and e are 8 μm...... 39

Figure 4.1 (a) Synthesis route for PAni-ES and PAni-EB. (b) FTIR result for PAni-EB...... 47

Figure 4.2 (a) SEM image of the CNT. (b) Raman spectroscopy profile for CNT. (c) Powder

XRD result for CNT...... 48

Figure 4.3 (a) Assembly of nanocomposite sensor. (b) Test setting up...... 48

Figure 4.4 Photo images of Pristine epoxy, epoxy-PAni, epoxy-CNT samples before and

after acid immersion...... 49

Figure 4.5 SEM images of samples' cross sections before and after acid immersion

respectively. (a,b): Pristine epoxy samples. (c,d): Epoxy-PAni samples and (e,f): Epoxy-

CNT samples...... 50

Figure 4.6 Resistance values of the sensors as a function of time. (a) Overall curves. (b)

A zoom-in figure for 0-200s...... 51

Figure 4.7 (a) Response time of different sensors. (b) Final resistance of different samples.

Figure 4.8 Modeling result: the effect of conductance (A’) of the attacking chemicals on the

electrical response...... 54

Figure 4.9 Modeling result: the effect of diffusion coefficient (D) of attacking chemicals on

the electrical response...... 55

Figure 4.10 Modeling result: relationship between diffusion coefficient and sensing response

time...... 56

Figure 5.1 Sensing mechanism of CNT based nanocomposite sensor...... 60

Figure 5.2 (a) Nanocomposite sensor assembly. (b) Testing setup...... 62

Figure 5.3 (a) SEM image of the CNT. (b) SEM image of the carbon black. (c) XRD results

for CNT and carbon black. (d) Optical image of cross section of the nanocomposite

sensor showing the dispersion of CNT is good...... 66

Figure 5.4 Typical SEM cross section images of (a, c) pre-treated and (b, c) post-treated

(a, b) CNT and (c, d) carbon black sensors...... 67

Figure 5.5 Recorded resistance signals of the carbon black nanocomposite sensors as a

function of time. (a) Using normal scale for x-axis. (b) Using logarithm scale for

x-axis...... 68

Figure 5.6 Recorded resistance signals of the CNT nanocomposite sensors as a function of

time. (a) Using normal scale for x-axis. (b) Using logarithm scale for x-axis...... 69

Figure 5.7 The effect of effective diffusion coefficient (De) of attacking chemicals on the

Figure 5.8 The effect of nano-additive aspect ratio and diffusion coefficient ratio. The arrow

shows the increment of additive aspect ratio from 200 to 40000 with an interval of

2000...... 72

Figure 5.9 Comparison of (a) experimental result and (b) Simulation result...... 73

Figure 6.1 Working scheme of RFID system...... 75

Figure 6.2 The assembled RFID system...... 76

Figure 6.3 Modification the RFID tag with CNT...... 76

Figure 6.4 Sensing behavior of the CNT modified RFID tag...... 77

xii

LIST OF TABLES

Table 2.1 Diameter ranges of As-Spun, Stabilized, and Carbonized fibers...... 22

xiii

LIST OF EQUATIONS

Equation 3.1 The force balance on the phase boundary...... 35

Equation 3.2 The Van der Waals force between two similar cylindrical surfaces...... 35

Equation 3.3 The Young-Laplace equation...... 36

Equation 4.1 The Fick's second law...... 45

Equation 4.2 A simple one-dimensional solution for the Fick's second law...... 45

Equation 4.3 The relationship between conductance and concentration of ions...... 45

Equation 4.4 The overall resistance represent by the sliced layer's resistance...... 46

Equation 4.5 The relationship between resistance and ion concentration...... 46

Equation 4.6 The expression of overall resistance...... 46

Equation 4.7 A transfromation of overall resistance...... 46

Equation 5.1 The Fick's second law with effective diffusion coefficient...... 63

Equation 5.2 The express the concentration of ion by time and location...... 63

Equation 5.3 The expression of the effective diffusion coefficient...... 63

Equation 5.4 The expression of the geometrical parameter x...... 64

Equation 5.5 The expression of x for CNT...... 64

Equation 5.6 The expression of overall conductivity by ion concentration...... 64

Equation 5.7 The expression of overall conductivity by individuals...... 64

Equation 5.8 The expression of overall resistance by individuals...... 65

xiv

LIST OF ABBREVIATIONS AND NOTATIONS

AE Acoustic Emission

AgNW Silver Nanowire

AlN Aluminum Nitride

ATR Attenuated Total Reflectance

BNNS Boron Nitride Nanosheet

CB Carbon Black

CNF Carbon Nanofiber

CNT Carbon Nanotube

CVD Chemical Vapor Deposition

DMA Dynamic Mechanical Analysis

DMF N,N-Dimethylformamide

DSC Differential Scanning Calorimeter

EIT Electrical Impedance Tomography

EMI Electromagnetic Interference

ESP Electrospinning

FHC Fricke-Hamilton-Crosser

FTIR Fourier-Transform Infrared Spectroscopy

FWHM Full Width at Half Maximum

LED Light Emitting Diode

MOF Metal-Organic Framework

NFC Near Field Communication

OPS Fiber Optics Based Spectrum

PAN Polyacrylonitrile

PAni Polyaniline

PAni-EB Polyaniline Emerald Base

PAni-ES Polyaniline Emerald Salt

PU Polyurethane

PVP Polyvinylpyrrolidone

RFID Radio Frequency Identification

SCC Stress Corrosion Cracking

SEM Scanning Electron Microscope

SHM Structural Health Monitoring

TGA Thermal Gravity Analysis

TiO2 Titanium Oxide

TMA Thermal Mechanical Analysis

VATRM Vacuum Assisted Transfer Resin Molding

XRD X-Ray Diffractometer

ZnO Zinc Oxide

xvi

1. CHAPTER I

INTRODUCTION

1.1 Nanocomposites

Composite materials consist of two or more components with distinct properties. By combining different components with a proper processing procedure, the product uses the advantages of each[1-3]. Nanocomposite includes components of nanometric scale (1-1000 nm).

For example, carbon nanomaterials, metallic nanoparticles, nanofibers, nano-sheets and other nanostructures[2, 3]. With careful designs of materials selection, processing methods[4, 5], structural design[6, 7] and post-treatment procedures, not only the advantages of components could be well expressed, but also new and functional properties might be developed.

Currently, due to its excellent design feasibility, nanocomposite is one of the most explored material are in thermal management, sensors, biomedical, structural, environmental and energy applications[8-13], etc. Usually, nanomaterials used for nanocomposites include metallic nanomaterials, ceramic nanostructures, and polymeric nanostructures. The matrix used for nanocomposite are generally polymers. By mixing the nano-additives and polymer, functionalized nanocomposites can be prepared. On the other hand, the nano-sized additives are not the only precursor for nanocomposite[14]. Molecular additives and polymer can be the precursor for nanocomposites, for example, the emerging MOF (metal-organic framework) nanocomposite[15].

Conductive application, such as sensor, is one of the major fields of functionalized nanocomposite[16, 17]. By adding conductive nano-additives into polymer matrix, the polymer becomes conductive when percolation threshold is reached: with the addition of conductive nano- additive, there is a critical concentration of the nano-additives in which the conductivity of nanocomposite will increase significantly[18-20]. As a visual concept, conductive path could be

well illustrated as shown in figure 1.1. (If not specify mentioned, the word conductivity in the following chapters means electrical conductivity.)

Figure 1.1 Dispersion (aggregation) of nano additive and formation of conductive path in nanocomposite.

The relationship between conductivity and aggregation of nanoparticles were widely investigated[19, 20]. The mechanical property of the nanocomposite decrease dramatically when the content of nano-additives reaches a critical point[21]. As a consequence, a balance between dispersion and aggregation was investigated. Aggregation of nano particles is a common phenomenon in nanocomposites. However, such a phenomenon is not well studied on both experimental and theoretical aspects.

1.2 Nano-additives

Common nano-additives for nanocomposites including metallic nanomaterials, carbon nanomaterials and ceramic nanomaterials. For example, silver nanoparticle, gold nanoparticle, silicon nanoparticle, carbon nanotube (CNT), graphene, carbon black (CB) (figure 1.2), boron nitride nanosheet (BNNS), quantum dots, ZnO nanowire[22], etc. Nano additives have many advantages. CNT has a low density, high conductivity, high aspect ratio and high stability under

ambient conditions. Graphene has a high surface area and high conductivity. With multiple selection of nano additives, functionalized nanocomposite can be designed. One of the most focused area is conductive nanocomposite functionalization[21, 23]. In order to tailor the electrical properties, four kinds of nano additives are usually considered: metal, semi-conductors

(silicon, transition metal oxide/nitride, such as ZnO, TiO2, AlN, etc.), carbon nanomaterials, and conductive/conjugated polymers[22].

b c

Figure 1.2 Carbon nanomaterials: (a) Carbon nanotube; (b) Graphene; (c) Carbon black[24].

Metallic nano-additives such as silver nanoparticle, nickel nanoparticle, have a high density and normally causes aggregation or agglomeration problems during processing. Carbon

nanomaterials have a much lower density which is normally less than 2.1 g/cm3 of graphite, except that the nanodiamond is about 3.1 g/cm3. The density match between carbon nanomaterials and polymer solutions allows an easier processing procedure. However, the molecular interaction between carbon nanomaterials and polymers is always challenging[25, 26] because of surface tension or high concentration of additives led to the rise in viscosity. The conductivity of conjugated polymers is usually in semi-conductive range, 10-5~104 S/cm[27, 28].

By doping with various chemicals, these polymers could be changed from insulator to good conductor. For example, polyaniline (PAni) has two forms which are emeraldine base (PAni-EB) and emeraldine salt (PAni-ES). The transformation between PAni-EB and PAni-ES could be controlled by adjusting the pH[27, 29, 30]. As a result, these conjugated polymers are widely used for sensing applications, such as humidity sensor, acid sensor and base sensor. However, such a promising material encounters the miscibility issue[31-33]. Most conjugated polymers contain large-conjugated molecular structure, the rigidity of the polymer chain is very high and which make them hard to dissolve in common solvents.

One way to increase miscibility is to reduce the molecular size, and match the solubility parameter. Isotropic pitch is an amorphous material with high carbonaceous content and its molecular weight is about 400 g/mol. By using Soxhlet extraction, soluble pitch content can be extracted and mixed with polymers. However, the molecular size of pitch is small and in a disk- like shape, the formed mixture is not conductive. Normally, viscosity of polymer solution comes from the entanglement of polymer chains. Higher molecular weight usually means a higher entanglement possibility. However, as a disk-like molecule, pitch has low intramolecular interaction (such as entanglement and hydrogen bonding) which causes the pitch solution to have a low viscosity[34]. Such a viscosity is not favorable for processing methods like spin coating and electrospinning[34]. Some researchers used solvents to improve the solubility of pitch and its miscibility in polymers[35-37]. Their carbonized nanofiber product shows a high porosity and low mechanical properties.

Currently, studies about sensors based on nanomaterials were focused on semi- conductive nanomaterials, metal nanoparticles, quantum dots and two-dimension materials.

Carbon nanomaterials such as carbon nanotube, carbon nanofiber (CNF), carbon black and graphene are commonly used. CNF could be classified into two categories: short nanofiber and continuous nanofiber[38, 39]. The short CNF has remarkable mechanical properties and the continuous CNF has various advantages, and could be used as electrode[40], carrier[41] and scaffold[42]. The major difference between these nano-additives are shape and geometry. CNT and CNF could be considered as one-dimension material. Graphene is a typical two-dimension material. CB could be considered either zero-dimension or three-dimension material based on the scale of interest.

1.3 Nano Fabrication

The fabrication method is an important way to define the final properties of nanocomposite. Traditional methods include spin coating, electrospinning (ESP)[43], spray coating, casting, chemical vapor deposition (CVD)[44], ion sputtering, melt processing, in-situ polymerization[33], and other common composite processing methods, such as vacuum assisted resin transfer molding (VATRM)[45].

Generally, spin coating is based on high speed rotation and solidification of polymer solution[46, 47]. The high speed rotation causes a fast evaporation of the solvent in polymer solution. Together with the centrifugal force, the spin coater spins out the polymer solution and forms a thin film on the substrate[47]. The spin coating process consists of 3 steps: solution dropping, solution thinning and solution drying. The controlling parameters in spin coating are solution viscosity, ambient humidity, solution evaporation speed[47, 48]. The morphology of obtained thin film is a complex product of parameters. For example, high viscosity requires a higher spinning speed. The evaporation speed also increases with the spinning speed. On the other hand, high ambient humidity affects the thin film morphology when applying a lower

spinning speed. As a result, the affinity among solution, moisture, polymer and nano additive, temperature, and pressure should be considered in the design of spin coating process[48, 49].

ESP process is based on the electrostatic force between positive and negative electrical charge[43, 50]. During the process, polymer solution spins out from orifice of the container and jets out semi-solid fiber, then the fiber is stretched and solidified in the air while dropping. As a result, a thin film consists of innumerous nanofibers could be collected. The controlling parameters for spin coating and ESP are very similar. Solution viscosity, applied voltage and ambient conditions are the most important factors. However, in the ESP process, the introducing of high voltage electric field results a complex multi-physics situation[43]. As a result, additional considerations should be given to the electric and magnetic properties of the spinning solution.

Such as the surface charge distribution and conductivity of the spinning solution. Spray coating is by using high pressure air to spray solution onto a surface[51-53]. Spray coating is good at forming a thin layer of coating in a large area at a lower cost and a faster speed. However, the balance between viscosity, mold geometry and curing/solidification time should be well controlled.

a thinning of solution drying of solution

solution

spinning substrate spinning substrate spinning substrate

b solution pump

high voltage Electrospinning setup electric field

rotating drum

Figure 1.3 Illustration of (a) spin coating and (b) electrospinning experiment setup.

Thin film could be made from spin coating, ESP, spray coating, CVD, ion sputtering and in-situ polymerization methods. By controlling the rheology properties of nanocomposite mixture and time of evaporation solvent, thin films from nanometers to hundreds of micrometers could be prepared from a spin coater or ESP setup. Nano-additive affects the polymer solution’s rheology properties significantly. Parameters such as geometry, shape and density of the nano-additive, interaction among nano-additive, solvent and matrix are key factors. Carbon nanomaterials such as CNT, graphene and carbon black could be used to study the effect of geometry. Functionalized material could be used to study the interaction between additive and matrix[54, 55]. The solvent evaporation speed is another major factor which affects the morphology of spun film/fiber[56,

57]. The balance between solvent evaporation and bubble nucleation-growth affects the porosity and pore size in polymer matrix. In ESP process, due to the large specific surface area of the

nanofiber, solvent’s diffusion and evaporation are much faster than the formation of vapor bubble

(phase separation) and normally results non-porous structure. However, for spin coating process, by controlling the selection of solvent and spin speed, various type of thin films can be made from non-porous to highly porous. Moreover, with the addition of nano-additives, evaporation speed and bubble nucleation-growth could be further controlled.

Another effect of nano fabrication methods on nanocomposites is the formation of anisotropic structures[58, 59]. The scale of nano fabrication is similar with the size of nano additives. For example, by using ESP method, CNT can be embedded inside the electrospun nanofiber. Due to the confinement effect of nanofiber, the embedded CNT is always aligned with the nanofiber. As a result, the prepared nanofiber mat is always anisotropic in mechanical, electrical and thermal properties.

1.4 Conductive Network

The conductive nanocomposite is based on the formation of conductive network[60, 61].

Generally, the conductivity of nanocomposite increases with the addition of conductive nano additives. However, the relationship between conductivity and concentration of nano additive is not linear. Percolation threshold theory is most widely applied to describe the relationship[61,

62]. The conductivity increases slowly with the addition of conductive nano additive. However, at a certain percentage region, the conductivity increases dramatically with the addition of conductive nano additive. It is at this composition region, the conductive network forms.

However, with additional nano additive, the conductive network is saturated. As a result, the conductivity does not increase much more.

More than one method can be used to form the conductive network. However, most methods are just adding more and more conductive nano additive to ensure a good conductivity.

1.5 Structural Health Monitoring

Structural health monitoring (SHM) is the method of online/real-time/in-situ monitoring of the health status of infrastructures[63, 64]. Especially, in the composite area, SHM become

more and more important[65]. The failure of composite structure usually starts with micro-cracks formation. In order to avoid unexpected catastrophic failure, it is important to monitor the structure integrity of composite. Common SHM method involves acoustic, optical, and electrical method[66-68]. The SHM can either use distributed sensors or tomographic method. However, most of the methods provide only 2D analysis. Sometimes, a 3D monitoring of structural health is also needed. For example, building vibration in seismic effect[69, 70] and chemical penetration or crack propagation in composite acid storage[71].

Electrical method is the easiest and cost-effective method to be applied in to composite structure. However, the electrical or sensing signal received is complex to be interpreted[67].

Optical method might be the most accurate and sensitive method. However, the optical system is not cost effective[72] and an additional training is required for the result analyst to understand optical signals. Additionally, for a composite structure, any outlet wire is not favored. A wireless sensing system is needed for the SHM of composite structure.

1.6 Thesis Statement

The goal of this research is to develop nanocomposite based conductive sensors for chemical penetration problem of composite and investigate the mechanism of chemical penetration. Such a nanocomposite based sensor will be achieved with materials such as molecular additive, conductive nanofiller and conductive polymer and appropriate manufacturing methods such as electrospinning, pyrolysis and in situ polymerization. This thesis will involve the following components:

1. The addition of pitch based carbonaceous additive. Carbon yield, modulus and conductivity of the pyrolyzed carbon nanofiber will be improved due to the pitch could form a larger and perfected graphitic structure easier than PAN.

2. Using of proper conductive nanofiller. Proper additive could tune the electrical properties of nanocomposite. Conductive nanofiller forms conductive network in nanocomposite which is the key of its sensing ability.

3. Conductive polymer has response to different chemicals. Due to the doping effect, the use of conductive polymer endow the designed sensor has selectivity to different chemicals.

Additionally, phase separation induced conductive network can be tailored with changing the polymer blend’s composition.

4. Diffusion of ions. With the help of the developed sensor, the diffusion of different ions was noticed and which is the key mechanism of the observed sensing result. By modeling the diffusion process, the sensing behavior can be understood.

5. Wireless sensing. Wireless sensing design does not require metallic wire go in-and-out of composite structure. Based on the change of conductivity of nanocomposite in modified RFID tag, upon chemical attacking, the RFID gives response quickly.

The target of this research is the development of sensor for chemical penetration in composite structure.

2. CHAPTER II

THE EFFECT OF CARBONACEOUS MOLECULAR ADDITIVE ON ELECTROSPUN

CARBON NANOFIBER

2.1 Background

Molecular additives are the mostly used additive in chemical, food, environmental, automobile and medical industry. The biggest advantage for molecular additive is its dispersion is very easy. However, molecular additives have to be carefully designed for each system. In the case of carbon fiber, the precursor Rayon, polyacrylonitrile (PAN) and pitch has to be carefully chosen. As a result, additives are rarely use in the production of carbon fiber. Here in this chapter, the possibility of using pitch and PAN mixture is explored.

Pitch and polyacrylonitrile (PAN) are the most common precursors for the fabrication of high performance carbon fibers. However, reducing the cost of carbon fibers remains very challenging. Few attempts were made to increase carbon yield and reducing cost by using polymer blends or additives[73-76]. Use of carbon nanotubes[76-79], graphene ribbon[50], lignin[73, 76], and other polymers[74] were studied. As a result, multiple benefits were demonstrated and properties were improved. However, some challenges were unsolved such as, how to increase the carbon yield, reducing cost, and keep the mechanical properties unchanged[73]. Another approach of obtaining carbon fibers with less defects is to reduce the fiber diameter less than 2 m. By reducing the diameter, the defects concentration could be minimized and the fiber orientation could significantly enhance the physical properties of carbon fibers, such as thermal, electrical, and mechanical properties. Commercially, the carbon fiber’s diameter is in the range of 6–8 m. Using traditional wet and melt spinning process does not guarantee yet any reduction in fiber diameter. However, electrospinning is a promising method to fabricate nanosized fibers with tailored diameter length and orientation[43, 80]. All these features

and dimensions can be controlled using high voltage, spinning distance, and the appropriate syringe orifice[81]. Long and coworkers[82] and Bittner and coworkers[83] studied electrospinning of various small molecules. These small molecules have strong intermolecular interaction and behave like entangled polymer chains[34, 82, 83]. In contrast, the disk-shaped pitch molecules do not have such an intermolecular interaction, pitch solution could only be electro-sprayed into beads[34, 84, 85] or whiskers[35]. A binder for pitch molecules is needed to create the entanglement between molecules. An addition of polymer in pitch solution could solve this problem by increasing system viscosity. Yang and coworkers studied PAN and pitch blends via electrospinning process[35-37]. Their study showed spinning-ability was sensitive to the concentration of PAN and obtained microfibers with a diameter of 2 m. Murugesan et al.[86] and Yan et al.[87] studied thermal properties of carbon nanofiber produced by electrospinning of polyimide/pitch blends. The presence of pitch helped in the formation of carbon crystallite, and enhanced thermal conductivity, electrical conductivity, and modulus significantly[36, 37, 86-88].

The property of as-spun fiber will be highly affected by the degree of molecular alignment[80, 89, 90]. The key parameters that affect the alignment of polymer chain during the spinning process are macromolecular chemistry, viscosity, surface tension of solution and spinning condition[43, 80, 81]. Among these factors, the chemistry and viscosity of precursor are the critical parameters for the formation of thin and high performance carbon fiber. The rheological behavior of the precursor will dictate the feasibility of the spinning process. The rheology behavior of the blend solution will change drastically from polymer blends[91], nanofiber/polymer[92], and nanoparticle/polymer[93, 94]. Rheology property of the blend solution is sensitive to the physical properties of the materials and also depends on the surface tension and miscibility of the two materials. Elias et al. studied the effect of modified silica nano- spheres on an immiscible polymer blend system[95]. In their study, the polymer viscosity changed dramatically with the silica concentration. Many similar studies confirmed that a change in miscibility would lead to a great difference in their rheology behavior[96, 97].

In this chapter, miscibility of Pitch and PAN will be studied, especially the interaction at molecular level. In order to have a better understanding of rheology behavior of the blends, miscibility property will be investigated using differential scanning calorimeter (DSC) and

Fourier transform infrared spectroscope (FTIR). Miscibility behavior of pitch and polymer blends is a delicate issue because of the complex chemistry of pitch material[98, 99]. Pitch such as A240 is a reactive and isotropic pitch material with partial solubility in N,N-dimethylformamide

(DMF). Dickinson provided an average structure for A240 petroleum pitch: an aromatic multi- ring structure with short aliphatic side groups (figure 2.1)[100]. Solubility is also an interesting parameter for electrospinning process[101], DMF soluble pitch content will be used in this chapter to form a homogenous viscous solution. After proper stabilization and carbonization process, clean carbon nanofibers were successfully fabricated. Structural, mechanical, and conductive properties were studied, respectively.

Figure 2.1 Average chemical structure of A240 molecule.

2.2 Materials and Characterization

2.2.1 Materials and Sample Preparation

PAN (6% methyl acrylate copolymer, Mn = 100,000 g/mol) was purchased from

Scientific Polymer Inc. (Ontario, NY) and used as received. Pitch material (A240, Koppers) was refluxed in DMF by Soxhlet and filtrated before use. For the miscibility study, all mixtures were stirred at 60 ˚C for 2 days. The total concentration of PAN and pitch in DMF solvent was controlled to be 1 wt.%. The fraction ratios for pitch and PAN were labeled as pitchXPANY respectively. For example, pitch3PAN7 means a blend of 30 wt.% of pitch with 70 wt.% of PAN.

Then the obtained mixtures were dried in oven (60 ˚C) for 2 days. For the rheology study and fiber electrospinning, the concentration of PAN was kept at 7 wt.%. The parts of pitch on PAN were changed from 0 to 1. For example, a pitch0.5PAN1 sample means 0.5 part of pitch against 1 part of PAN.

2.2.2 Electrospinning Set Up

Electrospinning process was carried out with a syringe pump (New Era Pump Systems,

Inc., NE-300, Farmingdale, NY), voltage controller (Stanford Research Systems, Inc. Model

PS375, Sunnyvale, CA), and #17 needle. Voltage between needle and collector was 15 kV, distance between needle and collector was about 30 cm, feeding rate was about 3 μL/min. Fibers were collected on a cylindrical rotator (linear velocity, 7.5 m/s).

2.2.3 Stabilization and Carbonization

The prepared precursor fibers were thermally stabilized in air with a tube furnace

(Thermo Scientific, Blue, Waltham, MA) by a multistep procedure. The fibers were heated up to

250 ˚C then 280 ˚C at a rate of 2˚C/min, and the holding time is 2 h for each temperature stage.

After cooling to ambient temperature, the stabilized fibers were carbonized in argon atmosphere at 1000 ˚C for 30 min, the heating rate was 2 ˚C/min.

2.2.4 Characterization

DSC (Q2000, TA Instruments, New Castle, DE) was used to show the thermodynamic property of blends, glass transition temperature (Tg) was determined with a heating rate of 10

˚C/min in the range of -40 ˚C to 150 ˚C. Interaction and reaction of PAN and pitch was studied by a FTIR spectroscope (Nicolet is50-FTIR, Thermo Scientific, Waltham, MA), it was set to scan

128 times and a resolution of 4 cm-1 under attenuated total reflectance (ATR) mode. Rheology study was processed on an Anton Paar MCR-302 Modular Compact Rheometer at 25 ˚C with cone fixture. A valid torque data range for rheometer used in this experiment was between 0.01 mN⋅m and 200 mN⋅m. Carbon yield of the precursors were measured using a thermal gravity analysis (TGA, Q500, TA Instruments, New Castle, DE) in N2 atmosphere, the heating program was set to be ramping to 800 ˚C at 10 ˚C/min. Scanning electron microscope (SEM, ProX,

Phenom, Eindhoen, The Netherlands) was used to see the morphology of as-spun, stabilized, and carbonized fibers. Raman spectrum of stabilized and carbonized fiber mats was recorded with a

Renishaw in-Via Raman Microscope (633 nm laser). The thickness of carbonized fiber mats was measured by thermal mechanical analysis (TMA, Q400, TA instruments, New Castle, DE) in expansion mode, and mechanical properties were tested in Film/Fiber Tension mode at room temperature. Conductivity of carbonized fiber mat thin strips was tested with Multimeter/ Data acquisition system (Keithley 2700, Cleveland, OH) under optical microscope/probe station

(Signatone 1160 Series Probe Station, Gilroy, CA) at room temperature.

2.3 Results and Discussion

2.3.1 Thermodynamic Study on Glass Transition

The glass transition temperature of PAN was measured by using DSC and it is 102 ˚C.

However, the glass transition temperature of the pitch will be affected greatly by its chemical composition[102]. A240 was examined and it is obvious that Tg for pitch7PAN3 sample could not be determined easily (figure 2.2a). By differentiating the heat flow as function of temperature,

a peak could be identified (figure 2.2b) at 55 ˚C. Using the same analytical method, Tg of PAN was found at 106 ˚C. As shown in figure 2.2a, DSC heat flow curves of each sample showed a gradual change as function of temperature. Base on normal Tg determination method, it was hard to determine the exact Tg values for pitch3PAN7 and pitch5PAN5 blends. The interceptions of tangent lines are highly dependent on the choice of onset and offset point. This Tg determination method is not precise and rigorous against the definition of glass transition temperature, but it is indeed useful for determination of Tg in such a complex situation. Again, it is much easier to use the derivative analytical method as shown in figure 2.2b. Derivative heat flow curve of PAN showed two peaks, this may due to the effect of 6% methylacrylic content in PAN or different amorphous states of PAN[103]. Methylacrylate group could affect the surrounding acrylonitrile segments and increase amorphous form percentage. Methylacrylate group could also form hydrogen bonding with cyano group, it could act as an obstacle for the movement of segments and results in a higher peak at 111˚C. The gradual change of peaks in figure 2.2b indicates PAN and pitch are thermodynamically miscible.

The gradual change of Tg indicates that the interaction or reaction between pitch and

PAN caused their miscibility. As introduced, the A240 petroleum pitch is more reactive than any other mesophase pitch. The various effects of pitch analyzed above showed probably some interaction and/or chemical reaction occurred between PAN and pitch during the blending process. Even the blending process was under moderate temperature, the possibility of chemical reaction had to be considered.

Figure 2.2 DSC results of PAN, pitch, and their blends. (a) Glass transition of the blends; (b) Derivate curves of (a).

2.3.2 FTIR Study of PAN/Pitch Blends

In order to have a better understanding of the effect of mixing for these blends, FTIR was used to characterize functional groups’ differences in blends. Figure 2.3a shows C-H stretch region for PAN and pitch, peaks were normalized. For PAN and pitch raw materials: peaks at

-1 -1 -1 -1 3030 cm (unsaturated C-H), 2960 cm and 2850 cm (-CH3), 2930 cm (-CH2-) are related to

C-H stretch on PAN backbone, methacrylate units, and edge of pitch molecules. The gradually change shown in figure 2.3a indicates the blends were well prepared. Figure 2.3b shows smaller wavenumber region differences. Peak at 1730 cm-1 is attributed to C=O stretch from methacrylate units[104]. Peak at 1600 cm-1 indicates the aromatic system in pitch and the blends. Peak at 1450 cm-1 is C-H bend or methyl group of methacrylate. The peak around 1375 cm-1 indicates the short aliphatic groups or aldehydes groups of pitch and its blends, also the weak peak at 2735 cm-1 leads to the same assumption (aldehydes groups exists). Peak at 1170 cm-1 clearly shows the existence of acrylate in PAN and the blend. Figure 2.3c shows 2240 cm-1 peak, which is the fingerprint peak for -CN group.

Figure 2.3 FTIR results of PAN, pitch, and their blends. (a) 3000 cm-1 region for C-H. (b) 900– 1800 cm-1 region for C-O bonds and C-C bonds; (c) 2250 cm-1 region for cyano bond of PAN.

However, beside the peaks related to PAN and pitch, new peaks are found for the blends.

The strong peaks at 1667 cm-1, 1266 cm-1, 1252 cm-1, and 1100 cm-1 clearly showed that new bonds were formed during blending process. Peak at 1660 cm-1 could be a response from C=C or

C=N group, and there are two possibilities: (1) the heteroatom on pitch molecules reacted with the cyano group of PAN; (2) the reaction between pitch and PAN side groups changed the structure of pitch, and results in the 1667 cm-1 sharp peak. Both explanations are possible because of the appearance of 1266 cm-1, 1252 cm-1, and 1100 cm-1 peaks. These peaks may come from acrylates or aromatic esters[105]. As a conclusion, these peaks come from newly formed aromatic esters.

On the other hand, -CN group fingerprint peak (cyano stretching) does not change much

(about 1 cm-1) with the addition of pitch to PAN. As shown in figure 2.3c, the 2243 cm-1 normalized peak of PAN has a slight red shift and became broader after blending with pitch. Such a small shift was also observed by Goh et al.[106] and Yeo et al.[107] in the study of miscibility of PAN with other polymers. The red shift of cyano bond is possibly contributed by p-orbital cyano band[108]. In this case, the red shift probably came from the interaction between conjugated system on pitch and cyano bonds.

The formation and interaction of new bond and the interaction on cyano bond are supposed to be the cause of change of glass transition temperature. The new aromatic ester could create a larger free space than it used to. The larger free space would require less energy for PAN amorphous region to reach a higher energy state, in other words, a lower glass transition temperature. In summary, the reaction (aromatic ester) and/or interaction (cyano group and aromatic rings) contribute to the miscibility of PAN and pitch.

2.3.3 Rheology Study

The rheological characterization was carried out to control further the electrospinning process. Oroumei et al. used lignin and PAN to prepare carbon nanofiber via electrospinning[109]. In their study, the total concentration was set at 18 wt.%, however, the blend of 18 wt.% of lignin was less viscous which causes difficulties in making nanofibers. After spinning, they could only prepare lignin material in the form of beads rather than continuous fiber.

Figure 2.4 (a) Rheology result of PAN and the blends. (b) TGA result of PAN and the blends.

After several trials, the concentration of PAN was set to be 7 wt.% in order to reduce the formation of beads in electrospinning process. The ratio of pitch to PAN was controlled as 0.5/1,

1/1. Results are shown in figure 2.4a. The viscosity of these blends decreased with the increase in shear rate. As the concentration of pitch increase from 0 to 50 wt.% ratio (sample: PAN,

Pitch0.5PAN1, Pitch1PAN1), viscosity has not changed a lot. Especially, while shear rate becomes higher, less difference on viscosity could be recognized. As the limitation of instrument, the highest shear rate measured was 2000 s-1. However, for a real spinning process, the shear speed could be as high as 10 m/s (shear rate: 10,000 s-1)[81]. As a consequence, there should be no significant difference on viscosity for different samples during electrospinning process. Based on the result of rheology test, the electrospinning of PAN and pitch hybrid fibers were fabricated from PAN solution, Pitch0.5PAN1 sample solution, and Pitch1PAN1 sample solution.

2.3.4 TGA Analysis

The as-spun fibers were used for TGA test and results are shown in figure 2.4b. The carbon yield for PAN, Pitch0.5PAN1, and Pitch1PAN1 samples were calculated to be 0.46, 0.51, and 0.55. Same as predicted, the addition of pitch content results in a higher carbon yield of the

CNF. Different from bulk sample, in order to prepare the testing sample, the as-spun fibers were folded after peeled off from the electrospinning substrate. Air would be trapped in the sample, and cause the weight increase during the test. It is very interesting to see that the PAN sample is oxidized more than the pitch added samples. This phenomenon indicates that the degradation behavior of PAN was changed by adding pitch, and which might be caused by the interaction/reaction between pitch and PAN. The initial weight loss of Pitch1PAN1 sample compared to Pitch0.5PAN1 sample may due to the vaporization of small molecules in pitch.

2.3.5 PAN/Pitch As-spun Stabilized, and Carbonized Fibers

Cylindrical fibers were successfully prepared via electrospinning process. Optical microscope images (figure 2.5a, b, c) and SEM images (figure 2.5d, e, f) show the morphology of as-spun fibers. Diameters of all as-spun fibers are generally ranged from 500 to 700 nm (Table

2.1). In figure 2.5(a–c), the color of as-spun fibers become darker with an increase of pitch concentration, from white to red or brown. Most of the fibers had smooth surface. However, due

to the difficulties of fully eliminate tiny insoluble pitch particles (less than 1 lm), some particles were embedded inside as-spun fiber (figure 2.5e, f). The insoluble pitch particles were smaller than fiber diameter, but observable in some regions on the nanofiber surface under microscopes.

Processing procedures are essential for fabricating of carbon nanofibers, such as stabilization and carbonization. In order to avoid potential fusion of fibers, a two-step stabilization process was chosen. The stabilization process at 250 ˚C was for PAN and 280 ˚C was for pitch. Heating rate is also important in avoiding fusion of fibers, here we chose it to be 2˚C/min. Stabilized and carbonized fiber morphology are shown in figure 6 and smooth, clean nanofibers were obtained.

In figure 2.6, carbon nanofibers surfaces are all quite smooth and the diameters are uniform (table

2.1).

Table 2.1 Diameter ranges of As-Spun, Stabilized, and Carbonized fibers. As-spun fibers (nm) Stabilized fibers (nm) Carbonized fibers (nm)

PAN 843±116 747±77 427±70

Pitch0.5PAN1 863±187 534±69 389±39

Pitch1PAN1 795±103 584±70 453±62

Figure 2.5 Optical (a–c) and SEM (d–f) images of PAN (a,d), Pitch0.5PAN1 (b,e), Pitch1PAN1 (c,f) as-spun fibers.

Figure 2.6 SEM images of stabilized (a–c) and carbonized (d–f) PAN (a,d), Pitch0.5PAN1 (b,e), Pitch1PAN1 (c,f) fibers.

2.3.6 Raman Analysis

Shown in figure 8a, Raman results show the clearly difference between different carbon nanofibers. The two broad peaks at about 1350 cm-1 and 1580 cm-1 are the D band and G band.

Normally, D band and G band are referred to the disordered structure (sp3) and graphene-like

ordered structure (sp2). Thus, the sp3 hybrid bonds could contribute to the inter-planar connection and interlock and which could contribute to a higher strength. On the other hand, with the increase of pitch content, the normalized D band become smaller and G band become narrower. It is easier to compare the valley between D band and G band, by increasing pitch content, the valley becomes deeper. Which means the sub-peak at around 1490 cm-1 become smaller and 1580 cm-1 peak (G band) become sharper. According to a previous study, the sub-peak at 1490 cm-1 attribute to amorphous graphitic structure[110]. Decrease of 1490 cm-1 peak indicates less amorphous content appears in pitch added carbon nanofiber.

Figure 2.7b shows sharpness of G band and ratio of G band to D band. In order to have a better illustration of shaper of the G peak, here we define the sharpness of a peak to be the ratio of its normalized height to Full Width at Half Maximum (FWHM). No obvious change in ID/IG could be found on G band, on contrary, sharpness of G band increased with the addition of pitch content obviously. The sharpness of G band suggests the graphite structure becomes larger, thicker, and more uniform. And which generally means higher modulus and conductivity.

Figure 2.7 Raman spectroscopy results. (a) Difference on D band and G Band; (b) sharpness of G band and D/G ratio.

2.3.7 Mechanical and Conductivity Properties

A test for single fiber segment is truly useful for fundamental study of the nature of carbon nanofiber[111]. However, thus studies cannot present the role of defects and the property

in a larger scale. In this chapter, mechanical tests were performed for fiber mats by using TMA in film/fiber tension mode. As shown in figure 2.8, modulus of Pitch1PAN1 carbon nanofiber mat was higher than that of PAN, increased by 20%. Same trend showed on the conductivity result,

Pitch1PAN1 carbon nanofiber had a higher than PAN-based carbon nanofiber. And it also should be noted that, with an additional increasing of pitch content (Pitch1PAN1), the physical properties did not follow a linear increasing. The higher modulus and conductivity suggest better graphite structure (larger) and this agree with the conclusion which made by mechanical test and

Raman spectrum.

Figure 2.8 Conductivity and modulus of the carbon nanofibers.

2.4 Conclusion

Based on DSC results, the addition of pitch lowers the overall Tg of PAN. The sample blends exhibit two Tg values. At this level of study, we can only assume that thermodynamically pitch and PAN are partially miscible. This assumption was confirmed by FTIR results. The interaction of cyano band and the new band at 1660 cm-1 showed the p (cyano)–p (pitch

conjugate) interaction and reactions relate to transesterification. Rheology study was helpful to have a better understanding of electrospinning process. The viscosity of the blends did not change significantly as the amount of pitch increased which provides a stable condition during electrospinning process.

In this chapter, homogenous carbon nanofibers were produced using all various mixtures between PAN and pitch. The final product, the carbon fibers from pitch added precursor showed an enhancement on mechanical property and electrical conductivity. Raman spectrum confirmed the increase of ordered structure in PAN/pitch based carbon nanofibers by analyzing the sharpness of G band. As a result, the addition of pitch increases the degree of alignment because of the high amount of liquid crystal present in the pitch and eventually it enhances the physical properties of the carbon nanofibers.

3. CHAPTER III

SELF-ASSEMBLY AND SURFACE TENSION INDUCED FRACTAL CONDUCTIVE

NETWORK IN TERNARY POLYMER SYSTEM

3.1 Background

Polymer blends has complex phase behavior depending on their affinity to each other.

The incompatibility between polymers is either a draw back or an advantage for functionalization.

The phase

Conductive nanocomposites are used widely in various applications, such as sensing, flexible electronics, tissue engineering, and electromagnetic shielding[71, 112-114]. Traditional conductive polymers are made using conductive fillers. Carbon nanotube (CNT), silver nanowire

(AgNW), graphene, and carbon black are commonly used fillers to form a conductive network.

The content and dispersion of fillers are key factors in improving the electrical properties[115-

119]. The formation of aggregates and their fractal dimensions has a big impact on both electrical and mechanical properties[120, 121]. The phase separation can also be used as a tool of forming a conductive network[122]. The efficiency of conductivity can be increased dramatically if one of the phases in a continuous phase separation is conductive[123, 124]. On the other hand, the mechanical property of the polymer blend can be tuned by changing the composition as well[125].

Polyaniline (PAni) is the most well studied conductive polymer which could be synthesized in the form of nanowire[126, 127], nanofiber[128, 129], nanosheet[130], and many other nano-structures[131]. The conductivity of PAni comes from its doping[131], and it is usually used for sensing applications. Adding acid or base, PAni can be changed from emerald base to emerald salt and vice versa. PAni is a brittle material due to the rigid polymer chain structure and has a poor compatibility with other polymers[132]. Polymer mixing is one of the major methods to soften polymers, the poor compatibility of PAni brings an additional difficulty

to use PAni. Some efforts were carried out using the in situ synthesis of PAni and found great use in sensing applications[129, 133, 134]. However, this approach is limited to the laboratory scale.

Polyvinylpyrrolidone (PVP) is another alternative polymer that can be dissolved in solvents such as water, acetone, ethanol, and N,N-dimethylformamide (DMF). It is widely used as an adhesive material due to its good affinity to common surfaces. PVP can be used as a stabilizer during the synthesis of colloidal PAni[135]. The formation of colloidal PAni significantly increases its processing ability[136, 137]. As a result, by synthesizing PAni in PVP, the polymer blend becomes both flexible and conductive.

However, the conductive mechanism in CNT-based nanocomposites relies on the formation of a conductive network[118]. Previously, we have developed CNT/epoxy resin-based nanocomposites for sensing of acid penetration in polymer and composite structures[71]. A similar PAni/epoxy sensor was used as a comparison to illustrate the function mechanism of the sensor. However, the compatibility between PAni and epoxy resin was an issue and limiting factor[138]. An efficient and conductive polymer blend or nanocomposite must be found. Indeed, it seems that the dried PAni/PVP blends are lacking in flexibility. However, wet PAni/PVP blends are viscous and pliant. Adding a third component to a PAni/PVP blend may form a material with the best compromise. The ideal polymer should be able to hold the structure in order to avoid viscous flow and be flexible enough. PVP can be incorporated into polyurethane

(PU). The formed PVP/PU semi-interpenetrating polymer network restricts the ability of PVP to absorb moisture[139].

In this chapter, balance between surface tension, van der Waals force, and friction were found to be key in the formation of a fractal conductive network[140]. By changing the composition of this ternary polymer blend, we observed the formation of dendritic fractal conductive network and scattered conductive network. First, we synthesized PAni in PVP solution with a modified recipe. The prepared PAni/PVP mixture was dried partially in an oven.

The third phase, PU, was then added to enhance the mechanical property. At last, we applied the

1-5-10 sample, which has the lowest conductivity, onto a nonwoven fabric. A simple test using an origami crane was carried out and shows that the material can be conductive and flexible.

3.2 Materials and Characterization

3.2.1 Synthesis of PAni

Polyaniline was synthesized using a modified chemical oxidation method[141]. Aniline

(2 mL) was mixed with hydrochloric acid (100 mL, 1 mol/L) at room temperature.

Polyvinylpyrrolidone (Mw, 10,000 g/mol) was dissolved into the solution. Then, 30 mL of H2O2

(30 wt.%, stabilized ACS) was added in order to initiate the polymerization reaction. The solution was vigorously stirred until its color changed into dark green and let it sit for one night.

3.2.2 Preparation of Polymer Blends

The prepared PAni/PVP compound was then dried and re-dissolved with DMF. Then, polyurethane (PU, ESTANE 5708) was added and mixed for 12 h. The polymer blend was dried in an oven. The thin film sample was prepared by re-dissolve the prepared ternary blend into

DMF and cast onto glass slide at room temperature. The concentration of this solution was about

0.5 g/mL. The bulk sample was made by pressing the material between two glass slides. The samples are noted with the sequence of PAni-PVP-PU. For example, 1-5-10 represents the sample consist of 1 part of PAni, 5 parts of PVP, and 10 parts of PU.

3.2.3 Characterization of Polymer Blends

The morphology of the final mixture was examined using a Zeiss Axio vert.1 optical microscope. Samples were vacuum-dried at room temperature and sputtered with gold before characterized by a scanning electron microscope (SEM). The electrical conductivity was measured with a Keithley 2700 Multimeter by the 2-point method. Before each test, the sample was baked at 60 °C oven for 30 min in order to drive out the moisture. After being taken out of the oven, the sample was quenched and the electrical signal was recorded using a Multimeter.

3.2.4 Fabrication of Flexible Conductive Fabric and Test

A volume of 2 mL of DMF was added to 1 g of the 1-5-10 sample in order to reduce its viscosity. Then, the solution was dropped onto the nonwoven fabric (Kimwipes paper, Kimtech).

The coated fabric was then dried in an oven at 80 °C. After drying, the fabric was placed above a hydrochloric acid beaker for the doping of PAni. The doped nonwoven fabric was then cut and folded for testing. The fabrics were sputtered with gold before being observed with SEM (Pro X,

Phenom). The SEM measurement was taken quickly under a lower voltage in order not to jeopardize the instrument. Conductivity was measured with a low resistivity meter (MCP-T610,

Mitsubishi Chemical Analytech) by the 4-point probe method.

3.3 Results and Discussion

3.3.1 Morphology

First, control groups of PAni blends were prepared. The 1-0-1 thin film sample in figure

3.1a shows PAni and PU are immiscible that the darker phase is PAni and the other one is the PU phase. The 1-1-0 thin film sample (figure 3.1d) has a scattered fractal network morphology.

However, in figure 3.1b and .1e, a typical phase separation morphology could not be identified because all phases were very small and hard to distinguish. SEM images (figure 3.1c and 3.1f) show PAni was mixed well with other polymers and formed a nanowire/matrix morphology.

Figure 1c shows the SEM morphology of the PAni/PU mixture (1-0-1 sample). PAni forms a 3D nanostructure which consists of nanowire and covers on the PU granular surface, which also confirms the poor miscibility between PAni and PU. However, in the in situ polymerized

PAni/PVP sample (1-1-0 sample), nanostructures can be seen as well. Different from the

PAni/PU sample, it seems the PAni and PVP are mixed well and that no obvious phase boundary can be identified.

Figure 3.1 Optical (a, b, d, e) and SEM (c, f) images for thin film (a, d) and bulk (b, c, e, f) of the 1-0-1 sample (a, b, c) and the 1-1-0 sample (d, e, f). The scale bars in optical images are 100 μm and for SEM images are 10 μm.

Figure 3.2 illustrates the morphology results of the ternary polymer blends. Figure 3.2a shows a concentrated PAni network form in the 1-1-1 sample. The 1-1-1 sample has the same

PAni/ PVP composition, and PAni nanoparticles were aggregated. A few aggregates could be identified in the optical image of 1-5-10 sample (figure 3.2d), which are noted by red arrows.

Bicontinuous phase separation occurred in this sample, in which the darker phase is the

PAni/PVP phase. However, in the optical image of the 1-10-5 thin film sample (figure 3.2g), no

PAni aggregation could be observed. By decreasing PU content and increasing PVP content, the

PU phase becomes isolated and round (figure 2g red arrow), which indicates that the change in morphology was attributed to surface tension. For the bulk samples, various morphologies were identified. A typical phase separation morphology can be found in figure 3.2b and 3.2e. The high contrast image suggests PAni/PVP and PU are two major phases with high surface curvature.

Figure 3.2d is similar to figure 3.2e that the bicontinuous phase separation occurs. However, in figure 3.2h, sample 1-10-5, only one phase could be identified. This is due to the fact that the

PVP phase absorbs a large amount of moisture and then covers the surface of the PU phase. Also,

the smooth phase boundary and homogeneous PVP phase suggest PAni exists as a colloid in the

PVP matrix.

Figure 3.2 Optical (a, b, d, e, g, h) and SEM (c, f, i) images for thin film (a, d, g) and bulk (b, c, e, f, h, i) of the 1-1-1 sample (a, b, c), 1-5-10 sample (d, e, f), and 1-10-5 sample (e, h, i). The scale bars in optical images are 100 μm and for SEM images are 10 μm.

The comparison of figure 3.1d and figure 3.2a (which both contain the same PAni/PVP ratio) reveals that the 1-1-0 sample formed a scattered fractal network; however, the 1-1-1 sample formed a concentrated fractal network (dendritic network). By adding the PU phase, surface tension that formed on the PU and PVP boundary forced the PVP phase to be spherical. The aggregation of PAni nanoparticles is blocking this movement. As a result, the scattered network

(figure 3.1d) becomes concentrated (figure 3.2a). The SEM result confirms this assumption. In figure 3.2c, a clear phase separation and assembly of nanowires or nanorods could be identified.

It is easy to identify the brighter phase is the PAni/PVP phase due to the clear phase boundary

and the nanorod feature. This morphology also helps us to identify the phases in figure 3.2f and

3.2i. The assembly of the nanorod structure in the PVP phase in figure 3.2c and corresponds to the fractal structure in figure 3.2a. Interestingly, the orientation of PAni nanorods in figure 3.2c is about 40°. As this angle is only seen in this location of sample, we believe this is a self-assembly and flow-induced feature. For example, figure 3.3e shows a more random-like feature.

The transformation of PAni from colloid to nanorod is dependent on its composition. It was found that the dopant of PAni may be a key role for its self-assembly behavior. However, in our case, the hydrochloric acid was used to eliminate the effect of side chains[142]. Generally, the in situ polymerized PAni exists as a colloidal nanoparticle in the PVP phase[143]. At the beginning of the polymerization process, the PVP sustains a certain amount and size of PAni colloid nanoparticles. As the polymerization process progresses, the formation of PAni colloids becomes more dominant. At a critical composition, the amount of PVP cannot stabilize the system and PAni colloids collapse into nanostructures[144]. It is known the PVP stabilizer wraps the polymerized PAni colloid nanoparticle. When there is less stabilizer in the mixture, there is a higher probability of collision between colloids. As a consequence, PAni nanostructures are grown. Other stabilizers, such as ethyl(hydroxyethyl)-cellulose, have a similar effect on the morphology of PAni. With additional sonication, the ethyl-(hydroxyethyl)-cellulose disintegrates and a tree-like fractal morphology of polyaniline can be observed. Once the concentration of the stabilizer is increased, it becomes harder to separate the stabilizer from PAni colloids[145].

Compared to other samples in figure 3.2, the 1-1-1 sample has a higher PAni to PVP ratio.

Figure 3.3 Phase separation situations in the observed ternary polymer blends. (a) Van der Waals force and surface tension are balanced with friction. (b) Van der Waals force balanced with surface tension. (c) Surface tension dominates. (d) Illustration of the balance between forces. (e) SEM image of part a.

Figure 3.3 shows the observed three types of phase separation. When equilibrium was achieved between the van der Waals force and surface tension, the morphology of phase separation changes is shown in figure 3.3a. As PAni exceeded the concentration threshold, nanowires of PAni were aggregated and formed the “backbone” of the PVP phase (figure 3.3e).

At the same time, the PU phase pushed the PVP/PU phase boundary to be spherical. As a comparison, with low concentration, PAni behaved as a colloid and dispersed well in the PVP phase (figure 3.3b). In this case, the combination of van der Waals force and surface tension led to better control of phase separation. When the surface tension is a determining factor, “island” shaped phase separation occurred (figure 3.3c). Figure 3.3e shows a self-assembly of nanowire inside a phase separation morphology, which is represented in figure 3.3a.

Figure 3.3d shows the force balance on the phase boundary. Generally, the formation of polyaniline defines the morphology of phases. Colloidal PAni/PVP composites have much higher mobility than PAni nanorods. The PAni nanorods flowed with the deformation of the PVP phase and formed patterns. There was a force balance on the phase boundary, which involved the

interaction of nanoparticles, the friction between nanoparticles and polymers, and the surface tension of the phase boundary. As shown in figure 3.3d, it can be described as:

퐹⃑⃑⃑푣푤⃑⃑⃑ + 휇 + ∫ ∆⃑⃑⃑⃑푃⃑ 푑푆 = 0 Equation 3.1

where Fvw is the interaction between PAni particles, which is the van der Waals force in this case. Introduced by Langbein, while nanowires’ radii are much larger than the distance between them, the van der Waals force between two similar cylindrical surfaces can be described as[146]:

ℎ퐿푅 퐹 ∝ − Equation 3.2 푣푤 푑5/2

where h is the Planck constant, L is the length of the cylindrical particle, R is the radius of the cylindrical particle, and d is the distance between particles. The particle radius R ranges from

0.83 to 1.25 μm. The distance d was measured to be 350 nm in average from figure 3.2c and 3.3e.

This equation can be interpreted as the van der Waals force being inversely proportional to d2.5.

As the van der Waals force decreases with an increase of the particles’ distance, it is easy to understand that the interaction between the cylindrical particles is much lower than the interaction between colloidal particles.

The second term, μ, in the first equation is the obstructer term, which consists of static friction between PAni particles[147] and friction/viscosity between PAni particles and the PVP matrix[148]. The friction between PAni particle and PVP is measurable. However, the static friction between PAni particles is impossible to be obtained in this case. The friction between the

PAni particle and PVP can be considered as the viscosity of the PVP phase.

The third term in the first equation describes the effect of surface tension between PVP and PU. S is the surface area, ΔP is the pressure difference between the PVP phase and the PU phase, which can be described from the Young−Laplace equation:

2훾 ∆푃 = Equation 3.3 푅푐

where γ is surface tension and Rc is the minimum radius of the phase curvature. Overall, the unobtainable term, inter-particle friction μ, makes the equation unsolvable. However, when the PAni content is small and exists as colloids, the friction between PAni particles might be negligible. In this case, the surface tension difference between PVP and PU and viscosity of PVP dominate the phase behavior. As a result, the polymer blends behave like common immiscible polymer blends. If the values of the friction and interaction terms are obtained, the minimal radius of phase curvature could be calculated and a prediction of phase morphology becomes possible.

3.3.2 Conductivity and Stability

The conductivity of 17 samples was tested with plate electrodes. Tested points are illustrated in figure 3.4a. Based on previous synthesis practices, the samples in the brittle

(granular) region (figure 3.4b) are no different. As PVP is easy to absorb moisture to be viscous and PU is elastic, more data points are focused with less PAni content to identify the transition from viscous to elastic. In practice, we always need a material neither too sticky nor too bouncy.

The transition between viscous and elastic is important for the design of the material. On the other hand, absorption of moisture is more favorable for conductive applications, such as moisture sensor. Also, the doping of PAni is mostly using hydrophilic Lewis acids.

Figure 3.4b shows the mechanical property of the polymer blend abstractly. It can be considered that, with over 70% of PVP, the material is always absorbing a large amount of

moisture and becomes viscous. However, while the PU content is more than 70 wt.%, the PU phase becomes matrix and dominates the blends’ physical property. While the PAni content is more than 70 wt.%, the sample becomes discrete which is hard to bond together. Shown in figure

3.4b, the “plastic” region indicates the material has an appropriate viscosity, elasticity, and conductivity. The polymer blends in this region are viscous enough to be torn apart and rebound, they are elastic enough to hold their geometry, and they are conductive enough to be used as a conductive material.

Figure 3.4 (a) Sample points prepared and measured in experiments. (b) General physical property of the polymer blends. (c) Initial and (d) stable conductivity of polymer blends. The unit of scale bar on the right side is log ohm and (c, d) are using the same scale. (e) Stability of conductivity for various polymer blends.

Generally, an increase in the concentration of PU results in a reduction in conductivity.

However, increasing the concentration of PAni and PVP will increase the overall conductivity.

The conductivity profiles of polymer blends are shown in figure 3.4c−e. PVP and PAni can absorb moisture from the air, which can result in these materials being unstable[149]. The samples were tested immediately after being taken out from an oven and cooled. Results are shown in figure 3.4e. It could be observed that most samples exhibited good stability against the absorption of moisture. This might be due to the addition of PVP to act as a moisture reservoir

which could store water, which prevented the blend from deforming. The absorbed moisture might also be helpful for the ionization of acids.

3.3.3 Conductive Fabric

In order to demonstrate the use at the device level, we dipped a nonwoven fabric in the solution of 1-5-10 to form a conductive coating on the fabric. The coating is uniform and flexible with a low conductivity value (figure 3.4e and 3.5). A conductive origami crane was applied as conductive media (figure 3.5c) to light up 2 light-emitting diodes (LEDs). The applied voltage was only 3.0 V. The conductivity of nonwoven fabric was about 0.1 S/m. Besides, the prepared sample shows high flexibility and non-flaking wear. The 1-5-10 sample has a much lower conductivity compared to other blends, which means the conductive performance of other samples will be better. What’s more, PAni is widely used as a sensing material in acid and ammonia sensing. In this case, this flexible conductive fabric design is promising for chemical sensing and electromagnetic radiation applications[150] as well.

Figure 3.5 (a) Flexible 1-5-10 sample treated nonwoven fabric mat. (b) Origami crane made from the fabric in part a. (c) Test of the origami crane as flexible and conductive fabric. (d) SEM image of pristine nonwoven fabric mat. (e) SEM image of coated nonwoven fabric mat. The scale bars in parts d and e are 8 μm.

3.4 Conclusion

In summary, the morphology of a PAni/PVP/PU ternary polymer system was studied and the phase separation was observed. PAni was synthesized using an in situ polymerization method which results in colloidal PAni or PAni nanowires. PAni nanowires self-assembled into scattered fractal networks. After adding PU, a concentrated PAni/PVP phase occurred. Such a phenomenon was attributed to the balance between blocking force and van der Waals force (surface tension).

When the surface tension is the determining factor, “island” shaped phase separation was formed.

When surface tension and van der Waals force were both determining factors, bicontinuous phase

separation occurred. When the forces were in equilibrium, a fractal network structure formed.

Conductivity profiles and the morphology of these polymer blends were evaluated. The viscosity of polymer blends ranged from a viscous liquid to elastic solid, and the resistance varied from 102

Ω to 107 Ω. Most of the blend systems were stable in air, and the flexibility and stability of phase structure was helpful in maintaining a stable conductive network.

4. CHAPTER IV

CNT AND POLYANILINE BASED SENSORS FOR THE DETECTION OF ACID

PENETRATION IN POLYMER COMPOSITE

4.1 Background

With the invention, improvement and popularization of modern microscopy technologies, such as SEM and TEM, researchers start understand materials in nano scale. As a result, nano additives are widely studied in the last 2 decades. The most studied nano additives are nano-clay, nano-sheet, nanotubes, nano-whisker, nanofiber, nano-sphere, and nanofoam. Nano materials are generally having a high aspect ratio or large specific surface area. For example, carbon nanotube has a high aspect ratio and the specific surface area of graphene is very high. Due to the nano size nature of nanomaterials, additional functions may be found. Such as catalytic or quantum effect properties. As a result, nanomaterials are being used for various applications.

In the last decade, polymer matrix composites used in aircraft fuselages, marine body vehicles and chemical storages have been rapidly developed[151-154]. Most of these composites are subjected to a moisture environment and are designed to resist aggressive chemicals[155,

156]. This type of direct exposure generally causes swelling and hydrolysis of the matrix.

However, it also results in the bleaching of fiber reinforcement[157]. In order to avoid catastrophic failure within the material, it is critical to monitor the diffusion of chemicals into the composite using in-situ embedded sensors. Damage in composites can severely impact their performance and should be quickly identified in a real time to avoid structural failures[158].

Current methods to track moisture or chemical diffusion into composites are based on weight and volume change measurements[153, 159, 160]. This method might be helpful to understand better the diffusion mechanism. However, it is not practical for monitoring the composite during operation in real time. Some in- situ/real-time structural health monitoring (SHM) methods are based on damage and deformation detection of large composite parts using electrical impedance

tomography (EIT)[161, 162], acoustic emission (AE)[163, 164], and fiber optics based spectrum

(OPS) method[72, 165]. However, these methods are time consuming and involve complex analyses, which are costly. Electrical resistance detection methods have rarely been used for polymer composites because most polymers are dielectric. In the case of detecting chemical penetration or diffusion, the attacking foreign ions act as conductive media and lead to the change of the composite's electrical properties. As a result, direct electrical resistance measurements might be a very convenient method for monitoring the diffusion of species into polymer composites.

The performance of composites under hydrothermal conditions and a stressed-corrosive environment have been well studied[166]. Stress corrosion cracking (SCC) is one of the major topics in composite durability. The applied stress accelerates the initialization and crack propagation in aggressive environments such as acids[167, 168], bases[72, 152, 169] and saline mediums[169, 170]. SCC has been widely studied[157, 163, 171-174] because it leads to catastrophic results. Researchers point out that during the ‘corrosion’ of the composite, the diffusion of acid occurs first, followed by hydrolysis[175, 176], and then by the bleaching of fibrous reinforcement[157]. Due to the fact that hydrogen ions have a smaller mass, it could diffuse faster in polymer materials compared to water molecules. As a consequence, the detection of acid penetration becomes essential for the prediction and prevention of composite failure.

However, as mentioned above, most current experimental and theoretical studies about SCC and the effect of moisture are based on post-testing methods (weight or volume measurement). The diffusion of moisture in polymer matrix is usually considered as a Fickian diffusion process which could be described by Fick's second law[177]. However, with conductive nano-additives, the conductivity profile can be changed. Conductive nano-additives such as conductive carbon nanotubes (CNT), graphene, carbon black and conjugated polymers, are widely used to improve electrical properties of the polymers. Habis et al.[118] simulated the effect of aggregation of conductive nanofillers. They point out that the aggregation of conductive nanoparticles is

responsible for the formation of conductive paths and reduce the nano-additive percolation threshold. Recent experimental[178-181] and theoretical[182] studies on conductive nanocomposite and the related sensors[183-186] showed a great potential of conductive nano- additives on sensing performance. Due to the effect of degradation of the matrix and bleaching of the fiber, the general Fickian diffusion model cannot fully predict the behavior of moisture absorption. Non-Fickian models were used to describe the effect of fillers on matrix degradation[187, 188]. The research from Minelli et al. indicated that the diffusion process also relates to the geometry and orientation of the nanofillers[189]. They established a model for nano-clay or graphene-like materials reinforced with polymer nanocomposites. The result shows that oriented obstacles could alter the mass transport of attacking agents.

As mentioned previously, degradation of the matrix and bleaching of fiber reinforcement are two other main issues which impact composite performance. The crosslinking and degradation of polymer matrices could be considered as the two sides of a reversible reaction.

Acids and alkalines are good catalysts for the hydrolysis degradation of general thermosetting polymers[175, 176]. The bond dissociation of epoxy resin starts with ester or ether bonds which are formed by an epoxy group, followed with nitrogen functional groups[175]. Ion exchange is the major mechanism for the bleaching of glass fiber[157, 190]. As a result, the glass fiber may crack and loose its strength. Monitoring the ions' migration in the composite is interesting, but in practice, it is not relevant. In this chapter, conductive sensors capable of detecting changes in terms of electrical conductivity when subjected various environments were designed. CNT were used as conductive nanofillers to reduce the response time during sensing operations. Conjugated polymers are another option if you wish to monitor the changing progressively. Conversely, polyaniline (PAni) was used because of its conjugated structure and was used as benchmark material for sensing chemicals. The transformation of insulated PAni to conductive PAni consumes active hydrogen ions or other Lewis acid cations, which give the material its chemo- electrical sensing properties.

4.2 Materials and Methods

4.2.1 Materials and Sample Preparation

PAni were synthesized by blending method[191, 192] (figure 4.1a). PAni were synthesized using a mixture of aniline and ammonium persulfate in 1 M hydrochloric acid in ice bath and stirred overnight. An over doze of ammonia solution was added to transfer the emerald salt (PAni-ES) to polyaniline emerald base (PAni-EB). The PAni-EB was then filtrated and washed with distilled water, ethanol and acetone for several times. The product was dried in air, and dark violet/blue powder could be obtained. Saturated PAni-EB/DMF solution was prepared by adding enough PAni-EB into DMF and stirred for 24hrs. Solubility of PAni-EB was measured by drying supernatant of the centrifuged saturated PAni-EB solution and the result is 16.69mg/ml.

The saturated PAni-EB/DMF solution was then added to Epon 862. Concentration of PAni-EB in epoxy was controlled by adding different volume of PAni-EB/DMF solution to Epon 862. After stirring and drying the blend, Jeffamine D230 was then added to prepare the epoxy-PAni sample.

The percentages of PAni-EB in epoxy are about to be 0.5 wt.% and 1 wt.%.

Epoxy thermosetting resin was prepared from Epon 862 epoxy and Jeffamine D230 diamine, the recipe is 5:2 by weight. CNT (Pyrograf, PR25, figure 4.2a) was sonicated in DMF solvent overnight and then added to the Epon 862 to stir overnight. The percentage of CNT was calculated in weight percent. Then, Jeffamine D230 was added to the blend system and stirred until a high viscosity was achieved. The prepared thermoset resin and blends were then used to prepare a metal electrode embedded with natural fiber mat reinforced composite (figure 4.3a) using the vacuum bag curing method. The use of reinforcement is aiming at maintaining a steady distance between electrode and nanocomposite surface.

4.2.2 Testing and Characterization Methods

A Nicolet is50 FTIR (Fourier-Transform Infrared Spectroscope) was used to characterize the synthesized PAni-EB. X-ray diffractometer (XRD, Rigaku, using Cu K radiation, 40 kV, 44 mA) and Raman microscope (Renishaw, with 633 nm laser) were used to determine the

crystallinity of CNT. Scanning Electron Microscope (SEM, ProX, Phenom) was used to study the morphology of samples. Samples were gold sputter coated before observation. Pristine epoxy, epoxy-PAni, and epoxy-CNT specimens were tested with the setup showing in figure 4.3b for sensing measurement. The system consists of an acid cell, a copper electrode, a multi-meter based signal record setup and the specimen. The specimens were immerged in a 1 mol/L H3PO4 solution and resistance between the probes were recorded using a Multimeter (2700, Keithley). The recording interval was about 0.1s. H3PO4 was used in this experiment due to safety issue.

4.2.3 Modeling Method

A simple model for the resistant response based on the Fickian diffusion model was established.

Based on Fick’s second law:

휕푐 휕2푐 = 퐷 Equation 4.1 휕푥 휕푡2

‘c’ represents the concentration of the phosphoric acid solution, ‘t’ is the time, and ‘D’ is the diffusion coefficient of acid or moisture. For a simple one-dimensional diffusion problem, one of the solution of equation 3.1 could be written as[177]:

푥2 퐴 − 4퐷푡 푐 = 1 푒 Equation 4.2 푡 ⁄2

Here, ‘A’ is an arbitrary constant, and ‘x’ is the distance between the point to the surface of the composite. The resistance of system could also be expressed as:

퐶푒푙푙 휅 = = 푐 ∙ 훬 Equation 4.3 푅 푚

In this equation, ‘k’ is the reciprocal of specific resistance, ‘Cell’ is the cell constant. Cell constant could be described as the proportion of resistance and specific resistance of the

electrolyte. As an inherent parameter of a specific electrolyte system, we considered it to be a constant value during the test[193]. ‘R’ is the resistance of the system; ‘m’ is molar conductivity of the solution which considered to be a constant. The resistance was calculated from molar conductivity and an integral of gradual resistance equations (4.3-4.6). As phosphoric acid could hydrolyze out multiple hydrogen ions, the molar conductivity of the 1 mol/L phosphoric acid solution was considered to be constant during the test.

푥 푅 = ∫ 푅푥푑푥 Equation 4.4 0

퐶푒푙푙 푅푥 = Equation 4.5 푐 ∙ 훬푚

1 1 퐶푒푙푙 푡2 푥2 푅 = ∫ ∙ 푒4퐷푡 푑푥 Equation 4.6 0 훬푚 퐴

In these equations, ‘x’ is the thickness of epoxy; ‘D’ is a diffusion coefficient related constant. equation (6) is a variation of Dawson's integral and which is hard to obtain a general solution. With the help of MapleTM (2016.0, Maplesoft, a division of Waterloo Maple Inc.,

Waterloo, Ontario.), we could rewrite it as:

1 퐴′ 휋푡 ∙ 푒푟푓 0.5푥√− √ ( 퐷푡) 푅 = Equation 4.7 1 √− 퐷푡

where A’ is a unified constant to replace ‘Cell/(m A)’ and the ‘erf’ function is the Error

Function. Also, in order to have a brief illustration, we omitted the constant term. Obviously, this

is an imaginary equation which does not have a general solution. Here we use numerical method to obtain its numerical solution in a suitable range.

4.3 Result and Discussion

4.3.1 Materials and Morphology

PAni-EB were examined by FTIR in ATR mode. The result in Fig. 1b shows it has the typical PAni vibrations and stretches peaks. 1587 cm-1 peak is attributed to quinoid (Q) stretch;

1490 cm-1 peak is attribute to benzenoid ring stretch (B); 1142 cm-1 peak is attributed to B-N=Q structure[194]. The broad 1142 cm-1 peak suggests a mixture of B-NH+=Q, B-N=Q and similar structures. 1286 cm-1 peak is considered to be the C-N stretch in aromatic amine structure. Such a result could confirm we made a typical PAni-EB material. CNT were examined by Raman spectroscopy and powder XRD. Raman shifts and XRD peaks in figure 4.2b and c shows the crystallinity of CNT is high and it's a typical multiwall CNT material. The low ratio of D band

(1318cm-1) to G band (1580 cm-1) suggests it is multiwall CNT[195]. Also the sharp (002) peak and clear appearance of (100), (101), (004) and (110) peaks confirms it is multiwall CNT and its high crystallinity[38]. The high crystallinity is not only good for its conductivity property, it also endows the material high strength and modulus.

Figure 4.1 (a) Synthesis route for PAni-ES and PAni-EB. (b) FTIR result for PAni-EB.

Figure 4.2 (a) SEM image of the CNT. (b) Raman spectroscopy profile for CNT. (c) Powder XRD result for CNT.

Photo images of prepared and tested composite samples are shown in figure 4.4. The pristine epoxy post-test sample shows typical swelling wrinkles. After acid immersion, the sample had swollen and the color had changed. The PAni samples initially exhibited a blue-violet color and once in contact with the acid solution the color had turned into green (conductive material). However, in the case of CNT based nanocomposites, no obvious color change was observed and no swelling was identified.

Figure 4.3 (a) Assembly of nanocomposite sensor. (b) Test setting up.

Figure 4.4 Photo images of Pristine epoxy, epoxy-PAni, epoxy-CNT samples before and after acid immersion.

SEM analysis on samples were showed in figure 4.5. Figure. 4.5a and b shows the cross section of epoxy sample before and after acid immersion. The surface had changed from smooth to porous. Figure 4.5c and d shows epoxy-PAni sample morphology. As it is known, PAni has a poor miscibility with common organic chemicals, thus we consider the PAni-EB is immiscible with epoxy. As we have dissolved PAni into DMF and then mixed with epoxy, the PAni in epoxy matrix might form micro- or nano-phase separation. PAni-ES is considered to have a lower compatibility with epoxy. The nanosized bumps on the sample cross section surface might be the PAni-ES domains. Figure 4.5e and f shows CNT added epoxy samples before and after acid immersion. Before acid immersion of the CNT- epoxy sample, a large amount of CNT tips could be observed. However, after acid immersion, more CNT was exposed. This phenomenon may due to the change of elasticity of the epoxy.

Figure 4.5 SEM images of samples' cross sections before and after acid immersion respectively. (a,b): Pristine epoxy samples. (c,d): Epoxy-PAni samples and (e,f): Epoxy-CNT samples.

4.3.2 Sensing Data

The resistance test results are shown in figure 4.6 and 4.7. We defined the response time in which the conductive value does not change dramatically anymore. It's very clear that there is a difference in response time between the pristine epoxy, epoxy-PAni and epoxy-CNT samples.

The pristine epoxy samples show slow response time and multiple response stages (figure 4.6a).

The epoxy-PAni samples also show two stages. As an example, for epoxy-PAni 0.5 wt.% sample, the first stage ends at about 100s, and the second stage ends at about 1000s. The first is supposed to be the response to hydrogen ions and the second stage is supposed to be the response to water.

In figure 4.7c, we can identify the stages clearly. However, for the CNT based samples, only one response could be observed (figure 4.6b). For the epoxy-PAni samples, as we added PAni-EB, the response time was reduced. However, in the CNT based sample, the addition of additives did not change response time significantly but it changed the final conductivity of the material.

Figure 4.6 Resistance values of the sensors as a function of time. (a) Overall curves. (b) A zoom- in figure for 0-200s.

Figure 4.7 (a) Response time of different sensors. (b) Final resistance of different samples. (c) Repeatability of different sensors.

Figure 4.7 shows the response time (figure 4.7a), final resistance (figure 4.7b) data of all samples and repeatability (figure 4.7c) of the tests. Three curves from each sample were listed in figure 4.7c. The difference in response time and final resistance could be explained by their conductivity properties. PAni is always considered as an immiscible material in most solvents and polymers. Even when it was mixed in DMF, the PAni-EB in the epoxy-PAni sample tends to form a PAni “rich domain” and an epoxy “rich domain”. The diffusion of hydrogen ions through the free volume of polymers is faster than water. Especially since the PAni-EB has a rigid backbone and tends to form larger free volume in the blend. The PAni-EB has another effect on the diffusion of hydrogen ion. It could react with hydrogen ions and result into the conductive

PAni-ES. However, in order to transform into the conductive PAni-ES, PAni-EB consumes hydrogen ions. This will cause a delay on the response time compared to the epoxy-CNT sample.

The formation of the conductive path in CNT added sample attributes to the bridging effect of hydrogen ions. Theoretically, due to the tunneling effect, in order to form a conductive path, two

CNTs have to touch or should be very close to each other (around several nanometers' distance or less)[17, 196], which requires a high concentration of CNT. In most cases, the CNTs were aggregated. The penetration and diffusion of hydrogen ions were responsible for connecting the

CNT aggregates to form the continuous conductive network. Meanwhile, due to the hydrophobic property of CNT, a gap between epoxy and CNT might form. This gap would be a pathway for hydrogen ion and water. As a result, the response time for epoxy-CNT sample is very short. In figure 4.7c, we can notice that the repeatability for pristine epoxy sample is much worse than that of others and the change of resistance goes through a fairly long duration. This phenomenon could be an additional indirect evidence that PAni and CNT have important effect on the diffusion of ions and water.

4.3.3 Modeling Result

The effects of A’ and ‘D’ were investigated in the modeling, where A’ could represent the effect of the conductance of the ions, and ‘D’ is the diffusion coefficient of the ions with a unit of

cm2/s. The thickness of epoxy, ‘x’, was assumed to be 1 mm according to our experiment.

Showing in figure. 4.8, the conductance of ions (A’) mainly tunes final resistance of the system.

The diffusion coefficient was set to be 10-9 cm2/s. The higher conductance of the ions, the lower final resistance. Also, by increasing the A’ in magnitude from 1 to 104, the resistance reduces from 103 to 10-2, which is relatively linear. Conductivity values for common chemicals are in the range of 1-100 mS·m2/mol, which means the values are in a small range. As the A’ is a unified term, and the cell constant of the system could not be obtained, we can assume the value of A’ in the experiment is in a small range. As a consequence, the final resistance should fall in a small range too. This could explain why the final resistance showing in figure 4.7b falls in a small range. Based on the result in figure 4.7b, we can estimate the value of A’ correspond to our experiment is about to be 0.01.

Figure 4.8 Modeling result: the effect of conductance (A’) of the attacking chemicals on the electrical response.

Figure 4.9 Modeling result: the effect of diffusion coefficient (D) of attacking chemicals on the electrical response.

Figure 4.9 shows the effect of the diffusion coefficient where A’ was set to be. As the diffusion coefficient increases, the response time increases. The intervals of D-1 (s/cm2) for the curves in figure 4.9 are 2×108 s/cm2 in the range of 108~109 s/cm2 and 2×109 s/cm2 in the range of 109~4×109 s/cm2. Figure 4.10 shows the relationship between diffusion coefficient ‘D’ and responding time. In order to obtain the responding time, 135˚ tangent lines were applied to the curves and the abscissas of the tangent points was used as the responding time. With the increase of the diffusion coefficient, the responding time decreases exponentially. Compare to figure 4.8, which the response time did not change significantly through the change of A’, the diffusion coefficient for epoxy-PAni sample could be estimated to be 2×10-9 cm2/s.

Figure 4.10 Modeling result: relationship between diffusion coefficient and sensing response time.

This modeling is based on the assumption that the material system is homogenous.

However, for a polymeric system, the mechanism of diffusion is complex. Especially for a system contains reactive chemical (PAni) and/or nanoparticle (CNT), the diffusion behavior is far more complex. The relationship between electrical resistance and the different parameters which were investigated in this model is enough to have a better understanding of the sensor behavior.

4.4 Conclusion

In this chapter, we have carried out two nanocomposite formulations: one is based on polyaniline and the second uses CNT as additives. Numerous samples with different concentrations of additives were fabricated. The samples were connected to a Multimeter and then were immerged into an acid solution. The resistance of the samples was measured as a function of time. It appears as the amount of nano-additives increased, the conductivity increased

and the response time towards acid penetration was shorter. However, the CNT based nanocomposite showed faster response time and the result was very consistent. In order to understand what was occurred during these tests, a kinetic study was carried out for the responding behavior. The sensing mechanism was depicted using a Fickian model and the experimental and theoretical data were in agreement. Indeed, the penetration and diffusion of hydrogen ions were responsible in connecting the CNT aggregates by forming a continuous conductive network.

Mechanical properties of composite infrastructure are sensitive to the defects. Our designed sensors are either based on epoxy-CNT nanocomposite or polymer blends. This design will endow a great compatibility between sensor and the structure compare to other embedded sensor. However, in this design, an online signal receiving system is still required. For future study, wireless sensor system and mobile easy-scan system is highly preferred. Potential applications for our design will not be limited to durability of civil infrastructure[197]. As the resin system could be easily alternated, the nanocomposite could be designed for applications such as in vivo biochemistry monitoring[198], EMI shielding[199], Environmental remediation[200, 201], structural application[200, 202], energy harvest[203] and wearable sensing[204].

5. CHAPTER V

FABRICATION OF HIGH PERFORMANCE NANOCOMPOSITE BASED CHEMICAL

SENSOR USING LOW CONCENTRATION ADDITIVES

5.1 Background

Considering carbon nanomaterials for sensing the chemical penetration of composite structure, the sensing mechanism was proposed in the last chapter that the diffusion of ion increase the conductivity of nanocomposite. However, the diffusion mechanism was not clear since the diffusion along nanoparticle is more complex than the diffusion phenomenon in polymer. In this case, a comprehensive model should be considered in order to illustrate the effect of diffusion on the conductivity change of nanocomposite. In this chapter, the sensing behavior based on low concentration of CNT and carbon black was explored experimentally and theoretically.

Composites are used as structural materials. Sometimes they are exposed to aggressive chemicals such as fuel, acid, alkaline solution, saline and sewage[155, 156, 205]. Although these composite structures are designed to last for decades, it is important to monitor their health[162,

165, 206]. Swelling and hydrolysis of the matrix are two general results of this kind of exposure[157, 207, 208]. Swelling leads to deformation of structure and hydrolysis leads to direct failure. However, the corrosive chemicals also result a bleaching of fiber glass reinforcement[157, 209]. In order to avoid these degradations, an in situ embedded sensor is critical to monitor the diffusion of aggressive chemicals into composite[165].

Micro-damage in composites can propagate and develop into macroscopic cracks quickly[210, 211]. In order to avoid catastrophic failures, the identification of micro-damage should be detected in a real time[212, 213]. The micro-damage is generally caused by stress concentration and materials’ degradation[214]. Structural health monitoring (SHM) used for crack and deformation detection, are based on electrical, acoustic and optical principles[215,

216]. For example, electrical impedance tomography[206] and fiber Bragg grating sensors[217,

218]. Currently, studies of the effect of aggressive chemicals on composites are based on weight and volume measurement. In conjunction with Fickian diffusion modeling, this approach is helpful for understanding the diffusion mechanism[209]. However, these methods are not effective in terms of time, convenience and cost. The penetration of aggressive chemicals always changes the dielectric property of the polymer matrix[71, 219]. In this case, measuring the conductivity of polymer matrix might be a promising method for monitoring the chemical penetration process.

Chemical attacking of composites by aggressive chemicals result in severe damage for both polymer matrix and fibrous reinforcement[209, 220]. Especially under hydrothermal and stressed conditions. The Stress corrosion cracking (SCC) has been the focus for decades in understanding the interaction between chemicals and composites[157, 171]. Using acid as an example, the penetration of acid occurs initially during diffusion process. After the penetration of acid and water, hydrolysis of polymer matrix occurs. Consequently, the mechanical properties of the materials are impacted. It is important to monitor the penetration of acid species in composite.

Current SCC studies are relying on post-testing methods, for example, weight measurement and volume measurement. These studies have shown that Fickian diffusion theory can be applied in

SCC and water diffusion problems[157, 187]. However, in the case of nanocomposite, the

Fickian diffusion theory is not fully explored. Together with other remained problems, such as hydrolysis of matrix and bleaching of fibers, non-Fickian diffusion models were proposed[188].

Minelli et al. have suggested that the geometry and orientation of nano-additives would affect the diffusion process[189]. Usually, the use of nanomaterials would increase the diffusion speed of attacking chemicals. However, many studies have shown that two dimensional materials, such as nano-clay, graphene and boron nitride, are excellent in slowing downing the diffusion process[221, 222].

Our previous study shows the dispersion of conductive nano-additives affects the physical property of Nanocomposite[119]. Especially, the formation of conductive network in nanocomposite significantly affects its conductivity behavior[118]. In this case, we have designed a sensor using low concentration nano-additives for detecting the chemical penetration in composite structure[71, 219]. A standard and traditional conductive sensor relies on the formation of conductive network[33]. However, the conductive network cannot be established with a few conductive nano- additives and the nanocomposite is basically insulating. However, with the chemical penetration process, the attacking ions become the bridging units for the network as shown in figure 5.1. As a result, the conductivity of nanocomposite increases upon the penetration and which could be used as an indicator for monitoring of penetration. Additionally, the diffusion speeds of ions are different. As a result, the complexity of the sensing behavior is expected.

Figure 5.1 Sensing mechanism of CNT based nanocomposite sensor.

On the other hand, carbon black is the mostly used carbon nanomaterials to construct a conductive nanocomposite. However, the formation of carbon black based conductive network is not efficient compare to carbon nanotube (CNT). As a high aspect ratio material, CNT could form the conductive network in a very efficient way[22, 118]. In this chapter, we prepared CNT and carbon black based nanocomposite sensors and investigated the effect of nano-additives’

geometry on sensing behaviors. Also the complex sensing behavior was explained by the modeling result.

5.2 Materials and Methods

5.2.1 Materials and Sample Preparation

The CNT material (PR-25) were purchased from Pyrograf Products, Inc. The diameter of this CNT is about 150 nm and the length is up to 50 m[38]. The diameter of carbon black

(Vulcan XC 72R) is about 300 nm. The epoxy resin system consists of 100 parts of Epon 862

(epoxy) and 35 parts of Jeffamine D230 (hardener). The nanocomposite sensors were prepared with the following procedures: (1) mix a calculated amount of CNT with 20 g of epoxy and 2 g of acetone for 12 hours; (2) add 7 g Jeffamine D230 into the mixture and mix for an additional 30 min; (3) cast the nanocomposite resin onto prepared non-woven fabric layer by layer as shown in figure 5.2a; (4) seal and cure the samples in a vacuum bag for 2 days. The use of non-woven fabric layer is aiming at keeping the distance between electrode and the surface of nanocomposite steady. In order to ensure the thickness of samples are about the same, all procedures are carefully controlled. Due to the difficulties in processing, the sensor with 5wt% of CNT is not successfully prepared. The samples are noted by their concentration and additive name. For example, the nanocomposite with 0.5wt% of carbon black is noted as 0.5CB, the nanocomposite with 1wt% of

CNT is noted as 1CNT.

Figure 5.2 (a) Nanocomposite sensor assembly. (b) Testing setup.

5.2.2 Testing and Characterization Methods

Scanning Electron Microscope (SEM, ProX, Phenom) was used to study the morphology of raw material and prepared sensors. A Zeiss Vert. A1 optical microscope was used to characterize the polished cross section of made sensor. X-ray diffractometer (XRD, Rigaku, Cu

K, 40kV, 44mA, =1.5418 Å) was used to examine the crystallinity of CNT and carbon black.

The sensing behavior testing setup is shown in figure 5.2b. The sensing system consists of an acid tank, a copper counter electrode, a multi-meter based signal recording system and the sensor. The testing acid solution was using 1mol/L H3PO4 for safety concerns. The resistance between counter electrode and the sensor was recorded using the Multimeter (model 2700, Keithley). The recording intervals are chosen to be 0.1 s, 1 s and 5 s for different purpose. 0.1 s and 1 s intervals

were used for determine the response time. The 5 s interval was used for recording the long period sensing behavior.

5.2.3 Modeling Method

The modeling method is directly based on Fick’s second law (equation 5.1):

휕푐 휕2푐 = 퐷 Equation 5.1 휕푋 푒 휕푡2

Where the c is the concentration of ion, t is the time, X is the distance between the calculation point to the surface of the nanocomposite sensor, De is the effective diffusion coefficient of ion or water in nanocomposite. As a result, with a known effective diffusion coefficient, the relationship between the concentration c and time t could be established in a one dimensional space as equation 5.2:

푐 = 푓(푡, 푋) Equation 5.2

The Fricke-Hamilton-Crosser (FHC) model is used to integrate diffusion coefficients of different phases into a single value, the effective diffusion coefficient[177]. It assumes the dispersion of nano-additives is homogenous. The effective diffusion coefficient (De) for a nanocomposite system, which consists of part a and part b, can be expressed as:

푥푣푎(휑 − 1) + 휑 + 푥 퐷푒 = 퐷푏 Equation 5.3 푣푎(1 − 휑) + 휑 + 푥

where v is volume fraction of each content, Db is the diffusion coefficient of b phase,  equals to Da/Db and x is a function of Da/Db. If the additive is spherical, then the x=2. Here, the x can be expressed with an empirical relationship[177]:

3 2 − 1 (푝푟표푙푎푡푒 푒푙푙𝑖푝푠표𝑖푑) 휓 3 푥 = − 1 (표푏푙푎푡푒 푒푙푙𝑖푝푠표𝑖푑) Equation 5.4 휓1.5 3 − 1 (푠푝ℎ푒푟푒) { 휓1

The , sphericity of a particle. is defined as the ratio of the surface area of a volume- equivalent sphere to that of the particle[177]. As a result, for CNT, the x can be expressed as:

3 1 (6푙) ⁄2 푟 ⁄3 푥 = ∙ Equation 5.5 2 푟 + 푙

Where the l is length of CNT and r is the radius of CNT.

The concentration of ions determines the conductivity of nanocomposite (). Here, we calculate the conductivity of each thin layer (dx) since we treat the concentration of ion in this layer (dx) as a constant. As a result, the relationship between concentration of ion and conductivity of the material could be expressed as:

푥 푥 휎 = ∫ 푓(푐)푑푥 = ∑ 푓(푐푖) Equation 5.6 0 푖=0

In this model, we assume the effect of different ions are independent. For example, H+, phosphoric ions and water molecules has independent effects. The overall conductivity is the sum of each:

휎푡표푡푎푙 = ∑ 휎 Equation 5.7

With the relationship between conductivity and resistance of material R=L/A, where L is the length of material and A is cross section area. By combining the listed equations, the overall resistance of the nanocomposite could be expressed as:

1 1 푅 = = 푡표푡푎푙 1 1 1 1 Equation 5.8 + + ⋯ + ∑ 푅1 푅2 푅푛 푓(푡, 푋)

The n represents the number of ions/molecules considered in the simulation. In this model, we took three ions/molecules into consideration.

5.3 Result and Discussion

5.3.1 Material and Morphology

The CNT and carbon black raw materials were examined using SEM and the morphology of these materials are shown in figure 5.3a and 5.3b. The raw CNT material are entangled due to its high aspect ratio. As a result, the sonication process is necessary for this CNT to be well dispersed in polymer. Generally, the carbon black material which used in this chapter has a diameter of 300 nm with a spherical geometry. Figure 5.3c shows the XRD result of these materials. The sharp (002) peak of CNT can be easily identified at 26.0˚. The layer inter-spacing d(002) is 3.43 Å for CNT and 3.58 Å for carbon black. These values are found similar to other studies[223, 224]. Considering the electronic conductivity of carbon nanomaterials is highly dependent on intra-layer transportation, CNT has a higher conductivity than carbon black material. Both CNT and carbon black have high crystallinity which ensures their high conductivity. Figure 3d shows an enlarged optical image of figure 5.2a. Due to its small size, the

CNT cannot be seen clearly using optical microscope. However, figure 5.3d clearly shows the

CNT is well dispersed in the epoxy matrix.

Figure 5.4 shows SEM images of sensor cross sections. Both pre-testing and post-testing images for CNT and carbon black are presented. CNT threads and carbon black bumps on the cross section surface can be easily identified in the pre-testing images. However, the carbon black cannot be identified in the post-testing image and the CNT threads are pulled out from the cross section surface. Both post-testing images show the cross section becomes rougher, which

suggests the epoxy becomes more plastic than it used to be. This could be explained by the degradation of epoxy upon the penetration of acid.

Figure 5.3 (a) SEM image of the CNT. (b) SEM image of the carbon black. (c) XRD results for CNT and carbon black. (d) Optical image of cross section of the nanocomposite sensor showing the dispersion of CNT is good.

Figure 5.4 Typical SEM cross section images of (a, c) pre-treated and (b, c) post-treated (a, b) CNT and (c, d) carbon black sensors.

5.3.2 Sensing Behavior

The sensing behavior of various carbon black and CNT samples were tested and recorded. Figure 5.5 shows the sensing behavior of carbon black samples with different nano- additive concentration. The figure on the left uses logarithmic scale for Y axis and normal scale for X axis. However, the figure on the right uses logarithmic scale for both axes. This figure setting is also used for the results of CNT based sensors.

Figure 5.5 Recorded resistance signals of the carbon black nanocomposite sensors as a function of time. (a) Using normal scale for x-axis. (b) Using logarithm scale for x-axis.

Obviously, the sensing behavior of carbon black base sensor differs depend on the nano- additive concentration. The resistivity of samples drops progressively through the testing period.

Especially, the 0.1 wt.% carbon black samples (0.1CB) show multiple stages in the resistance- time curve. The first drop, represents a typical ion diffusion induce resistance change. Our previous model matches this sensing behaviour perfectly. However, the previous model could not explain the multiple stages’ behavior very well. It was different from 0.5CB sample, with a higher concentration of CB, the sensing response became much faster. In this case, we had to use logarithm X axial to illustrate the sensing behavior of 1CB, 2CB, and 5CB samples. In figure

5.5b, the sensing curves of 1CB was well distinguished from 0.5CB curve. 0.5CB samples had a response time ranges from tens to hundreds seconds. However, the response time is shorter for

1CB. On the other hand, the 2nd and 3rd stages in the sensing behavior of 0.1CB samples are more clear than other sample. This behavior indicates the addition of nano-additive affects the appearance of stages. Carbon black could be well dispersed into epoxy resin while CNT may be affected by entanglement and aggregation.

Figure 5.6 shows the sensing behavior of CNT based samples. The addition of CNT changes resistance-time curve dramatically compared to 0.5CB sample. Almost all samples show an instant response. The multiple stages still could be found in some samples. Especially, for the

0.1CNT sample (the green curve), a short and smooth dropping curve could be found. This indicates that the conductive network is just formed or nearly formed in 0.1CNT sample.

Additionally, based on our previous model which simulates the formation of conductive network in nanocomposite, the percolation threshold for CNT nanocomposite was around 0.1 wt.% to 0.5 wt.%. At 0.1 wt.%, the probability for the formation of conductive network was 40%. At 0.5 wt.%, the probability for the formation of conductive network was 90%. In this case, it is reasonable that upon attacking of ionic chemicals, the resistance drops quickly as shown in figure

5.6. As a comparison, the sensing behavior of 1CNT and 2CNT were given as well. The sensing response was faster than 0.1CNT one.

Figure 5.6 Recorded resistance signals of the CNT nanocomposite sensors as a function of time. (a) Using normal scale for x-axis. (b) Using logarithm scale for x-axis.

5.3.3 Modeling Result

An analytical solution was obtained from the previous modeling work (Chapter IV).

However, the analytical approach leaded to a less flexible model. In this model, we solved the

Fick’s second law numerically from the very beginning. Illustrated in figure 5.7, the resistance curve becomes more flat with a lower diffusion coefficient value. Based on some similar assumptions, reasonable results were obtained. For example, in order to avoid a time consuming model, we assumed the diffusion coefficient was constant. Although the ions may only diffuse

through the nano-additive surface rather than the volume, a simple diffusion value was given for each kind of material. With the diffusion of acid species, the epoxy resin will degrade overtime.

In microscopic point of view, the cross linked polymer chains are broken slowly which leads a higher free volume of polymer. The cross linked polymer chains could also be considered as nano sieves. With broken polymer chains, the sieve holes become larger. As a result, the diffusion speed of larger ions is dependent on polymer degradation. At different penetration depth of the polymer, the diffusion speed is different. However, currently, there is no such a model could take the degradation of polymer into consideration due to the unknown degradation degree and how does it affect the diffusion. In this case, we consider the diffusion of three different ion species is independent to each other.

Based on the FHC model, we investigated the effect of nano-additive concentration, nano-additive aspect ratio, diffusion coefficient ratio between different phases () (figure 5.8).

Normally, with a higher nano-additive concentration the effective diffusion coefficient becomes higher. However, when the diffusion coefficient ratio  is less than 1, the addition of nano- additive leads to a smaller effective diffusion coefficient. In the other words, when the additive becomes an obstacle for the diffusion, the more additive results a slower diffusion behavior. In practice, graphene and hexagonal boron nitride nano-sheets are studied as effective additive for reducing moisture permeability of polymer. The aspect ratio of nano-additive is another controlling parameter. In this model, the aspect ratio ranges from 200 to 40000 with an increment of 1000. The increment is shown as the arrows pointed in figure 8. The effect of increment of aspect ratio is more significant when the aspect ratio has a smaller value. This result indicates the change from carbon black to CNT is more significant than increase the length of CNT for their effect on effective diffusion coefficient. However, based on the previous study, a longer CNT could increase the probability to form the conductive network. As a result, the sensing probability could be promoted by the longer CNT.

Figure 5.8 The effect of nano-additive aspect ratio and diffusion coefficient ratio. The arrow shows the increment of additive aspect ratio from 200 to 40000 with an interval of 2000.

The effect of different parameters shown in figure 5.8 is only affecting the effective diffusion coefficient. The diffusion model is further controlled by the Fickian diffusion model. As we assumed the three species have independent diffusion effects, with equation 5.8, the result of overall sensing behavior is illustrated in figure 5.9b. Figure 5.9a shows three typical sensing behavior curves for pristine epoxy resin based sensor, CNT and carbon black based nanocomposite sensors. Obviously, this model could present the multiple stages’ curve. However, the curve shape for the 2nd and 3rd transition region is far more complex than the ideal Fickian diffusion behavior. As a result, the diffusion dependency of different ions and molecules is doubted. A far more sophisticated experiment should be done to investigate the ion/molecule diffusion dependency question.

Figure 5.9 Comparison of (a) experimental result and (b) Simulation result.

5.4 Conclusion

In this chapter, low concentration additives were used for the fabrication of nanocomposite for sensing of chemical penetration. Carbon black and CNT were used as conductive additives to form the conductive network in the nanocomposite. At low concentration, the nanocomposite was not well conductive. However, with the penetration of ions, the resistance of nanocomposite drops and this change could be easily detected. The detected sensing behavior of nanocomposite was also confirmed by our modeling work. The modeling work shows there are three major parameters that affect the sensing behavior: (1) The aspect ratio of the conductive nano-additive. (2) The diffusion speed of chemicals in different region. (3) The different diffusion speed of the ions and molecules.

This chapter is focused on carbon black and CNT. However, this model also explains why the graphene and boron nitride nanosheet materials could be a good candidate for moisture barrier nanocomposite.

6. CHAPTER VI

CARBON NANO MATERIAL BASED CHEMICAL SENSING: WIRELESS DESIGN

6.1 Background

Various sensing methods are used for structural health monitoring (SHM) of composite structures. For example, acoustic emission method (AE), electrical impedance tomography (EIT), fiber optics based spectrum methods. However, most of these methods are developed for damage or crack detection. Currently, only one method is used for detecting the chemical penetration in composite or polymer materials. Kubouchi et al modified fiber optics by replacing a short section of optical fiber with pH sensitive resin. With the attacking of acid or base, the obtained IR and

UV-vis spectrum changes. However, this method involves cost-ineffective optical system and the interpretation of spectrum requires skilled workers. As a result, a new cost effective and easy understood method should be developed.

Previous chapters introduced an electrical method for the detection of chemical penetration, which only involves resistance measurement. In the porotype design, metallic conductive wire is required to connect the sensor with resistance measuring instrument. It’s known to all that the fatigue of metal limits its use. On the surface of composite structure, the metal wire is easily bended and could fatigue quickly in its service. In this case, a wireless sensor is required.

Wireless sensor can adopt many technologies, for example, infrared, radio, micro-wave, blue tooth, tetra Hz and ultra-sonic. Power source of the sensor is one of the most important limitation. Common power source choices are solar cell and embedded battery. However, composite structures operate for decades. A power source less sensor is needed. As a result, passive power based sensor is a good choice. Overall, passive radio frequency identification

(RFID) sensor is chosen. The power of passive RFID comes from the electro-magnetic induction

(figure 6.1).

Antennae

RFID tag Reader

computer

Figure 6.1 Working scheme of RFID system.

The passive RFID only consists of a microchip and an antenna. Such a design has been applied in many industries such as logistical tracking, personal identification, and electronic transaction. Additionally, the RFID based near field communication (NFC) technology is available in most of smart phones. Therefore, monitoring RFID signals with a mobile device is convenient. The size of RFID varies and can be easily designed with special geometry.

6.2 Experiment Setup

The RFID system is based on Arduino-RFID kit, which is consist of an Arduino UNO R3 board, a RC522 board, USB cable, a LCD screen, connection wires and 13.54 MHz RFID tags.

The connection is illustrated as below (figure 6.2 and Appendix A). The corresponding code can be found in Mendeley datasets [225] and Appendix B. The commercial radio frequency identification tags were modified with knife and the CNT/DMF suspension as figure 6.3. The

sensors were tested with RC522 board from Arduino RFID kit. The RFID were tested with volatile organic acid (formic acid) to verify whether the sensor works or not.

RC522 board reader connection to computer Testing card Arduino board

LCD

Figure 6.2 The assembled RFID system.

Figure 6.3 Modification the RFID tag with CNT.

6.3 Sensing Result

The sensing result are shown in figure 6.4, by adding volatile organic acid, the sensor turned from off to on and which make the RFID to be a sensor. As the amount of volatile acid

was not precisely controlled, the sensing signals differs. Since a simple RFID reader was used, only 3 levels of signal can be apparently identified in figure 6.4. As shown in figure 6.4, when sensing target (acid) is not added, the CNT modified RFID tag has an impedance which does not match with the chip and result a high return loss and low signal strength (stage 1). When a fraction acid is added, the impedance of the sensor change which result a signal strength in stage

2. However, when the chemical starts to vaporize, impedance of the sensing section start to change and match the impedance of RFID system, which leads to an increasing of signal to stage

3. As the acid vaporizes more, the match on impedance break and cause the signal strength to decrease.

Figure 6.4 Sensing behavior of the CNT modified RFID tag.

6.4 Conclusion

A wireless sensor based on NFC and RFID technology was developed. The sensing system was a proof as a valid solution for sensing standard chemicals. This approach shows a great potential in chemical sensing area. Additionally, the RFID can be designed as failure sensor which means the chemical/physical damage causes the active/inactive statue change and be recognized as the signal. As a result, the RFID sensor could be designed not only for chemical

sensing, but also useful for structural sensing, environmental hazard monitoring, and explosives detection.

7. CHAPTER VII

CONCLUSION AND PERSPECTIVE

In this study, various nano-additives were developed and used to create novel nanocomposites for sensing applications. Molecular, polymeric and carbon nano-additive were used in three main experimental designs:

(1) The A240 petroleum pitch was used to enhance the conductivity and modulus of carbonized pitch/PAN nanofiber by improving its crystallinity. Based on DSC results, the addition of pitch lowered the overall Tg of PAN. The sample blends exhibited two Tg values. At this level of study, we can only assume that pitch and PAN are partially miscible. From infrared spectroscopy, the interaction of the cyano band (2250cm-1) and the new band at 1660 cm-1 was attributed to the (cyano)- (pitch conjugate) interaction and reactions, and the reaction relate to transesterification. This may have contributed to the miscibility of PAN and pitch. A rheology study was helpful in giving a better understanding of electrospinning process. In this study, homogenous carbon nanofibers were produced using all of the various mixtures between PAN and pitch. The final product, the carbon nanofibers produced from blended pitch/PAN precursor, showed an enhancement in tensile modulus electrical conductivity. The addition of pitch increased the degree of alignment because of the high amount of liquid crystal present in the pitch and eventually it enhances the physical properties of the carbon nanofibers. It was assumed that the extraordinary improvement in conductivity is induced by the anisotropic distribution of pitch through fiber axial since the pitch rich region generally results a higher conductivity than PAN rich region.

(2) Next, the morphology of a PAni/PVP/PU ternary polymer system was studied. PAni was synthesized using an in situ polymerization method which results in colloidal PAni or PAni nanowires. PAni nanowires self-assembled into scattered fractal networks. After adding PU to this blend, a concentrated PAni/PVP phase occurred. Such a phenomenon was attributed to the

balance between blocking force and van der Waals force (surface tension). When the surface tension is the determining factor, the “sea-island” shaped phase separation occurs. The surface tension and van der Waals force were two determining factors in the formation of bi-continuous phase separation. When the forces were in equilibrium, a fractal network structure formed which could significantly affect the electrical conduction behavior. The viscosity of polymer blends ranged from a viscous liquid to elastic solid, and the resistance varied from 102 Ω to 107 Ω. Most of the blend systems were stable in air. The flexibility and stability of phase structure was helpful in maintaining a stable conductive network.

(3) In the third study, we developed two nanocomposite formulations: one is based on polyaniline and the second uses CNT as additives. Both of these were embedded in epoxy to form a conductive network that would be sensitive to the transport of ions through the bulk. The samples were connected to a multimeter and then immerged into an acid solution. The resistance of the samples was measured as a function of time. It appears as the amount of nano-additives increased, the conductivity increased and the response time towards acid penetration was shorter.

However, the CNT based nanocomposite showed faster response time and the result was very consistent. In order to understand what was occurred during these tests, a kinetic study was carried out to model the system response and gain insight. The sensing mechanism was depicted using a Fickian model and the experimental and theoretical data were in agreement. Indeed, the penetration and diffusion of hydrogen ions were responsible for connecting the CNT aggregates by forming a continuous conductive network.

The long term mechanical durability of polymer matrix composites used in infrastructure are highly sensitive to defects such as foreign debris, voids, or embedded non-uniformities. Our designed sensors are either based on epoxy-CNT nanocomposite or polymer blends. This design will optimize the compatibility between the sensor and the structure compared to other embedded metal or ceramic sensors. However, ideally, in this design, a wireless signal receiving system is still required in order to eliminate protruding wires. Therefore, for future study, a wireless sensor

system and mobile easy-scan system is recommended as a key area of development. Potential applications for our design will not be limited to durable civil infrastructure projects. As the resin system could be easily tailored to other organic systems, the nanocomposite could be designed for applications such as in vivo biochemistry monitoring., EMI shielding in electronic systems, environmental remediation, structural health monitoring applications, and wearable sensing.

Future work at the device level would be crucial to implement the wireless sensing concept.

REFERENCES

[1] Chung Deborah D. L. Composite Materials: Science and Applications: Springer-Verlag

London; 2010.

[2] Lau Alan Kin-tak, Hussain Farzana, Lafdi Khalid. Nano- and Biocomposites: CRC Press;

2017.

[3] Ajayan Pulickel M., Schadler Linda S., Braun Paul V. Nanocomposite Science and

Technology: Wiley‐VCH Verlag GmbH & Co. KGaA; 2004.

[4] Modesti Michele, Lorenzetti Alessandra, Bon D, Besco Stefano. Thermal behaviour of compatibilised polypropylene nanocomposite: Effect of processing conditions. Polymer

Degradation and Stability. 2006;91(4):672-80.

[5] Dennis H. Rayn, Hunter Douglas L., Chang Dohoon, Kim Seokho, White James Lindsay, Cho

Jae Whan, Paul Donald. R. Effect of melt processing conditions on the extent of exfoliation in organoclay-based nanocomposites. Polymer. 2001;42(23):9513-22.

[6] Zhu Yanping, Chen Gao, Zhong Yijun, Zhou Wei, Shao Zongping. Rationally Designed

Hierarchically Structured Tungsten Nitride and Nitrogen-Rich Graphene-Like Carbon

Nanocomposite as Efficient Hydrogen Evolution Electrocatalyst. Advanced Science.

2018;5(2):1700603.

[7] Li Ran, Hou Pengkun, Xie Ning, Ye Zhengmao, Cheng Xin, Shah Surendra P. Design of SiO

2 /PMHS hybrid nanocomposite for surface treatment of cement-based materials. Cement and

Concrete Composites. 2018;87:89-97.

[8] Azeez Asif Abdul, Rhee Kyong Yop, Park Soo Jin, Hui David. Epoxy clay nanocomposites – processing, properties and applications: A review. Composites Part B: Engineering.

2013;45(1):308-20.

[9] Zhao Xin, Lv Lu, Pan Bingcai, Zhang Weiming, Zhang Shujuan, Zhang Quanxing. Polymer- supported nanocomposites for environmental application: A review. Chemical Engineering

Journal. 2011;170(2-3):381-94.

[10] Huang Zheng-Ming, Zhang Yanzhong, Kotaki Masaya, Ramakrishna Seeram. A review on polymer nanofibers by electrospinning and their applications in nanocomposites. Composites

Science and Technology. 2003;63(15):2223-53.

[11] Azizi Samir My Ahmed Said, Alloin Fannie, Dufresne Alain. Review of Recent Research into Cellulosic Whiskers, Their Properties and Their Application in Nanocomposite Field.

Biomacromolecules. 2005;6(2):612-26.

[12] Boccaccini Aldo R., Erol Melek, Stark Wendelin J., Mohn Dirk, Hong Zhongkui, Mano João

F. Polymer/bioactive glass nanocomposites for biomedical applications: A review. Composites

Science and Technology. 2010;70(13):1764-76.

[13] Satarkar Nitin S., Biswal Dipti, Hilt J. Zach. Hydrogel nanocomposites: a review of applications as remote controlled biomaterials. Soft Matter. 2010;6(11):2364.

[14] Zhou You, Yan Bing. A responsive MOF nanocomposite for decoding volatile organic compounds. Chemical Communications. 2016;52(11):2265-8.

[15] Hussain Mian Zahid, Schneemann Andreas, Fischer Roland A., Zhu Yanqiu, Xia Yongde.

MOF Derived Porous ZnO/C Nanocomposites for Efficient Dye Photodegradation. ACS Applied

Energy Materials. 2018;1(9):4695-707.

[16] Zhang Wen-Hui, Zhang Wei-De. Fabrication of SnO2–ZnO nanocomposite sensor for selective sensing of trimethylamine and the freshness of fishes. Sensors and Actuators B:

Chemical. 2008;134(2):403-8.

[17] Hu Ning, Karube Yoshifumi, Yan Cheng, Masuda Zen, Fukunaga Hisao. Tunneling effect in a polymer/carbon nanotube nanocomposite strain sensor. Acta Materialia. 2008;56(13):2929-36.

[18] Amjadi Morteza, Kyung Ki-Uk, Park Inkyu, Sitti Metin. Stretchable, Skin-Mountable, and

Wearable Strain Sensors and Their Potential Applications: A Review. Advanced Functional

Materials. 2016;26(11):1678-98.

[19] Bauhofer Wolfgang, Kovacs Josef Z. A review and analysis of electrical percolation in carbon nanotube polymer composites. Composites Science and Technology. 2009;69(10):1486-

98.

[20] Schmidt Ronald H., Kinloch Ian A., Burgess Andrew N., Windle Alan H. The Effect of

Aggregation on the Electrical Conductivity of Spin-Coated Polymer/Carbon Nanotube Composite

Films. Langmuir. 2007;23(10):5707-12.

[21] Yeh Meng-Kao, Tai Nyan-Hwa, Liu Jia-Hau. Mechanical behavior of phenolic-based composites reinforced with multi-walled carbon nanotubes. Carbon. 2006;44(1):1-9.

[22] Liu Chang, Fang Qichen, Wang Daoyuan, Yan Chao, Liu Faqian, Wang Ning, Guo Zhanhu,

Jiang Qinglong. Carbon and Boron Nitride Nanotubes: Structure, Property and Fabrication. ES

Materials & Manufacturing. 2019;3:2-15.

[23] Yang Shin-Yi, Lin Wei-Ning, Huang Yuan-Li, Tien Hsi-Wen, Wang Jeng-Yu, Ma Chen-Chi

M., Li Shin-Ming, Wang Yu-Sheng. Synergetic effects of graphene platelets and carbon nanotubes on the mechanical and thermal properties of epoxy composites. Carbon.

2011;49(3):793-803.

[24] Cadiff University. HRTEM: Experimental Results. 2019. Access by: 2019.2019.03.09. http://sites.cardiff.ac.uk/hrtem/experimental-results/tem-imaging-with-gatan-digital-camera/

[25] Tao Fangfang, Nysten Bernard, Baudouin Anne-Christine, Thomassin Jean-Michel, Vuluga

Daniela, Detrembleur Christophe, Bailly Christian. Influence of nanoparticle–polymer interactions on the apparent migration behaviour of carbon nanotubes in an immiscible polymer blend. Polymer. 2011;52(21):4798-805.

[26] Chen Jie, Shi Yun-yun, Yang Jing-hui, Zhang Nan, Huang Ting, Chen Chen, Wang Yong,

Zhou Zuo-wan. A simple strategy to achieve very low percolation threshold via the selective

distribution of carbon nanotubes at the interface of polymer blends. Journal of Materials

Chemistry. 2012;22(42):22398.

[27] McQuade D. Tyler, Pullen Anthony E., Swager Timothy M. Conjugated Polymer-Based

Chemical Sensors. Chemical Reviews. 2000;100(7):2537-74.

[28] Swager Timothy M. 50th Anniversary Perspective: Conducting/Semiconducting Conjugated

Polymers. A Personal Perspective on the Past and the Future. Macromolecules.

2017;50(13):4867-86.

[29] Fratoddi Ilaria, Venditti Iole, Cametti Cesare, Russo Maria Vittoria. Chemiresistive polyaniline-based gas sensors: A mini review. Sensors and Actuators B: Chemical.

2015;220:534-48.

[30] Anderson Mark R., Mattes Benjamin R., Reiss Howard, Kaner Richard B. Conjugated

Polymer Films for Gas Separations. Science. 1991;252(5011):1412-5.

[31] Ramaprasad A. T., Rao Vijayalakshmi. Morphology and Miscibility of Chitin-Polyaniline

Blend. Current Science. 2017;112(12):2415.

[32] Tikish Tekalign A., Kumar Ashok, Kim Jung Yong. Study on the Miscibility of Polypyrrole and Polyaniline Polymer Blends. Advances in Materials Science and Engineering. 2018;2018:1-5.

[33] Liu Chang, Lafdi Khalid. Self-Assembly and Surface Tension Induced Fractal Conductive

Network in Ternary Polymer System. ACS Applied Polymer Materials. 2019;1(3):493-9.

[34] Singer Julia C., Ringk Andreas, Giesa Reiner, Schmidt Hans-Werner. Melt Electrospinning of Small Molecules. Macromolecular Materials and Engineering. 2015;300(3):259-76.

[35] Kim Bo-Hye, Wazir Arshad Hussain, Yang Kap-Seung, Bang Yun-Hyuk, Kim Sung-Ryong.

Molecular structure effects of the pitches on preparation of activated carbon fibers from electrospinning. Carbon letters. 2011;12(2):70-80.

[36] Bui Nhu-Ngoc, Kim Bo-Hye, Yang Kap Seung, Dela Cruz Marilou E., Ferraris John P.

Activated carbon fibers from electrospinning of polyacrylonitrile/pitch blends. Carbon.

2009;47(10):2538-9.

[37] Kim Bo-Hye, Bui Nhu-Ngoc, Yang Kap-Seung, Cruz Marilou E. Dela, Ferraris John P.

Electrochemical Properties of Activated Polyacrylonitrile/pitch Carbon Fibers Produced Using

Electrospinning. Bulletin of the Korean Chemical Society. 2009;30(9):1967-72.

[38] Guadagno Liberata, Raimondo Marialuigia, Vittoria Vittoria, Vertuccio Luigi, Lafdi Khalid,

De Vivo Biagio, Lamberti Patrizia, Spinelli Giovanni, Tucci Vincenzo. The role of carbon nanofiber defects on the electrical and mechanical properties of CNF-based resins.

Nanotechnology. 2013;24(30):305704.

[39] Liu Chang, Lafdi Khalid. Fabrication and characterization of carbon nanofibers from polyacrylonitrile/pitch blends. Journal of Applied Polymer Science. 2017;134(42):45388.

[40] Lin Chunfu, Hu Lei, Cheng Chuanbing, Sun Kai, Guo Xingkui, Shao Qian, Li Jianbao,

Wang Ning, Guo Zhanhu. Nano-TiNb2O7/carbon nanotubes composite anode for enhanced lithium-ion storage. Electrochimica Acta. 2018;260:65-72.

[41] Mallakpour Shadpour, khodadadzadeh Leila. Ultrasonic-assisted fabrication of starch/MWCNT-glucose nanocomposites for drug delivery. Ultrasonics Sonochemistry.

2018;40:402-9.

[42] Gorain Bapi, Choudhury Hira, Pandey Manisha, Kesharwani Prashant, Abeer Muhammad

Mustafa, Tekade Rakesh Kumar, Hussain Zahid. Carbon nanotube scaffolds as emerging nanoplatform for myocardial tissue regeneration: A review of recent developments and therapeutic implications. Biomedicine & Pharmacotherapy. 2018;104:496-508.

[43] Greiner Andreas, Wendorff Joachim H. Electrospinning: A Fascinating Method for the

Preparation of Ultrathin Fibers. Angewandte Chemie International Edition. 2007;46(30):5670-

703.

[44] Guellati Ouanassa, Bégin Dominique, Antoni Frederic, Moldovan Simona, Guerioune

Mohamed, Pham-Huu Cuong, Janowska Izabela. CNTs’ array growth using the floating catalyst-

CVD method over different substrates and varying hydrogen supply. Materials Science and

Engineering: B. 2018;231:11-7.

[45] Abusrea Mahmoud R., Han Seung-Wook, Arakawa Kazuo, Choi Nak-Sam. Bending strength of CFRP laminated adhesive joints fabricated by vacuum-assisted resin transfer molding.

Composites Part B: Engineering. 2019;156:8-16.

[46] Shojaeiarani Jamileh, Bajwa Dilpreet S., Stark Nicole M. Spin-coating: A new approach for improving dispersion of cellulose nanocrystals and mechanical properties of poly (lactic acid) composites. Carbohydrate Polymers. 2018;190:139-47.

[47] Hall David B., Underhill Patrick, Torkelson John M. Spin coating of thin and ultrathin polymer films. Polymer Engineering & Science. 1998;38(12):2039-45.

[48] Washo Basil. D. Rheology and Modeling of the Spin Coating Process. IBM Journal of

Research and Development. 1977;21(2):190-8.

[49] Jiang Peng, McFarland Michael J. Large-Scale Fabrication of Wafer-Size Colloidal Crystals,

Macroporous Polymers and Nanocomposites by Spin-Coating. Journal of the American Chemical

Society. 2004;126(42):13778-86.

[50] Matsumoto Hidetoshi, Imaizumi Shinji, Konosu Yuichi, Ashizawa Minoru, Minagawa Mie,

Tanioka Akihiko, Lu Wei, Tour James M. Electrospun Composite Nanofiber Yarns Containing

Oriented Graphene Nanoribbons. ACS Applied Materials & Interfaces. 2013;5(13):6225-31.

[51] Krebs Frederik C. Fabrication and processing of polymer solar cells: A review of printing and coating techniques. Solar Energy Materials and Solar Cells. 2009;93(4):394-412.

[52] Choi Dong Yun, Kang Hyun Wook, Sung Hyung Jin, Kim Sang Soo. Annealing-free, flexible silver nanowire–polymer composite electrodes via a continuous two-step spray-coating method. Nanoscale. 2013;5(3):977-83.

[53] Huang Ping-Kang, Yeh Jien-Wei, Shun Tao-Tsung, Chen Swe-Kai. Multi-Principal-Element

Alloys with Improved Oxidation and Wear Resistance for Thermal Spray Coating. Advanced

Engineering Materials. 2004;6(12):74-8.

[54] Prencipe Giuseppe, Tabakman Scott M., Welsher Kevin, Liu Zhuang, Goodwin Andrew P.,

Zhang Li, Henry Joy, Dai Hongjie. PEG Branched Polymer for Functionalization of

Nanomaterials with Ultralong Blood Circulation. Journal of the American Chemical Society.

2009;131(13):4783-7.

[55] Oh Jung Kwon, Park Jong Myung. Iron oxide-based superparamagnetic polymeric nanomaterials: Design, preparation, and biomedical application. Progress in Polymer Science.

2011;36(1):168-89.

[56] De Vrieze Sander, Van Camp Tamara, Nelvig Anna, Hagström Bengt, Westbroek Phillipe,

De Clerck Karen. The effect of temperature and humidity on electrospinning. Journal of Materials

Science. 2009;44(5):1357-62.

[57] Wu Xiang-Fa, Salkovskiy Yury, Dzenis Yuris A. Modeling of solvent evaporation from polymer jets in electrospinning. Applied Physics Letters. 2011;98(22):223108.

[58] Guo Yongqiang, Xu Genjiu, Yang Xutong, Ruan Kunpeng, Ma Tengbo, Zhang Qiuyu, Gu

Junwei, Wu Yalan, Liu Hu, Guo Zhanhu. Significantly enhanced and precisely modeled thermal conductivity in polyimide nanocomposites with chemically modified graphene via in situ polymerization and electrospinning-hot press technology. Journal of Materials Chemistry C.

2018;6(12):3004-15.

[59] Li Wan-Ju, Mauck Robert L., Cooper James A., Yuan Xiaoning, Tuan Rocky S. Engineering controllable anisotropy in electrospun biodegradable nanofibrous scaffolds for musculoskeletal tissue engineering. Journal of Biomechanics. 2007;40(8):1686-93.

[60] Yu Choongho, Kim Yeon Seok, Kim Dasaroyong, Grunlan Jaime C. Thermoelectric

Behavior of Segregated-Network Polymer Nanocomposites. Nano Letters. 2008;8(12):4428-32.

[61] Du Fangming, Scogna Robert C., Zhou Wei, Brand Stijn, Fischer John E., Winey Karen I.

Nanotube Networks in Polymer Nanocomposites: Rheology and Electrical Conductivity.

Macromolecules. 2004;37(24):9048-55.

[62] Li Jing, Kim Jang-Kyo. Percolation threshold of conducting polymer composites containing

3D randomly distributed graphite nanoplatelets. Composites Science and Technology.

2007;67(10):2114-20.

[63] Farrar Charles R., Worden Keith. An introduction to structural health monitoring.

Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering

Sciences. 2007;365(1851):303-15.

[64] Ko Jinming, Ni Yiqing. Technology developments in structural health monitoring of large- scale bridges. Engineering Structures. 2005;27(12):1715-25.

[65] Diamanti K., Soutis Constantinos. Structural health monitoring techniques for aircraft composite structures. Progress in Aerospace Sciences. 2010;46(8):342-52.

[66] Rosania Colleen L., Chang Fu-Kuo. Damage Diagnosis in Composite Materials under

Applied Load with Guided Waves-based SHM. In Structural Health Monitoring 2017. Standford

University. Conference Year. DOI: 10.12783/shm2017/3886.

[67] García-Macías Enrique, D'Alessandro Antonella, Castro-Triguero Rafael, Pérez-Mira

Domingo, Ubertini Filippo. Micromechanics modeling of the uniaxial strain-sensing property of carbon nanotube cement-matrix composites for SHM applications. Composite Structures.

2017;163:195-215.

[68] Chen Tao, He Yuting, Du Jinqiang. A High-Sensitivity Flexible Eddy Current Array Sensor for Crack Monitoring of Welded Structures under Varying Environment. Sensors.

2018;18(6):1780.

[69] Pierdicca Alessio, Clementi Francesco, Maracci Diletta, Isidori Daniela, Lenci Stefano.

Vibration-Based SHM of Ordinary Buildings: Detection and Quantification of Structural

Damage. In ASME Proceedings | 27th Conference on Mechanical Vibration and Noise.

Conference Year. V008T13A98.

[70] Wu Wen-Hwa, Wang Sheng-Wei, Chen Chien-Chou, Lai Gwolong. Assessment of environmental and nondestructive earthquake effects on modal parameters of an office building based on long-term vibration measurements. Smart Materials and Structures. 2017;26(5):055034.

[71] Liu Chang, Sergeichev Ivan, Akhatov Iskander, Lafdi Khalid. CNT and polyaniline based sensors for the detection of acid penetration in polymer composite. Composites Science and

Technology. 2018;159:111-8.

[72] Gotou Tomohiro, Katagiri Nobutatsu, Sakai Tetsuya, Kubouchi Masatoshi, Tsuda Ken.

Detection of environmental acid penetrated in FRP using optical fiber. In: Proceedings of the 16th

International Conference on Composite Materials. Kyoto, Conference 2007.07.09, Conference

2007. p. MoBA1-01.

[73] Ding Rui, Wu Hongchao, Thunga Mahendra, Bowler Nicola, Kessler Michael R. Processing and characterization of low-cost electrospun carbon fibers from organosolv lignin/polyacrylonitrile blends. Carbon. 2016;100:126-36.

[74] Yang Ying, Le TrungHieu, Kang Feiyu, Inagaki Michio. Polymer blend techniques for designing carbon materials. Carbon. 2017;111:546-68.

[75] Zhou Zhengping, Liu Kunming, Lai Chuilin, Zhang Lifeng, Li Juanhua, Hou Haoqing,

Reneker Darrell H., Fong Hao. Graphitic carbon nanofibers developed from bundles of aligned electrospun polyacrylonitrile nanofibers containing phosphoric acid. Polymer. 2010;51(11):2360-

[76] Liu H. Clive, Chien An-Ting, Newcomb Bradley A., Liu Yaodong, Kumar Satish.

Processing, Structure, and Properties of Lignin- and CNT-Incorporated Polyacrylonitrile-Based

Carbon Fibers. ACS Sustainable Chemistry & Engineering. 2015;3(9):1943-54.

[77] Newcomb Bradley A., Giannuzzi Lucille A., Lyons Kevin M., Gulgunje Prabhakar V.,

Gupta Kishor, Liu Yaodong, Kamath Manjeshwar, McDonald Kenneth, Moon Jaeyun, Feng Bo,

Peterson G. P., Chae Han Gi, Kumar Satish. High resolution transmission electron microscopy study on polyacrylonitrile/carbon nanotube based carbon fibers and the effect of structure development on the thermal and electrical conductivities. Carbon. 2015;93:502-14.

[78] Ju Young-Wan, Choi Gyoung-Rin, Jung Hong-Ryun, Lee Wan-Jin. Electrochemical properties of electrospun PAN/MWCNT carbon nanofibers electrodes coated with polypyrrole.

Electrochimica Acta. 2008;53(19):5796-803.

[79] Chae Han Gi, Sreekumar T. V., Uchida Tetsuya, Kumar Satish. A comparison of reinforcement efficiency of various types of carbon nanotubes in polyacrylonitrile fiber. Polymer.

2005;46(24):10925-35.

[80] Richard-Lacroix Marie, Pellerin Christian. Molecular Orientation in Electrospun Fibers:

From Mats to Single Fibers. Macromolecules. 2013;46(24):9473-93.

[81] Andrady Anthony L. Science and Technology of Polymer Nanofibers. Hoboken, New Jersy:

John Wiley & Sons, Inc.; 2008.

[82] McKee Matthew G., Layman John M., Cashion Matthew P., Long Timothy E. Phospholipid

Nonwoven Electrospun Membranes. Science. 2006;311(5759):353-5.

[83] Singh Gurvinder, Bittner Alexander M., Loscher Sebastian, Malinowski Nikola, Kern Klaus.

Electrospinning of Diphenylalanine Nanotubes. Advanced Materials. 2008;20(12):2332-6.

[84] de Gans Berend-Jan, Duineveld Paul. C, Schubert Ulrich. S. Inkjet Printing of Polymers:

State of the Art and Future Developments. Advanced Materials. 2004;16(3):203-13.

[85] Fong Hao., Chun Iksoo., Reneker Darrell. H. Beaded nanofibers formed during electrospinning. Polymer. 1999;40(16):4585-92.

[86] Murugesan Murali, Zandén Carl, Luo Xin, Ye Lilei, Jokubavicius Valdas, Syväjärvi Mikael,

Liu Johan. A carbon fiber solder matrix composite for thermal management of microelectronic devices. J Mater Chem C. 2014;2(35):7184-7.

[87] Yan Han, Mahanta Nayandeep K., Wang Bojie, Wang Shanshan, Abramson Alexis R.,

Cakmak Miko. Structural evolution in graphitization of nanofibers and mats from electrospun polyimide–mesophase pitch blends. Carbon. 2014;71:303-18.

[88] Shi Zhiqiang, Chong Chuanbin, Wang Jing, Wang Chengyang, Yu Xuewen. Electrospun pitch/polyacrylonitrile composite carbon nanofibers as high performance anodes for lithium-ion batteries. Materials Letters. 2015;159:341-4.

[89] Huang Cchaobo., Chen Shuiliang., Reneker Darrell. H, Lai Chuilin., Hou Haoqing. High-

Strength Mats from Electrospun Poly(p-Phenylene Biphenyltetracarboximide) Nanofibers.

Advanced Materials. 2006;18(5):668-71.

[90] Su Run, Zhong Ganji, Fu Qiang, Zhang Lifeng, Fong Hao, Zhu Lei. Polarity-induced ferroelectric crystalline phase in electrospun fibers of poly(vinylidene fluoride)/polyacrylonitrile blends. Journal of Materials Research. 2012;27(10):1389-98.

[91] Kalra Vibha, Kakad Prashant A., Mendez Sergio, Ivannikov Timur, Kamperman Marleen,

Joo Yong Lak. Self-Assembled Structures in Electrospun Poly(styrene-block-isoprene) Fibers.

Macromolecules. 2006;39(16):5453-7.

[92] Shao Shijun, Zhou Shaobing, Li Long, Li Jinrong, Luo Chao, Wang Jianxin, Li Xiaohong,

Weng Jie. Osteoblast function on electrically conductive electrospun PLA/MWCNTs nanofibers.

Biomaterials. 2011;32(11):2821-33.

[93] Song Kunlin, Wu Qinglin, Zhang Zhen, Ren Suxia, Lei Tingzhou, Negulescu Ioan I., Zhang

Quanguo. Porous Carbon Nanofibers from Electrospun Biomass Tar/Polyacrylonitrile/Silver

Hybrids as Antimicrobial Materials. ACS Applied Materials & Interfaces. 2015;7(27):15108-16.

[94] Cao Li, Hou Yanwen, Lafdi Khalid, Urmey Kirk. Fluorescent composite scaffolds made of nanodiamonds/polycaprolactone. Chemical Physics Letters. 2015;641:123-8.

[95] Elias-Birembaux Lama Hélène, Fenouillot Francoise., Majeste Jean-Charles, Cassagnau

Philippe. Morphology and rheology of immiscible polymer blends filled with silica nanoparticles.

Polymer. 2007;48(20):6029-40.

[96] Yang Zhiyi, Han Chang Dae. Rheology of Miscible Polymer Blends with Hydrogen

Bonding. Macromolecules. 2008;41(6):2104-18.

[97] Yu Wei, Wu Youjun, Yu Ruobing, Zhou Chixing. Dynamic rheology of the immiscible blends of liquid crystalline polymers and flexible chain polymers. Rheologica Acta.

2005;45(2):105-15.

[98] Li Chun-Zhu, Wu Fan, Cai Hai-Yong, Kandiyoti Rafael. UV-Fluorescence Spectroscopy of

Coal Pyrolysis Tars. Energy & Fuels. 1994;8(5):1039-48.

[99] Gargiulo Valentina, Apicella Barbara, Alfè Michela, Russo Carmela, Stanzione Fernando,

Tregrossi Antonio, Amoresano Angela, Millan Marcos, Ciajolo Anna. Structural Characterization of Large Polycyclic Aromatic Hydrocarbons. Part 1: The Case of Coal Tar Pitch and

Naphthalene-Derived Pitch. Energy & Fuels. 2015;29(9):5714-22.

[100] Dickinson Eric M. Average structures of petroleum pitch fractions by 1H13C n.m.r. spectroscopy. Fuel. 1985;64(5):704-6.

[101] Luo Chaojie, Stride Eleanor., Edirisinghe Mohan. Mapping the Influence of Solubility and

Dielectric Constant on Electrospinning Polycaprolactone Solutions. Macromolecules.

2012;45(11):4669-80.

[102] Sakai Mototsugu, Kida Tohoru, Inagaki Michio. Creep and glass transition behaviour of fractionated petroleum pitches: the influence of molecular weight and its distribution. Journal of

Materials Science. 1984;19(8):2651-63.

[103] Zeng Jian-Bing, Zhu Qun-Ying, Lu Xi, He Yi-Song, Wang Yu-Zhong. From miscible to partially miscible biodegradable double crystalline poly(ethylene succinate)-b-poly(butylene succinate) multiblock copolymers. Polym Chem. 2012;3(2):399-408.

[104] Dalton Stephen, Heatley Frank, Budd Peter M. Thermal stabilization of polyacrylonitrile fibres. Polymer. 1999;40(20):5531-43.

[105] Lin Rong-Hsien. In situ FTIR and DSC investigation on cure reaction of liquid aromatic dicyanate ester with different types of epoxy resin. Journal of Polymer Science Part A: Polymer

Chemistry. 2000;38(16):2934-44.

[106] Goh Suat Hong, Lee Swee Yong, Yeo Y. T., Zhou Xu, Tan Kuang Lee. Miscibility and specific interactions in polyacrylonitrile/ poly(p-vinylphenol) blends. Macromolecular Rapid

Communications. 1999;20(3):148-51.

[107] Yeo Y. T., Goh Suat Hong, Lee Swee Yong. Miscibility of Polyacrylonitrile with Tertiary

Amide Polymers. Journal of Macromolecular Science, Part A. 1997;34(4):597-603.

[108] Bashir Zahir, Church Stephen P., Waldron Dennis. Interaction of water and hydrated crystallization in water-plasticized polyacrylonitrile films. Polymer. 1994;35(5):967-76.

[109] Oroumei Azam, Fox Bronwyn, Naebe Minoo. Thermal and Rheological Characteristics of

Biobased Carbon Fiber Precursor Derived from Low Molecular Weight Organosolv Lignin. ACS

Sustainable Chemistry & Engineering. 2015;3(4):758-69.

[110] Jawhari T., Roid A., Casado J. Raman spectroscopic characterization of some commercially available carbon black materials. Carbon. 1995;33(11):1561-5.

[111] Arshad Salman N., Naraghi Mohammad, Chasiotis Ioannis. Strong carbon nanofibers from electrospun polyacrylonitrile. Carbon. 2011;49(5):1710-9.

[112] Son Donghee, Kang Jiheong, Vardoulis Orestis, Kim Yeongin, Matsuhisa Naoji, Oh Jin

Young, To John W. F., Mun Jaewan, Katsumata Toru, Liu Yuxin, McGuire Allister F., Krason

Marta, Molina-Lopez Francisco, Ham Jooyeun, Kraft Ulrike, Lee Yeongjun, Yun Youngjun, Tok

Jeffrey B. H., Bao Zhenan. An integrated self-healable electronic skin system fabricated via dynamic reconstruction of a nanostructured conducting network. Nature Nanotechnology.

2018;13(11):1057-65.

[113] Du Yuzhang, Ge Juan, Li Yannan, Ma Peter X., Lei Bo. Biomimetic elastomeric, conductive and biodegradable polycitrate-based nanocomposites for guiding myogenic differentiation and skeletal muscle regeneration. Biomaterials. 2018;157:40-50.

[114] Chizari Kambiz, Arjmand Mohammad, Liu Zhe, Sundararaj Uttandaraman, Therriault

Daniel. Three-dimensional printing of highly conductive polymer nanocomposites for EMI shielding applications. Materials Today Communications. 2017;11:112-8.

[115] Pang Huan, Xu Ling, Yan Ding-Xiang, Li Zhong-Ming. Conductive polymer composites with segregated structures. Progress in Polymer Science. 2014;39(11):1908-33.

[116] Li Heng, Minus Marilyn L. On the formation of potential polymer-nanotube blends by liquid-solid phase separation. Polymer. 2017;131:179-92.

[117] Amjadi Morteza, Pichitpajongkit Aekachan, Lee Sangjun, Ryu Seunghwa, Park Inkyu.

Highly Stretchable and Sensitive Strain Sensor Based on Silver Nanowire–Elastomer

Nanocomposite. ACS Nano. 2014;8(5):5154-63.

[118] Al Habis Nuha, Liu Chang, Dumuids Jean-Baptiste, Lafdi Khalid. Intelligent design of conducting network in polymers using numerical and experimental approaches. RSC Advances.

2016;6(97):95010-20.

[119] Al-Ajrash Saja M. Nabat, Lafdi Khalid, Vasquez Erick S., Chinesta Francisco, Le

Coustumer Philippe. Experimental and Numerical Investigation of the Silicon Particle

Distribution in Electrospun Nanofibers. Langmuir. 2018;34(24):7147-52.

[120] Hermanson Kevin D., Lumsdon Simon O., Williams Jacob P., Kaler Eric W., Velev Orlin

D. Dielectrophoretic Assembly of Electrically Functional Microwires from Nanoparticle

Suspensions. Science. 2001;294(5544):1082-6.

[121] Pashayi Kamyar, Fard Hafez Raeisi, Lai Fengyuan, Iruvanti Sushumna, Plawsky Joel,

Borca-Tasciuc Theodorian. Self-constructed tree-shape high thermal conductivity nanosilver networks in epoxy. Nanoscale. 2014;6(8):4292-6.

[122] Wang Jun, Reyna-Valencia Alejandra, Favis Basil D. Assembling Conductive PEBA

Copolymer at the Continuous Interface in Ternary Polymer Systems: Morphology and Resistivity.

Macromolecules. 2016;49(14):5115-25.

[123] Zhang Liying, Cui Tingting, Cao Xiao, Zhao Chengji, Chen Quan, Wu Lixin, Li Haolong.

Inorganic-Macroion-Induced Formation of Bicontinuous Block Copolymer Nanocomposites with

Enhanced Conductivity and Modulus. Angewandte Chemie International Edition.

2017;56(31):9013-7.

[124] Chopade Sujay A., Au Jesus G., Li Ziang, Schmidt Peter W., Hillmyer Marc A., Lodge

Timothy P. Robust Polymer Electrolyte Membranes with High Ambient-Temperature Lithium-

Ion Conductivity via Polymerization-Induced Microphase Separation. ACS Applied Materials &

Interfaces. 2017;9(17):14561-5.

[125] Haase Martin F., Sharifi-Mood Nima, Lee Daeyeon, Stebe Kathleen J. In Situ Mechanical

Testing of Nanostructured Bijel Fibers. ACS Nano. 2016;10(6):6338-44.

[126] Stejskal Jaroslav, Spirkova Milena, Riede Andrea, Helmstedt Martin, Mokreva Pavlina,

Prokes Jan. Polyaniline dispersions 8. The control of particle morphology. Polymer.

1999;40(10):2487-92.

[127] Yu Pingping, Li Yingzhi, Zhao Xin, Wu Lihao, Zhang Qinghua. Graphene-Wrapped

Polyaniline Nanowire Arrays on Nitrogen-Doped Carbon Fabric as Novel Flexible Hybrid

Electrode Materials for High-Performance Supercapacitor. Langmuir. 2014;30(18):5306-13.

[128] Zhao Yibo, Wei Huige, Arowo Moses, Yan Xingru, Wu Wei, Chen Jianfeng, Wang Yiran,

Guo Zhanhu. Electrochemical energy storage by polyaniline nanofibers: high gravity assisted oxidative polymerization vs. rapid mixing chemical oxidative polymerization. Physical Chemistry

Chemical Physics. 2015;17(2):1498-502.

[129] Baker Christina O., Huang Xinwei, Nelson Wyatt, Kaner Richard B. Polyaniline nanofibers: broadening applications for conducting polymers. Chemical Society Reviews.

2017;46(5):1510-25.

[130] Zhang Miao, Liu Lin, Ju Xiaohua, He Teng, Chen Ping. Molten salt assisted synthesis of microporous polyaniline nanosheets with superior gas sorption properties. Microporous and

Mesoporous Materials. 2018;267:100-6.

[131] Ma Yong, Hou Chunping, Zhang Hao, Qiao Mingtao, Chen Yanhui, Zhang Hepeng, Zhang

Qiuyu, Guo Zhanhu. Morphology-dependent electrochemical supercapacitors in multi- dimensional polyaniline nanostructures. Journal of Materials Chemistry A. 2017;5(27):14041-52.

[132] Pan Wei, Yang Shenglin, Li Guang, Jiang Jianming. Relationship Between Electrical

Conductivity And Phase Morphology Of Polyaniline/polyacrylonitrile And

Polyaniline/polystyrene Blends. International Journal of Polymeric Materials and Polymeric

Biomaterials. 2006;54(1):21-35.

[133] Dudem Bhaskar, Mule Anki Reddy, Patnam Harishkumar Reddy, Yu Jae Su. Wearable and durable triboelectric nanogenerators via polyaniline coated cotton textiles as a movement sensor and self-powered system. Nano Energy. 2019;55:305-15.

[134] Kumar Vanish, Mahajan Rashmi, Kaur Inderpreet, Kim Ki-Hyun. Simple and Mediator-

Free Urea Sensing Based on Engineered Nanodiamonds with Polyaniline Nanofibers Synthesized in situ. ACS Applied Materials & Interfaces. 2017;9(20):16813-23.

[135] Miao Zhuling, Wang Peiyu, Zhong AiMing, Yang Minfeng, Xu Qin, Hao Shirong, Hu

Xiaoya. Development of a glucose biosensor based on electrodeposited gold nanoparticles– polyvinylpyrrolidone–polyaniline nanocomposites. Journal of Electroanalytical Chemistry.

2015;756:153-60.

[136] Wang Liping, Vivek Raju, Wu Weifeng, Wang Guowu, Wang Jin-Ye. Fabrication of Stable and Well-Dispersed Polyaniline–Polypyrrolidone Nanocomposite for Effective Photothermal

Therapy. ACS Biomaterials Science & Engineering. 2018:1880–90.

[137] Wu James Chun-Cheng, Ray Sudip, Gizdavic-Nikolaidis Marija, Jin Jianyong, Cooney

Ralph P. Effect of polyvinylpyrrolidone on storage stability, anti-oxidative and anti-bacterial properties of colloidal polyaniline. Synthetic Metals. 2016;217:202-9.

[138] Tiitu Mari, Talo Anja, Forsén Olof, Ikkala Olli. Aminic epoxy resin hardeners as reactive solvents for conjugated polymers: polyaniline base/epoxy composites for anticorrosion coatings.

Polymer. 2005;46(18):6855-61.

[139] Karabanova Lyudmyla V., Mikhalovska Sergey V., Sergeeva Lyudmila M., Meikle Steve,

Helias Michael, Lloyd Andrew William. Semi-interpenetrating polymer networks based on

polyurethane and poly(vinyl pyrrolidone) obtained by photopolymerization: Structure-property relationships and bacterial adhesion. Polymer Engineering and Science. 2004;44(5):940-7.

[140] Shin Suyong, Gu Ming-Long, Yu Chin-Yang, Jeon Jongseol, Lee Eunji, Choi Tae-Lim.

Polymer Self-Assembly into Unique Fractal Nanostructures in Solution by a One-Shot Synthetic

Procedure. Journal of the American Chemical Society. 2017;140(1):475-82.

[141] Wang Yangyong, Jing Xinli, Kong Junhua. Polyaniline nanofibers prepared with hydrogen peroxide as oxidant. Synthetic Metals. 2007;157(6-7):269-75.

[142] Nandan Bhanu, Chen Hsin-Lung, Liao Chien-Shiun, Chen Show-An. Self-Assembly and

Crystallization in a Supramolecular Hairy Rod Polymer from the Complex of Polyaniline with ω-

Methoxy Poly(ethylene oxide) Phosphates. Macromolecules. 2004;37(25):9561-70.

[143] Stejskal Jaroslav, Kratochvíl Pavel, Helmstedt Martin. Polyaniline Dispersions. 5.

Poly(vinyl alcohol) and Poly(N-vinylpyrrolidone) as Steric Stabilizers. Langmuir.

1996;12(14):3389-92.

[144] Boeva Zhanna A., Sergeyev Vladimir. G. Polyaniline: Synthesis, properties, and application. Polymer Science Series C. 2014;56(1):144-53.

[145] Chattopadhyay Dipankar, Banerjee Surjo, Chakravorty Dipankar, Mandal Broja M.

Ethyl(hydroxyethyl)cellulose Stabilized Polyaniline Dispersions and Destabilized Nanoparticles

Therefrom. Langmuir. 1998;14(7):1544-7.

[146] Langbein Dieter. Van der Waals attraction between cylinders, rods or fibers. Physik der

Kondensierten Materie. 1972;15(1):61-86.

[147] Sheehan Paul E., Lieber Charles M. Friction between van der Waals Solids during Lattice

Directed Sliding. Nano Letters. 2017;17(7):4116-21.

[148] Choi Jihoon, Cargnello Matteo, Murray Christopher B., Clarke Nigel, Winey Karen I.,

Composto Russell J. Fast Nanorod Diffusion through Entangled Polymer Melts. ACS Macro

Letters. 2015;4(9):952-6.

[149] Yoshikawa Hitoshi, Hino Tetsuo, Kuramoto Noriyuki. Effect of temperature and moisture on electrical conductivity in polyaniline/polyurethane (PANI/PU) blends. Synthetic Metals.

2006;156(18-20):1187-93.

[150] Biswas Sourav, Panja Sujit S., Bose Suryasarathi. Tailored distribution of nanoparticles in bi-phasic polymeric blends as emerging materials for suppressing electromagnetic radiation: challenges and prospects. Journal of Materials Chemistry C. 2018;6(13):3120-42.

[151] Artero-Guerrero Jose A., Pernas-Sánchez Jesus, López-Puente Jorge, Varas David. On the influence of filling level in CFRP aircraft fuel tank subjected to high velocity impacts. Composite

Structures. 2014;107:570-7.

[152] Eesaee Mostafa, Shojaei Akbar. Effect of nanoclays on the mechanical properties and durability of novolac phenolic resin/woven glass fiber composite at various chemical environments. Composites Part A: Applied Science and Manufacturing. 2014;63:149-58.

[153] Ma Chen-Chi M., Lee Chang-Lun, Tai Nyan-Hwa. Chemical resistance of carbon fiber- reinforced poly(ether ether ketone) and poly(phenylene sulfide) composites. Polymer

Composites. 1992;13(6):435-40.

[154] Wong Kok, Rudd Chris, Pickering Steve, Liu XiaoLing. Composites recycling solutions for the aviation industry. Science China Technological Sciences. 2017;60(9):1291-300.

[155] Lindgren Mari, Wallin Markus J., Kakkonen Markus, Saarela Olli J., Vuorinen Jyrki. The influence of high-temperature sulfuric acid solution ageing on the properties of laminated vinyl- ester joints. International Journal of Adhesion and Adhesives. 2016;68:298-304.

[156] Fiore Vincenzo, Calabrese Luigi, Proverbio Edoardo, Passari Rocco, Valenza Antonino.

Salt spray fog ageing of hybrid composite/metal rivet joints for automotive applications.

Composites Part B: Engineering. 2017;108:65-74.

[157] Solis-Ramos Euripides, Kumosa Maciej. Synergistic effects in stress corrosion cracking of glass reinforced polymer composites. Polymer Degradation and Stability. 2017;136:146-57.

[158] Polimeno Umberto, Meo Michele. Detecting barely visible impact damage detection on aircraft composites structures. Composite Structures. 2009;91(4):398-402.

[159] Gentry Russell, Bank Lawrence C., Barkatt Aaron, Prian Luca. Accelerated Test Methods to Determine the Long-Term Behavior of Composite Highway Structures Subject to

Environmental Loading. Journal of Composites Technology and Research. 1998;20(1):38.

[160] Nazemi Mohammad Karim, Valix Marjorie. Evaluation of acid diffusion behaviour of amine-cured epoxy coatings by accelerated permeation testing method and prediction of their service life. Progress in Organic Coatings. 2016;97:307-12.

[161] Tallman Tyler N., Wang Kon-well. Damage and strain identification in multifunctional materials via electrical impedance tomography with constrained sine wave solutions. Structural

Health Monitoring: An International Journal. 2016;15(2):235-44.

[162] Seppänen Aku, Hallaji Milad, Pour-Ghaz Mohammad. A functionally layered sensing skin for the detection of corrosive elements and cracking. Structural Health Monitoring: An

International Journal. 2016;16(2):215-24.

[163] Fujii Yoshimichi, Ramakrishna Seeram, Hamada Hiroyuki. Estimation of Durability of

GFRP Laminates Under Stress-Corrosive Environments Using Acoustic Emission. In:

Proceedings of ISTM international. West Conshohocken, PA, Conference, Conference 1996. p.

190-203.

[164] Romhány Gábor, Czigány Tibor, Karger-Kocsis Jozsef. Failure Assessment and Evaluation of Damage Development and Crack Growth in Polymer Composites Via Localization of Acoustic

Emission Events: A Review. Polymer Reviews. 2017;57(3):397-439.

[165] Gotou Tomohiro, Noda Masashi, Tomiyama Tomonori, Sembokuya Hideki, Kubouchi

Masatoshi, Tsuda Ken. In situ health monitoring of corrosion resistant polymers exposed to alkaline solutions using pH indicators. Sensors and Actuators B: Chemical. 2006;119(1):27-32.

100

[166] Wang Jun, GangaRao Hota, Liang Ruifeng, Liu Weiqing. Durability and prediction models of fiber-reinforced polymer composites under various environmental conditions: A critical review. Journal of Reinforced Plastics and Composites. 2015;35(3):179-211.

[167] Kawada Hiroyuki, Srivastava Varsha Kumar. The effect of an acidic stress environment on the stress-intensity factor for GRP laminates. Composites Science and Technology.

2001;61(8):1109-14.

[168] Kumosa Lucas S., Armentrout Daniel, Kumosa Maciej. An evaluation of the critical conditions for the initiation of stress corrosion cracking in unidirectional E-glass/polymer composites. Composites Science and Technology. 2001;61(4):615-23.

[169] Abanilla Maria Araceli, Li Yan, Karbhari Vistasp M. Durability characterization of wet layup graphite/epoxy composites used in external strengthening. Composites Part B: Engineering.

2005;37(2-3):200-12.

[170] Tanks Jonathon D., Sharp Stephen R., Harris Devin K. Kinetics of in-plane shear degradation in carbon/epoxy rods from exposure to alkaline and saline environments. Composites

Part B: Engineering. 2017;110:204-12.

[171] Megel Marc, Kumosa Lucas S., Ely Thomas, Armentrout Daniel, Kumosa Maciej.

Initiation of stress-corrosion cracking in unidirectional glass/polymer composite materials.

Composites Science and Technology. 2001;61(2):231-46.

[172] Hogg Paul. J., Hull Derek. Micromechanisms of crack growth in composite materials under corrosive environments. Metal Science. 2013;14(8-9):441-9.

[173] Hogg Paul. J. Factors affecting the stress corrosion of GRP in acid environments.

Composites. 1983;14(3):254-61.

[174] Hogg Paul. J. A model for stress corrosion crack growth in glass reinforced plastics.

Composites Science and Technology. 1990;38(1):23-42.

[175] Kasehagen Leo J., Haury Idelette, Macosko Christopher W., Shimp David A. Hydrolysis and blistering of cyanate ester networks. Journal of Applied Polymer Science. 1997;64(1):107-13.

101

[176] Sembokuya Hideki, Negishi Yoshiyuki, Kubouchi Masatoshi, Tsuda Ken. Corrosion

Behavior of Epoxy Resin Cured with Different Amount of Hardener in Corrosive Solutions.

Journal of the Society of Materials Science, Japan. 2003;52(9Appendix):230-4.

[177] Crank John. The Mathematics Of Diffusion. Oxford: Clarendon Press; 1975.

[178] Liu Hu, Huang Wenju, Gao Jiachen, Dai Kun, Zheng Guoqiang, Liu Chuntai, Shen

Changyu, Yan Xingru, Guo Jiang, Guo Zhanhu. Piezoresistive behavior of porous carbon nanotube-thermoplastic polyurethane conductive nanocomposites with ultrahigh compressibility.

Applied Physics Letters. 2016;108(1):011904.

[179] Wang Caifeng, Zhao Min, Li Jun, Yu Jiali, Sun Shaofan, Ge Shengsong, Guo Xingkui, Xie

Fei, Jiang Bo, Wujcik Evan K., Huang Yudong, Wang Ning, Guo Zhanhu. Silver nanoparticles/graphene oxide decorated carbon fiber synergistic reinforcement in epoxy-based composites. Polymer. 2017;131:263-71.

[180] Sun Kai, Xie Peitao, Wang Zhongyang, Su Tongming, Shao Qian, Ryu JongEun, Zhang

Xihua, Guo Jiang, Shankar Akash, Li Jianfeng, Fan Runhua, Cao Dapeng, Guo Zhanhu. Flexible polydimethylsiloxane/multi-walled carbon nanotubes membranous metacomposites with negative permittivity. Polymer. 2017;125:50-7.

[181] Cheng Chuanbing, Fan Runhua, Wang Zhongyang, Shao Qian, Guo Xingkui, Xie Peitao,

Yin Yansheng, Zhang Yuliang, An Liqiong, Lei Yanhua, Ryu Jong Eun, Shankar Akash, Guo

Zhanhu. Tunable and weakly negative permittivity in carbon/silicon nitride composites with different carbonizing temperatures. Carbon. 2017;125:103-12.

[182] Zhao Jinbo, Wu Lili, Zhan Chuanxing, Shao Qian, Guo Zhanhu, Zhang Liqun. Overview of polymer nanocomposites: Computer simulation understanding of physical properties. Polymer.

2017;133:272-87.

[183] Liu Hu, Dong Mengyao, Huang Wenju, Gao Jiachen, Dai Kun, Guo Jiang, Zheng

Guoqiang, Liu Chuntai, Shen Changyu, Guo Zhanhu. Lightweight conductive

102

graphene/thermoplastic polyurethane foams with ultrahigh compressibility for piezoresistive sensing. Journal of Materials Chemistry C. 2017;5(1):73-83.

[184] Liu Hu, Huang Wenju, Yang Xinru, Dai Kun, Zheng Guoqiang, Liu Chuntai, Shen

Changyu, Yan Xingru, Guo Jiang, Guo Zhanhu. Organic vapor sensing behaviors of conductive thermoplastic polyurethane–graphene nanocomposites. Journal of Materials Chemistry C.

2016;4(20):4459-69.

[185] Liu Hu, Li Yilong, Dai Kun, Zheng Guoqiang, Liu Chuntai, Shen Changyu, Yan Xingru,

Guo Jiang, Guo Zhanhu. Electrically conductive thermoplastic elastomer nanocomposites at ultralow graphene loading levels for strain sensor applications. Journal of Materials Chemistry C.

2016;4(1):157-66.

[186] Liu Hu, Gao Jiachen, Huang Wenju, Dai Kun, Zheng Guoqiang, Liu Chuntai, Shen

Changyu, Yan Xingru, Guo Jiang, Guo Zhanhu. Electrically conductive strain sensing polyurethane nanocomposites with synergistic carbon nanotubes and graphene bifillers.

Nanoscale. 2016;8(26):12977-89.

[187] Starkova Olesja A., Buschhorn Samuel. T., Mannov Evgenij, Schulte Karl, Aniskevich

Andrey. Water transport in epoxy/MWCNT composites. European Polymer Journal.

2013;49(8):2138-48.

[188] Brethous R., Colin Xavier, Fayolle Bruno, Gervais M. Non-Fickian behavior of water absorption in an epoxy-amidoamine network. 2016;1736:020070.

[189] Minelli Matteo, Baschetti Marco Giacinti, Doghieri Ferruccio. Analysis of modeling results for barrier properties in ordered nanocomposite systems. Journal of Membrane Science.

2009;327(1-2):208-15.

[190] Shokrieh Mahmood, Nasir Vahid, Karimipour Hadis. A micromechanical study on longitudinal strength of fibrous composites exposed to acidic environment. Materials & Design.

2012;35:394-403.

103

[191] Tzou Koutai, Gregory Richard. A method to prepare soluble polyaniline salt solutions — in situ doping of PANI base with organic dopants in polar solvents. Synthetic Metals.

1993;53(3):365-77.

[192] Stejskal Jaroslav, Sapurina Irina. Polyaniline: Thin films and colloidal dispersions (IUPAC

Technical Report). Pure and Applied Chemistry. 2005;77(5):815-26.

[193] Olthuis Wouter, Streekstra Wim, Bergveld Piet. Theoretical and experimental determination of cell constants of planar-interdigitated electrolyte conductivity sensors. Sensors and Actuators B: Chemical. 1995;24(1-3):252-6.

[194] Ping Zhao. In situ FTIR–attenuated total reflection spectroscopic investigations on the base–acid transitions of polyaniline. Base–acid transition in the emeraldine form of polyaniline. J

Chem Soc, Faraday Trans. 1996;92(17):3063-7.

[195] Ivanova Marianna V, Lamprecht Constanze, Loureiro M Jimena, Huzil J Torin, Foldvari

Marianna. Pharmaceutical characterization of solid and dispersed carbon nanotubes as nanoexcipients. International Journal of Nanomedicine. 2012:403.

[196] Bao Weishun, Meguid Shaker. A., Zhu Zhenghong, Weng Geroge. J. Tunneling resistance and its effect on the electrical conductivity of carbon nanotube nanocomposites. Journal of

Applied Physics. 2012;111(9):093726.

[197] Fernandes Pedro, Sena-Cruz José, Xavier José, Silva Patrícia, Pereira Eduardo, Cruz José.

Durability of bond in NSM CFRP-concrete systems under different environmental conditions.

Composites Part B: Engineering. 2018;138:19-34.

[198] Lynch Jerome P., Gupta Sumit, Loh Kenneth J. Noncontact tomography and a pH-sensitive nanocomposite for monitoring osseointegrated prosthesis interfaces. 2017;10168:101681K.

[199] Zhang Kai, Li Gen-Hui, Feng La-Mei, Wang Ning, Guo Jiang, Sun Kai, Yu Kai-Xin, Zeng

Jian-Bing, Li Tingxi, Guo Zhanhu, Wang Ming. Ultralow percolation threshold and enhanced electromagnetic interference shielding in poly(l-lactide)/multi-walled carbon nanotube

104

nanocomposites with electrically conductive segregated networks. Journal of Materials Chemistry

C. 2017;5(36):9359-69.

[200] Huang Jiangnan, Cao Yonghai, Shao Qian, Peng Xiangfang, Guo Zhanhu. Magnetic

Nanocarbon Adsorbents with Enhanced Hexavalent Chromium Removal: Morphology

Dependence of Fibrillar vs Particulate Structures. Industrial & Engineering Chemistry Research.

2017;56(38):10689-701.

[201] Ma Yanli, Lv Ling, Guo Yuanru, Fu Yujie, Shao Qian, Wu Tingting, Guo Sijie, Sun Kai,

Guo Xingkui, Wujcik Evan K., Guo Zhanhu. Porous lignin based poly (acrylic acid)/organo- montmorillonite nanocomposites: Swelling behaviors and rapid removal of Pb (II) ions. Polymer.

2017;128:12-23.

[202] Zhao Min, Meng Linghui, Ma Lichun, Ma Lina, Yang Xiaobing, Huang Yudong, Ryu Jong

Eun, Shankar Akash, Li Tingxi, Yan Chao, Guo Zhanhu. Layer-by-layer grafting CNTs onto carbon fibers surface for enhancing the interfacial properties of epoxy resin composites.

Composites Science and Technology. 2018;154:28-36.

[203] Liu Tao, Yu Kun, Gao Lina, Chen Hui, Wang Ning, Hao Luhan, Li Tingxi, He Hongcai,

Guo Zhanhu. A graphene quantum dot decorated SrRuO3 mesoporous film as an efficient counter electrode for high-performance dye-sensitized solar cells. Journal of Materials Chemistry A.

2017;5(34):17848-55.

[204] Li Yahong, Zhou Bing, Zheng Guoqiang, Liu Xianhu, Li Tingxi, Yan Chao, Cheng

Chuanbing, Dai Kun, Liu Chuntai, Shen Changyu, Guo Zhanhu. Continuously prepared highly conductive and stretchable SWNT/MWNT synergistically composited electrospun thermoplastic polyurethane yarns for wearable sensing. Journal of Materials Chemistry C. 2018;6(9):2258-69.

[205] Ma Peng-Cheng, Zhang Yi. Perspectives of carbon nanotubes/polymer nanocomposites for wind blade materials. Renewable and Sustainable Energy Reviews. 2014;30:651-60.

105

[206] Rashetnia Reza, Hallaji Milad, Smyl Danny, Seppänen Aku, Pour-Ghaz Mohammad.

Detection and localization of changes in two-dimensional temperature distributions by electrical resistance tomography. Smart Materials and Structures. 2017;26(11):115021.

[207] Prolongo Silvia Gonzalez, Gude Maria Rodriguez, Ureña Alejandro. Water uptake of epoxy composites reinforced with carbon nanofillers. Composites Part A: Applied Science and

Manufacturing. 2012;43(12):2169-75.

[208] Lu Tianyi, Solis-Ramos Euripides, Yi Yunbo, Kumosa Maciej. Particle removal mechanisms in synergistic aging of polymers and glass reinforced polymer composites under combined UV and water. Composites Science and Technology. 2017;153:273-81.

[209] Tanks Jonathon D., Arao Yoshihiko, Kubouchi Masatoshi. Diffusion kinetics, swelling, and degradation of corrosion-resistant C-glass/epoxy woven composites in harsh environments.

Composite Structures. 2018;202:686-94.

[210] Zhou Haili, Li Chao, Zhang Liquan, Crawford Bryn, Milani Abbas S., Ko Frank K. Micro-

XCT analysis of damage mechanisms in 3D circular braided composite tubes under transverse impact. Composites Science and Technology. 2018;155:91-9.

[211] Glud Jens Ammitzboll, Carraro Paolo A., Quaresimin Marino, Dulieu-Barton Janice. M.,

Thomsen Ole Thomsen, Overgaard Lars Christian. A damage-based model for mixed-mode crack propagation in composite laminates. Composites Part A: Applied Science and Manufacturing.

2018;107:421-31.

[212] Luder Daniel, Ariely Shmuel, Yalin Moti. Stress corrosion cracking and brittle failure in a fiber-reinforced plastic (FRP) insulator from a 400 kV transmission line in humid environment.

Engineering Failure Analysis. 2019;95:206-13.

[213] Kusano Masahiro, Kanai Takafumi, Arao Yoshihiko, Kubouchi Masatoshi. Degradation behavior and lifetime estimation of fiber reinforced plastics tanks for hydrochloric acid storage.

Engineering Failure Analysis. 2017;79:971-9.

106

[214] Naebe Minoo, Abolhasani Mohammad Mahdi, Khayyam Hamid, Amini Abbas, Fox

Bronwyn. Crack Damage in Polymers and Composites: A Review. Polymer Reviews.

2016;56(1):31-69.

[215] Wang Guantao, Wang Yong, Zhang Peipei, Zhai Yujiang, Luo Yun, Li Liuhe, Luo Sida.

Structure dependent properties of carbon nanomaterials enabled fiber sensors for in situ monitoring of composites. Composite Structures. 2018;195:36-44.

[216] Tuloup Corentin, Harizi Walid, Aboura Zoheir, Meyer Yann, Kamel Khellil, Lachat Remy.

On the use of in-situ piezoelectric sensors for the manufacturing and structural health monitoring of polymer-matrix composites: A literature review. Composite Structures. 2019;215:127-49.

[217] Todd Michael, Gregory William, Key Christopher, Yeager Michael, Ye Jordan. Composite

Laminate Fatigue Damage Detection and Prognosis Using Embedded Fiber Bragg Gratings.

ASME Proceedings, Structural Health Monitoring. 2018:V002T05A11.

[218] Tian Zhenhua, Yu Lingyu, Sun Xiaoyi, Lin Bin. Damage localization with fiber Bragg grating Lamb wave sensing through adaptive phased array imaging. Structural Health

Monitoring. 2018;18(1):334-44.

[219] Liu Chang, Lafdi Khalid. Environmental Monitoring of Composite Durability Use Multiple

Sensing Technologies. In: Proceedings of CAMX 2018 – The Composites and Advanced

Materials Expo. Dallas, Conference, Conference 2018. p. TP18-0586.

[220] Saharudin Mohd Shahneel, Atif Rasheed, Shyha Islam, Inam Fawad. The degradation of mechanical properties in polymer nano-composites exposed to liquid media – a review. RSC

Advances. 2016;6(2):1076-89.

[221] Li Xuyang, Bandyopadhyay Parthasarathi, Kshetri Tolendra, Kim Nam Hoon, Lee Joong

Hee. Novel hydroxylated boron nitride functionalized p-phenylenediamine-grafted graphene: an excellent filler for enhancing the barrier properties of polyurethane. Journal of Materials

Chemistry A. 2018;6(43):21501-15.

107

[222] Peretz Damari Sivan, Cullari Lucas, Nadiv Roey, Nir Yiftach, Laredo Dalia, Grunlan

Jaime, Regev Oren. Graphene-induced enhancement of water vapor barrier in polymer nanocomposites. Composites Part B: Engineering. 2018;134:218-24.

[223] Kharissova Oxana V., Kharisov Boris I. Variations of interlayer spacing in carbon nanotubes. RSC Adv. 2014;4(58):30807-15.

[224] Pulidindi Indra Neel. Development and exploitation of carbon materials from plant sources.

Chennai, India: Indian Institute of Technology Madras; 2009. 312.

[225] Author, RFID (RC522) simple signal strength detection via changing antenna gain.

Mendeley. Arduino code. 2018. 10.17632/n6gn3k4dkn.1. https://data.mendeley.com/datasets/n6gn3k4dkn/1

108

APPENDIX A

Arduino-RFID-RC522-LCD Connections

I2C LCD adaptor

Arduino UNO

RC522

109

APPENDIX B

Arduino Code for Wireless Sensing

#include

constexpr uint8_t RST_PIN=9; constexpr uint8_t SS_PIN=10;

MFRC522 mfrc522(SS_PIN, RST_PIN);

LiquidCrystal_I2C lcd(0x27,20,4); // set the LCD address to 0x27 for a 16 chars and 2 line display

void setup() {

Serial.begin(9600);

while (!Serial);

SPI.begin();

mfrc522.PCD_Init();

mfrc522.PCD_DumpVersionToSerial();

Serial.println(F("Scan PICC to see UID, SAK, Type, data blocks and signal level"));

lcd.init(); // initialize the lcd

lcd.backlight();

lcd.setCursor(1,0);

}

110

void loop() {

mfrc522.PCD_SetAntennaGain(0x01<<4);

if (mfrc522.PICC_IsNewCardPresent() == 1) {

lcd.setCursor(0,0);

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 1 ");

return;

}

else

mfrc522.PCD_SetAntennaGain(0x02<<4);

if (mfrc522.PICC_IsNewCardPresent() == 1){

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 2 ");

return;

}

else

mfrc522.PCD_SetAntennaGain(0x03<<4);

if (mfrc522.PICC_IsNewCardPresent() == 1){

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 3 ");

return;

}

else

mfrc522.PCD_SetAntennaGain(0x04<<4);

if (mfrc522.PICC_IsNewCardPresent() == 1){

111

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 4 ");

return;

}

else

mfrc522.PCD_SetAntennaGain(0x05<<4); if (mfrc522.PICC_IsNewCardPresent() == 1){

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 5 ");

return;

}

else

mfrc522.PCD_SetAntennaGain(0x06<<4); if (mfrc522.PICC_IsNewCardPresent() == 1){

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 6 ");

return;

}

else

mfrc522.PCD_SetAntennaGain(0x07<<4); if (mfrc522.PICC_IsNewCardPresent() == 1){

lcd.print(mfrc522.PCD_GetAntennaGain());

Serial.println("Level 7 ");

return;

}

else

112

lcd.print("N/A");

Serial.println("N/A ");

return;

}

113